Acta Orthopaedica issue 3 / 2020 by Acta

3/20

* Sprowson AP et al. Bone Joint J 2016; 98-B: 1534–1541

www.heraeus-medical.com

90th anniversary

Vol. 91, No. 3, 2020 (pp. 221–363)

Bone cement with gentamicin ttamicin i i and clindamycin

ACTA ORTHOPAEDICA

69 %

reduction of deep infections in hip hemiarthroplasty after fractured neck of femur *

Volume 91, Number 3, June 2020

TF-IORT200080.indd 1

27-05-2020 19:07:17

Acta Orthopaedica is owned by the Nordic Orthopaedic Federation and is the official publication of the Nordic Orthopaedic Federation

E DITORIAL O F FICE

Acta Orthopaedica Department of Orthopedics Lund University Hospital SE–221 85 Lund, Sweden E-mail: acta.ort@med.lu.se Homepage: http://www.actaorthop.org

EDITOR

THE FOUNDATION BOARD OF

Anders Rydholm Lund, Sweden

THE NORDIC O RTHOPAEDIC F EDERATION AND A CTA O RTHOPAEDICA

DEPUTY EDITOR

Peter A Frandsen Odense, Denmark CO-EDITORS

Li Felländer-Tsai Stockholm, Sweden Nils Hailer Uppsala, Sweden Ivan Hvid Oslo, Norway Urban Rydholm Lund, Sweden Bart A Swierstra Wageningen, The Netherlands Eivind Witsø Trondheim, Norway Rolf Önnerfält Lund, Sweden

Peter Frandsen Denmark Ragnar Jonsson Iceland Heikki Kröger Finland Anders Rydholm Sweden Kees Verheyen the Netherlands

WEB EDITOR

Magnus Tägil Lund, Sweden S TATISTICAL EDITOR

Jonas Ranstam Lund, Sweden P RODUCTION MANAGER

Kaj Knutson Lund, Sweden

Vol. 91, No. 3, 2020

SUBSCRIPTION INFORMATION Acta Orthopaedica [print 1745-3674, online 1745-3682] is a peerreviewed journal, published six times a year plus supplements by Taylor & Francis on behalf of Nordic Orthopaedic Federation.

Airfreight and mailing in the USA by agent named WN Shipping USA, 156-15, 146th Avenue, 2nd Floor, Jamaica, NY 11434, USA. Periodicals postage paid at Jamaica NY 11431.

Annual Institutional Subscription, Volume 91, 2020

US Postmaster: Send address changes to Acta Orthopaedica, WN Shipping USA, 156-15, 146th Avenue, 2nd Floor, Jamaica, NY 11434, USA.

$1,291

£798

€1,035

The subscription fee purchases an online subscription. The price includes access to current content and back issues to January 1997 (if available). Printed copies of the journal are provided on request as a free supplementary service accompanying an online subscription. Supplements to the journal are also included in the subscription price. For more information, visit the journal’s website: http://www.tandfonline.com/IORT Manuscripts should be uploaded at http://www.manuscriptmanager.com/ao/ for further handling at: Acta Orthopaedica Editorial Office, Department of Orthopaedics, Lund University Hospital, SE-221 85 Lund, Sweden Correspondance concerning copyright and permissions should be sent to: Maria Montzka, Portfolio Manager – Medicine P.O. Box 3255, SE-103 65 Stockholm, Sweden, Tel: +46 (0)760 14 24 68. Fax: +46 (0)8 440 80 50. E-mail: maria.montzka@informa.com Ordering information: Please contact your local Customer Service Department to take out a subscription to the Journal: USA, Canada: Taylor & Francis, Inc., 530 Walnut Street, Suite 850, Philadelphia, PA 19106, USA. Tel: +1 800 354 1420; Fax: +1 215 207 0050. UK/ Europe/Rest of World: T&F Customer Services, Informa UK Ltd, Sheepen Place, Colchester, Essex, CO3 3LP, United Kingdom. Tel: +44 (0) 20 7017 5544; Fax: +44 (0) 20 7017 5198; Email: subscriptions@tandf.co.uk Dollar rates apply to all subscribers outside of Europe. Euro rates apply to all subscribers in Europe except the UK and Republic of Ireland. If you are unsure which applies, contact Customer Services. All subscriptions are payable in advance and all rates include postage. Journals are sent by air to the USA, Canada, Mexico, India, Japan and Australasia. Subscriptions are entered on an annual basis, i.e., January to December. Payment may be made by sterling check, US dollar check, euro check, international money order, National Giro, or credit card (Amex, Visa and Mastercard). Back issues: Taylor & Francis retains a two-year back issue stock of journals. Older volumes are held by our official stockists to whom all orders and enquiries should be addressed: Periodicals Service Company, 351 Fairview Ave., Suite 300, Hudson, New York 12534, USA. Tel: +1 518 537 4700; fax: +1 518 537 5899; e-mail: psc@ periodicals.com.

Subscription records are maintained at Taylor & Francis Group, 4 Park Square, Milton Park, Abingdon, OX14 4RN, United Kingdom.

Copyright © 2020 The Author(s). Published by Taylor & Francis on behalf of the Nordic Orthopedic Federation. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Non-Commercial License (https://creativecommons.org/licenses/by-nc/3.0 . Informa UK Limited, trading as Taylor & Francis Group makes every effort to ensure the accuracy of all the information (the “Content”) contained in its publications. However, Informa UK Limited, trading as Taylor & Francis Group, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Informa UK Limited, trading as Taylor & Francis Group. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Informa UK Limited, trading as Taylor & Francis Group shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. Terms & Conditions of access and use can be found at http://www.tandfonline. com/page/terms-and-conditions Indexed/abstracted in: Allied and Complementary Medicine Library (Amed); ASCA (Automatic Subject Citation Alert); Biological Abstracts; Chemical Abstracts; Cumulative Index to Nursing and Allied Health Literature(CINAHL); Current Advances in Ecological and Environmental Sciences; Current Contents/Clinical Medicine; Current Contents/Life Sciences; Developmental Medicine and Child Neurology; Energy Research Abstracts; EMBASE/ Excerpta Medica; Faxon Finder; Focus On: Sports Science & Medicine; Health Planning and Administration; Index Medicus/MEDLINE; Index to Dental Literature; Index Veterinarius; INIS Atomindex; Medical Documentation Service; Nuclear Science Abstracts (Ceased); Periodicals Scanned and Abstracted. Life Sciences Collection; Research Alert; Science Citation Index; SciSearch; SportSearch; Uncover Veterinary Bulletin. Printed in England by Henry Ling

Acta Orthopaedica

ISSN 1745-3674

Vol. 91, No. 3, June 2020 Editorial Pandemic pressure: policy, politics, profession, and rapid publication Acta Orthopaedica 90-year anniversary Guest editorials – Acta Orthopaedica 90-year anniversary LIA in arthroplasty – the history of a single-center observational study leading to implementation in general clinical practice Enhanced Recovery after arthroplasty surgery Uncemented science at its best! How registry data can improve outcomes from joint replacement – a seminal paper Reflections on the RSA guidelines Should length of stay in hospital be the endpoint in arthroplasty? Understanding fracture populations by epidemiology A blueprint for global registries from Norway Hip, knee and revision hip replacement – are they as clinically and cost effective as we think? Patient satisfaction after total knee replacement—still a challenge Guest editorial ERAS guidelines for hip and knee replacement – need for reanalysis of evidence and recommendations ? Hip Inferior stabilization of cementless compared with cemented dualmobility cups in elderly osteoarthrosis patients: a randomized controlled radiostereometry study on 60 patients with 2 years’ follow-up 2-year results of an RCT of 2 uncemented isoelastic monoblock acetabular components: lower wear rate with vitamin E blended highly cross-linked polyethylene compared to ultra-high molecular weight polyethylene Low-dose CT-based implant motion analysis is a precise tool for early migration measurements of hip cups: a clinical study of 24 patients Greater early migration of a short-stem total hip arthroplasty is not associated with an increased risk of osseointegration failure: 5th-year results from a prospective RSA study with 39 patients, a follow-up study How do EQ-5D-3L and EQ-5D-5L compare in a Swedish total hip replacement population? Increased risk for dislocation after introduction of the Continuum cup system: lessons learnt from a cohort of 1,381 THRs after 1-year follow-up Postoperative 30-day complications after cemented/hybrid versus cementless total hip arthroplasty in osteoarthritis patients > 70 years: A multicenter study from the Lundbeck Foundation Centre for Fast-track Hip and Knee replacement database and the Danish Hip Arthroplasty Register The role of bone cement for the development of intraoperative hypotension and hypoxia and its impact on mortality in hemiarthroplasty for femoral neck fractures 14-year hip survivorship after periacetabular osteotomy: a follow-up study on 1,385 hips Hip and knee Influence of fast-track programs on patient-reported outcomes in total hip and knee replacement (THR/TKR) at Swedish hospitals 2011–2015: an observational study including 51,169 THR and 8,393 TKR operations Prognostic factors for inpatient functional recovery following total hip and knee arthroplasty: a systematic review

221 222

L Felländer-Tsai A Rydholm

224

H Kehlet

226 228 230

R G H H Nelissen A Lübbeke K Tucker

232 234 236 238 240

B G Pijls L Nordsletten A P Launonen S Graves A Blom

241

Jan Verhaar

243

H Kehlet and S G Memtsoudis

246

S Tabori-Jensen, S B Mosegaard, T B Hansen, and M Stilling

254

J H J van Erp, J R A Massier, J J Halma, T E Snijders, and A de Gast

260

C Brodén, O sandberg, O Sköldenberg, H Stigbrand, M Hänni, J W Giles, R Emery, S Lazarinis, A Nyström, and H Olivecrona T Floerkemeier, S Budde, G v. Lewinski, H Windhagen, C Hurschler, and M Schwarze

266

272

T Eneqvist, S Nemes, J Kärrholm, K Burström, and O Rolfson

279

O A Pakarinen, P S Neuvonen, A R P Reito, and A P Eskelinen

286

M Lindberg-Larsen, P B Petersen, C C Jørgensen, S Overgaard, and H Kehlet + Lundbeck Foundation Center for Fast-track Hip and Knee Arthroplasty Collaborating Group

293

F Olsen, M Hård af Segerstad, B Nellgård, E Houltz, and S-E Ricksten

299

J B Larsen, I Mechlenburg, S S Jakobsen, T M Thilleman, and K Søballe

306

U Berg, A W-Dahl, O Rolfson, E Nauclér, M Sundberg, and A Nilsdotter

313

N Hewlett-Smith, R Pope, J Furness, V Simas, and W Hing

Shoulder Patients undergoing shoulder arthroplasty for failed nonoperative treatment of proximal humerus fracture have low implant survival and low patient-reported outcomes: 837 cases from the Danish Shoulder Arthroplasty Registry Hand Percutaneous needle fasciotomy in Dupuytren contracture: a registerbased, observational cohort study on complications in 3,331 treated fingers in 2,257 patients Ankle and foot Reduced incidence and economic cost of hardware removal after ankle fracture surgery: a 20-year nationwide registry study Mid-term results of hindfoot arthrodesis with a retrograde intramedullary nail in 24 patients with diabetic Charcot neuroarthropathy Oncology Modified Harrington’s procedure for periacetabular metastases in 89 cases: a reliable method for cancer patients with good functional outcome, especially with long expected survival Infection Surgical site infection after hip fracture – mortality and risk factors: an observational cohort study of 1,709 patients Single-stage debridement with implantation of antibiotic-loaded calcium sulphate in 34 cases of localized calcaneal osteomyelitis Correspondence Postoperative mortality after a hip fracture over a 15-year period in Denmark: a national register study Errata Deep learning in fracture detection: a narrative review (DOI 10.1080/17453674.2019.1711323) Consensus statement for perioperative care in total hip replacement and total knee replacement surgery: Enhanced Recovery After Surgery (ERAS®) Society recommendations (DOI 10.1080/17453674.2019.1683790) Information to authors (see http://www.actaorthop.org/)

319

I Mechlenburg, S Rasmussen, D Unbehaun, A Amundsen, and J V Rasmussen

326

L H Therkelsen, S T Skov, M Laursen, and J Lange

331

N Partio, T T Huttunen, H M Mäenpää, and V M Mattila

336

M Ersin, M Demirel, M Chodza, F Bilgili, and O I Kiliçoglu

341

G Kask, J Nieminen, V van Iterson, M Naboistsikov, T-K Pakarinen, and M K Laitinen

347 353

C T Pollmann, F A Dahl, J H M Røtterud, J-E Gjertsen, and A Årøen N Jiang, X-q Zhao, L Wang, Q-r Lin, Y-j Hu, and B Yu

360

C Rogmark versus O Gundel

362

P H S Kalmet, S Sanduleanu, S Primakov, G Wu, A Jochems, T Refaee, A Ibrahim, L v. Hulst, P Lambin, and M Poeze T W Wainwright et al.

363

Acta Orthopaedica 2020; 91 (3): 221

221

Editorial

Pandemic pressure: policy, politics, profession, and rapid publication

Health care is under siege. Covid-19 has caused a disaster where health care needs by far transcend available resources at the rip curl of the pandemic. Infectious diseases have wreaked havoc before. Lessons have been learnt and sometimes forgotten, despite the sufferings and tragedies of patients and heroic tasks by the healthcare workforce. While the virus is raging, the economy is in free fall with an unprecedented impact on politics and society. In many ways it has taken the fear of the current pandemic to remind us that many infectious diseases were once so common and deadly, with no choice but to accept the toll. Death by pandemic was dreadful and depressing yet a natural part of life in the world until just some decades ago. But previously polio, malaria, yellow fever, HIV, and cholera have impacted the history of many nations though at a lower speed, apart from, more recently, Ebola and SARS. New dangers have been unleashed by globalization and rapid travel. Recent and saddening datasets featuring the deaths of medical staff have been published after the initial early reports from China. Not only has ethical and moral stress from having to prioritize scarce resources soared, but also the horror-filled insight that front-line doctors and staff are most at risk. The understanding of the threat that Covid-19 poses to doctors and healthcare staff has initially lagged behind but is now fueling the debate and controversy on state-of-the-art personal protective equipment (PPE) in light of limited supplies. With worries regarding PPE, insurance issues for sick and dead doctors and healthcare staff have also surfaced. Treatment of severely ill Covid-19 patients has mandated major reorganization of health care, upscaling emergency medicine and trauma while elective surgery is down-prioritized and put on hold. Managing the backlog will create repercussions for a substantial duration. As Covid-19 is spread mainly through airborne transmission, unprotected first-line medical staff and also anesthetists intubating Covid-19 patients are thought to be at particular risk. There is also ongoing debate on the presence of airborne virus transmission in orthopedic surgery caused by power tools and operating room ventilation systems potentially causing viral wind tunnels. The exit strategies from society lockdowns run in parallel with the global hunt for coronavirus drugs and a vaccine. A promising way of identifying candidate drugs and vaccine can-

didates is by crunching huge amounts of data and many artificial intelligence companies are now providing their services to scientists. Open sharing of data will be necessary to move forward at speed. And it will be a challenge for pharmaceutical companies to balance commercial versus moral aspects while traversing borders in a collective effort to combat the Covid-19 pandemic. Rapid media reports and cases of disinformation add to the desperation. Covid-19 has distorted markets, alliances, and policies. Politicians have to seek guidance for decisions while opinions are brokered by an array of experts (including selfappointed ones), government agencies, and non-government organizations in search of stable evidence. World leaders have not always shown signs of cooperation in public-health efforts to fight Covid-19 and the thin line to a blame-game is unfortunately sometimes crossed. Warnings of the risks of a pandemic have in many instances not been considered when performing risk analyses and planning for supplies and infrastructure. Being in frightening new territory alongside anxiety and fear, it gives solace to remember the huge medical advancements that vaccines and new drugs have enabled over time. During the wait we must acknowledge the selfless contributions by all professionals in medicine, nursing, allied health, research, and also volunteers fighting the war against Covid-19. This is well framed by the transcendent sentence:“ To cure sometimes, to relieve often, to comfort always.” Rapid dissemination of new knowledge about coronavirus is important; major scientific journals have introduced fasttrack publication of clinical features of the disease, potential drugs and vaccines, effectiveness of preventive measures such as quarantine and even isolation of entire countries. ACTA will rapidly electronically publish online relevant orthopedic information related to Covid-19 within an accelerated editorial and review process; please submit material online, category “Corona” (ManuscriptManager). April 5, 2020 Li Felländer-Tsai, Co-Editor

© 2020 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group, on behalf of the Nordic Orthopedic Federation. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. DOI 10.1080/17453674.2020.1753162

222

Acta Orthopaedica 2020; 91 (3): 222–223

Editorial

Acta Orthopaedica 90-year anniversary Acta Orthopaedica, a nonprofit, Open Access journal, was founded in 1930 and is owned by the Nordic Orthopaedic Federation (NOF). The NOF is the federation of the national orthopedic societies of Denmark, Estonia, Finland, Iceland, Lithuania, The Netherlands, Norway, and Sweden representing over 5,500 orthopedic surgeons. Acta enables the mission of NOF, i.e., dissemination of evidence-based medicine, education and research. Approximately 600 manuscripts are submitted to Acta each year, and of these around 100 are published after editorial and outside/external peer review. Articles come from all over the world but submissions and publications from the NOF countries are most common. Acta publishes numerous arthroplasty register studies—the first registers (Knee 1975 and Hip 1979, were started in Sweden)—and radiostereometry investigations (developed for orthopedic purposes in Sweden in 1974). Several of these register and radiostereometry studies are highly cited: see the top-10 cited articles below. During recent years Acta has also published an increasing number of randomized clinical trials. This issue of Acta celebrates our 90-year anniversary with this overview of Acta’s development during the last 20 years and a compilation of the 10 most cited research articles during the last 20 years, each of them accompanied by a Guest editorial written by invited experts. Acta 2000–2020 The Open Access (OA) movement started in the 1990s and gained momentum in the early 2000s when PubMed Central (PMC) and OA BioMed Central were launched. Acta soon realized that OA had come to stay and in 2005 started, first of orthopedic journals, OA publication without cost for authors and readers. At this time we also made all Acta articles published since start 1930 electronically accessible cost-free. When we introduced OA we changed the journal’s name from Acta Orthopaedica Scandinavica to Acta Orthopaedica to emphasize Acta’s international character. We foresaw that with OA we would lose subscriptions and in 2018 we introduced an Article Processing Charge (APC), which is common to most other journals that offer OA publication. APC relates to articles emanating from outside NOF. All Nordic orthopedic surgeons are members of NOF and pay a subscription for Acta, which means that they pay no APC. At the same time, we also stopped the print edition and

Acta became an electronic-only journal. Accepted articles are, however, still compiled into 6 annual issues (each comprising around 20 original articles) presented as an e-zine and distributed to all collective NOF member subscribers. All Acta articles are, since 2005, distributed under the terms of the Creative Commons Attribution Non-commercial Licence. This means that Acta has no copyright on your article; you as an author, and others, are free to use it for all noncommercial purposes. In 2009 Acta started online prepublication; soon after acceptance the article is available on PubMed (Epub ahead of print) eliminating the previous delay caused by the wait for the print publication. In the same year Acta also started voluntary nonanonymous reviewing (also, the names of authors are open to the reviewers); the reviewers are named and thanked in the published article. There is an option for reviewers to remain anonymous, which is, however, rarely used: almost all > 200 annual reviewers sign their reports. We have a strong feeling this open system has improved the quality of the reviews, an observation also made by other journals. Inappropriate use of statistics in biomedicine (including misuse/misunderstanding of p-values) leading to false conclusions has become increasingly apparent and discussed during the 2000s. It has been named as one of the explanations for the so-called replication crisis within science. Already in 1993 Acta included an experienced biostatistician among its editors, since then scrutinizing statistics in all manuscripts. He has also written many Editorials on the proper use of statistical methods and helped numerous authors to improve their manuscripts, thus implying well-founded conclusions. Several authors have expressed their appreciation for this help. In 2019 Acta decided to publish Registered Reports, i.e., the methods and proposed analyses of planned studies are published as a journal article after peer-review has confirmed a meaningful study applying appropriate methods. Once the study is completed—and adheres to the initially approved proposal—publication in Acta is guaranteed irrespective of the study outcome. We also decided to consider for publication manuscripts previously uploaded to a non-commercial preprint server, such as MedRxiv. We do not consider posting on a preprint server to be duplicate publication and this will not jeopardize consideration for publication in Acta after the ordinary review process.

Acta Orthopaedica 2020; 91 (3): 222–223

Topics to be considered during 2020 • Make the review process transparent, i.e., make the complete preprint correspondence between authors, editors, and external reviewers electronically available together with the published article (a system already applied by some journals). • Encourage publication of studies with (unexpected) null results (to avoid publication bias). • Encourage publication of replication studies (cf. current biomedicine with a serious replication problem: the reproducibility crisis). There is a need for more systematic replication studies, regardless of their outcome. This applies also to clinical studies; in some way register studies of joint replacements are already close to replication studies—the same implants are studied in different populations managed at different hospitals.

223

Top-10 articles 2000–2020 These 10 articles, published since 2000, were ranked (from 2,721 articles) by the fraction ”Total number of cites/Years from publication”, to compensate for increasing possibilities of citation with time. The articles were cited around > 400– 200 times. The abstract of each article is reproduced and each of them is accompanied by a Guest editorial written by invited experts. Acta thanks them for their excellent work.

Anders Rydholm on behalf of all editors

Top-10 articles, 2000–2020 in descending order of number of citations Local infiltration analgesia: a technique for the control of acute postoperative pain following knee and hip surgery. A case study of 325 patients. Dennis R Kerr and Lawrence Kohan. Acta Orthop 2008; 79(2): 174-83. Predictors of length of stay and patient satisfaction after hip and knee replacement surgery: Fast-track experience in 712 patients. Henrik Husted, Gitte Holm, and Steffen Jacobsen. Acta Orthop 2008; 79(2): 168-73. Uncemented and cemented primary total hip arthroplasty in the Swedish Hip Arthroplasty Register: evaluation of 170,413 operations. Nils P Hailer, Göran Garellick, and Johan Kärrholm. Acta Orthop 2010; 81(1): 34-41. Long-term registration has improved the quality of hip replacement: a review of the Swedish THR Register comparing 160,000 cases. Peter Herberts and Henrik Malchau. Acta Orthop Scand 2000; 71(2): 111-21. Guidelines for standardization of radiostereometry (RSA) of implants. Edward R Valstar, Richie Gill, Leif Ryd, Gunnar Flivik, Niclas Börlin, and Johan Kärrholm. Acta Orthop 2005; 76(4): 563-72. Why still in hospital after fast-track hip and knee arthroplasty? Henrik Husted, Troels H Lunn, Anders Troelsen, Lissi Gaarn-Larsen, Billy B Kristensen, and Henrik Kehlet. Acta Orthop 2011; 82(6): 679-84. The epidemiology of proximal humeral fractures. Charles M Court-Brown, Ashima Garg and Margaret M McQueen. Acta Orthop Scand 2001; 72(4): 365–71. The Norwegian Arthroplasty Register: 11 years and 73,000 arthroplasties. Leif I Havelin, Lars B Engesaeter, Birgitte Espehaug, Ove Furnes, Stein A Lie, and Stein E Vollset. Acta Orthop Scand 2000; 71(4): 337-53. Effectiveness of hip or knee replacement surgery in terms of quality-adjusted life years and costs. Pirjo Räsänen, Pekka Paavolainen, Harri Sintonen, Anna-Maija Koivisto, Marja Blom, Olli-Pekka Ryynänen, and Risto P Roine. Acta Orthop 2007; 78(1): 108-15. Patient satisfaction after knee arthroplasty: a report on 27,372 knees operated on between 1981 and 1995 in Sweden. Otto Robertsson, Michael Dunbar, Thorbjörn Pehrsson, Kaj Knutson, and Lars Lidgren. Acta Orthop Scand 2000; 71(3): 262–7.

224

Acta Orthopaedica 2020; 91 (3): 224–225

Local infiltration analgesia: a technique for the control of acute postoperative pain following knee and hip surgery A case study of 325 patients Dennis R Kerr and Lawrence Kohan Acta Orthopaedica 2008; 79 (2): 174–183 DOI 10.1080/17453670710014950

Background We have developed a multimodal technique for the control of pain following knee and hip surgery, called “local infiltration analgesia” (LIA). It is based on systematic infiltration of a mixture of ropivacaine, ketorolac, and adrenaline into the tissues around the surgical field to achieve satisfactory pain control with little physiological disturbance. The technique allows virtually immediate mobilization and earlier discharge from hospital. Patients and methods In this open, nonrandomized case series, we used LIA to manage postoperative pain in all 325 patients presenting to our service from Jan 1, 2005 to Dec 31, 2006 for elective hip resurfacing (HRA), primary total hip replacement (THR), or primary total knee replacement arthroplasty (TKR). We recorded pain scores, mobilization times, and morphine usage for the entire group.

Results Pain control was generally satisfactory (numerical rating scale pain score range 0–3). No morphine was required for postoperative pain control in two-thirds of the patients. Most patients were able to walk with assistance between 5 and 6 h after surgery and independent mobility was achieved 13–22 h after surgery. Orthostatic hypotension, nausea, and vomiting were occasionally associated with standing for the first time, but other side effects were unremarkable. 230 (71%) of the 325 patients were discharged directly home after a single overnight stay in hospital. Interpretation Local infiltration analgesia is simple, practical, safe, and effective for pain management after knee and hip surgery. ■

Guest editorial

LIA in arthroplasty – the history of a single-center observational study leading to implementation in general clinical practice Although local anesthetic wound infiltration has been widely used for more than 50 years in different surgical procedures, the observational study by Kerr and Kohan (2008) in 325 patients undergoing 3 different hip and knee arthroplasties led to an increased attention to the very simple, surgeon-administrated technique of high-volume local anesthetic wound infiltration (LIA), with claimed effects on pain, mobilization, opioid use and hospital stay. This study was a non-RCT performed by 2 physicians in a small private hospital in Sydney, factors that usually argue against the value of such reports. Nevertheless, in this case the article further stimulated others to do original and more scientific research including RCT’s and later followed by several systematic reviews. However, these reviews unfortunately often suffered from methodological problems, since the efficacy of LIA (with NSAID) was not compared to appropriate control groups including systemic NSAIDs and paracetamol (Andersen et al. 2014). Also, although the original suggestion that ketorolac and epinephrine should be added in the solution in some way was rational, it was not proven in later studies to be necessary or essential compared to systemic

NSAIDs/Cox 2 inhibitors (Spreng et al. 2010, Schotanus et al. 2017). Whatsoever, history has shown that the observational study by Kerr and Kohan (2008) stimulated much research to argue that the technique is simple, safe and effective in TKA, but apparently not in THA (Andersen et al. 2014, Tan et al. 2019, Wainwright et al. 2020). The reasons for this discrepancy between TKA and THA still needs to be clarified but may be related to wound size and the fact that pain is less following a THA compared with TKA. Concomitant with the availability of more scientific data on LIA, the concept of fast-track THA and TKA was undergoing much progress. Since optimal patient management is a prerequisite for enhanced recovery and early mobilization, LIA most probably may have played an important role in this process, although a direct association is difficult to prove. The suggestion that further efficacy can be obtained by a wound catheter is controversial (Andersen et al. 2014, Wainwright et al. 2020) although this author personally observed an effect in a patient the day after a knee replacement in

Acta Orthopaedica 2020; 91 (3): 224–225

the orthopedic hospital in Sydney when visiting Kerr and Kohan. Although safe, simple and effective, however, there is still a demand to take the time to do the infiltration meticulously to cover all important wound layers and probably also to do further studies on which tissue layers and places to be most important to infiltrate to obtain sufficient analgesia (Andersen et al. 2014). In this context, more information is needed on the additional analgesic effect of performing a surgeon-administered “blind” saphenous block when being inside the joint. Finally, although many studies are available, we need more data from comparing or combining LIA with different peripheral nerve blocks (Zhang et al. 2018, Sardana et al. 2019). In conclusion, this study is a nice example of an idea that subsequently has been well-documented to be useful and furthermore implemented in clinical practice in many centers around the world as well as in clinical guidelines.(Soffin et al. 2019a and b, Wainwright et al. 2020). Henrik Kehlet Rigshospitalet, Section of Surgical Pathophysiology, Copenhagen, Denmark E-mail: henrik.kehlet@regionh.dk

Andersen L O, Kehlet H. Analgesic efficacy of local infiltration analgesia in hip and knee arthroplasty: a systematic review. Br J Anaesth 2014; 113(3): 360-74.

225

Kerr D R, Kohan L. Local infiltration analgesia: a technique for the control of acute postoperative pain following knee and hip surgery: a case study of 325 patients. Acta Orthop 2008; 79(2): 174-83. Sardana V, Burzynski J M, Scuderi G R. Adductor canal block or local infiltrate analgesia for pain control after total knee arthroplasty? A systematic review and meta-analysis of randomized controlled trials. J Arthroplasty 2019; 34(1): 183-9. Schotanus M G M, Bemelmans Y F L, van der Kuy P H M, Jansen J, Kort N P. No advantage of adrenaline in the local infiltration analgesia mixture during total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2017; 25(9): 2778-83. Soffin E M, Gibbons M M, Ko C Y, Kates S L, Wick E, Cannesson M, Scott M J, Wu C L. Evidence review conducted for the Agency for Healthcare Research and Quality Safety Program for Improving Surgical Care and Recovery: focus on anesthesiology for total knee arthroplasty. Anesth Analg 2019a; 128(3): 441-53. Soffin E M, Gibbons M M, Ko C Y, Kates S L, Wick E C, Cannesson M, Scott M J, Wu C L. Evidence review conducted for the Agency for Healthcare Research and Quality Safety Program for Improving Surgical Care and Recovery: focus on anesthesiology for total hip arthroplasty. Anesth Analg 2019b; 128(3): 454-65. Spreng U J, Dahl V, Hjall A, Fagerland M W, Raeder J. High-volume local infiltration analgesia combined with intravenous or local ketorolac+morphine compared with epidural analgesia after total knee arthroplasty. Br J Anaesth 2010; 105(5): 675-82. Tan N L, Gotmaker R, Barrington M J. Impact of local infiltration analgesia on the quality of recovery after anterior total hip arthroplasty: a randomized, triple-blind, placebo-controlled trial. Anesth Analg 2019; 129(6): 1715-22. Wainwright T W, Gill M, McDonald D A, Middleton R G, Reed M, Sahota O, Yates P, Ljungqvist O. Consensus statement for perioperative care in total hip replacement and total knee replacement surgery: Enhanced Recovery After Surgery (ERAS) Society recommendations. Acta Orthop 2020; 91(1): 3-19. Zhang L K, Ma J X, Kuang M J, Ma X L. Comparison of Periarticular local infiltration analgesia with femoral nerve block for total knee arthroplasty: a meta-analysis of randomized controlled trials. J Arthroplasty 2018; 33(6): 1972-8.e4.

226

Acta Orthopaedica 2020; 91 (3): 226–227

Predictors of length of stay and patient satisfaction after hip and knee replacement surgery Fast-track experience in 712 patients Henrik Husted, Gitte Holm, and Steffen Jacobsen Acta Orthopaedica 2008; 79 (2): 168–173 DOI 10.1080/17453670710014941

Background and purpose Very few studies have focused on patient characteristics that influence length of stay (LOS) in fast-track total hip (THR) and knee arthroplasty (TKR). The aim of this prospective study was to identify patient characteristics associated with LOS and patient satisfaction after total hip and knee replacement surgery. Patients and methods Between September 2003 and December 2005, 712 consecutive, unselected patients (440 women) with a mean age of 69 (31–91) years were admitted for hip and knee replacement surgery at our specialized fast-track joint replacement unit. Epidemiological, physical, and perioperative parameters were registered and correlated to LOS and patient satisfaction.

Results 92% of the patients were discharged directly to their homes within 5 days, and 41% were discharged within 3 days. Age, sex, marital status, co-morbidity, preoperative use of walking aids, pre- and postoperative hemoglobin levels, the need for blood transfusion, ASA score, and time between surgery and mobilization, were all found to influence postoperative outcome in general, and LOS and patient satisfaction in particular. Interpretation We identified several patient characteristics that influence postoperative outcome, LOS, and patient satisfaction in our series of consecutive fast-track joint replacement patients, enabling further attention to be paid to certain aspects of surgery and rehabilitation. ■

Guest editorial

Enhanced Recovery after arthroplasty surgery Terminology for accelerated postoperative functional recovery of patients after arthroplasty surgery has changed from fast track surgery, rapid recovery surgery, enhanced recovery after surgery (ERAS), as well as initiatives like getting it right first time (GRIFT). All aim to have the most favorable treatment course (i.e. optimal benefit/risk balance) for the patient after surgery, in the shortest length of hospital stay (LOS). But enhanced recovery is more than just reducing length of hospital stay, admitting patients the day of surgery, mobilize them within 3-6 hours after surgery and get patients in control of daily living activities. The aim of this perioperative and rehabilitation enhanced recovery period is to “cure” a patient from its pathology or symptoms including the stress invoked by the treatment as such. For that matter, great diversity exists after arthroplasty surgery, some patients perform excellent, and are referred to as examples of “best practice” outcome, while others are worse after surgery not expected by the patient or clinician. The initiative of enhanced recovery after surgery is based on (patho)physiological principles on how a human being responds to surgical stress. Henrik Kehlet, a Danish surgeon, has addressed this complex pathophysiological phenomenon in the perioperative period extensively since the late 80’s up to present day in collaboration with all surgical disciplines

(Kehlet 1997). A such he used existing basic knowledge to change thinking in clinical practice. Evidence-based medicine is the conscientious, explicit and ethical use of best available evidence to have optimal care of individual patients. For that matter knowledge on pathophysiological principles, recovery after stress, the impact of stress on the patient was nothing knew, but the combination throughout the overall perioperative period was rather new. The surgical stress response originates at the surgical site with an inflammatory response (IL-6, PGE), katabolic/anabolic hormonal release/ body heat loss which cause more stress hormone release within the feedback loop, which causes more stress etc. The basic idea is to reduce these effects of surgical stimuli by optimizing the individual patient before, during and after surgery. Perioperative blood management is only one important factor, which was evaluated in the largest RCT on blood management in over 2,500 patients (So-Osman et al. 2014). The integral approach on this and other perioperative factors on the patient recovery are addressed in the multimodal approach on pathophysiological responses after surgery in patients. A such enhanced recovery after surgery (ERAS), as terminology perfectly explains this complex process, is then the repair of one part of the musculoskeletal organ (e.g. hip, knee, spine etc.) of a patient. It needs a multidis-

Acta Orthopaedica 2020; 91 (3): 226–227

ciplinary collaboration between surgeon, anaesthesiologists, nursing staff, rehabilitation and the patient. Henrik Husted, Gitte Holm, and Steffen Jacobsen evaluated predictors for length of stay and patient satisfaction after an accelerated hip and knee arthroplasty regimen (Husted et al. 2008). Their conclusion on the predictors for good outcome was “Rather it is the sum of information given before and during the admission – regarding the intended regimen, the intended short LOS, and the motivation of the patient to actively participate – that results in shorter LOS compared to conventional surgical tracks”. They combined several modalities of knowledge at that time (2003–2006), including extensive education to the patient and their family, in 712 unselected patients. The active participation of the patient is his or hers recovery was sensed to be important. During the last decade the patient has been transformed from a passive human being undergoing a treatment into a more active stakeholder, engaging into his or hers recovery. Engagement in possible treatment options with the consulting clinician, also stimulated the development of patient reported outcome (PRO) measures as well its value as outcome measures. Since Husted’s paper (Husted et al. 2008), a multitude of articles on enhanced recovery have been written. Recently data from over 400,000 hip and 400,000 knee arthroplasty patients from the National Joint Registry (NJR) of the UK & Wales showed that across the country during 10 years a trend towards shorter LOS, and a decrease of 90 days complications to 1.6% irrespective of the implementation of ERAS (Judge et al. 2020). Notably, quality of life scores did not show clinical important differences between ERAS and non-ERAS hospitals. Comparable results were found with data from the Swedish hip (SHAR) and knee (SKAR) registry (Berg et al. 2020). Nevertheless it seems obvious that the focus of the last 2 decades on enhanced recovery in a multimodal approach is very likely to be instrumental to all (also non-ERAS) hospital protocols. Standard clinical practice as such does not exist, it is an adaptive process but evidence on the enhanced recovery has been evaluated recently (Wainwright et al. 2020).

227

In 2020 enhanced recovery after surgery protocols can be considered as least as good as conventional care for patients, since conventional care as such is not similar to conventional care of 2000, but is today comparable to or almost similar to enhanced recovery. Henrik Kehlet integrated his ideas and research, “logos”, on the pathophysiological response of the human body to a surgical intervention, by his and other authors “pathos” the topic of enhanced recovery after surgery and thus rapid functional recovery became viral. Essentially, the aim is “ethos” recovery of our patients. Rob G H H Nelissen Leiden University Medical Center, Albinusdreef 2, 2333 ZA Leiden, the Netherlands E-mail: r.g.h.h.nelissen@lumc.nl

Berg U, W-Dahl A, Rolfson O, Nauclér E, Sundberg M, Nilsdotter A. Influence of fast-track programs on patient-reported outcomes in total hip and knee replacement (THR/TKR) at Swedish hospitals 2011-2015: an observational study including 51,169 THR and 8,393 TKR operations. Acta Orthop 2020: 91(3): 306-12. Husted H, Holm G, Jacobsen S. Predictors of length of stay and patient satisfaction after hip and knee replacement surgery: fast-track experience in 712 patients. Acta Orthop 2008; 79(2): 168-173. Judge A, Carr A, Price A, Garriga C, et al , The Impact of the enhanced recovery pathway and other factors on outcomes and costs following hip and knee replacement: routine data study. Southampton (UK): NIHR Journals Library; 2020 Jan. Health Services and Delivery Research. Kehlet H. Multimodal approach to control postoperative pathophysiology and rehabilitation. Br J Anaesth 1997; 78(5): 606-17. So-Osman C, Nelissen R G, Koopman-van Gemert A W, Kluyver E, Pöll R G, Onstenk R, Van Hilten J A, Jansen-Werkhoven T M, van den Hout W B, Brand R, Brand A. Patient blood management in elective total hip- and knee-replacement surgery (Part 1): a randomized controlled trial on erythropoietin and blood salvage as transfusion alternatives using a restrictive transfusion policy in erythropoietin-eligible patients. Anesthesiology 2014; 120(4): 839-51. Wainwright T W, Gill M, McDonald D A, Middleton R G, Reed M, Sahota O, Yates P, Ljungqvist O. Consensus statement for perioperative care in total hip replacement and total knee replacement surgery: enhanced recovery after surgery (ERAS) society recommendations. Acta Orthop 2020; 91(1): 3-19.

228

Acta Orthopaedica 2020; 91 (3): 228–229

Uncemented and cemented primary total hip arthroplasty in the Swedish Hip Arthroplasty Register Evaluation of 170,413 operations Nils P Hailer, Göran Garellick, and Johan Kärrholm Acta Orthopaedica 2010; 81 (1): 34–41 DOI 10.3109/17453671003685400

Background and purpose Since the introduction of total hip arthroplasty (THA) in Sweden, both components have most commonly been cemented. A decade ago the frequency of uncemented fixation started to increase, and this change in practice has continued. We therefore analyzed implant survival of cemented and uncemented THA, and whether the modes of failure differ between the two methods of fixation. Patients and methods All patients registered in the Swedish Hip Arthroplasty Register between 1992 and 2007 who received either totally cemented or totally uncemented THA were identified (n = 170,413). Kaplan-Meier survival analysis with revision of any component, and for any reason, as the endpoints was performed. Cox regression models were used to calculate risk ratios (RRs) for revision for various reasons, adjusted for sex, age, and primary diagnosis. Results Revision-free 10-year survival of uncemented THA was lower than that of cemented THA (85% vs. 94%, p < 0.001). No age or diagnosis groups benefited from the use of uncemented fixation. Cox regression analysis confirmed that uncemented THA had a higher risk of revision for any reason (RR = 1.5, 95% CI: 1.4–1.6) and for aseptic loosening (RR = 1.5, CI: 1.3–1.6). Unce-

mented cup components had a higher risk of cup revision due to aseptic loosening (RR = 1.8, CI: 1.6–2.0), whereas uncemented stem components had a lower risk of stem revision due to aseptic loosening (RR = 0.4, CI: 0.3–0.5) when compared to cemented components. Uncemented stems were more frequently revised due to periprosthetic fracture during the first 2 postoperative years than cemented stems (RR = 8, CI: 5–14). The 5 most common uncemented cups had no increased risk of revision for any reason when compared with the 5 most commonly used cemented cups (RR = 0.9, CI: 0.6–1.1). There was no significant difference in the risk of revision due to infection between cemented and uncemented THA. Interpretation Survival of uncemented THA is inferior to that of cemented THA, and this appears to be mainly related to poorer performance of uncemented cups. Uncemented stems perform better than cemented stems; however, unrecognized intraoperative femoral fractures may be an important reason for early failure of uncemented stems. The risk of revision of the most common uncemented cup designs is similar to that of cemented cups, indicating that some of the problems with uncemented cup fixation ■ may have been solved.

Guest editorial

Uncemented science at its best! What makes a paper successful, meaning in this case much cited? It has been suggested that a short title, as here, is one of the keys to success. More seriously, the paper by Hailer, Garellick and Kärrholm (2010) addressed a relevant, complex and controversial question of high clinical utility: the surgeon’s choice of fixation of a primary total hip arthroplasty and the influence of this decision on short- to mid-term implant survivorship and revision causes. And although the evidence has evolved, the topic is still much discussed today. This is illustrated by the fact that the paper has been cited more than 200 times, with an average citation of over 20 each year, and no sign of declining interest in the last years. The authors based their study on a very large number of operations (more than 170,000 primary THAs operated 1992– 2007) from the Swedish Hip Arthroplasty Register and on a robust data collection and analysis methodology. The introduction and discussion sections were comprehensive, multifaceted, thought provoking and well written. As always in clinical research, studying the influence of a specific factor/exposure on an outcome has to take into account its multifactorial con-

text, and has to be aware of the fact that it is conducted in a given time and place. Here the influence of implant fixation on prosthesis longevity was evaluated in the interplay with the individual prosthesis brand(s), the surgeon’s experience and patient characteristics. Moreover, their outcome comparison took into account that choice of fixation might influence short-, mid-, long-term survival differently and might lead to different revision cause patterns. The authors found that THAs with all cemented fixation had overall a much better survival (allcause revision) at 2, 10 and 15 years than all uncemented fixation, with 10-year survival of 85% for uncemented and 94% for cemented THAs. No age group benefitted from the use of uncemented fixation. Nonetheless, the study did not conclude with a “one size fits all” message. Instead they pointed out that uncemented cups were “the Achilles’ tendon of uncemented THA” and that uncemented stems performed better overall than cemented stems. However, use of uncemented stems was associated with an eight times greater risk of stem revision for periprosthetic fracture in the first two years after surgery. Others have studied the same question in the following years. Among them I will point out two publications from the same

Acta Orthopaedica 2020; 91 (3): 228–229

place (Scandinavia), but different time periods. The first study (Mäkelä et al. 2014) evaluated 347,899 primary THAs operated 1995–2011 from the Nordic Arthroplasty Register Association and found a lower overall 10-year survival (all-cause revision) of 91.4–93.5% for all uncemented vs. 93.4–95.0% for all cemented THAs with an improvement in survival for both types of fixation from the late nineties to the early 2000 years. Looking at age, the survival disadvantage of all uncemented fixation was present in patients aged 65 years or older, but not in younger patients. Periprosthetic fracture was a more common cause of revision after uncemented (27%) than after cemented fixation (4%). The second study (Dale et al. 2020) evaluated 66,995 primary THAs operated in Norway 2005–2017 overall and by strata of fixation, age and sex. The 10-year survival (all-cause revision) was 94% with all uncemented and 95% with all cemented fixation. The authors stated: “We found good overall survival for common, contemporary, well-documented primary THAs regardless of fixation method: cemented, uncemented, reverse hybrid, or hybrid fixation. However, uncemented THAs had a slightly higher overall risk of revision compared with cemented THAs. This difference was mainly caused by an increased risk of periprosthetic fracture and dislocation after uncemented THA, in particular when used in elderly women.” Summarizing the development from the early nineties until most recently: First, 10-year survival of primary THA with all uncemented fixation has improved since the study by Hailer et al. Second, there are still slightly more revisions overall with all uncemented vs. all cemented fixation, mainly happening during the first year(s). Third, the well-known, repeatedly highlighted issue of a much higher incidence of – especially intraoperative – periprosthetic fractures with the use of uncemented stems has not been solved (Abdel et al. 2016, Dale et al. 2020, Hailer et al. 2010, Mäkelä et al. 2014). This complication has serious consequences including substantially higher mortality and stem revision rates early on and up to 10 years following the uncemented primary THA (Lamb et al. 2019a). Solutions have been proposed consisting among

229

others of avoidance of uncemented fixation in elderly patients, and in particular in women over 75 years of age (Abdel et al. 2016, Dale et al. 2020, Lamb et al. 2019b). At a time when the personalized/stratified medicine approach for improved patient-centred care is on many lips, here is an area to apply it to. The paper by Hailer, Garellick, and Kärrholm paved the way. Anne Lübbeke Division of Orthopaedic Surgery and Traumatology, Geneva University Hospitals, Switzerland and Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, University of Oxford, UK E-mail: anne.lubbekewolff@hcuge.ch

Abdel M P, Watts C D, Houdek MT, Lewallen D G, Berry D J. Epidemiology of periprosthetic fracture of the femur in 32 644 primary total hip arthroplasties: a 40-year experience. Bone Joint J 2016; 98-B(4): 461-7. Dale H, Børsheim S, Kristensen T B, Fenstad A M, Gjertsen J E, Hallan G, Atle S, Furnes O. Fixation, sex, and age: highest risk of revision for uncemented stems in elderly women — data from 66,995 primary total hip arthroplasties in the Norwegian Arthroplasty Register. Acta Orthop 2020; 91(1): 33-41. Hailer N P, Garellick G, Kärrholm J. Uncemented and cemented primary total hip arthroplasty in the Swedish Hip Arthroplasty Register: evaluation of 170,413 operations. Acta Orthop 2010; 81(1): 34-41 Lamb J N, Matharu G S, Redmond A, Judge A, West R M, Pandit H G. Patient and implant survival following intraoperative periprosthetic femoral fractures during primary total hip arthroplasty: an analysis from the national joint registry for England, Wales, Northern Ireland and the Isle of Man. Bone Joint J 2019a; 101-B(10): 1199-208. Lamb J N, Matharu GS, Redmond A, Judge A, West R M, Pandit H G. Risk factors for intraoperative periprosthetic femoral fractures during primary total hip arthroplasty. An analysis from the National Joint Registry for England and Wales and the Isle of Man. J Arthroplasty 2019b; 34(12): 3065-73. Mäkelä K, Matilainen M, Pulkkinen P, Fenstad A M, Havelin L, Engesaeter L, Furnes O, Pedersen A B, Overgaard S, Kärrholm J, Malchau H, Garellick G, Ranstam J, Eskelinen A. Failure rate of cemented and uncemented total hip replacements: register study of combined Nordic database of four nations. BMJ 2014; 348: f7592.

230

Acta Orthopaedica 2020; 91 (3): 230–231

Long-term registration has improved the quality of hip replacement A review of the Swedish THR Register comparing 160,000 cases Peter Herberts and Henrik Malchau Acta Orthop Scand 2000; 71 (2): 111–121 DOI: 10.1080/000164700317413067

ABSTRACT – The Swedish Hip Register has defined the epidemiology of total hip replacement in Sweden. Most hip implants are fully cemented. Serious complications and rates of revision have declined significantly despite an increasing number of patients at risk. During the past 5 years, only 8–9% of hip replacements are revisions. Aseptic loosening with or without osteolysis is the major problem and constitutes 71% of the revisions, but the incidence had decreased three times during the past 15

years to less than 3% at 10 years. The effectiveness of the surgical technique is the most important factor for reducing the risk of revision because of aseptic loosening, but choice of implant is also important. In practice, total hip replacement in Sweden has improved, as judged by information from this Register about individualized patient risks, implant safety, and the greater efficacy of surgical and cementing techniques.

Guest editorial

How registry data can improve outcomes from joint replacement – a seminal paper Charnley, McKee and Ling, arguably the pioneers of Total Hip Replacement (THR), all recommended that there should be some sort of register for hip replacements. Perceptively, Sir John wrote in 1972: “Serious consideration should be given to establishing a Central Register to keep a finger on the pulse of total implant surgery on a nation-wide basis.” However, it was Peter Herberts and Lennart Ahnfelt, following the example set by the Swedish Knee Arthroplasty Register (SKAR 1975), who set up the world’s first hip register in 1979, the Swedish Hip Arthroplasty Registry (SHAR). Nowadays we all realise how registers have formed the basis for the multitude of observational studies that have been conducted over the past 40+ years. What we have also, increasingly realized, is that collecting data is only half the job. It is what can be learnt from the data that is so important. In their article: “Long-term registration has improved the quality of hip replacement” Herberts and Malchau (2000) reviewed the Swedish THR Register comparing the 160,000 cases contained within it. Not only did they show how huge amounts of observational data, can be collected, they introduced us to a methodology of its interrogation and ways of showing its value They made it clear from the outset that it was their primary aim to make hip replacement safer and more reliable for patients. Another aim was to evaluate the performance

of surgeons and implants whilst getting the information into the public domain, so as to influence health care policy. They argued that registry data, properly analysed, was more likely to give a true picture of the effectiveness of the operation than a clinical trial or RCT (Randomized Control Trial). Importantly, they argued that registry data will include the results from all the surgeons doing the operation and not just from those in the originating centre. The tools they used are second nature to us now. Revision of the implant was the trigger for further analysis whether it be surgical performance, the type of implant, the use of cement (and type of cement) etc. The risks of infection and its control were drawn out of the database besides the other common indications for revision. They defined revision whilst accepting that there are other definitions than their own. Anyone who wants to make a registry work properly, will know that the stakeholders and all the people who contribute data, must gain benefit from it. It is pointless having a registry without “feedback”. Herberts and Malchau included interesting and pertinent demographic data. They showed graphs on trends, just as we all do nowadays. Importantly, they showed how, with SHAR, the revision rate in Sweden dropped. Importantly for many of us, they compared their “Post Registry” revision rate with reported revision rates in other countries.

Acta Orthopaedica 2020; 91 (3): 230–231

Over the past 20 years there has been much debate as to which statistical methodology is the best for registry observational data. They used KM (Kaplan–Meier) which set the benchmark and has hardly ever been superseded. The hallmarks of a first-class registry, as defined by ISAR (The International Society of Arthroplasty Registries) are completeness (all hospitals taking part), compliance (all operations being up-loaded) and linkability (the ability to link primary to revision operations in the same patient). Herberts and Malchau pioneered the importance of these standards whilst commenting on the importance they attached to validation. Validation is another of the standards expected by ISAR to qualify a registry to be in the first division. It is often difficult, time consuming and expensive but an effective method is discussed in the text. Cost would have inevitably been a question that any reader would have been thinking about whilst reading this article in 2000. The authors pointed out that the costs of revision operations that could be prevented by surgeons taking heed of the findings reported in the article, would far outweigh the costs of the register and the expense of data analysis. They were fortunate that the advent of the registry was well timed in terms of the IT revolution that we all witnessed, towards the end of the last century. It has made data collection much easier, more accurate and much easier to analyse. There must have been thousands of surgeons who read this article when it was published in 2000 and who were immediately struck by its importance. Surely, it can be assumed that it would not just have been UK surgeons, who had learned from the article that their revision rate was probably 3 times the Swedish rate, that their country also needed a joint replacement registry. There can be no doubt that SHAR and this paper accelerated the introduction of many of the registries we have today.

231

What is even more impressive is that when one reads the article again, very nearly 20 years later we still use virtually the same standards and methodology and reporting methods in most of the world’s registries today. Of course, it is not just hips nowadays but all joints. The basics has been copied by other subspecialities in orthopaedics and increasingly in other medical specialities. At the time this paper was revolutionary. There are often some who will dismiss/ridicule a novel idea which doesn’t sit comfortably with their own beliefs. It helped their cause that both Herberts and Malchau have always been held in the highest regard in terms of honesty and diligence, throughout the world. This article, which many will now regard as seminal, set the pattern for the reporting of registry data, the methodology that should be used and introduced many of the important definitions that we have been using ever since. Keith Tucker Norwich England Chairman of ODEP and the Beyond Compliance Advisory Group United Kingdom E-mail: KTUCKER77@aol.com

Charnley J. Internal Publication of the Wrightington Hip Centre No. 39, July 1972. Herberts P, Malchau H. Long-term registration has improved the quality of hip replacement: a review of the Swedish THR Register comparing 160,000 cases. Acta Orthop Scand 2000; 71(2): 111-21.

232

Acta Orthopaedica 2020; 91 (3): 232–233

Guidelines for standardization of radiostereometry (RSA) of implants Edward R Valstar, Richie Gill , Leif Ryd, Gunnar Flivik , Niclas Börlin and Johan Kärrholm Acta Orthopaedica 2005; 76 (4): 563–572 DOI 10.1080/17453670510041574

ABSTRACT There is a need for standardization of radiostereometric (RSA) investigations to facilitate comparison of outcome reported from different research groups. In this document, 6 research centers have agreed upon standards for terminology, description and use of RSA arrangement including radiographic set-up and techniques. Consensus regarding minimum requirements for marker stability and scatter, choice of coordinate systems, and preferred way of describing prosthetic micromotion is of special interest. Some notes on data interpretation are also presented. Validation of RSA should be standardized by preparation of protocols for assessment of accuracy and precision. Practical issues related to loading of the joint by weight bearing

or other conditions, follow-up intervals, length of followup, radiation dose, and the exclusion of patients due to technical errors are considered. Finally, we present a checklist of standardized output that should be included in any clinical RSA paper. This document will form the basis of a detailed standardization protocol under supervision of ISO and the European Standards Working Group on Joint Replacement Implants (CEN/TC 285/WG4). This protocol will facilitate inclusion of RSA in a standard protocol for implant testing before it is released for general use. Such a protocol—also including other recognized clinical outcome parameters—will reduce the risk of implanting potentially inferior prostheses on a large scale. ■

Guest editorial

Reflections on the RSA guidelines Roentgen stereogrammetric analysis/radiostereometric analy- reporting quality increased almost 3-fold in the period 2006sis (RSA) is an internationally accepted and validated surro- 2011 compared to the period before the RSA guidelines were gate marker for long-term primary joint replacement outcome published. The RSA guidelines also formed the foundation for in terms of aseptic loosening and its positive effect on patient the ISO standard ISO 16087:2013: Implants for surgery — safety even echoes through in national joint registries (Ryd Roentgen stereophotogrammetric analysis for the assessment 1986, Nelissen et al. 2011, Kärrholm 2012). The history of of migration of orthopaedic implants (ISO 16087:2013). This RSA dates back to the time when X-rays were discovered, ISO standard facilitated further optimization and professionwhen Davidson and Hedley determined the 3-D position of alization of RSA research. Additionally, some of the recoma pin that was radiographed on the same radiograph by two mendations in the RSA guidelines on e.g. the use of signed separate x-ray sources (Davidson and Hedley 1897). Modern values, the rationale for using (or not using) Maximal Total RSA dawned in the 1970s and has first been reported by Göran Point Motion (MTPM) and the timing of the first postoperaSelvik in his thesis (Selvik 1989). In 2005 a landmark paper tive examinations are still very relevant today. Copyright© Taylor & Francis 2005. ISSN 1745–3674. Printed in Sweden – all rights reserved. The importance of adequate reporting is being recognized in the field of RSA was published by an international group of RSA experts: “Guidelines for standardization of radioste- across all fields of health care research. Reporting guidelines reometry (RSA) of implants”, also referred to as “The RSA are considered vital for achieving and maintaining high stanguidelines” (Valstar et al. 2005). Presently, this paper counts dards in reporting healthcare research and avoiding waste in over 400 citations in google scholar (accessed 19-02-2020). the production and reporting of research (Altman and Simera Although this number of citations is impressive, the measur- 2016). As such, reporting guidelines specify the minimum able impact on the reporting quality of RSA studies is even information that is needed for a reader to get a clear and commore impressive. Since its publication, the reporting quality plete picture of what was done, what was found and what the of RSA studies has greatly improved; Madanat et al. (2014) results mean, so the study can be completely understood, rephave shown that the proportion of RSA studies with high licated, appraised and the results be interpreted in the correct © 2020 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group, on behalf of the Nordic Orthopedic Federation. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. DOI 10.1080/17453674.2020.1763568

Acta Orthopaedica 2020; 91 (3): 232–233

context (Altman and Simera 2016). It is important to realize that although high reporting quality is required to judge the methodological quality of a study, high reporting quality does not mean the methodological quality is high as well. Perfectly reported studies may be poorly designed and poorly executed and vice versa. For example, a lost to follow-up of 50% in a study indicates poor methodological quality. However, when this fact is clearly reported, indicating high reporting quality, it allows the readers to adequately appraise the methodological study quality. Hence, high reporting quality is a prerequisite to judge the methodological quality of a study. Regarding RSA studies, the reporting quality has improved, but further improvement is possible and necessary (Madanat et al. 2014). In a recent review on migration in total knee replacements it became apparent that RSA studies could benefit from further improvement in the reporting of especially the migration results and precision as determined by double examinations: only 19 of 53 included studies reported precision as determined by original double examinations (Pijls et al. 2018). Clinicians, researchers, clinical guide line developers, systematic reviewers and patients would greatly benefit from standardized and complete reporting of prosthetic migration e.g. the mean migration, the number of RSA examinations for each follow-up and detailed description of the type of prosthesis and fixation method. To further improve reporting and transparency it should become standard practice to register RSA studies (including case series and cohorts) in a trial registry e.g. at clinicaltrials. gov before start of the study and data-analysis, as is common practice for randomized controlled trials. Although study registration is not yet compulsory for RSA cohorts and caseseries, proper study registration of such studies would ensure assessment of publication bias especially when the migration of new prostheses exceed the unacceptable thresholds.

233

The RSA guidelines have greatly enhanced reporting of RSA studies and have paved the way for further improvements especially regarding the reporting of migration results and study registration in trial registries. Bart G Pijls Department of Orthopaedics, Leiden University Medical Center, Leiden, The Netherlands E-mail: b.g.c.w.pijls@lumc.nl

Altman D G, Simera I. A history of the evolution of guidelines for reporting medical research: the long road to the EQUATOR Network. J R Soc Med 2016; 109(2): 67-77. Davidson J M, Hedley W S. A method of precise localisation and measurement by means of roentgen rays. Lancet 1897; 65(16): 1001. ISO 16087:2013(E). Implants for surgery: roentgen stereophotogrammetric analysis for the assessment of migration of orthopaedic implants. Geneva: International Organization for Standardization; 2013. Kärrholm J. Radiostereometric analysis of early implant migration – a valuable tool to ensure proper introduction of new implants. Acta Orthop 2012; 83(6): 551-2. Madanat R, Mäkinen T J, Aro H T, Bragdon C, Malchau H. Adherence of hip and knee arthroplasty studies to RSA standardization guidelines. A systematic review. Acta Orthop 2014; 85(5): 447-55. Nelissen R G, Pijls B G, Kärrholm J, Malchau H, Nieuwenhuijse M J, Valstar E R. RSA and registries: the quest for phased introduction of new implants. J Bone Joint Surg Am 2011; 93(Suppl 3): 62-5. Pijls B G, Plevier J W, Nelissen R G. RSA migration of total knee replacements. Acta Orthop 2018; 89(3): 320-28. Ryd L. Micromotion in knee arthroplasty. A roentgen stereophotogrammetric analysis of tibial component fixation. Acta Orthop; 1986; Suppl 220: 1-80. Selvik G. Roentgen stereophotogrammetry. A method for the study of the kinematics of the skeletal system. Acta Orthop 1989; Suppl 232: 1-51. Valstar E R, Gill R, Ryd L, Flivik G, Borlin N, Kärrholm J. Guidelines for standardization of radiostereometry (RSA) of implants. Acta Orthop 2005; 76(4): 563-72.

234

Acta Orthopaedica 2020; 91 (3): 234–235

Why still in hospital after fast-track hip and knee arthroplasty? Henrik Husted, Troels H Lunn, Anders Troelsen, Lissi Gaarn-Larsen, Billy B Kristensen, and Henrik Kehlet Acta Orthopaedica 2011; 82 (6): 679–684 DOI 10.3109/17453674.2011.636682

Background and purpose Length of stay (LOS) following total hip and knee arthroplasty (THA and TKA) has been reduced to about 3 days in fast-track setups with functional discharge criteria. Earlier studies have identified patient characteristics predicting LOS, but little is known about specific reasons for being hospitalized following fast-track THA and TKA. Patients and methods To determine clinical and logistical factors that keep patients in hospital for the first postoperative 24–72 hours, we performed a cohort study of consecutive, unselected patients undergoing unilateral primary THA (n = 98) or TKA (n = 109). Median length of stay was 2 days. Patients were operated with spinal anesthesia and received multimodal analgesia with paracetamol, a COX-2 inhibitor, and gabapentin—with

opioid only on request. Fulfillment of functional discharge criteria was assessed twice daily and specified reasons for not allowing discharge were registered. Results Pain, dizziness, and general weakness were the main clinical reasons for being hospitalized at 24 and 48 hours postoperatively while nausea, vomiting, confusion, and sedation delayed discharge to a minimal extent. Waiting for blood transfusion (when needed), for start of physiotherapy, and for postoperative radiographic examination delayed discharge in one fifth of the patients. Interpretation Future efforts to enhance recovery and reduce length of stay after THA and TKA should focus on analgesia, prevention of orthostatism, and rapid recovery of muscle function.

Guest editorial

Should length of stay in hospital be the endpoint in arthroplasty? This is, interestingly, one of the 10 most cited papers in the history of Acta after year 2000 (Husted et al. 2011). Interestingly, since length of stay (LOS) is not the most important parameter in arthroplasty: freedom of pain, normalized function and longevity are the ultimate goals. Why is then LOS of such interest? Hospital beds are a limited resource in many parts of the world, irrespective of payer system. LOS has therefore come under surveillance, to the degree that day care arthroplasty has become common in certain hospitals (Hartog et al. 2015). Remember that it is not more than 15 years ago since patients stayed in hospital for 1 to 2 weeks after total joint arthroplasty (TJA). The study on 207 patients undergoing hip or knee arthroplasty registered 2 times a day whether fulfillment of each ofthe discharge criteria had been obtained, and detailed reason(s) for not being discharged. Husted et al. found that in a fast track system pain, dizziness, and general weakness were the main reasons for not being discharged after 24 and 48 hours in 80% of patients. Median LOS was 2 days, and 95% were discharged after 3 days. Waiting for blood transfusion, start of physiotherapy, and for postoperative radiographic examination delayed the discharge for 20%. The first factors can be seen as patient related, while the last ones are hospital factors. The hospital factors could be organizationally removed, while patient factors probably could not be changed. The authors had previously shown that readmissions were not increased by

the fast-track system. The authors themselves concluded that the findings offered the possibility of safe reduction of LOS after fast-track hip or knee arthroplasty. Now, nearly 10 years after its publication, it can be discussed whether being highly cited is equivalent to being an important scientific paper? The study was non-selective in including all patients scheduled for TJA in a 6 months period, thereby it was valid to all patients treated at Hvidovre hospital, and maybe to all patient in Denmark and Scandinavia. It was published in a period when LOS was rapidly decreasing due to implementation of fast-track surgery around the world. Husted et al. studied why some patients were in hospital while others had returned home, a topic which interested all researchers in hospital logistics and post-operative analgesia. The 176 citing papers are mostly on rapid recovery and analgesia. The study reached a peak with 18 citations in 2018. The most surprising citation was in pediatric urology, but also that study was on enhanced recovery after surgery (Haid et al. 2020). Husted and Kehlet have been the pioneers in rapid recovery in Scandinavia, with numerous publications on analgesia (which is a prerequisite for rapid discharge), and recently outpatient total joint surgery (Gromov et al. 2019). The value of the 2011 paper has perhaps been mostly to pave the way for this unthought possibility just 15 years ago, leaving hospital with a new hip or knee the same day as you went in through the hospital doors.

Acta Orthopaedica 2020; 91 (3): 234â&#x20AC;&#x201C;235

Lars Nordsletten Division of Orthopaedic Surgery, Oslo University Hospital, Oslo, Norway E-mail: lars.nordsletten@medisin.uio.no

235

Gromov K, Jorgensen C C, Petersen P B, Kjaersgaard-Andersen P, Revald P, Troelsen A, Kehlet H, Husted H. Complications and readmissions following outpatient total hip and knee arthroplasty: a prospective 2-center study with matched controls. Acta Orthop 2019; 90(3): 281-5. Haid B, Lusuardi L, Oswald J. [Enhanced recovery after surgery â&#x20AC;&#x201C; a concept, also in pediatrics. Urologe A 2020; 59(3): 294-9. Hartog Y M, Mathijssen N M, Vehmeijer S B. Total hip arthroplasty in an outpatient setting in 27 selected patients. Acta Orthop 2015; 86(6): 667-70. Husted H, Lunn T H, Troelsen A, Gaarn-Larsen L, Kristensen B B, Kehlet H. Why still in hospital afte fast-track hip and knee arthroplasty? Acta Orthop 2011; 82(6): 679-84.

236

Acta Orthopaedica 2020; 91 (3): 236–237

The epidemiology of proximal humeral fractures Charles M Court-Brown, Ashima Garg and Margaret M McQueen Acta Orthop Scand 2001; 72 (4): 365–371 DOI:10.1080/000164701753542023

ABSTRACT – We present a 5-year prospective study of the epidemiology of 1,027 proximal humeral fractures. These fractures, which tend to occur in fit elderly persons, have a unipolar age distribution and the highest age-specific incidence occurs in women between 80 and 89 years of age. The commonest was the B1.1 impacted valgus fracture, found in one-fifth of the cases in this series, a type that is not included in the Neer classification.

We used both Neer and AO classifications. The AO classification proved to be more comprehensive because in the Neer classification, half of the fractures are minimally displaced and almost nine-tenths fall into only three categories. In the AO classification, the B1.1, A2.2, A3.2 and A1.2 sub-groups comprise over half of all proximal humeral fractures, while the AO type C fractures occur in only 6%. We suggest that the literature does not adequately reflect the spectrum of proximal humeral fractures.

Guest editorial

Understanding fracture populations by epidemiology In orthopedics, we tend to classify and quantify different traumas in order to focus our efforts and resources on the specific areas of need. To gain a perspective, we use epidemiology as a fallback instead of trusting on our own feelings of prevalence. There are several ways to access the analysis performed in epidemiological studies. For example, hospital discharge registries are usually nationwide or form part of a subsample of a larger area patient cohort; they give a reliable overview of the data attained. There are, however, several problems with such registries. One major problem is the inability to access patientlevel information. This leads to only the overall numbers of incidences and rates of interventions being provided, and a subsequent lack of incident classification and patient reported outcome measures (PROMs). Another way to attain the data is a cohort sample of the catchment population of a specific area. The problem with this method is the relative rarity of the specific cases (i.e., fractures). In addition, it is time consuming to accumulate a large enough cohort for analysis. Furthermore, the data provided are usually retrospective in nature and can lead to deficiencies and heterogeneity of the dataset. The study by Court-Brown et al. (2001) reported the 5-year data of all consecutive proximal humeral fracture patients, including both out-patients and in-patients, treated at the Trauma Unit of a hospital in Edinburgh, Scotland (CourtBrown et al. 2001). At the time, the catchment area for the fracture population was 700,000, which could be estimated to be sufficient for a meaningful cohort for relevant analysis of the fractures. The strength of the study that makes it unique is its prospective nature. As a result, the data acquired were

comprehensive with detailed patient demographics and radiographic images for fracture classification. The incidence data implied full coverage and stratified by age and were included adjacent, patient-matched fracture classification. The study raised several important issues. For example, the fractures that occurred in young patients were mainly related to sports or high velocity injuries, whereas those that occurred in patients 30 years and older were mostly caused by falling from a standing height. The authors could show the peak incidence for older patients, which was also associated with more severe fracture patterns than in younger patients. Furthermore, they were able to classify all the patients using the Neer and AO classification systems (Neer 1970, Müller et al. 1990). Of the 2 systems used, the AO-classification was shown to be more comprehensive. The authors acknowledged, however, that there were problems in intra- and interobserver reliability, which was later rated as moderate for both the Neer and AO classifications (Papakonstantinou et al. 2016). However, they raised this important issue for discussion: “Neer three- and four-part fractures and fracture dislocations and the AO type C fractures occurred with an incidence of 13% and 6%, respectively, indicating that the literature does not adequately respect the spectrum of proximal humeral fractures” In our smaller, retrospective cohort study from year 2015, we found a similar rate of fractures classified with Neer for three- and four-part fractures with an incidence of 19% and 7%, respectively (Launonen et al. 2015). Therefore, the veracity of the statement still stands. Naturally, such multipart frac-

Acta Orthopaedica 2020; 91 (3): 236â&#x20AC;&#x201C;237

tures cause the most difficulty for clinicians when trying to make treatment decisions. The Court-Brown study has aged well, and it can still be considered as a benchmark of proximal humeral fracture epidemiology with large prospective patient cohorts. In the clinical and research field, we need detailed epidemiological data so that we can focus on the key issues. However, the data provided in the study need an update since the age distribution has changed over time, which may affect the fracture distribution. Although the epidemiology and fracture classification do not help us decide on the best treatments for patients, they do help us to understand the disparity between the common and the rare and act as a baseline in the guidelines (Brorson et al. 2012). Thus, they help us to better focus our efforts and resources on the specific areas of need. Antti P Launonen Department of Orthopaedics and Traumatology, Tampere University Hospital, Teiskontie, Tampere, Finland E-mail: antti.launonen@pshp.fi

237

Brorson S, Olsen B S, Frich LH, Jensen SL, Sorensen A K, Krogsgaard M, Hrobjartsson A. Surgeons agree more on treatment recommendations than on classification of proximal humeral fractures. BMC Musculoskelet Disord 2012; 13: 114. doi: 10.1186/1471-2474-13-114. Court-Brown C M, Garg A, McQueen M M. The epidemiology of proximal humeral fractures. Acta Orthop Scand. 2001; 72(4): 365-71. doi: 10.1080/000164701753542023. Launonen A P, Lepola V, Saranko A, Flinkkila T, Laitinen M, Mattila V M. Epidemiology of proximal humerus fractures. Arch Osteoporos 2015; 10(1): 209. doi: 10.1007/s11657-015-0209-4. MĂźller M E, Nazarian S, Koch P, Schatzker J. The comprehensive classification of fractures of long bones. 1 ed. Springer-Verlag, Berlin Heidelberg 1990. Neer C S, 2nd. Displaced proximal humeral fractures. I. Classification and evaluation. J Bone Joint Surg Am 1970; 52(6): 1077-89. Papakonstantinou M K, Hart M J, Farrugia R, Gabbe B J, Kamali Moaveni A, van Bavel D, Page R S, Richardson M D. Interobserver agreement of Neer and AO classifications for proximal humeral fractures. ANZ J Surg 2016; 86(4): 280-4. doi: 10.1111/ans.13451.

238

Acta Orthopaedica 2020; 91 (3): 238–239

The Norwegian Arthroplasty Register 11 years and 73,000 arthroplasties Leif I Havelin, Lars B Engesæter, Birgitte Espehaug, Ove Furnes, Stein A Lie and Stein E Vollset Acta Orthop Scand 2000; 71 (4): 337–353 DOI: 10.1080/000164700317393321

ABSTRACT – In 1985, the Norwegian Orthopaedic Association decided to establish a national hip register, and the Norwegian Arthroplasty Register was started in 1987. In January 1994, it was extended to include all artificial joints. The main purpose of the register is to detect inferior results of implants as early as possible. All hospitals participate, and the orthopedic surgeons are supposed to report all primary operations and all revisions. Using the patient’s unique national social security number, the revision can be linked to the primary operation, and survival analyses of the implants are done. In general, the survival analyses are performed with the Kaplan-Meier method or using Cox multiple regression analysis with adjustment for possible confounding factors such as age, gender, and diagnosis. Survival probabilities can be calculated for each of the prosthetic components. The end-point in the analyses is revision surgery, and we can assess the rate of revision due to specific causes like aseptic loosening, infection, or dislocation. Not only survival, but also pain, function, and satisfaction have been registered for subgroups of patients. We receive reports about more than 95% of the prosthesis operations. The register has detected inferior implants 3 years after their introduction, and several uncemented prostheses were abandoned during the early 1990s due to our documentation of poor performance. Further, our results also contributed to withdrawal of

the Boneloc cement. The register has published papers on economy, prophylactic use of antibiotics, patients’ satisfaction and function, mortality, and results for different hospital categories. In the analyses presented here, we have compared the results of primary cemented and uncemented hip prostheses in patients less than 60 years of age, with 0–11 years’ follow-up. The uncemented circumferentially porous- or hydroxyapatite (HA)-coated femoral stems had better survival rates than the cemented ones. In young patients, we found that cemented cups had better survival than uncemented porous-coated cups, mainly because of higher rates of revision from wear and osteolysis among the latter. The uncemented HA-coated cups with more than 6 years of follow-up had an increased revision rate, compared to cemented cups due to aseptic loosening as well as wear and osteolysis. We now present new findings about the six commonest cemented acetabular and femoral components. Generally, the results were good, with a prosthesis survival of 95% or better at 10 years, and the differences among the prosthesis brands were small. Since the practice of using undocumented implants has not changed, the register will continue to survey these implants. We plan to assess the mid- and long-term results of implants that have so far had good short-term results.

Guest editorial

A blueprint for global registries from Norway In 2000 ACTA published a landmark article authored by the leads of the Norwegian Arthroplasty Registry. Reviewing it today it is easy to understand why it is one of the 10 most cited ACTA articles of the last 20 years. The Norwegian Arthroplasty Registry commenced hip data collection in 1987 and expanded to include other arthroplasty procedures in 1994. In 2000 it was one of a small number of quality national arthroplasty registries with sufficient follow up to provide meaningful data. Others included the Swedish Knee, Swedish Hip and Finnish registries. Publications by these foundation registries particularly in the 10 or so years prior had brought world attention to the value of registry data. By the late 1990’s and early 2000’s many other national registries were being established or planned. The importance of this publication was that it was a blueprint for developing registries on how to successfully implement and manage a national arthroplasty registry.

The article describes the purpose, methodology, governance, and achievements of the Norwegian Arthroplasty Registry. It provided a rational for why registries should be established and how beneficial change could be achieved. Selective examples of registry analysis were also provided that highlighted the value and importance of this data. Registries are quality assurance mechanisms designed to monitor and provide real time stakeholder feedback. Their purpose is to ensure continuous quality improvement. Assessing and reporting comparative prosthesis performance is one of those important quality assurance activities. The article justifies the need for this because at that time it was common practice for surgeons to implant “undocumented” prostheses i.e. prostheses without evidence. Regulators did not require prosthesis specific premarket clinical evaluation and post market clinical trials were rarely undertaken. A number of examples of registry identified outlier performance were

Acta Orthopaedica 2020; 91 (3): 238–239

provided including the Norwegian Registry “Boneloc” identification. Despite the litany of prosthesis specific problems reported by registries since that time disappointingly there has been little change to regulatory premarket evaluation requirements and the use of “undocumented” prostheses continues. Performance also varies by prosthesis class. Evidence was presented on reduced revision rates in younger patients associated with some designs of cementless compared to cemented femoral stems. This and the identification of specific characteristics of cementless designs associated with success was important information at that time. Comments on the need to better evaluate metal on metal bearings and the potential good outcomes of crosslinked polyethylene were eerily predictive and reflect the informed commentary, expertise and understanding of the Norwegian Arthroplasty team. Critical to registry success is surgeon engagement. A strategy recommended by the Norwegian Registry was the provision of regular reports to participating hospitals. Surgeons

239

being informed of their comparative performance either personal or through practice hospital data is now known to be a major incentive for ongoing participation and consequently most registries provide this type of data. Successful engagement also requires surgeon confidence in data accuracy and completeness, analytical techniques as well as transparency and accountable governance. The article provided specific information on each of these issues. It is only high-quality publications that are frequently cited. They clearly and succinctly define the issues and provide real and possible solutions, grow knowledge and provide unexpected insights. They often stand the test of time. This article has all of these characteristics and it is just as valuable and relevant today as when it was published 20 years ago. Stephen Ellis Graves Email: segraves@aoanjrr.org.au

240

Acta Orthopaedica 2020; 91 (3): 240

Effectiveness of hip or knee replacement surgery in terms of quality-adjusted life years and costs Pirjo Räsänen, Pekka Paavolainen , Harri Sintonen, Anna-Maija Koivisto, Marja Blom, Olli-Pekka Ryynänen and Risto P Roine Acta Orthopaedica 2007; 78 (1): 108–115 DOI 10.1080/17453670610013501

Background Concurrent head-to-head comparisons of healthcare interventions regarding cost-utility are rare. The concept of favorable cost-effectiveness of total hip or knee arthroplasty is thus inadequately verified. Patients and methods In a trial involving several thousand patients from 10 medical specialties, 223 patients who were enrolled for hip or knee replacement surgery were asked to fill in the 15D health-related quality of life (HRQoL) survey before and after operation. Results Mean (SD) HRQoL score (on a 0–1 scale) increased in primary hip replacement patients (n = 96) from 0.81 (0.084) preoperatively to 0.86 (0.12) at 12 months (p < 0.001). In revision hip replacement (n = 24) the corresponding scores were 0.81 (0.086) and 0.82

(0.097) respectively (p = 0.4), and in knee replacement (n = 103) the scores were 0.81 (0.093) and 0.84 (0.11) respectively (p < 0.001). Of 15 health dimensions, there were statistically significant improvements in moving, usual activities, discomfort and symptoms, distress, and vitality in both primary replacement groups. Mean cost per quality-adjusted life year (QALY) gained during a 1-year period was € 6,710 for primary hip replacement, € 52,274 for revision hip replacement, and € 13,995 for primary knee replacement. Interpretation Hip and knee replacement both improve HRQoL. The cost per QALY gained from knee replacement is twice that gained from hip replacement. ■

Guest editorial

Hip, knee and revision hip replacement – are they as clinically and cost effective as we think? Modern medicine is expensive and thus societal choices around resource allocation are important to ensure maximal benefit. Total hip and knee replacement are two of the commonest elective surgical treatments performed in the developed world and have been shown to be effective in relieving pain and improving function in those with advanced osteoarthritis. However, treatments need to be both clinically effective and cost effective in order to justify their widespread use, particularly in healthcare systems that are subsidised by general taxation, as is the practice in most European countries. Comparing different treatments for different conditions is both difficult and contentious. Quality adjusted life years (QALYs) is one method of doing this and is predicated on the idea that the treatment cost of delivering a quantifiable unit of improvement in health can be measured. This will allow us to both contrast the value of treatments and ascertain levels of “willingness to pay” for health gain. Rasanen and colleagues from Finland measured the quality of life gained following both primary and revision hip replacement and primary knee replacement by cost incurred to achieve that gain. Primary hip and knee replacement improved the mean quality of life scores of patients, but improvements after revision hip replacement were neither clinically nor statistically significant. Furthermore, the cost per unit of improvement was twice as high for TKR as THR

and nearly eight times as high for revision compared to primary total hip replacement. This paper is a landmark as it elegantly demonstrates the effectiveness and cost effectiveness of both hip and knee replacements, whilst showing what many surgeons suspected that knee replacement results in slightly lower gains, compared to hip replacements, at higher costs. Probably the most interesting finding is that revision hip replacement results in little or no improvement in overall quality of life despite being very expensive. This finding should cause surgeons to reflect on their practice. Certainly, many revisions are performed to prevent symptoms worsening (such as after fracture or infection) and this analysis compares before and after rather than the sequelae of treatment versus no treatment. However, many revisions, particularly after knee replacement, are performed to treat stable conditions such as persistent pain. Even though the work presented here was published in 2007, we as a community have not yet established the utility, efficacy and wisdom of performing revision arthroplasty for stable conditions affecting pain and function. The findings of Rasanen and colleagues are as pertinent today as they were in 2007. Ashley Blom Head of Bristol Medical School, University of Bristol, UK E-mail: Ashley.Blom@bristol.ac.uk

Acta Orthopaedica 2020; 91 (3): 241–242

241

Patient satisfaction after knee arthroplasty A report on 27,372 knees operated on between 1981 and 1995 in Sweden Otto Robertsson, Michael Dunbar, Thorbjörn Pehrsson, Kaj Knutson and Lars Lidgren Acta Orthop Scand 2000; 71 (3): 262–267 DOI:10.1080/000164700317411852

ABSTRACT – During a validation process of the Swedish Knee Arthroplasty Register (SKAR), living registered patients were sent a questionnaire to ask if they had been reoperated on. This gave an opportunity to pose a simple four-point question with respect to patient satisfaction which 95% of patients answered. We analyzed the answers of patients operated on between 1981 and 1995 and found that only 8% of the patients were dissatisfied regarding their knee arthroplasty 2–17 years postoperatively. The satisfaction rate was constant, regardless of when the operation had been performed during the 15-year period. The proportion of satisfied patients was affected by the preoperative diagnosis, patients operated on for a long-standing disease more often being satisfied than those with a short disease-duration. There was no differ-

ence in proportions of satisfied patients, whether they had primarily been operated on with a total knee arthroplasty (TKA) or a medial unicompartmental arthroplasty (UKA). For TKAs performed with primary patellar resurfacing, there was a higher ratio of satisfied patients than for TKAs not resurfaced, but this increased ratio diminished with time passed since the primary operation. Unrevised knees had a higher proportion of satisfied patients than knees that had been subject to revision, and among patients revised for medial UKA, the proportion of satisfied patients was higher than among patients revised for TKA. We conclude that satisfaction after knee arthroplasty is stable and long-lasting in unrevised cases and that even after revision most patients are satisfied.

