Acta Orthopaedica Vol. 89, Issue 5 October 2018

5/18 ACTA ORTHOPAEDICA Vol. 89, No. 5, 2018 (pp. 475–593)

Volume 89, Number 5, October 2018 ISSN 1745-3674

IORT_89_05_Cover.indd 1

09-10-2018 12:16:36

IORT_89_05_Cover.indd 2

09-10-2018 12:16:39

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. 89, No. 5, 2018

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.

Annual Institutional Subscription, Volume 89, 2018 $1,173 £725 €939 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: Håkan Pårup P.O. Box 3255, SE-103 65 Stockholm, Sweden, Tel: +46 (0)8 440 80 40. Fax: +46 (0)8 440 80 50.

E-mail: hakan.parup@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.

Airfreight and mailing in the USA by agent named Air Business Ltd, c/o Worldnet Shipping Inc., 156-15, 146th Avenue, 2nd Floor, Jamaica, NY 11434, USA. Periodicals postage paid at Jamaica NY 11431.

US Postmaster: Send address changes to Acta Orthopaedica, Air Business Ltd, c/o Worldnet Shipping Inc., 156-15, 146th Avenue, 2nd Floor, Jamaica, NY 11434, USA. Subscription records are maintained at Taylor & Francis Group, 4 Park Square, Milton Park, Abingdon, OX14 4RN, United Kingdom.

Copyright © 2018 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. 89, No. 5, October 2018 Guest editorial Tech-trends in orthopedics 2018 Annotation High-dose glucocorticoid before hip and knee arthroplasty: To use or not to use—that’s the question Technical note Surgery guided by mixed reality: presentation of a proof of concept Hip Early gain in pain reduction and hip function, but more complications following the direct anterior minimally invasive approach for total hip arthroplasty: a randomized trial of 100 patients with 5 years of follow up Good stability of a cementless, anatomically designed femoral stem in aging women: a 9-year RSA study of 32 patients Surgeon’s experience level and risk of reoperation after hip fracture surgery: an observational study on 30,945 patients in the Norwegian Hip Fracture Register 2011–2015 Higher cartilage wear in unipolar than bipolar hemiarthroplasties of the hip at 2 years: A randomized controlled radiostereometric study in 19 fit elderly patients with femoral neck fractures Hemiarthroplasty and total hip arthroplasty in 30,830 patients with hip fractures: data from the Dutch Arthroplasty Register on revision and risk factors for revision Hip and knee Time-driven activity-based cost of outpatient total hip and knee arthroplasty in different set-ups 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 The course of pain and function in osteoarthritis and timing of arthroplasty: the CHECK cohort Knee Structural abnormalities detected by knee magnetic resonance imaging are common in middle-aged subjects with and without risk factors for osteoarthritis Minimal important change values for the Oxford Knee Score and the Forgotten Joint Score at 1 year after total knee replacement No detrimental effect of ligament balancing on functional outcome after total knee arthroplasty: a prospective cohort study on 129 mechanically aligned knees with 3 years’ follow-up Anterior distal femoral hemiepiphysiodesis can reduce fixed flexion deformity of the knee: a retrospective study of 83 knees Ankle and foot Re-arthrodesis after primary ankle fusion: 134/1,716 cases from the Swedish Ankle Registry Ten cold clubfeet Oncology Incidence and demographics of giant cell tumor of bone in The Netherlands: First nationwide Pathology Registry Study Midterm risk of cancer with metal-on-metal hip replacements not increased in a Finnish population

475

M Gordon

477

H Kehlet and V Lindberg-Larsen

480

T M Gregory, J Gregory, J Sledge, R Allard, and O Mir

484

B H Brismar, O Hallert, A Tedhamre, and J U Lindgren

490

E Aro, J J Alm, N Moritz, K Mattila, and H T Aro

496

A L Authen, E Dybvik, O Furnes, and J-E Gjertsen

503

W Figved, S Svenøy, S M Röhrl, J Dahl, L Nordsletten, and F Frihagen

509

S Moerman, N M C Mathijssen, W E Tuinebreijer, A J H Vochteloo, and R G H H Nelissen

515

H Husted, B B Kristensen, S E Andreasen, C S Nielsen, A Troelsen, and K Gromov U Berg, E Bülow, M Sundberg, and O Rolfson

522

528

M G J Gademan, H Putter, W B van den Hout, M Kloppenburg, S N Hofstede, S C Cannegieter, R G H H Nelissen, and P J Marang–van de Mheen

535

J Kumm, A Turkiewicz, F Zhang, and M Englund

541

L H Ingelsrud, E M Roos, B Terluin, K Gromov, H Husted, and A Troelsen E Aunan and S M Röhrl

548 555

N Stiel, K Babin, E Vettorazzi, S Breyer, N Ebert, M Rupprecht, R Stuecker, and A S Spiro

560

A Henricson, L Jehpsson, Å Carlsson, and B E Rosengren

565

R B Giesberts, E E G Hekman, G J Verkerke, and P G M Maathuis

570

A J Verschoor, J V M G Bovée, M J L Mastboom, P D S Dijkstra, M A J van de Sande, and H Gelderblom E Ekman, I Laaksonen, A Eskelinen, P Pulkkinen, E Pukkala, and K Mäkelä

575

Miscellaneous Growth of Cutibacterium acnes is common on osteosynthesis material of the shoulder in patients without signs of infection An increase in myeloid cells after severe injury is associated with normal fracture healing: a retrospective study of 62 patients with a femoral fracture

580 585

A Both, T O Klatte, A Lübke, H Büttner, M J Hartel, L G Grossterlinden, and H Rohde L Hesselink, O W Bastian, M Heeres, M ten Berg, A Huisman, I E Hoefer, W W van Solinge, L Koenderman, K J P van Wessem, L P H Leenen, and F Hietbrink

Case report Surgical management of obturator neuropathy with a concomitant acetabular labral tear — a case report

591

S Kanezaki, A Sakai, E Nakamura, and S Uchida

Information to authors

594

Acta Orthopaedica 2018; 89 (5): 475–476

475

Guest editorial

Tech-trends in orthopedics 2018 A trend is a direction in which something is developing; in medical specialties this can be viewed as the phase before something becomes evidence-based medicine. Early adopters are those that start using a technology as soon as it becomes available, i.e., individuals that are sensitive to trends. Within the medical field, orthopedics has a long track record of being an early adopter.

Unfortunately, discriminating between positive and negative trends can be difficult; while the anterolateral approach for hip fractures (Enocson et al. 2009) has become evidencebased medicine, other trends such as resurfacing arthroplasties (Reito et al. 2017) and primary surgery for clavicle fractures (Ban et al. 2016) have failed. As increasing amount of innovation occurs in the digital space, it is important that we transfer the lessons from surgical trends to these innovations.

Augmented reality In this issue of Acta Orthopaedica, Gregory et al. (2018) show us a glimpse of how we may perform surgery a few years down the road. Their article explores the use of mixed reality (also known as augmented reality) for surgery, a technology in which a computer-generated image is superimposed on top of the visual field. This is different from its sibling, virtual reality (VR), where the user is completely immersed in the computergenerated reality. Earliest mentions of augmented reality in PubMed go back to the mid-1990s (Lavallée et al. 1995), but we have only recently gained the technology that can live up to the original visions.

While Gregory et al.’s paper shows the first stumbling steps, it feels quite plausible that this could be just as common as cordless power tools in the years to come. The field where augmented reality probably has the strongest foothold is neurosurgery (Kersten-Oertel et al. 2013), but unfortunately the evidence of whether outcomes actually improve is scarce (Meola et al. 2017). It will be interesting to see if augmented reality will become a trend—at the moment the jury is still out; it is even uncertain whether they have assembled.

Computer-assisted surgery Computer-assisted surgery (CAS) has, contrary to augmented reality, been both widely implemented and tested, especially for knee surgery. By mapping CT/MRI scans to the actual bone the system can navigate for the surgeon, thereby allowing improved implant positioning and smaller incisions (Dutton et al. 2008). An interesting randomized controlled trial from Petursson et al. (2017) showed no benefit in regard

to RSA migration patterns, i.e. no signs of improved implant survival despite better positioning. Recently, they published patient reported outcomes from the same RCT where CAS patients were better on some subscales (Petursson et al.

2018), this should though be viewed with caution as knee function was a secondary outcome and the subscales were never mentioned in the ClinicalTrials.gov registration. Large registry cohorts have, though, not been able to clearly demonstrate the benefits (Roberts et al. 2015, Dyrhovden et al.

2016) and there are reports of falling utilization rates (Gholson et al. 2017).

Some believe that patient-specific guides will succeed CAS. By skipping the cumbersome mapping of the CT/MRI to the bone structure, you get patient specific 3D-printed saw guides that can both reduce the surgery time and improve accuracy.

Unfortunately, this has also failed to translate into any tangible patient benefits (Chareancholvanich et al. 2013, Victor et al. 2014, Leeuwen et al. 2018). The trend is certainly looking grim for these types of technologies.

Robot-assisted surgery One interesting development is the development of robotassisted orthopedic surgery. At the moment we are still far from fully autonomous robots, but simply assisting the human could be an efficient way of providing accuracy (Marchand et al. 2018). The long history of robotic surgery publications (Kwoh et al. 1988, Bach et al. 2002) suggests that the trend has some difficulty catching on. A quick glance at other surgical fields shows that the robot trend has certainly made a huge impact; the Da Vinci surgical system continues to grow year on year (Peters et al. 2018).

Artificial intelligence One of the strongest general tech-trends recently is the revival of neural networks, also known as deep learning, a form of artificial intelligence (AI). Through my work in applying AI for interpreting radiographs (Olczak et al. 2017) I am certainly biased, but I believe there is great potential in the technology. For instance, Chung et al. (2018) recently showed how it could be used for classifying humerus fractures, providing hope for solving the low classification reproducibility (Audigé et al. 2004). It is hoped it could also make the classifications more clinically relevant (Shehovych et al. 2016). There is also interesting work for classifying knee osteoarthritis (Tiulpin et al. 2018), the authors of which have released their dataset for anyone to experiment with (open data).

© 2018 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/4.0) DOI 10.1080/17453674.2018.1518806

476

Artificial intelligence is, however, not limited to image interpretations. The technology is about finding structure in data; it is similar to regular statistics but does this on an entirely different scale. It is already being implemented for augmented reality (Pollefeys 2017) and can in theory enhance anything that analyzes patterns. At the same time, there are indications that clinical applications struggle to deliver (Ross and Swetlitz 2017). The struggles suggest that we still have a lot to learn and, based on my own experience, it takes time to appreciate the full complexity. For instance, an orthopedic surgeon is well aware that a fracture is not a question of yes/no, but has almost infinite subtle interpretations.

Final thoughts We know that predicting the next big thing is hard (Denrell and Fang 2010), but at the same time it is interesting to survey the area and think of what direction we want the future to take.

There are even some who believe that we are the ones who shape the future. Most of the things mentioned in this paper will require great efforts and a great number of people; fortunately, it has never been easier to participate in this endeavor.

Max Gordon Department of Clinical Sciences at Danderyd Hospital Karolinska Institute, Stockholm, Sweden Email: max@gforge.se

Audigé L, Bhandari M, Kellam J. How reliable are reliability studies of fracture classifications?A systematic review of their methodologies. Acta Orthop 2004; 75(2): 184-94.

Bach C M, Winter P, Nogler M, Göbel G, Wimmer C, Ogon M. No functional impairment after Robodoc total hip arthroplasty. Acta Orthop Scand 2002; 73(4): 386-91.

Ban I, Nowak J, Virtanen K, Troelsen A. Overtreatment of displaced midshaft clavicle fractures. Acta Orthop 2016; 87(6): 541-5.

Chareancholvanich K, Narkbunnam R, Pornrattanamaneewong C. A prospective randomised controlled study of patient-specific cutting guides compared with conventional instrumentation in total knee replacement. Bone Joint J 2013; 95-B(3): 354-9.

Chung S W, Han S S, Lee J W, Oh K-S, Kim N R, Yoon J P, Kim J Y, Moon S H, Kwon J, Lee H-J, Noh Y-M, Kim Y. Automated detection and classification of the proximal humerus fracture by using deep learning algorithm.

Acta Orthop 2018; 89(4): 468-73.

Denrell J, Fang C. Predicting the next big thing: success as a signal of poor judgment. Rochester, NY: Social Science Research Network; 2010 Jun. Report No.: ID 1621800.

Dutton A Q, Yeo S-J, Yang K-Y, Lo N-N, Chia K-U, Chong H-C. Computerassisted minimally invasive total knee arthroplasty compared with standard total knee arthroplasty: a prospective, randomized study. J Bone Joint Surg (Am) 2008; 90(1): 2-9.

Dyrhovden G S, Fenstad A M, Furnes O, Gøthesen Ø. Survivorship and relative risk of revision in computer-navigated versus conventional total knee replacement at 8-year follow-up. Acta Orthop 2016; 87(6): 592-9.

Acta Orthopaedica 2018; 89 (5): 475–476

Enocson A, Hedbeck C-J, Tidermark J, Pettersson H, Ponzer S, Lapidus LJ. Dislocation of total hip replacement in patients with fractures of the femoral neck: a prospective cohort study of 713 consecutive hips. Acta Orthop 2009; 80(2): 184-9.

Gholson J J, Duchman K R, Otero J E, Pugely A J, Gao Y, Callaghan J J. Computer navigated total knee arthroplasty: rates of adoption and early complications. J Arthroplasty 2017; 32(7): 2113-19.

Gregory T M, Gregory J, Sledge J, Allard R, Mir O. Surgery guided by mixed reality: presentation of a proof of concept. Acta Orthop 2018; 89(5): 480483.

Kersten-Oertel M, Jannin P, Collins D L. The state of the art of visualization in mixed reality image guided surgery. Comput Med Imaging Graph 2013; 37(2): 98-112.

Kwoh Y S, Hou J, Jonckheere E A, Hayati S. A robot with improved absolute positioning accuracy for CT guided stereotactic brain surgery. IEEE Trans Biomed Eng 1988; 35(2): 153-60.

Lavallée S, Cinquin P, Szeliski R, Peria O, Hamadeh A, Champleboux G, Troccaz J. Building a hybrid patient’s model for augmented reality in surgery: a registration problem. Comput Biol Med 1995; 25(2): 149-64.

Leeuwen J A M J, Snorrason F, Röhrl S M. No radiological and clinical advantages with patient-specific positioning guides in total knee replacement. Acta Orthop 2018; 89(1): 89-94.

Marchand R C, Sodhi N, Khlopas A, Sultan A A, Higuera CA, Stearns K L, Mont M A. Coronal correction for severe deformity using robotic-assisted total knee arthroplasty. J Knee Surg 2018; 31(01): 2-5.

Meola A, Cutolo F, Carbone M, Cagnazzo F, Ferrari M, Ferrari V. Augmented reality in neurosurgery: a systematic review. Neurosurg Rev 2017; 40(4): 537-48.

Olczak J, Fahlberg N, Maki A, Razavian A S, Jilert A, Stark A, Sköldenberg O, Gordon M. Artificial intelligence for analyzing orthopedic trauma radiographs. Acta Orthop 2017; 88(6): 581-6.

Peters B S, Armijo P R, Krause C, Choudhury S A, Oleynikov D. Review of emerging surgical robotic technology. Surg Endosc 2018; 32(4): 1636-55.

Petursson G, Fenstad A M, Gøthesen Ø, Dyrhovden G S, Hallan G, Röhrl S M, Aamodt A, Furnes O. Computer-assisted compared with conventional total knee replacement: A multicenter parallel-group randomized controlled trial. J Bone Joint Surg 2018; 100(15): 1265-74.

Petursson G, Fenstad A M, Gøthesen Ø, Haugan K, Dyrhovden G S, Hallan G, Röhrl S M, Aamodt A, Nilsson K G, Furnes O. Similar migration in computer-assisted and conventional total knee arthroplasty. Acta Orthop 2017; 88(2): 166-72.

Pollefeys M. Second version of HoloLens HPU will incorporate AI coprocessor for implementing DNNs. Microsoft Res Blog. 2017. https://www. microsoft.com/en-us/research/blog/second-version-hololens-hpu-willincorporate-ai-coprocessor-implementing-dnns/ Reito A, Lehtovirta L, Lainiala O, Mäkelä K, Eskelinen A. Lack of evidence: the anti-stepwise introduction of metal-on-metal hip replacements. Acta Orthop 2017; 88(5): 478-83.

Roberts T D, Clatworthy M G, Frampton C M, Young S W. Does computer assisted navigation improve functional outcomes and implant survivability after total knee arthroplasty? J Arthroplasty 2015; 30(9, Suppl.): 59-63.

Ross C, Swetlitz I. IBM pitched Watson as a revolution in cancer care. It’s nowhere close. STAT News. 2017. https://www.statnews.com/2017/09/05/ watson-ibm-cancer/ Shehovych A, Salar O, Meyer C, Ford D. Adult distal radius fractures classification systems: essential clinical knowledge or abstract memory testing?

Ann R Coll Surg Engl 2016; 98(8): 525-31.

Tiulpin A, Thevenot J, Rahtu E, Lehenkari P, Saarakkala S. Automatic knee osteoarthritis diagnosis from plain radiographs: a deep learning-based approach. Sci Rep 2018; 8(1): 1727.

Victor J, Dujardin J, Vandenneucker H, Arnout N, Bellemans J. Patientspecific guides do not improve accuracy in total knee arthroplasty: a prospective randomized controlled trial. Clin Orthop Relat Res 2014; 472(1): 263-71.

Acta Orthopaedica 2018; 89 (5): 477–479

477

Annotation

High-dose glucocorticoid before hip and knee arthroplasty: To use or not to use—that’s the question During the last decade, “fast-track” or “enhanced recovery” programs have been introduced in hip and knee arthroplasty (THA/TKA) with a continuing improvement in recovery and consequently reduction in length of stay (LOS) (Zhu et al.

2017). Importantly, LOS may depend on early organ dysfunction, appearance of complications, or organizational factors.

Early postoperative organ dysfunction including pain, orthostatic intolerance, cardiovascular and thromboembolic morbidity, cognitive disturbances, nausea and vomiting, sleep disturbances, fatigue, and loss of muscle function may all depend on factors involved in the global surgical stress response. Two main mechanisms include the neuro-endocrine responses and the inflammatory-immunological responses, where the latter may be most important to determine post THA/TKA recovery (Gaudilliere et al. 2014). Of the various pharmacological interventions to reduce the inflammatory responses, glucocorticoids are the most powerful (de la Motte et al. 2014, Steinthorsdottir et al. 2017, Toner et al. 2017). Notably, systemic perioperative glucocorticoid administration has been the subject of 7 systematic reviews and/or meta-analyses in THA/ TKA since December 2016, summarizing that glucocorticoids may reduce nausea and vomiting as well as acute early postoperative pain. Interestingly, all of the 7 reviews (Hartman et al. 2017, Li et al. 2017, Liu et al. 2017, Meng and Li 2017, Yue et al. 2017, Li et al. 2018, Mohammad et al. 2018) claim to be the first and even more interestingly they include between 4 and 14 studies, thereby questioning the search methodology. Although all reviews agree that additional glucocorticoid may provide more analgesia, a critical assessment of efficacy was not done in relation to basic anesthetic/analgesic principles with different regimes in the different RCTs and a dosefinding analysis was only discussed in one review, but was not possible (Yue et al. 2017). Overall, most pain studies have used a limited dose of glucocorticoid (about 6–12 mg dexamethasone), but the most detailed studies discussed below have used a higher dose (125 mg methylprednisolone/~ 25 mg dexamethasone). Nevertheless, a more critical update on the use of perioperative glucocorticoid in THA and TKA may be appropriate in the light of more recent findings of specific pathophysiological responses to surgery and safety aspects.

Surgical pathophysiology The anti-inflammatory effects of high-dose systemic glucocorticoid are well documented by reduced responses in IL-6

and CRP (Yue et al. 2017). Consequently, there seems to be less postoperative fatigue (Lunn et al. 2011, 2013), which may provide an excellent start to postoperative recovery combined with the documented analgesic and anti-nausea and vomiting effects. Recently, other interesting effects on postoperative THA/TKA responses have shown that the usual degradation of endothelial barrier function after TKA was attenuated by preoperative high-dose (125 mg) methylprednisolone (~ 25 mg dexamethasone) (Lindberg-Larsen et al. 2017b) while there was no effect on the usual early (48 h) pronounced loss of quadriceps function (Lindberg-Larsen et al. 2017a), despite an inhibition of the inflammatory response. In addition, the endogenous anti-inflammatory protein pentraxin-3 (PTX-3) was significantly increased by preoperative high-dose glucocorticoids (Lindberg-Larsen et al. 2018b), although the clinical consequences regarding a potential reduction of infectious complications remain to be studied. The conventional effect of glucocorticoid on glucose homeostasis has also been assessed in a recent very detailed study on the different components of glucose homeostasis (hyperglycemia, insulin resistance, C-peptide response) and showed a transient impairment in glucose homeostasis lasting for only about 24–48 hours after 125 mg preoperative methylprednisolone (Lindberg-Larsen et al. 2018a). The surgery-induced increased thrombogenic responses may also be attenuated by perioperative high-dose glucocorticoid (Yue et al. 2017), although the potential positive effects on thromboembolic complications (Li et al. 2018) need further evaluation. Other aspects of early postoperative recovery after THA and TKA such as sleep disturbances (Krenk et al. 2012), and early cognitive dysfunction (Krenk et al. 2014) have not been elucidated so far. Interestingly, in one RCT, 125 mg methylprednisolone reduced early fatigue despite subjective reports on sleep quality on the first night being reduced (Lunn et al. 2011). With regard to early mobilization, preoperative 125 mg methylprednisolone did not improve orthostatic intolerance after THA despite a reduced inflammatory response (Lindberg-Larsen et al. 2018c). Finally, the pathophysiological effects of perioperative high-dose glucocorticoid on the early recovery phase from the operating theatre to the postoperative recovery unit (PACU) and discharge to the ward may potentially shorten the PACU stay due to improved opioid-sparing analgesia and PONV reduction, but calls for future detailed PACU discharge criteria studies (Steinthorsdottir et al. 2017).

© 2018 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/4.0) DOI 10.1080/17453674.2018.1475177

478

In summary, preoperative high-dose glucocorticoid may attenuate some, but not all undesirable pathophysiological responses to surgery. Consequently, detailed dose-finding studies and whether repeat dosing may further enhance recovery is required, as few and inconclusive data are available.

Finally, the delicate balance between necessary and undesirable inflammatory response needs attention, since an inflammatory response is required for wound healing etc., but an exaggerated response may delay recovery (Gaudilliere et al.

2014).

Safety aspects Overall, there seem to be no safety issues in using a highdose of glucocorticoid perioperatively in the surgical literature (Toner et al. 2017). However, due to the small-scale RCTs in THA/TKA (Hartman et al. 2017, Li et al. 2017, Liu et al. 2017, Meng and Li 2017, Yue et al. 2017, Li et al. 2018, Mohammad et al. 2018), there was no final conclusion with regard to safety aspects. More recently, a large prospective comparative multicenter cohort study in TKA was analyzed including about 1,500 consecutive patients after introduction of high-dose glucocorticoid vs. about 2,400 patients in the period before (Jørgensen et al. 2017). These high-volume data showed no risk signals from using a high dose (125 mg methyl­prednisolone) regarding different types of medical complications or superficial wound or deep infections. However, no conclusive data are available from insulin-dependent diabetics (LindbergLarsen et al. 2018a) and no safety studies are available evaluating higher doses or repeat dosing or consequences on prosthesis fixation. Based upon all 7 systematic reviews, no risk of glucocorticoid-induced psychosis has been reported.

Future strategies and conclusions Due to the increased interest in short-stay and outpatient THA and TKA (Vehmeijer et al. 2018), and given the abovementioned efficacy and safety aspects, the question is whether perioperative high-dose glucocorticoid may further enhance early recovery and facilitate a safe outpatient setup. A recent study in unselected patients in the Danish government-run healthcare system (Gromov et al. 2017), and including preoperative 125 mg methylprednisolone in addition to standard multimodal analgesia, showed that for about 15% of patients undergoing THA and TKA this could be performed in an outpatient setting. However, the most interesting finding from that study was the question why the remaining 85% of patients did not recover early, again emphasizing the many early recovery problems mentioned in the introduction of our article. Thus, the major future challenge that remains to be answered is whether an individualized and potentially higher dose of glucocorticoid with more pronounced anti-inflammatory effects can be safe and further enhance early recovery and reduce organ dysfunction in these 85% of patients. Interestingly, a large safety non-RCT cohort study (Jørgensen et al. 2017) showed that the proportion of patients staying > 4 days was reduced by 50%

Acta Orthopaedica 2018; 89 (5): 477–479

by high-dose glucocorticoid. Also, future studies should focus on specific “high-risk” pain responders like pain catastrophizers, preoperative opioid users, etc. (Gilron et al. 2018) Finally, it remains to be evaluated whether the recovery benefits are correlated with the magnitude of reduction in the inflammatory responses or whether the approach should be used only in certain “high-inflammatory” responders.

In conclusion, the previous and the more recent data suggest several major beneficial recovery effects of the anti-inflammatory effects of perioperative high-dose glucocorticoid in THA and TKA and so far without safety issues. Consequently, it may at this time be appropriate to recommend its use on a “routine” basis at least in TKA, but calling for more data from specific “risk” groups (THA, high-pain responders, diabetics, infected revisions, etc.) although the optimal dose needs to be defined. Perioperative use of high doses of glucocorticoids may therefore represent a further step to enhance recovery after “fast-track” or “enhanced recovery” THA and TKA (Kehlet 2013).

Both authors have contributed equally to the preparation of the manuscript.

No conflict of interest declared.

Henrik Kehlet 1,2 and Viktoria Lindberg-Larsen 1,2 1 Section

of Surgical Pathophysiology, Rigshospitalet, Copenhagen University; 2 The Lundbeck Foundation Centre for Fast-track Hip and Knee replacement, Copenhagen, Denmark Email: henrik.kehlet@regionh.dk

de la Motte L, Kehlet H, Vogt K, Nielsen C H, Groenvall J B, Nielsen H B, Andersen A, Schroeder T V, Lonn L. Preoperative methylprednisolone enhances recovery after endovascular aortic repair: a randomized, doubleblind, placebo-controlled clinical trial. Ann Surg 2014; 260(3): 540-9.

Gaudilliere B, Fragiadakis G K, Bruggner R V, Nicolau M, Finck R, Tingle M, Silva J, Ganio E A, Yeh C G, Maloney W J, Huddleston J I, Goodman S B, Davis M M, Bendall S C, Fantl W J, Angst M S, Nolan G P. Clinical recovery from surgery correlates with single-cell immune signatures. Sci Transl Med 2014; 6(255): 255ra131.

Gilron I, Carr D B, Desjardins P J, Kehlet H. Current methods and challenges for acute pain clinical trials. Pain Reports 2018 (Epub). doi: 10.1097/ PR9.0000000000000647 Gromov K, Kjaersgaard-Andersen P, Revald P, Kehlet H, Husted H. Feasibility of outpatient total hip and knee arthroplasty in unselected patients. Acta Orthop 2017; 88(5): 516-21.

Hartman J, Khanna V, Habib A, Farrokhyar F, Memon M, Adili A. Perioperative systemic glucocorticoids in total hip and knee arthroplasty: a systematic review of outcomes. J Orthop 2017; 14(2): 294-301.

Jørgensen C C, Pitter F T, Kehlet H. Safety aspects of preoperative highdose glucocorticoid in primary total knee replacement. Br J Anaesth 2017; 119(2): 267-75.

Kehlet H. Fast-track hip and knee arthroplasty. Lancet 2013; 381(9878): 1600-2.

Acta Orthopaedica 2018; 89 (5): 477–479

Krenk L, Jennum P, Kehlet H. Sleep disturbances after fast-track hip and knee arthroplasty. Br J Anaesth 2012; 109(5): 769-75.

Krenk L, Kehlet H, Hansen T B, Solgaard S, Soballe K, Rasmussen L S. Cognitive dysfunction after fast-track hip and knee replacement. Anesth Analg 2014; 118(5): 103-40.

Li X, Sun Z, Han C, He L, Wang B. A systematic review and meta-analysis of intravenous glucocorticoids for acute pain following total hip arthroplasty.

Medicine (Baltimore). 2017; 96(19): e6872.

Li D, Wang C, Yang Z, Kang P. Effect of intravenous corticosteroids on pain management and early rehabilitation in patients undergoing total knee or hip arthroplasty: a meta-analysis of randomized controlled trials. Pain Pract 2018; 18(4): 487-99.

Lindberg-Larsen V, Bandholm T Q, Zilmer C K, Bagger J, Hornsleth M, Kehlet H. Preoperative methylprednisolone does not reduce loss of kneeextension strength after total knee arthroplasty: a randomized, doubleblind, placebo-controlled trial of 61 patients. Acta Orthop 2017a ;88(5): 543-9.

Lindberg-Larsen V, Ostrowski S R, Lindberg-Larsen M, Rovsing M L, Johansson P I, Kehlet H. The effect of pre-operative methylprednisolone on early endothelial damage after total knee arthroplasty: a randomised, double-blind, placebo-controlled trial. Anaesthesia 2017b; 72(10): 1217-24.

Lindberg-Larsen V, Kehlet H, Bagger J, Madsbad S. Preoperative high-dose methylprednisolone and glycemic control early after total hip and knee arthroplasty: a randomized, double-blind, placebo-controlled trial. Anesth Analg 2018a (in press).

Lindberg-Larsen V, Kehlet H, Pilely K, Bagger J, Rovsing M L, Garred P. Preoperative methylprednisolone increases plasma Pentraxin 3 early after total knee arthroplasty: a randomized, double-blind, placebo-controlled trial. Clin Exp Immunol 2018b; 191(3): 356-62.

Lindberg-Larsen V, Petersen P B, Jans O, Beck T, Kehlet H. Effect of preoperative methylprednisolone on orthostatic hypotension during early mobilization after total hip arthroplasty. Acta Anaesthesiol Scand 2018c.

(Epub) doi: 10.1111/aas.13108

479

Liu X, Liu J, Sun G. Preoperative intravenous glucocorticoids can reduce postoperative acute pain following total knee arthroplasty: a meta-analysis.

Medicine (Baltimore). 2017; 96(35): e7836.

Lunn T H, Kristensen B B, Andersen L O, Husted H, Otte KS, Gaarn-Larsen L, Kehlet H. Effect of high-dose preoperative methylprednisolone on pain and recovery after total knee arthroplasty: a randomized, placebo-controlled trial. Br J Anaesth 2011; 106(2): 230-8.

Lunn T H, Andersen L O, Kristensen B B, Husted H, Gaarn-Larsen L, Bandholm T, Ladelund S, Kehlet H. Effect of high-dose preoperative methylprednisolone on recovery after total hip arthroplasty: a randomized, double-blind, placebo-controlled trial. Br J Anaesth 2013; 110(1): 66-73.

Meng J, Li L. The efficiency and safety of dexamethasone for pain control in total joint arthroplasty: a meta-analysis of randomized controlled trials.

Medicine (Baltimore). 2017; 96(24): e7126.

Mohammad H R, Hamilton T W, Strickland L, Trivella M, Murray D, Pandit H. Perioperative adjuvant corticosteroids for postoperative analgesia in knee arthroplasty. Acta Orthop 2018; 89(1): 71-6.

Steinthorsdottir K J, Kehlet H, Aasvang E K. Surgical stress response and the potential role of preoperative glucocorticoids on post-anesthesia care unit recovery. Minerva Anestesiol 2017; 83(12): 1324-31.

Toner A J, Ganeshanathan V, Chan M T, Ho K M, Corcoran T B. Safety of perioperative glucocorticoids in elective noncardiac surgery: a systematic review and meta-analysis. Anesthesiology 2017; 126(2): 234-48.

Vehmeijer S B W, Husted H, Kehlet H. Outpatient total hip and knee arthroplasty. Acta Orthop 2018; 89(2): 141-4.

Yue C, Wei R, Liu Y. Perioperative systemic steroid for rapid recovery in total knee and hip arthroplasty: a systematic review and meta-analysis of randomized trials. J Orthop Surg Res 2017; 12(1): 100.

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.

480

Acta Orthopaedica 2018; 89 (5): 480–483

Technical note

Surgery guided by mixed reality: presentation of a proof of concept Thomas M GREGORY 1,2, Jules GREGORY 3, John SLEDGE 4, Romain ALLARD 1, and Olivier MIR 2,5 1 Department

of Orthopaedic Surgery, Avicenne Teaching Hospital, Assistance Publique—Hôpitaux de Paris, University Paris-Seine-Saint-Denis, Sorbonne Paris Cité, Bobigny, France; 2 Moveo Institute, University Paris-Seine-Saint-Denis, Sorbonne Paris Cité, Bobigny, France; 3 Department of Radiology, Beaujon Teaching Hospital, Assistance Publique—Hôpitaux de Paris, University Paris-Diderot, Clichy, France; 4 Department of Orthopedic Surgery, Lafayette Hospital, Lafayette, LA, USA; 5 Department of Ambulatory Care, Gustave Roussy Cancer Campus, Villejuif, France Correspondence: tms.gregory@gmail.com Submitted 2018-05-01. Accepted 2018-07-03.

Surgical innovation aims to improve both patient outcome and safety. Among many others, computer-based solutions such as 3-D anatomical reconstruction and computer-based procedures planning have shown clear benefits to patients (Beckmann et al. 2009, Hernandez et al. 2017, Saragaglia et al. 2018). The next step in the usage of 3D anatomical information is to go from planning to per-operative usage. This can be done using augmented reality (AR), a technique that allows the superimposing of a digital image on top of the visual field, i.e., augmenting reality. As virtual elements in the form of holograms align with the reality this is commonly referred to as mixed reality (MR) as it enhances reality.

Recently, Microsoft has developed a new concept of an MR headset (HoloLens), equipped with motors, that allows the hand user to interact with the headset by oral command or simple gesture.

The promise of such technology in the surgical field is huge, as it allows the surgeon (i) to gain access to computer-based solutions in real time during the procedure while remaining totally sterile, (ii) to gain access to 3-D holograms related to the patient imaging or the surgical technique, and (iii) to remotely interact with colleagues located outside the theatre.

Here, we present the first use of such a system in a surgical environment and focus on the latest trends of the rapidly developing connection between MR and surgery.

Methods Patient The patient was an 80-year-old woman (height 155 cm, weight 52 kg) with advanced arthritis of the shoulder combined with rupture of the tendons covering the joint, motivating the implantation of a shoulder prosthesis.

Technical setup The HoloLens headset is a self-contained, holographic com-

puter, enabling its user to engage with her/his digital content and interact with holograms in the world around her/him.

The technical characteristics of the device are depicted in the supplementary Appendix. No specific installation time in the operating room is required, the device being wireless.

Data preparation The raw DICOM files are too large to be downloaded into the headset. Therefore, the DICOM data stored in the hospital PACS were instantly loaded and preprocessed in a Cloud Unit (Azure Graphic Processing Unit—Microsoft Corp, Redmond, WA, USA). During the surgery, the DICOM data are made available for the headset through a dedicated radiological holographic application (TeraRecon Holoportal—https:// www.terarecon.com/) connected via WiFi to the Cloud Unit.

Surgical procedure A standard reverse shoulder arthroplasty was performed. The challenge in this surgery was the limited bone stock of the vault (due to the patient’s anatomy with a Walch A2-type glenoid).

Navigational technique There was no calibration between HoloLens and the navigation system, as the MR headset has a capacity to drag the holograms manually. Subsequently, the holographic 3-D reconstruction of the scapula was manually positioned in such a way that the visible part of the bone matched with the corresponding part of the hologram. The surgeon was then able to see the hidden part of the scapula in a holographic mode.

Funding and potential conflicts of interest Funding source: MOVEO Foundation. Dr Mir has acted as a consultant for Amgen, BMS, Eli-Lilly, Lundbeck, MSD, Novartis, Pfizer, Roche, Servier and Vifor. Dr Gregory has acted as a consultant for Evolutis. The other authors have no conflict of interest to disclose.

© 2018 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/4.0) DOI 10.1080/17453674.2018.1506974

Acta Orthopaedica 2018; 89 (5): 480–483

Figure 1. Outside view of the headset during the surgical procedure.

481

Figure 2. View from the headset during the surgical procedure.

Results The prosthesis was implanted with the aid of this MR headset at Avicenne Hospital in Bobigny, France, on December 5, 2017. As described above, the surgeon was able to access the patient’s medical data combined with the data of the operative technique, which were transmitted into his operating headset in real time during the intervention, e.g., patient scapula 3-D reconstruction extracted from her CT scan, the planning of the positioning of the glenoid and the whole operative technique developed in 3-D Figure 3. Postoperative CT scan of the glenoid (top left: sagittal view; bottom left: coronal view; holograms according to the planned right: axial view). position. Hence, the surgeon was able to compare, stage by stage, what he was doing with what and required dedicated time to comment on every step. On the had to be done. The surgeon was also able to drag the 3-D other hand, a better understanding of the patient’s anatomy reconstruction of the scapula, scaled 100%, right before his as allowed by the use of the headset probably contributed to eyes, by simple gestures in front of the headset (Figure 1) in time-saving. approximately 3 minutes. He was then able to superimpose the Postoperative CT scan showed adequate position of the CT scan scapula reconstruction with manual adjustment onto prosthesis (Figure 3). The patient experienced no peri- or postthe visible part of the bone (Figure 2 and see video at https:// operative complication and was discharged home 3 days after youtu.be/oQu1rGt6ym4). surgery (similar to the routine practice in our institution).

The surgeon also shared what he saw in his headset with Postoperative visit on day 45 confirmed the lack of postop4 other surgeons in the USA and in the UK, who were able erative complications. to send information in the operator’s field of vision, via Skype, throughout the intervention (which was taking place in France). This is the first time that such complete usage of the immersive and collaborative aspects has been implemented in Discussion surgery of the shoulder.

Compared with a classical surgical procedure, the use of an The total duration of surgery was 90 minutes, similar to the MR headset might provide an improvement in outcomes for mean duration of such a procedure in the operator’s experience both surgeon and patient, without reducing the safety of the (without the use of HoloLens). The duration of the procedure procedure. Indeed, thanks to 3-D holograms, the keyframes was prolonged due to the fact that it was broadcast on the Web of the operation as well as critical organs (nerves, arteries...)

482

are shown in real time by the surgeon. This provides a gain in accuracy and safety in the procedurem which finally might result in time-saving and correct positioning of the implant (Léger et al. 2017, Vávra et al. 2017). Notably, the technical setup of the device and data preparation before surgery had short duration (as explained above) but might be time-consuming for first-time users.

The improvement of the surgical procedure also lies in the opportunity for the surgeon to interact with colleagues and ask them for advice or warnings. The operation field is often small, and the surgeon’s gestures as well as an accurate view of the patient’s anatomy are restricted to only the surgeon and his assistant. Thanks to this new technology, the surgeon’s view is projected onto a screen, and accessible to many more people. This provides constant feedback on his/her decision, as well as being a potential educational tool.

The safety of the patient seems not be compromised by this new device. Indeed, the operating time was not lengthened, and no major adverse events occurred. On the contrary, the security of the surgery is probably improved, because the surgeon can see the patient’s anatomy directly, and other physicians can give helpful comments. As an illustration, Opolski et al. (2017) performed revascularization of a chronic coronary total occlusion assisted by AR glasses without adverse events.

This improvement does not incur a loss of comfort and focus. The weight of the device is low (579 grams, see supplementary Appendix), and the use of AR devices of comparable weight did not cause pain or tiredness (Sinkin et al. 2016, Léger et al. 2017). The amount of information is controlled in real time by the surgeon, so that the cognitive load is tolerable. The voice- and image-activated command is easy to use, without latency or excessive repetition (Léger et al. 2017). The image quality, contrast and stability are good enough not to cause tiredness or motion sickness. During the videoconferencing, no cut-offs occurred; thus, the constant of interaction and procedure safety was guaranteed. Finally, the battery has sufficient durability for most surgical procedures (up to 5.5 hours, see supplementary Appendix).

When a new technology appears, patient and staff approval is also a fundamental element. Muensterer et al. (2014) reported that their colleagues, staff, families and patients had a positive response to such technology.

Importantly, issues of data protection will have to be addressed for greater use of this technology (Muensterer et al. 2014, Vávra et al. 2017). AR interest in surgical procedures is strongly increasing, as shown by the large number of published studies. Reviews of the literature, randomized controlled trials, and single-center tests are multiplying (Muensterer et al. 2014, Sinkin et al. 2016, Léger et al. 2017, Opolski et al. 2017, Vávra et al. 2017, Pulijala et al. 2018). Various AR and MR devices are currently being tested, and data are convergeing toward the same results: procedure performance, and effectiveness as a learning tool. The divergence mainly

Acta Orthopaedica 2018; 89 (5): 480–483

involves surgeon comfort and satisfaction, and is related to the AR device itself (Table).

From an educational perspective, while the use of screens in the surgical theater has exploded over the last few years, the vision of high-quality AR for surgery (Ponce et al. 2014) or medical training has not yet been realized. To date, AR-based simulation is considered a promising approach in the surgical curriculum, allowing not only technical performance evaluation, but also telementoring.

Because the gap between research and wide application is mainly a question of cost, things could change rapidly in the next years thanks to cheaper AR devices. As mentioned in the ASiT position paper (Milburn et al. 2012), regional disparities in the availability of simulators and financial restraints are theoretical limitations to their wide use. We believe that recent, inexpensive devices will help lessen these expected difficulties. Importantly, dedicated teaching time should be made available within existing and future projects (Milburn et al. 2012).

In summary, we believe that our case report illustrates the fact that surgical practice and education can derive significant benefits from the implementation of AR and MR tools in daily practice. We believe that such an immersive and cost-effective approach would improve the training capacity of orthopedic surgery simulations.

The authors wish to thank Dr April Amstrong (Hershey Medical Center, Pennsylvania, US), Professor Roger Emery and Dr Peter Reilly (Imperial College, London, UK), and Professor Young Lae Moon (Teaching Hospital of Chosun, South Korea) for their support and contribution to this work.

Acta thanks Max Gordon for help with peer review of this study.

Beckmann J, Stengel D, Tingart M, Götz J, Grifka J, Lüring C. Navigated cup implantation in hip arthroplasty. Acta Orthop 2009; 80: 538-44.

Hernandez D, Garimella R, Eltorai A E M, Daniels A H. Computer-assisted orthopaedic surgery. Orthop Surg 2017; 9: 152-8.

Léger É, Drouin S, Collins D L, Popa T, Kersten-Oertel M. Quantifying attention shifts in augmented reality image-guided neurosurgery. Healthcare Technol Lett 2017; 4:1 88-92.

Milburn J A, Khera G, Hornby S T, Malone P S C, Fitzgerald J E F. Introduction, availability and role of simulation in surgical education and training: review of current evidence and recommendations from the Association of Surgeons in Training. Int J Surg 2012; 10: 393-8.

Muensterer O J, Lacher M, Zoeller C, Bronstein M, Kübler J. Google Glass in pediatric surgery: an exploratory study. Int J Surg 2014; 12: 281-9.

Opolski M P, Debski A, Borucki B A, Staruch A D, Kepka C, Rokicki J K, Sieradzki B, Witkowski A. Feasibility and safety of augmented-reality glass for computed tomography-assisted percutaneous revascularization of coronary chronic total occlusion: a single center prospective pilot study.

J Cardiovasc Comput Tomogr 2017; 11: 489-96.

Ponce B A, Menendez M E, Oladeji L O, Fryberger C T, Dantuluri P K. Emerging technology in surgical education: combining real-time augmented reality and wearable computing devices. Orthopedics 2014; 37: 751-7.

Pulijala Y, Ma M, Pears M, Peebles D, Ayoub A. Effectiveness of immersive virtual reality in surgical training: a randomized control trial. J Oral Maxillofac Surg 2018; 76: 1065-72.

Acta Orthopaedica 2018; 89 (5): 480–483

483

Current published data on AR in surgery Study

Outcomes

Main findings

Vávra et al. 2017

Review of the literature: augmented reality can presently improve the results of surgical procedures?

– AR is an effective tool for training and skill assessment of surgery residents, other medical staff, or students – The interest of surgeons is increasing – The performance is comparable to traditional techniques – The time required for completing a procedure has been reduced while using any form of AR – Inattentional blindness: the surgeon does not see an unexpected object which suddenly appears in his field of view – The amount of information is increasing and may be distracting – The latency of the whole system is also of concern – Long-term wear comfort is an issue – AR projections produce simulator sickness Sinkin et al. 2016 Ease of use – Average ease of image capture = 3.11/5 Quality of images – Average ease of video capture = 3.22/5 Gaze disruption – Average ease of using the wink feature = 1.89 Distraction – Quality of image and video = 3.89 and 3.67/7 Comfort and satisfaction to wear: Average for comfort = 4.56/7, Average for satisfaction = 3.78/7 – 33% of users felt the device to be a distraction from the case Pulijala et al. 2017 Self-assessment scores of trainee confidence before and after the intervention. Novice surgical residents – The study group participants showed greater perceived self-confidence levels Opolski et al. 2017 Operators’ satisfaction assessed by a 5-point Likert scale – The voice-activated co-registration and review of images were feasible and highly rated by operators (4.7/5 points) – There were no major adverse events – More frequent selection of the first-choice stiff wire and lower contrast exposure – Success rates and safety outcomes remained similar between the two groups Muensterer et al. 2014 AR Glass was worn daily for 4 weeks to identify and evaluate daily activities with potential applicability – Wearing Glass throughout the day was well tolerated – Colleagues, staff, families, and patients overwhelmingly had a positive response to Glass – Low battery endurance – Poor overall audio quality – Long transmission latency combined with interruptions and cut-offs during internet videoconferencing Léger et al. 2017 Time taken to perform the task. Attention shifts: – There were significant reductions in terms of the time taken to perform the task, and attention shifts Feelings about accuracy, intuition, comfort, perceived cognitive load – Users felt that the system was easier to use and that their performance was better

Saragaglia D, Rubens-Duval B, Gaillot J, Lateur G, Pailhé R. Total knee arthroplasties from the origin to navigation: history, rationale, indications.

Int Orthop 2018; Mar 27. doi: 10.1007/s00264-018-3913-z. [Epub ahead of print] Sinkin J C, Rahman O F, Nahabedian M Y. Google Glass in the operating room: the plastic surgeon’s perspective. Plast Reconstr Surg 2016; 138: 298-302.

Vávra P, Roman J, Zonča P, Ihnát P, Němec M, Kumar J, Habib N, ElGendi A. Recent development of augmented reality in surgery: a review. J Healthc Eng 2017; 2017: 4574172.

484

Acta Orthopaedica 2018; 89 (5): 484–489

Early gain in pain reduction and hip function, but more complications following the direct anterior minimally invasive approach for total hip arthroplasty: a randomized trial of 100 patients with 5 years of follow up B Harald BRISMAR 1, Ola HALLERT 2, Anna TEDHAMRE 3, and J Urban LINDGREN 1 1 Department of Clinical Sciences, Intervention and Technology, Karolinska Intitutet, Department of Orthopedics, Karolinska University Hospital, Stockholm; 2 Department of Orthopedics, Södertälje Hospital, Södertälje; 3 Department of Orthopedics, St Göran Hospital, Stockholm, Sweden

Correspondence: harald.brismar@sll.se Submitted 2017-11-02. Accepted 2018-06-20.

Background and purpose — The minimally invasive direct anterior (DA) approach for total hip arthroplasty (THA) is supposed to reduce surgical tissue trauma. We hypothesized that patients operated with the DA technique would have less postoperative pain and better hip function compared with a group operated with a conventional direct lateral (DL) approach.

Patients and methods — 100 patients with hip osteoarthritis scheduled for THA were equally randomized to surgery through either DA or DL. Pain was assessed on a VAS scale, hip function with TUG, 10mWT, HHS, and quality of life with EQ-5D. Patients were followed up after the first 3 days, 8 weeks, and at 1 and 5 years postoperatively.

Results — The DA group registered less pain with activity on the second day (VAS 42 vs. 55), performed TUG 6 seconds faster on the third day and had 8 points higher HHS and higher EQ-5D index (0.86 vs 0.78) at 8 weeks; all differences were statistically significant. No clinically relevant differences between groups in pain, hip function, or quality of life were seen at 1 or 5 years. 7 surgical approach related complications appeared in the DA group, none in the DL. Interpretation — The results indicate that the presumably less traumatic approach results in reduced immediate postoperative pain and better hip function and higher quality of life in the early postoperative period. However, this positive effect is not seen at later time points. Instead, complications appear to be over-represented, thus questioning the use of the method.

A continuous ambition towards further improvement of functional outcome and shorter hospital stay has led to techniques aiming at reducing surgical tissue trauma in total hip arthroplasty (THA). The minimally invasive direct anterior (DA) single incision approach uses an internervous plane without detachment of tendons or splitting of muscles. A number of observational cohort studies have been published describing the technique (Kennon et al. 2003, Siguier et al. 2004, Matta et al. 2005). Several reports indicate less pain in the immediate postoperative period following DA-THA compared with other approaches (Goebel et al. 2012, Barrett et al. 2013, Mjaaland et al. 2015) and higher quality of life measures up to 1 year postoperatively were found in 1 study (Restrepo et al. 2010).

Potential risks associated with minimally invasive surgical techniques are malpositioning of implants and unintended tissue damage due to a reduced surgical field. A higher complication rate for minimally invasive hip arthroplasty has been reported from the Swedish hip arthroplasty registry (Hailer et al. 2012) and also in a study focusing on DA-THA (Spaans et al. 2012). However, others have reported excellent results using the DA technique (Siguier et al. 2004, Restrepo et al.

2010). MRI studies following DA-THA have shown less gluteus muscle fatty atrophy at 6–12 months postoperatively (Bremer et al. 2011, De Anta-Díaz et al. 2016) compared with DL-THA. Also, a cadaveric dissection study implies less gluteus medius damage following DA (van Oldenrijk et al. 2010).

We compared the outcome following DA-THA with a standard direct lateral approach (DL) for THA focusing on pain, hip function, quality of life, and complications. Our primary hypothesis was that a surgical technique not involving intentional detachment of muscle insertions (the DA technique) would lead to less pain in the immediate postoperative period.

© 2018 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/4.0) DOI 10.1080/17453674.2018.1504505

Acta Orthopaedica 2018; 89 (5): 484–489

485

Table 1. Characteristics of patients who had a total hip arthroplasty through a direct anterior (DA) approach or a direct lateral (DL) approach. Values are median and inter-quartile range or proportion of patients

Patient characteristics

DA DL n = 50 n = 50

Sex (F / M) Age (years) Smoker (yes / no) Weight (kg) Height (cm) Body mass index ASA grade (1 / 2 / 3) 1 hip OA / 2 hip OA / 2 hip THA Side (R / L)

32 / 18 66 (58–74) 6 / 42 80 (68–90) 168 (164–175) 27 (24–29) 12 / 30 / 8 26 / 17 / 7 30 / 20

33 / 17 67 (60–76) 6 / 42 76 (64–86) 167 (162–175) 27 (24–30) 10 / 35 / 5 25 / 20 / 5 30 / 20

Our secondary hypothesis was that hip function and quality of life would be better following DA-THA. No difference in complications was expected. At the start of this study, no large randomized trial on this subject had been published.

Patients and methods This was a single-center randomized parallel group study with equal randomization to 2 groups. The study was conducted at the orthopedic department, Karolinska University Hospital, Huddinge, Sweden. Patients with hip osteoarthritis referred for hip arthroplasty were, after consent, informed and asked for participation in the study. Exclusion criteria were dementia, neuromuscular disorders, alcohol/drug abuse, and previous hip surgery on the affected side. 100 consecutive patients were randomly allocated by a computer program to either surgery through a direct anterior (DA) single incision approach (Rachbauer 2005) or a direct lateral (DL) transgluteal approach (Duparc et al. 1997), inclusion period November 2006 to April 2008. There were 50 patients in each group.

Patient preoperative characteristics were similar between the 2 groups (Table 1). Each patient was given a consecutively numbered sealed envelope containing an allocation paper put into the envelope by an independent research manager not further involved in the study. The surgeon and patient were blinded prior to the opening of the envelope.

Briefly, the DA was carried out with the patient supine on a standard operating table allowing angulation at the level of the hip. The skin was incised at a point 2 fingerbreadths lateral to the anterior sciatic spine and extended 8–10 cm distally. The tensor fascia lata and gluteus medius muscles were retracted laterally and the sartorius and rectus muscles medially exposing the capsule. A special offset acetabular reamer and an offset broach handle were used. The DL was performed with the patient in a lateral decubitus position. Access to the joint was gained through a 10–20 cm long skin incision centered over the greater trochanter, splitting the fascia lata/glu-

teus maximus and detachment of the caudal 2/3 of the gluteus medius and the entire gluteus minimus tendon insertions.

Finally, the capsule was excised anteriorly. The muscle tendons were reattached to the trochanter by osteosutures following implantation.

All patients had uncemented implants (Accolade stem and Trident PSL cup, Stryker, Kalamazoo, MI, USA). 92 patients received spinal anesthesia (47 DA and 45 DL) and 8 general anesthesia (3 DA and 5 DL). 2 surgeons performed all procedures. Both surgeons were well acquainted with both techniques having performed at least 40 independent procedures with the DA technique and using the DL technique for several years before the onset of the study.

Time for surgery was from skin incision to skin closure. The anesthesia nurse estimated perioperative blood loss from suction and surgical swabs. Blood samples were drawn preoperatively, and at 6 hours and 2 days postoperatively; C-reactive protein (CRP), interleukin-6 (IL-6), and hemoglobin levels were analyzed.

All patients were treated postoperatively according to the same pain management protocol including a regular long-acting morphine analog the first day (oxycodone 10 mg 2 times daily), regular paracetamol (1 g 4 times daily) and shortduration morphine (oxycodone or morphine) on demand.

The long-acting dose was adjusted with regard to the previous day’s morphine consumption. The total sum of equipotent doses of oral morphine consumed 3 days postoperatively was estimated (10 mg oral oxycodone = 20 mg oral morphine, 10 mg iv morphine = 30 mg oral morphine). Patients were asked to keep track of how many days after discharge from hospital they continued to use morphine.

Pain in the operated hip at rest and during activity was estimated by the patients on a VAS scale. VAS was recorded preoperatively, on the morning of day 1 and day 2, at 8 weeks, 1 year, and 5 years postoperatively. Hospital stay was defined as the number of postoperative nights in the hospital and in those cases where rehabilitation was needed an additional 3 days were added as a default. No patient was sent to a rehabilitation facility earlier than the fourth postoperative day. All patients were given the same postoperative physiotherapy regime focusing on independent rising from bed and walking with aids. Discharge criteria were: ability to walk 20 meters independently and adequate pain management with oral analgesics.

Timed up and go (TUG) (Kennedy et al. 2002) and 10-meter walk test (10mWT) (Salbach et al. 2001) were performed preoperatively, on day 3, and at 8 weeks, 1 year, and 5 years postoperatively. EQ-5D and Harris Hip Score were assessed at 8 weeks, 1 year, and 5 years postoperatively.

Patients or their caretakers were asked at 8 weeks, 1 year, and 5 years if any complication had occurred related to the hip surgery. 99 patients had 8 weeks’ follow-up (1 patient, DL group, missed the 8-week follow-up). One patient had died at 1 year (DA group) and 99 came to follow-up. At 5 years 87 patients had complete follow-up (DA 45 and DL 42). 3 patients in

486

each group had died, 1 had been revised (DA group), 1 declined further follow-up (DL) and 5 patients were unable to respond to questions due to dementia or frailty (DA 1 and DL 4).

The 13 patients who did not come to followup were all checked with a phone call and/or medical chart investigation.

Acta Orthopaedica 2018; 89 (5): 484–489

Table 2. Surgical and early postoperative outcome parameters for patients who had a total hip arthroplasty through a direct anterior (DA) approach or a direct lateral (DL) approach

Outcome parameters Operation time (min) a Intraoperative blood loss (mL) a IL-6, 6 h postoperatively (pg/mL) b CRP 2 days postoperatively (mg/mL) b Morphine consumption (mg) 3 days a Postoperative days on morphine a Postoperative hospital nights a

DA DL n = 50 n = 50 p-value 101 (87–112) 325 (200–500) 56 (46–67) 211 (191–231) 140 (110–197) 12 (6–17) 3 (3–4)

80 (71–86) < 0.001 c 300 (250–450) 0.7 c 77 (60–94) 0.05 d 238 (219–257) 0.06 d 175 (140–234) 0.02 c 14 (8–18) 0.05 c 4 (3–5) 0.006 c

Statistics Categorical data were summarized using frequency counts or percentages and continuous data were presented as means and 95% confia Median and inter-quartile range. dence intervals or as medians and inter-quarb Mean and 95% confidence intervals. tile ranges (Q1–Q3). A linear mixed model, c Mann–Whitney U test, d T-test. which allows for missing data, was used to analyze repeated measurements of VAS, EQ5-D, 10mWT, TUG, and Harris Hip Score VAS pain with activity (mm) (HHS) taken during the hospital stay and at VAS pain at rest (mm) 80 80 follow up until 1 year. The fixed factors in the Direct anterior approach model were Group (DA and DL), Time (2, 3 70 70 Direct lateral approach or 4 time points) and the Group x Time inter60 60 action. In case of a significant interaction, 50 simple main effects tests were performed, 50 i.e., effects of 1 factor holding the levels of 40 40 the other factor fixed. We used the default 30 method ddfm = between in SAS in order to 30 compute the degrees of freedom. The most 20 20 appropriate covariance structure for our out10 Direct anterior approach come measures was unstructured and hetero- 10 Direct lateral approach geneous compound symmetry. The variables 0 0 Preop.

Day 1 Day 2 Day 3 Preop.

Day 1 Day 2 Day 3 TUG and 10mWT were ln-transformed prior to these analyses to compensate for their posi- Figure 1. VAS for patients who had a total hip arthroplasty through a direct anterior (DA) approach or a direct lateral( DL) approach preoperatively and at 3 time points posttively skewed distribution. Student’s t-test for operatively. The DA group had lower VAS pain at rest 2 days (p = 0.03) and lower VAS independent samples and the Mann–Whitney pain with activity 2 days and 3 days postoperatively than the DL group (p = 0.002 and p U test were used in analyzing data at 1 time = 0.02). Between- and within-group comparisons were performed using a linear mixed model. Circles indicate DA means, boxes indicate DL means, error bars indicate the 95% point. Fisher’s exact test was used to analyze confidence intervals of the DA and DL means. categorical data. A p-value < 0.05 was considered statistically significant.

A sample size of 100 patients was chosen in order to get statistical power of at least 80%, assuming a standard deviation of 20 for VAS registration. The primary endpoint was pain Results in the first postoperative days. The minimal clinically impor- All patients were operated according to their allocated method tant difference was set to 13 for VAS. 5-year results were ana- and there were no intraoperative complications. Hospital stay lyzed separately since the number of patients under study had was 1 day shorter in the DA group (Table 2) and only 1 patient diminished considerably and unequally between groups. had to be sent to a rehabilitation facility compared with 5 in We used Statistica 10.0 (www.statsoft.com), StatSoft (www. the DL group. statsoft.com), and SAS System 9.1 (SAS Institute, Cary, NC, Patients in the DA group registered less pain during activity USA) for the analyses. and at rest on the second day postoperatively and also the third day during activity. There was no clinically relevant differEthics, funding, and potential conflicts of interest ence at any other time point (Figure 1). 13 patients in the DA The Central Ethical Review Board, Stockholm approved the group did not require any extra, oral or parenteral, morphine protocol (2005/1508-31) and informed consent was obtained in the first 3 postoperative days compared with 5 patients in from all participants. Stryker unconditionally sponsored the the DL group (p = 0.07). The DA group had a 20% lower morstudy. The authors declare no conflicts of interest. phine analgesic consumption measured as the total amount of

Acta Orthopaedica 2018; 89 (5): 484–489

487

Timed Up and Go Test (seconds)

10 Meter Walk Test (seconds)

50

50 Direct anterior approach Direct lateral approach

40

40

30

30

20

20

10

10

8 weeks 1 year 5 years DA DL DA DL DA DL

n = 50 n = 49 n = 49 n = 50 n = 45 n = 42

Direct anterior approach Direct lateral approach 0

Preop. 3 days 8 weeks 1 year 3 years

0

Table 3. Pain in the operated hip at rest and during activity (VAS) at 3 time points postoperatively following total hip arthroplasty through a direct anterior (DA) approach or a direct lateral (DL) approach.

Values are absolute number of patients at each time period with VAS = 0, VAS 1–30 and VAS > 30

Preop. 3 days 8 weeks 1 year 3 years

Figure 2. Timed up and go (TUG) and 10-meter walk test (10mWT) for patients who had a total hip arthroplasty through a direct anterior (DA) approach or a direct lateral (DL) approach preoperatively and at 4 time points postoperatively. The DA group performed TUG faster than the DL group 3 days postoperatively (p = 0.04). Between- and withingroup comparisons were performed using a linear mixed model. Circles indicate DA means, boxes indicate DL means, error bars indicate the 95% confidence intervals of the DA and DL means.

VAS at rest 0 36 31 45 41 38 38 1–30 11 15 4 6 6 3 > 30 3 3 0 3 1 1 VAS during activity 0 26 21 44 41 36 33 1–30 21 22 4 6 7 7 > 30 3 6 1 3 2 2

differences between groups at any other time point (Figure 2).

Both groups improved in HHS and EQ-5D Harris Hip score EQ-5D index from preoperative levels to 8 weeks and 1 year 100 1.0 postoperatively. At 8 weeks the DA group had 8 points higher HHS and 0.08 higher EQ5-D 80 0.8 means than the DL group. At 1- and 5-year follow-ups no clinically relevant differences were 60 0.6 seen between groups (Figure 3).

At 5-year follow up 7 surgical approach 40 0.4 related complications had appeared in the DA group, but none in the DL group (p = 0.01). In the DA group 2 early dislocations were handled 20 0.2 by closed reduction, neither of them recurred; Direct anterior approach Direct anterior approach Direct lateral approach Direct lateral approach 1 early deep infection was resolved following 0 0 Preop. 8 weeks 1 year 5 years Preop. 8 weeks 1 year 5 years open irrigation, retention of the prosthesis, and antibiotics for 6 months; 1 patient dislocated Figure 3. Harris hips score (HHS) and EQ-5D index for patients who had a total hip arthroplasty through a direct anterior (DA) approach or a direct lateral (DL) approach 22 months postoperatively and was revised at preoperatively and at 3 time points postoperatively. The DA group had lower HHS and 2 years and 9 months due to a pseudotumor; 1 EQ-5D 8 weeks postoperatively than the DL group (p = 0.002 and p = 0.009). Betweenpatient developed hip pain 4 years postoperaand within-group comparisons were performed using a linear mixed model. Circles indicate DA means, boxes indicate DL means, error bars indicate the 95% confidence tively and was diagnosed with an iliopsoas cyst, intervals of the DA and DL means. revision was planned at 5 years; 1 patient had a late dislocation at 4 years and 7 months postopmorphine consumed by the third day and quit oral morphine eratively and was handled by closed reduction, and, finally, 1 analgesics 2 days earlier than the DL group (Table 2). There patient developed instability–subluxations at 1.5 years postwas no important difference between groups regarding pain at operatively, but did not find the disability severe enough to rest or during activity at 8 weeks, 1 year, and 5 years postop- motivate revision surgery. Postoperative radiographs were eratively (Table 3). analyzed in all patients with instability and all but one had In the DA group mean Il-6 was 27% lower at 6 hours post- cups positioned within Lewinnek’s safe zone (Lewinnek et al. operatively and mean CRP was 11% lower 2 days postopera- 1978). That patient had a cup with 54 degrees of inclination tively (Table 2). and 16 degrees anteversion; he had 1 dislocation and no later TUG and 10mWT deteriorated from preoperative values to recurrent instability problem. In the DL group, 1 patient devel3 days postoperatively but already at 8 weeks postoperatively oped hyperesthesia from the femoral cutaneous nerve of the they had improved and continued to improve until 1 year. The opposite, unoperated leg, probably originating from pressure DA group performed TUG 6 seconds faster than the DL group from the table support during surgery. 1 deep venous thromboat 3 days postoperatively. There were no clinically relevant sis occurred within 3 months in each group. 1 patient in each

488

group had disturbed wound healing. Both healed uneventfully following treatment with oral antibiotics.

Discussion In this randomized trial, we found an initial beneficial effect on postoperative pain and hip function, probably due to reduced tissue trauma. However, at later follow-up, that effect disappeared. An unexpectedly high number of complications occurred in the DA group, presumably related to a more demanding surgical technique.

Pain Patients had less pain following the DA procedure in the immediate postoperative period. Even though the DA group consumed less total morphine in the first 3 days postoperatively, pain at rest was the same between the 2 groups and notably pain during activity was lower in the DA group.

Patients in the DA group also had a shorter postoperative period on opiate analgesics and a shorter hospital stay. It is, therefore, reasonable to believe that the presumably less traumatic minimal invasive technique actually caused less pain.

The result is in line with previous reports (Goebel et al. 2012, Mjaaland et al. 2015). Not surprisingly, the pain did not differ between groups after the immediate postoperative period, in concordance with other studies (Restrepo et al. 2010, Reichert et al. 2015).

Surgical trauma A technique not involving tendon or muscular division is theoretically attractive, although one has to bear in mind the risk of unintentional muscular trauma due to either excessive traction of tissues or accessory releases in order to gain surgical access. In an attempt to objectively estimate the surgical trauma we measured Il6 and CRP, inflammatory markers reported to be elevated proportionally to the amount of trauma (Watt et al. 2015). Following THA Il-6 has been found to peak rapidly 6 hours postoperatively, while CRP rises more slowly with a peak 2 days postoperatively (Wirtz et al. 2000). The 27% lower Il-6 levels in the DA group could reflect reduced surgical trauma, while the 11% lower CRP levels are more difficult to interpret. Both values varied considerably between individuals within groups and have in a recent study been reported to correlate badly to the postoperative functional outcome (Poehling-Monaghan et al. 2017).

Hip function Mean TUG was 6 seconds shorter in the DA group at 3 days and mean HHS 8 points higher at 8 weeks indicating early better hip function. Since there were no clinically important differences between groups at later time points, it seems that the presumed lesser muscular trauma following the DA procedure did not substantially affect hip function in the long

Acta Orthopaedica 2018; 89 (5): 484â&#x20AC;&#x201C;489

run. This is in line with previous reports (Mayr et al. 2009, Restrepo et al. 2010, Parvizi et al. 2016).

Quality of life The long-term benefit of reduced surgical trauma and less initial postoperative pain is unclear. In the immediate rehabilitation period, pain and function of the operated joint will certainly affect measures of quality of life such as EQ-5D, illustrated by the mean 0.08 higher EQ-5D at 8 weeks in the DA group. Later, at 1- and 5-year follow up, potential differences in hip function must be quite substantial compared with other factors not related to the previous hip surgery in order to have an impact on daily life.

Complications 7 of 50 patients in the DA group had an approach-related complication within 5 years from surgery. 4 were directly related to instability. Malposition of the cups did not seem to explain the dislocations. One may speculate that excessive mobilization of the femur in order to be able to insert the femoral implant during a DA procedure could negatively affect posterior stabilizing structures and thereby hip stability. There have been previous reports on high complication rates following MISTHA procedures (Spaans et al. 2012, Hailer et al. 2012, MĂźller et al. 2014) and longer learning curves than expected (Spaans et al. 2012) indicating that MIS-THA can be demanding. The higher complication rate in the DA group could represent a still ongoing learning curve. There are reports of more favorable results (Kennon et al. 2003, Siguier et al. 2004) suggesting that a high surgical volume and meticulous improvement in surgical technique probably decrease rates of complications, but in view of our results one may question whether the method is too demanding to master in relation to the potential advantages.

Limitations Although this was a randomized trial the result may have been affected by bias since patients and caregivers could not be blinded as the location of the scar revealed the approach. At the time of the study, the method was not widespread, possibly reducing the risk of different expectations linked to either method. Patients were treated by the same protocol and the results were recorded by a physiotherapist not involved in the recruitment of patients. It is possible that the functional tests used were not demanding enough to detect differences between groups at later time points.

Conclusion In our hands, the positive early effects of the minimally invasive direct anterior technique do not justify further use of the technique unless complications can be reduced to a similar rate seen using the conventional direct lateral technique. The hypothetical advantage of having undamaged tissue around the hip following a minimally invasive direct anterior procedure could not be proven in our study.

