Acta Orthopaedica Vol. 89 Issue 6, December 2018 by Acta

6/18 ACTA ORTHOPAEDICA

Vol. 89, No. 6, 2018 (pp. 595–701)

Volume 89, Number 6, December 2018 ISSN 1745-3674

Keep up-to-date with the latest research developments in Orthopaedics. Start subscribing to article alerts for Acta Orthopaedica 2017 IMPACT FACTOR:

3.080*

Log on to the Acta Orthopaedica homepage:

www.tandfonline.com/IORT Click on the New content alerts

button, at the top of the page (under the journal title)

Official publication by the Nordic Orthopaedic Federation *ÂŠ2017 Clarivate Analytics, 2017 release of the Journal Citation ReportsÂŽ

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. 6, 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. 6, December 2018 Annotation Patient-reported outcome after hip dislocation in primary total hip arthroplasty is virtually unknown: a systematic literature review Hip Early recovery trajectories after fast-track primary total hip arthroplasty: the role of patient characteristics Myocardial infarction following fast-track total hip and knee arthroplasty—incidence, time course, and risk factors: a prospective cohort study of 24,862 procedures Time of admission and mortality after hip fracture: a detailed look at the weekend effect in a nationwide study of 55,211 hip fracture patients in Norway Low-molecular-weight heparin for hip fracture patients treated with osteosynthesis: should thromboprophylaxis start before or after surgery? An observational study of 45,913 hip fractures reported to the Norwegian Hip Fracture Register Does the physical activity profile change in patients with hip dysplasia from before to 1 year after periacetabular osteotomy? Validated repeatability of patient-reported outcome measures following primary total hip replacement: a mode of delivery comparison study with randomized sequencing 18F-FDG-PET uptake in non-infected total hip prostheses Knee Higher mid-term revision rates of posterior stabilized compared with cruciate retaining total knee arthroplasties: 133,841 cemented arthroplasties for osteoarthritis in the Netherlands in 2007–2016 Blood cytokine pattern and clinical outcome in knee arthroplasty patients: comparative analysis 5 years after standard versus “hypoallergenic” surface coated prosthesis implantation Children Pelvic obliquity and measurement of hip displacement in children with cerebral palsy Knee joint sagittal plane movement in cerebral palsy: a comparative study of 2-dimensional markerless video and 3-dimensional gait analysis Timing for Ponseti clubfoot management: does the age matter? 90 children (131 feet) with a mean follow-up of 5 years Innovative treatment for pes cavovarus: a pilot study of 13 children Miscellanous Abaloparatide versus teriparatide: a head to head comparison of effects on fracture healing in mouse models High complication rate after extendible endoprosthetic replacement of the proximal tibia: a retrospective study of 42 consecutive children Vancomycin concentrations in the cervical spine after intravenous administration: results from an experimental pig study Trauma simulation training: a randomized controlled trial evaluating the effectiveness of the Imperial Femoral Intramedullary Nailing Cognitive Task Analysis (IFINCTA) tool Case reports Distal femoral shortening osteotomy for treatment of sciatic nerve palsy after total hip arthroplasty — a report of 3 cases Blount’s disease successfully treated with intraepiphyseal osteotomy with elevation of the medial plateau of the tibia—a case report with 65 years’ follow-up Information to authors (see http://www.actaorthop.org/)

595

L L Hermansen, M H Haubro, B L Viberg, and S Overgaard

597

J T Porsius, N M C Mathijssen, L C M Klapwijk–van Heijningen, J C van Egmond, M Melles, and S B W Vehmeijer P B Petersen, H Kehlet, and C C Jørgensen on behalf of the Lundbeck Foundation Center for Fast-track Hip and Knee Replacement Collaborative Group A Asheim, S M Nilsen, M Toch-Marquardt, K S Anthun, L G Johnsen, J H Bjørngaard

603 610 615

S Leer-Salvesen, E Dybvik, L B Engesæter, O E Dahl, and J-E Gjertsen

622

J Sandell Jacobsen, K Thorborg, P Hölmich, L Bolvig, S Storgaard Jakobsen, K Søballe, and I Mechlenburg C V E Carpenter, J Blackburn, J Jackson, A W Blom, A Sayers, M R Whitehouse

628 634

S J Gelderman, P C Jutte, R Boellaard, J J W Ploegmakers, D Vállez García, G A Kampinga, A W J M Glaudemans, and M Wouthuyzen-Bakker

640

A Spekenbrink-Spooren, L N van Steenbergen, G A W Denissen, B A Swierstra, R W Poolman, and R G H H Nelissen

646

P Thomas, P Hisgen, H Kiefer, U Schmerwitz, A Ottersbach, D Albrecht, B Summer, and C Schinkel

652

G Hägglund, M Goldring, M Hermanson, and E Rodby-Bousquet E Pantzar-Castilla, A Cereatti, G Figari, N Valeri, G Paolini, U Della Croce, A Magnuson, and J Riad

656 662

Y-B Liu, S-J Li, L Zhao, B Yu, and D-H Zhao

668

I Sanpera Jr, G Frontera-Juan, J Sanpera-Iglesias, and L Corominas-Frances

674

M Bernhardsson and P Aspenberg

678

P Tsagozis, M Parry, and R Grimer

683

M Bue, P Hanberg, M Tøttrup, M B Thomassen, H BirkeSørensen, T M Thillemann, T L Andersson, and K Søballe R Bhattacharyya, K Sugand, B Al-Obaidi, I Sinha, R Bhattacharya, and C M Gupte

689

696

B Puliero, W G Blakeney, Y Beaulieu, A Roy, and P-A Vendittoli

699

T Terjesen and D Anticevic

Acta Orthopaedica 2018; 89 (6): 595–596

595

Annotation

Patient-reported outcome after hip dislocation in primary total hip arthroplasty is virtually unknown: a systematic literature review Most patients have good to excellent outcomes after total hip arthroplasty (THA) due to osteoarthritis (OA). However, severe complications do still occur, and hip dislocation remains one of the most common reasons for revision surgery in the first postoperative years (Bozic et al. 2015, Singh et al. 2016). The incidence of hip dislocation after primary THA ranges from 1% to 10% (Dargel et al. 2014, Jorgensen et al. 2014, Petis et al. 2015, Zhang et al. 2015). Implant malposition or loosening is an obvious reason for surgical intervention after hip dislocations. In cases with no clear etiology, the non-surgical treatment is often prolonged, and the effect of the dislocation on daily activities and the subjective hip symptoms become more essential. The outcome after revision surgery due to recurrent dislocations is also not encouraging, as 10–34% of the revised patients re-dislocate (Wetters et al. 2013, Jo et al. 2015, Yoshimoto et al. 2017). In order to advise these patients properly, it is important to know the impact of 1 or recurrent dislocations on the patient’s quality of life and self-experienced hip function. This will contribute to an improved decision-making process for the patients, with no obvious cause for their hip dislocation. Thus, we conducted a systematic review of studies comparing patient-reported outcomes (PROs) in patients with a primary THA due to OA with and without episodes of hip dislocation. Method This review is reported with respect to the Preferred Reporting Items for Systematic reviews and Meta-Analysis (PRISMA) statements (Moher et al. 2009). The protocol was based upon the Preferred Reporting Items for Systematic reviews and Meta-Analysis Protocols (PRISMA-P) guidelines (Moher et al. 2015) and registered in the International Prospective Register of Systematic Reviews (PROSPERO) database (CRD42017076125). We searched Pubmed, Embase, SveMed, and Cochrane databases for relevant literature from the origin of each database and up to September 1, 2017. The reference list in each of the included studies was scanned for additional eligible studies. Studies were included in the review if the following criteria were fulfilled: • Study designs: randomized controlled trials (RCT), nonrandomized controlled clinical trials, prospective/retrospective cohort, and case-control studies (level of evidence 1–3). • Participants: OA as a primary diagnosis. • Outcomes: PRO after dislocation, either completely patient-

reported measurements or measures, where the patient part was reported separately from the clinician’s evaluation. • Studies published in English, German, and Scandinavian language. • No “prior to revision surgery” studies were included (selected patients). Risk of bias within cohort and case-control studies were assessed using the Critical Appraisal Skills Programme (CASP 2017) tool for relevant study designs. We performed the quality assessment of the eligible studies before data were extracted and this was carried out by 2 reviewers. Results We identified 3,460 unique studies using our broad search query. Throughout the title/abstract screening, 3,432 studies were excluded. We assessed full text of the remaining 28 articles for eligibility and, of these, only 2 studies (Forsythe et al. 2007, Edmunds and Boscainos 2011) met the inclusion criteria. No further studies were identified after screening the reference lists of the 2 included studies. The study by Forsythe et al. reported no statistically significant differences in the reduced WOMAC and SF-12 scores between the 32 patients with dislocation and the 64 patients without dislocation. However, the group without dislocation was significantly more satisfied postoperatively. Edmund and Boscainos primarily compared the anterolateral and posterior approaches. The combined results from patients with or without dislocations were not presented. The authors simply concluded that patients with dislocation lose approximately 5 points in total HHS, compared with non-dislocators. 3 of these points were represented by the function score since the HHS is only partly patient reported. No statistics were performed. A meta-analysis was not possible, since 4 different patient reported outcome measures (PROMs) had been used in the 2 included studies. Since only 2 studies met the inclusion criteria, comparing PROs in THA patients with/without hip dislocation, we aimed to extend the scope of the present review, and also to present papers covering PROM after dislocation, without comparisons. However, we found no additional studies presenting PROMs after a dislocation episode exclusively in THA patients with OA as the primary diagnosis. Discussion The goal of this systematic review was to provide valuable information regarding patient experience after dislocating a

© 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 10.1080/17453674.2018.1518428

596

primary THA. Our review revealed that knowledge of patientreported quality of life and subjective hip function after dislocation is merely non-existent. Efforts are ongoing to raise the use of PROMs in orthopedics from study to registry level across Europe. This will enable future prospective studies to evaluate the subjective importance of various complications (Paulsen 2014, Rolfson et al. 2016). A challenge though, is that closed reduction of prosthesis dislocation without revision is not reported in hip registries. Presumably, quality of life must be affected and continuously decreases with recurrent events. Likewise, confidence and trust in hip function and stability is impaired. These statements need scientific support. We are planning a larger scale study to identify differences in PRO for patients with single and recurrent dislocations. The original paper is unpublished and available on request to the corresponding author.

Lars L Hermansen 1–3, Martin H Haubro 1, Bjarke L Viberg 4, and Søren Overgaard 2,3 1 Department

of Orthopaedics, Hospital of South West Jutland, Esbjerg; 2 The Orthopaedic Research Unit, Department of Orthopaedic Surgery and Traumatology, Odense University Hospital, Odense; 3 Department of Clinical Research, Odense University Hospital, Odense; 4 Department of Orthopaedic Surgery and Traumatology, Kolding Hospital, Kolding; Correspondence: lars.lykke.hermansen@rsyd.dk

Acta Orthopaedica 2018; 89 (6): 595–596

Bozic K J, Kamath A F, Ong K, Lau E, Kurtz S, Chan V, Vail T P, Rubash H, Berry D J. Comparative epidemiology of revision arthroplasty: failed THA poses greater clinical and economic burdens than failed TKA. Clinical Orthop Relat Res 2015; 473(6): 2131-8. CASP. Critical Appraisal Skills Programme. CASP (Case-control and Cohort) Checklist. [online] Available at: http://www.casp-uk.net/casp-tools-checklists; 2017. Accessed: November 3, 2017. Dargel J, Oppermann J, Bruggemann G P, Eysel P. Dislocation following total hip replacement. Dtsch Arztebl Int 2014; 111(51-52): 884-90. Edmunds C T, Boscainos P J. Effect of surgical approach for total hip replacement on hip function using Harris Hip scores and Trendelenburg’s test. A retrospective analysis. Surgeon 2011; 9(3): 124-9. Forsythe M E, Whitehouse S L, Dick J, Crawford R W. Functional outcomes after nonrecurrent dislocation of primary total hip arthroplasty. J Arthroplasty 2007; 22(2): 227-30. Jo S, Jimenez Almonte J H, Sierra R J. The cumulative risk of re-dislocation after revision THA performed for instability increases close to 35% at 15 years. J Arthroplasty 2015; 30(7): 1177-82. Jorgensen C C, Kjaersgaard-Andersen P, Solgaard S, Kehlet H, Lundbeck Foundation Centre for Fast-track H, Knee Replacement Collaborative G. 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. Moher D, Liberati A, Tetzlaff J, Altman D G. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. J Clin Epidemiol 2009; 62(10): 1006-12. Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M, Shekelle P, Stewart L A. Preferred reporting items for systematic review and metaanalysis protocols (PRISMA-P) 2015 statement. Syst Rev 2015; 4: 1. Paulsen A. Patient reported outcomes in hip arthroplasty registries. Dan Med J 2014;61(5):B4845. Petis S, Howard J L, Lanting B L, Vasarhelyi E M. Surgical approach in primary total hip arthroplasty: anatomy, technique and clinical outcomes. Can J Surg 2015; 58(2): 128-39. Rolfson O, Bohm E, Franklin P, Lyman S, Denissen G, Dawson J, Dunn J, Eresian Chenok K, Dunbar M, Overgaard S, Garellick G, Lubbeke A, Patient-Reported Outcome Measures Working Group of the International Society of Arthroplasty R. Patient-reported outcome measures in arthroplasty registries. Report of the Patient-Reported Outcome Measures Working Group of the International Society of Arthroplasty Registries Part II. Recommendations for selection, administration, and analysis. Acta Orthop 2016; 87(eSuppl 362): 9-23. Singh J, Politis A, Loucks L, Hedden D R, Bohm E R. Trends in revision hip and knee arthroplasty observations after implementation of a regional joint replacement registry. Can J Surg 2016; 59(5): 304-10. Wetters N G, Murray T G, Moric M, Sporer S M, Paprosky W G, Della Valle C J. Risk factors for dislocation after revision total hip arthroplasty. Clinical Ortho Relat Res 2013; 471(2): 410-16. Yoshimoto K, Nakashima Y, Aota S, Kaneuji A, Fukui K, Hirakawa K, Nakura N, Kinoshita K, Naito M, Iwamoto Y. Re-dislocation after revision total hip arthroplasty for recurrent dislocation: a multicentre study. Int Orthop 2017; 41(2): 253-8. Zhang D, Chen L, Peng K, Xing F, Wang H, Xiang Z. Effectiveness and safety of the posterior approach with soft tissue repair for primary total hip arthroplasty: a meta-analysis. Orthop Traumatol Surg Res 2015; 101(1): 39-44.

Acta Orthopaedica 2018; 89 (6): 597–602

597

Early recovery trajectories after fast-track primary total hip arthroplasty: the role of patient characteristics Jarry T PORSIUS 1, Nina M C MATHIJSSEN 2, Lisette C M KLAPWIJK-VAN HEIJNINGEN 3, Jeroen C VAN EGMOND 2, Marijke MELLES 1, and Stephan B W VEHMEIJER 2 1 Faculty of Industrial Design Engineering, Delft University of Technology, Delft; 2 Department 3 Department of Physiotherapy, Reinier de Graaf Groep, Delft, the Netherlands

of Orthopedics, Reinier de Graaf Groep, Delft;

Correspondence: N.Mathijssen@rdgg.nl Submitted 2018-04-10. Accepted 2018-07-23.

Background and purpose — Little is known about heterogeneity in early recovery after primary total hip arthroplasty (THA). Therefore, we characterized subgroups of patients according to their hip function trajectory during the first 6 weeks after THA in a fast-track setting. Patients and methods — 94 patients (median age 65 years [41–82], 56 women) from a single hospital participated in a diary study. Patients recorded their severity of hip problems (Oxford Hip Score, OHS) weekly for 6 weeks after THA. Latent class growth modelling (LCGM) was used to identify patients with the same hip function trajectory and to compare these subgroups on patient characteristics. Results — LCGM revealed a fast (n = 17), an average (n = 53), and a slow (n = 24) recovery subgroup. Subgroups differed on the estimated weekly growth rate during the first 2 weeks (fast: 9.5; average: 5.3; slow: 2.7), with fewer differences between groups in the last 4 weeks (fast: 0.90; average: 2.0; slow: 1.7). Patients in the slow recovery group could be characterized as women of older age (mean age = 69) who rated their health as lower preoperatively, needed more assistance during recovery, and were less satisfied with the outcomes of the surgery. Interpretation — We identified distinct recovery trajectories in the first 6 weeks after fast-track primary THA which were associated with patient characteristics.

Between 7% and 15% of patients are dissatisfied with the results of total hip arthroplasty (THA) (Espehaug et al. 1998, Adie et al. 2012, Palazzo et al. 2014). Various studies indicate that preoperative patient characteristics, such as age (Dowsey et al. 2010), radiographic osteoarthritis (OA) severity (Keurentjes et al. 2013) and psychological factors (Judge et al. 2011), affect the outcome of THA. It is therefore important to better understand the heterogeneity in responses to THA in order to improve the quality of care for all patients. Most THA studies investigate postoperative functioning at 6 weeks, 3 months, or 1-year follow-up. However, less is known about differences in hip functioning during the first 6 weeks after discharge. We recently found that patients with a primary THA operated in a fast-track setting improved in self-reported pain and function during the first 6 weeks to clinically meaningful levels and were highly satisfied with the result (Klapwijk et al. 2017). However, a visual inspection of our data also revealed considerable heterogeneity in improvement. A valuable statistical method to improve insight into heterogeneous data is latent class growth modelling (LCGM) (Nagin and Odgers 2010). LCGM has been successfully applied to study responses to treatments in a variety of conditions (e.g. breast cancer [Smoot et al. 2016] and cardiovascular disease [Arthur et al. 2013]). In the present exploratory study, we aimed to characterize subgroups of patients according to their hip function trajectory in the first 6 weeks after primary THA in a fast-track setting. First, we used LCGM to study whether distinct early recovery patterns after THA can be identified. Second, we investigated whether early recovery patterns are associated with pre- and postoperative patient characteristics.

598

Patients and methods Study design and setting This study used data gathered in a recent prospective cohort study, which was conducted at the Reinier de Graaf Hospital (Delft, the Netherlands) (Klapwijk et al. 2017). Data were gathered with the aim to describe patients’ experiences with THA in a fast-track setting during the first 6 weeks after discharge. Patients were treated in an outpatient setting with the intention of leaving the hospital on the same day as surgery, except when 1 of the following exclusion criteria applied: ASA > 2; cardiovascular history; insulin-dependent diabetes mellitus and insufficient support from a caring person at home during the first postoperative night; a preference for an inpatient setting. All patients were operated through the anterior supine intermuscular approach (ASI) by the same orthopedic surgeon. Discharge criteria for all patients were functional and evaluated by a doctor, a physiotherapist, and a nurse. Patients had to be able to walk 30 meters with crutches or walker, to climb stairs if they were able to walk with crutches, to dress independently, and to go to the toilet independently. Adequate pain relief had to be achieved by means of oral medication before discharge, assessed with a numeric rating scale (NRS, 1–10) for pain below 3 at rest and below 5 during mobilization. Last, the wound had to be dry, or nearly dry, and the patient should not experience dizziness or nausea. A diary method was used for data collection; patients received a booklet after discharge with a variety of questions regarding their recovery, which they had to complete every day. To reduce patient burden not all questions were assessed daily. Preoperatively (approximately 1 week before surgery) a subset of questions was probed as part of standard routine. Participants Patients who were planned for primary unilateral THA between February 2015 and October 2015 were considered eligible for inclusion. Exclusion criteria were an insufficient command of the Dutch language, being mentally disabled, and a prosthesis in another joint of the ipsilateral or contralateral lower limb placed within 6 months before THA surgery. Of the 144 eligible patients, 43 did not participate in the study (n = 6, co-morbidities; n = 1, insufficient command of Dutch; n = 7, logistical reasons; n = 29, declined participation in the study). Of the 101 included patients, 7 were excluded from analyses for various reasons. Outcome measure The main outcome measure of this study was reported severity of problems with the operated hip, as measured with the Oxford Hip Score (OHS). Higher scores (0–48) indicate better functioning and less pain. Patients had to complete the OHS questionnaire at 1 week before surgery and once every week starting on the 7th day after discharge.

Acta Orthopaedica 2018; 89 (6): 597–602

Patient profiling We assessed a variety of patient characteristics to explore whether the identified subgroups based on OHS trajectories match the profile of a “typical” patient. These variables can be divided into preoperative or pre-discharge characteristics, healthcare utilization and assistance during the 6-week recovery period, and postoperative patient-reported experience measures (PREMs) after the 6-week recovery period. Preoperative or pre-discharge characteristics Preoperative demographic and patient characteristics included age, sex, BMI, and whether the patient lived alone. Preoperative radiographic OA severity was assessed by the last author (SV) using the Kellgren–Lawrence (K–L) grading (0–4) system. The Charlson comorbidity index (CCI) and the American Society of Anesthesiologists (ASA) physical status classification system (1–4) were used as co-morbidity measures. Health-related quality of life was measured with the EQ-5D (5L). All 5 individual domain scores (1–5) were used, where a higher score indicated more problems, as well as the overall health Visual Analog Scale (0–100; higher score indicated better health). Length of stay (LOS) in hospital was recorded, as well as surgical complications. Healthcare utilization and assistance Once a week patients recorded in their diary whether they had a physiotherapy session, a visit to the doctor, whether they needed aids (e.g., a walking stick or walker) to walk inside or outside their house and whether they received help (e.g., professionally or from a partner) with personal hygiene or with housekeeping. Pain medication was recorded daily. Postoperative PREMs At the end of the 6-week recovery period patients completed the PREMs in their diary. They indicated on an 11-point scale how satisfied they were with the results of surgery (0–10). On several 5-point scales they indicated their satisfaction with the general result of surgery, current pain, activities of daily living, and their quality of life. In addition, we asked whether they would opt for THA surgery again if needed (yes/no) and if they would recommend it to others (yes/no). Patient’s attitude towards fast-track THA, provided support during recovery, and the information provided were also assessed through yes/no questions. Statistics To investigate heterogeneity of patients’ responses to THA, we conducted LCGM analyses using a group-based trajectory modelling approach (Nagin and Odgers 2010) on the OHS assessed at 7 time-points. Robust full-information maximum likelihood estimation (Mplus version 6.12; https://www.statmodel.com/version6.12.shtml) was used to handle missing data. LCGM is an extension to latent growth curve models, which are commonly used to assess inter-individual differ-

Acta Orthopaedica 2018; 89 (6): 597–602

ences in intra-individual change (Duncan and Duncan 2004). In standard latent growth curve models 1 intercept (initial value) and 1 slope (change) is estimated to describe overall growth in a sample. In contrast, LCGM allows an investigator to estimate separate intercepts and slopes for a number of unobserved classes (i.e., subgroups). The goal of LCGM analyses is to identify different clusters in such a way that homogeneity is maximized within classes and heterogeneity between classes (Hoeksma and Kelderman 2006). Previous analyses on the same data showed a non-linear improvement in the OHS with a different rate of change from week 2 onwards (Klapwijk et al. 2017). We used a piecewise growth model specification (Chou et al. 2004) to accommodate for the non-linear growth. In such a model a growth process is cut into 2 (or more) pieces representing the average growth rate in a specific time period. In our case the 1st slope represents the average weekly growth rate between baseline and week 2, and the 2nd slope represents the average weekly growth rate between week 2 and 6. In line with the groupbased trajectory modelling approach, no random effects were estimated. We compared 1- to 4-class models and used conventional criteria to assess their relative statistical fit (BIC, BLRT, Entropy) (Nylund et al. 2007). Based on statistical fit, as well as theoretical and clinical interpretability of these models, 1 model was chosen as a basis to explore patient profiles. The most likely class membership based on posterior probabilities of the chosen model was used to assign every patient to the appropriate class. To compare classes, averages or percentages were calculated for all patient profiling characteristics. To assist the profiling process, we ran multinomial logistic regression analyses with class membership as outcome, and the specific characteristic in question (e.g., age) as predictor. As reference class the largest class will be chosen. Statistically significant differences (p < 0.05) should be interpreted as exploratory results and are displayed for descriptive purposes. Ethics, funding, and potential conflicts of interest The study protocol was assessed by the regional Medical Ethical Committee and no ethical approval was necessary, as the study did not fall under the scope of the Medical Research Involving Human Subjects Act. However, all included patients gave their written informed consent. We thank the Foundation for Scientific Research of Reinier de Graaf Groep, Delft, for funding this study. SV has a consultant contract with Zimmer Biomet. The authors have no further competing interests to declare.

Results Patient characteristics Patients were median 65 years (41–82) old and 56 were women (Table 1). Almost all patients were classified as ASA I

599

or II. Of all 94 patients 80 had no missing values on the OHS, 13 had a missing value on 1 occasion, and 1 on 2 occasions. Early recovery trajectories As can be seen from the BIC values and the outcomes of the BLRT, model fit improved by adding classes up to the 4-class model (Table 2, see Supplementary data). The entropy values all exceeded the minimum of 0.8, indicating that the classes were sufficiently distinct from each other. When accounting for the fit of the various models, as well as the clinical interpretation based on the intercept and slopes, we preferred the 3-class model. The 2-class model only distinguishes between an average and faster initial growth group and does not identify a—theoretically expected—slower recovery group. In the 3-class model this latter group is identified (Class 3), as well as an average (Class 2) and fast recovery group (Class 1), and statistical fit improved. Even though fit improved even further in a 4-class model, clinical interpretability did not improve. In this model the average recovery subgroup in the 3-class model was divided into 2 separate subgroups (Class 1 and 2) differing only marginally from each other in their initial growth rate. We therefore chose the 3-class model for further analyses. The average change in OHS from baseline until week 6 for the fast recovery group was 23, for the average recovery group 19, and for the slow recovery group 9.9 (Figure). The biggest difference between the classes became apparent in the first 2 weeks of recovery. Both the fast (Class 1) and average recovery group (Class 2) showed a steep initial increase in reported functioning of the operated hip in the first 2 weeks, levelling off in the subsequent 4 weeks. The levelling off phase was a little more pronounced in the fast recovery group. In stark contrast, the slow recovery group showed a slow and steady improvement, lacking the exponential increase in the first 2 weeks. In addition, growth seemed to already level off after week 4. It is also notable that the classes were similar in their preoperative baseline OHS. Differences between classes in patient characteristics There were distinct differences between the 3 different early recovery trajectory classes in preoperative characteristics, healthcare utilization and assistance during recovery, and postoperative PREMs. Average recovery patient profile The largest class of patients (Class 2, average recovery) can be characterized as slightly overweight patients (average BMI = 27) with an average age of 65, having an average preoperative movement-evoked NRS pain score of 7 and staying on average 1 night in hospital after surgery. Patients in this class on average did not report mental health problems preoperatively or problems with self-care, but they did report slight problems with mobility, usual activities, and pain. For the majority of the 6-week recovery period they

600

Acta Orthopaedica 2018; 89 (6): 597–602

Table 1. Pre- and postoperative characteristics of patients according to class membership

Class 1: Class 2: Class 3: Fast Average Slow recovery recovery recovery (n = 17) (n = 53) (n = 24)

Total (n = 94)

Preoperative or pre-discharge characteristics Female sex, n 8 29 19 a 56 Age, mean (SD) 59 (8.6) a 65 (8.3) 69 (7.3) a 65 (8.7) BMI, mean (SD) 27 (5.3) 27 (3.5) 28 (4.2) 27 (4.1) ASA score, n 1 9 18 7 34 2 8 32 15 55 3 0 3 2 5 Charlson comorbidity index, n 0 11 28 9 48 1 5 15 7 27 ≥ 2 1 10 8 19 Kellgren–Lawrence classification > 3 , n 12 41 16 69 a a Outpatient surgery, n 13 25 4 42 Length of stay in nights, median (IQR) 0.0 (0.5) a 1.0 (1.0) 1.0 (1.0) b 1.0 (1.0) Operated side, left, n 12 a 20 12 44 Discharged to home, n 17 52 22 91 Living alone, n 1 6 5 12 Additional prostheses, n 3 6 9 a 18 No surgical complications, n 17 51 22 90 Oxford Hip Score, mean (SD) 24 (7.5) 24 (8.5) 23 (9.4) 24 (8.5) EQ–5D, median (IQR) mobility (1–5) 2.0 (0.0) 2.0 (0.0) 2.0 (0.0) 2.0 (0.0) self-care (1–5) 1.0 (1.0) 1.0 (1.0) 1.0 (1.0) 1.0 (1.0) usual activities (1–5) 2.0 (1.0) 2.0 (0.0) 2.0 (0.0) 2.0 (0.0) pain/discomfort (1–5) 2.0 (0.0) 2.0 (1.0) 2.0 (1.0) 2.0 (1.0) mental health (1–5)) 1.0 (0.0) 1.0 (0.0) 1.0 (0.0) 1.0 (0.0) VAS (0–100), mean (SD) 70 (18) 71 (17) 61 (26) a 69 (20) NRS pain in rest (0–10), mean (SD) 4.4 (2.6) 4.6 (2.4) 6.0 (2.5) a 4.9 (2.5) NRS pain movement-evoked (0–10), mean (SD) 6.5 (2.0) 7.0 (1.9) 7.2 (1.9) 7.0 (1.9) Healthcare utilization and assistance during recovery Number of weeks, median (IQR) with attended physical therapy sessions 5.0 (1.0) 6.0 (1.0) 6.0 (1.0) 6.0 (1.0) with a doctor’s visit 0 0 0 0 with pain medication usage 3.0 (2.0) b 5.0 (3.0) 6.0 (1.0) a 5.0 (3.0) with personal hygiene assistance 2.0 (2.0) a 3.0 (4.0) 6.0 (2.8) a 3.5 (4.0) with housekeeping assistance 5.0 (3.0) 6.0 (1.0) 6.0 (1.0) 6.0 (1.0) b when walking aids were required inside 3.0 (2.5) 5.0 (2.0) 6.0 (1.0) 5.0 (3.0) when walking aids were required outside 2.0 (2.5) b 4.0 (2.0) 5.0 (1.0) a 4.0 (2.0) Postoperative PREMs c Satisfaction, median (IQR) with result THA (0–10) 10 (0.0) 10 (1.0) 8.0 (3.0) b 10 (2.0) with general result THA (1–5) 5.0 (0.5) 5.0 (1.0) 4.0 (0.0) b 5.0 (1.0) with current pain (1–5) 5.0 (0.5) 4.0 (1.0) 4.0 (2.0) b 4.0 (1.0) with activities of daily living (1–5) 5.0 (1.0) 4.0 (0.0) 4.0 (1.0) b 4.0 (1.0) b with quality of life (1–5) 5.0 (1.0) 4.0 (1.0) 3.5 (1.0) 4.0 (1.0) Would have THA again if needed, n 17 50 20 87 Would recommend to others, n 17 50 21 88 Positive attitude fast-track THA, n 15 47 18 80 a Positive attitude assistance during recovery, n 16 48 18 82 Positive attitude provided information, n 9 40 17 66 a p < 0.05, compared with the average recovery class b p < 0.01, compared with the average recovery class c PREMs: Patient reported experience measures

used walking aids, attended weekly physical therapy sessions, and took pain medication. After 6 weeks they were highly satisfied with the results of the surgery. In the subsequent analyses we compared the other smaller classes with this dominant class.

2. 2.

Fast recovery patient profile The patients in Class 1 (fast recovery) were younger (mean age = 59) and were operated more often in an outpatient setting. During the 6-week recovery they used pain medication and required walking aids for a shorter period (median = 3

Acta Orthopaedica 2018; 89 (6): 597–602

601

Oxford Hip Score (0–48) 48 42 36 30 24 18 12 Class 1: Fast recovery Class 2: Average recovery Class 3: Slow recovery

6 0

Preop.

Weeks after index operation

Trajectories of 6-week recovery after THA in a 3-class model (mean OHS [95% CI]).

weeks) than patients in the average trajectory class (median = 5 weeks). Slow recovery patient profile Patients in Class 3 (slow recovery) were more often female (four-fifths) and older (mean age = 69) than patients classified in the average recovery trajectory. They were less often operated on in an outpatient setting and rated their overall health as lower and reported more hip pain at rest before the surgery. More patients already had another prosthesis, and although not statistically significant, co-morbidities (ASA and CCI) were more common. Compared with the average recovery class patients required more weeks with personal hygiene assistance (median = 6 weeks) and longer use of walking aids outside (median = 5 weeks). After the 6-week recovery period patients were less satisfied with the results of the surgery and were also less positive about the support provided during recovery.

Discussion In this prospective cohort study we identified 3 different trajectories of early recovery after primary unilateral THA in a fast-track setting. In the slow recovery group, healthcare utilization and assistance during recovery was higher, in combination with lower satisfaction of provided care. Preoperatively, health perception was more negative and patients reported more pain at rest. Together with being older and having more co-morbidities, clinicians might describe this group of patients as being more “difficult” (Hahn et al. 1996, Jackson and Kroenke 1999) or challenging. The more complex preoperative clinical picture may have played a role in the slower recovery after THA. Pre-

vious work has for instance shown that preoperative pain at rest (as opposed to movement-evoked pain) is associated with chronic pain after THA (Sayers et al. 2016). Pain at rest could be an indication of altered central pain modulation (Baert et al. 2016), which might have played a role in this patient subgroup. At the other end of the spectrum the subgroup of patients displaying a fast recovery presented a very different profile. The majority of these patients left hospital on the day of surgery. In line with their pronounced improvement in pain and function during the first 2 weeks after discharge, they used pain medication and walking aids for a much shorter period of time than the average patient subgroup. Their quick recovery may in part be explained by having fewer co-morbidities and their younger age. However, these patients did not rate their health as better, or have less pain preoperatively. Perhaps being scheduled for the outpatient treatment and meeting the discharge criteria to leave on the same day as the surgery may have instilled more positive recovery expectations. Positive outcome expectations can enhance the effects of medical treatments (Crow et al. 1999) and may have led to the quicker recovery we observed in this subgroup of patients. Our results underscore the need for tailored care and provide directions to achieve such an outcome. Patients fitting the slow recovery trajectory were less satisfied with the surgery and less positive about the support provided during recovery. Although we are not able to tell what kind of support this subgroup of patients wanted, their higher healthcare utilization suggests unfulfilled needs when it comes to pain management and assistance with daily activities. This is in line with previous work stressing the importance of postoperative pain management after THA to improve patient satisfaction (Bergés et al. 2006). It may be of value to investigate this subgroup’s needs for care management and cost-effective ways for improvement. The most important limitations concern the generalizability of our findings and that we did not try to identify potentially important preoperative psychosocial predictors of early recovery. Being an exploratory study with a relatively small sample size, replication of our findings in other larger samples in different settings is essential. Previous studies have pointed to the importance of psychosocial predictors of responses to arthroplasty, such as pain catastrophizing or psychological distress (Vissers et al. 2012, Lewis et al. 2015). A recent controlled cohort study demonstrated a positive effect of providing psychological support on well-being and physiotherapy objectives after THA (Tristaino et al. 2016). When looking at the broader treatment spectrum for OA, a recent meta-analysis calculated that 75% of all treatment effects were due to contextual effects (i.e., placebo effects) (Zou et al. 2016). This is in line with studies showing that patients with higher preoperative pain relief expectations showed better outcomes after arthroplasty (Gandhi et al. 2009, Judge et al. 2011). The only preoperative psychosocial predic-

602

tor in our study was self-rated mental health, measured with 1 question from the EQ-5D questionnaire. We did not find a difference between the groups; however, a more comprehensive measurement of psychosocial factors might have revealed more differences between the recovery groups. In summary, 3 distinct subgroups of patients were identified with markedly different recovery rates. By comparing these groups we were able to develop preliminary patient profiles of early recovery. Our study demonstrates the necessity to further explore the specific needs of these different subgroups, as well as clinical and psychosocial predictors of early recovery. Supplementary data Table 2 is available available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674. 2018.1519095

The authors are grateful to all the patients for their efforts in completing their diaries, and especially thank N. de Esch for her help during the study. JP performed data analysis, wrote and revised the manuscript, NM designed the study, supported data analysis and reviewed the manuscript. LK supported data collection and reviewed the manuscript. JE helped composing the diary, performed data collection, and reviewed the manuscript. MM supported the data analysis and reviewed the manuscript. SV designed the study, operated on all patients and reviewed the manuscript. Acta thanks Henrik Husted for help with peer review of this study.

Adie S, Dao A, Harris I A, Naylor J M, Mittal R. Satisfaction with joint replacement in public versus private hospitals: a cohort study. ANZ J Surg 2012; 82(9): 616-24. Arthur H M, Blanchard C, Gunn E, Kodis J, Walker S, Toner B. Exercise trajectories of women from entry to a 6-month cardiac rehabilitation program to one year after discharge. Biomed Res Int 2013; 2013: 121030. Baert I A C, Lluch E, Mulder T, Nijs J, Noten S, Meeus M. Does pre-surgical central modulation of pain influence outcome after total knee replacement? A systematic review. Osteoarthr Cartilage 2016; 24(2): 213-23. Bergés I M, Ottenbacher K J, Smith P M, Smith D. Perceived pain and satisfaction with medical rehabilitation after hospital discharge. Clin Rehabil 2006; 20(8): 724-30. Chou C P, Yang D, Pentz M A, Hser Y I. Piecewise growth curve modeling approach for longitudinal prevention study. Comput Stat Data An 2004; 46(2): 213-25. Crow R, Gage H, Hampson S, Hart J, Kimber A, Thomas H. The role of expectancies in the placebo effect and their use in the delivery of health care: a systematic review. Health Technol Assess 1999; 3(3): 1-96. Dowsey M M, Liew D, Stoney J D, Choong P F M. The impact of obesity on weight change and outcomes at 12 months in patients undergoing total hip arthroplasty. Med J Australia 2010; 193(1): 17-21.

Acta Orthopaedica 2018; 89 (6): 597–602

Duncan T E, Duncan S C. An introduction to latent growth curve modeling. Behav Ther 2004; 35(2): 333-63. Espehaug B, Havelin L I, Engesaeter L B, Langeland N, Vollset S E. Patient satisfaction and function after primary and revision total hip replacement. Clin Orthop Relat Res 1998; (351): 135-48. Gandhi R, Davey J R, Mahomed N. Patient expectations predict greater pain relief with joint arthroplasty. J Arthroplasty 2009; 24(5): 716-21. Hahn S R, Kroenke K, Spitzer R L, Brody D, Williams J B W, Linzer M, deGruy F V. The difficult patient. J Gen Intern Med 1996; 11(1): 1-8. Hoeksma J B, Kelderman H. On growth curves and mixture models. Infant Child Dev 2006; 15(6): 627-34. Jackson J L, Kroenke K. Difficult patient encounters in the ambulatory clinic: clinical predictors and outcomes. Arch Intern Med 1999; 159(10): 1069-75. Judge A, Cooper C, Arden N K, Williams S, Hobbs N, Dixon D, Günther K-P, Dreinhoefer K, Dieppe P A. Pre-operative expectation predicts 12-month post-operative outcome among patients undergoing primary total hip replacement in European orthopaedic centres. Osteoarthr Cartilage 2011; 19(6): 659-67. Keurentjes J C, Fiocco M, So-Osman C, Onstenk R, Koopman-Van Gemert A W M M, Pöll R G, Kroon H M, Vliet Vlieland T P M, Nelissen R G. Patients with severe radiographic osteoarthritis have a better prognosis in physical functioning after hip and knee replacement: a cohort-study. Baradaran HR, editor. PLoS One 2013; 8(4): e59500. Klapwijk L C M, Mathijssen N M C, Van Egmond J C, Verbeek B M, Vehmeijer S B W. The first 6 weeks of recovery after primary total hip arthroplasty with fast track. Acta Orthop 2017; 88(2): 140-4. Lewis G N, Rice D A, McNair P J, Kluger M. Predictors of persistent pain after total knee arthroplasty: a systematic review and meta-analysis. Br J Anaesth 2015; 114(4): 551-61. Nagin D S, Odgers C L. Group-based trajectory modeling in clinical research. Annu Rev Clin Psychol 2010; 6(1): 109-38. Nylund K L, Asparouhov T, Muthén B O. Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Struct Equ Modeling 2007; 14(4): 535-69. Palazzo C, Jourdan C, Descamps S, Nizard R, Hamadouche M, Anract P, Boisgard S, Galvin M, Ravaud P, Poiraudeau S. Determinants of satisfaction 1 year after total hip arthroplasty: the role of expectations fulfilment. BMC Musculoskelet Disord 2014; 15(1): 53. Sayers A, Wylde V, Lenguerrand E, Beswick AD, Gooberman-Hill R, Pyke M, Dieppe P, Blom AW. Rest pain and movement-evoked pain as unique constructs in hip and knee replacements. Arthritis Care Res 2016; 68(2): 237-45. Smoot B, Cooper BA, Conley Y, Kober K, Levine JD, Mastick J, Topp K, Miaskowski C. Differences in limb volume trajectories after breast cancer treatment. J Cancer Surviv 2016; 10(4): 772-82. Tristaino V, Lantieri F, Tornago S, Gramazio M, Carriere E, Camera A. Effectiveness of psychological support in patients undergoing primary total hip or knee arthroplasty: a controlled cohort study. J Orthop Trauma 2016; 17(2): 137-47. Vissers M M, Bussmann J B, Verhaar J A N, Busschbach J J V, Bierma-Zeinstra S M A, Reijman M. Psychological factors affecting the outcome of total hip and knee arthroplasty: a systematic review. Semin Arthritis Rheum 2012; 41(4): 576-88. Zou K, Wong J, Abdullah N, Chen X, Smith T, Doherty M, Zhang W. Examination of overall treatment effect and the proportion attributable to contextual effect in osteoarthritis: meta-analysis of randomised controlled trials. Ann Rheum Dis 2016; 75: 1964-70.

Acta Orthopaedica 2018; 89 (6): 603–609

603

Myocardial infarction following fast-track total hip and knee arthroplasty—incidence, time course, and risk factors: a prospective cohort study of 24,862 procedures Pelle B PETERSEN 1, Henrik KEHLET 1,2, and Christoffer C JØRGENSEN 1,2 on behalf of the Lundbeck Foundation Center for Fast-track Hip and Knee Replacement Collaborative Group* 1 Section

for Surgical Pathophysiology, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark; 2 Lundbeck Foundation Center for Fast-track Hip and Knee Arthroplasty Correspondence: Pelle.Petersen.01@regionh.dk * Members of the Lundbeck Foundation Centre for Fast-track Hip and Knee collaborative group: Frank Madsen, Department of Orthopedics, Aarhus University Hospital, Aarhus; Torben B. Hansen, Department of Orthopedics, Regional Hospital Holstebro and University of Aarhus, Holstebro; Henrik Husted, Department of Orthopedics, Hvidovre Hospital, Hvidovre; Mogens B. Laursen, Aalborg University Hospital Northern Orthopaedic Division, Aalborg; Lars T. Hansen, Department of Orthopedics, Sydvestjysk Hospital Esbjerg/Grindsted, Grindsted; Per Kjærsgaard-Andersen, Department of Orthopedics, Vejle Hospital, Vejle; Soren Solgaard, Department of Orthopedics, Gentofte University Hospital, Copenhagen; Niels Harry Krarup, Department of Orthopedics, Viborg Hospital, Viborg, Denmark Submitted 2018-05-01. Accepted 2018-07-12.

Background and purpose — Acute myocardial infarction (MI) is a leading cause of mortality following total hip and knee arthroplasty (THA/TKA). The reported 30-day incidence of MI varies from 0.3% to 0.9%. However, most data derive from administration and insurance databases or large RCTs with potential confounding factors. We studied the incidence of and potential modifiable risk factors for postoperative MI in a large, multicenter optimized “fasttrack” THA/TKA setting. Patients and methods — A prospective cohort study was conducted on consecutive unselected elective primary unilateral THA and TKA, using prospective information on comorbidities and complete 90-day follow-up from the Danish National Patient Registry. Evaluation of discharge summaries and medical records was undertaken in cases of suspected MI. Logistic regression analyses were carried out for identification of preoperative risk factors. Results — Of 24,862 procedures with a median length of stay 2 (IQR 2–3) days, 30- and 90-day incidence of MI was 31 (0.12%) and 48 (0.19%). Preoperative risk factors for MI ≤ 30 days were age > 85 years (OR 7.4, 95% CI 2.3– 24), insulin-dependent diabetes mellitus (IDDM) (3.6, CI 1.1–12), cardiovascular disease (2.4, CI 1.1–5.0) and hypercholesterolemia (2.3, CI 1.1–5.1). Of 31 patients with MI ≤ 30 days 9 were treated with vasopressors for intraoperative hypotension and 27 had postoperative anemia. Interpretation — Fast-track THA and TKA had a low 30-day MI incidence. Focus on patients with age > 85, IDDM, cardiovascular disease, and hypercholesterolemia may further reduce the 30-day incidence of MI. The role of postoperative anemia and intraoperative hypotension are other areas for further improvement

Acute myocardial infarction (MI) is a serious complication after total hip and knee arthroplasty (THA/TKA). Despite improvements in perioperative optimization and enhanced recovery throughout the last decades (Kehlet 2013), MI remains one of the most frequent causes of mortality following THA and TKA (Malviya et al. 2011, Kirksey et al. 2012, Bemenderfer et al. 2017, Jørgensen et al. 2017). Recent studies investigating the occurrence of postoperative MI found varying incidences both in hospital (0.2%–0.7%) (Menendez et al. 2015, Bemenderfer et al. 2017, Urban et al. 2017), after 30 days (0.3%–0.9%) (Singh et al. 2011, Belmont Jr et al. 2014, Pedersen et al. 2014, Thornqvist et al. 2014), and after 6 weeks (0.5% for THA and 0.2% for TKA) (Lalmohamed et al. 2012). In contrast, a recent large observational study on fast-track THA and TKA has shown a low 0.2% incidence of MI (Jørgensen et al. 2016b) and another observational study found a marked decrease in 30-day incidence of MI from around 0.9% before to 0.4% after implementation of a fast-track protocol (Khan et al. 2014). Furthermore, a recent study on early thromboembolic events after fast-track THA and TKA reported an incidence of only 0.07% for MI either in hospital or within the first postoperative week, with a minor increase to 0.10% within 30 days of surgery (Jørgensen et al. 2016a). However, most of the established evidence on incidence and risk factors for MI following THA and TKA derive from administrative databases relying on diagnostic codes or insurance data (Kirksey et al. 2012, Lalmohamed et al. 2012, Belmont Jr et al. 2014, Thornqvist et al. 2014, Menendez et al. 2015, Bemenderfer et al. 2017, Urban et al. 2017). These databases lack detailed information on perioperative care, which limits the ability to identify potential pathophysiological events leading to MI. Additionally, a novel comparison of

604

NSQIP data and administrative databases has shown a low specificity of around 81% for both, and with chart review of a single complication as the “gold standard” (Etzioni et al. 2018). Consequently, we present results from a large prospective multicenter database on consecutive unselected elective primary THA and TKA in established fast-track settings with complete 90-day follow-up and chart review (Jørgensen and Kehlet 2013). We investigated the incidence of postoperative MI, potential modifiable preoperative risk factors, perioperative events disposing to the occurrence of MI, and MI related mortality after fast-track THA and TKA.

Patients and methods This was an observational cohort study on prospectively collected consecutive unselected primary THA and TKA patients between January 2010 and October 2015. All patients were treated at 8 Danish orthopedic centers (> 200 procedures per year) with established fast-track protocols and reporting to the Lundbeck Foundation Centre for Fast-track Hip and Knee Replacement database (LCDB) (Jørgensen and Kehlet 2013). The fast-track setup included regional anesthesia, multimodal opioid-sparing analgesia, in-hospital thromboprophylaxis, and early mobilization with discharge to own home (> 95%) when fulfilling a set of functional discharge criteria (Jørgensen et al. 2013). Data on preoperative comorbidity and patient demographics were prospectively collected by patient reported questionnaires. Length of stay (LOS) (postoperative nights in hospital, including transfer to other departments and hospitals), 90-day readmission (≥ 1 night in hospital, and potentially surgically related) and mortality (Jørgensen and Kehlet 2013) were obtained from the Danish National Patient Registry (DNPR) (Schmidt et al. 2015). In case of LOS ≥ 5 days, readmission, or death within 90 days discharge summaries were analyzed. Furthermore, in all cases of suspected MI in the discharge summary or a diagnostic code (ICD10) related to MI in the DNPR, the complete medical records were reviewed. MI was classified according to the proposed EPCO definitions (Jammer et al. 2015). Additionally, in cases with MI ≤ 30 days results from blood samples and intraoperative anesthesia notes were acquired in order to identify any potential association between intraoperative events, anemia, and postoperative MI. Pre- and postoperative anemia were defined as hemoglobin (HB) < 13 g/dL for both male and females (Muñoz et al. 2017) and intraoperative hypotension was defined as an intraoperative systolic blood pressure < 90 mmHg classified as dichotomous outcome (Sessler et al. 2017). Furthermore, in case of death the complete medical records on primary admission and eventual readmissions were obtained to establish cause of death and potential relation to index surgery. Finally, we obtained and analyzed discharge papers of patients with new postoperative administrations of anticoagulants through the Danish National Database of Reimbursed Prescriptions (Johannesdot-

Acta Orthopaedica 2018; 89 (6): 603–609

tir et al. 2012), to establish indication for the prescription. All discharge papers and medical records were initially analyzed by either the first (PBP) (patients from October 2013–November 2015) or the senior author (CCJ) (January 2010–September 2013). In cases of doubt the papers were evaluated by all authors until agreement was reached. Outcomes Primary outcome was 30-day incidence of MI. Secondarily we investigated 90-day incidence, time course, and risk factors for MI. Statistics No pre-study power analysis was conducted as we included all available procedures from the LCDB. Continuous data are presented as means (SD) or medians (IQR) and analyzed using Student’s t-test and non-parametric testing as appropriate. Categorical data are presented as actual number ((%) if n > 100, 95% confidence interval (CI)), and tested using χ2 or Fisher’s-exact test. We calculated risk factors with CI by multivariable logistic regression analyses in 4 models for MI ≤ 30 days and 4 similar models for MI ≤ 90 days. Variable selection was done based on clinically acknowledged risk factors (Mantilla et al. 2011, Belmont Jr et al. 2014) and directed acyclic graphs to reduce possible bias (Shrier and Platt 2008). Due to the risk of mediation between cardiac disease, anticoagulant treatment, and hypercholesterolemia and between anticoagulant treatment, anemia, and cardiac disease, we constructed 4 models. The 1st included age (grouped in 5-year intervals), sex, diabetes, cardiovascular disease, and anemia. The 2nd included anticoagulant treatment but excluded cardiovascular disease and anemia. The 3rd included age, sex, diabetes, and hypercholesterolemia and the 4th included age, sex, diabetes, hypercholesterolemia, and cardiovascular disease. All variables were included as direct effects. Furthermore, to check for potential bias by including patients with more than 1 procedure (Bryant et al. 2006, Ravi et al. 2013) we performed a sensitivity analysis by analyzing incidence and risk factors from multiple logistic regression model 4 among the independent observations (patients with only 1 procedure). Statistical significance was defined as p < 0.05. All analyses were carried out using SPSS v.22 (IBM Corp, Armonk, NY, USA) and http://vassarstats.net/prop1.html for calculation of CI for proportions. Ethics, registration, funding, and potential conflicts of interest No ethical approval was necessary due to the non-interventional study design. Permission to store and review patient data without prior informed consent was obtained from the Danish Data protection Agency (RH-2017-132) and the Danish National Board of Health (3-3013-56/2/EMJO). Furthermore, LCDB is registered on ClinicalTrials.gov (ID: NCT01515670) as an ongoing study registry for this type of studies.

Acta Orthopaedica 2018; 89 (6): 603–609

605

Table 1. Preoperative patient characteristics. Values are number (%) unless otherwise stated

Total number of THA and TKA in DNPR between January 2010 and October 2015 n = 28,150 Excluded (n = 3,288): – fractures, non-elective procedures or previous surgery on the same joint within 90 days, 1,615 – severe congenital disordes, infection or cancer, 134 – simultaneous or staged bilateral procedures within 90 days, 647 – other (e.g. non-Danish citizens and no standard fast track procedure, 34 – procedures not in the Lundbeck Centre Database, 858 Primary elective unilateral THA/TKA with completed preoperative questionnaire n = 24,862

Figure 1. Flow of patients. DNPR = Danish National Patient Registry.

The study was supported by a PhD grant to PBP from Lundbeckfonden (R230-2017-166), which took no part in organizing, analyzing, or writing the manuscript. Outside the submitted work HK is a member of the advisory board at “Rapid Recovery” by Zimmer Biomet and CCJ received speakers’ fees from “Rapid Recovery” by Zimmer Biomet.

Results We included 24,862 procedures in 22,291 patients (97% of performed procedures) (Figure 1). Mean age was 68 (11) years and 58% were females (Table 1). Median LOS was 2 (2–3) days and 7.3% (CI 7.0–7.6) had a LOS ≥ 5 days. 30- and 90-day readmission rates were 5.6% (CI 5.3–5.8) and 7.6% (CI 7.3–7.9), respectively. Overall mortality was 0.15% (CI 0.11–0.21) within 30 days and 0.31% (0.25–0.39) within 90 days. The 30-day incidence of MI was 31 (0.12%, CI 0.08–0.17) and time to MI was median 4 (2–8) days after surgery, with 13 (0.05%, CI 0.03–0.09) during primary admission. Of the MIs, 6 were diagnosed by rise in troponin combined with ECG abnormalities and 7 by troponin and symptoms of ischemia. In the remaining 18 patients, 5 MIs were verified by coronary angiography, 12 by percutaneous coronary intervention, and 1 by radiological evidence of new loss of viable myocardium. In addition, 17 (0.07%, CI 0.04–0.11) MIs occurred during days 30–90 resulting in a 90-day incidence of 48 (0.19%, CI 0.14–0.25) (Figure 2), similar between THA (n = 29, 0.22%, CI 0.15–0.32) and TKA (n = 19, 0.16%, CI 0.10–0.25) (p = 0.3). Median time to MI was 11 (3–56) days for MI ≤ 90 days. Prior to readmission with MI 1 patient had a readmission due to anemia 2 days prior to admission with MI (8 days postoperatively, LOS 1). No additional readmissions prior to MI occurred in the 47 remaining patients. 90-day mortality was 77 (0.31%, CI 0.25–0.39), of which proven or suspected MI was the primary cause of death in 5

Characteristic

TKA THA 11,873 (47.8) 12,989 (52.2)

Age a (years) 68 (10) 68 (11) < 50 559 (4.7) 857 (6.6) 50–60 2,135 (18) 1,926 (15) 61–65 1,790 (15) 1,808 (14) 66–70 2,374 (20) 2,669 (21) 71–75 2,279 (19) 2,378 (18) 76–80 1,641 (14) 1,874 (14) 81–85 818 (6.9) 1,025 (7.9) > 85 277 (2.3) 452 (3.5) Female sex 7,236 (61) 7,221 (56) BMI < 18.5 44 (0.4) 132 (1.0) 18.5–24.9 2,305 (19) 4,411 (34) 25.0–29.9 4,528 (38) 5,309 (41) 30.0–34.9 3,118 (26) 2,335 (18) 35.0–39.9 1,298 (11) 549 (4.2) ≥ 40 524 (4.4) 172 (1.3) Missing 56 (0.5) 81 (0.6) Smoking 1,542 (13) 1,955 (15) Missing 86 (0.7) 87 (0.7) Alcohol > 2 units daily 807 (6.8) 1,028 (7.9) Missing 101 (0.9) 108 (0.8) Social situation Living with others 7,918 (67) 8,597 (66) Living alone 3,858 (33) 4,246 (33) Nursing home. etc. 53 (0.4) 104 (0.8) Missing 44 (0.4) 42 (0.3) Preoperative anemia b 2,907 (25) 3,159 (24) Missing 121 (1.0) 234 (1.8) Use of walking aid 2,482 (21) 3,267 (25) Missing 232 (2.0) 0 (0.0) Hypertension 5,809 (49) 5,699 (44) Missing 62 (0.5) 55 (0.4) Hypercholesterolemia 4,779 (40) 4,656 (36) Missing 70 (0.6) 69 (0.5) Pharmacologically treated cardiac disease 1,691 (14) 1,740 (13) Missing 130 (1.1) 81 (0.6) pulmonary disease 1,042 (8.8) 1,109 (8.5) Missing 78 (0.7) 58 (0.4) psychiatric disorder 1,053 (8.9) 1,002 (7.7) Missing 91 (0.8) 61 (0.5) Anticoagulant treatment c 745 (6.3) 777 (6.0) Missing 0 (0.0) 0 (0.0) Prior cerebral stroke 744 (6.3) 752 (5.8) Missing 179 (1.5) 170 (1.3) Diabetes Insulin-dependent 300 (2.5) 202 (1.6) Non-insulin-dependent 1,222 (10) 899 (6.9) Missing 29 (0.2) 17 (0.1) BMI = body mass index; THA = total hip arthroplasty; TKA = total knee arthroplasty. a mean (SD); b Hb < 13 g/dL c Vitamin K antagonists or new oral anticoagulants.

(0.02%) patients. Of the 5 deaths caused by proven or suspected MI, 3 occurred during primary admission (postoperative days 2, 3, and 4) and 2 were after discharge (days 15 and 43). The 90-day mortality rate for postoperative MI patients

606

Acta Orthopaedica 2018; 89 (6): 603–609

Figure 2. Cumulated incidence of MI in 24,862 THA and TKA procedures.

was 6.3% (CI 2.2–17). 20 (0.08%, CI 0.05–0.12) patients died in their own home of unknown cause, with a median time to death of 64 (37–84) days. When analyzing risk factors for MI ≤ 30 days, age of > 85 years (OR 6.9–8.3) and insulin-dependent diabetes mellitus (IDDM) (OR 3.6–5.0) were statistically significantly associated with increased risk of MI regardless of the type of regression model. Preoperative cardiac disease (OR 2.4–2.8) and hypercholesterolemia (OR 2.3–2.7) were also significantly associated with MI in both models including these covariates (Table 2). In contrast, sex, use of anticoagulants, and preoperative anemia were not significantly associated with MI (Table 2). When assessing risk factors for MI ≤ 90 days the results

were much alike although the association with hypercholesterolemia was statistically not significant when adjusting for cardiac disease (p = 0.06), while male sex was significantly associated with MI ≤ 90 days (OR 2.3–2.5) in all 4 models (Table 3). In-hospital postoperative hemoglobin (Hb) levels were available in 28 of the 31 patients with MI ≤ 30 days. Of these, 27 were anemic with a mean postoperative Hb of 9.9 (1.7) g/ dL, and with a mean postoperative Hb reduction of 3.6 g/dL (CI 2.8–4.3). 10 MI patients with postoperative anemia also had preoperative anemia. Perioperative anesthesia notes were available in 24 cases with MI ≤ 30 days; 18 procedures were performed under spinal anesthesia. 11 patients developed intraoperative hypotension < 90 mm/Hg, 6 during spinal anesthesia and 5 during general anesthesia. 9 of the intraoperative hypotensive patients required treatment with vasopressors, 5 in the spinal anesthesia group and 4 in the general anesthesia group. No other cardiac events were recorded during the surgical procedure. Sensitivity analysis on the 19,793 patients with only 1 procedure showed similar results as mentioned above in the entire cohort. Thus, in the 19,793 patients LOS was median 2 (2–3) days, 7.6% (CI 7.2–7.9) had LOS ≥ 5 days and 7.9 % (CI 7.5–8.3) had a readmission within 90 days. 26 patients (0.13%, CI 0.9–0.19) developed MI within 30 days and 41 (0.21%, 0.16–0.29) within 90 days. Risk factors for MI ≤ 30 days were age > 85 years OR 6.8 (CI 1.9–24, p = 0.003), IDDM OR 4.5 (CI 1.3–16, p = 0.02), hypercholesterolemia OR 2.4 (CI 1.0–5.7, p = 0.05) and cardiovascular disease OR 2.3 (CI 1.0–5.2, p = 0.05). When analyzing the 5,069 patients with more than 1 procedure there was no difference. Median LOS was 2 (CI 2–3) days, 6.1% (CI 5.5–6.8%) had LOS ≥ 5

Table 2. Multivariable logistic regressions for preoperative risk factors for MI within 30 days of surgery Variable

Model 1 OR (95% CI) p-value

Model 2 OR (95% CI) p-value

Model 3 OR (95% CI) p-value

Model 4 OR (95% CI) p-value

Age 0.009 0.003 0.002 < 50 NA NA NA NA 50–60 NA NA NA NA 61–65 0.25 (0.03–2.1) 0.2 0.24 (0.03–2.0) 0.2 0.25 (0.03–2.1) 0.2 0.26 (0.03–2.2) 66–70 1 Ref 1 Ref 1 Ref 1 71–75 1.1 (0.35–3.3) 0.9 1.1 (0.36–3.4) 0.9 1.1 (0.36–3.4) 0.9 1.1 (0.35–3.3) 76–80 1.3 (0.43–4.2) 0.6 1.5 (0.47–4.5) 0.5 1.4 (0.47–4.5) 0.5 1.3 (0.43–4.2) 81–85 2.7 (0.85–8.4) 0.09 3.0 (0.94–8.3) 0.06 3.0 (0.95–9.3) 0.06 2.7 (0.85–8.5) > 85 6.9 (2.2–21.9) 0.001 7.9 (2.5–24.7) < 0.001 8.3 (2.6–26.3) < 0.001 7.4 (2.3–23.5) Male sex 1.6 (0.79–3.3) 0.2 1.8 (0.86–3.6) 0.1 1.8 (0.87–3.6) 0.1 1.6 (0.79–3.33) Diabetes 0.04 0.03 0.07 Insulin-dependent 4.5 (1.3–15.3) 0.02 5.0 (1.5–16.7) 0.009 3.7 (1.1–12.6) 0.04 3.6 (1.1–12.4) Non-insulin-dependent 0.69 (0.16–2.9) 0.6 1.3 (0.58–2.8) 0.6 0.59 (0.14–2.5) 0.5 0.56 (0.13–2.4) Cardiovascular disease 2.8 (1.3–5.8) 0.006 – – – – 2.4 (1.1–5.0) 1.2 (0.5–2.5) 0.7 Anemia a Hypercholesterolemia – – – – 2.7 (1.3–5.9) 0.01 2.3 (1.1–5.1) Anticoagulant treatment b – – 1.4 (0.50–4.2) 0.5 – – –

0.006 0.2 Ref 0.9 0.6 0.09 0.001 0.2 0.07 0.04 0.4 0.02 0.03 –

Number of patients included in analysis differs from the total number of patients due to missing data. NA: not available due to lack of data. a Hb < 13 g/dL for both men and women. b Vitamin K antagonists or new oral anticoagulants (NOAC).

