Acta Orthopaedica, Volume 91, Issue 4 by Acta

4/20 ACTA ORTHOPAEDICA

LONGER IMPLANT SURVIVAL. WITH THE RIGHT BONE CEMENT.

23%

lower revision risk* with PALACOSÂ® R+G compared to other bone cements

* Calculated difference of cumulated revision rates in hip arthroplasty at 13 years of implantation

www.heraeus-medical.com

10034

NJR Data Supplier Feedback (summary reports); Cumulative revision rates (2007â&#x20AC;&#x201C;2020) status February 2020. Current report accessible at http://herae.Us/njr-data We thank the patients and staff of all the hospitals in England, Wales, Northern Ireland and the Isle of Man who have contributed data to the National Joint Registry. We are grateful to the Healthcare Quality Improvement 3DUWQHUVKLS +4,3 WKH 1-5 6WHHULQJ &RPPLWWHH DQG VWDII DW WKH 1-5 &HQWUH IRU IDFLOLWDWLQJ WKLV ZRUN 7KH YLHZV H[SUHVVHG UHSUHVHQW WKRVH RI +HUDHXV 0HGLFDO *PE+ DQG GR QRW QHFHVVDULO\ UHÆ&#x192;HFW WKRVH RI WKH 1DWLRQDO -RLQW Registry Steering Committee or the Health Quality Improvement Partnership (HQIP) who do not vouch for how the information is presented.

Vol. 91, No. 4, 2020 (pp. 365â&#x20AC;&#x201C;499)

The element of success in joint replacement

Volume 91, Number 4, August 2020

2006_19476_AZ_PALACOS_EOS_2.0_Acta_Orthop_Europe_215x280_EN.indd 1 TF-IORT200094.indd 1

14.07.20 13:39

26-08-2020 14:56:40

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

E DITORIAL O F FICE

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

EDITOR

THE FOUNDATION BOARD OF

Anders Rydholm Lund, Sweden

THE NORDIC O RTHOPAEDIC F EDERATION AND A CTA O RTHOPAEDICA

DEPUTY EDITOR

Peter A Frandsen Odense, Denmark CO-EDITORS

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

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

WEB EDITOR

Magnus Tägil Lund, Sweden S TATISTICAL EDITOR

Jonas Ranstam Lund, Sweden P RODUCTION MANAGER

Kaj Knutson Lund, Sweden

Vol. 91, No. 4, 2020

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

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

Annual Institutional Subscription, Volume 91, 2020

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

$1,291

£798

€1,035

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

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

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

Acta Orthopaedica

ISSN 1745-3674

Vol. 91, No. 4, August 2020 Hip arthroplasty Complications and re-revisions after revisions of 528 metal-onmetal hips because of adverse reaction to metal debris Patient-acceptable symptom state for the Oxford Hip Score and Forgotten Joint Score at 3 months, 1 year, and 2 years following total hip arthroplasty: a registry-based study of 597 cases Repeated cobalt and chromium ion measurements in patients with bilateral large-diameter head metal-on-metal ReCap-M2A-Magnum total hip replacement Long-term outcomes of the hip shelf arthroplasty in adolescents and adults with residual hip dysplasia: a systematic review Effective radiation dose in radiostereometric analysis of the hip Hip fracture No association between waiting time to surgery and mortality for healthier patients with hip fracture: a nationwide Swedish cohort of 59,675 patients Is there a reduction in risk of revision when 36-mm heads instead of 32 mm are used in total hip arthroplasty for patients with proximal femur fractures? A matched analysis of 5,030 patients with a median of 2.5 years’ follow-up between 2006 and 2016 in the Nordic Arthroplasty Register Association Mortality and revision rate of cemented and uncemented hemiarthroplasty after hip fracture: an analysis of the Dutch Arthroplasty Register (LROI) Knee Increases in the rates of primary and revision knee replacement are reducing: a 15-year registry study across 3 continents A matched comparison of revision rates of cemented Oxford Unicompartmental Knee Replacements with Single and Twin Peg femoral components, based on data from the National Joint Registry for England, Wales, Northern Ireland and the Isle of Man Higher risk of revision for partial knee replacements in low absolute volume hospitals: data from 18,134 partial knee replacements in the Dutch Arthroplasty Register Why are patients still in hospital after fast-track, unilateral unicompartmental knee arthroplasty Ankle, foot Socioeconomic position is associated with surgical treatment of open fractures of the lower limb: results from a Swedish population-based study 12-year survival analysis of 322 Hintegra total ankle arthroplasties from an independent center Effectiveness of hallux valgus surgery on patient quality of life: a systematic review and meta-analysis Slipped capital femoral epiphysis Fate of patients with slipped capital femoral epiphysis (SCFE) in later life: risk of obesity, hypothyroidism, and death in 2,564 patients with SCFE compared with 25,638 controls Amputation SwedeAmp—the Swedish Amputation and Prosthetics Registry: 8-year data on 5762 patients with lower limb amputation show sex differences in amputation level and in patient-reported outcome Tumor Can MRI differentiate between atypical cartilaginous tumors and high-grade chondrosarcoma? A systematic review

365

O Lainiala, A Reito, J Nieminen, and A Eskelinen

372

V P Galea, L H Ingelsrud, I Florissi, D Shin, C R Bragdon, H Malchau, K Gromov, and A Troelsen

378

S Pietiläinen, H Mäntymäki, T Vahlberg, A Reito, A Eskelinen, P Lankinen, and K Mäkelä

383 390

K Willemsen, C J Doelman, A S Y Sam, P R Seevinck, R J B Sakkers, H Weinans, and B C H van der Wal I F Blom, L A Koster, B Ten Brinke, and N M C Mathijssen

396

K Greve, K Modig, M Talbäck, E Bartha, and M Hedström

401

G Tsikandylakis, J N Kärrholm, G Hallan, O Furnes, A Eskelinen, K Mäkelä, A B Pedersen, S Overgaard, and M Mohaddes

408

B J Duijnisveld, K L M Koenraadt, L N van Steenbergen, and S B T Bolder

414

P L Lewis, S E Graves, O Robertsson, M Sundberg, E W Paxton, H A Prentice, and A W-Dahl H R Mohammad, G S Matharu, A Judge, and D W Murray

420

426

I van Oost, K L M Koenraadt, L N van Steenbergen, S B T Bolder, and R C I van Geenen

433

C B Jensen, A Troelsen, C S Nielsen, N K S Otte, H Husted, and K Gromov

439

Y Granberg, K T Lundgren, and E K Lindqvist

444 450

M J Zafar, T Kallemose, M Benyahia, L B Ebskov, and J Ø Penny L E Hernández-Castillejo, V Martínez Vizcaíno, M GarridoMiguel, I Cavero-Redondo, D P Pozuelo-Carrascosa, and C Álvarez-Bueno

457

Y D Hailer

464

I Kamrad, B Söderberg, H Örneholm, and K Hagberg

471

C Deckers, M J Steyvers, G Hannink, H W B Schreuder, J W J de Rooy, and I C M van der Geest

Evaluation RCT Assessing variability and uncertainty in orthopedic randomized controlled trials Case report, Technical note External iliac artery injury following total hip arthroplasty via the direct anterior approachâ&#x20AC;&#x201D;a case report New 3-dimensional implant application as an alternative to allograft in limb salvage surgery: a technical note on 10 cases Correspondence RSA of the Symax hip stem Information to authors (see http://www.actaorthop.org/)

479

L Raittio and A Reito

485

E Burlage, J G Gerbers, B R H Geelkerken, and W C Verra

489

J W Park, H G Kang, J H Kim, and H-S Kim

497

H T Aro and S Nazari-Farsani versus D S M G Kruijntjens, L Koster, B L Kaptein, J J C Arts, and R H M ten Broeke

Acta Orthopaedica 2020; 91 (4): 365–371

365

Complications and re-revisions after revisions of 528 metal-on-metal hips because of adverse reaction to metal debris Olli LAINIALA, Aleksi REITO, Jyrki NIEMINEN, and Antti ESKELINEN

Coxa Hospital for Joint Replacement and Faculty of Medicine and Health Technologies, Tampere University, Tampere, Finland Correspondence: olli.lainiala@tuni.fi Submitted 2019-06-25. Accepted 2020-02-09

Background and purpose — There is limited amount of evidence about optimal revision indications, technique, and implants when performing revision surgery for metal-onmetal (MoM) hip replacements due to adverse reaction to metal debris (ARMD). We assessed which factors are related to re-revisions and complications after a revision of MoM hip arthroplasty because of ARMD. We also aimed to provide information on optimal implants for these revisions. Patients and methods — 420 MoM total hip arthroplasties (THA) and 108 MoM hip resurfacings were implanted and later revised at our institution. We used Cox regression to analyze the factors associated with re-revisions and complications after a revision for ARMD. Results — A re-revision was performed on 27 THAs (6%) and 9 resurfacings (8%). The most common indication for re-revision was recurrent dislocation (20 hips, 4%). Complications not leading to re-revision were seen in 21 THAs (5%) and 6 resurfacings (6%). The most common complication was dislocation treated with closed reduction in 13 hips (2%). Use of revision head size > 36mm was associated with decreased risk for dislocations. Presence of pseudotumor, pseudotumor grade, pseudotumor size, or the choice of bearing couple were not observed to affect the risk for re-revision. Non-linear association was observed between preoperative cobalt and risk for re-revision. Interpretation — As dislocation was the most frequent post-revision complication, large head sizes should be used in revisions. Because size or type of pseudotumor were not associated with risk of re-revision, clinicians may have to reconsider, how much weight is put on the imaging findings when deciding whether or not to revise. In our data blood cobalt was associated with risk for re-revision, but this finding needs further assessment.

Adverse reaction to metal debris (ARMD) is the most common reason for failure of metal-on-metal (MoM) hip replacements (Australian Orthopaedic Association [AOA] 2018, National Joint Registry [NJR] 2018). Despite the large number of revisions, there is no consensus when a MoM hip should be revised for ARMD, how extensively debridement should be done and which implants to use (Matharu et al. 2018a). In the initial studies describing the revisions of MoM hips for ARMD the re-revision and complication rates of MoM hip revisions were high, especially in those revised for pseudotumors (Grammatopoulos et al. 2009, de Steiger et al. 2010). This led to recommending early revisions for ARMD to prevent additional tissue damage (Haddad et al. 2011, Medicines and Healthcare products Regulatory Agency [MHRA] 2012, Hannemann et al. 2013). Recently a National Joint Registry (NJR) based study reported increased risk of revision for high BMI, head and liner only revision, ceramic-on-ceramic (CoC) bearing surface, and acetabular bone grafting (Matharu et al. 2017b). Many of the factors associated with complications and re-revision after ARMD revisions, such as blood metal ion levels and cross-sectional imaging, cannot be analyzed from registry data, and studies about these have been called out by a recent review article (Matharu et al. 2018a). Several guidelines exist concerning criteria for consideration of MoM hip revisions (Hannemann et al. 2013, MHRA 2017, US Food and Drug Administration [FDA] 2019). Most guidelines advise against revising those with normal imaging findings, low blood metal ion levels, and minor symptoms, and suggest considering revisions in those with major symptoms, large soft tissue lesions, and extremely high blood metal ion levels. Between these extremities falls the “gray area of ARMD,” for which cases the guidelines suggest individual

© 2020 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group, on behalf of the Nordic Orthopedic Federation. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. DOI 10.1080/17453674.2020.1748351

366

evaluation. We aimed to provide information to help decisionmaking with these patients. In this study, we report the reasons for re-revision and complications after revisions of MoM hip due to ARMD. We also analyze the risk factors for re-revision and complications. We hypothesized that large solid/mixed type pseudotumors and high blood metal ion levels would be associated with poor revision results, and that use of large head size would decrease the risk for post-revision complications. We also were interested in whether some bearing surface(s) performed better than others, or whether there is a difference between liner-only and cup revision, as those are variables that surgeons can choose.