Acta Orthopaedica 2018; 89 (5): 484–489

Design of the study, recruitment of patients, and performance of the operations were done by HB, OH, and UL. Follow up by HB and AT. The data was analyzed by HB and UL. The paper was written by HB and reviewed

by OH, AT and UL. The authors would like to thank Elisabeth Berg, LIME, KI for statistical support.

Acta thanks Stefan Bolder and Lars Nordsletten for help with peer review of this study.

Barrett W P I, Turner S E, Leopold J P. Prospective randomized study of direct anterior vs postero-lateral approach for total hip arthroplasty. J Arthroplasty 2013; 28(9): 1634-8.

Bremer A K, Kalberer F, Pfirrmann C W, Dora C. Soft-tissue changes in hip abductor muscles and tendons after total hip replacement: comparison between the direct anterior and the transgluteal approaches. J Bone Joint Surg (Br) 2011; 93-B: 886-9.

De Anta-Díaz B, Serralta-Gomis J, Lizaur-Utrilla A, Benavidez E, LópezPrats F A. No differences between direct anterior and lateral approach for primary total hip arthroplasty related to muscle damage or functional outcome. Int Orthop 2016; 40(10): 2025-30.

Duparc F, Thomine J M, Dujardin F, Durand C, Lukaziewicz M, Muller J M, Freger P. Anatomic basis of the transgluteal approach to the hip-joint by anterior hemimyotomy of the gluteus medius. Surg Radiol Anat 1997; 19: 61-7.

Goebel S, Steinert A F, Schillinger J, Eulert J, Broscheit J, Rudert M, Nöth U. Reduced postoperative pain in total hip arthroplasty after minimalinvasive anterior approach. Int Orthop 2012; 36(3): 491-8.

Hailer N P, Weiss R J, Stark A, Kärrholm 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.

Kennedy D, Stratford P W, Pagura S M, Walsh M, Woodhouse L J. Comparison of gender and group differences in self-report and physical performance measures in total hip and knee arthroplasty candidates. J Arthroplasty 2002; 17: 70-7.

Kennon R E, Keggi J M, Wetmore R S, Zatorski L E, Huo M H, Keggi K J. Total hip arthroplasty through a minimally invasive anterior surgical approach. J Bone Joint Surg (Am) 2003; 85-A(Suppl): 39-48.

Lewinnek G E, Lewis J L, Tarr R, Compere C L, Zimmerman J R. Dislocations after total hip-replacement arthroplasties. J Bone Joint Surg Am 1978; 60(2): 217-20.

Matta J M, Shahrdar C, Ferguson T. Single-incision anterior approach for total hip arthroplasty on an orthopaedic table. Clin Orthop Relat Res 2005; (441): 115-24.

489

Mayr E, Nogler M, Benedetti M G, Kessler O, Reinthaler A, Krismer M, Leardini A. A prospective randomized assessment of earlier functional recovery in THA patients treated by minimally invasive direct anterior approach: a gait analysis study. Clin Biomech 2009; 24(10): 812-18.

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.

Müller D A, Zingg P O, Dora C. Anterior minimally invasive approach for total hip replacement: five-year survivorship and learning curve. Hip Int 2014; 24(3): 277-83.

Parvizi J, Restrepo C, Maltenfort M G. Total hip arthroplasty performed through direct anterior approach provides superior early outcome: results of a randomized, prospective study. Orthop Clin North Am 2016; 47(3): 497-504.

Poehling-Monaghan K L, Taunton M J, Kamath A F, Trousdale R T, Sierra R J, Pagnano M W. No correlation between serum markers and early functional outcome after contemporary THA. Clin Orthop Relat Res 2017; 475(2): 452-62.

Rachbauer F. Minimally invasive total hip arthroplasty via direct anterior approach. Orthopade 2005; 34: 1103-4, 1106-8, 1110.

Reichert J C, Volkmann M R, Koppmair M, Rackwitz L, Lüdemann M, Rudert M, Nöth U. Comparative retrospective study of the direct anterior and transgluteal approaches for primary total hip arthroplasty. Int Orthop 2015; 39(12): 2309-13.

Restrepo C, Parvizi J, Pour A E, Hozack W J. Prospective randomized study of two surgical approaches for total hip arthroplasty. J Arthroplasty 2010; 25(5): 671-9 .

Salbach N M, Mayo N E, Higgins J, Ahmed S, Finch L E, Richards C L. Responsiveness and predictability of gait speed and other disability measures in acute stroke. Arch Phys Med Rehabil 2001; 82: 1204-12.

Siguier T, Siguier M, Brumpt B. Mini-incision anterior approach does not increase dislocation rate. Clin Orthop Relat Res 2004; (426): 164-73.

Spaans A J, van den Hout J A, Bolder S B. High complication rate in the early experience of minimally invasive total hip arthroplasty by the direct anterior approach. Acta Orthop 2012; 83(4): 342-6. van Oldenrijk J, Hoogland P V, Tuijthof G J, Corveleijn R, Noordenbos T W, Schafroth M U. Soft tissue damage after minimally invasive THA. Acta Orthop 2010; 81(6): 696-702.

Watt D G, Horgan P G, McMillan D C. Routine clinical markers of the magnitude of the systemic inflammatory response after elective operation: a systematic review. Surgery 2015; 157: 362-80.

Wirtz D C, Heller K D, Miltner O, Zilkens K W, Wolff J M. Interleukin-6: a potential inflammatory marker after total joint replacement. Int Orthop 2000; 24: 194-6.

490

Acta Orthopaedica 2018; 89 (5): 490–495

Good stability of a cementless, anatomically designed femoral stem in aging women: a 9-year RSA study of 32 patients Erik ARO 1, Jessica J ALM 1, Niko MORITZ 1, Kimmo MATTILA 2, and Hannu T ARO 1 1 Department

of Orthopaedic Surgery and Traumatology, Turku University Hospital and University of Turku, Turku; 2 Department of Diagnostic Imaging, Turku University Hospital, Turku, Finland Correspondence: hannu.aro@utu.fi Submitted 2018-03-05. Accepted Accepted: 2018-05-28.

Background and purpose — We previously reported a transient, bone mineral density (BMD)-dependent early migration of anatomically designed hydroxyapatite-coated femoral stems with ceramic–ceramic bearing surfaces (ABGII) in aging osteoarthritic women undergoing cementless total hip arthroplasty. To evaluate the clinical significance of the finding, we performed a follow-up study for repeated radiostereometric analysis (RSA) 9 years after surgery.

Patients and methods — Of the 53 female patients examined at 2 years post-surgery in the original study, 32 were able to undergo repeated RSA of femoral stem migration at a median of 9 years (7.8–9.3) after surgery. Standard hip radiographs were obtained, and the subjects completed the Harris Hip Score and Western Ontario and McMaster Universities Osteoarthritis Index outcome questionnaires.

Results — Paired comparisons revealed no statistically significant migration of the femoral stems between 2 and 9 years post-surgery. 1 patient exhibited minor but progressive RSA stem migration. All radiographs exhibited uniform stem osseointegration. No stem was revised for mechanical loosening. The clinical outcome scores were similar between 2 and 9 years post-surgery.

Interpretation — Despite the BMD-related early migration observed during the first 3 postoperative months, the anatomically designed femoral stems in aging women are osseointegrated, as evaluated by RSA and radiographs, and exhibit good clinical function at 9 years.

The femoral component of the ABG (Anatomique Benoist Girard) hip prosthesis with mirrored geometries for left and right hips was designed to accommodate the natural anatomy of the proximal femur. The 3 goals were to create postoperative stability, ensure osseointegration of the proximal part of the stem coated with hydroxyapatite (HA), and promote proximal load transfer to prevent stress shielding (Van Rietbergen and Huiskes 2001, Van Der Wal et al. 2008). Surgical preparation, consisting of distal reaming and proximal broaching, is less forgiving with anatomic stem designs (Khanuja et al.

2011, Giebaly et al. 2016). Although the ABG-II stem is prone to periprosthetic fractures in the early postoperative period (Mäkelä et al. 2010, Thien et al. 2014, Catanach et al. 2015), it has demonstrated a high survival rate in prospective studies (Epinette et al. 2013, Herrera et al. 2013) and received acceptance in the Dutch national evaluation based on a 10-year survival of 91–93% in the Australian implant register (Poolman et al. 2015). A similar 10-year survival rate is noted in the Finnish Arthroplasty Register (2018).

Considering the original design concepts, it is not surprising if the clinical performance of the ABG-II stem deviates in postmenopausal women due to the mismatch with the proximal femoral endosteal geometry. The proximal femur of postmenopausal women frequently undergoes age-related widening of the intramedullary canal (Noble et al. 1995, Casper et al. 2012) and endosteal trabeculation of the cortical bone (Zebaze et al. 2010) with a frequent (> 60–70%) presence of low bone mineral density (BMD) (Glowacki et al.

2003, Mäkinen et al. 2007). Indeed, we observed an increased migration of ABG-II stems in female patients with low BMD during the first 3 months but not thereafter (Aro et al. 2012).

In logistic regression analyses, signs of continued migration related to BMD, age, and canal flare index were also noted. In response to these observations, we invited the same cohort for the re-evaluation of stem stability. The participants underwent repeated radiostereometric analysis (RSA) of stem migration during the 2- and 9-year postoperative periods.

© 2018 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/4.0) DOI 10.1080/17453674.2018.1490985

Acta Orthopaedica 2018; 89 (5): 490–495

491

Table 1. Baseline demographics and clinical characteristics No. of cases Age, years BMI (SD) WHO classification of bone mineral density Normal bone (T-score ≥ –1.0 Osteopenia (–2.5 ≤ T-score < –1.0) Osteoporosis (T-score < –2.5) Dorr classification Type A Type B Type C WOMAC score, (SD) HHS score, (SD)

32 62 (41–78) 31 (6) 11 16 5 17 12 3 49 (16) 52 (15)

Inclusion n = 61

2-year RSA n = 53

9-year RSA n = 32

Excluded (n = 8): – periprosthetic fracture, 4 – absence of markers, 2 – concurrent use of corticosteroids and bisphosphonates, 2 Excluded (n = 21): – dead, 4 – ceramic head/liner revision, 3 – withdrawn for poor general condition, 9 – declined to participate, 4 – attended the visit but did not undergo RSA, 1

Figure 1. Patient flow throughout the study. Data from the first 2 years are presented in more detailed in a previous study (Aro et al. 2012).

Patients and methods Study design This is a follow-up study of the 2-year single-center RSA study on female patients with hip osteoarthritis (Aro et al.

2012). The subjects underwent a preoperative screening of skeletal status (Mäkinen et al. 2007). The study protocol, inclusion and exclusion criteria, and the screening process have been reported previously (Aro et al. 2012). Patients underwent cementless total hip arthroplasty at the Turku University Hospital between August 2003 and March 2005.

The surgery was performed using an anterolateral Hardinge approach. All patients received a custom-modified CE-certified ABG-II hip prosthesis with an anatomically designed hydroxyapatite-coated femoral stem with a non-modular neck, a hydroxyapatite-coated hemispherical acetabular cup, and ceramic-ceramic bearing surfaces (Stryker) (Aro et al. 2012, Finnilä et al. 2016). After surgery, patients were instructed to use crutches and partial weight-bearing up to 6 weeks.

The original cohort included 39 skeletally healthy patients with no history of osteoporosis treatment or any other medication affecting bone metabolism and a group of 14 patients with ongoing osteoporosis or corticosteroid treatment. The latter group did not differ in the 2-year RSA pattern of stem migration and were included in the current study. A cohort of 32 patients was able to attend the visit for repeated RSA imaging at 9 years (Table 1). The clinical outcome of the original cohort of 53 patients could be assessed (Figure 1). The 3 failure cases of ceramic-ceramic bearings have been discussed previously (Finnilä et al. 2016).

RSA The stem was marked by the manufacturer with 6 RSA beads (Figure 2). During surgery, multiple RSA markers (n = 4–7) were inserted into the greater and lesser trochanters. The center of the femoral head served as 1 additional marker.

A standardized RSA technique was applied throughout the study. RSA imaging was performed using the uniplanar technique (Kärrholm et al. 1997), and image analysis was

Figure 2. Schematic drawing of the prosthesis with 6 tantalum RSA stem markers and the coordinate system for RSA analysis of 3D micromotion of the femoral stem. In order to explicitly record the direction of micromotion, the directions are marked with + and – signs for both the translations and rotations.

performed using UmRSA software (RSA Biomedical Innovations AB, Umeå, Sweden) according to RSA guidelines (Valstar et al. 2005). RSA was applied to access femoral stem micromotions along and around 3 axes (x, y, z) (Figure 2) and comparison was made with the baseline position. Since the translations along the x-axis and z-axis are difficult to interpret, the focus was on the translation along the y-axis and rotations around the 3 axes. Stability and adequate distribution of the markers were assessed by calculating the mean error of rigid body fitting (ME) and the condition number (CN) (Valstar et al. 2005). At 9 years, the mean ME was 0.27, and the mean CN was 71. The precision of the measurements was validated in each patient via double examination. As recommended (Ranstam et al. 2000, Derbyshire et al. 2009), the precision for each axis (Table 2) was calculated as standard deviations multiplied by the Student’s t distribution with n degrees of freedom for the mean value of differences between the double examinations.

Clinical outcome Standard 2-plane hip radiographs were obtained at 2 and 9 years. Radiographic signs of osseointegration were evaluated and classified according to Engh et al. (1990). Hip function was evaluated using the Harris Hip Score (HHS) and West-

492

Acta Orthopaedica 2018; 89 (5): 490–495

Table 2. Precision of radiostereometric analysis based on double examinations

Y-axis translation (mm)

X-axis rotation (°)

2

15

0

10

–2

5

Clinical precision 0.13 0.22 0.36 0.53 1.95 0.19

–4

0

Clinical precision = 2.05 x standard deviation for the mean difference between the double examinations.

–6

Translation (mm) Rotation (°) x-axis y-axis z-axis x-axis y-axis z-axis

ern Ontario and McMaster Universities Osteoarthritis Index (WOMAC) (SooHoo et al. 2007).

Sample size and statistics Four patients were excluded from the RSA analysis due to inconsistent time-related values of translation or rotation related to either high ME (range 0.53–2.30) or high CN (185).

Of the excluded patients, 2 were characterized by exceptionally high BMI (38–48). Thus, the RSA analysis included 28 patients out of 32 fulfilling the requirement for the minimum number of subjects (15–25 per group) (Valstar et al. 2005).

Due to the limited group size, no subgroup analysis of patients with normal or low BMD was performed. HHS and WOMAC data were normally distributed and the significance of differences between 2 and 9 years were tested by paired t-test. Nonparametric tests were applied in the statistical analysis of RSA data. Data analysis was performed using SPSS software (IBM SPSS version 25.0; IBM Corp, Armonk, NY, USA). P-values < 0.05 were considered significant.

0

0.25

0.5

1

2

9

–5

0

0.25

0.5

Y-axis rotation (°)

Z-axis rotation (°)

10

2

5

1

2

9

1

2

9

0

0 –2 –5 –4

–10 –15

–6 0

0.25

0.5

1

2

9

0

0.25

0.5

Years after index operation

Figure 3. The migration pattern of individual femoral stems (n = 28) during the 9-year follow-up. 1 patient exhibited continuous x-axis rotation (yellow-filled markers).

Results

Based on the measurements of translation and rotation on the 3 axes, RSA exhibited no statistically significant femoral stem migration between 2 and 9 years after surgery (Table 3). The low precision of the y-axis rotation (Table 2) caused intraindividual variation of the measured internal rotation around the y-axis (Figure 3). As a comparison, the high precision of the y-axis translation (subsidence) was associated with conEthics, funding, and potential conflicts of interests stant values (Figure 3).

The study extension was approved by the Ethics CommitAfter the cessation of early migration during the first 3 tee of the Hospital District of Southwest Finland. Informed postoperative months, 1 patient exhibited progressive, albeit consent was obtained from all patients. The current study and minor, posterior tilt (rotation around the x-axis) (Figure 3, data the original study were supported financially by Competitive points marked in yellow). Migration was 0.5° between 2 years Research Funding of the Hospital District of Southwest Fin- and 9 years, close to the RSA precision for the rotation around land. The original study was also financially supported by the the x-axis.

Academy of Finland and Stryker Inc., which took no part in Patient-reported outcome scores exhibited no statistically the organization of the study, in analysis of the results, or writ- significant changes between 2 and 9 years. The mean HHS ing of the manuscript. No competing interests were declared. of the cohort (n = 32) was 87 (95% CI 81–91) at 2 years and 85 (CI 78–91) at 9 years. The mean WOMAC was 15 (CI 11–20) at 2 years Table 3. Femoral stem migration between 2 and 9 years and 19 (CI 12–26) at 9 years.

All of the stems were classified as a

2 years 9 years Difference stable (overall score > 10) according

median (range) median (range) (95% CI of median) p-value b to the mean fixation and stability score Translation compared with baseline, mm (Engh et al. 1990). The average score x-axis 0.02 (–0.46 to 2.58) 0.00 (–0.48 to 2.80) –0.02 (–0.08 to 0.07) 0.7 of the cohort (n = 32) was 21 (18–27). y-axis –0.73 (–4.67 to 0.05) –0.73 (–4.53 to 0.11) –0.05 (–0.08 to 0.03) 0.2 z-axis –0.25 (–1.03 to 2.17) –0.29 (–1.79 to 1.52) –0.07 (–0.17 to 0.13) 0.3 The patient with suspected stem migraRotation compared with baseline, degrees tion (Figure 3) did not show radiographic x-axis 0.18 (–3.23 to 11.04) 0.39 (–3.68 to 11.06) 0.02 (–0.14 to 0.18) 0.6 signs of implant loosening and exhibited y-axis 0.34 (–10.38 to 4.00) 0.24 (–10.76 to 4.12) 0.09 (–0.69 to 0.73) 0.9 z-axis –0.16 (–1.32 to 1.03) –0.21 (–2.29 to 1.05) 0.02 (–0.17 to 0.12) 1.0 a high (21) mean fixation and stability score. There was no radiolucency at the a Range of follow-up 7.8–9.3 years. b Related-samples Wilcoxon signed rank test. bone–implant interference, and all stems

Acta Orthopaedica 2018; 89 (5): 490–495

exhibited endosteal bone bridging (“spot welds”) as a sign of osseointegration. The spot welds were constantly located at the cortico-metaphyseal junction corresponding to the border between the HA-coated proximal stem and the polished distal stem. During the 9-year follow-up of the cohort, none of the stems were revised for loosening or infection.

To explore the clinical significance, the patient with suspected continuous stem migration was followed for an additional 5 years based on the review of electronic chart records.

During the follow-up including hip radiographs at 14 years, the patient did not develop stem loosening.

Discussion Our previous results (Aro et al. 2012) questioned the early stability of uncemented ABG-II stems in aging women, but the current study revealed good stability at 9 years postoperatively. Based on the clinical and radiographic evaluations and RSA performed at 9 years, apart from 1 possible exception, ABG-II stems were “well-fixed without migration,” fulfilling the classical mechanical definition of osseointegration as discussed by Ryd (2006). ABG-II stems also exhibited “definite roentgenographic signs of bone ingrowth” (Engh et al. 1990).

Thus, the stems appeared to have successfully osseointegrated during the first 2 years. In 1 patient, we could not exclude the possibility of stem loosening, which typically represents a continuous migration pattern over time (Kärrholm 2012).

However, the extended follow-up did not demonstrate any clinical consequence. Thus, failed osseointegration or late loosening seem to be unlikely.

The critical question relates to the upper limit and time frame for acceptable early migration of uncemented femoral stems. As stated previously (Kärrholm 2012), the stems should preferably not migrate at all. However, many designs migrate, and the maximum time limit for subsidence appears to be 1 year. Consistent with the results from previous 5-year (Wolf et al. 2010, Weber et al. 2014) and 6-year (Callary et al. 2012) studies on non-anatomical femoral stem designs, the current study suggests that limited early migration is not harmful. An uncemented femoral component designed specifically for total hip arthroplasty of elderly hip fracture patients has demonstrated pronounced early migration in patients with low periprosthetic BMD, but no component migrated after 3 months (Sköldenberg et al. 2011). In contrast to previous studies of ABG-I prostheses in male and female patients less than 65 years of age (Thien et al. 2007, Nysted et al. 2014), our study focused exclusively on aging women who suffer more frequently from age-related anatomic abnormalities of the proximal femur compared with men (Noble et al. 1995).

Our original analysis demonstrated that preoperative BMD dictates subsidence (translation along the y-axis) to a certain extent during the first 3 months after surgery (Aro et al. 2012).

The measured subsidence in patients with normal BMD (mean

493

0.5 mm) (SD 0.5) and low BMD (1.1 mm) (SD 0.9) seemed to be within acceptable ranges because the clinical outcome scores of the patients with normal and low BMD were similar at 2 years, and the current analysis suggested uniform clinical recovery and implant healing. Notably, the RSA-measured subsidence differs from the clinical term of stem subsidence, which is evident on plain radiographs and carries a natural risk of impaired functional outcome. In our cohort, 3 patients exhibited excessive stem subsidence (> 3 mm) during the first 3–6 months (Figure 3). Interestingly, an analysis of potential risk factors for excessive subsidence may be more difficult than expected (White et al. 2012).

The histological analysis of retrieved ABG-II specimens has demonstrated an almost complete resorption of HA coating as a function of implantation time. The new bone incorporated with HA coating may be replaced by growing new bone (Tonino et al. 2009). We could not recognize any signs of late mechanical loosening of the stems. This finding is consistent with the reported acceptable 10-year survival of the stem (Epinette et al. 2013, Herrera et al. 2013) in addition to data from Australian and Finnish national registers. Interestingly, the stems constantly formed spot welds at the lower border of the HA-coated proximal stem. The site of spot welds corresponds to the region of the cortico-metaphyseal junction, which obviously transfers the load from the prosthesis to the cortical bone. This correlation is not precisely according to the original design (Van Rietbergen and Huiskes 2001, Van Der Wal et al. 2008), which aimed to achieve full loading of the HA-coated region of the stem.

As a limitation, not all subjects participated in the repeated examination at 9 years. However, we were able to track the clinical outcomes of all 53 participants of the original cohort, and none of the stems were revised for mechanical loosening or infection. In this respect, the study group of the current 9-year evaluation was not biased. Our study was also weak in the detection of the main stem rotation around the y-axis. We paid attention to the quality of our RSA imaging setup both for marker-based and markerless RSA of cementless femoral stems (Mäkinen et al. 2004, Nazari-Farsani et al. 2016). In our hands, the precision for measurement of y-axis rotation has been at the same level as the marker-based RSA (0.71°) and model-based RSA (0.79 degrees). These values are higher or approximately at the same level as the precision values reported in recent clinical trials (Li et al. 2014, Weber et al.

2014). Collaborating with our previous analysis of the ABG-II stem (Aro et al. 2012), the precision for the measurement of y-axis rotation was suboptimal (1.95 degrees). This finding could be explained by the fact that the femoral head was poorly visible due to the use of ceramic-ceramic articulation.

The visibility of the RSA markers on the medial surface of the stem was not good either. The analysis would also benefit from the implantation of an increased number of bone markers. The number of markers is one of the critical determinants of RSA precision (Kärrholm et al. 1997). In a patient with a

494

low number (4 or 5) bone markers, the loss of even 1 marker, for example due to late bone resorption of the calcar region, may jeopardize the spatial contribution. The analysis of individual migration patterns demonstrated the consequences caused by the imprecision. The data on stem rotation around the y-axis suffered from the noise caused by the low precision throughout the follow-up visits.

In conclusion, the anatomically designed uncemented femoral stem, which exhibited BMD-dependent early migration in aging women, currently exhibits signs of osseointegration, as evaluated by RSA and radiographic images, and acceptable clinical outcome scores in our 9-year prospective extension study.

EA collected data and wrote the manuscript, JJA collected data, NM supervised RSA, KM supervised radiographic imaging, and HTA initiated the study, operated on patients, and wrote the manuscript. All the authors approved the final manuscript.

The authors would like to thank Satu Honkala, RN, for the study coordination.

Acta thanks Johan Kärrholm and Stephan Maximilian Röhrl for help with peer review of this study.

Aro H T, Alm J J, Moritz N, Mäkinen T J, Lankinen P. Low BMD affects initial stability and delays stem osseointegration in cementless total hip arthroplasty in women. Acta Orthop 2012; 83(2): 107-14.

Callary S A, Campbell D G, Mercer G E, Nilsson K G, Field J R. The 6-year migration characteristics of a hydroxyapatite-coated femoral stem. A radiostereometric analysis study. J Arthroplasty 2012; 27(7): 1344-8.

Casper D S, Kim G K, Parvizi J, Freeman T A. Morphology of the proximal femur differs widely with age and sex: relevance to design and selection of femoral prostheses. J Orthop Res 2012; 30(7): 1162-6.

Catanach M J M, Sorial R M, Eslick G D. Thirteen-year outcomes in the anatomique Benoist Girard II hip prosthesis. ANZ J Surg 2015; 85(4): 255-9.

Derbyshire B, Prescott R J, Porter M L. Notes on the use and interpretation of radiostereometric analysis. Acta Orthop 2009; 80(1): 124-30.

Engh C A, Massin P, Suthers K E. Roentgenographic assessment of the biologic fixation of porous-surfaced femoral components. Clin Orthop Relat Res 1990; 257: 107-28.

Epinette J A, Asencio G, Essig J, Llagonne B, Nourissat C. Clinical results, radiological findings and survival of a proximally hydroxyapatite-coated hip ABG II stem at a minimum of ten years’ follow-up: results of a consecutive multicentre study of 1148 hips in 1053 patients. Bone Joint J 2013; 95-B(12): 1610-16.

Finnish Arthroplasty Register (FAR). Endonet Report 2018. Available from: https://www2.thl.fi/endo/report/#index Finnilä S, Moritz N, Svedström 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. Acta Orthop 2016; 87(1): 48-54.

Giebaly D E, Twaij H, Ibrahim M, Haddad F S. Cementless hip implants: an expanding choice. Hip Int 2016; 26(5): 413-23.

Glowacki J, Hurwitz S, Thornhill T S, Kelly M, LeBoff M S. Osteoporosis and vitamin-D deficiency among postmenopausal women with osteoarthritis undergoing total hip arthroplasty. J Bone Joint Surg 2003; 85-A(12): 2371-7.

Acta Orthopaedica 2018; 89 (5): 490–495

Herrera A, Mateo J, Lobo-Escolar A, Panisello J J, Ibarz E, Gracia L. Longterm outcomes of a new model of anatomical hydroxyapatite-coated hip prosthesis. J Arthroplasty 2013; 28(7): 1160-6.

Khanuja H S, Vakil J J, Goddard M S, Mont M A. Cementless femoral fixation in total hip arthroplasty. J Bone Joint Surg Am 2011; 93(5): 500-9.

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, 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.

Mäkelä K T, Eskelinen A, Paavolainen P, Pulkkinen P, Remes V. Cementless total hip arthroplasty for primary osteoarthritis in patients aged 55 years and older. Acta Orthop 2010; 81(1): 42-52.

Mäkinen T J, Koort J K, Mattila K T, Aro H T. Precision measurements of the RSA method using a phantom model of hip prosthesis. J Biomech 2004; 37(4): 487-93.

Mäkinen T J, Alm JJ , Laine H, Svedström E, Aro H T. The incidence of osteopenia and osteoporosis in women with hip osteoarthritis scheduled for cementless total joint replacement. Bone 2007; 40(4): 1041-7.

Nazari-Farsani S, Finnilä S, Moritz N, Mattila K, Alm J J, Aro H T. Is modelbased radiostereometric analysis suitable for clinical trials of a cementless tapered wedge femoral stem? Clin Orthop Relat Res 2016; 474(10): 2246-53.

Noble P C, Box G G, Kamaric E, Fink M J, Alexander J W, Tullos H S. The effect of aging on the shape of the proximal femur. Clin Orthop Relat Res 1995; 316: 31-44.

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. Acta Orthop 2014; 85(2): 152-8.

Poolman R W, Verhaar J A N, Schreurs B W, Bom L P A, Nelissen R G H H, Koot H W J, Goosen J H M, Verheyen C C P M. Finding the right hip implant for patient and surgeon: the Dutch strategy—empowering patients. Hip Int 2015; 25(2): 131-7.

Ranstam J, Ryd L, Onsten I. Accurate accuracy assessment: review of basic principles. Acta Orthop Scand 2000; 71(1): 106-8.

Ryd L. The Gothenburg osseointegrated titanium (GOT) hip - a new addition to the orthopedic arsenal - whether osseointegrated or not. Acta Orthop 2006; 77(4): 547–8.

Van Rietbergen B, Huiskes R. Load transfer and stress shielding of the hydroxyapatite-ABG hip: a study of stem length and proximal fixation. J Arthroplasty 2001; 16(8 Suppl. 1): 55-63.

Sköldenberg O G, Salemyr M , Bodén H S, Lundberg A, Ahl T E, Adolphson P Y. A new uncemented hydroxyapatite-coated femoral component for the treatment of femoral neck fractures: two-year radiostereometric and bone densitometric evaluation in 50 hips. Bone Joint J 2011; 93-B(5): 665-77.

SooHoo N F, Vyas R M, Samimi D B, Molina R, Lieberman J R. Comparison of the responsiveness of the SF-36 and WOMAC in patients undergoing total hip arthroplasty. J Arthroplasty 2007; 22(8): 1168-73.

Thien T M, Ahnfelt L, Eriksson M, Strömberg C, Kärrholm J. Immediate weight bearing after uncemented total hip arthroplasty with an anteverted stem: a prospective randomized comparison using radiostereometry. Acta Orthop 2007; 78(6): 730-8.

Thien T M, Chatziagorou G, Garellick G, Furnes O, Havelin L I, Mäkelä K, Overgaard S, Pedersen A, Eskelinen A, Pulkkinen P, Kärrholm 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(19): e167.

Tonino A J, Van Der Wal B C H, Heyligers I C, Grimm B. Bone remodeling and hydroxyapatite resorption in coated primary hip prostheses. Clin Orthop Relat Res 2009; 467(2): 478-84.

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.

Acta Orthopaedica 2018; 89 (5): 490–495

Van Der Wal B C H, Rahmy A, Grimm B, Heyligers I, Tonino A. Preoperative bone quality as a factor in dual-energy X-ray absorptiometry analysis comparing bone remodelling between two implant types. Int Orthop 2008; 32(1):3 9-45.

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.

White C A, Carsen S, Rasuli K, Feibel R J, Kim P R, Beaulé P E. High incidence of migration with poor initial fixation of the Accolade stem. Clin Orthop Relat Res 2012; 470(2): 410-17.

495

Wolf O, Mattsson P, Milbrink J, Larsson S, Mallmin H. Periprosthetic bone mineral density and fixation of the uncemented CLS stem related to different weight bearing regimes: a randomized study using DXA and RSA in 38 patients followed for 5 years. Acta Orthop 2010; 81(3): 286–91.

Zebaze R M, Ghasem-Zadeh A, Bohte A, Iuliano-Burns S. Intracortical remodelling and porosity in the distal radius and post-mortem femurs of women: a cross sectional study. Lancet 2010; 375: 1729-36.

496

Acta Orthopaedica 2018; 89 (5): 496–502

Surgeon’s experience level and risk of reoperation after hip fracture surgery: an observational study on 30,945 patients in the Norwegian Hip Fracture Register 2011–2015 Ane L AUTHEN 1, Eva DYBVIK 2, Ove FURNES 1,2, and Jan-Erik GJERTSEN 1,2

1 Department of Clinical Medicine, University of Bergen, Bergen; 2 Norwegian Hip Fracture Register, Department of Orthopaedic Surgery, Haukeland University Hospital, Bergen, Norway Correspondence: jan-erik.gjertsen@helse-bergen.no Submitted 2018-03-16. Accepted 2018-05-09.

Background and purpose — In Norway, surgeons with variable levels of surgical experience manage hip fractures.

We investigated whether the experience level of the surgeon affected the risk of reoperation after hip fracture surgery.

Patients and methods — This is an observational register-based study of 30,945 hip fractures reported to the Norwegian Hip Fracture Register in the period 2011–2015.

An experienced surgeon was defined as a surgeon with more than 3 years’ experience in hip fracture treatment. If more than 1 surgeon performed the procedure, the most experienced surgeon defined the level of experience. Relative risks of reoperations were calculated with Cox regression analyses with adjustments for age groups, sex, and ASA class.

Results — Overall, patients operated by an inexperienced surgeon had a higher risk of reoperation compared with an experienced surgeon (RR = 1.2 (95% CI 1.1–1.4)).

Displaced femoral neck fractures (FNFs) had higher risk of reoperation regardless of operation method when managed by an inexperienced surgeon compared with an experienced surgeon (RR = 1.7 (95% CI 1.4–2.1)). Sub-analyses of other fracture types and operation methods showed no statistically significant differences between the 2 groups of experience.

Interpretation — FNFs operated by surgeons with less than 3 years’ experience in fracture treatment had a small increased risk of reoperation. The study indicates that experienced surgeons should manage displaced FNFs, and fractures operated with hemiprosthesis.

Approximately 9,000 hip fractures are operated annually in Norway. Roughly 25% of these patients die within the first year after the fracture (Figved et al. 2009, Gjertsen et al.

2010). It is therefore important to optimize the treatment and decrease the risk of reoperation in these patients.

Surgery for a hip fracture is most often done during the daytime, often by a consultant surgeon assisting a resident surgeon. In some cases, however, the resident surgeon operates without the assistance of a more experienced colleague. There are no guidelines in Norway as to whether hip fracture operations should be managed by surgeons that are more experienced and the practice differs from hospital to hospital.

Earlier studies indicate a direct relationship between surgeon volume and outcome in different surgical disciplines (Figved et al. 2009, Aquina et al. 2015, Damle et al. 2016, Kelly et al. 2016). Significant coherence has also been demonstrated between surgeon volume and outcome (Browne et al.

2009). 1 large register study reported that interns and junior residents performed half of all fracture-related surgery and that one-third of primary operations performed by junior residents were unsupervised (Andersen et al. 2014). Some studies have addressed surgeon experience in regards to the outcome after hip fracture surgery, though the results are ambiguous.

1 clinical study reported a higher reoperation rate among socalled demanding hip fracture procedures, if residents had not been supervised (Palm et al. 2007). Another study found no statistically significant difference between residents and consultant surgeons in reoperation rates, but higher mortality after procedures performed by residents (Khunda et al. 2013).

The Norwegian Hip Fracture Register (NHFR) has since 2005 collected nationwide information on all hip fractures in Norway (Gjertsen et al. 2008). In the present study, we used data from the NHFR to investigate whether there were any differences in reoperation rates between hip fracture operations performed by an experienced surgeon compared with an inexperienced surgeon.

© 2018 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/4.0) DOI 10.1080/17453674.2018.1481588

Acta Orthopaedica 2018; 89 (5): 496–502

Patients and method The Norwegian Hip Fracture Register (NHFR) Since 2005, surgeons managing hip fractures have voluntarily reported all hip fracture operations to the NHFR on a standardized 1-page questionnaire. Patient information registered includes age, sex, ASA classification, cognitive function, and national ID number (Gjertsen et al. 2008). A reoperation was defined as any secondary operation performed due to complications after the primary operation including both major reoperations such as re-osteosynthesis or secondary prosthesis and minor reoperations such as removal of implants, soft tissue debridement for infection, and closed reduction of a dislocated prosthesis. The national ID number allows link of an eventual reoperation to the former operation and linkage to death and emigration by Statistics Norway. In this way, the NHFR monitors the outcome of the operation. Furthermore, classification of the fracture, type of operation, cause and type of reoperation, information on implants, and duration of the procedure is registered.

The surgeons classified the intracapsular femoral neck fractures (FNFs) as undisplaced (Garden 1 or 2) or displaced (Garden 3 or 4). Extracapsular fractures were divided into basocervical FNFs, trochanteric fractures, and subtrochanteric fractures. The trochanteric fractures were further classified into 2-part trochanteric fractures (AO/OTA A1), multifragment trochanteric fractures (AO/OTA A2) or intertrochanteric fractures (AO/OTA A3). In 2011, information on surgeon’s experience was added to the questionnaire. Surgeon’s experience in fracture surgery is classified into more or less than 3 years. The question regarding surgeon’s experience is answered by the operating surgeon after each operation. The name and position of the surgeon is not registered in the database. There is no information on the number of procedures performed by the individual surgeon and each surgeon’s experience level is defined by number of years performing fracture surgery. If more than 1 surgeon performed the procedure, the most experienced surgeon defined the level of experience. For the time-period 2008–2014, a completeness analysis of the NHFR was conducted by comparing the registry with the Norwegian patient registry (NPR) (Havelin et al. 2015, 2016). Completeness of primary operations in the NHFR was 91–94% for hemiprosthesis and 80–86% for osteosynthesis. The completeness for reoperations after hemiarthroplasty and osteosynthesis has been found to be 68% and 65% respectively when compared with the NPR (Furnes et al. 2017).

Data and study sample Between January 1, 2011 and December 31, 2015, 36,538 hip fractures were reported to the NHFR on the new questionnaire including data on surgeons’ experience level. 30,945 cases were eligible for the study (Figure 1).

497

Hip fractures reported to the NHFR 2011–2015 n = 36,538 Excluded (n = 5,593): – surgeon’s experience data missing, 2,201 – patients < 60 years, 1,802 – incomplete data a, 1,590 Study population n = 30,945 Surgeon’s experience < 3 years n = 5,821

Surgeon’s experience > 3 years n = 25,124

Figure 1. Description of selection of patient study group from the NHFR. a Incomplete data on fracture type, operation type, ASA classification or operation time.

Statistics The Pearson chi-square test was used to compare categorical variables in the independent groups. Student’s t-test was used to compare continuous variables. The significance level was set at 0.05. Cox survival analyses were done to calculate risk of reoperations and risk of death. In all regression models surgical experience > 3 years was used as the reference. 95% confidence intervals (CIs) were calculated for relative risks.

Adjustments were done for age groups, sex, and ASA class.

The proportional hazards assumption was investigated visually by use of log-minus-log plots. For displaced femoral neck fractures treated with hemiarthroplasty the curves crossed each other after 60 days. We therefore performed additional Cox regression analyses with the follow-up divided into 2 time-periods. The first period ran from the day of surgery until 60 days postoperatively and the second period commenced at 60 days and ran until December 31, 2015. The proportional hazard was fulfilled within the 2 time-periods. Sub-analyses were done for different fracture types and operation methods.

Further, the Cox model was used to construct survival curves for all fractures and for displaced femoral neck fractures with adjustments for age group, physical status (ASA score), and sex. Adjustments for patients operated on both sides were not done, since an earlier study has shown that this will not alter the conclusion for the entered covariates (Lie et al. 2004).

The analyses were performed using IBM-SPSS version 24 software (IBM Corp, Armonk, NY, USA) and the cmprsk Library in the statistical package “R” (https://cran.r-project. org/Package=cmprsk).

Ethics, funding, and potential conflicts of interest The NHFR has permission from the Norwegian Data Inspectorate to collect patient data based on written consent from the patients. (Permission issued January 3, 2005; reference Number 2004/1658-2 SVE/-). Informed consent from patients is entered in the medical records at each hospital. The Norwegian Hip Fracture Register is financed by the Western Norway

498

Acta Orthopaedica 2018; 89 (5): 496–502

Table 1. Characteristics of all patients according to surgeon’s experience

Distribution (%) 0

10

20

30

40

50

60

70

80

90 100

Total number

Factor

< 3 years

> 3 years

Total number 5,821 25,124 Mean age (SD) at fracture 83 (8.6) 83 (8.5) Women, n (%) 4,113 (71) 17,625 (70) ASA score, n (%):

ASA 1–2 2,239 (39) 8,605 (34) ASA 3–5 3,528 (62) 16,519 (66) Operation time according to fracture type, minutes (SD): All fractures 57 (28) 64 (32) Undisplaced FNFs 32 (18) 32 (23) Displaced FNFs 69 (29) 72 (26) Trochanteric 2-fragmented 58 (21) 49 (22) Trochanteric multi-fragmented 63 (25) 60 (30) Inter- /subtrochanteric 79 (32) 91 (41) Operation time according to type of operation, minutes (SD): Screw osteosynthesis 30 (13) 25 (13) Hemiarthroplasty 78 (25) 76 (25) Sliding hip screw 62 (23) 60 (30) Long intramedullary nail 86 (36) 93 (41)

p-value

Screws Hemiarthroplasty

0.08 a 0.4 b < 0.001 b

Hip Compression Screw (HCS) HCS with lateral support plate Short intramedullary nail Long intramedullary nail

< 0.001 a 0.3 a < 0.001 a < 0.001 a < 0.001 a < 0.001 a < 0.001 a 0.06 a 0.001 a 0.01 a

ASA: American Society of Anesthesiologists.

FNF: femoral neck fracture. a Independent samples t-test.

b Pearson chi-square test.

Regional Health Authority (Helse-Vest). No competing interests declared.

Results Baseline characteristics The average age of the patients was the same, 83 years, for those operated by experienced surgeons and for those operated by inexperienced surgeons. Patients operated by experienced surgeons had statistically significantly more comorbidity than patients operated by inexperienced surgeons (ASA class 3–5: 66% vs. 62 %, respectively) (Table 1).

Other > 3 years experience

< 3 years experience

Figure 2. Proportion of procedures performed by experienced surgeons and inexperienced surgeons.

The patients were divided into subgroups to investigate the risk of reoperation for the different fracture types and operation methods (Table 2).

Fraction of operations performed by inexperienced surgeons Experienced surgeons participated in 25,124 (81%) of the total number of 30,945 hip fracture operations. The proportion of experienced surgeons for the different surgical methods was 70–92% (Figure 2). The highest fraction of experienced surgeons was found for hemiarthroplasties and operations with a long intramedullary (IM) nail. Screw osteosyntheses and operations with a hip compression screw were the procedures most likely to be performed by inexperienced surgeons.

Change in experience level in the period 2011–2015 In the period 2011–2015, an increasing proportion of operations was performed by experienced surgeons (Figure 3). This tendency was present for almost all operation methods. However, for osteosyntheses with hip compression screw with/ without lateral support plate the proportion of procedures performed by experienced surgeons increased from 2011 to 2013, but decreased again from 2013 to 2015 (Figure 3).

Table 2. Characteristics of patients according to fracture type

Factor Total number Mean age at fracture Women, n (%) ASA score, n (%): ASA 1–2 ASA 3–5 Duration of surgery, minutes Experienced surgeons, n (%) Median follow-up, years FNF: femoral neck fracture.

Undisplaced FNFs 4,220 80 2,292 (69)

1,730 (41) 2,490 (59) 32 2,954 (70) 1.4

Displaced Trochanteric Trochanteric Intertrochanteric/ FNFs 2-fragmented multifragmented subtrochanteric 13,098 5,077 5,000 83 83 84 9,023 (69) 3,554 (70) 3,750 (75)

4,453 (34) 1,777 (35) 1,700 (34) 8,645 (66) 3,300 (65) 3,300 (66) 72 52 61 11,919 (91) 3,503 (69) 3,750 (75) 1.3 1.2 1.2

2,303 83 1,750 (76) 806 (35) 1,497 (65) 89 2,027 (88) 1.2

Acta Orthopaedica 2018; 89 (5): 496–502

499

and shorter for screw osteosyntheses and sliding hip screws when the operation was performed by an experienced surgeon (Table 1).

Proportion (%) of procedures performed by experienced surgeons 100 95 90

Screws

85

Hemiarthroplasty

80

Hip Compression Screw (HCS)

75

HCS with lateral support plate

70

Short intramedullary nail

65

Long intramedullary nail

60

Other

55 50

2011

2012

2013

2014

2015

Figure 3. Changes in proportion of procedures performed by experienced surgeons.

Duration of surgery The duration of surgery was longer for displaced femoral neck fractures and sub-/intertrochanteric fractures and shorter for 2-fragmented and multi-fragmented trochanteric fractures when the operation was performed by an experienced surgeon (Table 1). Further, duration of surgery was longer for IM nails

Reoperations There was an increased risk of reoperation for patients operated by an inexperienced surgeon compared with patients operated by an experienced surgeon (5.3% vs. 4.2%, RR = 1.2 (CI 1.1–1.4)) (Table 3 and Figure 4a).

For undisplaced femoral neck fractures (FNFs) the risk of reoperation was similar between inexperienced and experienced surgeons (Table 3). Also, when performing sub-analyses including only screw osteosynthesis for those fractures, reoperation rates due to failure of the osteosynthesis (RR = 0.93, p = 0.7) or due to avascular necrosis of the femoral head (RR = 1.3, p = 0.3) were similar between inexperienced and experienced surgeons There were no reoperations after hemiprosthesis for undisplaced femoral neck fractures performed by inexperienced surgeons. Accordingly, it was not possible to calculate risk estimates for this group.

Some fracture types and certain operation methods had a higher risk of reoperation after operation by inexperienced

Table 3. Reoperation risk for surgeon experience level: all fracture types/all operation methods

Operation method < 3 years’ experience > 3 years’ experience RR a

Total Reop. (%) Total Reop. (%)

All fracture types: All operation methods 5,821 403 (6.9) 25,124 1,411 (5.6) 1.2 Hemiarthroplasty 1045 53 (5.1) 11,619 470 (4.0) 1.3 Screw osteosynthesis 1,379 194 (14) 3,209 415 (13) 1.1 Hip compression screw 2,619 115 (4.4) 6,257 325 (5.2) 0.87 Short intramedullary nail 508 21 (4.1) 2,329 100 (4.3) 0.94 Long intramedullary nail 168 9 (5.4) 1,343 67 (5.0) 1.04 Undisplaced FNFs: All operation methods 1,267 141 (11) 2,953 296 (10) 1.1 Screw osteosynthesis 1,178 139 (12) 2,519 273(11) 1.1 Hemiprosthesis 47 0 (0) 334 16 (4.8) c Displaced FNFs:

All operation methods 1,206 109 (9.0) 11,892 608 (5.1) 1.7 Screw osteosynthesis 198 54 (27) 675 137 (20) 1.4 Hemiprosthesis 982 52(5.3) 11,046 442 (4.0) 1.3 Trochanteric 2-fragmented fractures:

All operation methods 1,581 41 (2.6) 3,496 93 (2.7) 0.97 Hip compression screw 1,278 30 (2.3) 2,482 62 (2.5) 0.95 Intramedullary nail 271 8 (3.0) 948 28 (3.0) 0.95 Trochanteric multifragmented fractures: All operation methods 1,229 65 (5.3) 3,771 198 (5.3) 1.0 Hip compression screw 949 53 (5.6) 2,394 127 (5.3) 1.1 Intramedullary nail 264 11 (4.2) 1,290 65 (5.0) 0.81 Intertrochanteric and subtrochanteric fractures: All operation methods 287 24 (8.4) 2,016 127 (6.3) 1.3 Hip compression screw 173 17 (9.8) 767 68 (8.9) 1.1 Intramedullary nail 109 7 (6.4) 1,219 57 (4.8) 1.5

FNF: femoral neck fracture. a RR relative risk for reoperation. Surgeon’s experience > 3 years set to 1. b Cox regression analyses with adjustments for age group, gender, and ASA class. c RR could not be calculated due to no reoperations.

95% CI p-value b

1.1–1.4 0.95–1.7 0.94–1.3 0.70–1.1 0.59–1.5 0.52–2.1

0.001 0.1 0.2 0.2 0.8 0.9

0.92–1.4 0.91–1.4 c

1.4–2.1 1.0–1.9 0.99–1.8

0.67–1.4 0.61–1.5 0.43–2.1

0.2 0.3 c < 0.001 0.04 0.06

0.77–1.4 0.80–1.5 0.43–1.5

0.9 0.6 0.5

0.87–2.1 0.62–1.8 0.67–3.3

0.2 0.8 0.3

0.9 0.8 0.9

500

A

Acta Orthopaedica 2018; 89 (5): 496–502

B

C

D

Figure 4. Cox regression curves for implant survival after different fracture types and operation methods adjusted for age groups, gender, and ASA class. A. All fractures/all operations methods (RR = 1.2 (1.1–1.4), p = 0.001). B. Displaced femoral neck fractures/all operation methods (RR = 1.7 (1.4–2.1), p < 0.001). C. Displaced femoral neck fractures/screw osteosynthesis (RR = 1.4 (1.0–1.9), p = 0.001). D. Displaced fractures/ hemiprosthesis (RR = 1.3 (0.99–1.8), p = 0.06).

surgeons (Table 3). For displaced FNFs there was a higher risk of reoperation after operations performed by inexperienced surgeons (RR = 1.7 (CI 1.4–2.1)) (Table 3 and Figure 4b). Sub-analysis showed that displaced FNFs operated with screw osteosynthesis by inexperienced surgeons had a higher risk of reoperation (RR = 1.4 (CI 1.0–1.9)) (Table 3 and Figure 4c). Further, displaced FNFs operated with hemiprosthesis had more reoperations after operation by an inexperienced surgeon when analyzing the whole study period, but the difference was not statistically significant (RR = 1.3 (CI 0.99–1.8)) (Table 3 and Figure 4d). However, there was an increased risk of reoperation in the first 60 days after hemiarthroplasty for displaced femoral neck fracture performed by inexperienced surgeons (RR = 1.5 (CI 1.1–2.0)). After 60 days no statistically significant difference could be found (RR = 0.64 (CI 0.20–2.1)). There was an increased risk of reoperation due to dislocation in the first 60 days after hemiprosthesis operation by an inexperienced surgeon (RR = 2.0 (CI 1.1–3.9)). There was no difference in risk of reoperation due to infection after hemiprosthesis performed by inexperienced surgeons compared with experienced surgeons (RR = 1.1, p = 0.8).

For other fracture types and operation methods no statistically significant differences in risk of reoperation could be found between inexperienced and experienced surgeons (see Table 3).

When performing analyses with adjustments also for duration of surgery similar results were found.

Mortality 30-day mortality was 8.2% after operation by an experienced surgeon and 7.9% after operation by an inexperienced surgeon (RR = 0.98 (CI 0.89–1.1)). 1-year mortality was 24% after surgery performed by an experienced surgeon and 25% after operation by an inexperienced surgeon (RR = 1.0 (CI 0.99–1.1)).

Discussion We found statistically significant more reoperations after hip fracture operations performed by inexperienced surgeons compared with experienced surgeons, which was also the case in some subgroups of fractures and surgical methods. Inexperienced surgeons could perform operations for undisplaced FNFs without an increased risk of reoperation. On the other hand, displaced FNFs treated with screw osteosynthesis by inexperienced surgeons resulted in an increased risk of reoperation. Further, displaced FNFs operated with hemiarthroplasty by inexperienced surgeons showed an increased risk for both reoperation of any cause and reoperation due to dislocation the first 60 days postoperatively.

Findings in an observational, register-based study are less conclusive then those from randomized clinical trials. However, if confounders are corrected for, the level of evidence in validated national registry studies should be respected, reflecting the actual real-life national status. A considerable strength of this study was the large study sample, including more than 30,000 hip fractures from a whole country. Thus, the external validity is high.

Another strength of this study is the high reporting rate to the NHFR by Norwegian hospitals. When compared with the Norwegian Patient Register (NPR), approximately 90% of all hip fractures operations performed in Norway during the period 2011–2014 were registered in the NHFR (Havelin et al. 2016). Registration of reoperations, on the other hand, had a lower completeness, which gives an inaccurate estimation of the risk for reoperation in total. When compared with the NPR, 70% of reoperations were registered in the NHFR (Furnes et al. 2017). This annual report suggests that there is some degree of uncertainty connected to these registrations.

However, we have no indication of systematic underreporting, or that the completeness of reporting of reoperations is differ-

Acta Orthopaedica 2018; 89 (5): 496–502

ent in the 2 groups compared, so the relative risk of reoperations should not be affected.

It can be argued that 1 or 2 years of experience perhaps is sufficient to be defined as an experienced surgeon due to the high number of hip fractures, and thus the 3-year limit is too high. Additionally, it is likely that the number of years of experience in fracture treatment does not necessarily reflect the volume of hip fracture operations that a surgeon has performed. Therefore, within both groups there can be some variation in how experienced the surgeons actually are in performing the different operation methods.

Another limitation was that experienced surgeons performed a majority of the procedures. The numbers managed by experienced surgeons was also increasing every year throughout the period we studied. Accordingly, it was difficult to find statistically significant differences in reoperations related to surgeons’ experience because the comparison basis was unbalanced. However, this trend is very positive in the surgical care of these elderly and frail patients.

In the NHFR, there is no information on radiological data.

Accordingly, the quality of the primary operations (e.g., implant position and fracture reduction) cannot be assessed.

Further, it is not possible to assess whether an eventual reoperation was caused by a new trauma or due to failure of an osteosynthesis.

Palm et al. (2007) did a prospective study including 600 patients with proximal femoral fracture and assessed the influence of surgeon’s experience and supervision on reoperation rate. In a multivariate analysis with patient demographics, their results showed unsupervised junior registrars to have a higher reoperation rate among what they defined as technically demanding procedures (primarily displaced femoral neck fractures and comminuted trochanteric fractures). Also, the group then introduced mandatory supervision for some procedures and a “driver’s license” for others (Palm et al.

2012). This is in accordance with our results.

Shervin et al. (2007) reviewed existing literature on the association between surgeon procedure volume and improved patient outcomes in orthopedic surgery. Their result suggested an association between higher surgeon volume and lower rates of hip dislocation. In addition, they found that surgeon volume was strongly related to revision of arthroplasties.

Khunda et al. (2013) reviewed the records of 761 patients who underwent surgery for proximal femoral fracture, to determine whether surgeon’s experience and volume was associated with 6-month mortality and reoperation rate. They concluded that the mortality rate within 6 months was 80% higher in patients operated by inexperienced surgeons. However, the patients operated by inexperienced surgeons were older, and the fractures were generally more complex in the consultant group, which could be a considerable weakness of that study. They found no association between increased reoperation risk and surgery performed by inexperienced surgeons.

501

Browne et al. (2009) and co-workers did a retrospective cohort study, including 97,894 patients surgically treated for hip fracture. The study addressed whether surgeon’s volume was associated with mortality or nonfatal morbidity. The mortality rate and incidence of transfusion, pneumonia, and decubitus ulcer was higher for patients managed by surgeons with low volume (< 7 procedures/year). Further, operations performed by low-volume surgeons were associated with nonfatal morbidity and longer hospital stay.

In the present study, the mortality was similar for patients managed by experienced and inexperienced surgeons. However, more patients with severe comorbidity (ASA 3 or ASA 4) and more patients with complex fractures were managed by experienced surgeons, and even if comorbidity and fracture type were adjusted for, there is a possibility that these differences still might have influenced the results.

Bjorgul et al. (2011) did a prospective study, including 1,780 hip fracture procedures. Their aim was to identify and characterize learning curves in hip fracture surgery. The results showed that mean operating time decreased for 4 different surgical procedures, though at different rates. This indicated unique learning curves for the 4 different procedures. This is in contrast to our findings where the durations of surgery for experienced surgeons compared with inexperienced surgeons were similar for hemiarthroplasty and longer for long IM nail.

A possible explanation for this might be that the operations performed by experienced surgeons were more technically demanding. Less technically demanding operations, such as screw osteosyntheses and sliding hip screws, had a shorter operating time when performed by experienced surgeons.

Explanations and interpretations Few studies have addressed surgical experience and reoperation risk, and they lack consensus. We found an association between low experience with fracture surgery and higher reoperation risk after hemiarthroplasty and screw osteosynthesis for displaced FNFs. A similar increased risk for reoperation could not be found for undisplaced FNFs. An explanation could be that inexperienced surgeons do not manage to reduce the displaced FNFs properly or have problems in positioning the osteosynthesis material correctly when treating these fractures with osteosynthesis.

Correct anteversion of the femoral stem and correct leg length are factors important for stability of a hemiprosthesis.

Experience is required to assess this correctly intraoperatively.

A national registry can find otherwise hidden associations, e.g., between surgical inexperience and a higher reoperation rate, which merits new clinical studies with a higher degree of details such as, e.g., exact surgical technique and optimal reduction.

Today, surgeons with low experience perform a considerable proportion of hip fracture operations in Norway. The differences in risk of reoperations we found imply that experienced surgeons should manage some hip fracture types and opera-

502

tion methods. Based on our results the number needed to harm (NNH) for displaced FNFs was 25 (i.e., surgery performed by inexperienced surgeons resulted in 1 extra reoperation for every 25 operations compared with operations performed by experienced surgeons). If junior surgeons perform surgery for displaced femoral neck fractures, supervision by more experienced surgeons should be mandatory. For less technically demanding operations, junior surgeons could operate alone after they have obtained a “driver’s license” for that procedure, as described by Palm et al. (2012).

Conclusion Our findings suggest that surgeon’s experience has an impact on the risk of reoperation, in particular in some fracture types and operation methods. Experienced surgeons should manage or supervise all displaced femoral neck fractures, regardless of operation type, and all hemiprostheses for hip fractures.

The authors would like to thank the orthopedic surgeons in Norway for loyally reporting data on hip fracture operations to the NHFR. Our study was planned and designed by ALA and JEG. ED and JEG performed the statistical analyses. ALA wrote the manuscript. All authors participated in the interpretation of data, and critical revision of the manuscript.

Acta thanks Henrik Palm and Martyn Parker for help with peer review of this study.

Andersen M J, Gromov K, Brix M, Troelsen A. The Danish Fracture Database can monitor quality of fracture-related surgery, surgeons’ experience level and extent of supervision. Dan Med J 2014; 61(6): A4839.

Aquina C T, Probst C P, Kelly K N, Iannuzzi J C, Noyes K, Fleming F J, Monson J R T. The pitfalls of inguinal herniorrhaphy: surgeon volume matters. Surgery 2015; 158(3): 736-46.

Bjorgul K, Novicoff W M, Saleh K J. Learning curves in hip fracture surgery.

Int Orthop 2011; 35(1): 113-9.

Acta Orthopaedica 2018; 89 (5): 496–502

Browne J A, Pietrobon R, Olson S A. Hip fracture outcomes: does surgeon or hospital volume really matter? J Trauma 2009; 66(3): 809-14.

Damle R N, Flahive J M, Davids J S, Sweeney W B, Sturrock P R, Maykel J A, Alavi K. Surgeon volume correlates with reduced mortality and improved quality in the surgical management of diverticulitis. J Gastrointest Surg 2016; 20(2): 335-42.

Figved W, Opland V, Frihagen F, Jervidalo T, Madsen J E, Nordsletten L. Cemented versus uncemented hemiarthroplasty for displaced femoral neck fractures. Clin Orthop Relat Res 2009; 467(9): 2426-35.

Furnes O, Engesæter L, Hallan G, Fjeldsgaard K, Gundersen T, Gjertsen J E, Fenstad A M, Dybvik E, Bartz-Johannessen C. Annual report Norwegian National Advisory Unit on Arthroplasty and Hip Fractures; 2017.

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): 58393.

Gjertsen J E, Vinje T, Engesaeter L B, Lie S A, Havelin L I, Furnes O, Fevang J M. Internal screw fixation compared with bipolar hemiarthroplasty for treatment of displaced femoral neck fractures in elderly patients. J Bone Joint Surg Am 2010; 92(3): 619-28.

Havelin L I, Furnes O, Engesaeter L B, Fenstad A M, Bartz-Johannessen C, Fjeldsgaard K, Wiig O, Dybvik E. The Norwegian Arthroplasty Register: annual report; 2015. ISBN: 978-82-91847-20-7. ISSN: 1893-8914.

Havelin L I, Furnes O, Engesaeter L B, Fenstad A M, Bartz-Johannessen C, Dybvik E, Fjeldsgaard K, Gundersen T. The Norwegian Arthroplasty Register: annual report; 2016. ISBN: 978-82-91847-21-4. ISSN: 1893-8914.

Kelly E C, Winick-Ng J, Welk B. Surgeon experience and complications of transvaginal prolapse mesh. Obstet Gynecol 2016; 128(1): 65-72.

Khunda A, Jafari M, Alazzawi S, Mountain A, Hui A C. Mortality and reoperation rate after proximal femoral fracture surgery by trainees. J Orthop Surg (Hong Kong) 2013; 21(1): 87-91.

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.

Palm H, Jacobsen S, Krasheninnikoff M, Foss N B, Kehlet H, Gebuhr P; Hip Fracture Study Group. Influence of surgeon’s experience and supervision on re-operation rate after hip fracture surgery. Injury 2007; 38(7): 775-9.

Palm H, Krasheninnikoff M, Holck K, Lemser T, Foss N B, Jacobsen S, Kehlet H, Gebuhr P. A new algorithm for hip fracture surgery: reoperation rate reduced from 18% to 12% in 2,000 consecutive patients followed for 1 year. Acta Orthop 2012; 83(1): 26-30.

Shervin N, Rubash H E, Katz J N. Orthopaedic procedure volume and patient outcomes: a systematic literature review. Clin Orthop Relat Res 2007; 457: 35-41.

Acta Orthopaedica 2018; 89 (5): 503–508

503

Higher cartilage wear in unipolar than bipolar hemiarthroplasties of the hip at 2 years: A randomized controlled radiostereometric study in 19 fit elderly patients with femoral neck fractures Wender FIGVED 1, Stian SVENØY 1,2, Stephan M RÖHRL 2,3, Jon DAHL 2, Lars NORDSLETTEN 2,3, and Frede FRIHAGEN 2,3 1 Orthopaedic

Department, Bærum Hospital, Vestre Viken Hospital Trust; 2 Division of Orthopaedic Surgery, Oslo University Hospital; 3 Institute of Clinical Medicine, Faculty of Medicine, University of Oslo, Norway Correspondence: wender.figved@vestreviken.no Submitted 2018-03-13. Accepted 2018-05-02.

Background and purpose — The use of unipolar hemi­ arthroplasties for femoral neck fractures is increasing in some countries due to reports of higher reoperation rates in bipolar prostheses. On the other hand, it has been proposed that bipolar hemiarthroplasties have clinical advantages and less cartilage wear than unipolar hemiarthroplasties. We compared cartilage wear between bipolar and unipolar hemi­ arthroplasties using radiostereometric analyses (RSA), in patients aged 70 years or older.

Patients and methods — 28 ambulatory, lucid patients were randomized to treatment with a unipolar or a bipolar hemiarthroplasty for an acute femoral neck fracture. Migra­ tion of the prosthetic head into the acetabulum was measured using RSA. Secondary outcomes were Harris Hip Score (HHS), and EQ-5D scores. Patients were assessed at 3, 12. and 24 months.

Results — 19 patients were available for follow-up at 2 years: mean proximal penetration was 0.83 mm in the unipo­ lar group and 0.24 mm in the bipolar group (p = 0.01). Mean total point movement was 1.3 mm in the unipolar group and 0.95 mm in the bipolar group (p = 0.3). Median HHS was 78 (62–96) in the unipolar group and 100 (70–100) in the bipolar group (p = 0.004). Median EQ-5D Index Score was 0.73 (0.52–1.00) in the unipolar group and 1.00 (0.74–1.00) in the bipolar group (p = 0.01). Median EQ-5D VAS was 70 (50–90) in the unipolar group and 89 (70–95) in the bipolar group (p = 0.03) Interpretation — Patients with unipolar hemiarthroplas­ ties had higher proximal cartilage wear and lower functional outcomes. Unipolar hemiarthroplasties should be used with caution in ambulatory, lucid patients.

For displaced femoral neck fractures, total hip arthroplasty (THA) may be the best option for healthier, active patients (Hopley et al. 2010, Burgers et al. 2012), while unipolar or bipolar hemiarthroplasty is the most common treatment in elderly patients (Miller 2013, Rogmark and Leonards­ son 2016). A unipolar hemiarthroplasty (UHA) articulates between the large metal head and the acetabulum, while a bipolar hemiarthroplasty (BHA) also articulates between an inner metal head and the polyethylene of a larger head with an outer metal shell.

A systematic review of 10 randomized controlled trials (RCTs) found similar or better results for BHA compared with UHA in hip function, hip pain, and quality of life, and no differences in mortality, reoperation, dislocation, and com­ plications. Furthermore, BHA showed less cartilage wear at 1 year, but no differences at 4 months, 2 years, and 4 years (Jia 2015). No studies have shown a clear correlation between cartilage erosion and clinical manifestations of the hip joint.

1 RCT using radiostereometric analyses (RSA) of cartilage wear in hemiarthroplasties showed increased wear in the UHA group at 2 years (Jeffcote et al. 2010). Decision-making is still difficult due to contradictory results of clinical trials, price dif­ ferences in some markets, and the possibility of variances in properties between different hemiarthroplasty components.

We compared wear between a UHA and a BHA up to 2 years, using RSA and functional outcome scores, in patients 70 years and older with femoral neck fractures, with a hypoth­ esis of equivalence between the groups.

Patients and methods The trial was conducted at the orthopedic department of Baerum Hospital, Norway. Recruitment was from October

© 2018 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/4.0) DOI 10.1080/17453674.2018.1475899

504

Acta Orthopaedica 2018; 89 (5): 503–508

Patient characteristics at the time of enrollment

Characteristics Median age (range) Female sex Mean preoperative Harris Hip Score (SD) Mean preoperative EQ-5D Index (SD) Mean preoperative EQ-5D VAS (SD) Median outer head size (range)

Unipolar Bipolar (n = 14) (n = 14) 81 (70–90) 11 94 (6) 0.90 (0.12) 72 (17) 49 (45–53)

Assessed for eligibility: Femoral neck fractures presented to the hospital during the trial period (n = 271) Did not meet inclusion criteria (n = 223): – aged < 65, 9 – undisplaced fracture, 71 – previous fracture of the same hip, 2 – previous symptomatic hip pathology, 4 – fracture caused by malignant disease, 3 – ongoing infectious disease, 4 – mental impairment, 60 – dependent on walking aids, 39 – living in institution, 31

80 (70–89) 11 96 (4) 0.91 (0.11) 79 (16) 48 (46–54)

2009 to April 2011. Patients aged 70 years or older with a displaced intracapsular femoral neck fracture were eligible for inclusion. They had to be living independently and be able to walk without aids.

Patients with cognitive impairment, previous symp­ tomatic hip pathology such as osteoarthritis, a frac­ ture caused by malignant disease, or ongoing infec­ tious disease were excluded. Randomization was performed using a computer random number gen­ erator. Allocation was done by the surgeon on call using sealed envelopes. 28 patients were randomized to treatment with a cementless UHA or BHA for an acute femoral neck fracture (Table). Patients were followed at 3 months, 1 year, and 2 years. 19 patients were available for follow-up at 2 years (Figure 1).

Not included (n = 20): – refused to participate, 12 – surgeon on call failed to include, 4 – outside of hospital catchment area, 4 Randomized (n = 28) Allocated to unipolar hemiarthroplasty (n = 14) Received allocated treatment (n = 14)

Allocated to bipolar hemiarthroplasty (n = 14) Received allocated treatment (n = 14)

Follow-up at 3 months Outpatient clinic and RSA (n = 13) Converted to THA due to dislocation (n = 1) Lost to follow-up (n = 0)

Follow-up at 3 months Outpatient clinic and RSA (n = 12) Reoperated due to infection (n = 1) Withdrawn from trial (n = 1) Lost to follow-up: (n = 0)

Follow-up at 1 year Outpatient clinic and RSA (n = 12) Dead (n = 1) Lost to follow-up (n = 0)

Follow-up at 1 year Outpatient clinic and RSA (n = 10) Dead (n = 2) Lost to follow-up (n = 0)

Follow-up at 2 years Outpatient clinic and RSA (n = 10) Withdrawn from trial (n = 2) Lost to follow-up (n = 0)

Follow-up at 2 years Outpatient clinic and RSA (n = 9) Converted to THA due to dislocation (n = 1) Lost to follow-up (n = 0)

Intervention Patients were operated with a hemiarthroplasty Completed follow-up at 2 years (n = 10) Completed follow-up at 2 years (n = 9) using an uncemented press-fit hydroxyapatiteCompletely lost to follow-up (n = 0) Completely lost to follow-up (n = 0) Converted to THA due to dislocation (n = 1) Reoperated due to infection (n = 1) coated femoral stem (Corail, DePuy Orthopaedics Dead (n = 1) Withdrawn from trial (n = 1) Withdrawn from trial (n = 2) Converted to THA due to dislocation (n = 1) Inc, Warzaw, IN, USA). The BHA group received Analyzed (n = 10) Dead (n = 2) a 28 mm cobalt chromium head and a bipolar head Analyzed (n = 9) (Self-Centering™ Bipolar, DePuy Orthopaedics Inc, Warzaw, IN, USA). The UHA group received Figure 1. The CONSORT diagram shows the recruitment and flow of patients throughout the trial. a modular unipolar head (Modular Cathcart Unipo­ lar, DePuy Orthopaedics Inc, Warzaw, IN, USA).