Guest editorial

Patient satisfaction after total knee replacement—still a challenge In the 1990s patients reported outcome measures (PROMS) were developed to reduce the risk of bias if outcome is rated by the surgeon. When the Swedish Knee Registry sent out a mail in 1999 to validate their registry to check the revision status of the patients, they included a simple question “How satisfied are you with your knee replacement?” 95% of all patients responded and were clearly less positive than expected. Robertsson et al. (2000) reported that 17% of total knee replacement (TKR) patients were either dissatisfied or uncertain with respect to the outcome. This is lower than satisfaction after total hips replacement (THR). Since this key publication, the rate of satisfaction has been studied in many other groups of patients and found to be consistent in many countries. Only 4 out of 5 patients are satisfied after TKR (Bourne et al. 2010, Dunbar et al. 2013, Bryan et al. 2018). Identifying the causes of dissatisfaction is important in order to improve patient selection for TKR, adjust treatment strategies and to support or treat dissatisfied patients with their residual complaints. Sociodemographic, preoperative, operative, and postoperative factors have been studied in large reviews. No specific single leading factor has been found, but patients expectations, higher function before surgery,

lower stage of arthritic disease, complications, poor resolution of pain, and lower improvement of knee function were more common in dissatisfied patients (Gunaratne et al. 2017). Patients with a better preoperative mental function were more often satisfied (Vissers et al. 2010, 2012). However, in almost all studies it was found that unfulfilled expectations were the main reason for dissatisfaction. Many studies advised improving patient information and education preoperatively (ConnerSpady et al. 2020, Ghomrawi et al. 2020). Tilbury et al. (2016) reported that in dissatisfied patients unfulfilled expectations were found for “improvement walking ability middle long distances” (40%), “being able to kneel down” (47%) and “being able to squat”(44%). 20 years ago in Acta Orthopaedica Robertsson’ s publication (Robertsson et al. 2000), opened the eyes of orthopedic world: there was a discrepancy between patient and surgeon satisfaction after TKR. Unmet expectations are a main source of patient dissatisfaction and patients have the right to be informed about the limitations that current replacement techniques have. Over the past 2 decades, new knee implants have been introduced as well as new techniques including; computer assisted surgery, patients specific guides and alternative

242

alignment techniques. However, in unbiased studies none of these techniques and implants have shown a significant improvement of patient satisfaction. The gap between the satisfaction rates of THR and TKR may be caused by the more complex nature of the knee joint compared to the hip. The anatomy of the knee ligaments and the individual form and size of femur, tibia and patella may be better addressed with a customized patient specific prosthesis implanted with a surgical robot to optimize precision (Namin et al. 2019, Robinson et al. 2019). Both developments are underway and may lead to a paradigm shift in TKR necessary to overcome the high percentage of dissatisfied patients. It is very important to analyze patients experiences when introducing these techniques. Based on the expected considerable increase of costs of the TKR procedure health economics also need to be studied. Until real improvements are achieved, we orthopedic surgeons should be humble and realistic. TKR is a good, but not ideal, option for patients with significant complaints due to end-stage arthritis. We need to be careful in young patients, those with unbearable pain for which narcotics are used, and patients who want to resume high level sports activities. Reduction of pain and improvement of function may be expected but some complaints may persist. There are also possible complications including infection and thrombosis, which occur in less than 5 % of patients, but may create more problems than preoperatively. Pain relief and improving physical function are the main aims of TKR. Expectations should be explicitly addressed before surgery; a lesson now 20 years old, yet still true today. Jan Verhaar Department of Orthopaedics and Sports Medicine, Erasmus University Medical Center Rotterdam, the Netherlands E-mail: j.verhaar@erasmusmc.nl

Acta Orthopaedica 2020; 91 (3): 241–242

Bourne R B, Chesworth B M, Davis A M, Mahomed N N, Charron KD. Patient satisfaction after total knee arthroplasty: who is satisfied and who is not? Clin Orthop Relat Res 2010; 468(1): 57-63. Bryan S, Goldsmith L J, Davis J C, Hejazi S, MacDonald V, McAllister P, Randall E, Suryaprakash N, Wu A D, Sawatzky. Revisiting patient satisfaction following total knee arthroplasty: a longitudinal observational study. BMC Musculoskelet Disord 2018; 19(1): 423. Conner-Spady B L, Bohm E, Loucks L, Dunbar M J, Marshall D A, Noseworthy T W. Patient expectations and satisfaction 6 and 12 months following total hip and knee replacement. Qual Life Res 2020; 29(3): 705-19. Dunbar M J, Richardson G, Robertsson O. I can’t get no satisfaction after my total knee replacement: rhymes and reasons. Bone Joint J 2013; 95-B(11 Suppl A): 148-52. Ghomrawi H M K, Lee L Y, Nwachukwu B U, Jain D, Wright T, Padgett D, Bozic K J, Lyman S. Preoperative expectations associated with postoperative dissatisfaction after total knee arthroplasty: a cohort study. J Am Acad Orthop Surg 2020; 28(4): e145-e150. Gunaratne R, Pratt D N, Banda J, Fick D P, Khan R J K, Robertson B W. Patient dissatisfaction following total knee arthroplasty: a systematic review of the literature. J Arthroplasty 2017; 32(12): 3854-60. Namin A T, Jalali M S, Vahdat V, Bedair H S, O’Connor M I, Kamarthi S, Isaacs J A. Adoption of new medical technologies: the case of customized individually made knee implants. Value Health 2019; 22(4): 423-30. Robertsson O, Dunbar M, Pehrsson T, Knutson K, Lidgren L. Patient satisfaction after knee arthroplasty: a report on 27,372 knees operated on between 1981 and 1995 in Sweden. Acta Orthop Scand 2000; 71(3): 262-7. Robinson P G, Clement N D, Hamilton D, Blyth M J G, Haddad F S, Patton J T. A systematic review of robotic-assisted unicompartmental knee arthroplasty: prosthesis design and type should be reported. Bone Joint J 2019; 101-B(7): 838-47. Tilbury C, Haanstra T M, Leichtenberg C S, Verdegaal S H, Ostelo R W, de Vet H C, Nelissen R G, Vliet Vlieland T P. Unfulfilled expectations after total hip and knee arthroplasty surgery: there is a need for better preoperative patient information and education. J Arthroplasty 2016; 31(10): 213945. Vissers M M, Bussmann J B, Verhaar J A, Busschbach J J, Bierma-Zeinstra S M, Reijman M. Psychological factors affecting the outcome of total hip and knee arthroplasty: a systematic review. Semin Arthritis Rheum 2012; 41(4): 576-88. Vissers M M, de Groot I B, Reijman M, Bussmann J B, Stam H J, Verhaar J A. Functional capacity and actual daily activity do not contribute to patient satisfaction after total knee arthroplasty. BMC Musculoskelet Disord 2010; 11: 121.

Acta Orthopaedica 2020; 91 (3): 243–245

243

Guest editorial

ERAS guidelines for hip and knee replacement – need for reanalysis of evidence and recommendations ? In a laudable effort, 8 multidisciplinary authors produced a consensus statement for perioperative care in hip and knee replacement (THA/TKA) recently published in Acta Orthopaedica. The primary goal was to evaluate “efficacy of individual items of the perioperative treatment pathway to expedite the achievement of discharge criteria” (Wainwright et al. 2020). In this context, the consensus group addressed 17 topics that were assessed based on a proposed systematic review of the literature from January 1966 to October 2018, with the purpose to reflect evidence at the time of writing in January 2019. While this type of effort is important in order to provide the wider perioperative and anesthesia community with guidance, its publication comes with a great deal of responsibility, as practitioners will use this information to model their practice. Therefore, clinicians need to not only read the conclusions, but be aware of the sources of evidence they are based on. In this context, it must be acknowledged that this consensus document provides an enormous amount of valuable information, however, a number of points should be considered. Specifically, a closer look at the 17 considered recommendations may lead to questions regarding the search methodology as well as the concept of including so many factors, which may be of limited importance in fulfilling the primary aim of the consensus report, namely the earlier achievement of discharge criteria. It must be asked if the inclusion of this large number of interventions is indeed feasible and/or justifiable for such a project. This concern is based on the fact that the information considered was to a large extent not procedure specific nor was it – in many cases – derived from studies conducted during a time that is compatible with documented modern ERAS clinical practice in fasttrack THA/TKA. Another factor to consider is that the group’s focus was heavily weighted towards specific clinical studies, thus frequently ignoring information gained from other literature sources, such as population-based investigations, which when considered can provide valuable information. As to the proposed recommendations, preadmission patient optimization regarding smoking and alcohol, although arguably appropriate, is not based on hard scientific, procedurespecific data. Indeed, much of the information is based on transferable evidence from other procedures and not supported from the few fully implemented THA/TKA ERAS data. Preoperative information is obviously to be implemented

according to common practice in most countries, but is not a unique procedure-specific item on the specified discharge criteria or documented to enhance discharge in THA/TKA. Preoperative fasting has been emphasized repeatedly in all previous ERAS guidelines. Further, it is a well-established fact that oral intake does not have to be withheld for long. However, there is no procedure-specific evidence from THA/ TKA where the patient in modern practice care is admitted on the morning before surgery. The same applies to preoperative carbohydrate treatment, which is not recommended as an essential or routine intervention. To include pre-anesthetic anxiolytic medication is neither procedure-specific nor evidence-based and is obviously not necessary in other types of modern surgery. The recommendations for general vs. neuraxial anesthesia are discussed rather superficially and lead to the conclusion that either technique can be used. Here, the recommendation conflicts with recent (and previous) data suggesting that neuraxial approaches may be preferable (Memtsoudis et al. 2019a). However, the latter conclusion is one that stems from the evaluation of a much broader literature base, including population data that considers complication risk. The topics of spinal opioids and epidurals get to the right conclusion not to be recommended, although the literature basis for this suggestion is superficially presented with regard to side effects. Additionally, the important and large topic of pain management, which is crucial to achieving early mobilization and shortening discharge time (the primary aim for the guideline) is mostly based on old references in which a length of stay (LOS) > 2 days was prevalent in most studies. This fact and the rather limited literature review does therefore not seem to appropriately support the current clinical practice. The same applies to the multimodal analgesia section which also suffers from the surprising lack of discussion of the use of preoperative high-dose steroids (2 references mentioned, of which 1 is not with high-dose) despite the availability of several systematic reviews and RCT’s with rather promising results published before manuscript submission. Further, there is a lack of references to critical systematic reviews on pain management (Hojer Karlsen et al. 2015, Karlsen et al. 2017) and an absence of discussion of whether these data apply to a fast-track THA/TKA setup. Next, the section of surgical traditions that potentially impair early recovery such a urinary catheterization, use of tourni-

244

quets, drains, etc. is appropriately discussed although again with limited references. This is important as these care principles are still widely used and should be eliminated – as quoted – to achieve a shorter LOS and recovery time. The section on normothermia is not based on THA/TKA data and represents standard of care in modern surgical practice. The same applies to antibiotic prophylactic practice which is irrelevant for achieving early discharge criteria, but obviously important for post-discharge outcome. The issue of anti-thrombotic prophylaxis, which presently is controversial and ranges from weeks of prophylaxis with expensive and potentially side-effect laden modern potent anticoagulants vs. the simple use of Aspirin, falls into a similar category. Importantly, the anti-thrombotic prophylaxis discussion is turning away from the primary issue of the guideline to shorten the time before discharge. The recommendation for early return to normal diet is – as mentioned – a component of all ERAS pathways, but specifically of limited relevance to THA and TKA procedures, which increasingly can be performed on an outpatient or 1-day basis. The section on continuous improvement and audit is obvious and important, but is again not procedure-specific with regards to recommendations for an optimal THA/TKA ERAS program. Although, the recommended “interventions” have 17 stronggrade recommendations, only the ones suggesting the use of local infiltration analgesia in TKA, tranexamic acid, multimodal opioid-sparing analgesia and early mobilization are “active” components. In contrast, most of the other 17 components are not based on new scientific data to show that they may lead to a reduced LOS from previously 6–8 days to less than 2 days, or even to be performed on an outpatient basis. Finally, busy clinicians may have a hard time interpreting the many guidelines on perioperative interventions in THA/ TKA. For example, the recently published US guidelines (available as Epub in June/July 2018, but in print in March 2019) (Soffin et al. 2019a, b) are quite different from the present THA/TKA recommendations (Wainwright et al. 2020), although the former also suffer from the lack of a critical literature update (Kehlet and Joshi 2019). In summary, despite the major and laudable efforts made by the multidisciplinary team when putting together this important information, there is still room for improvement with a more critical reanalysis of the data. The future discussion should focus on which components of care are most important in order to achieve a length of stay between 0 and 2 days, including a more important evaluation of the pathophysiology of recovery in general and in THA and TKA patients with certain characteristics in particular (high-pain responders, high inflammatory responders, psychosocial factors, etc.) (Memtsoudis et al. 2019b, Wainwright and Kehlet 2019). In addition, since major progress has been achieved with ERAS programs in THA and TKA, as evidenced by the reported shortening of LOS and several types of complications, the major questions that remain to be answered include the issue of post-discharge functional recovery, which to date is addressed by a paucity

Acta Orthopaedica 2020; 91 (3): 243–245

of literature (but apparently was not the focus of the article). However, post-discharge recovery is per definition part of the concept of “enhanced recovery after surgery”. In this context, there is not only a lack of good scientific data, but a need for updated guidelines for post-discharge functional recovery (Wainwright and Kehlet 2019). Hopefully, future procedure-specific ERAS guidelines will be divided more specifically into items that constitute transferable evidence/routine perioperative care in modern surgery vs. well-documented interventions to achieve an improved and fully implemented procedure-specific ERAS program to enhance total recovery, shorten LOS, and decrease complications (Kehlet and Joshi 2019). This would allow groups of experts to focus their analytic efforts and discussions on procedure specific issues and address the many challenges that still remain unaddressed. Finally, ERAS recommendations may benefit from a clearer separation of components that are necessary for early recovery and the reduction of LOS vs. those that need to be considered for post-discharge functional recovery. Henrik Kehlet1 and Stavros G Memtsoudis2 1 Rigshospitalet,

Section of Surgical Pathophysiology, Copenhagen, Denmark E-mail: henrik.kehlet@regionh.dk 2 Department of Anesthesiology, Critical Care, and Pain Management, Hospital for Special Surgery, New York, USA Department of Healthcare Policy and Research, Weill Cornell Medical College, New York, USA Department of Anesthesiology, Perioperative Medicine and Intensive Care Medicine, Paracelsus Medical University, Salzburg, Austria

Hojer Karlsen A P, Geisler A, Petersen P L, Mathiesen O, Dahl J B. Postoperative pain treatment after total hip arthroplasty: a systematic review. Pain 2015; 156(1): 8-30. Karlsen A P, Wetterslev M, Hansen S E, Hansen M S, Mathiesen O, Dahl J B. Postoperative pain treatment after total knee arthroplasty: A systematic review. PLoS One 2017; 12(3): e0173107. Kehlet H, Joshi G P. Anesthesia in enhanced recovery pathways for hip and knee arthroplasty: where is the evidence? Anesth Analg 2019; 128 (4): e52. Memtsoudis S G, Cozowicz C, Bekeris J, Bekere D, Liu J, Soffin E M, Mariano E R, Johnson R L, Hargett M J, Lee B H, Wendel P, Brouillette M, Go G, Kim S J, Baaklini L, Wetmore D, Hong G, Goto R, Jivanelli B, Argyra E, Barrington M J, Borgeat A, De Andres J, Elkassabany N M, Gautier P E, Gerner P, Gonzalez Della Valle A, Goytizolo E, Kessler P, Kopp S L, Lavand’Homme P, MacLean C H, Mantilla C B, MacIsaac D, McLawhorn A, Neal J M, Parks M, Parvizi J, Pichler L, Poeran J, Poultsides L A, Sites B D, Stundner O, Sun E C, Viscusi E R, Votta-Velis E G, Wu C L, Ya Deau J T, Sharrock N E. . Anaesthetic care of patients undergoing primary hip and knee arthroplasty: consensus recommendations from the International Consensus on Anaesthesia-Related Outcomes after Surgery group (ICAROS) based on a systematic review and meta-analysis. Br J Anaesth 2019a; 123(3): 269-87.

Acta Orthopaedica 2020; 91 (3): 243â&#x20AC;&#x201C;245

Memtsoudis S G, Poeran J, Kehlet H. Enhanced recovery after surgery in the United States: from evidence-based practice to uncertain science? JAMA 2019b; 321(11): 1049-50. Soffin E M, Gibbons M M, Ko C Y, Kates S L, Wick E, Cannesson M, Scott M J, Wu C L. Evidence review conducted for the Agency for Healthcare Research and Quality Safety Program for Improving Surgical Care and Recovery: focus on anesthesiology for total knee arthroplasty. Anesth Analg 2019a; 128(3): 441-53. Soffin E M, Gibbons M M, Ko C Y, Kates S L, Wick E C, Cannesson M, Scott M J, Wu C L. Evidence review conducted for the Agency for Healthcare

245

Research and Quality Safety Program for Improving Surgical Care and Recovery: focus on anesthesiology for total hip arthroplasty. Anesth Analg 2019b; 128(3): 454-65. Wainwright T W, Kehlet H. Fast-track hip and knee arthroplasty â&#x20AC;&#x201C; have we reached the goal? Acta Orthop 2019; 90(1): 3-5. Wainwright T W, Gill M, McDonald D A, Middleton R G, Reed M, Sahota O, Yates P, Ljungqvist O. Consensus statement for perioperative care in total hip replacement and total knee replacement surgery: Enhanced Recovery After Surgery (ERAS) Society recommendations. Acta Orthop 2020; 91(1): 3-19.

246

Acta Orthopaedica 2020; 91 (3): 246–253

Inferior stabilization of cementless compared with cemented dualmobility cups in elderly osteoarthrosis patients: a randomized controlled radiostereometry study on 60 patients with 2 years’ follow-up Steffan TABORI-JENSEN 1,2, Sebastian Breddam MOSEGAARD 1,2, Torben B HANSEN 1,2, and Maiken STILLING 1,2 1 University

Clinic for Hand, Hip and Knee Surgery, Regional Hospital West Jutland, Holstebro; 2 Department of Clinical Medicine, Aarhus University, Aarhus, Denmark Correspondence: steffan.jensen@rm.dk Submitted 2019-06-04. Accepted 2019-12-09.

Background and purpose — Elderly patients may benefit from a dislocation low-risk dual-mobility (DM) articulation in total hip arthroplasty, but the best cup fixation method is unknown. We compared cup migration for cemented and cementless DM cups using radiostereometry. Patients and methods — In a patient-blinded randomized trial, 60 patients (33 female) with osteoarthritis were allocated to cemented (n = 30) or cementless (n = 30) Avantage DM cup fixation. Criteria were age above 70 years, and T-score above –4. We investigated cup migration, periprosthetic bone mineral density (BMD), and patient-reported outcome measures (PROMs) until 24 months postoperative follow-up. Results — At 24 months mean proximal cup migration was 0.11 mm (95% CI 0.00–0.23) for cemented cups and 0.09 mm (CI –0.09 to 0.28) for cementless cups. However, cementless cups generally migrated more than cemented cups at 12 and 24 months. Cemented cups had no measurable migration from 3 months’ follow-up, while cementless cups had not yet stabilized at 24 months in all rotations. Cementless cups showed statistically significantly more maximum total point motion (MTPM) at 12- and 24-month follow-up compared with cemented cups in patients with low systemic BMD (p = 0.01). Periprosthetic BMD changes did not statisticially significantly correlate to proximal migration in either cup fixation group (p > 0.05). PROMs improved similarly in both groups. Interpretation — Cemented cups were well fixed at 3 months. The cementless cups migrated more in patients with low BMD, showed an inconsistent pattern of migration, and migrated in different directions during the first and second year without tendency to stabilization. Cemented fixation of the Avantage DM cup seems safer in elderly patients

The most common indication for revision of a conventional primary total hip arthroplasty (THA) is aseptic loosening of the components (SHAR 2016, NJR 2017, DHAR 2018). Implant fixation method (i.e., cemented or cementless) in primary THA seems mainly based on the surgeon’s preference and national trends. The Danish Hip Replacement Registry report shows a decrease in the use of cemented cup fixation in osteoarthrosis (OA) patients above 70 years (DHAR 2018). This trend has also been described in the United Kingdom (UK) and Australian Joint Registries, while in Sweden and Norway cemented cup fixation is still the preferred fixation method in elderly patients (SHAR 2016, NAR 2017, NJR 2017). The dual-mobility (DM) concept, with 2 articulation surfaces and increased jump distance, may decrease the dislocation rate and increase range of motion compared with standard single mobility (SM) THAs. The long-term survival and the best fixation method of the newer Avantage Reload DM cup in elderly patients is currently unknown but retrospective studies on other types of primary DM THAs suggest acceptable survival rates (Batailler et al. 2017) . Excessive early (2-year) implant micromotion measured with radiostereometric analysis (RSA) is a strong predictor for later implant loosening and poor survival (Karrholm et al. 1997, Nieuwenhuijse et al. 2012, Pijls et al. 2012), and our primary aim was to investigate the early RSA-measured migration of cemented and cementless Avantage DM cups in elderly (> 70 years old) OA patients until 24 months’ followup. Secondary endpoints included systemic and periprosthetic bone mineral density (BMD) measurements, and clinical outcome scores.

Acta Orthopaedica 2020; 91 (3): 246–253

ENROLLMENT

247

Assessed for eligibility n = 90 Excluded (n = 30): – not meeting inclusion citeria, 2 – declined to participate, 28 Randomized n = 60

ALLOCATION

Allocated to cemented DMC (n = 30) Received allocated interventiion (n = 30)

Allocated to cementless DMC (n = 30) Received allocated interventiion (n = 30)

FOLLOW-UP

Excluded due to preoperative T-score –4.3 n=1

Excluded RSA data due to poor marker distribution n=2 ANALYSIS

Analyzed with RSA: – 3 months, 29 – 1 year, 29 – 2 years, 28

Analyzed with RSA: – 3 months, 28 – 1 year, 28 – 2 years, 28

Analyzed with DXA: – 3 months, 28 – 1 year, 28 – 2 years, 29

Analyzed with DXA: – 3 months, 27 – 1 year, 30 – 2 years, 30

PROM measurements: – preoperative, 29 – postoperative, 29 – 3 months, 29 – 1 year, 29 – 2 years, 29

PROM measurements: – preoperative, 30 – postoperative, 30 – 3 months, 30 – 1 year, 30 – 2 years, 30

Figure 1. CONSORT flow diagram showing the inclusion/exclusion process until 2-year follow-up.

Patients and methods Design and patients Between November 2014 and January 2018, we performed a Level I prospective, randomized, patient-blinded, parallel group trial at Regional Hospital West Jutland, Holstebro, Denmark. Inclusion criteria were primary osteoarthrosis, age above 70 years, informed consent, and only 1 hip operated and with adequate bone quality for insertion of a cementless acetabular component as judged by the surgeons on radiographs and intraoperatively. Exclusion criteria were vascular or neuromuscular disease in the operated leg, fracture sequelae, avascular necrosis of the femoral head, alcohol abuse, and severe osteoporosis (T-score ≤ –4.0). 60 patients (33 female) were included and block randomized (using a computerized algorithm) to either cemented (n = 30) or cementless (n = 30) cup fixation (Figure 1). There is no “acceptance limit” for “low systemic BMD/T-score.” We arbitrarily chose a T-score limit of –4.0 and thus excluded patients with severe osteoporosis. Prosthesis, surgery, and rehabilitation The Avantage Reload cemented and cementless DM cup (ZimmerBiomet, Warsaw, IN, USA) has been commercially available since 2005. The external surface of the cemented

Avantage Reload metal shell has a bright polish (Ra max 0.4 µm), and the inner articulate surface is highly polished. Vacuum-mixed Palacos R+G bone cement (Heraeus Medical, Wehrheim, Germany) was used for cemented fixation. The cementless Avantage Reload metal shell has a double coating with a projection vacuum plasma (VPS) titanium coating (Ra > 15 µm) and synthetic hydroxyapatite (HA) (150 ± 50 µm) to create a rough surface finish (Ra > 11 µm). Exeter highlypolished stems (Stryker Corporation, Kalamazoo, MI, USA) with vacuum mixed Palacos R+G bone cement (Heraeus Medical, Wehrheim, Germany) were used in all patients. A 28-mm chrome-cobalt femoral head was used in all cases. Vitamin E-infused highly cross-linked polyethylene liner (GUR 1050) was used in both cemented and cementless cups. All liners were vacuum-packed and gamma sterilized with a minimum of 25 kGy. All patients were operated by 1 of 2 highly experienced orthopedic hip surgeons. The sequentially numbered sealed envelopes were hidden from investigators until directly prior to surgery to prevent bias. On the day of surgery, a sealed randomization envelope was opened to allocate the patient to either cemented or cementless cup fixation. Prophylactic cefuroxime 1.5 g was administrated intravenously before surgery and twice postoperatively with an 8-hour interval. After bone preparation, 6–8 tantalum beads (1 mm) were inserted into the periacetabular bone during surgery. All patients were operated by a posterolateral approach and received the same rehabilitation program, allowing full weight-bearing immediately after surgery. Radiostereometric analysis Stereoradiographs were obtained within the first postoperative 2 days (mean 1.1, range 1–14) and at 3, 12, and 24 months after surgery. All examinations were performed with the patient in a supine position with a uniplanar calibration box (Carbon Box 19, RSAcore, Leiden, The Netherlands) located underneath the examination table. The anatomical axis of the leg was parallel to the y-axis of the calibration box. Cup migration was evaluated on all 3 follow-up stereoradiographs with the postoperative stereoradiograph as the baseline reference. The radiostereometric analysis was performed with ModelBased RSA version 4.10 software (RSAcore, Leiden, The Netherlands) using computer-aided design (CAD) implant models provided by the manufacturer (ZimmerBiomet, Warsaw, IN, USA). We measured cup migration (center of 3D model points) in the coordinate system of the calibration box as described in the Guidelines for RSA of Implants (Valstar et al. 2005) (Figure 2). Total translation (TT) and total rotation (TR) were both calculated using Pythagoras’ theorem (sqrt (x2 + y2 + z2). The mean condition number (CN) of the bone marker model was 83 (SD 47) and the rigid body error (ME) was mean 0.24 (SD 0.06). A minimum of 3 bone markers was accepted and the cut-off points for CN and ME were maintained at 150 and 0.35, respectively (Valstar et al. 2005).

248

Acta Orthopaedica 2020; 91 (3): 246–253

Proximal (+)

Posterior (–)

Anterior tilt (+)

Lateral (–)

Medial (+)

Adduction (+)

Anterior (+)

Z Anteversion (+)

Distal (–)

Figure 2. Illustration of directions, translation, and rotations for Avantage DM cup.

Figure 3. Wilkinson regions of interest (ROI) 1–4. Only the bone areas within yellow lines are included in the analysis.

All patients were subject to double examinations at the 3-month RSA examination according to guidelines (Table 1, see Supplementary data) (Valstar et al. 2005). The position of the fitted implant CAD model on the postoperative stereoradiograph pose-estimation served as inclination and anteversion estimates and were read from the ModelBased RSA software (RSAcore, Leiden, The Netherlands).

Statistics and sample size The primary endpoint was proximal cup migration at the 24-month follow-up (Pijls et al. 2012). Linear mixed-effect models were used to determine whether cup fixation and BMD had a significant effect on cup migration at 3-, 12-, and 24-month follow-up. This was used as it takes the correlation of measurements on the same patient into account and includes all patients/missing values effectively. The analysis was modelled as a function of fixation with the interaction of time and fixation as fixed effects. Bonferroni correction was applied to migration data (primary end-point), when linear mixed models showed significant p-values. Model estimates are reported as means with 95% confidence intervals (CI). The secondary endpoints were measurements of periprosthetic BMD, clinical outcomes of HHS, OHS, and EQ-5D, and VAS (rest and activity) for pain. Subgroup analyses (mixed model) were performed between cup fixation (cemented/cementless) and cup migration when stratified to normal (T-score ≥ –1.0) or low (T-score < –1.0) preoperative BMD. Student’s t-test was used for normally distributed data. When data were not normally distributed according to a Shapiro–Wilks test, a non-parametric (Mann–Whitney) test was used. Data were analyzed as of the date of the last data collection (January 2018). Proximal cup acceptance migration thresholds at 24 months’ follow-up was assessed according to Pijls et al. (2012). There are no previous migration data for the Avantage DM cup that could be used for sample size calculation. The pilot study included both cemented and cementless cups in patients (n = 5) older than 70 years. A pre-study sample size calculation using 2-sample mean test for a minimal relevant proximal cup migration difference of 0.2 mm (Pijls et al. 2012) with a mean cup migration of 0.2 mm (SD 0.27) (pilot study), power 80%, alpha 0.05, estimated 29 patients in each group. 30 patients per group were included to account for dropout. Statistical significance was set at 0.05. Stata version 13.1 (StataCorp, College Station, TX, USA) was used for statistical analysis.

Dual-energy X-ray absorptiometry (DXA) scans Preoperatively all patients underwent spine and dual hip DXA scan to determine systemic BMD and T-score. The mean time from preoperative DXA scan to surgery was 20 days (14–71). Postoperatively (within 4 days after surgery) and at 3, 12, and 24 months after surgery, quantitative measurements of the periprosthetic BMD (g/cm2) was acquired with DXA scans using a GE Lunar iDXA scanner (General Electric, Chicago, IL, USA), and analyses were performed using enCORE version 16 software (https://www.encore.com/). Patients were placed in supine position with the body parallel to the examination table and the feet fixed to a device that kept the halluces pointing straight up. Analysis was performed according to Wilkinson 4 regions of interest (ROI) and precision ranged from 3% to 13% in cemented cup fixation and 3% to 6% in cementless cup fixation (Table 2, see Supplementary data) (Wilkinson et al. 2001). A template was applied to the baseline scan, and the ROIs were subsequently copied to align with the bone-border on follow-up scans. ROI 2 and 3 were adjusted in height on the baseline scan depending on the cup size (each ROI was 1 half cup height) and ROIs 1 and 4 had fixed sizes (Figure 3). Clinical outcome measures and complications Clinical outcome measures were assessed by Harris Hip Score (HHS), Oxford Hip Score (OHS) (Paulsen et al. 2012), patient-reported quality of life (EQ-5D) (Brooks 1996), and visual analog scale (VAS) for hip pain preoperatively and at 3, 12, and 24 months after surgery.

Acta Orthopaedica 2020; 91 (3): 246–253

249

Table 3. Descriptive baseline characteristics of the patients, implants, and surgery. Values are mean (range) unless otherwise specified

Table 4. Translations along and rotations about the x-, y-, and z-axis for cemented and cementless cups presented as mean (95% CI) Axis

Cemented Cementless (n = 29) (n = 30)

Sex, male/female, n Age at operation Implant side, right/left, n Cup size (mm) Cup inclination (°) Cup anteversion (°) Preoperative T-score BMI ASA class

14/15 75 (70–82) 16/13 48.7 (44–54) 49.2 (36.2–61.0) 11.5 (1.2–26.2) –1.01 (–2.9–1.8) 28 (23–39) 2.0 (1–3)

Cemented Cementless (n = 28) (n = 28)

Translations (mm) x-axis 3 months –0.01 (–0.17 to 0.14) 0.08 (–0.19 to 0.36) 12 months –0.03 (–0.21 to 0.15) 0.16 (–0.20 to 0.51) 24 months –0.01 (–0.22 to 0.20) 0.23 (–0.20 to 0.66) y-axis 3 months 0.08 (0.00 to 0.16) 0.15 (0.02 to 0.27) 12 months 0.09 (0.01 to 0.18) 0.12 (–0.02 to 0.26) 24 months 0.11 (0.00 to 0.23) 0.09 (–0.09 to 0.28) z-axis 3 months 0.16 (0.00 to 0.32) 0.31 (0.00 to 0.62) 12 months 0.15 (–0.01 to 0.31) 0.36 (0.03 to 0.69) 24 months 0.23 (0.02 to 0.44) 0.39 (0.03 to 0.75) Total translation 3 months 0.49 (0.34 to 0.64) 0.79 (0.49 to 1.10) 12 months 0.56 (0.37 to 0.76) 0.88 (0.51 to 1.25) 24 months 0.65 (0.44 to 0.87) 0.98 (0.54 to 1.42) Rotations (°) x-axis 3 months 0.34 (0.01 to 0.66) 0.01 (–0.48 to 0.51) 12 months 0.52 (0.15 to 0.89) 0.64 (–0.01 to 1.30) 24 months 0.29 (–0.05 to 0.63) 0.04 (–0.63 to 0.70) y-axis a 3 months 0.23 (0.26 to 0.72) 1.08 (0.34 to 1.82) 12 months 0.30 (–0.25 to 0.85) 1.74 (0.91 to 2.57) b 24 months 0.18 (–0.37 to 0.73) 1.10 (0.42 to 1.78) z-axis 3 months –0.35 (–0.60 to 0.03) –0.07 (–0.60 to 0.46) 12 months –0.40 (–0.75 to –0.05) –0.33 (–0.92 to 0.26) 24 months –0.35 (–0.76 to 0.05) –0.01 (–0.69 to 0.68) Total rotation a 3 months 1.52 (1.12 to 1.90) 2.23 (1.55 to 2.92) 12 months 1.80 (1.40 to 2.24) 3.00 (2.20 to 3.80) b 24 months 1.72 (1.30 to 2.13) 2.57 (1.83 to 3.30) MTPM (mm) a 3 months 1.14 (0.86 to 1.42) 1.81 (1.26 to 2.36) 12 months 1.30 (1.00 to 1.60) 2.24 (1.64 to 2.85) b 24 months 1.36 (1.00 to 1.73) 2.16 (1.44 to 2.87)

13/17 75 (70–83) 15/15 52.8 (48–58) 43.5 (28.9–59.7) 11.7 (0.7–26.3) –1.12 (–3.1–2.3) 29 (22–38) 1.8 (1–3)

Ethics, funding, and potential conflicts of interest The study was conducted in accordance with the Helsinki Declaration. Patients gave informed consent before entering the study. The study was approved by the Central Danish Regional Committees on Biomechanical Research Ethics (Journal no. 1-10-72-209-14) and the study was registered with ClinicalTrials.gov (NCT02404727). ZimmerBiomet Inc. and The Danish Rheumatism Association supported the study financially but had no influence on the manuscript or publication. The authors have no conflicts of interest.

Results The baseline demographics of all patients are presented in Table 3. Radiostereometric analysis Translations and rotations, including TT, TR, and MTPM (mean and CI), are presented in Table 4, and statistically significant migrations are presented in Figure 4. Cemented cups showed no statistically significant translation (p > 0.3) or rota-

MTPM: maximum total point motion. a Denotes Bonferroni adjusted p-values for multiple testing. b Statistically significant difference between cemented and cementless cup fixation.

Mean y-axis rotation (°) with 95% CI

Mean total rotation (°) with 95% CI

Mean MTPM (mm) with 95% CI

Cemented Cementless

p < 0.01

p = 0.02

p < 0.01

–1 Post- 3 operative

Follow-up, months

–1 Post- 3 operative

Follow-up, months

–1 Post- 3 operative

Figure 4. Significant migration in cementless cups compared with cemented in y-axis, TR, and MTPM.

Follow-up, months

250

Acta Orthopaedica 2020; 91 (3): 246–253

Table 5. Cup rotations and MTPM between follow-ups within each cup fixation group presented as mean difference (95% CI) Rotations (°)

Cemented

Cementless

x-axis 3–12 months –0.18 (–0.49 to 0.13) –0.63 (–0.95 to 0.31) a 12–24 months 0.23 (–0.08 to 0.55) 0.61 (0.28 to 0.93) a y-axis 3–12 months –0.07 (–0.44 to 0.29) –0.66 (–1.03 to 0.28) a 12–24 months 0.14 (–0.22 to 0.51) 0.64 (0.26 to 1.01) a z-axis 3–12 months 0.09 (–0.15 to 0.32) 0.26 (0.01 to 0.50) 12–24 months –0.08 (–0.32 to 0.15) –0.33 (–0.60 to –0.08) a Total translation 3–12 months –0.25 (–0.62 to 0.12) –0.75 (–1.13 to 0.36) a 12–24 months 0.07 (–0.31 to 0.44) 0.42 (0.04 to 0.80) a a Statistically significant within-group difference from one follow-up to next follow-up.

tion (p > 0.2) during the 24-month follow-up time. Cementless cups had no statistically significant translations (p > 0.20) during 24-month follow-up, but showed rotation about all orthogonal axes and in TR during the 24-month followup (Table 5). Cemented cups migrated 0.54 (CI 0.44–0.65) and cementless cups migrated 1.08 (CI 0.71–1.46) MTPM between 12 months and 24 months with a mean difference of –0.54 (CI –0.94 to –0.14, (p = 0.01). By 24 months, 21/28 of the cemented cups showed proximal cup migration < 0.2 mm, 7/28 were between 0.2 and 1.0 mm, and no cemented cups had proximal cup migration > 1.0 mm. By 24 months, 18/28 of the cementless cups showed proximal cup migration < 0.2 mm, 9/28 were between 0.2 and 1.0 mm, and 1/28 cementless cup showed > 1.0 mm proximal cup migration. When stratifying patients into 2 subgroups based on preoperative systemic BMD (normal and low BMD), we found no within-subgroup difference in proximal cup migration between cemented and cementless cup fixation at any follow-up (p > 0.3; Figure 5, see Supplementary data). The mean 24-month proximal cup migration in the normal BMD group was 0.05 mm (CI –0.08– 0.18) for cemented cups and

Table 6. BMD change in the 4 ROIs around the acetabular component presented as mean (95% CI) percentage change from the baseline values at 3, 12, and 24 months ROI

Follow-up

Cemented

ROI 1 ROI 2 ROI 3 ROI 4

Postoperative 3 months 12 months 24 months Postoperative 3 months 12 months 24 months Postoperative 3 months 12 months 24 months Postoperative 3 months 12 months 24 months

ref 0 (–2 to 2) 1 (–1 to 3) 3 (1 to 5) ref –7 (–13 to –1) –1 (–6 to 6) –7 (–13 to –1) ref –5 (–10 to 0) –2 (–7 to 3) –1 (–6 to 4) ref –5 (–8 to –1) –7 (–10 to –3) –9 (–13 to –6)

Cementless ref –1 (–4 to 1) –2 (–4 to 0) –2 (–4 to 0) a ref –4 (–11 to 2) –12 (–19 to –6) a –11 (–17 to –4) ref 4 (–1 to 9) a 4 (–1 to 8) 4 (–1 to 9) ref –1 (–4 to 3) –2 (–6 to 2) –1 (–4 to 3) a

ROI: Regions of interest according to Wilkinson et al. (2001). a Statistically significant difference between cemented and cementless fixation.

Mean MTPM (mm) with 95% CI in normal BMD group

0.07 mm (CI –0.17–0.32) for cementless cups (Figure 5, see Supplementary data). Mean 24-month proximal cup migration in the low BMD group was 0.18 mm (CI 0.05–0.31) for cemented cups and 0.11 mm (CI –0.07–0.29) for cementless cups (Figure 5, see Supplementary data). The postoperative inclination angle was higher in cemented cups compared with cementless cups (p = 0.01) (Table 3). The postoperative anteversion angle did not differ between the two fixation methods (p = 0.9) (Table 3). Further analyses on BMD groups showed that MTPM was statistically significantly higher at 12- and 24-month follow-up in cementless cups compared with cemented cups in the low BMD group (p = 0.01, Figure 6), which could be explained by a higher cup migration in x-translation (p = 0.04 at 24 months), y- rotation (p < 0.001, p = 0.03, at 12 and 24 months respectively), and z-rotation (p = 0.04 at 24 months). Likewise, TT and TR was higher for cementless cups compared with cemented cups in the low BMD groups Mean MTPM (mm) with 95% CI in at 12 and 24 months (all p < 0.03). low BMD group

Cemented Cementless

p = 0.01

–1 Post- 3 operative

p = 0.01

Follow-up, months

–1 Post- 3 operative

Follow-up, months

Figure 6. MTPM migration in normal and low BMD groups based on cup fixation.

Percentage change in periprosthetic BMD Percentage BMD changes are presented in Table 6. We did not find a statistically significant correlation between percentage BMD change and proximal cup migration in cemented or cementless cups during follow-up (p > 0.06). However, this does not imply that there is no true correlation between percentage BMD change and proximal cup migration. The strongest correlation between proximal cup migration and percentage BMD was found at 24 months in the cementless group in ROI 4 (CI –0.68–0.00, p = 0.06).

Acta Orthopaedica 2020; 91 (3): 246–253

Clinical outcome measures and complications Clinical results (HHS, OHS, EQ-5D, VAS) were similar between the 2 fixation groups during 24-month follow-up (Table 7, see Supplementary data). The 2 groups showed similar clinical improvement from preoperative to 24-month follow-up. The biggest difference between fixation groups was found in VAS at activity, mean diff = 1.0 (CI 0.02–2.1, p > 0.07). 1 patient (cementless cup) underwent revision surgery (liner and femoral head exchange) 3 months after the index surgery due to an intraprosthetic dislocation. 2 weeks after revision, the patient underwent a 1-stage debridement, washout, femoral head and liner exchange, and antimicrobial treatment for 6 weeks due to deep infection (Staphylococcus aureus). Here after, the patient had a well-functioning hip and continued regular RSA follow-up. There were no large articulation dislocations in either fixation group during the follow-up period.

Discussion To our knowledge, this is the first RSA study of the DM articulation concept in elderly OA patients comparing cemented and cementless cup fixation in a randomized study. The key findings were similar proximal cup migration between fixation groups, but cementless cups migrated more on absolute measures, migrated more in patients with low BMD, and had not stabilized at 24 months, whereas cemented cups were stable from 3 months. Radiostereometric analysis The association between early high proximal cup migration and the elevated risk of aseptic cup loosening and later revision have previously been described (Nieuwenhuijse et al. 2012, Pijls et al. 2012). In relation to Pijls’ (2012) thresholds for proximal cup migration, we identified 7 cemented cups (range 0.23–0.71 mm) and 9 cementless cups (range 0.2–0.75 mm) “at risk” of later revision in our study, but we observed no cemented cups and 1 cementless cup (1.16 mm) with “unacceptable” proximal migration (Pijls et al. 2012). In relation to Nieuwenhuijse’s definition we observed no cups exceeding 1.76 mm proximal migration and 1 cementless cup (6.39°) with abduction (z-axis) above 2.53° (Nieuwenhuijse et al. 2012). Cementless cups are inserted by under-reamed technique, and the initial rim-fit may be lost over time resulting in a final bottoming in the acetabulum (Rohrl et al. 2004). However, we saw 1 cementless cup with large proximal migration, and in general no measurable translation over time in the cementless group. Cementless cups did, however, have more rotation overall, over time, and in opposite directions before and after 12 months, as compared with cemented cups. A clinical significance level in relation to rotation measures has not been established. The difference in cup y-rotation

251

was evident at 3 months and was probably an effect of final osseointegration. Cemented cups were inserted with significantly higher inclination angle compared with cementless cups, which may be explained by surgeons’ preference for free-hand insertion of cemented Avantage DM cups due to affection of the cup– cement interface before cement curing with disconnection of the cup-guide. However, our findings suggest that bone fixation of cemented cups is less sensitive to increased cup angulation compared with cementless cups. This is also in line with a study on all-poly cemented and cementless cups (Kadar et al. 2012). RSA evaluations of elderly patients with OA treated with a primary THA are scarce. Direct comparisons with previous RSA reports are difficult due to alternative ways of presenting data, methods of fixation, marked differences in patient demographics, implant design, surgical approach, and follow-up time. Based on 24-month proximal migration as an indicator for primary stability our findings for cemented and cementless fixation methods are comparable, and in many cases lower than reported in other studies on cemented and cementless cup fixation in primary THA (Lazarinis et al. 2014, Salemyr et al. 2015, Finnila et al. 2016). Radiostereometric analysis and preoperative BMD status Cementless ceramic-on-ceramic THA has shown statistically significantly and clinically relevant higher proximal migration and migration until 12-month follow-up in women with low BMD (lower T-score limit of –3.5) compared with women with normal BMD at 24-month follow-up (Finnila et al. 2016). These findings are inconsistent with the present study where we observed similar proximal migration in the normal BMD and low BMD groups and no migration within fixation groups of the BMD subgroups during follow-ups. The mean 24-month proximal migration in our cementless group with low BMD of 0.11 mm (CI –0.07 to 0.29) was lower than reported in the study by Finnilä et al. (2016) of 0.29 mm (CI 0.20–0.39), suggesting early initial proximal cup stability with Avantage DM cups, even in the low BMD group, and in both cemented and cementless cup fixation. However, in the low BMD group, cementless cups showed statistically significantly more migration in MTPM, x-axis translation, y-axis rotation, TT and TR compared with cemented cups, suggesting that cementless cup fixation is not preferable in patients with preoperative low BMD. Only 1 study reports proximal cup migration in cemented cups with stratification to normal and low BMD but unclear definition of osteoporosis makes direct comparison troublesome (Digas et al. 2004). Periprosthetic BMD measurements Differences in BMD change in cemented and cementless fixation may be a result of different load transfer mechanisms leading to different bone remodeling profile (Digas et al.

252

2006). In cementless cups, forces are transmitted sideways to the periphery, rather than proximally, which leads to reduced load transfer in the most cranial/proximal area (Digas et al. 2006, Lazarinis et al. 2014) with local bone resorption caused by stress-shielding. This might explain the greater bone loss observed in ROI 1 and 2 of cementless cups compared with cemented cups in our study. Conversely, the increased BMD in ROI 3 and lesser BMD reduction in ROI 4 in cementless cups compared with cemented could be due to the increased traction forces in cementless cups acting as a stimulus for preservation of bone or even increase in BMD (Salemyr et al. 2015). The percentage BMD changes in the 2 cup fixation methods did not correlate to proximal cup migration. Clinical outcome measures There was no statistically significant difference in postoperative clinical evaluations (quality of life measured by EQ-5D, or hip status measured by HHS and OHS) between cemented and cementless cup groups. The 2-year clinical evaluations of cemented and cementless fixation translates to either very good or excellent end-results (Nilsdotter and Bremander 2011). Cup fixation method in the elderly There is no clear consensus on the choice of the cemented or cementless cup fixation method in elderly patients, and registry reports from the UK, Australia, Sweden, Norway, and Denmark reveal no clear overall tendency regarding cup fixation methods in the elderly (SHAR 2016, NAR 2017, NJR 2017, DHAR 2018). While many registries report a tendency towards more cups being inserted with cementless fixation, their superiority is not supported in the literature (Troelsen et al. 2013, Makela et al. 2014). Limitations and strengths The strength of this study is the patient-blinded randomized controlled study design and a large group available for migration analysis. RSA is a validated surrogate measure of later implant loosening, but other complications, i.e., wear-induced osteolysis or fractures in the cement mantle, may not be detected with early RSA (Nieuwenhuijse et al. 2012). Linear mixed-model analysis enabled us to use all the available data. A high number of radiographs were available for analysis, and the number of patients with available 24-month RSA measurements was equal in both fixation groups. Only 2 patients in the cementless group were excluded due to poor marker distribution. In the cemented group 1 patient was excluded due to a mistake in identification of severe preoperative osteoporosis (preoperative T-score of –4.3) and the 24-month follow-up RSA radiographs of 1 patient were accidentally lost. BMD is typically higher around cemented cups due to opacity of the cement and therefore net BMD measurements are not direct comparable between cementless and cemented cups, and for this reason we compared percentage BMD changes (Jayasuriya and Wilkinson 2014). The high inclusion rate improves the external validity of

Acta Orthopaedica 2020; 91 (3): 246–253

our study, but results can be generalized only to patients older than age 70 years with a T-score above –4. Conclusion At minimum 2 years’ follow-up cemented and cementless DM cups had similar proximal mean migration below the recommended acceptance threshold, yet more cups in the cementless group migrated above the acceptance levels, which is of clinical importance. The cementless cups migrated more in patients with low BMD, showed an inconsistent pattern of migration, and migrated in different directions during the first and second year without tendency to stabilization. On the other hand, cemented cups were well fixed at 3 months. Cemented fixation of the Avantage DM cup seems safer in elderly patients; however, long-term studies are warranted. Supplementary data Tables 1, 2, 7, and Figure 5 are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674.2020.1720978

ST-J, TBH, and MS designed the study and validated the methods. ST-J performed all the follow-up examinations. SBM and ST-J performed the statistical analyses. ST-J wrote the initial manuscript, and all authors revised it. Note: All product names are for identification purposes only, and may be trademarks of their respective owners. Acta thanks Søren Kold and Johan Kärrholm for help with peer review of this study.

Batailler C, Fary C, Verdier R, Aslanian T, Caton J, Lustig S. The evolution of outcomes and indications for the dual-mobility cup: a systematic review. Int Orthop 2017; 41(3): 645-59. doi: 10.1007/s00264-016-3377-y. Brooks R. EuroQol: the current state of play. Health Policy 1996; 37(1): 53-72. DHAR. The Danish Hip Arthroplasty Registry. National annual report 2018; 2018. Digas G, Thanner J, Anderberg C, Karrholm J. Bioactive cement or ceramic/ porous coating vs. conventional cement to obtain early stability of the acetabular cup: randomised study of 96 hips followed with radiostereometry. J Orthop Res 2004; 22(5): 1035-43. doi: 10.1016/j.orthres.2003.09.012. Digas G, Karrholm J, Thanner J. Different loss of BMD using uncemented press-fit and whole polyethylene cups fixed with cement: repeated DXA studies in 96 hips randomized to 3 types of fixation. Acta Orthop 2006; 77(2): 218-26. doi: 10.1080/17453670610045948. Finnila S, Moritz N, Svedstro M E, Alm J J, Aro H T. Increased migration of uncemented acetabular cups in female total hip arthroplasty patients with low systemic bone mineral density: a 2-year RSA and 8-year radiographic follow-up study of 34 patients. Acta Orthop 2016; 87(1): 48-54. doi: 10.3109/17453674.2015.1115312. Jayasuriya R L, Wilkinson J M. Pelvic periprosthetic bone mineral density measurement around cemented vs cementless acetabular prostheses. J Clin Densitom 2014; 17(1): 116-20. doi: 10.1016/j.jocd.2013.09.005. Kadar T, Furnes O, Aamodt A, Indrekvam K, Havelin L I, Haugan K, Espehaug B, Hallan G. The influence of acetabular inclination angle on the penetration of polyethylene and migration of the acetabular component: a prospective, radiostereometric study on cemented acetabular components. J Bone Joint Surg Br 2012; 94(3): 302-7. doi: 10.1302/0301-620x.94b3.27460.

Acta Orthopaedica 2020; 91 (3): 246â&#x20AC;&#x201C;253

Karrholm J, Herberts P, Hultmark P, Malchau H, Nivbrant B, Thanner J. Radiostereometry of hip prostheses: review of methodology and clinical results. Clin Orthop Relat Res 1997; (344): 94-110. Lazarinis S, Milbrink J, Mattsson P, Mallmin H, Hailer N P. Bone loss around a stable, partly threaded hydroxyapatite-coated cup: a prospective cohort study using RSA and DXA. Hip Int 2014; 24(2): 155-66. doi: 10.5301/ hipint.5000104. Makela K T, Matilainen M, Pulkkinen P, Fenstad A M, Havelin L, Engesaeter L, Furnes O, Pedersen A B, Overgaard S, Karrholm J, Malchau H, Garellick G, Ranstam J, Eskelinen A. Failure rate of cemented and uncemented total hip replacements: register study of combined Nordic database of four nations. BMJ 2014; 348: f7592. doi: 10.1136/bmj.f7592. NAR. Norwegian National Registry for Arthroplasty annual report 2017; 2017. Nieuwenhuijse M J, Valstar E R, Kaptein B L, Nelissen R G. Good diagnostic performance of early migration as a predictor of late aseptic loosening of acetabular cups: results from ten years of follow-up with Roentgen stereophotogrammetric analysis (RSA). J Bone Joint Surg Am 2012; 94(10): 874-80. doi: 10.2106/JBJS.K.00305. Nilsdotter A, Bremander A. Measures of hip function and symptoms: Harris Hip Score (HHS), Hip Disability and Osteoarthritis Outcome Score (HOOS), Oxford Hip Score (OHS), Lequesne Index of Severity for Osteoarthritis of the Hip (LISOH), and American Academy of Orthopedic Surgeons (AAOS) Hip and Knee Questionnaire. Arthritis Care Res (Hoboken) 2011; 63(Suppl 11): S200-7. doi: 10.1002/acr.20549. NJR. National Joint Registry for England, Wales, Northern Ireland and the Isle of Man. Annual report 2017; 2017.

253

Paulsen A, Odgaard A, Overgaard S. Translation, cross-cultural adaptation and validation of the Danish version of the Oxford hip score: assessed against generic and disease-specific questionnaires. Bone Joint Res 2012; 1(9): 225-33. doi: 10.1302/2046-3758.19.2000076. Pijls B G, Nieuwenhuijse M J, Fiocco M, Plevier J W, Middeldorp S, Nelissen R G, Valstar E R. Early proximal migration of cups is associated with late revision in THA: a systematic review and meta-analysis of 26 RSA studies and 49 survival studies. Acta Orthop 2012; 83(6): 583-91. doi: 10.3109/17453674.2012.745353. Rohrl S M, Nivbrant B, Strom H, Nilsson K G. Effect of augmented cup fixation on stability, wear, and osteolysis: a 5-year follow-up of total hip arthroplasty with RSA. J Arthroplasty 2004; 19(8): 962-71. Salemyr M, Muren O, Eisler T, Boden H, Chammout G, Stark A, Skoldenberg O. Porous titanium construct cup compared to porous coated titanium cup in total hip arthroplasty: a randomised controlled trial. Int Orthop 2015; 39(5): 823-32. doi: 10.1007/s00264-014-2571-z. SHAR. The Swedish Hip Arthroplasty Register annual report 2016; 2016. Troelsen A, Malchau E, Sillesen N, Malchau H. A review of current fixation use and registry outcomes in total hip arthroplasty: the uncemented paradox. Clin Orthop Relat Res 2013; 471(7): 2052-9. doi: 10.1007/s11999013-2941-7. Valstar E R, Gill R, Ryd L, Flivik G, Borlin N, Karrholm J. Guidelines for standardization of radiostereometry (RSA) of implants. Acta Orthop 2005; 76(4): 563-72. doi: 10.1080/17453670510041574. Wilkinson J M, Peel N F, Elson R A, Stockley I, Eastell R. Measuring bone mineral density of the pelvis and proximal femur after total hip arthroplasty. J Bone Joint Surg Br 2001; 83(2): 283-8.

254

Acta Orthopaedica 2020; 91 (3): 254–259

2-year results of an RCT of 2 uncemented isoelastic monoblock acetabular components: lower wear rate with vitamin E blended highly cross-linked polyethylene compared to ultra-high molecular weight polyethylene Joost H J VAN ERP 1,2, Julie R A MASSIER 1,2, Jelle J HALMA 1, Thom E SNIJDERS 1, and Arthur DE GAST 1,2 1 Clinical Orthopedic Research Center—mN, Zeist, the Netherlands, 2 Department of Orthopedic Surgery, Diakonessenhuis, Utrecht, the Netherlands Correspondence: : jverp@diakhuis.nl Submitted 2019-11-05. Accepted 2019-12-16.

Background and purpose — The long-term survival of arthroplasty components may be limited by polyethylene wear-related problems such as periprosthetic osteolysis and aseptic loosening. Highly cross-linked polyethylene (HXLPE) blended with vitamin E was introduced to improve oxidative stability and to avoid long-term embrittlement. This study clinically compares the tribological behavior and clinical outcome of vitamin E blended HXLPE with ultrahigh molecular weight polyethylene (UHMWPE) in an isoelastic monoblock cup for uncemented total hip arthroplasty. Patients and methods — In this randomized controlled trial (RCT), 199 patients were included: 102 patients received the vitamin E blended HXLPE cup, 97 patients the UHMWPE cup. Clinical and radiographic parameters were obtained preoperatively, directly postoperative and at 3, 12, and 24 months. Wear rates were compared using the mean linear femoral head penetration (FHP) rate. Results — 188 patients (94%) completed the 2-year follow-up. Mean patient satisfaction was higher in the vitamin E blended HXLPE group (8.9 [1]) than in in the control group (8.5 [2], p = 0.03). The Harris Hip Score (HHS) was higher in the vitamin E blended HXLPE group (95 [8]) than in the control group (92 [11], p = 0.3). The FHP rate was lower in the vitamin E blended HXLPE group: 0.046 mm/year compared with 0.056 mm/year in the control group (p = 0.05). No adverse reactions associated with the clinical application of vitamin E blended HXLPE were observed during followup, with an excellent 2-year survival to revision rate of 98% for both cups. Interpretation — This study shows the superior performance of the HXLPE blended with vitamin E acetabular cup with lower linear femoral head penetration rates and better clinical results compared with the UHMWPE acetabular cup after 2 years.

Periprosthetic osteolysis and aseptic loosening, secondary to wear of the conventional UHMWPE in the acetabular cup, may limit the survivorship of total hip arthroplasty (Willert et al. 1990, Sochart 1999, Harris 2001). Highly cross-linked polyethylene (HXLPE) was introduced in the late 1990s as a response to the high wear rates of UHMWPE causing aseptic loosening and subsequent revision surgery (Kurtz et al. 2011). The use of HXLPE in THA has been shown to reduce wear significantly (Muratoglu et al. 2001b, Martell et al. 2003b, Bragdon et al. 2005, Snijders et al. 2020). However, cross-linking of polyethylene through the use of ionizing radiation forms free radicals, which leads to a decrease of long-term oxidative stability, causing embrittlement of the HXLPE (Sutula et al. 1995, Collier et al. 1996). Adding an antioxidant, like vitamin E, to the polyethylene might diminish this process (Burton and Ingold 1981, Oral et al. 2004). Several studies have shown good in-vitro and -vivo results of the vitamin E stabilized HXLPE in terms of wear rates and mechanical properties (Halma et al. 2015, Nebergall et al. 2017, Scemama et al. 2017, Zijlstra et al. 2017, Lambert et al. 2019). Besides, vitamin E blended HXLPE cups fit larger femoral head sizes due to a thinner liner, which could reduce the risk of dislocation (Zijlstra et al. 2017). However, no results of a large randomized controlled trial (RCT) with vitamin E blended HXLPE cups are available yet. This RCT compares wear rates between the uncemented vitamin E blended HXLPE acetabular cup and the UHMWPE acetabular cup after 2 years of follow-up. The primary outcome is the linear femoral head penetration (FHP) rate as a measure of wear. Secondary outcomes are the effect of increased head size, clinical performance, and complication rate between the 2 cups.

Acta Orthopaedica 2020; 91 (3): 254–259 255

Table 1. Demographics. Values are n (%) unless otherwise specified

Patients and methods Study design This single-blinded RCT was carried out at the Diakonessenhuis Hospital, a medium-size general hospital in Utrecht/ Zeist, the Netherlands. Between 2011 and 2014, 199 patients were included, and randomly allocated in 2 groups. Allocation of patients was done using an internet randomization system (ALEA, FormsVision, Abcoude, The Netherlands). After randomization, 102 patients received an uncemented vitamin E blended HXLPE cup (RM uncemented monoblock pressfit Vitamys cup, Mathys Ltd, Bettlach, Switzerland) and 97 received a conventional UHMWPE cup (RM uncemented monoblock pressfit, Mathys Ltd, Bettlach, Switzerland). Baseline characteristics and indications for THA are given in Table 1. All patients scheduled for primary THA aged between 20 and 85 years with primary osteoarthritis, osteoarthritis due to hip dysplasia, rheumatoid arthritis, avascular necrosis of the femoral head, or trauma, and willing to participate, were eligible. Patients with an ASA score ≥ 3 were excluded. After enrollment, baseline characteristics, Harris Hip Score (HHS), numeric rating scale (NRS) score for rest and load pain, and patient satisfaction were documented. Patients were scheduled for clinical and radiological followup on the first day postoperatively, and at 3, 12, and 24 months. At each follow-up HHS and NRS score, as well as complications or adverse reactions, were documented and anteroposterior radiographs (AP) of the pelvis in supine position were taken. Follow-ups were performed by 2 independent investigators who were blinded for the intervention (authors JvE and TS). In order to reduce bias, the patient was blinded and the surgeon did not perform the follow-up measurements. The endpoint of this study was set at 6 years’ follow-up or any case of revision or death. Sample size calculation An RCT by Dorr et al. (2005) found a lower mean wear rate at 2 years in the group treated with the HXLPE cup (0.150 mm [SD 0.09]) than in the group with an UHMWPE cup (0.191 mm [SD 0.09]). Sample size calculation was based on a mean difference of wear rate of 0.040, with a standard deviation of 0.09 mm. The sample size was determined on the basis of an overall a-error of 0.05 and a statistical power of 80% (ß = 0.20). This resulted in a sample size of 80 patients per group, which was increased to 100 patients per group to account for possible withdrawal and loss-to-follow-up. Subjects were able to leave the study at any time or reason, without any consequences. The investigators were able to withdraw a subject from the study for urgent medical reasons. There was no replacement of subjects after withdrawal. Procedure Critical aspects of the surgical procedure were standardized for both groups. Surgeons required knowledge of the allocated

Vitamin E Total HXLPE a UHMWPE b Factor n = 199 n = 102 n = 97 Age, mean (SD) 65 (5) 66 (5) Weight, mean (SD) 79 (13) 78 (13) Female sex 141 (71) 77 (75) Diagnosis Primary osteoarthritis 191 (96) 98 (96) Secondary osteoarthritis 2 (1) 1 (1) Inflammatory arthritis 1 (1) Femoral head avascular necrosis 1 (1) Hip dysplasia 4 (2) 3 (3)

65 (5) 79 (14) 64 (66) 93 (96) 1 (1) 1 (1) 1 (1) 1 (1)

a Vitamin E diffused highly cross-linked polyethylene. b UHMWPE: ultra-high molecular weight polyethylene.

treatment, and therefore were not blinded for the intervention. They were informed about the cup type during surgery. The procedures were performed by 7 orthopedic surgeons, each with vast experience in uncemented THA. Alumina ceramic femoral prosthetic heads (BIONIT2, Mathys Ltd, Bettlach, Switzerland) of 28, 32, or 36 mm were used. An uncemented hydroxyapatite coated stem (Twinsys, Mathys Ltd, Bettlach, Switzerland) was implanted in all cases. Key aspects of pre- and postoperative care were protocolled in order to ensure similar perioperative regimens. All patients received cefazolin prophylaxis during the 24 hours perioperatively and thromboprophylaxis with low molecular weight heparin for 6 weeks postoperative. Both groups followed the same rehabilitation regimen, starting on the first day after surgery. Radiological assessment Radiological assessment was performed by 1 of the authors (JM) according to a standardized form. The inclination of the acetabular component was measured. If present, then the radiolucency and osteolysis around the cup and stem and the affected zone, as described by DeLee and Charnley (1976), was scored. The Brooker classification was used for grading heterotopic ossifications (Brooker et al. 1973). The FHP measurements were performed on a PACS workstation with a high-resolution monitor using View Pro-X software version 4.0.8.9 (Rogan-Delft, Veenendaal, The Netherlands). All measurements were performed using a computerassisted method for measuring the FHP of all-polyethylene cups (modified dual-circle technique) previously described by Geerdink et al. (2008). AP pelvic radiographs were calibrated using the size of the prosthetic femoral head as reference. The FHP rate in mm/year was calculated by subtracting the distance measured at 2 years’ follow-up from the distance measured at 1 year and subsequently divided by a correction for the follow-up interval in days. Furthermore, the total FHP was measured by subtracting the distance measured at 24 months

256

Acta Orthopaedica 2020; 91 (3): 254–259

Table 2. Surgery specifics. Values are n (%)

Total number of randomized patients n = 199

Vitamin E Total HXLPE a UHMWPE b Factor n = 199 n = 102 n = 97 Approach Direct lateral Posterolateral Anterolateral Head size 28 mm 32 mm 36 mm Inclination < 35° 35–40° 41–45° 46–50° > 50° a, b See

98 (49) 85 (43) 16 (8)

45 (44) 46 (45) 11 (11)

53 (55) 39 (40) 5 (5)

22 (11) 112 (56) 65 (33)

2 (2) 35 (34) 65 (64)

20 (21) 77 (79) –

16 (8) 47 (24) 54 (27) 51 (26) 31 (16)

5 (5) 20 (19) 26 (26) 31 (30) 20 (20)

11 (11) 27 (28) 28 (29) 20 (21) 11 (11)

Allocated to and received vitamin E blended HXLPE cup n = 102 Excluded (n = 5): – revised, 2 – lost to follow-up, 3 Included in the 2-year follow-up n = 97

Allocated to and received UHMWPE cup n = 97 Excluded (n = 6): – revised, 2 – lost to follow-up, 4 Included in the 2-year follow-up n = 91

Figure 1. Summary of follow-up. 1 patient had died due to unrelated diseases, 4 patients were untraceable, and 2 patients left the study because of severe comorbidity. Furthermore, radiographical data were unavailable for 3 patients in the conventional HXLPE group and 5 in the vitamin E blended HXLPE group.

Table 1.

from the distance measured on the first day postoperatively, at 3 months, or at 12 months and subsequently divided by the correction for the follow-up interval in days. Statistics Statistical analysis was performed using SPSS Statistics, version 23.0 (IBM Corp., Armonk, NY, USA). The distribution of the data was checked using the Kolmogorov–Smirnov test. The Pearson chi-square test was used to test for significant differences between groups in head size, surgical approach, and radiographic specifics (Brooker classification, radiographic lucency around the cup and stem, and osteolysis around the stem). Furthermore, multivariable analyses were used to test for significant relationships of these variables with the FHP rate within 1 group. A paired t-test was used to test for significant differences in both groups between the FHP rate in the first year and in the second year. An independent t-test was used to test for significant differences in FHP rate between the two groups. Ethics, registration, data sharing, funding, and potential conflict of interests The procedures performed in this study, involving human participants, were in accordance with the ethical standards of the institutional and/or national research committee, with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standard and the CONSORT statement. All patients gave informed written consent. The protocol was approved by the local institutional review board and registered at Central Commission Human-Related research (CCMO) Registry as HipVit trial (NL 32832.100.10, R-10.17D/HIPVIT 1). Data are available from the corresponding author on reasonable request. This study was funded by the Clinical Orthopedic Research Foundation, Diakonessenhuis Zeist. The authors declare no relevant conflict of interest.

Results 188/199 patients (94%) completed the 2-year follow-up. Patient characteristics and surgery specifics are presented for both groups (Tables 1 and 2). 7 patients (4%) were lost to follow-up (Figure 1). 4 patients (2%) were excluded because they underwent revision surgery. In the vitamin E blended HXLPE group 2 patients underwent revision surgery: 1 for a deep infection and 1 for recurrent instability. In the UHMWPE group, 2 patients underwent revision surgery: 1 for cup malpositioning and 1 for recurrent instability. The head sizes used in the vitamin E blended HXLPE group were larger than the head sizes in the control group (p < 0.001). Bivariate analysis showed no statistically significant differences in heterotopic ossifications (p = 0.4), radiographic lucency (p = 0.3), or osteolysis (p = 0.8) around the stem. The vitamin E blended HXLPE cups showed less radiographic lucency around the cup (p = 0.04). The mean time to the 2-year follow-up visit was 27 (SD 5.4) months. Femoral head penetration The FHP rates were normally distributed. The total FHP after 2 years was 0.27 mm in the vitamin E blended HXLPE cup and 0.28 mm in the UHMWPE cup (Table 3). The FHP rate from 1 to 2 years, thus excluding the bedding-in time of 1 year, was 0.046 and 0.056 mm/year for the vitamin E blended HXLPE cup and UHMWPE cup respectively (Figure 2). An independent t-test showed a significant lower FHP rate of the vitamin E blended HXLPE cup after 2 years (p = 0.05) (Table 3). FHP rates for each head size in both groups are given in Table 4, see Supplementary data). There was a statistically significant difference between the FHP rate in the first and second year in both groups (vitamin E blended HXLPE cup: p < 0.001, UHMWPE cup: p < 0.001). There were no statistically significant differences in wear rates between head sizes of same cup type (Table 4, see Supplementary data).

Acta Orthopaedica 2020; 91 (3): 254–259

257

Table 3. Femoral head penetration rate (mm). Values are mean (SD) [95%CI]

Femoral head penetration (mm) 0.7

Follow-up (months) 0–3 0–12 12–24 0–24

Vitamin E HXLPE a UHMWPE b p-value c 0.16 (0.1) [0.15–0.19] 0.24 (0.1) [0.21–0.27] 0.046 (0.03) [0.040–0.052] 0.27 (0.1) [0.25–0.30]

a, b See Table 1. c p < 0.05 = statistically

0.16 (0.1) [0.13–0.18] 0.25 (0.1) [0.22–0.27] 0.056 (0.04) [0.048–0.065] 0.28 (0.1) [0.26–0.31]

FHP rate (mm/year): vitamin E blended HXLPE UHMWPE

0.6

Total FHP (mm): vitamin E blended HXLPE

0.5

0.5 0.9 0.05 0.8

UHMWPE

0.4 0.3 0.2

significant.

0.1 0 0

Table 5. Numeric Rating Scale (NRS) and Harris Hip Scores (HHS) at the 2-year follow-up. Values are mean (SD)

0.4 (1) 0.9 (2) 8.7 (1) 94 (10)

a, b See Table 1. c p < 0.05 = statistically

0.3 (1) 0.7 (2) 8.9 (1) 95 (8)

0.4 (1) 1.1 (2) 8.5 (2) 92 (11)

Months after index operation

Figure 2. Total femoral head penetration (FHP) in mm and mean FHP rate in mm/year.

Vitamin E Total HXLPE a UHMWPE b Factor n = 180 n = 94 n = 87 p-value c NRS rest pain NRS load pain NRS satisfaction HHS

Score 10

0.6 0.06 0.03 0.03

9 8 7 HSS/10 – vitamin E blended HXLPE

significant.

HSS/10 – UHMWPE NRS satisfaction – vitamin E blended HXLPE

NRS satisfaction – UHMWPE NRS load pain – vitamin E blended HXLPE NRS load pain– UHMWPE

NRS load pain – vitamin E blended HXLPE NRS rest pain– UHMWPE

3 2

Clinical outcomes NRS and Harris Hip scores at baseline and at each follow-up are presented in Figure 3. The scores for both groups at 2-year follow-up are specified in Table 5. The HHS was higher in the vitamin E blended HXLPE group (p = 0.03), as was the NRS score for patient satisfaction (p = 0.03). Specified results for both groups at baseline, 3-, and 12-months’ follow-up can be found in Tables 6, 7, and 8, see Supplementary data. The NRS scores for rest and load pain were not significantly different between groups (p = 0.6, p = 0.06). Complications No adverse reactions associated with the clinical application of vitamin E-blended HXLPE were observed. In 2 patients with a vitamin E blended HXLPE cup, a perioperative fracture of the femur occurred. In one patient with a UHMWPE cup, a postoperative fracture of the acetabulum occurred. All patients were treated nonoperatively and recovered completely without further complications. Within the 2 years’ follow-up 13 complications occurred (Table 9, see Supplementary data). All patients recovered completely, except for 2 patients who continued to have mild paresthesia, caused by neuropraxia of the sciatic nerve. The 2-year survival to revision rate is 98% for both groups. The 2-year survival rate for aseptic loosening is 100% in both groups.

1 0

Months after index operation

Figure 3. Harris Hip Scores (HHS) and Numeric Rating Scale (NRS) scores.

Discussion This RCT is the first to compare 2 similar uncemented acetabular monoblock cups in terms of wear rates; one made of vitamin E blended HXLPE and one of conventional UHMWPE. The vitamin E blended HXLPE showed a lower (p = 0.05) FHP rate of 0.046 mm/year compared with the conventional HXLPE cup (0.056 mm/year). No adverse reactions or abnormal mechanical behavior, concerning the clinical application of vitamin E, were observed after 2 years postoperatively. Previous in-vitro and -vivo studies have shown that vitamin E blended HXLPE has improved wear rates and protects against oxidation and material embrittlement (Halma et al. 2015, Nebergall et al. 2017, Scemama et al. 2017, Zijlstra et al. 2017, Lambert et al. 2019). The vitamin E blended HXLPE showed a statistically significant superior mean linear FHP rate of 0.046 mm/year compared with the conventional HXLPE cup (0.056 mm/year). Many studies show that a substantial amount of the FHP occurs within the first year of a

258

component being in vivo, due to bedding-in of the liner or creep of the polyethylene (Sychterz et al. 1999, McCalden et al. 2005). Therefore, we attributed the first year of linear FHP to the creep of the vitamin E blended HXLPE liner (Rochcongar et al. 2018, Snijders et al. 2020). The total FHP in our study was 0.27 mm in the vitamin E blended HXLPE cup and 0.28 mm in the UHMWPE cup. This is within the range of results reported in current literature (Salemyr et al. 2015, Shareghi et al. 2015). After the bedding-in time, a steady-state penetration rate is reached. Our study showed a statistically significant difference between the FHP rate in the first and second year in both groups, suggesting that steady-state penetration has not been reached yet. Therefore, at this point, it is impossible to differentiate between bedding-in and actual wear. Several studies have already shown superior performance of vitamin E addition to HXLPE for protection against oxidation, preserving mechanical properties, with reduced wear debris and bacterial adhesion compared with the conventional UHMWPE cups, suggesting that the addition of vitamin E may prevent osteolysis, implant loosening, and eventually revision surgery (Scemama et al. 2017, Yamamoto et al. 2017, Rochcongar et al. 2018, Lambert et al. 2019). A cohort study by Halma et al. (2015) measured a similar mean FHP rate of a vitamin E blended HXLPE cup of 0.055 mm/year in 107 patients after 2 years. A small RCT by Rochcongar et al. (2018), comparing the same cups in 62 patients, found a statistically significant difference (p < 0.001) between the steadystate FHP rate in a vitamin E blended HXLPE cup (0.020 mm/ year) and a UHMWPE cup (0.058 mm/year) after 3 years of follow-up, using radiostereometric analysis (RSA). The slightly lower wear rates for the vitamin E blended HXLPE cup in the study by Rochcongar et al. could be explained by the extended period of 3 years’ follow-up. Due to the extended follow-up, the influence of creep on the wear will reduce and this results in lower measured wear rates. Considering that our short-term study has a mean follow-up time of 2 years and that steady state might not have been reached yet, the wear rate is expected to decrease further in the long term. We observed no adverse reactions or abnormal mechanical behavior, concerning the clinical application of vitamin E blended HXLPE in an isoelastic monoblock cup for uncemented THA, after 2 years postoperatively. This is consistent with 2 clinical studies, which show no adverse reactions at intermediate follow-up (Halma et al. 2015, Snijders et al. 2020). Therefore, clinical application of vitamin E blended HXLPE could be considered safe. Yet, uncertainties remain about the long-term application of vitamin E. This study has several limitations. First, only a linear, 2-dimensional, penetration measurement was performed, using the View Pro-X software instead of the 3D RSA measuring method (Callary et al. 2015). Nevertheless, this software has been found to be reliable in clinical practice for the assessment of linear FHP to determine the polyethylene wear, with good inter- and intra-class reliability (Martell et

Acta Orthopaedica 2020; 91 (3): 254–259

al. 2003a, Geerdink et al. 2008). Second, patients in the vitamin E blended HXLPE group received larger head sizes, but, similar to other studies, multivariate analysis showed no statistically significant different effect of head size on the wear in both cups (Lachiewicz et al. 2009, 2016) (Table 4, see supplementary data). Some studies found larger head sizes to cause more wear (Jasty et al. 1997, Cooper and Della Valle 2014). Therefore, the larger head sizes in the vitamin E group can only cause an underestimation of the effect. However, the marked reduction in wear with the introduction of HXLPE has diminished this concern (Muratoglu et al. 2001a, Estok et al. 2007, Cooper and Della Valle 2014). Furthermore, our study shows statistically significant superior clinical results for the vitamin E blended HXLPE cup compared with the UHMWPE cup. However, both cups show excellent outcomes in terms of NRS and HHS score. Therefore, these results are not considered to be clinically significant and could be influenced by the greater head size of the vitamin E blended HXLPE cup. Due to the availability of the vitamin E blended HXLPE cup on the Dutch market, surgeons only have a choice of larger head sizes in most similar cup sizes. In its manufacturing process thinner polyethylene liners are applied because less wear is expected. This brings the advantage of the potential lower risk of dislocation with the use of larger heads in the vitamin E blended HXLPE cups compared with the UHMWPE cups (Cooper and Della Valle 2014). In conclusion, this RCT shows excellent and superior shortterm results of the vitamin E blended HXLPE compared with a UHMWPE control group in terms of wear rate. Further follow-up is needed to assess whether the protection against oxidation due to the vitamin E results in lower wear rates and subsequently less osteolysis and aseptic loosening, compared with conventional UHMWPE in the long term. Supplementary data Tables 4 and 6–9 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/ 17453674.2020.1730073

Study design: AdG, JH. Patient inclusion: JH. Follow-up of patients: TS, JvE. Measurements: JM. Analysis: JvE, JM. Drawing up of the manuscript: JvE, JM. Critical review of the manuscript: JvE, TS, AdG, JH. Acta thanks Henrik Malchau and Michael Charles Wyat for help with peer review of this study.

Bragdon C R, Jasty M, Muratoglu O K, Harris W H. Third-body wear testing of a highly cross-linked acetabular liner: the effect of large femoral head size in the presence of particulate poly(methyl-methacrylate) debris. J Arthroplasty 2005; 20(3): 379-85. Brooker A F, Bowerman J W, Robinson R A, Riley L H. Ectopic ossification following total hip replacement: incidence and a method of classification. J Bone Joint Surg Am 1973; 55(8): 1629-32.