Patients and methods We searched all the patients with primary MoM hip replacement implanted and later revised due to ARMD at our institution. We included all patients who were revised a minimum of 1 year before the data collection in September 2018 and had pre-revision cobalt (Co) and chromium (Cr) measurements and imaging available for the diagnosis of ARMD. The primary surgery (October 2001–August 2011) as well as revision surgery (September 2010–September 2017) of all the patients included was performed at our high-volume center. No referral patients were included. After the recall of Articular Surface Replacement (ASR, Depuy Orthopaedics, Warsaw, IN, USA) (Depuy Orthopaedics 2010), we established an intensified screening program for all ASR MoM hip replacements at our institution. Screening included whole blood Co and Cr measurements, physical examination and cross-sectional imaging with MRI as primary imaging modality. Initially, ultrasound was used only if there were contraindications for MRI, but later the use of ultrasound was increased due to lower cost. If the patient had progressive symptoms or elevating whole-blood metal ion levels, imaging was repeated. MoM hip replacement brands other than ASR were also included in the screening from 2012, but crosssectional imaging was performed only for the patients with symptoms or elevating whole-blood Co or Cr levels. Whole-blood Co and Cr measurements were performed with dynamic reaction cell inductively coupled plasma mass spectrometry (Agilent 7500cx, Agilent Technologies, Santa Clara, CA, USA). Preoperative MRIs were performed with Siemens Magnetom Avanto 1.5 T (Siemens Healthcare, Erlangen, Germany) or GE Signa HD 1.5 T (General Electric Healthcare, Waukesha, WI, USA) and ultrasound examinations were performed with Logiq E9 (GE Healthcare, USA). Imaging findings were graded according to a previously described system by musculoskeletal radiologists (Matthies et al. 2012). Revision surgery for ARMD was considered if (1) a thickwalled pseudotumor with atypical contents or solid-type pseudotumor was seen in imaging, or the patient had (2) elevated whole blood Co or Cr levels and hip symptoms despite normal

Acta Orthopaedica 2020; 91 (4): 365–371

imaging finding, or (3) continuously symptomatic hip or progressive symptoms regardless of normal imaging findings or whole blood metal concentrations, or (4) the patient had progressively increasing whole-blood metal ion levels, even without symptoms or findings in cross-sectional imaging. Co and Cr were considered as elevated if they exceeded 5 µg/L (Hart et al. 2011). The diagnosis of ARMD was based on intraoperative findings irrespective of preoperative working diagnosis. Failure was classified as being due to ARMD if metallosis was present or there was macroscopic synovitis in the joint, and/or a pseudotumor was found during revision and perioperatively there was no evidence of component loosening or periprosthetic fracture. Infection was ruled out by at least 5 bacterial cultures obtained during revision surgery. Histopathological samples were collected during revision to further support the validity of the intraoperative diagnosis. Revision was defined as surgery including a change of at least 1 component (stem, head, liner, or/and cup). In revision surgery of THA, if the stem was well fixed and correctly positioned it was retained and only the cup or liner revised. In resurfacing revisions, the femoral neck was cut and a stem implanted. The revision implants were chosen based on the surgeon’s preference. If a ceramic head was used, a titanium sleeve adapter was applied, even if there were no signs of corrosion in the taper. In 8 revisions with an unusually large pseudotumor extending into the intrapelvic region, additional resection of the intrapelvic pseudotumor was performed through an ilioinguinal approach to complement resection from posterior approach. After the revision, anteroposterior and lateral plain radiographs of the hip and anteroposterior pelvic radiographs were obtained, blood Co and Cr measurements were performed at 2, 6, and 12 months, and the patient was clinically evaluated by an orthopedic surgeon at 2 and 12 months, and thereafter at 2-year intervals. Acetabular inclination was measured from anteroposterior plain radiographs using ischial tuberosities as reference, and anteversion was measured from cross-table radiographs using the horizontal plane as a reference. Statistics Means (SD) are presented for normally distributed variables, and medians (ranges) for variables with non-Gaussian distribution. In Kaplan–Meier survival analysis the results are reported till the year with at least 20 hips at risk. A Cox regression model was used to analyze factors associated with rerevision, and proportional hazards assumption was analyzed using Schoenfeld’s residuals. No violation of proportional hazards assumption was met. Directed acyclic graphs (DAG) were used to guide the selection of variables for the model (Shrier and Platt 2008). As we had several variables of interest (revision head size, bearing surface, pseudotumor, type of revision, and pre-revision Co), we created DAGs for each variable to ensure appropriateness of multivariable analysis. Co level was analyzed as a linear variable, but also nonlinear

Acta Orthopaedica 2020; 91 (4): 365–371

Primary MoM hip replacement implanted October 2001 – August 2011 in Coxa Hospital n = 3,013 (2,520 patients) MoM hip revisions by September 2018 n = 793 Excluded (n = 183): – MoM hips implanted initially in a revision surgery, 70 – revisions performed after September 2017, 28 – revisions performed before systematic screening a, 85 MoM hip revisions with screening data and over 1 year follow-up n = 610 Excluded – revision for other reasons than ARMD (n = 82): – infection, 21 – aseptic loosening of the cup, 13 – periprosthetic fracture, 10 – aseptic loosening of the stem, 9 – femoral neck fracture, 9 – unexplained pain, 4 – impingment, 3 – aseptic loosening of resurfacing head, 3 – trochanteric fracture, 2 – malposition of the cup, 2 – recurrent dislocation, 2 – acetabular fracture, 2 – osteolysis (not yet fractured), 1 – rash (assumed as metal allergy), 1 MoM hip replacements revised for ARMD n = 528 (466 patients)

Figure 1. Flow chart of patient inclusion. MoM = metal-on-metal; Co = cobalt; Cr = chromium; ARMD = adverse reaction to metal debris. a No preoperative imaging or Co/Cr measurements.

relationships for Co were investigated fitting restricted cubic splines. Cox regression analysis was done by including prerevision Co level and appropriate covariates based on DAG using restricted cubic spline with 4 knots. HRs were plotted using median Co value as reference and subtracting the constant so that the width of the 95% confidence interval (CI) was zero for reference value. An ANOVA test was used to assess the statistical significance of nonlinearity. IBM SPSS Statistics version 25 (IBM Corp, Armonk, NY, USA) and R v3.2.1 (R Foundation for Statistical Computing, Vienna, Austria) were used for statistical analyses. Ethics, funding, and potential conflicts of interest This study was approved by the local ethics committee (registration IDs R11006 and R11195). This work was supported by the competitive research funds of Pirkanmaa Hospital District, Tampere, Finland, representing governmental funding. Individual potential conflicts of interests: OL: none. AR: Orion LTD, paid lecture. JN: none. AE: Zimmer Biomet, paid lectures; Depuy Synthes and Zimmer Biomet, institutional research support (not related to current study).

367

Table 1. Patients’ pre-revision characteristics Factor

Total hip arthroplasty n = 420

Women, n (%) 230 (55) Age at revision, mean (SD) 67 (8.5) Pre-revision Co, µg/L median (range) 11.8 (0.2–192) (IQR) (7.0–20) Pre-revision Cr, µg/L median (range) 7.9 (0.4–156) (IQR) (2.3–6.5) Any symptoms n (%) 272 (65) Imaged with MRI, n (%) 301 (72) ultrasound, n (%) 119 (28) Head size, median mm (range) 47 (28–59) Years between primary and revision surgery, mean (SD) 6.3 (2.4) BMI, mean (SD) 28 (4.7) ASA class at primary 1/2/3/4/NA 40/219/144/3/14

Hip resurfacing n = 108 72 (67) 57 (9.7) 11.8 (0.4–225) (2.1–36) 7.4 (0.8–125) (2.2–20) 79 (73) 83 (77) 25 (23) 49 (42–63) 6.7 (2.4) 28 (4.5) 42/51/10/2/3

SD = standard deviation; IQR = interquartile range; NA = not available.

Results 3,013 MoM hips in 2,520 patients were identified. By September 2018, 793 revisions of MoM hip replacements had been performed at our institution. As we included only hips revised for ARMD without any other indications for revision, 528 MoM hips in 466 patients were included in this study (Figure 1). 420 (80%) of the implants used in primary surgery of the revised hips were stemmed MoM total hip arthroplasties (THA) and 108 (20%) were hip resurfacings (Table 1). Implants used in primary surgeries are listed in Table 2 and implants used in revision surgeries in Table 3 (Supplementary data). Pre-revision cross-sectional imaging and whole blood Co and Cr measurements were available for all patients included in the study. Median time between the pre-revision imaging and revision surgery was 4.3 months (5 days–24 months). The pre-revision imaging findings are listed in Table 4. The median time between pre-revision Co and Cr measurements and revision was 4.8 months (2 days–19 months). For patients with stemmed MoM THAs, there were 374 cup revisions, 42 liner-only revisions, 1 revision with dual mobility cup system, and in 3 patients with stemmed MoM THA the stem component was also changed (2 cup revisions and 1 liner revision). In all 108 resurfacing revisions, the cup was changed and a stem implanted. Median head size of the revision components was 36 mm (28–48). Median acetabular inclination of the cups implanted at revision was 45° (SD 7) and acetabular anteversion 27° (SD 8). A re-revision requiring a change of any component was performed on 36 hips (36 patients, 27 involving THA [6%], 9 involving resurfacings [8%]). Median time from revision to re-revision was 4.6 months (7 days–7.3 years). There

368

Acta Orthopaedica 2020; 91 (4): 365–371

Table 4. Pre-revision imaging findings with MRI or ultrasound Total hip arthroplasty Hip resurfacing Size, cm Size, cm Grade n (%) median (range) n (%) median (range) 0 1 2A 2B 3

203 (48) 82 (20) 5.7 (1.2–25) 48 (11) 6.4 (3.0–17) 68 (16) 8.4 (3.6–30) 19 (5) 6.0 (2.4–13)

Total 420

50 (46) 22 (20) 8 (7) 25 (23) 3 (3)

4.9 (2.4–10) 7.5 (4.3–11) 7.9 (2.5–19) 10.7 (5.4–13)

108

Imaging findings are classified according to a previously described grading system (Matthies et al. 2012). Grade 0 represents normal imaging finding; grade 1 represents thin-walled fluid-filled pseudotumor; grade 2A a fluid-filled pseudotumor with thick or irregular walls; 2B a pseudotumor with atypical contents; and grade 3 a predominantly solid pseudotumor.

were complications not leading to change of the components involving 27 hips (21 THA [5%], 6 resurfacings [6%]). The median time from revision to complication was 4.8 months (47 days–7.6 years). The reasons for re-revisions and complications are listed in Table 5. Among the group of patients with complications not leading to revision and the re-revision group there was overlapping in 3 cases. Therefore, 60 (11%) hips experienced at least 1 complication that was or was not treated with revision surgery. The median post-revision follow-up time for those in follow-up without revision was 5.2 years (1.3–8.0). 21 hips (20 patients) were lost to follow-up as they either moved outside our hospital district, refused the follow-up or could not be contacted after mean follow-up of 2.7 years (1.1–5.0); all were included in the analyses until the last contact. 12 patients (12 hips) died during the study period (median follow-up 3.8 years [0.2–6.3]) for reasons not related to the prosthesis. The 7-year implant survivorship after revisions of stemmed MoM THA was 94% (CI 91–96) and after revisions of MoM resurfacing 91% (CI 86–97). In Cox regression analysis, none of the variables tested were observed to have statistically significant association with re-revisions (Table 6, Supplementary data). When Co was treated as a nonlinear variable, we observed a non-linear relationship in which pre-revision Co concentration between 20 and 90 µg/L was associated with increased risk for re-revision (Figure 2). As dislocation was the most common reason for re-revision as well as the most common complication, a separate analysis of risk factors for instability was performed. The instability group included those suffering a dislocation not leading to rerevision (13 hips) and those who underwent a re-revision due to recurrent dislocations/instability (20 hips). Thus 33 hips were included in the instability group. In both univariable and multivariable analyses revision head size of 36 mm or smaller was associated with increased risk for dislocation (Table 7, Supplementary data). As our study included patients with bilateral revisions, analyses in Tables 6 and 7 were re-performed with exclusion of

Table 5. Indications for re-revisions and complications after the revisions of metal-on-metal total hip arthroplasty (THA) and resurfacing arthroplasties Resur THA facing Re-revision indication Mechanical reasons Recurrent dislocation or instability after closed reduction 18 2 Periprosthetic femoral fracture 1 3 Acetabular fracture 2 0 Fracture of stem component 0 1 Aseptic loosening of stem 1 1 Non-mechanical reasons Recurrent ARMD 1 1 Infection 4 1 Total 27 9 Complication Mechanical complication Dislocation(s) treated with closed reduction Acetabular fracture a Trochanteric fracture b Fascial rupture c Hernia Non-mechanical complication Residual pseudotumor Pulmonary embolism Deep vein thrombosis Bowel occlusion Total

11 2 3 0 1 0 2 1 1 0 0 1 1 1 21

2 0 1 0 6

ARMD= adverse reaction to metal debris; a non-operative treatment; b operated with titanium plate without change of components; c re-suture without change of components.

Log Relative Hazard 3

–1

–2

–3

100

150

Pre-revision whole blood Co concentration (µg/L)

Figure 2. Non-linear association was seen between prerevision blood cobalt (Co) concentration and risk for revision. Preoperative Co between approximately 20 and 90 µg/L was associated with increased risk for re-revision. Hazard ratios are presented in relation to median Co of 11.8 µg/L. Gray area represents 95% confidence interval.

left hips of bilateral patients to rule out potential bias by clustered observations. This did not change the results (analyses not shown).