Both head options were available in 1 mm size increments. Outcomes and data collection The diameter of the femoral head was measured using full The primary outcome was migration of the prosthetic head circular measurement templates during surgery (Jeffery and into the acetabulum. Migration was measured with UmRSA Ong 2000), and the corresponding prosthetic head size was software (RSA BioMedical, Umea, Sweden) using an RSA chosen (Table). Arthroplasty was performed through a poste­ cage 43 containing Ta markers for creation of 3D coordinates rior approach with the patient in the lateral decubitus position, and built-in film cassette holders placed behind the patient. using spinal anesthesia. 5 or 6 1 mm tantalum (Ta) spherical Radiostereometric examinations were conducted using 2 fixed markers were inserted in the pelvis around the acetabulum, and X-ray tubes angled approximately 40 degrees in relation to 3 in the anterior superior iliac spine, using an UmRSA Injector each other. The center of the outer head was determined by (RSA BioMedical, Umea, Sweden) (Figure 2). 6 experienced semi-automatic edge detection of the metal shell in the BHA surgeons conducted the procedures. All patients were given group, and the surface of the unipolar head in the UHA group; preoperative intravenous cefalotin 2 g and a further 3 doses 1 experienced analyzer localized the edge with 4 points, and in the first 12 hours after the operation. All patients received the software automatically detected the remaining edge points 5000 IU low molecular weight heparin subcutaneously daily according to the pixel difference (Borlin et al. 2006, Figved for at least 10 days. Early mobilization was encouraged, with et al. 2012) (Figure 2). The motion of the center of the outer weight bearing as tolerated. head was calculated relative to the rigid body segment created

Acta Orthopaedica 2018; 89 (5): 503–508

505

ference of 0.5 mm. A sample size of 6 patients in each group was calculated to be sufficient, with a 2-sided 95% confi­ dence interval (CI) and 95% power, to establish equivalence.

A margin of equivalence of 0.5 mm and a range of –0.5 to 0.5 was predefined as an acceptable range for the CI of the difference in wear. To compensate for loss to follow-up, com­ plications, and mortality, we decided to include 28 patients.

To avoid analyzing RSA measurements of patients that were converted to a total hip arthroplasty or reoperated for infec­ tion, a per-protocol design was used. For RSA analyses, HHS and EQ-5D scores, we used the nonparametric independentsamples Mann–Whitney U test. SPSS version 24 for Macin­ tosh (IBM Corp, Armonk, NY, USA) was used for statistical analyses.

Figure 2. The image on the left is a conventional radiograph showing a bipolar hemiarthroplasty of the right hip. Tantalum markers have been implanted around the acetabulum, in the anterior superior iliac spine, and in the superior pubic ramus. To the right is a radiostereometric image with tantalum markers in the pelvis (numbered) and tantalum markers in the calibration cage. The center of the outer prosthetic head has been computer-calculated using edge detection.

by the Ta markers in the pelvis, in all 3 planes. Although pub­ lished after study completion, the trial complies with the ISO standard for RSA (ISO copyright office 2013). However, for comparison with previous trials, and clinical relevance, proxi­ mal penetration (Y-axis) and the total point movement (TPM) of the femoral head were reported as a surrogate for wear of the acetabular cartilage.

The RSA index radiographs were taken within 1 week post­ operatively. To determine the precision of the RSA measure­ ments, all examinations were conducted in the supine position and repeated within 1 hour, with repositioning of the patient between the scans. The precision was then calculated from the mean difference between the double examinations at all time intervals. For analyses of cartilage wear, double examinations of all patients at all time intervals were compared, and the mean result of the analyses was recorded. Hip function was rated with Harris Hip Score (HHS), ranging from 0 to 100 points covering a maximum of 44 points for absence of pain, 47 points for function, and 9 points for range of motion and absence of deformity. Health-related quality of life was rated by the patient-reported EQ-5D using VAS and index scores.

After inclusion but prior to surgery, all patients completed an HHS and an EQ-5D, instructed to recall and assess their prefracture status (Table).

Statistics Sample size calculation was conducted using the equivalence criterion and the extension of the CONSORT statement on non-inferiority and equivalence trials, and was based on an assumed precision of 0.2 mm of our RSA measurements. With no previously published values for a clinically relevant differ­ ence in cartilage wear, we chose an arbitrarily selected dif­

Ethics, registration, funding, and potential conflicts of interests The protocol was approved by the regional ethics com­ mittee (S-08619b) and registered at Clinicaltrials.gov (NCT00746876). Patients provided written informed consent prior to surgery. The study was conducted in compliance with the Helsinki Declaration, and the CONSORT Statement. The study was founded by the 2 participating hospitals. The first author received a research grant for this study of NOK 50,000 from Smith & Nephew, Norway. There are no other conflicts of interest to be reported by any of the authors.

Results The precision of the measurements expressed by the mean dif­ ference between 91 double examinations was 0.029 mm for the X-axis (99% CI –0.007 to 0.065), 0.028 mm for the Y-axis (99% CI –0.005 to 0.060), and 0.009 mm for the Z-axis (99% CI 0.04 to 0.06).

The distribution of the markers in the pelvis was assessed using the condition number which was below 150 in all but 3 examinations in 3 different patients, which were then excluded from analyses (mean 119; median 66 (25–1387)). The stability of the markers was assessed using the mean error of rigid body fitting, which was below 0.35 in all cases (mean 17; median 0.17 (0.006–0.345)) (ISO copyright office 2013).

Mean proximal penetration (Y-axis) at 3 months was 0.023 mm in the UHA group and 0.083 mm in the BHA group (CI –0.4 to 0.2), at 1 year 0.43 mm in the UHA group and 0.23 mm in the BHA group (CI –0.07 to 0.5), and at 2 years 0.83 mm in the UHA group and 0.24 mm in the BHA group (CI 0.1 to 1.0) (Figure 3). The CI interval for the mean difference at 2 years was above zero but exceeded the equivalence margin of 0.5 mm, indicating a superior, and not equivalent, result (Figure 4).

Mean TPM at 3 months was 0.71 mm in the UHA group and 0.60 mm in the BHA group (CI –0.4 to 0.6), at 1 year 1.0 mm in the UHA group and 0.86 mm in the BHA group (CI –0.5 to

506

Acta Orthopaedica 2018; 89 (5): 503–508

Vertical (Y-axis) migration (mm)

Total point motion (3D; mm)

2.0

2.0

1.8 1.6

Unipolar Bipolar

Figure 4. The graph shows the mean difference of proximal penetration (Y-axis) and TPM between the two groups at 2 years, in millimeters. Error bars indicate 95% confidence intervals (CIs) for the mean difference. Blue area indicates zone of equivalence, defined as ± 0.5 mm (Delta). The CI for the Y-axis lies above zero but exceeds the zone of equivalence, indicating a superior result. For TPM, the 95% CI includes zero and exceeds the zone of equivalence, indicating a nonsignificant result.

p = 0.4

1.8 1.6

1.4

p = 0.01

1.2 1.0

1.4

p = 0.6

1.2 1.0

0.8

0.8

p = 0.1

0.6

0.6

0.4

0.4

0.2

0.2

0.0

0.0 3 months

1 year

2 years

3 months

Time after index operation

1 year

2 years

Time after index operation

Figure 3. Graph showing mean migration in the vertical plane (Y-axis) and total point motion (3D migration) of patients at 3 months, 1 year, and 2 years.

Harris Hip Score

EQ-5D Index Score

EQ-5D VAS

100

1.0

100

80

0.8

80

0.6

60

60 p = 0.05

40 20 0

p = 0.001

p = 0.004

0.2

Unipolar Bipolar

Preop

3 months

p = 0.01

0.4

1 year

2 years

Time after index operation

0

Unipolar Bipolar

Preop

3 months

20

1 year

2 years

∆

0

0

∆

Y-axis p = 0.01

TPM p = 0.3 –1.0

–0.5

0.0

0.5

1.0

Difference in migration at 2 years (mm)

40

p = 0.003

Time after index operation

← unipolar better bipolar better →

p = 0.02

p = 0.03

Unipolar Bipolar

Preop

3 months

1 year

2 years

Time after index operation

Figure 5. Box plots of Harris Hip Score, EQ-5D Index Score, and EQ-5D Visual Analogue Scale (VAS) at all time intervals. Boxes represent the middle 50% of the data, with the central band representing the median.

The ends of the whiskers represent the minimum and maximum of all of the data. Statistical analyses conducted using the non-parametric Mann–Whitney U test.

0.7), and at 2 years 1.3 mm in the UHA group and 0.95 mm in the BHA group (CI –0.4 to 1.1) (Figure 3). The CI interval for the mean difference at 2 years included zero and exceeded the equivalence margin of 0.5 mm, indicating a statistically nonsignificant result (Figure 4).

Median HHS, EQ-5D Index Score and EQ-5D VAS was higher in the BHA group at all time intervals, and statistically significantly higher at 2 years (Figure 5).

Discussion In this trial, patients treated with UHA had higher proximal cartilage wear. The wear rate corresponds well with the only previously published similar RSA trial (Jeffcote et al. 2010).

TPM, mediolateral (X-axis), and anteroposterior (Z-axis) migration were similar between the groups. Thus, we only detected a difference in cartilage erosion in the proximal direction (Y-axis), corresponding to wear of cartilage in the dome of the acetabulum. The Y-axis might be the best way of measuring early cartilage erosion, being the loadbearing direction. Jeffcote et al. (2010), also found differences in TPM at 1 and 2 years’ follow-up, favoring BHA. 1 RCT found more

radiological cartilage wear in the UHA group during the first 12 months. The difference diminished over time, and was no longer statistically significant at 2 years and 4 years. The study also suggests the mechanism of the BHA ceases to function after some time, and behaves like a UHA (Inngul et al. 2013).

This was also proposed in earlier studies (Chen et al. 1989, Eiskjaer et al. 1989).

A recent meta-analysis of RCTs comparing unipolar versus bipolar hemiarthroplasty for displaced femoral neck frac­ tures did not find an advantage of bipolar prostheses (Jia 2015). However, the review lacks information on whether the same femoral stem was used in both groups of the included trials. Two recent RCTs including hemiarthroplasties did not list revision due to cartilage wear as a problem, during 5 to 7 years’ follow-up (Støen et al. 2014, Langslet et al. 2014).

Although many patients in these studies did not have a late radiograph taken, the follow-up was good, and patients in pain were addressed.

The difference in HHS, EQ-5D Index, and VAS scores in this trial should be interpreted with caution. The sample size calculation conducted for this study lacks power for these secondary outcomes, and they should therefore be considered subsidiary, with a high risk of a Type 1 error—a false-positive

Acta Orthopaedica 2018; 89 (5): 503–508

result. We found a surprisingly large difference in favor of the BHA group in all secondary outcomes at 1 and 2 years. In our trial, prosthetic head migration was the primary outcome measure, and the sample size is too low to show a trustworthy difference in any of the functional outcome scales used. Our trial, however, recruited a fit subgroup of patients with femo­ ral neck fractures, so good clinical results would be expected (Hebert-Davies et al. 2012, Mundi et al. 2014).

Decision-making is still difficult due to contradictory results of clinical trials, and the possibility of variances in properties between different hemiarthroplasty components: The first step towards a bipolar hemiarthroplasty was introduced by Chris­ tiansen in the late 1960s (Christiansen 1969). This prosthesis had a built-in trunnion bearing that allowed some movement between the stem and the head of the prosthesis. The results were promising (Soreide et al. 1975, Meyer 1981), but acetab­ ular protrusion remained a problem (Søreide et al. 1980). The first true bipolar model with a ball and socket joint between the femoral stem and the prosthetic head was the Bateman (1974) hemiarthroplasty. The bipolar design was then used in similar models such as the Giliberty, Monk, and Hastings.

Many series with short- and long-term follow-up showed less pain and decreased protrusion of the acetabulum than in previ­ ous reports on UHA (Devas and Hinves 1983, LaBelle et al.

1990, Wetherell and Hinves 1990, Haidukewych et al. 2002, Isotalo et al. 2002). However, no randomized controlled trials comparing UHA with the newer BHA models were conducted until much later.

Early radiological studies of interprosthetic motion in bipo­ lar hemiarthroplasties showed little or no movement between the stem and the head over time when analyzing passive motion of the hip without weight-bearing (Bochner et al.

1988, Hodgkinson et al. 1988, Chen et al. 1989). Later studies analyzing the interprosthetic movement during weight-bear­ ing have, however, shown a preserved movement of the inner joint during the stance phase of gait (Wada et al. 1997, Gaine et al. 2000). One recent RSA study has shown steady-state wear over time (Tsukanaka et al. 2017).

Cartilage wear may also be measured by the rate of revi­ sion surgery. In a Swedish register study, Leonardsson et al.

(2012) found a lower risk for reoperations caused by erosion in the bipolar HA, though the total revision rate was very low (0.17%). Counting all reasons for revision surgery, they found a higher risk of early reoperation following bipolar hemiar­ throplasty compared with unipolar. The Australian National Joint Replacement Registry nonetheless found that bipolar prostheses had a decreased risk of revision than unipolar, at least in younger patients (Rogmark and Leonardsson 2016).

In the study from Inngul et al. (2013), there was no differ­ ence in revision rates between unipolar HA and bipolar HA. There was also no correlation between cartilage wear and clinical scores (EQ-5D index score and HHS). Baker et al.

(2006) found frequently radiological erosion in UHAs in lucid patients. Still, only a few were surgically revised. In a

507

Cochrane review including 7 trials (857 participants, 863 frac­ tures), no differences were found between UHA and BHA. The review analyzed clinical scores and complications. How­ ever, several of the studies included few patients (Parker et al. 2010). Variations in inclusion criteria may influence out­ comes: It would be reasonable to stipulate that using UHA in community walking individuals (Baker et al. 2006) would certainly increase the rate of wear, compared with studies on those with very limited walking ability.

Several authors comparing UHA and BHA have discussed the issue of price differences between the 2, with the BHA usually being the more expensive implant. In our trial, how­ ever, the bipolar head plus the inner head used in the BHA group had a lower price than the unipolar head used in the UHA group. Also, when this study started, the use of THA in hip fracture patients was not common. Today, our study pop­ ulation of individuals living independently and able to walk without aids is not the group recommended to have unipolar hemiarthroplasties, but rather THA (Hopley et al. 2010, Burg­ ers et al. 2012).

In summary, we found that patients treated with BHA had lower proximal cartilage wear than patients with UHA. The BHA group showed superior clinical outcomes, but an uncer­ tain observation because of few patients. Unipolar hemiarthro­ plasties should be used with caution in self-ambulatory, lucid patients.

Idea conception: WF, LN. Inclusion, surgery, and follow-up: WF, SS. Imple­ mentation of RSA technique: JD. RSA analyses: SR. Data analyses. WF, SS, JD, LN, FF. Manuscript preparation: all authors.

Thanks are offered to Gunhild Olseng and Elisabeth Gunby for radiographic and RSA services in accordance with the highest standards, and to Jan Erik Madsen, Finnur Snorrason, and Asbjørn Hjall for important parts in con­ ceiving the idea and facilitating this trial.

Acta thanks Lennard Koster and Cecilia Rogmark for help with peer review of this study.

Baker R, Squires B, Gargan M, Bannister G. Total hip arthroplasty and hemi­ arthroplasty in mobile, independent patients with a displaced intracapsular fracture of the femoral neck: a randomized, controlled trial. J Bone Joint Surg Am 2006; 88(12): 2583-9.

Bateman J E. The classic: single-assembly total hip prosthesis—preliminary report. Clin Orthop Relat Res 1974; 441: 16-18.

Bochner R M, Pellicci P M, Lyden J P. Bipolar hemiarthroplasty for fracture of the femoral neck: clinical review with special emphasis on prosthetic motion. J Bone Joint Surg Am 1988; 70(7): 1001-10.

Borlin N, Rohrl S, Bragdon C. RSA wear measurements with or without markers in total hip arthroplasty. J Biomech 2006; 39(9): 1641-50.

Burgers P T, Van Geene A R, van den Bekerom M P, Van Lieshout E M, Blom B, Aleem I S, et al. Total hip arthroplasty versus hemiarthroplasty for dis­ placed femoral neck fractures in the healthy elderly: a meta-analysis and systematic review of randomized trials. Int Orthop 2012; 36(8): 1549-60.

508

Chen S C, Badrinath K, Pell L H, Mitchell K. The movements of the com­ ponents of the Hastings bipolar prosthesis: a radiographic study in 65 patients. J Bone Joint Surg Br 1989; 71(2): 186-8.

Christiansen T. A new hip prosthesis with trunnion-bearing. Acta Chir Scand 1969; 135(1): 43-6.

Devas M, Hinves B. Prevention of acetabular erosion after hemiarthroplasty for fractured neck of femur. J Bone Joint Surg Br 1983; 65(5): 548-51.

Eiskjaer S, Gelineck J, Søballe K. Fractures of the femoral neck treated with cemented bipolar hemiarthroplasty. Orthopedics 1989; 12(12): 1545-50.

Figved W, Dahl J, Snorrason F, Frihagen F, Rohrl S, Madsen J E, et al. Radio­ stereometric analysis of hemiarthroplasties of the hip: a highly precise method for measurements of cartilage wear. Osteoarthritis Cartilage 2012; 20(1): 36-42.

Gaine W J, Sanville P R, Bamford D J. The Charnley-Hastings bipolar pros­ thesis in femoral neck fractures: a study of dynamic motion. Injury 2000; 31(4): 257-63.

Haidukewych G J, Israel T A, Berry D J. Long-term survivorship of cemented bipolar hemiarthroplasty for fracture of the femoral neck. Clin Orthop Relat Res 2002; (403): 118-26.

Hebert-Davies J, Laflamme G-Y, Rouleau D; HEALTH and FAITH investi­ gators. Bias towards dementia: are hip fracture trials excluding too many patients? A systematic review. Injury 2012; 43(12): 1978-84.

Hodgkinson J P, Meadows T H, Davies D R, Hargadon E J. A radiological assessment of interprosthetic movement in the Charnley-Hastings hemiar­ throplasty. Injury 1988;1 9(1): 18-20.

Hopley C, Stengel D, Ekkernkamp A, Wich M. Primary total hip arthroplasty versus hemiarthroplasty for displaced intracapsular hip fractures in older patients: systematic review. BMJ 2010; 340(June 11): c2332-2.

Inngul C, Hedbeck C J, Blomfeldt R, Lapidus G, Ponzer S, Enocson A. Uni­ polar hemiarthroplasty versus bipolar hemiarthroplasty in patients with displaced femoral neck fractures: a four-year follow-up of a randomised controlled trial. Int Orthop 2013; 37(12): 2457-64.

ISO copyright office. Implants for surgery. Roentgen stereophotogrammetric analysis for the assessment of migration of orthopaedic implants [Internet].

2013. Available from: https://www.iso.org/standard/55662.html Isotalo K, Rantanen J, Aärimaa V, Gullichsen E. The long-term results of Lubinus interplanta hemiarthroplasty in 228 acute femoral neck fractures: a retrospective six-year follow-up. Scand J Surg 2002; 91(4): 357-60.

Jeffcote B, Li M-G, Barnet-Moorcroft A, Wood D, Nivbrant B. Roentgen stereophotogrammetric analysis and clinical assessment of unipolar versus bipolar hemiarthroplasty for subcapital femur fracture: a randomized pro­ spective study. ANZ J Surg 2010; 80(4): 242-6.

Jeffery J A, Ong T J. Femoral head measurement in hemiarthroplasty: assess­ ment of interobserver error using 3 measuring systems. Injury 2000; 31(3): 135-8.

Acta Orthopaedica 2018; 89 (5): 503–508

Jia Z. Unipolar versus bipolar hemiarthroplasty for displaced femoral neck fractures: a systematic review and meta-analysis of randomized controlled trials. J Orthop Surg Res 2015; 10(1): 1-8.

LaBelle L W, Colwill J C, Swanson A B. Bateman bipolar hip arthroplasty for femoral neck fractures: a five- to ten-year follow-up study. Clin Orthop Relat Res 1990; (251): 20-5.

Langslet E, Frihagen F, Opland V, Madsen J E, Nordsletten L, Figved W. Cemented versus uncemented hemiarthroplasty for displaced femoral neck fractures: 5-year followup of a randomized trial. Clin Orthop Relat Res 2014; 472(4): 1291-9.

Leonardsson O, Kärrholm J, Åkesson K, Garellick G, Rogmark C. Higher risk of reoperation for bipolar and uncemented hemiarthroplasty. Acta Orthop 2012; 83(5): 459-66.

Meyer S. Prosthetic replacement in hip fractures: a comparison between the Moore and Christiansen endoprostheses. Clin Orthop Relat Res 1981; (160): 57-62.

Miller B J. The trends in treatment of femoral neck fractures in the Medicare population from 1991 to 2008. J Bone Joint Surg Am 2013; 95(18): e1321.

Mundi S, Chaudhry H, Bhandari M. Systematic review on the inclusion of patients with cognitive impairment in hip fracture trials: a missed opportu­ nity? Can J Surg 2014; 57(4): E141-5.

Parker M J, Gurusamy K S, Azegami S. Arthroplasties (with and without bone cement) for proximal femoral fractures in adults. Cochrane Database of Systematic Reviews (Online). 2010; (6): CD001706.

Rogmark C, Leonardsson O. Hip arthroplasty for the treatment of displaced fractures of the femoral neck in elderly patients. Bone Joint J 2016; 98-B(3): 291-7.

Soreide O, Lerner A P, Thunold J. Primary prosthetic replacement in acute femoral neck fractures. Injury 1975; 6(4): 286-93.

Søreide O, Lillestøl J, Alho A, Hvidsten K. Acetabular protrusion follow­ ing endoprosthetic hip surgery: a multifactorial study. Acta Orthop Scand 1980; 51(6): 943-8.

Støen R Ø, Lofthus C M, Nordsletten L, Madsen J E, Frihagen F. Randomized trial of hemiarthroplasty versus internal fixation for femoral neck fractures: no differences at 6 years. Clin Orthop Relat Res 2014; 472(1): 360-7.

Tsukanaka M, Støen R Ø, Figved W, Frihagen F, Nordsletten L, Röhrl S M. Steady state acetabular cartilage wear after bipolar hemiarthroplasty: a case series of 10 patients with radiostereometric analysis. Hip Int 2017; 27(2): 193-7.

Wada M, Imura S, Baba H. Use of osteonics UHR hemiarthroplasty for frac­ tures of the femoral neck. Clin Orthop Relat Res 1997; (338): 172-81.

Wetherell R G, Hinves B L. The Hastings bipolar hemiarthroplasty for sub­ capital fractures of the femoral neck: a 10-year prospective study. J Bone Joint Surg Br 1990; 72(5): 788-93.

Acta Orthopaedica 2018; 89 (5): 509–514

509

Hemiarthroplasty and total hip arthroplasty in 30,830 patients with hip fractures: data from the Dutch Arthroplasty Register on revision and risk factors for revision Sophie MOERMAN 1, Nina M C MATHIJSSEN 1, Wim E TUINEBREIJER 2, Anne J H VOCHTELOO 3, and Rob G H H NELISSEN 4 1 Department of Orthopedic Surgery, Reinier de Graaf Gasthuis, Delft; 2 Department of Surgery/Traumatology, Erasmus MC, Rotterdam; 3 Centre for Orthopedic Surgery OCON, Hengelo; 4 Department of Orthopedic Surgery, Leiden University Medical Center, Leiden, the Netherlands

Correspondence: sophie.moerman@rdgg.nl Submitted 2018-03-19. Accepted 2018-06-11.

Background and purpose — In the Netherlands about 40% of hip fractures are treated with a hemiarthroplasty (HA) or a total hip arthroplasty (THA). Although these procedures are claimed to have fewer complications than osteosynthesis (i.e., reoperation), complications still occur. Analyses of data from national registries with adequate completeness of revision surgery are important to establish guidelines to diminish the risk for revision. We identified risk factors for revision.

Patients and methods — All patients older than 50 years of age with a hip fracture treated with arthroplasty by orthopedic surgeons and registered in the (national) Dutch arthroplasty register (LROI) were included in the study. In this register, patient characteristics and surgical details were prospectively collected. Revision surgery and reasons for revision were evaluated. A proportional hazard ratio model for revision was created using competing risk analysis (with death as competing risk).

Results — 1-year revision rate of HA was (cumulative incidence function [CIF] 1.6% (95% CI 1.4–1.8) and THA 2.4% (CI 2.0–2.7). Dislocation was the most common reason for revision in both groups (HA 29%, THA 41%). Male sex, age under 80 years, posterolateral approach, and uncemented stem fixation were risk factors for revision in both THA and HA. THA patients with ASA classification III/IV were revised more often, whereas revision in the HA cohort was performed more often in ASA I/II patients.

Interpretation — After arthroplasty of hip fractures, both a posterolateral approach and an uncemented hip stem have higher risks for revision surgery compared with an anterolateral approach and an cemented stem.

Arthroplasty surgery for acute hip fractures is performed in large numbers worldwide. In the Netherlands about 21,000 hip fractures occur annually (NVT and NOV 2016). In about 40% of these cases a hemi- (HA) or total hip arthroplasty (THA) is used (Opendisdata.nl 2017). Although these latter procedures are claimed to have fewer complications than osteosynthesis of the fractured hip, complications still occur (Gao et al. 2012). Analysis of observational data from national registries will more readily give data that can be of clinical value, but such studies are rare (Gillam et al. 2010, Leonardsson et al. 2012b, Gjertsen et al. 2014). A meta-analysis demonstrated a lower risk of reoperation and better function after THA compared with HA (Hopley et al. 2010); a more recent review found comparable outcomes between (bipolar) HA and THA (Wang et al. 2015). None of these studies used national registry data. Also, other issues like the use of a cemented or an uncemented stem, an unipolar or a bipolar HA, and what surgical approach is best to use still remain open (Leonardsson et al. 2012a, Gjertsen et al. 2014, Rogmark and Leonardsson 2016). Therefore, we performed an analysis into failure mechanisms (i.e., end-point revision surgery and reasons for revision) of hemiarthroplasties and total hip arthroplasty using data from the national Dutch Arthroplasty Register (LROI).

Patients and methods All acute hip fractures treated with a HA or a THA by orthopedic surgeons that were registered in the LROI between 2007 and 2017 were included in the study. Patient characteristics (sex, age at procedure, ASA classification, smoking, and BMI) and surgical details (approach, type of fixation, and type of

© 2018 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/4.0) DOI 10.1080/17453674.2018.1499069

510

Acta Orthopaedica 2018; 89 (5): 509–514

Table 1. Baseline characteristics and surgical details of patients with a hip fracture implant) are prospectively registered (van treated with a total hip arthroplasty (THA) or a hemiarthroplasty (HA) Steenbergen et al. 2015). All records in the LROI are linked by the encrypted citizen ser THA HA vice number unique to each Dutch inhabitant.

Factor n = 8,155 n = 22,675 Missing All revision operations during which components are replaced as well as reasons for reviSex, female 70% (5,672/8,141) 70% (15,938/22,644) 45 Age, mean (SD) 71 (9.2) 83 (7.7) b 12 sion are also registered in the database. The ASA, I/II 74% (5,710/7,743) 40% (8,855/22,001) b 1,085 citizen number allows these revisions to be Smoking a 17% (526/3,170) 8% (729/8,764) b 18,896 linked to the primary procedure. Reason(s) BMI, mean (SD) a 25 (7.3) 24 (9.4) b 17,062 Posterolateral approach 60% (4,790/8,046) 53% (11,860/22,462) b 322 for revision surgery are coded in the database Uncemented stem fixation 57% (4,584/8,036) 34% (7,578/22,442) b 352 with a multiple response variable set: disUnipolar HA 79% (17,123/21,685) 990 location, peri-prosthetic fracture, infection, a Smoking and BMI have been registered in the LROI database since 2014. loosening femoral component, loosening b P < 0.001. acetabular component, cup/liner wear, and other reasons.

For this study we included all registered patients older than Furthermore, CIFs for revision were made for each covari50 years of age, treated with a THA or HA for an acute hip able separated for HA and THA. Covariables used were sex, fracture. The LROI has a completeness for primary THA age (< 80 years vs. ≥ 80 years) (80 years was chosen since (independent of indication for THA) of 98%, and 88% for mean age was 80 years, range 50–107 years), ASA classificarevision arthroplasty (van Steenbergen et al. 2015). The com- tion (I/II vs. III/IV), smoking status (yes/no), normal weight pleteness of primary HA increased from 70% in 2013 to 88% (BMI 18.5–25) was compared with overweight (BMI 25–30), in 2015 (Landelijke Registratie Orthopedische Implantaten type of approach (posterolateral [53%] or not posterolateral 2013, 2015). In the Netherlands, HA for hip fracture is per- (anterolateral [12%], straight lateral [33%], and anterior formed by both orthopedic and trauma surgeons, THA for [2%]), and type of stem fixation (cemented vs. uncemented). acute fractures is performed only by orthopedic surgeons. As A hybrid THA was classified according to whether the stem the registration in LROI by trauma surgeons only started in was cemented or not, in order to be able to compare with 2014 and completeness is low, patients treated by trauma sur- HA. Finally, HA type of head (unipolar vs. bipolar head) was geons are not included in the current study. added to the analysis.

The Cox model in a multivariable approach with more Statistics covariables produces hazard ratios (HR) with 95% confidence Baseline characteristics for THA and HA are compared with intervals (CI). The estimated coefficients of the variables were Student’s t-test for continuous variables and the chi-square test tested if they were constant with time and if time interactions for categorical variables. We considered differences between were statistically significant. The variables were entered as groups to be statistically significant if the p-values were less time-varying covariables in the model when the proportional than 0.05. hazards assumption was violated. Separate proportional The high risk of mortality after arthroplasty surgery is an hazard models with hazard ratios (HR) are presented for HA important competing risk for revision operations. Due to and THA. the effect of the competing risk (in this case death) there is a chance of potential under- or overestimation of incidence Ethics, funding, and potential conflicts of interest of reoperations using a Kaplan–Meier analysis (Gillam et al. Ethical approval was not required for this study. The Depart2010, Keurentjes et al. 2012, van der Pas et al. 2017). If, for ment of Orthopaedic Surgery and the Orthopaedic Research example, an uncemented prostheses in this study was applied Foundation in Reinier de Graaf Hospital receive grants from to a healthier population with a lower incidence of death, the Zimmer Biomet. The company (Zimmer Biomet) was not probability of revision would be higher for that group. For this involved in this study. No conflicts of interest to declare. reason competing risk analysis was performed with STATA 11.2 (StataCorp, College Station, TX, USA) using the Cox model (Ranstam and Robertsson 2017). The estimated cumulative incidence functions (CIF) for revision are presented in graphs for both THA and HA. These CIFs were compared Results using the Pepe and Mori test for equality of CIF across groups 30,830 acute hip fractures treated with a HA or a THA were (Pepe and Mori 1993). Revision was defined as the exchange, registered in the LROI database between 2007 and 2017. In addition, or removal of one or more components as registered 22,675 fractures a HA was performed and in 8,155 a THA. in the LROI. Implant revision rate was calculated at 1 and 5 79% received a unipolar HA, 20% a bipolar HA, and 1% a monoblock HA (Table 1). years postoperatively.

Acta Orthopaedica 2018; 89 (5): 509–514

511

Table 3. Reasons for revision after hemiarthroplasty (HA) or total hip arthroplasty (THA) for hip fractures

Factor

HA THA n = 501 n = 298

Single reason for revision, n Dislocation, n (%) Peri-prosthetic fracture, n (%) Infection, n (%) Loosening of femoral component, n (%) Loosening of acetabular component or cup/liner wear, n (%) Other reasons, n (%) Multiple of above-mentioned reasons, n

435 128 (29) 58 (13) 68 (16) 15 (3)

228 94 (41) 28 (12) 26 (11) 25 (11)

n/a 166 (38) 66

18 (8) 37 (16) 70

Figure 1. Cumulative incidence function (CIF) of revision estimates from competing risks data (1 – survival) for patients treated with HA and THA (n = 30,830).

Revision rate 1-year revision rate in HA was (CIF [95% CI]) 1.6% (1.4–1.8) and 5-year 2.5% (2.3– 2.8). 1-year revision rate in THA was 2.4% (2.0–2.7) and 5-year 4.3% (3.8–4.8) (Figure 1, Table 2, see Supplementary data). Revision rate was higher in THA (p < 0.001).

Reasons for revision In 435 HA patients 1 reason for revision was given, in 66 patients multiple reasons were given (153 reasons in 66 patients). Dislocation, periprosthetic fracture, and infection were the most common reasons for revision.

In 228 THA patients 1 reason for revision was given, in 70 patients multiple reasons (156 reasons in 70 patients). Dislocation was the most common reason for revision (41%) (Table 3).

Risk factors for revision Male sex, age below 80 years, ASA classification I/II, a posterolateral approach, and uncemented fixation were risk factors for revision in HA in an univariable analysis risk (Table 4, Figure 2, see Supplementary data). A proportional hazard ratio model using all significant factors showed that male sex, age below 80 years, ASA I/II, a posterolateral approach, and uncemented fixation are risk factors for revision (Table 5). Age and ASA classification were time-varying covariables, meaning that the influence of these variables changes over time. For example, age is no risk factor for revision in the first year after the fracture but becomes one in the years thereafter.

Table 4. Factors associated with revision in hip fracture patients after hemiarthroplasty (HA) and total hip arthroplasty (THA) in a univariable analysis with a hazard analysis

Factor Sex, female (vs. male) Age, ≥ 80 (vs. < 80 years) ASA, III–IV (vs. I–II) Smoking, yes (vs. no) Weight, obesity (vs. normal BMI) Approach, non-posterolateral (vs. posterolateral) Stem fixation, cemented (vs. uncemented) Type of HA, bipolar (vs. unipolar)

HR

HA

95% CI

HR

THA

95% CI

0.78 b 0.65–0.94 0.55 b 0.46–0.65 0.84 0.70–1.01 1.40 0.90–2.18 0.90 0.67–1.22

0.61 b 0.48–0.77 0.44 b 0.29–0.67 1.37 a 1.06–1.76 1.70 a 1.02–2.83 1.37 0.86–2.17

0.67 b 0.56–0.80

0.68 a 0.54–0.88

0.61 b 0.51–0.73 0.73 a 0.57–0.93 0.91 0.73–1.14

HR = hazard ratio. a P < 0.05 b P < 0.001

Table 5. Factors associated with revision in hip fracture treated with a total hip arthroplasty (THA) or a hemiarthroplasty (HA) in a multivariable approach with hazards model with time-varying covariables

Factor Approach a, non-posterolateral (vs. posterolateral) Stem fixation a, cemented (vs. uncemented) ASA b, III–IV (vs. I–II) Age c, ≥ 80 (vs. < 80 years) Sex c, female (vs. male)

HR

HA

95% CI

HR

THA

95% CI

0.67 0.55–0.81

0.70 0.55–0.90

0.63 0.72 d 0.59 d 0.80

0.71 0.55–0.91 1.46 1.13–1.90 0.52 d 0.55–0.91 0.65 0.51–0.83

0.52–0.75 0.62–0.83 0.50–0.70 0.66–0.97

HR= hazard ratio. a Variables with direct effect on outcome. b Measured confounder with direct effect on choice of HA or THA. c Measured confounders with effect on ASA. d Time-varying covariables, Confounder with direct effect on revision: HA/THA choice (not accounted for by stratification).

512

Male sex, age below 80 years, smoking, a posterolateral approach, and uncemented stem fixation were risk factors for revision in THA in an univariable analysis. ASA classification was not a clear risk factor (p = 0.09) (Figure 2, Table 4). A proportional hazard ratio model showed that male sex, younger age, ASA III/ IV, a posterolateral approach, and an uncemented stem were associated with more revisions (Table 5). Age was a time-varying covariable meaning that the hazard of age on revision changes over time.

Specific reason for revision in factors associated with revision In both THA and HA a fracture as a reason for revision was more common in an uncemented prosthesis (HA 28% vs. 2%, THA 15% vs. 6%) (Table 6, see Supplementary data).

In HA, dislocation as a reason for revision was more common in younger patients (35% vs. 24%), ASA III/IV patients (35% vs. 24%), and a posterolateral approach (37% vs. 19%). A fracture was more common in older HA patients (18% vs. 9%).

Infection was more common amongst male patients (23% vs.

12%) and a cemented prosthesis (21% vs. 9%).

In THA dislocation as a reason for revision was more common in a cemented prosthesis (51% vs. 36%). A fracture as a reason for revision was more common in the male sex (THA 18% vs. 8%).

Discussion Revision rate of THA was higher compared with the revision rate of HA. The 5-year revision rate of an HA was 2.5% and 4.3% in THA, which is in contrast to the results from randomized trials, which showed no difference between HA and THA (van den Bekerom et al. 2010, Hedbeck et al. 2011) However, patients included in these randomized trials were less frail than the average hip fracture patients. The HA group in our registry study contained patients with more frailty (higher age, higher ASA classification) than the THA group, therefore the threshold for a surgeon to decide to revise was probably higher in the HA group.

In our study, dislocation was the most common reason for revision in both HA (29%) and THA (41%). Acetabular erosion (prevalence is 2–41%) is a theoretical indication to perform a revision in a painful HA (Baker et al. 2006). In the LROI, acetabular erosion as reason for revision cannot be registered. Patients who were revised for acetabular erosion were classified in the “other” category (38%). How many patients in this category had acetabular erosion is unclear.

Male sex and age below 80 years were risk factors for revision surgery in THA and HA. This in accordance with data from the Norwegian and British register (Stafford et al. 2012, Rogmark et al. 2014). Younger patients are likely to be more demanding regarding hip function after surgery, thus even revision for moderate postoperative complaints is more likely.

Acta Orthopaedica 2018; 89 (5): 509–514

Males have a higher occurrence of periprosthetic fractures, which may lead to a higher revision rate (Table 6) (Australian Orthopaedic Association National Joint Replacement Registry 2016).

In HA, ASA classification I/II was a risk factor for revision; however, in THA ASA classification III/IV was a risk factor for revision. This contradiction is probably explained by the selection bias of THA and HA. We believe THA patients with an ASA classification of III/IV are less frail than HA with an ASA classification of III/IV, while a surgeon will choose an HA in the frailest patients (i.e., shorter surgical time and less blood loss; Blomfeldt et al. 2007). These frail HA patients (ASA classification III/IV) are unlikely to undergo revision due to higher risks but also to lower demand on functionality of these patients. In THA these ASA classification III/IV patients have a higher risk of revision compared with ASA classification I/II. Comorbidities like diabetes mellitus might cause this higher rate of infection (Dale et al. 2011). A British and Norwegian register study has shown the same tendency for higher revision in higher category ASA patients in THA for hip fracture (Dale et al. 2011, Stafford et al. 2012).

A posterolateral approach was a risk factor for revision in both HA and THA. 2 large register studies showed that the posterolateral approach led to more dislocations (Leonardsson et al. 2012b, Rogmark et al. 2014). However, patient reported outcome measurements (PROMs) used in the registry study in Norway showed that the posterior approach gave less pain, fewer walking problems, and better QoL than the lateral approach (Kristensen et al. 2016). Using a dual mobility cup may reduce dislocation risk when using a posterolateral approach (Batailler et al. 2017, De Martino et al. 2017, TaboriJensen et al. 2018).

Uncemented stems were a risk factor for revision in both HA and THA. Peri-prosthetic fractures are more common in uncemented prostheses (both HA and THA), probably as a result of trying to create a press-fit situation in the weaker (osteoporotic) bone (Moerman et al. 2017). This increased risk of periprosthetic fracture in uncemented prostheses must be weighed against the potential complications of cementing such as bone cement implantation syndrome (BCIS) (Donaldson et al. 2009).

Bipolar prostheses are developed to reduce the risk of erosion of the acetabulum. We did not find any difference in revision hazards between unipolar and bipolar heads. 79% of the Dutch hip fracture patients treated with HA receive a unipolar head. Costs for bipolar heads in the Netherlands are about double the costs of unipolar heads. The Swedish register showed more reoperations with bipolar heads (Leonardsson et al. 2012b) and the Australian register found lower reoperation rates with bipolar heads (Gillam et al. 2010). Reasons for these conflicting data may be the difference in hemiarthroplasty populations in Australia, Sweden and the Netherlands.

The NICE guideline (NICE 2011) for hip fractures advises against use of monoblock prostheses. In our register only 164

Acta Orthopaedica 2018; 89 (5): 509–514

(0.8%) of all HA were monoblock prostheses. Therefore no analysis on these monoblock prostheses was performed.

Strengths and limitations This is the first nationwide Dutch study on HA and THA in acute hip fractures using data from the Dutch Arthroplasty Register (LROI). Previously the Scandinavian, British, and Australian registers have published their results (Gillam et al. 2010, Leonardsson et al. 2012b, Stafford et al. 2012, Gjertsen et al. 2014). The added value of these Dutch results is important, since each country has its own specific healthcare organization. As for the Netherlands, a quality mark for hip fractures was that surgery has to be performed within 24 hours of admittance which may cause differences in outcome between registers. Furthermore, this study includes both HA and THA data for acute hip fractures. Observational data studies for THA in hip fractures are sparse, thus knowledge on this subject has to be extended, since the proportion of hip fracture patients treated with THA is increasing. The proportion hazards model clearly assigns risk factors for revision, which is of clinical importance and may guide treatment of these often frail patients in order to minimize the perioperative risks.

A limitation of the study is the incomplete registration of HA for acute hip fractures (but still 88% completeness). Follow-up of hip fracture patients is limited because of the high mortality rate (1-year mortality is around 20%). There are a limited number of patient characteristics registered in our national registry. Alcohol use, for instance, was not registered although it influences revision rate (Johnston and Parker 2014, Kosola et al. 2017). Because of this limited number of patient characteristics, there is potential for residual confounding.

Furthermore, only revision operations in which components are replaced are registered in the database. Reoperations without component (re-)placement (like debridement of the wound and the prosthesis without liner exchange in the case of acute infection) are not registered in the LROI database.

In summary, revision rates in both HA and THA after an acute hip fracture are considerable. Avoidance of both an uncemented stem and a posterolateral approach may reduce the revision rate.

Supplementary data Tables 2 and 6, and Figure 2 are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674.2018.1499069

Design of study: SM, NM, RN, AV. SM and WT did the statistical analysis.

SM, NM, RN, and AV wrote the manuscript. All authors read and approved the final manuscript.

Acta thanks Torben B Hansen and other anonymous reviewers for help with peer review of this study.

513

Australian Orthopaedic Association National Joint Replacement Registry.

Annual Report 2016. https://aoanjrr.sahmri.com/annual-reports-2016.

Baker R P, Squires B, Gargan M F, Bannister G C. Total hip arthroplasty and hemiarthroplasty in mobile, independent patients with a displaced intracapsular fracture of the femoral neck (a randomized, controlled trial). J Bone Joint Surg 2006; 88(12): 2583-9.

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.

Blomfeldt R, Törnkvist H, Eriksson K, Söderqvist a, Ponzer S, Tidermark J. A randomised controlled trial comparing bipolar hemiarthroplasty with total hip replacement for displaced intracapsular fractures of the femoral neck in elderly patients. J Bone Joint Surg Br 2007; 89(2): 160-5.

Dale H, Skramm I, Lower H L, Eriksen H M, Espehaug B, Furnes O, Skjeldestad F E, Havelin L I, Engesaeter L B. Infection after primary hip arthroplasty. Acta Orthop 2011; 82(6): 646-54.

De Martino I, D’Apolito R, Soranoglou V G, Poultsides L A, Sculco P K, Sculco T P. Dislocation following total hip arthroplasty using dual mobility acetabular components: a systematic review. Bone Joint J 2017; 99B(1): 18-24.

Donaldson A J, Tomson H E, Harper N J, Kenny N W. Bone cement implantation syndrome (BCIS). Br J Anaesth 2009; 102(1): 12-22.

Dutch Arthroplasty Register (LROI). ROI.nl, site accessed on 03-12-2017.

Gao H, Liu Z, Xing D, Gong M. Which is the best alternative for displaced femoral neck fractures in the elderly?: a meta-analysis. Clin Orthop Relat Res 2012; 470(6): 1782-91.

Gillam M H, Ryan P, Graves S E, Miller L N, de Steiger R N, Salter A. Competing risks survival analysis applied to data from the Australian Orthopaedic Association National Joint Replacement Registry. Acta Orthop 2010; 81(5): 548-55.

Gjertsen J-E, Fenstad A M, Leonardsson O, Engesæter L B, Kärrholm J, Furnes O, Garellick G, Rogmark C. Hemiarthroplasties after hip fractures in Norway and Sweden: a collaboration between the Norwegian and Swedish national registries. Hip Int 2014; 24(3): 223-30.

Hedbeck C J, Enocson A, Lapidus G, Blomfeldt R, Tornkvist H, Ponzer S, Tidermark J. Comparison of bipolar hemiarthroplasty with total hip arthroplasty for displaced femoral neck fractures: a concise four-year follow-up of a randomized trial. J Bone Joint Surg Am 2011; 93(5): 445-50.

Hopley C, Stengel D, Ekkernkamp A, Wich M. Primary total hip arthroplasty versus hemiarthroplasty for displaced intracapsular hip fractures in older patients: systematic review. BMJ 2010; 340(7761): 1397.

Johnston L E, Parker M J. Hip fractures and chronic alcohol excess: a series of 7,023 cases. Hip Int 2014; 24(6): 644-9.

Keurentjes J C, Fiocco M, Schreurs B W, Pijls B G, Nouta K A, Nelissen R G H H. Revision surgery is overestimated in hip replacement. Bone Joint Res 2012; 1(10): 258-62.

Kosola J, Kaipia A, Laitinen M K, Nieminen J. Complications after surgical treatment of femoral neck fractures in men with alcohol dependence syndrome: retrospective register analysis of 154 cases. Arch Orthop Trauma Surg 2017; 137(7): 967-73.

Kristensen T B, Vinje T, Havelin L I, Engesaeter L B, Gjertsen J-E. Posterior approach compared to direct lateral approach resulted in better patientreported outcome after hemiarthroplasty for femoral neck fracture: 20,908 patients from the Norwegian Hip Fracture Register. Acta Orthop 2016; 87(1): 29-34.

Landelijke Registratie Orthopedische Implantaten. LROI-rapportage 2013; 2013.

Landelijke Registratie Orthopedische Implantaten. LROI-rapportage 2015; 2015.

Leonardsson O, Garellick G, Kärrholm J, Akesson K, Rogmark C. Changes in implant choice and surgical technique for hemiarthroplasty: 21,346 procedures from the Swedish Hip Arthroplasty Register 2005-2009. Acta Orthop 2012a; 83(1): 7-13.

Leonardsson O, Kärrholm J, Åkesson K, Garellick G, Rogmark C. Higher risk of reoperation for bipolar and uncemented hemiarthroplasty. Acta Orthop 2012b; 83(5): 459-66.

514

Moerman S, Mathijssen N M C, Niesten D D, Riedijk R, Rijnberg W J, Koëter S, Kremers van de Hei K, Tuinebreier W E, Molenaar T L, Nelissen R G H H, Vochteloo A J H. More complications in uncemented compared to cemented hemiarthroplasty for displaced femoral neck fractures: a randomized controlled trial of 201 patients, with one year follow-up.

BMC Musculoskelet Disord 2017; 18(1): 169.

NICE. Hip fracture: the management of hip fracture in adults. NICE clinical guideline 124; 2011.

NVT, NOV. Proximale femur fracturen (richtlijn). 2016.

Opendisdata.nl. DIS open data [Internet]. [cited 2017 Nov 23].

Pepe M S, Mori M. Kaplan-Meier, marginal or conditional probability curves in summarizing competing risks failure time data? Stat Med 1993;12(8):737-51.

Ranstam J, Robertsson O. The Cox model is better than the Fine and Gray model when estimating relative revision risks from arthroplasty register data. Acta Orthop 2017; 88(6): 578-80.

Rogmark C, Fenstad A M, Leonardsson O, Engesæter L B, Kärrholm J, Furnes O, Garellick G, Gjertsen J-E. Posterior approach and uncemented stems increases the risk of reoperation after hemiarthroplasties in elderly hip fracture patients. Acta Orthop 2014; 85(1): 18-25.

Rogmark C, Leonardsson O. Hip arthroplasty for the treatment of displaced fractures of the femoral neck in elderly patients. Bone Joint J 2016; 98-B(3): 291-7.

Acta Orthopaedica 2018; 89 (5): 509–514

Stafford G H, Charman S C, Borroff M J, Newell C, Tucker J K. Total hip replacement for the treatment of acute femoral neck fractures: results from the National Joint Registry of England and Wales at 3-5 years after surgery. Ann R Coll Surg Engl 2012; 94(3): 193-8.

Tabori-Jensen S, Hansen T B, Bøvling S, Aalund P, Homilius M, Stilling M. Good function and high patient satisfaction at mean 2.8 years after dual mobility THA following femoral neck fracture: a cross-sectional study of 124 patients. Clin Interv Aging 2018; 13: 615-21. van den Bekerom M P J, Hilverdink E F, Sierevelt I N, Reuling E M B P, Schnater J M, Bonke H, Goslings J C, van Dijk C N, Raaymakers E L F B. A comparison of hemiarthroplasty with total hip replacement for displaced intracapsular fracture of the femoral neck: a randomised controlled multicentre trial in patients aged 70 years and over. J Bone Joint Surg Br 2010; 92(10): 1422-8. van der Pas S L, Nelissen R G H H, Fiocco M. Patients with staged bilateral total joint arthroplasty in registries. J Bone Joint Surg 2017; 82: 1-8. van Steenbergen L N, Denissen G A W, Spooren A, van Rooden S M, van Oosterhout F J, Morrenhof J W, Nelissen R G H H. More than 95% completeness of reported procedures in the population-based Dutch Arthroplasty Register. Acta Orthop 2015; 86(4): 498-505 Wang F, Zhang H, Zhang Z, Ma C, Feng X. Comparison of bipolar hemiarthroplasty and total hip arthroplasty for displaced femoral neck fractures in the healthy elderly: a meta-analysis. BMC Musculoskelet Disord 2015; 16: 229.

Acta Orthopaedica 2018; 89 (5): 515–521

515

Time-driven activity-based cost of outpatient total hip and knee arthroplasty in different set-ups Henrik HUSTED 1, Billy B KRISTENSEN 2, Signe E ANDREASEN 1, Christian Skovgaard NIELSEN 1, Anders TROELSEN 1, and Kirill GROMOV 1 1 Department

of Orthopedic Surgery, Copenhagen University Hospital Hvidovre, Copenhagen; 2 Ambulatory Surgery Department, Copenhagen University Hospital Hvidovre, Copenhagen, Denmark Correspondence: henrikhusted@dadlnet.dk Submitted 2018-04-30. Accepted 2018-06-21.

Background and purpose — Length of stay (LOS) following total hip and knee arthroplasty (THA and TKA) has been reduced over the years due to fast-track. Short stays of 2 days in fast-track departments in Denmark have resulted in low total costs of around US$ 2,550. Outpatient THA and TKA is gaining popularity, albeit in a limited and selected group of patients; however, the financial benefit of outpatient arthroplasty remains unknown. We present baseline detailed economic calculations of outpatient THA and TKA in 2 different settings: one from the hospital and another from the ambulatory surgery department.

Patients and methods — Data from 6 patients (1 TKA, 1 uncemented THA, 1 cemented THA in each department) were collected prospectively using the Time Driven Activity Based Costing method (TDABC). Time consumed by different staff members involved in patient treatment in the perioperative period of outpatient THA and TKA was calculated in 2 different settings: one in the orthopedic department and one in the ambulatory surgery department.

Results — LOS was around 11 h in the orthopedic department and around 7 h in the ambulatory surgery department, respectively. TDABC revealed minor differences in the operative settings between departments and similar expenses occurred during the short stay of US$ 777 and US$ 746, respectively. Adding the preoperative preparation and postoperative follow-up resulted in total cost of US$ 951 and US$ 942 for the ward and the ambulatory surgery department, respectively.

Interpretation — Outpatient THA and TKA in hospital and ambulatory surgery departments results in similar cost using the TDABC method. Compared with the cost associated with 2-day stays, outpatient procedures are around twothirds cheaper provided no increase occurs in complications or readmissions.

Length of stay (LOS) in hospital following total hip and knee arthroplasty (THA and TKA) has been reduced over the years. In particular, fast-track THA and TKA combining evidence-based clinical features with organizational optimization has resulted in short LOS of 1–3 days for the majority of unselected patients as a positive spin-off from the main goal of reducing perioperative morbidity and mortality (Husted et al. 2010, 2016, Malviya et al. 2011, Husted 2012, Jørgensen et al. 2013a and b, Khan et al. 2014).

In the last decade, reports of patients operated with THA and TKA being discharged even faster have been published (Berger et al. 2009a, 2009b). However, concern has been expressed regarding the cost-benefit of this approach as the additional services provided by the additional personnel may outweigh the savings of a shorter stay (Berger et al. 2009b).

Since then, several reports have shown the feasibility of outpatient arthroplasty in selected patients (Hartog et al. 2015, Goyal et al. 2017, Gromov et al. 2017, Klein et al. 2017, Larsen et al. 2017, Meneghini et al. 2017).

Conventional inpatient stays have recently been reported to cost around US$ 30,000 in the US (Nichols and Vose 2016) and with expensive reimbursement at around US$ 26,000 on average in a Medicare population (Mechanic 2015). Fast-track THA and TKA with a 2-day stay was calculated at around US$ 2,500 in Denmark for comparison using the Time Driven Activity Based Costing method (TDABC), which represents an economical method using only staff-based costs of the procedures in the care-cycle (Andreasen et al. 2017).

A few studies have used various methods to determine the potential economic benefit of outpatient arthroplasty (Aynardi et al. 2014, Lovald et al. 2014, Huang et al. 2017) but their different set-ups including non-itemized billing, overhead and indirect costs, and 2 years’ compiled expense hinder comparison between hospitals as different reimbursement systems are used.

© 2018 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/4.0) DOI 10.1080/17453674.2018.1496309

516

Acta Orthopaedica 2018; 89 (5): 515â&#x20AC;&#x201C;521

Preop. outpatient visit #1: Referred to hospital: consultant + radiographs = agreement on surgery 20 min Preop. outpatient visit #2: Junior physician (paperwork, documentation) 30 min

Preop. outpatient visit #1: Referred to hospital: consultant + radiographs = agreement on surgery 20 min Blood samples 10 min Anesthetist clearance 10 min

Preop outpatient visit #3: PATIENT SEMINAR Surgeon: 30 min Physiotherapist: 15 min Nurse assistant: 90 min Anesthesia nurse: 15 min DAY OF SURGERY See Table 3 for activities and time used FOLLOW-UP Nurse outpatient visit 3 weeks postop.: 30 min Surgeon outpatient follow-up 3 months postop.: 10 min

Preop. outpatient visit #2: Junior physician (paperwork, documentation) 30 min

Blood samples 10 min

Anesthetist clearance information/preparation for outpatient surgery 20 min

Physiotherapist information/preparation exercises/crutches 15 min

DAY OF SURGERY See Table 3 for activities and time used FOLLOW-UP Nurse outpatient visit 3 weeks postop.: 30 min Surgeon outpatient follow-up 3 months postop.: 10 min

Flowchart of various procedures before, during, and after outpatient TJA in a central operation ward (left) and an ambulatory surgery department (right)

As several editorials have been addressing the need for detailed economic evaluation of the outpatient procedure itself (Argenson et al. 2016, Vehmeijer et al. 2018), and reimbursement systems vary between hospitals, we found it of interest to use the TDABC method to calculate the cost for the outpatient procedure in different outpatient set-ups allowing for comparison between hospitals and countries, regardless of reimbursement system.

Thus, we present baseline detailed economical calculations using TDABC of outpatient THA and TKA in a hospital and an ambulatory surgery department.

Methods Economic considerations Time-driven activity-based costing (TDABC) (Kaplan and Anderson 2004) is a method to calculate cost involving only estimates of 2 parameters: the time needed from staff members to perform a process or activity and the cost per minute of that staff member. Combining the different processes and the staff members involved, TDABC will reflect time and cost spent on operational processes in detail. The TDABC takes into account the amount of time spent by various staff members on patientrelated activities only, and the overhead value of staff-time spend not directly with patients or patient-related work (holidays, illness, breaks, research, and education) is neglected.

Prerequisites for the estimation of TDABC are the creation of process maps in the care cycle, a calculation of time spent

on the various specific processes and knowing the salaries of the staff members involved.

Adding other expenses in the care cycle for THA and TKA related to consumables, surgery (exclusive of implants), utensils, medicine, tests, and cleaning gives the total cost for the specific procedure. However, these expenses are not calculated or included here as they are not part of the TDABC method and may vary considerably between institutions. Also, fixed cost like buildings, equipment in surgery and rooms, heating and lighting, and administration will not be taken into account. By looking only at the direct costs, the comparison of cost incurred by different departments will be possible across potential differences in organizational set -up.

The care cycle of a total joint arthroplasty (TKA) (THA or TKA) is described in 2 different settings: 1 in the orthopedic department and 1 in the ambulatory surgery department (Figure). All procedures involving staff members were evaluated and timed: Time related to surgical procedures and anesthesia was collected from the surgical database, which gives exact time spent including detailed information on every procedure in the operating room. Other procedures were noted by both the staff members and by an independent observer to ensure completeness. Preoperative procedures differed slightly whereas the follow-up procedures were identical. A full care cycle is defined from the first preoperative visit to the final outpatient follow-up. All patients were seen by a nurse at an outpatient clinic at 2 (THA) or 3 (TKA) weeks postoperatively for staple removal. All patients were seen by the surgeon at 3 months postoperatively. Time spent on these procedures

Acta Orthopaedica 2018; 89 (5): 515–521

as for the preoperative procedures is an estimate based on the time slots and mean time spent on each patient in group Staff members CPM settings.

(Hvidovre) US$ To allow for the same case mix, 1 operation with a hybrid Orthopedic surgeon 2.00 Anesthesiologist 1.33 THA (cemented femur), 1 Orthopedic resident 1.20 operation with an uncemented Nurse (ward) 0.85 THA, and 1 operation with a Nurse (anesthesia) 0.94 Nurse (scrub, OR) 0.84 cemented tricompartmental Nurse assistant 0.75 TKA were timed in each setPhysiotherapist 0.88 ting and a “typical” course Cleaning 0.75 Radiologist 2.00 was ensured (no abnormal Secretary 0.75 procedures regarding logisLaboratory technician 0.85 tics or clinical procedures Porter 0.64 (anesthesia, surgery). Patients were not informed about nor consented to timing the various procedures as timing did not influence any procedure or content thereof and the 6 patients were chosen for type of surgery on days of timing. No randomization of patients or location of surgery was performed and patients were already scheduled for surgery on the 2 different lists. There was no choosing of a specific patient for the specific location by any factor except vacancy. Accurate times were registered and mean time for all operative procedures was calculated as there were slight differences (Table 3).

Information on salaries was obtained from the central hospital database and represents the average salary for each staff type (Table 1). Based on these calculations, TDABC was estimated for an operation in both settings.

Table 1. Cost per minute in US$ for various staff members involved

Treatment procedure All patients scheduled for outpatient THA or TKA were screened for eligibility (Gromov et al. 2017). Eligible patients were included and informed of the intended outpatient procedure. The surgery was performed under spinal analgesia in the OR and under general anesthesia in the ambulatory surgery department. THA was performed using a standard posterolateral approach with simple posterior soft-tissue repair and TKA was performed with a standard medial parapatellar approach without the use of tourniquet; LIA was given at the end of surgery. Postoperative radiographs were obtained in the OR in the ambulatory surgery department (in the radiographic department for the hospital patients), approved by the surgeon and given to the patient. Rivaroxaban was used as oral thromboprophylaxis starting 6 to 8 hours postoperatively and given for 2 days only. Mechanical thromboprophylaxis and extended oral thromboprophylaxis was not used. In Denmark the recommendations allow for oral thromboprophylaxis only during hospital stay if the patient is following a fast-track pathway (Jørgensen et al. 201b).

Patients in hospital bypass the post-anesthesia care unit (PACU) if blood loss is estimated less than 300 mL in order

517

to begin their functional recovery as early as possible—no patient in this study stayed in the PACU. Physiotherapy was started as soon as possible after surgery: after the spinal anesthesia had worn off or when patients were fit to mobilize. Patients were discharged if fulfilling the discharge criteria before 8 pm. The discharge criteria were: self-dependent, sit/stand from chair/toilet, steady gait with crutches, master stairs, stable vital signs, acceptable pain (VAS < 3 at rest and VAS < 5 on mobilization), postsurgical bleeding should be consistent with expected blood loss for the procedure and not require repeated dressing change, hemodynamically stable and showing no clinical signs of anemia (Husted 2012). All patients were discharged to their own homes without any additional assistance.

Pre- and postoperative outpatient visits were the same without differences in duration, except for the preoperative patient seminar, which was exclusively for the hospital patients, whereas the patients to be operated in the ambulatory surgery center had slightly longer preoperative preparation by the anesthetist. Patients operated in the ambulatory surgery department also received preoperative physiotherapy instructions, while physiotherapy was covered during the patient seminar for hospital patients.

Statistics Descriptive statistics is used reporting minutes spent on processes and activities and cost in US$.

Ethics, funding, and potential conflict of interest No approval from the National Ethics Committee was necessary as this was a non-interventional observational study.

Permission to store and review patient data was given by the Danish National Board of Health (j.nr:3-3013-56/1/HKR) and Danish Data Protection Agency (j.nr:20047-58-0015). No competing interests were declared.

Results All 6 patients fulfilled the discharge criteria on the day of surgery and were discharged. Patients on the ward stayed till 6 pm, whereas patients in the ambulatory surgery department stayed till 3 pm. Time used during the entire care cycle, and corresponding cost, is presented in Table 2.

Time used by staff members for the various procedures during the day of surgery is listed in Table 3 as is the TDABC for each procedure per staff member. Time for the different procedures varied a little between operations and between locations, reflecting the small variabilities even in a standard procedure: they accumulated to 701 minutes in total for the hospital patients versus 671 minutes for the ambulatory surgery department patients (Table 3). Mean cost for the 3 operations was US$ 777 on the ward and US$ 746 in the ambulatory surgery department (Table 3). Adding the time spent on preop-

518

Acta Orthopaedica 2018; 89 (5): 515–521

Table 2. Time (minutes) and corresponding TDABC (US$) used on preoperative preparation before operations in ambulatory surgery department (ASD) and central operation ward (COP) and post-discharge follow-up

Perioperative activities

ASD COP time TDABC time TDABC (min) (US$) (min) (US$)

Preop. visit 1: Surgeon 20 40 20 40 Total 20 40 20 40 Preop. visit 2: Resident 30 36 30 36 Nurse 30 26 30 26 Lab technician 10 9 10 9 Anesthetist 20 27 10 13 Physiotherapist 15 13 X 0 Total 105 110 80 83 Preop. visit 3 (patient seminar, n = 30): Surgeon X 30 60 Anesthetist nurse X 15 14 Physiotherapist X 15 13 Nurse assistant X 90 68 Total 0 150 155 Day of surgery: Total 671 746 701 777 Follow-up (nurse): Nurse 30 26 30 26 Total 30 26 30 26 Follow-up (surgeon): Surgeon 10 20 10 20 Total 10 20 10 20 Total 836 942 846 951

erative preparation and postoperative follow-up, a total of 846 versus 836 minutes were accumulated amounting to US$ 951 and US$ 942, respectively (Table 3).

Discussion In this prospective study we provide baseline detailed economic calculations using TDABC of outpatient THA and TKA in a hospital ward and in an ambulatory surgery department.

The time and money spent to do an outpatient THA or TKA varied only slightly between the hospital ward and the ambulatory surgery department (situated separately in the hospital).

Slightly different pathways and procedures were followed resulting in a total of around 14 hours of staff time used per procedure resulting in a difference of only US$ 9 in total. The amounts of US$ 951 and US$ 942 are of course significantly lower compared with the previously calculated US$ 2,550 for a 2-day fast-track stay (Andreasen et al. 2017), which is mostly attributable to the extra nights the patient spent in the hospital and the associated care.

All 6 patients in this study were discharged on the day of surgery. However, feasibility studies have shown discharge on the day of surgery (defined as intended within 12 h) to vary between 30% and 89% (Hartog et al. 2015, Goyal et al. 2017,

Gromov et al. 2017, Larsen et al. 2017). With no study reaching 100%, it is evident that precautions in the form of the possibility for an extended stay overnight are mandatory. The pathophysiological changes associated with the surgical stress response pose barriers to fulfilling functional discharge criteria early (i.e., on the day of surgery) and as these include pain, dizziness (orthostatic intolerance), and muscle weakness, a multimodal and multidisciplinary effort is needed but not successfully accomplished in all patients (Husted et al. 2011).

Logistically, bypassing the PACU or having very short stays there may require a reevaluation of discharge criteria from the PACU to facilitate outpatient arthroplasty (Lunn et al. 2012).

Occasionally, patients intended to be outpatients have to stay in hospital/be admitted to hospital. This is easily facilitated in a hospital ward where the bed is available, at least till next day depending on the booking of beds, and as nurses are present already no extra resources need be allocated. In the ambulatory surgery department (in our hospital), no imminent possibility for an overnight stay is present and hence this requires either staff to stay over or referral to a 24 h-manned ward, both requiring extra resources inflicting costs that are not accounted for in the presented TDABC. Hence, there are some potential hidden costs in performing outpatient arthroplasty not accounted for by the TDABC, which may actually render outpatient arthroplasty in the hospital ward cheaper, of course depending on the proportion of subsequent overnight stays, as the subsequent number of patients needing an inpatient stay may be “absorbed” in the (already existing) ward nurse manning. Also, the price for an “empty bed” for those patients operated in the hospital setting is not accounted for in the TDABC, which thus also presents a potential hidden cost.

Unfortunately, making estimations for such hidden costs in a TDABC is virtually impossible, as this would require many assumptions and potentially inaccurate estimations that would make the final result difficult to interpret and hinder comparison with other hospital and department set-ups.

TDABC analysis does not include overhead expenses.

While these costs can be substantial, including overheads may make comparison with other hospitals even more inaccurate as overhead expenses and their composition may vary greatly between the private and the public sector and even between hospitals within the same country.

