Acta Orthopaedica Vol. 90 Issue 6, December 2019 by Acta

6/19 ACTA ORTHOPAEDICA

Element of success in joint replacement

www.heraeus-medical.com

Vol. 90, No. 6, 2019 (pp. 507–626)

Proven for 60 years in more than 30 million procedures worldwide. *OREDObOHDGHU LQ FOLQLFDO HYLGHQFH ZLWK PRUH WKDQ VWXGLHV 7KLV makes PALACOS® ERQH FHPHQW ZKDW LW LV 7KH JROG VWDQGDUG DPRQJ bone cements, and the element of success in joint replacement.

Volume 90, Number 6, December 2019

IORT_90_06_COVER.indd 1

17-10-2019 10:22:33

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

E DITORIAL O F FICE

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

EDITOR

THE FOUNDATION BOARD OF

Anders Rydholm Lund, Sweden

THE NORDIC O RTHOPAEDIC F EDERATION AND A CTA O RTHOPAEDICA

DEPUTY EDITOR

Peter A Frandsen Odense, Denmark CO-EDITORS

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

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

WEB EDITOR

Magnus Tägil Lund, Sweden S TATISTICAL EDITOR

Jonas Ranstam Lund, Sweden P RODUCTION MANAGER

Kaj Knutson Lund, Sweden

Vol. 90, No. 6, 2019

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

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

Annual Institutional Subscription, Volume 90, 2019

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

$1,218

£753

€976

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

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

Copyright © 2019 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. 90, No. 6, December 2019 Guest editorial The (r)evolution of hyped innovations in orthopedic implants: can prudent introduction avoid throwing the baby out with the bathwater? Elbow, hand Use and outcome of 1,220 primary total elbow arthroplasties from the Australian Orthopaedic Association National Joint Arthroplasty Replacement Registry 2008–2018 Collagenase injections for Dupuytren disease: 3-year treatment outcomes and predictors of recurrence in 89 hands Hip Survival and revision causes of hip resurfacing arthroplasty and the Mitch proximal epiphyseal replacement: results from the Danish Hip Arthroplasty Register Has the threshold for revision surgery for adverse reactions to metal debris changed in metal-on-metal hip arthroplasty patients? A cohort study of 239 patients using an adapted risk-stratification algorithm Posterior and anterior tilt increases the risk of failure after internal fixation of Garden I and II femoral neck fracture Similar outcome of femoral neck fractures treated with Pinloc or Hansson Pins: 1-year data from a multicenter randomized clinical study on 439 patients Higher risk of cam regrowth in adolescents undergoing arthroscopic femoroacetabular impingement correction: a retrospective comparison of 33 adolescent and 74 adults General anesthesia might be associated with early periprosthetic joint infection: an observational study of 3,909 arthroplasties The effect of smoking on outcomes following primary total hip and knee arthroplasty: a population-based cohort study of 117,024 patients Knee Rates of knee arthroplasty in anterior cruciate ligament reconstructed patients: a longitudinal cohort study of 111,212 procedures over 20 years Equal tibial component fixation of a mobile-bearing and fixed-bearing medial unicompartmental knee arthroplasty: a randomized controlled RSA study with 2-year follow-up Cementing technique for primary knee arthroplasty: a scoping review All-polyethylene versus metal-backed posterior stabilized total knee arthroplasty: similar 2-year results of a randomized radiostereometric analysis study Anterior cruciate ligament reconstruction-related patient injuries: a nationwide registry study in Finland Less gap imbalance with restricted kinematic alignment than with mechanically aligned total knee arthroplasty: simulations on 3-D bone models created from CT-scans Children Routine radiographic follow-up is not necessary after physeal fractures of the distal tibia in children Femoral and pelvic osteotomies for severe hip displacement in nonambulatory children with cerebral palsy: a prospective population-based study of 31 patients with 7 years’ follow-up

507

J L C van Susante

511

J Viveen, M P J van den Bekerom, J N Doornberg, A Hatton, R Page, K L M Koenraadt, C Wilson, G I Bain, R L Jaarsma, and D Eygendaal J Nordenskjöld, A Lauritzson, A Åkesson, and I Atroshi

517

523

M Tang-Jensen, P Kjærsgaard-Andersen, T K Poulsen, S Overgaard, and C Varnumg

530

G S Matharu, F Berryman, D J Dunlop, A Judge, D W Murray, and H G Pandit

537

P Sjöholm, V Otten, O Wolf, M Gordon , G Karsten, O Sköldenberg, and S Mukka K Kalland, H Åberg, A Berggren, M Ullman, G Snellman, K B Jonsson, and T Johansson

542 547

T Arashi, Y Murata, H Utsunomiya, S Kanezaki, H Suzuki, A Sakai, and S Uchida

554

R Scholten, B Leijtens, G Hannink, E T Kamphuis, M P Somford, and J L C van Susante G S Matharu, S Mouchti, S Twigg, A Delmestri, D W Murray, A Judge, and H G Pandit

559

568

S G F Abram, A Judge, Tkhan, D J Beard, and A J Price

575

D Koppens, S Rytter, S Munk, J Dalsgaard, O G Sørensen, T B Hansen, and M Stilling

582

A M Refsum, U V Nguyen, J-E Gjertsen, B Espehaug, A M Fenstad, R K Lein, P Ellison, P J Høl, and O Furnes S Hasan, P J Marang-van de Mheen, B L Kaptein, R G H H Nelissen, and S Toksvig-Larsen

590 596 602

610 614

K-M Nyrhinen, V Bister, T Helkamaa, A Schlenzka, H Sandelin, J Sandelin, and A Harilainen W Blakeney, Y Beaulieu, M-O Kiss, C Rivière, and P-A Vendittoli

A Stenroos, J Kosola, J Puhakka, T Laaksonen, M Ahonen, and Y Nietosvaara T Terjesen

Case report Functional ambulation without lower-leg muscles or nerves — a case report with video An unexpected complication of nonoperative treatment for tibial posterior malleolus fracture: bony entrapment of tibialis posterior tendon – a case report Erratum Low revision rate despite poor functional outcome after stemmed hemiarthroplasty for acute proximal humeral fractures: 2,750 cases reported to the Danish Shoulder Arthroplasty Registry (Acta Orthop 2019; 90 (3): 196–201. DOI 10.1080/17453674. 2019.1597491) Information to authors (see http://www.actaorthop.org/)

622

J Parenti

624

T Amouyel, B Benazech, M Saab, N Sturbois-Nachef, C Maynou, and P Mertl

626

A Amundsen

Acta Orthopaedica 2019; 90 (6): 507–510

507

Guest editorial

The (r)evolution of hyped innovations in orthopedic implants: can prudent introduction avoid throwing the baby out with the bathwater? Multiple new orthopedic implant designs are introduced into the market yearly. The satisfying results already achieved with most available implants challenge efforts to create additional value without introducing new patient risks. In weighing the balance between stimulating the evolution and improvement of musculoskeletal implants/medical devices and avoiding the introduction of potential risks, innovations should be introduced in a regulated, safe manner to prevent harm to patients. In spite of a consensus framework (IDEAL), which recommends the phased introduction of new medical devices (McCulloch et al. 2009), this introductory process still appears to be quite complex. Recent innovations in orthopedic devices have often been embraced by both professionals and patients after popularity has increased from industrial marketing, which in turn has led to rapidly increasing clinical use. This phenomenon of the “Hype Cycle” is well known from consumer markets and may also well be applicable to illustrate the evolution of healthcarerelated innovations (Bortfeld and Marks 2013). In this model, product evolution is divided into 3 phases (Figure 1). First, the skeptical resistance phase goes from novel technology invention towards early clinical use. Second, in the hype phase the Expectations Phase 1

Phase 2

Phase 3

Peak of inflated expectations

Plateau of Slope of productivity enlightment Stagnation

Obsolescence Idea Engineering advances increase clinical feasibility

Time

Figure 1. The 3 phases of the “Hype Cycle”: 1. Skeptical resistance phase; 2. Hype phase; 3. Post-hype phase (modified from Bortfeld and Marks 2013).

new technology is subject to optimistic mass consumer uptake to a peak of inflated expectations; the transition from skeptical resistance to hype often happens abruptly (potential benefits outweigh potential risks). Third, the post-hype phase is when users gain a more realistic understanding of the strengths and weaknesses of the new technology. In fact, the post-hype phase is most interesting since it determines whether a sustainable plateau of productivity is achieved for an innovation or whether disillusion leads to stagnation or even obsolescence. Medical innovations are particularly vulnerable to a rapid drop into obsolescence in the post-hype phase since they are subject to multiple post-market surveillance and regulations because patients’ health is involved. As such, a hyped (market and technology driven) introduction of an orthopedic innovation may not only put patients at risk but may also negatively affect sustainable productivity of the innovation itself, ultimately increasing risks of eventually “throwing out the baby with the bath water.” The awareness that a hyped introduction in itself may have a negative effect on the evolution towards clinical use of an innovation may be at least as effective to avoid such “hype” as the introduction of all sorts of regulations by notified bodies. This cultural change may then improve patient safety whilst in the meantime not frustrating innovations. Lessons can be learnt from examples of the introductory process of orthopedic innovations in the past. Analysis of the number of yearly publications may reflect whether the respective innovations underwent a hyped introduction or not (phase 2) and whether this predisposed towards a plateau of productivity in phase 3 or towards obsolescence. Several recent examples can illustrate the case. I. Modular necks for THA Recurrent dislocation is one of the most important reasons for early revision of THA (Gerhardt et al. 2014). Modular necks were introduced in THA to improve restoration of hip geometry and reduce dislocation rates. Modular necks enhanced the opportunities for the surgeon to personalize anteversion, retroversion, varus and valgus orientation of the stem intraoperatively. Initially this innovation was reserved for revision stem designs and from the preliminary results encountered the concept was expanded towards the much larger market of primary THA. Strong industry marketing strategies resulted in rapid

© 2019 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group, on behalf of the Nordic Orthopedic Federation. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. DOI 10.1080/17453674.2019.1669115

508

Figure 2. Annual number of publications registered in PubMed for 4 orthopedic innovations. The curves seem to overlap with the course of introduction in the market. 1. Modular necks in THA. A peak of inflated expectations is visible in 2016 after a “hyped” introduction, followed by a steep decline as concerns about safety appeared. Obsolescence in primary THA is likely to occur. 2. Resurfacing hip arthroplasty (RHA). A similar peak is visible in 2012 after a rapid uptake in the market, followed by a gradual decline from concerns around metal-on-metal issues. Stagnation occurs and obsolescence may be inevitable. 3. Total disc replacement (TDR). The use of lumbar disc arthroplasty increased rapidly after a “hyped” introduction in early 2000, followed by a steep decline towards obsolescence as long-term safety issues outweighed potential benefits. However, for cervical disc arthroplasty a gradual increase in evidence and acceptance in clinical use can be observed. 4. Total hip arthroplasty (THA). There has been no “hype phase” and an annually increasing number of publications corresponds with the established position in the market and a growing plateau of productivity.

Acta Orthopaedica 2019; 90 (6): 507–510

Number of publications 60

Number of publications

Modular necks

Cervical disk Lumbar disk

50 30 40 20

30 20

10 10

1982 1986 1990 1994 1998 2002 2006 2010 2014 2018

Number of publications

1996

2000

2004

2008

2012

2016

Number of publications 3,000

200

RHA

THA 2,500

150

2,000

1,500

100

1,000 50 500

1978 1983 1988 1993 1998 2003 2008 2013 2018

market uptake and a typical hype phase was initiated. Subsequently, inflated expectations of reduced dislocation rates were not met, and instead concerns were raised about adverse reaction to metal debris (ARMD) from corrosion, third body wear at the taper junction and occasional implant fracture (Vundelinckx et al. 2013). The trend in annual publications on PubMed concerning taper modularity in primary THA typically followed the hyped introduction of this innovation and uptake into the market (Figure 2). The peak of inflated expectations appears to have been reached in 2016, followed by a rapid decline, associated with concerns regarding patient safety. Since the originally presumed benefits for patients did not seem to hold (Gerhardt et al. 2014) the use of modular necks in primary THA appears to reach obsolescence in phase 3 of the hype cycle. II. Resurfacing hip arthroplasty (RHA) This was designed for young and active patients and was reintroduced as a revolutionizing bone-preserving concept decades after resistance to an earlier introduction in the 1950s (phase 1). Following improved metallurgy in metal-on-metal (MoM) bearing, a hyped introduction followed from several different manufacturers (phase 2). At the same time the number of annual publications rapidly increased in the years 2006 to 2011 (Figure 2). Disillusion with the product (phase 3) was marked by reports of unexpectedly high metal ion release and high early revision rates (Smith et al. 2012). Sub-

1970 1976 1982 1988 1994 2000 2006 2012 2018

sequently, a rapid drop in annual publications coincided with a steep decline in clinical use from 2012 through 2018 (Figure 2). The RHA concept is at risk of reaching obsolescence in phase 3 of the hype cycle. One could argue whether this is entirely justified. In the disillusion phase concerns around MoM bearings were obtained from combined data of the RHA and large-head MoM THA. Only later did it appear that bearing issues had to be distinguished from trunnion issues in MoM THA and that the latter were clearly more disturbing. As such, MoM THA may have taken down the RHA in its fall, whereas there may still be a minor niche for RHA. In specialized high-volume centers excellent 10-year survival rates have been reported; however, real-world data from different joint registries still report relatively high revision rates for RHA. From randomized controlled trials comparing RHA with conventional THA we have learnt that, besides somewhat more natural weightbearing in gait analysis (Gerhardt et al. 2019), no clear benefit could be detected in clinical outcome for RHA. Concerns around the MoM bearing remain eminent and as such the potential for significant patient risks from metal-on-metal issues so far outweigh the established minimal potential benefits over conventional THA. III. Total disc replacement (TDR) Total vertebral disc replacement was introduced as an innovative implant design to treat degenerative disc disease. The the-

Acta Orthopaedica 2019; 90 (6): 507–510

oretical advantage of TDR over spinal fusion is that movement is preserved and that, as such, adjacent segment degeneration could be avoided or minimized. Again, this innovation was targeted for young and active patients. From the year 2000 both cervical and lumbar disc arthroplasties gained popularity followed by a rapid uptake in the market. Similar to the situation with modular necks and RHA a “hype cycle” could be observed where a steep increase in clinical use paralleled increasing numbers of annual publications (Figure 2). Subsequently, meta-analyses comparing both lumbar and cervical TDR with fusion (Ding et al. 2017, Findlay et al. 2018, Katsuura et al. 2019) reported at least equivalent clinical outcome for TDR in the short term. Some subtle benefits seemed to apply more to cervical TDR, but were not beyond the generally accepted clinical important differences. Concerns were also expressed that disadvantages may appear after years as uncertainty remains about degeneration of the prosthesis (Jacobs et al. 2013). Gradually inflated expectations subsided and in particular for lumbar disc arthroplasty the annual number of surgeries decreased dramatically from 3,059 to 420 from 2005 towards 2013 in the United States (Saifi et al. 2018a). Again, this decrease in clinical use coincided with a steep decline in the number of publications (Figure 2). Interestingly, for cervical disc replacement a different pattern can be described with, as opposed to the lumbar disc replacement, a slope of enlightenment in phase 3 towards a plateau of productivity in clinical practice (Figure 2). Contrary to lumbar TDR, the annual number of cervical TDRs increased 190% from 540 to 1,565 from 2006 towards 2013 in the United States (Saifi et al. 2018b). The annual number of publications followed this pattern or vice versa. The rationale behind this different pattern is difficult to understand from similar conclusions in meta-analyses for both cervical and lumbar TDR meta-analysis. Fortunately the increase in clinical use of cervical TDR is still monitored by an increase in scientific evidence, which will determine whether the potential clinical benefit will eventually outweigh the risk for (long term) complications and revision need. IV. Total hip arthroplasty (THA) The evolution of the THA, introduced as an innovation in orthopedics by Sir John Charnley (Charnley 1961), shows a distinct pattern. The innovation has been carefully incorporated into the market represented by gradually increasing numbers of implantations with careful monitoring of results. There was no “hype cycle” and subsequently the acceptance of the innovations could be monitored by a stepwise increase in the annual number of publications (Figure 2). Hype or (r)evolution In general there is a tendency to overestimate the benefit of an innovation in the short run and underestimate the potential new risks in the long run. From the examples presented we

509

have learnt that the curve of the annual number of publications of an innovation correlates with the evolution of its use. A “hyped” introduction may predispose to failure to meet with inflated expectations, followed by a steep decline in usage and subsequent banning of the innovation. Besides patient safety concerns, these findings suggest that it may also benefit the innovation itself and the manufacturers’ interest in implementing a stepwise introduction and avoiding hype. Hype may actually predispose to obsolescence. Several regulatory measures have already been taken to improve monitoring of the introduction of future orthopedic innovations (Howard 2016). However, regulatory changes only will probably lead to suffocation of innovation, slowing the progress of healthcare innovation and improvement, and potentially slowing value creation for society. Ideally, the first step with prudent introduction of innovations would be careful monitoring of a limited number of patients treated. Only after clinical success has been warranted at least at short-term follow-up, without the introduction of new complications, could subsequent expansion of clinical use be pursued. A randomized controlled trial (RCT) where the innovation is benchmarked against the “standard of care” remains the ideal choice to evaluate the true clinical value of an innovation. Since conventional RCTs are elaborate, expensive, and commonly limited in potential patient recruitment, registry-nested RCTs are likely to gain a more dominant role. Typically, in these trials a novel approach or implant is compared with the standard of care technique in a pragmatic multicenter setting where the study is incorporated in daily clinical practice and clinically relevant outcome parameters can be obtained from available national registry data. For example, the potential benefit of decreased dislocation rates of a dual mobility cup over conventional cups is currently evaluated in such way in the “Duality” and the “Redep” trial (clinicaltrials.gov NCT03909815 and NCT04031820, respectively). Besides, a cultural change is necessary to help stakeholders understand that a “hyped introduction” of innovations should be avoided at all times. A cultural shift is advocated from sales- or fashion-driven short-term attention towards sustainability of health in long-term gain in health and quality of life (Porter 2010). From the annual publication reports of recent innovations we have learnt that a rapid uptake of a new device in the market most likely predisposes to subsequent stagnation or even obsolescence in phase 3 as inflated expectations are not met or new risks appear. For this reason, marketing of an innovation by companies should focus on building scientific evidence and as such preserving the sustainability of a new device. Professionals in turn will have to embrace this approach and avoid rapid uptake of devices without solid evidence; the time of a “boys need toys” approach is over. And, importantly, patients will have to understand that “new is not allows better.” Parts of this manuscript were taken from the PhD thesis “Innovative Implant Design in Hip Arthroplasty” at the Radboud University Medical Centre, Nijmegen (Netherlands) by Davey M Gerhardt under co-supervision with

510

Marinus de Kleuver (Professor and Chair of the Department of Orthopedics). Data from the annual publication graph on TDR were kindly provided by Richard D Guyer, MD, Center for Disc Replacement at the Texas Back Institute.

Job L C VAN SUSANTE Department of Orthopedics, Rijnstate Ziekenhuis, Arnhem, The Netherlands Correspondence: jvansusante@rijnstate.nl

Bortfeld T, Marks L B. Hype cycle in radiation oncology. Int J Radiat Oncol Biolo Phys 2013; 86: 819-21. Charnley J. Arthroplasty of the hip: a new operation. Lancet 1961; 1: 1129-32. Ding F, Jia Z, Zhao Z, Xie L, Gao X, Ma D, Liu M. Total disc replacement versus fusion for lumbar degenerative disc disease: a systematic review of overlapping meta-analyses. Eur Spine J 2017; 26(3): 806-15. Findlay C, Ayis S, Demetriades A K. Total disc replacement versus anterior cervical discectomy and fusion: a systematic review with meta-analysis of data from a total of 3160 patients across 14 randomized controlled trials with both short- and medium- to long-term outcomes. Bone Joint J 2018; 100-B(8): 991-1001. Gerhardt D M, Bisseling P, de Visser E, van Susante J L C. Modular necks in primary hip arthroplasty without anatomical deformity: no clear benefit on restoration of hip geometry and dislocation rate. An exploratory study. J Arthroplasty 2014; 29(8): 1553-8. Gerhardt D M, Ter Mors T G, Hannink G, Van Susante J L C. Resurfacing hip arthroplasty better preserves a normal gait pattern at increasing walk-

Acta Orthopaedica 2019; 90 (6): 507â&#x20AC;&#x201C;510

ing speeds compared to total hip arthroplasty. Acta Orthop 2019; 90(3): 231-6. Howard J J. Balancing innovation and medical device regulation: the case of modern metal-on-metal hip replacements. Medical Devices 2016; 9: 267-75. Jacobs W, Van der Gaag N A, Kryut M C, Tuschel A, de Kleuver M, Peul W, Verbout A J, Oner F C. Total disc replacement for chronic discogenic low back pain: a Cochrane review. Spine 2013; 38(1): 24-36. Katsuura Y, York P J, Goto R, Yang J, Vaishnav A S, McAnany S, Albert T, Iver S, Gang C H, Qureshi S A. Sagittal reconstruction and clinical outcome using traditional ACDF, versus stand-alone ACDF versus TDR: a systematic and quantitative analysis. Spine 2019: June 2019 (Epub ahead of print). doi: 10.1097/BRS.0000000000003077 McCulloch P, Altman D G, Campbell W B, et al. No surgical innovation without evaluation: the IDEAL recommendations. Lancet 2009; 374: 1105-12. Porter M E. What is value in health care? N Engl J Med 2010; 363: 2477-81. Saifi C, Cazzulino A, Park C, Laratta J, Louie P K, Shillingford J N, Lehman R A, Howard A, Phillips F M. National trends for primary and revision lumbar disc arthroplasty throughout the United States. Global Spine J 2018a; 8(2): 172-7. Saifi C, Fein A W, Cazzulino A, Lehman R A, Phillips F M, An H S, Riew K D. Trends in resource utilization and rate of cervical disc arthroplasty and anterior cervical discectomy and fusion throughout the United States from 2006 to 2013. Spine J 2018b; 18(6): 1022-29. Smith A J, Dieppe P, Howard P W, Blom A W, National Joint Registry for England and Wales. Failure rates of metal-on-metal hip resurfacings: analysis of data from the National Joint Registry for England and Wales. Lancet 2012; 380: 1759-66. Vundelinckx B J, Verhelst L A and De Schepper J. Taper corrosion in modular hip prostheses: analysis of serum metal ions in 19 patients. J Arthroplasty 2013; 28(7): 1218-23.

Acta Orthopaedica 2019; 90 (6): 511–516

511

Use and outcome of 1,220 primary total elbow arthroplasties from the Australian Orthopaedic Association National Joint Arthroplasty Replacement Registry 2008–2018 Jetske VIVEEN 1, Michel P J VAN DEN BEKEROM 2, Job N DOORNBERG 1, Alesha HATTON 3, Richard PAGE 4,5, Koen L M KOENRAADT 6, Christopher WILSON 1, Gregory I BAIN 1, Ruurd L JAARSMA 1, and Denise EYGENDAAL 7,8 1 Department of Orthopedic and Trauma Surgery, College of Medicine and Public Health, Flinders University and Flinders Medical Centre, Adelaide, Australia; 2 Shoulder and Elbow Unit, Department of Orthopedic Surgery, Onze Lieve Vrouwe Gasthuis, Amsterdam, The Netherlands; 3 South Australian Health and Medical Research Institute (SAHMRI), Adelaide, South Australia; 4 Australian Orthopedic Association National Joint Replacement Registry (AOANJRR), Adelaide, SA, Australia; 5 Barwon Centre for Orthopaedic Research and Education (B-CORE), Barwon Health, St John of God Hospital and Deakin University, Geelong, Australia; 6 Foundation for Orthopedic Research, Care & Education, Amphia Hospital, Breda, The Netherlands; 7 Upper Limb Unit, Department of Orthopedic Surgery, Amphia Hospital, Breda, The Netherlands; 8 Department of Orthopedic Surgery, Amsterdam University Medical

Centers, Amsterdam The Netherlands Correspondence: jetskeviveen@gmail.com Submitted 2019-06-21. Accepted 2019-07-22.

Background and purpose — The Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) was analyzed to determine trends in use of primary total elbow arthroplasty (TEA), the types of prostheses used, primary diagnoses, reasons for and types of revision, and whether the primary diagnosis or prosthesis design influenced the revision rate. Patients and methods — During 2008–2018, 1,220 primary TEA procedures were reported of which 140 TEAs were revised. Kaplan–Meier estimates of survivorship were used to describe the time to first revision and hazard ratios (HR) from Cox proportional hazard models, adjusted for age and sex, were used to compare revision rates. Results — The annual number of TEAs performed remained constant. The 3 most common diagnoses for primary TEA were fracture/dislocation (trauma) (36%), osteoarthritis (OA) (34%), and rheumatoid arthritis (RA) (26%). The cumulative percentage revision for all TEAs undertaken for any reason was 10%, 15%, and 19% at 3, 6, and 9 years. TEAs undertaken for OA had a higher revision rate compared with TEAs for trauma (HR = 1.8, 95% CI 1.1–3.0) and RA (HR = 2.0, CI 1.3–3.1). The Coonrad-Morrey (50%), Latitude (30%), Nexel (10%), and Discovery (9%) were the most used prosthesis designs. There was no difference in revision rates when these 4 designs were compared. The most common reasons for revision were infection (35%) and aseptic loosening (34%). Interpretation — The indications for primary and revision TEA in Australia are similar to those reported for other registries. Revision for trauma is lower than previously reported.

Total elbow arthroplasty (TEA) designs have improved and the use of TEA has increased worldwide (Day et al. 2010). However, the procedure remains challenging and the results variable. A number of studies, including registry studies, have reported the outcomes of primary TEA. Although pain relief and improved function can be achieved in many patients, the complication and revision rates after TEA range from 20% to 62% (Reinhard et al. 2003, van der Lugt et al. 2005, Brinkman et al. 2007, Kim et al. 2011, Voloshin et al. 2011, Park et al. 2013) and are higher when compared with primary total hip arthroplasty (THA) and primary total knee arthroplasty (TKA) (Voloshin 2011). Revision rates vary depending on primary diagnosis, with in general less favorable results in TEA placed for posttraumatic sequalae (Fevang et al. 2009, Plaschke et al. 2014, Krukhaug et al. 2018). At 10 years TEA post trauma, prosthesis survival has been reported to be 60% while for RA it is reported to be 90% (Gill and Morrey 1998, Cil et al. 2008). The most common indications for revision surgery are symptomatic aseptic loosening, infection, polyethylene (PE) or bushing wear, and instability (Prkic et al. 2017, Geurts et al. 2019). Primary TEA procedures are uncommon, with 0.5 procedures per 100,000 persons in Australia in 2018, compared with primary TKA and THA at 218 and 131 procedures per 100,000 persons per year respectively (AOANJRR 2018). Nationwide registries are a valuable resource to assess the performance of this uncommon procedure. Prevalence and outcomes in TEA can be identified in a community-based setting with a larger number of procedures available for analysis compared with most other types of studies. To date there have been published reports on TEA from 5 registries. These

512

Acta Orthopaedica 2019; 90 (6): 511–516

Table 1. Registry studies of TEA No. of Mean obser- 10-year Registry No. of prosthesis Female vation time No. of survival Study (country) TEA designs (%) (years) revisions (%) Skytta et al. (2009) Jenkins et al. (2013) Plaschke et al. (2014) Fevang et al. (2009) Krukhaug et al. (2018) Nestorson et al. (2018)

Finland 1,457 Scotland 1,146 Denmark 324 Norway 562 Norway 838 Sweden 406

9 NR 7 9 13 7

87 74 82 80 78 90

8.2 NR 8.8 6 b 9 b 6

201 140 68 58 158 18

Most common revision reason n (%) Loosening a

83 90 Infection 81 Loosening a 85 Loosening a 81 Loosening a 90 Loosening a

95 (47) 86 (61) 39 (57) 19 (33) 66 (42) 7 (39)

Diagnosis RA RA, OA, trauma RA, OA, trauma RA, OA, trauma RA, OA, fracture c Trauma

a Aseptic loosening b Median. c Fracture sequelae

and acute fracture NR = not reported. RA = rheumatoid arthritis. OA = osteoarthritis.

include the Finnish (Skytta et al. 2009), Scottish (Jenkins et al. 2013), Danish (Plaschke et al. 2014) Norwegian (Fevang et al. 2009, Krukhaug et al. 2018) and Swedish (Nestorson et al. 2018) arthroplasty registries (Table 1). This study reports the use and outcomes of primary TEA from the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) and compares these results with other reported studies including registry studies. This includes: (1) the number of primary TEAs performed per year; (2) the most common indications for primary TEA; (3) the reasons they were revised; (4) the overall revision rate; and (5) the effect of primary diagnosis and type of prosthesis on the rate of revision.

Patients and methods Australian Orthopaedic Association National Joint Replacement Registry This study included all primary TEA procedures reported to the AOANJRR between January 1, 2008 and December 31, 2018. The AOANJRR commenced national data collection for TEA in 2007 and by 2017 94% of elbow arthroplasty procedures had been reported to the registry (AOANJRR 2018). Registry data are validated against health department recorded data through a sequential multi-level matching process. A matching program is run monthly to search for all primary and revision arthroplasty procedures recorded in the Registry that involve the same side and joint of the same patient, thus enabling each revision to be linked to the primary procedure. Data are also matched biannually with the Department of Health and Ageing’s National Death Index to obtain information on the date of death (AOANJRR 2018). When a bilateral primary TEA was performed, each TEA was considered separately. Demographic data including patient characteristics (age, sex, and since 2012 ASA score), primary diagnosis, fixation, and type of prosthesis are reported. Fixation included cemented, hybrid, and cement-

less. Prosthesis design was identified by brand and classified as linked, unlinked, or convertible. First revision rates and reasons for revision were determined. The effect of primary diagnosis and prosthesis type on the rate of revision was also determined. The AOANJRR defines a revision as any reoperation of a previous TEA replacement where one or more of the prosthetic components are replaced, removed, or another component is added. Statistics Kaplan–Meier estimates of survivorship were used to report the time to revision of a TEA, with censoring at the time of death or closure of the dataset at the end of December 2018. The unadjusted cumulative percentage revision (CPR), with 95% confidence intervals (CI), was calculated using unadjusted pointwise Greenwood estimates. Age and sex adjusted hazard ratios (HR) calculated from Cox proportional hazard models were used to compare the rate of revision between the groups. The assumption of proportional hazards was checked analytically for each model. If the interaction between the predictor and the log of time was statistically significant in the standard Cox model, then a time-varying model was estimated. Time points were selected based on the greatest change in hazard, weighted by a function of events. Time points were iteratively chosen until the assumption of proportionality was met and HRs were calculated for each selected time period. For the current study, if no time period was specified, the HR was calculated over the entire follow-up period. All tests were 2-tailed at 5% levels of significance. Statistical analysis was performed using SAS software version 9.4 (SAS Institute Inc., Cary, NC, USA). Ethics, funding, and potential conflicts of interest Since no individual patient characteristics were available, approval by the human ethics research committee was not required. No funding for this study was received. JV received an unrestricted Research Grant from the Marti-Keuning-Eckhardt Foundation, Amsterdam Movement Sciences, Jo Kolk Foundation, and Michael-van Vloten

Acta Orthopaedica 2019; 90 (6): 511–516

513

Table 2. Data on 1,220 primary TEA

Number of procedures 60

Factor

n (%)

Elbow class TEA without radial head component 1,177 (96) TEA including radial head component 43 (4) Prosthesis design Linked Coonrad-Morrey a 608 (50) Latitude a,b 344 (28) a Nexel 121 (10) Discovery a 111 (9) Mutars a 5 (< 1) Unlinked Latitude a,b 17 (1) a IBP 1 (< 1) Souter Strathclyde a 7 (1) Undefined Comprehensive 4 (< 1) Custom-made/other 2 (< 1) Fixation technique Cemented 1,119 (92) Hybrid (ulnar cemented) 65 (5) Hybrid (ulnar cementless) 32 (3) Cementless 4 (< 1) a

Coonrad-Morrey (Zimmer, Inc., Warsaw, IN, USA), Discovery (Biomet, Inc., Warsaw, IN, USA), Nexel (Zimmer, Inc., Warsaw, IN, USA), Mutars (Implantcast GmbH, Buxtehude, Germany), Latitude (Tornier, Montbonnot-Saint-Martin, France), IBP (Biomet Inc, Warsaw, IN), Souter Strathclyde (Stryker, Rutherford, NJ, USA). b The Latitude elbow prosthesis is a convertible design and can be placed either linked or unlinked.

Foundation. JND received an unrestricted Postdoc Research Grant from the Marti-Keuning-Eckhardt Foundation. MPJB declares that the OLVG Hospital receives research support from Wright/Tornier unrelated to this study.

Results Demographic characteristics There were 1,220 primary TEAs reported to the AOANJRR during the study period of which 140 were revised. The majority were female (73%). The mean age was 70 years (female 71 years and male 69 years). ASA score was available for 630 (52%) primary TEA procedures. The majority (59%) had an ASA score of 3 or 4. Primary TEA prostheses 9 different types of prostheses were used (Table 2). The most common types were the Coonrad-Morrey (Zimmer, Inc., Warsaw, IN, USA) (n = 608; 50%) followed by the Latitude (Tornier, Montbonnot-Saint-Martin, France) (n = 344

Trauma Osteoarthritis Rheumatoid arthritis

0 2008

2010

2012

2014

2016

2018

Procedure year

Figure 1. Primary total elbow replacement by primary diagnosis.

linked and n = 17 unlinked; 30%), the Nexel (Zimmer, Inc., Warsaw, IN, USA; the Nexel became available in Australia only in 2013) (n = 121; 10%), and the Discovery (Biomet Inc, Warsaw, IN, USA) (n = 111; 9%) (Table 2). Of the types of TEA prostheses used, 4 were linked, 1 was a convertible, and 2 were unlinked designs. 2 implants were classified as undefined, because they were custom-made designs. These implants were excluded from further analysis on linked versus unlinked designs. Almost all procedures used a linked design (n = 1,189, 98%). Most prostheses were cemented (n = 1,119; 92%). The radial head was replaced in a small number of procedures (n = 43). All involved the Latitude prosthesis. The radial head was replaced in only 12% of procedures when this device was used. The number of primary TEAs performed each year remained constant (Table 3, see Supplementary data). The most common primary diagnoses were trauma (n = 434, 36%), OA (n = 414, 34%), and RA (n = 318, 26%). The proportion of primary TEAs undertaken for trauma has increased in recent years and is now the most common reason (Figure 1). Revisions of primary TEA Of the 1,220 primary TEAs, 140 were revised. The CPR was 10%, 15%, and 19% at 3, 6, and 9 years, respectively (Table 4 and Figure 2). The revision rate varied depending on the primary diagnosis. Primary TEAs undertaken for OA were revised more frequently compared with both RA (entire period: HR = 2.0, CI 1.3–3.1) and trauma (entire period: HR = 1.8, CI 1.1–3.0). There was no statistically significant difference in the rate of revision when RA and trauma were compared (entire period: HR = 0.9, CI 0.5–1.6) (Figure 3).

Table 4. Yearly unadjusted cumulative percentage revision (CPR (CI)) of primary total elbow replacement (all diagnoses)

1 year

2 years

3 years

4 years

CPR (95% CI)

4 (3–5)

7 (5–8)

10 (8–12)

11 (9–14)

5 years

6 years

7 years

8 years

9 years

13 (11–15) 15 (13–18) 17 (14–20) 18 (15–21) 19 (16–22)

514

Acta Orthopaedica 2019; 90 (6): 511–516

Cumulative percent revision

Trauma Osteoarthritis Rheumatoid arthritis

Coonrad/Morrey Discovery Latitude Nexel

20 30

15 20 10

10 10

Years since primary procedure

Figure 2. Cumulative percentage revision of primary total elbow replacement (all diagnoses).

Years since primary procedure

Figure 3. Cumulative percentage revision of primary total elbow replacement by primary diagnosis.

There was no statistically significant difference in the rate of revision when a radial head was used (entire period: HR = 1.5, CI 0.7–2.9) (Figure 4, see Supplementary data). There was no statistically significant difference when linked and unlinked prostheses were compared (0–6 months: HR = 3.7, CI 0.9–15.6; > 6 months: HR = 0.8, CI 0.2–2.4) (Figure 5, see Supplementary data). Revision rates were similar for the 4 most used prostheses (Coonrad Morrey, Discovery, Latitude, and Nexel, Figure 6). The most common reasons for revision were infection (35%) and aseptic loosening (34%) (Table 5). The most common type of revision for primary TEA procedures without radial replacement undertaken for all diagnoses was of the humeral component (n = 32; 24%), followed by an elbow linking pin only (n = 25; 19%), ulnar component (n = 122; 17%), humeral/ulnar (n = 21; 16%), and cement spacer (n = 17; 13%) (Table 6, see Supplementary data). For primary TEA procedures with a radial head, the use of an ulnar component (n = 2; 22%), humeral/ulnar (n = 2; 22%), and radial head only (n = 2; 22%) were the most common types of revision (Table 6, see Supplementary data).

Discussion This is one of the largest studies on the use and outcome of contemporary primary TEA prostheses. The annual use did not change over the 10-year period; however, there was a change in indications for primary TEA with an increased use for trauma. This has been reported previously (Gay et al. 2012). A possible explanation for this increase is that it is being used more often as a salvage procedure in selective cases of complex, comminuted, intra-articular distal humerus fractures. Its use for this diagnosis has been reported to be associated with good results (Frankle et al. 2003, McKee et al. 2009, Barco et al. 2017, Nestorson et al. 2018).

Years since primary procedure

Figure 6. Cumulative percentage revision of primary total elbow replacement (all diagnoses). Only prostheses with over 100 procedures.

Number at risk at year 0 1 2 3 4 5 6 7 8 9 Coonrad-Morrey Discovery Latitude Nexel

608 539 478 392 336 267 204 149 100 59 111 95 83 67 55 48 46 31 22 13 361 284 229 176 147 120 94 72 52 34 121 96 65 37 18 8 0 0 0 0

The percentage of patients with RA is low compared with other studies with reports of up to 70% (Fevang et al. 2009, Jenkins et al. 2013, Plaschke et al. 2014, Stamp et al. 2017, Welsink et al. 2017, Krukhaug et al. 2018). The most recent Norwegian registry study identified a substantial decrease in the use of TEA for RA over the last decade (Krukhaug et al. 2018). This is likely due to the improved medical management of RA (Emery 2002, Korpela et al. 2004, Verstappen et al. 2006). The low proportion of RA patients in this study may also reflect this. The all-cause revision rate for all diagnoses combined reported in this study is comparable to other studies (Fevang et al. 2009, Plaschke et al. 2014, Krukhaug et al. 2018). The revision rate for trauma is similar to 1 recent report (Nestorson et al. 2018). These authors considered primary TEA as a reliable treatment option for the management of complex distal humeral fractures. Although these data are supportive of that conclusion, it is our view that the use of TEA for this diagnosis, while promising, needs to be considered with some caution. This is because higher revision rates in the longer term have been reported, particularly in younger patients with posttraumatic sequelae under 65 years of age (Cil et al. 2008). The low use of unlinked prostheses in Australia is notable. Unlinked prostheses have been popular in Europe (Fevang et al. 2009, Skytta et al. 2009, Jenkins et al. 2013, Plaschke et al. 2014, Krukhaug et al. 2018). There has, however, been an increase in the use of linked prostheses over the last decade (Krukhaug et al. 2018). Unlinked prostheses have been identified as having a higher risk of revision compared with linked designs (Plaschke et al. 2014, Geurts et al. 2019). In this study,

Acta Orthopaedica 2019; 90 (6): 511–516

Table 5. Revision diagnosis of primary total elbow replacement by type of primary (all diagnoses). Values are frequency Total elbow Revision diagnosis Total elbow and radial Infection 46 3 Loosening 44 3 Fracture 13 Malposition 3 Wear bushing 3 Implant breakage ulna 2 Instability 1 2 Progression of disease 2 Arthrofibrosis 1 Implant breakage humeral 1 Incorrect aizing 1 Lysis 1 Metal related pathology 1 Prosthesis dislocation 1 1 Wear ulna 1 Other 10 No. revision 131 9 No. primary 1,177 43

we were unable to identify a difference between linked and unlinked prostheses because of the low use of unlinked prostheses. The main prostheses used in Australia are the CoonradMorrey and the Latitude (linked version). The risk of revision for these 2 devices is the same. In fact, there were similar revision rates of the 4 most commonly used prostheses, which include the Nexel and the Discovery. The reasons for revision are similar to previous reports, with infection and aseptic loosening being the most common (Brinkman et al. 2007, Fevang et al. 2009, Plaschke et al. 2014, Prkic et al. 2017). The proportions of aseptic loosening, infection, and periprosthetic fracture in this study are comparable to most other studies (Table 1). Only Jenkins et al. (2013) reported an extremely high infection rate of 61%. However, it is uncertain whether this percentage is accurate, since no cases at all of aseptic loosening were reported in this study. This study has several limitations. No functional or patientreported outcomes data are available. In addition, detail on specific patient characteristics including individual comorbidities, other factors that may impact on outcome, and disease severity were not available. It was also not possible to separate acute management of trauma and later management of trauma into separate groups. In summary, the annual use of TEA over the last decade is stable and TEA remains an uncommon procedure. The indications for primary TEA in Australia are similar to those reported by other registries. There was a trend toward the increased use of TEA for trauma and a decrease in the proportion of TEAs undertaken for RA, while the number of TEAs placed for OA remained stable. The main reasons for revision surgery (infection and aseptic loosening) and overall revision rate of 19% at 9 years are comparable to other studies as well.

515

Primary diagnosis had a major impact on the risk of revision with procedures performed for OA having almost twice the risk compared with trauma and RA. The revision rate for TEA post trauma is lower than previously reported. The almost universal use of linked TEA designs is notable and is in contrast to the European experience. Supplementary data Tables 3 and 6 and Figures 4 and 5 are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674.2019.1657342

JV: Writing of the manuscript and data analysis. MPJB: Critical revision of the manuscript and data analysis. JND, CW, GIB, RLJ, and DE: Critical revision of the manuscript. AH: Statistical help and data analysis. RP: Critical revision of the manuscript. KLMK: Statistical help and data analysis. The authors would like to thank Dr. Sophie Rainbird and Prof. Stephen Graves for their valuable suggestions. Acta thanks Ilse Degreef and Hans Rahme for help with peer review of this study.

AOANJRR. Hip, knee & shoulder arthroplasty: 2018 Annual Report. Adelaide: AOA; 2018. Barco R, Streubel P N, Morrey B F, Sanchez-Sotelo J. Total elbow arthroplasty for distal humeral fractures: a ten-year-minimum follow-up study. J Bone Joint Surg Am 2017; 99(18): 1524-31. doi: 10.2106/JBJS.16.01222. Brinkman J M, de Vos M J, Eygendaal D. Failure mechanisms in uncemented Kudo type 5 elbow prosthesis in patients with rheumatoid arthritis: 7 of 49 ulnar components revised because of loosening after 2–10 years. Acta Orthop 2007; 78(2): 263-70. doi: 10.1080/17453670710013780. Cil A, Veillette C J, Sanchez-Sotelo J, Morrey B F. Linked elbow replacement: a salvage procedure for distal humeral nonunion. J Bone Joint Surg Am 2008; 90(9): 1939-50. doi: 10.2106/JBJS.G.00690. Day J S, Lau E, Ong K L, Williams G R, Ramsey M L, Kurtz S M. Prevalence and projections of total shoulder and elbow arthroplasty in the United States to 2015. J Shoulder Elbow Surg 2010; 19(8): 1115-20. doi: 10.1016/j.jse.2010.02.009. Emery P. Evidence supporting the benefit of early intervention in rheumatoid arthritis. J Rheumatol 2002; 66(Suppl.): 3-8. Fevang B T, Lie S A, Havelin L I, Skredderstuen A, Furnes O. Results after 562 total elbow replacements: a report from the Norwegian Arthroplasty Register. J Shoulder Elbow Surg 2009; 18(3): 449-56. doi: 10.1016/j. jse.2009.02.020. Frankle M A, Herscovici Jr D, DiPasquale T G, Vasey M B, Sanders R W. A comparison of open reduction and internal fixation and primary total elbow arthroplasty in the treatment of intraarticular distal humerus fractures in women older than age 65. J Orthop Trauma 2003; 17(7): 473-80. Gay D M, Lyman S, Do H, Hotchkiss R N, Marx R G, Daluiski A. Indications and reoperation rates for total elbow arthroplasty: an analysis of trends in New York State. J Bone Joint Surg Am 2012; 94(2): 110-17. doi: 10.2106/ JBJS.J.01128. Geurts E J, Viveen J, van Riet R P, Kodde I F, Eygendaal D. Outcomes after revision total elbow arthroplasty: a systematic review. J Shoulder Elbow Surg 2019; 28(2): 381-6. doi: 10.1016/j.jse.2018.08.024. Gill D R, Morrey B F. The Coonrad-Morrey total elbow arthroplasty in patients who have rheumatoid arthritis: a ten to fifteen-year follow-up study. J Bone Joint Surg Am 1998; 80(9): 1327-35.

516

Jenkins P J, Watts A C, Norwood T, Duckworth A D, Rymaszewski L A, McEachan J E. Total elbow replacement: outcome of 1,146 arthroplasties from the Scottish Arthroplasty Project. Acta Orthop 2013; 84(2): 119-23. doi: 10.3109/17453674.2013.784658. Kim J M, Mudgal C S, Konopka J F, Jupiter J B. Complications of total elbow arthroplasty. J Am Acad Orthop Surg 2011; 19(6): 328-39. Korpela M, Laasonen L, Hannonen P, Kautiainen H, Leirisalo-Repo M, Hakala M, Paimela L, Blafield H, Puolakka K, Mottonen T, Group FI-RT. Retardation of joint damage in patients with early rheumatoid arthritis by initial aggressive treatment with disease-modifying antirheumatic drugs: five-year experience from the FIN-RACo study. Arthritis Rheum 2004; 50(7): 2072-81. doi: 10.1002/art.20351. Krukhaug Y, Hallan G, Dybvik E, Lie S A, Furnes O N. A survivorship study of 838 total elbow replacements: a report from the Norwegian Arthroplasty Register 1994–2016. J Shoulder Elbow Surg 2018; 27(2): 260-9. doi: 10.1016/j.jse.2017.10.018. McKee M D, Veillette C J, Hall J A, Schemitsch E H, Wild L M, McCormack R, Perey B, Goetz T, Zomar M, Moon K, Mandel S, Petit S, Guy P, Leung I. A multicenter, prospective, randomized, controlled trial of open reduction– internal fixation versus total elbow arthroplasty for displaced intra-articular distal humeral fractures in elderly patients. J Shoulder Elbow Surg 2009; 18(1): 3-12. doi: 10.1016/j.jse.2008.06.005. Nestorson J, Rahme H, Adolfsson L. Arthroplasty as primary treatment for distal humeral fractures produces reliable results with regards to revisions and adverse events: a registry-based study. J Shoulder Elbow Surg 2018. doi: 10.1016/j.jse.2018.07.035. Park S E, Kim J Y, Cho S W, Rhee S K, Kwon S Y. Complications and revision rate compared by type of total elbow arthroplasty. J Shoulder Elbow Surg 2013; 22(8): 1121-7. doi: 10.1016/j.jse.2013.03.003. Plaschke H C, Thillemann T M, Brorson S, Olsen B S. Implant survival after total elbow arthroplasty: a retrospective study of 324 procedures performed

Acta Orthopaedica 2019; 90 (6): 511–516

from 1980 to 2008. J Shoulder Elbow Surg 2014; 23(6): 829-36. doi: 10.1016/j.jse.2014.02.001. Prkic A, Welsink C, The B, van den Bekerom M P J, Eygendaal D. Why does total elbow arthroplasty fail today? A systematic review of recent literature. Arch Orthop Trauma Surg 2017; 137(6): 761-9. doi: 10.1007/s00402-0172687-x. Reinhard R, van der Hoeven M, de Vos M J, Eygendaal D. Total elbow arthroplasty with the Kudo prosthesis. Int Orthop 2003; 27(6): 370-2. doi: 10.1007/s00264-003-0491-4. Skytta E T, Eskelinen A, Paavolainen P, Ikavalko M, Remes V. Total elbow arthroplasty in rheumatoid arthritis: a population-based study from the Finnish Arthroplasty Register. Acta Orthop 2009; 80(4): 472-7. doi: 10.3109/17453670903110642. Stamp L K, Haslett J, Chapman P, O’Donnell J, Raja R, Rothwell A, Frampton C, Hooper G. Rates of joint replacement surgery in New Zealand, 1999– 2015: a comparison of rheumatoid arthritis and osteoarthritis. J Rheumatol 2017; 44(12): 1823-7. doi: 10.3899/jrheum.170551. van der Lugt J C, Geskus R B, Rozing P M. Primary Souter-Strathclyde total elbow prosthesis in rheumatoid arthritis: surgical technique. J Bone Joint Surg Am 2005; 87(Suppl. 1, Pt 1): 67-77. doi: 10.2106/JBJS.D.02734. Verstappen S M, Hoes J N, Ter Borg E J, Bijlsma J W, Blaauw A A, van Albada-Kuipers G A, van Booma-Frankfort C, Jacobs J W. Joint surgery in the Utrecht Rheumatoid Arthritis Cohort: the effect of treatment strategy. Ann Rheum Dis 2006; 65(11): 1506-11. doi: 10.1136/ard.2005.049957. Voloshin I, Schippert D W, Kakar S, Kaye E K, Morrey B F. Complications of total elbow replacement: a systematic review. J Shoulder Elbow Surg 2011; 20(1): 158-68. doi: 10.1016/j.jse.2010.08.026. Welsink C L, Lambers K T A, van Deurzen D F P, Eygendaal D, van den Bekerom M P J. Total elbow arthroplasty: a systematic review. JBJS Rev 2017; 5(7): e4. doi: 10.2106/JBJS.RVW.16.00089.

Acta Orthopaedica 2019; 90 (6): 517–522

517

Collagenase injections for Dupuytren disease: 3-year treatment outcomes and predictors of recurrence in 89 hands Jesper NORDENSKJÖLD 1,4, Anna LAURITZSON 1,2, Anna ÅKESSON 3, and Isam ATROSHI 1,4 1 Department

of Orthopedics, Hässleholm-Kristianstad Hospitals, Hässleholm; 2 Department of Rehabilitation, Hässleholm Hospital, Hässleholm; 3 Clinical Studies Sweden—Forum South, Skåne University Hospital, Lund; 4 Department of Clinical Sciences—Orthopedics, Lund University, Lund, Sweden Correspondence: jesper.nordenskjold@skane.se Submitted 2019-02-18. Accepted 2019-07-22.

Background and purpose — Few prospective studies have reported the long-term effect durability of collagenase injections for Dupuytren disease. We assessed the 3-year treatment outcome of collagenase injections and predictors of recurrence. Patients and methods — We conducted a single-center prospective cohort study. Indication for collagenase injection was palpable Dupuytren’s cord and active extension deficit (AED) ≥ 20° in the metacarpophalangeal (MCP) and/or proximal interphalangeal (PIP) joint. From November 2012 through June 2013, we treated 86 consecutive patients (92 hands, 126 fingers). A hand therapist measured joint contracture before, 5 weeks, and 3 years after injection. The patients rated their treatment satisfaction. Primary outcome was proportion of treated joints with ≥ 20° AED worsening between the 5-week and 3-year measurements. We analyzed predictors of recurrence. Results — 3-year outcomes were available for 83 of the 86 patients (89 hands, 120 treated fingers). Between the 5-week and 3-year measurements, AED worsened by ≥ 20° in 17 MCP (14%) and 28 PIP (23%) joints. At 3 years, complete correction (passive extension deficit 0–5°) was present in 73% of MCP and 35% of PIP joints. Treatment of small finger PIP joint contracture, greater pretreatment contracture severity, and previous fasciectomy on the treated finger were statistically significant predictors of recurrence. Treatment satisfaction was rated as very satisfied or satisfied in 59 of 87 hands. No long-term treatment-related adverse events were observed. Interpretation — 3 years after collagenase injections for Dupuytren disease, improvement was maintained and treatment satisfaction reported in two-thirds of the treated hands, with no adverse events. Complete contracture correction was achieved in 3 of 4 MCP joints, but in only a third of the PIP joints.

Open surgery (limited fasciectomy) has been the most common treatment method for Dupuytren disease (Gerber et al. 2011, Dias et al. 2013, Liu et al. 2013, Nordenskjöld et al. 2017a). Collagenase injection and percutaneous needle fasciotomy are now established first-line treatment options (van Rijssen et al. 2012, Peimer et al. 2015). Advantages of these minimally invasive procedures compared with fasciectomy are safety (Krefter et al. 2017), quick recovery and lower cost (Atroshi et al. 2014). Treatment with collagenase comprises injection into the cord, followed by finger extension 1 to 2 days later. Since the initial multicenter randomized controlled trial (Hurst et al. 2009) the procedure of injection and finger manipulation (extension) has evolved. The use of local anesthesia before the extension procedure, to optimize contracture reduction, is considered standard practice (Manning et al. 2013). Furthermore, treating multiple joints in 1 session using 2 simultaneous injections (Coleman et al. 2014, Gaston et al. 2015) and injecting a higher dose (Atroshi et al. 2015, Grandizio et al. 2017) has been shown to be safe and effective. Few prospective studies have reported the long-term effect durability of collagenase treatment. The initial multicenter study has reported outcomes at 3 and 5 years (Peimer et al. 2013, 2015) and another study has reported outcomes at 5 years (Werlinrud et al. 2018), both studies using the original injection technique. We have previously reported treatment outcomes at 2 years using a modified injection technique (Lauritzson and Atroshi 2017). In the present study we assess 3-year treatment outcomes of collagenase injections using the modified injection technique, including assessment of predictors of recurrence and patient dissatisfaction.

518

Patients and methods Study design and eligibility criteria We conducted a prospective cohort study at 1 orthopedic department in Southern Sweden. This department is the only center that treats patients with Dupuytren disease in a region with 300,000 inhabitants and no patients are referred for treatment at other centers. The indication for treatment with collagenase injections was presence of a palpable cord and an active extension deficit (AED) of ≥ 20° in the metacarpo phalangeal (MCP) joint and/or proximal interphalangeal (PIP) joint. Patients From November 2012 through June 2013, we treated 86 consecutive patients (92 hands, 126 fingers) with collagenase injections. During the study period 34 patients (34 hands) were treated with limited fasciectomy in our department (see Supplementary data). All patients were asked to participate in a follow-up examination at a minimum of 3 years after first injection. The mean age at baseline was 72 years (52–91) and 64 patients were men. 11 patients (15 fingers) had previously been treated with surgery (limited fasciectomy) on the study fingers. 7 patients with contractures in 2 or more fingers received 2 simultaneous injections (2 vials). The small finger was most commonly treated (56), followed by the ring finger (49) and middle finger (17), whereas treatment of the index finger (3) and thumb (1) was uncommon. Treatment involved the right hand in 64 of the 92 treated hands. Intervention A hand surgeon injected collagenase into the cord using a modified method (Atroshi et al. 2015). After reconstituting collagenase with 0.39 mL of diluent, the surgeon injected all reconstituted collagenase (approximately 0.80 mg) in the cord, distributed in multiple spots along the palpable cord, from the PIP joint to the palmar crease. All palpable cords across the PIP joint were treated. Before the injection, local anesthesia (10 mL of 10 mg/mL mepivacaine buffered with sodium bicarbonate) was administered as a nerve block in the proximal palm (Nordenskjöld et al. 2017b). After injection, a nurse applied a soft dressing and the hand therapist gave the patient verbal and written instructions regarding edema prophylaxis (hand exercises, elevation during rest) and avoidance of heavy use of the hand. The surgeon performed finger manipulation 1 day (43 fingers) or 2 days (83 fingers) after collagenase injection, as schedule permitted. The surgeon injected local anesthetic (similar to that before collagenase injection) and after about 20 minutes performed finger manipulation by applying pressure with the thumb along the cord to disrupt it and then manipulating the MCP and PIP joints into maximum possible extension.

Acta Orthopaedica 2019; 90 (6): 517–522

Immediately after finger manipulation, the hand therapist applied a static splint with fingers in maximal possible extension and gave instructions on edema management and range of motion exercises. The patients were instructed to use the splint at night for 8 weeks. Skin tears after the manipulation procedure were treated with simple dressing and nurse followups, as described previously (Atroshi et al. 2015). The patients returned to the hand therapist after 1 week for splint adjustment. In any case where contracture correction was incomplete and the patient was willing to receive further treatment, the surgeon scheduled the patient for a second injection (minimum 30-day interval). Measurements Before and 5 weeks after treatment, an experienced hand therapist measured AED in the fingers with a hand-held metal goniometer and recorded the results in a standardized protocol. At 3 years after injection, a single experienced hand therapist measured both AED and passive extension deficit (PED) in the fingers. All measurements were done independently of the treating surgeon. All hands were examined for possible treatment-related complications at the 5-week and 3-year follow-up evaluations. The patients were also asked to rate their satisfaction with treatment outcome according to a 4-point scale (1: very satisfied, 2: satisfied, 3: neutral, 4: dissatisfied). We reviewed the electronic records of all patients to ascertain any subsequent surgery or other procedures on the study hand. Statistics The primary outcome was treatment effect non-durability defined as the proportion of patients that worsen by ≥ 20° in AED in a treated joint between the 5-week and the 3-year measurements. Recently an expert group on Dupuytren disease reached a consensus that defined recurrence as contracture ≥ 20° in a treated joint at 1 year compared with 6 weeks post-treatment (Kan et al. 2017). Furthermore, we considered this cut-off value to be of clinical importance since it has been used in the previous collagenase multicenter study (Peimer et al. 2015). We recorded the mean AED values for MCP and PIP joints for all treated fingers at baseline, 5 weeks, and 3 years. Analysis of PED values was possible only for the 3-year follow-up, because only AED was measured at baseline and at 5 weeks. We also calculated the proportion of joints with a PED value 0–5°, defined as complete correction by previous studies, including only treated joints with baseline AED ≥ 10° (to avoid overestimating the correction). Hyperextension was considered as 0° extension deficit. The changes in AED between baseline, 5 weeks, and 3 years were statistically tested with the paired t-test. Predictors of recurrent contracture (worsening by ≥ 20° in AED) at 3 years were analyzed with a mixed-effects logistic regression model adjusting for sex and age, and odds ratios (OR) were calculated separately for MCP and PIP joints. In this analysis we excluded thumb

Acta Orthopaedica 2019; 90 (6): 517–522

Table 1. Active extension deficit before (baseline), 5 weeks, and 3 years after injection. Values are mean degrees (SD) unless specified otherwise. Number of fingers (n) in 86 patients (83 patients at 3 years) Baseline 5 weeks 3 years n = 126 n = 126 n = 120 MCP PIP Total

Mean difference (CI) Baseline– 5 weeks– 5 weeks 3 years

44 (25) 9 (15) 12 (17) 35 (31–39) a –4 (–6 to –1) b 31 (29) 12 (17) 20 (24) 19 (15–22) a –9 (–12 to –6) a 75 (39) 21 (26) 32 (29) 54 (49–58) a –13 (–16 to –9) a

CI: 95% confidence interval, MCP: metacarpophalangeal, PIP: proximal interphalangeal. a p < 0.001, b p = 0.002.

and index finger joints due to small number of treated fingers (n = 3). Predictors of patient dissatisfaction at 3 years were analyzed with a Cox regression model adjusting for baseline factors and relative risks (RR) were calculated for total AED (MCP plus PIP). For patients who underwent limited fasciectomy after collagenase injection (n = 2) the preoperative contracture values were used in the analyses. We present the data as proportions and means with standard deviations (SD) or 95% confidence intervals (CI). A 2-sided p-value < 0.05 indicate statistical significance. The analyses were performed with Stata (version 14; StataCorp, College Station TX, USA) and SPSS (version 24; IBM Corp, Armonk, NY, USA). Ethics, funding, and potential conflicts of interest This research was reviewed by the Regional Ethical Review Board in Lund (2013/656) and was conducted in accordance with the Helsinki Declaration of 1975 as revised in 2000. All patients received verbal and written information about the study and gave informed consent. The research was supported by Region Skåne. Author IA was a member of an expert group on Dupuytren disease for Pfizer in 2012 and participated in meetings organized by Sobi. The manufacturer did not support the study or any of the authors.

Results At the 3-year follow-up, 3 patients (2 men) were deceased. Thus, 3-year outcomes were available for 83 of the patients (89 hands, 120 treated fingers). Joint contracture Between the 5-week and 3-year measurements, AED worsened by ≥ 20° in 17 MCP, 28 PIP, and either joint in 41 fingers. For all treated fingers, the mean total AED (SD) was 75° (39) before injection, 21° (26) at 5 weeks, and 32° (29) at 3 years. The corresponding values for MCP joints were 44° (25), 9° (15), and 12° (17), and for PIP joints 31° (29), 12° (17) and 20° (24), respectively (Table 1). At the 3-year follow-up complete correction (PED 0–5°) in treated joints with baseline

519

Table 2. Predictors of recurrence at 3 years. Mixed effects regression model (data from 117 treated fingers, excluding thumb and index finger) Variable Odds ratio (95% confidence interval) (referent category/unit) MCP PIP Sex (male) 2.2 (0.5–9) 4.1 (1.1–15) Age (per year) 1.03 (0.95–1.11) 1.08 (1.01–1.16) Finger (small) Ring 3.1 (1.1–9) 0.2 (0.05–0.7) a Middle 5.8 (1.3–27) 0.3 (0.05–1.9) Baseline contracture (per °) 1.06 (1.03–1.09) b 1.08 (1.05–1.12) b Recurrence after fasciectomy (no) 7.2 (1.4–39) c 3.2 (0.7–16) Injection-extension interval (1 day) 0.9 (0.3–3) 1.0 (0.3–2.9) MCP: metacarpophalangeal, PIP: proximal interphalangeal. = 0.01, b p < 0.001, c p = 0.02.

AED ≥ 10° was observed in 79 of 108 MCP and 25 of 72 of PIP joints. Treatment of small finger PIP joint contracture, greater severity of pretreatment contracture and previous fasciectomy on the treated finger were statistically significant predictors of contracture recurrence (Table 2). The OR (CI) for ring finger PIP joint contracture (vs. small finger) was 0.20 (0.05–0.72), for increasing baseline severity (per degree) of MCP contracture 1.06 (1.03–1.09) and of PIP contracture 1.08 (1.05–1.12), and for previous fasciectomy (MCP joint) 7.2 (1.4–39). Age and sex had no significant association with recurrence. Patient satisfaction Patients rated degree of satisfaction with outcome in 87 of the 92 treated hands: very satisfied in 39 hands, satisfied in 20 hands, neutral in 12 hands, and dissatisfied in 16 hands. Mean improvement in total AED (SD) from baseline to 3 years was 47° (34) among very satisfied and satisfied patients and 22° (22) among dissatisfied patients; mean difference 25° (CI 13–37). Patient dissatisfaction was higher with increasing pretreatment contracture severity (RR 1.02, CI 1.01–1.04) and with contracture recurrence (RR 1.03, CI 1.02–1.05). Age and sex had no significant association with dissatisfaction. Reinjections, subsequent surgery, and adverse events 4 patients received a second injection during the study period (at 4, 8, 12, and 24 months after the first injection, respectively). 2 other patients underwent limited fasciectomy (at 6 months and 3 years after injection, respectively). No other patients had any other surgical interventions during the study period. The occurrence of skin tears was recorded and published in a previous study (Atroshi et al. 2015). At the 5-week and the 3-year follow-up the hand therapist did not observe, and the patients did not report, any adverse events. No patient suffered from neurovascular injury, flexor tendon injury, infection, or complex regional pain syndrome during the study period.

520

Discussion This prospective cohort study of consecutive patients with Dupuytren disease treated with collagenase injection using a modified injection method, with near complete follow-up, has shown that improvement was maintained in two-thirds of the treated hands 3 years after treatment. Complete contracture correction, defined as PED 0–5°, was achieved in 3 of 4 MCP joints but in only a third of the PIP joints. Previous studies have also shown that PIP joints have a higher recurrence rate regardless of treatment method (van Rijssen et al. 2012, Peimer et al. 2015, Hansen et al. 2017). Furthermore, the analyses identified PIP joint contracture in the small finger, higher severity of pretreatment contracture, and previous surgical fasciectomy as predictors of worsening after initial correction. Although the majority of patients (about two-thirds) were satisfied with the treatment, 18% were dissatisfied. These results are similar to a previous study examining satisfaction after collagenase treatment (Bradley and Warwick 2016), and were correlated with treatment outcome and severity of pretreatment contracture. Identifying predictors of recurrence and patient dissatisfaction would be helpful for surgeons when informing patients with regard to treatment expectations. At 3 years, no adverse events were noted by the evaluating hand therapist or reported by the patients, indicating that collagenase injection is a safe treatment method in the long term. In the Collagenase Option for Reduction of Dupuytren Long-Term Evaluation of Safety Study (CORDLESS), 643 of 950 of the initial study participants could be evaluated 3 years after treatment (Peimer et al. 2013). At 3 years, “worsening” of contracture (defined as ≥ 20° increase in contracture in fully or partially corrected joints with or without a palpable cord, or subsequent treatment) was 28% for MCP joints and 58% for PIP joints. In our study, using the same definition, recurrence occurred in 14% of MCP joints and 23% of PIP joints. A study of 47 patients with isolated MCP joint contracture (McFarlane et al. 2016), treated with a single 0.58 mg collagenase injection, reported a recurrence (defined as contracture > 20°) in 12 patients at 2 years. A study of 68 patients with both MCP and PIP joint contractures evaluated 2 years after collagenase injection (Van Beeck et al. 2017) showed ≥ 20° increase in extension deficit in 11 of 39 MCP joints and 18 of 29 PIP joints. Treatment outcome in our study is better, which may have several explanations. First, the use of the modified method injecting multiple spots in the cord and use of a higher dose. Second, in comparison with the CORDLESS study, use of local anesthesia before finger extension procedure may enhance optimal contracture reduction. Furthermore, we have shown a high prevalence of skin tears after finger manipulation (40% of treated hands) in a previous study (Atroshi et al. 2015), but all wounds healed without complications. Therefore, our aim during the finger manipulation procedure is always to achieve the best possible contracture

Acta Orthopaedica 2019; 90 (6): 517–522

reduction despite skin tear occurrence. In our previous prospective study reporting the 2-year outcome of collagenase injections in 48 patients (Lauritzson and Atroshi 2017), recurrence (AED worsening ≥ 20°) was observed in 7 of 50 MCP joints, similar to the 3-year outcome, but was lower for the PIP joint (7 of 50 PIP joints). There are few randomized controlled trials comparing treatment methods for Dupuytren disease. A recent Cochrane review states that there is insufficient evidence to show the relative superiority of different surgical procedures (Rodrigues et al. 2015). In a randomized controlled trial comparing percutaneous needle fasciotomy and limited fasciectomy (van Rijssen et al. 2012), with recurrence defined as ≥ 30° in total PED from 6 weeks to 3 years, the authors reported recurrence in 4 of 46 treated hands in the limited fasciectomy group and 35 of 55 hands in the needle fasciotomy group. Using the same definition, recurrence in our study occurred in 19 of 89 hands. Randomized controlled trials comparing collagenase injection with percutaneous needle fasciotomy for isolated MCP joint contractures have been reported with up to 3-year followup. These studies show no statistically significant difference in treatment outcome, and the authors suggest an advantage for percutaneous needle fasciotomy, foremost based on presumed difference in initial treatment cost (Scherman et al. 2018, Strömberg et al. 2018). Our data show that during the study period 34 hands were treated with surgical fasciectomy at the study center, constituting 27% of all treatments for Dupuytren disease (see Supplementary data). A similar treatment trend of decreasing use of fasciectomy was observed in the United States after the introduction of collagenase treatment (Zhao et al. 2016). The finding may suggest that collagenase injections can replace limited fasciectomy to a large extent. Since surgical fasciectomy is associated with higher costs than treatment with collagenase (Atroshi et al. 2014, Sefton et al. 2018), the number of fasciectomy procedures performed has to be considered in an overall cost-effectiveness analysis of treating Dupuytren disease. In a randomized study comparing percutaneous needle fasciotomy and collagenase injection for isolated PIP joint contractures (Skov et al. 2017), treatment efficacy was similar at 2 years. However, in that study 28 of 29 of the patients treated with collagenase had a PIP joint contracture in the small finger, compared with only 15 of 21 patients in the needle fasciotomy group. In our study we have identified small-finger PIP joint contracture as a predictor of recurrence. This finding suggests that future randomized controlled trials should be stratified according to small finger involvement. The limitations of our study include a single center and moderate sample size, implying uncertain generalizability. In comparison with other studies using PED as the primary outcome measure, we measured AED before treatment and 5 weeks after treatment since it may be less examiner dependent (Nordenskjöld et al. 2018) and may be more relevant as a measure of hand function. Thus, our posttreatment AED

Acta Orthopaedica 2019; 90 (6): 517–522

values are conservative compared with posttreatment PED values in other studies. We added the measurement of PED at the 3-year evaluation to enable comparison with other studies. Another limitation is the lack of follow-up between 5 weeks and 3 years. Furthermore, patient-related outcome measures (PROMs) are limited to a satisfaction scale in this study, and more information regarding the patients’ view on collagenase treatment could have been obtained with the addition of more PROMs. However, the disabilities of the arm, shoulder, and hand (DASH) questionnaire has shown only modest responsiveness in Dupuytren disease (Rodrigues et al. 2016), and disease-specific measures, such as Unité Rhumatologique des Affections de la Main (URAM) scale (Beaudreuil et al. 2011) need further independent validation. The major strength of our study is the high participation rate at the 3-year evaluation, with data available for 100% of the treated hands of patients still living. The study center, an orthopedic department to which the vast majority of patients seeking care for Dupuytren disease are referred, enhances generalizability. Furthermore, all baseline and follow-up measurements of treatment outcomes were performed using a standardized protocol and independently of the treating surgeon. In summary, our prospective cohort study with near complete follow-up is 1 of the largest studies that reports collagenase treatment outcomes at 3 years. Complete contracture correction was maintained in 3 of 4 MCP joints, but only in a third of PIP joints. It shows that collagenase injection using a modified injection technique is a safe treatment method for Dupuytren disease, with treatment satisfaction reported in two-thirds of the treated hands. Supplementary data Supplementary data are available in the online version of this article, http://dx.doi.org/10.1080/17453674.2019.1663472

JN: Data acquisition and analysis, drafting of the manuscript. AL: Data acquisition. IA: Study design and conception, data acquisition and analysis, revision of the manuscript. Acta thanks Michel E H Boeckstyns, Joakim Strömberg, and David Warwick for help with peer review of this study.

