Acta Orthopaedica Vol. 89, Number 2, April 2018 by Acta

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

EDITORIAL OFFICE

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

BOARD

Anders Rydholm Lund, Sweden

Peter Frandsen, Denmark Morten Schultz Larsen, Denmark Juhana Leppilahti, Finland Karl-Ivar Lorentzen, Norway Katre Maasalu, Estonia Rob Nelissen, The Netherlands Ragnar Jonsson, Iceland Magnus Karlsson, Sweden Giedrius Kvederas, Lithuania Anders Rydholm, Sweden

DEPUTY EDITOR

Peter A Frandsen Odense, Denmark CO-EDITORS

Per Aspenberg Linköping, Sweden Nils Hailer Uppsala, Sweden Ivan Hvid Oslo, Norway Urban Rydholm Lund, Sweden Bart A Swierstra Nijmegen, The Netherlands Eivind Witsø Trondheim, Norway Rolf Önnerfält Lund, Sweden WEB EDITOR

DEPUTIES

Magnus Tägil Lund, Sweden

Kristina Åkesson Malmö, Sweden Jan-Magnus Björkenheim Helsinki, Finland Jon Fevang Bergen, Norway Rein Raie Tallin, Estonia Otto Robertsson Reykjavik, Iceland Kjeld Søballe Århus, Denmark Jan A N Verhaar Rotterdam, The Netherlands

STATISTICAL EDITOR

Jonas Ranstam Lund, Sweden PRODUCTION MANAGER

Kaj Knutson Lund, Sweden

Vol. 89, No. 2, 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. 2, April 2018 Hip and knee Outpatient total hip and knee arthroplasty: Facts and challenges Different competing risks models for different questions may give similar results in arthroplasty registers in the presence of few events: Illustrated with 138,234 hip (124,560 patients) and 139,070 knee (125,213 patients) replacements from the Dutch Arthroplasty Register Cemented total hip replacement in patients under 55 years: Good results in 104 hips followed up for > 22 years Do dual-mobility cups cemented into porous tantalum shells reduce the risk of dislocation after revision surgery? A retrospective cohort study on 184 patients The effect of bearing type on the outcome of total hip arthroplasty: Analysis of 209,912 primary total hip arthroplasties registered in the Dutch Arthroplasty Register Temporal trends in hip fracture incidence, mortality, and morbidity in Denmark from 1999 to 2012 High and rising burden of hip and knee osteoarthritis in the Nordic region, 1990–2015: Findings from the Global Burden of Disease Study 2015 Fast-tracking for total knee replacement reduces use of institutional care without compromising quality: A register-based analysis of 4 hospitals and 4,256 replacements Migration and clinical outcome of mobile-bearing versus fixedbearing single-radius total knee arthroplasty: A randomized controlled trial High occurrence of osteoarthritic histopathological features unaccounted for by traditional scoring systems in lateral femoral condyles from total knee arthroplasty patients with varus alignment Graft fixation influences revision risk after ACL reconstruction with hamstring tendon autografts: A study of 38,666 patients from the Scandinavian knee ligament registries 2004–2011 Slipped capital femoral epiphysis Comparison between modified Dunn procedure and in situ fixation for severe stable slipped capital femoral epiphysis: A retrospective study of 29 hips followed for 2–7 years Good inter- and intraobserver reliability for assessment of the slip angle in 77 hip radiographs of children with a slipped capital femoral epiphysis Early osteoarthritis after slipped capital femoral epiphysis: Cartilage degeneration, residual deformity and patient-reported outcome in 25 patients Oncology Spinal metastasis with neurologic deficits: Outcome of late surgery in patients primarily deemed not suitable for surgery Risk of cancer after primary total hip replacement: The influence of bearings, cementation and the material of the stem: A retrospective cohort study of 8,343 patients with 9 years average follow-up from Valdoltra Orthopaedic Hospital, Slovenia Miscellaneous Patient injury claims involving fractures of the distal radius: 208 compensated claims from the Finnish Patient Insurance Center Avoidable 30-day mortality analysis and failure to rescue in dysvascular lower extremity amputees: Implications for future treatment protocols Correspondence Highlighting the results of a trial by using appropriate inferential statistics: Commentary on “Specific exercises for subacromial pain” by Björnsson Hallgren et al. 2017

141 145

S B W Vehmeijer, H Husted, and H Kehlet S van der Pas, R Nelissen, and M Fiocco

152

M Kiran, L R Johnston, S Sripada, G G McLeod, and A C Jariwala A Brüggemann, H Mallmin, and N P Hailer

156 163

R M Peters, L N van Steenbergen, M Stevens, P C Rijk, S K Bulstra, and W P Zijlstra

170

C Jantzen, C M Madsen, J B Lauritzen, and H L Jørgensen

177

A A Kiadaliri, L S Lohmander, M Moradi-Lakeh, I F Petersson, and M Englund

184

K J Pamilo, P Torkki, M Peltola, M Pesola, V Remes, and J Paloneva

190

K T van Hamersveld, P J Marang-van de Mheen, H J L van der Heide, H M J van der Linden-van der Zwaag, E R Valstar, and R G H H Nelissen V P Mantripragada, N S Piuzzi, T Zachos, N A Obuchowski, G F Muschler, and R J Midura

197 204

A Persson, T Gifstad, M Lind, L Engebretsen, K Fjeldsgaard, J O Drogset, M Forssblad, B Espehaug, A B Kjellsen, and J M Fevang

211

G Trisolino, S Stilli, G Gallone, P Santos Leite, and G Pignatti

217

B Herngren, M Lindell, and G Hägglund

222

L Helgesson, P K Johansson, Y Aurell, C-J Tiderius, J Kärrholm, and J Riad

229

P Tsagozis and H C F Bauer

234

V Levašič, I Milošev, and V Zadnik

240

H Sandelin, E Waris, E Hirvensalo, J Vasenius, H Huhtala, T Raatikainen, and T Helkamaa C Wied, N B Foss, P T Tengberg, G Holm, A Troelsen, and M T Kristensen

246

251

M Saltychev and P Virolainen versus H C Björnsson Hallgren, L E Adolfsson, K Johansson, B Öberg, A Peterson, and T M Holmgren

Erratum Risk of cancer after primary total hip replacement: The influence of bearings, cementation and the material of the stem: A retrospective cohort study of 8,343 patients with 9 years average follow-up from Valdoltra Orthopaedic Hospital, Slovenia (Acta Orthop 2018; 89 (2): 234–239) Information to authors

252

253

V Levašič, I Milošev, and V Zadnik

Acta Orthopaedica 2018; 89 (2): 141–144

141

Outpatient total hip and knee arthroplasty Facts and challenges Stephan B W VEHMEIJER 1, 2, Henrik HUSTED 3, 5, and Henrik KEHLET 4, 5

1 Department

of Orthopedic Surgery, Reinier de Graaf Groep, Delft, The Netherlands; 2 Department of Orthopedic Surgery, Orthoparc, Bosch en Duin, The Netherlands; 3 Department of Orthopedic Surgery, Copenhagen University Hospital, Hvidovre, Denmark; 4 Section of Surgical Pathophysiology, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark; 5 Lundbeck Foundation Centre for Fast-track Hip and Knee Arthroplasty, Copenhagen, Denmark Correspondence: s.vehmeijer@rdgg.nl Submitted 2017-08-10. Accepted 2017-10-03.

ABSTRACT— As a result of the introduction of fast-track programs, the length of hospital stay after arthroplasty has decreased to a point where some patients meet the discharge criteria on the day of surgery. In several studies, well-established fast-track centers have demonstrated the feasibility of outpatient procedures in selected patients. However, in literature the term “outpatient” is sometimes also used for patients who spend one or more nights in hospital. We therefore propose to use “outpatient” solely for patients who are discharged to their own home on the day of surgery and do not have an overnight stay at either the hospital or another non-home facility. Also, several challenges need to be overcome before this becomes an established procedure. The combination of preoperative high-dose steroids and multimodal opioid-sparing analgesia has enhanced patient recovery after arthroplasty, but efforts to control undesirable pathophysiological responses will be a prerequisite to improve the success rate of an outpatient setting. Also, care must be taken to avoid extra activities or investments solely to enable discharge on the day of surgery. Further cost analyses will have to be performed to establish the true financial benefit of outpatient treatment. ■

In the past decades, fast-track programs have successfully been introduced in orthopedics, mainly in total hip and knee arthroplasty (THA and TKA). Thus, a combination of organizational and medical improvements in the pain and anesthetic, mobilization, and surgical protocols has led to enhanced recovery of patients after arthroplasty lowering morbidity and mortality. As a side effect of these programs, length of hospital stay (LOS) decreased to a point where a majority of total joint

arthroplasty (TJA) patients reached the essentially unchanged discharge criteria after the first postoperative night (Klapwijk et al. 2017). Although some patients reach these criteria on the day of surgery, outpatient TJA remains a psychological barrier in many institutions; in addition, reimbursement issues and concerns over safety prevent surgeons from allowing patients to go home on the day of surgery (Thienpont et al. 2015). Until recently, the reports of outpatient TJA have mostly been anecdotal, single surgeon or single institution based on selected patient populations (Berger et al. 2005, Dorr et al. 2010, Aynardi et al. 2014, Hartog et al. 2015, Kort et al. 2015). However, 2 more recent papers report on a multicenter randomized trial (Goyal et al. 2017) and a 2-center study with unselected patients (Gromov et al. 2017), confirming the feasibility of outpatient TJA in unselected populations. However, many challenges need to be overcome before it can be defined as an established treatment option and with more widespread recommendations.

Definition of outpatient THA and TKA Different definitions of outpatient arthroplasty are used in publications. In some reports a length of stay of < 23 hours is defined as outpatient (Sher et al. 2017), whereas others define outpatient as hospital discharge on the day of surgery (Nelson et al. 2017). The National Surgical Quality Improvement Program (NSQIP) used in many studies appears not to be a consistent entity. In a study by Bovonratwet et al. (2017) using the NSQIP, of the 529 THA patients who were registered as outpatients, only 63 (12%) were actually discharged on the

© 2017 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) DOI 10.1080/17453674.2017.1410958

11995 Vehmeijer D.indd 141

21-02-2018 15:01:48

142

day of surgery and of 890 patients undergoing TKA who were registered as outpatients, only 95 (11%) were discharged on the day of surgery. Current regulations in the United States allow for these observed patients to stay more than 1 night in hospital under observation status, despite being coded as outpatients (Bovonratwet et al. 2017). Some studies that report on outpatient TJA with data from the NSQIP use the outpatient variable, which may include patients who have been admitted for 1 night or more (Lovecchio et al. 2016, Courtney et al. 2017), whilst others use the LOS variable, including only patients with an LOS of 0 nights (Sher et al. 2017, Nelson et al. 2017). Clarity and uniformity is essential in publications on outpatient TJA. We would therefore propose to reserve the term “outpatient” solely for patients who are discharged to their own home on the day of surgery and who do not have an overnight stay at either the hospital or another non-home facility.

Why in hospital: pathophysiological considerations on postoperative recovery When discussing the possibilities for outpatient TJA, the basic question is what the reasons are for “staying in the hospital.” There seem to be 3 types of reasons: early organ dysfunction, appearance of complications, and organizational factors (Husted et al. 2011). Among the early organ dysfunctions, pain, nausea and vomiting, fatigue, weakness, and dizziness are the main complaints. In contrast, a well-implemented fasttrack approach has been demonstrated to decrease the risk of thromboembolic complications, cardiopulmonary complications and mortality, delirium, etc. Furthermore, a fast-track approach will attenuate the consequences of the otherwise well-established comorbidities (diabetes, cardiopulmonary, etc.) as risk factors for hospitalization and complications (Jørgensen et al. 2016). Since a study of unselected patients planned for TJA (Gromov et al. 2017) has shown that about 15–20% can be managed in an outpatient setting, the main question to be answered is: What are the reasons for delayed recovery and discharge in the remaining 85%? Obviously, organizational issues may account for a proportion of patients in which discharge is delayed. However, as mentioned earlier, the early organ dysfunctions leading to a risk of delayed recovery and/ or complications, which are principally mediated by the surgical stress responses (neuroendocrine/inflammatory/immunological), may be the most important determinant for delayed recovery in TJA (Gaudilliere et al. 2014). Consequently, the possibilities for future enhancement of a successful and safe outpatient TJA setting will require interventions to attenuate these responses. For this purpose, preoperative high-dose steroids (Jørgensen et al. 2017) in combination with well-established multimodal opioid-sparing analgesia are most promising. Despite these improvements, the future challenges will

11995 Vehmeijer D.indd 142

Acta Orthopaedica 2018; 89 (2): 141–144

include more attention to specific high-risk patient groups such as those receiving preoperative opioids, pain catastrophizers, and patients receiving psychopharmacological treatment (Greene et al. 2016, Jørgensen et al. 2016). Also, the mechanism for the pronounced early loss of quadriceps strength in TKA (which may lead to weakness and risk of falling) needs to be clarified. Neither surgery without the use of a tourniquet nor high-dose steroids have so far solved this problem (Lindberg-Larsen et al. 2017). Finally, the well-documented patient complaints concerning nausea, dizziness, and risk of syncope and falling may be related to early orthostatic intolerance (Jans and Kehlet 2017). Further research is required to define the relative role of autonomic nervous system disturbances, opioid use, and the inflammatory response. The risk of delirium has essentially been eliminated by the fast-track setup (Petersen et al. 2017), but further challenges and need for improvement are related to the pronounced early sleep disturbances that also may have an influence on subsequent pain responses (Chouchou et al. 2014). Finally, the risk of severe complications must be differentiated between “medical” and direct “surgical” complications, since the fast-track approach primarily focuses on improvements in medical morbidity in contrast to an initial “surgical” complication (hip dislocation, bleeding, etc.), which may be related to surgical expertise (Kehlet and Jørgensen 2016). However, the overall risk of mortality with the fast-track approach is currently extremely low (Jørgensen and Kehlet 2017) and not considered to be a relevant safety issue within a potential outpatient setting. In summary, efforts to control undesirable pathophysiological responses to TJA will be a prerequisite to improve the success rate of an outpatient setting. The main challenge remains to demonstrate the safety and positive patient recovery aspects of an outpatient setting vs. staying until the next day or even a little longer in the identified specific high-risk patient groups already mentioned.

Economic benefits of outpatient TJA In an economically challenged environment like today’s hospital system, the financial burden of an increasing number of arthroplasties needs to be addressed. Thus, it has been estimated that the number of hip and knee arthroplasties will increase by 75% in the years to come in the US alone (Kurtz et al. 2014) and similar projections have been made for Sweden (Nemes et al. 2014). Fast track has, apart from being clinically superior to more conventional pathways and resulting in less morbidity and mortality, also been shown to be financially attractive by lowering LOS. Thus, very low costs of around US$ 2,500 for a 2-day stay have been calculated in 2 Danish fast-track departments using the Time Driven Activity Based Costing method (Andreasen et al. 2017). The economic benefit of outpatient TJA is the lower cost associated with the reduced length of

21-02-2018 15:01:48

Acta Orthopaedica 2018; 89 (2): 141–144

stay (Aynardi et al. 2014, Lovald et al. 2014). Also, outpatient TJA can be performed in selected healthier patients in Ambulatory Surgical Centers (ASC), which may benefit from reduced overhead costs compared with inpatient hospitals (Parcells et al. 2016, Klein et al. 2017). However, in inpatient hospitals, the reduced cost associated with outpatient treatment is more difficult to assess. After all, a bed that would normally be used for a TJA patient will not always be filled with another patient on the orthopedic ward on that same night. In contrast, operating on the outpatient ward/ASC in hospitals that do not have any patients or hospital staff in the evening or at night would lead to cost reduction. Attention must, however, also be given to the costs that are incurred outside the ASC or the hospital. If, as a consequence of discharge on the day of surgery, patients are transferred to skilled nursing facilities instead of their own home, if extra physical therapy is indicated or if extra home care or home visits by nurses or physical therapists are needed, the potential financial benefits for ASCs or hospitals are outweighed by the additional activities in the postoperative period. Also, if discharge on the day of surgery leads to more readmissions or complications the potential financial benefits are nullified or worse. Therefore, further cost analyses will have to be performed to establish the true financial benefit of outpatient treatment. Consequently, the primary goal of physicians who implement an outpatient program should be to ensure patient safety and avoid extra activities or investments that are employed solely to enable discharge on the day of surgery unless a cost–benefit analysis is performed beforehand. Altogether, although healthcare budgets are under pressure, cost reduction or profit optimization should not be the main driver of outpatient treatment. Conclusions The published studies on outpatient TJA from Europe have all been from institutions that have a well-established fast-track protocol (Hartog et al. 2015, Kort et al. 2015, Gromov et al. 2017, Hoorntje et al. 2017). As a result of their programs, these hospitals have seen their length of stay gradually decrease to a point where outpatient TJA seemed feasible. This requires, however, a serious investment in time and resources and without this effort other hospitals should not commence an outpatient program. For most hospitals, outpatient TJA surgery should not be a goal in itself, but should rather be the result of a successful, already implemented fast-track program based on the concept “first better – then faster.” Only then will it not lead to an increased rate of complications and readmissions. Consequently, several challenges lie ahead focusing on organizational aspects, improving interventions to reduce the risk of organ dysfunction, safety issues, and economic consequences. HK, HH and SV planned the paper. SV drafted the manuscript. All authors revised and approved the final manuscript.

11995 Vehmeijer D.indd 143

143

SBWV has a consultant contract with Zimmer Biomet (ZB). HH and HK are adivisory board members in Rapid Recovery ZB. ZB had no involvement in this study. The authors declare no conflicts of interest.

Acta thanks David Houlihan-Burne and Steen Sameer Khan for help with peer review of this study.

Andreasen S E, Holm H B, Jørgensen M, Gromov K, Kjærsgaard-Andersen P, Husted H. Time-driven activity-based cost of fast-track total hip and knee arthroplasty. J Arthroplasty 2017; 32 (6): 1747-55. Aynardi M, Post Z, Ong A, Orozco F, Sukin D C. Outpatient surgery as a means of cost reduction in total hip arthroplasty: A case-control study. HSS J 2014; 10: 252-5. Berger R A, Sanders S, Gerlinger T, Della Valle C, Jacobs J J, Rosenberg A G. Outpatient total knee arthroplasty with a minimally invasive technique. J Arthroplasty 2005; 20 (7, Suppl 3): 33-8. Bovonratwet P, Webb M L, Ondeck N T, Lukasiewicz A M, Cui J J, McLynn R P, Grauer J N. Definitional differences of “outpatient” versus “‘inpatient” THA and TKA can affect study outcomes. Clin Orthop Relat Res 2017; 475 (12): 2917-25. Chouchou F, Khoury S, Chauny J M, Denis R, Lavigne G J. Postoperative sleep disruptions: A potential catalyst of acute pain? Sleep Med Rev 2014; 18 (3): 273-82. Courtney P M, Boniello A J, Berger R A. Complications following outpatient total joint arthroplasty: An analysis of a national database. J Arthroplasty 2017; 32 (5): 1426-30. Dorr L D , Thomas D J, Zhu J, Dastane M, Chao L, Long W T. Outpatient total hip arthroplasty. J Arthroplasty 2010; 25: 501-6. Gaudilliere B, Fragiadakis G K, Bruggner R V, Nicolau M, Finck R, Tingle M, Silva J, Ganio E A, Yeh C G, Maloney W J, Huddleston J I, Goodman S B, Davis M M, Bendall S C, Fantl W J, Angst M S, Nolan G P. Clinical recovery from surgery correlates with single-cell immune signatures. Sci Transl Med 2014; 6: 255ra131. Goyal N, Chen A F, Padgett S E, Tan T L, Kheir M M, Hopper R H Jr, Hamilton W G, Hozack W J. Otto Aufranc Award: A multicenter, randomized study of outpatient versus inpatient total hip arthroplasty. Clin Orthop Relat Res 2017; 475 (2): 364-72. Greene M E, Rolfson O, Gordon M, Annebrink K, Malchau H, Garellick G. Is the use of antidepressants associated with patient-reported outcomes following total hip replacement surgery. Acta Orthop 2016; 87 (5): 444-51. Gromov K, Kjærsgaard-Andersen P, Revald P, Kehlet H, Husted H. Feasibility of outpatient total hip and knee arthroplasty in unselected patients. Acta Orthop 2017; 88(5): 516-521. Hartog Y M, Mathijssen N M, Vehmeijer S B. Total hip arthroplasty in an outpatient setting in 27 selected patients. Acta Orthop 2015; 86 (6): 667-70. Hoorntje A, Koenraadt K L, Boevé M G, van Geenen R C. Outpatient unicompartmental knee arthroplasty: Who is afraid of outpatient surgery? Knee Surg Sports Traumatol Arthrosc 2017; 25 (3): 759-66. Husted H, Lunn T H, Troelsen A, Gaarn-Larsen L, Kristensen B B, Kehlet H. Why still in hospital after fast-track hip and knee arthroplasty? Acta Orthop 2011; 82 (6): 679-84. Jans Ø, Kehlet H. Postoperative orthostatic intolerance: A common perioperative problem with few available solutions. Can J Anaesth 2017; 64 (1): 10-15. Jørgensen C C, Kehlet H; Lundbeck Foundation Centre for Fast-track Hip and Knee Replacement Collaborative group. Time course and reasons for 90-day mortality in fast-track hip and knee arthroplasty. Acta Anaesthesiol Scand 2017a; 61 (4): 436-44. Jørgensen C C, Petersen M A, Kehlet H; Lundbeck Foundation Centre for Fast-Track Hip and Knee Replacement Collaborative Group. Preoperative prediction of potentially preventable morbidity after fast-track hip and knee arthroplasty: A detailed descriptive cohort study. BMJ Open 2016; 6 (1): e009813.

21-02-2018 15:01:48

144

Jørgensen C C, Pitter F T, Kehlet H; Lundbeck Foundation Center for Fasttrack Hip and Knee Replacement Collaborative Group. Safety aspects of preoperative high-dose glucocorticoid in primary total knee replacement. Br J Anaesth 2017b; 119 (2): 267-75. Kehlet H, Jørgensen C C. Advancing surgical outcomes research and quality improvement within an enhanced recovery program framework. Ann Surg 2016; 264 (2): 237-8. Klapwijk L C, Mathijssen N M, Van Egmond JC, Verbeek B M, Vehmeijer S B. The first 6 weeks of recovery after primary total hip arthroplasty with fast track. Acta Orthop 2017; 88 (2): 140-4. Klein G R, Posner J M, Levine H B, Hartzband M A. Same day total hip arthroplasty performed at an ambulatory surgical center: 90-day complication rate on 549 patients. J Arthroplasty 2017; 32 (4): 1103-6. Kort N P, Bemelmans Y F, Schotanus M G. Outpatient surgery for unicompartmental knee arthroplasty is effective and safe. Knee Surg Sports Traumatol Arthrosc 2015; July 1. [Epub ahead of print] Kurtz S M, Ong K L, Lau E, Bozic K J. Impact of the economic downturn on total joint replacement demand in the United States: Updated projections to 2021. J Bone Joint Surg Am 2014; 96 (8): 624-30. Lindberg-Larsen V, Bandholm T Q, Zilmer C K, Bagger J, Hornsleth M, Kehlet H. Preoperative methylprednisolone does not reduce loss of kneeextension strength after total knee arthroplasty: A randomized, doubleblind, placebo-controlled trial of 61 patients. Acta Orthop 2017; 88(5): 543-9. Lovald S T, Ong K L, Malkani A L, Lau E C, Schmier J K, Kurtz S M, Manley M T. Complications, mortality, and costs for outpatient and short-stay total knee arthroplasty patients in comparison to standard-stay patients. J Arthroplasty 2014; 29: 510-5.

11995 Vehmeijer D.indd 144

Acta Orthopaedica 2018; 89 (2): 141–144

Lovecchio F, Alvi H, Sahota S, Beal M, Manning D. Is outpatient arthroplasty as safe as fast-track inpatient arthroplasty? A propensity score matched analysis. J Arthroplasty 2016; 31 (Suppl): 197-201. Nelson S J, Webb M L, Lukasiewicz A M, Varthi A G, Samuel A M, Grauer J N. Is outpatient total hip arthroplasty safe? J Arthroplasty 2017; 32 (5): 1439-42. Nemes S, Gordon M, Rogmark C, Rolfson O. Projections of total hip replacement in Sweden from 2013 to 2030. Acta Orthop 2014; 85 (3): 238-43. Parcells B W, Giacobbe D, Macknet D, Smith A, Schottenfeld M, Harwood D A, Kayiaros S. Total joint arthroplasty in a stand-alone ambulatory surgical center: Short-term outcomes. Orthopedics 2016; 39 (4): 223-8. Petersen P B, Jørgensen C C, Kehlet H; Lundbeck Foundation Centre for Fast-track Hip and Knee Replacement Collaborative Group. Delirium after fast-track hip and knee arthroplasty: A cohort study of 6331 elderly patients. Acta Anaesthesiol Scand 2017; 61 (7): 767-72. Sher A, Keswani A, Yao D H, Anderson M, Koenig K, Moucha C S. Predictors of same-day discharge in primary total joint arthroplasty patients and risk factors for post-discharge complications. J Arthroplasty 2017; 32 (9S): S150-6. Thienpont E, Lavand’homme P, Kehlet H. The constraints on day-case total knee arthroplasty: The fastest fast track. Bone Joint J 2015; 97-B (10, Suppl A): 40-4.

21-02-2018 15:01:48

Acta Orthopaedica 2018; 89 (2): 145–151

145

Different competing risks models for different questions may give similar results in arthroplasty registers in the presence of few events Illustrated with 138,234 hip (124,560 patients) and 139,070 knee (125,213 patients) replacements from the Dutch Arthroplasty Register Stéphanie VAN DER PAS 1,2, Rob NELISSEN 3, and Marta FIOCCO 1,2

1 Mathematical

Institute, Leiden University, Leiden, The Netherlands; 2 Department of Medical Statistics and Bioinformatics, Leiden University Medical Center, Leiden, The Netherlands; 3 Department of Orthopaedics, Leiden University Medical Center, Leiden, The Netherlands Correspondence: svdpas@math.leidenuniv.nl Submitted 2017-07-01. Accepted 2017-12-08

Background and purpose — In arthroplasty registry studies, the analysis of time to revision is complicated by the competing risk of death. There are no clear guidelines for the choice between the 2 main adjusted analysis methods, cause-specific Cox and Fine– Gray regression, for orthopedic data. We investigated whether there are benefits, such as insight into different aspects of progression to revision, to using either 1 or both regression methods in arthroplasty registry studies in general, and specifically when the length of follow-up is short relative to the expected survival of the implants. Patients and methods — Cause-specific Cox regression and Fine–Gray regression were performed on total hip (138,234 hips, 124,560 patients) and knee (139,070 knees, 125,213 patients) replacement data from the Dutch Arthroplasty Register (median follow-up 3.1 years, maximum 8 years), with sex, age, ASA score, diagnosis, and type of fixation as explanatory variables. The similarity of the resulting hazard ratios and confidence intervals was assessed visually and by computing the relative differences of the resulting subdistribution and cause-specific hazard ratios. Results — The outcomes of the cause-specific Cox and Fine– Gray regressions were numerically very close. The largest relative difference between the hazard ratios was 3.5%. Interpretation — The most likely explanation for the similarity is that there are relatively few events (revisions and deaths), due to the short follow-up compared with the expected failurefree survival of the hip and knee prostheses. Despite the similarity, we recommend always performing both cause-specific Cox and Fine–Gray regression. In this way, both etiology and prediction can be investigated. ■

Competing risks methodology is beginning to take its rightful place in the arsenal of statistical methods for arthroplasty registry data (Gillam et al. 2011, Lacny et al. 2015, Wongworawat et al. 2015). The generally advanced age of arthroplasty patients necessitates competing risks techniques, which naturally incorporate the probability that a patient may die before experiencing revision, or before another outcome of interest occurs. For unadjusted analyses, the Aalen–Johansen estimator is typically used, which is a more general version of the Kaplan–Meier, capable of incorporating competing events (Aalen and Johansen 1978, Putter et al. 2007). It provides an estimate of the cumulative incidence function, which is defined as the probability of failing from a specific cause before time t. For adjusted analyses, 2 methods are typically considered: cause-specific Cox regression and Fine–Gray regression (Holt 1978, Fine and Gray 1999, Putter et al. 2007). The choice between these 2 methods is the focus of this paper. The assumptions underlying both methods are not in general compatible. Current practice is to select 1 of the 2 methods, e.g., Puchner et al. 2015, Wang et al. 2009, contrary to the recommendations of Latouche et al. (2013), if competing risks are adjusted for at all. The implications for the interpretation of analyses of large arthroplasty data-sets of these methodologies are still lacking in the orthopedic literature (Porcher 2015). Competing risks Traditional methods for estimating the time to revision of a joint implant include the Kaplan–Meier estimator and Cox proportional hazard models. These methods treat patients who

© 2018 The Author(s). Published by Taylor & Francis on behalf of the Nordic Orthopedic Federation. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Non-Commercial License (https://creativecommons.org/licenses/by-nc/3.0) DOI 10.1080/17453674.2018.1427314

11869 van der Pas D.indd 145

2/21/2018 5:37:00 PM

146

Acta Orthopaedica 2018; 89 (2): 145–151

Overview of statistical outcomemeasure methods Cause-specific Cox regression The Cox proportional hazards (PH) model is a default choice for modelling the effect of covariates on the hazard rate when there is no competing event. Causespecific Cox regression is a natural extension of standard Cox PH modelling for the competing risks setting, where a PH model is applied to each cause-specific hazard. The cause-specific hazard is the instantaneous rate of failure due to one Figure 1. Unadjusted cumulative incidences of revision (bottom) and death (top) after THR of the causes. All cause-specific hazards, in patients 70 years or younger or older than 70. Estimated by the Aalen–Johansen (1978) in our case the cause-specific hazards of estimator. revision and death, are estimated separately, by censoring all individuals who failed due to a cause other than the one considered. Thus, when the cause-specific hazard of revision is estimated, all patients who die before undergoing revision are considered censored. In the Cox model, the instantaneous risk of revision is compared among patients who are event free and in follow-up (that is, who have not yet experienced revision or the competing event of death at a particular time point). This model is appropriate if the interest is in understanding etiological or biological questions (Putter et al. Figure 2. Unadjusted cumulative incidences of revision (bottom) and death (top) after TKR in patients 70 years or younger or older than 70. Estimated by the Aalen–Johansen (1978) 2007). estimator. An advantage of cause-specific Cox regression is that it gives detailed insights die before experiencing revision as censored observations, into the relationship between a risk factor and each separate implying that their implants could still be revised, even though outcome. In our orthopedic setting, with revision and death as they have died. Methods that do not account for the compet- outcomes, such insights are of the form: “Is the revision risk ing risk of death will overestimate the probability of revision for a patient group (e.g., older patients) only decreased because (Putter et al. 2007, Keurentjes et al. 2012, Lacny et al. 2015), these patients are more likely to die before being eligible for which may influence medical decision-making. The impact revision, or is there a separate age-related effect?” A drawback of ignoring the competing risk of death on the results of the is that the results from cause-specific Cox regression do not analyses depends on the incidence of the competing event. In directly answer the question as to whether the revision risk is the case of revision surgery, the incidence of death is typi- decreased at all for patients with a certain characteristic (e.g., cally very high, as the patient population is on average elderly older patients), at least, not without combining the analyses (average age is 69 years for THR and TKR in The Nether- for both the hazard of revision and the hazard of death. The lands). This is illustrated in Figures 1 and 2. The competing effect of a covariate on the cause-specific hazard of revision risk of death is especially strong for patients older than 70: the can be quite unpredictable when expressed in terms of the cumulative incidence of revision 8 years after THR is 3.4%, cause-specific cumulative incidence function. For example, while that of death is 18%. The numbers for TKR are 3.3% a covariate may be associated with an increased hazard of revision, but the probability of revision may be unaffected or and 19% respectively. For adjusted analyses, typically either cause-specific Cox even decrease. One of the ways in which this can happen is regression or Fine–Gray regression are performed. We briefly if the covariate is associated with an even larger increase in review the 2 methods. We refer to Gillam et al. (2011) for the hazard of death. The reason for this is that the cumulative a comprehensive review and comparison of competing risks incidence of any event (for example, revision) depends on the cause-specific hazards of all events. It follows that the way in methods for arthroplasty registry data.

11869 van der Pas D.indd 146

2/21/2018 5:37:00 PM

Acta Orthopaedica 2018; 89 (2): 145–151

which the cause-specific cumulative incidence is associated with covariates might be different from the way in which the cause-specific hazard is associated with covariates. For example, old age could be associated with an increased hazard of revision, but with an even larger hazard of death, such that the probability of revision for older patients may turn out to be lower than that of younger patients. Fine–Gray regression model Fine–Gray regression resolves the most important drawback to cause-specific Cox regression, as the coefficients resulting from Fine–Gray regression do have a direct relationship with the cumulative incidence (Fine and Gray 1999). Although the value itself is hard to interpret (Andersen et al. 2012), if a covariate has a positive coefficient in the Fine–Gray model, then the cumulative incidence will be increased. Fine–Gray regression achieves this by assuming a proportional hazards model for a different hazard, namely the subdistribution hazard. The subdistribution hazard is the instantaneous risk of failing from a cause given that the individual has not failed from that cause. The difference with the cause-specific hazard is that the risk set for the subdistribution hazard includes individuals who have failed from other causes already (such as death, which is “competing” with the risk of revision). The hazard of revision is compared based on the subset of patients who have not yet experienced revision at a particular time point. A patient who dies remains in the risk set, contrary to the risk set for cause-specific Cox regression, where such a patient would be censored. While Fine–Gray regression allows direct assessment of the relationship between a covariate and the cumulative incidence of the cause of interest, the insight into the effect of a covariate on a cause-specific hazard instead of a probability is lost. A model that regresses on the cumulative incidence function is a proper tool for prognosis and medical decision-making, since it deals with the actual risk of events occurring over time (Gail and Pfeiffer 2005, Ambrogi et al. 2008). In our orthopedic setting, questions answered by Fine–Gray regression are of the form: “Is a certain group of patients (e.g., older patients) more or less likely to experience revision than other (e.g., younger) patients?” and these can be answered by only estimating the subdistribution hazard of revision, without need for combination with the subdistribution hazard of death. Relationship between cause-specific Cox and Fine–Gray regression Both regression methods can be used to obtain an estimate of the cumulative incidence function, through different hazard ratios. There is a little-known relationship between the subdistribution hazard and cause-specific hazard to which we draw attention in this paper. Taking revision as the end-point of interest, the following equality holds (Beyersmann and Schumacher 2007, Beyersmann and Scheike 2013):

11869 van der Pas D.indd 147

147

subdistribution hazard of revision = (overall survival) / (probability of not dying) × cause-specific hazard of revision Here, the overall survival is the probability of neither experiencing revision nor dying. If the probability of experiencing revision is low, then both quantities in the ratio will be close to each other, and thus the ratio will be close to 1. This in turn implies that the cause-specific hazard and subdistribution are almost the same. A similar expression holds for the subdistribution hazards and cause-specific hazards for death; when there are few deaths, the subdistribution and causespecific hazards will be almost the same. There is another situation in which the cause-specific and subdistribution hazard ratios for a covariate are similar, namely when a covariate only affects one of the cause-specific hazards (Grambauer et al. 2010). As we shall see in the analysis of data from the Dutch Arthroplasty register, equality will turn out to be relevant for the analysis of data from orthopedic registries with a relatively short follow-up. Purpose of study The purpose of the present paper is to use real orthopedic data to discuss the advantages and disadvantages of each method for arthroplasty registry studies with revision as the endpoint, to characterize the questions that can be answered by each method, and to draw attention to the little-known fact that Fine–Gray and cause-specific Cox regression may yield numerically very similar results when there are few revisions or death, or when a covariate affects only one of the causespecific hazards.

Patients and methods Comparison of the Dutch Arthroplasty Register data This is a national cohort study, using data on THRs and TKRs from the Dutch Arthroplasty Register (LROI), established in 2007. Completeness for hip arthroplasties was over 97%, and for knee arthroplasties over 96%, in 2012 and 2013 (Van Steenbergen et al. 2015). Inclusion criteria for this analysis were: • THR/TKR performed between January 1, 2008 and December 31, 2015; • Complete covariate information available; • Known diagnosis (i.e., “other” was excluded). 138,234 hips in 124,560 patients, and 139,070 knees in 125,213 patients were included. On both data sets, causespecific Cox regression and Fine–Gray regression were performed, with either revision or death as outcomes. The PH assumptions were checked by inspecting the Aalen–Johansen estimates of the cumulative incidences (for Fine–Gray) and the Nelson–Aalen estimates of the cumulative hazards. Besides a visual assessment of the similarity of the outcomes, for each variable the relative difference of the hazard ratios

2/21/2018 5:37:02 PM

148

Acta Orthopaedica 2018; 89 (2): 145–151

Table 1. Patient characteristics for the total hip replacement patients from the Dutch Arthroplasty Register included in the analysis, that is, for the patients with complete covariate information and no “other” diagnosis. Values are frequency (%) unless otherwise specified Characteristic THRs Uncemented fixation Cemented fixation Mean age, years 70 years or younger Older than 70 years Female Male ASA 1 ASA 2 ASA 3/4 Osteoarthritis Osteonecrosis Post-Perthes/dysplasia Rheumatoid/inflammatory arthritis Late posttraumatic

Characteristic 138,234 (in 124,560 patients) 94,225 (68) 44,009 (32) 68.9 67,310 (49) 70,924 (51) 92,571 (67) 45,663 (33) 35,144 (25) 86,450 (63) 16,640 (12) 126,404 (91) 4,031 (3) 3,262 (2) 1,422 (1) 3,115 (2)

was computed as: (subdistribution hazard ratio – cause-specific hazard ratio)/ cause-specific hazard ratio × 100% The analyses are adjusted for sex, age, ASA classification, diagnosis, and type of fixation. Age is categorized to “70 or younger” and “older than 70”. ASA scores 3 and 4 were grouped together. Patients with hybrid fixations were excluded, leaving only patients with cemented or uncemented fixation. The THR patients had 1 of 5 diagnoses: osteoarthritis, dysplasia or post-Perthes, rheumatoid or inflammatory arthritis, osteonecrosis, or late posttraumatic. The TKR patients had 1 of 4 diagnoses: osteoarthritis, rheumatoid or inflammatory arthritis, osteonecrosis, or posttraumatic arthritis (Tables 1 and 2). Median follow-up for both THR and TKR was 3.1 years, maximum was 8 years. Analyses were performed using R version 3.3.2 (R Core Team 2016). Ethics, funding, and potential conflicts of interest No research ethics committee approval was sought for secondary analysis of registry data in line with the guidelines of the Central Committee on Research Involving Human Subjects. No funding was received for this study. No competing interests were declared.

Results Tables 3 and 4 (see Supplementary data) state the estimated coefficients and standard errors for each competing outcome obtained by cause-specific Cox and Fine–Gray regression, for THR and TKR respectively. No violation of the PH assumptions was detected.

11869 van der Pas D.indd 148

Table 2. Patient characteristics for the total knee replacement patients from the Dutch Arthroplasty Register included in the analysis, that is, for the patients with complete covariate information and no “other” diagnosis. Values are frequency (%) unless otherwise specified

TKRs Uncemented fixation Cemented fixation Mean age, years 70 years or younger Older than 70 years Female Male ASA 1 ASA 2 ASA 3/4 Osteoarthritis Osteonecrosis Rheumatoid/inflammatory arthritis Posttraumatic arthritis

139,070 (in 125,215 patients) 7,594 (5) 131,476 (95) 69.0 73,337 (53) 65,733 (47) 91,921 (66) 47,149 (34) 26,784 (19) 93,454 (67) 18,832 (14) 134,043 (96) 603 (< 1) 2,298 (2) 2,126 (2)

The cause-specific hazard ratios resulting from cause-specific Cox, and the subdistribution hazard ratios resulting from Fine–Gray by exponentiating the coefficients in Tables 3 and 4 (see Supplementary data) are visualized, together with the 95% confidence intervals, in Figures 3, 4, 5, and 6, to allow for visual assessment of the similarity of the outcomes. The maximum relative difference of the hazard ratios was 3.5%. Covariates have the same effect on the cumulative incidence (estimated by the Fine–Gray model) and on the hazard (estimated by the cause-specific Cox) for THR and TKR data. Given 2 THR patients with the same characteristics except for fixation, results in Table 3 and Figures 3 and 4 show that cemented fixation has a statistically significant protective effect on the cumulative incidence (and on the rate) of revision compared with uncemented. Age has a statistically significant protective effect on revision. The effect of ASA score and diagnosis is more prominent for the end-point death than for revision.

Discussion Similarity of results on hip and knee replacement data Clinically, the risk factors found in Tables 3 and 4 (see Supplementary data) are in line with previous research, with uncemented fixation, younger age, male sex, higher ASA score, and posttraumatic procedures associated with a higher THR revision risk, and younger age and posttraumatic procedures associated with a higher TKR revision risk (Prokopetz et al. 2012, Jasper et al. 2016). We focus here on the methodological aspects of our results. The outcomes of the cause-specific Cox regression and Fine–Gray regression are numerically very similar (Figures 3, 4, 5, and 6). As explained in the methods

2/21/2018 5:37:02 PM

Acta Orthopaedica 2018; 89 (2): 145–151

149

Revision – Total Hip Replacement

Figure 3. Cause-specific hazard ratios and subdistribution hazard ratios for total hip replacement with revision as endpoint (dots), with 95% confidence intervals (lines).

Fixation (uncemented) Cemented Age (70 or younger) Older than 70 years Gender (female) Male ASA score (ASA 1) ASA 2 ASA 3/4 Diagnosis Osteonecrosis Post-Perthes/dysplasia Rheumatoid/infl. arthritis Late posttraumatic

Cause−specific HR (Cox) Subdistribution HR (Fine−Gray)

0.5

1.5

Death – Total Hip Replacement

Figure 4. Cause-specific hazard ratios and subdistribution hazard ratios for total hip replacement with death as endpoint (dots), with 95% confidence intervals (lines).

Cause−specific HR (Cox) Subdistribution HR (Fine−Gray)

0.5

1.5

2.5

3.5

Revision – Total Knee Replacement

Figure 5. Cause-specific hazard ratios and subdistribution hazard ratios for total knee replacement with revision as end-point (dots), with 95% confidence intervals (lines).

Fixation (uncemented) Cemented Age (70 or younger) Older than 70 years Gender (female) Male ASA score (ASA 1) ASA 2 ASA 3/4 Diagnosis Osteonecrosis Rheumatoid/infl. arthritis Posttraumatic

Cause−specific HR (Cox) Subdistribution HR (Fine−Gray)

0.5

1.5

Death – Total Knee Replacement

Figure 6. Cause-specific hazard ratios and subdistribution hazard ratios for total hip replacement with death as endpoint (dots), with 95% confidence intervals (lines).

11869 van der Pas D.indd 149

Fixation (uncemented) Cemented Age (70 or younger) Older than 70 years Gender (female) Male ASA score (ASA 1) ASA 2 ASA 3/4 Diagnosis Osteonecrosis Rheumatoid/infl. arthritis Posttraumatic 0.5

Cause−specific HR (Cox) Subdistribution HR (Fine−Gray)

1.5

2.5

3.5

2/21/2018 5:37:02 PM

150

section, such numerical similarity is expected when censoring is heavy. In the Dutch Arthroplasty Register data, censoring is very heavy indeed: there are 3,251 revisions and 5,813 deaths among the 138,234 transplanted hips, and 4,169 revisions and 5,610 deaths among the 139,070 transplanted knees. The low frequency of events is explained by the short amount of follow-up relative to the average survival of a hip or knee implant. Implications for clinical interpretation The similarity of the outcomes of both regression methods indicates that the answers to etiological and predictive questions are the same for the early follow-up phase. For example, TKR patients older than 70 years have a lower probability of revision than patients younger than 70 years (as we can conclude from the Fine–Gray regression with revision as outcome) and this is not just because they are more likely to die before experiencing revision (as we can conclude from the cause-specific Cox hazard ratio for revision). The same reasoning holds for the other covariates and outcomes. Added value of reporting cause-specific Cox and Fine–Gray regression outcomes The recommendation of Grambauer et al. (2010) and Latouche et al. (2013) is to report the outcome of both cause-specific Cox and Fine–Gray regression side by side, for all causes. When the results of the 2 analyses are not numerically close, different insights can be learned from each analysis. Causespecific Cox allows for separate assessment of the relationship between each covariate and each hazard of interest (in this case, of revision and death) and may thus provide more insight into the mechanisms leading to failure. Fine–Gray regression yields in a sense a summary, indicating the association between a covariate and the cumulative incidence of revision. This direct relationship cannot be directly determined from cause-specific Cox coefficients, because the effect of a covariate on a hazard can be very different from the effect on the corresponding cumulative incidence. For example, if a coefficient resulting from a cause-specific Cox analysis is positive for revision but even larger for death, then the net effect on the cumulative incidence may actually be negative. Fine–Gray regression would indicate without further computations that the effect is negative, but does not reveal that this is because the high hazard of death prevents the occurrence of revision. The benefits of both methods can be taken advantage of by presenting the outcomes of both analyses. While the benefits cannot be demonstrated on the data from the Dutch Arthroplasty Register, researchers analyzing data from older registries with longer follow-up may obtain additional insights by performing both cause-specific Cox and Fine–Gray regression. The cause-specific hazards model is more appropriate when etiological questions are of interest since it quantifies the event rate among individuals at risk of experiencing the event of interest (revision in this context). Fine–Gray is a

11869 van der Pas D.indd 150

Acta Orthopaedica 2018; 89 (2): 145–151

regression model for the cumulative incidence function and it should be used when prediction is the focus. Limitations There are several limitations to this study. The data are observational, and thus no causal conclusions can be drawn from the analyses performed. The amount of follow-up is short relative to the average survival of each implant. Due to the scarcity of events, Fine–Gray and cause-specific Cox are numerically similar in all comparisons. While that is the point to which we would like to draw attention, we would like to emphasize that, as a rule, Fine–Gray and cause-specific Cox regression will yield different results. Finally, if the effect of 1 of the covariates is time-dependent, a more careful analysis is required. We refer to Gillam et al. (2011) for discussion on this point. Link to ignored bilaterality As an aside, we remark on a connection between the observed similarity between cause-specific Cox regression and Fine– Gray regression, and the issue of incorporating bilateral patients in orthopedic studies. The impact of ignored bilaterality is commonly considered negligible (Ranstam et al. 2011). We would like to point out that the circumstances under which ignoring the presence of bilateral patients does not substantially affect the analyses are the same under which the outcomes of Fine–Gray regression and cause-specific Cox regression are numerically close: when there are few events, ignoring bilaterality is unlikely to affect the results (Robertsson and Ranstam 2003). Details concerning analysis in the presence of bilateral patients are discussed in Lie et al. (2004) and Van der Pas et al. (2017). Recommendations for statistical analysis of arthroplasty registry data For researchers faced with the choice between cause-specific Cox regression and Fine–Gray regression, we concur with the recommendations of Grambauer et al. (2010) and Latouche et al. (2013) to report the outcome of both Fine–Gray and causespecific Cox regression, but add the recommendation to only do so when the results are not numerically similar. Numerical closeness of the Fine–Gray and cause-specific Cox regressions is expected in many arthroplasty registry studies, because the survival of hip and knee prostheses is generally high. If the results of the 2 analyses are indeed similar, then presenting 1 of them suffices, with a brief remark indicating that both analyses were performed. In case of longer follow-up, less similarity between the 2 regression methods is expected. Again, we emphasize that the interpretation of the results based on the 2 models is different and that the research question should guide the choice between the 2 models. We therefore caution that the recent statement made by Ranstam and Robertsson (2017), based on simulated data, that cause-specific Cox is the best method for estimating relative revision risk, should not be interpreted as a guideline that cause-specific Cox is always

2/21/2018 5:37:04 PM

Acta Orthopaedica 2018; 89 (2): 145–151

the best option. Cause-specific Cox regression is most suitable for etiological questions, while Fine–Gray regression is more appropriate for prediction. Supplementary data Tables 3 and 4 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/ 17453674.2018.1427314

SP and MF conceived and designed the study. SP and MF conducted the analysis, and all authors interpreted the results. SP drafted the first version of the manuscript. All authors helped in revising the manuscript and gave their final approval of the submitted version. All authors had full access to the data and take responsibility for the integrity of the data and the accuracy of the data analysis. Acta thanks Ove Furnes and other anonymous reviewers for help with peer review of this study.

Aalen O O, Johansen S. An empirical transition matrix for non-homogeneous Markov chains based on censored observations. Scand J Stat 1978; 5: 14150. Ambrogi F, Biganzoli E, Boracchi P. Estimates of clinically useful measures in competing risks survival analysis. Stat Med 2008; 27: 6407-25. Andersen P K, Geskus R B, de Witte T, Putter H. Competing risks in epidemiology: possibilities and pitfalls. Int J Epidemiol 2012; 41: 861-70. Beyersmann J, Scheike T H. Classical regression models for competing risks. In: Klein JP, van Houwelingen HC, Ibrahim JG, Scheike TH, editors. Handbook of survival analysis. Boca Raton, FL: CRC Press; 2013. pp. 157-177. Beyersmann J, Schumacher M. Letter to the Editor: Misspecified regression model for the subdistribution hazard of a competing risk. Stat Med 2007; 26: 1649–51. Fine J P, Gray R J. A proportional hazards model for the subdistribution of a competing risk. JASA 1999; 94: 496-509. Gail M H, Pfeiffer R M. On criteria for evaluating models of absolute risk. Biostatistics 2005; 6: 227-39. Gillam M H, Salter A, Ryan P, Graves S E. Different competing risks models applied to data from the Australian Orthopaedic Association National Joint Replacement Registry. Acta Orthop 2011; 82: 513-20. Grambauer N, Schumacher M, Beyersmann J. Proportional subdistribution hazards modeling offers a summary analysis, even if misspecified. Stat Med 2010; 29: 875-84. Holt J. Competing risks analyses with special reference to matched pair experiments. Biometrika 1978; 65: 159-65. 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.

11869 van der Pas D.indd 151

151

Keurentjes J, Fiocco M, Schreurs B, Pijls B, Nouta K, Nelissen R. Revision surgery is overestimated in hip replacement. Bone Joint Res 2012; 26: 2389-430. Lacny S, Wilson T, Clement F, Roberts D J, Faris P D, Ghali A, Marshall D A. Kaplan–Meier survival analysis overestimates the risk of revision arthroplasty: a meta-analysis. Clin Orthop Relat Res 2015; 473: 3431-42. Latouche A L, Allignol A, Beyersmann J, Labopin M, Fine J P. A competing risks analysis should report results on all cause-specific hazards and cumulative incidence functions. J Clin Epidemiol 2013; 66: 648-53. Lie S A, Engesæter L B, Havelin L I , Gjessing H K, Vollset S E. Dependency issues in survival analyses of 55 782 primary hip replacements from 47 355 patients. Stat Med 2004; 23: 3227-40. Porcher R. CORR Insights®: Kaplan–Meier survival analysis overestimates the risk of revision arthroplasty: a meta-analysis. Clin Orthop Relat Res 2015; 473: 3443-5. Prokopetz J J Z, Losina E, Bliss R L, Wright J, Baron J A, Katz J N. Risk factors for revision of primary total hip arthroplasty: a systematic review. BMC Musculoskelet Disord 2012; 13: 251. Puchner S E, Kutscha-Lissberg P, Kaider A, Panotopoulos J, Puchner R, Böhler C, Hobusch G, Windhager R, Funovics P T. Outcome after reconstruction of the proximal tibia: complications and competing risk analysis. PLoS One 2015; 10: e0135736. Putter H, Fiocco M, Geskus RB . Tutorial in biostatistics: competing risks and multi-state models. Stat Med 2007; 26: 2389-430. R Core Team. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2016. Available from: https://www.R-project.org/. Ranstam J, Robertsson O. The Cox model is better than the Fine and Gray model when estimating relative revision risk from arthroplasty register data. Acta Orthop 2017; 88 (6): 578-580. Ranstam J, Kärrholm J, Pulkkinen P, Keijo M, Espehaug B, Pedersen A B, Mehnert F, Furnes O. Statistical analysis of arthroplasty data. II. Guidelines. Acta Orthop 2011; 82: 258-67. Robertsson O, Ranstam J. No bias of ignored bilaterality when analysing the revision risk of knee prostheses: analysis of a population based sample of 44590 patients with 55298 knee prostheses from the national Swedish Knee Arthroplasty Register. BMC Musculoskelet Disord 2003; 4: 1. Van der Pas S L, Nelissen R G H H, Fiocco M. Patients with staged bilateral total joint arthroplasty in registries: immortal time bias and methodological options. J Bone Joint Surg Am 2017; 99: e82(1-8). Van Steenbergen L N, Denissen G A W, Spooren A, van Rooden S M, van Oosterhout F J, Morrenhof J W, Nelissen R G H H. More than 95% completeness of reported procedures in the population-based Dutch Arthroplasty Register: external validation of 311,890 procedures. Acta Orthop 2015; 86: 498-505. Wang Y, Simpson J A, Wluka A E, Teichtahl A J, English D R, Giles G G, Graves S, Cicuttini F M. Relationship between body adiposity measures and risk of primary knee and hip replacement for osteoarthritis: a prospective cohort study. Arthritis Res Ther 2009; 11: R31. 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 Relat Res 2015; 473: 1173-6.

2/21/2018 5:37:04 PM

152

Acta Orthopaedica 2018; 89 (2): 152–155

Cemented total hip replacement in patients under 55 years Good results in 104 hips followed up for  22 years Manish KIRAN, Linda R JOHNSTON, Sankar SRIPADA, Gordon G MCLEOD, and Arpit C JARIWALA

University of Dundee, TORT Centre, Ninewells Hospital, Dundee, Scotland, UK. Correspondence: drmanishkiran@gmail.com Submitted 2017-11-10. Accepted 2017-12-05 .

Background and purpose — About 86,000 total hip replacements (THR) have been registered in patients under 55 years in the National Joint Registry of England and Wales (NJR). The use of uncemented implants has increased, despite their outcomes not having been proven to be significantly better than cemented implants in this registry. We determined the implant survivorship and functional outcomes of cemented THR in patients under 55 years at a minimum follow-up of 22 years. Patients and methods — 104 hips in 100 patients were included in this prospective study. Functional outcome was assessed using the Harris Hip Score and radiographs were assessed for implant failure and “at risk” of failure. Kaplan–Meier survivorship analysis was performed. Results — 89% of hips showed good to excellent results at final follow-up with a mean Harris Hip Score of 88 at a mean follow-up of 25 years. Revision was performed in 3/104 hips. 14 acetabular components and 4 femoral components were “at risk” of failure. The survivorship at minimum 22 years with revision for any reason as the end-point was 97% (95% CI 95–98). Interpretation — Cemented hip replacements perform well in young patients with good long-term functional and radiographic outcomes. ■

There has been a constant debate regarding the use of uncemented vs. cemented implants in total hip replacement (THR). Although meta-analyses have reported no substantial difference in patient satisfaction, functional, or radiological outcome, there has been an increasing use of uncemented implants, especially in patients under 55 years of age, according to the National Joint Registry (NJR) report 2015 (Abdulkarim et al. 2013, NJR report 2015). The NJR has recorded about 86,000 THRs in patients under 55 years from the time of its incep-

tion in 2003. The most common cemented implant used in our database was the Charnley stem/Ogee cup (Wrightington, UK). The long-term outcomes of Charnley total hip replacement (THR) in the elderly are well established (Neumann et al. 1994, Sochart and Porter 1997a, Sochart and Porter 1997b, Callaghan et al. 2000, Wroblewski et al. 2009). The primary aim of our study was to determine the implant survivorship and functional outcome of a cohort of cemented total hip replacements in patients under 55 years at a minimum followup of 22 years.

Patients and methods All patients under the age of 55 undergoing primary cemented THR between January 1990 and December 1995 were reviewed from our audit database. All uncemented and hybrid THRs and patient who were over 55 years were excluded. 71 Bicontact/Plasmacup uncemented THRs and 122 Exeter/ Trident hybrid THRs performed during the same period were excluded. Charnley/Ogee THR had been performed in 112 hips. 5 patients died and 3 patients were lost to follow-up leaving 104 hips in 100 patients for this analysis. The mean age of the patients was 48 (16–55) years. The mean follow-up was 25 (22–27) years. Demographic, clinical, and radiological follow-up data were recorded. All the surgeries were performed by 5 consultant orthopedic surgeons using the classic Charnley trochanteric approach in a supine position. Palacos bone cement (Zimmer Biomet, Warsaw, IN, USA) with gentamycin was used with modern techniques like retrograde cementation, and adequate femoral and acetabular preparation using lavage systems were used. The patients were reviewed postoperatively at 6 weeks, 3 months, 6 months, and 1 year by the surgeons. Subsequent

11992 Kiran D.indd 152

2/20/2018 11:03:03 AM

Acta Orthopaedica 2018; 89 (2): 152–155

Table 1. Demographic data Sex (M/F), n BMI, mean (SD) Normal, n Overweight, n Obese, n Very obese, n Bilateral THR

153

Table 2. Preoperative diagnosis 56/44 26.8 (4.8) 49 32 14 5 4

Diagnosis

Number of patients

Osteoarthritis Rheumatoid arthritis Avascular necrosis DDH Ankylosing spondylitis Fracture neck femur Total

60 9 6 21 2 2 100

Table 3. Harris Hip Score and pain profile Score

Pre-op 1 year 5 years 10 years 15 years Final

HHS, mean (SD) 47 (12) 92 (7) No pain, n 0 97 Mild, n 1 3 Moderate, n 50 0 Severe, n 45 0 Very severe, n 4 0

92 (8) 93 7 0 0 0

90 (11) 90 (10) 88 (9) 90 89 86 8 9 11 2 2 2 0 0 1 0 0 0

Table 4. Complications in 104 hips Complication

Corp, Armonk, NY, USA). Statistical differences in functional outcome were assessed by using Student’s t-test for independent samples. Paired t-test was used to analyze improvement post- surgery. The Kaplan–Meier method was used to determine survivorship of the implant with revision for any reason as the end point; p < 0.05 was used as a measure of statistical significance.

Number

Dislocation Peroperative femoral fracture Deep infection Superficial infection Aseptic loosening of acetabulum

3 2 1 1 1

reviews were at 3, 5, 10 years and every 4 to 5 years after that in the local arthroplasty audit clinic where the data were recorded by independent audit specialist practitioners. Harris Hip Score (HHS) was used as a measure of functional outcome and a score less than 70 was considered as a failure. The pain profile was recorded using a visual analog scale (VAS). Radiographs were taken at every follow-up. All the radiographs were analyzed for signs of implant failure, wear, loosening, heterotopic ossification, and fractures by 2 blinded observers, senior fellows with a specialist interest in hip surgery. Femoral loosening was defined by Harris classification as definite, probable, and possible (Harris et al. 1982). Progressive radiolucent lines in 2 or more Charnley DeLee zones were considered signs of loosening in the acetabulum (DeLee and Charnley 1976). Definite implant position change or migration and eccentric wear of the cup were also considered signs of implant failure. Brooker’s classification was used for heterotopic ossification (Brooker et al. 1973). Statistics The data were statistically analyzed using the Statistical Package for Social Sciences (SPSS) software version 20.0 (IBM

11992 Kiran D.indd 153

Kaplan–Meier survival curve.

Ethics, funding, and potential conflicts of interests The study received Caldicott Guardian ethical approval. There were no external sources of funding and none of the authors had any conflicts of interest.

Results The demographic data and preoperative diagnoses are given in Table 1 and 2. The mean HHS of the 104 hips at final follow-up was 88. On the basis of the HHS, the functional outcome was good to excellent results at final follow-up in 89% hips. BMI did not statistically significantly influence the final outcome (p = 0.09). There was a significant improvement in pain after surgery (p < 0.001), which was maintained at final follow-up (Table 3,). The most common complication was dislocation, occurring in 3 hips (Table 4). Revision surgery was performed in 3 hips. 1 hip was revised for aseptic loosening of the acetabular component at 15 years. The second hip underwent a 2-stage revision for deep hematogenous infection at 13 years. The third hip was revised for recurrent dislocation at 8 years. Radiographic analysis at final follow-up showed “at risk” signs of radiological loosening in 14 acetabular components and 4 femoral components. With revision surgery for any reason as the end-point, the THR survivorship at final follow-up using the Kaplan–Meier method was 97% (95% CI 95–98) (Figure). If we included the patients “at risk” of revision, the survivorship reduced to 82% (95% CI 79–85).

2/20/2018 11:03:03 AM

154

Discussion Over 700,000 total hip replacement operations have been reported to the National Joint Registry of England and Wales in the last 13 years, of which almost 86,000 have been performed in patients under 55 years (NJR 2015). The management of hip problems in these younger patients is challenging. Cemented, uncemented, hybrid THR, reverse hybrid, and hip resurfacing have been performed in these patients. The NJR report shows an increase in uncemented THR in this age group. This trend has also been seen in the Swedish, Norway, and New Zealand Joint Registries. In contrast, the use of hip resurfacing has reduced substantially since the recognition of problems with metal-on-metal bearings (NJR 2015). Cemented THR has been performed in the management of a wide range of pathologies in the young, namely osteoarthritis, rheumatoid arthritis, developmental dysplasia of the hip, avascular necrosis, ankylosing spondylosis, and juvenile idiopathic arthritis (Joshi et al. 1993, Sochart and Porter 1997b, Lehtimäki et al. 1997, Wroblewski et al. 2010). The mean HHS and number of patients with good to excellent results in our study is comparable with other series from specialist centers or teaching hospitals with long-term follow-up (Lehtimäki et al. 1997, Goodman et al. 2014). A majority of our patients were pain free with a well-functioning hip at final follow-up. In concurrence with other series, dislocation was our most common complication (Joshi et al. 1993). However, our rate of deep infection (0.9%) was less than other series with long-term follow-up, with rates of 1.2% to 8% (Wroblewski et al. 2009, Warth et al. 2014). The use of strict aseptic technique and antibiotic-loaded cement may have contributed to the low infection rate. Aseptic loosening has been reported to be one of the leading causes of failure and revision in this age group. Good cementing technique is essential. Early studies of cemented THR have reported low revision rates in patients followed up for less than 5 years (Halley and Charnley 1975, Bisla et al. 1976). Subsequently, higher rates of revision ranging from 12% to 37% at 15- to 20-year follow-up were reported (Joshi et al. 1993, Boeree and Bannister 1993, Caton and Prudhon 2011, Warth et al. 2014). Various authors have stressed on the importance of long-term follow-up as aseptic loosening is progressive (Eftekhar 1987, Wroblewski et al. 1992, Keener et al. 2003). Based on our number of revisions and number of cases “at risk,” we agree with these studies and suggest that regular long-term clinical and radiographic review should be standard practice. Although our revision rate of 3% was lower than the rates reported in other studies using cemented implants at 20 years, the number of hips radiographically “at risk” suggest that further follow-up might result in higher revision rates as seen in other series with longer follow-up. The Get it Right First Time (GIRFT) report by the British Orthopaedic Association (BOA) recognized the need to standardize the use of total hip replacement implants in the National Health Service with a move towards cemented implants (Briggs 2012). Additionally, transparency in the

11992 Kiran D.indd 154

Acta Orthopaedica 2018; 89 (2): 152–155

pricing of hip implants has been highlighted (Pennington et al. 2013, Briggs 2015). The GIRFT report found that the cost of uncemented implants was almost double that of the cemented implants. Similar results were found in the United States (Unnanuntana et al. 2009). The Swedish Joint Registry did not show any clear advantage of hybrid implants and the use of hip resurfacing has seen a steady decline in all registries since problems with metal-on-metal bearings have been recognized. The strength of our series is the prospective nature with complete long-term functional and radiological follow-up in a cohort of young patients using a single cemented hip implant. A relative weakness is the varied diagnoses included, which mirrors the variety of pathology in this age group that require THR. Additionally, the choice of implant was based on surgeon preference and training and the experience of the surgeon in using a particular implant may influence the final outcome. Our series provides evidence for the utility of established cemented hip implants, which may be used in also in young adults.