Selection criteria for outpatient arthroplasty as well as safety issues have been addressed mainly in large cohorts like the American College of Surgeons–National Surgical Quality Improvement Program (NSQIP), but unfortunately the data in the cohorts are blinded and rely on the hospitals’ definition of the outpatient procedure. As a study has shown this definition to be severely flawed by also including patients staying in hospital for “observation” for 1 or more nights (Bovonratwet et al. 2017)—and still counted as outpatients—the outcomes of the studies using NSQIP data on patient selection and safety should be interpreted with caution. Also, in some studies examining outpatient arthroplasty and looking at safety, the

Acta Orthopaedica 2018; 89 (5): 515–521

519

Table 3. Time (minutes) and corresponding TDABC (US$) used on various procedures in 3 operations (TKA = total knee arthroplasty, THAh = total hip arthroplasty hybrid, THAu = total hip arthroplasty uncemented) in ambulatory surgery department (ASD) and central operation ward (COP) during day of surgery

ASD1 ASD2 ASD3 time TDABC COP1 COP2 COP3 time TDABC TKA THAh THAu mean US$ TKA THAh THAu mean US$

Anesthetist: Anesthesia procedure Total Anesthetist nurse: Anesthesia prep. (+ iv lines) Anesthesia procedure Surgery Postoperative tasks Total Surgeon: Preparation outside OR Preparation inside OR Scrubbing Surgery Postop. radiographs Documentation Rounds Total Surgical assistant: Preparing prostheses, bed Preparing op table/positioning Scrubbing Surgery Postop. radiographs Taking patient to recovery room Total Scrub nurse: Preparation OR Scrubbing Prep. instruments Anesthesia procedure Sterilization, draping Surgery Cleaning instruments Total Floor nurse: Preparation OR Prep. instruments Anesthesia procedure Sterilization, draping Surgery Postop. radiographs Cleaning Total Radiologist: Postop radiographs Total Ward nurse: Prep. of patient Various incl. documentation Total Physiotherapist: Mobilization Total Total staff minutes

13 11 8

13 11 8 11 15

15 13 8

13 11 9

35 57 45

5 5 5

68 86 67 74 70

8 10 10

28 20 22

5 5 5

35 57 45

8 6 7

10 10 11

15 12 14

109 120 114 114 228

15 20 16

15 13 15

5 5 5

35 57 45

8 6 7

10 8 10

88 109 98 98 118

15 15 15

5 5 5

9 10 15

13 11 12

17 16 16

35 57 45

16 10 15

114 124 123 120 101

11 15 15

17 15 20

13 11 13

17 15 15

35 57 45

8 8 8

10 10 15

111 131 131 124 104

X X X

0 0 0 0 0

5 5 5

90 92 92

95 97 97 96 82

30 35 35

30 35 35 33 29 628 713 673 671 746

treatment continues after discharge as the hospital is “moved home with the patient” as both nurses and physiotherapists visit and treat the patient at home (Klein et al. 2017). Hence, valid data are lacking on safety in larger groups of patients and

15 8 16

15 8 16 13 17

35 35 10

15 10 10

43 48 39

5 5 9

98 98 68 88 83

10 10 10

35 25 27

5 5 5

43 48 39

X X X

5 5 5

15 15 15

113 108 101 107 214

20 10 10

23 15 13

5 5 5

43 48 39

X X X

13 15 15

104 93 82 93 112

10 15 15

5 10 5

10 20 7

15 8 16

13 20 10

43 48 39

13 10 10

109 131 102 114 96

10 15 12

20 10 7 15 8 16 13 20 10 43 48 39 X X X 7 5 10 108 106 94 103 87

10 10 10 10 10 10 10 20

22 15 10 102 100 80

124 115 90 110 94

60 60 70 60 60 70 63 55 741 729 633 701 777

also hindering economic cost–benefit evaluation of the true outpatient procedure with same-day discharge to home without nurses or physiotherapists visiting including 90-day complication/readmission rates in larger groups. Hence, safety

520

issues are still pending regarding outpatient versus inpatient arthroplasty and might add to cost, if more complications/ readmissions were to occur (Lovecchio et al. 2016). The current publications, either finding it safe (Courtney et al. 2017, Nelson et al. 2017, Courtney et al. 2018) or associated with more complications (Lovecchio et al. 2016, Arshi et al. 2017) may not answer the question of safety as they are based on register studies with poor or no control of the definition of “outpatient” (Bovonratwet et al. 2017). Also, reviews call for prospective cohort studies to determine outcome, safety, and cost efficiency of outpatient arthroplasty (Pollock et al. 2016, Hoffmann et al. 2018). If extra resources are needed with the sole purpose of allowing discharge on the day of surgery, cost–benefit analyses need to be performed including perioperative safety (Argenson et al. 2016, Thienpont et al. 2015, Vehmeijer et al. 2018).

We have not focused on potential complications, readmissions, or mortality in this study as the aim was to evaluate a care cycle for the standard arthroplasty outpatient. As the aim was to illuminate the staff-associated time and the cost thereof in outpatient arthroplasty—and not to compare the 2 pathways to find the cheapest—only 6 patients were included to allow calculations for the various procedures. No power calculation was performed or needed as description and calculation of the procedures allowing for comparison was intended.

No other studies using the TDABC method have been published for outpatient hip and knee arthroplasty. Other accounting methods exist and TDABC may underestimate the total cost by the inherent weaknesses addressed above and by not taking overheads, heating, linen, food, indirect costs etc. into account. However, the TDABC method is still recommended as it provides a care cycle-based identification of care processes and the associated staff minutes allowing for evaluation, comparison, and optimization (Palsis et al. 2018). Hence, our study provides a preliminary baseline for comparison with longer hospital stays (Akhavan et al. 2016, Andreasen et al.

2017) as well as more conventional tracks (Chen et al. 2015).

When larger prospective cohort studies focusing on morbidity and mortality in outpatients are published, the figures need to be included in more detailed economic analyses allowing for a more accurate cost–benefit analysis.

It can and should be debated whether TDABC is the best method to perform a detailed economic evaluation or if other more detailed analyses are warranted, i.e. the Patient and Family Centered Care Methodology and Practice (PFCCM/P).

This hybrid process map includes TDABC plus “behind-thescene activities” such as central sterile processing and billing, non-direct personnel time, and patient and family waiting time (DiGioia et al. 2016). However, we find the TDABC method reliable, reproducible, simple, and transparent to allow for comparison and strategic optimization of care processes. In a recent study from the US comparing TDABC for TKA in 29 hospitals, total personnel cost varied by a factor of 2.3 with nursing costs and length of stay being some of the more

Acta Orthopaedica 2018; 89 (5): 515–521

expensive drivers (Haas and Kaplan 2016). Discharge destination was another driver for total cost and hence the study supports evaluation of the logistical set-up, staff use, short stay, and discharge to home if feasible. The TDABC method is not without bias. Performance bias could occur as staff know time measurement is taking place and also external validity may be questioned as the 6 measured operations all represent flawless pathways without delays of any kind.

As the demand for THA and TKA is increasing in a financially challenged environment and an increasing number of procedures are likely to be performed on an outpatient basis, it is important to perform evaluation of pathways, of the staff minutes used, and of the economic expenses associated herewith.

HH, BBK, SEA, and KG conceived the study; all authors collected data and evaluated them. HH wrote the first draft of the manuscript, and all authors revised it and approved the final version to be published.

Acta thanks Ola Rolfson and other anonymous reviewers for help with peer review of this study.

Akhavan S, Ward L, Bozic K J. Time-driven activity-based costing more accurately reflects costs in arthroplasty surgery. Clin Orthop Relat Res 2016; 474(1): 8-15.

Andreasen S E, Holm H B, Jørgensen M, Gromov K, Kjærsgaard-Andersen P, Husted H. Time-driven activity-based cost of fast-track total hip and knee arthroplasty. J Arthroplasty 2017; 32(6): 1747-55.

Argenson J N, Husted H, Lombardi A Jr, Booth R E, Thienpont E. Global Forum: An international perspective on outpatient surgical procedures for adult hip and knee reconstruction. J Bone Joint Surg Am 2016; 98(13): e55.

Arshi A, Leong N L, D’Oro A, Wang C, Buser Z, Wang J C, Jones K J, Petrigliano F A, SooHoo N F. Outpatient total knee arthroplasty is associated with higher risk of perioperative complications. J Bone Joint Surg Am 2017; 99(23): 1978-86.

Aynardi M, Post Z, Ong A, Orozco F, Sukin D C. Outpatient surgery as a means of cost reduction in total hip arthroplasty: a case-control study. HSS J 2014; 10(3): 252-5.

Berger R A, Kusuma S K, Sanders S A, Thill E S, Sporer S M. The feasibility and perioperative complications of outpatient knee arthroplasty. Clin Orthop Relat Res 2009a; 467(6): 1443-9.

Berger R A, Sanders S A, Thill E S, Sporer S M, Della Valle C. Newer anesthesia and rehabilitation protocols enable outpatient hip replacement in selected patients. Clin Orthop Relat Res 2009b; 467(6): 1424-30.

Bovonratwet P, Webb M L, Ondeck N T, Lukasiewicz A M, Cui J J, McLynn R P, Grauer J N. Definitional differences of “outpatient” versus “inpatient” THA and TKA can affect study outcomes. Clin Orthop Relat Res 2017; 475(12): 2917-25.

Chen A, Sabharwal S, Akhtar K, Makaram N, Gupte C M. Time-driven activity based costing of total knee replacement surgery at a London teaching hospital. Knee 2015; 22: 640-5.

Courtney P M, Boniello A J, Berger R A. Complications following outpatient total joint arthroplasty: an analysis of a national database. J Arthroplasty 2017; 32(5): 1426-30.

Courtney P M, Froimson M I, Meneghini R M, Lee G C, Della Valle C J. Can total knee arthroplasty be performed safely as an outpatient in the Medicare population? J Arthroplasty 2018; 33(7S): S28-S31 DiGioia A M 3rd, Greenhouse P K, Giarrusso M L, Kress J M. Determining the true cost to deliver total hip and knee arthroplasty over the full cycle of care: preparing for bundling and reference-based pricing. J Arthroplasty 2016; 31(1): 1-6.

Acta Orthopaedica 2018; 89 (5): 515–521

Goyal N, Chen A F, Padgett S E, Tan T L, Kheir M M, Hopper R H Jr, 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.

Gromov K, Kjærsgaard-Andersen P, Revald P, Kehlet H, Husted H. Feasibility of outpatient total hip and knee arthroplasty in unselected patients. Acta Orthop 2017; 88(5): 516-21.

Haas D A, Kaplan R S. Variation in the cost of care for primary total knee arthroplasties. Arthroplasty Today 2016; 3(1): 33-7.

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.

Hoffmann J D, Kusnezov N A, Dunn J C, Zarkadis N J, Goodman G P, Berger R A. The shift to same-day outpatient joint arthroplasty: a systematic review. J Arthroplasty 2018; 33(4): 1265-1274 Huang A, Ryu J J, Dervin G. Cost savings of outpatient versus standard inpatient total knee arthroplasty. Can J Surg 2017; 60(1): 57-62.

Husted H. Fast-track hip and knee arthroplasty: clinical and organizational aspects. Acta Orthop 2012; 83(Suppl. 346): 1-39.

Husted H, Otte K S, Kristensen B B, Orsnes T, Kehlet H. Readmissions after fast-track hip and knee arthroplasty. Arch Orthop Trauma Surg 2010; 130(9): 1185-91.

Husted H, Lunn T H, Troelsen A, Gaarn-Larsen L, Kristensen B B, Kehlet H. Why still in hospital after fast-track hip and knee arthroplasty? Acta Orthop 2011; 82(6): 679-84.

Husted H, Jørgensen C C, Gromov K, Kehlet H; Lundbeck Foundation Center for Fast-track Hip and Knee Replacement Collaborative Group. Does BMI influence hospital stay and morbidity after fast-track hip and knee arthroplasty? Acta Orthop 2016; 87(5): 466-72.

Jørgensen C C, Kehlet H; Lundbeck Foundation Centre for Fast-track Hip and Knee Replacement Collaborative Group. Role of patient characteristics for fast-track hip and knee arthroplasty. Br J Anaesth 2013a; 110(6): 972-80.

Jørgensen C C, Jacobsen M K, Soeballe K, Hansen T B, Husted H, Kjærsgaard-Andersen P, Hansen L T, Laursen M B, Kehlet H. Thromboprophylaxis only during hospitalisation in fast-track hip and knee arthroplasty: a prospective cohort study. BMJ Open 2013b; 3(12): e003965.

Kaplan R S, Anderson S R. Time-driven activity-based costing. Harvard Bus Rev 2004; 82: 131-8.

Khan S K, Malviya A, Muller S D, Carluke I, Partington P F, Emmerson K P, Reed M R. Reduced short-term complications and mortality following enhanced recovery primary hip and knee arthroplasty: results from 6,000 consecutive procedures. Acta Orthop 2014; 85(1): 26-31.

Klein G R, Posner J M, Levine H B, Hartzband M A. Same day total hip arthroplasty performed at an ambulatory surgical center: 90-day complication rate on 549 patients. J Arthroplasty 2017; 32(4): 1103-6.

521

Larsen J R, Skovgaard B, Prynø T, Bendikas L, Mikkelsen L R, Laursen M, Høybye M T, Mikkelsen S, Jørgensen L B. Feasibility of day-case total hip arthroplasty: a single-centre observational study. Hip Int 2017; 27(1): 60-5.

Lovald S T, Ong K L, Malkani A L, Lau E C, Schmier J K, Kurtz S M, Manley M T. Complications, mortality, and costs for outpatient and short-stay total knee arthroplasty patients in comparison to standard-stay patients. J Arthroplasty 2014; 29(3): 510-15.

Lovecchio F, Alvi H, Sahota S, Beal M, Manning D. Is outpatient arthroplasty as safe as fast-track inpatient arthroplasty? A propensity score matched analysis. J Arthroplasty 2016; 31(Suppl. 9): 197-201.

Lunn T H, Kristensen B B, Gaarn-Larsen L, Husted H, Kehlet H. Postanaesthesia care unit stay after total hip and knee arthroplasty under spinal anaesthesia. Acta Anaesthesiol Scand 2012; 56(9): 1139-45.

Malviya A, Martin K, Harper I, Muller S D, Emmerson K P, Partington P F, Reed M R. Enhanced recovery program for hip and knee replacement reduces death rate. Acta Orthop 2011; 82(5): 577-81.

Meneghini R M, Ziemba-Davis M, Ishmael M K, Kuzma A L, Caccavallo P. Safe Selection of outpatient joint arthroplasty patients with medical risk stratification: the “Outpatient Arthroplasty Risk Assessment Score”. J Arthroplasty 2017; 32(8): 2325-31.

Mechanic R E. Mandatory Medicare bundled payment: is it ready for prime time? N Engl J Med 2015; 373(14): 1291-3.

Nelson S J, Webb M L, Lukasiewicz A M, Varthi A G, Samuel A M, Grauer J N. Is outpatient total hip arthroplasty safe? J Arthroplasty 2017; 32(5): 1439-42.

Nichols C I, Vose J G. Clinical outcomes and costs within 90 days of primary or revision total joint arthroplasty. J Arthroplasty 2016; 31(7): 1400-6.

Palsis J A, Brehmer T S, Pellegrini V D, Drew J M, Sachs B L. The cost of joint replacement: comparing two approaches to evaluating costs of total hip and knee arthroplasty. J Bone Joint Surg Am 2018; 100(4): 326-33.

Pollock M, Somerville L, Firth A, Lanting B. Outpatient total hip arthroplasty, total knee arthroplasty, and unicompartmental knee arthroplasty: a systematic review of the literature. JBJS Rev 2016; 4(12). pii: 01874474201612000-00004. doi: 10.2106/JBJS.RVW.16.00002.

Thienpont E, Lavand’homme P, Kehlet H. The constraints on day-case total knee arthroplasty: the fastest fast track. Bone Joint J 2015; 97-B(10 Suppl A): 40-4.

Vehmeijer S B W, Husted H, Kehlet H. Outpatient total hip and knee arthroplasty. Acta Orthop 2018; 89(2): 141-4.

522

Acta Orthopaedica 2018; 89 (5): 522–527

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 Urban BERG 1,2, Erik BÜLOW 1,3, Martin SUNDBERG 4,5, and Ola ROLFSON 1,3 1 Department

of Orthopaedics, Institute of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg; 2 Department of Surgery and Orthopaedics, Kungälv Hospital; 3 The Swedish Hip Arthroplasty Register; 4 Department of Orthopedics, Clinical Sciences, Lund University and Skåne University Hospital, Lund, Sweden; 5 The Swedish Knee Arthroplasty Register Correspondence: urban.berg@vgregion.se Submitted 2018-01-25. Accepted 2018-05-28.

Background and purpose — Fast-track care programs in elective total hip and knee replacement (THR/TKR) have been introduced in several countries during the last decade resulting in a significant reduction of hospital stay without any rise in readmissions or early adverse events (AE). We evaluated the risk of readmissions and AE within 30 and 90 days after surgery when a fast-track program was introduced in routine care of joint replacement at 8 Swedish hospitals.

Patients and methods — Fast-track care programs were introduced at 8 public hospitals in Västra Götaland region from 2012 to 2014. We obtained data from the Swedish Hip and Knee Arthroplasty Registers for patients operated with THR and TKR in 2011–2015. All readmissions and new contacts with the health care system within 3 months with a possible connection to the surgical intervention were requested from the regional patient register. We compared patients operated before and after the introduction of the fast-track program.

Results — Implementation of the fast-track program resulted in a decrease in median hospital length of stay (LOS) from 5 to 3 days in both THR and TKR. The total readmission rate < 90 days for THR was 7.2% with fasttrack compared with 6.7% in the previous program, and for TKR 8.4% in both groups. Almost half of the readmissions occurred without any AE identified. There was no statistically significant difference concerning readmissions or AE when comparing the programs.

Interpretation — Implementation of a fast-track care program in routine care of elective hip and knee replacement is effective in reducing hospital stay without increasing the risk of readmissions or adverse events within 90 days after surgery.

Fast-track care programs have been introduced in elective joint replacement in several countries during the last decade (Antrobus and Bryson 2011, Raphael et al. 2011, Fawcett et al. 2012, Husted 2012, Okamoto et al. 2016). Inventors and pioneers of the fast-track concept have reported considerable reduction of hospital length-of-stay (LOS) and high patient satisfaction without increased readmissions or adverse events within 90 days (Larsen et al. 2008, Husted et al. 2010a and b, Machin et al. 2013, Glassou et al. 2014, Khan et al. 2014).

A systematic review and meta-analysis of fast-track hip and knee arthroplasty (Zhu et al. 2017) concluded that the fasttrack concept reduces LOS and the incidence of complications but does not appear to impact the 30-day readmission rate.

However, most studies are observational cohort studies comparing the fast-track concept with historical data in providers dedicated to the new concept. Some uncertainty remains as to how to define fast-track as the care programs and clinical pathways are quite complex. The care processes in the cohorts not considered as fast-track differ in various aspects. Whilst LOS has gained much attention (Walters et al. 2016), one could argue for more focus on rapid recovery of function, and that short LOS should be balanced against any increase in morbidity and the cost of advanced follow-up (Thienpont et al. 2015, Lovecchio et al. 2016). One publication has reported a trend to higher infection rate after introducing the fast-track concept in THR (Amlie et al. 2016), but most studies are supporting equal or better results compared with conventional care in different outcome measures (Pilot et al. 2006, Husted et al.

2010a, b and c, Machin et al. 2013, Glassou et al. 2014, Khan et al. 2014, Stambough et al. 2015, Stowers et al. 2016, Delanois et al. 2017, Wilches et al. 2017, Zhu et al. 2017). It has been questioned whether the good results can be generalized and whether patient safety is ensured when the fast-track programs are broadly introduced in routine arthroplasty practice

© 2018 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/4.0) DOI 10.1080/17453674.2018.1492507

Acta Orthopaedica 2018; 89 (5): 522–527

523

THR in 8 public hospitals 2011–2015 n = 7,774

TKR in 8 public hospitals 2011–2015 n = 6,374

THR fast-track n = 3,915

TKR fast-track n = 3,430

Excluded Incomplete demographic data n = 58 Analyzed n = 3,857

THR not fast-track n = 3,859 Excluded Incomplete demographic data n = 135 Analyzed n = 3,724

Excluded Incomplete demographic data n = 17 Analyzed n = 3,413

TKR not fast-track n = 2,944 Excluded Incomplete demographic data n = 28 Analyzed n = 2,916

Figure 1. Patient allocation.

(Antrobus and Bryson 2011, Raphael et al. 2011, Jørgensen and Kehlet 2013). Our aim was to evaluate the risk of readmissions or adverse events (AE) within 30 and 90 days after surgery when a fast-track program was introduced in routine care of joint replacement in a defined region of Sweden.

Patients and methods To define the fast-track programs and the time of implementation a questionnaire was sent to hospitals performing elective hip and knee replacements in the Swedish Region Västra Götaland, a county council with a population of 1.7 million inhabitants. In 3 clinics without weekend service and exclusively patients with ASA 1–2, a care program based on the fast-track principles had already been implemented before 2011. These clinics were excluded from our study. In 8 public hospitals fast-track care programs were implemented between January 2012 and November 2014 at different times. We defined that fast-track was implemented when the following criteria for standard of care were met: (1) admission on the day of surgery, (2) mobilization within 3–6 hours after the operation, (3) functional discharge criteria in practice, and (4) an intended median length of stay (LOS) not more than 3 days. The patients were informed about the expected LOS, but the decision on discharge followed the functional ability and pain relief. However, regardless of whether the care program was defined as fast-track or not, the standard of care included written and oral structured information at a preoperative visit with a multiprofessional team 1–3 weeks before surgery, multimodal analgesia for pain relief, and tranexamic acid to reduce bleeding. Spinal anesthesia was routinely preferred supplemented by local infiltration analgesia in knee replacements but not in hips. 3 doses of cloxacillin were given on the day of surgery. The length of antithrombotic prophylaxis was 10 days in knees and 28–30 days in hips, but the antithrombotic drug varied between hospitals. No drains were used, a urinary catheter only in selected cases, and tourniquet in TKR was optional depending on the surgeon’s preference.

We collected data from the Swedish Knee and Hip Arthroplasty Registers and linked them to the regional patient regis-

ter. In the 8 hospitals 7,774 elective THRs and 6,374 TKRs for osteoarthritis were performed between 2011 and 2015. The LOS was defined as the number of days by using the date of discharge minus the date of admission as the formula for calculation. Data on readmissions and new contacts within 30 and 90 days after surgery were retrieved from the regional patient register. Not only were outpatient contacts at the hospitals requested but also contacts with primary health care. Adverse events were defined based on International Classification of Diseases (ICD-10) codes for diagnoses and the Nordic Medico-Statistical Committee (NOMESCO) Classification of Surgical Procedures codes for interventions. The code list has been elaborated by the Swedish Knee Arthroplasty Register (SKAR) in collaboration with the National Board of Health and Welfare to be used after knee replacements. Based on the same principles we elaborated a code list adapted for elective hip replacements. It includes all local complications, secondary fractures, and tendon ruptures in the lower extremity, thromboembolic events, myocardial infarction, pneumonia, gastro-duodenal ulcers, acute kidney injury, and urinary retention (Appendices 1 and 2, see Supplementary data).

The patients were divided into 2 groups depending on whether they were operated in a fast-track program or not (Figure 1). There were 3,915 THRs and 3,430 TKRs in fasttrack programs and 3,859 THRs and 2,944 TKRs in programs not considered as fast-track. No patients were excluded from the fast-track program after the implementation, and the intention was that all patients should follow the same clinical pathway and care program. If they stayed longer than 3 days, they were still included in the fast-track group.

Statistics Univariable and multivariable logistic regression analyses were used to evaluate the risk of readmissions and adverse events within 30 and 90 days. At low incidences of the outcome, odds ratios are good approximations for relative risks and can be interpreted as such (Davies et al. 1998). Relative risks were therefore approximated by odds ratios (OR) and estimated with a 95% confidence interval (CI). The results were considered statistically significant if observed p-values were smaller than 0.05. The statistical analyses were performed using R software

524

Acta Orthopaedica 2018; 89 (5): 522–527

Table 1. Demographics on total hip replacement patients and data on operations. Values are frequency and (percentage) unless ­otherwise stated

Factor

Not fast-track n = 3,859

Patients with complete data Mean LOS, days Median LOS, days (IQR) ASA 1–2 Age, mean (SD) Sex, female BMI, mean (SD) Posterolateral approach Direct lateral approach Cemented Uncemented Hybrid Reverse hybrid

3,724 5.8 5 (4–6) 3,138 (84) 69.5 (10.3) 2,186 (59) 27.5 (4.6) 1,021 (27) 2,688 (72) 2,582 (69) 675 (18) 150 (4.0) 317 (8.5)

Fast-track n = 3,915

p-value

3,857 3.7 < 0.001 3 (2–4) 3,395 (88) < 0.001 69.5 (10.1) 2,200 (57) 0.2 27.4 (4.6) 0.6 1,441 (37) 2,396 (62) 2,578 (67) 720 (19) 271 (7.0) 288 (7.5)

LOS: length of stay; IQR: interquartile range; ASA: American Society of Anesthesiologists; SD: standard deviation.

Table 2. Demographics on total knee replacement patients. Values are frequency and (percentage) unless ­otherwise stated

Factor

Not fast-track n = 2,944

Patients with complete data Mean LOS, days Median LOS, days (IQR) Age, mean (SD) Sex, female ASA 1–2 BMI, mean (SD)

Fast-track n = 3,430

2,916 5.4 5 (4–6) 69.5 (9.3) 1,734 (59) 2,493 (85) 29.2 (4.9)

p-value

3,413 3.3 < 0.001 3 (2–4) 68.8 (9.0) 0.002 1,952 (57) 0.08 3,035 (89) < 0.001 29.2 (4.60) 1.0

For abbreviations, see Table 1.

Percentage of THR and TKR in fast-track 100 THR TKR 80

60

package (version 3.4 or later; http://www.r-project.org) with the packages “tidyverse” (Hadley Wickham 2017) and “data. table” (Dowle and Srinivasan 2017).

Ethics, funding, and potential conflicts of interest The study was approved by the Regional Ethical Review Board in Gothenburg (Dnr 388-15, 2015-06-01 and 2015-0717, T 1107-16, 2016-12-15). Financial support was received from the Healthcare Committee, Region Västra Götaland. No competing interests were declared.

Results Implementation of the fast-track program reduced LOS with statistical significance for both THR and TKR. The mean LOS for hips decreased from 5.8 days to 3.7 and from 5.4 to 3.3 for knees. All hospitals except 1 achieved a median value of LOS 3 days or less compared with 5 days or more in the previous care program. After implementation of fast-track the capacity increased at the public hospitals and included more ASA 1–2 patients, who previously had been sent to private clinics or hospitals accepting only ASA 1–2. This can explain why ASA 1–2 was slightly more frequent in the fast-track group.

Demography, LOS, and surgical data are presented in Tables 1 and 2.

Fast-track care programs were implemented between January 2012 and November 2014 at different times in all 8 hospitals. Thus, in 2011 no patients were operated in the fast-track program and in 2015 all patients followed the program (Figure 2).

Most adverse events after surgery were identified at the hospital, either during the first hospital stay, at an unplanned

40

20

0 2011

2012

2013

2014

2015

Figure 2. Percentage of THR and TKR in fast-track programs 2011– 2015.

readmission, or at a new contact as outpatient not necessarily related to the joint replacement. However, 10–15% of the new contacts due to an adverse event after surgery were identified at a primary health center outside the hospital. The numbers of patients with the first identified AE at different levels in the health care system are presented in Tables 3 and 4 and the total number of readmissions in Table 5. Almost half of the readmissions were due to other reasons; no AE could be identified in association with the readmission.

Multivariable logistic regression showed no statistically significant influence of fast-track on readmission or adverse events. The OR of readmission after THR in the fast-track group was 1.2 (CI 0.9–1.5) within 30 days and 1.1 (CI 0.9– 1.3) within 90 days. The OR of AE within 30 days in the fasttrack group was 1.1 (CI 0.9–1.3) and 1.1 (0.9–1.2) within 90 days. For TKR the OR of readmission after TKA in the fasttrack program was 1.1 (CI 0.9–1.4) within 30 days and 1.1 (CI 0.9–1.3) within 90 days. The OR of AE after TKA was estimated at 1.1 (CI 0.9–1.3) within 30 days and 1.2 (1.0–1.4) within 90 days.

The overall complication rate was similar regardless of whether the fast-track program was applied or not both for the major local and general complications (Tables 6 and 7).

Acta Orthopaedica 2018; 89 (5): 522–527

525

Table 3. First care contact for adverse events (AEs) after total hip replacement. Values are frequency and (percentage)

Factor

Not fast-track n = 3,859

AE < 30 days Primary health care 13 (0.3) Hospital outpatient 82 (2.1) Hospitalization 156 (4.0) AE < 90 days Primary health care 35 (0.9) Hospital outpatient 110 (2.9) Hospitalization 194 (5.0)

Fast-track n = 3,915 25 (0.6) 82 (2.1) 162 (4.1) 46 (1.2) 107 (2.7) 196 (5.0)

Table 4. First care contact for adverse events (AEs) after total knee replacement. Values are frequency and (percentage)

Not fast-track n = 2,944

AE < 30 days Primary health care 17 (0.6) Hospital outpatient 70 (2.4) Hospitalization 126 (4.3) AE < 90 days Primary health care 29 (1.0) Hospital outpatient 115 (3.9) Hospitalization 156 (5.3)

Fast-track n = 3,430 29 (0.8) 97 (2.8) 140 (4.1) 54 (1.6) 162 (4.7) 173 (5.0)

Table 5. Total number of patients with readmissions < 30 and < 90 days after surgery. Values are frequency and (percentage)

THR, n All readmissions < 30 days Readmissions < 30 days with AE All Readmissions < 90 days Readmissions < 90 days with AE TKR, n All readmissions < 30 days Readmissions < 30 days with AE All readmissions < 90 days Readmissions < 90 days with AE

Not fast-track

Fast-track

3,859 168 (4.4) 97 (2.5) 260 (6.7) 141 (3.7) 2,944 159 (5.4) 75(2.5%) 246 (8.4) 120 (4.1)

3,915 196 (5.0) 111 (2.8) 281 (7.2) 151 (3.9) 3,430 193 (5.6) 93 (2.7) 288 (8.4) 139 (4.1)

Table 6. Adverse events < 90 days after total hip replacement according to ICD-10 codes. Values are frequency and (percentage)

Factor Codes Deep infection a Surgical site infection Hip dislocation Myocardial infarction Cerebrovascular event Deep venous thrombosis Pulmonary embolism Acute kidney injury Urinary retention Pneumonia Gastrointestinal ulcers Constipation a Prosthesis

Not fast-track n = 3,859

T84.5 T81.4 Combinations I21 I60–I65 I80 I26.0, I26.9 N17, N99.0 R33.9 J15–J18 K25–K27 K59.0

39 (1.0) 58 (1.5) 51 (1.3) 12 (0.3) 14 (0.4) 29 (0.8) 10 (0.3) 9 (0.2) 9 (0.2) 10 (0.3) 13 (0.3) 11 (0.3)

Fast-track n = 3,915 47 (1.2) 42 (1.1) 33 (0.8) 8 (0.2) 15 (0.4) 35 (0.9) 14 (0.4) 9 (0.2) 28 (0.7) 21 (0.5) 15 (0.4) 11 (0.3)

joint infection

Table 7. Adverse events < 90 days after total knee replacement according to ICD-10 codes. Values are frequency and (percentage)

Factor Codes Deep infection a Surgical site infection Knee stiffness with manipulation NGT19 Myocardial infarction Cerebrovascular event Deep venous thrombosis Pulmonary embolism Acute kidney injury Urinary retention Pneumonia Gastrointestinal ulcers Constipation

Not fast-track n = 2,944

Fast-track n = 3,430

T84.5 T81.4

49 (1.7) 32 (1.1)

43 (1.3) 42 (1.2)

M24.5 I21 I60–I65 I80 I26.0, I26.9 N17, N99.0 R33.9 J15–J18 K25–K27 K59.0

16 (0.5) 9 (0.3) 12 (0.4) 38 (1.3) 8 (0.3) 6 (0.2) 11 (0.4) 14 (0.5) 19 (0.6) 15 (0.5)

18 (0.5) 11 (0.3) 13 (0.4) 45 (1.3) 21 (0.6) 8 (0.2) 17 (0.5) 25 (0.7) 23 (0.7) 9 (0.3)

track program. 2 patients (0.1%) died within 90 days after TKR with fast-track and 6 patients (0.2%) with the previous care program.

Discussion Pulmonary embolism was slightly more frequent in the fasttrack group, particularly of TKR, but not statistically significant. The rates of clinical deep venous thrombosis (DVT) were almost the same. The number of patients with urinary retention was higher in the fast-track groups. We noticed that about 30% of them were treated at the health centers outside the hospitals.

However, for most AEs the differences are small, and we abstain from statistical analysis as the diagnosis of the reported events cannot be confirmed by medical records.

In the fast-track group 8 patients (0.2%) died within 90 days after THR compared with 11 patients (0.3%) without fast-

We did not find any statistically significant increase in readmissions or adverse events within 30 and 90 days after surgery when a fast-track program was implemented in routine care of elective total hip and knee replacement (THR and TKR) at 8 Swedish hospitals. As expected, the mean LOS decreased by more than 2 days in both THR and TKR—a reduction of more than one-third. The absence of statistically significant difference in AE frequency or readmission is not an evidence of equivalence. However, concerns about fast-track surgery causing increased AEs and readmission were not substantiated in our results. The absolute differences in AE frequencies between groups were 0.3% or smaller.

526

A recent study from Finland presented an increased rate of readmissions in one hospital after implementation of a fasttrack program (Pamilo et al. 2018). However, most other publications have reported unchanged rate of readmissions with fast-track total joint replacement (Husted et al. 2010b, Glassou et al. 2014, Khan et al. 2014).

One of the most severe and costly complications in joint replacement surgery is prosthetic joint infection (PJI). It has been questioned whether the fast-track program can increase the risk of PJI and the findings in a small sample from Norway (Amlie et al. 2016) raised concern leading to the program being abandoned after a short period. In contrast, a much larger study from Denmark (Glassou et al. 2014) reported that the risk of readmission due to infection may decrease over time after introduction of the new concept. In our study, the rate of PJI is slightly higher in the fast-track group for hips but lower for knees as defined by the ICD-10 code M845. The differences were not statistically significant. However, some uncertainty may remain regarding the real infection rate as the diagnosis is not confirmed by investigating medical records to know whether the criteria for PJI have been fulfilled. The aim of our study was to compare the groups and not to conclude the exact infection rate.

Among local complications, we found similar rates of knee stiffness requiring manipulation (0.5%) in the 2 groups. The manipulation rate is quite low, but longer follow-up may be needed (Husted et al. 2015). The rate of hip dislocation was low compared with other studies (Jorgensen et al. 2014) but probably reliable, as we have searched for all possible ICD-10 and NOMESCO codes in the regional patient register and not only readmissions. The rate was even lower in the fast-track cohort and we conclude that early discharge does not increase risk of hip dislocation.

It has been argued that early mobilization might reduce the risk of deep venous thrombosis (DVT) and the necessity of prolonged use of antithrombotic drugs has been questioned in the fast-track joint replacement setting (Husted et al. 2010c).

In our study, reduction of symptomatic DVT among fast-track patients could not be demonstrated. Contrary to the expected, we found a higher number of pulmonary embolisms among TKR patients operated in the fast-track program although this difference was not statistically significant. Some uncertainty remains as the difference in mobilization regime between the 2 groups is not evident, the figures are quite low, and confirmation of the diagnosis in medical records is lacking.

The rates of pneumonia, gastro-duodenal ulcers, and acute kidney injury were similar in the 2 groups. Postoperative urinary retention is common (Bjerregaard et al. 2015) but if the problem is resolved before discharge, it would not be identified as a complication. The diagnosis code for urinary retention was uncommon in this Swedish context. However, a very short length of hospital stay may increase the need for a new contact with the health system, and the higher rate in the fasttrack groups, especially in THR, may partly be explained by

Acta Orthopaedica 2018; 89 (5): 522â&#x20AC;&#x201C;527

the fact that more than one-third of the patients with urinary retention were identified as new contacts in the primary health care system shortly after discharge.

The overall rates of readmissions and AE were similar without any statistical significant difference between the groups.

Readmission rates and the incidence of AEs do not differ considerably from other studies (Husted et al. 2010b, Wolf et al. 2012, Zmistowski et al. 2013, Glassou et al. 2014) even if there are some differences in how the AEs are defined. The primary health care system was slightly more frequent as the first source of contact due to AEs in the fast-track group for both hips and knees but more than 85% of the AEs were identified and assessed at the hospitals as outpatients or inpatients.

The AEs identified in the primary health care system were dominated by medical events.

The mortality rate within 30 and 90 days after THR and TKR is generally low with fewer deaths after implementation of the fast-track program, but as the numbers are small they do not allow reliable conclusions.

Strengths and limitations The strength of this study is that we have investigated routine care in arthroplasty surgery without patient selection.

All hospitals had a defined date for the implementation of the fast-track according to our criteria. The national quality registers and patient register used have a very high completeness of data. The fact that also contacts in the primary health care system are included gives a more complete view of the short-term complications. However, we know that the coding in the health institutions is not always accurate. As we cannot confirm the accuracy of ICD-10 and NOMESCO codes by medical records some uncertainty remains concerning the correct incidence of specific complications. Another weakness is the difficulty in defining fast-track and in controlling all confounding factors in the clinical pathway and care process. The accuracy of LOS could be discussed as we had access only to the dates of admission and discharge and not the exact time point. Consequently, the LOS calculation may underestimate the real LOS if calculated by the hour, as the value of LOS is equal to the number of nights spent at the hospital. However, most publications report the LOS based on the same calculation (Husted et al. 2010a and b, Brock et al. 2017).

Conclusion Patient safety is preserved in the fast-track program but not better compared with other care programs for elective joint replacements in a Swedish context. To achieve reduction of adverse events more specific measures are needed.

Supplementary data Appendices 1 and 2 are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/ 17453674.2018.1492507

Acta Orthopaedica 2018; 89 (5): 522–527

The authors would like to thank Daniel Odin for help with parts of the statistical analyses.

UB, MS, and OR conceived and planned the study. UB and EB performed the statistical analyses. All authors discussed the results and commented on the manuscript, which was drafted by UB. Acta thanks Henrik Husted and other anonymous reviewers for help with peer review of this study.

Amlie E, Lerdal A, Gay CL, Hovik O, Nordsletten L, Dimmen S. A trend for increased risk of revision surgery due to deep infection following fast-track hip arthroplasty. Adv Orthop 2016; 2016: 7901953.

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): 891-6.

Bjerregaard L S, Bogo S, Raaschou S, Troldborg C, Hornum U, Poulsen A M, Bagi P, Kehlet H. Incidence of and risk factors for postoperative urinary retention in fast-track hip and knee arthroplasty. Acta Orthop 2015; 86(2): 183-8.

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.

Davies H T, Crombie I K, Tavakoli M. When can odds ratios mislead? BMJ (Clinical research ed.). 1998; 316(7136): 989-91.

Delanois R E, Gwam C, Mistry J B, Khlopas A, Chughtai M, Ramkumar P, Piuzzi N, Berger R, Bonutti P M, Malkani A L, Mont M A. Does length of stay influence how patients rate their hospitalization after total hip arthroplasty? Surg Technol Int 2017; 30: 393-8.

Dowle M, Srinivasan A. data.table - CRAN-R. 2017: https://cran.r-project. org/web/packages/data.table/data.table.pdf Fawcett W J, Mythen M G, Scott M J. Enhanced recovery: more than just reducing length of stay? Br J Anaesth 2012; 109(5): 671-4.

Glassou E N, Pedersen A B, Hansen T B. Risk of re-admission, reoperation, and mortality within 90 days of total hip and knee arthroplasty in fast-track departments in Denmark from 2005 to 2011. Acta Orthop 2014; 85(5): 493-500.

Husted H. Fast-track hip and knee arthroplasty: clinical and organizational aspects. Acta Orthop 2012; 83 (Suppl 346): 1-39.

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 2010a; 130(2): 263-8.

Husted H, Otte K S, Kristensen B B, Orsnes T, Kehlet H. Readmissions after fast-track hip and knee arthroplasty. Arch Orthop Trauma Surg 2010b; 130(9): 1185-91.

Husted H, Otte K S, Kristensen B B, Orsnes T, Wong C, Kehlet H. Low risk of thromboembolic complications after fast-track hip and knee arthroplasty.

Acta Orthop 2010c; 81(5): 599-605.

Husted H, Jorgensen C C, Gromov K, Troelsen A. Low manipulation prevalence following fast-track total knee arthroplasty. Acta Orthop 2015; 86(1): 86-91.

Jørgensen C C, Kehlet H. Role of patient characteristics for fast-track hip and knee arthroplasty. Br J Anaesth 2013; 110(6): 972-80.

527

Jorgensen C C, Kjaersgaard-Andersen P, Solgaard S, Kehlet H. Hip dislocations after 2,734 elective unilateral fast-track total hip arthroplasties: incidence, circumstances and predisposing factors. Arch Orthop Trauma Surg 2014; 134(11): 1615-22.

Khan S K, Malviya A, Muller S D, Carluke I, Partington P F, Emmerson K P, Reed M R. Reduced short-term complications and mortality following enhanced recovery primary hip and knee arthroplasty: results from 6,000 consecutive procedures. Acta Orthop 2014; 85(1): 26-31.

Larsen K, Hvass K E, Hansen T B, Thomsen P B, Soballe K. Effectiveness of accelerated perioperative care and rehabilitation intervention compared to current intervention after hip and knee arthroplasty: a before-after trial of 247 patients with a 3-month follow-up. BMC Musculoskelet Disord 2008; 9: 59.

Lovecchio F, Alvi H, Sahota S, Beal M, Manning D. Is outpatient arthroplasty as safe as fast-track inpatient arthroplasty? A propensity score matched analysis. J Arthroplasty 2016; 31(9 Suppl): 197-201.

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.

Okamoto T, Ridley R J, Edmondston S J, Visser M, Headford J, Yates P J. Day-of-surgery mobilization reduces the length of stay after elective hip arthroplasty. J Arthroplasty 2016; 31(10): 2227-30.

Pamilo K J, Torkki P, Peltola M, Pesola M, Remes V, Paloneva J. Reduced length of uninterrupted institutional stay after implementing a fast-track protocol for primary total hip replacement. Acta Orthop 2018; 89(1): 10-16.

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.

Raphael M, Jaeger M, van Vlymen J. Easily adoptable total joint arthroplasty program allows discharge home in two days. Can J Anaesth 2011; 58(10): 902-10.

Stambough J B, Nunley R M, Curry M C, Steger-May K, Clohisy J C. Rapid recovery protocols for primary total hip arthroplasty can safely reduce length of stay without increasing readmissions. J Arthroplasty 2015; 30(4): 521-6.

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.

Thienpont E, Lavand’homme P, Kehlet H. The constraints on day-case total knee arthroplasty: the fastest fast track. Bone Joint J 2015; 97-B(10 Suppl A): 40-4.

Walters M, Chambers M C, Sayeed Z, Anoushiravani A A, El-Othmani M M, Saleh K J. Reducing length of stay in total joint arthroplasty care. Orthop Clin North Am 2016; 47(4): 653-60.

Wilches C, Sulbaran J D, Fernandez J E, Gisbert J M, Bausili J M, Pelfort X. Fast-track recovery technique applied to primary total hip and knee replacement surgery: analysis of costs and complications. Rev Esp Cir Ortop Traumatol 2017; 61(2): 111-16.

Wolf B R, Lu X, Li Y, Callaghan J J, Cram P. Adverse outcomes in hip arthroplasty: long-term trends. J Bone Joint Surg Am 2012; 94(14): e103.

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.

Zmistowski B, Restrepo C, Hess J, Adibi D, Cangoz S, Parvizi J. Unplanned readmission after total joint arthroplasty: rates, reasons, and risk factors. J Bone Joint Surg Am 2013; 95(20): 1869-76.

528

Acta Orthopaedica 2018; 89 (5): 528–534

The course of pain and function in osteoarthritis and timing of arthroplasty: the CHECK cohort Maaike G J GADEMAN 1,2, Hein PUTTER 3, Wilbert B VAN DEN HOUT 3, Margreet KLOPPENBURG 4, Stefanie N HOFSTEDE 3, Suzanne C CANNEGIETER 2, Rob G H H NELISSEN 1, and Perla J MARANG–VAN DE MHEEN 3 1 Department 3 Department

of Orthopaedics, Leiden University Medical Center; 2 Department of Clinical Epidemiology, Leiden University Medical Center; of Biomedical Data Sciences;, Leiden University Medical Center; 4 Department of Rheumatology, Leiden University Medical Center, Leiden, The Netherlands Correspondence: m.g.j.gademan@lumc.nl Submitted 2018-02-02. Accepted 2018-06-11.

Background and purpose — It is unknown whether different trajectories of pain or function are associated with timing of total hip or knee arthroplasty (THA/TKA) in osteoarthritis (OA) patients. We investigated this association in early symptomatic OA patients.

Patients and methods — Data from the prospective Dutch CHECK cohort (patients with early hip/knee OA complaints) covering 9 years of follow-up were used. Pain and function were measured annually using the WOMAC questionnaires. Changes in pain/function over time were estimated using a linear mixed model adjusted for baseline age, sex, BMI, maximal Kellgren and Lawrence score, number of painful joints, and comorbidities. The same covariates were included in a Cox regression model, with time to first arthroplasty as event. Both were combined in a joint model to assess the association between changes in pain/function and time to arthroplasty.

Results — Of the 868 eligible patients, 84 received a TKA/THA during follow-up. Patients receiving arthroplasty were somewhat older, had a higher Kellgren and Lawrence score and worse WOMAC scores at baseline. Irrespective of receiving arthroplasty, about two-thirds of the patients showed at least 1 period of deterioration of pain/function (≥ 10 points WOMAC subscale). In approximately two fifths this deterioration was followed by another deterioration in the following year. Worse pain and function levels increased the hazard of receiving THA/TKA (1.08 [95% CI 1.06–1.10] for pain and 1.07 [CI 1.05–1.08] for function). Changes in pain or function over time were not associated with timing of THA/TKA Interpretation — Worse pain and function levels rather than long-term changes are associated with timing of THA/ TKA.

Although total knee arthroplasty (TKA) and total hip arthroplasty (THA) are effective interventions, their optimal timing is unknown. This is illustrated by the varying disease severity when surgery is performed across centres in Europe and Australia (Ackerman et al. 2009, Dieppe et al. 2009). Indication criteria for TKA/THA in guidelines acknowledge pain, function, radiological changes, and insufficient effect of nonoperative therapy (Jordan et al. 2003, Zhang et al. 2005, 2008, Gademan et al. 2016). Therefore the first step towards knowledge on optimal timing of TKA/THA is to investigate the time course of pain, function, and joint degeneration before surgery.

Several studies have shown that pain and function in knee osteoarthritis (OA) are persistent rather than progressive (Leffondre et al. 2004, Yusuf et al. 2011, Pisters et al. 2012, Collins et al. 2014, Wesseling et al. 2015). For example, in our Cohort Hip and Cohort Knee (CHECK) consisting of patients with pain and/or stiffness of knees and/or hips we showed previously that 3 stable pain trajectories were found with marginal, mild, and moderate pain (Wesseling et al. 2015). Similar results have been shown for function (Leffondre et al. 2004, Yusuf et al. 2011, Pisters et al. 2012, Riddle et al. 2013, Collins et al. 2014, Wesseling et al. 2015). If pain and function in OA are indeed stable and not progressive, optimal timing of THA/TKA would depend only on the prosthesis life span because the gain after surgery will be equal irrespective of timing. However, if the condition is progressive, one would expect TKA/THA patients to experience increased pain and declining function, particularly in the years before arthroplasty. Here gain after surgery may be less if preoperative pain/function is worse (Hofstede et al. 2016); patients with worse preoperative pain/function tend to have worse outcomes although they improve more than patients with fewer preoperative complaints. Only 1 previous study investigated trajectories of pain and function before arthroplasty and showed that mean pain increased and function decreased during the last 2.5

© 2018 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/4.0) DOI 10.1080/17453674.2018.1502533

Acta Orthopaedica 2018; 89 (5): 528–534

years before surgery (Riddle et al. 2013). However, there was no control group, so it is possible that patients not receiving arthroplasty reported similar changes in pain/function.

Therefore, we assessed whether different trajectories of pain or function are associated with timing of THA/TKA. To answer this question we combined the repeated measurements of pain and function among early symptomatic OA patients over 9 years, with a time-to-event analysis.

Patients and methods Study type and setting This is a nationwide Dutch observational prospective cohort study conducted in cooperation with 10 general and academic hospitals located in urbanized and semi-urbanized regions.

Population A detailed description of the CHECK cohort has been reported elsewhere (Wesseling et al. 2014). Eligible patients were recruited between 2002 and 2005 and had knee/hip pain, were between 45 and 65 years of age and within 6 months of their first GP visit for these complaints. There were no radiographic signs required to take part in the cohort. Exclusion criteria were: any other pathological condition than OA explaining the symptoms, comorbidity precluding physical evaluation or follow-up of 10 years, malignancy in the past 5 years, and inability to understand Dutch. We included 1,002 patients.

After inclusion, patients were divided into 2 groups: patients with relatively more severe symptoms (n = 861, Figure 1, see Supplementary data) visited the research centre each year for collection of clinical, radiological, and biochemical data; patients with mild symptoms (n = 141) visited the research centre at years 0, 2, 5, 8, and 10. Patients could shift to the more serious symptoms group, and were then measured annually as well (n = 79 after 5 years) (Wesseling et al. 2014).

Measurements Demographics At baseline patients reported BMI and the number of selfreported comorbidities according to the Statistics Netherlands questionnaire (Botterweck et al. 2001). Self-reported stiffness was assessed at baseline using the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) stiffness subscale (Bellamy et al. 1988). At baseline, 2, 5, and 8 years knee radiographs were taken in a weight-bearing semi-flexed posteroanterior view and for the hip in weight-bearing anteroposterior radiographs of the pelvis with hips in 15° internal rotation. Both knee and hip radiographs were read by observers blinded to all patient characteristics and the radiographs were scored according to Kellgren and Lawrence (K&L). In 38 participants radiographs were blindly scored by 5 trained observers (4 research assistants and 1 experienced general practitioner reader) in a paired fashion, with known sequence and interob-

529

server variability was tested (Cohen’s kappa = 0.60 for K&L ≥ 2 in knees at 5-year follow-up). The Cohen’s kappa is rather low because of the relatively low frequency of radiographic abnormalities. Therefore a prevalence-adjusted bias-adjusted kappa (PABAK) score was assessed. The PABAK score for reliability on progression of OA (KL score) in the knee from 0 to 5 years was 0.82, with a 90% average agreement. Similar results were found for the hip (Damen et al. 2014).

Pain and function At baseline, pain history was assessed by a rheumatologist.

Knee and hip pain were classified as present or absent and the total number of painful joints was calculated. Self-reported pain and function were assessed at each visit using the WOMAC pain/function subscales, each scored on a 5-point Likert scale (0 = no pain/good function (Bellamy et al. 1988).

The pain and function subscales comprise respectively 5 and 17 questions.

Arthroplasty During each visit patients were asked whether they had received a TKA/THA. If so, the date of surgery was noted.

Specific methods current study We included all measurements until 9 years of follow-up. Only patients with at least 2 measurements and no missing values for baseline characteristics were included (sex, age, BMI, maximal K&L, hip/knee joint pain, and comorbidities) (n = 868, i.e., 87%). The median follow-up was 9.0 (IQR 8.9–9.1) years. We used questionnaire data, data derived by a clinician from medical records, and radiographical data (Figure 1, see Supplementary data). Pain and function were assessed by the standardised WOMAC pain score ((total pain score) × (100/20)) and function score ((total function score) × (100/68)). To investigate whether THA/TKA patients more often experienced episodes of deteriorating pain or function prior to THA/TKA, compared with patients not receiving arthroplasty, we defined deteriorating pain or function as the first ≥ 10 points increase on the standardised WOMAC pain/function subscales compared with the year before. Further deterioration in the year thereafter was defined as any increase in WOMAC pain/function. When patients received multiple arthroplasties during follow-up, the first arthroplasty was taken as event, and their follow-up was censored. Baseline maximal K&L was defined as the highest K&L in the hip/knee joints at baseline. Maximal K&L during follow-up was defined as the highest K&L in the hip/knee in all measurements.

Statistics Differences in baseline characteristics between patients with and without a TKA/THA during follow-up were investigated with Student’s t-test (continuous outcomes) or a chi-square test (categorical outcomes). To depict the timing of arthroplasties, a Kaplan–Meier curve was estimated.

530

Acta Orthopaedica 2018; 89 (5): 528–534

Descriptive statistics of the episodes of deterioration were reported, e.g., proportion of patients with deterioration of pain/function among those with and without arthroplasty. Differences between groups were evaluated with logistic regression adjusted for time to follow-up.

In addition we fitted linear mixed-effects models with a random slope and random intercept per patient to the reported levels of pain and function. The random-effects part describes the time course of pain and function for each patient and takes into account the within-subject correlation of different measurements. Within these models we adjusted for age, sex, BMI, maximal K&L, number of painful joints, and comorbidities at baseline. As such we could estimate the adjusted course of pain and function for each patient separately. Pain and function are not assumed to be constant between successive measurements, so that direct inclusion in a time-dependent Cox model may produce biased results. Therefore, we included the estimated courses of pain from the linear mixed-effect model (each point in time can be estimated with this model) as a time-dependent covariate in the Cox model. The latter was done by joint modelling the longitudinal and survival data with the JM package (1.4-2) (R version 3.2.3) (Rizopoulos 2010) (Supplementary file). We used a piecewise constant baseline hazard. In the Cox model we adjusted for age, sex, BMI, maximal K&L, number of painful joints, and comorbidities at baseline.

Table 1. Baseline patient characteristics of patients who received total joint replacement somewhere during follow-up and the patients who did not

Sensitivity analysis We adjusted for maximal K&L during follow-up instead of baseline maximal K&L, as radiographic joint deterioration is an indication for THA/TKA. Second, we checked whether adding the slope of the linear mixed model to the joint model changed our results. As the slope refers to the long-term change over time in pain/function during 9 years of follow-up, we assessed whether these were associated with the likelihood of receiving arthroplasty.

included (n = 134), our population had similar baseline characteristics, with the exception of better WOMAC scores (Table 2, see Supplementary data). Of the 84 arthroplasties, 67 were implanted in the first 5 years (Figure 2, see Supplementary data). Mean WOMAC scores before arthroplasty were respectively 43.5 (SD 20.6) for function and 45.2 (19.2) for pain.

Ethics, funding, and potential conflicts of interest The medical ethics committees of all participating centres approved the study and all patients gave written informed consent. This study was funded by CHECK and by a separate grant (Dutch Arthritis Foundation (ARGON, BP12-3-401)).

This foundation did not play a role in the study’s design, conduct, or reporting. The authors have no conflicts of interest to declare.

Results Of the 868 included patients, 84 received TKA (n = 29) or THA (n = 55) during 9 years of follow-up. Patients receiving an arthroplasty were somewhat older at baseline, had higher K&L scores, and worse WOMAC scores compared with patients without arthroplasty (Table 1). None of the patients had a maximal K&L score ≥ 2. Compared with patients not

Total joint replacement No Yes n = 784 n = 84

Age (years) Sex: Male Female BMI Comorbidities Number of painful joints: 1 2 3 4 Maximal Kellgren and Lawrence score: 0 1 WOMAC standardized subscales: Pain Function Stiffness

56 (5.3)

58 (4.3)

p-value 0.001

161 (21) 18 (21) 623 (80) 66 (79) 26 (4.1) 27 (4.6) 1.9 (1.5) 1.7 (1.5)

0.9 0.4 0.3

230 (29) 338 (43) 114 (15) 102 (13)

32 (38) 32 (38) 11 (13) 9 (11)

0.4

291 (37) 493 (63)

11 (13) 73 (87)

< 0.001

24 (16) 22 (17) 32 (21)

33 (19) 32 (17) 38 (20)

< 0.001 < 0.001 < 0.01

WOMAC: Western Ontario and McMaster Universities Osteoarthritis Index. Continuous variables are shown as mean (SD), categorical variables are shown as number (percentage)

Episodes of deterioration of pain and function There was no difference in the percentage of patients that showed at least 1 period of deterioration of pain between patients who did or did not receive TKA/THA during followup: 56/85 patients versus 531/784. 13 of these 56 TKA/THA patients showed further deterioration of pain in the year thereafter against 88 of the 531 patients without arthroplasty.

Similar results were found for function; at least 1 period of deterioration of function before arthroplasty was found in 56 of 84 TKA/THA patients and in 540 of 784 patients without arthroplasty. This deterioration was followed by further deterioration in the following year in 14 of these 56 the patients receiving arthroplasty during follow-up and in 118 of the 540 patients who did not.

Hence, there are episodes of deterioration of pain/function in patients both with and without arthroplasty. However, patients without a prosthesis had more time to develop deterioration than patients who received an arthroplasty (median follow-up time 9.0 years [IQR 9.0–10.0] versus 4.5 years [IQR 3.3–6.3], p < 0.001). When corrected for follow-up duration, arthroplasty patients had higher odds on a first deterioration in pain

Acta Orthopaedica 2018; 89 (5): 528–534

531

Figure 3. Spaghetti plots of the course of pain and function, (left) individual WOMAC pain scores, and (right) individual WOMAC function scores. The blue lines represent the mean WOMAC score.

Table 3. Fixed effects of adjusted linear mixed-effects models describing the course of pain and function over time

Factor

Estimates for pain Beta (95% CI)

Time (years) –0.1 (–0.2 to 0.1) Sex a 3.8 (1.5 to 6.0) Age (years) 0.0 (–0.2 to 0.2) BMI 0.8 (0.6 to 1.0) Maximal Kellgren and Lawrence score 1.0 (–0.9 to 2.8) Number of painful joints 2.1 (1.2 to 3.0) Comorbidities 2.5 (1.9 to 3.1) a Men

Estimates for function Beta (95% CI) 0.3 (0.2 to 0.4) 2.6 (0.3 to 4.9) 0.1 (–0.1 to 0.3) 0.9 (0.7 to 1.1) 1.6 (–0.4 to 3.5) 2.4 (1.5 to 3.4) 2.8 (2.1 to 3.4)

as reference category

Table 4. Effect estimates of adjusted joint models for total hip and knee arthroplasty

Factor

Estimates for pain HR (95% CI)

Pain 1.08 (1.06–1.10) Function Sex a 0.61 (0.35–1.07) Age (years) 1.06 (1.01–1.10) BMI 0.94 (0.89–0.99) Maximal Kellgren and Lawrence score 2.96 (1.55–5.67) Number of painful joints 0.73 (0.57–0.93) Comorbidities 0.77 (0.66–0.91)

Estimates for function HR (95% CI) 1.07 (1.05–1.08) 0.76 (0.44: 1.31) 1.05 (1.01–1.10) 0.94 (0.90–1.00) 2.95 (1.54–5.64) 0.72 (0.56–0.92) 0.79 (0.68–0.92)

HR = hazard ratio. a Men as reference category

(odds ratio 3.2 [95%CI 1.4–7.2]) and function (odds ratio 2.3 [CI 1.1–5.0]) than patients without arthroplasty.

the hazard of receiving THA/TKA (1.08 [CI 1.06–1.10] for pain and 1.07 [CI 1.05–1.08] for function) (Table 4).

Course of pain/ function and timing of arthroplasty Individual pain and function trajectories are depicted in Figure 3. Patients had some variation in pain/function levels over time (deteriorations were often followed by improvements), but overall pain and function seemed to be stable during follow-up as depicted by the plotted line. In accordance with this, our adjusted mixed models showed stable pain levels over time, –0.1 (CI –0.2 to 0.1) points/year (Table 3), which is –0.5 points/decade (CI –2.1 to 0.5). Function significantly deteriorated over time by 0.3 (CI 0.2–0.4) points/year on a 100-point scale (Table 3), which is 2.6 points/decade (CI 1.6–4.2).

The estimates from the joint model showed that higher levels on WOMAC score of pain and function significantly increased

Sensitivity analysis Adjusting for maximal K&L during follow-up did not change our results (data not shown). When investigating the effect of the long-term changes of pain and function over time by adding the slope of the initial mixed models, the effect estimates of pain and function levels on receiving THA/TKA did not change (Table 5, see Supplementary data), nor did the slope itself increase the risk of receiving THA/TKA. The wide 95% CIs show that the model could not properly estimate the effect of the slope. Hence, long-term changes in pain or function over time did not affect the risk of receiving THA/TKA, when adjusted for other covariates and the level of pain/function.

532

Discussion In this cohort of patients with early OA symptoms at inclusion, we investigated whether pain and function changes were associated with receiving THA/TKA. During 9 years of follow-up one-tenth of the patients received an arthroplasty.

Approximately two-thirds of all patients showed at least 1 episode of deterioration of pain or function during followup. In about one-fifth these deteriorations were followed by another deterioration in the following year. At group level, pain and function remained fairly stable over time. We showed that higher pain and function levels were associated with an increased risk of receiving THA/TKA. Adding the individual long-term changes in pain or function over time did not affect the risk of receiving THA/TKA. Thus, it seems that pain and function levels rather than long-term changes are associated with timing of THA /TKA. At group level, we showed fairly stable levels of pain and function over 9 years of follow-up, suggesting that major debilitating variables in OA, like pain/functional loss, are persistent rather than worsening. This is in accordance with other studies. A study from the Osteoarthritis Initiative identified 5 relatively stable pain trajectories over 6 years of time in knee OA patients (Collins et al. 2014). These trajectories differed in severity, but all remained stable during follow-up and none showed considerable deterioration/improvement. Recently, in the same cohort White et al. (2016) showed 5 different trajectories of function over a period of 7 years. Overall, function remained stable, although a subgroup of 5% of the cohort showed progressive deterioration. However, the mean deterioration was only 13 out of 68 points on the WOMAC scale. A different study identified 4 trajectories in total WOMAC score in hip/knee OA: increasing scores (18% of the cohort), stable (40%), decreasing (24%), and unstable trajectories (18%) (Leffondre et al. 2004). Hence, for most patients, trajectories remained fairly stable in these studies, although in subgroups of patients improvement and deterioration were present.

The only study investigating trajectories of pain and function in arthroplasty patients showed that patients tend to worsen in function and pain in the last 2 years before arthroplasty (Riddle et al. 2013). However, they did not compare pain/function trajectories with patients not receiving arthroplasty. Our study showed that the long-term change in pain and function over time was not associated with arthroplasty, and thus not different from those not receiving arthroplasty, whereas levels of pain and function were. As patients receiving an arthroplasty reported worse function and more pain at baseline, the levels at which a patient presents him-/herself at first visit for OA complaints seem to determine mostly the risk of arthroplasty. This conclusion is strengthened by the finding that approximately 80% of the arthroplasties were conducted within the first 5 years. Moreover we assembled all patients at inception of symptoms, implying a rather similar disease stage

Acta Orthopaedica 2018; 89 (5): 528â&#x20AC;&#x201C;534

among all patients. As patients receiving THA/TKA showed more complaints at baseline than patients not receiving a prosthesis, these patients may represent a different patient group.

One should try to identify this group when commencing clinical care, so that early non-operative treatment can be better targeted.

However, besides modelling the natural course of OA, we also modelled surgeonsâ&#x20AC;&#x2122; and patientsâ&#x20AC;&#x2122; behaviour as they decide together that arthroplasty is warranted. If a patient is eager to have an arthroplasty he/she has a higher chance of receiving it than when the same patient is reluctant. Moreover, some surgeons will advise arthroplasty sooner than others. These variations might be amplified by the absence of clear indication criteria for THA/TKA (Ackerman et al. 2009, Dieppe et al. 2009, Gademan et al. 2016, Skou et al. 2016, Riddle and Perera 2017). As patients receiving arthroplasty may not have been the only ones needing arthroplasty, these variations may have diluted our analysis, leading to either an under- or an overestimation of our effect sizes. Nonetheless, by modelling clinical practice, our estimates represent real-life effect sizes.

Concerning optimal timing of arthroplasty, one could speculate that it is beneficial to postpone surgery when possible to reduce the risk of revision surgery. Patients with better preoperative function attain better postoperative functional levels than patients with worse preoperative function (Hofstede et al. 2016) but as we showed that patients seem to remain fairly stable over time, then lowering the risk of revision surgery by postponing the primary surgery might outweigh the risk of fast deterioration from a lifetime perspective. Nonetheless, no conclusions about timing can be based on this single study with early OA patients at baseline. Our results first need to be validated in other OA cohorts.

Limitations of our study include that we had no information on pharmacological or other non-operative treatment. Different treatment strategies might lead to differences in pain/ function trajectories. However, as THA/TKA is the end-stage intervention, both surgeon and patient were convinced that THA/TKA was the ultimate treatment option. Furthermore, patients were treated according to the Dutch OA guidelines.

Therefore we expect this to have only small effects on our results. Second, patients had early OA symptoms at baseline and therefore represent a different patient group than those seen by an orthopedic surgeon. This could make our results less generalizable to orthopedic practice. However, by the time the patients received arthroplasty, they were seen by an orthopedic surgeon. Moreover, an advantage of this population is that we could assess the course of pain/function from first complaints onwards, giving a complete view of the course of OA complaints. Moreover, by including all patients at inception of their complaints, selection bias was avoided. Nonetheless our results should be validated in another OA population. Third, our analysis concerns actual THA/TKA decisions, which may not be indicative of optimal THA/TKA decision-making.

Fourth, often knee and hip OA are reported separately, but in

Acta Orthopaedica 2018; 89 (5): 528–534

the current study we could not make a such a distinction as patients often had both hip and knee complaints. Finally, we are aware that adjusting for maximal K&L during follow-up is not a statistically preferred method (maximal K&L should have been added as a time-varying covariate). However, this method would complicate our analysis and the interpretation of our results. Furthermore, adjusting for the maximal K&L most likely reflects the maximum effect of adjustment.

In summary, pain and function levels rather than long-term pain and function changes over time (which can be seen as the progression rate) are associated with timing of THA/TKA in early OA patients.

Supplementary data Tables 2 and 5, and Figures 1–2 are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674.2018.1502533

CHECK is funded by the Dutch Arthritis Foundation under the lead of a steering committee comprising 16 members with expertise in different fields of OA chaired by Prof. Dr. J.W.J. Bijlsma, coordinated by J. Wesseling, and performed within Erasmus Medical Center Rotterdam; Kennemer Gasthuis Haarlem; Leiden University Medical Center; Maastricht University Medical Center; Martini Hospital Groningen/Allied Health Care Center for Rheumatolgy and Rehabilitation Groningen; Medical Spectrum Twente Enschede/ Ziekenhuisgroep Twente Almelo; Reade/VU Medical Center Amsterdam; St. Maartens-kliniek Nijmegen; University Medical Center Utrecht and Wilhelmina Hospital Assen. In addition, the current study was funded by a separate grant (Dutch Arthritis Foundation (ARGON, BP12-3-401)). This foundation did not play a role in the study’s design, conduct, or reporting.

The authors would like to thank Loïc Ferrer and Dimitris Rizopoulos for their help with the R package JM. The members of the ARGON consortium are thanked for their help and advice (http://www.artroseresearch.nl).

MG, HP, and PM contributed to the conception and design of the study. MK contributed to the acquisition of the data. MG and HP performed the statistical analysis. MG drafted the first version of the study. All authors interpreted the data and revised the manuscript.

Acta thanks Max Gordon and Sten Rasmussen for help with peer review of this study.

Ackerman I N, Dieppe P A, March L M, Roos E M, Nilsdotter A K, Brown G C, Sloan K E, Osborne R H. Variation in age and physical status prior to total knee and hip replacement surgery: a comparison of centers in Australia and Europe. Arthritis Rheum 2009; 61(2): 166-73. doi: 10.1002/ art.24215.

Bellamy N, Buchanan W W, Goldsmith C H, Campbell J, Stitt L W. Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol 1988; 15(12): 1833-40.

Botterweck A, Frenken F, Janssen S, Rozendaal L, de Vree M, Otten F. Plausibility new measurements general health and lifestyles. [Article in Dutch].

Heerlen: Centraal Bureau voor de Statistiek; 2001.

533

Collins J E, Katz J N, Dervan E E, Losina E. Trajectories and risk profiles of pain in persons with radiographic, symptomatic knee osteoarthritis: data from the osteoarthritis initiative. Osteoarthritis Cartilage 2014; 22(5): 62230. doi: S1063-4584(14)01010-3 [pii]; 10.1016/j.joca.2014.03.009 [doi].

Damen J, Schiphof D, Wolde S T, Cats H A, Bierma-Zeinstra S M, Oei E H. Inter-observer reliability for radiographic assessment of early osteoarthritis features: the CHECK (cohort hip and cohort knee) study. Osteoarthritis Cartilage/OARS, Osteoarthritis Research Society 2014; 22(7): 969-74. doi: 10.1016/j.joca.2014.05.007.

Dieppe P, Judge A, Williams S, Ikwueke I, Guenther K P, Floeren M, Huber J, Ingvarsson T, Learmonth I, Lohmander L S, Nilsdotter A, Puhl W, Rowley D, Thieler R, Dreinhoefer K, Group ES. Variations in the pre-operative status of patients coming to primary hip replacement for osteoarthritis in European orthopaedic centres. BMC Musculoskeletal Disord 2009; 10: 19. doi: 10.1186/1471-2474-10-19.

Gademan M G, Hofstede S N, Vliet Vlieland T P, Nelissen R G, Marang-van de Mheen P J. Indication criteria for total hip or knee arthroplasty in osteoarthritis: a state-of-the-science overview. BMC Musculoskeletal Disord 2016; 17(1): 463. doi: 10.1186/s12891-016-1325-z.

Hofstede S N, Gademan M G, Vliet Vlieland T P, Nelissen R G, Marang-van de Mheen P J. Preoperative predictors for outcomes after total hip replacement in patients with osteoarthritis: a systematic review. BMC Musculoskeletal Disord 2016; 17: 212. doi: 10.1186/s12891-016-1070-3.