Acta Orthopaedica 2018; 89 (6): 603–609

607

Table 3. Multivariable logistic regressions for preoperative risk factors for MI within 90 days of surgery Variable

Model 1 OR (95% CI) p-value

Model 2 OR (95% CI) p-value

Model 3 OR (95% CI) p-value

Model 4 OR (95% CI) p-value

Age < 0.001 < 0.001 < 0.001 < 0.001 < 50 NA NA NA NA 50–60 0.13 (0.02–1.0) 0.05 0.12 (0.02–0.90) 0.04 0.13 (0.02–1.0) 0.05 0.14 (0.02–1.1) 0.06 61–65 0.14 (0.02–1.1) 0.06 0.13 (0.02–1.0) 0.05 0.14 (0.02–1.0) 0.06 0.14 (0.02–1.1) 0.06 66–70 1 Ref 1 Ref 1 Ref 1 Ref 71–75 0.78 (0.31–2.0) 0.6 0.81 (0.32–2.0) 0.6 0.81 (0.32–2.0) 0.6 0.78 (0.31–1.9) 0.6 76–80 1.2 (0.51–2.9) 0.7 1.3 (0.56–3.14) 0.5 1.3 (0.57–3.2) 0.5 1.2 (0.52–2.9) 0.6 81–85 2.2 (0.89–5.4) 0.09 2.5 (1.0–6.0) 0.05 2.5 (1.0–6.2) 0.04 2.3 (0.9–5.5) 0.07 > 85 5.0 (1.9–12.9) 0.001 5.9 (2.4–15.1) < 0.001 6.3 (2.5–16.0) < 0.001 5.5 (2.2–12.1) < 0.001 Male sex 2.3 (1.3–4.3) 0.005 2.5 (1.4–4.5) 0.002 2.5 (1.4–4.5) 0.002 2.3 (1.3–4.1) 0.006 Diabetes 0.003 0.002 0.006 0.006 Insulin-dependent 4.4 (1.7–11.5) 0.002 5.0 (1.9–12.9) 0.001 3.9 (1.5–10.2) 0.006 3.9 (1.5–10.1) 0.006 Non-Insulin-dependent 0.42 (0.10–1.7) 0.2 0.47 (0.11–1.9) 0.3 0.38 (0.09–1.6) 0.2 0.36 (0.09–1.5) 0.2 Cardiovascular disease 3.0 (1.7–5.4) < 0.001 – – – – 2.7 (1.4–4.9) 0.001 Anemia a 1.2 (0.62–2.2) 0.6 Hypercholesterolemia – – – – 2.2 (1.2–4.0) 0.01 1.8 (0.99–3.4) 0.06 Anticoagulant treatment b – – 1.7 (0.73–3.7) 0.2 – – – – For footnotes, see Table 2.

days, 90-day readmission rate was 6.6% (CI 6.0–7.4%) and 30-day incidence of MI was 0.10% (CI 0.04–0.23%) and was similar compared with the entire cohort. When comparing the 2 subgroups, 30-day MI incidence was similar (p = 0.7). However, patients with multiple procedures had a lower proportion of LOS ≥ 5 days (p < 0.001) and readmissions within 90 days (p = 0.003).

Discussion In this prospective cohort study on 24,862 consecutive unselected primary elective fast-track THA and TKAs we found a 30-day MI incidence of 0.12% with increased risk for age > 85, IDDM, cardiovascular disease, and hypercholesterolemia in multivariable analysis. Furthermore, perioperative anemia and intraoperative hypotension was frequent among patients with postoperative MI. Our MI incidence is lower than reported in studies on administrative databases reporting 30-day incidences of 0.25%– 0.9% (Singh et al. 2011, Lalmohamed et al. 2012, Belmont Jr et al. 2014, Pedersen et al. 2014, Thornqvist et al. 2014). A recent comparison of administrative data and the clinical NSQIP registry have shown a low specificity of ≈ 81% for both sources of data compared with chart review (Etzioni et al. 2018), potentially underestimating the true incidence. Nonetheless, we found an in-hospital incidence rate of only 0.05%, considerably lower than previous data from such administrative and clinical registries with comparable cohorts but variable follow-up (Pulido et al. 2008, Malviya et al. 2011, Pedersen et al. 2014, Bemenderfer et al. 2017, Urban et al. 2017). Although, the numbers we found are comparable to previous fast-track THA and TKA studies (Malviya et al. 2011, Khan et

al. 2014), previous studies lack detailed information on potential risk factors and events preceding MI. Median time to MI ≤ 30 days was 4 (2–8) days and 11 (3–56) days for MI ≤ 90 days with increased cumulated incidence within the first 3 weeks. This is in concordance with previous data from the few THA- and TKA-specific studies, which report increased risk of MI within the first 2 weeks and up to 6 weeks for THA (Lalmohamed et al. 2012, Belmont Jr et al. 2014). When evaluating preoperative risk factors for MI ≤ 90 days, age > 85 years was found to be the most important risk factor for MI with OR of about 6. That increasing age is a risk factor for MI has been demonstrated in several studies (Belmont Jr et al. 2014, Menendez et al. 2015, Urban et al. 2017); however, we were unable to demonstrate any significant association between age and MI until age > 85 years. That IDDM, cardiovascular disease, and hypercholesterolemia were found to be associated with MI in our study is unsurprising as these are well-known risk factors (Mantilla et al. 2011, Belmont Jr et al. 2014, Menendez et al. 2015). However, that male sex was clearly associated with 90- but not 30-day MI has to our knowledge not been shown previously. A previous investigation on early thromboembolic events including some of the same patients as our study suggested postoperative anemia as a possible pathogenic mechanism (Jørgensen et al. 2016a). Preoperative anemia is well known to dispose to postoperative anemia and blood transfusions (Goodnough et al. 2011), including in fast-track THA and TKA (Jans et al. 2018). In our study, preoperative anemia was not associated with MI. However, 12 of the 48 patients with MI had preoperative and 27 had postoperative anemia. Although conflicting results on anemia as a risk factor for postoperative MI have been discussed previously (Mantilla et al. 2011, Menendez et al.

608

2015), anemia can and should be prevented both pre- and postoperatively (Goodnough et al. 2013, Belmont Jr et al. 2014, Pujol-Nicolas et al. 2017), and increased focus on blood management remains a potential “low hanging fruit” for further improvement. We also identified 11 cases of intraoperative hypotension of which 9 cases were treated with vasopressors. This occurred in 5 of 6 MI patients having general anesthesia in contrast to 6 of 18 patients having spinal anesthesia. Several studies from unspecific non-cardiac surgery have shown an association between intraoperative hypotension and risk of MI and mortality (Walsh et al. 2013, Roshanov et al. 2017, Sessler et al. 2017). Thus, an increased focus on avoiding a supply–demand mismatch during surgery may possibly further decrease the risk of MI. This hypothesis is also backed by a positive effect of an individualized blood pressure management strategy lowering postoperative organ dysfunction (Futier et al. 2017), although not specifically addressing MI. In addition, a recent study on etiology of MI after non-cardiac surgery found about 75% of MIs to be “demand” ischemia (type 2 myocardial infarctions) (Garcia et al. 2014). However, we do not know the incidence of intraoperative hypotension in patients who did not develop MI and, thus, further research on intraoperative hypotension and postoperative MI in THA and TKA is needed. Our study has some methodological limitations. We only obtained health records from patients with LOS ≥ 5 days, readmission, and/or death within 90 days. Thus, a case of MI during primary admission with total LOS < 5 days and without a diagnostic code of MI would be missed. However, that a patient with MI would be discharged within 4 days of surgery without a diagnostic code and no postoperative prescriptions of anticoagulants is unlikely. Although the LCDB aims at including all patients, 3% of the eligible population from the DNPR was missing in the LCDB (see Figure 1). However, we have previously shown similar postoperative outcomes in these patients (Jørgensen et al. 2013). Data on readmissions were obtained from the DNPR, which has shown varying reliability with regards to specific diagnostic codes (Vest-Hansen et al. 2013, Egholm et al. 2016) but > 99% overall reliability with regards to somatic admission (Schmidt et al. 2015). Consequently, we obtained discharge summaries on all patients with LOS ≥ 5 and medical records including results from blood samples and intraoperative anesthesia records in case of MI to ensure best available information on incidence and pathophysiological mechanisms (Etzioni et al. 2018). The declining autopsy rate in Denmark limits the availability of specific cause of death as seen in our study by 20 deaths of unknown causes. Consequently, although we obtained all healthcare records of patients dying ≤ 90 days of surgery to establish potential relation to surgery, we may underestimate the true MI incidence. Nonetheless, most deaths of unknown reason occurred > 30 days postoperatively, and even if all 20 deaths were due MI, the 90-day MI rate would remain < 0.3%. The lack of information on postoperative anemia and intraop-

Acta Orthopaedica 2018; 89 (6): 603–609

erative hypotension in patients not developing MI precluded relevant statistical analyses. Likewise, Hb values at readmission were not consistently available, thus precluding conclusions on possible effects of post-discharge anemia and MI. Furthermore, our study design also impedes comparison on risk of MI with a non-surgical group as no matched control group was made. Finally, about 10% of the patients had a second procedure during the study period, which may limit external validity of our results (Bryant et al. 2006). However, the incidence of and risk factors for MI among the patients with only 1 procedure or those with multiple primary arthroplasties was not different from the entire cohort. Additionally, leaving out patients with multiple primary arthroplasties may also introduce additional bias to postoperative outcomes (Ravi et al. 2013). Nonetheless, the inclusion of consecutive “real-life” patients from 8 different centers throughout Denmark without any selection bias increases the generalizability within fast-track THA and TKA. In summary, the incidence of MI ≤ 30 days after fast-track THA and TKA was only 0.12%, increasing to 0.19% at 90 day. Age > 85 years, IDDM, hypercholesterolemia, and cardiovascular disease were associated with MI ≤ 30 days. The roles of perioperative anemia and intraoperative hypotension as potential modifiable risk factors of developing MI remain areas for further research.

PBP: Designing the work; acquisition, analysis and interpretation of data; drafting and revising the manuscript. HK: Designing the work, interpretation of data, revising the manuscript. CCJ: Designing the work, acquisition and interpretation of data, revising the manuscript. Acta thanks Keijo T Mäkelä and other anonymous reviewers for help with peer review of this study.

Belmont Jr P J, Goodman G P, Kusnezov N A, Magee C, Bader J O, Waterman B R, Schoenfeld A J. Postoperative myocardial infarction and cardiac arrest following primary total knee and hip arthroplasty: rates, risk factors, and time of occurrence. J Bone Joint Surg 2014; 96: 2025-31. Bemenderfer T B, Rozario N L, Moore C G, Karunakar M A. Morbidity and mortality in elective total hip arthroplasty following surgical care improvement project guidelines. J Arthroplasty 2017; 32(8): 2359-62. Bryant D, Havey T C, Roberts R, Guyatt G. How many patients? How many limbs? Analysis of patients or limbs in the orthopaedic literature: a systematic review. J Bone Joint Surg 2006; 88(1): 41-5. Egholm G, Bøtker H E, Maeng M. Evaluation of algorithms for registry-based detection of acute myocardial infarction following percutaneous coronary intervention. Clin Epidemiol 2016; 415-23. Etzioni D A, Lessow C, Bordeianou L G, Kunitake H, Deery S E, Carchman E, Papageorge C M, Fuhrman G, Seiler R L, Ogilvie J, Habermann E B, Chang Y H H, Money S R. Postoperative myocardial infarction in administrative data vs clinical registry: a multi-institutional study. J Am Coll Surg 2018; 226(1): 14-21. Futier E, Lefrant J-Y, Guinot P-G, Godet T, Lorne E, Cuvillon P, Bertran S, Leone M, Pastene B, Piriou V, Molliex S, Albanese J, Julia J-M, Tavernier B, Imhoff E, Bazin J-E, Constantin J-M, Pereira B, Jaber S. Effect of individualized vs standard blood pressure management strategies on postoperative organ dysfunction among high-risk patients undergoing major surgery. JAMA 2017; 318(14): 1346-57.

Acta Orthopaedica 2018; 89 (6): 603–609

Garcia D, Alexander JH, Wallentin L, Wojdyla DM, Thomas L, Hanna M, Al-Khatib S M, Dorian P, Ansell J, Commerford P, Flaker G, Lanas F, Vinereanu D, Xavier D, Hylek E M, Held C, Verheugt F W, Granger C B, Lopes R D. Management and clinical outcomes in patients treated with apixaban versus warfarin undergoing procedures. Blood 2014; 124(25): 3692-9. Goodnough L T, Maniatis A, Earnshaw P, Benoni G, Beris P, Bisbe E, Fergusson D A, Gombotz H, Habler O, Monk T G, Ozier Y, Slappendel R, Szpalski M. Detection, evaluation, and management of preoperative anaemia in the elective orthopaedic surgical patient: NATA guidelines. Br J Anaesth 2011; 106(1): 13-22. Goodnough L T, Levy J H, Murphy M F. Concepts of blood transfusion in adults. Lancet 2013; 381(9880): 1845-54. Jammer I, Wickboldt N, Sander M, Smith A, Schultz M J, Pelosi P, Leva B, Rhodes A, Hoeft A, Walder B, Chew M S, Pearse R M. Standards for definitions and use of outcome measures for clinical effectiveness research in perioperative medicine: European Perioperative Clinical Outcome (EPCO) definitions: a statement from the ESA-ESICM joint taskforce on perioperative outcome measur. Eur J Anaesthesiol 2015; 32(2): 88-105. Jans Ø, Nielsen C S, Khan N, Gromov K, Troelsen A, Husted H. Iron deficiency and preoperative anaemia in patients scheduled for elective hip- and knee arthroplasty: an observational study. Vox Sang 2018; 113(3): 260-7. Johannesdottir S A, Horváth-Puhó E, Ehrenstein V, Schmidt M, Pedersen L, Sørensen H T. Existing data sources for clinical epidemiology: the Danish National Database of Reimbursed Prescriptions. Clin Epidemiol 2012; 4: 303-13. Jørgensen C C, Kehlet H. Role of patient characteristics for fast-track hip and knee arthroplasty. Br J Anaesth 2013; 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 2013; 3(12): e003965. Jørgensen C C, Kehlet H, Lundbeck Foundation Centre for Fast-track Hip and Knee replacement collaborative group. Early thromboembolic events ≤ 1week after fast-track total hip and knee arthroplasty. Thromb Res 2016a; 138: 37-42. Jørgensen C C, Petersen M A, Kehlet H. Preoperative prediction of potentially preventable morbidity after fast-track hip and knee arthroplasty: a detailed descriptive cohort study. BMJ Open. 2016b ;6(1): e009813. Jørgensen C C, Kehlet H, Soeballe K, Hansen T B, Husted H, Laursen M B, Hansen L T, Kjærsgaard-Andersen P, Solgaard S, Krarup N H. Time course and reasons for 90-day mortality in fast-track hip and knee arthroplasty. Acta Anaesthesiol Scand 2017; 61(4): 436-44. Kehlet H. Fast-track hip and knee arthroplasty. Lancet 2013; 381(9878): 1600-2. 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. Kirksey M, Lin Chiu Y, Ma Y, Gonzalez Della Valle A, Poultsides L, Gerner P, Memtsoudis S G. Trends in in-hospital major morbidity and mortality after total joint arthroplasty: United States 1998-2008. Anesth Analg 2012; 115(2): 321-7. Lalmohamed A, Vestergaard P, Klop C, Grove E L, de Boer A, Leufkens H G M, van Staa T P, de Vries F. Timing of acute myocardial infarction in patients undergoing total hip or knee replacement. Arch Intern Med 2012; 172(16): 1229-35. 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. Mantilla C B, Wass C T, Goodrich K A, Johanns C J, Kool M L, Zhu X, Corredor J A, Warner D O, Joyner M J, Berry D J, Schroeder D R, Sprung J. Risk for perioperative myocardial infarction and mortality in patients undergoing hip or knee arthroplasty: the role of anemia. Transfusion 2011; 51(1): 82-91.

609

Menendez M E, Memtsoudis S G, Opperer M, Boettner F, Gonzalez Della Valle A. A nationwide analysis of risk factors for in-hospital myocardial infarction after total joint arthroplasty. Int Orthop 2015; 39(4): 777-86. Muñoz M, Acheson A G, Auerbach M, Besser M, Habler O, Kehlet H, Liumbruno G M, Lasocki S, Meybohm P, Rao Baikady R, Richards T, Shander A, So-Osman C, Spahn D R, Klein A A. International consensus statement on the peri-operative management of anaemia and iron deficiency. Anaesthesia 2017; 72(2): 233-47. Pedersen A B, Mehnert F, Sorensen H T, Emmeluth C, Overgaard S, Johnsen S P. The risk of venous thromboembolism, myocardial infarction, stroke, major bleeding and death in patients undergoing total hip and knee replacement: a 15-year retrospective cohort study of routine clinical practice. Bone Joint J 2014; 96 B(4): 479-85. Pujol-Nicolas A, Morrison R, Casson C, Khan S, Marriott A, Tiplady C, Kotze A, Gray W, Reed M. Preoperative screening and intervention for mild anemia with low iron stores in elective hip and knee arthroplasty. Transfusion 2017; 57(12): 3049-57. Pulido L, Parvizi J, Macgibeny M, Sharkey P F, Purtill J J, Rothman R H, Hozack W J. In hospital complications after total joint arthroplasty. J Arthroplasty 2008; 23(6): 139-45. Ravi B, Croxford R, Hawker G. Exclusion of patients with sequential primary total joint arthroplasties from arthroplasty outcome studies biases outcome estimates: a retrospective cohort study. Osteoarthritis Cartilage 2013; 21(12): 1841-8. Roshanov P S, Rochwerg B, Patel A, Salehian O, Duceppe E, Belley-Côté E P, Guyatt G H, Sessler D I, Le Manach Y, Borges F K, Tandon V, Worster A, Thompson A, Koshy M, Devereaux B, Spencer F A, Sanders R D, Sloan E N, Morley E E, Paul J, Raymer K E, Punthakee Z, Devereaux P J. Withholding versus continuing angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers before noncardiac surgery. Anesthesiology 2017; 126(1): 16-27. Schmidt M, Schmidt S A J, Sandegaard J L, Ehrenstein V, Pedersen L, Sørensen H T. The Danish National patient registry: a review of content, data quality, and research potential. Clin Epidemiol 2015; 7: 449-90. Sessler D I, Meyhoff C S, Zimmerman N M, Mao G, Leslie K, Vásquez S M, Balaji P, Alvarez-Garcia J, Cavalcanti A B, Parlow J L, Rahate P V, Seeberger M D, Gossetti B, Walker S A, Premchand R K, Dahl R M, Duceppe E, Rodseth R, Botto F, Devereaux P J. Period-dependent associations between hypotension during and for 4 days after noncardiac surgery and a composite of myocardial infarction and death. Anesthesiology 2017; 128(2): 317-27. Shrier I, Platt R W. Reducing bias through directed acyclic graphs. BMC Med Res Methodol 2008; 8(1): 70. Singh J A, Jensen M R, Harmsen W S, Gabriel S E, Lewallen D G. Cardiac and thromboembolic complications and mortality in patients undergoing total hip and total knee arthroplasty. Ann Rheum Dis 2011; 70(12): 2082-8. Thornqvist C, Gislason G H, Køber L, Jensen P F, Torp-Pedersen C, Andersson C. Body mass index and risk of perioperative cardiovascular adverse events and mortality in 34,744 Danish patients undergoing hip or knee replacement. Acta Orthop 2014; 85(5): 456-62. Urban M K, Wolfe S W, Sanghavi N M, Fields K, Magid S K. The incidence of perioperative cardiac events after orthopedic surgery: a single institutional experience of cases performed over one year. HSS J 2017; 13(3): 248-54. Vest-Hansen B, Riis A H, Christiansen C F. Registration of acute medical hospital admissions in the Danish National Patient Registry: a validation study. Clin Epidemiol 2013; 5: 129-33. Walsh M, Devereaux P J, Garg AX, Kurz A, Turan A, Rodseth RN, Cywinski J, Thabane L, Sessler D I. Relationship between intraoperative mean arterial pressure and clinical outcomes after noncardiac surgery. Anesthesiology 2013; 119(3): 507-15.

610

Acta Orthopaedica 2018; 89 (6): 610–614

Time of admission and mortality after hip fracture: a detailed look at the weekend effect in a nationwide study of 55,211 hip fracture patients in Norway Andreas ASHEIM 1,2,a, Sara Marie NILSEN 1,a, Marlen TOCH-MARQUARDT 3,a, Kjartan Sarheim ANTHUN 3,4, Lars Gunnar JOHNSEN 5,6, Johan Håkon BJØRNGAARD 3 1 Center for Health Care Improvement, St Olav’s Hospital, Trondheim; 2 Norwegian University of Science and Technology, Department of Mathematical Sciences, Trondheim; 3 Norwegian University of Science and Technology, Department of Public Health and Nursing, Trondheim; 4 Department of Health Research, SINTEF Digital, Trondheim; 5 Department of Orthopaedic Surgery, St Olav’s Hospital, Trondheim; 6 Norwegian University of Science and Technology, Department of Neuromedicine, Trondheim, Norway a Joint first author Correspondence: johan.h.bjorngaard@ntnu.no Submitted 2018-06-14. Accepted 2018-09-06.

Background and purpose — There are numerous studies on the weekend effect for hip fracture patients, with conflicting results. We analyzed time of admission and discharge, and the association with mortality and length of hospital stay in more detail. Patients and methods — We used data from 61,211 surgically treated hip fractures in 55,211 patients, admitted to Norwegian hospitals 2008–2014. All patients were aged 50 years or older. Data were analyzed with Cox and Poisson regression. Results — Mortality within 30 days did not differ substantially by day of admission, although admissions on Sundays and holidays had a slightly increased mortality. The hazard ratios were 1.1 (95% confidence interval [CI] 0.97– 1.2) for Sundays, and 1.2 (CI 0.98–1.4) for holidays, relative to Mondays. For patients admitted between 6:00 am and 7:00 am the hazard ratio was 1.4 (CI 1.1–1.8) relative to patients admitted between 2:00 pm and 3:00 pm. Discharges during weekends and holidays were associated with a substantial higher mortality than weekday discharges. Patients admitted from Friday to Sunday generally stayed in hospital for a shorter time than patients admitted during other days. Interpretation — Our results indicate that the discussion on weekday versus weekend admission effects might have distracted attention from other important factors, such as time of day of admission, and day of discharge from hospital treatment.

Despite improvements in health care organization and medical technology, 1 out of 10 hip fracture patients die within 30 days after admission (Giannoulis et al. 2016, Søgaard et al. 2016, Johansen et al. 2017). With hip fractures being acute, an event occurring out-ofhours will pose a challenge to the health care system. While increased mortality for patients admitted at weekends relative to weekdays has been reported by some studies (Thomas et al. 2014, Kristiansen et al. 2016), others have not found this (Foss and Kehlet 2006, Daugaard et al. 2012, Boylan et al. 2015, Nandra et al. 2017, Nijland et al. 2017, Sayers et al. 2017). These studies had limited information on the volume of admissions by day of the week and time of day, and most included few cases. Most of these studies are limited to aggregation of information on timing into binary indicators of weekdays versus weekends, and out-of-hours versus working hours (Sayers et al. 2017). However, these are just two of several patterns of time variation, which as well are likely to vary by clinical pathways and health care settings (Bray et al. 2016). A recent Danish study (Sayers et al. 2017) on hip fractures, for example, reported a 17% higher mortality for out-of-hours discharges and 50% higher mortality for discharges on Sundays. By using a large, nationwide cohort, we were able to describe the timing of admission and discharge, and the subsequent association with mortality and length of hospital stay.

Patients and methods We retrieved a nationwide cohort of 61,211 hip fractures from 55,211 unique patients, discharged from January 1, 2008 to December 31, 2014, from the Norwegian patient registry. All Norwegian hospital trusts are obliged to submit information © 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 10.1080/17453674.2018.1533769

Acta Orthopaedica 2018; 89 (6): 610–614

611

about their activity to this registry. Each patient has a unique, anonymous identification number throughout the observation period (Hassani et al. 2015). Information on the date of death from the National Registry in Norway was provided through the Norwegian patient registry. Population-level data (e.g., population size) were obtained from Statistics Norway. Two Norwegian validation studies have shown high accuracy for hip fracture incidents (Høiberg et al. 2014, Helgeland et al. 2016), the data were also assessed as adequate for calculating 30-day mortality after hip fracture. We identified hip fracture patients through a combination of ICD-10 codes and the Nordic Medico-Statistical Committee (NOMESCO) Classification of Surgical Procedures codes (Bay-Nielsen 2010). Admitted patients with ICD-10 codes S72.0x, S72.1x, or S72.2x as primary diagnosis and 1 or more NOMESCO codes, NFBxy (x = 0–9, y = 0–2) or NFJxy (x = 0–9, y = 0–2) during their hospitalization, were included in the study. This definition has previously shown high accuracy, excluding for example hospitalizations for rehabilitation (Høiberg et al. 2014). Patients whose hospitalization commenced with a planned admission were excluded. We also excluded re-hospitalizations within 30 days from the previous hospitalization, to avoid influence from several hospitalizations for the same hip fracture. We included patients who were registered with 1 of the relevant ICD-10 codes, but none of the procedure codes if they died during the hospitalization, to include patients who died before surgery. Due to the Norwegian coding practice, we cannot capture the non-operatively treated fractures. We selected patients aged 50 years and older. See also Figure 1 (Supplementary data) for a flow chart showing the selection criteria. Ward and hospital transfers were considered part of the hospitalization if subsequent admissions were closer in time than 8 hours. We considered the length of hospitalization as the continuous time between admission and discharge, and not measured as discrete number of days. The construction of hospitalizations (episodes of care) was according to Hassani et al. (2015).

lowed for 30 days or until time of death, whichever occurred first. We used time of admission as explanatory variable, with admission hour, day-of-the-week, month, and year. The analysis was adjusted for sex, age, and day-after-holiday. We used Mondays and hour 2:00 to 3:00 pm as reference categories, because these are categories with the largest volumes. For the analyses of discharge day and mortality, we used the time from admission as time scale, while the risk of mortality was measured from the day of discharge to death or end of follow-up.

Time variables The analyses were performed using calendar variables as explanatory variables: hour, day-of-the-week/holiday, and year. Hour of admission was categorized by the number of whole hours after midnight. For hour of discharge, we used a coarser categorization, with 4 hours per category. Day-ofthe-week was extended with a category named “holiday,” such that public holidays are coded as a category separate from Mondays to Sundays. The models were adjusted for age in years, without any categorization, and sex.

70% of all hip fracture incidents were among women, and the average age in our cohort was 82 years (SD 10). The 30-day case mortality after hospital admission was 8.0 %. The average and median length of hospitalization were 8.6 days and 6.4 days, respectively. Table 2 (Supplementary data) shows characteristics of the portion of the cohort that died within 30 days of admission. We see a 6-year higher average age among these patients, and an overrepresentation of men. For patients who were discharged alive and died within 30 days, the length of hospitalization was around 2 days shorter compared with those who survived. In our cohort, the hip fracture incidence rate was 5.2 hip fractures per 1,000 persons per year. Table 3 (Supplementary data) includes yearly breakdowns of the incidence rate, suggesting a slight decrease from 2008 to 2014, while the 30-day case fatality remained stable. The arrival

Statistics Mortality We used Cox regression to analyze mortality. All models used time from admission as the time scale. Patients were fol-

Length of hospitalization The association between time of admission and length of hospitalization was analyzed using Poisson regression with robust standard errors for patient identification, to account for dependence in the data. Admission hour, day-of-the-week, month, and year were used as explanatory variables, and the model was adjusted for sex, age, and day-after-holiday. The association between day of admission and length of hospitalization was estimated using the average marginal effect. The analyses were performed using Stata version 14 (StataCorp, College Station, TX, USA) and RStudio version 10.5 (https://www.rstudio.com). Rates of patients in/out were computed as densities with respect to time-of-day and day-of-theweek. Detailed results from the different regression models are presented in Table 1 (Supplementary data). Ethics, funding, and potential conflicts of interest The study was approved by the Regional Committee of Ethics in Medical Research (2016/2158-1). Participant consent was not required. Authors SMN and MTM were funded by the Norwegian research council (Grant number 256579). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors declare that they have no competing interests.

Results

612

Acta Orthopaedica 2018; 89 (6): 610–614

Figure 2. Day and time of admission (left panel) and discharge (right panel).

patterns were similar for all weekdays, with 41% of patients arriving between 12:00 pm and 6:00 pm (Figure 2). Saturdays and Sundays had markedly fewer patients arriving during the day, but somewhat more admissions during night-hours. Sunday had the lowest number of hip fracture admissions. For breakdowns of case fatality, sex, and age and primary diagnosis by day and hour of admission, Tables 4 and 6 (Supplementary data). Patient characteristics, sex, age, and type of fracture, vary to some extent with hour of admission but are apparently stable between days of admission. There was a tendency to higher hip fracture occurrence during the winter months December to March (Table 7, Supplementary data). The average age of patients admitted during the winter months was about 1 year lower than for the other months.

Admissions

Admission hour

Discharges

Monday (n = 8,614)

(n = 11,654)

Tuesday (n = 8,750)

(n = 10,935)

Wednesday (n = 8,681)

(n = 10,334)

Thursday (n = 8,367)

(n = 10,182)

Friday (n = 8,797)

(n = 11,530)

Saturday (n = 8,108)

(n = 2,824)

Sunday (n = 7,570)

(n = 1,994)

Holiday (n = 2,324)

The rates of patients discharged every weekday are depicted in Figure 2. All days of the week have similar discharge patterns, where the patients were mainly discharged from 12:00 am until 6:00 pm. 74% of discharges occurred in this time span. Saturdays and Sundays had markedly fewer discharges than the other days. Holidays were almost identical to Sundays, while Mondays and Fridays had the largest volume of discharges. For information on case fatality, sex, and age by day of discharge see Table 5 (Supplementary data). Figures 3 and 4 show the average relative difference in hazard ratios of mortality within 30 days of admission, by day and hour of admission. There were only slight differences in mortality between admissions on weekdays. Compared with admissions on Mondays, admissions on Sundays and holidays had a slightly increased mortality. The hazard ratios were 1.1 (95% confidence interval [CI] 0.97–1.2) for Sundays, and 1.2 (CI 0.98–1.4) for holidays, compared with Monday admissions. Mortality after day of discharge was lowest for Monday discharges. There was about a 10% increased mortality for all weekday discharges as compared with Mondays. Compared with discharges on Mondays, discharges on Saturdays, Sundays, and holidays had a substantially increased mortality. The hazard ratio was 1.5 (CI 1.3–1.8) for Saturday discharges, 1.6 (CI 1.3–2.0) for Sunday discharges, and 1.4 (CI 1.1–1.8) for holiday discharges, compared with Monday discharges. Time of day of admissions was largely unrelated to mortality, except for a very marked spike in the early morning. Admissions between 6:00 am and 8:00 am had a 40% increased risk,

(n = 1,758)

-20 -10 0 10 20 30 40 50 60 70 80 90 100

Relative change in hazard ratio Figure 3. Mortality within 30 days after admission, by day of admission and discharge. Relative change in hazard ratios with 95% confidence interval is displayed. Number of admissions and discharges per category are displayed on the left and right, respectively. Adjusted for sex, age, admission hour, day-after-holiday, month, and year.

0 (n = 1,837) 1 (n = 1,361) 2 (n = 1,063) 3 (n = 947) 4 (n = 770) 5 (n = 705) 6 (n = 705) 7 (n = 736) 8 (n = 1,038) 9 (n = 2,069) 10 (n = 3,224) 11 (n = 3,941) 12 (n = 4,343) 13 (n = 4,483) 14 (n = 4,574) 15 (n = 4,342) 16 (n = 3,972) 17 (n = 3,668) 18 (n = 3,550) 19 (n = 3,100) 20 (n = 3,078) 21 (n = 2,947) 22 (n = 2,474) 23 (n = 2,284) -30 -20 -10 0

10 20 30 40 50 60 70 80

Relative change in hazard ratio Figure 4. Mortality within 30 days after admission, by hour of admission. Relative change in hazard ratios with 95% confidence interval is displayed. Number of admissions per hour is displayed on the left. Adjusted for sex, age, day-of-the-week, day-after-holiday, month, and year.

Acta Orthopaedica 2018; 89 (6): 610â&#x20AC;&#x201C;614

613

Monday (n = 8,750) Tuesday (n = 8,681) Wednesday (n = 8,367) Thursday (n = 8,797) Friday (n = 8,108) Saturday (n = 7,570) Sunday (n = 8,614) Holiday (n = 2,324) 0

Length of hospitalization (days) Figure 5. Estimated length of hospitalization according to day of admission, with 95% margins. Number of admissions per category in the data are displayed on the left. Adjusted for sex, age, admission hour, day-after-holiday, month, and year.

and admissions between 8:00 am and 11:00 am show a somewhat elevated risk. Figure 5 shows the estimated length of hospitalization by day of admission. Patients admitted on weekends had shorter stays, varying between 7.7 and 8.1 days. Patients admitted on weekdays except Fridays stayed longer (between 8.1 and 8.6 days), and those admitted on Fridays had a slightly shorter stay.

Discussion Our results suggest that there might be other timing effects in addition to the well-studied weekend effect. We find slightly increased mortality for patients admitted on Sundays and holidays, but not Saturdays, whereas early morning admissions and weekend discharges are associated with much higher mortality. The finding of fewer hip fracture admissions at weekends is in line with a study from the UK (Johansen and Boulton 2017), which reports this as particularly true for in-hospital fractures. Because it is an acute condition, admissions for hip fractures are not likely to be postponed until weekdays if the fracture occurs during weekends. This is reflected in the fact that the number of admissions was not higher on Mondays compared with other weekdays. There was increased mortality for patients admitted between 6:00 and 8:00 am. With this finding we have, to some extent, pinpointed a time of day when frail and severely ill patients are especially vulnerable. We did not have data on time of fracture or surgery, which would be a natural starting point for investigations into the causes of this finding. Our finding of increased mortality for early morning arrivals has not been reported before, since studies usually consider a dichotomous

outcome of out-of-hours versus working-hours (Bray et al. 2016). The finding of markedly higher mortality for patients discharged on weekends or holidays is in line with another recent study (Sayers et al. 2017), reporting 50% higher mortality for patients discharged on Sundays compared with other days of the week. While fewer patients are discharged during weekends and holidays, weekend discharges still account for about 20% of the discharge volume from an ordinary day in our material. The association could be confounded by discharges of frail patients to primary health services. However, we did not have any data on discharge destination to investigate this. Alternative explanations for the elevated mortality could be sought in reduced hospital and community staffing, which could impact the timeliness of discharge, accuracy of discharge instructions, and medication reconciliation (Cloyd et al. 2015). Discharging patients from hospital is a complex task, which requires clear communication among physicians, nurses, case managers, social workers, and the family (Hesselink et al. 2012). The shorter length of stay for patients admitted on weekends, and to a certain extent Fridays, may be due to discharge routines, e.g., discharging patients before the weekend. An average stay of around 8 days implies that patients admitted on Fridays and weekends would, on average, be under consideration for discharge the following Friday. We have not investigated changes over time in detail. However, it has been pointed out that despite a decline in the hip fracture rate over the past 15 years, the absolute number of fractures may increase as the number of older individuals grows (SĂ¸gaard et al. 2016). A recent Danish study (Pedersen et al. 2017) on incidence rates of hip fracture and mortality found that the decline in incidence was not seen in all patient groups. Our findings suggest the importance of well-organized hospital treatment of hip fracture patients at any time of week and day. The results indicate the need for a wider scope on timing effects. The narrow focus on weekday versus weekend effects might have distracted attention from important factors related to time of day, but also discharge from hospital treatment. Supplementary data Figure 1, and Tables 1â&#x20AC;&#x201C;7 with descriptive statistics and results from regression analyses are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/ 17453674.2018.1533769 Availability of data and material The data of this study are available from the Norwegian patient registry, but restrictions apply to availability. These data were used under license for the current study, and are not publicly available.

614

AA, JHB, SMN, and MTM conceived the study. AA performed the data analysis and visualizations. AA, SMN, and MTM wrote the first draft of the manuscript. All authors interpreted the data and edited the manuscript. AA, SMN, and MTM are joint first authors. Acta thanks Klaus Hindsø and Alma B Pedersen for help with peer review of this study.

Bay-Nielsen H. NOMESCO Classification of Surgical Procedures: Nordic Medico-Statistical Committee (NOMESCO); 2010. [Available from: https://norden.diva-portal.org/smash/get/diva2:970547/FULLTEXT01. pdf]. Boylan M R, Rosenbaum J, Adler A, Naziri Q, Paulino C B. Hip fracture and the weekend effect: does weekend admission affect patient outcomes? Am J Orthop 2015; 44(10): 458-64. Bray B D, Cloud G C, James M A, Hemingway H, Paley L, Stewart K, Tyrrell P J, Wolfe C D A, Rudd A G. Weekly variation in health-care quality by day and time of admission: a nationwide, registry-based, prospective cohort study of acute stroke care. Lancet 2016; 388(10040): 170-7. Cloyd J M, Chen J, Ma Y, Rhoads K F. Association between weekend discharge and hospital readmission rates following major surgery. JAMA Surg 2015; 150(9): 849-56. Daugaard C L, Jorgensen H L, Riis T, Lauritzen J B, Duus B R, van der Mark S. Is mortality after hip fracture associated with surgical delay or admission during weekends and public holidays? A retrospective study of 38,020 patients. Acta Orthop 2012; 83(6): 609-13. Foss N B, Kehlet H. Short-term mortality in hip fracture patients admitted during weekends and holidays. Br J Anaesth 2006; 96(4): 450-4. Giannoulis D, Calori G M, Giannoudis P V. Thirty-day mortality after hip fractures: has anything changed? Eur J Orthop Surg Traumatol 2016; 26(4): 365-70. Hassani S, Lindman A S, Kristoffersen D T, Tomic O, Helgeland J. 30-day survival probabilities as a quality indicator for Norwegian Hospitals: Data management and snalysis. PLoS One 2015; 10(9): e0136547. Helgeland J, Kristoffersen D T, Skyrud K D, Lindman A S. Variation between hospitals with regard to diagnostic practice, coding accuracy, and casemix: a retrospective validation study of administrative data versus medical records for estimating 30-day mortality after hip fracture. PLoS One 2016; 11(5): e0156075.

Acta Orthopaedica 2018; 89 (6): 610–614

Hesselink G, Flink M, Olsson M, Barach P, Dudzik-Urbaniak E, Orrego C, Toccafondi G, Kalkman C, Johnson J K, Schoonhoven L, Vernooij-Dassen M, Wollersheim H. Are patients discharged with care? A qualitative study of perceptions and experiences of patients, family members and care providers. BMJ Qual Saf 2012; 21(Suppl 1): i39-i49. Høiberg M P, Gram J, Hermann P, Brixen K, Haugeberg G. The incidence of hip fractures in Norway: accuracy of the national Norwegian patient registry. BMC Musculoskelet Disord 2014; 15: 372. Johansen A, Boulton C. Hip fracture presentations are less frequent at weekends. Injury 2017; 48(12): 2885-6. Johansen A, Golding D, Brent L, Close J, Gjertsen J E, Holt G, Hommel A, Pedersen A B, Rock N D, Thorngren K G. Using national hip fracture registries and audit databases to develop an international perspective. Injury 2017; 48(10): 2174-9. Kristiansen N S, Kristensen P K, Nørgård B M, Mainz J, Johnsen S P. Offhours admission and quality of hip fracture care: a nationwide cohort study of performance measures and 30-day mortality. Int J Qual Health Care 2016; 28(3): 324-31. Nandra R, Pullan J, Bishop J, Baloch K, Grover L, Porter K. Comparing mortality risk of patients with acute hip fractures admitted to a major trauma centre on a weekday or weekend. Sci Rep 2017; 7(1): 1233. Nijland L M G, Karres J, Simons A E, Ultee J M, Kerkhoffs G, Vrouenraets B C. The weekend effect for hip fracture surgery. Injury 2017; 48(7): 153641. Pedersen A B, Ehrenstein V, Szepligeti SK, Lunde A, Lagerros Y T, Westerlund A, Tell G S, Sørensen H T. Thirty-five-year trends in first-time hospitalization for hip fracture, 1-year mortality, and the prognostic impact of comorbidity: a Danish nationwide cohort study, 1980–2014. Epidemiology 2017; 28(6): 898-905. Sayers A, Whitehouse M R, Berstock J R, Harding K A, Kelly M B, Chesser T J. The association between the day of the week of milestones in the care pathway of patients with hip fracture and 30-day mortality: findings from a prospective national registry—the National Hip Fracture Database of England and Wales. BMC Med 2017; 15(1): 62. Søgaard A J, Holvik K, Meyer H E, Tell G S, Gjesdal C G, Emaus N, Grimnes G, Schei B, Forsmo S, Omsland T K. Continued decline in hip fracture incidence in Norway: a NOREPOS study. Osteoporos Int 2016; 27(7): 2217-22. Thomas C J, Smith R P, Uzoigwe C E, Braybrooke J R. The weekend effect: short-term mortality following admission with a hip fracture. Bone Joint J 2014; 96-B(3): 373-8.

Acta Orthopaedica 2018; 89 (6): 615–621

615

Low-molecular-weight heparin for hip fracture patients treated with osteosynthesis: should thromboprophylaxis start before or after surgery? An observational study of 45,913 hip fractures reported to the Norwegian Hip Fracture Register Sunniva LEER-SALVESEN 1, Eva DYBVIK 2, Lars B ENGESÆTER 1,2, Ola E DAHL 3,4, and Jan-Erik GJERTSEN 1,2 1 Department of Clinical Medicine, University of Bergen, Bergen, Norway; 2 Norwegian Arthroplasty Register, Department of Orthopaedic Surgery, Haukeland University Hospital, Bergen, Norway; 3 Innlandet Hospital Trust, Elverum, Norway; 4 Thrombosis Research Institute, London, UK Correspondence: sunniva.leer-salvesen@helse-bergen.no Submitted 2018-04-09. Accepted 2018-07-12.

Background and purpose — Controversies exist regarding thromboprophylaxis in orthopedic surgery. We studied whether the thromboprophylaxis in hip fracture patients treated with osteosynthesis should start preoperatively or postoperatively. Data were extracted from the nationwide Norwegian Hip Fracture Register (NHFR). The risks of postoperative deaths, reoperations, and intraoperative bleeding were studied within 6 months after surgery. Patients and methods — After each operation for hip fracture in Norway the surgeon reports information on the patient, the fracture, and the operation to the NHFR. Cox regression analyses were performed with adjustments for age group, ASA score, sex, duration of surgery, and year of surgery. During the period 2005–2016, 96,599 hip fractures were reported to the register. Only osteosyntheses where low-molecular-weight heparin (LMWH) were given and with known information on preoperative start of the prophylaxis were included in the analyses. Dalteparin and enoxaparin were used in 58% and 42% of the operations respectively (n = 45,913). Results — Mortality (RR = 1.01, 95% CI 0.97–1.06) and risk of reoperation (RR = 0.99, CI 0.90–1.08) were similar comparing preoperative and postoperative start of LMWH. Postoperative start reduced the risk of intraoperative bleeding complications compared with preoperative start (RR = 0.67, CI 0.51–0.90). Interpretation — The initiation of LMWH did not influence the mortality or the risk of reoperation in hip fracture patients treated with osteosynthesis. Postoperative start of LMWH could possibly decrease the risk of intraoperative bleeding.

Among elderly hip fracture patients, vascular events caused by thrombosis are common and the use of chemical thromboprophylaxis is a well-established routine in the management of these patients. On the other hand, the risk of intraoperative bleeding also represents a major concern that may increase both the duration of surgery and risk of reoperation, anemia and transfusions (Vera-Llonch et al. 2006). The potential complications following both bleeding and thromboembolic events can in turn prolong the need for hospitalization and rehabilitation. The Norwegian Hip Fracture Register (NHFR) has collected nationwide information on hip fractures since 2005 (Gjertsen et al. 2008). In a previous article based on data from the NHFR we found that postoperative start of LMWH increased both mortality and risk of reoperation compared with preoperative start after femoral neck fractures treated with hemiarthroplasty (Leer-Salvesen et al. 2017). On the other hand, preoperative start of the thromboprophylaxis did not increase the risk of intraoperative bleeding complications or reoperation due to hematoma. Knowledge concerning the administration of thromboprophylaxis in hip fracture patients treated with osteosynthesis is, on the contrary, still sparse. By using data in the nationwide NHFR we wanted to compare the effect of preoperative start versus postoperative start of thromboprophylaxis in hip fracture patients treated with osteosynthesis. Primary endpoints were mortality and reoperations in the first 6 months after surgery and intraoperative bleeding complications. Secondary endpoints were reoperations due to infection or hematoma.

616

Acta Orthopaedica 2018; 89 (6): 615–621

Patients with hip fractures (n = 96,599) Excluded (n = 20,372): – no drug prophylaxis, 1,124 – no information on prophylaxis, 451 – no information on start of prophylaxis, 17,578 – not LMWH-prophylaxis, 1,219 Hip fractures with known start of LMWH prophylaxis (n = 76,227) Excluded (n = 30,314) Prostheses and other procedures Osteosyntheses (n = 45,913)

Screws (n = 14,985)

Hip compression screw (n = 21,764)

Intramedullary nail (n = 9,164)

Figure 1. Flow chart for patients included in the study.

Patients and methods The NHFR started registration of primary operations and reoperations for all hip fractures in Norway in 2005 (Gjertsen et al. 2008). Compared with the Norwegian Patient Registry, the completeness of primary operations in the NHFR is 88% for osteosyntheses (Furnes et al. 2018). After each operation the surgeon fills in a one-page paper form. The form includes information on age, sex, cognitive function, type of fracture (intracapsular femoral neck fractures classified as Garden 1–2 or 3–4; trochanteric fractures classified as 2-fragmented [AO/OTA A1], multi-fragmented [AO/OTA A2], or intertrochanteric [AO/OTA A3]; and subtrochanteric fractures), ASA class, and duration of surgery. The form further provides information on the chemical thromboprophylaxis given (if used or not, which drug, dosage, and whether the first dose was given preoperatively or postoperatively). The surgeon should also report intraoperative complications, including major bleeding complications, to the register. A reoperation was defined as any secondary surgery following the primary operation. The cause of reoperation is reported by the surgeon on a similar paper form to that used for the primary operation. In the period 2005–2016, 96,599 primary operations for hip fractures were reported to the NHFR (Figure 1). 15 different types of drugs for prophylaxis had been used. However, low-molecular-weight heparin (LMWH) dominated entirely. To obtain a more homogeneous study group, operations with drugs other than LMWH were excluded. 45,913 patients operated with osteosynthesis were included. Dalteparin (Fragmin, Pfizer) was used in 58% (26,469 operations) and enoxaparin (Klexane, Sanofi-Aventis) was used in 42% (19,444 operations).

Statistics Survival analyses were performed using the Kaplan–Meier and Cox regression methods. Patients who died or emigrated during follow-up were identified from files provided by Statistics Norway, and the follow-up time for these patients was censored at the date of death or emigration. Only the first 6 postoperative months were included in the analyses as this period was considered most relevant in the present investigation. The Cox multiple regression model was used to compare the relative risks of postoperative death and revision (failure-rate ratios) among patients where the thromboprophylaxis started preoperatively compared with postoperatively, with adjustments for possible influences of sex, duration of surgery, age of the patient at surgery (grouped in decades), ASA classification and year of surgery. Analyses on mortality and risk of reoperations within 7 days postoperatively, within 30 days postoperatively, and within 6 months postoperatively were performed. In addition to reoperations for any reason, we also investigated reoperations due to infection or hematoma and reported intraoperative bleeding complications. When investigating reoperations due to hematomas, patients with concurrent other causes for reoperation were excluded from the analyses. 95% confidence intervals (CIs) were calculated for relative risks. Number needed to harm (NNH) was calculated and defined as the number of patients treated with preoperative start of LMWH in order to cause 1 intraoperative bleeding complication because of preoperative start compared with postoperative start assuming a direct causal effect. The number needed to harm was calculated as an inverse value of the risk difference between the preoperative and postoperative start of LMWH. Sub-analyses The osteosyntheses were divided into 3 sub-groups to investigate the most frequently used surgical procedures: screw osteosynthesis (14,985 operations, 33%), hip compression screw (21,764 operations, 47%), and intramedullary nail (9,164 operations, 20%). Further, sub-analyses were performed for ASA classes 1–2 and 3–5. Lastly, duration of surgery was investigated: Group 1 with a short duration of surgery (less than 16 minutes for screws, less than 40 minutes for nails or compression screws, 25% of the study populations), group 2 with a median duration of surgery (16–30 minutes for screws, 40–75 minutes for compression screws and 40–84 minutes for nails, 25–75% of the study populations), and group 3 with a long duration of surgery (more than 30 minutes for screws, more than 75 minutes for compression screws and more than 84 minutes for nails, 25% of the study populations). Assessments of proportionality in the Cox models were performed using log minus log plots of the adjusted survival curves, and the proportionality assumptions were fulfilled. We used the statistical software packages IBM SPSS® Statistics,

Acta Orthopaedica 2018; 89 (6): 615–621

Table 1. Patients included in the study

617

Relative distribution (%) 100 Preoperative start

Factor

Start of prophylaxis Preoperative Postoperative

Hip fractures with osteosynthesis, n (%) 20,563 (45) 25,350 (55) Mean duration of surgery a 51 (33) 51 (34) 60 Mean age at fracture (years) (SD) 80 (12) 80 (13) Women (%) 69 69 ASA groups, n (%) ASA 1 1,363 (6.6) 1,857 (7.3) 40 ASA 2 7,015 (34) 8,793 (35) ASA 3 10,362 (50) 12,860 (51) ASA 4 1,475 (7) 1,520 (6) 20 ASA 5 34 (0.2) 27 (0.1) Missing 314 (1.5) 293 (1.2) Type of surgery b Postoperative start Screws, n (%) 6,781 (45) 8,204 (55) 0 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Mean duration of surgery a 26 (14) 26(14) Year of operation Hip compression screw, n (%) 9,939 (46) 11,825 (54) Mean duration of surgery a 62 (31) 61 (30) Figure 2. Timeline demonstrates the development in start of thromboIntramedullary nail, n (%) 3,843 (42) 5,321 (58) prophylaxis from 2005 to 2016 for the patients observed in the study Mean duration of surgery a 66 (38) 66 (41) (n = 45,913). Hip fracture patients operated with osteosynthesis with a Values are minutes (SD) b Screws include operations

with Olmed screws. Hip compression screws include operations with a dynamic hip screw (DHS) with or without a support plate. Intramedullary nails include long and short nails with or without the use of an interlocking screw.

version 24.0 for Windows (IBM Corp, Armonk, NY, USA) and the statistical package R, version 3.4.0, (http://www.Rproject.org), for the statistical analyses. P-values < 0.05 were considered statistically significant. 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 patientsis entered in the medical records at each hospital. The Norwegian Hip Fracture Register is financed by the Western Norway. The authors declare no competing interests.

Results Table 1 presents baseline information on the patients included. Thromboprophylaxis was given preoperatively in 45% (20,563 operations) and postoperatively in 55% (25,350 operations). When comparing patients with a preoperative versus postoperative start of LMWH, no statistically significant differences in age, sex, comorbidity, or duration of surgery were found. There was an increase in postoperative initiation of LMWH during the studied period (Figure 2). Mortality Overall mortality after 6 months was 19% (8,751 of 45,913 patients). No statistically significant difference in mortality

known start of LMWH thromboprophylaxis (dalteparin or enoxaparin).

between preoperative and postoperative start of LMWH could be found. The results were consistent after 7, 30, and 180 days (Table 2, Figure 3). Reoperations After 6 months 4.5% (2,067 of 45,913 operations) had been reoperated. There were 115 reoperations (0.3%) due to infection. Only 19 reoperations due to hematoma were reported to the register. No statistically significant differences in reoperations for any cause, reoperations due to infection, or reoperation due to hematoma could be found when comparing a preoperative versus postoperative start of thromboprophylaxis (Table 2, Figure 4). Intraoperative bleeding complications 1,294 (3.0%) intraoperative complications were reported after osteosyntheses. 208 (16% of all reported complications) were intraoperative bleedings. Postoperative start of LMWH decreased the risk of intraoperative bleeding complications compared with preoperative start (RR = 0.67 (CI 0.51–0.90), NNH = 434) (Table 3). Type of osteosynthesis The hip fractures were reviewed in subgroups based on the type of osteosynthesis performed (Table 4). There was an increased 30-day mortality risk after operation with hip compression screw when LMWH was initiated postoperatively compared with preoperatively (RR = 1.10; CI 1.00–1.21). For other types of osteosyntheses, the startup time of LMWH prophylaxis had no statistically significant influence on mortality 7, 30, or 180 days after surgery. After operation with intramedullary nail there was an increased 180-day risk of reoperation due to infection after postoperative start of throm-

618

Acta Orthopaedica 2018; 89 (6): 615–621

Table 2. Mortality and risk of reoperation 180 days postoperatively after osteosynthesis for hip fracture

Total, n (%)

Start of prophylaxis, n (%) Preoperative Postoperative RR a 95% CI

7 days postoperatively (n = 45,913) Mortality 1,050 (2.3) 481 (2.3) 569 (2.2) 1.02 Reoperations 190 (0.4) 76 (0.4) 114 (0.4) 1.18 Reoperation due to infection 5 (0.0) 1 (0.0) 4 (0.0) 3.50 Reoperation due to hematoma 9 (0.0) 3 (0.0) 6 (0.0) 1.86 30 days postoperatively (n = 45,913) Mortality 3,534 (7.7) 1,606 (7.8) 1,928 (7.6) 1.05 Reoperations 627 (1.4) 275 (1.3) 352 (1.4) 1.12 Reoperation due to infection 66 (0.1) 30 (0.1) 36 (0.1) 0.88 Reoperation due to hematoma 18 (0.0) 6 (0.0) 12 (0.0) 2.04 180 days postoperatively (n = 45,913) Mortality 8,751 (19) 4,049 (20) 4,702 (19) 1.01 Reoperations 2,067 (4.5) 966 (4.7) 1,101 (4.3) 0.99 Reoperation due to infection 115 (0.3) 49 (0.2) 66 (0.3) 1.02 Reoperation due to hematoma 19 (0.0) 6 (0.0) 13 (0.1) 2.18

p-value

0.90–1.16 0.88–1.58 0.38–32.0 0.46–7.49

0.7 0.3 0.3 0.4

0.98–1.12 0.95–1.32 0.54–1.44 0.71–5.82

0.2 0.2 0.6 0.2

0.97–1.06 0.90–1.08 0.70–1.49 0.78–6.18

0.6 1.0 0.9 0.1

a Cox

relative revision risk (RR) (with preoperative start of prophylaxis as reference) is given with adjustments for possible influences of sex, ASA class, age group of the patient at surgery, duration of surgery, and year of surgery.

Table 3. Risk of intraoperative bleeding complications after osteosynthesis (n = 45,913) in hip fractures receiving screws (n = 14,985), hip compression screws (n = 21,764), or medullary nails (n = 9,164) Treatment

Intraoperative bleeding, n (%) Preop. start Postop. start RR a (95% CI) p-value

Osteosynthesis Screws HCS Intramedullary nail

118 (0.6) 3 (0.0) 102 (1.0) 13 (0.3)

90 (0.4) 2 (0.0) 71 (0.6) 17 (0.3)

0.67 0.43 0.64 1.06

(0.51–0.90) (0.07–2.68) (0.47–0.87) (0.50–2.26)

0.007 0.4 0.004 0.9

Risk Risk Preop.– Postop. difference NNH b 0.0060 – 0.0037 0.0004 – 0.0002 0.0108 – 0.0063 0.0036 – 0.0034

0.0023 434 0.0002 4,940 0.0045 222 0.0002 4,684

a See Table 2. b NNH: Number

of patients treated with preoperative start of LMWH in order to cause one intraoperative bleeding complication because of preoperative LMWH start compared with postoperative LMWH start if there is a direct causal effect. The NNH was calculated as an inverse value of the risk difference (RD) between the methods [1/ (risk preoperative start–risk postoperative start)].