Acta Orthopaedica 2020; 91 (4): 365–371

Discussion A large number of revisions have been performed on MoM hip replacements due to ARMD all over the world (NJR 2018, AOA 2018); however, there is still a limited amount of information regarding on which patients the revision should be performed. We aimed to contribute to the understanding concerning revisions of MoM hips due to ARMD. A strength of our study is our unselected patient group, as our hospital is a primary center and all the primary surgeries and revisions were performed at our institution. Therefore, referral patients do not cause selection bias in this study. The revision threshold may have been lower in our institution compared with some other institutions, and our previously reported survivals are lower compared with implant registries, possibly due to higher proportion of ASR MoM hip replacements (Lainiala et al. 2019), which should be accounted for when comparing these results with results from other centers. Our study has some limitations. 1st, as this is retrospective study and patients from several surgeons were included, there might have been variation in the threshold for revision surgery and the surgical technique (for example the extent of tissue resection and choice of implants used). Further, there were several brands of stemmed MoM THAs and MoM hip resurfacings as well as revision implant brands used. For studying the clinical significance of preoperative factors, the use of a single revision implant brand would have been the best option. On the other hand, using similar numbers of metal-on-polyethylene (MoP), ceramic-on-ceramic (CoC), and ceramic-onpolyethylene (CoP) bearing surfaces with a single cup brand in a randomized setting would have allowed us to compare the performance of the different revision bearing surfaces. 2nd, the amount of re-revisions and complications is only 1 perspective. Oxford Hip Scores have also been registered at our institution, but we decided to concentrate on the objective endpoints in this study, and the patient-reported outcome measures will be reported in future. We observed re-revisions requiring change of any component in 36 hips (7%) and a complication not requiring change of components in 27 (5%) hips. The 7-year implant survivorship of 94% after revision of stemmed THAs and 91% after revision of hip resurfacings is better than the overall 7-year re-revision rate of 14.2% for all bearing surfaces and indications in the NJR registry (NJR 2018). 1 of the earliest hip resurfacing cohorts that raised concerns about the complications after ARMD revisions has reached follow-up of 10 years and a re-revision rate of 38% was reported (Grammatopoulos et al. 2009, Matharu et al. 2017c). Poor early results led several authors to recommend early intervention (Grammatopoulos et al. 2009, De Smet et al. 2011, Su and Su 2013). Since then, re-revision rates between 0% and 4% and complication rates between 0% and 10% have been reported for revisions of MoM hip resurfacings (Eswaramoorthy et al. 2009, De Smet

369

et al. 2011, Gross and Liu 2014), and 8-22% re-revision rates and 8-38% complication rates for MoM THAs (Munro et al. 2014, Stryker et al. 2015, Wyles et al. 2014, van Lingen et al. 2015, Jennings et al. 2019). Many cohorts have described dislocation as a common problem after a revision of MoM hip replacement (Grammatopoulos et al. 2009, De Smet et al. 2011, Munro et al. 2014, Stryker et al. 2014, van Lingen et al. 2015, Jennings et al. 2019). Despite the early warnings about poor results of ARMD revisions, a recent study from the NJR registry showed that the re-revision rates were lower in MoM hips revised for ARMD compared with those revised for nonARMD (Matharu et al. 2018b). A larger head size has been suggested to decrease the risk for poor outcome (Matharu et al. 2019), which is in line with our result: head sizes larger than 36 mm decreased the risk for dislocation. A recent study reported solid pseudotumors with abductor deficiency to be associated with post-revision complications (Liow et al. 2016), but neither our study nor a study by Matharu et al. (2019) found evidence of an association between cross-sectional imaging findings and revision results. Guidelines have put weight on the type and the size of the soft tissue abnormalities when considering revision (Hannemann et al. 2013, MHRA 2017, FDA 2019), but it seems that mixed or solid-type pseudotumors do not necessarily cause a high risk of complications after revision. Of course, pseudotumors are not the only type of lesions related to ARMD and muscle deficiency and osteolysis need to be considered. A CoC bearing surface is reported to be associated with risk for poor outcome in 2 recent British studies (Matharu et al. 2017b, 2019), but neither our study nor an Australian registry-based study (Wong et al. 2015) found a difference between different bearing surfaces used for MoM revisions. Ceramic heads are used at our institution to minimize metal release from the trunnion– taper junction, and nowadays our bearing surface of choice is CoP, as use of CoC is associated with occasional squeaking (McDonnell et al. 2013, Salo et al. 2017). Using CoP bearings with head size > 36 mm may lead to a very thin polyethylene liner, and we certainly try to avoid this—especially if the patient is young and active. Currently, our policy is to use a CoP bearing mainly with a 36 mm head. In patients with a very large cup size that allows usage of > 36 mm heads with adequate thickness of the polyethylene liner, > 36 mm heads can be considered. However, if satisfactory stability cannot be achieved with the CoP bearing, then we would choose either a constrained liner or dual mobility bearing. Jennings et al. (2019) reported higher median Co and Cr for patients with post-revision complications compared with those without complications. In our study, the association was non-linear, and only Co 20–90 μg/L was associated with an increased risk for re-revision. A few recent studies observed no association with preoperative metal ion levels and poor outcome (Liow et al. 2016, Matharu et al. 2019). The possible association between whole-blood metal ion concentrations and revision results is clearly complex, needs further investigation

370

and no single metal ion value can be given as a threshold for revision. An explanation as to why higher Co values were not associated with increased risk might be that extremely high Co levels have led to revision with lesser imaging findings and symptoms compared with only slightly or moderately elevated whole blood Co levels. Recent studies (Matharu et al. 2017a, 2017b, 2019) observed increased risk for poor outcome in patients with selective component revision (some of the components retained). We did not observe a difference between THAs treated with head and liner exchange, and those with the cup revised. Therefore, we still consider head and liner revision to be a viable option in a subset of patients with a well-fixed and positioned modular cup. Conclusion Dislocation is the most frequent post-revision complication after ARMD revisions and using larger head sizes than 36 mm decreases the risk for dislocation. Neither the size nor the grade of pseudotumor were associated with the outcome of revision, but this should be further evaluated with inclusion of other variables describing tissue damage. We recommend using a CoP bearing with as large a head size as feasible and choosing either a constrained liner or dual mobility bearing if satisfactory stability cannot be reached with the CoP bearing. Supplementary data Tables 2, 3, 6, and 7 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/ 17453674.2020.1748351

The authors would like to thank Ella Lehto RN, and Heli Kupari RN, Coxa Hospital for Joint Replacement, Tampere, Finland, for maintaining their study database. They also thank the orthopedic surgeons from Coxa Hospital who performed the surgeries analyzed in this study. They thank the musculoskeletal radiologists who performed the ultrasounds and analyzed MRI studies. The authors also thank the physiotherapists from Coxa Hospital. OL: study design, literature search, data collection and analysis, interpretation of data, statistics, writing and revision of the manuscript, and final approval. AR: study design, statistical analysis, interpretation of data, writing and revising the manuscript, and final approval. JN and AE: study design, interpretation of data and statistics, writing and revision of the manuscript, and final approval. Acta thanks Richard N de Steiger and Wierd P Zijlstra for help with peer review of this study.

Australian Orthopaedic Association, National Joint Replacement Registry. Annual Report 2018. https://www.aoa.org.au/docs/default-source/annualreports/aoa-annual-report-2017-18_web.pdf?sfvrsn=c6dec704 (accessed November 4, 2019). 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): 259-69. de Steiger R N, Miller L N, Prosser G H, Graves S E, Davidson D C, Stanford T E. Poor outcome of revised resurfacing hip arthroplasty. Acta Orthop 2010; 81(1): 72-6.

Acta Orthopaedica 2020; 91 (4): 365–371

Depuy Orthopaedics. ASR recall; 2016. https://www.jnjmedicaldevices.com/ en-US/news-events/depuy-orthopaedics-voluntarily-recalls-asrtm-hip-system-depuy (accessed November 4, 2019). Eswaramoorthy V K, Biant L C, Field R E. Clinical and radiological outcome of stemmed hip replacement after revision from metal-on-metal resurfacing. J Bone Joint Surg Br 2009; 91(11): 1454-8. Grammatopoulos G, Pandit H, Kwon Y, 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. Gross T P, Liu F. Outcomes after revision of metal-on-metal hip resurfacing arthroplasty. J Arthroplasty 2014; 29(9 Suppl.): 219-23. Haddad F S, Thakrar R R, Hart A J, Skinner J A, Nargol A V F, Nolan J F, Gill H S, Murray D W, Blom A W, Case C P. Metal-on-metal bearings: the evidence so far. J Bone Joint Surg Br 2011; 93(5): 572-9. Hannemann F, Hartmann A, Schmitt J, Lutzner J, Seidler A, Campbell P, Delaunay C P, Drexler H, Ettema H B, Garcia-Cimbrelo E, Huberti H, Knahr K, Kunze J, Langton D J, Lauer W, Learmonth I, Lohmann C H, Morlock M, Wimmer M A, Zagra L, Gunther K P. European multidisciplinary consensus statement on the use and monitoring of metal-on-metal bearings for total hip replacement and hip resurfacing. Orthop Traumatol Surg Res 2013; 99(3): 263-71. Hart A J, Sabah S A, Bandi A S, Maggiore P, Tarassoli P, Sampson B, Skinner J A. 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. Jennings J M D P T, White S, Martin J R, Yang C C, Miner T M, Dennis D A. Revisions of modular metal-on-metal THA have a high risk of early complications. Clin Orthop Relat Res 2019; 477(2): 344-50. Lainiala O S, Reito A P, Nieminen J J, Eskelinen A P. Declining revision burden of metal-on-metal hip arthroplasties. J Arthroplasty 2019; 34(9): 2058,2064.e1. Liow M H L, Dimitriou D, Tsai T, Kwon Y. Preoperative risk factors associated with poor outcomes of revision surgery for “pseudotumors” in patients with metal-on-metal hip arthroplasty. J Arthroplasty 2016; 31(12): 2835-42. Matharu G S, Judge A, Murray D W, Pandit H G. Outcomes following revision surgery performed for adverse reactions to metal debris in non-metalon-metal hip arthroplasty patients: analysis of 185 revisions from the national joint registry for England and Wales. Bone Joint Res 2017a; 6 (7): 405-13. Matharu G S, Judge A, Pandit H G, Murray D W. Which factors influence the rate of failure following metal-on-metal hip arthroplasty revision surgery performed for adverse reactions to metal debris? An analysis from the National Joint Registry for England and Wales. Bone Joint J 2017b; 99-B(8): 1020-7. Matharu G S, Pandit H G, Murray D W. Poor survivorship and frequent complications at a median of 10 years after metal-on-metal hip resurfacing revision. Clin Orthop Relat Res 2017c; 475 (2): 304-14. 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, Judge A, Murray D W, Pandit H G. Outcomes after metal-onmetal hip revision surgery depend on the reason for failure: a propensity score-matched study. Clin Orthop Relat Res 2018b; 476 (2): 245-58. 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-on-metal hip arthroplasty patients with adverse reactions to metal debris: a retrospective cohort study of 346 revisions. J Arthroplasty 2019; 34(7): 1483-91. Matthies A K, Skinner J A, Osmani H, Henckel J, Hart A J. Pseudotumors are common in well-positioned low-wearing metal-on-metal hips. Clin Orthop Relat Res 2012; 470(7): 1895-906. McDonnell S M, Boyce G, Bare J, Young D, Shimmin A J. The incidence of noise generation arising from the large-diameter Delta Motion ceramic total hip bearing. Bone Joint J 2013; 95-B(2): 160-5.

Acta Orthopaedica 2020; 91 (4): 365â&#x20AC;&#x201C;371

Medicines and Healthcare products Regulatory Agency (MHRA). Medical device alert: All metal-on-metal (MoM) hip replacements: updated advice for follow-up (MDA/2017/018); 2017. https://Assets.publishing.service. gov.uk/media/5954ca1ded915d0baa00009b/MDA-2017-018_Final.pdf (accessed November 4, 2019). Medicines and Healthcare products Regulatory Agency (MHRA). Medical device alert: All metal-on-metal (MoM) hip replacements (MDA/2012/036); 2012. http://www.mhra.gov.uk/home/groups/dts-bs/ documents/medicaldevicealert/con155767.pdf (accessed November 4, 2019). Munro J T, Masri B A, Duncan C P, Garbuz D S. High complication rate after revision of large-head metal-on-metal total hip arthroplasty. Clin Orthop Relat Res 2014; 472(2): 523-8. doi: 10.1007/s11999-013-2979-6. National Joint Registry for England, Wales, Northern Ireland and the Island of Man. Annual Report; 2018. https://reports.njrcentre.org.uk/Portals/0/PDFdownloads/NJR%2015th%20Annual%20Report%202018.pdf (accessed November 4, 2019). Salo P P, Honkanen P B, Ivanova I, Reito A, Pajamaki J, Eskelinen A. High prevalence of noise following delta ceramic-on-ceramic total hip arthroplasty. Bone Joint J 2017; 99-B(1): 44-50.

371

Shrier I, Platt R W. Reducing bias through directed acyclic graphs. BMC Med Res Methodol 2008; 8: 70. Stryker L S, Odum S M, Fehring T K, Springer B D. Revisions of monoblock metal-on-metal THAs have high early complication rates. Clin Orthop Relat Res 2015; 473(2): 469-74. doi: 10.1007/s11999-014-3791-7. Su E P, Su S L. Surface replacement conversion: tesults depend upon reason for revision. Bone Joint J 2013; 95-B(11 Suppl. A): 88-91. US Food and Drug Administration (FDA). Metal-on-metal hip implants: information for Orthopaedic Surgeons. https://www.fda.gov/medicaldevices/metal-metal-hip-implants/information-orthopaedic-surgeons. 2019 (accessed November 4, 2019). van Lingen C P, Ettema H B, Bosker B H, Verheyen C C P M. Revision of a single type of large metal head metal-on-metal hip prosthesis. Hip Int 2015; 25(3): 221-6. Wong J M, Liu Y L, Graves S, de Steiger R. What is the re-revision rate after revising a hip resurfacing arthroplasty? Analysis from the AOANJRR. Clin Orthop Relat Res 2015; 473(11): 3458-64. Wyles C C, Van Demark R E 3rd, Sierra R J, Trousdale R T. High rate of infection after aseptic revision of failed metal-on-metal total hip arthroplasty. Clin Orthop 2014; 472(2): 509-16.