Atroshi I, Strandberg E, Lauritzson A, Ahlgren E, Walden M. Costs for collagenase injections compared with fasciectomy in the treatment of Dupuytren’s contracture: a retrospective cohort study. BMJ Open 2014; 4(1): e004166. Atroshi I, Nordenskjöld J, Lauritzson A, Ahlgren E, Waldau J, Walden M. Collagenase treatment of Dupuytren’s contracture using a modified injection method: a prospective cohort study of skin tears in 164 hands, including short-term outcome. Acta Orthop 2015; 86(3): 310-15. Beaudreuil J, Allard A, Zerkak D, Gerber R A, Cappelleri J C, Quintero N, Lasbleiz S, Bernabe B, Orcel P, Bardin T, Group US. Unité Rhumatologique des Affections de la Main (URAM) scale: development and validation of a tool to assess Dupuytren’s disease-specific disability. Arthritis Care Res 2011; 63(10): 1448-55.

521

Bradley J, Warwick D. Patient satisfaction with collagenase. J Hand Surg Am 2016; 41(6): 689-97. Coleman S, Gilpin D, Kaplan F T, Houston A, Kaufman G J, Cohen B M, Jones N, Tursi J P. Efficacy and safety of concurrent collagenase clostridium histolyticum injections for multiple Dupuytren contractures. J Hand Surg Am 2014; 39(1): 57-64. Dias J, Bainbridge C, Leclercq C, Gerber R A, Guerin D, Cappelleri J C, Szczypa P P, Dahlin L B. Surgical management of Dupuytren’s contracture in Europe: regional analysis of a surgeon survey and patient chart review. Int J Clin Pract 2013; 67(3): 271-81. Gaston R G, Larsen S E, Pess G M, Coleman S, Dean B, Cohen B M, Kaufman G J, Tursi J P, Hurst LC. The efficacy and safety of concurrent collagenase clostridium histolyticum injections for 2 Dupuytren contractures in the same hand: a prospective, multicenter study. J Hand Surg Am 2015; 40(10): 1963-71. Gerber R A, Perry R, Thompson R, Bainbridge C. Dupuytren’s contracture: a retrospective database analysis to assess clinical management and costs in England. BMC Musculoskelet Disord 2011; 12: 73. Grandizio L C, Akoon A, Heimbach J, Graham J, Klena J C. The use of residual collagenase for single digits with multiple-joint Dupuytren contractures. J Hand Surg Am 2017; 42(6): 472 e1-e6. Hansen K L, Werlinrud J C, Larsen S, Ipsen T, Lauritsen J. Difference in success treating proximal interphalangeal and metacarpophalangeal joints with collagenase: results of 208 treatments. Plast Reconstr Surg Glob Open 2017; 5(4): e1275. Hurst L C, Badalamente M A, Hentz V R, Hotchkiss R N, Kaplan F T, Meals R A, Smith T M, Rodzvilla J, Group CIS. Injectable collagenase clostridium histolyticum for Dupuytren’s contracture. N Engl J Med 2009; 361(10): 968-79. Kan H J, Verrijp F W, Hovius S E R, van Nieuwenhoven C A, Dupuytren Delphi G, Selles R W. Recurrence of Dupuytren’s contracture: a consensus-based definition. PloS One 2017; 12(5): e0164849. Krefter C, Marks M, Hensler S, Herren D B, Calcagni M. Complications after treating Dupuytren’s disease: a systematic literature review. Hand Surg Rehabil 2017; 36(5): 322-9. Lauritzson A, Atroshi I. Collagenase injections for Dupuytren’s disease: prospective cohort study assessing 2-year treatment effect durability. BMJ Open 2017; 7: e012943. Liu W, O’Gorman D B, Gan B S. Operative trends and physician treatment costs associated with Dupuytren’s disease in Canada. Can J Plast Surg 2013; 21(4): 229-33. Manning C J, Delaney R, Hayton M J. Efficacy and tolerability of Day 2 manipulation and local anaesthesia after collagenase injection in patients with Dupuytren’s contracture. J Hand Surg Eur 2013; 39(5): 466-71. McFarlane J, Syed A M, Sibly T F. A single injection of collagenase clostridium histolyticum for the treatment of moderate Dupuytren’s contracture: a 2 year follow-up of 47 patients. J Hand Surg Eur 2016; 41(6): 664-5. Nordenskjöld J, Englund M, Zhou C, Atroshi I. Prevalence and incidence of doctor-diagnosed Dupuytren’s disease: a population-based study. J Hand Surg Eur 2017a; 42(7): 673-677. Nordenskjöld J, Walden M, Kjellin A, Franzen H, Atroshi I. Benefit of local anesthesia in reducing pain during collagenase injection for Dupuytren’s contracture. Plast Reconstr Surg 2017b; 140(3): 565-9. Nordenskjöld J, Broden S, Atroshi I. Examiners’ influence on the measured active and passive extension deficit in finger joints affected by Dupuytren disease. BMC Med Res Methodol 2018; 18(1): 120. Peimer C A, Blazar P, Coleman S, Kaplan F T, Smith T, Tursi J P, Cohen B, Kaufman G J, Lindau T. Dupuytren contracture recurrence following treatment with collagenase clostridium histolyticum (CORDLESS study): 3-year data. J Hand Surg Am 2013; 38(1): 12-22. Peimer C A, Blazar P, Coleman S, Kaplan F T, Smith T, Lindau T. Dupuytren contracture recurrence following treatment with collagenase clostridium histolyticum (CORDLESS [Collagenase Option for Reduction of Dupuytren Long-Term Evaluation of Safety Study]): 5-year data. J Hand Surg Am 2015; 40(8): 1597-605.

522

Rodrigues J N, Becker G W, Ball C, Zhang W, Giele H, Hobby J, Pratt A L, Davis T. Surgery for Dupuytren’s contracture of the fingers. Cochrane Database Syst Rev 2015 (12): CD010143. Rodrigues J, Zhang W, Scammell B, Russell P, Chakrabarti I, Fullilove S, Davidson D, Davis T. Validity of the Disabilities of the Arm, Shoulder and Hand patient-reported outcome measure (DASH) and the Quickdash when used in Dupuytren’s disease. J Hand Surg Eur 2016; 41(6): 589-99. Scherman P, Jenmalm P, Dahlin L B. Three-year recurrence of Dupuytren’s contracture after needle fasciotomy and collagenase injection: a two-centre randomized controlled trial. J Hand Surg Eur 2018; 43(8): 836-40. Sefton A K, Smith B J, Stewart D A. Cost comparison of collagenase clostridium histolyticum and fasciectomy for treatment of Dupuytren’s contracture in the Australian Health System. J Hand Surg Asian Pac 2018; 23(3): 336-41. Skov S T, Bisgaard T, Sondergaard P, Lange J. Injectable collagenase versus percutaneous needle fasciotomy for Dupuytren contracture in proximal interphalangeal joints: a randomized controlled trial. J Hand Surg Am 2017; 42(5): 321-8.

Acta Orthopaedica 2019; 90 (6): 517–522

Strömberg J, Ibsen Sorensen A, Friden J. Percutaneous needle fasciotomy versus collagenase treatment for Dupuytren Contracture: a randomized controlled trial with a two-year follow-up. J Bone Joint Surg Am 2018; 100(13): 1079-86. Van Beeck A, Van den Broek M, Michielsen M, Didden K, Vuylsteke K, Verstreken F. Efficacy and safety of collagenase treatment for Dupuytren’s disease: 2-year follow-up results. Hand Surg Rehabil 2017; 36(5): 346-9. van Rijssen A L, ter Linden H, Werker P M. Five-year results of a randomized clinical trial on treatment in Dupuytren’s disease: percutaneous needle fasciotomy versus limited fasciectomy. Plast Reconstr Surg 2012; 129(2): 469-77. Werlinrud J, Hansen K, Larsen S, Lauritsen J. Five-year results after collagenase injection for Dupuytren disease. J Hand Surg Eur 2018; 43(8): 841-7. Zhao J, Hadley S, Floyd E, Earp E, Blazar P. The impact of collagenase clostridium histolyticum introduction on Dupuytren treatment patterns in the United States. J Hand Surg Am 2016; 41(10): 963-8.

Acta Orthopaedica 2019; 90 (6): 523–529

523

Survival and revision causes of hip resurfacing arthroplasty and the Mitch proximal epiphyseal replacement: results from the Danish Hip Arthroplasty Register Maja TANG-JENSEN 1,2, Per KJÆRSGAARD-ANDERSEN 1, Thomas K POULSEN 1, Søren OVERGAARD 2,3, and Claus VARNUM 1 1 Department

of Orthopaedic Surgery, Vejle Hospital, Vejle; 2 Department of Clinical Research, University of Southern Denmark, Odense; 3 Department of Orthopaedic Surgery and Traumatology, Odense University Hospital, Denmark Correspondence: majatangen@gmail.com Submitted 2018-12-04. Accepted 2019-06-04.

Background and purpose — The Mitch proximal epiphyseal replacement (PER) was developed to preserve proximal femoral bone and minimize femoral neck fracture associated with hip resurfacing arthroplasty (HRA). We studied the survival and risk of revision of HRA compared with cementless metal-on-polyethylene (MoP) total hip arthroplasty (THA) and the survival and risk of revision of the Mitch PER compared with MoP THA. Patients and methods — Using propensity score, we matched 1,057 HRA to 1,057 MoP THA and 202 Mitch PER to 1,010 MoP THA from the Danish Hip Arthroplasty Register. To estimate the relative risk (RR) of revision, we used regression with the pseudo-value approach and treated death as a competing risk. Results — The cumulative incidence for any revision of HRA at 10 years’ follow-up was 11% (95% confidence interval [CI] 9.1–13) and 6.4% (CI 5.8–7.0) for MoP THA. The RR of any revision was 1.5 (CI 1.1–2.1) for HRA at 10 years’ follow-up. By excluding the ASR components, the RR of revision at 10 years was 1.2 (CI 0.8–1.7). The cumulative incidence of revision was 9.6% (CI 4.2–18) for Mitch PER and 5.4% (CI 5.1–5.7) for MoP THA at 8 years. The RR of revision was 2.0 (CI 0.9–4.3) for Mitch PER at 8 years’ follow-up. Interpretation — The HRA had increased risk of revision compared with the MoP THA. When excluding ASR, the HRA group had similar risk of revision compared with MoP THA. The Mitch PER did not have a statistically significant increased risk of revision compared with MoP THA.

Metal-on-metal (MoM) hip resurfacing arthroplasty (HRA) is still used in younger and more active patients in some countries (Marshall et al. 2014). Disadvantages of MoM and HRA include early implant failure with specific designs, especially the ASR Hip System from DePuy (de Steiger et al. 2011), proximal femoral bone resorption, femoral neck fracture (Marshall et al. 2014), and adverse reactions to metal debris (ARMD) with reports of pseudotumors (Pandit et al. 2008, Langton et al. 2011). These factors have led to a clear drop in the use of MoM bearings in general, and since 2012 Danish national guidelines on MoM and HRA have advocated to stop the use of these implants and very few have been inserted since. The Mitch proximal epiphyseal replacement (PER) (Figure 1) was developed by Finsbury Orthopaedics and first used in 2005. It was developed to solve the problems with femoral neck fractures of HRA and secondary proximal femoral bone resorption as it was designed to preserve patient bone

Figure 1. The Mitch proximal epiphyseal replacement (PER). © 2019 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group, on behalf of the Nordic Orthopedic Federation. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. DOI 10.1080/17453674.2019.1646201

524

and strengthen the femoral neck. It has gone through extensive computer simulation studies using the finite element method, a powerful tool to investigate the mechanical behavior of structural components. The results showed that the risk of femoral neck fracture was not influenced by the presence of the revised implant and the femoral neck strength increased after implantation from 9% to 49% compared with a previous study (Martelli et al. 2011). To our knowledge, there are no clinical studies of the Mitch PER or any similar prosthesis. Further, there are no nationwide Danish mid-term results on HRA. Therefore, we examined the survival and risk of revision of HRA and Mitch PER compared with a control group of cementless MoP THA based on nationwide data from several registers in Denmark. Further, we studied the revision risk of different designs of HRAs and the causes of revision.

Patients and methods This Danish register-based study with prospectively collected data was based on data from the Danish Hip Arthroplasty Register (DHR), the Civil Registration System (CRS), and the Danish National Patient Registry (DNPR). The population in Denmark is approximately 5.8 million and every citizen is entitled to tax-funded “free” health care. Data sources The DHR is a nationwide population-based clinical database, containing prospectively collected data on primary THA and revisions. The DHR was established in 1995, validated in 2004 (Pedersen et al. 2004) and has almost complete coverage as reporting is compulsory. The completeness for primary THA is 98% and 95% for revisions using the DNPR as a reference (Danish Hip Arthroplasty Registry 2018). The CRS is an administrative register founded in 1968. It holds information on vital status, sex, date of birth, and residence on all persons residing in Denmark (and Greenland from 1972). Every Dane is given a unique 10-digit identification number at birth that allows unambiguous linkage between all medical databases in Denmark. The CRS is updated daily, is virtually complete with a prevalence of disappeared persons around 0.3%, and is checked systematically for errors (Schmidt et al. 2014). We used data from the CRS to account for censoring due to emigration or death. Missing persons and changed CRS number were treated as emigrated as we have the exact date they went missing or changed CRS number. Since 1977, the DNPR has collected data on non-psychiatric patient visits from Danish hospitals and from 1978 with nationwide coverage. Diagnosis is classified according to the Danish version of the International Classification of Diseases (ICD) and since 1993 the 10th edition (ICD-10) has been used (Schmidt et al. 2015). We used the DNPR to identify patients with comorbidity and determine the Charlson Comorbidity Index (CCI), based on diagnoses registered in the DNPR. It

Acta Orthopaedica 2019; 90 (6): 523–529

Primary THAs with > 1 year follow-up from the Danish Hip Arthroplasty Register n = 36,817 HRA n = 1,076

Mitch PER n = 202

MoP THA n = 35,539

Excluded (n = 5,934): – lateral approach, 1,222 – anterior approach, 52 – other, not posterior, approach, 4 – missing data on approach, 363 – femoral head > 36 mm, 1,205 – femoral head < 32 mm, 2,840 – dual mobility acetabular systems, 151 – THA femoral stem with HRA femoral head, 30 – long femoral stem, 1 – misclassification on emigration or dead at time of operation, 66 THAs with complete information eligible for propensity score matching n = 30,883 HRA n = 1,056

Mitch PER n = 202

Matched groups (1:1): HRA MoP THA Total no. 1,056 1,056 Revised 111 67 Emigrated 7 6 Dead 34 79

MoP THA n = 29,625

Matched groups (1:5): Mitch PER MoP THA Total no. 202 1,010 Revised 13 46 Emigrated 1 7 Dead 11 57

Figure 2. Flow diagram: inclusion of hips with hip resurfacing arthroplasty (HRA), Mitch proximal epiphyseal replacement (PER), or cementless metal-on-polyethylene total hip arthroplasty (MoP THA) in the study population.

contains 19 major disease categories and brings the comorbidities down to 1 single numeric score (Thygesen et al. 2011). For each patient at time of surgery, the CCI was classified into 3 groups: low (CCI 0), medium (CCI 1 to 2), high (CCI 3 or more). The coverage and completeness of the 19 Charlson conditions have all been validated and have an overall positive predictive value of 98 % (Thygesen et al. 2011). Study population (Figure 2) This study is reported according to the RECORD guidelines. From the DHR, we identified patients with HRA, Mitch PER, or primary cementless MoP THAs with highly crosslinked polyethylene and minimum 1 year of follow-up (n = 36,817). The first HRAs were followed from 2005 and Mitch PER from 2008 until end of the study period in 2016. THAs with missing information on approach or approach other than posterior (n = 1,641) were excluded as these approaches are not commonly used. THAs with a femoral head size smaller than 32 mm (n = 1,222) were excluded, because they have a greater risk of dislocation (Kostensalo et al. 2013). Further, also THAs with femoral head sizes larger than 36 mm (n = 1,205) were excluded due to increased volumetric polyethylene wear (Cooper and Della Valle 2014). Patients with a dual mobility acetabular cup (n = 151), a THA stem combined with a HRA femoral head (n = 30), or with long femoral stem (n = 1) were excluded, as were patients registered as emigrated or dead at date of surgery (n = 66).

Acta Orthopaedica 2019; 90 (6): 523–529

Definitions Patients entered the study on the date of primary surgery and were followed until revision, death, emigration, or end of study period (October 24, 2016), whichever came first. Revision was defined as a new surgical intervention including partial or complete removal or exchange of the implant. Revision for any reason was considered as primary endpoint and aseptic loosening, dislocation, femoral fracture, and “other” revision causes were considered secondary endpoints. Time since operation was chosen as the underlying timescale in the time-to-event analysis and death was considered a competing risk. Patients with cementless MoP THA were used as reference, as cementless MoP bearings were considered standard. Statistics Patients with HRA and Mitch PER were matched to patients with MoP THA using a propensity score calculated on sex, age (as a continuous variable), year of surgery (as categorical variable), osteoarthritis (OA) as diagnosis, and CCI score as these may influence the outcome (Johnsen et al. 2006, Deleuran et al. 2015, Danish Arthroplasty Registry 2018). We used nearest-neighbor matching with no replacement of controls to simplify the statistical analysis. The balance in baseline variables was examined using standardized differences, where an absolute value below 10% was regarded as balanced (Austin 2014). The number of matched controls for HRA and Mitch PER were determined based on the balance of the baseline variables. We found that 1 THA for every HRA and 5 THAs for every Mitch PER gave a standardized difference below 10% for most variables except for age at surgery (11%) and year of surgery (16%) in the HRA group. Descriptive statistics were used for the presentation of demographic data and procedure characteristics. The chisquare test was used to compare proportions, and the 2-sample Wilcoxon rank-sum test was used to compare ages and followup time because of skewness of these distributions. Median and interquartile range (IQR) are given for age and followup time. Cumulative incidence of any revision was computed using the Aalen–Johansen estimator accounting for competing risk. The Aalen–Johansen method estimates the patient’s risk of undergoing a revision as a function of time since operation (Ranstam et al. 2011, Andersen and Keiding 2012, Lacny et al. 2015). Multivariable regression based on the pseudo-value observation (Klein et al. 2007) was calculated at the pre-specified time-points 2, 4, 6, and 8 years after surgery for Mitch PER and 2, 4, 6, 8, and 10 years for HRA. Once the pseudo-observations had been computed, a model for relative risk (RR) for the uncensored data was applied via generalized estimating equation. In practice, the generalized estimating equation can be obtained in a generalized linear model for the pseudoobservations (Parner and Andersen 2010). We performed stratified analysis on sex, age, OA as diagnosis, comorbidity,

525

Table 1. Patient- and surgery-related characteristics for the patients who received hip resurfacing arthroplasty (HRA) or cementless metal-on-polyethylene total hip arthroplasty (MoP THA)

Factor

MoP THA MoP THA HRA (matched) (full cohort) Stand. n = 1,056 n = 1,056 n = 29,625 diff. a p-value

Sex 0.05 0.2 Female 287 (27) 290 (28) 16,096 (54) Male 769 (73) 766 (73) 13,529 (46) Age at operation 0.10 < 0.001 < 49 332 (31) 285 (27) 1,939 (7) 50–59 463 (44) 360 (34) 4,432 (15) 60–69 248 (24) 345 (33) 11,724 (40) 70–79 13 (1) 61 (6) 9,278 (31) > 80 0 (0) 5 (1) 2,252 (8) Diagnosis –0.10 < 0.001 Primary OA 907 (86) 867 (82) 25,236 (85) Trauma 12 (1) 46 (4) 1,742 (6) Femoral head osteonecrosis 3 (0) 24 (2) 718 (2) Arthritis 6 (1) 10 (1) 299 (1) Childhood hip disorders 120 (11) 88 (8) 1,372 (5) Other 8 (1) 21 (2) 258 (1) Year of surgery 0.20 0.01 2005 24 (2) 20 (2) 26 (0) 2006 213 (20) 141 (13) 223 (1) 2007 190 (18) 206 (20 665 (2) 2008 164 (16) 180 (17) 1,125 (4) 2009 218 (21) 221 (21) 2,010 (7) 2010 168 (16) 201 (19) 3,054 (10) 2011 66 (6) 70 (7) 3,792 (13) 2012 13 (1) 17 (2) 4536 (15) Charlson comorbidity index at surgery 0.06 0.4 Low 904 (86) 881 (83) 23,543 (80) Medium 120 (11) 137 (13) 4414 (15) High 32 (3) 38 (4) 1668 (6) Values are numbers of patients and percentages (%) within each group. a Standardized difference and p-value are between HRA and matched MoP THA.

and on the different designs for HRA. Stratified analysis was performed at 10 years for HRA and 8 years for Mitch PER. Any p-value < 0.05 was considered significant and 95% confidence intervals (CI) were calculated. Statistical analyses were carried out using Stata statistical software, release 14.2 (StatCorp, College Station, TX, USA). Ethics, funding, and potential conflicts of interest This study was approved by the Danish Data Protection Agency (journal no. 2008-58-0035). Research was funded by grants from Lillebaelt hospital and the Southern Region of Denmark. No conflict of interest was reported by the authors.

Results Description of study population (Tables 1 and 2) 29,625 cementless MoP THA, 1,056 HRAs, and 202 Mitch PER with complete information on sex, age, diagnosis,

526

Acta Orthopaedica 2019; 90 (6): 523–529

Table 2. Patient- and surgery-related characteristics for the patients who received the Mitch proximal epiphyseal replacement (PER) or cementless metal-on-polyethylene total hip arthroplasty (MoP THA)

Factor

MoP THA MoP THA Mitch PER (matched) (full cohort) Stand. n = 202 n = 1,010 n = 29,625 diff. a p-value

Sex 0.06 0.4 Female 50 (25) 223 (22) 16,096 (54) Male 152 (75) 787 (78) 13,529 (46) Age at operation 0.09 < 0.001 < 49 51 (25) 267 (26) 1,939 (7) 50–59 96 (48) 291 (29) 4,432 (15) 60–69 49 (24) 373 (37) 1,1724 (40 70–79 5 (3) 77 (8) 9,278 (31) > 80 1 (1) 2 (0) 2,252 (8) Diagnosis 0.03 0.1 Primary OA 168 (83) 858 (84) 25,236 (85) Trauma 7 (4) 31 (3) 1,742 (6) Femoral head osteonecrosis 13 (6) 28 (3) 718 (2) Arthritis 1 (1) 9 (1) 299 (1) Childhood hip disorders 12 (6) 73 (7) 1,372 (5) Other 1 (1) 16 (2) 258 (1) Year of surgery: 0.05 0.9 2008 23 (11) 117 (11) 1,125 (4) 2009 77 (38) 374 (38) 2,010 (7) 2010 67 (33) 326 (34) 3,054 (10) 2011 35 (17) 193 (17) 3,792 (13) Charlson comorbidity index at surgery –0.02 1.0 Low 180 (90) 908 (90) 23,543 (80) Medium 19 (9) 88 (9) 4,414 (15) High 3 (2) 14 (1) 1,668 (6) Values are numbers of patients and percentages (%) within each group. a Standardized difference and p-value are between Mitch PER and matched MoP THA.

comorbidity, year of surgery, surgical approach, and femoral head size were included in the analyses. Among patients with HRA, the median follow-up was 8 (IQ 6.4–9.4) years for HRA and 8 (IQR 6.1–9.0) years for matched MoP THA (p = 0.05). The median follow-up was 7 (IQR 6–8) years for Mitch PER and 7 (IQR 6–7) years for the matched MoP THA (p = 0.003). The median age was 55 (IQR 48–60) years for HRAs and 56 (IQR 50–61) years for Mitch PER. The most common diagnosis was OA, accounting for 86% of HRA and 83% of Mitch PER. HRA was used between 2005 and 2012 and Mitch PER between 2008 and 2011. HRAs were used in 15 different hospitals and Mitch PER in only 1 hospital. Risk of any revision of HRA At 10 years’ follow-up, the cumulative incidence of any revision of HRA was 11% (9–13) and 6.4% (6–7) of MoP THA (Figure 3). At 4, 6, 8, and 10 years there was a statistically significantly higher RR of any revision of HRA compared with MoP THA (Table 3). Most HRA revisions occurred in 2012 and 2013, accounting for 34 % (n = 37) and 16 % (n = 17), respectively.

Table 3. Relative risk (RR) of any revision with 95% confidence interval (CI) for hip resurfacing arthroplasty (HRA), unmatched cementless metal-on-polyethylene total hip arthroplasty (MoP THA), and propensity-matched MoP THA Relative risk of revision Patients at before after the start of matching matching Factor the period RR (95% CI) RR (95% CI) 0–2-year follow-up HRA 1,056 MoP THA 1,056 Unmatched MoP THA 29,625 2–4-year follow-up HRA 1,020 MoP THA 1,010 Unmatched MoP THA 27,891 4–6-year follow-up HRA 978 MoP THA 977 Unmatched MoP THA 27,085 6–8-year follow-up HRA 944 MoP THA 942 Unmatched MoP THA 26,565 8–10-year follow-up HRA 922 MoP THA 917 Unmatched THA 26,343

0.8 (0.6–1.1) 1.0 (0.6–1.6) – 1 (ref.) 1 (ref.) – 1.5 (1.2–2.0) 1.5 (1.0–2.2) – 1 (ref.) 1 (ref.) – 2.4 (1.9–3.2) 1.7 (1.2–2.3) – 1 (ref.) 1 (ref.) – 2.9 (2.1–4.0) 1.6 (1.2–2.2) – 1 (ref.) 1 (ref.) – 2.9 (1.8–4.8) 1.5 (1.1–2.1) – 1 (ref.) 1 (ref.) –

Risk of any revision of Mitch PER The cumulative incidence of any revision of the Mitch PER was 10% (4–18%) and 5.4% (5–6%) for MoP THA at 8 years (Figure 4). There was a not statistically significant difference in the RR of any revision for the Mitch PER compared with MoP THA at 8 years (Table 4, see Supplementary data). Among all the revisions of the Mitch PER, 41% (n =5) were in 2013. Stratified analysis HRAs had a statistically significantly higher risk of revision for women (2, CI 1–3), for patients younger than 60 years (2, CI 1–2), for patients diagnosed with OA (2, CI 1–2), and for patients with no comorbidity (2, CI 1–2) compared with THAs at 10 years’ follow-up. There was no significant difference in risk of revision for men, patients aged 60 years or older, patients with other diagnoses than OA, or any comorbidity (CCI score > 0). For different designs of HRAs at 10 years’ follow-up (Figure 5), the RR of revision for any reason was higher for the ASR component compared with MoP THA (3, CI 2–5) (Table 5, see Supplementary data). After excluding patients with ASR components, the cumulative incidence of any revision at 10 years’ follow-up was 11 (CI 9–12). The cumulative incidence for ASR alone was 23 (CI 17–29) The RR of revision for any reason was not statistically significantly different (1, CI 1–2) for the HRA compared with

Acta Orthopaedica 2019; 90 (6): 523–529

527

Cumulative revision incidence

0.15

0.30

MoP THA HRA

MoP THA Mitch PER

0.10

MoP THA Recap BHR ASR Durom

0.25

0.20

0.10

0.15

0.05

0.10

0.05

0.00

Years postoperatively

Figure 3. Cumulative incidence for any revision for hip resurfacing arthroplasty (HRA) and propensity score matched cementless metal-on-polyethylene total hip arthroplasty (MoP THA).

Figure 4. Cumulative incidence for any revision for Mitch proximal epiphyseal replacement (PER) and propensity score matched cementless metal-on-polyethylene total hip arthroplasty (MoP THA).

Figure 5. Cumulative incidence for any revision for different designs of resurfacing arthroplasty and propensity score matched cementless metal-on-polyethylene total hip arthroplasty (MoP THA).

Number at risk:

Year

0 2 4 6 8 10

Year 0

Year

MoP THA 1,056 1,010 975 819 431 82 HRA 1,056 1,020 978 834 473 159

MoP THA 1,010 MITCH PER 202

972 195

940 188

673 138

81 8

MoP THA 1,056 1,010 975 819 431 82 Recap 548 528 511 436 236 69 ASR 286 281 272 217 94 19 Durom 177 166 152 139 103 44 BHR 45 45 43 42 40 27

MoP THA at 10 years’ follow-up, when excluding patients with ASR components. Regarding Mitch PER, there was no statistically significant difference in risk of revision for women, men, patients younger than 60 years, patients older than 60 years, diagnosed with OA, with other diagnoses than OA, and no or any comorbidity compared with MoP THA at 8 years’ follow-up. Causes of revision (Table 6, see Supplementary data) Among HRA revisions, pain, femoral fracture, and “other reasons” dominated, whereas only a few were revised due to dislocation compared with the MoP THAs (p < 0.001). At 10 years’ follow-up, the RR of revision due to femoral fracture was higher for HRA than MoP THA (3, CI 1–9) but the RR of revision due to dislocation (0, CI 0–0.5) was lower. We found no statistically significant difference in RR of revision due to aseptic loosening. The majority of Mitch PER were revised due to femoral fracture. At 8-year follow-up, the RR of revision due to fracture was markedly higher for Mitch PER than MoP THA (34, CI 4–279) and the RR of revision due to dislocation (0, CI 0–0.4) and “other causes” (0, CI 0–0) was lower. We found no statistically significant difference in RR of revision due to aseptic loosening.

0 2 4 6 8 10

Discussion In this nationwide, register-based study from DHR, we found higher overall revision of HRA compared with MoP THA. However, if patients with ASR were excluded from the study population the RR of revision for any reason was similar for HRA and uncemented MoP THA. The risk of revision of Mitch PER was not statistical significantly different compared with that of MoP THA after 8 years. Most of the revisions were performed in the early period after the Danish national guidelines advocated discontinuing the use of MoM bearings. Revision for any reason In the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) (Australian Orthopaedic Association 2017), the cumulative incidence of revision for any reason was 9.6 (CI 9.1–10) for HRA after 10 years’ followup, which is lower than in our study. This could be explained by differences in the use of specific component designs. In our study, 16% of the HRAs were ASR while in Australia only 7% of the HRAs were ASR. In the National Joint Registry for England, Wales, Northern Ireland and the Isle of Man (NJR) (National Joint Registry 2017), the cumulative percentage probability of revision was 11.2 (10.8–11.5). In the NJR, Kaplan–Meier estimates were used, which could account for

528

the increased risk of revision, since the Kaplan–Meier estimator may overestimate the risk of revision by not accounting for death as a competing risk (Gillam et al. 2010). The Mitch PER had a lower cumulative incidence of revision at 8 years than the 8-year cumulative incidence for the HRAs in our study (10%; CI 8.5–12%), which could indicate that Mitch PER might have a survival rate similar to some of the HRAs. Causes of revision We found that the most common cause of revision for HRA was “other causes” and pain, followed by loosening and fracture of the femoral neck, which are not comparable to the AOANJRR (where it was aseptic loosening, metal-related pathology, and fracture of the femoral neck) or to the NJR (where pain, metal-related pathology, and aseptic loosening were the most common causes). The reason for this could be that revisions due to “metalrelated pathology” such as ARMD have not been registered in the DHR and would likely be classified as “other causes” or “pain.” The Mitch PER was designed to protect bone of the proximal femur, to minimize the risk of femoral neck fracture. However, the most common cause of revision of the Mitch PER was fracture of the femoral neck followed by aseptic loosening. This could indicate that the Mitch PER may not have the femoral bone protecting abilities as previously believed. Strengths and limitations The strengths of our study include the population-based design with prospectively collected data, the large sample size, and the complete follow-up that limits possible selection bias. Further, the DHR is a medical database with independently registered data with moderate to high validity (Danish Hip Arthroplasty Registry 2018). Although the DHR has been validated, prosthetic joint infection is the only revision cause that has been validated (Gundtoft et al. 2016). Hence, misclassifications of revisions cannot be fully accounted for. There are several limitations that should be considered when interpreting the results. We excluded 433 THAs with missing data on surgical approach (n = 367), or death or emigration (n = 66). However, we assume that they would have no influence on the results in this large study population. Even though the use of cumulative competing risk analysis and the regression with the pseudo-value approach is based on assumption of independent observations, we chose not to exclude bilateral hips because a previous published guideline reported that the bilateral issue in register settings has little practical consequence when the outcome studied is revision (Ranstam et al. 2011). Even though we used propensity score matching to account for several confounders, there is still the possibility of unmeasured confounding. We did not include femoral head size in the propensity score matching even though it is a well-doc-

Acta Orthopaedica 2019; 90 (6): 523–529

umented risk factor (Smith et al. 2012), as THA has smaller femoral head size than HRA and thereby could be considered a proxy for the different groups in this study. Further, the DHR does not contain any information on potential confounders such as blood concentrations of chromium and cobalt ion levels, height, weight, BMI, physical activity before and after surgery, or medication before and after surgery. We used nearest-neighbor matching to match HRAs to MoP THA but could not meet our criteria of a standardized difference below 10% for year of surgery (19%). However, this should not have any impact on the results, as the difference is small and the surgical technique for THA surgery has not been significantly altered for several years. In our stratified analysis we divided the HRAs and Mitch PER into smaller groups. This might leave room for a type II error due to the small numbers in each group, especially the Mitch PER group. The increased risk of revision for any reason found among the HRAs might be influenced by the tendency to revise HRAs earlier because of the increased focus on ARMD and pseudotumors since January 2012 when a documentary concerning the dangers of MoM bearings was released in Denmark. This might be the reason for the increased number of revisions seen in 2012 and 2013 for the HRAs and Mitch PER and could shorten the survival compared with THAs. Surgeons may have been more prone to do revision surgery due to pressure from different stakeholders including orthopedic surgeons, administrative systems, patients, press, and industry. Or the increased risk of revision could be real and have no relation to the increased focus. All surgeries with the Mitch PER were performed at the same hospital and by one and the same surgeon. This could influence the results depending on the surgeon’s preferences (confounding by indication) and threshold for revision. While the RR of revision for any reason at 8 years was not significant, the rising RR might indicate that the risk of revision for any reason may become significant with longer follow-up. Conclusion We showed that the HRA had an increased risk of revision compared with the MoP THA at 10 years’ follow-up. When excluding ASR, the HRA group had a similar risk of revision compared with MoP THA. Mitch PER did not have a statistically significantly increased risk of revision, but as the RR is increasing every 2nd year together with the broad confidence interval, this might indicate that with longer follow-up the results could have shown a statistically significantly increased revision risk. The most common cause of revision of the Mitch PER was femoral fracture. Hence, this prosthesis does not protect the femur from fracture as computer simulation has previously indicated. We found revisions of HRA during the whole follow-up even 10 years after implantation, which is why we suggest that these patients should be followed clinically.

Acta Orthopaedica 2019; 90 (6): 523–529

Supplementary data Tables 4–6 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674. 2019.1646201 All the authors designed the study protocol. MTJ and CV collected the data. The analyses were planned by all the authors and they were carried out by MTJ. MTJ wrote the initial draft of the manuscript, which was critically revised by all the authors. Acta thanks Harald Brismar and Ola Rolfson for help with peer review of this study.

Andersen P K, Keiding N. Interpretability and importance of functionals in competing risks and multistate models. Stat Med 2012; 31(11–12): 1074-88. Austin P C. The use of propensity score methods with survival or time-toevent outcomes: reporting measures of effect similar to those used in randomized experiments. Stat Med 2014; 33(7): 1242-58. Australian Orthopaedic Association. Austrialian National Joint Replacement Registry Annual Report; 2017. Cooper H J, Della Valle C J. Large diameter femoral heads: is bigger always better? Bone Joint J 2014; 96-B(11 Suppl. A): 23-6. Danish Arthroplasty Registry. Danish Annual Report 2018; 2018. de Steiger R N, Hang J R, Miller L N, Graves S E, Davidson D C. Five-year results of the ASR XL Acetabular System and the ASR Hip Resurfacing System: an analysis from the Australian Orthopaedic Association National Joint Replacement Registry. J Bone Joint Surg Am 2011; 93(24): 2287-93. Deleuran T, Vilstrup H, Overgaard S, Jepsen P. Cirrhosis patients have increased risk of complications after hip or knee arthroplasty. Acta Orthop 2015; 86(1): 108-13. Gillam M H, Ryan P, Graves S E, Miller L N, de Steiger R N, Salter A. Competing risks survival analysis applied to data from the Australian Orthopaedic Association National Joint Replacement Registry. Acta Orthop 2010; 81(5): 548-55. Gundtoft P H, Pedersen A B, Schonheyder H C, Overgaard S. Validation of the diagnosis “prosthetic joint infection” in the Danish Hip Arthroplasty Register. Bone Joint J 2016; 98-B(3): 320-5. Johnsen S P, Sorensen H T, Lucht U, Soballe K, Overgaard S, Pedersen A B. Patient-related predictors of implant failure after primary total hip replacement in the initial, short- and long-terms: a nationwide Danish follow-up study including 36,984 patients. J Bone Joint Surg Br 2006; 88(10): 1303-8. Klein J P, Logan B, Harhoff M, Andersen P K. Analyzing survival curves at a fixed point in time. Stat Med 2007; 26(24): 4505-19.

529

Kostensalo I, Junnila M, Virolainen P, Remes V, Matilainen M, Vahlberg T, Pulkkinen P, Eskelinen A, Makela K T. Effect of femoral head size on risk of revision for dislocation after total hip arthroplasty: a population-based analysis of 42,379 primary procedures from the Finnish Arthroplasty Register. Acta Orthop 2013; 84(4): 342-7. 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. Langton D J, Joyce T J, Jameson S S, Lord J, Van Orsouw M, Holland J P, Nargol A V, De Smet K A. Adverse reaction to metal debris following hip resurfacing: the influence of component type, orientation and volumetric wear. J Bone Joint Surg Br 2011; 93(2): 164-71. Marshall D A, Pykerman K, Werle J, Lorenzetti D, Wasylak T, Noseworthy T, Dick D A, O’Connor G, Sundaram A, Heintzbergen S, Frank C. Hip resurfacing versus total hip arthroplasty: a systematic review comparing standardized outcomes. Clin Orthop Relat Res 2014; 472(7): 2217-30. Martelli S, Taddei F, Cristofolini L, Schileo E, Rushton N, Viceconti M. A new hip epiphyseal prosthesis: design revision driven by a validated numerical procedure. Med Eng Phys 2011; 33(10): 1203-11. National Joint Registry. National Annual Report 2017; 2017. Pandit H, Glyn-Jones S, McLardy-Smith P, Gundle R, Whitwell D, Gibbons C L, Ostlere S, Athanasou N, Gill H S, Murray D W. Pseudotumours associated with metal-on-metal hip resurfacings. J Bone Joint Surg Br 2008; 90(7): 847-51. Parner E T, Andersen P K. Regression analysis of censored data using pseudoobservations. Stata J 2010; 10(3): 408-22. Pedersen A, Johnsen S, Overgaard S, Soballe K, Sorensen H T, Lucht U. Registration in the Danish Hip Arthroplasty Registry: completeness of total hip arthroplasties and positive predictive value of registered diagnosis and postoperative complications. Acta Orthop Scand 2004; 75(4): 434-41. Ranstam J, Karrholm J, Pulkkinen P, Makela K, Espehaug B, Pedersen A B, Mehnert F, Furnes O, group N s. Statistical analysis of arthroplasty data, II: Guidelines. Acta Orthop 2011; 82(3): 258-67. 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. Smith A J, Dieppe P, Howard P W, Blom A W, National Joint Registry for E, Wales. Failure rates of metal-on-metal hip resurfacings: analysis of data from the National Joint Registry for England and Wales. Lancet 2012; 380(9855): 1759-66. Thygesen S K, Christiansen C F, Christensen S, Lash T L, Sorensen H T. The predictive value of ICD-10 diagnostic coding used to assess Charlson comorbidity index conditions in the population-based Danish National Registry of Patients. BMC Med Res Methodol 2011; 11: 83.

530

Acta Orthopaedica 2019; 90 (6): 530–536

Has the threshold for revision surgery for adverse reactions to metal debris changed in metal-on-metal hip arthroplasty patients? A cohort study of 239 patients using an adapted risk-stratification algorithm Gulraj S MATHARU 1,2,3, Fiona BERRYMAN 3, David J DUNLOP 3, Andrew JUDGE 1,2, David W MURRAY 1, and Hemant G PANDIT 1,4 1 Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences University of Oxford, Nuffield Orthopaedic Centre, Oxford; 2 Musculoskeletal Research Unit, Translational Health Sciences, Bristol Medical School, University of Bristol; 3 The Royal Orthopaedic Hospital, Birmingham; 4 Leeds Institute of Rheumatic and Musculoskeletal Medicine, Chapel Allerton Hospital and University of Leeds, Leeds, UK

Correspondence: gsm@doctors.org.uk Submitted 2019-05-01. Accepted 2019-08-03.

Background and purpose — A risk-stratification algorithm for metal-on-metal hip arthroplasty (MoMHA) patients was devised by US experts to help clinicians make management decisions. However, the proposed algorithm did not cover all potential patient or surgical abnormalities. Therefore we adapted the US risk-stratification algorithm in MoMHA patients revised for adverse reactions to metal debris (ARMD) to determine the variability in the revision threshold, and also whether high-risk patients had inferior outcomes following revision. Patients and methods — We analysed 239 MoMHA revisions for ARMD between 2001 and 2016 from 2 centres with pre-revision blood metal ions and imaging. Patients were stratified (low risk, moderate risk, high risk) using prerevision factors (implant, radiographic, blood metal ions, cross-sectional imaging) by adapting a published algorithm. The risk categories for each factor were assessed against revision year, revision centre, and post-revision outcomes (re-revision surgery, and any poor outcome). Results — Compared with hips revised before 2012, hips revised from 2012 onwards included more high-risk implants (44% vs. 17% pre-2012), high-risk radiographic features (85% vs. 69% pre-2012), and low-risk metal ions (41% vs. 19% pre-2012). 1 centre more frequently revised patients with high-risk implants (48% vs. 14%) and low-risk blood metal ions (45% vs. 15%) compared with the other. All these comparisons were statistically significant (p < 0.05). With the limited sample size available, implant, radiographic, blood metal ion, and cross-sectional imaging risk groups did not statistically significantly affect the rates of re-revision surgery or frequency of poor outcomes post-revision.

Interpretation — When applying the adapted risk-stratification algorithm the threshold for ARMD revision changed over time, presumably due to increasing evidence, patient surveillance, and investigation since 2012. Lower blood metal ion thresholds were used from 2012 for ARMD revisions; however, there was evidence that centres attached different importance to metal ions when managing patients. High-risk patients did not have inferior outcomes following ARMD revision.

Metal-on-metal hip arthroplasty (MoMHA) in the form of stemmed total hip arthroplasty (THA) and hip resurfacing were used in large volumes, but due to high implant failure rates they have since been abandoned (Smith et al. 2012a, 2012b). Adverse reactions to metal debris (ARMD) represent the commonest revision indication (Matharu et al. 2016c, 2017b). Many worldwide regulatory authorities have proposed guidance on managing MoMHA patients (MHRA 2012, 2017, FDA 2013). However, these guidelines are complex, and in some cases are contradictory and not supported by evidence (Matharu et al. 2018c). Managing MoMHA patients therefore remains difficult and sometimes controversial, even in multidisciplinary teams (Berber et al. 2016). In 2014, Kwon et al. published a risk-stratification algorithm for managing MoMHA patients. This consensus statement was devised by numerous expert US surgeons using their experience and the available evidence, as there was limited high-quality evidence to produce formal management guidelines. The proposed algorithm provided clinicians with infor-

Acta Orthopaedica 2019; 90 (6): 530–536

mation to risk stratify patients (low-, moderate-, and highrisk groups) according to their pre-revision factors, including implant, blood metal ions, and imaging (Kwon et al. 2014), given that often patients have pre-revision factors in different risk categories (Berber et al. 2016, Hussey et al. 2016). However, this algorithm was not comprehensive enough to account for all potential patient or surgical abnormalities. To our knowledge 1 study has attempted to risk stratify patients using this algorithm, with the authors observing that the blood metal ion risk stratification was useful for distinguishing between revised and unrevised patients with recalled articular surface replacement (ASR) devices (Hussey et al. 2016). How this algorithm performs in non-ASR implants is unknown. Furthermore, it is not known how this risk-stratification algorithm relates to outcomes following ARMD revision, with limited data on any prognostic factors of outcome following ARMD revision available (Matharu et al. 2018a). Such information would help define the threshold for recommending revision, and when counselling patients regarding the outcomes associated with further procedures. Given the risk-stratification algorithm proposed by Kwon et al. (2014) did not cover all potential patient or surgical abnormalities, we applied an adapted version of the algorithm to a large cohort of MoMHA patients who had all already undergone revision for ARMD at 2 centres. Using this adapted algorithm we assessed whether: (1) the revision threshold for ARMD changed over time, (2) the revision threshold differed between centres, and (3) whether patients at higher risk prior to revision had inferior outcomes following ARMD revision.

Patients and methods We performed a retrospective cohort study of prospectively collected data from 2 specialist UK arthroplasty centres (Nuffield Orthopaedic Centre, Oxford and the Royal Orthopaedic Hospital, Birmingham). The study included patients with large-diameter (36 mm or above) MoMHAs undergoing revision surgery for ARMD between January 2001 and March 2016. Cases were identified from prospectively maintained institutional databases described previously (Matharu et al. 2014, 2016b, 2016c, 2017a). This study was registered with each institution’s review board, with all patients reviewed according to institutional follow-up protocols. There were 346 revisions performed for ARMD, confirmed intraoperatively and histopathologically, which were eligible for this study. Comprehensive details of this cohort including the definitions for ARMD, preoperative investigations, intraoperative findings at revision, follow-up after revision surgery, and the outcomes following revision have been described (Matharu et al. 2019). Briefly, both centres were tertiary units with 16 surgeons performing all cases. All patients underwent clinical examination and radiographic

531

assessment (standardized anteroposterior pelvic radiographs +/– lateral hip radiograph), and most underwent blood cobalt and chromium ion sampling and cross-sectional imaging (ultrasound and/or metal artefact reduction sequence magnetic resonance imaging). The decision to perform revision surgery was made by the patient’s surgeon based on symptoms and/or investigative findings. After revision, patients were reviewed, usually annually, which included examination, radiographs, and completion of the Oxford Hip Score (OHS) questionnaire. Further investigations were performed in symptomatic patients (including those with new groin/ thigh pain, clicking/clunking, limping, instability), with these tests including bloods (inflammatory markers and metal ions) and cross-sectional imaging. The threshold for performing investigations in symptomatic patients following ARMD revision was considered on a case-by-case basis at the discretion of each surgeon, and where appropriate by the multidisciplinary team. The present study includes only the ARMD revision patients with both pre-revision blood metal ions and cross-sectional imaging available. We applied an adapted version of the algorithm proposed by Kwon et al. (2014) to stratify patients revised for ARMD into 3 risk groups (low, moderate, and high) based on their pre-revision factors, namely implant, radiographic, blood metal ions, and cross-sectional imaging. We used the same risk stratification for blood metal ions and cross-sectional imaging as proposed by Kwon et al. (2014). However, the risk stratification for implant and radiographic factors required adaptation from the original publication due to difficulties in applying the proposed algorithm. For implant factors, large-diameter modular THAs and recalled implants appeared in both the moderaterisk and high-risk stratification groups in the proposed algorithm. As the failure rates for large-diameter THAs are greater than those for hip resurfacing, and that recalled implants should be considered high risk, we assigned these 2 features to the high-risk category (Langton et al. 2011, Smith et al. 2012a, 2012b). For radiographic factors, the original algorithm did not fully define a suboptimally positioned acetabular component, as information on version was lacking. We considered acetabular components malpositioned if 1 or both parameters were outside the recommended optimal zone (inclination 35°–55° and anteversion 10°–30°) (Grammatopoulos et al. 2010). The adapted algorithm for implant, radiographic, blood metal ions, and cross-sectional imaging risk stratification used in the present study is summarised in Table 1. When assessing revision thresholds over time it was necessary to group patients by the year of ARMD revision. In 2012 the Medicines & Healthcare products Regulatory Agency (MHRA) issued a Medical Device Alert for all MoMHAs (MHRA 2012) in light of the high-profile reports of increased revision rates for both large-diameter modular THAs and hip resurfacing (Smith et al. 2012a, 2012b). Therefore patients were categorised as revised before 2012 or from January 2012 onwards.

532

Acta Orthopaedica 2019; 90 (6): 530–536

Table 1. Adaptation of the risk stratification of metal-on-metal hip arthroplasty patients originally proposed by Kwon et al. (2014) Pre-revision factor

Low risk

Moderate risk

Implant – Non-recalled hip resurfacing in – All other non-recalled hip men under 50 years with OA resurfacing implants Radiographic – Optimal acetabular component – Optimal acetabular component position a position a – No implant osteolysis or loosening – No implant osteolysis or loosening Blood metal ions b – Both under 3 ppb – Either or both between 3 and 10 ppb Cross-sectional – Within normal limits – ARMD without muscle/bone imaging involvement – Simple cystic lesions or small cystic lesions without thick walls

High risk – Large-diameter (≥ 36 mm) modular THA – Any recalled implants – Suboptimal acetabular component position a – Any implant osteolysis and/or loosening – Either or both above 10 ppb – ARMD with muscle/bone involvement – Solid or mixed lesions – Cystic lesions with thick walls

ARMD = adverse reactions to metal debris; OA = osteoarthritis; ppb = parts per billion; THA = total hip arthroplasty. Optimal position defined in methods section. Chromium and cobalt

a b

The 2 outcomes of interest following ARMD revision were: (1) re-revision surgery, and (2) a poor outcome. Re-revision surgery was defined as removal, exchange, or addition of any implant. A poor outcome was defined as 1 or more of the following: intraoperative complication, postoperative complication, further surgery or procedure (including re-revision), mortality within 90 days of surgery, and poor OHS (less than 27 out of 48) (Murray et al. 2007). Statistics The level set for statistical significance for all analyses was p < 0.05. For numerical data either the median and interquartile range (IQR), or the mean and standard deviation (SD) or range were used depending on data distribution. The effect of the implant, radiographic, blood metal ion, and cross-sectional imaging risk groups on (1) the year of revision, (2) the centre performing revision, and (3) post-revision outcomes (re-revision surgery and poor outcomes) were assessed using either the chi-square test or 2-sided Fisher’s exact test. The latter was used only when any cell had an expected frequency under 5. Implant survival analysis was performed using the Kaplan– Meier method using re-revision surgery as the endpoint. Patients not undergoing re-revision were censored at latest follow-up or death. Cox regression models including adjustment for differences in age and sex were used to examine the effect of the radiographic, blood metal ion, and cross-sectional imaging risk groups on the rates of re-revision surgery. These models were presented as hazard ratios with 95% confidence intervals (CI). The proportional hazards assumption was assessed using scaled Schoenfeld residuals and satisfied for all regression analyses. Ethics, funding, and potential conflicts of interest This study did not require ethical approval as all metal-onmetal hip arthroplasty patients were reviewed as part of each institution’s routine follow-up arrangements, which were

adapted in response to published recommendations from the United Kingdom MHRA, and following revision all patients were reviewed as per the standard institutional protocols. Therefore no patients were specifically recalled for the study. The study was funded by Arthritis Research UK (grant reference number 21006), the Royal Orthopaedic Hospital Hip Research and Education Charitable Fund, and the Orthopaedics Trust. This paper presents independent research funded/ supported by the National Institute for Health Research (NIHR) Leeds Biomedical Research Centre (BRC). The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care. GSM has received financial support for metal-on-metal hip research work from Arthritis Research UK, the Orthopaedics Trust, the Royal College of Surgeons of England, and the Royal Orthopaedic Hospital Hip Research and Education Charitable Fund. GSM has also received personal fees for undertaking medicolegal work on metal-on-metal hips for Leigh Day. FB and DJD receive institutional research funding from Smith & Nephew Orthopaedics UK. FB’s salary is paid by a grant from Smith & Nephew Orthopaedics UK. AJ has received consultancy fees from Freshfields Bruckhaus Deringer, and is a paid member of the data safety and monitoring board for Anthera Pharmaceuticals. DWM and HGP are paid consultants for Zimmer Biomet, and both receive institutional research funding from Zimmer Biomet. HGP is also a paid consultant for Kennedys Law, Bristol Myers Squibb, Depuy Synthes, Medacta Int, and Meril Life. He has received institutional research grants from UKIERI, Charnley Trust, Depuy Synthes, Glaxo Smith Kline, and NIHR.

Results There were 239 MoMHAs revised for ARMD with blood metal ions and cross-sectional imaging performed prior to revision surgery that were eligible for inclusion (Table 2).

Acta Orthopaedica 2019; 90 (6): 530–536

533

Table 2. Demographics of metal-on-metal hip arthroplasty patients undergoing revision surgery. Values are frequency (percentage) unless otherwise specified Pre-revision details

Cohort revised (n = 239)

Mean age at revision, years (SD) Female sex (%) Bilateral MoM hips any (%) Bilateral MoM hips revised for ARMD (%) Mean time to revision in years for ARMD (SD) Primary and revision centre same Primary implant type Hip resurfacing Total hip arthroplasty Primary implant design BHR Other (THR or HR) Conserve HR Corail-Pinnacle THR Synergy BHR THR Primary implant head size < 46 mm 46 mm > 46 mm Symptoms Local symptoms Systemic symptoms Blood metal ions: Median cobalt, µg/l (IQR) Median chromium, µg/l (IQR) Radiographs Mean cup inclination in degrees (SD) Mean cup version in degrees (SD) Cup malposition Stem/head malposition Loose cup Loose stem Lysis cup Lysis stem Neck thinning Impingement Heterotopic ossification Any cross-sectional imaging Any abnormality (% of those with imaging) Pseudotumors (PT) PT numbers PT consistency (% of all PT) Cystic Mixed Solid PT location (% of all PT) Anterior ± lateral Posterior ± lateral Anterior + posterior ± lateral Other Median PT volume, cm3 (IQR) Other image abnormalities Effusion Muscle atrophy/damage Tendon abnormality/damage Bursal distension/thickening

60 (11) 168 (70) 93 (39) 35 (15) 7 (3) 182 (76) 150 (63) 89 (37) 104 (44) 51 (21) 29 (12) 29 (12) 26 (11) 80 (42) 60 (31) 52 (27) 221 (93) 2 (0.8) 1.9 (0.7–8.0) 3.5 (1.6–8.3) 49 (11) 19 (10) 135 (57) 3 (1) 9 (4) 11 (5) 101 (42) 42 (18) 34 (14) 1 (0.4) 20 (8) 202 (85) 163 (68) 71 (44) 83 (52) 7 (4) 64 (40) 48 (30) 29 (18) 20 (12) 45 (13–130) 44 (18) 17 (7) 13 (5) 24 (10)

BHR = Birmingham Hip Resurfacing; HR = hip resurfacing; IQR = interquartile range; PT = pseudotumor; SD = standard deviation; THA = total hip arthroplasty. Note that micrograms per litre (µg/l) and parts per billion (ppb) are equivalent units of measure (see Table 1).

Table 3. Adapted risk-stratification group in relation to year of revision surgery. Values are frequency (percentage) Risk- Revised Revised stratification before 2012 2012 onwards group Overall (n = 58; 24%) (n = 181; 76%) p-value Implant 0.001 Low 11 (5) 3 (5) 8 (4) Moderate 138 (58) 45 (78) 93 (51) High 90 (38) 10 (17) 80 (44) Radiographic 0.01 Low/moderate 46 (19) 18 (31) 28 (16) High 193 (81) 40 (69) 153 (85) Blood metal ions 0.001 Low 86 (36) 11 (19) 75 (41) Moderate 92 (39) 23 (40) 69 (38) High 61 (25) 24 (41) 37 (20) Cross-sectional imaging 0.4 Low 43 (18) 7 (12) 36 (20) Moderate 106 (44) 26 (45) 80 (44) High 90 (38) 25 (43) 65 (36)

Revision thresholds over time Of the 239 hips, 181 (76%) were revised from 2012 onwards and 58 (24%) were revised before 2012. Hips revised from 2012 onwards were more likely to have high-risk implants (high-risk implants 44% from 2012 onwards vs. 17% pre2012; p = 0.001 for chi-square test of year of surgery vs. implant risk category), high-risk radiographic features (85% vs. 69% pre-2012; p = 0.01), and low-risk blood metal ions (41% vs. 19% pre-2012; p = 0.001) (Table 3). A statistically significant difference could not be demonstrated in the crosssectional imaging risk of hips revised from 2012 onwards compared with before 2012 (p = 0.4). Revision thresholds at different centres Of the 239 hips, 166 (70%) were revised at centre 1, and 73 (30%) were revised at centre 2. Hips revised at centre 1 were significantly more likely to have high-risk implants (48% vs. 14%; p < 0.001) and low-risk blood metal ions (45% vs. 15%; p < 0.001) compared with centre 2 (Table 4, see Supplementary data). The radiographic risk (p = 0.1) and cross-sectional imaging risk (p = 0.5) of hips revised for ARMD were not statistically significantly different between the 2 centres. Effect of risk stratification on outcomes after ARMD revision Mean follow-up after revision was 5 years (1–16). During follow-up 22 hips (9%) needed re-revision surgery for any indication, and 92 hips (39%) had a poor outcome. The cumulative implant survival following ARMD revision at 5 years and 7 years was 89% (CI 83–93; 65 hips at risk) and 86% (CI 75–92; 23 hips at risk) respectively. A difference could not be demonstrated between the various risk categories (implant, radiographic, blood metal ion,

534

and cross-sectional imaging risk) and the frequency of rerevision surgery or the frequency of poor outcomes (Table 5, see Supplementary data). In addition, there was no statistically significant difference in the hazard ratios between the various risk categories (implant, radiographic, blood metal ion, and cross-sectional imaging risk) and the rates of re-revision surgery (Table 6, see Supplementary data).

Discussion We applied an adapted version of the current risk-stratification algorithm (Kwon el al. 2014) to a large cohort of MoMHA patients revised for ARMD at 2 tertiary European centres over a 15-year period. There was evidence that the threshold for performing revision surgery for ARMD has changed over time but also differed between centres. However, with the limited study sample available we found no evidence that patients considered high risk pre-revision subsequently experienced worse outcomes following ARMD revision surgery compared with moderate-risk and low-risk patients. We observed that revisions performed from 2012 onwards were more likely to include high-risk implants, high-risk radiographic features, and low-risk blood metal ions. However, cross-sectional imaging risk was similar before and after 2012. Differences we observed between the revision thresholds used over time are likely to reflect increasing evidence, patient surveillance, and investigation in more recent years. The 2012 MHRA alert (MHRA 2012) and registry studies highlighting increased revision rates for MoMHAs (Smith et al. 2012a, 2012b) changed how these patients were managed worldwide, with evidence that revision rates have increased since regular surveillance was recommended (Matharu et al. 2016c, 2017b, 2018b). Around 2012 was also the time when it was widely recognised that large-diameter MoM THAs had universally high revision rates, as prior to this the problems reported were mainly in hip resurfacings (Pandit et al. 2008, Grammatopoulos et al. 2009). These large-diameter THAs are a high-risk group in the adapted algorithm, thus explaining why we observed more high-risk implants revised since 2012. As the evidence evolved, the importance of optimal acetabular orientation for MoMHA success was recognised (Grammatopoulos et al. 2010). This again explains why more high-risk radiographic features were seen from 2012, with most high-risk features in our series being due to suboptimal acetabular orientation rather than other adverse radiographic features. As the understanding of blood metal ions improved, some surgeons gradually started revising symptomatic MoMHA patients with lower blood metal ions on the premise that early revision would improve subsequent outcomes (Grammatopoulos et al. 2009, De Smet et al. 2011). We observed similar findings with lower blood metal ion thresholds used from 2012. The revision threshold varied between centres, with 1 centre more commonly revising high-risk implants and low-

Acta Orthopaedica 2019; 90 (6): 530â&#x20AC;&#x201C;536

risk blood metal ions compared with the other. Difference in implant risk reflects 1 centre performing high numbers of large-diameter MoM THA (high risk in the adapted algorithm) in addition to hip resurfacing (Matharu et al. 2016b, 2017b), whilst the other centre exclusively performed resurfacing using non-recalled designs (low risk and moderate risk) (Matharu et al. 2016c). Our observation that centres attached different importance to metal ions when managing patients ultimately requiring revision is supported by previous findings. 10 MoMHA clinical scenarios were used to examine how multidisciplinary teams from 6 experienced international centres managed problematic MoMHA patients, with agreement being inconsistent when patients had raised or rising blood metal ions (Berber et al. 2016). This is not surprising given many studies have proposed that a variety of different blood metal ion thresholds below the MHRA recommended 7 parts per billion (ppb) limit are better for managing MoMHA patients (Hart et al. 2011, Van Der Straeten et al. 2013). Furthermore, recent evidence established that the primary function of blood metal ions was for identifying patients at low risk of ARMD rather than for diagnosing ARMD (Matharu et al. 2016a, 2016b, 2017a). There is also great variability in the blood metal ion concentrations of MoMHAs revised for ARMD at different centres (De Smet et al. 2011, Liddle et al. 2013, Pritchett 2014), which supports our findings. 1 study reported pre-revision blood metal ions from as low as a median of 4 ppb (De Smet et al. 2011), whilst another centre reported ions ranging between 17 and 136 ppb in ARMD revisions (Pritchett 2014). With the limited study sample available and the relatively low number of re-revisions, the implant, radiographic, blood metal ion, and cross-sectional imaging risk groups did not statistically significantly affect outcomes following ARMD revision in our study. It is intuitive that MoMHA patients with grossly raised blood metal ions and those with substantial osteolysis and/or tissue destruction on imaging should have inferior outcomes following revision compared with patients without such features, especially given the poor outcomes reported after the early ARMD revisions (Grammatopoulos et al. 2009, De Smet et al. 2011). Although some single-centre studies have identified predictors of poor outcomes following ARMD revision, such as solid ARMD lesions, these studies were small and underpowered (De Smet et al. 2011, Liddle et al. 2013, Matharu et al. 2014). Our larger study, which used an adapted risk-stratification algorithm proposed by expert surgeons, suggests that patients considered high risk pre-revision do not necessarily have inferior outcomes following ARMD revision compared with individuals with lower pre-revision risk. However, given the original algorithm required adaptation we question its clinical utility, and we also acknowledge that our modified algorithm was not perfect. Therefore we recommend further research to develop clinically useful algorithms for managing patients with problematic MoMHAs and also to inform thresholds for recommending revision surgery

Acta Orthopaedica 2019; 90 (6): 530–536

in MoMHA patients, if indeed such thresholds truly exist. In the interim it is recommended that surgeons continue to make decisions on an individual case basis and by using the best available evidence. This study has limitations. Its retrospective nature may introduce potential bias, for example when assessing the cross-sectional imaging reports; however, undertaking a prospective study to answer the same questions would take many years. We acknowledge that focusing on patients with ions and imaging is a limitation; however, this was inevitable given the retrospective nature of the study and that the diagnosis and investigation of ARMD evolved over time (Grammatopoulos et al. 2009, De Smet et al. 2011). It was necessary to modify some of the initially proposed risk categories, such as implant factors, given the potential overlap between risk groups. This may be considered a limitation. However, it would otherwise have been impossible to apply the algorithm clinically given the original algorithm did not comprehensively cover all possibilities, and the original authors did recognise that their algorithm would evolve over time (Kwon et al. 2014). Similarly, it was not possible to assign each patient to 1 global risk category given they often had pre-revision factors in different risk groups. This limitation of the algorithm was recognised previously (Hussey et al. 2016), and we propose that more information was available by assessing each pre-revision category (implant, radiographic, blood metal ions, and cross-sectional imaging) separately. Patients were grouped by year of revision because there were not enough patients undergoing surgery each calendar year for meaningful analysis. Although this may obscure some detail regarding when changes occurred over time, such analysis would only be possible with registry data, which would lack most of the prerevision data (ions and imaging). Some of the proportions/frequencies assessed varied between the different risk subgroups even though a statistically significant difference could not be demonstrated (Tables 3–5). It is acknowledged this may be a reflection that the sample size and/or number of events was too low to detect a difference. Finally, our findings apply only to MoMHA patients revised for ARMD, and not to asymptomatic patients with MoMHAs given we did not include such a comparator. It was never the intention to include a non-revised patient group given the study aims, and a previous study has already assessed the original algorithm in non-revised patients (Hussey et al. 2016). However, it is important to acknowledge that without applying the risk-stratification groups to patients who did not undergo revision surgery, it is not possible to conclusively demonstrate that the threshold for revision surgery changed over time. In summary, when applying the adapted risk-stratification algorithm we found the threshold for revision surgery changed over time, which is likely to reflect the increasing evidence, patient surveillance, and investigation of MoMHA patients since 2012. Although lower blood metal ion thresholds have been used since 2012 for ARMD revisions, the centres stud-

535

ied attached different importance to metal ions when managing patients, which is consistent with previous findings. With the sample size available we found no evidence that patients considered high risk pre-revision experienced worse outcomes following ARMD revision surgery compared with moderate-risk and low-risk patients. Therefore further research is required to inform thresholds for recommending revision surgery in MoMHA patients. Supplementary data Tables 4–6 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674. 2019.1659661

All authors were involved with the study design, data collection, interpretation, manuscript revision, and approval. GSM and AJ analysed the data. DJD and DWM performed the surgeries. GSM wrote the manuscript draft and is the guarantor of the data. The authors would also like to thank Jo Brown and Lesley Brash who have collected outcome data on patients reported in this study. Acta thanks Pepijn Bisseling and Aleksi Reito for help with peer review of this study.

Berber R, Skinner J, Board T, Kendoff D, Eskelinen A, Kwon Y M, Padgett D E, Hart A, ISCCoMh. International metal-on-metal multidisciplinary teams: do we manage patients with metal-on-metal hip arthroplasty in the same way? An analysis from the International Specialist Centre Collaboration on MOM Hips (ISCCoMH). Bone Joint J 2016; 98-B (2): 179-86. De Smet K A, Van Der Straeten C, Van Orsouw M, Doubi R, Backers K, Grammatopoulos G. Revisions of metal-on-metal hip resurfacing: lessons learned and improved outcome. Orthop Clin North Am 2011; 42 (2): 25969, ix. FDA (U.S. Food and Drug Administration). Medical Devices. Metal-on-metal hip implants. Information for orthopaedic surgeons; 2013. http://www.fda. gov/MedicalDevices/ProductsandMedicalProcedures/ImplantsandProsthetics/MetalonMetalHipImplants/ucm241667.htm. Grammatopoulos G, Pandit H, Kwon Y M, Gundle R, McLardy-Smith P, Beard D J, Murray D W, Gill H S. Hip resurfacings revised for inflammatory pseudotumour have a poor outcome. J Bone Joint Surg Br 2009; 91 (8): 1019-24. Grammatopoulos G, Pandit H, Glyn-Jones S, McLardy-Smith P, Gundle R, Whitwell D, Gill H S, Murray D W. Optimal acetabular orientation for hip resurfacing. J Bone Joint Surg Br 2010; 92 (8): 1072-8. Hart A J, Sabah S A, Bandi A S, Maggiore P, Tarassoli P, Sampson B, J A S. Sensitivity and specificity of blood cobalt and chromium metal ions for predicting failure of metal-on-metal hip replacement. J Bone Joint Surg Br 2011; 93 (10): 1308-13. Hussey D K, Madanat R, Donahue G S, Rolfson O, Bragdon C R, Muratoglu O K, Malchau H. Scoring the current risk stratification guidelines in follow-up evaluation of patients after metal-on-metal hip arthroplasty: a proposal for a metal-on-metal risk score supporting clinical decision-making. J Bone Joint Surg Am 2016; 98 (22): 1905-12. Kwon Y M, Lombardi A V, Jacobs J J, Fehring T K, Lewis C G, Cabanela M E. Risk stratification algorithm for management of patients with metal-onmetal hip arthroplasty: consensus statement of the American Association of Hip and Knee Surgeons, the American Academy of Orthopaedic Surgeons, and the Hip Society. J Bone Joint Surg Am 2014; 96 (1): e4.