Table 4. Mortality and risk of reoperation 180 days postoperatively after osteosynthesis in hip fractures receiving screws, hip compression screw, or medullary nails

Total, n (%)

Start of prophylaxis, n (%) Preoperative Postoperative RR a 95% CI

Screws (n = 14,985) Mortality 2,776 (19) 1,353 (21) 1,423 (18) 1.01 Reoperations 1,226 (8.2) 593 (9.1) 633 (8.0) 0.97 Reoperation due to infection 20 (0.1) 11 (0.2) 9 (0.1) 0.76 Reoperation due to hematoma 4 (0.0) 0 (0.0) 4 (0.1) – Hip compression screw (n = 21,764) Mortality 4,264 (20) 1,942 (20) 2,322 (20) 1.01 Reoperations 580 (2.7) 266 (2.7) 314 (2.7) 0.97 Reoperation due to infection 77 (0.4) 35 (0.4) 42 (0.4) 0.92 Reoperation due to hematoma 11 (0.1) 4 (0.0) 7 (0.1) 1.9 Intramedullary nail (n = 9,164) Mortality 1,711 (19) 754 (20) 957 (18) 0.96 Reoperations 261 (2.8) 107 (2.8) 154 (2.9) 1.2 Reoperation due to infection 18 (0.2) 3 (0.1) 15 (0.3) 3.7 Reoperation due to hematoma 4 (0.1) 2 (0.1) 2 (0.0) 1.2 a See Table

p-value

0.92–1.08 0.86–1.09 0.30–1.9 –

1.0 0.6 0.6 -

0.98–1.1 0.82–1.2 0.58–1.4 0.49–7.6

0.2 0.7 0.7 0.3

0.87–1.1 0.91–1.5 1.04–13 0.14–9.8

0.4 0.2 0.04 0.9

Acta Orthopaedica 2018; 89 (6): 615–621

619

Discussion When comparing preoperative versus postoperative start of LMWH for hip fracture patients treated with osteosynthesis, no differences in mortality or risk of reoperation were found. Preoperative start of LMWH was found to give more intraoperative bleeding complications for patients treated with hip compression screws, but not for patients receiving intramedullary nails or screws. Whether chemical prophylaxis should start preoperatively or postoperatively is controversial (Ettema et al. 2009, Borgen et Figure 3. Postoperative mortality for hip fracFigure 4. Risk of reoperation for hip fracture patients treated with osteosynthesis. ture patients treated with osteosynthesis. al. 2013). In Europe the LMWH prophylaxis has traditionally started before hip fracture surgery (Ettema et al. 2009), while in North boprophylaxis compared with preoperative start (RR = 3.7; America a postoperative initiation has been common (Kearon CI 1.04–13.2). No other statistically significant differences in and Hirsh 1995, Gomez-Outes et al. 2012, Lassen et al. 2012). risk for reoperation due to any cause, reoperation due to infec- This divergent practice between the continents may be contintion, or reoperation due to hematoma could be found (Table ued based on traditional consequences. The fear of bleeding4). After operation with hip compression screw there was a related complications has been most critical for surgeons in decreased risk of intraoperative bleeding complications after North America due to medico-legal issues. In Europe, on the postoperative start of LMWH compared with preoperative other hand, such complications have been the common responstart (RR = 0.64; CI 0.47–0.87) (see Table 3). sibility of the department and the main focus has been to prevent local and systemic thromboembolic events. Due to the ASA classification tremendous costs of antithrombotic drug development, sevPatients were stratified into ASA classes 1–2 and ASA classes eral companies have recently developed a common regimen 3–5. For both subgroups, no statistically significant differ- for both continents. Nevertheless, timing in relation to surences in mortality or risk of reoperation could be found within gery has been considered important for the efficacy-to-safety 180 days of follow-up between preoperative and postoperative balance in any pharmaceutical anticoagulant program. Trials start of LMWH (data not shown). funded by the industry have primarily focused on detecting thromboses with mandatory radiology following the surgical Duration of surgery intervention. Unfortunately, some studies have been criticized Patients treated with intramedullary nail with long duration of for underestimating the challenges bleeding and wound comsurgery (> 84 minutes, upper quartile) had the most marked plications can present following surgery (Parvizi et al. 2007, increased risk of reoperation due to infection after postopera- Lachiewicz 2009, Dahl et al. 2010). Second, the reported trials tive start of LMWH compared with preoperative start (RR = have been designed to show potential favorable effects of new 8.2; CI 1.03–65). The startup time of LMWH did not statis- experimental regimes versus established regimes (Yoshida et tically significantly influence the risk of reoperation due to al. 2013). In contrast, our register study compares the same infection in patients treated with intramedullary nails with compounds when investigating preoperative versus postoperashorter operation time or for other osteosyntheses (data not tive start of thromboprophylaxis. To our knowledge, there no study of this size has been conducted investigating the startup shown). For patients treated with hip compression screw the risk time of thromboprophylaxis in hip fracture surgery. When studying femoral neck fractures treated with hemiarof intraoperative bleeding complication decreased after postoperative start of LMWH with long duration of surgery (> throplasty, preoperative start of thromboprophylaxis reduced 75 minutes, upper quartile) (RR = 0.68; CI 0.47–0.99). The mortality within 6 months of surgery (Leer-Salvesen et al. startup time of LMWH did not influence the risk of intra- 2017). This favorable effect of preoperative LMWH adminisoperative bleeding complication in patients treated with hip tration was most pronounced in the first postoperative weeks. compression screw with shorter operation time (data not Nevertheless, the preoperative effect was also robust over time shown). and no catch-up effect was noticed during 6 months of observation. In the present study investigating hip fractures treated with osteosynthesis, no such protective effect of preoperative

620

start of LMWH could be detected, either for the whole group or for subgroups of patients receiving screw osteosynthesis, hip compression screw, or intramedullary nail. In the frail elderly undergoing hip fracture, the primary trauma and subsequent surgery have been shown to significantly impact the immediate and long-term mortality (Talsnes et al. 2011, 2013). As shown in our previous study, preoperative administration of LMWH may contribute to reducing mortality following hemiarthroplasty (Leer-Salvesen et al. 2017). Importantly, osteosynthesis for hip fractures seem to induce less trauma and thrombin-driven vascular complications as compared with surgery with hemiprosthesis. Thus, the start of prophylaxis in relation to surgery might be less important when conducting osteosynthesis procedures. Preoperative start of thromboprophylaxis did not increase the risk of reoperation after osteosyntheses compared with postoperative start. In previous discussions, the risk of reoperation has in particular been brought forward as an argument to start thromboprophylaxis postoperatively (Lassen et al. 2012). This argument has partly been based on the fear of intraoperative bleeding complicating the surgical intervention. The fear of bleeding might also explain the gradual shift from preoperative to postoperative initiation of LMWH observed during the last decade. Relevantly, our study did demonstrate a decreased risk of intraoperative bleeding when the LMWH was initiated postoperatively compared with preoperatively for patients receiving osteosynthesis. In contrast, we did not find a decreased risk of intraoperative bleeding when the LMWH was initiated postoperatively in femoral neck fractures treated with hemiarthroplasty (Leer-Salvesen et al. 2017). Postoperative start of LMWH decreased the risk of intraoperative bleeding in connection with hip compression screws. However, in patients operated with screw osteosynthesis or intramedullary nail a postoperative start of LMWH did not influence the risk of intraoperative bleeding. Screw osteosynthesis and intramedullary nail for hip fractures are most often performed as mini-invasive surgery, which may explain why risk of intraoperative bleeding complications is not affected by LMWH. Since displaced femoral neck fractures in the elderly are most often treated with an arthroplasty (Gjertsen et al. 2017), patients treated with screw osteosynthesis are younger than patients treated with osteosynthesis for extracapsular hip fractures. Less preoperative bleeding and younger age might explain why the start of LMWH administration does not significantly affect the risk of mortality, reoperation, or bleeding when using screw fixation. Extracapsular hip fractures may be treated with hip compression screw or intramedullary nail. The complexity of the fracture will affect the risks of infection, bleeding, and reoperation. Intramedullary nails are increasingly used for complex trochanteric and subtrochanteric fractures. Postoperative start of LMWH compared with preoperative start increased the risk of reoperations due to infection after intramedullary nails. Hip fractures with time-consuming intramedullary nail-

Acta Orthopaedica 2018; 89 (6): 615â&#x20AC;&#x201C;621

ing (more than 84 minutes, upper quartile) had a more than 8 times increased risk of reoperation due to infection after postoperative start of LMWH compared with preoperative start. A possible explanation could be that preoperative start of thromboprophylaxis induces bleeding earlier, which allows hemostasis or hematoma evacuation during surgery. Contrariwise, postoperative start of thromboprophylaxis LMWH may postpone traumatic bleeding and produce a late hematoma, which predispose to infection. A longer duration of surgery is often related to complex fractures, other concurrent intraoperative complications, or less experienced surgeons. These factors may also influence the risk of infection and potentiate the protective effect of a preoperative LMWH. However, due to limited number of patients in the sub-studies the results may be interpreted with caution. Strength and limitations Our study is not a randomized controlled study (RCT) and may consequently be categorized as a hypothesis-creating study. However, when examining rare adverse events and large patient numbers, observational register studies remain an achievable method compared with RCTs. Nevertheless, this is, to our knowledge, the first study of its kind comparing preoperative versus postoperative benefit of the same compounds. The strength in our study is the inclusion of data from all surgical units treating hip fractures in one country. Accordingly, the external validity of the results is high. The inclusion of multiple hospitals may influenceour results. However, as we believe each hospital has routines in thromboprophylaxis administration, the inclusion of all hospitals has not been used as a confounder in our analyses. The register is dependent on volunteer reporting from individual surgeons. According to earlier coverage analyses, reoperations are more frequently left unreported to the register than primary operations, and such a tendency may affect our endpoints (Wiik et al 2014). It is, however, unlikely that differences in the reporting of reoperations after preoperative versus postoperative initiation of LMWH exist. Data concerning the start of LMWH are entered immediately after surgery by the responsible surgeon. Unfortunately, the NHFR only receives limited information concerning the usage of thromboprophylaxis in treatment of hip fractures. The surgeons do not report the exact moment in time of LMWH administration before and after surgery. Therefore, consequences of preoperative versus postoperative LMWH startup regimes are still partly unknown. The exact duration of LMWH treatment may not always be possible to predict immediately after surgery as later events or complications may alter the originally planned duration of treatment. Furthermore, after each operation the surgeon has to interpret whether the intraoperative bleeding was so extensive that it should be reported as a complication. Therefore, the findings regarding intraoperative bleeding must be interpreted with caution. Even though the study has weaknesses, the results are based on a large number of patients as well as consistent reporting of LMWH initiation to the register.

Acta Orthopaedica 2018; 89 (6): 615–621

Conclusion Our data strongly indicate that preoperative compared with postoperative administration of LMWH does not influence mortality or risk of reoperation in hip fracture patients treated with osteosynthesis. However, postoperative start of LMWH does decrease the risk of reported intraoperative bleeding complications for operations with hip compression screw, but not with intramedullary nail or screw osteosynthesis.

The loyal reporting from all orthopedic surgeons made our studies possible. The manuscript was produced by close teamwork between all authors. SLS and ED performed the statistical analyses. All authors contributed to the study design and interpretation of results. Acta thanks Gerold Labek and other anonymous reviewers for help with peer review of this study.

Borgen P O, Dahl O E, Reikeras O. Biomarkers of coagulation and fibrinolysis during cemented total hip arthroplasty with pre- versus postoperative start of thromboprophylaxis. Thrombosis 2013; 2013: 563217. Dahl O E, Quinlan D J, Bergqvist D J, Eikelboom J W. A critical appraisal of bleeding events reported in venous thromboembolism prevention trials of patients undergoing hip and knee arthroplasty. J Thromb Haemost 2010; 8: 1966-75. Ettema H B, Mulder M C, Nurmohamed M T, Buller H R, Verheyen C C. Dutch orthopedic thromboprophylaxis: a 5-year follow-up survey. Acta Orthop 2009; 80(1): 109-12. Furnes O, Gjertsen JE, Hallan G, Visnes H, Gundersen T, Fenstad AM, Dybvik E, Bartz-Johannessen C. Annual report. Norwegian Arthroplasty Register 2018; http://nrlweb.ihelse.net/Rapporter/Rapport2018.pdf. p. 234 Gjertsen J E, Engesaeter L B, Furnes O, et al. The Norwegian Hip Fracture Register: experiences after the first 2 years and 15,576 reported operations. Acta Orthop 2008; 79(5): 583-93.

621

Gjertsen J E, Dybvik E, Furnes O, Fevang J M, Havelin L I, Matre K, Engesæter L B. Improved outcome after hip fracture surgery in Norway. Acta Orthop 2017; 88(5): 505-11. Gomez-Outes A, Terleira-Fernandez A I, Suarez-Gea M L, Vargas-Castrillon E. Dabigatran, rivaroxaban, or apixaban versus enoxaparin for thromboprophylaxis after total hip or knee replacement: systematic review, metaanalysis, and indirect treatment comparisons. BMJ 2012; 344:e 3675. Kearon C, Hirsh J. Starting prophylaxis for venous thromboembolism postoperatively. Arch Intern Med 1995; 155:3 66-72. Lachiewicz P F. Comparison of ACCP and AAOS guidelines for VTE prophylaxis after total hip and total knee arthroplasty. Orthopedics 2009; 32: 74-8. Lassen M R, Gent M, Kakkar A K, et al. The effects of rivaroxaban on the complications of surgery after total hip or knee replacement: results from the RECORD programme. J Bone Joint Surg Br 2012; 94(11): 1573-8. Leer-Salvesen S, Dybvik E, Dahl O E, Gjertsen J-E, Engesæter L B. Postoperative start compared to preoperative start of low-molecular-weight heparin increases mortality in patients with femoral neck fractures. Acta Orthop 2017; 88(1): 48-54 Parvizi J, Ghanem E, Joshi A, Sharkey P F, Hozack W J, Rothman R H. Does “excessive” anticoagulation predispose to periprosthetic infection? J Arthroplasty 2007; 22: 24-8. Talsnes O, Hjelmstedt F, Dahl O E, Pripp A H, Reikeras O. Clinical and biochemical prediction of early fatal outcome following hip fracture in the elderly. Int Orthop 2011; 35(6): 903-7. Talsnes O, Vinje T, Gjertsen J E, et al. Perioperative mortality in hip fracture patients treated with cemented and uncemented hemiprosthesis: a register study of 11,210 patients. Int Orthop 2013; 37(6): 1135-40. Vera-Llonch M, Hagiwara M, Oster G. Clinical and economic consequences of bleeding following major orthopedic surgery. Thromb Res 2006; 117(5): 569-77. Wiik R, Baste V, Gjertsen J E, Engesæter L B Dekningsgradsanalyse for Nasjonalt Hoftebruddregister 2008-2012. Helsedirektoratet 13/11218. https:// helsedirektoratet.no/Lists/Publikasjoner/Attachments/675/Dekningsgradsanalyse-for-najonalt-hoftebruddregister-2008-2012-IS-2210.pdf Yoshida Rde A, Yoshida W B, Maffei F H, El Dib R, Nunes R, Rollo H A. Systematic review of randomized controlled trials of new anticoagulants for venous thromboembolism prophylaxis in major orthopedic surgeries, compared with enoxaparin. Ann Vas Surg 2013; 27(3): 355-69.

622

Acta Orthopaedica 2018; 89 (6): 622–627

Does the physical activity profile change in patients with hip dysplasia from before to 1 year after periacetabular osteotomy? Julie SANDELL JACOBSEN 1,2, Kristian THORBORG 3, Per HÖLMICH 3, Lars BOLVIG 4, Stig STORGAARD JAKOBSEN 5, Kjeld SØBALLE 5,6, and Inger MECHLENBURG 5–7 1 Department of Physiotherapy and Department of Research in Rehabilitation and Health Promotion, Faculty of Health Sciences, VIA University College, Aarhus; 2 Department of Physiotherapy and Occupational Therapy, Aarhus University Hospital, Aarhus; 3 Sports Orthopaedic Research CenterCopenhagen (SORC-C), Department of Orthopaedic Surgery, Copenhagen University Hospital, Amager-Hvidovre, Hvidovre; 4 Department of Radiology, Aarhus University Hospital, Aarhus; 5 Department of Orthopaedic Surgery, Aarhus University Hospital, Aarhus; 6 Department of Clinical Medicine, Aarhus University, Aarhus, Denmark; 7 Department of Public Health, Aarhus University, Aarhus, Denmark

Correspondence: jsaj@via.dk Submitted 2018-04-23. Accepted 2018-09-03.

Background and purpose — Knowledge of physical activity profiles among patients with hip dysplasia is lacking. We investigated whether patients with hip dysplasia change physical activity profile from before to 1 year after periacetabular osteotomy. Furthermore, we investigated associations between change in accelerometer-based physical activity and change in self-reported participation in preferred physical activities (PA). Patients and methods — Physical activity was objectively measured at very low to high intensity levels with accelerometer-based sensors. Subjectively, PA was recorded with Copenhagen Hip and Groin Outcome Score (HAGOS) in 77 patients. Associations between the 2 were analyzed with simple linear regression analyses. Results — Changes in accelerometer-based physical activity ranged from –2.2 to 4.0 % points at all intensity levels from baseline to 1-year follow-up. These changes represent very small effect sizes (–0.16 to 0.14). In contrast, self-reported PA showed a statistically and clinically relevant increase of 22 (CI 14–29) HAGOS PA points 1 year post-surgery. Associations between change in accelerometer-based physical activity and change in self-reported PA were, however, not statistically significant and correspond to a percentage change in physical activity of only –0.87% to 0.65% for a change of 10 HAGOS PA points. Interpretation — Patients with hip dysplasia do not seem to change physical activity profile 1 year post-surgery if measured with objective accelerometer-based sensors. This is interesting as self-reported PA indicates that patients’ ability to participate in physical activity increases, suggesting that this increased self-reported participatory capacity is not manifested as increased objectively measured physical activity.

Hip dysplasia is a pathological development of the hip joint, presenting in early adulthood. The clinical presentation of symptomatic hip dysplasia is pain caused by intra-articular injury (Nunley et al. 2011) and coexisting muscle–tendon pain (Jacobsen et al. 2018b). The pain prevents patients from participating in their preferred physical activities (Jacobsen et al. 2013), including higher intensity activities such as running (Novais et al. 2013). Periacetabular osteotomy (PAO) is the preferred joint-preserving treatment for patients with symptomatic hip dysplasia (Troelsen et al. 2008). PAO has been shown to result in pain relief, restoration of function, and high hip survival rates (Lerch et al. 2017). Furthermore, previous studies on selfreported physical activity have shown that patients increase their ability to participate in their preferred physical activities 1 year after PAO (Jacobsen et al. 2014, Klit et al. 2014, Khan et al. 2017) and spend more time on physical activities at higher intensities (Novais et al. 2013). Discrepancies between selfreported physical activity and objectively measured physical activity have been reported (Sallis and Saelens 2000, Harding et al. 2014). However, it is important to recognize that selfreported physical activity captures the capability to complete a given task, while the ability to measure the actual performance is limited (Smuck et al. 2017). Accelerometer-based measures of physical activity allow us to evaluate if patients with hip dysplasia change objectively measured physical activity profile, alongside with subjective measured levels of improved pain, physical function, and quality of life after PAO. We investigated whether patients with hip dysplasia change physical activity profile from before to 1 year after periacetabular osteotomy, measured by accelerometer-based sensors and self-reported physical activity. Furthermore, we investigated associations between change in accelerometer-based physical activity and change in self-reported participation in preferred physical activities (PA).

Acta Orthopaedica 2018; 89 (6): 622–627

Patients and methods Patients 100 consecutive patients with unilateral and bilateral symptomatic hip dysplasia were included in this study from May 2014 to August 2015 at the Department of Orthopedics, Aarhus University Hospital, Denmark. Inclusion criteria to participate in this study was Wiberg’s center-edge (CE) angle < 25 degrees (Wiberg 1939), groin pain for at least 3 months, and scheduled PAO. Patients with known co-morbidities and a history of previous surgical interventions affecting the hip function were excluded. Further details on study design was reported in our previous studies on the same study population, reporting ultrasonography-detected abnormalities and prevalence of muscle–tendon pain prior to PAO (Jacobsen et al. 2018a, b). Baseline characteristics Baseline characteristics were recorded at baseline by standardized questions. Pain at rest was measured on a numeric rating scale (NRS), and radiographic acetabular angles and Tönnis’s osteoarthritis grade were measured by a single rater using standing anteroposterior radiographs. Hospital charts were used to retrieve data on previous surgeries and co-morbidities. Periacetabular osteotomy 2 experienced orthopedic surgeons used the minimally invasive approach for PAO. The surgical procedure involves reorientation of the acetabulum through 3 separate osteotomies and correction of the insufficient coverage of the femoral head. Compared with other PAO procedures, the minimally invasive approach involves minor incision of the soft tissue and enables early mobilization (Troelsen et al. 2008). Post-surgery, all patients underwent an in-hospital standardized rehabilitation program, and only partial weight bearing with a maximum load of 30 kg was allowed in the first 6–8 weeks. Patients were discharged after approximately 2 days. Accelerometer-based physical activity The patients’ habitual physical activity profile was objectively measured with a commercial tri-axial accelerometer-based sensor (Gulf Coast Data Concepts, Waveland, MS, USA) during 7 consecutive days (working and leisure days) at baseline and 1 year after PAO. The sensor was taped to the lateral non-affected upper leg of patients using hypo-allergenic double-sided tape (3M, Maplewood, MN, USA). The position of the sensor was at the mid-thigh between the trochanter and lateral condyle, with the y-axis of the sensor along the axis of the femur (van Laarhoven et al. 2016). The sensor was worn during waking hours for a minimum of 8 hours per day. It was removed for the night and during both showering and swimming activities. Data from –6g to 6g were sampled at 50 Hz.

623

The raw activity data were transferred to a computer and processed using a validated algorithm (Lipperts et al. 2017). Data were visually divided into separate days using a MatLab (https://www.mathworks.com/products/matlab.html) script designed for the purpose, and non-worn days were excluded. All data were calibrated manually by selecting a period of level walking in the dataset of each day to adapt variations in height, morphology, sensor placement, walking style, and speed (Lipperts et al. 2017). After calibration, the data were run through an algorithm based on a decision tree, and divided into resting, standing, cycling, level walking, walking on stairs (upstairs and downstairs), and running. Further details are described by Lipperts et al. (2017). In addition, the algorithm constructed an intensity variable on the distribution of the average signal intensity in 10-second intervals grouped into 4 intensity levels. The intensity levels were: (i) Very low intensity such as sitting and standing (0–0.05g), (ii) Low intensity such as standing and shuffling (0.05–0.1g), (iii) Moderate intensity such as slow and normal walking (0.1–0.2g), (iv) High intensity such as brisk walking, running, and jumping (> 0.2g). Self-reported physical activity Self-reported physical activity was subjectively measured as preferred physical activity participation (PA) using the Copenhagen Hip and Groin Outcome Score (HAGOS) (Thorborg et al. 2011), and completed by the patients at baseline and 1 year after PAO. Besides participation in physical activities, the HAGOS outcome score captures pain, symptoms, physical function in daily living (ADL), physical function in sports and recreation, and quality of life on a scale from 0 to 100 points, where 100 points indicate the highest outcome. HAGOS is a valid, reliable, and responsive outcome measure developed for young patients with longstanding hip and groin pain (Thorborg et al. 2011). Additionally, time spent on preferred PA per week was recorded at baseline and 1 year after PAO. Sample size consideration The purpose of this study was to investigate whether patients change physical activity profile from before to 1 year after surgery. We considered moderate effect sizes to be of importance, which required 85 patients for detecting an effect size of 0.5 from before to 1 year after surgery. This calculation was based on a power of 90% and a significance level of 5%. Statistics Parametric continuous outcome measures were reported as means (SD) if normally distributed; otherwise reported as medians with either interquartile ranges (IQR) or 95% confidence intervals (CI). Normality was assessed with histograms and probability plots. Categorical data were reported as number of events and percentage with CI. Time spent on physical activities was normalized to total accelerometer-based wear time before surgery and 1 year after surgery. Number

624

Acta Orthopaedica 2018; 89 (6): 622–627

Table 1. Baseline characteristics of 97 consecutive patients with hip dysplasia

Informed consent n = 100 Missing baseline data: Error in output from accelerometer-based sensor n=3 Baseline activity data n = 97 Lost to follow-up (n = 17): – postponed surgery, 6 – time and transport, 4 – serious disease unrelated to PAO, 3 – injuries unrelated to PAO, 2 – non-union of the pubic bone, fracture, 1 – emigrated, 1 Missing follow-up data: Did not return accelerometer n=3 Complete follow-up data but missing baseline data due to error in output from accelerometer-based sensor n=1

Paired analysis n = 77

Follow-up activity data n = 78

Figure 1. Flowchart of the study process among 100 consecutive patients with symptomatic hip dysplasia scheduled for periacetabular osteotomy (PAO).

of events at baseline were normalized to average daily accelerometer-based wear time at baseline, and the number of events at 1-year follow-up were normalized to average daily accelerometer-based wear time at 1-year follow-up. Differences between baseline and 1-year follow-up were tested with a paired t-test if assumptions were met and otherwise tested with the non-parametric Wilcoxon signed-rank test. Effect sizes were calculated from the paired t-test as Cohen’s d on the formula: t statistic/√(n) for parametric data and otherwise from the non-parametric Wilcoxon signed-rank test on the formula: z statistic/√(n). Furthermore, simple linear regression analyses were performed assessing associations between change in objectively measured physical activity at 4 intensity levels (very low, low, moderate, and high intensity) and change in subjectively measured HAGOS PA from baseline to 1 year post-surgery. Crude ß-coefficients were estimated, and the assumptions (independent observations, linear association, constant variance of residuals, and normal distribution of residuals) for the linear regression analyses were met. The crude ß-coefficients refer to the slope of the regression line (the increase in objectively measured physical activity for every unit increase in subjectively measured HAGOS PA). STATA 14.0 (StataCorp, College Station, TX, USA) software package was used for data analysis and significant results were presented if p < 0.05.

Outcomes Baseline Men 15 Age, years (SD) 30 (9) BMI, kg/m2 (SD) 23 (3) CE angle, degrees (SD) 17 (4.7) AI angle, degrees (SD) 14 (4.8) Tönnis osteoarthritis grade 1 3 Bilateral hip dysplasia 86 Hours spent on self-reported preferred physical activities/week, n <2.5 13 2.5 to <5 18 5 to <10 40 ≥10 26 Self-reported preferred physical activities, n Fitness 27 Running 19 Team sports 13 Gymnastics 6 Horseback riding 6 Bicycling 5 Dancing 3 Other 16 No one 2 Abbreviations: SD (standard deviation); BMI (body mass index), CE (center-edge), AI (Tönnis’ acetabular index).

Ethics, registration, funding, and potential conflict of interests This study was done in accordance with the STROBE statement and in compliance with the Helsinki Declaration. The study was notified to the Central Denmark Region Committee on Biomedical Research Ethics on January 14, 2014 (5/2014). The Danish Data Protection Agency gave permission for the handling of personal data (1-16-02-47-14), and the protocol was registered at ClinicalTrials.gov (20140401PAO). All patients gave informed written consent. This study was funded by the Danish Rheumatism Association [grant number: A3280], the Aase and Ejnar Danielsen Fund [grant number: 10-000761/LPJ], and the Fund of Family Kjaersgaard, Sunds [date: 01-05-2018]. The authors have no competing interests.

Results 100 consecutively enrolled patients gave informed consent and were included in this study. Data on objective accelerometer-based physical activity were lost on 3 patients at baseline. At 1-year follow-up 17 patients were lost to follow-up and objective accelerometer-based physical activity data were lost on further 3 patients (Figure 1). Baseline characteristics of the included patient are reported in Table 1. Pain at rest measured on an NRS was 3.0 (1.5–4.0)

Acta Orthopaedica 2018; 89 (6): 622–627

625

Table 2. Accelerometer-based physical activity and self-reported physical activity in hip dysplasia (n = 77) Outcomes

Baseline 1-year follow-up mean (SD) mean (SD)

Difference mean % points (CI) Effect size

Percent of time measured by accelerometer-based sensors Very low intensity 74 (9.1) 76 (8.9) 1.6 (–0.89 to 4.0) Low intensity 14 (5.1) 14 (5.3) –0.66 (–2.2 to 0.89) Moderate intensity 6.8 (2.9) 6.4 (2.6) –0.40 (–1.1 to 0.31) High intensity 4.6 (2.0) 4.3 (2.0) –0.32 (–0.77 to 0.13) Self-reported physical activity HAGOS PA a 23 (24) 45 (33) 22 (14 to 29) Hours in PA/week b,c 0.0 (0.0–2.3) 1.0 (0.0–3.0) –

p-value

0.14 –0.096 –0.13 –0.16

0.2 0.4 0.3 0.2

0.67 0.18

<0.001 0.1

a HAGOS (Copenhagen Hip and Groin Outcome Score, 0–100 points), b PA (preferred physical activity participation) c Non-parametric data presented as median (interquartile range).

Table 3. General physical activity profile based on objective data on patients with hip dysplasia Outcomes

Baseline (n = 97) median (CI)

1-year follow-up (n = 78) median (CI)

Cadence as steps/min 99 (98–100) 100 (98–102) Events/day, n Total steps 7,404 (6,645–8,418) 7,925 (6,637–8,612) Steps (level) 6,923 (6,192–7,709) 7,322 (6,081–8,217) Steps (up) 266 (194–403) 235 (171–313) Steps (down) 155 (134–183) 146 (123–169) Total wear time, h/day 14 (14–15) 15 (14–15) Time as percentage Resting 64 (61–68) 63 (59–66) Standing 23 (22–27) 26 (23–27) Walking 11 (9.9–13) 11 (9.3–13) Cycling 0.15 (0.063–0.33) 0.084 (0.046–0.18) Running 0.011 (0.0042–0.020) 0.0078 (0.0040–0.025)

Change in accelerometer-based PA 0.10

0.05

–0.05 –50

100

Change in self-reported HAGOS PA

Figure 2. Scatter plot of the prediction of change in accelerometerbased physical activity (PA) at high intensity as a linear function of change in self-reported preferred physical activity participation (PA) measured by the Copenhagen Hip and Groin Outcome Score (HAGOS) from baseline to 1 year post-surgery.

at baseline, and at 1-year follow-up it was 0.0 (0.0–1.3). A sensitivity analysis of the 17 patients lost to follow-up revealed no differences in age, sex, pain, and radiological angles compared with the analyzed patients.

erate effect size. Similarly, a statistically non-significant change in hours spent on preferred physical activities at 1-year follow-up was present, represented by a very small effect size (Table 2).

Accelerometer-based physical activity The accelerometer-based sensor was worn for a median of 7 days (3–8) at both baseline and 1-year follow-up; only 6 patients had worn the sensor for less than 5 days. Changes in objectively measured physical activity ranging from –2.2 to 4.0 % points, across all intensity levels (covering very low, low, moderate, and high intensity) from baseline to 1-year follow-up (Table 2). This was, however, not statistically significant, and represents very small effect sizes ranging from –0.16 to 0.14. A general physical activity profile of the patients is reported in Table 3.

Associations between accelerometer-based physical activity and self-reported physical activity Associations between change in accelerometer-based physical activity and change in HAGOS PA ranged from –0.00087 to 0.00065, covering very low, low, moderate, and high intensity from baseline to 1-year follow-up. These were, however, not statistically significant and correspond to a percentage change in physical activity of only –0.87% to 0.65% for a change of 10 HAGOS PA points. The individual associations were as follows: Very low intensity: ß = –0.00011 (CI –0.00087 to 0.00065), p = 0.8. Low intensity: ß = –0.000022 (CI –0.00050 to 0.00046), p = 0.9. Moderate intensity: ß = 0.000044 (CI –0.00018 to 0.00027), p = 0.7. High intensity: ß = 0.000072 (CI –0.000066 to 0.00021), p = 0.3 (Figure 2).

Self-reported physical activity Subjectively measured HAGOS PA increased statistically significantly from baseline to 1-year follow-up with a mod-

626

Discussion Using accelerometer-based sensors, we found that the objectively measured physical activity profile was the same before and 1 year after PAO at all the measured physical activity intensity levels, indicated by very low effect sizes. In contrast, subjectively measured physical activity showed a moderate increase, and this was higher than the minimal important change (MIC) of HAGOS PA (Thomeé et al. 2014). Our linear regression analyses showed absence of any statistically significant and clinically relevant associations between objectively and subjectively measured physical activity as a change of 10 HAGOS PA points would maximally result in a change of physical activity of 1%. In a systematic review, Tudor-Locke et al. (2011) reported minimum international recommendations for physical activity in healthy populations. The minimum number of steps for adults was 7–8,000 steps per day with an average cadence of 100 steps per minute indicating moderate intensity walking (Tudor-Locke et al. 2011). The general physical profile (Table 3) of our patients shows that their overall level of physical activity was at level with reference values for both steps and cadence at baseline and at the 1-year follow-up, indicating that the patients followed current public health guidelines. Similarly, an American study reported daily strides by a step watch in adults with hip dysplasia and in healthy adults. No differences were seen between the 2 groups (Harris-Hayes et al. 2012). The average number of strides per day was 4,627 (i.e., 9,254 steps) among the adults with hip dysplasia, similar to the average number of steps reported among our patients. Additionally, physical activity was grouped into 4 different intensity levels based on strides per minute. No differences between the groups were found, and the effect sizes were very low (Harris-Hayes et al. 2012). Overall, these 4 intensity levels were comparable to the 4 intensity levels for physical activity in our study, and the results were similar to our results. This suggests that the physical activity profile among Danish patients with hip dysplasia is comparable to that of healthy American adults both before and 1 year after PAO, when measured objectively with accelerometer-based sensors. This finding is in line with observations among patients undergoing total hip arthroplasty. A previous study measured physical activity with objective accelerometer-based sensors in 19 patients (mean age 69 years) with hip osteoarthritis before and 6 months after total hip arthroplasty (Harding et al. 2014). Similar to our findings, no significant increase in objectively measured physical activity was found after surgery. Self-reported physical activity after surgery has been reported in 6 previous studies of patients with hip dysplasia (Novais et al. 2013, Jacobsen et al. 2014, Beaulé et al. 2015, Clohisy et al. 2017, Hara et al. 2017, Khan et al. 2017). In 1 prospective, multicenter study, the physical activity level was measured subjectively with the University of California, Los Angeles

Acta Orthopaedica 2018; 89 (6): 622–627

(UCLA) activity scale in 371 patients (mean age 25 years) with hip dysplasia (Clohisy et al. 2017). The patients’ physical activity level had increased statistically significantly 2 years post-surgery, alongside with improvement in self-reported pain, physical function, and quality of life. These results are in line with the results of 4 other previous studies in patients with hip dysplasia reporting statistically significantly and clinically relevant increases in self-reported physical activity levels after surgery (Novais et al. 2013, Beaulé et al. 2015, Hara et al. 2017, Khan et al. 2017). In a 6th study, a statistically significantly and clinically relevant increase in HAGOS PA was reported among 32 patients (mean age 34 years) with hip dysplasia 1 year after PAO compared with baseline (Jacobsen et al. 2014). Overall, the results of the 6 previous studies showed statistically significantly and clinically relevant increases in subjectively measured physical activity levels after surgery (Novais et al. 2013, Jacobsen et al. 2014, Beaulé et al. 2015, Clohisy et al. 2017, Hara et al. 2017, Khan et al. 2017); this corroborates our findings. Difficulties in participating in preferred physical activities are reported as important components of disability among patients with hip pain (Novais et al. 2013). We observed increased HAGOS PA and unchanged objectively measured physical activity 1 year post-surgery. The discrepancy probably exists because HAGOS PA and accelerometer-based methods measure different aspects of physical activity. Accelerometer-based methods provide an objective measure of physical activity, while HAGOS PA measures the patients’ self-perceived ability to participate in preferred physical activity. Moreover, age, employment status, culture, and geography have been shown to be associated with activity level (Hirata et al. 2006, Sechriest II et al. 2007, Althoff et al. 2017). Changing one’s physical activity profile is complex (Harding et al. 2014), because it is under the influence of one’s lifestyle and thus both patients and health professionals need to be aware that PAO alone is unlikely to change the physical activity profile from before to 1 year after PAO. Limitations We included no healthy controls and estimates of both accelerometer-based physical activity and self-reported physical activity could therefore not be compared with estimates in healthy controls. The negative impact of this is presumably small since it was possible to compare our findings with the findings of a previous study that included a control group (Jacobsen et al. 2014). Processing of the accelerometer-based physical activity data does involve some measure of subjectivity as the researcher has to calibrate the algorithm manually (Lipperts et al. 2017). To reduce potentially systematic effects, we processed all data in pairs, and in this way processing of baseline data was followed by processing of 1-year follow-up data in each patient. Still, we cannot rule out that the subjective element may increase the overall variability. The intensity variable shows the percentage of time at 4 different intensity

Acta Orthopaedica 2018; 89 (6): 622–627

levels, and other thresholds may have given other results. However, a previous study on young and middle-aged adults with hip disorders found similar results, even though their intensity levels were based on strides per minute (Harris-Hayes et al. 2012). This similarity suggests that our findings are valid. Summary Patients with hip dysplasia do not seem to change physical activity profile 1 year post-surgery if measured with objective accelerometer-based sensors. This is interesting as selfreported PA indicates that patients’ ability to participate in physical activity increases, suggesting that this increased participatory capacity is not manifested as increased objectively measured physical activity. These findings are similar to another study on patients undergoing total hip arthroplasty where no increase in accelerometer-based physical activity was found after surgery. All authors took part in planning the study and writing of the present manuscript. IM, KT, PH, and LB concentrated on the design and use of methods. JSJ tested the participants. KS and SSJ were responsible for inclusion of patients, and SSJ made the radiologic evaluations. JSJ coordinated all elements and made the first draft of the manuscript. The authors would like to thank Bernd Grimm, Matthijs Lipperts, and Marianne Tjur for their invaluable help and advice carrying out the accelerometer analysis. Furthermore, they would like to thank Charlotte Møller Sørensen and Gitte Hjørnholm Madsen for their assistance in collecting accelerometer data from the patients. Acta thanks Jan Erik Madsen and other anonymous reviewers for help with peer review of this study.

Althoff T, Sosič R, Hicks J L, King A C, Delp S L, Leskovec J. Large-scale physical activity data reveal worldwide activity inequality. Nature 2017; 547 (7663): 336-9. Beaulé P E, Dowding C, Parker G, Ryu J-J. What factors predict improvements in outcomes scores and reoperations after the Bernese periacetabular osteotomy? Clin Orthop Relat Res 2015; 473 (2): 615-22. Clohisy J C, Ackerman J, Baca G, Baty J, Beaulé P E, Kim Y-J, Millis M B, Podeszwa D A, Schoenecker P L, Sierra R J, Sink E L, Sucato D J, Trousdale R T, Zaltz I. Patient-reported outcomes of periacetabular osteotomy from the Prospective ANCHOR Cohort Study. J Bone Joint Surg Am 2017; 99 (1): 33-41. Hara D, Hamai S, Fukushi J-I, Kawaguchi K-I, Motomura G, Ikemura S, Komiyama K, Nakashima Y. Does participation in sports affect osteoarthritic progression after periacetabular osteotomy? Am J Sports Med 2017; 45 (11): 2468-75. Harding P, Holland A E, Delany C, Hinman R S. Do activity levels increase after total hip and knee arthroplasty? Clin Orthop Relat Res 2014; 472 (5): 1502-11. Harris-Hayes M, Steger-May K, Pashos G, Clohisy J C, Prather H. Stride activity level in young and middle-aged adults with hip disorders. Physiother Theory Pract 2012; 28 (5): 333-43. Hirata S, Ono R, Yamada M, Takikawa S, Nishiyama T, Hasuda K, Kurosaka M. Ambulatory physical activity, disease severity, and employment status in adult women with osteoarthritis of the hip. J Rheumatol 2006; 33 (5): 939-45.

627

Jacobsen J S, Nielsen D B, Sørensen H, Søballe K, Mechlenburg I. Changes in walking and running in patients with hip dysplasia. Acta Orthop 2013; 84 (3): 265-70. Jacobsen J S, Nielsen D B, Sørensen H, Søballe K, Mechlenburg I. Joint kinematics and kinetics during walking and running in 32 patients with hip dysplasia 1 year after periacetabular osteotomy. Acta Orthop 2014; 85 (6): 592-9. Jacobsen J S, Bolvig L, Hölmich P, Thorborg K, Jakobsen S S, Søballe K, Mechlenburg I. Muscle-tendon-related abnormalities detected by ultrasonography are common in symptomatic hip dysplasia. Arch Orthop Trauma Surg 2018a; 138 (8): 1059-67. Jacobsen J S, Hölmich P, Thorborg K, Bolvig L, Jakobsen S S, Søballe K, Mechlenburg I. Muscle–tendon-related pain in 100 patients with hip dysplasia: prevalence and associations with self-reported hip disability and muscle strength. J Hip Preserv Surg 2018b; 5 (1): 39-46. Khan O H, Malviya A, Subramanian P, Agolley D, Witt J D. Minimally invasive periacetabular osteotomy using a modified Smith–Petersen approach. Bone Joint J 2017; 99-B (1): 22-8. Klit J, Hartig-Andreasen C, Jacobsen S, Søballe K, Troelsen A. Periacetabular osteotomy: sporting, social and sexual activity 9–12 years post surgery. Hip Int 2014; 24 (1): 27-31. van Laarhoven S N, Lipperts M, Bolink S A A N, Senden R, Heyligers I C, Grimm B. Validation of a novel activity monitor in impaired, slow-walking, crutch-supported patients. Ann Phys Rehabil Med 2016; 59 (5-6): 308-13. Lerch T D, Steppacher S D, Liechti E F, Tannast M, Siebenrock K A. Onethird of hips after periacetabular osteotomy survive 30 years with good clinical results, no progression of arthritis, or conversion to THA. Clin Orthop Relat Res 2017; 475 (4): 1154-68. Lipperts M, van Laarhoven S, Senden R, Heyligers I, Grimm B. Clinical validation of a body-fixed 3D accelerometer and algorithm for activity monitoring in orthopaedic patients. J Orthop Transl 2017; 11: 19-29. Novais E N, Heyworth B, Murray K, Johnson V M, Kim Y-J, Millis M B. Physical activity level improves after periacetabular osteotomy for the treatment of symptomatic hip dysplasia. Clin Orthop Relat Res 2013; 471 (3): 981-8. Nunley R M, Prather H, Hunt D, Schoenecker P L, Clohisy J C. Clinical presentation of symptomatic acetabular dysplasia in skeletally mature patients. J Bone Joint Surg Am 2011; 93 (Suppl 2): 17-21. Sallis J F, Saelens B E. Assessment of physical activity by self-report: status, limitations, and future directions. Res Q Exerc Sport 2000; 71 (Suppl 2): 1-14. Sechriest II V F, Kyle R F, Marek D J, Spates J D, Saleh K J, Kuskowski M. Activity level in young patients with primary total hip arthroplasty: a 5-year minimum follow-up. J Arthroplasty 2007; 22 (1): 39-47. Smuck M, Tomkins-Lane C, Ith M A, Jarosz R, Kao M-C J. Physical performance analysis: a new approach to assessing free-living physical activity in musculoskeletal pain and mobility-limited populations. PLoS One 2017; 12 (2): e0172804. Thomeé R, Jónasson P, Thorborg K, Sansone M, Ahldén M, Thomeé C, Karlsson J, Baranto A. Cross-cultural adaptation to Swedish and validation of the Copenhagen Hip and Groin Outcome Score (HAGOS) for pain, symptoms and physical function in patients with hip and groin disability due to femoro-acetabular impingement. Knee Surgery, Sport Traumatol Arthrosc 2014; 22 (4): 835-42. Thorborg K, Hölmich P, Christensen R, Petersen J, Roos E M. The Copenhagen Hip and Groin Outcome Score (HAGOS): development and validation according to the COSMIN checklist. Br J Sports Med 2011; 45 (6): 478-91. Troelsen A, Elmengaard B, Søballe K. A new minimally invasive transsartorial approach for periacetabular osteotomy. J Bone Joint Surg Am 2008; 90 (3): 493-8. Tudor-Locke C, Craig C L, Aoyagi Y, Bell R C, Croteau K A, De Bourdeaudhuij I, Ewald B, Gardner A W, Hatano Y, Lutes L D, Matsudo S M, Ramirez-Marrero F A, Rogers L Q, Rowe D A, Schmidt M D, Tully M A, Blair S N. How many steps/day are enough? For older adults and special populations. Int J Behav Nutr Phys Act 2011; 8 (1): 80. Wiberg G. Studies on dysplastic acetabula and congenital subluxation of the hip joint. Acta Orthop Scand 1939; (Suppl 58): 1-132.

628

Acta Orthopaedica 2018; 89 (6): 628–633

Validated repeatability of patient-reported outcome measures following primary total hip replacement: a mode of delivery comparison study with randomized sequencing Charlotte V E CARPENTER 1, Julia BLACKBURN 2, John JACKSON 3, Ashley W BLOM 1, Adrian SAYERS 1, Michael R WHITEHOUSE 1 1 Musculoskeletal

Research Unit, Bristol Medical School, 1st Floor Learning & Research Building, Southmead Hospital, Bristol; 2 Department of Trauma and Orthopaedics, Bristol Royal Infirmary, Upper Maudlin St, Bristol; 3 Department of Trauma and Orthopaedics, North Devon District Hospital, Raleigh Park, Barnstaple, UK Correspondence: adrian.sayers@bristol.ac.uk Submitted 2017-10-09. Accepted 2018-08-19.

Background and purpose — Patient-reported outcome measures (PROMs) are used to understand better the outcomes after total hip replacement (THR). These are administered in different settings using a variety of methods. We investigated whether the mode of delivery of commonly used PROMs affects the reported scores, 1 year after THR. Patients and methods — A prospective test–retest mode comparison study with randomized sequence was done in 66 patients who had undergone primary THR. PROMs were administered by 4 modes: self-administration, face-toface interview, telephone interview, and postal questionnaire. PROMs included: Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), Oxford Hip Score (OHS), EQ5D-3L (EQ5D), and Self-Administered Patient Satisfaction Scale (SAPS). Linear regression was used to estimate relationships between the mean scores for PROMs by mode. Individual paired differences by mode were calculated, relationships between modes were identified, and results adjusted by time delay and participant age. Results — There was no statistically significant difference between the mean PROM scores recorded for each mode of delivery for each score. Statistically significant differences in the individual paired differences were detected between modes for the WOMAC stiffness subscale, OHS, EQ5D, and SAPS. OHS difference in individual paired means between face-to-face and telephone interview exceeded the minimal clinically important difference. Interpretation — PROMs mode of administration can affect the recorded results. Modes should not be mixed and may not be comparable between studies. It should not be assumed that different modes will obtain the same results and where not already established this should be checked by researchers before use.

There are various metrics used to judge the success or failure of total hip replacement (THR). Hard endpoints such as revision of the THR and mortality are popular as they are easy to define, but such outcomes fail to take account of the degree of relief of symptoms experienced by the patient, i.e., soft endpoints (Wylde and Blom 2011). To better understand the outcomes after THR, patient-reported outcome measures (PROMs) have been widely adopted, and typically these PROMs are focused around domains such as pain, function, and stiffness. Their use has become routine and widespread, for example, the UK Department of Health’s National PROMs program (http:// www.hscic.gov.uk/proms) administers PROMs prior to and 6 months after intervention for procedures such as THR, total knee replacement, hernia repair, and varicose vein surgery. PROMs questionnaires are administered and completed in a number of different settings using a variety of methods (Tourangeau et al. 2000). Common modes of delivery include paper based, face-to-face, telephone, and computer delivered, with responses being self-recorded or assisted by a third party. With the evolution of technology, the boundaries between modes of delivery are now becoming blurred. In addition, preoperative and postoperative assessments are frequently performed using different modes of administration, and in many research studies a mixture of modes is used to ensure data completeness (Dillman et al. 2009). When modes of delivery are mixed, it is important to understand whether the mode of delivery of the questionnaire affects the psychometric properties of the score (Honaker 1988). If different modes of delivery result in scores that are not equivalent, then these modes should not be mixed in a single study design. Factors that may be associated with the magnitude of difference include context, content, and the population studied (Hood et al. 2012). Systematic review of mode comparisons

Acta Orthopaedica 2018; 89 (6): 628–633

Table 1. Randomization sequence

Invited participants n = 135 Refused n = 69 Number recruited n = 66 Lost to follow up n=0 Data analysed n = 66

629

Group 1

Time point 1 in clinic 1. Self-administered 2. Face-to-face Time point 2 Telephone interview Time point 3 Postal

Group 2

Group 3

Group 4

1. Face-to-face 2. Self-administered

1. Self-administered 2. Face-to-face

1. Face-to-face 2. Self-administered

Postal

Telephone interview

Postal

Study flowchart.

has shown that modes are vulnerable to bias when comparison is made between an interviewer being involved and selfcompletion (Hood et al. 2012). In a large population study, telephone administration yielded more positive health-related quality of life estimates than self-administration (Hanmer et al. 2007). However, multiple item scales are less prone to bias and differences between modes have ameliorated as technologies such as telephone- and computer-based completion have become commonplace (Gwaltney et al. 2008). We investigated whether the mode of questionnaire delivery influences test scores in commonly used PROMS in primary THR. Study design — Patients were invited to participate in a test– retest study of 4 PROMs using 4 modes of delivery, 1 year following THR, using a randomized crossover design.

Patients and methods A prospective mode comparison cohort study was conducted in a single NHS tertiary orthopedic center. In order to ensure that patients had reached a steady state in terms of their outcome following surgery, patients who were 1 year following THR were invited to participate (Lenguerrand et al. 2016). Patients were eligible for inclusion in the study if they had undergone primary THR for any indication 1 year previously. Exclusion criteria were patients who had undergone revision THR, patients who were unwilling or unable to provide informed consent, and patients who were unable to understand or complete questionnaires in English (Study flow chart, Figure). Patients were recruited to the study by written invitation sent 1 week before their outpatient clinic appointment. Recruitment occurred between June 2014 and October 2015. 66 patients, who indicated they wished to participate, were consented and randomized to the order in which they would receive the questionnaires. Participants were asked to complete a set of 4 questionnaires: the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC); the EQ5D-3L health questionnaire (EQ5D); the Self-Administered Patient Satisfaction Scale (SAPS); and the Oxford Hip Score (OHS). 4 modes of ques-

tionnaire completion were used: self-administered in clinic; face-to-face interview in clinic; telephone interview; and postal questionnaire. For full details of PROMS questionnaires, questions, and structure of questionnaires please see Appendix. The sets of PROMs were delivered by 4 modes: selfadministered in clinic and face-to-face interviewer led, both completed during the outpatient clinic appointment; and later via telephone interview and self-administered by post. Participants completing PROMs self-administered in clinic were asked to complete the set of questionnaires using pen and paper, without assistance. During the face-to-face interview, each question in the set was asked by a member of the research team and the questionnaires were completed by the researcher based on the verbal responses of the participant. Randomization The participants were randomized to the sequence in which mode of completion was done (Table 1). Participants were randomized in permuted blocks to 1 of 4 groups with 3 times block size. Statistics Sample size A sample size of 60 was calculated, based on the OHS. The minimal clinically important difference (MCID) for the OHS is 5 (Beard et al. 2015). Therefore, for 80% power and a 2-sided 5% significance, a sample size of 52 is required to allow the study to detect if there is a significant difference between the scores by different modes. This was rounded up to 60 participants to account for loss to follow-up. Missing values Missing values were identified before statistical analysis. For calculation of the PROM scores, missing values were dealt with according to the user guides for each score. For WOMAC, if 2 or more pain, both stiffness, or 4 or more physical function items were omitted, the participants’ responses were deemed invalid and the deficient subscale was not used for analysis. According to OHS guidance, if more than 2 questions were unanswered, a score was not calculated. If 1 or 2 questions were unanswered, the mean value of other responses

630

was substituted for the missing value. An EQ5D index was calculated from the EQ5D-3L scores. EQ5D questionnaires with missing values were excluded. There is no specific way of dealing with missing data for SAPS; however, there were no missing values. Analysis There are 2 possible methods to analyze mode comparison studies. The first is to assume independence on each occasion a questionnaire is completed. This analytical standpoint is typically described as investigation of between-population differences. In this case, linear regression was used to estimate the between-population mean difference of each questionnaire by delivery mode of delivery. The second method of analysis investigates within-individual (paired) differences of each questionnaire by mode of delivery. In this case, the paired difference between each mode was calculated and linear regression was used to estimate any within individual differences between modes of delivery. Within-patient analyses were further adjusted for the time between completion of questionnaires (time delay) and age of the participant, i.e., < 70 or ≥ 70 years and old. P-values are reported without adjustment for multiple testing (Perneger 1998). Data were analyzed using STATA (version 14, StataCorp, College Station, TX, USA). Ethics, registration, funding, and potential conflicts of interest Ethical approval was granted for the study by the West of Scotland Research Ethics Service on April 24, 2014 (Ref: 14/WS/0062). This study was supported by the NIHR Biomedical Research Centre at the University Hospitals Bristol NHS Foundation Trust and the University of Bristol. The views expressed in this publication are those of the author(s) and not necessarily those of the NHS, the National Institute for Health Research, or the Department of Health. Adrian Sayers is funded by an MRC Strategic Skills Fellowship (MR/ L01226X/). No competing interests are declared.

Results 66 participants (median age 69 (IQR 62–77), 33 male) consented to participate (Figure). Missing data Missing data were observed for each of the WOMAC, OHS, and EQ5D measures. There were no incomplete SAPS questionnaires. More missing data were observed for the OHS than for any of the other scores (Table 2, see Supplementary data). This resulted in the exclusion of 5 OHS questionnaires, 3 of which were when the questionnaire was self-administered in clinic. In comparison, the WOMAC score was insufficiently completed for the inclusion of all 3 subscales in 3 cases and 1 EQ5D questionnaire was excluded. Scores were left unan-

Acta Orthopaedica 2018; 89 (6): 628–633

swered most frequently when self-administered in clinic (29 of 2,970 data points). Between-population mode differences There were no statistically significant differences among mean scores by mode of delivery for any PROM investigated (Table 3, see Supplementary data). Timing between administration There was no difference between administrations 1 and 2, as these were undertaken on the same day, before and after an outpatient appointment. The median time difference between administration 2 and 3 was 7 (IQR 3–8) days and 6 (IQR 3–7) between timepoints 3 and 4. The median time between administrations 1 and 4 was 14 (IQR 8–17) days. Within-individual (paired) mode differences WOMAC subscales When the WOMAC subscales were considered, there was a difference in the individual paired differences observed for the stiffness subscale between the modes of delivery (Table 4). This persisted when adjustment was made for age and the time delay between modes of delivery. Young patients were more likely to give a higher (worse) score when the score was completed in clinic by face-to-face interview or self-administered than when the score was completed by postal or telephone modes. The WOMAC function subscale revealed a similar pattern with higher (worse) scores given when the score was completed in clinic by face-to-face interview or self-administered and when the form was self-administered and delivered by post compared with telephone interview, but this difference disappeared when adjusting for the time delay between modes and age. OHS The individual paired differences between the OHS scores for different modes showed a statistically significant difference between postal and telephone scores when unadjusted and adjusted for time delay. However, this may not be clinically relevant as the difference is below the MCID of 5. When the OHS was adjusted for age and the time delay between modes of delivery, lower (worse) scores were given when the form was completed in clinic by face-to-face interview (–7.35, 95% CI –11 to –4) or self-administered by post (–1.24, 95% CI –2.4 to –0.07) compared with telephone interview completion. EQ5D When the EQ5D was considered, no differences in the individual paired differences were seen. However, when adjusted for time delay between questionnaires, higher (better) scores were seen in the telephone and postal groups compared with the other modes of delivery. When adjusting for the age of the respondents, this difference did not persist. When age

Acta Orthopaedica 2018; 89 (6): 628–633

631

Table 4. Individual paired differences by mode F2F—SI F2F—P F2F—T Model dif. SE 95% CI p-value dif. SE 95%CI p-value dif. SE 95% CI p-value WOMAC Stiffness 1 0 0.60 2.1 (–3.6 to 4.8) 0.8 5.95 1.8 (2.3 to 9.6) 0.002 2 0 –0.83 1.8 (–4.3 to 2.7) 0.6 5.28 1.6 (2.0 to 8.6) 0.002 3 0 13.5 5.6 (2.2 to 25) 0.02 5.87 2.5 (0.9 to 11) 0.02 WOMAC Pain 1 0 1.59 1.0 (–0.36 to 3.5) 0.1 1.61 1.0 (–0.35 to 3.5) 0.1 2 0 1.31 1.0 (–0.66 to 3.3) 0.2 1.78 1.3 (–0.77 to 4.3) 0.2 3 0 4.91 3.3 (–1.7 to 12) 0.1 1.66 2.0 (–2.3 to 5.6) 0.4 WOMAC Function 1 0 0.81 1.6 (–2.4 to4.1) 0.6 3.15 1.5 (0.21 to 6.1) 0.04 2 0 –0.07 1.4 (–3.0 to 2.8) 1 2.69 1.4 (–0.06 to 5.4) 0.06 3 0 0.54 4.9 (–9.2 to 10) 0.9 2.72 2.1 (–1.5 to 6.9) 0.2 OHS 1 –1.41 1.6 (–4.7 to 1.8) 0.4 –1.18 1.2 (–3.5 to 1.2) 0.3 –2.20 1.1 (–4.5 to 0.08) 0.06 2 –1.41 1.6 (–4.6 to 1.8) 0.4 –1.09 1.2 (–3.5 to 1.3) 0.4 –2.26 1.3 (–4.8 to 0.24) 0.08 3 –1.40 1.6 (–4.6 to 1.8) 0.4 5.34 3.9 (–2.5 to 13) 0.2 –7.35 1.7 (–11 to –4) <0.001 EQ5D index 1 0 –0.01 0.02 (–0.06 to 0.03) 0.6 –0.04 0.02 (–0.09 to 0.01) 0.1 2 0 –0.02 0.02 (–0.06 to 0.01) 0.2 –0.05 0.03 (–0.1 to –0.001) 0.05 3 0 –0.03 0.07 (–0.17 to 0.10) 0.6 –0.05 0.04 (–0.13 to 0.03) 0.2 SAPS 1 4.29 1.8 (0.80 to 7.8) 0.02 0.20 1.4 (–2.6 to 3.0) 0.9 –0.48 1.4 (–3.2 – 2.3) 0.7 2 4.30 1.7 (0.80 to 7.8) 0.02 –0.10 1.4 (–2.9 to 2.7) 0.9 0.11 1.5 (–2.8 to 3.0) 0.9 3 4.26 17 (0.77 to 7.7) 0.02 –4.26 4.7 (–14 to 5.2) 0.4 –2.10 2.2 (–6.5 to 2.3) 0.4 Model 1 = Individual paired differences (IPD), Model 2 = IPD + time delay, Model 3= IPD + time delay + age. F2F: face to face in clinic; SI: Self-administered in clinic; P: postal; T: telephone interview. Dif: Individual paired difference: 95% CI: 95% confidence interval.

Table 2. Continued SI—P SI—T P—T Model dif. SE 95% CI p-value dif. SE 95%CI p-value dif. SE 95% CI p-value WOMAC Stiffness 1 0.60 2.1 2 –0.83 1.8 3 –4.19 2.1 WOMAC Pain 1 1.58 1.0 2 1.31 1.0 3 0.45 1.2 WOMAC Function 1 0.81 1.6 2 –0.07 1.4 3 –0.24 1.8 OHS 1 0.96 1.0 2 1.06 1.0 3 0.96 1.3 EQ5D index 1 –0.01 0.02 2 –0.02 0.02 3 –0.02 0.03 SAPS 1 –3.77 0.93 2 –3.69 0.92 3 –4.10 1.2

(–3.6 to 4.8) (–4.3 to 2.7) (–8.4 to 0.001)

0.8 0.6 0.05

5.95 5.28 5.87

1.8 1.6 2.5

(2.3 to 9.6) (2.0 to 8.6) (0.9–11)

0.002 0.002 0.002

4.62 1.5 4.94 1.6 3.04 2.1

(1.7 to 7.6) (1.8 to 8.1) (–1.1 to 7.2)

0.003 0.003 0.1

(–0.36 to 3.5) (–0.66 to 3.3) (–2.0 to 3.0)

0.1 0.2 0.7

1.62 1.78 1.66

1.2 1.3 2.0

(–0.73 to 4.0) (–0.77 to 4.3) (–2.3 to 5.6)

0.2 0.2 0.4

0.71 0.8 0.42 0.8 0.30 1.1

(–0.91 to 2.3) (–1.2 to 2.0) (–1.9 to 2.5)

0.4 0.6 0.8

(–2.4 to 4.1) (–3.0 to 2.8) (–3.9 to 3.4)

0.6 1 0.9

3.15 2.69 2.72

1.5 1.4 2.1

(0.21 to 6.1) (–0.06 to 5.4) (–1.5 to 6.9)

0.04 0.06 0.2

2.61 1.3 2.62 1.3 2.11 1.8

(0.11 to 5.1) (–0.02 to 5.3) (–1.1 to 5.6)

0.04 0.05 0.2

(–1.1 to 3.0) (–1.0 to 3.1) (–1.6 to 3.6)

0.4 0.2 0.5

–0.69 –0.63 –0.29

1.0 1.0 1.7

(–2.6 to 1.3) (–2.8 to 1.5) (–3.6 to 3.1)

0.5 0.6 0.9

(–0.06 to 0.03) 0.6 (–0.06 to 0.01) <0.001 (–0.07 to 0.03) 0.4

–0.04 –0.05 –0.05

0.02 (–0.09 to 0.01) 0.1 0.03 (–0.10 to –0.001) 0.05 0.04 (–0.13 to 0.03) 0.2

–0.02 0.02 (–0.06 to 0.02) –0.02 0.02 (–0.07 to 0.02) 0.00 0.03 (–0.06 to 0.06)

0.5 0.3 1.0

(–5.6 to –1.9) (–5.5 to –1.9) (–6.4 to –1.8)

–4.90 –4.77 –5.28

0.95 (–6.8 to –3.0) 1.0 (–6.8 to –2.7) 1.5 (–8.3 to –2.2)

–0.30 0.57 (–1.4 to 0.84) –0.21 0.60 (–1.4 to 1.0) 0.09 0.81 (–1.5 to 1.7)

0.6 0.7 0.9

<0.001 0.002 0.001

<0.001 <0.001 0.001

–1.71 0.44 (–2.6 to –0.83) <0.001 –1.81 0.46 (–2.7 to –0.88) <0.001 –1.24 0.59 (–2.4 to –0.07) 0.04

Model 1 = Individual paired differences (IPD), Model 2 = IPD + time delay, Model 3= IPD + time delay + age. F2F: face to face in clinic; SI: Self-administered in clinic; P: postal; T: telephone interview. Dif: Individual paired difference: 95% CI: 95% confidence interval.

632

was categorized into young (< 70) or old (≥ 70), the difference persisted in the older group, suggesting older patients are more likely to report a higher (better) score for the EQ5D when completed by telephone interview or self-administered and delivered by post. SAPS There was a statistically significant reduction in the individual paired differences for the SAPS between self-administration in clinic and face-to-face interview, and between self-administration in clinic and delivery by post or telephone interview. This difference persisted when adjusting for time delay between modes and age. Each point on the SAPS scale has a value of 6.25%. The difference between groups was roughly 4%, thus the effect was less than 1 point in 1 response, and of unlikely clinical significance.