372

Acta Orthopaedica 2020; 91 (4): 372–377

Patient-acceptable symptom state for the Oxford Hip Score and Forgotten Joint Score at 3 months, 1 year, and 2 years following total hip arthroplasty: a registry-based study of 597 cases Vincent P GALEA 1, Lina Holm INGELSRUD 2, Isabella FLORISSI 1, David SHIN 1, Charles R BRAGDON 1,3, Henrik MALCHAU 1,3, Kirill GROMOV 2, and Anders TROELSEN 2 1 Harris Orthopaedic Laboratory, Massachusetts General Hospital, Boston, MA, USA; 2 Department of Orthopedic Surgery, Copenhagen University Hospital Hvidovre, Hvidovre, Denmark; 3 Department of Orthopaedic Surgery, Harvard Medical School, Boston, MA, USA Correspondence: dshin4@mgh.harvard.edu Submitted 2019-09-24. Accepted 2020-02-19.

Background and purpose — Patient-acceptable symptom states (PASS) represent the level on a patient-reported outcome measure (PROM) at which patients are satisfied with postoperative outcomes. We defined the PASS for the Oxford Hip Score (OHS) and Forgotten Joint Score (FJS-12) at 3-month, 1-year, and 2-year intervals after primary total hip arthroplasty (THA). Patients and methods — Between July 2018 and April 2019, primary THA patients in an academic medical center’s registry completed the OHS, FJS-12, and a satisfaction anchor question at 3-month (n = 230), 1-year (n = 180), or 2-year (n = 187) postoperative intervals. PASS thresholds were derived with receiver operating characteristic analysis using the 80% specificity method. 95% confidence intervals (CI) were calculated using 1,000 non-parametric bootstrap replications. Results — 74%, 85%, and 86% of patients reported having a satisfactory symptom state at 3 months, 1, and 2 years after surgery, respectively. At 3-month, 1-year, and 2-year intervals, PASS thresholds were 34 (CI 31–36), 40 (CI 36–44), and 39 (CI 35–42) points for the OHS and 59 (CI 54–64), 68 (CI 61–75), and 69 (CI 62–75) points for the FJS-12. Interpretation — PASS thresholds varied with time for both the OHS and the FJS-12, with lower 3-month compared with 1-year and 2-year thresholds. These PASS thresholds represent OHS and FJS-12 levels at which the average patient is satisfied with THA outcomes, helping to interpret PROMs and serving as clinically significant benchmarks and patient-centered outcomes for research.

Patient-reported outcome measures (PROMs) are commonly used to evaluate preoperative and postoperative symptom states of patients undergoing procedures such as total hip arthroplasty (THA) (Rolfson et al. 2016). Although measures such as revision or infection rates may reliably identify significant outliers in arthroplasty outcomes, the absence of such negative outcomes is not sufficient to determine whether the outcome of a procedure was satisfactory from a patient’s point of view (American Academy of Orthopedic Surgeons 2018). Within arthroplasty, there is a focus on joint-specific PROMs, but even between these PROMs there remains variation in the ways in which joint-related health is measured. The Oxford Hip Score (OHS) and the Forgotten Joint Score (FJS-12) are 2 such PROMs. The OHS assesses hip pain and function, and has been widely used in hip arthroplasty since its development in 1996 (Dawson et al. 1998). The FJS-12, designed in 2012, is a joint-specific questionnaire that focuses on the patient’s awareness of the affected joint (Behrend et al. 2012). 3 studies comparing these 2 PROMs found a smaller ceiling effect (proportion of respondents achieving the maximum score) in the FJS-12 compared with the OHS, suggesting that the FJS-12 may be better at distinguishing between patients with good postoperative outcomes in comparison with the OHS within their respective constructs (Hamilton et al. 2016, 2017, Larsson et al. 2019). The patient acceptable symptom state (PASS) is the threshold on a PROM most closely associated with patient satisfaction, which is assessed on a separate questionnaire (Tubach et al. 2005, Sayers et al. 2017). PASS values allow for the interpretation of PROMs within the context of a given treatment,

Acta Orthopaedica 2020; 91 (4): 372–377

and they may fulfil a variety of roles: as clinically significant benchmarks for procedures, as clinically relevant, patient-centered outcomes for research, and as guides for physicians to contextualize a patient’s postoperative symptom state. Although 2 studies have presented PASS values for the OHS following THA, they have not been externally validated (Judge et al. 2012, Keurentjes et al. 2014). Furthermore, these studies did not investigate the time-dependence of the PASS. The PASS may change within the first year of surgery in accordance with changes in patient expectations during rehabilitation. 1 study established the OHS PASS 6 months after arthroplasty, while the other derived the PASS on a cohort of patients ranging between 1.5 and 6 years following THA. Another study applied a composite questionnaire-based satisfaction anchor criterion to establish an OHS value associated with patient satisfaction 1 year following THA of 37.5 points (Hamilton et al. 2018). To our knowledge, while no THA PASS values have been established for the FJS-12, a composite anchor questionnaire-based “successful treatment” anchor was used by 1 study to establish a threshold value of 74 and 70 points at 1- and 2-year intervals following THA, respectively (Rosinsky et al. 2019). We defined PASS values for the OHS and FJS-12 at 3 months, 1 year, and 2 years following primary THA.

Patients and methods Study design, patients, and data sources We performed a prospective observational cohort study analyzing data from the arthroplasty registry of a tertiary academic medical center in Denmark. Starting in March 2013, all patients undergoing primary THA due to osteoarthritis at this institution were asked to complete preoperative, 3-month, 1-year, and 2-year OHS and FJS-12 as part of the institutional registry’s data-collection process. Beginning in July 2018, all THA patients were asked to answer an additional question about satisfaction with their postoperative symptom state at each postoperative time point. These PROMs and satisfaction questions were administered electronically—patients who were unable to complete PROMs electronically were instead mailed the questionnaires. Patients were included in this study’s analysis if they had completed all of the OHS, FJS12, and satisfaction question at any of the 3-month, 1-year, or 2-year intervals postoperatively. Patients unable to speak or read Danish, refusing to participate in the data collection, or otherwise failing to complete a PROM battery for at least 1 time point were excluded from analysis. As the satisfaction question was administered beginning in July of 2018, only PROM batteries completed between July 2018 and April 2019 were included in this analysis. Each patient and their PROMs completed during this time were subsequently categorized into 3-month, 1-year, or 2-year postoperative interval cohorts (Figure 1).

373

Data collection period July 2018 – April 2019 2016

July

December

7 months THA patients n = 274 Missing PROMs n = 87 (32%) 2-year cohort with complete PROMs n = 187 (68%)

May

October

March

7 months THA patients n = 266 Missing PROMs n = 86 (32%) 1-year cohort with complete PROMs n = 180 (68%)

August

2019

11 months THA patients n = 372 Missing PROMs n = 142 (38%) 3-month cohort with complete PROMs n = 230 (62%)

Figure 1. Patient selection flowchart. Patient-reported outcome measures (PROMs) were collected from patients during the data collection period from July 2018 to April 2019, and were categorized as 3-month, 1-year, and 2-year postoperative interval PROMs. Patients unable to speak or read Danish or refusing to participate in the data collection or otherwise failing to complete the PROM responses were excluded from analysis.

Questionnaires The OHS is a 12-item questionnaire that measures a patient’s pain and physical ability. Each question of the OHS allows responses ranging from 0 (worst) to 4 (best), which are scaled and summed to provide a composite score ranging from 0 (worst) to 48 (best) points (Dawson et al. 1998). The FJS-12 is a questionnaire that assesses a patient’s awareness of his/her joint. This questionnaire was developed to assess patients’ awareness of their artificial joint following total joint arthroplasty. The PROM includes 12 questions that are each answered on a 5-level Likert scale. The FJS-12 generates a score ranging from 0 to 100 points, with a higher score indicating that the patient is less aware of the affected joint when undergoing daily activities (Behrend et al. 2012). The satisfaction question was: “Taking into account all the activities you have during your daily life, your level of pain, and also your functional impairment, do you consider that your current state is satisfactory?” (Tubach et al. 2005). Possible answers are “Yes” and “No.” This question served as the PASS anchor in our derivation analyses. Statistics Patient demographics and surgical variables are presented as median (interquartile range [IQR]) for non-parametric distributions, as mean (range) for parametric distributions, and as number (proportion) for categorical variables. The correlation of the OHS and FJS-12 to the anchor question was visualized with boxplots and assessed via point-biserial coefficients. The SPSS Statistics Version 24.0 (IBM Corp, Armonk, NY, USA) software package was used for all analyses. Methods of anchor-based PASS derivation 3 different methods were used to derive PASS thresholds for the OHS and for the FJS-12 at 3 months, 1 year, and 2 years postoperatively. The primary method of PASS derivation was the anchor-based 80% specificity method, which has been

374

Acta Orthopaedica 2020; 91 (4): 372–377

Table 1. Descriptive data for the 3-month, 1-year, and 2-year cohorts. Values are count (%) unless otherwise specified Factor

3-month 1-year 2-year n = 230 n = 180 n = 187

Oxford Hip score

Forgotten Joint score

100

Patient demographics Age (years), mean (SD) 68 (11) 68 (11) Female (vs. male) 140 (61) 111 (62) BMI, mean (SD) 27 (5) 30 (6) Severe (vs. mild/moderate) OA a 46 (20) 38 (21) No/mild (vs. moderate/severe) systemic disease b 44 (19) 32 (18) Implant characteristics c Cemented femoral component 92 (40) 68 (38)

67 (11) 118 (63) 27 (5) 32 (17) 41 (22) 65 (35)

a Osteoarthritis graded according to Tönnis. b According to American Society of Anesthesiologists (ASA) Score. c All acetabular components were uncemented across all cohorts.

Table 2. Patient-reported outcomes for the 3-month, 1-year, and 2-year cohorts. Values are median (interquartile range) unless otherwise specified Factor

3-month 1-year 2-year

Oxford Hip Score preoperative 22 (17–28) 23 (17–29) 22 (18–27) postoperative 39 (30–43) 45 (38–48) 44 (36–47) Forgotten Joint Score preoperative 15 (4–29) 17 (7–29) 17 (4–27) postoperative 71 (50–86) 83 (58–96) 81 (55–96) Reporting satisfactory symptom state, n (%) a 170 (74) 153 (85) 161 (86) a

Referring to PASS transition item described in methods.

previously shown to be the most reliable method of PASS derivation (Aletaha et al. 2009, Kvamme et al. 2010). By this method, the PASS is the numerical value on the PROM below which 80% of dissatisfied patients are correctly identified. To derive 95% confidence intervals (CI) for these PASS values, PASS values were calculated for 1,000 non-parametric bootstrapped samples of each study subcohort, and by deriving the 2.5 and 97.5 quantiles therein (Campbell 1999). 2 additional methods of PASS derivation were performed as sensitivity analyses. The 1st of these methods is the Youden method (Youden 1950), which identifies the PASS as the coordinate on the ROC curve for which there is the highest combination of sensitivity and specificity. The 2nd of these methods is the 75th percentile method (Tubach et al. 2005), which defines the PASS as the numerical value on a PROM scale beyond which 75% of patients reported satisfaction with the outcome of their procedure. Ethics, funding, and potential conflicts of interest The institutional arthroplasty registry supplying data for this analysis was approved by the national data protection agency

24 40 12

3-month

1-year

2-year

3-month

1-year

2-year

Figure 2. Boxplots depicting differences in Oxford Hip Scores (left panel) and Forgotten Joint Scores (right panel) between those who reported being in a satisfactory symptom state (green) and those who did not (red) for the 3-month, 1-year, and 2-year cohorts. Horisontal lines are median, boxes interquartile range (IQR), whiskers range, ● ouliers (> 1.5 x IQR), and * extreme ouliers (> 3 x IQR).

in Denmark, where approval from the IRB is not required for registry-based studies that exclusively examine PROMs. The study was conducted in accordance with the Declaration of Helsinki. This study was fully funded by the orthopedic departments of 2 institutions and an orthopedics research lab. The authors declare no potential conflicts of interest.

Results Demographic and implant data for the cohorts at each postoperative time point are presented in Table 1. Despite being composed of different patients, the 3-month, 1-year, and 2-year postoperative patient cohorts were comparable across the demographic variables assessed. OHS values, FJS-12 values, and the proportion of patients reporting a satisfactory symptom state are presented in Table 2. At 3 months postoperatively, 74% of patients reported having satisfactory symptoms and this proportion was 85% and 86% at 1 and 2 years postoperatively, respectively. The mean OHS was 39 points at 3 months, and 45 and 44 points at 1 and 2 years postoperatively, respectively. Similarly, the mean FJS-12 value increased from 71 points at 3 months to 83 and 81 points at 1 and 2 years postoperatively. Correlation between the PROMs and the satisfaction anchor The point-biserial coefficients between the OHS and the satisfaction item were 0.47 for the 3-month cohort, 0.50 for the 1-year cohort, and 0.45 for the 2-year cohort. The point-biserial coefficients between the FJS-12 and the transition item were 0.51 for the 3-month cohort, 0.53 for the 1-year cohort, and 0.56 for the 2-year cohort. FJS-12 and OHS values were lower for most patients who answered “No” to the satisfaction question, when compared with those who answered “Yes”— this held true across all time-point cohorts (Figure 2).