536

Langton D J, Jameson S S, Joyce T J, Gandhi J N, Sidaginamale R, Mereddy P, Lord J, Nargol A V. Accelerating failure rate of the ASR total hip replacement. J Bone Joint Surg Br 2011; 93 (8): 1011-16. Liddle A D, Satchithananda K, Henckel J, Sabah S A, Vipulendran K V, Lewis A, Skinner J A, Mitchell A W, Hart A J. Revision of metal-on-metal hip arthroplasty in a tertiary center: a prospective study of 39 hips with between 1 and 4 years of follow-up. Acta Orthop 2013; 84 (3): 237-45. Matharu G S, Pynsent P B, Sumathi V P, Mittal S, Buckley C D, Dunlop D J, Revell P A, Revell M P. Predictors of time to revision and clinical outcomes following revision of metal-on-metal hip replacements for adverse reaction to metal debris. Bone Joint J 2014; 96-B (12): 1600-9. Matharu G S, Berryman F, Brash L, Pynsent P B, Dunlop D J, Treacy R B. Can blood metal ion levels be used to identify patients with bilateral Birmingham Hip Resurfacings who are at risk of adverse reactions to metal debris? Bone Joint J 2016a; 98-B (11): 1455-62. Matharu G S, Berryman F, Brash L, Pynsent P B, Treacy R B, Dunlop D J. the effectiveness of blood metal ions in identifying patients with unilateral Birmingham Hip Resurfacing and Corail-Pinnacle metal-on-metal hip implants at risk of adverse reactions to metal debris. J Bone Joint Surg Am 2016b; 98 (8): 617-26. Matharu G S, Judge A, Murray D W, Pandit H G. Prevalence of and risk factors for hip resurfacing revision: a cohort study into the second decade after the operation. J Bone Joint Surg Am 2016c; 98 (17): 1444-52. Matharu G S, Berryman F, Judge A, Reito A, McConnell J, Lainiala O, Young S, Eskelinen A, Pandit H G, Murray D W. Blood metal ion thresholds to identify patients with metal-on-metal hip implants at risk of adverse reactions to metal debris: an external multicenter validation study of Birmingham Hip Resurfacing and Corail-Pinnacle implants. J Bone Joint Surg Am 2017a; 99 (18): 1532-39. Matharu G S, Nandra R S, Berryman F, Judge A, Pynsent P B, Dunlop D J. Risk factors for failure of the 36 mm metal-on-metal Pinnacle total hip arthroplasty system: a retrospective single-centre cohort study. Bone Joint J 2017b; 99-B (5): 592-600. Matharu G S, Eskelinen A, Judge A, Pandit H G, Murray D W. Revision surgery of metal-on-metal hip arthroplasties for adverse reactions to metal debris. Acta Orthop 2018a; 89 (3): 278-88. Matharu G S, Hunt L P, Murray D W, Howard P, Pandit H G, Blom A W, Bolland B, Judge A. Is the rate of revision of 36 mm metal-on-metal total hip arthroplasties with Pinnacle acetabular components related to the year of

Acta Orthopaedica 2019; 90 (6): 530â&#x20AC;&#x201C;536

the initial operation? An interrupted time-series analysis using data from the National Joint Registry for England and Wales. Bone Joint J 2018b; 100-B (1): 33-41. Matharu G S, Judge A, Pandit H G, Murray D W. Follow-up for patients with metal-on-metal hip replacements: are the new MHRA recommendations justified? BMJ 2018c; 360: k566. Matharu G S, Berryman F, Dunlop D J, Revell M P, Judge A, Murray D W, Pandit H G. No threshold exists for recommending revision surgery in metal-onmetal hip arthroplasty patients with adverse reactions to metal debris: a retrospective cohort study of 346 revisions. J Arthroplasty 2019; 34 (7): 1483-1491. MHRA. Medical device alert: MDA/2017/018: all metal-on-metal (MoM) hip replacementsâ&#x20AC;&#x201D;updated advice for follow-up of patients; June 29, 2017. https://www.gov.uk/drug-device-alerts/all-metal-on-metal-mom-hipreplacements-updated-advice-for-follow-up-of-patients. MHRA. Medicines & Healthcare products Regulatory Agency (MHRA). Medical Device Alert: all metal-on-metal (MoM) hip replacements. MDA/ 2012/036; 2012. http://www.mhra.gov.uk/. Murray D W, Fitzpatrick R, Rogers K, Pandit H, Beard D J, Carr A J, Dawson J. The use of the Oxford hip and knee scores. J Bone Joint Surg Br 2007; 89 (8): 1010-4. Pandit H, Glyn-Jones S, McLardy-Smith P, Gundle R, Whitwell D, Gibbons C L, Ostlere S, Athanasou N, Gill H S, Murray D W. Pseudotumours associated with metal-on-metal hip resurfacings. J Bone Joint Surg Br 2008; 90 (7): 847-51. Pritchett J W. One-component revision of failed hip resurfacing from adverse reaction to metal wear debris. J Arthroplasty 2014; 29 (1): 219-24. Smith A J, Dieppe P, Howard P W, Blom A W, National Joint Registry for England and Wales. Failure rates of metal-on-metal hip resurfacings: analysis of data from the National Joint Registry for England and Wales. Lancet 2012a; 380 (9855): 1759-66. Smith A J, Dieppe P, Vernon K, Porter M, Blom A W, National Joint Registry of England and Wales. Failure rates of stemmed metal-on-metal hip replacements: analysis of data from the National Joint Registry of England and Wales. Lancet 2012b; 379 (9822): 1199-204. Van Der Straeten C, Grammatopoulos G, Gill H S, Calistri A, Campbell P, De Smet K A. The 2012 Otto Aufranc Award: The interpretation of metal ion levels in unilateral and bilateral hip resurfacing. Clin Orthop Relat Res 2013; 471 (2): 377-85.

Acta Orthopaedica 2019; 90 (6): 537–541

537

Posterior and anterior tilt increases the risk of failure after internal fixation of Garden I and II femoral neck fracture Pontus SJÖHOLM 1, Volker OTTEN 1, Olof WOLF 2, Max GORDON 3, Gustav KARSTEN 1, Olof SKÖLDENBERG 3, and Sebastian MUKKA 1 1 Department of Surgical and Perioperative Sciences at Umeå University, Umeå; 2 Section of Orthopaedics, Department of Surgical Sciences, Uppsala University, Uppsala; 3 Departments of Orthopedics and Clinical Sciences at Danderyd Hospital, Karolinska Institute, Stockholm, Sweden Correspondence: pontus.sjoholm@umu.se Submitted 2019-03-01. Accepted 2019-06-13.

Background and purpose — Preoperative posterior tilt of the femoral head as seen on lateral radiographs has been reported to affect the risk of fixation failure in cases of minimally displaced femoral neck fractures (Garden I–II). We investigated radiological risk factors of treatment failure. Patients and methods — We included 417 patients (68% women, median age: 78 years (50–108) with a minimally displaced femoral neck fracture (Garden I–II) treated with internal fixation in a retrospective cohort study. The patients were followed for 3.4 years (2–14). Data on age, sex, housing, cognitive impairment, implant angulation, pre- and postoperative tilt, hip complications, and reoperations were recorded. The risk of fixation failure was assessed using Cox proportional hazards regression analysis. Results — The overall reoperation rate was 17%, and the rate of treatment failure (fixation failure, nonunion, avascular necrosis, or posttraumatic osteoarthritis) was 13%. Cox proportional hazard analysis revealed an increased risk of treatment failure with a preoperative posterior tilt of at least 20° and a preoperative anterior tilt greater than 10°. A failure occurred in 13 of the 65 patients with a posterior tilt of at least 20° and in 5 of the 9 patients with an anterior tilt greater than 10°. Interpretation — A preoperative posterior tilt of 20° and an anterior tilt greater than 10° in cases of Garden I and II femoral neck fractures increase the risk of fixation failure necessitating additional surgery. In this group of patients, there is a need for future interventional studies regarding the feasibility of primary hip arthroplasty.

Several authors have raised doubts regarding the results of the internal fixation of minimally displaced femoral neck fractures (FNFs) (Rogmark et al. 2009, Gjertsen et al. 2011). In elderly patients, reoperation rates ranging from 8% to 19% have been reported (Onativia et al. 2018). Several authors have proposed that preoperative posterior tilt of the femoral head increases the risk of reoperation after internal fixation of minimally displaced FNFs (Palm et al. 2009, Clement et al. 2013, Dolatowski et al. 2016). However, Lapidus et al. (2013) were unable to reproduce these findings. Surgeons need a reliable predictor that can be used to identify patients who are at risk of treatment failure after internal fixation. We further investigated the value of posterior or anterior tilt and comorbidities for predicting fixation failure and avascular necrosis (AVN) after internal fixation of minimally displaced FNFs. We hypothesized that the risk of failure would increase with increasing preoperative tilt of the fracture.

Patients and methods Study design and setting This retrospective cohort study included patients treated with closed reduction and internal fixation of minimally displaced FNFs between 2003 and 2015 at the Orthopaedic Department of Umeå University Hospital, Sweden. Umeå University Hospital is a third-level university hospital with a catchment area of emergency care for approximately 160,000 inhabitants. The STROBE guidelines were followed. Participants and data collection We included all patients with an age greater than 50 years and acute minimally displaced FNF (Garden I–II) treated with

538

internal fixation between 2003 and 2015 in our department. Patients were followed until February 2019 or until death. Data were collected retrospectively throughout the study period by a combination of in-hospital surgical and medical records at admission. Patient data included age, sex, cognitive impairment (yes/no, classified by the treating surgeon; temporary confusion was not classified as cognitive impairment), use of a walking aid, sheltered living (yes/no), surgical treatment, and date of death. Radiographic analysis The anteroposterior (AP) radiographs were used to classify fractures according to the Garden system as minimally displaced (Garden I–II) or displaced (Garden III–IV) (Nilsson et al. 1993). The preoperative tilt angle of the femoral head was measured on a lateral radiograph of the hip using the method described by Palm et al. (2009). Implant positioning on the postoperative AP pelvic radiograph was divided into 2 categories (inclination of ≤ 125° and > 125°). All images were digitally acquired using the Picture Archiving and Communication System (PACS, Impax, Agfa, Antwerp, Belgium), and all measurements were performed by PS who was not blinded to the outcome. Implants and surgery Internal fixation was carried out with the patient on a fracture table. Before sterile draping, the fracture underwent closed reduction if angulated and amenable to closed reduction with the aid of an image intensifier. Internal fixation was performed with 2 pins (Hansson Pins, Swemac, Sweden). On the AP projection, the caudal pin was aimed to extend from the level of the lesser trochanter laterally along the medial inferior cortex of the femoral neck to approximately 1 cm from the border between the bone and cartilage of the head. The cranial pin was positioned parallel at 1 to 2 cm dorsoproximal from the distal pin. The pins were parallel and positioned in the central or posterior third of the femoral head and neck. Neither capsulotomy nor joint aspiration was performed. Antibiotic prophylaxis with cloxacillin was given on the day of surgery. Low-molecular-weight heparin was administered postoperatively for 14 days. The aim was to mobilize all patients on the first postoperative day under the supervision of a physiotherapist. Outcomes Treatment failure was defined as fixation failure, nonunion, avascular necrosis (AVN), or posttraumatic osteoarthritis. Major reoperation was defined as hip replacement or excision arthroplasty. Minor reoperation was defined as pin removal. Complications were defined as all registered complications related to the hip, i.e., fixation failure, nonunion, AVN, posttraumatic osteoarthritis, and peri-implant fracture.

Acta Orthopaedica 2019; 90 (6): 537–541

Statistics We used a Cox proportional hazards regression model with time since primary surgery as the dependent variable with time to whichever of the following events first occurred: reoperation, death, or uneventful end of the observational time. We evaluated the proportional hazards assumption using Schoenfeld residuals. In addition to the exposure variable (posterior and anterior tilt) we selected age, sex, sheltered housing or nursing home, and cognitive impairment as covariates for fragility. As surgical factors may impact the risk for failure, we also included the inclination angle (Nyholm et al. 2018) and postoperative tilt of the pins to adjust for surgical factors that might affect the risk of failure. We performed an unadjusted relative risk calculation. Pre- and postoperative tilt, implant angulation, and age were modeled as continuous variables where each was tested for non-linearity using ANOVA, and, if significant, they were modeled using restricted cubic splines, where the number of knots, which determines flexibility of the spline, was chosen using the Bayesian information criterion (BIC), also known as Schwarz’s information criterion. We used R version 3.5.2 in combination with the rms package (v. 5.1–2; R Foundation for Statistical Computing, Vienna, Austria) for the survival analysis. Ethics, data sharing, funding, and potential conflicts of interest The study was conducted in accordance with the ethical principles of the Helsinki Declaration and was approved by the Ethics Committee of Umeå University (entry number dnr 2017-489-32M and 2016-447-31M). Data will be made available on request to the corresponding author. The study was funded by grants from the regional agreement on medical training and clinical research (ALF) between Västerbotten County Council and Umeå University. The authors declare no competing interests.

Results Patients and descriptive data 417 patients (median age: 78 years [50–108] with a mean follow-up of 3.4 years [2–14]), 68% of whom were women, were included in the study (Figure 1, Table 1). The mortality rate was high, with an overall 1-year mortality rate of 20% and a 2-year mortality rate of 33%. Complications The overall complication rate during the study period was 14.5%, including 4% fixation failure, 2% nonunion, 6% AVN, 0.5% posttraumatic osteoarthritis, and 2% peri-implant fracture. All fixation failures occurred during the first year, all cases of AVN and the 2 cases of posttraumatic osteoarthritis occurred during the second year or later.

Acta Orthopaedica 2019; 90 (6): 537–541

539

Count

Undisplaced femoral neck fractures between 2003 and 2015 n = 560

100

Hazard ratio

No failure Failure

Excluded – other treatment (n = 111): – hemiarthroplasty, 64 – total hip arthroplasty, 12 – twin hook, 34 – conservative treatment, 1 Treated with internal fixation n = 439

2.2

1.8 60

1.5 1.2

Excluded (n = 22): – missing preoperative lateral radiograph, 8 – poor radiographic quality, 2 – contralateral femoral neck fracture, 12

1.0 20

0.82 0

Preoperative lateral radiographs analyzed (n = 417)

–20°

0°

20°

Figure 2. Diagram displaying the distribution of patients and treatment failure (n = 417).

Preoperative tilt: < –10°, 9 (2%) –10° to 20°, 348 (83%) > 20°, 60 (14%)

Figure 1. Flowchart of patient selection.

Table 1. Characteristics of patients (n = 417) Age a b Women Cognitive impairment b Sheltered housing b Preoperative tilt a Missing b Postoperative tilt a Missing b Implant angulation a ≤ 125° b > 125° b Missing b

40°

Dorsal tilt

78 (10) 285 (68) 165 (40) 153 (37) 10° (10) 0 5° (8) 5 (4) 141° (7) 7 (2) 399 (96) 11 (3)

a Mean (SD). b n (%).

–10°

0°

10°

20°

30°

Dorsal tilt

Figure 3. Plot of the risk of treatment failure. Anterior tilt corresponds to negative values on the X-axis. Adjusted for sex, sheltered housing, cognitive impairment, implant angulation, and postoperative tilt. Light blue area corresponds to 95% CI.

Table 2. Treatment failures and reoperations

Age a Women b Failure b Type of failure, n Fixation failure Nonunion Avascular necrosis (AVN) Posttraumatic osteoarthritis Reoperation b Type of reoperation, n Removal of osteosynthesis Hemiarthroplasty Total hip arthroplasty Revision osteosynthesis Peri-implant fracture Girdlestone

< –10° (n = 9)

–10° to 20° (n = 348)

> 20° (n = 60)

80 (4) 8 (89) 5 (56)

78 (10) 242 (70) 35 (10)

78 (10) 35 (58) 13 (22)

2 1 2 0 5 (56) 0 3 2 0 0 0

8 7 18 2 49 (14) 16 6 17 2 6 2

6 2 5 0 15 (25) 3 4 7 0 1 0

a Mean (SD). b n (%).

Reoperations 17% of all patients underwent reoperation, including minor procedures, such as screw removal, during the study period (Figure 2 , Table 2). 2 patients with nonunion and 2 patients with AVN did not undergo surgery due to their general health status (Table 2).

Discussion

Risk of treatment failure 13% of patients underwent major reoperation and/or were diagnosed with fixation failure, nonunion, AVN, or posttraumatic osteoarthritis. The relative risk (RR) of treatment failure for patients with a preoperative tilt of more than 20° was 2.2 (95% CI 1.2–3.8) and the RR for an anterior tilt of greater than –10° was 5.5 (CI 2.8–11) compared with patients with a preoperative tilt of –10° to 20° (Figure 3, Table 3).

In this retrospective cohort study, we found that the risk of treatment failure was higher for the one-fifth of patients with a preoperative anterior tilt of at least 10° or posterior tilt of at least 20°. We suggest that a thorough preoperative assessment of the lateral radiograph be performed in cases of a minimally displaced FNF in elderly patients to define who should be considered for arthroplasty rather than internal fixation. The impact of posterior tilt has been debated previously, and studies have presented various results (Palm et al. 2009, Clem-

540

Acta Orthopaedica 2019; 90 (6): 537–541

Table 3. Cox proportional hazard model for covariates associated with treatment failure adjusted for sex, sheltered housing, cognitive impairment, implant angulation, and postoperative tilt Variable Age Sex Female Male Sheltered housing: No Yes Cognitive impairment: No Yes Implant angulation Preoperative tilt: –20° –10° 0° 10° 20° 30° 40° Postoperative tilt

Crude HR (2.5%–97.5%)

Adjusted HR (2.5%–97.5%)

1.0 (1.0–1.0) 1.0 Ref. 0.7 (0.3–1.3) 1.0 Ref. 0.5 (0.2–1.0) 1.0 Ref. 0.7 (0.4–1.4) 1.0 (1.0–1.0 5.8 (1.7–19) 2.9 (1.4–6.3) 1.5 (1.1–2.1) 1.0 Ref. 1.5 (1.2–2.0) 3.3 (1.7–6.4) 7.0 (2.4–21) 1.0 (1.0–1.0)

1.0 (1.0–1.1) 1.0 Ref. 0.6 (0.3–1.3) 1.0 Ref. 0.4 (0.2–1.1) 1.0 Ref. 0.9 (0.4–2.0) 1.0 (1.0–1.0) 7.9 (1.8–35) 3.6 (1.4–9.3) 1.6 (1.1–2.5) 1.0 Ref. 1.5 (1.1–2.1) 3.4 (1.6–7.3) 7.5 (2.2–25) 1.0 (1.0–1.1)

HR = hazard ratio; n = 417.

ent et al. 2013, Lapidus et al. 2013, Dolatowski et al. 2016). The differences among the conclusions might be attributed to differences in the categorization of measurements, definition of outcomes and reoperation, number of included patients, and follow-up durations. Among patients with a posterior tilt exceeding 20°, Palm et al. (2009), Lapidus et al. (2013), and Dolatowski et al. (2016) reported a reoperation rate of 56%, 10%, and 19%, respectively. In concordance with the report by Dolatowski et al., we found a similar failure rate. Palm et al. (2009) included 3 peri-implant fractures as failures; in contrast, Lapidus et al. (2013) analyzed reoperation rates and excluded 5 patients for whom revision surgery was indicated but not performed due to medical comorbidities. Clement et al. (2013) included screw removal due to local discomfort, thus increasing the failure rate. An anterior tilt of at least 10° has not been described and published in the peer-reviewed literature as a risk factor of treatment failure, although this phenomenon has been proposed in a book discussing internal fixation for FNFs by Manninger and coauthors (2007). We found, somewhat surprisingly, that a posterior tilt up to 10 degrees gave the lowest risk for failure. This might be due to an increase in compression and stability at the fracture site. Fracture reduction did not decrease the risk of treatment failure, which is in agreement with the results previously reported (Palm et al. 2009, Dolatowski et al. 2016). A more pronounced tilt might cause a predisposition to greater instability due to comminution of calcar femorale and thus lead to an increased risk of disrupted healing tendency for redisplacement and mechanical failure (Alho et al. 1992).

In a recent study by Nyholm et al. (2018), the risk factors of reoperation were analyzed in patients undergoing osteosynthesis for a Garden I–II or Garden III–IV FNF with parallel implants. The authors concluded that insufficient reduction, varus implant positions (≤ 125°), and femoral head cartilage perforation were the only surgical factors influencing the risk of reoperation. Thus, implant placement in minimally displaced femoral neck fractures might be of less importance than previously suggested. In our study, we found no statistically significant correlation between the risk of reoperation and postoperative tilt or varus implant positions. Hip arthroplasty is the treatment of choice for elderly patients with displaced FNFs (Rogmark and Leonardsson 2016). In elderly patients, failed internal fixation necessitating a revision hip arthroplasty is a severe complication. Also, salvage arthroplasty following failed internal fixation has inferior outcomes compared with primary hip arthroplasty (Blomfeldt et al. 2006, Frihagen et al. 2007). In groups with a high risk of fixation failure, such as patients with an anteriorly or posteriorly tilted FNF (Garden I–II), hip arthroplasty is a feasible option that could improve the surgical outcome. In a recently published randomized controlled trial comparing hemiarthroplasty with screw fixation for Garden I–II FNFs, the authors found similar hip function, as measured by the Harris hip score (Dolatowski et al. 2019). However, regarding secondary outcomes, hemiarthroplasty led to improved mobility and fewer major reoperations. The authors also performed a post-hoc analysis of those patients allocated to internal fixation and found that those with a preoperative posterior tilt of more than 20° had a higher risk of healing-related complications compared with those with a preoperative tilt of 20° or less. Further large register-based randomized controlled trials focusing on high-risk groups are of interest for establishing whether the reduced reoperation rate and improved mobility lead to improved patient-reported outcomes. Inherent flaws of the retrospective observational study design and the lack of regular or scheduled follow-ups limit the impact of our study. There are failures that were not identified due to the lack of scheduled follow-ups and would consequently increase the failure rate. We found a tendency towards the risk of treatment failure being lower for patients living in sheltered housing, most likely explained by the high mortality rate and the fragility of patients unfit to actively seek healthcare services. This reduces the identification of failure and masks the breadth of issues related to the internal fixation of minimally displaced FNFs. All measurements were performed by 1 rater. We did not perform any validation of the measurements; however, the inter- and intra-rater reliability of posterior tilt measurements in cases of minimally displaced FNFs has previously been reported to be excellent although the repeatability and agreement by the minimal detectable change were found to be 14° (Palm et al. 2009, Dolatowski and Hoelsbrekken 2017).

Acta Orthopaedica 2019; 90 (6): 537–541

In summary, patients with a minimally displaced FNF (Garden I–II) with a preoperative posterior tilt of more than 20° or an anterior tilt greater than 10° have an increased risk of fixation failure necessitating a salvage procedure. Primary hip arthroplasty is a feasible option for this group of patients. The authors thank Olle Svensson at the Section of Orthopedics, Department of Surgical and Perioperative Sciences at Umeå University, for his work with Umanhip. PS collected and analyzed data, and wrote the manuscript. VO, OW, and OS wrote the manuscript. MG performed the statistical analysis and wrote the manuscript. GK collected data and reviewed the manuscript. SM initiated and supervised the study, collected data, performed the statistical analysis, and wrote the manuscript. Acta thanks Filip C Dolatowski, Frede Frihagen and Jan-Erik Gjertsen for help with peer review of this study.

Alho A, Benterud J G, Rønningen H, Høiseth A. Prediction of disturbed healing in femoral neck fracture: radiographic analysis of 149 cases. Acta Orthop Scand 1992; 63(6): 639-44. Blomfeldt R, Tornkvist H, Ponzer S, Soderqvist A, Tidermark J. Displaced femoral neck fracture: comparison of primary total hip replacement with secondary replacement after failed internal fixation: a 2-year follow-up of 84 patients. Acta Orthop 2006; 77(4): 638-43. Clement N D, Green K, Murray N, Duckworth A D, McQueen M M, CourtBrown C M. Undisplaced intracapsular hip fractures in the elderly: predicting fixation failure and mortality. A prospective study of 162 patients. J Orthop Sci 2013; 18(4): 578-85. Dolatowski F C, Hoelsbrekken S E. Eight orthopedic surgeons achieved moderate to excellent reliability measuring the preoperative posterior tilt angle in 50 Garden-I and Garden-II femoral neck fractures. J Orthop Surg Res 2017; 12(1): 133. Dolatowski F C, Adampour M, Frihagen F, Stavem K, Erik Utvåg S, Hoelsbrekken S E. Preoperative posterior tilt of at least 20° increased the risk

541

of fixation failure in Garden-I and -II femoral neck fractures. Acta Orthop 2016; 87(3): 252-6. Dolatowski F C, Frihagen F, Bartels S, Opland V, Benth J S, Talsnes O, Hoelsbrekken S E, Utvag S E. Screw fixation versus hemiarthroplasty for nondisplaced femoral neck fractures in elderly patients: a multicenter randomized controlled trial. J Bone Joint Surg Am 2019; 101(2): 136-44. Frihagen F, Madsen J E, Aksnes E, Bakken H N, Maehlum T, Walloe A, Nordsletten L. Comparison of re-operation rates following primary and secondary hemiarthroplasty of the hip. Injury 2007; 38(7): 815-19. Gjertsen J E, Fevang J M, Matre K, Vinje T, Engesæter L B. Clinical outcome after undisplaced femoral neck fractures. Acta Orthop 2011; 82(3): 268-74. Lapidus L J, Charalampidis A, Rundgren J, Enocson A. Internal fixation of Garden I and II femoral neck fractures: posterior tilt did not influence the reoperation rate in 382 consecutive hips followed for a minimum of 5 years. J Orthop Trauma 2013; 27(7): 386-90. Manninger J, Bosch U, Cserháti P, Fekete K, Kazár G. Internal fixation of femoral neck fractures: an atlas. Salzburg, Austria: Springer Science & Business Media; 2007. ISBN 978-3-211-68585-3. Nilsson L T, Johansson A, Stromqvist B. Factors predicting healing complications in femoral neck fractures: 138 patients followed for 2 years. Acta Orthop Scand 1993; 64(2): 175-7. Nyholm A M, Palm H, Sandholdt H, Troelsen A, Gromov K. Osteosynthesis with parallel implants in the treatment of femoral neck fractures: minimal effect of implant position on risk of reoperation. J Bone Joint Surg Am 2018; 100(19): 1682-90. Onativia I J, Slullitel P A, Dilernia F D, Viezcas J M G, Vietto V, Ramkumar P N, Buttaro M A, Piuzzi N S. Outcomes of nondisplaced intracapsular femoral neck fractures with internal screw fixation in elderly patients: a systematic review. Hip Int 2018; 28(1): 18-28. Palm H, Gosvig K, Krasheninnikoff M, Jacobsen S, Gebuhr P. A new measurement for posterior tilt predicts reoperation in undisplaced femoral neck fractures: 113 consecutive patients treated by internal fixation and followed for 1 year. Acta Orthop 2009; 80(3); 303-7. Rogmark C, Leonardsson O. Hip arthroplasty for the treatment of displaced fractures of the femoral neck in elderly patients. Bone Joint J 2016; 98-B(3): 291-7. Rogmark C, Flensburg L, Fredin H. Undisplaced femoral neck fractures—no problems? A consecutive study of 224 patients treated with internal fixation. Injury 2009; 40(3): 274-6.

542

Acta Orthopaedica 2019; 90 (6): 542–546

Similar outcome of femoral neck fractures treated with Pinloc or Hansson Pins: 1-year data from a multicenter randomized clinical study on 439 patients Kristine KALLAND 1, Henrik ÅBERG 2, Anna BERGGREN 3, Michael ULLMAN 4, Greta SNELLMAN 2, Kenneth B JONSSON 2, and Torsten JOHANSSON 5 1 Department of Orthopedic Surgery, Nyköping Hospital, Nyköping, and Department of Clinical and Experimental Medicine, Linköping University, Linköping; 2 Department of Orthopedic Surgery, Institution of Surgical Sciences, Uppsala University, Uppsala; 3 Department of Orthopedic Surgery, Falu Hospital, Falun; 4 Department of Orthopedics, Sahlgrenska University Hospital, Gothenburg/Mölndal, Institute of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg, Gothenburg; 5 Department of Orthopedics, Norrköping, and Department of Clinical and Experimental Medicine, Linköping University, Linköping, Sweden Correspondence: kristine.kalland@regionsormland.se Submitted 2019-03-13. Accepted 2019-07-06.

Background and purpose — There are few reports on the efficiency of the Hansson Pinloc System (Pinloc) for fixation of femoral neck fractures. We compare Pinloc with the commonly used Hansson Pin System in a randomized clinical trial. The primary outcome measure is non-union or avascular necrosis within 2 years. We now report fracture failures and reoperations within the first year. Patients and methods — Between May 2014 and February 2017, 439 patients were included in the study. They were above 50 years of age and treated for a femoral neck fracture at 9 orthopedic departments in Sweden. They were randomized to either Pinloc or Hansson pins. The fractures were grouped as (a) non-displaced regardless of age, (b) displaced in patients < 70 years, or (c) ≥ 70 years old, but deemed unfit to undergo arthroplasty. Follow-up with radiographs and outpatient visits were at 3 and 12 months. Failure was defined as early displacement/non-union, symptomatic segmental collapse, or deep infection. Results — 1-year mortality was 11%. Of the 325 undisplaced fractures, 12% (21/169) Pinloc and 13% (20/156) Hansson pin patients had a failure during the first year. The reoperation frequencies were 10% (16/169) and 8% (13/156) respectively. For the 75 patients 50–69 years old with displaced fractures, 11/39 failures occurred in the Pinloc group and 11/36 in the Hansson group, and 8/39 versus 9/36 patients were reoperated. Among those 39 patients ≥ 70 years old, 7/21 failures occurred in the Pinloc group and 4/18 in the Hansson group. Reoperation frequencies were 4/21 for Pinloc and 3/18 for the Hansson pin patients. No statistically significant differences were found in any of the outcomes between the Pinloc and Hansson groups. Interpretation — We found no advantages with Pinloc regarding failure or reoperation frequencies in this 1-year follow-up.

Internal fixation is the preferred choice for treating undisplaced femoral neck fractures and can also be used for displaced femoral neck fractures depending on age or function. There is still a high revision frequency of approximately 11% in undisplaced femoral neck fractures (Gjertsen et al. 2011) and approximately 27% in patients 55–70 years old with displaced fractures (Bartels et al. 2018). It is therefore important to evaluate new implants designed for better internal fixation. Pinloc (Figure 1) is a development of the commonly used Hansson pins and represents a new concept. Pinloc consists of 3 cylindrical parallel pins with hooks, connected through a fixed angle interlocking plate. The locking plate is not fixed to the femoral cortex, which allows for compression of the fracture along the femoral neck. Biomechanical laboratory studies with composite bone block have shown greater stiffness, torque at failure, and absorbed total energy at failure when fixed with Pinloc compared with 2 Hansson pins (Brattgjerd et al. 2018). Torsional stability is thought to be beneficial for healing of femoral neck fractures (Ragnarsson and Kärrholm 1992) and could possibly mean less pain for the patient and thus facilitate rehabilitation. Additionally, the lateral plate in the Pinloc implant could reduce local soft tissue irritation compared with the use of protruding pins or screw heads,

Figure 1. Hansson Pinloc System.

Acta Orthopaedica 2019; 90 (6): 542–546

which has been proposed to contribute to local pain. We have found only 1 clinical Pinloc study: a retrospective study on 40 patients with no control group (Yamamoto et al. 2019). We report fracture failure and subsequent reoperation frequencies at 1 year in the Swedish Pinloc Study, a randomized controlled study comparing Pinloc to 2 standard Hansson pins in the treatment of femoral neck fractures. Failure was defined as deep infection, early displacement, non-union, and symptomatic segmental collapse. The study is designed for 2 years’ follow-up, but we find it ethically important to provide 1-year results. If patients allocated to Pinloc would have a reduced risk of failure or reoperation compared with Hansson pins, it would be wise to implement its use immediately. If not, a widespread introduction of this implant should be avoided while awaiting the final results of 2-year follow-up, including patient-related outcome measures.

543

ENROLLMENT

Assessed for eligibility n = 1,054 Excluded (n = 516): – did not meet inclusion criteria, 14 – declined to participate, 3 – other reasons, surgeon’s preference, 499 Randomized n = 538 ALLOCATION

Allocated to Hansson Pinloc (n = 274): – received allocated intervention, 273 – did not receive allocated intervention (reason unknown), 1

FOLLOW-UP

Lost to follow-up (n = 1): (moved to other region/country)

Lost to follow-up (n = 4): (moved to other region/country)

Discontinued intervention (n = 43): (deemed unfit for follow-up by patient/ relative, withdrawn consent)

Discontinued intervention (n = 46): (deemed unfit for follow-up by patient/ relative, withdrawn consent)

Patients and methods Subjects The study is a prospective randomized controlled trial including participants from 9 orthopedic departments in Sweden. The trial lasted from May 7, 2014 to February 25, 2017. The inclusion of undisplaced fractures ended on September 10, 2016 when 325 patients had been included. All patients aged 50 years and above, who were admitted to the trial hospitals with a femoral neck fracture considered for internal fixation, were eligible for participation in the study. At randomization, patients were stratified according to orthopedic department and fracture type: undisplaced/displaced, and if displaced, according to age (Figure 2). Patients with prior inclusion in the study presenting with a fracture in the contralateral hip were not included in the study with the new fracture. The modified Garden classification (Oakes et al. 2003), including lateral view radiographs, was used to classify fractures. Garden I–II fractures were considered undisplaced, whereas Garden III–IV fractures were displaced. The patients were given oral and written information concerning the trial and provided written or oral consent to participate in the study. In cases of morbidity or mental dysfunction, where the patient was not able to give consent, a proxy (relative or caretaker) granted permission for participation. Patients were randomized in the operating room after fracture reduction, using a digital randomization platform, to receiving either Pinloc or Hansson pins. Clinical study protocol Each hospital had a surgeon in charge of the study and data collection. Anteroposterior and lateral view radiographs of the hip were taken pre- and postoperatively. All patients were allowed full weight-bearing postoperatively. During the initial hospital admission, information was obtained regarding social conditions, ADL, ASA score,

Allocated to Hanson pins (n = 264): – received allocated intervention, 260 – did not receive allocated intervention (misinterpretation of randomization program, peroperative change in treatment strategy), 4

ANALYSIS

Analyzed (n = 229)

Analyzed (n = 210)

Excluded from analysis (n = 45)

Excluded from analysis (n = 54)

Figure 2. Flow chart of patient enrollment.

smoking, and use of medications (Table 1), as well as the patient-reported outcome measures (PROMs) WOMAC (Western Ontario and McMaster Universities Osteoarthritis Index), and EQ-5D-3L (EuroQol). Patient follow-up consisted of an outpatient visit including radiographs at 3 and 12 months, PROMs and a TUG test (Timed Up and Go test) (PROMs and TUG data not used in this report). In cases where patients did not attend follow-up appointments, information was obtained either by telephone or by review of patient charts. Failure was defined as early displacement, non-union, avascular necrosis (symptomatic segmental collapse), or deep infection. Reoperation was defined as revision surgery with all causes, except removal of an implant due to local pain, as this is considered a less serious complication. The diagnosis of failure and decision to perform further surgery was made locally, at the discretion of the treating surgeon. Statistics Proportion (chi-square test, Fisher’s exact test) was used to compare deaths, failures, and reoperation frequencies. Power analyses showed that to detect a reduction of failures from 40% to 20% for patients with displaced fractures treated with Pinloc, 64 patients were needed in each group for a power of 80%. As the failure rate of undisplaced fractures is lower and less studied, power analysis was not conducted on failure or reoperation. As we expected a 1-year mortality of 30%, we calculated that 43% more patients needed to be included in the study to reach sufficient numbers of patients. Deceased study persons were included in the analysis until death.

544

Acta Orthopaedica 2019; 90 (6): 542–546

Table 1. Demography of patients operated with Pinloc (P) or Hansson pins (H). Values are frequency (percent) unless otherwise stated Displaced Displaced Undisplaced age 50–69 years age ≥ 70 years Factor P (n = 169) H (n = 156) P (n = 39) H (n = 36) P (n = 21) H (n = 18) Female 129 (76) 115 (74) Male 40 (24) 41 (26) Age, median (IQR) 80 (73–86) 80 (71–87) BMI, mean (SD) 24 (4) 23 (4) Dementia 31 (18) 19 (12) Smoking 21 (12) 21 (13) Corticosteroids 12 (7) 7 (4)

19 14 15 11 20 22 6 7 59 (56–64) 62 (58–65) 84 (78–87) 82 (77–88) 25 (4) 26 (5) 25 (4) 25 (4) 0 1 7 5 11 13 2 0 2 0 1 2

Table 2. Failures and local pain. Values are frequency (percent) Displaced Displaced Undisplaced age 50–69 years age ≥ 70 years Factor P (n = 169) H (n = 156) P (n = 39) H (n = 36) P (n = 21) H (n = 18) Infection Early displacement/ non-union Symptomatic segmental collapse New fracture Local pain Total

1 (1)

8 (5)

15 (10)

9 (5) 3 (2) 3 (2) 1 (1) 30 (18) 30 (19) 51 50

2 2 3 1 0 0 0 0 12 17 3 4 23 28 10 8

Östergötland. The authors declare no conflicts of interest.

Results 538 patients were randomized and 439 patients were included in the trial (325 undisplaced and 114 displaced femoral neck fractures) (Figure 2). Patient demographics were similar in the Pinloc and Hansson pin groups (Table 1). The risk of fracture failure varies greatly between displaced and undisplaced fractures as well as between younger patients and those treated with internal fixation due to medical impairments. For these reasons, the data analyses of these groups are presented separately.

Undisplaced fractures No statistically significant difference was found in the failure frequency at 1 year between Table 3. Indication for reoperation. Values are frequency (percent) Pinloc (12%, 21/169), and Hansson pins (13%, 20/156) Displaced Displaced (Table 2). At the 1-year follow Undisplaced age 50–69 years age ≥ 70 years Factor P (n = 169) H (n = 156) P (n = 39) H (n = 36) P (n = 21) H (n = 18) up, 16/169 patients in the Pinloc group and 13/156 patients in the Infection 1 (1) 0 0 0 0 0 Hansson group had undergone Early displacement/ non-union 6 (4) 10 (6) 7 8 3 3 subsequent surgery (other than Symptomatic segmental extraction of the implant only). collapse 6 (4) 2 (1) 1 2 1 0 The indications for reoperation New fracture 3 (2) 1 (1) 0 0 0 0 Local pain 15 (9) 6 (4) 4 4 1 0 were similar between groups Total 31 19 12 14 5 3 (Table 3). Implant removal due to local pain was done in 9% (15/169) of Pinloc and 4% Proportions analyzed with a chi-square test or Fisher’s exact (6/156) of Hansson cases respectively (Table 4). 1-year mortest were used to compare deaths, failures, and reoperation tality was 12% in both groups. frequencies. P-values < 0.05 were considered to be statistiDisplaced fractures, age 50–69 cally significant. The failure frequency at 1 year was similar in the Pinloc Ethics, registration, funding, and potential conflicts of (11/39) and Hansson (11/36) groups. The reoperation freinterest quency and implant removal due to local pain was similar The study protocol was approved by the regional ethics com- between the groups (Table 3). 1 study person in the Pinloc and mittee in Linköping 2013-10-16 (dnr 2013/327-31). The study 1 in the Hansson group had died at 1 year. complies with the World Medical Association Declaration of Helsinki Ethical Principles for Medical Research Involving Displaced fractures, age ≥ 70 Human Subjects. The study was registered at ClinicalTrials. Failure frequencies and subsequent surgery were similar gov (identifier NCT02776631) and was funded by Region between the groups (Table 2 and Table 3). 1/21 Pinloc versus

Acta Orthopaedica 2019; 90 (6): 542–546

545

frequency of segmental collapse was higher in the Pinloc group, but not statistically significant. Displaced Displaced Undisplaced age 50–69 years age ≥ 70 years This shows that there may not Factor P (n = 169) H (n = 156) P (n = 39) H (n = 36) P (n = 21) H (n = 18) be a clear correlation between biomechanical and clinical THA/HA a 11 (7) 12 (8) 8 8 4 3 Re-osteosynthesis b 3 (2) 1 (1) 0 0 0 0 studies (Viberg et al. 2017) and Wound revision 1 (1) 0 0 0 0 0 highlights the importance of Girdlestone 1 (1) 0 0 1 0 0 evaluating new implants in ranExtraction c 15 (9) 6 (4) 4 5 1 0 Total 31 19 12 14 5 3 domized clinical trials before general implementation. a THA = total hip arthroplasty, HA = hemiarthroplasty b Re-osteosynthesis: extraction of implant replaced with another internal fixation implant. This study has several weakc Extraction: removal of implant only (due to pain). nesses. A new device or method has a learning curve. Pinloc is a new concept and the surgical 0/18 Hansson implants were extracted (Table 4). 2 patients in procedure is more demanding for the surgeon than most other the Pinloc group and 6 in the Hansson group had died at the implants. Furthermore, in this multicenter study, over 100 dif1-year follow-up. ferent surgeons performed the operations. All had experience with Hansson pins or Olmed screws, but limited practice with Pinloc. Half (499 of 1,047) of the patients who fulfilled the inclusion criteria were never included because the surgeon on Discussion call chose not to randomize them. No specific reason had to We found similar fracture failure and reoperation frequencies be given by the surgeon, but common explanations were that in the Pinloc and Hansson groups. Several internal fixation the operation would take more time with Pinloc, the surgeon devices for closed reduction and percutaneous fixation have did not feel comfortable with Pinloc, or they simply forgot been developed, among them: single nails with a side plate; about the study. The evaluation of radiographs, segmental colthe sliding hip screw; paired screws or pins; triple screws or lapse, early displacement, and non-union may differ between pins; and pins with hooks or flanges. However, the results treating surgeons. The threshold for performing revision surhave remained about the same for all designs. Bhandari et al. gery may also vary between surgeons, hospitals, and implants (2017) compared a sliding hip screw with cancellous screws (Alho et al. 1998). The study was unblinded since the surin the FAITH trial, but found no statistically significant dif- geons reviewing radiographs for failures were aware of the ference in the risk of reoperations. A systematic review of type of implant allocated. numerous implants for internal fixation of femoral neck fractures showed no statistically significant differences between Summary implants regarding fracture healing complications, reopera- The preliminary data of the 1-year results from this RCT, tions, and mortality (Parker and Gurusamy 2017). This sug- show no statistically significant difference in the frequency of gests that there are aspects of the healing process of femoral failures or reoperations between Pinloc and the Hansson pins neck fractures that we do not understand. in patients over 50 years of age with an undisplaced or a disDifferent methods and implants may come with specific placed femoral neck fracture. As of today we see no benefit in benefits and risks. The most common implants for internal the use of Pinloc over Hansson pins in femoral neck fractures. fixation in Sweden (Hansson pins and Olmed screws) may not offer enough fracture stability, even when optimal reduction AB: performed surgery, and collection of data. GS: performed surgery, and and implant position are achieved. Hansson pins are considediting of manuscript. HÅ: performed surgery, data collection, data interered easier to use by surgeons and the implant positioning pretation, statistical analysis, and editing of manuscript. KJ: study design, performed surgery, and editing of manuscript. KK: performed surgery, data is generally better than for AO screws (Mjørud et al. 2006). collection, data interpretation, and drafting of manuscript. MU: performed Although theoretically appealing, the advantages of Pinloc did surgery, collection of data, and editing of manuscript. TJ: idea, study design, not translate into better healing conditions in either undisplaced supervision, performed surgery, collection of data, and editing of manuscript. or displaced fractures in our clinical study. The increased stability of the Pinloc (Brattgjerd et al. 2018), may come at the The authors thank Dr Jens Nilsson, Department of Orthopedics, Helsingcost of increased intraosseous pressure caused by the 3 pins in borg, Dr Matthias Fassbender, Department of Orthopedics, Eskilstuna, Dr the femoral head. Moreover, the 3 pins connected through the Björn Werner, Department of Orthopedics, Norrköping, Dr Kristbjörg Siglateral plate may put increased stress on the subtrochanteric urdadottir, Department of Orthopedics, Falun, and Dr Håkan Ledin, Department of Orthopedics, Motala for assistance in recruiting patients and data region, leading to a higher rate of subtrochanteric fractures. collection. Swemac is thanked for allowing us to use its image in Figure 1. However, these theories were not supported by our study. The Table 4. Type of reoperation. Values are frequency (percent)

546

Acta thanks Bjarke Viberg for help with peer review of this study. Alho A, Austdal S, Benterud J G, Blikra G, Lerud P, Raugstad T S. Biases in a randomized comparison of three types of screw fixation in displaced femoral neck fractures. Acta Orthop Scand 1998; 69(5): 463-8. Bartels S, Gjertsen J E, Frihagen F, Rogmark C, Utvåg S E. High failure rate after internal fixation and beneficial outcome after arthroplasty in treatment of displaced femoral neck fractures in patients between 55 and 70 years. Acta Orthop 2018; 89(1): 53-8. Bhandari M et al. Fracture fixation in the operative management of hip fractures (FAITH): an international, multicentre, randomised controlled trial. Lancet 2017; 389(10078): 1519-27 Brattgjerd J E, Loferer M, Niratisairak S, Steen H, Strømsøe K. Increased torsional stability by a novel femoral neck locking plate: the role of plate design and pin configuration in a synthetic bone block model. Clin Biomech 2018; 55: 28-35. Gjertsen J E, Fevang J M, Matre K, Vinje T, Engesæter L B. Clinical outcome after undisplaced femoral neck fractures. Acta Orthop 2011; 82(3): 268-74.

Acta Orthopaedica 2019; 90 (6): 542–546

Mjørud J, Skaro O, Solhaug J H, Thorngren K-G. A randomised study in all cervical hip fractures: osteosynthesis with Hansson hook-pins versus AOscrews in 199 consecutive patients followed for two years. Injury 2006; 37(8): 768–77. Oakes D A, Jackson K R, Davies M R, Ehrhart K M, Zohman G L, Koval K J, Lieberman J R. The impact of the garden classification on proposed operative treatment. Clin Orthop Relat Res 2003; (409): 232-40. Parker M J, Gurusamy K S. Internal fixation implants for intracapsular hip fractures in adults (Review). Cochrane Database of Sytematic Reviews 2001, Issue 4. Art. No.: CD001467. Ragnarsson J I, Kärrholm J. Factors influencing postoperative movement in displaced femoral neck fractures: evaluation by conventional radiography and stereoradiography. J Orthop Trauma 1992; 6(2): 152-8. Viberg B, Rasmussen K M V, Overgaard S, Rogmark C. Poor relation between biomechanical and clinical studies for the proximal femoral locking compression plate. Acta Orthop 2017; 88(4): 427-4. Yamamoto T, Kobayashi Y, Nonomiya H. Undisplaced femoral neck fractures need a closed reduction before internal fixation. Eur J Orthop Surg Traumatol 2019; 29(1): 73-8.

Acta Orthopaedica 2019; 90 (6): 547–553

547

Higher risk of cam regrowth in adolescents undergoing arthroscopic femoroacetabular impingement correction: a retrospective compari son of 33 adolescent and 74 adults Tomoya ARASHI 1, Yoichi MURATA 1, Hajime UTSUNOMIYA 1, Shiho KANEZAKI 1, Hitoshi SUZUKI 2, Akinori SAKAI 2, and Soshi UCHIDA 1 1 Wakamatsu

Hospital of University of Occupational and Environmental Health, Japan; 2 University of Occupational and Environmental Health, Japan Correspondence: soushi@med.uoeh-u.ac.jp Submitted 2019-03-18. Accepted 2019-08-12.

Background and purpose — The current literature does not clarify the predictors of cam regrowth and poor clinical outcome following hip arthroscopic femoroacetabular impingement (FAI) correction surgery. Therefore, we investigated the prevalence and risk factors of cam regrowth following arthroscopic FAI correction surgery in skeletally immature patients compared with skeletally mature patients. Patients and methods — 33 teenagers (36 hips as 4 underwent bilateral hip arthroscopies, average age 16.7 [SD 1.6] years, 21 boys [22 hips], 12 girls [14 hips]) undergoing arthroscopic FAI correction and 74 adult controls (74 hips, average age 41 [SD 12] years, 42 men [42 hips], 32 women [32 hips]) were retrospectively reviewed. Postoperative radiographs were obtained, and cam regrowth was evaluated. Clinical characteristics, radiographic findings, arthroscopic findings, and procedures between skeletally immature (SI) and mature (SM) patients were compared. Average followup period was 28 months in the SI group and 24 months in the SM group. Results — Preoperatively, 27 of 36 hips were SI, having either a Risser sign grade ≤ 4 of iliac apophysis or open physes of the proximal femur. Cam regrowth was present in 4 of 27 SI hips. The number of cam regrowth cases was significantly higher in SI patients than in SM patients (0/74 hips). 6 patients required revision hip arthroscopic surgeries (4 men: FAI recurrence due to cam regrowth; 2 women: capsulolabral adhesions). At the last follow-up, the mean modified Harris hip score and nonarthritic hip score were significantly improved postoperatively. Interpretation — 4 of 27 SI hips (95% CI 0.04–0.3) had bone regrowth after cam resection arthroscopically. Our case series showed a non-negligible risk of cam regrowth in SI patients, especially in male patients and patients aged approximately 15 years.

Femoroacetabular impingement (FAI) is today regarded as the most common cause of hip pain in young athletes, resulting from the abutment between the acetabulum and the bump at the femoral head–neck junction (Ganz et al. 2003). Specifically, a cam deformity may be associated with cartilage delamination and labral tear, predisposing to osteoarthritis. Larger alpha angle in cam deformity is the most important predictor of osteoarthritis risk in FAI patients (Agricola et al. 2013a, 2014, Saberi Hosnijeh et al. 2017). Arthroscopic management of FAI in adolescents offers higher patient-reported outcome scores than in adults (Byrd et al. 2016, Fabricant et al. 2012, Tran et al. 2013). Recent studies have shown that potential risk factors for poor clinical outcomes of hip arthroscopic management for FAI are hip dysplasia, older age, and joint space narrowing at surgery (Fukui et al. 2015, Nepple et al. 2011, Nicholls et al. 2011). Although arthroscopic hip revision surgery for a residual cam deformity yielded substantially improved outcome measures, these were inferior to those after primary arthroscopic FAI correction surgery (Larson et al. 2014). Therefore we investigated the incidence and risk factors of cam regrowth after arthroscopic correction in FAI patients. We hypothesized that our results would reveal favorable clinical outcomes following hip arthroscopy for adolescent FAI patients and that the risk for cam regrowth in skeletally immature (SI) FAI patients would be higher than that in skeletally mature (SM) FAI patients (Akel et al. 2013, Kuo et al. 2016).

Patients and methods 355 patients who underwent hip arthroscopies at our institution from January 2009 to December 2014 were enrolled in this study. We defined adolescents as persons aged between

548

Acta Orthopaedica 2019; 90 (6): 547–553

Hips that underwent hip arthroscopy (n = 355): Non-adolescent hips n = 247

Adolescent hips n = 108

Excluded (n = 166): – dysplasia,66 – bilateral, 55 – osteoarthritis, 35 – revision, 10

Enrolled in the study n = 81

Enrolled in the study n = 40

Lost to follow-up n=7 Non-adolescent hips analyzed n = 74

Excluded (n = 68): – dysplasia, 54 – revision, 8 – osteoarthritis, 1 – lumbar or knee disorder, 1 – synovial osteochondromatosis, 1 – psychogenic problem, 1 – presence of capsular laxity, 1 – additional hip surgery, 1

Lost to follow-up n=4

Adolescent hips analyzed (n = 36): – skeletally immature, 27 – skeletally mature, 9

Figure 1. Flowchart showing the recruitment process for patients with femoroacetabular impingement in this study.

13 and 20 years based on past reports (Fabricant et al. 2012, Sink et al. 2008). Of the remaining 107 potential candidates for this study, ultimately 37 patients (40 hips) were enrolled; 4 patients were lost to follow-up; finally, the outcomes of 33 adolescents (36 hips) who underwent arthroscopy for FAI were investigated (Figure 1). The Risser ossification scale for skeletal maturity was used to evaluate maturity of the pelvis (Bitan et al. 2005). Patients with either a Risser sign of grade ≤ 4 or an open physis of the proximal femur were diagnosed with skeletal immaturity. Patient demographics, radiographs, operative details, validated preoperative, and postoperative modified Harris hip score (MHHS) and nonarthritic hip score (NHS) were collected retrospectively. To compare the clinical outcomes of adolescents and adults, 74 adult patients were recruited from the same cohort during the same time period. Those who were diagnosed with dysplasia or osteoarthritis were excluded. The indication for arthroscopic FAI correction surgery was based on the physical examination and radiographs of symptomatic patients. The clinical inclusion criteria were refractory groin pain after a minimum of 3 months of nonoperative treatment, including activity modification, physical therapy, and nonsteroidal anti-inflammatory agents; restricted hip range of motion (ROM) (flexion < 105° and/or restricted internal rotation in flexion < 20°); and a positive impingement test. All patients underwent diagnostic intra-articular local anesthesia, which resulted in immediate relief from symptoms in all patients. This effect was temporary (Kalberer 2008, Yamasaki et al. 2015); therefore, we also performed the anterior impingement test and flexion–adduction–internal-rotation test before surgery (Shanmugaraj et al. 2018, Troelsen et al. 2009). Radiographic evidence of a cam deformity included alpha angle > 55° or head–neck offset ratio < 0.14 in at least radiographic view or the presence of a cam lesion on CT or

Figure 2. Surgical findings: (a) s motorized burr was utilized to decompress the AIIS and trim the rim; (b) The detached labrum was fixed with suture anchors; (c) cam osteochondroplasty was performed; and the procedure was completed with (d) complete capsular closure using Ultra-braids.

MRI (Clohisy et al. 2008). The alpha angle was measured on plain radiographs. We used the highest alpha angle of the 2 views including modified Dunn view and cross-table lateral view for each hip (Notzli et al. 2002). The radiographic FAI subtype was additionally classified as an isolated cam, an isolated pincer, or a combined FAI. Intra-articular pathological abnormalities, including acetabular labral and chondral lesions, were evaluated by gadolinium-enhanced 1.5 Tesla MR arthrography or 3 Tesla MRI. We determined the inter-observer and intra-observer reproducibility of these radiographic parameters. For intraobserver reliability, a single hip surgeon measured each radiograph 3 times, with a minimum interval of 1 week between measurements. For inter-observer reliability, the radiographs were independently reviewed and measured by 2 hip surgeons, who were blinded to the clinical data and details of radiology reports. Intraclass correlation coefficients (ICCs) and corresponding 95% confidence intervals (CI) were calculated to quantify inter-observer and intra-observer reliability for continuous variables. The weighted k-value was used to determine a broken Shenton line and Tonnis classification. k-values and ICCs of 1.0 were indicative of perfect agreement, and the strength of agreement was interpreted as the following ICC values: 0.80 almost perfect; 0.61–0.80 substantial; 0.41–0.60 moderate; and 0.21–0.40 fair. Based on the standards for the k-statistic proposed by Landis and Koch, our measurements were in substantial agreement (Landis and Koch 1977).

Acta Orthopaedica 2019; 90 (6): 547–553

No cam regrowth

549

Cam regrowth

Preoperative

Postoperative

Final follow-up

Figure 3. Representative radiographs of cam regrowth.

Surgical technique (Figure 2) Supine hip arthroscopy was performed on a traction table with a well-padded perineal post under general anesthesia. Intraarticular pathological abnormalities, including labral tearing and cartilage damage, were assessed by introducing 3 portals: an anterolateral portal, a mid-anterior portal (MAP), and a proximal mid-anterior portal (PMAP). An inter-portal capsular cut was performed to improve the access of the scope and surgical instruments. Then, anteroinferior iliac spine (AIIS) decompression and rim trimming were performed using a motorized round burr to prevent damage to the acetabular labrum if necessary (a), to create a surface for labral healing. The detached labrum was repaired with suture anchors (Gryphon BR, Johnson & Johnson, Raynam, MA, USA) (b). After releasing the traction, the peripheral compartment was evaluated for the presence of a cam lesion. If cam impingement was significant, femoral osteochondroplasty was performed with a 5.5 mm motorized round bur with dynamic confirmation of impingement-free ROM (c). We confirmed that femoral osteoplasty for cam lesion was performed appropriately during surgery under fluoroscopic guidance. Finally, capsular closure through the MAP and PMAP was performed as previously described (d). Postoperative recovery Patients were instructed to avoid full weight-bearing to preserve the repaired labrum and capsule for the first 2 weeks. In cases

of a microfracture during surgery, weight-bearing limitations were extended to 6 weeks. Gentle passive ROM exercises such as circumduction were initiated during the 1st week under the supervision of a physical therapist. Continuous passive motion exercises were used to avoid adhesive capsulitis by positioning the hip in 0° to 90° flexion for up to 4 hours a day for 2 weeks. Endurance strengthening was commenced only after the achievement of maximum ROM, good gait stability, and movement. Patients were allowed to progress to physical activity only after demonstrating symmetric passive ROM, achieving a normal gait pattern, and reporting complete pain resolution. Arthroscopic findings At the time of surgery, the condition of the acetabular rim cartilage was evaluated and classified using the Multicenter Arthroscopy Hip Outcome Research Network (MAHORN) classification (Safran and Hariri 2010). The labrum and ligamentum teres were assessed for tears, and the presence of femoral head chondral lesions was reported using the International Cartilage Research Society (ICRS) classification system (Outerbridge 1961). Patient-reported outcome (PRO) scores Patients completed a comprehensive subjective questionnaire, including the NHS (out of 100 points) (Christensen et al. 2003) and MHHS (out of 100 points) (Byrd 2011), which assessed “pain” and “function,” respectively, to document outcomes.

550

Postoperative radiographs Postoperative radiographs from several viewpoints were taken for each patient at 6 months, 1 year, and 2 years. AP pelvic and false-profile radiographs were obtained as part of the protocol to monitor for osteoarthritis progression, heterotopic ossification, and cam or pincer lesion development. AP pelvic radiographs with the collinear alignment of the symphysis and coccyx were also obtained. In addition, the Dunn or modified Dunn view, cross-table lateral views, frogleg lateral views, and false-profile views of the hip, before surgery and at annual followup, were obtained.

Acta Orthopaedica 2019; 90 (6): 547–553

Table 1. Comparison between adolescent patients and adult controls. Data are pre sented as mean (SD) [95% CI] or number and percentage (%) [95% CI] Factor

Adolescent Adult (n = 36) (n = 74)

Age at surgery 16.7 (1.7) [16.1–17.3] Male sex 22 (61) [40%–80%] Body mass index 21 (3.4) [20–23] Lateral center-edge angle (°) 32 (5.4) [30–34] Alpha angle (°) 63 (13.6) [58–67] Femoral neck-shaft angle (°) 131 (3.7) [130–132] AIIS type 1 a 13 (36) [20%–52%] Acetabular cartilage delamination, MAHORN III–V 7 (19) [8.0%–37%] Femoral head cartilage damage, ICRS grade 4 2 (6) [0%–13%] Ligamentum teres pathology 3 (8) [0%–18%] Final follow-up, MHHS 98 (4.6) [96–99] Final follow-up, NHS 97 (5.7) [95–99] Cam regrowth 4 (11) [0.91%–29%] Revision surgery required 6 (17) [3.9%–29%]

41 (12) [38–43] 42 (57) [45%–68%] 22 (2.6) [22–23] 35 (6.4) [33–36] 63 (12.2) [60–66] 131 (3.9) [131–132] 34 (46) [34%–58%] 20 (27) [17%–37%] 9 (12) 11 (15) 97 (4.5) 93 (11) 0 (0) 0 (0)

[4.5%–20%] [6.6%–23%] [96–98] [91–96]

Cam regrowth Cam regrowth was evaluated based on a AIIS (anteroinferior iliac spine) type was diagnosed using 3-dimensional computed postoperative plain radiographs with the tomography (Hetsroni et al. 2012). cross-table lateral view and modified Dunn ICRS = International Cartilage Repair Society; view. To evaluate the alpha angle precisely, MAHORN = Multicenter Arthroscopy of the Hip Outcome Research Network. MHHS = modified Harris hip score we utilized the same view of radiographs NHS = nonarthritic hip score and postoperative radiographs. An alpha angle bigger than that measured by radiographs just after surgery was defined as cam regrowth (Figure 3). Patients were evaluated immediately postoperatively, and postoperatively at 6 months, 1 year, and 2 years. The decision to perform revision arthroscopy was based on the patient’s symptom, image evaluation Results findings including cam regrowth, residual AIIS impingement, Comparison between adolescents and adults and labral re-tear, and physical examination findings. If cam The cam regrowth rate and revision surgery rate were signifiregrowth was found on radiographs images, we performed cantly higher in the adolescent patients than in the adult consecondary cam osteoplasty for revision arthroscopy. trol group. All other data were similar (Table 1). Statistics Outcome data were analyzed using the t-test, paired t-test, or Fisher’s exact test, comparing the adult and adolescent groups. 95% CI and IQR were calculated. Statistical analyses were performed using the SPSS software package version 21 (IBM Corp, Armonk, NY, USA). Power analysis was performed using the method of Degen et al. (2017). Assuming that postoperative outcomes were compared dependently (paired t-test), the effect size was calculated as d = 1.58, and with the actual statistical power of 0.87, 6 adolescents were required to achieve statistically significant improvement (alpha = 0.05).

Bony maturity and patient demographics Bony maturity evaluation revealed 27 hips as SI (16 males), having either an open physis of the proximal femur or a Risser grade ≤ 4. For those that were SI at the first operation, 15 patients were SM after 6 months, 4 patients were SM at 1 year, 3 patients were SM after 2 years, 2 patients were SM after 3 years, and 2 patients were still skeletally immature during follow-up, and in 1 patient radiographs were missing. The average follow-up period for SI patients was 24 (SD 13) (IQR 15–32) months, and for SM patients 40 (SD 16) (IQR 36–48) months.

Ethics, funding, and potential conflicts of interest This study was approved by the institutional review board of University of Occupational and Environmental Health (approval no: H29-005), and by the local institutional review board with a blinded reviewer (approval no. H28-223). SU is a paid consultant for Smith & Nephew and ZimmerBiomet and receives research or institutional support from Smith & Nephew/Pfizer.

Arthroscopic findings No statistically significant differences in the rate of severe acetabular cartilage delamination (MAHORN III–V) was found between the SM group and SI group. Moreover, no significant difference in the rate of femoral head cartilage damage (ICRS grades 2–4) was noted between the 2 groups. Both debridement and labral repair were performed in all cases. Cam osteochondroplasty was performed in all cases in the SI and SM groups.

Acta Orthopaedica 2019; 90 (6): 547–553

551

2 female patients required revision surgery because of adhesive capsulitis, female and SM male patients, in general, were more * 80 80 likely to maintain excellent clinical outcomes * * than SI male patients for a minimum of 2 years postoperatively. 60 60 Degen et al. (2017). indicated no evidence of a difference in follow-up survey 40 40 scores between adult and adolescent groups. Although the improvement in NHS scores 20 20 was statistically significant at the final folPreoperative score low-up, further long-term clinical follow-up Final score 0 0 is necessary to ensure that these improved Immature Mature Immature Mature outcome scores are maintained in the adolesFigure 4. Patient-reported outcome scores: clinical outcome scores of the MHHS and NHS cent population. Our findings are similar to scores are presented with in a box and whisker plot comparing preoperative and final fol those of previous studies on hip arthroscopy low-up values; p < 0.001, paired t-test. Red line is median, box is IQR, whiskers are range, showing it is a safe procedure, with excellent and ● and are outliers. clinical outcomes in teenagers in the presence of FAI (Philippon et al. 2012). A retrospective Outcomes (PRO scores) case series study of adolescent hips reported that short-term At follow-up, MHHS and NHS significantly improved in both improvement of PRO scores without any complications was possible in selected adolescent athletes (Fabricant et al. 2012). groups (Figure 4). A multicenter case series study of 34 patients (41 hips) under Cam regrowth and revision rate 18 years old showed that hip arthroscopy for treating cam-type 6 patients subsequently underwent revision hip arthroscopy, impingement resulted in return to sports, significant improveincluding 4 patients who underwent revision surgery because ment in PRO scores, and no complications (Tran et al. 2013). of FAI recurrence due to cam regrowth and 2 because of cap- Moreover, in a cohort study of 122 hips, superior outcomes sulolabral adhesion. Cam regrowth was noted in 4 hips of 4 were noted in adolescents (≤ 18 years old) to those in the conpatients in the SI group. No cam regrowth occurred in the trol group (> 18 years old) after arthroscopic management of bilateral cases. The rate of cam regrowth was significantly FAI (Byrd et al. 2016). Kocher et al. (2005) evaluated 54 hips higher in the SI than in the SM group. All 4 patients with in 42 patients younger than 18 years with a minimum followcam regrowth required revision surgery for the recurrence of up of 1 year and reported significant improvement in HHS impingement resulting from cam regrowth. All 4 were boys, from 53 to 83, with[AQ2] of patients showing improvement. and the initial arthroscopy was performed at an average age Similarly, our study found that arthroscopic rim trimming, of 15.6 (range 15.4–15.8) years (Figure 5). In addition, revi- labral repair, and cam osteoplasty provide excellent clinical sion hip arthroscopic surgeries were performed in 2 female outcomes for the adolescent FAI. patients to release adhesions; both were diagnosed with adheIn our study, we found cam regrowth in 4/27 hips of male sion at the capsulolabral junction and the osteotomy site. adolescents who required revision surgery after hip arthrosConsequently, patients with cam regrowth had a significantly copy. Previous studies reported no cam regrowth after hip higher rate of revision surgery after initial hip arthroscopy (4 arthroscopic surgery (Gupta et al. 2014, Perets et al. 2017). in 6 patients with cam regrowth vs. 0% in patients without The study by Perets et al. (2017) included only female patients. cam regrowth). In agreement with the study, our findings showed no cam regrowth in adolescent female patients. Recently, Gupta et al. (2014) demonstrated no cam occurrence after arthroscopic FAI surgery in adolescents. They also compared the mean Discussion alpha angle and mean of femoral neck offset at 2 weeks after Our results support the hypothesis that risk factors for poor surgery with those at the final follow-up. In that study, cam clinical outcome resulting from cam regrowth include skeletal regrowth was defined as an alpha angle at the final follow-up immaturity and male sex. The main findings of our study were bigger than that measured by radiographs just after surgery that the rate of cam regrowth, revision surgery rate, and mean in each patient. This slight difference in definition may have NHS at the final follow-up were statistically significantly caused a discrepancy in cam recurrence rate. higher in the adolescent group than in the adult group. 4 SI Predictors of poor clinical outcomes such as joint-space male patients had significant cam regrowth causing recurrent narrowing on radiographs, a prominent AIIS, residual cam impingement and required subsequent surgeries. Although impingement, and capsular laxity after a capsular cut folModified Harris Hip Score

Non Arthritic Hip Score

100

552

lowing hip arthroscopy (Nepple et al. 2011, Nicholls et all. 2011), and developmental dysplasia of the hip (Domb et al. 2014, Larson et al. 2014, Fukui et al. 2015, Sardana et al. 2015) have been reported in previous studies. Cam impingement has been reported as a modifiable risk factor and it has also been stated that the early recognition and treatment of this condition prevents arthritic progression in this population (Agricola et al. 2013b). We found no recurrence of cam lesion in the adult control group, whereas cam regrowth was seen in 4/27 hips of adolescent patients. Further, adolescent male FAI patients had a higher rate of cam regrowth requiring revision arthroscopy. Perhaps the fact that these patients underwent initial hip arthroscopy and cam osteochondroplasty at age 15 is crucial. We believe that male sex and age of 15 years at the time of surgery are risk factors for cam regrowth because adolescent male patients around this age have more active bone formation than adults or female patients of similar age. Moreover, the possibility of cam regrowth with mechanical stress on the unfused proximal femoral physis should be considered. These preoperative risk factors should be considered during surgical planning and discussions with male adolescent patients, who are candidates for arthroscopic FAI correction. Agricola et al. (2014) prospectively investigated the development of cam deformity in a series of 63 male preprofessional soccer players. They stated that cam deformities in young male soccer players gradually developed during skeletal maturation and were likely to stabilize upon epiphyseal closure. Our study has some limitations. First, we lack a nonoperative treatment control group. Since we found no case of cam recurrence in adult patients, the present study focused more on the adolescent population. A study with a larger population may be necessary to allow for additional statistical analyses. Second, arthroscopic findings included chondral lesions, which may have influenced the results. Third, we used MHHS and NHS as the primary clinical outcome assessment scores. Although MHHS shows a significant ceiling effect and iHot should be considered, it was not available during this period (Griffin et al. 2012). Fourth, the sample size was relatively small. However, power analysis suggested that the number of patients was sufficient for this study. Further studies are needed to evaluate the longer-term clinical outcomes of surgical procedures and in a larger number of patients. Fifth, we evaluated cam regrowth on plain radiographs, the projections of which, however, may not be consistent. For evaluation under the same condition we considered performing CT, but this was unrealistic because many patients were concerned about radiation. Sixth, there were 3 bilateral cases in the adolescent group, and our statistical method did not explicitly account for the dependency between bilateral hips in the same patient. In summary, 4 of 27 SI hips had bone regrowth after CAM resection arthroscopically, a non-negligible risk, especially in male patients and those aged approximately 15 years.