Discussion We investigated whether the mode of delivery of commonly used PROMs affected the results reported by patients who had undergone primary total hip replacement. There were no statistically significant between population differences among mean scores in any PROMs investigated. We observed more missing data for the OHS than other PROMs. This included 3 OHS questionnaires that had no responses completed. The OHS was the last questionnaire in the series. The higher missing data count in this score may be due to survey fatigue, or perhaps, most simply, that patients omitted the last page of the questionnaire (Porter et al. 2004). Most missing data were seen in the self-administered in clinic group across all PROMs. In a systematic review of mode of administration of surveys, Bowling (2005) described 13 sources of bias that may affect the results obtained by different modes. That study concluded that non-response is likely to be influenced by mode of administration, with a higher non-response reported in postal than face-to-face and postal than telephone, suggesting that premature termination is less likely in the presence of a motivating interviewer. This finding is echoed by Wood and McLauchlan (2006) and by Fitzpatrick et al. (2000) in the response rates to the OHS, with highest responses achieved by face-to-face and selfadministered questionnaires and lowest with postal responses. Both studies reported that question 6 of the OHS (“In the past 4 weeks, for how long have you been able to walk before pain from your hip becomes severe (with or without a stick)?”) was the one most frequently left unanswered. WOMAC subscales did not reveal any statistically significant difference in mean scores across mode of delivery. When adjusted for time delay and the age of the participant, young patients showed a small propensity to worse scores on the WOMAC stiffness subscale when the score was completed in

Acta Orthopaedica 2018; 89 (6): 628–633

clinic by face-to-face interview or self-administered. This difference was between 5% and 13%, below the MCID of 25% and is therefore not likely to be clinically significant (Quintana et al. 2005). No statistically significant differences were seen in WOMAC pain scales by any mode of delivery. These findings are similar to those of Bellamy et al. (2002), which showed no difference between telephone and onsite administration for the WOMAC knee score and electronic versus paper surveys for patients with hip and knee OA in 2002. OHS gave a higher (better) score for telephone than postal scores or face-to-face interviews, with a difference between face-to-face and telephone of 7 points, which is in excess of the MCID and therefore could be clinically significant (Murray et al. 2007). Older participants may give better scores for EQ5D if recorded by post or telephone interview. However, quantifying the magnitude is difficult with the EQ5D index. A small reduction in satisfaction was seen when SAPS was completed by self-administered in clinic when compared with face-to-face interview and telephone interviews and postal responses. These findings suggest a small propensity to better score responses in telephone questionnaires than other modes. Telephone interviews may be subject to biases including social desirability bias, yes-saying, and interviewer bias (Bowling 2005). Indeed, it has been shown that satisfaction may be more positive if surveys are presented aurally than visually (Dillman et al. 2009). Health-related quality of life scores have been shown to be consistent when mode of administration was the same but telephone administration of EQ5D yields more positive results than self-administered in an older group with both US and UK weighting (Hanmer et al. 2007). Hays et al. (2009) found that the maximum effect size between postal versus telephone administration of the EQ5D was 0.5. Wood and McLauchlan (2006) found no difference in mode of administration between postal delivery and interview of the OHS at 10-year follow-up after THR. The OHS showed only a small increase in mode effect average, 1.2, in telephone versus postal administration, which was not deemed clinically relevant (Messih et al. 2014). However, in a meta-analysis of mode of administration of PROMS, self-completion and assisted completion produced equivalent scores overall but results were influenced by the setting in which questionnaires were completed (Rutherford et al. 2016). We found that mode of administration of PROMs 1 year after THR may have small effects on the results obtained. Participants in our study were randomized according to the order of modes; therefore, we do not believe practice effects are likely to explain these differences. Telephone and postal responses were collected following self-administration in clinic and face-to-face interview in clinic. As such, a time delay between these sets was introduced. This had an effect in some cases but was adjusted for within the analysis. Although participants were encouraged to complete the self-administered PROMs themselves, whether they received help or support from friends, family members, or carers was not docu-

Acta Orthopaedica 2018; 89 (6): 628–633

mented. Interviewer-led PROMs were completed by several different researchers, which may introduce bias to the answers obtained. Comorbidity and the indication for THR were not investigated as part of this study. Our findings may be generalizable to patients who have undergone THR but not those awaiting THR and only apply to English versions of these PROMs. We did not investigate electronic modes of delivery of these PROMs, such as via mobile phone, hand-held devices in clinic and automated telephone responses. In an increasingly digital age, these modes of delivery are increasingly used and the effects are as yet unknown in this population. The small variations in responses to PROMs delivered by different modes of administration are not likely to have a significant impact on the results in smaller studies. However, in large longitudinal studies where the timing of questionnaires may vary, these small biases may induce statistically significant chance findings. To ensure bias is minimized in studies using PROMs assessment after THR, we recommend that modes of administration are, wherever possible, not mixed in a single study. If multiple modes are used it is important to distinguish between modes of administration, and avoiding mixing self-administered and interviewer-led PROMs may minimize the effects. When outcomes are collected by different modes of administrations this should be acknowledged, and care should be used when interpreting results. Whilst using different modes in the same study may be useful in minimizing missing data in clinical studies, it is important to recognize it is not a panacea, and the primary response is still missing. Supplementary data Appendix and Tables 2 and 3 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/ 17453674.2018.1521183

MRW: Conception and design of the study. JB and AS: Design and acquisition of the data. JB, and JJ: Acquisition of data. CC, and AS: Analysis of data. All authors interpreted data and wrote the report. CC, JB, AS, MW contributed equally to this work The following people contributed to the paper through the acquisition of data: Christopher Woodrow, Harriet Mitchell, Sophie Stanger, Samantha Dixon, Nicolas Toosi, and Charlotte Howie. Acta thanks Sarah Whitehouse and Nienke Wolterbeek for help with peer review of this study

633

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. Bellamy N, Campbell J, Hill J, Band P. A comparative study of telephone versus onsite completion of the WOMAC 3.0 osteoarthritis index. J Rheumatol 2002; 29(4): 783-6. Bowling A. Mode of questionnaire administration can have serious effects on data quality. J Public Health 2005; 27(3): 281-91. Dillman D A, Phelps G, Tortora R, Swift K, Kohrell J, Berck J, Messer B L. Response rate and measurement differences in mixed-mode surveys using mail, telephone, interactive voice response(IVR) and the Internet. Social Sci Res 2009; 38(1): 1-18. Fitzpatrick R, Morris R, Hajat S, Reeves B, Murray D W, Hannen D, Rigge M, Williams O, Gregg P. The value of short and simple measures to assess outcomes for patients of total hip replacement surgery. Quality Saf Health Care 2000; 9(3): 146-50. Gwaltney C J, Shields A L, Shiffman S. Equivalence of electronic and paperand-pencil administration of patient-reported outcome measures: a metaanalytic review. Value Health 2008; 11(2): 322-33. Hanmer J, Hays R D, Fryback D G. Mode of administration is important in US national estimates of health-related quality of life. Med Care 2007; 45(12): 1171-9. Hays R D, Kim S, Spritzer K L, Kaplan R M, Tally S, Feeny D, Liu H, Fryback D G. Effects of mode and order of administration on generic healthrelated quality of life scores. Value Health 2009; 12(6): 1035-9. Honaker L M. The equivalency of computerized and conventional MMPI administration: a critical review. Clin Psychol Rev 1988; 8(6): 561-77. Hood K, Robling M, Ingledew D, Gillespie D, Greene G, Ivins R, Russell I, Sayers A, Shaw C, Williams J. Mode of data elicitation, acquisition and response to surveys: a systematic review. Southampton: NIHR Health Technology Assessment programme: Executive Summaries; 2012. Lenguerrand E, Wylde V, Gooberman-Hill R, Sayers A, Brunton L, Beswick A D, Dieppe P, Blom A W. Trajectories of pain and function after primary hip and knee arthroplasty: the ADAPT cohort study. PloS One 2016; 11(2): e0149306. Messih M A, Naylor J, Descallar J, Manickam A, Mittal R, Harris I. Mail versus telephone administration of the Oxford knee and hip scores. J Arthroplasty 2014; 29(3): 491-4. Murray D, Fitzpatrick R, Rogers K, Pandit H, Beard D, Carr A, Dawson J. The use of the Oxford hip and knee scores. J Bone Joint Surg (Br) 2007; 89-B(8): 1010-14. Perneger T V. What’s wrong with Bonferroni adjustments. BMJ 1998; 18,316(7139): 1236-8. Porter S R, Whitcomb M E, Weitzer W H. Multiple surveys of students and survey fatigue. New Directions for Institutional Research 2004; 2004(121): 63-73. Quintana J, Escobar A, Bilbao A, Arostegui I, Lafuente I, Vidaurreta I. Responsiveness and clinically important differences for the WOMAC and SF-36 after hip joint replacement. Osteoarthritis Cartilage 2005; 13(12): 1076-83. Rutherford C, Costa D, Mercieca-Bebber R, Rice H, Gabb L, King M. Mode of administration does not cause bias in patient-reported outcome results: a meta-analysis. Qual Life Res 2016; 25(3): 559-74. Tourangeau R, Rips L J, Rasinski K. The psychology of survey response. Cambridge: Cambridge University Press; 2000. Wood G, McLauchlan G. Outcome assessment in the elderly after total hip arthroplasty. J Arthroplasty 2006; 21(3): 398-404. Wylde V, Blom A. The failure of survivorship. J Bone Joint Surg (Br) 2011; 93-B(5): 569-70.

634

18F-FDG-PET

Acta Orthopaedica 2018; 89 (6): 634–639

uptake in non-infected total hip prostheses

Stefan J GELDERMAN 1, Paul C JUTTE 2, Ronald BOELLAARD 1, Joris J W PLOEGMAKERS 2 , David VÁLLEZ GARCÍA 1, Greetje A KAMPINGA 3, Andor W J M GLAUDEMANS 1, and Marjan WOUTHUYZEN-BAKKER 3 1 Medical

Imaging Center, Department of Nuclear Medicine and Molecular Imaging, University of Groningen, University Medical Center Groningen, Groningen; 2 Department of Orthopaedic Surgery, University of Groningen, University Medical Center Groningen, Groningen; 3 Department of Medical Microbiology and Infection Prevention, University of Groningen, University Medical Center Groningen, Groningen, the Netherlands Correspondence: s.j.gelderman@umcg.nl Submitted 2018-08-19. Accepted 2018-08-23.

Background and purpose — 18F-fluorodeoxyglucose positron emission tomography (FDG-PET) can be used in the diagnostic work-up of a patient with suspected periprosthetic joint infection (PJI) but, due to a lack of accurate interpretation criteria, this technique is not routinely applied. Since the physiological uptake pattern of FDG around a joint prosthesis is not fully elucidated, we determined the physiological FDG uptake in non-infected total hip prostheses. Patients and methods — Patients treated with primary total hip arthroplasty (1995–2016) who underwent a FDG-PET/CT for an indication other than a suspected PJI were retrospectively evaluated. Scans were both visually and quantitatively analyzed. Semi-quantitative analysis was performed by calculating maximum and peak standardized uptake values (SUVmax and SUVpeak) by volume of interests (VOIs) at 8 different locations around the prosthesis. Results — 58 scans from 30 patients were analyzed. In most hips, a diffuse heterogeneous uptake pattern around the prosthesis was observed (in 32/38 of the cemented prostheses, and in 16/20 of the uncemented prostheses) and most uptake was located around the neck of the prosthesis. The median SUVmax in the cemented group was 2.66 (95% CI 2.51–3.10) and in the uncemented group 2.87 (CI 2.65–4.63) (Median difference = –0.36 [CI –1.2 to 0.34]). In uncemented prostheses, there was a positive correlation in time between the age of the prosthesis and the FDG uptake (rs = 0.63 [CI 0.26–0.84]). Interpretation — Our study provides key data to develop accurate interpretation criteria to differentiate between physiological uptake and infection in patients with a prosthetic joint.

Diagnosing a periprosthetic joint infection (PJI) can be a clinical challenge, especially for chronic and low-grade infections. The preoperative diagnostic work-up, e.g. serum C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), synovial culture, and several synovial biomarkers can be false negative or false positive (Hozack et al. 1991, Ivanèeviae et al. 2002, Koh et al. 2017). The currently applied diagnostic criteria as described by the Musculoskeletal Infection Society (MSIS) to diagnose a PJI do not entail advanced nuclear imaging (Parvizi et al. 2011). However, given the difficulty of the diagnosis of PJI and the rapidly evolving field of nuclear diagnostics, it may be beneficial to add this technology to the diagnostic possibilities. Several nuclear medicine imaging techniques are available to assist in diagnosing or excluding PJI, but the choice of which technique to use first depends mostly on local expertise, availability of techniques, and costs. The labelled white blood cell (WBC) scintigraphy is currently considered as the gold standard, because of its high specificity for infection (Jutte et al. 2014) but has its limitations. The technique is time-consuming for both patient and personnel, is not available in all hospitals, and patients with leukopenia are not suitable for the test due to a low labelling efficacy (Glaudemans et al. 2013, Palestro 2015). A good alternative for WBC scintigraphy could be 18F-fluorodeoxyglucose positron emission tomography (FDG-PET), which can be combined with computed tomography (CT) for exact anatomical localization (the CT part) of the accumulation of FDG (Boellaard et al. 2015). FDG-PET is widely available, more comfortable for the patient, provides better possibilities for quantification, and results in higher resolution images compared with WBC scintigraphy (Love et al. 2005, Verberne et al. 2016). However, an important limitation

Acta Orthopaedica 2018; 89 (6): 634–639

of FDG-PET/CT when compared with WBC scintigraphy is its lack of discrimination between infection and inflammation secondary to a physiological reaction to the foreign body (the implant). In addition, it is not known how long physiological FDG uptake around a prosthesis remains after primary arthroplasty (Delank et al. 2006, Glaudemans et al. 2013). Consequently, diagnostic interpretation criteria for declaring an FDG-PET/CT positive or negative for PJI are lacking and, as a consequence, FDG-PET/CT is not routinely performed for the diagnosis of PJI. Therefore, we determined the physiological FDG uptake in non-infected total hip prostheses.

Patients and methods Patients Data of patients who underwent a primary total hip arthroplasty between January 1995 and December 2016 and underwent an FDG-PET/CT scan in the time period after the primary arthroplasty for reasons other than a suspected PJI (mostly for oncological reasons) were collected retrospectively. Patients were included only if they had no clinical signs (e.g., pain at rest) of a PJI at the time the FDG-PET/CT scan was performed. The included patients were divided into 2 groups, cemented and uncemented. FDG-PET/CT imaging All FDG-PET/CT scans were performed with an EARL accredited scanner, Siemens Biograph MCT 64 or 40 slice (Knoxville, TN, USA). Patients were instructed to fast for at least 6 hours prior to the administration of FDG. Blood glucose levels were measured before injection of the FDG and had to be below 11 mmol/L. The acquisition of the scan started after a resting time of approximately 60 minutes after 3 MBq/ kg FDG administration. All scan acquisition parameters were according to existing guidelines from the European Association of Nuclear Medicine (EANM) (Boellaard et al. 2015). All analyses were performed on EARL reconstructed images. Image interpretation and analysis Both visual and semi-quantitative analysis was performed to evaluate the FDG uptake surrounding the prosthesis. For the visual analysis, FDG uptake was categorized in 4 categories: diffuse heterogeneous (uptake around the prosthesis with different intensities), diffuse homogeneous (relative constant uptake around the prosthesis), focal (1 spot with high intensity), and no uptake. In addition, when there was extension of FDG uptake in the soft tissues surrounding the prosthesis this was recorded. For the semi-quantitative analysis, volumes of interests (VOIs) were created in each scan at 8 different locations around the prosthesis (tip of the prosthesis, femur–prosthesis interface, greater and lesser trochanter, neck of the prosthesis, lateral and medial acetabulum, and interface between the

635

prosthesis and the median part of the acetabulum). From these VOIs the maximum and peak standardized uptake value (SUV max and SUVpeak, which are values for the amount of FDG uptake) were determined. The mean of the SUVmax and SUVpeak at the 8 regions was calculated, as a representative of the uptake of the whole prosthesis (meanSUVmax and meanSUVpeak). These values were corrected for background SUV, using the mean SUV (SUVmean) of the liver (meanSUVmax/liver ratio and meanSUVpeak/liver ratio). Finally, the correlation of the 4 variables between the age of the prosthesis and FDG uptake was determined. Statistics IBM SPSS statistics version 20 (IBM Corp, Armonk, NY, USA) was used for statistical analysis. Additionally, StatsDirect version 3 (https://www.statsdirect.co.uk/) was used for calculating the confidence intervals (CI) for both the difference in mean values and the correlation coefficient. Since in clinical practice a PJI can occur in any joint prosthesis, regardless of the age of the prosthesis, all values were considered practically relevant and included in the CI. Continuous variables were presented as means (SD) or (95% confidence intervals—CI). Not normally distributed data were presented with the median and interquartile ranges (IQR) or with the CI of the mean. Differences between groups were calculated with the Mann–Whitney U-test. Correlation between the age of the prosthesis and the SUV uptake was calculated with Spearman’s rho. Differences in uptake between the locations were calculated with Friedman’s ANOVA. In the case of a statistically significant result, Wilcoxon’s signed rank test was used for post hoc calculation. A p-value < 0.05 was considered significant, without correction for multiple comparisons. Effect size estimates (r) for Mann–Whitney U and Wilcoxon signed rank test was calculated as r = z ⁄ √N (Fritz et al. 2012), with z being the standardized test statistics and N the sample size. Effect sizes were interpreted as large (> 0.5), medium (0.3– 0.5), and small (< 0.3). Moreover, to overcome the problem of correlated observations, additional analyses were done with only the last available scan of each patient. Ethics, funding, and potential conflicts of interest Ethical permission was acquired for this study from the UMCG medical ethical committee on 21 November 2017 (number M17.221053). None of the authors has had funding that might pose a conflict of interest in connection with the submitted article.

Results Patients Finally, 33 patients, with 42 total hip prostheses and 58 FDGPET/CT scans, were included (Figure 1). The indication for 37

636

Acta Orthopaedica 2018; 89 (6): 634–639

Cemented total hip arthroplasty n = 1,909

FDG-PET/CT scan n = 18,735

Crossmatch n = 76 arthroplasties

Uncemented total hip arthroplasty n = 1,112

Crossmatch n = 35 arthroplasties

Excluded (n = 31): – scan prior to arthroplasty, 18 – revision before scan, 2 – suspected PJI, 2 – suspected looosening, 1 – polyarthritis, 3 – other reasons, 5

Excluded (n = 15): – scan prior to arthroplasty, 5 – revision before scan, 5 – suspected PJI, 1 – suspected looosening, 3 – other reasons, 3

Remaining n = 45 arthroplasties n = 58 scans

Remaining n = 20 arthroplasties n = 28 scans Excluded (n = 8): – prosthesis partially scanned, 8

Excluded (n = 20): – prosthesis partially scanned, 20 Analyzed cemented total hip prostheses n = 38 scans

Analyzed uncemented total hip prostheses n = 20 scans

Figure 1. Flow chart of the patient inclusion process.

Table 1. Visual FDG-uptake pattern in non-infected hip prostheses Factor Diffuse heterogeneous Diffuse homogeneous Focal No uptake

Cemented Uncemented n = 38 n = 20 32 16 2 2 2 0 2 2

FDG-PET/CT scans was because of a malignancy, the commonest being lung carcinoma. In the non-oncological group, scans were performed in patients with increased inflammatory parameters (n = 8) or other indications (n = 4). Similar FDG uptake was found between oncological (meanSUVmax = 2.57 [IQR 2.08–3.21]) and non-oncological indicated scans (meanSUVmax = 3.01 [IQR 2.32–3.95]). The cemented group consisted of 38 prostheses scanned in 21 patients (median age 73 years [IQR 66–79]). 19 of these prostheses were scanned once in 15 different patients. 7 cemented hip prostheses in 5 patients were scanned twice, with a median time between the 2 scans of 1 year (IQR 0.4– 3.9). 1 hip prosthesis was scanned 5 times within a duration of approximately 4 years. The uncemented group consisted of 20 scans from 12 patients (median age 52 years [IQR 50–67]). 9 patients (11 prostheses) were scanned once. 2 patients with 3 prostheses were scanned twice (median time between 2 scans: 0.2 years [IQR 0.1–0.2]), and 1 hip prosthesis in a different patient was scanned 3 times within a duration of 9 months.

Figure 2. Example of 18F-FDG uptake in a non-infected total hip prosthesis. (A) Coronal FDG-PET image showing FDG uptake around the prosthesis, most prominent at the lateral side of the collum; (B, C, D) Coronal fused FDG-PET/CT images at different slices showing the uptake around the prosthesis; (E) Transaxial fused FDG-PET/CT image showing prominent physiological uptake at the lateral side of the cup of the prosthesis.

Visual analysis Visually, 39/42 prostheses showed uptake of FDG around the prosthesis. Diffuse heterogeneous uptake (Figure 2) was the most common uptake pattern in both the cemented group (32/38) and the uncemented group (16/20) (Table 1). No FDG uptake was seen in the periprosthetic soft tissues surrounding the prosthesis, except for the soft tissue at the height of the greater trochanter.

Acta Orthopaedica 2018; 89 (6): 634–639

637

Mean SUVmax cemented prostheses

Mean SUVmax uncemented prostheses

Table 2. SUVmax measured at 8 different locations around the prosthesis

0 0

Years after index operation

Tip Shaft Greater trochanter Lesser trochanter Neck Acetabulum lateral Acetabulum median Acetabulum medial

Cemented group Uncemented group Median SUVmax (IQR) Median SUVmax (IQR) 2.0 (1.6–2.4) 0.8 (0.6–1.1) 3.3 (2.5–4.0) 2.8 (2.3–3.5) 4.2 (2.9–5.4) 3.1 (2.5–4.4) 2.0 (1.6–2.6) 2.7 (2.0–3.3)

2.2 (1.1–3.1) 1.1 (0.9–2.0) 2.2 (1.7–4.2) 2.6 (1.7–4.2) 4.6 (3.1–10.4) 3.2 (2.4–6.3) 1.9 (1.4–3.4) 3.2 (2.1–6.5)

Figure 3. The relationship between the mean SUVmax and the age of the prosthesis in cemented (left panel) and uncemented hip prostheses (right panel).

Semi-quantitative analysis The median value of the meanSUVmax in the cemented group was 2.7 (CI 2.5–3.1) and in the uncemented group 2.9 (CI 2.7– 4.6) (median difference –0.36 [CI –1.2 to 0.34]). Also, without correlated observations the groups were similar. Other variables did not differ statistically significantly between groups either: meanSUVpeak (r = –0.10, p = 0.4), meanSUVmax/liver ratio (r = –0.25, p = 0.06), meanSUVpeak/liver ratio (r = –0.26, p = 0.05). The background uptake (SUVmean) in the liver was for the cemented group 2.9 (IQR 2.6–3.3), for the uncemented group 2.7 (IQR 2.3–3.0). The age of the patient did not statistically significantly influence the FDG uptake in either the cemented group (rs = 0.20, p = 0.2) or the uncemented group (rs = 0.10, p = 0.7). The SUVmax at the 8 different locations around the prosthesis is given in Table 2. The highest uptake was observed around the neck of the prosthesis and was significantly different compared with the other locations in both the cemented (r = [range –0.57 to –0.87], p < 0.001) and the uncemented group (r = [range –0.74 to –0.88], p < 0.001). Furthermore, other locations with increased uptake were observed at the lateral and medial acetabulum and the greater and lesser trochanter. The tip of the prosthesis showed minimal uptake. In addition, minimal uptake was observed at the femur–prosthesis interface of uncemented hip prostheses. FDG uptake in this location was absent in 34/38 cemented hip prostheses. The uptake between the cemented and uncemented prostheses was significantly different at the femur–prosthesis interface location (r = –0.37, p = 0.01) (r = –0.29, p = 0.06 without correlated observations). FDG uptake at other locations was similar between the cemented and uncemented prostheses. Relation FDG uptake and age of the prosthesis The median time between the hip arthroplasty and the FDGPET/CT scan was 3.9 years (IQR 1.2–9.1) in the cemented group and 6.6 years (IQR 1.6–15) in the uncemented group. In the cemented group, no relation was found between FDG uptake and the age of the prosthesis. However, a moderate relationship was found between FDG uptake and the age of the prosthesis in the uncemented group (Figure 3). This rela-

tion was observed for all 4 SUV variables (meanSUVmax: rs = 0.63 [CI 0.26–0.84], p = 0.003; meanSUVpeak: rs = 0.61 [CI 0.24–0.83], p = 0.004; meanSUVmax/liver ratio: rs = 0.63 [CI 0.26–0.84], p = 0.003; meanSUVpeak/liver ratio: rs = 0.65 [CI 0.29–0.85], p = 0.002). This was also observed in the results without correlated observations (meanSUVmax: rs = 0.55, p = 0.04). The FDG uptake in the uncemented prostheses remains stable for the early years, but an increase was observed after approximately 13 years. In general, the cemented total hip prostheses showed a mild decline in FDG uptake in the first 5 years after implantation (rs = –0.43, p = 0.03 [rs = –0.56, p = 0.05 without correlated observations]). After 5 years, the cemented prostheses show a relatively stable FDG uptake pattern over time.

Discussion The aim of our study was to determine the physiological FDG uptake in non-infected total hip prostheses. We found a diffuse heterogeneous uptake pattern in most prostheses with median SUVmax values of around 2.5, with the highest uptake observed around the neck. FDG uptake was positively related with the age of uncemented prostheses, but this finding was not observed in cemented prostheses. Our data could aid in the development of accurate interpretation criteria for FDG-PET/ CT to diagnose a PJI. An interesting finding was the increase of FDG uptake with time in the uncemented hip prostheses, which was observed after approximately 13 years after the primary implant. This FDG increase may be due to the development of wear particles (third particle disease) and/or subtle aseptic loosening, although no differences in implant survival between cemented and uncemented total hip prosthesis are described in the literature (Abdulkarim et al. 2013). It is important to note that the uncemented prostheses older than 13 years of age all comprised one brand of prosthesis: Biomet Mallory Head, a fulllength hydroxyapatite-coated titanium stem combined with an uncemented titanium shelled cup. Therefore, we cannot rule

638

out that the prosthesis design itself, instead of the age of the prosthesis, resulted in higher FDG uptake. Unfortunately, we do not have data from this type of prosthesis at earlier time points. Nevertheless, imaging interpreters should be aware that FDG uptake in non-infected older prostheses of this type may increase in time and may lead to false-positive results regarding diagnosis of a PJI. Most studies that have been conducted so far on FDG uptake in hip prostheses evaluated visual uptake patterns associated with loosening or infection. In accordance with our findings, these studies concluded that physiological uptake is mostly seen around the neck of the prosthesis (Zhuang et al. 2002, Vanquickenborne et al. 2003), which theoretically may be explained by wear of components and an adverse tissue reaction. In addition, previous studies indicate that an infection is highly suspected if there is uptake at the femur–prosthesis interface. This finding is also in agreement with our data demonstrating that FDG uptake is almost absent in non-infected prostheses (Zhuang et al. 2001, Chacko et al. 2002, Manthey et al. 2002). Unlike previous studies, we also observed uptake at the lateral and medial acetabulum, which may be due to differences in the camera systems that were used. Previous studies used an FDG-PET without the addition of a CT scan. With the addition of a CT scan, higher resolution images are acquired and anatomical structures are better visualized. Currently, the interpretation criteria of FDG-PET developed by Reinartz et al. (2005) is most often applied. The authors conducted a SUVmean analysis on the location with the highest uptake found around the prostheses. In that study, the SUVmean was 3.3 (SD 1.6) in prostheses without an infection, 5.0 (SD 3.0) in prostheses with aseptic loosening, and 5.9 (SD 2.7) in infected prostheses. With this analysis, the SUVmean between aseptic loosening and an infected prosthesis was similar. Unlike Reinartz et al. (2005), we performed a mean SUV analysis at 8 locations around the prosthesis, which may improve its diagnostic accuracy. Future studies should address the SUV values of these different locations in prostheses with aseptic loosening and infection. This is the first study that compared the FDG uptake between cemented and uncemented hip prostheses and analyzed physiological FDG uptake over time. Furthermore, SUV analyses were performed at 8 different locations around the prostheses. With this number of locations, a more precise uptake around the prosthesis is noted than in previous studies and contributes to a better understanding of physiological FDG uptake. Future studies should focus on FDG uptake in septic and aseptic loosened total hip prostheses, to further develop adequate interpretation criteria for diagnosing PJI using FDG-PET/CT scan. In summary, our findings provide insight into the physiological FDG uptake in non-infected total hip prostheses. In both cemented and uncemented total hip prostheses a diffuse uptake pattern is present with most uptake seen around the neck of the prostheses and no uptake in the soft tissues surrounding the prostheses. In addition, the level of FDG uptake

Acta Orthopaedica 2018; 89 (6): 634–639

in non-infected uncemented total hip prostheses is influenced by the age and probably the type of prosthesis. These findings may aid in the development of accurate interpretation criteria for FDG-PET/CT to better differentiate between (physiological) reactive inflammation due to a foreign body and a PJI.

SG: including patients, analysis of the data, interpretation and presentation of the results, main author. AG, PJ, and MW-B: study design, writing of the manuscript and supervising SG in analyzing and presenting the data. RB: analyzing the data and critical review of the manuscript. JP and GK: critical review of the manuscript. DVG: statistical analysis. Acta thanks Zohar Keidar and Samantha Terry for help with peer review of this study.

Abdulkarim A, Ellanti P, Motterlini N, Fahey T, O’Byrne J M. Cemented versus uncemented fixation in total hip replacement: a systematic review and meta-analysis of randomized controlled trials. Orthop Rev 2013; 5(1): e8. Boellaard R, Delgado-Bolton R, Oyen WJ, Giammarile F, Tatsch K, Eschner W, Verzijlbergen F J, Barrington S F, Pike L V, Weber W A, Stroobants S, Delbeke D, Donohoe K J, Holbrook S, Graham M M, Testanera G, Hoekstra O S, Zijlstra J, Visser E, Hoekstra C J, Pruim J, Willemsen A, Arends B, Kotzerke J, Bockisch A, Beyer T, Chiti A, Krause B J. FDG PET/CT: EANM procedure guidelines for tumour imaging: version 2.0. Eur J Nucl Med Mol Imag 2015; 42(2): 328-54. Chacko T K, Zhuang H, Stevenson K, Moussavian B, Alavi A. The importance of the location of fluorodeoxyglucose uptake in periprosthetic infection in painful hip prostheses. Nucl Med Commun 2002; 23(9): 851-5. Delank K S, Schmidt M, Michael J P, Dietlein M, Schicha H, Eysel P. The implications of 18F-FDG PET for the diagnosis of endoprosthetic loosening and infection in hip and knee arthroplasty: results from a prospective, blinded study. BMC Muscoskel Disord 2006; 7(1): 20. Fritz C O, Morris P E, Richler J J. Effect size estimates: current use, calculations, and interpretation. J Exp Psychol Gen 2012; 141(1): 2. Glaudemans A W J M, Galli F, Pacilio M, Signore A. Leukocyte and bacteria imaging in prosthetic joint infection. Eur Cell Mater 2013; 25: 61-77. Hozack W J, Balderston R A, Rothman R H, Gluckman S J, Maslack M M, Booth R E Jr. Evaluation of the painful prosthetic joint: relative value of bone scan, sedimentation rate, and joint aspiration. J Arthroplasty 1991; 6(3): 237-44. Ivanèeviae V, Perka C, Hasart O, Sandrock D, Munz D L. Imaging of lowgrade bone infection with a technetium-99m labelled monoclonal antiNCA-90 Fab’ fragment in patients with previous joint surgery. Eur J Nucl Med 2002; 29: 547-51. Jutte P C, Lazzeri E, Sconfienza L M, Cassar-Pullicino V, Trampuz A, Petrosillo N, Signore A. Diagnostic flowcharts in osteomyelitis, spondylodiscitis and prosthetic joint infection. Q J Nucl Med Mol Imaging 2014; 58(1): 2-19. Koh I J, Han S B, In Y, Oh K J, Lee D H, Kim T K. The leucocyte esterase strip test has practical value for diagnosing periprosthetic joint infection after total knee arthroplasty: a multicenter study. J Arthroplasty 2017; 32(11): 3519-23. Love C, Tomas M B, Tronco G G, Palestro C J. FDG PET of infection and inflammation 1. Radiographics 2005; 25(5): 1357-68. Manthey N, Reinhard P, Moog F, Knesewitsch P, Hahn K, Tatsch K. The use of [18 F]fluorodeoxyglucose positron emission tomography to differentiate between synovitis, loosening and infection of hip and knee prostheses. Nucl Med Commun 2002; 23(7): 645-53. Palestro C J. Radionuclide imaging of osteomyelitis. Semin Nucl Med 2015; 45(1): 32-46.

Acta Orthopaedica 2018; 89 (6): 634â&#x20AC;&#x201C;639

Parvizi J, Zmistowski B, Berbari E F, Bauer T W, Springer B D, Della Valle C J, Garvin K L, Mont M A, Wongworawat M D, Zalavras C G. New definition for periprosthetic joint infection: from the Workgroup of the Musculoskeletal Infection Society. Clin Orthop Relat Res 2011; 469(11): 2992-4. Reinartz P, Mumme T, Hermanns B, Cremerius U, Wirtz D C, Schaefer W M, Niethard F, Buell U. Radionuclide imaging of the painful hip arthroplasty: positron-emission tomography versus triple-phase bone scanning. J Bone Joint Surg Br 2005; 87(4): 465-70. Vanquickenborne B, Maes A, Nuyts J, Van Acker F, Stuyck J, Mulier M, Verbruggen A, Mortelmans L. The value of 18FDG-PET for the detection of infected hip prosthesis. Eur J Nucl Med Mol Imaging 2003; 30(5): 705-15.

639

Verberne S J, Raijmakers P G, Temmerman O P. The accuracy of imaging techniques in the assessment of periprosthetic hip infection: a systematic review and meta-analysis. J Bone Joint Surg Am 2016; 98(19): 1638-45. Zhuang H, Duarte P S, Pourdehnad M, Maes A, Van Acker F, Shnier D, Garino J P, Fitzgerald R H, Alavi A. The promising role of 18F-FDG PET in detecting infected lower limb prosthesis implants. J Nucl Med 2001; 42(1): 44-8. Zhuang H, Chacko T K, Hickeson M, Stevenson K, Feng Q, Ponzo F, Garino J P, Alavi A. Persistent non-specific FDG uptake on PET imaging following hip arthroplasty. Eur J Nucl Med Mol Imaging 2002; 29(10): 1328-33.

640

Acta Orthopaedica 2018; 89 (6): 640–645

Higher mid-term revision rates of posterior stabilized compared with cruciate retaining total knee arthroplasties: 133,841 cemented arthroplasties for osteoarthritis in the Netherlands in 2007–2016 Anneke SPEKENBRINK-SPOOREN 1, Liza N VAN STEENBERGEN 1, Geke A W DENISSEN 1, Bart A SWIERSTRA 2, Rudolf W POOLMAN 3, and Rob G H H NELISSEN 4 1 Dutch

Arthroplasty Register (Landelijke Registratie Orthopedische Implantaten), ’s-Hertogenbosch; 2 Department of Orthopaedic Surgery, Sint Maartenskliniek, Nijmegen; 3 Department of Orthopaedic Surgery, Joint Research, OLVG Amsterdam; 4 Department of Orthopaedic Surgery, Leiden University Medical Center, Leiden, The Netherlands Correspondence: aspekenbrink@orthopeden.org Submitted 2017-12-21. Accepted 2018-07-14.

Background and purpose — The preference for a cruciate retaining (CR) or posterior stabilized (PS) TKA (total knee arthroplasty) system varies greatly between Dutch hospitals, independent of patient characteristics. We examined mid-term revision rates for men and women of different age categories. Patients and methods — We included all 133,841 cemented fixed-bearing primary CR or PS TKAs for osteoarthritis reported in the Dutch Arthroplasty Register (LROI) in 2007–2016. Revision procedures were defined as minor when only insert and/or patella were revised and as major when fixed components (tibia and femur) were revised or removed. 8-year all-cause revision rates of CR and PS TKAs were calculated using competing-risk analyses. Cox-regression analyses were performed, adjusted for age at surgery, sex, ASA -score, and previous operations. Results — PS TKAs were 1.5 (95% CI 1.4–1.6) times more likely to be revised within 8 years of the primary procedure, compared with CR TKAs. When stratified for sex and age category, 8-year revision rate of PS TKAs in men < 60 years was 13% (CI 11–15), compared to 7.2% (CI 6.1– 8.5) of CR TKAs. Less prominent differences were found in older men and women. For men < 60 years differences were found for minor (CR 1.8% (CI 1.4-2.5); PS 3.7% (CI 3.0–4.7)) and major revisions (CR 4.2% (CI 3.3–5.3); PS 7.0% (CI 5.6–8.7)). Interpretation — Patients who received a cemented fixed-bearing primary PS TKA for osteoarthritis are more likely to undergo either a minor or a major revision within 8 years. This is especially prominent for younger men.

In the Netherlands the number of TKAs increased from over 21,000 in 2012 to almost 25,000 in 2016 (LROI 2017). After 8 years, close to 95% of primary TKAs in the Netherlands are still not revised (LROI 2017). The vast majority of TKA systems are either cruciate retaining (CR) or posterior stabilized (PS) designs. The proportion of PS TKA systems has increased over the years in the Netherlands without any scientific explanation. Theoretically, in patients with a non-functional posterior cruciate ligament (PCL), a PS TKA system has to be used. In patients with a functional PCL, the decision on which design should be used depends largely on the preference and training of the surgeon (Jacobs et al. 2005, van den Boom et al. 2009). In the literature, there is no consensus regarding the outcome of CR compared with PS TKAs. Several studies showed no differences (either for survival, or for functional, clinical, and radiological outcome parameters) for CR compared with PS TKA systems (Jacobs et al. 2005, Verra et al. 2015, Jasper et al. 2016). Other studies concluded that CR TKAs showed better survival rates (Abdel et al. 2011, Vertullo et al. 2017). Data from registries with high completeness and validity (van Steenbergen et al. 2015, LROI 2017) could be valuable when they have valid classification of the type of design (i.e., CR or PS). We evaluated the 8-year revision rates for all CR and PS TKA systems in the Dutch Arthroplasty Register, to determine the association between patient characteristics and revision rates, as well as the characteristics of these revision procedures. Furthermore, an analysis of the reason for revision is performed.

Patients and methods Dutch Arthroplasty Register The Dutch Arthroplasty Register (LROI) is a nationwide population-based register that includes information on joint arthro© 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 10.1080/17453674.2018.1518570

Acta Orthopaedica 2018; 89 (6): 640–645

plasties in the Netherlands since 2007 (van Steenbergen et al. 2015). The LROI is initiated by the Netherlands Orthopaedic Association (NOV). Nearly all Dutch orthopedic surgeons are members of this society. The LROI is well supported by these members, resulting in a completeness of 99% of primary knee arthroplasties and 98% of knee revision arthroplasties and a high overall validity in 2016 (LROI 2017). Data collection The LROI database contains information on patient, operation, and prosthesis characteristics (van Steenbergen et al. 2015). For each component a product number is registered. Prosthesis characteristics are derived from an implant library, which contains several core characteristics, including name, type (e.g., PS or CR TKA) and material of prostheses. These data are supplied by implant manufacturers or distributors in the Netherlands, resulting in a database containing almost 36,000 different implant components. If a product number is not present in the database when registered by a local hospital, the specific manufacturer is contacted to add their data to the implant library. A primary knee arthroplasty is defined as the first implantation of a prosthesis. Knee revision arthroplasty is defined as any exchange (placement, replacement, or removal) of 1 or more components of the prosthesis. Revision procedures were categorized as minor if only the patella and insert were revised (excluding patella additions) and as major when at least the femur or tibia component was revised. Patella additions were studied as a separate category. Date of death of patients was obtained on a regular basis from Vektis, the national insurance database, which records vital status of Dutch citizens (Vektis 2017). For the present study, we selected all 133,841 cemented fixed bearing primary TKAs for osteoarthritis in the period 2007–2016 with a PS or CR TKA system. Overall physical condition was scored using the ASA score (I–IV). The median follow-up was 4 years (0–10). Statistics Survival time was calculated as the time between primary arthroplasty and first revision arthroplasty for any reason (Nelissen et al. 1992), death of the patient, or end of the study follow-up (January 1, 2017). 8-year all cause crude cumulative incidence of revision of CR and PS TKAs was calculated using competing risk analysis, where death was considered to be a competing risk (Andersen et al. 2012, Lacny et al. 2015, Wongworawat et al. 2015). Multivariable Cox regression analyses were performed to compare the adjusted revision rates between the types of TKA system. Adjustments were made for age at surgery, sex, ASA score, and previous operations. Reasons for revision were compared between the type of TKA system (PS and CR TKA system) using a chi-square test. All statistical analyses were done using SPSS version 23.0 (IBM Corp, Armonk, NY, USA).

641

Table 1. Patient characteristics of patients who received a cemented fixed-bearing primary total knee arthroplasty for osteoarthritis with cruciate retaining (CR) or posterior stabilized (PS) knee system in the Netherlands in 2007–2016

CR TKA (n = 60,546)

PS TKA (n = 73,295)

Mean age (SD) 69 (9.3) 69 (9.4) Age (%) < 50 2 2 50–59 14 15 60–69 36 36 70–79 35 35 ≥ 80 13 12 Sex (%) Male 34 34 Female 66 66 ASA score (%) I 19 18 II 67 68 III–IV 14 14 Previous surgery a (%) Yes 32 34 No 67 65 Unknown 1 1

Total (n = 133,841) 69 (9.3) 2 15 36 35 12 34 66 19 67 14 33 66 1

a Includes

meniscectomy, arthroscopy, osteotomy, osteosynthesis, ligament reconstruction, synovectomy, and other previous surgery. TKA: total knee arthroplasty.

For the 95% confidence intervals (CI), we assumed that the number of observed cases followed a Poisson distribution. Funding and conflicts of interest This research did not receive any grants. None of the authors declare any competing interests.

Results Patient characteristics 133,481 cemented fixed-bearing primary TKAs for osteoarthritis were registered by 100 hospitals in the period 2007– 2016. Of these TKAs, 45% were a CR and 55% were a PS TKA. Patient characteristics were similar between CR and PS TKA systems (Table 1). A patellar component was placed during the primary procedure in 18% (n = 10,668) of CR TKAs and in 29% (n = 20,831) of PS TKAs. The proportion of CR TKAs per hospital ranged from 0% to 100%, with 20 hospitals using CR TKA systems in over 95% of all TKAs and 24 hospitals using PS TKA systems in over 95% of cases. Revision rates Of all CR TKA systems, 4.2% (CI 4.0–4.4) were revised within 8 years, while 6.1% (CI 5.8–6.3) of PS TKAs were revised (Figure 1). Of all CR TKA systems 12.7% (CI 12.2– 13.2) of the patients died and in 11.1% (CI 10.6–11.6) of primary PS TKAs the patients died within 8 years of the primary procedure.

642

Acta Orthopaedica 2018; 89 (6): 640–645

Cumulative revision rate (%)

Cumulative major revision rate in women (%)

Cumulative major revision rate in men (%)

Posterior stabilized

PS, < 60

CR, < 60 5

PS, < 60 PS, 60–69

4 Cruciate retaining

PS, 60–69 3

CR, 60–69 PS, ≥ 70

2 1

CR, ≥ 70

Years after index operation

Figure 1. Cumulative revision rates of cemented cruciate retaining or posterior stabilized fixed bearing total knee arthroplasties for osteoarthritis by type of knee system in the Netherlands in 2007–2016 (n = 133,841). Dotted lines represent the 95% confidence interval.

Crude model n HRcrude CI

3 CR, 60–69

PS, ≥ 70

1 0

CR, ≥ 70 0

Years after index operation

Figure 3. Cumulative major revision rates of cemented cruciate retaining or posterior stabilized fixed bearing total knee arthroplasties for osteoarthritis by type of knee system per age category, stratified for sex, in the Netherlands in 2007–2016 (n = 133,841). CR: cruciate retaining TKA system; PS: posterior stabilized TKA system

Table 2. Multivariate survival analyses of cemented cruciate retaining (CR) or posterior stabilized (PS) fixed bearing total knee arthroplasties for osteoarthritis in the Netherlands in 2007–2016

CR, < 60 4

Adjusted model a HRadj CI

TKA system CR (ref) 60,546 1.0 1.0 PS 73,295 1.4 b 1.4–1.5 1.5 b 1.4–1.6 Age < 50 2,851 2.5 b 2.2–2.9 2.4 b 2.1–2.8 50–59 19,552 1.6 b 1.4–1.7 1.4 b 1.4–1.7 60–69 (ref) 47,982 1.0 1.0 70–79 46,864 0.8 b 0.7–0.9 0.8 b 0.8–0.9 ≥ 80 16,368 0.5 b 0.4–0.6 0.5 b 0.5–0.6 Sex Female (ref) 87,660 1.0 1.0 Male 45,905 1.1 c 1.0–1.1 1.0 0.9–1.1 ASA score I (ref) 24,298 1.0 1.0 II 89,657 0.9 d 0.9–1.0 1.1 c 1.0–1.2 III–IV 18,164 1.0 0.9–1.1 1.3 b 1.2–1.4 Previous operations e No (ref) 84,447 1.0 1.0 Yes 42,503 1.4 b 1.3–1.5 1.2 b 1.1–1.3 TKA: total knee arthroplasty; HRcrude: crude hazard ratio; HRadj: adjusted hazard ratio. a Adjusted for age at surgery, sex, ASA score, and previous operations. b p < 0.001; c p = 0.02; d p = 0.04. e Includes meniscectomy, arthroscopy, osteotomy, osteosynthesis, ligament reconstruction, synovectomy, and other previous surgery.

Overall, the risk of a TKA being revised within 8 years after the primary procedure strongly depended on age, with TKAs in patients under 50 years of age being 2.4 (CI 2.1–2.8) times more likely to be revised, compared with TKAs in patients aged 60–69 years. TKA systems in patients aged over 80 years

Table 3. Multivariate survival analyses of cemented fixed bearing primary cruciate retaining (CR) or posterior stabilized (PS) total knee arthroplasty for osteoarthritis in the Netherlands in 2007–2016

TKA system

Primary Revisions Crude model Adjusted model a TKAs (n) (n) HRcrude CI HRadj CI

Any type of revision CR (ref) 60,546 1,718 PS 73,295 2,659 Minor revisions c CR (ref) 60,546 333 PS 73,295 713 Patella addition d CR (ref) 60,546 402 PS 73,295 555 Major revisions e CR (ref) 60,546 911 PS 73,295 1,238

1.0 1.4 b 1.4–1.5

1.0 1.5 b 1.4–1.6

1.0 2.0 b 1.7–2.2

1.0 2.0 b 1.7–2.3

1.0 1.3 b 1.1–1.5

1.0 1.3 b 1.2–1.4

For abbreviations, see Table 2 a Adjusted for age at surgery, sex, ASA score, and previous operations. b p < 0.001. c Only insert and/or patella exchange (excluding patella addition). d Only patella addition. A patellar component was placed during the primary procedure of 17.7% (n = 10,668) of cruciate retaining TKAs and of 28.5% (n = 20,831) of posterior stabilized TKAs. e Revision of at least femur or tibia.

were half as likely to be revised (HR 0.5; CI 0.5–0.6). Furthermore, TKAs in patients who underwent a previous surgery to the same knee were 1.2 (CI 1.1–1.3) times more likely to be revised within 8 years (Table 2). PS TKAs were 1.5 (CI 1.4–1.6) times more likely to be revised within 8 years, compared with CR TKA systems (Table 3). After excluding revisions for infection, a revision was 1.4 (CI 1.3–1.5) times more likely to be performed on a

Acta Orthopaedica 2018; 89 (6): 640–645

643

primary PS TKA. When stratified for sex and age category, the difference in cumulative revision rates was predominant for younger males. Of PS TKA systems in males under 60 years of age, 13% (CI 11–15) were revised within 8 years of the primary procedure, compared with 7.2% (CI 6.1–8.5) of CR TKA systems. This difference was not present in younger females (CR 8.2% (CI 7.2–9.3); PS 9.4% (CI 8.5–11)) (Figure 2, see Supplementary data). Regarding general health (ASA score), the difference in 8-year cumulative revision between CR and PS TKA systems was most prominent in patients with ASA score II (Table 4, see Supplementary data).

instability (CR 41%; PS 37%), followed by infection, which was more often the reason for minor revisions in PS TKAs, compared with CR TKAs. Loosening of the tibial component was the most common reason for performing a major revision, though it was more common on a PS (41%) compared with a CR TKA system (27%; p < 0.001). Furthermore, instability was more often a reason for performing a major revision on a CR (36%) compared with a PS TKA system (23%; p < 0.001) (Table 6, see Supplementary data).

Revision characteristics Of all primary cemented fixed-bearing CR or PS TKAs for osteoarthritis in the period 2007–2016, 4,377 (CR 1,718; PS 2,659) TKAs were revised within 8 years of the primary procedure. Of these revisions, 49% were major (CR 53%; PS 47%) and 24% were minor revisions (CR 19%; PS 27%). Furthermore, 22% of all revisions were a patella addition (CR 23%; PS 21%). A minor revision was 2.0 (CI 1.7–2.3) times more likely to be performed on a primary PS TKA, compared with a primary CR TKA within 8 years of the primary procedure. After excluding revisions for infection, a minor revision was 1.8 (CI 1.5–2.1) times more likely to be performed on a primary PS TKA. A major revision was 1.3 (CI 1.2–1.4) times more likely to be performed on a PS TKA system compared with a CR TKA system (Table 3). When zooming in on patient characteristics, in general, CR TKAs had higher revision rates for female patients compared with male patients (Table 5, see Supplementary data). Male patients who received a PS TKA system significantly more often underwent a minor revision procedure (2.0% CI 1.7– 2.3) compared with male patients who received a CR TKA system (1.1% CI 0.9–1.3). These results were predominant for younger males (PS TKAs 3.7% (CI 3.0–4.7); CR TKAs 1.8% (CI 1.4–2.5)). In females, minor revision rates between CR and PS TKA systems also differed for younger patients (PS TKAs:= 2.7% (CI 2.1–3.4); CR TKAs:= 1.2% (CI 0.9–1.6)). This difference was less prominent. Furthermore, major revisions were more common in younger males who received a PS TKA system (7.0% CI 5.6–8.7), compared with younger males who received a CR TKA system (4.2% CI 3.3–5.3; Table 5, see Supplementary data). The differences in revision rates between CR and PS TKA systems increased as the number of years after the primary TKA increased. In females, major revision rates between CR and PS TKA systems did not differ for patients under 60 years of age (Figure 3).

Discussion

Reasons for revision Patellar pain was the most common reason for revision within 8 years for both CR and PS TKA systems. Reasons for revision were different for minor compared with major revisions. The most common reason for performing a minor revision was

Main findings and previous research We found higher mid-term minor and major revision rates in patients with PS TKA systems, especially in males younger than 60 years. These results are not in line with several previous review studies comparing CR and PS TKA designs (Jacobs et al. 2005, Verra et al. 2015, Jasper et al. 2016). Possible reasons for this difference might be non-consecutive study series with subsequent selection bias, smaller study sizes, and patients lost to follow-up in these previous studies, emphasizing the importance of analyzing population-based data. However, our results confirm results of other previous studies (Abdel et al. 2011, Vertullo et al. 2017). As life expectancy is growing and the mean age of patients undergoing primary arthroplasty continues to decrease, revision rate and survival of implants are becoming increasingly important (Carr et al. 2012, Pabinger et al. 2013, Hamilton et al. 2015, Shah et al. 2017). The number of patients currently living with a TKA suggests a large potential revision healthcare burden (Hamilton et al. 2015). International literature has shown that revision rates differ for a number of patient characteristics, including age, sex, general health, physical activity, BMI, and smoking (McCalden et al. 2013, Pabinger et al. 2013, Apold et al. 2014, Jasper et al. 2016, Kunutsor et al. 2016, LROI 2017). Preoperative decision-making is a complex process for both the surgeon and patient (Carr et al. 2012). In particular age and weight are important indicators for surgery. Patients aged less than 55 years or with preoperative morbid obesity have more variable outcomes after knee replacement than those older than 55 years and those with a lower BMI (Carr et al. 2012). A recent study showed that patients aged under 60 years had an increased risk for revision of their TKA, with time to revision reaching a peak around 5 years after implantation (Bayliss et al. 2017). Grade of preoperative radiographic destruction in combination with high age have been reported as important factors for preoperative decision-making (Verra et al. 2015). However, patient characteristics do not appear to be the main motive when choosing between a CR and PS TKA system in the Netherlands, since large variation is seen between hospitals. In patients with a functional PCL, the decision on which design should be used depends largely on the preference and

644

training of the surgeon (Jacobs et al. 2005, van den Boom et al. 2009). Our study shows, however, that mid-term revision rates are higher for PS TKA systems, especially for younger males. This implies a need to consider these patient characteristics when choosing the type of TKA system. The incidence of TKA is increasing (LROI 2017), and is projected to increase rapidly in the near future (Carr et al. 2012), particularly in young patients (Kurtz et al. 2009). As the number of primary TKAs grows, the number of revisions is expected to increase as well (Ethgen et al. 2004, Hamilton et al. 2015). Although overall improvement is seen in patients’ health and function, outcome of knee revision arthroplasty is worse than that of the primary procedure, with even worse results for early revisions in young patients (Hardeman et al. 2012). Survival rates of revision surgery are consistently reported at around 80% at 10 years (Hardeman et al. 2012, Hamilton et al. 2015). Younger patients are also more likely to undergo a reoperation, without revision of TKA system components (Zmistowski et al. 2011). Large discrepancies exist between the extensiveness of revision arthroplasties. The difference in mid-term revision rates between CR and PS TKA systems was more prominent for minor revisions (i.e., insert or patella revision). This may partially be a consequence of hospital policy or surgeon preference. This does not, however, explain the sex difference. For either minor or major revisions instability was a frequent reason for revision, indicating an important motivation for deciding on the preference for CR and PS TKA systems. The higher frequency of loosening of the tibial component as a reason for major revision in PS TKAs, compared with CR TKAs, might possibly be the result of higher shear forces on the tibial component in the PS TKA models. Moreover, after removing the PCL the flexion gap is usually bigger and inserts tend to be slightly thicker in PS models than in CR models, which may affect shear forces. However, these are only speculations as an observational study cannot disclose causative mechanisms. The higher frequency of infection as reason for revision in PS TKA systems compared with CR TKA systems was unexpected. Although types of revisions are considered to be small, these findings are of importance for the results of TKAs in terms of the risk for complications, the patient’s perspective on the results, and health costs. Strengths and limitations Using a generic matrix on design features across all implants enables analyses of conceptual design issues of implants. With the help of the implant library, the register shows which component is registered for every reference number. This reduces the risk of registration errors. Furthermore, the LROI contains a large population-based nationwide database of primary TKAs, with a completeness of nearly 100% (van Steenbergen et al. 2015, LROI 2016). This is the first study on this subject with this many included procedures in the known literature. Registry studies like this study have the advantage

Acta Orthopaedica 2018; 89 (6): 640–645

of analyzing population-based data, avoiding selection bias, and minimizing misclassification error. Nevertheless, registry data also have their drawbacks since only limited variables are registered and causality cannot be proven due to its observational nature. Data from registries should accompany prospective cohort studies like randomized controlled trials, which can collect a large number of variables. An analysis of patient characteristics determining the choice for a CR or PS TKA design would be interesting, but also difficult, if not impossible, to undertake. The latter stresses even more the importance of our conclusions: which TKA, CR or PS, is appropriate for a typical patient. A limitation of this study is that there is probably residual confounding due to BMI, Charnley score, and smoking. BMI is known to be associated with TKA revision and mortality rates (Carr et al. 2012, Hamilton et al. 2015). The same is true for the degree of osteoarthritis in lower extremity joints (i.e., Charnley score) and smoking. These confounding factors have been recorded only since July 2013 and could therefore not be included in the present study. Furthermore, a TKA has proven to be a successful procedure, resulting in a small proportion of revision procedures. Despite the large database of primary TKAs, more data are needed to study characteristics of these revision procedures in more detail. This study was unable to study patient characteristics sufficiently in relation to reasons for minor or major revisions by type of TKA system, particularly in younger patients. Infections are only registered in the LROI when 1 of the prosthesis components is revised. Furthermore, only (suspected) prosthetic joint infections as reason for revision were registered, since this is registered at the time of operation and not based on microbiology results. As had been shown earlier, implant registries largely underscore prosthetic joint infections (Gundtoft et al. 2015). When excluding revisions where infection was registered to be a reason for revision, however, revision rates were still significantly higher for PS TKAs, compared with CR TKAs. Further research is recommended on the effect of revision procedures on the patient’s perspective on the results. Patientreported outcome measures are important in the evaluation of revisions, since an important indication is pain. Data on pain in the postoperative period were not available for our study population (LROI 2017). Furthermore, research is warranted on high-flexion TKAs as a factor influencing revision rates. This prosthesis characteristic could not currently be detected. Finally, only 1 prosthesis characteristic was studied, while a complex interaction between factors has become increasingly evident. An understanding of this interplay between prosthesis characteristics on the outcome of TKA survival is needed to assist surgeons to optimize outcomes for patients (Carr et al. 2012). Summary Revision rates within 8 years of the primary procedure were 1.5 times higher for PS TKA compared with CR TKA systems.

Acta Orthopaedica 2018; 89 (6): 640–645

This was more prominent in male patients under 60 years of age. Although this effect was largest in minor revisions, it was also found for major revisions. Our data may give guidance for surgical decision-making for specific patient groups when choosing a CR or PS TKA system. Supplementary data Figure 2 and Tables 4–6 are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674.2018.1518570 The authors contributed to: (1) study design and study protocol, (2) gathered data, (3) analyzed data, (4) initial draft, and (5) final draft. AS and LS contributed to: 1–5; GD, BS, and RP contributed to: 2, 4, 5; RN contributed to: 1, 2, 4, 5.

Acta thanks Asgeir Gudnason and Mika Niemeläinen for help with peer review of this study.

Abdel M P, Morrey M E, Jensen M R, Morrey B F. Increased long-term survival of posterior cruciate-retaining versus posterior cruciate-stabilizing total knee replacements. J Bone Joint Surg Am 2011; 93(22): 2072-8. Andersen P K, Geskus R B, de Witte T, Putter H. Competing risks in epidemiology: possibilities and pitfalls. Int J Epidem 2012; 41(3): 861-70. Apold H, Meyer H E, Nordsletten L, Furnes O, Baste V, Flugsrud G B. Risk factors for knee replacement due to primary osteoarthritis, a population based, prospective cohort study of 315,495 individuals. BMC Musculoskelet Disord 2014; 15: 217. Bayliss L E, Culliford D, Monk A P, Glyn-Jones S, Prieto-Alhambra D, Judge A, Cooper C, Carr A J, Arden N K, Beard D J, Price A J. The effect of patient age at intervention on risk of implant revision after total replacement of the hip or knee: a population-based cohort study. Lancet 2017; 389(10077): 1424-30. Carr A J, Robertsson O, Graves S, Price A J, Arden N K, Judge A, Beard D J. Knee replacement. Lancet 2012; 379(9823): 1331-40. Ethgen O, Bruyere O, Richy F, Dardennes C, Reginster J Y. Health-related quality of life in total hip and total knee arthroplasty: a qualitative and systematic review of the literature. J Bone Joint Surg Am 2004; 86-A(5): 963-74. Gundtoft P H, Overgaard S, Schønheyder H C, Møller J K, KjærsgaardAndersen P, Pedersen A B. The “true” incidence of surgically treated deep prosthetic joint infection after 32,896 primary total hip arthroplasties: a prospective cohort study. Acta Orthop 2015; 86(3): 326-34. Hamilton D F, Howie C R, Burnett R, Simpson A H, Patton J T. Dealing with the predicted increase in demand for revision total knee arthroplasty: challenges, risks and opportunities. Bone Joint J 2015; 97-B(6): 723-8. Hardeman F, Londers J, Favril A, Witvrouw E, Bellemans J, Victor J. Predisposing factors which are relevant for the clinical outcome after revision total knee arthroplasty. KSSTA 2012; 20(6): 1049-56. Jacobs W C, Clement D J, Wymenga A B. Retention versus removal of the posterior cruciate ligament in total knee replacement: a systematic literature review within the Cochrane framework. Acta Orthop 2005; 76(6): 757-68.

645

Jasper L L, Jones C A, Mollins J, Pohar S L, Beaupre L A. Risk factors for revision of total knee arthroplasty: a scoping review. BMC Musculoskelet Disord 2016; 17: 182. Kunutsor S K, Whitehouse M R, Blom A W, Beswick A D. Patient-related risk factors for periprosthetic joint infection after total joint arthroplasty: a systematic review and meta-analysis. PloS One 2016; 11(3): e0150866. Kurtz S M, Lau E, Ong K, Zhao K, Kelly M, Bozic K J. Future young patient demand for primary and revision joint replacement: national projections from 2010 to 2030. Clin Orthop Rel Res 2009; 467(10): 2606-12. Lacny S, Wilson T, Clement F, Roberts D J, Faris P D, Ghali W A, Marshall D A. Kaplan–Meier survival analysis overestimates the risk of revision arthroplasty: a meta-analysis. Clin Orthop Rel Res 2015; 473(11): 343142. LROI. LROI Report. 2016: Blik op Uitkomsten [in Dutch]. In: http://wwwlroi-rapportagenl/2016 accessed 10-02-2018 (Eds. Spekenbrink-Spooren A, Van Steenbergen L N, Denissen G A W, Rijnsburger E, Van der Togt C R); 2016. LROI. LROI Report. 2017 (Eds. Spekenbrink-Spooren A, Van Steenbergen L N, Denissen G A W, Rijnsburger E, Van der Togt C R); 2017. http://www. lroi-rapportage.nl/2017. McCalden R W, Robert C E, Howard J L, Naudie D D, McAuley J P, MacDonald S J. Comparison of outcomes and survivorship between patients of different age groups following TKA. J Arthroplasty 2013; 28(8 Suppl): 83-6. Nelissen R G, Brand R, Rozing P M. Survivorship analysis in total condylar knee arthroplasty. A statistical review. J Bone Joint Surg Am 1992; 74(3): 383-9. Pabinger C, Berghold A, Boehler N, Labek G. Revision rates after knee replacement: cumulative results from worldwide clinical studies versus joint registers. Osteoarthritis Cartilage. 2013;21(2):263-8. Shah S H, Schwartz B E, Schwartz A R, Goldberg B A, Chmell SJ. Total knee arthroplasty in the younger patient. J Knee Surg 2017; 30(6): 555-9. van den Boom L G, Brouwer R W, van den Akker-Scheek I, Bulstra S K, van Raaij J J. Retention of the posterior cruciate ligament versus the posterior stabilized design in total knee arthroplasty: a prospective randomized controlled clinical trial. BMC Musculoskelet Disord 2009; 10: 119. van Steenbergen L N, Denissen G A, Spooren A, van Rooden S M, van Oosterhout F J, Morrenhof J W, Nelissen R G. More than 95% completeness of reported procedures in the population-based Dutch Arthroplasty Register. Acta Orthop 2015; 86(4): 498-505. Vektis. http://www.vektis.nl. 2017. Verra W C, Boom L G, Jacobs W C, Schoones J W, Wymenga A B, Nelissen R G. Similar outcome after retention or sacrifice of the posterior cruciate ligament in total knee arthroplasty. Acta Orthop 2015; 86(2): 195-201. Vertullo C J, Lewis P L, Lorimer M, Graves S E. The effect on long-term survivorship of surgeon preference for posterior-stabilized or minimally stabilized total knee replacement: an analysis of 63,416 prostheses from the Australian Orthopaedic Association National Joint Replacement Registry. J Bone Joint Surg Am. 2017;99(13):1129-39. Wongworawat M D, Dobbs M B, Gebhardt M C, Gioe T J, Leopold S S, Manner P A, Rimnac C M, Porcher R. Editorial: Estimating survivorship in the face of competing risks. Clin Orthop Rel Res 2015; 473(4): 1173-6. Zmistowski B, Restrepo C, Kahl L K, Parvizi J, Sharkey P F. Incidence and reasons for nonrevision reoperation after total knee arthroplasty. Clin Orthop Rel Res 2011; 469(1): 138-45.