Acta Orthopaedica 2020; 91 (4): 372–377

375

Table 3. PASS analysis at each follow-up interval for the Oxford Hip Score and Forgotten Joint Score using 3 different methods to determine threshold values PASS values PROM ROC curves Methods a Cohort AUC (95% CI) p-value A B C Oxford Hip Score 3-month 0.84 (0.77–0.90) 1-year 0.92 (0.90–0.99) 2-year 0.90 (0.83–0.98) Forgotten Joint Score 3-month 0.87 (0.82–0.93) 1-year 0.95 (0.90–0.99) 2-year 0.94 (0.82–0.99)

PASS value—OHS

PASS value—FJS 100

48 OHS FJS

60 24

< 0.001 < 0.001 < 0.001

34 40 39

33 38 38

34 41 40

< 0.001 < 0.001 < 0.001

59 68 69

57 66 69

59 69 71

a Methods: A

= 80% specificity (the primary PASS analysis method); B = Youden; C = 75th percentile. PASS: Patient acceptable symptom state. PROM: Patient reported outcome measure. ROC: Receiver operating characteristic. AUC: area under the curve.

PASS thresholds Based on the primary method of PASS derivation, PASS threshold values were found to be 34 (CI 31–36), 40 (CI 36–44), and 39 (CI 35–42) points on the OHS and 59 (CI 54–64), 68 (CI 61–75), and 69 (CI 62–75) points on the FJS12, for the 3-month, 1-year, and 2-year cohorts, respectively. There was minimal variation in PASS thresholds for both PROMs when comparing the 3 methods of derivation (Table 3). The 3-month PASS thresholds were observed to be lower than those of the 1-year and 2-year cohorts for both the OHS and the FJS-12 (Figure 3).

Discussion We derived PASS values for the OHS, a well-established and widely used PROM, and for the FJS-12, a newer and promising PROM, in the early follow-up period after THA. PASS thresholds were 34, 40, and 39 points for the OHS and 59, 68, and 69 points for the FJS-12, for the 3 months, 1-year, and 2-year cohorts, respectively. PROMs provide an objective way to measure a patient’s subjective experience (Gagnier 2017). The many PROMs used in orthopedics ask different questions of patients and quantify different constructs, ranging from general health to joint pain, joint function, and joint awareness. In addition to assessing distinct constructs, each PROM can be applied to evaluate treatments that have different goals and therefore different expected results. While PROMs include the patient perspective, the heterogeneity of both the available PROMs as well as their different applications complicates their interpretation. The PASS can provide valuable insight into the interpretation of PROMs for clinicians and researchers alike by identifying

40 12

3-month

1-year

2-year

Figure 3. Patient acceptable symptom state (PASS) thresholds for each patient reported outcome measure at 3 months, 1 year, and 2 years after total hip arthroplasty calculated using the 80% specificity method. Error bars represent 95% confidence intervals calculated using 1,000 non-parametric bootstrap replications. Possible FJS-12 values range from 0 to 100 points, while OHS values range from 0 to 48 points.

the value on a PROM scale at which patients consider their symptom state to be satisfactory following a given procedure. This, in turn, enables clinicians to contextualize the PROM scores of their patients, and provides researchers with a clinically relevant, patient-centered benchmark for the evaluation of surgical success. 2 previous studies have derived PASS values for the OHS at intervals of 6 months and 3 years after THA (Judge et al. 2012, Keurentjes et al. 2014). The 3-year PASS derivation suffered from a wide range of sample PROM follow-up intervals, ranging from 1.5 to 6 years—such a spread of time points is antithetical to the concept of the PASS as a timedependent measure. Although the 3-year study also derived and compared PASS values from subsets of pre- and post-3year PROMs, it was not able to consider PROMs collected prior to 1 year after surgery (Keurentjes et al. 2014). The PASS is likely to change within the first year of surgery as patients undergo rehabilitation. A combination of time points including those earlier than 1.5 years may better map across the typical recovery period for THA. The study deriving values at 6 months was also not able to assess the potential time-dependence of the PASS, given that only 1 time-point was studied (Judge et al. 2012). In addition, that study was limited by the use of a continuous anchor (satisfaction visual analogue scale), which was arbitrarily dichotomized at the midpoint. Our study considered 3 time intervals across the early follow-up period after THA and found that PASS values were subject to change between 3 months and 1 year. Similar results were found when comparing our 3-month PASS value with the previously derived 6-month value (35). So, too, our 1- and 2-year PASS values were found to be similar to those presented by the 3-year study (42). Hamilton et al. (2018) also established threshold values for the OHS at 1 year fol-

376

lowing THA representing “treatment success”—the authors used a composite questionnaire-based anchor, obtaining a threshold value of 37.5 points for the OHS for “success.” Furthermore, Rosinsky et al. (2019), using another composite questionnaire-based successful treatment anchor for the FJS-12, found threshold values of 74 and 70 points at 1- and 2-year intervals compared with those of this study at 68 and 69 points at 1- and 3-year intervals following THA. These similarities suggest that the PASS may be somewhat robust to variations in derivation methods and criteria. Giesinger et al. (2019) established normative values for the PROM across the US general population, which exceeded PASS thresholds derived in this study at all postoperative time intervals, indicating that restoration of symptom states to those of the general population may not be required for patients to be satisfied with THA outcomes. Our study is not without limitations. 1st, because of data collection constraints, our 3-month, 1-year, and 2-year cohorts comprised different patients. If we had been able to follow a single cohort of patients over time, we could have better assessed longitudinal PROM changes using paired analyses. Nevertheless, a comparison of demographic and implant variables showed the 3 cohorts to be similar. 2nd, our patients were sourced from a single Danish institution, and therefore the generalizability of our PASS values may be limited given that PROM results may vary across different international regions. However, both the OHS and the FJS-12 have been found to have comparable psychometric properties across different language versions (Paulsen et al. 2012, Harris et al. 2016, Shadid et al. 2016, Hamilton et al. 2017, Klouche et al. 2018). Furthermore, the patient demographics in our study are comparable to other registry settings (National Joint Registry 2016, American Joint Replacement Registry 2017, AOANJRR 2017). PASS values are meant to represent the result with which the average patient is likely to be satisfied. A registrybased study, including patients treated by a variety of providers with a variety of implants, is an ideal way to determine such a value. External validation of the PASS values derived in this study may prove useful for assessing the generalizability of these thresholds to a broader population of patients. Additionally, future studies might consider deriving PASS values at later time-points in order to gauge whether age-related decline influences the PASS. Given the typical age range of THA patients, extending PASS analysis up to 10 years or longer may offer valuable insight. Our study is the first to present PASS thresholds for the FJS-12 after THA. The OHS PASS values derived were found to be similar to values presented by other studies, but they provide clearer evidence of the changes in PASS over time. We found that both OHS and FJS-12 PASS thresholds increase between 3 months and 1 year, but not between 1 and 2 years. These PASS thresholds represent the level on the OHS and FJS-12 where patients undergoing THA find their symptom state acceptable.

Acta Orthopaedica 2020; 91 (4): 372–377

VPG, LHI, and AT conceived the research question for the study and took part in drafting and revising the manuscript. CRB, HM, KG, and AT designed the experiment and took part in revising the manuscript. VPG, IF, and DS performed the statistical analysis of the data and drafted the manuscript. Acta thanks David F Hamilton and Nienke Wolterbeek for help with peer review of this study.

Aletaha D, Funovits J, Ward M M, Smolen J S, Kvien T K. Perception of improvement in patients with rheumatoid arthritis varies with disease activity levels at baseline. Arthritis Rheum 2009; 61(3): 313-20. American Academy of Orthopedic Surgeons. AAOS Registry Program Patient-Reported Outcome Measures Guide; 2018. American Joint Replacement Registry. Annual Report; 2017. AOANJRR. Hip and Knee Arthroplasty—Annual Report; 2017. Behrend H, Giesinger K, Giesinger J M, Kuster M S. The “forgotten joint” as the ultimate goal in joint arthroplasty: validation of a new patient-reported outcome measure. J Arthroplasty 2012; 27(3): 430-6. Campbell M K. Bootstrapping: estimating confidence intervals for cost-effectiveness ratios. QJM. 1999; 92(3): 177-82.[AQ1] Dawson J, Fitzpatrick R, Murray D, Carr A. Questionnaire on the perceptions of patients about total knee replacement. J Bone Joint Surg Br 1998; 80(1): 63-9. Gagnier J J. Patient reported outcomes in orthopaedics. J Orthop Res 2017; 35(10): 2098-108. Giesinger J M, Behrend H, Hamilton D F, Kuster M S, Giesinger K. Normative values for the Forgotten Joint Score-12 for the US general population. J Arthroplasty 2019; 34(4): 650-5. Hamilton D F, Giesinger J M, MacDonald D J, Simpson A H R W, Howie C R, Giesinger K. Responsiveness and ceiling effects of the Forgotten Joint Score-12 following total hip arthroplasty. Bone Joint Res 2016; 5(3): 87-91. Hamilton D F, Loth F L, Giesinger J M, Giesinger K, MacDonald D J, Patton J T, Simpson A H R W, Howie C R. Validation of the English language Forgotten Joint Score-12 as an outcome measure for total hip and knee arthroplasty in a British population. Bone Joint J 2017; 99-B(2): 218-24. Hamilton D F, Loth F L, MacDonald D J, Giesinger K, Patton J T, Simpson A H, Howie C R, Giesinger J M. Treatment success following joint arthroplasty: defining thresholds for the Oxford Hip and Knee Scores. J Arthroplasty 2018; 33(8): 2392-7. Harris K, Dawson J, Gibbons E, Lim C, Beard D, Fitzpatrick R, Price A. Systematic review of measurement properties of patient-reported outcome measures used in patients undergoing hip and knee arthroplasty. Patient Relat Outcome Meas 2016; 7: 101-8. Judge A, Arden N K, Kiran A, Price A, Javaid M K, Beard D, Murray D, Field R E. Interpretation of patient-reported outcomes for hip and knee replacement surgery. J Bone Joint Surg Br 2012; 94-B(3): 412-18. Keurentjes J C, Van Tol F R, Fiocco M, So-Osman C, Onstenk R, KoopmanVan Gemert A W M M, Poll R G, Nelissen R G H H. Patient acceptable symptom states after total hip or knee replacement at mid-term follow-up: thresholds of the Oxford hip and knee scores. Bone Joint Res 2014; 3(1): 7-13. Klouche A, Giesinger J M, Sariali EH. Translation, cross-cultural adaption and validation of the French version of the Forgotten Joint Score in total hip arthroplasty. Orthop Traumatol Surg Res 2018; 104(5): 657-61. Kvamme M K, Kristiansen I S, Lie E, Kvien T K. Identification of cutpoints for acceptable health status and important improvement in patient-reported outcomes, in rheumatoid arthritis, psoriatic arthritis, and ankylosing spondylitis. J Rheumatol 2010; 37(1): 26-31. Larsson A, Rolfson O, Kärrholm J. Evaluation of Forgotten Joint Score in total hip arthroplasty with Oxford Hip Score as reference standard. Acta Orthop 2019; 90(3): 253-7.

Acta Orthopaedica 2020; 91 (4): 372–377

National Joint Registry. National Joint Registry 14th Annual report. Natl Jt Regist Reports 2016; (2017). Paulsen A, Odgaard A, Overgaard S. Translation, cross-cultural adaptation and validation of the Danish version of the Oxford hip score: assessed against generic and disease-specific questionnaires. Bone Joint Res 2012; 1(9): 225-33. Rolfson O, Eresian Chenok K, Bohm E, Lübbeke A, Denissen G, Dunn J, Lyman S, Franklin P, Dunbar M, Overgaard S, Garellick G, Dawson J, Patient-Reported Outcome Measures Working Group of the International Society of Arthroplasty Registries. Patient-reported outcome measures in arthroplasty registries. Acta Orthop 2016; 87 (eSuppl 362): 3-8. Rosinsky P J, Chen J W, Lall A C, Shapira J, Maldonado D R, Domb B G. Can we help patients forget their joint? Determining a threshold for successful outcome for the Forgotten Joint Score. J Arthroplasty 2019; 35(1): 153-9.

377

Sayers A, Wylde V, Lenguerrand E, Gooberman-Hill R, Dawson J, Beard D, Price A, Blom A W. A unified multi-level model approach to assessing patient responsiveness including: return to normal, minimally important differences and minimal clinically important improvement for patient reported outcome measures. BMJ Open 2017; 7(7): e014041. Shadid M B, Vinken N S, Marting L N, Wolterbeek N. The Dutch version of the Forgotten Joint Score: test–retesting reliability and validation. Acta Orthop Belg 2016; 82(1): 112-18. Tubach F, Ravaud P, Baron G, Falissard B, Logeart I, Bellamy N, Bombardier C, Felson D, Hochberg M, Van Der Heijde D, Dougados M. Evaluation of clinically relevant states in patient reported outcomes in knee and hip osteoarthritis: rhe patient acceptable symptom state. Ann Rheum Dis 2005; 64(4): 34-7. Youden W J. Index for rating diagnostic tests. Cancer 1950; 3(1): 32-5.