MK, LJ, SS, and AJ conceived and designed the protocol. LJ and SS collected the data. MK, LJ, and GM analyzed the data and performed statistical analysis. MK and AJ drafted the manuscript. LJ, SS, and GM proof-read the manuscript.

Acta thanks Maziar Mohaddes and other anonymous reviewers for help with peer review of this study.

Abdulkarim A, Ellanti P, Motterlini N, Fahey T, O’Byrne JM. Cemented versus uncemented fixation in total hip replacement: a systematic review and meta-analysis of randomized controlled trials. Orthop Rev (Pavia) 2013; 5(1): e8. Bisla R S, Ranawat C S, Inglis A E. Total hip replacement in patients with ankylosing spondylitis with involvement of the hip. J Bone Joint Surg Am 1976; 58(2): 233-8. Boeree N R, Bannister G C. Cemented total hip arthroplasty in patients younger than 50 years of age. Ten- to 18-year results. Clin Orthop Relat Res 1993; (287): 153-9. Briggs T. Getting it Right First Time report (2012). http://gettingitrightfirsttime.com (accessed December 15, 2016). Briggs T. https://www.boa.ac.uk/wp-content/ uploads/ 2015/ 06/RaisingTransparency-letter-orthopaedic-prostheses-hip-and-knee-prices.pdf (accessed December 15, 2016). Brooker A F, Bowerman J W, Robinson R A, Riley L H Jr. Ectopic ossification following total hip replacement: incidence and a method of classification. J Bone Joint Surg Am 1973; 55(8): 1629-32. Callaghan J J, Albright J C, Goetz D D, Oleiniczak J P, Johnston R C. Charnley total hip arthroplasty with cement. J Bone Joint Surg Am 2000; 82-A: 487. Caton J, Prudhon J L. Over 25 years survival after Charnley’s total hip arthroplasty. Int Orthop 2011; 35(2): 185-8. DeLee J G, Charnley J. Radiological demarcation of cemented sockets in total hip replacement. Clin Orthop Relat Res 1976; (121): 20. Eftekhar N S. Long-term results of cemented total hip arthroplasty. Clin Orthop Relat Res 1987; (225): 207-17.

2/20/2018 11:03:03 AM

Acta Orthopaedica 2018; 89 (2): 152–155

Goodman S B, Hwang K, Imrie S. High complication rate in revision total hip arthroplasty in juvenile idiopathic arthritis. Clin Orthop Relat Res 2014; 472(2): 637-44. Halley D K, Charnley J. Results of low friction arthroplasty in patients thirty years of age or younger. Clin Orthop Relat Res 1975; (112): 180-91. Harris W H, McCarthy J C Jr, O’Neill D A. Femoral component loosening using contemporary techniques of femoral cement fixation. J Bone Joint Surg Am 1982; 64 (7): 1063-7. Joshi A B, Porter M L, Trail I A, Hunt L P, Murphy J C M, Hardinge K. Longterm results of Charnley low-friction arthroplasty in young patients. J Bone Joint Surg Br 1993; 75(B): 616-23. Keener J D, Callaghan J J, Goetz D D, Pederson D R, Sullivan P M, Johnston R C. Twenty-five-year results after Charnley total hip arthroplasty in patients less than fifty years old: a concise follow-up of a previous report. J Bone Joint Surg Am 2003; 85-A (6): 1066-72. Lehtimäki M Y, Lehto M U, Kautiainen H, Savolainen H A, Hämäläinen M M. Survivorship of the Charnley total hip arthroplasty in juvenile chronic arthritis: a follow-up of 186 cases for 22 years. J Bone Joint Surg Br 1997; 79 (5): 792-5. Neumann L, Freund K, Sorenson K H. Long-term results of Charnley total hip replacement: review of 92 patients at 15 to 20 years. J Bone Joint Surg Br 1994; 76-B: 245-51. New Zealand Joint Registry report 2015. http://www.nzoa.org.nz/nz-jointregistry (accessed December 15, 2016). NJR 12th Annual report. 2015. http://www.njrcentre.org.uk (accessed December 15, 2016). Norwegian Arthroplasty Register report 2015. http://nrlweb.ihelse.net (accessed December 15, 2016).

11992 Kiran D.indd 155

155

Pennington M, Grieve R, Sekhon J S, Gregg P, Black N, van der Meulen J H. Cemented, cementless, and hybrid prostheses for total hip replacement: cost effectiveness analysis. BMJ 2013; 346: 1026. Sochart D H, Porter M L. The long term results of Charnley low-friction arthroplasty in young patients who have congenital dislocation, degenerative osteoarthrosis or rheumatoid arthritis. J Bone Joint Surg Am 1997a; 79 (11): 1599-1617. Sochart D H, Porter M. Long-term results of total hip replacement in young patients who had ankylosing spondylitis: eighteen- to thirty-year results with survivorship analysis. J Bone Joint Surg Am 1997b; 79 (8): 1181-9. Swedish Hip Arthroplasty Register report 2014. http://www.shpr.se (accessed December 15, 2016). Unnanuntana A, Dimitroulias A, Bolognesi M P, Hwang K L, Goodman S B, Marcus R E. Cementless femoral prostheses cost more to implant than cemented femoral prostheses. Clin Orthop Relat Res 2009; 467 (6): 154651. Warth L C, Callaghan J J, Liu S S, Klaassen A L, Goetz D D, Johnston R C. Thirty-five-year results after Charnley total hip arthroplasty in patients less than fifty years old: a concise follow-up of previous reports. J Bone Joint Surg Am 2014; 96 (21): 1814-9. Wroblewski B M, Siney P D. Charnley low-friction arthroplasty in the young patient. Clin Orthop Relat Res 1992; (285): 45-7. Wroblewski B M, Siney P D, Fleming P. Charnley low-frictional torque arthroplasty: follow-up for 30 to 40 years. J Bone Joint Surg Br 2009; 91-B( 4): 447. Wroblewski B M, Purbach B, Siney P D, Fleming P A. Charnley low-friction arthroplasty in teenage patients: the ultimate challenge. J Bone Joint Surg Br 2010; 92 (4): 486-8.

2/20/2018 11:03:03 AM

156

Acta Orthopaedica 2018; 89 (2): 156–162

Do dual-mobility cups cemented into porous tantalum shells reduce the risk of dislocation after revision surgery? A retrospective cohort study on 184 patients Anders BRÜGGEMANN, Hans MALLMIN, and Nils P HAILER

Department of Surgical Sciences/Section of Orthopedics, Uppsala University, Uppsala, Sweden Correspondence: anders.bruggemann@surgsci.uu.se Submitted 2017-08-25. Accepted 2017-11-11

Background and purpose — Dual-mobility cups (DMCs) reduce the risk of dislocation and porous tantalum (TM) shells show favorable osseointegration after acetabular revision surgery, yet the combination of these implants has not been studied. We hypothesized that (1) cementing a DMC into a TM shell decreases the risk of dislocation; (2) DMCs cemented into TM shells are not at greater risk of re-revision; (3) liberation of tantalum ions is marginal after use of this combined technique. Patients and methods — We investigated the outcome in 184 hips (184 patients) after acetabular revision surgery with TM shells, fitted either with DMCs (n = 69), or with standard polyethylene (PE) liners (n = 115). Chart follow-up was complete for all patients, and the occurrence of dislocations and re-revisions was recorded. 20 were deceased, 50 were unable to attend follow-up, leaving 114 for assessment of hip function after 4.9 (0.5–8.9) years, radiographs were obtained in 99, and tantalum concentrations in 84 patients. Results — 1 patient with a DMC had a dislocation, whereas 14 patients with PE liners experienced at least 1 dislocation. 11 of 15 re-revisions in the PE group were necessitated by dislocations, whereas none of the 2 re-revisions in the DMC group was performed for this reason. Hence, dislocation-free survival after 4 years was 99% (95% CI 96–100) in the DMC group, whereas it was 88% (CI 82–94, p = 0.01) in the PE group. We found no radiographic signs of implant failure in any patient. Mean tantalum concentrations were 0.1 µl/L (CI 0.05–0.2) in the DMC group and 0.1 µg/L (CI 0.05–0.2) in the PE group. Interpretation — Cementing DMCs into TM shells reduces the risk of dislocation after acetabular revision surgery without jeopardizing overall cup survival, and without enhancing tantalum release. ■

Dislocation after hip revision surgery occurs in between 3 and 10% of patients (Swedish Hip Arthroplasty Register 2016). It is the most common reason for re-revision after hip revision surgery during the first 2 years, accounting for 35% of all re-revisions (Springer et al. 2009). Porous tantalum (TM) shells perform well in hip revision surgery, with a low risk of re-revision due to aseptic loosening, but dislocation seems to be an issue after the use of this device (Lakstein et al. 2009, Skyttä et al. 2011, Borland et al. 2012, Kremers et al. 2012, Batuyong et al. 2014, Beckmann et al. 2014, Mohaddes et al. 2015, Konan et al. 2016, Brüggemann et al. 2017). Dual-mobility cups (DMCs) reduce the risk of dislocation after revision hip surgery (Hailer et al. 2012, Gonzalez et al. 2017), and DMCs cemented into TM shells have occasionally been used in an attempt at reducing the risk of dislocation after hip revision surgery, with seemingly satisfactory results (Brüggemann et al. 2017). On the other hand, concerns related to the stability of the DMC inside a TM shell remain, and the concept of cementing DMCs into TM shells has not been systematically investigated. Very little is known regarding the liberation of tantalum ions from TM shells, apart from 1 case report by Babis et al. (2014) that describes grossly elevated serum levels of tantalum after failed revision making use of a TM device. The tantalum concentration in that case was 20 µg/L, which is much higher than the reference interval of 0.008–0.010 µg/L that is derived from a healthy population without implants (Rodushkin et al. 2004). The question remains whether measurable tantalum liberation occurs after the use of TM shells, either from the interface between a DMC and the TM shell, or from the TMbone interface. We hypothesized that cementing a DMC into a TM shell decreases instability. We further hypothesized that cementing

12021 Bru ggemann D.indd 156

2/20/2018 11:07:54 AM

Acta Orthopaedica 2018; 89 (2): 156–162

of DMCs into TM shells results in a stable construct, and does not lead to increased tantalum ion release when compared with standard PE liners within a TM shell.

Patients and methods Implant survival and dislocation rates We identified all acetabular revisions using a TM shell (Zimmer Biomet, Warsaw, IN) during 2008–2016 in our local arthroplasty register. Whenever a patient had revision in both hips, only the hip that was revised first was included (Ranstam et al. 2011). 184 hips were thus included, and we divided this cohort into 2 subgroups. DMC group: 69 hips were treated with a DMC (Avantage; Zimmer Biomet, Warsaw, IN) cemented into a larger TM shell (TM Modular, Trilogy TM or TM Revision shell, Zimmer Biomet, Warsaw, IN). PE group: 115 hips received a standard polyethylene (PE) liner inside a TM shell, either as a standard snap-fit liner when a smaller TM shell was used (Continuum, TM Modular, or Trilogy TM; Zimmer Biomet, Warsaw, IN), or a PE liner cemented into a larger TM shell (TM Modular, Trilogy TM, or TM Revision shell; Zimmer Biomet, Warsaw, IN). PE liners were cemented according to the manufacturer’s instructions. To ensure that no dislocations treated with closed reduction went unnoticed, we accessed all patient charts and searched charts for all closed reductions, thus ensuring complete chart follow-up for all patients. Radiographs Acetabular defect size prior to the index procedure was assessed using the Paprosky classification for all but 1 hip by analyzing radiographs obtained prior to the index procedure (Paprosky et al. 1994). 2 observers (AB, EK) analyzed standard anteroposterior pelvic radiographs and classified bone loss. When in disagreement, consensus was reached under the guidance of the senior author (NPH). We assessed cup migration at follow-up using the method described by Nunn et al. (1989).

157

Table 1. Study population PE (n = 115) Sex: Male Female Paprosky: I IIA IIB IIC IIIA IIIB Reason for index procedure: Loosening Dislocation Infection Other Number of revisions: First-time revision Previously revised Bone graft: No Yes Augments: 0 1 2 Stem revised at index procedure: No Yes

DMC (n = 69)

56 59

34 35

22 16 19 28 12 17

5 10 9 8 18 19

98 3 6 8

56 2 6 5

88 27

46 23

84 26

49 16

109 4 2

57 11 1

78 37

43 26

Note: We found only a statistically significant difference between groups concerning Paprosky classification (p = 0.003) and the use of augments (p = 0.01). Missing data concerning Paprosky classification (1 case in PE group) as well as bone grafting (4 cases in DMC group, 5 cases in PE group).

Characteristics of the study population There were 94 females and 90 males with a mean age of 67 (35–88) years. The main reason for the index procedure was aseptic loosening in both groups. There was a difference between the DMC group and the PE group regarding the severity of the acetabular defects, with a higher proportion of Paprosky grade III A and B defects in the DMC group. Reasons for the index procedure and number of revisions prior to the index procedure were similar between groups (Table 1).

the cup component and reaming for the TM shell, morselized bone graft (42 cases) or augments (18 cases) were used when necessary to fill acetabular defects, a TM shell with appropriate dimensions was impacted into the acetabulum, and this was fixed with additional screws whenever necessary (in 114 cases, median 3 (1–9) screws). Cup positioning within the “safe zones” as proposed by Lewinnek et al. (1978) was aimed for. DMCs with an external diameter at least 14 mm smaller than the external diameter of the implanted TM shells were used to ensure sufficient amounts of bone cement (Palacos R+G; Heraeus Kulzer GmbH, Wehrheim, Germany) between the DMC and the TM shell, i.e., the minimal external diameter for the TM shells in the DMC group was 58 mm, since the smallest DMC had an external diameter of 44 mm. Postoperatively obtained radiographs confirmed adequate positioning of the components. Patients were allowed full weight-bearing postoperatively, even in cases with transfemoral approaches.

Surgical procedures The surgical approach was direct lateral in all but 13 cases where the transfemoral approach was chosen. Revision of the stem at index procedure occurred in 63 cases. After removal of

Clinical follow-up In our department, clinical and radiographic follow-up after acetabular revision surgery is scheduled 3–4 months and 1 year postoperatively. For this study, an additional follow-

12021 Bru ggemann D.indd 157

2/20/2018 11:07:54 AM

158

Acta Orthopaedica 2018; 89 (2): 156–162

Acetabular revisions using a TM shell during 2008–2016 n = 184 DMC n = 69 Excluded (n = 25): – deceased, 5 – severe comorbidity, 13 – declined participation, 7 Available for clinical follow-up (n = 44): – follow-up at our department, 25 (tantalum analysis, radiographs, HHS, HOOS, EQ5D) – follow-up at external unit, 13 (radiographs, HHS, HOOS, EQ5D) – follow-up with questionnaire only, 6 (HOOS, EQ5D)

PE liner n = 115 Excluded (n = 45): – deceased, 15 – severe comorbidity, 24 – declined participation, 6 Available for clinical follow-up (n = 70): – follow-up at our department, 59 (tantalum analysis, radiographs, HHS, HOOS, EQ5D) – follow-up at external unit, 2 (radiographs, HHS, HOOS, EQ5D) – follow-up with questionnaire only, 9 (HOOS, EQ5D)

Figure 1. Description of the inclusion of the study population.

up visit was offered. Of the 184 patients, 20 were deceased at final follow-up, and 37 patients were excluded from the clinical follow-up due to significant comorbidity, leaving 127 patients eligible for clinical follow-up. 13 patients declined to participate in this study, leaving 114 of 184 patients for clinical follow-up (Figure 1). All patients received the Hip disability and Osteoarthritis Outcome Score (HOOS) and EQ5D via surface mail and were asked to return the forms. 15 patients from outside our county were scheduled for follow-up at the orthopedic department closest to their place of residence. For these patients, an analysis of tantalum ion concentration was not possible. The HOOS and EQ5D questionnaires were filled in by 114 patients. Hip function was assessed using the Harris Hip Score (HHS) in 99 hips. Standard anteroposterior and lateral radiographs were taken in 99 hips and analyzed by 2 observers regarding cup abduction and anteversion angles, migration of the implant, screw breakage, osteolysis, or radiolucent lines. Osteolysis, migration exceeding 2 mm, screw breakage, or radiolucent lines > 30% of the bone-implant interface were considered radiographic signs of loosening. Tantalum concentrations were measured in 84 of 184 patients. Whole-blood samples were obtained, the same batch of cannulas and tubes (10-mL polypropylene with sodium heparin) was used, and samples were analyzed by a certified laboratory (ALS Scandinavia AB, Luleå, Sweden) using inductively coupled plasma sector field mass spectrometry (ICP-SFMS) with a tantalum detection threshold of 0.05 µg/L. Statistics Continuous data were described as medians with ranges or means with standard deviation (SD). Estimation uncertainty was approximated by 95% confidence intervals (CI). Categorical data were summarized in cross-tables and the chi-square

12021 Bru ggemann D.indd 158

test was used to investigate differences between groups. We calculated cumulative unadjusted component survival using the Kaplan–Meier method, with the endpoints dislocation, or rerevision due to instability, due to aseptic loosening, or for any reason. The Mantel–Haenszel log-rank test estimated statistical difference in survival between groups. We fitted Cox exploratory multivariable regression models to determine hazard ratios (HR) adjusted for the confounders age, sex, and acetabular defect size (assessed by the Paprosky classification). P-values < 0.05 were considered statistically significant. All data were analyzed using the R software version 3.3.1 and RStudio version 0.99.993 with the “Gmisc” and “rms” packages (R Core Team 2016).

Ethics, funding, and potential conflicts of interest This study was conducted in accordance with the Helsinki declaration and was approved by the local ethics committee (Regionala Etikprövningsnämnden Uppsala, entry no. 2014/108, April 16, 2014, and entry no. 2014/108/3, April 19, 2017). The study was in part financed by Zimmer. However, neither Zimmer nor its employees took any part in study design, collecting the data, analysis of the data, or writing of the manuscript.

Results Dislocation rates There was 1 dislocation in the DMC group compared with 14 dislocations in the PE group. Of the 14 dislocated hips in the PE group 12 were uncemented snap-fit liners; femoral head size was 28 mm in 9 cases and 32 mm in 5. Head size was not statistically significantly associated with dislocation. 6 of the 14 dislocated hips in the PE group underwent stem revision at index procedure, as did the only dislocated hip in the DMC group. The dislocated hip in the DMC group remained dislocated because the patient’s general health was too poor to proceed with further surgery. 11 of the 14 patients with dislocations in the PE group underwent re-revision due to persistent instability, 1 patient was treated with closed reduction and remained dislocation-free until follow-up, 1 patient was diagnosed with a periprosthetic joint infection, and 1 patient was re-revised due to aseptic loosening that had caused the dislocation. Dislocation-free survival after 4 years was thus 99% (CI 96–100) in the DMC group, whereas it was 88% (CI 82–94, p = 0.01, Figure 2) in the PE group. The adjusted HR for the occurrence of dislocation was 0.11 (CI 0.01–0.8, p = 0.03) for the DMC group.

2/20/2018 11:07:54 AM

Acta Orthopaedica 2018; 89 (2): 156–162

159

Table 2. Description of the re-revised study population PE (n = 15)

Figure 2. Kaplan–Meier survival curves with the endpoint dislocation with shaded area indicating CI (p = 0.01, derived from Mantel– Haenszel log-rank test). Numbers at risk for both subgroups are given above the X-axis.

Analysis of postoperative radiographs showed that mean anteversion was 13 (range 0–35) degrees, mean inclination was 46 (range 35–58) degrees, and cranialization of the center of rotation was mean 3 (range –11 to 13) mm in the group of dislocated implants. Implant survival Within the DMC group, 2 re–revisions occurred, both due to aseptic loosening at the interface between the TM shell and bone, not between the DMC and the TM shell. In the PE group, there were altogether 15 re-revisions. Apart from the 13 re-revisions accounted for above, 1 additional hip was re-revised due to aseptic loosening, and 1 was revised due to persistent pain related to a cancellous screw penetrating the inner pelvic cortex into the iliopsoas muscle (Table 2). For the endpoint re-revision due to dislocation we thus obtained 100% survival for the DMC group after 4 years, whereas it was 89% (95% CI 84–95, p = 0.006, Figure 3) in the PE group. With rerevision for any reason as the endpoint, 4-year survival within the DMC group was 96% (CI 90–100), whereas it was 87% (CI 81–93) for the PE group (p = 0.03, Figure 4). With rerevision due to aseptic loosening as the endpoint, the DMC group had a 4-year survival of 96% (CI 90–100), while it was 98% (CI 95–100) for the PE group (p = 0.5). Clinical follow-up The mean tantalum concentrations were 0.1 µl/L (CI 0.05–0.2) in the DMC group and 0.1 µg/L (CI 0.05–0.2) in the PE group, with 3 of 25 patients below detection limit in the DMC group, and 5 of 59 patients below detection limit in the PE group. The maximal tantalum concentration of 0.5 µg/L was measured in a patient in the PE group; the maximal concentration in the DMC group was 0.3µg/L. After a mean of 4.9 (range 0.5–8.9) years, none of the hips that underwent radiographic examination showed signs of migration above the predefined threshold, osteolysis, breakage of screws, or radiolucent lines exceeding 30% of the cup circumference. Overall, hip function was satisfactory, with a

12021 Bru ggemann D.indd 159

Sex: Male Female Paprosky: I IIA IIB IIC IIIA IIIB Reason for index procedure: Loosening Dislocation Infection Other Number of revisions: First-time revision Previously revised Bone graft: No Yes Stem revised at index procedure: No Yes Head size: 22 28 32 Other Reason for re-revision: Aseptic loosening Dislocation Pain Infection

DMC (n = 2)

6 9

1 1

2 1 1 8 2 1

0 0 1 0 0 1

12 1 0 2

2 0 0 0

11 4

2 0

11 3

0 2

9 6

2 0

0 9 4 2

1 1 0 0

2 11 1 1

2 0 0 0

Note: There was no statistically significant difference between the 2 groups. Missing data for 1 patient in the PE group concerning bone grafting.

mean HHS of 77 (25–100). In the HOOS subdimensions there were only small differences between the 2 groups that were neither statistically nor clinically significant. There were small and statistically not significant differences between groups in global health, as assessed by EQ5D (Tables 3 and 4).

Discussion Recurrent dislocation after hip revision surgery is a frequent and devastating complication; thus technical concepts that reduce joint instability after acetabular revision surgery are warranted. In our present study, no re-revision was performed due to instability after the use of a DMC in conjunction with a TM shell, indicating that cementing DMCs into TM shells may be an adequate way to address joint instability after hip revision surgery. It is noteworthy that there is a selection bias to the disadvantage of the DMC group, since the combination of a DMC inside a TM shell was chosen when the risk of

2/20/2018 11:07:59 AM

160

Acta Orthopaedica 2018; 89 (2): 156–162

Figure 3. Kaplan–Meier survival curves with the endpoint re-revision for instability with shaded area indicating CI (p = 0.006, derived from Mantel–Haenszel log-rank test). Numbers at risk for both subgroups are given above the X-axis.

Figure 4. Kaplan–Meier survival curves with the endpoint re-revision for any reason with shaded area indicating CI (p = 0.03, derived from Mantel–Haenszel log-rank test). Numbers at risk for both subgroups are given above the X-axis.

Table 3. EQ5D subdimensions

Table 4. HOOS subdimensions

EQ5D Mobility: No problems Some problems Extreme problems Anxiety/depression: No problems Some problems Extreme problems Usual activities: No problems Some problems Extreme problems Pain/discomfort: No problems Some problems Extreme problems Self-care: No problems Some problems Extreme problems

PE (n = 69)

DMC (n = 44)

11 44 13

17 23 4

16 20 33

23 13 8

14 22 32

25 11 8

10 39 20

13 24 7

16 12 41

33 4 7

Pain: Symptoms: ADL: Sport/recreation: QOL:

Note: 1 patient in the PE group did not answer all questions.

instability was considered higher than average, and because the combination of a DMC with a TM shell was more often used in more severe acetabular defects. It should be kept in mind that the majority of hips within the PE group that dislocated had received uncemented snap-fit liners. Smaller head size seemed not to be associated with an increased risk of dislocation, and we found no obvious malpositioning of the cup in the cases that suffered from dislocation. It has been argued that procedures involving isolated cup revision are associated with a higher risk of dislocation than procedures where cup and stem revision are combined (Mohaddes et al. 2013), but in our material this could not be confirmed. The risk of dislocation is substantial after hip revision surgery—also after the use of TM shells. Several studies have shown that recurrent dislocation is an issue after the use of

12021 Bru ggemann D.indd 160

HOOS

group

mean

range

PE DMC PE DMC PE DMC PE DMC PE DMC

70 43 70 44 70 44 70 44 70 44

70 73 68 71 67 69 44 45 52 58

28 25 25 27 28 27 33 30 30 30

0–100 5–100 10–100 10–100 0–100 12–100 0–100 0–100 0–100 0–100

Note: 1 patient did not answer all questions. No statistically significant difference was found.

TM shells (Unger et al. 2005, Lakstein et al. 2009, Kremers et al. 2012). Most notably, the authors of a study on 827 patients with TM revision shells (Skyttä et al. 2011) concluded that “the most common reason for revision was dislocation of the prosthesis with or without malposition of the socket (60%)”. In our study, dislocation was the main reason for re-revision when using PE liners in conjunction with TM shells. However, when patients within the PE group were re-revised with a DMC due to recurrent dislocations (7 patients), no further dislocations occurred. Of the 4 remaining patients that underwent re-revision due to instability but did not receive a DMC, 1 patient remained unstable. Cementing a DMC into a TM shell could thus also be considered suitable in cases of recurrent dislocations after acetabular revision surgery. Despite corrosion, metallosis, and adverse tissue reactions to metal implants being hotly debated, we were unable to find systematic studies on tantalum ion release after the use of TM shells. We therefore determined serum concentrations of tantalum in 84 patients as a measure of tantalum release, and found a 10-fold increase compared with the reference value of 0.008–0.010 µg/L obtained in a healthy population without any implants, but about a 100-fold lower mean con-

2/20/2018 11:07:59 AM

Acta Orthopaedica 2018; 89 (2): 156–162

centration than in the cited case report (Babis et al. 2014). It should be kept in mind that our measurements are derived from patients with radiographically stable constructs, and the extensive increase in tantalum concentrations observed after a failed revision THA was not expected in our patients. Tantalum concentrations were not statistically significantly higher in patients who had received a DMC cemented into a TM shell when compared with those who had PE liners. Whether the elevated tantalum concentrations observed in both groups we investigated are of clinical importance is an open question, and no reference group with hip implants made of materials other than tantalum was available. Our results indicate a satisfactory clinical outcome in terms of patient-reported outcome measures (PROM), and they are similar to those described by other authors (Flecher et al. 2008, Batuyong et al. 2014, Konan et al. 2016). PROM were slightly better in the DMC group, but the observed differences compared with the PE group were small. The minimal clinically important difference for the scores reported here is higher than the estimated difference between our 2 groups, indicating that these findings lack clinical relevance (Berliner et al. 2016, Singh et al. 2016). Strengths and weaknesses of the study There are several limitations to this study. Since it is a retrospective cohort study we faced the common problems of missing data, underreporting, relatively large loss to clinical follow-up, and possible confounders that we cannot control for. Whether the actual acetabular defects are correctly represented by the radiographic assessment that was based on Paprosky’s classification is unknown because operative notes were not detailed enough on this topic. PROM and functional outcome scores were only established at clinical follow-up but not at baseline, and we are therefore unable to conclude whether the chosen treatment improved our patients’ quality of life. Mean follow-up was shorter for the DMC group than for the PE group because the technique of cementing a DMC into a TM shell gradually replaced the previously common use of PE liners. Due to the stepwise introduction of different implants and due to different sizes of acetabular defects the Continuum, the TM Modular, and the Trilogy TM cup as well as the TM revision shell are included in this study. However, we investigated porous tantalum as a method of fixation— rather than a specific cup design. The 2 investigated groups are not equal with respect to the severity of acetabular defects, and surgeons may have chosen a DMC in cases considered “more difficult” for other reasons than acetabular defect size, introducing residual confounding. We fitted exploratory multivariable regression models to estimate the adjusted risk of dislocation or re-revision for various reasons, but the scarcity of events inflated confidence intervals and estimation uncertainty was therefore large. Instability causing re-revision generally precludes re-revision due to aseptic loosening as an endpoint, which implies that the risk

12021 Bru ggemann D.indd 161

161

of re-revisions due to aseptic loosening in the PE group could be higher than our estimate. There are, however, certain strengths to this study. Our cohort of 184 patients with information on Paprosky classification is relatively large. Our analysis of all patients’ charts—even charts kept at external units—enabled us to draw conclusions not only regarding the endpoint re-revision due to instability, but also regarding the endpoint dislocation. We also have reliable data concerning demographics, re-revisions, and deaths. Furthermore, clinical follow-up was available for 114 hips, a rather large cohort as regards acetabular revision surgery. Our evaluation of tantalum concentrations in 84 patients is the first of its kind. This study is also the first systematic analysis of TM shells combined with DMCs, comparing a new technique with the previously recommended procedure based on the use of standard PE liners. Conclusion We believe that the technique of cementing a DMC into a TM shell is a valuable treatment option in acetabular revision surgery, especially for patients that are at higher risk of dislocation. The new technique seems to be safe since we found no increased risk of implant loosening, and no increased liberation of tantalum ions.

All authors were involved in writing the manuscript. AB collected data under the guidance of HM. AB and NPH analyzed the radiographs. AB and NPH performed the statistical analysis.

We would like to thank Elin Kramer for her efforts during follow-up and analysis of radiographs. Special thanks to our department colleagues associate professor Jan Milbrink and professor Olle Nilsson for introducing this new treatment strategy. Acta thanks Martin Clauss and Marc Nijhof for help with peer review of this study. Babis G C, Stavropoulos N A, Sasalos G, Ochsenkuehn–Petropoulou M, Megas P. Metallosis and elevated serum levels of tantalum following failed revision hip arthroplasty: A case report. Acta Orthop 2014; 85 (6): 677-80. Batuyong E D, Brock H S, Thiruvengadam N, Maloney W J, Goodman S B, Huddleston J I. Outcome of porous tantalum acetabular components for Paprosky type 3 and 4 acetabular defects. J Arthroplasty 2014; 29 (6): 1318-22. Beckmann N A, Weiss S, Klotz M C, Gondan M, Jaeger S, Bitsch R G. Loosening after acetabular revision: Comparison of trabecular metal and reinforcement rings. A systematic review. J Arthroplasty 2014; 29 (1): 229-35. Berliner J L, Brodke D J, Chan V, SooHoo N F, Bozic K J. John Charnley Award: Preoperative patient-reported outcome measures predict clinically meaningful improvement in function after THA. Clin Orthop Relat Res 2016; 474 (2): 321-9. Borland W S, Bhattacharya R, Holland J P, Brewster N T. Use of porous trabecular metal augments with impaction bone grafting in management of acetabular bone loss. Acta Orthop 2012; 83(4): 347-52. Brüggemann A, Fredlund E, Mallmin H, Hailer N P. Are porous tantalum cups superior to conventional reinforcement rings? Acta Orthop 2017; 88 (1): 35-40.

2/20/2018 11:08:01 AM

162

Flecher X, Sporer S, Paprosky W. Management of severe bone loss in acetabular revision using a trabecular metal shell. J Arthroplasty 2008; 23(7): 949-55. Gonzalez A I, Bartolone P, Lubbeke A, Dupuis Lozeron E, Peter R, Hoffmeyer P, Christofilopoulos P. Comparison of dual-mobility cup and unipolar cup for prevention of dislocation after revision total hip arthroplasty. Acta Orthop 2017; 88 (1): 18-23. Hailer N P, Weiss R J, Stark A, Karrholm J. Dual-mobility cups for revision due to instability are associated with a low rate of re-revisions due to dislocation: 228 patients from the Swedish Hip Arthroplasty Register. Acta Orthop 2012; 83 (6): 566-71. Konan S, Duncan CP, Masri B A, Garbuz D S. Porous tantalum uncemented acetabular components in revision total hip arthroplasty: A minimum tenyear clinical, radiological and quality of life outcome study. Bone Joint J 2016; 98-B (6): 767-71. Kremers H M, Howard J L, Loechler Y, Schleck C D, Harmsen W S, Berry D J, Cabanela M E, Hanssen A D, Pagnano M W, Trousdale R T, Lewallen D G. Comparative long-term survivorship of uncemented acetabular components in revision total hip arthroplasty. J Bone Joint Surg Am 2012; 94 (12): e82. Lakstein D, Backstein D, Safir O, Kosashvili Y, Gross A E. Trabecular metal cups for acetabular defects with 50% or less host bone contact. Clin Orthop Relat Res 2009; 467 (9): 2318-24. Lewinnek G E, Lewis J L, Tarr R, Compere C L, Zimmerman J R. Dislocations after total hip-replacement arthroplasties. J Bone Joint Surg Am 1978; 60(2): 217-20. Mohaddes M, Garellick G, Karrholm J. Method of fixation does not influence the overall risk of rerevision in first-time cup revisions. Clin Orthop Relat Res 2013; 471 (12): 3922-31. Mohaddes M, Rolfson O, Karrholm J. Short-term survival of the trabecular metal cup is similar to that of standard cups used in acetabular revision surgery. Acta Orthop 2015; 86 (1): 26-31.

12021 Bru ggemann D.indd 162

Acta Orthopaedica 2018; 89 (2): 156–162

Nunn D, Freeman M A, Hill P F, Evans S J. The measurement of migration of the acetabular component of hip prostheses. J Bone Joint Surg Br 1989; 71 (4): 629-31. Paprosky W G, Perona P G, Lawrence J M. Acetabular defect classification and surgical reconstruction in revision arthroplasty: A 6-year follow-up evaluation. J Arthroplasty 1994; 9 (1): 33--44. R Core Team. R: A language and environment for statistical computing. 2016. http://www.R-project.org/. Ranstam J, Kärrholm J, Pulkkinen P, Mäkelä K, Espehaug B, Pedersen A B, Mehnert F, Furnes O. Statistical analysis of arthroplasty data, II: Guidelines. Acta Orthop 2011; 82 (3): 258-67. Rodushkin I, Engstrom E, Stenberg A, Baxter D C. Determination of lowabundance elements at ultra-trace levels in urine and serum by inductively coupled plasma-sector field mass spectrometry. Anal Bioanal Chem 2004; 380(2): 247-57. Singh J A, Schleck C, Harmsen S, Lewallen D. Clinically important improvement thresholds for Harris Hip Score and its ability to predict revision risk after primary total hip arthroplasty. BMC Musculoskeletal Disorders 2016; 17: 256. Skyttä E T, Eskelinen A, Paavolainen P O, Remes V M. Early results of 827 trabecular metal revision shells in acetabular revision. J Arthroplasty 2011; 26 (3): 342-5. Springer B D, Fehring T K, Griffin W L, Odum S M, Masonis J L. Why revision total hip arthroplasty fails. Clin Orthop Relat Res 2009; 467 (1): 166-73. Swedish Hip Arthroplasty Register. Annual Report 2016, https:// registercentrum.blob.core.windows.net/shpr/r/-rsrapport-2016SJirXXUsb.pdf” Unger A S, Lewis R J, Gruen T. Evaluation of a porous tantalum uncemented acetabular cup in revision total hip arthroplasty: Clinical and radiological results of 60 hips. J Arthroplasty 2005; 20 (8): 1002-9.

2/20/2018 11:08:01 AM

Acta Orthopaedica 2018; 89 (2): 163–169

163

The effect of bearing type on the outcome of total hip arthroplasty Analysis of 209,912 primary total hip arthroplasties registered in the Dutch Arthroplasty Register Rinne M PETERS 1,2, Liza N VAN STEENBERGEN 3, Martin STEVENS 2, Paul C RIJK 1, Sjoerd K BULSTRA 2, and Wierd P ZIJLSTRA 1

1 Department of Orthopedic Surgery, Medical Center Leeuwarden; 2 Department of Orthopedic Surgery, University of Groningen, University Groningen; 3 Dutch Arthroplasty Register (Landelijke Registratie Orthopedische Implantaten), ‘s Hertogenbosch, The Netherlands

Medical Center

Correspondence: rinnepeters@gmail.com Submitted 2017-07-12. Accepted 2017-10-31.

Background and purpose — Alternative bearing surfaces such as ceramics and highly crosslinked polyethylene (HXLPE) were developed in order to further improve implant performance of total hip arthroplasties (THAs). Whether these alternative bearing surfaces result in increased longevity is subject to debate. Patients and methods — Using the Dutch Arthroplasty Register (LROI), we identified all patients with a primary, non-metal-onmetal THA implanted in the Netherlands in the period 2007–2016 (n = 209,912). Cumulative incidence of revision was calculated to determine differences in survivorship of THAs according to bearing type: metal-on-polyethylene (MoPE), metal-on-HXLPE (MoHXLPE), ceramic-on-polyethylene (CoPE), ceramic-onHXLPE (CoHXLPE), ceramic-on-ceramic (CoC), and oxidizedzirconium-on-(HXL)polyethylene (Ox(HXL)PE). Multivariable Cox proportional hazard regression ratios (HRs) were used for comparisons. Results — After adjustment for confounders, CoHXLPE, CoC, and Ox(HXL)PE resulted in a statistically significantly lower risk of revision compared with MoPE after 9 years follow-up (HR = 0.8–0.9 respectively, compared with HR = 1.0). For small (22–28 mm) femoral head THAs, lower revision rates were found for CoPE and CoHXLPE (HR = 0.9). In the 36 mm femoral head subgroup, CoC-bearing THAs had a lower HR compared with MoHXLPE (HR = 0.7 versus 1.0). Crude revision rates in young patients (< 60 years) for CoHXLPE, CoC, and Ox(HXL)PE (HR = 0.7) were lower than MoPE (HR = 1.0). However, after adjustment for case mix and confounders these differences were not statistically significant. Interpretation — We found a mid-term lower risk of revision for CoHXLPE, CoC, and Ox(HXL)PE bearings compared with traditional MoPE-bearing surfaces. ■

Increased activity of patients and a younger age at the time of the primary procedure have sparked the development of alternative bearing surfaces in total hip arthroplasty (THA) such as ceramics, highly-crosslinked-polyethylene (HXLPE), and metal-on-metal articulations (MoM), in order to further improve survival and implant performance (Mihalko et al. 2014, Varnum et al. 2015). Currently, aseptic loosening of the acetabular component is the most frequent cause of revision after THA with a metal-on-polyethylene (MoPE) counterface (LROI annual report 2015, Norwegian Arthroplasty Register 2016). Osteolysis with subsequent loosening of components can be generated by polyethylene (PE) particles as a result of PE wear (Varnum et al. 2015). Therefore, the use of alternative bearing surfaces has become more common over the last 2 decades. It is unknown whether the survivorship of these implants is better compared with the traditional MoPE bearings they sought to replace. Studies which compared the survival of different bearing surfaces attained variant conclusions. The Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) demonstrated superior results of HXLPE, ceramics, and ceramicized metal (or oxidized zirconium) in terms of increased longevity of the THA compared with standard PE (Annual Report AOANJRR 2016). A systematic review and network meta-analysis of randomized controlled trails demonstrated similar survivorship among ceramic-on-ceramic (CoC), ceramic-on-polyethylene (CoPE), ceramic-on-highlycrosslinked-polyethylene (CoHXLPE) and metal-on-highlycrosslinked-polyethylene (MoHXLPE) bearings, and inferior results for MoM and MoPE bearing implants (Yin et al. 2015). Whether these alternative bearing materials, in combination with larger heads, have indeed resulted in increased survival

11913 Peters D.indd 163

21-02-2018 15:24:31

164

rates, however, remains to be proven. Using nationwide data from the LROI, we assessed survivorship of CoC, CoHXLPE, MoHXLPE, CoPE, and oxidized-zirconium-on-(highly crosslinked)-polyethylene (Ox(HXL)PE) bearings in THA in the Netherlands, compared with MoPE.

Patients and methods Data sources The LROI, initiated by the Dutch Orthopaedic Association in 2007, is a nationwide population-based registry covering all hospitals in the Netherlands. This inter-institutional database has a completeness of 98% for primary THA and 88% for hip revision arthroplasty (van Steenbergen et al. 2015). The LROI contains prospectively collected data on primary and revision arthroplasty. Patient characteristics are recorded at the time of the primary procedure. In addition, surgical variables such as procedure and implant information are registered in the LROI. Implant information is supplied by all manufacturers, and is collected at the time of the procedure using stickers that could be attached to a registration form. Thereafter, prosthesis characteristics are derived from an implant library within the LROI, which contains several core characteristics of all prostheses used in the Netherlands, including name and type of the prosthesis, manufacturer, material, and femoral head size (van Steenbergen et al. 2015). Data from the LROI are matched with the national insurance database on healthcare (Vektis 2017), in order to obtain information on the vital status and date of death of registered patients. Data collection Eligible patients were registered in the LROI as having received a primary THA in a Dutch hospital, from the start of the registry in 2007 until the end of the follow-up period on December 31, 2016 (n = 227,107). A patient can be registered twice, having undergone a bilateral hip replacement. A primary THA is defined as the first implantation of a prosthesis, to replace a hip joint (van Steenbergen et al. 2015). Given their now known higher failure rates, THAs with a MoM bearing surface were excluded (n = 5,359) (Drummond et al. 2015, Nederlandse Orthopaedische Vereniging 2015, Rieker 2017). Patients with unknown prosthesis components or patients for whom 1 of the components was not registered were excluded (n = 11,836). The final cohort contained 209,912 THAs. The mean length of follow-up was 3.9 years, with a maximum of 9.9 years. Types of bearing surface Hip arthroplasty articulation was differentiated based on the bearing surface of the femoral head and the inlay or monoblock cup. Metal-on-standard-polyethylene was used as reference bearing type. All other bearing surfaces, except for ceramicon-polyethylene, were considered as an alternative bearing

11913 Peters D.indd 164

Acta Orthopaedica 2018; 89 (2): 163â&#x20AC;&#x201C;169

type. The following groups were discerned: metal-on-polyethylene (MoPE), metal-on-highly-crosslinked-polyethylene (MoHXLPE), ceramic-on-polyethylene (CoPE), ceramic-onhighly-crosslinked-polyethylene (CoHXLPE), ceramic-onceramic (CoC), and oxidized-zirconium-on-(highly-crosslinked)-polyethylene (Ox(HXL)PE). Due to small group sizes, prostheses with an oxidized-zirconium-on-standard-PE (OxPE) and oxidized-zirconium-on-highly-crosslinked-polyethylene (OxHXLPE) were analyzed together. For demographics on all registered patients see Table 5, supplementary data. Categories for these explanatory variables were, similar to previous studies, classified using data from the LROI (Peters et al. 2016, Zijlstra et al. 2017). Procedure and implant information (surgical variables) were retrieved, e.g., fixation technique, surgical approach, and reason for revision. Statistics Group comparisons were made using a chi-square test to test for differences in patient and prosthesis characteristics. Survival time (with 95% confidence interval (CI)) was calculated as the time from primary THA to first revision arthroplasty for any reason, death of the patient, or the end of followup. Cumulative crude incidence of revision was calculated using competing risk analysis, where death was considered to be a competing risk (Lacny et al. 2015, Wongworawat et al. 2015). The consequence of using Kaplanâ&#x20AC;&#x201C;Meier is that the probability of revision will be overestimated (Putter et al. 2007, Keurentjes et al. 2012). Crude cumulative revision percentages within 5 and 9 years were calculated. In addition, revision rates within 9 years according to the reason for revision were estimated for different bearing types. Differences were compared using a chi-square test. In order to test for differences in revision rates between subgroups, hazard ratios were calculated using multivariable Cox proportional hazards regression analyses adjusting for possible confounding variables. The following confounders were entered into our analysis: age, sex, ASA score, diagnosis, previous operation to the affected hip joint, fixation technique, femoral head diameter, surgical approach, and period of surgery. For all covariates added, the proportional hazards assumption was checked by inspecting log-minus-log curves (JĂ¤msen et al. 2014). Differences in revision rate for the different bearing types in patients younger than 60 or with different sizes of femoral head were assessed using multivariable Cox proportional hazards regression analyses. Due to small numbers (1,451 cases, 38 revision procedures) for the subgroup of 38 mm femoral head components, multivariable regression analysis of this subgroup was not feasible. P-values < 0.05 were considered statistically significant. All analyses were performed using SPSS for Windows version 23.0 (IBM Corp, Armonk, NY, USA).

21-02-2018 15:24:31

Acta Orthopaedica 2018; 89 (2): 163–169

165

Table 1. Reasons for revision or reoperation in revised THAs performed in 2007–2016 in the Netherlands (n = 6,515)

Revision within follow-up period

MoPE (n = 1,023) n %

MoHXLPE (n = 890) n %

CoPE (n = 1,186) n %

CoHXLPE (n = 1,649) n %

Dislocation Loosening of acetabulum Infection Loosening of femur Periprosthetic fracture Cup/liner wear Girdlestone Periarticular ossification Other

391 190 163 145 106 27 35 12 158

248 97 165 213 166 17 32 19 133

393 171 180 262 118 30 44 36 182

498 162 330 323 283 29 52 23 249

38 19 16 14 10 2.6 3.4 1.2 15

28 11 19 24 19 1.9 3.6 2.1 15

33 14 15 22 9.9 2.5 3.7 3.0 15

30 9.8 20 20 17 1.8 3.2 1.4 15

CoC (n = 454) n % 91 46 51 112 42 15 14 8 133

20 10 11 25 9.3 3.3 3.1 1.8 29

Ox(HXL)PE (n = 262) n % 60 39 35 69 59 9 9 2 39

23 15 13 26 23 3.4 3.4 0.8 15

Total c (n = 5,464) n % 1681 705 924 1124 774 127 186 100 894

31 a 13 a 17 a 21 a 14 a 2.3 3.4 1.8 16 a

C – ceramic, HXL – highly crosslinked, M – metal, Ox – oxidized zirconium, PE – polyethylene. a p < 0.001 between different bearing types. b A patient may have more than 1 reason for revision or reoperation. As such, the total is over 100%.

Table 2. Crude cumulative incidence of revision in THAs performed in 2007–2016 in the Netherlands

Revision for any reason

MoPE (n = 37,351) % (CI)

MoHXLPE (n = 32,867) % (CI)

CoPE (n = 40,109) % (CI)

CoHXLPE (n = 70,175) % (CI)

CoC (n = 17,625) % (CI)

Ox(HXL)PE (n = 11,785) % (CI)

5 year 9 year

2.7 (2.5–2.9) 3.9 (3.6–4.2)

3.3 (3.1–3.5) 4.2 (3.8–4.6)

3.0 (2.8–3.2) 4.0 (3.7–4.3)

2.9 (2.7–3.0) 4.0 (3.6–4.4)

2.8 (2.5–3.0) 4.1 (3.4–4.9)

2.5 (2.2–2.8) 3.5 (3.0–4.1)

For abbreviations, see Table 1.

Results The most frequently employed bearing surface between 2007 and 2016 was CoHXLPE (n = 70,175), followed by CoPE (n = 40,109), MoPE (n = 37,351), MoHXLPE (n = 32,867), CoC (n = 17,625), and Ox(HXL)PE (n = 11,785) (Table 5, Supplementary data). Reasons for revision The most common reason for revision was dislocation (31%), followed by femoral loosening (21%), and infection (17%) (Table 1). Revision due to dislocation was more frequently registered in THAs with a MoPE bearing surface (38%) compared with other bearing types, but less frequent in CoC and Ox(HXL)PE. Revision due to femoral loosening was more frequently registered in CoC (25%), and Ox(HXL)PE (26%). Periprosthetic fractures which necessitated revision were less frequently registered in MoPE (10%), CoPE (10%), and CoC (9%) THAs compared with other bearings (Table 1). Overall crude cumulative incidence of revision In total, 5,464 THAs were revised within the follow-up period. The overall, unadjusted 5- and 9-year cumulative incidence of revision for traditional MoPE THAs were respectively 2.7% (CI 2.5–2.9) and 3.9% (3.6–4.2) (Table 2, Figure 1). After 5 years, MoHXLPE showed a higher cumulative incidence of revision compared with MoPE. At 9 years, there were no dif-

11913 Peters D.indd 165

Figure 1. Cumulative incidence of revision according to bearing type of THA in the Netherlands in the period 2007–2016.

ferences in crude revision rate between the various bearings (Table 2). For MoHXLPE, crude hazard ratio (HR) for revision was higher than for MoPE (HR = 1.18; CI 1.08–1.29) (Table 3). Other bearing couples did not display improved crude revision rates over MoPE. Overall multivariable (case-mix adjusted) revision rates Since the risk of revision can be influenced by case-mix, prosthesis, and operation characteristics, we performed multivari-

21-02-2018 15:24:31

166

Acta Orthopaedica 2018; 89 (2): 163–169

Table 3. Multivariable survival analyses of patients with a THA in the period 2007–2016 in the Netherlands (n = 209,912)

Articulation a MoPE MoHXLPE CoPE CoHXLPE CoC Ox(HXL)PE

Crude hazard ratio for revision (CI) 1.0 1.18 (1.08–1.29) d 1.08 (0.99–1.17) 1.08 (1.00–1.17) 1.01 (0.91–1.13) 0.94 (0.82–1.08)

Adjusted hazard ratio for revision b (CI) 1.0 0.98 (0.88–1.09) 0.99 (0.90–1.08) 0.87 (0.79–0.96) c 0.82 (0.71–0.94) c 0.81 (0.70–0.94) c

a For abbreviations, see Table 1. b Adjusted for age at surgery, sex,

ASA score, diagnosis, previous operation, fixation, head diameter, surgical approach, and period. c p < 0.05. d p < 0.001.

Table 4. Multivariable survival analysis of patients with different femoral head components Femoral head Articulation

n (revisions)

22–28 mm (n = 73,114) MoPE 27,423 (843) MoHXLPE 7,236 (256) CoPE 22,165 (660) CoHXLPE 14,188 (367) CoC 1,406 (42) Ox(HXL)PE 696 (17) 32 mm (n = 96,330) MoPE 9,908 (179) MoHXLPE 17,248 (377) CoPE 17,888 (525) CoHXLPE 40,496 (877) CoC 3,279 (99) Ox(HXL)PE 7,511 (158) 36 mm (n = 39,017) MoPE 13 (0) MoHXLPE 8,124 (253) CoPE 56 (1) CoHXLPE 15,490 (405) CoC 11,756 (280) Ox(HXL)PE 3,578 (87)

Crude hazard Adjusted hazard ratio for revision (CI) ratio a (CI) 1.0 1.3 (1.2–1.5)c 1.0 (0.9–1.1) 1.1 (1.0–1.2) 1.0 (0.7–1.4) 0.8 (0.5–1.4)

1.0 1.1 (1.0–1.3) 0.9 (0.8–1.0) b 0.9 (0.7–1.0) b 0.8 (0.6–1.1) 0.7 (0.5–1.2)

1.0 1.4 (1.1–1.6) b 1.5 (1.3–1.8) c 1.3 (1.1–1.6) b 1.5 (1.2–1.9) b 1.1 (0.9–1.4)

1.0 1.1 (0.9–1.3) 1.3 (1.1–1.6) b 1.0 (0.9–1.2)

n.a. d 1.0 n.a. d 1.0 (0.8–1.1) 0.8 (0.6–0.9) b 0.9 (0.7–1.2)

n.a. d 1.0 n.a. d 0. 9 (0.8–1.1) 0.7 (0.6–0.9) b 0.9 (0.7–1.1)

0.9 (0.8–1.2)

a Adjusted

for sex, ASA score, diagnosis, previous operation, fixation, surgical approach, and period. b p < 0.05. c p < 0.001. d n.a. = not applicable; hazard ratios and confidence intervals for the MoPE and CoPE articulation were not applicable due to small number of revisions.

Figure 2. Cumulative incidence of revision according to bearing type for patients aged younger than 60.

able survival analyses, adjusted for age, sex, ASA, diagnosis, previous operation, fixation, head diameter, surgical approach, and period of surgery. These analyses showed that CoHXLPE, CoC, and Ox(HXL)PE had a 13–19% lower risk of revision compared with MoPE (respectively HR = 0.87; CI 0.8–1.0, HR = 0.82; CI 0.7–0.9, and HR = 0.81; CI 0.7–0.9) (Table 3). Revision rate in young patients (< 60 years) In patients under 60 years, THAs with a CoHXLPE, CoC, and Ox(HXL)PE bearing surface were less frequently revised compared with traditional MoPE THAs (respectively HR = 0.73; CI 0.60–0.88, HR = 0.68; CI 0.55–0.85, and HR = 0.74; CI 0.56– 0.98 versus HR = 1.0) (Figure 2). However, after adjustment for case mix and confounders, revision rates were similar. Revision rates and femoral head size Subgroup analyses for different femoral head sizes were performed. For small femoral head components (22–28 mm), the adjusted analyses demonstrated statistically significant lower revision rates for CoPE and CoHXLPE compared with MoPE

11913 Peters D.indd 166

(HR = 0.9 vs. 1.0) (Table 4). Furthermore, CoC and Ox(HXL) PE demonstrated numerically lower revision rates, which, however, were not statistically significant. For 32 mm femoral heads the adjusted analyses showed a higher risk for revision for patients with CoPE bearing surface (HR = 1.3, CI 1.1–1.6) (Table 4). In the 36 mm femoral head subgroup, CoC bearing THAs had a significantly lower hazard ratio compared with MoHXLPE (HR = 0.7 vs. 1.0) (Table 4). The hazard ratios and associated CI for the MoPE and CoPE articulation were not applicable due to small numbers (number of revisions: MoPE 0, CoPE 1). The overall risk of revision with 22–28 mm heads was 18% higher than 32 mm heads (HR = 1.2; CI 1.1–1.3), and 36 mm heads yielded a 11% higher risk over a 32 mm head (HR = 1.1; CI 1.0–1.2) (data not shown in Table). Conventional versus highly-crosslinked-polyethylene Adjusted overall hazard ratios were similar between THAs with highly-crosslinked-polyethylene acetabular components compared with standard PE (Table 8, Supplementary data). However, revisions due to loosening of the acetabular component or liner wear were less frequently observed with HXLPE THAs compared with traditional PE (respectively 10% and 1.8% vs. 17% and 2.7%). Revision due to recurrent dislocation was performed more frequently in THAs with conven-

21-02-2018 15:24:32

Acta Orthopaedica 2018; 89 (2): 163–169

tional PE (35%) versus HXLPE (29%) (Table 9, Supplementary data).

Discussion There is an ongoing interest in alternative bearing surfaces in THA in order to further improve survivorship and reduce the risk of revision surgery. We found a statistically significant benefit in mid-term revision rates for CoHXLPE, CoC, and Ox(HXL)PE bearings compared with a traditional MoPE bearing surface. Furthermore, stratified analyses for small femoral heads (22–28 mm) demonstrated lower revision rates for CoPE and CoHXLPE bearings. For THAs with a 36 mm femoral head, CoC resulted in a lower risk for revision. It has been hypothesized that modern bearing surfaces such as ceramics, oxidized-zirconium, and HXLPE articulations can decrease revision rates compared with traditional MoPE THAs. Historically, aseptic loosening is the most frequent cause of revision in THA (LROI annual report 2015, Norwegian Arthroplasty Register 2016). Over time, wear of the polyethylene liner in a traditional MoPE counterface can generate an adverse local host response, which can result in periprosthetic osteolysis and subsequent aseptic loosening of components (Hu et al. 2015, Varnum et al. 2015). This process is even more relevant in young patients with increased activity demands. Alternative bearing surfaces were introduced in order to reduce PE wear. Ceramic is harder and offers more scratch resistance than cobalt-chrome, which improves lubrication through a low friction coefficient, resulting in excellent wear resistance and low osteolysis rate (Wang et al. 2013, Hu et al. 2015). A meta-analysis of RCTs reporting on the comparison between CoC and MoPE bearing surfaces concluded that CoC resulted in lower revision rates, osteolysis, loosening of components and dislocation, despite more squeaking (Hu et al. 2015). Well-documented drawbacks for ceramic components include high cost and adverse events, such as intra- or postoperative ceramic fractures, and audible squeaking (Hu et al. 2015, Wyles et al. 2015). In the Danish Arthroplasty Registry incidences of ceramic head and liner fractures of respectively 0.28% and 0.17% have been reported (Varnum et al. 2015). Ceramicized metal or oxidized zirconium (Oxinium, Smith & Nephew, Memphis, TN, USA) for femoral heads was developed during the 1980s in an attempt to reduce PE wear. Oxidized-zirconium femoral head components consist of a 5 µm-thick ceramic layer on the metal alloy core, which makes it more resistant to fractures compared with alumina ceramic heads (Jassim et al. 2015). Data from the AOANJRR demonstrated the lowest revision rates for ceramicized-metal-onHXLPE with a 10-year follow-up. The cumulative incidence of revision was 3.2% (2.9–3.7) compared with 6.3% (6.1–6.6) for traditional MoPE bearing after 10 years. However, these results need to be interpreted with caution since the ceram-

11913 Peters D.indd 167

167

icized-metal-HXLPE bearing is a single-company product with a small number of femoral stem and acetabular component combinations, which may have a confounding effect on the outcome (Annual Report AOANJRR 2016). HXLPE was developed to decrease wear in traditional PE liners and subsequently decrease the incidence and severity of osteolysis. Mall et al. (2011) compared the incidence of osteolysis in conventional PE versus HXLPE in young patients (under 50 years of age) undergoing primary THA using radiographs and computed tomography: HXLPE diminished the incidence of osteolysis by 92% compared with conventional PE. The AOANJRR demonstrated that HXLPE had a lower rate of revision compared with non-HXLPE. The difference increased with time and at 15 years the cumulative percentage of revision is 5.6% for HXLPE and 11% for non-HXLPE THAs. Fewer revisions for loosening and dislocation were observed. Other registries, e.g., Kaiser Permanente and NJR, did not report on differences in survival between THAs with conventional and highly-crosslinked PE components, but did also show advantages of ceramics. In the Netherlands, we found a similar overall risk for revision for HXLPE and conventional PE THAs with a short-term follow-up. A similar shift in reasons for revision was observed in the Netherlands. Revisions due to loosening of the acetabular component or liner wear were less frequently observed in HXLPE THAs compared with traditional PE. Revision due to recurrent dislocations was performed more frequently in THAs with conventional PE compared with HXLPE. This can be explained by a preferential use of larger femoral head components in THAs with HXLPE (data not shown). In addition, Jassim et al. (2015) found that the effect of using an HXLPE liner was more important in reducing component wear than the choice of the femoral head bearing (either ceramic or cobalt-chromium). In the Netherlands, revision due to dislocation was more frequently encountered in MoPE THAs (38%) compared with other bearing types, which could be related to a high proportion of small femoral head components (22–28mm) in this group (73%) (Table 5, Supplementary data). Femoral loosening as reason for revision was more frequently registered in CoC (25%) and Ox(HXL)PE (26%) THAs. Theoretically, this could be associated with the large proportion of uncemented THAs in these bearing type groups (respectively 89% and 55%). Periprosthetic fractures which necessitate revision were less common in MoPE (10%), CoPE (10%), and CoC (9%) THAs compared with other bearings. Theoretically, this could be explained by a large proportion of cemented fixations in THAs with MoPE and CoPE bearings. In our dataset, metal-on-metal THAs were excluded. National Arthroplasty Registry data have demonstrated inferior results for large-diameter MoM THAs. The use of these articulations has been associated with wear-related adverse events, e.g., soft tissue inflammatory reactions to metal debris, including inflammatory pseudotumors and aseptic lympho-

21-02-2018 15:24:32

168

cytic vasculitis-associated lesions (Drummond et al. 2015, Nederlandse Orthopaedische Vereniging 2015, Rieker 2017). We performed a detailed analysis in order to assess the influence of bearing surface on survival of the THAs in young (< 60 years), generally more active patients (n = 34,204). We found a statistically significantly lower crude cumulative incidence of revision for advanced bearing surfaces such as CoHXLPE, CoC, and Ox(HXL)PE, over MoPE. However, after adjustment for confounding variables, no statistically significant differences at mid-term follow-up were found. This trend favoring the use of ceramics, HXLPE, and oxidized-zirconium components was consistent with results in patients aged under 55 years in the AOANJRR (Annual Report AOANJRR 2016). We performed further subgroup analyses to assess the influence of bearing type in THAs with different femoral head components. Our results from patients with a small femoral head component demonstrate a reduced risk of revision for CoPE, CoHXLPE, CoC, and Ox(HXL)PE, compared with MoPE after correction for confounding variables. Although this phenomenon was visible for all alternative bearing surfaces, only CoPE and CoHXLPE demonstrated statistically significant differences. In the large femoral head component (36 mm) subgroup, significantly lower revision rates for CoC THAs were determined compared with the MoHXLPE reference bearing surface. Theoretically, the benefits of advanced bearing surfaces with more wear-resistant characteristics would increase with increasing size of the femoral head components since large femoral heads might cause more PE wear and taper corrosion. Respectively, the use of HXLPE and ceramic or oxidizedzirconium heads may presumably lead to less wear and taper corrosion (Ries and Pruitt 2005, Zijlstra et al. 2017). Our study should be interpreted with its limitations in mind. Possible differences in survival are expected to be found in the long term. Our study has limited follow-up with a mean follow-up of 3.9 years and a maximum of 9.9 years. We acknowledge that variation in bearing type may result in possible differences in survival due to wear or loosening of components that will not be detected within our follow-up. Second, national registry studies are based on observational data and therefore cannot infer causality. Furthermore, our data limit the ability to comment on the effect of individual components, which may be an unknown confounder. However, a prosthesis-specific analysis of frequently registered stem components did demonstrate a similar trend of superior results for THA with advanced bearing surfaces. Lastly, comparing different bearing surfaces inherently results in a confounding by indication bias, which cannot be discounted. This phenomenon was also present in our data, but was statistically corrected for by multivariable Cox proportional regression analysis. In summary, based on nationwide arthroplasty registry data, the use of a ceramic-on-highly-crosslinked-polyethylene (CoHXLPE), ceramic-on-ceramic (CoC), and oxidizedzirconium-on-(highly-crosslinked)-polyethylene (Ox(HXL)

11913 Peters D.indd 168

Acta Orthopaedica 2018; 89 (2): 163–169

PE) bearing surfaces resulted in significantly better mid-term survival rates compared with traditional MoPE in the Netherlands. Supplementary data Tables 5–9 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674. 2017.1405669 The authors contributed to: (1) study design and study protocol, (2) gathered data, (3) analyzed data, (4) initial draft, and (5) final draft. RMP and WPZ. contributed to: (1), (2), (3), (4), (5); LNS contributed to: (2), (3), (4), (5); MS contributed to: (1), (2), (3), (4), (5); PCR contributed to: (1), (4), (5); SKB contributed to: (1), (4), (5).