646

Acta Orthopaedica 2018; 89 (6): 646–651

Blood cytokine pattern and clinical outcome in knee arthroplasty patients: comparative analysis 5 years after standard versus “hypoallergenic” surface coated prosthesis implantation Peter THOMAS 1, Philipp HISGEN 2, Hartmuth KIEFER 3, Ulf SCHMERWITZ 3, Andreas OTTERSBACH 4, Dominique ALBRECHT 4,5, Burkhard SUMMER 1, and Christian SCHINKEL 2 1 Department

of Dermatology and Allergology, Ludwig-Maximilians-University, Munich, Germany; 2 Department of Trauma Surgery and Orthopaedics, Klinikum Memmingen, Germany (Academic Teaching Hospital of Ludwig-Maximilians-University Munich); 3 Clinic of Trauma Surgery and Orthopaedics, Bünde, Germany; 4 Clinic for Orthopaedic Surgery, Brig, Switzerland; 5 Clinic of Trauma Surgery and Orthopaedics, St Gallen, Switzerland Correspondence: Peter.Thomas@med.uni-muenchen.de Submitted 2018-03-26. Accepted 2018-07-24.

Background and purpose — Metal sensitivity might provoke complications after arthroplasty. Correspondingly, coated “hypoallergenic” implants are of interest but longterm follow-up data are missing. Thus, we assessed immunological and clinical parameters in such patients. Patients and methods — 5 years’ follow-up data were obtained from 3 centers, which used either a standard total knee replacement (TKR) or the identical implant with multilayer surface zirconium nitride based coating. Of the 196 patients (mean age 68 years (44–84), 110 females) 97 had arthroplasty with a coated surface, and 99 were treated by a standard TKR of the same type. Investigations were Knee Society Score (KSS), Knee injury and Osteoarthritis Outcome Score (KOOS), radiographic analysis, and cytokine measurement in peripheral blood. Pro- and anti-inflammatory cytokines were evaluated by cytometric beads assay and RT-PCR. Results — Survival rate (Kaplan–Meier) was 98% for coated and 97% for uncoated implants after 5 years. Mechanical axis and KSS pain score (42 vs. 41 (0–50)) were comparable. Most serum cytokine levels were comparable, but mean interleukin-8 and interleukin-10 levels were higher in the group with an uncoated implant. IL-8: 37 (SD 7.5) pg/ mL vs. 1.1 (SD 4.3) (p < 0.001); IL-10: 3.6 (SD 2.5) vs. 0.3 (SD 1.8) pg/mL (p < 0.001). Interpretation — There was similar clinical outcome 5 years after standard and surface-coated TKR. In peripheral blood there was an increased pro-inflammatory status, i.e., significant elevation of IL-8 and the anti-inflammatory IL-10, after standard uncoated prosthesis. Any long-term effects of these cytokine changes are unknown.

Revision rates after total knee replacement (TKR) are about 9.5% in Germany and about 8.4% in the USA for the year 2011 (Wengler et al. 2014). The pathogenesis of implant failure is multifactorial (Sharkey et al. 2002, Bourne et al. 2010), and adverse reactions to metal ions released from arthroplasty represent one such factor. Both unspecific (“innate”) immune reactions and specific T lymphocytic sensitization with subsequent hypersensitivity mechanisms may contribute to arthroplasty intolerance reactions in patients (Mahendra et al. 2009, Reito et al. 2016, Hallab et al. 2017). The clinical spectrum encompasses both local and systemic effects, being often unspecific like: pain, swelling, effusion, reduced range of motion, impaired wound healing, eczema. Diagnostic findings include osteolysis, lymphocyte-dominated peri-implant inflammation, and in extreme situations soft tissue necrosis or pseudotumor formation. In 2017 Hallab et al. framed the statement “a consensus of studies agree the dominant form of this response is due to innate reactivity to implant debris danger signaling” (Hallab et al. 2017). Thus, the corresponding release of cytokines and chemokines might open the way to implant failure. Examples are: chemokines like IL-8, CXCL10, or monocyte chemotactic protein-1 (Lawrence et al. 2014), danger signaling associated inflammasome pathway with IL-1ß/IL-18 formation (Yang et al. 2015) or TIRAP/Mal mediated particle-induced osteolysis (Bechtel et al. 2016). Correspondingly, various approaches have been discussed to diminish the metal ion exposure in arthroplasty patients, such as reducing wear by modifying modularity or surface modification (Bolognesi et al. 2015, Morlock 2015, Thomas et al. 2015). It might be speculated that the mostly inflammatory pathologic reactions to metal ions and debris represent stages of “broken tolerance” of the implant-exposed organism. However, the underlying mechanisms leading to immune homeo-

Acta Orthopaedica 2018; 89 (6): 646–651

stasis are also not well understood in patients with symptom-free arthroplasty. We compared 2 groups of TKR patients at 5-year follow-up who had had the identical type of prosthesis with one exception: In half of the patients a variant was implanted with multilayer Advanced Surface (AS) coating (zirconium nitride (ZrN)) to reduce metal ion release. First, we assessed the clinical performance of the 2 variants of TKR in these patients. Second, we investigated immune reactivity to TKR with a focus on the systemic response, i.e., the response pattern of selected cytokines and chemokines in peripheral blood. The latter was done because, in contrast to in vitro studies often using cell lines, such an approach is rarely performed. We hypothesized that a local yet “subclinical” inflammatory state might be also recognizable by this approach.

647

Characteristics of the 196 patients with primary TKR and of the 40 individuals without metal implant Group IV Group I Group II Group III healthy uncoated a coated b preoperative c controls d Patients Sex (female/male) Age median (years) Smoker (formely smokers) History of skin metal allergy Atopy e a b c d e

99 97 20 20 54/45 56/41 17/3 14/6 67 (44–84) 70 (48–82) 60 (46–84) 61 (52–63) 5 (11) 10 (19) 2 2 0 6 14 3 11 4 12 3

Uncoated = primary knee arthroplasty with e.motion. Coated = primary knee arthroplasty with multilayer coated, i.e., AS e.motion. Gonarthrosis patients, prior to knee arthroplasty (at present no metal implant) Healthy patients without metal implant. Atopy = history of hay fever and/or allergic asthma and/or atopic dermatitis.

Patients and methods Patients and controls All 322 patients who received a TKR (e.motion UC, Aesculap AG, Tuttlingen, Germany) in 2007 were invited for mid-term follow-up (around 5 years). Of these, 41 patients were lost to follow-up. For 78 patients, who were not willing or able to participate in the follow-up, at least information on prosthesis status could be obtained and in 7 patients 1 or more implant components had been exchanged. 196 patients (mean age 68 years [44–84], 110 females) with primary knee arthroplasty took part in the study (Table). The follow-up examinations were done in 3 centers: (a) Klinikum Memmingen (Germany, center 1); here all patients received exclusively the surface-coated variant of the knee implant. (b) Lukas Hospital Bünde (Germany, center 2) and Hospital Oberwallis Brig (Switzerland, center 3); here patients routinely received the uncoated variant of the knee implant. In centers 2 and 3, if patients had a history of metal allergy they received the surface-coated implant. These few patients are not included in this evaluation. All standard knee implants used were CoCr alloys and were cemented. Only patients from center 3 had received retropatellar resurfacing, but no difference in cytokine levels was found between the two centers (2 and 3) with uncoated implants. The follow-up assessment encompassed clinical, radiological, and immunological examination. Analysis of blood samples was done at the Department of Dermatology and Allergology, Ludwig-Maximilians-University (LMU) Munich. The Table gives the patients’ characteristics. In addition to the arthroplasty patients, 40 individuals without a metal implant volunteered as blood donors for cytokine

control assessment: 20 healthy individuals (median age 61 years [52–63], 14 females; 3 with a history of metal allergy) without a metal implant; 20 individuals with knee osteoarthritis prior to primary arthroplasty (median age 60 years [46–84], 17 females; 14 with a history of metal allergy). The 20 individuals prior to arthroplasty were patients presented by their orthopedic surgeons to the Department of Dermatology and Allergology of LMU for allergy diagnostics. Thus in this group patients with an allergy background are overrepresented. However, we thought that pre-existing “allergy status” would not alter preoperative blood cytokine levels. Clinical and radiological assessments A detailed physical examination was done. In addition, functional and symptom assessment was performed by Knee Society Score (KSS) and Knee injury and Osteoarthritis Outcome Score (KOOS). Radiological evaluation was done using standard weight-bearing knee radiographs in two planes and in load-bearing leg axis to evaluate prosthesis positioning and presence of radiolucencies. Analysis of blood cytokine levels and expression From each patient 2 blood samples were obtained. First sample — After blood drawing, serum was obtained from this sample by centrifugation and immediately frozen. 190 cryopreserved serum samples (98 patients with uncoated, 92 patients with coated arthroplasty) could be analyzed. The blinded samples were assessed for the presence and concentration of 10 cytokines by a multiplex cytometric bead assay (CBA; BD Biosciences, Heidelberg, Germany) via flow cytometry using a FACS canto (BD Biosciences, Heidelberg, Germany). The panel of cytokines included—with the restriction of sometimes overlapping functions—inflammatory (IL-4, IL-5, IL-6, IL-17A, IL-1ß, IFNγ, TNFα), chemoattractant (IL-8, CXCL10), and immune-regulating (IL-10) factors. The respective detection limits were < 0.5 pg/mL.

648

Acta Orthopaedica 2018; 89 (6): 646–651

IL-8 [pg/mL]

IL-6 [pg/mL]

IL-8 [pg/mL] 40

IL-6 [pg/mL]

10 0

uncoated

coated

2 c

uncoated

coated

preoperative

IL-10 [pg/mL]

CXCL10 [pg/mL]

IL-10 [pg/mL]

controls

IL-8 [PCR Units]

0.25

0.15 40 0.10

uncoated

coated

uncoated

coated

40 2 1

IL-10 [PCR Units] 80

0.20

preoperative

CXCL10 [pg/mL]

preoperative

controls

preoperative

controls

Figure 1. (left panel) Serum cytokine levels in patients with the uncoated (n = 98) and coated (n = 92) TKR (e.motion/e.motion-AS). (right panel) Serum cytokine levels in 20 patients prior to primary arthroplasty (“preoperative”) and 20 healthy individuals without implant (“controls”). The cytokines IL8, IL-6, the immunoregulatory IL-10, and the T-cell chemoattractant CXCL10 were assessed. Concentrations are given in pg/mL. b p < 0.01; c p < 0.001; standard deviation is given.

Second sample — A peripheral blood sample with EDTA was stabilized with RNAlater (Ambion, Carlsbad, USA) and cryopreserved. Then RNA was isolated with the RiboPure Blood Kit (Ambion, Carlsbad, USA). After reverse transcription into cDNA (Transcriptor First strand cDNA Synthesis Kit, Roche, Mannheim, Germany) gene expression of 5 different genes (IFNγ, Foxp3, IL-10, IL-8, CASP1) was analyzed by semiquantitative realtime PCR in a LightCycler 1.5. Statistics The statistical packages SPSS (Version 19.0; SPSS Inc, Chicago, IL, USA) and SAS (Version 9.4; SAS Institute, Cary, NC, USA) were used. Differences in the blood cytokine levels were assessed using Student’s t-test. Prosthesis survival was calculated by Kaplan–Meier analysis. Ethics, funding, and potential conflicts of interest The study was approved by the local ethics committee (approval number 186-12). Patient examination and blood sampling was done after obtaining informed consent. Funding was by a research grant from Aesculap AG. There was no further conflict of interest.

Results Orthopedic and radiologic outcomes The average time of follow-up in the 196 arthroplasty patients was 5.7 years (54–77 months). The mean BMI was 28 (20–44) and was comparable between the groups. The KSS were similar between the centers and groups: Center 1 (patients with surface-coated TKR): mean 84 (SD 15). Center 2 (patients

0.05 0.00

uncoated

coated

uncoated

coated

Figure 2. Relative expression levels of mRNA for IL-10 and IL-8 in peripheral blood of patients with the uncoated (n = 98) and coated (n = 92) TKR (e.motion/e.motion-AS). c p < 0.001; standard deviation is given. NS: not significant.

with uncoated TKR): mean 82 (SD 19). Center 3 (patients with uncoated TKR): mean 85 (SD 15). Radiolucent lines were investigated by plain radiographs. No complete loosening was observed. The revision free implant survival rates after 5 years (Kaplan–Meier) were comparable with 98% for coated and 97% for uncoated implants. Cytokine levels and expression in peripheral blood Blood samples of 190 patients could be evaluated. Marked differences were found for 3 of the 10 cytokines. There was a clear-cut difference between patients with uncoated and coated arthroplasty in circulating level of: (i) IL-8 (37 pg/mL [SD 5.6] vs. 1.2 pg/mL [SD 2.1]; p < 0.001); (ii) IL-10 (3.6 pg/mL [SD 8.9] vs. 0.25 pg/mL [SD 0.5]; p < 0.001); (iii) and to a lesser extent IL-6 (2.9 pg/mL [SD 2.0] vs. 2.3 pg/mL [SD 3.4]; p < 0.05). Levels of CXCL10 did not differ statistically significantly (64.9 pg/mL, SD 39.0 vs. 55.6 pg/mL, SD 29.9). The osteoarthritis control patients (prior to arthroplasty) also showed a less pronounced IL-8 and CXCL-10 elevation as compared with the healthy controls. With regard to IL-8 and IL-10, the serum cytokine levels of the healthy controls were similar to those of the patients with a coated arthroplasty. When comparing results of patients with allergy background (here mostly “atopy”) with those without such background, no major difference was found in the different groups. The data are summarized in Figure 1. The statistically significant difference of the IL-8 content in the serum samples of arthroplasty patients was paralleled by a significant difference in the IL-8 expression on RNA level (uncoated: 0.18 [SD 0.02] vs. coated: 0.09 [SD 0.007]). RNA analysis was also done for Foxp3 and IL-10 with the intention to assess regulatory cell marker, for a caspase (casp1) respon-

Acta Orthopaedica 2018; 89 (6): 646–651

sible for cleaving pro IL-1ß into its active form and for IFNγ. Expression of all 4 of these was, however, similar in the 2 patient groups. The results of IL-8 and IL-10 expression are shown in Figure 2.

Discussion It is believed that after wound healing and tissue regeneration most arthroplasty patients adapt to the implant and regain quality of life. However, tissue homeostasis around the “foreign body” may be disturbed by inflammatory mediator response to wear and metal ions. Discussion was triggered by some patients developing extreme adverse reactions to metalon-metal hip arthroplasty, i.e., pseudotumor formation (Hutt et al. 2016). Jämsen et al. (2017) described the correlations between macrophage polarizing cytokines, osteoclast activity, and toll-like receptors—particularly TLR4 in tissue around aseptically loosened hip implants. Among other things, they reported that IL-8 expression correlates with early revision time. Apart from potential toxic or unspecific immunostimulatory mechanisms, metal allergy could also play a role since enhanced metal sensitivity has been found in patients with failed arthroplasty (Hallab et al. 2001, Thomas et al. 2009, Granchi et al. 2012). Thus, use of alternative materials or surface-coated CoCrMo alloys could be an alternative for metalsensitive patients (Bader et al. 2008). In our study we describe comparable good clinical 5-year results in patients with the same CoCrMo-based TKR in uncoated and surface-coated versions. The second major conclusion of our study is that patients with uncoated, well-performing TKR nevertheless present with a subclinical proinflammatory cytokine pattern in the peripheral blood. Unexpectedly, we found that serum IL-8 was substantially elevated in standard uncoated TKR patients. Since in center 1 the surface-coated variant of TKR was used in all patients, the patient series consisted of 6/97 with and 91/97 without known metal allergy. This approach avoids a selection bias regarding metal-allergic patients. The clinical performance reflected by the KSS score was almost identical. This is in line with the Kaplan–Meier implant survival rate of 98% (coated) and 97% (uncoated). These numbers are in accordance with published data on 5-year survival rates of TKR ranging between 88 and 98% (Lee et al. 2013). The rationale for determination of selected cytokines and chemokines in peripheral blood was: first, to avoid joint aspirates in symptom-free patients and second, systemic cytokine pattern in peripheral blood is described to reflect a variety of local tissue conditions (for example rheumatoid arthritis, malignant tumor, hepatic or lung inflammatory disease) (Fang et al. 2016, Lin et al. 2016, Matz et al. 2016, Tan et al. 2017). We selected an array of partly innate proinflammatory mediators (IL-1ß, IL-4, IL-5, IL-6, IL-8), the specific T-cell chemoattractant CXCL10 and the counterregulatory immune “dampening” IL-10. Also, when comparing with the “normal”

649

situation, i.e., osteoarthritis patients prior to knee arthroplasty and healthy individuals without osteoarthritis and without implant, a significantly higher blood level of IL-8 was found in patients with uncoated CoCrMo-based TKR. Despite an allergy background being overrepresented in the 20 preoperative controls, this did not influence the “unspecific blood cytokine profile”. In addition, molecular analysis of peripheral blood cells showed enhanced expression of IL-8 in the patients with uncoated TKR. When considering conditions with potential immune activation/production of systemic proinflammatory mediators, elevated IL-8 levels are typically found in various settings: exercise stimulates increases in the circulating concentrations of several cytokines, namely IL-8 (Peake et al. 2015); in peripheral arterial disease acute phase proteins, matrix metalloproteases, and cytokines—in particular IL-8—are considered as potential biomarkers of pathogen-related inflammation (Signorelli et al. 2014); the degree of elevated serum IL-8 is related to the extent of potential inflammation and fibrosis in chronic liver disease (Nobili et al. 2004); elevated serum IL-8 levels are found for several types of cancer and are discussed as a prognostic marker for malignant disease (Palena et al. 2012, Jin et al. 2014). Already in 2005 Tanaka et al. had reported two observations: serum and synovial fluid IL-8 levels were significantly correlated in hip arthroplasty patients—and elevated IL-8 levels correlated with aseptic loosening of hip arthroplasty (Tanaka et al. 2005). 12 years later, in 2017 Jämsen et al. in retrieval tissue samples not only observed the “protective” IL-4 polarized M2 macrophage type and the IFNγ governed M1 macrophage type adding to aseptic loosening, but also found that “IL-8 expression correlates with early revision time” (Jämsen et al. 2017). Metal ions, in particular cobalt, greatly enhance in vitro IL-8 production by various cell types (Ninomiya et al. 2013 , Lawrence et al. 2014). Recent investigations have shown that cobalt and nickel induce such IL-8 expression via TLR4 activation (Anjum et al. 2016, Lin et al. 2016). In addition, as reported by Lawrence et al. (2014), cobalt ions promote the expression of CXCL10, a T-lymphocyte chemokine. Similarly to IL-6 (Wang et al. 2014, Ettinger et al. 2015) as contributor to autoimmune and inflammatory diseases, the chemokine IL-8 has multiple proinflammatory effects. As an intravascular danger signal it enables trafficking of leukocytes to sites of tissue injury (by chemotaxis and expression of vascular adhesion molecules), it promotes neovascularization, it enables epithelial–mesenchymal transition (both in wound healing/ tissue regeneration and fibrosis/cancer progression) and it prolongs ongoing inflammation. With regard to prolonged periimplant inflammatory response, Anjum et al. (2016) presented an interesting working model. Cobalt ions released from arthroplasty activate TLR4 on immune cells and endothelial cells, resulting in cytokine and chemokine (including IL-8) release that leads among others to inflammatory cell recruitment from the circulation.

650

Were our patients with “standard”, i.e., uncoated, TKR and high IL-8 levels at risk of earlier revision? When planning the study—yet unaware of such results—we had wondered if also some specific, i.e., T-cell related, marker or cytokines reflecting regulatory mechanism were demonstrable in our patients. For the latter aspects we assessed IL-10 for the following reasons: Martire-Greco et al. (2013) had reported that addition of IL-10 reduced LPS-induced TNFa- and IL-8 secretion of human neutrophils in vitro; IL-10 was found to reflect a potent anti-inflammatory, immune-dampening response for example in pregnancy allowing “fetal tolerance” (Santner-Nanan et al. 2013), in successful immunotherapy with pollen allergens (Jutel and Akdis 2011) or symptom-free metal (titanium) dental implant patients (Thomas et al. 2013). Apart from its tolerogenic role IL-10 can also play a role in immune-mediated diseases by promoting autoantibody production and cytotoxic T-cell responses (Geginat et al. 2016). Even if we tend to consider the here observed IL-10 production as immunoregulatory, such a pivotal role of IL-10 together with enhanced IL-8 production needs attention in future investigations. Cassuto et al. (2017) have recently reported their results of repeated measuring of plasma biomarkers in hip arthroplasty over a period of almost 2 decades. Among their findings, there was a striking peak of IL-8 around 5–6 years post-surgery and a late rise of RANKL. In this study the focus was set on bone remodeling/ healing phenomena. Nevertheless such a late proinflammatory mediator peak accords with our findings. Unfortunately IL-10 levels were not assessed by Cassuto et al. (2017). Since CXCL10 is described as a chemoattractant for the recruitment of activated T-cells (Wang et al. 2016), i.e., thus prolonging or increasing adaptive immune response, we assessed its serum levels as well. The 2 arthroplasty patient groups had more elevated levels with the highest, but statistically non-significant, increase in uncoated arthroplasty. Expression of Foxp3 as a marker of regulatory T-cells could again be found in both patient groups, with not significantly higher values in the uncoated TKR group. Immune reactivity to implanted metal arthroplasty is an area of growing research intensity. At the same time the understanding of immunological responses to it and of cytokine homeostasis in such patients still remains incomplete. Our study, to our knowledge, is the first comparative analysis of serum cytokine responses and clinical performance in patients with standard and surface-coated TKR. The observed differences, i.e., the proinflammatory yet clinically inapparent and controlled condition in the standard TKR patient, should trigger further work to identify additional factors contributing to implant failure in given patients. Subsequent studies should include blood sampling at more time points and also followup study examinations with the 10- and 15-year results should be planned. In summary, we could observe differences in cytokine patterns between patients with coated and uncoated knee joint implants but with similar clinical outcomes at the 5-year fol-

Acta Orthopaedica 2018; 89 (6): 646–651

low-up. Identifying the complex interplay of individual “biomediators” may add to the understanding of why implants might fail. Thus, the long-term cumulative effect of those cytokine changes should be a matter for further research.

The authors thank Prof. Lindenmayr for initiating the coated TKA concept at Memmingen Hospital. TP: study leader, manuscript preparation; PH, HK, US, AO, DA: clinical work, patient recruitment, and examination (in the different study centers); SB: cytokine measurement, data analysis, statistics; SC: clinical work, patient examination, manuscript discussion. Acta thanks Christian P Delaunay and Stephan Söder for help with peer review of this study.

Anjum S A, Lawrence H, Holland J P et al. Effect of cobalt-mediated Tolllike receptor 4 activation on inflammatory responses in endothelial cells. Oncotarget 2016; 7(47): 76471-8. Bader R, Bergschmidt P, Fritsche A, et al. Alternative materials and solutions in total knee arthroplasty for patients with metal allergy. Orthopade 2008; 37(2): 136-42. Bechtel C P, Gebhart J J, Tatro J M et al. Particle-induced osteolysis is mediated by tirap/mal in vitro and in vivo: dependence on adherent pathogenassociated molecular patterns. J Bone Joint Surg Am 2016; 98(4): 285-94. Bolognesi M P, Ledford C K. Metal-on-metal total hip arthroplasty: patient evaluation and treatment. J Am Acad Orthop Surg 2015; 23(12): 724-31. Bourne R B, Chesworth B, Davis A et al. Comparing patient outcomes after THA and TKA: is there a difference? Clin Orthop Relat Res 2010; 468(2): 542-6. Cassuto J, Folestad A, Gothlin J et al. The key role of proinflammatory cytokines, matrix proteins, RANKL/OPG and Wnt/beta-catenin in bone healing of hip arthroplasty patients. Bone 2017; 107: 66-77. Ettinger M, Calliess T, Kielstein J T, et al. Circulating biomarkers for discrimination between aseptic joint failure, low-grade infection, and high-grade septic failure. Clinical Infect Dis 2015; 61(3): 332-41. Fang Y, Saiyin H, Zhao X et al. IL-8-positive tumor-infiltrating inflammatory cells are a novel prognostic marker in pancreatic ductal adenocarcinoma patients. Pancreas 2016; 45(5): 671-8. Geginat J, Larghi P, Paroni M, et al. The light and the dark sides of Interleukin-10 in immune-mediated diseases and cancer. Cytokine Growth Factor Rev 2016; 30: 87-93. Granchi D, Cenni E, Giunti A et al. Metal hypersensitivity testing in patients undergoing joint replacement: a systematic review. J Bone Joint Surg Br 2012; 94(8): 1126-34. Hallab N. Metal sensitivity in patients with orthopedic implants. J Clin Rheumatol 2001; 7(4): 215-18. Hallab N J, Jacobs J J. Chemokines associated with pathologic responses to orthopedic implant debris. Front Endocrinol 2017; 8: 5. Hutt J, Lavigne M, Lungu E, et al. Comparison of whole-blood metal ion levels among four types of large-head, metal-on-metal total hip arthroplasty implants: a concise follow-up, at five years, of a previous report. J Bone Joint Surg Am 2016; 98(4): 257-66. Jämsen E, Kouri V P, Ainola M, et al. Correlations between macrophage polarizing cytokines, inflammatory mediators, osteoclast activity, and tolllike receptors in tissues around aseptically loosened hip implants. J Biomed Mater Res Part A 2017; 105(2): 454-63. Jin W J, Xu J M, Xu W L et al. Diagnostic value of interleukin-8 in colorectal cancer: a case-control study and meta-analysis. World J Gastroenterol 2014; 20(43): 16334-42.

Acta Orthopaedica 2018; 89 (6): 646â&#x20AC;&#x201C;651

Jutel M, Akdis C A. Immunological mechanisms of allergen-specific immunotherapy. Allergy 2011; 66(6): 725-32. Lawrence H, Deehan D, Holland J et al. The immunobiology of cobalt: demonstration of a potential aetiology for inflammatory pseudotumours after metal-on-metal replacement of the hip. Bone Joint J 2014; 96-B(9): 11727. Lee D H, Lee D K, Shin Y S et al. Mid-term outcomes of floating platform mobile-bearing total knee arthroplasty under navigational guidance with a minimum 4-year follow-up. J Arthroplasty 2013; 28(10): 1801-5. Lin C H, Chung C A, Wong J H, et al. Involvement of L-type Ca(2)(+) channel and toll-like receptor-4 in nickel-induced interleukin-8 gene expression. Environ Toxicol 2016; 31(1): 5-12. Mahendra G, Pandit H, Kliskey K et al. Necrotic and inflammatory changes in metal-on-metal resurfacing hip arthroplasties. Acta Orthop 2009; 80(6): 653-9. Martire-Greco D, Rodriguez-Rodrigues N, Landoni V I et al. Interleukin-10 controls human peripheral PMN activation triggered by lipopolysaccharide. Cytokine 2013; 62(3): 426-32. Matz M, Lorkowski C, Fabritius K, et al. The selective biomarker IL-8 identifies IFTA after kidney transplantation in blood cells. Transplant Immunol 2016; 39: 18-24. Morlock M M. The taper disaster: how could it happen? Hip Int 2015; 25(4): 339-46. Ninomiya J T, Kuzma S A, Schnettler T J et al. Metal ions activate vascular endothelial cells and increase lymphocyte chemotaxis and binding. J Orthop Res 2013; 31(9): 1484-91. Nobili V, Marcellini M, Giovannelli L, et al. Association of serum interleukin-8 levels with the degree of fibrosis in infants with chronic liver disease. J Pediatr Gastroenterol Nutr 2004; 39(5): 540-4. Palena C, Hamilton D H, Fernando R I. Influence of IL-8 on the epithelialmesenchymal transition and the tumor microenvironment. Future Oncol 2012; 8(6): 713-22. Peake J M, Della Gatta P, Suzuki K et al. Cytokine expression and secretion by skeletal muscle cells: regulatory mechanisms and exercise effects. Exercise Immunol Rev 2015; 21: 8-25. Reito A, Lainiala O, Elo P et al. Prevalence of failure due to adverse reaction to metal debris in modern, medium and large diameter metal-on-metal hip replacements: the effect of novel screening methods: systematic review and metaregression analysis. PLoS One 2016; 11(3): e0147872.

651

Santner-Nanan B, Straubinger K, Hsu P, et al. Fetal-maternal alignment of regulatory T cells correlates with IL-10 and bcl-2 upregulation in pregnancy. J Immunol 2013; 191(1): 145-53. Sharkey P F, Hozack W J, Rothman R H et al. Insall Award paper. Why are total knee arthroplasties failing today? Clin Orthop Relat Res 2002; (404): 7-13. Signorelli S S, Fiore V, Malaponte G. Inflammation and peripheral arterial disease: the value of circulating biomarkers (Review). Int J Mol Med 2014; 33(4): 777-83. Tan Y, Qi Q, Lu C, et al. Cytokine imbalance as a common mechanism in both psoriasis and rheumatoid arthritis. Mediators Inflamm 2017; 2017: 2405291. Tanaka R, Yasunaga Y, Hisatome T et al. Serum interleukin 8 levels correlate with synovial fluid levels in patients with aseptic loosening of hip prosthesis. J Arthroplasty 2005; 20(8): 1049-54. Thomas P, Braathen L R, Dorig M, et al. Increased metal allergy in patients with failed metal-on-metal hip arthroplasty and peri-implant T-lymphocytic inflammation. Allergy 2009; 64(8): 1157-65. Thomas P, Iglhaut G, Wollenberg A et al. Allergy or tolerance: reduced inflammatory cytokine response and concomitant IL-10 production of lymphocytes and monocytes in symptom-free titanium dental implant patients. Biomed Res Int 2013; 2013: 539834. Thomas P, Summer B. Diagnosis and management of patients with allergy to metal implants. Expert Rev Clin Immunol 2015; 11(4): 501-9. Wang M, Ketheesan N, Peng Z. Investigations of wear particles and selected cytokines in human osteoarthritic knee joints. Proc Inst Mech Eng H 2014; 228(11): 1176-82. Wang F, Cai R, He D et al. Serum IFN-gamma-inducible chemokines CXCL9 and CXCL10 are elevated in nonimmediate drug hypersensitivity reactions. Asian Pac J Allergy Immunol 2016; 34(3): 236-41. Wengler A, Nimptsch U, Mansky T. Hip and knee replacement in Germany and the USA: analysis of individual inpatient data from German and US hospitals for the years 2005 to 2011. Dtsch Arztebl Int 2014; 111(23-24): 9. Yang C A, Chiang B L. Inflammasomes and human autoimmunity: a comprehensive review. J Autoimmunity 2015; 61: 1-8.

652

Acta Orthopaedica 2018; 89 (6): 652–655

Pelvic obliquity and measurement of hip displacement in children with cerebral palsy Gunnar HÄGGLUND 1,2, Mikael GOLDRING 1, Maria HERMANSON 3, and Elisabet RODBY-BOUSQUET 1,4 1 Department

of Clinical Sciences, Lund University, Lund; 2 Department of Orthopaedics, Skane University Hospital, Lund; 3 Department of Surgery, Sahlgrenska University Hospital/Östra, Göteborg; 4 Centre for Clinical Research, Uppsala University, Region Västmanland, Västerås, Sweden Correspondence: gunnar.hagglund@med.lu.se Submitted 2018-05-10. Accepted 2018-07-23.

Background and purpose — Pelvic obliquity, common in individuals with cerebral palsy (CP), changes the muscle force vector on the hip joint and probably affects the risk of hip dislocation. We evaluated a new method for measurement of hip displacement in CP that takes the pelvic obliquity into account: the pelvic adjusted migration percentage (PAMP). Children and methods — From the Swedish surveillance program for cerebral palsy (CPUP), the first pelvic radiograph of 268 children < 18 years in southern Sweden during a 3-year period were evaluated. Pelvic obliquity, PAMP, and the migration percentage (MP) were measured. 50 radiographs were randomly selected for analysis of interrater reliability by three raters using the intraclass correlation coefficient (ICC). The correlations between PAMP/MP and pelvic obliquity were analyzed with Pearson correlation coefficients. Results — The interrater reliability for all 3 measurements was high (ICCs 0.88–0.97). The correlation between the high side of the pelvic obliquity and the difference between right and left hip displacement was higher for PAMP (r = 0.70) than for MP (r = 0.41). Interpretation — The new PAMP measurement showed high interrater reliability and a higher correlation with pelvic obliquity than MP. We suggest the use of PAMP at least in hips with a pelvic obliquity exceeding 5°.

Hip displacement in cerebral palsy (CP) is caused by abnormal forces of the femoral head on the acetabulum (Kalen and Bleck 1985). Spasticity, muscle weakness, and/or poor positioning are the underlying causes of the altered magnitude and direction of the forces (Kalen and Bleck 1985, Miller et al. 1999). In a computer modelling analysis, Miller et al. (1999) showed that in a normal hip the force of the hip muscles is

directed to the superomedial part of the acetabulum. In a spastic hip the force is increased, and the imbalance causes an abnormal position of the hip in adduction and internal rotation. This ‘spastic position’ of the hip shifts the direction of forces to the posterosuperolateral aspect of the acetabulum. Because of the changed force direction, the femoral head is at risk for lateral displacement, usually measured as an increased migration percentage (MP) (Reimers 1980) (Figure 1a). Pelvic obliquity (PO) is a common deformity in individuals with CP (Porter et al. 2007). PO can be measured as the angle between the horizontal plane and Hilgenreiner’s line, or a line between the acetabular teardrops, the lowest prominence of the ischial tuberosities, or the highest prominence of the iliac crest (Figure 1b). The hip on the high side of the PO is positioned in adduction with correspondingly more lateral direction of the muscle force vector. The acetabulum on the high side becomes more vertical, which uncovers the femoral head. The hip on the lower side is in abduction and the acetabulum is more horizontal with increased coverage of the femoral head. MP is the gold standard for measuring hip displacement and is used in most hip surveillance programs (Dobson et al. 2002, Hägglund et al. 2014). In MP, the lateral displacement is related to Hilgenreiner’s line. This means that the measurement of MP is not related to a PO. Therefore, we evaluated a new measurement method of hip displacement that accounts for the PO, a pelvic adjusted migration percentage (PAMP), by relating the displacement to a horizontal line instead of Hilgenreiner’s line (Figure 1c).

Children and methods In CPUP—the Swedish surveillance program for CP—more than 95% of all children with CP are followed with standardized repeated examinations (Alriksson-Schmidt et al. 2017). The gross motor function is classified by the child’s physio-

Acta Orthopaedica 2018; 89 (6): 652–655

Figure 1a. Pelvic obliquity in a 12-year-old girl with cerebral palsy in GMFCS level IV. Measurement of migration percentage (MP). MP = a/b × 100. MP = 46% in the left hip and 32% in the right hip.

653

Figure 1b. Measurement of pelvic obliquity (PO). PO = 15°, left side elevated.

therapist according to the Gross Motor Function Classification System (GMFCS) (Palisano et al. 1997). GMFCS is an agerelated, 5-level system in which children at level I are the least affected. CPUP also includes standardized repeated anteroposterior pelvic radiographic examinations (Hägglund et al. 2014). Children at GMFCS levels III–V are examined radiographically once a year up to 8 years of age and those at level II are examined at 2 and 6 years of age. Children at level I are not examined radiographically provided that the physiotherapist’s reports show a normal pain-free range of hip motion. After 8 years of age, the children are followed-up individually based on the previous radiographic and clinical findings. In the present study, all radiographs of children < 18 years in southern Sweden (Skåne and Blekinge with 1.4 million inhabitants) during the 3-year period from July 1, 2014 to June 31, 2017 were analyzed. In children with repeated radiographic examinations during the study period, the first radiograph was used. Children operated for scoliosis or with varus osteotomy of the proximal femur before the first examination were excluded. On all radiographs, PO, MP, and PAMP were measured. The difference in hip displacement between the right and left hip, measured as both MP and PAMP, was analyzed in relation to the degree of PO. 50 radiographs were randomly selected for analysis of interrater reliability. 3 raters (GH, MG, MH) independently measured the PO, MP, and PAMP. All measurements were sent separately to a 4th blinded person (ERB) for statistical analyses. Statistics Interrater reliability was evaluated using the intraclass correlation coefficient (ICC) and 95% confidence interval (CI) with a 2-way random model, and absolute agreement for single measures (McGraw and Wong 1996). Scatter plots with fitted

Figure 1c. Measurement of pelvic adjusted migration percentage (PAMP). PAMP = 59% in the left hip and 15% in the right hip.

values for the difference between right and left hip in hip displacement, measured as MP, PAMP, and PO, were prepared. Correlations between MP/PAMP and PO were investigated with Pearson correlation coefficients (r) and 95% bootstrap CIs were added based on 1,000 bootstrap samples. The result was interpreted according to Evans (1996); r = 0.00–0.19, negligible; 0.20–0.39, weak; 0.40–0.59, moderate; 0.60–0.79, strong; and 0.90–1.00, very strong correlation. The statistical analyses were performed using IBM SPSS Statistics version 24 (IBM Corp, Armonk, NY, USA). Ethics, funding, and potential conflicts of interest The study was approved by the Medical Research Ethics Committee at Lund University (LU-443-99). The study was funded by Linnea and Josef Carlssons foundation and Stiftelsen för bistånd åt rörelsehindrade i Skåne. The authors declare no conflicts of interest.

Results During the 3-year period, 307 children (167 boys) were examined radiographically. Of these, 39 children were excluded: 32 children (17 girls) had been operated with varus osteotomy; 6 children (4 girls) had received surgery for scoliosis; and 1 girl was operated with both varus osteotomy and spine surgery before the first examination. The remaining 268 individuals were included in the analysis. The age at examination and GMFCS distribution are presented in Table 1. There was high interrater reliability for all 3 measurements, with the lower levels of the 95% CI above 0.7 (Table 2). The highest ICC was seen for PO measurement (ICC 0.97). Measurement of PAMP showed slightly higher intraclass correlation coefficients for both the right and the left side (ICC 0.91; 0.93) than MP (ICC 0.88; 0.89).

654

Acta Orthopaedica 2018; 89 (6): 652–655

Table 1. Distribution of age and Gross Motor Function Classification System (GMFCS) levels Age (years) < 3 3–5 6–8 9–11 12–14 15–18 Total

GMFCS level III IV

Table 3. Difference between PAMP and MP on the high side of pelvic obliquity Pelvic obliquity (degrees)

Total

18 8 10 17 53 18 14 17 14 63 17 9 15 12 53 5 12 15 10 42 3 7 15 8 33 1 10 8 5 24 62 60 80 66 268

ICC

95% CI

Pelvic obliquity MP right MP left PAMP right PAMP left

0.97 0.88 0.89 0.91 0.93

0.96–0.98 0.78–0.94 0.78–0.94 0.83–0.95 0.86–0.96

11 9 (0 to 24)

cases. The difference between PAMP and MP on the high side of the PO increased with degree of obliquity (Table 3). There was a statistically significant (p < 0.001) correlation between the MP or PAMP in the right minus the left hip and the PO. Pearson’s correlation coefficient was 0.41 (CI 0.28–0.54) for the comparison with MP and 0.70 (CI 0.61–0.79) for PAMP (Figures 3 and 4).

Discussion The difference between the new (PAMP) and traditional (MP) method of measuring hip displacement in CP is the use of a horizonal line instead of Hilgenreiner’s line. This study showed that PAMP had a slightly higher interrater reliability, and a substantially higher correlation with PO than the MP. When the reference points for Hilgenreiner’s line are indistinct, it is easier to draw a horizontal line. Heidt et al. (2015) analyzed PO in 98 individuals with CP at 19 years (15–24) of age. They found a high inter- and intrarater reliability for PO measurement and a strong correlation between PO and hip morphology. PO and PAMP depend on the position of the child on the examination table. The table is often narrow and it seems reasonable that the child is positioned in the middle line of

In the cohort of 268 children, the PO ranged from 0 to 18 degrees. The left side was higher in 86 cases, the right in 126, and in 56 radiographs the pelvis was neutral (Figure 2). The MP/PAMP was higher on the high side of the PO in most cases. Of the 64 radiographs with PO > 5 degrees, 51 had a higher MP and 61 a higher PAMP on the high side compared with the lower side of the PO. The MP was equal on both hips in 3 cases and lower on the high side of the PO in 10 cases. The PAMP was equal on both hips in 1 case and lower in 2

For abbreviations, see Table 2

MP: migration percentage; PAMP: pelvic adjusted migration percentage; p < 0.001 for all measurements.

Number of radiographs

≥10

5–9

Number of radiographs 148 57 PAMP–MP mean difference (90% percentiles) 1.9 (0 to 7) 6.5 (0 to 14) median difference (range) 0 (–4 to 10) 6 (–5 to 16)

Table 2. Interrater reliability estimated by intraclass correlation coefficient (ICC) Factor

1–4

MP right minus left hip (%)

PAMP right minus left hip (%)

100

–50

50 40 30 20 10

–100 >9

–8

–6

–4

–2

Pelvic obliquity (°)

Figure 2. Number of pelvic radiographs related to the degree of PO. Negative value = left side of pelvis elevated, positive value = right side of pelvis elevated.

–100 –20

–10

–20

–10

Pelvic obliquity (°)

Figure 3. Difference between MP in the right and left hip related to PO. Negative PO value = left side of pelvis elevated, positive PO value = right side of pelvis elevated.

Figure 4. Difference between PAMP on the right and left hip related to PO. Negative PO value = left side of pelvis elevated, positive PO value = right side of pelvis elevated.

Acta Orthopaedica 2018; 89 (6): 652–655

the table. The CPUP manual states that the child should be positioned with the legs straight. The radiographic examinations in this study were performed in 11 different radiology departments. Therefore, the results are likely to represent the situation in routine care and could be generalized to hip surveillance programs in different settings. Porter et al. (2007) found no correlation between the side of hip displacement, defined as MP > 33%, and high/low side of PO in a study of 747 individuals with CP. Lonstein and Beck (1986) reported similar results with no correlation between the high side of the PO and the side of hip displacement, defined as MP > 33%, and no increase in hip displacement with increasing PO among 304 individuals with CP classified as dependent sitters. In both these studies, PO was measured in the sitting position. A PO caused by asymmetric hip abduction is only present in the lying or standing position. In the sitting position, restricted abduction will rotate the pelvis backwards, or the legs will deviate into adduction, but it cannot force the pelvis into PO in the sitting position, which might explain the absence of correlation found in these studies. Letts et al. (1984) did a longitudinal follow-up study of all radiographs, both sitting and supine, from infancy to the teen ages in 22 children with hip displacement, PO, and scoliosis. All hip displacements occurred on the high side of the PO. There are limitations to this study. This was a descriptive and psychometric evaluation of a new measurement of hip displacement based on radiographs from 1 occasion for each child. To demonstrate a stronger causal relationship between PAMP and hip dislocation than with MP and hip dislocation, a longitudinal study comparing the 2 measures as risk factors is required. However, with knowledge of the cause of hip dislocation and the force vectors of the hip joint as described by Miller et al. (1999), there is theoretical support that PAMP better reflects the risk of hip dislocation. On radiographs with no or small PO, the MP is equal to the PAMP. In higher degrees of PO, MP probably underestimates the risk of hip dislocation on the high side and overestimates the risk on the lower side. No child in this series had a PO that exceeded 18 degrees. Therefore, we cannot make conclusions on the relationship between MP and PAMP in cases with larger pelvic asymmetries. The results of this study were based on all children with CP in the area examined in the CPUP program. However, hips operated with varus osteotomy and spinal surgery were excluded; therefore, the study was made on a selected cohort with relatively few children in GMFCS V. In summary, measurement of hip displacement using the new PAMP showed a high interrater reliability, and a higher

655

correlation with PO than the regular measurement of MP. The measurement of PAMP is in agreement with known consequences of forces on the hip joint and is likely to provide a more accurate reflection of the risk of hip dislocation in hips with a PO. The PAMP has the potential of being more accurate in hip surveillance. We suggest the use of PAMP instead of regular MP at least in hips with a pelvic obliquity exceeding 5 degrees.

Study design: GH, MG, MH, ERB. Data collection: GH. Data analysis: GH, MG, MH, ERB. Manuscript preparation: GH, MG, MH, ERB. Acta thanks Freeman Miller and Terje Terjesen for help with peer review of this study.

Alriksson-Schmidt A I, Arner M, Westbom L, Krumlinde-Sundholm L, Nordmark E, Rodby-Bousquet E, et al. A combined surveillance program and quality register improves management of childhood disability. Disabil Rehabil 2017; 39: 830–6. Dobson F, Boyd R N, Parrott J, Nattrass G R, Graham H K. Hip surveillance in children with cerebral palsy: impact on the surgical management of spastic hip disease. J Bone Joint Surg Br 2002; 84: 720–6. Evans J D. Straightforward statistics for the behavioral sciences. Pacific Grove, CA: Brooks/Cole Publishing; 1996. Hägglund G, Alriksson-Schmidt A, Lauge-Pedersen H, Rodby-Bousquet E, Wagner P, Westbom L. Prevention of dislocation of the hip in children with cerebral palsy: 20-year results of a population-based prevention programme. Bone Joint J 2014; 96-B: 1546–52. Heidt C, Hollander K, Wawrzuta J, Molesworth C, Willoughby K, Thomason P, et al. The radiological assessment of pelvic obliquity in cerebral palsy and the impact on hip development. Bone Joint J 2015; 97-B: 1435–40. Kalen V, Bleck E E. Prevention of spastic paralytic dislocation of the hip. Dev Med Child Neurol 1985; 27: 17–24. Letts M, Shapiro L, Mulder K, Klassen O. The windblown hip syndrome in total body cerebral palsy. J Pediatr Orthop 1984; 4: 55–62. Lonstein J E, Beck K. Hip dislocation and subluxation in cerebral palsy. J Pediatr Orthop 1986; 6: 521–6. McGraw K O, Wong S P. Forming inferences about some intraclass correlation coefficients. Psychological Methods 1996; 1: 30–46. Miller F, Slomczykowski M, Cope R, Lipton G E. Computer modeling of the pathomechanics of spastic hip dislocation in children. J Pediatr Orthop 1999; 19: 486–92. Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi B. Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dev Med Child Neurol 1997; 39: 214–23. Porter D, Michael S, Kirkwood C. Patterns of postural deformity in nonambulant people with cerebral palsy: what is the relationship between the direction of scoliosis, direction of pelvic obliquity, direction of windswept hip deformity and side of hip dislocation? Clin Rehabil 2007; 21: 1087–96. Reimers J. The stability of the hip in children: a radiological study of the results of muscle surgery in cerebral palsy. Acta Orthop Scand 1980; (Suppl 184): 1–100.

656

Acta Orthopaedica 2018; 89 (6): 656–661

Knee joint sagittal plane movement in cerebral palsy: a comparative study of 2-dimensional markerless video and 3-dimensional gait analysis Evelina PANTZAR-CASTILLA 1, Andrea CEREATTI 2,3, Giulio FIGARI 2,4, Nicolò VALERI 3, Gabriele PAOLINI 4, Ugo DELLA CROCE 2,4, Anders MAGNUSON 5, and Jacques RIAD 6 1 Department 3 Department

of Orthopedics, Örebro University Hospital, Örebro, Sweden; 2 Department of Biomedical Sciences, University of Sassari, Sassari, Italy; of Electronics and Telecommunications, Politecnico di Torino, Torino, Italy; 4 GPEM srl, Alghero, Italy; 5 Clinical Epidemiology and Biostatistics, School of Medical Sciences, Örebro University, Örebro, Sweden; 6 Department of Orthopedics, Skaraborg Hospital Skövde and Mölndal Hospital Sahlgrenska, Gothenburg, Sweden Correspondence: evelina.pantzar-castilla@regionorebrolan.se Submitted 2018-05-14. Accepted 2018-08-10.

Background and purpose — Gait analysis is indicated in children with cerebral palsy (CP) to identify and quantify gait deviations. One particularly difficult-to-treat deviation, crouch gait, can progress in adolescence and ultimately limit the ability to ambulate. An objective quantitative assessment is essential to early identify progressive gait impairments in children with CP. 3-dimensional gait analysis (3D GA) is considered the gold standard, although it is expensive, seldom available, and unnecessarily detailed for screening and follow-up. Simple video assessments are time-consuming when processed manually, but more convenient if used in conjunction with video processing algorithms; this has yet been validated in CP. We validate a 2-dimensional markerless (2D ML) assessment of knee joint flexion/extension angles of the gait cycle in children and young adults with CP. Patients and methods — 18 individuals, mean age 15 years (6.5–28), participated. 11 had bilateral, 3 unilateral, 3 dyskinetic, and 1 ataxic CP. In the Gross Motor Function Classification System, 6 were at level I, 11 at level II, and 1 at level III. We compared 2D ML, using a single video camera with computer processing, and 3D GA. Results — The 2D ML method overestimated the knee flexion/extension angle values by 3.3 to 7.0 degrees compared with 3D GA. The reliability within 2D ML and 3D GA was mostly good to excellent. Interpretation — Despite overestimating, 2D ML is a reliable and convenient tool to assess knee angles and, more importantly, to detect changes over time within a follow-up program in ambulatory children with CP.

Gait analysis is indicated in children with cerebral palsy (CP) to identify gait deviations, evaluate treatment, and follow the natural history (Bell et al. 2002). One of the most critical and frequent gait deviations in bilateral CP is crouch gait, defined as the inability to fully extend the knee in stance (Gage et al. 1996, Rodda and Graham 2001, Miller 2005). Even if crouch gait is tolerated in the young child, in adolescence with growth, it may progress and ultimately limit the ability to ambulate (Opheim et al. 2013). Several gait variables are relevant in crouch gait (Perry 1992, Benedetti et al. 1998, Sangeux and Armand 2015) (Figure 1). Even if there is treatment for crouch gait the challenge is to identify progressive gait impairments in children with CP early and choose effective treatment (DeLuca 1991, Gage et al. 1996, Rodda et al. 2004, Narayanan 2007). Hence, a systematic quantitative description of gait function is essential (Perry 1992, Benedetti et al. 1998, Sangeux and Armand 2015). Marker-based 3-dimensional gait analysis (3D GA) provides a dynamic assessment of gait and is considered the gold standard to quantify lower limb movements, however seldom available since it is expensive and requires a dedicated space (Eastlack et al. 1991, Perry 1992, Mackey et al. 2003). In addition, for screening purposes and for follow-up, the level of detail provided by 3D GA might be unnecessary; knee-joint sagittal angles may suffice as a valid and useful preliminary assessment (Maathuis et al. 2005, Kawamura et al. 2007, Narayanan 2007). To quantify and partially reduce examiner subjectivity, visual scoring systems have been proposed (Narayanan 2007). The Edinburgh Visual Gait Score was specifically developed for assessing gait in CP and has proven to have good reliability and validity (Read et al. 2003, Rathinam et al. 2014, Del Pilar Duque Orozco et al. 2016). However, joint angles, used to score gait, require manual identification of specific anatomi-

Acta Orthopaedica 2018; 89 (6): 656–661

657

Laboratory, Skövde. 3 participants were not able to perform the gait analysis due to cognitive and visual impairments, and 2 were excluded for technical reasons. Gait data were collected for 18 participants, mean age 15 years (6.5–28), 4 females and 14 males. 11 participants had bilateral CP, 3 had unilateral CP, 3 had dyskinetic CP, and 1 had ataxic CP. 6 were classified as GMFCS level I, 11 were at level II, and 1 was at level III.

Figure 1. The 4 knee-flexion and extension variables selected for the analysis (stance and swing phase in percentage of the gait cycle on the X-axis). 1: Knee flexion at initial contact (0% of the gait cycle) important for step length, which can be limited by hamstring spasticity and/or short hamstring muscles; 2: maximum knee flexion at loading response (0–40% of the gait cycle) manages force absorption; 3: minimum knee flexion in stance (25–75% of the gait cycle) describes degree of crouch; 4: and maximum knee flexion in swing (50–100% of the gait cycle) contributes to foot clearance.

cal landmarks on multiple videotaped sequences. Clearly, this operation can be quite time-consuming and heavily relies on the operator’s skills and experience (Read et al. 2003, Ong et al. 2008, Gupta and Raja 2012, Del Pilar Duque Orozco et al. 2016). Some of the limitations associated with visual video assessments can be overcome by using automatic video processing (Ugbolue et al. 2013, Castelli et al. 2015, Saner et al. 2017). However, most of the 2-dimensional markerless (2D ML) gait analysis methods proposed in the literature have been applied and validated only on healthy individuals (Surer et al. 2011, Sandau et al. 2014, Castelli et al. 2015, Saner et al. 2017); therefore, the results cannot be generalized to populations with gait disorders, as deviations from typical gait patterns might interfere with the performance and feasibility of the method. To address the above-mentioned clinical requirements, we developed a 2D ML method which takes advantage of a singlecolor red–green–blue (RGB) camera combined with a depth infrared sensor (RGB-D). The proposed method is an improvement of a recent markerless method developed and validated on healthy adults (Castelli et al. 2015, Saner et al. 2017). We evaluated the validity of this newly developed 2D ML gait analysis method compared with 3D GA for the estimation of knee flexion and extension angles during gait in children and young adults with CP.

Patients and methods Population From the medical records in Örebro and Skaraborg, individuals with CP were identified. Those eligible for inclusion were sent a letter with information and invitation to participate. Inclusion criteria were CP and GMFSCS level I–III. 23 were willing to participate, and visited the Skaraborg Gait Analysis

Data collection The participants walked at a self-selected speed on a 7-meter walkway wearing a T-shirt, underwear, and colored ankle socks. The RGB-D camera was positioned laterally to the walkway, and a homogeneous green background was rigged. The image coordinate system of the video camera was aligned to the sagittal plane. The data acquisition and synchronization was performed simultaneously for 2D ML and 3D GA. 5 gait trials per participant and side were recorded. 2-dimensional markerless gait analysis (2D ML) An RGB-D system was used (Kinect 2 for Xbox One [Microsoft Corp, Redmond, WA, USA]), RGB images with resolution 1920×1080 pixels at 30 frames/second and depth image of 512×424 pixels at 30 frames/second. The estimation of knee flexion and extension angles required the implementation of the following steps (see Appendix in Supplementary data for a detailed description): 1. Image pre-processing calibration: includes optical distortion correction and allows RGB and depth data to obtain a matrix of colored 3D pixels for each frame (Figure 2a). 2. Segmentation: subtraction to separate the participant from the background (Figure 2b). 3. Participant-specific multi-segmental model: a lower-limb kinematic model composed of the foot, shank, and thigh segments connected by cylindrical hinges. To compensate for real and apparent deformations of the body segments, a multiple anatomical landmark calibration was implemented (Figure 2c). 4. Joint center tracking: trajectories of ankle, knee, and hip joints centers were estimated during the gait cycle by minimizing the distance between the measured and modelled body-segment points (Figure 2d). Depth data were exploited to handle segments overlapping during walking. 5. Flexion and extension knee joint angle: orientation of the tibia and femur were determined from joint center trajectories and expressed as percentage of the gait cycle. 3-dimensional marker-based gait analysis (3D GA) 3D GA images were captured at 100 frames/second using a 12-camera stereo-photogrammetric system (Oqus 400 Qualisys medical AB, Gothenburg, Sweden). 38 retro-reflective spherical markers were attached to the participants according to the modified Helen-Heyes model. Calculation of joint angles was performed using the Visual 3D software (C Motion Inc., USA).

658

Acta Orthopaedica 2018; 89 (6): 656–661

Figure 2. Procedures needed for the estimation of the knee flexion and extension angles with the 2-dimensional markerless video system: image pre-processing calibration (a), segmentation (b), participant-specific multisegmental model (c), and joint-center tracking (d).

Knee joint gait variables For each participant and trial, the knee flexion and extension angles were estimated for 2D ML and 3D GA (see Figure 1). Time synchronization between 2D ML and 3D GA curves was performed. The following knee variables were extracted from the 2D ML and 3D GA curves: • Knee flexion at initial contact (0% of the gait cycle), is important to achieve an effective “prepositioning of the foot” and “adequate step length.” These are two prerequisites of normal gait, depending on the ability to extend the knee before initial contact in late swing phase (Perry 1992, Benedetti et al. 1998). • Maximum knee flexion at loading response (0–40% of the gait cycle), is controlled mainly by the quadriceps muscle with eccentric- and isometric muscle contraction to stabilize the knee, and for shock absorption (Sangeux and Armand, 2015). • Minimum knee flexion in stance (25–75% of the gait cycle), provides a measure of crouch gait and is influenced by fixed knee flexion, hamstring spasticity and/or short hamstring muscles, and foot stability, affecting a third prerequisite of normal gait: “stability in stance” (Perry 1992). • Maximum knee flexion in swing (50–100% of the gait cycle), may be limited and delayed by prolonged rectus femoris muscle activity in swing, influencing both “adequate step length” and “clearance in swing” (Perry 1992). The decreased knee flexion in swing is often presented as toe-drag, not to be confused with limited ankle dorsiflexion in swing (Sangeux and Armand 2015). Stastistics For each variable, the mean value of 3 trials was calculated for each participant and side. Normal distribution was assessed. Bland–Altman plots with 95% limits of agreements (LoA) were used to compare 2D ML and 3D GA methods for the selected variables. In the Bland–Altman plot, the Spearman correlation coefficient was used to further evaluate if differences between methods were correlated to the mean values of knee flexion. Unpaired t-test with 95% confidence intervals (CI) was used to quantify potential systematic mean differences between the 2D ML and 3D GA methods. For each vari-

able, method reliability (2D ML and 3D GA) was evaluated with intraclass correlation (ICC) based on absolute agreement and 2-way random effects. Based on the CI of the ICC estimate, values less than 0.5 indicate poor reliability, values between 0.5 and 0.75 indicate moderate reliability, values between 0.75 and 0.9 indicate good reliability, and values greater than 0.90 indicate excellent reliability (Koo and Li 2016). A p-value less than 0.05 were regarded as statistically significant, and the Bonferroni method was used to correct for multiple testing. All statistical analyses were performed in SPSS for Windows, version 22 (IBM Corp, Armonk, NY, USA). Ethics, funding, and potential conflict of interest The study was approved by the regional ethical review board in Gothenburg, Sweden (approval number 660-15). The projected was funded by national hospital research departments, FoU Region of Örebro Län and FoU Skaraborg and Vg Region. We have no financial or other conflicts of interest that might bias our work.

Results Overall, the mean statistical differences of the gait variables as estimated by 2D ML and 3D GA (2D ML minus 3D GA) ranged from +3.3 to +7.0 degrees (CI 1.3–9.0) (Table 1). These differences remained statistically significant after Bonferroni correction except for 2 gait variables (Initial contact and Loading response, left side) that, after Bonferroni correction, were not statistically significant using the Spearman correlation coefficient. The agreement of the 2D ML and 3D GA estimates for knee extension in stance, left side, and knee flexion in swing, left side, are illustrated with Bland–Altman plots (Figures 3 and 4). The reliability of 2D ML and 3D GA is reported in Table 2. The 2D ML measurements revealed moderate to excellent reliability, with ICC CI between 0.62 and 0.98. The 3D GA measurements showed similar reliability except for the variable Loading response, where the ICC CI was 0.41–0.89, indicating poor reliability. The resulting variables exhibited moderate to excellent results in the 3D

Acta Orthopaedica 2018; 89 (6): 656–661

659

Table 1. Variables, side, and mean angles for the 2-dimensional markerless (2D ML) method and 3-dimensional gait analysis (3D GA) Gait variables (°) and side

2D ML mean (SD)

3D GA Difference Correlation mean (SD) mean (95% CI) p-value a 95% LoA coefficient p-value b

Knee flexion at initial contact Right 26.1 (9.5) 19.1 (9.8) 7.0 (5.0–9.0) < 0.001 –0.9 to 14.8 0.16 Left 26.8 (10.6) 20.6 (8.5) 6.3 (4.2–8.2) < 0.001 –1.6 to 14.1 0.47 Knee flexion at loading response Right 32.4 (9.5) 26.3 (8.7) 6.1 (4.2–8.0) < 0.001 –1.1 to 13.4 0.22 Left 34.0 (9.3) 27.7 (7.1) 6.3 (4.2–8.2) < 0.001 –1.6 to 14.1 0.49 Knee flexion in stance Right 17.3 (11.7) 10.6 (12.5) 6.7 (4.9–8.3) < 0.001 –0.2 to 13.5 –0.19 Left 17.5 (9.1) 12.2 (9.1) 5.3 (3.1–7.6) < 0.001 –3.5 to 14.1 –0.052 Knee flexion in swing Right 60.3 (11.4) 55.4 (9.8) 5.0 (2.7–7.3) < 0.001 –4.1 to 13.9 0.26 Left 59.6 (8.4) 56.4 (7.2) 3.3 (1.3–5.2) < 0.001 –4.5 to 11.1 0.31

0.5 0.05 0.4 0.04 0.4 0.8 0.3 0.2

Mean difference calculated by subtracting the 3D GA angle from the 2D ML angle in degrees (SD, standard deviation; CI, confidence interval; LoA, limits of agreement). a p-value of the mean difference. b p-value of the Spearmen correlation coefficient.