Acta Orthopaedica 2020; 91 (3): 254–259 259

Burton G W, Ingold K U. Autoxidation of biological molecules. 1. The antioxidant activity of vitamin E and related chain-breaking phenolic antioxidants in vitro. J Am Chem Soc 1981; 103(21): 6472-7. Callary S A, Solomon L B, Holubowycz O T, Campbell D G, Munn Z, Howie D W. Wear of highly crosslinked polyethylene acetabular components. Acta Orthop 2015; 86(2): 159-68. Collier J P, Sperling D K, Currier J H, Sutula L C, Saum K A, Mayor M B. Impact of gamma sterilization on clinical performance of polyethylene in the knee. J Arthroplasty 1996; 11(4): 377-89. Cooper H J, Della Valle C J. Large diameter femoral heads. Bone Joint J 2014; 96-B(11 Suppl. A): 23-6. DeLee J G, Charnley J. Radiological demarcation of cemented sockets in total hip replacement. Clin Orthop Relat Res 1976; (121): 20-32. Dorr L D, Wan Z, Shahrdar C, Sirianni L, Boutary M, Yun A. Clinical performance of a Durasul highly cross-linked polyethylene acetabular liner for total hip arthroplasty at five years. J Bone Joint Surg 2005; 87(8): 1816. Estok D M, Burroughs B R, Muratoglu O K, Harris W H. Comparison of hip simulator wear of 2 different highly cross-linked ultra high molecular weight polyethylene acetabular components using both 32- and 38-mm femoral heads. J Arthroplasty 2007; 22(4): 581-9. Geerdink C H, Grimm B, Vencken W, Heyligers I C, Tonino A. The determination of linear and angular penetration of the femoral head into the acetabular component as an assessment of wear in total hip replacement: a comparison of four computer-assisted methods. J Bone Joint Surg Br 2008; 90(7): 839-46. Halma J J, Eshuis R, Vogely H C, van Gaalen S M, de Gast A. An uncemented iso-elastic monoblock acetabular component: preliminary results. J Arthroplasty 2015; 30(4): 615-21. Harris W H. Wear and periprosthetic osteolysis: the problem. Clin Orthop Relat Res 2001; (393): 66-70. Jasty M, Goetz D D, Bragdon C R, Lee K R, Hanson A E, Elder J R, Harris W H. Wear of polyethylene acetabular components in total hip arthroplasty: an analysis of one hundred and twenty-eight components retrieved at autopsy or revision operations. J Bone Joint Surg Am 1997; 79(3): 349-58. Kurtz S M, Gawel H A, Patel J D. History and systematic review of wear and osteolysis outcomes for first-generation highly crosslinked polyethylene. Clin Orthop Relat Res 2011; 469(8): 2262-77. Lachiewicz P F, Heckman D S, Soileau E S, Mangla J, Martell J M. Femoral head size and wear of highly cross-linked polyethylene at 5 to 8 years. Clin Orthop Relat Res 2009; 467(12): 3290-6. Lachiewicz P F, Soileau E S, Martell J M. Wear and osteolysis of highly crosslinked polyethylene at 10 to 14 years: the effect of femoral head size. Clin Orthop Relat Res 2016; 474(2): 365-71. Lambert B, Neut D, van der Veen H C, Bulstra S K. Effects of vitamin E incorporation in polyethylene on oxidative degradation, wear rates, immune response, and infections in total joint arthroplasty: a review of the current literature. Int Orthop 2019; 43(7): 1549-57. Martell J M, Berkson E, Berger R, Jacobs J. Comparison of two and threedimensional computerized polyethylene wear analysis after total hip arthroplasty. J Bone Joint Surg Am 2003a; 85(6): 1111-7. Martell J M, Verner J J, Incavo S J. Clinical performance of a highly crosslinked polyethylene at two years in total hip arthroplasty: a randomized prospective trial. J Arthroplasty 2003b; 18(7 Suppl. 1): 55-9. McCalden R W, Naudie D D, Yuan X, Bourne R B. Radiographic methods

for the assessment of polyethylene wear after total hip arthroplasty. J Bone Joint Surg Am 2005; 87(10): 2323-34. Muratoglu O K, Bragdon C R, O’Connor D, Perinchief R S, Estok D M, Jasty M, Harris W H. Larger diameter femoral heads used in conjunction with a highly cross-linked ultra-high molecular weight polyethylene: a new concept. J Arthroplasty 2001a; 16(8 Suppl. 1): 24-30. Muratoglu O K, Bragdon C R, O’Connor D O, Jasty M, Harris W H. A novel method of cross-linking ultra-high-molecular-weight polyethylene to improve wear, reduce oxidation, and retain mechanical properties. Recipient of the 1999 HAP Paul Award. J Arthroplasty 2001b; 16(2): 149-60. Nebergall A, Greene M E, Laursen M B, Nielsen P T, Malchau H, Troelsen A. Vitamin E diffused highly cross-linked polyethylene in total hip arthroplasty at five years: a randomised controlled trial using radiostereometric analysis. Bone Joint J 2017; 99-B(5): 577-84. Oral E, Wannomae K K, Hawkins N, Harris W H W H, Muratoglu O K O K. Alpha-tocopherol-doped irradiated UHMWPE for high fatigue resistance and low wear. Biomaterials 2004; 25(24): 5515-22. Rochcongar G, Buia G, Bourroux E, Dunet J, Chapus V, Hulet C. Creep and wear in vitamin E-infused highly cross-linked polyethylene cups for total hip arthroplasty. J Bone Joint Surg 2018; 100(2): 107-14. Salemyr M, Muren O, Ahl T, Bodén H, Chammout G, Stark A, Sköldenberg O. Vitamin-E diffused highly cross-linked polyethylene liner compared to standard liners in total hip arthroplasty: a randomized, controlled trial. Int Orthop 2015; 39(8): 1499-505. Scemama C, Anract P, Dumaine V, Babinet A, Courpied J P, Hamadouche M. Does vitamin E-blended polyethylene reduce wear in primary total hip arthroplasty: a blinded randomised clinical trial. Int Orthop 2017; 41(6): 1113-18. Shareghi B, Johanson P-E, Kärrholm J. Femoral head penetration of vitamin E-infused highly cross-linked polyethylene liners: a randomized radiostereometric study of seventy hips followed for two years. J Bone Joint Surg Am 2015; 97(16): 1366-71. Snijders T E, Halma J J, Massier J R A, van Gaalen S M, de Gast A. The survivorship of the uncemented Iso-Elastic monoblock acetabular component at a mean of 6-year follow-up. HSS J 2020; 16(1): 15-22. Sochart D H. Relationship of acetabular wear to osteolysis and loosening in total hip arthroplasty. Clin Orthop Relat Res 1999; (363): 135-50. Sutula L C, Collier J P, Saum K A, Currier B H, Currier J H, Sanford W M, Mayor M B, Wooding R E, Sperling D K, Williams I R. The Otto Aufranc Award: Impact of gamma sterilization on clinical performance of polyethylene in the hip. Clin Orthop Relat Res 1995; (319): 28-40. Sychterz C J, Engh C A J, Yang A, Engh C A. Analysis of temporal wear patterns of porous-coated acetabular components: distinguishing between true wear and so-called bedding-in. J Bone Joint Surg Am 1999; 81(6): 821-30. Willert H G, Bertram H, Buchhorn G H. Osteolysis in alloarthroplasty of the hip: the role of ultra-high molecular weight polyethylene wear particles. Clin Orthop Relat Res 1990; (258): 95-107. Yamamoto K, Tateiwa T, Takahashi Y. Vitamin E-stabilized highly crosslinked polyethylenes: the role and effectiveness in total hip arthroplasty. J Orthop Sci 2017; 22(3): 384-90. Zijlstra W P, De Hartog B, Van Steenbergen L N, Scheurs B W, Nelissen R G H H. Effect of femoral head size and surgical approach on risk of revision for dislocation after total hip arthroplasty. Acta Orthop 2017; 88(4): 395-401.

260

Acta Orthopaedica 2020; 91 (3): 260–265

Low-dose CT-based implant motion analysis is a precise tool for early migration measurements of hip cups: a clinical study of 24 patients Cyrus BRODÉN 1,2, Olof SANDBERG 3, Olof SKÖLDENBERG 2, Hampus STIGBRAND 4, Mari HÄNNI 5, Joshua W GILES 6, Roger EMERY 7, Stergios LAZARINIS 8, Andreas NYSTRÖM 8, and Henrik OLIVECRONA 9 1 Department of Surgery and Cancer, Imperial College London, London, UK; 2 Department of Clinical Sciences, Danderyd Hospital, Karolinska Institutet, Stockholm, Sweden; 3 Sectra, Linköping, Sweden; 4 Department of Orthopedic Surgery, Länssjukhuset, Gävle, Sweden; Center for Research and Development, Uppsala University/County Council of Gävleborg, Sweden; 5 Department of Surgical Sciences, Section of Radiology, Uppsala University Hospital, Uppsala, Sweden; 6 Department of Mechanical Engineering, University of Victoria, Victoria, BC, Canada; 7 Department of Orthopaedic Surgery, St Mary’s Hospital, London, UK; 8 Department of Orthopedics, Institute of Surgical Sciences, Uppsala University Hospital, Uppsala, Sweden; 9 Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden Correspondence: cyrus.broden@gmail.com Submitted 2019-04-22. Accepted 2020-01-17.

Background and purpose — Early implant migration is known to be a predictive factor of clinical loosening in total hip arthroplasty (THA). Radiostereometric analysis (RSA) is the gold standard used to measure early migration in patients. However, RSA requires costly, specialized imaging equipment and the image process is complex. We determined the precision of an alternative, commercially available, CT method in 3 ongoing clinical THA studies, comprising 3 different cups. Materials and methods — 24 CT double examinations of 24 hip cups were selected consecutively from 3 ongoing prospective studies: 2 primary THA (1 cemented and 1 uncemented) and 1 THA (cemented) revision study. Precision of the CT-based implant motion analysis (CTMA) system was calculated separately for each study, using both the surface anatomy of the pelvis and metal beads placed in the pelvis. Results — For the CTMA analysis using the surface anatomy of the pelvis, the precision ranged between 0.07 and 0.31 mm in translation and 0.20° and 0.39° for rotation, respectively. For the CTMA analysis using beads the precision ranged between 0.08 and 0.20 mm in translation and between 0.20° and 0.43° for rotations. The radiation dose ranged between 0.2 and 2.3 mSv. Interpretation — CTMA achieved a clinically relevant and consistent precision between the 3 different hip cups studied. The use of different hip cup types, different CT scanners, or registration method (beads or surface anatomy) had no discernible effect on precision. Therefore, CTMA without the use of bone markers could potentially be an alternative to RSA to measure early migration.

Early migration within the first 2 years after implantation of hip and knee implants is a strong indicator of future clinical loosening (Kärrholm 2012, Pijls et al. 2012, Klerken et al. 2015). The gold standard to assess early migration in orthopedic implants is radiostereometry (RSA) (Valstar et al. 2005). There are some limitations of the standard RSA technique such as the dependency on using tantalum beads, the specialized laboratories and expensive imaging equipment, the need for specially trained personnel to conduct the examinations, and the unavoidable loss of data due to tantalum beads being occluded or hidden (Kaptein et al. 2005). Over the last decades, improvements in CT hardware and software have increased CT scan quality while also reducing the effective dose. This combined with increases in computer processing power means that CT now can form the basis for an alternative method to measure implant migration while mitigating the aforementioned limitations of RSA. Early phantom studies and clinical pilots have indicated that with methods using CT scans, a precision comparable to that of RSA can be achieved (Brodén et al. 2016, Olivecrona et al. 2016, Scheerlinck et al. 2016) Lately, commercial image-registration software named “CTbased implant Motion Analysis” (CTMA, Sectra, Linköping) has been developed to assess implant migration using standard clinical CT scanners. This article explores the CTMA tool in a clinical setting, using 3 different types of hip cups (2 cemented and 1 uncemented) drawn from different centers in Sweden to evaluate the precision of this technique. We estimated the precision of the CTMA technique in 2 different settings: first, omitting bone markers relying solely on the pelvic anatomy for the image registration; and second, using bone markers for the image registration.

Acta Orthopaedica 2020; 91 (3): 260–265

261

Figure 1. Implant types used in the studies: (a) the uncemented TOP cup used in the Uppsala study, (b) a cemented Muller Exceed ABT polyethylene cup from the Danderyd study, (c) the graft-compressing titanium shell and a cemented Lubinus cross-linked UHMWP polyethylene cup used in the Gävle study.

Materials and methods Subjects and clinical setup Study subjects were included from 3 ongoing clinical studies in different regions within Sweden: Uppsala University Hospital in Uppsala, Danderyds Hospital in Stockholm, and Gävle Hospital in Gävle. 24 hip implants were included in this study. For every hip, 2 consecutive CT examinations were conducted (i.e., a double CT examination). A different CT scanner, CT protocol, and implant was used at each hospital as specified below. 1. Uppsala University Hospital Study setting CT examinations from 5 randomly selected patients were included from an ongoing prospective cohort study. The study explores bone mineral density around a hip stem and migration of hip components investigated by RSA. Patients received an uncemented hip prosthesis with the CFP (Collum Femoris Preserving) stem and the TOP (Trabeculae-Oriented Pattern cup), provided by Waldemar Link Gmbh (Hamburg, Germany) (Figure 1) Perioperatively, beads were inserted in the pelvic and femoral bones. RSA was used to analyze component migration, and clinical follow-up was performed at 3, 12, and 24 months postoperatively. At the 7-year follow up, a CT double examination was performed. More details of this study can be found in the 2-year follow-up publication (Lazarinis et al. 2013). CT examinations The 2 consecutive CT scan were obtained with a 2x64, dual source CT scanner (Somatom Definition Flash, Siemens, Forchheim, Germany). Between scans the patient was repositioned on the CT table. Details of CT protocol settings are described in Table 1.

2. Danderyds Hospital Study setting CT examinations from 9 randomly selected patients were included from an ongoing randomized study. The study compared proximal migration of 2 types of cemented cups: an Argon gas-sterilized Polyethylene (PE) group and a Vitamin E treated PE group (Muller Exceed ABT, Biomet, Warsaw, IN, USA) (Figure 1). Patients who were included had a diagnosis of primary osteoarthritis planned for total hip arthroplasty. Patients were followed with RSA at 3 months and at 1, 2, and 5 year(s) postoperatively. A double CT examination was also performed at 3 months postoperatively. Tantalum beads were inserted in the pelvis and cup. For more details, the study protocol has been published (Sköldenberg et al. 2016). CT examinations 2 consecutive CT scans were obtained with a 128-detector CT scanner (Discovery CT750HD, GE Healthcare, Chicago, IL, USA). Between scans the patient was repositioned on the CT table. The CT settings are described in Table 1. 3. Gävle Hospital Study setting CT examinations from 10 randomly patients were included from an ongoing revision THA study on early migration. Included patients had THA revision surgery due to cup loosening and acetabular osteolysis. Exclusion criteria were systemic diseases affecting the skeleton, and/or medication with known effect on bone metabolism. The mean age at surgery was 73 years (49–87). Acetabular reconstruction was performed using an impaction bone graft with a titanium compressing shell and a cemented cup (Lubinus cross-linked UHMW-PE, Waldemar Link, Hamburg, Germany) (Figure 1). Tantalum beads were spread in the pelvis. CT scans were performed postoperatively, after 6 weeks, and after 2 years.

262

Acta Orthopaedica 2020; 91 (3): 260–265

Figure 2. Processing schematic for CTMA. (Step 1) First an optimization of bone thresholding is performed manually. (Step 2) Thereafter the pelvic bone is defined here by the surface of the pelvic anatomy; a first registration is subsequently performed. (Step 3) A manual thresholding of metal is performed. (Step 4) The implant rigid body (the cup) is defined. The user indicates the region of interest for the second registration in the 2 datasets, i.e., the implant. Next the second registration occurs. Green color indicates a successful registration.

measured migration by the CTMA system reflected the random errors and therefore the precision of the CTMA method. To assess the mea Tube Slice Rotation time kVp a current Exposure thickness Increments Pitch sured motion of the cup relative Site of CT scanner (V) (mA) (mAs) (mm) (mm) (s) to the pelvis by the software the following steps were performed Uppsala University Hospital b 120 23 23 0.6 0.6 0.9 1 Danderyd’s Hospital c 120 10 10 0.625 0.312 0.98 1 (Figure 2): Gävle Hospital d 120 80–500 e 55 0.5 0.5 0.81 0.5 1. The CT volumes were imported a kVp: Kilovoltage peak is the peak voltage applied to the X-ray tube. into the CTMA system, and an b Somatom Definition Flash, Siemens, Forchheim, Germany. optimization of bone or bead c Discovery CT750HD, GE Healthcare, Chicago, IL, USA. thresholding was performed d Aquilon One CT scanner (Toshiba). e Dynamic tube current. manually. The same threshold was used for all examinations from a given cohort. 2. The reference rigid body (pelvic bone) was defined in 2 CT examinations separate CT examinations and registered to obtain a visual 2 consecutive CT scans were obtained at 6 weeks postoperaoverlap of the bone of the 2 volumes. tively with a 160 detector Aquilon One CT scanner (Toshiba). In between the double CT examinations the patient was repo- 3. A manual thresholding of metal was performed using the same setting for all patients from a given cohort. sitioned. Details of CT protocol settings are described in 4. The implant rigid body (the cup) was defined in 2 separate Table 1. CT examinations and registered to obtain a visual overlap The definition of precision of the implant of the 2 volumes. The software calculated the motion that has occurred for the The precision of a measurement is defined by the “degree to which repeated measurement under unchanged conditions implant relative to the pelvic bone between these 2 CT volshow the same results” (Sköldenberg et al. 2014). In the datas- umes described in the CT-based coordinate system. It resulted ets used in this study, a double CT examination was performed in a visual output in the form of registered 3D volumes as in which 2 CT scans were conducted on the same day for each well as numerical migration values expressed in 6 degrees of patient. The CT scans were performed under the same condi- freedom (rotations around and translation along x, y, z, in a CT tions consecutively with the patient standing up in between. DICOM coordinate system). The CTMA procedure described earlier (steps 1–4) was The assumption is that no movement between cup and pelvis occurs between these 2 examinations. Therefore, any measure- repeated twice. The first series was performed by defining the ment other than zero is attributed to the errors of the method. reference rigid body using the surface anatomy of the pelvis without the use of bone markers (Figure 2). The second series Image analysis and evaluation procedure was performed using tantalum beads in the pelvic bone to To assess the precision of the CTMA system, an estimation of define the reference rigid body. the random errors of the method was made. For both series, the implant rigid body was defined either by Since no movement had occurred between cup and pelvis in the metallic surface of the implant or the metallic threads and/ between the 2 CT examinations in 1 double examination, the or beads in the implant depending on the implant type. Table 1. CT settings from 3 clinical trials. A bone filter has been used for all scans without any metal artefact reduction algorithm.

Acta Orthopaedica 2020; 91 (3): 260–265

During the analysis, the consistency of the registered rigid bodies was verified manually with a color-coded feedback mechanism that illustrates any change in the transformation of either rigid body due to deviation of tantalum beads, bone morphology changes, or implant deformation. Green color indicated a successful registration. Radiation The effective dose of the CT method was estimated for each dataset. The dose length product (DLP) indicates the overall energy delivered along the scan length. DLP is then multiplied by the constant (k) of the pelvis that provides the conversion from mGy · cm to mSv, which is the unit of effective dose based on the tissue exposed (AAPM, 2008).

263

Table 2. Precision of CTMA of different cups using the pelvic anatomy without beads for registration Hospital

Translation, mm Rotation, ° n x y z TT x y z

Uppsala University Hospital 5 0.07 0.13 0.31 0.20 Danderyd’s Hospital 9 0.23 0.11 0.08 0.16 Gävle Hospital 10 0.12 0.31 0.15 0.22

0.37 0.22 0.39 0.31 0.28 0.29 0.28 0.20 0.23

TT: total translation.

Table 3. Precision of CTMA of different cups using bone markers for registration Hospital

Translation, mm Rotation, ° n x y z TT x y z

Uppsala University Hospital 5 0.11 0.09 0.08 0.03 Danderyd’s Hospital 9 0.08 0.13 0.13 0.13 Gävle Hospital 10 0.10 0.20 0.14 0.12

0.43 0.23 0.22 0.25 0.21 0.29 0.20 0.22 0.26

TT: total translation.

Statistics Statistical analyses were performed in SPSS Statistics, version 25 (SPSS Hong Kong 1804, Westlands Centre, Westlands Road, Quarry Bay, HK). Data from each type of implant were tested separately for normality using the Kolmogorov– Smirnoff test. We estimated the precision of the method using the standard deviation of double measurements and the critical values that encompass 95% of Student’s t-distribution with n-1 degrees of freedom, where n represents the number of patients (Sköldenberg and Odquist 2011). Ethics, funding, and potential conflict of interest The Uppsala study was approved by the ethics committee of Uppsala University (Dnr 2007/105/2). The Karolinska study was approved by the ethics committee of Karolinska Institute (No. 2011/2003-31/1). The Gävle study was approved by the ethics committee of Uppsala University (Dnr 2015/228). No funding was received for this study. HO and CB have received consultancy fees from Sectra; Olof Sandberg is a full-time employee at Sectra.

Results For the CTMA technique using the surface of the pelvic anatomy, the precision ranged between 0.07 and 0.31 mm in translation and between 0.20° and 0.39° for rotation (Table 2). For the CTMA analysis using beads in the bone, the precision ranged between 0.08 and 0.20 mm in translation and between 0.20° and 0.43° for rotation (Table 3). For the visual feedback, no movement within the prosthesis and the bone was visualized between any of the sets of 2

CT scans forming a double examination, which agrees with our assumption that no true migration had occurred during patient repositioning in the CT scanner. The effective dose of CT scans differed between the clinical trials: the mean effective dose was 0.7 mSv for Uppsala; 0.2 mSv for Danderyd’s Hospital and 2.3 mSv for Gävle Hospital. Artefacts could be detected in the CT scans, but no movement artefacts were present.

Discussion This study determined the clinical precision of the CTMA method in 3 ongoing clinical THA studies with variable protocols (e.g., implant type, CT scanner, and CT protocol). The precision of the technique using pelvic anatomy ranged between 0.07 and 0.31 mm in translation and 0.20° and 0.39° for rotation while the precision of the technique using beads implanted in bone ranged between 0.08 and 0.20 mm in translation and between 0.20° and 0.43° for rotations. Precision values of standard RSA have been estimated to range between 0.15 and 0.60 mm for translation and between 0.3° and 2° for rotations in the lower limb, which is slightly less precise than our findings in this study (Kärrholm et al. 1997). Micromotion of more than 1.2 mm has been linked to higher revision risk (Kärrholm et al. 1994). Our results indicate that CTMA is more precise than this threshold, making it capable of detecting clinically important micromotion. The precision of CTMA of the 3 different cups used in this study did not differ markedly, despite there being differences in geometry and implant material. In addition, each type of implant was scanned with a different CT scanner and CT scanning proto-

264

col. The technique also demonstrated an insensitivity to x, y, and z directions as the precision values differed only slightly between x, y, and z for translation and rotation. Scheerlink et al. (2016) previously developed a CT method to assess early migration in hip prostheses in both a clinical and an experimental setting. Their results in vitro were as accurate and precise as RSA in the lower limb, with a precision better than 0.09 mm and 0.14°. In a clinical setting, Scheerlink et al. presented values such as mean absolute error to estimate precision. Since our methodology for estimating precision was different in this CTMA study, the comparison is difficult. However, while our study reports results using low-dose CT (which is more challenging but crucial for widespread clinical application), the Scheerlink paper reports exposure of the CT scans of 3.1 to 8.2 mSv, which limits the use of their technique. This is an important difference as the European Commission, in its guidance on exposure in medical and biomedical research, “Radiation protection 99” (European Commission, 1998), gives approximate values of between 0.5 and 10 mSv as the target per person for an entire study of this type. In contrast, our method achieved substantially lower radiation doses across different CT scanners and protocols (0.2–2.3 mSv). The quality of the CT images improved with the higher dosage and the segmentation of the image was easier; however, all collected data could easily be used for the CTMA analysis and the precision did not improve with higher dosage. This effective dose could be compared to RSA examination that delivers a dose of 0.15 mSv and a normal pelvic radiograph of 0.7 mSv (Valstar 2001, Boettner et al. 2016) However, in practice, RSA examination involves additional retakes and the effective dose would often be higher than the quoted 0.15 mSv. This study only included acetabular components, thus further studies are needed to evaluate whether the reported precision level is consistent for other implants. For example, the femoral stem has a larger volume, is more asymmetric, and the material is metallic. These three factors should diminish registration errors. At the same time the femur is a more elongated and symmetrical bone compared with the pelvis. As a result of these conflicting factors, it is probably not possible to extrapolate from the current results. Another aspect to consider in future studies is the efficacy of the CTMA software in following component migration over time. The registration process of the bone could, in theory, be more complicated because of the changes in the bone morphology over time. However, the software has a feature that identifies differences between the CT volumes and, for the purposes of registration, uses only portions of the bone morphology that are common in each scan. An implant motion analysis system based on low-dose CT scans and free of bone tantalum markers opens up the possibility for simpler implant migration studies. This could be used to create more rigorous quality control of new implants before widespread market introduction, while minimizing impediments to innovation.

Acta Orthopaedica 2020; 91 (3): 260–265

The CTMA addresses some of the limitation of RSA such as the need for special personnel to perform the image acquisition, the need for an expensive RSA laboratory, and the problems of obscured markers. However, CTMA expertise is needed to perform some of the manual steps of the image analysis. Each image registration takes between 10 and 20 minutes to perform. In conclusion, our study demonstrates in 3 ongoing clinical studies that CT-based migration measurement achieves a clinically relevant precision without the need for bone markers. It demonstrates that this technique can be used on different types of cup implants using CT volumes with different CT scanners at different effective doses and still maintain the clinically relevant precision needed to follow early migration of cups in hip arthroplasty.

CB, HO, OSK, and JG were responsible for the study conception and designed the study. HS, MH, OSK, SL, and AN recruited the patients and assisted CB with the data collection. CB and OSA analyzed the data. CB, OSA, and HO drafted the manuscript with continuous feedback from HS, MH, SL, AN, and JG. RE supervised the study from the design to the writing of the manuscript. Acta thanks Stephan Maximilian Röhrl for help with peer review of this study.

AAPM. Diagnostic Imaging Council CT Committee. The measurement, reporting, and management of radiation dose in CT. 2008 Jan. Report No.: 96. Boettner F, Sculco P, Lipman J, Renner L, Faschingbauer M. A novel method to measure femoral component migration by computed tomography: a cadaver study. Arch Orthop Trauma Surg 2016; 136(6): 857-63. Brodén C, Olivecrona H, Maguire G Q, Noz M E, Zeleznik M P, Sköldenberg O. Accuracy and precision of three-dimensional low dose CT compared to standard RSA in acetabular cups: an experimental study. BioMed Res Int 2016; 2016: 5909741. European Commission. Radiation protection 99. Guidance on medical exposures in medical and biomedical research. Brussels: European Commission; 1998. Kaptein B L, Valstar E R, Stoel B C, Rozing P M, Reiber J H C. A new type of model-based Roentgen stereophotogrammetric analysis for solving the occluded marker problem. J Biomech 2005; 38(11): 2330-4. Kärrholm J. Radiostereometric analysis of early implant migration: a valuable tool to ensure proper introduction of new implants. Acta Orthop 2012; 83(6): 551-2. Kärrholm J, Borssén B, Löwenhielm G, Snorrason F. Does early micromotion of femoral stem prostheses matter? 4–7-year stereoradiographic follow-up of 84 cemented prostheses. J Bone Joint Surg Br 1994; 76(6): 912-17. Kärrholm J, Herberts P, Hultmark P, Malchau H, Nivbrant B, Thanner J. Radiostereometry of hip prostheses: review of methodology and clinical results. Clin Orthop 1997; (344): 94-110. Klerken T, Mohaddes M, Nemes S, Kärrholm J. High early migration of the revised acetabular component is a predictor of late cup loosening: 312 cup revisions followed with radiostereometric analysis for 2–20 years. Hip Int J Clin Exp Res Hip Pathol Ther 2015; 25(5): 471-6. Lazarinis S, Mattsson P, Milbrink J, Mallmin H, Hailer N P. A prospective cohort study on the short collum femoris-preserving (CFP) stem using RSA and DXA: primary stability but no prevention of proximal bone loss in 27 patients followed for 2 years. Acta Orthop 2013; 84(1): 32-9.

Acta Orthopaedica 2020; 91 (3): 260–265

Olivecrona H, Maguire G Q, Noz M E, Zeleznik M P, Kesteris U, Weidenhielm L. A CT method for following patients with both prosthetic replacement and implanted tantalum beads: preliminary analysis with a pelvic model and in seven patients. J Orthop Surg 2016; 11:27. Pijls B G, Nieuwenhuijse M J, Fiocco M, Plevier J W, Middeldorp S, Nelissen R G, Valstar E R. Early proximal migration of cups is associated with late revision in THA: a systematic review and meta-analysis of 26 RSA studies and 49 survival studies. Acta Orthop 2012; 83(6): 583-91. Scheerlinck T, Polfliet M, Deklerck R, Van Gompel G, Buls N, Vandemeulebroucke J. Development and validation of an automated and marker-free CT-based spatial analysis method (CTSA) for assessment of femoral hip implant migration: in vitro accuracy and precision comparable to that of radiostereometric analysis (RSA). Acta Orthop 2016; 87(2): 139-45. Sköldenberg O, Odquist M. Measurement of migration of a humeral head resurfacing prosthesis using radiostereometry without implant marking: an

265

experimental study. Acta Orthop 2011; 82(2): 193-7. Sköldenberg O, Eisler T, Stark A, Muren O, Martinez-Carranza N, Ryd L. Measurement of the migration of a focal knee resurfacing implant with radiostereometry. Acta Orthop 2014; 85(1): 79-83. Sköldenberg O, Rysinska A, Chammout G, Salemyr M, Muren O, Bodén H, Eisler T. Migration and head penetration of Vitamin-E diffused cemented polyethylene cup compared to standard cemented cup in total hip arthroplasty: study protocol for a randomised, double-blind, controlled trial (E1 HIP). BMJ Open. 2016; 6(7):e 010781. Valstar E. Digital roentgen stereophotogrammetry: development, validation, and clinical application [doctoral dissertation]. Leiden, Netherlands: Leiden University; 2001. Valstar E R, Gill R, Ryd L, Flivik G, Börlin N, Kärrholm J. Guidelines for standardization of radiostereometry (RSA) of implants. Acta Orthop 2005; 76(4): 563-72.

266

Acta Orthopaedica 2020; 91 (3): 266–271

Greater early migration of a short-stem total hip arthroplasty is not associated with an increased risk of osseointegration failure: 5th-year results from a prospective RSA study with 39 patients, a follow-up study Thilo FLOERKEMEIER 1*, Stefan BUDDE 1*, Gabriela v. LEWINSKI 1, Henning WINDHAGEN 1, Christof HURSCHLER 2, and Michael SCHWARZE 2 1 Department of Orthopaedic Surgery, Hannover Medical School, Germany; 2 Laboratory for Biomechanics and Biomaterials, Hannover Medical School * Shared first authorship. Correspondence: schwarze.michael@mh-hannover.de Submitted 2019-09-30. Accepted 2020-02-10.

Background and purpose — Short-stem hip arthroplasty has been a viable alternative to standard stems for the treatment of hip osteoarthritis for over 10 years. This study assessed whether a correlation existed between a greater initial increase in implant migration and inferior clinical outcomes at 5 years postoperatively. Results on these patients after 2 years have been published previously. Patients and methods — Radiostereometry and clinical scoring were undertaken after surgery and at 3, 6, 12, and 24 months, and 5 years postoperatively. The migration and the clinical outcomes data from the patients with initial migrations at 3 months above the 75th percentile (≥ 75% group) were compared with those with migrations at 3 months of less than the 75th percentile (< 75% group). Results — Between 3 months and 5 years after surgery, the mean resultant implant migrations were 0.40 mm (SD 0.32) in the ≥ 75% group and 0.39 mm (SD 0.25) in the < 75% group. The mean Harris Hip Scores and SF-36 physical scores at 5 years postoperatively were 100 (SD 0.4) and 44 (SD 12), respectively, for the ≥ 75% group and 99 (SD 2) and 50 (SD 10), respectively, for the < 75% group. The differences between the patient groups were not statistically significant. Interpretation — There was no correlation between a greater initial migration and inferior clinical outcomes at 5 years postoperatively. Despite a greater initial migration, there were no risks of early aseptic loosening and inferior midterm clinical outcomes associated with a short-stem implant with a primary metaphyseal anchorage.

Short-stem hip prostheses are commonly used to treat hip osteoarthritis, especially among younger patients (Thorey et al. 2013). The proposed advantages of using these prostheses include more physiological proximal load transfers to the surrounding bone that potentially reduce stress shielding and provide better options should revision surgery become necessary (Floerkemeier et al. 2017). Furthermore, implanting a shortstem prosthesis with a partial resection of the femoral neck is particularly compatible with minimally invasive surgery, which not only minimizes the length of the skin incision, but, more importantly, reduces muscle and tendon damage (Mjaaland et al. 2015). Findings from studies involving radiostereometric analysis (RSA) of short stems with primary metaphyseal fixations have suggested greater migration during the first postoperative months, followed by stabilization (Budde et al. 2016, Schwarze et al. 2018). Data describing the migration characteristics of this specific short-stem implant beyond 2 years of follow-up have not been reported (Schwarze et al. 2018). This previous study focused on the effect of surgical approach and did not find an influence on migration or clinical results. Furthermore, it remains questionable whether a greater initial migration of short-stem implants during the first postoperative weeks represents a risk factor for later loosening. In addition, it is unclear whether a greater initial migration is associated with inferior long-term functional outcomes. Hence, this study assessed whether a greater initial migration at the first follow-up assessment after a short-stem total hip arthroplasty (a) signifies a further increase in migration at the midterm follow-up assessment, which would indicate a risk of later aseptic loosening, and (b) is associated with inferior subsequent clinical outcomes and pain.

Acta Orthopaedica 2020; 91 (3): 266–271

Patients and methods The previous study described 2-year follow-up data from 60 patients who underwent total hip arthroplasty using a Metha short-stem prosthesis (Schwarze et al. 2018). To obtain 5-year follow-up data, the same patient cohort was followed for an additional 3 years. The initial 2-year randomized controlled study enrolled 60 patients (34 women and 26 men) with a mean age of 59 years (36–72) and a mean BMI of 26 (21–37) between February 2010 and June 2013. These patients provided written informed consent prior to the operation and before they participated in the 5-year follow-up study. The study’s original inclusion criteria were patients aged 30–75 years with progressive osteoarthritis of the hip that was confirmed by analyzing plain radiographs. The study’s exclusion criteria were neurological disorders, cardiovascular disorders affecting ambulation, sensorimotor disorders, previous bone surgery on the affected joint, allergic reactions to the implant materials, revision surgery, and an unwillingness to participate in the extension of the study. The present study analyzed the use of the Metha short-stem implant (Aesculap AG, Tuttlingen, Germany), which is representative of the partial neck-sparing group of short-stem implants (Lombardi et al. 2009). Likewise, Feyen and Shimmin (2014) classified this implant as a short-stem implant with a subcapital osteotomy. All of the patients received appropriately sized (sizes 1–7) cementless short-stem hip implants. The stem was made of a titanium forged alloy (Ti4Al6V) coated with pure titanium and a 20-µm layer of calcium phosphate in the proximal region, and had a polished tip. The patients included in this study received the nonmodular version of the implant that had a caput-collum-diaphyseal angle of 120°, 130°, or 135°. A Plasmacup SC press-fit acetabular component (Aesculap AG, Tuttlingen, Germany) was used with either a polyethylene or a ceramic insert; ceramic heads of 32-mm diameter were used exclusively. Surgery was performed by 1 of 5 experienced surgeons using either a conventional transgluteal lateral Hardinge approach or an anterolateral modified Watson-Jones approach. The anesthesia protocol, insertion of the tantalum beads that provided a reference for the migrations calculated using RSA, postoperative pain management protocol, rehabilitation program, and the discharge criteria have been described previously (Schwarze et al. 2018). The precision of the RSA setup was determined through duplicate examinations of 15 patients (Table 1, see Supplementary data). The 5-year follow-up appraisals between February 2015 and June 2018 involved examinations of the RSA images and clinical investigations that comprised assessments using the Harris Hip Score (HHS), the physical functioning scale of the Short Form-36 (SF-36), the mental health domain of the SF-36, perceived pain measured in a visual analog scale (VAS), and radiological examinations. A score of 0 on the

267

VAS pain scale corresponds to “no pain,” while 10 corresponds to “maximum pain imaginable.” Implant migration was calculated relative to the first postoperative images that were captured at 3–10 days postoperatively. The implant surface models required for model-based RSA were obtained using reverse engineering techniques. The calibration box’s coordinate system served as the reference for the migration measurements, where x was positive in the medial direction, y was positive in the cranial/proximal direction, and z was positive in the anterior direction. In addition to migration along these axes, implant migration relative to the surrounding bone was calculated as the magnitude of the resultant movement vector (which is (Tx²+Ty²+Tz²)0.5) of the implant’s geometric center, and it was, therefore, always positive. The RSA parameters and procedures (Table 1, see Supplementary data) were defined in accordance with standard guidelines (ISO 2013). As relatively large initial migrations of implants were observed at 3 months postoperatively in the initial study, which had a 24-month follow-up duration, the patient population was divided into 2 groups for further analysis. One group comprised patients with initial resultant migrations below the 75th percentile, that is, < 1.43 mm, (the < 75% group), and the other group comprised patients with initial resultant migrations that were equal to or above the 75th percentile, that is, ≥ 1.43 mm (the ≥ 75% group). These groups were compared to investigate relationships between the extent of the initial resultant migration and the midterm clinical outcomes. Statistics The clinical scores in the groups with low (< the 75th percentile) and high (≥ the 75th percentile) initial migration were compared at each follow-up using two-sided Student’s t-tests with significance levels of α = 0.05. To compare clinical scores and resultant implant migration between follow-up intervals within each group, we used paired t-tests that used a Bonferroni-corrected initial significance level of 0.05. Ethics, registration, funding, and potential conflicts of interest The local ethics committee approved this study (Amendment to Institutional Review Board No. 4565, February 2015). The initial 2-year randomized controlled study was registered in the German Clinical Trials Register (DRKS00010421). The study was funded by Aesculap AG Tuttingen, Germany. The authors TF, GvL, and HW are paid instructors in lecture courses for Aesculap AG.

Results Of the 60 patients initially recruited, 11 were excluded from the analysis at the 24-month follow-up stage for a variety of reasons (Schwarze et al. 2018). A further 10 of the 49 remain-

268

Acta Orthopaedica 2020; 91 (3): 266–271

Resultant migration, mm

Included n = 60

Excluded (n = 11): – insufficient post-operative image quality/marker position, 5 – withdrawal of consent, 1 – other reasons, 5

Temporarily lost to follow-up/analysis (n = 8): – RSA technical issues, 2 – no show, 4 – regulatory issues, 2 3-month follow-up n = 41

Temporarily lost to follow-up/analysis (n = 10): – RSA technical issues, 1 – no show, 2 – regulatory issues, 7 6-month follow-up n = 39

l l

5 Migration high low

l l l

l l

3 l

ll l l

Months postoperatively

Resultant migration, mm n=

10 31

13 26

0 3 6

Months postoperatively

Resultant migration from 3 months, mm 10 33

9 32

8 31

13 26

10 33 l

9 32

8 31

Temporarily lost to follow-up/analysis (n = 6): – RSA technical issues, 0 – no show, 2 – regulatory issues, 4

1.0 4 0.8

1-year follow-up n = 43

Temporarily lost to follow-up/analysis (n = 8): – RSA technical issues, 2 – no show, 6 – regulatory issues, 0

3 0.6 l

0.4 l l

2-year follow-up n = 41 0

Temporarily lost to follow-up/analysis (n = 10): – RSA technical issues, 4 – no show, 6 – regulatory issues, 0 5-year follow-up n = 39

ll l

Migration high low

0.2

Months postoperatively

Migration high low

Months postoperatively

Figure 2. Box-plots showing resultant implant migration (A) in all patients during followup, (B) during follow-up, categorized according to the initial migration percentile, (C) in all patients in relation to the first postoperative follow-up, and (D) in relation to the 3-month follow-up. ■ represents the ≥ 75% group, and ■ represents the < 75% group.

Figure 1. Flowchart of the 60 study participants and the number of patients at each follow-up assessment. The technical issues associated with the radiostereometric analysis included marker occlusion, caused by unusual positioning of ovarian/testicular radiation protection, and insufficient image quality.

ing patients were lost to follow-up at the 5-year follow-up stage (Figure 1). 252 RSA image pairs were analyzed successfully. Implant migration The 2 criteria that defined the quality of the RSA images, namely, the rigid-body error (median: 0.140 mm [0.017– 0.316]) and the condition number (median: 24 [13–118]), were within acceptable ranges (< 0.35 mm and < 120, respectively). At 60 months postoperatively, the mean (standard deviation [SD]) resultant migration was 1.12 mm (SD 1.21; 0.15–5.05), which did not statistically significantly differ from the migrations determined at the earlier follow-up intervals, namely, 3, 6, 12, and 24 months (Figure 2). The mean initial

resultant migrations were 2.71 mm (SD 1.56) in the ≥ 75% group and 0.47 mm (SD 0.30) in the < 75% group (Figure 2). At 60 months postoperatively, the resultant migrations were 3.08 mm (SD 1.41) in the ≥ 75% group and 0.61 mm (SD 0.33) in the < 75% group. With the 3-month follow-up used as baseline, the resultant migrations were 0.41 mm (SD 0.32) in the ≥ 75% group and 0.39 mm (SD 0.25) in the < 75% group after 60 months. The results did not statistically significantly differ between groups at any follow-up. The major contribution to the resultant migration is subsidence in the distal direction, followed by a lateral translation for the ≥ 75% group and a posterior translation for the < 75% group. The largest rotations were observed about the proximal–distal axis (Table 2, see Supplementary data). Clinical outcomes Both groups’ clinical scores increased postoperatively and plateaued at 12 months postoperatively (Figure 3). At 5

Acta Orthopaedica 2020; 91 (3): 266–271

269

increasing migration and an increased risk of aseptic loosening. Furthermore, our data did Migration not reveal an association between a greater high low initial migration and inferior midterm clini80 60 cal outcomes. We observed that the prosthesis was stable after the first 3 months and 60 that migration did not increase significantly 40 over the 5-year follow-up period. Thus, we 40 did not find that a greater initial migration led to a higher risk of early aseptic loos20 Migration ening or insufficient osseointegration. In 20 high fact, good secondary osseointegration and low n= 13 36 10 31 13 26 10 33 9 32 8 31 excellent clinical outcomes were observed, 0 0 0 3 6 12 24 60 0 3 6 12 24 60 regardless of the magnitude of the initial Months postoperatively Months postoperatively migration. VAS pain score SF–36–M score Several publications from studies that 10 used highly accurate RSA have described Migration high the migration of uncemented total hip low 8 implants (Table 3, see Supplementary data). 60 When comparing studies in relation to the migration patterns associated with different 6 stem systems, it must be remembered that 40 there are several ways to express implant 4 migration data using, for example, the maximum total point motion, orthogonal 20 2 translations and rotations, or the resultant Migration migration. Subsidence, which is defined as high low distal translation along the y-axis, was the 0 0 0 3 6 12 24 60 0 3 6 12 24 60 major contributor to the resultant migration Months postoperatively Months postoperatively observed among the patients in our study, Figure 3. Clinical scores at each follow-up interval. There were no significant differences and it was used to undertake more robust between the groups. ■ represents the ≥ 75% group and ■ represents the < 75% group. (Harris Hip score; SF-36-P: Physical functioning scale of the Short Form-36; SF-36-M: comparisons with the data from previously Mental health domain of the Short Form-36; VAS: Visual analog scale pain score). published studies (Table 3, see Supplementary data). Compared with other uncemented stem systems, the Metha short-stem system years postoperatively, the mean (SD) HHSs for the ≥ 75% seems to possess less primary stability. The key conclusion and < 75% groups were 100 (0.4) and 99 (2), respectively, that can be drawn is that the Metha short-stem implant with a and the mean (SD) VAS scores for the ≥ 75% and < 75% primary metaphyseal anchorage provides stability and enables groups were 0.8 (1.0) and 1.3 (1.6), respectively. The mean osseointegration after an initial phase of about 3 months. Even (SD) physical functioning scale of the SF-36 and the mental among patients with greater initial migrations of ≥ 1–2 mm health domain of the SF-36 were 44 (12) and 57 (2) for the immediately after the operation, secondary osseointegration ≥ 75% group and 50 (10) and 51 (10) for the < 75% group at is possible. Kärrholm et al. (1994) established a method for predict5 years postoperatively. Overall, there were no statistically significant differences between the groups in relation to the ing long-term aseptic loosening that was based on the migramean HHSs and the SF-36 and VAS scores, and the extent of tion of hip stems that exceeded 1.2 mm after 2 years. This the initial migration had no significant effects on the clinical method was established using cemented long-stem implants that have different underlying fixation principles compared scores (Figure 3). with uncemented short-stem devices, because the bone-loading patterns differ and initial settling of the implant into the compacted bone can be expected (Salemyr et al. 2015). Our Discussion study’s results suggest that a value of 1.2 mm after 2 years is The RSA findings were similar between the < 75% group and not directly applicable for short stems. These findings confirm the ≥ 75% group with respect to additional migration; there- the conclusion given by Kroell et al. (2009) that the initial fore, a greater initial migration was not associated with steadily migration patterns associated with short-stem implants and SF–36–P score

Harris Hip score

100

l l

270

the subsequent secondary stabilization may not predict longterm survival. Similar to our findings, these authors found that excellent secondary osseointegration may occur even among patients with initial migrations of > 2.0 mm within the first 3 months. In our first study, the stem survival rate was 97% after 2 years (n = 2 revisions due to infection). After 5 years, no further revisions were necessary. The cases in the ≥ 75% group did not show any clinical signs of loosening at 5 years’ followup. The stems of these patients have been proven to be stable without clinical or radiological hint of further subsidence or loosening. Good clinical outcomes were evident at 60 months regarding both the HHS and SF-36 scores. Whilst other publications found a constant development of the SF-36 physical score towards the 5-year follow-up (Nebergall et al. 2016), the ≥ 75% group revealed a drop at 5 years in that score compared with 2 years. However, due to the small sample size in the group, this should not be overrated, since the difference for the < 75% group is not statistically significant. Regarding the specific stem of our study, several authors have published data describing midterm clinical outcomes at similar followup intervals after implanting the Metha short-stem device: Thorey et al. (2013) showed that after a mean follow-up duration of 5.8 years, the mean (SD) HHS had increased from 46 (17) preoperatively to 90 (5) postoperatively, the mean Hip dysfunction and Osteoarthritis Outcome Score (HOOS) had improved from 55 (16) preoperatively to 89 (10) postoperatively, and the Kaplan–Meier survival rate was about 98%. Lacko et al. (2014) showed that within a group that received Metha short-stem implants, the mean preoperative and postoperative HHSs were 42 (10) and 94 (5), respectively, and subsidence of the stem had occurred in 1 patient. Wittenberg et al. (2013) presented clinical and radiological data from 250 patients who received Metha stem devices, and they showed that with a mean follow-up duration of 4.9 years, the average HHS was 97 points, the 5-year Kaplan–Meier survival rate was 96.7%, and 85% of the patients were very satisfied, 14% were satisfied, and 1% were dissatisfied with the treatment. Our study has several limitations. First, the postoperative RSA image set was captured after initial weight-bearing. Thus, it is possible that initial settling of the implant had occurred previously, which was not considered in the data presented. Unfortunately, organizational factors in the clinic meant that no other procedure was possible. Second, a high number of the patients dropped out of the study during follow-up. On average only two-thirds of our patient population were examined at each follow-up. Nonetheless, compared with other RSA studies, the number of patients is quite high and in accordance with requirements (ISO 2013) (Table 3, see Supplementary data). In summary, our data demonstrated that there were no statistically significant differences between the group of patients with minor initial implant migrations and the group of patients with greater initial implant migrations with respect to addi-

Acta Orthopaedica 2020; 91 (3): 266–271

tional migration over 5 years. Therefore, a greater initial migration was not associated with an increased risk of aseptic loosening and subsequent insufficient osseointegration. Furthermore, this study’s findings did not demonstrate an association between a greater initial implant migration and inferior midterm clinical outcomes. Overall, the migration of the Metha short-stem implant generated promising data in this study, without hints of aseptic loosening. Supplementary data Tables 1–3 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2020. 1732749

TF: patient inclusion, conception of the study, data analysis, reviewing manuscript. SB: patient inclusion, writing of the manuscript, data analysis. GvL, HW: conception of the study, data analysis, surgery, reviewing manuscript. CH: conception of the study, reviewing manuscript. MS: conception and writing of the manuscript, data analysis and statistics, X-ray image processing. Acta thanks Bernhard Flatøy, Rami Madanat, Johan Kärrholm, and Arne Lundberg for help with peer review of this study.

Acklin Y P, Jenni R, Bereiter H, Thalmann C, Stoffel K. Prospective clinical and radiostereometric analysis of the Fitmore short-stem total hip arthroplasty. Arch Orthop Trauma Surg 2016; 136(2): 277-84. Aro E, Alm J J, Moritz N, Mattila K, Aro H T. Good stability of a cementless, anatomically designed femoral stem in aging women: a 9-year RSA study of 32 patients. Acta Orthop 2018; 89(5): 490-5. Budde S, Seehaus F, Schwarze M, Hurschler C, Floerkemeier T, Windhagen H, Noll Y, Ettinger M, Thorey F. Analysis of migration of the Nanos® short-stem hip implant within two years after surgery. Int Orthop 2016; 40(8): 1607-14. Campbell D, Mercer G, Nilsson K G, Wells V, Field J R, Callary S A. Early migration characteristics of a hydroxyapatite-coated femoral stem: an RSA study. Int Orthop 2011; 35(4): 483-8. Edmondson M, Ebert J, Nivbrant O, Wood D. Prospective randomised clinical trial assessing subsidence and rotation, using radiostereometric analysis, of two modular cementless femoral stems (Global K2 and Apex). J Orthop 2014; 11(2): 96-102. Ferguson R J, Broomfield J A, Malak T T, Palmer A J R, Whitwell D, Kendrick B, Taylor A, Glyn-Jones S. Primary stability of a short bone-conserving femoral stem. Bone Joint J 2018; 100-B(9): 1148-56. Feyen H, Shimmin A J. Is the length of the femoral component important in primary total hip replacement? Bone Joint J 2014; 96-B(4): 442-8. Flatøy B, Röhrl S M, Bøe B, Nordsletten L. No medium-term advantage of electrochemical deposition of hydroxyapatite in cementless femoral stems. Acta Orthop 2016; 87(1): 42-7. Floerkemeier T, Schwarze M, Hurschler C, Gronewold J, Windhagen H, Von Lewinski G, Budde S. The influence of tribological pairings and other factors on migration patterns of short stems in total hip arthroplasty. Biomed Res Int 2017; 2017. Hjorth M H, Stilling M, Søballe K, Nielsen P T, Christensen P H, Kold S. Preparation of the femoral bone cavity in cementless stems: broaching versus compaction. Acta Orthop 2016; 87(6): 575-82. Hoornenborg D, Sierevelt I N, Spuijbroek J A, Cheung J, van der Vis H M, Beimers L, Haverkamp D. Does hydroxyapatite coating enhance ingrowth and improve longevity of a Zweymuller type stem? A double-blinded randomised RSA trial. HIP Int 2018; 28(2): 115-21.

Acta Orthopaedica 2020; 91 (3): 266–271

ISO. RSA-ISO-16087-2013. Implants for surgery – Roentgen stereophotogrammetric analysis for the assessment of migration of orthopaedic implants; International Organization for Standardization, 2013. Jerosch J. Short stems are different: a classification of short stem designs. Orthopädische und Unfallchirurgische Prax 2012; 1: 304-12. Klein L J, Puretic G, Mohaddes M, Kärrholm J. Similar clinical results and early subsidence between the Collum Femoris Preserving and the Corail stem: a randomized radiostereometric study of 77 hips with 2 years’ followup. Acta Orthop 2019; 90(3): 202-8. Kroell A, Beaul P, Krismer M, Behensky H, Stoeckl B, Biedermann R. Aseptic stem loosening in primary THA: migration analysis of cemented and cementless fixation. Int Orthop 2009; 33(6): 1501-5. Kärrholm J, Borssén B, Löwenhielm G, Snorrason F. Does early micromotion of femoral stem prostheses matter? 4-7-year stereoradiographic follow-up of 84 cemented prostheses. J Bone Joint Surg Br 1994; 76(6): 912–7. Lacko M, Filip V, Cellár R, Vaško G. [Our experience with the metha short hip stem]. Acta Chir Orthop Traumatol Cech 2014; 81(1): 70-6. Lazarinis S, Mattsson P, Milbrink J, Mallmin H, Hailer N P. A prospective cohort study on the short collum femoris-preserving (CFP) stem using RSA and DXA. Primary stability but no prevention of proximal bone loss in 27 patients followed for 2 years. Acta Orthop 2013; 84(1): 32-9. Lombardi A V, Berend K R, Adams J B. A short stem solution: through small portals. Orthopedics 2009; 32(9). Mahmoud A N, Kesteris U, Flivik G. Stable migration pattern of an ultrashort anatomical uncemented hip stem: a prospective study with 2 years radiostereometric analysis follow-up. HIP Int 2017; 27(3): 259-66. Matejcic A, Vidovic D, Nebergall A, Greene M, Bresina S, Tepic S, Malchau H, Hodge W A. New cementless fixation in hip arthroplasty: a radiostereometric analysis. Hip Int 2015; 25(5): 477-83. McCalden R W, Korczak A, Somerville L, Yuan X, Naudie D D. A randomised trial comparing a short and a standard-length metaphyseal engaging cementless femoral stem using radiostereometric analysis. Bone Joint J 2015; 97-B(5): 595-602. Mjaaland K E, Kivle K, Svenningsen S, Pripp A H, Nordsletten L. Comparison of markers for muscle damage, inflammation, and pain using minimally invasive direct anterior versus direct lateral approach in total hip arthroplasty: a prospective, randomized, controlled trial. J Orthop Res 2015; 33(9): 1305-10. Nebergall A K, Rolfson O, Rubash H E, Malchau H, Troelsen A, Greene M E. Stable fixation of a cementless, proximally coated, double wedged, double tapered femoral stem in total hip arthroplasty: a 5-year radiostereometric analysis. J Arthroplasty 2016; 31(6): 1267-74.

271

Nieuwenhuijse M J, Valstar E R, Nelissen R G H H. 5-year clinical and radiostereometric analysis (RSA) follow-up of 39 CUT femoral neck total hip prostheses in young osteoarthritis patients. Acta Orthop 2012; 83(4): 334-41. Nysted M, Foss O A, Klaksvik J, Benum P, Haugan K, Husby O S, Aamodt A. Small and similar amounts of micromotion in an anatomical stem and a customized cementless femoral stem in regular-shaped femurs. A 5-year follow-up randomized RSA study. Acta Orthop 2014; 85(2): 152-8. Röhrl S M, Li M G, Pedersen E, Ullmark G, Nivbrant B. Migration pattern of a short femoral neck preserving stem. Clin Orthop Relat Res 2006; 448:73-8. Rutherford M, Khan R J K, Fick D P, Haebich S, Nivbrant O, Kozak T. Randomised clinical trial assessing migration of uncemented primary total hip replacement stems, with and without autologous impaction bone grafting. Int Orthop 2019; 43(12): 2715-23. Salemyr M, Muren O, Ahl T, Bodén H, Eisler T, Stark A, Sköldenberg O. Lower periprosthetic bone loss and good fixation of an ultra-short stem compared to a conventional stem in uncemented total hip arthroplasty. Acta Orthop 2015; 86(6): 659-66. Schwarze M, Budde S, von Lewinski G, Windhagen H, Keller M C, Seehaus F, Hurschler C, Floerkemeier T. No effect of conventional vs. minimally invasive surgical approach on clinical outcome and migration of a short stem total hip prosthesis at 2-year follow-up: a randomized controlled study. Clin Biomech 2018; 51:105-12. Sesselmann S, Hong Y, Schlemmer F, Hussnaetter I, Mueller L A, Forst R, Tschunko F. Radiostereometric migration measurement of an uncemented Cerafit® femoral stem: 26 patients followed for 10 years. Biomed Eng/ Biomed Tech 2018; 63(6): 657-63. Simpson D J, Kendrick B J L, Hughes M, Glyn-Jones S, Gill H S, Rushforth G F, Murray D W. The migration patterns of two versions of the Furlong cementless femoral stem: a randomised, controlled trial using radiostereometric analysis. J Bone Joint Surg Br 2010; 92(10): 1356-62. Thorey F, Hoefer C, Abdi-Tabari N, Lerch M, Budde S, Windhagen H. Clinical results of the metha short hip stem: a perspective for younger patients? Orthop Rev (Pavia) 2013; 5(4): e34. Weber E, Sundberg M, Flivik G. Design modifications of the uncemented Furlong hip stem result in minor early subsidence but do not affect further stability: a randomized controlled RSA study with 5-year follow-up. Acta Orthop 2014; 85(6): 556-61. Wittenberg R H, Steffen R, Windhagen H, Bücking P, Wilcke A. Five-year results of a cementless short-hip-stem prosthesis. Orthop Rev (Pavia) 2013; 5(1): e4.

272

Acta Orthopaedica 2020; 91 (3): 272–278

How do EQ-5D-3L and EQ-5D-5L compare in a Swedish total hip replacement population? Ted ENEQVIST 1–3, Szilárd NEMES 1,2, Johan KÄRRHOLM 1–3, Kristina BURSTRÖM 4–6, and Ola ROLFSON 1–4 1 Swedish Hip Arthroplasty Register, Gothenburg; 2 Department of Orthopedics, Institute of Clinical Sciences, the Sahlgrenska Academy, University of Gothenburg, Gothenburg; 3 Department of Orthopedics, Sahlgrenska University Hospital Gothenburg; 4 Health Outcomes and Economic Evaluation Research Group, Stockholm Centre for Healthcare Ethics, Department of Learning, Informatics, Management and Ethics, Karolinska Institutet, Stockholm; 5 Equity and Health Policy Research Group, Department of Public Health Sciences, Karolinska Institutet, Stockholm; 6 Health Care Services, Region Stockholm, Stockholm, Sweden Correspondence: ted.eneqvist@gmail.com Submitted 2018-11-30. Accepted 2020-02-19.

Background and purpose — To better detect small changes in postoperative outcome following total hip replacement (THR), the Swedish Hip Arthroplasty Register (SHAR) has decided to change from the EQ-5D-3L (3L) to the EQ5D-5L (5L). To enable comparison of results obtained with use of the 2 versions of EQ-5D, transferal of results between the questionnaires used is necessary. We assessed the measurement properties of the EQ-5D-5L compared with the EQ-5D-3L, preoperatively and 1-year postoperatively in a Swedish THR population. Patients and methods — Patients eligible for elective THR during 2015 in Western Sweden were invited to the study. With a 2-week separation, the 3L and 5L questionnaires were administered to patients before and 1 year after surgery. Comparing the 2 versions of the EQ-5D, we investigated redistribution of responses, ceiling and floor effects, EQ VAS correlations (Spearman’s rank correlation coefficient, rs), and EQ VAS scores for different severity levels by dimension (univariable ordinary least square regression). Results — The additional severity levels of the 5L version were frequently used on both measurement occasions (preoperative mobility 5%, self-care 17%, usual activities 20%, pain 5% and anxiety 3%, postoperative mobility 6%, self-care 5%, usual activities 8%, pain 9%, and anxiety 5%). Ceiling effects of the 3L version diminished overall by 7% using the 5L version. The correlations between the 2 EQ VAS scores obtained with the 3L and 5L instruments were strong both pre- (rs = 0.71) and postoperatively (rs = 0.87). Estimated EQ VAS scores for different levels of severity were consistent for all dimensions except for the mobility dimension of the preoperative 5L version and the anxiety dimension in the postoperative 5L version.

Interpretation — Our findings support that the 5L has a higher resolution than the 3L version regarding description of health-related quality of life in patients undergoing THR in Sweden. The EQ VAS scores for different levels of severity agree well between the EQ-5D versions. This could potentially be used to develop a crosswalk value set for transforming 3L to 5L responses in this patient group.

1 of the most commonly used HRQoL instruments is the EQ-5D-3L (3L), which has 5 dimensions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression) and 3 severity levels (no problems, some problems, and confined to bed [mobility], unable [self-care, usual activities] and severe problems [pain/discomfort, anxiety/depression]) (Rabin and Charro 2001). It is a short questionnaire and recognized as valid for use in many disease groups and for many conditions (Buchholz et al. 2018), including total hip replacement (THR) patients (Rabin and Charro 2001, Devlin et al. 2013). However, 3L has been questioned due to profound ceiling effects, low sensitivity, and the lack of descriptive richness. As it has limited ability to measure small but clinically relevant changes in the outcome following interventions, its usefulness to assess interventions has been debated (Sullivan et al. 2005). These limitations have been reported for both the general population and specific patient groups (Sullivan et al. 2005, Janssen et al. 2013, Buchholz et al. 2018), including THR patients (ConnerSpady et al. 2015, Greene et al. 2015). Among THR patients, the 3L particularly exhibits difficulties in assessing outcome in the mobility dimension, since the options no problems, some problems, and confined to bed limit its use in describing

Acta Orthopaedica 2020; 91 (3): 272–278

the limitations in mobility commonly experienced by patients with hip disorders. These patients typically experience limping, limited range of hip joint motion, and impaired walking capacity—and often require different aids for mobility—but they are seldom confined to bed. Similarly, the response levels for the dimensions self-care and usual activities (no problems, some problems, and unable) limit the range of responses for individuals with moderate to severe disability (Wolfe and Hawley 1997, Conner-Spady et al. 2015). In response to the critique of the 3L, the 5L, which has 5 severity levels (no problems [mobility, self-care, usual activities]/no pain [pain]/not anxious or depressed [anxiety/ depression]; some problems [mobility, self-care, usual activities]/slight pain [pain]/slightly [anxiety/depression]; moderate problems [mobility, self-care, usual activities]/moderate pain [pain]/moderately [anxiety/depression]; severe problems [mobility, self-care, usual activities]/severe pain [pain]/ severely [anxiety/depression]; confined to bed [mobility]/ unable to wash or dress [self-care]/unable to do [usual activities]/extreme pain [pain]/extremely [anxiety/depression]) has been developed (Herdman et al. 2011). The purpose is that the increased number of response levels provides a more accurate profile of the patient’s health. 5L has been compared with the 3L in several studies and has been reported to be valid, to decrease ceiling effects, and to increase discriminatory power in multiple populations (Pickard et al. 2007, Janssen et al. 2013, Scalone et al. 2013, Hinz et al. 2014, Buchholz et al. 2018), including THR patients (Conner-Spady et al. 2015, Greene et al. 2015). We asked participants to complete the 3L and 5L with 2 weeks’ separation, both preoperatively and 1 year following THR. Comparing the 2 versions of the EQ-5D, we investigated response rates, redistribution into other severity levels, ceiling and floor effects, EQ VAS score correlations, and EQ VAS scores for different severity levels for all 5 dimensions.

Patients and methods EQ-5D-3L and EQ-5D-5L In the 3L descriptive system an individual’s health state is composed of 1 level for each dimension, so if an individual answers level 1 on each dimension, the health profile is “11111” (full health). The 3L descriptive system yields 243 possible health states (35). In the 5L, the severity levels are: no, slight, moderate, severe, and extreme problems yielding 3,125 unique health states (55). The EQ-5D index is an overall measure of HRQoL, which is calculated by applying weights given by a specific value set. The derived value sets differ between the 3L and 5L, so the indexes are not directly comparable between the 3L and 5L. The 3L is part of the standard follow-up procedures of patients both pre- and postoperatively in several arthroplasty registries (Devlin et al. 2010, Rolfson et al. 2011, Lawless et al. 2012, Greene et al. 2015, LROI

273

PREOPERATIVELY

EQ-5D-5L sent together with time of THR info

Approximately 2 weeks

Sent directly to SHAR and registered into a separate database

POSTOPERATIVELY

EQ-5D-3L usual practice Collected by SHAR affiliated secretary a, registered into SHAR

EQ-5D-3L usual practice Collected and registered into SHAR at outpatient visit

Approximately 2 weeks

EQ-5D-5L sent by SHAR affiliated secretary Sent directly to SHAR and registered into a separate database

Figure 1. Preoperative and 1-year postoperative procedures for collecting the EQ-5D-3L and -5L questionnaires. a Non-respondents were reminded after 1 month according to regular procedure.

report 2016). In 2017, the Swedish Hip Arthroplasty Register (SHAR) started to collect PROMs using the 5L instrument (Kärrholm et al. 2018). Both EQ-5D questionnaires comprise a visual analog scale (EQ VAS), where the patient assesses his/her overall health status from 0 (worst imaginable health state) to 100 (best imaginable health state). Patient selection We invited all patients eligible for THR during 2015 on the basis of primary hip OA at any of the 7 publicly funded hospitals performing THR in the Western region of Sweden. The included hospitals were: Sahlgrenska University Hospital, Södra Älvsborg Hospital, Kungälv Hospital, Norra Älvsborg Hospital, Alingsås Hospital, Skövde Hospital (Kärnsjukhuset), and Lidköping Hospital. For patients on the waiting list for THR, the standard care includes a preparatory preoperative visit to the hospital. At least 2 weeks prior to the preoperative visit, waiting list coordinators at the respective hospitals sent a letter to the patients including general information regarding their upcoming surgery. Patients were informed about the study in this letter and invited to participate. The letter included the 5L version of the EQ-5D questionnaire and a preaddressed return envelope. Upon completion of the 5L version of the EQ-5D, the questionnaire was sent directly to SHAR and registered in a separate 5L database. No reminders were sent out. Following standard practice of the routine PROMs program of the SHAR, all hospitals ask patients to complete the 3L questionnaire at the preoperative visit (Clement et al. 2011) (Figure 1). 1 year following the index procedure, the SHAR-affiliated secretary at each hospital sent the postoperative 3L questionnaire to all patients. Responses were returned to the hospitals and registered in the SHAR PROMs database by administrative staff. Non-respondents were reminded after 1 month according to regular procedure. Patients received the 5L version together with information about the study

274

Acta Orthopaedica 2020; 91 (3): 272–278

EQ-5D-3L

EQ-5D-5L

EQ-5D-3L

EQ-5D-5L

EQ-5D-3L

EQ-5D-5L

None

Moderate

Extreme

Minimal

Many SAME

Minimal

Many NEW

Many INCONSISTENT

Figure 2. Possible redistribution of responses between EQ-5D-3L and -5L. Inconsistency = choosing an answer 2 levels from the first.

and an envelope pre-addressed to SHAR 2 weeks after the 1-year postoperative 3L responses had been registered. When received at SHAR the 5L data was entered to the 5L database by administrative staff (Figure 1). Statistics Response rates We calculated response rates for the pre- and postoperative 3L and 5L versions of the questionnaires for the whole group of patients and for each hospital. We compared differences in response rates between the 3L and 5L questionnaires, both pre- and postoperatively. Redistribution of responses For patients who completed both 3L and 5L questionnaires preoperatively (n = 524) or 1 year postoperatively (n = 508), the responses of the severity level were compared by dimension. The responses were defined as the same, new, or inconsistent. The same answer was defined as the same level between the 2 questionnaires, for example level 1 in 3L and level 1 in 5L. New was defined as 1 level change between the questionnaires, for example answer at level 3 in 3L and level 2 in 5L. Inconsistent was defined as change by 2 severity levels or more between the questionnaires (Janssen et al. 2008), for example answer at level 3 in 3L and at level 1 in 5L (Figure 2). Ceiling and floor effects For the patients who completed both 3L and 5L questionnaires preoperatively or 1 year postoperatively, we calculated the proportion of responses of no problems by separate dimensions and overall to investigate ceiling effects. Similarly, floor effects were investigated by calculating the proportion of patients with extreme problems in separate dimensions and overall. Ceiling and floor effects were considered present if > 15% of the patients reported the best (ceiling) or worst (floor) severity levels (Lim et al. 2015). Strength of the association and agreement between EQ-5D-3L and EQ-5D-5L EQ VAS score To assess and illustrate the agreement of EQ VAS measurements with the 2 EQ VAS scores obtained with the EQ-5D-3L and EQ-5D-5L questionnaires for patients who completed both 3L and 5L questionnaires preoperatively and 1 year postoperatively, we used Bland–Altman plots for differences (Bland and Altman 2010).

To determine the convergence of the EQ VAS scores obtained by the 3L and 5L questionnaires, we used Spearman’s rank correlation coefficient (rs). The strength of the correlations between the 2 questionnaire EQ VAS scores was defined as absent (rs < 0.20), weak (0.20 ≤ rs < 0.35), moderate (0.35 ≤ rs < 0.50), or strong (rs ≥ 0.50) (Myers and Well 2003). EQ VAS scores For the patients who completed both 3L and 5L questionnaires preoperatively or 1 year postoperatively, we used univariable ordinary least squares (OLS) regression models to estimate EQ VAS scores for the different levels of severity of each dimension. The preoperative EQ VAS score was regressed onto the preoperative 3L and 5L dimensions in separate computations. The same calculations with use of postoperative data were repeated for estimation of the postoperative EQ VAS scores. Each dimension’s level 1 was defined as reference, and the estimated regression coefficients denote the mean difference in EQ VAS scores between patients who reported level 2 and level 1, and level 3 and level 1 and so on. The premise with the regression analysis was that coefficients should have negative signs, and the magnitude of coefficients should increase with the levels. All statistical analyses were performed in R (R Foundation for Statistical Computing, Vienna, Austria). Ethics, funding, and potential conflicts of interests Ethical review approval was obtained from the Regional Ethical Review Board in Gothenburg, Sweden, registration number 516-14. TE has received funding for the study from the Felix Neubergh foundation. OR and JK were funded by grants from the Swedish state under the agreement between the Swedish government and the county councils, the ALF agreement (ALFGBG-700781). The authors report no conflicts of interest.

Results Response rates 1,567 patients received a THR during 2015 at the included 7 hospitals and were available for the study (Table 1). Of these 1,182 (75%) responded to the preoperative 3L version, 767 (49%) responded to the preoperative 5L version and 524 (33%) answered both questionnaires. For the 1-year followup, there were 1,554 patients available as 13 patients were not available due to early reoperation or death. 1,400 (89%) responded to the 1-year 3L version and 508 (32%) responded to the 5L version. Postoperatively, the response rate also differed between the hospitals with an 83–93% response rate for the 3L questionnaire while the proportion of patients who had completed the 5L and both versions ranged from 19% to 49% (Table 2, see Supplementary data).

Acta Orthopaedica 2020; 91 (3): 272–278

275

Table 1. Demographics pre- and 1 year postoperatively THR Factor

n = 1,554

Age, mean (SD) Female sex, n (%) ASA, n (%) 1 2 3 4 BMI, mean (SD) EQ VAS preoperatively, mean (SD) EQ VAS postoperatively, mean (SD) EQ-5D index preoperatively, mean (SD) EQ-5D index postoperatively, mean (SD)

70 (11) 878 (57) 340 (22) 967 (63) 229 (15) 1 (0.1) 28 (4.8) 57 (22) 74 (21) 0.40 (0.3) 0.76 (0.3)

PREOPERATIVE Mobility Self-care Usual activities Pain Anxiety/depression

Same

POSTOPERATIVE

New Inconsistent

Mobility Self-care Usual activities Pain Anxiety/depression 0

90 100

Relative distribution (%)

Figure 3. Redistributions of answers between EQ-5D-3L and -5L for all patients.

Table 3. Percentage of patients reporting severe problems (floor effect) and no problems (ceiling effect) EQ-5D dimensions

Preoperative patients (n = 524) Severe problems No problems (floor effect) (ceiling effects) 3L 5L 3L 5L

Mobility Self-care Usual activities Pain/discomfort Anxiety/depression Overall

Postoperative patients (n = 508) Severe problems No problems (floor effect) (ceiling effects) 3L 5L 3L 5L

0 2 6 2 0.6 0.6 67 30 11 9 27 4 44 4 0.6 0.6 3 0.4 53 39 0 0 1 0.4

Redistribution of responses Preoperatively, a large proportion used new severity levels in the mobility (61%), self-care (41%), usual activities (46%), and pain/discomfort (54%) dimensions in the 5L version (Figure 3). At the 1-year follow-up, patients most frequently reported no problems in all dimensions for both versions (Table 3). Inconsistencies (a response 2 or more levels away from their first response) were reported both pre- and postoperatively. Preoperatively, inconsistencies were most frequently reported in the dimensions self-care (17%) and usual activities (20%). Inconsistencies were less frequent postoperatively compared to preoperatively (Figure 3). Ceiling and floor effects Preoperatively, both versions presented ceiling effects in the self-care and anxiety/depression dimensions, as did usual activities in the 3L version. There were almost no ceiling effects preoperatively (1% for the 3L and 0.4% for the 5L questionnaire). Postoperatively, all dimensions in both 3L and 5L presented ceiling effects but to a lesser extent in the 5L version and the overall ceiling effect differed by 7 percentage units. There were no floor effects except for the 3L pain/discomfort dimension (Table 3).

0 0.2 0.2 0.2 2 1 5 0.2 2 0.2 0 0

51 44 91 76 71 46 39 35 73 69 32 25

Strength of the association and agreement between EQ-5D-3L and EQ-5D-5L EQ VAS score The Bland–Altman plots indicated that there was good agreement between the EQ VAS score measured with the 3L and the 5L questionnaires (Figure 4). Most data points were in between the limits of agreement. This was more accentuated for the postoperative measurement. Highest disagreement was observed at the middle of the EQ VAS scale (at 50), indicating “undecided” patients. The correlations between the EQ VAS scores obtained with the 3L and 5L from patients who filled out both questionnaires were strong: rs = 0.71 preoperatively and rs = 0.87 postoperatively (Figure 4). Data that disagreed between the 2 measurements tended to have small estimates, which reinforces the indicative value of the analysis. EQ VAS scores The estimated EQ VAS scores and confidence intervals are presented graphically in Figure 5 (see Supplementary data). Due to the rarity of level 4 and 5 responses to the 5L questionnaire, these 2 levels were merged. With 1 notable exception (mobility), the EQ VAS scores for different levels of severity in the preoperative data conformed

276

Acta Orthopaedica 2020; 91 (3): 272–278

sion on the EQ VAS scores for both versions of the EQ-5D questionnaires. The evolution of increasing or decreasing problems in all EQ-5D dimensions was to a 75 75 great extent mirrored in the self-rated EQ VAS. With 1 exception, the scores for different levels of severity had intuitive order within and between the 2 different ver50 50 sions. The assumption that scores follow a consistent pattern, i.e., that scores for level 1 of the 5L would be 25 25 higher than the corresponding level for the 3L, is consistent with developing studies of 5L value sets (Devlin 0 0 et al. 2018). Scores for level 3 on the 5L were generally 0 25 50 75 100 0 25 50 75 100 Preoperative EQ VAS 3L Postoperative EQ VAS 3L similar to corresponding mid-level of the 3L version. The 5L level 4/5 scores mostly had similar values as Preoperative EQ VAS 3L – EQ VAS 5L Postoperative EQ VAS 3L – EQ VAS 5L 60 level 3 in the 3L version. We expected this pattern since the 2 levels had to be collapsed as the extreme severity 40 30 levels were rarely used in our data. Value set studies have demonstrated that level 5 in the 5L version commonly have higher weights than 0 0 level 3 in the 3L version (Leidl and Reitmeir 2017, Devlin et al. 2018). However, it would not be appro–30 –40 priate to compare the scores provided in our study with valuation studies as methodologies differ vastly –60 (Oppe et al. 2014). Similar to the developers of the 0 25 50 75 100 0 25 50 75 100 Mean of measurements Mean of measurements German experience-based value set for the 5L, we Figure 4. Strength of association and agreement between the EQ VAS scores used EQ VAS to establish scores for the health states obtained with the EQ-5D-3L and -5L questionnaires approximately 2 weeks (Leidl and Reitmeir 2017). In experience-based valuapart. The diameter of the points is proportional to the number of patients reportations, patients value their own current health state as ing that particular score. Note the different scaling of the y-axis in the Bland– Altman plots to improve visibility. opposed to health states described to them (Burstrom et al. 2014). Due to the limited number of particiwell to the expected pattern (Figures 4 and 5, see Supplemen- pants, we used univariate regression models for each dimentary data). Level 1 response on the 5L questionnaire generally sion to estimate how different levels of severity project on resulted in higher estimated EQ VAS than level 1 response on the EQ VAS. Nevertheless, it is a strength to have 3L and 5L the 3L questionnaire. Level 2 responses on the 5L question- responses from the same patients, which allows for a unique naire were positioned between level 1 and level 2 on the 3L comparison of severity measured by the EQ VAS of different version. Level 3 on the 5L and level 2 on the 3L version gen- health states on the same scale. erally resulted in similar scores. The merged level 4 and level Similar to our study, Greene et al. (2015) investigated differ5 responses on the 5L version mostly took values close to the ences in responses between 3L and 5L questionnaires before level 3 response of the 3L version. and 1 to 6 years after THR. The authors stated that the 5L version is “extremely valuable in identifying preoperative health states, but appears a little less so postoperative.” Although not as strong as our results, they reported strong correlations Discussion between the EQ VAS reported in the 2 versions of EQ-5D The 5L was introduced to improve the instrument’s sensitivity both pre- and postoperatively. Thus, the strong correlation and to reduce ceiling effects, as compared with the 3L (Herdman between the 2 versions indicates convergent validity of the 2 et al. 2011). Our study confirms this intention in patients with EQ VAS versions. This warrants the use of EQ VAS of the hip joint disorders before and 1 year after THR. A vast majority 2 versions to compare the influence of different health states of patients used a new response option of the 5L version in 1 or on self-assessed health status. Furthermore, convergent validmore dimensions. Although all dimensions presented some ceil- ity between the 3L and 5L versions has been well established ing effects 1 year after surgery, the proportion of patients with regarding the health profile (Golicki et al. 2015). We found “11111” decreased from 32% to 25% with the 5L version. These similar redistribution of responses as reported by Greene et al. observations are consistent with previous research investigating (2015) in similar patients. The pre- and postoperative ceiling measurement properties of 3L and 5L in orthopedic patients effects followed a similar pattern to the results presented here, (Conner-Spady et al. 2015, Greene et al. 2015). The novelty of although our results suggest an even more pronounced reducour work is the projection of response levels for each dimen- tion of the ceiling effect with the 5L. Preoperative EQ VAS 5L

Postoperative EQ VAS 5L

100

Acta Orthopaedica 2020; 91 (3): 272–278

Conner-Spady et al. (2015) investigated the validity and reliability of the 5L version compared with the 3L version in 176 patients with OA referred for hip or knee replacement with a similar study set-up to ours. Compared with our preoperative result, these authors found a smaller proportion of patients who used new severity levels in the 5L version for all 5 dimensions. The authors concluded that the added levels in the 5L version provided stronger evidence of validity compared with the 3L version for patients with hip and knee OA referred for total joint replacement, in particular for the 3 dimensions that are particularly relevant to this patient population: mobility, usual activities, and pain/discomfort. The studies by Green et al. (2015) and Connor-Spady et al. (2015) suggest that the 5L has the ability to better discriminate health states in patients eligible for THR. This feature may improve health outcome assessment following THR. THR is predominantly an elective procedure with the main purpose to reduce pain and gain mobility and HRQoL. If the dimensions of pain/discomfort, usual activities, and mobility are not improved by the intervention, it is likely that the patient is not satisfied with the outcome of the procedure (Anakwe et al. 2011, Clement et al. 2011). Diminished ceiling effects enable a more accurate outcome assessment so that changes in care can be better monitored. The most commonly used method to investigate differences between the 3L and 5L versions is to administer both questionnaires in the same session with demographic and/or other questionnaires in between the 2 versions (Janssen et al. 2013, Scalone et al. 2013 , Craig et al. 2014). However, Janssen et al. (2008) found that patients tended to avoid using the intermediate level in the 5L questionnaire if the 3L version was administered first at the same sitting. To possibly avoid this bias, we decided to administer the 2 questionnaires with a 2-week gap in between. Similar to Greene et al. (2015), we found higher rates of inconsistent responses in the preoperative pain/discomfort and mobility dimensions than seen in other studies (Janssen et al. 2008, Scalone et al. 2013). Due to natural fluctuations in hip symptoms, the 2-week time span likely contributed to the inconsistent responses both in separate dimensions and in EQ VAS. This is a limitation that could be caused by properties of the instrument itself and/or by a true change in health state. We are aware that HRQoL measures such as EQ VAS are subjective measures and show temporal variability. Thus, we cannot expect that EQ VAS measurement at different time points should have complete agreement. However, estimates in our study are based on the presumption that most of variability in EQ VAS score at follow-up can be explained by the preceding measurement. As the 2 measurements were only 2 weeks apart there should not be substantial disagreement/ or trends in disagreement in the Bland–Altman plots, and the strength of relationship between the EQ VAS scores at the 2 time points should be strong. There is also a possible bias in the preoperative selection of patients, since no reminder was sent out after the preoperative

277

collection of the 5L version. This was not considered possible because of too short a time period until surgery. A reminder would therefore be likely to reach the patient after the THR procedure, which most certainly would have influenced the health status and the answers from the patient. Nonetheless, the health profile might have differed between the responders and non-responders. If so, we think that this difference is small when measured with EQ VAS, with comparatively low impact on the results of this study. Another limitation is that we did not randomize the order in which the respondents completed the 2 versions. There was a larger redistribution of responses preoperatively when the patients responded to the 5L version first compared with postoperatively when the order of questionnaires was the opposite. However, we believe this was not a result of the order of responses but rather an effect of the improved health state with a large proportion having no problems. The response rate for the 5L version was generally much lower than for the 3L version. We do not interpret this as a reluctance among patients to respond to a more comprehensive questionnaire. The 3L version was collected through a well-established routine as opposed to additional collection of 5L. We have no information on whether patients actually received the 5L version. Logistical challenges at the participating hospitals for distributing the additional 5L version may well explain the poorer response rate as exemplified by large differences between the different hospitals. Both the pre- and postoperative response rates of the 3L version were in all hospitals, except 1, close to the average response rate of the 3L in the SHAR (Kärrholm et al. 2018). This is a considerably higher response rate to the questionnaire than in similar studies (Conner-Spady et al. 2015, Greene et al. 2015). Conclusion The results indicate that 5L describes HRQoL in more detail in patients undergoing THR in Sweden. The EQ VAS scores for different levels of severity agree well between the 3L and the 5L questionnaires. Our research will be directed into the development of a crosswalk value set for transforming 3L to 5L responses in this patient group to be able to follow up over time. Supplementary data Table 2 and Figure 5 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/ 17453674.2020.1746124

TE, SN, and OR conceived the study and acquired ethical approval. Statistical analyses were performed by TE and SN. TE drafted the manuscript. All authors contributed to the analysis plan, interpretation of the results, and all reviewed the final submitted manuscript. Acta thanks Ines Buchholz and Nienke Willigenburg for help with peer review of this study.