Acta thanks Richard de Steiger and other anonymous reviewers for help with peer review of this study.

Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR). Annual report 2016. Dutch Arthroplasty Register (LROI). Annual report. LROI report, 2015. Drummond J, Tran P, Fary C. Metal-on-metal hip arthroplasty: A review of adverse reactions and patient management. J Funct Biomater 2015; 6 (3): 486-99. Hu D, Tie K, Yang X, Tan Y, Alaidaros M, Chen L. Comparison of ceramicon-ceramic to metal-on-polyethylene bearing surfaces in total hip arthroplasty: A meta-analysis of randomized controlled trials. J Orthop Surg Res 2015; 10: 22. Jämsen E, Eskelinen A, Peltola M, Mäkelä K. High early failure rate after cementless hip replacement in the octogenarian. Clin Orthop Rel Res 2014; 472 (9): 2779-89. Jassim S S, Patel S, Wardle N, Tahmassebi J, Middleton R, Shardlow D L, Stephen A, Hutchinson J, Haddad F S. Five-year comparison of wear using oxidised zirconium and cobalt-chrome femoral heads in total hip arthroplasty: A multicentre randomised controlled trial. Bone Joint J 2015; 97-B (7): 883-9. Keurentjes J C, Fiocco M, Schreurs B W, Pijls B G, Nouta K A, Nelissen R G H H. Revision surgery is overestimated in hip replacement. Bone Joint Res 2012; 1 (10): 258-62. 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 Relat Res 2015; 473 (11): 3431-42. Mall N A, Nunley R M, Zhu J J, Maloney W J, Barrack R L, Clohisy J C. The incidence of acetabular osteolysis in young patients with conventional versus highly crosslinked polyethylene. Clin Orthop Relat Res 2011; 469 (2): 372-81. Mihalko W M, Wimmer M A, Pacione C A, Laurent M P, Murphy R F, Rider C. How have alternative bearings and modularity affected revision rates in total hip arthroplasty? Clin Orthop Relat Res 2014; 472 (12): 3747-58. Nederlandse Orthopaedisch Vereniging. Advies Metaal-op-Metaal Heupprothesen per 1 augustus 2015. Norwegian Arthroplasty Register. Norwegian National Advisory Unit on Arthroplasty and Hip Fractures (NNAUoAaHF). Annual Report, 2016. Peters R M, van Steenbergen L N, Bulstra S K, Zeegers A V C M, Stewart R E, Poolman R W, Hosman A H. Nationwide review of mixed and nonmixed components from different manufacturers in total hip arthroplasty. Acta Orthop 2016; 87 (4): 356-62. Putter H, Fiocco M, Geskus R B. Tutorial in biostatistics: Competing risks and multi-state models. Stat Med 2007; 26 (11): 2389-430.

21-02-2018 15:24:32

Acta Orthopaedica 2018; 89 (2): 163â&#x20AC;&#x201C;169

Rieker C B. Tribology of total hip arthroplasty prostheses: What an orthopaedic surgeon should know. EFORT Open Rev 2017; 1 (2): 52-7. Ries M D, Pruitt I. Effect of cross-linking on the microstructure and mechanical properties of ultra-high molecular weight polyethylene. Clin Orthop Relat Res 2005; 440: 149-56. Varnum C, Pedersen A B, Kjaersgaard-Andersen P, Overgaard S. Comparison of the risk of revision in cementless total hip arthroplasty with ceramic-onceramic and metal-on-polyethylene bearings. Acta Orthop 2015; 86 (4): 477-84. 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. Wang S, Zhang S, Zhao Y. A comparison of polyethylene wear between cobalt-chrome ball heads and alumina ball heads after total hip arthroplasty: A 10-year follow-up. J Orthop Surg Res 2013; 8:20.

11913 Peters D.indd 169

169

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 Relat Res 2015; 473 (4): 1173-6. Wyles C C, Jimenez-Almonte J H, Murad M H, Murad M H, NorambuenaMorales G A, Cabanela M E, Sierra R J, Trousdale R T. There are no differences in short- to mid-term survivorship among total hip-bearing surface options: A network meta-analysis. Clin Orthop Relat Res 2015;473 (6): 2031-41. Yin S, Zhang D, Du H, Du H, Yin Z, Qiu Y. Is there any difference in survivorship of total hip arthroplasty with different bearing surfaces? A systematic review and network meta-analysis. Int J Clin Exp Med 2015; 8 (11): 21871-85. Zijlstra W P, De Hartog B, Van Steenbergen L N, Scheurs B W, Nelissen R G H H. Effect of femoral head size and surgical approach on risk of revision for dislocation after total hip arthroplasty, Acta Orthop 2017; 88 (4): 392-401.

21-02-2018 15:24:32

170

Acta Orthopaedica 2018; 89 (2): 170–176

Temporal trends in hip fracture incidence, mortality, and morbidity in Denmark from 1999 to 2012 Christopher JANTZEN 1, Christian M MADSEN 1, Jes B LAURITZEN 1, and Henrik L JØRGENSEN 2

1 Department of Orthopaedic Surgery, Bispebjerg Hospital, University of Copenhagen; 2 Department of Clinical Biochemistry, Bispebjerg Hospital, University of Copenhagen, Denmark Correspondence: Christopherjantzen@gmail.com Submitted 2017-11-01. Accepted 2017-12-09

Background and purpose — While development in hip fracture incidence and mortality is well examined, none has yet looked at the temporal trends regarding prevalence of co-morbidities. Therefore we investigated changes in incidence of first hip fracture, co-morbidity prevalence, 30 day- and 1-year mortality in hip fracture patients in the Danish population during the period 1999 to 2012. Patients and methods — Patients > 18 years admitted with a fractured hip in Denmark between 1996 and 2012 were identified with data for the period 1999–2012 being analyzed regarding prevalence of co-morbidities, incidence, and mortality. Results — 122,923 patients were identified. Incidence in the whole population declined but sex-specific analysis showed no changes for men. For the whole study population, 30-day and 1-year mortality remained unchanged. Age at time of first hip fracture also remained unchanged. Of the included co-morbidities a decrease in prevalence of malignancy and dementia in women was found while there was an increase in the prevalence of all remaining co-morbidities, except hemi- or paraplegia for both sexes, rheumatic diseases for women, and for men diabetes with complications, myocardial infarction, AIDS/HIV, and malignancy. Interpretation — While hip fracture incidence declined for women it was unchanged for men; likewise, 30-day and 1-year mortality rates together with age at first fracture remained unchanged. When these results are compared with the relatively large increase in the prevalence of co-morbidities, it does not seem likely that the increased disease burden is affecting either the incidence or the mortality. ■

Patients sustaining a hip fracture are known to have increased mortality and morbidity compared with the general population and as the population gets older an increased incidence of hip fractures could be expected (Kanis 1993). Studies on the development in hip fracture incidence have found diverging results with some reporting decreasing incidence (Kannus et al. 2006, Abrahamsen and Vestergaard 2010, Lippuner et al. 2011, Nilson et al. 2013, Jean et al. 2013, Korhonen et al. 2013) and others increasing incidence (Mann et al. 2008). Also, hip fracture patients have an excess mortality (Haentjens et al. 2010) with risk factors including co-morbidity (Barone et al. 2009). While the development in hip fracture incidence and mortality has already been examined, few studies have looked at the development in co-morbidity and it is unclear whether hip fracture incidence and mortality change with the prevalence of co-morbidities. The purpose of this descriptive study was to investigate the changes in incidence of first hip fracture, co-morbidity, 30-day and 1-year mortality in hip fracture patients in the Danish population during the period 1999 to 2012.

Patients and methods Study population Data were collected from the Danish National Patient Registry (DNPR) on all patients above 18 years, sustaining a hip fracture (ICD-10 codes: DS720, DS721 or DS722) during the period January 1, 1996 to December 31, 2012. For patients sustaining more than 1 hip fracture during the period, only the first was included and used as index fracture for survival analysis. 154,062 patients sustained a hip fractures during

11562 Jantzen D.indd 170

21-02-2018 17:07:58

Acta Orthopaedica 2018; 89 (2): 170–176

this period. To make sure that only the first hip fracture was included and thus reducing the risk of finding a false decrease in hip fracture incidence and co-morbidities, data for the first 3 years were excluded as a washout period. As such, only data from the period 1999 to 2012 were analyzed resulting in the inclusion in the study of 122,923 patients with a hip fracture. Data collected on the individual patient included age, sex, fracture type, co-morbidity, and time of death. The co-morbidities found in the Charlson Co-morbidity Index (CCI) were included in the study with data on co-morbidities registered prior to the fracture being retrieved from the DNPR and coded as described by Quan et al. (2005). The following co-morbidities were thus included: myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, dementia, chronic pulmonary disease, rheumatic disease, peptic ulcer disease, moderate or severe liver disease, mild liver disease, diabetes without chronic complication, diabetes with chronic complication, hemiplegia or paraplegia, metastatic solid tumor, any malignancy, AIDS/HIV, and renal disease. National patient registry data In Denmark, all citizens have a unique, non-reusable, 10-digit civil registration number (CRN) assigned by the Civil Registration System (CRS). The CRS contains demographic information on all citizens residing legally in Denmark including vital status and emigration. The CRN is used in all public records making it possible to link different information from national registries for the unique individual. This provides excellent traceability and allows almost complete follow-up (Schmidt et al. 2014). Through the use of the CRN, all contacts and admissions to Danish hospitals are registered in the DNPR which contains information on all non-psychiatric hospital admissions dating back to 1977 and has since 1995 also included data on outpatient visits and psychiatric admissions (Lynge et al. 2011). For this study, data on discharge diagnosis or secondary diagnostic codes were used in the form of International Disease Classification 10 (ICD-10) codes. When using data from registries 2 important measures of quality are validity and completeness, which vary with the different diagnosis codes (Schmidt et al. 2015). To our knowledge no completeness study has been performed on hip fractures reported to the DNPR, but since data completeness depends on hospitalization patterns and diagnostic accuracy, one would expect that conditions such as hip fracture, which should always lead to a hospital encounter, are registered consistently in the DNPR with a high level of completeness. Data on the Danish population size for the different years during the period 1999–2012 were obtained from http://www. dst.dk/da/Statistik/statistikbanken. Statistics All included continuous variables were non-normally distributed and were thus analyzed using Mann–Whitney U-tests.

11562 Jantzen D.indd 171

171

The Cochran–Armitage trend test was used to test for trends in development in prevalence of co-morbidities. The development in incidence and mortality over time was assessed using negative binomial regression analysis since overdispersion was found when performing Poisson regression analysis, with patients sustaining a fracture being removed from the population at risk for the purpose of calculating incidence rates. Person-time was calculated based on the number of individuals at risk on January 1 each year in Denmark assuming all were followed for the subsequent year. P-values < 0.05 were considered significant. Statistical analyses were conducted using SAS version 9.4 (SAS Institute, Cary, NC, USA). Ethics, funding, and potential conflicts of interest According to Danish law, ethical committee approval is not required for this type of observational database study. The data were obtained through secure remote access to Statistics Denmark (ref. 704670). The study was approved by the Danish Data Protection Agency. No funding was received for this study. No competing interest was declared for any of the authors.

Results 122,923 patients sustained their first hip fracture during the study period (Table 1). From 1999 to 2012, the hip fracture incidence in the whole population declined from 182 to 137 per 100,000 (p < 0.001) but gender-specific analysis showed that a decrease was evident only for women (256 to 181 per 100,000, p < 0.001) since the changes for men were not statistically significant (107 to 93 per 100,000, p = 0.09) (Figure 1). At the same time, the median age at time of first hip fracture remained largely unchanged (Figure 2). For the whole population, 30-day together with 1-year mortality remained unchanged (9.7% to 10.3%, p = 0.9 and 28.2% to 27.4%, p = 0.4, respectively) with sex-specific analysis showing the same trends for men (13.7% to 12.5%, p = 0.5 and 33.5% to 31.3%, p = 0.5) and women (8.1% to 9.1%, p = 0.9 and 26.0% to 25.5% p = 0.2) (Figures 3 and 4). Of the included co-morbidities a trend with a decrease in prevalence of malignancy and dementia in women was found while an increase in the prevalence of all remaining co-morbidities, except hemi- or paraplegia for both genders, rheumatic diseases for women, and for men diabetes with complications, myocardial infarction, AIDS/HIV, and malignancy was found, with the largest increase in prevalence seen for congestive heart failure, moderate to severe liver disease, and renal disease, with the largest increments found for congestive heart failure (men: 6.5% to 10.7%, women: 5.9% to 13.1%), moderate to severe liver disease (men: 0.1% to 0.7%, women: 0.1% to 0.4%) and renal disease (men: 0.4% to 2.0%, women: 0.3% to 1.1%) (Table 2, see Supplementary data).

21-02-2018 17:07:58

172

Acta Orthopaedica 2018; 89 (2): 170–176

Table 1. Basic characteristics. Values are frequency (%) unless otherwise stated Alive after 30 days Dead within 30 days Number 110,702 (90.1) Male 32,968 (29.8) Female 77,734 (70.2) Median age (range) 80 (18–107) Median Charlson-score (range) 1 (0–15) Myocardial infarction 2,451 (2.2) Congestive heart failure 10,229 (9.2) Peripheral vascular disease 2,648 (2.4) Cerebrovascular disease 7,114 (6.4) Dementia 4,381 (4.0) Chronic pulmonary disease 3,964 (3.6) Rheumatic disease 1,364 (1.2) Peptic ulcer disease 3,778 (3.4) Mild liver disease 705 (0.6) Diabetes without chronic complication 1,065 (1.0) Diabetes with chronic complication 2,283 (2.1) Hemiplegia or paraplegia 300 (0.3) Renal disease 746 (0.7) Any malignancy 2,862 (2.6) Moderate or severe liver disease 252 (0.2) Metastatic solid tumor 665 (0.6) AIDS/HIV 14 (0.0)

12,221 (9.9) 5,175 (42.4) 7,046 (57.6) 86 (18–109) 1 (0–19) 530 (4.3) 1,490 (12.2) 412 (3.4) 1,044 (8.5) 849 (7.0) 635 (5.2) 174 (1.4) 568 (4.7) 82 (0.7) 148 (1.2) 290 (2.4) 17 (0.1) 204 (1.7) 475 (3.9) 32 (0.3) 166 (1.4) 0 (0.0)

p-value

Alive after 1 year Dead within 1 year

–

88,541 (72.0) 25,497 (28.8) 63,044 (71.2) 79 (18–107) 0 (0–16) 1,722 (1.9) 7,772 (8.8) 1,978 (2.2) 5,369 (6.1) 2,896 (3.3) 2,973 (3.4) 1,099 (1.2) 2,832 (3.2) 575 (0.7) 815 (0.9) 1,773 (2.0) 250 (0.3) 483 (0.6) 2,074 (2.3) 198 (0.2) 293 (0.3) 13 (0.0)

< 0.001 < 0.001 < 0.001 < 0.001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 0.4 < 0.0001 0.7 0.008 0.02 0.006 < 0.0001 < 0.0001 0.5 < 0.0001 0.2

34,382 (28.0) 12,646 (36.8) 21,736 (63.2) 84 (18–111) 1 (0–19) 1,259 (3.7) 3,947 (11.5) 1,082 (3.2) 2,789 (8.1) 2,334 (6.8) 1,626 (4.7) 439 (1.3) 1,514 (4.4) 212 (0.6) 398 (1.2) 800 (2.3) 67 (0.2) 467 (1.4) 1,263 (3.7) 86 (0.3) 528 (1.5) 1 (0.0)

p-value – < 0.001 < 0.001 < 0.001 < 0.001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 0.6 < 0.0001 0.5 0.0002 0.0004 0.007 < 0.0001 < 0.0001 0.4 < 0.0001 0.08

Hip fracture incidence per 100,000 – Overall

Hip fracture incidence per 100,000 – Men

Hip fracture incidence per 100,000 – Women

300

280

260

240

220

200

180

160

140

120

100

0 2000

2002

2004

2006

2008

2010

2012

2000

2002

2004

2006

2008

2010

2012

0 2000

2002

2004

2006

2008

2010

2012

Figure 1. Overall and sex-specific hip fracture incidence. O: fitted values. X: observed values. P-values and 95% confidence intervals are derived from negative binomial regression. Overall p < 0.0001, men p = 0.09, and women p < 0.0001,

Overall

Men

Women

Median age and interquartile range

2000

2002

2004

2006

2008

2010

2012

2000

2002

2004

2006

2008

2010

2012

2000

2002

2004

2006

2008

2010

2012

Figure 2. Overall and sex-specific median age and interquartile range at time of first hip fracture.

11562 Jantzen D.indd 172

21-02-2018 17:07:58

Acta Orthopaedica 2018; 89 (2): 170–176

173

30-day mortality (%) – Overall

30-day mortality (%) – Men

30-day mortality (%) – Women

0 2000

2002

2004

2006

2008

2010

2012

0 2000

2002

2004

2006

2008

2010

2012

2000

2002

2004

2006

2008

2010

2012

Figure 3. Overall and sex-specific 30-day mortality. O: fitted values. X: observed values. P-values and 95% confidence intervals are derived from negative binomial regression. Overall p = 0.9, men p = 0.5, and women p = 0.9.

1-year mortality (%) – Overall

1-year mortality (%) – Men

1-year mortality (%) – Women

0 2000

2002

2004

2006

2008

2010

2012

0 2000

2002

2004

2006

2008

2010

2012

2000

2002

2004

2006

2008

2010

2012

Figure 4. Overall and sex-specific 1-year mortality. O: fitted values. X: observed values. P-values and 95% confidence intervals are derived from negative binomial regression. Overall p = 0.4, men p = 0.5, and women p = 0.2.

Discussion Hip fractures represent the most serious complication of osteoporosis in terms of morbidity, mortality, disability, and medical costs (Melton III 1993) and in 1993 it was estimated that the incidence of hip fractures would increase rapidly towards 2016 (Kanis 1993). This hypothesis was further supported by another study in which it was projected that the increasing size of the elderly population would lead to a 50% increase in the number and cost of osteoporotic fractures by 2025 (Burge et al. 2007). While this tendency towards an increased incidence is supported by data from Austria (Mann et al. 2008), other studies from Denmark, Sweden, Canada, Switzerland, and Finland have shown a decline (Kannus et al. 2006, Abrahamsen and Vestergaard 2010, Lippuner et al. 2011, Jean et al. 2013, Korhonen et al. 2013, Nilson et al. 2013). In line with these studies, we found that the incidence of hip

11562 Jantzen D.indd 173

fractures during the period 1999 to 2012 declined in the general population by 25%. Earlier data have shown that hip fractures occur more commonly in women than in men (Seeman 1995). This difference is also supported by our results showing a higher incidence of hip fractures for women during the whole period and even though the reduction in incidence was higher for women than for men the incidence was still twice as high in 2012 (181 vs. 93 per 100,000). Different explanations for the declining incidence have been proposed and include more widespread preventive measures, diagnosis, and treatment of osteoporosis (Jean et al. 2013) which in part could explain why the incidence in the Austrian population was found to increase since the lack of a structured nationwide osteoporosis prevention approach has been reported (Mann et al. 2008). While anti-osteoporotic treatment has been proposed as one of the major contributors to the declining inci-

21-02-2018 17:07:59

174

dence, it has also been suggested that the decrease is too large to be explained by the extent of anti-osteoporotic treatment with other potential explanations being changes in smoking habits, obesity, improved general health, and vitamin D supplements (Abrahamsen and Vestergaard 2010). But even though a decreasing incidence has been found during the last decade in Denmark it has been proposed in a recent study by Rosengren et al. (2017), in which age-period-cohort effects for hip fractures were analyzed, that increasing fracture rates would be expected due to aging of more recent birth cohorts with a higher relative risk of sustaining a hip fracture. Hip fracture patients constitute one of the most vulnerable groups of patients with a high level of co-morbidity, with earlier data showing that 90% of male hip fracture patients suffered from some sort of chronic disease with two-thirds having a disease affecting sensory or motor function and onetenth having chronic alcoholism (Huuskonen et al. 1999). Likewise, different co-morbidities have been found to influence the risk of sustaining a hip fracture. It has thus been shown that patients suffering from systemic lupus erythematosus (Wang et al. 2013), Parkinsonâ&#x20AC;&#x2122;s disease (Pouwels et al. 2013), dementia (Jorgensen et al. 2014), former stroke (Pouwels et al. 2009), heart failure, peripheral atherosclerosis, ischemic heart disease (Sennerby et al. 2009), hemodialysis, liver cirrhosis, prior fracture, and osteoporosis all have an increased risk of sustaining a hip fracture (Lin et al. 2014). For COPD and asthma diverging results have been found with some reporting an increased risk of sustaining a fracture (Vestergaard et al. 2007) while others have not (Dam et al. 2010). The same applies to diabetes, where one study found an increased risk for patients suffering from type-2 diabetes (Schwartz et al. 2011) while another study found an increased risk only in patients with type-1 diabetes (Hothersall et al. 2014). Also, it has been found that patients who sustained a hip fracture were more likely to be women, living in long-term institutional care, using neuroleptics, dependent in activities of daily living (ADL), with a history of previous stroke with hemiparesis, Parkinsonism, and/or lower BMI than those who did not sustain a fracture after a fall on the hip (Willig et al. 2003). Since different co-morbidities have been found to increase the risk of sustaining a hip fracture, a general reduction in the disease burden in the population could be hypothesized as one of the reasons for the reduced incidence (Jorgensen et al. 2014) but, contrary to this, our results on the temporal development of the 17 included co-morbidities showed a trend with a decrease in prevalence of malignancy and dementia in women, while there was an increase in the prevalence of all remaining co-morbidities, except hemi- or paraplegia for both sexes, rheumatic diseases for women, and for men diabetes with complications, myocardial infarction, AIDS/HIV, and malignancy, with the largest increase in prevalence seen for congestive heart failure, moderate to severe liver disease, and renal disease. Of the included co-morbidities, the most preva-

11562 Jantzen D.indd 174

Acta Orthopaedica 2018; 89 (2): 170â&#x20AC;&#x201C;176

lent in 1999 were congestive heart failure (men 6.5%, women 5.9%), cerebrovascular disease (men 5.7%, women 4.8%), dementia (men 3.4%, women 3.2%), and chronic pulmonary disease (men 4.2%, women 2.5%) and during the period these co-morbidities were also subjected to the largest increments resulting in these also being the most prevalent in 2012 (congestive heart failure: men 10.7%, women 13.1%, cerebrovascular disease: men 7.4%, women 6.9%, dementia: men 4.0%, women 3.9%, and chronic pulmonary disease: men 5.1%, women 3.8%). While an increased disease burden would be expected to lead to an increase in incidence of first hip fracture, this was not evident in our study. Reasons for the lack of increase in the incidence despite an increased disease burden could be better treatment of the individual diseases or more intensive diagnostic measures resulting in earlier detection and treatment of the diseases and thus less impact on the individual patient. A potential reason for the higher disease rate we found could be increased age at the time of the first hip fracture but the median age of patients sustaining a hip fracture remained almost unchanged at 81 for the whole population, 82 for women, and between 76 and 78 for men. Therefore, changes in age at the time of fracture do not seem to be a probable explanation for the changes in co-morbidity, which instead could be a consequence of more extensive diagnostic measures in the hospital setting in the years prior to the fracture. During the study period, 30-day and 1-year mortality for the whole study population and for both sexes remained unchanged with no significant changes. At the same time a higher mortality was evident for men at both 30 days and 1 year. This difference in mortality has been shown earlier in several studies (Kellie and Brody 1990, Myers et al. 1991, Jacobsen et al. 1992). This is further supported by a metaanalysis showing an excess mortality in both men and women after a hip fracture with higher mortality in men compared with women at all ages (Haentjens et al. 2010). The effect of co-morbidities on 30-day mortality has previously been investigated and an increased risk of early death was found in patients suffering from central nervous system diseases (dementia, Parkinsonâ&#x20AC;&#x2122;s, hemiplegia), diabetes, circulatory disorders, nutritional deficiencies, COPD, chronic renal disease, and other chronic diseases (liver, pancreas, intestine) (Barone et al. 2009). We also found a significant association between low socioeconomic status and the risk of early death. As such, one would expect a concomitant increase in mortality rates with the increase in prevalence of the co-morbidities increases. One explanation for why this phenomenon was not observed in our study could be better and more effective treatment of the different co-morbidities and/or better and faster management of hip fracture leading to a smaller effect on the mortality rates. A limitation of this study is that it is only concerned with the first hip fracture and as such data on incidence represent this and not the total incidence during the period, which

21-02-2018 17:08:01

Acta Orthopaedica 2018; 89 (2): 170–176

would be expected to be higher and, even though a washout period of 3 years was included, there is a small risk that the fracture included is not the first but the second fracture or a complication from an earlier fracture. Also, while a decrease in incidence was found during the study period it is important to remember that the data shown are based on ICD-10 codes given on discharge and since no completeness study has been performed it is possible that the incidences shown could be influenced by lacking or inappropriate coding. Another limitation includes data on co-morbidity, which are based on ICD-codes given to the patient on discharge. As such, a reason for the increase could be more focus on allocating codes to the patients on discharge at the end of the study period, but against this are results from a recent Danish study (Jørgensen et al. 2015) on patients undergoing first-time coronary angiography during the period 2000 to 2009. In this study a general tendency towards lower prevalence of co-morbidities was found and when these results are compared with ours it seems likely that the increase in disease burden found in our study is real and not based on a change in coding practice. In addition, a correlation between time of inclusion and length of window of registration for co-morbidities is present in the study with patients included later having a longer window. While this would lead to a hypothetical increased risk of accumulating co-morbidities for the patients included towards the end of the study period we are sure that any substantial co-morbidities would lead to contact with the healthcare system within a 3-year period and as such the risk would be minimally increased due to the effect of the initial 3-year data washout period in the study. Also, since our data do not contain information on changes in mortality and disease burden in the background population, it is not possible to conclude whether or not the changes are isolated to the hip fracture population or are a result of changes in the background population. The strengths of the study include the large number of patients in the study, covering all Danish hip fracture patients, and the use of the Danish national registries, which ensure that data are collected unbiasedly. As the study covers an entire population, it also increases the generalizability of the results and it is likely that the trends observed in the Danish population match those of other Western countries. In summary, this observational study shows that during the period 1999 to 2012 the incidence of first hip fracture has declined for women but remained unchanged for men. For 30-day mortality, rates have increased for women even though there are no changes in 1-year mortality. On the contrary, 30-day mortality for men is unchanged while 1-year mortality has declined. When these results are compared with the relatively large increase in the prevalence of the different co-morbidities it does not seem likely that the increased disease burden is affecting the incidence or the mortality.

11562 Jantzen D.indd 175

175

Supplementary data Table 2 is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674. 2018.1428436 CJ: Statistical analyses, wrote the paper. CMM, JBL: Critical review of the manuscript. HLJ: Statistical analyses, critical review of the manuscript. Acta thanks Jan-Erik Gjertsen and Björn Erik Rosengren for help with peer review of this study.

Abrahamsen B, Vestergaard P. Declining incidence of hip fractures and the extent of use of anti-osteoporotic therapy in Denmark 1997–2006. Osteoporos Int 2010; 21(3): 373-80. Barone A P, Fusco D, Colais P, D’Ovidio M, Belleudi V, Agabiti N, Sorge C, Davoli M, Perucci C A. Effects of socioeconomic position on 30-day mortality and wait for surgery after hip fracture. Int J Qual Health Care 2009; 21(6): 379-86. Burge R, Dawson-Hughes B, Solomon DH, Wong JB, King A, Tosteson A. Incidence and economic burden of osteoporosis-related fractures in the United States, 2005–2025. J Bone Miner Res 2007; 22(3): 465-75. Dam T T, Harrison S, Fink H A, Ramsdell J, Barrett-Connor E. Bone mineral density and fractures in older men with chronic obstructive pulmonary disease or asthma. Osteoporos Int 2010; 21(8): 1341-9. Haentjens P, Magaziner J, Colon-Emeric C S, Vanderschueren D, Milisen K, Velkeniers B, Boonen S. Meta-analysis: excess mortality after hip fracture among older women and men. Ann Intern Med 2010; 152(6): 380-90. Hothersall E J, Livingstone S J, Looker H C, Ahmed S F, Cleland S, Leese G P, Lindsay R S, McKnight J, Pearson D, Philip S, Wild S H, Colhoun H M. Contemporary risk of hip fracture in type 1 and type 2 diabetes: a national registry study from Scotland. J Bone Miner Res 2014; 29(5): 1054-60. Huuskonen J, Kroger H, Arnala I, Alhava E. Characteristics of male hip fracture patients. Ann Chir Gynaecol 1999; 88(1): 48-53. Jacobsen S J, Goldberg J, Miles T P, Brody J A, Stiers W, Rimm A A. Race and sex differences in mortality following fracture of the hip. Am J Public Health 1992; 82(8): 1147-50. Jean S, O’Donnell S, Lagace C, Walsh P, Bancej C, Brown J P, Morin S, Papaioannou A, Jaglal S B, Leslie W D. Trends in hip fracture rates in Canada: an age-period-cohort analysis. J Bone Miner Res 2013; 28(6): 1283-89. Jørgensen M E, Andersson C, Olsen A M, Juel K, Mortensen P E, Jørgensen E, Tilsted H H, von Kappelgaard L M, Torp-Pedersen C, Gislason G H. Danish trends in pharmacotherapy, comorbidities, and demographics in patients referred for coronary angiography: what changed during a decade? Eur Heart J Cardiovasc Pharmacother 2015; 1(3): 157-65. Jorgensen T S, Hansen AH, Sahlberg M, Gislason G H, Torp-Pedersen C, Andersson C, Holm E. Falls and comorbidity: the pathway to fractures. Scand J Public Health 2014; 42(3): 287-94. Kanis J A. The incidence of hip fracture in Europe. Osteoporos Int 1993; 3 Suppl 1: 10-5. Kannus P, Niemi S, Parkkari J, Palvanen M, Vuori I, Jarvinen M. Nationwide decline in incidence of hip fracture. J Bone Miner Res 2006; 21(12): 18368. Kellie S E, Brody J A. Sex-specific and race-specific hip fracture rates. Am J Public Health 1990; 80(3): 326-8. Korhonen N, Niemi S, Parkkari J, Sievanen H, Palvanen M, Kannus P. Continuous decline in incidence of hip fracture: nationwide statistics from Finland between 1970 and 2010. Osteoporos Int 2013; 24(5): 1599-1603. Lin Z Z, Wang J J, Chung C R, Huang P C, Su B A, Cheng K C, Chio C C, Chien C C. Epidemiology and mortality of hip fracture among patients on dialysis: Taiwan National Cohort Study. Bone 2014; 64: 235-9.

21-02-2018 17:08:01

176

Lippuner K, Popp A W, Schwab P, Gitlin M, Schaufler T, Senn C, Perrelet R. Fracture hospitalizations between years 2000 and 2007 in Switzerland: a trend analysis. Osteoporos Int 2011; 22(9): 2487-97. Lynge E, Sandegaard J L, Rebolj M. The Danish National Patient Register. Scand J Public Health 2011; 39(7 Suppl): 30-3. Mann E, Icks A, Haastert B, Meyer G. Hip fracture incidence in the elderly in Austria: an epidemiological study covering the years 1994 to 2006. BMC Geriatr 2008; 8: 35. Melton L J, III. Hip fractures: a worldwide problem today and tomorrow. Bone 1993; 14 Suppl 1: S1-S8. Myers A H, Robinson E G, Van Natta M L, Michelson J D, Collins K, Baker S P. Hip fractures among the elderly: factors associated with in-hospital mortality. Am J Epidemiol 1991; 134(10): 1128-37. Nilson F, Moniruzzaman S, Gustavsson J, Andersson R. Trends in hip fracture incidence rates among the elderly in Sweden 1987-2009. J Public Health (Oxf) 2013; 35(1): 125-31. Pouwels S, Lalmohamed A, Leufkens B, de Boer A, Cooper C, van ST, de Vries F. Risk of hip/femur fracture after stroke: a population-based casecontrol study. Stroke 2009; 40(10): 3281-5. Pouwels S, Bazelier M T, de Boer A, Weber W E, Neef C, Cooper C, de Vries F. Risk of fracture in patients with Parkinsonâ&#x20AC;&#x2122;s disease. Osteoporos Int 2013; 24(8): 2283-90. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi J C, Saunders L D, Beck C A, Feasby T E, Ghali W A. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005; 43(11): 1130-9.

11562 Jantzen D.indd 176

Acta Orthopaedica 2018; 89 (2): 170â&#x20AC;&#x201C;176

Rosengren B E, Bjork J, Cooper C, Abrahamsen B. Recent hip fracture trends in Sweden and Denmark with age-period-cohort effects. Osteoporos Int 2017; 28(1): 139-49. Schmidt M, Pedersen L, Sorensen H T. The Danish Civil Registration System as a tool in epidemiology. Eur J Epidemiol 2014; 29(8): 541-9. Schmidt M, Schmidt S A, Sandegaard J L, Ehrenstein V, Pedersen L, Sorensen H T. The Danish National Patient Registry: a review of content, data quality, and research potential. Clin Epidemiol 2015; 7: 449-90. Schwartz AV , Vittinghoff E, Bauer D C, Hillier T A, Strotmeyer E S, Ensrud K E, Donaldson M G, Cauley J A, Harris T B, Koster A, Womack C R, Palermo L, Black D M. Association of BMD and FRAX score with risk of fracture in older adults with type 2 diabetes. JAMA 2011; 305(21): 218492. Seeman E. The dilemma of osteoporosis in men. Am J Med 1995; 98(2A): 76S-88S. Sennerby U, Melhus H, Gedeborg R, Byberg L, Garmo H, Ahlbom A, Pedersen N L, Michaelsson K. Cardiovascular diseases and risk of hip fracture. JAMA 2009; 302(15): 1666-73. Vestergaard P, Rejnmark L, Mosekilde L. Fracture risk in patients with chronic lung diseases treated with bronchodilator drugs and inhaled and oral corticosteroids. Chest 2007; 132(5): 1599-1607. Wang S H, Chang Y S, Liu C J, Lai C C, Chen W S, Chen T J, Wang S J. Association of systemic lupus erythematosus with a higher risk of cervical but not trochanteric hip fracture: a nationwide population-based study. Arthritis Care Res (Hoboken) 2013; 65(10): 1674-81. Willig R, Luukinen H, Jalovaara P. Factors related to occurrence of hip fracture during a fall on the hip. Public Health 2003; 117(1): 25-30.

21-02-2018 17:08:01

Acta Orthopaedica 2018; 89 (2): 177–183

177

High and rising burden of hip and knee osteoarthritis in the Nordic region, 1990–2015 Findings from the Global Burden of Disease Study 2015 Aliasghar A KIADALIRI 1, L Stefan LOHMANDER 1, Maziar MORADI-LAKEH 2, Ingemar F PETERSSON 1,3, and Martin ENGLUND 1,4

1 Lund University, Faculty of Medicine, Department of Clinical Sciences-Lund, Orthopedics, Clinical Epidemiology Unit, Lund, Sweden; 2 Preventive Medicine and Public Health Research Center, Department of Community Medicine, Iran University of Medical Sciences, Tehran, Iran; 3 Epidemiology and Register Centre South, Skåne University Hospital Lund, Lund, Sweden; 4 Clinical Epidemiology Research and Training Unit, Boston University School of Medicine, Boston, MA, USA Correspondence: aliasghar.ahmad_kiadaliri@med.lu.se Submitted 2017-05-17. Accepted 2017-10-30.

Background and purpose — Osteoarthritis (OA) imposes a substantial burden on individuals and societies. We report on the burden of knee and hip OA in the Nordic region. Patients and methods — We used the findings from the 2015 Global Burden of Diseases Study to explore prevalence, years lived with disability (YLDs), and disability-adjusted life-years (DALYs) of OA in the 6 Nordic countries during 1990–2015 (population of about 27 million in 2015). Results — During 1990–2015, the number of prevalent OA cases increased by 43% to 1,507,587 (95% uncertainty interval [UI] 1,454,338–1,564,778) in the region. OA accounted for 1.3% (UI 1.0–1.7) of YLDs in 1990, increasing to 1.6% (UI 1.2–2.0) in 2015. Of 315 causes studied, OA was the 15th leading cause of YLDs, causing 52,661 (UI 34,056–77,499) YLDs in 2015; of these 23% were attributable to high body mass index. The highest relative importance of OA was reported for women aged 65–74 years (8th leading cause of YLDs in 2015). Among the top 30 leading causes of YLDs in the region, OA had the 5th greatest relative increase in total YLDs during 1990–2015. From 1990 to 2015, increase in age-standardized YLDs from OA in the region was slightly lower than increase at the global level (7.5% vs. 10.5%). OA was, however, responsible for a higher proportional burden of DALYs in the region compared with the global level. Interpretation — The OA burden is high and rising in the Nordic region. With population growth, aging, and the obesity epidemic, a substantial rise in the burden of OA is expected and should be addressed in health policies. ■

Painful osteoarthritis (OA) of the peripheral joints is often associated with physical disability and deterioration in healthrelated quality of life, translating into a substantial burden on individuals and societies (Hiligsmann and Reginster 2013, Cross et al. 2014, Kiadaliri et al. 2016). Although OA may affect all joints, the knee, hip, and hand are most commonly affected. OA incidence and prevalence has increased over recent decades and this is generally attributed to aging of the population and rising prevalence of obesity (Nguyen et al. 2011, Neogi and Zhang 2013, Rahman et al. 2014, Yu et al. 2015). In addition, occurrence of OA among younger active people has been reported to be increasing (Leskinen et al. 2012, Yu 2015). The rising incidence and prevalence of OA implies that its burden will increase and impose significant pressure on healthcare systems worldwide. An up-to-date and accurate estimate of the burden of OA and its burden in relation to other diseases can aid informed decision-making by health authorities. In addition, monitoring and comparison of the country-specific burden would provide useful insights to assess health systems’ performance and benchmark a country against others. However, to do so, consistent and comparable data across diseases and geographies over times are required. The Global Burden of Disease (GBD) Study (Murray and Lopez 1996) aims to respond to this requirement by estimating comprehensive and internally consistent estimates of mortality and disability from major diseases, injuries, and risk factors. In the latest iteration of the GBD study, GBD 2015, the burden of 315 causes including OA has been estimated for 195 countries during 1990–2015 (GBD 2015 DALYs and HALE Collaborators 2016, GBD 2015 Disease and Injury Incidence and Preva-

11730 Kiadaliri D.indd 177

21-02-2018 15:35:16

178

lence Collaborators 2016, GBD 2015 Mortality and Causes of Death Collaborators 2016). The GBD 2015 reported the estimates for the Nordic region as a whole. In the current study, we aimed to explore, for the first time, prevalence and disability due to OA in the Nordic region and across countries in the region between 1990 and 2015 using the findings from the GBD 2015.

Methods The GBD 2015 estimated the burden of 315 causes of diseases and injuries and 79 risk factors for 195 countries, 7 superregions, and 21 regions. The GBD uses 3 metrics to quantify the burden of a disease: years of life lost (YLLs) due to premature death, years lived with disability (YLDs) for the non-fatal health loss, and disability-adjusted life years (DALYs) as a measure of total burden (fatal and non-fatal). YLLs are calculated by multiplying the number of deaths from a cause in each age group by the reference life expectancy at the average age of death for those who die in that age group (GBD 2015 Mortality and Causes of Death Collaborators 2016). YLDs for each age, sex, location, and calendar year are computed by multiplying prevalence of a given cause sequela by disability weights for that sequela (GBD 2015 Disease and Injury Incidence and Prevalence Collaborators 2016). DALYs are calculated as the sum of YLLs and YLDs. 1 DALY is equivalent to 1 healthy year of life lost due to a specific disease or injury. In the GBD 2015, the OA reference case definition was symptomatic OA of the hip or knee radiologically confirmed as Kellgre–Lawrence grade 2–4 (GBD 2015 Disease and Injury Incidence and Prevalence Collaborators 2016). OA was assumed to be non-fatal and no mortality and no YLLs were estimated. For non-fatal health loss, the GBD 2013 systematic reviews of epidemiological measures for OA were updated in 2014. A list of all data points used in GBD 2015 is available from IHME Global Health Data Exchange (http://ghdx. healthdata.org/gbd-2015/data-input-sources). Statistics The detailed information on data and statistical analysis of the GBD 2015 study are available in the web appendix of a previous publication (GBD 2015 Disease and Injury Incidence and Prevalence Collaborators 2016, http://www.thelancet. com/cms/attachment/2090344431/2075672336/mmc1.pdf). In short, the data from systematic reviews were assembled using DisMod-MR 2.1, a Bayesian meta-regression tool to estimate prevalence by location, age, sex, and year accounting for study level and predictive covariates. The sequence of estimation in DisMod-MR 2.1 occurs at 5 levels (called “cascade”): global, super-region, region, country and, where applicable, subnational geographical unit. An initial run is made with all data points for all locations, age, sex, and time periods to get an initial fit at the global level. The estimates from the global fit

11730 Kiadaliri D.indd 178

Acta Orthopaedica 2018; 89 (2): 177–183

with adjustments for fixed effects (adjusting for study characteristics and predictive covariates) and random effects by geography is passed on to each of 7 super-region fits by sex and year (1990, 1995, 2000, 2005, 2010, and 2015) as priors and confronted with data from the super-region, sex, and year to make an initial fit for each super-region. For OA, the data from studies that reported OA based on radiographic only, self-reported OA with pain, reporting a physician diagnosis of OA, and symptomatic OA without radiographic confirmation were included and adjusted as study-level covariates in the modelling. The mean body mass index (BMI) was included as a predictive covariate. Then estimates from the super-region are passed on similarly to 21 region fits by sex and year and in turn the region estimates determine the priors for country estimates, and the country estimates determine the prior being passed down to subnational locations, where applicable. This cascade allows the GBD estimates to be made for locations with sparse or no primary data. In addition, disease-by-disease value priors may be set. In the case of OA, a prior value of no incidence and prevalence before the age of 30, zero remission, and no excess mortality were assumed. Based on the severity of OA, 3 sequelae (mild, moderate, and severe OA) were considered. Severity was classified according to the Western Ontario and McMaster Universities Arthritis Index (WOMAC) and the lay description of these was used to elicit disability weights. A random effects meta-analysis model based on the data from 4 studies from 3 regions was used to determine the proportion of people within each of the OA severity levels (based on frequentist methods). All measures were age-standardized using the GBD world population age standard. For each measure, 95% uncertainty intervals (UI) were calculated by taking the 2.5 and 97.5 centile values of 1,000 draws of the posterior distribution of that measure. It should be noted that as YLLs for OA was assumed to be zero, the YLDs is equal to DALYs. We obtained the estimates of the GBD 2015 on the burden of OA for the Nordic region and the 6 countries in the region (Denmark, Finland, Greenland, Iceland, Norway, and Sweden) for 1990–2015 at 5-year increments (1990, 1995, 2000, 2005, 2010, and 2015) from the Institute for Health Metrics and Evaluation interactive visualization tools (http:// vizhub.healthdata.org/gbd-compare). The region had a population of around 23 million in 1990 and 27 million in 2015. We reported the burden of OA attributable to high BMI (≥ 22.5) (GBD 2015 Risk Factors Collaborators 2016). We also compared the observed YLDs of OA with those expected based on the socio-demographic index (SDI). The SDI was constructed in the GBD 2015 for each location-year as a measure of overall development. It is based on the geometric mean of income per capita, average years of schooling among people aged 15 years or older, and total fertility rate (GBD 2015 DALYs and HALE Collaborators 2016). In the Nordic region in 2015 the SDI values were as follow: 0.76 for Greenland, 0.89 for Sweden, 0.89 for Finland, 0.91 for Denmark, and 0.94 for

21-02-2018 15:35:16

Acta Orthopaedica 2018; 89 (2): 177–183

179

Table 1. Age-standardized prevalence (%) of osteoarthritis in the countries of the Nordic region and the world Women Location Global Nordic region Denmark Finland Greenland Iceland Norway Sweden

1990

2015

Percent change 1990–2015

1990

Men 2015

3.8 (3.7–3.9) 3.7 (3.6–3.9) 4.3 (4.1–4.5) 3.9 (3.7–4.0) 4.4 (4.2–4.6) 4.1 (3.9–4.3) 3.6 (3.5–3.8) 3.4 (3.2–3.5)

4.1 (3.9–4.2) 4.0 (3.9–4.2) 4.6 (4.4–4.8) 4.1 (3.9–4.2) 4.8 (4.6–5.0) 4.4 (4.2–4.7) 3.9 (3.7–4.1) 3.8 (3.6–3.9)

6.6 (5.6–7.6) 8.4 (6.6–10.2) 7.2 (3.8–10.9) 4.8 (1.0–8.7) 8.1 (4.7–12.0) 8.2 (3.9–13.0) 7.4 (3.4–11.8) 11.9 (8.3–15.8)

2.6 (2.6–2.7) 2.9 (2.8–3.1) 3.5 (3.3–3.6) 3.0 (2.9–3.1) 3.6 (3.4–3.7) 3.2 (3.1–3.4) 2.9 (2.8–3.1) 2.6 (2.5–2.7)

2.8 (2.8–2.9) 3.2 (3.1–3.3) 3.7 (3.6–3.9) 3.2 (3.1–3.3) 3.9 (3.7–4.0) 3.5 (3.3–3.7) 3.2 (3.0–3.3) 2.8 (2.7–3.0)

Percent change 1990–2015 7.6 (6.5–8.5) 8.1 (6.3–9.8) 8.1 (4.6–11.8) 6.4 (2.7–10.3) 7.9 (4.5–11.2) 9.1 (4.9–13.6) 8.0 (3.8–11.7) 8.7 (5.1–11.9)

Values in parentheses show 95% uncertainty interval.

Prevalence (%) – women

Prevalence (%) – men

25 Global Nordic region

Global Nordic region

0 30–34 40–44 50–54 60–64 70–74 80– 35–39 45–49 55–59 65–69 75–79

30–34 40–44 50–54 60–64 70–74 80– 35–39 45–49 55–59 65–69 75–79

Age groups

global level, even though it was estimated to be higher in Denmark, Greenland, and Iceland and lower in Sweden compared with the global level (Table 1). Among men, the age-standardized prevalence of OA in all Nordic countries but Sweden was higher than the global average. Between 1990 and 2015, the age-standardized prevalence of OA increased by 7.5% (UI 6.1–8.9) in the region, which was slightly higher than the observed increase at the global level. In both sexes, the highest prevalence was observed in people 75–79 years old (Figure 1).

Age groups

YLDs Between 1990 and 2015, total YLDs of OA increased by 38% (UI 35–40) to 30,974 (UI 19,987–45,631) in women; and by 53% (UI 49–55) to 21,687 (UI 14,064–32,040) in men in the Nordic region. In both sexes, the increases in total YLDs from 1990–2015 in the region were smaller than the global rate of increase (Figure S1, Supplementary data). In the Nordic region, OA contributed to 1.33% (UI 1.00–1.67) of total YLDs in 1990 and this increased to 1.61% (UI 1.22–2.01) in 2015 (ranging from 1.37% [UI 1.03 to 1.72] in Greenland to 1.85% [UI 1.40–2.31] in Denmark, Figure 2). Of 315 causes studied, OA was ranked as the 15th leading cause of total YLDs in the region in 2015 (Figure 3). In addition, OA was among 10 leading causes of YLDs in those aged 55–74 years in 2015 (Figure S2, Supplementary data). While overall OA had a higher relative importance in women (14th leading cause of YLDs) than men (20th leading cause of YLDs), in those aged 30–50 years the opposite was observed. Among the top 30 leading causes of YLDs in the region, OA had the 5th greatest relative increase in total YLDs between 1990 and 2015 (Table 2). In both sexes, the highest and lowest age-standardized YLD rates of OA were seen in Greenland and Sweden, respectively (Figure S3, Supplementary data). The ratio of age-standardized YLDs rate in the Nordic region to the global level was

Figure 1. Age- and sex-specific prevalence (%) of osteoarthritis in the Nordic region and the world, 2015.

Norway. We also decomposed the changes in total YLDs from OA between 1990 and 2015 into population growth, aging, and change in age- and sex-specific YLD rates of OA. Ethics, funding, and potential conflicts of interest Ethics approval not applicable; data are publicly available. We would like to acknowledge support from the Swedish Research Council, Crafoord Foundation, Greta and Kocks Foundation, Österlunds Foundation, the Faculty of Medicine Lund University, Governmental Funding of Clinical Research within National Health Service (ALF) and Region Skåne. No conflicts of interest declared.

Results Prevalence The number of prevalent OA cases increased from 1,051,897 (UI 1,017,093–1,089,896) in 1990 to 1,507,587 (UI 1,454,338–1,564,778) in 2015 in the Nordic region (an increase of 43% [UI 41–45]). The age-standardized prevalence of OA for women in the region was comparable to the

11730 Kiadaliri D.indd 179

21-02-2018 15:35:16

180

Acta Orthopaedica 2018; 89 (2): 177–183

Percent of total YLDs – women

Percent of total YLDs – men

Percent of total DALYs – women

Percent of total DALYs – men

2.0

1.2

1.6

1.0

0.8

1.2

1.2 0.6

0.6

0.8

0.8 0.4

0.4

0.2

0 1995

1990

2000

2005

2010

2015

1995

1990

2000

2005

2010

2015

0 1990

1995

2000

2005

2010

2015

0 1990

1995

2000

2005

2010

2015

Figure 2. Burden of osteoarthritis as a proportion of total years lived with disability (YLDs) and disability-adjusted life years (DALYs) 1990–2015, by sex and location.

Location Global Nordic region Denmark Finland Greenland Iceland Norway Sweden

Women 1990 2015 15 12 17 14 16 12 18 15 28 17 20 14 17 14 16 14

rank slightly increased

YLDs Men 1990 2015 27 17 25 20 22 15 27 18 33 20 29 17 26 22 26 22

Both 1990 2015 16 13 18 15 17 13 23 15 32 18 21 13 17 15 20 16

DALYs Men 1990 2015 101 66 59 39 51 34 63 40 87 58 55 38 56 39 56 43

Women 1990 2015 60 35 33 26 32 24 35 27 71 42 36 21 35 27 33 26

rank moderately increased

Both 1990 2015 81 42 41 29 39 27 46 30 84 53 43 28 45 32 44 31

rank substantially increased

Figure 3. The rank of osteoarthritis among 315 causes in terms of total years lived with disability (YLDs) and disability-adjusted life years (DALYs) in 1990 and 2015, by sex and location (lower number indicates higher relative importance).

Total YLDs – women

YLDs per 100,000 800

6,000

Total YLDs – men

YLDs per 100,000 800

6,000 Number Rate

Number Rate 4,500

600

4,500

600

3,000

400

3,000

400

1,500

200

1,500

200

30–34 40–44 50–54 60–64 70–74 80– 35–39 45–49 55–59 65–69 75–79

Age groups

Figure 4. Total and age-specific rates of years lived with disability (YLDs) from osteoarthritis in the Nordic region in 2015, by sex and age groups.

0.65 in women and 0.73 in men. Between 1990 and 2015, the age-standardized YLDs rate of OA rose by 7.5% (UI 6.0–9.1) in the region, statistically significantly lower than the observed increase at the global level (10.5%, UI 9.1–12.0, Figure S1, Supplementary data). Across the Nordic countries, the largest and smallest increases in age-standardized YLD rates of OA were reported for women in Sweden and Finland, respectively.

11730 Kiadaliri D.indd 180

In 2015, the women-to-men age-standardized YLDs rate ratio was 1.27 in the region (ranging from 1.22 in Denmark to 1.32 in Sweden), which was slightly lower than the global level (1.43). Among women, the observed agestandardized YLD rates of OA were lower than expected on the basis of SDI alone in all Nordic countries, but among men this was the case only in Finland, Greenland, and Sweden in 2015 (Figure S4, Supplementary data). While the highest YLD rates were observed for people aged 75–79 years in both sexes, those women aged 80+ years and men aged 65–69 years had the highest number of YLDs (Figure 4). YLD rates were higher in women compared with men in all age groups except those aged 30–39 years. In the Nordic region in 2015, 23.7% and 22.4% of total YLDs of OA were attributable to high BMI in women and men, respectively, higher than the global average (Figure S5, Supplementary data). The decomposition analysis suggested that in all Nordic countries but Sweden and Norway, population aging made the highest contribution to the rise in total YLDs from 1990 to 2015 (Table 3, Supplementary data).

DALYs Between 1990 and 2015, the burden of OA as a proportion of total DALYs increased from 0.69% (UI 0.49–0.93) to 0.97% (UI 0.70–1.30) in women and from 0.38% (UI 0.26–0.52) to 0.63% (UI 0.45–0.85) in men in the Nordic region (Figure 2). While globally OA was ranked as 35th (66th) leading cause of total DALYs in women (men) in 2015, it had a higher rela-

21-02-2018 15:35:18

Acta Orthopaedica 2018; 89 (2): 177–183

181

Table 2. The burden of OA in comparison with 30 leading causes of years lived with disability (YLDs) for both sexes combined in the Nordic region in 2015

Rank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Cause Low back pain Neck pain Major depressive disorder Age-related and other hearing loss Diabetes mellitus Migraine Anxiety disorders Other musculoskeletal disorders Falls Iron-deficiency anemia Asthma Alzheimer’s disease and other dementias Edentulism and severe tooth loss Chronic obstructive pulmonary disease Osteoarthritis Schizophrenia Dysthymia Bipolar disorder Dermatitis Other mental and substance use disorders Alcohol use disorders Medication overuse headache Benign prostatic hyperplasia Rheumatoid arthritis Psoriasis Atrial fibrillation and flutter Other sense organ diseases Ischemic heart disease Other unintentional injuries Other congenital birth defects

Total YLDs (95% UI), 2015

Percent change1990–2015 in total crude age-standardized YLDs YLD rates YLD rates

341,720 (243,165–471,373) 257,221 (173,423–347,621) 171,717 (115,456–234,977) 148,248 (101,265–208,190) 138,677 (95,904–190,884) 132,779 (81,995–196,494) 114,250 (78,464–154,855) 104,139 (69,148–144,842) 92,541 (63,146–129,745) 82,713 (55,250–120,348) 73,429 (47,637–102,761) 66,549 (47,170–88,334) 61,886 (41,423–86,253) 55,530 (46,435–65,762) 52,661 (34,056–77,499) 46,913 (34,275–58,739) 42,564 (28,457–61,162) 39,793 (24,723–58,761) 39,510 (27,057–54,714) 38,861 (27,055–52,365) 38,585 (25,998–54,459) 38,539 (25,260–54,435) 37,653 (24,423–53,434) 37,311 (26,047–49,975) 36,829 (25,740–49,858) 36,593 (24,776–49,774) 35,211 (22,070–51,023) 31,912 (21,923–43,427) 30,357 (19,917–43,616) 29,647 (12,960–72,466)

16.5 24.1 11.5 37.0 56.1 11.6 11.8 40.4 –3.5 24.2 –7.6 39.6 30.8 25.0 43.3 17.4 18.2 12.2 12.5 17.0 3.1 16.3 45.6 15.6 20.2 62.5 29.0 –1.7 –33.4 99.0

1.6 8.3 –2.7 19.6 36.2 –2.6 –2.4 22.5 –15.8 8.4 –19.4 21.8 14.2 9.1 25.1 2.5 3.1 –2.1 –1.9 2.1 –10.1 1.5 27.0 0.9 4.9 41.8 12.5 –14.2 –41.9 73.6

–2.4 1.5 –3.0 4.4 19.6 –0.2 –1.1 12.9 –25.9 13.1 –18.1 0.9 0.5 –3.9 7.5 –0.8 –0.3 0.9 0.0 0.6 –10.5 –1.7 6.9 –9.9 1.2 21.4 2.2 –26.5 –48.2 77.2

UI = uncertainty interval.

tive importance in the region (ranked as 26th and 39th leading cause of total DALYs in women and men, respectively (Figure 3).

Discussion This report is, to our knowledge, the first comparative investigation of the OA burden in the Nordic region. The burden of OA consistently rose in the region from 1990 through 2015, mainly due to population aging. The burden peaked in the 75–79 age group and was higher in women than in men. In addition, OA was among the 10 leading causes of YLDs in those aged 55–74 years in 2015. While the prevalence of OA in the region was comparable to the global level, the agestandardized estimate of OA burden was lower than the global average. Moreover, the region experienced slightly lower increases in OA burden than the global average between 1990 and 2015. On the other hand, while OA was globally ranked as 42nd leading cause of DALYs, it had higher relative importance in the Nordic region (29th leading cause). Across the countries in the region, Greenland had the highest age-stan-

11730 Kiadaliri D.indd 181

dardized OA burden and the highest proportion of OA burden from total YLDs was seen in Denmark. There were persistent increases in prevalence and burden of OA across all Nordic countries from 1990 through 2015. The aging population alongside the limited effect of treatments for OA are possible explanations for this increasing trend. Furthermore, marked increases in overweight and obesity in the region (Asgeirsdottir and Gerdtham 2016) could explain part of the rise in OA burden. However, as a recent study noted (Wallace et al. 2017), increases in longevity and obesity cannot fully explain recent increases in OA burden and recent environmental changes (e.g., physical inactivity and high intake of refined carbohydrates) might have played an important role. Moreover, the role of advances in diagnosis and medicalization should not be ruled out as more people are recognizing OA as a treatable condition, not just a given for getting old. The expected future increases in life expectancy and prevalence of overweight and obesity imply that the OA burden will continue to rise in the coming decades. A recent study reported rises in incidence of knee arthroplasty across Nordic countries between 1997 and 2012 (Niemelainen et al. 2017). Moreover, the rate of hospital admissions due to knee and hip OA

21-02-2018 15:35:21

182

increased by 50% and 32% between 1998 and 2015 in Sweden (http://www.socialstyrelsen.se). Furthermore, healthcare consultation for OA and primary knee arthroplasty are projected to rise in coming decades in Sweden (Turkiewicz et al. 2014, Nemes et al. 2015). These rising trends highlight the need to incorporate prevention and treatment of OA and sports-related injuries as a priority in health policies in the region. At the population level, raising public awareness about the importance of a healthy lifestyle (e.g., maintain physical fitness and an ideal weight, follow a balanced diet), improving occupational safety and ergonomics, and promoting public awareness of OA should be part of any preventive initiative (Mody and Brooks 2012, Woolf et al. 2012, Bruyere et al. 2014, Briggs et al. 2016, Moradi-Lakeh et al. 2017). At patient level, informing patients about the nature of the disease and treatment goals and importance of self-management, providing lifestyle recommendations (exercise and weight loss), and implementation of integrated models of patient-centered multidisciplinary care have been suggested in order to address the rising burden of OA (Mody and Brooks 2012, Woolf et al. 2012, Bruyere et al. 2014, Briggs et al. 2016). Along these lines, OA prevention and management programs aiming to offer adequate information and exercise to people with knee or hip OA have been initiated in Sweden (Better management of patients with OsteoArthritis, https://boa.registercentrum.se/), Denmark (Good Life with osteoArthritis in Denmark, https://www.glaid.dk/) and Norway (ActiveOA—Active living with osteoarthritis, http://aktivmedartrose.no/) and establishing similar programs in other Nordic countries is highly recommended. The magnitude and trend of OA burden varied by sex and country. Incidence, prevalence, and severity of OA is higher in women compared with men (Srikanth et al. 2005, Neogi and Zhang 2013), and any preventive and therapeutic intervention should consider this sex difference. In addition, while the changes in age-standardized YLD rates of OA were comparable between sexes, men observed greater increases in total YLDs than women did over the observation period. This might be partly due to a higher gain in life expectancy in men compared with women over the study period in the Nordic region (life expectancy at birth increased by 6.0 years in men and 4.2 years in women from 1990 through 2015 (GBD 2015 Mortality and Causes of Death Collaborators 2016)). Differences in population genetic predisposition and lifestyle, in epidemiologic and clinical profile of OA, in prevalence of OA risk factors, and in the healthcare system might partially explain variation in OA burden across the Nordic countries (Croft 1996, Petersson and Jacobsson 2002). Further investigations are required to explore the observed variations in the magnitude and secular trends of OA burden in the region. In addition to the general limitations of the GBD 2015, which have been acknowledged (GBD 2015 DALYs and HALE Collaborators 2016, GBD 2015 Disease and Injury Incidence and Prevalence Collaborators 2016, GBD 2015 Risk Factors Collaborators 2016), several OA-specific limitations should

11730 Kiadaliri D.indd 182

Acta Orthopaedica 2018; 89 (2): 177–183

be mentioned. The reference case definition based on Kellgren–Lawrence grades 2–4 is a conservative definition of OA (Cross et al. 2014) excluding a substantial part of early-stage OA patients from the estimates. The distribution of people across the OA severity levels was determined mainly using the data from high-income regions with no adjustment for potential changes over time. In addition, the brief lay descriptions used to elicit disability weights were mainly focused on major functional consequences and symptoms that could not fully capture the impact of OA on other aspects of health and well-being. The burden of OA was limited to knee and hip OA and other sites were not included. All these limitations likely led to underestimation of the true OA burden in this study. Nevertheless, these results provide an up-to-date comparative assessment of OA burden in the Nordic region of importance to policy-makers and health professionals. Summary The burden of OA is high and rising in the Nordic region. While among women prevalence of OA was comparable to the global average, men had a higher prevalence than the global level. There were persistent increases in the age-standardized rates of OA burden with time. Furthermore, OA was responsible for a higher proportional burden of total DALYs in the Nordic region compared with the global level. Among the top 30 leading causes of YLDs in the region, OA had the 5th greatest relative increase in total YLDs from 1990 to 2015. In all Nordic countries but Denmark the observed OA burden was lower than expected based on sociodemographic status, but worryingly this gap between the observed and expected number of YLDs is closing. Current trends in population growth, aging, and prevalence of overweight and obesity will lead to a further increase in OA burden in coming years and call for incorporating OA prevention and treatment as a priority in health policies in the Nordic region. Supplementary data Table 3 and Figures S1–S5 are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674.2017.1404791.

AAK participated in the design, acquisition of data, analysis, and interpretation of results and drafting the manuscript. LSL, MM-L, IFP, and ME participated in interpretation of results, and revision of the manuscript. Acta thanks Lars Vatten and Maiju Welling for help with peer review of this study.

Asgeirsdottir T L, Gerdtham U G. Health behavior in the Nordic countries. Nordic J Health Economics 2016; 1: 28-40.