Table 2. Comparison of estimates from the 2-dimensional markerless (2D ML) method and 3-dimensional gait analysis (3D GA) Gait variables 2D ML 3D GA Side ICC (95% CI) ICC (95% CI) Knee flexion at initial contact Right 0.85 (0.67–0.94) 0.87 Left 0.96 (0.92–0.96) 0.95 Knee flexion at loading response Right 0.83 (0.62–0.93) 0.73 Left 0.93 (0.84–0.97) 0.95 Knee flexion in stance Right 0.89 (0.74–0.96) 0.88 Left 0.93 (0.85–0.97) 0.97 Knee flexion in swing Right 0.92 (0.82–0.97) 0.90 Left 0.94 (0.87–0.98) 0.94

(0.70–0.95) (0.89–0.98)

Difference in minimum knee flexion in stance (2D ML – 3D GA)

Difference in maximum knee flexion in swing (2D ML – 3D GA)

15 Upper LoA 11.1

Mean 5.3

(0.73–0.95) (0.93–0.99)

–5

The 18 participants completed 3 trials for each side.

(0.41–0.89) (0.89–0.98)

(0.78–0.96) (0.87–0.98)

Upper LoA 14.1

Mean 3.3

0 Lower LoA –3.5

Lower LoA –4.5 –5

Mean minimum knee flexion in stance ((2D ML + 3D GA)/2)

Mean maximum knee flexion in swing ((2D ML + 3D GA)/2)

Figure 3. Bland–Altman plot for minimum knee flexion in stance on the left side (2D ML, 2-dimensional markerless; 3D GA, 3-dimensional gait analysis).

Figure 4. Bland–Altman plot for maximum knee flexion in swing on the left side (2D ML, 2-dimensional markerless 3D GA, 3-dimensional gait analysis).

GA estimates, with ICC CI between 0.70 and 0.98 (Koo and Li 2016).

Discussion We found a mean difference between the 2D ML and 3D GA estimates of knee joint angles ranging from + 3.3 to +7.0 degrees. For both the 2D ML and 3D GA methods, the withinmethod reliability was mostly good to excellent. The 95% LoA reveal how well the 2D ML and 3D GA methods agree. The LoA for all gait variables showed an expect-

edly large range (from 4.5 to 15 degrees), resulting from the high gait variability observed over the heterogeneous subject cohort analyzed. Several factors can explain the differences found in the knee estimated by 2D ML and 3D GA. First, the anatomical coordinate system definition for the femur and tibia differs between the methods. Second, we compared the knee flexion and extension angles obtained by projecting the participant movement in the sagittal plane (2D ML) with a 3D joint kinematic description based on Euler angle decomposition (3D GA). Third, 2D ML tracked the body segment movement based on the segment contours, whereas 3D GA is based on skin marker trajectories; therefore different soft tissue

660

artefacts should be expected, and their effects on the knee kinematics should be different. However, despite the unavoidable differences found between the methods, it is important to highlight that for both 2D ML and 3D GA the within-method reliability was mostly good to excellent. There is consensus in the literature that the knee joint is one of the more difficult to assess using the Edinburgh video Visual Gait Score reflecting a degree of agreement with the measurements obtained using a 3D GA of 47%–63% (Read et al. 2003) and 52–71% (del Pilar Duque Orozco et al. 2016). Both studies showed the usefulness of a video-based scoring system but with some limitations, as the previous experience of the observer and the training were both demonstrated to play a role in the agreement. Concurrent validity of most previous 2D ML methods has been assessed only on healthy participants (Sandau et al. 2014, Castelli et al. 2015, Saner et al. 2017). Sandau et al. (2014) applied their 2D ML method on 10 healthy adults and found a general overestimation of knee joint kinematics (mean error ± SD) equal to 2.8 ± (1.9) degrees for knee flexion and extension when compared against a 3D GA system. Castelli et al. (2015) found in 10 healthy participants an average root-mean-square difference between the knee joint angles estimated by 2D ML and 3D GA, over the entire gait cycle, of up to 4.4 degrees. In a recent study, Saner et al. (2017) proposed a method based on a low-cost webcam technology for knee sagittal kinematics estimation, which showed good to excellent reliability but lacked sufficient reliability during knee-stance. Unfortunately, no concurrent validation against a 3D GA system was performed. Furthermore, it should be noted that whereas participants in our study walked at a self-selected speed on a walkway with no constriction to arm movement, in the study reported by Saner et al. (2017) participants walked on a treadmill at a set speed and held their hands and arms up on a handrail, a setup that inevitably influences walking patterns and is difficult to apply to individuals with CP. There is no exact definition or classification of crouch. Miller (2005) proposed 3 different intervals: mild (10–25 degrees), moderate (26–45 degrees), and severe (over 45 degrees). Others defined severe crouch as over 30 degrees of flexion in mid-stance (Rodda and Graham 2001). From a clinical point of view, we think the 2D ML assessment can prove useful, for example in stance phase. A change from mild to moderate, or moderate to severe crouch, could be detected with the 2D ML system despite overestimating range of motion with a mean of 3.3–7.0 degrees, since the within-method reliability was mostly good to excellent. Certainly, using the 2D ML system, flexion increase of 10 degrees in mid-stance could be detected and was considered a significant change motivating a full 3D GA and a clinical evaluation for possible treatment to avoid progress of crouch. In Sweden, a systematic follow-up program for children with CP (CPUP) has been conducted since the 1990s (Hägg lund et al. 2005). According to the 2013 yearly CPUP report,

Acta Orthopaedica 2018; 89 (6): 656–661

61% of children with CP ambulate independently (GMFCS levels I and II) and 8% need support when walking (GMFCS level III) (CPUP yearly report 2013, Palisano et al. 1997). This means that ambulatory children constitute a substantial number where gait deterioration over time may be expected (Opheim et al. 2013). However, no dynamic assessment of gait in children with CP is currently included in the CPUP, and hence an objective tool to detect and follow the development of gait impairment is lacking. It is important to highlight that the purpose of the proposed 2D ML method is not to replace 3D GA, but rather to make an additional tool available to a greater number of children with CP, which could make longitudinal follow-up possible. A case report (Butler et al. 2016) and previously, Bell et al. (2002) showed that, using 3D GA, it was possible to follow and detect changes in the gait patterns in CP over several years. 3D GA has thus shown itself to be useful in follow-up, and therefore it is reasonable to assume that 2D ML can also be used for longterm follow-up and screening. A disadvantage with the 2D ML system is obviously the lack of frontal and rotational plane assessment, which is especially important in CP, where the rotational plane deformity often needs to be corrected with bony surgery. Nevertheless, when progression of deviation is identified with the 2D ML system, a more comprehensive assessment with 3D GA, as mentioned above, may be indicated. The 2D ML method we propose is composed of inexpensive, commercially available equipment and requires limited handson for processing. Overall, when the clinical interest is limited to kinematic analysis in the sagittal plane, the 2D ML system may represent a valid alternative to traditional 3D GA by making the gait analysis low-cost and avoiding the use of skin markers. The 2D ML method offers a reliable quantification of movement. It is a simpler and quicker measurement than 3D GA but, at the same time, more valuable than the visual analysis based on manual anatomical landmark digitalization. With the relatively subtle involvement of CP, we were unable to study whether those individuals with more severe gait deviations would show similar results. The rotational profile of the participants in this study was not assessed, which naturally could influence the results. We present, and assessed, the knee joint kinematics only, although the proposed 2D ML method can also estimate hip and ankle joint kinematics. However, a full validation of the lower limb kinematics was beyond the scope of our study. In summary, despite overestimating, 2D ML is a convenient tool that could be used to assess knee joint angles in the sagittal plane and, more importantly, detect changes over time in a follow-up program in ambulatory children with CP. Supplementary data Appendix is available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674. 2018.1525195

Acta Orthopaedica 2018; 89 (6): 656–661

EPC: Main investigator, assessment of medical records, recruitment of participants, performing statistical analysis, and writing the manuscript. AC: Organizing video assessments, interpretation of the 2D ML assessments, and writing the manuscript with special emphasis on methods. GF, GP, NV: Main responsibility for 2D ML video assessments. UDC: Participated in writing and editing the manuscript. AM: Participated in performing the statistical analysis and writing of the manuscript. JR: Planning and organization, recruitment of participants and writing the manuscript

Bell K J, Ounpuu S, DeLuca P A, Romness M J. Natural progression of gait in children with cerebral palsy. J Pediatr Orthop 2002; 22(5): 677-82. Benedetti M G, Catani F, Leardini A, Pignotti E, Giannini S. Data management in gait analysis for clinical applications. Clinical Biomechanics 1998; 13(3): 204-15. Butler E E, Steele K M, Torburn L, Gamble J G, Rose J. Clinical motion analyses over eight consecutive years in a child with crouch gait: a case report. J Med Case Rep 2016; 10: 157. Castelli A, Paolini G, Cereatti A, Della Croce U. A 2D Markerless gait analysis methodology: validation on healthy subjects. Comput Math Methods Med 2015; 2015: 186780. CPUP. Uppföljningsprogram för Cerebral pares: Årsrapport. 2013. Available from: http://cpup.se/wp-content/uploads/2014/03/Arsrapport2013_1.pdf. Del Pilar Duque Orozco M, Abousamra O, Church C, Lennon N, Henley J, Rogers K J, Sees J P, Connor J, Miller F. Reliability and validity of Edinburgh visual gait score as an evaluation tool for children with cerebral palsy. Gait Posture 2016; 49: 14-18. DeLuca P A. Gait analysis in the treatment of the ambulatory child with cerebral palsy. Clin Orthop Relat Res 1991; (264): 65-75. Eastlack M E, Arvidson J, Snyder-Mackler L, Danoff J V, McGarvey C L. Interrater reliability of videotaped observational gait-analysis assessments. Phys Ther 1991; 71(6): 465-72. Gage J R, DeLuca P A, Renshaw T S. Gait analysis: principle and applications with emphasis on its use in cerebral palsy. Instr Course Lect 1996; 45: 491-507. Gupta S, Raja K. Responsiveness of Edinburgh Visual Gait Score to orthopedic surgical intervention of the lower limbs in children with cerebral palsy. Am J Phys Med Rehabil 2012; 91(9): 761-7. Hägglund G, Andersson S, Duppe H, Lauge-Pedersen H, Nordmark E, Westbom L. Prevention of dislocation of the hip in children with cerebral palsy: the first ten years of a population-based prevention programme. J Bone Joint Surg Br 2005; 87(1): 95-101. Kawamura C M, de Morais Filho M C, Barreto M M, de Paula Asa S K, Juliano Y, Novo N F. Comparison between visual and three-dimensional gait analysis in patients with spastic diplegic cerebral palsy. Gait Posture 2007; 25(1): 18-24. Koo T K, Li M Y. A guideline of selecting and reporting intraclass correlation coefficients for reliability research. J Chiropr Med 2016; 15(2): 155-63.

661

Maathuis K G, van der Schans C P, van Iperen A, Rietman H S, Geertzen J H. Gait in children with cerebral palsy: observer reliability of Physician Rating Scale and Edinburgh Visual Gait Analysis Interval Testing scale. J Pediatr Orthop 2005; 25(3): 268-72. Mackey A H, Lobb G L, Walt S E, Stott N S. Reliability and validity of the Observational Gait Scale in children with spastic diplegia. Dev Med Child Neurol 2003; 45(1): 4-11. Miller F. Cerebral palsy [Elektronisk resurs]. New York: Springer Science+Business Media; 2005. Narayanan U G. The role of gait analysis in the orthopaedic management of ambulatory cerebral palsy. Curr Opin Pediatr 2007; 19(1): 38-43. Ong A M, Hillman S J, Robb J E. Reliability and validity of the Edinburgh Visual Gait Score for cerebral palsy when used by inexperienced observers. Gait Posture 2008; 28(2): 323-6. Opheim A, McGinley J L, Olsson E, Stanghelle J K, Jahnsen R. Walking deterioration and gait analysis in adults with spastic bilateral cerebral palsy. Gait Posture 2013; 37(2): 165-71. Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi B. Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dev Med Child Neurol 1997; 39(4): 214-23. Perry J. Gait analysis, normal and pathological function. Thorofare, NJ: SLACK; 1992. Rathinam C, Bateman A, Peirson J, Skinner J. Observational gait assessment tools in paediatrics: a systematic review. Gait Posture 2014; 40(2): 279-85. Read H S, Hazlewood M E, Hillman S J, Prescott R J, Robb J E. Edinburgh visual gait score for use in cerebral palsy. J Pediatr Orthop 2003; 23(3): 296-301. Rodda J, Graham H K. Classification of gait patterns in spastic hemiplegia and spastic diplegia: a basis for a management algorithm. Eur J Neurol 2001; 8 (Suppl 5): 98-108. Rodda J M, Graham H K, Carson L, Galea M P, Wolfe R. Sagittal gait patterns in spastic diplegia. J Bone Joint Surg Br 2004; 86(2): 251-8. Sandau M, Koblauch H, Moeslund T B, Aanaes H, Alkjaer T, Simonsen E B. Markerless motion capture can provide reliable 3D gait kinematics in the sagittal and frontal plane. Med Eng Phys 2014; 36(9): 1168-75. Saner R J, Washabaugh E P, Krishnan C. Reliable sagittal plane kinematic gait assessments are feasible using low-cost webcam technology. Gait Posture 2017; 56: 19-23. Sangeux M, Armand S. Kinematic deviations in children with cerebral palsy. In: Orthopedic management of children with cerebral palsy: a comprehensive approach. Hauppauge, NY: Nova Science Publishers; 2015. Surer E, Cereatti A, Grosso E, Della Croce U. A markerless estimation of the ankle-foot complex 2D kinematics during stance. Gait Posture 2011; 33(4): 532-7. Ugbolue U C, Papi E, Kaliarntas K T, Kerr A, Earl L, Pomeroy V M, Rowe P J. The evaluation of an inexpensive, 2D, video based gait assessment system for clinical use. Gait Posture 2013; 38(3): 483-9.

662

Acta Orthopaedica 2018; 89 (6): 662–667

Timing for Ponseti clubfoot management: does the age matter? 90 children (131 feet) with a mean follow-up of 5 years Yu-Bin LIU 1,2,3, Song-Jian LI 1, Li ZHAO 2,3, Bo YU 1, and Da-Hang ZHAO 3 1 Department

of Orthopaedics, Zhujiang Hospital of Southern Medical University, Guangzhou; 2 Ying-Hua Medical Group of Bone and Joint Healthcare in Children, Shanghai; 3 Department of Pediatric Orthopaedics, Xin-Hua Hospital affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai, China Correspondence: (Li Zhao) orthzl@126.com Submitted 2018-01-29. Accepted 2018-08-22.

Background and purpose — There are still controversies as to the age for beginning treatment with the Ponseti method. We evaluated the clinical outcome with different age at onset of Ponseti management for clubfoot. Patients and methods — 90 included children were divided into 3 groups in terms of age at start of treatment. The difference in treatment-related and prognosis-related variables including presentation age, initial Pirani and Dimeglio score, casts required, relapse rates, final Dimeglio score, and international clubfoot study group score (ICFSG) was analyzed. Results — Age between 28 days and 3 months at start of treatment method was associated with fewer casts required, lower relapse rate, and lower final ICFSG score (p < 0.05). Early treatment before 28 days of age required more casts and had a higher relapse rate (p < 0.05). The highest ICFSG scores were found in the ages between 3 and 6 months (p < 0.05). After propensity score matching, age between 28 days and 3 months was demonstrated to have a lower finial ICFSG score. Linear regression models showed that presentation age was positively correlated with final ICFSG score, and was identified as the only independent prognostic risk factor. Interpretation — There was lower rate of relapse and better clinical outcome when treatment was initiated at age between 28 days and 3 months. With the Ponseti method, clubfeet may not need urgent treatment.

Idiopathic clubfoot appears with an incidence of 5/104 in China (Yi et al. 2013). Nowadays, the Ponseti method has been widely accepted as the treatment of choice, and its safety and efficacy has been extensively demonstrated around the world (Zhao et al. 2014a, Liu et al. 2016). Generally, treatment with the Ponseti method is started within the first few weeks of life (Ponseti 1996, Dobbs and Gurnett 2009, Zhao et al. 2014b, Liu et al. 2016). In the study by Alves et al. (2009), the patients were divided into 2 groups (younger or older than 6 months) and no difference was found in the number of casts, tenotomies, success in terms of rate of initial correction, rate of recurrence, or rate of tibialis anterior transference. Iltar et al. (2010) reported that casting treatment beginning later than age 1 month or with an affected foot ≥ 8 cm in length had better treatment outcome. Zionts et al. (2016) reported that the age at the onset of treatment did not appreciably influence the cast phase of treatment or the initiation of post-corrective bracing, with the exception of cast slippage. Awang et al. (2014) have demonstrated that the total number of castings required to treat clubfoot was determined by the severity of clubfoot but not by the weight and age of patients. Other authors have reported that children presenting late with clubfoot can also be successfully treated by the Ponseti method (Bor et al. 2006, Lourenco and Morcuende 2007, Haj Zargar Bashi et al. 2016). It is still unclear if the treatment outcome is related to the age when the treatment was initiated. We assessed whether age at start of treatment influences the number of casts, tenotomies, the correction rates, recurrence rates, ankle dorsiflexion after treatment, final Demeglio and international clubfoot study group score (ICFSG) with a mean follow-up of 5 years.

Acta Orthopaedica 2018; 89 (6): 662–667

663

Idiopathic clubfoot treated in our hospital using the Ponseti method October 2007 to December 2012 n = 188 cases (268 feet) Excluded Lost to follow-up or still under treatment n = 77 cases (107 feet) Finished the whole protocol of the Ponseti method n = 111 cases (161 feet) Excluded Presentation age > 6 months n = 21 cases (30 feet) Presentation age < 6 months n = 90 cases (131 feet) Group I: 30 cases (43 feet) Group II: 36 cases (56 feet) Group III: 24 cases (32 feet) Excluded Radiographs lost or non-eligible n = 31 cases (44 feet) Evaluated using the International clubfoot study group (ICFSG) score system n = 59 cases (87 feet)

Figure 1. Flowchart of patients with idiopathic clubfoot.

Patients and methods Children who had been diagnosed with idiopathic clubfeet, and who were treated with the Ponseti method, were retrospectively reviewed in the study. The inclusion criteria entailed the following: first presented in our hospital at age less than 6 months, finished the whole protocol of the Ponseti method, and at least 4 years of post-treatment follow-up. Children were excluded from this study for the following reasons: postural, syndromic, and neurological clubfeet; previous treatment before referral; lost to follow-up at our institution before reaching 4 years of age; data lost or incomplete; declined to participate. We reviewed the records of 188 patients (268 idiopathic clubfeet) treated consecutively between October 2007 and December 2012. 98 children were excluded from this study for the reasons outlined in Figure 1. 90 children (131 clubfeet) were divided into 3 groups according to the age at initiation of treatment, Group I (the initial treatment age younger than 28 days), Group II (age more than 28 days but less than 3 months) and Group III (age more than 3 months but less than 6 months). For each case included in the study, the gestational and presentation age, sex, side of clubfoot, Pirani and Dimeglio scores were recorded before initial treatment. We strictly followed the protocol outlined in the Ponseti method. All the clubfoot cases were treated by a single orthopedist (LZ). A long leg cast was applied after the first evaluation in the outpatient clinic. All children were treated without anesthesia or sedative. The special technician in our clinics was responsible for the removal of the cast, and parents were

Figure 2. Foot-abduction brace was applied in our clinics. The brace consists of a bar with the shoes attached at 60–70° of abduction on the affected side and 30–40° on the normal side.

not recommended to remove the cast themselves. Evaluation using Pirani score and Demeglio score was performed each time after removal of the cast. After the last cast, the procedure of percutaneous Achilles tenotomy (PAT) is indicated if ankle dorsiflexion is less than 15° and the foot abducted to 60–70° without pronation. Post PAT long-leg cast was applied for 3 weeks. When the full correction had been achieved, a foot abduction orthosis (FAO, Figure 2) was prescribed to maintain the foot in the corrected position. The brace protocol was in full-time use for the first 3 months, then 16 to 18 hours until the children were 2 years old, then 14 to 16 hours until 4 years old. Relapses were treated with repeated manipulation and casting with or without PAT, followed by the use of the foot-abduction brace. Tibialis anterior tendon transfer (TATT) was performed in children with dynamic supination of the forefoot during the swing phase of gait. Wearing of the brace was required for an additional year after the removal of the last cast in relapsed cases identified at older than 4 years of age. We defined relapse as recurrence of any component of the deformities including adductus, varus, cavus, and equinus requiring further intervention, including either repeated manipulation and casting or surgical intervention. Noncompliance was defined as not wearing the brace for at least 75% of the number of hours prescribed (Liu et al. 2016), as reported by the parents. The mean follow-up was 5 years (4–8). Statistics Comparisons of 3 groups in terms of Pirani and Dimeglio scores, number of casts before PAT, ankle dorsiflexion degrees, mean follow-up, and ICFSG scores were performed using ANOVA. The chi-square test for comparison of multiple rate among 3 groups was applied for comparing variables such as the rate of PAT, initial correction rate, compliance rate, and relapse rate. Univariable and stepwise multivariable linear regressions were employed to estimate the correlations of each clinical characteristics with final ICFSG scores. The propensity scores were estimated using a logistic regression model that included the following eight covariates: gender, bilateral or unilateral, initial Pirani score, initial Dimeglio score, number of casts before PAT, PAT or not, brace

664

Acta Orthopaedica 2018; 89 (6): 662–667

Figure 3. Relapse was identified at 4 years of age for poor brace compliance. Right clubfoot: Dimeglio score = 8. (A) Forefoot adductus, (B) No obvious varus deformity, (C) Lateral edge curve, (D) Equinus deformity, (E) Standing anteroposterior views: reduced talocalcaneal angle, (F) Standing lateral views: reduced talocalcaneal angle and flat-top talus.

Figure 4. The same patient after repeated Ponseti treatment: ICFSG = 4. (A–C) Morphology evaluation = 1, (D) Muscle function = 0, (E) Radiologic evaluation = 3, (F) Gait analysis: dynamic function = 0.

compliance, and relapse or not. The matching approach was 1:1 nearest neighbor. Absolute standardized differences for all baseline covariates of < 10% were accepted as adequate balance. All statistical analyses were performed with SPSS® (IBM SPSS Statistics, version 24; IBM, Armonk, NY, USA) and twang R library (available at: cran.r-project.org/web/ packages/twang/index.html). A 2-tailed test with p < 0.05 was considered significantly different.

Ethics, funding, and potential conflicts of interest Informed consent was obtained from all the parents. This study was approved by the institutional ethics committee (approval number XHEC-D-2017-061) and was performed in accordance with the Declaration of Helsinki. This work was supported by National Natural Science Foundation of China (No. 81802215). No competing interests were declared.

Acta Orthopaedica 2018; 89 (6): 662–667

665

ment due to relapse. Noncompliance was associated with 3.3 times greater odds of relapse (16 of 33) in comparison with compliance (14 of 94). The included 90 cases (131 feet) that Group I Group II Group III had finished the whole recommended protocol of the Ponseti method and were followed up until 4 years of age with comNo. of cases (feet) 30 (43) 36 (56) 24 (32) plete information collected in our database. There were no Age at presentation 13 days 47 days 4.5 months range (2–27) (29–90) (3–6) preterm infants, and the mean gestational age was 39.4 weeks Sex (range, 37–41 weeks) at birth. Male 25 29 18 Of the 90 cases (131 feet), the mean age at the onset of treatFemale 5 7 6 Side ment was 48 days (2 days to 6 months) with a mean Pirani Bilateral 13 20 8 score of 4.5 points (2.5–6) and a mean Dimeglio score of 13 Unilateral 17 16 16 c points (6–17). 78 children (115 feet) had the PAT procedure, Initial Pirani score 4.8 (0.9) 4.3 (1.1) 4.3 (0.7) Initial Dimeglio score c 14 (3.5) 13 (3.1) 13 (2.7) and the initial correction rate was 87%. Cast slippage was No. of casts before PAT c 4.5 (1.6) a 3.7 (1.2) 4.2 (1.4) a observed in 3 cases (4 feet) for Group I with severe equinus d PAT rate deformity. No skin necrosis or ulcer, rocker bottom deformity, Yes 28 (40) 29 (46) 22 (28) No 2 (3) 7 (10) 2 (4) or neurovascular compromise post-tenotomy were observed. Initial correction rate d 14 children reported the brace was used < 75% of the preYes 22 (33) 33 (51) 20 (26) scribed time. Relapses occurred in 13 children (21 feet) with No 8 (10) 3 (5) 4 (6) Brace compliance d a mean relapse age of 34 months (11 months to 5 years). The Yes 25 (35) 30 (46) 1 8 (26) most common relapse was presented with adductus and equiNo 5 (8) 6 (10) 6 (6) d Relapse rate nus (5 cases, 8 feet [Figure 3]), followed by adductus (4 cases, Yes 7 (14) 4 (6) b 2 (2) 8 feet). Relapses were treated with a second series of manipNo 23 (29) 32 (50) 22 (30) ulation and casting, followed by the use of a foot-abduction c Dorsal ﬂexion (°) 13.7 (9.2) 15.2 (8.3) 14.8 (7.9) brace. 3 children (4 feet) required a second tendo Achilles Mean follow-up (years) c 4.9 (1.2) 4.7 (0.7) 4.9 (0.7) Final Dimeglio score c 4.3 (1.3) 4.1 (0.8) 4.1 (0.6) tenotomy. 1 child (1 foot) required a TATT procedure. ICFSG score d 24 (34) 21 (35) 14 (18) 59 children (87 feet) were evaluated using the scoring system c a,b Total score 4.6 (1.4) 3.9 (2.2) 6.3 (1.1) of the International Clubfoot Study Group score (ICFSG). 31 Group I (≤ 28 days), Group II (> 28 days to ≤ 3 months), children (44 feet) were excluded from this evaluation because Group III (> 3 months to ≤ 6 months); of non-weight-bearing radiographs. The outcome of clinical a p < 0.05 compared with Group II. b p < 0.05 compared with Group I. evaluation in terms of foot morphology, function, and radic values are mean (SD) ology was evaluated as good or excellent for all 3 groups d values are cases (feet) according to the ICFSG rating system (Figure 4). The influence of age at the beginning of treatment for clubfoot with Ponseti management is presented in Table 1. There was no statistically significant difference in the severity of Results clubfoot deformity at initial presentation, as assessed by both Of 188 cases (268 feet), 61 cases (86 feet) were lost to fol- the Pirani and Dimeglio scoring systems. A statistically signiflow-up with mean brace-wearing time of 11 months (ranges, icant difference was found in the number of casts (Group I and 0–40 months), and 16 cases (21 feet) were still under treat- Group III) required before the PAT procedure in comparison with Group II. There was no statistically significant difference in the PAT rates, Table 2. Linear regression of ICFSG score with clinical characteristics (n = 59) initial correction rates, compliance rates, ankle dorsiflexion degrees after treat Univariable linear regression Multivariable linear regression ment, mean follow-up, and finial DimegVariables β (95% CI) p-value β (95% CI) p-value lio score among the three groups. The Presentation age (days) 0.017 (0.007 to 0.027) 0.002 0.019 (0.009–0.029) < 0.001 relapse rate was lower in Group II in comSex 0.18 (–1.3 to 1.7) 0.8 – – parison with Group I, and a statistically Side –0.14 (–1.2 to 0.9) 0.8 – – significant difference was found between Initial Pirani score –0.51 (–1.0 to –0.02) 0.04 –0.11 0.4 Initial Dimeglio score –0.16 (–0.3 to –0.01) 0.04 –0.22 0.08 them (p < 0.05). After completion of the No. of casts 0.18 (–0.2 to 0.6) 0.3 – – whole protocol of the Ponseti method, the PAT –1.23 (–2.8 to 0.3) 0.1 – – mean scores of Group III were the highest Compliance 0.47 (–0.9 to 1.8) 0.5 – – Relapse –0.04 (–1.5 to 1.5) 1.0 – – in comparison with other groups (Group I and Group II) with a statistically signifiICFSG: International Clubfoot Study Group score. cant difference between them. Table 1. Demographic data with reference to age when clubfoot management was initiated

666

Acta Orthopaedica 2018; 89 (6): 662–667

Table 3. Demographic data after propensity score matching

Group I

Group II

Group III

No. of cases (feet) 15 (25) 15 (24) 10 (16) Age at presentation 13 days 45 days 4.6 months range (3–27) (31–87) (3–6) Sex Male 12 13 8 Female 3 2 2 Side Bilateral 10 9 6 Unilateral 5 6 4 Initial Pirani score c 4.6 (0.9) 4.5 (1.2) 4.1 (0.7) c Initial Dimeglio score 14 (3.9) 13 (3.9) 14 (1.2) No. of casts before PAT c 4.3 (1.8) 3.5 (0.9) 4.4 (1.0) PAT rate d Yes 13 (22) 13 (22) 10 (16) No 2 (3) 2 (2) 0 (0) Brace compliance d Yes 13 (21) 13 (21) 6 (12) No 2 (4) 2 (3) 4 (4) d Relapse rate Yes 5 (10) 2 (3) 2 (2) No 10 (15) 13 (21) 8 (14) Propensity score c 0.5 (0.2) 0.5 (0.2) 0.6 (0.2) c ICFSG score 4.9 (1.6) 3.9 (2.2) 6.8 (1.1) a,b For footnotes see Table 1.

The correlations of clinical characteristics with final ICFSG score were analyzed using univariable and stepwise multivariable linear regression models (Table 2). In the cases with bilateral clubfeet who had different initial Pirani score, Dimegliso score, or required number of casts for correction, the greater score or bigger number was used accordingly. As for univariable linear regression analysis, the results revealed that final ICFSG score was positively correlated with presentation age (p = 0.002), but negatively associated with initial Pirani score and initial Dimeglio score (p = 0.04, p = 0.04, respectively). After the stepwise multivariable linear regression analysis with adjustment for covariates, presentation age was found to be positively associated with final ICFSG score with statistical significance (p < 0.001). For further analysis of the cause–effect relations between presentation age and final ICFSG score, the propensity score matching method was applied to adjust for potential confounding factors. After propensity score matching, the covariates such as sex, bilateral or unilateral, initial Pirani score, initial Dimeglio score, number of casts before PAT, PAT or not, brace compliance, and relapse or not, were well balanced among groups. The results showed that the mean ICFSG scores of Group III were the highest in comparison with the other groups (Group I and Group II) with a statistically significant difference, and Group II was found to have the lowest final ICFSG score after completing Ponseti method treatment (Table 3).

Discussion The proposal that congenital clubfoot should be treated soon after birth has been widely accepted (Dobbs et al. 2004, Morcuende et al. 2004, Dobbs and Gurnett 2009, Zhao et al. 2014a, Liu et al. 2016). Ponseti (1996) also suggested that initial treatment should begin in the first few weeks of life to take advantage of the more favorable viscoelastic properties of the connective tissues in the newborn. The upper limit age for Ponseti method management is unclear. Several authors have reported that neglected clubfoot cases or patients presenting at an older age could also be successfully managed by the Ponseti method (Lourenco and Morcuende 2007, Spiegel et al. 2009, Khan and Kumar 2010, Ayana et al. 2014). In a recently published study in which 11 neglected patients with a mean age of 11 years (6–19) were treated with a modified Ponseti method, 17 out of 18 feet achieved a good result with no need for further surgery (Haj Zargar Bashi et al. 2016). It is still unknown whether the age at initiation of treatment influences the clinical outcome and the rate of relapse. We found that start of treatment age between 28 days and 3 months was identified with fewer casts required before PAT procedure, lower relapse rate, and better final ICFSG score. Stepwise multivariable linear regression models identified presentation age as the only independent prognostic risk factor for final clinical outcome. Earlier treatment with presentation age less than 28 days required more casts before the PAT procedure and had a higher relapse rate during the process of Ponseti method management. This finding is in contrast to previous recommendations that treatment should be started soon (7 to 10 days) after birth (Ponseti 1996, Dobbs et al. 2004, Liu et al. 2016). Our finding supports the notion that patients with presentation age after the first month achieved a better clinical outcome than those with earlier presentation (Iltar et al. 2010). The variance was attributed to the long age span in the older infant group (> 30 days, but < 1 year of age). The less satisfactory results in newborns may be attributed to small feet and partial Achilles tenotomy. Karami et al. (2015) demonstrated that complete Achilles tenotomy for newborns was identified in only 7/16 feet with a mean percentage of cut area of 77% measured by ultrasound. Partial Achilles tenotomy may be high risk for the late relapse of deformity in Group I. The highest ICFSG score was obtained in Group III (age more than 3 months), and the score was mainly dependent on the results of radiographic evaluation when weight-bearing. We presume that age at initial treatment of more than 3 months may influence alignment of the tarsal bones after Ponseti method treatment. This may be attributed to reduced response of soft tissue to the Ponseti maneuver due to the decreased viscoelastic properties of the connective tissues along with the infant’s age. In some cases treated using the Ponseti method, we found the mild deformity of residual varus, which was characterized radiographically by a decreased talocalcaneal

Acta Orthopaedica 2018; 89 (6): 662–667

angle including views in both standing anteroposterior and lateral positions. This is consistent with the results reported by Cooper and Dietz (1995). However, the final Dimeglio score was similar among these 3 groups. We presume that the foot morphological evaluation was not completely consistent with its radiographic assessment when weight-bearing. Many studies suggested that non-compliance with the FAO brace was closely associated with relapse of the deformity (Ponseti 1996, Dobbs et al. 2004, Morcuende et al. 2004, Dobbs and Gurnett 2009, Zhao et al. 2014b, Liu et al. 2016). Our findings indicated that there is no influence of age at initiation of treatment on brace adherence among 3 groups. Ramirez et al. (2011) reported that there was no significant difference in brace adherence or rate of relapse in patients who began treatment before or after 1 month of age. Zionts et al. (2012) also found no difference in brace adherence between infants who began treatment before and after 3 months of age. In contrast, Bor et al. (2009) found that children who began treatment before 28 days of age had poorer brace compliance than those who began treatment later. In our cases with relapse, the most common type of relapse was identified with adductus and equinus in the present study (Figure 4). This is in contrast to Ponseti (1996), who noted slight equinus and varus deformity as the first appearance in the relapse cases, often without increased adductus and cavus of the forefoot. As Ponseti reported, the relapse was caused by the same pathology that initiated the deformity, and the soft tissue abnormality may be an important cause of the deformity with an excess of collagen synthesis in the Achilles tendon, the posterior tibial tendon, and medial and posterior tarsal ligaments (Ponseti 1996). The retraction of connective tissue of the foot and ankle first induced adductus of the forefoot, followed by equinus deformity due to the kinematic coupling relationship between the tarsal bones. Supposing that the relapse was identified at a very early stage, the recurrence of varus deformity was not often observed in our children. We suppose the retraction of soft issue in the relapse cases to be the initial factor for abnormal alignment of the tarsal bones. We acknowledge some limitations of our study. The study was retrospective and a relatively small number of cases were included. More randomized controlled trials or large-scale case-control studies are required for further validation. In summary, treatment initiated between 28 days and 3 months of age produced lower rate of relapse and better clinical outcomes. YBL and SJL contributed equally to this article. YBL contributed substantially to the acquisition of data and drafted the article. SJL and BY performed the statistical analysis. LZ contributed to the conception and design. DHZ contributed to the case collection in clinical practice. The authors are grateful to Prof. Weiyu Han from the Clinical Medical Research Center and department of Orthopaedics, Zhujiang Hospital of Southern Medical University, for help in statistical analysis. Acta thanks Ole Rahbek for help with peer review of this study.

667

Alves C, Escalda C, Fernandes P, Tavares D, Neves M C. Ponseti method: does age at the beginning of treatment make a difference? Clin Orthop Relat Res 2009; 467(5): 1271–7. Awang M, Sulaiman A R, 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. Ayana B, Klungsoyr P J. Good results after Ponseti treatment for neglected congenital clubfoot in Ethiopia: a prospective study of 22 children (32 feet) from 2 to 10 years of age. Acta Orthop 2014; 85(6): 641–5. Bor N, Herzenberg J E, Frick S L. Ponseti management of clubfoot in older infants. Clin Orthop Relat Res 2006; (444): 224–8. Bor N, Coplan JA, Herzenberg J E. Ponseti treatment for idiopathic clubfoot: minimum 5-year followup. Clin Orthop Relat Res 2009; 467(5): 1263–70. Cooper D M, Dietz F R. Treatment of idiopathic clubfoot: a thirty-year follow-up note. J Bone Joint Surg Am 1995; 77(10): 1477–89. Dobbs M B, Gurnett C A. Update on clubfoot: etiology and treatment. Clin Orthop Relat Res 2009; 467(5): 1146–53. Dobbs M B, Rudzki J R, 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-A(1): 22–7. Haj Zargar Bashi R, Baghdadi T, Ramezan Shirazi M, Abdi R, Aslani H. Modified Ponseti method of treatment for correction of neglected clubfoot in older children and adolescents: a preliminary report. J Pediatr Orthop B 2016; 25(2): 99-103. Iltar S, Uysal M, Alemdaroglu K B, Aydogan N H, Kara T, Atlihan 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. Karami M, Dehghan P, Moshiri F, Shamami M S. Effect of unintentional partial Achilles tenotomy on Ponseti clubfoot management outcomes. J Pediatr Orthop B 2015; 24(1): 1–5. Khan S A, Kumar A. Ponseti’s manipulation in neglected clubfoot in children more than 7 years of age: a prospective evaluation of 25 feet with long-term follow-up. J Pediatr Orthop B 2010; 19(5): 385–9. Liu Y, Zhao D, Zhao L, Li H, Yang X. Congenital clubfoot: early recognition and conservative management for preventing late disabilities. Indian J Pediatr 2016; 83(11): 1266–74. Lourenco A F, Morcuende J A. Correction of neglected idiopathic club foot by the Ponseti method. J Bone Joint Surg Br 2007; 89-B(3): 378–81. Morcuende J A, Dolan L A, Dietz F R, Ponseti I V. Radical reduction in the rate of extensive corrective surgery for clubfoot using the Ponseti method. Pediatrics 2004; 113(2): 376–80. Ponseti I V. Congenital clubfoot: fundamentals of treatment. Oxford, New York: Oxford University Press 1996. Ramirez N, Flynn J M, Fernandez S, Seda W, Macchiavelli R E. Orthosis noncompliance after the Ponseti method for the treatment of idiopathic clubfeet: a relevant problem that needs reevaluation. J Pediatr Orthop 2011; 31(6): 710–15. Spiegel D A, Shrestha O P, Sitoula P, Rajbhandary T, Bijukachhe B, Banskota A K. Ponseti method for untreated idiopathic clubfeet in Nepalese patients from 1 to 6 years of age. Clin Orthop Relat Res 2009; 467(5): 1164–70. Yi L, Zhou G X, Dai L, Li K S, Zhu J, Wang Y. [A descriptive epidemiological study on congenital clubfoot in China during 2001 to 2010]. Sichuan Da Xue Xue Bao Yi Xue Ban 2013; 44(4): 606–9. Zhao D, Li H, Zhao L, Liu J, Wu Z, Jin F. Results of clubfoot management using the Ponseti method: do the details matter? A systematic review. Clin Orthop Relat Res 2014a; 472(4): 1329–36. Zhao D, Liu J, Zhao L, Wu Z. Relapse of clubfoot after treatment with the Ponseti method and the function of the foot abduction orthosis. Clin Orthop Surg 2014b; 6(3): 245–52. Zionts L E, Frost N, Kim R, Ebramzadeh E, Sangiorgio S N. Treatment of idiopathic clubfoot: experience with the Mitchell–Ponseti brace. J Pediatr Orthop 2012; 32(7): 706–13. Zionts L E, Sangiorgio S N, Cooper S D, Ebramzadeh E. Does clubfoot treatment need to begin as soon as possible? J Pediatr Orthop 2016; 36(6): 558–64.

668

Acta Orthopaedica 2018; 89 (6): 668–673

Innovative treatment for pes cavovarus: a pilot study of 13 children Ignacio SANPERA JR 1, Guillem FRONTERA-JUAN 2, Julia SANPERA-IGLESIAS 3, and Laura COROMINAS-FRANCES 1 1 Pediatric Orthopedic Department, Hospital Universitari Spain; 3 Croydon University Hospital, Croydon, UK

Son Espases, Palma de Mallorca, Spain; 2 Research Unit, Hospital Universitari Son Espases,

Correspondence: isanpera@gmail.com Submitted 2017-10-04. Accepted 2018-04-23.

Background and purpose — Pes cavovarus (PCV) is a complex deformity, frequently related to neurological conditions and associated with foot pain, callosities, and walking instability. The deformity has the tendency to increase during growth. Orthotic treatment is ineffective and surgery may be troublesome. We present the preliminary results of a new mini-invasive surgical technique for correction of this foot deformity. Patients and methods — We operated on 13 children (24 feet), age 7–13 years. In 7 children the deformity was neurological in origin. The surgical technique included a dorsal hemiepiphysiodesis of the 1st metatarsal, and a plantar fascia release. The clinical deformity, hindfoot flexibility, and foot callosities, together with a radiological assessment (Meary angle, calcaneal pitch, and talo-calcaneal angle), was done pre- and postoperatively. At final check-up, after a median of 28 months (12–40), the Oxford Ankle Foot Questionnaire for children (OXAFQ-C) was used to assess patient satisfaction. The primary outcome was the hindfoot varus correction. Results — All the operated feet improved clinically and radiologically. Heel varus improved from a mean 6º preoperatively to 5º valgus postoperatively. In those children where treatment was initiated at a younger age, full correction was achieved. Footwear always improved. Interpretation — This treatment may offer a less aggressive alternative in the treatment of PCV in young children and may eventually reduce the amount of surgery needed in the future.

Pes cavovarus (PCV) is a complex deformity of the foot produced by the combination of forefoot pronation and inversion of the hindfoot. The etiology of the deformity is reported to be neurological in a majority of cases; Charcot–Marie–Tooth (CMT), a hereditary motor and sensory neuropathy, being the most common cause. This condition is characterized by foot weakness and deformity (Pareyson and Marchesi 2009). The foot deformity is the result of unbalanced forces applied on a growing foot, leading to a progressive deformity. Although unanimous agreement does not exist, the initial event, at least in Charcot–Marie–Tooth, appears to be denervation of the intrinsic muscles of the foot, resulting in clawing toes and an increase in the height of the foot’s arch, leading to an equinus deformity of the forefoot over the hindfoot. Once the deformity is established it tends to progress. The rate of progression and severity will ultimately depend on the form of Charcot–Marie–Tooth, its causative gene and the type of mutation (Pareyson and Marchesi 2009, Burns et al. 2012). The deformity tends to appear at the beginning of the second decade of life; the heel varus is initially flexible, but in most patients will progress and becomes stiff by the end of the same decade (Aktas and Sussman 2000). The deformity results in foot pain, callosities, ankle instability, and difficulties with footwear (Burns et al. 2009, Wicart 2012, Dreher et al. 2014). Common treatment for pes cavovarus includes tendon transfers, bone osteotomies located at either the fore- or the hindfoot, and combinations of all of them, determined by the treating surgeon and the severity of the condition. Hemiepiphysiodesis produces bone reshaping by asymmetrically interfering with bone growth, allowing in this way deformity correction and sparing bone osteotomies. Although the mechanism of application is simple, our understanding of the physiology of this mechanism is limited (Corominas-Frances et al. 2015, Gotliebsen et al. 2016).

Acta Orthopaedica 2018; 89 (6): 668–673

669

Eligible pes cavovarus patients during study period November 2012 to Novemeber 2015 with inclusion criteria: – uni- or bilateral pes cavovarus – positive Coleman test – skeletal immature – evidence of progression Patients, n = 19 Feet n, = 33 Refused treatment: Patients, n = 4 Feet, n = 7 Accepted treatment: Patients, n = 15 Feet, n = 26

Figure 2. (Left) Preoperative picture of a 9-year-old child standing, seen from the back. The lines drawn show the presence of a heel varus deformity of the hindfoot. (Right) Same child in a similar position, 18 months after surgery. Note the changes in the heel position, now in a valgus position.

Lost to follow-up: Patients, n = 1 Feet, n = 1 Failure of treatment: Patients, n = 1 Feet, n = 1 Available for final evaluation: Patients, n = 13 Feet, n = 24

Figure 1. Flowchart of data collection. Inclusion criteria are explained, and number of patients recruited and excluded are detailed.

The aim of this study was to assess whether the use of dorsal hemiepiphyseal arrest of the 1st metatarsal (1MMT), when combined with a percutaneous fascia plantar release, would stop deterioration or even achieve correction of the hindfoot deformity in the cavovarus foot.

Patients and methods During November 2012 to November 2015, 15 children (10 girls) were recruited for a pilot study. This was a proof of concept study to ascertain the effect, safety, and variations that could be expected from the treatment before starting a major trial. In 9 cases the cause of the pes cavovarus was neurological whilst in the remaining 6 children no etiological cause was found. 11 children had bilateral surgery and 4 children were only operated on 1 foot. The reasons behind single foot surgery were: first, asymmetrical involvement (2 feet), and second, previous foot surgery on the contralateral side (2 patients). The average age for girls at surgery was 10 years (7–13) and for boys 11 years (9–12) for boys. For the study, we used a number of selection criteria. Patients who fulfilled the criteria were offered the possibility

of attending the study and parents signed an informed consent. 4 suitable candidates declined to participate (Figure 1). All patients were assessed both clinically and radiologically. Clinically, the foot deformity was monitored dynamically and statically. Pictures from the back in standing position and during the Coleman block test were taken. The heel angle was measured on the photographs, drawing a line bisecting the heel and a second one bisecting the calf and measuring the angle of confluence (Figure 2). Foot callosities were recorded. Patients and caregivers were questioned about the presence of pain and disabilities in daily life. Preoperative front and lateral standing foot radiographs were obtained in all patients. The following parameters were analyzed: Meary’s angle and calcaneal pitch on the lateral film and 1st metatarsal-talus angle on standing AP view (Figure 3). The primary outcome was the correction of the heel varus, and the secondary outcome the change in the measured angles and callosities. At final examination, at a mean of 28 months (12–40) after surgery new clinical photographs were obtained and compared with preoperative images (Figure 2). Radiological values at pre-surgery and at final check-up were also compared (see Figure 3). A foot and ankle questionnaire was handed out to evaluate the extent of disability produced by their foot condition. We used the OXAFQ-C (Morris et al. 2008, 2009), which assesses 3 fields: physical (maximum score 24), school and play (score 16), and emotional (score 16). It also evaluates the limitation that children experience with the use of footwear (maximal score 4). Only 7 patients fulfilled the OXAFQ-C preoperatively, and their responses were compared with the scores achieved at final follow-up. Finally, patients were specifically questioned about how their footwear usability had changed with surgery, and those who did not complete the OXAFQ-C preoperatively were asked to score, numerically and retrospectively, their footwear difficulty. Of the 15 patients (26 feet) operated on, 13 patients (24 feet) were available for follow-up (13 of neurological origin and 11

670

Acta Orthopaedica 2018; 89 (6): 668â&#x20AC;&#x201C;673

of unknown etiology). 2 patients (2 feet) were lost to follow-up: 1 of them moved away and the second had an unnoticed misplacement of the hemiepiphysiodesis device, resulting in failure. Additional surgeries included: 2 feet with Jones tendon transfer for symptomatic first toe claw deformity, and 1 foot with a simultaneous tendo Achilles lengthening for a fixed equinus. As some children did not have same-day surgery on both feet, the follow-up time was referred to feet rather than to patients (in 3 bilateral cases each foot was operated at different stages with an interval range from 3 to 12 months). At the moment of the final check-up 5 patients had already reached skeletal maturity. Statistics Descriptive analysis was performed calculating mean and standard deviation (SD) for continuous variables, with their 95% confidence interval (CI). To measure the effect of hemiepiphysiodesis we calculated the difference between the basal angle and the angle at last follow-up, with their 95% confidence intervals (CI). Figure 3. Radiographs of both feet taken preoperatively at 7 years 8 months (top), and postoperatively (bottom, same patient 18 months after surgery), to evaluate correction. On each As for the foot questionnaire, the issue radiograph the following angles were measured: on the AP standing view, the talusâ&#x20AC;&#x201C;1st metaof unilateral/bilateral involvement was not tarsal angle, and on the lateral standing view Mearyâ&#x20AC;&#x2122;s angle and the calcaneal pitch. Note that considered as patients were evaluated as a the initial involvement was asymmetrical, with the right foot more involved than the left. By 18 months, both feet show signs of hypercorrection. whole, independent of whether they had single or bilateral foot surgery. All analyses were performed with IBM SPSS Statistics v.20 (IBM Corp, Armonk, NY, USA). Ethics, funding and conflict of interest Approval from the Local Institution Review Board was obtained, with number CI-185-17 issued on April 1, 2017. The parents of the children signed an informed consent to enter the study. No funding was obtained. There are no conflicts of interest to declare.

Figure 4. Operative pictures. A: Incision for the percutaneous plantar release. B: A dorsal incision for insertion of the plate is made; the physis is located with the help of a needle and confirmed under an image intensifier. C: detail of the inserted plate.

Surgical technique At the level of the internal tuberosity of the os calcis, on the medial foot border, a 2-centimeter incision was made (Figure 4) in an oblique fashion following the inferior border of the calcaneum. The fibers of the plantar fascia and the foot sole muscles were identified by a blunt instrument and percutaneously sectioned. A periosteal elevator was then introduced through the hole and used to disinsert any possible remaining fibers. Through a dorsal 1.5 cm incision centered over the proximal physis of the 1st metatarsal, the extensor hal-

Acta Orthopaedica 2018; 89 (6): 668–673

671

Table 1. Changes observed in the different radiologically measured angles preoperatively and at final follow-up

Heel angle (°) 15

Mean difference 95% CI

Right foot Change in Meary’s angle –6.8 –9.4 to –4.5 Change in talus–1st metatarsal angle –8.2 –12.5 to –4.0 Change in calcaneal pitch –3.6 –5.8 to –1.4 Left foot Change in Meary’s angle –7.1 –10.3 to –4.4 Change in talus–1st metatarsal angle –9.2 –16.3 to –3.7 Change in calcaneal pitch –2.2 –4.5 to 0.2

–5

–10 Pre

Post

Left side

Pre

Post

Results are presented according to the side of involvement.

Right side

Figure 5. Box plot showing the evolution of the clinical varus deformity after the surgery. The heel angle is reflected at 2 different periods in time, preoperatively (blue) and at final follow-up (green). The results are show by side of foot involvement (please note that positive values in the heel angle line represents varus deformity at the heel, while negative values represent valgus). All the operated feet decreased their varus deformity. The red lines reflect the median heel angle for each group. The bottom of the box represents the 25th percentile and the upper the 75 percentile. The T lines extending from the boxes are the innerfences or whiskers and extend 1.5 the heigth of the box or if no case/row has a value in this range to the minimum and maximum values. The point represents an outlier.

Table 2. Results of foot questionnaire: comparison between the different domains of the 7 patients who completed a questionnaire preoperatively and at final follow-up

lucis longus tendon was identified, and retracted medially. Using a 26-gauge needle the physis was located by punction and its position confirmed with fluoroscopy. Next, a 2-hole plate and screws were inserted bridging the physis, with the proximal screw placed first because is the most technically demanding (1.5 mm diameter) and once the position was ascertained the second screw was place. Finally, the position of the device was confirmed under fluoroscan in both anteroposterior and lateral positions. When in doubt an oblique view was used. A below-knee plaster was applied with the forefoot in maximal dorsiflexion over the hindfoot and was maintained for a period of 4 weeks. Full weight-bearing was allowed as soon as the patient felt comfortable.

All the radiological parameters improved postoperatively (Table 1). The OXAFQ-C at final follow-up showed an average score of 54 over a maximum of 60. Of the 7 children with available OXAFQ-C preoperatively, all improved (Table 2). The footwear usability changed from a preoperative average of 1.3 to a postoperative average of 3.5 out of 4 possible (no difficulties).

Results 3 patients had surgical complications. In 2 feet the proximal screw was wrongly placed and needed new surgery to reposition the screw. 1 foot had a rupture of the plate. Clinical assessment showed a preoperative mean varus deformity of 6º, which reversed to a mean 5º valgus of the hindfoot. Postoperatively, only one patient had a residual varus, shifting from an initial 12º varus to 3º varus. Altogether the mean correction achieved was 10º (CI 7.6–11) (Figure 5). At the final follow-up examination, all the callosities had disappeared, except in one patient where its size had considerably shrunk.

Factor

Mean score Mean score Mean preoperatively postoperatively difference 95 % CI

Physical domain School and play Emotional

15 9 12

21 13 14

–5.9 –3.4 –2.4

–9.2 to –2.5 –7.4 to –0.6 –5.2 to 0.4

Discussion The cavovarus foot is not strictly a diagnosis, but rather the description of a deformity secondary to different etiologies and with diverse pathomechanics, which makes its treatment complex (Paulos et al. 1980, Mosca 2001, Ward et al. 2008, Wicart 2012). Even in the same individuals, foot involvement can be highly variable and therefore requires an individualized approach for each foot (Burns et al. 2012). Few reports have dealt with the treatment of PCV at a young age. Due to its natural course, it is at this stage where intervention stands a better chance of being successful. Most authors agree that non-operative treatment has no role in this condition (Paulos et al. 1980, Burns et al. 2010), other than for symptomatic relief, and surgery is nearly always indicated. However, there are no clear indications for surgery, often leading to unpredictable results. Pes cavovarus should be considered, at least in the initial stages, as a forefoot problem with adaptive changes of the

672

Figure 6. Despite the plate being located at the base of the 1st metatarsal, part of the correction occurs on the surrounding bones. The arrow points at the changes seen at the navicular bone in a patient 2 years after surgery. Note the enlargement of the navicular on its plantar aspect.

midfoot and hindfoot (Paulos et al. 1980, Wicart and Seringe 2006, Mubarak and Van Valin 2009). Many authors have used an osteotomy at the base of the 1st metatarsal to achieve correction of the fall of the 1st metatarsal and consequently the hindfoot varus, with good results (Mubarak and Van Valin 2009). However, others, as the apex of the deformity is located at the medial cuneiform (Mosca 2001), prefer the use of an opening plantar wedge osteotomy at the first cuneiform, despite the fact that the effect over the hindfoot varus may be less obvious (Viehweger et al. 2005). 1st metatarsal osteotomies in children are controversial, as the osteotomy has to be located in the diaphysis to avoid physeal damage, and this may lead to a crooked correction far away from the apex of the deformity. Addressing the forefoot deformity at the physis of the 1st metatarsal allows for a progressive correction of the deformity (with the subsequent adaptive changes on the surrounding tissues) and, on the other hand, the correction point is located at close proximity to the apex of the deformity. Moreover, our radiological study suggests that part of the deformity correction happens in the adjacent tarsal bones, which is likely due to the foot pressure changes (Figure 6). The use of hemiepiphysiodesis of the proximal physis of the 1st metatarsal has previously been successfully used for treatment of hallux valgus. According to the literature, the proximal 1st metatarsal physis closes at around the age of 14.2 years in girls and 15.6 years in boys (Greene et al. 2017). There are 2 reasons for using a temporary rather than a defini-

Acta Orthopaedica 2018; 89 (6): 668â&#x20AC;&#x201C;673

tive hemiepiphysiodesis. First to be able to reverse the effect in case of overcorrection (Figure 3), and second to have an immediate effect, as opposed to the ablation hemiepiphysiodesis (Ross and Zionts 1997, Stevens 2007). Muscle balancing is certainly the most difficult part of the treatment of PCV. There is no consistent pattern of muscle imbalance (Mubarak and Van Valin 2009) and evaluation in a young child may be extremely difficult. There is general agreement that muscle imbalance, at least in CMT disease, starts with progressive weakness of the intrinsic foot muscles, resulting in hyperextension of the 1st metatarsophalangeal joint and, consequently, through a windlass effect in a height increase of the longitudinal arch of the foot. Percutaneous plantar release at this initial stage would not only lengthen the short intrinsic muscle, but would also facilitate the hemiepiphysiodesis effect by decreasing the plantar pressures. Another theoretical advantage of early PCV correction is to avoid the abnormal muscle growth associated with a deformed position. In order for the muscle to grow normally, it requires overstretching of its fibers (Gough and Shortland 2012). If the foot remains in a constant varus position, the tibial muscles will never reach full length, increasing the varus deforming forces at the hindfoot. The opposite would apply to the overlengthened foot evertors. This might explain why in our cases we have not required further tendon transfers. However, as the condition in most PCVs is progressive, it is likely that further muscle rebalancing will be needed. Whether the plantar fascial release is responsible on its own for the correction achieved may be discussed. However, multiple reports have shown that although radical plantar release may correct cavus, it is unable to reverse hindfoot varus (Paulos et al. 1980, Sherman and Westin 1981). Furthermore, the role of the hemiepiphysiodesis in valgus correction was clearly proved by our case with a foot deformity relapse after plate breakage. 2 patients underwent 3 additional surgeries. As none of these procedures has proven to correct the hindfoot varus deformity, we did not exclude these patients from the final analysis. The most frequent complication was proximal screw misplacement. Many reasons could explain this difficulty: the small size of the proximal epiphysis, the oblique position of the physis to the ground, and, last but not least, the superposition of multiple anatomical structures with the proximal physis of the 1st metatarsal on the lateral view radiographs. In 3 feet the screw was incorrectly positioned. In 2 instances the misplacement was detected immediately postoperatively and corrected, but in the remainder it went undetected and determined the failure of the procedure. 1 patient who was operated bilaterally had a relapse in 1 of the feet. Radiographs revealed breakage of the plate. Another possible adverse event is overcorrection of the deformity, which can be observed in younger patients and is a good reason to consider the use of a temporary arrest rather than a definitive one (Figure 3).

Acta Orthopaedica 2018; 89 (6): 668–673

The limitations of our study are the number of patients enrolled and the short follow-up. Nevertheless, it is long enough to demonstrate that, with minimal surgery, a normal or near-normal shaped foot can be reached at skeletal maturity. This procedure is not intended to be the definitive treatment for these children as the condition is progressive, and there is actually no effective treatment for most of the causes. Nevertheless, this procedure allows for control of the deformity, and makes the foot less symptomatic, as reflected by the improvement in footwear and the OXAFQ-C. The use of a material not specifically designed for this type of surgery is another limitation and may explain some of the complications encountered such as plate breakage (due to inadequate trimming of the plate). In summary, we believe that this technique may be useful for the treatment of pes cavovarus during childhood, although its use should be limited to cavovarus associated with weakness or the idiopathic form. The technique follows the principles laid down by Mosca (2001) for the treatment of pes cavus. First, the segmental deformities should be corrected while preserving joint motion (1st metatarsal hemiepiphysiodesis corrects the fall of the metatarsal). Next, the remaining muscle forces should be balanced out (achieved by plantar fascia release). Third, this leaves open the possibility of reasonable treatment options because the foot remains almost untouched.

Design of the study: IS, GFJ. Data collection: IS, LCF. Data analysis: IS, GFJ, JSI, LCF. Writing the paper: IS, GFJ, JSI Acta thanks Hanne Hedin and Line Kjeldgaard Pedersen for help with peer review of this study.

Aktas S, Sussman M D. The radiological analysis of pes cavus deformity in Charcot Marie Tooth disease. J Pediatr Orthop B 2000; 9(2): 137-40. Burns J, Ryan M M, Ouvrier R A. Evolution of foot and ankle manifestations in children with CMT1A. Muscle Nerve 2009; 39(2): 158-66. Burns J, Scheinberg A, Ryan M M, Rose K J, Ouvrier R A. Randomized trial of botulinum toxin to prevent pes cavus progression in pediatric Charcot– Marie–Tooth disease type 1A. Muscle Nerve 2010; 42(2): 262-7. Burns J, Ouvrier R, Estilow T, Shy R, Laurá M, Eichinger K, et al. Symmetry of foot alignment and ankle flexibility in paediatric Charcot–Marie–Tooth disease. Clin Biomech 2012; 27(7): 744-7.