Acta Orthopaedica 2019; 90 (6): 547–553

SU, TA, and YM had the major role in elaborating the plan, formulating the manuscript, and interpreting the data. AS, SK, and HU contributed in data collection, formulating the plans, and writing the manuscript. The authors would like to thank Editage (www.editage.jp) for English-language editing and reviewing of this manuscript. Acta thanks Hanne Hedin and Søren Overgaard 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 [Research Support, Non-U.S. Gov’t] 2013a; 72(6): 918-23. doi: 10.1136/annrheumdis-2012-201643. Agricola R, Waarsing J H, Arden N K, Carr AJ, Bierma-Zeinstra SM, Thomas GE, Weinans H, Glyn-Jones S. Cam impingement of the hip: a risk factor for hip osteoarthritis. Nat Rev Rheumatol 2013b; 9(10): 630-4. doi: 10.1038/nrrheum.2013.114. Agricola R, Heijboer M P, Ginai A Z, Roels P, Zadpoor A A, Verhaar J A, Weinans H, Waarsing J H. A cam deformity is gradually acquired during skeletal maturation in adolescent and young male soccer players: a prospective study with minimum 2-year follow-up. Am J Sports Med 2014; 42(4): 798-806. doi: 10.1177/0363546514524364. Akel I, Songur M, Karahan S, Yilmaz G, Demirkiran H G, Tumer Y. Acetabular index values in healthy Turkish children between 6 months and 8 years of age: a cross-sectional radiological study. Acta Orthop Traumatol Turc 2013; 47(1): 38-42. Bitan F D, Veliskakis K P, Campbell B C. Differences in the Risser grading systems in the United States and France. Clin Orthop Relat Res 2005; (436): 190-5. Byrd J W. Arthroscopic hip surgery for the treatment of femoroacetabular impingement. Orthopedics 2011; 34(3): 186. doi: 10.3928/0147744720110124-17. Byrd J W, Jones K S, Freeman C R. Surgical outcome of pincer femoroacetabular impingement with and without labral ossification. Arthroscopy 2016; 32(6): 1022-9. doi: 10.1016/j.arthro.2015.12.042. 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. Clohisy J C, Carlisle J C, Trousdale R, Kim Y-J, Beaule P E, Morgan P, Steger-May K, Schoenecker P L, Millis M. Radiographic evaluation of the hip has limited reliability. Clin Orthop Relat Res 2008; 467(3): 666-75. doi: 10.1007/s11999-008-0626-4. Degen R M, Mayer S W, Fields K G, Coleman S H, Kelly B T, Nawabi D H. Functional outcomes and cam recurrence after arthroscopic treatment of femoroacetabular impingement in adolescents. Arthroscopy 2017; 33(7): 1361-9. doi: 10.1016/j.arthro.2017.01.044. Domb B G, Stake C E, Lindner D, El-Bitar Y, Jackson T J. Revision hip preservation surgery with hip arthroscopy: clinical outcomes. Arthroscopy 2014; 30(5): 581-7. doi: 10.1016/j.arthro.2014.02.005. Fabricant P D, Heyworth B E, Kelly B T. Hip arthroscopy improves symptoms associated with FAI in selected adolescent athletes. Clin Orthop Relat Res 2012; 470(1): 261-9. doi: 10.1007/s11999-011-2015-7. Fukui K, Briggs K K, Trindade C A, Philippon M J. Outcomes after labral repair in patients with femoroacetabular impingement and borderline dysplasia. Arthroscopy 2015; 31(12): 2371-9. doi: 10.1016/j.arthro.2015.06.028. Ganz R, Parvizi J, Beck M, Leunig M, Notzli H, Siebenrock K A. Femoroacetabular impingement: a cause for osteoarthritis of the hip. Clin Orthop Relat Res 2003; (417): 112-20. Griffin D R, Parsons N, Mohtadi N G, Safran M R. A short version of the International Hip Outcome Tool (iHOT-12) for use in routine clinical practice. Arthroscopy 2012; 28(5): 611-6; quiz 6-8. doi: 10.1016/j. arthro.2012.02.027.

Acta Orthopaedica 2019; 90 (6): 547–553

Gupta A, Redmond J M, Stake C E, Finch N A, Dunne K F, Domb B G. Does the femoral cam lesion regrow after osteoplasty for femoroacetabular impingement? Two-year follow-up. Am J Sports Med 2014; 42(9): 214955. doi: 10.1177/0363546514541782. Hetsroni I, Larson C M, Dela Torre K, Zbeda R M, Magennis E, Kelly B T. Anterior inferior iliac spine deformity as an extra-articular source for hip impingement: a series of 10 patients treated with arthroscopic decompression. Arthroscopy 2012; 28(11): 1644-53. doi: 10.1016/j.arthro. 2012.05.882. Kalberer F. Ischial spine projection into the pelvis: a new sign for acetabular retroversion. Clin Orthop Relat Res 2008; 466(3): 677-83. Kocher M S, Kim Y J, Millis M B, Mandiga R, Siparsky P, Micheli L J, Kasser J R. Hip arthroscopy in children and adolescents. J Pediatr Orthop 2005; 25(5): 680-6. doi: 01241398-200509000-00024 [pii]. Kuo F C, Kuo S J, Ko J Y. Overgrowth of the femoral neck after hip fractures in children. J Orthop Surg Res 2016; 11(1): 50. doi: 10.1186/s13018-0160387-9. Landis J R, Koch G G. The measurement of observer agreement for categorical data. Biometrics 1977; 33(1): 159-74. Larson C M, Giveans MR, Samuelson K M, Stone R M, Bedi A. Arthroscopic hip revision surgery for residual femoroacetabular impingement (FAI): surgical outcomes compared with a matched cohort after primary arthroscopic FAI correction. Am J Sports Med 2014; 42(8): 1785-90. doi: 10.1177/0363546514534181. Nepple J J, Carlisle J C, Nunley R M, Clohisy J C. Clinical and radiographic predictors of intra-articular hip disease in arthroscopy. Am J Sports Med 2011; 39(2): 296-303. doi: 0363546510384787 [pii] 10.1177/0363546510384787. Nicholls A S, Kiran A, Pollard T C, Hart D J, Arden C P, Spector T, Gill H S, Murray D W, Carr A J, Arden N K. The association between hip morphology parameters and nineteen-year risk of end-stage osteoarthritis of the hip: a nested case-control study. Arthritis Rheum 2011; 63(11): 3392-400. doi: 10.1002/art.30523. 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.

553

Outerbridge R E. The etiology of chondromalacia patellae. J Bone Joint Surg Br 1961; 43-B: 752-7. Perets I, Gupta A, Chaharbakhshi E O, Ashberg L, Hartigan D E, Close M R, Domb B G. Does bony regrowth occur after arthroscopic femoroplasty in a group of young adolescents? Arthroscopy 2017; 33(5): 988-95. doi: 10.1016/j.arthro.2017.01.023. Philippon M J, Ejnisman L, Ellis H B, Briggs K K. Outcomes 2 to 5 years following hip arthroscopy for femoroacetabular impingement in the patient aged 11 to 16 years. Arthroscopy 2012; 28(9): 1255-61. doi: S07498063(12)00088-6 [pii] 10.1016/j.arthro.2012.02.006. Saberi Hosnijeh F, Zuiderwijk M E, Versteeg M, Smeele H T, Hofman A, Uitterlinden A G, Agricola R, Oei E H, Waarsing J H, Bierma-Zeinstra S M, van Meurs J B. Cam deformity and acetabular dysplasia as risk factors for hip osteoarthritis. Arthritis Rheumatol 2017; 69(1): 86-93. doi: 10.1002/ art.39929. Safran M R, Hariri S. Hip arthroscopy assessment tools and outcomes. Oper Tech Orthop 2010; 20(4): 264-77. Sardana V, Philippon M J, de Sa D, Bedi A, Ye L, Simunovic N, Ayeni O R. Revision hip arthroscopy indications and outcomes: a systematic review. Arthroscopy 2015; 31(10): 2047-55. doi: 10.1016/j.arthro.2015.03.039. Shanmugaraj A, Shell J R, Horner N S, Duong A, Simunovic N, Uchida S, Ayeni O R. How useful is the flexion–adduction–internal rotation test for diagnosing femoroacetabular impingement. Clin J Sport Med 2018. [Epub ahead of print] doi: 10.1097/jsm.0000000000000575. Sink E L, Gralla J, Ryba A, Dayton M. Clinical presentation of femoroacetabular impingement in adolescents. J Pediatr Orthop 2008; 28(8): 806-11. doi: 10.1097/BPO.0b013e31818e194f. Tran P, Pritchard M, O’Donnell J. Outcome of arthroscopic treatment for cam type femoroacetabular impingement in adolescents. ANZ J Surg 2013; 83(5): 382-6. doi: 10.1111/j.1445-2197.2012.06197.x. Troelsen A, Mechlenburg I, Gelineck J, Bolvig L, Jacobsen S, Soballe K. What is the role of clinical tests and ultrasound in acetabular labral tear diagnostics? Acta Orthop 2009; 80(3): 314-8. doi: 10.3109/17453670902988402. Yamasaki T, Yasunaga Y, Shoji T, Izumi S, Hachisuka S, Ochi M. Inclusion and exclusion criteria in the diagnosis of femoroacetabular impingement. Arthroscopy 2015; 31(7): 1403-10. doi: 10.1016/j.arthro.2014.12.022

554

Acta Orthopaedica 2019; 90 (6): 554–558

General anesthesia might be associated with early periprosthetic joint infection: an observational study of 3,909 arthroplasties Ruben SCHOLTEN 1, Borg LEIJTENS 1, Gerjon HANNINK 2, Ed T KAMPHUIS 3, Matthijs P SOMFORD 1, and Job L C VAN SUSANTE 1 1 Rijnstate 3 Rijnstate

Ziekenhuis, Department of Orthopedics, Arnhem; 2 Radboud University Medical Center, Department of Operating Rooms, Nijmegen; Ziekenhuis, Department of Anesthesiology, Arnhem, the Netherlands Correspondence: JvanSusante@rijnstate.nl Submitted 2019-03-28. Accepted 2019-06-25.

Background and purpose — Periprosthetic joint infection (PJI) remains a devastating complication following total knee or total hip arthroplasty (TKA/THA). Nowadays, many studies focus on preventive strategies regarding PJI; however, the potential role of anesthesia in the development of PJI remains unclear. Patients and methods — All consecutive patients undergoing elective primary unilateral TKA or THA from January 2014 through December 2017 were included. Exclusion criteria included femoral fractures as the indication for surgery and previously performed osteosynthesis or hardware removal on the affected joint. Age, sex, BMI, ASA classification, type of arthroplasty surgery, type of anesthesia, duration of surgery, smoking status, and intraoperative hypothermia were recorded. Propensity score-matched univariable logistic regression analysis was used to control for allocation bias. Results — 3,909 procedures consisting of 54% THAs and 46% TKAs were available for analysis. 42% arthroplasties were performed under general anesthesia and 58% under spinal anesthesia. Early PJIs were observed in 1.7% of the general anesthesia group and in 0.8% in the spinal anesthesia group. The multivariable logistic regression model demonstrated an odds ratio for PJI of 2.0 (95% CI 1.0–3.7) after general anesthesia relative to the propensity score-matched patients who received spinal anesthesia. Interpretation — These results suggest a potential association between general anesthesia and early PJI. Future research using large-scale data is required to further elucidate this clinically relevant association.

Periprosthetic joint infection (PJI) is responsible for up to 25% of failed TKA and 15% of failed THA (Bozic et al. 2009, 2010). Despite the increasing awareness of certain patient characteristics that influence the risk of PJI (Kunutsor et al. 2016), the role of procedure-related factors, such as the type of anesthesia, remains to be elucidated (Berbari et al. 2017). Remarkably, the notion that anesthesia may influence the immune response was suggested as early as 1903 (Moudgil 1986). In the late 1970s and ’80s several reviews identified the ability of anesthetic agents to influence a wide variety of specific and non-specific host defenses (Moudgil 1986). However, the clinical relevance and the exact role of anesthesia in the pathogenesis of postoperative infections remains unclear (Moudgil 1986, Cruz et al. 2017). Several studies have suggested spinal anesthesia to reduce the risk for surgical site infection (SSI) when compared with general anesthesia in THA and TKA (Liu et al. 2013, Pugely et al. 2013); however, other studies found no clear difference between the 2 types of anesthesia and their influence on SSI (Basques et al. 2015, Helwani et al. 2015, Kopp et al. 2015). Nevertheless, a recent systematic review suggested that regional anesthesia seems to decrease the risk of SSI when compared with general anesthesia (Zorrilla-Vaca et al. 2016). Despite several clues pointing to general anesthesia predisposing to infection, no studies assessing the role of anesthesia during THA and TKA with well-defined definitions of PJI have been performed. Therefore, we investigated the relationship between type of anesthesia (i.e., spinal or general) and PJI following THA or TKA in a large-scale observational cohort study.

Acta Orthopaedica 2019; 90 (6): 554–558

Patients and methods All consecutive patients undergoing elective primary unilateral TKA or THA for osteoarthritis in a single general teaching hospital from January 2014 through December 2017 were retrieved from the hospital’s prospective database. Subsequent exclusion criteria were proximal femoral fracture or acetabular fracture as the indication for primary surgery and concomitant or previous hardware removal on the affected joint. Data were recorded regarding the patient’s age, sex, ASA classification, BMI, smoking behavior, type of arthroplasty surgery, type of anesthesia, surgery duration, intraoperative hypothermia, and length of stay. Over the course of the study period a similar surgical technique was used, and no changes to the surgery protocol were implemented. Patients received prophylactic administration of 2 grams cefazolin 15 to 60 minutes prior to skin incision or tourniquet inflation, followed by 3 administrations of 1 gram after surgery with an 8-hour interval. All THAs were performed by, or under direct supervision of, 1 of 7 hip surgeons. Accordingly, all TKAs were performed by 1 of 4 knee surgeons. Several residents or trainees participated in most surgeries. All TKA patients underwent surgery while using a tourniquet, which was inflated 15 to 60 minutes after infusion of the prophylactic cefazolin and deflated after applying a pressure bandage over the affected knee. Only patients with primary implant models and no revision models were included. All TKAs were cemented and performed using a medial parapatellar arthrotomy. THA was performed using a posterolateral approach. Both cemented and uncemented THA were performed with a patient age cut-off point below 75 years for uncemented THA. The bone cement (Palacos R+G; Heraeus Medical, Hanau, Germany) used in both TKA and THA contained 0.75 grams of gentamicin per 61.2 grams of powder. The decision to apply either general or spinal anesthesia during the arthroplasty was at the discretion of 1 of the senior anesthesiologists and based on the patient’s personal preference. Patients were extensively informed about both general and spinal anesthesia, after which they could indicate their preference. To correct for potential allocation bias introduced through this selection procedure, propensity score-based matching of cases was performed (please refer to Statistics section for further information). Surgical duration was defined as the time between skin incision and closure. The core temperature was measured at the tympanic membrane in the operating room directly after wound closure. Prior to discharge patients were closely monitored for signs of potential postoperative infection. Following discharge, all patients were subjected to protocolized surveillance of infection in the outpatient clinic for at least 3 months after surgery. In case of prolonged wound drainage (> 10 days), suspected

555

(superficial) surgical site infection (SSI), or superficial wound breakdown, surgical debridement, with antibiotics and implant retention (DAIR), was performed. During DAIR, 6 periprosthetic tissue biopsies were always obtained and subsequently cultured. Superficial SSI was defined according to the Infectious Centers of Disease Control (CDC) guidelines with the presence of: (1) purulent incisional drainage, (2) positive culture of aseptically obtained fluid or tissue from the superficial wound, (3) local signs and symptoms of pain or tenderness, swelling, and erythema after the incision is opened by the surgeon (unless culture negative), or (4) diagnosis of SSI by the attending surgeon or physician based on their experience and expert opinion (Horan et al. 1992). Until final cultures results were obtained up to 10 days after DAIR, patients were treated with intravenous antibiotics (flucloxacillin, 6g/day via continuous intravenous infusion). PJI was diagnosed according to the major Musculoskeletal Infection Society (MSIS) criteria by means of 2 or more tissue cultures demonstrating growth of an identical pathogen (Parvizi and Gehrke 2014). If PJI was diagnosed, antibiotic therapy was adjusted accordingly in consultation with the attending microbiologist. The primary outcome of this study was the incidence of PJI within 3 months of surgery. Statistics Multiple imputation by chained equations procedures was used for missing values to increase precision and to avoid bias (van Buuren 2018). We generated 25 independent imputed datasets, as current guidance recommends that 1 imputation should be performed per percent of incomplete observations (White et al. 2011). Smoking behavior and hypothermia had 4.1% and 24% missing values, respectively, whereas other variables had less than 0.1% missing values (Table 1). A difference for the risk for early PJI between cases that received spinal and those that received general anesthesia might be biased by confounding. A particularly important type of confounding in this case is “confounding by indication,” which occurs when the clinical indication for selecting a particular intervention also affects the outcome. For example, patients with more severe comorbidities (e.g., CVD) are more likely to receive general anesthesia, but they are also more likely to develop early PJI. Another type of confounding is “confounding by association”, which occurs when both exposure (i.e. type of anesthesia) and outcome (i.e. early PJI) are associated with a third variable. For example, BMI is associated both with type of anesthesia and with increasing risk of early PJI. In order to adjust for potential confounding baseline characteristics, we matched patients based on their propensity scores (Rubin 1997). The propensity score was defined as the probability of receiving general anesthesia during TJA dependent on a case’s recorded baseline characteristics. Propensity scores were estimated independently for each imputed dataset, using a logistic regression model with

556

Acta Orthopaedica 2019; 90 (6): 554â&#x20AC;&#x201C;558

Table 1. Distribution of patient characteristics and missing data among the general anesthesia and spinal anesthesia groups. Values are frequency (%) unless otherwise stated Factor

Spinal General Cumulative anesthesia Missing anesthesia Missing missing (n = 2,279) data (%) (n = 1,630) data (%) data (%)

Covariates Male sex Age BMI ASA score 1 ASA score 2 ASA score 3

Age, mean (SD) 70 (9.5) Male sex 789 (35) BMI, mean (SD) 28.7 (4.7) ASA 1 348 (15) ASA 2 1,614 (71) ASA 3 306 (13) ASA 4 11 (0.5) Active smoker 231 (11) TKA 1,082 (48) 2014 674 (30) 2015 591 (26) 2016 488 (21) 2017 526 (23) Mean surgery duration, min (SD) 59 (16) Hypothermia 87 (5.6) PJI 19 (0.8)

0 67 (10) 0 597 (37) 0 29.7 (5.2) 0 221 (14) 0 1,049 (64) 0 344 (21) 0 15 (0.9) 101 (4.4) 223 (14) 0 716 (44) 0 286 (18) 0 391 (24) 0 518 (32) 0 435 (27) 0 724 (32) 0

62 (16) 54 (3.8) 28 (1.7)

0 0 0 0 2 (0.1) 2 (0.0) 1 (0.1) 1 (0.0) 1 (0.1) 1 (0.0) 1 (0.1) 1 (0.0) 1 (0.1) 1 (0.0) 61 (3.7) 162 (4.1) 0 0 0 0 0 0 0 0 0 0 0 0 223 (14) 947 (24) 0 0

ASA score 4 TKA 2014 2015 2016 2017

Unadjusted Adjusted

Smoking 0.0

0.1

0.2

0.3

Standardized absolute mean differences Covariate balance before (unadjusted) and after (adjusted) propensity score matching. Standardized differences less than 10% (dashed line) indicate an appropriate balance (Austin 2010).

Percentages are displayed as valid (calculated through discarding missing data) percentages. BMI: body mass index, TKA: total knee arthroplasty, PJI: periprosthetic joint infection.

type of anesthesia as the dependent variable in relation to the following baseline characteristics: age, sex, BMI, ASA classification, smoking status, THA or TKA surgery, and year of surgery. The selection of which variables to include in our analyses in order to minimize bias was done using directed acyclic graphs based on the approaches described by Shrier and Pearl (Shrier and Platt 2008, Pearl 2010). A 1:1 optimal matching algorithm was applied without replacement to match exposed and non-exposed cases on their corresponding propensity scores within a caliper of 0.2 standard deviation of the logit of the propensity score (Austin 2011). A 1:1 matching on propensity score was used as it has been shown that it tends to minimize bias compared with many-to-1 matching on propensity score (Austin 2010). The balance between the two groups after matching was checked graphically and descriptively. A standardized difference of less than 10% indicates an appropriate balance (Austin 2011). Standardized differences (difference in means divided by the pooled standard deviation) of the baseline characteristics for a randomly selected matched dataset are provided in Table 2 (see Supplementary data). On each of the 25 imputed and propensity score-matched datasets, a univariable logistic regression analysis with PJI within 3 months after surgery as the dependent variable and type of anesthesia as independent variable was performed. The resulting estimates were pooled using Rubinâ&#x20AC;&#x2122;s rule (Rubin 1997). Statistical analyses were performed using R 3.5.2 (R Foundation for Statistical Computing, Vienna, Austria).

Ethics, funding, and potential conflicts of interest The local institutional review board approved this study (study number: 2018-1276). No funding was received. No conflicts of interest were declared.

Results Between January 1, 2014 and December 31, 2017, 4,026 primary unilateral total hip and knee arthroplasties were performed. 58 THAs and 59 TKAs were excluded due to previous or concomitant hardware removal, leaving 3,909 joints consisting of 2,111 (54%) hips and 1,798 (46%) knees available for analysis. Among all eligible patients, 1,630 (42%) arthroplasties were performed under general anesthesia and 2,279 (58%) arthroplasties were performed under spinal anesthesia. Apart from the DAIR procedures, 17 cases underwent revision surgery within 3 months of primary TJA (Table 3, see Supplementary data). None of these cases were eventually diagnosed with early PJI. 47 early PJIs were diagnosed through 2 or more positive intraoperative tissue cultures, obtained during DAIR, demonstrating an identical pathogen. 28 (1.7%) PJIs occurred in the general anesthesia group and 19 (0.8%) in the spinal anesthesia group. The covariate balance before and after propensity scorebased matching is displayed in the Figure and Table 2 (see

Acta Orthopaedica 2019; 90 (6): 554–558

Supplementary data). In the 1,630 patients who received general anesthesia 28 (1.7%) PJIs occurred, while in the 1,630 matched participants who received spinal anesthesia, 13–15 (0.8–0.9%) PJIs occurred, depending on imputation set. The odds ratio for early PJI was estimated to be 2.0 (95% CI 1.0–3.7) for patients who received general anesthesia compared with matched patients who received spinal anesthesia. Although no longer statistically significant, subsequent subgroup analysis addressing THA and TKA separately showed similar odds ratios (THA 2.1 [CI 0.99–4.6] and TKA 2.0 [CI 0.53–7.9]).

Discussion Over the past decade, several studies have suggested that spinal anesthesia decreases the risk for SSI after TJA when compared with general anesthesia (Liu et al. 2013, Pugely et al. 2013). However, this remains debated since many conflicting results have been reported and there is a paucity of highquality studies using objective criteria for SSI (Basques et al. 2015, Helwani et al. 2015, Kopp et al. 2015). The distinction between (superficial) SSI and early PJI in orthopedic surgery is far from straightforward yet clinically important. In 1999, the Centers for Disease Control (CDC) formulated definitions for superficial, deep incisional, and organ/space SSI (Mangram et al. 1999). However, there are no procedures or tests to reliably allow differentiation between these subtypes of SSI (Amanatullah et al. 2019). Furthermore, diagnostic criteria for superficial SSI such as tenderness, redness, localized swelling, and local heat are subject to interobserver variability (Allami et al. 2005). Therefore, previous studies addressing the effect of anesthesia on SSI yield less reliable results compared with this study using objectified PJI as the primary outcome measure. The IDSA guidelines dictate vigorous surgical treatment for (suspected) SSI following TJA including surgical debridement and rinsing of the implant (Osmon et al. 2013). In previous studies these guidelines were not applied and as such the diagnosis of actual early PJI was not reliably established. To our knowledge this is the first study using the IDSA guidelines where an association between the type of anesthesia and the incidence of objectified early PJI (using the major MSIS criteria through the availability of at least 6 periprosthetic tissue cultures in every case with suspected infection) is shown. Our results indicate an increased risk of early PJI following TJA under general anesthesia, illustrated by an odds ratio of 2 (CI 1.0–3.7). Although no longer statistically significant, subgroup analysis for the type of arthroplasty (THA or TKA) demonstrated similar confidence intervals, which indicate these results are robust and do not seem to depend on type of arthroplasty. So far, the mechanism by which general anesthesia might increase, or spinal anesthesia might reduce the incidence of infection is not fully understood. However, increased tissue

557

oxygenation (through reduced postoperative pain and the direct vasodilatory effect of spinal anesthesia) has been suggested as a potential mechanism in the past (Sessler 2010). Alongside these beneficial effects on tissue oxygenation, neuraxial anesthesia is also associated with reduced blood loss, a reduced requirement for blood transfusions, and a reduced incidence of hyperglycemia. All these factors are known for their immunosuppressive effects (Guay 2006, Gottschalk et al. 2014). Besides the suggestion of protective effects of spinal anesthesia, several aesthetic agents that are commonly used in general anesthesia may significantly inhibit leukocyte chemotactic migration, phagocytosis, lymphocyte function, inflammation, or even directly support bacterial growth in case of contamination (Moudgil 1986). Furthermore, studies comparing general and spinal anesthesia showed that in spinal anesthesia these immunosuppressive effects were minimal (Whelan and Morris 1982). On the other hand, one could speculate regarding a potential negative effect of spinal anesthesia on the incidence of early PJI induced by intraoperative hypothermia, which has been associated with an increased incidence of SSIs in other surgical specialties and is more prevalent during spinal anesthesia (Scholten et al. 2018). However, despite the latter, spinal anesthesia was still associated with a reduced risk of early PJI in our study. Limitations First, due to the observational nature of the study, confounding (by indication) cannot be precluded. To control for this potential confounding, we matched patients based on propensity scores. Although matching of patients was successfully performed based on a subset of baseline characteristics, differences could theoretically still exist in unmeasured covariates (e.g., diabetes mellitus, rheumatoid arthritis, and anticoagulant usage in this study) resulting in residual confounding. Anticoagulant therapy, for example, is generally considered to be a contra-indication for the application of spinal anesthesia. This may have caused allocation of anticoagulant users to the general anesthesia group. However, since protocols for perioperative interruption of anticoagulant use (with or without bridging) are readily available and mandatory regarding elective TJA in our clinic, anticoagulant therapy is unlikely to cause allocation of patients towards the general anesthesia group. Furthermore, both diabetes and rheumatoid arthritis do not influence the choice for either spinal or general anesthesia in our hospital. Another limitation is the fact that our data are sourced from 1 hospital only. Therefore, the major question remains whether our data and the conclusions drawn will be reproducible in studies on, for example, national joint registries. However, on the other hand this last limitation warranted that complete follow-up could be guaranteed and that no PJIs could have been missed.

558

In conclusion, this is the first study to suggest a potential association between general anesthesia and an increased risk of early PJI. This clinically relevant finding should encourage the setting up of future research using (multi-center) randomized large-scale data and national joint registry studies. Supplementary data Tables 2 and 3 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674. 2019.1644069

RS: Study setup, literature research, study design, data collection, writing of the manuscript. BL: Study design, literature research, writing of the manuscript. GH: Study design, statistical analysis, writing of the manuscript. ETK: Literature research, writing of the manuscript. MPS: Study setup, study supervision, writing of the manuscript. JLCS: Study setup, study supervision, writing of the manuscript. Acta thanks Anna Stefánsdóttir for help with peer review of this study.

Allami M K, Jamil W, Fourie B, Ashton V, Gregg P J. Superficial incisional infection in arthroplasty of the lower limb: interobserver reliability of the current diagnostic criteria. J Bone Joint Surg Br 2005; 87(9): 1267-71. doi: 10.1302/0301-620x.87b9.16672. Amanatullah D, Dennis D, Oltra E G, Marcelino Gomes L S, Goodman S B, Hamlin B, Hansen E, Hashemi-Nejad A, Holst D C, Komnos G, Koutalos A, Malizos K, Martinez Pastor J C, McPherson E, Meermans G, Mooney J A, Mortazavi J, Parsa A, Pecora J R, Pereira G A, Martos M S, Shohat N, Shope A J, Zullo S S. Hip and knee section, diagnosis, definitions: Proceedings of International Consensus on Orthopedic Infections. J Arthroplasty 2019; 34(2s): S329-s37. doi: 10.1016/j. arth.2018.09.044. Austin P C. Statistical criteria for selecting the optimal number of untreated subjects matched to each treated subject when using many-to-one matching on the propensity score Am J Epidemiol 2010; 172(9):1092-7. doi: 10.1093/aje/kwq224. Austin P C. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat 2011; 10(2): 150-61. doi: 10.1002/pst.433. Basques B A, Toy J O, Bohl D D, Golinvaux N S, Grauer J N. General compared with spinal anesthesia for total hip arthroplasty. J Bone Joint Surg Am 2015; 97(6): 455-61. doi: 10.2106/jbjs.n.00662. Berbari E, Segreti J, Parvizi J, Berrios-Torres s i. future Research Opportunities in Peri-Prosthetic Joint Infection Prevention. Surg Infect (Larchmt) 2017; 18(4): 409-12. doi: 10.1089/sur.2017.065. Bozic K J, Kurtz S M, Lau E, Ong K, Vail T P, Berry D J. The epidemiology of revision total hip arthroplasty in the United States. J Bone Joint Surg Am 2009; 91(1): 128-33. doi: 10.2106/jbjs.h.00155. Bozic K J, Kurtz S M, Lau E, Ong K, Chiu V, Vail T P, Rubash H E, Berry D J. The epidemiology of revision total knee arthroplasty in the United States. Clin Orthop Relat Res 2010; 468(1): 45-51. doi: 10.1007/s11999009-0945-0. Cruz F F, Rocco P R, Pelosi P. Anti-inflammatory properties of anesthetic agents. Crit Care 2017; 21(1): 67. doi: 10.1186/s13054-017-1645-x. Gottschalk A, Rink B, Smektala R, Piontek A, Ellger B, Gottschalk A. Spinal anesthesia protects against perioperative hyperglycemia in patients undergoing hip arthroplasty. J Clin Anesth 2014; 26(6): 455-60. doi: 10.1016/j. jclinane.2014.02.001.

Acta Orthopaedica 2019; 90 (6): 554–558

Guay J. The effect of neuraxial blocks on surgical blood loss and blood transfusion requirements: a meta-analysis. J Clin Anesth 2006; 18(2): 124-8. doi: 10.1016/j.jclinane.2005.08.013. Helwani M A, Avidan M S, Ben Abdallah A, Kaiser D J, Clohisy J C, Hall B L, Kaiser H A. Effects of regional versus general anesthesia on outcomes after total hip arthroplasty: a retrospective propensity-matched cohort study. J Bone Joint Surg Am 2015; 97(3): 186-93. doi: 10.2106/jbjs.n.00612. Horan T C, Gaynes R P, Martone W J, Jarvis W R, Emori T G. CDC definitions of nosocomial surgical site infections, 1992: a modification of CDC definitions of surgical wound infections. Infect Control Hosp Epidemiol 1992; 13(10): 606-8. Kopp S L, Berbari E F, Osmon D R, Schroeder D R, Hebl J R, Horlocker T T, Hanssen A D. The impact of anesthetic management on surgical site infections in patients undergoing total knee or total hip arthroplasty. Anesth Analg 2015; 121(5): 1215-21. doi: 10.1213/ane.0000000000000956. Kunutsor S K, Whitehouse M R, Blom A W, Beswick A D. Patient-related risk factors for periprosthetic joint infection after total joint arthroplasty: a systematic review and meta-analysis. PLoS One 2016; 11(3): e0150866. doi: 10.1371/journal.pone.0150866. Liu J, Ma C, Elkassabany N, Fleisher L A, Neuman M D. Neuraxial anesthesia decreases postoperative systemic infection risk compared with general anesthesia in knee arthroplasty. Anesth Analg 2013; 117(4): 1010-6. doi: 10.1213/ANE.0b013e3182a1bf1c. Mangram A J, Horan T C, Pearson M L, Silver L C, Jarvis W R. Guideline for prevention of surgical site infection, 1999. Hospital Infection Control Practices Advisory Committee. Infect Control Hosp Epidemiol 1999; 20(4): 250-78; quiz 79-80. doi: 10.1086/501620. Moudgil G C. Update on anaesthesia and the immune response. Can Anaesth Soc J 1986; 33(3 Pt 2): S54-60. Osmon D R, Berbari E F, Berendt A R, Lew D, Zimmerli W, Steckelberg J M, Rao N, Hanssen A, Wilson W R. Executive summary: diagnosis and management of prosthetic joint infection: clinical practice guidelines by the Infectious Diseases Society of America. Clin Infect Dis 2013; 56(1): 1-10. doi: 10.1093/cid/cis966. Parvizi J, Gehrke T. Definition of periprosthetic joint infection. J Arthroplasty 2014; 29(7): 1331. doi: 10.1016/j.arth.2014.03.009. Pearl J. Understanding propensity scores. In: Causality: models, reasoning, and inference. Cambridge/New York: Cambridge University Press; 2010. p. 348-52. Pugely A J, Martin C T, Gao Y, Mendoza-Lattes S, Callaghan J J. Differences in short-term complications between spinal and general anesthesia for primary total knee arthroplasty. J Bone Joint Surg Am 2013; 95(3): 193-9. doi: 10.2106/jbjs.k.01682. Rubin D B. Estimating causal effects from large data sets using propensity scores. Ann Intern Med 1997; 127(8 Pt 2): 757-63. Scholten R, Leijtens B, Kremers K, Snoeck M, Koeter S. The incidence of mild hypothermia after total knee or hip arthroplasty: a study of 2600 patients. J Orthop 2018; 15(2): 408-11. doi: 10.1016/j.jor.2018.03.014. Sessler D I. Neuraxial anesthesia and surgical site infection. Anesthesiology 2010; 113(2): 265-7. doi: 10.1097/ALN.0b013e3181e2c1ed. Shrier I, Platt R W. Reducing bias through directed acyclic graphs. BMC Med Res Methodol 2008; 8: 70. doi: 10.1186/1471-2288-8-70. van Buuren S. Flexible imputation of missing data. Second ed. Boca Raton, FL: Chapman & Hall/CRC; 2018. Whelan P, Morris P J. Immunological responsiveness after transurethral resection of the prostate: general versus spinal anaesthetic. Clin Exp Immunol 1982; 48(3): 611-8. White I R, Royston P, Wood A M. Multiple imputation using chained equations: issues and guidance for practice. Stat Med 2011; 30(4): 377-99. doi: 10.1002/sim.4067. Zorrilla-Vaca A, Grant M C, Mathur V, Li J, Wu C L. The impact of neuraxial versus general anesthesia on the incidence of postoperative surgical site infections following knee or hip arthroplasty: a meta-analysis. Reg Anesth Pain Med 2016; 41(5): 555-63. doi: 10.1097/aap.0000000000000437.

Acta Orthopaedica 2019; 90 (6): 559–567

559

The effect of smoking on outcomes following primary total hip and knee arthroplasty: a population-based cohort study of 117,024 patients Gulraj S MATHARU 1,2, Sofia MOUCHTI 2, Sarah TWIGG 3,4, Antonella DELMESTRI 1, David W MURRAY 1, Andrew JUDGE 1,2,5, and Hemant G PANDIT 1,4 1 Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, University of Oxford, Nuffield Orthopaedic Centre, Oxford; 2 Musculoskeletal Research Unit, Translational Health Sciences, Bristol Medical School, University of Bristol; 3 Bradford Teaching Hospitals NHS Foundation Trust, St Luke’s Hospital, Bradford; 4 Leeds Institute of Rheumatic and Musculoskeletal Medicine, Chapel Allerton Hospital and University of Leeds, Leeds; 5 National Institute for Health Research Bristol Biomedical Research Centre (NIHR Bristol BRC), University Hospitals Bristol NHS Foundation

Trust, University of Bristol, Southmead Hospital, Bristol, UK Correspondence: gulraj.matharu@bristol.ac.uk Submitted 2019-02-15. Accepted

Background and purpose — Smoking is a modifiable risk factor that may adversely affect postoperative outcomes. Healthcare providers are increasingly denying smokers access to total hip and knee arthroplasty (THA and TKA) until they stop smoking. Evidence supporting this is unclear. We assessed the effect of smoking on outcomes following arthroplasty. Patients and methods — We identified THAs and TKAs from the Clinical Practice Research Datalink, which were linked with datasets from Hospital Episode Statistics and the Office for National Statistics to identify outcomes. The effect of smoking on postoperative outcomes (complications, medications, revision, mortality, patient-reported outcome measures [PROMs]) was assessed using adjusted regression models. Results — We studied 60,812 THAs and 56,212 TKAs (11% smokers, 33% ex-smokers, 57% non-smokers). Following THA, smokers had an increased risk of lower respiratory tract infection (LRTI) and myocardial infarction compared with non-smokers and ex-smokers. Following TKA, smokers had an increased risk of LRTI compared with non-smokers. Compared with non-smokers (THA relative risk ratio [RRR] = 0.65; 95% CI = 0.61–0.69; TKA RRR = 0.82; CI = 0.78–0.86) and ex-smokers (THR RRR = 0.90; CI = 0.84–0.95), smokers had increased opioid usage 1-year postoperatively. Similar patterns were observed for weak opioids, paracetamol, and gabapentinoids. 1-year mortality rates were higher in smokers compared with non-smokers (THA hazard ratio [HR] = 0.37, CI = 0.29–0.49; TKA HR = 0.52, CI = 0.34–0.81) and ex-smokers (THA HR = 0.53, CI = 0.40–0.70). Long-term revision rates were not increased in smokers. Smokers had improvement in PROMs compared with preoperatively, with no clinically important difference in postoperative PROMs between smokers, non-smokers, and ex-smokers.

Interpretation — Smoking is associated with more medical complications, higher analgesia usage, and increased mortality following arthroplasty. Most adverse outcomes were reduced in ex-smokers, therefore smoking cessation should be encouraged before arthroplasty.

Total hip arthroplasty (THA) and total knee arthroplasty (TKA) are well established clinically and are cost-effective interventions for treating symptomatic arthritis (Learmonth et al. 2007). These procedures are commonly performed worldwide, with numbers predicted to increase (Culliford et al. 2015). The UK National Health Service is currently under unprecedented financial pressures (Daily Telegraph 2018). In the UK, 197 clinical commissioning groups (CCGs) have the authority and funding to commission healthcare services for their communities. In recent years over half of CCGs have rationed THA and TKA to reduce healthcare expenditure; therefore patients with certain perceived risk factors (like smokers, or those with a high BMI) have been denied access to arthroplasty (Daily Telegraph 2018). A recent report highlighted the severity of the problem with almost 1,700 requests for THA and TKA rejected by CCGs between 2017 and 2018, which represented a 45% increase from the previous year with some CCGs rejecting almost all requests received (Iacobucci 2018a). These actions leave many patients in considerable pain for prolonged periods despite a clinically effective intervention being available, and it appears patients are increasingly accessing arthroplasty in the private sector (Iacobucci 2018a). Thus, the longstanding problem of health inequalities between socioeconomic groups is perpetuated. Rationing of THA and TKA has been strongly discouraged by surgical bodies (Royal College of Surgeons 2016, British

560

Orthopaedic Association [BOA] 2017). NICE (2017) recommends patient-specific factors (including age, sex, smoking, obesity, and comorbidities) should not be barriers to referral for arthroplasty. Smoking is a modifiable risk factor that is often perceived to adversely affect outcomes following surgery. However, there is insufficient evidence to support the CCGsâ&#x20AC;&#x2122; stance of denying current smokers access to arthroplasty. Studies of arthroplasty patients have observed that, compared with nonsmokers, smokers have increased wound complications, deep infection, chest infection, implant revision, hospital readmission, and mortality (Singh 2011, Duchman et al. 2015, Singh et al. 2015, Teng et al. 2015, Bohl et al. 2016, Tischler et al. 2017). However, these observations were not consistent between the different studies, with some studies reporting no effect of smoking on these same outcome measures (Inoue et al. 1999, Malik et al. 2004, Sadr Azodi et al. 2008, Musallam et al. 2013, Maoz et al. 2015, Cunningham et al. 2017, Sahota et al. 2018). The inconsistent findings of studies into the effect of smoking on outcomes following arthroplasty might be explained by their numerous limitations including the analysis of small cohorts, not separating THA and TKA patients for analysis, limited or no adjustment for confounding factors, providing only shortterm outcomes, and not assessing the effect of previous smoking on outcomes. Furthermore, many studies have not reported on important outcomes, like postoperative analgesia usage and patient-reported outcome measures (PROMs), which have not been assessed in large cohorts. The latter is pertinent given that clinically meaningful improvement in PROMs following arthroplasty is key in deciding whether or not to recommend joint replacement to patients (Wallace et al. 2014). Therefore it is difficult to support the implementation of a policy that denies access to arthroplasty for smokers based on current evidence. This population-based cohort study assesses the effect of smoking and cessation of smoking on postoperative outcomes following THA and TKA. For completeness we have studied complications, medication usage, hospital readmission, revision surgery, mortality, and PROMs in smokers, ex-smokers, and non-smokers.

Patients and methods Patients were initially identified using the Clinical Practice Research Datalink (CPRD) GOLD, which has been described previously (Bayliss et al. 2017). CPRD represents one of the largest databases of longitudinal primary care medical records worldwide. It contains anonymized patient data from 4% of the current UK population (over 2 million patients from 269 contributing practices) (Herrett et al. 2015). Practicesâ&#x20AC;&#x2122; spread ensures CPRD is representative of the wider UK population for age, sex, and ethnicity. Read Codes are used to enter clinical information (medical history, prescription data, hospital admissions, and interventions), which are standard clinical terminologies used within UK primary care (Benson 2011).

Acta Orthopaedica 2019; 90 (6): 559â&#x20AC;&#x201C;567

CPRD therefore provides a detailed record of both primary and secondary care (Bayliss et al. 2017). The validity and quality of data captured within CPRD have been previously well described (Herrett et al. 2015). A systematic review of validation studies assessing the validity of diagnoses in CPRD identified a large number of studies across a wide range of over 183 different diagnoses and overall estimates of validity were high (Herrett et al. 2010). Aspects of data quality in English primary care are enhanced by the Quality and Outcomes Framework, an incentive payment program for primary care physicians, which encourages recording of key data items (for example smoking status). Where available, primary care records from CPRD were linked to secondary care admission records from Hospital Episodes Statistics Admitted Patient Care data (HES) and to the Office for National Statistics (ONS) database. HES uses International Classification of Diseases 10th revision (ICD-10) records diagnoses and the Office of Population Censuses and Surveys version 4 (OPCS-4) procedures to record diseases, complications, interventions, and procedures from secondary care. From April 1, 2009, HES provided PROMs data before and 6 months following THA and TKA (see below). The ONS provides data on all-cause mortality. Population All patients aged 18 years and older in CPRD with a diagnostic code for primary THA or TKA between January 1, 1995 and January 28, 2017 were identified using previously validated Read Codes (Culliford et al. 2012, 2015). Patients were eligible for inclusion if their record was labelled acceptable by CPRD quality control (Herrett et al. 2015), approved for CPRD, HES, ONS linkage, and if the patient was registered with their general practice for at least 12 months (n = 136,410). Patients were excluded if data on the exposure variable were missing (n = 19,386) leaving 117,024 patients for analysis (Appendix 1). The study exposure, covariates, and outcomes were identified from the various linked databases using ICD-10 codes, OPCS-4 operation codes, Product Code lists for prescribed medications and Read Codes. Exposure The study exposure was patient smoking status as classified in CPRD at the time of arthroplasty: current smoker, ex-smoker, and non-smoker. Studies specifically assessing the quality of smoking data within CPRD demonstrate prevalence estimates for current smoking and non-smoking that are similar to those from nationally representative surveys, although former smoking may be under-recorded (Booth et al. 2013). Covariates CPRD contains numerous patient-related covariates, which were subsequently adjusted for. These included age, sex, BMI, socioeconomic status, alcohol consumption, year of surgery, and pre-existing comorbidities (including cardiovascular,

Acta Orthopaedica 2019; 90 (6): 559–567

respiratory, and cerebrovascular diseases, renal failure, cancer, inflammatory arthritis, diabetes). BMI was categorized as underweight (≥10 and <18.5); normal (≥18.5 and <25 ); overweight (≥25 and <30 ); obese class I (≥30 and <35 ); obese class II (≥35 and <40 ); and obese class III (≥40 and ≤60 ). Socioeconomic status was classified using the Index of Multiple Deprivation (IMD), as described previously (Conrad et al. 2018). Patients were divided into 10 equal groups ranked from 1 (least deprived area) to 10 (most deprived area) (Department for Communities and Local Government [DCLG] 2015). The Charlson Comorbidity Index was calculated for each patient based on pre-existing comorbidities. Preoperative PROMs were available for a subgroup of patients (detailed below). Outcomes Postoperative outcomes of interest were: complications, mortality, medication usage, hospital readmission, revision surgery, and PROMs. All outcome data were collected using a combination of CPRD and HES apart from mortality, which was obtained from CPRD and ONS. Complications were recorded within 6 months of surgery, and included cerebrovascular events, myocardial infarction, ischemic heart disease, deep vein thrombosis, pulmonary embolism, lower respiratory tract infection (LRTI), urinary tract infection, and wound infection. Medications prescribed (which included analgesia requirements) and hospital readmissions within 1 year of surgery were recorded. Revision surgery was defined as removal or exchange or addition of any implant(s), within 20 years of surgery. Mortality was assessed 1 year postoperatively. PROMs at 6 months following surgery were available in a subgroup of patients, which are collected as part of a national PROMs program. These joint-specific (Oxford Hip Score [OHS] and Oxford Knee Score [OKS]) PROMs are validated measures for assessing outcomes following arthroplasty. They are scored from 0 (worst) to 48 (best). The change in score following arthroplasty was calculated by subtracting the preoperative score from the 6-month postoperative score. The minimally important clinical differences are 5 points for the OHS and 4 points for the OKS (Beard et al. 2015). The validity of coding of complications has been previously assessed and known to be good (Wallace et al. 2014). Mortality rates within CPRD are comparable to rates in the National Joint Registry (NJR) annual reports (NJR 2018). Data on revision and readmission come from linked data from HES records, and the validity of coding between NJR and HES records has previously been described in NJR annual reports and data quality audits (NJR 2018). Furthermore validation studies of joint replacement records in CPRD and HES showed high levels of agreement (Hawley et al. 2016). Statistics We assessed the effect of smoking status on binary outcomes (complications, medication usage, and readmissions) by fitting a generalized linear model with a binomial error structure

561

and a log link function (log-logistic model), where results are presented as relative risk ratios (RRR). Models were adjusted for potential confounding factors (age, sex, BMI, IMD, alcohol consumption, year of surgery, and the Charlson Comorbidity Index). Cumulative implant and patient survival rates following arthroplasty were determined using Kaplan–Meier estimates. Patients who were alive with an arthroplasty not requiring revision surgery were censored on the study end date. The associations between smoking status with implant and patient survival rates were explored using Cox regression analysis, with models adjusted for confounders (see above). The confounders we adjusted for were selected a priori given that they have been shown to affect the study exposure, outcomes, and/or access to healthcare (Hunt et al. 2013, 2014, Wallace et al. 2014, Kunutsor et al. 2016, AOANJRR 2018, Edwards et al. 2018). Proportional hazards assumptions were assessed and satisfied for all regression analyses. We used an ANCOVA linear regression model to look at predictors of the obtained 6-month OHS and OKS, adjusting for the baseline score. To look at changes in scores between baseline and follow-up we fitted a repeated measures regression model where the outcome was expanded to include the preoperative and 6-month postoperative OHS or OKS. Interaction terms were fitted between the predictor variable and time, to describe the change in OHS and OKS over time across categories of the predictor variable smoking. We used robust standard errors with the sandwich variance estimator given there was evidence of heteroscedasticity. All models were based on complete case analysis. All statistical analyses were performed with Stata Statistical Software release version IC 15.0, 2017 (StataCorp, Stata College Station, TX, USA). Ethics, funding, and potential conflicts of interest Ethical approval was not required for this study. The Clinical Practice Research Datalink (CPRD) Group has obtained ethical approval from a National Research Ethics Service Committee for all purely observational research with anonymized CPRD data (i.e., studies that do not include patient involvement). The study was approved by the Independent Scientific Advisory Committee (ISAC) for MHRA Database Research, protocol number 17_104R. Funding was received for this study from departmental funds held at the University of Leeds, England. AJ is supported by the NIHR Biomedical Research Centre at the University Hospitals Bristol NHS Foundation Trust and the University of Bristol. GSM has received financial support for other research work from Arthritis Research UK, the Orthopaedics Trust, Royal College of Surgeons of England, and Royal Orthopaedic Hospital Hip Research and Education Charitable Fund. GSM has also received personal fees for undertaking medicolegal work for Leigh Day. SM, ST, and AD have no relevant conflicts of interest. AJ has received consultancy fees from Freshfields Bruckhaus Deringer, and is a paid member of the data safety and monitoring board for Anthera Pharmaceuticals. DWM

562

Acta Orthopaedica 2019; 90 (6): 559–567

Table 1. Demographics of patients undergoing total hip arthroplasty. Values are frequency (%) unless otherwise stated Confounder

Smoker Non-smoker Ex-smoker

Table 2. Demographics of patients undergoing total knee arthroplasty. Values are frequency (%) unless otherwise stated Confounder

Smoker Non-smoker Ex-smoker

Total 7,543 (12) 34,271 (56) 18,998 (31) Male 3,059 (41) 11,343 (33) 9,343 (49) Female 4,484 (59) 22,928 (67) 9,655 (51) Age, mean (SD) 63 (12) 70 (11) 70 (10) BMI Underweight 228 (3.7) 526 (1.8) 227 (1.4) Normal 2,118 (34) 8,692 (30) 3,913 (23) Overweight 2,217 (36) 11,771 (40) 6,866 (41) Obese class I 1,140 (19) 5,852 (20) 4,015 (24) Obese class II 365 (5.9) 1,849 (6.3) 1,356 (8.1) Obese class III 110 (1.8) 607 (2.1) 392 (2.3) Missing 1,365 4,974 2,229 Charlson score 1 year a 0 6,962 (92) 31,437 (92) 16,874 (89) 1 325 (4.3) 1,398 (4.1) 1,052 (5.5) 2 206 (2.7) 1,148 (3.3) 803 (4.2) ≥ 3 50 (0.7) 288 (0.8) 269 (1.4) Deprivation Index rank 1 416 (9.6) 3,446 (17) 1,574 (14) 2 418 (9.7) 2,862 (14) 1,453 (13) 3 436 (10) 2,697 (13) 1,444 (13) 4 469 (11) 2,552 (12) 1,387 (12) 5 488 (11) 2,590 (13) 1,480 (13) 6 418 (9.7) 1,908 (9.2) 1,092 (9.4) 7 389 (9.0) 1,635 (7.9) 995 (8.6) 8 402 (9.3) 1,283 (6.2) 882 (7.6) 9 426 (9.9) 959 (4.6) 675 (5.8) 10 461 (11) 756 (3.7) 599 (5.2) Missing 3,220 13,583 7,417 Alcohol consumption Yes 4,691 (80) 21,151 (78) 13,029 (84) No 898 (15) 5,628 (21) 1,854 (12) Ex 243 (4.2) 530 (1.9) 555 (3.6) Missing 1,711 6,962 3,560 Year of surgery 1995–2000 738 (9.8) 2,980 (8.7) 736 (3.9) 2001–2010 4,291 (57) 18,864 (55) 10,710 (56) 2011–2016 2,514 (33) 12,427 (36) 7,552 (40) Preoperative Oxford Hip Score median (IQR) 16 (11–21) 18 (13–24) 18 (12–24)

Total 5,101 (9.1) 31,961 (57) 19,150 (34) Male 2,597 (51) 10,910 (34) 10,997 (57) Female 2,504 (49) 21,051 (66) 8,153 (43) Age, mean (SD) 64 (10) 70 (10) 70 (9) BMI Underweight 44 (1.0) 138 (0.5) 63 (0.4) Normal 844 (20) 4,811 (17) 2,300 (13) Overweight 1,628 (38) 10,567 (38) 6,696 (39) Obese class I 1,144 (26) 7,744 (28) 5,179 (30) Obese class II 488 (11) 3,526 (13) 2,162 (13) Obese class III 178 (4.1) 1,390 (4.9) 855 (5.0) Missing 775 3,785 1,895 Charlson score 1 year a 0 4,646 (91) 28,994 (91) 16,985 (89) 1 249 (4.9) 1,451 (4.5) 1,062 (5.5) 2 161 (3.2) 1,190 (3.7) 857 (4.5) ≥ 3 45 (0.9) 326 (1.0) 246 (1.3) Deprivation Index rank 1 253 (8.2) 2,932 (15) 1,496 (13) 2 272 (8.8) 2,396 (12) 1,407 (12) 3 276 (8.9) 2,427 (12) 1,459 (12) 4 310 (10) 2,455 (13) 1,431 (12) 5 358 (12) 2,363 (12) 1,451 (12) 6 312 (10) 1,907 (9.7) 1,208 (10) 7 320 (10) 1,709 (8.7) 1,033 (8.7) 8 328 (11) 1,445 (7.3) 982 (8.2) 9 310 (10) 1,092 (5.5) 764 (6.4) 10 357 (12) 964 (4.9) 705 (5.9) Missing 2,005 12,271 7,214 Alcohol consumption Yes 3,129 (79) 19,436 (75) 13,348 (85) No 660 (17) 5,809 (23) 1,859 (12) Ex 173 (4.4) 550 (2.1) 570 (3.6) Missing 1,139 6,166 3,373 Year of surgery 1995–2000 357 (7.0) 1,917 (6.0) 524 (2.7) 2001–2010 3,074 (60) 18,115 (57) 11,101 (58) 2011–2016 1,670 (33) 11,929 (37) 7,525 (39) Preoperative Oxford Knee Score median (IQR) 17 (12–23) 19 (14–25) 20 (14–25)

a Charlson index score based on comorbidities 1 year prior to index date.

receives royalties from Zimmer Biomet. DWM and HGP are paid consultants for Zimmer Biomet, and both receive institutional research funding from Zimmer Biomet. HGP is also a paid consultant for Kennedys Law, Bristol Myers Squibb, Depuy Synthes, Medacta Int and Meril Life. He has received institutional research grants from UKIERI, Charnley Trust, Depuy Synthes, Glaxo Smith Kline and NIHR.

Complications Following THA, smokers had an increased risk of myocardial infarction and LRTI compared with both non-smokers and exsmokers (Figure 1; Appendix 2 and 3). Smokers had a similar risk of wound infection and thromboembolic events compared with non-smokers and ex-smokers following THA as well as TKA. Following TKA, only LRTI was more commonly observed in smokers compared with non-smokers (RRR = 0.66, CI 0.52–0.83) (Figure 2; Appendix 3 and 4).

Results

Medication usage Within 1 year of THA, smokers had increased use of opioids compared with non-smokers (RRR = 0.65; CI 0.61–0.69) and ex-smokers (RRR = 0.90; CI 0.84–0.95) (Figure 1; Appendix 2 and 3). Smokers also had increased use of weak opioids,

We studied 117,024 patients undergoing arthroplasty (60,812 THAs and 56,212 TKAs) with details available on smoking status (12,644 (11%) smokers, 38,148 (33%) ex-smokers, and 66,232 (57%) non-smokers) (Tables 1 and 2).

Acta Orthopaedica 2019; 90 (6): 559–567

563

Relative Risk Ratio (95% CI)

Exposure

Relative Risk Ratio (95% CI)

Exposure

Cerebro−vascular disease Non−smoker Ex−smoker

Deep vein thrombosis Non−smoker Ex−smoker

Ischemic heart disease Non−smoker Ex−smoker

Lower respiratory tract infection Non−smoker Ex−smoker

Myocardial infarction Non−smoker Ex−smoker

Pulmonary embolism Non−smoker Ex−smoker

Wound infection Non−smoker Ex−smoker

Urinary tract infection Non−smoker Ex−smoker

Hospital readmission Non−smoker Ex−smoker

Gabapentinoids Non−smoker Ex−smoker

NSAID Non−smoker Ex−smoker

Opioids Non−smoker Ex−smoker

Weak opioids Non−smoker Ex−smoker

Paracetamol Non−smoker Ex−smoker

0.2

Figure 1. Forest plot for complications and medication usage following total hip arthroplasty by smoking status. The respective relative risk ratios and 95% confidence intervals are provided in Appendix 3.

Figure 2. Forest plot for complications and medication usage following total knee arthroplasty by smoking status. The respective relative risk ratios and 95% confidence intervals are provided in Appendix 3.

paracetamol, and gabapentinoids compared with both nonsmokers and ex-smokers following THA. Similar patterns of increased analgesia use in smokers were observed following TKA (Figure 2; Appendix 3 and 4).

and ex-smokers (HR = 0.53, CI 0.40–0.70) (Figure 3). Following TKA, 1-year mortality rates were higher in smokers compared with non-smokers only (HR = 0.52, CI 0.34–0.81),

Readmission The risk of hospital readmission in smokers following THA was higher compared with non-smokers, but not compared with ex-smokers (Figure 1). The risk of hospital readmission was not affected by smoking status following TKA (Figure 2). Revision surgery The risk of revision up to 20 years following THA was similar in smokers compared with non-smokers (hazard ratio (HR) = 1.1; CI 0.88–1.5) and ex-smokers (HR = 1.1; CI 0.84–1.5). The risk of revision following TKA was similar in smokers compared with non-smokers (HR = 1.2; CI 0.90–1.6) and exsmokers (HR 1.1; CI = 0.78–1.4).

Cumulative probability of death (%) 3.0 Smoker Non-smoker Ex-smoker

2.5

2.0

1.5

1.0

0.5

0 0

Mortality Following THA, 1-year mortality rates were higher in smokers compared with non-smokers (HR = 0.37, CI 0.29–0.49)

90 120 150 180 210 240 270 300 330 360

Days since index operation

Figure 3. Cumulative probability of mortality up to 1 year following total hip arthroplasty.

564

Acta Orthopaedica 2019; 90 (6): 559–567

Cumulative probability of death (%) 1.50 Smoker Non-smoker Ex-smoker

1.25

Oxford Hip Score

Oxford Knee Score

Smoker Non-smoker Ex-smoker

38 36

1.00

0.75

0.50

0.25

0 0

90 120 150 180 210 240 270 300 330 360

Days since index operation

Figure 4. Cumulative probability of mortality up to 1 year following total knee arthroplasty.

Smoker Non-smoker Ex-smoker

14 0

Months since index operation

Figure 5. Estimation of the mean predicted preoperative (0 months) and postoperative (6 months) Oxford Hip Score by smoking status for patients receiving total hip arthroplasty.

but not compared with ex-smokers (HR = 0.71, CI 0.46–1.1) (Figure 4). Patient-reported outcome measures PROMs were available for 10,009 (8.6%) patients. Smokers had improvement in PROMs compared with the preoperative scores (Figures 5 and 6, and Appendix 5). Smokers undergoing THA and TKA had lower postoperative PROMs compared with non-smokers and ex-smokers; however, these differences were not clinically meaningful. Following THA, smokers had lower postoperative OHSs compared with non-smokers (mean 2.5 points; CI 1.5–3.5) and ex-smokers (mean 1.8 points; CI 0.79–2.9). Following TKA, smokers had lower postoperative OKSs compared with non-smokers (mean 3.2 points; CI 2.0– 4.5) and ex-smokers (mean 2.9 points; CI 1.7–4.2).

Discussion Smoking was associated with an increased risk of medical complications, increased analgesia usage, and higher mortality following arthroplasty. The increased risk of LRTI (Bohl et al. 2016) and cardiovascular complications (Ockene and Miller 1997, Hunt et al. 2017) in smokers undergoing arthroplasty in this study was similar to previous observations. However, there was no increased risk of cardiovascular complications following TKA in smokers compared with ex-smokers and non-smokers, which could represent selection bias for undergoing these procedures. Contrary to some studies (Duchman et al. 2015, Bohl et al. 2016, Sahota et al. 2018), we found no increase in wound infections in smokers following arthroplasty, and the risks of venous thromboembolism were similar between groups. The

Months since index operation

Figure 6. Estimation of the mean predicted preoperative (0 months) and postoperative (6 months) Oxford Knee Score by smoking status for patients receiving total knee arthroplasty.

differences in cohort size and study design may explain some of these findings, including our separate analyses for THA and TKA, and differences in follow-up period. Previous studies have assessed outcomes at 30 days (Duchman et al. 2015, Sahota et al. 2018) despite needing at least 90 days to report postoperative infections (Centers for Disease Control 2018). Arthroplasty patients who smoked had higher 1-year mortality rates compared with non-smokers and ex-smokers (the latter for THA only), with similar observations reported previously (Singh et al. 2011, Clement et al. 2012). However, it is recognized that mortality rates for smokers in the general population are known to be 2 to 3 times higher compared with nonsmokers (Thun et al. 2013, Carter et al. 2015). Increased opioid usage in smokers following arthroplasty has been reported (Kim et al. 2017, Cryar et al. 2018). Within 1 year of arthroplasty we observed increased use of paracetamol, non-steroidal anti-inflammatory drugs, weak opioids (including codeine), strong opioids (including morphine), and gabapentinoids in smokers compared with non-smokers. This is concerning given the current worldwide opioid epidemic (Brat et al. 2018) coupled with the projected increase in arthroplasty (Culliford et al. 2015). This poses significant public health risks including the development of opioid dependence and opioid-related deaths (Brat et al. 2018). All clinicians must therefore remain cognizant of the increased analgesic needs of smokers following arthroplasty and use non-opioid medications when possible. Smokers had a similar risk of long-term revision (at up to 20 years postoperatively) compared with non-smokers and exsmokers. Other studies observing an increased revision risk have focused on short-term outcomes (Duchman et al. 2015, Teng et al. 2015). Furthermore, smokers obtained clinically meaningful PROM improvement following arthroplasty, with

Acta Orthopaedica 2019; 90 (6): 559â&#x20AC;&#x201C;567

postoperative PROMs comparable with non-smokers and exsmokers. Other smaller studies have suggested that smoking does not adversely influence postoperative PROMs (Fisher et al. 2007, Khan et al. 2009). Our findings therefore suggest that arthroplasty is clinically effective in smokers, which is important to recognize given the increasing pressures from some healthcare providers to deny arthroplasty to patients who continue to smoke (RCS 2016, BOA 2017, Iacobucci 2018a). Most adverse outcomes, namely complications and mortality, in smokers (versus non-smokers) were not seen in ex-smokers. This suggests that stopping smoking prior to arthroplasty may reduce postoperative risks. Fewer studies have assessed the effect of cessation of smoking (ex-smoking) on outcomes following arthroplasty (Singh 2011). These previous studies generally suggest that ex-smoking can still be associated with complications and mortality following arthroplasty. However, this variance could again relate to methodological differences between our study and previous work. The question of whether or not smokers should quit before arthroplasty raises an important ethical dilemma. On one hand there is a proven and clinically effective operation available that can substantially reduce the pain and disability associated with hip and knee arthritis. On the other hand, by proceeding with elective arthroplasty in current smokers there are increased risks related to postoperative medical complications, mortality, and analgesia usage, which are all arguably avoidable. An early randomized controlled trial reported fewer postoperative complications in smokers who either quit or reduced their smoking by 50% prior to THA and TKA compared with those continuing to smoke (Moller et al. 2002). Subsequently strong evidence has been published regarding the benefits of various smoking cessation methods before surgery (Myers et al. 2011, Thomsen et al. 2014). The latest Cochrane review, which included 13 trials assessing smoking cessation, concluded that preoperative smoking interventions, which provide behavioral support and offer nicotine replacement therapy, can increase cessation in the short term and may reduce postoperative morbidity (Thomsen et al. 2014). Although the optimal preoperative intensity remains unknown, the authors suggested that interventions beginning 4 to 8 weeks before surgery, including weekly counselling and nicotine replacement therapy, were more likely to have an impact on complications and on longâ&#x20AC;&#x201C;term smoking cessation. On the basis of our data and the existing literature we recommend that healthcare professionals actively promote smoking cessation preoperatively in patients eligible for arthroplasty, as this will prevent the increased surgical risk associated with smoking and ultimately will improve patient safety. Preoperative smoking cessation also has the advantage of promoting long-term smoking abstinence (Rigotti et al. 2008, Thomsen et al. 2009). However, it is somewhat concerning that smoking cessation funding is currently being substantially reduced or removed altogether in some areas of England (Iacobucci

565

2018b), which will undoubtedly influence who has access to this evidence-based support and the quality of it. Eligible arthroplasty patients who are unable to quit smoking despite undergoing cessation therapy will also continue to pose a dilemma. Although there is no clear evidence as to what to do in these circumstances, it is advised that the surgeon reconsiders the indication for surgery and balances the benefits and risks together with the patient, anesthesiologist, and other relevant healthcare professionals. Strengths and limitations This is by far the largest study assessing the effect of smoking on outcomes following arthroplasty. We suspect the findings are generalizable to many Western populations. Contrary to other studies, we have importantly considered outcomes in ex-smokers, subdivided the cohort into THA and TKA, and explored short-term and long-term postoperative outcomes, including PROMs, implant and patient survival, and medication usage. Therefore our findings provide the most comprehensive picture of the outcomes patients will achieve following arthroplasty based on smoking status, and they provide useful information for healthcare professionals when counselling patients regarding the relative risks associated with each arthroplasty procedure. Using observational data means we cannot infer causality. Although the validity of clinical diagnoses in CPRD is high (Herrett et al. 2010), it is possible that some inaccuracies in coding exist within the dataset analyzed. We acknowledge that a limitation of using routinely available data from primary care is that information on smoking status is captured broadly in only 3 categories. Also, more detailed information on this exposure (cigarettes per day, tar and tobacco content etc.) was not available. Thus it was not possible to analyze any potential doseâ&#x20AC;&#x201C;response relationship related to the outcomes. Given the length of time since stopping smoking was not known, by categorizing the short-term quitters together with long-term quitters it is possible this could increase the complication rate for the total ex-smoking group, thereby reducing a difference between the ex-smoking and smoking groups. We also recognize that former smoking may be under-recorded in CPRD, which may influence the findings in this particular group (Booth et al. 2013). Some variables had missing data, which may have affected the findings. Most incomplete variables only had a small proportion of missing data. However, the later introduction of PROMs into HES (and availability of only 6-month postsurgery PROMs) means that much fewer patients had PROMs available. Thus care should be taken when interpreting results relating to PROMs. Some outcome measures, such as revision, can be underestimated (Sabah et al. 2016), but there is no reason to suspect any underestimation would differ between the exposure groups. Although we identified differences in all-cause mortality between smoking groups, data were not available on cause-specific mortality; therefore we cannot

566

comment on the causes of death and whether they were smoking related. Furthermore, given how the CPRD population was selected, our findings may not apply to other populations worldwide with different patient characteristics and/or healthcare practices. Conclusions This large observational cohort study demonstrated that smokers gain benefit from arthroplasty, with clinically meaningful PROM improvement and no increased revision risk. However, smoking was associated with more medical complications, analgesia usage, and death following arthroplasty. Most adverse outcomes were reduced in ex-smokers compared with smokers, suggesting that preoperative smoking cessation may improve outcomes following arthroplasty and thus should be encouraged by healthcare professionals. Supplementary data Appendices 1–5 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/ 17453674.2019.1649510

GSM, DWM, AJ, and HGP conceived and designed the study. GSM, ST, AD, DWM, AJ, and HGP contributed to acquiring the data. AD cleaned the data. SM and AJ did the statistical analysis. All authors contributed to interpreting the data and findings. The manuscript was initially drafted by GSM, SM, and ST, with all other authors involved in revising the manuscript. AJ is the guarantor of the data. This study is based in part on data from the CPRD obtained under license from the UK Medicines and Healthcare products Regulatory Agency. However, the interpretation and conclusions contained in this study are those of the authors alone. This article presents independent research supported by the National Institute for Health Research (NIHR) Leeds Biomedical Research Centre (BRC). The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care. Acta thanks Johanna Adami and Hanne Tönnesen for help with peer review of this study.

AOANJRR. Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) Hip, Knee & Shoulder Arthroplasty Annual Report; 2018.: https://aoanjrr.sahmri.com/annual-reports-2018. Bayliss L E, Culliford D, Monk A P, Glyn-Jones S, Prieto-Alhambra D, Judge A, Cooper C, Carr A J, Arden N K, Beard D J, Price A J. The effect of patient age at intervention on risk of implant revision after total replacement of the hip or knee: a population-based cohort study. Lancet 2017; 389(10077):1424-30. Beard D J, Harris K, Dawson J, Doll H, Murray D W, Carr A J, Price A J. Meaningful changes for the Oxford hip and knee scores after joint replacement surgery. J Clin Epidemiol 2015; 68(1): 73-9. Benson T. The history of the Read Codes: the inaugural James Read Memorial Lecture 2011. Inform Prim Care 2011; 19(3): 173-82. Bohl D D, Sershon R A, Fillingham Y A, Della Valle C J. Incidence, risk factors, and sources of sepsis following total joint arthroplasty. J Arthroplasty 2016; 31(12): 2875-79.e2.