21-02-2018 15:35:21

Acta Orthopaedica 2018; 89 (2): 177–183

Briggs A M, Cross M J, Hoy D G, Sanchez-Riera L, Blyth F M, Woolf A D, March L. Musculoskeletal health conditions represent a global threat to healthy aging: A report for the 2015 World Health Organization World Report on Ageing and Health. Gerontologist 2016; 56 (Suppl 2): S243-55. doi: 10.1093/geront/gnw002. Bruyere O, Cooper C, Pelletier J P, Branco J, Luisa Brandi M, Guillemin F, Hochberg M C, Kanis J A, Kvien T K, Martel-Pelletier J, Rizzoli R, Silverman S, Reginster J Y. An algorithm recommendation for the management of knee osteoarthritis in Europe and internationally: A report from a task force of the European Society for Clinical and Economic Aspects of Osteoporosis and Osteoarthritis (ESCEO). Semin Arthritis Rheum 2014; 44 (3): 253-63. doi: 10.1016/j.semarthrit.2014.05.014. Croft P. The occurrence of osteoarthritis outside Europe. Ann Rheum Dis 1996; 55 (9): 661-4. Cross M, Smith E, Hoy D, Nolte S, Ackerman I, Fransen M, Bridgett L, Williams S, Guillemin F, Hill C L, Laslett L L, Jones G, Cicuttini F, Osborne R, Vos T, Buchbinder R, Woolf A, March L. The global burden of hip and knee osteoarthritis: Estimates from the global burden of disease 2010 study. Ann Rheum Dis 2014; 73 (7): 1323-30. doi: 10.1136/annrheumdis-2013-204763. GBD 2015 DALYs and HALE Collaborators. Global, regional, and national disability-adjusted life-years (DALYs) for 315 diseases and injuries and healthy life expectancy (HALE), 1990–2015: A systematic analysis for the Global Burden of Disease Study 2015. Lancet 2016; 388 (10053): 1603-58. doi: 10.1016/S0140-6736 (16)31460-X. GBD 2015 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990-2015: A systematic analysis for the Global Burden of Disease Study 2015. Lancet 2016; 388 (10053): 1545-602. doi: 10.1016/S0140-6736 (16)31678-6. GBD 2015 Mortality and Causes of Death Collaborators. Global, regional, and national life expectancy, all-cause mortality, and cause-specific mortality for 249 causes of death, 1980–2015: A systematic analysis for the Global Burden of Disease Study 2015. Lancet 2016; 388 (10053): 1459544. doi: 10.1016/S0140-6736 (16)31012-1. GBD 2015 Risk Factors Collaborators. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks, 1990–2015: A systematic analysis for the Global Burden of Disease Study 2015. Lancet 2016; 388 (10053): 1659-724. doi: 10.1016/S0140-6736 (16)31679-8. Hiligsmann M, Reginster J Y. The economic weight of osteoarthritis in Europe. Medicographia 2013; 35: 197-202. Kiadaliri A A, Lamm C J, de Verdier M G, Engstrom G, Turkiewicz A, Lohmander L S, Englund M. Association of knee pain and different definitions of knee osteoarthritis with health-related quality of life: a population-based cohort study in southern Sweden. Health Qual Life Outcomes 2016; 14 (1): 121. doi: 10.1186/s12955-016-0525-4. Leskinen J, Eskelinen A, Huhtala H, Paavolainen P, Remes V. The incidence of knee arthroplasty for primary osteoarthritis grows rapidly among baby boomers: A population-based study in Finland. Arthritis Rheum 2012; 64 (2): 423-8. doi: 10.1002/art.33367.

11730 Kiadaliri D.indd 183

183

Mody G M, Brooks P M. Improving musculoskeletal health: Global issues. Best Pract Res Clin Rheumatol 2012; 26 (2): 237-49. doi: 10.1016/j.berh. 2012.03.002. Moradi-Lakeh M, Forouzanfar M H, Vollset S E, et al.. Burden of musculoskeletal disorders in the Eastern Mediterranean Region, 1990–2013: Findings from the Global Burden of Disease Study 2013. Ann Rheum Dis 2017. doi: 10.1136/annrheumdis-2016-210146. Murray C, Lopez A. The global burden of disease: A comprehensive assessment of mortality and disability from diseases, injuries and risk factors in 1990 and projected to 2020. Cambridge, MA: Harvard University Press, 1996. Nemes S, Rolfson O, A W D, Garellick G, Sundberg M, Karrholm J, Robertsson O. Historical view and future demand for knee arthroplasty in Sweden. Acta Orthop 2015; 86 (4): 426-31. doi: 10.3109/17453674.2015.1034608. Neogi T, Zhang Y. Epidemiology of osteoarthritis. Rheum Dis Clin North Am 2013; 39 (1): 1-19. doi: 10.1016/j.rdc.2012.10.004. Nguyen U S, Zhang Y, Zhu Y, Niu J, Zhang B, Felson D T. Increasing prevalence of knee pain and symptomatic knee osteoarthritis: Survey and cohort data. Ann Intern Med 2011; 155 (11): 725-32. doi: 10.7326/0003-4819155-11-201112060-00004. Niemelainen M J, Mäkelä K T, Robertsson O, W–Dahl A, Furnes O, Fenstad A M, Pedersen A B, Schroder H M, Huhtala H, Eskelinen A. Different incidences of knee arthroplasty in the Nordic countries. Acta Orthop 2017: Acta Orthop 2017; 88(2): 173-8. doi: 10.1080/17453674.2016.1275200. Petersson I F, Jacobsson L T. Osteoarthritis of the peripheral joints. Best Pract Res Clin Rheumatol 2002; 16 (5): 741-60. Rahman M M, Cibere J, Goldsmith C H, Anis A H, Kopec J A. Osteoarthritis incidence and trends in administrative health records from British Columbia, Canada. J Rheumatol 2014; 41 (6): 1147-54. doi: 10.3899/ jrheum.131011. Srikanth V K, Fryer J L, Zhai G, Winzenberg T M, Hosmer D, Jones G. A meta-analysis of sex differences prevalence, incidence and severity of osteoarthritis. Osteoarthritis Cartilage 2005; 13 (9): 769-81. doi: 10.1016/j. joca.2005.04.014. Turkiewicz A, Petersson I F, Bjork J, Hawker G, Dahlberg L E, Lohmander L S, Englund M. Current and future impact of osteoarthritis on health care: A population-based study with projections to year 2032. Osteoarthritis Cartilage 2014; 22 (11): 1826-32. doi: 10.1016/j.joca.2014.07.015. Wallace I J, Worthington S, Felson D T, Jurmain R D, Wren K T, Maijanen H, Woods R J, Lieberman D E. Knee osteoarthritis has doubled in prevalence since the mid-20th century. Proc Natl Acad Sci USA 2017; 114 (35): 93326. doi: 10.1073/pnas.1703856114. Woolf A D, Erwin J, March L. The need to address the burden of musculoskeletal conditions. Best Pract Res Clin Rheumatol 2012; 26 (2): 183-224. doi: 10.1016/j.berh.2012.03.005. Yu D, Peat G, Bedson J, Jordan K P. Annual consultation incidence of osteoarthritis estimated from population-based health care data in England. Rheumatology (Oxford) 2015; 54 (11): 2051-60. doi: 10.1093/rheumatology/ kev231.

21-02-2018 15:35:21

184

Acta Orthopaedica 2018; 89 (2): 184–189

Fast-tracking for total knee replacement reduces use of institutional care without compromising quality A register-based analysis of 4 hospitals and 4,256 replacements Konsta J PAMILO 1, Paulus TORKKI 2, Mikko PELTOLA 3, Maija PESOLA 1, Ville REMES 4, and Juha PALONEVA 1

1 Department

of Orthopaedics and Traumatology, Central Finland Hospital, Jyväskylä, 2 Aalto University, Helsinki, 3 Centre for Health and Social Economics CHESS, National Institute for Health and Welfare, Helsinki, 4 Pihlajalinna Group, Helsinki, Finland Correspondence: konsta.pamilo@ksshp.fi Submitted 2017-09-19. Accepted 2017-10-17.

Background and purpose — Fast-tracking shortens the length of the primary treatment period (length of stay, LOS) after total knee replacement (TKR). We evaluated the influence of the fasttrack concept on the length of uninterrupted institutional care (LUIC) and other outcomes after TKR. Patients and methods — 4,256 TKRs performed in 4 hospitals between 2009–2010 and 2012–2013 were identified from the Finnish Hospital Discharge Register and the Finnish Arthroplasty Register. Hospitals were classified as fast track (Hospital A) and non-fast track (Hospitals B, C and D). We analyzed length of uninterrupted institutional care (LUIC), LOS, discharge destination, readmission, revision, manipulation under anesthesia (MUA) and mortality rate in each hospital. We compared these outcomes for TKRs performed in Hospital A before and after fast-track implementation and we also compared Hospital A outcomes with the corresponding outcomes for the other 3 hospitals. Results — After fast-track implementation, median LOS in Hospital A fell from 5 to 3 days (p < 0.001) and (median) LUIC from 7 to 3 (p < 0.001) days. These reductions in LOS and LUIC were accompanied by an increase in the discharge rate to home (p = 0.01). Fast-tracking in Hospital A led to no increase in 14and 42-day readmissions, MUA, revision or mortality compared with the rates before fast-tracking, or with those in the other hospitals. Of the 4 hospitals, LOS and LUIC were most reduced in Hospital A. Interpretation — A fast-track protocol reduces LUIC and LOS after TKR without increasing readmission, complication or revision rates.

The aim of fast-tracking is to optimize the whole treatment protocol, leading eventually to shorter length of stay (LOS) without compromising treatment quality (Husted 2012). For selected patients, even same-day discharge after TKR is feasible (Gromov et al. 2017, Hoorntje et al. 2017). Fast-track TKR is not associated with higher readmission, reoperation, manipulations under anesthesia (MUA) or mortality rates (Husted et al. 2010b, 2014, Glassou et al. 2014, Wied et al. 2015, Winther et al. 2015, Jørgensen et al. 2017). In Finnish hospitals, LOS and length of uninterrupted institutional care (LUIC) after TKR have universally decreased over the past decade (Pamilo et al. 2015). In previous fasttrack studies, the overall reduction in LOS, even without fasttracking, has rarely been taken into account (Glassou et al. 2014). Apart from studies on LOS conducted only on hospitals directly discharging to home, 100% of patients (Husted et al. 2010b, 2011a, Jørgensen et al. 2013a), no reports have been published on total length of uninterrupted institutional care (LUIC) after fast-track TKR. It is important to enhance the efficiency of these procedures, i.e., lower their economic impact, without compromising their outcomes (Andreasen et al. 2016). By combining Finnish Arthroplasty Register and hospital discharge register data and benchmarking data from 4 different hospitals, we evaluated the effect of introducing fast-tracking on LUIC, LOS, discharge destination, readmissions, early revision, MUA and mortality rates after TKR.

■

12089 Pamilo D.indd 184

20-02-2018 18:02:41

Acta Orthopaedica 2018; 89 (2): 184–189

Patients and methods For this study, we selected 4 similar Finnish public central hospitals, all with some teaching responsibilities, from a benchmarking database maintained by Nordic Healthcare Group Ltd (NHG). Implementation of a fast-track protocol started in September 2011 in Hospital A, which soon after that date fulfilled all the fast-track criteria. The other hospitals (Hospitals B, C, and D) did not meet the fast-track criteria to the same extent. For fast-track criteria and characteristics of the hospitals, see Pamilo et al. (2017). A hospital was classified as a fast-track hospital if it fulfilled all the fast-track criteria as evaluated from answers to a written questionnaire sent to each hospital in the study. Patient education and information in Hospital A was planned to give the patient all the information needed to enable early discharge. Preoperative education included patient education seminars and an outpatient session with an orthopedic surgeon and a nurse. Written standardized information was given to all patients and included a phone number to be called in case of any questions. This study is based on the PERFECT hip and knee replacement databases (Mäkelä et al. 2011), which collect data from the Finnish Hospital Discharge Register (FHDR) and the Finnish Arthroplasty Register (FAR), cause-of-death statistics (Statistics Finland) and drug prescription and drug reimbursement registers (Social Insurance Institution). All public and private hospitals in Finland are obliged to report all surgical procedures to the FHDR. In comparison with the FHDR, the FAR coverage for primary knee replacements in the 4 target hospitals during the study period was 91% in Hospital A, 96% in Hospital B, 81% in Hospital C and 97% in Hospital D (Institute for Health and Welfare 2017). We evaluated LOS, LUIC, discharge destination, presence at home 1 week post-surgery, readmissions, revisions, MUAs and mortality during 2 2-year periods, 1 before (2009–2010) and 1 after (2012–2013) fasttrack implementation in Hospital A. Patients were followed up until the end of 2015. The results for Hospital A were also compared with those for the other hospitals (Hospitals B, C and D). However, the readmission and MUA rates were not compared with those of the other hospitals due to variation in the readmission and MUA criteria. For definition and calculation of LOS and LUIC, see Pamilo et al. (2017). Inclusion criteria The study population was formed by selecting patients from the FHDR according to the WHO International Classification of Diseases (ICD-10 2010) and applying the following criteria: M17.0/M17.1 for primary osteoarthritis (OA) of the knee. The codes for primary TKR were NGB20, NGB30, NGB40 and NGB50, according to the NOMESCO classification of surgical procedures, Finnish version. The accuracy of the diagnosis of primary OA was double-checked against the relevant data in the FAR. It should be noted that the length of

12089 Pamilo D.indd 185

185

the surgical treatment period, the length of institutional care, and unscheduled readmissions were evaluated for total knee replacements—not patients. Exclusion criteria TKRs performed for secondary OA and revisions were excluded (Appendix 1). A diagnosis of secondary knee OA was noted retrospectively from the beginning of 1987. A patient was excluded from the study if a diagnosis of secondary knee OA had been recorded in the Hospital Discharge Register between the beginning of 1987 and the day of the operation. Patients listed in the Social Insurance Institution database as eligible for reimbursement for the sequelae of transplantation, uremia requiring dialysis, rheumatoid arthritis, or connective tissue disease were excluded from the study. We also excluded patients who were not Finnish citizens or were residents of the autonomous region of Åland. Readmission Readmission was recorded if the patient had been readmitted after discharge to any ward in any hospital in Finland during the first 14 or 42 days from the index operation. Direct transfer to another hospital was not counted as a readmission. Only the first readmissions for any reason after the index operation (also readmissions not directly related to the index TKR operation) were included in the study. Revision and MUA A search for revision surgery on the same knee after TKR was conducted using codes NGC00–NGC99 and for MUA using code NGT60. A search for removal of the total prosthesis from the knee was made in the FAR. Patients were followed up until the end of 2015. Only first revisions within 1 year and first MUA of the same knee within 6 months of the primary TKR were included. Non-standardized indications for MUA were flexion < 90 degrees or unsatisfactory flexion. Discharge destination Some patients are admitted to hospital from other social and welfare institutions and therefore are unlikely to be discharged home. Thus, only patients who came from home to hospital for their TKR were included in the discharge destination analyses. The percentage of patients who were at home 1 week after TKR was also analyzed irrespective of the hospital discharge destination. Statistics The same statistical procedures were used as in Pamilo et al. (2017). Ethics, funding, and potential conflicts of interest Permission for the study was obtained from each register and from each study hospital. No ethics permission was required

20-02-2018 18:02:41

186

Acta Orthopaedica 2018; 89 (2): 184–189

LOS (days)

LUIC (days)

7 2009–2010 2012–2013

A discharged 5% of the TKR patients home on the first postoperative day. Despite the post-fast-tracking reduction in LOS, Hospital A’s discharge destination rates to home increased (from 66% to 75%) (p = 0.01). However, Hospitals B and C, with longer LOS, continued to discharge more TKR patients directly home than Hospital A (p < 0.001). Hospital D showed similar LOS (3 days; CI 3–5) and discharge rate (71%) to home as Hospital A after fast-tracking.

Episode Median LUIC in Hospital A was 7 (CI 3–24) days A B C D A B C D Hospital Hospital before fast-tracking and 3 (CI 2–20) days (p < 0.001) Figure 1. Median length of stay in days (LOS; left panel) and of uninterrupted instithereafter (Figure 1). After fast-track implementation, tutional care (LUIC; right panel) in 2 2-year periods for primary total knee arthromedian LUIC was shorter in Hospital A than hospital plasty in 4 different hospitals. Hospital A was defined as a fast-track hospital after C (5 days; CI 4–22) (p < 0.01) but not significantly 2011. shorter than in Hospitals B (4 days; CI 3–14) or D (3 days; CI 3–14). The percentage of patients at home a week after to perform this registry study. No grants were received to con- TKR increased from 48% before fast-tracking to 75% thereafter duct this study. No conflicts of interest are declared. in Hospital A (p < 0.001). After fast-tracking in Hospital A, this percentage was higher only in Hospital B (84%, p < 0.001).

Results 4,256 TKRs meeting the inclusion but not exclusion criteria were identified from the FHDR and FAR. Of these, 437 were performed in Hospital A before, and 624 after, implementation of the fast-track protocol. The corresponding numbers in the other hospitals were 367 and 442 in Hospital B, 501 and 514 in Hospital C, and 641 and 730 in Hospital D. No statistically significant age or sex differences were observed before or after fast-tracking in Hospital A, or between hospital A and the other hospitals. Primary hospital stay Before implementing fast-tracking, the median LOS in Hospital A was 5 (CI 3–9) days: thereafter, it fell to 3 (CI 1–5) days (p < 0.001) (Figure 1). After fast-tracking, LOS was statistically significantly shorter in Hospital A than in Hospitals B (4 days; CI 3–14) (p < 0.001) or C (4 days; CI 3–6) (p < 0.05). Unlike the other study hospitals, after fast-tracking, Hospital

Quality and complications In Hospital A, the rate of revision TKR (within 1 year after the primary operation) was 1.1% (CI 0.0–2.2) between 2009 and 2010 and 2.4% (CI 1.4–3.4) in patients operated between 2012 and 2013 (NS). No statistically significant differences in revision rates were observed before or after the implementation of fast-tracking in Hospital A between the 4 hospitals (Table 1). The rate of MUA (during the first 6 months after the primary operation) was 6.4% (CI 5.1–7.8) before and 5.9% (CI 4.8–7.0) after fast-tracking in Hospital A. Unscheduled readmissions and mortality In Hospital A, the 14-day readmission rate was 2.4% (CI 1.1–3.6) before and 1.6% (CI 0.5–2.8) after fast-tracking, and the corresponding 42-day readmission rates were 6.0% (CI 3.9–8.2) and 6.1% (CI 4.3–7.9). The reasons for readmission recorded in the hospital discharge register are given in Table 2 (see Supplementary data).

Table 1. Adjusted revision rates and mortality during 1 year in 2-year periods for primary total knee arthroplasty in four different hospitals a

Hospital

TKR n

A B C D

437 367 501 641

12089 Pamilo D.indd 186

2009–2010 Revision Mortality rate (%) (95% CI) rate (%) (95 % CI) 1.1 1.8 1.4 1.7

(0.0–2.2) (0.5–3.1) (0.3–2.5) (0.8–2.7)

0.8 0.8 0.8 0.8

(0.7–0.9) (0.8–0.9) (0.8–0.8) (0.7–0.9)

TKR n 624 442 514 730

2012–2013 Revision Mortality rate (%) (95% CI) rate (%) (95% CI) 2.4 1.8 1.4 2.7

(1.4–3.4) (0.6–3.1) (0.3–2.5) (1.7–3.6)

0.7 0.7 0.7 0.7

(0.6–0.8) (0.7–0.8) (0.4–0.9) (0.6–0.8)

fast-track protocol was implemented in Hospital A in September 2011.

20-02-2018 18:02:41

Acta Orthopaedica 2018; 89 (2): 184–189

Mortality at 1 year after TKR in Hospital A was 0.8% (CI 0.7–0.9) before and 0.7% (CI 0.6–0.7) after fast-tracking (Table 1). Mortality rates were similar between the hospitals.

Discussion The aim of this study was to evaluate the effect of a fast-track protocol on LOS and LUIC after TKR. Median LOS and LUIC both decreased along with an increase in the discharge rate directly to home and without any significant change in readmission, revision surgery or MUA. We have recently reported similar findings for THA (Pamilo et al. 2017) Validity of the data The level of completeness and accuracy in the FHDR is satisfactory (Sund 2012) and the coverage of FAR is good (Institute for Health and Welfare 2017). The strength of our study is the inclusion of data from all the private and public hospitals in Finland. Thus, all revisions, MUAs and readmissions were included in the analyses. Only 1 hospital (A) in our study had fully implemented the fast-track protocol. In addition to fasttracking, the changes in the studied parameters may also in part be explained by other factors, such as other processual changes and differences in the annual arthroplasty volume of surgeons. LOS Several factors have been reported to affect LOS: surgeon volume, hospital volume, time between surgery and mobilization, process standardization (such as fast-track programs), operation day and patient-related factors (Judge et al. 2006, Mitsuyasu et al. 2006, Bozic et al. 2010, Husted et al. 2010a, Paterson et al. 2010, Styron et al. 2011, Pamilo et al. 2015, Jans et al. 2016, Mathijssen et al. 2016). An annual decline in LOS after TKR, even in the absence of a fast-track protocol, has been reported (Cram et al. 2012, Pamilo et al. 2015). The same observation was also made in the hospitals studied here. The effect of this annual decline in LOS has not usually been taken into account in earlier fast-track studies (Husted et al. 2010b, den Hartog et al. 2013, Winther et al. 2015). Thus, it can be argued either that the effect of fast-tracking on LOS has been overestimated in those studies or that non-fast-track hospitals have adopted some of the features of fast-tracking, resulting in shorter LOS. The latter possibility was also discussed by Glassou et al. (2014) in their study. In line with our previous report on fast-track THR (Pamilo et al. 2017), we found in this study that fast-track implementation in Hospital A resulted in a statistically significant decrease in LOS and LUIC. Our finding of a median LOS of 3 days accords with previous reports on LOS after fast-track TKR (Husted et al. 2010b, 2016, Glassou et al. 2014, Winther et al. 2015, Pitter et al. 2016). After fast-tracking, median LUIC in our study was 3 days, which mimics the results of studies of hospitals discharging all their patients directly home (Husted

12089 Pamilo D.indd 187

187

et al. 2010b, 2011a, Jørgensen et al. 2013a). The other hospitals in our study had implemented some elements of the fasttrack protocol (Pamilo et al. (2017). However, median LOS and LUIC decreased statistically significantly only in Hospital A, which had systematically and comprehensively implemented fast-tracking to its full extent. Further, while LOS was shorter in Hospital A after fast-track implementation than in Hospitals B or C, LUIC was statistically significantly shorter only when compared with Hospital C. Discharge destination Patient expectation, one of the most important factors predicting discharge destination (Halawi et al. 2015), presents a challenge for preoperative patient education. Discharging TKR patients to a skilled care facility has been associated with higher readmission rates (Keswani et al. 2016, McLawhorn et al. 2017). The economic wisdom of discharging patients to an extended institutional care facility instead of allowing longer LOS has also been disputed (Sibia et al. 2017). 1 earlier fasttrack study reported a discharge rate to home after TKR of 80%, both before and after fast-tracking (Winther et al. 2015). In our study, the discharge destination rate to home increased statistically significantly (66% to 75%) after fast-tracking, as also did the proportion of patients at home 1 week after surgery. The last-mentioned accords with our previous report after THR (Pamilo et al. 2017). Hospitals B and C, in which LOS was longer, nevertheless discharged more TKR patients directly home than either Hospitals A or D. Hospitals B and C, unlike A and D, were aiming at short stay throughout the study period via patient education. Unscheduled readmissions Unscheduled readmissions are widely used as a marker of quality of care. However, comparison of readmission rates between studies is difficult, because definitions of readmission, and diagnoses, vary between studies. Moreover, readmissions to other hospitals have not been included in all the previous studies (Ramkumar et al. 2015). A recent systematic review found the readmission rate after TKR to be 3.3% within 30 days and 9.7% within 90 days, with surgical site infection as the leading reason (Ramkumar et al. 2015). Although we included all events that required care in any hospital and in any ward, our finding of a 42-day readmission rate (6%) with no increase after fast-tracking is in line with previous fasttrack reports (Jørgensen et al. 2013b, Husted et al. 2016). Revision and MUA The revision rate after fast-track TKR has been reported to be between 1.4% and 2% within 90 days and 3.3% within one year (Husted et al. 2008, 2011b, Glassou et al. 2014, Winther et al. 2015). In line with Glassou et al. (2014), no significant difference was observed in revision rates before and after fasttracking in Hospital A or between Hospital A’s pre- and postfast-tracking revision rates and those of the other 3 hospitals.

20-02-2018 18:02:46

188

In our earlier study, we found no association between short LOS and increased risk for MUA (Pamilo et al. 2015). Moreover, in line with our present results, no increase in the incidence of MUA rates after fast-tracking has been reported (Husted et al. 2015, Wied et al. 2015). Mortality Death after TKR is relatively rare event and not always surgery-related (Jørgensen et al. 2017). An enhanced recovery program has been found to be associated with a significant or nearly significant reduction in mortality after TKR and THR (Malviya et al. 2011, Savaridas et al. 2013, Khan et al. 2014). However, for patients with a comorbidity burden at the time of surgery mortality risk has not declined (Glassou et al. 2017). In our study, the 1-year mortality rate was 0.7% after fast-tracking. This is a little lower than the 1-year mortality 1.3% reported by Savaridas et al. (2013), but their study also included THR patients. Other studies have reported 90-day mortality rates of 0.2%–0.5% after fast-track THR and TKR (Husted et al. 2010b, Malviya et al. 2011, Khan et al. 2014, Glassou et al. 2017, Jørgensen et al. 2017). Summary Process standardization by fast-tracking protocols offers an opportunity to substantially reduce LUIC and LOS. In addition, implementation of fast-tracking increases the discharge rate to home. Fast-track protocols do not appear to increase complication or revision rates. Supplementary data Table 2 and Appendices 1 and 2 are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674.2017.1399643

KJP, PT, MiP, MaP, VR, and JP wrote the manuscript. PT and MiP performed the data analysis. All contributed to the conception and design of the study, to critical analyses of the data, to interpretation of the findings, and to critical revision of the manuscript.

Acta thanks Per Kjaersgaard-Andersen for help with peer review of this study.

Andreasen S E, Holm H B, Jørgensen M, Gromov K, KjærsgaardAndersen P, Husted H. Time-driven activity-based cost of fasttrack total hip and knee arthroplasty. J Arthroplasty 2016; 32: 1747-55. Bozic K J, Maselli J, Pekow P S, Lindenauer P K, Vail T P, Auerbach A D. The influence of procedure volumes and standardization of care on quality and efficiency in total joint replacement surgery. J Bone Joint Surg Am 2010; 92 (16): 2643-52. Cram P, Lu X, Kates S L, Singh J A, Li Y, Wolf B R. Total knee arthroplasty volume, utilization, and outcomes among Medicare beneficiaries, 1991–2010. JAMA 2012; 308 (12): 1227-36. den Hartog Y M, Mathijssen N M, Vehmeijer S B. Reduced length of hospital stay after the introduction of a rapid recovery protocol for primary THA procedures. Acta Orthop 2013; 84 (5): 444-7.

12089 Pamilo D.indd 188

Acta Orthopaedica 2018; 89 (2): 184–189

Glassou E N, Pedersen A B, Hansen T B. Risk of re-admission, reoperation, and mortality within 90 days of total hip and knee arthroplasty in fast-track departments in Denmark from 2005 to 2011. Acta Orthop 2014; 85 (5): 493-500. Glassou E N, Pedersen A B, Hansen T B. Is decreasing mortality in total hip and knee arthroplasty patients dependent on patients’ comorbidity? Acta Orthop 2017; 88 (3): 288-93. Gromov K, Kjærsgaard-Andersen P, Revald P, Kehlet H, Husted H. Feasibility of outpatient total hip and knee arthroplasty in unselected patients. Acta Orthop 2017; 88 (5): 516-21. Halawi M J, Vovos T J, Green C L, Wellman S S, Attarian D E, Bolognesi M P. Patient expectation is the most important predictor of discharge destination after primary total joint arthroplasty. J Arthroplasty 2015; 30 (4): 539-42. Hoorntje A, Koenraadt K L M, Boevé M G, van Geenen R C I. Outpatient unicompartmental knee arthroplasty: Who is afraid of outpatient surgery? Knee Surgery, Sport Traumatol Arthrosc 2017; 25 (3): 759-66. Husted H. Fast-track hip and knee arthroplasty: Clinical and organizational aspects. Acta Orthop 2012; 83 (Suppl 346): 1-39. Husted H, Holm G, Jacobsen S. Predictors of length of stay and patient satisfaction after hip and knee replacement surgery: Fast-track experience in 712 patients. Acta Orthop 2008; 79 (2): 168-73. Husted H, Hansen H C, Holm G, Bach-Dal C, Rud K, Andersen K L, et al. What determines length of stay after total hip and knee arthroplasty? A nationwide study in Denmark. Arch Orthop Trauma Surg 2010a; 130 (2): 263-8. Husted H, Otte K S, Kristensen B B, Orsnes T, Kehlet H. Readmissions after fast-track hip and knee arthroplasty. Arch Orthop Trauma Surg 2010b; 130 (9): 1185-91. Husted H, Lunn T H, Troelsen A, Gaarn-Larsen L, Kristensen B B, Kehlet H. Why still in hospital after fast-track hip and knee arthroplasty? Acta Orthop 2011a; 82 (6): 679-84. Husted H, Troelsen A, Otte K S, Kristensen B B, Holm G, Kehlet H. Fasttrack surgery for bilateral total knee replacement. J Bone Jt Surg Br 2011b; 93 (3): 351-6. Husted H, Jørgensen C C, Gromov K, Troelsen A. Low manipulation prevalence following fast-track total knee arthroplasty. Acta Orthop 2015; 86 (1): 86-91. Husted H, Jørgensen C C, Gromov K, Kehlet H. Does BMI influence hospital stay and morbidity after fast-track hip and knee arthroplasty? Acta Orthop 2016; 87 (5): 466-72. ICD-10. International statistical classification of diseases and related health problems 10th revision [Internet] 2010 [cited 2017 Sep 15]. Available from: http://apps.who.int/classifications/icd10/browse/2010/en Institute for Health and Welfare. Finnish Arthroplasty Register [Internet] 2017 [cited 2017 May 15]. Available from: https://www.thl.fi/far//#data/cphd Jans O, Bandholm T, Kurbegovic S, Solgaard S, Kjaersgaard-Andersen P, Johansson P I, et al. Postoperative anemia and early functional outcomes after fast-track hip arthroplasty: A prospective cohort study. Transfusion 2016; 56 (4): 917-25. Jørgensen C C, Kehlet H, Group LFC for FH and KRC. Fall-related admissions after fast-track total hip and knee arthroplasty: Cause of concern or consequence of success? Clin Interv Aging 2013a; 8: 1569-77. Jørgensen C C, Kehlet H, Group LFC for FH and KRC. Role of patient characteristics for fast-track hip and knee arthroplasty. Br J Anaesth 2013b; 110 (6): 972-80. Jørgensen C C, Kehlet H, Soeballe K, Hansen T B, Husted H, Laursen M B, et al. Time course and reasons for 90-day mortality in fast-track hip and knee arthroplasty. Acta Anaesthesiol Scand 2017; 61 (4): 436-44. Judge A, Chard J, Learmonth I, Dieppe P. The effects of surgical volumes and training centre status on outcomes following total joint replacement: Analysis of the Hospital Episode Statistics for England. J Public Health (Oxf) 2006; 28 (2): 116-24. Keswani A, Tasi M C, Fields A, Lovy A J, Moucha C S, Bozic K J. Discharge destination after total joint arthroplasty: An analysis of postdischarge outcomes, placement risk factors, and recent trends. J Arthroplasty 2016; 31 (6): 1155-62.

20-02-2018 18:02:46

Acta Orthopaedica 2018; 89 (2): 184–189

Khan S K, Malviya A, Muller S D, Carluke I, Partington P F, Emmerson K P, et al. 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. Mäkelä K T, Peltola M, Sund R, Malmivaara A, Häkkinen U, Remes V. Regional and hospital variance in performance of total hip and knee replacements: a national population-based study. Ann Med 2011; 43 (Suppl 1): S31-8. Malviya A, Martin K, Harper I, Muller S D, Emmerson K P, Partington P F, et al. Enhanced recovery program for hip and knee replacement reduces death rate. Acta Orthop 2011; 82 (5): 577-81. Mathijssen N M, Verburg H, van Leeuwen C C, Molenaar T L, Hannink G. Factors influencing length of hospital stay after primary total knee arthroplasty in a fast-track setting. Knee Surg Sports Traumatol Arthrosc 2016; 24 (8): 2692-6. McLawhorn A S, Fu M C, Schairer W W, Sculco P K, MacLean C H, Padgett D E. Continued inpatient care after primary total knee arthroplasty increases 30-day post-discharge complications: A propensity score-adjusted analysis. J Arthroplasty 2017; 32 (9S): 113-18. Mitsuyasu S, Hagihara A, Horiguchi H, Nobutomo K. Relationship between total arthroplasty case volume and patient outcome in an acute care payment system in Japan. J Arthroplasty 2006; 21 (5): 656-63. Pamilo K J, Peltola M, Paloneva J, Mäkelä K, Häkkinen U, Remes V. Hospital volume affects outcome after total knee arthroplasty: A nationwide registry analysis of 80 hospitals and 59,696 replacements. Acta Orthop 2015; 86 (1): 41-7. Pamilo K J, Torkki P, Peltola M, Pesola M, Remes V, Paloneva J. Reduced length of uninterrupted institutional stay after implementing a fast-track protocol for primary total hip replacement. Acta Orthop 2017; Sep 7: 1-7. [Epub ahead of print]

12089 Pamilo D.indd 189

189

Paterson J M, Williams J I, Kreder H J, Mahomed N N, Gunraj N, Wang X, et al. Provider volumes and early outcomes of primary total joint replacement in Ontario. Can J Surgery/Journal Can Chir 2010; 53 (3): 175-83. Pitter F T, Jørgensen C C, Lindberg-Larsen M, Kehlet H. Postoperative morbidity and discharge destinations after fast-track hip and knee arthroplasty in patients older than 85 years. Anesth Analg 2016; 122 (6): 1807-15. Ramkumar P N, Chu C T, Harris J D, Athiviraham A, Harrington M A, White D L, et al. Causes and rates of unplanned readmissions after elective primary total joint arthroplasty: A systematic review and meta-analysis. Am J Orthop 2015; 44 (9): 397-405. Savaridas T, Serrano-Pedraza I, Khan S K, Martin K, Malviya A, Reed M R. Reduced medium-term mortality following primary total hip and knee arthroplasty with an enhanced recovery program. A study of 4,500 consecutive procedures. Acta Orthop 2013; 84 (1): 40-3. Sibia U S, Turcotte J J, MacDonald J H, King P J. The cost of unnecessary hospital days for Medicare joint arthroplasty patients discharging to skilled nursing facilities. J Arthroplasty 2017; 32 (9): 2655-7. Styron J F, Koroukian S M, Klika A K, Barsoum W K. Patient vs provider characteristics impacting hospital lengths of stay after total knee or hip arthroplasty. J Arthroplasty 2011; 26 (8): 1412-18. Sund R. Quality of the Finnish Hospital Discharge Register: A systematic review. Scand J Public Health 2012; 40 (6): 505-15. Wied C, Thomsen M G, Kallemose T, Myhrmann L, Jensen L S, Husted H, et al. The risk of manipulation under anesthesia due to unsatisfactory knee flexion after fast-track total knee arthroplasty. Knee 2015; 22 (5): 419-23. Winther S B, Foss O A, Wik T S, Davis S P, Engdal M, Jessen V, et al. 1-year follow-up of 920 hip and knee arthroplasty patients after implementing fast-track. Acta Orthop 2015; 86 (1): 78-85.

20-02-2018 18:02:46

190

Acta Orthopaedica 2018; 89 (2): 190–196

Migration and clinical outcome of mobile-bearing versus fixed-bearing single-radius total knee arthroplasty A randomized controlled trial Koen T VAN HAMERSVELD 1, Perla J MARANG-VAN DE MHEEN 2, Huub J L VAN DER HEIDE 1, Henrica M J VAN DER LINDEN-VAN DER ZWAAG 1, Edward R VALSTAR 1,3, and Rob G H H NELISSEN 1

1 Department 3 Department

of Orthopaedics, Leiden University Medical Center, Leiden; 2 Medical Decision Making, Leiden University Medical Center, Leiden; of Biomechanical Engineering, Faculty of Mechanical, Maritime, and Materials Engineering, Delft University of Technology, Delft, the

Netherlands Correspondence: ktvanhamersveld@lumc.nl Submitted 2017-06-26. Accepted 2017-11-28.

Background and purpose — Mobile-bearing total knee prostheses (TKPs) were developed in the 1970s in an attempt to increase function and improve implant longevity. However, modern fixedbearing designs like the single-radius TKP may provide similar advantages. We compared tibial component migration measured with radiostereometric analysis (RSA) and clinical outcome of otherwise similarly designed cemented fixed-bearing and mobilebearing single-radius TKPs. Patients and methods — RSA measurements and clinical scores were assessed in 46 randomized patients at baseline, 6 months, 1 year, and annually thereafter up to 6 years postoperatively. A linear mixed-effects model was used to analyze the repeated measurements. Results — Both groups showed comparable migration (p = 0.3), with a mean migration at 6-year follow-up of 0.90 mm (95% CI 0.49–1.41) for the fixed-bearing group compared with 1.22 mm (95% CI 0.75–1.80) for the mobile-bearing group. Clinical outcomes were similar between groups. 1 fixed-bearing knee was revised for aseptic loosening after 6 years and 2 knees (1 in each group) were revised for late infection. 2 knees (1 in each group) were suspected for loosening due to excessive migration. Another mobile-bearing knee was revised after an insert dislocation due to failure of the locking mechanism 6 weeks postoperatively, after which study inclusion was preliminary terminated. Interpretation — Fixed-bearing and mobile-bearing singleradius TKPs showed similar migration. The latter may, however, expose patients to more complex surgical techniques and risks such as insert dislocations inherent to this rotating-platform design. ■

Mobile-bearing total knee prostheses (TKPs) were developed in the late 1970s in an attempt to increase function and improve implant longevity. The bearing was designed to articulate with both a congruent femoral component and a flat non-constrained tibial component, thereby minimizing both contact stresses at the implant–bone interface and polyethylene wear, which should ultimately reduce the occurrence of mechanical loosening (Callaghan et al. 2001, Mahoney et al. 2012). The first—implant developer—long-term survival studies of such designs showed promising high survival rates and good clinical performance (Buechel et al. 2001, Callaghan et al. 2001, Buechel 2002, 2004). Contrarily, no superior results compared with fixed bearings were seen in a number of trials, large registry-based studies and meta-analyses (Pagnano et al. 2004, Namba et al. 2011, Mahoney et al. 2012, van der Voort et al. 2013, Australian Orthopaedic Association National Joint Replacement Registry 2015, Hofstede et al. 2015). Several trials assessing the migration pattern with radiostereometric analysis (RSA) found no superiority of either design on tibial component fixation (Hansson et al. 2005, Henricson et al. 2006, Pijls et al. 2012a, Tjornild et al. 2015) and even questioned whether the mobile bearing truly stays mobile in vivo (Garling et al. 2007). Furthermore, mobile-bearing arthroplasty is considered technically more challenging as less optimal ligament balancing increases the risk of insert dislocations, requiring revision surgery (Cho et al. 2010, Fisher et al. 2011, Namba et al. 2012). Nevertheless, the mobile-bearing design is marketed as an appealing choice for especially young and active patients who demand maximum function and implant longevity (Jolles et al. 2012, Mahoney et al. 2012, Tjornild et al. 2015).

11845 Hamersveld D.indd 190

20-02-2018 12:08:47

Acta Orthopaedica 2018; 89 (2): 190–196

Over time, modern TKPs have substantially improved in design, quality of materials (particularly the polyethylene) and fixation methods. In contrast to most conventional designs that have several axes of femoral rotation during flexion, the femoral component of the ‘single-radius’ TKP rotates about a single axis and should thereby reduce contact stress (Molt et al. 2012, Wolterbeek et al. 2012). The fixed-bearing variant of this single-radius design allows for some axial rotation during deep flexion with minimal constraint forces (Molt et al. 2012). Thus, the theoretical advantages of this fixed-bearing single-radius design might come close to the concepts of mobile-bearing designs, but without the associated risks like insert dislocations. There are to our knowledge no studies comparing mobilebearing and fixed-bearing single-radius TKPs, except for a previous report on 1-year migration and kinematics on the first 20 patients of this trial (Wolterbeek et al. 2012). We now present medium-term follow-up results of all included patients and compare tibial component migration and clinical outcomes of similarly designed mobile-bearing and fixed-bearing cemented single-radius TKPs.

Patients and methods This randomized controlled trial was conducted at the Leiden University Medical Center (an academic tertiary referral center) between April 2008 and February 2010. Patients received either mobile-bearing or fixed-bearing components of an otherwise similarly designed cemented posterior stabilized Triathlon TKP (Stryker, Mahwah, NJ, USA). The rotating-platform mobile-bearing design additionally has a locking O-ring, which allows axial rotation about a central post (Wolterbeek et al. 2012). The arthroplasties were performed by three experienced knee surgeons or under their direct supervision, using the appropriate guidance instruments following the manufacturer’s instructions. In all patients, the components were cemented first, after which the insert was mounted. Pulsatile lavage of the osseous surface was undertaken before applying bone cement (Palacos R cement, Heraeus-Kulzer GmbH, Hanau, Germany). For more details regarding patients, randomization and prostheses, see Wolterbeek et al. (2012). Follow-up Baseline characteristics, including the Knee Society Score (KSS) and hip–knee–ankle angle (HKA) measurements (with varus < 180°) were assessed 1 week before surgery. Postoperative evaluations including RSA radiographs were performed the first or second day after surgery, before weight bearing. Subsequent RSA and clinical examinations including KSS scores were scheduled at 6 months, 1 year and annually thereafter. HKA measurements were repeated at the 1-year followup.

11845 Hamersveld D.indd 191

191

Radiostereometric analysis To accurately measure tibial component migration, radiostereometric analysis measurements were performed according to the RSA guidelines (ISO 16087:2013(E) 2013). At each examination, the patient was in a supine position with the calibration cage (Carbon Box, Leiden, The Netherlands) under the table in a uniplanar setup. Migration was analyzed using Model-based RSA, version 4 (RSAcore, LUMC, Leiden, the Netherlands). Positive directions along and about the orthogonal axes are: medial on transverse (x-)axis, cranial on longitudinal (y-)axis and anterior on sagittal (z-)axis for translations and anterior tilt (x-axis), internal rotation (y-axis) and valgus tilt (z-axis) for rotations (Valstar et al. 2005). The maximum total point motion (MTPM), which is the length of the translation vector of the point on the tibial component that has moved most, was defined as the primary outcome. Sample size RSA measurement error of less than 0.5 mm was expected (Valstar et al. 2005). If the true difference in MTPM between fixed-bearing and mobile-bearing TKPs is 0.5 mm, 17 patients were required to detect this difference with alpha 0.05 and power 0.80. To account for loss to follow-up, the intention was to randomize 20 patients to each group. Statistics The original primary endpoint (Wolterbeek et al. 2012) was registered as a difference in MTPM between groups after 1-year follow-up on the first 20 enrolled patients. For this medium-term follow-up analysis, we changed the primary endpoint—prior to data analysis—to a difference in MTPM between groups of all included patients after 6 years of followup, as 6-year data were available at the time of data analysis. To provide unbiased comparisons between groups, the main approach to analyze the results was the intention-to-treat analysis (groups according to allocation). In case of switches between groups so that patients were not treated as randomized, thereby diluting the treatment effect, an as-treated analysis (groups according to received type of prosthesis) was also performed. The first postoperative radiographs were taken as reference for the migration measurements. We used repeated measures analysis of variance with a linear mixed-effects model to analyze the migration measurements. This is the recommended technique to model repeated measurements as it takes the correlation of measurements performed on the same subject into account and includes all patients in the analysis while dealing effectively with missing values (DeSouza et al. 2009, Ranstam et al. 2012, Nieuwenhuijse et al. 2013). The difference in migration between groups is only tested once after 6-year follow-up to safeguard against multiple testing and is modelled as a function of time and the interaction of time with type of prosthesis (fixed effects). A random-intercepts term is used (random effect) and remaining variability is modelled with

20-02-2018 12:08:47

192

Acta Orthopaedica 2018; 89 (2): 190–196

ENROLLMENT

Randomized (n = 52 TKPs) Excluded (n = 6): – FB with insufficient amount of markers, 3 – MB with insufficient amount of markers, 3

ment. This study was partially funded by a single unrestricted grant from Stryker. The sponsor did not take any part in the design, conduct, analysis, and interpretations stated in the final manuscript.

Results ALLOCATION Allocated to ﬁxed bearing (n = 23): – received allocated treatment, 23

Allocated to mobile bearing (n = 23): – received allocated treatment, 18 – received ﬁxed-bearing TKPs, 5

FOLLOW-UP Lost to follow-up (n = 11): (intention-to-treat) – 1 revised after 3 years (infection) – 3 died after 0.5, 3 and 5 years – 7 withdrew after 1, 3, 4 and 4 after 5 years

Lost to follow-up (n = 8): (intention-to-treat) – 3 were revised after: 5 weeks (insert dislocation) 1 year (infection) 6 years (aseptic loosening, received FB, 6-year RSA images were made) – 2 died after 4 and 5 years – 1 withdrew after 2 years – 2 refused 6-year examination

ANALYSIS Analyzed: at 0.5, 1, 2, 3, 4, 5, 6 years n = 23, 22, 21, 21, 19, 18, 12

52 knees were eligible in 48 patients (Figure 1). 6 patients (3 of both groups) were excluded due to an insufficient number of bone markers placed in the proximal tibia, resulting in unmeasurable RSA images. Thus 23 fixed-bearing and 23 mobile-bearing TKPs could be used in the intention-to-treat analysis. During the 6-year followup, 5 patients died, 4 revisions were performed (see below), 1 patient withdrew dissatisfied with his knee function, and 9 patients withdrew or refused to visit the clinic for reasons not related to the knee prosthesis. This resulted in 299 valid RSA radiographs used for the migration analysis. Baseline characteristics did not differ between groups (Table 1).

Analyzed: at 0.5, 1, 2, 3, 4, 5, 6 years n = 22, 22, 21, 20, 20, 19, 16

RSA and clinical outcomes The precision of RSA measurements was assessed with 34 double examinations (Table Figure 1. CONSORT flow diagram. FB = fixed-bearing, MB = mobile-bearing, TKPs = total knee prostheses. 2). There were no statistically significant differences in mean migration between groups during 6 years of follow-up (Figure 2 and Table 4, see a heterogeneous autoregressive order 1 covariance structure. For revised and lost cases, RSA measurements were included Supplementary data). Migration remained similar between in the analysis up to the last follow-up. MTPM was log-trans- groups when excluding five components with high migration formed during statistical modelling as it was not normally dis- profiles (Figure 2). tributed. The secondary (clinical) outcomes, namely KSS scores, flexion, and extension, were analyzed with a similar linear Table 1. Baseline demographic characteristics. Values are mean mixed-effects model. The standard errors of KSS knee score (SD) unless otherwise indicated and extension were corrected via the sandwich estimator using Fixed bearing Mobile bearing a generalized estimating equations approach, as these outcome Outcome (n = 23 TKPs) (n =23 TKPs) measures were not normally distributed and a log-transformation did not result in a normal distribution. To illustrate the Age 68.0 (9.6) 67.5 (10.1) Body mass index 30.1 (6.2) 29.8 (6.2) directions of migration, descriptive data of the translations and Female sex, n 16 19 rotations along and about the orthogonal axes are presented Diagnosis, n but not tested for significance. Osteoarthritis 17 13 Rheumatoid arthritis 5 10 IBM SPSS Statistics 23.0 (IBM Corp, Armonk, NY, USA) Hemophilic arthropathy 1 0 was used for all analyses, and significance was set at p < 0.05. Ethics, registration, funding, and potential conflicts of interest The trial was performed in compliance with the Declaration of Helsinki and Good Clinical Practice guidelines, and approved by the local ethics committee prior to enrollment (entry no. P07.205, retrospectively registered at ClinicalTrials.gov, NCT02924961). All patients gave informed consent. Reporting of the trial was in accordance with the CONSORT state-

11845 Hamersveld D.indd 192

ASA classification, n I II III Hip–knee–ankle angle Preoperative Postoperative Knee Society Score Knee Score Function Score

3 17 3

2 15 6

177 (6) 178 (4)

180 (8) 178 (4)

49.3 (8.9) 45.7 (22.6)

47.2 (18.3) 35.9 (21.8)

ASA: American Society of Anesthesiologists

20-02-2018 12:08:47

Acta Orthopaedica 2018; 89 (2): 190â&#x20AC;&#x201C;196

193

Table 2. Precision of RSA measurements (upper limits of the 95% CI around zero motion) Tibial component

Transverse

Longitudinal

Sagittal

Translation (mm) Rotation (Â°)

0.05 0.21

0.04 0.45

0.14 0.11

Both groups showed comparable translations and rotations along and around the 3 orthogonal axes, and high migration of individual components was seen in almost any direction (Figure 3). 5 components showed excessive migration (Figure 2 and Figure 3), of which 2 were revised for septic loosening (late infections of a mobile-bearing knee with

Figure 2. Mean maximum total point motion and 95% CI for the groups alone (top) and mean and 95% CI for the groups with solid red lines for the revised components and dashed red lines for the components suspected for loosening excluded from the groups (bottom). One component revised due to a mobile-bearing insert dislocation is not shown separately, as this complication occurred before 6 months of follow-up. *Analyzed as mobile-bearing TKP in intention-to-treat analysis but received fixed-bearing TKP. LFU = lost to follow-up.

11845 Hamersveld D.indd 193

Staphylococcus aureus after 1 year and a fixed-bearing with a Candida albicans after 3 years) and 1 fixed-bearing (randomized in the mobile-bearing group) was revised for aseptic loosening after 6 years (Table 3 #35, see Supplementary data). The other 2 were suspected for aseptic loosening of which 1 mobile-bearing knee was postponed for revision surgery (Figure 4, see Supplementary data) and 1 fixed-bearing, placed in an 81-year-old female with osteoarthritis, was lost to follow-up after 1 year. This patient visited the outpatient clinic after 6 years of follow-up with severe knee complaints, show-

Figure 3. Descriptive data showing the translations in mm (left side) and rotations in degrees (right side) of the transverse axis (top), longitudinal axis (middle) and sagittal axis (bottom) for both groups (mean and 95% CI). Similar to Figure 2, the revised components (solid red lines) and the 2 components suspected for loosening (dashed red lines) are drawn separately.

20-02-2018 12:08:47

194

ing a progressive varus alignment of the tibial component (HKA 174° at 1 year versus 168° at the 6-year follow-up), but refused further RSA examinations and treatment (other than a knee brace) due to age and comorbidities. The secondary outcome scores (KSS scores, flexion, and extension) showed no statistical differences in improvement over time between the two groups (Table 5, see Supplementary data). Adverse events Besides the 5 components with excessive migration already stated, 1 patient withdrew due to dissatisfaction. This 47-yearold man with secondary osteoarthritis due to hemophilic arthropathy had a preoperative knee flexion of 85° and a flexion contracture of 15°; postoperatively, his knee flexion did not improve after receiving a fixed-bearing design. 1 mobilebearing knee was revised due to an insert dislocation, which occurred 5 weeks after surgery (Figure 5, see Supplementary data). Dislocation of a Stryker mobile bearing was not described in the literature at that time and thus necessitated thorough investigations. Patient inclusion was put on hold until the manufacturer had evaluated the reason for this insert dislocation. Incorrect intraoperative mounting of the insert on the tibial post possibly damaged the tibial insert locking mechanism, although the exact cause of the failed locking mechanism remains unclear. For this reason, patient recruitment of this study was stopped preliminarily after 18 out of the intended 20 mobile-bearing TKPs were implanted. As-treated analysis Intraoperatively, 1 of the surgeons (who performed 37 of the study procedures) deemed 5 knees unsuitable for the allocated mobile-bearing insert and fixed-bearing components were used instead. The as-treated population therefore included 28 fixed-bearing and 18 mobile-bearing TKPs (see Figure 1). The reasons for the deviations and the outcome in these patients are given in Table 3 (see Supplementary data). All primary and secondary outcome results were comparable in the as-treated analysis and subsequently did not alter conclusions (Tables 4–5, see Supplementary data).

Discussion While migration measured by RSA and clinical outcomes of mobile-bearing and fixed-bearing designs of the single-radius TKP were comparable after 6 years, some of the complications experienced are inherent to the mobile-bearing design. In 5 cases, suboptimal gap balancing during mobile-bearing surgery resulted in the decision to switch to fixed-bearing TKPs, as is recommended in the literature (Bhan and Malhotra 2003). Especially if bone resections and soft-tissue releases are performed conservatively in cases with compromised (peri-)articular tissue, insertion of the mobile bearing onto the central post of the baseplate in a perpendicular verti-

11845 Hamersveld D.indd 194

Acta Orthopaedica 2018; 89 (2): 190–196

cal manner can be technically challenging. Forcing the insert onto the post from a different angle can damage the locking mechanism, which possibly occurred in 1 procedure and, if so, instigated an insert dislocation necessitating revision surgery. Several explanations have been suggested for the discrepancies between the theoretically expected superior outcome and actual clinical results of mobile-bearing TKPs. First, it is questionable whether the mobile-bearing component truly is mobile in vivo. Garling et al. (2007) performed a fluoroscopic study using a different rotating-platform TKP (NexGen LPS, Zimmer Biomet, Winterthur, Switzerland) and found limited rotation of the mobile bearing. Among other explanations, the authors hypothesized that this might be caused by (1) polyethylene-on-metal impingement due to a mismatch of the location of the fixed pivot point in the rotating-platform design and the actual tibiofemoral rotation point, or (2) due to fibrous tissue formation between the mobile bearing and the baseplate (Garling et al. 2007). However, in a previous report on a subset of our study population (Wolterbeek et al. 2012), kinematic analysis with step-up and lunge motions showed that overall the mobile-bearing insert followed the femoral component movement as intended by its design, but not in all patients. Second, dislocation of the mobile bearing is a serious complication requiring revision surgery. Historically, this complication was mainly seen in the old mobile meniscalbearing designs (Namba et al. 2011), while insert dislocations in rotating-platform designs are rare nowadays (Huang et al. 2002, Thompson et al. 2004, Fisher et al. 2011). At the time (2008–2010) of patient inclusion for the current study, there were no reports on dislocation of the mobile-bearing insert with similar locking mechanisms as used in the Triathlon TKP. Thus our study was stopped awaiting results of thorough investigations. A case report on a bearing dislocation was later reported, describing failure of the locking O-ring identical to the Triathlon locking mechanism (Kobayashi et al. 2011). Testing the mode of failure during revision surgery in our case resulted in similar conclusions: once the O-ring of the insert has been damaged, flexing the knee can lead to liftoff and anterior dislocation of the insert. This was most easily observed while testing the knee intraoperatively with external rotation force. Third, several authors have addressed the effect of surgical procedure volumes, with superior results being attained by high-volume centers (Baker et al. 2013, Critchley et al. 2012, Lau et al. 2012, Liddle et al. 2016). Good clinical results reported in single-surgeon series may not be realized in low-volume centers or centers treating patients with diverse demographic factors (Namba et al. 2012). In our academic center, all participating surgeons were experienced in performing both mobile-bearing and fixed-bearing total knee arthroplasties and often performed surgery in patients with secondary osteoarthritis due to rheumatoid arthritis and other inflammatory diseases, which was also the case in a high proportion of the included patients. Nevertheless, the number of adverse events observed in this study was much higher than

20-02-2018 12:08:49

Acta Orthopaedica 2018; 89 (2): 190–196

reported in other clinical (RSA) studies performed in our center. Although this could be due to chance, a learning-curve effect with this new design may have contributed to some of the complications and intraoperative decisions to deviate from the randomized treatment allocation. A limitation of this study is that patient inclusion was prematurely terminated for patient safety after the mobilebearing dislocation, before reaching the intended 20 patients in this study arm. This did not compromise the number of patients needed to have sufficient power on the primary outcome in the first 5 years of follow-up, as only 17 patients were required according to the sample size calculation. This was not the case at 6 years (with less than 17 TKPs available for analysis in both groups). However, as the patients lost in the sixth postoperative year had stable migration patterns, it is unlikely that migration at 6 years would substantially differ from the pattern depicted in Figure 2. Contrarily, results of the clinical outcomes should be interpreted with caution, given the lower accuracy and precision of these measurements. However, large meta-analysis studies comparing mobilebearing with fixed-bearing TKPs found no differences in clinical outcomes either (van der Voort et al. 2013, Hofstede et al. 2015). Another limitation is the duration of follow-up. Although early tibial component migration measured through RSA is a proven predictor of late loosening (Ryd et al. 1995, Pijls et al. 2012b), one can hypothesize about various mechanisms affecting migratory patterns at different time intervals. However, results of an RSA study with long-term follow-up (> 10 years) revealed no changes in migration patterns of mobile-bearing and fixed-bearing prostheses after the first 2 years (Pijls et al. 2012a). In summary, fixed-bearing single-radius TKPs showed similar migration compared with the mobile-bearing TKPs, while the latter may expose patients to more complex surgical techniques and risks such as insert dislocations inherent to this rotating-platform design. Supplementary data Tables 3–5 and Figures 4 and 5 and are available as supplementary data in the online version of this article, http://dx.doi. org/10.1080/17453674.2018.1429108

The study was designed by EV and RN. Surgeries were performed by HH, HL, and RN. Data collection and RSA analysis were performed by KH. Statistical analysis was done by KH and PM. KH, PM, EV, and RN interpreted the data and wrote the initial draft manuscript. KH, PM, HH, HL, and RN critically revised and approved the manuscript.

Acta thanks Anders Henricson and Kaj Knutson for help with peer review of this study. Australian Orthopaedic Association National Joint Replacement Registry. Annual Report 2015. Adelaide: AOA; 2015.

11845 Hamersveld D.indd 195

195

Baker P, Jameson S, Critchley R, Reed M, Gregg P, Deehan D. Center and surgeon volume influence the revision rate following unicondylar knee replacement: an analysis of. 23,400 medial cemented unicondylar knee replacements. J Bone Joint Surg Am. 2013;95(8):702-9. Bhan S, Malhotra R. Results of rotating-platform, low-contact-stress knee prosthesis. J Arthroplasty 2003; 18(8): 1016-22. Buechel F F, Sr. Long-term followup after mobile-bearing total knee replacement. Clin Orthop Relat Res 2002; (404): 40-50. Buechel F F, Sr. Mobile-bearing knee arthroplasty: rotation is our salvation! J Arthroplasty 2004; 19(4 Suppl 1): 27-30. Buechel F F, Sr, Buechel F F, Jr, Pappas M J, D’Alessio J. Twenty-year evaluation of meniscal bearing and rotating platform knee replacements. Clin Orthop Relat Res 2001; (388): 41-50. Callaghan J J, Insall J N, Greenwald A S, Dennis D A, Komistek R D, Murray D W, Bourne R B, Rorabeck C H, Dorr L D. Mobile-bearing knee replacement: concepts and results. Instr Course Lect 2001; 50: 431-49. Cho W S, Youm Y S, Ahn S C, Sohn D W. What have we learned from LCS mobile-bearing knee system? Knee Surg Sports Traumatol Arthrosc 2010; 18(10): 1345-51. Critchley R J, Baker P N, Deehan D J. Does surgical volume affect outcome after primary and revision knee arthroplasty? A systematic review of the literature. Knee 2012; 19(5): 513-8. DeSouza C M, Legedza A T, Sankoh A J. An overview of practical approaches for handling missing data in clinical trials. J Biopharm Stat 2009; 19(6): 1055-73. Fisher D A, Bernasek T L, Puri R D, Burgess M L. Rotating platform spinouts with cruciate-retaining mobile-bearing knees. J Arthroplasty 2011; 26(6): 877-82. Garling E H, Kaptein B L, Nelissen R G, Valstar E R. Limited rotation of the mobile-bearing in a rotating platform total knee prosthesis. J Biomech 2007; 40 Suppl 1: S25-30. Hansson U, Toksvig-Larsen S, Jorn L P, Ryd L. Mobile vs. fixed meniscal bearing in total knee replacement: a randomised radiostereometric study. Knee 2005; 12(6): 414-18. Henricson A, Dalen T, Nilsson K G. Mobile bearings do not improve fixation in cemented total knee arthroplasty. Clin Orthop Relat Res 2006; 448: 114-21. Hofstede S N, Nouta K A, Jacobs W, van Hooff M L, Wymenga A B, Pijls B G, Nelissen R G, Marang-van de Mheen P J. Mobile bearing vs fixed bearing prostheses for posterior cruciate retaining total knee arthroplasty for postoperative functional status in patients with osteoarthritis and rheumatoid arthritis. Cochrane Database Syst Rev 2015; 2: CD003130. Huang C H, Ma H M, Liau J J, Ho F Y, Cheng C K. Late dislocation of rotating platform in New Jersey Low-Contact Stress knee prosthesis. Clin Orthop Relat Res 2002; (405): 189-94. ISO 16087:2013(E). Implants for surgery—roentgen stereophotogrammetric analysis for the assessment of migration of orthopaedic implants. Geneva, Switzerland: International Organization for Standardization; 2013. Jolles B M, Grzesiak A, Eudier A, Dejnabadi H, Voracek C, Pichonnaz C, Aminian K, Martin E. A randomised controlled clinical trial and gait analysis of fixed- and mobile-bearing total knee replacements with a five-year follow-up. J Bone Joint Surg Br 2012; 94(5): 648-55. Kobayashi H, Akamatsu Y, Taki N, Ota H, Mitsugi N, Saito T. Spontaneous dislocation of a mobile-bearing polyethylene insert after posterior-stabilized rotating platform total knee arthroplasty: a case report. Knee 2011; 18(6): 496-8. Lau R L, Perruccio A V, Gandhi R, Mahomed N N. The role of surgeon volume on patient outcome in total knee arthroplasty: a systematic review of the literature. BMC Musculoskelet Disord 2012; 13: 250. Liddle A D, Pandit H, Judge A, Murray D W. Effect of surgical caseload on revision rate following total and unicompartmental knee replacement. J Bone Joint Surg Am 2016; 98(1): 1-8. Mahoney O M, Kinsey T L, D’Errico T J, Shen J. The John Insall Award: no functional advantage of a mobile bearing posterior stabilized TKA. Clin Orthop Relat Res 2012; 470(1): 33-44.