673

Corominas-Frances L, Sanpera I, Saus-Sarrias C, Tejada-Gavela S, SanperaIglesias J, Frontera-Juan G. Rebound growth after hemiepiphysiodesis: an animal-based experimental study of incidence and chronology. Bone Joint J 2015; 97: 862-8. Dreher T, Wolf S I, Heitzmann D, Fremd C, Klotz M C, Wenz W. Tibialis posterior tendon transfer corrects the foot drop component of cavovarus foot deformity in Charcot–Marie–Tooth disease. J Bone Joint Surg Am 2014; 96(6): 456-62. Gough M, Shortland A P. Could muscle deformity in children with spastic cerebral palsy be related to an impairment of muscle growth and altered adaptation? Dev Med Child Neurol 2012; 54(6): 495-9. Gotliebsen M, Shiguetomi-Medina J M, Rahbek O, Moller-Madsen B. Guided growth: mechanism and reversibility of modulation. J Child Orthop 2016; 10: 471-7. Greene J D, Nicholson A D, Sanders J O, Cooperman D R, Liu R W. Analysis of serial radiographs of the foot to determine normative values for the growth of the 1st metatarsal to guide hemiepiphysiodesis for immature hallux valgus. J Pediatr Orthop 2017; 37(5): 338-43. Morris C, Doll HA, Wainwright A, Theologis T, Fitzpatrick R. The Oxford ankle foot questionnaire for children: Scaling, reliability and validity. J Bone Joint Surg Br. 2008; 90(11): 1451-6. Morris C, Doll H, Davies N, Wainwright A, Theologis T, Willett K, Fitzpatrick R. The Oxford ankle foot questionnaire for children: responsiveness and longitudinal validity. Qual Life Res 2009; 18(10): 1367-76. Mosca V S. The cavus foot. J Pediatr Orthop 2001; 21(4): 423-4. Mubarak S J, Van Valin S E. Osteotomies of the foot for cavus deformities in children. J Pediatr Orthop 2009; 29(3): 294-9. Pareyson D, Marchesi C. Diagnosis, natural history, and management of Charcot–Marie–Tooth disease. Lancet Neurol 2009; 8(7): 654-67. Paulos L, Coleman S S, Samuelson K M. Pes cavovarus. Review of a surgical approach using selective soft-tissue procedures. J Bone Joint Surg Am 1980; 62(6): 942-53. Ross T K, Zionts L E. Comparison of different methods used to inhibit physeal growth in a rabbit model. Clin Orthop Relat Res 1997; (340): 236-43. Sherman F C, Westin G W. Plantar release in the correction of deformities of the foot in childhood. J Bone Joint Surg Am 1981; 63(9): 1382-9. Stevens P M. Guided growth for angular correction: a preliminary series using a tension band plate. J Pediatr Orthop 2007; 27(3): 253-9. Viehweger E, Jacquemier M, Launay F, Giusiano B, Bollini G. First cuneiform osteotomy alters hindfoot architecture. Clin Orthop Relat Res 2005; 441: 356-65. Ward C M, Dolan L A, Bennett D L, Morcuende J A, Cooper R R. Long-term results of reconstruction for treatment of a flexible cavovarus foot in Charcot–Marie–Tooth disease. J Bone Joint Surg Am 2008; 90(12): 2631-42. Wicart P. Cavus foot, from neonates to adolescents. Orthop Traumatol Surg Res 2012; 98(7): 813-28. Wicart P, Seringe R. Plantar opening-wedge osteotomy of cuneiform bones combined with selective plantar release and Dwyer osteotomy for pes cavovarus in children. J Pediatr Orthop 2006; 26(1): 100-8.

674

Acta Orthopaedica 2018; 89 (6): 674–677

Abaloparatide versus teriparatide: a head to head comparison of effects on fracture healing in mouse models Magnus BERNHARDSSON and Per ASPENBERG

Orthopaedics, Department of Clinical and Experimental Medicine, Faculty of Health Sciences, Linköping University, Linköping, Sweden Correspondence: magnus.bernhardsson@liu.se Submitted 2018-04-09. Accepted 2018-08-09.

Background and purpose — Teriparatide accelerates fracture healing in animals and probably in man. Abaloparatide is a new drug with similar although not identical effects on the teriparatide receptor. Given at 4 times the teriparatide dose in a human osteoporosis trial, abaloparatide increased bone density more than teriparatide, and both reduced fracture risk. We investigated in mice whether abaloparatide stimulates fracture healing, and if it does so with the suggested dose effect relation (4:1). Patients and methods — In a validated mouse model for metaphyseal healing (burr hole with screw pull-out), 96 mice were randomly allocated to 11 groups: control (saline), teriparatide or abaloparatide, where teriparatide and abaloparatide were given at 5 different doses each. In a femoral shaft osteotomy model, 24 mice were randomly allocated to 3 groups: control (saline), teriparatide (15 µg/kg) or abaloparatide (60 µg/kg). Each treatment was given daily via subcutaneous injections. Results were evaluated by mechanical testing and microCT. Results — In the metaphyseal model, a dose-dependent increase in screw pull-out force could be seen. In a linear regression analysis (r = 0.78) each increase in ln(dose) by 1 (regardless of drug type) was associated with an increase in pull-out force by 1.50 N (SE 0.18) (p < 0.001). Changing drug from teriparatide to abaloparatide increased the force by 1.41 N (SE 0.60; p = 0.02). In the diaphyseal model, the callus density was 23% (SD 10), 38% (SD 10), and 47% (SD 2) for control, for teriparatide and abaloparatide respectively. Both drugs were significantly different from controls (p = 0.001 and p = 0.008), but not from each other. Interpretation — Both drugs improve fracture healing, but in these mouse models, the potency per µg of abaloparatide seems only 2.5 times that of teriparatide, rather than the 4:1 relation chosen in the clinical abaloparatide–teriparatide comparison trial.

Teriparatide is a truncated parathyroid hormone acting on the teriparatide receptor. It activates and stimulates osteoblasts and is an approved drug for osteoporosis. Teriparatide also stimulates fracture healing in a large number of animal models, and there are ample data to suggest it does also in humans, although a pivotal trial is missing. Teriparatide also activates osteoclasts, even though its anabolic effect on osteoblasts dominates its effects (Aspenberg and Johansson 2010, Aspenberg et al. 2016, Campbell et al. 2015). Abaloparatide is a new peptide that also binds the teriparatide receptor with effects similar to teriparatide. Due to differences in details of this binding, osteoclast activation is described as less pronounced (Boyce et al. 2018). In a large randomized trial comparing placebo, teriparatide, and abaloparatide in humans, the latter drug showed a stronger increase in bone density in the total hip region, the femoral neck, and the lumbar spine (Leder et al. 2015). Both drugs also reduced fracture risk. Abaloparatide was given daily at a dose of 80 µg and teriparatide at 20 µg. The choice of these doses was not motivated. To our knowledge, no studies on abaloparatide and fracture healing have been published. This animal study addresses 2 questions: Does abaloparatide stimulate fracture healing similar to teriparatide, and does the dose relation 80 versus 20 µg represent a similar relation in potency?

Methods Overview 96 mice received bilateral drill holes in their proximal tibiae, where a steel screw was inserted in one of them. The mice received daily injections of either saline, teriparatide, or abaloparatide of different dosage (n = 8 per group). The mice were killed after 10 days and the tibias harvested. The pull-out force of the steel screw was measured using a material-testing

Acta Orthopaedica 2018; 89 (6): 674–677

machine and the bone formation in the drill hole was measured using microCT. Another 24 mice received a mid-diaphyseal osteotomy in the femoral bone, stabilized by intramedullary nailing. The mice received a daily dose of either saline, teriparatide, or abaloparatide (n = 8 per group) for 28 days before they were killed. The material properties of the healed femur were measured by a 3-point bending test and the bone formation in the healing callus was measured using microCT. Animals 120 male C57BL/6 mice with a mean weight of 26 (SD 1.3) g were used. The animals were kept 4 per cage and given ad libitum access to food and water. Surgical procedure: tibia — following procedure has been described elsewhere (Sandberg and Aspenberg 2015). The mice were anesthetized with isoflurane gas and received a subcutaneous injection of 0.2 mg/kg oxytetracycline as infection prophylaxis and 0.1 mg/kg buprenorphine for postoperative analgesia. Both hind legs were shaved and cleaned with chlorhexidine. Under aseptic conditions, a 5 mm incision was made below the knee, at the anterio-medial side of the tibia on both legs. The muscle covering the medial side of the tibia was scraped aside to expose the bone surface. A drill hole into the proximal tibia was made by hand, using a 0.4 mm diameter syringe needle approximately 1.0 mm below the growth plate. A custom-made screw (Ti6A14V grade 2, thread size M 0.7) was inserted into the hole in the right leg. The skin was then sutured. Surgical procedure: femur — the following procedure has been described elsewhere (Sandberg and Aspenberg 2015). The preoperative preparation was the same as in the above section, with the difference that only the left leg was shaved and cleaned. A 7 mm longitudinal incision was made along the lateral thigh. The flexor and extensor muscles were separated bluntly, exposing the lateral femoral surface. The knee was opened laterally, and the patella was luxated medially. A 0.4 mm syringe needle was inserted as an intramedullary nail into the femur through the distal condyles. Before the needle was inserted, the tip of the needle was bent to act as a hook. Once inserted, the needle was rotated to fasten the hook into the proximal end of the femur, preventing the distal end of the needle from slipping out into the knee joint. The mid-femur was then exposed, and using a custom-designed pair of tongs the midshaft was cut. A suture was inserted to hold the patella in place before the skin was sutured. Drug treatment The mice were randomized to different treatment groups after surgery. The tibia-mice were randomly allocated to 11 different groups: control (saline), teriparatide at the following doses: 0.56; 1.67; 5; 15; or 45 µg/kg. or abaloparatide at 2.2; 6.7; 20; 60; or 180 µg/kg. The femur-mice were randomly allocated to 3 different groups: control (saline), teriparatide (15 µg/kg), or

675

abaloparatide (60 µg/kg). Treatment was administrated daily via subcutaneous injections in volumes of 0.03–0.05 mL. The animals were killed by cervical dislocation after sedation with isoflurane. The tibiae or femur were harvested for further analysis. Mechanical testing The following mechanical evaluations of tibia and femur has been described elsewhere (Sandberg and Aspenberg 2015). The mechanical properties of screw fixation and the femurs were analyzed using a computerized materials-testing machine (100R; DDL Inc., Eden Prairie, MN). The pull-out force of the screws was measured by a holder that was attached to the head of the screw and to the cross-head. The tibia was then fixed to the machine by a suture thread, looped around the bone next to the screw. The cross-head speed was 0.01 mm/s. Force at failure was recorded and defined as pull-out force. Stiffness was calculated by the machine after manual definition of the linear part of the load deformation curve, and energy was automatically calculated after a 10% drop from peak load. For the femurs, the intramedullary needle was removed before the femur was mounted on a stand, with both ends resting with 6 mm between supporting points with the lateral aspect upwards. The central part of the callus was then pushed downwards by a cross-head, which was lowered with a speed of 0.05 mm/s. The force at failure, stiffness, and energy uptake were recorded as described above. microCT The left tibiae, with a hole in the proximal metaphysis, and the femurs were analyzed with microCT (Skyscan 1174, v. 2; Bruker, Aarteselaar, Belgium). Topographic images of the bones with an isotropic voxel size of 11.2 μm were acquired at energy settings of 50 kV and 800 μA, using an aluminum filter of 0.25 mm, rotation step of 0.4°, and frame field averaging of 3 in a 180° scan. The images were reconstructed with NRecon (Skyscan, v. 1.6.8.0; Aarteselaar, Belgium) and corrected for ring artifacts and beam hardening. For the drill holes, a volume of interest was defined as a cylinder with a diameter of 0.3 mm extending 1.0 mm into the bone marrow cavity, starting at the endosteal side of the cortical bone. For the femurs, a volume of interest was defined as a 2 mm long segment of the central callus, delineated manually, which excluded the old cortex and medullar cavity. All definitions of volume of interest was done by a blinded examiner (MB), as was all other handling of the specimens. Calibration of bone mineral density was carried out by scanning 2 hydroxyapatite phantoms of known density (0.25 and 0.75 g/cm3). Analysis of total bone volume (BV/TV) were performed in CTAn (Skyscan, v. 1.10; Aarteselaar). Statistics The peak pull-out force of the tibial screw was pre-defined as the primary effect variable. In a linear regression analysis,

676

Acta Orthopaedica 2018; 89 (6): 674–677

Results from mechanical and morphometric analysis of femoral shaft fractures after 28 days of healing. Values are mean (SD) Group Force at failure (N) Stiffness (MPa) Energy uptake (Nmm) BV/TV (%)

Linear regression of mechanical data from metaphyseal screws in tibiae, 10 days after insertion.

pull-out force (N) was the dependent variable, and Ln(dose) (µg) and drug type were independent. Ln(dose) was a continuous variable, and drug was coded as 0 (teriparatide) or 1 (abaloparatide). The density (BV/TV) of the bone inside the burr hole was analyzed in the same way as the pull-out force. Ethics, funding, and potential conflicts of interest The study was approved by the Regional Ethics Committee for Animal Experiments and the animals were treated according to the institutional guidelines for care and treatment of laboratory animals. Funding was received from the Swedish Research Council (2031-47-5). Per Aspenberg has collaborated scientifically with Eli Lilly Corp, but received no grants or funding during the last 3 years.

Control Teriparatide Abaloparatide (n = 7) (15 µg/kg; n = 8) (60 µg/kg; n = 7) 26 (11) 20 (11) 14 (6.9) 23 (9.8)

31 (15) 27 (17) 12 (3.6) 38 (10)

36 (11) 27 (9.0) 14 (4.6) 47 (2.3)

N (SE 0.60; p = 0.02). This corresponds to dose increase by a factor of 2.5. Bone volume per total volume (BV/TV) in the tibial regions of interest showed slightly larger variation. In the linear regression analysis (r = 0.62), each increase in ln dose by 1 was associated with an increase in density by 10 percentage units (SE 1.7) regardless of drug type (p < 0.001). Changing drug from teriparatide to abaloparatide, regardless of dose, increased the density by 3 percentage units, but this was not statistically significant (p = 0.6). Diaphyseal model In the diaphyseal model, confidence intervals were too large for meaningful interpretation of 3-point bending tests (Table), but the bone density (BV/TV) of the callus was 23% (SD 10), 38% (SD 10), and 47% (SD 2) for control, teriparatide, and abaloparatide respectively. Both drugs produced significantly denser callus than controls (p = 0.001 and p = 0.008), but they were not significantly different from each other (Table).

Discussion Results Exclusions In the metaphyseal model, 3 mice (1 control, 2 abaloparatide; 6.67 and 20 µg/kg) died postoperatively due to injuries from fighting amongst themselves. 2 screws (1 control and 1 teriparatide 1.67 µg/kg) fell out during harvesting before mechanical testing, probably due to poor insertion during surgery. In the diaphyseal model, 2 specimens (1 control, 1 abaloparatide) were damaged during harvesting, before mechanical testing. Metaphyseal model In the metaphyseal model, the control pull-out force was 5.8 (2.2) N. There was a dose-dependent increase up to 14 (SD 2.5) N at the highest dose of abaloparatide (Figure). In a linear regression analysis (r = 0.78) each increase in ln dose by 1 (regardless of drug type) was associated with an increase in pull-out force by 1.5 N (SE 0.18; p < 0.001). Changing drug from teriparatide to abaloparatide increased the force by 1.4

Abaloparatide stimulated bone healing in both the diaphyseal and metaphyseal models. To our knowledge, abaloparatide has not been previously studied in a fracture-healing context. The higher dose of abaloparatide versus teriparatide used by Ledeer et al. (2015) in the clinical comparison trial (4:1) only partly corresponded to a difference in potency (2.5:1) in our mouse models. Our study has several weaknesses. Due to the assumption that abaloparatide would be 4 times more potent per µg, the highest dose was used only for this drug, and the lowest only for teriparatide in the metaphyseal model. The principal part of the study is the metaphyseal model, which has been thoroughly validated (Bernhardsson et al. 2015), while the diaphyseal model was used only as a complement. The pull-out force measurements provided the data with least random variation and comprised the predetermined primary outcome variable. The tibial density measurements support the mechanical results of a dose response but weaken the conclusion that abaloparatide is the more potent drug per µg. The mechanical evaluation of the diaphyseal model had large confidence intervals, while bone

Acta Orthopaedica 2018; 89 (6): 674–677

density measurements allowed the conclusions of a positive effect of both drugs also in diaphyseal healing. It can be disputed whether screw pull-out really reflects cancellous bone fracture healing. It is, however, clear that the force reflects the strength of the bone that has regenerated and grasped the screw threads during the 10 days after screw insertion, so it can always be stated that the force is a measure of local cancellous bone formation in response to trauma. This is at least very similar to cancellous bone healing (Bernhardsson et al. 2015). We are unaware of species differences in receptor binding characteristics, which could be different between mouse and man. Such differences might explain the differences in dose relations between the species. Even though this was a mouse study of fracture healing, it is tempting to speculate about the interpretation of the clinical abaloparatide teriparatide comparison trial (Leder et al. 2015). Possibly, the difference in effect between abaloparatide and teriparatide in this clinical trial partly reflects that abaloparatide was given at a higher dose, and that a moderate dose reduction would have eliminated the difference. This possibility is supported by the fact that the abaloparatide group also had more treatment discontinuations due to mild side effects, such as nausea, than teriparatide and placebo. In conclusion, abaloparatide and teriparatide stimulated bone regeneration in mouse fracture models to a similar extent.

677

Both authors were involved in designing the study, analyzing the data, preparing and approving the submitted manuscript. MB conducted the experiments and acquired the data. Acta thanks Ming Ding and Sten Rasmussen for help with peer review of this study.

Aspenberg P, Johansson T. Teriparatide improves early callus formation in distal radial fractures. Acta Orthop 2010; 81(2): 234-6. Aspenberg P, Malouf J, Tarantino U, García-Hernández P A, Corradini C, Overgaard S, Stepan J J, Borris L, Lespessailles E, Frihagen F, Papavasiliou K, Petto H, Caeiro J R, Marin F. Effects of teriparatide compared with risedronate on recovery after pertrochanteric hip fracture: results of a randomized, active-controlled, double-blind clinical trial at 26 weeks. J Bone Joint Surg Am 2016; 98(22): 1868-78. Bernhardsson M, Sandberg O, Aspenberg P. Experimental models for cancellous bone healing in the rat. Acta Orthop 2015; 86(6): 745-50. Boyce E G, Mai Y, Pham C. Abaloparatide: review of a next-generation parathyroid hormone agonist. Ann Pharmacother 2018; 52(5): 462-72. Campbell E J, Campbell G M, Hanley D A. The effect of parathyroid hormone and teriparatide on fracture healing. Expert Opin Biol Ther 2015; 15(1): 119-29. Leder B Z, O’Dea L S, Zanchetta J R, Kumar P, Banks K, McKay K, Lyttle C R, Hattersley G. Effects of abaloparatide, a human parathyroid hormonerelated peptide analog, on bone mineral density in postmenopausal women with osteoporosis. J Clin Endocrinol Metab 2015; 100(2): 697-706. Sandberg, O H, Aspenberg P. Glucocorticoids inhibit shaft fracture healing but not metaphyseal bone regeneration under stable mechanical conditions. Bone Joint Res 2015; 4(10): 170-5.

678

Acta Orthopaedica 2018; 89 (6): 678–682

High complication rate after extendible endoprosthetic replacement of the proximal tibia: a retrospective study of 42 consecutive children Panagiotis TSAGOZIS 1,2, Michael PARRY 1, and Robert GRIMER 1 1 The Royal Orthopaedic Hospital, Bristol Road South, Birmingham, UK; 2 Department of Orthopaedic Surgery, Karolinska University Hospital and Karolinska Institute, Stockholm, Sweden Correspondence: panagiotis.tsagkozis@sll.se Submitted 2018-02-26. Accepted 2018-09-04.

Background and purpose — The long-term outcome of reconstruction with extendible prostheses after resection of tumors the proximal tibia in children is unknown. We investigated the functional outcome, complication rate and final limb salvage rate after this procedure. Patients and methods — 42 children who had a primary extendible replacement of the proximal tibia for bone tumor with a Stanmore implant between 1992 and 2013 were identified in the department’s database. All notes were reviewed to identify the oncological and functional outcomes, the incidence of complications and the rate of amputation. 20 children were alive at final follow-up. Median follow-up time was 6 years and minimum follow-up for surviving patients was 3 years. Results — The overall limb salvage rate was 35/42; amputation was needed in 7 children. 15 implants were revised with a new implant. The Musculoskeletal Tumor Society Score was 73% (40–93) at final follow-up. The overall complication rate was 32/42. Soft tissue problems were the most common mode of complication, noted in 15 children, whereas structural failure and infection occurred in 12 children each. Use of prostheses with non-invasive lengthening was associated with a higher infection rate as compared with conventional ones (4/6 vs. 8/36) and inferior limb survival. Interpretation — Extendible replacements of the proximal tibia allow for limb salvage and satisfactory late functional outcome but have a high rate of complications. The use of non-invasive lengthening implants has not shown any benefit compared with conventional designs and is, rather, associated with higher risk for infection and amputation.

Extendible (growing) prostheses for surgical management of children with tumors of the lower extremity have been used since the mid-1970s. Initial designs allowed for extension through the replacement of modular parts, and later through the use of a screw-based extension mechanism (Henderson et al. 2012, Ruggieri et al. 2013, Albergo et al. 2017). More recently, non-invasive lengthening models have been introduced that rely on the presence of an external electromagnetic field that controls an internal motor or a spring mechanism for elongation (Gupta et al. 2006, Gilg et al. 2016). The use of extendible prostheses for the reconstruction of the proximal tibia has not been described in detail. The proximal tibia is the second most common site for primary skeletal sarcomas and contributes half of the overall growth of the tibia, and one-quarter of the growth of the lower limb. The ability to maintain length of this segment is important following tumor excision (Albergo et al. 2017). Early-generation implants and reconstruction of the extensor mechanism with a mesh, which is thought to predispose to infection, carried a high complication risk (Grimer et al. 2000). We investigated the outcome after primary reconstruction of the proximal tibia with modern, commercially available extendible endoprostheses in 42 children, using muscle flap reconstruction of the extensor mechanism in all children and non-invasive implants in 6 children.

Patients and methods This is a retrospective study of 42 consecutive children operated at the Royal Orthopaedic Hospital (Bristol Road South, Birmingham, UK) between 1992 and 2013 with extendible proximal tibia endoprosthesis for bone tumor. The implants were manufactured by Stanmore (Stanmore Implants Worldwide Ltd, Elstree, UK).

Acta Orthopaedica 2018; 89 (6): 678–682

Children were identified from the institution’s prospectively maintained database. All notes were reviewed. Children who had had previous surgery in the area (except for biopsy) were not included. There were 23 girls and 19 boys with a mean age of 10 years (6–13). The most common diagnosis was osteosarcoma (35 children) followed by Ewing’s sarcoma in 5, and other primary bone sarcomas in 2. Mean tumor size was 9 cm (5–17). Lung metastases at presentation were seen in 2 children, bone metastases in 1, and combined bone and lung metastases in another 2. All implants were custom-made. 36 used minimally invasive lengthening by a modified Allen key and 6 were noninvasive designs, relying on an external drive unit that generates a rotating electromagnetic field. Mean implant length was 17 cm (8–24) and similar between minimally invasive (mean length 17 cm) and non-invasive (mean length 16 cm) implants. 4 implants were surface treated with silver (Agluna, Accentus Medical, Oxford, UK). Implants were cemented into the intramedullary cavity of the remaining tibia, whereas the stem inserted in the proximal femur was uncemented in order to allow growth of the distal femoral physis. Surgical margins were clear in 41 children and there was a microscopically positive margin in 1 child. Reconstruction of the extensor mechanism and coverage of the fascial defect was done using a medial gastrocnemius flap. Antibiotic prophylaxis varied: intravenous cephalosporins were used in the first decade while a combination of flucloxacillin with gentamycin for 24 hours is the current regimen. Thromboembolism prophylaxis consisted of pneumatic compression devices on the contralateral leg as well as thromboembolic deterrent stockings. Children were allowed to fully weight-bear immediately. Mobilization was routinely done without the use of any brace or cast. Complications were described according to the system adopted by the International Society of Limb Salvage (ISOLS), and survival analysis was done according to the Kaplan–Meier method (Henderson et al. 2014). Functional outcomes were recorded using the modified scoring system of the Musculoskeletal Tumor Society (MSTS) (Enneking et al. 1993). Leg-length discrepancy was measured on calibrated whole-length radiographs of both extremities. There was no blinding in the evaluation of data. Oncological followup was standardized according to the European Society for Medical Oncology guidelines with clinical examination and chest radiographs every 3 months for the first 2 years, every 6 months up to the 5th year after surgery, and then annually for another 5 years. Implants were controlled for another 2 years with annual radiographs. Median follow-up time was 6 years and minimum follow-up for surviving patients was 3 years. Patients who had received amputation were included in the survival analysis. Statistics Statistical analysis was performed using SPSS (version 20,

679

Overall survival 1.0

0.8

0.6

0.4

0.2

0 0

100

120

Follow-up period (months)

Figure 1. Kaplan–Meier analysis for overall survival of 42 children who underwent proximal tibial replacement for bone tumors.

IBM Corp, Armonk, NY, USA) and STATA (version 13) software (StataCorp LLC, College Station, TX, USA). Survival curves were prepared according to the Kaplan–Meier method, and compared using the log-rank and modified Wilcoxon–Breslow test. Competing risk analysis was done using the method of Pepe and Mori. Analysis of possible prognostic factors was performed using the Cox proportional hazard test. Pearson’s chi-square (χ2) test was used for comparisons between groups. All tests were double-sided, and a p-value of ≤ 0.05 was considered significant. In parentheses, we present range of values unless otherwise stated. When 95% confidence intervals are shown in parentheses they are stated as CI. The core facility of the Statistics Department of the Karolinska Institute was consulted for data analysis. Ethics, funding, and potential conflicts of interest The work is a retrospective study that complies with Institutional Guidelines. No funding was received for this work. The authors declare no conflict of interest.

Results Oncological outcome Of the 42 children, 20 were alive at final follow-up. Overall survival (OS) was 63% (CI 47–79) at 5 years and 53% (CI 36–70) at 10 years (Figure 1). OS was not affected by tumor type. Local recurrence occurred in 2 children. Functional outcome Mean MSTS score was 57% (30–87) at 1 year and 73% (40–93) at final follow-up. Mean range of movement was 72° (15–130) at 1 year and 89° (25–120) at final follow-up. The mean extension lag at 1 year was 32° (0–90), diminishing to 17° (0–50) at final follow-up. Radiographs allowing final leg length measurement were available in 25 children and mean

680

Acta Orthopaedica 2018; 89 (6): 678–682

extensor mechanism due to persistent extensor deficit and 1 had excision of scar tissue and correction of a fixed flexion deformity with revision of the implant at a later time-point. 0.8 0.8 Type II failure (symptomatic aseptic loosening) was documented in 1 child, who had his 0.6 0.6 implant revised for this reason 9 years after insertion. 0.4 0.4 Type III (structural) failure occurred in 12 children. Failures of the expansion mechanism of the prosthesis were seen in 7 children, 5 of 0.2 0.2 whom were revised to another growing prosthesis; the remaining 2 refused reoperation. 0 0 Periprosthetic fractures occurred in 4 children. 0 20 40 60 80 100 120 0 20 40 60 80 100 120 Follow-up period (months) These were managed by cast treatment in 1, Follow-up period (months) Figure 2. (left panel) Kaplan–Meier analysis for prosthesis survival of 42 children who open reduction and internal fixation in 1, and underwent proximal tibial replacement for bone tumors. (right panel) Minimally invasive revision of the prosthesis in 1, while amputaimplants had a marginally superior survival as compared with non-invasive ones (p = 0.02 according to the modified Wilcoxon method and p = 0.2 according to the log-rank tion was required in 1 patient due to poor bone test, showing that minimally invasive implants have better survival only in the early follow- quality and the presence of severe soft-tissue up period). contractions. The anti-rotation lug failed in 1 child, whose implant was also revised. Type IV failure (infection) was seen in 12 children. Of Univariate Cox regression analysis of different complication types on prosthesis survival (PS) as well as limb survival (LS) of 42 chilthese, 5 eventually underwent amputation, as the infection dren who underwent proximal tibial replacement for bone tumors could not be controlled. 4 children underwent 2-stage revi(hazards ratio with 95% confidence intervals) sion, 1 of them twice. None of these 4 children ended in amputation. 2 children had a single-stage revision for infecType of Number of Hazard Hazard tion, 1 being successfully managed and 1 ending in amputacomplication complications ratio PS (CI) ratio LS (CI) tion. 6 children developed infections that were managed with I (soft tissue) 15 3 (1–9) 6 (1–37) antibiotics and soft-tissue procedures, but 4 of them eventuII (loosening) 1 0.6 (0.1–4.8) 1 (0–1000) ally required amputation as the infection could not be conIII (structural) 12 0.7 (0.2–2.1) 0.2 (0.1–1.6) trolled. There was a higher incidence of infection in the case IV (infectious) 12 3.5 (1.1–10) 14 (2–106) V (recurrence) 2 1.1 (0.3–5.0) 2.3 (0.3–19) of non-invasive designs (4/6 vs. 8/36, p = 0.05). The incidence of infection was similar between silver-coated implants and those not silver coated (11/36 vs. 1/6, p = 1). There was no diftibia lengthening was 1.5 cm (0–7). In 12/25 of cases tibia ference either between prophylactic antibiotic regimens used. Type V failure (local recurrence) was documented in 2 chillengthening was less than 0.5 cm, and in 10 of these this was attributed to local complications including infection and other dren. Of these, 1 was treated by amputation, the other by local soft-tissue problems (wound complications, stiffness, pain, excision and adjuvant radiotherapy. Type VI failure (growth arrest, joint deformity) was noted joint contraction) that often resulted in prolonged treatment and secondary operations that impeded lengthening, whereas in 1 of the children, who had a 6 cm limb shortening due to mechanical failure of the prosthesis was responsible for the growth arrest at the level of the distal femoral growth plate. remaining 2 cases of non-lengthening. The final average leg Prosthesis survival, limb salvage rate, and use of length difference at final follow-up was 1.8 cm (0–6). non-invasive implants Complications Prosthesis survival was 55% (CI 37–73) at 5 years and 25% Complications of any kind were observed in 32/42 children. (CI 7–43) at 10 years (Figure 2). 22/42 implants failed due to There was no correlation between the age of the patient and revision or amputation of the limb. The impact of the different types of complication on prosthesis survival is shown in the incidence of complications. Type I complications (soft tissue problems) arose in 15 chil- the Table. Infection was the sole factor predisposing to infedren. Although intensive physiotherapy was offered to all chil- rior prosthesis survival in univariate analysis. Silver-coated dren, 9 of them eventually developed persistent stiffness and implants were not associated with superior prosthesis survival underwent 1 or more manipulations of the joint under general (p = 0.8). There was a marginal difference regarding prosthesis anesthesia. 4 children experienced wound dehiscence and/or survival among minimally invasive and non-invasive implants, necrosis. 1 patient underwent Trevira tube augmentation of the with the latter showing a higher failure rate in the early followProsthesis survival

Prosthesis survival

1.0

Minimally invasive Non-invasive

Acta Orthopaedica 2018; 89 (6): 678–682

681

Limb survival

1.0

the growing individual poses challenges: the skeletal growth must be taken into consideration when planning reconstruction. The use 0.8 0.8 of extendible prostheses has gained popularity as it allows immediate mobilization and improved cosmesis for the patient (Henderson 0.6 0.6 et al. 2012, Picardo et al. 2012, Schinhan et al. 2015). For tumors arising in the proximal tibia, 0.4 0.4 however, concerns are that poor soft tissue coverage may increase the risk of infection and stiffness, as well as difficulty in reconstruction 0.2 0.2 of the extensor mechanism resulting in restricMinimally invasive Non-invasive tion in the range of movement (Albergo et al. 0 0 2017). Previous study in a limited number of 0 20 40 60 80 100 120 0 20 40 60 80 100 120 Follow-up period (months) Follow-up period (months) children, who were not included in this report, Figure 4. (left panel) Limb survival according to Kaplan–Meier, for 42 children who highlighted the high risk for complications underwent proximal tibial replacement for bone tumors. (right panel) Minimally invasive after extendible endoprosthetic reconstruction implants were associated with superior limb survival (p = 0.003 according to the log-rank for proximal tibia tumors (Grimer et al. 2000). test and p < 0.001 according to the modified Wilcoxon method). However, this option is chosen for many children as it allows for retaining a cosmetically up period (Figure 2). This association was not observed in a near-normal limb, which is not the case for amputation or competing risk survival analysis with death as a competing rotationplasty. On the other hand, rotationplasty (but not an above-knee amputation) allows for an active lifestyle includfactor (Figure 3, Supplementary data). Amputation was performed in 7 children. Kaplan–Meier ing sport activities, which are not compatible with endoprosanalysis showed a limb survival of 85% (CI 72–98) at 5 years thetic reconstruction due to the risk for mechanical failures. and 75% (CI 58–92) at 10 years (Figure 4). Infection was the We present our experience from the largest series of patients leading cause of amputation (5 children), whilst local recur- in the medical literature, showing that limb salvage is feasible rence accounted for 1 amputation and 1 was attributed to non- but at the expense of a high complication rate. This study has a number of limitations. First, it is a retrohealing fracture. The development of a postoperative infection was associated with significantly inferior limb survival (p = spective case series: prospective randomized studies are not 0.01). Other type of complications did not have a significant possible given the rarity of diagnosis. Second, for the same effect on limb survival in our series (Table). Furthermore, pos- reason, the study expands over 3 decades. However, all operasible confounding factors such as sex (p = 0.7 for prosthesis tions were done with similar implants from a single manufacsurvival and p = 0.8 for limb survival) and age (p = 0.3 for turer, following a common surgical technique and postoperaprosthesis survival and p = 0.3 for limb survival) had no statis- tive aftercare. Follow-up was also standardized. The functional outcome, as judged by the MSTS score, was tically significant effect. There was a higher incidence of infection in the case of non- satisfactory at final follow-up, although it was rather poor at 1 invasive designs (4/6 vs. 8/36, p = 0.05). The higher infec- year after surgery. The prostheses were also capable of retaintion rate observed in children with non-invasive implants was ing a good range of movement of the knee joint. Leg-length disreflected in a higher risk of requiring a secondary amputation. crepancy was mainly due to inability to lengthen the prosthesis, Use of minimally invasive implants was associated with a often due to local complications such as soft-tissue problems superior limb salvage rate as compared with non-invasive ones and infection requiring secondary operations and prolonged (Figure 4), even though this correlation did not reach statisti- medical treatment, rather than mechanical failure or femoral cal significance in a competing risk analysis with death as a growth disturbance. We found a high rate of complications in competing factor (Figure 5, Supplementary data). approximately 3 out of 4 children. Although soft-tissue prob lems were the most common type of complication, and often required repeated operations and resulted in a poor functional outcome, they do not usually pose a threat to the limb. On the Discussion other hand, infection remains the main concern, encountered Limb salvage is the preferred mode of treatment for malignant in almost one-third of children, and was the main threat for tumors of the extremities, since it results in superior functional amputation. The poor soft tissue envelope of the anterior tibia outcome, offering at the same time equivalent local control predisposes to the high rate of infection, which is high even and overall survival compared with amputation (Simon et when a gastrocnemius flap is used for soft-tissue reconstrucal. 1986, Rougraff et al. 1994). However, management of tion. Silver coating did not have a protective effect in our study,

682

in contrast to previous data, which is probably attributable to the low number of patients in this report (Hardes et al. 2010, Wafa et al. 2015). Once an infection developed, the chance of it being controlled with either a 1-stage or 2-stage revision was 0.8. For those treated with more conservative procedures or suppressive antibiotics, the risk of requiring an amputation to control the infection rose to 0.6. We also observed a high risk for structural failures, the commonest one being inability to lengthen the prosthesis or maintain lengthening due to faults in the extension mechanism. Other studies have also documented a high incidence of mechanical failures with expandable prostheses, highlighting the need for improvement in the design and manufacture of the implants (Eckardt et al. 1993, Ruggieri et al. 2013, Cipriano et al. 2015). We observed an increased rate of infection with a higher risk for amputation in the case of non-invasive implants, although the number of non-invasive implants in our series is too small to draw safe conclusions. Non-invasive implants are bulkier in order to accommodate the lengthening mechanism, which may explain their higher risk of infection. Size of the implant is probably the most critical factor regarding susceptibility to infection, as opposed to other sites with superior soft tissue coverage where non-invasive implants have shown low risk for type IV failure (Picardo et al. 2012, Schinhan et al. 2015, Gilg et al. 2016). We do not feel that antibiotic prophylaxis, which was the same as in any megaprosthetic reconstruction, was the main factor that contributed to the high infection rate. Advances in implant design, offering smaller implants, may decrease the risk for complications. Moreover, we believe that avoidance of soft-tissue tension during wound closure is crucial for avoiding infection. The prosthesis survival curve illustrates the fact that in most cases the initial implant is eventually revised. Although the prostheses used in our study are manufactured to be able to function as the final implant, in practice they are often revised to an adult-type implant at some stage. In summary, our results suggest that extendible endoprosthetic reconstruction of the tibia allows for limb preservation in most children but is unfortunately offset by a high risk of complications. In the majority of cases this risk is acceptable to the children and their families, but the risk for secondary amputation must be explained. Use of non-invasive prostheses has not resulted in a reduction of infections, contrary to surgeons’ expectations. Further developments in implant design are required to improve outcome, reducing the risk of both mechanical and infective complications. Other methods might also be considered, such as biological reconstruction or the use of conventional implants combined with epiphysiodesis in children near adolescence, as well as rotationplasty or amputation. Supplementary data Figures 3 and 5 are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/ 17453674.2018.1534320

Acta Orthopaedica 2018; 89 (6): 678–682

Study conception and design: RG. Data analysis: PT. Manuscript preparation. PT, MP, RG Acta thanks Gerard Schaap and Reinhard Windhager for help with peer review of this study.

Albergo J I, Gaston C L, Aponte-Tinao L A, Ayerza M A, Muscolo D L, Farfalli G L, Jeys L M, Carter S R, Tillman R M, Abudu A T, Grimer R J. proximal tibia reconstruction after bone tumor resection: are survivorship and outcomes of endoprosthetic replacement and osteoarticular allograft similar? Clin Orthop 2017; 475(3): 676-82. Cipriano C A, Gruzinova I S, Frank R M, Gitelis S, Virkus W W. Frequent complications and severe bone loss associated with the Repiphysis expandable distal femoral prosthesis. Clin Orthop 2015; 473(3): 831-8. Eckardt J J, Safran M R, Eilber F R, Rosen G, Kabo J M. Expandable endoprosthetic reconstruction of the skeletally immature after malignant bone tumor resection. Clin Orthop 1993; (297): 188-202. Enneking W F, Dunham W, Gebhardt M C, Malawar M, Pritchard D J. A system for the functional evaluation of reconstructive procedures after surgical treatment of tumors of the musculoskeletal system. Clin Orthop 1993; (286): 241-6. Gilg M M, Gaston C L, Parry M C, Jeys L, Abudu A, Tillman R M, Carter S R, Grimer R J. What is the morbidity of a non-invasive growing prosthesis? Bone Joint J 2016; 98-B(12): 1697-703. Grimer R J, Belthur M, Carter S R, Tillman R M, Cool P. Extendible replacements of the proximal tibia for bone tumours. J Bone Joint Surg Br 2000; 82(2): 255-60. Gupta A, Meswania J, Pollock R, Cannon S R, Briggs T W R, Taylor S, Blunn G. Non-invasive distal femoral expandable endoprosthesis for limb-salvage surgery in paediatric tumours. J Bone Joint Surg Br 2006; 88(5): 649-54. Hardes J, von Eiff C, Streitbuerger A, Balke M, Budny T, Henrichs M P, Hauschild G, Ahrens H. Reduction of periprosthetic infection with silver-coated megaprostheses in patients with bone sarcoma. J Surg Oncol 2010; 101(5): 389-95. Henderson E R, Pepper A M, Marulanda G, Binitie O T, Cheong D, Letson G D. Outcome of lower-limb preservation with an expandable endoprosthesis after bone tumor resection in children. J Bone Joint Surg Am 2012; 94(6): 537-47. Henderson E R, O’Connor M I, Ruggieri P, Windhager R, Funovics P T, Gibbons C L, Guo W, Hornicek F J, Temple H T, Letson G D. Classification of failure of limb salvage after reconstructive surgery for bone tumours: a modified system including biological and expandable reconstructions. Bone Joint J 2014; 96-B(11): 1436-40. Picardo N E, Blunn G W, Shekkeris A S, Meswania J, Aston W J, Pollock R C, Skinner J A, Cannon S R, Briggs T W. The medium-term results of the Stanmore non-invasive extendible endoprosthesis in the treatment of paediatric bone tumours. J Bone Joint Surg Br 2012; 94(3): 425-30. Rougraff B T, Simon M A, Kneisl J S, Greenberg D B, Mankin H J. Limb salvage compared with amputation for osteosarcoma of the distal end of the femur: a long-term oncological, functional, and quality-of-life study. J Bone Joint Surg Am 1994; 76(5): 649-56. Ruggieri P, Mavrogenis A F, Pala E, Romantini M, Manfrini M, Mercuri M. Outcome of expandable prostheses in children. J Pediatr Orthop 2013; 33(3): 244-53. Schinhan M, Tiefenboeck T, Funovics P, Sevelda F, Kotz R, Windhager R. Extendible prostheses for children after resection of primary malignant bone tumor: twenty-seven years of experience. J Bone Joint Surg Am 2015; 97(19): 1585-91. Simon M A, Aschliman M A, Thomas N, Mankin H J. Limb-salvage treatment versus amputation for osteosarcoma of the distal end of the femur. J Bone Joint Surg Am 1986; 68(9): 1331-7. Wafa H, Grimer R J, Reddy K, Jeys L, Abudu A, Carter S R, Tillman R M. Retrospective evaluation of the incidence of early periprosthetic infection with silver-treated endoprostheses in high-risk patients: case-control study. Bone Joint J 2015; 97-B(2): 252-7.

Acta Orthopaedica 2018; 89 (6): 683–688

683

Vancomycin concentrations in the cervical spine after intravenous administration: results from an experimental pig study Mats BUE 1,2, Pelle HANBERG 1,2, Mikkel TØTTRUP 3,4, Maja B THOMASSEN 2, Hanne BIRKE-SØRENSEN 2, Theis M THILLEMANN 2,4, Torben L ANDERSSON 5, and Kjeld SØBALLE 2,4 1 Department 3 Department 5 Department

of Orthopaedic Surgery, Horsens Regional Hospital, Horsens; 2 Orthopaedic Research Unit, Aarhus University Hospital, Aarhus; of Orthopaedic Surgery, Randers Regional Hospital, Randers; 4 Department of Orthopaedic Surgery, Aarhus University Hospital, Aarhus; of Clinical Biochemistry, Aarhus University Hospital, Aarhus, Denmark Correspondence: matsbue6@rm.dk Submitted 2018-04-17. Accepted 2018-07-03.

Background and purpose — Vancomycin may be an important drug for intravenous perioperative antimicrobial prophylaxis in spine surgery. We assessed single-dose vancomycin intervertebral disc, vertebral cancellous bone, and subcutaneous adipose tissue concentrations using microdialysis in a pig model. Material and methods — 8 female pigs received 1,000 mg of vancomycin intravenously as a single dose over 100 minutes. Microdialysis probes were placed in the C3–C4 intervertebral disc, C3 vertebral cancellous bone, and subcutaneous adipose tissue, and vancomycin concentrations were obtained over 8 hours. Venous blood samples were obtained as reference. Results — Ranging from 0.24 to 0.60, vancomycin tissue penetration, expressed as the ratio of tissue to plasma area under the concentration-time curve from 0 to the last measured value, was incomplete for all compartments. The lowest penetration was found in the intervertebral disc. The time to a mean clinically relevant minimal inhibitory concentration (MIC) of 4 μg/mL was 3, 17, 25, and 156 min for plasma, subcutaneous adipose tissue, vertebral cancellous bone, and the intervertebral disc, respectively. In contrast to the other compartments, a mean MIC of 8 μg/mL was not reached in the intervertebral disc. An approximately 3-times longer elimination rate was observed in the intervertebral disc in comparison with all the other compartments (p < 0.001), and the time to peak drug concentration was higher for all tissues compared with plasma Interpretation — Preoperative administration of 1,000 mg of vancomycin may provide adequate vancomycin tissue concentrations with a considerable delay, though tissue penetration was incomplete. However, in order also to achieve adequate intervertebral disc concentrations in all individuals and accommodating a potentially higher MIC target, supplemental application of vancomycin may be necessary.

Postoperative surgical site infections following spine surgery can have devastating complications such as mechanical fixation failure, neurological injury, pseudoarthrosis, and sepsis (Fang et al. 2005). The reported incidence of postoperative spondylodiscitis ranges between 1% and 4% and may be even higher when instrumentation is applied (Gaynes et al. 2001, Deyo et al. 2004, Fang et al. 2005, Smith et al. 2011). Perioperative antimicrobial prophylaxis plays a key role in lowering the risk of postoperative spondylodiscitis. The antimicrobial effect relies not only on its sensitivity against the invading bacteria, but also on adequate target site concentrations. For vancomycin, tissue targets for prevention of postoperative spondylodiscitis have not been established. Nonetheless, it seems prudent to maintain tissue concentrations that, as a minimum, exceed minimal inhibitory concentrations (MIC) of relevant bacteria throughout surgery and until a few hours after the incision is closed (Mangram et al. 1999, Whiteside 2016). Vancomycin may become increasingly important as perioperative antimicrobial prophylaxis in spine surgery due to the increasing incidence of methicillin-resistant Staphylococcus aureus (MRSA) infections (Al-Nammari et al. 2007, Gouliouris et al. 2010). However, routine use of vancomycin is not indicated due to the risk of resistance development (Mangram et al. 1999, Bryson et al. 2016). Incomplete vancomycin penetration has recently been reported for a number of tissues (Bue et al. 2018a, b). Given the avascular nature of the intervertebral disc, vancomycin penetration may be inadequate for optimal prevention of postoperative spondylodiscitis. It is challenging to determine antimicrobial spine tissue concentrations. Intervertebral disc and bone concentrations of vancomycin have almost only been evaluated using bone and disc tissue samples and discectomy (Scuderi et al. 1993, Conaughty et al. 2006, Landersdorfer et al. 2009, Komatsu et al. 2010). However, these methods suffer from methodological challenges, which makes it difficult to relate the results

684

to relevant pharmacokinetic/pharmacodynamic endpoints (Landersdorfer et al. 2009, Pea 2009). Recently, microdialysis has evolved as a promising method for sampling various antimicrobials in different types of tissues, including bone and the intervertebral disc (Stolle et al. 2004, Bue et al. 2015, Tottrup et al. 2015, Hanberg et al. 2016, Bue et al. 2018a, b). We assessed single-dose vancomycin concentrations in the C3–C4 intervertebral disc, the C3 vertebral cancellous bone, and subcutaneous adipose tissue using microdialysis in a pig model mimicking a perioperative situation. Tissue penetration ratios, expressed as the ratio of tissue to plasma area under the concentration-time curve from 0 to the last measured value (AUCtissue/AUCplasma), and time to mean MICs of 2, 4, and 8 μg/mL were the primary endpoints. The secondary endpoints were pharmacokinetic parameters: the area under the concentration-time curves (AUC0–last), peak drug concentration (Cmax), time to Cmax (Tmax), and half-life (T1/2).

Material and methods This study was conducted at the Institute for Clinical Medicine, Aarhus University Hospital, Aarhus, Denmark. All chemical analyses were performed at the Department of Clinical Biochemistry, Aarhus University Hospital, Aarhus, Denmark. Overview 8 female pigs were included in the study (Danish Landrace Breed; weight 78–82 kg). Vancomycin was administered intravenously as a single dose of 1,000 mg over 100 min, and sampling was conducted over 8 hours starting at the beginning of vancomycin infusion. Vancomycin concentrations were obtained using microdialysis in the C3–C4 intervertebral disc, the C3 vertebral cancellous bone, and subcutaneous adipose tissue. Anesthesia and surgical procedures The pigs were kept under general anesthesia using a combination of fentanyl (0.35–0.5 mg/h, continuous infusion) and propofol (500–600 mg/h, continuous infusion) during the surgery and the sampling period. Arterial pH was monitored throughout the study and kept in the range of 7.36–7.47 by regulating ventilation. Blankets were used to keep the core temperatures within the range of 36.2–39.1ºC. Immediately after induction of anesthesia, the surgical procedures were initiated. With the pig in supine position, and under fluoroscopic guidance, the C2–C4 vertebrae were exposed via an anterolateral incision. At approximately 45º to the sagittal plane, a drill hole with a diameter of 2 mm and a depth of 25 mm was created in the middle of C3. Parallel to this drill hole, a Kirschner wire with a fixating device (PEBAX, M Dialysis AB, Stockholm, Sweden) was drilled into the caudal part of C2. A microdialysis probe (membrane length 20 mm) was inserted into a splittable introducer with the membrane protruding approximately 30 mm from the tip

Acta Orthopaedica 2018; 89 (6): 683–688

of the introducer. This probe-introducer setup was then placed in the drill hole in C3 and fixed to the fixating device; an endo clip was then attached to the introducer. These steps were taken to avoid subsequent displacement of the probe. At the same angle, a splittable introducer with a needle was introduced into the intervertebral disc between C3 and C4 parallel to, and in the middle of, the adjacent endplates. After the annulus fibrosus was penetrated, the needle was retracted, and the introducer was carefully advanced into the nucleus pulposus until resistance from the opposite wall of the annulus fibrosus was felt. A microdialysis probe (membrane length 10 mm) was then placed in the introducer, and the splittable introducer was retracted until the entire membrane of the probe was exposed in the intervertebral disc. The probe was attached to the introducer with endo clips. Fluoroscopy was used to assess correct location of the probes in the C3 vertebral body and the C3–C4 intervertebral disc. In addition to the bone and intervertebral disc probes, a subcutaneous adipose tissue probe (membrane length 20 mm) was placed in the lateral part of the right thigh, based on the manufacturer’s guidelines. Microdialysis and sampling procedures A detailed description of microdialysis can be found elsewhere (Muller 2002, Joukhadar and Muller 2005). Briefly, microdialysis is a probe-based method that allows for serial sampling of water-soluble molecules from the extracellular fluid in the tissue of interest by means of a semipermeable membrane at the tip of the microdialysis probe (Joukhadar et al. 2001, Hutschala et al. 2013). Due to continuous perfusion of the probe, a non-equilibrium diffusion of molecules following the concentration gradient will occur. Consequently, the concentration in the dialysate represents only a fraction of the true tissue concentration, which is expressed as the relative recovery. Thus, relative recovery must be determined to calculate the absolute tissue concentrations. In the present study, all the microdialysis probes were individually calibrated at the end of the study using the retrodialysis by drug method (Scheller and Kolb 1991). The relative recovery was calculated using the following Equation 1: Relative recovery (%) = 100 × ( 1 – Cdialysate/Cperfusate)

(1)

where Cdialysate is the concentration (μg/mL) in the dialysate and Cperfusate is the concentration (μg/mL) in the perfusate. Absolute, extracellular concentrations (μg/mL), Ctissue, were calculated by correcting for relative recovery using the following Equation 2: Ctissue = 100 × Cdialysate/Relative recovery (%)

(2)

The microdialysis system consisted of CMA 107 precision pumps (M Dialysis AB, Stockholm, Sweden) and CMA 70 probes (membrane length 20 mm and 10 mm, molecular cutoff 20 kilo Daltons). All the microdialysis probes were perfused with 0.9% NaCl at a perfusion rate of 1 μL/min throughout the sampling time. Following a 30-min tissue equilibration

Acta Orthopaedica 2018; 89 (6): 683–688

685

Pharmacokinetic parameters for plasma, subcutaneous adipose tissue, vertebral cancellous bone, and intervertebral disc Tissue

AUC0–last (min μg/mL)

Cmax (μg/mL)

Tmax (min)

Plasma (unbound) Subcutaneous adipose tissue Vertebral cancellous bone Intervertebral disc p-value a

7,880 (7164–8597) b 4,719 (4002–5436) 3,677 (2960–4393) c 1,983 (1237–2729) d < 0.001

40.0 (35.7–44.3) b 18.0 (14.1–21.9) 12.3 (9.6–15.0) c 6.6 (3.6–9.6)4 < 0.001

75 (61–89) 113 (96–129) 159 (134–184) 270 (187–353) –

T1/2 (min) 325 (99–552) 224 (197–250) 271 (227–315) 933 (527–1,339) e < 0.008

AUCtissue/AUCplasma 0.60 (0.48–0.72) 0.46 (0.40–0.53) 0.24 (0.17–0.31)

Values are given as means (95% confidence interval). AUC0–last, area under the concentration-time curve from 0 to the last measured value; Cmax, peak drug concentration; Tmax, time to Cmax; T1/2, half-life at β-phase; AUCtissue/AUCplasma, tissue penetration expressed as the ratio of AUCtissue/AUCplasma. a Overall comparison using F test for plasma (unbound), subcutaneous adipose tissue, vertebral cancellous bone, and intervertebral disc. b p < 0.001 for all comparisons between plasma and the other compartments. c p < 0.01 for comparison with subcutaneous adipose tissue. d p < 0.01 for comparison between intervertebral disc and the other compartments. e p < 0.001 for all comparisons between intervertebral disc and the other compartments.

period after placement of the microdialysis probes, 1,000 mg of vancomycin was administered intravenously over 100 min starting at time zero. For the first 2 hours, dialysates were harvested at 40-min intervals and, thereafter, at 60-min intervals for the following 6 hours, resulting in a total of nine samples over 8 hours. The relative recovery-corrected dialysate concentrations were ascribed to the midpoint of each sampling interval. Venous blood samples were harvested from a central venous catheter at the midpoint of the same 9 sampling intervals. When the last dialysate was collected, all the probes were individually calibrated using 0.9% NaCl with vancomycin at a concentration of 300 μg/mL by collecting 60-min samples. This high calibration concentration of vancomycin was chosen to minimize the influence of the residual local tissue concentrations. After the last dialysates were collected, the pigs were killed using pentobarbital. The dialysates were immediately placed in a –80°C freezer until analysis. The venous blood samples were stored at 5°C for a maximum of eight hours before being centrifuged at 3,000 g for 10 minutes. The plasma aliquots were then frozen and stored at –80°C until analysis. Chemical analysis of vancomycin The vancomycin concentrations in the dialysates were quantified using ultra-high-performance liquid chromatography as previously described (Bue et al. 2015). The quantification limit was defined as the lowest concentration with intra-run CV < 20%; it was found to be 0.05 µg/mL. The free concentration of vancomycin in plasma was measured with a homogeneous enzyme immunoassay technique using the Siemens Chemistry XPT platform (Advia Chemistry, Erlangen, Germany). Intrarun (total) imprecisions for the assay were ±1.2 µg/mL (2SD) at 6.6 µg/mL and ±3.7 µg/mL (2SD) at 29.1 µg/mL. Pharmacokinetic analysis and statistics Using Microsoft Excel (Microsoft Corp, Redmond, WA, USA), the time to mean clinically relevant MICs of 2, 4, and

8 μg/mL was estimated using linear interpolation. The pharmacokinetic parameters, AUC0–last, Cmax, Tmax, and T1/2, were determined separately for each compartment for each pig by non-compartmental analysis using the pharmacokinetic-series of commands in Stata (v. 14.1, StataCorp LLC, College Station, TX, USA). The AUC0–last was calculated using the trapezoidal rule. The tissue AUC0–last to plasma AUC0–last ratio (AUCtissue/AUCplasma) was calculated as a measure of tissue penetration. Cmax was calculated as the maximum of all the recorded concentrations and Tmax was calculated as the time to Cmax. T1/2 was calculated as ln(2)/λeq, where λeq is the terminal elimination rate constant estimated by linear regression of the log concentration on time. These pharmacokinetic parameters were obtained in all 4 compartments from the same pig and a mixed model for repeated measurements had compartments as fixed effect and subject identification variable as a random effect was applied. Also, distinct residual variance was assumed within each compartment. The normality of the residuals was estimated using a quantile–quantile (QQ) plot for the residuals and the homogeneity of the residual variance was checked by plotting residuals vs. best linear unbiased prediction estimates. The normality of the estimated random effects was checked using a QQ plot of the estimated random effects. A correction for degrees of freedom due to small sample size was handled using the Kenward–Roger approximation method. The F-test was used to determine the overall comparisons between the compartments and the t-test was used to determine pairwise comparisons. A p-value < 0.05 was considered to be significant. Statistical analyses were also performed using Stata. Values below the lower limit of quantification were set to zero. The means and 95% CI of AUC0–last, Cmax, Tmax, and T1/2 are presented in the Table. Ethics, funding, and potential conflicts of interest The study was approved by the Danish Animal Experiments Inspectorate and carried out according to existing laws (license No. 2017/15-0201-01184).

686

Acta Orthopaedica 2018; 89 (6): 683–688

Figure 1. Representative fluoroscopic image showing the location of the microdialysis (MD) probes in a sagittal view: the Kirschner wire with the fixating device in the C2 vertebral body, C3 and C4 vertebral body, C3–C4 intervertebral disc (IVD), the gold thread within the microdialysis probe membrane tip in the vertebral cancellous bone (VCB) and intervertebral disc.

Figure 2. Mean concentration-time profiles for plasma, subcutaneous adipose tissue, vertebral cancellous bone, and the intervertebral disc. Bars represent 95% confidence intervals. MICs of 2, 4, and 8 μg/mL are also inserted.

This work was supported by unrestricted grants from the Augustinus Foundation, the Lippmann Foundation, the Knud and Edith Eriksens Memorial Foundation, the Søster and Verner Lipperts Foundation, and the Health Research Fund of Central Denmark Region. No competing interests were declared.

for vertebral cancellous bone, 18 μg/mL (14–22) for subcutaneous adipose tissue, and 40 μg/mL (36–44) for plasma. The Tmax findings revealed delayed tissue penetration, particularly to the intervertebral disc and vertebral cancellous bone. Furthermore, T1/2 was approximately 3 times longer in the intervertebral disc in comparison with the other compartments (p < 0.001). Finally, AUC0–last and Cmax were lower in the intervertebral disc than in the vertebral cancellous bone (p < 0.01).

Results All 8 pigs completed the study. Except for one malfunctioning intervertebral disc probe, data were obtained from all probes. Fluoroscopy confirmed correct placement of all of the probes (Figure 1). Mean (SD) relative recovery for the intervertebral disc, vertebral cancellous bone, and subcutaneous adipose tissue was 16% (6), 36% (4), and 33% (5), respectively. Tissue penetration (AUCtissue/AUCplasma) of vancomycin (95% CI) was incomplete for the subcutaneous adipose tissue 0.60 (0.48–0.72), vertebral cancellous bone 0.46 (0.40–0.53), and intervertebral disc 0.24 (0.17–0.31). After 15 min, a mean concentration of 2 μg/mL (MIC) was reached in all compartments. The time to a mean MIC of 4 μg/mL was 3, 17, 25, and 156 min for plasma, subcutaneous adipose tissue, vertebral cancellous bone, and the intervertebral disc, respectively. A mean MIC of 8 μg/mL could not be reached in the intervertebral disc, whereas it was reached after 7, 37, and 81 min in plasma, subcutaneous adipose tissue, and vertebral cancellous bone, respectively. The vancomycin tissue and plasma concentration-time profiles are shown in Figure 2. The pharmacokinetic parameters are presented in the Table. Cmax (95% CI) was 6.6 μg/ mL (3.6–9.6) for the intervertebral disc, 12 μg/mL (9.6–15)

Discussion To our knowledge, this is the first study to investigate singledose vancomycin intervertebral disc and vertebral cancellous bone concentrations using microdialysis. Insufficient perioperative antimicrobial target site penetration might play an important role for the rather high incidence of postoperative spondylodiscitis. A key finding of this study was therefore incomplete and delayed intervertebral disc and vertebral cancellous bone penetration of vancomycin, with the lowest and most delayed penetration found in the intervertebral disc. However, using standard recommendations for prevention of surgical site infections and planktonic MICs of commonly encountered bacteria in spine surgery (0.5–4 μg/mL), adequate mean concentrations were achieved in all compartments, although a considerable delay was found in the intervertebral disc (Gouliouris et al. 2010, EUCAST 2017). Advantageously for a perioperative prophylactic setting, it should be noted that an approximately 3 times longer vancomycin elimination rate was found in the intervertebral disc in comparison with the other compartments. Thus, if vancomycin is administrated in due time, adequate intervertebral disc concentrations may be

Acta Orthopaedica 2018; 89 (6): 683â&#x20AC;&#x201C;688

sustained throughout even long surgical procedures and for some time after. This makes vancomycin attractive because it continues to kill bacteria even after surgery has ended. On the other hand, our data suggest a rather narrow or no margin at all in individuals with low intervertebral disc concentrations to bacteria exhibiting high MICs. Moreover, in the case of MRSA, increasing vancomycin MICs have been demonstrated over the last decades (Steinkraus et al. 2007). The traditional target recommendations for prevention of surgical site infections lack scientific evaluation and may in fact be insufficient for spine surgery, particularly when considering the possible devastating complications of infection in spine surgery. Higher and prolonged tissue concentrations cannot be achieved by means of increasing intravenous vancomycin doses as this is restricted by toxicity (Rybak et al. 2009). Consequently, our findings call for some considerations regarding vancomycin dosing, timing, and administration in the perioperative spine setting. Preoperative administration of 1,000 mg of vancomycin may provide adequate vancomycin tissue concentrations with a considerable delay. However, in order also to achieve adequate intervertebral disc concentrations in all individuals and accommodating a potentially higher MIC target, supplemental application (e.g. as perioperative powder) of vancomycin may be necessary (Murphy et al. 2017). Incomplete and heterogeneous tissue distribution of antimicrobials has been demonstrated for a diverse combination of drugs and tissues and under different conditions, including infections (Joukhadar et al. 2001, Joukhadar and Muller 2005, Schintler et al. 2009, Hutschala et al. 2013, Tottrup et al. 2015, 2016). In our study, the tissue penetration ratios for vancomycin were lower for all tissues in comparison with an analogous cefuroxime study (Hanberg et al. 2016). These findings suggest that the choice of antimicrobial prophylaxis should not only be based on the characteristics of the infectious bacteria and plasma pharmacokinetics, but also on the tissue pharmacokinetics for the specific drug and the specific setting. In terms of infected tissue, it has previously been shown that vancomycin bone penetration may decrease with progression of an infection (Bue et al. 2018a). This emphasizes the need for spine tissue pharmacokinetic studies of relevant antimicrobials under different conditions. The penetration of vancomycin into a human intervertebral disc may vary from our findings in the study, as juvenile pigs (aged 5 months) differ from adult humans in several ways (Alini et al. 2008). First, blood vessels in the human annulus fibrosus are only present in the first part of life; thereafter, the perfusion relies upon only diffusion from the endplates (Roberts et al. 2006, Gouliouris et al. 2010). Second, the intervertebral disc is thinner in pigs than in humans, indicating shorter diffusion distances (Alini et al. 2008). Third, the body mass and weight-bearing impact on the vertebral bodies and intervertebral disc differs between humans and pigs. Moreover, higher vancomycin concentrations have been demonstrated in pig bone and tissue concentrations in comparison with male

687

patients undergoing total knee replacement surgery (Bue et al. 2015, Bue et al. 2018b). Until now, vancomycin bone and intervertebral disc concentrations have only been assessed using bone and disc tissue samples and discectomy (Scuderi et al. 1993, Conaughty et al. 2006, Landersdorfer et al. 2009, Komatsu et al. 2010). In contrast, microdialysis allows for serial sampling of the unbound extracellular concentrations of drug in bone and in the intervertebral disc, and it provides dynamic concentrationtime profiles (Joukhadar and Muller 2005, Hanberg et al. 2016). Therefore, the pharmacokinetic parameters obtained by microdialysis are useful for evaluating pharmacodynamic/ pharmacokinetic targets. In addition to the inherent inter-species limitations of a porcine study, a certain methodological aspect of the microdialysis approach should be considered when evaluating microdialysis data. Thus, to obtain absolute tissue concentrations, the actual measured concentrations are corrected for relative recovery. This leads to a magnification of the variations associated with pre-analytical sample handling and chemical assay. These variations will increase exponentially as the relative recovery decreases. The variances in plasma and tissue pharmacokinetics found in the present study were comparable, indicating acceptable precision of the measurements within the biological variation. In summary, vancomycin penetration into healthy pig intervertebral disc and vertebral cancellous bone was found to be incomplete and delayed, with the lowest and most delayed penetration to the intervertebral disc. However, applying standard recommendations for prevention of postoperative spondylodiscitis, preoperative administration of 1,000 mg of vancomycin may provide adequate vancomycin tissue concentrations with a considerable delay. Nonetheless, in order also to achieve adequate intervertebral disc concentrations in all individuals and accommodating a potentially higher MIC target, supplemental application of vancomycin may be necessary. Validation of these findings in a clinical setting is warranted. MB, PH, MT, MBT, HBS, TMT, and KS initiated and designed the study. MB, MBT, and PH conducted the surgery and MB placed all the probes. MB, MBT, and PH collected the data. TLA performed the chemical analyses. Statistical analysis and interpretation of data was done by MB, PH, MT, HBS, TMT, TLA, and KS. All authors drafted and revised the manuscript. Acta thanks Volker Alt and Ivar Rossvoll for help withÂ peer review of this study.