Jordan K M, Arden N K, Doherty M, Bannwarth B, Bijlsma J W, Dieppe P, Gunther K, Hauselmann H, Herrero-Beaumont G, Kaklamanis P, Lohmander S, Leeb B, Lequesne M, Mazieres B, Martin-Mola E, Pavelka K, Pendleton A, Punzi L, Serni U, Swoboda B, Verbruggen G, Zimmerman-Gorska I, Dougados M, Standing Committee for International Clinical Studies Including Therapeutic Trials E. EULAR Recommendations 2003: an evidence based approach to the management of knee osteoarthritis: report of a Task Force of the Standing Committee for International Clinical Studies Including Therapeutic Trials (ESCISIT). Ann Rheum Dis 2003; 62(12): 1145-55.

Leffondre K, Abrahamowicz M, Regeasse A, Hawker G A, Badley E M, McCusker J, Belzile E. Statistical measures were proposed for identifying longitudinal patterns of change in quantitative health indicators. J Clin Epidemiol 2004; 57(10): 1049-62. doi: S0895-4356(04)00089-7 [pii]; 10.1016/j.jclinepi.2004.02.012 [doi].

Pisters M F, Veenhof C, van Dijk G M, Heymans M W, Twisk J W, Dekker J. The course of limitations in activities over 5 years in patients with knee and hip osteoarthritis with moderate functional limitations: risk factors for future functional decline. Osteoarthritis Cartilage 2012; 20(6): 503-10. doi: S1063-4584(12)00063-5 [pii]; 10.1016/j.joca.2012.02.002 [doi].

Riddle D L, Perera R A. Appropriateness and total knee arthroplasty: an examination of the American Academy of Orthopaedic Surgeons appropriateness rating system. Osteoarthritis Cartilage/OARS, Osteoarthritis Research Society 2017; 25(12): 1994-8. doi: 10.1016/j.joca.2017.08.018.

Riddle D L, Perera R A, Stratford P W, Jiranek W A, Dumenci L. Progressing toward, and recovering from, knee replacement surgery: a five-year cohort study. Arthritis Rheum 2013; 65(12): 3304-13. doi: 10.1002/art.38139 [doi].

Rizopoulos D. JM: An R package for the joint modelling of longitudinal and time-to-event data. J Statistical Software 2010; 35(9): 1-33.

Skou S T, Roos E M, Laursen M B, Rathleff M S, Arendt-Nielsen L, Simonsen O, Rasmussen S. Criteria used when deciding on eligibility for total knee arthroplasty: between thinking and doing. The Knee 2016; 23(2): 300-5. doi: 10.1016/j.knee.2015.08.012.

Wesseling J, Boers M, Viergever M A, Hilberdink W K, Lafeber F P, Dekker J, Bijlsma J W. Cohort profile: Cohort Hip and Cohort Knee (CHECK) study.

Int J Epidemiol 2014. doi: dyu177 [pii];10.1093/ije/dyu177 [doi].

Wesseling J, Bastick A N, ten Wolde S, Kloppenburg M, Lafeber F P, BiermaZeinstra S M, Bijlsma J W. Identifying trajectories of pain severity in early symptomatic knee osteoarthritis: a 5-year followup of the Cohort Hip and Cohort Knee (CHECK) Study. J Rheumatol 2015; 42(8): 1470-7. doi: jrheum.141036 [pii];10.3899/jrheum.141036 [doi].

534

White D K, Neogi T, Nguyen US, Niu J, Zhang Y. Trajectories of functional decline in knee osteoarthritis: the Osteoarthritis Initiative. Rheumatology 2016; 55(5): 801-8. doi: 10.1093/rheumatology/kev419.

Yusuf E, Bijsterbosch J, Slagboom P E, Kroon H M, Rosendaal F R, Huizinga T W, Kloppenburg M. Association between several clinical and radiological determinants with long-term clinical progression and good prognosis of lower limb osteoarthritis. PLoS One 2011; 6(10): e25426. doi: 10.1371/ journal.pone.0025426 [doi]; PONE-D-11-05738 [pii].

Zhang W, Doherty M, Arden N, Bannwarth B, Bijlsma J, Gunther KP, Hauselmann HJ, Herrero-Beaumont G, Jordan K, Kaklamanis P, Leeb B, Lequesne M, Lohmander S, Mazieres B, Martin-Mola E, Pavelka K, Pendleton A, Punzi L, Swoboda B, Varatojo R, Verbruggen G, ZimmermannGorska I, Dougados M, Therapeutics ESCfICSI. EULAR evidence based recommendations for the management of hip osteoarthritis: report of a task force of the EULAR Standing Committee for International Clinical Studies Including Therapeutics (ESCISIT). Ann Rheum Dis 2005; 64(5): 669-81. doi: 10.1136/ard.2004.028886.

Acta Orthopaedica 2018; 89 (5): 528â&#x20AC;&#x201C;534

Zhang W, Moskowitz RW, Nuki G, Abramson S, Altman RD, Arden N, Bierma-Zeinstra S, Brandt KD, Croft P, Doherty M, Dougados M, Hochberg M, Hunter DJ, Kwoh K, Lohmander LS, Tugwell P. OARSI recommendations for the management of hip and knee osteoarthritis, Part II: OARSI evidence-based, expert consensus guidelines. Osteoarthritis Cartilage/OARS, Osteoarthritis Research Society 2008; 16(2): 137-62. doi: 10.1016/j.joca.2007.12.013.

Acta Orthopaedica 2018; 89 (5): 535–540

535

Structural abnormalities detected by knee magnetic resonance imaging are common in middle-aged subjects with and without risk factors for osteoarthritis Jaanika KUMM 1, Aleksandra TURKIEWICZ 2, Fan ZHANG 2, and Martin ENGLUND 2,3 1 Department of Radiology, University of Tartu, Tartu, Estonia; 2 Lund University, Faculty of Medicine, Department of Clinical Sciences Lund, Orthopaedics, Clinical Epidemiology Unit, Lund, Sweden; 3 Clinical Epidemiology Research and Training Unit, Boston University School of Medicine, Boston, MA, USA Correspondence: jaanika.kumm@kliinikum.ee Submitted 2018-02-03. Accepted 2018-05-14.

Background and purpose — Few data are available regarding structural changes present in knees without radiographically evident osteoarthritis (OA). We evaluated the prevalence of findings suggestive of knee OA by magnetic resonance imaging (MRI) in middle-aged subjects without radiographic OA with or without OA risk factors.

Patients and methods — 340 subjects from the Osteoarthritis Initiative, aged 45–55 years (51% women) with Kellgren–Lawrence grade 0 in both knees, who had 3T knee MR images were eligible. 294 subjects had risk factors and 46 were without risk factors. MR images were assessed using the MOAKS scoring system.

Results — At least 1 MR-detected feature was found in 96% (283/294) of subjects with risk factors and in 87% (40/46) of those without. Cartilage damage (82%), bone marrow lesions (60%), osteophytes (45%), meniscal body extrusion (32%), and synovitis–effusion (29%) were the most common findings in subjects with risk factors, while cartilage damage (67%), osteophytes (46%), meniscal body extrusion (37%), and bone marrow lesions (35%) were most common in subjects without. The prevalence of any abnormality was higher in subjects with OA risk factors than in subjects without (prevalence ratio adjusted for age and sex 1.3 [95% CI 1.1–1.6]), so was prevalence of subchondral cysts and bone marrow lesions. MR-detected structural changes were more frequent in patellofemoral joints.

Interpretation — Our findings highlight the great challenge in distinguishing pathological features of early knee OA from what could be considered part of “normal ageing.”

Bone marrow lesions were more frequently found in subjects with multiple OA risk factors.

Recently, a high prevalence of magnetic resonance imaging (MRI) detected osteophytes, cartilage damage, and other OA-related pathological features have been identified in the tibiofemoral joint of knees without any evidence of OA on conventional radiographs (Hayes et al. 2005, Englund et al.

2008, Davies-Tuck et al. 2009, Javaid et al. 2010, Roemer et al. 2011, Guermazi et al. 2012, Hayashi et al. 2014, Sharma et al. 2014). This new knowledge has highlighted the difficulty in distinguishing true pathological features of early OA from those potentially related to what could be considered “normal ageing” of the joint.

There is an inevitable risk of interpreting incidental findings on MRI as related to early OA. It is therefore critical to gain further insights into the presence of anatomical alterations that could be considered suggestive of early OA in persons without radiographic evidence of knee OA to improve the understanding about the pre-radiographic stage of the disease.

Thus, our aim was to determine the prevalence of a broad spectrum of knee joint structural features on MRI in Osteoarthritis Initiative (OAI) participants (aged 45–55 years) who had radiographically normal knees (Kellgren and Lawrence grade 0 in both knees). Importantly, we stratified the study cohort into those having common risk factors for knee OA vs. those without risk factors. We report the structural changes in both the tibiofemoral and the patellofemoral joint and provide estimates of the prevalence of tibiofemoral and patellofemoral OA based on an MRI definition (Hunter et al. 2011a).

Patients and methods Study sample Subjects were sampled from the Osteoarthritis Initiative (OAI) database. Details for this multicenter longitudinal observational study are available for public access at http://www.oai. ucsf.edu/. Our criteria for sampling were: (i) age 45–55 years;

© 2018 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/4.0) DOI 10.1080/17453674.2018.1495164

536

(ii) no evidence of radiographic tibiofemoral OA of either knee at baseline based on the central readings (weight-bearing fixed-flexion posterior-anterior knee radiographs, Kellgren and Lawrence grade = 0); and (iii) available knee MRIs at baseline (and at 24, 48, and 72 months’ follow-up) (the followup images were not used for the current study). All subjects that fulfilled the above criteria were selected for the study. The subjects were drawn from the OAI incidence, progression, and “non-exposed” reference cohort. The subjects in the incidence and progression cohort had risk factors for OA like obesity, previous knee trauma and/or surgery, family history of total knee joint replacement, presence of Heberden’s nodes and repetitive knee bending but were given Kellgren–Lawrence grade 0 at baseline in both knees when assessed by the central reading facility in Boston. The reference cohort had no such risk factors, and no pain, aching, or stiffness in either knee in the past year before the baseline examination.

340 subjects met the above criteria. Among them were 294 subjects with OA risk factors (278 and 16 subjects from the incidence and progression cohort, respectively) and 46 individuals without OA risk factors (the reference cohort).

Knee joint radiography protocol The standing fixed-flexion knee radiography protocol was used to assess OA in the tibiofemoral joints. All participants had bilateral standing knee films obtained in posterior-anterior projection with knees flexed to 20–30 degrees and feet internally rotated 10 degrees. The degree of knee flexion and foot rotation was fixed using a plexiglass positioning frame (SynaFlexer, Synarc, Inc, San Francisco, CA, USA).

MRI protocol and assessment of knee joint structural abnormalities The MR images of subjects’ knees were obtained at the 4 clinical study centers using Siemens Trio 3.0-Tesla MRI scanners and quadrature transmit–receive extremity radio-frequency coils as previously detailed (Kumm et al. 2016).

One radiologist (JK), who was blinded to subject characteristics, interpreted the images for osteophytes, cartilage damage, BMLs including cysts, meniscal tears, synovitis–effusion, Hoffa synovitis, ligament abnormalities, and popliteal cysts and scored these features semi-quantitatively according to the Magnetic Resonance Imaging Osteoarthritis Knee Score (MOAKS) (Hunter et al. 2011b). In MOAKS, the knee joint is divided into 14 articular sub-regions for scoring articular cartilage, into 15 sub-regions for scoring bone marrow lesions (BMLs) and into 12 regions for scoring osteophytes. In the present study, cartilage damage, and bone marrow lesions were scored in 14 sub-regions, excluding the tibial subspinous subregion. Also, for the purpose of this study, we only report findings from right knees.

For further analysis, we also divided knee joints into three different articular compartments—patellofemoral, medial, and lateral tibiofemoral.

Acta Orthopaedica 2018; 89 (5): 535–540

Osteophytes were scored using a 0–3 scale (grade 0 = none; grade 1 = small; 2 = medium; 3 = large). Osteophytes were considered as present when MOAKS grade ≥ 1. Cartilage damage was graded for the size of cartilage loss as a percentage of surface area as related to the size of subregion surface; and for the percentage of loss in the subregion that is fullthickness loss using 0–3 scales. For the size of cartilage loss, the following grading system was used: 0 = none; 1 = < 10%; 2 = 10–75%; 3 = > 75% of sub-region of cartilage surface area; and for the percentage of full-thickness cartilage loss: 0 = none; 1 = < 10%; 2 = 10–75%; and 3 = > 75% of subregion of cartilage surface area. Cartilage damage graded as grade ≥ 1.0 was considered as having cartilage damage. Bone marrow lesions were scored as hyperintense signal within the trabecular bone on T2-weighted fat-suppressed images using a 0–3 scale: 0 = none; 1 = < 33%; 2 = 33–66%; and > 66% of subregional volume. Subchondral cysts were assessed as a percentage of bone marrow lesion that is cyst using a 0–3 scale: 0 = none; 1 = < 33%; 2 = 33–66%; and 3 = > 66% of the lesion. We considered bone marrow lesions and cysts present if graded ≥ 1. Meniscal damage (meniscal tears and/ or destruction) was assessed in the anterior horn, body, and posterior horn of the menisci. Meniscal tears were considered present if intra-meniscal increased signal communicated with the superior, inferior, or free edge of the meniscal surface on at least 2 consecutive images (De Smet et al. 2006). Another observer (FZ, orthopedic surgeon) measured meniscal body extrusion to the closest 0.1 mm using mid-coronal images and Sante DICOM Editor (64-bit) software (http://www.santesoft. com/win/sante-dicom-editor/sante-dicom-editor.html) (Zhang et al. 2016). Intra-observer reliability for meniscal measurements ranged from 0.6 to 1. We defined meniscal extrusion as extrusion = > 3mm. Synovitis-effusion was graded using a 0–3 scale: grade 0 = none; grade 1 = small; grade 2 = medium; grade 3 = large. Synovitis-effusion was present if graded ≥ 1. Hoffa synovitis and popliteal cysts were scored as absent/ present.

30 randomly selected subjects’ knees were reassessed. Intraobserver reliability measured as kappa coefficients was as follows: 0.4 (CI 0.1–0.8) for the detection of osteophytes, 0.7 (CI 0.4–1.0) for synovitis–effusion, 0.9 (CI 0.8–1.0) for meniscal damage, and 0.8 (CI 0.5–1.0) for popliteal cysts. All knees included in the reliability assessment were graded consistently in both readings for the presence of cartilage damage and bone marrow lesions (kappa = 1.0).

Statistics We calculated the prevalence of MRI features in right knees.

As the individuals with risk factors were oversampled in the OAI cohort, to avoid bias we present separate estimates for those with and without risk factors. The results are presented as percentage with 95% confidence intervals (CI) calculated according to the Agresti–Coull method (Agresti and Coull 1998). We also determined the prevalence of MRI-defined

Acta Orthopaedica 2018; 89 (5): 535–540

537

Table 1. Characteristics of the study participants. Values are frequency and (percentage) unless otherwise stated

Subjects with Reference OA risk factors cohort (n = 294) (n = 46)

Age, mean (SD) Women BMI a: < 25

25–29 ≥ 30 Family history of joint replacement b Heberden’s nodes, either hand b Engage in at least 1 knee bending activity b Knee symptoms (pain, aching, or stiffness), either knee b

50 (2.9) 146 (50) 104 (35) 114 (39) 70 (24) 53 (18) 55 (19) 203 (69)

50 (3.3) 27 (59) 26 (57) 20 (44) 0 (0) 0 (0) 2 (4) 19 (41)

273 (93)

0 (0)

a Data on BMI (body mass index) is missing for 6 individuals. b As assessed in eligibility interviews at baseline.

knee OA based on the method described by Hunter et al.

(2011a). To estimate the prevalence ratio among persons with and without OA risk factors we used a Poisson regression model with robust standard errors and general estimating equations to account for correlation between different features within the same knee. The model was adjusted for age and sex.

Additionally, we used a Poisson regression model with robust standard errors to estimate adjusted (for age and sex) prevalence ratio of the specific features.

Ethics, funding, and conflicts of interest The OAI has been approved by the institutional review boards for the University of California, San Francisco and the four OAI clinical centers (University of Pittsburgh; Ohio State University; University of Maryland, Baltimore; Memorial Hospital of Rhode Island). All subjects have given informed consent to participate in the study.

JK reports a grant from the Osteoarthritis Research Society International (young investigator travel award). ME reports grants from the Swedish Research Council, Kock Foundations, the Swedish Rheumatology Association, Österlund Foundation, the Faculty of Medicine, Lund University, Sweden, and Governmental Funding of Clinical Research within the National Health Service (ALF). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors declare no conflicts of interest.

Results The mean (SD) age of study subjects, n = 340 (173; 51% women), was 50 years (3.0) (range 45–55) years, and the mean (SD) BMI was 27 (4.4) (18–32) (Table 1).

Prevalence in subjects with risk factors for knee OA 283 (96%) subjects with knee OA risk factors but without radiographic knee OA had at least 1 abnormality present (Table 2). The most common findings were cartilage damage (82%, CI 77–86), bone marrow lesions (60%, CI 54–65), osteophytes (45%, CI 39–50), Hoffa synovitis (44%, CI 39–50), and subchondral cysts (41%, CI 35–46). MRI features like meniscal extrusions (23%, CI 19–28), synovitis–effusion (29%, CI 24–35), popliteal cysts (28%, CI 23–34), and meniscal damage (19%, CI 15–24) were also quite frequently encountered in these individuals with radiographically normal knees. The prevalence of ligamentous lesions was low (1%, CI 0–4).

In the joint and compartment-specific analysis, the prevalence of cartilage damage was quite similar in the patellofemoral and tibiofemoral joints (199/294; 68% and 189/294; 64%, respectively; Table 2). The prevalence of cartilage damage was similar also in the medial and lateral compartments of the tibiofemoral joint (120/294; 41% and 127/294; 43%, respec-

Table 2. Prevalence of MRI structural abnormalities in the right knees of the study subjects with knee OA risk factors. Values are frequency and (percentage)

BMI a

Overall Women Men < 25 25–29 ≥ 30 MRI feature (n = 294) (n = 146) (n = 148) (n = 104) (n = 114) (n = 70) Cartilage damage Tibiofemoral Tibiofemoral medial Tibiofemoral lateral Patellofemoral Osteophytes Tibiofemoral Tibiofemoral medial Tibiofemoral lateral Patellofemoral Bone marrow lesions Tibiofemoral Tibiofemoral medial Tibiofemoral lateral Patellofemoral Subchondral cysts Tibiofemoral Tibiofemoral medial Tibiofemoral lateral Patellofemoral Meniscal damage Meniscal extrusion Synovitis effusion Hoffa synovitis Popliteal cysts a Data

240 (82) 189 (64) 120 (41) 127 (43) 199 (68) 131 (45) 56 (19) 43 (15) 30 (10) 124 (42) 176 (60) 91 (31) 54 (18) 43 (15) 138 (47) 119 (40) 51 (17) 31 (11) 22 (8) 93 (32) 55 (19) 68 (23) 86 (29) 130 (44) 83 (28)

125 (86) 115 (78) 90 (62) 99 (67) 55 (38) 65 (44) 60 (41) 67 (45) 109 (75) 90 (61) 56 (38) 75 (51) 28 (19) 28 (19) 21 (14) 22 (15) 15 (10) 15 (10) 52 (36) 72 (49) 91 (62) 85 (57) 40 (27) 51 (34) 23 (16) 31 (21) 18 (12) 25 (17) 75 (51) 63 (43) 67 (46) 52 (35) 24 (16) 27 (18) 14 (10) 17 (11) 10 (7) 12 (8) 55 (38) 38 (26) 20 (14) 35 (24) 32 (22) 36 (24) 38 (26) 48 (32) 58 (40) 72 (49) 42 (29) 41 (28)

on BMI is missing for 6 individuals.

80 (77) 65 (63) 48 (46) 43 (41) 64 (62) 35 (34) 21 (20) 17 (16) 9 (9) 33 (32) 59 (57) 34 (33) 24 (23) 11 (11) 43 (41) 38 (37) 17 (16) 12 (12) 5 (5) 28 (27) 25 (24) 21 (20) 30 (29) 48 (46) 33 (32)

94 (82) 70 (61) 46 (40) 45 (39) 80 (70) 58 (51) 20 (18) 13 (11) 11 (10) 54 (47) 64 (56) 32 (28) 17 (15) 18 (16) 56 (49) 46 (40) 19 (17) 10 (9) 10 (9) 40 (35) 17 (15) 31 (27) 31 (27) 48 (42) 32 (28)

61 (87) 49 (70) 23 (33) 34 (49) 52 (74) 36 (51) 14 (20) 12 (17) 9 (13) 35 (50) 49 (70) 22 (31) 13 (19) 11 (16) 38 (54) 32 (46) 13 (19) 9 (13) 5 (7) 24 (34) 9 (13) 14 (20) 23 (33) 32 (46) 16 (23)

538

Acta Orthopaedica 2018; 89 (5): 535–540

Table 3. Prevalence of MRI structural abnormalities in the right knees of the study subjects without knee OA risk factors (reference cohort). Values are frequency and (percentage)

BMI

Overall Women Men < 25 25–29 MRI feature (n = 46) (n = 27) (n = 19) (n = 26) (n = 20) Cartilage damage Tibiofemoral Tibiofemoral medial Tibiofemoral lateral Patellofemoral Osteophytes Tibiofemoral Tibiofemoral medial Tibiofemoral lateral Patellofemoral Bone marrow lesions Tibiofemoral Tibiofemoral medial Tibiofemoral lateral Patellofemoral Subchondral cysts Tibiofemoral Tibiofemoral medial Tibiofemoral lateral Patellofemoral Meniscal damage Meniscal extrusion Synovitis effusion Hoffa synovitis Popliteal cysts

31 (67) 21 (46) 15 (33) 11 (24) 26 (57) 21 (46) 5 (11) 4 (9) 1 (2) 20 (43) 16 (35) 5 (11) 2 (4) 3 (7) 14 (30) 9 (20) 3 (7) 1 (2) 2 (4) 6 (13) 5 (11) 10 (22) 8 (17) 14 (30) 12 (26)

21 (78) 15 (56) 12 (44) 8 (30) 19 (70) 14 (52) 4 (15) 3 (11) 1 (4) 13 (48) 9 (33) 3 (11) 1 (4) 2 (7) 8 (30) 7 (26) 3 (11) 1 (4) 2 (7) 4 (15) 2 (7) 4 (15) 7 (26) 8 (30) 7 (26)

10 (53) 6 (32) 3 (16) 3 (16) 7 (37) 7 (37) 1 (5) 1 (5) 0 (0) 7 (37) 7 (37) 2 (11) 1 (5) 1 (5) 6 (32) 2 (11) 0 (0) 0 (0) 0 (0) 2 (11) 3 (16) 6 (32) 1 (5) 6 (32) 5 (26)

18 (69) 13 (50) 10 (38) 6 (23) 16 (62) 11 (42) 4 (15) 3 (12) 1 (4) 10 (38) 9 (35) 3 (12) 1 (4) 2 (8) 8 (31) 7 (27) 3 (12) 1 (4) 2 (8) 4 (15) 2 (8) 6 (23) 4 (15) 10 (38) 7 (27)

13 (65) 8 (40) 5 (25) 5 (25) 10 (50) 10 (50) 1 (5) 1 (5) 0 (0) 10 (50) 7 (35) 2 (10) 1 (5) 1 (5) 6 (30) 2 (10) 0 (0) 0 (0) 0 (0) 2 (10) 3 (15) 4 (20) 4 (20) 4 (20) 5 (25)

tively). On the other hand, osteophytes, bone marrow lesions, and subchondral cysts were observed more frequently in the patellofemoral joint.

Prevalence in subjects without knee OA risk factors (reference cohort) 40 (87%) knees of subjects from the reference cohort had at least 1 type of structural abnormality. The most common MRI findings were cartilage damage (67%, CI 53–79), osteophytes (46%, CI 32–60), Hoffa synovitis (30%, CI 19–45), meniscal extrusions (22%, CI 12–36), and bone marrow lesions (35%, CI 23–49).

Cartilage damage, osteophytes, and bone marrow lesions were again observed more often in the patellofemoral joints than in the tibiofemoral compartments; the difference was biggest in the case of osteophytes (Table 3).

Differences in the prevalence between the subjects with and without knee OA risk factors The prevalence of any abnormality was higher (prevalence ratio adjusted for age, sex, and BMI 1.3 [CI 1.1–1.6]) in subjects with knee OA risk factors than in subjects without.

In this study sample, with the exception of osteophytes and meniscal extrusions, all other MRI structural abnormalities were observed more often in subjects who had knee OA risk factors compared with those without. The prevalence ratios

(CI; adjusted for age and sex) were for subchondral cysts 2.1 (1.2–3.8); bone marrow lesions 1.7 (1.2–2.6); meniscal tears 1.6 (0.7–3.8); synovitis–effusion 1.7 (0.9–3.2); Hoffa synovitis 1.4 (0.9–2.2); cartilage lesions 1.2 (1.0–1.5); popliteal cysts 1.1 (0.6–1.8); meniscal extrusion 1.1 (0.6–1.9); and osteophytes 1.0 (0.7–1.4) in subjects with knee OA risk factors compared with those without.

The prevalence of MRI-defined OA in the tibiofemoral joint was 23% (CI 19–28) in subjects with OA risk factors and 9% (CI 3–21) in those without. The corresponding proportions for the patellofemoral joint were 35% (CI 30–41) and 35% (CI 23–49), respectively.

The prevalence of MRI-defined tibiofemoral OA tended to be higher (prevalence ratio adjusted for age and sex 2.6 [CI 1.0–6.9]) in subjects with knee OA risk factors than in subjects without. The corresponding prevalence ratio for the patellofemoral joint was 1.0 (CI 0.7–1.5).

Discussion In contrast to several prior studies we examined the prevalence of knee joint structural features indicative of early knee OA on MRI in subjects aged 45 to 55 years without the evidence of knee OA on conventional radiographs (Hayes et al. 2005, Englund et al. 2008, Davies-Tuck et al. 2009, Javaid et al.

2010, Roemer et al. 2011, Guermazi et al. 2012, Hayashi et al. 2014, Sharma et al. 2014). At least 1 MRI-detected feature was found in 96% of the subjects with knee OA risk factors and in 87% of those without OA risk factors.

Cartilage damage was the most common MRI feature (82%), followed by BMLs (60%) and osteophytes (45%). Cartilage damage was equally highly prevalent in tibiofemoral and patellofemoral joints (64% and 68%, respectively), whereas BMLs and osteophytes were more common in patellofemoral joints in that age group (bone marrow lesions 47% and 31%; osteophytes 42% and 19% in patellofemoral and tibiofemoral joints, respectively). The prevalence of any abnormality was 32% higher in subjects with knee OA risk factors than in subjects without. Subchondral cysts, BMLs, meniscal tears, synovitis–effusion, Hoffa synovitis, cartilage damage, and popliteal cysts were observed more often in subjects with knee OA risk factors compared with the reference cohort. However, these differences reached statistical significance only in case of BMLs and subchondral cysts.

In the majority of prior knee OA studies the MRI features indicative of OA have been examined in older subjects with already radiographically evident knee OA (Raynauld et al.

2006, Bruyere et al. 2007, Madan-Sharma et al. 2008, Cibere et al. 2011, Crema et al. 2014, de Lange-Brokaar et al. 2016).

Much less is known about the pre-radiographic stage of the disease when the basic radiographic features of knee OA— joint space narrowing and osteophytes—are not yet visible.

In the pre-radiographic stage of OA, the structural features

Acta Orthopaedica 2018; 89 (5): 535–540

indicative of OA can be assessed by MRI as it has a greater sensitivity than radiography for the detection of OA-related bone and soft tissue changes (Felson et al. 1987, Bruyere et al.

2007). Still, there are concerns regarding its specificity.

According to the prior population-based observational study (Framingham Osteoarthritis Study) on 710 subjects with a mean age of 62 years and without radiographic knee OA, MRI detected lesions indicative of tibiofemoral OA were found in most subjects, regardless of knee pain. Among them 74% had osteophytes, 69% had cartilage damage, and 52% had BMLs (Guermazi et al. 2012). In another cohort study (OAI) on 849 subjects with a mean age of 60 years and without radiographic knee OA, 76% had cartilage damage, 61% bone marrow lesions, 21% meniscal tears, and 14% meniscal extrusion (Sharma et al. 2014). According to the Multicenter Osteoarthritis Study (MOST) on subjects aged 50–79 years without radiographic knee changes but at high risk of OA, again a high percentage of structural abnormalities was found by MRI (tibiofemoral or patellofemoral cartilage damage in 67–81%, bone marrow lesions in 55–75%, and osteophytes in as many as 99–100% of studied individuals) (Javaid et al. 2010). MRI features indicative of OA in our sample of radiographically normal subjects with a mean age of at least 10 years younger than in the prior studies were similarly highly prevalent.

We found that MRI-detected cartilage damage, osteophytes, BMLs, and subchondral cysts were more frequent in the patellofemoral than the tibiofemoral joint. Our findings are in accordance with the recent study by Lankhorst et al. who suggested that OA is more likely to start in the patellofemoral joint and then progress to the combined involvement of both the patellofemoral and tibiofemoral joint (Lankhorst et al. 2017). In another study the authors reported that isolated patellofemoral abnormalities were more common than isolated tibiofemoral abnormalities using MRI. Moreover, when mixed disease was present, the patellofemoral joint had more severe pathologies (Stefanik et al. 2013).

It has been reported that BMLs are strongly associated with knee pain (Felson et al. 2001, Yusuf et al. 2011) and might precede the development of radiographic knee OA (Javaid et al. 2010). In our study, BMLs were indeed significantly more prevalent in subjects with risk factors for knee OA compared with those without risk factors (60% and 35%, respectively), especially in the patellofemoral joints.

Our study has limitations. Only right knees were included in the analyses. Still, it is unlikely to affect our overall conclusion. Only central OAI readings for posteroanterior tibiofemoral joint radiographs were available at baseline. Further, we cannot exclude the potential for selection bias by including only subjects with MRI scans at all 4 time points, but is likely to affect both study cohorts (with and without risk factors) to a similar degree. Also, our findings cannot be generalized to the adult population younger than 45 years.

In summary, we found that MRI features indicative of knee OA are common in middle-aged subjects without radiographic

539

knee OA. This was true irrespective of whether the knee OA risk factors were present or not. Subchondral cysts and BMLs were among the most frequent findings in subjects with multiple OA risk factors. The clinical importance of MRI findings related to OA remains vague. It is unclear to what extent these frequently encountered MRI findings represent early knee OA or whether they may perhaps be considered as part of “normal ageing” of the joint.

JK read the knee MRIs, participated in the analysis of data, interpreted the results, and drafted the first version of the manuscript. FZ measured meniscus extrusion, interpreted the results, and revised the manuscript for important intellectual content. AT analyzed the data, interpreted the results, and revised the manuscript. ME conceived the study, interpreted the results, and revised the manuscript. All authors approved the final version of the manuscript.

The authors would like to express their sincere gratitude to the OAI for the free use of public access data. They would like to acknowledge support from the Osteoarthritis Research Society International and the scholarship which made it possible for Jaanika Kumm to come to Lund University to perform the study.

Acta thanks Kirill Gromov for help with peer review of this study.

Agresti A, Coull B A. Approximate is better than “exact” for interval estimation of binomial proportions. American Statistician 1998; 52(2): 119-26.

Bruyere O, Genant H, Kothari M, Zaim S, White D, Peterfy C, Burlet N, Richy F, Ethgen D, Montague T, Dabrowski C, Reginster J-Y. Longitudinal study of magnetic resonance imaging and standard X-rays to assess disease progression in osteoarthritis. Osteoarthritis Cartilage 2007; 15: 98-103.

Cibere J, Sayre E C, Guermazi A, Nicolaou S, Kopec J A, Esdaile J M, Thorne A, Singer J, Wong H. Natural history of cartilage damage and osteoarthritis progression on magnetic resonance imaging in a population-based cohort with knee pain. Osteoarthritis Cartilage 2011; 19: 683-8.

Crema M D, Nevitt M C, Guermazi A, Felson D T, Wang K, Lynch J A, Marra M D, Torner J, Lewis C E, Roemer F W. Progression of cartilage damage and meniscal pathology over 30 months is associated with an increase in radiographic tibiofemoral joint space narrowing in persons with knee OA: the MOST study. Osteoarthritis Cartilage 2014; 22: 1743-7.

Davies-Tuck M L, Wluka A E, Wang Y, English D R, Giles G G, Cicuttini F. The natural history of bone marrow lesions in community-based adults with no clinical knee osteoarthritis. Ann Rheum Dis 2009; 68: 904-8. de Lange-Brokaar B J E, Bijsterbosch J, Kornaat P R, Yusuf E, Ioan-Facsinay A, Zuurmond A-M, Kroon H M, Meulenbelt I, Bloem J L, Kloppenburg M. Radiographic progression of knee osteoarthritis in association with MRI abnormalities in both the patellofemoral and tibiofemoral joint. Osteoarthritis Cartilage 2016; 24: 473-9.

De Smet A A, Tuite M J. Use of the “two-slice-touch” rule for the MRI diagnosis of meniscal tears. Am J Roentgenol 2006; 187(4): 911-14.

Englund M, Guermazi A, Gale D, Hunter D J, Aliabadi P, Clancy M, Felson D T. Incidental meniscal findings on knee MRI in middle-aged and elderly persons. N Engl J Med 2008; 359: 1108-15.

Felson D T, Naimark A, Anderson J, Kazis L, Castelli W, Meenan R F. The prevalence of knee osteoarthritis in the elderly. The Framingham Osteoarthritis Study. Arthritis Rheum 1987; 30: 914-18.

Felson D T, Chaisson C E, Hill C L, Totterman S M, Gale M E, Skinner K M, Kazis L, Gale D R. The association of bone marrow lesions with pain in knee osteoarthritis. Ann Intern Med 2001; 134: 541-9.

540

Guermazi A, Niu J, Hayashi D, Roemer F W, Englund M, Neogi T, Aliabadi P, McLennan C E, Felson D T. Prevalence of abnormalities in knees detected

by MRI in adults without knee osteoarthritis: population based observational study (Framingham Osteoarthritis Study). BMJ 2012; 345: e5339.

Hayashi D, Felson D T, Niu J, Hunter D J, Roemer F W, Aliabadi P, Guermazi A. Pre-radiographic osteoarthritic changes are highly prevalent in the medial patella and medial posterior femur in older persons: Framingham OA Study. Osteoarthritis Cartilage 2014; 22:7 6-83.

Hayes C W, Jamadar D A, Welch G W, Jannausch M L, Lachance L L, Capul D C, Sowers M R. Osteoarthritis of the knee: comparison of MR imaging findings with radiographic severity measurements and pain in middle-aged women. Radiology 2005; 237: 998-1007.

Hunter D J, Arden N, Conaghan P G, Eckstein F, Gold G, Grainger A, Guermazi A, Harvey W, Jones G, Hellio Le Graverand M P, Laredo J D, Lo G, Losina E, Mosher T J, Roemer F, Zhang W, OARSI OA Imaging Working Group. Definition of osteoarthritis on MRI: the results of a Delphi exercise. Osteoarthritis Cartilage 2011a; 19(8): 963-9.

Hunter D J, Guermazi A, Lo G H, Grainger A J, Conaghan P G, Boudreau R M, Roemer F W. Evolution of semi-quantitative whole joint assessment of knee OA: MOAKS (MRI Osteoarthritis Knee Score). Osteoarthritis Cartilage 2011b; 19: 990-1002.

Javaid M K, Lynch J A, Tolstykh I, Guermazi A, Roemer F, Aliabadi P, McCulloch C, Curtis J, Felson D, Lane N E, Torner J, Nevitt M. Preradiographic MRI findings are associated with onset of knee symptoms: the most study. Osteoarthritis Cartilage 2010; 18: 323-8.

Kumm J, Roemer F W, Guermazi A, Turkiewicz A, Englund M. Natural history of intrameniscal signal intensity on knee MR images: six years of data from the Osteoarthritis Initiative. Radiology 2016; 278(1): 164-71.

Lankhorst N E, Damen J, Oei E H, Verhaar J A N, Kloppenburg M, BiermaZeinstra S M A, van Middelkoop M. Incidence, prevalence, natural course and prognosis of patellofemoral osteoarthritis: the Cohort Hip and Cohort Knee study. Osteoarthritis Cartilage 2017; 25(5): 647-53.

Acta Orthopaedica 2018; 89 (5): 535â&#x20AC;&#x201C;540

Madan-Sharma R, Kloppenburg M, Kornaat P R, Botha-Scheepers S A, Le Graverand M P, Bloem J L, Watt I. Do MRI features at baseline predict radiographic joint space narrowing in the medial compartment of the osteoarthritic knee 2 years later? Skeletal Radiol 2008; 37: 805-11.

Raynauld J P, Martel-Pelletier J, Berthiaume M J, Beaudoin G, Choquette D, Haraoui B, Tannenbaum H, Meyer J M, Beary J F, Cline G A, Pelletier J P. Long term evaluation of disease progression through the quantitative magnetic resonance imaging of symptomatic knee osteoarthritis patients: correlation with clinical symptoms and radiographic changes. Arthritis Res Ther 2006; 8: R21.

Roemer F W, Guermazi A, Felson D T, Niu J, Nevitt M C, Crema M D, Lynch J A, Lewis C E, Torner J, Zhang Y. Presence of MRI-detected joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at 30-month follow-up: the MOST study. Ann Rheum Dis 2011; 70: 1804-9.

Sharma L, Chmiel J S, Almagor O, Dunlop D, Guermazi A, Bathon J M, Eaton C B, Hochberg M C, Jackson R D, Kwoh C K, Mysiw W J, Crema M D, Roemer F W, Nevitt M C. Significance of preradiographic magnetic resonance imaging lesions in persons at increased risk of knee osteoarthritis. Arthritis Rheum 2014; 66: 1811-19.

Stefanik J J, Niu J, Gross K D, Roemer F W, Guermazi A, Felson D T. Using magnetic resonance imaging to determine the compartmental prevalence of knee joint structural damage. Osteoarthritis Cartilage 2013; 21: 695-9.

Yusuf E, Kortekaas M C, Watt I, Huizinga T W, Kloppenburg M. Do knee abnormalities visualised on MRI explain knee pain in knee osteoarthritis?

A systematic review. Ann Rheum Dis 2011; 70: 60-7 Zhang F, Kumm J, Svensson F, Turkiewicz A, Frobell R, Englund M. Risk factors for meniscal body extrusion on MRI in subjects free of radiographic knee osteoarthritis: longitudinal data from the Osteoarthritis Initiative.

Osteoarthritis Cartilage 2016; 24: 801-6.

Acta Orthopaedica 2018; 89 (5): 541–547

541

Minimal important change values for the Oxford Knee Score and the Forgotten Joint Score at 1 year after total knee replacement Lina H INGELSRUD 1,2, Ewa M ROOS 2, Berend TERLUIN 3, Kirill GROMOV 1, Henrik HUSTED 1, and Anders TROELSEN 1 1 Department

of Orthopaedic Surgery, Copenhagen University Hospital Hvidovre, Copenhagen, Denmark; 2 Department of Sports Science and Clinical Biomechanics, University of Southern Denmark, Odense, Denmark; 3 Department of General Practice and Elderly Care Medicine, VU University Medical Center, Amsterdam, Netherlands Correspondence: lina.holm.ingelsrud@regionh.dk Submitted 2018-01-17. Accepted 2018-05-03.

Background and purpose — Interpreting changes in Oxford Knee Score (OKS) and Forgotten Joint Score (FJS) following total knee replacement (TKR) is challenged by the lack of methodologically rigorous methods to estimate minimal important change (MIC) values. We determined MIC values by predictive modeling for the OKS and FJS in patients undergoing primary TKR. Patients and methods — We conducted a prospective cohort study in patients undergoing TKR between January 2015 and July 2016. OKS and FJS were completed preoperatively and at 1 year postoperatively, accompanied by a 7-point anchor question ranging from “better, an important improvement” to “worse, an important worsening.” MIC improvement values were defined with the predictive modeling approach based on logistic regression, with patients’ decisions on important improvement as dependent variable and change in OKS/FJS as independent variable. Furthermore, the MICs were adjusted for high proportions of improved patients.

Results — 333/496 (67.1%) patients with a median age of 69 years (61% female) had complete data for OKS, FJS, and anchor questions at 1 year postoperatively. 85% were importantly improved. Spearman’s correlations between the anchor and the change score were 0.56 for OKS, and 0.61 for FJS. Adjusted predictive MIC values (95% CI) for improvement were 8 (6–9) for OKS and 14 (10–18) for FJS. Interpretation — The MIC value of 8 for OKS and 14 for FJS corresponds to minimal improvements that the average patient finds important and aids in our understanding of whether improvements after TKR are clinically relevant.

Patient-reported outcome measures (PROM) are increasingly advocated as primary outcome measures in clinical trials, as well as quality of care assessment in arthroplasty registries (FDA and HHS 2009, Rolfson et al. 2016). The Oxford Knee Score (OKS) and Forgotten Joint Score (FJS) are 2 such commonly used outcome measures developed for patients undergoing TKR (Giesinger et al. 2014, Harris et al. 2016, Thomsen et al. 2016). However, interpreting whether changes in OKS and FJS scores are clinically meaningful is challenging because statistically significant improvements are not necessarily clinically meaningful (King 2011).

The concept of minimal important change (MIC) is defined as the smallest change in a PROM considered important by a notional average patient (Terluin et al. 2015). MIC values for PROMs may differ depending on the patient population, intervention, follow-up time etc. It is therefore necessary to determine context-specific MIC threshold values for specific PROMs that may improve the translation of PROM scores into clinical relevance (King 2011).

No previous studies have estimated MIC values for the FJS. MIC values for OKS ranging from 7 to 9 at 6 months following a TKR, and 4.3 to 5 at 12 months follow-up have been suggested (Clement et al. 2014, Beard et al. 2015). However, the applicability of the reported MIC values depends on the definition of MIC and the methodological approach used to determine the MIC values. As different methodological approaches yield different MIC values (Terluin et al. 2015), additional studies are needed to further establish MIC thresholds for the OKS. We therefore determined MIC values for the OKS and FJS at 1 year after a TKR. These MIC values are intended for interpretation of within-group mean improvements and for use as responder thresholds when interpreting whether improvements differ between intervention groups, including the surveillance of treatment outcome in registries and clinical studies as a supplement to implant survivorship.

© 2018 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/4.0) DOI 10.1080/17453674.2018.1480739

542

Patients and methods Study design and setting The study is a prospective observational cohort study using data from 1 Danish hospital’s local arthroplasty registry. From March 2013, all patients scheduled for joint replacement at the hospital were asked to complete an electronic questionnaire at their preoperative visit at the hospital. At 1 year postoperatively, patients received an email with a link to an electronic follow-up questionnaire. If the questionnaire was not completed after 2 reminder emails with a 2-week interval, a paper version of the questionnaire was sent by postal mail.

Participants Data from all patients with knee replacement surgery performed between January 1, 2015 and July 31, 2016 were extracted from the registry. Patients who had revision surgery and unicompartmental arthroplasty were excluded. To avoid multiple observations on patients having had surgery performed on both knees within the data extraction period, or simultaneous bilateral surgery, we randomly selected 1 of the observations to be included in the analyses.

Questionnaires Preoperative and 1-year postoperative questionnaires that were extracted from the registry included the OKS, FJS, and an anchor question. The OKS and FJS were developed to evaluate patient-relevant outcomes after TKR. Previous studies have established adequate validity, reliability, and responsiveness characteristics for the OKS and FJS in the TKR population (Behrend et al. 2012, Harris et al. 2016, Thomsen et al. 2016, Hamilton et al. 2017). The OKS includes 12 items about knee pain and function that are summed to a total score of 0–48 (worst–best) (Murray et al. 2007). The FJS includes 12 items about the patient’s knee awareness in different daily life activities. The original version that was developed in 2012 (Behrend et al. 2012), and cross-culturally adapted into Danish, asked patients “Are you aware of your artificial knee…” (Thomsen et al. 2016). The FJS was later shown to be responsive in measuring change in knee awareness from before to after a TKR (Thienpont et al. 2016, Hamilton et al. 2017). Therefore, to enable measurement of change from before to after surgery in our study, the question in the preoperative form was adapted to “Are you aware of your knee…”, in accordance with the version used in the British population (Hamilton et al. 2017). A total score of 0 to 100 is calculated, with higher scores reflecting higher ability to forget the knee joint (Behrend et al. 2012). Missing items in OKS and FJS were handled in accordance with each respective user-guide (Murray et al. 2007, Behrend et al. 2012).

Furthermore, at 1 year postoperatively, the patients’ experienced degree and importance of change were obtained by asking the anchor question: “How are your knee problems

Acta Orthopaedica 2018; 89 (5): 541–547

Table 1. Response options to minimal important change anchor question and classification into importantly improved or not Classification of importance / Response options Importantly improved Better, an important improvement Somewhat better, but enough to be an important improvement Not importantly improved Very small change, not enough to be an important improvement About the same Very small change, not enough to be an important deterioration Somewhat worse, but enough to be an important deterioration Worse, an important deterioration

now compared to prior to your operation?” Patients responded on a 7-point scale ranging from “better, an important improvement” to “worse, an important deterioration” (Jaeschke et al.

1989). Patients were classified as being importantly improved when answering “better, an important improvement” or “somewhat better, but enough to be an important improvement” (Table 1).

Statistics Patient demographics were presented as median (interquartile range (IQR)) for non-normally distributed continuous variables and number (proportion) for categorical variables. OKS and FJS change score distributions across anchor response categories were investigated with boxplots. The association between the OKS and FJS change score and anchor responses were investigated with Spearman’s correlation. R version 3.4.0 (https://www.r-project.org/) was used for all analyses.

Anchor-based MIC approach MIC values were estimated using anchor-based approaches that involve anchoring of the OKS and FJS change scores to the anchor question responses. Several anchor-based methods are advocated and there is no consensus on which method provides the best estimate of MIC values (King 2011). Due to its methodological advantages, our primary method for estimating MIC values is the newly described predictive modeling method (MICpred) (Terluin et al. 2015). This method is more precise than the commonly used receiver operating characteristics (ROC) method and less dependent on the correlation between the PROM score and the anchor responses (Terluin et al. 2015). Furthermore, it enables adjustment for the bias that incurs when the proportion of improved patients is smaller or larger than 50%. This adjusted MICpred has been shown to equal the mean of the latent individual MICs. Hence, the adjusted MICpred represents the amount of change that the notional average patient considers to be minimally important. The predictive modeling method is based on a logistic regression, using the dichotomized anchor response (importantly improved or not) as dependent variable and the PROM change score as independent variable. The change in PROM

Acta Orthopaedica 2018; 89 (5): 541–547

543

Table 2. Patient preoperative demographics. Values are median (interquartile range), unless otherwise stated

Eligible for extraction 560 surgeries (519 patients) Excluded: – revision surgeries, 21 surgeries (18 patients) – unicompartmental arthroplasty, 6 surgeries (5 patients) 533 surgeries (496 patients) Randomly excluded surgeries (n = 37): – bilateral surgery, 19 – 2 surgeries within the data extraction period, 18

Age Female, n (%) Body mass index Oxford Knee Score Forgotten Joint Score

Patients with Patients with complete data incomplete data n = 333 n = 163 p-value a 69 (61–73) 203 (61) 29 (26–33) 23 (17–27) 14 (6–27)

68 (61–75) 119 (73) 30 (26–33) 21 (15–25) 14 (7–23)

0.7 0.01 0.8 0.01 0.6

a Wilcoxon

496 surgeries (496 patients)

Signed Rank test for continuous variables and chi-square test for dichotomous variables.

Excluded (n = 163 patients): – missing 1-year form, 139 – missing anchor, 24 Patients with complete data for either OKS or FJS 333 (67%)

Complete OKS data 330 (67%) Complete FJS data 328 (66%)

Figure 1. Flow chart.

that corresponds to a Likelihood Ratio of 1 is estimated as the MIC value. With a likelihood ratio of 1, the posttest odds of being importantly improved are the same as the pretest odds of improvement. The adjustment for the large proportion of improved patients was performed with the equation MICadjusted = MICpred – (0.090 + 0.103 × Cor) × SDchange ×

log-odds(imp).

Cor is the point biserial correlation between the PROM change score and the anchor, SDchange is the SD of the change score, and log-odds(imp) is the natural logarithm of (proportion improved/[1 – proportion improved]). Bootstrap replications (n = 1,000) were used to determine 95% confidence intervals (CI) for adjusted MICpred (Terluin et al. 2017).

To enable comparison with other commonly described methods, we also estimated MIC values with the mean change (MICMeanChange) and ROC (MICROC) methods. With the mean change method the MIC value corresponds to the mean change in PROM in the subgroup of patients responding “somewhat better, enough to be importantly improved” (Jaeschke et al. 1989). We calculated 95% CI for the ­MICMeanChange as Meanchange ± 1.96 (SDchange /(√n)), with n and SDchange corresponding to the subgroup “somewhat better.” With the ROC method, the MIC value is the change in PROM score that with the least degree of misclassification, according to the Youden criterion, discriminates patients from being importantly improved or not. Bootstrap replications (n = 1,000) were used to determine 95% CI for MICROC (Terwee et al. 2010).

Baseline dependency To investigate whether preoperative severity impacted on MICpred values, an interaction term between the preopera-

tive PROM score and change in PROM score was included in each respective logistic regression model (Terluin et al.

2015). Effect modification of MICpred was considered present if interaction terms had p-values < 0.05.

Ethics, funding, and potential conflicts of interest The local arthroplasty registry was approved by the national data protection agency (Journal number HVH-2012-048).

In Denmark, approval from the ethical committee is not required for register-based studies involving only questionnaire data. The study was conducted in accordance with the WMA Declaration of Helsinki. The study was fully funded by the orthopedic department at the hospital. The authors declare that there are no potential conflicts of interest in relation to this study.

Results Participants After excluding patients who had undergone revision surgery or unicompartmental arthroplasty, 496 unique patients were registered with a primary TKA, of which 139 were excluded because they had not answered the 1-year follow-up form.

Complete data for anchor questions and for either the OKS or FJS were available for 333/496 (67%) patients (Figure 1).

Preoperative patient characteristics Patients with complete data had a median (IQR) age of 69 (61–73) years and 61% were female. These patients differed from the patients with incomplete data, with 12% fewer females (p = 0.01) and a 2-point higher median preoperative OKS (p = 0.01). Other preoperative characteristics were comparable (Table 2).

Descriptive data The overall percentage of patients reporting important improvements was 85%, while 8% reported being either unchanged, or perceiving too small improvement or deterio-

544

Acta Orthopaedica 2018; 89 (5): 541–547

Forgotten Joint Score change (%)

Oxford Knee Score change (%) 50

100

40 30

50

20 10

0

0 –10

–50

–20 better n = 226 (69%)

somewhat very small better improvement n = 54 n = 22 (16%) (7%)

same n=3 (1%)

very small somewhat deterioration worse n=1 n = 10 (0.3%) (3%)

worse n = 14 (4%)

better n = 225 (69%)

somewhat very small better improvement n = 52 n = 22 (16%) (7%)

same n=3 (1%)

very small somewhat deterioration worse n=1 n = 10 (0.3%) (3%)

worse n = 15 (5%)

Figure 2. OKS and FJS change scores by anchor questions response categories ranging from “better, an important improvement” to “worse, an important deterioration.” Horizontal bars present the median, the box the interquartile range, and the whiskers the maximum and minimum scores.

Table 3. MIC improvement values determined with the predictive modeling approach adjusted for the proportions of improved patients, the mean change method and the ROC method

Oxford Knee Score Forgotten Joint Score

Predictive Mean change modeling approach a method ROC method b MICpred (CI ) MICMeanChange (CI c) MICROC (CI b) Sensitivity Specificity 8 (6–9) 14 (10–18)

a adjusted for the proportions improved. b 95% confidence intervals (CI) calculated c 95% confidence intervals (CI) calculated

to the subgroup “somewhat better.”

10 (8–1) 23 (17–28)

9 (6–15) 17(11–29)

0.83 0.85

0.74 0.88

using 1,000 bootstrap replications, reported as 0.025–0.975 quantiles. as Meanchange ± 1.96 (SDchange/(√n)), where n and SDchange correspond

ration to be of importance, and the final 8% reported being importantly deteriorated. OKS and FJS change scores for each of the anchor response categories are presented in Figure 2.

MIC improvement values The correlations between the anchor question and the change in OKS and FJS were 0.56 and 0.61, respectively.

MICpred improvement values adjusted for the large proportions of improved patients were 8 (CI 6–9) for OKS and 14 (CI 10–18) for FJS. Unadjusted MIC values determined with the mean change method and the ROC method were higher with wider 95% CI than the adjusted MICpred values (Table 3).

Baseline dependency Interaction terms between preoperative and change in OKS and FJS scores were statistically not significant (p = 0.1 and p = 0.9, respectively), suggesting no baseline dependency of MICpred values.

Discussion Summary of findings In this prospective single-center study we propose estimates for meaningful improvement in OKS and FJS at one year after primary TKR. The MICpred values of 8 for the OKS and 14 for the FJS reflect the smallest improvement needed to be considered important by a notional average patient one year after a TKR. The majority of patients (85%) experienced important improvements in their knee problems, while 8% considered themselves to be unchanged and 8% to have importantly deteriorated after TKR. Relation to previous studies No previous studies have determined MIC values for the FJS with which we can compare our findings. Our MIC value for the OKS lies in the range of those proposed by Beard et al. (2015). They used data on 94,015 patients from the NHS PROMS data set, and found that 6 months after surgery

Acta Orthopaedica 2018; 89 (5): 541â&#x20AC;&#x201C;547

changes larger than 6.5 calculated with the ROC method and 9.2 points calculated with the mean change method were considered clinically meaningful. The similarity in MIC estimates between our studies suggests that the thresholds for important improvements may not vary much between 6 and 12 months after TKR. Conversely, OKS MIC values proposed by Clement et al. (2014) were smaller, ranging from 4.3 to 5 at 12 months after a TKR. Their anchor question, a 5-point Likert scale of satisfaction with functional improvement and pain relief, and statistical approach, a simple linear regression, differed from our study, which may explain the discrepancy in MIC estimates (Clement et al. 2014). Although our proposed MIC values are in the same range as those from the study by Beard et al. (2015), methodological differences may explain the variation in MIC estimates that have been found (Terwee et al. 2010).

MIC estimations vary with methodology In accordance with a previous study, we found that MIC values differ with methodology used (Ingelsrud et al. 2018). For both the OKS and the FJS we found the largest MIC values with the mean change method, and after adjusting for the large proportion of improved patients the smallest MIC values were found with the predictive modeling method.

Although the mean change method is appealing because it is intuitive and easily calculated, it is criticized because only data from a subgroup of the population sample are used (Terwee et al. 2010, King 2011). Furthermore, MICMeanChange values are not considered appropriate as responder criteria because, assuming normal distribution of scores, only half of the patients in the subgroup used to calculate the threshold value would be correctly classified as being importantly improved (McLeod et al. 2011).

Conversely, with the ROC method, all data points are used in the MIC estimation, but simulation studies have shown it to be less precise and more susceptible to errors than the predictive modeling method (Terluin et al. 2015). As an example, the optimal ROC cut-off of 8.5 in our study was associated with smaller degrees of misclassification (specificity: 0.74 and sensitivity: 0.83) than the cut-off of 6.5 found by Beard et al.

(2015) (specificity: 0.64 and sensitivity: 0.65). We consider the discrepancy between these MICROC values to result from the impreciseness of the ROC method, probably due to random fluctuations in the samples. The ROC methodâ&#x20AC;&#x2122;s impreciseness is also revealed from the wider CI in our study as compared with the CI for the MICpred (Table 3) (Terluin et al. 2015).

Finally, the ROC method has been shown to yield the same result as the predictive modeling method when the change scores under study are perfectly normally distributed (Terluin et al. 2015). However, MICROC values cannot be adjusted for the biased overestimation that results from proportions of improved patients being larger than 50%. The predictive modeling method is therefore preferred due to its strengths that include higher precision than the ROC method, and the ability

545

to adjust for the overestimation resulting from proportions of improved patients being larger than 50% (Terluin et al. 2017).

Limitations of our study Limitations of our study include the risk of selection bias since almost 30% of the patients did not return their 1-year follow-up questionnaires. The non-responders were more often female and had a 2-point lower median OKS score (see Table 2). However, these differences are considered small, and as the responders with complete data are otherwise comparable to non-responders with regards to age, BMI, and preoperative knee awareness, we do not expect that our MIC values would differ had we had a higher response rate. Additionally, patients in our cohort were comparable to patients included in the Danish Knee Arthroplasty Registry. The mean age reported by the national registry for patients undergoing a primary TKR has been 67 to 69 years and the cumulated proportion of females has been 61% since 1997, which supports the representativeness of our cohort (Odgaard et al.

2016). Furthermore, possible confounding factors of the MIC values could be mental and medical comorbidities, socioeconomic characteristics, and radiographic osteoarthritis severity and pattern. However, since the patients in our cohort include diversity of these characteristics, we consider our results to be generalizable to other cohorts and registry settings where the sample diversity is assumed similar.

The risk of recall bias has previously been pointed out as a limitation when using anchor questions to estimate MIC values. Recall bias is considered to be present when the anchor responses are more highly correlated to the PROM follow-up score than the change score (Guyatt et al. 2002, King 2011).

However, Terluin et al. (2017) in a simulation study showed that after adjusting for the proportions of improved patients exceeding 50%, the bias introduced by increasing the dependency on the follow-up score was very small (Terluin et al.

2017). We therefore do not consider recall bias a limitation of our MICpred estimates. Another limitation of the anchorbased approach is the bias caused by response shift. Response shift implies that patientsâ&#x20AC;&#x2122; own judgments of their health state changes throughout the follow-up period, resulting in paradoxical responses to the anchor questions, compared to the changes seen in the PROM. The effects of response shift on MIC estimations and how to handle it are, however, not clear (Schwartz et al. 2017).

A further important acknowledgment is that the MIC for improvement cannot serve to estimate that of deterioration in knee problems (Crosby et al. 2003). Even though 7% of the patients considered themselves to be importantly deteriorated after surgery, the absolute number of deteriorated patients was too low to enable the calculation of MIC values for deterioration.

Lastly, while the FJS was originally intended to evaluate the postoperative cross-sectional outcome of joint replacement surgery, subsequent studies have reported high responsiveness

546

to change from before to after surgery when using an adapted version of the questionnaire (Thienpont et al. 2016, Hamilton et al. 2017). Although the validity and reliability characteristics of the Danish version used in our study were determined only in patients at 1 to 4 years postoperatively, we consider the changes made to the questionnaire to enable measurement of the preoperative knee awareness to be minor and that a new validation study of the Danish version does not seem needed since the changes are in line with other language versions of the FJS. Implications of findings Terluin et al. (2017) demonstrated that the adjusted MICpred represents the mean of the latent individual MICs in a sample.

Thus, the presented adjusted MICpred values can be used to interpret mean change improvements within one group of patients at 1 year after a TKR. In national registries and longitudinal cohort studies, improvements exceeding 8 on the OKS and 14.0 on the FJS reflect improvements that are considered important by a notional average patient. To enable comparison between groups in randomized clinical trials or other comparative study types, the adjusted MICpred values may be used as thresholds for treatment response. These thresholds can be used in responder analyses to calculate the numbers and proportions of responders in each treatment arm (McLeod et al.

2011). The MIC values are not applicable for comparing mean change improvements between groups, or to be used in corresponding sample size estimations. Such minimal important difference (MID) values require careful consideration depending on the specific context under study and how to decide on MID values for a specific study is much discussed (Cook et al. 2015). We do not consider that MID values can be derived from longitudinal data from only 1 cohort of patients.

When evaluating individual patients’ improvement in the clinic, the adjusted MICpred values should not be considered absolute thresholds, but may be used in shared decisionmaking as references to what a notional average patient finds important (King 2011).

The presented MIC values are considered context-specific and should not be transferred to other patient populations, time-points, or interventions. However, the presented MIC values are considered applicable to other TKR cohorts with comparable demographic characteristics as in our cohort.

Finally, the MIC represents the smallest improvement that is needed to be considered important by patients, but it does not necessarily represent the best possible outcome or the full potential of the treatment. Other determinants of outcome may therefore also be relevant to evaluate in the comparison of different interventions.

Conclusion The MIC for improvement values of 8 for the OKS and 14 for the FJS can be used to interpret longitudinal within-group score changes, or as responder criteria when comparing

Acta Orthopaedica 2018; 89 (5): 541–547

improvements between 2 groups at 1 year after TKR. In addition to improving the interpretation of results from research studies, the MIC values may also aid in monitoring quality of treatment through national registries.

Design of the study: LHI, AT, ER. Analyses: LHI. Interpretation of results: LHI, ER, BT, KG, HH, AT. Manuscript preparation: LHI, AT. Manuscript review and final acceptance of manuscript: LHI, ER, BT, KG, HH, AT. The authors would like to thank the staff at the orthopedic department for managing the local database at a daily basis. They also thank statisticians Thomas Kallemose and Håkon Sandholdt for statistical assistance.

Acta thanks Madeleine King and Annette W-Dahl for help with peer review of this study.

Beard D J, Harris K, Dawson J, Doll H, Murray D W, Carr A J, Price A J. Meaningful changes for the Oxford hip and knee scores after joint replacement surgery. J Clin Epidemiol 2015; 68(1): 73-9.

Behrend H, Giesinger K, Giesinger J M, Kuster M S. The “forgotten joint” as the ultimate goal in joint arthroplasty: validation of a new patient-reported outcome measure. J Arthroplasty 2012; 27(3): 430-6.e1.

Clement N D, MacDonald D, Simpson A H. The minimal clinically important difference in the Oxford knee score and Short Form 12 score after total knee arthroplasty. Knee Surgery, Sport Traumatol Arthrosc 2014; 22(8): 1933-9.

Cook J A, Hislop J, Altman D G, Fayers P, Briggs A H, Ramsay C R, Norrie J D, Harvey I M, Buckley B, Fergusson D, Ford I, Vale L D. Specifying the target difference in the primary outcome for a randomised controlled trial: guidance for researchers. Trials 2015; 16(1): 12.

Crosby R D, Kolotkin R L, Williams G R. Defining clinically meaningful change in health-related quality of life. J Clin Epidemiol 2003; 56(5): 395407.

FDA, HHS. Patient-report outcome measures: Guidance for industry use in medical product development to support labeling claims. Silver Spring, MD: FDA; 2009.

Giesinger K, Hamilton D F, Jost B, Holzner B, Giesinger J M. Comparative responsiveness of outcome measures for total knee arthroplasty. Osteoarthritis Cartilage 2014; 22(2): 184-9.

Guyatt G H, Norman G R, Juniper E F, Griffith LE . A critical look at transition ratings. J Clin Epidemiol 2002; 55(9): 900-8.

Hamilton D F, Loth F L, Giesinger J M, Giesinger K, MacDonald D J, Patton J T, Simpson A H, Howie C R. Validation of the English language Forgotten Joint Score-12 as an outcome measure for total hip and knee arthroplasty in a British population. Bone Joint J 2017; 99-B(2): 218-24.

Harris K, Dawson J, Gibbons E, Lim C, Beard D, Fitzpatrick R, Price A. Systematic review of measurement properties of patient-reported outcome measures used in patients undergoing hip and knee arthroplasty. Patient Relat Outcome Meas 2016; 7: 101-8.

Ingelsrud L H, Terwee C B, Terluin B, Granan L-P, Engebretsen L, Mills K A G, Roos E M. Meaningful change scores in the Knee Injury and Osteoarthritis Outcome Score in patients undergoing anterior cruciate ligament reconstruction. Am J Sports Med 2018; 46(5): 1120-8.

Jaeschke R, Singer J, Guyatt G H. Measurement of health status: ascertaining the minimal clinically important difference. Control Clin Trials 1989; 10(4): 407-15.

King M T. A point of minimal important difference (MID): a critique of terminology and methods. Expert Rev Pharmacoecon Outcomes Res 2011; 11(2): 171-84.

McLeod L D, Coon C D, Martin S A, Fehnel S E, Hays R D. Interpreting patient-reported outcome results: US FDA guidance and emerging methods. Expert Rev Pharmacoecon Outcomes Res 2011; 11(2): 163-9.

Acta Orthopaedica 2018; 89 (5): 541–547

Murray D W, Fitzpatrick R, Rogers K, Pandit H, Beard D J, Carr A J, Dawson J. The use of the Oxford hip and knee scores. J Bone Joint Surg Br 2007; 89: 1010-14.

Odgaard A, Emmeluth C, Schrøder H, Kappel A, Lamberg A, Troelsen A, Pedersen A B, Kyndesen S. Danish Knee Arthroplasty Register Annual Report, 2016. Copenhagen; 2016.

Rolfson O, Eresian Chenok K, Bohm E, Lübbeke A, Denissen G, Dunn J, Lyman S, Franklin P, Dunbar M, Overgaard S, Garellick G, Dawson J. Patient-reported outcome measures in arthroplasty registries: Report of the Patient-Reported Outcome Measures Working Group of the International Society of Arthroplasty Registries, Part I: Overview and rationale for patient-reported outcome measures. Acta Orthop 2016; 87(362): 3-8.

Schwartz C E, Powell V E, Rapkin B D. When global rating of change contradicts observed change: examining appraisal processes underlying paradoxical responses over time. Qual Life Res 2017; 26(4): 847-57.

Terluin B, Eekhout I, Terwee C B, de Vet H C. Minimal important change (MIC) based on a predictive modeling approach was more precise than MIC based on ROC analysis. J Clin Epidemiol 2015; 68: 1388-96.

547

Terluin B, Eekhout I, Terwee C B. The anchor-based minimal important change, based on receiver operating characteristic analysis or predictive modeling, may need to be adjusted for the proportion of improved patients.

J Clin Epidemiol 2017; 83: 90-100.

Terwee C B, Roorda L D, Dekker J, Bierma-Zeinstra S M, Peat G, Jordan K P, Croft P, de Vet H C W. Mind the MIC: large variation among populations and methods. J Clin Epidemiol 2010; 63(5): 524-34.

Thienpont E, Vanden Berghe A, Schwab P E, Forthomme J P, Cornu O. Joint awareness in osteoarthritis of the hip and knee evaluated with the ‘Forgotten Joint’ Score before and after joint replacement. Knee Surg Sport Traumatol Arthrosc 2016; 24(10): 3346-51.

Thomsen M G, Latifi R, Kallemose T, Barfod K W, Husted H, Troelsen A. Good validity and reliability of the forgotten joint score in evaluating the outcome of total knee arthroplasty. Acta Orthop 2016; 87(3): 280-5.

548

Acta Orthopaedica 2018; 89 (5): 548–554

No detrimental effect of ligament balancing on functional outcome after total knee arthroplasty: a prospective cohort study on 129 mechanically aligned knees with 3 years’ follow-up Eirik AUNAN ¹ and Stephan M RÖHRL ²

¹ Department of Orthopaedic Surgery, Sykehuset Innlandet Hospital Trust, Lillehammer; ² Orthopaedic Department, Oslo University Hospital, Oslo, Norway Correspondence: eirik.aunan@sykehuset-innlandet.no Submitted 2018-04-22. Accepted 2018-05-21.

Background and purpose — In the classical mechanical alignment technique, ligament balancing is considered a prerequisite for good function and endurance in total knee arthroplasty (TKA). However, it has been argued that ligament balancing may have a negative effect on knee function, and some authors advocate anatomic or kinematic alignment in order to reduce the extent of ligament releases. The effect of the trauma induced by ligament balancing on functional outcome is unknown; therefore, the aim of this study was to investigate this effect.

Patients and methods — 129 knees (73 women) were investigated. Mean age was 69 years (42–82), and mean BMI was 29 (20–43). Preoperatively 103 knees had a varus deformity, 21 knees had valgus deformity, and 5 knees were neutral. The primary outcome measure was the Knee injury and Osteoarthritis Outcome Score (KOOS). Secondary outcome measures were the Oxford Knee Score (OKS) and patient satisfaction (VAS). All ligament releases were registered intraoperatively and outcome at 3 years’ follow-up in knees with and without ligament balancing was compared Results — 86 knees were ligament balanced and 43 knees were not. Ligament-balanced varus knees had more preoperative deformity than varus knees without ligament balancing (p = 0.01). There were no statistically significant differences in outcomes between ligament-balanced and non-ligamentbalanced knees at 3 years’ follow-up. No correlation was found between increasing numbers of soft tissue structures released and outcome.