20-02-2018 12:08:49

196

Molt M, Ljung P, Toksvig-Larsen S. Does a new knee design perform as well as the design it replaces? Bone Joint Res 2012; 1(12): 315-23. Namba R S, Inacio MC , Paxton E W, Robertsson O, Graves S E. The role of registry data in the evaluation of mobile-bearing total knee arthroplasty. J Bone Joint Surg Am. 2011; 93(Suppl 3): 48-50. Namba R S, Inacio M C, Paxton E W, Ake C F, Wang C, Gross T P, MarinacDabic D, Sedrakyan A. Risk of revision for fixed versus mobile-bearing primary total knee replacements. J Bone Joint Surg Am 2012; 94(21): 1929-35. Nieuwenhuijse M J, van der Voort P, Kaptein B L, van der Linden-van der Zwaag H M, Valstar E R, Nelissen R G. Fixation of high-flexion total knee prostheses: five-year follow-up results of a four-arm randomized controlled clinical and roentgen stereophotogrammetric analysis study. J Bone Joint Surg Am 2013; 95(19): e1411-11. Pagnano M W, Trousdale R T, Stuart M J, Hanssen A D, Jacofsky D J. Rotating platform knees did not improve patellar tracking: a prospective, randomized study of 240 primary total knee arthroplasties. Clin Orthop Relat Res 2004; (428) :221-7. Pijls B G, Valstar E R, Kaptein B L, Nelissen R G. Differences in long-term fixation between mobile-bearing and fixed-bearing knee prostheses at ten to 12 yearsâ&#x20AC;&#x2122; follow-up: a single-blinded randomised controlled radiostereometric trial. J Bone Joint Surg Br 2012a;94(10):1366-71. Pijls B G, Valstar E R, Nouta KA, Plevier JW, Fiocco M, Middeldorp S, Nelissen RG. Early migration of tibial components is associated with late revision: a systematic review and meta-analysis of 21,000 knee arthroplasties. Acta Orthop 2012b; 83(6): 614-24.

11845 Hamersveld D.indd 196

Acta Orthopaedica 2018; 89 (2): 190â&#x20AC;&#x201C;196

Ranstam J, Turkiewicz A, Boonen S, Van Meirhaeghe J, Bastian L, Wardlaw D. Alternative analyses for handling incomplete follow-up in the intentionto-treat analysis: the randomized controlled trial of balloon kyphoplasty versus non-surgical care for vertebral compression fracture (FREE). BMC Med Res Methodol 2012; 12: 35. Ryd L, Albrektsson B E, Carlsson L, Dansgard F, Herberts P, Lindstrand A, Regner L, Toksvig-Larsen S. Roentgen stereophotogrammetric analysis as a predictor of mechanical loosening of knee prostheses. J Bone Joint Surg Br 1995; 77(3): 377-83. Thompson N W, Wilson D S, Cran G W, Beverland D E, Stiehl J B. Dislocation of the rotating platform after low contact stress total knee arthroplasty. Clin Orthop Relat Res 2004; (425): 207-11. Tjornild M, Soballe K, Hansen P M, Holm C, Stilling M. Mobile- vs. fixedbearing total knee replacement. Acta Orthop 2015; 86(2): 208-14. Valstar E R, Gill R, Ryd L, Flivik G, Borlin N, Karrholm J. Guidelines for standardization of radiostereometry (RSA) of implants. Acta Orthop 2005; 76(4): 563-72. van der Voort P, Pijls B G, Nouta K A, Valstar E R, Jacobs W C, Nelissen R G. A systematic review and meta-regression of mobile-bearing versus fixed-bearing total knee replacement in 41 studies. Bone Joint J 2013; 95-B (9): 1209-16. Wolterbeek N, Garling E H, Mertens B J, Nelissen R G, Valstar E R. Kinematics and early migration in single-radius mobile- and fixed-bearing total knee prostheses. Clin Biomech (Bristol, Avon) 2012; 27(4): 398-402.

20-02-2018 12:08:49

Acta Orthopaedica 2018; 89 (2): 197–203

197

High occurrence of osteoarthritic histopathological features unaccounted for by traditional scoring systems in lateral femoral condyles from total knee arthroplasty patients with varus alignment Venkata P MANTRIPRAGADA 1, Nicolas S PIUZZI 1,2,4, Terri ZACHOS 2, Nancy A OBUCHOWSKI 3, George F MUSCHLER 1,2, and Ronald J MIDURA 1

1 Department of Biomedical Engineering, Lerner Research Institute, Cleveland Clinic, Cleveland, USA; 2 Department of Orthopedic Surgery, Cleveland Clinic, Cleveland, USA; 3 Department of Quantitative Health Science, Cleveland Clinic, Cleveland, USA; 4 Instituto Universitario del Hospital Italiano de Buenos Aires, Buenos Aires, Argentina Correspondence: muschlg@ccf.org Submitted 2017-05-31. Accepted 2017-10-10.

Background and purpose — A better understanding of the patterns and variation in initiation and progression of osteoarthritis (OA) in the knee may influence the design of therapies to prevent or slow disease progression. By studying cartilage from the human lateral femoral condyle (LFC), we aimed to: (1) assess specimen distribution into early, mild, moderate, and severe OA as per the established histopathological scoring systems (HHGS and OARSI); and (2) evaluate whether these 2 scoring systems provide sufficient tools for characterizing all the features and variation in patterns of OA. Patients and methods — 2 LFC osteochondral specimens (4 x 4 x 8 mm) were collected from 50 patients with idiopathic OA varus knee and radiographically preserved lateral compartment joint space undergoing total knee arthroplasty. These were fixed, sectioned, and stained with HE and Safranin O-Fast Green (SafO). Results — The histopathological OA severity distribution of the 100 specimens was: 6 early, 62 mild, 30 moderate, and 2 severe. Overall, 45/100 specimens were successfully scored by both HHGS and OARSI: 12 displayed low OA score and 33 displayed cartilage surface changes associated with other histopathological features. However, 55/100 samples exhibited low surface structure scores, but were deemed to be inadequately scored by HHGS and OARSI because of anomalous features in the deeper zones not accounted for by these systems: 27 exhibited both SafO and tidemark abnormal features, 16 exhibited only SafO abnormal features, and 12 exhibited tidemark abnormal features. Interpretation — LFC specimens were scored as mild to moderate OA by HHGS and OARSI. Yet, several specimens exhibited deep zone anomalies while maintaining good surface structure,

inconsistent with mild OA. Overall, a better classification of these anomalous histopathological features could help better understand idiopathic OA and potentially recognize different subgroups of disease. ■

Idiopathic osteoarthritis (OA) was initially perceived as a mechanical “wear and tear” process of articular cartilage that increased with age. Now, OA is recognized as whole-organ disease process influenced by multiple factors including body weight, diet, joint stability, alignment, joint and meniscus shape, and associated changes in local and systemic inflammatory mediators, genetic factors, and innate immunity (Greene and Loeser 2015, Orlowsky and Kraus 2015, Malfait 2016, Varady and Grodzinsky 2016). A better understanding of the patterns and variation in initiation and progression of OA in the knee could influence the design and patient-specific selection of therapies to prevent or slow the progression of OA. The key challenge encountered in this process is our incomplete understanding of the underlying disease mechanism (Li et al. 2013). Although the order in which different OA phenomena occur is unclear, some of these changes can be readily identified by 2 validated cartilage histopathological scoring assessments (Rutgers et al. 2010): (1) the histologic/histochemical grading system (HHGS) (Mankin et al. 1971) and (2) the advanced Osteoarthritis Cartilage Histopathology Assessment System (OARSI) (Pritzker et al. 2006).

11773 Mantripragada D.indd 197

21-02-2018 15:41:32

198

Acta Orthopaedica 2018; 89 (2): 197–203

Figure 1. Sample of lateral femoral condyle (black star) (right knee shown in example) was obtained (A, B). The orientation of the condyle was noted in the operating room (A = anterior; P = posterior; M = medial; L = lateral) and placed in an in-house fabricated miter box in AP orientation and cut into 4 mm thick arches through the region of the femur that is weight bearing in extension (C). The central region of 1 arch was cut into 5 specimens (4 mm x 4 mm), the second (Lateral) and fourth (Medial) were processed for histology (D). Each specimen was paraffin embedded, sectioned, and stained with HE (E), and SafO-FG (F) for analysis.

Previous studies have demonstrated that varus knee OA subjects have higher loading on the medial than lateral compartment, with joint space width (JSW) preserved in the latter (Kumar et al. 2013). A reduced joint load in people with knee OA is related to a slower progression of the disease (Sritharan et al. 2017). Thus, lower loads in the lateral compartment may result in slower OA progression at this site. Based on these observations, we designed this study to obtain cartilage specimens from the lateral femoral condyle (LFC) from patients undergoing total knee arthroplasty (TKA) with idiopathic OA varus knee, which presented relatively preserved JSW, with 2 aims: (1) to assess specimen distribution into early, mild, moderate, and severe OA as per the established histopathological scoring systems (HHGS and OARSI); and (2) to evaluate whether the HHGS and OARSI scoring systems provide sufficient tools for characterizing all the features and variation in patterns of primary OA progression.

Methods 50 patients (30 men) with a mean age of 63 (37–80) years and mean BMI of 31 (18–49) with varus knees scheduled for TKA were recruited after getting their informed consent. Inclusion criteria required patients with a diagnosis of idiopathic OA (primarily medial compartment and/or patellofemoral disease) exhibiting a relatively preserved lateral compartment (JSW: 2–10 mm, median: 6 mm in the lateral compartment) based on preoperative weight-bearing anterior-posterior radiographs taken in full extension and 30° of flexion. Patients were excluded if they had secondary arthritis related to systemic inflammatory arthritis (e.g. rheumatoid arthritis, psoriatic arthritis); history of autoimmune disorders, gout or pseudogout, previous surgery to the index knee, current or previous treatment with systemic glucocorticoids or osteotropic medication; current treatment or treatment of cancer within previous 2 years; known or suspected infection; and osteonecrosis.

11773 Mantripragada D.indd 198

Cartilage procurement During TKA, the LFC was collected after making the distal femoral cut and the AP orientation was noted. All included LFC specimens presented with Grade 0 (normal), Grade I (cartilage with softening and swelling), or Grade II (a partial-thickness defect with fissures on the surface that do not reach subchondral bone (SCB) or exceed 1.5 cm in diameter) macroscopic Outerbridge classifications (Outerbridge 1961). 2 osteochondral specimens (4 × 4 × 8 mm) were prepared by placing the condyle in an in-house fabricated miter box in AP orientation. The miter box had evenly spaced slots, 4 mm apart on the top edge and cartilage arches were cut using a razor blade from the weight-bearing center portion of the LFC; 1 was located medial (L1) and 1 lateral (L2) to the LFC midline (Figure 1). The centers of these two samples were separated by 10 mm. Histological sample processing and digital imaging 100 osteochondral specimens were processed from 50 patients. Immediately after surgical retrieval, specimens were fixed for 48 h at 4 °C (Scott 1989). 5µ thick paraffin sections were cut and stained with freshly prepared HE or SafraninO and fast green (SafO). The embedded tissue was cut in the plane perpendicular to the surface of the cartilage to obtain a representative overview of the tissue structure and thickness (Figure 1). Two adjacent sections per stain were digitally imaged at 10x and used for scoring using HHGS and advanced OARSI systems. Disease severity distribution for all specimens HHGS and OARSI score distribution was divided into 5 equal bins to classify specimens into early, mild, moderate, and severe OA (Figure 2, dotted lines). Percentage of specimens indicating features unaccounted for by either scoring systems We classified samples as presenting unaccounted for features if

21-02-2018 15:41:32

Acta Orthopaedica 2018; 89 (2): 197–203

199

Results

Figure 2. Correlation plot of average HHGS versus average OARSI scores for specimens from lateral and medial locations. Moderate correlation was observed between the 2 systems (Spearman’s coefficient = 0.806). The total HHGS and OARSI scores were divided into 5 equal bins as indicated by the dotted lines, and classified as early, mild, moderate, and severe to assess specimen distribution. The shaded regions on the bottom and left represent the regions containing the 50% of samples with scores below the median for each scoring system (median HHGS = 4.8, OARSI = 5.5) where correlation between the systems is less robust. For illustration, the trend is represented by a red line from a fitted linear regression of OARSI on HHGS.

the samples presented with low structure score (HHGS structure = 1), but had substantial changes in SafO staining (HHGS SafO score > 2) and tidemark features (HHGS tidemark score = 0 or 1). Of the specimens binned by the aforementioned criteria, in order to quantify the percentage of samples presenting with the unaccounted for SafO and tidemark histological features, we systematically filtered samples as per the above conditions, and then individually classified these samples into yes/no decisions (yes = shows at least 1 of the unaccounted for features; no = does not show any features) based on at least 2 of the 3 reviewers’ agreement. Statistics To assess the association between histopathology scores and subjects’ age and sex, generalized linear models were fit, with the histopathology score as the dependent variable and age and sex as independent variables. Spearman’s correlation coefficients were estimated to assess the association of HHGS and OARSI mean total scores. 95% confidence intervals (CI) were constructed using Fisher’s z-transformation (for each side separately) or the percentile bootstrap (when lateral and medial sides were pooled). Ethics, funding, and potential conflicts of interest This study was approved by the Institutional Review Board committee of the Cleveland Clinic (Protocol:13641) and was supported by a National Institute of Health grant (R01AR063733) awarded to GFM. The authors declare no conflict of interest.

11773 Mantripragada D.indd 199

Disease severity distribution The distribution of histopathology severity exhibited by this LFC sample cohort appeared strongly skewed to earlier stages of OA, though none were scored as normal or as severe OA (Figure 2). It was noted that 6% of our samples were considered early, 62% were mild, 30% were moderate, and 2% were binned as severe OA. The histopathology scores were distributed similarly across age and sex. When the scores from all the readers were averaged, the Spearman’s coefficient was moderate (0.8, CI:0.8, 0.9) (Figure 2). The correlation between HHGS and OARSI scores below the median values (4.8 and 5.5, respectively) was calculated to be weaker than the full range. Features unaccounted for by either scoring system Using both HHGS and OARSI histopathological scoring systems, 45/100 specimens were categorized adequately: 12/100 displayed low scores (structure = 1) and 33/100 displayed surface degradation (structure > 2) along with other histopathological changes (Figure 3). However, 55/100 specimens were inadequately scored by either of the 2 histopathological systems, especially in regard to SafO staining and tidemark changes, which were not associated with substantial surface changes (structure = 1 or less). Of the 55 specimens, 27/55 displayed both abnormal SafO and tidemark features, 16/55 displayed abnormal SafO features only, and 12/55 displayed abnormal tidemark features only. For instance, Figure 4A shows a loss of SafO stain in the top half of the tissue that is not associated with surface erosion or fissures. Figure 4B shows tissue necrosis/matrix degradation in the radial zone, accompanied by some loss of SafO stain in the inter-territorial matrix region that is not associated with obvious signs of surface erosion. Figure 4C shows loss of SafO stain in the inter-territorial matrix, mainly confined to the bottom half of the tissue section (well within the radial zone). Complete loss of SafO stain is seen in the top region of this sample, but no major surface erosion or fissures are observed. Figure 4D shows varying staining patterns seen in the territorial matrix region and no SafO stain observed in some inter-territorial regions in the radial zone. Figure 4E shows that there are some changes starting to appear near the tidemark even when the rest of the cartilage features appear relatively normal. The number of specimens in this study that exhibited at least 1 of the above-described SafO features represented 43/100 specimens. Only HHGS scoring evaluates tidemark features, while OARSI does not apply scoring criteria for the tidemark. The current HHGS scoring system has a score of either 0 or 1 that is determined by a blood-vessel breach, but we have found other tidemark features not considered by HHGS. Figure 5A shows the formation of multiple tidemarks, and Figure 5B shows multiple tidemarks that are breached by multiple blood

21-02-2018 15:41:34

200

Acta Orthopaedica 2018; 89 (2): 197–203

Figure 3. Representative HE and SafO stained images of cartilage specimens obtained from the lateral femoral condyle. Top panel images indicate histopathological features found in “normal” cartilage, showing normal surface and cells (A), uniform safraninO staining (B) and single undulating tidemark not breached by blood vessels (C). Bottom panel images suggest typical osteoarthritic cartilage features, where surface shows fissures (D), hypercellularity, and cloning associated with fissures (D, E) and tidemark breached by blood vessels (F).

vessels. Figure 5C exhibits an unspecified tissue composition deposited near the tidemark, which stains significantly differently from normal hyaline cartilage and bone tissue. Figure 5D shows the formation of bone tissue well within the hyaline cartilage region resulting in the appearance of cartilage–bone–cartilage–bone interleaved tissue layers accompanied by multiple tidemarks. The number of specimens in this study that exhibited at least 1 of the above-described features represented 39/100 specimens.

Figure 4. Representative images of cartilage obtained from the lateral femoral condyle that indicate the unaccounted histopathological safraninO features: (A) loss of SafO stain in the top half of the tissue that is not associated with much surface erosion or fissures; (B) tissue necrosis/degradation in the radial zone, accompanied by some loss of SafO stain in the inter-territorial matrix region; (C) loss of SafO stain in the inter-territorial matrix, mainly confined to the bottom half of the tissue section; (D) varying staining patterns seen in the territorial matrix region and no SafO stain observed in some inter-territorial regions in the radial zone; (E) SafO staining loss near the tidemark even when the rest of the cartilage features appear relatively normal. The table indicates the total HHGS scores for each individual specimen, along with structure score (S), cell score (C), safraninO/fast green score (Saf), tidemark score (T). 43% of sample cohort presented with at least one feature.

11773 Mantripragada D.indd 200

21-02-2018 15:41:35

Acta Orthopaedica 2018; 89 (2): 197–203

Figure 5. Representative images of cartilage obtained from the lateral femoral condyle that indicate the unaccounted tidemark-related features as seen in the HE stained sections: (A) multiple tidemarks; (B) multiple tidemarks that are breached by multiple blood vessels; (C) unknown tissue composition deposited near the tidemark, which stains significantly differently from normal hyaline cartilage and bone tissue; (D) formation of bone tissue well within the hyaline cartilage region resulting in the appearance of cartilage–bone–cartilage–bone interleaved tissue layers accompanied by multiple tidemarks. The table indicates the total HHGS scores for each individual specimen, along with structure score (S), cell score (C), safraninO/fast green score (Saf), tidemark score (T). 39% of the sample cohort presented with at least 1 feature.

Other noteworthy observations of the histological features of osteochondral specimens from LFC include: (1) infrequent occurrence of pannus-like synovial tissue overgrowth on cartilage surface (2%); (2) relatively low incidence of hypocellularity (6%), often in the deep zone even when upper-zone chondrocytes were reasonably well preserved; and (3) frequent observation of hypercellularity in the top third of the hyaline cartilage tissue, though it was not uncommon for this to appear in the middle or deep zones as well.

Discussion Human idiopathic OA is a gradually progressing and disabling condition, with a combination of disease stages and cellular responses that are incompletely understood (Goldring and Goldring 2016). This, and the fact that mostly severe human OA cartilage specimens are readily available for research, makes understanding of the disease mechanism and prediction of the disease development a challenging task. The main finding of this study was that histological assessment of osteochondral specimens from human LFC in varus

11773 Mantripragada D.indd 201

201

knees, with relatively preserved lateral compartment JSW, revealed that 45/100 specimens were categorized adequately using 2 traditional histopathological scoring systems: 12/100 displayed low HHGS and OARSI scores and 33/100 displayed surface degradation along with other histopathological changes. However, 55/100 of the specimens exhibited histopathological features in the deep zone (e.g. loss of chondrocytes, loss of SafO stained matrix, multiple tidemarks, and SCB eruptions into the deep zone), while exhibiting good surface structure, and were not adequately assessed with the 2 current scoring systems. Inability to account for these deep-zone histopathological features along with an over-emphasized assessment of surface structure in the current scoring systems resulted in both HHGS and OARSI categorizing these specimens into early to mild OA categories, which we contend are quite possibly erroneous. We believe that the cartilage samples from this relatively large cohort of patients with idiopathic OA bring to attention histopathological features that we still need to understand, particularly if recent contentions of OA subtypes are valid (Karsdal et al. 2014, Wyatt et al. 2016). The cartilage procurement method in this study presents suitable samples exhibiting the abovementioned histopathological features, and may help to better understand early, mild, and moderate histopathological OA changes (Sritharan et al. 2017). Our findings contribute to new details and provide well-documented evidence of unreported histopathological changes observed during OA progression, especially in cartilage tissue compartments that are under low mechanical impact. A high prevalence of deep-zone and SCB alterations were seen despite the absence of substantial surface structure degradation. We are currently developing a modified HHGS system that should accommodate all of the anomalous histopathological features we observed. Careful assessment of these specimens revealed that a large subset of our cohort exhibited unaccounted for extra-cellular matrix degradation features and tidemark-related features (43% and 39%, respectively). When assessing the specimens using the traditional scoring systems, the major difference between the HHGS and OARSI system is that OARSI does not take into account changes in the deep zone independent of any observed changes in the superficial zone. Thus, tidemark alterations, death of chondrocytes in the deep zone, loss of SafO staining in the deep zone, and SCB alterations are not accounted for in OARSI if there are no associated surface structural changes. Since over half of our samples exhibited these deep-zone alterations without substantial alteration in surface structure, OARSI was unable to render a fully accurate histopathological score for these types of samples. None of the features described as abnormal features in our study have been demonstrated in normal cartilage histology (Pauli et al. 2012) and thus are definitely attributes of disease. Using the HHGS score range distribution for early, mild, moderate, and severe OA (Table) defined by Ostergaard et al. (1999), none of our samples were considered as early

21-02-2018 15:41:38

202

Acta Orthopaedica 2018; 89 (2): 197–203

Distribution of specimen cohort into early, mild, moderate, and severe OA as per the Ostergaard et al. (1999) classification of HHGS scores. OARSI score range was extrapolated from HHGS scores assuming linear correlation OA stage

HHGS Specimen score distribution

Early Mild Moderate Severe

<2 2–5 6–9 10–14

0 58% 42% 0

OARSI score

Specimen distribution

< 3.4 3.4–8.6 8.6–15.4 15.4–24

14% 78% 8% 0

OA, 58% ranked as mild, 42% ranked as moderate, and none were ranked as severe. Assuming a linear correlation between HHGS and OARSI, as determined in our study and noted by others (Ostergaard et al. 1999, Pauli et al. 2012), we calculated the OARSI score distribution range for our sample cohort and found that 14% would be considered early OA, 78% considered mild OA, 8% considered moderate OA, and none considered severe OA (Table). Many of the features we found have been mentioned in the literature to be associated with various stages and subtypes of OA progression, yet to date none of these features have been considered in the existing scoring systems. Loss of territorial matrix and inter-territorial matrix surrounding deepzone chondrocytes has been reported (Maldonado and Nam 2013). In the mechanically loaded joint space, changes have been reported to start near the superficial zone, as it is most susceptible to mechanical injury (Caramés et al. 2012). Yet, a majority of our observations in regard to territorial and interterritorial matrix loss were confined to the bottom half of the cartilage thickness. Another speculated mechanism of loss of matrix near the bottom half of the cartilage is due to changes in thickness, volume, and stiffness of SCB (Sritharan et al. 2017). We also had LFC specimens that demonstrated a near complete loss of SafO staining in the top half, but the cartilage surface remained relatively free of fissures, likely a result of the predicted lower mechanical loads in the lateral compartment of the knee (Kumar et al. 2013, Scott et al. 2013). In regard to features seen near the tidemark, it has been reported that a duplication of tidemark is associated with vascular invasion (Bullough 2004). Such a process is usually followed by an advancement of calcified cartilage into the deep zone of articular cartilage. In the late stages of OA, the penetration of vascular elements into the hyaline cartilage zone leads to bone formation around these blood vessels. Loss of proteoglycans in the deep zone has been associated with subsequent invasion into these tissue locations by blood vessels (Mapp and Walsh 2012). Chondrocytes also have a high tendency to be metabolically activated in proteoglycandepleted matrix (Suri and Walsh 2012). As a result, bloodvessel breach is more feasible, followed by bone formation long term (Maes 2017).

11773 Mantripragada D.indd 202

To date, structure-modifying treatments for OA have been disappointing (Karsdal et al. 2016). Karsdal et al. advocated for the need to segregate patients with different OA subtypes to pair them with an optimal mode of treatment to yield an effective intervention. A possible reason for poor translation to clinical practice is that many preclinical studies involve posttraumatic OA, which accounts for only 12% of symptomatic OA (Brown et al. 2006) and exhibits a different pathophysiology from idiopathic OA. If the current contention of OA subtypes is valid, then future improvements to OA histological scoring systems will need to incorporate a better balance and definitions for changes in cellular response (Poole 1997, Lotz et al. 2010), extracellular matrix (Favero et al. 2015), tidemark and SCB (Lane and Bullough 1980). Our study provides a robust and consistent model to study human idiopathic OA in the knee to contribute to the present challenges. Our study has multiple limitations, the major one being inaccessibility to normal human cartilage sample to compare our observations. The level of patello-femoral disease and integrity of the anterior and posterior cruciate ligaments were not routinely assessed intraoperatively. Future studies and analysis will need to focus on better characterizing the histopathological features by immunohistochemical staining for specific antigens like collagen type I, II, or X. A broader array of staining techniques may be considered to enhance the depth of data regarding underlying biological function at a cell and matrix level (Changoor et al. 2011). Gene expression and protein analysis can also be performed to correlate with histopathological observations and finally correlating the characteristics of the progenitors in terms of their biological potential to regenerate cartilage. Conclusions Some suggested areas of improvement for assessing OA pathological features include: (1) a better balance between the different features of the scoring systems so no one parameter (structure, cells, SafO, tidemark) is over-weighted with respect to the other parameters; (2) clear and specific instructions to distinguish the scores within each of the 4 major conventional parameters; and (3) provide a validated, online image database and simplified diagrammatic representations of these images so raters of all experience levels would achieve a better consensus opinion and decrease scoring disparities. Implementations of these approaches should potentially increase standardization among histopathological scoring systems, increase precision and rigor, and allow documentation of unaccounted for anomalies that might shed light on subgroup identification or different patterns of progression of OA.

The authors would like to acknowledge the histological processing skills of Edward Uhl in the Histochemistry Core Facility in the Biomedical Engineering Department (Lerner Research Institute, Cleveland Clinic).

21-02-2018 15:41:40

Acta Orthopaedica 2018; 89 (2): 197–203

Conception of study: GM, RJM. Study design, acquisition of data, data analyses: VPM, NP, TZ, RJM. Statistical analysis and interpretation: NO, VPM. Drafting of the article: VPM. Critical revising for important intellectual content: NP, RJM, GFM.

Acta thanks Anders Troelsen and other anonymous reviwers for help with peer review of this study.

Brown T D, Johnston R C, Saltzman C L, Marsh J L, Buckwalter J A. Posttraumatic osteoarthritis: A first estimate of incidence, prevalence, and burden of disease. J Orthop Trauma. 2006; 20 (10): 739-44. Bullough P G. The role of joint architecture in the etiology of arthritis. Osteoarthritis Cartilage 2004; 12 (Suppl.): 2-9. Caramés B, Taniguchi N, Seino D, Blanco F J, D’Lima D, Lotz M. Mechanical injury suppresses autophagy regulators and pharmacologic activation of autophagy results in chondroprotection. Arthritis Rheum 2012; 64 (4): 1182-92. Changoor A, Tran-Khanh N, Méthot S, Garon M, Hurtig M B, Shive M S, Buschmann M D. A polarized light microscopy method for accurate and reliable grading of collagen organization in cartilage repair. Osteoarthritis Cartilage 2011; 19 (1): 126-35. Favero M, Ramonda R, Goldring M B, Goldring S R, Punzi L. Early knee osteoarthritis: Figure 1. RMD Open [Internet] 2015; 1 (Suppl 1): e000062. Goldring S R, Goldring M B. Changes in the osteochondral unit during osteoarthritis: Structure, function and cartilage-bone crosstalk. Nat Rev Rheumatol 2016; 12 (11): 632-44. Greene M A, Loeser R F. Aging-related inflammation in osteoarthritis. Osteoarthritis Cartilage 2015; 23 (11): 1966-71. Karsdal M A, Christiansen C, Ladel C, Henriksen K, Kraus V B, Bay-Jensen A C. Osteoarthritis: A case for personalized health care? Osteoarthritis Cartilage 2014; 22 (1): 7-16. Karsdal M A, Michaelis M, Ladel C, Siebuhr A S, Bihlet A R, Andersen J R, Guehring H, Christiansen C, Bay-Jensen A C, Kraus V B. Disease-modifying treatments for osteoarthritis (DMOADs) of the knee and hip: Lessons learned from failures and opportunities for the future. Osteoarthritis Cartilage 2016; 24 (12): 2013-21. Kumar D, Manal K T, Rudolph K S. Knee joint loading during gait in healthy controls and individuals with knee osteoarthritis. Osteoarthritis Cartilage 2013; 21 (2): 298-305. Lane L B, Bullough P G. Age-related changes in the thickness of the calcified zone and the number of tidemarks in adult human articular cartilage. J Bone Joint Surg Br 1980; 62 (3): 372-5. Li G, Yin J, Gao J, Cheng TS, Pavlos NJ, Zhang C, Zheng MH. Subchondral bone in osteoarthritis: insight into risk factors and microstructural changes. Arthritis Res Ther 2013; 15(6): 223. Lotz M K, Otsuki S, Grogan S P, Sah R, Terkeltaub R, D’Lima D. Cartilage cell clusters. Arthritis Rheumatol 2010; 62 (8): 2206-18.

11773 Mantripragada D.indd 203

203

Maes C. Signaling pathways effecting crosstalk between cartilage and adjacent tissues. Seminars in cell and developmental biology: The biology and pathology of cartilage. Semin Cell Dev Biol 2017; 62: 16-33. Maldonado M, Nam J. The role of changes in extracellular matrix of cartilage in the presence of inflammation on the pathology of osteoarthritis. Biomed Res Int 2013; 2013: 284873. Malfait A M. Osteoarthritis year in review 2015: Biology. Osteoarthritis Cartilage 2016; 24 (1): 21-6. Mankin H J, Dorfman H, Lippiello L, Zarins A. Biochemical and metabolic abnormalities in articular cartilage from osteo-arthritic human hips. J Bone Joint Surg Am 1971; 53 (3): 523-7. Mapp P I, Walsh D A. Mechanisms and targets of angiogenesis and nerve growth in osteoarthritis. Nat Rev Rheumatol 2012; 8 (7): 390-8. Orlowsky E W, Kraus V B. The role of innate immunity in osteoarthritis: When our first line of defense goes on the offensive. J Rheumatol 2015; 42 (3): 363-71. Ostergaard K, Andersen C B, Petersen J, Bendtzen K, Salter D M. Validity of histopathological grading of articular cartilage from osteoarthritic knee joints. Ann Rheum Dis 1999; 58 (4): 208-13. Outerbridge R. The etiology of chondromalacia patellae. J Bone Joint Surg Br 1961; 43-B: 752-7. Pauli C, Whiteside R, Heras F L, Nesic D, Koziol J, Grogan S P, Matyas J, Pritzker K P H, D’Lima D D, Lotz M K. Comparison of cartilage histopathology assessment systems on human knee joints at all stages of osteoarthritis development. Osteoarthritis Cartilage 2012; 20 (6): 476-85. Poole C A. Articular cartilage chondrons: Form, function and failure. J Anat 1997; 191 ( Pt 1): 1-13. Pritzker K P H, Gay S, Jimenez S A, Ostergaard K, Pelletier J P, Revell K, Salter D, van den Berg W B. Osteoarthritis cartilage histopathology: Grading and staging. Osteoarthritis Cartilage 2006; 14 (1): 13-29. Rutgers M, van Pelt M J P, Dhert W J A, Creemers L B, Saris D B F. Evaluation of histological scoring systems for tissue-engineered, repaired and osteoarthritic cartilage. Osteoarthritis Cartilage 2010; 18 (1): 12-23. Scott J E. Cetylpyridinium chloride as a fixative for glycosaminoglycans in histologic sections. Arch Dermatol 1989; 125 (7): 1002. Scott C E H, Nutton R W, Biant L C. Lateral compartment osteoarthritis of the knee: Biomechanics and surgical management of end-stage disease. Bone Joint J 2013; 95 B (4): 436-44. Sritharan P, Lin Y C, Richardson S E, Crossley K M, Birmingham T B, Pandy M G. Musculoskeletal loading in the symptomatic and asymptomatic knees of middle-aged osteoarthritis patients. J Orthop Res 2017; 35 (2): 321-30. Suri S, Walsh D A. Osteochondral alterations in osteoarthritis. Bone 2012; 51 (2): 204-11. Varady N H, Grodzinsky A J. Osteoarthritis year in review 2015: Mechanics. Osteoarthritis Cartilage 2016; 24 (1): 27-35. Wyatt L A, Moreton B J, Mapp P I, Wilson D, Hill R, Ferguson E, Scammell B E, Walsh D A. Histopathological subgroups in knee osteoarthritis. Osteoarthritis Cartilage 2016; 25 (1): 14-22.

21-02-2018 15:41:40

204

Acta Orthopaedica 2018; 89 (2): 204–210

Graft fixation influences revision risk after ACL reconstruction with hamstring tendon autografts A study of 38,666 patients from the Scandinavian knee ligament registries 2004–2011 Andreas PERSSON 1,2, Tone GIFSTAD 3,4, Martin LIND 5 , Lars ENGEBRETSEN 6,7,8, Knut FJELDSGAARD 1, Jon Olav DROGSET 3,4, Magnus FORSSBLAD 9, Birgitte ESPEHAUG 10, Asle B KJELLSEN 1, and Jonas M FEVANG 1

1 Department of Orthopaedic Surgery, Haukeland University Hospital, Bergen, Norway; 2 University of Bergen, Faculty of Medicine and Dentistry, Department of Clinical Medicine, Bergen, Norway; 3 Department of Orthopaedic Surgery, Trondheim University Hospital, Trondheim, Norway; 4 Norwegian University of Science and Technology, Trondheim, Norway; 5 Department of Orthopaedics, Aarhus University Hospital, Aarhus, Denmark; 6 Department of Orthopaedic Surgery, Oslo University Hospital, Oslo, Norway; 7 Faculty of Medicine, University of Oslo, Oslo, Norway; 8 Oslo Sports Trauma Center, Norwegian School of Sport Sciences, Oslo, Norway; 9 Stockholm Sports Trauma Research Center, Karolinska Institutet, Stockholm, Sweden; 10 Centre for Evidence-Based Practice, Bergen University College, Bergen, Norway Correspondence: Andreas.persson@helse-bergen.no Submitted 2017-04-06. Accepted 2017-10-23.

Background and purpose — A large number of fixation methods of hamstring tendon autograft (HT) are available for anterior cruciate ligament reconstruction (ACLR). Some studies report an association between fixation method and the risk of revision ACLR. We compared the risk of revision of various femoral and tibial fixation methods used for HT in Scandinavia 2004–2011. Materials and methods — A register-based study of 38,666 patients undergoing primary ACLRs with HT, with 1,042 revision ACLRs. The overall median follow-up time was 2.8 (0–8) years. Fixation devices used in a small number of patients were grouped according to design and the point of fixation. Results — The most common fixation methods were Endobutton (36%) and Rigidfix (31%) in the femur; and interference screw (48%) and Intrafix (34%) in the tibia. In a multivariable Cox regression model, the transfemoral fixations Rigidfix and Transfix had a lower risk of revision (HR 0.7 [95% CI 0.6–0.8] and 0.7 [CI 0.6–0.9] respectively) compared with Endobutton. In the tibia the retro interference screw had a higher risk of revision (HR 1.9 [CI 1.3–2.9]) compared with an interference screw. Interpretation — The choice of graft fixation influences the risk of revision after primary ACLR with hamstring tendon autograft. ■

The most commonly used grafts in Scandinavia for anterior cruciate ligament reconstruction (ACLR) are hamstring tendon autografts (HT) or patellar tendon autografts (Granan et al. 2009). There are multiple devices available on the

market for fixation of the graft. Most devices have been evaluated mechanically tested with acceptable results (Ahmad et al. 2004, Milano et al. 2006, Aga et al. 2013). Numerous clinical studies have found similar objective or subjective outcomes comparing different fixation techniques (Laxdal et al. 2006, Rose et al. 2006, Moisala et al. 2008, Myers et al. 2008, Harilainen and Sandelin 2009, Drogset et al. 2011, Frosch et al. 2012, Gifstad et al. 2014). Hence, there is no definite recommendation for the best fixation technique and the surgeon’s choice of fixation is likely to be influenced by personal experience, local traditions, and possibly marketing from the industry. A recent study (Persson et al. 2015) from the Norwegian Knee Ligament Registry (NKLR) identified combinations of fixations for HT with increased risk of revision at 2 years. In addition, a higher risk of revision when using cortical buttons compared with transfemoral or intratunnel fixations in the femur was observed. These findings call into question the increasing use of cortical buttons for HT fixation (Ahlden et al. 2012). In addition, Andernord et al. (2014) found a reduced risk of early revision when a metal interference screw was used to fixate semitendinosus grafts in the tibia. This study further investigates the risk of revision for the most common fixation techniques and devices in HT reconstructions during the period 2004–2011, using a combined dataset from all 3 Scandinavian ACL registries (the NKLR, the Swedish National Anterior Cruciate Ligament Registry, and the Danish Knee Ligament Reconstruction Register).

11411 Persson D.indd 204

21-02-2018 15:44:44

Acta Orthopaedica 2018; 89 (2): 204–210

Materials and methods Data sources The Scandinavian knee ligament registries were established in 2004–2005 and are similar in design (Granan et al. 2008, Ahlden et al. 2012, Rahr-Wagner and Lind 2016). Patient-specific data (sex, age, previous surgery/injuries to index or contralateral knee), surgical details (graft choice, fixation choice, potential treatment of other ligament injuries or meniscal/cartilage injuries) and intraoperative findings (meniscal and cartilage injuries and signs of arthrosis) are reported at the time of surgery. Patients are followed prospectively with revision ACLR, subsequent surgery to the index knee, and patient-reported outcome (Knee Injury and Osteoarthritis Outcome Score at 1, 2, 5, and 10 years follow-up) as endpoints. The report rates to the registries are similar, from 86% to ≥ 90% (Ytterstad et al. 2012, Rahr-Wagner et al. 2013a, www.aclregister.nu 2014). This study includes all 38,666 patients registered from the start of the Scandinavian registries up to December 31, 2011, with a primary ACLR with an HT. The following data were considered in the study: date of primary and potential revision reconstruction, patient age and sex, fixation of the graft in femur and tibia, activity at primary injury, location (right/left knee), meniscal injury or treatment (yes/no), cartilage injury (yes/no), medial collateral injury (yes/no), and other concomitant injuries (fractures, nerve injuries, and vascular injuries). Patients with concomitant ligament injuries treated surgically were not included. Exposure We analyzed the revision rate and risk dependent on what tibial and femoral fixation device was used in the primary ALCR. The fixation device in the femur and tibia was either registered by the catalogue number of each device by using the unique bar-code sticker delivered by the manufacturer, or reported manually by the surgeon with either the registered trademark name of the device or a description of the fixation design, such as interference screw. Devices used in fewer than 500 patients were grouped according to their design and point of graft fixation. The femoral devices in the dataset were grouped as: cortical fixation (Endobutton [Smith & Nephew] or other), transfemoral fixation (Rigidfix [DePuy Mitek], Transfix [Arthrex] or other), interference screw, or other/unknown. The tibial devices in the dataset were grouped as: cortical fixation, interference screw, Intrafix (DePuy Mitek), retro interference screw, Rigidfix (DePuy Mitek), or other/unknown. Statistics Statistical analyses were performed using SPSS Statistics software version 22 (SPSS Inc, IBM Corp, Armonk, NY, USA). All tests were 2-sided with a 0.05 significance level. Unadjusted cumulative implant revision curves were established using Kaplan–Meier estimates and crude 2- and 5-year revision percentages are reported. Unadjusted and adjusted hazard ratios (HR) with 95% confidence intervals (CI) were

11411 Persson D.indd 205

205

estimated in Cox regression analyses. The multivariable analyses were stratified for country. The assumption of proportional hazards of the Cox regression model was evaluated with log–log plot and was found suitable. All survival analyses were performed with revision as the endpoint. No data were received on death or emigration. Patients were at risk and followed until revision or end of study. Confounding factors Patient age (5-year categories) at the time of the primary reconstruction, sex, meniscal injury to 1 or both menisci (yes/ no), cartilage injury (yes/no), and activity at primary injury (pivoting activity [soccer, team handball, alpine activities]/ other activities) were considered as possible confounding factors as these are potential risk factors for revision and may also influence the choice of fixation method. Further, none of the factors were considered as possible mediating variables. Additional analyses showed that meniscal injury and cartilage injury was not associated with, and thus did not inform, the fixation method. They were therefore not entered into the multivariable Cox regression analysis. Additional adjustment was made for corresponding fixation in the tibia when analyzing femoral fixations and for corresponding fixation in the femur when analyzing tibial fixations. Ethics, funding, and potential conflicts of interest Informed consent has been signed by all the participants in the NKLR, and the NKLR is approved by the Norwegian Data Inspectorate. No written consent is necessary in Denmark and Sweden for national healthcare registries. The study was funded by a grant from the Norwegian Orthopedic association. LE has received course honoraria from Smith & Nephew, a fellowship grant from Arthrex to his institution, royalties for making of tools from Arthrex, and travel/accommodation expenses covered or reimbursed by Smith & Nephew for Multiligament course in Vail.

Results The mean age at surgery was 28 years, and 57% were men. The median time from initial injury to the time of primary ACLR was 8 months (range 0–45 years). The most commonly used fixations in the femur were the Endobutton and Rigidfix, used in 14,106 and 12,041 patients respectively. The most commonly used tibial fixations were interference screw and Intrafix, used in 18,640 and 13,014 patients respectively. The median overall follow-up time was 2.8 (1.8–4.3) years (Table 1). The most commonly used combinations of fixations (femoral x tibial) were Rigidfix x Intrafix and Endobutton x Interference screw, used in 8,023 and 8,006 patients respectively (Table 2). The use of femoral fixation with Endobutton increased during the entire study period while the usage of Rigidfix decreased after a peak in 2007 (Figure 1). The use of tibial

21-02-2018 15:44:44

206

Acta Orthopaedica 2018; 89 (2): 204–210

Table 1. Patients’ characteristics and baseline epidemiology. Values are percentages unless otherwise specified Femoral fixation

Cortical fixation Endobutton Other

Transfemoral fixation Rigidfix Transfix Other

n Age, mean (SD) a,b Pivoting activity c Male MCL injury Menisc injury Cartilage injury Other injury Follow-up, mean (SD) b

14,106 27 (10) 66 56 2.5 41 21 0.4 2.2 (1.8)

4,352 28 (11) 66 58 2.1 44 21 0.5 2.5 (1.7)

Tibial fixation

Cortical Interference fixation screw

n Age, mean (SD) a,b Pivoting activity c Male MCL injury Meniscal injury Cartilage injury Other injury Follow-up, mean (SD) b

4,814 27 (11) 65 55 3.4 43 19 0.5 3.2 (2.0)

a At time of surgery. b Years. c At primary injury (soccer,

team handball, alpine activities).

18,640 28 (10) 66 58 2.3 43 24 0.4 2.7 (1.9)

12,041 29 (10) 66 57 1.9 38 20 0.7 3.7 (1.8)

520 28 (10) 72 57 3.7 42 28 1.3 5.4 (2.0)

3,453 28 (10) 67 59 1.2 43 20 0.3 3.0 (1.8)

Retro interIntrafix ference screw Rigidfix

Other/ unknown

13,014 29 (11) 67 58 1.6 37 18 0.8 3.3 (1.9)

3,652 28 (10) 66 59 3.2 42 29 0.5 3.9 (1.8)

Interference Other/ screw unknown

508 27 (10) 63 58 0.8 46 30 1.0 3.3 (1.8)

867 27 (10) 59 54 3.6 37 25 0.2 4.1 (1.7)

542 28 (11) 64 54 3.3 40 23 0.4 2.9 (2.1)

823 27 (11) 66 57 5.1 45 29 0.9 2.8 (2.3)

Table 2. Combinations of fixations used in more than 500 patients Fixations ( femoral x tibial) Endobutton x interference screw Endobutton x intrafix Endobutton x cortical fixation Other cortical x interference screw Other cortical x cortical fixation Other cortical x Intrafix Rigidfix x Intrafix Rigidfix x interference screw Rigidfix x Rigidfix Transfix x interference screw Interference screw x interference screw Other combinations (used in less than 500 patients) Total

n 8,006 3,154 2,541 1,856 1,483 948 8,023 2,661 825 3,123 2,859 3,187 38,666

fixation with interference screw increased after 2006 while the use of Intrafix decreased after a peak in 2006 (Figure 2). Revision rate during the first postoperative year was low (Figures 3 and 4). The 5-year revision rate according to femoral fixation was 5.0% (CI 4.4–5.7) for Endobutton, 3.4% (CI 3.0–3.8) for Rigidfix, and 3.5% (CI 2.8–4.1) for Transfix. For tibial fixation the 5-year revision rate was 4.2% (CI 3.7–4.6) for interference screw, 4.0% (CI 3.0–3.8) for Intrafix, and 2.5% (CI 1.4–3.7) for Rigidfix (Figures 3, 4 and Table 3). In the multivariable analysis, the HR for revision was 0.7

11411 Persson D.indd 206

Figure 1. Femoral fixation in Scandinavia 20015–2011.

Figure 2. Tibial fixation in Scandinavia 20015–2011.

21-02-2018 15:44:44

Acta Orthopaedica 2018; 89 (2): 204–210

207

Figure 3. Cumulative revision curve for femoral fixations.

Figure 4. Cumulative revision curve for tibial fixations.

Table 3. Crude revision rates for patients within the examined groups of fixations at 2 and 5 years

Fixation point and group Femoral fixation a Cortical fixation Endobutton Other Transfemoral fixation Rigidfix Transfix Other Interference screw Other/unknown Tibial fixation b Cortical fixation Interference screw Intrafix Retro interference screw Rigidfix Other/unknown

n (revisions)

Revision rate (CI) % 2 years 5 years

Table 4. Results (hazard ratios – HR) from the Cox regression models with revision as endpoint Fixation point and group

14,106 (342) 4,352 (115)

2.7 (2.4–3.1) 2.2 (1.7–2.7)

5.0 (4.4–5.7) 4.5 (3.6–5.4)

12,041 (316) 3,652 (100) 520 (32) 3,453 (119) 542 (18)

1.7 (1.4–1.9) 1.5 (1.1–1.9) 4.2 (2.5–6.0) 2.7 (2.1–3.3) 2.7 (1.1–4.2)

3.4 (3.0–3.8) 3.5 (2.8–4.1) 6.1 (4.0–8.3) 5.2 (4.2–6.2) 5.4 (2.7–8.0)

4,814 (159) 18,640 (462) 13,014 (355) 508 (27) 867 (18) 823 (21)

2.8 (2.3–3.3) 2.2 (2.0–2.5) 1.9 (1.6–2.1) 3.4 (1.7–5.1) 1.3 (0.4–2.0) 1.8 (0.6–2.9)

4.6 (3.8–5.3) 4.2 (3.7–4.6) 4.0 (3.6–4.5) 6.7 (4.1–9.3) 2.5 (1.4–3.7) 4.7 (2.7–6.8)

Femoral fixation Cortical fixation Endobutton Other Transfemoral fixation Rigidfix Transfix Other Interference screw Other/unknown Tibial fixation Cortical fixation Interference screw Intrafix Retro interference screw Rigidfix Other/ unknown

HR (CI)

Adjusted HR (CI) a

Ref. 0.9 (0.8–1.2)

Ref. 0.8 (0.7–1.1)

0.7 (0.6–0.8) 0.7 (0.5–0.8) 1.2 (0.9–1.8) 1.1 (0.9–1.3) 1.1 (0.7–1.7)

0.7 (0.6–0.8) 0.7 (0.6–0.9) 1.1 (0.7–1.6) 1.1 (0.9–1.4) 1.1 (0.7–1.9)

1.1 (1.0–1.4) Ref. 0.9 (0.8–1.1) 1.8 (1.2–2.6) 0.6 (0.3–0.9) 1.1 (0.7–1.6)

1.1 (0.9–1.4) Ref. 1.0 (0.9–1.2) 1.9 (1.3–2.9) 0.9 (0.5–1.4) 1.0 (0.6–1.5)

a Adjusted

Log-rank test for difference in overall revision between groups: a p-value < 0.001 b p-value = 0.001

for the Rigidfix (CI 0.6–0.8) and Transfix (CI 0.6–0.9) groups compared with the Endobutton group and 1.9 (CI 1.3–2.9) for the group with the tibial fixation retro interference screw compared with the interference screw group (Table 4).

Discussion In this large multiregistry-based study from the Scandinavian countries, the main finding was that the HR for revision was reduced by 30% when using transfemoral fixation with Rigidfix or Transfix compared with cortical fixation with Endobutton, independent of the tibial fixation used. The hamstring tendon autograft was fixed with the cortical fixation Endobut-

11411 Persson D.indd 207

analysis model stratified for country (Sweden, Denmark, Norway) and adjusted for gender, age at surgery (5-year categories), activity at primary injury, and corresponding fixation in tibia or femur.

ton in one-third of the patients, with increasing use during the last years of the study period. These results are in line with the recent findings of increased risk of revision within 2 years for cortical fixation compared with transfemoral fixation from the NKLR (Persson et al 2015). One can question the clinical relevance of a minor difference in revision risk. However, when clinical outcome after revision ACLR may be worse than after primary ACLR (Battaglia et al. 2007, Grassi et al. 2016), we believe the differences are of interest. Previously, a variety of outcomes have been studied in clinical studies comparing different fixation devices and techniques (Drogset et al. 2005, Rose et al. 2006, Capuano et al. 2008, Moisala et al. 2008) with similar outcomes in the examined

21-02-2018 15:44:46

208

groups. However, there are a few clinical, biomechanical, and anatomical studies that have reported differences between different graft fixations in the femur. A recent meta-analysis by Browning et al. (2017) included 41 clinical level 1–4 studies comparing clinical outcome for patients treated with an ACLR with 4-strand hamstring autograft using either suspensory or aperture fixation. They found better arthrometric stability and fewer graft ruptures using suspensory compared with aperture fixation at a minimum of 2-year follow-up; however, they included graft fixation in the femur with cross-pins in the suspensory group. In a clinical trial of double-bundle ACLR, Ibrahim et al. (2015) found that 4 out of 32 patients with ACL grafts that were fixed in the femur with cortical fixation had > 5 mm of postoperatively instrumented knee laxity compared with 0 out of 34 patients with transfemoral fixation at a mean follow-up of 2.5 years. They found no difference between the 2 groups in the Lachman and pivot-shift test. Frosch et al. (2012) compared, in a prospective non-randomized study, femoral fixation with bioabsorbable interference screws in 31 cases and bioabsorbable Rigidfix in 28 cases. They found similar subjective results but less side-to-side anterior translation as measured with a KT-1000 arthrometer in the cases with femoral fixation using Rigidfix. Biomechanical studies most frequently investigate graftfixation complex stiffness, pull-out strength, or graft–fixation complex lengthening after cyclic loading. Laxity of the graft– fixation complex and graft–tunnel motion might disturb the biologic incorporation of the graft in the bone tunnel (Hoher et al. 1998), leading to a weaker reconstruction. In a cadaver model measuring graft–fixation complex stiffness in double-looped semitendinosus grafts, To et al. (1999) found the stiffness of the graft and fixation complex to be dependent on the fixation method rather than the graft, with decreased stiffness when using a suture loop and a cortical button. Höher et al. (1998) found up to 3 mm of graft-tunnel motion when using a titanium button and polyester tape to fix quadruple hamstring grafts within the femoral bone tunnel. To further investigate the histological insertion point or the graft itself there is a need for more studies where samples are collected from revision ACLRs. There has been a debate as to whether the surgical technique for femoral tunnel drilling affects the clinical outcome. Both Rigidfix and Transfix are likely to mainly have been fixed through a transtibial technique (TT) for drilling the femoral tunnel. TT has been shown to have a lower risk of revision compared with the anteromedial (AM) technique in a previous register study (Rahr-Wagner et al. 2013b). The authors argued that it could be due to the increased load on an anatomic reconstructed graft, due to potential problems with a shorter femoral tunnel or as a result of the surgeon’s learning curve when the new AM technique was introduced. However, they did not adjust for graft fixation in their analysis. Liu et al. (2015) found, in a systematic review, superior results for the AM technique based on physical examination, and it is possible that the

11411 Persson D.indd 208

Acta Orthopaedica 2018; 89 (2): 204–210

mentioned difference in revision risk could be due to an unknown confounder, such as the graft fixation. A change from transfemoral devices to cortical fixation has previously been reported from the Swedish ACL registry, probably as a result of the focus on anatomic ACL reconstruction using the AM technique (Ahlden et al. 2012). This tendency is also clear in our study. Among the investigated tibial fixation devices the retro interference screw was the only device with a statistically significantly higher risk of revision compared with the interference screw. The retro interference screw (available in titanium or degradable poly-L-lactic acid [PLLA]) is placed retrogradely into the tibial bone tunnel from inside the joint. Although poor results have been reported in a previous biomechanical study (Scannell et al. 2015), and the possible risk of failure when using PLLA screws (Drogset et al. 2005, Persson et al. 2015) could explain the increased revision risk for the retro interference screw found in this study, we interpret the results with caution due to the small sample size. Further, we did not have data defining the material of the included retro interference screws and thus may not know whether this could have contributed to the inferior results. A limited number of register studies have been conducted on the current topic. Andernord et al. (2014) found a statistically significant lower incidence of revision surgery when a metal interference screw was used in semitendinosus tendon autograft reconstructions compared with a bioabsorbable interference screw, AO screw, metal interference screw + staple, or Intrafix registered in the Swedish National Anterior Cruciate Ligament Registry 2005–2011. This was, however, not found in the group with a combined semitendinosus and gracilis graft, which was used in four-fifths of the patients, in line with our results. Strengths and weaknesses The most important strength in this study is the large sample size of the groups investigated. A randomized controlled trial is difficult to conduct with enough statistical power to investigate a rare endpoint such as revision ACLR (Naylor and Guyatt 1996). A sample size calculation shows that 1,000 patients are needed in each group to detect a statistically significant difference in 2-year revision rates of 2.4% and 4.7%, equivalent to a hazard rate ratio of 2 (with a 2-sided 0.05 level and power of 80%). In general, prospective registry-based cohort studies are considered to be hypothesis-generating and not proving causality. However, in modern observational studies where potential biases are considered, estimates of treatment effects may be similar to those found in randomized controlled trials (Benson and Hartz 2000). Therefore, we believe our study to have a good methodology to investigate the risk of failure for different surgical techniques, such as choice of fixation method for the graft. The baseline data of the Norwegian registry have been shown to be congruent with other registries (Maletis et al.

21-02-2018 15:44:47

Acta Orthopaedica 2018; 89 (2): 204–210

2011, Granan et al. 2012). Accordingly, we believe the results to be applicable not only to the countries where the study was conducted, but to a general orthopedic community. We acknowledge the existing weaknesses of this study. For the smallest patient groups our results might be influenced by hospital-dependent revision rates. Experienced surgeons at large-volume clinics might be more prone to revise patients, and could have a different fixation choice for the primary ALCR than surgeons in low-volume clinics. These surgeons could also attract more high-level athletes with a higher risk of re-injury. We have no complete data on the surgeons’ experience, the postoperative rehabilitation protocol, graft size, activity level of the patient, if TT or AM technique was used for femoral drilling, or if the hamstring tendons are semitendinosus grafts or a combination of semitendinosus and gracilis, which are factors that potentially could influence the risk of revision. The use of revision surgery as the endpoint is a robust outcome measure, but it does not include patients with subjective or objective graft failures that have not undergone revision surgery. Although the number of graft failures is probably greater than the number of patients reaching our endpoint, we believe the observed differences are valid. In addition, we have no reason to believe that patients in certain fixation groups would be more prone to seek clinical attention and be considered for revision surgery. We do not have the data on why the patients were revised, which could potentially differ between fixation groups. We have no data on death or emigration, which potentially could bias our results as a competing risk to revision. With a mean age of 28 years in the population, occurrence of death in the follow-up is likely to be low. We do not believe that occurrence of emigration would differ between the groups. Further, we do not have data on possible bilateral observations included. Even though the occurrence is probably not different amongst the groups investigated, this might have biased our results. Summary Although that the cause of revision ACLR is often multifactorial, the results from this study suggest that there could be substantial differences in revision risk dependent on what fixation method is used in hamstring autograft ACL reconstructions.The results illustrate the need for continuous multiregister cooperation with fixation devices registered by catalogue number to allow for early detection of possible implant failures. All authors contributed to the planning of the project, interpreting results, draft revision, and approval of the manuscript. AP, TG, and BE did the statistical analysis. The authors would like to thank all colleagues for reporting primary ACLRs and revisions to the registries. Special thanks are extended to the staff of the registries for data processing and quality assurance.

11411 Persson D.indd 209

209

Acta thanks Brian Devitt and Jos van Raaij for help with peer review of this study. Aga C, Rasmussen M T, Smith S D, Jansson K S, LaPrade R F, Engebretsen L, Wijdicks C A. Biomechanical comparison of interference screws and combination screw and sheath devices for soft tissue anterior cruciate ligament reconstruction on the tibial side. Am J Sports Med 2013; 41 (4): 841-48. Ahlden M, Samuelsson K, Sernert N, Forssblad M, Karlsson J, Kartus J. The Swedish National Anterior Cruciate Ligament Register: A report on baseline variables and outcomes of surgery for almost 18,000 patients. Am J Sports Med 2012; 40 (10): 2230-35. Ahmad C S, Gardner T R, Groh M, Arnouk J, Levine W N. Mechanical properties of soft tissue femoral fixation devices for anterior cruciate ligament reconstruction. Am J Sports Med 2004; 32 (3): 635-40. Andernord D, Bjornsson H, Petzold M, Eriksson B I, Forssblad M, Karlsson J, Samuelsson K. Surgical predictors of early revision surgery after anterior cruciate ligament reconstruction: Results from the Swedish National Knee Ligament Register on 13,102 patients. Am J Sports Med 2014; 42 (7): 1574-82. Battaglia M J 2nd, Cordasco F A, Hannafin J A, Rodeo S A, O’Brien S J, Altchek D W, Cavanaugh J, Wickiewicz T L, Warren R F. Results of revision anterior cruciate ligament surgery. Am J Sports Med 2007; 35 (12): 2057-66. Benson K, Hartz A J. A comparison of observational studies and randomized, controlled trials. N Engl J Med 2000; 342 (25): 1878-86. Browning W M 3rd, Kluczynski M A, Curatolo C, Marzo J M. Suspensory versus aperture fixation of a quadrupled hamstring tendon autograft in anterior cruciate ligament reconstruction: A meta-analysis. Am J Sports Med 2017; 45 (10): 2418-27. Capuano L, Hardy P, Longo U G, Denaro V, Maffulli N. No difference in clinical results between femoral transfixation and bio-interference screw fixation in hamstring tendon ACL reconstruction: A preliminary study. Knee 2008; 15 (3): 174-79. Drogset J O, Grontvedt T, Tegnander A. Endoscopic reconstruction of the anterior cruciate ligament using bone-patellar tendon-bone grafts fixed with bioabsorbable or metal interference screws: A prospective randomized study of the clinical outcome. Am J Sports Med 2005; 33 (8): 1160-65. Drogset J O, Straume L G, Bjorkmo I, Myhr G. A prospective randomized study of ACL-reconstructions using bone-patellar tendon-bone grafts fixed with bioabsorbable or metal interference screws. Knee Surg Sports Traumatol Arthrosc 2011; 19 (5): 753-59. Frosch S, Rittstieg A, Balcarek P, Walde T A, Schuttrumpf J P, Wachowski M M, Sturmer K M, Frosch K H. Bioabsorbable interference screw versus bioabsorbable cross pins: Influence of femoral graft fixation on the clinical outcome after ACL reconstruction. Knee Surg Sports Traumatol Arthrosc 2012; 20 (11): 2251-56. Gifstad T, Drogset J O, Grontvedt T, Hortemo G S. Femoral fixation of hamstring tendon grafts in ACL reconstructions: The 2-year follow-up results of a prospective randomized controlled study. Knee Surg Sports Traumatol Arthrosc 2014; 22 (9): 2153-62. Granan L P, Bahr R, Steindal K, Furnes O, Engebretsen L. Development of a national cruciate ligament surgery registry: The Norwegian National Knee Ligament Registry. Am J Sports Med 2008; 36 (2): 308-15. Granan L P, Forssblad M, Lind M, Engebretsen L. The Scandinavian ACL registries 2004–2007: Baseline epidemiology. Acta Orthop 2009; 80 (5): 563-7. Granan L P, Inacio M C, Maletis G B, Funahashi T T, Engebretsen L. Intraoperative findings and procedures in culturally and geographically different patient and surgeon populations: An anterior cruciate ligament reconstruction registry comparison between Norway and the USA. Acta Orthop 2012; 83 (6): 577-82. Grassi A, Ardern C L, Marcheggiani Muccioli G M, Neri M P, Marcacci M, Zaffagnini S. Does revision ACL reconstruction measure up to primary surgery? A meta-analysis comparing patient-reported and clinician-reported outcomes, and radiographic results. Br J Sports Med 2016; 50 (12): 716-24.

21-02-2018 15:44:47

210

Harilainen A, Sandelin J. A prospective comparison of 3 hamstring ACL fixation devices—Rigidfix, BioScrew, and Intrafix—randomized into 4 groups with 2 years of follow-up. Am J Sports Med 2009; 37 (4): 699-706. Hoher J, Moller H D, Fu F H. Bone tunnel enlargement after anterior cruciate ligament reconstruction: Fact or fiction? Knee Surg Sports Traumatol Arthrosc 1998; 6 (4): 231-40. Ibrahim S A, Abdul Ghafar S, Marwan Y, Mahgoub A M, Al Misfer A, Farouk H, Wagdy M, Alherran H, Khirait S. Intratunnel versus extratunnel autologous hamstring double-bundle graft for anterior cruciate ligament reconstruction: A comparison of 2 femoral fixation procedures. Am J Sports Med 2015; 43 (1): 161-8. Laxdal G, Kartus J, Eriksson B I, Faxen E, Sernert N, Karlsson J. Biodegradable and metallic interference screws in anterior cruciate ligament reconstruction surgery using hamstring tendon grafts: Prospective randomized study of radiographic results and clinical outcome. Am J Sports Med 2006; 34 (10): 1574-80. Liu A, Sun M, Ma C, Chen Y, Xue X, Guo P, Shi Z, Yan S. Clinical outcomes of transtibial versus anteromedial drilling techniques to prepare the femoral tunnel during anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc 2015; [Epub ahead of print] Maletis G B, Granan L P, Inacio M C, Funahashi T T, Engebretsen L. Comparison of community-based ACL reconstruction registries in the US and Norway. J Bone Joint Surg Am 2011; 93 (Suppl 3): 31-6. Milano G, Mulas P D, Ziranu F, Piras S, Manunta A, Fabbriciani C. Comparison between different femoral fixation devices for ACL reconstruction with doubled hamstring tendon graft: A biomechanical analysis. Arthroscopy 2006; 22 (6): 660-8. Moisala A S, Jarvela T, Paakkala A, Paakkala T, Kannus P, Jarvinen M. Comparison of the bioabsorbable and metal screw fixation after ACL reconstruction with a hamstring autograft in MRI and clinical outcome: A prospective randomized study. Knee Surg Sports Traumatol Arthrosc 2008; 16 (12): 1080-6. Myers P, Logan M, Stokes A, Boyd K, Watts M. Bioabsorbable versus titanium interference screws with hamstring autograft in anterior cruciate ligament reconstruction: A prospective randomized trial with 2-year follow-up. Arthroscopy 2008; 24 (7): 817-23.