278

Anakwe R E, Jenkins P J, Moran M. Predicting dissatisfaction after total hip arthroplasty: a study of 850 patients. J Arthroplasty 2011; 26(2): 209-13. doi: 10.1016/j.arth.2010.03.013. Bland J M, Altman D G. Statistical methods for assessing agreement between two methods of clinical measurement. Int J Nurs Stud 2010; 47(8): 931-6. doi: 10.1016/j.ijnurstu.2009.10.001. Buchholz I, Janssen M F, Kohlmann T, Feng Y-S. A systematic review of studies comparing the measurement properties of the three-level and five-level versions of the EQ-5D. PharmacoEconomics 2018; 36(6):645-66. doi: 10.1007/s40273-018-0642-5. Burstrom K, Sun S, Gerdtham U-G, Henriksson M, Johannesson M, Levin L-A, Zethraeus N. Swedish experience-based value sets for EQ-5D health states (Report) (Author abstract). Qual Life Res 2014; 23(2): 431. doi: 10.1007/s11136-013-0496-4. Clement N D, Muzammil A, Macdonald D, Howie C R, Biant L C. Socioeconomic status affects the early outcome of total hip replacement. J Bone Joint Surg Br 2011; 93(4): 464. doi: 10.1302/0301-620X.93B4.25717. Conner-Spady B L, Marshall D A, Bohm E, Dunbar M J, Loucks L, Al Khudairy A, Noseworthy T W. Reliability and validity of the EQ-5D-5L compared to the EQ-5D-3L in patients with osteoarthritis referred for hip and knee replacement. Qual Life Res 2015; 24(7): 1775-84. doi: 10.1007/ s11136-014-0910-6. Craig B M, Pickard A S, Lubetkin E I. More common, but less severe: differences between EQ-5D 3-level, youth, and 5-level versions. J Clin Epidemiol 2014; 67(1): 93-9. doi: 10.1016/j.jclinepi.2013.07.011. Devlin N J, Parkin D, Browne J. Patient–reported outcome measures in the NHS: new methods for analysing and reporting EQ–5D data. Health Econ 2010; 19(8): 886-905. doi: 10.1002/hec.1608. Devlin N J, Krabbe P F M. The development of new research methods for the valuation of EQ-5D-5L. Eur J Health Econ 2013; 14(Suppl 1): S1. doi: 10.1007/s10198-013-0502-3. Devlin N J, Shah K K, Feng Y, Mulhern B, van Hout B. Valuing health-related quality of life: an EQ-5D-5L value set for England. Health Econ 2018; 27(1): 7-22. doi: 10.1002/hec.3564. Golicki D, Niewada M, Buczek J, KarliAska A, Kobayashi A, Janssen M F, Pickard A S. Validity of EQ-5D-5L in stroke. Qual Life Res 2015; 24(4): 845. doi: 10.1007/s11136-014-0834-1. Greene M E, Rader K A, Garellick G, Malchau H, Freiberg A A, Rolfson O. The EQ-5D-5L improves on the EQ-5D-3L for health-related quality-of-life assessment in patients undergoing total hip arthroplasty. Clin Orthop Relat Res 2015; 473(11): 3383. doi: 10.1007/s11999-0144091-y. Herdman M, Gudex C, Lloyd A, Janssen M, Kind P, Parkin D, Bonsel G, Badia X. Development and preliminary testing of the new five-level version of EQ-5D (EQ-5D-5L). Qual Life Res 2011; 20(10): 1727. doi: 10.1007/s11136-011-9903-x. Hinz A, Kohlmann T, Stöbel-Richter Y, Zenger M, Brähler E. The quality of life questionnaire EQ-5D-5L: psychometric properties and normative

Acta Orthopaedica 2020; 91 (3): 272–278

values for the general German population. Qual Life Res 2014; 23(2): 443-7. doi: 10.1007/s11136-013-0498-2. Janssen M F, Birnie E, Haagsma J A, Bonsel G J. Comparing the standard EQ–5D three–level system with a five–level version. Value Health 2008; 11(2): 275-84. doi: 10.1111/j.1524-4733.2007.00230.x. Janssen M F, Pickard A S, Golicki D, Gudex C, Niewada M, Scalone L, Swinburn P, Busschbach J. Measurement properties of the EQ-5D-5L compared to the EQ-5D-3L across eight patient groups: a multi-country study. Qual Life Res 2013; 22(7): 1717. doi: 10.1007/s11136-012-0322-4. Kärrholm J, Mohaddes M, Odin D, Vinblad J, Rogmark C, Rolfson O. The Swedish Hip Arthroplasty Register Annual Report 2017 [Internet]. The Swedish Hip Arthroplasty Register; 2018. Available from: http://refdocs. registercentrum.se/10.18158/BkOffx7U4 Lawless B, Greene M, Slover J, Kwon Y-M, Malchau H. Does age or bilateral disease influence the value of hip arthroplasty? Clin Orthop Relat Res 2012; 470(4): 1073-8. doi: 10.1007/s11999-011-2118-1. Leidl R, Reitmeir P. An experience-based value set for the EQ-5D-5L in Germany. Value Health 2017; 20(8): 1150-6. doi: 10.1016/j.jval.2017.04.019. Lim C R, Harris K, Dawson J, Beard D J, Fitzpatrick R, Price A J. Floor and ceiling effects in the OHS: an analysis of the NHS PROMs data set. BMJ Open 2015; 5(7): e007765. doi: 10.1136/bmjopen-2015-007765. LROI. Annual report Dutch Arthroplasty Register. [Annual report]; 2016. Myers J L, Well A D. Research design and statistical analysis, 2nd ed. Mahwah, NJ: Lawrence Erlbaum; 2003. p 508. Oppe M, Devlin N J, van Hout B, Krabbe P F, de Charro F. A program of methodological research to arrive at the new international EQ-5D-5L valuation protocol. Value Health 2014; 17(4) (445-453). Pickard S A, De Leon C M, Kohlmann C T, Cella C D, Rosenbloom C S. Psychometric comparison of the standard EQ-5D to a 5 level version in cancer patients. Med Care 2007; 45(3): 259-63. doi: 10.1097/01. mlr.0000254515.63841.81. Rabin R, Charro F D. EQ-SD: a measure of health status from the EuroQol Group. Ann Med 2001; 33(5): 337-43. Rolfson O, Karrholm J, Dahlberg L E, Garellick G. Patient-reported outcomes in the Swedish Hip Arthroplasty Register: results of a nationwide prospective observational study. J Bone Joint Surg Br 2011; 93(7): 867-75. doi: 10.1302/0301-620x.93b7.25737. Scalone L, Ciampichini R, Fagiuoli S, Gardini I, Fusco F, Gaeta L, Del Prete A, Cesana G, Mantovani L G. Comparing the performance of the standard EQ-5D 3L with the new version EQ-5D 5L in patients with chronic hepatic diseases. Qual Life Res 2013; 22(7): 1707-16. doi: 10.1007/s11136-012-0318-0. Sullivan W P, Lawrence F W, Ghushchyan F V. A national catalog of preference-based scores for chronic conditions in the United States. Med Care 2005; 43(7): 736-49. doi: 10.1097/01.mlr.0000172050.67085.4f. Wolfe F, Hawley D J. Measurement of the quality of life in rheumatic disorders using the EuroQol. Rheumatol 1997; 36(7): 786-93. doi: 10.1093/ rheumatology/36.7.786.

Acta Orthopaedica 2020; 91 (3): 279–285 279

Increased risk for dislocation after introduction of the Continuum cup system: lessons learnt from a cohort of 1,381 THRs after 1-year follow-up Oskari A PAKARINEN, Perttu S NEUVONEN, Aleksi R P REITO, and Antti P ESKELINEN

Coxa Hospital for Joint Replacement, and Faculty of Medicine and Health Technology, University of Tampere, Tampere, Finland Correspondence: oskari.pakarinen@tuni.fi Submitted 2019-12-31. Accepted 2020-02-20.

Background and purpose — The introduction of new total hip replacements (THRs) is known to be associated with an increased risk for complications. On completion of a competitive procurement process, a new uncemented cup system was introduced into general use at our institution in 2016. We launched this study after the introduction to assess (1) the incidence of early dislocations of the old (Pinnacle) and the new (Continuum) cup systems, and (2) whether the cup design would affect the risk for dislocation. Patients and methods — We assessed the incidence of dislocations after 1,381 primary THRs performed at our institution during 2016. Also, the effect of the cup system (Pinnacle, Continuum with neutral liner, Continuum with elevated rim liner) on dislocation rates was analyzed using a multivariable regression model. Results — 47 (3.4%) early dislocations were identified. The incidence of dislocations was 1.3% for the Pinnacle, 5.1% for the Continuum with neutral liner, and 1.2% for the Continuum with elevated rim liner. The Continuum with neutral liner was found to have an increased risk for dislocations compared with the Pinnacle (aOR 5, 95% CI 1.4–17). However, when an elevated rim liner was used with the Continuum, the risk for dislocation between the Continuum and the Pinnacle was similar. Interpretation — Our results emphasize the need for both careful consideration before the introduction of new implants and the systematic monitoring of early outcomes thereafter. The elevated rim liner should be preferred for use with the Continuum cup because of the poor coverage of the neutral liner that may result in dislocations.

Dislocation is one of the most common complications after total hip replacement (THR) with an incidence rate between 0.4% and 4.1% in recent studies (Blom et al. 2008, Itokawa et al. 2013, Ravi et al. 2014, Klasan et al. 2019). Dislocation is often multiple factorial. Many patient-related (e.g., age and BMI), perioperative (e.g., cup malposition), and implantrelated (e.g., femoral head size) risk factors for dislocation have been reported (Howie et al. 2012, Danoff et al. 2016, Seagrave et al. 2017). Femoral neck fracture, osteonecrosis, and posterior approach increase the risk for revision due to dislocations (Hailer et al. 2012). Jumping distance (JD) is defined as the degree of lateral translation of the femoral head center required for the hip to dislocate (Sariali et al. 2009). A smaller JD theoretically increases the risk for dislocation. Femoral head offset, cup inclination and anteversion angles, and femoral head size affect JD (Hamilton et al. 2015). Larger femoral heads decrease the risk for dislocation (Howie et al. 2012), which is partly due to the larger JD (Crowninshield et al. 2004). Nonetheless, a larger femoral head does not guarantee better implant survivorship (Tsikandylakis et al. 2018). The coverage of the acetabular liner is another factor that affects JD. With decreased liner coverage, the JD also decreases. At our institution, new hip implants were introduced in April 2016. Over the following months, the Pinnacle uncemented porous-coated cup system (DePuy, Warsaw, IN, USA) was gradually replaced by the Continuum, an ultra-porous uncemented cup (Zimmer Biomet, Warsaw, IN, USA). We assessed (1) the dislocation incidences for the different cup systems, and (2) whether the cup design would affect the risk for dislocation even after the confounding factors had been adjusted. Secondarily, we assessed the possible effect of the learning curve associated with the introduction of new hip implants.

280

Acta Orthopaedica 2020; 91 (3): 279–285

Alcohol use

Age

All primary THRs performed at Coxa Hospital in 2016 n = 1,438

Dementia Excluded Missing postoperative data or follow-up shorter than 1 year because of early death or revision without previous dislocation n = 57

BMI Co-operation

Sex

Included n = 1,381

Pinnacle n = 299

Neutral liner n = 205

Other liner n = 95

Tissue loosening

Cup position Hip hypermobility

Other cup systems a n = 155

Neutral liner n = 566

Femoral neck fracture

Continuum n = 927

Elevated rim liner n = 171

Range of motion

Head size

Other liner n = 190

Figure 1. Flow chart of the study population.

Patients and methods Our institution is a high-volume academic tertiary referral center. Currently, more than 1,800 primary THRs are performed annually. In this study, the analyzed data are based on the 1,438 primary THRs performed at our institution between January 1, 2016 and December 31, 2016. Patients who underwent primary THR in both hips during the study period were analyzed as 2 separate operations. The baseline data were obtained from our institution’s electronic health records (EHR) and joint replacement database. 1,347 operations were performed by 30 different surgeons using the posterior approach, while 91 operations were performed using the anterior approach, all performed by the same surgeon. The EHRs of the patients were investigated between December 2017 and August 2018, giving us a minimum of 1-year followup (Figure 1). From our joint replacement database, we collected the following patient demographics and operative details: age, sex, BMI, primary diagnosis, hip range of motion, Charnley classification, surgical approach, femoral head size and model, acetabular cup model, and liner model. From the EHR, we recorded information on all dislocations and revisions that had occurred after primary THR. Those hips that were revised during the first 12 months for reasons other than dislocations were excluded, unless they had dislocated before the revision. We also checked whether these patients had had recurrent dislocations after the first dislocation, and whether the dislocation/dislocations led to revision surgery during the followup. In addition, we recorded whether the first dislocation was defined as posterior or anterior in the EHR. The learning curve effect was assessed by comparing the radiographic results of the acetabular cup positioning of the Pinnacle and Continuum cup systems. Cup malposition is an important risk factor for dislocation (Biedermann et al. 2005). An independent observer (JS) measured cup anteversion and

Jumping distance

Dislocation

Cup/liner design

Figure 2. The directed acyclic graph. Factors indicated in red were included in multivariable logistic regression analyses.

inclination angles from postoperative plain radiographs (AP pelvis) using a previously published technique (Reito et al. 2012). In this study, we defined the “safe zone” for cup position as 10°–25° anteversion and 30°–50° inclination angles, which are considered to be approximately the safest position for the acetabular component (Elkins et al. 2015, Danoff et al. 2016). Statistics 95% confidence intervals (CI) for proportions were calculated using the Wilson score interval. Univariable logistic regression was used to estimate the association of baseline risk factors with dislocation. Log-transformed regression coefficients corresponding to unadjusted odds ratio (OR) were reported with CI based on the Wald statistic. Multivariable logistic regression was fitted to estimate the adjusted ORs for dislocation. All potential factors associated with dislocation were included. We used a directed acyclic graph (DAG) in the variable selection to minimize the bias (Figure 2). Based on the DAG, in the final model, ORs for the THR designs were adjusted for age, BMI, diagnosis, sex, cup position, ROM, and head size. 2 different multivariable models were fitted. In the first model, Pinnacle cups were used as a reference with which the Continuum with neutral liner and the Continuum with elevated rim liner were compared. In the second model, we compared the Pinnacle cups with the Continuum cups with neutral liner only. All analyses were performed using IBM SPSS software, version 25 (IBM Corp, Armonk, NY, USA). Ethics, funding, and potential conflicts of interest In accordance with Finnish regulations, informed patient consent was not required as the patients were not contacted. This work was supported by the competitive research funds of Pirkanmaa Hospital District, Tampere, Finland, representing governmental funding. The authors have the following poten-

Acta Orthopaedica 2020; 91 (3): 279–285 281

Table 1. Patient demographics and perioperative details. Values are count (%) unless otherwise specified Factor Pinnacle Continuum Age, mean (SD) 64 (11) 66 (11) BMI, mean (SD) 29 (4.7) 29 (5.4) Follow-up, mean years (SD) 2.0 (0.3) 1.5 (0.2) Female sex 169 (57) 537 (58) Diagnosis Primary osteoarthritis 267 (89) 793 (86) Femoral neck fracture 6 (2.0) 33 (3.6) Osteonecrosis 8 (2.7) 25 (2.7) Developmental dysplasia of the hip 9 (3.0) 19 (2.0) Rheumatoid arthritis 1 (0.3) 17 (1.8) Other 8 (2.7) 40 (4.3) Charnley classification Unilateral hip OA 132 (62) 445 (64) Bilateral hip OA 76 (36) 205 (30) Disability because of other diseases 4 (1.9) 42 (6.1) Head size 36 mm 221 (77) 750 (81) 32 mm 66 (23) 158 (17) 28 mm 0 (0.0) 4 (0.4) Cup position, median (range) Anteversion 26 (2.6–49) 22 (3.8–53) Inclination 45 (29–62) 46 (20–71) Inside safe zone Anteversion 116 (39) 651 (71) Inclination 246 (82) 645 (70) Both 101 (34) 491 (53)

tial conflicts of interest to declare. OP: None. PN: Zimmer Biomet, educational courses. AR: Orion, paid lectures. AE: Zimmer Biomet, paid lectures; Depuy Synthes and Zimmer Biomet, institutional research support (not related to current study).

Results The median follow-up was 1.6 years (1.1–2.5). The mean age of patients was 66 years (SD 11) and the mean BMI was 28 (SD 5). 58% of patients were female. Primary osteoarthritis (83%) was the most common reason for operation, followed by femoral neck fracture (5%), osteonecrosis (3%), and developmental dysplasia of the hip (2%). More specific patient characteristics and operative details comparing patients in the Pinnacle and Continuum groups are presented in Table 1. After 1,381 operations, 47 dislocations occurred (incidence of 3.4%, CI 2.6–4.5). Of the 47 dislocations, 31 were posterior and 11 anterior. The direction of dislocation was unknown in 5 cases. In addition, 20 hips dislocated only once, 20 hips twice, and 11 hips 3 or more times. Over half of the dislocations (24/47) led to revision surgery during the follow-up period. Moreover, 18 of the 24 hips revised for dislocation had had at least 2 dislocations prior to revision. In 4 cases, 2 or more revision surgeries were required.

Table 2. Risk factors for dislocation in univariable logistic regression analysis Variable n Odds ratio (95% CI) p-value Cup/liner design Pinnacle 299 Reference Continuum with neutral liner 566 4.0 (1.4–11) 0.01 with elevated rim liner 171 0.9 (0.2–4.8) 0.9 Age (+ 10 years) 1.4 (1.0–1.8) 0.03 Sex Male 582 Reference Female 799 1.1 (0.6–1.9) 0.8 BMI 1.0 (1.0–1.1) 0.8 Diagnosis Primary osteoarthritis 1,149 Reference Femoral neck fracture 67 3.9 (1.7–9.3) 0.002 Other 165 1.5 (0.7–3.4) 0.3 Hip mobility Internal rotation (+ 5°) 1.3 (1.0–1.5) 0.02 Total ROM (+ 30°) 0.9 (0.8–1.1) 0.3 Femoral head size: 36 mm 1,061 Reference 32 mm 248 1.0 (0.4–2.2) 1.0 Approach Posterior 1,295 Reference Anterior 86 1.5 (0.5–4.1) 0.5 Cup position Anteversion 10°–25° 849 Reference Anteversion < 10° or > 25° 516 0.8 (0.4–1.4) 0.4 Inclination 30°–50° 1,011 Reference Inclination < 30° or > 50° 364 1.2 (0.6–2.2) 0.6

The incidence of dislocation was 4/299 with the Pinnacle cup system (1.3%, CI 0.5–3.4) and 37/927 with the Continuum cup system (4.0%, CI 2.9–5.5). When the Continuum cup was used with neutral liner, the incidence of dislocation was 29/566 (5.1%, CI 3.6–7.3), while the Continuum combined with elevated rim liner resulted in an incidence of 2/171 (1.2%, CI 0.3–4.2). The Pinnacle used with neutral liner had a dislocation incidence of 3/204 (1.5%, CI 0.5–4.2). In univariable logistic regression analysis (Table 2), the Continuum combined with neutral liner increased the risk for dislocation compared with the Pinnacle. However, the dislocation risk was similar when the Continuum combined with elevated rim liner was compared with the Pinnacle. The risk for dislocation using the Continuum with neutral liner compared with the Pinnacle cup remained elevated in the first multivariable analysis. When the Continuum with elevated rim liner was compared with the Pinnacle, however, no difference was observed. In the second multivariable analysis, the neutral liner of the Continuum was compared with the neutral liner of the Pinnacle with the same confounders adjusted. In that analysis, the Continuum’s neutral liner increased the risk for dislocation compared with the Pinnacle’s neutral liner (Table 3). 53% of the Continuum cups and 34% of the Pinnacle cups were positioned inside the safe zone by both anteversion and inclination angles (Table 1). When the Continuum was used

282

Acta Orthopaedica 2020; 91 (3): 279–285

Table 3. Risk for dislocation using the Pinnacle cup system, the Continuum cup system with neutral liner, or the Continuum cup system with elevated rim liner Cup/liner design n aOR (95% CI) Pinnacle 299 Reference Continuum with neutral liner 566 4.8 (1.4–17) Continuum with elevated rim liner 171 1.2 (0.2–7.8)

p-value 0.01 0.8

Pinnacle with neutral liner 204 Reference Continuum with neutral liner 566 5.3 (1.2–24) 0.03 Adjusted for age, sex, BMI, primary diagnosis, total range of motion, femoral head size, cup anteversion angle, and cup inclination angle.

with neutral liner, in 17/29 dislocations, the cup was positioned in the safe zone, while neither of the 2 dislocations with the Pinnacle, and none of the 4 dislocations with the Continuum combined with elevated rim liner, occurred inside the safe zone (Figure 3).

Discussion In this study, the overall incidence of dislocations at our highvolume academic hospital was 3.4%. Moreover, more than half of the dislocations led to early revision. The incidence of dislocation was unexpectedly high in patients operated on with the Continuum cup and neutral liner (5.1%). Using either the Pinnacle cup (1.3%) or the Continuum cup with an elevated rim liner (1.2%), the dislocation rate was in accordance with our hospital’s quality data over previous years (data not shown). The Continuum cup combined with neutral liner significantly increased the risk for dislocation compared with the Pinnacle cup system, while no statistically significant difference could be detected when the Continuum cup with elevated rim liner was compared with the Pinnacle cup. The Continuum’s neutral liner was also found to increase the risk for dislocation compared with the Pinnacle’s neutral liner. This study was designed to assess the possible increase in the incidence of dislocation after the change of THR implants at our institution. Before the competitive procurement process, we had mainly used the Pinnacle uncemented porous-coated cup system with neutral polyethylene liner in primary THRs. Starting in April 2016 and over the following few months, the Continuum ultra-porous uncemented cup gradually replaced the Pinnacle as our main cup system. The neutral liner was a natural choice to use with the Continuum because the Pinnacle had been successfully used with neutral liner for many years at our institution. As both cup systems were used as our main system, it is unlikely that the Continuum’s higher dislocation rate was because it had been used in more difficult operations. After the introduction of the Continuum cup as the main cup system, some issues with hip stability were noted in clinical work. Later, it was discovered that the Continuum’s neutral

Figure 3. Cup position for the Pinnacle cup (top panel), the Continuum cup with neutral liner (middle panel) and with elevated rim liner (bottom panel). Dislocated hips are marked with red dots.

liner had substantially worse coverage compared with the Pinnacle’s neutral liner (Figure 4). The weak coverage of the Continuum neutral liner reduces the JD, and thus explains the high risk for dislocation. The Continuum’s elevated rim liner increases the JD compared with the neutral liner, and therefore decreases the risk for dislocation. These findings led to the launch of the current study. Previously, trabecular metal cups have been reported to have a higher risk for revision than other uncemented cups (Laaksonen et al. 2018). Since 2011, however, the Continuum cup has also been found to have a higher than anticipated risk for revision in Australian registry data (AOANJRR Annual Report 2011). In the latest AOANJRR annual report, the biggest difference between the Continuum and other total conventional hip prostheses revisions was in the rate of revisions for dislocation; 1.4% of all Continuum cups had had a revi-

Acta Orthopaedica 2020; 91 (3): 279–285 283

Figure 4. Difference in coverage of the Pinnacle’s and Continuum’s neutral liners. The same ceramic head was first placed in the Pinnacle’s neutral liner (left) and the line at the rim was marked with a pen. Then, the head was moved into the same-sized Continuum’s neutral liner (right).

sion for dislocation, while in other total conventional hips the rate in time-matched comparison was 0.8%. Dislocation was the most common reason for revision with the Continuum, as 38% of all the Continuum revisions were performed due to dislocation compared with 23% for other total conventional hip prostheses (AOANJRR Annual Report 2019). In a recent Finnish register-based study, Continuum cups had a greater risk for revision, and risk for revision because of dislocation compared with other uncemented cups. Moreover, the use of neutral liner with the Continuum cup increased the risk for revision due to dislocation compared with the use of elevated rim liner with the Continuum cup (Hemmila et al. 2019). Our results are clearly in accordance with these recent findings. As previous studies and register reports indicate, concerns had previously been raised about the reliability of the Continuum cup system, especially regarding the general risk for revision and also the risk for revision due to dislocation. However, other recent studies have concentrated on the risk of revision due to dislocation in register-based settings. To the best of our knowledge, this is the first clinical cohort study that shows an increased risk for dislocations with the Continuum cup system. Moreover, our study is the first to provide causality for the previously observed higher risk for revisions due to dislocations with the Continuum cup, i.e., the poor coverage of its neutral liner that leads to a smaller JD (Figure 4). In our study, however, the risk for dislocation with the Continuum was successfully minimized by using an elevated rim liner that increases the JD. Therefore, we currently prefer to use the elevated liner with the Continuum cup still used at our hospital. We have also learned that when using the Continuum with an elevated rim liner, the anteversion angle should not be exaggerated due to the risk of posterior impingement between the stem neck and the rim of the liner, which can lead to anterior dislocation. In patients who are at high risk for dislocation after THR, we currently favor dual-mobility cups (Harwin et al. 2017). Although the Continuum cup system was used at our institution on a small scale before the competitive procurement process of 2016, the high dislocation rate could be partially explained by the learning curve. The introduction of new implants in joint replacement surgery has been reported to be

associated with a learning curve (Peltola et al. 2013), but there are also less pronounced results (Magill et al. 2016, Mohaddes et al. 2016). On average, the Pinnacle cup was generally positioned at a slightly higher anteversion angle than the Continuum. However, this did not prevent dislocation, as in all dislocations with the Pinnacle the anteversion angle was over 25°. Overall, the Continuum was better positioned than the Pinnacle. As the Continuum has a rougher (macro-textured) surface than the Pinnacle, we have learned that it is better to ream the acetabulum either line-to-line (in most cases), or even perform slight over-reaming (in hard, sclerotic bone) when using the Continuum. With the Pinnacle, on the other hand, we usually under-ream by 1 mm and use line-to-line reaming only in patients with hard, sclerotic acetabular bone. Clearly, the optimal cup position did not provide any protection from dislocation when the Continuum cup was used with neutral liner because most of the dislocations occurred in patients whose cups were positioned inside the safe zone. In comparison, the only 2 dislocations with the Continuum combined with elevated rim liner occurred with very unorthodox cup positions. These results indicate that the high dislocation rates of the Continuum combined with neutral liner cannot be explained by poor cup positioning. However, there may well be other factors that are affected by the learning curve that we are unaware of, so its effect cannot be completely excluded. In our directed acyclic graph, we did not find any factors that could have affected the choice of cup or liner design. That is because the cup system was changed from the Pinnacle to the Continuum after the competitive procurement process, and at first we mainly used the neutral liner with the Continuum as we had done with the Pinnacle for many years. However, as we observed several dislocations in patients with well-positioned Continuum cups (and neutral liner), we then gradually started to shift from neutral liner to elevated rim liner. At this point, there were still not enough data available to see whether this phenomenon was real or just a chance finding. Thus, the patient characteristics should not have affected the choice of cup system or liner model. Nevertheless, in multivariable regression analysis, we adjusted all the factors that were thought to affect the risk for dislocation and measured them in our data. With or without the adjustment, the difference in dislocation risk (in favor of the Pinnacle) remained obvious. There is always a possibility that shortly after the introduction of the Continuum the elevated rim liner may have been preferred in a small number of patients with slight instability perioperatively. However, if this has caused bias in our material, it is in favor of the neutral liner, not the elevated rim liner. We acknowledge a few weaknesses in this study. The retrospective study design may enable information bias in the data that we cannot control. Also, adjusting the right confounding factors is not easy and, even when using the DAG, it is just a subjective view. Moreover, we did not have data on the patients’ alcohol use or neurodegenerative disorders, which are known to affect the risk for dislocation (Espehaug

284

et al. 1997, Gausden et al. 2018). Furthermore, because the study cohorts were not randomized, there is always a risk for selection bias. The follow-up period was also rather short, but the majority of dislocations occur within 1 year after surgery (Blom et al. 2008, Meek et al. 2008). In addition, most revisions for dislocation are also performed within 1 year of the primary operation (Hailer et al. 2012). Still, a longer followup period would have provided more specific information about the long-term incidence and consequences of dislocations. The strengths of this study are the comprehensive and consistent data from 1 high-volume center that enabled the specific analysis of perioperative factors, such as liners, as there were no differences between hospitals as confounding factors. Indeed, the investigation of the patients’ case records enabled a more exact identification of all the possible dislocations. A clinical cohort study has an obvious advantage over register studies when there is a need to find and to prove causality. The use of DAG also limits the selection bias resulting from a collider (Shrier and Platt 2008). In conclusion, a marked increase in the incidence of THR dislocations occurred at our high-volume academic joint replacement center after the introduction of a new cup system. This finding can be attributed to the decreased jumping distance and surprisingly low coverage of the neutral polyethylene liners used in the Continuum cup system. Therefore, when a new THR design is brought into general use, the risks for early complications should be considered, and the outcomes carefully monitored, even in high-volume centers with experienced surgeons. There may be design-specific pitfalls in contemporary hip replacement designs that are difficult to tackle if surgeons are unaware of them in advance. We recommend using an elevated rim liner with the Continuum cup system.

OP gathered the data, performed the analyses, and wrote the core of the manuscript. PN co-supervised the project and revised the manuscript. AR coordinated the statistical analyses and revised the manuscript. AE planned the study, revised the manuscript, and supervised the project. The authors would like to thank Ari Lehtinen for permission to use his photographs in Figure 6, Jaakko Seppänen for measuring the anteversion and inclination angles from the plain radiographs, and Peter Heath for the language editing of the manuscript. Acta thanks Stephen Ellis Graves and Maziar Mohaddes for help with peer review of this study.

Australian Orthopaedic Association, National Joint Replacement Registry. Annual report 2011. p 167. https://aoanjrr.sahmri.com/documents/ 10180/44800/Annual+Report+2011 Australian Orthopaedic Association, National Joint Replacement Registry. Annual report 2019. Protheses investigations. Hips. Total hips. Total conventional hips. Re-identified and still used. Continuum. p 5. https://aoanjrr. sahmri.com/documents/10180/669533/Continuum+Acetabular.pdf,

Acta Orthopaedica 2020; 91 (3): 279–285

Biedermann R, Tonin A, Krismer M, Rachbauer F, Eibl G, Stockl B. Reducing the risk of dislocation after total hip arthroplasty: the effect of orientation of the acetabular component. J Bone Joint Surg Br 2005; 87(6): 762-9. Blom A W, Rogers M, Taylor A H, Pattison G, Whitehouse S, Bannister G C. Dislocation following total hip replacement: the Avon Orthopaedic Centre experience. Ann R Coll Surg Engl 2008; 90(8): 658-62. Crowninshield R D, Maloney W J, Wentz D H, Humphrey S M, Blanchard C R. Biomechanics of large femoral heads: what they do and don’t do. Clin Orthop Relat Res 2004; (429): 102-7. Danoff J R, Bobman J T, Cunn G, Murtaugh T, Gorroochurn P, Geller J A, Macaulay W. Redefining the acetabular component safe zone for posterior approach total hip arthroplasty. J Arthroplasty 2016; 31(2): 506-11. Elkins J M, Callaghan J J, Brown T D. The 2014 Frank Stinchfield Award: The “landing zone” for wear and stability in total hip arthroplasty is smaller than we thought: a computational analysis. Clin Orthop Relat Res 2015; 473(2): 441-52. Espehaug B, Havelin L I, Engesaeter L B, Langeland N, Vollset S E. Patientrelated risk factors for early revision of total hip replacements: a population register-based case-control study of 674 revised hips. Acta Orthop Scand 1997; 68(3): 207-15. Gausden E B, Parhar H S, Popper J E, Sculco P K, Rush B N M. Risk factors for early dislocation following primary elective total hip arthroplasty. J Arthroplasty 2018; 33(5): 1567-1571.e2. Hailer N P, Weiss R J, Stark A, Karrholm J. The risk of revision due to dislocation after total hip arthroplasty depends on surgical approach, femoral head size, sex, and primary diagnosis: an analysis of 78,098 operations in the Swedish Hip Arthroplasty Register. Acta Orthop 2012; 83(5): 442-8. Hamilton W G, Parks N L, Huynh C. Comparison of cup alignment, jump distance, and complications in consecutive series of anterior approach and posterior approach total hip arthroplasty. J Arthroplasty 2015; 30(11): 1959-62. Harwin S F, Mistry J B, Chughtai M, Khlopas A, Gwam C, Newman J M, Higuera CA, Bonutti P M, Malkani A L, Kolisek F R, Delanois R E, Mont M A. Dual mobility acetabular cups in primary total hip arthroplasty in patients at high risk for dislocation. Surg Technol Int 2017; 30: 251-8. Hemmila M, Karvonen M, Laaksonen I, Matilainen M, Eskelinen A, Haapakoski J, Puhto A-P, Kettunen J, Manninen M, Makela K T. Survival of 11,390 Continuum cups in primary total hip arthroplasty based on data from the Finnish Arthroplasty Register. Acta Orthop 2019; 90(4): 312-7. Howie D W, Holubowycz O T, Middleton R. Large femoral heads decrease the incidence of dislocation after total hip arthroplasty: a randomized controlled trial. J Bone Joint Surg Am 2012; 94(12): 1095-102. Itokawa T, Nakashima Y, Yamamoto T, Motomura G, Ohishi M, Hamai S, Akiyama M, Hirata M, Hara D, Iwamoto Y. Late dislocation is associated with recurrence after total hip arthroplasty. Int Orthop 2013; 37(8): 145763. Klasan A, Neri T, Oberkircher L, Malcherczyk D, Heyse T J, Bliemel C. Complications after direct anterior versus Watson-Jones approach in total hip arthroplasty: results from a matched pair analysis on 1408 patients. BMC Musculoskelet Disord 2019; 20(1): 77. Laaksonen I, Lorimer M, Gromov K, Eskelinen A, Rolfson O, Graves S E, Malchau H, Mohaddes M. Trabecular metal acetabular components in primary total hip arthroplasty. Acta Orthop 2018; 89(3): 259-64. Magill P, Blaney J, Hill J C, Bonnin M P, Beverland D E. Impact of a learning curve on the survivorship of 4802 cementless total hip arthroplasties. Bone Joint J 2016; 98-B(12): 1589-96. Meek R M D, Allan D B, McPhillips G, Kerr L, Howie C R. Late dislocation after total hip arthroplasty. Clin Med Res 2008; 6(1): 17-23. Mohaddes M, Björk M, Nemes S, Rolfson O, Jolbäck P, Kärrholm J. No increased risk of early revision during the implementation phase of new cup designs. Acta Orthop 2016; 87 (eSuppl 362): 31-6. Peltola M, Malmivaara A, Paavola M. Hip prosthesis introduction and early revision risk: a nationwide population-based study covering 39,125 operations. Acta Orthop 2013; 84(1): 25-31.

Acta Orthopaedica 2020; 91 (3): 279â&#x20AC;&#x201C;285 285

Ravi B, Jenkinson R, Austin P C, Croxford R, Wasserstein D, Escott B, Paterson J M, Kreder H, Hawker G A. Relation between surgeon volume and risk of complications after total hip arthroplasty: propensity score matched cohort study. BMJ 2014; 348:g3284. Reito A, Puolakka T, Paakkala A, Pajamaki J. Assessment of inter- and intra-observer reliability in the determination of radiographic version and inclination of the cup in metal-on-metal hip resurfacing. Int Orthop 2012; 36(3): 519-25. Sariali E, Lazennec J Y, Khiami F, Catonne Y. Mathematical evaluation of jumping distance in total hip arthroplasty: influence of abduction angle, femoral head offset, and head diameter. Acta Orthop 2009; 80(3): 277-82.

Seagrave K G, Troelsen A, Madsen B G, Husted H, Kallemose T, Gromov K. Can surgeons reduce the risk for dislocation after primary total hip arthroplasty performed using the posterolateral approach? J Arthroplasty 2017; 32(10): 3141-6. Shrier I, Platt R W. Reducing bias through directed acyclic graphs. BMC Med Res Methodol 2008; 8:70. Tsikandylakis G, Karrholm J, Hailer N P, Eskelinen A, Makela K T, Hallan G, Furnes O N, Pedersen A B, Overgaard S, Mohaddes M. No increase in survival for 36-mm versus 32-mm femoral heads in metal-on-polyethylene THA: a registry study. Clin Orthop Relat Res 2018; 476(12): 2367-78.

286

Acta Orthopaedica 2020; 91 (3): 286–292

Postoperative 30-day complications after cemented/hybrid versus cementless total hip arthroplasty in osteoarthritis patients > 70 years A multicenter study from the Lundbeck Foundation Centre for Fast-track Hip and Knee replacement database and the Danish Hip Arthroplasty Register Martin LINDBERG-LARSEN 1,2, Pelle Baggesgaard PETERSEN 3, Christoffer Calov JØRGENSEN 1,3, Søren OVERGAARD 2, and Henrik KEHLET 1,3, + Lundbeck Foundation Center for Fast-track Hip and Knee Arthroplasty Collaborating Group 1 Lundbeck Foundation Centre for Fast-track Hip and Knee Arthroplasty; 2 Orthopaedic Research Unit, Department of Orthopaedic Surgery and Traumatology, Odense University Hospital, Department of Clinical Research, University of Southern Denmark; 3 Section for Surgical Pathophysiology, Rigshospitalet, Copenhagen, Denmark Members of the Lundbeck Foundation Centre for Fast-track Hip and Knee Collaborative Group: Frank Madsen, Department of Orthopedics, Aarhus University Hospital, Aarhus, Denmark; Torben B Hansen, Department of Orthopedics, Regional Hospital Holstebro and University of Aarhus, Holstebro, Denmark; Kirill Gromov, Department of Orthopedics, Hvidovre Hospital, Hvidovre, Denmark; Mogens Laursen, Aalborg University Hospital Northern Orthopaedic Division, Aalborg, Denmark; Lars T Hansen, Department of Orthopedics, Sydvestjysk Hospital Esbjerg/Grindsted, Grindsted, Denmark; Per Kjærsgaard-Andersen, Department of Orthopedics, Vejle Hospital, Vejle, Denmark; Soren Solgaard, Department of Orthopedics, Gentofte University Hospital, Copenhagen, Denmark; Niels Harry Krarup, Department of Orthopedics, Viborg Hospital, Viborg, Denmark; Jens Bagger, Department of Orthopaedic Surgery, Copenhagen University Hospital Bispebjerg, Copenhagen NV, Denmark. Correspondence: martin.lindberg-larsen@rsyd.dk Submitted 2019-11-08. Accepted 2020-03-06.

Background and purpose — The use of cementless total hip arthroplasty (THA) in elderly patients is debated because of increased risk of early periprosthetic femoral fractures. However, cemented femoral components carry a risk of bone cement implantation syndrome. Hence, we compared in-hospital complications, complications leading to readmission and mortality ≤ 30 days postoperatively between hybrid/ cemented (cemented femoral component) vs. cementless THA in osteoarthritis patients > 70 years. Patients and methods — This is a prospective observational cohort study in 9 centers from January 2010 to August 2017. We used 30-day follow-up from the Danish National Patient Registry, patient records, and data from the Danish Hip Arthroplasty Register. Only THAs performed as a result of osteoarthritis were included. Results — 3,368 (42%) of the THAs were cemented/ hybrid and 4,728 (58%) cementless. The in-hospital complication risk was 7.7% after cemented/hybrid vs. 5.3% after cementless THA (< 0.001), statistically not significant when adjusting for comorbidities (p = 0.1). There were similar risks of complications causing readmission (5.7% vs. 6.2%) and mortality ≤ 30 days (0.2% vs. 0.3%). 15 cases (0.4%) of pulmonary embolism (PE) were found after cemented/ hybrid vs. 4 (0.1%) after cementless THA (p = 0.001); none occurred within 24 hours postoperatively. 2 of the PEs after cementless THA led to mortality. Cemented/hybrid THA

remained significantly associated with risk of PE (RR 3.9, p = 0.02), when adjusting for comorbidities. BMI > 35 was associated with highest risk of PE (RR 5.7, p = 0.003). The risk of periprosthetic femoral fracture was 0.2% after cemented/hybrid vs. 1.5% after cementless THA (p < 0.001) and the risk of dislocations was 1.2% after cemented/hybrid THA vs. 1.8% after cementless THA (p = 0.04). Interpretation — The higher risk of PE after cemented/ hybrid THA and higher risk of periprosthetic femoral fractures and dislocations after cementless THA highlights that both medically and surgically complications are related to fixation technique and have to be considered.

The use of a cementless fixation technique in total hip arthroplasties (THA) is generally preferred compared with a cemented technique in patients with hip osteoarthrosis and age < 70 years in most Western countries. The use of cementless implants in THA has increased in all age groups including the elderly (AOANJRR 2018, DHR 2018, NJR 2018) despite the fact that cementless femoral components are associated with increased risk of early periprosthetic femoral fractures, especially in patients > 70 years (Makela et al. 2014, Thien et al. 2014, Lindberg-Larsen et al. 2017, Tanzer et al. 2018).

Acta Orthopaedica 2020; 91 (3): 286–292 287

Primary THA in patients > 70 years registered in DNPR and LCDB n = 9,037 Excluded (n = 941): – no DHR data, 168 – indication not osteoarthritis a, 722 – no data on fixation technique, 17 – reverse hybrid THA b, 34 Analyzed THAs (n = 8,096): – cementless, 4,728 – cemented/hybrid, 3,368

Figure 1. Flow-chart inclusion. DNPR: Danish National Patient Register; LCDB: Lundbeck Foundation Centre for Fast-track THA and TKA Database; DHR: Danish Hip Arthroplasty Register. a Including fracture, rheumatoid arthritis, tumor cases, and sequelae after fracture. b Cemented acetabular component and cementless femoral component.

A concern regarding the use of cemented femoral components is the risk of the so-called bone cement implantation syndrome. The syndrome occurs at the time of bone cementation and prosthesis insertion where the high intra-medullary pressure forces medullary fat emboli into the circulation resulting in risk of pulmonary embolism (Orsini et al. 1987, Donaldson et al. 2009, Segerstad et al. 2019). The clinical presentation of this syndrome may range from hypoxia to fulminatory pulmonary and systemic marrow embolization, right ventricular failure, and circulatory collapse (Issack et al. 2009). The incidence has been difficult to determine due to the variation of severity of the syndrome. In a series of 1,016 cemented hemiarthroplasty patients, the incidence of the most severe grade of the syndrome (grade 3) was 1.7% and associated with mortality of 88%. Risk factors for the syndrome were ASA grade III–IV, chronic obstructive pulmonary disease, and medication with diuretics or warfarin (Olsen et al. 2014). However, the exact incidence and consequences of bone cement implantation syndrome after total hip arthroplasty remains unknown. As occurrence of severe bone cement implantation syndrome would likely influence postoperative length of hospital stay, readmission risk, or mortality, we compared in-hospital complications, complications leading to readmission, and mortality within 30 days postoperatively between hybrid/cemented (cemented femoral component) versus cementless THA in osteoarthritis patients > 70 years.

Patients and methods This is a prospective observational cohort study on patients from the Lundbeck Foundation Centre for Fast-track THA and TKA database (LCDB 2019). The STROBE guidelines for reporting observational studies were followed. Data on 9,037 primary unselected elective THA procedures performed in patients > 70 years were prospectively collected between February 2010 and August 2017 in 9 departments reporting to

the LCDB. Supplementary data on fixation technique, duration of surgery, indication, and previous surgeries were available from the Danish Hip Arthroplasty Registry (DHR 2018) in 98% of cases. Hence, 8,096 primary elective THA procedures performed in patients with primary osteoarthritis were available for analysis after exclusions (Figure 1). The indication for either cemented/hybrid or cementless THA was based on surgeon and center preference and varied between centers. The percentage of cemented/hybrid THA ranged from 4% to 95% with great interdepartmental variation. The procedures were divided into cemented/hybrid (cemented femoral components) versus cementless THA (cementless femoral and acetabular components). Reverse hybrid THA procedures (cementless femoral component and cemented acetabular component) were excluded. Patients registered in the DHR as operated on for indications other than osteoarthritis were excluded. Furthermore, patients having simultaneous bilateral procedures or an additional major arthroplasty procedure within 90 days were excluded (Figure 1). The Lundbeck Foundation Centre for Fast-track THA and TKA Database (LCDB) The LCDB is a prospective database on unselected consecutive primary hip and knee arthroplasty procedures from 9 contributing centers initiated in January 2010 in 6 Danish centers, and an additional 3 centers joined the collaboration in 2012, 2013, and 2014 (LCDB 2019). The LCDB contains data on preoperative comorbidity and patient characteristics such as pharmacologically treated cardiopulmonary disease, diabetes, previous venous thromboembolism (VTE), alcohol consumption, smoking habits, and living conditions (Table 1). Data are prospectively collected from patients within 1 month before surgery using self-completed questionnaires with staff available for assistance. Validation of the consistency of the preoperative patient questionnaire has been performed previously using matched patient medical records (Pitter et al. 2016). The study period for the current updated detailed analysis was from January 2010 until August 2017. All collaborating centers adhered to similar fast-track protocols, including use of spinal anesthesia (≈90%), multi-modal opioid-sparing analgesia, early mobilization, and discharge to own home based on functional discharge criteria. Thromboprophylaxis was prescribed according to local guidelines including thromboprophylaxis administered 6–8 hours after surgery, consisting of either rivaroxaban (Xarelto, Bayer Pharma, Berlin, Germany) 10 mg/day, enoxaparin (Klexane, Sanofi-Aventis, Paris, France) 4,000 IU/day, dalteparin (Fragmin, Pfizer Health Care, New York, USA) 5,000 IU/day, or fondaparinux (Arixtra, GlaxoSmithKline, London, UK) 2.5 mg/day. Thromboprophylaxis was used only during hospitalization if LOS was ≤ 5 day. If LOS > 5 days recommendations on duration varied in the study period. From 2010 to 2016 international guidelines with thromboprophylaxis for up to 35 days for

288

Acta Orthopaedica 2020; 91 (3): 286–292

Table 1. Patient characteristics. Values are number (%) unless otherwise specified Cemented/ hybrid THA a Patient characteristics n = 3,368 (42%) Mean age (range) Female sex Mean BMI (range) missing Use of walking aid missing Smoking missing Alcohol > 2 units/day missing Type 1 diabetes missing Type 2 diabetes missing Pharmacologically treated cardiovascular disease missing Pharmacologically treated pulmonary disease missing Prior cerebral stroke missing History of VTE missing Hereditary VTE missing Hypercholesterolemia missing Anticoagulative treatment missing Preoperative anemia b missing Hypertension missing Pharmacologically treated psychiatric disorder missing a Hybrid = cemented b Hb < 13 g/dL.

79 (71–100) 2,311 (69) 26 (14–64) 24 (0.7) 1,374 (41) 78 (2.3) 341 (10) 23 (0.7) 114 (3.4) 36 (1.1) 25 (0.7) 22 (0.7) 319 (9.5) 20 (0.6)

Cementless THA n = 4,728 (58%) 76 (71–97) 2,641 (56) 26 (15–61) 29 (0.6) 1,421 (30) 98 (2.1) 448 (9.5) 27 (0.6) 375 (7.9) 26 (0.5) 24 (0.5) 25 (0.5) 463 (9.8) 11 (0.2)

664 (20) 42 (1.2)

831 (18) 40 (0.8)

362 (11) 24 (0.7) 219 (6.5) 65 (1.9) 261 (7.7) 95 (2.8) 362 (11) 347 (10) 1,422 (42) 25 (0.7) 689 (21) 32 (1.0) 1,193 (35) 23 (0.7) 1,994 (59) 11 (0.3)

421 (8.9) 26 (0.5) 350 (7.4) 60 (1.3) 367 (7.8) 60 (1.3) 470 (9.9) 623 (13) 1,780 (38) 24 (0.5) 850 (18) 49 (1.0) 1,313 (28) 34 (0.7) 2,689 (57) 20 (0.4)

259 (7.7) 28 (0.8)

303 (6.4) 23 (0.5)

femoral component

THA: total hip arthroplasty VTE: venous thromboembolism

THA were recommended and from 2016 the Danish recommendation changed to 6–10 days (RADS 2016). Compression stockings or intermittent pneumatic compression devices were not used. Pulsatile lavage was used as a standard procedure in cemented THA surgery in the participating centers. Outcomes Complications within 30 days postoperatively leading to length of hospital stay (LOS) > 4 days, readmission, and mortality were analyzed. All patients with a LOS of > 4 days had their medical records examined to determine the reason for prolonged LOS and in-hospital complications during primary admission. Using the patients’ unique Danish social security numbers, we obtained information on LOS, 30-day

readmissions, and mortality from the Danish National Patient Registry. As reporting to the Danish National Patient Registry is mandatory for hospitals to receive reimbursement, an almost complete follow-up (> 99%) is assured (Schmidt et al. 2015). All unplanned readmissions with an overnight hospital stay within 30 days postoperatively were evaluated to determine postoperative complications after discharge. Discharge records or patient records of readmitted patients were scrutinized and relation to index surgery was evaluated. Causes of mortality within 30 days were evaluated using patient records. In the case of pulmonary embolism (PE) in the discharge summary the entire medical record was evaluated to confirm the occurrence of PE and identify any possible causative events. The PE diagnosis was in all cases confirmed by spiral computed tomography (CT). Statistics Overall comparative analysis of the groups was performed using Student’s t-test for normally distributed data and crude comparisons of proportions were done using a chi-square test. Analysis of potential risk factors associated with in-hospital complications and PE was performed using a multivariable Poisson regression model with robust error variance (Zou 2004). Cases with missing data were excluded from analysis. A sensitivity analysis on risk of PE, periprosthetic fractures, and dislocations after excluding 165 (85+80) cemented/hybrid THA from 2 centers performing < 10% cemented/hybrid THA and 14 cementless THA from 1 center performing < 10% cementless THA was performed. Results are given as relative risk (RR) estimates or percentages with 95% confidence intervals (CI). Any p-value of < 0.05 was considered significant. Analysis was done using SPSS version 22 (IBM Corp, Armonk, NY, USA). Ethics, registration, funding, and potential conflicts of interests According to Danish law no approval from the regional ethics committee was required as the study was non-interventional. Permission to obtain and store data without informed consent was obtained from the Danish Patient Safety Authority (3-3013-56/2/EMJO) and the Danish Data Protection Agency (RH-2014-132). The LCDB is registered as an ongoing study registry on clinicaltrials.gov (NCT01515670). This study was funded by the Lundbeck Foundation Centre for Fast-Track Hip and Knee Arthroplasty, Copenhagen, Denmark. No competing interests were declared.

Results 8,096 primary elective THA procedures performed in 7,786 osteoarthritis patients > 70 years were analyzed (Figure 1). In 42% of the procedures a cemented or hybrid (cemented femoral component) fixation technique was used and 58%

Acta Orthopaedica 2020; 91 (3): 286–292 289

Table 2. Postoperative 30-day outcome. Values are count (%) Outcomes Mortality In-hospital complications Complications causing readmission Pulmonary embolism Periprosthetic femoral fractures Dislocations

Cemented/ Cementless hybrid THA THA p-value n = 3,368 n = 4,728 8 (0.2) 260 (7.7)

15 (0.3) 0.5 252 (5.3) < 0.001

193 (5.7) 15 (0.4) 7 (0.2) 42 (1.2)

292 (6.2) 0.4 4 (0.1) 0.001 70 (1.5) < 0.001 87 (1.8) 0.04

Cardiac Pulmonary Venous thromboembolism Cerebral Renal/fluid related Urological Gastrointestinal Anemia Other medical complications Pain/mobilization problems

Cementless Cemented/hybrid

Hip dislocation Periprosthetic fracture Oozing/hematoma Periprosthetic infection Other surgical complications 0

Table 3. Multivariable Poisson regression analysis of potential risk factors influencing the risk of in-hospital complications Potential risk factor

RR (95 % CI)

Cemented/hybrid THA Pharmacologically treated cardiovascular disease Pharmacologically treated pulmonary disease History of venous thromboembolism BMI ≥ 35 Use of mobility aid Age (per 10 years)

1.1 (0.9–1.3) 1.5 (1.2–1.8) 1.4 (1.1–1.8) 1.6 (1.2–2.0) 2.1 (1.5–2.9) 2.1 (1.8–2.6) 2.1 (1.8–2.5)

RR: Relative risk; BMI: body mass index.

were performed using a cementless technique. There were more females in the cemented/hybrid group (69%) compared with the cementless group (56%), but otherwise demographics were similar between groups (Table 1). The in-hospital complication rate was 7.7% after cemented/ hybrid THA vs. 5.3% after cementless THA (< 0.001) (Table 2), but statistically not significant (RR 1.1, p = 0.2) when adjusting for preoperative comorbidity (Table 3). The most frequent in-hospital complication causing prolonged LOS > 4 days was pain and mobilization problems after both cemented/ hybrid and cementless THA (1.5%, n = 50 and 1.1%, n = 50 respectively). All in-hospital complications are presented in Figure 2. The 30-day overall readmission risk was 6.5% after cemented/hybrid vs. 7.0% after cementless THA. When subtracting readmissions due to suspected but disproven VTE/ infections the readmission risk was 5.7% after cemented/ hybrid vs. 6.2% after cementless THA (p = 0.4). The most frequent complications causing readmission were hip dislocations (0.8%, n = 27), periprosthetic joint infection (0.7%, n = 24), and cardiac complications (0.7%, n = 22) after cemented/ hybrid THA whereas hip dislocations (1.6%, n = 78) and periprosthetic femoral fractures (1.0%, n = 46) were most frequent after cementless THA (Figure 3). 7 cases of PE during primary admission and 8 PEs leading to readmission ≤ 30 days were found after cemented/hybrid THA compared with 2 cases of PE during primary admission and 2 PE leading to readmission after cementless THA.

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

Complications (%)

Figure 2. In-hospital complications causing LOS > 4 days.

Cardiac Pulmonary Venous thromboembolism Cerebral Renal/fluid related Urological Gastrointestinal Anemia Other medical complications Pain/mobilization problems Fall related Susp. DVT/infection, not found

Cementless Cemented/hybrid

Hip dislocation Periprosthetic fracture Oozing/hematoma Periprosthetic infection Other surgical complications 0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

Complications (%)

Figure 3. Complications causing readmission ≤ 30 days postoperatively.

Pulmonary embolism (PE) and time course from surgery as well as information on thromboprophylaxis at time of PE are presented in Table 4. 2 of the PEs after cementless THA led to mortality. Hence, 15 cases (0.4%) of PE were found after cemented/hybrid vs. 4 (0.1%) after cementless THA within 30 days postoperatively (p = 0.001). None of the PEs occurred within 24 hours postoperatively. Cemented/hybrid THA remained significantly associated (RR 3.9, p = 0.022) with risk of PE in the regression model (Table 5). 2 cases of periprosthetic femoral fractures during primary admission and 5 leading to readmission ≤ 30 days were found after cemented/hybrid THA compared with 24 periprosthetic femoral fractures during primary admission and 46 leading to readmission after cementless THA (Figures 2–3). Hence, 7 cases (0.2%) of periprosthetic femoral fractures were found after cemented/hybrid vs. 70 (1.5%) after cementless THA within 30 days postoperatively (p < 0.001). In 1 of 7 cases of periprosthetic fractures after cemented/hybrid THA revision surgery was needed and the remaining 6 cases were treated nonoperatively. In contrast, 39 of 70 cases of periprosthetic fracture after cementless THA needed revision surgery and the

290

Acta Orthopaedica 2020; 91 (3): 286–292

Table 4. Pulmonary embolism (PE) and time course from surgery Days from surgery to PE

Place of PE diagnosis

Cemented/hybrid (n = 15) 2 Primary admission 2 Primary admission 2 Primary admission 2 Primary admission 3 Readmission 3 Primary admission 4 Primary admission 7 Readmission 8 Readmission 10 Readmission 13 Readmission 14 Readmission 23 Primary admission 26 Readmission 27 Readmission Cementless (n = 4) 1 Primary admission 2 Readmission 14 Primary admission 17 Readmission

Ongoing Preoperative thrombo- Previous anticoagulant PE-related prophylaxis VTE therapy a Age mortality Yes Yes Yes Yes Yes Yes No Yes Yes Yes No No No No No Yes Yes Yes Yes No

85 75 86 81 77 78 82 84 92 73 71 87 79 84 84

Yes Yes Yes Yes No

73 78 Yes 71 Yes 89

a All

3 patients treated with warfarin pre- and postoperatively. VTE: venous thromboembolism.

Table 5. Multivariable Poisson regression analysis of potential risk factors influencing the risk of pulmonary embolism (PE) Potential risk factor

RR (95% CI)

Cemented and hybrid THA Pharmacologically treated cardiovascular disease Pharmacologically treated pulmonary disease History of venous thromboembolism BMI ≥ 35 Use of mobility aid Age (per 10 years)

3.9 (1.2–13) 1.9 (0.7–5.3) 1.0 (0.2–4.8) 2.0 (0.6–6.6) 5.7 (1.6–20) 1.6 (0.6–4.4) 1.6 (0.6–4.2)

remainder were treated nonoperatively (25 cases) or treated at primary surgery with wiring/osteosynthesis (6 cases). The risk of dislocations was 1.2% (15 during primary admission and 27 causing readmission) after cemented/hybrid THA vs. 1.8% (9 during primary admission and 78 causing readmission) after cementless THA (p = 0.04). A sub-analysis on risk of dislocations comparing groups based on acetabular fixation was performed and the risk of dislocations was similar: 1.6% (5 during primary admission and 3 causing readmission) after the use of cemented vs. 1.6% (19 during primary admission and 102 causing readmission) after cementless acetabular components. The postoperative mortality ≤ 30 days was 0.2% after cemented/hybrid vs. 0.3% after cementless THA (p = 0.5). Time course and causes of 30-day mortality are presented in Table 6.

A sensitivity analysis on risk of PE, periprosthetic fractures, and dislocations after excluding 165 (85+80) cemented/hybrid THA from 2 centers performing < 10% cemented/ hybrid THA and 14 cementless THA from 1 center performing < 10% cementless THA revealed no differences from the risk estimates presented above.

Discussion The use of cementless THA implants in patients > 70 years is debatable due to the increased risk of early periprosthetic fractures (Makela et al. 2014, Lindberg-Larsen et al. 2017, Tanzer et al. 2018). One of the arguments for the use of cementless implants may be the risk of bone cement implantation syndrome (Olsen et al. 2014) and the potentially associated mortality after cementation (McMinn et al. 2012, Garland et al. 2017). However, the actual impact of the use of cement on the overall risk of postoperative

Table 6. Mortality, time course from surgery and cause of death Days from surgery to death

Place of death

Cemented/hybrid THA (n = 8) 2 Primary admission 3 Primary admission 4 Home 8 Primary admission 13 Readmission 22 Readmission 25 Primary admission 29 Readmission Cementless THA (n = 15) 2 Readmission 2 Readmission 3 Primary admission 5 Readmission 5 Home 6 Primary admission 8 Primary admission 13 Readmission 13 Home 14 Readmission 14 Primary admission 14 Primary admission 18 Readmission 23 Readmission 23 Readmission

Age

Cause

86 Ileus 76 Respiratory insufficiency 100 Cardiac arrest (unknown) 79 Stroke 77 Myocardial infarction 90 Renal insufficiency 79 Stroke 88 Endocarditis 78 Pulmonary embolism 71 Cardiac arrest 78 COPD, sepsis, ARI 82 UTI, sepsis 82 Cardiac arrest (arrhythmia) 79 Pneumonia 87 Stroke 88 Stroke 74 Unknown 71 Pulmonary embolism 87 Pneumonia 86 Myocardial infarctionI 71 Unknown 90 Pneumonia 94 Pneumonia

COPD: chronic obstructive pulmonary disease; ARI: acute renal insufficiency; UTI: urinary tract infection

Acta Orthopaedica 2020; 91 (3): 286–292 291

complications has not been described in detail previously. In this study, we report a higher risk of in-hospital complications after cemented/hybrid THA (cemented femoral component) compared with cementless THA (7.7% vs. 5.3%); however, this was unrelated to fixation technique when adjusting for preoperative patient comorbidities. Furthermore, an increased risk of 30-days postoperative PE after cemented/hybrid compared with cementless THA was found, and also when adjusting for comorbidities. The higher risk of PE can hardly be directly explained by cementation as none happened on the day of surgery or the first day after and it seems unlikely that a potential low-grade bone cement implantation syndrome during surgery can affect the cardiopulmonary system, making the patient more at risk of PE later. The lack of a direct and timely relationship between cement implantation and the occurrence of postoperative PE suggests that other factors, such as obesity and delayed mobilization, may also contribute (Table 4). Furthermore, many of the PE events happened despite ongoing thromboprophylaxis, indicating that these events happen in high-risk patients, requiring further investigations as discussed in detail by Petersen et al. (2019). We therefore feel that the term “bone cement implantation syndrome” should be used with more caution in this context. Whether the patients suffering PE in our study had detectable bone cement implantation syndrome intraoperatively cannot be confirmed in this study. Hence, our finding of clinically documented postoperative PE of 0.4% after cemented/ hybrid THA is significantly higher than after cementless THA, but still much lower compared with the paraclinically reported incidence (28% [all grades]) of bone cement implantation syndrome (Olsen et al. 2014). Postoperative mortality in relation to the use of a cemented fixation technique in THA has been analyzed on a larger scale using data from the Swedish Hip Arthroplasty Register (n = 178,784), finding a minimally increased relative risk of early mortality < 15 days postoperatively after cemented/hybrid THA compared with a matched control group from the general population, but the relative risk of mortality reversed > 15 days postoperatively (Garland et al. 2017). We found similar mortality rates between groups, but a larger cohort is needed to detect potentially statistically significant differences in mortality. The higher risk of periprosthetic fractures after cementless THA, especially in the elderly (Makela et al. 2014, Thien et al. 2014, Lindberg-Larsen et al. 2017), is confirmed in our study, amounting to 1.5% after cementless THA compared with 0.2% after cemented/hybrid THA. Periprosthetic fractures may cause protracted mobilization with limited weightbearing and when revision surgery is needed the procedures are complex and often associated with a high risk of complications and mortality (Bhattacharyya et al. 2007, Griffiths et al. 2013). Our study includes only 30 days’ follow-up and hence more periprosthetic fractures, especially after cementless THA, with associated patient morbidity may be recognized later (Lindberg-Larsen et al. 2017).

The higher risk of postoperative dislocations after cementless THA compared with a cemented/hybrid technique is in line with the higher revision rates due to dislocations reported in the Danish Hip Arthroplasty Register over the past decades after a cementless fixation technique for the femoral and acetabular components (DHR 2018). A higher risk of dislocations after cementless THA has previously been explained by more variation in the positioning of the cementless acetabular component (Chawda et al. 2009). However, we found that the risk of dislocation was similar (1.6%) when comparing cementless versus cemented acetabular components in our cohort in contrast to previous findings, supporting that the risk of dislocation is multifactorial. The high proportion of cementless THA procedures (58%) performed in our population > 70 years is surprising, but in line with 53% cementless THAs performed in Denmark in 2017 overall in the same age group (DHR 2018). This finding confirms the overall Danish trend of using cementless THA in all age groups. A limitation of this study is confounding by indication as no similar indication criteria for the use of a cemented fixation technique existed between centers. We tried to limit this by performing a sensitivity analysis on risk of PE, periprosthetic fractures, and dislocations after excluding cemented/hybrid THA from 2 centers performing < 10% cemented/hybrid THA and cementless THA from 1 center performing < 10% cementless THA. However, the results from the sensitivity analysis did not differ from the main analysis. Although the registered preoperative patient comorbidity was similar, other confounding factors such functional status and bone quality (Dorr type) (Nash and Harris 2014) might have influenced surgeons’ choice. The type of cementless femoral component used may also have influenced the risk of periprosthetic fractures (Gromov et al. 2017) and this was not analyzed in the present study. Furthermore, the few PE events might lead to overfitting of the multivariable Poisson risk-factor analysis, potentially introducing a “type II error.” Another limitation of the study is the observational nature that excludes the possibility to report on causality. However, our study provides detailed information on complications leading to prolonged LOS or readmissions after cemented/hybrid THA versus cementless THA, which is necessary for shared decision-making between surgeons and patients. A strength of this study includes prospective recording of numerous baseline characteristics, thus minimizing recall bias, an unselected consecutive patient population, and follow-up through a high-quality nationwide register, thereby ensuring data completeness (Schmidt et al. 2015). Other strengths are the standardized multicenter fast-track setup, and the relatively short and recent study period. Furthermore, not relying on only diagnostic codes but using review of discharge notes and patient records for specific causes of morbidity ensures > 99% follow-up for somatic readmissions (Schmidt et al. 2015) and eliminates the dependency on the questionable reliability

292

of a discharge diagnosis within the DNPR (Severinsen et al. 2010, Sundboll et al. 2016, Bedard et al. 2018). In conclusion, we report a higher risk of PE after cemented/ hybrid THA, but the risk of periprosthetic femoral fractures and dislocations was higher after cementless THA. These findings highlight that, both medically and surgically, complications are related to fixation technique and have to be considered in the decision-making on an individual level.

MLL, PBP, CJ, SO, and HK wrote the protocol. MLL, PBP, and CJ undertook all data gathering. MLL performed all analyses. MLL wrote the first draft of the manuscript and all authors including the collaborators revised the draft. Acta thanks Steffen Breusch and B Willem Schreurs for help withÂ peer review of this study.

AOANJRR. Australian Orthopaedic Association National Joint Replacement Registry, Annual Report 2018. https://aoanjrr.sahmri.com/. Bedard N A, Pugely A J, McHugh M A, Lux N R, Bozic K J, Callaghan J J. Big data and total hip arthroplasty: how do large databases compare? J Arthroplasty 2018; 33: 41-45 e3. Bhattacharyya T, Chang D, Meigs J B, Estok D M 2nd, Malchau H. Mortality after periprosthetic fracture of the femur. J Bone Joint Surg Am 2007; 89: 2658-62. Chawda M, Hucker P, Whitehouse S L, Crawford R W, English H, Donnelly W J. Comparison of cemented vs uncemented acetabular component positioning using an imageless navigation system. J Arthroplasty 2009; 24: 1170-3. DHR. The Danish Hip Arthroplasty Register, Annual Report 2018. http:// www.dhr.dk/. Donaldson A J, Thomson H E, Harper N J, Kenny N W. Bone cement implantation syndrome. Br J Anaesth 2009; 102: 12-22. Garland A, Gordon M, Garellick G, Karrholm J, Skoldenberg O, Hailer N P. Risk of early mortality after cemented compared with cementless total hip arthroplasty: a nationwide matched cohort study. Bone Joint J 2017; 99-B: 37-43. Griffiths E J, Cash D J, Kalra S, Hopgood P J. Time to surgery and 30-day morbidity and mortality of periprosthetic hip fractures. Injury 2013; 44: 1949-52. Gromov K, Bersang A, Nielsen C S, Kallemose T, Husted H, Troelsen A. Risk factors for post-operative periprosthetic fractures following primary total hip arthroplasty with a proximally coated double-tapered cementless femoral component. Bone Joint J 2017; 99-B: 451-57. Issack P S, Lauerman M H, Helfet D L, Sculco T P, Lane J M. Fat embolism and respiratory distress associated with cemented femoral arthroplasty. Am J Orthop 2009; 38: 72-6. LCDB. Lundbeck Foundation Centre for Fast-track Hip and Knee Arthroplasty; 2019. http://www.FTHK.dk. Lindberg-Larsen M, Jorgensen C C, Solgaard S, Kjersgaard A G, Kehlet H. Increased risk of intraoperative and early postoperative periprosthetic femoral fracture with uncemented stems. Acta Orthop 2017; 88: 390-94.

Acta Orthopaedica 2020; 91 (3): 286â&#x20AC;&#x201C;292

Makela K T, Matilainen M, Pulkkinen P, Fenstad A M, Havelin L, Engesaeter L, Furnes O, Pedersen A B, Overgaard S, Karrholm J, Malchau H, Garellick G, Ranstam J, Eskelinen A. Failure rate of cemented and uncemented total hip replacements: register study of combined Nordic database of four nations. BMJ 2014; 348: f7592. McMinn D J, Snell K I, Daniel J, Treacy R B, Pynsent P B, Riley R D. Mortality and implant revision rates of hip arthroplasty in patients with osteoarthritis: registry based cohort study. BMJ 2012; 344: e3319. Nash W, Harris A. The Dorr type and cortical thickness index of the proximal femur for predicting peri-operative complications during hemiarthroplasty. J Orthop Surg (Hong Kong) 2014; 22: 92-5. NJR. National Joint Registry for England and Wales (NJR England Wales), Annual Report 2018. http://www.njrcentre.org.uk/njrcentre/default.aspx. Olsen F, Kotyra M, Houltz E, Ricksten S E. Bone cement implantation syndrome in cemented hemiarthroplasty for femoral neck fracture: incidence, risk factors, and effect on outcome. Br J Anaesth 2014; 113: 800-6. Orsini E C, Byrick R J, Mullen J B, Kay J C, Waddell J P. Cardiopulmonary function and pulmonary microemboli during arthroplasty using cemented or non-cemented components: the role of intramedullary pressure. J Bone Joint Surg Am 1987; 69: 822-32. Petersen P B, Jorgensen CC, Kehlet H. Venous thromboembolism despite ongoing prophylaxis after fast-track hip and knee arthroplasty: a prospective multicenter study of 34,397 procedures. Thromb Haemost 2019; 119: 1877-85. Pitter F T, Jorgensen C C, Lindberg-Larsen M, Kehlet H. Postoperative morbidity and discharge destinations after fast-track hip and knee arthroplasty in patients older than 85 years. Anesth Analg 2016; 122: 1807-15. RADS. Treatment recommendation for prevention of venous thromboembolism in hip and knee surgery. The Danish Council for the Use of Expensive Hospital Medicines; 2016. Schmidt M, Schmidt S A, Sandegaard J L, Ehrenstein V, Pedersen L, Sorensen H T. The Danish National Patient Registry: a review of content, data quality, and research potential. Clin Epidemiol 2015; 7: 449-90. Segerstad M H A, Olsen F, Patel A, Houltz E, Nellgard B, Ricksten S E. Pulmonary haemodynamics and right ventricular function in cemented vs uncemented total hip arthroplasty: a randomized trial. Acta Anaesthesiol Scand 2019; 63: 298-305. Severinsen M T, Kristensen S R, Overvad K, Dethlefsen C, Tjonneland A, Johnsen S P. Venous thromboembolism discharge diagnoses in the Danish National Patient Registry should be used with caution. J Clin Epidemiol 2010; 63: 223-8. Sundboll J, Adelborg K, Munch T, Froslev T, Sorensen H T, Botker H E, Schmidt M. Positive predictive value of cardiovascular diagnoses in the Danish National Patient Registry: a validation study. BMJ Open 2016; 6: e012832. Tanzer M, Graves S E, Peng A, Shimmin A J. Is cemented or cementless femoral stem fixation more durable in patients older than 75 years of age? A comparison of the best-performing stems. Clin Orthop Relat Res 2018; 476: 1428-37. Thien T M, Chatziagorou G, Garellick G, Furnes O, Havelin L I, Makela K, Overgaard S, Pedersen A, Eskelinen A, Pulkkinen P, Karrholm J. Periprosthetic femoral fracture within two years after total hip replacement: analysis of 437,629 operations in the Nordic Arthroplasty Register Association database. J Bone Joint Surg Am 2014; 96: e167. Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004; 159: 702-6.

Acta Orthopaedica 2020; 91 (3): 293–298

293

The role of bone cement for the development of intraoperative hypotension and hypoxia and its impact on mortality in hemiarthroplasty for femoral neck fractures Fredrik OLSEN, Mathias HÅRD AF SEGERSTAD, Bengt NELLGÅRD, Erik HOULTZ, and Sven-Erik RICKSTEN

Department of Anesthesiology and Intensive Care Medicine, Institute of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg and Sahlgrenska University Hospital, Gothenburg, Sweden Correspondence: sven-erik.ricksten@gu.se Submitted 2019-06-10. Accepted 2020-02-14.

Background and purpose — The bone cement implantation syndrome characterized by hypotension and/or hypoxia is a well-known complication in cemented arthroplasty. We studied the incidence of hypotension and/or hypoxia in patients undergoing cemented or uncemented hemiarthroplasty for femoral neck fractures and evaluated whether bone cement was an independent risk factor for postoperative mortality. Patients and methods — In this retrospective cohort study, 1,095 patients from 2 hospitals undergoing hemiarthroplasty with (n = 986) and without (n = 109) bone cementation were included. Pre-, intra-, and postoperative data were obtained from electronic medical records. Each patient was classified for grade of hypotension and hypoxia during and after prosthesis insertion according to Donaldson’s criteria (Grade 1, 2, 3). After adjustments for confounders, the hazard ratio (HR) for the use of bone cement on 1-year mortality was assessed. Results — The incidence of hypoxia and/or hypotension was higher in the cemented (28%) compared with the uncemented group (17%) (p = 0.003). The incidence of severe hypotension/hypoxia (grade 2 or 3) was 6.9% in the cemented, but not observed in the uncemented group. The use of bone cement was an independent risk factor for 1-year mortality (HR 1.9, 95% CI 1.3–2.7), when adjusted for confounders. Interpretation — The use of bone cement in hemiarthroplasty for femoral neck fractures increases the incidence of intraoperative hypoxia and/or hypotension and is an independent risk factor for postoperative 1-year mortality. Efforts should be made to identify patients at risk for BCIS and alternative strategies for the management of these patients should be considered.

The bone cement implantation syndrome (BCIS) is a wellrecognized and potentially fatal complication of orthopedic surgery involving pressurized bone cement (Donaldson et al. 2009). The syndrome is mostly noted in cemented hemiarthroplasty after displaced femoral neck fractures, but is also found in total hip and knee replacement surgery (Byrick et al. 1986, Clark et al. 2001). This syndrome is characterized by hypoxia, systemic hypotension, pulmonary hypertension, arrhythmias, loss of consciousness, and cardiac arrest (Clark et al. 2001, Kotyra et al. 2010). The pathophysiology of BCIS is unclear, but anaphylaxis, inflammatory, thermic and complement activation (Dahl et al. 1988) have all been implicated to induce BCIS (Donaldson et al. 2009). Studies employing invasive hemodynamic monitoring and perioperative ultrasound imaging have revealed subclinical pulmonary embolisms and hemodynamic changes, not detected in standard intra- and postoperative monitoring (Orsini et al. 1987, Bisignani et al. 2008, Kotyra et al. 2010). Until recently, the incidence of BCIS in cemented hemiarthroplasty for hip fractures has been unknown, mainly because a consensual definition of the BCIS syndrome has been lacking. Donaldson et al. (2009) defined a severity classification of BCIS (Grade 1, 2, and 3). In a previous study on patients undergoing cemented hemiarthroplasty for hip fractures, we found that the incidence for all grades of BCIS was 28%, with a huge impact on early and late mortality (Olsen et al. 2014). In this study, we evaluated the role of the cementation, per se, for the development of hypotension and hypoxia and its impact on mortality in patients undergoing hemiarthroplasty for femoral neck fractures. To enable this, a multitude of risk factors influencing mortality were collected in order to iso-

294

late the cementation effect. Our hypothesis was that the use of bone cement is an independent risk factor for postoperative mortality.

Patients and methods Patients undergoing hemiarthroplasty for femoral neck fracture in 2 hospitals in the south-west of Sweden, from January 2008 until July 2011, were included. The 1st cohort of patients was operated at Sahlgrenska University Hospital/Mölndal, which routinely performs cemented hemiarthroplasty in patients with femoral neck fracture. The second cohort was operated at Alingsås Regional Hospital, which routinely performs hemiarthroplasty without cement in patients with femoral neck fracture. The operative treatments were standard procedures at the respective hospitals. Patients were excluded when having a failed internal fixation or pathological fractures. When surgery was performed > 1 during the study period, only the initial operation was included. From the review of the medical records, data were obtained retrospectively by 1 of the investigators (FO), for pre-existing comorbidity, medication, and biochemical markers. Besides sex, age, preadmission residence (assisted living), functional status (reduced mobility), ASA risk score, and type of anesthesia (regional or general), we collected data regarding preoperative cardiac history and presence of coexisting diseases, including liver disease, renal failure (defined as a documented history of renal failure and/or a serum creatinine >150 µmol/L), diabetes mellitus, previous stroke, peripheral vascular disease, hypertension, chronic obstructive pulmonary disease, cancer, dementia, arrhythmias, anemia, and current drug therapy. The selection of these variables was guided by a comprehensive review of the literature and subject-matter knowledge. Thus, in the present study we were able to retrieve from medical records 85% of previously described demographic factors and comorbidity associated with postoperative mortality (Maxwell et al. 2008, Juliebo et al. 2010, Olsen et al. 2014, Sheehan et al. 2016). In all patients, anesthesia charts were reviewed for arterial oxygen saturation, mean arterial blood pressure, and heart rate. At the involved hospitals, these variables are recorded, routinely, immediately before induction of anesthesia and every 5th minute during the operation. In patients receiving cemented hemiarthroplasty, the cementation process is marked on the anesthesia chart. These variables were obtained on 4 occasions: (a) immediately prior to induction of anesthesia, (b) every 5th minute, for a period of 10–15 minutes before prosthesis insertion, (c) every 5th minute, for a period of at least 15 minutes after prosthesis insertion (with or without implantation of bone cement), and (d) on arrival at the post-anesthesia recovery unit. The lowest systolic blood pressure recorded within 15 minutes after prosthesis insertion was compared with pre-insertion values and used to score the severity of hypoxia/hypotension.

Acta Orthopaedica 2020; 91 (3): 293–298

Each patient was classified as having no hypotension/hypoxia (Grade 0) or Grade 1, 2, or 3 hypotension/hypoxia, according to the criteria of Donaldson et al., where Grade 1 was defined as moderate hypoxia (arterial oxygen saturation < 94%) or moderate hypotension (a decrease in systolic arterial pressure (SAP) > 20%), Grade 2 included severe hypoxia (arterial oxygen saturation < 88%) or severe hypotension (a decrease in SAP > 40%) or unexpected loss of consciousness, and finally Grade 3 was defined as a cardiovascular collapse requiring cardiopulmonary resuscitation Statistics For descriptive purposes, categorical variables are presented with n (%), and continuous variables with mean (SD). In a univariable Cox proportional hazard regression analysis, we screened for important risk factors that were significantly associated with 1-year mortality (see Table 3). Unadjusted hazard ratios (HR) with 95% confidence interval (CI) were calculated for each risk factor. In addition, the following previously described risk factors (see above) for mortality were added: diabetes, stroke, and previous myocardial infarction. Thus, regarding the causality of the use of cement on 1-year mortality, the following 21 confounding variables were adjusted for in a multivariate Cox proportional hazard regression model; sex, age, assisted living, reduced mobility, ASA score, liver disease, renal failure, diabetes, stroke, arteriosclerosis, peripheral vascular disease, previous myocardial infarction, congestive heart failure, cancer, arrythmia, anemia (hemoglobin < 100g/L), use of diuretics, statins, antiplatelets drugs or ß-adrenergic blockers, and dementia. The adjusted hazard ratio for the variable bone cementation with CI is presented. The variable hypotension and/or hypoxia was considered a mediator variable and was not adjusted for in the Cox regression. None of the variables were considered colliders. The assumption of Cox proportional hazard was checked by using scaled Schoenfeld residuals and found satisfactory. We also performed 2 sensitivity analyses to assess the robustness of our model and the role of residual confounding. In 1 sensitivity analysis, we performed a Cox regression including only the strongest covariates (n = 9) as found by us and others (see above), i.e., age, assisted living, ASA risk score, liver disease, renal failure, diabetes, stroke, myocardial infarction, and dementia. In another sensitivity analysis, we used the E-value methodology, as described by VanderWeele and Ding (2017), which estimates the minimum strength of association (risk ratio, 95% CI) that an unmeasured or uncontrolled confounding would need to have (on a risk ratio scale) with both the treatment and outcome to fully explain away an association between treatment and outcome. A large E-value indicates that considerable unmeasured confounding would be necessary to explain away an effect estimate. All tests were 2-tailed and conducted at 5% significance level. Statistical analysis was performed using SPSS version 25 (IBM Corp, Armonk, NY, USA).