Interpretation — We did not find any negative effect of the trauma induced by ligament balancing on knee function after 3 years.

Symmetric ligament balancing, creating equal and rectangular gaps, has traditionally been considered a prerequisite for good function and endurance in total knee arthroplasty (TKA) (Sharkey et al. 2002, Matsuda et al. 2005, Graichen et al.

2007, Delport et al. 2013). The need for and the extent of ligament balancing is influenced by patient-dependent factors and surgical factors. The most important patient-dependent factors are the degree of knee deformity and the status of the ligaments and other soft tissues around the knee. The predominant surgical factors are the alignment goal, and whether a measured resection technique or a gap technique is used.

3 different principles for alignment exist. Classical mechanical alignment (Insall et al. 1985), anatomic alignment (­Hungerford and Krackow 1985), and kinematic alignment (Hollister et al. 1993, Eckhoff et al. 2005). In mechanical alignment, the aim is to place the center of the femoral and tibial components at the mechanical axis of the lower extremity and the joint line perpendicular to the mechanical axis. In contrast, anatomic and kinematic alignment aim to reestablish the patient’s natural premorbid alignment, that is with the mechanical axis passing on average 8 mm medial to the joint center and the joint line in 2°–3°varus (Paley 2003). Consequently, by using anatomical or kinematic alignment in a varus knee, less angular correction of the bone is needed and the extent of medial ligament releases is reduced. However, the scientific support for anatomical and kinematic alignment is currently scarce and mechanical alignment remains the gold standard (Abdel et al. 2014, Gromov et al. 2014).

The extent of ligament balancing can also be reduced by using a gap technique rather than a measured resection technique. When a measured resection technique is used, ligament balancing is performed both in extension and in flexion. In contrast, with a classical gap technique, ligament balancing is performed only in extension (Insall and Easley 2001).

© 2018 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/4.0) DOI 10.1080/17453674.2018.1485283

Acta Orthopaedica 2018; 89 (5): 548–554

Hence, the extent of ligament releases in varus knees can be reduced by aiming at anatomical or kinematic alignment and/ or by using a gap technique. Nevertheless, a possible downside is that the knee will be left with the mechanical axis passing medially to the center of the knee and the joint line in varus. In return, this will lead to uneven distribution of loads through the medial and lateral compartments of the knee and increased share forces on the interfaces between implants and bone. These factors may possibly threaten the longevity of the prosthetic knee (Ritter et al. 2011, Kim et al. 2014).

The exercise of ligament balancing induces an additional surgical trauma to the knee and it could be hypothesized that this trauma is deleterious to functional outcome after TKA. Each surgeon must choose between mechanical, anatomic, or kinematic alignment techniques and between measured resection and gap technique. The effect of the trauma induced by ligament balancing on functional outcome after TKA has not been described in the literature. However, it is a crucial factor to consider when the surgeon will decide whether to perform ligament balancing or not, and which alignment strategy and gap-balancing strategy to use. Therefore, we investigated the effect of the trauma imposed by ligament balancing on functional outcome after TKA.

Patients and methods All patients participating in another prospective, randomized, and double-blind study comparing TKA with and without patellar resurfacing (Aunan et al. 2016) were included in this study. Inclusion criteria were patients less than 85 years old scheduled for TKA because of osteoarthritis. Exclusion criteria were knees with severe deformity defined as: bone deformity to such a degree that the bone cuts would damage the ligamentous attachments on the epicondyles; ligament laxity without a firm end-point or to such a degree that ligament releases on the concave side would result in a need for more than 20 mm polyethylene thickness; the combination of bone deformity and ligament laxity resulting in the need for more than 20 mm polyethylene thickness. Excluded were also knees with posterior cruciate deficiency, inflammatory arthritis, and severe medical disability limiting the ability to walk or to fill out the patient-recorded outcome documents. Also excluded were patients with patellar thickness less than 18 mm measured on calibrated digital radiographs, isolated patello-femoral arthrosis, knees with secondary osteoarthritis (except for meniscal sequelae), and knees with previous surgery on the extensor mechanism. 2 patients died before the 3-year followup. In these patients, outcome scores 1 year after the operation were carried forward.

Standard radiographs and standing hip–knee–ankle (HKA) radiographs were taken preoperatively and at follow-up. A knee was considered in neutral alignment when the mechanical axis of the lower extremity passed through the center of the

549

tibial spines of the knee and any deviation was termed varus or valgus deformity according to the definitions recommended

by Paley (2003).

Surgical technique All knees were operated through a standard midline incision and a medial para-patellar arthrotomy, using a posterior cruciate-retaining prosthesis (NexGen, Zimmer, Warsaw, IN, USA) and measured resection technique. Classical mechanical alignment was aimed for by setting the valgus angle of the femoral component at 5–8 degrees, depending on the hip– knee–femoral shaft angle (HKFS) as measured on preoperative HKA radiographs.

Rotation of the femoral component was decided with the clinical rotational axis (CRA) method, described by Aunan et al. (2017). The tibial component was aligned to the medial third of the tibial tubercle or with a modified self-seeking technique. Ligament balancing was performed using the technique described by Whiteside and colleagues (Whiteside 1999, Whiteside et al. 2000). The aims of the ligament balancing were medial and lateral laxities of 1–3 mm in both extension and 90° of flexion, and equal and rectangular flexion and extension gaps. The indication for ligament balancing was laxities outside these limits. If an important difference in the height of the flexion and extension gap was still observed after ligament balancing, the gaps were corrected according to the contingency table described by Mont et al. (1999). Medial and lateral ligament laxity in extension and 90° of flexion was measured with the spatula method (Aunan et al. 2012, 2015).

This method has demonstrated a very high inter-rater reliability with an intraclass correlation coefficient equal to 0.88.

Outcome measures The primary outcome measure was the Knee injury and Osteoarthritis Outcome Score (KOOS) (Roos and Toksvig-Larsen 2003). Secondary outcome measures were the Oxford Knee Score (Dawson et al. 1998) and patient satisfaction measured on a visual analog scale (VAS). The primary and secondary outcome measures were recorded preoperatively and at 3 years of follow-up. VAS was recorded at 3 years.

First all ligament releases were registered intraoperatively.

Second, outcome scores at 3 years’ follow-up in knees with and without ligament balancing was compared. Third, the change in outcome scores from preoperative to the 3-year follow-up in each group was compared. Fourth, the change in outcome scores for varus knees and valgus knees was analyzed separately. Finally, the correlation between increasing number of ligament releases and functional outcome for all ligament-balanced knees was estimated.

Statistics A post hoc sample size calculation was performed with the OpenEpi, Version 3 (http://www.openepi.com/Menu/OE_ Menu.htm), open source calculator. The minimal clinically

550

Acta Orthopaedica 2018; 89 (5): 548–554

Table 1. Baseline data for knees with and without ligament balancing. Values are mean (range) unless otherwise specified

Factor

With Without ligament ligament balancing balancing (n = 86) (n = 43)

Without ligament balancing, 28 With ligament balancing, 1

Eligible knees, 129

p-value

All knees: Age 69 (42–81) 70 (53–82) 0.4 a BMI 29 (23–43) 29 (20–38) 0.8 a Women/men, n 50/36 23/20 0.7 b Patellar resurfacing yes/no, n 40/46 23/20 0.5 b Varus knees: Number of knees 75 28 Age 70 (48–81) 70 (53–82) 0.9 a BMI 29 (23–43) 30 (22–38) 0.4 a Women/men, n 41/34 13/15 0.5 b Deformity c 10° (4.4) 2–22 7° (5.1) 1–21 0.01 a Patellar resurfacing yes/no, n 38/37 16/12 0.7 b Valgus knees: Number of knees 10 11 Age 65 (42–79) 72 (63–82) 0.1 a BMI 32 (26–38) 28 (20–34) 0.06 a Women/men, n 9/1 8/3 0.6 b Deformity c 5° (3.2) 2–13 7° (3.0) 3–13 0.3 a Patellar resurfacing yes/no, n 2/8 6/5 0.2 b Neutral knees: Number of knees 1 4 Age 69 70 (65–79) BMI 32 30 (25–34) Women/men, n 0/1 2/2 Patellar resurfacing yes/no, n 0/1 1/3 a Independent samples b Fisher’s exact test. c Mean (SD) and range

With ligament balancing, 75 Varus knees, 103

Neutral knees, 5

Valgus knees, 21

Without ligament balancing, 4 With ligament balancing, 10 Without ligament balancing, 11

Number of knees with and without ligament balancing in different alignment groups.

Table 2. Frequency of soft tissue releases in 86 ligament-balanced knees

Structure released MCL, anterior part MCL, posterior part Medial posterior capsule Semimembranosus Pes anserinius Popliteus tendon Lateral collateral ligament Tractus ileotibialis Posterior-lateral corner.

Lateral posterior capsule Posterior cruciate ligament Total

Varus Valgus Neutral knees knees knees 57 47 11 2 0 5 1 0 0 0 33

2 a 1 1 a 0 0 0 0 0 0 0 4 0 1 0 4 0 2 0 4 0 3 0

156 21 1

a Compensatory

releases.

MCL: Medial collateral ligament.

t-test.

important difference (MCID) in KOOS was set at 10 points and the mean SD of all KOOS sub-scores at 3 years was set at 16.

The ratio of sample sizes was set at 0.5, the 2-sided CI at 95%, and the power at 90%. Given these data, the total sample size was calculated to be 122 with 41 in one group and 81 in the other.

Data were checked visually for normality based on histograms. Means or median values are presented depending on the distribution of data. Comparison of mean and median values was performed using the independent-samples t-test for normally distributed data and the Mann–Whitney U-test for skewed variables. Fisher’s exact test was used when analyzing categorical variables. The association between the number of ligaments released and outcome was estimated with Spearman’s correlation analysis. A significance level of 5% was used and the analyses were performed with IBM SPSS 22 software (IBM Corp, Armonk, NY, USA).

Ethics, funding, and potential conflicts of interest The patients included in this study was recruited from another randomized and double-blind trial that was

approved by the Regional Committee of Research Ethics at the University of Oslo (REK: 1.2007.952) and registered at ClinicalTrials.gov (identifier: NCT00553982). Later additions to the protocol was approved by the same committee (ID number: S-07172d 1.2007.952) and (2010/1678 D 33-07172b 1.2007.952 with changes 05.03.2012). All the patients signed an informed consent form. The first author has received funding from Sykehuset Innlandet Hospital trust. There are no conflicts of interest.

Results 129 knees were investigated (Table 1). Preoperatively 103 knees had a varus deformity, 21 knees had valgus deformity, and 5 knees were neutral (Figure). Ligament-balanced varus knees had statistically significantly more preoperative deformity than varus knees without ligament balancing. No other statistically significant differences in baseline data were observed.

86 knees were ligament balanced and 43 knees were not.

In the ligament-balanced knees, mean 2 (1–4) ligament structures were released per knee (Table 2).

Acta Orthopaedica 2018; 89 (5): 548–554

551

Table 3. Median (IQR) values for functional outcome for ligamentbalanced and non-ligament-balanced knees at 3 years follow-up

Factor

Without ligament balancing (n = 43)

With ligament balancing (n = 86)

p-value a

KOOS: Pain 92 (17) 97 (19) 0.3 Symptoms 89 (14) 93 (14) 0.9 ADL 93 (24) 94 (24) 0.7 Sport/recreation 70 (45) 65 (41) 0.9 QOL 88 (38) 88 (27) 0.9 Oxford Knee Score 56 (10) 57 (7) 0.3 Patient satisfaction 98 (10) 98 (10) 0.6 a Mann-Whitney

U test.

KOOS: Knee injury and Osteoarthritis Outcome Score, 0–100.

Best score is 100. ADL: Activities of daily living. QOL: Knee related quality of life.

Oxford knee score, 12-60. Best score is 60.

Table 4. Mean (SD) change in outcome scores for all knees (N = 129) from baseline to the 3 years follow up in ligament-balanced and non-ligament-balanced knees

Factor

Without ligament balancing (n = 43)

With ligament balancing (n = 86)

p-value a

KOOS: Pain 42 (18) 48 (19) 0.09 Symptoms 36 (17) 37 (20) 0.7 ADL 38 (19) 42 (21) 0.3 Sport/recreation 48 (27) 49 (30) 0.8 QOL 55 (22) 58 (25) 0.5 Oxford Knee Score 18 (7) 20 (8) 0.4 a Independent

samples t-test.

Abbreviations: See Table 3.

There were no statistically significant differences in outcome scores between ligament-balanced and non-ligamentbalanced knees at 3 years’ follow-up (Table 3), or in change of outcome score from baseline to follow-up between the 2 groups (Table 4). When varus and valgus knees were investigated separately, still no difference between ligament-balanced and non-ligament-balanced knees was observed (Table 5). No correlation was found between increasing numbers of soft tissue structures released on the one hand and KOOS, OKS or patient satisfaction on the other.

Discussion Our findings indicate that the surgical trauma imposed by ligament balancing does not have a detrimental effect on knee function as assessed 3 years after the operation. The majority

Table 5. Mean (SD) change in outcome scores from baseline to the 3 years follow up for varus-deformed and valgus-deformed knees in ligament-balanced and non-ligament-balanced knees

Factor

Without ligament balancing (n = 43)

With ligament balancing (n = 86)

Varus knees (n = 103), n 28 75 KOOS: Pain 46 (19) 49 (18) Symptoms 37 (16) 36 (20) ADL 40 (21) 41 (20) Sport/recreation 52 (26) 50 (29) QOL 60 (20) 58 (25) Oxford Knee Score 20 (8) 20 (8) Valgus knees (n = 21), n 11 10 KOOS: Pain 35 (12) 45 (26.)

Symptoms 38 (12) 41 (18) ADL 37 (11) 45 (22) Sport/recreation 44 (25) 42 (33) QOL 49 (20) 56 (31) Oxford Knee Score 15 (5) 19 (11)

p-value a

0.6 0.9 0.8 0.7 0.7 1.0 0.3 0.7 0.3 0.8 0.6 0.3

a Independent

samples t-test.

Abbreviations: See Table 3.

of the ligament-balanced knees had more deformity at baseline than the non-ligament-balanced knees, indicating a less favorable prognosis. Nevertheless, despite multiple releases in many knees, we could not find any negative effect of ligament balancing.

It is well documented that as much as one-fifth of TKA patients are unsatisfied with their TKA (Bourne et al. 2010).

The majority of TKAs have until now been aligned according to the principle of mechanical alignment. However, it has been shown that most native knees are slightly varus-aligned (Paley 2003) and that one-third of men and one-fifth of women have constitutional varus knees with a natural mechanical alignment ≥ 3° degrees varus (Bellemans et al. 2012). Based on this information, it has been speculated that one reason for dissatisfaction with TKA can be that mechanical alignment does not recreate the patient’s premorbid natural alignment (Bellemans et al. 2012, Lee et al. 2017), and that the increased need for ligament balancing in mechanically aligned varus knees can be detrimental to the functional outcome (Bellemans et al.

2012, Gu et al. 2014). Our findings do not support this theory, indicating that the need for additional soft tissue releases is not a valid argument against mechanical alignment in TKA. Kinematic alignment reduces the need for ligament and other soft tissue releases in 2 different ways: first, in traditional mechanical ligament balancing the goal is to obtain rectangular and equal flexion and extension gaps (Sharkey et al. 2002, Matsuda et al. 2005, Graichen et al. 2007, Delport et al. 2013). In kinematic alignment theory, the aim is to restore the native laxity of the knee ligaments (Lee et al.

2017). Native knee ligament laxity is more pronounced later-

552

ally than medially and more laxity is present in flexion than in extension (Tokuhara et al. 2004, Van Damme et al. 2005, Nowakowski et al. 2012). Consequently, by preserving these native properties the need for medial soft tissue releases in a varus-deformed knee is reduced as compared with traditional mechanical balancing. Second, in kinematically and anatomically aligned TKAs the need for soft tissue releases in varus deformed knees is reduced because less correction of the varus deformity is needed, thus less tension is generated in the medial soft tissues.

The degree of ligament balancing in flexion can also be reduced if a gap technique is used instead of a measured resection technique (Insall and Easley 2001). However, in a varus knee this will lead to external rotation of the femoral component and varus alignment in flexion. In a valgus knee, it will result in internal rotation of the femoral component and potential maltracking of the patella and valgus deformity in flexion.

Mechanical alignment is still considered a gold standard (Abdel et al. 2014, Gromov et al. 2014) but anatomic and kinematic alignment have gained increasing popularity in the last decade (Lee et al. 2017) and there is an ongoing debate as to what is the best alignment goal (Lee et al. 2017, Young et al. 2017). Classical mechanical alignment was introduced in order to secure equal distribution of loads between the medial and lateral compartments of the knee and to reduce shear forces at the interfaces between implants and bone (Insall et al. 1985). However, some recent studies have failed to show a relationship between coronal plane alignment and prosthetic survival (Parratte et al. 2010, Bonner et al. 2011).

Therefore, in the hope of improving knee function after TKA growing enthusiasm for anatomic and kinematic alignment has emerged. Nevertheless, an important matter to consider is the ability of current surgical techniques to reach the exact alignment goal. Although outliers from the mechanical axis ≥ 3° may be acceptable (Parratte et al. 2010, Bonner et al. 2011), the same amount of divergence in varus from the natural axis is probably not compatible with long-term survival and good knee function. Consequently, in order to prevent unacceptable outliers, the use of anatomic or kinematic alignment presumes surgical techniques with a high degree of accuracy and precision. Another limitation to the kinematic alignment theory is that replication of normal alignment and ligament laxity does not necessarily lead to more natural knee joint kinematics in TKA. It must be remembered that almost all total knee designs sacrifice 1 or both cruciate ligaments. The lack of well-functioning cruciate ligaments has profound impact on knee kinematics (Scanian and Andriacchi 2017), and nonanatomic prosthetic design features are needed to compensate for the lack of the cruciate ligament(s) and secure stability. It is therefore the authors’ opinion that, in the current context, the term kinematic alignment is too optimistic.

There are some limitations to this study. First, when the study population was subdivided into varus- and valgus-

Acta Orthopaedica 2018; 89 (5): 548–554

deformed knees (Table 5) the subsequent comparisons between ligament balanced and non-ligament balanced knees are underpowered, increasing the risk of a type 2 error. However, we observed no trends in favor of the non-ligamentbalanced knees. Second, we do not know how the ligamentbalanced knees would have performed without ligament balancing. Nevertheless, the fact that no differences between the groups were found in change in scores (Δ-values) (Tables 4 and 5) and that no correlation was found was found between increasing numbers of released soft tissue structures and outcome suggests that no real difference between the groups exists. Although an RCT could have been preferred, given the huge amount of literature pointing out the importance of proper ligament balancing in deformed knees with soft tissue contractures, it is our opinion that an RCT on this population would be unethical. Third, ligament balancing was performed according to the methods described by Whiteside et al. (Whiteside 1999, Whiteside et al. 2000). The results of our study are therefore not valid for other ligament-balancing techniques. Finally, optimal ligament balancing has until recently been unknown. Some earlier reports that compared lax and tight TKAs found better functional outcomes in lax knees (Edwards et al. 1988, Kuster et al. 2004). However, during the last decade different research groups have come to conclusions or recommendations that seem to resemble each other. For example, Heesterbeek et al. (2008) recommended 0.7–3.9° valgus laxity and 0.2–5.4° varus laxity in extension. In flexion they recommended between 0° and 7.1° varus laxity and between 0° and 5.5° valgus laxity. Bellemans et al. (2010) assumed ligament balance successful when 2–4 mm medial–lateral joint line opening was obtained in extension and 2–6 mm in flexion. Okamoto et al. (2014) concluded that the extension gap needs more than 1 mm laxity to avoid postoperative flexion contracture in a clinical study. Our research group studied the effect of ligament laxity measured intraoperatively on functional outcome at 1-year follow-up (Aunan et al. 2015). We concluded that medial laxity more than 2 mm in extension and 3 mm in flexion should be avoided in neutral and valgus-aligned knees and that the lateral side is more forgiving. These findings are supported by a recent study by Ismailidis et al. (2017) that found a positive effect on postoperative flexion and patient satisfaction in knees where the flexion gap exceeded the extension gap by 2.5 mm. Furthermore, Tsukiyama et al.

(2017) reported that medial rather than lateral knee instability correlates with inferior patient satisfaction and knee function after TKA. In summary, the potential detrimental effect of the surgical trauma imposed by ligament balancing is an important determinant that must be considered when surgeons choose between different principles for alignment and gap balancing.

It is also a crucial factor in cases where the need for ligament releases is debatable. We did not find any negative effect of ligament balancing on knee function after 3 years.

Acta Orthopaedica 2018; 89 (5): 548–554

EA: conception, design, data collection, analysis, interpretation, and writing of manuscript. SMR: Revision and approval of the manuscript.

Acta thanks Kirill Gromov and Kjell G Nilsson for help with peer review of this study.

Abdel M P, Oussedik S, Parratte S, Lustig S, Haddad F S. Coronal alignment in total knee replacement: historical review, contemporary analysis, and future direction. Bone Joint J 2014; 96-B(7): 857-62.

Aunan E, Kibsgard T, Clarke-Jenssen J, Röhrl S M. A new method to measure ligament balancing in total knee arthroplasty: laxity measurements in 100 knees. Arch Orthop Trauma Surg 2012; 132(8): 1173-81.

Aunan E, Kibsgard T J, Diep L M, Röhrl S M. Intraoperative ligament laxity influences functional outcome 1 year after total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2015; 23(6): 1684-92.

Aunan E, Naess G, Clarke-Jenssen J, Sandvik L, Kibsgard T J. Patellar resurfacing in total knee arthroplasty: functional outcome differs with different outcome scores. Acta Orthop 2016; 87(2): 158-64.

Aunan E, Ostergaard D, Meland A, Dalheim K, Sandvik L. A simple method for accurate rotational positioning of the femoral component in total knee arthroplasty. Acta Orthop 2017; 88(6): 657-63.

Bellemans J, Vandenneucker H, Van Lauwe J, Victor J. A new surgical technique for medial collateral ligament balancing: multiple needle puncturing.

J Arthroplasty 2010; 25(7): 1151-6.

Bellemans J, Colyn W, Vandenneucker H, Victor J. The Chitranjan Ranawat award: is neutral mechanical alignment normal for all patients? The concept of constitutional varus. Clin Orthop Relat Res 2012; 470(1): 45-53.

Bonner T J, Eardley W G, Patterson P, Gregg P J. The effect of post-operative mechanical axis alignment on the survival of primary total knee replacements after a follow-up of 15 years. J Bone Joint Surg Br 2011; 93(9): 1217-22.

Bourne R B, Chesworth B M, Davis A M, Mahomed N N, Charron K D. Patient satisfaction after total knee arthroplasty: who is satisfied and who is not? Clin Orthop Relat Res 2010; 468(1): 57-63.

Dawson J, Fitzpatrick R, Murray D, Carr A. Questionnaire on the perceptions of patients about total knee replacement. J Bone Joint Surg Br 1998; 80(1): 63-9.

Delport H, Labey L, De Corte R, Innocenti B, Vander Sloten J, Bellemans J. Collateral ligament strains during knee joint laxity evaluation before and after TKA. Clin Biomech (Bristol, Avon) 2013; 28(7): 777-82.

Eckhoff D G, Bach J M, Spitzer V M, Reinig K D, Bagur M M, Baldini T H, Flannery N M. Three-dimensional mechanics, kinematics, and morphology of the knee viewed in virtual reality. J Bone Joint Surg Am 2005; 87(Suppl 2): 71-80.

Edwards E, Miller J, Chan K H. The effect of postoperative collateral ligament laxity in total knee arthroplasty. Clin Orthop Relat Res 1988; (236): 44-51.

Graichen H, Strauch M, Katzhammer T, Zichner L, von Eisenhart-Rothe R. [Ligament instability in total knee arthroplasty—causal analysis]. Orthopade 2007; 36(7): 650,652-6.

Gromov K, Korchi M, Thomsen M G, Husted H, Troelsen A. What is the optimal alignment of the tibial and femoral components in knee arthroplasty?

Acta Orthop 2014; 85(5): 480-7.

Gu Y, Roth JD, Howell SM, Hull ML. How frequently do four methods for mechanically aligning a total knee arthroplasty cause collateral ligament imbalance and change alignment from normal in white patients? AAOS exhibit selection. J Bone Joint Surg Am 2014; 96(12): e101.

Heesterbeek P J, Verdonschot N, Wymenga A B. In vivo knee laxity in flexion and extension: a radiographic study in 30 older healthy subjects. Knee 2008; 15(1): 45-9.

Hollister A M, Jatana S, Singh A K, Sullivan W W, Lupichuk A G. The axes of rotation of the knee. Clin Orthop Relat Res 1993; (290): 259-68.

553

Hungerford D S, Krackow K A. Total joint arthroplasty of the knee. Clin Orthop Relat Res 1985; (192): 23-33.

Insall J, Easley M. Surgical techniques and instrumentation in total knee arthroplasty. In: Insall J, Scott W, editors. Surgery of the knee. New York: Churchill Livingstone; 2001. p. 1553-620.

Insall J N, Binazzi R, Soudry M, Mestriner L A. Total knee arthroplasty. Clin Orthop Relat Res 1985; (192): 13-22.

Ismailidis P, Kuster M S, Jost B, Giesinger K, Behrend H. Clinical outcome of increased flexion gap after total knee arthroplasty: can controlled gap imbalance improve knee flexion? Knee Surg Sports Traumatol Arthrosc 2017; 25(6): 1705-11.

Kim Y H, Park J W, Kim J S, Park S D. The relationship between the survival of total knee arthroplasty and postoperative coronal, sagittal and rotational alignment of knee prosthesis. Int Orthop 2014; 38(2): 379-85.

Kuster M S, Bitschnau B, Votruba T. Influence of collateral ligament laxity on patient satisfaction after total knee arthroplasty: a comparative bilateral study. Arch OrthopTrauma Surg 2004; 124(6): 415-17.

Lee Y S, Howell S M, Won Y Y, Lee O S, Lee S H, Vahedi H, Teo S H. Kinematic alignment is a possible alternative to mechanical alignment in total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2017; 25(11): 3467-79. doi: 10.1007/s00167-017-4558-y. Epub 2017 Apr 24.

Matsuda Y, Ishii Y, Noguchi H, Ishii R. Varus-valgus balance and range of movement after total knee arthroplasty. J Bone Joint Surg Br 2005; 87(6): 804-8.

Mont M A, Delanois R, Hungerford D S. Balancing and alignment: surgical techniques on how to achieve soft-tissue balancing. In: Lotke PA, Garino JP, editors. Revision total knee arthroplasty. Philadelphia: LippincottRaven; 1999. p. 173-86.

Nowakowski A M, Majewski M, Muller-Gerbl M, Valderrabano V. Measurement of knee joint gaps without bone resection: “physiologic” extension and flexion gaps in total knee arthroplasty are asymmetric and unequal and anterior and posterior cruciate ligament resections produce different gap changes. J Orthop Res 2012; 30(4): 522-7.

Okamoto S, Okazaki K, Mitsuyasu H, Matsuda S, Mizu-Uchi H, Hamai S, Tashiro Y, Iwamoto Y. Extension gap needs more than 1-mm laxity after implantation to avoid post-operative flexion contracture in total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2014; 22(12): 3174-80.

Paley D. Principles of deformity correction. 1st ed. Berlin, Heidelberg: Springer; 2003.

Parratte S, Pagnano M W, Trousdale R T, Berry D J. Effect of postoperative mechanical axis alignment on the fifteen-year survival of modern, cemented total knee replacements. J Bone Joint Surg Am 2010; 92(12): 2143-9.

Ritter M A, Davis K E, Meding J B, Pierson J L, Berend M E, Malinzak R A. The effect of alignment and BMI on failure of total knee replacement. J Bone Joint Surg Am 2011; 93(17): 1588-96.

Roos E M, Toksvig-Larsen S. Knee injury and Osteoarthritis Outcome Score (KOOS): validation and comparison to the WOMAC in total knee replacement. Health Qual Life Outcomes 2003; 1(1): 17.

Scanian S F, Andriacchi T P. Human movement and anterior cruciate ligament function: anterior cruciate ligament injury and gait mechanics. In: (Noyes F R, ed) Noyes’ knee disorders: surgery, rehabilitation, clinical outcomes.

2nd ed. Philadelphia: Elsevier; 2017. p. 125-36.

Sharkey P F, Hozack W J, Rothman R H, Shastri S, Jacoby S M. Insall Award paper: Why are total knee arthroplasties failing today? Clin Orthop Relat Res 2002; 404: 7-13.

Tokuhara Y, Kadoya Y, Nakagawa S, Kobayashi A, Takaoka K. The flexion gap in normal knees: an MRI study. J Bone Joint Surg Br 2004; 86(8): 1133-6.

Tsukiyama H, Kuriyama S, Kobayashi M, Nakamura S, Furu M, Ito H, Matsuda S. Medial rather than lateral knee instability correlates with inferior patient satisfaction and knee function after total knee arthroplasty. Knee 2017; 24(6): 1478-84.

554

Van Damme G, Defoort K, Ducoulombier Y, Van Glabbeek F, Bellemans J, Victor J. What should the surgeon aim for when performing computerassisted total knee arthroplasty? J Bone Joint Surg Am 2005; 87(Suppl 2): 52-8.

Whiteside L A. Selective ligament release in total knee arthroplasty of the knee in valgus. Clin Orthop Relat Res 1999; 367: 130-40.

Acta Orthopaedica 2018; 89 (5): 548â&#x20AC;&#x201C;554

Whiteside L A, Saeki K, Mihalko W M. Functional medial ligament balancing in total knee arthroplasty. Clin Orthop Relat Res 2000; 380: 45-57.

Young S W, Walker M L, Bayan A, Briant-Evans T, Pavlou P, Farrington B. The Chitranjan S. Ranawat Award: No difference in 2-year functional outcomes using kinematic versus mechanical alignment in TKA: a randomized controlled clinical trial. Clin Orthop Relat Res 2017; 475(1): 9-20.

Acta Orthopaedica 2018; 89 (5): 555–559

555

Anterior distal femoral hemiepiphysiodesis can reduce fixed flexion deformity of the knee: a retrospective study of 83 knees Norbert STIEL 1,2, Kornelia BABIN 1,2, Eik VETTORAZZI 3, Sandra BREYER 1,2, Nicola EBERT 1,2, Martin RUPPRECHT 1,2, Ralf STUECKER 1,2 and Alexander S SPIRO 1,2 1 Department

of Pediatric Orthopaedic Surgery, Children’s Hospital Hamburg-Altona, Hamburg; 2 Department of Orthopaedics, University Medical Center Hamburg-Eppendorf, Hamburg; 3 Department of Medical Biometry and Epidemiology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany Correspondence: a.spiro@uke.de Submitted 2017-10-08. Accepted 2018-05-02.

Background and purpose — Fixed knee flexion deformity in children is a common problem in various diseases including myelomeningocele and cerebral palsy. Until now, only a few studies focusing on the surgical procedure of anterior distal femoral hemiepiphysiodesis have been published.

We analyzed outcome and correction rate in the largest case series to date of patients treated by staples or 8-plates.

Patients and methods — We reviewed the medical records of all patients with fixed knee flexion deformity who were treated with anterior distal femoral hemiepiphysiodesis using either staples or 8-plates between the years 2002 and 2017 (73 patients; 130 knees). 49 patients (83 knees) had completed treatment with implant removal at the time of full correction of the deformity or at skeletal maturity and were included. The average age at operation was 12 years (6–20). Patients were assigned to 3 different groups based on their diagnosis: cerebral palsy, myelomeningocele, and the “other” group.

Results — Mean fixed knee flexion deformity improved from 21° (10–60°) to 8° (0–50°) (p < 0.001) with an average correction rate of 0.44° per month (range –2.14° to 1.74°).

The correction rate per month was lowest for patients with cerebral palsy (0.20°), followed by the myelomeningocele group (0.50°), and the “other” group (0.58°). Implant loosening occurred in 10% of the treated knees with consecutive re-implantation in 5% of the cases.

Interpretation — Anterior distal femoral hemiepiphysiodesis is an effective and safe method for the treatment of fixed knee flexion deformity in children. The optimal timing depends on the remaining individual growth potential, the underlying disease, and the extent of the deformity.

Fixed knee flexion deformity in children is a common problem in various diseases including arthrogryposis, myelomeningocele, and cerebral palsy (Williams et al. 1993, Wren et al. 2005, van Bosse et al. 2007, van der Krogt et al. 2007).

With shortening of muscles with or without spasticity children develop contractures and even bony deformities followed by decreased endurance, knee pain, and progressive crouch gait (Young et al. 2010). Even in non-ambulators knee flexion deformity interferes with activities of daily living (Williams et al. 1993, Murray and Fixsen 1997, Moen et al. 2005, Devalia et al. 2007). Even moderate deformities respond poorly to nonoperative treatment like casting, bracing, physical therapy, or local application of botulinum toxin (Molenaers et al. 2006, Westberry et al. 2006, Carbonell et al. 2007). Common surgical procedures include distal femoral extension osteotomy, arthrodiastasis using external fixators, and soft tissue release (Beals 2001, Saraph et al. 2002, Carbonell et al. 2007, Devalia et al. 2007). These procedures are associated with neurovascular risk, infection risk, prolonged postoperative immobilization, fractures, knee instability, and recurrent deformity with continued growth (Devalia et al. 2007, van Bosse et al. 2007, de Morais Filho et al. 2008).

In 2001 Kramer and Stevens published their first data on anterior distal femoral stapling for fixed knee flexion deformity (Kramer and Stevens 2001). Temporary anterior distal femoral hemiepiphysiodesis can be performed by stapling or the implantation of 8-plates. There are few published studies focusing on this surgical procedure. These studies demonstrated an improvement of the deformities combined with a low complication rate due to its less invasive character (Kramer and Stevens 2001, Klatt and Stevens 2008, Palocaren et al. 2010, Macwilliams et al. 2011, Al-Aubaidi et al. 2012).

However, none of these studies looked at the correction rates

© 2018 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/4.0) DOI 10.1080/17453674.2018.1485418

556

Acta Orthopaedica 2018; 89 (5): 555–559

Eligible cases 73 patients / 130 knees Excluded: – follow-up < 12 months, 9 patients / 16 knees – hamstring lengthening, 7 patients / 13 knees Follow-up 57 patients / 101 knees No implant removal 10 patients / 18 knees

Implant removal 49 patients / 83 knees

Open physis at removal 23 patients / 34 knees a No recurrent deformity 15 patients / 22 knees

Closed physis at removal 28 patients / 49 knees

Recurrent deformity 3 patients / 3 knees Repeat intervention 3 patients / 3 knees

Figure 1. Patients and follow-up. a 4 patients (7 knees) had no follow-up after removal and 1 patient (2 knees) had a permanent epiphysiodesis at the time of implant removal. Some patients had implant removal on one side only, although both sides had been treated.

according to the underlying diseases after anterior distal femoral hemiepiphysiodesis.

We reported our early results by using staples and 8-plates in 20 patients in 2012 (Spiro et al. 2012). This is a follow-up study involving all 73 patients who were treated with this procedure at our institution over a 15-year period. We assessed the outcome, complications, and correction rates in relation to the underlying disease or etiology.

Patients and methods

less than 12 months of predicted growth remaining; and (4) patients who had additional soft-tissue procedures around the knee, such as hamstring lengthening or capsular release.

The flexion contracture angle, defined as the angle between the neutral position corresponding to 0° and the maximum extension of the knee, was measured with a goniometer.

Anteroposterior and lateral radiographs were taken of all knees to document open distal femoral physis preoperatively.

From January 2002 to July 2017, 73 patients (43 males) were treated by anterior distal femoral hemiepiphysiodesis.

49 patients (83 knees) had completed treatment with implant removal at the time of full correction of the deformity or at skeletal maturity and were included. The average age at operation was 12 years (6–20). The mean age was higher in male than in female patients (13.1 vs. 10.5 years; p < 0.001). All patients were under 16 years of age at the time of surgery, except 1 patient who suffered from de Grouchy syndrome.

This patient had a chronological age of 20 years, but radiographs showed open distal femoral physis on both sides.

Growth modulation by anterior distal femoral hemiepiphysiodesis was performed on 130 knees (57 bilateral, 16 unilateral fixations). The angle of contracture varied from 10° to 60°. 68 patients were treated with staples and 5 patients with 8-plates. Implants (staples or plates) were removed when full knee extension was achieved or when skeletal maturity had occurred.

Group assignment The patients who had completed treatment were assigned to 3 groups according to their diagnosis. Most common diagnoses included cerebral palsy and myelomeningocele (Table 1).

Surgical technique (Figure 2) All procedures were performed under general anesthesia with the patient supine on a radiolucent table using tourniquets and identification of the distal femoral physis by radiography.

Longitudinal incisions of approximately 3 cm were performed

The medical records of all patients with fixed knee flexion deformity who had been treated with either staples or 8-plates between the years 2002 and 2017 at our institution were reviewed (Figure 1). Inclusion criteria were: (1) fixed knee flexion deformity exceeding 10°, not responding to non-operative treatment (physi- Table 1. Group assignment cal therapy, bracing, casting); (2) temporary growth modulation by anterior distal femoral

No. of Age at Time to Etiology of knee patients Age at surgery implant removal implant removal hemiepiphysiodesis; (3) implant removal at the flexion deformity (knees) years (range) years (range) months (range) time of full correction of the deformity or at

skeletal maturity independent of full correction; Cerebral palsy 14 (25) 13 (9–16) 15 (10–21) 38 (6–72) Myelomeningocele 17 (31) 12 (9–15) 14 (10–18) 31 (12–52) and (4) consistent radiographs preoperatively, a Other 18 (27) 12 (6–20) 14 (8–23) 27 (9–63) postoperatively, at the time of implant removal, Total 49 (83) 12 (6–20) 15 (8–23) 32 (6–72) and during follow-up. Clinical and radiographic follow-up examinations were usually performed a In the “other” group, 3 patients had a congenital knee flexion deformity, 2 suffered every 6 months after surgery. Exclusion critefrom arthrogryposis multiplex congenita, and 2 developed a contracture after distal femoral fracture. Each of the following diseases occurred once in the “other” group: ria were: (1) patients with less than 12 months’ central core myopathy, de Grouchy syndrome, transverse spinal cord syndrome after follow-up; (2) dynamic flexion deformities due embolization, multiple epiphyseal dysplasia, intraspinal lipoma, VACTERL syndrome, to spasticity or contracture of the hamstrings; congenital knee dislocation, Larsen syndrome, popliteal pterygium syndrome, Omenn syndrome, and skeletal dysplasia.

(3) nearly closed distal femoral physis with

Acta Orthopaedica 2018; 89 (5): 555–559

557

was inserted to mark the physis. Thereafter Blount staples were implanted over the physis in a 45° oblique direction. The implanted staples were placed at a distance of 5 mm from the ridge of the femoral sulcus in order to prevent problems with patellofemoral articulation. 8-plates were inserted by using a K-wire instead of a needle for marking the physis. The plates were positioned over the K-wire and fixed using two cannulated screws. Correct position of the implants was verified by fluoroscopy in 2 planes.

After surgery all patients were allowed full weight-bearing and started knee movement immediately as tolerated.

a

b

Statistics Patient characteristics are reported as mean (range). Due to the presence of dependent parameters taken on both sides (right and left knee), linear mixed models for analysis were used.

As the level of significance alpha was set to 0.05. All analyses were performed using R software version 3.3.3 (https:// www.r-project.org/).

Ethics, funding, and potential conflicts of interest No ethical authorization was required for this type of study.

There was no funding. None of the authors have any conflicts of interests or financial disclosures to declare.

Results (Table 2) c

d

e

f

Figure 2. 10-year-old boy with myelomeningocele. Before (a, b) and after (c, d) anterior distal femoral hemiepiphysiodesis. Initial radiographs (a, b) show an avulsion fracture of the lower patellar pole in addition. The small patellar fragment was also resected during surgery (c, d). Radiographs taken 3 years after hemiepiphysiodesis and before the implants were removed (e, f). Flexion deformity improved from 10° extension deficiency to 0°.

on the lateral and medial side of the patella, centered on the physis. An arthrotomy was performed on both sides. After identification of the physis by radiography, a 20-gauge needle

All included patients (49 patients; 83 knees) had completed treatment with implant removal at the time of full correction of the deformity or at skeletal maturity. The average follow-up was 46 months (12–78) after implant removal. The average preoperative fixed knee flexion deformity was 21° (range 10°– 60°), which improved to 8° (0°–50°). The mean correction of fixed knee flexion deformity was 13° (95% CI (10°–15°), p < 0.001) at the point of implant removal.

Implants were removed after 32 months (6–72) on average, according to a mean correction rate of 0.44° (CI 0.32°–0.60°), p < 0.001) per month. There were no statistically significant differences in the correction of the flexion deformity between males and females (p = 0.1). The highest correction rate per month was found for the patients in the “other” group (0.60°), followed by the myelomeningocele group (0.52°). The lowest correction rate per month was found for the patients with cerebral palsy (0.20°). With regard to the correction rate per month, a difference could be found between the cerebral palsy group and the “other” group. Patients with cerebral palsy had a lower correction rate (p = 0.03). We found a correlation between age at the time of surgery and the degree of correction. The improvement of the flexion deformity decreased with each year of age: delta = 1.22° (CI 0.22°–2.23°), p = 0.02.

At the time of implant removal the physis was still open in 23 patients (34 knees). Follow-up examinations to control for rebound deformity were performed for an average of

558

Acta Orthopaedica 2018; 89 (5): 555–559

Table 2. Results of hemiepiphysiodesis. Values are degrees

Knee flexion

Preop.

At removal Correction per Group (range) (range) month (range)

Cerebral palsy 21 (10–50) 11 (0–45) 0.20 (–2.1 to 1.2) Myelomeningocele 20 (10–60) 7 (0–50) 0.52 (–0.30 to 1.7) Other a 21 (10–45) 8 (0–45) 0.60 (0 to 1.3) Total a

21 (10–60)

8 (0–50)

0.44 (–2.1 to 1.7)

See Table 1.

21 months. 3 patients (3 knees) had a recurrent knee flexion contracture with re-stapling at a time when the distal femoral physis was still open. In 8 of the 49 patients (13 knees) the angle of knee flexion deformity remained unchanged or increased.

1 patient developed a hematoma after staple removal and needed revision surgery. There were no other complications during or after surgeries such as wound infection, neurovascular injury, pathological fractures, or reactive synovitis. Implant loosening occurred in 10% of the treated knees with consecutive re-implantation in 5% of the cases.

Discussion In contrast to other surgical procedures, anterior distal femoral hemiepiphysiodesis is a minor operation for fixed knee flexion deformity, it has lower complication risks, and does not require immobilization (Kramer and Stevens 2001). Since Blount and Clarke first mentioned this new method for controlling bone growth by physeal stapling, it has been used successfully to modulate limb-length and correct varus/valgus deformities of the knee in children (Blount and Clarke 1949, Mielke and Stevens 1996, Gorman et al. 2009). The principles of anterior distal femoral hemiepiphysiodesis by using either staples or 8-plates have been described in only a limited number of studies (Kramer and Stevens 2001, Klatt and Stevens 2008, Palocaren et al. 2010, Spiro et al. 2010).

In 2012 we published preliminary data on anterior distal femoral hemiepiphysiodesis. 20 patients with an average age of 13 years were analyzed. We found an improvement in fixed knee flexion deformity from 22° preoperatively to 7° after an average follow-up of 3 years (Spiro et al. 2012). This is a follow-up report with a longer follow-up of the already published data as well as 53 new patients. To our knowledge this is the largest number of patients so far reported in the literature and treated with that procedure. Mean fixed knee flexion deformity improved from 21° (10°–60°) to 8° (10°–50°) with a correction rate of 0.44° per month. Similar results after this procedure have been published by other authors. Klatt and Stevens (2008) found an improvement from 23° to 8° in 23 patients by using 8-plates. Some of their patients had ham-

string lengthening in addition. Patients who had fixed knee flexion deformity and hamstring contracture with an increased popliteal angle were treated by anterior distal femoral hemiepiphysCI iodesis and hamstring lengthening in our study.

We excluded these patients and all patients who –0.06 to 0.46 0.28 to 0.75 had capsular release in addition to anterior distal 0.37 to 0.84 femoral hemiepiphysiodesis in order to avoid any bias. The advantage of anterior distal femoral hemiepiphysiodesis is to avoid complications associated with soft tissue release.

Palocaren et al. (2010) used 8-plates for fixed knee flexion deformity in 10 children (16 knees) with arthrogryposis. Average deformity was 60° preoperatively and was corrected to 33°. Neither screw migration and loosening nor implant breakage was seen in the series of Klatt and Stevens (2008). Palocaren et al. reported only 1 case with plate loosening and consecutive implant removal. Implant loosening occurred in 10% of the treated knees with re-implantation in 5% of the knees in our study. This low proportion of implant loosening after stapling is acceptable. However, in patients with poor bone quality 8-plates for fixation may be preferable.

There have been no detailed analyses of the effectiveness of anterior distal femoral hemiepiphysiodesis focusing on the underlying etiology. The underlying diagnoses had an impact on the correction rate per month when comparing patients with cerebral palsy and patients with “other” diseases. With respect to our 2 largest groups, myelomeningocele and cerebral palsy, the average fixed knee flexion deformity improved from 21° preoperative to 11° in patients with cerebral palsy (correction rate 0.20° per month) and from 20° preoperative to 7° in patients with myelomeningocele (correction rate 0.52° per month). Although patients with myelomeningocele had a higher correction rate than patients with cerebral palsy in our series, the difference was not statistically significant (p = 0.08).

According to patient age at the time of surgery, the improvement of knee flexion deformity decreased by each year of age in this study. Based on these results, we recommend early treatment of fixed knee flexion deformity by anterior distal femoral hemiepiphysiodesis (at around 10 years of age).

Al-Aubaidi et al. (2012) reported one case of stress-related supracondylar femur fracture in a cohort of 8 neuromuscular patients treated with this method. None of our patients developed fractures, neural damage, or deep wound infection. In 3 patients recurrence of deformity was seen requiring repeated correction by hemiepiphysiodesis. In 1 case, after the staple removal, a hematoma developed which needed revision. None of our patients showed impairment of knee flexion or knee stability during the entire follow-up period.

Limitations of our study include its retrospective design.

In anterior distal femoral hemiepiphysiodesis the remaining growth potential and therefore the bone age is important.

Because bone age was not available for all patients, chrono-

Acta Orthopaedica 2018; 89 (5): 555â&#x20AC;&#x201C;559

logical age was used. Another weakness of the study is the fact that popliteal angle and patella position were not evaluated in each case. Having a large group of patients with long-term follow-up until implant removal is a strength of our study.

In summary, this minimally invasive technique with immediate mobilization and short operation time seems to be an effective treatment for fixed knee flexion deformity in patients with a wide spectrum of underlying diagnoses. We believe that the optimal timing of anterior distal femoral hemiepiphysiodesis especially depends on the remaining individual growth potential, the underlying disease, and the extent of the deformity. A slight overcorrection of fixed knee flexion deformity (about 5 degrees) should be considered in patients with substantial growth potential at the time of implant removal to avoid recurrent knee flexion deformity.

NS and ASS: study design, collection and interpretation of data, and writing. KB, SB, NE, KR and MR: collection of data. EV: statistical analyses, interpretation of data, correction of manuscript. RS: study design and correction of manuscript.

Acta thanks Martin Gottliebsen and Niels Wisbech Pedersen for help with peer review of this study.

Al-Aubaidi Z, Lundgaard B, Pedersen N W. Anterior distal femoral hemiepiphysiodesis in the treatment of fixed knee flexion contracture in neuromuscular patients. J Child Orthop 2012; 6: 313-18.

Beals R K. Treatment of knee contracture in cerebral palsy by hamstring lengthening, posterior capsulotomy, and quadriceps mechanism shortening. Dev Med Child Neurol 2001; 43: 802-5.

Blount W P, Clarke G R. Control of bone growth by epiphyseal stapling; a preliminary report. J Bone Joint Surg Am 1949; 31A: 464-78.

Carbonell P G, Valero J V, Fernandez P D, Vicente-Franqueira J R. Monolateral external fixation for the progressive correction of neurological spastic knee flexion contracture in children. Strategies Trauma Limb Reconstr 2007; 2: 91-7. de Morais Filho M C, Neves D L, Abreu F P, Juliano Y, Guimaraes L. Treatment of fixed knee flexion deformity and crouch gait using distal femur extension osteotomy in cerebral palsy. J Child Orthop 2008; 2: 37-43.

Devalia K L, Fernandes J A, Moras P, Pagdin J, Jones S, Bell M J. Joint distraction and reconstruction in complex knee contractures. J Pediatr Orthop 2007; 27:402-7.

Gorman T M, Vanderwerff R, Pond M, Macwilliams B, Santora S D. Mechanical axis following staple epiphysiodesis for limb-length inequality. J Bone Joint Surg Am 2009; 91-A: 2430-9.

559

Klatt J, Stevens P M. Guided growth for fixed knee flexion deformity. J Pediatr Orthop 2008; 28: 626-31.

Kramer A, Stevens P M. Anterior femoral stapling. J Pediatr Orthop 2001; 21: 804-7.

Macwilliams B A, Harjinder B, Stevens P M. Guided growth for correction of knee flexion deformity: a series of four cases. Strategies Trauma Limb Reconstr 2011; 6: 83-90.

Mielke C H, Stevens P M. Hemiepiphyseal stapling for knee deformities in children younger than 10 years: a preliminary report. J Pediatr Orthop 1996; 16: 423-9.

Moen T, Gryfakis N, Dias L, Lemke L. Crouched gait in myelomeningocele: a comparison between the degree of knee flexion contracture in the clinical examination and during gait. J Pediatr Orthop 2005; 25: 657-60.

Molenaers G, Desloovere K, Fabry G, De Cock P. The effects of quantitative gait assessment and botulinum toxin on musculoskeletal surgery in children with cerebral palsy. J Bone Joint Surg Am 2006; 88-A: 161-70.

Murray C, Fixsen J A. Management of knee deformity in classical arthrogryposis multiplex congenita (amyoplasia congenita). J Pediatr Orthop B 1997; 6: 186-91.

Palocaren T, Thabet A M, Rogers K, Holmes L, Jr, Donohoe M, King M M, Kumar S J. Anterior distal femoral stapling for correcting knee flexion contracture in children with arthrogryposi: preliminary results. J Pediatr Orthop 2010; 30: 169-73.

Saraph V, Zwick E B, Zwick G, Steinwender C, Steinwender G, Linhart W. Multilevel surgery in spastic diplegia: evaluation by physical examination and gait analysis in 25 children. J Pediatr Orthop 2002; 22: 150-7.

Spiro A S, Babin K, Lipovac S, Rupprecht M, Meenen N M, Rueger J M, Stuecker R. Anterior femoral epiphysiodesis for the treatment of fixed knee flexion deformity in spina bifida patients. J Pediatr Orthop 2010; 30: 85862.

Spiro A S, Stenger P, Hoffmann M, Vettorazzi E, Babin K, Lipovac S, Kolb J P, Novo De Oliveira A, Rueger J M, Stuecker R. Treatment of fixed knee flexion deformity by anterior distal femoral stapling. Knee Surg Sports Traumatol Arthrosc 2012; 20: 2413-18.

Van Bosse H J, Feldman D S, Anavian J, Sala D A. Treatment of knee flexion contractures in patients with arthrogryposis. J Pediatr Orthop 2007; 27: 930-7.

Van Der Krogt M M, Doorenbosch C A, Harlaar J. Muscle length and lengthening velocity in voluntary crouch gait. Gait Posture 2007; 26: 532-8.

Westberry D E, Davids J R, Jacobs J M, Pugh L I, Tanner S L. Effectiveness of serial stretch casting for resistant or recurrent knee flexion contractures following hamstring lengthening in children with cerebral palsy. J Pediatr Orthop 2006; 26: 109-14.

Williams J J, Graham G P, Dunne K B, Menelaus M B. Late knee problems in myelomeningocele. J Pediatr Orthop 1993; 13: 701-3.

Wren T A, Rethlefsen S, Kay R M. Prevalence of specific gait abnormalities in children with cerebral palsy: influence of cerebral palsy subtype, age, and previous surgery. J Pediatr Orthop 2005; 25: 79-83.

Young J L, Rodda J, Selber P, Rutz E, Graham H K. Management of the knee in spastic diplegia: what is the dose? Orthop Clin North Am 2010; 41: 561-77.

560

Acta Orthopaedica 2018; 89 (5): 560–564

Re-arthrodesis after primary ankle fusion: 134/1,716 cases from the Swedish Ankle Registry Anders HENRICSON 1, Lars JEHPSSON 2, Åke CARLSSON 2, and Björn E ROSENGREN 2 1 Department of Orthopedic Surgery, Falun Central Hospital, Falun, 2 Department of Clinical Sciences and Orthopedic Surgery, Skåne University Hospital, Malmö, Sweden Correspondence: anders.henricson@ltdalarna.se Submitted 2017-09-17. Accepted 2018-04-25.

Background and purpose — Arthrodesis is the most common treatment of severe ankle arthritis. Large studies on the occurrence of re-arthrodesis are few, especially with information in terms of risk. We used the National Swedish Ankle Registry to assess incidence and risk factors for rearthrodesis.

Patients and methods — In the Registry, we examined the occurrence of re-arthrodesis in 1,716 patients with a primary ankle arthrodesis. We also analyzed associations between the re-arthrodesis risk and sex, diagnosis, and surgical method.

Results — The risk of first re-arthrodesis at 2.5 years was 7.4% and the rate at 9 years 7.8%. The risk following arthroscopic surgery with fixation by screws was 15%, which is statistically significantly higher than the 8% following the gold standard technique with open screw fixation, the 5% following fixation by intramedullary nailing, and the 3% following fixation by plate and screws. Patients with either idiopathic osteoarthritis or posttraumatic arthritis had a higher risk of re-arthrodesis than patients with rheumatoid arthritis. We could not find that the risk of re-arthrodesis was associated with sex.

Interpretation — In Sweden, the re-arthrodesis risk varied by primary technique and was especially high after arthroscopic surgery. Reasons are unknown but poor surgical technique and/or surgeon inexperience may contribute, as may patient selection.

Ankle arthrodesis is the most common surgical treatment of severe ankle arthritis. The most frequent complication after ankle arthrodesis is non-union; the reported rate varies between 0% and 31% (see review articles by Nihal et al. 2008, and Yasui et al. 2016a). Studies concerning different non-union rates related to fixation methods or techniques are sparse.

We examined potential differences in re-arthrodesis risk related to sex, diagnosis, and different surgical methods in a large series of primary ankle fusions followed in the Swedish Ankle Registry.

Patients and methods Arthrodesis procedures of the ankle have been reported to the National Swedish Ankle Registry since 2008. The procedurebased coverage is now 96%. The different kinds of fixation and surgical techniques as well as sex, date of birth, date of surgery, and preoperative diagnoses are reported to the registry. Regarding re-arthrodesis the date of secondary surgery and fixation method are registered and also whether a second or third attempt to fuse the ankle has been undertaken.

We identified 1,773 primary ankle arthrodeses from 2008 to 2015. 54 departments had reported between 1 and 203 ankle arthrodeses to the registry; 4 departments reported only 1 arthrodesis each.

45 patients had undergone bilateral ankle arthrodesis. The second primary arthrodesis in these patients was excluded from analysis. 12 patients had less than 6 months’ interval between the primary arthrodesis and the re-arthrodesis and were excluded. Thus 1,716 primary ankle arthrodeses (and patients) were included in this study.

© 2018 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/4.0) DOI 10.1080/17453674.2018.1488208

Acta Orthopaedica 2018; 89 (5): 560–564

561

Table 1. Distribution of primary arthrodeses by sex, diagnosis, and method of fixation. Values are number of cases

Diagnosis

Post- Neuro- traumatic logical Method of fixation Total Men Women OA arthritis RA Diabetes disorders Other

All 1,716 935 781 460 729 213 39 66 209 Open screw fixation 888 496 392 288 430 59 2 21 88 Intramedullary nail 473 249 224 76 145 115 25 37 75 External fixation 48 29 19 6 20 1 7 1 13 Plate and screws 87 53 34 22 37 3 5 5 15 Arthroscopic screws 183 93 90 66 90 11 2 14 Percutaneous screws 23 7 16 3 18 2 Other methods a 14 8 6 2 4 6 2 a The

group of “Other methods” consisted of 5 percutaneous intramedullary retrograde nailings, 3 arthroscopic intramedullary retrograde nailings, 1 arthroscopic external fixation, 1 fixation by staples, and 4 with missing data.

Table 2. Surgical methods in other diagnoses. Values are number of cases

Post- Talar Unspecified Hemo- Congenital Acquired infectious Psoriatic osteo- secondary chroma- Osteo Methods of fixation Total deformities deformities arthritis arthritis necrosis osteoarthritis Instability tosis chondritis Other a

All 209 36 27 31 14 13 16 10 8 7 47 Open screw fixation 88 13 12 14 5 3 10 6 2 6 17 Intramedullary nail 75 17 14 9 8 5 4 2 4 12 External fixation 13 3 1 1 4 4 Plate and screws 15 3 4 1 1 1 1 4 Arthroscopic screws 13 3 1 1 1 1 6 Percutaneous screws 2 2 Other methods 3 1 2 a

Rare and occasional diagnoses

We examined the registry records for any re-arthrodesis procedures until December 2016 and estimated the rate of revision (irrespective of follow-up time) by taking the number of first revisions divided by the number of primary procedures undertaken. At 2.5 years of follow-up 93% of all re-arthrodeses had been undertaken and this time-point was chosen for risk estimation. We also analyzed the association between re-arthrodesis occurrence and sex, diagnosis, and surgical method.

Statistics We used chi-square test to compare risk of re-arthrodesis related to sex, different surgical methods, and diagnoses. We considered a p-value < 0.05 as statistically significant. Additionally, we undertook a survival analysis (revision-free survival following the primary procedure) of different methods using Kaplan–Meier curves and the log-rank test.

Funding and potential conflicts of interest The Swedish Ankle Registry is partly maintained by funds from the Swedish Association of Local Authorities and Regions. The authors received no financial support for the

research, authorship, and publication of this article. No conflicts of interest declared.

Results Of the 1,716 patients included, 935 were men (median age 50 years [15–87]) and 781 women (median age 53 years [15– 91]). There were 540 tibio-talocrural (TTC) primary arthrodeses, 485 (90%) of these performed by retrograde intramedullary nailing. The mean follow-up time was 4.0 years (1.0–9.0).

Posttraumatic arthritis and open screw fixation were the most common diagnosis and treatment, respectively (Table 1).

Congenital deformities were the most common entity among “other diagnoses” (Table 2). Information concerning bone grafting was found in 1,295 of the 1,510 cases using open techniques (in 276 cases no bone grafting was used, in 678 harvested from the ankle, in 125 from the tibia, in 100 from the iliac crest, and in 16 allograft was used).

We identified a re-arthrodesis in 134 of the 1,716 (7.8%) ankles with a primary arthrodesis (Tables 3 and 4). The median

562

Acta Orthopaedica 2018; 89 (5): 560–564

Table 3. Distribution of re-arthrodeses at 2.5-year follow-up by sex, diagnosis, and method of fixation. Values are number of cases

Diagnosis

Post- traumatic Septic Method of fixation Total Men Women OA arthritis RA arthritis Other

All 101 59 42 31 52 5 1 12 Open screw fixation 59 34 25 17 32 3 7 Intramedullary nail 19 16 3 5 9 1 4 External fixation 2 2 2 Plate and screws 2 2 1 1 Arthroscopic screws 18 6 12 8 9 1 Percutaneous screws 1 1 1 Other methods 0 Survival analysis of different methods of fixation.

Table 4. Total number of arthrodeses and re-arthrodeses, and rate of re-arthrodeses per primary method of fixation

Patients Re-arthro Patients with re- desis Years of Method of fixation at risk, arthrodesis, rate, follow-up

n n % mean (range)

All 1,716 134 7.8 4.0 (1.0–9.0) Open screw fixation 888 77 8.6 4.1 (1.0–9.0) Intramedullary nail 473 25 5.3 4.6 (1.0–9.0) Plate and screws 87 3 3.4 3.7 (1.0–6.8) External fixation 48 3 6.3 4.6 (1.2–8.9) Arthroscopic screws 183 25 14.0 3.5 (1.0–8.8) Percutaneous screws 23 1 4.3 5.7 (1.6–8.8) Other methods 14 0 0 4.0 (1.6–8.8)

Table 6. Life table: number at risk at the beginning of each interval

Years from surgery Method of fixation 0 1 2 3 4 5 6 7

Arthroscopic screws 182 140 96 69 41 23 10 2 Intramedullary nail 467 424 345 264 188 125 84 50 Open screw fixation 884 772 586 424 286 180 107 53 Plate and screws 86 71 55 41 27 14 4 0

The overall 2.5-year risk of re-arthrodesis was 7.4% but risks differed depending on fixation method. Arthroscopic screw fixation was associated with higher risk compared with open surgery with screws, plate and screws, and retrograde intramedullary nailing (Table 5). time between primary surgery and re-arthrodesis was 1.4 (0.5– Survival analysis of primary ankle arthrodesis with re6.9) years. 8 (5%) patients underwent a second re-arthrodesis. arthrodesis as end-point showed the lower survival rate of We found similar rates of re-arthrodesis in men (8.1%) and re-arthrodesis for arthroscopic screw fixation compared with women (7.4%) (p = 0.7), in cases with (7.2%) and without open screw fixation, plate and screws, and retrograde intrabone grafting (6.4%) (p = 0.6), and in cases with primary talo- medullary nailing (log-rank p < 0.001) (Figure, Table 6). crural arthrodesis (TC) (8.6%) and TTC (6.1%) (p = 0.8).

In patients with rheumatoid arthritis the rate of re-arthrodesis was 8.6% in those who had undergone TCs and 6.1% in those who had undergone TTCs. Table 5. Patients with primary arthrodesis with follow-up for 2.5 years, risk of reWe found a higher risk for re-arthrodesis at 2.5 arthodesis, and RR (risk ratio) for revision (cases) by primary method compared with arthroscopic screws years in patients with idiopathic (RR 3.3 (95%

CI 1.3–8.4)) or post-traumatic osteoarthritis

(RR 3.3 (1.3–8.2)) than in patients with rheu Patients with

follow-up for Non- matoid arthritis (4.2%).

Method of fixation 2.5 years Cases cases Risk (%) Risk ratio (95% CI) a

All 1,385 101 1,284 7.4 Open screw fixation 716 59 657 8.2 0.55 (0.34–0.91) Intramedullary nail 407 19 388 4.7 0.31 (0.17–0.58) Plate and screws 64 2 62 3.1 0.21 (0.05–0.88) External fixation 42 2 40 4.8 0.32 (0.07–1.3) Arthroscopic screws 121 18 121 14.9 Ref. Percutaneous screws 21 1 20 4.8 0.32 (0.05–2.3) Other methods 14 0 14 0 0

a Compared with fixation with arthroscopic screws (Ref.).

Discussion In this study from the National Swedish Ankle Registry the overall re-arthrodesis risk at 2.5 years was 7.4% but varied depending on primary technique; substantially, arthroscopic ankle arthrodesis with screws was associated

Acta Orthopaedica 2018; 89 (5): 560–564

with a higher risk for re-arthrodesis than open technique with screws, plate and screws, and retrograde intramedullary nailing.

When comparing the results of screw arthrodesis with open and with arthroscopic technique, Yasui et al. (2016b) in 8,474 patients could not detect any differences between these techniques concerning re-arthrodesis rate. Comparison of arthroscopic screw fixation versus open screw fixation was also studied by Myerson and Quill (1991), Nielsen et al. (2008), and Quayle et al. (2018). They found non-union rates of 2–6% with the arthroscopic technique and of 0–17% with the open technique. Quayle et al. (2018) reported a re-arthrodesis rate of 3% with open technique and 0% with arthroscopic technique.

Re-arthrodesis rates are usually somewhat lower than nonunion rates since some patients with a non-union do not need, want, or are unable to undergo reoperation (Collman et al.

2006, Nielsen et al. 2008, Abicht and Roukis 2013, Chayalon et al. 2015, Quayle et al. 2018).

Results of arthroscopic ankle arthrodesis have been evaluated in several reports with follow-up of mean 3 (1–5.5) years and non-union rates of 0–13% in series of 23–104 patients (Saragas 2004, Winson et al. 2005, Collman et al. 2006, Gougoulias et al. 2007, Nielsen et al. 2008, Dannawi et al. 2011, Townshend et al. 2013, Jain et al. 2016, Duan et al. 2017, Kolodziej et al. 2017). A re-arthrodesis rate of 7% was found in 8,474 ankles by Yasui et.al. (2016b). The re-arthrodesis rate varies between other smaller series (Collman et al. 2006, Gougoulias et al. 2007, Nielsen et al. 2008, Dannawi et al. 2011, Townshend et al. 2013, Jain et al. 2016, Kolodziej et al. 2017).

The re-arthrodesis rate (2.5-year risk 15%) for arthroscopic technique was in our study considerably higher than in most of the above-mentioned reports. We speculate that this may partly be because many of these procedures are performed by general foot and ankle surgeons rather than surgeons specially trained in arthroscopic techniques. The 183 arthroscopic ankle fusions in our study were performed by 31 different surgeons and of these 13 surgeons had performed only 1 procedure. We found only 4 surgeons that had performed more than 10 arthroscopic arthrodeses. These findings indicate that arthroscopic ankle fusions are more demanding, have a longer learning curve, and require a higher volume for successful outcome.

The re-arthrodesis rate with open screw technique (2.5-year risk 8%) in our study is in the same range as other studies that report rates of 3–17% in cohorts of 30–49 patients and with the use of 2 or 3 screws for fixation (Maurer et al.1991, Moeckel et al. 1991, Nielsen et al. 2008, Townshend et al.

2013). With a 4-screw technique, Zwipp et al. (2010) found a re-arthrodesis rate of only 1 in 72 patients. The Swedish Ankle Registry does not include data on the number of screws that have been used in the open screw technique.

The revision rate (2.5-year risk 3.1%) after plate fixation in our study is consistent with other reports in the literature.

Prissel et al. (2017) had 2 non-unions with a single-column anterior plate in 47 patients. Using anterior double plating

563

Plaass et al. (2009) found no non-unions or re-arthrodeses in 29 patients. The Swedish Ankle Registry does not collect data on which plate or how many screws are used in a specific case.

Maurer et al. (1991) and Moeckel et al. (1991) compared the external fixation technique with open screw fixation and found non-union rates of 17% and 21% when external fixation was used and 0% and 5% with the open screw technique and re-arthrodesis rates of 17% and 18% in the external fixation groups. However, these reports are old and the 4.8% revision risk at 2.5 years in our study following external fixation is probably the result of a more modern and refined external fixation technique.

Retrograde intramedullary nailing is often used in more difficult cases with, for example, diabetic arthropathy, neuropathy, large deformities, or rheumatoid arthritis. In our study with 25 re-arthrodeses (5%) out of 473 arthrodesis procedures with intramedullary nailing, half of the patients with re-arthrodesis were in these categories. The 2.5-year risk of revision following fixation with intramedullary nailing was 4.8%. Non-union rates below 10% in series with such diagnoses are reported by Fazal et al. (2006), Chettiar et al. (2011), Lucas y Hernandez et al. (2015), Thomas et al. (2015), Richter and Zech (2016).

A 12% non-union rate was found in a study by Pelton et al.

(2006). Re-arthrodesis rates of 0–5% were found by Pelton et al. (2006), Fazal et al. (2006), Lucas y Hernandez et al.

(2015), Thomas et al. (2015), and Richter and Zech (2016).

Fenton et al. (2014) demonstrated many complications with retrograde intramedullary nailing and had 9 non-unions out of 55 ankles; 3 of their 52 patients with postoperative deep infection underwent below-knee amputations. Fazal et al. (2006) also reported 1 below-knee amputation after deep infection. In the present data from the Swedish Ankle Registry no amputations were reported.