11411 Persson D.indd 210

Acta Orthopaedica 2018; 89 (2): 204–210

Naylor C D, Guyatt G H. Users’ guides to the medical literature, X: How to use an article reporting variations in the outcomes of health services. The Evidence-Based Medicine Working Group. JAMA 1996; 275 (7): 554-8. Persson A, Kjellsen A B, Fjeldsgaard K, Engebretsen L, Espehaug B, Fevang J M. Registry data highlight increased revision rates for endobutton/biosure HA in ACL reconstruction with hamstring tendon autograft: A nationwide cohort study from the Norwegian Knee Ligament Registry, 2004–2013. Am J Sports Med 2015; 43 (9): 2182-8. Rahr-Wagner L, Lind M. The Danish Knee Ligament Reconstruction Registry. Clin Epidemiol 2016; 8: 531-5. Rahr-Wagner L, Thillemann T M, Lind M C, Pedersen A B. Validation of 14,500 operated knees registered in the Danish Knee Ligament Reconstruction Register: Registration completeness and validity of key variables. Clin Epidemiol 2013a; 5: 219-28. Rahr-Wagner L, Thillemann T M, Pedersen A B, Lind M C. Increased risk of revision after anteromedial compared with transtibial drilling of the femoral tunnel during primary anterior cruciate ligament reconstruction: Results from the Danish Knee Ligament Reconstruction Register. Arthroscopy 2013b; 29 (1): 98-105. Rose T, Hepp P, Venus J, Stockmar C, Josten C, Lill H. Prospective randomized clinical comparison of femoral transfixation versus bioscrew fixation in hamstring tendon ACL reconstruction: A preliminary report. Knee Surg Sports Traumatol Arthrosc 2006; 14 (8): 730-8. Scannell B P, Loeffler B J, Hoenig M, Peindl R D, D’Alessandro D F, Connor P M, Fleischli J E. Biomechanical comparison of hamstring tendon fixation devices for anterior cruciate ligament reconstruction, Part 2: Four tibial devices. Am J Orthop 2015; 44 (2): 82-5. To J T, Howell S M, Hull M L. Contributions of femoral fixation methods to the stiffness of anterior cruciate ligament replacements at implantation. Arthroscopy 1999; 15 (4): 379-87. www.aclregister.nu. Swedish ACL Register. Annual Report 2014. Ytterstad K, Granan L P, Ytterstad B, Steindal K, Fjeldsgaard K A, Furnes O, Engebretsen L. Registration rate in the Norwegian Cruciate Ligament Register: Large-volume hospitals perform better. Acta Orthop 2012; 83 (2): 174-8.

21-02-2018 15:44:47

Acta Orthopaedica 2018; 89 (2): 211–216

211

Comparison between modified Dunn procedure and in situ fixation for severe stable slipped capital femoral epiphysis A retrospective study of 29 hips followed for 2–7 years Giovanni TRISOLINO 1, Stefano STILLI 1, Giovanni GALLONE 1, Pedro SANTOS LEITE 2, and Giovanni PIGNATTI 3

1 Department

of Pediatric Orthopaedics and Traumatology, Rizzoli Orthopaedic Institute, Bologna, Italy; 2 Department of Orthopaedics, Centro Hospitalar do Porto – Hospital de Santo António, Porto, Portugal; 3 Department of Revision Surgery of Hip Prosthesis and Development of New Implants, Rizzoli Orthopaedic Institute, Bologna, Italy Correspondence: giovanni.trisolino@ior.it Submitted 2017-07-05. Accepted 2017-11-18.

Background and purpose — The best treatment option for severe slipped capital femoral epiphysis (SCFE) is still controversial. We compared clinical and radiographic outcomes of modified Dunn procedure (D) and in situ fixation (S) in severe SCFE. Patients and methods — We retrospectively compared D and S, used for severe stable SCFE (posterior sloping angle (PSA) > 50°) in 29 patients (15 D; 14 S). Propensity analysis and inverse probability of treatment weights (IPTW) to adjust for baseline differences were performed. Patients were followed for 2–7 years. Results — Avascular necrosis (AVN) occurred in 3 patients out of 15, after D, causing conversion to total hip replacement (THR) in 2 cases. In S, 1 hip developed chondrolysis, requiring THR 3 years after surgery. 3 symptomatic femoroacetabular impingements (FAI) occurred after S, requiring corrective osteotomy in 1 hip, and osteochondroplasty in another case. The risk of early re-operation was similar between the groups. The slippage was corrected more accurately and reliably by D. The Nonarthritic Hip Score was similar between groups, after adjusting for preoperative and postoperative variables. Interpretation — Although D was superior to S in restoring the proximal femoral anatomy, without increasing the risk of early re-operation, some concern remains regarding the potential risk of AVN in group D.

(Loder et al. 2012). Nonetheless, severe SCFE may lead to functional impairment, femoroacetabular impingement (FAI), and progression to osteoarthritis (Ganz et al. 2003). Although the optimal treatment of these cases remains controversial, some authors agree that a severe displacement should be corrected during the same operation (Ziebarth et al. 2009, Novais et al. 2015, Sikora-Klak et al. 2017, Trisolino et al. 2017). In recent years, the modified Dunn procedure (D) by means of surgical hip dislocation (Leunig et al. 2007, Ziebarth et al. 2009) has gained in popularity. The technique stabilizes the epiphysis and corrects the deformity in a single intervention, by restoring the femoral head–neck anatomy, possibly avoiding FAI sequelae. Furthermore, in experienced hands, D is safe, with low complication rates (Leunig et al. 2007, Ziebarth et al. 2009, 2013, Slongo et al. 2010, Skin et al. 2011, Novais et al. 2015). Nonetheless, some authors report increased rates of avascular necrosis (AVN) of the femoral head and major complications, requiring revision surgery and total hip replacement (THR) in the short follow-up (Alves et al. 2012, Sankar et al. 2013, Souder et al. 2014, Javier et al. 2017, Sikora-Klak et al. 2017). We compared the clinical and radiographic outcomes of D and S in severe stable SCFE.

■

The slipped capital femoral epiphysis (SCFE) generally requires surgical management to stabilize the epiphysis, achieving early fusion of the proximal femoral physis and avoiding further displacement and deformity. In situ screw fixation (S) remains the most common treatment for stable SCFE, regardless of the degree of deformity

Patients and methods Study design We retrospectively studied prospectively gathered data, comparing D and S in severe stable SCFE. The D was introduced in 2011 in our center; the learning curve of a senior surgeon (a well-trained hip specialist, who habitually performs approxi-

11875 Trisolino D.indd 211

21-02-2018 15:51:27

212

Acta Orthopaedica 2018; 89 (2): 211–216

Patients with SCFE who underwent surgery between January 2011 and July 2015 114 hips in 108 patients Excluded Patients with mild to moderate SCFE (PSA < 50°) a 65 hips in 63 patients Patients with moderate to severe SCFE (PSA > 50°) 49 hips in 45 patients Excluded Patients with unstable SCFE b 11 hips in 11 patients Patients with stable moderate to severe SCFE 38 hips in 34 patients Excluded (9 hips in 5 patients): – underwent Imhauser ITO, 3 – underwent CRIF, 6 Eligible for study (29 hips in 29 patients): – underwent modified Dunn procedure (D group), 15 – underwent in situ fixation (S group), 14

Figure 1. Flow-chart of eligibility criteria. SCFE = slipped capital femoral epiphysis. PSA = posterior sloping angle. ITO = inter-trochanteric osteotomy. CRIF = closed reduction and internal fixation. a All cases with mild to moderate SCFE underwent S. b 8 patients were treated with CRIF, 1 patient with D, 2 patients with S.

mately 300 hip operations per year) was supervised by 2 surgeons, highly experienced in D, who also participated in the majority of the operations. A local registry, with continuous data on patients affected by SCFE, was created and implemented. Our center is a tertiary referral hospital for pediatric orthopedics and traumatology, treating 15 to 25 new cases of SCFE per year. Eligibility Patients with severe stable SCFE (posterior sloping angle (PSA) > 50°), undergoing D, were included in the study. During the same period a cohort of patients undergoing S to address severe SCFE, according to surgeons’ preference and experience, was also included into the study. Because the surgeons took turns on duty, the choice of procedure depended on the available surgeon and his/her personal preference, thus the patients were not randomized. We excluded mild to moderate SCFE (PSA < 50°) and unstable SCFE, based on Loder’s criteria (Loder et al. 1993). We also excluded cases treated by different techniques such as closed reduction and internal fixation (CRIF) or intertrochanteric osteotomy (ITO) (Figure 1). No cases with associated comorbidities were identified in this series (for clinical data on all patients, see Table 1, Supplementary data).

11875 Trisolino D.indd 212

Table 2. Patients’ demographic and preoperative data Clinical parameters Male/female, n Side L/R, n Age at surgery (years) (SD) Time elapsed from initial symptoms (months) BMI (SD) BMI-for-age (percentile) Preoperative PSA (SD; °) Follow-up (SD)

D group

S group

p-value

11/4 7/8 13.9 (2.3)

11/3 10/4 13.0 (1.0)

1.0 0.2 0.4

12 (9) 24 (4) 81 (21) 68 (11) 3.7 (1.1)

6 (5) 24 (4) 82 (23) 62 (9) 4.9 (1.2)

0.05 0.9 0.9 0.2 < 0.01

BMI = body mass index. BMI-for-age percentiles were calculated according to the World Health Organization charts and tables (www.who.int/growthref/ who2007) PSA = posterior sloping angle.

From January 2011 to July 2015, 108 patients (114 hips) affected by SCFE were admitted to our department. Among them 29 hips in 29 patients were eligible for inclusion in the study. 15 patients (15 hips) underwent D and 14 patients (14 hips) underwent S (Table 2). The mean follow-up was 4.3 (range 2–6.5) years. Surgical technique Modified Dunn procedure The modified Dunn procedure was performed according to the previously described technique (Leunig et al. 2007, Ziebarth et al. 2009). Postoperatively, patients follow a non-weightbearing protocol for 6 weeks followed by protected weightbearing with crutches for 6 more weeks. In situ fixation Via an anterolateral skin incision a guidewire was introduced, under fluoroscopic guidance, in the anterior aspect of the femoral neck aiming toward the center of the femoral head in the AP and lateral projections. 2 4.5 mm fully threaded screws were placed, transfixing the femoral physis. Patients were allowed partial weight-bearing with crutches for 6 weeks. Follow-up Patients were followed for a median 4 (2–7) years. The Nonarthritic Hip Score (NAHS) was used to assess clinical and functional outcomes of patients at the latest follow-up (Christensen et al. 2003). The PSA and the alpha angle on a frog-lateral view of the hip were used to assess the degree of correction and the radiographic presence of residual FAI deformity in the 2 groups. Statistics Continuous data were expressed as mean (SD), whereas categorical and ordinal data were expressed as proportions and 95% confidence interval (CI). Normality was tested using the chi-square test for categorical variables and the Kolmogorov–

21-02-2018 15:51:27

Acta Orthopaedica 2018; 89 (2): 211–216

Smirnov test for continuous variables. Differences in baseline and outcome characteristics between D and S were tested using Fisher’s exact test for categorical variables and the Student’s t-test (normal data distribution) or the Mann–Whitney U-test (skewed data distribution) for continuous variables. Exploratory univariable and multivariable analyses with Cox regression and general linear models were performed to assess the impact of the different treatments on the final outcomes. Results were presented as crude and adjusted relative risk (RR), for the dichotomous outcome variables and crude and adjusted mean for the continuous outcome variables; CI were reported for both. To account for potential selection biases between the two groups, a propensity analysis was performed (Pirracchio et al. 2012). For each patient, we estimated propensity scores (PS) for receiving D or S, using a binary logistic model that included baseline variables (body mass index (BMI)-for-age percentile, according to the World Health Organization charts and tables; see www.who.int/growthref/who2007), time elapsed from initial symptoms and preoperative PSA). We included in the PS model only baseline variables hypothetically related to the outcome. The balance of the PS was checked observing the overlap in the range of propensity scores across the two treatments and comparing the quintiles (Garrido et al. 2014). 1 extreme case not overlapping in the common support (minima and maxima comparison) was excluded from the adjusted analysis. T-test showed no statistically significant differences in covariate means between the 2 treatment groups after matching. Examining treatment effects on the outcome across PS quintiles, we did not observe any association between the outcome and the probability of receiving either treatment, which means that there is no evidence of unmeasured bias (Brookhart et al. 2013). Propensity scores were then used to derive inverse probability of treatment weights (IPTW), with the inverse of the propensity score for patients with D and the inverse of 1 minus the propensity score for patients with S. Then, the IPTW was used to adjust the RR for early revision surgery in the two groups. A p-value of < 0.05 was considered to be statistically significant, and all reported p-values are 2-sided. We used Excel (Microsoft, Redmond, WA, USA) and SPSS (version 22.0; IBM Corp, Armonk, NY, USA) for analysis. Ethics, funding, and potential conflicts of interest The study was authorized by the local Ethical Committee. The data for this investigation were collected and analyzed in compliance with the procedures and policies set forth by the Helsinki Declaration, and all patients gave their informed consent to data treatment. No competing interests declared.

213

groups regarding sex distribution, age at presentation, BMI and BMI-for-age percentile, time elapsed from initial symptoms, and preoperative PSA. The mean follow-up was shorter in D (3.7 (2–5.3) years) compared with S (4.9 (2.9–6.5) years), since the D was increasingly practiced in the latest years. The follow-up duration was considered as an independent covariate in regression analysis, as the difference between groups could influence the outcomes. In D, 3/15 patients (20%; CI 5–49) developed AVN, 3 to 6 months after surgery. 2 of them underwent THR, 1.4 and 1.8 years after D respectively, while another patient was scheduled for the same procedure at the latest follow-up visit. 1 patient developed mild asymptomatic heterotopic ossifications. No patients in the S group developed AVN In S, 1 case (7%; CI 4–36) developed chondrolysis 3 months after surgery, with severe pain, stiffness, and limping that did not regress at 3 years’ follow-up; the patient finally underwent THR elsewhere, 3 years after S. 3 other cases (21%; CI 6–51) developed symptomatic FAI with pain, stiffness, poor range of motion, and limping: one patient underwent sub-capital osteotomy by mean of surgical dislocation 1.2 year after S; the patient further developed septic nonunion, requiring THR 3 years after S. Another patient underwent femoral osteochondroplasty and labral repair 2.9 years after S. Finally, 1 patient was scheduled for corrective osteotomy but refused the operation at the latest follow-up visit. The overall rate of re-operation was 17% (13% (CI 4–38) in D; 21% (CI 6–51) in S). The risk of early re-operation was similar in both groups. The slippage was corrected more accurately and reliably by D (Figure 2). The mean lateral alpha angle (50° (15) versus 89° (13); p-value < 0.0001) and the median PSA (9° (8) versus 38° (20); p-value < 0.0001) were lower in the D group compared with the S group (Table 3, see Supplementary data). The mean difference remained statistically significant after adjustment for IPTW and follow-up duration. The Nonarthritic Hip Score (NAHS) was similar between groups at the latest follow-up, even after adjusting for IPTW and follow-up duration (Table 4, see Supplementary data), No interactions between NAHS (total and subscales) and preoperative or postoperative variables were found (all p-values for interaction tests > 0.05). 11 patients were able to return to a low to moderate impact sport activity (bicycling, swimming, aerobics, running), while 8 patients could return to highimpact non-professional sport activity (soccer); there was a similar distribution between groups.

Discussion Results No statistically significant differences were found between

11875 Trisolino D.indd 213

We describe our initial experience with the modified Dunn procedure (D) , comparing this technique with in situ fixation (S), which is the traditional treatment for stable SCFE.

21-02-2018 15:51:27

214

Acta Orthopaedica 2018; 89 (2): 211–216

Figure 2. Case no. 4. Severe stable SCFE in a 14-year-old boy treated by modified Dunn procedure: a–b: preoperative radiographs; c–d: postoperative radiographs; e–f: 3 years’ follow-up after hardware removal. Mild heterotopic ossification can be observed at the latest follow-up.

A first issue is the lack of a clear indication for subcapital realignment among the different types and grades of SCFE. Some authors encourage the use of this procedure in unstable SCFE, due to a substantial decrease of AVN (Leunig et al. 2007, Slongo et al. 2010, Sankar et al. 2013, Ziebarth et al. 2013), while other authors warn about a doubled risk of AVN following D in unstable hips (Alves et al. 2012). Some authors advocate the use of D only in cases of severe SCFE (Madan et al. 2013, Novais et al. 2015), because of the general favorable prognosis of S in mild SCFE, while other authors suggest D also in cases of mild SCFE, to avoid FAI-induced chondrolabral damage (Ziebarth et al. 2013). Thus an evidence-based algorithm of treatment is still lacking. We started our experience treating cases of severe stable SCFE. We hypothesized that the potential risk of AVN could be balanced by the expected benefit of a substantial reduction of the residual impingement. To test our hypothesis, we matched cases of D with cases of S. S is the most widely accepted treatment for SCFE at any grade of severity: the technical ease, the low risk of complications, and the potential role of bone remodeling, with spontaneous improvement of the deformity, confirmed this technique as the gold standard of SCFE treatment for years (Loder et al. 2012). As expected, the most frequent complication we encountered in the D group was AVN, which occurred in 3/15 hips but

11875 Trisolino D.indd 214

in none of 14 hips operated with S. With the small numbers available, the difference was not statistically significant, but is still a concern. The issue of the learning curve in D has been highlighted in other studies. Upasani et al. (2014) reported an important association between the surgeon’s volume and experience in performing D and the likelihood of major complications. In our series the rate of AVN was consistent with previous reports of surgeons at the beginning of the learning curve (Alves et al. 2012, Sankar et al. 2013, Souder et al. 2014, Upasani et al. 2014, Javier et al. 2017, Sikora-Klak et al. 2017), confirming that D is technically demanding. Conversely, in a previous report, we excluded a substantial impact of surgeon’s volume and experience in Imhauser ITOs, as 16 different surgeons performed 53 procedures across the years (Trisolino et al. 2017). The most important difference between the 2 procedures is the restoration of the head–neck anatomy, provided by D. We observed better radiographic outcomes in D. Whether or not some remodeling was observed also after S, this effect was inconstant and incomplete, leaving an important deformity in the majority of cases, causing symptomatic FAI in 3/14 patients, leading to early re-operation or THR. FAI has been associated with increased pain, reduced ROM, chondrolabral damage, and early hip osteoarthritis (Ganz et al. 2008, Agricola et al. 2013, Castaneda et al. 2013). Also,

21-02-2018 15:51:27

Acta Orthopaedica 2018; 89 (2): 211–216

the residual deformity in SCFE has been associated with early chondrolabral damage (Lee et al. 2013, Klit et al. 2014, Wylie et al. 2015). FAI following the surgical treatment of SCFE often requires early revision surgery, even THR, similarly to AVN. 2 other retrospective non-matched studies compared D and S in the treatment of severe stable SCFE (Souder et al. 2014, Novais et al. 2015), while 1 retrospective study compared D with ITO (Sikora-Klak et al. 2017). Souder et al. (2014) compared 64 stable SCFE treated by S and 10 stable SCFE treated by D and reported 2 cases of AVN in D and no cases in S. The authors concluded that stable SCFE should be better treated with S to minimize the risk of AVN, postponing the correction of the deformity, if need be. Conversely, Novais et al. (2015) compared 15 severe stable SCFE, treated by D, with a historical series of 15 severe stable SCFE treated by S. They found a better radiographic correction after D and no substantial differences in rate of AVN between the two groups but an increased number of re-operations and worse clinical outcomes following S. Recently Sikora-Klak et al. (2017) retrospectively compared 14 D and 12 Imhauser ITO, performed for high-grade stable SCFE. They found 3/14 AVN in the D group, but no difference in the rate of complications and a slightly higher rate of reoperations following ITOs. In their experience, ITO allowed acceptable correction of the deformity, compared with D. Our study has several limitations. The retrospective design and lack of randomization introduced potential selection biases. In particular, the follow-up duration (> 1 year shorter in the D group) may underestimate the rate of re-operation in D. Conversely the minimum follow-up of 2 years, adequate to estimate the incidence of some postoperative complications (AVN, nonunion, slip progression), could be insufficient to evaluate the effect of FAI on the onset of early osteoarthritis. An increase of re-operations and THR could be expected in S, because of the high rate of severe FAI (Agricola et al. 2013). The sample size is underpowered to evaluate, for example, the differences in AVN incidence. Finally, the absence of MRI confirmation of the femoral head vitality does not allow any definitive conclusion to be drawn regarding the onset of AVN as a complication of the treatment rather than the disease. Nevertheless, based on our results, we can affirm that D is superior to S in restoring the proximal femoral anatomy, without increasing the rate of revision surgery at short-term follow-up. Additional studies are needed to assess the longterm effectiveness of D, and also in comparison with extraarticular osteotomies, which are technically easier, showing low rates of major complications and satisfactory long-term results (Trisolino et al. 2017). Supplementary data Tables 1, 3, and 4 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/ 17453674.2018.1439238.

11875 Trisolino D.indd 215

215

GT, SS and GP conceived and planned the project. GT carried out the implementation of the database, analyzed the data, performed the calculations, interpreted the results with critical feedback by all authors, and wrote the manuscript with input from all authors. GG and PL contributed to the data collection and database implementation, data analysis and calculations, and helped in writing and editing the manuscript.

We would like to thank Dr Elettra Pignottti for her valuable assistance in the statistical analysis.

Acta thanks Randall Loder and other anonymous reviewers for help with peer review of this study.

Agricola R, Heijboer M P, Bierma-Zeinstra S M, Verhaar J A, Weinans H, Waarsing J H. Cam impingement causes osteoarthritis of the hip: A nationwide prospective cohort study (CHECK) Ann Rheum Dis 2013;72 (6): 918-23. Alves C, Steele M, Narayanan U, Howard A, Alman B, Wright J G. Open reduction and internal fixation of unstable slipped capital femoral epiphysis by means of surgical dislocation does not decrease the rate of avascular necrosis: A preliminary study. J Child Orthop 2012; 6 (4): 277-83. doi: 10.1007/s11832-012-0423-1. Brookhart M A, Wyss R, Layton J B, Stürmer T, Propensity score methods for confounding control in nonexperimental research. Circ Cardiovasc Qual Outcomes 2013; 6 (5): 604-11. doi: 10.1161/CIRCOUTCOMES.113.000359. Castaneda P, Ponce C, Villareal G, Vidal C. The natural history of osteoarthritis after a slipped capital femoral epiphysis/the pistol grip deformity. J Pediatr Orthop 2013; 33 (Suppl 1): S76-82. Christensen C P, Althausen P L, Mittleman M A, Lee J A, McCarthy J C. The nonarthritic hip score: Reliable and validated. Clin Orthop Relat Res 2003; (406): 75-83. Dunn D M. The treatment of adolescent slipping of the upper femoral epiphysis. J Bone Joint Surg Br 1964; 46: 621-9. Ganz R, Parvizi J, Beck M, Leuning M, Notzli H, Siebenrock K A. Femoroacetabular impingement: A cause for osteoarthritis of the hip. Clin Orthop Relat Res 2003; 417: 112-20. Ganz R, Leunig M, Leunig-Ganz K, Harris W H. The etiology of osteoarthritis of the hip: An integrated mechanical concept. Clin Orthop Relat Res 2008; 466 (2): 264-72. Garrido M M, Kelley A S, Paris J, Roza K, Meier D E, Morrison R S, Aldridge M D. Methods for constructing and assessing propensity scores. Health Serv Res 2014; 49 (5): 1701-20. doi: 10.1111/1475-6773.12182. Javier M J, Victoria A, Martín D, Gabriela M, Claudio A F. Treatment of slipped capital femoral epiphysis with the modified Dunn procedure: A multicenter study. J Pediatr Orthop 2017; Jan 18. doi: 10.1097/ BPO.0000000000000936. Klit J, Gosvig K, Magnussen E, Gelineck J, Kallemose T, Søballe K, Troelsen A. Cam deformity and hip degeneration are common after fixation of a slipped capital femoral epiphysis. Acta Orthop 2014; 85(6): 585-91. Lee C B, Matheney T, Yen Y M. Case reports: acetabular damage after mild slipped capital femoral epiphysis. Clin Orthop Relat Res 2013; 471 (7): 2163-72. doi: 10.1007/s11999-012-2715-7. Leunig M, Slongo T, Kleinschmidt M, Ganz R. Subcapital correction osteotomy in slipped capital femoral epiphysis by means of surgical hip dislocation. Oper Orthop Traumatol 2007; 19: 389-410. Loder R T, Richards B S, Shapiro P S, et al. Acute slipped capital femoral epiphysis: The importance of physeal stability. J Bone Joint Surg Am 1993; 75-A: 1134-40.

21-02-2018 15:51:28

216

Loder R T, Dietz F R. What is the best evidence for the treatment of slipped capital femoral epiphysis? J Pediatr Orthop 2012; 32 (Suppl 2): S158-65. doi: 10.1097/BPO.0b013e318259f2d1. Madan S S, Cooper A P, Davies A G, et al. The treatment of severe slipped capital femoral epiphysis via the Ganz surgical dislocation and anatomical reduction: A prospective study. Bone Joint J 2013; 95-B: 424-9. Novais E N, Hill M K, Carry P M, et al. Modified Dunn procedure is superior to in situ pinning for short-term clinical and radiographic improvement in severe stable SCFE. Clin Orthop Rel Res 2015; 473: 2108-17. Pirracchio R, Resche-Rigon M, Chevret S. Evaluation of the propensity score methods for estimating marginal odds ratios in case of small sample size. BMC Med Res Methodol 2012; 12: 70. doi: 10.1186/1471-2288-12-70. Sankar W N, Vanderhave K L, Matheney T, Herrera-Soto J A, Karlen J W. The modified Dunn procedure for unstable slipped capital femoral epiphysis: A multicenter perspective. J Bone Joint Surg Am 2013; 95: 585-91. doi: 10.2106/JBJS.L.00203 Sikora-Klak J, Bomar J D, Paik C N, Wenger D R, Upasani V. Comparison of surgical outcomes between a triplane proximal femoral osteotomy and the modified Dunn procedure for stable, moderate to severe slipped capital femoral epiphysis. J Pediatr Orthop 2017; Mar 10. doi: 10.1097/ BPO.0000000000000968. Slongo T, Kakaty D, Krause F, et al. Treatment of slipped capital femoral epiphysis with a modified Dunn procedure. J Bone Joint Surg Am 2010; 92 (18): 2898-908. Skin E, Beaule P E, Sucato D, et al. Multicenter study of complications following surgical dislocation of the hip. J Bone Joint Surg Am 2011; 93 (12): 1132-6.

11875 Trisolino D.indd 216

Acta Orthopaedica 2018; 89 (2): 211â&#x20AC;&#x201C;216

Souder C D, Bomar J D, Wenger D R. The role of capital realignment versus in situ stabilization for the treatment of slipped capital femoral epiphysis. J Pediatr Orthop 2014; 34 (8): 791-8. doi: 10.1097/BPO.0000000000000193. Trisolino G, Pagliazzi G, Di Gennaro G L, Stilli S. Long-term results of combined epiphysiodesis and Imhauser intertrochanteric osteotomy in SCFE: A retrospective study on 53 hips. J Pediatr Orthop 2017; 37 (6): 409-15. doi: 10.1097/BPO.0000000000000695. Upasani V V, Matheney T H, Spencer S A, Kim Y J, Millis M B, Kasser J R. Complications after modified Dunn osteotomy for the treatment of adolescent slipped capital femoral epiphysis. J Pediatr Orthop 2014; 34 (7): 661-7. doi: 10.1097/BPO.0000000000000161. Wylie J D, Beckmann J T, Maak T G, Aoki S K. Arthroscopic treatment of mild to moderate deformity after slipped capital femoral epiphysis: intraoperative findings and functional outcomes. Arthroscopy 2015; 31(2): 24753. doi: 10.1016/j.arthro.2014.08.019. Ziebarth K, Zilkens C, Spencer S, Leunig M, Ganz R, Kim Y J. Capital realignment for moderate and severe SCFE using a modified Dunn procedure. Clin Orthop Relat Res 2009; 467 (3): 704-16. doi: 10.1007/s11999008-0687-4. Ziebarth K, Leunig M, Slongo T, Kim Y J, Ganz R. Slipped capital femoral epiphysis: Relevant pathophysiological findings with open surgery. Clin Orthop Relat Res 2013; 471 (7): 2156-62. doi: 10.1007/s11999-013-28189. Ziebarth K, Milosevic M, Lerch T D, Steppacher S D, Slongo T, Siebenrock K A. High survivorship and little osteoarthritis at 10-year follow-up in SCFE patients treated with a modified Dunn procedure. Clin Orthop Relat Res 2017; 475 (4): 1212-28. doi: 10.1007/s11999-017-5252-6.

21-02-2018 15:51:28

Acta Orthopaedica 2018; 89 (2): 217–221

217

Good inter- and intraobserver reliability for assessment of the slip angle in 77 hip radiographs of children with a slipped capital femoral epiphysis Bengt HERNGREN 1, 2, Mikael LINDELL 2, and Gunnar HÄGGLUND 1

1 Department of Orthopedics, Lund University, Lund, Sweden; 2 Department of Orthopedics, Ryhov County Hospital, Jönköping, Sweden Correspondence: bengt.herngren@med.lu.se Submitted 2017-07-30. Accepted 2017-10-30.

Background and purpose — The decision on and the outcome of treatment for a slipped capital femoral epiphysis (SCFE) depend on the severity of the slip. In 2015, web-based registration was introduced into the Swedish Pediatric Orthopedic Quality (SPOQ) register. To determine whether the inclusion of commonly used methods in Sweden for radiographic measurement of SCFE (the calcar femorale [CF] method and the Billing method) is justified, we measured the inter- and intraobserver reliability of these 2 measurements. We also evaluated the internationally more commonly used head-shaft angle (HSA) method. Material and methods — 4 observers with different levels of experience with radiographic measurements analyzed 77 routine preoperative hip radiographs of children with SCFE. Inter- and intraobserver reliability was evaluated. Results — The interobserver reliability analysis for the 4 observers showed for CF an ICC of 0.99 (CI 0.97–0.99) and for Billing an ICC of 0.99 (CI 0.98–0.99). The interobserver reliability analysis for 2 observers showed for HSA an ICC of 0.98 (CI 0.97–0.99). Intraobserver reliability (2 observers) showed a mean difference below 1° for all 3 methods and with a 95% limit of agreement not exceeding ±6.8°. Interpretation — We found good reliability for both intra- and interobserver measurements of all 3 methods used for the assessment of the slip angle on routine preoperative lateral hip radiographs. ■

Slipped capital femoral epiphysis (SCFE) is the most common hip disorder in children aged 9–15 years (Loder 1996, Lehm-

ann et al. 2006). SCFE is caused by the displacement between the epiphysis and the metaphysis of the proximal femur. The epiphysis remains in the acetabulum while the femur usually rotates outward and in extension (Jerre 1995, Loder 2001). The recommended method of treatment (Souder et al. 2014, Loder 2017) and the outcome (Kocher et al. 2004, Larson et al. 2010, Terjesen and Wensaas 2017) depend on the severity of the slip. The reliability of the methods used to measure the slip angle is therefore important. Begun in 2015, the Swedish Pediatric Orthopedic Quality register (SPOQ, www.spoq.se) is now a web-based registration tool for 5 pediatric orthopedic conditions. For SCFE, the surgeon is requested to register certain variables including the preoperative slip angle. 2 methods for measurement of the slip angle in SCFE are used in the SPOQ register: the calcar femorale (CF) method (Hansson et al. 1988) and the Billing method (Billing et al. 2002). The aim of this study was to determine whether the reliability of the CF and Billing methods justifies their use in the register. We hypothesized that the intraobserver and interobserver reliability for these methods would warrant their use in SPOQ. We also included a comparison between the CF and the internationally more commonly used head–shaft angle (HSA) method (Southwick 1967).

Material and methods Conventional radiographs from 94 consecutively registered children with SCFE included in the SPOQ register during 2013 and 2014 were assessed. The radiographs used were routine preoperative examinations from all Swedish hospitals

11961 Herngren D.indd 217

21-02-2018 15:52:47

218

Acta Orthopaedica 2018; 89 (2): 217–221

Eligible radiographs n = 94 Excluded Only pelvic AP n=6 Remaining radiographs n = 88 Lauenstein view n = 61

Figure 1. Slip angle measured using the calcar femorale method in a Lauenstein view. 1. Identify the calcar femorale (cf) and the lesser trochanter ➀. 2. From the level of the lesser trochanter draw a line ➁ three cm in a proximal direction parallell to the calcar femorale. 3. Extend a line ➂ parallell to line ➁ up through the femoral neck. 4. Define a line ➃ through the physeal anterior and posterior margins. 5. Draw a line ➄ perpendicular to line ➃. 6. Slip angle ➅.

Figure 2. Slip angle measured using the Billing method in Billing lateral view with the patient positioned according to the figure to the right. 1. Draw a line ➀ along the anterior cortex of the proximal femur. Extend the line up through the femoral head and neck. 2. Draw a line ➁ along hte anterior border of the femoral neck. 3. Draw the bisector ➂ to lines ➀ and ➁. 4. Define a line ➃ through the physeal anterior and posterior margins. 5. Draw a line ➄ perpendicular to line ➂. 6. The slip angle ➅ is the angle between lines ➃ and ➄.

Figure 3. Slip angle measured using the lateral head–shaft angle method in a Lauenstein view. 1. Draw a line ➀, parallell with the proximal femoral shaft, further up into the femoral neck. 2. Define a line ➁ through the physeal anterior and posterior margins. 3. Draw a line ➂ perpendicular to line ➁. 4 Lateral head–shaft angle ➃.

11961 Herngren D.indd 218

Excluded (n = 11): – poor image quallity, 9 – < 2 cm of proximal femur, 2 Billing view n = 27

Lauenstein view n = 50

Figure 4. Study material.

that treated children with SCFE during this period. 1 fourthyear resident in orthopedics (ML – Observer 1), 1 specialist in orthopedics (BH – Observer 2), and 2 specialists in pediatric radiology (Observers 3 and 4) acted as observers. To obtain both presumptive normal hips and hips with SCFE in the study material, we chose the right hip for assessment for every second radiograph in the consecutive list irrespective of whether it was a hip with SCFE or a normal hip. The webbased instructions, available through the SPOQ, were used (Figures 1 and 2) together with a similar instruction on how to measure the lateral HSA (Figure 3). For the Lauenstein view (both hips), horizontal rotational alignment with an obturator index between 0.7 and 1.8 (Tönnis 1976) and at least 2 cm of the proximal femur below the lesser trochanter had to be included (Lehmann et al. 2013). For the Billing lateral view, radiographs were accepted if the lesser trochanter was not protruding posteriorly or anteriorly. A correct rotational alignment was emphasized by the developer of this method to be a crucial factor (Billing et al. 2002). According to these criteria, 50 radiographs in the Lauenstein view and 27 in the Billing lateral view were included in the analysis (Figure 4). Interobserver reliability The orthopedic resident and the orthopedic specialist (Observers 1 and 2), respectively, used their standard picture archiving and communication system (PACS). The orthopedic specialist was experienced in the use of both the CF and Billing methods. The orthopedic resident had no previous experience in any method. The pediatric radiologists (Observers 3 and 4) assessed all radiographs using their standard PACS. They were both experienced in using the Billing method but not the CF or HSA methods. The radiographs were all given a unique number in a list that did not follow any alphabetical order or pattern according to age, date, or sex. The observers were blinded to the measurements made by the other observers, radiographic reports,

21-02-2018 15:52:47

Acta Orthopaedica 2018; 89 (2): 217–221

information from medical records, or their own previous measurements. They were allowed to use their own preferred screen settings. Standardization of the measurements was performed before the study by thorough discussion and interpretation of the instructions. The orthopedic resident measured 20 different pelvic radiographs under supervision of Observer 2. We used these instructions for a single measurement for all 4 readers for the Billing and CF methods. Intraobserver reliability Observers 1 and 2 also measured the lateral HSA and assessed the radiographs twice each following the Billing method, the CF method, and the HSA. An interval of at least 6 weeks was used between the repeated measurements for analysis of intraobserver reliability. Statistics We assumed a t-distribution for a sample size of < 50 and a normal distribution for a sample size of ≥ 50. The effect size was set to 3° with 90% power and with a confidence level of 99%. The expected standard deviation was derived from a similar study (Carney and Liljenquist 2005). Intraobserver variation for each of the measurements was assessed using the mean difference, with its 95% limits of agreement (Bland and Altman 1986, Lehmann et al. 2013). For the purpose of graphic presentation, we plotted the differences against the mean measurements (Bland–Altman plots). Interobserver variation for 2 observers measuring HSA was assessed using the intraclass correlation coefficient (ICC) and 95% confidence interval (CI) with 2-way random and absolute agreement for single measures. The first measurements were used for both observers (McGraw and Wong 1996). For the 4 observers measuring Billing and CF, interobserver reliability was evaluated using the intraclass correlation coefficient (ICC) and CI with 2-way random and absolute agreement for average measures. The first measurements were used for all observers (McGraw and Wong 1996, Hermanson et al. 2017). When comparing the HSA with the CF method, we used the first measurements for both methods. For statistical analysis, the variability was described using the Bland–Altman method, with its 95% limits of agreement (Sedgwick 2013). IBM SPSS Statistics for Windows version 24 (IBM Corp, Armonk, NY, USA) was used for the statistical analyses. Ethics, funding, and potential conflicts of interest Ethical approval was authorized by the Regional Ethical Review Board in Lund, Sweden (registration number 2013/87). Informed consent was obtained from all participants and from one parent or guardian. Funding was received from the Swedish Association of Local Authorities and Regions (SKL), and the Futurum Academy for Health and Care, Jönköping County Council, Jönköping. The authors declare no conflicts of interest.

11961 Herngren D.indd 219

219

Table 1. Intraobserver variation between first and second measurements of slip angle (°)

Intraobserver Observer 1 – Billing Observer 2 – Billing Observer 1 – CF Observer 2 – CF Observer 1 – HSA Observer 2 – HSA

Subjects

Difference mean (SD)

95% limits of agreement

27 27 50 50 50 50

–0.8 (2.9) –0.2 (1.9) 0.0 (2.6) 0.2 (1.4) 0.1 (1.1) 0.4 (3.3)

–6.5 to 5.0 –4.0 to 3.5 –5.1 to 5.1 –2.3 to 3.0 –2.1 to 2.3 –6.0 to 6.8

Table 2. Difference (°) between HSA and CF measurement

HSA – CF

Subjects

Difference mean (SD)

95% limits of agreement

Observer 1 Observer 2

50 50

5.9 (4.8) 3.4 (4.8)

–3.5 to 15.3 –6.1 to 12.9

Results The mean slip angles for the different methods used were: 23° (3° to 59°) for the Billing method, 23° (–8° to 81°) for the CF method, and 26° (–7° to 89°) for the HSA method. The interobserver reliability analysis for 4 observers showed for CF an ICC of 0.99 (CI 0.97–0.99) and for Billing an ICC of 0.99 (CI 0.98–0.99). The interobserver reliability analysis for 2 observers showed for HSA an ICC of 0.98 (CI 0.97–0.99). Intraobserver reliability analysis for 2 observers showed a mean difference between the first and second measurement of less than one degree for all three methods. The 95% limits of agreement ranged between –6.5° and 6.8° (Table 1). Bland–Altman plots for HSA and CF visualize the proximity achieved between the first and second measurements (Figures 5 and 6). The mean difference between the first measurements of HSA and CF was below 6° for 2 observers (Table 2).

Discussion We found good inter- and intraobserver reliability for all 3 methods for assessing the slip angle on routine preoperative hip radiographs. The HSA method showed an acceptable inter- and intraobserver reliability. The HSA method produced on average a higher value for the slip angle compared with the CF method. The 95% limit of agreement between the 2 methods also showed a rather wide range of 19° for both observers. On the other hand, the HSA method for observer 1 together with the CF method for observer 2 showed the highest intraobserver

21-02-2018 15:52:50

220

Acta Orthopaedica 2018; 89 (2): 217–221

Figure 5. Intraobserver variation (°) for HSA – Observer 1 (left panel). The solid line represents the mean value and the dotted lines show the limits for 2 standard deviations above and below the mean value.

Figure 6. Intraobserver variation (°) for CF – Observer 2 (right panel). The solid line represents the mean value and the dotted lines show the limits for 2 standard deviations above and below the mean value.

reliability. These aspects should all be considered when comparing reports using either of these methods. In Sweden, the Billing method (Billing et al. 2002) and the CF method (Hansson et al. 1988) are frequently used even though the accuracy of the Billing method for the measurement of a severe slip has been questioned (Loder 2001). An advantage of the CF method is that the CF remains in an unchanged position after remodeling and is identifiable even in adulthood (Harty 1957, Griffin 1982, Hansson et al. 1988); this provides a method for detecting SCFE after growth plate closure (Hansson et al. 1988). Variability in the radiographic technique can affect the measurement of the slip angle on the Lauenstein view (Jerre 1950, Loder 2001, Carney and Liljenquist 2005). Multiplanar computerized tomography is probably the most reproducible method to assess the slip angle in SCFE (Cohen et al. 1986, Gelberman et al. 1986, Guzzanti and Falciglia 1991, Monazzam et al. 2013) but this technique is not currently an established routine examination in Swedish hospitals for children suspected to have SCFE. Loder et al. (1999) used Lauenstein radiographs of 48 hips with SCFE (38 children), and 4 observers measured the lateral HSA. They reported no influence of observer experience, no statistically significant difference between the observers and an interobserver variability of ±12 degrees. Carney and Liljenquist (2005) used 3 observers to test the variability of the lateral HSA using Lauenstein radiographs of 108 hips (55 with SCFE and 53 normal). They reported an intraobserver variability for the HSA of ±5.9 degrees and concluded that a single observer should document at least a 12-degree change between 2 radiographs to ensure a true change. We found an inter- and intraobserver variability that was comparable with these previous results. In our study, 11 Lauenstein radiographs did not meet the technical image criteria. Other investigators have also described an inability to obtain reproducible radiographs because of variability in limb position caused by osseous deformities through the physis and/or children experiencing pain (Cohen et al. 1986). Jones et al. (2017) showed by comparing Lauenstein views with 361 simulated models from CT scans that a small

11961 Herngren D.indd 220

error in positioning could cause a greater than 10° error in the reported lateral HSA. Clear instructions for the measurement procedure can probably compensate for differences in professional experience provided that the radiographic technique is of acceptable quality. As a consequence of our study, the 4 observers together prepared an updated instruction for all Swedish hospitals on how to achieve a correct Lauenstein view: the hips should be in maximal abduction, the knees flexed to 90°, the plantar aspects of the feet placed together with the lateral aspects of the feet resting against the table, absence of significant asymmetry in the appearance of the obturator foramina (Tönnis 1976), the central beam through the most cranial part of the pubic symphysis, and with a minimum of 5 cm of the femur below the lesser trochanter included in the radiograph. Our findings indicate that, independent of the experience of the observer, the inter- and intraobserver reliability values for the methods in this study are acceptable for routine use in a national quality register for SCFE. We will consider the inclusion of the HSA as an alternative measurement method for the SPOQ register. Limitations The severity of the slips in our study was less than that previously reported in similar studies (Loder et al. 1999, Carney and Liljenquist 2005, Lehmann et al. 2013) and this may have influenced our results for both the intra- and interobserver variability. We could not blind the radiographs to personal identity numbers because of the need for secure storage of patient information. To compensate, the radiographs were all given a unique number in a list that did not follow any alphabetical order or pattern according to age, date, or sex. We also used a minimum of 6 weeks between the radiographic assessments. Fewer Billing lateral views than Lauenstein views (27 and 50, respectively) were included in this study. Acknowledgements: Bo Rolander, statistician at the Futurum Academy for Health and Care, Jönköping County Council, Sweden. Håkan Bostrom and Hanna Hebelka Bolminger, specialists in pediatric radiology at Drottning Silvias Barn- och Ungdomssjukhus, Göteborg, Sweden.

21-02-2018 15:52:50

Acta Orthopaedica 2018; 89 (2): 217–221

Study design: BH and ML; data collection: BH and ML; data analysis: BH, ML, and GH; manuscript preparation: BH, ML, and GH.

Acta thanks Anders Wensaas and other anonymous reviewers for help with peer review of this study.

Billing L, Bogren H G, Wallin J. Reliable X-ray diagnosis of slipped capital femoral epiphysis by combining the conventional and a new simplified geometrical method. Pediatric Radiol 2002; 32 (6): 423-30. Bland J M, Altman D G. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1 (8476): 307-10. Carney B T, Liljenquist J. Measurement variability of the lateral head–shaft angle in slipped capital femoral epiphysis. J Surg Orthop Adv 2005; 14 (4): 165-7. Cohen M S, Gelberman R H, Griffin P P, Kasser J R, Emans J B, Millis M B. Slipped capital femoral epiphysis: Assessment of epiphyseal displacement and angulation. J Pediatr Orthop 1986; 6 (3): 259-64. Gelberman R H, Cohen M S, Shaw B A, Kasser J R, Griffin P P, Wilkinson R H. The association of femoral retroversion with slipped capital femoral epiphysis. J Bone Joint Surg Am 1986; 68 (7): 1000-7. Griffin J B. The calcar femorale redefined. Clin Orthop Relat Res 1982; (164): 211-14. Guzzanti V, Falciglia F. Slipped capital femoral epiphysis: Comparison of a roentgenographic method and computed tomography in determining slip severity. J Pediatr Orthop 1991; 11 (1): 6-12. Hansson L I, Hägglund G, Ordeberg G, Sandstrom S. The calcar femorale as a landmark in hip physiolysis. Acta Orthop Scand 1988; 59 (2): 134-8. Harty M. The calcar femorale and the femoral neck. J Bone Joint Surg Am 1957; 39 (3): 625-30. Hermanson M, Hägglund G, Riad J, Rodby-Bousquet E. Inter- and intra-rater reliability of the head–shaft angle in children with cerebral palsy. J Child Orthop 2017; 11 (4): 256-62. Jerre T. A study in slipped upper femoral epiphysis: With special reference to the late functional and roentgenological results and to the value of closed reduction [thesis]. Acta Orthop Scand 1950; Suppl 6. Jerre R. Physiolysis of the hip: Epidemiology, diagnosis and long-term follow-up [thesis]. Institue of Surgical Sciences, Gothenburg 1995. Jones C E, Cooper A P, Doucette J, Buchan L L, Wilson D R, Mulpuri K, d’Entremont A G. Southwick angle measurements and SCFE slip severity classifications are affected by frog-lateral positioning. Skeletal Radiol 2017 Aug 24 [Epub ahead of print].

11961 Herngren D.indd 221

221

Kocher M S, Bishop J A, Weed B, Hresko M T, Millis M B, Kim Y J, Kasser J R. Delay in diagnosis of slipped capital femoral epiphysis. Pediatrics 2004; 113 (4): e322-5. Larson A N, Yu E M, Melton L J, Peterson H A, Stans A A. Incidence of slipped capital femoral epiphysis: A population-based study. J Pediatr Orthop B 2010; 19 (1): 9-12. Lehmann C L, Arons R R, Loder R T, Vitale M G. The epidemiology of slipped capital femoral epiphysis: an update. J Pediatr Orthop 2006; 26 (3): 286-90. Lehmann T G, Vetti N, Laborie L B, Engesaeter I O, Engesaeter L B, Rosendahl K. Intra- and inter-observer repeatability of radiographic measurements for previously slipped capital femoral epiphysis at skeletal maturity. Acta Radiol 2013; 54: 587-91. Loder R T. The demographics of slipped capital femoral epiphysis: An international multicenter study. Clin Orthop Relat Res 1996; (322): 8-27. Loder R T. Effect of femur position on the angular measurement of slipped capital femoral epiphysis. J Pediatr Orthop 2001; 21 (4): 488-94. Loder R T. Slipped capital femoral epiphysis: A spectrum of surgical care and changes over time. J Child Orthop 2017; 11 (2): 154-9. Loder R T, Blakemore L C, Fairley F A, Laidlaw A T. Measurement variability of slipped capital femoral epiphysis. Journal of Orthopaedic Surgery 1999; 7 (1): 33-42. McGraw K O, Wong S P. Forming inferences about some intraclass correlation coefficients. Psychological Methods 1996; 1 (1): 30-46. Monazzam S, Dwek J R, Hosalkar H S. Multiplanar CT assessment of femoral head displacement in slipped capital femoral epiphysis. Pediatric Radiol 2013; 43 (12): 1599-605. Sedgwick P. Limits of agreement (Bland–Altman method). BMJ 2013; 346: f1630. Souder C D, Bomar J D, Wenger D R. The role of capital realignment versus in situ stabilization for the treatment of slipped capital femoral epiphysis. J Pediatr Orthop 2014; 34 (8): 791-8. Southwick W O. Osteotomy through the lesser trochanter for slipped capital femoral epiphysis. J Bone Joint Surg Am 1967; 49 (5): 807-35. Terjesen T, Wensaas A. Prognostic factors for long-term outcome of chronic slipped capital femoral epiphysis treated with fixation in situ. J Child Orthop 2017; 11 (2): 114-9. Tönnis D. Normal values of the hip joint for the evaluation of X-rays in children and adults. Clin Orthop Relat Res 1976; (119): 39-47.

21-02-2018 15:52:50

222

Acta Orthopaedica 2018; 89 (2): 222–228

Early osteoarthritis after slipped capital femoral epiphysis Cartilage degeneration, residual deformity and patient-reported outcome in 25 patients Lukas HELGESSON 1, Peter Kälebo JOHANSSON 2, Ylva AURELL 2, Carl-Johan TIDERIUS 3, Johan KÄRRHOLM 4, and Jacques RIAD 1

1 Department of Orthopaedics, Skaraborgs Hospital, Skövde, 2 Department of Radiology, Mölndal Hospital, Sahlgrenska, 3 Department of Orthopaedics, Lund University Hospital, 4 Department of Orthopaedics, Mölndal Hospital, Sahlgrenska, Sweden Correspondence: Jacques.riad@vgregion.se Submitted 2017-07-23. Accepted 2017-10-13.

Background and purpose — Slipped capital femoral epiphysis (SCFE) results in a more or less pronounced deformity of the proximal femur, sometimes causing impingement and early osteoarthritis. We studied early osteoarthritis after SCFE and the association with deformity and self-reported hip function, pain, and quality of life. Patients and methods — 9 women and 16 men, mean age 32 (21–50) years, 19 with unilateral and 6 with bilateral SCFE, participated. All patients had primarily been operated by pin or screw with no attempt at reposition of the slip. Hips were examined by delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC), which quantifies and locates cartilage degeneration. Plain radiographs were used to measure deformity as determined by the alpha angle. Outcome was assessed by Oxford hip score, Hip Groin Outcome score and EQ-5D-Visual scale. Results — In the 19 unilateral SCFE, on the slip side dGEMRIC mean value was 533 ms (SD 112, range 357–649) versus mean 589 ms (SD 125, range 320–788) on the non-slip side, (p = 0.01). The dGEMRIC correlated negatively to the alpha angle, correlation coefficient (CC) = –0.60, (p = 0.002). Oxford hip score, pain, and EQ-5D-Visual scale correlated to dGEMRIC CC = 0.43 (p = 0.03), CC = 0.40 (p = 0.05), and CC = 0.49 (p = 0.01) respectively. Interpretation — After SCFE, even relatively mild residual hip deformity can be associated with cartilage degeneration. A high alpha angle was associated with worse cartilage status. The Oxford hip score identified symptoms even though our patients had not previously sought medical care after the index operation. Quality of life showed strong inverse correlation with cartilage degeneration. Objective assessment of early cartilage degeneration may be useful for treatment decisions and follow-up. ■

Slipped capital femoral epiphysis (SCFE) results in a more or less pronounced deformity of the hip joint. Severe slips presumably pose a higher risk of early osteoarthritis (OA) (Hagglund et al. 1988, Loder et al. 2006, Rahme et al. 2006) The goal when treating SCFE is to prevent further slip by stabilizing the epiphysis. This can be done with a smooth pin with a hook device or a short threaded screw to allow further growth and remodeling, or with a threaded screw with compression to close the physis (Hansson 1982, Aronsson et al. 2006). The long-term goal is to achieve congruent hip joints that function painlessly at a high level without subsequent OA. For severe deformities, secondary surgical treatment has become more common (Azegami et al. 2013). Open surgery with dislocation of the hip, or arthroscopic surgery, can be performed to remove any bony prominences that cause impingement. Additional surgical options aim to re-orient the femoral head to avoid impingement and provide better congruity of the hip joint. Some of these procedures are extensive and entail high complication risks (Azegami et al. 2013). Sensitive diagnostic tools are needed to identify early cartilage degeneration after SCFE, both in the natural course of the disease and as indication for and evaluation of surgical treatment. Delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC) has been used since the late 1990s to study the glycosaminoglycan (GAG) content of articular cartilage (Bittersohl et al. 2015, Zilkens et al. 2015). Low cartilage GAG content results in a low T1 signal within the cartilage, (referred to as the dGEMRIC value), indicating degeneration (Bashir et al. 1996, van Tiel et al. 2016). Several studies of the hip and knee have reported early signs of cartilage degeneration before radiographic changes occur (Tiderius et al. 2007, Zilkens et al. 2011b). In addition, dGEMRIC has shown a correlation with clinically relevant parameters, such as pain and function (Kim et al. 2003).

11842 Riad D.indd 222

20-02-2018 12:19:15

Acta Orthopaedica 2018; 89 (2): 222–228

Individuals with SCFE age 20–50 identified in the region n = 288 Excluded No peroperative radiographs n = 181 Had peroperative radiographs n = 107 Excluded (n = 58): – declined to participate, 3 – did not respond to invitation, 55 Accepted participation n = 49 Excluded (n = 24): – remaining osteosynthesis, 19 – other secondary hip surgery, 3 – did not attend examination, 2 Participants, all with unilateral slip at presentation n = 25 Operated on slip side only (n = 13): – pin fixation, 10 – screw fixation, 3 Operated on both sides (n = 6): – pin fixation, 5 – screw fixation, 1 Later slip on contralateral side (n = 6): – pin fixation, 3 – screw fixation, 3

Figure 1. Recruitment and treatment of the 25 subjects with SCFE.

It appears that the deformity arising from SCFE influences the risk of subsequent OA; however, few studies have addressed the subject (Castaneda et al. 2013). We wanted to use dGEMRIC to detect signs of early OA, since the symptoms of early cartilage degeneration are vague and conventional radiographs do not identify early cartilage changes. We studied the development of early osteoarthritis after SCFE and possible associations between deformity (degree of slip) and self-reported hip function, pain, and quality of life. Our hypotheses were: 1, After SCFE there is development of early osteoarthritis associated with the degree of remaining deformity. 2, Self-reported hip function, pain, and quality of life are associated with early signs of osteoarthritis after SCFE. We hoped to identify prognostic variables that would enable development of rational treatment plans to improve the longterm outcome after SCFE.

11842 Riad D.indd 223

223

Patients and methods Participants Individuals in Västra Götalandregionen previously diagnosed with and treated surgically for SCFE were identified through medical records and Sweden’s National Patient Register. Inclusion criteria were SCFE, skeletal maturity, peroperative radiographs available, and no remaining pin or screw fixation. Exclusion criteria were other injuries or diseases affecting the hip or lower extremity, and subsequent surgical procedure other than pin removal. Those between 20 and 50 years of age received a letter with information about the study and an invitation to participate. From the medical records, data on sex, age, symptoms, side of involvement, and radiographic confirmation of a slip were recorded. 25 subjects, mean age 32 (21–50) years (9 women) participated (Figure 1). Follow-up time from initial diagnosis was mean 19 (11–35) years. At initial presentation in childhood all were unilateral SCFE and were treated with Hansson pin or short threaded screw fixation. Prophylactic fixation was performed in 6 patients at the initial surgery on the non-slip side. 6 others later developed a contralateral slip, which was operatively treated (Figure 1). In summary, there were 19 unilateral and 6 bilateral SCFE. Cartilage quality assessment dGEMRIC, a magnetic resonance imaging technique with contrast enhancement, was performed on a 1.5 Tesla MRI system (Siemens Aera, Erlangen, Germany) with a body array coil (Body 18). Participants were given gadolinium contrast i.v. (Magnevist 0.5 mmol/mL, Bayer Pharmaceuticals, Berlin, Germany) at a dose of 0.4 mL per kilogram body weight. After the injection, the participants walked for 15 minutes, then waited seated for another 45 minutes. During this time they answered the questionnaires and had conventional hip radiographs. MRI imaging was thus performed 1 hour after contrast injections based on previous results of contrast agent diffusion into hip joint cartilage (Tiderius et al. 2007). The dGEMRIC protocol included a post-contrast 3D T1-weighted gradient echo sequence with 2 flip-angles and a built-in B1 correction (3D VFA). A coronal projection including both hips with feet in neutral position was performed with the following parameters: TR 15.00 milliseconds (ms), TE 4.77 ms, flip angles 5º and 26º, FoV 320x320 mm, matrix 352x352, slice thickness 0.9 mm, 1 excitation, bandwidth 140Hz/Px, and scan time 9.35 min. MapIt software (Siemens, Erlangen, Germany) was used to calculate T1 values expressed in ms in cartilage regions of interest (ROI) on a workstation with Syngovia software (Syngovia Version B10B, Siemens, Erlangen, Germany). 3 ROIs for each hip were drawn manually on the coronal images, centrally, dorsally, and ventrally spaced 3.6 mm apart (Figure 2). The most cranial part of the cartilage on both the acetabular and the femoral side was included in the ROI, and the lateral border was the basis of the labrum and the

20-02-2018 12:19:15

224

Acta Orthopaedica 2018; 89 (2): 222–228

Figure 2. T1 region of interest (ROI) drawn on the initial, T1-gradient echo image, including the central cartilage.

Figure 3. dGEMRIC region of interest (ROI). The same ROI as in Fgure 2, with the corresponding dGEMRIC image.

medial border was the central fovea of the acetabulum. The ROI was drawn on the initial T1-gradient echo image that had the best anatomical delineation and automatically copied to the corresponding color-coded dGEMRIC image including an ROI of the cartilage (Figure 3). The dGEMRIC T1 value was given as mean (1 SD) by the Syngovia software for each area assessed. Remaining metal implants from previous operations was an exclusion criterion in this study to avoid unpredictable influence from artefacts. All magnetic resonance images were assessed by consensus readings by 2 radiologists (YA and PKJ) blinded to all other data. 10 randomly selected cases were assessed a second time 3 months later, also in consensus readings, to evaluate the reliability of the measurements.

cally active young to middle-aged patients, has also been used to study femoral-acetabular impingement (Thorborg et al. 2011). HAGOS covers 6 dimensions: symptoms (e.g., clicking, paresthesia, difficulties taking a long step), pain, function in daily living, sports and recreation, physical activities, and hip-related effects on quality of life. The score ranges from worst to best, 0–100. We also used the EQ-5D Visual Analogue Scale (EQ5D-VAS), a standardized measure of health, which records self-rated health on a 20-cm vertical scale with endpoints labelled “the best health you can imagine” and “the worst health you can imagine” (EuroQol 1990).

Radiographs to assess deformity Deformity of the proximal femur was assessed with conventional radiographs obtained at the same visit as the dGEMRIC measurement, using the lateral projection according to Dunn (Meyer et al. 2006). The alpha angle describes the sphericity of the femoral head and reveals possible impingement (Notzli et al. 2002). Alpha angles were classed thus: normal sphericity (< 50 degrees), mild deformity (50–60 degrees), and severe deformity (> 60 degrees) (Zilkens et al. 2011a). The head shaft angle (HSA) according to Southwick quantifies the posterior slip and was used to subgroup the 50 hips into: mild slip (HSA < 30 degrees), moderate (HSA 30–60 degrees), and severe (has > 60 degrees) (Southwick 1967). Patient-reported hip function, pain, and quality of life Hip function was measured with the widely used, diseasespecific, validated Oxford Hip Score (OHS), which includes 12 questions graded from 1 to 5 (Dawson et al. 1996). The Copenhagen hip and groin outcome score (HAGOS), developed to assess long-standing hip and/or groin pain in physi-

11842 Riad D.indd 224

Statistics In the 6 patients with bilateral SCFE, we selected the more severely deformed hip (higher alpha angle), for the comparative analysis. The dGEMRIC values and alpha angles were normally distributed and a paired sample t-test was used. Non-parametric statistical calculations were used for the remaining analysis since the clinical outcome measures were scale data. Correlations were analyzed using the Spearman rank correlation. All tests were 2-tailed and statistical significance was set at p < 0.05. The data were analyzed using SPSS Statistics version 22 (IBM Corp, Armonk, NY, USA). Ethics, funding, and potential conflicts of interest The study was approved by the Human Research Ethics Committee of the Medical Faculty at the University of Gothenburg, Sweden (Dnr 904-13), in accordance with the ethical standards of the Helsinki Declaration. Funding was received from the state hospital research unit and orthopaedic department. The authors declare no conflicts of interest.

20-02-2018 12:19:16

Acta Orthopaedica 2018; 89 (2): 222–228

225

Table 1. Degree of deformity of the proximal femur at follow-up with measurements of alpha angle and head shaft angle

dGEMRIC value 800

Slip hips n = 31

700

Alpha angle: < 50° (normal) 50–60° (mild) > 60° (severe) Head shaft angle: Mild Moderate Severe

600

500

400

Non-slip hips n = 19

14 5 12

15 2 2

27 4 –

19 – –

300 20

Alpha-angle

Table 2. Results from the 3 questionnaires, Oxford Hip Score, HAGOS, and the EQ-5D Visual Scale, and correlation with the dGEMRIC values for the 25 subjects

Figure 4. Scatter plot of alpha angle and dGEMRIC value (ms). The mean dGEMRIC value (ms) of the three regions—ventral, central, and dorsal—on the slip side in 19 subjects with unilateral SCFE and the most deformed side in 6 subjects with bilateral SCFE. n = 25.