378

Acta Orthopaedica 2020; 91 (4): 378–382

Repeated cobalt and chromium ion measurements in patients with bilateral large-diameter head metal-on-metal ReCap-M2A-Magnum total hip replacement Sakari PIETILÄINEN 1, Heikki MÄNTYMÄKI 2, Tero VAHLBERG 3, Aleksi REITO 4, Antti ESKELINEN 4, Petteri LANKINEN 1, and Keijo MÄKELÄ 1 1 Department

of Orthopaedics and Traumatology, Turku University Hospital and University of Turku; 2 Department of Orthopaedics, Tampere University Hospital and University of Turku; 3 Department of Biostatistics, University of Turku; 4 Coxa Hospital for Joint Replacement, Tampere, Finland Correspondence: sakari.pietilainen@tyks.fi Submitted 2019-12-23. Accepted 2020-03-23.

Background and purpose — Whole-blood (WB) chromium (Cr) and cobalt (Co) measurements are vital in the follow-up of metal-on-metal total hip replacement (MoM THR) patients. We examined whether there is a substantial change in repeated WB, Co, and Cr levels in patients with bilateral ReCap-M2A-Magnum THR. We also specified the number of patients exceeding the safe upper limit (SUL) of WB Co and Cr in the repeated measurement. Patients and methods — We identified 141 patients with bilateral ReCap-M2A-Magnum THR operated in our institution. 61 patients had repeated WB metal ion measurements with bilateral MoM implants still in situ in the second measurement. The mean time elapsing from the first measurement (initial measurement) to the second (control measurement) was 1.9 years (SD = 0.6, range 0.2–3.5). We used earlier established SUL levels for bilateral implants by Van Der Straeten et al. (2013). Results — The median (range) Co and Cr values decreased in the repeated measurement from 2.7 (0.6–25) to 2.1 (0.5–21) and 2.6 (0.8–14) to 2.1 (0.5–18) respectively. In 13% of the patients Co levels exceeded the SUL in the initial measurement and the proportion remained constant, at 13%, in the repeated measurement. In 5% of the patients, Cr levels were above SUL in the initial measurement and an equal 5% in the control measurement. Interpretation — Repeated WB metal ion levels did not increase in patients with bilateral ReCap-M2A-Magnum THR with a mean 1.9-year measurement interval. Long-term development of WB metal ion levels is still unclear in these patients.

More than 20,000 metal-on-metal (MoM) hip replacements were performed in Finland during 2000–2015 (Finnish Arthroplasty Register). Currently, there are still thousands of patients with a MoM THR in situ. Whole-blood (WB) metal ion measurements are an essential part of the follow-up of MoM patients, even though they do not solely identify failing implants alone (De Smet et al. 2008, Hart et al. 2014, Reito et al. 2016). While there is no agreed universal WB metal ion level that indicates revision surgery or predicts the outcome, different health authorities have suggested diverse follow-up protocols for the monitoring of MoM patients (Hannemann et al. 2013, MHRA 2017, US Food and Drug Administration (FDA) 2019). Furthermore, some MoM implants have better survival rates than others, which makes risk evaluation even more difficult (Matharu et al. 2016, MHRA 2017, Kasparek et al. 2018, Donahue et al. 2019). The evaluation of patients with bilateral MoM THR is even more challenging. Patients with bilateral MoM implants often present higher levels of Co and Cr than patients with a unilateral device (Van Der Straeten et al. 2013, Reito et al. 2014, 2016). Only a few studies have assessed blood metal ion levels in patients with bilateral MoM THR. Reito et al. (2016) evaluated ion level changes in bilateral ASR THR, and ASR (DePuy, Warsaw, IN, USA) hip resurfacing arthroplasty (HRA) patients. Both WB Co and Cr were substantially higher in the ASR THR cohort in the repeated measurement (Reito et al. 2016). However, metal ion levels were not able to distinguish failing MoM components from well-functioning hips in patients with bilateral ASR THR (Reito et al. 2016, Donahue et al. 2019).

Acta Orthopaedica 2020; 91 (4): 378–382

ReCap-M2A-Magnum was the most common MoM THR in Finland (Finnish Arthroplasty Register). We have previously reported that repeated metal ion measurements in unilateral ReCap-M2A-Magnum patients at a mean 2-year time interval did not show any increase (Mäntymäki et al. 2019). We performed a retrospective comparative study to further investigate the role of repeated WB metal ion measurements in patients with bilateral M2A-ReCap-Magnum THR. Our main objectives were to investigate: 1. Is there a substantial change in the WB Co and Cr level during a follow-up period? 2. How large proportion of patients’ measurements exceed the safe upper limits (SUL) of WB Co and Cr levels in the repeated measurement (thresholds WB Co 5.0 µg/L and Cr 7.4 µg/L) (Van Der Straeten et al. 2013).

Patients and methods A screening program for MoM hips was launched at our institution to detect patients with adverse reactions to metal debris (ARMD). The screening was performed in consensus with the follow-up protocol recommended by the Finnish Arthroplasty Society (Finnish Arthroplasty Society 2015). The screening included anteroposterior and lateral radiographs of the hip, WB Cr and Co ion measurements, and Oxford Hip Score (OHS) questionnaire. Furthermore, if patients had poor or moderate OHS score, or elevated Cr or Co WB concentration (beyond 5 ppb), they were referred for MARS (magnetic artefact reduction sequence) MRI. Patients with poor or moderate OHS or elevated WB ion measurements were also clinically evaluated by a senior orthopedic surgeon in an outpatient clinic. If patients had severe hip symptoms (pain, clicking, swelling) or if a pseudotumor was detected in MRI, revision surgery was considered. In addition to this, if an asymptomatic patient had WB metal ion levels above 10 ppb, revision surgery was considered to minimize the risk of Co poisoning. Patients who were not admitted for revision surgery were scheduled for annual or biannual visits in our outpatient clinic. A ReCap-M2A-Magnum THR was used in 1,329 operations (1,188 patients) at our institution from 2005 to 2012. For this study we identified patients with bilateral ReCap-M2A-Magnum THR. Overall 141 patients (282 hips) had bilateral M2AReCap-Magnum THR. Of these 141 patients we identified 62 patients with at least 2 WB Co and Cr ion measurements. Of these, 3 patients had unilateral revision surgery during the follow-up period. 1 patient was revised due to aseptic loosening of the femoral component, and another for acute-onset infection. However, both patients still had both MoM bearing surfaces in situ after the revision surgery, and they remained in our study group. One patient was excluded because of unilateral revision surgery, where MoM bearing surfaces were converted to conventional ones. After this exclusion we had

379

61 bilateral (31 females) ReCap-M2A Magnum THR patients (122 hips) in our study group. The mean age of patients was 60 years (SD 9.7) at the time of the first hip arthroplasty. The mean femoral head size was 50 mm (SD = 3.4) and the mean acetabular inclination 44 degrees (SD = 6.3). The study period concerning primary operations was from 2005 to 2012. The follow-up data concerning ion measurements were collected from the patients until 2017. All participating patients had their blood samples taken from the antecubital vein using a 21-gauge BD Vacutainer® Eclipse™ blood collection needle (Becton, Dickinson & Co, Franklin Lakes, NJ, USA). The first 10 mL tube of blood was used for analysis of standard laboratory tests such as C-reactive protein and erythrocyte sedimentation rate measurement. The second blood sample was taken in Vacuette® NH trace elements tube (Greiner Bio-One GmbH, Kremsmünster, Austria) containing sodium heparin. Cobalt and chromium analyses from whole blood were performed using an accredited method with inductively coupled plasma mass spectrometry (ICP-MS, VITA Laboratory, Helsinki, Finland in collaboration with Medical Laboratory of Bremen, Germany). The detection limit for Cr was 0.2 ppb and for Co 0.2 ppb. The intra-assay variation for WB Cr and Co were 2.2% and 2.7% and inter-assay variation were 6.7% and 7.9%, respectively. Statistics 61 patients with bilateral ReCap-M2a-Magnum THR met the criteria with at least 2 repeated metal ion measurements. The mean time elapsing from the first metal ion assessment (initial measurement) to the second (control measurement) was 1.9 years (SD 0.6, range 0.2–3.5). The time elapsing from the second hip replacement to the first (initial) metal ion measurement was considered as the follow-up time. Mean follow-up time from the second operation to the initial measurement was 4.7 years (1.9–9.0). Patients were divided into follow-up time interval groups according to the time elapsing from the second operation to the first metal ion assessment. The individual change in 2 consecutive metal ion measurements from the same patient was modelled using a random coefficient model. Log-transformed ion values were used in conditional models due to positively skewed distribution of ion levels. Results are expressed as geometric means for better interpretability. SUL values for WB Co were 5.0 ppb and WB Cr 7.4 ppb as reported earlier (Van Der Straeten et al. 2013). P-values lower than 0.05 in a 2-tailed test were considered statistically significant. The change over a 1.9-year measurement interval was calculated and plotted as frequency distributions for both metal ions separately. Ethics, funding, and potential conflicts of interest The study was based on the national recommendation for systematic screening of MoM THR patients given by the Finnish Arthroplasty Society (2015). It was a register study, and

380

Acta Orthopaedica 2020; 91 (4): 378–382

Differences in WB Co and Cr levels (ppb)

ChangeChange in Co values in Co (ppb) values (ppb)

ChangeChange in Cr values in Cr (ppb) values (ppb)

< 0.001 –5

–5

Initial Control p-value

WB Co, n = 61 geometric mean median (range) WB Cr, n = 61 geometric mean median (range)

2.8 2.7 (0.60–25)

2.2 2.1 (0.50–21)

2.8 2.6 (0.80–14)

2.3 2.1 (0.50–18)

< 0.007

There was a statistically significant decrease in repeated WB Co and –10 Cr ion values.

–10 2 0 4 2 6 4 8 6 10 8 12 1014 1216 14

Initial Co Initial values Co (ppb) values (ppb)

–10

–10 2 0 4 2 6 4 8 6 10 8 12 1014 1216 14

Figure 2. Changes in Cobalt (Co) and Chromium (Cr) ion levels compared to initial measurement Whole blood Co concentration (ppb)

Whole blood Cr concentration (ppb)

Log Log of Co ofvalues Co values (ppb)(ppb)

First measurement Second measurement

Log Log of Crofvalues Cr values (ppb)(ppb)

3.0 3.0

2.0 2.0

1.0 1.0

0.0 0.0

5 3

4 3

2 1

1 0

p = 0.01 p = 0.6 n = 11 n = 10

< 3.0

p = 0.6 p = 0.06 p = 0.3 n = 15 n = 12 n = 13

3.0–3.9 4.0–4.9 5.0–5.9 6.0–6.9

Years to first measurement

p = 0.01 p = 0.08 p = 0.4 n = 11 n = 10 n = 15

< 3.0

p = 0.1 p = 0.04 n = 12 n = 13

3.0–3.9 4.0–4.9 5.0–5.9 6.0–6.9

–1.0 –1.0

Years to first measurement

InitialInitial

–1.0 –1.0

Control Control

InitialInitial

Control Control

Figure 1. Geometric mean whole blood Co values (left) and Cr levels (right) divided across the follow-up time before initial measurement.

Figure 3. Spagetti plots for Co and Cr values at initial and control measurements. Values are naturally log-transformed.

the patients were not directly contacted. Therefore, approval by the local ethical committee was not needed. Data sharing is not possible. No benefits in any form have been received related directly or indirectly to this article. Outside this study, HM has received travel/accommodation expenses from DePuy Synthes. AE received research funding from Zimmer Biomet and DePuy Synthes and consultancy fees from Zimmer Biomet. AR reports personal fees from a paid lecture. SP, PL, KTM, and TV have nothing to disclose.

value above the SUL in both measurements. 1 patient had his Cr value below the SUL in the first measurement and above the SUL in the repeated measurement. In a similar manner, 1 patient had his Cr value above the SUL in the first measurement, but below the SUL in the repeated measurement. The Co and Cr levels decreased over time and stayed mostly below the SUL if the initial value was low. The exceptions were those with high values already in the initial measurements (Figure 2). Spaghetti plots for individual Co and Cr values at initial and control measurements are presented in Figure 3. Values are naturally log-transformed.