Patients with rheumatoid arthritis had an overall lower rearthrodesis rate than other diagnoses in our study. Reports concerning ankle fusion in rheumatoid arthritis are sparse. The re-arthrodesis rate of 4% in our study (213 primary fusions) is similar to other studies. Fujimori et al. (1999) found no nonunions in 19 patients and Anderson et al. (2005a) found a 4% non-union rate in 26 patients. These studies used retrograde intramedullary nails; in our study more than half of the cases with rheumatoid arthritis underwent this procedure (Table 1). With an open screw technique Anderson et al. (2005b) in 22 rheumatoid patients found non-union in 7 patients and 5 patients had a re-arthrodesis. The authors discouraged this technique in this patient group. In our study the re-arthrodesis rate following open screw fixation was 5% in patients with rheumatoid arthritis.

Diabetes is considered a risk factor for non-union in ankle arthrodesis (Thevendran et al. 2012). Perlman and Thordarson (1999) found a high rate of non-union in diabetic patients while Chalayon et al. (2015) and Jain et al. (2016) were unable to identify any increased risk. In our study there were no rearthrodeses in the 39 patients with diabetes.

564

Our study has several limitations. In a registry study, there is always a risk of incomplete reporting. However, the Swedish Ankle Registry has a coverage of 96%. Furthermore, certain details, such as the number of screws and which plates were used, are not presently documented in the registry. The strength is that a registry, especially a nationwide one, covers many procedures, centers, and surgeons of varying experience and proficiency, giving a reasonable depiction of real-life patient care and results.

In summary, in this study from the national Swedish Ankle Registry the 2.5-year re-arthrodesis risk was 7% but varied depending on primary technique. Arthroscopic fusions had a high re-arthrodesis risk (15%). Reasons for this are unknown but poor surgical technique and/or surgeon inexperience may contribute, as may patient selection. Arthroscopic ankle fusion may also be more demanding, have a longer learning curve, and require a higher volume for successful outcome.

Design of study: AH, ÅC, BR, data collection: AH, ÅC, statistics: LJ, data interpretation: all authors revised the manuscript.

Acta thanks Markus Knupp and Cees Verheyen reviewers for help with peer review of this study.

Abicht B P, Roukis TS . Incidence of nonunion after isolated arthroscopic ankle arthrodesis. Arthroscopy 2013; 29(5): 549-54.

Anderson T, Linder L, Rydholm U, Montgomery F, Besjakov J, Carlsson Å. Tibio-calcaneal arthrodesis as a primary procedure using a retrograde intramedullary nail: a retrospective study of 26 patients with rheumatoid arthritis. Acta Orthop 2005a;76(4):580-7.

Anderson T, Maxander P, Rydholm U, Besjakov J, Carlsson Å. Ankle arthrodesis by compression screws in rheumatoid arthritis: primary nonunion in 9/35 patients. Acta Orthop 2005b; 76(6): 884-90.

Chalayon O, Wang B, Blankenhorn B, Jackson JB, Beals T, Nickisch F, Saltzman CL. Factors affecting the outcome of uncomplicated primary ankle arthrodesis. Foot Ankle Int 2015; 36(10): 1170-9.

Chettiar K, Halder S, Bowman N, Cottam H, Armitage A, Skyrme A. Tibio-talo-calcaneo fusion using a locked intramedullary nail. Foot Ankle Surg 2011; 17: 228-32.

Collman D R, Kaas M H, Schuberth J M. Arthroscopic ankle arthrodesis: factors influencing union in 39 consecutive patients. Foot Ankle Int 2006; 27(12): 1079-85.

Dannawi Z, Nawabi D H, Patel A, Leong J J H, Moore D J. Arthroscopic ankle arthrodesis: are results reproducible irrespective of per-operative deformity? Foot Ankle Surg 2011; 17: 294-9.

Duan Z, Yang L, Yin L. Arthroscopic arthrodesis for ankle arthritis without bone graft. J Orthop Surg Res 2017; 11: 154-61.

Fazal M A, Garrido E, Williams R L. Tibio-talar-calcaneal arthrodesis by retrograde intramedullary nail and bone grafting. Foot Ankle Surg 2006; 12: 185-90.

Fenton P, Bali N, Matheshwari R, Youssef B, Meda K. Complications of tibiotalar-calcaneal fusion using intramedullary nails. Foot Ankle Surg 2014; 20: 268-71.

Fujimori J, Yoshino S, Koiwa M, Nakamura H Shiga H, Nagashima S. Ankle arthrodesis in rheumatoid arthritis using an intramedullary nail with fins.

Foot Ankle Int 1999; 20: 485-90.

Gougoulias N E, Agathangelidis F G, Parsons S W. Arthroscopic ankle arthrodesis. Foot Ankle Int 2007; 28(6): 695-706.

Acta Orthopaedica 2018; 89 (5): 560–564

Jain S K, Tiernan D, Kearns S R. Analysis of risk factors for failure of arthroscopic ankle fusion in a series of 52 ankles. Foot Ankle Surg 2016; 22: 91-6.

Kolodziej L, Sadlik B, Sokolowski S, Bohatyrewicz A. Results of arthroscopic arthrodesis with fixation using two parallel headless compression screws in a heterogenic group of patients. Open Orthop J 2017; 11: 37-44.

Lucas y Hernandez J, Abad J, Remy S, Darcel V, Chauveaux D, Laffenetre O. Tibiotalocalcaneal arthrodesis using a straight intramedullary nail. Foot Ankle Int 2015; 36(5): 539-46.

Maurer R C, Cimino W R, Cox C V, Satow G K. Transarticular cross-screw fixation: a technique of ankle arthrodesis. Clin Orthop Relat Res 1991; 268: 56-64.

Moeckel B H, Patterson B M, Inglis A E, Sculco T P. Ankle arthrodesis: a comparison of internal and external fixation. Clin Ortop Relat Res 1991; 268: 78-83.

Myerson M S, Quill G. Ankle arthrodesis: a comparison of an arthroscopic and an open method of treatment. Clin Orthop Relat Res 1991; (268): 84-95.

Nielsen K K, Linde F, Jensen N C. The outcome of arthroscopic and open surgery ankle arthrodesis: a comparative retrospective study on 107 patients.

Foot Ankle Surg 2008; 14: 153-7.

Nihal A, Gellman R E, Embil J M, Trepman E. Ankle arthrodesis. Foot Ankle Surg 2008; 14: 1-10.

Pelton K, Hofer J K, Thordarson D B. Tibiocalcaneal 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): 491-7.

Plaass C, Knupp M, Barg A, Hintermann B. Anterior double plating for rigid fixation of isolated tibiotalar arthrodesis. Foot Ankle Int 2009; 30(7): 631-9.

Prissel M A, Simpson G A, Sutphen S A, Hyer C F, Berlet G C. Ankle arthrodesis: a retrospective analysis comparing single column, locked anterior plating to crossed lag screw technique. J Foot Ankle Surg 2017; 56(3): 453-6.

Quayle J, Shafaty R, Khan M A, Ghosh K, Sakellariou A, Gougoulias N. Arthroscopic versus open ankle arthrodesis. Foot Ankle Surg 2018; 24: 137-42.

Richter M, Zech S. Tibiotalocalcaneal arthrodesis with a triple-bend intramedullary nail (A3)-2-year follow-up in 60 patients. Foot Ankle Surg 2016; 22: 131-8.

Saragas N P. Results of arthroscopic arthrodesis of the ankle. Foot Ankle Surg 2004; 10: 141-3.

Thevendran G, Younger A, Pinney S. Current concepts review: risk factors for nonunions in foot and ankle arthrodesis. Foot Ankle Int 2012; 33(11); 1031-40.

Thomas A E, Guyver P M, Taylor J M, Czipri M, Talbot N J, Sharpe J T. Tibiotalocalcaneal arthrodesis with a compressive retrograde nail: a retrospective study of 59 nails. Foot Ankle Surg 2015; 21: 202-5.

Townshend D, Di Silvestro M, Krause F, Penner M, Younger A, Glazebrook M, Wing K. Arthroscopic versus open ankle arthrodesis: a multicenter comparative case series. J Bone Joint Surg (Am) 2013; 95-A(2): 98-102.

Winson I G, Robinson D E, Allen P E. Arthroscopic ankle arthrodesis. J Bone Joint Surg (Br) 2005; 87(3): 343-7.

Yasui Y, Hannon C P, Seow D, Kennedy J G. Ankle arthrodesis: a systematic approach and review of the literature. World J Orthop 2016a; 7(11): 700-8.

Yasui Y, Vig K S, Murawski C D, Desai P, Savage-Eliott I, Kennedy J G. Open versus arthroscopic ankle arthrodesis: a comparison of subsequent procedures in a large database. J Foot Ankle Surg 2016b; 55: 777-81.

Zwipp H, Rammelt S, Endres T, Heineck J. High union rates and function scores at midterm followup with ankle arthrodesis using a four screw technique. Clin Orthop Relat Res 2010; 468: 958-68.

Acta Orthopaedica 2018; 89 (5): 565–569

565

Ten cold clubfeet Robert B GIESBERTS 1, Edsko E G HEKMAN 1, Gijsbertus J VERKERKE 1,2, and Patrick G M MAATHUIS 3 1 University of Twente, Department of Biomechanical Engineering, Enschede; 2 University of Groningen, University Medical Center Groningen, Department of Rehabilitation Medicine, Groningen; 3 University of Groningen, University Medical Center Groningen, Department of Orthopaedic Surgery, Groningen, the Netherlands Correspondence: r.b.giesberts@utwente.nl Submitted 2018-03-08. Accepted 2018-05-28.

Background and purpose — Idiopathic clubfeet are commonly treated with serial manipulation and casting, known as the Ponseti method. The use of Plaster of Paris as casting material causes both exothermic and endothermic reactions.

The resulting temperature changes can create discomfort for patients.

Patients and methods — In 10 patients, we used a digital thermometer with a data logger to measure belowcast temperatures to create a thermal profile of the treatment process.

Results — After the anticipated temperature peak, a surprisingly large dip was observed (Tmin = 26°C) that lasted 12 hours.

Interpretation — Evaporation of excess water from a cast might be a cause for discomfort for clubfoot patients and subsequently, their caregivers.

The common treatment of idiopathic clubfoot (talipes equinovarus) consists of serial manipulation and casting, known as the Ponseti method (Ponseti 2008). The treatment is started in the first weeks after birth and includes on average 5 cast changes, often followed by a percutaneous Achilles tenotomy.

The deformity is successfully corrected in over 90% of all cases (Morcuende et al. 2004). An abduction orthosis is worn for several years to prevent relapse (Dobbs et al. 2004).

The Ponseti method dictates toe-to-groin casts with the knee flexed to prevent slipping of the cast (Ponseti 2008, Maripuri et al. 2013). Plaster of Paris (PoP) is the accepted casting material, because it is inexpensive, easily obtained, and it can be easily molded (Aydin et al. 2015, Pittner et al. 2008). The use of PoP as casting material involves both an exothermic (hot) setting reaction and an endothermic (cold) drying process.

Bandage rolls impregnated with PoP are soaked in lukewarm water before being wrapped around the foot. The PoP has an exothermic chemical reaction with water. Many studies have focused on the extent of the generated heat and the effects of the associated temperature peak, which appears minutes after application (Shuler and Bates 2013, Burghardt et al. 2014).

Thermal injuries can occur if the temperature exceeds 40°C for longer periods of time, but can be prevented by optimizing the number of layers used, cast type and brand, dipping water temperature, and cast padding thickness (Halanski et al. 2007, Hutchinson and Hutchinson 2008, Shuler and Bates 2013).

Only a few layers of cast tape are needed to keep the tiny clubfeet in position, therefore thermal injuries are uncommon in the Ponseti method.

Evaporation of the excess water in the cast is an endothermic process, meaning that it extracts energy from its surroundings, thereby cooling it. The required energy for the evaporation of the excess water—the latent heat—is drawn in part from the patient. The effect of this endothermic reaction in the treatment of clubfoot has not been studied before. However, anecdotal information from parents suggests that some children struggle to keep warm with the wet and cold casts (Bridgens and Kiely 2010, van Doorn 2016).

We assessed the below-cast temperature profile in the treatment of clubfoot with the Ponseti method, in terms of both temperature drop and the extent of this drop.

Patients and methods Children 10 children with idiopathic clubfoot, who were younger than 3 months old, and who had had no form of prior treatment were selected for this study (Table 1). From the children who

© 2018 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/4.0) DOI 10.1080/17453674.2018.1493046

566

Acta Orthopaedica 2018; 89 (5): 565–569

Table 1. Child characteristics. Unless stated otherwise values are presented as mean (SD)

Skin Coban tape

Characteristic Children Age (range) Boys/girls Bilateral Pre-treatment Pirani score Pre-treatment Diméglio score

Value

Thermometer

10 7 (2–30) days 9/1 6 4.3 (0.7) 14.0 (1.6)

presented with bilateral clubfoot only 1 foot was included for measurements.

Sensors A 1-wire DS1825 digital thermometer (Maxim Integrated, San Jose, CA, USA) was used to measure the temperature under the cast. The accuracy of the sensor is ± 0.5°C (Maxim Integrated 2005).

Protocol After manipulation of the foot, the orthopedist placed the thermometer on the plantar aspect of the foot. A strip of Coban tape (3M, Neuss, Germany) was used between the skin and the thermometer and wires (Figure 1). The thermometer was connected to an electronic data logging system with a small battery and SD card, enclosed in 3D-printed casing. After placing the sensor, the foot was treated as usual following the Ponseti method.

The cast technician cut 1 roll of Gypsona (BSN Medical GmbH, Hamburg, Germany) in half, dipped it in a sink filled with lukewarm (20–25°C) water and applied it to the clubfoot while the orthopedist maintained the position of the foot. Care was taken not to use more than 3 layers of cast tape. Wires were routed distally and, as soon as the plaster had set, the casing was attached to the cast laterally with a single layer of Peha-haft bandage (Hartmann, Heidenheim, Germany) (Figure 1a). Parents were instructed to elevate the legs of the child with a rolled-up towel, as protocol directs.

At the next scheduled weekly meeting the cast and thermometer were removed. While the parents bathed their child, the sensors were left in a stable position for at least 15 minutes for calibration after which the data were copied to a computer and the battery replaced. The foot was carefully inspected for any signs of skin damage and the protocol was repeated until the final cast before tenotomy.

Water content The same measurement protocol was used to cast 1 rubber clubfoot model (MD Orthopaedics, Inc., Wayland, IA, USA).

The weight of all components was documented and the total weight of the rubber foot plus cast was monitored for several days. This provided an estimation of the excess water content in the plaster cast and the time it takes to evaporate.

Padding Plaster cast

Data logger Thermometer

(a)

(b)

Peha-haft

Figure 1. Thermometer location. (a) Location of the thermometer on the foot, below the cast. The plaster is represented in blue. The casing for the data logging system was attached to the cast with a layer of Peha-haft bandage. (b) Schematic representation of the layers between the skin and the thermometer.

The latent heat of evaporation of water is described as Q = m·∆Hev, with m the measured mass of the excess water and ∆Hev(25 °C) = 2,442 J/g (Rajput 2010).

Data collection and data processing The system was programmed to store time and temperature data continuously (0.16 Hz) for the first 4 hours, after which the system entered a low power state (85 µA) to save battery power. 4 times per hour it woke from this state to collect data for 10 seconds. Each 10 s period resulted in 2 temperature measurements over which the average was taken. The system featured a button to wake up after the cast was removed.

Data was processed using Matlab version R2016b. Potential sensor errors were filtered using the median filter medfilt1 (Pratt 2007). The start of the measurement (t0) was defined as the moment the plaster cast touched the foot. The height of the temperature peak (Tmax) was calculated as the maximum value over the first 3 hours, the final equilibrium temperature (Tend) as the mean value over the final 3 days of measurement, and the lowest temperature (Tmin) as the minimum value after Tmax.

The beginning of the temperature dip was defined as the moment the temperature reached a value smaller than 1 standard deviation of Tend. The moment the temperature reached a value within 1 standard deviation of Tend was defined as the end of the temperature dip.

The 3 sets of temperature data (Tmax, Tmin and Tend) were tested for normality using the Shapiro-Wilk test. The Student t-tests were used to test the difference between temperatures for statistical significance.

Ethics, funding, and potential conflicts of interest The medical ethical evaluation committee of the UMCG reviewed the study in accordance with the declaration of Helsinki, and declared that it did not meet the criteria as stated by the Medical Research Involving Human Subjects Act (WMO), and therefore did not require their approval (document number M16.196266).

Acta Orthopaedica 2018; 89 (5): 565–569

567

Table 2. Molecular ratios CaSO4 (H2O)1/2 + 1½H2O Molecular mass (mol-1) Mass (mol) Mass (g)

→ CaSO4(H2O)2 + H2O + heat

141.15 g 1½×18.015 g 172.17 g 18.015 g 1,433 kJ 0.52 0.78 0.52 Excess a 0.52 75 14 89 21 740 kJ

a The

21 g of excess water measured fits the prediction based on the molecular ratios of the exothermic setting reaction.

Temperature (°C)

Temperature (°C)

40

40

35

Tmax

Tend

Tmax > Tmin (p < 0.001)

Tmin < Tend (p < 0.001)

35

SD 30

30

Tmin 25

25

dip duration 20

t0

12

24

Time (h)

36

20

48

Figure 2. Results from the temperature measurements. The orange line indicates the average temperature over all measurements, the blue area the standard deviation (SD).

Table 3. Temperature data

Tmax Tmin Tend

Time (hours) median (IQR) 0.15 (0.1–0.8) 4.9 (4.0–7.3) 11.0 (9.6–15.8)

a Temperature

Temperature (°C) mean (SD) [95%CI] 35.3 (1.6) [34.7–35.8] 26.1 (3.0) [25.0–27.2] 34.9 (1.4) [34.4–35.5]

Normality p-value

a

0.5 0.2 0.09

data was tested for normality using Shapiro–Wilk tests.

After obtaining informed written consent from both parents, the children were included in the study. Financial support was provided by Stichting voor de Technische Wetenschappen [grant P12-03]. No competing interests declared.

Results Complications in the form of pressure marks were encountered in 4 cases. In 2 of those cases further measurements were canceled but the data obtained were included for analysis.

Technical malfunctioning occurred in 5 measurements (disconnected sensors, disconnected thermometer, faulty measurement protocol, empty battery). 29 successful measure-

Tmax

Tmin

Tend

Figure 3. Difference in the measured temperature. Tmax represents the temperature peak shortly after casting, Tmin the lowest temperature reached after the initial peak, and Tend the equilibrium temperature.

ments were performed on 10 clubfeet, giving more than 3,000 hours of data.

Clubfoot model—water content The measurement with a clubfoot model showed that approximately 35 g of water was used for one roll of PoP bandage of 75 g. Over the course of 2 days the water content in the cast decreased to reach an equilibrium of 14 g. This means that 14 g of the original water content attached to the PoP to form hard gypsum and the remaining 21 g of excess water evaporated. This ratio fits well within the molecular ratio of the chemical reaction, which also predicts the generation of 740 kJ of heat (Table 2). The required energy for evaporation of 21 g of excess water is 51 kJ. Clubfoot measurements All temperature measurements showed a clear temperature peak after the application of the plaster cast (Tmax). None of the below-cast temperatures ever exceeded 40°C and no thermal injuries were observed (Figure 2 and Table 3).

After the peak, a decrease to a lower temperature was observed (p < 0.001) (Figure 3). On average, this temperature dip lasted 12 (range 3.2–30) hours followed by a gradual increase to a stable final temperature (Tend).

568

Acta Orthopaedica 2018; 89 (5): 565–569

Figure 4. Gypsona temperature profile. Measurements from BSN Medical show the steep increase and subsequent gradual decline of the temperature after application. Copied with permission from BSN Medical.

All 3 sets of temperature data were normally distributed (Table 3). There was no statistical difference between the temperature peak and the equilibrium temperature (p = 0.3).

Discussion This is the first study to present long-term temperature measurements underneath the cast in the treatment of clubfeet. A surprisingly large temperature dip was observed caused by the evaporation of excess water. In several cases temperatures as low as 22°C were observed and they stayed low for more than 12 hours.

Evaluation of the results Our measurements showed a temperature peak after 9 minutes (Table 3). BSN Medical’s own measurements with Gypsona show a peak after 12 minutes (Figure 4). Burghardt et al.

(2014) found a peak temperature after 9 minutes when using 16 layers of a different brand of plaster.

Immediately after reaching the peak, the cast started to cool. This is also visible in BSN Medical’s measurement in which the temperature drops below the ambient temperature after approximately 1 hour (Figure 4). Depending on environmental influences it can take 2 to 3 days for the cast to dry completely (Hayter 2015). Indeed, in our water content measurement the cast reached a stable weight after 2 days. However, on the clubfeet it took an average of 11 hours before the temperature reached normal levels again. This suggests that the plaster dried much quicker when placed on the child’s foot and that some of the child’s energy was used to evaporate the water of the plaster cast.

Casting-induced hypothermia In adults, 1 case study exists which describes observed hypothermia after the application of a plaster jacket for immobilization of the cervical spine (Vale 1969). Caregivers of children with a clubfoot have mentioned that their children can be a bit prickly in the first 24 h after a new cast is applied (Bridgens

and Kiely 2010, van Doorn 2016). The observed cold period might be one of the causes for this behavior.

It would take 51 kJ to evaporate the excess water of 1 cast.

Part of this energy will be extracted from the environment, but a substantial part of it must be produced by the child. However, the thermoregulatory system of newborns is not yet fully developed, and they experience great difficulties when facing a cold environmental challenge (Pierro et al. 2012). Moreover, newborns have a relatively large surface area, so much heat is lost in relation to their heat-producing volume.

A newborn’s total energy expenditure in the first month after birth is 1.0–1.3 MJ/day (FAO 2001, Olgaher and Forsum 2003). Up to 8% of this energy—approximately 100 kJ/day— is used for thermogenesis and thermoregulation (Pierro et al.

2012). Relatively, the required 51 kJ for the evaporation of excess water is quite substantial and a mild form of castinginduced hypothermia in the treatment of clubfoot seems plausible, especially with bilateral clubfeet in vulnerable patients such as prematurely born children. In those cases, postponing treatment at least until after the first month might be beneficial (Alves et al. 2009, Awang et al. 2014, Íltar et al. 2010).

Given that newborns are naturally well monitored by their parents, that hypothermia in newborns is easily detected, and that no such case studies exist for clubfoot, it is unlikely that casting induces severe or moderate hypothermia. Still, having cold feet for half a day undoubtedly is uncomfortable and plausibly affects the newborn’s mood.

Limitations We acknowledge limitations of our study. First, we did not measure whether the wet and cold casts affect the child’s core temperature. If they do, a more prominent effect would be expected in bilateral cases. However, the number of children in our study does not allow for any meaningful test for subgroup differences, and no trend could be identified.

Second, the sensors were placed on the skin on the plantar aspect of the foot. So the measurements reflect only the temperature of that part of the foot and not the core temperature of the patient. Hypothermia was neither measured

Acta Orthopaedica 2018; 89 (5): 565–569

nor diagnosed based on clinical symptoms. Future research should include measuring the core temperature during the first 24 hours after casting to study the possibility of casting-induced hypothermia.

Clinical implications So far only anecdotal information exists regarding cold clubfeet (Bridgens and Kiely 2010, van Doorn 2016). In some cases, especially with bilateral clubfeet in vulnerable children, it might be better to postpone treatment to prevent any harmful effects of the wet and cold casts. This is an extra argument to defer treatment in premature babies for several weeks, as stated by the First European consensus meeting on Ponseti clubfoot treatment’s agreement (Böhm and Sinclair 2013).

Preventing a clubfoot from getting cold would be possible if less excess water needed to be evaporated. However, squeezing the water out of the plaster to use less water results in hotter plaster and increased risk of burns: the excess water is needed to release the heat from the exothermic reaction (Kaplan 1981).

Using synthetic cast tape as an alternative casting material could be a viable option to decrease the discomfort of cold feet. The chemical reaction and the structure of the material both cause the drying process to be shorter than when using PoP. In summary, after an anticipated temperature peak, a surprisingly large temperature dip was observed. More research is needed to assess whether the Ponseti method affects the child’s core temperature.

The authors would like to thank the parents of all participants for their trust and patience, and the medical staff of the Gipskamer for all their help and advice. Special thanks are offered to Maaike Vos, Els Huizenga, and Rohan Choudhari for performing the measurements.

RG conceived and designed the experiments, performed the experiments, analyzed the data, and wrote the paper. EH and GV critically edited and reviewed the test protocol, data interpretation, and the manuscript. PM performed the experiments, critically edited and reviewed the test protocol, data interpretation, and the manuscript.

Acta thanks Naomi Davis and Henrik Wallander for help with peer review of this study.

Alves C, Escalda C, Fernandes P, Tavares D, Neves C. Ponseti method: does age at the beginning of treatment make a difference? Clin Orthop Relat Res 2009; 467(5): 1271-7. doi: 10.1007/s11999-008-0698-1 Awang M, Sulaiman AR, Munajat I, Fazliq M E. Influence of age, weight, and Pirani score on the number of castings in the early phase of clubfoot treatment using Ponseti method. Malays J Med Sci 2014; 21(2): 40-3.

Aydin B K, Sofu H, Senaran H, Erkocak O F, Acar M A, Kirac Y. Treatment of clubfoot with Ponseti method using semirigid synthetic Softcast. Medicine 2015; 94(47): e2072. doi: 10.1097/MD.0000000000002072

569

Böhm S, Sinclair M. Report of the 1st European consensus meeting on Ponseti clubfoot treatment. J Child Orthop 2013; 7(3): 251-4. doi: 10.1007/ s11832-013-0497-4 Bridgens J, Kiely N. Current management of clubfoot (congenital talipes equinovarus). BMJ 2010; 340(6): 308-12. c355. doi: 10.1136/bmj.c355 Burghardt R D, Anderson J G, Reed R A, Herzenberg J E. Exothermic properties of plaster-synthetic composite casts. J Child Orthop 2014; 8(2): 193201. doi: 10.1007/s11832-014-0563-6 Dobbs M B, Rudzki J, Purcell D B, Walton T, Porter K R, Gurnett C A. Factors predictive of outcome after use of the Ponseti method for the treatment of idiopathic clubfeet. J Bone Joint Surg Am 2004; 86(1): 22-7.

FAO. Human energy requirements. Report of a joint FAO/WHO/UNU expert consultation. Rome: Food and Agriculture Organization of the United Nations (FAO); 2001.

Halanski M A, Halanski A D, Oza A, Vanderby R, Munoz A, Noonan K J. Thermal injury with contemporary cast-application techniques and methods to circumvent morbidity. J Bone Joint Surg Am 2007; 89(11): 236977. doi: 10.2106/JBJS.F.01208 Hayter F. Casting directives. 2015 [cited 2018 Mar 7]; Available from: http:// slideplayer.com/slide/3475355/ Hutchinson M J, Hutchinson M R. Factors contributing to the temperature beneath plaster or fiberglass cast material. J Orthop Surg Res 2008; 3: 10. doi: 10.1186/1749-799X-3-10 Kaplan S. Burns following application of plaster splint dressings. J Bone Joint Surg Am 1981; 63(4): 670-2. Íltar S, Uysal M, Alemdaroğlu KB, Aydoğan NH, Kara T, Atlıhan D. Treatment of clubfoot with the Ponseti method: should we begin casting in the newborn period or later? J Foot Ankle Surg 2010; 49(5): 426-31. doi: 10.1053/j.jfas.2010.06.010 Maripuri S N, Gallacher P D, Bridgens J, Kuiper J H, Kiely N T. Ponseti casting for club foot above-or below-knee? Bone Joint J 2013; 95(11): 1570-4. doi: 10.1302/0301-620X.95B11.31899 Maxim Integrated. DS1825 Programmable Resolution 1-Wire Digital Thermometer With 4-Bit ID (Rev 020105) [document on the internet]. San Jose, CA: Maxim Integrated; 2005 [cited 2018 May 30]. Available from: https://datasheets.maximintegrated.com/en/ds/DS1825.pdf.

Morcuende J A, Dolan L A, Dietz F R, Ponseti IV. Radical reduction in the rate of extensive corrective surgery for clubfoot using the Ponseti method.

Pediatrics 2004; 113(2): 376-80.

Olgaher A, Forsum E. Total energy expenditure, body composition and weight gain in moderately preterm and full-term infants at term postconceptional age. Acta Paediatr 2003; 92(11): 1327-34. doi: 10.1080/08035250310005396 Pierro A, De Coppi P, Eaton, S. Neonatal physiology and metabolic considerations. In: (Coran A G, Adzick N S, eds.) Pediatric surgery. 7th ed. Philadelphia: Mosby; 2012. p. 89-107. doi: 10.1016/B978-0-323-07255-7.00006-4 Pittner D E, Klingele K E, Beebe A C. Treatment of clubfoot with the Ponseti method: a comparison of casting materials. J Pediatr Orthop 2008; 28(2): 250-3. doi : 10.1097/BPO.0b013e318164f8e7 Ponseti I V. Congenital clubfoot: fundamentals of treatment. 2nd ed. New York: Oxford University Press; 2008.

Pratt WK. Digital image processing. 4th ed. Hoboken, NJ: Wiley; 2007.

Rajput R. Thermal engineering [Internet]. Bengaluru: Laxmi Publications Pvt Ltd; 2010. Available from Google Books (https://books.google.nl/ books?id=65gxCX2dC84C) Shuler F D, Bates CM. Skin temperatures generated following plaster splint application. Orthopedics 2013; 36(5):3 64-7. doi: 10.3928/0147744720130426-06 Vale R. Post-operative accidental hypothermia. Anaesthesia 1969; 24(3): 449-52. van Doorn E. A design for a new clubfoot treatment: the parents’ view. BSc [thesis]. Enschede: University of Twente; 2016 [Cited 2018 Jan 4]. Available from: http://essay.utwente.nl/70558/

570

Acta Orthopaedica 2018; 89 (5): 570–574

Incidence and demographics of giant cell tumor of bone in The Netherlands: First nationwide Pathology Registry Study Arie J VERSCHOOR 1, Judith V M G BOVÉE 2, Monique J L MASTBOOM 3, P D Sander DIJKSTRA 3, Michiel A J VAN DE SANDE 3, and Hans GELDERBLOM 1 1 Department 3 Department

of Medical Oncology, Leiden University Medical Center, Leiden; 2 Department of Pathology, Leiden University Medical Center, Leiden; of Orthopedic Surgery, Leiden University Medical Center, Leiden, The Netherlands Correspondence: a.j.verschoor@lumc.nl Submitted 2018-01-04. Accepted 2018-05-14.

Background and purpose — Giant cell tumors of bone (GCT-B) are rare, locally aggressive tumors characterized by an abundance of giant cells. Incidence studies for GCT-B are rare. This is the first study using a fully automated 100% covering pathology database, the nationwide Dutch Pathology Registry (17 million inhabitants), PALGA, to calculate incidence rates for GCT-B. Patients and methods — From PALGA, all pathology excerpts were retrieved for patients diagnosed with GCT-B, giant cell tumors of tenosynovium, and giant cell tumors of soft tissue between January 1, 2009 and December 31, 2013.

The incidence of GCT-B was calculated.

Results — In total, 8,156 excerpts of 5,922 patients were retrieved; these included 138 first GCT-B diagnosis. For GCT-B the incidence was 1.7 per million inhabitants per year with a male to female ratio of 1:1.38 and a median age of 35 years (9–77). Most common localization was the femur (35%), followed by the tibia (18%). No differences in localization according to age and sex were found. The incidence rate of local recurrence was 0.40 per million inhabitants per year.

Interpretation — This is the first nationwide study reporting the incidence of GCT-B, based on a nationwide pathology database with 100% coverage of pathology departments.

Current incidence calculations are based only on doctordriven registries. We confirmed that GCT-B is a rare disease with an incidence that is slightly higher than previously published. The relatively young median age of patients and the high incidence of recurrence stresses the importance of developing more effective treatments for this disease.

Giant cell tumor of bone (GCT-B) is a locally aggressive neoplasm composed of sheets of mononuclear cells admixed with uniformly distributed large osteoclast-like giant cells, primarily affecting the metaphysis of long bones (Athanasou et al.

2013). These cells express receptors of nuclear factor kappaB ligand (RANKL) (Atkins et al. 2000, Roux et al. 2002).

GCT-B are rare; however, the incidence is not exactly known and is for example not stated in the World Health Organization (WHO) classification of Tumors of Soft Tissue and Bone (Athanasou et al. 2013). The incidence was recently estimated at between 1.03 and 1.33 per million per year based on cancer registries in Australia, Japan and Sweden (Table 1) (Liede et al. 2014, Amelio et al. 2016). Median age of onset ranges between 20 and 40 years with an equal distribution between the sexes or a slight female predominance (Athanasou et al.

2013, Amelio et al. 2016).

Patients with GCT-B typically present with pain, swelling, and often decreased joint movement. In 5–30% of patients a pathologic fracture is noted (Athanasou et al. 2013, van der Heijden 2014b). Although this tumor rarely metastasizes, it is known to be locally aggressive, which may result in joint destruction and, uncommonly, neurological deficit in axial tumors (Athanasou et al 2013). Treatment options are curettage, curettage with an adjuvant treatment, or resection with joint replacement (van der Heijden 2014a). In GCT-B the local recurrence rate is 6–42% (Balke et al. 2008, van der Heijden 2014b). Recently, denosumab, a human IgG2 monoclonal antibody against RANKL, was registered for use in GCT-B and showed tumor response in 2 phase II studies (Thomas et al. 2010, Chawla et al. 2013).

According to our knowledge, current literature on GCT-B contains incidence calculations based solely on cancer registry studies. These studies are doctor-driven with a risk of underreporting (Table 1) (Liede et al. 2014, Amelio et al. 2016). In this study we use the non-profit nationwide network and reg-

© 2018 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/4.0) DOI 10.1080/17453674.2018.1490987

Acta Orthopaedica 2018; 89 (5): 570–574

571

Table 1. Review of all available incidence data on GCT-B Incidence

per million Article Country Type inhabitants Liede et al. 2014 a Amelio et al. 2016 a Current study

Sweden, Australia, Japan Sweden The Netherlands

Age Percentage median (range) men

Doctor-driven

1.03–1.33

20–40 (na)

na

Doctor-driven Nationwide pathology registry

1.3 1.66

34 (10–88) 35 (9–77)

48 42

na: not available. The study by Liede et al. does not report an exact median age, but only a median age group. a These studies probably also included patients with a giant cell tumor of the small bones of hands or feet and patients with central giant cell granulomas of the jaw.

istry of histo- and cytopathology in the Netherlands, PALGA. This fully automated nationwide database contains all pathology reports in the Netherlands (17 million inhabitants) (Casparie et al. 2007). In an effort not to miss GCT-B cases, our search included the following giant cell containing tumors: GCT-B, tenosynovial giant cell tumors, and giant cell tumors of soft tissue.

We calculated the incidence, demographics, and localizations of GCT-B in a nationwide pathology database study between January 1, 2009 and December 31, 2013.

Patients and methods Patients PALGA covers all pathology reports of all pathology laboratories in the Netherlands since 1993 (Casparie et al. 2007).

Patient registration in PALGA is based on social service number and thereby multiple reports of one patient will be grouped and not lead to double registration of one patient.

Excerpts matching our search criteria were retrieved from PALGA, encoded either as giant cell tumor of bone (PALGA code m9250*) or as giant cell tumor of tenosynovium (m9252*) or pigmented villonodular synovitis (m9252*) or giant cell tumor of soft tissue (m9251*) and terms separately used as free text between January 1, 2009 and December 31, 2013 (Casparie et al. 2007). In our search, giant cell tumors of soft tissue and tenosynovium were included, to be as comprehensive as possible. Additionally, for all these patients historical excerpts were retrieved matching our search criteria.

When date of first diagnosis met our 5-year timeframe, the patient was included for incidence calculations. Patients with a giant cell tumor of the small bones of the hands or feet were excluded, since these are considered a separate entity according to the current WHO classification of tumors of soft tissue and bone (Forsyth and Jundt 2013). Patients with a GCT-B affecting the mandible were also excluded, because these are probably central giant cell granulomas of the jaw (Jaffe 1953).

Based on the combination of the historical and current excerpts we could calculate the incidence of first local recur-

rences during the 5-year study period. It is essential to note that this is not the same as the incidence of recurrences for those patients diagnosed during these 5 years of study. To calculate the latter, a longer interval between the study period and the moment of reporting would be necessary.

Excerpts contained an encrypted patient identification number (allowing for identification of multiple excerpts of one patient), data on age and sex, date of arrival of the histological tissue, and the conclusion of the pathology report. AJV extracted the data and uncertain pathology conclusions in the reports were discussed with HG and JVMGB. Disaggregated incidence rate calculations for localized and diffuse type are necessary in giant cell tumors of tenosynovium. The PALGA database lacks information on tumor type (i.e., whether a giant cell tumor of tenosynovium is a localized type or a diffuse type as this is a combined diagnosis of radiological and pathological examinations), therefore additional chart review in giant cell tumors of tenosynovium was performed and published elsewhere (Mastboom et al.

2017).

Data collection Anonymized data were collected on age, sex, year of diagnosis, localization, GCT type, and date of local recurrence.

Statistics For statistical analysis, the Statistical Package for the Social Sciences (SPSS) version 23.0.0 (IBM Corp, Armonk, NY, USA) was used. Incidence of GCT-Bs was calculated per million inhabitants per year and standardized for 5-year age groups and sex for the Dutch population in 2012, as published by the Central Bureau of Statistics (CBS) (Statistics Netherlands 2014). Incidence standardized to the WHO standard population for 5-year age groups was also calculated (Ahmad et al. 2001).

We estimated first recurrences, defined as biopsied lesions or surgically treated recurrences, as the registry contained only reports for histological specimens. Both 95% confidence intervals (CI) for incidence rates (Mid-P exact) and frequencies (Wilson score) were calculated using www.openepi.com.

572

Acta Orthopaedica 2018; 89 (5): 570–574

Incidence per million inhabitants

Patients from PALGA between 2009 and 2013 n = 5,922

3.0

Excluded patients (n = 5,756): – diagnosis was not GCT, 848 – second opinion for foreign patients, 9 – uncertain diagnosis, 236 – GCT of tenosynovium, 4,663 Patients with a diagnosis suggesting GCT-B n = 166 Patients with a diagnosis suggesting a first GCT-B n = 151

Patients with a diagnosis suggesting a first recurrence of GCT-B n = 33

Figure 1. Diagram showing inclusion and exclusion of all patients with GCT in the Netherlands between 2009 and 2013.

Table 2. Overview of incidence rates

GCT-B total Long bones Axial

Crude incidence per million inhabitants per year (CI) 1.7 (1.4–1.9) 1.3 (1.1–1.6) 0.35 (0.21–0.45)

Age median (range) 35 (9–77) 35 (9–77) 38 (17–73)

1.5

0.5

Excluded because of localisation: – hand and foot, 2 – mandibular, 1

Patients with a diagnosis suggesting a first GCT-B n = 138

2.0

1.0

Patients with a diagnosis suggesting a first recurrence of GCT-B n = 36

Excluded because of localisation: – hand and foot, 10 – mandibular, 3

2.5

Percentage men (CI) 42 (34–50) 41 (33–51) 46 (29–65)

0.0 2009

2010

2011

2012

2013

Year of study

Figure 2. Crude incidence rates of GCT-B in the Netherlands with 95% confidence interval.

Age 0–4 5–9 10–14 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 75–79 80–84 85–89 90–94 95–99 0

10

20

30

Frequency during 5 years

Figure 3. Age distribution of GCT-B in the Netherlands.

Ethics, funding, and potential conflicts of interest As pathology excerpts were fully anonymized, no ethics approval was necessary for this study. This work was supported by Daiichi-Sankyo with an unconditional financial grant. There are no potential conflicts of interest.

Results Search results From PALGA, 8,156 excerpts of 5,922 patients were retrieved matching the search criteria (Figure 1). Of these 5,922 patients, 5,756 patients were excluded. 151 new cases of GCT-B were identified; however, 13 of these new cases were actually not GCT-B, but either giant cell tumors of the small bones of hands or feet or central giant cell granulomas of the jaw. 15 patients were only diagnosed with a first recurrence during the study period and had a primary tumor before the study period.

This resulted in 138 cases with a crude incidence ranging from 1.3 to 2.1 per million inhabitants per year (mean:

1.7; standard deviation (SD) 0.3; CI 1.4–2.0), age and sex corrected incidence ranged between 1.3 and 2.1 per million inhabitants per year (mean: 1.7; SD 0.3; CI 1.4–1.9), and the WHO standardized incidence was 1.4 to 2.3 per million inhabitants per year (mean: 1.7; SD 0.3; CI 1.4–2.0. Table 2 and Figure 2). 42% of patients were male (CI 34–50). The median age of patients was 35 years (9–77). 6% were below 18 years of age. The age distribution of GCT-B seems to be bimodal with a peak incidence between 20 and 39 and between 50 and 59 years (Figure 3). Most affected localization was the femur (35%), followed by the tibia (18%) (Table 3). During these 5 years, 33 patients were diagnosed with a first recurrence.

Consequently, crude incidence of pathology confirmed first recurrence was 0.40 per million inhabitants per year (CI 0.28– 0.55). Median age of patients with a first recurrence was 32 years (10–63). Median time between first diagnosis and first recurrence was 23 months (range 2–142). 30 % of the recurrences occurred within 1 year, 24% in the second year, and 30 % in the third year after diagnosis. Most common localization of recurrence were the femur and tibia (both 30%; Table 4).

Acta Orthopaedica 2018; 89 (5): 570–574

573

Table 3. Frequencies of localizations of first GCT-B Localization

Absolute frequency

Femur Tibia Radius Fibula Spine Ulna Pelvis Humerus Mastoid Patella Scapula Other * Unknown Total a

Percentage (CI)

48 35 (27–43) 25 18 (12–25) 14 10 (6–16) 13 9 (6–15) 13 9 (6–15) 5 4 (2–8) 5 4 (2–8) 2 1 (0–5) 2 1 (0–5) 2 1 (0–5) 2 1 (0–5) 3 2 (1–6) 4 3 (1–7) 138 100

Other: 4th rib, maxilla and petrous bone.

Incidence rates for the long bones and the axial skeleton were 1.3 (CI 1.1–1.6) and 0.31 (CI 0.21–0.45) per million inhabitants per year, respectively (1 patient could not be allocated to one of the groups). WHO standardized incidence rates were 1.40 and 0.31 per million per year. Incidence of recurrence were 0.32 (CI 0.21–0.47) and 0.07 (CI 0.03–0.15) per million per year. The median age of patients for the 2 groups was 35 (9–77) and 38 (17–73) years. Percentage of males was 41% (CI 33–51) and 46% (CI 29–65) respectively.

The localization of the tumors did not differ according to sex. Only 1 of the patients below 18 years of age had an axial localization of his GCT-B. However, this difference could be attributed to the low incidence of axial GCT-B. During the 5-year study period, only 1 malignant GCT-B was reported.

Discussion This is the first study on GCT-B incidence, based on a fully automated pathology database covering 100% of pathology reports in the Netherlands. Our calculated GCT-B incidence shows a higher number, compared with previously reported incidence rates (Table 1) (Liede et al. 2014, Amelio et al.

2016). In addition, the study by Amelio et al. does not seem to exclude giant cell tumors of the small bones of the hands or feet and central giant cell granulomas of the jaw, suggesting that the actual incidence of giant cell tumors of bone in this study is actually lower (Amelio et al. 2016). However, this is not exactly stated in the paper, but derived from the graphs showing localizations. For the study by Liede et al. (2014), no data on localization were reported. The higher incidence may be explained by the use of the 100% covering Dutch nationwide pathology database PALGA. Incidence of GCT-B seems to decrease slightly during the 5 years of study; this could be attributed to a normal variation in the low absolute count of

Table 4. Localization of recurrences Localization

Absolute frequency

Percentage (CI)

Femur 10 30 (17–47) Tibia 10 30 (17–47) Spine 4 12 (5–27) Radius 3 9 (3–24) Humerus 2 6 (2–20) Fibula 1 3 (1–15) Ulna 1 3 (1–15) Scapula 1 3 (1–15) Pelvis 1 3 (1–15) Total 33 100

GCT-B per year. As expected, most GCT-Bs were localized in the lower, weight-bearing extremities, as described in previous studies (Liede et al. 2014, Amelio et al. 2016). Both slight female preponderance and age distribution are comparable to these earlier reports (Liede et al. 2014, Amelio et al. 2016).

Age distribution seems to be bimodal (between 20 and 39 years and 50 and 59 years of age), which is comparable to data described by Liede et al. (2014). However, this variation could also be due to the small number.

GCT-Bs generally affect young patients (median age 35 years) and 6% are younger than 18 years. This is lower compared with the 14% in the Swedish study and the approximately 8% in Japan (manually calculated, based on published data) (Liede et al. 2014, Amelio et al. 2016). Reporting bias could be the cause of the higher percentage of patients < 18 years with a GCT-B in the Swedish and Japanese registries, because these are doctor-driven registries.

Although we do not calculate an exact incidence rate of first local recurrences for the patients diagnosed between 2009 and 2013 in this study, we calculated an incidence of all first recurrences during this time period of 0.40 per million inhabitants per year. This results in a rate of recurrence of approximately 24% (although the denominator is not exactly known), which is lower compared with the Swedish study (recurrence rate 41%) (Liede et al. 2014, Amelio et al. 2016). 2 retrospective cohort studies reported rates of recurrence between 6 and 42% (Balke et al. 2008, van der Heijden 2014b). The relatively higher recurrence rate in the Swedish study could be attributed to an effect of reporting bias (patients with a recurrence will have a higher chance of being registered). The differences in recurrence could be caused by different treatment strategies, which cannot be calculated in these studies due to a lack of data in our study and the other studies (van der Heijden 2014b).

Compared with the other studies, the reported incidence of malignant GCT-B (1 patient in 5 years, 1 of 138 patients) during these years was much lower than in the study by Liede et al. (27/337). We have no explanation for this difference.

In the future, additional nationwide studies are needed to calculate a more accurate worldwide incidence, because at the

574

moment incidence rates only for countries in North and West Europe are available. Furthermore, the incidence calculations should include information on the incidence of GCT-B subdivided into long bones and the axial skeleton.

In summary, this study is the first to report incidence of GCT-B based on a 100% coverage nationwide pathology database. These incidence numbers are of value for research and healthcare planning.

AJV, JVMGB, HG designed the study. AJV did the data collection and primary analysis of the data. All authors interpreted the data. AJV wrote the manuscript. All authors critically reviewed the manuscript and approved the final version.

Acta thanks Peter Holmberg Jørgensen and Claus Lindkær Jensen for help with peer review of this study.

Ahmad O B, Boschi-Pinto C, Lopez A D, Murray C J L, Lozano R, Inoue M. Age standardization of rates: a new WHO standard. Geneva: WHO; 2001.

Amelio J M, Rockberg J, Hernandez R K, Sobocki P, Stryker S, Bach B A, Engellau J, Liede A. Population-based study of giant cell tumor of bone in Sweden (1983–2011). Cancer Epidemiol 2016; 42: 82-9. doi: 10.1016/j. canep.2016.03.014.

Athanasou N A, Bansal M, Forsyth R, Reid R P, Sapi Z. Giant cell tumour of bone. In: (Fletcher C D, Bridge J A, Hogendoorn P C, Mertens F, eds.)

WHO classification of tumours of soft tissue and bone. Lyon: IARC; 2013. pp. 321-4.

Atkins G J, Haynes D R, Graves S E, Evdokiou A, Hay S, Bouralexis S, Findlay D M. Expression of osteoclast differentiation signals by stromal elements of giant cell tumors. J Bone Miner Res 2000; 15(4): 640-9. doi: 10.1359/jbmr.2000.15.4.640.

Balke M, Ahrens H, Streitbuerger A, Koehler G, Winkelmann W, Gosheger G, Hardes J. Treatment options for recurrent giant cell tumors of bone. J Cancer Res Clin Oncol 2008; 135(1): 149-58. doi: 10.1007/s00432-0080427-x.

Casparie M, Tiebosch A T, Burger G, Blauwgeers H, van de Pol A, van Krieken J H, Meijer G A. Pathology databanking and biobanking in The Netherlands, a central role for PALGA, the nationwide histopathology and cytopathology data network and archive. Cell Oncol 2007; 29(1): 19-24.

Acta Orthopaedica 2018; 89 (5): 570–574

Chawla S, Henshaw R, Seeger L, Choy E, Blay JY, Ferrari S, Kroep J, Grimer R, Reichardt P, Rutkowski P, Schuetze S, Skubitz K, Staddon A, Thomas D, Qian Y, Jacobs I. Safety and efficacy of denosumab for adults and skeletally mature adolescents with giant cell tumour of bone: interim analysis of an open-label, parallel-group, phase 2 study. Lancet Oncol 2013; 14(9): 901-8. doi: 10.1016/S1470-2045(13)70277-8.

Forsyth R, Jundt G. Giant cell lesion of the small bones. In: (Fletcher C D, Bridge J A, Hogendoorn P C, Mertens F, eds). WHO Classification of tumours of soft tissue and bone. Lyon: IARC; 2013. p. 320.

Jaffe H L. Giant-cell reparative granuloma, traumatic bone cyst, and fibrous (fibro-osseous) dysplasia of the jawbones. Oral Surg Oral Med Oral Pathol 1953; 6(1): 159-75.

Liede A, Bach B A, Stryker S, Hernandez R K, Sobocki P, Bennett B, Wong S S. Regional variation and challenges in estimating the incidence of giant cell tumor of bone. J Bone Joint Surg Am 2014; 96(23): 1999-2007. doi: 10.2106/jbjs.n.00367.

Mastboom M J L, Verspoor F G M, Verschoor A J, Uittenbogaard D, Nemeth B, Mastboom W J B, Bovée J V M G, Dijkstra P D S, Schreuder H W B, Gelderblom H, Van de Sande M A J. Higher incidence rates than previously known in tenosynovial giant cell tumors. Acta Orthop 2017: 88(6): 688-94. doi: 10.1080/17453674.2017.1361126.

Roux S, Amazit L, Meduri G, Guiochon-Mantel A, Milgrom E, Mariette X. RANK (receptor activator of nuclear factor kappa B) and RANK ligand are expressed in giant cell tumors of bone. Am J Clin Pathol 2002; 117(2): 210-6. doi: 10.1309/bpet-f2pe-p2bd-j3p3.

Statistics Netherlands 2014. http://statline.cbs.nl/StatWeb/publication/?

DM=SLNL&PA=7461BEV&D1=0&D2=1-2&D3=101-120&D4=5362&VW=T Thomas D, Henshaw R, Skubitz K, Chawla S, Staddon A, Blay J Y, Roudier M, Smith J, Ye Z, Sohn W, Dansey R, Jun S. Denosumab in patients with giant-cell tumour of bone: an open-label, phase 2 study. Lancet Oncol 2010; 11(3): 275-80. doi: 10.1016/s1470-2045(10)70010-3. van der Heijden L, Dijkstra P D, van de Sande M A, Kroep J R, Nout R A, van Rijswijk C S, Bovee J V, Hogendoorn P C, Gelderblom H. The clinical approach toward giant cell tumor of bone. Oncologist 2014a; 19(5): 55061. doi: 10.1634/theoncologist.2013-0432 van der Heijden L, van der Geest I C, Schreuder H W, van de Sande M A, Dijkstra P D. Liquid nitrogen or phenolization for giant cell tumor of bone?: a comparative cohort study of various standard treatments at two tertiary referral centers. J Bone Joint Surg Am 2014b; 96(5): e35. doi: 10.2106/jbjs.m.00516.

Acta Orthopaedica 2018; 89 (5): 575–579

575

Midterm risk of cancer with metal-on-metal hip replacements not increased in a Finnish population Elina EKMAN 1, Inari LAAKSONEN 1, Antti ESKELINEN 2, Pekka PULKKINEN 3, Eero PUKKALA 4, and Keijo MÄKELÄ 1

1 Department

of Orthopaedics and Traumatology, Turku University Hospital, Turku; 2 Coxa Hospital for Joint Replacement, Tampere; 3 Department of Public Health, Helsinki University, Helsinki; 4 Faculty of Social Sciences, University of Tampere, Tampere, and the Finnish Cancer Registry, Institute for Statistical and Epidemiological Cancer Research, Helsinki, Finland Correspondence: elina.ekman@tyks.fi Submitted 2018-02-17. Accepted 2018-05-15.

Background and purpose — Metal-on-metal (MoM) total hip arthroplasty (THA) and hip resurfacing arthroplasty (HRA) have been widely used during the early 21st century.

We assessed the midterm risk of cancer of patients treated with modern MoM hip implants compared with patients with non-MoM hip implants and the general Finnish population with special interest in soft tissue sarcomas and basalioma due to the findings of our previous report.

Patients and methods — All large-diameter head MoM THAs and hip resurfacings performed in Finland between 2001 and 2010 were extracted from the Finnish Arthroplasty Register (10,728 patients). Patients who underwent conventional THA formed the non-MoM reference cohort (18,235 patients). Data on cancer cases up to 2014 were extracted from the Finnish Cancer Registry. The relative risk of cancer in the general population was expressed as the ratio of observed to expected number of cases, i.e., standardized incidence ratio (SIR). Poisson regression analysis was used to compare the cancer risk between the cohorts. The mean follow-up was 7.4 years (1–14) in the MoM cohort and 8.4 years (1–14) in the non-MoM cohort.

Results — The overall risk of cancer in the MoM cohort was comparable to the general Finnish population (SIR 0.9, 95% CI 0.9–1.0). Risk of basalioma in the MoM cohort was higher than in the general Finnish population (SIR 1.2, CI 1.1–1.4) and higher than in the non-MoM cohort in the stratified regression analysis (RR 1.2, CI 1.0–1.4, p = 0.02). The SIR of soft-tissue sarcoma in the MoM cohort was 1.4 (CI 0.6–2.8); the incidence was same as in the non-MoM cohort.

Interpretation — Metal-on-metal hip implants are not associated with an increased overall risk of cancer during midterm follow-up.

Second-generation large-diameter head (LDH) MoM THA and HRA gained popularity at the beginning of the 21st century (FAR, AOANJRR 2016, NJR 2016). Metal particles emanate as a result of corrosion and wear of metal-on-metal (MoM) hip implants and can disseminate throughout the body. These particles can be found in several organs including lymphatic tissue, bone marrow, liver, and spleen (Case et al. 1994, Shea et al. 1997, Urban et al. 2000, Shimmin and Back 2005). Metal debris from hip implants has been associated with chromosomal aberrations and DNA damage (Case et al. 1996, Bonassi et al. 2000, Daley et al. 2004, Polyzois et al. 2012, Sarhadi et al. 2015). Wear particles are released both from MoM and conventional metal-on-polyethylene (MoP) bearings and from the trunnion of MoM THA (Pastides et al.

2013). However, the risk of cancer was not increased after conventional MoP THA or after earlier used first-generation MoM THA (Visuri et al. 1996, 2010).

Previous studies have found no increase in the overall risk of cancer after second-generation MoM hip arthroplasty in the short term when compared with other bearing types (Mäkelä et al. 2012, Smith et al. 2012, Brewster et al. 2013, Lalmohamed et al. 2013). However, follow-up time in these studies is relatively short as cancer takes years to develop. A recent meta-analysis could not find causative relationship between second-generation MoM implants and cancer risk (Christian et al. 2014). At the same time, a previous Finnish study reported an increased incidence of basalioma and soft tissue sarcoma in patients treated with MoM implants compared with patients treated with non-MoM implants (Mäkelä et al.

2014). Sarcoma is a severe, life-threatening disease and due to these earlier findings we felt obligated to update our data with longer follow-up.

© 2018 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/4.0) DOI 10.1080/17453674.2018.1487202

576

Acta Orthopaedica 2018; 89 (5): 575–579

Table 1. Number of patients (n) according to age at operation, and number of person-years (PY) according to the age at follow-up. The non-metal-on-metal cohort consisted of implants with metal-on-polyethylene, ceramic-on-polyethylene, and ceramic-on-ceramic bearing surfaces

Metal-on-metal cohort Non-metal-on-metal cohort Total Men Women Men Women Men Women Age n PY n PY n PY n PY n PY n PY 0–9 1 1 – – – – – – 1 1 - 10–19 5 16 3 16 – – – – 5 16 3 16 20–29 25 136 16 69 4 17 7 29 29 153 23 98 30–39 159 558 67 309 30 119 20 116 189 677 87 425 40–49 741 3,746 471 2,078 143 730 157 578 884 4,476 628 2656 50–59 2,277 11,787 1,643 8,578 850 3,880 922 4,416 3,127 15,667 2,565 12,994 60–69 2,257 19,652 1,581 14,090 2,260 13,980 2,739 16,099 4,517 33,632 4,320 30,189 70–79 763 9,403 594 6,773 3,044 25,854 5,397 39,047 3,807 35,257 5,991 45,820 80– 65 1,276 48 1,036 698 13,610 1,953 33,884 763 14,886 2,001 34,920 Total

6,293 46,573

4,423 32,948

7,029 58,190 11,195 94,168

13,322 104,763 15,618 127,116

PY: person-years

Table 2. Baseline characteristics for the metal-on-metal cohort and for the non-metal-on-metal cohort

Factor

Metal-on- Non-metal-onmetal cohort metal cohort

Mean age (SD) Number of women, n (%) Primary osteoarthritis, n (%) Secondary osteoarthritis, n (%)

59 (10) 4,426 (41) 9,901 (92) 827 (8)

71 (9) 11,202 (61) 17,683 (97) 552 (3)

We have now updated our earlier results on risk of cancer (Mäkelä et al. 2012, 2014) in patients treated with primary MoM hip implants compared with patients treated with primary non-MoM THAs and the general Finnish population in midterm follow-up, specifically the risk of sarcoma, basalioma, and skin melanoma.

Patients and methods In Finland virtually all cancer cases are recorded in the population-based Finnish Cancer Registry (Teppo et al. 1994, Pukkala et al. 2018) and 98% of primary total hip implants are recorded in the Finnish Arthroplasty Register (FAR). LDH MoM THAs and hip HRAs performed in Finland between 2001 and 2010 were extracted from the FAR and formed the MoM cohort. Patients who underwent metal-on-polyethylene, ceramic-on-polyethylene, or ceramic-on-ceramic THA during the study period formed the non-MoM reference cohort. All of these study subjects were followed-up until December 31, 2014 for emigration and vital status via the Population Registry, and for cancer incidence via Finnish Cancer Registry through a personal identity code. None of the patients were lost to follow-up. While forming the study population we included the first hip implant in every patient. If a patient

received another hip implant later, only those who had similar bearings in both sides were included.

There were 10,728 patients in the MoM cohort and 18,235 patients in the non-MoM THA cohort included in this study; 46% of the patients were men (Tables 1 and 2). The number of person-years at follow up was 79,521 for the MoM cohort and 152,358 for the non-MoM cohort. Of all our patients 497 (4.6%) had bilateral MoM implants. The mean follow-up was 7.4 years (0–14) in the MoM cohort and 8.4 years (0–14) in the non-MoM cohort.

Statistics For both cohorts (MoM and non-MoM) the person-years at risk were calculated within stratification of sex, calendar period (2001–05 and 2006–10), 5-year age groups, and follow-up time (< 2, 2–5, and > 5 years since the operation).

The expected number of each type of cancer within each stratum was calculated by multiplying the person-years in the stratum by the stratum’s age, sex, and calendar-periodspecific cancer incidence rate for the Finnish population.

The total expected numbers of cancers were summed over the strata. The cancer risk relative to the Finnish population, i.e., standardized incidence ratio (SIR), was expressed as the ratio of observed to expected number of cases. For the 95% confidence intervals (CI), we assumed that the number of observed cases followed a Poisson distribution. The study population and detailed information on the implant types have been presented in detail in earlier publications (Mäkelä et al. 2012, 2014).

Poisson regression analysis was used to estimate the relative cancer risk between the MoM and non-MoM cohorts for soft tissue sarcomas, melanoma and basalioma. Soft-tissue sarcoma and basalioma were chosen for Poisson regression due to the earlier results of Mäkelä et al. (2012, 2014) and skin melanoma due to earlier results with conventional THA (Visuri et al. 2003, Onega et al. 2006). The risk estimate (incidence rate ratio) was

Acta Orthopaedica 2018; 89 (5): 575–579

577

Table 3. Observed numbers of cancer cases, the expected numbers of cancer cases approximated from the Finnish population, and standardized incidence ratios with 95% confidence intervals—according to site—are given for the metal-onmetal cohort and for the non-metal-on-metal cohort. The latter cohort consisted of implants with metal-on-polyethylene, ceramic-on-polyethylene, and ceramic-on-ceramic bearing surfaces

Primary site

Obs

Metal-on-metal cohort Non-metal-on-metal cohort Exp SIR 95% CI % of cancer Obs Exp SIR 95% CI % of cancer

All sites 915 973 0.9 0.9–1.0 Stomach 23 21 1.1 0.7–1.7 Colon 48 55 0.87 0.6–1.2 Lung 61 95 0.64 0.5–0.8 a Corpus uteri 24 22 1.1 0.7–1.6 Prostate 239 216 1.1 1.0–1.2 Kidney 31 29 1.1 0.7–1.5 Bladder 32 41 0.8 0.5–1.1 Soft-tissue sarcoma 8 6 1.4 0.6–2.8 Non-Hodgkin lymphoma 37 38 1.0 0.7–1.4 Hodgkin lymphoma 2 2 0.9 0.1–3.1 Multiple myeloma 13 12 1.1 0.6–1.8 Leukemia 17 18 1.0 0.6–1.5 Melanoma 38 36 1.1 0.8–1.5 Basalioma 306 246 1.2 1.1–1.4 a

9 2851 2852 1.0 1.0–1.0 16 0.2 75 74 1.0 0.8–1.3 0.4 0.4 187 199 0.9 0.8–1.1 1 0.6 203 260 0.78 0.7–0.9 a 1 0.2 78 82 1.0 0.8–1.2 0.4 2 478 461 1.0 1.0–1.1 3 0.3 83 82 1.0 0.8–1.2 0.5 0.3 131 128 1.0 0.9–1.2 0.7 0.07 20 17 1.2 0.7–1.8 0.1 0.3 118 108 1.1 0.9–1.3 0.7 0.02 3 4 0.7 0.1–2.0 0.02 0.1 36 41 0.9 0.6–1.2 0.2 0.2 60 58 1.0 0.8–1.3 0.3 0.4 105 87 1.2 1.0–1.5 0.6 3 913 878 1.0 1.0–1.1 5

Obs: observed number of cancer cases; Exp: expected number of cancer cases based on cancer incidence in the comparable Finnish population; SIR: standardized incidence ratio; CI: confidence interval; % of cancer: the percentage of patients diagnosed with a certain cancer during the follow-up. a p < 0.001

adjusted for age (0–49,50–59, 60–69, 70–79, 80+) and followup time (< 2, 2–5, and > 5 years since the operation). Poisson regression analysis was checked for over-dispersion. The level of statistical significance was set at p < 0.05.

Ethics, funding, and potential conflicts of interest Ethical approval: 13.6.2017, Dnor THL/926/5.05.00/2017.

This research received no funding. The authors declare no conflicts of interest.

Results The overall risk of cancer in patients treated with MoM hip implants was slightly lower than in the general Finnish population (SIR 0.9, 95% CI 0.9–1.0) (Table 3).

There were 8 soft-tissue sarcomas in the MoM cohort during the follow-up period (SIR 1.4, CI 0.6–2.8) (Table 2). The risk of soft-tissue sarcoma in the MoM cohort was the same than that in the non-MoM cohort (RR 0.9, CI 0.4–2.0, p = 0.8).

Incidence of basalioma in the MoM cohort was higher than in the general Finnish population (SIR 1.2, CI 1.1–1.4; p < 0.001) (Table 3) and also higher than that of the non-MoM cohort (RR 1.2, CI 1.0–1.4, p = 0.02).

The SIR of skin melanoma in the MoM cohort was 1.1 (CI 0.8–1.5) and that in the non-MoM cohort was 1.2 (CI 1.0– 1.5) (Table 3). Risk of melanoma in the MoM cohort was not higher than that in the non-MoM cohort (RR 0.9, CI 0.6–1.4, p = 0.7).

Discussion We found that the overall midterm risk of cancer was not increased in patients treated with MoM hip implants when compared wit the general Finnish population in midterm follow-up. This is in line with previous short term follow-up studies on second-generation MoM hip implants (Mäkelä et al. 2012, Smith et al. 2012, Brewster et al. 2013, Lalmohamed et al. 2013, Mäkelä et al. 2014). The slightly lower overall risk of cancer in the MoM group can be influenced by the fact that MoM patients tend to be young and possibly healthier than the average population, which might cause some selection bias. A recent study from Slovenia including only THAs found a slightly higher risk of overall cancer in patients treated with MoM bearing when compared with the general population or the non-MoM patients (Levasic et al. 2018). In that study the specific cancer types that had higher prevalence in the MoM cohort compared with the general population were skin cancers excluding melanoma and prostate cancer. Comparably, we found higher risk for basalioma in our MoM cohort. This confirmation of our results from another country is an interesting finding, and needs further research. The study cohort size in the study by Levasic et al. was smaller than ours (338 MoM THAs). Prostate cancer risk was not increased in the MoM cohort in our study when compared with the general Finnish population.

Although MoM hip implants are associated with local pseudotumors, adverse local tissue reactions (ALTR), and possible genetic alterations, based on earlier literature it seems that the

578

risk for systemic tumors is not elevated after MoM THA (Case et al. 1996, Pandit et al. 2008, Ollivere et al. 2009, Langton et al. 2010). Our findings are in line with these studies. Sarhadi et al. (2015) studied DNA extracted from periprosthetic tissues of 20 MoM patients undergoing hip revision surgery because of ALTR. They found genetic alterations in 6 patients and a liposarcoma in 1 patient.

In our previous short-term follow-up study of this same study population the risk of soft-tissue sarcomas was elevated in the MoM group compared twith the non-MoM-group (Mäkelä et al. 2014). Furthermore, in that study all sarcomas were diagnosed during the last 4 years of the follow-up, raising a concern that during longer follow-up soft-tissue sarcomas might be overrepresented in the MoM cohort and that there might be a causative relationship between metal wear debris and softtissue sarcomas. However, in the current study only 1 additional soft tissue sarcoma was observed during the additional follow-up years 2012–14, and the incidence was similar in the MoM patient population compared with the general Finnish population and similar to the risk in the non-MoM group.

Nonetheless, the overall number of sarcomas is small and we plan to report further follow-up. To our knowledge, there are no other studies reporting increased incidence of soft-tissue sarcomas in patients treated with metal-on-metal hip implants.

Incidence of basalioma was higher in the MoM cohort than in the non-MoM cohort and also increased when compared with the Finnish population. A similar finding has previously been reported only with conventional total hip replacements (Brewster et al. 2013). The majority of the previous studies on MoM hip implants either exclude non-melanoma skin cancer or include basaliomas in the category of other skin cancers and the data on basalioma incidence in patients treated with MoM implants is limited (Visuri et al. 2006, Smith et al. 2012, Lalmohamed et al. 2013). Due to its benign nature, basalioma is traditionally not included in the official national cancer statistics. In Finland only the first basalioma for each person is recorded in the Finnish Cancer Registry (Pukkala et al. 2018).

This may bias our results since patients treated with HRA are generally younger than patients treated with conventional THA, and it may therefore be more likely for them to be diagnosed with basalioma for the first time during our follow-up.

We found that the incidence of skin melanoma was not elevated in the MoM cohort compared with the general Finnish population. Earlier studies have found conflicting evidence concerning conventional non-MoM THAs’ association with melanoma incidence. Some studies have reported higher melanoma incidence in patients treated with non-MoM implants than in the general population (Nyrén et al. 1995, Olsen et al.

1999, Visuri et al. 2003, 2006) while others have found no difference (Visuri et al. 2010, Levasic et al. 2018). No increase in the risk of melanoma was found for patients treated with a MoM resurfacing device (Brewster et al. 2013).

The study by Brewster et al. (2013) found an increased risk of multiple myeloma and other immunoproliferative neo-

Acta Orthopaedica 2018; 89 (5): 575–579

plasms in THA patients during the first 4 years after arthroplasty. However, their study did not differentiate MoM bearings from other types of bearings and the study also included patients with rheumatic conditions, which are known to increase the risk of immunoproliferative neoplasms (Isomäki et al. 1978). Our study found no excess risk of myeloma in the MoM hip implant patients.

We acknowledge that our study has several limitations.