Results Cartilage quality assessment The reliability testing of the dGEMRIC measurements revealed excellent interclass correlation coefficients for the 3 regions of interest measured by the 2 observers (YA and PKJ). In the ventral region ICC was 0.98 (95% CI 0.95–0.99), in the central region ICC was 0.96 (CI 0.90–0.98), and in the dorsal region ICC was 0.99 (CI 0.97–0.99). As there were no statistically significant differences in dGEMRIC values for the ventral, central, and dorsal regions of the hip joints, we calculated the mean of the 3 regions for each hip. In the compiled group of 25 hips (19 unilateral slips and the 6 most severe slips in the bilateral group) the mean dGEMRIC value was 532 ms (SD 114, range 347–728). When analyzing the 19 patients with unilateral SCFE we found a significant difference (p = 0.01) in dGEMRIC value comparing the slip side, mean 533 ms (SD 112, range 357–649) with the non-slip side, mean 589 ms (SD 125, range 320–788). For the dGEMRIC correlated (Spearman rank correlation) to the alpha angle, correlation coefficient (CC) = –0.60 (p = 0.002), meaning that the higher the alpha angle or the worse the deformity, the lower was the dGEMRIC value, equivalent to cartilage injury/degeneration (Figure 4). Radiographs to assess deformity The degree of deformity of the proximal femur was assessed at follow-up on plain radiographs on both sides, corresponding to 50 hips (31 hips with SCFE and 19 judged as normal). The deformity was relatively mild, with a mean alpha angle of

11842 Riad D.indd 225

Descriptive statistics median range IOR Oxford Hip Score HAGOS pain HAGOS symptom HAGOS ADL HAGOS Sport/Rec HAGOS PA HAGOS QOL EQ-5D VAS

46 92 88 100 94 88 85 90

36–48 50–100 54–100 50–100 50–100 50–100 40–100 50–100

43–48 84–100 73–100 82–100 78–100 75–100 65–100 70–97

Correlation with dGEMRIC values coefficient p-value 0.43 0.40 0.30 0.36 0.34 0.36 0.38 0.49

0.03 0.05 0.1 0.08 0.09 0.08 0.06 0.01

IOR: 25th to 75th percentile

54 (24–83) degrees on the slip side (31 hips) (Table 1). For the alpha angle correlated (Spearman rank correlation) with the HSA, CC = 0.50 (p < 0.001). In the 19 patients with unilateral SCFE, we found a mean alpha angle of 54 (30–79) degrees on the slip side and 45 (37– 62) on the non-slip side (p = 0.01). The difference in HSA, 16 versus 10 degrees, was not significant (p = 0.09). Patient-reported hip function, pain, and quality of life There was an association between both hip function and overall quality of life and cartilage degeneration. Oxford hip score, pain, and EQ-5D-VAS correlated with dGEMRIC CC = 0.43 (p = 0.03), CC = 0.40 (p = 0.05), and CC = 0.49 (p = 0.01), respectively (Table 2).

Discussion We found cartilage degeneration on the slip side after SCFE even in cases with relatively mild deformity. A high alpha angle, i.e., femoral-acetabular impingement, was strongly associated with signs of cartilage degeneration Patientreported hip function and quality of life were inversely associated with signs of early osteoarthritis.

20-02-2018 12:19:19

226

Cartilage quality assessment The quality of the cartilage did not differ between the ventral, central, and dorsal region in the individual hips, which was surprising since we had expected cartilage changes as a consequence of femoral-acetabular impingement anteriorly. Both the dGEMRIC values and the alpha angles on the slip side revealed less pronounced changes in our patients (mean age 32 years) than in a previous study, where the mean age was 23 years (Zilkens et al. 2011a). In a group of patients with femoral-acetabular impingement who underwent hip arthroscopy (Chandrasekaran et al. 2015), the mean preoperative dGEMRIC value was 426 ms, as opposed to 532 ms in our study group. Kim et al. (2003) used dGEMRIC to study 37-year-olds with non-dysplastic and non-symptomatic hips, and considered values of 570 (SD 90) ms as normal. Any value more than 2 SD lower (< 390 ms) was defined as OA (Kim et al. 2003). Their study aimed to predict outcome after pelvic osteotomy for hip dysplasia, and they found a steeply increased risk for failure in patients with values below 390 ms. Cunningham et al. (2006) arrived at a similar cut-off of 370 (SD 88) ms predicting failure and 498 ± 105 ms for satisfactory results (Cunningham et al. 2006). In our study, 5 of the total 31 slip hips had a dGEMRIC value below 390 ms. Even though we found a statistically significant difference between the slip and non-slip side in patients with unilateral slip, dGEMRIC values displayed high inter-individual variation between sides. This could reflect the different degrees of slip between patients, but possibly also differences in physical activity, which has been reported to influence dGEMRIC values (Roos and Dahlberg 2005). By scanning both hips simultaneously, we eliminated an important methodological issue, i.e., the risk of ongoing contrast medium diffusion in between scanning of separate hips (Tiderius 2007). Considering the variability within subjects it is possible that the dGEMRIC method might be most useful to follow progression within the same patient rather than between patients. We found a relatively high correlation between deformity as assessed from the alpha angle, and dGEMRIC measurements. Both Carney et al. (1991) and Jerre (1950) stated that the remaining deformity, referring to the degree of slip measured with the HSA, predisposes to early OA. The exact mechanism underlying OA development in the hip joint is not known; OA arising through a femoral-acetabular impingement mechanism might have another course than that acting in patients with a more pronounced deformity directly related to the degree of slip. Are the slip deformity and the alpha-angle deformity 2 different assessments describing different pathologies? Which pathology develops into a “whole joint disease”? Radiographs to assess deformity We noted mean alpha angle of 54 degrees on the slip side, whereas Zilkens et al. (2011a) reported this to be 65, which implies that our study group had relatively mild deformities.

11842 Riad D.indd 226

Acta Orthopaedica 2018; 89 (2): 222–228

Even so, we observed a clear difference in alpha angle comparing the slip and non-slip side in the 19 patients with unilateral SCFE. We used the lateral head shaft angle according to Southwick to assess the degree of slip; however, a previous reliability study revealed that a change of 12 degrees was necessary before observer variability could be ruled out (Carney and Liljenquist 2005). In addition, the reproducibility of the Lauenstein projection was low, which became clear when we assessed the primary radiographs at the time of presentation in childhood, and we observed marked variations in the degree of flexion, abduction, and rotation. Therefore, we could not rely on comparisons between the initial deformity and that at long-term follow-up. Regarding the relatively mild deformity at this long-term follow-up, it should be noted that all hips had deliberately been treated with smooth pins or short threaded screws to allow for further growth. Both longitudinal growth and remodeling can be expected to have occurred (Holmdahl et al. 2016, Ortegren et al. 2016). Patient-reported hip function, pain, and quality of life We found a correlation between cartilage quality and hip-specific reported outcome including function, OHS, and HAGOS, as well as a correlation with the EQ-5D-VAS, reflecting quality of life. Even comparatively small differences in alpha angle and dGEMRIC value influenced the patient-reported outcome after SCFE, a finding we have not seen so clearly illustrated before. Previous studies have mainly been performed on patients with ongoing symptoms seeking medical care and the threshold for when to expect symptoms related to morphological changes remains unknown (Chandrasekaran et al. 2015, Sansone et al. 2016). Our study group of relatively young patients showed subtle changes in dGEMRIC values and mild deformity after SCFE. We do not know if these patients’ expectations regarding what is normal in terms of activity and ability to participate in sports differ from those of older populations with confirmed OA or the young elite athletes on whom these patient-reported outcome measures have been used before (Sansone et al. 2016). Sansone et al. (2016) studied 75 patients with femoral-acetabular impingement 2 years after arthroscopic treatment, and reported improvement according to the HAGOS questionnaire and also in the EQ-5D. Chandrasekaran et al. (2015) studied patients undergoing hip arthroscopy, 1 group with dGEMRIC value below and one above 323, i.e., 1 SD below the cohort mean of 426 ms. While both groups had significantly improved Harris Hip Scores 2 years after arthroscopy, the group with dGEMRIC value above the cut-off 323 ms also showed improvement in the Hip Outcome Score Activities for ADL, Sport-specific Subscale, and the visual analogue scale for pain (Chandrasekaran et al. 2015). In our study, no participant had a dGEMRIC value under 323 ms. The lowest value was 347 ms. However, longer

20-02-2018 12:19:23

Acta Orthopaedica 2018; 89 (2): 222–228

follow-up might reveal progression of OA and clinical manifestations (Zilkens et al. 2011a). The correlation between the dGEMRIC and EQ-5D results was quite unexpected. We can only speculate that these patients, who had not previously sought contact with the medical care system for symptoms, nonetheless reported lower quality of life due to problems related to their previous SCFE. It is worth noting that the EQ-5D-VAS was the last questionnaire the participants completed, which might have motivated them to report some adverse effect of their SCFE. We found no previous publication testing HAGOS or OHS in SCFE, but in a study that used HAGOS on patients with hip dysplasia, the patients scored lower than controls on every dimension (Jacobsen et al. 2013). According to Impellizzeri et al. (2015) OHS can also be used to assess femoral-acetabular impingement. Limitations Physical activity has been suggested to influence dGEMRIC measurements, possibly owing to cartilage adapting to increased mechanical demands (Roos and Dahlberg 2005). We had no data on the physical activity of our participants, which is a limitation. The number of participants is relatively small, and we lacked information on the initial degree of slip. More comprehensive assessment of the orientation of the impingement/alpha angle, with MRI or CT scanning, could have been advantageous. Summary With even relatively mild residual deformity, a high alpha angle after SCFE can be associated with cartilage degeneration. Symptoms that are too small to prompt patients to contact the health care facilities can nonetheless be identified by the Oxford Hip Score. Quality of life appears strongly and inversely associated with cartilage degeneration. Objective assessment of early cartilage degeneration may be useful in treatment and follow-up after SCFE.

All authors contributed with the planning, methodology, protocol, data analyses and writing of the manuscript. LH and JR obtained the ethical approval, recruited participants, organized the visit to the radiology department, and collected the data and performed the statistical analysis, as well. YA and PKJ in addition assessed the dGEMRIC images and performed the reliability test.

Acta thanks Stine Hangaar and Jasper van Tiel for help with peer review of this study. Aronsson D D, Loder R T, Breur G J, Weinstein S L. Slipped capital femoral epiphysis: current concepts. J Am Acad Orthop Surg 2006; 14 (12): 666-79. Azegami S, Kosuge D, Ramachandran M. Surgical treatment of femoroacetabular impingement in patients with slipped capital femoral epiphysis: A review of current surgical techniques. Bone Joint J 2013; 95-B (4): 445-51.

11842 Riad D.indd 227

227

Bashir A, Gray M L, Burstein D. Gd-DTPA2 as a measure of cartilage degradation. Magn Reson Med 1996; 36 (5): 665-73. Bittersohl B, Hosalkar H S, Hesper T, Tiderius C J, Zilkens C, Krauspe R. Advanced imaging in femoroacetabular impingement: Current state and future prospects. Front Surg 2015; 2: 34. Carney B T, Liljenquist J. Measurement variability of the lateral head-shaft angle in slipped capital femoral epiphysis. J Surg Orthop Adv 2005; 14 (4): 165-7. Carney B T, Weinstein S L, Noble J. Long-term follow-up of slipped capital femoral epiphysis. J Bone Joint Surg Am 1991; 73 (5): 667-74. Castaneda P, Ponce C, Villareal G, Vidal C. The natural history of osteoarthritis after a slipped capital femoral epiphysis/the pistol grip deformity. J Pediatr Orthop 2013; 33 (Suppl 1): S76-82. Chandrasekaran S, Vemula S P, Lindner D, Lodhia P, Suarez-Ahedo C, Domb B G. Preoperative delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC) for patients undergoing hip arthroscopy: Indices are predictive of magnitude of improvement in two-year patientreported outcomes. J Bone Joint Surg Am 2015; 97 (16): 1305-15. Cunningham T, Jessel R, Zurakowski D, Millis M B, Kim Y J. Delayed gadolinium-enhanced magnetic resonance imaging of cartilage to predict early failure of Bernese periacetabular osteotomy for hip dysplasia. J Bone Joint Surg Am 2006; 88 (7): 1540-8. Dawson J, Fitzpatrick R, Carr A, Murray D. Questionnaire on the perceptions of patients about total hip replacement. J Bone Joint Surg Br 1996; 78 (2): 185-90. EuroQol. EuroQol – a new facility for the measurement of health-related quality of life. The EuroQol Group. Health Policy 1990; 16 (3): 199-208. Hagglund G, Hansson L I, Ordeberg G, Sandstrom S. Bilaterality in slipped upper femoral epiphysis. J Bone Joint Surg Br 1988; 70 (2): 179-81. Hansson L I. Osteosynthesis with the hook-pin in slipped capital femoral epiphysis. Acta Orthop Scand 1982; 53 (1): 87-96. Holmdahl P, Backteman T, Danielsson A, Karrholm J, Riad J. Continued growth after fixation of slipped capital femoral epiphysis. J Child Orthop 2016; 10 (6): 643-50. Impellizzeri F M, Mannion A F, Naal F D, Leunig M. Validity, reproducibility, and responsiveness of the Oxford Hip Score in patients undergoing surgery for femoroacetabular impingement. Arthroscopy 2015; 31 (1): 42-50. Jacobsen J S, Nielsen D B, Sorensen H, Soballe K, Mechlenburg I. Changes in walking and running in patients with hip dysplasia. Acta Orthop 2013; 84 (3): 265-70. Jerre T. A study in slipped capital femoral epiphysis: With special reference to late functional and roentgenological results and to value closed reduction. Acta Orthop Scand 1950; 6 (Suppl): 5-155. Kim Y J, Jaramillo D, Millis M B, Gray M L, Burstein D. Assessment of early osteoarthritis in hip dysplasia with delayed gadolinium-enhanced magnetic resonance imaging of cartilage. J Bone Joint Surg Am 2003; 85-A (10): 1987-92. Loder R T, Starnes T, Dikos G, Aronsson D D. Demographic predictors of severity of stable slipped capital femoral epiphyses. J Bone Joint Surg Am 2006; 88 (1): 97-105. Meyer D C, Beck M, Ellis T, Ganz R, Leunig M. Comparison of six radiographic projections to assess femoral head/neck asphericity. Clin Orthop Relat Res 2006; 445:181-5. Notzli H P, Wyss T F, Stoecklin C H, Schmid M R, Treiber K, Hodler J. The contour of the femoral head–neck junction as a predictor for the risk of anterior impingement. J Bone Joint Surg Br 2002; 84 (4): 556-60. Ortegren J, Bjorklund-Sand L, Engbom M, Tiderius C J. Continued growth of the femoral neck leads to improved remodeling after in situ fixation of slipped capital femoral epiphysis. J Pediatr Orthop 2016. DOI: 10.1097/ BPO.0000000000000797 Rahme D, Comley A, Foster B, Cundy P. Consequences of diagnostic delays in slipped capital femoral epiphysis. J Pediatr Orthop B 2006; 15 (2): 93-7.

20-02-2018 12:19:23

228

Roos E M, Dahlberg L. Positive effects of moderate exercise on glycosaminoglycan content in knee cartilage: A four-month, randomized, controlled trial in patients at risk of osteoarthritis. Arthritis Rheum 2005; 52 (11): 3507-14. Sansone M, Ahlden M, Jonasson P, Thomee C, Sward L, Collin D, Baranto A, Karlsson J, Thomee R. Outcome of hip arthroscopy in patients with mild to moderate osteoarthritis: A prospective study. J Hip Preserv Surg 2016; 3 (1): 61-7. Southwick W. Osteotomy through the lesser trochanter for slipped capital femoral epiphysis. J Bone Joint Surg Am 1967; 49-A (5): 807-35. Thorborg K, Holmich 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. Tiderius C J, Jessel R, Kim Y J, Burstein D. Hip dGEMRIC in asymptomatic volunteers and patients with early osteoarthritis: The influence of timing after contrast injection. Magn Reson Med 2007; 57 (4): 803-5.

11842 Riad D.indd 228

Acta Orthopaedica 2018; 89 (2): 222â&#x20AC;&#x201C;228

van Tiel J, Kotek G, Reijman M, Bos P K, Bron E E, Klein S, Nasserinejad K, van Osch G J, Verhaar J A, Krestin G P, Weinans H, Oei E H. Is T1rho mapping an alternative to delayed gadolinium-enhanced MR imaging of cartilage in the assessment of sulphated glycosaminoglycan content in human osteoarthritic knees? An in vivo validation study. Radiology 2016; 279 (2): 523-31. Zilkens C, Bittersohl B, Jager M, Miese F, Schultz J, Kircher J, Westhoff B, Krauspe R. Significance of clinical and radiographic findings in young adults after slipped capital femoral epiphysis. Int Orthop 2011a; 35 (9): 1295-301. Zilkens C, Miese F, Bittersohl B, Jager M, Schultz J, Holstein A, Kim Y J, Millis M B, Mamisch T C, Krauspe R. Delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC), after slipped capital femoral epiphysis. Eur J Radiol 2011b; 79 (3): 400-6. Zilkens C, Tiderius C J, Krauspe R, Bittersohl B. Current knowledge and importance of dGEMRIC techniques in diagnosis of hip joint diseases. Skeletal Radiol 2015; 44 (8): 1073-83.

20-02-2018 12:19:23

Acta Orthopaedica 2018; 89 (2): 229–233

229

Spinal metastasis with neurologic deficits Outcome of late surgery in patients primarily deemed not suitable for surgery Panagiotis TSAGOZIS 1,2 and Henrik C F BAUER 1,2

1 Section of Orthopedics, Department of Molecular Medicine and Surgery, Karolinska Institute; 2 Department of Orthopedics, Karolinska University Hospital, Stockholm, Sweden Correspondence: panagiotis.tsagkozis@sll.se Submitted 2017-06-09. Accepted 2017-11-06.

Background and purpose — A significant number of patients with spinal metastases are treated non-surgically, but may need surgical treatment at a later stage due to progression of symptoms. Therefore, we investigated the need for late surgical decompression in patients with spinal metastasis who were initially deemed as non-surgical candidates, as well as the outcome of late surgery. Patients and methods — 116 patients who were referred to the orthopedic oncology department between 2002 and 2011 due to spinal metastasis with neurologic symptoms were deemed to be non-surgical candidates. The primary reason was minor neurologic deficits in 40 patients (M) and short survival (S) in 76 patients. Results — 8 patients underwent a late operation due to progression of the neurologic symptoms, all of them belonged to group M. M-patients with a modified Bauer score of less than 2 had both an inferior survival as well as a higher risk for late surgery. Postoperative improvement in neurologic function was noted in 5/8 operated patients, whilst 2 patients had stationary symptoms and 1 deteriorated. Interpretation — The need for late surgery arises in a minority of patients with spinal metastasis primarily treated non-surgically, and only in patients with minor neurologic compromise rather than poor general condition. An established prognostic score (modified Bauer) can be used to guide decision-making. Late surgical decompression is effective in restoring the neurologic status

both retrospective and prospective studies have demonstrated a superior outcome of combined surgical decompression and radiotherapy as compared with radiotherapy only (Sundaresan et al. 1995, Harrington 1997, Patchell et al. 2005, Falicov et al. 2006, Ibrahim et al. 2008, Quan et al. 2011). However, surgical treatment is not offered to all patients. Surgery is generally not justified in patients with short expected survival, although it is hard to define who is terminally ill. Some may even overcome a period of poor condition due to optimal supportive care. Some SM patients have a favorable oncological prognosis due to modern treatment but present with minor neurologic deficits. Thus, there may be a progression of the neurologic symptoms of SM patients initially regarded as non-candidates for surgery. It is unknown how many of these patients need to undergo late surgery, if there are any factors associated with a higher risk for late surgery, and whether the outcome of this intervention is comparable to that of primary decompression. We investigated the incidence and outcome of late surgical decompression in 116 consecutive patients who were referred to our tertiary center with a query regarding surgical intervention due to SM with neurologic deficits, but were initially not operated on.

■

We reviewed the prospectively collected database of our department and identified 360 patients between 2002 and 2011 who were referred with a query regarding their suitability for surgery because of SM with neurologic deficits. Of these, 116 (81 male) were deemed as inappropriate for surgery. Our primary indication for surgery is significant neurologic deficits (Jansson and Bauer 2006), but factors such as the condition of the patient, expected survival, and the number of lumbar levels

Radiotherapy has been the standard treatment for decades in cases of neurologic deficits due to spinal metastasis (SM) as early studies did not show any benefit of surgery (Young et al. 1980, Dea and Fisher 2014). There has been a shift in the management of SM in the past 20 years, since data from

Patients and methods

11793 Tsagozis D.indd 229

21-02-2018 15:53:47

230

engaged by the tumor, as well as the expected morbidity of surgical decompression, are also taken into account. Decision was individualized and taken by 1 or more senior grade consultants with informed consent of the patient, relying on clinical judgment rather than the use of a formal scoring system with exact predetermined criteria. The decision not to operate was mainly based on the absence of significant neurologic compromise, i.e. Frankel D or near normal neurology, despite a good prognosis in 40 (34%) patients (minor neurology, group M). Short expected survival and poor overall condition, multi-level tumor, or a combination of the above was noted in 76 (66%) patients (short-survival group S). Mean age at referral was 68 (28–89) years and mean follow-up was 15 months. At last follow-up 3 patients were still alive. The majority of patients in the S-group (n = 34) had prostate cancer, whereas 12 had lung cancer, 7 breast cancer, 5 renal cancer, and 18 had other malignancies. In the M-group, 15 patients had prostate cancer, 6 myeloma, 4 breast cancer, 3 lung cancer, and 2 had other malignancies. 94 patients presented with compression of the thoracic spine, 19 of the lumbar, and 3 of the cervical spine. Imaging was by MRI or CT. A pathological fracture was documented in 21 patients. 9% of the patients had a Frankel A score at admittance, 9% Frankel B, 26% Frankel C, 35% Frankel D, and 22% Frankel E. Of the patients belonging to the M-group, 20 had Frankel-D status at diagnosis and 20 only minor deficits, which were best classified to the Frankel E grade. Of patients in the S-group, 10 had Frankel A status at diagnosis, 10 Frankel B, 28 Frankel C, 22 Frankel D and 6 Frankel E. All but 5 patients had had postoperative radiotherapy (3 because of poor general condition and 2 because adjuvant treatment relied on systemic chemotherapy). Postoperative radiotherapy was given after wound healing (3–6 weeks). In half of the patients the given dose was 20 Gy (4 Gy x 5), in 20% 16 Gy (8 Gy x 2), in 18% the dose exceeded 20 Gy (the most common being 30 Gy given in 10 fractions) and in 12% of cases a single dose of 8 Gy was given. Operated patients were allowed to mobilize without any restrictions. Routine antibiotic and antithrombotic prophylaxis were given. Neurologic deficits were evaluated according to the Frankel scale. To evaluate the mean neurologic outcome, patients were given an arithmetic score (Frankel A arbitrarily corresponding to 1 and E to 5), and the result after treatment was calculated by subtracting the first score from the last. The neurologic outcome was assessed at 2 time-points, the first at a median of 3 (1–70) weeks after referral, and the second at a median of 17 (1–636) weeks. Similar neurologic outcome was noted between the first and the second time-point. The Bauer score (Bauer and Wedin 1995), as modified by Leithner, was used in survival analysis (Leithner et al. 2008). Statistics Statistics analysis was carried out using SPSS software (version 20; SPSS Inc, Chicago, IL, USA) and Stata (version 13;

11793 Tsagozis D.indd 230

Acta Orthopaedica 2018; 89 (2): 229–233

Figure 1. Kaplan–Meier curve of overall survival of 116 patients initially treated non-surgically due to spinal metastasis. Main indications for conservative treatment were either absence of significant neurologic compromise while oncological prognosis was good (M-group), or poor prognosis and overall condition, widespread disease, or a combination of the above (S-group). Median overall survival of patients in the S-group was 3 months (CI 2–5), whereas median overall survival of patients in the M-group was 15 months (CI 6–24), p < 0.001.

StataCorp, College Station. TX, USA). Categorical variables were studied using the chi-square (χ2) test. The Kaplan–Meier method was used for survival analyses and comparisons were done using the log-rank test. End-points were death and the incidence of secondary surgical decompression. Since the Kaplan–Meier method is based on the assumption of noninformative censoring, which is not fulfilled when studying the late surgery rate, the result was validated using competitive risk analysis with the method of Pepe and Mori. All tests were double-sided, and a p-value of ≤ 0.05 was considered significant. 95% confidence intervals are presented in parentheses. Confidence intervals are given in order to better describe the inferential uncertainty. The core facility of the Statistics Department of the Karolinska Institute was consulted in the analysis of the data. Ethics, funding, and potential conflicts of interest The study fulfilled Institutional Review Board requirements. The authors did not receive any funding and declare no potential conflicts of interest.

Results Oncological outcome of all patients The median overall survival of the 116 patients was 5.4 (3.2–6.8) months: two-thirds of the patients were alive after 3 months, half after 6 months, and one-third after 12 months. Only 3 of the 116 patients were alive at last follow-up. Patients in M-group had a median survival of 15 (6–24) months, whereas these in S-group had a median survival of 3 (1–5) months (p < 0.001) (Figure 1). The 3-month and 12-month sur-

21-02-2018 15:53:47

Acta Orthopaedica 2018; 89 (2): 229â&#x20AC;&#x201C;233

231

Figure 2. Flowchart showing the neurologic outcome of 114/116 patients with spinal metastasis who were initially treated non-surgically because of either absence of significant neurologic compromise whilst oncological prognosis is good (minor neurology), or poor prognosis and condition, widespread disease, or a combination of the above (poor prognosis). Neurologic status at a median of 2.5 weeks according to Frankel score: stable, improvement, or deterioration.

vival for M-patients was 85% and 58% respectively. Among S-patients, the 3 month-survival rate was 53%, and only 14% survived for 12 months. Neurologic outcome and incidence of late surgery Of the 116 patients who were initially assessed for their suitability to undergo surgery and subsequently managed nonoperatively, sufficient data for analysis of the neurologic outcome were available for 114 patients. Radiotherapy alone was efficient in retaining or even improving the neurologic function in the vast majority of patients (approximately 85% of the entire cohort). 8 patients underwent surgery at a later stage due to progression of the neurologic deficits; all them belonged to M-group (p < 0.001), the risk of undergoing late surgery for the M-group being 21% (Figure 2). Of the remaining 32 patients in the M-group who did not undergo late surgery, 28 had stable neurologic status, 3 improved, and 1 deteriorated.

Outcome of late surgery in the 8 patients initially not operated on 7/8 patients presented with more substantial compression at the same level, whereas 1 who received radiotherapy for cord compression at the 6th thoracic vertebra presented 4 years later with compression at the level of the 3rd thoracic vertebra and underwent laminectomy (Table 1). 5/8 patients experienced an improvement, 2/8 remained relatively stable, and 1/8 deteriorated. Of the 5 patients who benefited from surgery, ambulation was restored in 4/5. There was 1 postoperative complication, a wound break-down that required muscle transposition for coverage of the defect, which eventually healed. There were no deaths within the 30-day postoperative period. Value of the Bauer prognostic score in predicting the risk for late surgical decompression Prognostic scores have been widely used to estimate survival in patients with SM. Progression of a neurologic deficit may be associated with survival in 2 seemingly opposite ways: an aggressive tumor, with a poorer prognosis, may also be locally aggressive, but, on the other hand, longer survival with an indolent tumor may give it the necessary time to progress locally. We evaluated the Bauer score regarding its efficacy in estimating survival as well as risk for late surgery in M-patients. We found that a score of more than 2 was associated with superior overall survival and a lower risk for surgery (Figure 3). A separate competing-risk analysis of the rate of late surgical decompression with death being the competing event is given in Figure 4 (see Supplementary data).

Discussion During the past 2 decades, various studies have established the advantages of surgical decompression for treatment of patients with neurologic deficits because of SM. However,

Table 1. Details and outcome of 8 patients with spinal metastasis who were initially treated non-operatively but underwent late surgery due to progression of neurologic deficits

11793 Tsagozis D.indd 231

Months to secondary surgery

Neurologic status at last follow-up

Case

Age, sex

Histology

Level

67, male

Prostate

Thoracic

Frankel B

56, male

Prostate

Thoracic

Frankel E

3 4 5 6 7 8

57, female 54, male 74, male 72, male 46, female 69, female

Breast Lung Prostate Prostate Breast Lung

Thoracic Lumbar Thoracic Thoracic Thoracic Lumbar

1 4 2 1 14 1

Frankel E Frankel D Frankel D Frankel D Frankel E Frankel D

Comments Posterior decompression. Neurologic deterioration postoperatively Compression at new level. Posterior decompression. Wound dehiscence 3 weeks postoperatively â&#x20AC;&#x201C; Posterior decompression Posterior decompression Posterior decompression Anterior decompression and fixation Posterior decompression and fixation

21-02-2018 15:53:48

232

Acta Orthopaedica 2018; 89 (2): 229–233

Figure 3. Kaplan–Meier curve of overall survival (left panel) in 40 patients who were initially treated non-surgically for spinal metastasis, due to absence of significant neurologic deficits while oncological prognosis was good (M-group). Median overall survival of patients with Bauer score 0–2 was 9 months (CI 3–6), whereas in those with Bauer score 3–4 this was 45 months (CI 26–64). Kaplan–Meier curve of incidence of late surgery (right panel) in the same patients, depending on their Bauer score (p = 0.006).

in everyday clinical praxis the management of the individual patient remains a challenge. Many factors such as the degree and duration of the neurologic deficits, presence of compression at more than 1 level, the radiosensitivity of the tumor, and the volume of the soft-tissue component causing compression, the condition and expected survival of the patient, and how extensive the spinal metastasis is should be considered (Dea and Fisher 2014). Survival prognostic tools have been described and may help in decision-taking, since many of them also provide guidelines regarding optimal treatment (Tokuhashi et al. 1990, Bauer and Wedin 1995, Tomita et al. 2001, Leithner et al. 2008). Treatment choice is thus a complex individualized approach. We have always considered the presence of significant neurologic deficits as the primary indication for surgery (Jansson and Bauer 2006). Nonetheless, a significant proportion of these patients, approximately one-third according to our experience, are deemed as non-candidates for surgery. The presence of only minor neurologic deficits despite a good prognosis was the main reason in approximately 1/3 patients (M-group), whereas approximately 2/3 patients (S-group) were not operated because of a dismal prognosis. The management of a patient belonging to the M-group is a dilemma, especially in cases where there is a moderate epidural soft-tissue component and the neoplasm is radioresistant. Some authors advocate an aggressive en bloc resection in selected patients with solitary spinal metastasis and a good prognosis, especially in the case of certain histological diagnoses (Boland et al. 1982, Chataigner and Onimus 2000). However, the value of this approach has been disputed by others (Bilsky et al. 2009). Undoubtedly, initial non-operative treatment may eventually result in progression of the neurologic symptoms, necessitating late surgical intervention. In our series, late surgery was needed in approximately 1 out of 5 M-patients. We consider this to be an acceptable proportion, especially in view of the effect of radiotherapy as a sole treatment in preserving or even improving function in over 80% of the M-group. It should be noted that the oncological outcome

11793 Tsagozis D.indd 232

of the M-group was poor, with a median overall survival of 12 months, which explains the low incidence of secondary surgical intervention. Importantly, the outcome of late surgery was good, comparable to that achieved after primary surgery (Jansson and Bauer 2006). Thus, late decompression when needed appears a safe option, although we acknowledge that the number of patients in this subgroup is not large enough to draw safe conclusions and that no direct comparison with a primarily operated group was made. A primary clinical decision not to proceed to surgery was not altered in any patient with a poor prognosis, which highlights the fact that this clinical decision is usually accurate and such patients’ condition is usually irreversible. The modified Bauer score could identify patients at risk for secondary surgery. M-patients with a superior oncological prognosis had a lower risk for late surgery, possible due to the more indolent local progression of their tumor. This characteristic obviously outweighs the fact that patients with an aggressive tumor have a shorter life-span for a secondary neurologic compromise to occur. This information is valuable for both the physician and the patient during shared decision-making. In summary, our results show that among patients treated non-surgically because of SM with neurologic deficits, late surgical intervention may be necessary only in these who are assigned to this treatment due to minor neurologic deficits and not short expected survival, especially if they have a modified Bauer score of less than 2. However, initial non-surgical treatment is a safe option, since the overall risk is low and secondary decompression appears effective in restoring neurologic function. Supplementary data Figure 4 is available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674. 2017.1412193 PT: Data analysis and manuscript preparation. HB: study design and manuscript review.

21-02-2018 15:53:48

Acta Orthopaedica 2018; 89 (2): 229–233

Acta thanks John Healey and Cumhur Öner for help with peer review of this study. Bauer H C, Wedin R. Survival after surgery for spinal and extremity metastases. Prognostication in 241 patients. Acta Orthop Scand 1995; 66 (2): 143-6. Bilsky M H, Laufer I, Burch S. Shifting paradigms in the treatment of metastatic spine disease. Spine 2009; 34 (22 Suppl): S101-7. Boland P J, Lane J M, Sundaresan N. Metastatic disease of the spine. Clin Orthop 1982; (169): 95-102. Chataigner H, Onimus M. Surgery in spinal metastasis without spinal cord compression: Indications and strategy related to the risk of recurrence. Eur Spine J 2000; 9 (6): 523-7. Dea N, Fisher C G. Evidence-based medicine in metastatic spine disease. Neurol Res 2014; 36 (6): 524-9. Falicov A, Fisher C G, Sparkes J, Boyd M C, Wing P C, Dvorak M F. Impact of surgical intervention on quality of life in patients with spinal metastases. Spine 2006; 31 (24): 2849-56. Harrington K D. Orthopedic surgical management of skeletal complications of malignancy. Cancer 1997; 80 (8 Suppl): 1614-27. Ibrahim A, Crockard A, Antonietti P, Boriani S, Bünger C, Gasbarrini A, Grejs A, Harms J, Kawahara N, Mazel C, Melcher R, Tomita K. Does spinal surgery improve the quality of life for those with extradural (spinal) osseous metastases? An international multicenter prospective observational study of 223 patients. J Neurosurg Spine 2008; 8 (3): 271-8.

11793 Tsagozis D.indd 233

233

Jansson K-A, Bauer H C F. Survival, complications and outcome in 282 patients operated for neurological deficit due to thoracic or lumbar spinal metastases. Eur Spine J 2006; 15 (2): 196-202. Leithner A, Radl R, Gruber G, Hochegger M, Leithner K, Welkerling H, Rehak P, Windhager R. Predictive value of seven preoperative prognostic scoring systems for spinal metastases. Eur Spine J 2008; 17 (11): 1488-95. Patchell R A, Tibbs P A, Regine W F, Payne R, Saris S, Kryscio R J, Mohiuddin M, Young B. Direct decompressive surgical resection in the treatment of spinal cord compression caused by metastatic cancer: A randomised trial. Lancet 2005; 366 (9486): 643-8. Quan G M Y, Vital J-M, Aurouer N, Obeid I, Palussière J, Diallo A, Pointillart V. Surgery improves pain, function and quality of life in patients with spinal metastases: A prospective study on 118 patients. Eur Spine J 2011; 20 (11): 1970-8. Sundaresan N, Sachdev V P, Holland J F, Moore F, Sung M, Paciucci P A, Wu L T, Kelligher K, Hough L. Surgical treatment of spinal cord compression from epidural metastasis. J Clin Oncol 1995; 13 (9): 2330-5. Tokuhashi Y, Matsuzaki H, Toriyama S, Kawano H, Ohsaka S. Scoring system for the preoperative evaluation of metastatic spine tumor prognosis. Spine 1990; 15 (11): 1110-13. Tomita K, Kawahara N, Kobayashi T, Yoshida A, Murakami H, Akamaru T. Surgical strategy for spinal metastases. Spine 2001; 26 (3): 298-306. Young R F, Post E M, King G A. Treatment of spinal epidural metastases: Randomized prospective comparison of laminectomy and radiotherapy. J Neurosurg 1980; 53 (6): 741-8.

21-02-2018 15:53:48

234

Acta Orthopaedica 2018; 89 (2): 234–239

Risk of cancer after primary total hip replacement: The influence of bearings, cementation and the material of the stem A retrospective cohort study of 8,343 patients with 9 years average follow-up from Valdoltra Orthopaedic Hospital, Slovenia Vesna LEVAŠIC 1,2, Ingrid MILOŠEV 1,3, and Vesna ZADNIK 4

1 Valdoltra

Orthopaedic Hospital, Ankaran; 2 University of Ljubljana, Faculty of Medicine, Ljubljana, 3 Jožef Stefan Institute, Ljubljana; 4 Institute of Oncology Ljubljana, Ljubljana, Slovenia Correspondence: VZadnik@onko-i.si Submitted 2017-08-30. Accepted 2017-12-09

Background and purpose — Despite the increasing number of total hip replacements (THRs), their systemic influence is still not known. We have studied the influence of specific features of THRs—the bearing surface, the use of bone cement and the material of the stem—on the cancer incidence. Patients and methods — In a retrospective cohort study we identified 8,343 patients with THRs performed at Valdoltra Hospital from September 1, 1997 to December 31, 2009. Patient data were linked to national cancer and population registries. The standardized incidence ratios (SIR) and Poisson regression relative risks (RR) were calculated for all and specific cancers. Results — General cancer risk in our cohort was comparable to the population risk. Comparing with population, the risk of prostate cancer was statistically significantly higher in patients with metal-on-metal bearings (SIR = 1.35); with metal-on-polyethylene bearings (SIR = 1.30), with non-cemented THRs (SIR = 1.40), and with titanium alloy THRs (SIR = 1.41). In these last 3 groups there was a lower risk of hematopoietic tumors (SIR = 0.69; 0.66 and 0.66 respectively). Risk of kidney cancer was significantly higher in the non-metal-on-metal, non-cemented, and titanium alloy groups (SIR = 1.30; 1.46 and 1.41 respectively). Risk of colorectal and lung cancer was significantly lower in the investigated cohort (SIR = 0.82 and 0.83, respectively). Risk for all cancers combined as well as for prostate and skin cancer, shown by Poisson analysis, was higher in the metal-on-metal group compared with non-metal-on-metal group (RR = 1.56; 2.02 and 1.92, respectively). Interpretation — Some associations were found between the THRs’ features, especially a positive association between metalon-metal bearings, and specific cancers. ■

Total hip replacement (THR) implants are manufactured mainly from metallic materials in combination with polymers and/or ceramics. Since life expectancy is becoming prolonged, more people are being exposed to implanted metallic materials. During the lifetime of an endoprosthesis, dissolved metal ions and solid metal nanoparticles are formed and released into the surrounding tissue and blood (Milošev et al. 2005). Some of the metals used to manufacture THRs are recognized by the International Agency for Research on Cancer (IARC) to be cancerogenic (IARC 1999). Several studies have analyzed the connection between THRs and cancer incidence (Gillespie et al. 1988, Visuri and Koskenvuo 1991, Mathiesen et al. 1995, Nyrén et al. 1995, Visuri et al. 1996, Paavolainen et al. 1999, Signorello et al. 2001, Visuri et al. 2010, Brewster et al. 2013). To our knowledge the present study is the first in which the effect of the characteristics of endoprostheses on the incidence of cancer has been investigated. Thus, we assessed whether any of the features of the THRs, such as bearing surface between head and cup, cementation, or material of the stem influenced the cancer risk.

Patients and methods 8,343 consecutive patients with primary total THR implanted for the first time in the period from September 1, 1997 to December 31, 2009 were followed from the date of the first THR until the date of each cancer diagnosis (some patients had more than 1 cancer diagnosis in the study period), the date of death or the end of follow-up on January 1, 2015. NonSlovenian patients, patients undergoing an operation on the contralateral hip, and patients with diagnosis of rheumatoid arthritis were excluded.

12041 LEVAS¦îIC¦î D.indd 234

20-02-2018 12:22:42

Acta Orthopaedica 2018; 89 (2): 234–239

The data on patients were taken from the hospital computer database from September 1, 1997 to January 1, 2002 and, later, from the Valdoltra Arthroplasty Registry. These data are checked by hospital analytics service since payment for the material used is controlled by the State Health Insurance Company. From 2002 the Valdoltra Arthroplasty Registry has been running and its completeness is 100% (Levašič et al. 2009). Using the name, surname, sex, and birth date the cohort was linked with the Slovenian population-based cancer registry (CRS), which is filled by obligatory reporting. The quality and completeness indices are monitored and reported routinely so the CRS adequately covers the entire population (Zadnik et al. 2017). The vital status of the patients was checked by linkage with the Central Population Register, the central repository and processing of data concerning state citizens, which is the basis for establishing e-Government services. Diagnoses of cancer were classified as in the International Classification of Diseases version 10 (ICD-10). The analysis was stratified by cancer site: All sites (C00–C96); Melanoma (C43); Corpus uteri (C54); Prostate (C61); Haematolymphatic cancers (C81– C96); Stomach (C16); Lung and bronchi (C33–C34); Breast (C50); Lymphomas (C81–C85); Gastrointestinal tract (C15– C20); Urotract (C64–C67); Kidney and renal pelvis (C64– C65); Bladder (C67); Liver (C22); Central nervous system (C70–C72); Plasmacytoma (C90); Leukemias (C91–C95); Colon and rectum (C18–C20); Liver and gall bladder (C22– C23); Pancreas (C25); Skin cancer other than melanoma (C44) according to ICD-10. The characteristics of the THR implants were classified using the Valdoltra Implant Library, which records all the prostheses used in Valdoltra from September 1, 1997 onwards together with their characteristics—the parts of the prosthesis used, cementation use, the material, and size. The prostheses in our cohort were categorized one by one into subgroups according to (1) bearing type—metal-on-metal (MoM), metal-onpolyethylene (MoP), ceramic-on-ceramic (CoC), and ceramicon-polyethylene (CoP), and then divided into 2 groups: MoM and non-MoM, (2) use of bone cement for fixation and (3) material of the stem components: titanium (Ti) alloy, cobalt chromium (CoCr) alloy or stainless steel (Fe-alloy). Statistics The ratio of observed to expected numbers of cancer was expressed as the standardized incidence ratio (SIR). The expected number of cases was calculated by the formula: expected = ∑ ni Ri for each 5-year age-group i, where ni is the number of person-years at risk in our cohort and Ri is age-specific cancer incidence rate in Slovenia. The person-years were calculated by summing all the follow-up days. The SIR can be interpreted as an estimate of the relative risk of cancer in a particular category in comparison with the national rates. The 95% confidence interval (CI) is reported.

12041 LEVAS¦îIC¦î D.indd 235

235

The crude and to bearing surface, cementation, and stem material mutually adjusted relative risks of cancers for relevant sub-cohorts were assessed by Poisson regression analysis. The reference categories were: MoM in bearings, Ti-alloy in stem material, and non-cemented THRs in use of bone cement. The statistical analysis was made using the STATA13 program (StataCorp, College Station, TX, USA). Results with a p-value of less than 0.05 were considered statistically significant. Ethics, funding, and potential conflicts of interest The study was approved by the Republic of Slovenia National Medical Ethics Committee on April 5, 2013, N° 117/02/13. The study was funded by the Valdoltra Orthopaedic Hospital but no other support, financial or other, was received for this study. No competing interests declared.

Results Participants 8,343 patients were included in the study: 3,260 men and 5,083 women. The mean age at the time of THR was 65 years, 63 for men and 67 for women. None of the patients were lost to follow-up. In the follow-up period, 1,405 cancers were observed in this cohort. 2,101 patients died before the end of the follow-up period. The follow-up time ranged from 1 day (death of patient) to 17 years with a mean follow up of 9.0 years. The sub-cohorts for bearing surface analysis included 338 patients with MoM bearings, 5,909 with MoP, 1,323 with CoC, and 773 with CoP bearings. The sub-cohorts for the presence of bone cement included 6,966 patients with non-cemented THR and 1,377 patients with cemented components (900 cemented THR, 257 hybrid, and 220 reverse hybrid THR). The sub-cohorts for stem material included 7,316 patients with stems made of Ti-alloy, 850 of CoCr-alloy, and 177 of Fe-alloy (Table 1). The numbers of cancer cases and person years by subcohort are listed in Table 2. The cumulative follow-up time of the cohort was 77,075.56 person-years. SIR analysis For all cancer sites the risk in the whole cohort was the same as that in the general Slovenian population (SIR = 0.98, CI 0.93–1.03). The risk was higher for prostate cancer (SIR = 1.4, CI 1.2–1.6) and for kidney with renal pelvis cancer (SIR = 1.4, CI 1.0–1.8). The risk was lower for lung and bronchi cancers (SIR = 0.83, CI 0.69-0.98), for hematolymphatic cancers due to the lower risk of leukemias (SIR = 0.41, CI 0.24–0.71) and for gastrointestinal tract cancers due to the occurrence of colorectal cancers (SIR = 0.82, CI 0.70–0.96). In patients fitted with a Ti-alloy stem there was a higher risk of prostate (SIR = 1.4, CI 1.2–1.6) and kidney cancers (SIR = 1.4, CI 1.0–1.9) and a lower risk of leukemia (SIR = 0.34, CI 0.24–0.71) and colorectal cancer (SIR = 0.83, CI 0.71–0.98).

20-02-2018 12:22:42

236

Acta Orthopaedica 2018; 89 (2): 234–239

Table 1. Number of patients in sub-cohorts (bearing surface between head and cup, cementation, material of the stem) by age at operation Group Subgroup

Age group 50–59 60–69

70–79

≥ 80

Total

591

1,631

3,059

2,664

267

8,343

12 39 69 11

43 233 274 41

141 763 535 192

124 2,128 407 400

18 2,481 38 127

0 265 0 2

338 5,909 1,323 773

119 12

580 11

1,596 35

2,91 149

1,713 951

48 219

6,966 1,377

122 9 0

585 5 1

161 21 0

298 67 12

194 586 138

79 162 26

7,316 850 177

0–40

40–49

131

Table 2. Number of cancer cases and number of person-years according to sub-cohorts (bearing surface between head and cup, cementation, material of the stem) Group/Subgroup

All Bearing MoM MoP CoC CoP Use of bone cement Non-cemented Cemented Stem material Ti-alloy CoCr-alloy Fe-alloy

In the MoM and in the MoP bearings groups, analysis showed a significantly higher risk of prostate cancer (SIR = 2.4, CI 1.4–4.1) and (SIR = 1.4, CI 1.2–1.6), respectively. Further, in the non-MoM group there was a significantly higher kidney cancer risk (SIR = 1.3, CI 1.0–1.8). The MoM group for hematolymphatic cancers did not differ from that in the Slovenian population (SIR = 0.45, CI 0.11–1.8) and SIR was even significantly lower in MoP (SIR = 0.76, CI 0.59–0.98) and in all other non-MoM bearings (SIR = 0.69, CI 0.54–0.87). In the MoP group there was also a lower risk of gastrointestinal cancer (SIR = 0.73, CI 0.63–0.86). In the non-cemented THR cohort there was a higher risk of prostate (SIR = 1.4, CI 1.2–1.6) and of kidney (SIR = 1.5, CI 1.1–2.0) cancers, but a lower risk of hematolymphatic cancers (SIR = 0.66, CI 0.50–0.86), especially of leukemia (SIR = 0.33, CI 0.16–0.65). There was a lower risk in the noncemented group of colorectal cancers (SIR = 0.83, CI 0.72– 0.95). Gastrointestinal cancers were, on the whole, lower in both cemented (SIR = 0.83, CI 0.72–0.95) and non-cemented (SIR = 0.83, CI 0.72–0.95) groups. Because of the small sample size, it was not possible to calculate the SIR for each cancer site by prosthesis features. This was the case for stem material—CoCr-alloy for liver, Fe-alloy for liver, liver with gall bladder, central nervous system, plasmacytoma and leukemia, and for bearings—MoM for corpus uteri, liver, liver with gall bladder, central nervous system, plasmacytoma, and leukemias (Tables 3 and 4, see Supplementary data). Poisson analysis Under Poisson analysis, not mutually adjusted to THRs characteristics, the risk for all types of cancer is shown to be higher in the MoM group (RR = 1.56, CI 1.23–1.95) for prostate cancer (RR = 2.02, CI 1.17–3.48) and for skin cancers other than melanoma (RR = 1.92, CI 1.19–3.10). The risk does not

12041 LEVAS¦îIC¦î D.indd 236

Bearing MoM MoP CoC CoP Use of bone cement Non-cemented Cemented Stem material Ti-alloy CoCr-alloy Fe-alloy

No. cancers Person-years 76 1,003 188 138

4,378.52 51,673.26 13,340.13 7,683.65

1,156 249

65,624.09 11,451.47

122 156 29

68,694.83 6,932.79 1,447.95

depend on the stem material or bone cement usage (Table 5, see Supplementary data). As the relative risks for specific cancer sites are mutually adjusted to bearing surface, cementation, and stem material (Table 6, see Supplementary data) the risk for all types of cancer is again significantly higher in the MoM group (RR = 1.54, CI 1.22–1.93), for prostate cancer (RR = 2.01, CI 1.20– 3.57) and for skin cancers other than melanoma (RR = 2.01, CI 1.24–3.25). There is less prostate cancer in femoral stems from CoCr-alloy (RR = 0.32, CI 0.12–0.87). A significantly higher risk exists of central nervous system cancer in the cemented THR group (RR = 5.87, CI 1.21–28.46). Similarly to that in the SIR analysis, it was not possible to calculate relative risks for each cancer site using prosthesis features because of small sample size.

Discussion Key results Comparing the groups with the general population using SIR analysis showed the general cancer risk in the whole cohort was comparable to that in the population. The prostate cancer risk was higher for the MoM and MoP bearings. The same was true for both the non-cemented THRs and for the Ti-alloy THRs. The risk of kidney cancer was higher in the non-MoM, non-cemented, and Ti-alloy groups. The risks of colorectal cancer and lung cancer were lower for the complete cohort. That for hematolymphatic cancers for MoM SIR was the same while, for the non-MoM, non-cemented, and Ti-alloy groups, SIR was even lower. Poisson analysis showed that risk for all types of cancer was higher in the MoM group compared with non-MoM and so too for prostate cancer and skin cancers. These results suggest that the bearing surface is the main feature of THR influencing the cancer risk.

20-02-2018 12:22:42

Acta Orthopaedica 2018; 89 (2): 234–239

Possible study improvements Cohort expansion or follow-up time elongation would improve the power of our analysis for certain less common cancer sites. Inclusion of further confounders (multiple joint replacements, other than RA concomitant diseases…) would give additional strength to the causality results. Interpretation The question as to the influence of the hip implant on the human body is probably as old as the THR procedure itself. The toxicity of the metals used in orthopedic prostheses was questioned back in 1981 (Rae 1981). Some bearing surfaces, such as MoM, were investigated by Huk et al. since they generate both metal particles and ions (Huk et al. 2004). Arthroplasty registries, especially in Scandinavian countries, have already been used to obtain a better insight into this problem. Several studies and meta-analyses have been carried out, but with differing results (Visuri and Koskenvuo 1991, Paavolainen et al. 1999, Signorello et al. 2001, Tharani et al. 2001, Visuri et al. 2003, Onega et al. 2006, Visuri et al. 2010, Smith et al. 2012, Wagner et al. 2012, Brewster et al. 2013, Mäkelä et al. 2014). The surrounding tissue reacts to wear debris by undergoing a local tissue reaction in which the aseptic loosening is thought to be due to the response of macrophages and to involve hypersensitivity (Milošev 2006). There is a growing consensus that metal-induced DNA damage may lead to carcinogenesis. Keegan et al. (2007) analyzed many kinds of systemic toxicology, especially of Al, Cr(VI), Co, Ni, and V(V) and outlined the ‘potential hazards’ of circulating metals that include potential harmful effects on immunity, reproduction, the kidney, developmental toxicity, the nervous system, and carcinogenesis. The International Agency for Research on Cancer (IARC) has classified Cr-(VI) and Ni-(II) as being carcinogenic, metallic Ni and soluble Co as possible carcinogens, and metallic Cr, Cr-(III) compounds and implanted orthopedic alloys as being unclassifiable. Possible pathogenetic mechanisms could involve direct mutagenic effects of metal ions or, in cases of hematolymphatic cancers, chronic antigenic stimulation of lymphocytes (IARC 2006). Ni is a highly allergenic element, Co is directly toxic, while bone cement has methylmethacrylate (MMA) as one of the main substances—not classifiable as to its carcinogenicity to humans by IARC 1994 (IARC 1994). But the monomer MMA is a lipid solvent and therefore could pass the hemato-encephalic membrane and affect the central nervous system. This would explain one of our results of Poisson analysis which indicates that use of cement increases risk of cancer in the central nervous system. Some local neurotoxicity has been diagnosed in dental technicians from contact exposure (Seppäläinen and Rajaniemi 1984). Willert and Semlitsch (1977) reported the development of a foreign-body reaction to wear debris, consisting of macrophages and foreign-body giant cells. It is possible that the bearing type influences local and, eventually, also systemic

12041 LEVAS¦îIC¦î D.indd 237

237

host defenses (Trebše et al. 2014). In the study of Milošev (2006) the histological picture of periprosthetic tissue of revised THR patients showed lymphocytic aggregates in MoP patients, but they were less frequent and of much lower intensity than those in MoM patients. MoM implants are associated with more adverse events because of metal-ion-induced, T-cell-mediated delayed-type hypersensitivity (Konttinen and Pajarinen 2012). Adverse tissue reactions in MoM THRs can be systemic or local. Higher serum and solid organ metal ion levels may, theoretically, have carcinogenic and teratogenic potential (Lohmann 2014). Metal ions can activate toll-like receptor signaling, so they can also function as haptens, activating the adaptive immune system (Pajarinen et al. 2014). The metal nanoparticles that are released from both MoM and MoP bearings result in a postoperative increase in metal ion levels at various organ sites. It is hypothesized that metalinduced DNA damage may lead to hematopoetic, prostate endometrial cancer as well as malignant melanoma (Polyzois et al. 2012). A meta-analysis by Visuri et al. (2003) of 6 Nordic cohorts showed that the overall incidence of cancer was reduced in gastrointestinal (GIT) cancers (stomach, colon, and rectum). All cohorts showed a reduced incidence of lung cancer. The incidence was increased for endometrial cancer; prostate cancer was slightly increased in all cohorts. All forms of hematopoietic cancers, except leukemia, were slightly increased. GIT cancers are affected by use of nonsteroidal anti-inflammatory drugs (NSAIDs), which can prevent the development of colorectal cancers (Muscat et al. 1994, Hawk et al. 1999). Patients who are treated with joint replacement have often, for many years, consumed NSAIDs and, in many cases, continue to take them as painkillers for pain from other locations as well. The higher risk for prostate cancer resulting in our SIR analysis also has also been found in other studies (Signorello et al. 2001, Visuri et al. 2010, Brewster et al. 2013). Possible reasons could be found in Watson et al. (2010). These authors claim that certain metal ions have the ability to bind to estrogen receptors. These xenestrogens (metalloestrogens) can facilitate both synthesis and regulated secretion of prolactin, a growth factor for target tissues such as breast and prostate. The incidence of prostate cancer could increase by detection bias and in the latter part of the follow-up may reflect a real increase in risk (Wagner et al. 2011). Another hypothesis is that metal-induced DNA damage may lead to carcinogenesis, including hematopoetic, prostate, and endometrial cancer and malignant melanoma (Lewis and Sunderman Jr 1996). Obesity is a risk factor for endometrial cancer, renal parenchyma cancer, and gallbladder cancer in women (Olsen et al. 1999). Since the average BMI of patients at the Valdoltra hospital is in the range of overweight people, this could affect the final result (Levašič and Milošev 2017). Ni and Co are rapidly excreted by the kidney, but Cr is concentrated in the epithelial cells of the proximal renal tubules, so this could harm the kidney tissue (Oliveira et al. 2006).

20-02-2018 12:22:42

238

The reduction in the number of respiratory tract cancers could be explained by reduced smoking, assuming that people who decide to undergo the THR have a healthier lifestyle (inclusion bias). One of the connections could be that the risk of osteoarthritis of the hip is lower in male smokers than in non-smokers (Makela et al. 2012). As smaller numbers of cases of lung cancer have been found in several studies, another mechanism not yet proven could provide the explanation alternatively (Mathiesen et al. 1995, Lewold et al. 1996, Visuri et al. 1996, Paavolainen et al. 1999, Signorello et al. 2001, Goldacre et al. 2005, Visuri et al. 2010, Makela et al. 2012). Hematolymphatic cancers—leukemic cancers and lymphomas—have been associated with McKee-Farrar (MoM) prostheses but the risk decreased after the first year following surgery (Visuri and Koskenvuo 1991). The incidence of lymphoma and of leukemia can be associated with rheumatoid arthritis (RA) or its treatment (Gillespie et al. 1988). SIR analysis in our cohort showed fewer leukemias and, in consequence, fewer hematolymphatic cancers, but only in nonMoM bearings, non-cemented THR, and the Ti-stem. This is most probably due to the exclusion of patients with RA from the cohort to avoid the bias of the higher percentage of RA patients since they are more prone to receive artificial joints than the normal population. Unlike in some other studies, as reported by Makela et al. (2012), we did not find a higher incidence of sarcomas. In the whole cohort we found 5 sarcomas, none of which was at the site of endoprosthesis. Generalizability From all the features of THRs we studied the main associations were found between the bearing surface, especially metal-on-metal bearings, and specific cancers. Absolute causality remains to be confirmed by other means; however, MoM bearings should be used only in individually chosen cases. Our results are transferrable to other countries, since the Slovenian cancer incidence rates applied in our analysis are comparable to the population rates in other European countries (Zadnik et al. 2017). Supplementary data Tables 3–6 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674. 2018.1431854

All authors contributed to the writing of the paper. VL wrote the first draft, VZ undertook the study design and contributed to the statistical analysis, and IM contributed to the section on materials. Acta thanks Robert Grimer and Gerold Labek for help with peer review of this study.

12041 LEVAS¦îIC¦î D.indd 238

Acta Orthopaedica 2018; 89 (2): 234–239

Brewster D H, Stockton D L, Reekie A, Ashcroft G P, Howie C R, Porter D E, Black R J. Risk of cancer following primary total hip replacement or primary resurfacing arthroplasty of the hip: a retrospective cohort study in Scotland. Br J Cancer 2013; 108(9): 1883-90. Gillespie W J, Frampton C M A, Henderson R J. The incidence of cancer following hip replacement. J Bone Joint Surg Br 1988; 70-B(4): 539-42. Goldacre M J, Wotton CJ, Seagroatt V, Yeates D. Cancer following hip and knee arthroplasty: record linkage study. Br J Cancer 2005; 92(7): 1298301. Hawk E, Lubet R, Limburg P. Chemoprevention in hereditary colorectal cancer syndromes. Cancer 1999; 86(S11): 2551-63. Huk O L, Catelas I, Mwale F, Antoniou J, Zukor D J, Petit A. Induction of apoptosis and necrosis by metal ions in vitro. J Arthroplasty 2004; 19(8, Suppl. 3): 84-7. IARC. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans: Some industrial chemicals. IARC Expert Committee on Cancer Statistics, editor. Lyon: World Health Organization, International Agency for Research on Cancer; 1994. IARC. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans: Surgical implants and other foreign bodies. Lyon, France: World Health Organization, International Agency for Research on Cancer; 1999. IARC. IARC Monographs on the Evaluation of the Carcinogenic Risks to Humans: Cobalt in hard metals and cobalt sulfate, gallium arsenide, indium phosphide, and vanadium pentoxide. Lyon, France; Geneva, Switzerland: International Agency for Research on Cancer; 2006. Keegan G M, Learmonth I D, Case C P. Orthopaedic metals and their potential toxicity in the arthroplasty patient: a review of current knowledge and future strategies. J Bone Joint Surg Br. 2007; 89-B(5): 567-73. Konttinen YT, Pajarinen J. Surgery: Adverse reactions to metal-on-metal implants. Nat Rev Rheumatol 2012; 9(1): 5-6. Levašič V, Milošev I. Valdoltra Arthroplasty Registry Report [Internet]. http:// www.ob-valdoltra.si/sl/international. 2017 [cited 2017 Jul 5]. Levašič V, Pišot V, Milošev I. Arthroplasty register of the Valdoltra Orthopaedic Hospital and implant retrieval program. Zdr Vestn 2009; 78 (Suppl II): 71-8. Lewis C G, Sunderman Jr F W. Metal carcinogenesis in total joint arthroplasty: animal models. Clin Orthop 1996; (329): S264-S268. Lewold S, Olsson H, Gustafson P, Rydholm A, Lindgren L. Overall cancer incidence not increased after prosthetic knee replacement: 15,551 patients followed for 66,622 person years. Int J Cancer 1996; 68(1): 30-3. Lohmann C H. Metallic debris from metal-on-metal total hip arthroplasty regulates periprosthetic tissues. World J Orthop 2014; 5(5): 660. Makela K T, Visuri T, Pulkkinen P, Eskelinen A, Remes V, Virolainen P, Junnila M, Pukkala E. Risk of cancer with metal-on-metal hip replacements: population based study. BMJ. 2012; 345(jul25 2): e4646-e4646. Mäkelä K T, Visuri T, Pulkkinen P, Eskelinen A, Remes V, Virolainen P, Junnila M, Pukkala E. Cancer incidence and cause-specific mortality in patients with metal-on-metal hip replacements in Finland: a populationbased study with a mean follow-up of 4.6 (1-11) years. Acta Orthop 2014; 85(1): 32-8. Mathiesen E, Ahlbom A, Bermann G, Lindgren J U. Total hip replacement and cancer. J Bone Jt Surg Br 1995; 77-B(3): 345-50. Milošev I. Survivorship and retrieval analysis of Sikomet metal-on-metal total hip replacements at a mean of seven years. J Bone Joint Surg Am 2006; 88(6): 1173. Milošev I, Pišot V, Campbell P. Serum levels of cobalt and chromium in patients with Sikomet metal-metal total hip replacements. J Orthop Res 2005; 23(3): 526-35. Muscat J, Stellman S D, Wynder E L. Nonsteroidal antiinflammatory drugs and colorectal cancer. Cancer 1994; 74(7): 1847-54. Nyrén O, McLaughlin J K, Gridley G, Ekbom A, Johnell O, Fraumeni J F, Adami H-O. Cancer risk after hip replacement with metal implants: a population-based cohort study in Sweden. J Natl Cancer Inst 1995; 87(1): 28-33.

20-02-2018 12:22:42

Acta Orthopaedica 2018; 89 (2): 234–239

Oliveira H, Santos T M, Ramalho-Santos J, de Lourdes Pereira M. Histopathological effects of hexavalent chromium in mouse kidney. Bull Environ Contam Toxicol 2006; 76(6): 977-83. Olsen J H, McLaughlin J K, Nyrén O, Mellemkjaer L, Lipworth L, Blot W J, Fraumeni J F. Hip and knee implantations among patients with osteoarthritis and risk of cancer: a record-linkage study from Denmark. Int J Cancer 1999; 81(5): 719-722. Onega T, Baron J, MacKenzie T. Cancer after total joint arthroplasty: a metaanalysis. Cancer Epidemiol Biomarkers Prev 2006; 15(8): 1532-7. Paavolainen P, Pukkala E, Pulkkinen P, Visuri T. Cancer incidence in Finnish hip replacement patients from 1980 to 1995: a nationwide cohort study involving 31,651 patients. J Arthroplasty 1999; 14(3): 272-280. Pajarinen J, Jamsen E, Konttinen Y T, Goodman S B. Innate immune reactions in septic and aseptic osteolysis around hip implants. J Long Term Eff Med Implants [Internet]. 2014 [cited 2017 Jul 28]; 24(4). Polyzois I, Nikolopoulos D, Michos I, Patsouris E, Theocharis S. Local and systemic toxicity of nanoscale debris particles in total hip arthroplasty: nanoparticles toxicity and total hip arthroplasty. J Appl Toxicol 2012; 32(4): 255-69. Rae T. The toxicity of metals used in orthopaedic prostheses. An experimental study using cultured human synovial fibroblasts. J Bone Joint Surg Br 1981; 63-B(3): 435-40. Seppäläinen A, Rajaniemi R. Local neurotoxicity of methyl methacrylate among dental technicians. Am J Ind Med 1984; 5(6): 471-7. Signorello L B, Ye W, Fryzek J P, Lipworth L, Fraumeni J F, Blot W J, McLaughlin J K, Nyrén O. Nationwide study of cancer risk among hip replacement patients in Sweden. J Natl Cancer Inst 2001; 93(18): 14051410. Smith A J, Dieppe P, Porter M, Blom A W, on behalf of the National Joint Registry of England and Wales. Risk of cancer in first seven years after metal-on-metal hip replacement compared with other bearings and general population: linkage study between the National Joint Registry of England and Wales and hospital episode statistics. BMJ 2012; 344(apr03 1): e2383-e2383.