Acta Orthopaedica 2020; 91 (3): 293–298

Ethics, data-sharing, funding, and potential conflicts of interest The study was approved by the Gothenburg Regional Ethics Committee and written informed consent was waived by the Ethics Committee (IRB#887-16). Data may be shared on reasonable request. The study was supported by Swedish State Support for Clinical Research (LUA-ALF #75130) and the Gothenburg Medical Society. None of the authors report any conflict of interests.

Results 1,159 patients were enrolled into the study. 64 patients were excluded from analysis due to registration errors (n = 30), lack of perioperative documentation (n = 3), and for indication other than acute fracture (n = 31). Thus, 109 patients were included in the uncemented group and 986 patients were included in the cemented group for analysis. Baseline characteristics are given in Table 1. The cemented group was older and had a higher ASA classification and a higher incidence of dementia, compared with the uncemented group. The use of spinal anesthesia was more common in the uncemented group (96% vs. 86%). Preoperative medication with ß-adrenergic blockers, diuretics, and calcium antagonists was more frequent in the cemented group. Preoperative serum creatinine was higher in the cemented compared with the uncemented group. The incidence and grade of hypoxia and/or hypotension, according to Donaldson’s criteria, were higher and more pronounced, respectively, in the cemented group (p = 0.003) (Table 2). Thus, in patients with cemented hemiarthroplasty, 28% had symptoms of hypoxia/hypotension (Grade 1–3), compared with 17% in those with uncemented hemiarthroplasty. The incidence of severe hypotension/hypoxia, i.e. Grades 2 and 3, was 7% in the cemented group, while no patients in the uncemented group experienced severe hypotension and/or hypoxia. Early postoperative mortality (< 48 hours) was 0% and 2% in the uncemented and cemented groups, respectively (p = 0.001). 30-day mortality was 3% in the uncemented and 9% in the cemented group (p = 0.03). One-year mortality was 13% in the uncemented group and 29% in the cemented, respectively (p < 0.001). Univariable hazard ratios (HR) were calculated for a multitude of possible risk factors correlated to mortality (Table 3). We found that several factors indicating cardiovascular and respiratory comorbidity had high HRs (Table 3). Male sex, age, ASA preoperative risk score, assisted living, reduced mobility, liver disease, renal failure, peripheral vascular disease, congestive heart failure, cancer, dementia, and arrhythmia all showed high hazard ratios in the univariable analysis. Regarding pre-existing medication, the use of beta-blockers, diuretics, and anti-platelet drugs were all significantly associ-

295

Table 1. Baseline characteristics. Values are frequency (%) unless otherwise specified Variable

Uncemented Cemented (n = 109) (n = 986)

Sex Female 80 (73) 704 (71) Male 29 (27) 282 (29) Age at surgery (SD) 82 (8) 85 (6) Assisted living 48 (41) 570 (56) Reduced mobility 38 (32) 422 (43) ASA preoperative risk score: 1 6 (5.6) 21 (2.1) 2 60 (56) 384 (39) 3 41 (38) 532 (54) 4 1 (0.9) 49 (5.0) Spinal anesthesia 104 (96) 847 (86) Medical history Liver disease 0 (0.0) 10 (1.0) Renal failure 2 (1.8) 36 (3.7) Diabetes 11 (10) 133 (14) Stroke 16 (15) 183 (19) Peripheral vascular disease 1 (0.9) 25 (2.5) Arteriosclerosis 3 (2.8) 21 (2.1) Hypertension 37 (34) 410 (42) Angina pectoris 16 (15) 139 (14) Previous myocardial infarction 11 (10) 111 (11) Congestive heart failure 7 (6.4) 112 (11) Chronic obstructive pulmonary disease 11 (10) 124 (12) Cancer 11 (10) 66 (6.7) Dementia 19 (17) 259 (26) Arrhythmia 21 (19) 220 (22) Medication β-adrenergic blocker 29 (27) 366 (37) Diuretics 22 (20) 367 (37) Antiplatelet drugs 41 (38) 422 (43) Organic nitrates 14 (13) 139 (14) Calcium antagonists 11 (10) 198 (20) ACE inhibitors 19 (17) 235 (24) Insulin 4 (3.7) 66 (6.7) Warfarin 6 (5.5) 67 (6.8) Statins 9 (8.3) 138 (14) Preoperative hemoglobin (g/L) (SD) 127 (14) 125 (15) Serum creatinine (µmol/L) (SD) 72 (25) 86 (41) Renal failure is defined as serum creatinine > 150 µmol/L. Diabetes includes both types I and II. ACE = angiotensin converting enzyme. Table 2. Incidence of intraoperative hypotension/hypoxia. Values are frequency (%) Grade of hypotension/hypoxia

0 1 2 3

Uncemented Cemented (n = 109) (n = 986) 90 (83) 18 (17) 0 (0.0) 0 (0.0)

711 (72) 204 (21) 51 (5.2) 17 (1.7)

p = 0.003

ated with an increased HR. Conversely, the use of statins was associated with significantly reduced HR (0.6). When investigating laboratory data, lower levels of hemoglobin (< 100

296

Acta Orthopaedica 2020; 91 (3): 293–298

Table 3. Univariable unadjusted hazard ratios (HR) for 1-year mortality Variables

HR (95% CI)

p-value

Adjusted cumulative survival 1.0

0.9

Uncemented Cemented

0.8 Cemented vs. uncemented 2.5 (1.5–4.2) 0.001 Male 1.5 (1.2–1.9) < 0.001 Age at surgery (per 1 year) 1.07 (1.05–1.09) < 0.001 0.7 Assisted living 2.7 (2.2–3.3) < 0.001 Reduced mobility 1.7 (1.4–2.1) < 0.001 0.6 ASA (1–2 vs. 3–4) 2.1 (1.7–2.7) < 0.001 Regional anesthesia (spinal) 1.1 (0.8–1.5) 0.79 Medical history 0.5 Liver disease 2.5 (1.1–5.7) 0.03 0 90 180 270 360 Renal failure 2.4 (1.5–3.8) < 0.001 Days after index operation Diabetes 0.98 (0.69–1.4) 0.9 Stroke 1.3 (0.96–1.7) 0.1 Cumulative 1-year survival, adjusted for covariates, after hemiarthro Peripheral vascular disease 1.8 (1.0–3.2) 0.05 plasty for femoral neck fracture with (red) or without (blue) the use of Arteriosclerosis 1.0 (0.91–3.2) 0.1 bone cement. Hypertension 0.93 (0.74–1.2) 0.5 Angina pectoris 1.3 (0.92–1.7) 0.2 Previous myocardial infarction 1.3 (0.94–1.8) 0.1 and with group, above the confounders already in the model, Congestive heart failure 1.9 (1.4–2.6) < 0.001 Chronic obstructive pulmonary is needed to fully explain away our effect estimate of 1.9. Sev disease 0.95 (0.67–1.4) 0.8 eral models for variable selection were performed including Cancer 1.7 (1.2–2.4) 0.002 clinically important risk factors, showing consistent hazard Dementia 2.3 (1.8–2.8) < 0.001 Arrhythmia 1.6 (1.2–2.0) < 0.001 ratios and E values (Table 4, See Supplementary data). Medication ß-blockers 1.5 (1.2–1.8) 0.001 Diuretics 1.7 (1.5–2.2) < 0.001 Antiplatelet drugs 1.5 (1.2–1.9) 0.001 Discussion Nitrates 1.3 (0.97–1.8) 0.08 Calcium antagonists 0.92 (0.68–1.2) 0.6 In the present investigation on patients undergoing cemented ACE inhibitors 1.0 (0.78–1.3) 0.9 Insulin 1.1 (0.68–1.7) 0.8 or uncemented hemiarthroplasty for femoral neck fractures, Warfarin 0.82 (0.49–1.4) 0.4 we used the classification system proposed by Donaldson et Statins 0.60 (0.40–0.88) 0.01 al, (2009), to evaluate the role of bone cementation for the Preoperative hemoglobin < 100 g/L 1.8 (1.1–2.9) 0.04 development of intraoperative hemodynamic and pulmonary serum creatinine > 150 µmol/L 2.3 (1.5–3.5) < 0.001 derangement and its impact on postoperative mortality. The Grade of hypotension/hypoxia main findings were that the incidence of hypoxia and/or hypo (2–3 vs. 0–1) 4.1 (3.0–5.8) < 0.001

Diabetes includes both types I and II. ASA; ACE = angiotensin converting enzyme. Unadjusted hazard ratios were calculated for each variable.

g/L) and elevated serum creatinine levels (> 150 µmol/L) were associated with increased HR. The presence of intraoperative severe hypotension and/or hypoxia (Grade 2 and 3) and the use of bone cement were both univariably associated with higher mortality (p < 0.001, and p = 0.001, respectively) In the Cox multivariate regression analysis, hypotension and/or hypoxia was not included as it was considered as a mediator variable. After adjustments for confounders (see above) in the multivariate Cox regression analysis, the use of bone cementation was associated with significantly increased HR for 1-year mortality (1.9, CI 1.3–2.7, Figure). In the 1st sensitivity analysis, including the most important confounders, the use of bone cementation was associated with significantly increased HR for 1-year mortality (1.8, CI 1.2–2.6). The calculated E-value was 3.1, i.e., a risk ratio of 3.1 (95% CI 1.9–4.8) for unmeasured confounding, both with outcome

tension was higher in the cemented compared with the uncemented group. In contrast, patients undergoing uncemented hemiarthroplasty, severe intraoperative hypotension, and/or hypoxia (Grades 2 or 3) were not observed. Furthermore, the use of bone cement was an independent risk factor for 1-year postoperative mortality. To our knowledge, this is the first study demonstrating the link between the use of bone cement, per se, with intraoperative hemodynamic and/or pulmonary instability and increased postoperative mortality. It has been suggested that the BCIS is caused by an increase in intra-medullary pressure caused by prosthesis insertion and cementation, which will force medullary content into the circulation causing lung embolization (de Froidmont et al. 2014). This will, in turn, induce hypoxia, pulmonary vasoconstriction, and an increase in right ventricular (RV) afterload and eventually RV failure with systemic hypotension. Pulmonary embolization has been demonstrated in several studies by the use of transesophageal echocardiography in patients undergoing both uncemented and cemented total hip arthroplasty starting during reaming of the femur and acetabulum until the end of the procedure (Lafont et al. 1994).

Acta Orthopaedica 2020; 91 (3): 293–298

Pulmonary embolization has been shown to be particularly marked when placing the femoral component in a cementing procedure (Kotyra et al. 2010). However, it is not immediately evident that pulmonary embolization in cemented hip arthroplasty triggers intraoperative hemodynamic and/or pulmonary instability, as in those small studies episodes of lung embolization caused only minimal changes in pulmonary and systemic hemodynamics with few adverse clinical events (Kotyra et al. 2010). In the present study, however, including more than 1,000 patients undergoing hemiarthroplasty for femoral neck fractures, the use of bone cement independently predicted 30-day mortality. Our data support the results from a recent study by White et al. (2016), showing that a fall in in blood pressure, intraoperatively, is significantly associated with a higher 5-day and 30-day mortality in patients undergoing hip fracture surgery. These findings are in line with recent studies showing that intraoperative hypotension is associated with increased postoperative morbidity and 30-day mortality in patients undergoing non-cardiac surgery (Walsh et al. 2013, Monk et al. 2015). Thus, the combination of a reduced perfusion pressure and arterial oxygen content, intraoperatively, seems to have an important role in the development of postoperative organ injury and mortality. To investigate changes in systemic and pulmonary hemodynamics, Kotyra et al. (2010) inserted pulmonary artery catheters in anesthetized patients operated on with cemented hemi-arthroplasty after an acute hip fracture. A clear temporal association with cementation and a distinct 45% increase in pulmonary vascular resistance (PVR) was found. The increase in PVR was clinically apparent only in 1/15 patients in that study, suggesting that embolization from the medullary canal is present in all cases and that the patients’ capability of handling a sudden increase in PVR determines the degree of hemodynamic instability, expressed clinically, heavily influenced by comorbidity and age. An analysis of the degree of hypotension and/or hypoxia in patients undergoing uncemented hemiarthroplasty for femoral fractures applying the Donaldson criteria has, to our knowledge, not previously been published. Severe hemodynamic and/or pulmonary instability was not seen in any of the 109 patients undergoing uncemented hemiarthroplasty, despite the fact that this group consisted of patients with a high ASA score (ASA 3 ≈ 40%) and a high incidence of chronic obstructive pulmonary disease and congestive heart failure, all being associated with severe intraoperative hypotension and/ or hypoxia in cemented hemiarthroplasty (Olsen et al. 2014). This strongly suggests that the bone cement itself is required to induce severe intraoperative hypotension and/or hypoxia with negative consequences on postoperative survival. We found that early (< 48 hours) postoperative mortality was 2% in the cemented group, a result somewhat higher than in previous studies showing an early mortality of 0.8–1.3% after cemented hemiarthroplasty. In a previous study, we iden-

297

tified risk factors for the development of BCIS and found that BCIS Grade 2 or 3 was associated with a 16-fold increase in 30-day mortality (Olsen et al. 2014). Thus, patients at risk for developing intraoperative hypotension and/or hypoxia are more likely to succumb early, if cement is used. Costain et al. (2011) could demonstrate that mortality from cemented hemiarthroplasty occurs mainly in the perioperative period. Thereafter the survival did not differ from the group not receiving bone cement. However, to perform preventive measures in patients undergoing cemented hemiarthroplasty to lower the risk of developing BCIS and to decrease early mortality, efforts should be made to identify patients at risk for BCIS. Furthermore, one should question the use of cemented hemiarthroplasty in elderly, male patients with a high ASA score and compromised cardiac and renal function, despite the fact that the alternative, cementless hemiarthroplasty, is associated with an increased rate of postoperative surgical complications and a worse biomechanical performance (Khan et al. 2002, Parker et al. 2010, Li et al. 2013, Ekman et al. 2019). The risk of postoperative morbidity and mortality, as well as patient satisfaction and the surgical result, all have to be evaluated when choosing which procedure to pursue. Our results are at variance with prospective randomized trials on patients undergoing hemiarthroplasty for femoral neck fracture showing no difference in 1-year mortality between cemented and uncemented hemiprosthesis (Parker et al. 2010, Li et al. 2013, Talsnes et al. 2013). Although we made efforts to adjust for differences in baseline demographic characteristics, preoperative risk score, type of anesthesia, comorbidity, medication, and renal function, we cannot rule out the possibility that there might be residual unknown confounding variables not recorded by us that to some extent could explain the difference in one-year mortality between the two groups. Our study is limited by its retrospective design. Therefore, we were bound by the quality of data presented in the medical records from both hospitals. In order to avoid inter-observer variability, the extraction of data in both hospitals was performed by 1 of the investigators (FO). The study groups varied in size by a factor of 10. A post-hoc power analysis revealed, however, that the smaller study group having uncemented hemiarthroplasty was adequate to reach our desired endpoints. Furthermore, we have no information on the cause of death. In addition, although we have adjusted for several known confounders, confounding may persist. Several potential unmeasured confounders are not accounted for in our model; these include, but are not limited to, surgeon and anesthesiologist seniority, time of admission, BMI, and smoking amongst others. In addition, surgical delay, an important systemic factor for mortality (Sheehan et al. 2016), was not available to us at the time of data collection and should also be regarded as an unmeasured confounder. On the other hand, we believe that the evidence of association between bone cementation and 1-year mortality seems reasonably robust to residual confounding, as substantial unmeasured confounding would be

298

necessary to negate the observed association. We are unaware of any potential confounder with such a strong risk (HR ≥ 3), above those already in the model, that is not already included in our model. The strength of the study is that we have analyzed individual anesthesia charts of more than 1,000 patients undergoing hemiarthroplasty for femoral neck fractures with or without bone cementation, identifying patients with more or less severe hemodynamic and pulmonary instability at the time around prosthesis insertion. In conclusion, in this retrospective study, on patients undergoing hemiarthroplasty for femoral neck fractures, we found that the use of bone cement was an independent risk factor for 1-year postoperative mortality. Cemented hemiarthroplasty is the preferred technique for displaced neck of femur fracture from an orthopedic surgeon’s perspective. Thus, with our results at hand, it is pivotal for a successful outcome that these patients are discussed preoperatively in an interdisciplinary forum focusing on risk identification and intraoperative surgical strategy, to mitigate the risk for intraoperative hypotension and/or hypoxia and mortality. Supplementary data Table 4 is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2020. 1745510 FO: Planned and designed the study, collected and analyzed data, and wrote up the first draft. MHaS: Planned and designed the study, analyzed data, and wrote up the first draft. BN: Interpreted data and revised the manuscript. EH and S-ER: Planned and designed the study, interpreted data, and revised the manuscript. Acta thanks Jan-Erik Gjertsen and Iain Moppett for help with peer review of this study.

Bisignani G, Bisignani M, Pasquale G S, Greco F. Intraoperative embolism and hip arthroplasty: intraoperative transesophageal echocardiographic study. J Cardiovasc Med 2008; 9(3): 277-81. Byrick R J, Forbes D, Waddell J P. A monitored cardiovascular collapse during cemented total knee replacement. Anesthesiology 1986; 65(2): 213-15. Clark D I, Ahmed A B, Baxendale B R, Moran C G. Cardiac output during hemiarthroplasty of the hip. J Bone Joint Surg Br 2001; 83(3): 5. Costain D J, Whitehouse S L, Pratt N L, Graves S E, Ryan P, Crawford R W. Perioperative mortality after hemiarthroplasty related to fixation method: a study based on the Australian Orthopaedic Association National Joint Replacement Registry. Acta Orthop 2011; 82(3): 275-81. Dahl O E, Molnar I, Vinje A, Rø Johs S, Kierulf P, Andersen Å-B W, Dalaker K, Prydz H. Studies on coagulation, fibrinolysis, kallikrein-kinin and complement activation in systemic and pulmonary circulation during hip arthroplasty with acrylic cement. Thrombosis Res 1988; 50(6): 875-84. Donaldson A J, Thomson H E, Harper N J, Kenny N W. Bone cement implantation syndrome. Br J Anaesth 2009; 102(1): 12-22.

Acta Orthopaedica 2020; 91 (3): 293–298

Ekman E, Laaksonen I, Isotalo K, Liukas A, Vahlberg T, Mäkelä K. Cementing does not increase the immediate postoperative risk of death after total hip arthroplasty or hemiarthroplasty: a hospital-based study of 10,677 patients. Acta Orthop 2019; 90(3): 270-4. de Froidmont S, Bonetti LR, Villaverde RV, del Mar Lesta M, Palmiere C. Postmortem findings in bone cement implantation syndrome–related deaths: Am J of Forensic Med Pathol. 2014; 35(3): 206–11. Juliebo V, Krogseth M, Skovlund E, Engedal K, Wyller T B. Medical treatment predicts mortality after hip fracture. J Gerontol A Biol Sci Med Sci 2010; 65(4): 442-9. Khan R, MacDowell A, Crossman P, Datta A, Jallali N, Arch B, Keene G. Cemented or uncemented hemiarthroplasty for displaced intracapsular femoral neck fractures. Int Orthop 2002; 26(4): 229-32. Kotyra M, Houltz E, Ricksten S-E. Pulmonary haemodynamics and right ventricular function during cemented hemiarthroplasty for femoral neck fracture: pulmonary haemodynamics and right ventricular function. Acta Anaesthesiol Scand 2010; 54(10): 1210-6. Lafont N D, Kostucki W M, Marchand P H, Michaux M N, Boogaerts J G. Embolism detected by transoesophageal echocardiography during hip arthroplasty. Can J Anaesth 1994; 41(9): 850-3. Li T, Zhuang Q, Weng X, Zhou L, Bian Y. Cemented versus uncemented hemiarthroplasty for femoral neck fractures in elderly patients: a metaanalysis. PLoS ONE 2013; 8(7): e68903. Maxwell M J, Moran C G, Moppett I K. Development and validation of a preoperative scoring system to predict 30 day mortality in patients undergoing hip fracture surgery. Br J Anaesth 2008; 101(4): 511-7. Monk T G, Mangione M P, Nguyen J D, Hammermeister K E. Association between intraoperative hypotension and hypertension and 30-day postoperative mortality in noncardiac surgery. Anesthesiology 2015; 123:13. Olsen F, Kotyra M, Houltz E, Ricksten S-E. Bone cement implantation syndrome in cemented hemiarthroplasty for femoral neck fracture: incidence, risk factors, and effect on outcome. Br J Anaesth 2014; 113(5): 800-6. Orsini E C, Byrick R J, Mullen J B, Kay J C, Waddell J P. Cardiopulmonary function and pulmonary microemboli during arthroplasty using cemented or non-cemented components: the role of intramedullary pressure. J Bone Joint Surgery Am 1987; 69(6): 822-32. Parker M J, Gurusamy K S, Azegami S. Arthroplasties (with and without bone cement) for proximal femoral fractures in adults. Cochrane Database Syst Rev 2010; (6):CD001706. doi: 10.1002/14651858.CD001706.pub4. Sheehan K J, Sobolev B, Chudyk A, Stephens T, Guy P. Patient and system factors of mortality after hip fracture: a scoping review. BMC Musculoskelet Disord 2016; 17: 166. Talsnes O, Vinje T, Gjertsen J E, Dahl O E, Engesæter L B, Baste V, Pripp A H, Reikerås O. Perioperative mortality in hip fracture patients treated with cemented and uncemented hemiprosthesis: a register study of 11,210 patients. Int Orthop 2013; 37(6): 1135-40. VanderWeele T J, Ding P. Sensitivity Analysis in observational research: introducing the E-value. Ann Intern Med 2017; 167(4): 268. Walsh M, Devereaux P J, Garg A X, Kurz A, Turan A, Rodseth R N, Cywinski J, Thabane L, Sessler D I. Relationship between intraoperative mean arterial pressure and clinical outcomes after noncardiac surgery toward an empirical definition of hypotension. Anesthesiology 2013; 119(3): 507-15. White S M, Moppett I K, Griffiths R, Johansen A, Wakeman R, Boulton C, Plant F, Williams A, Pappenheim K, Majeed A, Currie C T, Grocott M P W. Secondary analysis of outcomes after 11,085 hip fracture operations from the prospective UK Anaesthesia Sprint Audit of Practice (ASAP-2). Anaesthesia 2016; 71(5): 506-14.

Acta Orthopaedica 2020; 91 (3): 299–305 299

14-year hip survivorship after periacetabular osteotomy: a follow-up study on 1,385 hips Josefine Beck LARSEN, Inger MECHLENBURG, Stig Storgaard JAKOBSEN, Theis Munchholm THILLEMAN, and Kjeld SØBALLE

Department of Orthopaedic Surgery, Aarhus University Hospital, Aarhus N, Denmark Correspondence: josefinebl@clin.au.dk Submitted 2019-07-12. Accepted 2020-01-03.

Background and purpose — Few studies have evaluated the long- and mid-term outcomes after minimally invasive periacetabular osteotomy (PAO). We investigated: (1) the long-term hip survival rate after PAO; (2) the risk of complications and additional surgery after PAO; and (3) the hip function at different follow-up points. Patients and methods — We reviewed 1,385 hips (1,126 patients) who underwent PAO between January 2004 and December 2017. Through inquiry to the Danish National Patient Registry we identified conversions to total hip arthroplasty (THA) and complications after PAO. We evaluated the Hip disability and Osteoarthritis Outcome Score (HOOS) obtained preoperatively, and at 6 months, 2-, 5-, and 10-years’ follow-up. Results — 73 of the 1,385 hips were converted to THA. The overall Kaplan–Meier hip survival rate was 80% (95% CI 68–88) at 14 years with a mean follow-up of 5 years (0.03–14). 1.1% of the hips had a complication requiring surgical intervention. The most common additional surgery was removal of screws (13%) and 11% received a hip arthroscopy. At the 2-year follow-up, HOOS pain improved by a mean of 26 points (CI 24–28) and a HOOS pain score > 50 was observed in 86%. Interpretation — PAO preserved 4 of 5 hips at 14 years, with higher age leading to lower survivorship. The PAO technique was shown to be safe; 1.1% of patients had a complication that demanded surgical intervention. The majority of the patients with preserved hips have no or low pain. The operation is effective with a good clinical outcome.

Periacetabular osteotomy (PAO) is the most common surgical procedure to treat symptomatic hip dysplasia (Ganz et al. 1988, Clohisy et al. 2009). Previous studies have reported a 10-year hip survivorship of 78–95% in patients undergoing PAO. These studies, however, only include a small number of hips and surgical procedures performed during the surgical learning curve (Steppacher et al. 2008, Matheney et al. 2009, Hartig-Andreasen et al. 2012, Albers et al. 2013, Lerch et al. 2017, Ziran et al. 2018). In addition to hip survivorship, several studies have investigated the risk of complications following PAO. It has been estimated that early serious complications occurred in 6–37% of patients (Clohisy et al. 2009). Delayed complication rates suggested that 9% of patients had major complications requiring surgical or arthroscopic intervention, including nonunion, hematoma/deep infection, revision PAO, heterotopic ossification, intraoperative fractures, osteotomy, or sciatic nerve damage (Wells et al. 2018b). To our knowledge, only a few studies have evaluated the long-term complications after PAO (Wells et al. 2018b). Moreover, conversion to total hip arthroplasty (THA) may not be sufficient to describe the outcome after PAO, since patients may not want a THA or surgeons may not recommend it. Patient-reported outcomes (PRO) can therefore supplement the evaluation of the outcome after PAO. Previous studies have used different PROs to identify a failure after PAO, including the Merle d’Aubigné–Postel score < 15 or the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) ≥ 10 (Matheney et al. 2009, Hartig-Andreasen et al. 2012, Albers et al. 2013, Lerch et al. 2017, Wells et al. 2018a). In this study, we used the Hip disability and Osteoarthritis Outcome Score (HOOS). This study determines (1) long-term hip survival rate after PAO; (2) risk of complications after PAO; (3) hip function using HOOS at different follow-up points.

300

Patients and methods Institutional database Demographic, clinical, and PROs were collected in an institutional database that contains data from patients undergoing PAO at Aarhus University Hospital, Denmark and Mølholm Private Hospital, Denmark from 1998. The database was created in 2010 and includes prospectively gathered data from that time. Data on patients operated from 2004 to 2010 were retrospectively entered into the database in 2014. Data on all patients included: sex, age at surgery, and right or left hip on which PAO was performed. Since 2010, preoperative demographic data have included: height, weight, BMI, educational level, pain measured on a visual analogue scale (VAS) during rest and during activity, center–edge (CE) angle, acetabular index (AI) angle, degree of osteoarthritis (OA), impingement test score, and previous treatment in the same hip. PRO was systematically gathered prospectively from 2010 and included HOOS obtained preoperatively and at 6 months, 2, 5, and 10 years postoperatively. For patients operated in 2004 and onwards, HOOS was available at 10 years postoperatively. For patients operated in 2009 and onwards, HOOS was available at 5 years postoperatively. The Danish National Patient Registry The Danish National Patient Registry (DNPR) is a national registry established in 1976 containing information on all contacts, surgical procedures, and admissions for patients treated at Danish hospitals. The DNPR contains information on dates of admission, and discharge diagnoses according to the International Classification of Diseases (ICD 10). Furthermore, the registry holds data on dates and types of surgical procedures, according to the Health Care Classification System, e.g. SKS codes for the hip (KNF) and pelvis (KNE) (Sundhedsdatastyrelsen 2019, see also Schmidt et al. 2014). The patient data from our institutional database were merged with data from the DNPR regarding information on (1) THA, and (2) complications including deep vein thrombosis (DVT) and pulmonary embolism 1 month after PAO and information on which surgical procedures had been performed on the pelvis and hip. Furthermore, information on death and emigration was collected, where the DNPR allows for almost complete followup (Schmidt et al. 2014). Because most operations require a visit to the hospital and hospital encounters are registered consistently in the DNPR, a high level of completeness is expected (Schmidt et al. 2015). Since the DNPR contains all this information it allows for complete follow-up on all patients. Study population Patients were identified from our institutional database on patients undergoing PAO. 1,721 surgically treated hips between January 2004 and December 2017 were eligible for inclusion in this study (Figure 1). The vast majority of

Acta Orthopaedica 2020; 91 (3): 299–305

Eligible hips (patients) n = 1,721 (1,476) Excluded hips (n = 336): – reverse PAO, 194 – femur osteotomy, 4 – Legg-calvé-Perthes disease, 5 – congenital hip dysplasia, 4 – CE angle > 25°, 12 – no Danish civil registration number, 117 Included for survival analysis and complication rates n = 1,385 (1,126) Excluded hips (n = 17): – died, 4 (3) – emigrated, 13 (11) HOOS analysis: – hips operated after 2004, 1,385 – hips operated after 2009, 1,113 – hips operated after 2010, 1,017

Figure 1. Flowchart showing the study population, reasons for exclusion, and number of patients included for survival analysis, complication rates, and Hip disability and Osteoarthritis Outcome Score (HOOS) analysis. PAO = periacetabular osteotomy. CE angle = center–edge.

operations were performed by the senior author (KS); the rest were performed (parts of or entirely) by 4 surgical fellows, thus minimizing the surgical learning curve. Concomitant hip arthroscopy is not performed at our institution. The exclusion criteria were reverse PAO, femur osteotomy, persons without a Danish civil registration number (patients from abroad), Legg–Calvé–Perthes disease, and congenital hip dislocation. 1,385 hips (1,126 patients) met the inclusion criteria. During the study 3 patients died unrelated to the operation (4 hips) and 11 patients emigrated (13 hips). Indications for PAO surgery throughout the study were (1) symptomatic dysplasia of the hip with persistent hip pain and reduced function, (2) CE angle according to Wiberg < 25°, (3) pelvic bone maturity, (4) absence of hip subluxation, (5) internal rotation > 15°, and (6) hip flexion > 110°. Contraindications for PAO were (1) OA (this contraindication has gradually changed to exclude any OA above Tönnis Grade 1 [Tönnis 1987] from PAO surgery), (2) reduced ROM indicating joint degeneration, (3) lack of hip congruence and (4) BMI > 30. From 2016 the inclusion criteria for surgery changed for (1) OA = 0 (Tönnis 1987) (4) BMI ≤ 25 and age ≤ 45 years. The PAOs were performed using the minimally invasive transartorial approach developed by the senior author and described in detail by Troelsen et al. (2008). Figure 2 illustrates a typical postoperative radiograph. During hospitalization, the patients started a physiotherapist-supervised exercise program that continued after discharge supplemented by a home-based exercise program; the patients were allowed 30 kg weight-bearing for 6–8 weeks postoperatively and then full weight-bearing was allowed (Mechlenburg et al. 2018).

Acta Orthopaedica 2020; 91 (3): 299–305 301

Figure 2. Minimally invasive periacetabular osteotomy showing osteotomy of pubis, ischium, and ilium and redirection of acetabulum and 2 screws for fixation.

Outcomes Time to THA surgery was the primary outcome in the hip survival analysis. Follow-up started on the day of PAO and was considered a failure if the patient had undergone THA. Complications after PAO were investigated based on ICD-10 codes and surgical procedures registered in the DNPR. Complications were categorized into serious medical conditions including DVT and pulmonary embolism (DI802, DI803(F), DI809, DI828, DI829, DT817D, DT817C, DI260, DI269) 1 month following PAO and complications requiring surgical intervention and additional surgery in the hip and pelvis (KNE, KNF) to end of study. Furthermore, the categories were categorized into relevant overall descriptions since not all codes are mutually exclusive. The PRO in this study was HOOS, which was obtained preoperatively and at 6 months, 2, 5, and 10 years postoperatively. If HOOS pain score at 2 years was ≤ 50, it was considered a failure, similar to previously used WOMAC pain ≥ 10 (Hartig-Andreasen et al. 2012). Rate of responders with a difference ≥ 9 between preoperative score and follow-up score at 2 years was estimated from a minimally clinically important difference (MCID) of 9 on the HOOS subscale pain (Clohisy et al. 2017). HOOS is a validated measurement for patients suffering from OA (Beyer et al. 2008). HOOS consists of 40 items and assesses 5 separate patient relevant dimensions: pain (10 items), symptoms (5 items), activities of daily living (ADL) (17 items), sport/recreation function (4 items), and hip-related quality of life (QoL) (4 items). Responses to items are given using a 5-point Likert scale (no, mild, moderate, severe, and extreme). The HOOS score on each subscale is a score from 0, indicating extreme problems, to 100 indicating no problems. Missing data are treated as such; provided at least 50% of the items are completed within a subscale, a mean score can be calculated and in the case of 2 answers, the smallest one was selected (indicating worse score).

Statistics Normally distributed data are presented as means (range), non-normally distributed data are presented as medians with interquartile ranges (IQR), and categorical data are presented as numbers with percentages. The cumulative hip survivorship was calculated using Kaplan–Meier survival analysis with THA as the endpoint. Cox’s proportional hazard regression analysis was used to compute hazard ratios and 95% confidence intervals (CI) for sex and age. The hips were analyzed as independent observations (Lie et al. 2004). Censorship was conducted at emigration, death, or end of study, whichever came first. The PAO was categorized as a failure if a patient reported a HOOS pain score ≤ 50. HOOS development over time was tested using univariate ANOVA. HOOS pain at 2 years was compared with preoperative HOOS using a paired t-test. At 2 years, the rate of responders was estimated using a MCID of 9 points. All statistical analyses were performed using STATA/IC, version 15.1 (StataCorp LLC, College Station, TX, USA). For all risk estimates, a 95% CI was estimated. The level of significance was set at p < 0.05. Ethics, funding, and potential conflicts of interest In an accordance with the General Data Protection Regulation in European countries, the Danish Data Protection Agency gave permission to handle the personal data (case no. 1-16-02-626-18). As the study was based on registry data, ethical approval is not needed according to Danish law. This study received no funding and the authors had no conflicts of interest.

Results Demographics Median patient age was 32 years (13–59) and the proportion of men was 15%. Demographic and preoperative data are presented in Table 1. Hip survival In the study period, 73 of the 1,385 hips were converted to THA. The mean time from PAO to THA surgery was 5 years (0.6–14) after PAO. The mean follow-up time was 5 years (0.03–14). The Kaplan–Meier analysis with THA defined as failure showed a cumulative hip survival rate of 80% (CI 68–88) at 14 years for the entire cohort of 1,385 hips (Figure 3). The cumulative hip survival rate at 2-, 5- and 10-years’ follow-up was 99% (CI 98–99), 96% (CI 94–97), and 90% (CI 87–92), respectively. The hazard ratio for conversion to THA for women compared with men was 0.84 (CI 0.44–1.60). The hazard ratio for conversion to THA for the age group < 20 compared with the age groups 20–40 and > 40 was 1.4 (CI 0.6–3.4) and 2.5 (1.03–6.0), respectively (Figure 4).

302

Acta Orthopaedica 2020; 91 (3): 299–305

Table 1. Demographic and preoperative data concerning the 1,385 hips operated with periacetabular osteotomy. Values are n (%) unless otherwise specified

Type of complication

Covariates Value Age at time of surgery median (IQR) 32 (23–40) range 13–59 Sex female 1,175 (85) male 210 (15) Side of operation right 770 (56) left 615 (44) Additional covariates included after 2010 Educational level (n = 617) 1. General certificate of secondary education 101 (16) 2. Upper secondary school leaving 98 (16) 3. Vocational upper secondary education 91 (15) 4. Short-cycle higher education 80 (13) 5. Medium-cycle higher education 91 (15) 6. Bachelor education 92 (15) 7. Long-cycle higher education 60 (9) 8. PhD program 4 (1) Body mass index (n = 617) mean (range) 23 (16–34) Osteoarthritis score a (n = 592) Grade 0 571 (95) Grade 1 or 2 21 (5) VAS pain, median (IQR) (n = 618) pain at rest 35 (19–56) pain during activity 75 (61–86) Impingement test (n = 592) positive 562 (95) negative 30 (5) Previous treatment in the same hip (n = 592) yes 27 (5) no 565 (95) Center–edge angle (°) (n = 592) median (IQR) 19 (15–20) range –10 to 25 Acetabular index angle (°) (n = 592) median (IQR) 15 (12–20) range 0–40 a Score

according to Tönnis 1987 If data were not available for all 1,385 hips, n is stated. IQR = interquartile range VAS = visual analogue scale

Complications and additional surgery after PAO Complications and additional surgeries are listed in Table 2 and were found in 257 hips (243 patients) in the entire cohort. A medical condition (0.4%) consisting of DVT was observed in 6 hips (6 patients), but no pulmonary embolism was observed 1 month after PAO. The most common additional surgery was screw removal (13%) and hip arthroscopy (11%). Hip arthroscopy included resection of CAM and cartilage, partial

Table 2. Complications or additional surgery after periacetabular osteotomy (PAO) surgery on 1,385 hips n (% of entire cohort)

Period min–max

Additional surgery Screw removal 173 (12.5) 0.3–11 years Hip arthroscopy 154 (11.1) 0.4–9 years Total hip arthroplasty 73 (5.3) 0.6–14 years Open cheilectomy 1 (0.07) 6 years Complications Nonunion 6 (0.4) 0.5–7 years Superficial wound infection 3 (0.2) 23–89 days Revision PAO 2 (0.1) < 1 month after PAO 1 (0.07) 3 days > 1 month after PAO 1 (0.07) 0.7 years Bleeding from corona mortis (coiled) 1 (0.07) 0 days Open exploration of soft tissue 1 (0.07) 1 year Unrelated to PAO surgery 11 (0.9) Z-plasty of the iliotibial band 8 (0.6) 0.8–10 years Soft tissue biopsy 1 (0.07) 6 years Femoral fracture 1 (0.07) 6 years Tumor excision 1 (0.07) 7 years

synovectomy, and repair of labrum. The rare complications (1.1%) requiring surgical intervention are presented in Table 2. 4 hips (4 patients) had concomitant screw removal with revision PAO, open cheilectomy, treatment of non-union, and exploration of soft tissue, respectively. 25 hips (25 patients) had concomitant hip arthroscopy with screw removal. 60 hips (55 patients) had more than 1 complication or additional surgery. Of the 73 hips with conversion to THA, 45 hips (42 patients) had a complication or additional surgery prior to conversion to THA. Hip disability and Osteoarthritis Outcome Score The HOOS preoperatively and at follow-up is presented in Table 3. Each subscale shows that the preoperative score increased up to 6 months postoperatively and then plateaued from 6 months to 10 years (Figure 5). At 2 years’ follow-up (n = 624), 86% of the preserved hips scored no pain or low pain. From preoperatively to 2 years’ follow-up, the HOOS

Table 3. Hip disability and Osteoarthritis Outcome Score (HOOS) preoperatively and at follow-up. Values are median (IQR) Follow-up time HOOS Preop. 6 months 2 years 5 years 10 years subscale n = 599 (59%) n = 656 (64%) n = 643 (70%) n = 528 (73%) n = 197 (95%) Pain Symptoms ADL Sport/rec. QoL

53 (40–65) 50 (35–65) 65 (47–78) 44 (25–56) 31 (19–44)

83 (68–93) 75 (60–85) 89 (76–97) 69 (50–88) 56 (38–75)

83 (65–95) 75 (55–90) 91 (74–97) 75 (50–94) 63 (38–81)

83 (63–95) 75 (55–90) 91 (71–97) 75 (50–94) 63 (44–81)

ADL = activities of daily living QoL = hip-related quality of life Completeness of HOOS is presented as n (% of possible answers).

78 (63–93) 70 (50–85) 84 (66–96) 69 (44–88) 63 (44–81)

Acta Orthopaedica 2020; 91 (3): 299–305 303

Hip survival

1.00

0.75

0.50

0.25

0.25 < 20 years 20−40 years > 40 years

0 0

Number at risk: 1,385

1,111

Follow−up (years)

675

180

0 0

100

40 ADL Pain Symptom Sport QoL

Follow−up (years)

Figure 4. Kaplan–Meier survivorship curve, with conversion to total hip arthroplasty as endpoint for 1,385 hips after periacetabular osteotomy divided according to the age groups < 20 years, 20–40 years, > 40 years at surgery (left) and for each sex (right). Each decrease corresponds to a conversion to total hip arthroplasty. Log-rank test between age groups showed a p-value of 0.03 indicating a significant difference in survival and between the 2 sexes a p-value of 0.6 indicating no significant difference.

HOOS, median

0 Pre 0.5

Figure 3. Kaplan–Meier survivorship curve (with CI) with conversion to total hip arthroplasty as endpoint for 1,385 hips after periacetabular osteotomy. Each decrease corresponds to a conversion to total hip arthroplasty. The number of hips at risk remaining for every year of follow-up is given below the x-axis. The hip survival rate is 80% (CI 68–88) at 14 years.

female male

Years after surgery

Figure 5. Graph showing the median scores preoperatively and at follow-up times for each Hip disability and Osteoarthritis Outcome Score subscale. Each Hip disability and Osteoarthritis Outcome Score showed a significant development over time for all subscales from preoperatively to 5 years with a p-value < 0.001.

pain score (n = 427) improved significantly by 26 points (CI 24–28). At 2 years’ follow-up, 77% of the preserved hips reported an improvement of at least 9 points in HOOS pain compared with the preoperative score.

Discussion The aim of this study was to describe the outcome after PAO surgery by estimating hip survival rate, complications, and additional hip surgery rates, as well as reporting hip func-

tion. At 14 years’ follow-up, we found a survival rate of 80% (CI 68–88%). The survival analysis showed a statistically significant difference in survival between age groups; thus, the risk of THA increases with age. In this cohort, complications such as DVT had a rate of 0.4% and the risk of revision PAO was 0.1%. The rates showed that the most common additional surgery was screw removal (13%) and additional hip arthroscopy (11%). HOOS clearly demonstrated that most patients experienced a significant improvement in HOOS pain score from preoperatively to follow-up at 2 years with a mean difference of 26 (CI 24–28). At 2 years’ follow-up, 86% of the preserved hips had no pain or a low pain score, defined by HOOS pain > 50. The Kaplan–Meier analysis showed a survival rate of 80% (CI 68–88) at 14 years. This is slightly better than that reported by Steppacher et al. (2008), who found a 15-year survival rate of 77% and worse than that found by Wells et al. (2018a) of 92% at 15 years. At 10 years’ follow-up, our study found a survival rate of 90% (CI 87–92). This corresponds to results in previous studies ranging from 78% to 95% (Matheney et al. 2009, Hartig-Andreasen et al. 2012, Albers et al. 2013, Grammatopoulos et al. 2016, Ziran et al. 2018, Wells et al. 2018a). The difference in survival rates in our study compared with other studies could be because these studies included a small number of hips (between 68 and 401) (Table 4, see Supplementary data) compared with the 1,385 hips included in this study. The difference could also be explained by some studies being conducted on patients who were operated during the surgical learning curve, leading to an underestimation of the survival rate (Steppacher et al. 2008, Hartig-Andreasen et al. 2012, Albers et al. 2013, Grammatopoulos et al. 2016,

304

Acta Orthopaedica 2020; 91 (3): 299–305

Ziran et al. 2018). The high survival rate found by Wells et al. (2018a), could be explained by the fact that 13% (22 hips) were lost to follow-up, and these hips could potentially have had a THA, leading to an overestimation in the hip survival rate. We have 100% follow-up. Consistent with the literature, we found that increasing age was a predictor for failure of PAO with an HR of 2.5 (CI 1.0–6.0). This supports an upper age limit for patients undergoing PAO (Steppacher et al. 2008, Matheney et al. 2009, Hartig-Andreasen et al. 2012, Albers et al. 2013, Ziran et al. 2018). Moreover, our survival rate (80%) surpasses the 15-year implant survival rate after THA which was reported to be 64% in young patients below 35 treated for symptomatic hip dysplasia (Swarup et al. 2016). Among our patients, 0.4% experienced DVT and no pulmonary embolism was experienced within 1 month after PAO. This corresponds well with results reported by Clohisy et al. (2017) where 0.5% experienced pulmonary embolism and 0.3% experienced DVT. Zaltz et al. (2014) found that 3 patients (1.5%) experienced DVT within the first 10 weeks after PAO surgery. The higher rate could be because indications for PAO other than symptomatic hip dysplasia were included. Clohisy et al. (2009) found that the most common moderate complication was the removal of symptomatic hardware. This is supported by our study, where screw removal was undertaken in 13% and was the most common additional surgery. 0.1% had a revision PAO; Wells et al. (2018b) found that 2% had revision PAO. 0.4% at our institution had a nonunion requiring reoperation. Wells et al. (2018b) found that 3% had a nonunion requiring open reduction. We do not perform concomitant arthroscopy at the time of PAO, and therefore found it relevant to investigate how many patients underwent hip arthroscopy after PAO. 11% of our patients underwent hip arthroscopy. Clohisy et al. (2017) found that 2% underwent hip arthroscopy due to persistent pain; however, 18% of the included patients had concomitant hip arthroscopy with their PAO surgery. Matheney et al. (2009) found that 11% underwent arthroscopy after PAO at a mean of 7 years, which is similar to the numbers found in the present study. We found that 86% of the preserved hips had no pain or a low pain score. This rate corresponds well to previous results defined by a WOMAC score ≤ 10, with 84–88% preserved hips (Matheney et al. 2009, Hartig-Andreasen et al. 2012, Wells et al. 2018a) and a Merle d’Aubigné–Postel score < 15 ranging from 91–94% preserved hips (Albers et al. 2013, Lerch et al. 2017). The median HOOS score on all subscales showed a significant development over time. The mean change from preoperative to 2 years’ follow-up was 26 points (CI 24–28) for the HOOS pain score, which is in line with the 28 points change from the preoperative to the mean 2.6 years’ follow-up for HOOS pain reported by Clohisy et al. (2017).

patients had additional hip surgery. The inclusion of PRO in this study is a strength, allowing for a secondary endpoint and for clinicians to make a more distinct conclusion regarding the success and failure of PAO. Furthermore, our study involved a large number of patients undergoing PAO. Merging our institutional database with the DNPR allowed complete follow-up on all patients and a high level of completeness (Schmidt et al. 2014, 2015). Despite the high level of completeness, there is a risk of information bias. The DNPR uses SKS codes but there are potential differences that arise over time and among hospital departments, since codes are not mutually exclusive (Schmidt et al. 2015). Furthermore, a limitation in this study is that it includes only DVT, pulmonary embolism, additional surgery, and complications requiring surgical intervention in the hip. Another limitation concerns the data from the institutional database. Because the database was created in 2010, and patients operated before 2010 were first imputed in 2014, there might thus be a lack of completeness in the variables prospectively gathered after 2010. This limits the analysis on the HOOS data, where a small number of patients with 5-year follow-up had completed both the pre- and postoperative HOOS and none of the patients with 10-years’ follow-up had completed a preoperative HOOS.

Strengths and limitations The long-term follow-up of 14 years allowed us to assess complications of the PAO surgery and describe how many

Albers C E, Steppacher S D, Ganz R, Tannast M, Siebenrock K A. Impingement adversely affects 10-year survivorship after periacetabular osteotomy for DDH. Clin Orthop Relat Res 2013; 471(5): 1602-14. doi: 10.1007/ s11999-013-2799-8.

Summary To our knowledge, this is the largest prospective follow-up study on outcomes after PAO surgery. In conclusion, PAO preserved 4 of 5 hips at 14-years’ follow-up. Furthermore, the minimally invasive PAO technique is safe (1.1% requiring reoperation) and was also shown to be effective with good clinical results. This study demonstrates that carefully selected patients will demonstrate good survivorship and that higher age leads to lower survivorship. Furthermore, the most common additional surgery was screw removal and hip arthroscopy. The majority of patients with preserved hips had no or low pain. Supplementary data Table 4 is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2020. 1731159

JBL took part in planning of the study and writing the present manuscript under guidance from IM, TMT, SSJ, KS. Acta thanks Jan Erik Madsen and Rafael Sierra for help with peer review of this study.

Acta Orthopaedica 2020; 91 (3): 299–305 305

Beaulé P E, Dowding C, Parker G, Ryu J J. What factors predict improvements in outcomes scores and reoperations after the Bernese periacetabular osteotomy? Clin Orthop Relat Res 2015; 473(2): 615-22. doi: 10.1007/ s11999-014-3980-4. Beyer N, Thorborg K, Vinther A. Translation and cross-cultural adaptation of the Danish version of the Hip Dysfunction and Osteoarthritis Outcome Score 2.0 (HOOS 2.0). Available at: http://www.koos.nu/index.html 2008. Accessed February 15, 2019. Clohisy J C, Schutz A L, St John L, Schoenecker P L, Wright R W. Periacetabular osteotomy: a systematic literature review. Clin Orthop Relat Res 2009; 467(8): 2041-52. doi: 10.1007/s11999-009-0842-6. Clohisy J C, Ackerman J, Baca G, Baty J, Beaule P E, Kim Y J, Millis M B, Podeszwa D A, Schoenecker P L, Sierra R J, Sink E L, Sucato D J, Trousdale R T, Zaltz I. Patient-reported outcomes of periacetabular osteotomy from the Prospective ANCHOR Cohort Study. J Bone Joint Surg Am 2017; 99(1): 33-41. doi: 10.2106/JBJS.15.00798. Dahl L B, Dengsø K, Bang-Christiansen K, Petersen M M, Stürup J. Clinical and radiological outcome after periacetabular osteotomy: a cross-sectional study of 127 hips operated on from 1999–2008. Hip Int 2014; 24(4): 36980. doi: 10.5301/hipint.5000129. Ganz R, Klaue K, Vinh T S, Mast J W. A new periacetabular osteotomy for the treatment of hip dysplasias. Technique and preliminary results. Clin Orthop Relat Res 1988; (232): 26-36. doi: 10.1097/00003086-198807000-00006. Grammatopoulos G, Wales J, Kothari A, Gill H S, Wainwright A, Theologis T. What is the early/mid-term survivorship and functional outcome after Bernese periacetabular osteotomy in a pediatric surgeon practice? Clin Orthop Relat Res 2016; 474(5): 1216-23. doi: 10.1007/s11999-015-4386-7. Hartig-Andreasen C, Troelsen A, Thillemann TM, Søballe K. What factors predict failure 4 to 12 years after periacetabular osteotomy? Clin Orthop Relat Res 2012; 470(11): 2978-87. doi: 10.1007/s11999-012-2386-4. Isaksen K F, Roscher E K, Iversen K S, Eitzen I, Clarke-Jenssen J, Nordsletten L, Madsen J E. Preoperative incipient osteoarthritis predicts failure after periacetabular osteotomy: 69 hips operated through the anterior intrapelvic approach. Hip Int 2018: 1120700018804786. doi: 10.1177/1120700018804786. Lerch T D, Steppacher S D, Liechti E F, Tannast M, Siebenrock K A. One-third of hips after periacetabular osteotomy survive 30 years with good clinical results, no progression of arthritis, or conversion to THA. Clin Orthop Relat Res 2017; 475(4): 1154-68. doi: 10.1007/s11999-016-5169-5. Lie S A, Engesaeter L B, Havelin L I, Gjessing H K, Vollset S E. Dependency issues in survival analyses of 55,782 primary hip replacements from 47,355 patients. Stat Med 2004; 23(20): 3227-40. doi: 10.1002/sim.1905. Matheney T, Kim Y J, Zurakowski D, Matero C, Millis M. Intermediate to long-term results following the Bernese periacetabular osteotomy and predictors of clinical outcome. J Bone Joint Surg Am 2009; 91(9): 2113-23. doi: 10.2106/jbjs.G.00143.

Mechlenburg I, Jorgensen P B, Stentz-Olesen K, Tjur M, Grimm B, Soballe K. Leg power, pelvic movement and physical activity after periacetabular osteotomy. A prospective cohort study. Acta Orthop Belg 2018; 84(2): 16371. Schmidt M, Pedersen L, Sorensen H T. The Danish Civil Registration System as a tool in epidemiology. Eur J Epidemiol 2014; 29(8): 541-9. doi: 10.1007/s10654-014-9930-3. Schmidt M, Schmidt S A J, Sandegaard J L, Ehrenstein V, Pedersen L, Sørensen H T. The Danish National Patient Registry: a review of content, data quality, and research potential. J Clin Epidemiol 2015; 7:449-90. doi: 10.2147/CLEP.S91125. Steppacher S D, Tannast M, Ganz R, Siebenrock K A. Mean 20-year followup of Bernese periacetabular osteotomy. Clin Orthop Relat Res 2008; 466(7): 1633-44. doi: 10.1007/s11999-008-0242-3. Sundhedsdatastyrelsen. National Health IT SKS-Browser. Available at: http:// www.medinfo.dk/sks/brows.php. Accessed May 17, 2019. Swarup I, Marshall A C, Lee Y Y, Figgie M P. Implant survival and patientreported outcomes after total hip arthroplasty in young patients with developmental dysplasia of the hip. Hip Int 2016; 26(4): 367-73. doi: 10.5301/ hipint.5000354. Tönnis D. Congenital dysplasia and dislocation of the hip in children and adults. Berlin: Springer Verlag; 1987. Troelsen A, Elmengaard B, Søballe K. A new minimally invasive transsartorial approach for periacetabular osteotomy. J Bone Joint Surg Am 2008; 90(3): 493-8. doi: 10.2106/JBJS.F.01399. Troelsen A, Elmengaard B, Søballe K. Medium-term outcome of periacetabular osteotomy and predictors of conversion to total hip replacement. J Bone Joint Surg Am 2009; 91(9): 2169-79. doi: 10.2106/JBJS.H.00994. Wells J, Schoenecker P, Duncan S, Goss C W, Thomason K, Clohisy J C. Intermediate-term hip survivorship and patient-reported outcomes of periacetabular osteotomy: the Washington University experience. J Bone Joint Surg Am 2018a; 100(3): 218-25. doi: 10.2106/jbjs.17.00337. Wells J, Schoenecker P, Petrie J, Thomason K, Goss C W, Clohisy J C. Are complications after the Bernese periacetabular osteotomy associated with subsequent outcomes scores? Clin Orthop Relat Res 2018b. doi: 10.1097/ corr.0000000000000566. Zaltz I, Baca G, Kim YJ, Schoenecker P, Trousdale R, Sierra R, Sucato D, Sink E, Beaule P, Millis MB, Podeszwa D, Clohisy JC. Complications associated with the periacetabular osteotomy: a prospective multicenter study. J Bone Joint Surg Am. 2014; 96(23): 1967-74. doi: 10.2106/jbjs.N.00113. Ziran N, Varcadipane J, Kadri O, Ussef N, Kanim L, Foster A, Matta J. Tenand 20-year survivorship of the hip after periacetabular osteotomy for acetabular dysplasia. J Am Acad Orthop Surg 2018. doi: 10.5435/jaaosd-17-00810.

306

Acta Orthopaedica 2020; 91 (3): 306–312

Influence of fast-track programs on patient-reported outcomes in total hip and knee replacement (THR/TKR) at Swedish hospitals 2011–2015: an observational study including 51,169 THR and 8,393 TKR operations Urban BERG 1,2, Annette W-DAHL 3,4, Ola ROLFSON 1,5,6, Emma NAUCLÉR 6, Martin SUNDBERG 3,4, and Anna NILSDOTTER 1,5 1 Department of Orthopaedics, Clinical Sciences, Sahlgrenska Academy, Gothenburg University; 2 Department of Surgery and Orthopaedics, Kungälv Hospital; 3 Department of Orthopedics, Clinical Sciences Lund, Lund University; 4 The Swedish Knee Arthroplasty Register; 5 Department of Orthopedics Sahlgrenska University Hospital; 6 The Swedish Hip Arthroplasty Register, Sweden

Correspondence: urban.berg@vgregion.se Submitted 2019-11-11. Accepted 2020-01-24.

Background and purpose — Fast-track care programs have been broadly introduced at Swedish hospitals in elective total hip and knee replacement (THR/TKR). We studied the influence of fast-track programs on patient-reported outcomes (PROs) 1 year after surgery, by exploring outcome measures registered in the Swedish arthroplasty registers. Patients and methods — Data were obtained from the Swedish Knee and Hip Arthroplasty Registers and included TKR and THR operations 2011–2015 on patients with osteoarthritis. Based on questionnaires concerning the clinical pathway and care programs at Swedish hospitals, the patients were divided in 2 groups depending on whether they had been operated in a fast-track program or not. PROs of the fast-track group were compared with not fast-track using regression analysis. EQ-5D, EQ VAS, Pain VAS, and Satisfaction VAS were analyzed for both THR and TKR operations. The PROMs for TKR also included KOOS. Results — The differences of EQ-5D, EQ VAS, Pain VAS, and Satisfaction VAS 1 year after surgery were small but all in favor of fast-track for both THR and TKR, also in subscales of KOOS for TKR except KOOS QoL. However, the effect sizes as measured by Cohens’ d formula were < 0.2 for all PROs, in both THR and TKR. Interpretation — Our results indicate that the fast-track programs may be at least as good as conventional care from the perspective of PROs 1-year postoperatively.

Fast-track care programs in elective total hip and knee replacement (THR and TKR) were introduced in Europe at the beginning of the 2000s (Husted et al. 2006, Pilot et al. 2006). Using evidence-based methods in preparation and perioperative care aims to reduce surgical and psychological stress and accelerate recovery after surgery (Kehlet et al. 2008). The care concept has been spread worldwide (Antrobus and Bryson 2011, Christelis et al. 2015, Stowers et al. 2016), resulting in short perioperative hospital stay, and is considered to be safe and well tolerated by patients (Machin et al. 2013, Zhu et al. 2017, Deng et al. 2019, Wainwright and Kehlet 2019). During the last few years an increasing number of ambulatory arthroplasties have been performed as outpatients with maintained safety and short-term outcome (Goyal et al. 2017), Gromov et al. 2019, Coenders et al. 2020). The patients’ experiences and degree of satisfaction have been explored in qualitative studies (Specht et al. 2016, Strickland et al. 2018) and by selfmade questionnaires concerning satisfaction rating of the care (Husted and Holm 2006, Specht et al. 2015, Winther et al. 2015). Patient reported outcomes (PROs) after THR and TKR with fast-track programs have been reported using both generic and disease-specific questionnaires (Larsen et al. 2010, 2012, Winther et al. 2015). The follow-up periods have been of different length and only a few of them had a control group with standard care (Larsen et al. 2008, Machin et al. 2013). The PROs with fast-track programs have been compared with PROs from an age- and sex-matched population (Larsen et

Acta Orthopaedica 2020; 91 (3): 306–312 307

al. 2010, 2012), and the THR Primary THR for OA in Sweden 2011–2015 Primary TKR for OA in Sweden 2011–2015 patients but not TKR patients n = 68,258 n = 59,268 reached the population level Incomplete PROM data TKR at 62 hospitals not participating after 12 months. In a study n = 17,089 in the PROM program of the SKAR from Norway the PROs after n = 47,374 12 months were lower than the Complete PROM data (n = 51,169): – not fast-track program, 19,237 TKR at 15 hospitals participating matched population level but – fast-track program, 27,615 in PROM program of the SKAR similar to register-based aver– unknown care program, 4,317 n = 11,894 age gain in general health (EQIncomplete PROM data 5D) in THR patients (Winther n = 3,501 et al. 2015). Brock et al. (2017) studied the length of stay and Figure 1. Flow chart of the study. THR = total Complete PROM data (n = 8,393): – not fast-track program, 3,450 its impact on WOMAC and hip replacement, TKR = total knee replace– fast-track program, 4,943 ment, and OA = osteoarthritis. SF-36 1 year after surgery. They found a slight improvement of SF-36 associated with shorter LOS but no significant in and out of bed independently, dress, go to the toilet, walk influence on WOMAC. The question remains whether PROs with crutches, and have sufficient pain treatment. The hos1 year after THR and TKR are better with fast-track or not pitals in the fast-track cohort reported a median LOS of 2–4 compared with conventional care programs. days and the hospitals in the not fast-track group 4–7 days. In Sweden, fast-track programs have been broadly impleWe got information from 64 of 83 Swedish hospitals permented for hip and knee replacements during 2011–2015. We forming hip replacements during the period 2011–2015. studied the influence of the fast-track programs on PROs in Operations at hospitals not responding to the questionnaire elective THR and TKR 1 year after surgery by exploring the were used as a 3rd cohort to get an overview of THR on a PROs registered in the Swedish hip and knee arthroplasty reg- national level. The cohort with unknown care program for THR consisted of operations at 19 different hospitals, mostly isters (SHAR and SKAR). low-volume hospitals. Thus, there were 3 cohorts of THR, 1 with fast-track, 1 defined as not fast-track, and 1 cohort with unknown care program. Patients and methods For knee replacements the care programs could be defined Source of data (Figure 1) as fast-track or not fast-track at all 15 hospitals participating in Since 2008 all Swedish hospitals performing elective THR the PROM program based on the questionnaires to the hospihave participated in the PROM program of the SHAR with tals. Consequently, the TKR operations could be divided into data preoperatively and postoperatively after 1 year. Data 2 cohorts, fast-track and not fast-track. completeness with both pre- and postoperative PROMs is around 75%. The SKAR has a PROM project, which started Data collection in the Region Skåne, in the south of Sweden, as a pilot proj- Data were obtained from the SKAR and SHAR and included ect in 2008. From 2013 an increasing number of orthopedic TKR operations (NGB29, NGB39, and NGB49) and THR clinics outside the pilot region have joined the project, and in operations (NFB29, NFB39, NFB49, and NFB62) on patients 2015 there were 15 Swedish hospitals performing TKR par- with osteoarthritis in the knee (M17.0–M17.5) and the hip ticipating. PROM data are collected preoperatively and 1 year (M16.0–M16.9) during the period 2011–2015. Every operation was counted even if patients were operated bilaterally. postoperatively with a completeness of more than 70%. For THRs, PROM data were collected from SHAR using the Definition of cohorts generic health status measure EQ-5D (EuroQolGroup 1990) The study was based on questionnaires concerning the clini- with 3 levels of the dimensions mobility, self-care, usual activcal pathway and care program in elective THR and TKR at ities, pain/discomfort, and anxiety/depression. In addition, the Swedish hospitals (Kärrholm et al. 2016). The survey aimed visual analogue scale (VAS) with a range from 0 to 100 was to define when a fast-track program had been introduced. The used for general health, pain, and satisfaction with surgery. operations at hospitals responding to the questionnaire were For general health (EQ VAS) the score 0 represents the worst divided into 2 groups depending on whether the operations and 100 the best. For Pain VAS and Satisfaction VAS the were made in a fast-track program or not. The definition of best score is 0 and 100 the worst outcome. Delta values were fast-track was based on the following criteria: (1) admission used to measure improvement by comparing the preoperative on the day of surgery; (2) mobilization within 3–6 hours after values with the values 1 year postoperatively. The satisfacoperation; and (3) functional discharge criteria in practice tion (VAS) score was categorized into 5 groups: very satis(Berg et al. 2018). The functional criteria were: ability to get fied (0–20), satisfied (21–40), neither dissatisfied nor satisfied

308

Acta Orthopaedica 2020; 91 (3): 306–312

Table 1. Demographics and surgical data on THR operations 2011– 2015 in patients with osteoarthritis and complete PROM data. Values are n (%) unless otherwise specified Not fast-track Variable n = 19,237

Care program Fast-track n = 27,615

Unknown n = 4,317

Age, mean (SD) Female sex BMI, mean (SD) Year of operation 2011 2012 2013 2014 2015 Charnley class A B C Fixation Cemented Hybrid Uncemented Reverse hybrid Resurfacing Incision Direct lateral Posterior Other

69 (10) 11,071 (58) 28 (5)

68 (10) 15,537 (56) 27 (4)

66 (10) 2,269 (53) 27 (4)

5,998 (31) 4,845 (25) 3,820 (20) 2,879 (15) 1,695 (9)

3,373 (12) 4,571 (17) 5,793 (21) 6,642 (24) 7,236 (26)

1,170 (27) 1,085 (25) 796 (18) 683 (16) 583 (14)

9,259 (48) 2,330 (12) 7,648 (40)

13,885 (50) 3,362 (12) 10,368 (38)

2,112 (49) 514 (12) 1,691 (39)

13,469 (70) 313 (2) 3,278 (17) 2,130 (11) 47 (0)

17,447 (63) 998 (4) 5,220 (19) 3,840 (14) 95 (0)

1,811 (42) 73 (2) 1,551 (36) 768 (18) 114 (3)

6,897 (36) 12,015 (63) 325 (2)

14,643 (53) 12,965 (47) 4 (0)

1,940 (45) 2,334 (54) 43 (1)

(41–60), dissatisfied (61–80), and very dissatisfied (81–100). For TKRs the same outcome measures were explored, but according to the PROM project in the SKAR it also included the Knee injury and Osteoarthritis Outcome Score (KOOS) (Roos et al. 1998) with the 5 subscales Pain, Other Symptoms, Activity in Daily Life function (ADL), Sport and recreation function (Sport/Rec), and Knee related Quality of Life (QoL). All subscales have a range from 0 to 100 where the highest scores represent the best outcomes. Statistics and data analysis The 3 cohorts (fast-track, not fast-track, and unknown care program) of THR and the 2 cohorts of TKR operations, respectively, were presented with descriptive statistics on demographic and surgical data. The EQ-5D index (Burstrom et al. 2014, Nemes et al. 2015), EQ VAS, Pain VAS, and Satisfaction VAS postoperative scores 1 year after surgery were

compared between the not fast-track and the fast-track groups using regression analysis in 3 steps. First a univariable analysis without adjustments was made. In order to reduce bias in the effects of fast track on PRO outcomes after 1 year, 2 multivariable analyses were undertaken: the 1st of the 2 multivariable analyses included adjustment for patient factors such as age, sex, BMI, Charnley category, and the preoperative scores. These factors may influence how patients report their health status 1 year after surgery (Gordon et al. 2013). Finally, the adjustment also included type of fixation and incision in THR. For TKR type of anesthesia, use of tourniquet, and operation time were included. The significance of each covariate was tested (Wald’s test) before being included in the models. The regression coefficients were presented with 95% confidence interval (CI). The effect sizes (standardized mean differences) for the difference between fast-track and not fast-track as measured by the change from pre- to 1-year post-operation in PROs were calculated using Cohen’s d formula (Cohen 1992). Statistical analyses were performed using R version 3.6.1 (R Foundation for Statistical Computing, Vienna, Austria). Ethics, funding, and potential conflicts of interest Ethical approval was given by the Regional Ethical Review Board in Gothenburg (Dnr: Exp. 2019-01-10, 2019-00559/109518). No funding and no competing interests were declared.

Results PROs in THR The demographic variables were similar in the fast-track and not fast-track group. However, the proportion of cemented fixation and posterior approach was higher in the not fast-track group (Table 1). In the cohort with unknown care program representing less than 10% of the THRs in Sweden, there were more males, the mean age was lower, and the proportion of uncemented fixations was higher. The pre- and postoperative PROs for THRs were similar in the 3 cohorts, with only small differences between the groups (Table 2). Complete data with numbers of operations within each severity level of the EQ-5D-3L are presented in Table 3, see Supplementary data. Compared with the not fast-track group a slightly higher proportion of very satisfied patients (VAS 0–20) was seen

Table 2. Mean values (SD) and change (Delta (SD)) in PROs in THRs with complete data preoperatively and 1 year postoperatively Preoperatively 1 year postoperatively Delta PRO Not fast-track Fast-track Unknown Not fast-track Fast-track Unknown Not fast-track Fast-track Unknown EQ-5D index 0.74 (0.11) 0.73 (0.11) 0.74 (0.11) EQ VAS 57 (22) 58 (22) 57 (22) Pain VAS 63 (15) 63 (15) 66 (16) Satisfaction VAS

0.88 (0.11) 76 (20) 14 (18) 16 (20)

0.88 (0.11) 0.89 (0.11) 78 (20) 78 (20) 13 (17) 12 (18) 14 (20) 14 (20)

0.14 (0.13) 0.15 (0.13) 0.15 (0.13) 20 (26) 19 (26) 21 (25) –49 (22) –50 (22) –54 (22)

Acta Orthopaedica 2020; 91 (3): 306–312 309

Table 4. Categories of satisfaction with the THR operation. Values are n (%) Satisfaction (VAS) Not fast-track Very satisfied (0–20) 13,835 (72) Satisfied (21–40) 3,284 (17) Neither dissatisfied nor satisfied (41–60) 1,237 (6.4) Dissatisfied (61–80) 529 (2.7) Very dissatisfied (81–100) 352 (1.8)

Fast-track

Unknown

20,581 (75) 4,207 (15)

3,303 (77) 595 (14)

1,643 (5.9) 775 (2.8) 409 (1.5)

244 (5.7) 97 (2.2) 78 (1.8)

Table 5. Demographics and surgical data on TKR operations 2011– 2015 in patients with osteoarthritis and complete PROM data. Values are n (%) unless otherwise specified Variable

Care program Not fast-track Fast-track n = 3,450 n = 4,943

adjustments for all variables the deviations were equal or larger. However, the differences were small, < 2 on the scale from 0 to 100 for PROMs assessed by the VAS scale (Figure 2). The effect sizes were < 0.2 for all PROs, indicating small effects of the care program.

Age, mean (SD) 69 (9) 69 (9) Female 2,047 (59) 2,753 (56) BMI, mean (SD) 29 (5) 29 (4) Year of operation 2011 928 (27) 79 (2) 2012 491 (14) 590 (12) 2013 567 (16) 1,099 (22) 2014 668 (19) 1,503 (30) 2015 796 (23) 1,672 (34) Charnley class A 854 (25) 1,008 (20) B 1,178 (34) 1,857 (38) C 1,418 (41) 2,078 (42) Type of anesthesia General 722 (21) 1,610 (33) Spinal 2,639 (77) 3,272 (66) Other 83 (2) 51 (1) Tourniquet 2,302 (67) 2,408 (49) LIA 3,338 (97) 4,856 (98) Operation time minutes, mean (SD) 75 (22) 60 (24) LIA = Local infiltration analgesia with/without catheter

in the group with fast-track program, but the proportion of patients considered as very satisfied and satisfied (VAS 0–40) was similar in the 3 cohorts, at 89–90% (Table 4). In the regression analyses without adjustments the deviations from the reference were in favor of fast-track; with

Figure 2. Multivariable regression analysis of PROs 1 year after THR/ TKR with 95% CI. Regression coefficients for fast-track (not fast-track reference). Adjustments for age, sex, BMI, Charnley class, preop scores, and year of operation. For THR also adjustment for implant fixation method and surgical approach.

PROs in TKR In the fast-track group there was a higher proportion of operations with general anesthesia, tourniquet usage was less common, and the mean operation time was shorter. The demographics of the 2 cohorts were similar, with a slightly higher proportion of males in the fast-track group (Table 5). The improvement preoperatively to 1 year postoperatively was considerable regarding pain and health-related quality of life (HRQoL) in both cohorts, but the differences in postoperative mean scores between the cohorts were small, though slightly better in the fast-track group (Table 6). Each of the 5 questions of the EQ-5D is presented in Table 7, see Supplementary data. The proportions of very satisfied and satisfied (VAS 0–40) were 86% in the fast-track group and 83% in the not fast-track group (Table 8). The difference was larger in the category of very satisfied patients (VAS 0–20), with 72% in the fast-track group compared with 62% in the group with a care program defined as not fast-track.

Table 6. Mean values (SD) and change (Delta (SD)) in PROs in TKRs with complete data preop and 1 year postoperatively PRO EQ5D index EQ VAS Pain VAS Satisfaction VAS

Preoperatively Not fast-track Fast-track 0.78 (0.11) 67 (22) 63 (18) 21 (23)

0.77 (0.11) 66 (23) 64 (17) 17 (23)

1 year postoperatively Not fast-track Fast-track 0.87 (0.11) 0.88 (0.11) 76 (20) 77 (20) 20 (20) 17 (20)

Delta Not fast-track Fast-track 0.10 (0.12) 9 (24) –43 (24)

0.11 (0.13) 11 (25) –47 (25)

310

Table 8. Categories of satisfaction with the TKR operation. Values are n (%) Satisfaction (VAS) Not fast-track Fast-track Very satisfied (0–20) 2,131 (62) 3,570 (72) Satisfied (21–40) 720 (21) 670 (14)) Neither dissatisfied nor satisfied (41–60) 345 (10) 386 (7.8) Dissatisfied (61–80) 157 (4.6) 197 (4.0) Very dissatisfied (81–100) 97 (2.8) 120 (2.4)

Acta Orthopaedica 2020; 91 (3): 306–312

Score

Change in score

100

20 10

Postoperative not fast-track Postoperative fast-track Preoperative not fast-track Preoperative fast-track

Symptoms Pain

ADL

Sport/Rec

QoL

Not fast-track Fast-track

Symptoms Pain

ADL

Sport/Rec

QoL

Figure 3. Mean scores (left panel) and improvement (right panel) of the KOOS subscales (0–100) preoperatively and 1-year postoperatively.

Preoperative and postoperative mean scores in the subscales of KOOS for both cohorts are illustrated in Figure 3. The improvement, preoperatively to 1 year postoperatively (Delta values) was considerable in all subscales in both cohorts (Figure 3), and the adjusted regression estimate of the effect of care process was in favor of fast-track in all subscales except for the subscale KOOS QoL. The differences between the groups were small (1–2 points) both pre- and 1 year postoperatively. A table with mean scores of the KOOS subscales is available in the Supplementary data (Table 9). The multivariable regression analysis points to a favorable influence of fast-track in all subscales and outcome measures compared with not fast-track in TKR, but the differences were small (Figure 2). The effect sizes were < 0.2 as for THR. In the Supplementary data a table with adjustments for all procedure-specific variables is available (Table 10).

Discussion We found that the fast-track cohort had slightly better PROs in both THR and TKR patients concerning pain, satisfaction, and all dimensions of health outcome. The KOOS scores were statistically significantly better in the fast-track group of TKR patients in all subscales except for the subscale QoL (Figure 2). However, the differences were small, and the clinical relevance may be questioned. In the regression analysis with adjustments the year of operation had a limited impact on the outcome, and the results in THR and TKR were more influenced by Charnley class C and preoperative score than the care program (Tables 10 and 11, see Supplementary data). In all cohorts of THR and TKR there were patients operated bilaterally during the period 2011–2015; most of them had a 2-stage operation. Consequently, the outcomes of two operations were recorded in the same patient, but as there are large similarities between unilateral and second 2-stage bilateral

THR (Bülow, Nemes and Rolfson, personal communication 2019) we consider that bilaterality does not have a practical influence on the analysis of PROM data in our study. Previous studies aiming to evaluate PROs after THR or TKR in fast-track programs by using generic and disease-specific questionnaires had shorter follow-up periods (Larsen et al. 2008, Machin et al. 2013) or did not have a control group with standard care (Larsen et al. 2010, 2012, Winther et al. 2015). In our study the Swedish value set was used to calculate the EQ-5D index (Burstrom et al. 2014), which is not the same as the value sets used in other countries. Consequently, it is difficult to compare our results with previous publications from hospitals outside Sweden. In our study disease-specific questionnaires were not used for hips, but Larsen et al. (2010) found that generic and disease-specific outcomes such as EQ-5D and Harris Hip Score were strongly associated in THR. In THR it has been reported that uncemented fixation is associated with better PROs 1 year after THR (Rolfson et al. 2016) and in the fast-track cohort of our study there were in fact more uncemented and hybrid fixations. Posterior approach is also associated with better PROs compared with direct lateral approach (Lindgren et al. 2014), but the fast-track group had contrarily a higher proportion of direct lateral approach. This may explain why the difference in favor of fast-track was slightly larger after adjustment for surgical approach. In the fast-track cohort of TKR the use of a tourniquet was less common, but the influence of a tourniquet on functional outcome is inconclusive according to previous studies (Ledin et al. 2012, Zhang et al. 2014, Harsten et al. 2015, Zhou et al. 2017). Ledin et al. (2012) showed in a small randomized study an improved range of motion (ROM) persisting after 2 years and Harsten et al. (2015) showed no effect regarding postoperative pain and muscle strength when a tourniquet was not used while Zhang et al. (2014) and Zhou et al. (2017) found a better ROM in early stage after surgery when not using a tourniquet. We have no evidence from the literature or from our findings that variables related to the surgical intervention in TKR may

Acta Orthopaedica 2020; 91 (3): 306–312 311

have an influence on the PROMs 1 year after surgery. Thus, in our regression analysis we have preferred to present the results with adjustment for the demographic variables, preop scores, and year of operation but not for the surgical variables. The interpretation of results should always be made with caution. In a multinational evaluation of minimum clinical important improvement (MCII) of generic outcomes with scores from 0 to 100 used in rheumatic diseases including hip and knee osteoarthritis, the conclusion was that an absolute improvement of 15 out of 100 or 20% relative improvement was considered as MCII (Tubach et al. 2012). The MCII differs between diagnoses and whether the scores are in the upper or lower range. It has also been suggested that a change of 8 points on the KOOS subscale score ranging from 0 to 100 could be the minimal perceptible clinical improvement (MPCI) (Roos and Lohmander 2003). However, if the differences are too small to be considered as clinically relevant in an individual, it does not mean that the improvement on group level is without importance. Hip and knee replacements are powerful interventions with a high effect size, especially regarding pain and physical function (Jones and Pohar 2012) and we cannot expect more than a small additional improvement of outcome when the care programs are optimized. The responsiveness and ceiling effects of the PROs are also factors to consider in detecting small changes (Greene et al. 2015). Further, outcome may also differ depending on the PROM used (W-Dahl et al. 2014). It is difficult to define specific factors or procedures in the fast-track clinical pathway with the strongest positive impact on PROs 1 year postoperatively. Preoperative information from a multi-professional team and clearly communicated functional discharge criteria may reduce the anxiety and mental stress. The logistic frame aiming for a short LOS stimulates all professionals to contribute in preoperative preparation as well as coaching during the hospital stay and postoperatively to achieve the goal of rapid recovery.

1 year postoperatively is multifactorial and variables such as socioeconomic factors and mental health have not been explored. Fast-track lacks an internationally accepted clear definition. We have used a definition mainly based on logistical criteria in the care of patients undergoing THR and TKR (Berg et al. 2018), but in the fast-track philosophy there are also care principles that are not applied in exactly the same way in all hospitals, for example the use of anti-thrombotic medication. The absence of exact data on LOS is a weakness, but the reported median LOS from the hospitals indicates that fast-track hospitals had a shorter LOS (2–4 days) than the not-fast-track hospitals (4–7 days).

Strengths and limitations Our study explores the influence of fast-track programs in elective THR and TKR in Sweden during a period of nationwide implementation with register data for 5 years. All Swedish hospitals performing hip replacements are included in the PROM program of the SHAR, and this study gives an overview of the influence of fast-track programs on PROs in elective THR. In TKR 15 hospitals of 77 performing knee replacements participated in the PROM project during this period; however, according to register data the demographics and surgical data in the studied group were similar to TKRs in hospitals not participating. The data completeness of the PROM programs was more than 70% for both THR and TKA. A limitation of the study is that bias can persist, due to imbalance in potential confounding factors outside the care program, which may affect the results. The influence on PROs

Acta thanks Henrik Kehlet and Siri Bjørgen Winther for help with peer review of this study.

Conclusion Fast-track care programs in hip and knee replacements broadly introduced at Swedish hospitals during the period 2011–2015 are associated with slightly better PROs 1 year after the operation compared with programs defined as not fast-track. The small differences may not be of clinical relevance, but our results indicate that fast-track programs may be at least as good as the conventional care from the perspective of PROs. Further studies are needed to identify which factors are the most important in order to refine and further improve the clinical pathway and care process based on the principles of fasttrack. Supplementary data Tables 3, 7, and 9–11 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/ 17453674.2020.1733375 UB, AN, OR, AWD, and MS conceived and planned the study. UB and EN performed the statistical analyses. All authors discussed the results and commented on the manuscript, which was drafted by UB.