Acta Orthopaedica 2019; 90 (6): 559–567

Booth H P, Prevost A T, Gulliford M C. Validity of smoking prevalence estimates from primary care electronic health records compared with national population survey data for England, 2007 to 2011. Pharmacoepidemiol Drug Saf 2013; 22(12): 1357-61. Brat G A, Agniel D, Beam A, Yorkgitis B, Bicket M, Homer M, Fox K P, Knecht D B, McMahill-Walraven C N, Palmer N, Kohane I. Postsurgical prescriptions for opioid naive patients and association with overdose and misuse: retrospective cohort study. BMJ 2018; 360: j5790. British Orthopaedic Association. Arbitrary barriers: rationing of orthopaedic services; 9 August 2017. https://www.boa.ac.uk/wp-content/ uploads/2017/08/BOA-Statement-Rationing-of-Orthopaedic-Services.pdf. Carter B D, Abnet C C, Feskanich D, Freedman N D, Hartge P, Lewis C E, Ockene J K, Prentice R L, Speizer F E, Thun M J, Jacobs E J. Smoking and mortality: beyond established causes. N Engl J Med 2015; 372(7): 631-40. Centers for Disease Control and Prevention: Surgical site infection (SSI) event; January 1, 2018. https://www.cdc.gov/nhsn/pdfs/pscmanual/9pscssicurrent.pdf. Clement N D, Jenkins P J, Brenkel I J, Walmsley P. Predictors of mortality after total knee replacement: a ten-year survivorship analysis. J Bone Joint Surg Br 2012; 94(2): 200-4. Conrad N, Judge A, Tran J, Mohseni H, Hedgecott D, Crespillo A P, Allison M, Hemingway H, Cleland J G, McMurray J J V, Rahimi K. Temporal trends and patterns in heart failure incidence: a population-based study of 4 million individuals. Lancet 2018; 391(10120): 572-80. Cryar K A, Hereford T, Edwards P K, Siegel E, Barnes C L, Mears S C. Preoperative smoking and narcotic, benzodiazepine, and tramadol use are risk factors for narcotic use after hip and knee arthroplasty. J Arthroplasty 2018; 33(9): 2774-79. Culliford D J, Maskell J, Kiran A, Judge A, Javaid M K, Cooper C, Arden N K. The lifetime risk of total hip and knee arthroplasty: results from the UK general practice research database. Osteoarthritis Cartilage 2012; 20(6): 519-24. Culliford D, Maskell J, Judge A, Cooper C, Prieto-Alhambra D, Arden N K, Group C O S. Future projections of total hip and knee arthroplasty in the UK: results from the UK Clinical Practice Research Datalink. Osteoarthritis Cartilage 2015; 23(4): 594-600. Cunningham D J, Kavolus J J, 2nd, Bolognesi M P, Wellman S S, Seyler T M. Common medical comorbidities correlated with poor outcomes in hip periprosthetic infection. J Arthroplasty 2017; 32(9S): S241-S45 e3. Daily Telegraph. NHS faces £1bn deficit and widespread shortages of staff. Daily Telegraph February 21, 2018. https://www.telegraph.co.uk/ news/2018/02/21/nhs-faces-1bn-deficit-widespread-shortages-staff/. Department for Communities and Local Government (DCLG). The English Index of Multiple Deprivation; 2015. https://assets.publishing.service.gov. uk/government/uploads/system/uploads/attachment_data/file/465791/English_Indices_of_Deprivation_2015_-_Statistical_Release.pdf. Duchman K R, Gao Y, Pugely A J, Martin C T, Noiseux N O, Callaghan J J. The effect of smoking on short-term complications following total hip and knee arthroplasty. J Bone Joint Surg Am 2015; 97(13): 1049-58. Edwards H B, Smith M, Herrett E, MacGregor A, Blom A, Ben-Shlomo Y. The effect of age, sex, area deprivation, and living arrangements on total knee replacement outcomes: a study involving the United Kingdom National Joint Registry dataset. JBJS Open Access 2018; 3(2): e0042. Fisher D A, Dierckman B, Watts M R, Davis K. Looks good but feels bad: factors that contribute to poor results after total knee arthroplasty. J Arthroplasty 2007; 22(6 Suppl. 2): 39-42. Hawley S, Delmestri A, Judge A, Edwards C J, Cooper C, Arden N K, PrietoAlhambra D. Total hip and knee replacement among incident osteoarthritis and rheumatoid arthritis patients within the UK Clinical Practice Research Datalink (CPRD) compared to hospital episode statistics (HES): a validation study. Pharmacoepidemiol Drug Saf 2016; 25(Suppl. S3): 251-51. Herrett E, Thomas S L, Schoonen W M, Smeeth L, Hall A J. Validation and validity of diagnoses in the General Practice Research Database: a systematic review. Br J Clin Pharmacol 2010; 69(1): 4-14. Herrett E, Gallagher A M, Bhaskaran K, Forbes H, Mathur R, van Staa T, Smeeth L. Data resource profile: Clinical Practice Research Datalink (CPRD). Int J Epidemiol 2015; 44(3): 827-36.

Acta Orthopaedica 2019; 90 (6): 559â&#x20AC;&#x201C;567

Hunt L P, Ben-Shlomo Y, Clark E M, Dieppe P, Judge A, MacGregor A J, Tobias J H, Vernon K, Blom A W, National Joint Registry for England W, Northern I. 90-day mortality after 409,096 total hip replacements for osteoarthritis, from the National Joint Registry for England and Wales: a retrospective analysis. Lancet 2013; 382(9898): 1097-104. Hunt L P, Ben-Shlomo Y, Clark E M, Dieppe P, Judge A, MacGregor A J, Tobias J H, Vernon K, Blom A W, National Joint Registry for E, Wales. 45-day mortality after 467,779 knee replacements for osteoarthritis from the National Joint Registry for England and Wales: an observational study. Lancet 2014; 384(9952): 1429-36. Hunt L P, Ben-Shlomo Y, Whitehouse M R, Porter M L, Blom A W. The main cause of death following primary total hip and knee replacement for osteoarthritis: a cohort study of 26,766 deaths following 332,734 hip replacements and 29,802 deaths following 384,291 knee replacements. J Bone Joint Surg Am 2017; 99(7): 565-75. Iacobucci G. Nearly 1700 requests for knee and hip surgery were rejected in England last year. BMJ 2018a; 362: k3002. Iacobucci G. Stop smoking services: BMJ analysis shows how councils are stubbing them out. BMJ 2018b; 362: k3649. Inoue K, Ushiyama T, Tani Y, Hukuda S. Sociodemographic factors and failure of hip arthroplasty. Int Orthop 1999; 23(6): 330-3. Khan L A, Cowie J G, Ballantyne J A, Brenkel I J. The complication rate and medium-term functional outcome after total hip replacement in smokers. Hip Int 2009; 19(1): 47-51. Kim S C, Choudhry N, Franklin J M, Bykov K, Eikermann M, Lii J, Fischer M A, Bateman B T. Patterns and predictors of persistent opioid use following hip or knee arthroplasty. Osteoarthritis Cartilage 2017; 25(9): 1399-406. Kunutsor S K, Whitehouse M R, Blom A W, Beswick A D, Team I. Patientrelated risk factors for periprosthetic joint infection after total joint arthroplasty: a systematic review and meta-analysis. PLoS One 2016; 11(3): e0150866. Learmonth I D, Young C, Rorabeck C. The operation of the century: total hip replacement. Lancet 2007; 370(9597): 1508-19. Malik M H, Gray J, Kay P R. Early aseptic loosening of cemented total hip arthroplasty: the influence of non-steroidal anti-inflammatory drugs and smoking. Int Orthop 2004; 28(4): 211-13. Maoz G, Phillips M, Bosco J, Slover J, Stachel A, Inneh I, Iorio R. The Otto Aufranc Award: Modifiable versus nonmodifiable risk factors for infection after hip arthroplasty. Clin Orthop Relat Res 2015; 473(2): 453-9. Moller A M, Villebro N, Pedersen T, Tonnesen H. Effect of preoperative smoking intervention on postoperative complications: a randomised clinical trial. Lancet 2002; 359(9301): 114-17. Musallam K M, Rosendaal F R, Zaatari G, Soweid A, Hoballah J J, Sfeir P M, Zeineldine S, Tamim H M, Richards T, Spahn D R, Lotta L A, Peyvandi F, Jamali F R. Smoking and the risk of mortality and vascular and respiratory events in patients undergoing major surgery. JAMA Surg 2013; 148(8): 755-62. Myers K, Hajek P, Hinds C, McRobbie H. Stopping smoking shortly before surgery and postoperative complications: a systematic review and metaanalysis. Arch Intern Med 2011; 171(11): 983-9. National Joint Registry (NJR) for England, Wales, Northern Ireland and the Isle of Man. 15th Annual Report; 2018. http://www.njrreports.org.uk/Portals/0/PDFdownloads/NJR 15th Annual Report 2018.pdf.

567

NICE. Osteoarthritis care and management. Clinical Guideline 177. February 2014; updated 2017. Ockene I S, Miller N H. Cigarette smoking, cardiovascular disease, and stroke: a statement for healthcare professionals from the American Heart Association. American Heart Association Task Force on Risk Reduction. Circulation 1997; 96(9): 3243-7. Rigotti N A, Munafo M R, Stead L F. Smoking cessation interventions for hospitalized smokers: a systematic review. Arch Intern Med 2008; 168(18): 1950-60. Royal College of Surgeons of England. Smokers and overweight patients: soft targets for NHS savings? 2016. https://www.rcseng.ac.uk/library-andpublications/rcs-publications/docs/smokers-soft-targets/. Sabah S A, Henckel J, Koutsouris S, Rajani R, Hothi H, Skinner J A, Hart A J. Are all metal-on-metal hip revision operations contributing to the National Joint Registry implant survival curves?: a study comparing the London Implant Retrieval Centre and National Joint Registry datasets. Bone Joint J 2016; 98-B(1): 33-9. Sadr Azodi O, Adami J, Lindstrom D, Eriksson K O, Wladis A, Bellocco R. High body mass index is associated with increased risk of implant dislocation following primary total hip replacement: 2,106 patients followed for up to 8 years. Acta Orthop 2008; 79(1): 141-7. Sahota S, Lovecchio F, Harold R E, Beal M D, Manning D W. The effect of smoking on thirty-day postoperative complications after total joint arthroplasty: a propensity score-matched analysis. J Arthroplasty 2018; 33(1): 30-5. Singh J A. Smoking and outcomes after knee and hip arthroplasty: a systematic review. J Rheumatol 2011; 38(9): 1824-34. Singh J A, Houston T K, Ponce B A, Maddox G, Bishop M J, Richman J, Campagna E J, Henderson W G, Hawn M T. Smoking as a risk factor for short-term outcomes following primary total hip and total knee replacement in veterans. Arthritis Care Res (Hoboken) 2011; 63(10): 1365-74. Singh J A, Schleck C, Harmsen W S, Jacob A K, Warner D O, Lewallen D G. Current tobacco use is associated with higher rates of implant revision and deep infection after total hip or knee arthroplasty: a prospective cohort study. BMC Med 2015; 13: 283. Teng S, Yi C, Krettek C, Jagodzinski M. Smoking and risk of prosthesisrelated complications after total hip arthroplasty: a meta-analysis of cohort studies. PLoS One 2015; 10(4): e0125294. Thomsen T, Tonnesen H, Moller A M. Effect of preoperative smoking cessation interventions on postoperative complications and smoking cessation. Br J Surg 2009; 96(5): 451-61. Thomsen T, Villebro N, Moller A M. Interventions for preoperative smoking cessation. Cochrane Database Syst Rev 2014; (3): CD002294. Thun M J, Carter B D, Feskanich D, Freedman N D, Prentice R, Lopez A D, Hartge P, Gapstur S M. 50-year trends in smoking-related mortality in the United States. N Engl J Med 2013; 368(4): 351-64. Tischler E H, Matsen Ko L, Chen A F, Maltenfort M G, Schroeder J, Austin M S. Smoking increases the rate of reoperation for infection within 90 days after primary total joint arthroplasty. J Bone Joint Surg Am 2017; 99(4): 295-304. Wallace G, Judge A, Prieto-Alhambra D, de Vries F, Arden N K, Cooper C. The effect of body mass index on the risk of post-operative complications during the 6 months following total hip replacement or total knee replacement surgery. Osteoarthritis Cartilage 2014; 22(7): 918-27.

568

Acta Orthopaedica 2019; 90 (6): 568–574

Rates of knee arthroplasty in anterior cruciate ligament reconstructed patients: a longitudinal cohort study of 111,212 procedures over 20 years Simon G F ABRAM 1,2, Andrew JUDGE 1–4, Tanvir KHAN 1,2,5, David J BEARD 1,2, and Andrew J PRICE 1,2 1 Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, University of Oxford; 3 Musculoskeletal Research Unit, University of Bristol; 4 NIHR Biomedical Research Centre,

University of Nottingham, UK Correspondence: simon.abram@ndorms.ox.ac.uk Submitted 2019-03-08. Accepted 2019-06-13.

Background and purpose — Long-term rates of knee arthroplasty in patients with anterior cruciate ligament (ACL) injury who undergo ligament reconstruction (ACLr) are unclear. We determined this risk of arthroplasty through comparison with the general population. Patients and methods — All patients undergoing an ACLr in England, 1997–2017, were identified from national hospital statistics. Patients subsequently undergoing a knee arthroplasty were identified and survival analysis was performed (survival without undergoing knee arthroplasty). A Cox proportional hazards model was used to identify factors associated with knee arthroplasty. Relative risk of knee arthroplasty (total or partial) in comparison with the general population was determined. Results — 111,212 ACLr patients were eligible for analysis (mean age 29; 77% male). Overall, 0.46% (95% confidence interval [CI] 0.40–0.52) ACLr patients underwent knee arthroplasty within 5 years, 0.97% (CI 0.82–1.2) within 10 years, and 1.8% (CI 1.4–2.3) within 15 years. Knee arthroplasty risk was greater in older age groups and women. In comparison with the general population, the relative risk of undergoing arthroplasty at a younger age (at time of arthroplasty) was elevated: at 30–39 years (risk ratio [RR] 20; CI 11–35), 40–49 years (RR 7.5; CI 5.5–10), and 50–59 years (RR 2.5; CI 1.8– 3.5), but not 60–69 years (RR 1.7; CI 0.93–3.2). Interpretation — Patients sustaining an ACL injury who undergo ACLr are at elevated risk of subsequent knee arthroplasty in comparison with the general population. Although the absolute rate of arthroplasty is low, the risk of arthroplasty at a younger age is particularly elevated. When the outcome of shared decision-making is ACLr, this data will help inform patients and clinicians about the long-term risk of requiring knee arthroplasty.

Oxford, UK; 2 NIHR Biomedical Research Centre, Bristol; 5 Faculty of Medicine & Health Sciences,

Approximately 25–50% of patients with anterior cruciate ligament (ACL) injuries subsequently undergo ACL reconstruction surgery (Frobell et al. 2010, Collins et al. 2013, Nordenvall et al. 2014). This corresponds to a rate of ACL reconstruction thst has been reported at approximately 45/100,000 population per year in the United States and 24/100,000 in the United Kingdom (Buller et al. 2015, Abram et al. 2019). ACL reconstruction may either be performed early to stabilize the knee and prevent a further pivoting injury or may be delayed and performed only in patients with knee instability despite physiotherapy (Frobell et al. 2010). A key concern is that up to half of patients with a history of ACL injury develop signs of radiographic osteoarthritis within 10–15 years (Lohmander et al. 2007, Ajuied et al. 2014). ACL reconstruction and rehabilitation aims to stabilize the knee to reduce the risk of further injury with additional damage to the chondral surfaces and menisci (Kay et al. 2018, Mok et al. 2019). Many surgeons believe that ACL reconstruction is protective against osteoarthritis for the ACL-injured patient, notwithstanding the impact from the original injurious episode (Marx et al. 2003). Indeed, there is some reported evidence from a cohort of young, active individuals that ACL reconstruction reduces the risk of further meniscal and cartilage damage, in comparison with nonoperative management (Dunn et al. 2004). A previous meta-analysis (Ajuied et al. 2014) indicated that radiographic osteoarthritis may develop less frequently in ACL-injured knees managed with reconstruction in comparison with nonoperative treatment, but results are conflicting, with 1 large cohort finding no such association (Nordenvall et al. 2014). The use of radiological grading as an outcome or indicator of osteoarthritis is relatively subjective and may not reflect a patient’s symptoms (Parry et al. 2017).

Acta Orthopaedica 2019; 90 (6): 568â&#x20AC;&#x201C;574

Knee arthroplasty is a powerful and obvious surrogate marker for severe osteoarthritis, combining the severity of symptoms with radiological assessment (Bruyere et al. 2008, Raynauld et al. 2011, Carr et al. 2012). Population intervention rates for ACL reconstruction and knee arthroplasty have increased over the last 20 years in England but association between these interventions is unknown (Abram et al. 2019). Studies utilizing knee arthroplasty rates are scarce; however, case-control studies have suggested that ACL injury may be associated with up to a 7-times greater odds of knee arthroplasty (Leroux et al. 2014, Khan et al. 2018). Due to the limitations of these case-control studies, we determined the long-term risk (up to 15 years) of knee arthroplasty in patients with a history of ACL injury and undergoing surgical reconstruction, from an analysis of 20 years of longitudinal data from the complete National Health Service database for England, UK.â&#x20AC;&#x192;

Patients and methods Study design, setting, and data sources We performed a longitudinal cohort study utilizing the national healthcare records for England, UK. National Hospital Episode Statistics (HES) data were acquired for the purposes of this study for the period between April 1, 1997 and March 31, 2017 (NHS Digital; application reference: DARSNIC-68703). HES includes all National Health Service (NHS) care episodes, whether delivered in NHS hospitals or independent treatment centers, and also privately funded patients treated within NHS England hospitals. Surgical procedures, primary and secondary diagnoses, demographic and geographic data are recorded. Mortality data from the Office for National Statistics (ONS) mortality dataset (all in-hospital or community deaths) was used to adjust the number at risk when performing survival analysis (see below). Procedures and participants Patient HES records were identified from Classification of Surgical Operations and Procedures (OPCS-4) codes associated with inpatient care episodes for ACL reconstruction (see Supplementary data: Appendix 1). The first ACL reconstruction per patient was included, contralateral or revision procedures were not included, and cases undergoing multi-ligament reconstruction, ACL repair, or synthetic ligament surgery were excluded. Controls An arthroplasty control population was extracted for the year 2016â&#x20AC;&#x201C;17: all patients undergoing arthroplasty in this year without a history of ACL reconstruction in previous years (records of ACL injury specifically were not available). This control population was analyzed only in the secondary outcome analysis of relative risk of knee arthroplasty for the ACL

569

reconstruction population in comparison with this population. The population rate of arthroplasty for the control population was determined from the number of arthroplasties performed per age group population using mid-year population estimates available from the ONS. Outcomes The primary outcome was the rate of knee arthroplasty following ACL reconstruction. All HES and ONS records (including prior and subsequent hospital admission records) were then extracted for each patient with OPCS-4 codes for either ACL reconstruction or knee arthroplasty. Laterality coding was available to enable matching of procedures both by patient and by knee side (left vs. right). For patients undergoing ACL reconstruction, concurrent chondral or meniscal procedure codes were identified and included for comparison with isolated ACL reconstruction procedures. Per patient, only the first (primary) ACL reconstruction was included. Patients undergoing simultaneous knee arthroplasty and ACL reconstruction in the same hospital episode were excluded as these were not considered to be relevant to this study. Secondary outcomes were the relative odds of knee arthroplasty by a range of patient factors (defined below), and the relative risk of knee arthroplasty versus the control population defined above. Confounders The following potential confounding variables were analyzed: time (year of treatment), sex, age group, year of treatment, Charlson comorbidity index (derived with maximum 5-year diagnosis code lookback period), index of multiple deprivation (quintile derived from regional factors in England including average income, employment, education, housing, and crime; 1 = least deprived area, 5 = most deprived), rurality, ethnicity, and any concurrent chondral or meniscal surgery. These variables were selected a priori due to their potential importance in determining outcome, treatment choices or eligibility, or access to healthcare. Statistics In accordance with ONS and NHS Digital guidance, small numbers were suppressed where required. The rates of knee arthroplasty following ACL reconstruction at 5, 10, and 15 years were calculated as the absolute proportion and reported with the 95% confidence interval (CI) for the population sample size. A Cox proportional hazards model was used to calculate hazard ratios for subsequent arthroplasty over time, stratified by the confounders described in the previous section. Unadjusted and covariate-adjusted hazard ratios were calculated. Cases missing essential data (age, sex, procedure date, procedure laterality) were excluded from the study during data cleaning. Cases missing nonessential data (index of multiple deprivation, ethnicity, rurality) were included except for analyses adjusting for these specific variables.

570

Risk over time was also analyzed with Kaplan–Meier survival analysis to estimate and graphically report the long-term risk of undergoing knee arthroplasty (total or partial) using the full cohort data up to 20 years following ACL reconstruction: overall and stratified by patient age at the time of reconstruction and by sex. To determine the relative risk of knee arthroplasty after ACL reconstruction in comparison with the general population, the absolute rate of knee arthroplasty in 2016–17 was calculated for patients with and without a history of previous ACL reconstruction (records of ACL injury specifically were not available in this database). For the ACL population, the numerator was the number of matched, same side, knee arthroplasty procedures. For the general population, the numerator was all other knee arthroplasty patients without a history of ACL reconstruction. The denominator for the ACL population was all living patients in 2016–17 without a prior history of knee arthroplasty in the index knee. For the general population, denominator data were extracted from the ONS national population estimates. The relative annual risk (risk ratio) of knee arthroplasty for these respective populations within the most recent years of data (2016–17) was then calculated. To aid interpretation and for clinical relevance, both the absolute and relative risk estimates were stratified according to the age of the patient in 2016–17, irrespective of the year of previous ACL reconstruction, where applicable. Stata v15.1 (StataCorp, College Station, TX, USA) was used to perform all analyses. Confidence intervals are reported at the 95% level. Ethics, funding, and potential conflicts of interest The study was approved by the NHS Digital Independent Group Advising on the Release of Data (IGARD) committee (application reference: DARS-NIC-68703). No other ethical approval was required. This report is independent research supported by the National Institute for Health Research (NIHR), Oxford Biomedical Research Centre (BRC), and Bristol Biomedical Research Centre (BRC). The views expressed in this publication are those of the authors and not necessarily those of the NHS, the National Institute for Health Research or the Department of Health and Social Care. All authors have completed the Unified Competing Interest form (available on request from the corresponding author). AJ has received consultancy fees from Freshfields Bruckhaus Deringer (on behalf of Smith & Nephew Orthopaedics Limited), and is a member of the Data Safety and Monitoring Board (which involved receipt of fees) from Anthera Pharmaceuticals, Inc. All other authors declare no conflicts of interest.

Results Between April 1, 1997 and March 31, 2017, 124,489 patients underwent ACL surgery, of which 111,212 patients were

Acta Orthopaedica 2019; 90 (6): 568–574

Inpatient care episodes for ACLr between April 1, 1997 and March 31, 2017 n = 133,270 (124,489 patients) Excluded – data cleaning (n = 4,578): – missing side, 4,206 – data coding errors/simultaneous arthroplasty, 372 Excluded – non-index procedures (patient level): contralateral/revision ACLr n = 8,407 Excluded – concurrent procedures (n = 9,073): – ACL repair or synthetic, 1,495 – multi-ligament reconstruction, 7,578 Patient level cohort n = 111,212 patients/ACLr 5-year cohort (n = 46,402): knee arthroplasty, 212 patients 10-year cohort (n = 14,292): knee arthroplasty, 139 patients 15-year cohort (n = 3,726): knee arthroplasty, 66 patients

Figure 1. Flow chart illustrating extraction of patient level cohort.

included in the analysis, with a mean age of 29 (SD 10) (Figure 1) and mean follow up 5.9 years (range 0–20 years; SD 4.2) during which time 0.54% (600/111,212; CI 0.50–0.58) underwent subsequent arthroplasty. There was a greater proportion of male patients (77%) and the most common age group was 20–29 years (43%) (Table 1, see Supplementary data). Overall, 0.46% (212/46,402; CI 0.40–0.52) ACL reconstruction patients underwent knee arthroplasty within 5 years, 0.97% (139/14,292; CI 0.82–1.2) within 10 years, and 1.8% (66/3,726; CI 1.4–2.3) within 15 years (Table 1, see Supplementary data). The risk was greatly elevated in patients who were older at the time of their index ACL reconstruction (Table 2, Figure 2). In comparison with patients aged 20–29, patients aged 30–39 at the time of their ACL reconstruction were 6.2 times (CI 4.4–8.8), and patients 40–49 were 19 times (CI 13–26), more likely to undergo subsequent knee arthroplasty. The risk of subsequent knee arthroplasty was also elevated in female patients (Figure 3) and patients from the most deprived regions, but lower in patients of Asian ethnicity in comparison with White ethnicity (Table 2). In comparison with isolated ACL reconstruction, concurrent chondral surgery did not significantly alter the rate of subsequent arthroplasty (adjusted HR 1.5; CI 0.92–2.3), but the rate of subsequent arthroplasty was lower in patients with a record of concurrent meniscal surgery (adjusted HR 0.41; CI 0.30–0.56) (Table 2). The absolute risk of undergoing knee arthroplasty with a history of ACL reconstruction versus without a history of ACL reconstruction is summarized in Table 3. The absolute risks in the ACL reconstruction cohort were low, ranging from 0.04% (CI 0.02–0.06) per year between the ages of 30 and 39 years, to 0.72% (CI 0.34–1.3) per year between the ages of 60 and 69 years. Relative to the general population of patients without a history of ACL reconstruction, patients with a history of

Acta Orthopaedica 2019; 90 (6): 568–574

Table 2. Hazard ratios (subsequent TKA within maximum of 20 years) of ACL reconstruction Factor

Risk of subsequent TKA Unadjusted Adjusted HR (CI) HR (CI)

Sex Male 1.0 1.0 Female 2.4 (2.1–2.9) 1.5 (1.3–1.8) Age (years) < 20 – – 20–29 1.0 1.0 30–39 6.2 (4.4–8.8) 6.2 (4.4–8.8) 40–49 20 (14–28) 19 (13–26) 50–59 46 (32–66) 42 (29–61) ≥ 60 – – Year of ACL reconstruction per year 1.0 (1.0–1.1) 1.0 (0.99–1.0) Charlson comorbidity index per unit 1.1 (1.0–1.1) 1.0 (0.96–1.1) Index of multiple deprivation (quintile) 1 = least 1.0 1.0 2 1.1 (0.9–1.5) 1.2 (0.90–1.5) 3 1.1 (0.9–1.5) 1.2 (0.94–1.6) 4 1.2 (0.95–1.6) 1.6 (1.2–2.1) 5 = most 1.4 (1.1–1.8) 2.0 (1.5–2.6) Rurality Urban 1.0 1.0 Rural 1.2 (0.94–1.4) 1.1 (0.86–1.3) Ethnicity White 1.0 1.0 Asian 0.35 (0.18–0.67) 0.40 (0.21–0.77) Black – – Mixed – – Other – – Concurrent: Isolated ACLr 1.0 1.0 Chondral surgery 2.0 (1.3–3.2) 1.5 (0.92–2.3) Meniscal surgery a 0.39 (0.29–0.53) 0.41 (0.30–0.56) TKA = total or partial knee arthroplasty. HR = hazard ratio; CI = 95% confidence interval. a with or without concurrent chondral surgery.

571

Table 3. Rates and relative risk of undergoing TKA by age at TKA in 2016 (with versus without a history of ACL reconstruction) Age at TKA

Prior ACLr Annual rate TKA/105 (CI)

Without prior ACLr Annual rate TKA/105 (CI)

30–39, n % 40–49, n % 50–59, n % 60–69, n % Overall, n a %

37 (20–63) 0.04 (0.02–0.06) 186 (134–252) 0.19 (0.13–0.25) 384 (269–531) 0.38 (0.27–0.53) 717 (345–1,315) 0.72 (0.34–1.32) 148 (120–180) 0.15 (0.12–0.18)

1.9 (1.6–2.2) 0.00 (0.00–0.00) 25 (24–26) 0.02 (0.02–0.03) 151 (148–153) 0.15 (0.15–0.15) 414 (408–419) 0.41 (0.41–0.42) 133 (131–134) 0.13 (0.13–0.13)

Relative risk RR (CI) 20 (11–35) 7.5 (5.5–10) 2.5 (1.8–3.5) 1.7 (0.9–3.2) 1.1 (0.9–1.4)

TKA = total or partial knee arthroplasty; ACLr = anterior cruciate ligament reconstruction; CI = 95% confidence interval. a 30–69 years

ACL reconstruction were considerably more likely to subsequently undergo knee arthroplasty at a younger age, at 30–39 years (risk ratio [RR] 20; CI 11–35), 40–49 years (RR 7.5; CI 5.5–10), and 50–59 years (RR 2.5; CI 1.8–3.5), but not 60–69 years (RR 1.7; CI 0.9–3.2). Overall, for all patients (30–69 years), the relative risk was not significantly elevated (RR 1.1; CI 0.9–1.4) (Table 3).

Discussion Principal findings In this nationwide retrospective cohort study, we found that 1.8% of patients undergo knee arthroplasty within 15 years of ACL reconstruction. Annual rates of subsequent knee arthro-

K–M cumulative knee arthroplasty risk (%)

20 Age groups 20–29 30–39 40–49 50–59

Sex female male

9 10 11 12 13 14 15

Years after index operation

Figure 2. Kaplan–Meier cumulative risk of undergoing knee arthroplasty following ACL reconstruction by age group. Age group < 20 years and ≥ 60 suppressed due to small numbers; shaded areas represent 95% confidence intervals.

9 10 11 12 13 14 15

Years after index operation

Figure 3. Kaplan–Meier cumulative risk of undergoing knee arthroplasty following ACL reconstruction by sex. Age group < 20 years and ≥ 60 suppressed due to small numbers; shaded areas represent 95% confidence intervals.

572

plasty at a young age were elevated in comparison with the general population, suggesting an association with progressive osteoarthritis. The rate of arthroplasty in patients aged 30–39 years, when undergoing arthroplasty, was 21 times higher than anticipated for the general population but the annual rate of arthroplasty in this age group is still very low at 0.04% per year. Risk ratios were lower in increasing age groups and, overall, there was no significant difference in annual arthroplasty rates when all patients (30–69 years) were analyzed together. Comparison with other studies 2 case-control studies have reported rates of knee arthroplasty following ACL “injury” and ACL reconstruction, respectively, at 7 times the odds of the general population (Leroux et al. 2014, Khan et al. 2018). Both studies had a number of limitations related to study design and control group matching. The ACL “injury” study could not identify those patients who had undergone surgical interventions and neither study could identify knee “laterality,” precluding matching of the affected knee and potentially leading to overestimation of arthroplasty rates. Overall (all patients aged 30–69 years), patients with a history of ACL reconstruction did not have a significantly elevated annual risk of arthroplasty in comparison with expected rates for the general population. In comparison, however, the annual rate of arthroplasty was elevated in younger age groups to a maximum of 20 times greater risk of undergoing knee arthroplasty at a young age (30–39 years). These findings suggest that ACL injury is an independent risk factor for early osteoarthritis development and may accelerate progression to severe osteoarthritis in susceptible individuals. There remains uncertainty over whether ACL reconstruction decreases the risk of osteoarthritis following ACL injury (Ajuied et al. 2014, Nordenvall et al. 2014). ACL reconstruction, once performed, may reduce the risk of further pivoting injuries and associated chondral or meniscal damage (Dunn et al. 2004, Fithian et al. 2005, Lohmander et al. 2007). The timing of ACL reconstruction was unknown in our cohort and, being observational, this study cannot determine at a patient level whether the decision to undergo an ACL reconstruction was protective or not. Considering the comparison of our findings with the higher rate of knee arthroplasty in the ACL injury case-control study, the results cannot be interpreted as a comparison of nonoperative care versus operative care of a ruptured ACL (Khan et al. 2018). The case-control study could not distinguish patients who had undergone ACL reconstruction from those who had not, and there is a considerable risk of confounding by indication—for example, patients with pre-existing osteoarthritis would be unlikely to undergo ACL reconstruction but would still have been included (Frobell et al. 2013, Khan et al. 2018). Given the low absolute rate of arthroplasty, however, a very large sample size would be required to determine the relative risk of arthroplasty after nonoperative versus operative management of ACL injury in a randomized study. This would

Acta Orthopaedica 2019; 90 (6): 568–574

likely be cost-prohibitive and therefore our findings currently represent the best available evidence in this area. In revision ACL reconstruction, meniscal and chondral pathology has been shown to be associated with inferior patient-reported outcomes (Webster et al. 2018). However, our study found no association between concurrent chondral surgery and subsequent knee arthroplasty. We observed reduced rates of knee arthroplasty in patients who had undergone simultaneous meniscal surgery (partial meniscectomy or repair). These meniscal procedures were combined due to small numbers of repairs, particularly earlier in the study period, and some dual coding – meaning variation in outcome by the specific type of meniscal procedure could not be determined. The lower rate of arthroplasty in this group was surprising, but similar to the findings following ACL reconstruction with simultaneous meniscal debridement reported in a previous case-control study (Leroux et al. 2014). These findings do, however, conflict with relative risk assumptions based on meniscectomy case series where patients with meniscal injury are known to be at greater risk of radiographic arthritic progression (Suter et al. 2016). In contrast to a previous study, year of treatment did not influence the rate of arthroplasty, suggesting that changes over time in ACL reconstruction techniques, or greater use of first-line nonoperative management strategies, did not significantly alter the observed rate of arthroplasty at a population level (Leroux et al. 2014). This is an interesting observation as, over the same study time period, the rate of ACL reconstruction being performed in England has risen by 1,100% to current rates of 24/105 population per year (Abram et al. 2019). Therefore, despite a rapid rise in intervention rate, rates of subsequent knee arthroplasty have remained stable for this population. This may suggest that, rather than changing indications for ACL reconstruction over this time, the increased intervention rate could represent a correction from previous national under-provision of the procedure or, alternatively, rising rates of injury. Indeed, the rate of intervention in England remains lower than rates in other countries that are as high as 52/105 in Australia (Janssen et al. 2012). Female sex was found to be associated with a higher rate of arthroplasty and previously female patients have also been shown to be at higher risk of knee arthroplasty after general knee arthroscopy in data from the United States (Boyd and Gradisar 2016). In addition to age and sex, patient-reported ethnicity also influenced this outcome, with a higher rate of arthroplasty in patients reporting white ethnicity in comparison with Asian ethnicity. The reason for this observation is uncertain but would be in accordance with previous studies indicating differences in healthcare access and care-seeking behavior in association with socioeconomic, cultural, occupational, and psychological factors (Adamson et al. 2003, Judge et al. 2010). There has been one previous clinical trial investigating the role of nonoperative management for ACL injuries and there

Acta Orthopaedica 2019; 90 (6): 568–574

is another ongoing clinical management strategy trial in the United Kingdom (Frobell et al. 2010, 2013, Beard et al. 2016). Changing treatment practices have not, so far, altered the observed rate of knee arthroplasty following ACL reconstruction at a population level in England, but these studies may lead to the development of new treatment guidance. Recently, there has been increased scrutiny on the need for individualized patient consent in clinical practice (Chan et al. 2017). For patients that do require ACL reconstruction, our study presents important new evidence to inform patients and clinicians of the risk of later requiring knee arthroplasty following ACL reconstruction, clearly stratified by patient factors including age group and sex. Strengths and limitations This paper comprises the largest cohort of ACL reconstruction patients that has been reported and, using national longitudinal cohort data, we have been able to report precisely the rate of subsequent arthroplasty, stratified by patient-specific risk factors, and have determined the relative risk of arthroplasty in comparison with the general population. An observational cohort study such as this is, however, unable to determine whether ACL reconstruction exerted any protective effect against the development of end-stage osteoarthritis. It is unclear whether the key driver of the risk of osteoarthritis following ACL rupture is the original injury or damage to the knee from subsequent pivoting instability episodes (Dunn et al. 2004, Lohmander et al. 2007, Kay et al. 2018, Mok et al. 2019). Other unmeasured factors that may determine the risk of progression to osteoarthritis include genetic, biochemical, and biomechanical factors (Lohmander et al. 2007). In some circumstances, ACL reconstruction may be performed to stabilize a knee to facilitate a partial or total knee arthroplasty and these individuals are unlikely to be representative of the ACL “injury” population (Krishnan and Randle 2009, Weston-Simons et al. 2012). For this reason, we excluded patients undergoing simultaneous knee arthroplasty and ACL reconstruction. The intraarticular ligament reconstruction codes used to identify ACL reconstruction will also have captured posterior cruciate reconstruction procedures. These procedures could not be separately identified but are very rare in other series, and therefore these procedures are unlikely to have materially altered the findings of this study (Årøen et al. 2013). Despite data cleaning to exclude patient procedures missing a side or with date coding errors, it is inevitable that some other coding errors will have persisted. For HES data, although the specific coding accuracy of ACL reconstruction and arthroplasty procedures has not been determined, other fields from which the Charlson comorbidity index is derived and records of serious vascular diagnoses have been shown to correlate strongly with primary care records in England (Wright et al. 2012, Crooks et al. 2015). Radiographic or patient-reported outcome data were not available for analysis. This is an important consideration, but

573

knee arthroplasty does represent an objective marker of severe symptomatic osteoarthritis that is clinically relevant and of high importance for patients (Lohmander et al. 2007). Given the reluctance to perform arthroplasty, especially at a young age, it must be noted, however, that the rate of knee arthroplasty will be lower than the rate of radiographic osteoarthritis and also the overall healthcare burden of symptomatic knee pathology in this population. Conclusion Patients with a history of ACL injury and ACL reconstruction are at an increased risk of subsequent knee arthroplasty, especially at a younger age, in comparison with the general population; however, the absolute rate of arthroplasty is low. The relative risk of knee arthroplasty had these patients been managed nonoperatively is unknown and further work is required to refine treatment recommendations following ACL injury. When ACL reconstruction is undertaken, our work will help to inform patients and clinicians of the risk of the undesirable long-term outcome of knee arthroplasty. Supplementary data Table 1 and Appendix are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/ 17453674.2019.1639360

SA: concept, methodology, analysis, writing and editing paper. AJ: methodology, analysis, editing paper. TK: editing paper. DB: editing paper. AP: concept, methodology, editing paper. Acta thanks Ville Mattila and Olof Sköldenberg for help with peer review of this study.

Abram S G F, Price A J, Judge A, Beard D J. Anterior cruciate ligament (ACL) reconstruction and meniscal repair rates have both increased in the past 20 years in England: hospital statistics from 1997 to 2017. Br J Sports Med 2019; pii: bjsports-2018-100195. [Epub ahead of print] Adamson J, Ben-Shlomo Y, Chaturvedi N, Donovan J. Ethnicity, socio-economic position and gender: do they affect reported health-care seeking behaviour? Soc Sci Med 2003; 57(5): 895-904. Ajuied A, Wong F, Smith C, Norris M, Earnshaw P, Back D, Davies A. Anterior cruciate ligament injury and radiologic progression of knee osteoarthritis. Am J Sports Med 2014; 42(9): 2242-52. Årøen A, Sivertsen E A, Owesen C, Engebretsen L, Granan L P. An isolated rupture of the posterior cruciate ligament results in reduced preoperative knee function in comparison with an anterior cruciate ligament injury. Knee Surgery, Sport Traumatol Arthrosc 21(5): 1017-22. Beard D J, Cook J, Campbell H, Monk A, Wilson C, O’Leary S, Jackson W, Davies L, Carr A, Price A, Haddad F, Barker K, Lamb S. The ACL SNNAP Trial: ACL surgery necessity in non acute patients. NIHR HTA; 2016. Source: https://www.hra.nhs.uk/planning-and-improving-research/ application-summaries/research-summaries/the-acl-snnap-trial-acl-surgery-necessity-in-non-acute-patients/ Boyd J A, Gradisar I M. Total knee arthroplasty after knee arthroscopy in patients older than 50 years. Orthopedics 2016; 39(6): 1-4. Bruyere O, Pavelka K, Rovati LC, Gatterová J, Giacovelli G, Olejarová M, Deroisy R, Reginster J Y. Total joint replacement after glucosamine sul-

574

phate treatment in knee osteoarthritis: results of a mean 8-year observation of patients from two previous 3-year, randomised, placebo-controlled trials. Osteoarthr Cartilage 2008; 16(2): 254-60. Buller L T, Best M J, Baraga M G, Kaplan L D. Trends in anterior cruciate ligament reconstruction in the United States. Orthop J Sport Med 2015; 3(1): 232596711456366. Carr A J, Robertsson O, Graves S, Price A J, Arden N K, Judge A, Beard D J. Knee replacement. Lancet 2012; 379(9823): 1331-40. Chan S W, Tulloch E, Cooper E S, Smith A, Wojcik W, Norman J E. Montgomery and informed consent: where are we now? BMJ 2017; 2224(May): j2224. Collins J E, Katz J N, Donnell-Fink L A, Martin S D, Losina E. Cumulative incidence of ACL reconstruction after ACL injury in adults. Am J Sports Med 2013; 41(3): 544-9. Crooks C J, West J, Card T R. A comparison of the recording of comorbidity in primary and secondary care by using the Charlson Index to predict shortterm and long-term survival in a routine linked data cohort. BMJ Open 2015; 5(6): 1-9. Dunn W R, Lyman S, Lincoln A E, Amoroso P J, Wickiewicz T, Marx R G. The effect of anterior cruciate ligament reconstruction on the risk of knee reinjury. Am J Sports Med 2004; 32(8): 1906-14. Fithian D C, Paxton E W, Stone M Lou, Luetzow W F, Csintalan R P, Phelan D, Daniel D M. Prospective trial of a treatment algorithm for the management of the anterior cruciate ligament-injured knee. Am J Sports Med 2005; 33(3): 335-46. Frobell R B, Roos E M, Roos H P, Ranstam J, Lohmander L S. A randomized trial of treatment for acute anterior cruciate ligament tears. N Engl J Med 2010; 363(4): 331-42. Frobell R B, Roos H P, Roos E M, Roemer F W, Ranstam J, Lohmander L S. Treatment for acute anterior cruciate ligament tear: five year outcome of randomised trial. BMJ 2013; 346(jan24_1): f232. Janssen K W, Orchard J W, Driscoll T R, van Mechelen W. High incidence and costs for anterior cruciate ligament reconstructions performed in Australia from 2003–2004 to 2007–2008: time for an anterior cruciate ligament register by Scandinavian model? Scand J Med Sci Sport 2012; 22(4): 495-501. Judge A, Welton N J, Sandhu J, Ben-Shlomo Y. Equity in access to total joint replacement of the hip and knee in England: cross sectional study. BMJ 2010; 341: c4092 Kay J, Memon M, Shah A, Meng Y, Kristian Y, Devin S. Earlier anterior cruciate ligament reconstruction is associated with a decreased risk of medial meniscal and articular cartilage damage in children and adolescents: a systematic review and meta- analysis. Knee Surgery, Sport Traumatol Arthrosc 2018; 0(0): 0. Khan T, Alvand A, Prieto-Alhambra D, Culliford D J, Judge A, Jackson W F, Scammell B E, Arden N K, Price A J. ACL and meniscal injuries increase the risk of primary total knee replacement for osteoarthritis: a matched case-control study using the Clinical Practice Research Datalink (CPRD). Br J Sports Med 2018; pii: bjsports-2017-097762. [Epub ahead of print]

Acta Orthopaedica 2019; 90 (6): 568–574

Krishnan S R, Randle R. ACL reconstruction with unicondylar replacement in knee with functional instability and osteoarthritis. J Orthop Surg Res 2009; 4(1): 1-5. Leroux T, Ogilvie-Harris D, Dwyer T, Chahal J, Gandhi R, Mahomed N, Wasserstein D. The risk of knee arthroplasty following cruciate ligament reconstruction: a population-based matched cohort study. J Bone Joint Surg Am 2014; 96(1): 2-10. Lohmander L S, Englund P M, Dahl L L, Roos E M. The long-term consequence of anterior cruciate ligament and meniscus injuries: Osteoarthritis. Am J Sports Med 2007; 35(10): 1756-69. Marx R G, Jones E C, Angel M, Wickiewicz T L, Warren R F. Beliefs and attitudes of members of the American Academy of Orthopaedic Surgeons regarding the treatment of anterior cruciate ligament injury. Arthroscopy 2003; 19(7): 762-70. Mok Y R, Wong K L, Panjwani T, Chan C X, Toh S J, Krishna L. Anterior cruciate ligament reconstruction performed within 12 months of the index injury is associated with a lower rate of medial meniscus tears. Knee Surgery, Sport Traumatol Arthrosc 2019; 27(1): 117-23. Nordenvall R, Bahmanyar S, Adami J, Mattila V M, Felländer-Tsai L. Cruciate ligament reconstruction and risk of knee osteoarthritis: the association between cruciate ligament injury and post-traumatic osteoarthritis: a population based nationwide study in Sweden, 1987–2009. PLoS One 2014; 9(8): e104681. Parry E, Ogollah R, Peat G. Significant pain variability in persons with, or at high risk of, knee osteoarthritis: preliminary investigation based on secondary analysis of cohort data. BMC Musculoskelet Disord 2017; 18(1): 1-11. Raynauld J-P, Martel-Pelletier J, Haraoui B, Choquette D, Dorais M, Wildi L M, Abram F, Pelletier J-P. Risk factors predictive of joint replacement in a 2-year multicentre clinical trial in knee osteoarthritis using MRI: results from over 6 years of observation. Ann Rheum Dis 2011; 70(8): 1382-8. Suter L G, Smith S R, Katz J N, Englund M, Hunter D J, Frobell R, Losina E. Projecting lifetime risk of symptomatic knee osteoarthritis and total knee replacement in individuals sustaining a complete anterior cruciate ligament tear in early adulthood. Arthritis Care Res (Hoboken) 2016; 69(2): 201-8. Webster K E, Feller J A, Kimp A, Devitt B M. Medial meniscal and chondral pathology at the time of revision anterior cruciate ligament reconstruction results in inferior mid-term patient-reported outcomes. Knee Surgery, Sport Traumatol Arthrosc 2018; 26(4): 1-6. Weston-Simons J S, Pandit H, Jenkins C, Jackson W F M, Price A J, Gill H S, Dodd C A F, Murray D W. Outcome of combined unicompartmental knee replacement and combined or sequential anterior cruciate ligament reconstruction: a study of 52 cases with mean follow-up of five years. Bone Joint J 2012; 94-B(9): 1216-20. Wright F L, Green J, Canoy D, Cairns B J, Balkwill A, Beral V. Vascular disease in women: comparison of diagnoses in hospital episode statistics and general practice records in England. BMC Med Res Methodol 2012; 12(1): 161.

Acta Orthopaedica 2019; 90 (6): 575–581

575

Equal tibial component fixation of a mobile-bearing and fixed-bearing medial unicompartmental knee arthroplasty: a randomized controlled RSA study with 2-year follow-up Daan KOPPENS 1, Søren RYTTER 1, Stig MUNK 1, Jesper DALSGAARD 1, Ole G SØRENSEN 2, Torben B HANSEN 1, and Maiken STILLING 1 1 Department

of Orthopedic Surgery, University Clinic for Hand, Hip and Knee Surgery, Hospital Unit West Holstebro, Denmark; 2 Department of Orthopedic Surgery, Aarhus University Hospital, Aarhus, Denmark Correspondence: daankopp@rm.dk Submitted 2019-03-15. Accepted 2019-06-18.

Background and purpose — Differences in stress distribution in a mobile-bearing and fixed-bearing unicompartmental knee arthroplasty (UKA) design might lead to a difference in fixation of the tibial component. We compared tibial component migration of a mobile-bearing (MB) UKA and a fixed-bearing (FB) UKA using radiostereometric analysis. Patients and methods — In a randomized, patientblinded clinical trial 62 patients received either the MB Oxford UKA or the FB Sigma UKA. The patients were followed for 24 months with radiostereometric analysis. Clinical outcome was assessed with Oxford Knee Score (OKS), RAND-36 and leg extension power. Results — Migration of the tibial components was similar between groups throughout follow-up. At 12 months, MTPM of the tibial component was 0.44 mm (95% CI 0.34– 0.55) for the MB group and 0.40 mm (CI 0.31–0.50) for the FB group. Between 12 and 24 months, the tibial components migrated with a median MTPM increase of 0.03 mm (CI –0.02 to 0.08) in the MB group and 0.03 mm (CI –0.02 to 0.07) in the FB group. Continuous migration of the tibial component was found for 1 MB UKA and 2 FB UKAs. Both groups showed similar and clinically relevant improvement in clinical outcome. Interpretation — MB and FB tibial components had similar good fixation and clinical improvement until 2 years. Based on this study, a low 5- to 10-year revision rate can be expected for both implants.

Unicompartmental knee arthroplasty (UKA) has shown good clinical outcome and implant survival for patients with medial osteoarthritis (OA) (Cheng et al. 2013, Peersman et al. 2015). The mobile-bearing (MB) medial Oxford UKA (Zimmer Biomet, Bridgend, UK) is a well-documented UKA and offers good functional results (Pandit et al. 2011, 2015), and a low 10-year revision rate of 7% and 15-year revision rate of 11% (Mohammad et al. 2018). The fixed-bearing (FB) medial Sigma UKA (DePuy International, Leeds, UK) offers 5-year revision rates between 4.7% and 5.6% in national arthroplasty registries (AOANJRR 2018, NJR 2018). Longterm results of the Sigma UKA are unknown. In 30–40% of UKA revisions, the reason is aseptic loosening (AOANJRR 2018, SKAR 2018). A fully congruous bearing design of the MB UKA results in low contact stress. The stress of the femur on the tibia occurring during movement is transformed into an evenly distributed compressive stress at the tibial implant/bone interface. Possible disadvantages of an MB design are backside wear and dislocation of the bearing. The concave bearing design of the FB UKA results in higher contact stress, resulting in shear stress and unevenly distributed compressive stress at the bone/ implant interface during loaded knee motion (Goodfellow et al. 2015). These differences in design and stress loading on the tibial bone could affect tibial component fixation. Implant fixation can be evaluated as component migration by radiostereometric analysis (RSA), which is a predictor for late implant loosening (Ryd et al. 1995, Pijls et al. 2012b, 2018). Low early implant migration has been related to low 5and 10-year revision rates in national registries (Pijls 2012b, 2018).

576

Acta Orthopaedica 2019; 90 (6): 575–581

Assessed for eligibility n = 180 Excluded (n = 115): – did not meet inclusion criteria, 53 – declined to participate, 34 – other reasons, 28 Randomized n = 65

logistic reasons. If a patient was excluded during the inclusion period of the study, an extra patient was included to maintain the power of the study. The study is reported in accordance with the CONSORT guidelines as well as the guidelines for standardization of RSA (Valstar et al. 2005) and the ISO standard for RSA (ISO 2013).

Surgery and implants The MB phase 3 Oxford medial UKA consists of a 2-pegged femoral component with a spheriAllocated to Oxford UKA (n = 33): Allocated to SigmaUKA (n = 32): – received allocated intervention, 31 – received allocated intervention, 31 cal articulation, a fully congruous mobile bear– did not receive allocated intervention – did not receive allocated intervention ing, and a tibial component with a flat articu(missing anterior cruciate ligament), 2 (no tantalum beads inserted), 1 lation surface and a keel at the non-articulating surface. The FB Sigma medial UKA consists of Lost to follow-up (n = 0) Lost to follow-up (n = 0) a 2-pegged femoral component with a large posDiscontinued intervention (n = 0) Discontinued intervention (n = 1): – deep infection, 1 terior condyle radius, a concave fixed bearing, and a tibial component with a keel and a peg at Excluded from analysis due to < 3 Excluded from analysis due to < 3 the non-articulating surface. Both UKAs were visible tantalum beads: visible tantalum beads: implanted with Palacos bone cement (Heraeus Femur Tibia Femur Tibia Holding GmbH, Hanau, Germany). 2 ortho4 months 18 2 4 months 3 1 12 months 17 3 12 months 5 2 pedic surgeons (SM and JD) experienced with 24 months 17 4 24 months 5 2 FB and MB UKA performed the surgeries. The manufacturer’s instructions were followed, and Figure 1. Consort 2010 flow diagram. a minimally invasive approach was used. During surgery, 4 to 6 1-mm tantalum beads were Besides implant survival, patient satisfaction and knee inserted in the periprosthetic femoral and tibial bone in order function are important clinical outcomes after knee surgery. to accommodate future RSA analysis. All patients followed a Implants introduced to the market should offer at least the fast-track program (Koppens et al. 2018). same clinical outcome as established implants. We compared the MB Oxford UKA and the FB Sigma UKA Primary outcome with tibial component migration as the primary outcome and Radiostereometric analysis clinical outcome scores as a secondary outcome. We hypoth- A previously described standardized RSA set-up (Koppens et esized that there was no difference in migration or clinical out- al. 2018) with the patient supine was used to obtain stereoradiographs on the first postoperative day, and at 4, 12, and 24 come between implants. months. An auto-positioning, direct-digital roentgen system (AdoraRSA suite, NRT, Aarhus, Denmark) was used. 2 ceiling-fixed, synchronized roentgen tubes (Varian Medical SysPatients and methods tems, Palo Alto, CA, USA) were positioned 100 cm above the Between January 2014 and November 2015, a randomized, calibration box at an angle of 40° to each other. Digital image patient-blinded clinical trial was performed. Patients with detectors (Canon, CXDi-701C Wireless; Canon Europe, primary medial OA of the knee were assessed for eligibility Uxbridge, UK) were placed behind the calibration box. Digi(Figure 1). tal radiographs were stored in DICOM format at a resolution The inclusion criteria were patients above 18 years of of 160 µm pixel pitch and a 16-bit grey-scale resolution in a age, who were eligible for medial UKA (Murray et al. 1998, picture archiving and communication system (PACS). DePuy International 2009). The exclusion criteria were All analyses were performed with Model-Based RSA softinflammatory arthritis, contralateral knee prosthesis, dissemi- ware version 4.11 by use of computer-aided design (CAD) nated malignant disease, serious systemic disease, female models (RSAcore, Leiden, The Netherlands). The upper limit patients of reproductive age, and patients unable to give writ- of mean error rigid body fitting was 0.35 mm, and 120 for the condition number (Valstar et al. 2005, ISO 2013). If migration ten informed consent. Patients were randomized to receive the Oxford UKA (MB analysis was not possible due to occluded markers or primary group) or the Sigma UKA (FB group). Randomization was analysis showed a high condition number (> 80), a patientdone in blocks of 10 patients, generated via www.random.org/ specific marker configuration model (MC model) of the bone lists. Opaque envelopes were drawn 1 day before surgery for markers was constructed if possible and applied in the analy-

Acta Orthopaedica 2019; 90 (6): 575–581

sis (Kaptein et al. 2005). An MC model for the tibial bone was used to analyze 3 tibial components in the MB group and 4 tibial components in the FB group. An MC model for the femoral bone was used to analyze 6 femoral components in the MB group and 7 femoral components in the FB group. Patients with less than 3 visible markers were excluded. The postoperative stereoradiograph served as reference examination. The y-axis of the calibration box was parallel to the anatomical axis of the leg. Signed translations along and rotations around the x-, y-, and z-axis were defined as Tx (lateral/medial), Ty (distal/proximal), and Tz (posterior/anterior) and as Rx (flexion/extension), Ry (external/internal), and Rz (abduction/adduction) (Valstar et al. 2005). Total translation (TT) for the center of gravity of the implant was defined as: √(Tx2 × Ty2 × Tz2) For small rotations, total rotation (TR) can be defined as (Kaptein et al. 2007): √(Rx2 × Ry2 × Rz2) Maximal total point motion (MTPM) was defined as the translation vector of the point in the CAD model that had the greatest motion (Valstar et al. 2005). Continuous migration was defined as MTPM more than 0.2 mm between 12 and 24 months (Ryd et al. 1995). Precision of RSA The precision of the measurements was based on double examinations on all patients taken at 12 months’ follow-up. The postoperative stereoradiograph was used as a reference in the migration analysis. The bias was defined as the mean difference in translation along and the rotation about the three axes between the double examinations. The precision was defined as the standard deviation (SD) of the difference (SDdif). The expected clinical precision was represented as the prediction interval (PI) and defined as 1.96 x SDdif. Pooled data were comparable to precision data from the literature (Tables 1 and 2, see Supplementary data) (Stilling et al. 2011, Molt et al. 2012, Pijls et al. 2012a, Koppens et al. 2018). Secondary outcome PROMs The Oxford Knee Score (OKS) (Murray et al. 2007) and a general health questionnaire (RAND-36) (Hays and Morales 2001) were obtained before surgery, and at 4, 12, and 24 months after surgery. OKS is a 12-item questionnaire, with scores ranging from 0 (worst) to 48 (best) (Odgaard and Paulsen 2009). RAND-36 was scored using the RAND scoring rules, ranging from 0 (lowest) to 100 (highest) (Laucis et al. 2015). Summary scores for physical functioning, role limitations caused by physical health problems, pain, and general health perception were given. Leg-extension power Functional outcome was measured as the leg-extensor power (LEP) (Bassey and Short 1990) using the leg-extensor power

577

rig (Bio-Med International, Nottingham, UK) (Barker et al. 2012, Munk et al. 2012). Both legs were tested before surgery and at 24 months after surgery, and the operated leg was further tested at 1 and 12 months after surgery. Patients performed a minimum of 5 repetitions and a maximum of 10 repetitions. The session was stopped if the patient had reached his or her maximum, defined as 2 attempts with a lower score than the previous or if the patient reported pain in the knee (Munk et al. 2012). The maximum recorded measurement was used in the analysis. LEP is expressed as power per kg of body weight (W/kg). Statistics Sample size A generally accepted threshold for migration is the difference in MTPM between 12 and 24 months > 0.2 mm (Ryd et al. 1995). To detect a 0.2 mm difference in MTPM we needed 22 patients in each group (power 90%, alpha 0.05, SD 0.2 mm) (Kendrick et al. 2015). To anticipate dropouts, 30 patients were included in each group. RSA All RSA data were assessed using mixed-model analysis (MMA) (Ranstam 2012). Assumptions concerning the data distribution were ensured, using mixed-model residual QQplots, fitted vs. residuals plots and histograms. A likelihoodratio test was used to detect differences between models, a Wald test to detect differences within the model. If a difference within the model was found, pairwise comparisons were used to specify the differences. Translations and rotations were shown as mean and 95% confidence intervals (CI). MTPM, TT, and TR were not normally distributed and were therefore analyzed on a logarithmic scale (median and CI reported). To accommodate comparison in the literature, mean and CI was also presented (MMA without logarithmic transformation). PROMs OKS was analyzed using mixed-model analysis (Ranstam 2012). The minimal clinically important difference (MCID) was defined as 9 points within groups, and 5 points between groups (Beard et al. 2015). Leg-extension power LEP data of the operated leg were analyzed using mixed model analysis. LEP data of the operated leg and the contralateral leg preoperatively and at 24 months were analyzed using paired t-tests (Barker et al. 2012). All data gathered on excluded patients were included in the analysis up to the time of exclusion (Figure 1). Statistical significance was assumed at p < 0.05. Intercooled Stata version 13.1 (StataCorp, College Station, TX, USA) was used for statistical analysis.

578

Acta Orthopaedica 2019; 90 (6): 575–581

Table 3. Summary of baseline characteristics Factor

Table 8. Oxford Knee Score (mean (CI)), ranging from 0 (worst) to 48 (best)

MB UKA (n = 33) FB UKA (n =32)

Mean age (range) Male/female sex Mean weight, kg (SD) Mean height, cm (SD) Mean BMI (SD) Mean Oxford Knee Score (SD) RAND-36 (SD) physical functioning pain general health

64 (50–78) 16/17 87 (15) 171 (10) 29 (4) 26 (4.8)

61 (47–79) 17/15 89 (13) 173 (9) 30 (4) 28 (7.1)

50 (17) 65 (44) 74 (18)

53 (19) 72 (38) 75 (14)

Ethics, registration, funding, and potential conflicts of interest The study was approved by the Central Denmark Region Committee on Biomedical Research Ethics (journal no. 1-10-72-591-12; issue date 12-03-2013) and the Danish Data Protection Agency (journal no. 1-16-02-82-13; issue date 22-05-2013). The study was conducted in accordance with the Helsinki Declaration and registered with ClinicalTrials.gov (NCT03434600). CAD implant models were provided from the implant companies. DK received a public grant for VIP salary from the Health Research Fund of Central Denmark Region. The other authors had no conflict of interest.

Time Preoperative 4 months 12 months 24 months

MB UKA

FB UKA

26 (24–28) 38 (35–40) 42 (40–44) 40 (37–43)

28 (26–30) 37 (34–39) 41 (39–43) 41 (38–44)

some migration over time (Tables 4 and 5, see Supplementary data, Figure 2). Between 4 and 24 months, the tibial components showed lift-off of mean 0.05 mm (CI 0.02–0.08) in the MB group and mean 0.04 mm (CI 0.01–0.07) in the FB group. Between 4 and 12 months, the tibial components showed posterior rotation of mean –0.18° (CI –0.29 to –0.08) in the MB group and mean –0.21° (CI –0.31 to –0.11) in the FB group. Between 12 and 24 months, the tibial components migrated with a median MTPM increase of 0.03 mm (CI –0.02 to 0.08) in the MB group and 0.03 mm (CI –0.02 to 0.07) in the FB group. Continuous migration was found for 1 MB UKA and 2 FB UKAs.

Baseline patient characteristics are given in Table 3. 1 patient with a FB UKA was excluded 5 weeks after primary surgery due to deep infection.

Femoral component Translations and rotations of the femoral components were similar between groups throughout follow-up (Table 6, see Supplementary data). At 4 months, the FB group showed a median 0.46° (CI 0.20–0.63) higher TR than the MB group. Also, the FB group showed a median 0.20 mm (CI 0.04–0.30) higher MTPM than the MB group at 4 months. This difference in TR and MTPM for the femoral component remained throughout follow-up (Table 7, Figure 3, see Supplementary data).

Primary outcome—RSA Tibial component Migration of the tibial components was similar between groups throughout follow-up, although both groups showed

Secondary outcome OKS and RAND-36 OKS is shown in Table 8 and RAND-36 in Table 9 (see Supplementary data).

Results

Y-axis translation (mm)

X-axis rotation (°) 0.2

0.10

0.08

–0.2

0.06

–0.4

0.04

MTPM (mm) 1.0

0.8

0.6

0.2

–0.8

FB UKA

–0.02

MB UKA

–1.0

0.4

0.02

–0.6

Months after index operation

0 0

Months after index operation

Figure 2. (a) X-rotation, (b) Y-translation, and (c) maximal total point motion (MTPM) for the tibial component (median and 95% CI).

Acta Orthopaedica 2019; 90 (6): 575–581

LEP LEP was similar between groups throughout follow-up (Table 10, see Supplementary data). Preoperatively, the knee awaiting surgery had lower LEP than the contralateral knee. At 24 months, both limbs performed equally.

Discussion RSA The MB and FB group showed similar tibial component migration with migration primarily in the first 12 months after surgery, after which the tibial components stabilized. Early migration of tibial components can be expected in the first 12 months (Kendrick et al. 2015, Koppens et al. 2018), but stabilization between 12 and 24 months is important (Ryd et al. 1995). We thought a higher strain at the bone interface of the FB compared with the MB tibial components to be a potential risk of higher migration. However, this was not the case, and a difference in the design of the backside of the tibial components might explain this. The keel on the FB tibial component is wider than on the MB tibial component, and further there is an extra peg on the medial side of the FB tibial component, which provides extra stability. In both the MB and FB group, the tibial component showed lift-off from the tibial bone (translation on the y-axis) until 12 months, and thereafter stabilized. This can be explained by the posterior rotation (rotation around the x-axis) of tibial components in both groups seen between 4 and 12 months. However, the posterior rotation was less than 0.8° at 24 months, which has recently been suggested as an acceptable threshold. Tibial lift-off, subsidence, and especially posterior rotation were shown to be predictors for late loosening (Gudnason et al. 2017). Signed migration measures have the advantage that they differentiate in the direction of measured migration, whereas the MTPM gives an implant- and time-dependent summarized vector-based migration measure. Although interesting, these thresholds (Gudnason et al. 2017) were based on a historical cohort of 116 patients of which only 5 were failures, and 4 of these failures had the same implant design. Combined prospective RSA data from several centers with long-term follow-up of failures are needed to validate new thresholds to be used as predictors for loosening. Recently Pijls et al. (2018) re-defined migration thresholds based on MTPM measures, with acceptance thresholds of 0.50 mm MTPM at 6 months’ follow-up (early migration), 0.20 mm continuous MTPM between 6 and 12 months (stabilization phase I), and 0.20 mm continuous MTPM between 12 and 24 months (stabilization phase II). Both the MB and FB tibial components showed acceptable migration on group level during stabilization phases I and II. Continuous migration was shown only for 1 MB and 2 FB tibial components. These results are in line with the low registry-reported short- to midterm revision rates of both components, which is 6–8.4% for

579

MB UKA and 4.7–5.6% for FB UKA at 5 years’ follow-up (AOANJRR 2018, NJR 2018). Other RSA studies evaluating fixation of UKA have found tibial component migration comparable to our study. Kendrick et al. (2015) compared cemented and cementless Oxford UKA until 24 months’ follow-up and found equally low migration for both groups. For the cemented tibial component, some posterior rotation was seen between 3 and 6 months though not statistically significant. In a prospective cohort RSA study (Koppens et al. 2018), we have formerly reported low migration of the FB Sigma UKA until 2 years’ follow-up, and MTPM under the 1-year threshold defined by Pijls et al. (2012b). In this study, we also found some posterior rotation throughout the follow-up period, though below the 0.8° threshold (Gudnason et al. 2017). The femoral component of the FB group showed a slightly higher MTPM than the MB group at 4 months, after which both groups stabilized. No thresholds exist for migration of the femoral component; migration of the femoral component in the FB group was similar as previously reported (Koppens et al. 2018). Clinical outcome The MB and FB UKA had an equally good clinical outcome. Both groups experienced a statistically significant and clinically relevant (Beard et al. 2015) improvement in knee pain and function from poor preoperatively to good up to 12 months after surgery. This improvement was sustained up to 24 months after surgery. Comparable improvements have been shown in the literature for both MB UKA (Pandit et al. 2011, Kendrick et al. 2015) and FB UKA (Koppens et al. 2018). Overall, the general health improved equally in both groups after surgery. Clinically relevant improvements were shown after surgery for “physical function,” “role limitations due to physical health,” and “pain” (Keurentjes et al. 2014). An improvement in LEP over time was observed in both groups, and after 24 months there was no inter-limb difference (operated vs. non-operated leg) in LEP. Similar improvements in LEP have been seen in patients with UKA (Barker et al. 2012, Jorgensen et al. 2017). Limitations Some limitations should be noted. First, nearly 20% of eligible patients declined to participate in the study, which could have resulted in selection bias. However, this decline rate is not unusual for surgical trials (Thoma et al. 2010). Our results should therefore be generalizable to other similar clinics. Second, a number of the stereoradiographs were unsuitable for analysis due to occluded markers. This issue was partly solved by using an MC model (Kaptein et al. 2005). Third, non-weightbearing stereoradiographs were taken, which might have given an underestimation of tibial subsidence (Horsager et al. 2017). This is, though, not of influence on the comparison made in this study.

580

In summary, the MB and FB tibial components had similar good fixation and clinical improvement until 2 years, and therefore a low long-term (5–10 year) revision rate can be expected for both implants. Supplementary data Tables 1, 2, 4–7, 9, and 10 as well as Figure 3 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2019.1639965

DK, MS, SM, and TBH designed the study. SM and JD participated in data acquisition. DK, MS, and SR analyzed the data. DK, MS, SR, OGS, and TBH interpreted the data. DK wrote the initial manuscript. All authors critically revised the manuscript. The authors thank Anne Dorthe Riedel, Conni Abildgaard Lisbjerg, and Pia Toft Pedersen for their valued assistance in patient management and data collection. Acta thanks Bart L Kaptein and Ben J L Kendrick for help with peer review of this study.

AOANJRR. Australian Orthopaedic Association National Joint Replacement Registry: Annual Report 2018. Adelaide, Australia; 2018. Barker K L, Jenkins C, Pandit H, Murray D. Muscle power and function two years after unicompartmental knee replacement. The Knee 2012; 19(4): 360-4. doi: 10.1016/j.knee.2011.05.006. Bassey E J, Short A H. A new method for measuring power output in a single leg extension: feasibility, reliability and validity. Eur J Appl Physiol Occup Physiol 1990; 60(5): 385-90. Beard D J, Harris K, Dawson J, Doll H, Murray D W, Carr A J, Price A J. Meaningful changes for the Oxford hip and knee scores after joint replacement surgery. J Clin Epidemiol 2015; 68(1): 73-9. doi: 10.1016/j. jclinepi.2014.08.009. Cheng T, Chen D, Zhu C, Pan X, Mao X, Guo Y, Zhang X. Fixed- versus mobile-bearing unicondylar knee arthroplasty: are failure modes different? Knee Surg Sports Traumatol Arthrosc 2013; 21(11): 2433-41. doi: 10.1007/s00167-012-2208-y. DePuy International. Sigma high performance partial knee: unicondylar surgical technique. DePuy International; 2009. Goodfellow J, O’Connor J, Pandit H, Dodd C, Murray D. Unicompartmental arthroplasty with the Oxford knee. 2nd ed. Wolvercote, Oxford: Goodfellow Publishers, 2015. Gudnason A, Adalberth G, Nilsson K G, Hailer NP. Tibial component rotation around the transverse axis measured by radiostereometry predicts aseptic loosening better than maximal total point motion. Acta Orthop 2017; 88(3): 282-7. doi: 10.1080/17453674.2017.1297001. Hays R D, Morales L S. The RAND-36 measure of health-related quality of life. Ann Med 2001; 33(5): 350-7. Horsager K, Kaptein B L, RoMer L, Jorgensen P B, Stilling M. Dynamic RSA for the evaluation of inducible micromotion of Oxford UKA during step-up and step-down motion. Acta Orthop 2017:1-7. doi: 10.1080/17453674.2016.1274592. ISO. Implants for surgery: Roentgen stereophotogrammtric analysis for the assessment of migration of orthopaedic implants. Switzerland: ISO; 2013. Jorgensen P B, Bogh S B, Kierkegaard S, Sorensen H, Odgaard A, Soballe K, Mechlenburg I. The efficacy of early initiated, supervised, progressive resis-

Acta Orthopaedica 2019; 90 (6): 575–581

tance training compared to unsupervised, home-based exercise after unicompartmental knee arthroplasty: a single-blinded randomized controlled trial. Clin Rehabil 2017; 31(1): 61-70. doi: 10.1177/0269215516640035. Kaptein B L, Valstar E R, Stoel B C, Rozing P M, Reiber J H. A new type of model-based Roentgen stereophotogrammetric analysis for solving the occluded marker problem. J Biomech 2005; 38(11): 2330-4. doi: 10.1016/j. jbiomech.2004.09.018. Kaptein B L, Valstar E R, Stoel B C, Reiber H C, Nelissen R G. Clinical validation of model-based RSA for a total knee prosthesis. Clin Orthop Relat Res 2007; 464:205-9. Kendrick B J, Kaptein B L, Valstar E R, Gill H S, Jackson W F, Dodd C A, Price A J, Murray D W. Cemented versus cementless Oxford unicompartmental knee arthroplasty using radiostereometric analysis: a randomised controlled trial. Bone Joint J 2015; 97-B(2): 185-91. doi: 10.1302/0301620x.97b2.34331. Keurentjes J C, Fiocco M, Nelissen R G. Willingness to undergo surgery again validated clinically important differences in health-related quality of life after total hip replacement or total knee replacement surgery. J Clin Epidemiol 2014; 67(1): 114-20. doi: 10.1016/j. jclinepi.2013.04.010. Koppens D, Stilling M, Munk S, Dalsgaard J, Rytter S, Sorensen O G, Hansen T B. Low implant migration of the SIGMA((R)) medial unicompartmental knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2018; 26(6): 1776-85. doi: 10.1007/s00167-017-4782-5. Laucis N C, Hays R D, Bhattacharyya T. Scoring the SF-36 in orthopaedics: a brief guide. J Bone Joint Surg Am 2015; 97(19): 1628-34. doi: 10.2106/ jbjs.o.00030. Mohammad H R, Strickland L, Hamilton T W, Murray D W. Long-term outcomes of over 8,000 medial Oxford Phase 3 Unicompartmental Knees: a systematic review. Acta Orthop 2018; 89(1): 101-7. doi: 10.1080/17453674.2017.1367577. 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. doi: 10.1302/2046-3758.112.2000064. Munk S, Dalsgaard J, Bjerggaard K, Andersen I, Hansen T B, Kehlet H. Early recovery after fast-track Oxford unicompartmental knee arthroplasty. 35 patients with minimal invasive surgery. Acta Orthop 2012; 83(1): 41-5. doi: 10.3109/17453674.2012.657578. Murray D W, Goodfellow J W, O’Connor J J. The Oxford medial unicompartmental arthroplasty: a ten-year survival study. J Bone Joint Surg Br 1998; 80(6): 983-9. Murray D W, Fitzpatrick R, Rogers K, Pandit H, Beard D J, Carr A J, Dawson J. The use of the Oxford hip and knee scores. J Bone Joint Surg Br 2007; 89(8): 1010-14. doi: 10.1302/0301-620X.89B8.19424. NJR. National Joint Registry for England, Wales and Northern Ireland and the Isle of Man: 15th Annual Report 2018; 2018. Odgaard A, Paulsen A. Translation and cross-cultural adaptation of the Danish version of Oxford Knee Score (OKS). 2009. Pandit H, Jenkins C, Gill H S, Barker K, Dodd C A, Murray D W. Minimally invasive Oxford phase 3 unicompartmental knee replacement: results of 1000 cases. J Bone Joint Surg Br 2011; 93(2): 198-204. doi: 10.1302/0301620x.93b2.25767. Pandit H, Hamilton T W, Jenkins C, Mellon S J, Dodd C A, Murray D W. The clinical outcome of minimally invasive Phase 3 Oxford unicompartmental knee arthroplasty: a 15-year follow-up of 1000 UKAs. Bone Joint J 2015; 97-b(11): 1493-500. doi: 10.1302/0301-620x.97b11.35634. Peersman G, Stuyts B, Vandenlangenbergh T, Cartier P, Fennema P. Fixedversus mobile-bearing UKA: a systematic review and meta-analysis. Knee Surg Sports Traumatol Arthrosc 2015; 23(11): 3296-305. doi: 10.1007/ s00167-014-3131-1. Pijls B G, Valstar E R, Kaptein B L, Nelissen R G. Differences in longterm fixation between mobile-bearing and fixed-bearing knee prostheses at ten to 12 years’ follow-up: a single-blinded randomised controlled radiostereometric trial. J Bone Joint Surg Br 2012a; 94(10): 1366-71. doi: 10.1302/0301-620x.94b10.28858.