12041 LEVAS¦îIC¦î D.indd 239

239

Tharani R, Dorey F J, Schmalzried T P. The risk of cancer following total hip or knee arthroplasty. J Bone Joint Surg 2001; 83(5): 774-80. Trebše R, Levašič V, Milošev I, Kovač S. Does the bearing type influence the incidence of periprosthetic infections of the hip? CeraNews 2014; 12. Visuri T, Koskenvuo M. Cancer risk after McKee-Farrar total hip replacement. Orthopedics 1991; 14(2): 137-42. Visuri T, Pukkala E, Paavolainen P, Pulkkinen P, Riska E B. Cancer risk after metal on metal and polyethylene on metal total hip arthroplasty. Clin Orthop 1996; 329: S280-S289. Visuri T, Pukkala E, Pulkkinen P, Paavolainen P. Decreased cancer risk in patients who have been operated on with total hip and knee arthroplasty for primary osteoarthrosis: a meta-analysis of 6 Nordic cohorts with 73,000 patients. Acta Orthop Scand 2003; 74(3): 351-60. Visuri T, Pulkkinen P, Paavolainen P, Pukkala E. Cancer risk is not increased after conventional hip arthroplasty: a nationwide study from the Finnish Arthroplasty Register with follow-up of 24,636 patients for a mean of 13 years. Acta Orthop 2010; 81(1): 77-81. Wagner P, Olsson H, Lidgren L, Robertsson O, Ranstam J. Increased cancer risks among arthroplasty patients: 30 year follow-up of the Swedish Knee Arthroplasty Register. Eur J Cancer 2011; 47(7): 1061-71. Wagner P, Olsson H, Ranstam J, Robertsson O, Zheng M H, Lidgren L. Metal-on-metal joint bearings and hematopoetic malignancy: a review. Acta Orthop 2012; 83(6): 553-8. Watson C S, Jeng Y-J, Kochukov M Y. Nongenomic signaling pathways of estrogen toxicity. Toxicol Sci 2010; 115(1): 1-11. Willert H, Semlitsch M. Reactions of the articular capsule to wear products of artificial joint prostheses. J Biomed Mater Res 1977; 11: 157-64. Zadnik V, Primic Zakelj M, Lokar K, Jarm K, Ivanus U, Zagar T. Cancer burden in Slovenia with the time trends analysis. Radiol Oncol [Internet]. 2017 Jan 22 [cited 2017 Oct 24]; 51(1).

20-02-2018 12:22:42

240

Acta Orthopaedica 2018; 89 (2): 240–245

Patient injury claims involving fractures of the distal radius 208 compensated claims from the Finnish Patient Insurance Center Henrik SANDELIN 1, Eero WARIS 2, Eero HIRVENSALO 1, Jarkko VASENIUS 3, Heini HUHTALA 4, Timo RAATIKAINEN 2 and Teemu HELKAMAA 1

1 Department

of Orthopaedics and Traumatology, Helsinki University Central Hospital and University of Helsinki, Helsinki; 2 Department of Hand Surgery, Helsinki University Hospital and University of Helsinki, Helsinki; 3 Pohjola Hospital, Helsinki; 4 Faculty of Social Sciences, University of Tampere, Tampere, Finland Correspondence: henrik.sandelin@helsinki.fi Submitted 2017-06-26. Accepted 2017-11-14.

Background and purpose — Optimal treatment for distal radius fractures remains controversial, with a significant number of fractures resulting in complications and long-term morbidity. We investigated patient injury claims related to distal radius fractures to detect the critical steps in the treatment leading to avoidable adverse events Patients and methods — We analyzed all compensated patient injury claims in Finland between 2007 and 2011. Claims were collected from the Patient Insurance Center’s (PIC) nationwide claim register. Patients of all ages were included. Each claim decision, original patient records, and radiographs related to treatment were reviewed. Results — During the study period, the PIC received 584 claims regarding distal radius fractures, of which 208 (36%) were compensated. Pain and impaired wrist function were the most common subjective reasons to file claims among compensated patients. In 66/208 patients, more than 1 adverse event leading to patient injury was detected. The detected adverse events could be divided into 3 main groups: diagnostic errors (36%, n = 103), decision/planning errors (30%, n = 87), and insufficient technical execution (32%, n = 91). Issues related to malalignment were the main concerns in each group. Diagnostic errors were often related to incorrect assessment of the fracture (re)displacement (75%, n = 78). All of the decision-making errors concerned physicians’ decisions to accept unsatisfactory fracture alignment. The most common technical error was insufficient reduction (29%, n = 26). Interpretation — We identified avoidable adverse events behind patient injuries related to distal radius fracture treatment. This study will help physicians to recognize the critical steps in the treatment of this common fracture and enhance patient safety. ■

Distal radius fractures account for approximately 15% of all fractures treated in emergency departments (Chung and Spilson 2001) and the age distribution is bimodal (Flinkkilä et al. 2011, Wilcke et al. 2013). Despite increasing scientific evidence and published current care guidelines, there is a wide variation in the treatment practice for distal radius fractures (Egol et al. 2010, Lichtman et al. 2010, Arora et al. 2011, Costa et al. 2014, Distal radius fracture: Current Care Guidelines, 2016). It seems that subjective opinions (Walenkamp et al. 2016) and the specialty (Chung et al. 2011) of the physician influence treatment decisions. There is a wide variety of fracture patterns and varying degrees of experience among physicians treating distal radius fractures. Therefore, the treatment of distal radius fractures is understandably susceptible to adverse events and complications leading to patient injuries (Khan and Giddins 2010, Mahdavian Delavary et al. 2010, Statistics of Finnish Patient Insurance Center, 2014, Lutz et al. 2014, Mathews and Chung 2015). Although patient injuries usually represent the more severe end of adverse events, the mechanisms behind severe adverse events and adverse events in general are often similar. Patient injuries therefore provide important data to help prevent adverse events and to improve patient safety (Mikkonen 2004, Järvelin 2012). In accordance with the revised patient injury act (Patient Injury Act. 25.7.1986/585), the Finnish Patient Insurance Center (PIC) covers and handles all suspected patient injuries in public or private health care in Finland that occur during medical treatment. New Zealand and the Nordic countries, including Finland, use the no-fault patient insurance system, as opposed to the tort insurance system used in the United

11831 Sandelin D.indd 240

20-02-2018 12:33:22

Acta Orthopaedica 2018; 89 (2): 240–245

241

Patient injury claims recorded as distal radius fractures n = 596 Accepted claims n = 220

Denied claims n = 376

Claims excluded (n = 12): – registry error, 7 – medication error, 1 – decision pending, 4 Compensated claims n = 208 Non-operative treatment n = 159

Operative treatment n = 45

Both operative and non-operative treatment n=4

Figure 1. Flowchart of patient injury data collection.

Kingdom and United States (Mikkonen 2004). The primary task of both tort and no-fault systems is to determine whether patients’ claims are eligible for compensation, as well as the amount of monetary compensation (Järvelin 2012). In both systems, the consequences of the adverse event must always be severe enough to merit compensation. However, the nofault system does not aim to find out who is to blame and therefore only rarely do claims advance to legal courts. 7 compensation criteria are defined in the Patient Injury Act (Patient Injury Act. 25.7.1986/585) and the criterion most often used is the so-called “preventability rule”; i.e., is it likely that an experienced medical professional would have avoided the patient injury event by taking a different action. Exceptions to this rule include infections and unreasonable injuries that are generally unavoidable and therefore compensated in accordance with the “tolerability” rather than the “preventability” concept (Helkamaa et al. 2016). The main objective of this study was to investigate compensated avoidable patient injuries related to fractures of the distal radius and to determine the underlying reasons for these severe adverse events.

individually analyzed by an independent researcher (HS). We collected the following data: patient characteristics, detailed information concerning the treatment and reoperations, subjective reasons for filing claims, and reasons for compensation. Patient injury classifications All of the compensated patient injury claims we identified were compensated based on the “preventability rule” (see above). For injuries caused by medical management, we used the term “adverse event.” We further classified all adverse events into 5 subgroups based on our own analyses and the evaluation of the PIC’s external medical advisors for reasons of reimbursement: diagnostic errors, decision-making errors, technical errors, follow-up planning errors, and other errors (Figure 2). Statistics We estimated the nationwide number of distal radius fractures using the age- and sex-specific annual mid-population obtained from the Finnish Official Statistics (Statistics Finland) and the previously published age- and sex-specific incidence of distal radius fractures from Finland, which included patients treated in public as well as private clinics (Flinkkilä et al. 2011). We compared the risks for a compensated patient injury in each age group and for both sexes, using regression models and Poisson-based confidence intervals. All analyses were carried out using SPSS for Windows version 22.0 (IBM Corp, Armonk, NY, USA), STATA 13.0 (StataCorp, College Station, TX, USA) and Confidence Interval Analysis (CIA) 2.2 software. 95% confidence intervals (CI) were calculated using Wilson’s exact method. Ethics, funding, and potential conflicts of interest Approval for the study was obtained from the PIC and Finnish Ministry of Social Affairs and Health. Funding support was provided by the Finnish Medical Association, Patient Insurance Center (PIC), and Helsinki University Central Hospital. No competing interests declared.

Patients and methods

Results

Data collection We analyzed all patient injury claims concerning the treatment of distal radius fractures that had been filed between January 1, 2007 and December 31, 2011. Claims (n = 596) were collected from the national claim register (Figure 1). To find all the claims, we used the International Classification of Diseases, tenth revision (ICD-10) for diagnosis code S52.5 (distal radius fracture) and the Nordic Medico-Statistical Committees Classification of Surgical Procedures (NCSP) for procedure codes (NCJ40-99, NDJ40-99, NDK00-99). Each case (original patient records, radiographs, and claim decisions) was

Claims and claimants After reviewing all patient injury claims and excluding registry errors, we found 584 closed patient injury claims that concerned distal radius fracture treatment. Compensation was granted for 208 (36%) claimants (Figure 1 and Table 1), of whom 5 were children under 16 years of age, 119 were adults (age 16–64), and 84 were elderly over 65 years of age. Among compensated claimants, pain was the main reason to file a claim (67%), followed by impaired wrist function (62%), suspicion of incorrect treatment (38%), and malalignment of the wrist (31%). Reporting the reason for filing a claim was

11831 Sandelin D.indd 241

20-02-2018 12:33:22

242

Acta Orthopaedica 2018; 89 (2): 240–245

Table 1. Patient characteristics

Patients Female Mean age Injured side Right Injury type Leisure/sports Work related Traffic School/college Injury mechanism Low-energy trauma a High-energy trauma Occupation Retired Post-fracture physiotherapy received AO fracture classification A2 A3 B1 B2 B3 C1 C2 C3 Radiographs unavailable a Falling

Table 2. Claimants’ subjective reasons for filing a claim n

208 153 59

176 24 6 2

85 12 3 1

142 66

68 32

104 146

50 70

50 35 6 1 4 25 57 14 16

24 17 3 0.5 2 12 27 7 8

range

(8–90)

on the same level while standing or walking.

an official part of the claim filing process. Only 1 claimant had not specified a reason for the claim (Table 2). The average times for claimants to file and for PIC to handle the claims were 9 (1–36) months from the injury and 4 (1–17) months, respectively. Reasons for adverse events We detected 288 discrete adverse events in 208 compensated claimants. 66/208 patients had more than 1 adverse event during their treatment. All compensated patient injury claims concerned management of the index fracture and all were considered to be avoidable injuries by PIC’s medical advisors. We further classified adverse events into subgroups based on error types and their temporal occurrence during treatment (Figure 2). Diagnostic errors accounted for 103 (36%) adverse events. Three-fourths of the diagnostic error subtypes concerned failure to diagnose primary displacement of the fracture or re-displacement of an adequately reduced fracture during follow-up (n = 78, 75%). In these cases the physician explicitly failed to notice the displacement and unintentionally treated the fracture, as the fracture alignment would have been satisfactory. In 12 cases the diagnostic error was repeated in consecutive follow-ups. Decision-making errors accounted for 53 (18%) adverse events. In these cases the physicians correctly diagnosed the

11831 Sandelin D.indd 242

n Pain Impaired wrist function Incorrect treatment Visual deformity Loss of income/additional expenses Prolonged recovery time Poor doctor–patient relationship Cosmetic harm (e.g., scar) Mental stress Need of professional re-education Corneal erosion during anesthesia

139 129 79 65 48 33 12 10 7 1 1

% 67 62 38 31 23 16 6 5 3 0 0

207 claimants out of 208 had a total of 524 subjective reasons for claims.

fracture displacement, accepted it, and intentionally continued the treatment without any interventions (Figure 2, legend). Technical errors accounted for 91 (32%) adverse events. Half of the adverse events in non-operative treatment occurred because of inadequate casting technique or insufficient reduction of the fracture. As in the non-operative group, failure to reduce the fracture was a common problem among patients operated on, accounting for 27% of adverse events in this group (n = 14). Follow-up planning errors accounted for 34 (12%) adverse events. Inadequately timed or lacking follow-up visits in the early post-treatment period accounted for 59% of these adverse events (n = 20). In non-operative treatment, 55% of the adverse events occurred at the follow-up visits whereas in operative treatment 80% of adverse events occurred during the primary operation. Adverse events were considered compensable if the first 2 controls were missed altogether or if the 2-week control was missed or arranged too late when fracture union had already occurred. For age and sex, there were no statistically significant differences for the different subgroups or the combinations of errors. Evaluation of the study population From 2007 to 2011, we estimated that a total of 64,990 fractures of the distal radius occurred in Finland: approximately 13,000 fractures per year (Flinkkilä et al. 2011, Statistics Finland). There was a strong relationship between the number of compensated patient injuries and the total number of distal radius fractures (estimate) regarding 2 confounding factors, age and sex (Figure 3). The calculated risk for a compensated patient injury varied from 0.06% (CI 0.01–0.32) to 0.5% (CI 0.28–0.91) among different age groups and sexes. The differences between groups were not statistically significant. Based on our estimation, the risk for a patient injury, defined as a compensated PIC claim, in distal radius fracture treatment was 0.3% in Finland.

20-02-2018 12:33:22

Acta Orthopaedica 2018; 89 (2): 240–245

243

Compensated patient injury claims n = 208 Detected adverse events n = 288

Diagnostic errors (n = 103): – failure to diagnose redisplacement of fracture, 50 – failure to diagnose primary displacement of fracture, 28 – radiographs not taken, 13 – fracture missed on radiographs, 8 – other, 4

Decision-making errors (n = 53): – acceptance of unsatisfactory fracture alignment without diagnostic error a, 53

Technical errors n = 91

Operative treatment errors (n = 51): – inadequate reduction, 14 – nerve lesion b, 7 – intra-articulate plate screws, 7 – wrong surgical procedure, 6 – inadequate volar plate fixation, 5 – other, 12

Follow-up and planning errors (n = 34): – inadequate timed or missed follow-ups c, 20 – inadequate imaging, 4 – inadequate immobilisation, 3 – other, 7

Other errors n=7

Non-operative treatment errors (n = 40): – inadequate casting position, 12 – inadeequate reduction, 4 – unnecessary reduction, 4 – other, 12

Figure 2. Reasons for adverse events. Treatment injuries, referred to as adverse events in the text, were further classified into subgroups by error type and their consecutive occurrence during treatment. a Fracture was not primarily reduced or re-reduced during follow-up. Non-operative treatment was not switched to operative treatment or the patient was not referred to a specialist. b Distribution of 7 nerve lesions: Median nerve (3), superficial radial nerve (2), ulnar nerve (1), bone graft: lateral femoral cutaneous nerve (1). c Normal follow-up visits at 1, 2, and 5 weeks, with radiograph taken at 1 and 2 weeks.

Number of distal radius fractures ––––

Number of compensated patient injuries - - - -

14,000

50 Women Men Total

12,000

40 10,000 30

8,000 6,000

4,000 10 2,000 0

0 16–19 20–29 30–39 40–49 50–59 60–69 70–79 80–89 90–99

Age groups

Figure 3. Estimate of the total number of fractures of the distal radius and compensated claimants during the study period (2007–2011) in Finland. For men, the total number of fractures of the distal radius (continuous line) and compensated claimants (dashed line) was highest in the 50–59 years age group, whereas for women it was in the 70–79 years age group.

Discussion To our knowledge, this is the first nationwide analysis of compensated patient injury claims among patients treated for fractures of the distal radius. Our analysis of adverse events revealed that errors could be divided into 3 main groups based on the stage at which the adverse event occurred: diagnostic, decision/planning, and technical. Each group accounted for approximately one-third of all adverse events. All of these

11831 Sandelin D.indd 243

errors were avoidable. Over half of all adverse events comprised situations where the fracture displacement was misdiagnosed or where it was diagnosed correctly but mismanaged. At best, fractures of the distal radius heal completely without significant functional limitations. However, a significant number of distal radius fracture treatments result in complications and long-term morbidity (Friedman 2005). Several studies of complications after distal radius fracture treatment have focused on clinical outcome (McKay et al. 2001, Lutz et al. 2014). However, there is surprisingly little information about adverse events associated with distal radius fracture treatment leading to complications and even less information regarding why these adverse events occur and how they can be avoided (DeNoble et al. 2014, Mathews and Chung 2015). Diagnostic errors were the most common adverse events that we detected. Several previous medical reports have noted similar findings (Guly 2001, Brown et al. 2010, Saber Tehrani et al. 2013, Ring et al. 2014, Talbot et al. 2014). The majority of diagnostic errors (73%) that we identified in our study were due to physicians failing to assess displacement or redisplacement of the fracture. Diagnostic errors were recently acknowledged to be one of the most common and harmful patient safety problems (Singh and Graber 2015). Since 2008, working groups, conferences, and societies have been established to address this large and seldom-discussed problem (Croskerry 2012). A recent report by the National Academies of Sciences emphasized diagnostic error and proposed recommendations to reduce these adverse events (National Academies of Sciences 2015). In 2000, Svensson et al. (2000) had already addressed these problems and proposed many of the

20-02-2018 12:33:22

244

same recommendations for distal radius fracture treatment. Incorrect assessment of fracture (re)displacement is a common error that can be prevented by increasing the physician’s clinical expertise and following treatment recommendations stated in current care guidelines (Wyrick 2016). Checklists have also been shown to be promising interventions to reduce diagnostic errors in emergency room settings (Graber et al. 2014). A checklist for distal radius fractures could include: (1) instructions on how to correctly measure the radiological parameters, (2) the radiological criteria for non-operative and operative treatment according to the guidelines, (3) common pitfalls, (4) instructions for adequately consenting patients, especially patients older than 65 years with an unstable fracture, for whom surgical fixation is not recommended. Computer-aided diagnostic systems (CADx) for the characterization of distal radius fractures might also offer interesting new opportunities to help physicians interpret radiographs and accordingly reduce diagnostic errors in distal radius fracture treatment. Over half (56%) of adverse events in non-operative treatment occurred during early follow-up visits. This is important, because in recent years the rationalization of follow-up visits has been a trend in orthopedic outpatient clinics in Finland (Ovaska et al. 2016). Based on this finding the follow-up visit, especially of non-operatively treated distal radius fracture patients, may not be an ideal visit type to be reduced when the productivity of outpatient clinics is optimized. There are several limitations to our study to consider when results are interpreted. Studies of patient injuries have been considered to be vulnerable to a few sources of bias, including non-standardized sources of data, selection bias, and hindsight bias. In Finland, the PIC claims register constitutes a unique nationwide database that covers all claims filed by patients from public and private healthcare sectors. It contains all necessary information about care of the claimants, including medical records, radiographs, MRI and CT files, and laboratory results. Therefore, data collection, handling, and recording are optimally standardized. As the claim-filing process is patient derived and many factors affect the likelihood that a patient will file a claim after an adverse event (Bismark et al. 2006), there is potential selection bias in this study. According to some estimates, only 1–3% of all patients with a severe, compensable adverse event ever file a claim (Mikkonen 2004, Bismark et al. 2006). Thus the PIC claims register represents only a portion of all adverse events that qualify as patient injuries and any estimates of incidence based on reported claims must be assessed with caution. Furthermore, our data comprised only compensated patient injury claims, as we did not review denied claims. Numbers of noncompensated and total claims should thus be used with caution. For the risk analysis, we estimated the total number of fractures of the distal radius using the incidence reported by Flinkkilä et al. (2011). An accurate number of distal radius fractures nationwide is difficult to obtain, because these frac-

11831 Sandelin D.indd 244

Acta Orthopaedica 2018; 89 (2): 240–245

tures are treated at a variety of health facilities and there is no national register where the total number of all distal radius fractures is reported. However, the incidence for fractures of the distal radius reported in a large, Swedish registry-based study (Wilcke at al. 2013) is very similar to the incidence reported by Flinkkilä et al. (2011), upon which we based our estimate. Hindsight bias is a phenomenon in which the individuals evaluating claims are more eager to grant compensation when the consequences of the adverse event are more severe regardless of the quality of the treatment provided. One might assume that this kind of bias would be less likely in a system which functions based on the “preventability rule,” where only the quality of the given treatment is assessed. Furthermore, the external medical advisors’ assessments of the claimants’ treatment are subjective. However, the assessments, based on the then prevailing national current care guidelines, are made by experienced orthopedic surgeons or hand surgeons, with access to all patient records and radiographs. The strengths of this study include the nationwide study design and the thorough review of each claim, the original patient records, and radiological images. Furthermore, although a selective group of patients, the compensated claimants represent the general distal radius fracture population regarding age and sex. In summary, we describe here the typical avoidable patient injuries related to distal radius fracture treatment. We also categorize the types and the typical causes of these severe adverse events. The adverse events fall into 3 main groups: diagnostic errors, decision/planning errors, and technical errors. This study will hopefully help physicians to recognize the critical steps in the treatment of this common fracture, enhance patient safety, and diminish adverse events.

We thank Reima Palonen from the Finnish Ministry of Social Affairs and Health and Saija Lehtinen from the Patient Insurance Center for their assistance and help.

HS designed the study, collected the data, analyzed the data, reviewed the literature, and prepared the manuscript. TH and EW designed the study, analyzed the data and prepared the manuscript. EH, JV and TR contributed to study design, interpreting the results, and preparation of the manuscript. HH performed the statistical analyses.

Acta thanks Leif Hove and other anonymous reviewers for help with peer review of this study.

Arora R, Lutz M, Deml C, Krappinger D, Haug L, Gabl M. A prospective randomized trial comparing nonoperative treatment with volar locking plate fixation for displaced and unstable distal radial fractures in patients sixtyfive years of age and older. J Bone Joint Surg Am 2011; 93 (23): 2146-53. Bismark M M, Brennan T A, Davis P B, Studdert D M. Claiming behaviour in a no-fault system of medical injury: A descriptive analysis of claimants and non-claimants. Med J Aust 2006; 185 (4): 203-7.

20-02-2018 12:33:23

Acta Orthopaedica 2018; 89 (2): 240–245

Brown T W, McCarthy M L, Kelen G D, Levy F. An epidemiologic study of closed emergency department malpractice claims in a national database of physician malpractice insurers. Acad Emerg Med 2010: 17 (5): 553-60. Chung K C, Spilson S V. The frequency and epidemiology of hand and forearm fractures in the United States. J Hand Surg Am 2001; 26 (5): 908-15. Chung K C, Shauver M J, Yin H. The relationship between ASSH membership and the treatment of distal radius fracture in the United States Medicare population. J Hand Surg Am 2011; 36 (8): 1288-93. Costa M L, Achten J, Parsons N R, Rangan A, Griffin D, Tubeuf S, Lamb S E; DRAFFT Study Group. Percutaneous fixation with Kirschner wires versus volar locking plate fixation in adults with dorsally displaced fracture of distal radius: Randomised controlled trial. BMJ 2014; 349: g4807. Croskerry P. Perspectives on diagnostic failure and patient safety. Healthc Q 2012; 15 Spec No: 50-6. DeNoble P H, Marshall A C, Barron O A, Catalano L W 3rd, Glickel S Z. Malpractice in distal radius fracture management: An analysis of closed claims. J Hand Surg Am 2014; 39 (8): 1480-8. Distal radius fracture (online). Current Care Guidelines. Working group set up by the Finnish Medical Society Duodecim and the Finnish Orthopaedic Association. Helsinki: Finnish Medical Society Duodecim, 2016. Available from www.kaypahoito.fi. Egol K A, Walsh M, Romo-Cardoso S, Dorsky S, Paksima N. Distal radial fractures in the elderly: Operative compared with nonoperative treatment. J Bone Joint Surg Am 2010; 92 (9): 1851-7. Flinkkilä T, Sirnio K, Hippi M, Hartonen S, Ruuhela R, Ohtonen P, Hyvönen P, Leppilahti J. Epidemiology and seasonal variation of distal radius fractures in Oulu, Finland. Osteoporos Int 2011; 22 (8): 2307-12. Friedman S L. Complications in orthopedics: Distal radius fractures. Pp. 93. Rosemont, IL: American Academy of Orthopedic Surgeons; 2005. ISBN 0-89203-367. Graber M L, Sorensen A V, Biswas J, Modi V, Wackett A, Johnson S, Lenfestey N, Meyer A N, Singh H. Developing checklists to prevent diagnostic error in emergency room settings. Diagnosis (Berl) 2014; 1 (3): 223-31. Guly H R. Diagnostic errors in an accident and emergency department. Emerg Med J 2001; 18 (4): 263-9. Helkamaa T, Hirvensalo E, Huhtala H, Remes V. Patient injuries in primary total hip replacement. Acta Orthop 2016; 87 (3): 209-17. Järvelin J. Studies on filed and compensated claims for patient injuries [dissertation]. University of Helsinki; 2012. Available from http://urn.fi/ ISBN:978-952-245-750-9. Khan I H, Giddins G. Analysis of NHSLA claims in hand and wrist surgery. J Hand Surg Eur Vol 2010; 35 (1): 61-4. Lichtman D M, Bindra R R, Boyer M I, Putnam M D, Ring D, Slutsky D J, Taras J S, Watters W C 3rd, Goldberg M J, Keith M, Turkelson C M, Wies J L, Haralson R H 3rd, Boyer K M, Hitchcock K, Raymond L. Treatment of distal radius fractures. J Am Acad Orthop Surg 2010; 18 (3): 180-9. Lutz K, Yeoh K M, MacDermid J C, Symonette C, Grewal R. Complications associated with operative versus nonsurgical treatment of distal radius fractures in patients aged 65 years and older. J Hand Surg Am 2014; 39 (7): 1280-6.

11831 Sandelin D.indd 245

245

Mahdavian Delavary B, Cremers J E, Ritt M J. Hand and wrist malpractice claims in The Netherlands: 1993–2008. J Hand Surg Eur Vol 2010; 35 (5): 381-4. Mathews A L, Chung K C. Management of complications of distal radius fractures. Hand Clin 2015; 31 (2): 205-15. Mikkonen M. Prevention of patient injuries: The Finnish patient insurance scheme. Med Law 2004; 23 (2): 251-7. McKay S D, MacDermid J C, Roth J H, Richards R S. Assessment of complications of distal radius fractures and development of a complication checklist. J Hand Surg Am 2001; 26 (5): 916-22. National Academies of Sciences, Engineering, and Medicine. Improving diagnosis in health care. Washington, DC: National Academies Press; 2015. Ovaska M T, Nuutinen T, Madanat R, Mäkinen T J, Söderlund T. The role of outpatient visit after operative treatment of ankle fractures. Injury 2016; 47 (11): 2575-78. Patient Injury Act. 25.7.1986/585. Available from www.finlex.fi/fi/laki/ajantasa/1986/19860585. Ring J, Talbot C L, Clough T M. Clinical negligence in foot and ankle surgery: A 17-year review of claims to the NHS Litigation Authority. Bone Joint J 2014; 96-B (11): 1510-14. Saber Tehrani A S, Lee H, Mathews S C, Shore A, Makary M A, Pronovost P J, Newman-Toker D E. 25-Year summary of US malpractice claims for diagnostic errors 1986–2010: An analysis from the National Practitioner Data Bank. BMJ Qual Saf 2013; 22 (8): 672-80. Singh H, Graber M L. Improving diagnosis in health care: The next imperative for patient safety. N Engl J Med 2015; 373 (26): 2493-5. Statistics Finland. Available from http://pxnet2.stat.fi/PXWeb/pxweb/fi/StatFin/StatFin__ vrm__vaerak/073_vaerak_tau_109_fi.px/. Statistics of Finnish Patient Insurance Center, 2014. Available from www.pvk. fi/en/Statistics/. Svensson O, Ahrengart L, Ekholm C, Andersson G L, Höglund M, Jonsson U, Juhlin L, Kopylov P, Lagerström C, Lundborg G, Made C, Mallmin H, Räf L, Törnkvist H. Malpractice in connection with radius fractures must be reduced: Clear guidelines for treatment and follow-up are required. Lakartidningen 2000; 97 (15): 1800-4, 1807-9. Talbot C L, Ring J, Holt E M. Litigation relating to conditions affecting the shoulder and elbow: An analysis of claims against the National Health Service. Bone Joint J 2014; 96-B (5): 574-9. Walenkamp M M, Mulders M A, Goslings J C, Westert G P, Schep N W. Analysis of variation in the surgical treatment of patients with distal radial fractures in the Netherlands. J Hand Surg Eur Vol 2016; 42 (1): 39-44. Wilcke M K, Hammarberg H, Adolphson P Y. Epidemiology and changed surgical treatment methods for fractures of the distal radius: A registry analysis of 42,583 patients in Stockholm County, Sweden, 2004–2010. Acta Orthop 2013; 84 (3): 292-6. Wyrick J. Open reduction and internal fixation of a distal radius fracture with a volar locking plate: A case report. J Orthop Trauma 2016; e1-e4.

20-02-2018 12:33:23

246

Acta Orthopaedica 2018; 89 (2): 246–250

Avoidable 30-day mortality analysis and failure to rescue in dysvascular lower extremity amputees Implications for future treatment protocols Christian WIED 1, Nicolai B FOSS 2, Peter T TENGBERG 1, Gitte HOLM 1, Anders TROELSEN 1, Morten T KRISTENSEN 1,3

1 Department

of Orthopaedic Surgery, Copenhagen University Hospital Hvidovre; 2 Department of Anesthesiology and Intensive Care, Copenhagen University Hospital Hvidovre; 3 Physical Medicine and Rehabilitation Research-Copenhagen (PMR-C), Department of Physiotherapy, Copenhagen University Hospital Hvidovre, Denmark Correspondence: Chr.Wied@gmail.com Submitted 2017-11-23. Accepted 2017-12-12.

Background and purpose — An enhanced treatment program may decrease 30-day mortality below 20% after lower extremity amputations (LEA). The potential and limitations for further reduction are unknown. We analyzed postoperative causes of 30-day mortality, and assessed failure to rescue (FTR) rate in LEA patients who followed an enhanced treatment program. Patients and methods — Medical charts of 195 primary LEA procedures were reviewed independently by 3 of the authors, and deaths during hospitalization following amputation were classified according to consensus. Results — 31 patients died within 30 days after surgery. 4 deaths were classified as “definitely unavoidable,” 4 as “probably unavoidable,” and 23 as “FTR.” Patients who died had a higher incidence of sepsis, pneumonia, and acute myocardial infarction compared with those alive. A log binominal regression analysis adjusted for age, sex, ASA score, diabetes, nursing home admission, transfemoral amputation (TFA), and BMI showed that the risk of 30-day mortality was increased for TFA (RR = 2.3, 95% CI 1.1–4.8) and for patients with diabetes (RR = 2.7, 95% CI 1.3–5.6). The FTR rate (patients with 30-day mortality/all patients with a severe postoperative complication) was 30%. Of the FTR deaths, 20 at some point had active lifesaving care curtailed. Interpretation — Future initiatives should be directed at enhanced sepsis and pneumonia prophylactic actions, in addition to close monitoring of hemodynamics in anemic patients, with the potential to further reduce morbidity and mortality rates. ■

30-day mortality rates in excess of 30% have been reported in patients following a major dysvascular lower extremity

amputation (LEA) (Kristensen et al. 2012). Beyond atherosclerosis or diabetes, treatment is most often challenged by several competing co-morbidities and high age (Kristensen et al. 2012, Wied et al. 2016). Perioperative optimization may reduce morbidity and mortality. Thus a recent study reported decreased mortality rates following an enhanced in-hospital treatment program (Kristensen et al. 2016), but the potential and limitations of a further reduction in mortality are unknown. The mortality rates are frequently used to compare the quality of treatment between hospitals. However, this has recently been challenged since some deaths in hospital are inevitable. As stated by Silber et al. (1992, 2007), the mortality is related to the degree of illness and co-morbidity of patients receiving treatment and not necessarily the expression of differences in the quality of care. Failure to rescue (FTR)—the probability of death if experiencing a severe postoperative complication—is becoming increasingly popular as an indicator showing how well hospitals perform once the complications occur (Silber et al. 1992, 2007). The aim of this study is to analyze causes of 30-day mortality after LEA procedures and the FTR rate. We hypothesize that some of the 30-day deaths could be classified as FTR cases.

Patients and methods Hvidovre University Hospital Copenhagen has a catchment area of around 600,000 people, and performs all non-traumatic amputations in the area resulting in approximately 100 major dysvascular LEAs per year. The study is a single-center

11546 Wied D.indd 246

20-02-2018 12:26:00

Acta Orthopaedica 2018; 89 (2): 246–250

retrospective cohort study of patients with a primary LEA admitted between January 2013 and April 2015. All patients with primary LEA were assessed for inclusion if the limb was amputated because of arteriosclerosis or diabetic complications. Exclusion criteria were bilateral amputation procedures and re-amputations. All LEA patients follow a well-defined enhanced rehabilitation program, in an acute orthopedic ward (Kristensen et al. 2016). Preoperative management Patients referred from general practitioners, outpatient clinics, or emergency departments have their ankle-brachial blood pressure index measured, and their medical record is submitted to the vascular surgeons for the possibility for revascularization. If revascularization is deemed unobtainable by the Department of Vascular Surgery a supplementary measurement of skin perfusion pressure is performed to aid the decision on the amputation level. Senior consultants then review the indication for amputation. The patients are treated upon arrival at our department according to standardized fluid and transfusion protocols. The patients receive prophylactic antibiotics (dicloxacillin intravenously) before surgery and lowdose low-molecular-weight heparin after surgery. Intraoperative management The surgery is performed by trained residents or senior consultants. All transtibial amputations (TTA) procedures are performed approximately 12 cm below the knee joint, with sagittal flaps ad modum Persson (Persson 1974). The transfemoral amputation (TFA) procedure is performed with standard anterior and posterior skin flaps approximately 10 cm above the knee joint. Through-knee exarticulation (TKE) is rarely performed. The tissue vitality is assessed regularly during the operation, and the patient is informed of the risk of amputation at a more proximal level if tissue intraoperatively is deemed not to be vital. Spinal or general anesthesia is used during surgery. An ischiatic catheter with a continued infusion of (2 mg/ mL) ropivacaine at a rate of (4 mL/h) with a possible bolus of 5 mL and 30 min lockout time is placed during the TFA. A peripheral nerve catheter with continuous infusion is applied after surgery for TTA to provide extended analgesia for the first 4 postoperative days. Postoperative management Postoperatively the oral intake of fluids is supplemented with 1.000 mL standard rehydration fluid (isotonic Na-K-glucose or Ringer’s lactate) administered intravenously. Fluid balance is measured from daily body weight if possible, while hemoglobin and blood electrolytes and creatinine are measured preoperatively and until the fourth postoperative day. Hypovolemic patients are rehydrated with fluid at 20 mL/kg body weight. The unit uses a liberal blood transfusion trigger of 6 mmol/L for the first 4 days postoperatively.

11546 Wied D.indd 247

247

Patients are mobilized as soon as possible after surgery. Physiotherapy is started on postoperative day 1 and continued for 2–5 days during weekdays (for most patients, on a daily basis) until discharge (Kristensen et al. 2016). Data collection and management 3 of the authors reviewed the medical charts independently, and deaths within 30 days from amputation or during primary hospitalization were classified according to consensus, as: “Definitely unavoidable” being a result of either pre-amputation terminal disease or the patient refusing relevant postoperative care such as nutrition, fluids, and rehabilitation, or “probably unavoidable” being a result of pre-amputation acute life-threatening medical illness with predictable short life expectancy (< 1 month) (Foss and Kehlet 2005). The remaining deaths classified as FTR were defined as those that were related to possible avoidable postoperative complications. We noted whether the patients who died had a notation in their charts on restrictions in the level of active therapy (Foss and Kehlet 2005). Also, the triggering complication of death, based on the medical chart review, was registered. The FTR rate is calculated as patients with 30-day mortality/all patients with a severe postoperative complication (Henneman et al. 2013). The patients classified with “definitively” or “probably” unavoidable deaths were excluded from the FTR calculation. Patients who developed 1 of the following 8 complications while hospitalized were included in the FTR calculation: radiographically verified pneumonia or respiratory failure, sepsis, acute renal failure (postoperative creatinine > 200 µmol/L), stroke, gastrointestinal complication (ileus, hemorrhage), acute myocardial infraction, requirement for more than 3 blood transfusions within 72 hours from surgery, and re-amputation at a higher level within 30 days from index amputation. We were cautious not to register preexisting diseases as postoperative complications. The findings were double-checked by 2 independent researchers. Statistics Continuous data are presented as median values with interquartile ranges (IQRs) or mean values with standard deviations (SD). Categorical data are presented as numbers and were compared using the chi-square test or Fisher’s exact test in cases with cell counts of 5 or less. Log binominal regression analysis, with 30-day mortality as dependent variable, was used to assess the relative risk of variables with known influence on 30-day mortality (Barros and Hirakata 2003, Kristensen et al. 2012, Karam et al. 2013). The selected variables were: age, sex, ASA score, diabetes, nursing home admission, TFA, and BMI. A corresponding model evaluated the 30-day mortality of FTR patients with active care curtailed, compared with those who survived. Log binominal regression analysis was used due to the relatively few events, thereby eliminating the risk of an overestimation of the estimates from simple multivariable logistic regression analysis (Barros and Hirakata

20-02-2018 12:26:00

248

Acta Orthopaedica 2018; 89 (2): 246–250

Amputations assessed for eligibility (n = 243): – TTA, 86 – TKE, 5 – TFA, 150 – hip ex-articulation, 2 Excluded (n = 48): – bilateral amputations, 12 – re-amputations, 36 (7 of whom died within 30 days of their index amputation) re-amputations were conversions from: TTA, 26 TKE, 4 TFA, 6 Patients included in the study (n = 195) 31 patients with 30-day perioperative death

Figure 1. Flow chart of amputee patients at Hvidovre Hospital, from January 2013 to April 2015. TTA = transtibial amputation, TKE = through knee ex-articulation, TFA = transfemoral amputation.

2003). To our knowledge our study is the first analyzing mortality causality in this group of patients, which is why we were not able to assess variability or distribution before the data collection. Ethics, funding, and potential conflicts of interest The study was approved by the local ethics committee and registered with the regional data protection agency (04.12.2012) (j. no. 01975 HVH-2012-053). No benefits in any form have been received or will be received from any commercial party related directly or indirectly to the subject of this article. No competing interests were declared.

Table 1. Patient characteristics, n = 195. Values are number of patients unless otherwise stated

Factor Male sex Female sex Age in years, mean (SD) Own home Nursing home Body mass index, mean (SD) New mobility score, mean (SD) Diabetes type I or II Dementia ASA score > 2 Transtibial amputation Through-knee amputation Transfemoral amputation Intraoperative blood loss (mL), median (IQR) General anesthesia an bn

Survivors n = 164

Non-survivors n = 31

90 74 74 (12) 110 54 24.6 (6.5) 3.9 (2.8) a 68 22 135 71 5 88 300 (150–500) 46

19 12 78 (13) 22 9 23.5 (6.8) 2.5 (3.2) b 18 2 29 9 0 22 400 (200–675) 9

= 140. = 22.

22 patients who died in the TFA sub-group received statistically significantly more blood transfusions during the period from 2 days before surgery and until the fifth postoperative day [median of 3.0 units vs. 2.0 units (p = 0.02)] compared with those 88 patients alive in the TFA group. Estimation of FTR 4 deaths were classified as “definitely unavoidable,” 4 as “probably unavoidable,” and 23 as FTR, of which 20 patients at some point had active care curtailed (Table 3). 53 patients experienced at least 1 of the 8 defined complications during their stay at the hospital. The FTR rate was 30% (postoperative complications among all 195 patients may be found in Table 4, see Supplementary data).

Results 195 consecutive patients with a single primary major dysvascular LEA procedure were included in the study between January 2013 and April 2015 (Figure 1 and Table 1). Causes of death 31 of the 195 patients died in hospital within 30 days after surgery (median of 6 postoperative days (IQR: 4.5–10)): 22 post-TFA procedures and 9 post-TTA procedures (Table 1 and see Supplementary data). There was a statistically significant higher incidence of postoperative sepsis, pneumonia, and acute myocardial infarction among the patients who died compared with the survivors in univariable analysis. In an adjusted log binominal regression analysis TFA and diabetes were associated with an increased risk of 30-day mortality (TFA: RR = 2.3, CI = 1.1–4.8, p = 0.03 and diabetes: RR = 2.7, CI = 1.3–5.6, p = 0.01) (Table 2, see Supplementary data). The

11546 Wied D.indd 248

Discussion The 16% 30-day mortality rate in this study is comparable with previous European studies (Moxey et al. 2010) and it shows how severely ill patients are when admitted to the department of orthopedic surgery for a dysvascular-related primary LEA. 84% of the patients had an ASA score of 3–4, most likely the highest within the orthopedic specialty. Patients with diabetes or patients having TFA surgery performed all had an elevated risk of 30-day mortality. One-quarter of the deaths were highly expected in patients suffering preoperatively from the consequences of septic shock, severe sequelae after recent massive stroke, or unrecoverable respiratory failure. Nevertheless, they had surgery scheduled and performed. This is remarkable and leaves an impression of how selected patients could have a more digni-

20-02-2018 12:26:00

Acta Orthopaedica 2018; 89 (2): 246–250

249

Table 3. Classification of patients according to the potential for avoiding death. Values are number of patients unless otherwise stated

Factor Active care curtailed, no/yes Estimated cause of death: Stroke Acute myocardial infarction Pneumonia or respiratory failure Sepsis Renal failure Gastrointestinal ileus or hemorrhage Postop. days to death, median (IQR) a 22

Definitely unavoidable n=4

Probably unavoidable n=4

Failure to rescue n = 23

Total n = 31 a

0/4

1/3

3/20

4/27

0 0 2 2 0 0 7 (2.0–9.8)

1 0 0 2 0 1 5 (2.8–6.5)

3 6 5 4 3 2 7 (4.0–11)

4 6 7 8 3 3 6 (4.5–10)

transfemoral and 9 transtibial amputations.

fied death than is the case today if more attention were paid to the possibility of non-surgical palliative treatment. Thus, for some patients, limb ischemia/severe infections along with serious co-morbidities are part of the death process, and a nonsurgical approach might be more ethically correct. However, we acknowledge that an algorithm on how to select patients for non-surgical palliative treatment is difficult. If the mortality rate from postoperative acute myocardial infraction is to be reduced, it seems critical that hypoperfusion is avoided (Dhalla et al. 2010). Clinical estimation alone to guide blood transfusion is probably inadequate (Ram et al. 2014). Daily Hgb measurement is to be recommended, and a close regulation of LEA patients’ hemodynamics before and after surgery is important. Even with a perioperative standard for fluid therapy and antibiotic during surgery, sepsis and pneumonia remain significant challenges. It seems reasonable to consider extending antibiotic prophylaxis into the early postoperative days. Issues such as tissue hypoxia due to surgical stress could be insufficiently monitored on regular wards by blood pressure, heart rate, and saturation only. Different invasive and non-invasive techniques provide an enhanced assessment of optimal oxygen delivery, and individualized goal-directed hemodynamic therapy (Sobol and Wunsch 2011) is perhaps something that needs to be included more often in LEA surgery in the future. Since our study, to our knowledge, is the first to apply the FTR approach in LEA surgery, it is compared with hip fracture patients with similar demographics. Our patients had an exceptionally high FTR rate. Thus, a US register study on patients with hip fracture published in 2015 with a large patient population reported an FTR rate of 6% (Menendez and Ring 2015). The difference may be explained by different comorbidity patterns. The fact that 20 of the 23 patients classified as FTR deaths received less than maximum postoperative care gives cause for concern. Doctors curtailed active care on ethical grounds considering the current health status, mental status, and the pre-amputation function level. The decision to curtail active

11546 Wied D.indd 249

care was taken by the attending medical staff, and not influenced by the authors of this study. Going through the medical records it became evident how the decision not to refer to ICU or active revival was settled. Frequently it was through a conversation between the attending resident and a consultant in the situation of a quickly deteriorating patient. This is stressful and not the ideal time for such discussions. There are several limitations to this study. The percentage of patients classified as FTR could be too high, and the investigators could have overlooked some complications among the survivors. This could influence the FTR rate and to some extent explain the exceptionally high percentage. The usage of FTR is relatively new and what diagnoses should be included in the calculation is still being debated. A recent study evaluating the concept of FTR found that administrative data are inaccurate in the general hospital population and the FTR rate ought not to be used in its current state for comparative purposes (Horwitz et al. 2007). Another limitation is the retrospective classification of cause of death. However, the final classification was based on the consensus of 3 specialists, including the one responsible for the corresponding hip fracture study (Foss and Kehlet 2005). The data represent the biggest dataset on the subject and provide a background for further exploratory analysis of the subject. The data are partly qualitative, highlighting the fact that there is both a group where care is probably futile as well as a group where postoperative care is probably inadequate. Ideally a future study on this question should be prospective, with independent assessments preoperatively and postoperatively, preferably daily, but at least weekly. From our work and analysis process it became clear that future trials will gain considerably from the recommendations listed in the Appendix. In summary, the 30-day mortality within this study is comparable to previous studies and shows the potential in continuous research in perioperative strategies. Diabetic patients and patients with TFA have an elevated risk of 30-day mortality. Postoperative sepsis, pneumonia, and acute myocardial infarction are the main complications leading to death following

20-02-2018 12:26:01

250

LEA surgery. It seems apposite to recommend a future intensified focus on the perioperative optimization of LEA patients, and with room for improvements. Supplementary data The Appendix and Tables 2 and 4 are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674.2018.1430420

GH, NBF, and CWI collected the material. CWI, NBF, GH, PTT, AT, and MTK analyzed the data. CWI wrote the first draft and took care of revisions. CWI, PTT, GH, AT, NBF, and MTK contributed to the planning of the study, interpretation of the results, and preparation of the manuscript. The study has not been published and the manuscript is not being considered for publication elsewhere. None of the above-mentioned authors have any conflicts of interest directly related to this study.

Acta thanks Jan Larsson and other anonymous reviewers for help with peer review of this study.

Barros A J D, Hirakata V A. Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio. BMC Med Res Methodol 2003; 321–34. Dhalla N S, Adameova A, Kaur M. Role of catecholamine oxidation in sudden cardiac death. Fundam Clin Pharmacol 2010; 24:5 39–46. Foss N B, Kehlet H. Mortality analysis in hip fracture patients: implications for design of future outcome trials. Br J Anaesth 2005; 94(1): 24–9. Henneman D, Snijders H S, Fiocco M, Leersum N J, Kolfschoten N E, Wiggers T, Wouters M W J M, Tollenaar R A E M. Hospital variation in failure to rescue after colorectal cancer surgery: results of the Dutch Surgical Colorectal Audit. Ann Surg Oncol 2013; 20: 2117–23.

11546 Wied D.indd 250

Acta Orthopaedica 2018; 89 (2): 246–250

Horwitz L I, Cuny J F, Cerese J, Krumholz H M. Failure to rescue: validation of an algorithm using administrative data. Med Care 2007; 45(4): 283–7. Karam J, Shepard A, Rubinfeld I. Predictors of operative mortality following major lower extremity amputations using the National Surgical Quality Improvement Program public use data. J Vasc Surg 2013; 58: 1276–82. Kristensen M T, Holm G, Kirketerp-Møller K, Krasheninnikoff M, Gebuhr P. Very low survival rates after non-traumatic lower limb amputation in a consecutive series: what to do? Interact Cardiovasc Thorac Surg 2012; 14(5): 543–7. Kristensen M T, Holm G, Krasheninnikoff M, Jensen P S, Gebuhr P. An enhanced treatment program with markedly reduced mortality after a transtibial or higher non-traumatic lower extremity amputation. Acta Orthop 2016; 87: 1–6. Menendez M E, Ring D. Failure to rescue after proximal femur fracture surgery. J Orthop Trauma 2015; 29(3): 96–102. Moxey P W, Hofman D, Hinchliffe R J, Jones K, Thompson M M, Holt P J E. Epidemiological study of lower limb amputation in England between 2003 and 2008. Br J Surg 2010; 97(9): 1348–53. Persson B M. Sagittal incision for below-knee amputation in ischaemic gangrene. J Bone Joint Surg (Br) 1974; 56: 110–14. Ram G G, Suresh P, Vijayaraghavan P V. Surgeons often underestimate the amount of blood loss in replacement surgeries. Chin J Traumatol 2014; 17(4): 225–8. Silber J H, Williams S V, Krakauer H Schwartz J S. Hospital and patient characteristics associated with death after surgery: a study of adverse occurrence and failure to rescue. Med Care 1992; 30(7): 615–29. Silber J H, Romano P S, Rosen A K, Wang Y, Even-Shoshan O, Volpp K G. Failure-to-rescue: comparing definitions to measure quality of care. Med Care 2007; 45(10): 918–25. Sobol J B, Wunsch H. Triage of high-risk surgical patients for intensive care. Crit Care 2011; 15(2): 217. Wied C, Foss N B, Kristensen M T, Holm G, Kallemose T, Troelsen A. Surgical Apgar score predicts early complication in transfemoral amputees: retrospective study of 170 major amputations. World J Orthop 2016; 7(12): 832–8.

20-02-2018 12:26:01

Acta Orthopaedica 2018; 89 (2): 251

251

Correspondence

Highlighting the results of a trial by using appropriate inferential statistics Commentary on “Specific exercises for subacromial pain” by Björnsson Hallgren et al. 2017

Sir,—Subacromial pain is a very common cause of visits to an orthopedic clinic. We have read with interest the paper “Specific exercises for subacromial pain” by Björnsson-Hallgren et al. The question whether an exercise program specifically developed to treat subacromial pain is superior to a regular training regime is important. Recognizing the strengths of the study, we would like to discuss some issues concerning the approach used by the authors when reporting the results. The main results were reported as absolute numbers of exposed and unexposed patients in each group along with p-values. The exposure was a decision to agree to a surgical procedure at the end of follow-up. Such an approach may create a problem: absolute estimates describe a particular single sample but say nothing about the entire population of interest. Based on the reported results, we discover that 14 of 47 patients in the intervention group and 28 of 44 controls ended in a decision to undergo surgery. Additionally, we find out that the difference between these two particular groups was statically significant as shown by the p-value < 0.05. While such statistics are adequate to inform readers about certain differences between groups, several important questions remain unanswered: • How substantial was the difference between groups—how much higher risk did controls have regarding the surgery decision? • How wide could the possible boundaries of observed difference be if different random samples containing different individuals were drawn from the population of patients with subacromial pain? • How many patients have to be treated by a specific training program in order to avoid at least one surgery decision? The first question can be answered by calculating, e.g., a relative risk ratio (RR), the second by drawing a confidence interval (95% CI), and the third question can be answered by estimating a number needed to treat (NNT). These statistics can be calculated in Excel, or by the majority of statistical packages available on the market, or even by hand. Thus, the following statistics may be added to the report: • RR = 0.47 (meaning that the intervention group had around half the risk of ending up with a decision for surgery when compared with controls); • p-value of RR = 0.0025 (the difference between the two groups regarding the risk was statistically significant);

• 95% CI for RR = 0.27–0.77 (the risk ratio may vary within these limits if different patients are drawn from the population); • NNT = 3.0 (at least 3 patients need to be included in a specific exercise group in order to avoid 1 decision for surgery); • 95% CI for NNT is 1.9–6.9 (the NNT will probably fall into these limits if different patients are drawn from the population). In this way, the report may come to be more robust and precise, emphasizing substantially the conclusion made by the authors. We encourage researchers to report statistics that describe a population (“inferential statistics”) whenever possible. Mikhail Saltychev 1 and Petri Virolainen 2 Departments of 1 Physical and Rehabilitation Medicine and 2 Orthopedics, Turku University Hospital and University of Turku, Turku, Finland Email: mikhail.saltychev@gmail.com

Sir,—We thank M Saltychev and P Virolainen for their interest in our work. We appreciate their approach to the research question, which may provide an additional viewpoint and further demonstrates the differences between the study groups. We have controlled our data according to number needed to treat (NNT) and relative risk (RR) and confirmed that their calculations are accurate. These statistical results allow for a slightly different perspective on our results, which may be even clearer in presenting our results for the majority of readers treating subacromial pain patients. They further emphasize the importance of a specific exercise program and may very well be added to our results. We found their comments very helpful and valuable for our future research. Hanna Björnsson Hallgren, corresponding author Email: hanna.bjornsson.hallgren@regionostergotland.se Björnsson Hallgren H C, Adolfsson L E, Johansson K, Öberg B, Peterson A, Holmgren T M. Specific exercises for subacromial pain: Good results maintained for 5 years. Acta Orthop 2017; 88 (6): 600-605.

19-12539 Saltychev D.indd 251

20-02-2018 12:28:05

252

Acta Orthopaedica 2018; 89 (2): 252

Erratum

1 Valdoltra

Table 1 had some missing zeros and therefore some subgroups (Non-cemented and Ti-alloy) did not add up. A corrected version is presented below:

Table 1. Number of patients in sub-cohorts (bearing surface between head and cup, cementation, material of the stem) by age at operation Group Subgroup All Bearing MoM MoP CoC CoP Use of bone cement Non-cemented Cemented Stem material Ti-alloy CoCr-alloy Fe-alloy

Age group 50–59 60–69

70–79

≥ 80

Total

591

1,631

3,059

2,664

267

8,343

12 39 69 11

43 233 274 41

141 763 535 192

124 2,128 407 400

18 2,481 38 127

0 265 0 2

338 5,909 1,323 773

119 12

580 11

1,596 35

2,910 149

1,713 951

48 219

6,966 1,377

122 9 0

585 5 1

1,610 21 0

2,980 67 12

1,940 586 138

79 162 26

7,316 850 177

0–40

40–49

131

12041erratum D.indd 252

20-02-2018 18:07:07

Acta Orthopaedica 2018; 89

253

Information to authors Open Access/ Article Publication Charge Acta Orthopaedica, owned by the Nordic Orthopaedic Federation, is a non-profit, immediate Open Access journal, meaning your work is free for every-one to access online as soon as it is published (Gold OA). The standard research article publishing charge (APC) is EURO 1,750 (USD 2,000/GBP 1,520). For Case Reports the APC is EURO 875 (USD 1,000/GBP 760). Depending on your location these charges may be subject to local taxes. Note: There is no APC for members in the national orthopedic associations who are paying members of the Nordic Orthopaedic Federation (Denmark, Estonia, Finland, Iceland, Lithuania, Norway, and Sweden) Submission of manuscripts Manuscripts should be submitted online, www.actaorthop.org. Authors are encouraged to provide the names, addresses, and e-mail of potential reviewers. Note the following points: a. The manuscript should be prepared according to Uniform Requirements for Manuscripts submitted to Biomedical Journals (www.icmje.org/). American spelling is preferred. A manuscripts total word count should not exceed 3,300 (including Abstract (max. 250 words) but excluding title page, figure and table legends, and references). The word count should be added on the title page. b. Authors submitting a paper do so on the understanding that it has not been published and is not being considered for publication elsewhere. The authors should provide a statement about previous submissions and reports that might be regarded as duplicate publication of the same or a similar paper. Copies of that study should be submitted with the paper. c. Authors should provide a description of their individual contributions to the study. d. Interventional studies (defined by human participants being prospectively assigned to one or more health-related interventions to evaluate the effects on health outcomes) should be registered in a public trials register, and the registration ID should be presented under ‘Ethics’. In addition, manuscripts presenting randomized trials should comply with the CONSORT statement (www.consort-statement.org). The study protocol, approved by the appropriate ethics committee, a completed CONSORT checklist and flowchart should be submitted together with the manuscript. Animal in vivo experiments should comply with ARRIVE and a completed checklist (www.nvcr3s.org.uk/arriveguidelines) and a copy of the study protocol should be submitted together with the manuscript. e. Systematic reviews and meta-analyses of interventional studies should comply with the PRISMA statement (www.prisma-statement.org), and the PRISMA checklist and flowchart should be appended to the manuscript. For

systematic reviews and meta-analyses of observational studies the MOOSE checklist and flowchart (www.equatornetwork.org) should be used. f. Manuscripts presenting results on diagnostic accuracy should comply with the STARD-statement (www.stardstatement.org), and the STARD checklist and flowchart should be appended to the manuscript. g. Reports of observational studies should comply with the STROBE-statement (www.strobe-statement.org) and the STROBE checklist should be appended to the manuscript. h. Extensive tables, complementary figures, and other explanatory material should be given as Supplementary data which will be available in the online version of the article. i. All documents with original data of relevance to the submitted manuscript should be stored and be retrievable on request for a minimum of 10 years. Authors are encouraged to include a copy of raw data in electronic form and/or make these data available via the internet. j. Acta does not publish studies that only assess completeness, coverage, accuracy, and validity of data in registers. Such studies are considered part of normal register administration, and the results should be presented in the register’s annual report. k. Acta does not publish studies reporting merely the existence or organization of a new register. Furthermore, to qualify for publication an epidemiologic/register study must include a question assessed by use of the register data, for example: outcome related to different treatments, treatment changes over time, or incidence changes over time. Manuscript layout Title. The title should include information on the results of the investigation, including the number of patients, average follow-up, and animal or cadaver experiments. The first name, middle initial(s), and last name of each author should be given with department affiliations as well as the name and e-mail of the corresponding author. Abstract. The abstract should not exceed 250 words. It should consist of 4 parts: Background and purpose, Patients/ material/animals and methods, Results, and Interpretation. Introduction. The introduction should focus on the state of knowledge at the beginning of the study. The aims and main hypothesis of the study should be stated clearly. Only in exceptional cases should it exceed one typewritten page. Case reports. These should be short, include the Case history and Discussion without an Introduction. Patients/material/animals and methods. The selection of subjects and the inclusion or exclusion criteria should be described. Subjects who declined to participate, withdrawals and those with incomplete follow-up should be accounted for.

254

Describe in detail how the measurements were made and the techniques used. Statistics. All statistical methods should be mentioned. Unusual methods should be referenced. The tests should be two-sided, unless the use of one-sided tests is justified. No data should be removed, imputed, weighted, adjusted or trimmed unless they are specifically described and justified and its consequences are given. Hypothesis tests (p-values). These tests should be used, in combination with a defined effect size and when statistical power has been considered. Give p-values with real numbers, if these are greater than 0.001 (1 digit except zeros), otherwise use ‘p < 0.001’. Do not use ‘ns’, ‘p > 0.05’ or asterisks. In most instances, the 95% confidence intervals should be given, especially in exploratory analyses and when estimating the effects or differences. Ethics, registration, funding, and potential conflicts of interest. Approval by the local ethics committee should be stated with date of issue and registration number. Funding and potential conflicts of interest should be declared. Results. Do not give the same data in more than one way. When summarizing the data, always include measures of variability and the number of subjects. Give the median and range—e.g., 60 (35–70) years, the mean and standard deviation—59 (SD 15) years, and the frequencies for nominal data. The results of matched data should be given in relevant form (e.g., the distribution of pairwise differences). Percentages should not be used if the sample size is less than 100. Discussion. This section should contain a critical discussion of the results—e.g., the quality of the data (selection and information bias) and adequacy of the statistical analysis (confounding bias). It should also assess the relevant literature for or against the findings and if possible, the conclusions as regards clinical application or further research. Discuss, do not recapitulate, your results. Tables. Each table should be self-explanatory, with an adequate title and a logical presentation of the data. Abbreviate words in the columns and explain in footnotes. Each column heading for numerical data should include the unit of measurement. Use SI units. Avoid unnecessary decimals! It is seldom advisable to use more than 2 digits for biologic measurements. Digital illustrations should be used. A scanned picture or digital photo should usually be about 85 mm (i.e. 1000 pixels) wide with a resolution of 300 DPI (dots per inch), which gives a square picture an uncompressed file size of 4.5 MiB in color and 1 MiB in gray-scale. A black and white drawing or graph may be scanned in 800 DPI bitmap, i.e. 1 bit TIFF. The preferred format is a TIFF-file. The resolution of WEB illustrations (gif) is usually too poor. Color photos should be in CMYK colors. The cost of color will be paid by Acta. Highlights, arrows and letters may be added to digital photos but keep a clean version of the photo for the layout process. Acta prefers to add these as a separate layer.

Acta Orthopaedica 2018; 89

For digital graphs, use a graphics program that can export vectorized graphs as EMF, EPS or PDF files. Avoid frames around diagrams, diagrams with perspective drawing, and bar graphs or histograms (use tables). Symbols should be consistent throughout a series of figures. Each axis should be horizontally labeled, with a description of the variable it represents. Use sans serif letters (Arial or Helvetica). Make diagrams in black-and-white with gray or color areas, but avoid complex patterns. Author contributions. Describe in short what each author did (initials). Acknowledgments. Technical help or other help may be acknowledged. References. We prefer the references to be cited by name and year (chronologically) of publication in the text instead of giving their numbers, i.e. 1 author, 2 authors or author et al. and year. Thus, the list of references should be given in alphabetical order. In the reference list all authors should be included. References with same author and year should be separated by adding a letter to the year (a, b, c...) in the list and the text. Submission Create an Adobe Portable Document File (PDF) of your full manuscript. You may need the Adobe Acrobat Writer program or similar PDF creator program installed on your computer. We recommend that you write your manuscript as a Microsoft Word file and embed all images, tables, figures etc. within the one file. After inserting line numbers (see below) convert the Microsoft Word document to a PDF file by choosing the appropriate icon in the task bar of the Microsoft Word application. Alternatively, right hand click on the icon of your manuscript file and choose, convert to Adobe PDF. If you do not have the Adobe PDF writer installed on your computer, you can use the�“try it for free” feature found on the Adobe website at http://createpdf.adobe.com/ Important: Please insert line numbers in the Microsoft Word document before converting it to PDF by following the procedure below: • Windows 1. Switch to print layout view. 2. To add line numbers to part of a document, select the text you want to number. 3. On the File menu, click Page Setup, and then click the Layout tab. 4. Click Selected text in the Apply to box. 5. Click Line Numbers. 6. Select the Add line numbering check box, and then select the options you want. • Macintosh. Line number options is found by selecting Document layout –> Layout under the Format menu. By using the “Apply to”-option it is possible to select parts which will not be included in the feature.