Antrobus J D, Bryson G L. Enhanced recovery for arthroplasty: good for the patient or good for the hospital? Can J Anesth [Review] 2011; 58(10): 8916. Berg U, Bülow E, Sundberg M, Rolfson O. No increase in readmissions or adverse events after implementation of fast-track program in total hip and knee replacement at 8 Swedish hospitals: an observational before-and-after study of 14,148 total joint replacements 2011–2015. Acta Orthop 2018; 89 (5): 522-7. Brock T M, Baker P N, Rushton S, Bardgett M, Deehan D. Length of stay and its impact upon functional outcomes following lower limb arthroplasty. Knee Surg Sports Traumatol Arthrosc 2017; 25(9): 2676-81. Burstrom K, Sun S, Gerdtham U G, Henriksson M, Johannesson M, Levin L A, Zethraeus N. Swedish experience-based value sets for EQ-5D health states. Qual Life Res 2014; 23(2): 431-42. Christelis N, Wallace S, Sage C E, Babitu U, Liew S, Dugal J, Nyulasi I, Mutalima N, Tran T, Myles P S. An enhanced recovery after surgery program for hip and knee arthroplasty. Med J Aust 2015; 202(7):363-8.

312

Coenders M J, Mathijssen N M C, Vehmeijer S B W. Three and a half years’ experience with outpatient total hip arthroplasty. Bone Joint J 2020; 102b(1): 82-9. Cohen J. A power primer. Psychol Bull 1992; 112(1): 155-9. Deng Q F, Gu H Y, Peng W Y, Zhang Q, Huang Z D, Zhang C, Yu Y X. Impact of enhanced recovery after surgery on postoperative recovery after joint arthroplasty: results from a systematic review and meta-analysis. Postgrad Med J 2019; 94(1118): 678-93. EuroQolGroup T. EuroQol—a new facility for the measurement of healthrelated quality of life. Health Policy 1990; 16(3): 199-208. Gordon M, Paulsen A, Overgaard S, Garellick G, Pedersen A B, Rolfson O. Factors influencing health-related quality of life after total hip replacement: a comparison of data from the Swedish and Danish hip arthroplasty registers. BMC Musculoskeletal Disord 2013; 14: 316. Goyal N, Chen A F, Padgett S E, Tan T L, Kheir M M, Hopper R H, Hamilton W G, Hozack W J. Otto Aufranc Award: A multicenter, randomized study of outpatient versus inpatient total hip arthroplasty. Clin Orthop Relat Res 2017; 475(2): 364-72. Greene M E, Rader K A, Garellick G, Malchau H, Freiberg A A, Rolfson O. The EQ-5D-5L improves on the EQ-5D-3L for health-related quality-oflife assessment in patients undergoing total hip arthroplasty. Clin Orthop Relat Res 2015; 473(11): 3383-90. Gromov K, Jørgensen C C, Petersen P B, Kjaersgaard-Andersen P, Revald P, Troelsen A, Kehlet H, Husted H. Complications and readmissions following outpatient total hip and knee arthroplasty: a prospective 2-center study with matched controls. Acta Orthop 2019; 90(3): 281-5. Harsten A, Bandholm T, Kehlet H, Toksvig-Larsen S. Tourniquet versus no tourniquet on knee-extension strength early after fast-track total knee arthroplasty: a randomized controlled trial. The Knee 2015; 22(2): 126-30. Husted H, Holm G. Fast track in total hip and knee arthroplasty: experiences from Hvidovre University Hospital, Denmark. Injury 2006; 37(Suppl. 5): S31-5. Jones C A, Pohar S. Health-related quality of life after total joint arthroplasty: a scoping review. Clin Geriatr Med 2012; 28(3): 395-429. Kehlet H, Wilmore D W. Evidence-based surgical care and the evolution of fast-track surgery. Ann Surg 2008; 248(2): 189-98. Kärrholm J L H, Malchau H, Mohaddes M, Rogmark C, Rolfson O. The Swedish Hip Arthroplasty Register. Annual report 2015. Available from https://registercentrum.blob.core.windows.net/shpr/r/-rsrapport2015-B1dTcO3Pe.pdf. Larsen K, Sorensen O G, Hansen T B, Thomsen P B, Soballe K. Accelerated perioperative care and rehabilitation intervention for hip and knee replacement is effective: a randomized clinical trial involving 87 patients with 3 months of follow-up. Acta Orthop 2008; 79(2): 149-59. Larsen K, Hansen T B, Soballe K, Kehlet H. Patient-reported outcome after fast-track hip arthroplasty: a prospective cohort study. Health Qual Life Outcomes 2010; 8: 144. Larsen K, Hansen T B, Soballe K, Kehlet H. Patient-reported outcome after fast-track knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2012; 20(6): 1128-35. Ledin H, Aspenberg P, Good L. Tourniquet use in total knee replacement does not improve fixation, but appears to reduce final range of motion. Acta Orthop 2012; 83(5): 499-503. Lindgren J V, Wretenberg P, Karrholm J, Garellick G, Rolfson O. Patientreported outcome is influenced by surgical approach in total hip replacement: a study of the Swedish Hip Arthroplasty Register including 42,233 patients. Bone Joint J 2014; 96-b(5): 590-6.

Acta Orthopaedica 2020; 91 (3): 306–312

Machin J T, Phillips S, Parker M, Carrannante J, Hearth M W. Patient satisfaction with the use of an enhanced recovery programme for primary arthroplasty. Ann R Coll Surg Engl 2013; 95(8): 577-81. Nemes S, Burstrom K, Zethraeus N, Eneqvist T, Garellick G, Rolfson O. Assessment of the Swedish EQ-5D experience-based value sets in a total hip replacement population. Qual Life Res 2015; 24(12): 2963-70. Pilot P, Bogie R, Draijer W F, Verburg A D, van Os J J, Kuipers H. Experience in the first four years of rapid recovery: is it safe? Injury 2006; 37(Suppl. 5): S37-40. Rolfson O, Donahue G S, Hallsten M, Garellick G, Karrholm J, Nemes S. Patient-reported outcomes in cemented and uncemented total hip replacements. Hip Int 2016; 26(5): 451–7. Roos E M, Lohmander L S. The Knee injury and Osteoarthritis Outcome Score (KOOS): from joint injury to osteoarthritis. Health Qual Life Outcomes 2003; 1: 64. Roos E M, Roos H P, Lohmander L S, Ekdahl C, Beynnon B D. Knee Injury and Osteoarthritis Outcome Score (KOOS): development of a self-administered outcome measure. J Orthop Sports Phys Ther 1998; 28(2): 88-96. Specht K, Kjaersgaard-Andersen P, Kehlet H, Wedderkopp N, Pedersen B D. High patient satisfaction in 445 patients who underwent fast-track hip or knee replacement. Acta Orthop 2015; 86(6): 702-7. Specht K, Kjaersgaard-Andersen P, Pedersen B D. Patient experience in fasttrack hip and knee arthroplasty: a qualitative study. J Clin Nurs 2016; 25(56): 836-45. J Stowers M D, Manuopangai L, Hill A G, Gray J R, Coleman B, Munro J T. Enhanced recovery after surgery in elective hip and knee arthroplasty reduces length of hospital stay. ANZ J Surg 2016; 86(6): 475-9. Strickland L H, Kelly L, Hamilton T W, Murray D W, Pandit H G, Jenkinson C. Early recovery following lower limb arthroplasty: qualitative interviews with patients undergoing elective hip and knee replacement surgery. Initial phase in the development of a patient-reported outcome measure. J Clin Nurs 2018; 27(13-14): 2598-2608. Tubach F, Ravaud P, Martin-Mola E, Awada H, Bellamy N, Bombardier C, Felson D T, Hajjaj-Hassouni N, Hochberg M, Logeart I, Matucci-Cerinic M, van de Laar M, van der Heijde D, Dougados M. Minimum clinically important improvement and patient acceptable symptom state in pain and function in rheumatoid arthritis, ankylosing spondylitis, chronic back pain, hand osteoarthritis, and hip and knee osteoarthritis: results from a prospective multinational study. Arthritis Care Res 2012; 64(11): 1699-707. Wainwright T W, Kehlet H. Fast-track hip and knee arthroplasty: have we reached the goal? Acta Orthop 2019; 90(1): 3-5. W-Dahl A, Sundberg M, Lidgren L, Ranstam J, Robertsson O. An examination of the effect of different methods of scoring pain after a total knee replacement on the number of patients who report unchanged or worse pain. Bone Joint J 2014; 96-b(9): 1222-6. Winther S B, Foss O A, Wik T S, Davis S P, Engdal M, Jessen V, Husby O S. 1-year follow-up of 920 hip and knee arthroplasty patients after implementing fast-track. Acta Orthop 2015; 86(1): 78-85. Zhang W, Li N, Chen S, Tan Y, Al-Aidaros M, Chen L. The effects of a tourniquet used in total knee arthroplasty: a meta-analysis. J Orthop Surg Res 2014; 9(1): 13. Zhou K, Ling T, Wang H, Zhou Z, Shen B, Yang J, Kang P, Pei F. Influence of tourniquet use in primary total knee arthroplasty with drainage: a prospective randomised controlled trial. J Orthop Surg Res 2017; 12(1): 172. Zhu S, Qian W, Jiang C, Ye C, Chen X. Enhanced recovery after surgery for hip and knee arthroplasty: a systematic review and meta-analysis. Postgrad Med J 2017; 93(1106): 736-42.

Acta Orthopaedica 2020; 91 (3): 313–318

313

Prognostic factors for inpatient functional recovery following total hip and knee arthroplasty: a systematic review Nicola HEWLETT-SMITH 1,2, Rodney POPE 1,3, James FURNESS 1, Vini SIMAS 1, and Wayne HING 1 1 Faculty of Health Sciences and Medicine, Bond University, Gold Coast, Australia; 2 Allied 3 School of Community Health, Charles Sturt University, Albury, Australia

Health Department, The Wesley Hospital, Brisbane, Australia;

Correspondence: nicola.hewlett-smith@uchealth.com.au Submitted 2019-11-26. Accepted 2020-02-04.

Background and purpose — Essential for safe and timely hospital discharge, inpatient functional recovery following lower limb arthroplasty is also variable. A previous systematic review reported moderate and conflicting levels of evidence regarding patient-related predictors of inpatient recovery for primary total hip arthroplasty (THA). A systematic review of surgical prognostic factors for inpatient recovery following THA or total knee arthroplasty (TKA) is yet to be undertaken. We identified patient and surgical prognostic factors for inpatient functional recovery following THA and TKA; determined whether inpatient functional recovery varies between these procedures; and established whether validated outcome measures relevant to the patient’s functional requirements for hospital discharge are routinely assessed. Patients and methods — Critical Appraisal Skills Programme checklists assessed methodological quality, and a best-evidence synthesis approach determined the levels of evidence supporting individual prognostic factors. PubMed, CINAHL, Embase, Scopus, and PEDro databases were searched from inception to May 2019. Included studies examined patient or surgical prognostic factors and a validated measure of post-operative function within 2 weeks of primary, unilateral THA or TKA. Results — Comorbidity status and preoperative function are supported by a strong level of evidence for TKA. For THA, no strong level of evidence was found for patientrelated prognostic factors, and no surgical factors were independently prognostic for either arthroplasty site. Limited evidence supports fast-track protocols in the TKA population. Interpretation — Preoperative screening and optimization is recommended. Assessment of Enhanced Recovery Pathways using validated outcome measures appropriate for the early postoperative period is warranted.

The International Classification of Function, Disability and Health (WHO 2013) describes the interdependent relationship among function, activity, and participation. Following lower limb arthroplasty, functional recovery is key to the independent performance of fundamental activities of daily living (ADL) such as walking, transferring in and out of bed, and climbing stairs; achieving these milestones is necessary for safe and timely hospital discharge (Shields et al. 1995, Hoogeboom et al. 2015, Poitras et al. 2015). Inability to perform basic ADL increases the patient’s risk of social isolation, falls, and the need for additional resources such as rehabilitation and community services (Poitras et al. 2015). To promote rapid recovery, multimodal Enhanced Recovery Pathways (ERP) are increasingly used for lower limb arthroplasty (Scott et al. 2013). However, the success of these pathways is primarily assessed via non patient-centric measures including morbidity and mortality, readmission rates, length of stay (LOS), and organizational economic savings (Husted 2012). Functional recovery is linked to discharge destination, longer-term functional outcomes, quality of life (Elbaz et al. 2015), patient satisfaction (Scott et al. 2012), LOS, and associated costs (Husted et al. 2008, 2010, Ibrahim et al. 2013). However, few studies have specifically examined inpatient functional recovery as an outcome following lower limb arthroplasty, using valid measures. While studies have considered achievement of hospitalspecific functional discharge criteria, these constitute neither a standardized nor a validated outcome measure. Whilst LOS may be influenced by wide-ranging factors (Husted et al. 2008, 2010, 2011, Den Hertog et al. 2012, Napier et al. 2013, Elings et al. 2016), inpatient functional recovery is commonly thought to be primarily affected by patient and surgical factors. Surprisingly, inpatient functional recovery has been investigated by a single systematic review. Based on the results of 2

314

Acta Orthopaedica 2020; 91 (3): 313–318

Table 1. Eligibility criteria Criteria Inclusion

Exclusion

Population Humans undergoing primary elective total hip or knee arthroplasty Context Australian and international studies carried out in public and private hospital settings Language All languages Recency of publication All date periods preceding the search date Time period Studies examining outcomes in the early postoperative period (≤ 2 weeks postoperatively) Prognostic factors Studies examining the relationship between 1 or more surgical or patient-related prognostic factors and functional performance or patient-reported outcome measures Outcomes Studies examining at least 1 validated functional performance or patient reported outcome measure indicating postoperative functional recovery

Bilateral total hip or knee arthroplasty Unicompartmental knee arthroplasty Hip joint re-surfacing Arthroplasty performed secondary to fracture (trauma or pathological) Studies not carried out within a public or private hospital Articles not reporting primary research Studies where language translation was not possible. However, these studies were noted for completeness, prior to exclusion

studies, Elings et al. (2015) reported moderate and conflicting levels of evidence regarding the association between patientrelated factors and inpatient functional recovery. Therefore, this systematic review examines the evidence for patient and surgical prognostic factors for inpatient functional recovery following both total hip arthroplasty (THA) and total knee arthroplasty (TKA); determines whether inpatient functional recovery varies between these procedures; and identifies whether validated outcome measures relevant to the patient’s functional requirements for hospital discharge are routinely assessed. The identification of surgical prognostic factors may provide an opportunity to refine ERP, whilst patient-related factors may aid in identifying those at risk of delayed recovery, enabling medical optimization, prehabilitation, and early discharge planning (Oosting et al. 2016).

1) were established and applied to the search results during initial screening of titles and abstracts. Final selection of articles based on full text review was performed independently by 2 reviewers. Differences were resolved by consensus. Critical Appraisal Skills Programme (CASP) (2019a, 2019b, 2019c) checklists were used to address the methodological quality of the differing study designs and examine external validity, internal validity (bias), internal validity (confounding), and statistical power. To grade methodological quality, a scoring system was devised by the reviewers, and applied to each CASP checklist. Subsequently, Questions 7 and 8 of each checklist, and Question 12 of the Cohort Studies checklist were modified to elicit a “Yes,” “No,” or “Can’t tell” response (Appendix 2). For each checklist question, a “Yes” response scored 1, and a response of “Can’t tell” or “No” scored 0; for questions involving a 2-part answer, parts (a) and (b) were scored separately. Using this system, the CASP checklists for Randomized Controlled Trials, Case Control Studies, and Cohort Studies had a maximum possible score of 11, 12, and 14, respectively. Scores were converted to a percentage and ranges were determined (by the reviewers) to reflect methodological quality as follows: < 30% low quality, 31–65% medium quality, and > 65% high quality. Studies were independently appraised by 2 reviewers, and Cohen’s Kappa (κ) (Cohen 1960) assessed level of agreement; differences were resolved by discussion and consensus. Extracted data was tabulated, including: study design, context, sample size, demographics, arthroplasty site, prognostic factors, validated measures of postoperative functional recovery, and the time points at which these were assessed. Metaanalysis was not possible due to the methodological hetero-

Method The systematic review protocol was registered with PROSPERO (PROSPERO Registration: CRD42019136206), and reporting is in accordance with the PRISMA statement. A comprehensive search of PubMed, CINAHL, Embase, Scopus, and PEDro databases was undertaken on May 31, 2019. The search strategy included key search terms relating to prognostic factors, hip and knee arthroplasty, and function. Subject headings specific to individual databases were utilized, and wildcards employed. No date range or language filters were applied. The PubMed/MEDLINE search strategy is presented in Appendix 1. Reference lists were also examined to capture all potentially eligible publications. Eligibility criteria (Table

Studies where the postoperative time point at which outcome measures were assessed is not specified or was > 2 weeks Studies where the prognostic factors of interest pertained only to determining the efficacy of a treatment intervention, the specific properties of the prosthesis used or patient genetic, blood, or radiological markers

Acta Orthopaedica 2020; 91 (3): 313–318

Records identified through database search (n = 7,680): – Pubmed, 2,071 – CINAHL, 847 – Embase, 3,170 – Scopus, 729 – PEDro, 863

Additional records identified through other sources (n = 44)

Total number of records (n = 7,724) Records following deduplication (n = 4,528) Records excluded via title and abstract (n = 4,371) Full-text articles assessed for eligibility (n = 157) Full-text articles excluded, with reasons (n = 140): – population, 28 – context, 6 – follow-up time period, 65 – predictor variables, 5 – outcome measures, 33 – pooled results, 1 – duplicated results, 2 Studies included in qualitative synthesis (n = 17)

PRISMA flow diagram.

geneity of included studies, therefore a best evidence synthesis approach was employed. Evidence levels were ranked as follows: strong evidence is provided by ≥ 2 studies with low risk of bias and by generally consistent findings in all studies (≥ 75% of the studies reported consistent findings); moderate evidence is provided by 1 low risk of bias study and ≥ 2 moderate/high risk of bias studies or by ≥ 2 moderate/ high risk of bias studies and by generally consistent findings in all studies (≥ 75%); limited evidence is provided by ≥ 1 moderate/high risk of bias studies or 1 low risk of bias study and by generally consistent findings (≥ 75%); conflicting evidence is provided by conflicting findings (< 75% of the studies reported consistent findings) (Eijgenraam et al. 2018).

Results The search identified 7,724 records and, following screening, 17 studies were included (Figure). These incorporated 1,171 THA and 1,662 TKA procedures. 8 studies investigated THA, 8 TKA, and 1 both procedures (Table 2, see Supplementary data). 12 studies examined patient-related factors (Table 3, see Supplementary data) and 9 studies investigated surgical factors (Table 4, see Supplementary data). Numerous tools evaluated comorbidity status and preoperative function. Postoperative functional recovery was assessed via 14 different validated functional performance and patient-reported outcome mea-

315

sures (PROM). Assessment time points varied significantly between studies within the 2-week postoperative period. Critical appraisal results are presented in Appendix 3; 7 studies were rated as high methodological quality, 7 as medium quality, and 3 as low quality. There was strong level of agreement between the two reviewers’ judgements (κ = 0.944, p < 0.001). The best-evidence synthesis for prognostic factors for early functional recovery following THA and TKA is presented in Tables 5 and 6 respectively (see Supplementary data). The heterogeneity of outcome measures employed in the included studies is presented in Appendix 4. Only 7 studies utilized objective outcome measures that assess key functional tasks representative of ADL required for discharge. The Modified Barthel Index (MBI; Shah et al. 1989), Iowa Level of Assistance Scale (ILAS; Shields et al. 1995) and modified Iowa Level of Assistance Scale (mILAS; Oldmeadow et al. 2006) each assess tasks including bed or chair transfers, ambulation, stair climbing, and the amount of assistance required for their achievement. However, the mILAS was further modified (from that published by Oldmeadow et al. 2006) in 2 studies (Elings et al. 2016, van der Sluis et al. 2017) and only partially implemented in all 4 studies where it was assessed, potentially compromising its validity, the generalizability of results and also the ability to compare outcomes between studies. Morri et al. (2016) describe the scoring method for the ILAS inaccurately, casting doubt on the validity of its implementation.

Discussion This systematic review examines the evidence for patientrelated and surgical prognostic factors for inpatient functional recovery following THA and TKA; determines whether inpatient functional recovery varies depending on arthroplasty site; and identifies whether inpatient functional recovery was assessed using validated outcome measures relevant to the patient’s functional requirements for hospital discharge. The level of evidence for patient-related prognostic factors and inpatient functional recovery differs between THA and TKA populations. However, associations between timed and observational performance measures of preoperative physical function or comorbidity status (ASA grade) and inpatient recovery was evident for both arthroplasty sites. Conflicting evidence exists for body mass index (BMI) and age as prognostic factors in both arthroplasty populations. The role of sex was supported by limited evidence and conflicting evidence in TKA and THA studies, respectively. These results contrast to those published by Elings et al. (2015), which (based on 2 included studies) reported moderate-level evidence for preoperative ADL status, female sex, and BMI; and conflicting evidence for increased age, as prognostic factors of delayed inpatient recovery following THA. Moderate-level evidence indicated no association for ASA

316

grade; however, it should be noted this result was based on the findings of a single study. Greater comorbidity (Charnley class C), poorer preoperative functional performance (10meter walk test, Timed Up and Go [TUG]), and increased age were also confirmed prognostic factors of delayed functional recovery in a further study of 294 THA patients (Oosting et al. 2016), which did not meet inclusion criteria in this review due to some participants undergoing revision surgery. In summary, preoperative function has consistently been associated with early postoperative function following THA and TKA. The roles of increased comorbidity, older age, sex, and BMI must also be considered. The confirmation of these prognostic factors highlights the need for routine preoperative patient screening. Screening could be implemented conjointly with the decision to proceed to surgery, thus maximizing the preoperative window. Simple performance measures may identify patients potentially at risk of delayed recovery, providing the opportunity for preoperative medical and functional optimization, and prompt discharge planning (Elings et al. 2015, Oosting et al. 2016). Prehabilitation has been demonstrated to improve preoperative function (Swank et al. 2011, Clode et al. 2018) and may successfully be implemented via telerehabilitation (Doiron-Cadrin et al. 2019), thereby capturing patients with reduced access (Westby et al. 2010), whilst avoiding significant cost burden to both patients and healthcare organizations (Fusco and Turchetti 2016). This review did not identify any surgical factors that were independently prognostic for postoperative functional recovery. Although the overall methodological quality of studies examining surgical factors was of a moderate to high level, sample sizes were small (40â&#x20AC;&#x201C;67 participants) in 4 studies, and 3 studies did not report confidence intervals for their results. These results suggest that individual surgical factors may not significantly impact recovery and rather that ERP or Fast-track pathways, which address many aspects of the surgical pathway, are more effective in promoting early functional return. Further research is required to assess the impact of ERP using validated functional outcome measures. Differences in the pattern of inpatient recovery following THA and TKA require further research. A single study (Kennedy et al. 2006) modelled the recovery pattern for both sites of arthroplasty; however, the methodological quality of this study limits the generalizability of the results. Hierarchical linear modelling was used due to the varied patient numbers and lack of standardization of postoperative time points, and several confounding factors were not accounted for. LOS was reported in 12 studies and appears to range from 2 to 39 days, with 9 studies stating or implying the use of discharge criteria. Due to the heterogeneity of studies with regard to the presence or type of discharge criteria used, how rigorously the discharge criteria were implemented, and when and how functional recovery was assessed, there is insufficient evidence to determine whether inpatient functional recovery differs by arthroplasty site.

Acta Orthopaedica 2020; 91 (3): 313â&#x20AC;&#x201C;318

Validated tools for assessing short-term postoperative function following lower limb arthroplasty are lacking (Kimmel et al. 2016, Poitras et al. 2016). Currently there is no gold standard for evaluating functional recovery in acute hospital inpatients (Kimmel et al. 2016), which may explain the heterogeneity of outcome measures employed. Several PROMs including the Lower Extremity Function Scale, Knee injury and Osteoarthritis Outcome Score, and WOMAC are appropriate for assessing longer-term functional outcomes as they address more advanced functional activities (Poitras et al. 2016) however these activities are not achieved within the acute recovery phase and are not reflective of ADL required for hospital discharge. Low to moderate correlations are reported between PROMs and performance measures in the early postoperative period following THA and TKA (Mizner et al. 2011, Poitras et al. 2016). PROMs are subjective and may be influenced by many factors (Poitras et al. 2016), including perceived level of exertion (Mizner at al. 2011), anxiety, and expectations regarding recovery (Salmon et al. 2001b); therefore performance-based measures are necessary to objectively assess actual patient function (Mizner et al. 2011). However, performance measures should be clinically relevant, easily integrated into routine postoperative assessment, appropriate to the time point at which they are assessed, and implemented in a standardized manner to enable evaluation of patient outcomes across organizations. PROMs have been adopted by some National Joint Registries to record longer-term functional outcomes. Similar integration of standardized performance-based assessments could aid in generating a database of early postoperative functional outcomes, thus providing more pertinent information than LOS comparisons. A strength of this review is the broad search undertaken with few exclusion criteria to ensure all available evidence regarding patient-related and surgical prognostic factors and inpatient functional recovery following THA and TKA was captured. Studies published in all languages were considered for inclusion. There are also several limitations. The heterogeneity of outcome measures assessed and, additionally, the modification, or varied and partial implementation of valid outcome measures (in particular the mILAS) limits the comparison of results between studies. For this reason a metaanalysis was not possible. Not all included studies published results for their early postoperative time points. Moreover, not all studies reported 95% confidence intervals, therefore the significance of some results may be questioned. None of the included studies collected data within the last 4 years, thus the potential impact of more recent surgical advances including muscle-sparing surgical approaches and robotic-assisted surgery has not been assessed. For the purpose of screening, studies where joint ROM was the only postoperative outcome measure examined were excluded. Although a noted contributor, joint ROM alone is not sufficient to enable mobility or the performance of ADL.

Acta Orthopaedica 2020; 91 (3): 313–318

Conclusion Based on the findings of this review, there is strong level of evidence that comorbidity status determined by ASA grade, and preoperative functional status assessed by the TUG are prognostic factors for inpatient functional recovery following TKA. No strong level of evidence was found for patientrelated prognostic factors for inpatient recovery following THA. No surgical factors were found to be independent prognostic factors for inpatient recovery following either THA or TKA; however, limited evidence supports Fast-track protocols in the TKA population. Studies assessing inpatient functional recovery are heterogeneous. Variance in methodological quality, variables examined, outcome measures, and the time points at which they are assessed makes comparison of results difficult. With shorter LOS desirable, preoperative screening is recommended to identify patients at risk of delayed inpatient recovery enabling prehabilitation, medical optimization, and early discharge planning. Valid, standardized performance measures assessing basic functional tasks would assist in objectively determining patient readiness for discharge (Shields et al. 1995), evaluating the success of ERP interventions (Poitras et al. 2016), and enable benchmarking across organizations. Surgical advances in lower limb arthroplasty and their impact on impatient functional recovery are also worthy of investigation. Funding and potential conflicts of interest This research was supported by an Australian Government Research Training Program Scholarship. Each author certifies that he or she has no commercial associations that might pose a conflict of interest in connection with the submitted article. Data statement N H-S is registered with the data repository Open Science Framework. Supplementary data Tables 2–6 and Appendices 1–4 are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674.2020.1744852

Conception/planning: NH-S, RP, WH. Initial screening: NH-S. Full text review: NH-S, VS. Data extraction: NH-S, VS. Critical appraisal: NH-S, VS. Best evidence synthesis: NH-S, JF. Results: NH-S, RP, JF. Discussion/ conclusion: NH-S, RP, JF, WH. Formatting/editing: NH-S, RP, JF, WH. The authors wish to acknowledge Ross Ferguson (The Wesley Hospital) and David Honeyman (Bond University) for their support and contributions. Acta thanks Petri Virolainen and André Stark for help with peer review of this study.

317

Bade M J, Kittelson J M, Kohrt W M, Stevens-Lapsley J E. Predicting functional performance and range of motion outcomes after total knee arthroplasty. Am J Phys Med Rehabil 2014; 93: 579-85. Carli F, Clemente A, Asenjo J F, Kim D J, Mistraletti G, Gomarasca M, Morabito A, Tanzer M. Analgesia and functional outcome after total knee arthroplasty: periarticular infiltration vs continuous femoral nerve block. Br J Anaesthesia 2010; 105(2): 185-95. Carmichael N M E, Katz J, Clarke H, Kennedy D, Kreder H J, Gollish J, McCartney C J L. An intensive perioperative regimen of pregabalin and celecoxib reduces pain and improves physical function scores six weeks after total hip arthroplasty: a prospective, randomized controlled trial. Pain Res Manag 2013; 18(3): 127-32. Clode N J, Perry M A, Wulff L. Does physiotherapy prehabilitation improve pre-surgical outcomes and influence patient expectations prior to knee and hip joint arthroplasty? Int J Orthop Trauma Nurs 2018; 30: 14-19. Critical Appraisal Skills Programme. CASP (Case Control Study) Checklist. 2019a [online]. Available at: https://casp-uk.net Accessed: August 2019. Critical Appraisal Skills Programme. CASP (Cohort Study) Checklist. 2019b [online]. Available at: https://casp-uk.net Accessed: August 2019. Critical Appraisal Skills Programme. CASP (Randomized Controlled Trial) Checklist. 2019c [online]. Available at: https://casp-uk.net Accessed: August 2019. Cohen J. A coefficient of agreement for nominal scales. Educ Psychol Meas 1960; 20: 37-46. Den Hertog A, Gliesche K, Timm J, Mühlbauer B, Zebrowski S. Pathway-controlled fast-track rehabilitation after total knee arthroplasty: a randomized prospective clinical study evaluating the recovery pattern, drug consumption, and length of stay. Arch Orthop Trauma Surg 2012; 132: 1153-63. Doiron-Cadrin P, Kairy D, Vendittoli P-A, Lowry V, Poitras S, Desmeules F. Feasibility and preliminary effects of a tele- prehabilitation program and an in-person prehabilitation program compared to usual care for total hip or knee arthroplasty candidates: a pilot randomized controlled trial. Disabil Rehabil 2019; 13: 1-10. Eijgenraam S, Reijman M, Bierma-Zeinstra S, Van Yperen D, Meuffels D. Can we predict the outcome of arthroscopic partial meniscectomy? A systematic review. Br J Sports Med 2018; 52(8): 514-21. Elbaz A, Debbi E, Segal G, Mor A, Bar-Ziv Y, Benkovich V, Shasha N, Shoham-Blonder R, Fruchter G, Debi R. New approach for the rehabilitation of patients following total knee arthroplasty. Physiother 2015; 101(1): e412e413. Elings J, Hoogeboom T, van der Sluis G, van Meeteren N L U. What preoperative patient-related factors predict inpatient recovery of physical functioning and length of stay after total hip arthroplasty? A systematic review. Clin Rehabil 2015; 29(5): 477-92. Elings J, Van Der Sluis G, Goldbohm R A, Garre F G, De Gast A, Hoogeboom T, Van Meeteren N L. Development of a risk stratification model for delayed inpatient recovery of physical activities in patients undergoing total hip replacement. J Orthop Sports Phys Ther 2016; 46(3): 135-43. Fransen B L, Hoozemans M J M, Argelo K D S, Keijser L C M, Burger B J. Fast-track total knee arthroplasty improved clinical and functional outcome in the first 7 days after surgery: a randomized controlled pilot study with 5-year follow-up. Arch Orthop Trauma Surg 2018; 138: 1305-16. Fusco F, Turchetti G. Telerehabilitation after total knee replacement in Italy: cost-effectiveness and cost-utility analysis of a mixed telerehabilitationstandard rehabilitation programme compared with usual care. BMJ Open 2016; 6: e009964. Hoogeboom T, van Meeteren N, Schank K, Kim R, Miner T, Stevens-Lapsley J. Risk factors for delayed inpatient functional recovery after total knee arthroplasty. Biomed Res Int 2015; Article ID 167643, 5 pages. Husted H. Fast-track hip and knee arthroplasty: clinical and organisational aspects. Acta Orthop Oct 2012; 83 (Suppl. 346): 2-38 (thesis, incorporates 9 articles). Husted H, Holm G, Jacobsen S. Predictors of length of stay and patient satisfaction after hip and knee replacement surgery: fast-track experience in 712 patients. Acta Orthop 2008; 79(2): 168-73.

318

Husted H, Hansen H C, Holm G, Bach-Dal C, Rud K, Andersen K L, Kehlet H. What determines length of stay after total hip and knee arthroplasty? A nationwide study in Denmark. Arch Orthop Trauma Surg 2010; 130: 263-8. Husted H, Lunn T, Troelsen A, Gaarn-Larsen L, Kristensen B, Kehlet H. Why still in hospital after fast-track hip and knee arthroplasty? Acta Orthop 2011; 82(6): 679-84. Ibrahim M, Khan M, Nizam I, Haddad F. Peri-operative interventions producing better functional outcomes and enhanced recovery following total hip and knee arthroplasty: an evidence-based review. BMC Med 2013; 11: 37. Ilfeld B M, Le L T, Meyer R S, Mariano E R, Vandenbourne K, Duncan P W, Sessler D I, Kayser Enneking F, Shuster J J, Theriaque D W, Berry L F, Spadoni E H. Ambulatory continuous femoral nerve blocks decrease time to discharge after tricompartment total knee arthroplasty: a randomized, triple-masked, placebo-controlled study. Anesthesiol 2008; 108: 703-13. Kennedy D M, Hanna S E, Stratford P W, Wessel J, Gollish J D. Pre-operative function and gender predict pattern of functional recovery after hip and knee arthroplasty. J Arthroplasty 2006; 21(4): 559-66. Kennedy D M, Stratford P W, Robarts S, Gollish J D. Using outcome measure results to facilitate clinical decisions the first year after total hip arthroplasty. J Orthop Sports Phys Ther 2011; 41(4): 232-9. Kessler S, Kafer W. Overweight and obesity: two predictors for worse early outcome in total hip replacement? Obesity 2007; 15(11): 2840-5. Kimmel L, Elliott J, Sayer J, Holland A. Assessing the reliability and validity of a physical therapy measurement tool—the modified Iowa level of assistance scale—in acute hospital inpatients. Phys Ther 2016; 96: 176-82. Maiorano E, Bodini B D, Cavaiani F, Pelosi C, Sansone V. Length of stay and short-term functional outcomes after total knee arthroplasty: Can we predict them? The Knee 2017; 24: 116-20. Mizner R, Petterson S, Clements K, Zeni Jr J, Irrgang J, Snyder-Mackler L. Measuring functional improvement after total knee arthroplasty requires both performance-based and patient-report assessments. J Arthroplasty 2011; 26(5): 728-37. Morri M, Natali E, Tosarelli D. At discharge gait speed and independence of patients provides a challenge for rehabilitation after total joint arthroplasty: an observational study. Arch Physiother 2016; 6: 6. Napier R J, Spence D, Diamond O, O’Brien S, Walsh T, Beverland D E. Modifiable factors delaying early discharge following primary joint arthroplasty. Eur J Orthop Surg Traumatol 2013; 23: 665-69. Ogonda L, Wilson R, Archbold P, Lawlor M, Humphreys P, O’Brien S, Beverland D. A minimal-incision technique in total hip arthroplasty does not improve early postoperative outcomes: a prospective, randomized, controlled trial. J Bone Joint Surg 2005; 87(4): 701-10. Oldmeadow L B, Edwards E R, Kimmel L A, Kipen E, Robertson V J, Bailey M J. No rest for the wounded: early ambulation after hip surgery accelerates recovery. ANZ J Surg 2006; 76: 607-11. Oosting E, Hoogeboom T, Appekman-de Vries S, Swets A, Jaap J, van Meeteren N. Preoperative prediction of inpatient recovery of function after

Acta Orthopaedica 2020; 91 (3): 313–318

total hip arthroplasty using performance-based tests: a prospective cohort study. Disabil Rehabil 2016; 38(13): 1243-9. Poitras S, Wood K, Beaule P. Predicting early clinical function after hip or knee arthroplasty. Bone Joint Res 2015; 4: 145-51. Poitras S, Wood K, Savard J, Dervin G, Beaule P. Assessing functional recovery shortly after knee or hip arthroplasty: a comparison of the clinimetric properties of four tools. BMC Musculoskelet Disord 2016; 17: 478. Salmon P, Hall G M, Peerbhoy D. Influence of the emotional response to surgery on functional recovery during 6 months after hip arthroplasty. J Behav Medicine 2001a; 24(5): 489-502. Salmon P, Hall G M, Peerbhoy D, Shenkin A, Parker C. Recovery from hip and knee arthroplasty: patients’ perspective on pain, function, quality of life, and well-being up to 6 months postoperatively. Arch Phys Med Rehabil 2001b; 82: 360-6. Scott C E, Bugler K E, Clement N D, MacDonald D, Howie C R, Biant L C. Patient expectations of arthroplasty of the hip and knee. J Bone Joint Surg Br 2012; 94(7): 974-81. Scott N, McDonald D, Campbell J, Smith R, Carey A, Johnston I, James K, Breusch S. The use of enhanced recovery after surgery (ERAS) principles in Scottish orthopaedic units—an implementation and follow-up at 1 year, 2010–2011: a report from the Musculoskeletal Audit, Scotland. Arch Ortho Trauma Surg 2013; 133: 117-24. Shah S, Vanclay F, Cooper B. Improving the sensitivity of the Barthel Index for stroke rehabilitation. J Clin Epidemiol 1989; 42: 703-9. Shields R, Enloe L, Evans R, Smith K, Steckel S, Goldsmith C. Reliability, validity and responsiveness of functional tests in patients with total joint replacement. Phys Ther 1995; 75(3): 169-79. Swank A M, Kachelman J B, Bibeau W, Quesada P M, Nyland J, Malkani A, Topp R V. Prehabilitation before total knee arthroplasty increases strength and function in older adults with severe osteoarthritis. J Strength Cond Res 2011; 25(2): 318-25. Van der Sluis G, Goldbohm R A, Elings J E, Nijhuis-van der Sanden M W, Akkermans R P, Bimmel R, Hoogeboom T J, van Meeteren N L. Preoperative functional mobility as an independent determinant of inpatient functional recovery after total knee arthroplasty during three periods that coincided with changes in clinical pathways. Bone Joint J 2017; 99-B: 211-17. Wang A, Ackland T, Hall S, Gilbey H, Parsons R. Functional recovery and timing of hospital discharge after primary total hip arthroplasty. Aust NZ J Surg 1998; 68: 580-3. Westby M D, Backman C L. Patient and health professional views on rehabilitation practices and outcomes following total hip and knee arthroplasty for osteoarthritis: a focus group study. BMC Health Serv Res 2010; 10: 119. World Health Organization. How to use the ICF: a practical manual for using the International Classification of Functioning, Disability and Health (ICF): Exposure draft for comment. Geneva: WHO; October 2013.

Acta Orthopaedica 2020; 91 (3): 319–325

319

Patients undergoing shoulder arthroplasty for failed nonoperative treatment of proximal humerus fracture have low implant survival and low patient-reported outcomes: 837 cases from the Danish Shoulder Arthroplasty Registry Inger MECHLENBURG 1-3, Sigrid RASMUSSEN 1, Ditte UNBEHAUN 1, Alexander AMUNDSEN 4, and Jeppe Vejlgaard RASMUSSEN 4 1 Department

of Orthopaedic Surgery, Aarhus University Hospital; 2 Department of Clinical Medicine, Aarhus University; 3 Department of Public Health, Aarhus University; 4 Department of Orthopaedic Surgery, Herlev University Hospital, Denmark Correspondence: inger.mechlenburg@clin.au.dk Submitted 2019-12-10. Accepted 2020-01-23.

Background and purpose — When nonoperative treatment of proximal humerus fracture (PHF) fails, shoulder arthroplasty may be indicated. We investigated risk factors for revision and evaluated patient-reported outcome 1 year after treatment with either stemmed hemiarthroplasty (SHA) or reverse total shoulder arthroplasty (RTSA) after previous nonoperative treatment of PHF sequelae. Patients and methods — Data were derived from the Danish Shoulder Arthroplasty Registry and included 837 shoulder arthroplasties performed for PHF sequelae between 2006 and 2015. Type of arthroplasty, sex, age, and surgery period were investigated as risk factors. The Western Ontario Osteoarthritis of the Shoulder index (WOOS) was used to evaluate patient-reported outcome (0–100, 0 indicates worst outcome). Cox regression and linear regression models were used in the statistical analyses. Results — 644 patients undergoing SHA and 127 patients undergoing RTSA were included. During a mean followup of 3.7 years, 48 (7%) SHA and 14 (11%) RTSA were revised. Men undergoing RTSA had a higher revision rate than men undergoing SHA (hazard ratio [HR] 6, 95% confidence interval [CI] 2–19). 454 (62%) patients returned a complete WOOS questionnaire. The mean WOOS score was 53 for SHA and 53 for RTSA. Patients who were 65 years or older had a better WOOS score than younger patients (mean difference 7, CI 1–12). Half of patients had WOOS scores below 50. Interpretation — Shoulder arthroplasty for PHF sequelae was associated with a high risk of revision and a poor patientreported outcome. Men treated with RTSA had a high risk of revision.

Displaced proximal humerus fractures (PHF) have been treated nonoperatively, with a locking plate osteosynthesis or with a stemmed hemiarthroplasty (SHA) (Launonen et al. 2019) for many years. More recently, randomized controlled trials have reported similar functional outcome between nonoperative treatment, locking plate osteosynthesis, or SHA for Neer 3 and 4-part fractures (Olerud et al. 2011a, b, Boons et al. 2012, Fjalestad and Hole 2014), but with significantly higher risk of complications and reoperations after operative treatment (Handoll and Brorson 2015, Launonen et al. 2015, Rangan et al. 2015, Beks et al. 2018). This relatively new information may lead to a higher number and proportion of nonoperative treatments in the future. Fracture sequelae after nonoperative or operative treatment of PHF such as malunion, nonunion, humeral head necrosis, degeneration or tear of the rotator cuff, and secondary glenohumeral osteoarthritis can lead to severe disability with pain, stiffness of the shoulder, and functional impairment (Greiner et al. 2014, Mansat and Bonnevialle 2015, Brorson et al. 2017). The treatment of fracture sequelae is challenging and the functional outcome after surgery is often disappointing (Kristensen et al. 2018). SHA has been used for decades, but the design depends on intact rotator cuff function and the longevity may be short due to glenoid wear. The reverse total shoulder arthroplasty (RTSA) was initially used in patients with cuff tear arthropathy, but the indication has expanded to other diagnoses, including PHF sequelae (Han et al. 2016). The design of the RTSA does not depend on rotator cuff function, although rotation and stability are improved with intact subscapularis and infraspinatus function. We investigated risk factors for revision and evaluated patient-reported outcome 1 year after treatment with either SHA or RTSA in previous nonoperative treatment of PHF sequelae.

320

Patients and methods Data were derived from the Danish Shoulder Arthroplasty Registry (DSR), established in January 2004 to monitor and improve shoulder arthroplasty surgery. The registry contains information on primary and revision arthroplasties. Reporting to the DSR has been mandatory for all Danish hospitals and private clinics since 2006 (Rasmussen et al. 2012). The surgeon reports data electronically and patient-reported outcomes are collected by mail 12 months (10–14) after surgery using the Western Ontario Osteoarthritis of the Shoulder index (WOOS) (Rasmussen et al. 2012). The completeness of patients registered in the DSR was 93% during the study period (Danish Shoulder Arthroplasty Registry 2017). PHF sequelae were defined as fractures reported with nonunion, malunion (including fractures reported together with osteoarthritis), or humeral head necrosis. We included all patients with PHF sequelae reported to the DSR from January 1, 2006 to December 31, 2015. Fractures reported together with previous osteosynthesis were excluded. Revision A revision was defined as removal or exchange of any component or the addition of a glenoid component. The revision was linked to the primary procedure with use of the unique civil registration number assigned to all Danish citizens. The civil registration number is also used when information regarding patients who die or emigrate is derived from the Danish Civil Registration System. Patient-reported outcome The WOOS was used as patient-reported outcome. The WOOS contains 19 questions categorized into 4 domains: physical symptoms, sport and work, lifestyle, and emotions. The patient-reported results are indicated on a visual analogue scale ranging from 0 to 100. The total score ranges from 0 to 1,900 (1,900 indicates worst outcome). To simplify the presentation of the patient-reported results, the raw scores were converted into percentages, where 100 is the best. The Danish version of the WOOS has been culturally adapted and validated for patients with glenohumeral osteoarthritis (Rasmussen et al. 2013). In case of revision, death or emigration within 1 year, the WOOS score was registered as missing. Statistics Descriptive statistics were used to report demographic data and follow-up time. The Kaplan–Meier method was used to illustrate the estimated unadjusted survival rates with 95% confidence interval (CI). The Cox regression model was used to determine the hazard ratios (HR) of revision with a CI. Arthroplasty type, age, sex, and surgery period were included in the multivariate model and the linear regression model. We used 2 age categories: younger than 65 years, and 65 years or

Acta Orthopaedica 2020; 91 (3): 319–325

older. This categorization was applied based on the Danish retirement age, due to an expected change in the patient’s activity level. We used 2 surgery periods, 2006–2010 and 2011–2015. Patient data contributed with individual risk time until revision, emigration, death, or until December 31, 2015, whichever came first. Test of proportional-hazards assumption was considered fulfilled. Although it violates the assumption of independence, patients with bilateral shoulder arthroplasty procedures were included in the survival analysis as if they were independent (Ranstam et al. 2011). A linear regression model was used to estimate the predicted mean difference in the WOOS score. Arthroplasty type, age, sex, and surgery period were included in the multivariate model. A plot of residuals versus predicted values, plots of residuals versus independent variables, and a normal probability plot of the residuals were used to check whether the assumptions of linearity, independence, constant variance, and normality of the residuals were fulfilled. Characteristics of patients responding or not responding to the WOOS questionnaire were compared using the chisquare test for categorical variables and Student’s t-test for continuous variables. The level of statistical significance was set at p < 0.05 and all p-tests were 2-tailed. The analyses were performed using STATA 15.0 (StataCorp LP, College Station, TX, USA). Ethics, funding, data sharing, and potential conflicts of interest According to Danish law, ethics committee approval was not required. No funding was obtained for this study. As part of the Data Use Agreement at the Danish Shoulder Arthroplasty Registry, authors are not allowed to provide raw data. Upon reasonable request, the corresponding author will provide statistical programming codes used to generate the results. No potential conflicts of interests are declared.

Results Demographics 837 patients were treated with shoulder arthroplasty for sequelae after a previous nonoperatively treated PHF; 644 underwent SHA and 127 underwent RTSA (Figure 1). 2% of patients had bilateral shoulder arthroplasty performed. Women accounted for 71% of arthroplasties. Mean age was 70 (SD 11) years, and 69% of patients were older than 65 years. The reasons for sequelae were nonunion (67%), malunion (27%), or humeral head necrosis (7%) (Table 1). Risk of revision The median follow-up time was 3.2 years (IQR 1.3–6.1). 71 (8%) of the shoulder arthroplasties were revised (Figure 2). The 1-, 5-, and 10-year cumulative arthroplasty survival rates with CI for women were 97% (95–98), 92% (89–94), and

Acta Orthopaedica 2020; 91 (3): 319–325

321

Frequency

Cumulative arthroplasty survival

Cumulative survival – female patients

100

0.9

0.8

0.7

0.6

SHA RTSA

2006

2009

2012

0.5

2015

Year

Figure 1. Number of hemiarthroplasties (SHA) and reverse shoulder arthroplasties (RTSA) due to failed nonoperative treatment of proximal humerus fracture registered in the Danish Shoulder Arthroplasty Registry, 2006–2015.

0.5 0

Years efter surgery

Figure 2. Cumulative survival for all types of arthroplasties, 2006–2015

SHA RTSA 2

Years efter surgery Cumulative survival – male patients 1

0.9

91% (87–93) for SHA, and 94% (86–98) and 92% (84–97) for RTSA. For men 1-, 5-, and 10-year cumulative survival rates with CI were 94% (92–98), 92% (87–96), and 82% (65– 92) for SHA, and 80% (62–90) and 76% (57–87) for RTSA, respectively (Figure 3). There were no statistically significant differences in risk of revision among arthroplasty type, age, sex, or surgery period (Table 2). However, sex had a significant impact on the result of arthroplasty. Men treated with

0.8

0.7

0.6

0.5 0

Years efter surgery

Figure 3. Cumulative survival for SHA (blue) and RTSA (red) in women (upper panel) and men (lower panel).

Table 1. Demographics of the study population presented by other type of arthroplasties (Others), stemmed hemiarthroplasty (SHA), and reverse shoulder arthroplasty (RTSA). Values are n (%) unless otherwise specified Factor Sex Women Men Age < 65 ≥ 65 Age, mean (SD) Indication for surgery Malunion b Nonunion Caput necrosis Missing Period of surgery 2006–2010 2011–2015 Revision WOOS (completed) WOOS mean (SD) a

Others SHA RTSA Total a (n = 65) (n = 644) (n = 127) (n = 837) 44 (68) 21 (32)

462 (72) 182 (28)

89 (70) 38 (30)

595 (71) 242 (29)

19 (29) 46 (71) 69 (11)

206 (32) 438 (68) 70 (11)

32 (25) 95 (75) 71 (9.4)

258 (31) 579 (69) 70 (11)

33 (51) 23 (35) 9 (14) –

134 (21) 470 (73) 37 (6) 3 (0)

55 (43) 61 (48) 10 (8) 1 (1)

222 (27) 555 (67) 56 (7) 4 (0)

25 (38) 40 (62) 9 (14) 37 (66) 38 (26)

393 (61) 251 (39) 48 (7) 347 (61) 47.4 (26)

14 (11) 113 (89) 14 (11) 70 (66) 47.5 (26)

432 (52) 404 (48) 71 (8) 454 (62) 46.6 (26)

For 1 shoulder arthroplasty, there was missing information on the type of arthroplasty with which the patient was treated. b Malunion includes fractures reported together with osteoarthritis. WOOS: Western Ontario Osteoarthritis of the Shoulder index.

Table 2. Univariate and multivariate Cox regression model for revision of stemmed hemiarthroplasty (SHA), reverse shoulder arthroplasty (RTSA), sex, age, and period of surgery, (n = 771, revisions = 62) Factor

Univariate Multivariate HR (CI) p-value HR (CI) p-value

Type of arthroplasty SHA 1 RTSA 2.0 (1.1–3.7) 0.02 Sex Women 1 Men 1.4 (0.8–2.4) 0.2 Age < 65 1 ≥ 65 0.7 (0.4–1.2) 0.2 Period of surgery 2006–2010 1 2011–2015 1.6 (0.9–2.6) 0.1 SHA + women RTSA + men

1 1.1 (0.4–2.7)

0.8

1 1.0 (0.5–1.9)

1.0

1 0.7 (0.4–1.2)

0.2

1 1.3 (0.7–2.3) 1 3.5 (1.0–12)

HR (CI): Hazard ratio (95% confidence interval). SHA: stemmed hemiarthroplasty, RTSA: reverse shoulder arthroplasty.

0.4 0.05

322

Acta Orthopaedica 2020; 91 (3): 319–325

Table 3. Multivariate Cox regression model for revision of women (n = 551, revisions = 40) and men (n = 220, revisions = 22) adjusted for age and period of surgery Type of arthroplasty SHA RTSA

Women Men HR (CI) p-value HR (CI) p-value 1 1.0. (0.4–2.4) 0.9

1 6.0 (1.9–19)

0.003

For abbreviations, see Table 2.

RTSA had a higher risk of revision than men treated with SHA (HR 6.0, CI 1.9–19) (Table 3). The cumulative survival rates for men were significantly higher for SHA compared with RTSA (Figure 3). Overall, the most common reasons for revision were dislocation (28%), rotator cuff problems (17%), other reasons (includes pain with no other reasons reported or malposition) (16%), and infection (11%) (Table 4). Patient-reported outcome 0.5% of the patients emigrated or were foreign citizens. 7% of patients died and 5% of patients underwent revision within the first year after surgery. Thus, the WOOS questionnaire was sent to 732 patients; 62% of patients returned a complete WOOS, 6% returned an incomplete questionnaire, and 32% did not respond. 1 year after shoulder arthroplasty, the mean WOOS score was 53 (SD 26) for the entire cohort, 53 (SD 26) for patients treated with SHA and 53 (SD 26) for patients treated with RTSA (Figure 4). 49% of the patients had a WOOS score below 50. There was no significant difference between the WOOS score of SHA and RTSA in the multivariate regression model (mean difference = –1.9, CI –9.2 to 5.5) (Table 5.) Patients 65 years or older had a better WOOS score compared with younger patients (mean difference = 6.6, CI 1.1–12), but

Table 5. Univariate and multivariate linear regression model for mean difference (MD) in Western Ontario Osteoarthritis of the Shoulder score, type of arthroplasty, sex, age, and period of surgery, (n = 417) Factor

Type of arthroplasty SHA 0 RTSA –0.1 (–6.8 to 6.5) 1.0 Sex Women 0 Men 1.7 (–3.9 to 7.4) 0.6 Age < 65 0 ≥ 65 5.8 (0.5 to 11) 0.03 Period of surgery 2006–2010 0 2011–2015 2.1 (–3.0 to 7.1) 0.4

Dislocation Loosening Glenoid wear Infection Fracture Technical failure Rotator cuff problems Other reasons b Missing Total a

23 (2.7) 3 (0.4) 5 (0.6) 9 (1.1) 5 (0.6) 8 (1.0) 14 (1.7) 13 (1.6) 1 (0.1) 81 (9.7)

28 4 6 11 6 10 17 16 1 100

0 –1.9 (–9.2 to 5.5) 0.6 0 3.5 (–2.3 to 9.3) 0.2 0 6.6 (1.1 to 12)

0.02

0 2.4 (–3.1 to 8.0) 0.4

the differences were not considered clinically relevant (Table 5). There was no statistically significant difference in type of arthroplasty, sex, age, and surgery period between patients who responded or did not respond to the WOOS (Table 6); this indicates that the patients responding to the WOOS in this study were not selected.

Discussion The main findings in this study were the poor patient-reported outcomes and the low implant survival rate, especially for men undergoing RTSA. We found a 5-year cumulative survival rate of 76% for men undergoing RTSA, which is less than we would usually accept, particularly because revision of RTSA is complex and

Table 4. Reasons for revision for all types of arthroplasties (All), other types of arthroplasties (Others), stemmed hemiarthroplasty (SHA), and reverse shoulder arthroplasty (RTSA). Values are n, (percentage of primary arthroplasties), and percentage of revisions Reasons All a for revision (n = 837)

Univariate Multivariate MD (CI) p-value MD (CI) p-value

Frequency 40

Others SHA RTSA (n = 65) (n = 644) (n = 127) 3 (5) 27 0 (0) 0 0 (0) 0 2 (3) 18 0 (0) 0 1 (2) 9 2 (3) 18 2 (3) 18 1 (2) 9 11 (17) 100

12 (1.9) 21 2 (0.3) 4 5 (0.8) 9 6 (0.9) 11 5 (0.8) 9 7 (1.1) 12 12 (1.9) 21 8 (1.2) 14 0 (0) 0 57 (8.9) 100

For 1 shoulder arthroplasty, there was missing information on the type of arthroplasty with which the patient was treated. b Other reasons includes pain with no other reasons reported.

8 (6.3) 62 1 (0.8) 8 0 (0) 0 1 (0.8) 8 0 (0) 0 0 (0) 0 0 (0) 0 3 (2.4) 23 0 (0) 0 13 (10.2) 100

100

WOOS-score

Figure 4. Distribution of WOOS scores at the 1-year follow-up for all patients.

Acta Orthopaedica 2020; 91 (3): 319–325

323

Table 6. Comparison of patients responding or not responding to the Western Ontario Osteoarthritis of the Shoulder index (WOOS). Values are n (%) unless otherwise specified Factor

Responders Non-responders (n = 454) (n = 278)

p-value

Type of arthroplasty 0.5 SHA 347 (76) 222 (80) RTSA 70 (15) 36 (13) Others 37 (8) 19 (7) Sex 0.7 Women 329 (72) 198 (71) Men 125 (28) 80 (29) Age 0.7 < 65 145 (32) 85 (31) ≥ 65 309 (68) 193 (69) Age mean (SD) 69.4 (11.0) 70.1 (11.4) 0.1 Indication for surgery a 0.4 Malunion b 128 (28) 66 (24) Nonunion 293 (65) 191 (69) Caput necrosis 32 (7) 20 (7) Period of surgery 0.2 2006–2008 146 (32) 76 (27) 2009–2011 139 (31) 100 (36) 2012–2015 169 (37) 102 (37) a Responders, n = 453. Non-responders, n = 277, stemmed hemiarthroplasty (SHA), and reverse shoulder arthroplasty (RTSA). b Malunion includes fractures reported together with osteoarthritis.

challenging due to the design of the RTSA and limited glenoid bone stock (Brorson et al. 2017, Holton et al. 2017). In our study, dislocation was the indication for revision in two-thirds of patients with revised RTSA. The reason for dislocations cannot be deduced from this registry study. It may be related to difficulties in placing the humeral component appropriately with correct tensioning of the deltoid muscle due to changed bone morphology. It may also be related to malposition or reabsorption of the tubercles and thus the function of the infraspinatus and subscapularis muscles. This could add imbalance to the reversed design. The survival rate of SHA seems more promising, but it is difficult to compare the 2 arthroplasty types directly. The RTSA may have been used for the most severe cases and in patients with a rotator cuff problem. Thus, the survival rate of SHA would probably not be as good if it had been used in the same patients. Our findings indicate that RTSA is not the easy solution for patients with PHFS and based on the high risk of revision in general and dislocation in particular we suggest that RTSA for PHFS is performed by experienced surgeons only and that surgeons who perform the operation focus their attention on technical details. Information concerning patient-reported outcome after shoulder arthroplasty for failed, nonoperative treatment of PHF is sparse. A retrospective study reported the mean Constant score in 42 patients being operated with a RTSA due to malunion of a PHF and found that the Constant score increased from 20 points (0–52 points) preoperatively to 55 points (21–83 points) 4 years postoperatively (Raiss et al. 2016). The

complication rate was 10%. Raiss et al. also reported the Constant score in 32 patients operated with an RTSA due to nonunion of a PHF and found that the Constant score increased from 14 points (2–35 points) to 47 points (6–75 points) 4 years postoperatively (Raiss et al. 2014). There were, however, complications in 41% of patients leading to revision of 28% of the arthroplasties. The Constant score and the WOOS are not directly comparable but the results from these studies seem similar to our WOOS score of 53. Kristensen et al. (2018) evaluated patient-reported outcome in patients with PHF initially treated with osteosynthesis and later treated with shoulder arthroplasty. The authors used data from the DSR and found a WOOS score of 46 (SD 25). As in our study there was no difference in WOOS between SHA and RTSA. Thus, based on data from DSR there seems to be no difference in WOOS between patients previously treated nonoperatively and patients who have had previous osteosynthesis. We found that patients aged above 65 years achieved a significantly better WOOS score compared with younger patients. However, the difference was small and may not be clinically relevant. Poor patient-reported outcomes have also been reported for younger patients with osteoarthritis (Rasmussen et al. 2014). The reason is unknown and cannot be deduced from this registry study, but it may be related to higher expectations and higher functional demands. Overall, the patients achieved poor patient-reported outcomes 1 year after shoulder arthroplasty and 49% of patients reported WOOS scores < 50, interpreted as a clinical failure. Methodological considerations Although there was no difference in type of arthroplasty, sex, age, and surgery period between patients who responded or who did not respond to the WOOS, the relatively low number of patients responding to the WOOS may have affected the results. The WOOS was developed and validated for patients with osteoarthritis and the validity of applying WOOS as an outcome for patients with PHF sequelae has not been investigated. Although the WOOS is used in the Danish and Swedish shoulder registries and from 2020 introduced in the Finnish and Norwegian shoulder registries, the WOOS is not widely used, which makes comparison with results from other countries difficult. Although the correlation between the WOOS and the Constant score and the Oxford Shoulder Score is high (Rasmussen et al. 2012), direct comparability is not possible. The size of the study population and the number of events was relatively low, especially since the regression models included several variables. Furthermore, the follow-up time was short and loosening will usually not appear within 1 year postoperatively. The indications for undergoing SHA and RTSA might have been different, introducing selection bias. In addition, there is risk of wrong coding when registering to the DSR. Moreover, unknown confounders may have influenced the results. The literature reports that comorbidities, smoking, body mass index, drug and alcohol abuse are frequent in

324

patients with fracture sequelae and that these factors affected the outcome of shoulder arthroplasty (Murray et al. 2011, Werner et al. 2015, Hernandez et al. 2017). Unfortunately, the DSR does not collect this type of information. Neither does the DSR collect information on preoperative shoulder pain and function. Finally, not all patients are treated with arthroplasty after PHF sequelae and there is most likely a selection bias towards more healthy, active, and demanding patients. An unknown proportion of patients will not be offered or will not accept shoulder arthroplasty, even if they have a poor outcome. Hence, it is a strength of our study that revision rates are supplemented with patient-reported outcome, which shows that half of patients reported high levels of pain 1 year after shoulder arthroplasty. In conclusion, we found low survival rates in patients undergoing SHA and RTSA for PHF sequelae. In particular, men undergoing RTSA had a high risk of revision. The patientreported outcomes were poor with half of patients reporting WOOS score below 50 with no difference between SHA and RTSA or between men and women. The findings indicate that RTSA is not the easy solution for PHF sequelae. It is a technically demanding operation and we suggest that caution is warranted when considering treating PHF sequelae with RTSA in men.

All authors took part in the conception and design of the study and in the interpretation of the results. SR, DU, AA, and JVR prepared data from the Danish Shoulder Arthroplasty Registry. SR and DU performed the statistical analyses. All authors participated in the preparation of the manuscript. IM incorporated input from all authors and was responsible for writing the manuscript. Acta thanks Tore Fjalestad and Kaspar Tootsi for help with peer review of this study.

Beks R B, Ochen Y, Frima H, Smeeing D P J, van der Meijden O, Timmers T K, van der Velde D, van Heijl M, Leenen L P H, Groenwold R H H, Houwert R M. Operative versus nonoperative treatment of proximal humeral fractures: a systematic review, meta-analysis, and comparison of observational studies and randomized controlled trials. J Shoulder Elbow Surg 2018; 27(8): 1526-34. doi: 10.1016/j.jse.2018.03.009. Boons H W, Goosen J H, van Grinsven S, van Susante J L, van Loon C J. Hemiarthroplasty for humeral four-part fractures for patients 65 years and older: a randomized controlled trial. Clin Orthop Relat Res 2012; 470(12): 3483-91. doi: 10.1007/s11999-012-2531-0. Brorson S, Salomonsson B, Jensen S L, Fenstad A M, Demir Y, Rasmussen J V. Revision after shoulder replacement for acute fracture of the proximal humerus. Acta Orthop 2017; 88(4): 446-50. doi: 10.1080/17453674.2017.1307032. Danish Shoulder Arthroplasty Registry. Annual report from the Danish Shoulder Arthroplasty Registry. https://www.sundhed.dk/content/cms/3/4703_ dsr-aarsrapport-2019-til-offentliggoerelse.pdf Fjalestad T, Hole M O. Displaced proximal humeral fractures: operative versus non-operative treatment—a 2-year extension of a randomized controlled trial. Eur J Orthop Surg Traumatol 2014; 24(7): 1067-73. doi: 10.1007/s00590-013-1403-y.

Acta Orthopaedica 2020; 91 (3): 319–325

Greiner S, Uschok S, Herrmann S, Gwinner C, Perka C, Scheibel M. The metaphyseal bone defect predicts outcome in reverse shoulder arthroplasty for proximal humerus fracture sequelae. Arch Orthop Trauma Surg 2014; 134(6): 755-64. doi: 10.1007/s00402-014-1980-1. Han R J, Sing D C, Feeley B T, Ma C B, Zhang A L. Proximal humerus fragility fractures: recent trends in nonoperative and operative treatment in the Medicare population. J Shoulder Elbow Surg 2016; 25(2): 256-61. doi: 10.1016/j.jse.2015.07.015. Handoll H H, Brorson S. Interventions for treating proximal humeral fractures in adults. Cochrane Database Syst Rev 2015; (11): Cd000434. doi: 10.1002/14651858.CD000434.pub4. Hernandez N M, Chalmers B P, Wagner E R, Sperling J W, Cofield R H, Sanchez-Sotelo J. Revision to reverse total shoulder arthroplasty restores stability for patients with unstable shoulder prostheses. Clin Orthop Relat Res 2017; 475(11): 2716-22. doi: 10.1007/s11999-017-5429-z. Holton J, Yousri T, Arealis G, Levy O. The role of reverse shoulder arthroplasty in management of proximal humerus fractures with fracture sequelae: a systematic review of the literature. Orthop Rev (Pavia) 2017; 9(1): 6977. doi: 10.4081/or.2017.6977. Kristensen M R, Rasmussen J V, Elmengaard B, Jensen S L, Olsen B S, Brorson S. High risk for revision after shoulder arthroplasty for failed osteosynthesis of proximal humeral fractures. Acta Orthop 2018; 89(3): 345-50. doi: 10.1080/17453674.2018.1450207. Launonen A P, Lepola V, Flinkkila T, Laitinen M, Paavola M, Malmivaara A. Treatment of proximal humerus fractures in the elderly: a systemic review of 409 patients. Acta Orthop 2015; 86(3): 280-5. doi: 10.3109/17453674.2014.999299. Launonen A P, Fjalestad T, Laitinen M K, Lahdeoja T, Ekholm C, Wagle T, Mattila V M. Nordic Innovative Trials to Evaluate osteoPorotic Fractures (NITEP) Collaboration: the Nordic DeltaCon Trial protocol—non-operative treatment versus reversed total shoulder arthroplasty in patients 65 years of age and older with a displaced proximal humerus fracture: a prospective, randomised controlled trial. BMJ Open 2019; 9(1): e024916. doi: 10.1136/bmjopen-2018-024916. Mansat P, Bonnevialle N. Treatment of fracture sequelae of the proximal humerus: anatomical vs reverse shoulder prosthesis. Int Orthop 2015; 39(2): 349-54. doi: 10.1007/s00264-014-2651-0. Murray I R, Amin A K, White T O, Robinson C M. Proximal humeral fractures: current concepts in classification, treatment and outcomes. J Bone Joint Surg Br 2011; 93(1): 1-11. doi: 10.1302/0301-620x.93b1.25702. Olerud P, Ahrengart L, Ponzer S, Saving J, Tidermark J. Hemiarthroplasty versus nonoperative treatment of displaced 4-part proximal humeral fractures in elderly patients: a randomized controlled trial. J Shoulder Elbow Surg 2011a; 20(7): 1025-33. doi: 10.1016/j.jse.2011.04.016. Olerud P, Ahrengart L, Ponzer S, Saving J, Tidermark J. Internal fixation versus nonoperative treatment of displaced 3-part proximal humeral fractures in elderly patients: a randomized controlled trial. J Shoulder Elbow Surg 2011b; 20(5): 747-55. doi: 10.1016/j.jse.2010.12.018. Raiss P, Edwards T B, da Silva M R, Bruckner T, Loew M, Walch G. Reverse shoulder arthroplasty for the treatment of nonunions of the surgical neck of the proximal part of the humerus (type-3 fracture sequelae). J Bone Joint Surg Am 2014; 96(24): 2070-6. doi: 10.2106/jbjs.N.00405. Raiss P, Edwards T B, Collin P, Bruckner T, Zeifang F, Loew M, Boileau P, Walch G. Reverse shoulder arthroplasty for malunions of the proximal part of the humerus (type-4 fracture sequelae). J Bone Joint Surg Am 2016; 98(11): 893-9. doi: 10.2106/jbjs.15.00506. Rangan A, Handoll H, Brealey S, Jefferson L, Keding A, Martin B C, Goodchild L, Chuang L H, Hewitt C, Torgerson D. Surgical vs nonsurgical treatment of adults with displaced fractures of the proximal humerus: the PROFHER randomized clinical trial. JAMA 2015; 313(10): 1037-47. doi: 10.1001/jama.2015.1629. Ranstam J, Karrholm J, Pulkkinen P, Makela K, Espehaug B, Pedersen A B, Mehnert F, Furnes O. Statistical analysis of arthroplasty data, I: Introduction and background. Acta Orthop 2011; 82(3): 253-7. doi: 10.3109/17453674.2011.588862.

Acta Orthopaedica 2020; 91 (3): 319â&#x20AC;&#x201C;325

Rasmussen J V, Jakobsen J, Brorson S, Olsen B S. The Danish Shoulder Arthroplasty Registry: clinical outcome and short-term survival of 2,137 primary shoulder replacements. Acta Orthop 2012; 83(2): 171-3. doi: 10.3109/17453674.2012.665327. Rasmussen J V, Jakobsen J, Olsen B S, Brorson S. Translation and validation of the Western Ontario Osteoarthritis of the Shoulder (WOOS) index: the Danish version. Patient Relat Outcome Meas 2013; 4: 49-54. doi: 10.2147/ prom.S50976.

325

Rasmussen J V, Polk A, Sorensen A K, Olsen B S, Brorson S. Outcome, revision rate and indication for revision following resurfacing hemiarthroplasty for osteoarthritis of the shoulder: 837 operations reported to the Danish Shoulder Arthroplasty Registry. Bone Joint J 2014; 96-B(4): 519-25. doi: 10.1302/0301-620x.96b4.31850. Werner B C, Burrus M T, Begho I, Gwathmey F W, Brockmeier S F. Early revision within 1 year after shoulder arthroplasty: patient factors and etiology. J Shoulder Elbow Surg 2015; 24(12): e323-30. doi: 10.1016/j.jse.2015.05.035.

326

Acta Orthopaedica 2020; 91 (3): 326–330

Percutaneous needle fasciotomy in Dupuytren contracture: a registerbased, observational cohort study on complications in 3,331 treated fingers in 2,257 patients Laura Houstrup THERKELSEN 1, Simon Toftgaard SKOV 1, Malene LAURSEN 1, and Jeppe LANGE 2 1 Elective Surgery Centre, Silkeborg Regional Hospital, Denmark; 2 Department of Orthopaedic Surgery, Horsens Regional Hospital, Denmark Correspondence: lather@rm.dk Submitted 2019-09-16. Accepted 2020-01-21.

Background and purpose — Percutaneous needle fasciotomy (PNF) is a minimally invasive treatment option for Dupuytren contracture, which has gained momentum worldwide in recent years. However, evidence regarding safety and severe complications associated with PNF is sparse. Thus, we evaluated safety of a specific PNF method in the largest cohort reported in literature. Patients and methods — This is a single-center, register-based, observational study on PNF treatments between 2007 and 2015. The study cohort was identified by the Danish National Patient Registry, and diagnosis codes and procedure codes were used to identify potential severe postoperative complications such as: tendon rupture, nerve damage, infection, amputation, and reflex dystrophy for all index treatments. The Danish National Prescription Registry was used to identify non-hospital-treated infections. All index treatments and postoperative complications were verified by review of medical records. Results — 2,257 patients received PNF treatment of 3,331 treated finger rays. Median follow-up was 7.2 years (interquartile range: 4.9–9.5 years, range 1–12 years). 4 patients sustained flexor tendon rupture. 1 patient had digital nerve damage. 2 patients had an infection treated in hospital, while 31 patients received antibiotics in the primary sector for an infection or based on suspicion of an infection after PNF. None of the infections required surgical intervention. No finger amputations or ipsilateral upper limb reflex dystrophy cases were registered in relation to the procedure. Interpretation — Percutaneous needle fasciotomy for Dupuytren contracture is a safe procedure with a low rate of severe postoperative complications when a specific PNF method is applied.

Dupuytren contracture is a benign fibroproliferative disease that affects the aponeurotic fibers in the palm, causing permanent flexion contracture of the fingers. This extension deficit may imply disability in activities of daily living for the affected patients (Wilburn et al. 2013). There is no cure for Dupuytren contracture, but several symptomatic treatment options exist. However, there is no consensus as to the optimal treatment in individual cases (McMillan et al. 2017). Percutaneous needle fasciotomy (PNF) is performed as a minimally invasive surgical procedure under local anesthesia in an outpatient setting. A fine syringe needle is used to perforate the Dupuytren cords repeatedly until the finger can be extended. The PNF technique was reintroduced in the late 1970s by a group of French rheumatologists, and has been performed in Denmark at our center since 2006, according to the method of Lermusiaux and Debeyre (1979). 2-year results after the initial introduction of PNF treatment were promising (Rahr et al. 2011). Several randomized controlled trials have shown that PNF is not inferior to collagenase clostridium histolyticum (CCH) injection regarding complication rate and clinical outcome (Scherman et al. 2016, Skov et al. 2017, Strömberg et al. 2016, 2018). Even though PNF has shown reliable clinical results, evidence regarding complications associated with PNF is sparse. A potential risk of iatrogenic nerve and tendon damage during the procedure has been stated as a specific concern. We evaluated the safety and severe complications following the PNF method applied in our institution.

Acta Orthopaedica 2020; 91 (3): 326–330 327

Patients and methods Study design This was a single-center, register-based follow-up study with data validation by review of medical records. Study cohort The study cohort was established by identifying all patients receiving PNF treatment for Dupuytren contracture at Silkeborg Regional Hospital, Denmark, between January 1, 2007, and December 31, 2015 in the Danish National Patient Registry (DNPR) (data extraction specified in Table 1, see Supplementary data). 2007 was chosen as it was the first full year of PNF treatment in our department. The index procedure was defined as the patient’s first PNF treatment during the study period. Thereby, any individual patient is only represented once. Percutaneous needle fasciotomy procedure: the Silkeborg method Percutaneous needle fasciotomy (PNF) is performed in an outpatient setting. Following the standard disinfection procedure and sterile draping, 0.1–0.2 mL of 1% lidocaine with epinephrine is injected into chosen sites of the Dupuytren cord, using a small, 25-gauge syringe needle. The anesthetics is injected only intradermally to maintain full sensibility of the PNFtreated finger during the entire procedure. This specific procedural monitoring is believed essential to minimize the risk of nerve affection. If the patient experiences any kind of nerve affection following the injection of the local anesthetics before the procedure begins, or any nerve sensation during the procedure, the procedure is stopped, and patients with unreleased contractures are offered a repeated procedure another day. This concept of nerve protection was meticulously adhered to in all treated patients in this cohort, and we believe it of utmost importance when performing PNF. The technique, in which the cord is then weakened, is a combination of repeated needle-tip perforations into the cord, and cautious pendulum cutting of the cord, with a 25-gauge needle at a slow pace, and simultaneously passively stretching the finger to rupture the Dupuytren cord. To protect against nerve affection also during the procedure a small 25-gauge needle is always used. Several needles may be used, if dull. When the Dupuytren cord is sufficiently weakened, the finger can be stretched manually and any potential residual cord strings can be slowly, but forcefully ruptured. Additional local anesthetic may be required to achieve a final manipulation at the end of the treatment, but only after the perforation/cutting has definitely ended. The procedure is typically applied at 1–5 different sites, including Dupuytren cords affecting both metacarpophalangeal and proximal interphalangeal joints, depending on the severity of the contracture and the number of finger rays to be treated (Rahr et al. 2011, Skov et al. 2017).

The PNF procedures were performed by 4 orthopedic surgeons, 1 PNF-trained resident doctor, and 1 PNF-trained registered nurse. Data collection All identified procedures were verified by review of medical records to ensure that the index procedures were PNF. Simultaneously, baseline information on treated hand, finger, and joint affection, and any special perioperative circumstances, was collected from the medical records. Patients who died within 1 year after the index procedure were excluded due to the insufficient follow-up time. Likewise, we excluded patients residing abroad due to inability to follow-up. After establishing the study cohort, information on succeeding diagnosis codes and procedure codes relative to the index procedure was extracted from the DNPR to locate potential postoperative complications for a retrospective review in the medical records (data-extraction strategy specified in Tables 2 and 3, see Supplementary data). The dataset was time limited, assuming that complications such as flexor tendon ruptures, nerve damage, and amputations would present within 1 year, infections within 1 month, lesions within 3 months, and reflex dystrophy within 5 years from index procedure. The dataset was compiled on September 20, 2018. Presumably, all operations and severe complications are registered in the DNPR, because reimbursement for both government hospitals and private clinics relies on registration (Lynge et al. 2011). However, the primary sector is not required to report to DNPR. In order to avoid this information bias regarding the postoperative infection rate, we extracted data on anatomical therapeutic chemical (ATC) codes from the Danish National Prescription Registry, which holds information on prescribed antibiotics (specified in Supplementary data, Table 4). Thereby, all infections were most likely accounted for, covering the full spectrum from minor superficial infections to severe deep infections. The indications for prescribed antibiotics were audited nationwide in the medical records. If the information was unavailable in the hospital medical records, we contacted the patient’s general practitioner. Follow-up time was determined as time from index procedure to either date of data extract or date of death. Time of death and information on country of residence were extracted from the Danish Civil Registration System (Pedersen 2011). Statistics Data extracts from the National Health Administrative Registries were available on the Research Engine administrated by the Research Services of the Danish Health Data Authority. (Data management and statistical analysis were performed in the open source RStudio; R Foundation for Statistical Computing, Vienna, Austria, version 1.1.463.)

328

Acta Orthopaedica 2020; 91 (3): 326–330

Table 5. Distribution of treated hand, finger, and joint affection

Total study cohort n = 2,257 Complications following PNF verified by reviewing medical records n = 38

Factor Number Affected hand right left NS Affected finger little finger ring finger middle finger index finger thumb NS Affected joint metacarpophalangeal joint proximal interphalangeal joint distal interphalangeal joint NS

1,723 1,592 16 1,445 1,265 408 92 88 33

Flexor tendon rupture (n = 4): – double flexor tendon ruptur requiring further intervention, 3 – ruptur of the profound tendon, 1

Digital nerve damage n=1

Antibiotics (n = 33): – infection treated in hospital with antibiotic and wound care, 2 – antibiotic potentially prescribed to treat or prevent an infection, 31

Complications. n: number of patients.

2,696 1,713 144 164

NS: Not specified in medical record. The sum of treated hands and fingers (n = 3,331) exceeds the number of patients (n = 2,257) as some patients had several fingers treated in same session. Likewise, the sum of affected joints (n = 4,717) exceeds the number of treated fingers as more joints on the same finger were affected.

Registration, funding, and conflicts of interest Study approvals were obtained from the Danish Patient Safety Authority (3-3013-2261/1) for access to review of medical records and from the Danish Data Protection Agency (1-1602-711-17). According to Danish law, no further approvals were required. Data from medical records were collected under current regional data protection legislature, with acceptance from the Head of Department of the relevant departments or the patient’s general practitioner. The study was conducted in accordance with the Law of Processing Personal Data and the Danish Health Act. Funding was received from Rieck-Andersens Familiefond Aarhus, the Danish Rheumatism Association, and Elisabeth & Karl-Ejnar Nis-Hanssens Mindelegat. All authors declare no conflict of interest.

Results 2,308 patients were identified in the DNPR. All index PNF treatments were verified by review of medical records; 28 patients did not receive PNF treatment, 21 patients died within the first year after index procedure, and 2 patients had residence abroad. The study cohort included 2,257 patients with 3,331 treated finger rays. 1,777 (79%) were men. Mean age at index was 66 years (18–96). Median follow-up was 7.2 years (IQR 4.9–9.5, range 1–12 years). The distributions of treated hand, finger, and joint affection are shown in Table 5.

Complications (Figure) 110 of the 2,257 treated patients were identified to have a complication potentially related to the PNF treatment. 4 patients had flexor tendon rupture after PNF, double flexor tendon rupture (both superficial and deep flexor tendons) occurred in 3 cases, while 1 patient had only the profound flexor tendon ruptured. In the 3 cases of double tendon injury, the patients had tenorrhaphies subsequently. The double flexor tendon ruptures were registered 7, 8, and 9 days postoperative, respectively. The patient who ruptured the deep flexor tendon reported that it occurred 3 days after PNF and was treated nonoperatively. 1 patient sustained digital nerve damage, while 5 patients had temporary nerve symptoms that resolved within 2 weeks after the index procedure. None of the patients with transient or persisting nerve symptoms received further treatment, and no later contacts were registered regarding nerve laceration. 2 cases of infections were identified via the DNPR nationwide data. These patients were seen in the outpatient clinic, and received oral antibiotic treatment and simple wound care. 75 patients were registered in the Danish National Prescription Registry as receiving antibiotics within 1 month after PNF (including the above-mentioned 2 patients with postoperative infections). 42 patients received antibiotic treatment for other reasons, and 18 had antibiotics prescribed in general practice to treat or to prevent an infection following the PNF treatment. In 13 cases, there was no documentation available in the hospital records or in the general practitioner’s records specifying the indication for the antibiotic treatment. None of the infections required intravenously administered antibiotics or surgical treatment. 4 patients contacted our department after the index procedure with a skin rupture needing the bandage replacing but without signs of infection. In treatment of 70 finger rays (in 63 patients), the PNF procedure was discontinued due to paresthesia or accidental skin lesion. Unreleased contractures were offered a new procedure another day, but no further wound complication or nerve affection in relation to these cases were documented in the medical records. No complications led to amputation. However, 3 patients had the digit of the index procedure amputated within the first year due to insufficient effect of the PNF treatment. This was

Acta Orthopaedica 2020; 91 (3): 326â&#x20AC;&#x201C;330 329

not defined as a complication. No ipsilateral upper limb reflex dystrophy cases were registered in relation to the procedure.