Acta Orthopaedica 2019; 90 (6): 575â&#x20AC;&#x201C;581

Pijls B G, Valstar E R, Nouta K A, Plevier J W, Fiocco M, Middeldorp S, Nelissen R G. 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. doi: 10.3109/17453674.2012.747052. Pijls B G, Plevier J W M, Nelissen R. RSA migration of total knee replacements. Acta Orthop 2018: 1-9. doi: 10.1080/17453674.2018.1443635. Ranstam J. Repeated measurements, bilateral observations and pseudoreplicates, why does it matter? Osteoarthritis Cartilage 2012; 20(6): 473-5. doi: http://dx.doi.org/10.1016/j.joca.2012.02.011. 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.

581

SKAR. Swedish Knee Arthroplasty Register: Annual Report 2018. Malmo, Sweden: SKAR; 2018. Stilling M, Madsen F, Odgaard A, Romer L, Andersen N T, Rahbek O, Soballe K. Superior fixation of pegged trabecular metal over screw-fixed pegged porous titanium fiber mesh: a randomized clinical RSA study on cementless tibial components. Acta Orthop 2011; 82(2): 177-86. doi: 10.3109/17453674.2011.566139. Thoma A, Farrokhyar F, McKnight L, Bhandari M. Practical tips for surgical research: how to optimize patient recruitment. Can J Surg 2010; 53(3): 205-10. Valstar E R, Gill R, Ryd L, Flivik G, Borlin N, Karrholm J. Guidelines for standardization of radiostereometry (RSA) of implants. Acta Orthop 2005; 76(4): 563-72. doi: 10.1080/17453670510041574.

582

Acta Orthopaedica 2019; 90 (6): 582–589

Cementing technique for primary knee arthroplasty: a scoping review Anders M REFSUM 1, Uy V NGUYEN 1, Jan-Erik GJERTSEN 1,2, Birgitte ESPEHAUG 3, Anne M FENSTAD 2, Regina K LEIN 4, Peter ELLISON 2, Paul J HØL1,2, and Ove FURNES 1,2 1 Department of Clinical Medicine, Faculty of Medicine, University of Bergen, Bergen; 2 Department of Orthopaedic Surgery, Haukeland University Hospital, Bergen; 3 Centre for Evidence-Based Practice, Faculty of Health and Social Sciences, Western Norway University of Applied Sciences, Bergen; 4 Medical Library, University of Bergen, Bergen, Norway Correspondence: Ove.Furnes@uib.no Submitted 2018-12-22. Accepted 2019-07-12.

Background and purpose — The optimal cementing technique for primary total knee arthroplasty (TKA) remains unclear. We therefore performed a scoping review based on available studies regarding cementation technique in primary TKA and unicondylar knee arthroplasty (UKA). Patients and methods — A search in 3 databases identified 1,554 studies. The inclusion criteria were literature that studied cementing technique in primary TKA or UKA. This included cement application methods, full or surface cementing, applying cement to the bone and/or prosthesis, stabilization of the implant during curing phase, bone irrigation technique, drilling holes in the bone, use of suction, and the timing of cementation. 57 studies met the inclusion criteria. Results — The evidence was unanimously in favor of pulsatile lavage irrigation, drying the bone, and drilling holes into the tibia during a TKA. All studies concerning suction recommended it during TKA cementation. 7 out of 11 studies favored the use of a cement gun and no studies showed that finger packing was statistically significantly better than using a cement gun. There is evidence that full cementation should be used if metal-backed tibial components are used. Applying the cement to both implant and bone seems to give better cement penetration. Interpretation — There are still many knowledge gaps regarding cementing technique in primary TKA. There seems to be sufficient evidence to recommend pulsatile lavage irrigation of the bone, drilling multiple holes, and drying the bone before cementing and implant insertion, and applying cement to both implant and on the bone.

Aseptic loosening is the most common cause of revision after total knee arthroplasty (TKA) worldwide (Khan et al. 2016). Implant loosening appears to be a multifactorial event, but without preceding micromotion of the implant, aseptic loosening seems unlikely to occur (Goodman et al. 1994, Scuderi and Clarke 2005). Aseptic loosening may occur at the implant– cement interface (Kutzner et al. 2018), or at the bone–cement interface (Mann et al. 1997, Dahabreh et al. 2015). Studies have shown that sufficient cement penetration and thickness is important to prevent implant micromotion (Miskovsky et al. 1992). Penetration of cement into the cancellous bone at 1.5 mm or less usually leads to higher radiolucency and lower tensile strength, which is associated with early implant micromotion (Walker et al. 1984, Mann et al. 1997, Waanders et al. 2010). The cementing technique is multifactorial and includes: preparation of the bone before cementation; where, when, and how to apply the cement; and the curing and stabilization phase after installation (Endres and Wike 2011, Cawley et al. 2013). A study by Lutz and Halliday (2002) indicated a wide variation in cementing technique among orthopedic surgeons. This highlights the need for a general consensus based on evidence on how to cement a TKA, especially the tibial component, which has a 4 times higher risk of loosening than the femoral component in total knee arthroplasty (Furnes et al. 2007, Dyrhovden et al. 2017). We therefore performed a scoping review on available studies regarding cementing technique in primary TKA and UKA. Our aim was to investigate knowledge on cementing technique in primary knee arthroplasty and to identify eventual gaps in the knowledge that need more research.

Acta Orthopaedica 2019; 90 (6): 582–589

Method For this study, we followed the recommendations from the Cochrane Collaboration (Higgins et al. 2011) and the Methodological Framework to approach a scoping review (Arksey and O’Malley 2005). Research question From the aim we created 8 research questions: 1. What is the recommended cement application method? 2. Surface cementation or full cementation? 3. Should cement be applied to either bone or prosthesis or both? 4. What is the recommended irrigation method? 5. Is drilling holes into the tibial bone recommended? 6. Is peroperative suction recommended? 7. At which cement phase should cement be applied? 8. How should the implant be stabilized during the curing phase? Eligibility criteria We included all literature from our search on cementing technique in primary TKA and UKA where the topic was consistent with the formulated research questions. All study designs were included except for case reports. Literature that studied the use of tourniquet, patellar component, and mixing method of the cement were excluded to sharpen the scope of the study. Information sources The information search through the electronic databases OVID MEDLINE, OVID Embase, and Web of Science was last updated September 27, 2018 by 1 author (RKL). Subject headings for the specific database and free text terms were used with no restrictions to language, time, or format. The complete search strategies are shown in Appendix 1, see Supplementary data. Keywords and free text terms were decided and validated by 3 of the authors (UN, AR, and OF). Study selection The references were deduplicated in Endnote, and in addition manually by 2 of the authors (UN, AR). Obviously irrelevant studies were identified and excluded through title and abstract screening. 2 reviewers (UN, AR) independently screened the remaining studies and checked the full text versions of potential relevant studies. Data collection The reviewers developed a data extraction sheet based on the Cochrane Consumer and Communication review group’s data extraction template (Ryan et al. 2015) and pilot tested it on 3 studies regarding use of drilling holes. The result was discussed with the third reviewer (OF) for optimization and to decide which variables needed to be extracted from the studies.

583

Records identified through database searching n = 1,550

Additional records identified through other sources n=4

Records after duplicates removed n = 913 Records excluded after screening title and abstract n = 808 Full-text articles assessed for eligibility n = 105 Full-text articles with exclusion criteria (n = 48): – component studies, 11 – bone and soft tissue, 3 – wrong joint, 2 – poster, 1 – no cementing technique, 8 – coating, 1 – from textbooks, 9 – case-report, 1 – tourniquet, 2 – questionnaires, 2 – duplicate with different main author, 1 – letter to the editor, 1 – language, 1 a – revision studies, 5 Studies included in the review n = 57

Figure 1. Flow diagram of study inclusion. a Pujol et al. (2008).

Data items The parameters the reviewers (UN, AR, OF) agreed upon initially formed the aim of the study: study method, study design, demographics, follow-up period, level of evidence based after Oxford Centre for Evidence-based Medicine—Levels of Evidence (Howick et al. 2016), application method, preparation of the bone, cement type, prosthesis design, and outcome. Together, the 2 reviewers determined the studies’ Level of Evidence. The references were rated from I to V based on their study method. Animal and laboratory studies were regarded as mechanism-based reasoning or bench research and, therefore, graded as Level V (Howick et al. 2016). Any disagreements were resolved by consensus or through the third reviewer. Funding and potential conflicts of interest No funding was received. The authors declare no conflicts of interest.

Results Study selection, quality, and study characteristics Of 1,554 studies 105 articles were retrieved in full text (Figure 1). 57 articles met the inclusion criteria: animal studies (n =

584

Acta Orthopaedica 2019; 90 (6): 582–589

Studies included in the review n = 57

Type of studies (n = 59): a – animal studies, 3 – laboratory studies, 33 – patient studies, 23

Level of evidence (n = 59): a – I, 0 – II, 4 – III, 11 – IV, 8 – V, 36

Cement application method, 10 Surface versus full cementation, 19 Cement application area, 7 Bone irrigation, 9 Drilling holes, 5 Suction, 6 Cement properties and timing of cementation, 5 Stabilization of the implants during curing phase, 5

Figure 2. Included articles, study design, study quality, and inclusion groups. Some studies analyzed more than one parameter and were therefore categorized into several groups. a Walker et al. 1984 and Kanekasu et al. 1997 consists of 2 studies.

3), laboratory studies (n = 33), and clinical trials (n = 23). 2 studies had methods that met the inclusion criteria in 2 categories (Walker et al. 1984, Kanekasu et al. 1997). Only 4 studies were randomized controlled trials (RCTs). Characteristics of the included studies are summarized in Figure 2. 4,120 knees were included in the studies: 3,418 of them in patients, 501 in cadaver bones, and 111 sawbone knees. 25 of the knees were made from an experimental model (Bert and McShane 1998, Bucher et al. 2015). One study constructed a computer model of a female knee (Chong et al. 2011) and one study used finite element analysis (Cawley et al. 2012). Chapter 1. Cement application method Studies 10 studies conducted between 2003 and 2017 were reviewed (Table 1, see Supplementary data). The studies consisted of 3 clinical, 4 sawbone, and 3 cadaver studies. Mostly, the aim was to compare different application methods; cement gun, spatula, and finger packing or syringe use to achieve optimal cement penetration. 6 studies favored the use of a cement gun (Labutti et al. 2003, Kopec et al. 2009, Lutz et al. 2009, Vanlommel et al. 2011, Bucher et al. 2015, Schlegel et al. 2015a). 4 of these studies favored use of cement gun over finger packing when comparing cement penetration, clinical function score after operation, mechanical pull-out force, and the occurrence of postoperative radiolucent lines (RLL) (Kopec et al. 2009, Lutz et al. 2009, Vanlommel et al. 2011, Schlegel et al. 2015a). 4 studies favored the use of finger packing (Perez Mananes et al. 2012, Schlegel et al. 2014, Silverman et al. 2014, Han and Lee 2017). 2 of these studies favored finger packing over use of a cement gun when comparing cement penetration on human cadaver tibia (Schlegel et al. 2014, Silverman et al.

2014). Schlegel et al.(2014) also studied lift-off force. No studies favored the use of spatula over any other methods in terms of cement penetration. 1 study favored use of syringe over finger packing comparing cement penetration and RLL (Lutz et al. 2009). Comments In the 2 studies where finger packing was recommended over the usage of a cement gun (Schlegel et al. 2014, Silverman et al. 2014), the finger packing method was accompanied by factors that might be considered favorable, such as pulsatile lavage preparation of the tibial bone and cementing in doughy phase. The studies favoring finger packing still showed acceptable results, but it seems that use of a cement gun has shown better results in terms of surrogate outcomes. None of the studies could show any reduction in loosening rate when using a cement gun. In terms of the optimal cement penetration, Vanlommel et al. (2011) suggest a penetration between 3 and 5mm. Walker et al. (1984) concluded in their study that cement penetration over 1.5 mm is sufficient but suggested that ideally the penetration should be between 3 and 4 mm. None of the studies showed an increased loosening or revision rate with lower cement penetration, but Miller et al. (2014) concluded that a cement mantle over 3 mm is advisable to counteract cement decay over time. A pragmatic view would be to aim for between 3 and 5 mm cement penetration. Chapter 2. Surface versus full cementation Studies 19 studies were reviewed (Table 2, see Supplementary data). These studies consisted of 11 clinical studies (3 of them RCTs), 2 sawbone studies, 5 cadaver studies, and 1 computer study. The aims of these studies were either to compare the impact of full or surface cementation or to assess the quality of one of these methods. 2 studies showed a statistically significant difference when comparing surface cementation (SC) against full cementation (FC) favoring FC in terms of lift-off and rotation when using a metal-backed tibial model (Hyldahl 2003) and lift-off when using mobile-bearing prosthesis (Luring et al. 2006). 8 clinical and laboratory studies reported no statistically significant difference when comparing the 2 techniques. 1 study showed lower lift-off force using SC if the cement mantle was less than 3 mm, but no difference if the mantle was above that depth (Bert and McShane 1998). Other studies showed that FC gave a higher tibial bone resorption (Chong et al. 2011) and more micromotion (Skwara et al. 2009, Cawley et al. 2012). 2 studies showed that FC gave a higher stability and less strain compared to SC, especially in mobile bearing TKAs (Luring et al. 2006, Cawley et al. 2012). Finally, 1 study showed an excellent 10-year clinical result for both SC and FC, but found a lower revision rate for mechanical reasons in SC (Galasso et al. 2013), whereas Schlegel et

Acta Orthopaedica 2019; 90 (6): 582–589

al. (2015b) found no such difference. Case-control studies showed that both techniques could be sufficient over time, but without randomization, large number of patients, and longer follow-up this information was hard to assess (Galasso et al. 2013). However, an RCT using radiostereometric analysis (RSA) by Hyldahl et al. (2003, 2005a, 2005b) compared FC with SC in metal-backed and all-polyethylene tibial components of the AGC knee. The studies found that migration was reduced when using FC in metal-backed tibial components, but the migration was the same for all-polyethylene tibial components. The result in metal-backed components could not be confirmed by Saari et al. (2009) using the Profix metalbacked knee replacement. Comments The question of full versus surface cementation seems to be the most controversial and more clinical studies are needed. 3 studies showed that FC was better than SC (Bert and McShane 1998, Hyldahl 2003, Luring et al. 2006); meanwhile most of the other studies had the 2 techniques as equal. The different prosthesis designs, such as coating, roughness of the prosthesis surface, metal type, metal or all-poly tibial components, use of mobile bearing, and keel type probably influenced the results in the comparison of full versus surface cementing (Hyldahl 2003, Hyldahl et al. 2005a, 2005b, Luring et al. 2006, Saari et al. 2009). More clinical studies comparing both techniques in a standardized study method with different implants would be advisable to make progress on this topic. Chapter 3. Cement application area Studies 7 studies were reviewed (Table 3, see Supplementary data). These studies consisted of 2 clinical studies, 2 sawbone studies, 2 cadaver studies, and 1 porcine study. The aims of these studies were to assess the cement–bone interface strength, cement penetration depth and cement–mantle thickness regarding application of the cement to the bone only, implant only, or both. 4 studies favored application onto both the bone and prosthesis over application onto only either bone or prosthesis alone, where cement penetration and the length of the cement mantle was compared (Stannage et al. 2003, Vaninbroukx et al. 2009, Vanlommel et al. 2011, Wetzels et al. 2018). 1 study found no statistically significant difference comparing application onto bone versus bone and prosthesis, when studying properties of the cement interface and mechanical load to failure using a UKA model (Grupp et al. 2013). Another study favored cement application onto the prosthesis only over cement application onto the bone only, comparing percentage of cement penetration at different levels in porcine tibial bone (Bauze et al. 2004). Regarding the femoral component, 2 studies reported that cementation onto both the bone and the prosthesis was superior to cement application only to the

585

bone or prosthesis. However, in only 1 of them was the result statistically significant (Vaninbroukx et al. 2009). Comments At this point, a technique applying the cement to both implant and bone seems to be more favorable as supported by Vaninbroukx et al. (2009), Vanlommel et al. (2011), Han and Lee (2017), and Wetzels et al. (2018). More studies analyzing only this parameter are needed. These studies should also include the timing of application of cement to the implants and bone. Chapter 4. Bone irrigation Studies 9 studies were reviewed (Table 4, see Supplementary data). These studies consisted of 2 clinical studies and 7 cadaver studies. The aim of these studies was to compare different methods of preparing the bone before cementation. These methods were mainly irrigation with syringe, brush, lavage, or no preparation. 8 studies favored pulsatile lavage over manual syringe. Cement penetration depth, bone–cement interface strength, and pull-out force were statistically significantly increased when the bone was pulsatile lavaged compared with brushed or syringe lavaged (Ritter et al. 1994, Maistrelli et al. 1995, Clarius et al. 2009, Schlegel et al. 2011, Jaeger et al. 2012, Helwig et al. 2013, Schlegel et al. 2014, Boontanapibul et al. 2016, Scheele et al. 2017). One study found no difference between pulsatile lavage and cleaning with a surgical brush comparing cement penetration and a mechanical compression test using a UKA model (Scheele et al. 2017). 1 study found cleaning with pressurized CO2 in addition to pulsatile lavage to be significantly better than pulsatile lavage alone (Boontanapibul et al. 2016). Comments All 9 studies on irrigation methods of the bone concluded that pulsatile lavage was superior to irrigation by syringe. To achieve sufficient cement penetration depth and to reduce the occurrence of RLLs, a clean bone by pulsatile lavage and drying afterwards is crucial for the initial stability of the components (Schlegel et al. 2011). All included studies showed an improvement in either cement penetration or reduction in RLL. None of the studies showed reduction in the primary outcome loosening or revision rate. Our review showed that TKA studies regarding bone irrigation were unanimously in favor of pulsatile lavage irrigation, which therefore should be performed routinely in TKAs. Chapter 5. Drilling holes Studies 5 studies were reviewed (Table 5, see Supplementary data). These studies consisted of 2 clinical studies, 2 cadaver studies,

586

Acta Orthopaedica 2019; 90 (6): 582–589

and 1 dog study. In 2 studies, drilling holes were compared with no drilling holes (Miskovsky et al. 1992, van de Groes et al. 2013). The diameter of the drilling holes ranged from 2.4 to 4.5 mm. The numbers of holes were stated in 3 studies and the depth was mentioned in 4 out of 5 studies. All studies favored drilling holes into the tibial bone as this increased cement penetration, reduced occurrence of RLL, and increased bone–cement interface strength. No clinical studies examined or showed reduced loosening rate. None of the included studies discarded the measure of drilling holes into the bone due to negative effects. Only 1 of the studies compared different diameter of drilling holes and concluded that 4.5 mm diameter holes were superior to 2.0 mm holes in a sclerotic medial tibial plateau (Ahn et al. 2015).

cement–bone interface, RLL, and cement penetration. 1 study recommended that the cement should be applied in a doughy phase, comparing cement penetration and the use of a cement gun and finger packing (Silverman et al. 2014). 1 study highlighted the importance of application time when creating a cement–cement interface comparing mechanical bond strength and scanning electron microscope analysis (Park et al. 2001). 1 study concluded that a cement mantle over 3 mm is advisable to counteract decay over time comparing cement depth and contact fraction in post mortem TKAs (Miller et al. 2014). Dahabreh and colleagues’ study (2015) highlighted the diversity between cement brands and the study by Walker et al. analyzed many aspects to find the ideal cement penetration.

Comments The optimal number of holes, depth, and size should be further investigated and their clinical effect on loosening rate should be verified.

Comments It is important to use the manufacturers’ advice on cement curing, since different cement types have different properties (Kühn 2000, Dahabreh et al. 2015). In summary, to generate a strong bone–cement and cement–cement interlock the application should take place at around 2–3 minutes in a doughy/ application phase and the cement mantle should be at least 3 mm to weigh against the decay in the interlock over time (Miller et al. 2014). Park et al. (2001) show that creating a cement–cement interface was only 8% weaker than bulk cementation when created after 1 minute, whereas when created after 6 minutes was 42% weaker with only 50% bonding according to SEM analysis. After our literature search, Billi et al. (2019) published a laboratory study that recommended cementation of both the keel and undersurface of the tibial component, studying Palacos and Simplex cement. They also found that timing of cementation was important with improved pull-out force needed to separete the implant from the cement when the cement was applied on the implant in a sticky face 2 minutes after the start of mixing the Palacos cement and 3 minutes for the Simplex cement. The study also revealed that cementation in a dry condition gave higher pull-out force.

Chapter 6. Suction Studies 6 studies were reviewed (Table 6, see Supplementary data). These studies consisted of 3 clinical studies, 1 sawbone study, and 2 cadaver studies. The aim of these studies was to assess the effect of applying negative pressure to the tibial bone on cement penetration. The study by Banwart et al. (2000) compared negative pressure intrusion (NPI) against standard third-generation positive pressure intrusion (PPI) with no difference in cement penetration. The NPI technique was described similarly as a suction technique via Wolf needle and PPI was described as a standard third-generation cementing technique with a cement gun. All studies recommended using NPI but only 3 studies showed statistically significantly higher cement penetration using suction compared with no use of suction (Norton and Ayres 2000, Stannage et al. 2003, Bucher et al. 2015). No studies of suction has shown reduced loosening. Comments The use of suction in the tibia probably cannot replace a cement gun, but it might be a viable addition to optimize cement penetration depth if a tourniquet is not used. In this study suction and NPI were regarded as the same technique. Chapter 7. Cement properties and timing of cementation Studies 5 studies were reviewed (Table 7, see Supplementary data). These studies consisted of 1 clinical study, 3 cadaver studies, and 1 study that involved both a cadaver and a radiographic study (Walker et al. 1984). The aim of these studies was mainly to compare different cement application timings or cement phases and what effect these methods had on the

Chapter 8. Stabilization of the implants during curing phase Studies 5 studies were reviewed (Table 8, see Supplementary data). These studies consisted of 3 clinical studies, where 1 was an RCT, 2 were cadaver studies, and 1 was a porcine study. 1 study consisted of both a clinical and a cadaver study (Kanekasu et al. 1997). The aim of these studies was mainly to study different ways of keeping the prosthesis in position during the curing phase. 3 of the studies recommended using an external pressurizer to stabilize the implants during curing phase when compared against a manual method in a 2-stage cementation technique to increase cement penetration and stiffness (Kanekasu et al. 1997, Bauze et al. 2004, Diaz-Borjon et al. 2004). However, only 1 of these studies reported a statistically significant difference when using an experimental clamp, in the form of

Acta Orthopaedica 2019; 90 (6): 582–589

uniform stiffness of the fixation (Bauze et al. 2004). 1 study reported that with a single-stage cementation technique of UKA, a flexion angle of the knee of more than 45 degrees led to a tilting of the tibial component comparing femoral force application and cement penetration pressure (Jaeger et al. 2012). Single-stage cementing technique was superior to 2-stage cementing technique in 1 study, reducing the total number of RLLs (Guha et al. 2008). Comments Most surgeons do a single-stage cementing technique and extend the knee fully to apply pressure during the cement curing as described by Guha et al. (2008). But more evidence is needed to support this and also in which position the leg should be held when stabilizing the implant during the curing phase.

Discussion One of the most important findings in this scoping review was the heterogeneity between the studies. Comparability was limited due to different methods, materials, components, and parameters studied. 34 of the 57 included studies were laboratory studies and animal studies. The overall level of evidence seems low considering the potential impact on outcome. The most obvious gap in the literature is the lack of randomized clinical trials. We found only 4 RCTs and a lack of studies with revision or loosening as primary outcome. More research and especially solid RCTs are needed before one can find best practice. Summary Based on our scoping review the following guidelines for the cementing technique can be recommended: 1. A cement gun can be recommended to achieve optimal cement penetration and reduce occurrence of RLLs. The optimal cement penetration is not clearly defined but studies indicate between 3 and 5 mm. Applying cement by finger packing is a satisfactory method, while applying cement with a spatula was not advisable. 2. Full cementation should be applied on both the stem/keel and undersurface of the tibial component if using metalbacked components. All-poly tibial components can be cemented with surface cementation. 3. Cement should be applied to both implant and bone. Applying cement on only the bone or prosthesis should be avoided. 4. Pulsatile lavage irrigation and drying of the bone should be performed routinely in TKA to increase cement penetration depth and bone–interface strength. 5. Drilling holes into the sclerotic bone surface of the tibia can be recommended. 6. Suction in the tibial bone shows promising results in terms of cement penetration, but the evidence is insufficient to recommend use of suction routinely in TKA.

587

7. The cement should be applied in the cement’s application phase to both the femoral and tibial bone. 8. A single-stage cementation procedure is the recommended technique with the knee extended, keeping it as immobilized as possible. There is uncertainty on the degree of extension needed. 9. There is evidence from in vitro studies that applying the cement to the implant early, 2 minutes after mixing, increases the implant cement bonding, but no clinical studies support this. Supplementary data The Appendix and Tables 1–8 are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674.2019.1657333

AR and UN contributed equally as main authors. Our study was planned and designed by OF. BE and RL contributed with the scoping review methodology and database research, JEG with discussions, clinical insight, and manuscript preparation, PE and PH with biomechanical expertise. All authors participated in interpretation of the data and approved the final draft.

Acta thanks Steffen Breusch and Kaj Knutson for help with peer review of this study.

Ahn J H, Jeong S H, Lee S H. The effect of multiple drilling on a sclerotic proximal tibia during total knee arthroplasty. Int Orthop 2015; 39(6): 10778. doi: http://dx.doi.org/10.1007/s00264-014-2551-3. Allen M J, Leone K A, Lamonte K, Townsend K L, Mann K A. Cemented total knee replacement in 24 dogs: surgical technique, clinical results, and complications. Vet Surg 2009; 38(5): 555-67. doi: http://dx.doi. org/10.1111/j.1532-950X.2009.00528.x. Arksey H, O’Malley L. Scoping studies: towards a methodological framework. Int J Social Research Methodology 2005; 8(1): 19-32. doi: 10.1080/ 1364557032000119616. Arora J, Ogden A C. Osteolysis in a surface-cemented, primary, modular Freeman-Samuelson total knee replacement. J Bone Joint Surg Br 2005; 87(11): 1502-6. Banwart J C, McQueen D A, Friis E A, Graber C D. Negative pressure intrusion cementing technique for total knee arthroplasty. J Arthroplasty 2000; 15(3): 360-7. Bauze A J, Costi J J, Stavrou P, Rankin W A, Hearn T C, Krishnan J, Slavotinek J P. Cement penetration and stiffness of the cement–bone composite in the proximal tibia in a porcine model. J Orthop Surg 2004; 12(2): 194-8. Bert J M, McShane M. Is it necessary to cement the tibial stem in cemented total knee arthroplasty? Clin Orthop Relat Res 1998; (356): 73-8. Billi F, Kavanaugh A, Schmalzried H, Schmalzried T P. Techniques for improving the initial strength of the tibial tray–cement interface bond. Bone Joint J 2019; 101-B(1_Supple_A): 53-8. doi: 10.1302/0301-620X.101B1. BJJ-2018-0500.R1. Boontanapibul K, Ruangsomboon P, Charoencholvanich K, Pornrattanamaneewong C. Effectiveness testing of combined innovative pressurized carbon dioxide lavage and pulsatile normal saline irrigation to enhance bone cement penetration in total knee replacement: a cadaveric study. J Med Assoc Thai 2016; 99(11): 1198-202. Bucher T A, Butler M, Lee C, Eyres K S, Mandalia V, Toms A D. TKR without tourniquet: a laboratory study investigating the quality of the tibial cement mantle when using metaphyseal suction and cement gun. J Arthrosc Jt Surg 2015; 2(2): 62-6. doi: http://dx.doi.org/10.1016/j.jajs.2015.06.002.

588

Cawley D T, Kelly N, Simpkin A, Shannon F J, McGarry J P. Full and surface tibial cementation in total knee arthroplasty: a biomechanical investigation of stress distribution and remodeling in the tibia. Clin Biomech 2012; 27(4): 390-7. doi: http://dx.doi.org/10.1016/j.clinbiomech.2011.10.011. Cawley D T, Kelly N, McGarry J P, Shannon F J. Cementing techniques for the tibial component in primary total knee replacement. Bone Joint J 2013; 95-B(3): 295-300. doi: http://dx.doi.org/10.1302/0301-620X.95B3.29586. Chong D Y, Hansen U N, van der Venne R, Verdonschot N, Amis A A. The influence of tibial component fixation techniques on resorption of supporting bone stock after total knee replacement. J Biomech 2011; 44(5): 94854. doi: http://dx.doi.org/10.1016/j.jbiomech.2010.11.026. Clarius M, Hauck C, Seeger J B, James A, Murray D W, Aldinger P R. Pulsed lavage reduces the incidence of radiolucent lines under the tibial tray of Oxford unicompartmental knee arthroplasty: pulsed lavage versus syringe lavage. Int Orthop 2009; 33(6): 1585-90. doi: http://dx.doi.org/10.1007/ s00264-009-0736-y. Dahabreh Z, Phillips H K, Stewart T, Stone M. The effect of application time of two types of bone cement on the cement–bone interface strength. Eur J Orthop Surg Traumatol 2015; 25(4): 775-81. doi: http://dx.doi.org/10.1007/ s00590-014-1522-0. Diaz-Borjon E, Yamakado K, Pinilla R, Worland R L. Cement penetration using a tibial punch cement pressurizer in total knee arthroplasty. Orthopedics 2004; 27(5): 500-3. Dinh N L, Chong A C, Walden J K, Adrian S C, Cusick R P. Intrusion characteristics of high viscosity bone cements for the tibial component of a total knee arthroplasty using negative pressure intrusion cementing technique. Iowa Orthop J 2016; 36:161-6. Dyrhovden G S, Lygre S H L, Badawy M, Gothesen O, Furnes O. Have the causes of revision for total and unicompartmental knee arthroplasties changed during the past two decades? Clin Orthop Relat Res 2017; 475(7): 1874-86. doi: 10.1007/s11999-017-5316-7. Endres S, Wike A. Is cementing technique the cause of early aseptic looseningof the tibial component in total knee arthroplasty? A report of 22 failed tibial components. Orthop Rev 2011; 3(1): 20. doi: 10.4081/ or 2011; 3:e5. Furnes O, Espehaug B, Lie S A, Vollset S E, Engesaeter L B, Havelin L I. Failure mechanisms after unicompartmental and tricompartmental primary knee replacement with cement. J Bone Joint Surg Am 2007; 89(3): 519-25. Galasso O, Jenny J Y, Saragaglia D, Miehlke R K. Full versus surface tibial baseplate cementation in total knee arthroplasty. Orthopedics 2013; 36(2): e151-8. doi: http://dx.doi.org/10.3928/01477447-20130122-16. Goodman S B, Song Y, Doshi A, Aspenberg P. Cessation of strain facilitates bone formation in the micromotion chamber implanted in the rabbit tibia. Biomaterials 1994; 15(11): 889-93. Grupp T M, Pietschmann M F, Holderied M, Scheele C, Schroder C, Jansson V, Muller P E. Primary stability of unicompartmental knee arthroplasty under dynamic compression-shear loading in human tibiae. Clin Biomech 2013; 28(9-10): 1006-13. doi: http://dx.doi.org/10.1016/j.clinbiomech.2013.10.003. Grupp T M, Holderied M, Pietschmann M F, Schroder C, Islas Padilla A P, Schilling C, Jansson V, Muller P E. Primary stability of unicompartmental knee arthroplasty under dynamic flexion movement in human femora. Clin Biomech 2017; 41: 39-47. Guha A R, Debnath U K, Graham N M. Radiolucent lines below the tibial component of a total knee replacement (TKR): a comparison between single- and two-stage cementation techniques. Int Orthop 2008; 32(4): 453-7. doi: http://dx.doi.org/10.1007/s00264-007-0345-6. Han H S, Lee M C. Cementing technique affects the rate of femoral component loosening after high flexion total knee arthroplasty. Knee 2017; 24(6): 1435-41. Helwig P, Konstantinidis L, Hirschmuller A, Miltenberger V, Kuminack K, Sudkamp N P, Hauschild O. Tibial cleaning method for cemented total knee arthroplasty: An experimental study. Indian J Orthop 2013; 47(1): 18-22. doi: http://dx.doi.org/10.4103/0019-5413.106887. Higgins J, P T, Green S. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. handbook.cochrane.org. The Cochrane Collaboration; 2011.

Acta Orthopaedica 2019; 90 (6): 582–589

Hofmann A A, Goldberg T D, Tanner A M, Cook T M. Surface cementation of stemmed tibial components in primary total knee arthroplasty: minimum 5-year follow-up. J Arthroplasty 2006; 21(3): 353-7. doi: 10.1016/j. arth.2005.06.012. Howick J, Chalmers I, Glasziou P, Greenhalgh T. The Oxford levels of evidence 2. Oxford: Oxford Centre for Evidence-Based Medicine; 2016. Huiskes R, Sloof T J. Thermal injury of cancellous bone, following pressurised penetration of acrylic bone cement. Transactions of the annual meeting of the Ortopaedic Research Society. Las Vegas 1981; p.134. Hyldahl H. Fixation of the cemented tibial component: a radiostereometric analysis. Department of Clinical Science, Intervention and Technology [Thesis]. Stockholm: Karolinska; 2003. p. 45. Hyldahl H, Regner L, Carlsson L, Karrholm J, Weidenhielm L. All-polyethylene vs. metal-backed tibial component in total knee arthroplasty: a randomized RSA study comparing early fixation of horizontally and completely cemented tibial components, Part 1: Horizontally cemented components: AP better fixated than MB. Acta Orthop 2005a; 76(6): 769-77. Hyldahl H, Regner L, Carlsson L, Karrholm J, Weidenhielm L. All-polyethylene vs. metal-backed tibial component in total knee arthroplasty: a randomized RSA study comparing early fixation of horizontally and completely cemented tibial components, Part 2: Completely cemented components: MB not superior to AP components. Acta Orthop 2005b; 76(6): 778-84. doi: http://dx.doi.org/10.1080/17453670510045363. Jaeger S, Helling A, Bitsch R G, Seeger J B, Schuld C, Clarius M. The influence of the femoral force application point on tibial cementing pressure in cemented UKA: an experimental study. Arch Orthop Trauma Surg 2012; 132(11): 1589-94. doi: http://dx.doi.org/10.1007/s00402-012-1582-8. Jaeger S, Seeger J B, Schuld C, Bitsch R G, Clarius M. Tibial cementing in UKA: a three-dimensional analysis of the bone cement implant interface and the effect of bone lavage. J Arthroplasty 2013; 28(9 Suppl): 191-4. doi: http://dx.doi.org/10.1016/j.arth.2013.05.014. Kanekasu K, Yamakado K, Hayashi H. The clamp fixation method in cemented total knee arthroplasty: dynamic experimental and radiographic studies of the tibial baseplate clamper. Bull Hosp Jt Dis 1997; 56(4): 218-21. Khan M, Osman K, Green G, Haddad FS. The epidemiology of failure in total knee arthroplasty: avoiding your next revision. Bone Joint J 2016; 98-B(1 Suppl. A): 105-12. doi: 10.1302/0301-620X.98B1.36293. Kopec M, Milbrandt J C, Duellman T, Mangan D, Allan D G. Effect of hand packing versus cement gun pressurization on cement mantle in total knee arthroplasty. Can J Surg 2009; 52(6): 490-4. Kühn K-D. Bone cements: up-to-date comparison of physical and chemical properties of commercial materials. Berlin/Heidelberg Springer-Verlag; 2000. Kutzner I, Hallan G, Hol P J, Furnes O, Gothesen O, Figved W, Ellison P. Early aseptic loosening of a mobile-bearing total knee replacement. Acta Orthop 2018; 89(1): 77-83. doi: 10.1080/17453674.2017.1398012. Labutti R S, Bayers-Thering M, Krackow K A. Enhancing femoral cement fixation in total knee arthroplasty. J Arthroplasty 2003; 18(8): 979-83. Luring C, Perlick L, Trepte C, Linhardt O, Perlick C, Plitz W, Grifka J. Micromotion in cemented rotating platform total knee arthroplasty: cemented tibial stem versus hybrid fixation. Arch Orthop Trauma Surg 2006; 126(1): 45-8. Lutz M J, Halliday B R. Survey of current cementing techniques in total knee replacement. ANZ J Surg 2002; 72(6): 437-9. Lutz M J, Pincus P F, Whitehouse S L, Halliday B R. The effect of cement gun and cement syringe use on the tibial cement mantle in total knee arthroplasty. J Arthroplasty 2009; 24(3): 461-7. doi: http://dx.doi.org/10.1016/j. arth.2007.10.028. Maistrelli G L, Antonelli L, Fornasier V, Mahomed N. Cement penetration with pulsed lavage versus syringe irrigation in total knee arthroplasty. Clin Orthop Rel Res 1995; (312): 261-5. Mann K A, Ayers D C, Werner F W, Nicoletta R J, Fortino M D. Tensile strength of the cement–bone interface depends on the amount of bone interdigitated with PMMA cement. J Biomech 1997; 30(4): 339-46. doi: http:// dx.doi.org/10.1016/S0021-9290%2896%2900164-9.

Acta Orthopaedica 2019; 90 (6): 582–589

Mann K A, Miller M A, Khorasani M, Townsend K L, Allen M J. The dog as a preclinical model to evaluate interface morphology and micro-motion in cemented total knee replacement. Vet 2012; 25(1): 1-10. doi: http://dx.doi. org/10.3415/VCOT-11-01-0014. Matthews J J, Ball L, Blake S M, Cox P J. Combined syringe cement pressurisation and intra-osseous suction: an effective technique in total knee arthroplasty. Acta Orthop Belg 2009; 75(5): 637-41. Miller M A, Goodheart J R, Izant T H, Rimnac C M, Cleary R J, Mann K A. Loss of cement–bone interlock in retrieved tibial components from total knee arthroplasties. Clin Orthop Rel Res 2014; 472(1): 304-13. doi: http:// dx.doi.org/10.1007/s11999-013-3248-4. Miskovsky C, Whiteside L A, White S E. The cemented unicondylar knee arthroplasty: an in vitro comparison of three cement techniques. Clin Orthop Rel Res 1992; (284): 215-20. Norton M R, Eyres K S. Irrigation and suction technique to ensure reliable cement penetration for total knee arthroplasty. J Arthroplasty 2000; 15(4): 468-74. Park S H, Silva M, Park J S, Ebramzadeh E, Schmalzried T P. Cement–cement interface strength: influence of time to apposition. J Biomed Mater Res A 2001; 58(6): 741-6. doi: http://dx.doi.org/10.1002/jbm.10023. Pelt C E, Erickson J, Christensen B A, Widmer B, Severson E P, Evans D, Peters C L. The use of a modular titanium baseplate with a press-fit keel implanted with a surface cementing technique for primary total knee arthroplasty. BioMed Res Int 2014; 2014: 972615. doi: http://dx.doi. org/10.1155/2014/972615. Perez Mananes R, Vaquero Martin J, Villanueva Martinez M. An experimental study of bone cement penetration in total knee arthroplasty depending on cementing technique used. [in Spanish]. Trauma (Spain) 2012; 23(1): 48-58. Peters C L, Craig M A, Mohr R A, Bachus K N. Tibial component fixation with cement: full-versus surface-cementation techniques. Clin Orthop Relat Res 2003; (409): 158-68. doi: 10.1097/01.blo.0000058638.94987.20. Pujol N, Verdot F X, Chambat P. [Quality of tibial cementing in total knee arthroplasty: one or two phase cementing of the tibial and femoral implants]. Rev Chir Orthop Reparatrice Appar Mot 2008; 94(3): 241-6. doi: http://dx.doi.org/10.1016/j.rco.2007.09.005. Ritter M A, Herbst S A, Keating E M, Faris P M. Radiolucency at the bone– cement interface in total knee replacement: the effects of bone-surface preparation and cement technique. J Bone Joint Surg Am 1994; 76(1): 60-5. Rossi R, Bruzzone M, Bonasia D E, Ferro A, Castoldi F. No early tibial tray loosening after surface cementing technique in mobile-bearing TKA. Knee Surg Sports Traumatol Arthrosc 2010; 18(10): 1360-5. doi: http://dx.doi. org/10.1007/s00167-010-1177-2. Ryan R, Hill S, Prictor M, McKenzie J. Data extraction template for included studies. Cochrane Consumers and Communication Review Group; 2015. Saari T, Li M G, Wood D, Nivbrant B. Comparison of cementing techniques of the tibial component in total knee replacement. Int Orthop 2009; 33(5): 1239-42. doi: http://dx.doi.org/10.1007/s00264-008-0632-x. Scheele C, Pietschmann M F, Schroder C, Grupp T, Holderied M, Jansson V, Muller P E. Effect of lavage and brush preparation on cement penetration and primary stability in tibial unicompartmental total knee arthroplasty: an experimental cadaver study. Knee 2017; 24(2): 402-8. Schlegel U J, Siewe J, Delank K S, Eysel P, Puschel K, Morlock M M, de Uhlenbrock A G. Pulsed lavage improves fixation strength of cemented

589

tibial components. Int Orthop 2011; 35(8): 1165-9. doi: http://dx.doi. org/10.1007/s00264-010-1137-y. Schlegel U J, Puschel K, Morlock M M, Nagel K. An in vitro comparison of tibial tray cementation using gun pressurization or pulsed lavage. Int Orthop 2014; 38(5): 967-71. doi: http://dx.doi.org/10.1007/s00264-0142303-4. Schlegel U J, Bishop N E, Puschel K, Morlock M M, Nagel K. Comparison of different cement application techniques for tibial component fixation in TKA. Int Orthop 2015a; 39(1): 47-54. doi: http://dx.doi.org/10.1007/ s00264-014-2468-x. Schlegel U J, Bruckner T, Schneider M, Parsch D, Geiger F, Breusch S J. Surface or full cementation of the tibial component in total knee arthroplasty: a matched-pair analysis of mid- to long-term results. Arch Orthop Trauma Surg 2015b; 135(5): 703-8. doi: http://dx.doi.org/10.1007/s00402015-2190-1. Scuderi G R, Clarke H. Optimizing cementing technique. In: Total knee arthroplasty. A guide to get better performance. (Eds. Bellemans J, Ries M D, Victor J) Berlin/Heidelberg: Springer; 2005. Sharkey P F, Hozack W J, Rothman R H, Shastri S, Jacoby S M. Insall Award paper: Why are total knee arthroplasties failing today? Clin Orthop Relat Res 2002; (404): 7-13. Silverman E J, Landy D C, Massel D H, Kaimrajh D N, Latta L L, Robinson R P. The effect of viscosity on cement penetration in total knee arthroplasty, an application of the squeeze film effect. J Arthroplasty 2014; 29(10): 2039-42. doi: http://dx.doi.org/10.1016/j.arth.2014.05.010. Skwara A, Figiel J, Knott T, Paletta J R, Fuchs-Winkelmann S, Tibesku C O. Primary stability of tibial components in TKA: in vitro comparison of two cementing techniques. Knee Surg Sports Traumatol Arthrosc 2009; 17(10): 1199-205. doi: http://dx.doi.org/10.1007/s00167-009-0849-2. Stannage K, Shakespeare D, Bulsara M. Suction technique to improve cement penetration under the tibial component in total knee arthroplasty. Knee 2003; 10(1): 67-73. van de Groes S A, de Waal Malefijt M C, Verdonschot N. Influence of preparation techniques to the strength of the bone–cement interface behind the flange in total knee arthroplasty. Knee 2013; 20(3): 186-90. doi: http:// dx.doi.org/10.1016/j.knee.2012.08.002. Vaninbroukx M, Labey L, Innocenti B, Bellemans J. Cementing the femoral component in total knee arthroplasty: which technique is the best? Knee 2009; 16(4): 265-8. doi: http://dx.doi.org/10.1016/j.knee.2008.11.015. Vanlommel J, Luyckx J P, Labey L, Innocenti B, De Corte R, Bellemans J. Cementing the tibial component in total knee arthroplasty: which technique is the best? J Arthroplasty 2011; 26(3): 492-6. doi: http://dx.doi. org/10.1016/j.arth.2010.01.107. Walker P S, Soudry M, Ewald F C, McVickar H. Control of cement penetration in total knee arthroplasty. Clin Orthop Relat Res 1984; (185): 155-64. Wetzels T, van Erp J, Brouwer R W, Bulstra S K, van Raay J. Comparing cementing techniques in total knee arthroplasty: an in vitro study. J Knee Surg 2018; doi: 10.1055/s-0038-1669917 [Epub ahead of print] Waanders D, Janssen D, Mann K A, Verdonschot N. The mechanical effects of different levels of cement penetration at the cement–bone interface. J Biomech 2010; 43(6): 1167-75. doi: https://doi.org/10.1016/j.jbiomech.2009.11.033.

590

Acta Orthopaedica 2019; 90 (6): 590–595

All-polyethylene versus metal-backed posterior stabilized total knee arthroplasty: similar 2-year results of a randomized radiostereometric analysis study Shaho HASAN 1, Perla J MARANG-VAN DE MHEEN 2, Bart L KAPTEIN 1, Rob G H H NELISSEN 1, and Sören TOKSVIG-LARSEN 3 1 Department

of Orthopaedics, Leiden University Medical Center, Leiden, The Netherlands; 2 Department of Biomedical Data Sciences, Leiden University Medical Center, Leiden, The Netherlands; 3 Department of Orthopaedics, Hässleholm Hospital, Hässleholm, Sweden and Department of Clinical Sciences, Lund University, Lund, Sweden Correspondence: S.Hasan@lumc.nl Submitted 2019-05-06. Accepted 2019-07-23.

Background and purpose — The all-polyethylene tibial (APT) component, introduced in the early 1970s, was surpassed by metal-backed tibial (MBT) trays as the first choice for total knee arthroplasty (TKA). With improved polyethylene, the modern APT components can reduce costs, and have shown equivalent results in survivorship and early migration of the cruciate-retaining and cruciate-stabilizing designs. This study compares the 2-year migration of a similarly designed APT-posterior stabilized (PS) and a MBT-PS TKA, using radiostereometric analysis (RSA). Patients and methods — 60 patients were randomized to receive either an APT Triathlon PS or an MBT Triathlon PS TKA (Stryker, NJ, USA). Migration measured by RSA and clinical scores were evaluated at baseline and at 3, 12, and 24 months postoperatively. Repeated measurements were analyzed with a linear mixed model and generalized estimating equations. Results — The mean maximum total point movement (MTPM) at 3, 12, and 24 months was 0.41 mm (95% CI 0.33–0.50), 0.57 mm (0.44–0.70), and 0.56 mm (0.42–0.69) respectively in the MBT group and 0.46 mm (0.36–0.57), 0.61 mm (0.49–0.73), and 0.64 mm (0.50–0.77) in the APT group. 2 MBT and 1 APT implant were considered unstable at the 2-year follow-up. The KSS Knee score and KSS Function across 3, 12, and 24 months were comparable in both groups. Interpretation — For an APT-PS designed component, MTPM measured with RSA is comparable to the MBT-PS component after 2 years of follow-up. No differences in complications or clinical outcomes were found.

Despite the many advantages of the all-polyethylene tibial (APT) component, such as avoiding backside wear, preserving tibial bone, and lower costs, it accounts for only 0.1–13% of the total knee arthroplasties (TKA) registered (Norwegian Arthroplasty Register 2018, Swedish Knee Arthroplasty Register 2018). When TKA was introduced in the early 1970s, implants included APT components, but this design was soon replaced by a metal-backed tibial (MBT) component due to disappointing survival rates of the APT (Steinberg and Steinberg 2000, Browne et al. 2011). However, the APT is now regaining interest due to the higher costs of the MBT (Gioe and Maheshwari 2010, Chambers et al. 2016). Furthermore, APT has comparable results to MBT (Gioe and Maheshwari 2010). The advantage of the APT is that it preserves tibial bone as less resection is needed for the same polyethylene thickness, and that it avoids backside wear (Gioe and Maheshwari 2010, Browne et al. 2011, Cheng et al. 2012, Gustke and Gelbke 2017). Several studies have compared the outcomes of more recent APT designs with MBT in terms of survival, revision, and complications. Although reporting differing results, most studies found comparable survival rates of the APT and MBT (Browne et al. 2011, Cheng et al. 2011, Nouta et al. 2012b, Voss et al. 2016, Longo et al. 2017). Radiostereometric analysis (RSA) objectively measures migration of a prosthesis and can predict revision for aseptic loosening after 2 years (Ryd et al. 1995, Pijls et al. 2012). Few RSA studies comparing the APT and MBT have been conducted, showing less migration for the APT design in 1 study (Van Hamersveld et al. 2018), whilst others found no difference (Adalberth et al.

Acta Orthopaedica 2019; 90 (6): 590–595

2000, 2001, Norgren et al. 2004, Hyldahl et al. 2005), but these studies only included cruciate-retaining (CR) or condylar-stabilizing (CS) TKA and not posterior-stabilizing (PS) TKAs. The use of PS designed TKAs varies and is particularly popular in the United States and the Netherlands where it comprises 49% and 56% of all TKAs used, respectively (American Academy of Orthopedic Surgeons 2018, Dutch Arthroplasty Register 2018). The cam-post design of a PS insert could cause additional stress on the tibial component compared with a CR design (Garling et al. 2005, Molt and Toksvig-Larsen 2014). So apart from mixed results in studies with CR and CS designs, outcomes of these studies cannot be extrapolated to PS implants because of this cam-post design. A study comparing PS designed APT and MBT components is therefore needed. Hence we compared the migration of an APT- versus a MBT-PS designed prosthesis with up to 2-year follow-up using RSA.

Patients and methods This study was a randomized RSA trial comparing the APT-PS Triathlon Total Knee System with the MBT-PS Triathlon (Stryker, Warsaw, USA). Between November 2014 and June 2015, 60 consecutive patients were included and randomized to either an APT-PS or a MBT-PS component at the Hässleholm Hospital (Sweden). A blocked, computergenerated randomization scheme with a 1:1 ratio was used for randomization with a block size of 20. Patients were blinded to the treatment allocation and remained blinded throughout the study. Surgery was performed by 2 orthopedic surgeons who opened sealed opaque envelopes on the day of surgery. Clinical scores were assessed by blinded physical therapists. Inclusion criteria were patients with a painful knee resulting from osteoarthritis who were scheduled to undergo primary total knee surgery and were willing to sign an informed patient consent form. Main exclusion criteria were BMI > 40, a flexion or varus/valgus contracture > 15°, preoperative knee score > 70, and patients who could not make the follow-up visits because of living far away from the hospital. Prosthesis and surgical procedure The Triathlon APT is made from conventional polyethylene, sterilized with gamma radiation in vacuum and is packaged in nitrogen gas (N2Vac). The modular MBT component uses a highly cross-linked polyethylene insert (X3, Stryker Orthopaedics, Mahwah, NJ, USA). Patients were operated in concordance with the surgical protocol using a midline incision and a medial parapatellar approach. No tourniquet was used. Smartset GHV bone cement (DePuy CMW, Blackpool, UK) was applied only to the tibial baseplate. Perioperatively, 8 well-scattered tantalum beads (ø 0.8 mm; RSA Biomedical, Umeå, Sweden) were inserted into the tibial bone as reference

591

markers. 5 beads were inserted into the polyethylene insert of the MBT and in a similar position in the polyethylene of the APT. Patellae were reshaped. The postoperative regime included immediate full weight-bearing and there were no differences in postoperative treatment between the 2 groups. Outcome measures Primary outcome measure was prosthetic migration after 2 years measured by RSA defined as the maximum total point movement (MTPM), which is the length of the translational vector of the marker with the greatest migration in translation or rotation along the transverse, longitudinal, or sagittal axis. In concordance with the ISO 16087 Standard (ISO16087:2013(E), 2013), migration of a left-sided patient will be transformed to match the data of a right-sided patient to enable comparison between patients. Translations and rotations are expressed according to the right-hand screw rule. RSA radiographs were taken with the patient in supine position and the knee in a calibration cage using a biplanar technique at a 90-degree angle (Cage 10, RSA Biomedical, Umeå, Sweden). Radiographs were taken within 1–2 days postoperatively and at 3, 12, and 24 months. The first postoperative examination was taken as reference for subsequent examinations. At 12 months, double measurements were made to determine the precision of the examination. As no migration is expected between these 2 examinations performed at the same point in time, any migration measured will be the measurement error. The precision is expressed as the standard deviation of these measurements. Marker-based analysis using the software model-based RSA version 4.11 (RSAcore, Leiden, the Netherlands) was used. A mean error of rigid body fitting below 0.35 mm and a condition number below 120 were set as cut-off points. A marker configuration model was used if not enough markers were visible at any follow-up moment (Kaptein et al. 2005). Individual prostheses were considered stable if the increase in MTPM between 1 and 2 years postoperatively was ≤ 0.2 mm, and consequently any prosthesis with an MTPM increase of > 0.2 mm was considered as at risk for loosening (Ryd et al. 1995). Secondary outcome measures were the Knee Society Score (KSS), the Knee Osteoarthritis Outcome Score (KOOS) and the Forgotten Joint Score (FJS). The KSS and KOOS were measured preoperatively and at 3, 12, and 24 months. The FJS was measured at 3, 12, and 24 months. All scores ranged from 0 to 100 with higher scores indicating better scores. Sample size Sample size was calculated assuming that a difference of 0.3 mm for translation and 0.25° for rotation would be clinically relevant. 17 patients were needed in each group with an alpha of 0.05 and a power of 0.80. Taking into account that patients with inappropriate marking of the prosthesis or tibial bone will be excluded as well as possible patients lost to follow-up, 30 patients were included in each group.

592

ENROLLMENT

Acta Orthopaedica 2019; 90 (6): 590–595

Assessed for eligibility and randomized n = 60

ALLOCATION

Allocated to metal-backed tibia (n = 30) Received allocated interventiion (n = 27)

Allocated to all-polyethylene tibia (n = 30) Received allocated interventiion (n = 30)

FOLLOW-UP

Baseline, n = 29 – insufficient tibial markers, 1

Baseline, n = 27 – death by myocardial infarction, 1 – mismatching images, 1 – patient withdrawal, 1

3 months, n = 29 – insufficient tibial markers, 1

3 months, n = 27 – same as Baseline, 3

12 months, n = 28 – insufficient tibial markers, 1 – death by gastric tumor, 1

12 months, n = 27 – same as Baseline, 3

24 months, n = 29 – insufficient tibial markers, 1 – dead, 1 – patient withdrawal, 2

24 months, n = 23 – same as Baseline, 3 – patient withdrawal, 5 ANALYSIS

Baseline, n = 29 – insufficient tibial markers, 1 3 months, n = 26 – insufficient tibial markers, 1 – too few RSA-cage markers visible, 1 – too few markers visible, 1 – no radiographs, 1 12 months, n = 27 – insufficient tibial markers, 1 – dead, 1 – error of rigid body fitting >0.35, 1 24 months, n = 25 – insufficient tibial markers, 1 – dead, 1 – error of rigid body fitting >0.35, 1 – patient withdrawal, 2

Baseline, n = 27 – death by myocardial infarction, 1 – mismatching images, 1 – patient withdrawal, 1 3 months, n = 26 – same as Baseline, 3 – condition numer of prosthesis >120, 1 12 months, n = 27 – same as Baseline, 3 24 months, n = 22 – dead, 1 – mismatching images, 1 – error of rigid body fitting >0.35, 1 – patient withdrawal, 5

Figure 1. CONSORT flow chart.

Statistics Analyses were performed according to the intention-to-treat principle. MTPM, translations, rotations, and clinical outcome scores were analyzed with a linear mixed model if normally distributed. This model is recommended to analyze repeated measurements as it takes the within-subject correlation as well as the missing values into account (Ranstam et al. 2012). The model consisted of a group variable (APT versus MBT), a time variable (baseline, 3 months, 12 months, and 24 months), and an interaction term (fixed effects). An Autoregressive Order-1 covariance matrix was used to model remaining variability. The generalised estimating equations (GEE) approach was used if a normal distribution could not be obtained through transformation. This approach was needed for the analysis of MTPM, the KSS Knee score and the KOOS Sports subscore. Mean translations and rotations are reported per group at 3, 12, and 24 months. Mean scores of the KSS Knee, KSS Function, and the 5 subscales of the KOOS are reported per group preoperatively, and at 3, 12, and 24 months postoperatively. The mean FJS is reported at 3, 12, and 24 months postoperatively. P-values < 0.05 were considered statistically significant. Means are reported with 95% confidence intervals (CI). Analyses were performed with SPSS version 23 (IBM SPSS Statistics 23.0; IBM Corp, Armonk, NY, USA).

Table 1. Baseline demographic characteristics. Values are frequency unless otherwise stated Factor

Metal-backed All-polyethylene

Total

Patients 29 27 56 Age, mean years (SD) 68 (4) 68 (4) 68 (4) BMI, mean (SD) 28 (4) 29 (3) 28 (3) Sex Female 17 13 30 Male 12 14 26 ASA classification I 4 7 11 II 18 17 35 III 7 3 10 Surgeon #1 14 14 28 #2 15 13 28 Ahlbäck classification II 5 4 9 III 23 23 46 IV 1 0 1 HKA postoperative Varus (< 177°) 7 3 10 Neutral (177–183°) 15 17 32 Valgus (> 183°) 2 4 6 Missing a 5 3 8 SD = standard deviation, HKA = hip–knee–ankle angle. Some patients had no postoperative long-leg radiographs taken and HKA could not be assessed. a

Ethics, registration, funding, and potential conflicts of interest Approval of the Regional Ethical Review Board in Lund was obtained before recruitment (entry no. 2014/513). This study was registered at the ISRCTN Registry (ISRCTN10744502) and was conducted in accordance with the CONSORT statement. All patients provided informed consent. Stryker funded this study but did not take any part in the design, conduct, analysis, and interpretations stated in this paper.

Results 60 patients were included and randomized to either the APT-PS or the MBT-PS total knee prosthesis. After randomization, 4 patients were excluded. 56 patients were thus included in the analysis (Figure 1). During follow-up, 9 patients withdrew or had radiographs that could not be analyzed, leaving 47 patients for analysis at 2 years (Figure 1). Age, BMI, sex, ASA score, and Ahlbäck classification were similar at baseline. Each surgeon operated on approximately half of the patients in both groups. Postoperatively, the MBT implants seemed to be more in varus compared with the APT (Table 1). The mean MTPM across 3, 12, and 24 months was similar in both groups. The mean MTPM change from 12 to 24 months was –0.01 mm (CI –0.19 to 0.17) in the MBT group and 0.03 (CI –0.14 to 0.21) in APT group (Table 2; Figure

Acta Orthopaedica 2019; 90 (6): 590–595

593

Table 2. Mean (95% CI) MTPM in mm of the metal-backed tibial implant group (MBT) and the all-polyethylene tibial implant group (APT) at 3, 12, and 24 months follow-up

Mean MTPM (mm) 1.8

MBT Unstable MBT Unstable MBT APT Unstable APT

1.6 1.4

Time (months) MBT APT 3 0.41 (0.33–0.50) 0.46 (0.36–0.57) 12 0.57 (0.44–0.70) 0.61 (0.49–0.73) 24 0.56 (0.42–0.69) 0.64 (0.50–0.77)

1.2 1.0 0.8 0.6

Figure 2. Mean (95% CI) MTPM in mm of the metal-backed tibial implant group (MBT) and the all-polyethylene tibial implant group (APT) at 3, 12, and 24 months follow-up. The MTPM of the 3 unstable implants is plotted and all 3 show continuous migration between 12 and 24 months’ follow-up.

0.4 0.2 0 0

Translation along the longitudinal axis (mm)

Rotation about the longitudinal axis (°)

Rotation a about the transverse axis (°)

0.3

0.2

MBT APT

0.2

0.1

–0.1

–0.2

–0.3

–0.3 0

Months from index operation

Figure 3. Mean translation along the longitu dinal axis in mm with 95% confidence intervals. A positive value indicates tibial lift-off and a negative value indicates subsidence of the tibial implant.

Months from index operation

–0.4 0

Months from index operation

Figure 4. Mean rotation along the longitudinal axis in degrees with 95% confidence intervals. A positive value indicates internal rotation and a negative value indicates external rotation of the tibial implant.

2). 2 implants in the MBT and 1 in the APT group displayed > 0.2 mm MTPM between 1- and 2-year follow-up, and were considered unstable (Figure 2). The MBT group showed liftoff (positive), while the APT group showed tibial subsidence (negative) (Figure 3, Translation along the longitudinal axis). A different migration pattern between the groups was also visible in the rotation along the longitudinal axis, being external (negative) in the MBT and internal (positive) in the APT group (Figure 4, Rotation about the longitudinal axis). Other translations and rotations were similar between groups with backward tilting (negative) being the most prominent direction of migration in both groups (Figure 5, Rotation along the transverse axis; Figures 6–8, see Supplementary data). None of the patients were scheduled for revision surgery. 50 double measurements were made at 1-year follow-up. The precision of the measurements of the translations and rotations were 0.1 mm and 0.1 degrees. The mean condition number of the tibial bone and the prosthesis was 42 (range 20–108) and 40 (range (21–114), respectively. The mean error of rigid body fitting

Months from index operation

Figure 5. Mean rotation along the transverse axis in degrees with 95% confidence intervals. A positive value indicates forward tilting and a negative value indicates backward tilting of the tibial implant.

was 0.14 (range 0.04–0.34) and 0.08 (range 0.01–0.35) of the tibial bone and the prosthesis, respectively. The KSS Knee scores across 3, 12, and 24 months were similar in both groups. KSS Function score was also similar. Moreover, no statistically significant difference was found in the KOOS subscores or in the FJS (Table 3, see Supplementary data).

Discussion We found similar MTPM between the APT-PS and MBT-PS at 2-year follow-up; the translation and rotation along and about the 3 orthogonal axes were different for longitudinal translation and rotation. Van Hamersveld et al. (2018), who used a CS design, and other RSA studies on CR designs reported comparable MTPM values to those in our study (Adalberth et al. 2000, 2001; Norgren et al. 2004; Hyldahl et al. 2005). These findings suggest that, although PS implants most likely

594

experience different shear forces at the implant–bone interface, the MTPM values after 2-year follow-up are comparable to CR and CS designs. Furthermore, despite the relative elasticity of a full APT component, this did not result in a difference in migration compared to a MBT component. This may imply that the polyethylene insert within the metal baseplate gives enough peak stress absorption in the PS design. The difference in translation along the longitudinal axis was previously described by Adalberth et al. (2000) who compared a low-conforming APT and MBT with RSA, and concluded that this finding might be explained by an increase in tensile forces in the less flexible MBT (Bartel et al. 1982). In our study, the subsidence of the APT and lift-off of the MBT stabilized after 3 months. The difference in rotation about the longitudinal axis (i.e., internal/external rotation) between the MBT and APT in our study might be due to unmeasured differences between the groups such as the alignment of the tibial component. Another explanation might be the minor differences in the postoperative HKA between groups, but the groups are too small to draw any valid conclusion. We reported signed migration values in contrast to several other RSA studies. In order to allow comparison between RSA studies and to understand the direction of migration, reporting signed values is preferred as was previously suggested by Valstar et al. (2005). Gudnason et al. (2017) suggested that it was better to use the transversal rotation for analysis of RSA migration data as it was a better predictor for aseptic loosening than MTPM. The rotation in the transverse plane was posterior for both groups (Figure 5, Rotation along the transverse axis). The posterior rotation of the tibial implants in both groups could be due to anterior engagement of the cam-post mechanism of the PS design, which engages in extension. Banks et al. (2002) found that TKAs are frequently aligned in relative hyperextension, which might explain the rotation in the present study. Another factor contributing to the posterior rotation might be the singleradius design of the TKA used in our study, which might play a role as the center of rotation lies more posteriorly compared with multi-radius designs (D’Lima et al. 2001). Whether this migration pattern has clinical consequences remains unclear and should be studied further when longer follow-up data become available. The KSS Knee and Function scores increased postoperatively and were comparable in both groups during follow-up, which is consistent with previous studies (Adalberth et al. 2000, 2001). The KOOS subscales and the FJS also showed similar results. De Carvalho et al. (2013) used different clinical outcomes (the Oxford Knee Score, the Western Ontario and McMaster Universities Arthritis Index, and the Short form-12 scores), but also found no difference between groups. Our study with an all-polyethylene PS design failed to show superiority of either APT or MBT. Nevertheless, Chambers et al. (2016) estimated that a reduction of 42% in costs could be achieved if the APT were used. However, the actual costs of an implant differ widely and the total costs of TKA treatment

Acta Orthopaedica 2019; 90 (6): 590–595

consist of more than just the tibial component, including personnel, equipment, and space costs. In addition, the financial benefit of the APT might not outweigh the limitations as it cannot be coated and liner exchange is not possible. These factors may be among the reasons why orthopedic surgeons continue to use the MBT TKA as the implant of first choice even though some suggest that the APT could be an acceptable treatment in patients above 70 years of age or with rheumatoid arthritis (Nouta et al. 2012a). A limitation of this study is the lack of power to detect a difference in clinical scores between the two groups. RSA studies, in general, include small groups and probably fail to detect any differences due to this small sample size. Including more patients, however, would nullify the strength of RSA studies as it can measure migration with high precision and, therefore, only a small sample size is needed to assess the stability of implants. Another limitation is the difference in polyethylene as the polyethylene insert of the MBT tray was made of highly crosslinked polyethylene and the APT was made from conventional polyethylene. Ideally, the polyethylene in both implants would be the same, but this was not possible due to manufacturing limitations. In summary, the APT-PS TKA prosthesis has comparable migration to the MBT-PS TKA in terms of MTPM measured by RSA at 2 years of follow-up, even though there was a different pattern in longitudinal translation and rotation. No differences in complications or clinical outcomes were found between the two groups. Supplementary data Table 3 and Figures 6–8 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/ 17453674.2019.1668602

SH: Data collection, statistical analysis, interpretation of data. PM, BL: Statistical analysis, interpretation of data. RN: Interpretation of data. STL: Study design, coordination of study, interpretation of data. All authors critically revised the manuscript. The authors would like to thank Dr. Koen T. van Hamersveld for his contribution to the design of the study and the collection of data. Acta thanks Anders Henricson and Leif Ryd for help with peer review of this study. American Academy of Orthopaedic Surgeons (AAOS). Fifth AJRR Annual Report on Hip and Knee Arthroplasty Data, 2018. Retrieved from http:// ajrr.net/images/annual_reports/AAOS-AJRR-2018-Annual-Report-final. pdf?hsCtaTracking=c794b145-8b50-405a-af5c-666a0841a730%7C6996bb535b74-4d65-bb51-b24f37c55c9d. Accessed February 8, 2019. Adalberth G, Nilsson K G, Bystrom S, Kolstad K, Milbrink J. Low-conforming all-polyethylene tibial component not inferior to metal-backed component in cemented total knee arthroplasty: prospective, randomized radiostereometric analysis study of the AGC total knee prosthesis. J Arthroplasty 2000; 15(6): 783-92.