First, as with all registry-based studies there is a possibility of a selection bias. Registry-based studies have the advantage of reporting results from a large patient group and reporting so-called “real world data” but the disadvantage of possible confounding by indication (Freemantle et al. 2013). That is, the patients selected for THA or HRA may be for example healthier than the average population. Ideally this could be avoided by randomized controlled studies. Second, we did not have any blood metal ion measurements or imaging findings of the patients. It is theoretically possible that higher cancer risk might be associated with higher ion levels and our study is not able to detect such a subgroup. However, the findings of the meta-analysis by Christian et al. (2014) suggest that the concentrations and doses of Co/Cr required to induce a genotoxic or tumorigenic outcome are much higher than the systemic Co/Cr concentrations typically present in MoM hip implant patients. Third, even though our follow-up time now reaches the midterm point, genetic alterations might still happen or manifest later. In a recent meta-analysis Pijls et al.

(2016) reported a higher risk of mortality in patients with MoM THA compared with patients with non-MoM THA when the follow-up exceeded 10 years. No difference was noted with shorter follow-up. Considering this we plan to report longterm results from this same population.

In summary, patients treated with MoM hip implants had a comparable cancer risk to patients treated with non-MoM hip implants and the general Finnish population. They did not have increased risk for soft-tissue sarcoma or skin melanoma.

Only the incidence for basalioma was increased in the MoM cohort compared with the non-MoM cohort and compared with the general population.

KM designed and coordinated the study and helped to draft the manuscript.

EE collected the data and drafted the manuscript. EP calculated the statistics. KM, IL, PP, AE, EP and EE contributed to the interpretation of the data and results and to the preparation of the manuscript. All authors read and approved the final manuscript.

Acta thanks Bart G Pijls and other anonymous reviewers for help with peer review of this study.

AOANJRR. Annual Report 2016. AOANJRR; 2016.

Bonassi S, Hagmar L, Strömberg U, Huici Montagud A, Tinnerberg H, Forni A, Heikkilä P, Wanders S, Wilhardt P, Hansteen I L, Knudsen L E, Norppa H. Chromosomal aberrations in lymphocytes predict human cancer independently of exposure to carcinogens. Cancer Res 2000; 60(6): 1619-25.

Acta Orthopaedica 2018; 89 (5): 575–579

Brewster D H, Stockton D L, Reekie A, Ashcroft G P, Howie C R, Porter D E, Black R J. Risk of cancer following primary total hip replacement or primary resurfacing arthroplasty of the hip: a retrospective cohort study in Scotland. Br J Cancer 2013; 108(9): 1883-90.

Case C P, Langkamer V G, James C, Palmer M R, Kemp A J, Heap P F, Solomon L. Widespread dissemination of metal debris from implants. J Bone Joint Surg Br 1994; 76(5): 701-12.

Case C P, Langkamer V G, Howell R T, Webb J, Standen G, Palmer M, Kemp A, Learmonth I D. Preliminary observations on possible premalignant changes in bone marrow adjacent to worn total hip arthroplasty implants.

Clin Orthop Relat Res 1996; (329 Suppl): S269-79.

Christian W V, Oliver L D, Paustenbach D J, Kreider M L, Finley B L. Toxicology-based cancer causation analysis of CoCr-containing hip implants: a quantitative assessment of genotoxicity and tumorigenicity studies. J Appl Toxicol 2014: 939-67.

Daley B, Doherty A T, Fairman B, Case C P. Wear debris from hip or knee replacements causes chromosomal damage in human cells in tissue culture.

J Bone Joint Surg Br 2004; 86(4): 598-606.

FAR. The Finnish Arthroplasty Register (FAR).

Freemantle N, Marston L, Walters K, Wood J, Reynolds M R, Petersen I. Making inferences on treatment effects from real world data: propensity scores, confounding by indication, and other perils for the unwary in observational research. BMJ 2013; 347: f6409.

Isomäki H A, Hakulinen T, Joutsenlahti U. Excess risk of lymphomas, leukemia and myeloma in patients with rheumatoid arthritis. J Chronic Dis 1978; 31(11): 691-6.

Lalmohamed W, MacGregor AJ , de Vries F, Leufkens H, van Staa T. Patterns of risk of cancer in patients with metal-on-metal hip replacements versus other bearing surface types: a record linkage study between a prospective joint registry and general practice electronic health records in England.

PLoS One 2013; 8(7): e65891.

Langton D J, Jameson S S, Joyce T J, Hallab N J, Natu S, Nargol V F. Early failure of metal-on-metal bearings in hip resurfacing and large-diameter total hip replacement: a consequence of excess wear. J Bone Joint Surg Br 2010; 92(1): 38-46.

Levasic V, Milosev I, Zadnik V. Risk of cancer after primary total hip replacement: the influence of bearings, cementation and the material of the stem.

Acta Orthop 2018; 89(2): 234-9.

Mäkelä K T, Visuri T, Pulkkinen P, Eskelinen A, Remes V, Virolainen P, Junnila M, Pukkala E. Risk of cancer with metal-on-metal hip replacements: population based study. BMJ 2012; 345: e4646.

Mäkelä K T, Visuri T, Pulkkinen P, Eskelinen A, Remes V, Virolainen P, Junnila M, Pukkala E. Cancer incidence and cause-specific mortality in patients with metal-on-metal hip replacements in Finland. Acta Orthop 2014; 85(1): 32-8.

NJR. NJR 13th Annual Report. NJR 2016.

Nyrén O, McLaughlin J K, Gridley G, Ekbom A, Johnell O, Fraumeni J F, Adami H O. Cancer risk after hip replacement with metal implants: a population-based cohort study in Sweden. J Natl Cancer Inst 1995; 87(1): 28-33.

Ollivere B, Darrah C, Barker T, Nolan J, Porteous M J. Early clinical failure of the Birmingham metal-on-metal hip resurfacing is associated with metallosis and soft-tissue necrosis. J Bone Joint Surg Br 2009; 91(8): 1025-30.

Olsen J H, McLaughlin J K, Nyrén O, Mellemkjaer L, Lipworth L, Blot W J, Fraumeni J F. Hip and knee implantations among patients with osteoarthritis and risk of cancer: a record-linkage study from Denmark. Int J Cancer 1999; 81(5): 719-22.

579

Onega T, Baron J, MacKenzie T. Cancer after total joint arthroplasty: a metaanalysis. Cancer Epidemiol Biomarkers Prev 2006; 15(8): 1532-7.

Pandit H, Glyn-Jones S, McLardy-Smith P, Gundle R, Whitwell D, Gibbons C L M, Ostlere S, Athanasou N, Gill H S, Murray D W. Pseudotumours associated with metal-on-metal hip resurfacings. J Bone Joint Surg Br 2008; 90(7): 847-51.

Pastides P S, Dodd M, Sarraf K M, Willis-Owen C A. Trunnionosis: a pain in the neck. World J Orthop 2013; 4(4): 161-6.

Pijls B G, Meessen J M T A, Schoones J W, Fiocco M, van der Heide H J L, Sedrakyan A, Nelissen R G H H. Increased mortality in metal-on-metal versus non-metal-on-metal primary total hip arthroplasty at 10 years and longer follow-up: a systematic review and meta-analysis. PLoS One 2016; 11(6): e0156051 Polyzois I, Nikolopoulos D, Michos I, Patsouris E, Theocharis S. Local and systemic toxicity of nanoscale debris particles in total hip arthroplasty. J Appl Toxicol 2012; 32(4): 255-69.

Pukkala E, Engholm G, Hojsgaard Schmidt L, Storm H, Khan S, Lambe M, Pettersson D, Olafsdottir E, Tryggvadottir L, Hakanen T, Malila N, Virtanen A, Johannesen T, Laronningen S, Ursin G. Nordic Cancer Registries: an overview of their procedures and data comparability. Acta Oncol 2018; 57(4): 440-55.

Sarhadi V K, Parkkinen J, Reito A, Nieminen J, Porkka N, Wirtanen T, Laitinen M, Eskelinen A, Knuutila S. Genetic alterations in periprosthetic soft-tissue masses from patients with metal-on-metal hip replacement.

Mutat Res 2015; 781: 1-6.

Shea K G, Lundeen G A, Bloebaum R D, Bachus K N, Zou L. Lymphoreticular dissemination of metal particles after primary joint replacements. Clin Orthop Relat Res 1997; (338): 219-26.

Shimmin A J, Back D. Femoral neck fractures following Birmingham hip resurfacing: a national review of 50 cases. J Bone Joint Surg Br 2005; 87: 463-4.

Smith A J, Dieppe P, Porter M, Blom A W; National Joint Registry of England and Wales. Risk of cancer in first seven years after metal-on-metal hip replacement compared with other bearings and general population: linkage study between the National Joint Registry of England and Wales and hospital episode statistics. BMJ 2012; 344: e2383.

Teppo L, Pukkala E, Lehtonen M. Data quality and quality control of a population-based cancer registry: experience in Finland. Acta Oncol 1994; 33(4): 365-9.

Urban R M, Jacobs J J, Tomlinson M J, Gavrilovic J, Black J, Peoc’h M. Dissemination of wear particles to the liver, spleen, and abdominal lymph nodes of patients with hip or knee replacement. J Bone Joint Surg Am 2000; 82(4): 457-76.

Visuri T, Pukkala E, Paavolainen P, Pulkkinen P, Riska EB. Cancer risk after metal on metal and polyethylene on metal total hip arthroplasty. Clin Orthop Relat Res 1996; (329 Suppl): S280-9.

Visuri T, Pukkala E, Pulkkinen P, Paavolainen P. Decreased cancer risk in patients who have been operated on with total hip and knee arthroplasty for primary osteoarthrosis: a meta-analysis of 6 Nordic cohorts with 73,000 patients. Acta Orthop Scand 2003; 74(3): 351-60.

Visuri T, Pulkkinen P, Paavolainen P. Malignant tumors at the site of total hip prosthesis. analytic review of 46 cases. J Arthroplasty 2006: 311-23.

Visuri T, Pulkkinen P, Paavolainen P, Pukkala E. Cancer risk is not increased after conventional hip arthroplasty: a nationwide study from the Finnish Arthroplasty Register with follow-up of 24,636 patients for a mean of 13 years. Acta Orthop 2010; 81(1): 77-81.

580

Acta Orthopaedica 2018; 89 (5): 580–584

Growth of Cutibacterium acnes is common on osteosynthesis material of the shoulder in patients without signs of infection Anna BOTH 1*, Till O KLATTE 2*, Andreas LÜBKE 3, Henning BÜTTNER 1, Maximilian J HARTEL 2, Lars G GROSSTERLINDEN 4, and Holger ROHDE 1 1 Institut

für Medizinische Mikrobiologie, Virologie und Hygiene, 2 Klinik für Unfall- und Wiederherstellungschirurgie, 3 Institut für Pathologie, Universitätsklinikum Hamburg-Eppendorf, Hamburg, Germany, 4 Zentrum für Orthopädie, Unfall- und Wirbelsäulenchirurgie, Asklepios Klinik Altona, Hamburg, Germany * Authors contributed equally to work.

Correspondence: rohde@uke.de Submitted 2018-03-09. Accepted 2018-05-08.

Background and purpose — Cutibacterium acnes, formerly known as Propionibacterium acnes, is often isolated from deep tissues of the shoulder. It is recognized as an important causative agent of foreign-material associated infections. However, the incidence and significance of its detection in tissues from patients without clinical evidence for infection is unclear. We assessed the incidence of C. acnes colonization of osteosynthesis material in asymptomatic patients, and evaluated the short-term outcome in relation to the microbiological findings.

Patients and methods — We microbiologically analyzed osteosynthesis material of 34 asymptomatic patients after surgery on the clavicle. Material obtained from 19 asymptomatic patients after osteosynthesis of the fibula served as a control group. Patients were clinically followed up for 3–24 months after removal of the osteosynthesis material.

Results — Bacteria were recovered from devices in 29 of 34 patients from the clavicle group. 27 of 29 positive samples grew C. acnes. Isolation of C. acnes was more common in male than in female patients. No bacterial growth was observed on foreign material from patients in the fibula group. All patients remained asymptomatic at follow-up.

Interpretation — Growth of C. acnes is common on osteosynthesis material of the shoulder, especially in males.

Samples were positive irrespective of clinical signs of infection. Therefore, detection of C. acnes in this clinical setting is of questionable clinical significance. The high positivity rate in asymptomatic patients discourages routine sampling of material in cases without clinical evidence for infection.

Operative treatment of shoulder fractures has increased since the early 2000s (Persico et al. 2014). Infections after operative fracture-fixation occur with an incidence of roughly 3%, although large prospective studies are lacking. Similarly, prosthetic joint infections (PJI) of the shoulder occur with an incidence of 3%, comparable to PJI of the hip and knee (Kurtz et al. 2012, Bohsali et al. 2017).

In recent years, most likely due to improved diagnostic procedures, an increase in Cutibacterium acnes infections has been noted (Achermann et al. 2014, Shifflett et al. 2016).

Cutibacterium acnes, formerly Propionibacterium acnes, is a Gram-positive anaerobic rod-shaped bacterium, commensally inhabiting the pilosebaceous unit in humans. Though a part of the normal human skin microbiota, C. acnes is also implicated in biofilm-associated infections and inflammatory processes, such as prosthetic valve endocarditis, infection of breast or eye implants, and acne vulgaris (Aubin et al. 2014, Beylot et al. 2014, van Valen et al. 2016). Of note, C. acnes is increasingly recognized as an important pathogen in bone and joint infections (Achermann et al. 2014). The shoulder is the most commonly affected joint in C. acnes infections.

In fact, C. acnes has previously been isolated in one-quarter of revision arthroscopies for shoulder pain or stiffness after a first arthroscopy (Horneff et al. 2015). Even more strikingly, C. acnes has been identified as the most frequently isolated pathogen in prosthetic shoulder joint infections (Piper et al.

2009, Kadler et al. 2015).

C. acnes can lead to infections that are clinically evident, while some cases with detection of C. acnes are considered only mildly symptomatic or even asymptomatic. One study found 42% of joint fluid aspirates taken at primary shoulder replacement to be positive for C. acnes, and in another study 20% of deep tissue specimens acquired after shoulder arthroscopy grew C. acnes, both studies focusing on patients

© 2018 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/4.0) DOI 10.1080/17453674.2018.1489095

Acta Orthopaedica 2018; 89 (5): 580–584

with no clinical signs of infection (Levy et al. 2013, Chuang et al. 2015). Consequentially, there has been a debate over the implication of growth of C. acnes in tissue samples from shoulders in the absence of clinical signs of infection.

In this study we assessed the frequency of bacterial colonization of osteosynthesis material in healthy adults with no signs of foreign material infection and assessed self-reported outcomes at 3 to 24 months of patients with osteosynthesis material, which were culture positive, compared with those who had sterile devices.

Patients and methods Patient selection and sample collection Inclusion criteria were age over 18 years with need for removal of a fixation device of the clavicle or distal fibula without any clinical signs of infection. Except for the hook plate, which has to be removed routinely, the reason for removal of the osteosynthesis material was at the patient’s request, mainly due to foreign body sensation or for cosmetic reasons.

Patients in the clavicle group (n = 34) had suffered a closed clavicle fracture or a closed dislocation of the acromioclavicular (AC) joint due to trauma. In case of a lateral clavicle fracture (n = 7) or dislocation of the AC joint (n = 12), open reduction and fixation with a hook plate was performed. To avoid any complication due to the design of this plate routine removal was considered necessary after 4 months in the case of AC joint dislocation and after 6 months in the case of a lateral clavicle fracture. The other clavicle fractures (n = 15) were treated by open reduction and internal fixation with a standard plate.

All patients in the fibula group (n = 19) had a closed fracture of the distal fibula, which was treated with open reduction and internal fixation by plate and screws.

Exclusion criteria were development of pseudarthrosis, intake of antibiotics 2 weeks before removal of the fixation device for any reasons, presentation of fistula or abscess in the area of the fixation device, an open injury of the clavicle or fibula, any history of wound healing disturbance, or infection of the fixation device after the initial surgery.

Preoperatively physical examination was undertaken and C-reactive protein levels (CRP) in serum and leucocyte count were assessed in every patient.

Removal of the osteosynthesis material was performed in the clavicle and fibula group similarly. The surgical region was shaved where appropriate. For skin disinfection isopropanol was applied for 8 minutes. The skin was allowed to dry; thus isopropanol had a total exposure time of 10 minutes. After skin incision, the surgical blade was changed and the fixation device dissected. Two tissue specimens of around 0.5 x 0.5 x 0.5 cm were taken above and 1 below the fixation device. Plates and screws were separately transferred to sterile containers and immediately transported to the microbiologi-

581

cal laboratory for testing. Perioperative antibiotic prophylaxis in the form of 2 g intravenous cefazolin was given after all samples were taken.

Patients were seen postoperatively once at the first postoperative day for routine check-up. Finally, 3 to 24 months after surgery, participating patients were called by telephone and asked if any further pain, limitation of motion, foreign body sensation, or problems with wound healing had occurred.

Microbiology Tissue samples from under and above the fixation plate were transferred to a sterile mortar and covered with 1 mL of sterile phosphate-buffered saline (PBS). Tissues were homogenized and 100 µL of the suspension was plated on Columbia sheep blood agar, chocolate agar, Sabouraud agar, and Schaedler anaerobic agar (all Oxoid, Basingstoke, UK). Plates were incubated at 36°C in 5% CO2 or under anaerobic conditions.

Plates were read after 48 h, 7 days, and 14 days. Additionally, 2 mL of thioglycolate broth was inoculated with one drop of the tissue suspension and incubated likewise.

Osteosynthesis screws were transferred to a sterile 50 mL Falcon tube (Greiner, Frickenhausen, Germany) and plates were transferred to a sterilized implant box (Lock & Lock, Seoul, South Korea) in a biosafety cabinet with laminar airflow. Screws were covered with 10 mL and plates were covered with 50 mL of PBS, respectively. These air-tight sealed containers were vortexed at maximum speed for 30 s. Vessels containing the foreign material were treated in an ultrasound bath for 3 min at 40 kHz (BactoSonic; Bandelin GmbH, Berlin, Germany) followed by 30 s of vigorous shaking.

Sonicate fluid was transferred into new Falcon tubes and centrifuged at 3500 rpm for 10 min. 9 mL and 49 mL of supernatant, respectively, was removed and the pellet was resuspended by pipetting in the remaining 1 mL of fluid. Sheep blood agar, chocolate agar, Sabouraud agar, and Schaedler anaerobic agar were each inoculated with 100 µL of the sonicate fluid. Plates were incubated and read as described above.

For histological analysis, 11 tissue samples from below the fixation device from 9 clavicular group patients and 3 fibula group patients were placed into 4% formalin and embedded into paraffin blocks.

Statistics Categorical variables, such as surgical site, sex, restriction of movement, and foreign body sensation were tested for their association with having positive cultures from tissue or foreign material by Fisher’s exact test.

Ethics, funding, and potential conflicts of interest Internal review board approval was obtained (No: PV4696, Review Board of University Medical Center Hamburg-Eppendorf). All participants gave written informed consent. The study was supported by the Damp Foundation (project 201319). There are no conflicts of interest to declare.

582

Acta Orthopaedica 2018; 89 (5): 580–584

Overview of patients with growth of non-C. acnes bacterial species Patient

Growth on foreign material

3

C. acnes C. acnes S. saccharolyticus (low CFU a) S. saccharolyticus (low CFU a) C. acnes C. acnes K. rhizophila (low CFU a) Brevundimonas sp. (low CFU a) C. acnes C. acnes S. epidermidis S. epidermidis C. acnes C. acnes P. mirabilis (low CFU a) P. mirabilis (low CFU) S. capitis (low CFU a) S. epidermidis (low CFU a) C. acnes C. acnes S. saccharolyticus (low CFU a) S. epidermidis (low CFU a) No growth S. saccharolyticus S. saccharolyticus

6 9 30

36 18 45

Growth in peri-implant tissue

a Low CFU indicates colony count of under 10 CFU/mL in sonicate fluid. b Contamination was deemed likely if species were detected in only one

or below the fixation devices).

Results Patient characteristics 53 patients (median age 41 years [21–74], 36 men) undergoing removal of osteosynthesis material between February 2016 and September 2017 were included in the study. 34 patients underwent removal of osteosynthesis material from the clavicle and 19 patients had material removed from the fibula. The foreign material was removed after a median of 6 months (3–120) for clavicular osteosynthesis and 20 months (8–53) for fibular osteosynthesis. Median leucocyte count in blood samples at a mean of 14 days (1–67) before surgery was 6.8 Giga/L (3.7–12). Median C-reactive protein levels were < 5 mg/L (< 5–50).

Isolated organisms Tissue samples, as well as screws and fixating plate, were available for microbiological analysis in all patients. Bacterial growth was observed in 29/34 of tissue samples from clavicular osteosynthesis. 27 samples from the clavicle group grew C. acnes. Interestingly, only 1 out of 4 female patients undergoing removal of a fixation device from the clavicle had cultures positive for C. acnes, while 26 out of 30 male patients had cultures positive for C. acnes (p = 0.02). Cultures from sonication fluid of clavicular foreign material gave largely similar results to the tissue culture for C. acnes. In 1 patient C. acnes was found only on the foreign material but not in the surrounding tissue.

22 of those samples grew only C. acnes, while 5 showed growth of at least one additional organism (Table). There were no sex-specific differences in the growth of non-C. acnes isolates. In Patient 3 C. acnes was accompanied by S. saccha-

Interpretation

Contamination likely b (K. rhizophila; Brevundimonas sp.)

Contamination likely b (P. mirabilis; S. capitis, S. epidermidis)

Contamination likely b (S. saccharolyticus) Contamination likely b

of the analyzed materials (fixating plate, screws, tissues from above

rolyticus, and materials from Patient 9 showed growth of S. epidermidis in addition to C. acnes. In both patients 3 and 9 the respective second species was present in multiple materials and high colony counts.

Patient 30 showed additional growth of S. epidermidis and S. capitis on the osteosynthesis material and P. mirabilis in 1 tissue sample, while samples from Patient 6 showed additional growth of Brevundimonas sp. and Kocuria sp. Foreign materials of Patient 36 grew S. saccharolyticus in addition to C. acnes. The non-C. acnes species in Patients 6, 30, and 36 were grown in very low numbers, thus they might constitute contamination during removal from the body or microbiological workup (Table).

Samples from Patient 45 grew only S. saccharolyticus.

Osteosynthesis material from Patient 18 grew S. epidermidis in very low numbers, which might also, rather, constitute contamination.

Neither tissue samples nor osteosynthesis materials from any of the 19 patients who underwent removal of fibular osteosynthesis showed bacterial growth.

Patients whose specimens grew C. acnes were not more likely to have a leucocyte count or serum CRP above the median than patients with sterile specimens (p = 1.0).

Tissue samples from below the fixation device from 8 clavicular group patients and 3 fibula group patients were analyzed in histology. All samples showed signs of fibrosis and all but 1 showed small to moderate amounts of metal particles.

Two samples from clavicular group patients showed few to moderate signs of inflammation. Microscopy of tissue from Patient 51 revealed 25 polymorphonuclear leucocytes (PMN) per 10 high power fields (HPF). No systemic signs of infection were noted (blood leucocyte count: 3.7 Giga/L, CRP < 5 mg/L). There was no growth of aerobic or anaerobic bacteria

Acta Orthopaedica 2018; 89 (5): 580–584

in either tissue or on the foreign material. Histology of tissue from Patient 36 showed 55 PMN per 10 HPF. Blood leucocytes were normal at 6.4 Giga/L; CRP was < 5 mg/L. Cultures from foreign material and tissue grew C. acnes, additionally small amounts of S. saccharolyticus were cultured from the foreign material. Of the remaining 6 samples from the clavicular group, without signs of inflammation, 3 had positive cultures for C. acnes and 3 were culture negative. None of the 3 samples from the fibula group showed signs of inflammation.

During the follow-up telephone call at 3 to 24 months postoperatively, none of the patients complained about pain, limitation of motion, foreign body sensation, or problems with wound healing.

Discussion We found a high incidence of bacterial growth on osteosynthesis material of the clavicle in patients with no clinical signs of infection. In a control group of patients undergoing removal of osteosynthesis material devices of the fibula, no bacterial growth was observed. C. acnes was by far the most common organism isolated from the foreign material. Of note, materials from male patients were significantly more likely to have positive cultures for C. acnes compared with specimens from female patients. This finding is limited by the small number of women included in this study, which is due to the epidemiology of shoulder injuries (Kihlström et al. 2017). In support of our findings, however, others have found a higher incidence of C. acnes infection in men as compared with women (Berthelot et al. 2006, Millett et al. 2011). A possible explanation for this finding is that men are more commonly colonized with C. acnes in the pilosebaceous units of the head, neck, shoulders, and upper trunk (Kadler et al. 2015).

Prosthetic shoulder joints are at special risk for infection with C. acnes, which is much less common in knee or hip prosthetic joint infections. In fact, C. acnes was found to be the most commonly or second most commonly isolated organism in shoulder prosthetic joint infections in many studies or case series examining revision shoulder arthroplasties (Pottinger et al. 2012, Nelson et al. 2016, Singh et al. 2012, Wang et al. 2013). Interestingly, C. acnes has been implicated in socalled aseptic loosening of prosthetic joints and glenohumeral arthropathy, both without any overt signs of infection (Levy et al. 2013). It may be that C. acnes preferentially establishes slowly destructive infections that do not necessarily elicit an inflammatory response.

However, the interpretation of the pathogenic significance of C. acnes in tissue samples especially from the shoulder is challenging. Some reports suggest an elevated risk of contamination during incision or removal of the specimen in this location (Hudek et al. 2014), potentially promoted by the sometimes insufficient removal of C. acnes from the dermis by surgical skin preparation (Lee et al. 2014). Building on the

583

frequent identification of C. acnes in apparently non-infected joints, others even speculated that the organism might be able to colonize the joint without clinical signs of infection (Zeller et al. 2007, Chuang et al. 2015). In line with these observations made in shoulder arthroplasty, in our series of patients a high proportion of fixation devices of the clavicle were positive for C. acnes, but clinical signs of infection and histologic evidence suggestive of infection were lacking. Moreover, no significant difference in laboratory infection parameters in patients positive with C. acnes on the foreign material and those with negative specimens were found. Thus, the idea is supported that, due to its innocuous nature, C. acnes indeed can colonize even body sites and tissues that usually would be regarded as sterile, and may persist even for long periods without causing signs of infection. However, clinical significance is extremely difficult to assess as even manifest infections with C. acnes may show only slight signs of infection, i.e., systemic parameters such as leucocyte count, erythrocyte sedimentation rate, and C-reactive protein in serum may be normal (Uçkay et al. 2010, Piggott et al. 2016). Importantly, in our study, no significant histologic signs of infection were observed, and all patients were well at follow-up. Also important is that the presence of an implant has been found essential for persistent C. acnes infections in an animal model of foreign-body infection (Shiono et al. 2016). Therefore, it cannot be ruled out that C. acnes positive patients would have developed apparent infections at later time points if devices were not removed. However, in our experience and after reviewing the literature, complications of osteosynthesis after more than one year appear to be rare.

In summary, detection of C. acnes in tissue samples from osteosynthesis surgery of the clavicle must be interpreted with caution, even if high numbers of bacteria are recovered.

Microbiological sampling should always be accompanied by histopathological analysis, and possibly additional markers like α-defensin or C-reactive protein, allowing for complete assessment of possible infection. As long as no specific markers to differentiate between invasive and contaminating isolates are available, it is advisable not to send tissue for microbiological analysis if no clinical signs or symptoms suggestive of infection are present. Our data suggest that, otherwise, unnecessary antibiotic treatment of uninfected individuals could occur. Further studies addressing the role of C. acnes in surgery-related complications like pseudarthrosis are warranted.

AB, TOK, HR, and LGG designed the study, AB, TOK, HR, AL, HB, and MJH conducted the experiments and collected and interpreted the data. HR and AB wrote the manuscript.

The authors would like to thank Claudia Lehnert for the excellent technical support.

584

Acta thanks Lars Gunnar Johnsen and other anonymous reviewers for help with peer review of this study.

Achermann Y, Goldstein E J C, Coenye T, Shirtliff M E. Propionibacterium acnes: from commensal to opportunistic biofilm-associated implant pathogen. Clin Microbiol Rev 2014; 27(3): 419-40.

Aubin G G, Portillo M E, Trampuz A, Corvec S. Propionibacterium acnes, an emerging pathogen: from acne to implant-infections, from phylotype to resistance. Médecine Mal Infect 2014; 44(6): 241-50.

Berthelot P, Carricajo A, Aubert G, Akhavan H, Gazielly D, Lucht F. Outbreak of postoperative shoulder arthritis due to Propionibacterium acnes infection in nondebilitated patients. Infect Control Hosp Epidemiol 2006; 27(9): 987-90.

Beylot C, Auffret N, Poli F, Claudel J-P, Leccia M-T, Del Giudice P, et al.

Propionibacterium acnes: an update on its role in the pathogenesis of acne.

J Eur Acad Dermatology Venereol 2014; 28(3): 271-8.

Bohsali K I, Bois A J, Wirth M A. Complications of shoulder arthroplasty. J Bone Joint Surg Am 2017; 99(3): 256-69.

Chuang M J, Jancosko JJ , Mendoza V, Nottage W M. The incidence of Propionibacterium acnes in shoulder arthroscopy. Arthrosc J Arthrosc Relat Surg 2015; 31(9): 1702-7.

Horneff J G, Hsu J E, Voleti P B, O’Donnell J, Huffman G R. Propionibacterium acnes infection in shoulder arthroscopy patients with postoperative pain. J Shoulder Elb Surg 2015; 24(6): 838-43.

Hudek R, Sommer F, Kerwat M, Abdelkawi A F, Loos F, Gohlke F. Propionibacterium acnes in shoulder surgery: true infection, contamination, or commensal of the deep tissue? J Shoulder Elb Surg 2014; 23(12): 1763-71.

Kadler B K, Mehta S S, Funk L. Propionibacterium acnes infection after shoulder surgery. Int J Shoulder Surg 2015; 9(4): 139-44.

Kihlström C, Möller M, Lönn K, Wolf O. Clavicle fractures: epidemiology, classification and treatment of 2 422 fractures in the Swedish Fracture Register; an observational study. BMC Musculoskelet Disord 2017; 18(1): 82.

Kurtz S M, Lau E, Watson H, Schmier J K, Parvizi J. Economic burden of periprosthetic joint infection in the United States. J Arthroplasty 2012; 27(8): 61-65.e1.

Lee M J, Pottinger P S, Butler-Wu S, Bumgarner R E, Russ S M, Matsen F A. Propionibacterium persists in the skin despite standard surgical preparation. J Bone Joint Surg Am 2014; 96(17): 1447-50.

Levy O, Iyer S, Atoun E, Peter N, Hous N, Cash D, et al. Propionibacterium acnes: an underestimated etiology in the pathogenesis of osteoarthritis? J Shoulder Elb Surg 2013; 22(4): 505-11.

Acta Orthopaedica 2018; 89 (5): 580–584

Millett P J, Yen Y-M, Price C S, Horan M P, van der Meijden O A, Elser F. Propionibacterium acnes infection as an occult cause of postoperative shoulder pain: a case series. Clin Orthop Relat Res 2011; 469(10): 282430.

Nelson G N, Davis D E, Namdari S. Outcomes in the treatment of periprosthetic joint infection after shoulder arthroplasty: a systematic review. J Shoulder Elb Surg 2016; 25(8): 1337-45.

Persico F, Lorenz E, Seligson D. Complications of operative treatment of clavicle fractures in a Level I trauma center. Eur J Orthop Surg Traumatol 2014; 24(6): 839-44.

Piggott D A, Higgins Y M, Melia M T, Ellis B, Carroll K C, McFarland E G, et al. Characteristics and treatment outcomes of Propionibacterium acnes prosthetic shoulder infections in adults. Open Forum Infect Dis 2016; 3(1): ofv191.

Piper K E, Jacobson M J, Cofield R H, Sperling J W, Sanchez-Sotelo J, Osmon D R, et al. Microbiologic diagnosis of prosthetic shoulder infection by use of implant sonication. J Clin Microbiol 2009; 47(6): 1878-84.

Pottinger P, Butler-Wu S, Neradilek M B, Merritt A, Bertelsen A, Jette J L, et al. Prognostic factors for bacterial cultures positive for Propionibacterium acnes and other organisms in a large series of revision shoulder arthroplasties performed for stiffness, pain, or loosening. J Bone Joint Surg Am 2012; 94(22): 2075-83.

Shifflett G D, Bjerke-Kroll B T, Nwachukwu B U, Kueper J, Burket J, Sama A A, et al. Microbiologic profile of infections in presumed aseptic revision spine surgery. Eur Spine J 2016; 25(12): 3902-7.

Shiono Y, Ishii K, Nagai S, Kakinuma H, Sasaki A, Funao H, et al. Delayed Propionibacterium acnes surgical site infections occur only in the presence of an implant. Sci Rep 2016; 6(1): 32758.

Singh J A, Sperling J W, Schleck C, Harmsen W S, Cofield R H. Periprosthetic infections after total shoulder arthroplasty: a 33-year perspective. J Shoulder Elb Surg 2012; 21(11): 1534-41.

Uçkay I, Dinh A, Vauthey L, Asseray N, Passuti N, Rottman M, et al. Spondylodiscitis due to Propionibacterium acnes: report of twenty-nine cases and a review of the literature. Clin Microbiol Infect 2010; 16(4): 353-8.

Van Valen R, de Lind van Wijngaarden R A F, Verkaik N J, Mokhles M M, Bogers AJ J C. Prosthetic valve endocarditis due to Propionibacterium acnes. Interact Cardiovasc Thorac Surg 2016; 23(1): 150-5.

Wang B, Toye B, Desjardins M, Lapner P, Lee C. A 7-year retrospective review from 2005 to 2011 of Propionibacterium acnes shoulder infections in Ottawa, Ontario, Canada. Diagn Microbiol Infect Dis 2013; 75(2): 1959.

Zeller V, Ghorbani A, Strady C, Leonard P, Mamoudy P, Desplaces N. Propionibacterium acnes: an agent of prosthetic joint infection and colonization.

J Infect 2007; 55(2): 119-24.

Acta Orthopaedica 2018; 89 (5): 585–590

585

An increase in myeloid cells after severe injury is associated with normal fracture healing: a retrospective study of 62 patients with a femoral fracture Lillian HESSELINK 1, Okan W BASTIAN 1, Marjolein HEERES 1, Maarten TEN BERG 2, Albert HUISMAN 2, Imo E HOEFER 2, Wouter W VAN SOLINGE 2, Leo KOENDERMAN 3, Karlijn J P VAN WESSEM 1, Luke P H LEENEN 1, and Falco HIETBRINK 1 1 Department of Trauma Surgery, University Medical Center Utrecht, Utrecht; 2 Department of Clinical Chemistry and Hematology, University Medical Center Utrecht, Utrecht; 3 Laboratory for Translational Immunology and Department of Respiratory Medicine, University Medical Center Utrecht Wilhelmina Children’s Hospital, Utrecht, the Netherlands Correspondence: l.hesselink@umcutrecht.nl Submitted 2018-02-20. Accepted 2018-07-02.

Background and purpose — Nonunion is common in femoral fractures. Previous studies suggested that the systemic immune response after trauma can interfere with fracture healing. Therefore, we investigated whether there is a relation between peripheral blood cell counts and healing of femur fractures.

Patients and methods — 62 multi-trauma patients with a femoral fracture presenting at the University Medical Centre Utrecht between 2007 and 2013 were retrospectively included. Peripheral blood cell counts from hematological analyzers were recorded from the 1st through the 14th day of the hospital stay. Generalized estimating equations were used to compare outcome groups.

Results — 12 of the 62 patients developed nonunion of the femoral fracture. The peripheral blood-count curves of total leukocytes, neutrophils, monocytes, lymphocytes, and platelets were all statistically significantly lower in patients with nonunion, coinciding with significantly higher CRP levels during the first 2 weeks after trauma in these patients.

Interpretation — Patients who developed femoral nonunion after major trauma demonstrated lower numbers of myeloid cells in the peripheral blood than patients with normal fracture healing. This absent rise of myeloid cells seems to be related to a more severe post-traumatic immune response.

Nonunion has been reported in one-tenth of patients with femoral fractures. This risk further increases in cases of multiple fractures and open fractures, as frequently seen in multitrauma (Zura et al. 2016).

Local factors, such as severe soft tissue injury and reduced weight-bearing on the affected extremity, can impair bone healing (Karladani et al. 2001, Taitsman et al. 2009, Zura et al. 2016). In addition, an increasing body of evidence suggests that the systemic immune response can also influence bone healing (Bastian et al. 2011). For instance, blunt chest injury in an experimental setting or intraperitoneal injection of lipopolysaccharides, which are both models for systemic inflammation, impaired fracture healing in animal models (Reikerås et al. 2005, Claes et al. 2011, Recknagel et al. 2013). However, the exact mechanism underlying the fracture healing impairment after systemic inflammation remains unknown.

Secondary bone healing consists of at least 4 different stages: the inflammatory phase, soft callus formation, hard callus formation, and tissue remodeling. During the inflammatory phase, neutrophils and macrophages are recruited to the fracture hematoma within days up to a week after injury (Li et al. 2016, Loi et al. 2016). The inflammatory phase normally ends within a week, after which the formation of callus starts (Marsell and Einhorn 2011, Loi et al. 2016). Disruption of the inflammatory process, for example by sustained inflammation, may interfere with the consecutive stages of bone healing and, thereby, increase the risk of nonunion (Loi et al. 2016).

It is unclear if a correlation exists between the cellular systemic immune response after trauma and femoral fracture healing. Hence, we investigated whether peripheral blood cell counts differ between multi-trauma patients with normal and impaired fracture healing of the femur.

© 2018 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/4.0) DOI 10.1080/17453674.2018.1501974

586

Acta Orthopaedica 2018; 89 (5): 585–590

Patients and methods Study design and setting Patients were selected from the trauma registry database of the University Medical Center (UMC) Utrecht which collected data of all patients who were admitted to the trauma department. Patients admitted between January 1, 2007 and the December 31, 2013, were included. Hemoglobin concentrations and total number of leukocytes, neutrophils, monocytes, platelets, erythro­ cytes and lymphocytes were compared during the first 2 weeks after injury between multi-trauma patients with nonunion and multi-trauma patients with normal healing of the femoral fracture. In addition, the acute phase protein C-reactive protein (CRP) and reticulocyte counts were obtained. Outcome data concerning healing of the femoral fracture were obtained from the electronic medical record system 1 year after trauma.

Participants Multi-trauma patients ≥ 16 years of age with a femoral shaft or distal femoral fracture presenting in the emergency department (ED) of the UMC Utrecht were included.

Exclusion criteria were: (1) transfer to another hospital and (2) non-weight bearing of the affected extremity, for instance due to paresis, amputation, or severe head injury.

Also, patients were excluded if the healing outcome could not be determined due to death or loss to follow-up. Data concerning patient characteristics, trauma mechanism, injuries, and treatment were obtained from the trauma registry database and supplemented with information from the electronic medical record system.

Baseline characteristics. Data are shown as number (percentage), median [range] or mean [standard deviation]

Factor

Total Union Nonunion (n = 62) (n = 50) (n = 12)

Sex, male 46 (74) 38 (76) 8 (67) Age 32 [16–85] 30 [16–85] 40 [11.7] Injury Severity Score 25 [17–48] 26 [17–48] 24 [4.9] New Injury Severity Score 27 [17–50] 27 [17–50] 28 [6.2] Femur fracture localization Shaft 43 (69) 36 (72) 7 (58) Distal 19 (31) 14 (28) 5 (42) Type of femur fractures (AO): Simple extra-articular 27 (44) 29 (60) 6 (50) Complex extra-articular 21 (34) 10 (20) 3 (25) Intra-articular 14 (23) 11 (22) 3 (25) Soft tissue injury (Gustilo) 0: closed fracture 35 (57) 29 (58) 6 (50) 1: wound < 1cm 8 (13) 5 (10) 3 (25) 2: wound > 1 cm 6 (10) 5 (10) 1 (8) 3a: adequate soft tissue cover 4 (7) 3 (6) 1 (8) 3b: inadequate soft tissue cover 2 (3) 2 (4) 0 3c: associated arterial injury 0 0 0 Unknown 7 (11) 6 (12) 1 (8) Type of fixation External fixation + IMN 6 (10) 5 (10) 1 (8) External fixation + plates 9 (15) 7 (14) 2 (17) External fixation + screws 1 (2) 1 (2) 0 IMN 35 (57) 30 (60) 5 (42) Plates 10 (16) 6 (12) 4 (33) Screws 1 (2) 1 (2) 0 Number of surgical procedures 3 [1–18] 3 [1–13] 3 [1–18] Inflammatory complications Urinary tract infection 6 (10) 5 (10) 1 (8) Surgical site infection 3 (5) 2 (4) 1 (8) Pneumonia 10 (16) 9 (18) 1 (8) MODS 3 (5) 2 (4) 1 (8) Nonunion 13 (21) Atrophic 11 (18) 11 (92) Hypertrophic 1 (2) 1 (8) Infected nonunion 2 (3) 2 (17) ICU stay (days) 2 [0–68] 2 [0–46] 0 [0–68] Hospital stay (days) 20 [4–154] 20 [4–95] 16 [4–154]

p-value 0.5 0.2 0.3 0.6 0.5 0.5 0.8 0.7 1.0 0.5 0.2 1.0 1 1

1 1 1 0.4 0.1 1 0.4 1 0.5 0.7 0.5

0.2 0.3

Baseline variables of patients with nonunion are compared with baseline variables of patients with union with the use of Fisher’s exact test or Mann-Whitney U test as indicated.

IMN = intramedullary nailing, ICU = intensive care unit, MODS = multiple organ dysfunction syndrome.

Patients ≥ 16 years with ISS ≥ 16 and distal or shaft femoral fracture 01/01/2007 – 12/31/2013 n = 100 Excluded (n = 38): – dead < 1 month after trauma, 21 – no weight bearing on affected leg, 6 – lost to follow-up, 11 Patients included n = 62

Figure 1. Flowchart of patients who met inclusion/exclusion criteria for the study population. ISS = Injury Severity Score.

100 multi-trauma patients with a distal or shaft femoral fracture were enrolled in the study (Figure 1). Multi-trauma was defined as an Injury Severity Score (ISS) ≥ 16 (Baker et al.

1974). 11 patients were lost to follow-up, of which 8 were transferred to another hospital, 2 did not visit the outpatient clinic, and in one case the information in the patient record system was insufficient to determine healing outcome. 21 patients died.

Causes of death were severe traumatic injuries (n = 18), inflammatory complications (n = 2), and unknown (n = 1). 6 patients were excluded because of the inability to bear weight on the affected leg due to severe ipsilateral injuries, an amputation, severe head trauma, or spinal cord injury. Of the remaining 62 patients, union of the femoral fracture was seen in 50 patients (50/62) and nonunion was seen in 12 patients (Table).

Acta Orthopaedica 2018; 89 (5): 585–590

Procedures Hematological parameters were obtained from the Utrecht Patient Oriented Database (UPOD). Data were collected from the day patients arriving in the Emergency Department (ED) through the 14th day of their hospital stay. The technical details of the UPOD are described elsewhere (ten Berg et al. 2007).

In short, this database is an infrastructure of relational databases that allows (semi-)automated transfer, processing and storage of data, including administrative information, medical and surgical procedures, medication orders, and laboratory test results for all clinically admitted patients and patients attending the outpatient clinic of the UMC Utrecht since 2004.

The process and storage of data are in accordance with privacy and ethics regulations. UPOD data acquisition and data management are in line with current Dutch regulations concerning privacy and ethics and are approved by the institution’s medical ethics committee. Because no extra material, such as blood samples, was taken from patients, there was no requirement to obtain informed consent from individual patients. The data were analyzed anonymously. Routine hematological analysis was performed using the Cell-Dyn Sapphire hematology analyzer (Abbott Diagnostics, Santa Clara, CA, USA) (ten Berg et al. 2007). The reliability and validity of the laboratory results were monitored through routine quality control.

Variables and outcome measures The study outcome was femoral fracture healing. Union was defined as pain-free mobilization (clinical union) or bridging of 3 of the 4 cortices (radiological union) within 12 months after injury. Nonunion was defined as lack of radiological and clinical union within 12 months after trauma or a fracture which required a re-intervention to achieve union. Peripheral blood cell counts of multi-trauma patients with union were compared with peripheral blood cell counts of multi-trauma patients with nonunion. Soft tissue injury was scored according to the Gustilo classification (Gustilo et al. 1984).

Statistics Data were analyzed with IBM SPSS version 23 (IBM Corp, Armonk NY, USA). Descriptive statistics are presented as median (range) for non-normally distributed variables and mean (SD) for normally distributed variables. Comparison of baseline variables between outcome groups was performed with Fisher’s exact test for categorical variables or a Mann–Whitney U test for the continuous data that were not normally distributed. Statistical significance was defined as a p-value < 0.05. Since the design of the study is longitudinal with repeated measurements, we chose linear generalized estimating equations (GEE) to compare the development over time of hematological parameters between outcome groups.

This linear analysis was performed to analyze whether the course of hematological parameters differed between the outcome groups during the first 2 weeks after trauma. The GEE was used to account for within-subject correlation between

587

repeated measurements. Based on spaghetti plots we chose the autoregressive working correlation structure for platelets and the exchangeable working correlation structure for the other hematological parameters.

Ethics, funding, and potential conflicts of interest A waiver was provided by the institutional medical ethics committee for this study. In addition, in line with the academic hospital policy, an opt-out procedure is in place for use of patient data for research purposes. The process and storage of data are in accordance with privacy and ethics regulations.

Financial support was not received. No conflicts of interests were declared.

Results Demographics There were no statistically significant differences in sex, age, ISS, new injury severity score (NISS), fracture localization, type of fracture, soft tissue injury, type of fixation, number of surgical procedures, inflammatory complications, length of stay in intensive care unit (ICU), and total hospital stay between patients with normal and impaired fracture healing of the femur.

Peripheral blood cell counts (Figure 2) Neutrophil and leukocyte counts were similarly elevated in both outcome groups upon arrival in ED and decreased to normal values within 1 day. In patients with union, mean leukocyte, neutrophil, monocyte, and platelet counts rose above reference values in the second week after trauma. In contrast, leukocyte, neutrophil, monocyte, and platelet counts of patients with nonunion remained within reference values. Lymphocyte counts remained within reference values in patients with union and decreased to just below reference values in patients with nonunion. When compared with the union group, in the nonunion group there was an average change in leukocytes of –0.33/day (p = 0.03), neutrophils of –0.39/day (p = 0.03), monocytes of –0.03/day (p = 0.03), platelets of –21/day (p = 0.001) and lymphocytes of –0.04/day (p = 0.02). Hemoglobin and erythrocytes decreased after trauma and remained below reference values for both outcome groups.

Reticulocyte count and CRP level (Figure 3) Figure 3 show reticulocyte counts and CRP levels during the first 2 weeks after trauma for patients with union and nonunion of the femoral fracture. In both outcome groups, CRP levels rose to 200–300 within 3 days, and gradually decreased thereafter. After day 2, higher CRP levels were observed in patients who later developed nonunion than in patients with union. The average change in CRP levels in patients with nonunion was 7.4/day (p = 0.01) compared with CRP levels of patients with union. No statistically significant differences were observed

588

A

Acta Orthopaedica 2018; 89 (5): 585â&#x20AC;&#x201C;590

B

A

B

Figure 3. Reticulocyte count (A) and C-reactive protein (B) during the first 2 weeks after major trauma. Patients with union are depicted in green and patients with nonunion in red. Data are presented as mean with standard error of the mean. Grey bars represent reference values.

Discussion

C

D

E

F

G

Figure 2. Leukocyte count (A), neutrophil count (B), monocyte count (C), lymphocyte count (D), erythrocyte count (E), hemoglobin (F), and platelet count (G) during the first 2 weeks after major trauma. Patients with union are depicted in green and patients with nonunion in red.

Data are presented as mean with standard error of the mean.

Grey bars represent reference values.

between outcome groups in the number of infections and the number of severe infections leading to multiple organ dysfunction syndrome. Reticulocyte count rose in both outcome groups after trauma, and was similar in patients with union and patients with nonunion.â&#x20AC;&#x192;

Leukocytes, neutrophils, monocytes, and platelets were above reference values in patients with normal fracture healing during the second week after severe injury. Patients with nonunion did not exhibit such an increase in myeloid blood cells and exhibited a statistically significant, but minor, decrease in lymphocytes. Although CRP levels were elevated in both outcome groups, there was a small but statistically significant increase in CRP in the nonunion group compared with the union group.

An increase of myeloid cells after trauma, as seen in patients with union, has been described before (Manz and Boettcher 2014, Bastian et al. 2016b, Loftus et al. 2017). Moreover, a previous study found similar trends in peripheral blood cell counts in patients with and without union of their tibia fracture (Bastian et al. 2016b). This study additionally investigated CRP and reticulocyte count. CRP provided information on inflammation and reticulocyte count reflected the production of immature red blood cells from the bone marrow, and can thus be used as an indicator of bone marrow function (Livingston et al. 2003, Piva et al. 2015). There are different hypotheses that can explain the lack of leukocytosis and thrombocytosis in the nonunion group. First, persistent inflammation might suppress the bone marrow response (Livingston et al.

2003). However, the increase found in reticulocytes, which did not differ between outcome groups, suggests an adequate bone marrow response and makes this hypothesis less likely.

Second, the lack of leukocytosis and thrombocytosis might be caused by persistent extravasation of myeloid cells to the tissues, a process associated with inflammatory conditions (Hietbrink et al. 2006, Johansson 2014).

Both hypotheses regarding the lack of leukocytosis and thrombocytosis in nonunion patients without bone marrow suppression are based on sustained inflammation. This is supported by the finding that nonunion patients had slightly higher CRP levels, while there was no statistically significant

Acta Orthopaedica 2018; 89 (5): 585–590

difference in clinically evident infections between outcome groups. Previous studies have demonstrated that a local controlled inflammatory reaction is key to successful bone healing (Schell et al. 2017) and that sustained systemic inflammation after trauma can impair this process (Recknagel et al. 2011, Claes et al. 2011). The influx of leukocytes in the fracture hematoma is an essential step during the inflammatory phase in the first week after trauma (Marsell and Einhorn 2011, Bastian et al. 2016a). However, termination of this phase to prevent persistent inflammation seems to be at least as important (Schmidt-Bleek et al. 2012, Loi et al. 2016). Decreased numbers of myeloid cells in the blood of nonunion patients during the second week after trauma supports the hypothesis of enhanced extravasation of these cells after the inflammatory phase and thus persistent local inflammation. This is in line with previous studies showing a relation between impaired fracture healing and increased numbers of pro-inflammatory leukocytes in the fracture hematoma (Schmidt-Bleek et al.

2012) and decreased numbers of leukocytes in the peripheral blood during the second week after trauma (Bastian et al. 2016b). Furthermore, previous studies showed that both the reduction of neutrophils in the fracture hematoma, and the inhibition of extravasation by blocking the anaphylatoxin C5a, improved fracture healing in rats (Grogaard et al. 1990, Chung et al. 2006, Recknagel et al. 2012). Taken together, it is tempting to speculate that increased extravasation of myeloid cells can disturb fracture healing and that this is reflected by decreased numbers of myeloid cells in the peripheral blood.

It is not surprising that conditions which further enhance the post-traumatic immune response, such as open fractures and multiple injuries, are additional risk factors for nonunion.

However, we did not find statistically significant differences in soft tissue injury and injury severity between the two outcome groups. Other factors that can influence peripheral blood cell counts, such as infectious complications and the number of surgical procedures, were also not significantly different.

An important limitation of this study is that blood values were retrospectively obtained and were therefore not available for each patient at each time point. It is possible that blood was more frequently drawn from patients who were more severely injured or from patients who developed complications during hospital stay. Yet, we did not find a relation between ISS or complications and healing outcome, precluding a substantial bias.

In summary, multi-trauma patients who developed femoral nonunion after major trauma demonstrated lower numbers of myeloid cells in the peripheral blood than patients with normal fracture healing. Patients with union demonstrated leukocyte numbers above reference values in the second week after trauma, reflecting a normal physiological response.

These findings support the hypothesis that persistent systemic inflammation after major injury can affect physiological processes necessary for bone healing.

589

FH, LPHL, KJPW, LK, MH, and LH developed the original study design.

AH, MB, IEH, WWS, and LH contributed to the data acquisition. OWB, FH, and LH contributed to the analysis and drafted the paper.

Acta thanks Anita Ignatius and other anonymous reviewers for help with peer review of this study.

Baker S P, O’Neill B, Haddon W, Long W B. The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma 1974; 14(3): 187–96.

Bastian O, Pillay J, Alblas J, Leenen L, Koenderman L, Blokhuis T. Systemic inflammation and fracture healing. J Leukoc Biol 2011; 89(5): 669–73.

Bastian O W, Koenderman L, Alblas J, Leenen L P H, Blokhuis T J. Neutrophils contribute to fracture healing by synthesizing fibronectin+ extracellular matrix rapidly after injury. Clin Immunol 2016a; 164: 78–84.

Bastian O W, Kuijer A, Koenderman L, Stellato R K, van Solinge W W, Leenen L P H, et al. Impaired bone healing in multitrauma patients is associated with altered leukocyte kinetics after major trauma. J Inflamm Res 2016b; 9: 69–78.

Chung R, Cool J C, Scherer M A, Foster B K, Xian C J. Roles of neutrophilmediated inflammatory response in the bony repair of injured growth plate cartilage in young rats. J Leukoc Biol 2006; 80(6): 1272–80.

Claes L, Ignatius A, Lechner R, Gebhard F, Kraus M, Baumgärtel S, et al.

The effect of both a thoracic trauma and a soft-tissue trauma on fracture healing in a rat model. Acta Orthop 2011; 82(2): 223–7.

Grogaard B, Gerdin B, Reikerfis O. The polymorphonuclear leukocyte: has it a role in fracture healing? Arch Orthop Trauma Surg 1990; 109(5): 268–71.

Gustilo R B, Mendoza R M, Williams D N. Problems in the management of type III (severe) open fractures: a new classification of type III open fractures. J Trauma 1984; 24(8): 742–6.

Hietbrink F, Koenderman L, Rijkers G, Leenen L. Trauma: the role of the innate immune system. World J Emerg Surg 2006; 1: 15.

Johansson M W. Activation states of blood eosinophils in asthma. Clin Exp Allergy 2014; 44(4): 482–98.

Karladani A H, Granhed H, Kärrholm J, Styf J. The influence of fracture etiology and type on fracture healing: a review of 104 consecutive tibial shaft fractures. Arch Orthop Trauma Surg 2001; 121(6): 325–8.

Li H, Liu J, Yao J, Zhong J, Guo L, Sun T. Fracture initiates systemic inflammatory response syndrome through recruiting polymorphonuclear leucocytes. Immunol Res 2016; 64(4): 1053–9.

Livingston D H, Anjaria D, Wu J, Hauser C J, Chang V, Deitch E A, et al.

Bone marrow failure following severe injury in humans. Ann Surg 2003; 238(5): 748–53.

Loftus T J, Mohr A M, Moldawer L L. Dysregulated myelopoiesis and hematopoietic function following acute physiologic insult. Curr Opin Hematol 2017; 25(1): 37–43.

Loi F, Córdova L A, Pajarinen J, Lin T hua, Yao Z, Goodman S B. Inflammation, fracture and bone repair. Bone 2016; 86: 119–30.

Manz M G, Boettcher S. Emergency granulopoiesis. Nat Rev Immunol 2014; 14(5): 302–14.

Marsell R, Einhorn T. The biology of fracture healing. Injury 2011; 42(6): 551–5.

Piva E, Brugnara C, Spolaore F, Plebani M. Clinical utility of reticulocyte parameters. Clin Lab Med 2015; 35(1): 133–63.

Recknagel S, Bindl R, Kurz J, Wehner T, Ehrnthaller C, Knöferl M W, et al.

Experimental blunt chest trauma impairs fracture healing in rats. J Orthop Res 2011; 29(5): 734–9.

Recknagel S, Bindl R, Kurz J, Wehner T, Schoengraf P, Ehrnthaller C, et al. C5aR-antagonist significantly reduces the deleterious effect of a blunt chest trauma on fracture healing. J Orthop Res 2012; 30(4): 581–6.

590

Recknagel S, Bindl R, Brochhausen C, Göckelmann M, Wehner T, Schoengraf P, et al. Systemic inflammation induced by a thoracic trauma alters the cellular composition of the early fracture callus. J Trauma Acute Care Surg 2013; 74(2): 531–7.

Reikerås O, Shegarfi H, Wang JE , Utvåg S E. Lipopolysaccharide impairs fracture healing: an experimental study in rats. Acta Orthop 2005; 76(6): 749–53.

Schell H, Duda G N, Peters A, Tsitsilonis S, Johnson K A, Schmidt-Bleek K. The haematoma and its role in bone healing. J Exp Orthop 2017; 4(1): 5.

Schmidt-Bleek K, Schell H, Schulz N, Hoff P, Perka C, Buttgereit F, et al.

Inflammatory phase of bone healing initiates the regenerative healing cascade. Cell Tissue Res 2012; 347(3): 567–73.

Acta Orthopaedica 2018; 89 (5): 585–590

Taitsman L A, Lynch J R, Agel J, Barei D P. Risk factors for femoral nonunion after femoral shaft fracture. J Trauma 2009; 67(6): 1389–92. ten Berg M J, Huisman A, Van den Bemt P M L A, Schobben A F A M, Egberts A C G, van Solinge W W. Linking laboratory and medication data: new opportunities for pharmacoepidemiological research. Clin Chem Lab Med 2007; 45(1): 13–19.

Zura R, Xiong Z, Einhorn T, Watson J T, Ostrum R F, Prayson M J, et al.

Epidemiology of fracture nonunion in 18 human bones. JAMA Surg 2016; 151(11): e162775.

Acta Orthopaedica 2018; 89 (5): 591–593

591

Surgical management of obturator neuropathy with a concomitant acetabular labral tear — a case report Shiho KANEZAKI 1, Akinori SAKAI 2, Eiichiro NAKAMURA 2, and Soshi UCHIDA 1 1 Department

of Orthopedic Surgery, Wakamatsu Hospital for the University of Occupational and Environmental Health, 2 Department of Orthopedic Surgery, University of Occupational and Environmental Health, Fukuoka, Japan Correspondence: soushi@med.uoeh-u.ac.jp Submitted 2018-02-24. Accepted 2018-05-08.

A 42-year-old woman presented to our hospital with a 3-month history of right groin pain after an accident. While cycling she collided with a car, which hit her frontally on the right side. She was immediately transferred to the hospital. Pain did not improve with NSAIDs and was becoming progressively worse. She was unable to perform deep squats, walk for more than10 minutes or climb stairs. The pain was located in the anterior and medial aspect of the right groin.

The patient had limited hip motion with 80° of flexion and 20° of abduction. With the hip in 90° flexion to the supine position, the patient had 5° of external rotation and 10° of internal rotation. Pain was reproduced on flexion adduction and internal rotation (FADIR) test, a positive impingement test. The hip dial test was positive.

Manual muscle strength (MMT) of the adductor muscles of the right hip was fair (3/5). Strengths of other hip muscles were normal.

An anterior-posterior (AP) pelvis and Dunn radiological views showed a lateral center edge (LCE) angle of 33°, an alpha angle of 69°, an offset ratio of 0.09, acetabular roof angle of 8° and absence of the crossover sign (an indicator of Cam type femoroacetabular impingement [FAI]). No other major abnormalities were observed. A 3D CT showed that the femoral version was 27° and the acetabular version was 22°.

MRI confirmed an antero-superior labral tear and the presence of a Cam lesion (Figure 1A). The patient was diagnosed with Cam-type FAI and concomitant labral tear. Since her pain was refractory to rest, NSAIDs, 3 months of physiotherapy, and a trial of 5 mL of 1% lidocaine injection into the hip joint, hip arthroscopy surgery was performed. A tear of the anterosuperior labrum was confirmed (Figure 1B) and repaired using 3 suture anchors (Gryphon BR, DePuy Synthes, Monument, CO, USA) (Figure 1C).

After releasing traction, a dynamic impingement test confirmed significant impingement on abduction and flexion.

Osteochondroplasty was then performed (Figure 1D). After osteochondroplasty, an intraoperative dynamic impingement test confirmed removal of the impingement and an effectively sealed labrum. Capsular repair was performed as described previously (Murata et al. 2017). The patient completed a reha-

Figure 1. Coronal view of the affected hip and arthroscopic finding showing an antero-superior acetabular labral tear. (A) T2-star coronal view showing the high intensity of an acetabular labral tear. (B) Arthroscopic findings from an anterolateral portal revealed an anterosuperior labral tear. (C) Labral repair with suture anchors. (D) View of Cam impingement. *FH: femoral head. AC: acetabulum.

bilitation protocol and full weight-bearing was allowed at 5 weeks after surgery. Postoperative Dunn radiological views showed an alpha angle of 45° , and an offset ratio of 0.18.

At 4 months postoperatively, the patient began to complain of medial thigh paresthesia and at 12 months postoperatively she reported improved range of motion (ROM) but continued discomfort with the same medial groin pain as before arthroscopy. Pain was localized to the inner side of the thigh over the adductor muscle, starting 2 cm distal to the inguinal ligament (Figure 2A). An MRI scan was performed, confirming appropriate healing of the labrum without evidence of any abnormality. There was nothing changed on this MRI from the preoperative MRI in the obturator nerve or magnus area.

MMT of the affected hip adductor was still fair (3/5).

© 2018 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/4.0) DOI 10.1080/17453674.2018.1494118

592

Acta Orthopaedica 2018; 89 (5): 591–593

Figure 2. Image of obturator nerve pain and operative findings during obturator nerve release. (A) Patients complained of medial thigh pain and numbness. (B) The anterior branch of the obturator nerve was identified over the adductor brevis.

Neurolysis of the obturator nerve (ON) was performed. View of adductor longus (ADD.L); arrow pointing to the overlying fascia. (C) Note the nerve being trapped by overlying fascia. Arrow: overlying facia.

B

C

A

An obturator neuropathy was suspected. Although electromyography (EMG) showed no evidence of denervation in the short and long adductor muscles, we suspected an obturator nerve entrapment. Using ultrasound, we injected 5 mL of 1% lidocaine and 3.1 mL of corticosteroid into the interfascial space between the adjacent pectineus and adductor brevis muscles (Anagnostopoulou et al. 2008). The patient reported a temporary improvement in her hip pain confirming the diagnosis of obturator nerve entrapment. An open decompression of the obturator nerve was performed via an anterior approach according to Bradshow et al. (1997). The nerve was strangulated by the overlying adhesive fascia (Figure 2B–C).

The patient was allowed to fully weight-bear immediately after surgery. The groin pain improved. At 3 months postoperatively, the hip ROM was improved, with flexion of 120°, abduction of 30°, internal rotation of 40°, and external rotation of 45°. The patient returned to work and became pain-free.

Discussion We report a case of obturator neuropathy that presented with groin pain similar to that of FAI. Following the arthroscopic procedure, her clinical hip scores including ROM improved.

However, the patient continued to complain of medial groin pain and paresthesia. The persistent medial groin pain was likely resulting from an obturator entrapment that was initially masked by her FAI symptoms. Therefore, when her FAI was addressed, her symptoms did not resolve completely.

Some studies have reported the use of electromyography (EMG) in patients with obturator nerve entrapment. The study

by Bradshaw et al. (1997) on athletes with chronic pain secondary to obturator nerve neuropathy reported EMG findings indicative of nerve denervation. Sorenson et al. (2002) and Rigaud et al. (2007) have also described positive EMG findings, which are important in diagnosing obturator nerve neuropathy. However, obturator nerve entrapment can present with normal findings on EMG (Rigaud et al. 2008) and in our case, EMG was normal.

Generally, first-line treatment for obturator neuropathy is non-operative (Tipton 2008) If non-operative modalities fail, surgical intervention with decompression of the obturator nerve may be indicated. Bradshow et al. (1997) have shown that surgically releasing obturator nerve entrapment was effective in 32 athletes with persistent groin pain.

Several studies have shown obturator neuropathy as a potential complication associated with gynecological, urological, and orthopedic surgeries (DeHart and Riley 1999, Hakoiwa et al. 2011). Obturator neuropathy has not been reported as one of the complications after hip arthroscopy (Harris et al.

2013, Kowalczuk et al. 2013). However, we have experienced obturator neuropathy after hip arthroscopic procedures (data not published). Hip traction during the arthroscopy in our case could have worsened the obturator nerve dysfunction.

In summary, surgeons should be mindful that obturator nerve entrapment can present as a source of groin pain and mimic FAI clinically.

Acta thanks Hal Martin for help with peer review of this study.

Anagnostopoulou S, Kostopanagiotou G, Paraskeuopoulos T, Alevizou A, Saranteas T. Obturator nerve block: from anatomy to ultrasound guidance. Anesth Analg 2008; 106(1): 350; author reply -1. doi: 10.1213/01. ane.0000297272.42192.70.

Bradshaw C, McCrory P, Bell S, Brukner P. Obturator nerve entrapment: A cause of groin pain in athletes. Am J Sports Med 1997; 25(3): 402–8.

DeHart M M, Riley L H, Jr. Nerve injuries in total hip arthroplasty. J Am Acad Orthop Surg 1999; 7(2): 101–11.

Hakoiwa S, Hoshi T, Tanaka M, Mishima H. [Case of bilateral obturator neuropathy after Caesarean section]. Masui 2011; 60(6): 721–3.

Harris J D, McCormick F M, Abrams G D, Gupta A K, Ellis T J, Bach B R, Jr, Bush-Joseph C A, Nho S J. Complications and reoperations during and after hip arthroscopy: a systematic review of 92 studies and more than 6,000 patients. Arthroscopy 2013; 29(3): 589–95. doi: 10.1016/j. arthro.2012.11.003 Kowalczuk M, Bhandari M, Farrokhyar F, Wong I, Chahal M, Neely S, Gandhi R, Ayeni O R. Complications following hip arthroscopy: a systematic review and meta-analysis. Knee Surg Sports Traumatol Arthrosc 2013; 21(7): 1669–75. doi: 10.1007/s00167-012-2184-2.

Acta Orthopaedica 2018; 89 (5): 591–593

Murata Y, Uchida S, Utsunomiya H, Hatakeyama A, Nakamura E, Sakai A. A comparison of clinical outcome between athletes and non-athletes undergoing hip arthroscopy for femoroacetabular impingement. Clin J Sport Med 2017; 27(4): 349–56. doi: 10.1097/JSM.0000000000000367.

Rigaud J, Labat J J, Riant T, Bouchot O, Robert R. Obturator nerve entrapment: diagnosis and laparoscopic treatment. Technical case report.

Neurosurgery 2007; 61(1): E175; discussion E. doi: 10.1227/01. neu.0000279743.45801.7e.

593

Rigaud J, Labat J J, Riant T, Hamel O, Bouchot O, Robert R. Treatment of obturator neuralgia with laparoscopic neurolysis. J Urol 2008; 179(2): 590–4; discussion 594–5.

Sorenson E J, Chen J J, Daube J R. Obturator neuropathy: causes and outcome. Muscle Nerve 2002; 25(4): 605–7.

Tipton S. Obturator neuropathy. Curr Rev Musculoskelet Med 2008; 1(3-4): 234–7. doi: 10.1007/s12178-008-9030-7.

Acta Orthopaedica 2018; 89

594

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

Copyright © Informa Healthcare Ltd 2018. ISSN 1745–3674. Printed in Sweden – all rights reserved.

Acta Orthopaedica 2018; 89

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-

595

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.

Format: 215 x 280 mm

1 Rastereinheit: 7 mm

69 %

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

69

Bone cement with gentamicin Bone and clindamycin * Sprowson AP et al. Bone Joint J 2016; 98-B: 1534â&#x20AC;&#x201C;1541

PRODUCTS & SOLUTIONS YOU CAN TRUST

www.heraeus-medical.com 1712_13136_AZ_COPAL_G+C_Image_Acta_Orthop_Europe_215x280_EN.indd 1

19.12.17 09:22

5/18 ACTA ORTHOPAEDICA Vol. 89, No. 5, 2018 (pp. 475â&#x20AC;&#x201C;593)

Volume 89, Number 5, October 2018 ISSN 1745-3674

IORT_89_05_Cover.indd 1

09-10-2018 12:16:36