Discussion In this register-based follow-up study on 2,257 patients, we found low rates of severe postoperative complications including flexor tendon rupture (n = 4), nerve damage (n = 1), and infection (n = 33) following our PNF method. One of the key findings in this study is the extremely low prevalence of nerve damage identified. We acknowledge the risk of information bias due to insufficient registration, or that patients with a nerve affection are simply not seen by a doctor for various reasons and the consequences this would have for our possibility of identifying these patients. However, we truly believe our result to be genuine regarding the low prevalence of nerve affection, when the described PNF technique is applied. This is supported by the results in our randomized controlled trial (Skov et al. 2017), and also in other randomized controlled trials involving a similar PNF procedure (Scherman et al. 2016, StrĂśmberg et al. 2016). It is important to highlight a methodological aspect of the PNF procedures performed in this cohort: PNF is performed using a fine needle (25G), at a slow pace, and under low doses of local anesthetic allowing the full sensibility of the finger to be monitored during the procedure, to minimize the risk of digital nerve damage. We believe this is reflected in the low rate of nerve injury (Skov et al. 2017). In addition, the procedures were performed by only 6 individuals during the study period, thus a certain expertise is represented, and they were all specially trained in the PNF method. In 2018, a review stated that PNF has significantly lower complication rates compared with open fasciectomy and CCH injection (Elzinga and Morhart 2018). In 2017, a systematic review of 113 studies assessing the incidence of complications associated with different treatment options for Dupuytren contracture found a pooled complication rate of 19% after PNF (Krefter et al. 2017). They included both major and minor complications, and found that the highest prevalence of nerve and vessel lesions occurred following open surgery. This review did not consider the severity of the complications, and raised the question as to whether a skin tear, which can heal without further intervention, should be considered a complication. Our study is based on Danish administrative registries, merely investigating complications requiring contact with the healthcare system. Consequently, minor skin tears will not be registered. A single-center retrospective study on open surgeries for Dupuytren contracture performed between 1956 and 2006 at Erlangen University Hospital, Germany, was published in 2007 (Loos et al. 2007). The study included 2,919 patients, corresponding to the number of patients in our study (2,257). Nerve injuries were observed in 108 cases (3.7%), tendon injuries occurred in 4 cases (0.2%), skin necrosis was documented in

2.6% of all operations, and wound infection was observed in 94 patients (3.2%). The reported tendon injury risk after open surgery is similar to our findings after PNF (0.18%). But we found remarkably lower rates of nerve injuries (0.04%) and infections (1.5%). Also, we did not identify any cases of skin necrosis. However, the complication rate reported by Loos et al. may be underestimated as the authors looked only at data from their own local medical records. In contrast, we included nationwide data on potential PNF complications registered by other hospitals/departments. Comparing different treatment modalities for Dupuytren contracture based on available literature is a cumbersome task due to variations in follow-up time, different definitions of complications, and noncomparable study cohorts. However, by reviewing the literature (Table 6, see Supplementary data) in combination with our results, PNF appears to be a safe treatment modality for Dupuytren contracture. Dupuytren contracture is relatively common, and it is noticeable that the number of studies on PNF is limited, and many of them are small case series. In 13 of our cases, in which antibiotics were prescribed, there was no documented indication available in the hospital records or in the general practitionerâ&#x20AC;&#x2122;s records. In those cases, and with the intention of avoiding underestimation of the risk of infection following PNF treatment, we assumed that antibiotic prescription was associated with PNF treatment, leaving the total infection rate at 1.5%. Complex regional pain syndrome (CRPS) is known to be a complication associated with Dupuytren contracture interventions (Krefter et al. 2017). 1 patient in our cohort was registered with CRPS (ICD-10: DM89.0) almost 4 years after index procedure, but this was found to be related to a distal radius fracture and not the PNF treatment. We examined CRPS cases up to 5 years after the index procedure and included differential diagnoses, but found no other CRPS cases. Of course, there is a risk of patients being misdiagnosed or not diagnosed at all. However, if a patient in Denmark experiences severe symptoms resembling CRPS, we believe they will most likely be referred to a hospital clinic for an evaluation, and thereby become adequately registered. Study limitations and strengths The major strengths of this study are the size and complete nationwide follow-up, combined with data verification by manual review of medical records. However, there are limitations to this study. (1) This is a retrospective register-based study, and the complications following PNF were identified through the DNPR and the Danish National Prescription Registry. The data were therefore not prospectively targeted for this study. (2) Complications were identified based on a data extract of diagnosis codes, procedure codes, and ATC codes. Thus, complications reported with codes not included in our dataset will not be identified. (3) There is a risk of underestimating complications following PNF if they are inadequately

330

registered, or if patients with potential complication are not seen in hospital clinics or referred by the general practitioner, e.g., nerve affection with low demand in daily activities. Regarding infections, we extracted data covering all infections including patients treated in the primary sector. (4) In contrast to other studies, we did not report minor complications such as skin ruptures. Neither recurrence nor subsequently repeated treatment were regarded as complications. However, when we compare detailed results regarding severe complications, our cumulative complication rate of 1.7% seems similar to those other studies. Moreover, it is noteworthy that no skin lesions led to further wound complication, than could be managed by oral antibiotics and simple wound care. To our knowledge, this study represents the largest cohort with complete nationwide follow-up regarding safety and severe complications after PNF treatment in the literature. Percutaneous needle fasciotomy for Dupuytren contracture is a safe procedure with a low rate of severe postoperative complications when an appropriate PNF technique is applied (Skov et al. 2017). The importance of this study is further emphasized, as PNF is less expensive than all treatment alternatives for Dupuytren contracture, and it has shown to be costeffective (Baltzer and Binhammer 2013, Chen et al. 2011). Supplementary data Tables 1–4 and 6 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/ 17453674.2020.1726057

STS and JL designed the study. LHT collected and processed the data. All authors interpreted the results. LHT drafted the manuscript, and it was revised by all authors. The authors would like to thank Therkel Bisgaard, Per Søndergaard, Frederik Vang Hansen, Linda Bohmann Vinther Lauridsen, Morten Søndergaard, and Steen Olesen who have performed the PNF procedures in Silkeborg. Acta thanks Isam Atroshi, Peter Scherman and Maria Wilcke for help with peer review of this study.

Acta Orthopaedica 2020; 91 (3): 326–330

Baltzer H, Binhammer P A. Cost-effectiveness in the management of Dupuytren’s contracture. Bone Joint J 2013; 95-B: 1094-1100. doi:10.1302/0301620X.95B8.31822 Chen N C, Shauver M J, Chung K C. Cost-effectiveness of open partial fasciectomy, needle aponeurotomy, and collagenase injection for Dupuytren contracture. J Hand Surg Am 2011; 36A: 1826-34.e32. doi:10.1016/j. jhsa.2011.08.004 Elzinga K E, Morhart M J. Needle aponeurotomy for Dupuytren disease. Hand Clin 2018; 34: 331-44. https://doi.org/10.1016/j.hcl.2018.03.003 Krefter C, Marks M, Hensler S, Herren D B, Calcagni M. Complications after treating Dupuytren’s disease: a systematic literature review. Hand Surg Rehabil 2017; 36: 322-29. http://dx.doi.org/10.1016/j.hansur.2017.07.00 Lermusiaux J L, Debeyre N. Le traitement médical de la maladie de Dupuytren. L’Actualité Rhumatol Paris Expans Sci Publ 1979: 238-49. Loos B, Puschkin V, Horch R E. 50 years experience with Dupuytren’s contracture in the Erlangen University Hospital: a retrospective analysis of 2919 operated hands from 1956 to 2006. BMC Musculoskelet Disord 2007; 8: 60. doi:10.1186/1471-2474-8-60 Lynge E, Sandegaard J L, Rebolj M. The Danish National Patient Register. Scand J Public Health 2011; 39(7): 30-3. doi: 10.1177/1403494811401482 McMillan C, Yeung C, Binhammer P. Variation in treatment recommendations for Dupuytren disease. J Hand Surg Am 2017; 42: 963-70.e6. http:// dx.doi.org/10.1016/j.jhsa.2017.08.023 Pedersen C B. The Danish Civil Registration System. Scand J Public Health 2011; 39(7): 22-5. doi: 10.1177/1403494810387965 Rahr L, Søndergaard P, Bisgaard T, Baad-Hansen T. Percutaneous needle fasciotomy for primary Dupuytren’s contracture. J Hand Surg Eur 2011; 36(7): 548-52. doi: 10.1177/1753193411407245 Scherman P, Jenmalm P, Dahlin L B. One-year results of needle fasciotomy and collagenase injection in treatment of Dupuytren’s contracture: a twocentre prospective randomized clinical trial. J Hand Surg Eur 2016; 41(6): 577-82. doi: 10.1177/1753193415617385 Skov S T, Bisgaard T, Søndergaard P, Lange J. Injectable collagenase versus percutaneous needle fasciotomy for Dupuytren contracture in proximal interphalangeal joints: a randomized controlled trial. J Hand Surg Am 2017; 42(5): 321-8.e3. http://dx.doi.org/10.1016/j.jhsa.2017.03.003 Strömberg J, Ibsen-Sörensen A, Fridén J. Comparison of treatment outcome after collagenase and needle fasciotomy for Dupuytren contracture: a randomized, single-blinded, clinical trial with a 1-year follow-up. J Hand Surg Am 2016; 41(9): 873e880. http://dx.doi.org/10.1016/j.jhsa.2016.06.014 Strömberg J, Ibsen-Sörensen A, Fridén J. Percutaneous needle fasciotomy versus collagenase treatment for Dupuytren contracture: a randomized controlled trial with a two-year follow-up. J Bone Joint Surg Am 2018; 100(13): 1079-86. http://dx.doi.org/10.2106/JBJS.17.01128 Wilburn J, McKenna S P, Perry-Hinsley D, Bayat A. The impact of Dupuytren disease on patient activity and quality of life. J Hand Surg 2013; 38A: 1209-14. http://dx.doi.org/10.1016/j.jhsa.2013.03.036

Acta Orthopaedica 2020; 91 (3): 331–335

331

Reduced incidence and economic cost of hardware removal after ankle fracture surgery: a 20-year nationwide registry study Nikke PARTIO 1, Tuomas T HUTTUNEN 2,3, Heikki M MÄENPÄÄ 1, and Ville M MATTILA 1,3 1 Department

of Orthopaedics and Traumatology, Tampere University Hospital, Tampere; 2 Department of Emergency, Anesthesia and Pain Medicine, Tampere University Hospital; 3 Faculty of Medicine and Health Technology, Tampere University, Finland Correspondence: Nikke.partio@tuni.fi Submitted 2019-10-13. Accepted 2020-02-04.

Background and purpose — Open reduction and internal fixation (ORIF) is a treatment method for unstable ankle fractures. During recent years, scientific evidence has shed light on surgical indications as well as on hardware removal. We assessed the incidence and trends of hardware removal procedures following ORIF of ankle fractures. Patients and methods — The study covered all patients 18 years of age and older who had an ankle fracture treated with ORIF in Finland between the years 1997 and 2016. Patient data were obtained from the Finnish National Hospital Discharge Register. Results — 68,865 patients had an ankle fracture treated with ORIF in Finland during the 20-year study period between 1997 and 2016. A hardware removal procedure was performed on 27% of patients (n = 18,648). The incidence of hardware removal procedures after ankle fracture decreased from 31 (95% CI 29–32) per 100,000 person-years in the highest year 2001 (n = 1,247) to 13 (CI 12–14) per 100,000 person-years in 2016 (n = 593). Moreover, the proportion and number of removal operations performed within the first 3 months also decreased. The costs of removal procedures decreased from approximately €994,000 in 2001 to €472,600 in 2016. Interpretation — Removal of hardware after ankle surgery (ORIF) is a common operation with substantial costs. However, the incidence and cost of removals decreased during the study period, with a particular decrease in hardware removal operations within 3 months.

It is estimated that approximately 40% of all ankle fractures require surgical management, most commonly open reduction and internal fixation (ORIF) (Jensen et al. 1998). According to Kannus et al. (2016), an earlier increasing trend in the incidence of ankle fractures in Finland has steadied. Ankle fractures are associated with high costs related not only to the operation and subsequent hospitalization, but also to the duration of occupational disability (Stull et al. 2017). To reduce the costs of occupational disability, an early return to previous activities and the avoidance of secondary operations is crucial. In a recent study, Fenelon et al. (2019) found that 13% of patients who had had ankle fracture surgery in Ireland underwent hardware removal. The most common reason was planned removal (6%) followed by symptomatic hardware (6%), and infection (0.5%). The reasons for hardware removal include pain and soft tissue irritation, deep late infection, metal allergy or toxicity, hardware migration, metal failure, and secondary fracture (Bostman and Pihlajamaki 1996). The hardware removal rates reported by previous studies have varied between 12% and 80% (Richards et al. 1992, Sanderson et al. 1992, Bostman and Pihlajamaki 1996). While the removal of hardware after ankle fracture surgery is often a straightforward procedure, complication rates are still as high as 10–20% (Sanderson et al. 1992, Kasai et al. 2019). Patient satisfaction and symptomatic relief following ankle fracture surgery is also controversial (Jamil et al. 2008, Williams et al. 2012). Postoperative complications include infections, impaired wound healing, refractures, tissue and nerve damage, postoperative bleeding, and incomplete removal (Sanderson et al. 1992). We determined the incidence and trends in Finland of ankle fracture surgery and hardware removal after ORIF of ankle fractures on a national level. Additionally, we estimated the costs and economic burden of the removal and surgery itself.

332

Patients and methods Study design and data sources The Finnish National Hospital Discharge Register (FNHDR) was founded in 1967. It provides data on age, sex, domicile of the subject, duration of hospital stays, primary and secondary diagnosis, and surgical procedures performed during the hospital stay. The data collected by the FNHDR are mandatory for all hospitals, including private, public, and other institutions. The validity of the FNHDR has been proven to be excellent regarding both the coverage and accuracy of the database. In 1996, the Nordic Medico-Statistical Committee (NOMESCO) published the first printed edition of the NOMESCO Classification of Surgical Procedures (NCSP) followed by the Finnish translation (NCSP-F) in 1997. Study population We included all adult patients (≥ 18 years) with a diagnosis of ankle fracture (ICD-10 code) and who underwent ankle fracture surgery (ORIF) between January 1, 1997 and December 31, 2016. The procedural codes (according to the Finnish version of the NOMESCO classification) for the ankle fracture included NHJ10 (internal fixation of fracture of ankle using plate, wire, rod, cerclage, or pin) and NHU20 (hardware removal). The primary outcome was the incidence of operative treatment of ankle fractures and hardware removal. Since the FNHDR does not include laterality of the operation, only the patient’s 1st ankle fracture ORIF operation performed during the study period was included in the analysis. A secondary outcome was the cost of hardware removal as determined by diagnosis-related group-based (DRG-based) hospital payment system pricing. In Finland as well as in most Organization for Economic Co-operation and Development (OECD) countries, DRG-based hospital payment systems are being used. The basic idea of DRG-based hospital payment systems is that all patients treated by a hospital are classified into a limited number of DRGs that are supposed to be clinically meaningful and relatively homogeneous in their patterns of resource consumption. Statistics The trends for operatively treated ankle fracture and hardware removal (per 100,000 person-years) were based on the entire adult population of Finland rather than cohort or sample-based estimates. Mid-year population size was estimated by taking the geometric mean of year-end population sizes of consecutive years. Incidence density rate and operations total rate are presented with the 95% confidence interval (CI). Data were analyzed with R 3.5.3 (R Foundation for Statistical Computing, Vienna, Austria).

Acta Orthopaedica 2020; 91 (3): 331–335

Ethics, registration, funding, and potential conflicts of interest The study was approved by Finland: National Institute of Health and Welfare (THL): THL/89/5.05.00/2012. The study was funded by Finland’s government research and development foundation. Funding sources were not involved in study design, collection, analysis, interpretation, or completion. The authors declare no conflicts of interest regarding this study.

Results 68,865 adult patients (51% women) had an ankle fracture surgically treated with ORIF in Finland during the 20-year study period between 1997 and 2016. The mean age at the time of the 1st surgery in 1997 was 52 years in women and 44 years in men. In 2016, the corresponding ages were 57 years in women and 47 years in men. The total incidence of ankle fracture surgery was 81 (CI 78–83) per 100,000 person-years in 1997 (3,218 operations) and 74 (CI 71–76) per 100,000 person-years in 2016 (3,276 operations) (Figure 1). In men, the incidence was 87 (CI 83–91) per 100,000 person-years in 1997 and 72 (CI 68–75) per 100,000 person-years in 2016. In women, the corresponding figures were 80 (CI 76–84) per 100,000 person-years in 1997 and 82 (CI 79–86) per 100,000 person-years in 2016. During the 20-year study period, a total of 18,648 (27%) hardware removal procedures (52% women) were performed after primary ankle fracture surgery. The mean age at the time of the first hardware removal in 1997 was 53 years in women and 45 years in men. In 2016, the corresponding ages were 56 years in women and 57 years in men. The incidence of hardware removal procedures after ankle fracture surgery decreased from 31 (CI 29–32) per 100,000 person-years in the highest year, 2001 (n = 1,247), to 13 (CI 12–14) per 100,000 person-years in 2016 (n = 593) (Figure). In men, the incidence was 32 (CI 29–34) per 100,000 person-years in 2001 and 12 (CI 11–13) per 100,000 person-years in 2016. In women, the corresponding figures were 30 (CI 27–32) per 100,000 person-years in 2001 and 15 (CI 13–16) per 100,000 person-years in 2016. The mean time between ankle fracture surgery and hardware removal procedure was 14 months. 43% of the hardware removals occurred during 0–3 months after surgery, 8% during 3–6 months, 15% during 7–12 months, 21% during 13–24 months, 5% during 25–36 months, and 8% at over 36 months (Table). Hardware removals were conducted within 3 months for 538 (52%) of the 1,029 patients who underwent hardware removal in the highest year, 2007 (Table). Thereafter, the removal rate remained steady until 2012 when the removal rate decreased markedly from 49% to 26% in 2016. According to Finnish DRG-based hospital payment pricing, the direct costs for one ankle fracture surgery were €2,881 per patient in 2016. This amounted to an annual direct cost of

Acta Orthopaedica 2020; 91 (3): 331–335

333

Table 1. Frequencies and ratios of under 3 months hardware removals by year

Incidence/100,000 100

Year

ORIF Total hardware removal Hardware removal 0–3 months

0 1997

2000

2003

2006

2009

2012

2015

Year

Incidence of ORIF, total hardware removals and hardware removals under 3 months in Finland between 1997 and 2016.

ankle fracture surgery of approximately €9,500,000 in 2016. The DRG-based costs for one hardware removal are €797. Thus, the costs of removal procedures in 2016 were approximately €472,600, whereas the corresponding removal costs in 2001 were almost €994,000.

Discussion This study is the 1st nationwide study that shows a large number of ankle fracture surgeries (n = 68,865) as well as hardware removal procedures (27%; n = 18,648) during a 20-year study period. This study confirms the previous findings that the incidence of ankle fracture surgery is higher in young men and in older women (Kannus et al. 2016, Juto et al. 2018). The decreasing incidence of ankle fracture surgery might have been caused by the increasing knowledge of nonoperatively treated Weber-B type fractures, which can be either stable or unstable (Kortekangas et al. 2019). Stable ankle fractures can be treated non-surgically and account for about one half of all ankle fractures (Pakarinen et al. 2011, Van Schie-Van der Weert et al. 2012). Previous studies have reported 12–80% of hardware removal (Richards et al. 1992; Sanderson et al. 1992, Bostman and Pihlajamaki 1996). This variation might be due to cultural and treatment policy differences and may also be attributed to the different lengths of observation periods or analytic methods between studies. The recent lowest hardware removal rates (13%) were reported by Fenelon et al. (2019). However, the authors suggest that their results were an underestimation because of the retrospective nature of their study that could have led to a larger loss to follow-up. Additionally, they showed that the majority of removals were due to symptomatic hardware. In Another retrospective study reported that

Hardware removals 0–3 months Total

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Ratio (%)

239 350 68 343 941 37 372 1,032 36 269 971 28 528 1,247 42 511 1,195 43 403 1,019 40 364 820 44 550 1,151 48 475 1,028 46 538 1,029 52 512 1,034 50 534 1,005 53 446 924 48 472 954 50 456 941 49 378 945 40 267 807 33 195 661 30 152 593 26

about 17% of patients underwent hardware removal over a 3-year period (Naumann et al. 2016). Our study showed that 27% of patients underwent hardware removal. A notable number, 8% (n = 1,516), of hardware removal procedures were performed after 3 years. Therefore, we believe that previous studies have underestimated the hardware removal rate due to shorter follow-up. Moreover, we assume that most of the removals performed after more than 3 years were due to symptomatic hardware. Williams et al. (2018) reported improvement in function following ankle implant removal, but their sample size was small (43 patients) and there was no control group. Patient symptomatic relief after hardware removal is still controversial. Approximately 10% of all ankle fractures have concomitant syndesmotic injury. In 15–23% of operatively treated ankle fractures, a syndesmotic disruption necessitates surgical repair with a syndesmotic screw (Jensen et al. 1998, Egol et al. 2010). However, the need to remove this screw remains controversial. In his literature review in 2011 (including 7 studies between 2000 and 2010), Schepers (2011) reported that there is no need to routinely remove the syndesmotic screws. In a recent systematic review, Dingemans et al. (2016) also suggest that the current literature does not support the routine removal of syndesmotic screws. Furthermore, the complication rate for routine syndesmotic screw removal is about 20% (Schepers et al. 2011). Fenelon et al. (2019) showed in their study that 6% of all patients underwent planned hardware removal and that the majority of procedures were for the removal of a syndesmosis screw after a median time of 3 months. Our register study could not separate all hardware removals from only syndesmotic screw removal, but we assume that most removals

334

within 3 months were of syndesmotic screws. This assumption is supported by the fact that these procedures decreased markedly between 2011 and 2016, most likely due to the changed evidence suggesting syndesmotic screws need not be removed (Figure) (Schepers 2011). Hardware removal causes significant costs to patients, hospitals, and societies through the consumption of healthcare resources and absence from work. The total economic cost of removal is difficult to evaluate due to the multifactorial nature and financing of the Finnish healthcare system. However, Fenelon et al. (2019) reported the cost of hardware removal to be €1,113 per patient in Ireland. A study by Lalli et al. (2015) from the United States found the average cost of syndesmosis screw removal to be $3,579 ($287 to $9,981). In this figure, they included anesthesia, operating room and recovery room fees, as well as pharmacy, laboratory, and supplies costs. In our study, the DRG-based hospital payment system was used to reflect the cost of hardware removal (€794), but this sum does not include the costs of drug prescriptions, missed work, or loss of income. According to our results, the annual cost of hardware removal decreased due to a decrease in removal rates (n = 1,247 in 2001 and n = 593 in 2016). In other words, in 2001 the cost of hardware removals was approximately €994,000 but in 2016 it was €472,600. Bioabsorbable fixation materials have been suggested as a solution to the hardware removal problem. Good outcomes with bioabsorbable screws in ankle fractures without secondary surgery have been reported (Partio et al. 1992). In a metaanalysis including 4 studies comparing bioabsorbable and metallic screws, Wang et al. (2013) showed that all metallic screws were routinely removed 6–8 weeks after primary operation while only 2 symptomatic patients (3%) in the absorbable screw group needed re-surgery. In their meta-analysis, van der Eng et al. (2015) found no significant differences in the incidence of complications between patients treated with a polylactide acid (PLA)/polylevolactic acid (PLLA) screw and patients treated with a metallic syndesmotic screw. In the past, rapidly degrading polyglycolide acid (PGA) screws were associated with delayed inflammatory reactions, foreign body reaction, formation of a sinus, tract, or fistula, and osteolysis. However, there is no evidence that the PLA/PLLA combination has such problems (Bostman et al. 2005, Wang et al. 2013). The key advantage of using biomaterials is that hardware removal becomes unnecessary. Our study has some limitations. Importantly, the study is register-based and bilateral operations on the same patient cannot be differentiated from the registry data. Additionally, the reason for removal remained unclear because the reason for removal is poorly recorded using ICD coding. Also, patientrelated risk factors, such as smoking or alcohol consumption, are not recorded in the NHDR. A strength of the study is the accuracy and coverage of the NHDR database, which is collected from all Finnish hospitals. Indeed, the national cover-

Acta Orthopaedica 2020; 91 (3): 331–335

age of the NHDR provided a large population of surgically treated ankle fractures for a 20-year period. In summary, this 20-year nationwide study showed that 27% of patients underwent a hardware removal procedure after ankle fracture surgery. The number of routine syndesmosis screw removals seemed to decrease, resulting in lower economic costs. A substantial number of hardware removal procedures are being performed after the 3-year period.

All the authors made a substantial contribution to the conception of the study. NP planned the study and wrote the protocol. VM obtained permission from the Finnish National Hospital Discharge Register (FNHDR) and drafted and revised the article. TH planned the study, revised the protocol, and revised the article. HM planned the study and revised the article. All the authors critically revised the draft prepared by NP. The authors would like to thank Topias Koukkula, research assistant, for statistical groundwork. Acta thanks Anders Henricson and Dagmar Isabella Vos for help with peer review of this study.

Bostman O, Pihlajamaki H. Routine implant removal after fracture surgery: a potentially reducible consumer of hospital resources in trauma units. J Trauma 1996; 41(5): 846-9. Bostman O M, Laitinen O M, Tynninen O, Salminen S T, Pihlajamaki H K. Tissue restoration after resorption of polyglycolide and poly-laevolactic acid screws. J Bone Joint Surg Br 2005; 87(11): 1575-80. doi: 87-B/11/1575. Dingemans S A, S Rammelt, White T O, Goslings J C, Schepers T. Should syndesmotic screws be removed after surgical fixation of unstable ankle fractures? A systematic review. Bone Joint J 2016; 98-B(11): 1497-1504. doi: 98-B/11/1497. Egol K A, Pahk B, Walsh M, Tejwani N C, Davidovitch R I, Koval K J. Outcome after unstable ankle fracture: effect of syndesmotic stabilization. J Orthop Trauma 2010; 24(1): 7-11. doi: 10.1097/BOT.0b013e3181b1542c. Fenelon C, Murphy E P, Galbraith J G, Kearns S R. The burden of hardware removal in ankle fractures: how common is it, why do we do it and what is the cost? A ten-year review. Foot Ankle Surg 2019; 25(4): 546-9. doi: 10.1016/j.fas.2018.05.006. Jamil W, Allami M, Choudhury M Z, C Mann, Bagga T, Roberts A. Do orthopaedic surgeons need a policy on the removal of metalwork? A descriptive national survey of practicing surgeons in the United Kingdom. Injury 2008; 39(3): 362-7. doi: 10.1016/j.injury.2007.10.028. . Jensen Steen L, Andresen B K, Mencke S, Nielsen P T. Epidemiology of ankle fractures: a prospective population-based study of 212 cases in Aalborg, Denmark. Acta Orthop 1998; 69(1): 48-50. doi: 10.3109/17453679809002356. Juto H, Nilsson H, Morberg P. Epidemiology of adult ankle fractures: 1756 cases identified in Norrbotten County during 2009–2013 and classified according to AO/OTA. BMC Musculoskeletal Disord 2018; 1 (1): 441-x. doi: 10.1186/s12891-018-2326-x. Kannus P, Niemi S, Parkkari J, Sievanen H. Declining incidence of fallinduced ankle fractures in elderly adults: Finnish statistics between 1970 and 2014. Arch Orthop Trauma Surg 2016; 136(9): 1243-6. doi: 10.1007/ s00402-016-2524-7. Kasai T, Matsumoto T, Iga T, Tanaka S. Complications of implant removal in ankle fractures. J Orthop 2019; 16(3): 191-4. doi: 10.1016/j.jor.2019.02.017. Kortekangas T, Haapasalo H, Flinkkila T, Ohtonen P, Nortunen S, Laine H J, Jarvinen T L, Pakarinen H. Three week versus six week immobilisation for stable Weber B type ankle fractures: randomised, multicentre, non-

Acta Orthopaedica 2020; 91 (3): 331â&#x20AC;&#x201C;335

inferiority clinical trial. BMJ (Clinical Research ed.) 2019; 364: k5432. doi: 10.1136/bmj.k5432. Lalli T A, Matthews L J, Hanselman A E, Hubbard D F, Bramer M A, Santrock R D. Economic impact of syndesmosis hardware removal. Foot 2015; 25(3): 131-3. doi: 10.1016/j.foot201503001. Naumann M G, Sigurdsen U, Utvag S E, Stavem K. Incidence and risk factors for removal of an internal fixation following surgery for ankle fracture: a retrospective cohort study of 997 patients. Injury 2016; 47(8): 1783-8. doi: 10.1016/j.injury.2016.05.011. Pakarinen H J, Flinkkil T E, Ohtonen P P, Ristiniemi J Y. Stability criteria for nonoperative ankle fracture management. Foot Ankle Int 2011; 32(2): 141-7. doi: 10.3113/FAI.2011.0141. Partio E K, Bostman O, Hirvensalo E, Vainionpaa S, Vihtonen K, Patiala H, Tormala P, Rokkanen P. Self-reinforced absorbable screws in the fixation of displaced ankle fractures: a prospective clinical study of 152 patients. J Orthop Trauma 1992; 6(2): 209-15. doi: 10.1097/00005131-199206000-00013. Richards R H, Palmer J D, Clarke N M. Observations on removal of metal implants. Injury 1992; 23(1): 25-8. doi: //doi-orglibproxytun ifi/101016/0020-1383(92)90120-H. Sanderson P L, Ryan W, Turner P G. Complications of metalwork removal. Injury 1992; 23(1): 29-30. doi: 0020-1383(92)90121-8. Schepers T. To retain or remove the syndesmotic screw: a review of literature. Arch Orthop Trauma Surg 2011; 131(7): 879-83. doi: 10.1007/s00402010-1225-x.

335

Schepers T, Van Lieshout E M, de Vries M R, Van der Elst M. Complications of syndesmotic screw removal. Foot Ankle Int 2011; 32(11): 1040-4. doi: 10.3113/FAI.20111040. Stull J D, Bhat S B, Kane J M, Raikin S M. Economic burden of inpatient admission of ankle fractures. Foot Ankle Int 2017; 38(9): 997-1004. doi: 10.1177/1071100717709576. van der Eng D M, Schep N W, Schepers T. Bioabsorbable versus metallic screw fixation for tibiofibular syndesmotic ruptures: a meta-analysis. J Foot Ankle Surg 2015; 54(4): 657-62. doi: 101053/j.jfas.2015.03.014. Van Schie-Van der Weert E M, Van Lieshout E M, De Vries M R, Van der Elst M, Schepers T. Determinants of outcome in operatively and non-operatively treated Weber-B ankle fractures. Arch Orthop Trauma Surg 2012; 132(2): 257-63. doi: 10.1007/s00402-011-1397-z. Wang C, Ma X, Wang X, Huang J, Zhang C, Chen L. Internal fixation of distal tibiofibular syndesmotic injuries: a systematic review with metaanalysis. Int Orthop 2013; 37(9): 1755-63. doi: 10.1007/s00264-0131999-x. Williams A A, Witten D M, Duester R, Chou L B. The benefits of implant removal from the foot and ankle. J Bone Joint Surg Am 2012; 94(14): 1316-20. doi: 10.2106/JBJS.J.01756. Williams B R, McCreary D L, Chau M, Cunningham B P, Pena F, Swiontkowski M F. Functional Outcomes of symptomatic implant removal following ankle fracture open reduction and internal fixation. Foot Ankle Int 2018; 39(6): 674-80. doi: 10.1177/1071100718757719.

336

Acta Orthopaedica 2020; 91 (3): 336–340

Mid-term results of hindfoot arthrodesis with a retrograde intra medullary nail in 24 patients with diabetic Charcot neuroarthropathy Mehmet ERSIN, Mehmet DEMIREL, Mehmet CHODZA, Fuat BILGILI, and Onder Ismet KILIÇOGLU

Department of Orthopaedics and Traumatology, İstanbul School of Medicine, İstanbul University, İstanbul, Turkey Correspondence: onder.kilicoglu@gmail.com Submitted 2019-09-16. Accepted 2020-02-11.

Background and purpose — Hindfoot arthrodesis using retrograde intramedullary nailing assumes a critical role in limb salvage for patients with diabetic Charcot neuro arthropathy (CN). However, this procedure is compelling and fraught with complications in diabetic patients. We report the mid-term clinical and radiological outcomes of retrograde intramedullary nailing for severe foot and ankle deformity in patients with diabetic CN. Patients and methods — Hindfoot arthrodesis was performed using a retrograde intramedullary nail in 24 patients (15 females) with diabetic Charcot foot. The mean age of the patients was 62 years (33–82); the mean follow-up was 45 months (24–70). The primary outcomes were rates of fusion, limb salvage, and complications. Results — The overall fusion rate was 23/24, and none of the patients needed amputation. The rate of superficial wound infection was 4/24, and no deep infection or osteomyelitis was observed postoperatively. Interpretation — For selected cases of diabetic CN with severe foot and ankle deformity, hindfoot arthrodesis using a retrograde intramedullary nail seems to be a good technique in achieving fusion, limb salvage, and avoidance of complications.

Acute Charcot neuroarthropathy (CN) of the foot and ankle is a consequence of the combined neuropathic process including sensory, motor, and autonomic peripheral nerves, which is clinically characterized by obvious swelling, bone destruction, and final healing with severe bone deformity (Pinzur and Noonan 2005). Bone deformities in CN may cause several problems ranging from an ulcer to osteomyelitis and eventually amputation (Alvarez et al. 1994, Aktaş et al. 2016). Although any clinical condition that gives rise to sensory or autonomic neuropathy can result in this debilitating disorder, diabetes mellitus (DM) is the leading cause of CN (van der Ven et al. 2009). Many authors have highlighted the salvage role of hindfoot arthrodesis in nonbraceable, severe ankle and hindfoot deformities in patients with CN (Alvarez et al. 1994, Bennett et al. 2005, Pinzur and Noonan 2005, Pelton and Hofer 2006). Various methods, comprising intramedullary nailing (Pinzur and Noonan 2005), crossed compression screws, blade plate (Alvarez et al. 1994), and external fixation (Russotti et al. 1988) have been employed for hindfoot arthrodesis. Nonetheless, retrograde intramedullary nailing has been a universal method of yielding stability and generating a plantigrade, stable foot for CN (Pinzur and Kelikian 1997, Pinzur and Noonan 2005). Conversely, the general consensus in the literature is that in patients with diabetic CN, hindfoot and ankle fusions are compelling and fraught with infectious and noninfectious complications due to their neuropathy, diabetes, poor bone quality, and the high mechanical loadings of internal fixation implants (Pinzur and Kelikian 1997, Perlman and Thordarson 1999, Mendicino et al. 2004, Chahal et al. 2006, ter Gunne and Cohen 2009, Wukich et al. 2011, Caravaggi et al. 2012, Myers et al. 2012).

Acta Orthopaedica 2020; 91 (3): 336–340 337

We present the mid-term clinical and radiological outcomes of retrograde intramedullary nailing for severe foot and ankle deformity in patients with diabetic CN.

Patients and methods 26 consecutive patients with diabetic CN who underwent unilateral hindfoot arthrodesis using a retrograde intramedullary nail between 2009 and 2015 by a single surgeon were reviewed retrospectively. After excluding 2 patients due to lack of medical records, the remaining 24 were included in the present study. 2 different retrograde intramedullary nails were used, including the expert HAN (Synthes AG, Bettlach, Switzerland) in 23 patients and the Trigen Hindfoot Fusion Nail (Smith & Nephew, Memphis, TN, USA) in 1 patient. After institutional review board approval was obtained, all the patients enrolled in the study were recalled for final follow-up clinical and radiological evaluations. 24 patients (mean age, 62 years [33–82]; 15 females) were analyzed retrospectively based on the institution medical records including radiology, operative notes, information on demographic characteristics, discharge reports, progress notes, and final follow-up with radiographical and physical examination. The mean follow-up was 45 months (24–70). The average weight was 90 kg (65–120), and BMI was 33 (24–42). According to BMI, 18 patients were obese, 4 of whom were morbidly obese (BMI > 40). The underlying pathology was type 2 DM in 20 patients and type 1 in 4. The average duration of DM at the time of the surgery was 19 years (10–32), and all patients had uncontrolled DM with a mean HbA1c of 91 mmol/mol (10.5%). The preoperative comorbidities, as a well-established predictor of increased risk for surgical complications, were documented and graded according to the ASA system; all patients were grade III. The data regarding comorbidities consisted of history of smoking (n = 4), peripheral vascular disorder (n = 1), coronary artery disorder (n = 4), chronic renal failure (n = 2), morbid obesity (n = 4), and chronic obstructive pulmonary disease (n = 5). Furthermore, prior surgeries and previous forefoot amputation on the involved foot and ankle were documented as well. The primary indications for hindfoot arthrodesis were bone deformity accompanied by nonhealing ulcers or severe instability. Our preoperative management protocol included an ulcer-free approach and thorough glycemic regulation. To obtain sufficient fixation, all patients were typically operated in the late stage of coalescence phase (stage II) or consolidation phase (stage III) according to the modified Eichenholtz classification (Figure 1) (Rosenbaum and DiPreta 2015). Multilayer compression bandaging was applied in patients with lymphedema until the operation. Additionally, 5 patients underwent an off-loading regimen with a non-weight-bearing circular cast due to infected ulcers before the operation. Of

these, 2 patients were monitored in hospital with vacuumassisted therapy, local debridement, and antibiotic administration. All the patients were ulcer-free at the time of the surgery. The primary outcomes of the study were rates of fusion, limb salvage, and complication. Fusion was assessed radiographically and clinically in all cases. Solid fusion was identified as an osseous bridging across the arthrodesis site—at least 3 cortices in the serial radiographs. Recurrent deformity and an ability to bear weight on the affected foot with or without support were investigated in the physical examination. Intraand postoperative complications were recorded. Functional status and quality of life were additional secondary outcomes, which were measured using the AOFAS score including 40 points for pain, 50 points for function, and 10 points for alignment (Kitaoka et al. 1994), the Mazur Ankle Arthrodesis Score (Mazur et al. 1979), and the SF-36 Health Survey Questionnaire (Ware et al. 1998). The AOFAS scale was determined preoperatively and at the final follow-up examination; the Mazur ankle arthrodesis score and the SF-36 were conducted at the final follow-up. Moreover, the ambulatory status of patients was evaluated on admission and at the final follow-up examination and divided into 3 categories: ambulation with a wheelchair (nonambulation), ambulation with gait assistance devices (walker or crutches), and independent ambulation. Surgical technique The patients were positioned in either lateral decubitus or a supine position on a radiolucent operating table under general or spinal anesthesia. The C-arm was positioned to obtain imaging of subtalar and tibiotalar joints in both anteroposterior and lateral views. A pneumatic thigh tourniquet was routinely applied during the whole operation. Surgery started with a transfibular approach to the ankle. An osteotomy of the lateral malleolus, 2–4 cm proximal to the joint line, was performed using a sagittal saw with a second proximal shortening cut of at least 1 cm. In the preparation of articular surfaces, the distal articular surface of the tibia was resected using an oscillating saw in lateral to medial direction. The superior articular surface of the talus (if present) was resected parallel to the distal tibial cut while holding the foot in plantigrade position. Then, all the cartilage, necrotic or sclerotic bone was removed from the subtalar or tibiotalar joint using curettes until viable subchondral bone was achieved. Furthermore, to enhance the fusion, multiple holes were drilled in the subchondral bone. In 9 patients with necrosis or collapse of the talus, remnants of the talar body were resected completely to achieve a stable tibiocalcaneal contact. In 3 patients, Chopart’s joint fusion was performed using cannulated screws in addition to hindfoot fusion. After preparation of the surfaces, the ankle was positioned in neutral dorsiflexion–plantar flexion, and the foot was then realigned in neutral to 10° of valgus at the heel and mild pos-

338

Figure 1. A case of severe CN with talar collapse.

Acta Orthopaedica 2020; 91 (3): 336–340

Figure 2. Solid fusion after retrograde intramedullary nailing using the expert HAN.

terior displacement of the calcaneus on the tibia in all cases. In some patients, the realignment was maintained temporarily with Kirschner wires. In the later stage, the expert HAN in 23 patients and Trigen Hindfoot Fusion Nail in 1 patient were implanted according to their surgical technique guide. In tibiotalocalcaneal arthrodesis, the entry point was located at the intersection of lines along the center of the tibial canal in the lateral view and the lateral column of the calcaneus in the anteroposterior view. However, for tibiocalcaneal arthrodesis, the entry point was shifted slightly to a more posterior location in the calcaneus relative to the typical entry point. In all patients who underwent the expert HAN, 1 spiral blade and 1 locking screw were first inserted from the posterior aspect into the calcaneus, followed by manual compression using a hammer. In the patient who underwent the Trigen Hindfoot Fusion Nail, 2 locking screws were used. Local autogenous bone graft was inserted in 3 cases. Allograft was not used. In 19 cases, the system was locked with 2 static locking screws from the lateral into the tibial diaphysis, and talar locking was not routinely used. At the final step of the operation, in order to enhance the stability, the lateral half of the fibula osteotomized in the initial stage was fixed with 2 cortical screws as a live biological plate. As an additional procedure, Achilles tendon lengthening was needed for only 1 patient with a severe contracture. Finally, anatomical foot alignment was confirmed clinically and with the image intensifier. Postoperatively, a non-weight bearing below-knee circular cast was used for a minimum of 3 months. Then, full weightbearing with a ROM walker was allowed in all patients for at least 6 months postoperatively. Healing enhancement methods were not used in this case series. Ethics, funding, and potential conflicts of interest Institutional review board approval was obtained (registration number 2019-0306). No funding for this study was received.

Demographic data and preoperative comorbidities of the 24 patients Median age at time of surgery Male/female sex Mean follow-up, months Average weight BMI Average duration of DM, years Mean HbA1c Smoking, n Peripheral vascular disorder, n Coronary artery disorder, n Chronic renal failure, n Morbid obesity, n Chronic obstructive pulmonary disease, n

62 (33–82) 9/15 45 (24–70) 90 (65–120) 33 (24–42) 19 (10–32) 10.5 % 4 1 4 2 4 5

DM: diabetus mellitus; BMI: body mass index.

Each author certifies that he or she has no commercial associations that might pose a conflict of interest.

Results The demographic data and preoperative comorbidities of all the patients are summarized in the Table. Radiographic solid fusion was obtained with retrograde intramedullary nailing in 23 of 24 patients (Figure 2 and 3), with an average time to fusion of 10 months (6–14). In the remaining patient who underwent tibiotalocalcaneal arthrodesis there was no radiographic evidence of solid fusion at follow-up, but revision was not considered because the patient remained asymptomatic. All the patients were able to bear weight on the affected foot with or without support; physical examination revealed no recurrent deformity. No patient underwent amputation. During postoperative follow-up, four patients with superficial infection (4/24) were treated successfully with antibiotic and local debridement; none of the patients developed deep soft tissue

Acta Orthopaedica 2020; 91 (3): 336–340 339

Chahal et al. 2006, Wukich et al. 2011, Myers et al. 2012). Furthermore, according to our literature review, there has been little research (Pinzur and Kelikian 1997, Mendicino et al. 2004, Pinzur and Noonan 2005, Caravaggi et al. 2006, Pelton et al. 2006, Caravaggi et al. 2012, Chraim et al. 2018) to review the results of retrograde hindfoot arthrodesis nailing in diabetic patients with severe foot and ankle deformity. Therefore, the present study focused specifically on a certain group of patients with diabetic Charcot foot and reflected the results of intramedullary nailing for this challenging combination of disorders. To provide additional evidence of feasibility and effectiveness of retrograde hindfoot arthrodesis in such patients, the primary outcomes of our study included fusion, limb salvage, and complication rates, concernFigure 3. Solid tibiocalcaneal fusion using the Trigen Hindfoot Fusion Nail. ing which the existing literature presents differing results. A study reported fusion infection or osteomyelitis. In addition, there was no need for in 19 patients and a fusion time of 20 months at a followrevision surgery; however, the spiral blade in 1 patient and up of 12–31 months in 20 diabetic patients with CN, despite the calcaneal locking screw in 2 patients were removed due to 1 amputation (Pinzur and Kelikian 1997). Nevertheless, the authors stated that half of the patients, who needed talectomy retrocalcaneal bursitis. In the evaluation of functional status, the mean AOFAS to obtain acceptable tibiocalcaneal alignment, sustained sevscore increased from 40 (35–45) before surgery to 68 (42–86) eral complications such as ulcers and wound infections. A after surgery. The mean Mazur ankle arthrodesis score was 64 recent study (Chraim et al. 2018) referred to a high fusion rate, (13–88) at the final follow-up. In terms of SF-36, the mean 16 of 19 diabetic Charcot feet healed but with complications physical component score (PCS) was 70 (5–100), and the in 9 patients, and 3 patients progressed to amputation due to mean mental component score (MCS) was 76 (20–96) at the persisting infection and osteomyelitis. In our opinion, glycemic regulation before surgery may be final evaluation. Regarding the ambulatory status of patients, on admission 2 another important factor in preventing postoperative compatients could ambulate independently without an assistance plications in our case series since it has been elucidated that device; 11 patients required gait assistance devices (walker or increasing levels of Hgb A1c are closely associated with postcrutches), and 11 patients were unable to walk. At the final operative infections after hindfoot and/or ankle arthrodesis in follow-up visit, except for 1 patient who was non-ambulatory, diabetic population (Myers et al. 2012). Although we failed all remaining patients could ambulate with a rocker-bottom to attain the desired level of Hgb A1c in each patient postshoe. Of these, half were able to walk independently and the operatively, all blood glucose levels returned to normal preother half with the aid of a gait device. operatively in all patients with readjustment of anti-diabetic medication. Furthermore, considering patients’ high burden of comorbidities, substantive functional limitations according to ASA, and poor long-term glucose control, it may be difficult Discussion to obtain an adequate level of HbA1c. As an alternative approach to amputation, arthrodesis assumes We recommend utilizing second-generation IM nails as a critical role in limb salvage for Charcot patients (Pinzur and used in our study, which offer specific ankle- and foot-locking Kelikian 1997, Pinzur and Noonan 2005, van der Ven et al. options as well as a compression mode, to provide sufficient 2009). Although retrograde intramedullary nailing is a widely stability and axial compression on arthrodesis sites, thus proaccepted method of hindfoot and ankle fusion to yield sta- moting fusion (Klos et al. 2009, Yakacki et al. 2011). Conbility and generate a plantigrade and braceable foot (Pinzur versely, using such nail designs for hindfoot arthrodesis may and Kelikian 1997, Pinzur and Noonan 2005, Wukich et al. need advanced technical skills. 2011), the general consensus in the literature is that this proAlso, it is necessary to discuss the functional status and cedure is compelling but fraught with complications in dia- quality of life. The patients’ preoperative functional status and betic patients (Mendicino et al. 2004, Caravaggi et al. 2006, quality of life were obviously impaired, as reflected by the

340

AOFAS and SF-36 PCS and MCS. The mean AOFAS score was improved from 40 (35–45) preoperatively to 68 (42–86) postoperatively; the mean Mazur ankle arthrodesis score was 64 (13–88) at the final follow-up. The SF-36 PCS and MCS were 70 and 76 respectively at the final examination. Considering the AOFAS improvement following arthrodesis, hindfoot arthrodesis provided a significant amelioration in patients’ functional status and satisfaction. Some limitations should be mentioned when interpreting the findings of our study. The main limitations were the retrospective nature of the study and the relatively small number of patients. Nevertheless, the number of diabetic patients on whom we performed our analysis was consistent with most of the studies cited. Additionally, the follow-up period in this study was limited to the mid-term results. Ultimately, this study did not include a control group. Conclusion For selected cases of diabetic CN with severe foot and ankle deformity, hindfoot arthrodesis using a retrograde intramedullary nail results in fusion, limb salvage, and a significant amelioration of quality of life. In addition, a thorough preoperative management protocol, including an ulcer-free approach and glycemic regulation, may minimize the rate of postoperative complications. ME: Designing the trial, acquiring, analyzing, and interpreting the data, writing and editing the manuscript. MD: Designing the trial, acquiring, analyzing, and interpreting the data, writing and editing the manuscript. MC: Acquiring, analyzing, and interpreting the data, writing and editing the manuscript. FB: Acquiring and analyzing the data, editing the manuscript. OIK: Conception of the study, designing the trial, acquiring, analyzing, and interpreting the data, writing and editing the manuscript. Acta thanks Kaj Knutson for help with peer review of this study.

Aktaş Ş, Baktıroğlu S, Demir L, Kılıçoğlu Ö, Topalan M, Güven E, Mirasoğlu B, Yanar F. Intralesional application of epidermal growth factor in limbthreatening ischemic diabetic foot ulcers. Acta Orthop Traumatol Turc 2016; 50(3): 277-83. Alvarez R G, Barbour T M, Perkins T D. Tibiocalcaneal arthrodesis for nonbraceable neuropathic ankle deformity. Foot Ankle Int 1994; 15(7): 354-9. Bennett G L, Cameron B, Njus G, Saunders M, Kay D B. Tibiotalocalcaneal arthrodesis: a biomechanical assessment of stability. Foot Ankle Int 2005; 26(7): 530-6. Caravaggi C, Cimmino M, Caruso S, Dalla Noce S. Intramedullary compressive nail fixation for the treatment of severe Charcot deformity of the ankle and rear foot. J Foot Ankle Surg 2006; 45(1): 20-4.

Acta Orthopaedica 2020; 91 (3): 336–340

Caravaggi C M, Sganzaroli A B, Galenda P, Balaudo M, Gherardi P, Simonetti D, Ferraresi R, Farnetti A, Morandi A. Long-term follow-up of tibiocalcaneal arthrodesis in diabetic patients with early chronic Charcot osteoarthropathy. J Foot Ankle Surg 2012; 51(4): 408-11. Chahal J, Stephen D J, Bulmer B, Daniels T, Kreder H J. Factors associated with outcome after subtalar arthrodesis. J Orthop Trauma 2006; 20(8): 55561. Chraim M, Krenn S, Alrabai H, Trnka H, Bock P. Mid-term follow-up of patients with hindfoot arthrodesis with retrograde compression intramedullary nail in Charcot neuroarthropathy of the hindfoot. Bone Joint J 2018; 100(2): 190-6. Kitaoka H B, Alexander I J, Adelaar R S, Nunley J A, Myerson M S, Sanders M. Clinical rating systems for the ankle–hindfoot, midfoot, hallux, and lesser toes. Foot Ankle Int 1994; 15(7): 349-53. Klos K, Gueorguiev B, Schwieger K, Fröber R, Brodt S, Hofmann G O, Windolf M, Mückley T. Comparison of calcaneal fixation of a retrograde intramedullary nail with a fixed-angle spiral blade versus a fixed-angle screw. Foot Ankle Int 2009; 30(12): 1212-18. Mazur J M, Schwartz E, Simon S R. Ankle arthrodesis. Long-term follow-up with gait analysis. J Bone Joint Surg Am 1979; 61(7): 964-75. Mendicino R W, Catanzariti A R, Saltrick K R, Dombek M F, Tullis B L, Statler T K, Johnson B M. Tibiotalocalcaneal arthrodesis with retrograde intramedullary nailing. J Foot Ankle Surg 2004; 43(2): 82-6. Myers T G, Lowery N J, Frykberg R G, Wukich D K. Ankle and hindfoot fusions: comparison of outcomes in patients with and without diabetes. Foot Ankle Int 2012; 33(1): 20-8. Pelton K, Hofer J K, Thordarson D B. Tibiotalocalcaneal arthrodesis using a dynamically locked retrograde intramedullary nail. Foot Ankle Int 2006; 27(10): 759-63. Perlman M H, Thordarson D B. Ankle fusion in a high risk population: an assessment of nonunion risk factors. Foot Ankle Int 1999; 20(8): 491-6. Pinzur M S, Kelikian A. Charcot ankle fusion with a retrograde locked intramedullary nail. Foot Ankle Int 1997; 18(11): 699-704. Pinzur M S, Noonan T. Ankle arthrodesis with a retrograde femoral nail for Charcot ankle arthropathy. Foot Ankle Int 2005; 26(7): 545-9. Rosenbaum A J, DiPreta J A. Classifications in brief: Eichenholtz classification of Charcot arthropathy. New York: Springer; 2015. Russotti G M, Johnson K A, Cass J. Tibiotalocalcaneal arthrodesis for arthritis and deformity of the hind part of the foot. J Bone Joint Surg Am 1988; 70(9): 1304-7. ter Gunne A F P, Cohen D B. Incidence, prevalence, and analysis of risk factors for surgical site infection following adult spinal surgery. Spine 2009; 34(13): 1422-8. van der Ven A, Chapman C B, Bowker J H. Charcot neuroarthropathy of the foot and ankle. J Am Acad Orthop Surg 2009; 17(9): 562-71. Ware J E, Kosinski M, Gandek B, Aaronson N K, Apolone G, Bech P, Brazier J, Bullinger M, Kaasa S, Leplège A. The factor structure of the SF-36 Health Survey in 10 countries: results from the IQOLA Project. J Clin Epidemiol 1998; 51(11): 1159-65. Wukich D K, Shen J Y, Ramirez C P, Irrgang J J. Retrograde ankle arthrodesis using an intramedullary nail: a comparison of patients with and without diabetes mellitus. J Foot Ankle Surg 2011; 50(3): 299-306. Yakacki C M, Gall K, Dirschl D R, Pacaccio D J. Pseudoelastic intramedullary nailing for tibio–talo–calcaneal arthrodesis. Expert Rev Med Devices 2011; 8(2): 159-66.

Acta Orthopaedica 2020; 91 (3): 341–346 341

Modified Harrington’s procedure for periacetabular metastases in 89 cases: a reliable method for cancer patients with good functional outcome, especially with long expected survival Gilber KASK 1,3, Jyrki NIEMINEN 2, Vincent VAN ITERSON 3, Mihhail NABOISTSIKOV 1, Toni-Karri PAKARINEN 2, and Minna K LAITINEN 3 1 Department

of Orthopaedics and Traumatology, Tampere University Hospital, Tampere; 2 Coxa Hospital for Joint Replacement, Tampere; 3 Department of Orthopaedics and Traumatology, Helsinki University Hospital, University of Helsinki, Helsinki, Finland Correspondence: gilber.kask@pshp.fi Submitted 2019-09-30. Accepted 2020-02-04.

Background and purpose — The pelvis is the 3rd most common site of skeletal metastases. In some cases, periacetabular lesions require palliative surgical management. We investigated functional outcome, complications, and implant and patient survival after a modified Harrington’s procedure. Patients and methods — This retrospective cohort study included 89 cases of surgically treated periacetabular metastases. All patients were treated with the modified Harrington’s procedure including a restoration ring. Lesions were classified according to Harrington. Functional outcome was assessed by Harris Hip Score (HHS) and Oxford Hip Score (OHS). Postoperative complications, and implant and patient survival are reported. Results — The overall postoperative functional outcome was good to fair (OHS 37 and HHS 76). Sex, age, survival > 6 and 12 months, and diagnosis of the primary tumor affected functional outcome. Overall implant survival was 96% (95% Cl 88–100) at 1 year, 2 years, and 5 years; only 1 acetabular implant required revision. Median patient survival was 8 months (0–125). 10/89 patients had postoperative complications: 6 major complications, leading to revision surgery, and 4 minor complications. Interpretation — Our modified Harrington’s procedure with a restoration ring to achieve stable fixation, constrained acetabular cup to prevent dislocation, and antegrade iliac screws to prevent cranial protrusion is a reliable reconstruction for periacetabular metastases and results in a good functional outcome in patients with prolonged survival. A standardized procedure and low complication rate encourage the use of this method for all Harrington class defects.

The pelvis is the 3rd most common site for surgically treated skeletal metastases after the femur and humerus (Ratasvuori et al. 2013). In deciding whether and how to operate on periacetabular lesions, the estimated patient survival and size of the skeletal lesion should be considered. Expected survival is dependent on the type of primary tumor and metastatic burden. The mean survival of pathological fractures in the pelvic area is usually less than 2 years (Hansen et al. 2004, Ratasvuori et al. 2014). Periacetabular defects can be reconstructed in several ways depending on the extent. Harrington’s classification separates cases as follows: class I, the acetabular lateral cortices and superior and medial walls are intact; class II, the medial wall is deficient; class III, the lateral cortices, medial wall, and superior wall are all deficient; and class IV, there is wide destruction all the way to the wing of the ilium (Harrington 1981). Harrington also designed a method for reconstruction in cases in which the periacetabular bone presents extensive loss, as in classes III and IV. In this conventional procedure, antegrade pins (from the wing of the ilium to the acetabular dome) or retrograde pins (from the acetabular dome into the wing of the ilium and into the sacroiliac joint) are used. Other methods are also available for the reconstruction, such as filling metastatic cavities with bone cement (cementoplasty), acetabular cages, custom-made pelvic endoprostheses, and the “ice-cream cone” periacetabular prosthesis (Walker 1993, Harrington 1995, Fisher et al. 2011). The original Harrington’s procedure is rarely used any more, whereas some studies have other procedures, usually less invasive, e.g., no pins in the iliac crest (Tsagozis et al. 2015), using short screws or pins (Bernthal et al. 2015, Tsagozis et al. 2015), and not performing arthroplasty (Charles et al. 2017).

342

Acta Orthopaedica 2020; 91 (3): 341–346

Table 1. Patient characteristics. Values are number of cases unless otherwise specified Characteristics n Eligible cases Sex Female Male Mean/median follow-up months (range) Mean age at surgery years (range) Age > 60 ≤ 60 Metastatic load Multiple bone Solitary bone Bone, lung, and other Bone and lung ASA 2 3 4 Data missing Radiotherapy Preoperative Postoperative Pre- and postoperative None Primary malignant tumor Breast Prostate Renal cell Myeloma Lung Other Last status Dead due to cancer Dead due to treatment Dead due to other cause Dead due to unknown reason Alive

89 50 39 18/8 (0–125) 67 (27–94) 62 27 43 18 17 11 14 41 26 8 31 37 2 19

51 year old woman with primary tumor of sigma carcinoma with bone, lung, and liver metastasis. Harrington classification 2 (A). The modified Harrington’s procedure for periacetabular metastases (B). 3 cannulated screws are directed to the roof of the acetabulum and the periacetabular defect is supported by a restoration reinforcement ring. A cemented acetabular cup is implanted. The bone cement is used to augment the bone defects, reinforcement ring, and antegrade inserted screws.

28 13 11 10 7 20 55 3 1 10 20

Sample sizes in publications reporting conventional and modified Harrington’s technique are small, ranging from 19 to 51 patients (Harrington 1981, Nilsson et al. 2000, Tillman et al. 2008, Shahid et al. 2014, Charles et al. 2017), and the publications reporting functional outcomes are few (Nilsson et al. 2000, Charles et al. 2017). The aim of this study was to report the functional outcome, post-operative implant survival, including complications, and patient survival after modified Harrington’s procedure.

Patients and methods Study design The medical records from prospectively maintained hospital databases were reviewed retrospectively. 89 periacetabular metastasis cases treated surgically with the modified Harrington’s procedure at Helsinki University Hospital, Helsinki, and Coxa Replacement Hospital, Tampere, Finland, were

included in the analysis. Patients were included if they met the following inclusion criteria: age ≥ 18 years, diagnosis of metastatic disease in the pelvic periacetabular bone, and surgery for an impending fracture or existent pathological fracture of the periacetabular area between January 2006 and December 2018 (Table 1). Patients with (multiple) myeloma and lymphoma were also included because the surgical approach for these diseases is similar to the treatment of metastatic long bone disease. No predefined criteria were used by the surgeons to make the decision to operate; the patient and physician together decided whether to operate. Preoperative radiological assessment was performed by plain radiographs, CT, and/or MRI. Metastatic lesions were classified according to Harrington (Harrington 1981). Surgical procedure All procedures were performed by orthopedic oncologists, and a similar procedure was performed on all patients independent of Harrington’s classification. Patients were placed under spinal anesthesia, and a preoperative prophylactic antibiotic was used. In our modified Harrington’s technique, the procedure was started by a posterior approach. The hip was dislocated and the neck resected. Curettage of periacetabular metastases was performed, and then cannulated mostly fully threaded, occasionally partially threaded 7.3 mm screws were inserted from the iliac crest through a separate transverse iliac incision. The screws were directed to the roof of the acetabulum. The periacetabular defect was then supported by adding a Restoration GAP II reinforcement ring (Stryker, Mahwah,

Acta Orthopaedica 2020; 91 (3): 341–346 343

Table 2. Perioperative characteristics of the 89 patients. Values are number of cases unless otherwise specified Characteristics n Operation Primary only Revision Mean lesion size (cm) (range) Harrington’s classification Class I Class II Class III Class IV Preoperative embolization Yes No Data missing Mean operation blood loss (L) (range) Impending fracture Pathologic fracture Mean operation time (hr) (range) Type of prothesis Cemented regular Uncemented regular Cemented long-stem Uncemented long-stem Tumor prothesis Acetabulum component Constrained Normal Antegrade acetabular pins (n) 2 3 4 5 Head size (mm) 22 28 30 32 36

83 6 8.9 (4–25) 36 41 12 – 15 69 5 1.6 (0.1–5.7) 33 56 2.8 (2.2–7.2) 70 2 7 3 7 56 33 3 68 11 7 2 11 1 18 57

NJ, USA). A conventional normal or constrained cemented acetabular cup was implanted, followed by the femoral component, and bone cement was used to augment the bone defects, reinforcement ring, and antegrade inserted screws (Figure). Various components were used in both the femur and acetabulum throughout the study period (Table 2). Preoperative embolization was selectively performed in 15 cases. The majority of patients received postoperative radiation therapy. All patients were mobilized postoperatively, allowing immediate full weight-bearing. Postoperative clinical assessment was performed routinely after 2–3 months, 6 months, and 12 months. Functional outcome Functional outcome was assessed by the Harris Hip Score (HHS) and the Oxford Hip Score (OHS). To report the level of the functional outcome, we used the grading system by Marchetti et al. (2005). Functional outcomes based on HHS were: < 70, poor; 70–79, fair; 80–89, good; and 90–100, excellent.

The OHS is a patient-reported outcome measure (PROM) devised as a joint-specific instrument. A score > 41 is considered excellent, 34–41 good, 27–33 fair, and < 27 poor functional outcome (Kalairajah et al. 2005). To report a good functional outcome, we used the definition by Hamilton et al. (2018) of “treatment success” for THA patients based on an OHS threshold of 37.5 points. Questionnaires were not used for patients with < 2 months of postoperative follow-up or in bad general health, or when patients were confined to bed or disoriented by advanced disease. Complications Minor, major, and mechanical postoperative complications are reported. Complications were deemed major when surgical reintervention was needed. Mechanical complications were defined as an implant failure. Implant survival was defined as the time from the Harrington procedure to revision due to any cause. Statistics Statistical analyses were performed using SPSS Statistics 23.0 (IBM Corp, Armonk, NY, USA). A p-value < 0.05 was considered significant. The Kaplan–Meier method was applied for patient and implant survival. Patient survival rates were calculated from the date of surgery to the most recent follow-up or death, and implant survival to revision surgery due to a failure of the reconstruction. Between-group comparisons were performed using the log-rank test. Continuous variables are reported as means and 95% confidence intervals (CI). The chisquared test or Fisher’s exact test was used to compare variables between groups, and the Kruskal–Wallis test for means between groups. Linear regression analysis was performed to determine the relation in scores in a time-series analysis. Ethics, funding, data sharing plan, and potential conflicts of interest This retrospective cohort study was approved by the local chair of the audit department. This research received no specific grant or funding from any funding agency in the public, commercial, or not-for-profit sectors. The data are available from the corresponding author. No competing interests are declared.

Results Functional outcome Functional outcome data were collected from 53/89 patients. The OHS was determined for 18/89 patients and the HHS for 51/89. Data were missing for 36/89 patients: 14/89 survived < 2 months, and 22/89 could not answer the questions due to bad general health. Functional outcome measurements were performed on average at 1 year (OHS at 17 months and HHS at 9 months).

344

Acta Orthopaedica 2020; 91 (3): 341–346

Table 3. Functional outcome scores Characteristics

OHS score n mean p-value

Table 4. Complications of the Harrington’s procedure cases HHS score n mean p-value

Eligible cases 18 37 51 76 Sex 0.6 Female 8 38 26 82 Male 10 36 25 69 Age over 60 years 0.5 Yes 15 36 36 71 No 3 41 15 87 Metastatic load 1.0 Multiple bone 10 37 27 76 Solitary bone 6 37 10 76 Bone and other 2 37 14 76 ASA 0.2 2 3 46 9 86 3 5 32 24 74 4 4 38 12 70 Radiotherapy 0.5 Preoperative 5 34 16 76 Postoperative 8 40 27 76 Pre- and post. 0 – 1 94 None 5 36 7 73 Harrington’s classification 0.9 Class I 4 36 17 79 Class II 10 37 24 72 Class III 4 38 10 80 Primary malignant tumor 1.0 Breast 6 37 19 85 Prostate 3 40 9 72 Renal cell 2 33 6 55 Myeloma 3 35 8 75 Lung 6 37 19 85 Other 1 43 4 81 Survival over 6 months 0.4 Yes 16 36 38 79 No 2 42 13 67 Survival over 12 months 0.7 Yes 15 37 30 81 No 3 35 21 68 Operation Primary only 18 37 48 75 Revision – – 3 80 Acetabulum component 0.3 Constrained 14 36 38 76 Normal 4 41 13 75

0.01

Compli- Mechanical Hip joint Local tumor cations failure dislocation progression Infection Decubitus Major 1 2 Minor 0 2

0 2 1 1 2 0

Overall 1

4 1

0.001 1.0

0.1

0.8

0.3

0.04

0.03 0.01 0.7 0.8

The average postoperative OHS was 37 (good), and 6/18 of the patients had successful treatment (> 37.5 points) according to Hamilton’s definition (Hamilton et al. 2018). The average HHS was 76, but was statistically significantly better in female patients (82 versus 69; p = 0.01), patients aged < 60 years (87 versus 71; p = 0.001), and in patients with survival > 6 months (79 versus 67; p = 0.03) and > 12 months (81 versus 68; p = 0.01). The HHS was better in patients with prostate cancer (85; p = 0.04) and worse in patients with myeloma (55; p = 0.04) (Table 3). From linear regression analysis, we could not detect a statistically significant increase in scores and time. Implant survival and complications Overall implant survival was 96% (Cl 88–100) at 1 year, 2 years, and 5 years. Harrington’s classification (p = 0.9), radio-

therapy (p = 0.3), sex (p = 0.3), and type of primary tumor (p = 0.3) did not statistically significantly influence implant survival at univariable analysis. 10/89 postoperative complications occurred: 6 major, leading to revision surgery, and 4 minor, which were treated nonoperatively. 1 patient had mechanical failure of the initial construct due to constant dislocation, necessitating re-intervention at 6 months. The postoperative infection rate was 4/89 (Table 4). We observed fewer dislocations when using constrained cups (1/56 patient) when compared with normal cups (3/33 patients) (p = 0.04). Patient survival 55/89 of the patients died of cancer. 3 died due to treatment: 1 due to fatal pulmonary embolism and 2 due to consequences of deep infection and sepsis. Overall patient survival was 46% (Cl 35–57) at 1 year, 25% (Cl 14–35) at 2 years, and 16% (Cl 7–25) at 5 years. Median patient survival was 8 months (0–125). In patients with skeletal and lung metastases, the overall survival was 50% at 4 months and 23% at 1 year. The survival was significantly better in patients with skeletal metastases only, compared with patients with more disseminated disease (p < 0.001). We could not observe a statistically significant difference in survival between patients with solitary or multiple skeletal metastases, nor did we detect a difference between different Harrington’s classifications on patient survival.

Discussion We used our modified Harrington technique, namely the insertion of antegrade screws from the iliac crest and an antiprotrusion cup combined with THA, in all patients with periacetabular metastasis and pathological fractures regardless of the size of the lesion. Our results show that the reconstruction is durable, with low implant-related failure and complication rates, and good functional outcomes. Functional outcomes have been reported in a few studies, usually in publications with small sample sizes of Harrington or modified Harrington procedure (range 19 to 51 patients) (Harrington 1992, Nilsson et al. 2000, Tillman et al 2008, Shahid et al. 2014, Charles et al. 2017), or with nonspecific outcome measures (Tillman et al. 2008, Charles et al

Acta Orthopaedica 2020; 91 (3): 341–346 345

2017, Wegrzyn et al. 2018, Rowell et al. 2019). No studies have been published using PROMs. Most of the studies have used ClinROMs, such as the Musculoskeletal Tumor Society (MSTS) score (Harrington 1992, Allan et al. 1995) or the HHS (Wegrzyn et al. 2018), which was developed for the assessment of hip surgery outcomes and is intended to evaluate various hip disabilities and methods of treatment in an adult population. In accordance with the literature, our results shows that about half of periacetabular metastasis patients could ambulate in the community independently after such large reconstructive surgery (Wegrzyn et al. 2018, Rowell et al. 2019). The OHS is a validated, reliable, and well-established PROM for evaluating the outcomes of THA (Dawson et al. 1996). It is difficult to compare our results with the literature, as postoperative mobility has been reported in different ways (Harrington 1981, Marco et al. 2000, Shahid et al. 2014, Charles et al. 2017, Wegrzyn et al. 2018, Rowell et al. 2019). Based on Marchetti’s grading system (Marchetti et al. 2005), the functional outcome results were fair in our population when all patients were evaluated together (HHS score 76 out of 100). However, in young female, mainly breast cancer patients with good estimated survival, the scores were > 80, indicating significantly improved results compared with the overall study population. The average OHS was 37 in our study. Based on the classification of Kalairajah et al. (2005), in which a score of 34 to 41 indicates a good functional outcome, a good outcome is represented in our population. In a recent study, Hamilton et al. (2018) defined treatment success following THA based on an OHS threshold of 37.5 points. In our study, one-third of patients had an OHS exceeding this threshold. Again, the best OHSs in our study were seen in young female patients with long survival. Primary tumor and, most importantly, defect defined by Harrington’s classification did not affect the HHS or OHS. In accordance with the literature, our results showed that age strongly affects functional outcome, as the absolute score tends to decrease with age (Bremner-Smith et al. 2004), and a patient’s overall physical status is an important factor in estimating a good functional outcome. Estimating patient survival is of outmost importance, as reconstruction survival should be longer than patient survival. Since Harrington’s original publication, several modifications of the procedure have been introduced, but the idea of transferring the weight load from the acetabular joint to the intact bone in the ileum over the lytic lesion in the periacetabular area via threaded screws or pins and a reinforcement ring has remained the same. Compared with the literature, our implant survival was superior: 96% at 1, 2, and 5 years, as only 1 reconstruction required revision (Nilsson et al. 2000, Shahid et al. 2014, Tsagozis et al. 2015, Erol et al. 2016, Charles et al. 2017). In the literature, implant survival is rarely reported, but reported implant failure rates leading to revision surgery vary between 3% and 18%. Revisions are needed due to fractured rods, constant or repeated dislocations, and mechanical and technical errors similar to our case (Allan et al. 1995, Marco et

al. 2000, Bernthal et al. 2015, Colman et al. 2015). Complications after surgery vary considerably, with 11% in our study and published reports ranging from 16% to 33% (Nilsson et al. 2000, Shahid et al. 2014, Tsagozis et al. 2015, Charles et al. 2017, Krishnan et al. 2017). Complications are often divided into major complications requiring surgery and minor complications, which are treated nonoperatively. Deep infection and recurrent dislocation are common major complications requiring revision surgery. Local tumor progression, in addition to mechanical instability due to insufficient fixation, is more likely to occur in large lesions or long-term survivors (Nilsson et al. 2000, Shahid et al. 2014, Tsagozis et al. 2015, Charles et al. 2017). In accordance with the literature, dislocation and infection were the most common complications in our study, and we did not have any failures due to local tumor growth. The prognosis of many patients with metastatic bone disease, particularly those without visceral disease, has improved in recent years (Kimura 2018). Although advanced oncological treatment results in the prolonged survival of some patients, almost half of the patients die within the 1st year after surgery for pelvic metastases. Our patient survival rates are in accordance with the literature, with 1-year survival ranging between 42% and 49% (Nilsson et al. 2000, Shahid et al. 2014, Tsagozis et al. 2015). Visceral metastases decrease survival, whereas it is suggested that solitary skeletal metastases, at least in the extremities, be treated with margins to achieve improved survival (Ratasvuori et al. 2014). In our series, the skeletal metastasis was solitary in 18 patients, the surgical margin was always intralesional, and patient survival was similar to patients with more disseminated disease. Increased surgical margins in the periacetabular location result in more massive resection, with an increased rate of complications; therefore, based on our results, we select Harrington’s procedure with intralesional curettage as the method of choice for solitary bone metastases in the periacetabular region. Our study has several limitations, including those inherent to its retrospective design. The mobility and functional scores were not structurally assessed, and OHS and/or HHS were available for only 53/89 of the patients. In addition, 14/89 of the patients were deceased before the first follow-up and 22/89 were unable to come to a follow-up visit. However, the results can be used to estimate the probable functional outcome in patients with improved survival. A strength of this multicenter study was the large sample size with a homogeneous treatment method for periacetabular metastasis in contrast to previous ones, in which very heterogeneous technical methods were combined. A further strength of the paper was the use of well-validated tools for functional outcome measurement. In conclusion, functional outcome is mostly dependent on patient-related factors, as young female patients with long survival have good functional outcomes. The results indicate that our modified Harrington’s procedure with the Restoration GAP II ring to achieve stable fixation with bone cement, a constrained acetabular cup to prevent dislocation, and antegrade

346

iliac screws to prevent cranial protrusion is a reliable reconstruction for periacetabular metastases and achieves excellent implant survival in all patients. Though the complication number was 10/89, the reconstruction survival was surprisingly good, at 96% overall. Furthermore, the overall patient survival was 46% at 1 year. Patients with skeletal metastases only had improved survival compared with patients with visceral metastases.

MKL and GK concepted and designed the study. Data extraction was performed by MKL. Data analysis and interpretation were performed by GK and MKL. GK, MN, and MKL were major contributors in writing the article. Article drafting and revising were performed by MKL, JN, VVI, and TKP. The final version was approved by all authors. Acta thanks Johnny Keller and Panagiotis Tsagozis for help with peer review of this study.

Allan D G, Bell R S, Davis A, Langer F. Complex acetabular reconstruction for metastatic tumor. J Arthroplasty 1995; 10(3): 301-6. Bernthal N M, Price S L, Monument M J, Wilkinson B, Jones K B, Randall R L. Outcomes of modified Harrington reconstructions for nonprimary periacetabular tumors: an effective and inexpensive technique. Ann Surg Oncol 2015; 22(12): 3921-8. Bremner-Smith A T, Ewings P, Weale A E. Knee scores in a ‘normal’ elderly population. Knee 2004; 11(4): 279-82. Charles T, Ameye L, Gebhart M. Surgical treatment for periacetabular metastatic lesions. Eur J Surg Oncol 2017; 43(9): 1727-32. Colman M W, Karim S M, Hirsch J A, Yoo A J, Schwab J H, Hornicek F J, Raskin K A. Percutaneous acetabuloplasty compared with open reconstruction for extensive periacetabular carcinoma metastases. J Arthroplasty 2015; 30(9): 1586-91. Dawson J, Fitzpatrick R, Carr A, Murray D. Questionnaire on the perceptions of patients about total hip replacement. J Bone Joint Surg Br 1996; 78(2): 185-90. Erol B, Aydemir A N, Onay T, Topkar M O. Reconstruction of advanced periacetabular metastatic lesions with modified Harrington procedure. Acta Orthop Traumatol Turc 2016; 50(2): 178-85. Fisher N E, Patton J T, Grimer R J, Porter D, Jeys L, Tillman R M, Abudu A, Carter S R. Ice-cream cone reconstruction of the pelvis: a new type of pelvic replacement: early results. J Bone Joint Surg Br 2011; 93(5): 684-8. Hamilton D F, Loth F L, MacDonald D J, Giesinger K, Patton J T, Simpson A H, Howie C R, Giesinger J M. Treatment success following joint arthroplasty: defining thresholds for the Oxford Hip and Knee Scores. J Arthroplasty 2018; 33(8): 2392-7. Hansen B H, Keller J, Laitinen M, Berg P, Skjeldal S, Trovik C, Nilsson J, Walloe A, Kalen A, Wedin R. The Scandinavian Sarcoma Group Skeletal

Acta Orthopaedica 2020; 91 (3): 341–346

Metastasis Register. Survival after surgery for bone metastases in the pelvis and extremities. Acta Orthop Scand 2004; 75(311 Suppl.): 11-15. Harrington K D. The management of acetabular insufficiency secondary to metastatic malignant disease. J Bone Joint Surg Am 1981; 63(4): 653-64. Harrington K D. The use of hemipelvic allografts or autoclaved grafts for reconstruction after wide resections of malignant tumors of the pelvis. J Bone Joint Surg Am 1992; 74(3): 331-41. Harrington K D. Orthopaedic management of extremity and pelvic lesions. Clin Orthop Relat Res 1995; (312): 136-47. Kalairajah Y, Azurza K, Hulme C, Molloy S, Drabu K J. Health outcome measures in the evaluation of total hip arthroplasties: a comparison between the Harris Hip Score and the Oxford Hip Score. J Arthroplasty 2005; 20(8): 1037-41. Kimura T. Multidisciplinary approach for bone metastasis: a review. Cancers (Basel) 2018; 10(6). Krishnan C K, Han I, Kim H S. Outcome after surgery for metastases to the pelvic bone: a single institutional experience. Clin Orthop Surg 2017; 9(1): 116-25. Marchetti P, Binazzi R, Vaccari V, Girolami M, Morici F, Impallomeni C, Commessatti M, Silvello L. Long-term results with cementless Fitek (or Fitmore) cups. J Arthroplasty 2005; 20(6): 730-7. Marco R A, Sheth D S, Boland P J, Wunder J S, Siegel J A, Healey J H. Functional and oncological outcome of acetabular reconstruction for the treatment of metastatic disease. J Bone Joint Surg Am 2000; 82(5): 642-51. Nilsson J, Gustafson P, Fornander P, Ornstein E. The Harrington reconstruction for advanced periacetabular metastatic destruction: good outcome in 32 patients. Acta Orthop Scand 2000; 71(6): 591-6. Ratasvuori M, Wedin R, Keller J, Nottrott M, Zaikova O, Bergh P, Kalen A, Nilsson J, Jonsson H, Laitinen M. Insight opinion to surgically treated metastatic bone disease: Scandinavian Sarcoma Group Skeletal Metastasis Registry report of 1195 operated skeletal metastasis. Surg Oncol 2013; 22(2): 132-8. Ratasvuori M, Wedin R, Hansen B H, Keller J, Trovik C, Zaikova O, Bergh P, Kalen A, Laitinen M. Prognostic role of en-bloc resection and late onset of bone metastasis in patients with bone-seeking carcinomas of the kidney, breast, lung, and prostate: SSG study on 672 operated skeletal metastases. J Surg Oncol 2014; 110(4): 360-5. Rowell P, Lowe M, Sommerville S, Dickinson I. Is an acetabular cage and cement fixation sufficiently durable for the treatment of destructive acetabular metastases? Clin Orthop Relat Res 2019; 477(6): 1459-65. Shahid M, Saunders T, Jeys L, Grimer R. The outcome of surgical treatment for peri-acetabular metastases. Bone Joint J 2014; 96-B(1): 132-6. Tillman R M, Myers G J, Abudu A T, Carter S R, Grimer R J. The three-pin modified ‘Harrington’ procedure for advanced metastatic destruction of the acetabulum. J Bone Joint Surg Br 2008; 90(1): 84-7. Tsagozis P, Wedin R, Brosjo O, Bauer H. Reconstruction of metastatic acetabular defects using a modified Harrington procedure. Acta Orthop 2015; 86(6): 690-4. Walker R H. Pelvic reconstruction/total hip arthroplasty for metastatic acetabular insufficiency. Clin Orthop Relat Res 1993(294): 170-5. Wegrzyn J, Malatray M, Al-Qahtani T, Pibarot V, Confavreux C, Freyer G. Total hip arthroplasty for periacetabular metastatic disease: an original technique of reconstruction according to the Harrington classification. J Arthroplasty 2018; 33(8): 2546-55.