Acta Orthopaedica 2019; 90 (6): 590–595

Adalberth G, Nilsson K G, Bystrom S, Kolstad K, Milbrink J. All-polyethylene versus metal-backed and stemmed tibial components in cemented total knee arthroplasty: a prospective, randomised RSA study. J Bone Joint Surg Br 2001; 83(6): 825-31. Banks S A, Harman M K, Hodge W A. Mechanism of anterior impingement damage in total knee arthroplasty. J Bone Joint Surg Am 2002; 84-A(Suppl. 2): 37-42. doi:10.2106/00004623-200200002-00004 Bartel D L, Burstein A H, Santavicca E A, Insall J N. Performance of the tibial component in total knee replacement. J Bone Joint Surg Am 1982; 64(7): 1026-33. Browne J A, Gall Sims S E, Giuseffi S A, Trousdale R T. All-polyethylene tibial components in modern total knee arthroplasty. J Am Acad Orthop Surg 2011; 19(9): 527-35. Chambers M C, El-Othmani M M, Sayeed Z, Anoushiravani A, Schnur A K, Mihalko W M, Saleh K J. Economics of all-polyethylene versus metalbacked tibial prosthesis designs. Orthopedics 2016; 39(3 Suppl.): S61-6. doi:10.3928/01477447-20160509-18 Cheng T, Zhang G, Zhang X. Metal-backed versus all-polyethylene tibial components in primary total knee arthroplasty. Acta Orthop 2011; 82(5): 589-95. doi:10.3109/17453674.2011.618913 Cheng T, Pan X, Liu T, Zhang X. Tibial component designs in primary total knee arthroplasty: should we reconsider all-polyethylene component? Knee Surg Sports Traumatol Arthrosc 2012; 20(8): 1438-49. doi:10.1007/ s00167-011-1682-y D’Lima D D, Poole C, Chadha H, Hermida J C, Mahar A, Colwell C W Jr. Quadriceps moment arm and quadriceps forces after total knee arthroplasty. Clin Orthop Relat Res 2001; 392: 213-20. doi:10.1097/00003086200111000-00026 De Carvalho B R, Yassaie O S, Muir D C. Modular versus all-polyethylene tibial components: comparison of pre- and early post-operative patient scores in total knee replacement. ANZ J Surg 2013; 83(10): 784-7. doi:10.1111/ans.12270 Dutch Arthroplasty Register (LROI). Online LROI annual report 2018. Retrieved from http://www.lroi-rapportage.nl/media/pdf/PDF%20Online_ LROI_annual_report_2018.pdf. Accessed February 8, 2019. Garling E H, Valstar E R, Nelissen R G. Comparison of micromotion in mobile bearing and posterior stabilized total knee prostheses: a randomized RSA study of 40 knees followed for 2 years. Acta Orthop 2005; 76(3): 353-61. Gioe T J, Maheshwari A V. The all-polyethylene tibial component in primary total knee arthroplasty. J Bone Joint Surg Am 2010; 92(2): 478-87. doi:10.2106/jbjs.i.00842 Gudnason A, Adalberth G, Nilsson K G, Hailer N P. Tibial component rotation around the transverse axis measured by radiostereometry predicts aseptic loosening better than maximal total point motion. Acta Orthop 2017; 88(3): 282-7. doi: 10.1080/17453674.2017.1297001. Gustke K A, Gelbke M K. All-polyethylene tibial component use for elderly, low-demand total knee arthroplasty patients. J Arthroplasty 2017; 32(8), 2421-6. doi:10.1016/j.arth.2017.02.077 Hyldahl H, Regner L, Carlsson L, Karrholm J, Weidenhielm L. All-polyethylene vs. metal-backed tibial component in total knee arthroplasty: a randomized RSA study comparing early fixation of horizontally and completely cemented tibial components, Part 2: Completely cemented components: MB not superior to AP components. Acta Orthop 2005; 76(6): 778-84. doi:10.1080/17453670510045363 ISO16087:2013(E). Implants for surgery: Roentgen stereophotogrammetric analysis for the assessment of migration of orthopaedic implants. Geneva, Switzerland: International Organization for Standardization; 2013.

595

Kaptein B L, Valstar E R, Stoel B C, Rozing P M, Reiber J H. A new type of model-based Roentgen stereophotogrammetric analysis for solving the occluded marker problem. J Biomech 2005; 38(11): 2330-4. doi: 10.1016/j. jbiomech.2004.09.018 Longo U G, Ciuffreda M, D’Andrea V, Mannering N, Locher J, Denaro V. All-polyethylene versus metal-backed tibial component in total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2017; 25(11): 3620-36. doi:10.1007/s00167-016-4168-0 Molt M, Toksvig-Larsen S. Similar early migration when comparing CR and PS in Triathlon TKA: a prospective randomised RSA trial. Knee 2014; 21(5): 949-54. doi:10.1016/j.knee.2014.05.012 Norgren B, Dalen T, Nilsson K G. All-poly tibial component better than metalbacked: a randomized RSA study. Knee 2004; 11(3): 189-96. doi:10.1016/ s0968-0160(03)00071-1 Norwegian Arthroplasty Register. Norwegian National Advisory Unit on Arthroplasty and Hip Fractures: 2018 Annual Report. Retrieved from http://nrlweb.ihelse.net/eng/Rapporter/Report2018_english.pdf. Accessed February 8, 2019. Nouta K A, Pijls B G, Nelissen R G. All-polyethylene tibial components in TKA in rheumatoid arthritis: a 25-year follow-up study. Int Orthop 2012a; 36(3): 565-70. doi:10.1007/s00264-011-1341-4 Nouta K A, Verra W C, Pijls B G, Schoones J W, Nelissen R G. All-polyethylene tibial components are equal to metal-backed components: systematic review and meta-regression. Clin Orthop Relat Res 2012b; 470(12): 354959. doi:10.1007/s11999-012-2582-2 Pijls B G, Valstar E R, Nouta K A, Plevier J W, Fiocco M, Middeldorp S, Nelissen R G. Early migration of tibial components is associated with late revision: a systematic review and meta-analysis of 21,000 knee arthroplasties. Acta Orthop 2012; 83(6): 614-24. doi: 10.3109/17453674.2012.747052 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. doi:10.1186/1471-2288-12-35 Ryd L, Albrektsson B E, Carlsson L, Dansgard F, Herberts P, Lindstrand A et al. Roentgen stereophotogrammetric analysis as a predictor of mechanical loosening of knee prostheses. J Bone Joint Surg Br 1995; 77(3): 377-83. Swedish Knee Arthroplasty Register (SKAR). Annual Report 2018. Retrieved from http://www.myknee.se/pdf/SVK_2018_Eng_1.0.pdf. Accessed February 8, 2019. Steinberg D R, Steinberg M E. The early history of arthroplasty in the United States. Clin Orthop Relat Res 2000; 374: 55-89. Stryker. Triathlon All-Polyethylene Tibial Component Surgical Protocol Addendum, 2015. Retrieved from http://www.bizwan.com/_mydoc/stryker/ Knee/020%20Triathlon%20All-Polyethylene%20Tibial%20Component%20Surgical%20Protocol%20Addendum.pdf. Accessed May 6, 2019. Valstar E R, Gill R, Ryd L, Flivik G, Borlin N, Karrholm J. Guidelines for standardization of radiostereometry (RSA) of implants. Acta Orthop 2005; 76(4): 563-72. doi:10.1080/17453670510041574 Van Hamersveld K T, Marang-Van De Mheen P J, Nelissen R, Toksvig-Larsen S. Migration of all-polyethylene compared with metal-backed tibial components in cemented total knee arthroplasty. Acta Orthop 2018; 89(4): 412-7. doi:10.1080/17453674.2018.1464317 Voss B, El-Othmani M M, Schnur A K, Botchway A, Mihalko W M, Saleh K J. A meta-analysis comparing all-polyethylene tibial component to metalbacked tibial component in total knee arthroplasty: assessing survivorship and functional outcomes. J Arthroplasty 2016; 31(11): 2628-36. doi: 10.1016/j.arth.2015.08.035

596

Acta Orthopaedica 2019; 90 (6): 596–601

Anterior cruciate ligament reconstruction-related patient injuries: a nationwide registry study in Finland Kirsi-Maaria NYRHINEN 1, Ville BISTER 2, Teemu HELKAMAA 1, Arne SCHLENZKA 1, Henrik SANDELIN 3, Jerker SANDELIN 4, and Arsi HARILAINEN 4 1 Department of Orthopaedics and Traumatology, Helsinki University Central Hospital; 2 Department of Surgery, Hyvinkää Hospital, Hyvinkää; 3 Orthopaedic Department, Liverpool Hospital, Sidney, New South Wales, Australia; 4 ORTON Orthopaedic Hospital, Invalid Foundation, Helsinki,

Correspondence: kirsi-maaria.nyrhinen@hus.fi Submitted 2019-03-27. Accepted 2019-08-20.

Background and purpose — Treatment outcomes of anterior cruciate ligament (ACL) injuries are generally good, but complications after ACL reconstruction (ACLR) can result in long-lasting problems. Patient injury claims usually fall on the more severe end of the complication spectrum. They are important to investigate because they may reveal the root causes of adverse events, which are often similar regardless of the complication’s severity. Therefore, we analyzed ACL-related patient injuries in Finland, the reasons for these claims, causes of complications, and grounds for compensation. Patients and methods — We analyzed all claims filed at the Patient Insurance Centre (PIC) between 2005 and 2013 in which the suspected patient injury occurred between 2005 and 2010. This study also reviewed all original patient records and available imaging studies. General background data were obtained from the National Care Register for Social Welfare and Health Care (HILMO). Results — There were 248 patient injury claims, and 100 of these were compensated. Compensated claims were divided into 4 main categories: skill-based errors (n = 46), infections (n = 34), knowledge-based errors (n = 6), and others (n = 14). Of the compensated skill-based errors, 34 involved graft malposition, 26 of them involved the femoralside tunnel. All compensated infections were deep surgical site infections (DSSI). Interpretation — This is the first nationwide study of patient injuries concerning ACLRs in Finland. The most common reasons for compensation were DSSI and malposition of the drill tunnel. Therefore, it would be possible to decrease the number of serious complications by concentrating on infection prevention and optimal surgical technique.

Finland

Anterior cruciate ligament reconstruction (ACLR) has a high success rate, and most patients can return to sporting activities postoperatively with nearly normal knee function. Complications are rare but can be devastating. A new trauma, technical errors, graft failure, problems with the fixation methods, postoperative infections, and venous thrombosis are the most common complications and often lead to reoperations and prolonged rehabilitation (Schulz et al. 2007, Saper et al. 2014, Magnussen et al. 2015, Christensen and Miller 2018). However, the literature contains little information on ACLRrelated patient injury claims. A range of patient injuries in Finland have been previously investigated by several research groups, including total hip and knee arthroplasty, children’s tibial and femoral fractures, distal radius fractures, and fatal complications (Palmu et al. 2009, 2010, Järvelin et al. 2012, Hakala et al. 2014, Helkamaa et al. 2016, Sandelin et al. 2018). This study analyzes and describes ACL-related patient injuries in Finland and the root causes that lead to these complications.

Patients and methods This is a descriptive retrospective register study. Unlike other Nordic countries, there is no national anterior cruciate ligament (ACL) register in Finland. However, there is the Care Register for Social Welfare and Health Care (HILMO), which is maintained by the National Institute for Health and Welfare (THL). The HILMO collects national information from all healthcare sectors on patients who undergo an operation. This study obtained background data on ACL patients treated in Finland between 2005 and 2010 from the HILMO. Data on claims and claimants was obtained from

Acta Orthopaedica 2019; 90 (6): 596–601

Table 2. Criteria entitling patient to compensation according to the Finnish Patient Injury Act (Patient Injury Act 25.7.1986/585) 1. 2. 3. 4. 5. 6. 7.

Treatment injury Equipment-related injury Infection injury Accidental injury Injury from damage to healthcare facilities Injury due to delivery of pharmaceuticals Unreasonable injury

597

Number of procedures 4,000 open ACLR arthroscopic ACLR 3,000

2,000

1,000

the Patient Insurance Centre (PIC; www.pvk.fi/en/). ACL injuries were identified from this national register through the International Classification of Diseases 10th edition (ICD-10) codes (Table 1, see Supplementary data), and ACLRs were identified through the Nordic Medico-Statistical Committee (NOMESCO) procedure codes NGE30 and NGE35. The code NGE30 stands for “open ACLR,” and NGE35 stands for “arthroscopic ACLR.” The study period 2005 to 2010 was chosen to allow for the complete processing of all patient injury claims by the PIC. It can take 1 to 7 years to collect the necessary material, analyze it, and sometimes even re-analyze it. All patient injury claims obtained for this study were related to the ICD-10 code S83.5. They were analyzed by reviewing the claimants’ original patient records and imaging studies (when available). The PIC collects statistics and promotes research to support patient health and safety. Public and private healthcare units must have patient injury insurance. The insurance companies that offer this insurance must be members of the PIC. A single healthcare unit or a healthcare district cannot be a member of the PIC directly. Patient injuries have been centralized to the PIC in Finland. The PIC’s function is based on the Patient Injury Act. According to the Act, there are 7 criteria that entitle a patient to compensation (Table 2). At least 1 of the 7 criteria must be met to comprise a patient injury case (Patient Injury Act 1986). Finland and the other Nordic countries have a no-fault patient insurance system as opposed to the tort insurance system in the United Kingdom and United States (Palonen et al. 2005, Järvelin and Häkkinen 2012, Patient Insurance Centre n.d.). If a patient is unsatisfied with their care for any reason, the PIC offers an impartial estimation concerning that care. When the claim arrives at the PIC, all information regarding the care in question is collected. Using this information, an independent specialist estimates whether an experienced healthcare professional would have treated the patient differently and if, thereby, the event leading to the compensation claim could have been avoided. When infection or unreasonable injury is suspected, a specialist and the PIC evaluate whether the consequences of the complication are too much for the patient to tolerate. Therefore, the operation of the PIC is based on the following questions: was the injury preventable and was the consequence tolerated (tolerable disadvantage, temporary disadvantage, or permanent

2005 2006 2007 2008 2009 2010

Figure 1. ACL reconstructions or revisions (n) (procedure codes NGE30 and NGE35) between 2005 and 2010.

disadvantage) (Mikkonen 2004, Helkamaa et al. 2016, Patient Insurance Centre n.d.). In general, patients have 3 years to file a claim after treatment. If the claimant is not satisfied with the decision, he/she may refer the claim to the Patient Injury Board, which consists of several independent specialists. If the claimant is still dissatisfied, he/she can take the claim to the general court. This rarely occurs. The PIC compensates the true expenses that are due to the patient injury. These include the cost of hospital care, visits to outpatient clinics, rehabilitation, visiting nurse services, and laboratory and imaging studies. If there is a decline in work performance or loss of primary income, an estimation of work ability is made. In these cases, compensation is paid from the patient’s insurance to prevent a decline in income. In addition, the PIC defines the disability and harm that results from the patient’s injury as either transient or permanent and compensates this accordingly. The severity of the patient’s injury, their recovery time, their level of required care, and any patient-specific factors all affect the decision and amount of compensation (Patient Insurance Centre n.d.). Ethics, funding, and potential conflicts of interest Permission for this study was provided by the PIC and the Ethical Board of Helsinki University (376/13/03/02/2015). No funding was received. The authors have no conflicts of interest.

Results Background data: HILMO Between 2005 and 2010, 31,643 patients (64% men) were diagnosed with an ACL injury (S83.5). During the same period, 17,041 ACLRs or revision procedures were performed in Finland (Figure 1). It is impossible to separate revisions from primary reconstructions based on registry data alone because the NOMESCO procedure codes are the same. The

598

Acta Orthopaedica 2019; 90 (6): 596–601

Number of procedures

2,500

2,000

private clinic university hospital other public hospital

uncompensated compensated

1,500

30 1,000 20 500 0

10 0

2005 2006 2007 2008 2009 2010

Figure 2. Distribution of different healthcare units performing ACL surgery between 2005 and 2010 in Finland.

Figure 3. Patient injury claims during 2005–2010.

Table 3. Total number, sex distribution, and mean age of anterior cruciate ligament reconstruction (ACLR) patients (HILMO) and patient injury claimants (PIC) between 2005 and 2010: background data from the HILMO and research data from the PIC

Table 4. Reasons for patient complaints a

Sex Male Female Total

HILMO: ACLR patients PIC: patient injury claims n mean age (range) n mean age (range) 11,293 5,646 16,939

31 (4–76) 35 (6–77) 33 (4–77)

141 107 248

35 (9–71) 35 (8–75) 35 (8–75)

distribution of healthcare units is provided in Figure 2. The distribution of sex and the mean age of ACLR patients within the HILMO data and patient injury claimants in the PIC data are provided in Table 3. The research data: PIC This study found 248 filed patient injury claims between 2005 and 2013 that concerned a suspected ACL-related patient injury occurring between 2005 and 2010. Though injuries generally occurred during sports (n = 117), 67 occurred during leisure time and 41 were work-related. The remaining 23 were traffic accidents, accidents at home, or occurred under circumstances that remain unclear. 239 injuries were treated operatively. Of these, 231 were ACLRs, 2 were ACL avulsion fixations, and 6 were other surgical procedures. All ACLRs, compensated or not, were performed with arthroscopic assistance. A hamstring graft was used in 188 of these operations, a bone–tendon– bone graft (BTB) was used in 33, and different graft sources were used in the remaining cases. An anteromedial (AM) drilling technique was used in 111 operations, a transtibial (TT) drilling technique was used in 74, and it was impossible to confirm the drilling technique used in 45 as the data were missing or the surgery reports were incomplete. The average operation time was 72 min (33–191). Use of a prophylactic antibiotic was documented in 196 operations, most commonly a single dose of intravenous cefuroxime 1.5g (n = 137).

Reason n Pain 95 Financial difficulties (sick leave, unemployment, additional expenses) 60 Infection 56 Reoperation 53 Decline in work performance 48 Delay in care 45 Delay in diagnosis 38 Decline in general performance 31 Prolonged rehabilitation 29 Prolonged use of antibiotics 22 Edema 15 Instability 15 Numbness 13 Arthrosis 11 Deep venous thrombosis 10 a Patients

complained for 65 different experienced reasons. Each of those reasons have been collected and are presented in this table. Reasons that appeared less than 10 times are not presented.

Reasons to file a claim (Table 4) Figure 3 shows patient injury claims (n = 248) between 2005 and 2010 based on the search criteria. On average, 41 ACLRrelated patient injury claims were filed annually, and 17 claims were compensated. Compensated claims The PIC compensated 100/248 of all ACLR-related claims (Table 5, see Supplementary data). The 2 most common reasons for compensation were technical errors/skill-based errors (n = 46) and postoperative deep surgical site infections (DSSI) (n = 34). Only 6 claims were reimbursed due to knowledge-based errors. The remaining claims were single cases. The average age of the compensated claimants at the time of their injury was 33 years (14–75), and 55 of them were male.

Acta Orthopaedica 2019; 90 (6): 596–601

Skill-based errors (n = 46) Technical errors leading to instability, loss in a patient’s range of motion (ROM), loss of a graft, or reoperation are considered patient injuries. Of all compensated patient injury claims, 34/100 were due to graft malposition. The most common error was to the anterior femoral tunnel, which was compensated in 15 cases (Table 6, see Supplementary data). In 9 cases, an additional arthroscopy was performed. This occurred when a ruptured ACL was not reconstructed during the first arthroscopy (n = 5) or a broken instrument or fixation material needed to be removed after ACLR (n = 4). There were 6 claims concerning saphenous nerve problems after ACLR but only 2 were compensated. In 1 case, the BTB graft was damaged at the bone–ligament junction. It was estimated to be long enough without the bone block and was fixed with a screw to the femoral canal. During follow-up, the knee became loose and required a revision. Compensation was granted because, with a careful operation technique, the BTB graft would have remained intact and a revision could have been avoided. Infections (n = 34) All compensated infections were DSSI, and they comprised 34/100 of the compensated claims. The graft was lost in 20/34 of these infections. There was a mean of 1.6 (1–6) arthroscopic lavages per infection. The treatment period at the hospital after infection was on average 14 days (2–33). The most common bacteria were Staphylococcus aureus (n = 9), coagulase-negative coccus (n = 4), and Staphylococcus epidermidis (n = 3). In 13 infection cases, the bacteria could not be identified, despite the culture, or the information was missing. A hamstring graft was used in 30/34 cases, a BTB graft was used in 2/34 cases, and the surgery report was missing in 2/34 cases. A prophylactic antibiotic was used in 27/34 of the postoperative infections, while 3/34 did not receive any antibiotics before their operation. In 4/34 cases, information on prophylactic antibiotics was missing. Paid compensations The PIC paid €823,800 in compensation to 93 patients. In 7 of these cases, the documents concerning the paid compensations were missing. The median value per patient injury was €5,600 (€400–126,000). The largest compensation was for an infection that led to a 2-phase reconstruction. The minimum amount was paid to 3 patients who were considered to have suffered only transient discomfort due to additional arthroscopy (Table 7, see Supplementary data). Comparison of HILMO and PIC register data In a comparison of the total number of ACLRs (17,041) to the PIC materials containing 231 ACLRs, only 1.4% of ACLRs were sent to the PIC. There were 34 compensated DSSIs among the 17,041 ACLRs. According to these data, the risk of a DSSI severe enough to result in compensation was

599

0.2%. Furthermore, there were 56 infection-related claims. Therefore, the overall risk of infection was at least 0.3%.

Discussion This is the first nationwide study of ACL-related patient injury claims and compensation in Finland. Filed patient injuries usually entail more severe complications because these are more readily reported. These claims are important to investigate because they can reveal the root causes of complications that can often be similar regardless of severity. The most common grounds for compensation found were technical errors and infections. This study’s results agree with the current knowledge regarding ACLR-related complications. The femoral side is the most difficult side on which to drill and errors are, therefore, most likely to occur there (Sommer et al. 2000, Wright et al. 2010, Morgan et al. 2012, Chen et al. 2013). Most compensated graft malpositions (26/34) were due to a technical error on the femoral side. The length and tension of the graft is determined by the positioning of the femoral drill tunnel. There are criteria that guide the positions of both the femoral and the tibial drill tunnels (Bernard et al. 1997, Marchant et al. 2010, Kopf et al. 2012, Robin and Lubowitz 2014, Samitier et al. 2015, Robin et al. 2015). This study’s findings suggest that, simply by following existing standard surgical techniques, many ACLR-related patient injuries could be avoided. Infections were the second most common reason for compensation in this study. The incidence of infection after ACLR is 0.5–1%, according to contemporary literature (Westermann et al. 2017, Bansal et al. 2018). Postoperative infections can cause severe, long-lasting consequences. Antibiotics are used for several weeks, and the infected knee often requires arthroscopic lavages. Eventually, the graft is often lost and a 1- or 2-phase revision is needed. In the worst scenario, the infection leads to poor joint function and arthrosis (Vertullo et al. 2012). Compensation is granted if the postoperative infection is unexpected (the patient has no predisposing health conditions that push the infection risk above 2%) and the consequences are so severe that patient should not have to tolerate them. There were more infections among patients whose operations included hamstring grafts, as previously reported (Maletis et al. 2013, Gifstad et al. 2014, Okoroha et al. 2016, Westermann et al. 2017, Randsborg et al. 2018). Among the compensated claims, a hamstring graft was the most common. However, the exact percentage of hamstring grafts used in Finland during ACLRs is unknown. In Norway, this has led to an increased use of BTB grafts. However, Pérez-Prieto et al. (2016) and Phegan et al. (2016) have found that the likelihood of postoperative infection can be significantly reduced if the hamstring graft is pre-soaked in a vancomycin solution.

600

In 2018, Randsborg et al. published their study on 101 compensated claims as a result of ACLRs performed in Norway from 2005 to 2015. The most common reason for compensation was postoperative infection (39%), the second most common was inappropriate surgical technique (27%), and the third most common was delayed diagnosis (13%). Risk of postoperative infection increased when a hamstring graft was used. The Norwegian register reveals that the use of hamstring grafts decreased from 2005 to 2015. The reason for this is the higher risk of revision. This same trend was previously demonstrated by Persson et al. (2014) who published their results regarding the Norwegian Cruciate Ligament Registry between 2004 and 2012. In addition, technical errors, and especially an incorrect femoral tunnel placement, were common mistakes. These findings correspond with this study’s findings. Functional results between the BTB graft and the hamstring graft are similar despite the anterior knee pain, which is more common with a BTB graft. Risk of infection and risk of a graft failure increase with a hamstring graft (Gifstad et al. 2014, Okoroha et al. 2016). Other patient injuries detected in this study included diagnostic errors and treatment delays, which were clinically relevant errors. After an injury, patients often visit an emergency department or a general practitioner who does not have access to magnetic resonance imaging (MRI). A clinical diagnosis of an ACL rupture is difficult in an acute phase, and it is therefore difficult to estimate an acceptable delay. According to contemporary literature, ACL reconstruction should be performed within 1 to 6 months following the injury (Francis et al. 2001, Taketomi et al. 2018). However, determining a diagnosis is difficult and recommendations for scheduling surgery vary. Therefore, it can be challenging to estimate whether a patient has suffered too much because of a delay and a patient injury has occurred. This study has 2 major limitations. Due to the retrospective study design, this study suffers from the limitations of the registry. Finland does not have a national ACL register; therefore, it is difficult to determine the operation volumes of healthcare units or single surgeons. Furthermore, this study is unable to separate primary ACLRs from revisions or to assess the exact number of ACL patients because the ICD-10 codes do not distinguish between ACL and PCL injuries. In addition, the exact information on grafts used, fixation methods, drilling techniques, and rehabilitation programs at the national level remain unknown. To compensate for these shortcomings, this study combined 2 nationwide databases and analyzed the original patient records and imaging studies (when available) of all patient injuries. This allowed the study to gather more specific data and exclude any errors that commonly occur when only registry data are used. As previous studies have demonstrated, the coverage and accuracy of the HILMO is very good (Mattila et al. 2008, Sund 2012). Patients do not complain as often concerning mild complications as they do for severe complications. Although

Acta Orthopaedica 2019; 90 (6): 596–601

reported patient injuries usually fall on the more severe end of the complication spectrum, they can still be investigated to determine a wide range of complications. The root causes of adverse events are often similar regardless of the complication’s severity. Based on these data, this study calculated that the overall risk of infection after ACLR in Finland was at least 0.3%. Based on the average risk of infection after ACLR in the literature (Westermann et al. 2017, Bansal et al. 2018), this study can estimate that coverage of serious complications, such as infection, is fairly good according to PIC data. The rate of coverage may not be 100%, but it is far better than what was previously suggested by Pukk et al. (2003) who argued that patient injury claims would represent only 3% of patients who had a complication that would fulfill patient injury criteria. Therefore, the conclusions drawn from this registry study are more likely to be widely applicable and reliable. In addition, this study’s results regarding typical error types are in line with previous studies regarding Scandinavian ACL registry data (Persson et al. 2014, Randsborg et al. 2018). Conclusion Complications leading to filed patient injury claims are quite rare after ACL reconstructions, but they can lead to devastating consequences. According to this study, the best way to reduce ACLR-associated complications is the prevention of DSSIs, optimal femoral canal drilling, and optimal graft placement. Supplementary data Tables 1, 5, 6, and 7 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453 674.2019.1678233

K-MN: collection, analysis, and interpretation of data; writing of manuscript. VB: conception of study, collection, and interpretation of data; writing of manuscript. TH, HS: design of data matrix, writing of manuscript. AS: collection of data. JS, AH: conception of study, interpretation of data, writing of manuscript. The authors wish to thank Saija Lehtinen for the PIC material and Jutta Järvelin and Pia Peltola for the HILMO material. Acta thanks Jon Olav Drogset and Pelle Gustafson for help with peer review of this study.

Bansal A, Lamplot J D, VandenBerg J, Brophy R H. Meta-analysis of the risk of infections after anterior cruciate ligament reconstruction by graft type. Am J Sports Med 2018; 46(6): 1500-8. Bernard M, Hertel P, Hornung H, Cierpinski T. Femoral insertion of the ACL: radiographic quadrant method. Am J Knee Surg 1997; 10(1): 14-21 Chen J L, Allen C R, Stephens T E, Haas A K, Huston L J, Wright R., Feeley B T, the Multicenter ACL Revision Study (MARS) Group. Differences in mechanisms of failure, intraoperative findings, and surgical characteristics between single- and multiple-revision ACL reconstructions: a MARS cohort study. Am J Sports Med 2013; 41(7): 1571-8.

Acta Orthopaedica 2019; 90 (6): 596–601

Christensen J E, Miller M D. Knee anterior cruciate ligament injuries: common problems and solutions. Clin Sports Med 2018; 37(2): 265-80. Francis A, Thomas R D, McGregor A. Anterior cruciate ligament rupture: reconstruction surgery and rehabilitation. A nation-wide survey of current practice. Knee 2001; 8:13-8. Gifstad T, Foss O A, Engebretsen L, Lind M, Forssblad M, Albrektsen G, Drogset J O. Lower risk of revision with patellar tendon autografts compared with hamstring autografts: a registry study based on 45,998 primary ACL reconstructions in Scandinavia. Am J Sports Med 2014; 42:2319-28. Hakala T, Vironen J, Karlsson S, Pajarinen J, Hirvensalo E, Paajanen H. Fatal surgical or procedure-related complications: a Finnish registry-based study. World J Surg 2014; 38(4): 759-64. Helkamaa T, Hirvensalo E, Huhtala H, Remes V. Patient injuries in primary total hip replacement. Acta Orthop 2016; 87(3): 209-17. Järvelin J, Häkkinen U. Can patient injury claims be utilised as a quality indicator? Health Policy 2012; 104:155-62. Järvelin J, Häkkinen U, Rosenqvist G, Remes V. Factors predisposing to claims and compensations for patient injuries following total hip and knee arthroplasty. Acta Orthop 2012; 83(2): 190-6. Kopf S, Forsyth T, Wong A K, Tashman S, Irrgang J J, Fu F H. Transtibial ACL reconstruction technique fails to position drill tunnels anatomically: in vivo 3D CT study. Knee Surg Sports Traumatol Arthrosc 2012; 20(11): 2200-7. Magnussen R A, Trojani C, Granan L P, Neyret P, Colombet P, Engebretsen L, Wright R W, Kaeding C C, MARS Group; SFA Revision ACL Group. Patient demographics and surgical characteristics in ACL revision: a comparison of French, Norwegian, and North American cohorts. Knee Surg Sports Traumatol Arthrosc 2015; 23(8): 2339-48. Maletis G B, Inacio M C, Reynolds S, Desmond J L, Maletis M M, Funahashi T T. Incidence of postoperative anterior cruciate ligament reconstruction infections: graft choice makes a difference. Am J Sports Med 2013; 41(8): 1780-5. Marchant B, Noyes F, Barber-Westin S, Fleckenstein C. Prevalence of nonanatomical graft placement in a series of failed anterior cruciate ligament reconstruction. Am J Sport Med 2010; 38:1571-7. Mattila V M, Sillanpää P, Iivonen T, Parkkari J, Kannus P, Pihlajamäki H. Coverage and accuracy of diagnosis of cruciate ligament injury in the Finnish National Hospital Discharge Register. Injury 2008; 39:1373-6. Mikkonen, M. Prevention of patient injuries: the Finnish patient insurance scheme. Med Law 2004; 23: 251-7. Morgan J A, Dahm D, Levy B, Stuart M J, the MARS Study Group. Femoral tunnel malposition in ACL revision reconstruction. J Knee Surg 2012; 25(5): 361-8. Okoroha K R, Keller R A, Jung E K, Khalil L, Marshall N, Kolowich P A, Moutzouros V. Pain assessment after anterior cruciate ligament reconstruction: bone–patellar tendon–bone versus hamstring tendon autograft. Orthop J Sports Med 2016; 4(12): 2325967116674924. Palmu S, Paukku R, Mäyränpää M K, Peltonen J, Nietosvaara Y. Injuries as a result of treatment of tibial fractures in children: claims for compensation submitted to the Patient Insurance Centre in Finland. Acta Orthop 2009; 80(1): 78-82. Palmu S, Paukku R, Peltonen J, Nietosvaara Y. Treatment injuries are rare in children’s femoral fractures. Acta Orthop 2010; 81(6): 715-8. Palonen R, Nio A, Mustajoki P. Potilas- ja lääkevahingot – korvaaminen ja ennaltaehkäisy [Patient and medication injuries – compensation and prevention]. Jyväskylä: Talentum Media Oy; 2005 [in Finnish]. Patient Injury Act 25.7.1986/585. Available at www.fi nlex.fi /fi /laki/ajantasa/1986/19860585.

601

Patient Insurance Centre. Potilasvakuutuskeskus n.d. Available at www.pvk. fi/en/ (English version). Pérez-Prieto D, Torre-Claramunt R, Gelber P E, Shehata T M A, Pelfort X, Monllau J C. Autograft soaking in vancomycin reduces the risk of infection after anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc 2016; 24: 2724-8. Persson A, Fjeldsgaard K, Gjertsen J E, Kjellsen A B, Engebretsen L, Hole R M, Fevang J M. Increased risk of revision with hamstring tendon grafts compared with patellar tendon grafts after anterior cruciate ligament reconstruction: a study of 12,643 patients from the Norwegian Cruciate Ligament Registry, 2004–2012. Am J Sports Med 2014; 42: 285-91. Phegan M, Grayson J E, Vertullo C J. No infections in 1,300 anterior cruciate ligament reconstructions with vancomycin pre-soaking of hamstring grafts. Knee Surg Sports Traumatol Arthrosc 2016; 24:2729-35. Pukk K, Lundberg J, Penaloza-Pesantes R V, Bromels M, Gaffney F A. Do women simply complain more? National patient injury claims data show gender and age differences. Qual Manag Health Care 2003; 12:225-31. Randsborg P H, Bukholm I R K, Jakobsen R B. Compensation after treatment for anterior cruciate ligament injuries: a review of compensation claims in Norway from 2005 to 2015. Knee Surg Sports Traumatol Arthrosc 2018; 26(2): 628-33. Robin B N, Lubowitz J H. Disadvantages and advantages of transtibial technique for creating the anterior cruciate ligament femoral socket. J Knee Surg 2014; 27(5): 327-30. Robin B N, Jani S S, Marvil S C, Reid J B, Schillhammer C K, Lubowitz J H. Advantages and disadvantages of tibial, anteromedial portal and outside-in femoral tunnel drilling in single-bundle anterior cruciate ligament reconstruction: a systemic review. Arthroscopy 2015; 31(7): 1412-17. Samitier G, Marcano A I, Alentorn-Geli E, Cugat R, Farmer K W, Moser M W. Failure of anterior cruciate ligament reconstruction. Arch Bone JT Surg 2015; 3(4): 220-4. Sandelin H, Waris E, Hirvensalo E, Vasenius J, Huhtala H, Raatikainen T, Helkamaa T. Patient injury claims involving fractures of the distal radius. Acta Orthop 2018; 89(2): 240-5. Saper M, Stephenson K, Heisey M. Arthroscopic irrigation and debridement in the treatment of septic arthritis after anterior cruciate ligament reconstruction. Arthroscopy 2014; 30(6): 747-54. Schulz A P, Götze S, Schmidt H G, Jürgens C, Faschingbauer M. Septic arthritis of the knee after anterior cruciate ligament surgery: a stage-adapted treatment regimen. Am J Sports Med 2007; 35(7): 1064-9. Sommer C, Friederich N F, Müller W. Improperly placed anterior cruciate ligament grafts: correlation between radiological parameters and clinical results. Knee Surg Sports Traumatol Arthrosc 2000; 8(4): 207-3. Sund R. Quality of the Finnish Hospital Discharge Register: a systematic review. Scand J Public Health 2012; 40:505-15. Taketomi S, Inui H, Yamagami R, Kawaguchi K, Nakazato K, Kono K, Kawata M, Nakagawa T, Tanaka S. Surgical timing of anterior cruciate ligament reconstruction to prevent associated meniscal and cartilage lesions. J Orthop Sci 2018; 23(3): 546-51. Vertullo C J, Quick M, Jones A, Grayson J E. Surgical technique using presoaked vancomycin hamstring grafts to decrease the risk of infection after anterior cruciate ligament reconstruction. Arthroscopy 2012; 28:337-42. Westermann R, Anthony C A, Duchman K R, Gao Y, Pugely A J, Hettrich C M, Amendola N, Wolf B R. Infection following anterior cruciate ligament reconstruction: an analysis of 6,389 cases. J Knee Surg 2017; 30(6): 535-543. Wright R W, Huston L J, Spindler K P. Descriptive epidemiology of the multicenter ACL revision study (MARS) cohort. Am J Sports Med 2010; 38(10): 1979-86.

602

Acta Orthopaedica 2019; 90 (6): 602–609

Less gap imbalance with restricted kinematic alignment than with mechanically aligned total knee arthroplasty: simulations on 3-D bone models created from CT-scans William BLAKENEY 1,2, Yann BEAULIEU 1, Marc-Olivier KISS 1,3, Charles RIVIÈRE 4, and Pascal-André VENDITTOLI 1,3 1 Department

of Surgery, CIUSSS-de-L’Est-de-L’Ile-de-Montréal, Hôpital Maisonneuve Rosemont, Montréal, Québec, Canada; 2 Department of Surgery, Albany Health Campus, Albany, Australia; 3 Department of Surgery, Université de Montréal, Montréal, Québec, Canada; 4 Adult Reconstruction and Joint Replacement, South West London Elective Orthopaedic Centre, MSK-Lab—Imperial College London, London, UK Correspondence: pa.vendittoli@me.com Submitted 2019-06-28. Accepted 2019-09-10

Background and purpose — Mechanical alignment techniques for total knee arthroplasty (TKA) introduce significant anatomic alteration and secondary ligament imbalances. We propose a restricted kinematic alignment (rKA) protocol to minimize these issues and improve TKA clinical outcomes. Patients and methods — rKA tibial and femoral bone resections were simulated on 1,000 knee CT scans from a database of patients undergoing TKA. rKA was defined by the following criteria: independent tibial and femoral cuts within 5° of the bone neutral mechanical axis, with a resulting HKA within 3° of neutral. Imbalances in the extension space, flexion space at 90°, medial compartment and lateral compartment were calculated and compared with measured resection mechanical alignment (MA) results. 2 MA techniques were simulated for rotation using the surgical transepicondylar axis (TEA) and 3° to the posterior condyles (PC). Results — Extension space imbalances ≥ 3 mm occurred in 33% of TKAs with MA technique versus 8.3% with rKA (p < 0.001). Similarly, more frequent flexion space imbalance ≥ 3mm was created by MA technique (TEA 34% or 3° PC 15%) versus rKA (6.4%, p < 0.001). Using MA with TEA or PC, there were only 49% and 63% of the knees respectively with < 3 mm of imbalance throughout the extension and flexion spaces and medial and lateral compartments versus 92% using rKA (p < 0.001). Interpretation — significantly fewer imbalances are created using rKA versus MA for TKA. rKA may be the best compromise, by helping the surgeon to preserve native knee ligament balance during TKA and avoid residual instability, whilst keeping the lower limb alignment within a safe range.

Human lower limb anatomy varies widely, and pathological changes increase this variability further (Almaawi et al. 2017, Hirschmann et al. 2019b, Moser et al. 2019). A standardized, systematic approach, using right-angled femoral and tibial bone cuts (Mechanical Alignment) with the concept of parallel and equal flexion and extension gaps, was introduced early in the development of TKA (Freeman et al. 1973, Scuderi et al. 2001). As very few individuals have neutral femoral and tibial mechanical axes (0.1% of a population of 4,884 patients scheduled for TKA), MA leads to important anatomic alterations for many subjects (Bellemans et al. 2012, Almaawi et al. 2017). This results in unequal bone resections with resultant imbalances (Blakeney et al. 2019a). Multiple ligament release techniques and algorithms have been proposed to re-balance the joint gaps. This resulted in many surgeons thinking of TKA as a soft-tissue surgery to balance the gap modification linked to these standardized bone-cut orientations (Whiteside 2002). There is, however, debate as to whether the knee’s collateral ligament laxities are modified in knees with less than 15° of deformity (McAuliffe et al. 2017, 2019). Soft-tissue releases are technically demanding, unpredictable, and can even introduce further imbalance (Kumar and Dorr 1997). Extensive releases may change the position of the joint line (Yoshii et al. 1991), which may have an adverse effect on the knee’s range of movement or the function of the extensor mechanism (Walker and Garg 1991) and worsen the clinical outcome (Martin and Whiteside 1990, Unitt et al. 2008). TKA joint gap imbalance has been associated with abnormal kinematics, decreased range of motion, condylar lift-off, loosening, wear and is a frequent cause of revision surgery, with rates varying from 21% to 35% (Wasielewski et al. 1994, Dennis et al. 2010, Gustke et al. 2014, Le et al. 2014).

Acta Orthopaedica 2019; 90 (6): 602–609

The restoration and preservation of pre-arthritic knee anatomy and ligament laxities during TKA has gained interest in recent years (Hirschmann et al. 2019a). The kinematic alignment (KA) technique represents a resurfacing of articular surfaces, removing equivalent amounts of bone and cartilage to match implant thickness (Howell et al. 2013). Concerns remain about restoring extreme anatomies, which may not be compatible with current TKA prostheses and fixation methods. Some knee anatomies may be inherently biomechanically inferior, or may have been altered by trauma, tumors, childhood deformity, or previous surgery. Keeping in mind these uncertainties, the senior author (PAV) developed a restricted KA (rKA) protocol (Hutt et al. 2016). rKA aims to perform KA bone resections for most cases, but performing adjustments for patients outside a “safe range” defined by the following criteria: independent tibial and femoral cuts must be within 5° of the mechanical axis of the respective bone and the overall resulting Hip–Knee–Ankle angle (HKA) must fall within 3° of neutral. Using a database of 4,884 CT scans of lower limbs for patients scheduled for TKA, a previous study demonstrated that in half of osteoarthritic knees there was no difference from standard KA resections with the rKA protocol, and mean anatomical corrections of 0.5° for medial proximal tibial angle (MPTA) and 0.3° for lateral distal femoral angle (LDFA) were needed to fit 4,062 cases (83%) (Almaawi et al. 2017). The objective of this study was to calculate bone resection thicknesses and resulting imbalances in the flexion/extension spaces and medial/lateral compartments, by simulating rKA protocol on 3-D bone models created from 1,000 CT scans of patients undergoing TKA and to compare the imbalances with previously published measured resection mechanical alignment (MA) results (Blakeney et al. 2019a). The study hypothesis was that the rKA protocol would reduce imbalance in the extension and flexion spaces and in the medial and lateral compartments versus MA.

Material and methods The data from this study were taken from a database of 1,000 consecutive lower limb CT scans, on patients scheduled for TKA using patient-specific instrumentation (PSI) with the MyKnee system (Medacta International, Switzerland). Mean HKA from the supine CT scan was 177° (SD 5.0, range 164– 194). There were 730 (73%) varus cases and 270 (27%) valgus cases. We then calculated a computed HKA, which was the sum of LDFA and MPTA. Tibial and femoral bone resections were simulated according to our rKA protocol. The “safe range’’ is defined by the following criteria: independent tibial and femoral cuts within 5° of the bone’s neutral mechanical axis and a resulting HKA within 3° of neutral. The algorithm was applied in 2 steps. For knee anatomy that fell outside the proposed safe range,

603

the LDFA and MPTA were corrected independently by setting them to closest value: ±5° from neutral. After the independent femoral or tibial corrections, if HKA remained > 3° of varus or valgus (aiming to maintain the femoral flexion axis as closely as possible) we adjusted the MPTA to bring the HKA within the safe range of ±3°. The distal femoral and proximal tibial cut resections were set at 8 mm from the distal femoral condyles and 8 mm from the proximal tibial plateaus. If corrections to the MPTA or LDFA were required per above protocol, 1 resection was maintained at 8 mm and the other reduced accordingly. An 8 mm resection thickness was based on an implant size of 10 mm (bone +2 mm of cartilage) (Li et al. 2005). Equal medial and lateral posterior femoral resections of 8 mm thickness were simulated on all scans (no femoral rotation). After simulation of the bone cuts, the gap sizes were calculated as the sum of the femoral and tibial bone resections. Using a CT scan without the cartilage, the target bone resection was 16 mm (+ 2 x 2 mm for cartilage thicknesses corresponds to a total implant thickness of 20 mm). 4 gaps were measured: medial and lateral gaps in both extension and 90° of flexion. An “imbalance” was defined as the difference between 2 gaps. These imbalances are created when the protocol modifies the KA resection to stay within the safe range. A clinically important imbalance was considered to be 3 mm or greater. 4 imbalances were calculated: • extension space: medial gap in extension—the lateral gap in extension; • flexion space: medial gap in flexion—the lateral gap in flexion; • medial compartment: medial gap in extension—medial gap in flexion; • lateral compartment: lateral gap in extension—lateral gap in flexion. The mean was also calculated based on absolute values. We compared these results with previously reported results for MA technique using the same database of patients (Blakeney et al. 2019a). MA femoral rotation was assessed with 2 techniques: femur aligned with the surgical transepicondylar axis (TEA) or aligned with 3° of external rotation to the posterior condyles (PC). A resection plane, aligned with the posterior condyles (8 mm thickness medially and laterally) was rotated to the appropriate angle (TEA or PC) using a central pivot. Statistics Descriptive statistics were calculated to summarize patient anatomy and resection measures. To compare continuous variables between rKA and MA techniques, 2-sample t-tests for independent groups were used. Paired t-tests were used to compare continuous variables between PC and TEA techniques. All tests were 2-tailed, with a significance level of p < 0.001 (to allow for multiple comparisons). Chi-squared or McNemar tests were used to compare categorical data.

604

Acta Orthopaedica 2019; 90 (6): 602–609

Table 1. Lower limb alignment of pre-operative anatomy compared with after rKA. Values are mean (SD) [range] degrees.

Preoperative anatomy

After rKA

p-value

rKA angle modification whole cohort cases modified

HKA angle 180 (3.6) [168 to 191] 180 (2.2) [177 to 183] 0.6 0.0 (1.8) [–8.3 to 8.8] absolute values 1.0 (1.5) [0.0 to 8.8] LDFA –2.8 (2.4) [–9.8 to 5.8) –2.6 (2.1) [–5.0 to 5.0] < 0.001 0.2 (0.6) [–0.8 to 4.8] absolute values 0.2 (0.6) [0.0 to 4.8] MPTA 2.9 (2.6) [–8.9 to 9.9] 2.7 (1.7) [–5.0 to 5.0] < 0.001 –0.2 (1.6) [–8.3 to 8.8] absolute values 0.8 (1.4) [0.0 to 8.8] TEA angle 5.2 (1.8) [0.3 to 9.7)

–0.1 (2.5) [–8.3 to 8.8] 1.9 (1.7) [0.0 to 8.8) 0.4 (0.8) [–0.8 to 4.8] 0.4 (0.8) [0.0 to 4.8] –0.4 (2.2) [–8.3 to 8.8) 1.6 (1.7) [0.0 to 8.8]

HKA angle: hip-knee-ankle angle (computed as LDFA + MPTA). LDFA: lateral distal femoral angle. MPTA: medial proximal tibial angle. TEA angle: degrees of external rotation of the transepicondylar axis to the posterior condyles. rKA angle modification: rKA minus native anatomy.

Preoperative LDFA (°)

Postoperative LDFA (°)

–2

–4

–6

–8 –10

Extension space (mm) < 12 12–13 14–15 16 p-value

–8 10

–10 –2 –4 –6 –8 –10 10

Preoperative MPTA (°)

Table 2. Distribution of medial and lateral gap sizes in the extension space for MA and rKA techniques. Values are percentages

–2 –4 –6 –8 –10

Postoperative MPTA (°)

Figure 1. LDFA and MPTA comparing preoperative and postoperative distributions.

Ethics, funding, and potential conflicts of interest This article used anonymous data from an existing collection of CT scans and does not contain any studies with human participants performed by any of the authors. Informed consent for this type of study is not required. Funding was received from OMeGA Medical Grants Association fellowship support. The authors have no potential conflict of interest.

Results Lower limb alignment Table 1 presents the preoperative lower limb alignment and the resulting effects of the rKA protocol. With computed HKA (LDFA + MPTA), there were 521 (52%) varus and 479 (48%) valgus cases preoperatively versus 505 (51%) and 495 (49%) after rKA. Although there was no significant mean difference in HKA after rKA, rKA significantly modified the LDFA and MPTA compared with preoperative values (p < 0.001). With rKA, LDFA and MPTA were independently modified for 18% and 45% of cases respectively. The femoral valgus and tibial

Medial extension gap MA rKA

Lateral extension gap MA rKA

17 1.5 33 5.0 35 15 15 79 < 0.001

17 1.1 35 5.0 32 14 15 80 < 0.001

The gap size in extension is the sum of the distal femoral bone resection and tibial bone resection. Note: The aim is for a resection of 16 mm.

varus were reduced by a mean of 0.4° for both (absolute modification for femur 0.4° and tibia 1.6°). Modifications of both the LDFA and MPTA were needed in 10% of cases. Figure 1 shows the native LDFA and MPTA versus the resulting cut orientations after rKA protocol application. Extension space With rKA, the created gaps in the medial and lateral compartments were maintained within 2 mm (14–16 mm) for 94% of cases, compared with 50% and 48% with MA (McNemar test p < 0.001 and p < 0.001). In other cases, the gaps were reduced (Tables 2 and 3). The mean extension space imbalance was 0.8 mm with rKA and 2.4 mm for MA (p < 0.001, Table 3). There were significantly fewer cases having imbalance ≥ 3mm with rKA (8.3%) vs. MA (33%), and ≥ 5mm with rKA (1.5%) vs. MA (11%) (p < 0.001, Figure 2). Flexion space at 90° Mean created gaps were reduced significantly more with MA PC vs. rKA (medial and lateral, p < 0.001) and with MA TEA only in the lateral compartment (p < 0.001, Table 4). Using

Acta Orthopaedica 2019; 90 (6): 602–609

605

Table 3. Medial and lateral gaps modification in the extension space and resulting medio-lateral difference in mm for MA and rKA techniques. Values are mean (SD) [range]

Medial gap Lateral gap DML absolute values

–2.7 (1.9) [–8.9 to 0.0] –2.7 (1.9) [–9.5 to 0.0] 0.0 (3.0) [–9.5 to 8.9] 2.4 (1.9) [0.0 to 9.5]

rKA p-value –0.4 (1.0) [–6.5 to 0.0] < 0.001 –0.4 (1.0) [–7.1 to 0.0] < 0.001 0.0 (1.5) [–7.1 to 6.5] 0.7 0.8 (1.3) [0.0 to 7.1] < 0.001

The gap size modification is the sum of the distal femoral bone resection and tibial bone resection minus 16 mm (resection goal). DML: lateral gap minus medial gap; a negative value in the row represents a greater medial space than lateral space, whereas a positive value represents a greater lateral than medial space.

Distribution (%)

100

100 rKA MA

0–3

3–5

rKA MA – PC3° MA – TEA

Medio-lateral gap imbalance in the extension space (mm)

0–3

3–5

Medio-lateral gap imbalance in the flexion space (mm)

Figure 2. Distribution of medio-lateral gap imbalance in the extension space for rKA and MA techniques (p < 0.001).

Figure 3. Distribution of medio-lateral gap imbalance in the flexion space for rKA and MA with PC 3° (p < 0.001) or TEA (p < 0.001) techniques.

rKA, mean flexion space imbalance was 0.7 mm versus 1.6 mm for MA PC (p = 0.001) and 2.6 mm for MA TEA (p < 0.001). There were significantly fewer cases having imbalance ≥ 3mm with rKA (6.4%) vs. MA PC (15%) and MA TEA (34%), and imbalances ≥ 5 mm for rKA (1.1%) vs. MA PC (2.5%) or MA TEA (13%) (p < 0.001, Figure 3). Medial and lateral compartment imbalances With rKA, mean medial compartment imbalance was 0 mm vs. 1.4 mm with MA PC (p = 0.001) and 2.4 mm with MA TEA (p < 0.001) (Table 5). Mean lateral compartment imbalance was 0.2 mm with rKA vs. 1.8 mm with MA PC (p < 0.001) and 1.6 mm with MA TEA (p = 0.001). In 4.4% of rKA vs. 16% of MA PC and 33% MA TEA, there was a mismatch between flexion and extension gaps, with an extension gap too small and a flexion gap too large or vice versa (Table 6). This means, for example, that releasing a tight extension gap may increase an already loose flexion gap. Combined imbalances With rKA, the percentage of knees with space imbalances < 3mm in both extension and flexion was 92% vs. 63% with MA PC (p < 0.001) and 49% with MA TEA (p < 0.001) (Table 7). NB: Data analyzed for varus and valgus native HKA separately can be found in the Supplementary data.

Table 4. Medial and lateral gaps modification in the flexion space and resulting medio-lateral difference in mm for MA PC method, MA TEA method, and rKA techniques. Values are mean (SD) [range] Medial gap Lateral gap DML absolute values

MA PC method –1.3 (1.8) [–6.5 to 1.7] –1.4 (0.6) [–7.3 to –0.9] –0.1 (2.1) [–8.4 to 5.4] 1.6 (1.3) [0.0 to 8.4]

MA TEA method –0.4 (1.9) [–6.8 to 4.3] –2.4 (1.0) [–7.2 to –0.1] –2.0 (2.6) [–10 to 5.7] 2.6 (2.0) [0.0 to 10]

rKA –0.4 (1.0) [–6.5 to 0.0] –0.2 (0.8) [–6.0 to 0.0] 0.2 (1.4) [–6.1 to 6.5] 0.7 (1.2) [0.0 to 6.5]

p-value: rKA vs MA PC MA TEA < 0.001 < 0.001 < 0.001 < 0.001

0.07 < 0.001 < 0.001 < 0.001

DML, see Table 3. Table 5. Flexion–extension gap differences (DFE) in mm for the medial and lateral compartments for MA PC method, MA TEA method, and rKA techniques. Values are mean (SD) [range] Medial DFE absolute values Lateral DFE absolute values

MA PC method –1.4 (0.6) [0.9 to 6.6] 1.4 (0.6) [–6.6 to –0.9] –1.3 (1.8) [–6.5 to 1.7] 1.8 (1.3) [0.0 to 6.5]

MA TEA method –2.4 (1.0) [–8.2 to –0.3] 2.4 (1.0) [–8.2 to –0.3] –0.3 (1.5) [–6.7 to 4.4] 1.6 (1.1) [0.0 to 6.7]

rKA 0.0 (0.0) [–0.6 to 0.0] 0.0 (0.0) [0.0 to 0.6] –0.2 (0.5) [–3.7 to 0.0] 0.2 (0.5) [0.0 to 3.7]

p-value: rKA vs MA PC MA TEA < 0.001 < 0.001 < 0.001 < 0.001

<0 .001 < 0.001 0.005 < 0.001

DFE: extension gap minus flexion gap; a negative value represents a greater flexion space than extension space, whereas a positive value represents a larger extension than flexion space.

606

Acta Orthopaedica 2019; 90 (6): 602–609

Table 6. Percentage of knees with medial or lateral flexion-extension gap mismatch for MA PC method, MA TEA method, and rKA techniques

Medial compartment MA PC MA TEA rKA

Ext. gap < 15 mm and flex. gap ≥ 16 mm Ext. gap ≥ 16 mm and flex. gap < 15 mm Total

5.2 0 5.2

23 0 23

0 0 0

Table 7. Percentage of knees where the medio-lateral gap mismatch is present in both the extension and flexion spaces for MA PC method, MA TEA method and rKA techniques Gap mismatch ≤ 3 mm ≤ 5 mm > 5 mm

MA PC MA TEA 63 89 1.9

49 81 3.8

rKA

p-value: rKA vs MA PC MA TEA

92 99 1.1

< 0.001 < 0.001 0.1

< 0.001 < 0.001 < 0.001

Discussion This study demonstrated that rKA produced less imbalance than an MA technique for TKA. rKA significantly reduced the cases with imbalance ≥ 3mm created by MA technique in the extension and flexion spaces, and in the medial and lateral compartments. The extension space is created by the distal femoral and proximal tibial cut orientations and resection levels. The MA technique, using the most prominent joint surface as a reference for resection thickness (mostly the medial femoral condyle and lateral tibial plateau), will intrinsically tend to reduce the extension medial gap in varus knees and lateral gap in valgus knees (Blakeney et al. 2019a). In contrast, the KA technique aims to restore the pre-arthritic joint surface orientations by removing corresponding bone thickness to the implant thickness, thus re-creating native joint gaps. Using the rKA protocol, gaps will be modified in cases where the patient’s anatomy falls outside our safe range, requiring adjustments to be performed. In this study, rKA maintained the extension gaps within 14–16 mm (16 mm meaning no gap modification) for 94% on the medial compartment vs. 50% with MA (Table 2). On the lateral side, it was 94% with rKA vs. 48% with MA. Extension space balance was also significantly improved with rKA, where only 8.3% had an imbalance ≥ 3 mm vs. 33% with MA (Figure 2). This means that frequency and magnitude of soft-tissue release to balance the extension space would be significantly reduced with rKA TKA compared with MA. The flexion space is created by the tibial and posterior femoral cut orientations and resection levels. To obtain a balanced flexion space with MA, femoral external rotation should match the tibial cut orientation. With MA, using a 90° cut on

p-value: rKA vs MA PC MA TEA < 0.001 < 0.001 N/A N/A < 0.001 < 0.001

Lateral compartment MA PC MA TEA rKA 0 10 10

0 9.8 9.8

4.4 0 4.4

p-value: rKA vs MA PC MA TEA < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001

the tibial side reduces its anatomic varus by a mean of 3° (Bellemans et al. 2012). This is why using MA with PC 3° was shown to result in fewer flexion space imbalances versus TEA where the mean external rotation was 5° (Blakeney et al. 2019a). Since valgus knees frequently have tibial mechanical axes near neutral or in valgus (Alghamdi et al. 2014, Almaawi et al. 2017), increasing femoral external rotation resulted in even greater imbalance. With KA, bone resection equivalent to the implant thickness will be removed from the proximal tibia and the posterior condyles (neutral femoral rotation parallel to PC), thus maintaining joint flexion gaps. With our rKA protocol, gaps will be modified in cases where the patient’s MPTA fell outside the safe range of ±5º. We found that rKA significantly reduced flexion space imbalances in comparison with MA. In cases where an MPTA tibial adjustment is needed (generally reducing the varus), one could ask if we should apply some femoral external rotation accordingly to balance the flexion space. The senior author abandoned this practice early on with rKA, trying to favor maintenance of a femoral component aligned as closely as possible with the femoral flexion (cylindrical) axis (Eckhoff et al. 2007). In addition to medio-lateral space equilibration, flexion– extension gap symmetry is considered to be a goal in TKA. Our results show that the rKA technique creates significantly fewer medial and lateral compartment imbalances versus MA (Table 5). Moreover, using MA, there were a high number of cases with an overly tight flexion gap with an overly loose extension gap, or vice versa (Table 6). Attempting to correct the tightness of these knees in one position is likely to worsen the laxity in the other position. When TKA was first introduced, instrument precision was poor and implantation errors were frequent. There were many pitfalls to overcome, hence the focus was on implant survivorship, rather than reproducing normal knee function (Vendittoli and Blakeney 2017). To simplify and standardize, surgeons introduced the MA technique. This systematic, “one size fits all” approach does not respect the wide range of normal anatomy of the knee (Almaawi et al. 2017). Many studies have illustrated the detrimental effects of soft tissue imbalance on function and long-term survival (Daines and Dennis 2014, Le et al. 2014, Sharkey et al. 2014). It is not well established as to what constitutes the limits of a knee that is considered balanceable with soft-tissue release. Soft tissue balancing is further complicated when attempting to balance both a medial/

Acta Orthopaedica 2019; 90 (6): 602–609

Figure 4. Lower limb long radiographs showing a case with an LDFA of 11° and MPTA of 6°. Reproducing her lower limb alignment with KA technique (unrestricted) would leave her lower limb HKA in 5° of valgus. With rKA, correcting the femur to 5° and the tibia to 2° of varus would results in an HKA of 3° valgus.

lateral compartment and flexion/extension space imbalance, with some releases unpredictably affecting both imbalances. With a better understanding of normal knee anatomy and function, KA technique has been introduced to improve clinical results following TKA. KA aims to restore the prearthritic patient’s constitutional lower limb alignment and joint surface orientations. It is a joint resurfacing procedure with only exceptional soft tissues release (Howell et al. 2010, Riviere et al. 2017). KA TKA short-term clinical scores and functional evaluation are favorable (Courtney and Lee 2017, Niki et al. 2018, Takahashi et al. 2018, Blakeney et al. 2019b). The implant survivorship at 10 years of a cohort of 220 TKAs by Howell et al. (2018) is very promising. It should be considered, however, that knee anatomy varies widely (Almaawi et al. 2017). Many believe that we should not blindly repro-

607

duce all anatomies when performing KA TKA, as some may have deleterious effects on TKA mechanics and clinical outcomes. These extreme anatomies may be inherently mechanically inferior and considered pathological. A strong argument for the existence of patho-anatomies is the unilateral occurrence in some patients. On the other hand, creating a neutral mechanical axis in these patients with MA TKA would generate substantial alteration of the native knee anatomy with subsequent extensive soft-tissue release, severely modifying the physiological joint line orientation and knee mechanics. To address these concerns, rKA has been developed as an alternative solution to the unrestricted KA technique (Hutt et al. 2016, Almaawi et al. 2017) for situations when patients have atypical knee anatomy (Figure 4). There are still few short- and mid-term follow-up studies on KA TKAs (Howell et al. 2018), whereas MA TKAs have a long history of good survivorship (Font-Rodriguez et al. 1997, Gill et al. 1999, Rodricks et al. 2007). Moreover, the current KA studies include only a limited number of outlier anatomy cases. The rKA is a sound compromise; it reproduces the patient’s constitutional knee anatomy when within a safe range for 50% of cases, requires minor modifications for the rest of the cases, and brings back the extreme anatomies towards acceptable values, modifying their deformities to allow an implant orientation compatible with current materials and fixation methods (Almaawi et al. 2017). As shown in the current study, MPTA and LDFA were modified in the outlier cases by a mean of 0.4° and 1.6° respectively (Table 1). By adhering to the rKA boundaries, it is possible to reliably produce a prosthetic knee with component/knee/limb alignments that always fall within an evidence-based safe alignment range. In a simulation study including 4,884 knees, 17% of knees had very unusual anatomy, with both the femur and tibia articular orientations being in varus or valgus (Almaawi et al. 2017). As both bones contribute the same direction to the overall HKA deviation, using rKA the surgeon needs to decide which bone to correct to fall within the safe range. We believe that the femoral flexion axis plays the more important role in knee kinematics, therefore our practice is to preserve femoral anatomy as closely as possible and perform greater modifications on the tibial side. In our experience, ligamentous releases are usually not needed in cases with anatomic modifications of < 3°. In larger corrections, minimal releases can be added (usually, to a much lesser degree compared with MA) (Hutt et al. 2016). In addition, a study of gait analysis comparing patients operated on with rKA compared with MA technique demonstrated that the rKA patients had knee kinematics that were closer to healthy controls than MA patients (Blakeney et al. 2019b). These kinematic differences translated into a higher postoperative mean KOOS score in the KA group compared with the MA group (74 vs. 61, p = 0.03). This study has some limitations. The database did not provide demographic data or preoperative diagnosis. We do not know whether any extra-articular deformity was contributing

608

to the alignment. Our results represent only the gaps created by bone resections and do not include additional imbalances linked to physiologic and/or pathologic soft tissue laxity or contracture, or bone loss due to wear. These would impact the final gaps but could not be determined using our method. It is also arguable at what limit a space or compartment imbalance becomes “unbalanceable.” We limited our comparison of the rKA with the MA techniques and did not test the gap balance technique. Finally, we found significant differences for most statistical analyses presented. Our large data set implies a very high analysis power. On the other hand, it is difficult to determine the clinical significance of all measured differences. In the absence of further evidence from long-term studies of KA TKAs, some authors have cautioned against widespread adoption of the KA technique (Abdel et al. 2014). We believe the rKA protocol offers a satisfactory compromise, allowing re-creation of normal patient anatomy for the majority of cases, avoiding the excessive corrections and ligamentous releases required with MA, but preventing the extremes of implant positioning that a universal KA technique application may produce (Rivière et al. 2019). Supplementary data Data analyzed for varus and valgus native HKA separately can be found in the Supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2019. 1675126

WB was involved in data analysis, drafting and reviewing of the manuscript. YB was involved in data collection and analysis. CR and MK were involved with critical appraisal of the manuscript. PAV was the study designer and coordinator and helped with critical appraisal of the manuscript. The authors thank Gianluca Gabellini and Medacta, Lugano, Switzerland, for their help with providing the data and statistical analysis. Abdel M P, Oussedik S, Parratte S, Lustig S, Haddad F S. Coronal alignment in total knee replacement: historical review, contemporary analysis, and future direction. Bone Joint J 2014; 96-B(7): 857-62. doi: 10.1302/0301620X.96B7.33946. Alghamdi A, Rahme M, Lavigne M, Masse V, Vendittoli P A. Tibia valga morphology in osteoarthritic knees: importance of preoperative full limb radiographs in total knee arthroplasty. J Arthroplasty 2014; 29(8): 1671-6. doi: 10.1016/j.arth.2014.03.001. Almaawi A M, Hutt J R B, Masse V, Lavigne M, Vendittoli P-A. The impact of mechanical and restricted kinematic alignment on knee anatomy in total knee arthroplasty. J Arthroplasty 2017; 32(7): 2133-40. doi: 10.1016/j. arth.2017.02.028. Bellemans J, Colyn W, Vandenneucker H, Victor J. The Chitranjan Ranawat award: is neutral mechanical alignment normal for all patients? The concept of constitutional varus. Clin Orthop Relat Res 2012; 470(1): 45-53. doi: 10.1007/s11999-011-1936-5. Blakeney W, Beaulieu Y, Puliero B, Kiss M O, Vendittoli P A. Bone resection for mechanically aligned total knee arthroplasty creates frequent gap modifications and imbalances. Knee Surg Sports Traumatol Arthrosc 2019a. [Epub ahead of print] doi: 10.1007/s00167-019-05562-8.

Acta Orthopaedica 2019; 90 (6): 602–609

Blakeney W, Clement J, Desmeules F, Hagemeister N, Riviere C, Vendittoli P A. Kinematic alignment in total knee arthroplasty better reproduces normal gait than mechanical alignment. Knee Surg Sports Traumatol Arthrosc 2019b; 27(5): 1410-17. doi: 10.1007/s00167-018-5174-1. Courtney P M, Lee G C. Early outcomes of kinematic alignment in primary total knee arthroplasty: a meta-analysis of the literature. J Arthroplasty 2017; 32(6): 2028-32 e1. doi: 10.1016/j.arth.2017.02.041. Daines B K, Dennis D A. Gap balancing vs. measured resection technique in total knee arthroplasty. Clin Orthop Surg 2014; 6(1): 1-8. doi: 10.4055/ cios.2014.6.1.1. Dennis D A, Komistek R D, Kim R H, Sharma A. Gap balancing versus measured resection technique for total knee arthroplasty. Clin Orthop Relat Res 2010; 468(1): 102-7. doi: 10.1007/s11999-009-1112-3. Eckhoff D, Hogan C, DiMatteo L, Robinson M, Bach J. Difference between the epicondylar and cylindrical axis of the knee. Clin Orthop Relat Res 2007; (461): 238-44. doi: 10.1097/BLO.0b013e318112416b. Font-Rodriguez D E, Scuderi G R, Insall J N. Survivorship of cemented total knee arthroplasty. Clin Orthop Relat Res 1997; (345): 79-86. Freeman M A, Swanson S A, Todd R C. Total replacement of the knee using the Freeman-Swanson knee prosthesis. Clin Orthop Relat Res 1973; (94): 153-70. Gill G S, Joshi A B, Mills D M. Total condylar knee arthroplasty 16- to 21-year results. Clin Orthop Relat Res 1999; (367): 210-5. Gustke K A, Golladay G J, Roche M W, Elson L C, Anderson C R. A new method for defining balance: promising short-term clinical outcomes of sensor-guided TKA. J Arthroplasty 2014; 29(5): 955-60. doi: 10.1016/j. arth.2013.10.020. Hirschmann M T, Becker R, Tandogan R, Vendittoli P A, Howell S. Alignment in TKA: what has been clear is not anymore! Knee Surg Sports Traumatol Arthrosc 2019a; 27(7): 2037-9. doi: 10.1007/s00167-019-05558-4. Hirschmann M T, Moser L B, Amsler F, Behrend H, Leclercq V, Hess S. Phenotyping the knee in young non-osteoarthritic knees shows a wide distribution of femoral and tibial coronal alignment. Knee Surg Sports Traumatol Arthrosc 2019b; 27(5): 1385-93. doi: 10.1007/s00167-019-05508-0. Howell S M, Howell S J, Hull M L. Assessment of the radii of the medial and lateral femoral condyles in varus and valgus knees with osteoarthritis. J Bone Joint Surg Am 2010; 92(1): 98-104. doi: 10.2106/JBJS.H.01566. Howell S M, Papadopoulos S, Kuznik K T, Hull M L. Accurate alignment and high function after kinematically aligned TKA performed with generic instruments. Knee Surg Sports Traumatol Arthrosc 2013; 21(10): 2271-80. doi: 10.1007/s00167-013-2621-x. Howell S M, Shelton T J, Hull M L. Implant survival and function ten years after kinematically aligned total knee arthroplasty. J Arthroplasty 2018; 33(12): 3678-84. doi: 10.1016/j.arth.2018.07.020. Hutt J R, LeBlanc M A, Masse V, Lavigne M, Vendittoli P A. Kinematic TKA using navigation: surgical technique and initial results. Orthop Traumatol Surg Res 2016; 102(1): 99-104. doi: 10.1016/j.otsr.2015.11.010. Kumar P J, Dorr L D. Severe malalignment and soft-tissue imbalance in total knee arthroplasty. Am J Knee Surg 1997; 10(1): 36-41. Le D H, Goodman S B, Maloney W J, Huddleston J I. Current modes of failure in TKA: infection, instability, and stiffness predominate. Clin Orthop Relat Res 2014; 472(7): 2197-200. doi: 10.1007/s11999-014-3540-y. Li G, Park S E, DeFrate L E, Schutzer M E, Ji L, Gill T J, Rubash H E. The cartilage thickness distribution in the tibiofemoral joint and its correlation with cartilage-to-cartilage contact. Clin Biomech (Bristol, Avon) 2005; 20(7): 736-44. doi: 10.1016/j.clinbiomech.2005.04.001. Martin J W, Whiteside L A. The influence of joint line position on knee stability after condylar knee arthroplasty. Clin Orthop Relat Res 1990; (259): 146-56. McAuliffe M J, Roe J, Garg G, Whitehouse S L, Crawford R. The varus osteoarthritic knee has no coronal contractures in 90 degrees of flexion. J Knee Surg 2017; 30(4): 297-303. doi: 10.1055/s-0036-1584539. McAuliffe M J, Garg G, Orschulok T, Roe J, Whitehouse S L, Crawford R. Coronal plane laxity of valgus osteoarthritic knee. J Orthop Surg (Hong Kong) 2019; 27(1): 2309499019833058. doi: 10.1177/2309499019833058.