Results (Table) The geometric mean of WB Cr level decreased in the < 3-year and ≥ 6-year follow-up groups. The geometric mean of WB Co level decreased in the < 3-year group (Figure 1). Co values were below the SUL in 49 of the 61 patients in both metal ion measurements. 4 patients (6.6%) had their Co value below the SUL in the first measurement and above the SUL in repeated measurement. Similarly, 4 patients had their Co value above the SUL in the first measurement and below the SUL in the repeated measurement. Only 4 patients had Co ion values above SUL in both measurements. Cr values were below the SUL in 57 of the 61 patients in both metal ion measurements. Only 2 patients had their Cr

Initial CrInitial values Cr (ppb) values (ppb)

Discussion The motivation for performing this study was the lack of evidence of progress of metal ion levels in bilateral ReCap Magnum THR patients. We found that median or geometric mean WB Co and Cr levels in repeated metal ion measurements in bilateral ReCap-M2A-Magnum patients at a mean 1.9-year time interval did not show notable increase. However, our results cannot be applied to other MoM THR brands. Data concerning ion levels of patients with a ReCap Magnum THR are scarce. A strength of our study is that we

Acta Orthopaedica 2020; 91 (4): 378–382

are able to present novel information, which can be used in modifying follow-up schedules worldwide. We are not aware of any other studies concerning ion levels of bilateral ReCap Magnum THRs. A limitation of our study is that the follow-up time was short. Long-term WB ion levels in patients with bilateral ReCap-M2A-Magnum are not yet known. It is possible that a longer time range between the measurements such as 10 years might give different results. Also, the mean time interval between the WB ion measurements was only 1.9 years. Another limitation of our study is that patients with intense hip symptoms or a pseudotumor in the MRI may have been revised before any ion measurements. In addition, some of the patients who had substantially elevated WB Co or Cr levels after the initial measurement could have been admitted for revision surgery to decrease the risk of toxic effects of the metal ion levels. The literature concerning patients with bilateral MoM hip arthroplasty is limited. Van Der Straeten et al. (2013) studied a group of 453 patients with unilateral, and 139 patients with bilateral MoM hip arthroplasty. They compared WB Co and Cr levels in patients with a well-functioning MoM hip with those who had a poorly functioning MoM hip. They suggested a SUL value of 4.6 µg/L for Cr, and 4.0 µg/L for Co in patients with unilateral MoM hip. Accordingly, they suggested SUL values of 7.4 µg/L for Cr and 5.0 µg/L for Co in bilateral patients. They stated that WB ion values above this predicted problems in metal-on-metal resurfacings. Donahue et al. (2019) proposed an even lower SUL of 4.0 µg/L for both Co and Cr for patients with bilateral ASR HRA (DePuy, Warsaw, IN, USA). The lower SUL was supposedly because ASR HRA has inferior survival to other HRA models. In their study, a SUL of 4.0 µg/L was able to successfully differentiate wellfunctioning implants from poorly functioning implants with a sensitivity of 42% and specificity of 90%. However, they were unable to present reliable general SUL for MoM THA due to the inadequate cohort size. In our study we used SUL values suggested by Van Der Straeten et al. (2013), because their study included also other brands in addition to ASR hip prosthesis. The Finnish Arthroplasty Society recommends biannual metal ion measurements of MoM THA patients (Finnish Arthroplasty Society 2015). However, there are no clear guidelines on how to interpret ion concentrations and how high levels justify revision surgery. It seems that further research is needed to elucidate implant-specific WB metal ion level thresholds (Matharu et al. 2015). Sidaginamale et al. (2013) found that metal ion concentrations are reliable indicators of abnormal wear processes in MoM implants and the Co concentration threshold of 4.5 µg/L provided good sensitivity and specificity. Metal ion levels that should raise concern vary in different countries. In the UK, Canada and Europe values that cause alarm are between 2 ppb and 7 ppb (EFORT 2012, Health Canada 2012, MHRA 2017).

381

Reito et al. (2016) assessed a cohort of 76 patients with bilateral (ASR) hip resurfacings or with bilateral ASR XL THR with repeated WB ion measurements and with a median follow-up of 3.6 years. They reported no substantial difference in the HRA cohort (38 patients). However, patients with bilateral THR had a statistically significant increase in their WB Co and CR ion levels during this follow-up period (Co 8.3 µg/L vs. 12.6 µg/L, Cr 3.15 µg/L vs. 3.4 µg/L, both p < 0.001) between the 2 measurements. They therefore suggested that annual blood metal ion measurements on patients with bilateral high-risk MoM THR could be beneficial (Reito et al. 2016). We were not able to confirm this finding in patients with bilateral ReCap-M2A-Magnum THR. The poorer performance of the ASR device may explain the difference in WB ion level development compared with the ReCap Magnum THR (Seppanen et al. 2018). Matharu et al. (2017) recommended the use of different whole-blood (WB) metal ion thresholds for different implants in the follow-up MoM patients. Our current findings support this recommendation. Mäntymäki et al. (2019) studied a group of 319 patients with unilateral ReCap-M2A-Magnum THR with repeated metal ion measurements. They had a mean follow-up time of 5.5 years (1.8–9.3) and the mean time between the measurements was 2 years. A statistically significant decrease in both Co and Cr values was detected. Both Co and Cr concentrations remained within ± 1 ppb of their initial value in the majority of patients (86% for Co, 81% for Cr). They concluded that repeated metal ion measurements may not be necessary for patients with unilateral M2A-ReCap-Magnum THR patients with WB metal ion levels below the SUL. It seems that the same may hold true even in patients with bilateral devices. In summary, it is not necessary that patients with asymptomatic bilateral ReCap Magnum THR undergo metal ion level measurements at 2-year intervals. The optimal measurement interval is not yet known. Long-term metal ion level progression is not known either. Therefore, further research on the subject is needed.

AR and AE designed the protocol and methods. KTM performed the surgery and recorded the intraoperative data. AR and TV analyzed the data and did the statistics. SP, KTM, and TV collected the data. SP, HM, PL, KTM, and TV wrote the manuscript. All authors contributed to the revision of the manuscript. Acta thanks Pieter Bos and Anne Garland for help with peer review of this study.

Donahue G S, Galea V P, Laaksonen I, Connelly J W, Muratoglu O, Malchau H. Establishing thresholds for metal ion levels in patients with bilateral articular surface replacement hip arthroplasty. HIP Int 2019; 29(5): 475-80. EFORT. European Federation of National Associations of Orthopaedics and Traumatology (EFORT), E. H. S. E., Consensus statement “Current Evidence on the Management of Metal-on-Metal Bearings”, April 16, 2012. Finnish Arthroplasty Register. https://www.thl.fi/far/.

382

Finnish Arthroplasty Society. Recommendation for the use and follow-up of patients with metal-on-metal (MoM) hip arthroplasty; 2015. Hannemann F, Hartmann A, Schmitt J, Lützner J, Seidler A, Campbell P, Delaunay C P, Drexler H, Ettema H B, García-Cimbrelo E, Huberti H, Knahr K, Kunze J, Langton D J, Lauer W, Learmonth I, Lohmann C H, Morlock M, Wimmer M A, Zagra L, Günther K P. European multidisciplinary consensus statement on the use and monitoring of metal-on-metal bearings for total hip replacement and hip resurfacing. Orthop Traumatol Surg Res 2013; 99(3): 263-71. Hart A J, Sabah S A, Sampson B, Skinner J A, Powell J J, Palla L, Pajamäki K J J, Puolakka T, Reito A, Eskelinen A. Surveillance of patients with metalon-metal hip resurfacing and total hip prostheses: a prospective cohort study to investigate the relationship between blood metal ion levels and implant failure. J Bone Joint Surg Am 2014; 96(13): 1091-9. Health Canada. “Metal-on-metal hip implants—information for orthopaedic surgeons regarding patient management following surgery—for health professionals”; May 9, 2012. Kasparek M F, Renner L, Faschingbauer M, Waldstein W, Weber M, Boettner F. Predictive factors for metal ion levels in metal-on-metal total hip arthroplasty. Arch Orthop Trauma Surg 2018; 138(2): 281-6. Mäntymäki H, Lankinen P, Vahlberg T, Reito A, Eskelinen A, Mäkelä K. Repeated cobalt and chromium ion measurements in patients with largediameter head metal-on-metal ReCap-M2A-Magnum total hip replacement. Acta Orthop 2019; 90(3): 243-8. Matharu G S, Mellon S J, Murray D W, Pandit H G. Follow-up of metalon-metal hip arthroplasty patients is currently not evidence based or cost effective. J Arthroplasty 2015; 30(8): 1317-23. Matharu G S, Berryman F, Brash L, Pynsent P B, Treacy R B C, 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 2016; 98(8): 617-26.

Acta Orthopaedica 2020; 91 (4): 378–382

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 2017; 99(18): 1532-9. MHRA. Medical Device Alert: All metal-on-metal (MoM) hip replacements: updated advice for follow-up of patients; 2017. Reito A, Moilanen T, Puolakka T, Pajamäki J, Eskelinen A. Repeated metal ion measurements in patients with high risk metal-on-metal hip replacement. Int Orthop 2014; 38(7): 1353-61. Reito A, Lainiala O, Nieminen J, Eskelinen A. Repeated metal ion measurement in patients with bilateral metal on metal (ASRTM) hip replacements. Orthop Traumatol Surg Res 2016; 102(2): 167-73. Seppanen M, Laaksonen I, Pulkkinen P, Eskelinen A, Puhto A P, Kettunen J, Leskinen J, Manninen M, Makela K. High revision rate for large-head metal-on-metal THA at a mean of 7.1 years: a registry study. Clin Orthop Relat Res 2018; 476(6): 1223-30. Sidaginamale R P, Joyce T J, Lord J K, Jefferson R, Blain P G, Nargol A V F, Langton D J. Blood metal ion testing is an effective screening tool to identify poorly performing metal-on-metal bearing surfaces. Bone Joint Res 2013; 2(5): 84-95. De Smet K, De Haan R, Calistri A, Campbell P A, Ebramzadeh E, Pattyn C, Gill H S. Metal ion measurement as a diagnostic tool to identify problems with metal-on-metal hip resurfacing. J Bone Joint Surg Am 2008; 90(4): 202-8. 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. US Food and Drug Administration (FDA). Medical devices: Metal-on-metal hip implants, information for orthopaedic surgeons; 2019.

Acta Orthopaedica 2020; 91 (4): 383–389

383

Long-term outcomes of the hip shelf arthroplasty in adolescents and adults with residual hip dysplasia: a systematic review Koen WILLEMSEN 1, Christiaan J DOELMAN 1, Ali S Y SAM 1, Peter R SEEVINCK 2,3, Ralph J B SAKKERS 1, Harrie WEINANS 1,4, and Bart C H VAN DER WAL 1 1 Department of Orthopedics, University Medical Center Utrecht, Utrecht; 2 Department of Radiology, University Medical Center 3 MRIguidance BV, Utrecht; 4 Department of Biomechanical Engineering, Technical University Delft, Delft, The Netherlands

Utrecht, Utrecht;

Correspondence: k.willemsen-4@umcutrecht.nl Submitted 2019-08-02. Accepted 2020-03-06.

Background and purpose — The shelf arthroplasty was the regular treatment for residual hip dysplasia before it was substituted by the peri-acetabular osteotomy. Yet, evidence regarding the survival of shelf arthroplasty surgery has never been systematically documented. Hence, we investigated the survival time of the shelf procedure until revision to THA in patients with primary hip dysplasia. Factors that influenced survival and complications were also examined, along with the accuracy of correcting radiographic parameters to characterize dysplasia. Material and methods — The inclusion criteria were studies of human adolescents and adults (> 16 years) with primary or congenital hip dysplasia who were treated with a shelf arthroplasty procedure. Data were extracted concerning patient characteristics, survival time, complications, operative techniques, and accuracy of correcting radiographic parameters. Results — Our inclusion criteria were applicable to 9 studies. The average postoperative Center-Edge Angle and Acetabular Head Index were mostly within target range, but large variations were common. Kaplan–Meier curves (endpoint: conversion to THA) varied between 37% at 20 years’ follow-up and 72% at 35 years’ follow-up. Clinical failures were commonly associated with pain and radiographic osteoarthritis. Only minor complications were reported with incidences between 17% and 32%. Interpretation — The shelf arthroplasty is capable of restoring normal radiographic hip parameters and is not associated with major complications. When carefully selected on minimal osteoarthritic changes, hip dysplasia patients with a closed triradiate cartilage may benefit from the shelf procedure with satisfactory survival rates. The importance of the shelf arthroplasty in relation to peri-acetabular osteotomies needs to be further (re)explored.

The concept of shelf arthroplasty as a treatment for hip dysplasia was introduced by Franz König (1891); autologous bone is transplanted extra-articularly to extend the coverage of the femoral head by the acetabulum. Nowadays, shelf arthroplasty that relies on fibrocartilaginous changes of the capsule has mostly been replaced by treatments that reorient the patient’s own hyaline cartilage, the peri-acetabular osteotomy (PAO) being one of the most frequently used treatments (Clohisy et al. 2009). However, evidence proving the superiority of the PAO over shelf arthroplasty is lacking. A systematic review of Clohisy et al. (2009) including 13 studies concerning PAO treatment displayed conversion rates to THA between 0% and 17% during, respectively, an average follow-up of 3 and 11 years. Moreover, the PAO is a relatively invasive procedure that necessitates a long rehabilitation period, requires a long learning curve, and has major complication rates reaching as high as 37% (Clohisy et al. 2007). A systematic review concerning shelf arthroplasty survival in adolescent and adult patients has never been made. Therefore, the primary objective of this study is to systematically evaluate the long-term survival of shelf arthroplasty in adolescents and adults. As a secondary objective we evaluated factors that influence survival, the amount and type of complications, and the ability to correct radiologically dysplastic parameters to normal levels.

Material and methods For this systematic review, we consulted the databases Pubmed, Embase, and Cochrane, per search date of November 2019. The term ‘shelf’ was separately combined with the term ‘arthroplasty’ including all known synonyms to minimize the chance of missing articles (see Supplementary data).