Al-Nammari S S, Lucas J D, Lam KS . Hematogenous methicillin-resistant Staphylococcus aureus spondylodiscitis. Spine (Phila Pa 1976) 2007; 32(22): 2480-6. Alini M, Eisenstein S M, Ito K, Little C, Kettler A A, Masuda K, et al. Are animal models useful for studying human disc disorders/degeneration? Eur Spine J 2008; 17(1): 2-19. Bryson D J, Morris D L, Shivji F S, Rollins K R, Snape S, Ollivere B J. Antibiotic prophylaxis in orthopaedic surgery: difficult decisions in an era of evolving antibiotic resistance. Bone Joint J 2016; 98-B(8): 1014-19.

688

Bue M, Birke-Sorensen H, Thillemann T M, Hardlei T F, Soballe K, Tottrup M. Single-dose pharmacokinetics of vancomycin in porcine cancellous and cortical bone determined by microdialysis. Int J Antimicrob Agents 2015; 46(4): 434-8. Bue M, Hanberg P, Koch J, Jensen L K, Lundorff M, Aalbaek B, et al. Singledose bone pharmacokinetics of vancomycin in a porcine implant-associated osteomyelitis model. J Orthop Res 2018a; 36(4): 1093-1098 Bue M, Tottrup M, Hanberg P, Langhoff O, Birke-Sorensen H, Thillemann T M, et al. Bone and subcutaneous adipose tissue pharmacokinetics of vancomycin in total knee replacement patients. Acta Orthop 2018b; 89(1): 95-100 Conaughty J M, Chen J, Martinez O V, Chiappetta G, Brookfield K F, Eismont F J. Efficacy of linezolid versus vancomycin in the treatment of methicillin-resistant Staphylococcus aureus discitis: a controlled animal model. Spine (Phila Pa 1976) 2006; 31(22): E830-2. Deyo R A, Nachemson A, Mirza S K. Spinal-fusion surgery: the case for restraint. N Engl J Med 2004; 350(7): 722-6. EUCAST. European Committee on Antimicrobial Susceptibility Testing. European 2017; data from the EUCAST MIC distribution website. https:// mic.eucast.org/Eucast2/SearchController/search.jsp?action=init Fang A, Hu S S, Endres N, Bradford D S. Risk factors for infection after spinal surgery. Spine (Phila Pa 1976) 2005; 30(12): 1460-5. Gaynes R P, Culver D H, Horan T C, Edwards J R, Richards C, Tolson J S. Surgical site infection(SSI) rates in the United States, 1992â&#x20AC;&#x201C;1998: the National Nosocomial Infections Surveillance System basic SSI risk index. Clin Infect Dis 2001; 33(Suppl. 2): S69-77. Gouliouris T, Aliyu S H, Brown N M. Spondylodiscitis: update on diagnosis and management. J Antimicrob Chemother 2010; 65(Suppl 3): iii11-24. Hanberg P, Bue M, Birke Sorensen H, Soballe K, Tottrup M. Pharmacokinetics of single-dose cefuroxime in porcine intervertebral disc and vertebral cancellous bone determined by microdialysis. Spine J 2016; 16(3): 432-8. Hutschala D, Skhirtladze K, Kinstner C, Zeitlinger M, Wisser W, Jaeger W, et al. Effect of cardiopulmonary bypass on regional antibiotic penetration into lung tissue. Antimicrob Agents Chemother 2013; 57(7): 2996-3002. Joukhadar C, Frossard M, Mayer B X, Brunner M, Klein N, Siostrzonek P, et al. Impaired target site penetration of beta-lactams may account for therapeutic failure in patients with septic shock. Crit Care Med 2001; 29(2): 385-91. Joukhadar C, Muller M. Microdialysis: current applications in clinical pharmacokinetic studies and its potential role in the future. Clin Pharmacokinet 2005; 44(9): 895-913. Komatsu M, Takahata M, Sugawara M, Takekuma Y, Kato T, Ito M, et al. Penetration of linezolid into rabbit intervertebral discs and surrounding tissues. Eur Spine J 2010; 19(12): 2149-55. Landersdorfer C B, Bulitta J B, Kinzig M, Holzgrabe U, Sorgel F. Penetration of antibacterials into bone: pharmacokinetic, pharmacodynamic and bioanalytical considerations. Clin Pharmacokinet 2009; 48(2): 89-124. Mangram A J, Horan T C, Pearson M L, Silver L C, Jarvis W R. Guideline for prevention of surgical site infection, 1999. Centers for Disease Control and Prevention (CDC) Hospital Infection Control Practices Advisory Committee. Am J Infect Control 1999; 27(2): 97-132; quiz 3-4; discussion 96.

Acta Orthopaedica 2018; 89 (6): 683â&#x20AC;&#x201C;688

Muller M. Science, medicine, and the future: microdialysis. BMJ 2002; 324(7337): 588-91. Murphy E P, Curtin M, Shafqat A, Byrne F, Jadaan M, Rahall E. A review of the application of vancomycin powder to posterior spinal fusion wounds with a focus on side effects and infection: a prospective study. Eur J Orthop Surg Traumatol 2017; 27(2): 187-91. Pea F. Penetration of antibacterials into bone: what do we really need to know for optimal prophylaxis and treatment of bone and joint infections? Clin Pharmacokinet 2009; 48(2): 125-7. Roberts S, Evans H, Trivedi J, Menage J. Histology and pathology of the human intervertebral disc. J Bone Joint Surg Am 2006; 88 Suppl. 2: 10-14. Rybak M, Lomaestro B, Rotschafer J C, Moellering R Jr, Craig W, Billeter M, et al. Therapeutic monitoring of vancomycin in adult patients: a consensus review of the American Society of Health-System Pharmacists, the Infectious Diseases Society of America, and the Society of Infectious Diseases Pharmacists. Am J Health Syst Pharm 2009; 66(1): 82-98. Scheller D, Kolb J. The internal reference technique in microdialysis: a practical approach to monitoring dialysis efficiency and to calculating tissue concentration from dialysate samples. J Neurosci Methods 1991; 40(1): 31-8. Schintler M V, Traunmuller F, Metzler J, Kreuzwirt G, Spendel S, Mauric O, et al. High fosfomycin concentrations in bone and peripheral soft tissue in diabetic patients presenting with bacterial foot infection. J Antimicrob Chemother 2009; 64(3): 574-8. Scuderi G J, Greenberg S S, Banovac K, Martinez O V, Eismont F J. Penetration of glycopeptide antibiotics in nucleus pulposus. Spine (Phila Pa 1976) 1993; 18(14): 2039-42. Smith J S, Shaffrey C I, Sansur C A, Berven S H, Fu K M, Broadstone P A, et al. Rates of infection after spine surgery based on 108,419 procedures: a report from the Scoliosis Research Society Morbidity and Mortality Committee. Spine (Phila Pa 1976) 2011; 36(7): 556-63. Steinkraus G, White R, Friedrich L. Vancomycin MIC creep in non-vancomycin-intermediate Staphylococcus aureus (VISA), vancomycin-susceptible clinical methicillin-resistant S. aureus (MRSA) blood isolates from 2001-05. J Antimicrob Chemother 2007; 60(4): 788-94. Stolle L B, Arpi M, Holmberg-Jorgensen P, Riegels-Nielsen P, Keller J. Application of microdialysis to cancellous bone tissue for measurement of gentamicin levels. J Antimicrob Chemother 2004; 54(1): 263-5. Tottrup M, Bibby B M, Hardlei T F, Bue M, Kerrn-Jespersen S, Fuursted K, et al. Continuous versus short-term infusion of cefuroxime: assessment of concept based on plasma, subcutaneous tissue, and bone pharmacokinetics in an animal model. Antimicrob Agents Chemother 2015; 59(1): 67-75. Tottrup M, Bue M, Koch J, Jensen L K, Hanberg P, Aalbaek B, et al. Effects of implant-associated osteomyelitis on cefuroxime bone pharmacokinetics: assessment in a porcine model. J Bone Joint Surg Am 2016; 98(5): 363-9. Whiteside L A. Prophylactic peri-operative local antibiotic irrigation. Bone Joint J 2016; 98-B(1 Suppl A): 23-6.

Acta Orthopaedica 2018; 89 (6): 689–695

689

Trauma simulation training: a randomized controlled trial evaluating the effectiveness of the Imperial Femoral Intramedullary Nailing Cognitive Task Analysis (IFINCTA) tool Rahul BHATTACHARYYA 1, Kapil SUGAND 1, Bilal AL-OBAIDI 2, Ian SINHA 2, Rajarshi BHATTACHARYA 2, and Chinmay M GUPTE 1 1 MSK

Lab, Imperial College London, Charing Cross Hospital Campus, Laboratory Block, London; 2 St Mary’s Hospital, Imperial College Hospitals NHS Trust, London, UK Correspondence: rahulbhattacharyya09@gmail.com Submitted 2018-03-14. Accepted 2018-07-12.

Background and purpose — Cognitive task analysis (CTA) has been used extensively to train pilots and in other surgical specialties. However, the use of CTA within orthopedics is in its infancy. We evaluated the effectiveness of a novel CTA tool to improve understanding of the procedural steps in antegrade femoral intramedullary nailing. Material and methods — Design: A modified Delphi technique was used to generate a CTA from 3 expert orthopedic trauma surgeons for antegrade femoral intramedullary nailing. The written and audiovisual information was combined to describe the technical steps, decision points, and errors for each phase of this procedure. Validation: A randomized double-blind controlled trial was undertaken with 22 medical students (novices) randomized into 2 equal groups. The intervention group were given the CTA tool and the control group were given a standard operative technique manual. They were assessed using the validated “Touch Surgery™” application assessment tool on femoral intramedullary nailing. Results — The pre-test scores between the two groups were similar. However, the post-test scores were statistically significantly better in the intervention group compared with the control group. The improvement (post-test median scores) in the intervention group compared with the control group was 20% for patient positioning and preparation, 21% for femoral preparation, 10% for proximal locking, and 19% for distal locking respectively (p < 0.001 for all comparisons). Interpretation — This is the first multimedia CTA tool in femoral intramedullary nailing that is easily accessible, userfriendly, and has demonstrated significant benefits in training novices over the traditional use of operative technique manuals.

There are significant strains on surgical training due to working regulations, increased malpractice cases, and reduced training time in the operating theatre (Philibert et al. 2002, RAD 1382 Consensus Statement 2008). Innovative adjuncts to the traditional apprenticeship model are required to reduce the learning curve and improve utilization of theatre training time. Previous research has highlighted the benefits of virtual reality (VR) and cadaveric simulation in orthopedic training (Cannon et al. 2014, Rebolledo et al. 2015, Sugand et al. 2015a, Camp et al. 2016, Tomlinson et al. 2016). However, these simulators are expensive and are not readily accessible to trainees worldwide. Karam et al. (2013) showed that 25% of training programs in the USA do not have a dedicated simulation facility and 87% of training program directors reported a lack of sufficient funds as the main barrier. The burden of orthopedic trauma is greater than ever before with major trauma now affecting the elderly population in addition to the high-energy injuries in the younger age group (Kehoe et al. 2015). In this setting, newer, innovative, and easily accessible training adjuncts to the traditional apprenticeship model are necessary to help train competent orthopaedic trauma surgeons. It is well established that cognitive learning of motor skills is key to undertaking a successful practical procedure (Spencer 1978, Flin et al. 2007, Wingfield et al. 2015, Wallace et al. 2017). Wingfield et al. (2015) defined cognitive task analysis (CTA) as a process by which experts provide complex information to trainees in a logical and simplified manner. This makes it easier to visualize and digest procedural steps. It has been extensively used to train pilots (Wingfield et al. 2015), and in other surgical specialties (Sullivan et al. 2008, Luker et al. 2008, Arora et al. 2011, Smink et al. 2012). However, it is rela-

690

tively new to orthopedic training. Bhattacharyya et al. (2017) reported the effectiveness of the first CTA tool in orthopedic training on diagnostic knee arthroscopy. There are no studies in the literature that have investigated the efficacy of CTA to teach orthopedic trauma procedures, which forms a major part of the orthopedic training curriculum, especially in the early years of training (BOTA Collaborators and Rashid 2018). We evaluated the effectiveness of a novel CTA tool to improve understanding of the procedural steps in antegrade femoral intramedullary nailing. Null hypothesis There was no difference in objective test scores between novices learning the procedural steps of antegrade femoral intramedullary nailing from our CTA compared with the op-tech manual.

Material and methods We first designed a CTA tool and then evaluated its effectiveness to teach the procedural steps. Design of the Imperial Femoral Intramedullary Nailing Cognitive Task Analysis (IFINCTA) tool Selection of experts The inclusion criteria for experts were: consultant, fellowshiptrained trauma and orthopedic surgeons, who practice trauma as part of their routine practice in a designated major trauma centre in the UK. Experts who matched the above criteria and were prepared to devote the time required for the Delphi process were selected to design the CTA. They were all male; median age 43 years (41–45); median years in consultant practice: 8 (4–8); median years in formal teaching: 8 (4–8). 2/3 experts were part of the group who received the grant from the AO Foundation for developing simulation techniques in trauma. However, the content of this CTA tool had no influence on the results of the validation part of the study as the experts were not involved in objectively scoring the participants in the trial. Procedural steps and Delphi technique The modified Delphi technique was used to generate the final written information of the IFINCTA tool. This technique involved a combination of independent analysis and “face to face” review from the 3 experts over 3 rounds. Round 1: The experts were interviewed independently to generate a list of technical steps, cognitive decision points, and potential errors and solutions that are applicable to an antegrade femoral intramedullary nailing procedure. All the technical steps and decision points that were common to the 3 experts were identified in round 1. Round 2: The steps that were common to 2 experts or exclusively generated from 1 expert were listed separately. This list

Acta Orthopaedica 2018; 89 (6): 689–695

was reviewed by the experts independently and they either agreed or disagreed with these steps. The steps agreed upon by all experts in rounds 1 and 2 were listed at the end of round 2. Round 3: This involved a “face to face” meeting of the expert panel where they reviewed all the steps that were agreed upon in rounds 1 and 2. Only those steps that had 100% consensus from all 3 experts at the end of the 3 rounds were included in the final CTA. Audiovisual component 1 of the experts in our study recorded a video of an antegrade femoral intramedullary nailing procedure for a diaphyseal mid-shaft femoral fracture. This video was divided into separate clips reflecting each of the 7 phases. A voiceover was recorded later using a “free recall” technique with the same expert. In this technique, the same expert recorded an audio clip describing each phase of the procedure by watching the video clips retrospectively. This was superimposed on the video clips. The other 2 experts reviewed this video and a consensus was reached regarding the final content of the video so that it accurately reflected the written CTA master template. Creation of the IFINCTA tool The written information (Delphi method), video clips, and audio recordings were input into a “wizard” tool developed in our unit. This is a web-based tool used by experts to create CTAs. This led to the design of the IFINCTA tool (Figure 1). This tool has a split-screen format. The left side has all the written information on the technical steps, cognitive decision points, and potential errors and solutions. The right side includes the multimedia modalities (video clips, audio voice recordings). Experimental design to evaluate the effectiveness of the IFINCTA tool This was a double-blinded, randomized controlled trial. Participants Inclusion criteria: Undergraduate medical students having observed 5 or fewer antegrade femoral intramedullary nailing procedures, neither having performed this procedure nor being familiarized with similar CTA tools beforehand. Exclusion criteria: Graduates, having performed this procedure, having been exposed to “Touch SurgeryTM” in the past, and observing more than 5 antegrade femoral intramedullary nailing procedures. Recruitment: 30 undergraduate novices initially registered their interest in participating and completed a questionnaire assessing their experience in femoral intramedullary nailing. 8 novices could not attend due to other commitments coinciding with the timing of the study. The remaining 22 novices all matched the inclusion criteria and completed the study. All participants gave written consent to participate in the study. They were recruited between October 2016 and April 2017.

Acta Orthopaedica 2018; 89 (6): 689–695

Figure 1. Snapshot of the IFINCTA tool showing the written information of the CTA on the left and the multimedia video clips and audio recordings on the right for the “approach” phase of the procedure.

691

Trial “Touch surgery™” application assessment tool We used the “Touch Surgery™” application (Kinosis, London, UK) assessment tool for femoral intramedullary nailing validated by Sugand et al. (2015b). This is a cognitive, interactive VR application that utilizes computer animation to explain the operative steps of a surgical procedure. It is available on mobile and smart devices (free of cost). It has separate learning and assessment tools. In our study, we only used the validated assessment tool for 4 modules of this procedure (1) patient positioning and preparation, (2) femoral canal preparation, (3) proximal locking, (4) distal locking and closure). Points were scored for single best answer questions for each step of the procedure and for virtually completing manual steps (incision, drilling etc.) using fingerswipe interactions on the screen (Figure 2). The final score for each module is automatically calculated by the software and is scored as a percentage with no negative marking (Sugand et al. 2015b).

Data collection Intervention and control groups: The intervention group were given the IFINCTA tool developed in this study and the control group were given a standard operative technique manual for this procedure, which is representative of a common current Figure 2. “Touch Surgery™” assessment tool for antegrade femoral intramedullary nailing: (left panel) demonstrating single best answer questions for cognitive decision making; (right panel) showing the method used by learners prior to digital swipe interactions required to complete the procedure. performing this procedure in theatre. Objective testing: The participants were supervised during the assessments by an independent research supervisor and were Randomization blinded to whether they belonged to the intervention or the The participants were randomized into 2 equal groups (n = control group. All participants completed a pre-test for analy11) using the random number generator function on Microsoft sis of baseline scores to ensure that the groups were homogeExcel (Microsoft Corp, Redmond, WA, USA). This generated neous at baseline. They were then provided with their respecrandom numbers from 1 to 22. These numbers were stored tive learning tools for 1 week and instructed to use them for as in concealed envelopes prior to being assigned to either the long as necessary in order to feel prepared before returning for intervention or the control group by a person in the lab who the post-test. We analyzed the mean time (minutes) for which was not directly involved in conducting the trial. the participants in each group used their learning tools. At the end of the week the students completed the post-test using the

692

Acta Orthopaedica 2018; 89 (6): 689–695

Table 1. IFINCTA tool rating parameters and rating analysis No.

Parameter

1 2

The tool was easy to use The tool made the procedure easy to understand The written, visual and audio modalities being simultaneously present was useful It would be beneficial to use this tool prior to attending an operating theatre session on femoral intramedullary nailing You enjoyed using this tool

3 4

5 a

> 3 on Likert scale

Enrollment

Assessed for eligibility (n = 30) Excluded (n = 8): – not meeting inclusion criteria, 0 – declined to participate, 0 – other reasons, 8

Agreed a 10/11 11/11 11/11

11/11 9/11

Allocation Allocated to intervention (n = 11): – received allocated intervantion, 11 – did not receive allocated intervention, 0

Allocated to control (n = 11): – received allocated control, 11

Follow-up Lost to follow-up (n = 0) Discontinued intervention (n = 0)

Lost to follow-up (n = 0) Discontinued intervention (n = 0)

Analysis Analyzed (n = 11): – excluded from analysis, 0

Analyzed (n = 11): – excluded from analysis, 0

Figure 3. CONSORT diagram showing recruitment and follow-up of participants.

same assessment tool. All testing occurred in the college laboratory facilities. Primary outcome measures: • The difference in the post-test scores between the intervention and the control groups. • The difference between the pre-test and the post-test scores in the intervention and the control groups. Subjective testing (secondary outcome measure): The students in the intervention group (n = 11) rated the IFINCTA tool on 5 parameters (Table 1) using the Likert rating scale (1 = strongly disagree, 2 = disagree, 3 = neither disagree nor agree, 4 = agree, 5 = strongly agree). The 3 expert consultant orthopedic trauma surgeons who designed the IFINCTA tool agreed on these parameters. Statistics Sample size calculation: An a priori power calculation was conducted using mean percentage scores from a previous pilot study whereby the intervention group scored 80% vs. the control group which scored 55% (SD 19) in module 2 of the assessment tool with alpha = 5% and beta = 20% resulting in power = 80%. The sample size resulted in a minimum n = 9 per group. The difference of scores was also deemed significant according to the expert surgeons. Software: The data obtained from the validation study were input into the SPSS software version 23. The analysis using histograms, QQ plots, skewness, and kurtosis data showed that the data were nonparametric. Results: The median (interquartile range) scores for the 4 modules were calculated for both groups. The Mann–Whitney U test was used for independent data and the Wilcoxon signed rank test was used for paired data. A p-value of < 0.05 was considered to be statistically significant.

Ethics, registration, funding, and potential conflicts of interest Ethical approval was obtained from our institutional Medical Education Ethics Committee (MEEC—reference number: 1617-08). Funding was received from the AO Foundation, Switzerland, under the AO Strategy Fund Project “Multipurpose Virtual Surgical Simulator”. The AO Foundation currently collaborates with the Touch SurgeryTM group; however, there was no association between the design of the IFINCTA tool in this study and the Touch SurgeryTM femoral nailing module, which was developed earlier and independent of the AO Foundation. The authors declared no conflicts of interest.

Results Participant demographics, baseline characteristics, and flow through the study (Figure 3) All 22 medical students (median age 23 years, 17 males) who matched the inclusion criteria completed the study. The median age was 22.5 years. The median year of undergraduate studies was 4 (1–5) in the intervention group compared with 5 (1–5) in the control group. The median number of antegrade femoral intramedullary nailing procedures observed in the operating room prior to the study in the intervention group was 1 (0–5) compared with 1.5 (0–5) in the control group. The mean time (minutes) for which the participants used their learning tool in the intervention group was 55 (45–65) compared with 57 (35–70) minutes in the control group. ‘Touch Surgery™’ objective assessment Pre-test scores: There was no statistically significant differ-

Acta Orthopaedica 2018; 89 (6): 689–695

693

Table 2. Comparison of pre-test “Touch Surgery™” assessment scores between intervention and control groups. Values are median scores (interquartile range)

Module 1. Patient positioning and preparation 2. Femoral canal preparation 3. Proximal locking 4. Distal locking and closure a

Intervention Control (IFINCTA) “op tech”

p-value a

50 (40–55)

45 (25–55)

0.4

53 (50–59) 57 (50–60)

56 (38–62) 50 (50–57)

1 0.4

57 (53–63)

58 (51–63)

0.7

“Touch SurgeryTM” assessment scores 100

Mann–Whitney U test for median difference in scores 20

Table 3. Comparison of post-test “Touch Surgery™” assessment scores between intervention and control groups. Values are median scores (interquartile range)

Module 1. Patient positioning and preparation 2. Femoral canal preparation 3. Proximal locking 4. Distal locking and closure a

Intervention Control Difference (IFINCTA) “op tech” (%) p-value a 80 (75–82) 60 (45–72)

< 0.001

79 (67–84) 58 (48–67) 77 (70–82) 67 (52–72)

21 10

0.001 < 0.001

82 (81–86) 63 (53–78)

< 0.001

Mann–Whitney U test for median difference in scores

Table 4. Comparison between pre-test and post-test scores for the intervention and the control groups. Values are difference between pre-test and post-test median scores

Module

Intervention Control (IFINCTA) p-value a “op tech” p-value a

1. Patient positioning and preparation 2. Femoral canal preparation 3. Proximal locking 4. Distal locking and closure a

< 0.001

26 20

< 0.001 < 0.001

2 17

0.4 0.04

< 0.001

0.04

Wilcoxon signed rank test

ence in the median pre-test scores between the intervention and the control groups to demonstrate homogeneity of participants, and avoiding selection bias (Table 2). Post-test scores (primary outcome measure 1): The post-test scores were statistically significantly better in the intervention group compared with the control group for all 4 modules in the “Touch Surgery™” assessment tool for antegrade femoral intramedullary nailing (Figure 4 and Table 3). Comparison between the pre- and post-test scores for both groups (primary outcome measure 2): Both groups showed

IFINCTA Control

Module 1: Module 2: Patient positioning Femoral canal and preparation preparation

Module 3: Proximal locking

Module 4: Distal locking and closure

Figure 4. Box and whisker plot comparing post-test “Touch SurgeryTM’ assessment scores between the IFINCTA and control groups for modules 1–4. Red lines are median values, boxes are interquartile ranges, whiskers are ranges of the data sets, and circles are 2 outliers.

improvement in the post-test compared with the pre-test but this was much greater in the intervention group compared with the control group (Table 4). IFINCTA tool rating analysis (secondary outcome measure) All participants agreed (> 3 on the Likert scale) on parameters 2–4 on the rating analysis. 10/11 participants agreed that the tool was easy to use and 9/11 enjoyed using the tool (Table 1).

Discussion This study adds to previous evidence supporting the use of CTA in orthopedic training (Bhattacharyya et al. 2017). The tool utilizes written and audiovisual information simultaneously to describe each step of antegrade femoral intramedullary nailing. The learner is equipped with knowledge on the technical steps, cognitive decision-making behind performing each technical step, and potential errors and solutions involved in this procedure. This study showed substantially significantly better “posttest” scores in the group using the IFINCTA tool compared with the control group. Although there were improvements in both groups between the pre-test and post-test, the difference was much greater in the IFINCTA group compared with the control group. Our analysis rejected the null hypothesis and demonstrated that the IFINCTA is an effective tool to improve the cognitive understanding of an antegrade femoral intramedullary nailing procedure for novice learners. It is also superior

694

compared with a standard operative technique manual, which is the gold-standard resource for learning surgical procedures prior to operating on patients. Previous research has demonstrated benefits and a training effect of the “Touch Surgery™” application to teach femoral intramedullary nailing (Sugand et al. 2015a, 2015b). “Touch Surgery™” uses computerized animations and written information to teach procedural steps. There are no defined cognitive decision points and description of specific errors and solutions in this application. Furthermore, our tool has video clips (with superimposed audio voiceovers) recorded in an actual operating theatre demonstrating the entire procedure in realtime. The IFINCTA help the learners to truly appreciate the technical challenges in each procedural step involved and the cognitive reasoning behind performing each technical step, while also bringing to their attention potential errors and solutions using simultaneous written and audiovisual multimodal multimedia modalities. This makes our tool both educationally novel and technologically innovative. Studies have highlighted the effectiveness of VR and cadaveric simulation to teach orthopedic trauma procedures (Leong et al. 2008, Sugand et al. 2015a). However (Karam et al. 2013), these are expensive and not readily accessible to trainees. Our CTA tool is web based and easily accessible. It allows repeated sustained practice, which is key in simulation training. Using this tool prior to attending the operating theatre will provide learners a sound understanding of the technical steps and the decision-making involved in this procedure. This will reduce their initial steep phase of the learning curve and allow more effective utilization of operating theatre training time. This is key in the current environment of reduced training hours (Chikwe 2004) available to achieve competency and can pave the way for CTA-based learning of complex surgical procedures in the future. The subjective assessment of the tool indicated that the entire intervention group agreed that the tool aided their understanding of the procedure and was easy and enjoyable to use. The participants believed that this tool is a useful adjunct to learning in the operating room and they would prefer to use our tool prior to operating on patients. Strengths and limitations The strengths of this study are: it is a prospective, doubleblind randomized controlled trial with a comparable intervention and control group. We used a sound modified Delphi technique to design the IFINCTA tool. This tool has demonstrated significant benefits to improve the cognitive understanding of the procedure for novice students. Our goal is for learners to use this tool prior to attending the operating theatre to give them a robust foundation in the understanding of this procedure, which will complement their learning in theatre. The limitation of this study is that we have not yet analyzed the impact of this tool to improve operating theatre performance (i.e., transfer validity), which carries its

Acta Orthopaedica 2018; 89 (6): 689–695

own ethical implications. We were obliged to demonstrate the ability of our CTA in a simulation setting to improve cognitive understanding prior to using it in the operating theatre, which will form part of our future work. Another potential limitation was that gender differences amongst the participants or the experts was not accounted for in this study. Application With an ever-increasing rise of orthopedic trauma workload (Kehoe et al. 2015) and reduced training hours (Chikwe 2004), simulation-training techniques have increasingly been encouraged (Robbins et al. 2010). We have designed a user-friendly, accessible CTA learning tool that has demonstrated significant benefits in improving the knowledge of novice learners in antegrade femoral intramedullary nailing. We believe that similar CTA learning tools can be used in other orthopedic sub-specialties, thereby increasing its impact within orthopedic training. Conclusion The IFINCTA tool has shown significant benefits in the cognitive understanding of novices in antegrade femoral intramedullary nailing. We recommend using our user-friendly, online tool to provide a strong foundation and a shorter cognitive learning curve for novices prior to starting their apprenticeship training in the operating theatre.

Study and tool design: RB, KS, RB, and CMG. Data collection and analysis: RB, KS, and BA. CTA tool Delphi: IS, RB. and CMG. Writing and editing the manuscript: RB, KS, RB, and CMG. Thanks are offered to Mr Victor Lesk for his help in designing the wizard tool. Acta thanks Li Felländer-Tsai and Christian Krettek for help with peer review of this study.

Arora S, Aggarwal R, Sirimanna P, Moran A, Grantcharov T, Kneebone R, Sevdalis N, Darzi A. Mental practice enhances surgical technical skills: a randomized controlled study. Ann Surg 2011; 253(2): 265-70. Bhattacharyya R, Davidson D J, Sugand K, Bartlett M J, Bhattacharya R, Gupte C M. Knee arthroscopy simulation: a randomized controlled trial evaluating the effectiveness of the Imperial Knee Arthroscopy Cognitive Task Analysis (IKACTA) Tool. J Bone Joint Surg Am 2017; 99(19): e103. BOTA Collaborators, Rashid M S. An audit of clinical training exposure amongst junior doctors working in trauma & orthopaedic surgery in 101 hospitals in the United Kingdom. BMC Med Educ 2018; 18(1): 1. Camp C L, Krych A J, Stuart M J, Regnier T D, Mills K M, Turner N S. Improving resident performance in knee arthroscopy: a prospective value assessment of simulators and cadaveric skills laboratories. J Bone Joint Surg Am 2016; 98(3): 220-5. Cannon W D, Garrett W E, Hunter R E, Sweeney H J, Eckhoff D G, Nicandri G T, Hutchinson M R, Johnson D D, Bisson L J, Bedi A, Hill J A, Koh J L, Reinig K D. Improving residency training in arthroscopic knee surgery with use of a virtual-reality simulator: a randomized blinded study. J Bone Joint Surg Am 2014; 96(21): 1798-806.

Acta Orthopaedica 2018; 89 (6): 689â&#x20AC;&#x201C;695

Chikwe J. No time to train the surgeons. BMJ 2004; 328(7437): 418-19. Flin R, Youngson G, Yule S. How do surgeons make intraoperative decisions? Qual Saf Health Care 2007; 16(3): 235-9. Karam M D, Pedowitz R A, Natividad H, Murray J, Marsh J L. Current and future use of surgical skills training laboratories in orthopaedic resident education: a national survey. J Bone Joint Surg Am 2013; 95(1): e4. Kehoe A, Smith J E, Edwards A, Yates D, Lecky F. The changing face of major trauma in the UK. Emerg Med J 2015;32(12):911-15. Leong J J H, Leff D R, Das A, Aggarwal R, Reilly P, Atkinson H D E, Emery R J, Darzi A W. Validation of orthopaedic bench models for trauma surgery. J Bone Joint Surg Br 2008; 90(7): 958-65. Luker K R, Sullivan M E, Peyre S E, Sherman R, Grunwald T. The use of a cognitive task analysis-based multimedia program to teach surgical decision making in flexor tendon repair. Am J Surg 2008; 195(1): 11-15. Philibert I, Friedmann P, Williams W T, ACGME Work Group on Resident Duty Hours. Accreditation Council for Graduate Medical Education. New requirements for resident duty hours. JAMA 2002; 288(9): 1112-14. Rebolledo B J, Hammann-Scala J, Leali A, Ranawat A S. Arthroscopy skills development with a surgical simulator: a comparative study in orthopaedic surgery residents. Am J Sports Med 2015; 43(6): 1526-9. RAD1382 Consensus Statement. The impact of EWTD on delivery of surgical services: a consensus statement. Association of Surgeons of Great Britain and Ireland; 2008. ewtd_consensus_statement_rad1382_.pdf. Robbins L, Bostrom M, Craig E, Sculco T P. Proposals for change in orthopaedic education: recommendations from an orthopaedic residency directorsâ&#x20AC;&#x2122; peer forum. J Bone Joint Surg Am 2010; 92(1): 245-9.

695

Smink D S, Peyre S E, Soybel D I, Tavakkolizadeh A, Vernon A H, Anastakis D J. Utilization of a cognitive task analysis for laparoscopic appendectomy to identify differentiated intraoperative teaching objectives. Am J Surg 2012; 203(4): 540-5. Spencer F. Teaching and measuring surgical techniques: the technical evaluation of competence. Bull Am Coll Surg. 1978; 63(3): 9-12. Sugand K, Akhtar K, Khatri C, Cobb J, Gupte C. Training effect of a virtual reality haptics-enabled dynamic hip screw simulator. Acta Orthop 2015a; 86(6): 695-701. Sugand K, Mawkin M, Gupte C. Validating Touch SurgeryTM: A cognitive task simulation and rehearsal app for intramedullary femoral nailing. Injury 2015b; 46(11): 2212-16. Sullivan M E, Ortega A, Wasserberg N, Kaufman H, Nyquist J, Clark R. Assessing the teaching of procedural skills: can cognitive task analysis add to our traditional teaching methods? Am J Surg 2008; 195(1): 20-3. Tomlinson J E, Yiasemidou M, Watts A L, Roberts D J H, Timothy J. Cadaveric spinal surgery simulation: a comparison of cadaver types. Glob Spine J 2016; 6(4): 357-61. Wallace L, Raison N, Ghumman F, Moran A, Dasgupta P, Ahmed K. Cognitive training: how can it be adapted for surgical education? Surgeon. 2017; 15(4): 231-9. Wingfield L R, Kulendran M, Chow A, Nehme J, Purkayastha S. Cognitive task analysis: bringing olympic athlete style training to surgical education. Surg Innov 2015; 22(4): 406-17.

696

Acta Orthopaedica 2018; 89 (6): 696–698

Distal femoral shortening osteotomy for treatment of sciatic nerve palsy after total hip arthroplasty — a report of 3 cases Benjamin PULIERO 1, William G BLAKENEY 1, Yann BEAULIEU 1, Alain ROY 1, and Pascal-André VENDITTOLI 1,2 1 Department

of Surgery, CIUSSS-de-L’Est-de-L’Ile-de-Montréal, Hôpital Maisonneuve Rosemont. Montréal, Québec; 2 Department of Surgery, Université de Montréal, Québec, Canada Correspondence: pa.vendittoli@me.com Submitted 2018-06-14. Accepted 2018-08-10.

Total hip arthroplasty (THA) for developmental dysplasia of the hip (DDH) with associated femoral head dislocation is a challenging procedure, with a high complication rate. These include dislocation, nonunion or malunion of osteotomies, infection, implant loosening, and sciatic nerve injury (Rogers et al. 2012, Sonohata et al. 2016). In severe DDH cases, overlengthening of the lower limb may cause sciatic nerve injury, a devastating complication (De Fine et al. 2017). Secondary neurologic pain and associated muscle weakness may overshadow otherwise excellent arthroplasty reconstruction and lead to patient dissatisfaction. In order to avoid this complication, it has been suggested that a proximal or subtrochanteric femoral osteotomy be performed when lengthening exceeds 4 cm (Cameron et al. 1998). Other authors have suggested intraoperative monitoring to assess nerve function as a decision tool to determine the maximal or appropriate limb lengthening (Paavilainen 1997). When a sciatic nerve injury is identified postoperatively, there is debate over what is the optimal treatment. In 3 THA cases for DDH (2 patients), we describe an effective surgical technique for treating associated sciatic nerve injuries secondary to excessive nerve tension. A late shortening distal femoral osteotomy was performed as treatment for 2 hips and as a prophylactic procedure for the third case. Surgical technique The patient is placed in the supine position. The distal femur is exposed using a lateral approach. A Kirschner wire is inserted under fluoroscopic guidance and is used as a guide to create a path for the saw blade. The bone is scored longitudinally to mark femoral rotation. The osteotomy is then performed and the desired length of bone removed. The 2 bone segments are approximated and a 4-hole 90° blade-plate applied with a compression guide. Postoperatively, the patient is allowed partial weight-bearing initially, increasing to full weight-bearing at 12 weeks. Patient 1 An 18-year-old woman with bilateral DDH was referred to our arthroplasty clinic because of bilateral groin pain and increasing difficulty with walking. Radiographs showed bilateral

DDH type III Crowe classification (Figure 1a). There was no leg length discrepancy due to bilateral pathology. On preoperative templating, by placing the acetabular component in the true acetabulum, planned limb lengthening was 4 cm. The THA was performed using a posterior approach. A 42 mm monobloc acetabular cup (Zimmer Maxera; Zimmer Biomet, Warsaw, IN, USA) was placed in the true acetabulum to restore the hip center of rotation. The cup used is a monobloc, uncemented implant with a non-modular ceramic liner (Biolox Delta; CeramTec, Plochingen, Germany). Matching ceramic femoral head size was 32 mm. With medialization, cup coverage was sufficient without the requirement for bone graft. The femoral stem was a straight conical uncemented stem (Zimmer Wagner cone prosthesis). During the procedure, the surgeon (AR) noted a tight sciatic nerve, but thought it was acceptable and no femoral shortening was performed. Neither electromyographic monitoring nor wake-up tests are commonly used in our practice. On the day following surgery, the patient had paresthesia in the sciatic nerve distribution but motor function was intact. The postoperative radiographs (Figure 1b) showed a 5.5 cm leg lengthening which was thought to be the cause of the sciatic nerve paresthesia. As there was no motor palsy, the initial diagnosis was a neurapraxia. Therefore, the patient was monitored to see if there was any recovery of nerve function. During the next 5 weeks, there was no improvement of the symptoms with a persistence of paresthesia when the knee was in full extension. A diagnosis of axonotmesis was made and the decision was taken to perform a femoral osteotomy to reduce the nerve tension and potentially favor nerve recovery. To avoid compromising the stability of the hip reconstruction surgery, a distal femoral osteotomy was seen as a better option. At 6 weeks after the initial surgery, the distal femoral osteotomy procedure was performed (3 cm shortening). Within the immediate days after the osteotomy, there was a significant reduction in pain and paresthesia. Complete recovery of sciatic nerve function was obtained within 6 weeks. The osteotomy was fully united by 3 -months. At 6-year follow-up, the patient remains asymptomatic with excellent hip function, no pain or dislocation, and normal motion.

Acta Orthopaedica 2018; 89 (6): 696–698

697

Case 1 with bilateral developmental dysplasia of the hip. a. Preoperatively b. Postoperatively, THA with 5.5 cm leg lengthening. c. Bilateral THAs. d. Healed bilateral distal femoral 3 cm shortening osteotomies.

2 years after the left THA, the patient requested surgery for her right hip. At the time of surgery on that side, a distal femoral osteotomy was performed in the same setting, immediately prior to the THA (3 cm shortening). The THA was uncomplicated, with importantly no sciatic nerve problems postoperatively (Figures 1c and d). Patient 2 c A 35-year-old woman suffering from unilateral left DDH was referred to the same surgeon (AR). A THA was performed without using a shortening osteotomy. In the recovery room during the postoperative examination, a sciatic nerve palsy was diagnosed with a complete motor palsy. An excessive 6 cm lower limb lengthening was diagnosed to be the cause of the injury. The surgeon decided to take the patient back to the operating room the following day. The previously described distal femoral osteotomy technique was performed (shortening of 3 cm). 2 days after the osteotomy, there was complete recovery of the motor and sensory function of the sciatic nerve. At 2 years’ follow-up, there was complete union of the osteotomy and no functional consequences of the transient nerve injury.

Discussion Sciatic nerve palsy is a relatively common complication after THA in patients with DDH. When it occurs, the best treatment remains controversial. To our knowledge, this article is the first description of a distal femoral osteotomy performed for the treatment of a sciatic nerve injury after THA. This strategy resulted in successful treatment of sciatic nerve injuries in 2 THAs and was as an effective prophylactic option in another. One of the causes for sciatic nerve injury is excessive leg lengthening during THA. Higuchi et al. (2015) reported that

exceeding 5 cm of lengthening during THA is a risk factor for sciatic nerve injury. In their case series of high dysplasia of the hip, Zappe et al. (2014) reported that only 17 of 34 patients fully recovered from their nerve lesions. They noted that recovery of function was still possible after 2 years, but only in 3 out of 17 of their patients. Neurapraxia lesions are known to recover within 3–4 weeks. When there is no recovery after 1 month, axonotmesis or neurotmesis is present (Campbell 2008). When an acute causative element such as overlengthening is present, we do not recommend waiting for a formal diagnosis of axonotmesis. Intervention should be performed as soon as possible. Interestingly, our first case demonstrates that full recovery is still possible even with a late intervention (6 weeks). Our observations have some limitations. The few cases described do not permit us to recommend this technique for every case of DDH requiring leg lengthening. In our practice, leg lengthening when performing a THA should be a maximum of 3–4 cm, but sometimes errors occur during surgery that may result in the requirement for the described distal femoral shortening osteotomy. This solution should be kept as a preventive or salvage procedure, since we do not have sufficient data to support more widespread use presently. One of the surgical options for sciatic nerve palsy after THA is surgical neurolysis. Regev et al. (2015) showed that when no improvement was found after 6 months of observation, surgical neurolysis improved function and pain outcomes.

698

Another study by Chughtai et al. (2017) compared patients who underwent nerve decompression treatment with those who did not. Their results showed a significant improvement in motor strength and pain when a surgical decompression was performed. Their literature review demonstrated the benefit of surgical treatment (improvement in 75%) compared with nonoperative treatment (improvement in 33%). In our cases the etiology was most certainly an excessive leg lengthening, and neurolysis without addressing the lengthening was deemed inappropriate. In THA in high dysplasia of the hip, several techniques for proximal femoral osteotomies are described to reduce leg lengthening and tension on the sciatic nerve (Reikeras et al. 2010). Numerous complications have been reported with proximal or subtrochanteric osteotomies such as nonunion, malunion, fractures (Paavilainen et al. 1993), and sciatic nerve palsy (Sener et al. 2002, Rollo et al. 2017). The advantage of performing a distal femoral osteotomy is that the stability of the hip is not compromised. Distal femoral shortening osteotomy concomitant with THA has been described by Koulouvaris et al. (2008) and provided satisfactory results regarding postoperative complications, with only 1 case of malunion and no cases of nonunion reported in 24 hips. Clinical outcomes were also excellent and comparable to proximal osteotomies. In their series, however, the mean shortening of the femur was only 18 mm (10–25). Their reported technique also differs from 2 of our cases, in that they performed their osteotomy at the time of THA, not as treatment for a sciatic nerve injury postoperatively. Distal femoral shortening osteotomy also allows for correction of frontal knee deformity, which is often associated with DDH, by changing the cylinder shape of the removed piece of bone into a wedge. Cameron et al. (2015) reported that distal femoral opening-wedge osteotomy also improved knee pain and functional scores. Only 1 nonunion occurred in their series of 31 knees and no nerve injuries were reported. Distal femoral osteotomy as described in our surgical technique also avoids rotational malunion when performed. Our technique has several advantages compared with proximal femoral osteotomy. First, the surgery is simple and there is no need to modify the hip arthroplasty, which could compromise stability. Second, it avoids the risk of infection or periprosthetic fracture due to revision surgery. Similar techniques have been used in the absence of sciatic nerve palsy with few complications reported (Cameron et al. 2015). It also avoids the risk of metaphyseal nonunion seen with proximal femoral osteotomies. Lastly, it may be an opportunity to correct severe genu valgum, which is often present in the DDH patient, by performing a wedge osteotomy. When a sciatic nerve palsy is diagnosed after THA with an important leg lengthening, the use of distal femoral shortening osteotomy should be considered.

Acta Orthopaedica 2018; 89 (6): 696–698

BP was involved in data collection and drafting of the manuscript. WB and YB were reviewers of the draft manuscript. AR was the senior surgeon performing the surgeries. PAV was the main reviewer and helped with critical appraisal of the manuscript. Acta thanks Johan Kärrholm for help with peer review of this study.

Cameron H U, Eren O T, Solomon M. Nerve injury in the prosthetic management of the dysplastic hip. Orthopedics 1998; 21(9): 980-1. Cameron J I, McCauley J C, Kermanshahi A Y, Bugbee W D. Lateral Opening-wedge distal femoral osteotomy: pain relief, functionnal improvement, and survivorship at 5 years. Clin Orthop Relat Res 2015; 473(6): 2009-15. Campell WW. Evaluation and management of peripheral nerve injury. Clin Neurophysiol 2008; 119(9): 1951-65. Chughtai M, Khlopas A, Gwam C U, Elmallah R K, Thomas M, Nace J, Mont M A. Nerve decompression surgery after total hip arthroplasty: what are the outcomes? J Arthroplasty 2017; 32(4): 1335-9. De Fine M, Romagnoli M, Zaffagnini S, Pignatti G. Sciatic nerve palsy following tTotal hip replacement: are patients’ personal characteristics more important than limb lengthening? A systematic review. Biomed Res Int 2017; 2017: 8361071. Higuchi Y, Hasegawa Y, Ishiguro N. Leg lengthening of more than 5 cm is a risk factor for sciatic nerve injury after total hip arthroplasty for adult hip dislocation. Nagoya J Med Sci 2015; 77(3): 455-63. Koulouvaris P, Stafylas K, Sculco T, Xenakis T. Distal femoral shortening in total hip arthroplasty for complex primary hip reconstruction: a new surgical technique. J Arthroplasty 2008; 23(7): 992-8. Paavilainen T, Hoikka V, Paavolainen P. Cementless total hip arthroplasty for congenitally dislocated or dysplastic hips: technique for replacement with a straight femoral component. Clin Orthop Relat Res 1993; (297): 71-81. Paavilainen T. Total hip replacement for developmental dysplasia of the hip: how I do it. Acta Orthop Scand 1997; 68(1): 77-84. Regev G J, Drexler M, Sever R, Dwyer T, Khashan M, Lidar Z, Salame K, Rochkind S. Neurolysis for the treatment of sciatic nerve palsy associated with total hip arthroplasty. Bone Joint J 2015; 97-B(10): 1345-9. Reikerås O, Haaland J E, Lereim P. Femoral shortening in total hip arthroplasty for high developmental dysplasia of the hip. Clin Orthop Relat Res 2010; 468(7): 1949-55. Rogers B A, Garbedian S, Kuchinad R A, Backstein D, Safir O, Gross AE. Total hip arthroplasty for adult hip dysplasia. J Bone Joint Surg Am 2012; 94-A(19): 1809-21. Rollo G, Solarino G, Vicenti G, Picca G, Carrozzo M, Moretti B. Subtrochanteric femoral shortening osteotomy combined with cementless total hip replacement for Crowe type IV developmental dysplasia: a retrospective study. J Orthop Traumatol 2017; 18(4): 407-13. Sener N, Tözün I R, Aşik M. Femoral shortening and cementless arthroplasty in high congenital dislocation of the hip. J Arthroplasty 2002; 17(1): 41-8. Sonohata M, Kitajima M, Kawano S, Mawatari M. Nerve palsy after total hip arthroplasty without subtrochanteric femoral shortening osteotomy for a completely dislocated hip joint. Open Orthop J 2016; 10: 785-92. Zappe B, Glauser P M, Majewski M, Stöckli H R, Ochsner P E. Long-term prognosis of nerve palsy after total hip arthroplasty: results of two-year follow-ups and long-term results after a mean time of 8 years. Arch Orthop Trauma Surg 2014; 134(10): 1477-82.

Acta Orthopaedica 2018; 89 (6): 699–701

699

Case report

Blount’s disease successfully treated with intraepiphyseal osteotomy with elevation of the medial plateau of the tibia—a case report with 65 years’ follow-up Terje TERJESEN 1 and Darko ANTICEVIC 2 1 Orthopaedic

Department, Oslo University Hospital, Rikshospitalet, Oslo, Norway; 2 Department of Paediatric Orthopaedics, Zagreb Children’s Hospital, Zagreb, Croatia Correspondence: terje.terjesen@rikshospitalet.no Submitted 2018-06-11. Accepted 2018-07-02.

A girl born in 1938 with Blount’s disease and pronounced varus deformity of her right knee underwent osteotomies of the proximal tibia and fibula with overcorrection into valgus at the age of 8 years. Because of relapse, a similar operation with

overcorrection to 10° valgus was performed 4 years later. However, again she had relapse of her varus deformity (Figure 1). In 1951 (patient age 13 years) an intraepiphyseal osteotomy of the medial tibial condyle, with elevation of the medial tibial plateau, was performed. The surgical technique was precisely described by Støren (1969) (Figure 1). The medial tibial condyle was exposed through a medial longitudinal incision about 12 cm long. The joint was not opened, except for a minor slit for orientation. The chiseling was done parallel to the sloping epiphyseal line at safe proximal distance from the line. Care was also taken to avoid getting too close to the joint surface. Several thinbladed chisels, 3 cm wide, were used simultaneously. The chiseling proceeded to the midline, Figure 1 a. Preoperative Figure 1 b. Preoperative radiographs, without Figure 1 d. Radiograph 1 picture of the patient at the weight-bearing (left). During weight-bearing year after operative eleva- under radiographic control. age of 13 years, showing (right) varus is increased because the defec- tion of the medial tibial joint The medial epicondylar fragpronounced varus defor- tive medial condyle provides poor support for surface, showing consolida- ment was carefully lifted, the mity of her right leg. the femoral condyle. Reprinted from Støren tion and good correction of final elevation being carried out (1969). the varus deformity. with a wide chisel covering the whole fragment. With the knee Figure 1 c. Operative technique. The in full valgus, large bone transchiseling is done proximal to the plants from the iliac crest were epiphyseal line and proceeds to the midline (left). The fragment is slowly wedged under the fragment elevated, lifting the whole fragment in under maximal elevation. A one piece (middle). Solid bone grafts plaster cast was applied under from the iliac crest, large enough to force the fragment into maximal eleforced valgus and maintained vation, are wedged under the fragfor 12 weeks. Weight-bearing ment under maximal forced valgus of was avoided for 1 year. The the knee (right). © 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 10.1080/17453674.2018.1516179

700

Acta Orthopaedica 2018; 89 (6): 699–701

extension and 130° of flexion (5° and 140°, respectively, in her left knee). The radiographs showed moderate osteoarthritis (OA) of the knee. Long-leg standing radiographs showed satisfactory position with 2° varus in her right knee and 0° in the left (Figure 2). There was a moderate leg length discrepancy (LLD) with the right leg 1.5 cm shorter.

Discussion Figure 2 a. Pictures at 65 years of follow-up (patient age 78 years), showing slight varus deformity of the right leg.

Figure 2 b. At 65 years follow-up the patient had almost full knee flexion.

Figure 2 c. Radiographs at 65 years follow-up, showing moderate osteoarthritis of the right knee. Figure 2 d. Long-leg standing radiographs at 65 years follow-up, showing slight varus of the right knee and leg length discrepancy of 1.5 cm with the right leg shorter.

osteotomy healed without any complications and gave good correction of the deformity (Figure 1). The treatment was published as a case report in Acta Orthopaedica Scandinavica (Støren 1969), with a follow-up time of 18 years. At that time the patient had good function of the knee and no pain. We managed to locate the patient and in November 2016, with a follow-up of 65 years (patient age 78 years), she underwent a clinical and radiographic examination. She was satisfied with the outcome, had good gait function and no pain on her usual activities. She could do “everything” (skiing, bicycling, swimming), but felt tiredness in her leg and felt that the knee was a bit unstable when walking on rough ground. The mobility of the right knee was good (Figure 2), with –5° of

Blount’s disease (tibia vara) is a disorder of growth that affects the medial aspect of the proximal end of the tibia leading to a progressive varus deformity. Langenskiöld and Riska (1964) maintained that, without operative treatment, the varus deformity usually progresses; thus early osteotomy is imperative. They recommended curved osteotomy of the proximal tibial metaphysis and osteotomy of the proximal fibula at the age of 2–8 years; this procedure cured the condition permanently in most patients. Recurrence usually occurred if the osteotomies were performed after 8 years of age. This is in accordance with other studies that reported better results in younger patients (Hofmann et al. 1982, Ferriter and Shapiro 1987, Loder and Johnston 1987, Doyle et al. 1996, Chotigavanichaya et al. 2002). Treatment of pronounced deformity is controversial and difficult. An osteotomy to correct varus is inadequate to obtain and maintain satisfactory correction in cases of substantial depression of the medial tibial articular surface (Schoenecker et al. 1992). This is supported by our case, where recurrence occurred twice after this operation. When there is extreme sloping of the medial condyle, a transepiphyseal tibial osteotomy with elevation of the medial condyle and bone grafts in the open wedge was described by Langenskiöld and Riska (1964). Other studies have shown relatively good results after similar osteotomies with elevation of the medial tibial plateau (Sasaki et al. 1986, Gregosiewicz et al. 1989, Langenskiöld 1989, Schoenecker et al. 1992). In recent years, elevation osteotomy of the medial tibial condyle

Acta Orthopaedica 2018; 89 (6): 699–701

has been combined with gradual lengthening using external fixation with the Ilizarov frame or Taylor’s frame (Accadbled et al. 2003, Bar-On et al. 2008, Hefny and Shalaby 2010, Edwards et al. 2017). The present surgical technique (Støren 1969) differs from that reported by Langenskiöld and Riska (1964). First, the Støren technique, which is indicated only in cases with open medial tibial physis, is intraepiphyseal whereas Langenskiöld used a transepiphyseal technique. Second, solid bone grafts were taken from the iliac crest whereas Langenskiöld used bone grafts from the tibial diaphysis. Third, Støren did not combine the elevation osteotomy with additional procedures, whereas Langenskiöld performed lateral epiphysiodesis of the proximal tibia and fibula. An intraepiphyseal osteotomy similar to Støren’s technique was described by Siffert (1982), although he combined it with a high inverted “V” tibial osteotomy to correct “the existing varus”. The long-term outcome of our patient was good, with almost normal function and range of knee motion and moderate residual varus after follow-up of 65 years. The case reported by Siffert (1982) had a good outcome at 13 years’ follow-up. Intraepiphyseal elevation osteotomy should be reserved for older patients in whom tibial osteotomy has failed to prevent progressive deformity. Care should be taken to avoid disturbance of bone nutrition with resulting aseptic necrosis of the elevated fragment and to avoid a lesion of the adjacent tibial epiphyseal line with ensuing inhibition of growth (Støren 1969). A moderate LLD of 1.5 cm, with the affected leg shorter, was seen in our case. After unilateral transepiphyseal osteotomy, where premature closure of the physis occurs, LLD ranging from 1.5 cm to 6.8 cm was seen in 5 of 7 patients (Schoenecker et al. 1992). This indicates that the intraepiphyseal technique with preservation of the physis is preferable in cases with open physis. Severe varus deformity should be surgically corrected; if left untreated, OA predictably occurs early in life (Hofmann et al. 1982). It is, however, difficult to evaluate the association between deformity following Blount’s disease and OA, because follow-up in most studies is too short. The longest previous follow-up after elevation osteotomy of the tibial plateau seems to be in a patient aged 41 years (Langenskiöld 1989). The follow-up time of our patient was 65 years (patient age 78 years) and thus more than long enough for a proper evaluation of OA. Surprisingly, only moderate OA had developed during this long time. This shows that a good result at skeletal maturity in a patient with severe preoperative varus (Støren 1969) can remain good even with long follow-up if the deformity has been adequately corrected.

701

In summary, intraepiphyseal osteotomy with elevation of the medial tibial joint surface and bone grafts to the open wedge gave a very good outcome after a follow-up of 65 years. If accurate preoperative planning and surgical technique are carried out, the method can be recommended in older children with pronounced varus deformity.

The authors would also like to thank Øystein H. Horgmo, Oslo University Hospital, for help with the photos in Figure 2.

Accadbled F, Laville J-M, Harper L. One-step treatment for evolved Blount’s disease: four cases and review of the literature. J Pediatr Orthop 2003; 23: 747-52. Bar-On E, Weigl D M, Becker T, Katz K. Treatment of severe early onset Blount’s disease by an intra-articular and a metaphyseal osteotomy using the Taylor spatial frame. J Child Orthop 2008; 2: 457-61. Chotigavanichaya C, Salinas G, Green T, Moseley C F, Otsuka N. Recurrence of varus deformity after proximal tibial osteotomy in Blount disease: long-term follow-up. J Pediatr Orthop 2002; 22: 638-41. Doyle B S, Volk A G, Smith C F. Infantile Blount disease: long-term follow-up of surgically treated patients at skeletal maturity. J Pediatr Orthop 1996; 16: 469-76. Edwards T A, Hughes R, Monsell F. The challenges of a comprehensive surgical approach to Blount’s disease. J Child Orthop 2017; 11: 479-87. Ferriter P, Shapiro F. Infantile tibia vara: factors affecting outcome following proximal tibial osteotomy. J Pediatr Orthop 1987; 7: 1-7. Gregosiewicz A, Wosko I, Kandzierski G, Drabik Z. Double-elevating osteotomy of tibiae in the treatment of severe cases of Blount’s disease. J Pediatr Orthop 1989; 9: 178-81. Hefny H, Shalaby H. A safer technique for the double elevation osteotomy in severe infantile tibia vara. Strat Traum Limb Recon 2010; 5: 79-87. Hofmann A, Jones R E, Herring J. Blount’s disease after skeletal maturity. J Bone Joint Surg (Am) 1982; 64-A: 1004-9. Langenskiöld A, Riska E B. Tibia vara (osteochondrosis deformans tibiae): a survey of seventy-one cases. J Bone Joint Surg (Am) 1964; 46-A: 140520. Langenskiöld A. Tibia vara: a critical review. Clin Orthop Rel Res 1989; (246): 195-207. Loder R T, Johnston C E. Infantile tibia vara. J Pediatr Orthop 1987; 7: 63946. Sasaki T, Yagi T, Monji J, Yasuda K, Kanno Y. Transepiphyseal plate osteotomy for severe tibia vara in children: follow-up study of four cases. J Pediatr Orthop 1986; 6: 61-5. Schoenecker P L, Johnston R, Rich M M, Capelli A M. Elevation of the medial plateau of the tibia in the treatment of Blount disease. J Bone Joint Surg (Am) 1992; 74-A: 351-8. Siffert R S. Intraepiphyseal osteotomy for progressive tibia vara: case report and rationale of management. J Pediatr Orthop 1982; 2: 81-5. Støren H. Operative elevation of the medial tibial joint surface in Blount’s disease: one case observed for 18 years after operation. Acta Orthop Scand 1969; 40: 788-96.

Keep up-to-date with the latest research developments in Orthopaedics. Start subscribing to article alerts for Acta Orthopaedica 2017 IMPACT FACTOR:

3.080*

Log on to the Acta Orthopaedica homepage:

www.tandfonline.com/IORT Click on the New content alerts

button, at the top of the page (under the journal title)

Official publication by the Nordic Orthopaedic Federation *ÂŠ2017 Clarivate Analytics, 2017 release of the Journal Citation ReportsÂŽ

6/18 ACTA ORTHOPAEDICA

Vol. 89, No. 6, 2018 (pp. 595–701)

Volume 89, Number 6, December 2018 ISSN 1745-3674