Acta Orthopaedica 2020; 91 (3): 347–352 347

Surgical site infection after hip fracture – mortality and risk factors: an observational cohort study of 1,709 patients Christian T POLLMANN 1,2, Fredrik A DAHL 2,3, Jan Harald M RØTTERUD 1, Jan-Erik GJERTSEN 4,5, and Asbjørn ÅRØEN 1,2,6 1 Department of Orthopedic Surgery, Akershus University Hospital, Lørenskog; 2 Institute of Clinical Medicine, Campus Ahus, University of Oslo, Oslo; 3 Health Services Research Unit, Akershus University Hospital, Lørenskog; 4 Norwegian Hip Fracture Register, Department of Orthopedic Surgery, Haukeland University Hospital, Bergen; 5 Department of Clinical Medicine (K1), University of Bergen, Bergen; 6 Department of Sports Medicine; Norwegian School of Sport Sciences, Oslo, Norway Correspondence: christian.pollmann@ahus.no Submitted 2019-09-27. Accepted 2019-12-17.

Background and purpose — Surgical site infection (SSI) is a devastating complication of hip fracture surgery. We studied the contribution of early deep SSI to mortality after hip fracture surgery and the risk factors for deep SSI with emphasis on the duration of surgery. Patients and methods — 1,709 patients (884 hemi arthroplasties, 825 sliding hip screws), operated from 2012 to 2015 at a single center were included. Data were obtained from the Norwegian Hip Fracture Register, the electronic hospital records, the Norwegian Surveillance System for Antibiotic Use and Hospital-Acquired Infections, and the Central Population Register. Results — The rate of early (≤ 30 days) deep SSI was 2.2% (38/1,709). Additionally, for hemiarthroplasties 7 delayed (> 30 days, ≤ 1 year) deep SSIs were reported. In patients with early deep SSI 90-day mortality tripled (42% vs. 14%, p < 0.001) and 1-year mortality doubled (55% vs. 24%, p < 0.001). In multivariable analysis, early deep SSI was an independent risk factor for mortality (RR 2.4 for 90-day mortality, 1.8 for 1-year mortality, p < 0.001). In univariable analysis, significant risk factors for early and delayed deep SSI were cognitive impairment, an intraoperative complication, and increasing duration of surgery. However, in the multivariable analysis, duration of surgery was no longer a significant risk factor. Interpretation — Early deep SSI is an independent risk factor for 90-day and 1-year mortality after hip fracture surgery. After controlling for observed confounding, the association between duration of surgery and early and delayed deep SSI was not statistically significant.

Hip fractures, in usually frail, elderly patients, have high mortality rates of around 9% within 30 days (Sheikh et al. 2017) and up to 30% within 1 year (Lund et al. 2014). If a deep surgical site infection (SSI) ensues, a 1-year mortality rate of 50% (Edwards et al. 2008) has been reported. However, it is not clear to what extent this increased mortality rate is due to the infection and the treatment thereof and to what extent it is due to a more pronounced frailty which predisposed these patients to SSI (Belmont et al. 2014). Considering the serious consequences of SSI for hip fracture patients it is important to optimize modifiable risk factors. However, reported risk factors differ, ranging from operative delay to the lead surgeon’s experience, duration of surgery, choice of implant, and patient factors such as obesity (Harrison et al. 2012, Cordero et al. 2016, de Jong et al. 2017, Zajonz et al. 2019). Duration of surgery is a risk factor commonly focused upon. However, the question remains as to whether longer duration of surgery increases the risk of SSI by prolonging exposure to possible bacterial contamination (Stocks et al. 2010) or if prolonged duration of surgery represents a surrogate parameter for a difficult procedure or a complication as the main cause for an increased risk of SSI. In this observational cohort study, we investigated the contribution of early deep SSI to mortality after hip fracture surgery and risk factors for early and delayed deep SSI in hip fracture patients with particular emphasis on the role of duration of surgery.

348

Patients and methods Patients All patients 18 years of age or older who were operated with a hemiarthroplasty or a sliding hip screw for a non-pathologic fracture of the proximal femur at a single institution (Akershus university hospital [AUH]) from January 2012 through December 2015 and who were reported to the Norwegian Hip Fracture Register (NHFR) (Gjertsen et al. 2008) were included in this study (Figure 1). In patients who sustained 2 hip fractures during the study period (n = 92), only the 1st fracture was included in the analyses. Other data from the present cohort of hip fracture patients have previously been used in an observational study on the effect of fast-track hip fracture care on mortality (Pollmann et al. 2019). Data collection In Norway, hip fracture operations are prospectively reported to the NHFR (Gjertsen et al. 2008) by the surgeon on a 1-page questionnaire, which includes information on the time elapsed from fracture to surgery, cognitive impairment (“no”, “uncertain,” “yes”), ASA score, type of fracture, type of operation, type of anesthesia, pathological fractures, intraoperative complications (“no”/“yes” with supplemental free text), duration of surgery (time from incision to skin closure), and the surgeon’s experience (at least 1 surgeon present with > 3 years of experience in hip fracture surgery; “yes”/“no”). Using the unique 11-digit Norwegian personal identification number data from the NHFR and the electronic hospital records were linked deterministically. Surgical site infection SSIs after hemiarthroplasty and total arthroplasty of the hip are surveyed under the Norwegian Surveillance System for Antibiotic Use and Hospital-Acquired Infections (NOIS) with 30-day and 1-year follow-up. A questionnaire is sent to each patient or, in the case of cognitive impairment or institutionalization, to the primary health care provider. If the patient reports an SSI or a suspicion of SSI this has to be confirmed by a physician on the same questionnaire. In equivocal cases the electronic hospital records are scrutinized and/or the primary health care provider is contacted. Until 2014, cases of SSI were defined according to the American Centers for Disease Control and Prevention (Horan et al. 2008) while from 2014 onwards case definitions from the European Centre for Disease Prevention and Control have been applied (Dalli 2012). Concerning SSIs, both definitions are practically identical. Sliding hip screws are not monitored by NOIS, but the Department of Microbiology and Infection Control at AUH also surveys SSIs with 30-day follow-up in these patients using the same method and criteria. The completeness of follow-up was 99%. In this study, SSI within 30 days of the index operation

Acta Orthopaedica 2020; 91 (3): 347–352

Primary hip fracture operations from January 2012 through December 2015 in patients ≥ 18 years of age n = 2,634 Excluded (n = 925 ): – second fracture in patients who sustained bilateral fractures, 92 – pathologic fracture, 47 – not reported to the NHFR, 265 – surgical procedures not routinely surveyed for surgical site infection, 426 – no surgical site infections recorded after total hip arthroplasty, 95 Patients included (n = 1,709): – hemiarthroplasties, 884 – sliding hip screws, 825

Figure 1. Flowchart of patient inclusion. NHFR: Norwegian Hip Fracture Register.

is termed early SSI, while an SSI diagnosed between 30 days and 1 year from the index operation is termed delayed SSI. Mortality Mortality data from the Central Population Register are routinely imported into the electronic hospital records. There was no loss to follow-up regarding mortality. Mortality rates were calculated from the time of arrival at the hospital. Survival was censored at 1 year. Antibiotic prophylaxis Fixation with antibiotic-loaded bone cement (0.5 g gentamicin per 40 g cement) was used in all hemiarthroplasties. All patients received perioperative systemic antibiotic prophylaxis. Statistics Fisher’s exact test was used for unadjusted comparisons of proportions. Due to the relatively high mortality rates we chose risk ratios as the statistical effect measure (Davies et al. 1998) for the multivariable model analyzing the effect of early deep SSI on mortality. Since log-binomial regression did not converge, Poisson regression with robust variance was chosen as the statistical model (Barros and Hirakata 2003). Risk ratios (RR) are presented with 95% confidence intervals (CI). We considered survival analysis by Cox regression. However, the Schoenfeld residuals ph-test showed that the proportional hazards assumption was not met, which is also illustrated by the Kaplan–Meier survival curve (Figure 2). Age squared was not a significant risk factor for mortality indicating that the effect of age on mortality was linear in our cohort. Age was therefore included as a continuous variable in the regression models. The type of intraoperative complication, specified as free text, ranged widely from myocardial infarction to technical prob-

Acta Orthopaedica 2020; 91 (3): 347–352 349

Table 1. Patient characteristics. Values are n (%) unless otherwise specified Survivors Deceased Entire cohort at 90 days at 90 days No SSI SSI Factor (n = 1,709) (n = 1,459) (n = 250) (n = 1,664) (n = 45) Age, mean (SD) 82 (9.5) 81 (9.7) 85 (7.5) 82 (9.5) 81 (8.6) Female sex 1,166 (68) 1,019 (70) 147 (59) 1,138 (68) 28 (62) ASA 1 27 (1.6) 27 (1.9) – 27 (1.6) – 2 451 (26) 429 (29) 22 (8.8) 444 (27) 7 (16) 3 1,021 (60) 865 (59) 156 (62) 988 (59) 33 (73) 4 171 (10) 104 (7.1) 67 (27) 166 (10) 5 (11) 5 3 (0.2) 1 (0.1) 2 (0.8) 3 (0.2) – Not reported 36 (2.1) 33 (2.3) 3 (1.2) 36 (2.2) – Cognitive impairment No 1113 (65) 1009 (69) 104 (42) 1091 (66) 22 (49) Uncertain 178 (10) 143 (9.8) 35 (14) 169 (10) 9 (20) Yes 374 (22) 267 (18) 107 (43) 362 (22) 12 (27) Not reported 44 (2.6) 40 (2.7) 4 (1.6) 42 (2.5) 2 (4.4) Type of fracture Femoral neck undisplaced 68 (4.0) 62 (4.2) 6 (2.4) 64 (3.8) 4 (8.9) displaced 806 (47) 683 (47) 123 (49) 782 (47) 24 (53) Basocervical 54 (3.2) 45 (3.1) 9 (3.6) 54 (3.2) Trochanteric 2 fragments 341 (20) 301 (21) 40 (16) 334 (20) 7 (16) > 2 fragments 345 (20) 282 (19) 63 (25) 337 (20) 8 (18) Intertrochanteric 39 (2.3) 34 (2.3) 5 (2.0) 38 (2.3) 1 (2.2) Subtrochanteric 33 (1.9) 32 (2.2) 1 (0.4) 32 (1.9) 1 (2.2) Other 19 (1.1) 16 (1.1) 3 (1.2) 19 (1.1) – Not reported 4 (0.2) 4 (0.3) – 4 (0.2) –

Ethics, registration, funding, and potential conflicts of interest The Regional Ethics Committee South East deemed this study not to require approval (reference number 2015/409). Data were collected and handled in accordance with the requirements of the local data protection officer. The study was exempt from consent to participate. The Norwegian Data Inspectorate has approved the registration of data in the NHFR. The study was funded by research grants from Sophies Minde AS and from the Norwegian Orthopedic Association in cooperation with Heraeus and by the Department of Orthopedic Surgery, Akershus university hospital. The authors declare no conflicts of interest.

Results Patient characteristics are presented in Table 1; baseline data on the primary surgical treatment are given in Table 2 (see Supplementary data).

Surgical site infection The rate of early SSI for all included procedures Percentages are column percentages; (hemiarthroplasties and sliding hip screws) during SSI: early (sliding hip screws) and early and delayed (hemiarthroplasties) deep the study period was 2.2% (38/1,709) with a surgical site infection. variation of between 0.5% and 3.1% per calendar year. For hemiarthroplasties the rate of early SSI lems to nausea and vomiting, to name a few. Therefore, we was 2.4% (21/884) while it was 2.1% (17/825) for sliding hip made no attempt at further classification and intraoperative screws. The cumulative 1-year SSI rate (early and delayed) for complication was treated as a binary variable. Both the ASA hemiarthroplasties was 3.2% (28/884). All SSIs were classified score (≤ 2/≥ 3) and time from fracture to surgery (≤ 24 hours/> as deep and all but 1 patient with an infected hemiarthroplasty, 24 hours) were dichotomized and included as binary variables. who declined surgical treatment, were reoperated due to the SSI. Since there were relatively few cases of SSI, the analysis of risk factors for SSI was based on both early (sliding hip Early deep surgical site infection and mortality screws and hemiarthroplasties) and delayed (hemiarthroplas- 30-day mortality did not differ statistically significantly between ties only) SSIs to achieve a more robust statistical analysis. patients with or without early deep SSI (Table 3). However, Logistic regression was used to analyze the risk factors for 90-day mortality tripled and 1-year mortality more than douearly and delayed deep SSI. bled in patients with early deep SSI (Table 3). A Kaplan–Meier In all multivariable models, the variables to be adjusted for cumulative survival curve illustrates that the increased morwere chosen from directed acyclic graphs (DAG), which were tality in patients with early deep SSI becomes apparent from constructed using DAGitty (Textor et al. 2017). approximately 6 weeks postoperatively (Figure 2). We performed a sensitivity analysis for the effect of early The analysis of the causal association between early deep deep SSI on mortality by calculating the E-value. “The SSI and mortality was based on a DAG (Figure 3a, see SuppleE-value is the minimum strength of association on the risk mentary data) and confounders to be adjusted for were chosen ratio scale that an unobserved confounder would need to have from this DAG (Figure 3b, see Supplementary data). In this with both the exposure and the outcome, above and beyond model, obesity, diabetes mellitus, and smoking represent the measured covariates, to fully explain away a specific expo- unobserved confounders. In a multivariable analysis adjusted sure–outcome association” (VanderWeele and Ding 2017). for age, sex, cognitive impairment, ASA score, the occurrence A p-value < 0.05 was considered as statistically significant. of an intraoperative complication, and time from fracture to Data were analyzed with the SPSS statistical package version surgery, early deep SSI was an independent risk factor for both 25.0.0.1 (IBM Corp, Armonk, NY, USA). 90-day and 1-year mortality (Table 4).

350

Acta Orthopaedica 2020; 91 (3): 347–352

Table 3. Mortality (% and 95% CI) with and without early deep surgical site infection Early No SSI deep SSI Between-group Mortality (n = 1,671) (n = 38) difference p-value a 30-day 8.3 (7.1–9.8) 5.3 (0.9–19) –3.0 (–12 to 5.5) 0.8 90-day 14 (12–156) 42 (27–59) 28 (11 to 45) < 0.001 1-year 24 (22–26) 55 (39–71) 31 (14 to 49) < 0.001 SSI: surgical site infection. a Fisher’s exact test.

Table 4. Multivariable Poisson regression with robust variance of independent risk factors for 90-day and 1-year mortality Factor

90-day mortality Risk ratio (CI) p-value

1-year mortality Risk ratio (CI) p-value

Age Male sex Cognitive impairment uncertain yes ASA score ≥ 3 Intraoperative complication Time from fracture to surgery > 24 h Early deep SSI

1.04 (1.02–1.06) < 0.001 1.7 (1.4–2.2) < 0.001

1.03 (1.02–1.04) < 0.001 1.5 (1.3–1.8) < 0.001

1.5 (1.1–2.2) 2.2 (1.7–2.8) 3.0 (1.9–4.8)

0.02 < 0.001 < 0.001

1.3 (1.0–1.7) 1.8 (1.5–2.2) 2.5 (1.9–3.5)

0.05 < 0.001 < 0.001

1.1 (0.7–1.6)

0.7

1.2 (1.0–1.6)

0.1

1.2 (0.9–1.5) 2.4 (1.6–3.5)

0.2 < 0.001

1.0 (0.8–1.2) 1.8 (1.3–2.5)

0.9 < 0.001

Omitting the variables intraoperative complication and time from fracture to surgery reduces missing cases from 9.8% to 3.9% while the parameter estimates for the remaining variables remain practically unchanged. For the association between early deep SSI and mortality the E-values for the point estimate of the RR and for the lower bound of its CI were 4.2 and 2.6 for 90-day and 3.0 and 1.9 for 1-year mortality. Hence, an unobserved confounder that is associated with both early deep SSI and 90-day mortality by an RR of 4.2 each could explain away the observed RR of 2.4. An unobserved confounder that is associated with both early deep SSI and 90-day mortality by an RR of 2.6 each could move the lower bound of the CI to 1 (VanderWeele and Ding 2017). Risk factors for early and delayed deep surgical site infection In a univariable analysis, cognitive impairment, the occurrence of an intraoperative complication, and longer duration of surgery were statistically significantly associated with an increased risk of early and delayed deep SSI (Table 5, see Supplementary data). An ASA score ≥ 3 bordered on being a statistically significant risk factor for early and delayed deep SSI (Table 5, see Supplementary data).

Duration of surgery and early and delayed deep surgical site infection The analysis of the causal association between duration of surgery and early and delayed deep SSI was based on a DAG (Figure 4a, see Supplementary data). Figure 4b shows which variables have to be adjusted for to control for observed confounding. Obesity represents an unobserved confounder. In a multivariable analysis adjusted for the occurrence of an intraoperative complication and for surgeon’s experience the association between duration of surgery and early and delayed deep SSI is no longer statistically significant (Table 6, see Supplementary data).

Figure 4b. Directed acyclic graph depicting the adjustment for observed confounding of the association between duration of surgery and early exposure; ■ outcome; and delayed deep surgical site infection. ancestor of exposure; ancestor of outcome; ancestor of exposure and outcome (confounder); adjusted variable; unobserved; causal path; biasing path. ▼

SSI: surgical site infection; 9.8% missing.

Figure 2. Kaplan–Meier patient survival curves with 95% confidence intervals for patients with and without early deep surgical site infection. SSI: early deep surgical site infection.

Acta Orthopaedica 2020; 91 (3): 347–352 351

Discussion Early deep surgical site infection and mortality In our cohort, 90-day mortality tripled and 1-year mortality more than doubled in patients with early deep SSI compared with patients without SSI. In multivariable analysis, early deep SSI was an independent and important risk factor for both 90-day and 1-year mortality. Adjusting for age, cognitive impairment, and ASA score controls for a large part of frailty and the fact that SSI remained an independent risk factor for mortality indicates that SSI in itself increases the mortality rate. The 30-day mortality rate did not differ statistically significantly between patients with early deep SSI and without SSI in our cohort and was in fact slightly lower in patients with early deep SSI. Edwards et al. (2008) observed a similar phenomenon. As the authors pointed out, this observation might have been caused by survival bias. The increase in 1-year mortality in our cohort was similar to the one reported by Merrer et al. (2007) (50% vs. 20%) and by Edwards et al. (2008) (50% vs. 30%). The rate of deep SSI in our cohort was in the mid-range of earlier reported rates (Harrison et al. 2012, Sprowson et al. 2016, de Jong et al. 2017). Interestingly, we observed only deep SSIs as opposed to most other studies that report both deep and superficial SSIs (Merrer et al. 2007, Edwards et al. 2008, de Jong et al. 2017). What caused this discrepancy is unclear. Superficial SSIs may have been underreported in our cohort. Another possible explanation might be a difference in treatment strategy. It can be difficult to ascertain that an SSI is purely superficial and we might have a more aggressive approach revising SSIs that others might classify as superficial. Since a diagnosis of deep SSI made by the surgeon is one of the possible criteria that define a deep SSI (Horan et al. 2008, Dalli 2012), an aggressive revision policy could partly explain why no superficial SSIs were reported. Risk factors for early and delayed deep surgical site infection In the univariable analysis cognitive impairment, the occurrence of an intraoperative complication and an increasing duration of surgery were statistically significantly associated with an increased risk of early and delayed deep SSI, while the association with an ASA score ≥ 3 bordered on statistical significance. For clinical practice it is most relevant to identify modifiable risk factors for SSI. Pre-existing cognitive impairment can be considered non-modifiable, while delirium, which has a high incidence amongst hip fracture patients (Watne et al. 2014), and therefore probably accounts for some of the reported cognitive impairment in our cohort, may be preventable in some patients. A high ASA score might be modifiable if it is due to an acute condition or an acute deterioration of an existing ailment. However, most often the ASA score will not be modifiable. Some intraoperative complications may be preventable

with adequate preoperative planning and experienced staff; however, it is in the nature of complications that not all of them can be prevented or even foreseen. An association between a longer duration of surgery and SSI has been shown before in several other studies (Harrison et al. 2012, Daley et al. 2015, Cheng et al. 2017, de Jong et al. 2017). On this basis, some authors have advocated measures to reduce duration of surgery (Cheng et al. 2017), such as expeditious surgical technique (Daley et al. 2015). However, the question remains how much of this association is due to the prolonged exposure to possible microbial contamination (Stocks et al. 2010) and how much is due to a longer duration of surgery being an indicator of a more complex surgical procedure, an inexperienced surgeon, or an intraoperative complication. To try to approach this question, we used a DAG with duration of surgery as the exposure and SSI as the outcome. From this DAG we determined that controlling for the occurrence of an intraoperative complication and for surgeon’s experience would control for all the observed confounders in our cohort. In the corresponding logistic regression model duration of surgery was no longer an independent risk factor for SSI. The fact that controlling for the occurrence of an intraoperative complication and for surgeon’s experience eliminated the statistical significance of the duration of surgery cannot readily be interpreted as duration of surgery not having a direct influence on the risk of SSI. However, this finding highlights the uncertainty that the prolonged exposure to possible bacterial contamination is the main reason for an association between duration of surgery and SSI. De Jong et al. (2017) reported an increased risk of SSI after hemiarthroplasty of the hip for both short (< 45 minutes) and long (> 90 minutes) durations of surgery. This might support the notion that careless tissue handling (short durations of surgery) and intraoperative complications (long durations of surgery) might play an important role in the development of SSI. Our study has several strengths. With 1,709 patients the studied cohort is quite large. By using data from the NHFR, NOIS, the electronic hospital records, and the Central Population Register the cohort was well characterized. The loss to follow-up for SSI was small and no patients were lost to follow-up concerning mortality. The study also has limitations. It is a single-center study. However, approximately 8% of all hip fracture operations in Norway are performed at our institution making this a relevant sample of Norwegian hip fracture patients. The number of cases of SSI was small (38 early SSIs, 7 delayed SSIs), limiting the number of covariates that could be included in and the statistical power of the multivariable regression model analyzing the risk factors for SSI. Information on delayed SSIs was only available for patients operated with a hemiarthroplasty. Data on patients’ comorbidities was restricted to the ASA score and cognitive impairment, while no information was available on some known risk factors for SSI, such as diabetes mellitus (Tsang and Gaston 2013), obesity (Zajonz et al. 2019), or smoking (Durand et al. 2013). Obesity, in par-

352

ticular, represents an unobserved confounder in the association between duration of surgery and SSI. However, for the association between early deep SSI and mortality the E-values indicate that the evidence for causality is rather robust (VanderWeele and Ding 2017). The variable “intraoperative complication” comprises a wide range of different events, which makes a detailed analysis impossible. While a DAG can help to decide which variables to include in an analysis, it will always represent a simplification of reality. In conclusion, our results indicate that an early deep SSI has a clinically significant impact on mortality in hip fracture patients and, hence, that the prevention of SSI should be seen as an essential aspect of hip fracture treatment. While we found no easily modifiable risk factors for early and delayed deep SSI in our cohort, we highly recommend adherence to the existing guidelines for the prevention of SSI (Ban et al. 2017). Additional measures, such as the use of high-dose dualimpregnated antibiotic-loaded bone cement in hemiarthroplasties (Sprowson et al. 2016) might be considered. We question the common wisdom that a longer duration of surgery in itself is closely associated with an increased risk of SSI and suggest that the underlying reason for a longer duration of surgery might be at least equally as important. Supplementary data Tables 2, 5, and 6 and Figures 3 and 4a are available as supplementary data in the online version of this article, http://dx.doi. org/10.1080/17453674.2020.1717841

CP conceived the study and drafted the manuscript. CP and FD performed the statistical analyses. All authors interpreted the results, contributed to the discussion, and reviewed the manuscript. The authors would like to thank Johan Inge Halse for his critical appraisal of this manuscript and Eva Hansen Dybvik for preparing the data set from the NHFR. Acta thanks Jon Goosen and Thord von Schewelov for help with peer review of this study. Ban K A, Minei J P, Laronga C, Harbrecht B G, Jensen E H, Fry D E, Itani K M F, Dellinger E P, Ko C Y, Duane T M. Executive dummary of the American College of Surgeons/Surgical Infection Society surgical site infection guidelines—2016 update. Surg Infect 2017; 18(4): 379-82. Barros A J, Hirakata V N. Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio. BMC Med Res Methodol 2003; 3: 21. Belmont P J, Garcia E S J, Romano D, Bader J O, Nelson K J, Schoenfeld A J. Risk factors for complications and in-hospital mortality following hip fractures: a study using the National Trauma Data Bank. Arch Orthop Trauma Surg 2014; 134(5): 597-604. Cheng H, Chen B P, Soleas I M, Ferko N C, Cameron C G, Hinoul P. Prolonged operative duration increases risk of surgical site infections: a systematic review. Surg Infect 2017; 18(6): 722-35.

Acta Orthopaedica 2020; 91 (3): 347–352

Cordero J, Maldonado A, Iborra S. Surgical delay as a risk factor for wound infection after a hip fracture. Injury 2016; 47(Suppl. 3): S56-S60. Daley B J, Cecil W, Clarke P C, Cofer J B, Guillamondegui O D. How slow is too slow? Correlation of operative time to complications: an analysis from the Tennessee Surgical Quality Collaborative. J Am Coll Surg 2015; 220(4): 550-8. Dalli J. Case definitions of communicable diseases and special health issues. Brussels: European Commission; 2012. Available from: https://www.fhi. no/globalassets/dokumenterfiler/helseregistre/nois/ecdcs-kasusdefinisjoner-fullstendig-engelsk-versjon-av-kasusdefinisjonene-av-smittsommesykdommer-august-2012-pdf-.pdf (Accessed: 4 September 2019). Davies H T, Crombie I K, Tavakoli M. When can odds ratios mislead? BMJ 1998; 316(7136): 989-91. de Jong L, Klem T, Kuijper T M, Roukema G R. Factors affecting the rate of surgical site infection in patients after hemiarthroplasty of the hip following a fracture of the neck of the femur. Bone Joint J 2017; 99-b(8): 1088-94. Durand F, Berthelot P, Cazorla C, Farizon F, Lucht F. Smoking is a risk factor of organ/space surgical site infection in orthopaedic surgery with implant materials. Int Orthop 2013; 37(4): 723-7. Edwards C, Counsell A, Boulton C, Moran C G. Early infection after hip fracture surgery: risk factors, costs and outcome. J Bone Joint Surg Br 2008; 90(6): 770-7. Gjertsen J E, Engesaeter L B, Furnes O, Havelin L I, Steindal K, Vinje T, Fevang J M. The Norwegian Hip Fracture Register: experiences after the first 2 years and 15,576 reported operations. Acta Orthop 2008; 79(5): 583-93. Harrison T, Robinson P, Cook A, Parker M J. Factors affecting the incidence of deep wound infection after hip fracture surgery. J Bone Joint Surg Br 2012; 94(2): 237-40. Horan T C, Andrus M, Dudeck M A. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control 2008; 36(5): 309-32. Lund C A, Moller A M, Wetterslev J, Lundstrom L H. Organizational factors and long-term mortality after hip fracture surgery: a cohort study of 6143 consecutive patients undergoing hip fracture surgery. PLoS One 2014; 9(6): e99308. Merrer J, Girou E, Lortat-Jacob A, Montravers P, Lucet J C. Surgical site infection after surgery to repair femoral neck fracture: a French multicenter retrospective study. Infect Control Hosp Epidemiol 2007; 28(10): 1169-74. Pollmann C T, Rotterud J H, Gjertsen J E, Dahl F A, Lenvik O, Aroen A. Fast track hip fracture care and mortality: an observational study of 2230 patients. BMC Musculoskelet Disord 2019; 20(1): 248. Sheikh H Q, Hossain FS, Aqil A, Akinbamijo B, Mushtaq V, Kapoor H. A comprehensive analysis of the causes and predictors of 30-day mortality following hip fracture surgery. Clin Orthop Surg 2017; 9(1): 10-8. Sprowson A P, Jensen C, Chambers S, Parsons N R, Aradhyula N M, Carluke I, Inman D, Reed M R. The use of high-dose dual-impregnated antibioticladen cement with hemiarthroplasty for the treatment of a fracture of the hip: the Fractured Hip Infection trial. Bone Joint J 2016; 98-b(11): 1534-41. Stocks G W, Self S D, Thompson B, Adame X A, O’Connor D P. Predicting bacterial populations based on airborne particulates: a study performed in nonlaminar flow operating rooms during joint arthroplasty surgery. Am J Infect Control 2010; 38(3): 199-204. Textor J, van der Zander B, Gilthorpe M S, Liśkiewicz M, Ellison G T. Robust causal inference using directed acyclic graphs: the R package ‘Dagitty’. Int J Epidemiol 2017; 45(6): 1887-94. Tsang S T, Gaston P. Adverse peri-operative outcomes following elective total hip replacement in diabetes mellitus: a systematic review and meta-analysis of cohort studies. Bone Joint J 2013; 95-b(11): 1474-9. VanderWeele T J, Ding P. Sensitivity analysis in observational research: introducing the E-value. Ann Intern Med 2017; 167(4): 268-74. Watne L O, Torbergsen A C, Conroy S, Engedal K, Frihagen F, Hjorthaug G A, Juliebo V, Raeder J, Saltvedt I, Skovlund E, Wyller T B. The effect of a pre- and postoperative orthogeriatric service on cognitive function in patients with hip fracture: randomized controlled trial (Oslo Orthogeriatric Trial). BMC Med 2014; 12:63. Zajonz D, Brand A, Lycke C, Ozkurtul O, Theopold J, Spiegl U J A, Roth A, Josten C, Fakler J K M. Risk factors for early infection following hemiarthroplasty in elderly patients with a femoral neck fracture. Eur J Trauma Emerg Surg 2019; 45(2): 207-1.

Acta Orthopaedica 2020; 91 (3): 353–359

353

Single-stage debridement with implantation of antibiotic-loaded calcium sulphate in 34 cases of localized calcaneal osteomyelitis Nan JIANG 1, 2, Xing-qi ZHAO 2, Lei WANG 1, Qing-rong LIN 1, Yan-jun HU 1, and Bin YU 1, 2 1 Department

of Orthopaedics & Traumatology, Nanfang Hospital, Southern Medical University, Guangzhou; 2 Guangdong Provincial Key Laboratory of Bone & Cartilage Regenerative Medicine, Nanfang Hospital, Southern Medical University, Guangzhou, P R China Correspondence: yubin@smu.edu.cn Submitted 2019-10-03. Accepted 2020-02-28.

Background and purpose — The successful eradication of calcaneus infection with limb salvage remains a challenge. We describe the outcomes of cortical bone windowing followed by eggshell-like debridement and implantation of antibiotic-loaded calcium sulphate (CS) for localized (Cierny–Mader type III) calcaneal osteomyelitis (CO). Patients and methods — We report a retrospective study of 34 patients. Infection followed trauma or orthopedic surgery in 30 patients and hematogenous spread in 4 patients. 31 patients had a sinus tract, accompanied by a soft tissue defect in 3 patients. All patients received cortical bone windowing, debridement, multiple sampling, local implantation of vancomycin- and gentamicin-loaded CS, skin closure or flap coverage, and culture-specific systematic antibiotic treatment in a single-stage procedure. Patients were followed up for a median of 26 months. Results — Infection was eradicated in 29 patients after the single-stage surgery, and all of the 5 recurrent infections were cleared by repeated surgery without amputation. Other adverse events included 11 patients with aseptic wound leakage and 1 unrelated death. Compared with those before surgery, the median postoperative scores of the American Orthopaedic Foot & Ankle Society (AOFAS) ankle hindfoot scale (65 vs. 86 vs. 89) and the visual analog scale (VAS) for pain (6 vs. 3 vs. 1) improved at the 1-year and 2-year follow-up. Interpretation — This single-stage protocol, cortical bone windowing, and eggshell-like debridement combined with local implantation of antibiotic-loaded CS is effective in treating type III CO. However, the incidence of aseptic wound leakage is high.

Calcaneal osteomyelitis (CO) is an uncommon condition, which usually occurs following trauma, orthopedic surgery, diabetic ulcers, and hematogenous spread (Fukuda et al. 2010, Mooney et al. 2017). Infrequent causes are iatrogenic steroid injection and acupuncture (Waibel et al. 2019). The primary goals of CO management are eradicating infection, adequate and durable soft tissue coverage, and maximal maintenance of the foot function (Merlet et al. 2014). The successful treatment of CO with limb salvage remains a challenge, primarily being attributable to the unique anatomical structure and function of the calcaneus, with limited surrounding soft tissue and blood supply. Clinical efficacy remains unsatisfying, with high risks of infection relapse and amputation (Merlet et al. 2014, Sabater-Martos et al. 2019, Waibel et al. 2019). Calcium sulphate (CS), a novel local antibiotic vehicle, has been widely used for the treatment of chronic osteomyelitis with satisfying outcomes (Gauland 2011, Ferguson et al. 2014, Andreacchio et al. 2019). Compared with polymethylmethacrylate (PMMA), CS can carry a wider range of antibiotics and is completely biodegradable, thus not requiring second surgery for removal (Inzana et al. 2016). Although previous studies had reported local antibiotic-loaded CS implantation for CO treatment, their strategies differed, including a 2-stage surgery of debridement followed by autologous bone graft (Papagelopoulos et al. 2006), the Silo technique with CS/hydroxyapatite (Drampalos et al. 2018), and even calcanectomy (Walsh and Yates 2013). Nonetheless, the clinical experience of bone-preserving strategy for CO treatment remains limited. To describe the extent of the inflammatory process and determine treatment strategy, Cierny–Mader (C–M) classification (Cierny and Mader 2003) has been proposed, which consists

354

of anatomic type (I: medullary, II: superficial, III: localized, and IV: diffuse) and physiologic class (host A: good immune system and delivery, host B: compromised locally (BL) or systemically (BS) or both (BLS), and host C: not a surgical candidate due to poor systemic condition and prognosis). Here, we present a series of patients with C–M type III CO treated with a bone-preserving surgical protocol, cortical bone windowing, eggshell-like debridement, and local antibioticloaded CS implantation in a 1-stage surgery.

Patients and methods Study design and inclusion and exclusion criteria This retrospective study was conducted in Nanfang Hospital of Southern Medical University, a tertiary medical center in Southern China. Included patients were those presenting with type III CO of the C–M anatomy classification, following injury, orthopedic surgery, or hematogenous spread, and having inflammatory symptoms for over 10 weeks (Metsemakers et al. 2018a). Type III CO refers to a local infection of the cortical and cancellous bone of the calcaneus, which does not affect the stability of the bone. 2 experienced surgeons made the C–M classification for each included patient independently and disputes were resolved by a third assessor. The establishment of CO diagnosis is based on at least 1 of the following confirmatory criteria (Metsemakers et al. 2018b): supportive histology, microbiological cultures from at least two suspected sites revealing the same pathogen, a definite sinus tract connecting directly the bone or the implant, or wound purulent drainage, or intraoperative pus. The exclusion criteria were diabetic foot-associated CO, patients with renal failure, calcium metabolism disorders, and a known allergy to CS, vancomycin, or gentamicin. In addition, patients who refused to receive this protocol were also excluded. Electronic medical records of 59 consecutive patients with chronic CO were screened between January 2013 and December 2017. However, 25 patients were excluded for the following reasons: 15 patients with type II, 6 with type IV, and 4 with type III, including 2 with diabetic foot and 2 with hypercalcemia. The remaining 34 patients (27 males) involving 34 limbs were included. The mean age at diagnosis was 41 years (3–67). The median follow-up time was 26 months (12–68). Preoperative management After admission, patients underwent routine physical examinations, imaging, and laboratory tests. Microbiology samples from the sinus tract were not taken. Antibiotics were stopped for at least 2 weeks prior to surgery. Surgical management After anesthesia, a tourniquet was routinely placed at the upper thigh. The surgical approach was selected based on the

Acta Orthopaedica 2020; 91 (3): 353–359

infection location and range, as suggested by the sinus tract location and/or radiographs/MRI scans. The sinus tract was first excised, then necrotic and devascularized soft tissues surrounding the infected calcaneus were debrided. Thereafter, the devitalized cortical bone was windowed and removed using an osteotome, to expose the calcaneus cancellous bone. A curette was used to remove the infected cancellous bone and its surrounding osseous tissues. However, the “calcaneus shape” was carefully protected during debridement and intraoperative radiographs were taken when necessary. After multiple samples were collected and sent for culture and histology, patients received empirical intravenous antibiotics (cephalosporins or alternatively clindamycin). Finally, the infected cancellous and surrounding bone were removed, leaving a cavity in the residual calcaneus as an eggshell. Subsequently, the cavity was soaked in 0.05% aqueous chlorhexidine solution 3 times for 5 minutes, followed by pulsed irrigation of 2–4 L of sterile saline solution. The cavity was dried and packed with gauze. After changing the gloves and drapes, the antibiotic-loaded CS was made with a mixture of vancomycin, gentamicin, and CS powder (Stimulan Rapid Cure; Biocomposites Ltd, Staffordshire, UK). The mixture ratio was 500 mg vancomycin with 5 mL CS dissolved in 2 mL/80 mg gentamicin with an additional 0.6 mL of sterile water. After shaping and solidification, the antibiotic-loaded CS block was filled into the cavity and the surrounding soft tissues. Primary skin closure was either achieved directly or the wound was covered by a local sural neurovascular flap or using the skin-stretching technique (Figure). Postoperative management After surgery, all of the patients were immobilized in a brace until the wound healed. During this period, patients received empirical antibiotics intravenously (cephalosporins or alternatively clindamycin), multimodal analgesia, and supportive treatment. Subsequent antibiotic selections were based on results of cultures and drug sensitivity outcomes of the samples taken intraoperatively. If the culture outcomes were negative, cephalosporins or clindamycin were continued. All patients received antibiotics intravenously for 2 weeks and orally for an additional 4 weeks. Rifampin and/or quinolones were added in patients suspected of having staphylococci and/or gram-negative bacteria-associated biofilm infections. Wound dressings were changed every 2–3 days. For patients who developed continuous drainage, the wound status was observed closely, serological levels of white blood cell count (WBC), erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP) were monitored, and multiple samples were cultured to exclude infection relapse. Patients received functional training with supervision. Partial weight-bearing was allowed after complete wound healing, and full weight-bearing was usually allowed 6–8 weeks after surgery.

Acta Orthopaedica 2020; 91 (3): 353–359

355

Patient with a sharp puncture injury history, who presented with a discharging sinus tract in the lateral side of the calcaneus (a). The axial radiograph indicates infection spreading to the rear and medial sides (b, red arrows). Therefore, we selected a semi-ring incision parallel to the pelma. First, the sinus tract was excised and the surrounding soft tissues were debrided (c). Second, the devitalized cortical bone was windowed and removed to expose the calcaneal cancellous bone (d). Third, a curette was used to remove the infected cancellous bone and its surrounding osseous tissues, leaving a cavity in the residual calcaneus, similar to an eggshell (e). Fourth, vancomycin- and gentamicin-loaded CS was implanted into the cavity and surrounding the soft tissue (f). Finally, the skin was closed primarily (g). A postoperative radiograph showed the locally implanted CS (h). The patient recovered well at 1-year follow-up (i and j).

Outcome parameters Outcome parameters included the infection eradication rate after the 1-stage procedure at 1-year follow-up, adverse events, the American Orthopaedic Foot & Ankle Society (AOFAS) ankle–hindfoot scale score, and the visual analog scale (VAS) for pain score (scale from 0 to 10) before surgery, at 1 year, and if possible at 2 years postoperatively. Statistics Descriptive statistics were conducted using the Statistical Product and Service Solutions (SPSS) 19.0 software (SPSS Inc., Chicago, IL, USA). Distributions of continuous variables were evaluated for normality using the Kolmogorov–Smirnov test initially. Then, data were expressed as mean (SD) and median (range) for normally and abnormally distributed variables, respectively. Ethics, funding, and potential conflicts of interest This study was approved by the Medical Ethics Committee of Nanfang Hospital, Southern Medical University. Informed consent from all included patients or their legal guardian was waived due to the retrospective design. Research was funded by a grant from the National Natural Science Foundation of China (grant number: 81802182). There is no conflict of interest to declare.

Results Patients According to the C–M host classification, 3 patients were type A and 31 were type B, including 26 patients as type BL, 1 as type BS, and the remaining 4 as type BLS hosts. 31 patients had at least 1 sinus tract, including 3 patients with accompanying soft tissue defect. There were 21 and 13 patients with CO on the left and right side of the body, respectively. 30 patients had posttraumatic osteomyelitis, whereas 4 had CO following hematogenous spread. 20 patients suffered from open fractures and 11 patients had infection after fracture fixation. Falling from a height was the most frequent injury type (10 patients), followed by a sharp puncture injury (7 patients) (Table 1). Inflammatory biomarkers and microbiology The median serological levels of WBC, ESR, and CRP were 7.2 (4.4–14) × 109/L, 16 (2–140) mm/h, and 4.7 (0.2–228) mg/L, with abnormal levels in 5/34, 16/33, and 14/31 patients, respectively. 14 of 34 patients had a positive pathogen culture outcome, with 12 patients having monomicrobial and 2 having polymicrobial infections. Pseudomonas aeruginosa (5 cases) was the most frequently detected bacteria, followed by Enterobacter

356

Acta Orthopaedica 2020; 91 (3): 353–359

Table 1. Clinical characteristics of the included patients No. Sex/age 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

M/38 M/60 M/29 F/41 M/50 M/64 F/51 M/52 M/10 M/67 M/63 M/45 F/67 M/53 M/54 F/51 M/56 M/30 M/20 M/10 M/33 M/12 M/25 F/60 M/36 F/59 F/65 M/3 M/22 M/29 M/43 M/18 M/48 M/46

a Sinus

Cause Iatrogenic factor Iatrogenic factor Falling injury Sharp puncture Crushing injury Falling from a height Falling from a height Traffic accident Sharp puncture Crushing injury Sharp puncture Falling from a height Sharp puncture Hematogenous spread Sharp puncture Falling from a height Frostbite injury Falling from a height Sharp puncture Hematogenous spread Traffic accident Sharp puncture Traffic accident Falling from a height Sprain injury Hematogenous spread Falling from a height Iatrogenic factor Falling from a height Falling from a height Falling injury Hematogenous spread Falling injury Falling from a height

Side R L R L L L L R L R L L R R L L R L L L R L L L R L L R L L L R R R

Culture outcome

Local comorbidity

Negative Sinus tract Stenotrophomonas maltophilia Sinus tract Pseudomonas aeruginosa Sinus tract Negative Sinus tract Staphylococcus aureus Sinus tract Polymicrobial Sinus tract a Negative Sinus tract Negative Sinus tract Pseudomonas aeruginosa Sinus tract Negative Sinus tract Pseudomonas aeruginosa No Escherichia coli Sinus tract Negative No Negative Sinus tract Negative Sinus tract Negative Sinus tract Negative Sinus tract Negative Sinus tract a Negative Sinus tract Enterobacter cloacae No Proteus mirabilis Sinus tract Pseudomonas aeruginosa Sinus tract Negative Sinus tract Staphylococcus aureus Sinus tract Enterobacter cloacae Sinus tract Polymicrobial Sinus tract Negative Sinus tract Negative Sinus tract Negative Sinus tract Negative Sinus tract Negative Sinus tract Negative Sinus tract Negative Sinus tract Pseudomonas aeruginosa Sinus tract a

tract + soft tissue defect

cloacae (2 cases) and Staphylococcus aureus (2 cases) (Table 1). All of the cultured microorganisms were sensitive to vancomycin and/or gentamycin. Infection eradication rate and adverse events The median CS volume implanted was 20 (5–40) mL. Skin closure was primarily achieved in 31 patients, with 2 covered by a local sural neurovascular flap and 1 by the skin-stretching technique. Infection was successfully eradicated in 29 patients following the 1-stage procedure at the 1-year follow-up. All 5 patients with recurrent infections received secondary debridement surgery, with 2 patients receiving wound irrigation, 2 receiving antibiotic-loaded CS implantation, and the remaining patient receiving local flap coverage. All of these 5 patients had recovered well at the median follow-up of 14 months (14–26). 11 patients had aseptic wound leakage, with a median duration of 31 days (16–52). Of these, 10 were managed with wound dressing and 1 underwent a secondary debridement surgery to remove the residual CS; all of these patients recovered well. In addition, 1 patient died of a cause unrelated to

CO or the surgery, as his wound and foot function had recovered well at the 1-year follow-up. Moreover, none of the patients had limb amputation or secondary fractures during follow-up (Table 2). AOFAS ankle–hindfoot scale score before and after surgery The median AOFAS ankle–hindfoot scale score of the 34 patients was 65 (39–82) before surgery, which increased to 86 (75–90) at the 1-year follow-up, and to 89 (78–90) at the 2-year follow-up after surgery (Table 2). VAS scores before and after surgery The median VAS score for pain was 6 (4–7) before surgery, which decreased to 3 (2–4) at the 1-year follow-up, and to 1 (0–2) at the 2-year follow-up postoperatively (Table 2).

Discussion

Successful treatment of CO with limb preservation remains a challenge for clinicians, not only because of the unique structure and function of the calcaneus, but also owing to the limited surrounding soft tissue coverage and poor blood supply. Previous studies had indicated that the risk of secondary below-knee amputation following limb-saving surgery in CO patients ranged from 4% to 20% (Baumhauer et al. 1998, Schade 2012, Van Riet et al. 2012, Walsh and Yates 2013, Babiak et al. 2016). Even a recent study (Waibel et al. 2019) showed that 50% of the CO patients had to undergo secondary below-knee amputation even if they had already received total calcanectomy. The optimal surgical strategy for CO remains controversial (Sabater-Martos et al. 2019). We have reported the outcomes of the cortical bone windowing followed by eggshell-like debridement and local vancomycin- and gentamicin-loaded CS for the treatment of patients with type III CO. Our findings can be summarized under the following 4 aspects. First, the infection eradication rate (29/34) after the singlestage surgery was similar to those of previous studies (Ferguson et al. 2014, Luo et al. 2016, Ferguson et al. 2017). Whether bone infection relapses or not after treatment is influenced by multiple factors, such as surgical and antibiotic strategies, pathogen species and virulence, and host immune status. Among these factors, debridement plays a vital role, which should not be limited by concerns of repairing osseous/soft tissue defects (Metsemakers et al. 2018a). The goal

Acta Orthopaedica 2020; 91 (3): 353â&#x20AC;&#x201C;359

357

ogous bone graft surgery, whereas in our study we only performed 1-stage surgery, facilitated by CS AOFAS score VAS pain score 1 year 2 years 1 year 2 years being totally biodegradable and Before after after Before after after not requiring a second surgery. It No. Sex/age Adverse events surgery surgery surgery surgery surgery surgery should be noted that, in this study, 1 M/38 Aseptic wound leakage 72 87 89 6 2 0 we included only non-diabetic foot 2 M/60 Nil 65 87 90 5 3 0 patients, which may be the primary 3 M/29 Nil 39 84 90 7 3 2 reason for our much lower ampu4 F/41 Nil 43 80 80 7 3 1 5 M/50 Aseptic wound leakage 65 87 89 6 3 2 tation rate compared with those of 6 M/64 Aseptic wound leakage 68 90 6 3 previous studies. 7 F/51 Infection relapse 58 86 6 3 Second, incidence of the adverse 8 M/52 Aseptic wound leakage 65 82 84 6 4 1 9 M/10 Aseptic wound leakage 68 90 90 6 2 1 events apart from infection relapse 10 M/67 Aseptic wound leakage 62 86 6 2 was 12/34, with aseptic wound 11 M/63 Unrelated death 72 86 5 2 leakage being the most fre12 M/45 Nil 72 89 90 6 3 1 13 F/67 Nil 82 86 87 6 4 2 quently observed. Previous stud14 M/53 Nil 61 84 87 6 2 1 ies reported that the proportion 15 M/54 Aseptic wound leakage 61 87 87 6 2 0 of aseptic wound leakage ranged 16 F/51 Nil 49 90 90 7 2 0 17 M/56 Nil 68 87 87 6 4 1 from 6/100 (McNally et al. 2016) 18 M/30 Aseptic wound leakage 65 86 6 3 to 7/21 (Humm et al. 2014), show19 M/20 Nil 79 90 90 4 3 0 ing an average incidence of 16% 20 M/10 Nil 51 90 90 6 2 0 21 M/33 Nil 46 84 7 3 (Ferguson et al. 2017), whereas, it 22 M/12 Nil 73 90 90 5 2 0 was 11/34 in our patients, higher 23 M/25 Infection relapse 72 84 84 7 3 2 than most of the previous studies. 24 F/60 Infection relapse 65 85 90 5 4 2 25 M/36 Nil 81 90 4 2 The difference may be primarily 26 F/59 Nil 72 84 6 2 attributed to 2 reasons. 1 is owing 27 F/65 Infection relapse 45 75 6 4 to the limited soft tissue cover28 M/3 Nil 71 90 6 2 29 M/22 Aseptic wound leakage 52 78 78 6 4 2 age as well as blood supply to the 30 M/29 Nil 56 84 6 2 calcaneus physiologically. In the 31 M/43 Infection relapse 68 86 6 3 case of bone infection, the local 32 M/18 Aseptic wound leakage 75 90 4 2 33 M/48 Nil 58 84 5 3 soft tissue might worsen, which 34 M/46 Aseptic wound leakage 58 89 6 2 aggravates the blood circulation locally. Thus, there is an increased AOFAS: American Orthopaedic Foot & Ankle Society; VAS: visual analog scale. risk of wound problem. The other reason may be related to the limof infection excision is to remove all the devitalized tissues, ited sample size of our study. To evaluate more accurately leaving behind healthy vascularized bone and managing the whether wound leakage risk after CS local use in the calcasubsequent dead space to prevent re-accumulation of hema- neus is higher than with CS implanted in other bones, studtoma that may become re-infected. It is reasonable to under- ies with larger sample sizes are warranted. In addition to the stand that the protocols for CO treatment include partial and wound leakage, local CS implantation may also bring other total calcanectomy, or even below-knee amputation (Waibel problems, such as heterotopic ossification (HO) (Kallala et al. et al. 2019). Although infection can be eradicated follow- 2018) and even hypercalcemia (Kallala and Haddad 2015), ing such radical surgeries, the foot function may be more or which appears to be more frequently found in patients with less impaired. The outcomes of a recent study (Waibel et al. prosthetic joint infections. Third, we found that, compared with those before sur2019) including 50 patients indicated that the secondary reamputation proportions of the CO patients following partial/ gery, the median AOFAS ankleâ&#x20AC;&#x201C;hindfoot scale and VAS total calcanectomy and below-knee amputation were 8/28, scores for pain improved at the 1-year and 2-year follow2/4, and 1/18, respectively, implying the potential problems ups, demonstrating the restoration of foot function and the of such surgical strategies. Considering the importance of the alleviation of pain. Such changes might be associated with weight-bearing function of the calcaneus, we tried to main- absolute infection eradication as well as preservation of tain the integrity of the external cortical shell of the calcaneus calcaneus integrity. In a recent study, Babiak et al. (2016) during debridement, similar to the approach of a previous compared the efficacy between bone-preserving strategies report (Papagelopoulos et al. 2006). However, they used a (debridement or drilling and collagen-gentamicin sponge) PMMA cement carrier and, thus, performed secondary autol- and radical surgical protocols (partial or total calcanecTable 2. Clinical efficacy of the included patients

358

tomy) for CO treatment. Outcomes revealed that the proportion of patients who were ambulatory before therapy retained their walking ability after the bone-preserving surgeries was higher than that of the patients receiving radical surgical interventions, suggesting that, on the premise of infection eradication, calcaneus integrity should be preserved as much as possible. Fourth, we found several unique characteristics of our patients. Initially, the proportion of patients with a sinus tract was much higher, demonstrating compromised local status in most CO patients. Additionally, even sharp puncture injury and iatrogenic injection or acupuncture may lead to calcaneal infection, although such causes are rare (Waibel et al. 2019). Despite the 2-week non-administration of antibiotics and the high number of patients with a drainage sinus tract, the positivity rate of the culture in our study was low, which may be associated with the suboptimal culture conditions. The most frequently detected pathogen in this cohort was Pseudomonas aeruginosa, which may be related to the humid environment surrounding the heel. Our study has 2 main limitations. First, the sample size is limited, therefore the results should be interpreted cautiously, and more patients should be recruited in future studies to obtain more precise conclusions. Second, we did not analyze the risk factors of aseptic wound leakage and infection relapse because of the study’s retrospective design. To better identify potential risk factors, well-designed studies are necessary. However, this study provides a novel bonepreserving surgical strategy for CO treatment, with satisfying efficacy. In conclusion, cortical bone windowing plus eggshell-like radical debridement with local implantation of antibioticloaded CS is an effective way to treat type III CO. However, patients receiving this treatment should be fully informed regarding the potential risk of aseptic wound leakage after local CS implantation. Future investigations may focus on the risk factors of aseptic wound leakage and infection relapse following local CS use as well as the efficacy of other substitute materials that could be implanted locally, such as bioactive glass.

NJ and XQZ contributed equally to this study. NJ, YJH, and BY designed the study. XQZ and QRL screened the medical record for patient inclusion. XQZ and LW collected the patient data. NJ and YJH conducted the data analysis. NJ and BY drafted the manuscript. All authors reviewed and revised the manuscript. The authors are grateful for the funding support from the National Natural Science Foundation of China and also would like to thank Editage for language editing and revision. Acta thanks Trude Basso and Martin McNally for help with peer review of this study.

Acta Orthopaedica 2020; 91 (3): 353–359

Andreacchio A, Alberghina F, Paonessa M, Cravino M, De Rosa V, Canavese F. Tobramycin-impregnated calcium sulfate pellets for the treatment of chronic osteomyelitis in children and adolescents. J Pediatr Orthop B 2019; 28(3): 189-95. Babiak I, Pedzisz P, Kulig M, Janowicz J, Maldyk P. Comparison of bone preserving and radical surgical treatment in 32 cases of calcaneal osteomyelitis. J Bone Jt Infect 2016; 1: 10-6. Baumhauer J F, Fraga C J, Gould J S, Johnson J E. Total calcanectomy for the treatment of chronic calcaneal osteomyelitis. Foot Ankle Int 1998; 19(12): 849-55. Cierny G 3rd, Mader J T, Penninck J J. A clinical staging system for adult osteomyelitis. Clin Orthop Relat Res 2003; (414): 7-24. Drampalos E, Mohammad H R, Kosmidis C, Balal M, Wong J, Pillai A. Single stage treatment of diabetic calcaneal osteomyelitis with an absorbable gentamicin-loaded calcium sulphate/hydroxyapatite biocomposite: the Silo technique. Foot (Edinb) 2018; 34: 40-4. Ferguson J Y, Dudareva M, Riley N D, Stubbs D, Atkins B L, McNally M A. The use of a biodegradable antibiotic-loaded calcium sulphate carrier containing tobramycin for the treatment of chronic osteomyelitis: a series of 195 cases. Bone Joint J 2014; 96-B(6): 829-36. Ferguson J, Diefenbeck M, McNally M. Ceramic biocomposites as biodegradable antibiotic carriers in the treatment of bone infections. J Bone Jt Infect 2017; 2(1): 38-51. Fukuda T, Reddy V, Ptaszek A J. The infected calcaneus. Foot Ankle Clin 2010;15(3): 477-86. Gauland C. Managing lower-extremity osteomyelitis locally with surgical debridement and synthetic calcium sulfate antibiotic tablets. Adv Skin Wound Care 2011; 24(11): 515-23. Humm G, Noor S, Bridgeman P, David M, Bose D. Adjuvant treatment of chronic osteomyelitis of the tibia following exogenous trauma using OSTEOSET®-T: a review of 21 patients in a regional trauma centre. Strategies Trauma Limb Reconstr 2014; 9(3): 157-61. Inzana J A, Schwarz E M, Kates S L, Awad H A. Biomaterials approaches to treating implant-associated osteomyelitis. Biomaterials 2016; 81: 58-71. Kallala R, Haddad F S. Hypercalcaemia following the use of antibiotic-eluting absorbable calcium sulphate beads in revision arthroplasty for infection. Bone Joint J 2015; 97-B(9): 1237-41. Kallala R, Harris W E, Ibrahim M, Dipane M, McPherson E. Use of Stimulan absorbable calcium sulphate beads in revision lower limb arthroplasty: safety profile and complication rates. Bone Joint Res 2018; 7(10): 570-9. Luo S, Jiang T, Yang Y, Yang X, Zhao J. Combination therapy with vancomycin-loaded calcium sulfate and vancomycin-loaded PMMA in the treatment of chronic osteomyelitis. BMC Musculoskelet Disord 2016; 17(1): 502. McNally M A, Ferguson J Y, Lau A C, Diefenbeck M, Scarborough M, Ramsden A J, Atkins B L. Single-stage treatment of chronic osteomyelitis with a new absorbable, gentamicin-loaded, calcium sulphate/hydroxyapatite biocomposite: a prospective series of 100 cases. Bone Joint J 2016; 98-B(9): 1289-96. Merlet A, Cazanave C, Dauchy F A, Dutronc H, Casoli V, Chauveaux D, De Barbeyrac B, Dupon M. Prognostic factors of calcaneal osteomyelitis. Scand J Infect Dis 2014; 46(8): 555-60. Metsemakers W J, Kuehl R, Moriarty T F, Richards R G, Verhofstad M H J, Borens O, Kates S, Morgenstern M. Infection after fracture fixation: current surgical and microbiological concepts. Injury 2018a; 49(3): 511-22. Metsemakers W J, Morgenstern M, McNally M A, Moriarty T F, McFadyen I, Scarborough M, Athanasou N A, Ochsner P E, Kuehl R, Raschke M, Borens O, Xie Z, Velkes S, Hungerer S, Kates S L, Zalavras C, Giannoudis P V, Richards R G, Verhofstad M H J. Fracture-related infection: a consensus on definition from an international expert group. Injury 2018b; 49(3): 505-10. Mooney M L, Haidet K, Liu J, Ebraheim N A. Hematogenous calcaneal osteomyelitis in children. Foot Ankle Spec 2017; 10(1): 63-8. Papagelopoulos P J, Mavrogenis A F, Tsiodras S, Vlastou C, Giamarellou H, Soucacos P N. Calcium sulphate delivery system with tobramycin for the treatment of chronic calcaneal osteomyelitis. J Int Med Res 2006; 34(6): 704-12. Sabater-Martos M, Sigmund I K, Loizou C, McNally M. Surgical treatment and outcomes of calcaneal osteomyelitis in adults: a systematic review. J Bone Jt Infect 2019; 4(3): 146-54.

Acta Orthopaedica 2020; 91 (3): 353â&#x20AC;&#x201C;359

Schade V L. Partial or total calcanectomy as an alternative to below-the-knee amputation for limb salvage: a systematic review. J Am Podiatr Med Assoc 2012; 102(5): 396-405. Van Riet A, Harake R, Stuyck J. Partial calcanectomy: a procedure to cherish or to reject? Foot Ankle Surg 2012; 18(1): 25-9.

359

Waibel F W A, Klammer A, Gotschi T, Uckay I, Boni T, Berli M C. Outcome after surgical treatment of calcaneal osteomyelitis. Foot Ankle Int 2019; 40(5): 562-7. Walsh T P, Yates B J. Calcanectomy: avoiding major amputation in the presence of calcaneal osteomyelitisâ&#x20AC;&#x201D;a case series. Foot (Edinb) 2013; 23(4): 130-5.

360

Acta Orthopaedica 2020; 91 (3): 360–361

Correspondence

Postoperative mortality after a hip fracture over a 15-year period in Denmark: a national register study Sir,—I have, with great interest, read the study by Gundel et al. (2020) that gives national results on hip fracture care in Denmark. Given the high-quality national registers in Denmark, my expectations were high. The fact that the 1-yearmortality was found to be high, 27%, and unchanged during a period of 15 years increased my attention further. Regrettably, my enthusiasm died down (no pun intended). The authors did not discuss any reason for the high mortality in terms of public health, quality of hospital and elderly care, psychological factors post-trauma leading to loss of self-preservation, etc. A Western country like Denmark with easily accessible public healthcare ought to have a lower mortality rate after hip fracture, in particular when fracture patients of all ages are included. The same mortality rate was found when studying only Finnish patients over 65 years (Panula et al. 2011). A 2011 urban cohort of all hip fracture patients over 20 years, in Malmö, Sweden, had a mortality rate of 24% (Hansson et al. 2015). A recent meta-analysis suggested 22% to be an expected mortality rate (Downey et al. 2019). Instead, the authors choose to investigate whether surgical methods have any association with the risk of death. They found “operation type other than total hip arthroplasty was … associated with postoperative mortality.” Being one of the authors cited in this question, I would like to stress the pronounced selection bias when using surgical methods as predictors. In our paper (Hansson et al. 2019), we did not discuss the difference in mortality between total hip arthroplasty and hemiarthroplasty as a clinically relevant result. We clearly explained this finding as a result of confounding and nothing else. Drawing conclusions and clinical perspectives on THA being associated with a reduced risk of dying reveals a lack of orthopedic experience in the author group. THA is recommended for healthier, active patients with good chances of long-term survival. This is only touched upon in the Danish paper. In addition, when comparing different operations, the authors have chosen “closed reposition” as reference. The options compared with closed reduction are “open reposition, external fixation, internal fixation, hemiarthroplasty, arthroplasty” and “other.” It makes no sense to use closed reposition either as a procedure of its own, or as the reference. Surgical treatment of a hip fracture starts with either closed or open reduction and is then followed by internal fixation.

Arthroplasty is the other main alternative. External fixation is extremely uncommon, here used in less than 1 per 1,000 of the patients, and may represent either multi-trauma cases or miscoding. Table 2 rewritten with internal fixation as reference would have provided the reader with relevant comparisons. The authors state: “Further studies are needed to shed light on whether the type of operative procedure influences the postoperative mortality following HF surgery and whether the prognosis of specific comorbid elderly patients can be modulated by choosing procedures with less trauma and thus less surgical stress.” This has been a major aim for numerous studies during the last 25 years (Rogmark and Johnell 2006), comparing the more strenuous arthroplasty procedure with a quick pinning or screwing of a displaced femoral neck fracture. Arthroplasty led to less pain, better function and healthrelated quality-of-life, and fewer reoperations, without any clear risk of increased mortality. Sometimes, more is more. Besides pointing out that those who are men, sick, and old have a worse prognosis after a hip fracture – well-known factors (Dahl 1980) – the paper is a good example of how an epidemiology paper has to be written in cooperation with researchers who are in clinical practice and with knowledge of the disease-specific literature. I hope the Köge group will continue to study outcomes after hip fracture, to the benefit of elderly Danes, and to improve Danish healthcare further. When doing so, I hope they involve co-workers with specific clinical expertise. Cecilia Rogmark Skane University Hospital/Lund University Malmö, Sweden E-mail: cecilia.rogmark@skane.se

Sir,—Thank you for the opportunity to comment on the letter from Associate Professor Cecilia Rogmark, regarding our paper “Postoperative mortality after a hip fracture over 15 years in Denmark: a national register study.” We appreciate that Dr Rogmark took her time to read and comment on our paper and the results, and the scientific debate about how to improve postoperative mortality. Below are our comments to the remarks by Dr Rogmark.

Acta Orthopaedica 2020; 91 (3): 360–361 361

Our primary aim with this study was to investigate whether there had been any change in the mortality of Danish hip fracture patients over a period of 15 years. These data have not previously been reported for Danish hip fracture patients. With the focus on enhanced recovery after surgery in other surgical patient groups, we see our data as relevant in the debate about lowering postoperative mortality and complications. With these epidemiological data, we are not able to draw any certain conclusion as to why mortality has not changed over time. We acknowledge that the quality of hospital and elderly care could likely influence mortality over a 15-year period, which could have been interesting to include in the analysis. Other clinical factors with potential importance to the outcome after hip fracture surgery are early epidural/type of anesthesia (Van Waesberghe et al. 2017) and time to surgery. Unfortunately, these data were not available in the study. Moreover, we agree that we could have done an extended analysis on the potential change in number of comorbidities in the population over these 15 years. As life expectancy increases and patients are living with more comorbidities (World Health Organization 2015), this could partly explain why we did not see a decrease in postoperative mortality over the 15-year period. Regarding the type of operation, we were aware of the potential confounding by indication. We did address this in the discussion when interpreting our results. We acknowledge that the type of surgery is chosen based on type of fracture, patient age, comorbidities, and daily level of function in order to give the patient the best outcome (Claus et al. 2008). In order to reduce confounding by indication, we adjusted for age and comorbidities in our multivariate analysis. Despite this adjustment, we still observed a statistically significant association between type of surgery and mortality. In this study, we chose to display the primary and first surgical intervention for hip fracture patients. However, we recognize that for patients with reposition as the first surgical intervention, the secondary surgical intervention could have been included in our analysis. Regarding choice of reference group in Table 2, we do not find it essential to change this for type of operation since hazard ratios and 95% confidential intervals are stated for each type of operation.

We acknowledge the extensive research on hip fracture surgery that has been done in the past 25 years, as pointed out by Dr Rogmark. With our paper, we bring epidemiological data on outcome after hip fractures over a 15-year period from Denmark into the broad discussion regarding enhanced recovery after surgery. Our goal is to encourage a debate on how to improve recovery and mortality after surgery. Ossian Gundel Center for Surgical Science, Department of Surgery, Zealand University Hospital, Koege, Denmark E-mail: qrb124@alumni.ku.dk

Dahl E. Mortality and life expectancy after hip fractures. Acta Orthop Scand 1980; 51(1-6): 163-70. Downey C, Kelly M, Quinlan J F. Changing trends in the mortality rate at 1-year post hip fracture-a systematic review. World J Orthop 2019; 10(3): 166. Gundel O, Thygesen L C, Gögenur I, Ekeloef S. Postoperative mortality after a hip fracture over a 15-year period in Denmark: a national register study. Acta Orthop 2020; 91(1): 58-62 Hansson S, Rolfson O, Åkesson K, Nemes S, Leonardsson O, Rogmark C. Complications and patient-reported outcome after hip fracture: a consecutive annual cohort study of 664 patients. Injury 2015; 46(11): 2206-11. Hansson S, Bülow E, Garland A, Kärrholm J, Rogmark C. More hip complications after total hip arthroplasty than after hemiarthroplasty as hip fracture treatment: analysis of 5,815 matched pairs in the Swedish Hip Arthroplasty Register. Acta Orthop 2019; Nov 18: 1-6. doi: 10.1080/17453674.2019.1690339. [Epub ahead of print] Jensen C M, Mainz H, Lamm M, Foss N B, Villadsen B, Ovesen O, Lind B, Lauritzen J B, Kirk G, Kristensen M T. Referenceprogram for Patienter med Hoftebrud. 2008. https://www.ortopaedi.dk/guidelines-2/referenceprogrammer/Referenceprogram_for_patienter_med_hoftebrud2008.pdf Panula J, Pihlajamäki H, Mattila V M, Jaatinen P, Vahlberg T, Aarnio P, Kivelä S L. Mortality and cause of death in hip fracture patients aged 65 or older: a population-based study. BMC Musculoskelet Disord 2011; 12(1): 105. Rogmark C, Johnell O. Primary arthroplasty is better than internal fixation of displaced femoral neck fractures: a meta-analysis of 14 randomized studies with 2,289 patients. Acta Orthop 2006; 77(3): 359-67. Van Waesberghe J, Stevanovic A, Rossaint R, Coburn M. General vs. neuraxial anaesthesia in hip fracture patients: a systematic review and metaanalysis. BMC Anesthesiol 2017; 17(1): 87. World Health Organization. WHO world report on ageing and health 2015. Geneva: World Health Organization; 2015.

362

Acta Orthopaedica 2020; 91 (3): 362

Erratum Deep learning in fracture detection: a narrative review Pishtiwan H S KALMET 1*, Sebastian SANDULEANU 2*, Sergey PRIMAKOV 2, Guangyao WU 2, Arthur JOCHEMS 2, Turkey REFAEE 2, Abdalla IBRAHIM 2–5, Luca v. HULST 1, Philippe LAMBIN 2,3, and Martijn POEZE 1,6 Acta Orthopaedica 2020; 91 (2): 215–220. DOI 10.1080/17453674.2019.1711323 Correction of affiliations: * Shared first authorship 1 Maastricht University Medical Center+, Department of Trauma Surgery, Maastricht; 2 The D-Lab: Decision Support for Precision Medicine, GROW— School for Oncology and Developmental Biology, Maastricht University Medical Center+, Maastricht; 3 Department of Radiology and Nuclear Medicine, GROW – School for Oncology and Developmental Biology, Maastricht University Medical Centre+, Maastricht, The Netherlands; 4 Division of Nuclear Medicine and Oncological Imaging, Department of Medical Physics, Hospital Center Universitaire De Liege, Liege, Belgium; 5 Department of Nuclear Medicine and Comprehensive diagnostic center Aachen (CDCA), University Hospital RWTH Aachen University, Aachen, Germany. 6 Nutrim School for Nutrition, Toxicology and Metabolism, Maastricht University, Maastricht, The Netherlands Correspondence: pishtiwan.kalmet@mumc.nl

Correction of conflict of interest statement Original statement: The authors declare that they have no conflict of interest. Corrected version: Conflict of interest The authors, except dr. Philippe Lambin, declare that they have no conflict of interest. Dr. Lambin reports, within and outside the submitted work, grants/sponsored research agreements from Varian medical, Oncoradiomics, ptTheragnostic, Health Innovation Ventures and DualTpharma. He received an advisor/ presenter fee and/or reimbursement of travel costs/external grant writing fee and/or in kind manpower contribution from Oncoradiomics, BHV, Merck and Convert pharmaceuticals. Dr Lambin has shares in the company Oncoradiomics SA and Convert pharmaceuticals SA and is co-inventor of two issued patents with royalties on radiomics (PCT/NL2014/050248, PCT/NL2014/050728) licensed to Oncoradiomics and one issue patent on mtDNA (PCT/EP2014/059089) licensed to ptTheragnostic/DNAmito, three non-patentable invention (softwares) licensed to ptTheragnostic/DNAmito, Oncoradi omics and Health Innovation Ventures.

Acta Orthopaedica 2020; 91 (3): 363

363

Corrigendum Consensus statement for perioperative care in total hip replacement and total knee replacement surgery: Enhanced Recovery After Surgery (ERAS®) Society recommendations Thomas W WAINWRIGHT et al. Correspondence: twainwright@bournemouth.ac.uk Acta Orthopaedica 2020; 91 (1): 3–19. DOI 10.1080/17453674.2019.1683790

Further to the publication of our article, in light of recent research (Kumin et al. 2019), and the results of forth coming trials, the wording used in regard to the non recommendation of forced air warming devices is currently too strong. We would therefore like to remove the following sentence and references from “Maintaining normothermia,” (page 10) “However, the use of forced air-warming is not recommended as there is evidence that this is associated with an increased risk of infection (McGovern et al. 2011, Koc et al. 2017)”. We provide an updated below. Maintaining normothermia The National Institute for Clinical Excellence (NICE) recommends the pre-warming of patients and to maintain the active warming of all adults undergoing surgery throughout the intraoperative phase (NICE, 2016). Multiple series suggest that normothermia should be targeted as part of the anesthetic care of hip and knee replacement patients. There are many methods described to conserve body temperature, including pre-warming and humidification of anesthetic gases, warm-

ing IV and irrigation fluids and forced air warming blankets and devices. In addition, the ambient temperature should be at least 21°C while the patient is exposed prior to active warming starting (NICE, 2016). Summary and recommendations—Normal body temperature should be maintained peri- and postoperatively through pre-warming and the active warming of patients intraoperatively Evidence level—High Recommendation grade—Strong

Kümin M, Deery J, Turney S, Price C, Vinayakam P, Smith A, Filippa A, Wilkinson-Guy L, Moore F, O’Sullivan M, Dunbar M, Gaylard J, Newman J, Harper CM, Minney D, Parkin C, Mew L, Pearce O, Third K, Shirley H, Reed M, Jefferies L, Hewitt-Gray J, Scarborough C, Lambert D, Jones CI, Bremner S, Fatz D, Perry N, Costa M, Scarborough M.. Reducing Implant Infection in Orthopaedics (RIIiO): Results of a pilot study comparing the influence of forced air and resistive fabric warming technologies on postoperative infections following orthopaedic implant surgery. J Hosp Infect 219; 103(4): 412-9.

Acta Orthopaedica 2020; 91 (3)

Information to authors

Open Access/ Article Publication Charge Acta Orthopaedica, owned by the Nordic Orthopaedic Federation, is a non-profit, immediate Open Access journal, meaning your work is free for every-one to access online as soon as it is published (Gold OA). The standard research article publishing charge (APC) is EURO 1,750 (USD 2,000/GBP 1,520). For Case Reports the APC is EURO 875 (USD 1,000/GBP 760). Depending on your location these charges may be subject to local taxes. Note: There is no APC for members in the national orthopedic associations who are paying members of the Nordic Orthopaedic Federation (Denmark, Estonia, Finland, Iceland, Lithuania, Norway, and Sweden) Submission of manuscripts Manuscripts should be submitted online (https://www.manuscriptmanager.net/ao). Submit a double line spaced and line numbered Word file including tables and figures at the end of the manuscript. Authors are encouraged to provide the names, addresses, and e-mail of potential reviewers. Note the following points: a. A manuscripts total word count should not exceed 3,300 (including Abstract (max. 250 words) but excluding title page, figure and table legends, and references). The word count should be added on the title page. American spelling is preferred. b. Authors submitting a paper do so on the understanding that it has not been published and is not being considered for publication elsewhere. The authors should provide a statement about previous publications that are similar to the submitted study. Copies of such studies should be submitted with the paper. c. Interventional studies (defined by human participants being prospectively assigned to one or more health-related interventions to evaluate the effects on health outcomes) should be registered in a public trials register before the enrolment of the first participant. The registration ID should be presented under ‘Ethics’. In addition, manuscripts presenting randomized trials should comply with the CONSORT statement (www.consort-statement.org). The study protocol, approved by the appropriate ethics committee, a completed CONSORT checklist and flowchart should be submitted together with the manuscript. Animal in vivo experiments should comply with ARRIVE and a completed checklist (www.nvcr3s.org.uk/ arrive-guidelines) and a copy of the study protocol should be submitted together with the manuscript. d. Systematic reviews and meta-analyses of interventional studies should comply with the PRISMA statement (www. prisma-statement.org), and the PRISMA checklist and flow-

chart should be appended to the manuscript. For systematic reviews and meta-analyses of observational studies the MOOSE checklist and flowchart (www.equator-network. org) should be used. e. Manuscripts presenting results on diagnostic accuracy should comply with the STARD-statement (www.stardstatement.org), and the STARD checklist and flowchart should be appended to the manuscript. f. Reports of observational studies should comply with the STROBE-statement (www.strobe-statement.org) and the STROBE checklist should be appended to the manuscript. g. Extensive tables, complementary figures, and other explanatory material should be given as Supplementary data which will be available in the online version of the article. h. All original data of relevance to the submitted manuscript should be stored and be retrievable on request for a minimum of 10 years. Authors are encouraged to make these data available via the internet. i. Acta does not publish studies that only assess completeness, coverage, accuracy, and validity of data in registers. Such studies are considered part of normal register administration, and the results should be presented in the register’s annual report. j. Acta does not publish studies reporting merely the existence or organization of a new register. Furthermore, to qualify for publication an epidemiologic/register study must include a question assessed by use of the register data, for example: outcome related to different treatments, treatment changes over time, or incidence changes over time. Manuscript layout Title. The title should include information on the results of the investigation, including the number of patients, average follow-up, and animal or cadaver experiments. The first name, middle initial(s), and last name of each author should be given with department affiliations as well as the name and e-mail of the corresponding author. Abstract. The abstract should not exceed 250 words. It should consist of 4 parts: Background and purpose, Patients/ material/animals and methods, Results, and Interpretation. Introduction. The introduction should focus on the state of knowledge at the beginning of the study. The aims and main hypothesis of the study should be stated clearly. Only in exceptional cases should it exceed 300 words. Case reports. These should be short, include the Case history and Discussion without an Introduction. Patients/material/animals and methods. The selection of subjects and the inclusion or exclusion criteria should be described. Subjects who declined to participate, withdrawals

Acta Orthopaedica 2020; 91 (3)

and those with incomplete follow-up should be accounted for. Describe in detail how the measurements were made and the techniques used. Statistics. All statistical methods should be mentioned. Unusual methods should be referenced. The tests should be 2-sided, unless the use of 1-sided tests is justified. No data should be removed, imputed, weighted, adjusted or trimmed unless they are specifically described and justified and its consequences are given. Hypothesis tests (p-values). These tests should be used, in combination with a defined effect size and when statistical power has been considered. Give p-values with real numbers, if these are greater than 0.001 (1 digit except zeros), otherwise use ‘p < 0.001’. Do not use ‘ns’, ‘p > 0.05’ or asterisks. In most instances, the 95% confidence intervals should be given, especially in exploratory analyses and when estimating the effects or differences. Ethics, registration, funding, and potential conflicts of interest. Approval by the local ethics committee should be stated with date of issue and registration number. Clinical trials should have been registered in a public trials register before start of the trial. Give registration number. Funding and potential conflicts of interest should be declared. Results. Do not give the same data in more than one way. When summarizing the data, always include measures of variability and the number of subjects. Give the median and range—e.g., 60 years (35–70), the mean and standard deviation—59 (SD 15) years, and the frequencies for nominal data. The results of matched data should be given in relevant form (e.g., the distribution of pairwise differences). Percentages should not be used if the sample size is less than 100. Discussion. This section should contain a critical discussion of the results—e.g., the quality of the data (selection and information bias) and adequacy of the statistical analysis (confounding bias). It should also assess the relevant literature for or against the findings and if possible, the conclusions as regards clinical application or further research. Discuss, do not recapitulate, your results. Tables. Each table should be self-explanatory, with an adequate title and a logical presentation of the data. Abbre-

viate words in the columns and explain in footnotes. Each column heading for numerical data should include the unit of measurement. Use SI units. Avoid unnecessary decimals! It is seldom advisable to use more than 2 digits for biologic measurements. Digital illustrations should be used. A scanned picture or digital photo should usually be about 85 mm (i.e. 1000 pixels) wide with a resolution of 300 DPI (dots per inch), which gives a square picture an uncompressed file size of 4.5 MiB in color and 1 MiB in gray-scale. A black and white drawing or graph may be scanned in 800 DPI bitmap, i.e. 1 bit TIFF. The preferred format is a TIFF-file. The resolution of WEB illustrations (gif) is usually too poor. Color photos should be in CMYK colors. The cost of color will be paid by Acta. Highlights, arrows and letters may be added to digital photos but keep a clean version of the photo for the layout process. Acta prefers to add these as a separate layer. For digital graphs, use a graphics program that can export vectorized graphs as EMF, EPS or PDF files. Avoid frames around diagrams, diagrams with perspective drawing, and bar graphs or histograms (use tables). Symbols should be consistent throughout a series of figures. Each axis should be horizontally labeled, with a description of the variable it represents. Axes should be equal in length to make the diagrams quadratic. Use sans serif letters (Arial or Helvetica). Make diagrams in color but avoid complex patterns. Author contributions. Describe in short what each author did (initials). Acknowledgments. Technical help or other help may be acknowledged. References. The number of references typically does not exceed 25 (except for meta-analyses or review articles). We prefer the references to be cited by name and year (chronologically) of publication in the text instead of giving their numbers, i.e. 1 author, 2 authors or author et al. and year. Thus, the list of references should be given in alphabetical order. In the reference list all authors should be included. References with same author and year should be separated by adding a letter to the year (a, b, c...) in the list and the text.

3/20

* Sprowson AP et al. Bone Joint J 2016; 98-B: 1534–1541

www.heraeus-medical.com

90th anniversary

Vol. 91, No. 3, 2020 (pp. 221–363)

Bone cement with gentamicin ttamicin i i and clindamycin

ACTA ORTHOPAEDICA

69 %

reduction of deep infections in hip hemiarthroplasty after fractured neck of femur *

Volume 91, Number 3, June 2020

TF-IORT200080.indd 1

27-05-2020 19:07:17