Acta Orthopaedica 2019; 90 (6): 602â&#x20AC;&#x201C;609

Moser L B, Hess S, Amsler F, Behrend H, Hirschmann M T. Native nonosteoarthritic knees have a highly variable coronal alignment: a systematic review. Knee Surg Sports Traumatol Arthrosc 2019; 27(5): 1359-67. doi: 10.1007/s00167-019-05417-2. Niki Y, Nagura T, Nagai K, Kobayashi S, Harato K. Kinematically aligned total knee arthroplasty reduces knee adduction moment more than mechanically aligned total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2018; 26(6): 1629-35. doi: 10.1007/s00167-0174788-z. Riviere C, Iranpour F, Auvinet E, Howell S, Vendittoli P A, Cobb J, Parratte S. Alignment options for total knee arthroplasty: a systematic review. Orthop Traumatol Surg Res 2017; 103(7): 1047-56. doi: 10.1016/j. otsr.2017.07.010. RiviĂ¨re C, Vigdorchik J M, Vendittoli P A. Mechanical alignment: the end of an era! Orthop Traumatol Surg Res 2019; Aug 1. pii: S18770568(19)30213-0. Rodricks D J, Patil S, Pulido P, Colwell C W. Press-fit condylar design total knee arthroplasty: fourteen to seventeen-year follow-up. J Bone Joint Surg Am 2007; 89(1): 89-95. doi: 10.2106/JBJS.E.00492. Scuderi G R, Scott W N, Tchejeyan G H. The Insall legacy in total knee arthroplasty. Clin Orthop Relat Res 2001; (392): 3-14 Sharkey P F, Lichstein P M, Shen C, Tokarski A T, Parvizi J. Why are total

609

knee arthroplasties failing today: has anything changed after 10 years? J Arthroplasty 2014; 29(9): 1774-8. doi: 10.1016/j.arth.2013.07.024. Takahashi T, Ansari J, Pandit H G. Kinematically aligned total knee arthroplasty or mechanically aligned total knee arthroplasty. J Knee Surg 2018; 31(10): 999-1006. doi: 10.1055/s-0038-1632378. Unitt L, Sambatakakis A, Johnstone D, Briggs T W, Balancer Study G. Short-term outcome in total knee replacement after soft-tissue release and balancing. J Bone Joint Surg Br 2008; 90(2): 159-65. doi: 10.1302/0301620X.90B2.19327. Vendittoli P A, Blakeney W. Redefining knee replacement. Orthop Traumatol Surg Res 2017; 103(7): 977-9. doi: 10.1016/j.otsr.2017.09.003. Walker P S, Garg A. Range of motion in total knee arthroplasty: a computer analysis. Clin Orthop Relat Res 1991; (262): 227-35. Wasielewski R C, Galante J O, Leighty R M, Natarajan R N, Rosenberg A G. Wear patterns on retrieved polyethylene tibial inserts and their relationship to technical considerations during total knee arthroplasty. Clin Orthop Relat Res 1994; (299): 31-43. Whiteside L A. Soft tissue balancing: the knee. J Arthroplasty 2002; 17(4 Suppl. 1): 23-7. Yoshii I, Whiteside L A, White S E, Milliano M T. Influence of prosthetic joint line position on knee kinematics and patellar position. J Arthroplasty 1991; 6(2): 169-77.

610

Acta Orthopaedica 2019; 90 (6): 610–613

Routine radiographic follow-up is not necessary after physeal fractures of the distal tibia in children Antti STENROOS 1, Jussi KOSOLA 1, Jani PUHAKKA 1, Topi LAAKSONEN 2, Matti AHONEN 2, and Yrjänä NIETOSVAARA 2 1 Department

of Orthopedics and Traumatology, Töölö Hospital, Helsinki University Hospital; 2 Department of Pediatric Orthopedics and Traumatology, Helsinki New Children’s Hospital; Finland Correspondence: antti.stenroos@helsinki.fi Submitted 2019-02-09. Accepted 2019-05-02.

Background and purpose — Unnecessary radiographic and clinical follow-ups are common in treatment of pediatric fractures. We hypothesized that follow-up radiographs are unnecessary to monitor union of physeal fractures of the distal tibia. Patients and methods — All 224 (147 boys) children under 16 years old treated for a physeal fracture of the distal tibia during a 5-year period (2010–14) in Helsinki Children’s Hospital were included in this study. Peterson type II fractures comprised 55% and transitional fractures (Tillaux and Triplane) 20% of all injuries. Fracture displacement and alignment was measured. Type and place of treatment was recorded. Number of follow-up radiographs and outpatient visits was calculated and their clinical significance was assessed. Results — 109 children had fractures with < 2 mm displacement and no angulation. The other 115 children’s mean fracture displacement was 6 mm (2–28). 54% of all children were treated by casting in situ in the emergency room, 20% with manipulation under anesthesia and 26% with surgery (internal 57, external fixation 2). Median 3 (1–7) follow-up appointments and median 3 (0–6) radiographs were taken. Follow-up radiographs at or before cast removal did not alter treatment in any of the patients. 223 patients’ fractures healed within 4–9 weeks in good alignment (≤ 5° angulation). Interpretation — Routine radiographic follow-up is unnecessary to monitor alignment and union of physeal fractures of the distal tibia.

Physeal fractures of the distal tibia represent around 5% of all fractures and 15–20% of physeal fractures in children (Peterson and Peterson 1972, Landin 1983). Appropriate treatment depends on fracture type and displacement, as well as on the age of the child (Cummings 2001, Leary et al. 2009). The aim of treatment is fracture union in good alignment without iatrogenic growth plate damage. Clinical and radiographic follow-up has been traditionally scheduled to monitor healing of physeal fractures of the distal tibia. The purpose of early clinical follow-up is to check that the cast is appropriate and that operative wounds are healing uneventfully. Radiographs have been taken to register fracture alignment and union, position and integrity of hardware, as well as development of a physeal bar or deformity in patients with growth plate damage. The clinical significance of these outpatient appointments and radiographic follow-up is, however, unclear, which has led Perry et al. (2018) to suggest that distal tibia fracture treatment protocols for children are among top clinical effectiveness research questions to date. Unnecessary outpatient visits and radiographs are a substantial economic burden and consume hospitals and families’ resources. We evaluated whether routine follow-up visits in the outpatient clinic before cast removal and radiographs to monitor healing of physeal fractures of distal tibia are necessary.

Patients and methods All children under 16 years old treated for a physeal fracture of the distal tibia during a 5-year period between 2010 and 2014 in the Children’s Hospital, Helsinki University Central Hospital were included in this study (n = 224, 147 boys). Patient

Acta Orthopaedica 2019; 90 (6): 610–613

611

III

Figure 1. The Peterson fracture classification. Etiology

The patients were divided into 3 groups based on their treatment: casting in situ in emergency department (ER), manipulation under anesthesia and casting (MUA), or surgical treatment in operation room (OR). The number of follow-up radiographs and outpatient visits were calculated and their clinical significance was assessed. Cast complications, infections, and reoperations were registered. The targeting, rate, and length of follow-up after fracture union to detect possible growth arrest was evaluated.

Skateboarding Traffic accidents Ball sports Ice hockey Alpine sports Gymnastics Playground 0

Number of patients

Figure 2. Etiology of distal tibial epiphyseal fractures.

identification was done from the hospital’s fracture register (KIDS fracture register). Less than 16-year-old residents of Helsinki are either transported directly or referred for treatment in Helsinki Children’s Hospital with very few exceptions. All patients admitted to the Children’s Hospital for fracture treatment are automatically registered in Kid’s Fracture Tool (New Children’s Hospital Helsinki and BCB Medical) in the emergency department. Likewise patients’ data from the operative theater, orthopedic wards, and outpatient clinic are included in the registry. Fracture type (Figure 1) was registered using Peterson classification for patients with open growth plates (Peterson and and Peterson 1972). Transitional fractures (growth plates partially closed at the time of fracture) were divided into Tillaux and Triplane fractures. Etiology of the fractures is presented in Figure 2. Fracture alignment was assessed from the digital radiographs taken at admission, after reduction, and during followups. Displacement was measured from radiographs as the greatest amount of displacement (mm) in any view. The angle between the shaft and the epiphysis was registered as varus or valgus (AP view), ante- or retrocurvatum (lateral view). Rotational alignment was recorded from patients’ case notes. All patients were followed until fracture union, which was defined as fracture site no longer painful and patient able bear full weight, and radiographic evidence of union (callus on ¾ cortices).

Ethics, funding, and potential conflicts of interest The ethics committee of Helsinki University Central Hospital approved the study protocol (approval identification number 67/E7/2002). This research received no specific grant from any funding agency. The authors declare no conflicts of interest.

Results 109 patients had fractures with < 2 mm displacement and no angulation. The remaining 115 patients’ fractures had ≥ 2 mm displacement (median 6 mm, range 2–28 mm) with varus angulation in 98 (median 6°, range 3°–29°) and valgus in 17 (median 7°, range 5°–22°) cases. 121 (54%) patients were treated by casting in ER, 44 (20%) with MUA, and 59 (26%) surgically (internal fixation 57, external fixation 2) (Table 1). The median number of outpatient appointments and radiographs after primary treatment (excluding follow-up to detect possible growth plate damage) was 3 (Figure 3). The first follow-up radiograph was taken at median 10 days (5–52) after the injury in all but one of the patients. All fractures united at mean 6 weeks (4–9). At the time of fracture union 172 (77%) patients had no, 52 (23%) patients had ≤ 5°, and 1 patient 6° of residual angulation. Follow-up radiographs to monitor alignment and union of the fractures did not alter any patient’s management. 3 reoperations were performed within 48 hours of primary surgery (inadequate reduction 1, hardware in the ankle joint 2). Furthermore 7 patients were treated surgically in our institution after inadequate primary treatment (3 ER, 2 MUA, 2 OR) elsewhere.

612

Acta Orthopaedica 2019; 90 (6): 610–613

Table 1. Basic characteristics of the study population (n = 224) according to different fracture classifications Factor

Cohort

Cast

Peterson: I 6 (3%) II 123 (55%) III 12 (5%) IV 21 (9%) V 16 (6%) Tillaux 17 (7%) Triplane 29 (13%) Mean age 12 (1–15) Leg cast (weeks) 6 (2–8) Radiographs 3 (0–6) Follow-ups 3 (1–7) Time of growth control (months) 7 (3.5–15)

MUA

Operative

3 2 80 29 6 4 11 6 5 7 8 4 11 (1–15) 12 (3–15) 6 (2–7) 6 (3–7) 2 (0–5) 3 (1–5) 2 (1–6) 3 (1–6)

1 14 2 10 5 10 17 13 (8–15) 6 (4–8) 2 (2–6) 3 (1–7)

6.5 (3.5–14) 7 (4–13)

8 (4–15)

Number of patients 100 80 60 40 20

Type

Total Treatment of monitored cohort Monitored Cast MUA Operative n = 224 n = 95 n = 40 n = 20 n = 35

Peterson: I 6 3 2 – 1 II 123 60 32 14 14 III 12 5 1 2 2 IV 21 7 1 – 6 V 16 7 2 2 3 Tillaux 17 2 1 – 1 Triplane 29 11 1 2 8

Discussion

Values are given as median (range).

Table 2. Patients whose growth plate function was monitored according to Peterson classification and method of treatment

Follow-up radiographs

Figure 3. Number of follow-up radiographs to monitor fracture alignment and union of fracture.

17 (8%) patients had cast-related complications (broken cast 7, pain with intact skin 6, full-thickness skin laceration 4). All 4 ulcers were detected at scheduled follow-up appointments (6–9 days from fracture). All these 4 children complained of pain, which led to plaster change during an outpatient visit. 3/59 patients treated surgically developed superficial wound infections that healed with oral antibiotics. 6 patients had additional unplanned surgery 5–33 months after their osteosynthesis (hardware problems 5, tendon release 1). Further follow-up appointments after fracture union were scheduled for 82 (47%) patients with Peterson type fractures and 13/46 patients with transitional fractures at median 7 months (3–15) from the injury to monitor function of the distal tibial growth plate. The mean primary displacement was higher in the 82 monitored patients vs. 96 non-monitored patients with Peterson type fractures (4.7 mm vs. 2.9 mm). Similar findings were found also concerning primary fracture angulation (4.7° vs. 3.6°). Premature physeal closure was diagnosed in 23/178 patients with Peterson type fractures (13%). 14 of these patients had surgery to correct either angular deformity or leg length discrepancy (Table 2).

To our knowledge, this is the first analysis of the clinical significance of routine follow-up outpatient visits and radiographs to monitor fracture alignment and union of physeal fractures of the distal tibia. The role of radiographic follow-up in adults’ ankle fractures has been questioned by several authors (Ghattas et al. 2013, McDonald et al. 2014, Ovaska et al. 2016), and it has recently been suggested by Ovaska et al. (2016) that radiographs should not be taken at the first outpatient visit at 2 weeks from primary treatment of adult ankle fractures without clinical signs of a complication. It seems that alignment of physeal fractures of the distal tibia is very unlikely to worsen after adequate primary treatment (Spiegel et al. 1978, Berson et al. 2000, Seel et al. 2011). Our findings support these earlier findings and strongly suggest that routine radiographic follow-up is unnecessary to monitor the alignment of adequately reduced and immobilized/fixed physeal fractures of the distal tibia. We could not find a reason to take a radiograph to document union of these pediatric fractures either, since all 224 fractures in our patients united within 1–2 months in satisfactory alignment. The reported risk of skin ulcers caused by inadequate casts and surgical wound infection rates after distal tibia fracture has been low (Spiegel et al. 1978, Leary et al. 2009, Seel et al. 2011). We had 4 full-thickness cast ulcers in the 222 patients who had a cast during their treatment, which were all detected during scheduled outpatient visits. This is, however, such a low rate (2%) that routine outpatient visits could perhaps be also abandoned before cast removal. Our findings concerning the proportion of skin problems and the need for cast change should be generalized with caution, because the quality of plaster application probably varies in different institutions. We believe that in our hospital information at the time of discharge on when to return to the emergency or outpatient clinic if potential cast problems arise, such as pain, is adequate. Our results concerning the rate of growth plate damage should be interpreted with caution, because follow-up was not

Acta Orthopaedica 2019; 90 (6): 610–613

continued in all patients after union of Peterson type fractures. It was not surprising that our surgeons have not had a uniform scheme to organize follow-up to monitor growth plate function in patients with physeal fractures of the distal tibia, considering there are no distinct guidelines in orthopedic textbooks or publications (Cummings 2001, Herring and Tachdjian 2014, Lovell and Weinstein 2014). Follow-up seems unnecessary for patients with Peterson type I and transitional type (Tillaux, Triplane) fractures, but on the other hand some patients with Peterson type II–V fractures might have benefited from further follow-up after fracture union. We are going to implement a new written protocol in order to reduce the number of unnecessary follow-ups, which represent a substantial cost to society and patients’ families. (Beiri et al. 2006, Vardy et al. 2014, Holm et al. 2016). All patients with physeal fractures of the distal tibia, irrespective of the method of treatment, are going to have 1 scheduled clinical follow-up at 4–6 weeks from the injury without routine radiographs. Further radiographic follow-up is arranged for all patients with Peterson type II–VI fractures at 6 months from the injury to monitor function of the distal tibial growth plate. Peterson type I fractures and transitional fractures are not routinely monitored after fracture union. In summary, routine follow-up radiographs to monitor alignment and union of physeal fractures of the distal tibia are unnecessary in patients whose fractures are immobilized, reduced, or fixed in satisfactory alignment. How to best target follow-up to monitor growth plate function in patients with Peterson types II–VI fractures should be assessed. AS: planning of the study design, data collection, and preparation of manuscript, JK data collection, preparation of manuscript, JP: data collection, preparation of manuscript, and planning of the study design TL: data collection, preparation of manuscript. MA: preparation of manuscript. YN: planning of the study design and preparation of manuscript. Acta thanks Z. Deniz Olgun and Gunnar Hägglund for help with peer review of this study.

613

Beiri A, Alani A, Ibrahim T, Taylor G J S. Trauma rapid review process: efficient out-patient fracture management. Ann R Coll Surg Engl 2006; 88(4): 408-11. Berson L, Davidson R S, Dormans J P, Drummond D S, Gregg J R. Growth disturbances after distal tibial physeal fractures. Foot Ankle Int 2000; 21(1): 54-8. Cummings R J. Distal tibial and fibular fractures: Rockwood and Wilkins’ fractures in children. 1st ed. (Eds: C A Rockwood, K E Wilkins, J H Beaty, J R Kasser) Lippincott Williams & Wilkins 2001; pp. 1122-66. Ghattas T N, Dart B R, Pollock A G A, Hinkin S, Pham A, Jones T L. Effect of initial postoperative visit radiographs on treatment plans. J Bone Joint Surg Am 2013; 95(9): e57-S1. Herring J A, Tachdjian M O. Tachdjian’s pediatric orthopaedics. 5th ed. Philadelphia: Elsevier Saunders 2014. Holm A G V, Lurås H, Randsborg P-H. The economic burden of outpatient appointments following paediatric fractures. Injury 2016; 47(7): 1410-13. Landin L A. Fracture patterns in children: analysis of 8,682 fractures with special reference to incidence, etiology and secular changes in a Swedish urban population 1950–1979. Acta Orthop Scand 1983; (Suppl. 202): 1-109. Leary J T, Handling M, Talerico M, Yong L, Bowe J A. Physeal fractures of the distal tibia: predictive factors of premature physeal closure and growth arrest. J Pediatr Orthop 2009; 29(4): 356-61. Lovell W W, Weinstein S L. Lovell and Winter’s pediatric orthopaedics. 7th ed. Philadelphia: Lippincott Williams & Wilkins 2014. McDonald M R, Bulka C M, Thakore R V, Obremskey W T, Ehrenfeld J M, Jahangir A A, et al. Ankle radiographs in the early postoperative period: do they matter? J Orthop Trauma 2014; 28(9): 538-41. 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): 2575-8. Perry D C, Wright J G, Cooke S, Roposch A, Gaston M S, Nicolaou N, et al. A consensus exercise identifying priorities for research into clinical effectiveness among children’s orthopaedic surgeons in the United Kingdom. Bone Joint J 2018; 100-B(5): 680-4. Peterson C A, Peterson H A. Analysis of the incidence of injuries to the epiphyseal growth plate. J Trauma 1972; 12(4): 275-81. Seel E H, Noble S, Clarke N M P, Uglow M G. Outcome of distal tibial physeal injuries. J Pediatr Orthop B 2011; 20(4): 242-8. Spiegel P G, Cooperman D R, Laros G S. Epiphyseal fractures of the distal ends of the tibia and fibula: a retrospective study of two hundred and thirtyseven cases in children. J Bone Joint Surg Am 1978; 60(8): 1046-50. Vardy J, Jenkins P J, Clark K, Chekroud M, Begbie K, Anthony I, et al. Effect of a redesigned fracture management pathway and “virtual” fracture clinic on ED performance. BMJ Open 2014; 4(6): e005282-2.

614

Acta Orthopaedica 2019; 90 (6): 614–621

Femoral and pelvic osteotomies for severe hip displacement in nonambulatory children with cerebral palsy: a prospective populationbased study of 31 patients with 7 years’ follow-up Terje TERJESEN

Department of Orthopaedic Surgery, Oslo University Hospital, Rikshospitalet, Oslo, Norway Correspondence: terje.terjesen@rikshospitalet.no Submitted 2019-03-20. Accepted 2019-09-18.

Background and purpose — There is no consensus regarding the optimal treatment of hip displacement in children with cerebral palsy (CP). This prospective study assessed the outcome of femoral and pelvic osteotomies for severe hip displacement in nonambulatory children and analyzed prognostic factors for outcome. Patients and methods — 31 nonambulatory children (20 boys), recruited from a population-based screening program, consecutively underwent unilateral (23) or bilateral (8) osteotomies and bilateral soft-tissue releases at a mean age of 6.1 years (2.2–9.9). The procedures were femoral varus osteotomy alone (20 hips) and combined Dega-type pelvic osteotomy and femoral osteotomy (19 hips). Final outcome was termed good if the patient had not undergone further bony surgery and migration percentage (MP) was < 50%. The mean follow-up time was 7.1 years (3.8–11). Results — The mean preoperative MP was 69% (36– 100). The outcome was good in 22 patients (29 hips) and poor in 9 patients (10 hips). Mean time to failure was 3.6 years (1.0–6.0). GMFCS level V and high MP 1-year postoperatively were statistically significant risk factors for poor final outcome. There was a higher rate of good outcome after combined osteotomies compared with isolated femoral osteotomy, but the difference was not statistically significant (p = 0.2). Interpretation — Better primary correction was obtained after combined femoral and pelvic osteotomies than after isolated femoral osteotomy, indicating that combined osteotomies are the preferred method in hips with the most severe degrees of displacement. Prophylactic femoral osteotomy of the contralateral non-subluxated hip is hardly indicated.

There is no consensus as to the optimal surgical treatment of hip displacement in children with cerebral palsy (CP). Preventive surgery with soft-tissue releases usually provides a satisfactory outcome in ambulatory children with moderate degrees of displacement, whereas the outcome deteriorates in nonambulatory children with more pronounced displacement (Shore et al. 2012, Terjesen 2017). In this group more radical surgery such as femoral and pelvic osteotomies is indicated. Although good results have been published after these procedures (McNerney et al. 2000, Oh et al. 2007), several matters are still under discussion: optimal age at operation, whether femoral osteotomies alone or combined femoral and pelvic osteotomies should be performed, and whether contralateral prophylactic femoral osteotomy should be done when the contralateral hip is not displaced (Valencia 2010, Shore and Graham 2017). The present study is a prospective population-based study of children enrolled in the Norwegian CP registry. Previously, data from this registry have been used to evaluate the ability of soft-tissue surgery to prevent deterioration of subluxation (Terjesen 2017). The present study evaluates the outcome of femoral and pelvic osteotomies in nonambulatory children with severe hip displacement. The aims of the study were to answer the following questions: 1. What is the outcome of reconstructive hip osteotomies combined with soft-tissue releases in nonambulatory children? 2. Is there any difference in outcome between hips with combined femoral and pelvic osteotomies and hips with isolated femoral osteotomy? 3. Are there any predictors for good and poor outcome? 4. Is prophylactic femoral osteotomy of the contralateral non-subluxated hip indicated?

Acta Orthopaedica 2019; 90 (6): 614–621

Patients and methods Patients in this prospective study were recruited from the screening program for children with CP in south-east Norway, designated CPOP (Cerebral Parese Oppfølgings Program). 31 children (20 boys) born during 2002–2006 had consecutively undergone 39 femoral and/or pelvic osteotomies (8 bilaterally) during the period 2007–2014. The mean age at surgery was 6.1 years (2.2–9.9). 23 patients had not been operated on previously, whereas 8 patients had previously undergone bilateral soft-tissue release, but had relapse of their hip displacement. All the children were nonambulatory. The functional classification according to the Gross Motor Function Classification System (GMFCS) (Palisano et al. 1997) was level IV in 8 children and level V in 23. 29 children had bilateral spastic CP (quadriplegia in 23 children and diplegia in 6) and 2 children had dyskinesia (variable muscle tone). 11 patients had intrathecal baclofen therapy. The hospital’s case records were insufficient for evaluation of hip pain because of missing information. Thus, the data on pain were prospectively registered, using the information from the yearly clinical examinations at the child habilitation centers. Physiotherapists filled out a standardized form, which included whether the patients or caregivers had noticed any pain during the last 4 weeks, and the location of pain was noted. Unfortunately, the side was not specified. Radiographic evaluation An anteroposterior radiograph of the pelvis and hip joints was taken with the child in the supine position. Care was taken to position the child correctly with the legs parallel and to avoid rotation of the pelvis. The radiographic measurements were performed by the author, who has many years of experience in evaluating radiographs of children’s hips. The following radiographic parameters were measured: migration percentage (Reimers 1980), acetabular index (Hilgenreiner 1925), and pelvic obliquity. Migration percentage (MP) is the percentage of the femoral head lateral to the acetabulum (lateral to Perkins’ line), measured parallel to Hilgenreiner’s line. The hips were classified as normal (MP < 33%), subluxation (MP 33–89%), and dislocation (MP ≥ 90%) (Reimers 1980). Acetabular index (AI) is the angle between the line through the medial and lateral edges of the acetabular roof and Hilgenreiner’s line. Pelvic obliquity was measured as the angle between the horizontal line and the line between the lowest points of the pelvic bones on the right and left side. Angles < 3° were not registered as pelvic obliquity. According to the study protocol a pelvic radiograph should be taken once a year. Thus, the progression in MP per year pre- and postoperatively could be assessed. Reduction in MP caused by the operation (“primary” correction) was defined as difference between preoperative MP and MP 1-year postoperatively.

615

Table 1. Association between types of osteotomy and various clinical and preoperative radiographic variables Combined Variables FVO a osteotomies b p-value Girls Boys GMFCS level IV GMFCS level V Primary operation Secondary operation Unilateral operation Bilateral operation Baclofen pump No baclofen pump Hip pain preoperatively No hip pain preoperatively Age at surgery (SD) Migration percentage (SD) preoperative 1 year postoperatively reduction Acetabular index preoperative (SD) Pelvic obliquity preoperative (SD)

7 6 4 9 12 1 7 6 2 11 5 8 5.2 (1.7)

4 14 4 14 11 7 16 2 9 9 5 13 6.7 (1.7)

0.07

63 (15) 27 (14) 36 (24) 30 (3.6) 3.1 (3.2)

75 (18) 19 (12) 56 (21) 30 (3.3) 4.6 (4.9)

0.02 0.09 0.008 0.8 0.3

0.6 0.05 0.03 0.05 0.5 0.02

a b

Femoral varus osteotomy. Femoral varus osteotomy and pelvic osteotomy. GMFCS = gross motor function classification system. SD = standard deviation.

Operative procedures Preoperatively, 16 patients had unilateral hip displacement (MP > 33%) and they underwent unilateral bony surgery. Thus, prophylactic contralateral osteotomies were not performed. 15 patients had bilateral displacement, of whom 8 patients underwent bilateral bony surgery. In the remaining 7 patients, unilateral bony surgery of the worst hip was performed. Choice of osteotomies (femoral osteotomy alone or combined femoral and pelvic osteotomies) varied according to the surgeons’ preferences, but usually combined osteotomies were chosen in complete dislocations and for the most severe subluxations. Proximal femoral varus osteotomy alone was performed in 13 patients (20 hips) and combined osteotomies were done in 18 patients (19 hips). Children with combined osteotomies had higher age at surgery, higher preoperative MP, better primary correction, more often intrathecal baclofen pump, and more often unilateral osteotomies (Table 1). The surgical procedures were done with the patient in the supine position, and an image intensifier was used. Femoral varus osteotomy was performed through a longitudinal lateral approach. A transverse femoral osteotomy just above the lesser trochanter was performed with an oscillating saw. The osteotomy was a combination of varization, derotation, and shortening. One Kirschner wire drilled transversely into the femur proximal to the osteotomy and one distal to the osteotomy were used as guide wires for derotation. A neck–shaft angle of 110–120° and derotation of about 30° were aimed

616

C Figure 1. A. Preoperative radiograph of a boy, aged 7.0 years and GMFCS level V, with severe subluxation of his right hip (MP 67%). B. 1 day after femoral osteotomy (varus, derotation and shortening), pelvic osteotomy, and bilateral soft tissue releases, showing good femoral head coverage bilaterally. The cortical bone segment removed from femur has been used as autograft in the open wedge of the pelvic osteotomy. C. 7.4 years postoperatively, at an age of 14.4 years, showing satisfactory femoral head coverage bilaterally, MP 19% (right hip) and 0% (left hip)

at. In 24 osteotomies a second transverse osteotomy distal to the first was performed, and a wedge of bone was removed, aiming for a femoral shortening of 1–2 cm. The osteotomy

Acta Orthopaedica 2019; 90 (6): 614–621

was fixed with a 110° pediatric locked compression plate (LCP; Synthes, Switzerland) (Rutz and Brunner 2010) in 30 osteotomies (Figure 1) and a 90° AO blade plate in 2. In the remaining 7 osteotomies a straight plate with 2 screws in each fragment was used, after the plate had been bent according to the planned varization. For the pelvic osteotomy, a modification of the incomplete transiliac Dega osteotomy was performed (Grudziak and Ward 2001) through a transverse anterior incision approximately 2 cm distal to the superior anterior iliac spine. The anterior part of the iliac apophysis was split and the inner and outer tables of the ilium were subperiosteally exposed. The osteotomy was performed with curved osteotomes. It started just above the anterior inferior iliac spine and proceeded posteriorly, keeping about 1.5 cm above the attachment of the joint capsule. The direction of the osteotomy was medially and inferiorly and ended just above the horizontal limb of the triradiate cartilage, leaving the posterior part of the cortex at the sciatic notch intact. A broad osteotome was used to lever open the osteotomy laterally and anteriorly. The bone graft from the femoral shortening was inserted in the open wedge (Figure 1). In the 23 patients without previous soft-tissue releases, simultaneous bilateral tenotomies of adductor longus, gracilis, and iliopsoas were performed, using a short oblique incision over the origin of adductor longus. Postoperatively, a plaster cast with moderate hip abduction was used for 6 weeks in 22 children and an abduction orthosis was used in 9 children. Assessment of outcome The outcome at the last follow-up was graded into 2 categories according to a modification of the classification of Shore et al. (2015), with both the patient (n = 31) and the hip (n = 39) as the unit of analysis. The outcome was termed good (“success”) when MP at the last follow-up was < 50%. If final MP was ≥ 50% and/or the patient had undergone subsequent bony surgery (pelvic and/or femoral osteotomies) to improve femoral head coverage, the outcome was poor (“failure”). In addition, the clinical outcome was assessed by the information on pain at the last CPOP follow-up. For those who underwent reoperation, pain evaluation at the last examination before reoperation was used. Statistics SPSS (version 25) was used for the statistical analysis (IBM, Armonk, NY, USA). Categorical variables were analyzed with the Pearson chi-square test. Continuous variables were analyzed with Student’s t-test. Potential risk factors for poor final outcome were estimated as relative risks using univariable Poisson loglinear regression and multivariable exact logistic regression. All tests were 2-sided. Differences were considered significant when the p-value was < 0.05. The percentage survival according to postoperative time (years) was described with a Kaplan–Meier plot.

Acta Orthopaedica 2019; 90 (6): 614–621

617

Table 2. Potential prognostic factors for poor radiographic outcome in 31 patients, estimated as relative risks (RR) for failure, using univariable Poisson loglinear regression Outcome Good Failure

Variables

Girls Boys GMFCS, level IV GMFCS, level V a Primary operation Secondary operation Unilateral operation Bilateral operation No pain preoperatively Pain preoperatively Combined osteotomies Femoral osteotomies Age (SD) Migration percentage (SD) preoperative 1 year postoperatively Acetabular index preop. (SD) Pelvic obliquity preop. (SD)

Statistical analysis RR (95% CI) p-value

7 15 8 14 15 7 18 4 15 7 15 7 6.4 (1.9)

4 5 0 9 8 1 5 4 6 3 3 6 5.4 (1.8)

Reference 0.7 (0.2–2.6) Reference OR 6.3 (0.8–∞) Reference 0.4 (0.1–2.9) Reference 2.3 (0.6–8.6) Reference 1.1 (0.3–4.2) Reference 2.8 (0.7–11.1) 0.8 (0.6–1.2)

72 (19) 21 (11) 30 (3.0) 3.4 (3.7)

69 (12) 35 (14) 31 (3.8) 5.9 (5.8)

1.0 (0.96–1.03) 1.05 (1.01–-1.1) 1.08 (0.88–1.3) 1.1 (0.94–1.3)

0.6 0.08 0.3 0.2 0.9 0.2 0.3 0.8 0.03 0.5 0.3

For abbreviations, see Table 1. a Exact logistic regression was used because one group contained the value “0.”

Table 3. Multivariable analysis of odds ratios for poor outcome, including variables with a p-value < 0.2 in univariable analysis, using multivariable exact logistic regression Variables GMFCS, level V Femoral osteotomy Migration percentage 1 year postop.

Odds ratio 95% CI 15 5.4 1.1

1.4–∞ 0.5–237 1.0–1.4

p-value 0.02 0.2 0.01

For abbreviations, see Table 1. CI = confidence interval.

Results C Figure 2. A. Preoperative radiograph of a girl, aged 8.1 years and GMFCS level V, with severe subluxation of her left hip (MP 74%). B. 6 weeks after femoral and pelvic osteotomies of the left hip and bilateral soft tissue releases, showing good femoral head coverage. C. 2.9 years postoperatively (age 11.0 years), showing relapse of subluxation of her left hip (MP 51%).

Ethics, funding, and potential conflicts of interest The study was approved by the Regional Committee of Medical Research Ethics (no. 2012/2258) and the hospital’s Privacy and Data Protection Officer. No external funding was received for this study and there are no conflicts of interest.

The mean follow-up time of children who have not undergone subsequent hip surgery was 7.1 years (3.8–11) and their mean age at the latest follow-up was 13.6 years (11.3–16.6). When the outcome of the worst hip was used in patients with bilateral bony procedures, the outcome was good in 22 patients and poor in 9 patients (Figure 2). When hip was used as the unit of analysis, a good outcome was achieved in 29 out of 39 hips. Potential risk factors for poor final outcome were analyzed with univariable Poisson loglinear regression (Table 2). When the variables with p-value < 0.2 were analyzed with multivariable exact logistic regression, GMFCS level V and high MP 1 year postoperatively were independent risk factors for poor outcome (Table 3). The rate of good outcome after combined osteotomies was higher (good results in 15 of 18 patients) than that after femoral osteotomy alone (good results in 7 of 13

618

Acta Orthopaedica 2019; 90 (6): 614–621

Table 4. Migration percentage pre- and postoperatively (mean values with 95% confidence intervals (CI) in parentheses) of all 39 hips and comparison between hips with good and poor radiographic outcome Outcome Migration percentage All hips Good Poor

Difference Mean (95% CI)

Preoperative Primary correction b 1 year postop 3 years postop 5 years postop Progression per year preoperative c postoperative Contralateral side preoperative last follow-up

0.8 13 – 13 –22 –32

69 (63–75) 46 (38–54) 23 (19–28) 30 (24–36) 33 (24–43)

69 (62–76) 49 (39–59) 20 (16–24) 25 (21–30) 25 (18–33)

68 (60–77) 36 (22–49) 33 (22–43) 47 (27–67) 58 (32–83)

15 (11–19) 2.9 (1–5)

14 (10–19) 16 (6–25) 0.7 (–0.4–1.8) 10 (5–15)

(–12 to 14) (–4 to 31) (–22 to –4) (–34 to –10) (–49 to –15)

–1.0 (–11 to 9) –9.5 (–15 to –4)

p-value a 0.9 0.1 0.008 0.04 0.001 0.8 0.002

29 (24–34) 26 (16–35)

a Student’s t-test. b Difference between migration c Progression of preoperatively

percentage preoperative and 1 year postoperatively. was calculated during a mean period of 2.9 years (0.8–5.8).

patients), but the difference was not statistically significant (p = 0.2). 2 patients died 3.7 and 6.2 years, respectively, after the index operation. The mean preoperative MP was 69% (36–100), 72% in hips with unilateral osteotomies and 65% in those with bilateral procedures. Only 4 of the operated hips had mild or moderate subluxation (MP 36–48%). The remaining 35 hips had severe subluxation (MP ≥ 50%) or complete dislocation. MP during the entire postoperative period was larger in hips with poor final outcome (Table 4). The progression in MP (mean change per year) was pronounced (15%) during the mean preoperative observation period of 2.9 years, with no statistically significant difference between hips with good and poor outcome. The mean postoperative progression was less than 1% per year in hips with good outcome and 10% per year in hips with poor final outcome (Table 4). In the 23 patients with unilateral osteotomies, the contralateral hip deteriorated in 3 patients. The surgical guidelines had not been followed in 1 of these patients, because contralateral soft-tissue releases had not been performed. 2 patients had subluxation with preoperative MP 48% and 51%, respectively, on the contralateral side, and these hips deteriorated postoperatively (Figure 3). Of the remaining 5 patients with contralateral subluxation (MP 36–50%) but no contralateral bony surgery, postoperative normalization occurred in 4 patients (MP < 33%) whereas subluxation remained unchanged (MP 38% preoperatively and 40% at follow-up) in 1 patient. None of the contralateral hips with preoperative MP < 33% and bilateral soft-tissue releases developed subluxation during follow-up. Preoperative hip pain had been noted in 10 of the 31 children. 1 year postoperatively these patients were painless, but 6 of the other patients had hip pain. At the last follow-up, pain was significantly more frequent in patients with poor radiographic outcome (5 of 9 patients) than in patients with good outcome (2 of 22 patients; p = 0.005).

Postoperative complications occurred in 4 patients. 2 children had failure of the femoral fixation and re-dislocation of the osteotomy 1–3 months postoperatively, which was treated with plate re-fixation in 1 child and plaster cast in the other. Decubitus ulcer of the heel because of pressure from the plaster and pneumonia occurred in 1 patient each. In addition to these early complications, femoral fracture of the operated extremity occurred 1–5 years postoperatively in 4 patients. The fractures were treated with plate fixation. There were no cases of avascular necrosis of the femoral head. The mean time to failure was 3.6 years (1.0–6.0). There was a shorter mean time to failure after femoral varus osteotomy (3.1 years) than after combined osteotomies (5.0 years), but the difference was not statistically significant (p = 0.2). Kaplan–Meier survival analysis with time to failure as endpoint showed that survival fell from 95% 1 year postoperatively to 74% at 6 years (Figure 4). No obvious reason for failure could be identified in 5 patients (6 hips). The reason was poor primary correction in 3 patients. 1 child had repeated femoral fractures of the operated side and also had respiratory problems; therefore, proximal femoral resection was performed after his third fracture. Reoperation has been performed in 5 hips (1 femoral osteotomy, 2 pelvic osteotomies, 1 combined pelvic and femoral osteotomies, and 1 proximal femoral resection). The mean period from index operation to reoperation was 3.6 years (2.3–6.3). 5 hips have so far not undergone further surgery; their mean MP at the last followup was 69% (51–100).

Discussion Reconstructive osteotomies are indicated in children with severe hip displacement, but there is no consensus as to which surgical strategy is preferable. Since the outcome deteriorates

Acta Orthopaedica 2019; 90 (6): 614–621

619

K–M failurefree survival 1.0

0.8

0.6

0.4

0.2

Hips at risk: 37 36 33 0

Years after index operation

Figure 4. Kaplan–Meier survival plot (% survival with 95% confidence intervals) in all 39 hips, with time to failure (reoperation or MP ≥ 50%) as “survival”.

C Figure 3. A. Preoperative radiograph of a girl, aged 5.7 years and GMFCS level V, with complete dislocation of her right hip (MP 100%) and subluxation of her left hip (MP 51%). B. 14 months after femoral and pelvic osteotomies of the right hip and bilateral soft tissue releases, showing slight subluxation of both hips (MP right hip 36% and left hip 37%). C. 7.3 years postoperatively (age 13.0 years), showing good position of right hip and deterioration of left hip (MP 64%).

with increasing length of follow-up and with decreasing functional capacity (Shore et al. 2015), nonambulatory children with a follow-up of > 5 years should be separately analyzed, as was done in Table 5 (see Supplementary data). The median failure rates were 31% (22–25) after isolated femoral osteotomy and 15% (9–20) after combined femoral and pelvic osteotomies. Previous reports have recommended combined osteotomies in patients with severe subluxation or dislocation and

pronounced dysplasia of the acetabulum (Brunner and Baumann 1997, McNerney et al. 2000, Sankar et al. 2006, Oh et al. 2007). This is in accordance with the strategy of the present study, where hips with combined osteotomies had larger preoperative displacement and better primary correction than those with femoral osteotomy alone. Pelvic osteotomy is, however, an extensive procedure and represents a large surgical trauma to a severely involved patient with respiratory or other medical problems. Therefore, the indications for pelvic osteotomy were based on the surgeon’s preference after discussion of benefits and risks with the parents and child neurologists. Shore et al. (2015) found better outcome at GMFCS levels I–III than in nonambulatory children at levels IV/V. Different results have been published regarding GMFCS level IV versus level V. Whereas failure rates at 5 years of 15% at GMFCS level IV and 24% at level V were reported by Shore et al. (2015), failure rates of 26% in both GMFCS levels IV and V were found by Zhang et al. (2014). In the present study, level V was a significant risk factor for poor outcome. The fact that no failures occurred after femoral osteotomies in children at GMFCS level IV indicated that femoral osteotomy alone should be considered in patients with this grade of function; however, the number of patients is too small to draw definite conclusions. In GMFCS level V, the outcome after femoral osteotomy was poor in 7 of 14 hips, indicating that combined osteotomies would be the preferred procedure. High preoperative MP was a risk factor for poor outcome in some studies (Oh et al. 2007, Settecerri and Karol 2000, Rutz et al. 2015) but neither in the present study nor in the multivariate analysis by Shore et al. (2015). High MP 1 year postoperatively, which was a significant risk factor in the present study, seems not to have been assessed in previous studies. Age was not a prognostic factor in the present study and in 2 previous studies (Settecerri and Karol 2000, Oh et al. 2007),

620

whereas young age (< 6 years) was a risk factor of failure in other studies (Brunner and Baumann 1997, Ruzbarsky et al. 2013, Shore et al. 2015). When the patients are followed in a population-based screening program, there is no reason to postpone surgery when the MP exceeds 40–50%, no matter the age of the patient. It was therefore not unexpected that the age of the present patients was in the lower range compared with other studies (Table 5). The need for concomitant open reduction is controversial. Some authors recommended open reduction (Jozwiak et al. 2000, McNerney et al. 2000, Sankar et al. 2006) whereas others did not perform open reduction (Settecerri and Karol 2000, Mallet et al. 2014). Capsulotomy and open reduction were not used in the present study because clinical experience and peroperative fluoroscopy indicated that the hips were sufficiently contained by osteotomies and soft-tissue releases. Opinion differs with regard to surgical strategy in patients with a normal contralateral hip. Whereas some studies recommended prophylactic, concurrent femoral osteotomy of the contralateral hip to obtain pelvic balance (Barakat et al. 2007, Oh et al. 2007) others recommended unilateral osteotomy only (Settecerri and Karol 2000, Larsson et al. 2012). After unilateral soft-tissue or bony surgery in 27 nonambulatory patients with no subluxation of the contralateral side, deterioration of the contralateral side occurred in 19 hips, leading the authors to caution against unilateral surgery (Noonan et al. 2000). When unilateral osteotomy was combined with bilateral softtissue releases, the rate of contralateral deterioration was only 8% (Larsson et al. 2012). This is in keeping with the present results. A decision analysis (Park et al. 2012) estimated that the contralateral hip should be prophylactically operated if the rate of later instability was ≥ 27%. Shukla et al. (2013) found that lack of contralateral soft-tissue release and preoperative MP > 25% were risk factors for contralateral subluxation. In the present study, no contralateral deterioration was seen when the preoperative MP was less than 33% and the surgical guidelines had been followed. Therefore, a reasonable conclusion would be that contralateral prophylactic femoral osteotomy is not indicated if bilateral soft-tissue releases are performed. There is a trend to deterioration of outcome with increasing follow-up time. Shore et al. (2015) reported failure rates of 24% at 5 years and 42% at 10 years after osteotomies in children at GMFCS level V. The mean time to failure in the present study was 3.6 years at GMFCS level V and there were no failures later than 6 years postoperatively. Zhang et al. (2014) reported a somewhat longer mean time to failure (5 years). If the hip remains stable during the first postoperative years, there is a relatively small risk of later deterioration. The mean postoperative progression in MP was only 0.7% per year in hips with good outcome in the present study, which is similar to the progression rate of 1.1% after combined osteotomies (Mallet et al. 2014). Pain relates strongly to quality of life and is therefore an important outcome parameter. A recent population-based

Acta Orthopaedica 2019; 90 (6): 614–621

study in children with CP found, at a mean age of 9.5 years, hip pain in 14% of children with normal hips and in about 60% of children with severe subluxation (Ramstad and Terjesen 2016). The present results showed that operative treatment had a good effect on hip pain, in agreement with previous studies (Barakat et al. 2007, Rutz et al. 2015). The reported rate of postoperative complications after hip reconstructions varies considerably, from 0% to 81% (Ruzbarsky et al. 2013). In a systematic review (Hesketh et al. 2016), the AVN rate varied from 0% to 46%. These rates are difficult to compare because different classifications were used and because of mixed populations of patients. The complication rate was moderate in the present study. There were no cases of avascular necrosis of the femoral head, which is in accordance with some other studies (Sankar et al. 2006, Mallet et al. 2014). Not performing capsulotomy and open reduction might have contributed to the low rate of avascular necrosis. There are some limitations of the study. First the number of patients was small, which could reduce the validity of the statistical analysis. The choice of osteotomies was not randomized, but based on the surgeon’s preference. Pain was reported crudely and without a validated scoring system. Pain assessment in severely involved CP patients is difficult because the children have limited ability to express the presence and intensity of pain. Information from caregivers is necessary, although such information could also be somewhat unreliable. The strengths of the study are that it is population-based and prospective, with a follow-up of all the children. Conclusions 1. The outcome was good in 22 of 31 patients (29 of 39 hips) after reconstructive hip osteotomies in nonambulatory children with severe hip displacement. 2. There was a higher rate of good outcome after combined osteotomies compared with isolated femoral osteotomy, but the difference was not statistically significant (p = 0.2). 3. Significant risk factors for a poor final outcome were markedly decreased functional level (GMFCS level V) and high migration percentage 1 year postoperatively. 4. Routine prophylactic femoral osteotomy of the contralateral non-subluxated hip is hardly indicated. Supplementary data Table 5 is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2019.1675928

The author would like to thank the physiotherapists of the child habilitation teams who coordinated the radiographic screening. He also thanks the statistician Are Hugo Pripp for help with the statistical analyses. Acta thanks Unni Narayanan and Erich Rutz for help with peer review of this study.

Acta Orthopaedica 2019; 90 (6): 614–621

Barakat M J, While T, Pyman J, Gargan M, Monsell F. Bilateral hip reconstruction in severe whole-body cerebral palsy: ten-year follow-up results. J Bone Joint Surg Br 2007; 89-B: 1363-8. Brunner R, Baumann J U. Long-term effects of intertrochanteric varus-derotation osteotomy on femur and acetabulum in spastic cerebral palsy: an 11- to 18-year follow-up study. J Pediatr Orthop 1997; 17: 585-91. Grudziak J S, Ward W T. Dega osteotomy for the treatment of congenital dislocation of the hip. J Bone Joint Surg Am 2001; 83-A: 845-54. Hesketh K, Leveille L, Mulpuri K. The frequency of AVN following reconstructive hip surgery in children with cerebral palsy: a systematic review. J Pediatr Orthop 2016; 36: e17-e24. Hilgenreiner H. Zur Frühdiagnose und Frühbehandlung der angeborenen Hüftgelenkverrenkung. Med Klin 1925; 21(38): 1425–9. Jozwiak M, Marciniak W, Piontek T, Pietrzak S. Dega’s transiliac osteotomy in the treatment of spastic subluxation and dislocation in cerebral palsy. J Pediatr Orthop B 2000; 9: 257-64. Larsson M, Hägglund G, Wagner P. Unilateral varus osteotomy of the proximal femur in children with cerebral palsy: a five-year follow-up of the development of both hips. J Child Orthop 2012; 6: 145-51. Mallet C, Ilharreborde B, Presedo A, Khairouni A, Mazda K, Pennecot G F. One-stage hip reconstruction in children with cerebral palsy: long-term results at skeletal maturity. J Child Orthop 2014; 8: 221-8. McNerney N P, Mubarak S J, Wenger D R. One-stage correction of the dysplastic hip in cerebral palsy with the San Diego acetabuloplasty: results and complications in 104 hips. J Pediatr Orthop 2000; 20: 93-103. Noonan K J, Walker T L, Kayes K J, Feinberg J. Effect of surgery on the nontreated hip in severe cerebral palsy. J Pediatr Orthop 2000; 20: 771-5. Noonan K J, Walker T L, Kayes K J, Feinberg J. Varus derotation osteotomy for the treatment of hip subluxation and dislocation in cerebral palsy: statistical analysis in 73 hips. J Pediatr Orthop 2001; 10: 279-86. Oh C-W, Presedo A, Dabney K W, Miller F. Factors affecting femoral osteotomy in cerebral palsy: a long-term result over 10 years. J Pediatr Orthop B 2007; 16: 23-30. Palisano R J, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi B. Development and validation of a gross motor function classification system for children with cerebral palsy. Dev Med Child Neurol 1997; 39: 214–23. Park M S, Chung C Y, Kwon D G, Sung K H, Choi I H, Lee K M. Prophylactic femoral varization osteotomy for contralateral stable hips in non-ambulant individuals with cerebral palsy undergoing hip surgery: decision analysis. Dev Med Child Neurol 2012; 54: 231-9.

621

Ramstad K, Terjesen T. Hip pain is more frequent in severe hip displacement: a population- based study of 77 children with cerebral palsy. J Pediatr Orthop B 2016; 25: 217-21. Reimers J. The stability of the hip in children. Acta Ortop Scand 1980; 51(Suppl. 184): 12–91. Rutz E, Brunner R. The pediatric LCP hip plate for fixation of proximal femoral osteotomy in cerebral palsy and severe osteoporosis. J Pediatr Orthop 2010; 30: 726-31. Rutz E, Vavken P, Camathias C, Haase C, Jünemann S, Brunner R. Long-term results and outcome predictors in one-stage hip reconstruction in children with cerebral palsy. J Bone Joint Surg Am 2015; 97-A: 500-6. Ruzbarsky J J, Beck N A, Baldwin K D, Sankar W N, Flynn J M, Spiegel D A. Risk factors and complications in hip reconstruction for nonambulatory patients with cerebral palsy. J Child Orthop 2013; 7: 487-500. Sankar W N, Spiegel D A, Gregg J R, Sennett B J. Long-term follow-up after one-stage reconstruction of dislocated hips in patients with cerebral palsy. J Pediatr Orthop 2006; 26: 1-7. Settecerri J J, Karol L A. Effectiveness of femoral varus osteotomy in patients with cerebral palsy. J Pediatr Orthop 2000; 20: 776-80. Shore B J, Yu X, Desai S, Selber P, Wolfe R, Graham H K. Adductor surgery to prevent hip displacement in children with cerebral palsy: the predictive role of the Gross Motor Function Classification System. J Bone Joint Surg Am 2012; 94-A: 326-34. Shore B J, Zurakowski D, Dufreny C, Powell D, Mathenay T H, Snyder B D. Proximal femoral varus derotation osteotomy in children with cerebral palsy. The effect of age, gross motor classification system level, and surgical volume on surgical success. J Bone Joint Surg Am 2015; 97-A: 202431. Shore B J, Graham H K. Management of moderate to severe hip displacement in nonambulatory children with cerebral palsy. J Bone Joint Surg Am 2017; 5(12): e4:1-12. Shukla P Y, Mann S, Braun S V, Gholve P A. Unilateral hip reconstruction in children with cerebral palsy: predictors for failure. J Pediatr Orthop 2013; 33: 175-81. Terjesen T. To what extent can soft-tissue releases improve hip displacement in cerebral palsy? A prospective population-based study of 37 children with 7 years’ follow-up. Acta Orthop 2017; 88: 695-700. Valencia F G. Management of hip deformities in cerebral palsy. Orthop Clin N Am 2010; 41: 549-59. Zhang S, Wilson N C, Mackey A H, Stott, N S. Radiological outcome of reconstructive hip surgery in children with gross motor function classification system IV and V cerebral palsy. J Pediatr Orthop B 2014; 23: 430-4.

622

Acta Orthopaedica 2019; 90 (6): 622–623

Functional ambulation without lower-leg muscles or nerves — a case report with video John PARENTI Department of Orthopaedic Surgery, Geisinger Medical Center, Danville, PA, USA Correspondence: jparenti@geisinger.edu Submitted 2019-03-11. Accepted 2019-04-15.

Our orthopedic service was called for an urgent consultation concerning a 56-year-old man who had been admitted to Intensive Care some 36 hours earlier because of syncope, cyanosis, profound orthostatic hypotension, severe hemoconcentration, and acute bilateral leg pain. His medical history was unremarkable, and the tentative diagnosis given was polycythemia of unknown etiology leading to venous thromboembolism, with a distributive shock pattern consistent with sepsis. In the wake of aggressive IV hydration while intubated, swelling and tenseness had developed in his lower extremities secondary to massive thrombosis in all limbs. The patient’s pressures were greater than 60 mmHg in all 4 compartments of both lower extremities, and they were elevated also in his forearms. Bilateral lower-extremity fasciotomies were performed immediately, and after the fascia were incised the muscles were poorly contractile and there was evidence of developing necrosis. 2 subsequent interventions in the days that followed entailed staged debridements to remove some necrotic musculature from both anterior and lateral lower-leg compartments. Based on the patient’s symptoms and clinical findings, the ICU staff eventually considered an alternative diagnosis of systemic capillary leak syndrome (SCLS). SCLS is a rare episodic disease of unknown etiology characterized by selfreversing episodes during which the endothelial cells that line the capillaries, usually in the extremities, separate for up to 3 days, causing a leakage of plasma mainly into the muscle compartments of arms and legs (Druey and Parikh 2017). The extravasation can be sufficiently massive to cause circulatory shock and compartment syndrome, especially when undiagnosed SCLS patients are treated as septic and are resuscitated aggressively, as happened here. In the span of 1 week, the patient recovered hemodynamic stability and his venous clots dissolved, CPK levels decreased, liver and kidney functions normalized, and peripheral edema resolved, whereupon he was extubated. These developments supported the diagnosis of SCLS. However, given the extent of rhabdomyolysis, our surgical team was uniformly pessimistic on the chances of salvaging the patient’s lower limbs. He was deemed too vulnerable to infection, accidental loss of blood circulation to the feet, and other complications and negative outcomes (e.g., reflex sympathetic dystrophy and paresthesia)

of limb salvage in the wake of acute compartment syndrome (von Keudell et al. 2015). Even in a best-case scenario, it was hard to imagine that the patient would be left with functional extremities. Bilateral above-the-knee amputations appealed to our team for their ability to enhance patient survival, reduce pain and disability, and shorten hospitalization. As per the conventional wisdom, limbs that would likely end up flail, painful, insensate, and nonfunctional would be inferior to an amputation and prosthetic fitting—especially true in the lower extremities, given that modern prosthetics have proven to be effective in the restoration of almost normal function (Russell et al. 1991). The patient was adamantly against amputation, however, arguing that his overall stabilization and good baseline health would see him through the necessary serial debridements. Since there were pulses in the patient’s feet, his general clinical picture was rapidly improving, and he was relatively young and free of comorbidities, there was no need for an urgent amputation. Nonaggressive debridements took 9 more surgeries over a 5-week period. At the end, no muscles or nerves survived below either knee, and both feet were immobile and insensate. The patient’s sequelae were pain and swelling in all extremities and major flexion contractures of the hands, because he had also suffered partial muscular and neurological damage in the forearms. He was transferred to an inpatient facility for several weeks of physical and occupational therapy, and additional weeks of outpatient hand therapy and physical rehabilitation followed his discharge to home. These events took place more than 13 years ago, during the winter of 2005–2006, and the patient is alive and well, with his SCLS under control since 2010. Within 6 months of discharge from the rehabilitation facility in February 2006, he was ambulatory and back to work as a full-time university professor. This entailed the return to a grueling weekly commute between 2 major East Coast cities using public transportation (buses, subways, and trains), while taking care of personal needs entirely on his own. By now he is 69 years old and continues to work and commute. He walks on paved surfaces with a surprisingly normal gait and takes very natural, long strides that belie his disability (Figure 1). He can ambulate up or down ramps and stairs with little difficulty, though he is justifiably unsteady on

Acta Orthopaedica 2019; 90 (6): 622–623

Figure 1. Patient standing at 13 years post-trauma.

soft, uneven, or slippery terrain (e.g., sand or snow). While he cannot run, he does not mind because he had always led a sedentary life and did not play sports. The initial swelling of his feet gradually dissipated with the use of compression stockings, and his limb pain subsided such that he quickly weaned himself from pain medication. For greater stability while ambulating outdoors, he uses a cane and a discreet, carbon-fiber ankle–foot orthosis for added ankle stability. While in the comfort of his 2 apartments, however, he manages so well that he neither uses the cane nor puts on his braces. Thus, he can get up from bed in the nighttime and visit the bathroom simply by propelling himself forward while maintaining balance. The patient has taken excellent care of his feet: they are free of deformities, scars, or skin changes. He has never needed to be treated for an infection in either foot. The physical condition of his legs and feet after more than 13 years of continuous, post-traumatic daily use is admirable. He can be watched ambulating with orthoses at 13 years post-trauma on the internet (https://www.youtube. com/watch?v=We_XXrQUmkw&feature=youtu.be), and without braces or cane (https://www.youtube.com/ watch?v=ojB1sk7ARhw&feature =youtu.be).

Discussion This patient provides an uplifting example of the power of personal determination, good habits (e.g., he is a non-diabetic, non-smoker), and physical and mental adaptation. We think of him as having been amputated, after all, but as having provided his own prostheses. His ability to balance himself is

623

akin to someone who has learned to walk on stilts. His knees, quadriceps, and hips are evidently doing some of the work that his calves and ankles used to perform, yet without causing any damaging wear and tear. The patient’s incredible adjustment to otherwise devastating loss is illustrated best by the fact that he has taken to driving his own car once again, having logged post-trauma over 150,000 miles of city and highway driving without incident. Evidently, his knees, quads, and eyes provide him the necessary information to distinguish between accelerator and brake pedals, and to regulate vehicle speed at short notice. To be sure, we have encountered sedentary patients who are satisfied with a poorly functional salvaged limb they can call their own (Attinger and Brown 2012). But the patient reported here rightly perceives his salvaged limbs to be very functional, and, as they have proven to be biomechanically sound and durable despite the passage of time, he has every reason to be happy with them. With the benefit of hindsight, amputation was not the better choice. Therefore, his case instigated our department to question the conventional wisdom concerning the lack of function in flail and insensate lower legs, as well as in rigid and insensate feet, to the extent that our service has since embraced the cause of limb salvage. We became much more cautious before making amputation recommendations, and we established a now busy wound-care center dedicated to the cause of limb preservation. This case is a vivid reminder that the decision to attempt limb salvage or to favor amputation is—and should always be—a patient-centered decision (Fiorito et al. 2012). Outcomes assessed from a patient’s perspective have the potential to be distinctly different from those reached by the treating surgeon (Momoh and Chung 2013). Therefore, we must resist the impulse to advocate solutions that appeal because they are relatively safe and expedient. Surgeons must make earnest efforts to understand patients’ medical history, lifestyle, and recovery potential before making assumptions and recommendations—especially as regards irreversible interventions. Attinger C E, Brown B J. Amputation and ambulation in diabetic patients: function is the goal. Diabetes Metab Res Rev 2012; 28(Suppl. 1): 93-6. DOI: 10.1002/dmrr.2236. Druey, K M, Parikh, S M. Idiopathic systemic capillary leak syndrome (Clarkson disease). J Allergy Clin Immunol 2017; 140(3): 663-70. DOI: 10.1016/j.jaci.2016.10.042. Fiorito J, Trinidad-Hernadez M, Leykum B, Smith D, Mills J L, Armstrong D G. A tale of two soles: sociomechanical and biomechanical considerations in diabetic limb salvage and amputation decision-making in the worst of times. Diabetic Foot Ankle 2012; 3(1). DOI: 10.3402/dfa.v3i0.18633. Momoh A O, Chung K C. Measuring outcomes in lower limb surgery. Clin Plast Surg 2013; 40(2): 323-9. DOI: 10.1016/j.cps.2012.10.007. Russell W L, Sailors D M, Whittle T B, Fisher D F Jr, Burns R P. Limb salvage versus traumatic amputation: a decision based on a seven-part predictive index. Ann Surg 1991; 213(5): 473-80. PMCID: PMC1358477. von Keudell A G, Weaver M J, Appleton P T, Bae D S, Dyer G S M, Heng M, Jupiter J B, Vrahas M S. Diagnosis and treatment of acute extremity compartment syndrome. Lancet 2015; 386(10000): 1299-1310. DOI: 10.1016/ S0140-6736(15)00277-9.

624

Acta Orthopaedica 2019; 90 (6): 624–625

An unexpected complication of nonoperative treatment for tibial posterior malleolus fracture: bony entrapment of tibialis posterior tendon – a case report Thomas AMOUYEL 1, Baptiste BENAZECH 2, Marc SAAB 1, Nadine STURBOIS-NACHEF 1, Carlos MAYNOU 1, and Patrice MERTL 2 1 Université

de Lille Nord de France, Service d’orthopédie 1, Hôpital Roger Salengro,Centre Hospitalier Universitaire de Lille, France; 2 Service orthopédie, Centre Hospitalo-Universitaire Amiens Picardie, 80480 Amiens, France Correspondence: thomas.amouyel@chru-lille.fr Submitted 2019-01-29. Accepted 2019-07-05.

A 41-year-old patient was referred to our center because of right medial ankle pain increasing for 3 months. He had, 10 years ago, had a displaced lateral malleolus fracture with an associated non-displaced posterior malleolus fracture but without a medial malleolus fracture. A fibular osteosynthesis without medial or posterior exploration was done at another hospital. Postoperatively, the patient remained non-weightbearing for 6 weeks with a cast. The patient recovered completely and returned to work as a fireman 4 months after the initial injury. The fibular osteosynthesis material was removed 1 year after the surgery. 10 years later, when referred to our center, he had increasing medial ankle swelling and pain, preventing him from working as a fireman. Physical examination revealed a medial retromalleolar swelling with local tenderness, but no flat-foot deformity. Testing of the tibialis posterior tendon (TPT) was positive: heel-rise test and strength assessment were painful but with no loss of strength. Radiographs showed a healed

Figure 1. Posterior tibial bony callus (arrow) of the right ankle.

lateral malleolus fracture in good alignment with a posterior tibial bony callus (Figure 1). A CT scan showed a medial retromalleolar bone tunnel containing the TPT of 15 mm length (Figure 2); the fibular and the posterior malleolus fractures were healed. An MRI scan showed tenosynovitis of the TPT (Figure 3), and thickening of the anterior talo-fibular ligament. We performed an open resection of the postero-medial part of the tunnel to release the TPT (Figure 4). The posteromedial part of the bony tunnel was resected and the TPT was released, inspected, and debrided (Figure 5). The TPT moved freely in its groove with no tendency to luxation. Bone wax was pressed into cancellous bone to prevent recurrence of the bony tunnel. The patient had a walking brace for 3 weeks and functional rehabilitation was started a few days after the surgery. At 6 weeks, the patient could walk with normal shoes and he was able to return to work after 3 months. At last follow-up (12

Figure 2. Ax CT scan, showing the tibialis posterior tendon (*) in a medial retromalleolar bone tunnel and the healed posterior malleolus fracture (red line).

Figure 3. Tibialis posterior tendon tenosynovitis (arrow) on fat-saturation gadolinium injected T1-weighted axial MRI.

Acta Orthopaedica 2019; 90 (6): 624–625

625

Figure 4. Tibialis posterior tendon identification above and below the medial malleolus (arrows), postero-medial part of the bony tunnel (*). (Right ankle, postero-medial approach, patient in supine position.)

Figure 5. Tibialis posterior tendon debridement after resection of the retro-malleolar bone tunnel. (Right ankle, postero-medial approach, patient in supine position.)

months), the patient had no pain and had returned to sport without physical limitation.

tenosynovitis. We found 1 similar case in the literature but the entrapment was not circumferential and it concerned a medial malleolus fracture treated nonoperatively (Khamaisy et al. 2012). The treatment and the outcomes were similar in each case, both patients returning fully to their former activities.

Discussion Tendon entrapment in bony callus is a rare complication of closed-reduction fracture management. Tendon is usually trapped directly in the fracture preventing its anatomical reduction, but it can also be engulfed in the growing osseous callus (Christodoulou et al. 2005, Erra et al. 2013). While displaced medial and lateral malleolus fractures are often operated on, allowing the diagnosis of the tendon entrapment, posterior malleolus fractures are often neglected or fixed with anterior to posterior screws through a percutaneous approach (Solan and Sakellariou 2017). Internal fixation seems recommended for posterior malleolus fractures involving more than 25% of the articular surface to achieve anatomical reduction (Gardner et al. 2011, Mingo-Robinet et al. 2011). Surgery via a postero-lateral or postero-medial approach allows for anatomical reduction and direct control of tendon and soft tissue entrapment, and thus reduces the risk of malunion. Recent research articles showed good results in patients with posterior malleolus synthesis by screw or buttress plate, without increasing the complication rate due to the postero-lateral approach (Verhage et al. 2016, Bali et al. 2017, Gougoulias and Sakellariou 2017). Structure entrapment is better known after upper limb fractures. Tendon entrapment has been reported rate in 1.3% of distal radius fractures involving particularly the extensor tendon and sometimes flexor tendon (Okazaki et al. 2009). Peripheral nerves can also be engulfed in fracture callus (Erra et al. 2013). In our case, the TPT retromalleolar groove was closed by the posterior malleolus fracture’s bony callus, but with no symptoms for almost 10 years. It probably became painful due to a conflict within the inextensible groove, resulting in a painful

TA did the research, wrote the manuscript and operated on the patient. BB did the follow-up. MS and NSN prepared the illustrations and translated the manuscript; CM and PM revised the manuscript. Acta thanks Mikko Ovaska for help with peer review of this study.

Bali N, Aktselis I, Ramasamy A, Mitchell S, Fenton P. An evolution in the management of fractures of the ankle: safety and efficacy of posteromedial approach for Haraguchi type 2 posterior malleolar fractures. Bone Joint J 2017; 99-B(11): 1496-501. Christodoulou A, Givissis P, Mavromatis I, Karkavelas G, Pournaras J. Fracture callus engulfing a peripheral nerve does not affect its function: an experimental study in rabbits. Clin Orthop Rel Res 2005; (433): 195-204. Erra C, Granata G, Liotta G, Podnar S, Giannini M, Kushlaf H, Hobson-Webb L D, Leversedge F J, Martinoli C, Padua L. Ultrasound diagnosis of bony nerve entrapment: case series and literature review. Muscle Nerve 2013; 48(3): 445-50. Gardner M J, Streubel P N, McCormick J J, Klein S E, Johnson J E, Ricci W M. Surgeon practices regarding operative treatment of posterior malleolus fractures. Foot Ankle Int 2011; 32(4): 385-93. Gougoulias N, Sakellariou A. When is a simple fracture of the lateral malleolus not so simple? How to assess stability, which ones to fix and the role of the deltoid ligament. Bone Joint J 2017; 99-B(7): 851-5. Khamaisy S, Leibner E D, Elishoov O. Tibialis posterior entrapment: case report. Foot Ankle Int 2012; 33(5): 441-3. Mingo-Robinet J, López-Durán L, Galeote J E, Martinez-Cervell C. Ankle fractures with posterior malleolar fragment: management and results. J Foot Ankle Surg 2011; 50(2): 141-5. Okazaki M, Tazaki K, Nakamura T, Toyama Y, Sato K. Tendon entrapment in distal radius fractures. J Hand Surg Eur 2009; 34(4): 479-82. Solan M C, Sakellariou A. Posterior malleolus fractures: worth fixing. Bone Joint J 2017; 99-B(11): 1413-19. Verhage S M, Boot F, Schipper I B, Hoogendoorn J M. Open reduction and internal fixation of posterior malleolar fractures using the posterolateral approach. Bone Joint J 2016; 98-B(6): 812-17.

626

Acta Orthopaedica 2019; 90 (6): 626

Erratum Low revision rate despite poor functional outcome after stemmed hemiarthroplasty for acute proximal humeral fractures: 2,750 cases reported to the Danish Shoulder Arthroplasty Registry Alexander AMUNDSEN et al. Correspondence: alexander.amundsen.03@regionh.dk Acta Orthopaedica 2019; 90 (3): 196â&#x20AC;&#x201C;201. DOI 10.1080/17453674.2019.1597491

Results ..... WOOS Erronous text: ..... A WOOS score below 30 and 50 was reported in 303 (11%) and 676 (25%) patients, respectively (Figure 2). These percentages have been calculated using the total number of cases in the study (2,750) and not by using the number of patients whom answered WOOS (1,525). The correction does not affect other sections of the article or change the conclusion. The corrected text should be: ..... A WOOS score below 30 and 50 was reported in 303 (20%) and 676 (44%) patients, respectively (Figure 2).

ÂŠ 2019 The Author(s). Published by Taylor & Francis on behalf of the Nordic Orthopedic Federation. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Non-Commercial License (https://creativecommons.org/licenses/by/4.0) DOI 10.1080/17453674.2019.1676517

6/19 ACTA ORTHOPAEDICA

Element of success in joint replacement

www.heraeus-medical.com

Vol. 90, No. 6, 2019 (pp. 507–626)

Volume 90, Number 6, December 2019

IORT_90_06_COVER.indd 1

17-10-2019 10:22:33