384

Acta Orthopaedica 2020; 91 (4): 383–389

Potentially relevant cohort studies retrieved up to November 2019 (n = 181): – PubMed, 191 – EMBASE, 90 – Cochrane, 0 Removal of duplicates n = 70 Unique publications n = 111 Studies excluded (n = 76): – no patient study (e.g. biomechanical) – population not eligible – not an original article (e.g. editorial, letter, review) – language criteria – no survival reported Full text obtained n = 35 Studies excluded (n = 16): – case-report or expert opinion – less than 8 years follow-up – multiple examinations of the same cohort (longest cohort included) – >50% secondary dysplasia – >50% combined treatment – outcome not eligible Relevant publications n=9 Other relevant articles using cross-referencing n=0 Publications included in the study n=9

Figure 1. From the 111 unique publications that were found in the systematic literature search, only 9 publications were eligible for this systematic review.

Obtained articles were imported into a RefWorks database (ProQuest, Ann Arbor, MI, USA). After removal of duplicates the abstracts were read separately by 2 authors (CD, AS) in search of the inclusion criteria (Figure 1). Inclusion criteria were studies reported in the English language, population human subjects with an average age of 16 years and older with mainly primary (congenital) hip dysplasia, treated with a shelf procedure, and with follow-up of at least 8 years. Studies concerning ≥ 50% secondary hip dysplasia, e.g., due to Down syndrome, Trevor’s disease, Perthes disease, or cerebral palsy were excluded. Studies that used ≥ 50% combined dysplasia treatments, e.g., additional osteotomies, were also excluded because the influence of the combined treatment on the results is not clear. In addition, studies with an average follow-up of less than 8 years, case reports, and reviews were excluded. Studies were excluded only when there was consensus between authors (KW, CD, AS). Finally, cross-referencing was done in the bibliographies of the included studies. Each published full article was reviewed separately by 3 of the authors (KW, CD, AS). Items reviewed included age, sex,

number of patients and hips, study type, level of evidence, type of shelf procedure, type of graft used, amount of patients who were lost to follow-up, combination with other treatments, previous operations, preoperative osteoarthritic state (with scale), failure definition, survival-rates, complications, used surgical indication, amount of conversions to total hip arthroplasty at final follow-up, and the change in hip score (with scale). If documented pre- and postoperatively, the 2 hip parameters (Center Edge Angle = CEA, and Acetabular Head Index = AHI) were also reviewed and displayed graphically. Furthermore, the Newcastle Ottawa Scale (NOS) was used to assess the quality of each study and the average between 2 observers (CD and AS) was documented (Tables 1 and 2). Preoperative advanced osteoarthritis was recorded and dichotomized because different scales were used: the Tönnis and Heinecke (1999), De Mourgues and Patte (1978), Japanese Orthopedic Association (Takatori et al. 2010) and Oxford Hip Scores (Dawson et al. 1996). Because of the heterogeneity of this parameter, we distinguished between mild and advanced osteoarthritis. Therefore, on every scale the level that corresponds to advanced osteoarthritis was identified after which the number of patients who were in an advanced state of osteoarthritis were identified (Table 1). Differences in extracted information were discussed between the 3 reviewers and consensus was reached regarding the aspect in question at all times. Authors of included studies were not contacted in the event of missing data. Funding and potential conflicts of interest KW and HW have received research grants from the European Government through the Prosperos project by Interreg VA Flanders—The Netherlands program, CCI grant no. 2014TC16RFCB046 and KW, HW, PS from the Dutch government through the Netherlands Organization for Scientific Research (NWO; Applied and Engineering Sciences research program, project number 15479) in relation to the submitted work. The funders of the study had no role in the study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication. HW has also received a research grant from the Dutch Arthritis Foundation outside the submitted work. PS has owner shares in MRIguidance BV not related to the submitted work. AS, CD, BW, and RS declare no competing interests.

Results 111 unique publications were found in the databases Pubmed, Cochrane, and Embase. 9 studies remained after inclusion and exclusion criteria were applied. Cross-referencing offered no additional articles, resulting in 9 studies analyzed in this study (Tables 1 and 2).

Acta Orthopaedica 2020; 91 (4): 383–389

385

Table 1. Study characteristics Combination Preop. Type of Analyzed with other Previous advanced NOS Study Level of shelf hips/ Male/ Mean age treatment operation OA OA Reference score design a evidence procedure patients female (range) n (%) n (%) scale b n (%) Bartoníček et al. (2012) 8 R IV Bosworth (1961) 25/18 1/17 Berton et al. (2010) 8 P III Modified Roy- 17/17 NR Camille (1968) Fawzy et al. (2005) 8 R IV NR 76/67 14/53 Hamanishi et al. (1992) 6.5 R IV Spitzy (1933) 124/113 12/101 Hirose et al. (2011) 7.5 R IV Mizuno (1970) 28/26 0/26 Migaud et al. (2004) 7 R IV NR 56/48 NR Nishimatsu (2002) 7 R IV Spitzy (1933) 119/108 3/105 Saito et al. (1986) 8 R IV Mizuno (1970) 27/24 3/21 Tanaka et al. (2018) 7 R IV Modified Spitzy 35/32 2/30 (1933)

31 (16–52) 34 (20–49)

0 (0) (100) c

2 (8) NR

33 (17–60) 24 (10–53) 34 (17–54) 32 (17–56) 25 (1–56) 25 (11–55) 31 (19–49)

6 (8) 33 (27) 6 (21) NR 27 (26) NR NR

≥ 7 8 (7) NR NR NR 11 (41) (0)

TH TH

2 (8) 4 (14)

MP >32 (42) NR NR JOA 0 (0) MP 32 (57) JOA 58 (48) NR 6 (22) TH 0 (0)

NOS = Newcastle Ottawa Scale for assessing study quality; NR = Not reported a Study design: P = prospective, R = retrospective b OA scales: JOA = Japanese Orthopedic Association (Takatori et al. 2010) and Oxford Hip Scores (Dawson et al. 1996) MP = De Mourgues and Patte (1978) TH = Tönnis and Heinecke (1999) c Diagnostic arthroscopy

Table 2. Study characteristics First author Study country

Analyzed Years Conversions hips/ follow-up to THA during patients mean (range) follow-up, n (%)

Clinical outcome Hip score Hip score scale a pre- (range) post- (range)

Final score (years)

Bartoníček Czech Republic 25/18 15 (10–23) 4 (16) HHS 68 (56–82) 90 (76–100) NR Berton France 17/17 16 (16–18) 8 (47) PMA NR NR NR Fawzy England 76/67 11 (6–14) 22 (30) OHS NR NR NR Hamanishi Japan 124/113 10 (5–25) 2 (2) JOA 73 NR 86 (10) Hirose Japan 28/26 25 (20–32) 5 (18) JOA 76 NR 92 (5) > 80 (20) Migaud France 56/48 17 (15–30) 25 (45) PMA NR NR NR Nishimatsu Japan 119/108 24 (15–41) 11 (9) JOA NR 80 68 (NR) Saito Japan 27/24 13 (5–19) 2 (7) b PMA 13 16 15 (18) Tanaka Japan 35/32 26 (16–36) 10 (28) JOA 82 > 90 86 (25)

Lost to follow-up n (%) NR 1/18 (6) NR NR 29/57 (51) 5/53 (9) NR 7/31 (23) NR

NR = Not reported a Outcome scales: HHS = Harris Hip Score (Harris 1969) JOA = Japanese Orthopaedic Association (Tanaka 1978, Takeda et al. 2006) PMA = The Postel–Merle d’Aubigné (Merle d’Aubigné 1990) OHS = Oxford Hip Score (Dawson et al. 1996) b Additional undefined surgery

All the studies, except for Berton et al. (2010), are observational retrospective cohort studies without a control group. Berton et al. is a prospective cohort that stratified for the existence of labral tears. In all studies autologous cortical bone was used and placed superiorly and extra-capsularly to create an extra weight-bearing area and increase joint stability (Nishimatsu et al. 2002, Migaud et al. 2004, Fawzy et al. 2005, Berton et al. 2010, Hirose et al. 2011, Bartoníček et al. 2012, Tanaka et al. 2018). The bone was harvested from the iliac crest (Nishimatsu et al. 2002, Migaud et al. 2004, Bartoníček et al. 2012), the iliac

inner (Fawzy et al. 2005) or outer (Hirose et al. 2011, Tanaka et al. 2018) fossa. Unicortical grafts were used by 2 studies (Migaud et al. 2004, Tanaka et al. 2018) and both uni- and bicortical grafts were used by 1 study (Fawzy et al. 2005). A tectoplasty was performed in 2 studies by raising a vertical flap and filling the space with cancellous bone (Nishimatsu et al. 2002, Hirose et al. 2011). Cancellous bone was packed above the shelf by 3 studies (Fawzy et al. 2005, Bartoníček et al. 2012, Tanaka et al. 2018). Migaud et al. (2004) contained the cortical shelf by securing it with a small bent plate. The operation time of 55 minutes (35–75) was only documented

386

Acta Orthopaedica 2020; 91 (4): 383–389

Table 3. Indications for the shelf procedure and negative survival predictors as suggested by the authors Reference

Surgical indication shelf

Significant negative survival factors

Bartoníček et al. (2012) Berton et al. (2010) Fawzy et al. (2005) Hamanishi et al. (1992) Hirose et al. (2011) Migaud et al. (2004) Nishimatsu et al. (2002) Saito et al. (1986) Tanaka et al. (2018)

Dysplastic centered hip, without osteoarthritic changes, even in patients who are 60 years old Age over 18 years, dysplastic hip, (0° < CE angle < 20°), hip centered with regard to the Shenton line Mild/moderate dysplasia, minimal secondary arthritis Age under 30, pre-/early osteoarthritis, stable hip joint, with intact or uninverted labrum Moderate dysplasia, without severe osteoarthritis; however, advanced osteoarthritis in combination with femoral valgus osteotomy might be possible If peri-acetabular osteotomy is not possible because of severe subluxation or incongruency Younger age (however not < 6 years) Age under 30, no or early degenerative change Moderate dysplasia without severe osteoarthritis

Aspherity, decentration, osteoarthritic changes. Osteoarthrosis, CE angle < 0°, subluxation, labral tears (in positive-angle acetabular dysplasia) Advanced osteoarthritis, moderate/severe incongruency Age above 30, bilateral dysplasia None found Severe dysplasia (CE angle < 15°), advanced stage osteoarthrosis Older age, advanced osteoarthritis, height of the shelf Age above 30, severe degenerative changes Incorrect graft placement (too high)

CE angle = center-edge angle.

Survival (%) with THA as endpoint 100 80 60 Berton Fawzy Hirose Migaud Tanaka

40 20 0

Years after index operation Figure 2. Survival of shelf arthroplasties with years to THA as endpoint. Data for these Kaplan–Meier survival analysis results were extracted from the articles.

by Bartoníček et al. (2012). Some studies combined the shelf arthroplasty in a minor part of their total population with a varus or valgus osteotomy of the proximal femur (8–27%) (Hamanishi et al. 1992, Nishimatsu et al. 2002, Hirose et al. 2011). Berton et al. (2010) combined the shelf procedure with diagnostic arthroscopy solely to image the labral condition. No surgical alterations were made. Preoperative indications varied widely (Table 3). Early arthritis secondary to dysplasia was used as indication in 3 studies (Hamanishi et al. 1992, Nishimatsu et al. 2002, Hirose et al. 2011). Pain was used as a preoperative indication by Fawzy et al. (2005) and Bartoníček et al. (2012). Radiographic parameters were used for preoperative indications by 4 studies (Migaud et al. 2004, Berton et al. 2010, Bartoníček et al. 2012, Tanaka et al. 2018); the diagnosis ‘congenital dislocation and subluxation of the hip’ was used by 1 study (Saito et al. 1986).

Kaplan–Meier survival analysis with THA as endpoint (Figure 2) was documented by 5 studies (Migaud et al. 2004, Fawzy et al. 2005, Berton et al. 2010, Hirose et al. 2011, Tanaka et al. 2018). Fawzy et al. (2005) analyzed 76 hips from 67 patients with an average age of 33. From those shelf procedures, 86% lasted 5 years, 70% lasted 7.5 years, and 46% lasted 10 years until revision to THA. However, many hips showed advanced narrowing of the joint space preoperatively with 32 hips graded as grade IV on the De Mourgues and Patte scale (1978) (> 50% joint space narrowing). When the 44 hips with preoperative grade 3 or less only were analyzed, they found a substantially higher survival percentage of 97% at 5 years and 75% at 10 years. Berton et al. (2010) used a prospective trial to investigate the effect of the CE angle and labral tears on the shelf arthroplasty survival in a small group of patients. From the 18 patients with an average age of 34 years, 8 hips were converted to a total hip replacement at 18 years’ follow-up. This was significantly higher in the group with labral tears with 7 hips (85%) converted in 18 years of follow-up, as compared with the group without labral tears with 1 hip (17%) converted in 18 years of follow-up. Migaud et al. (2004) analyzed 56 hips in 48 patients with an average age of 32 at the time of shelf arthroplasty. From their hips, 58% survived 15 years, and 37% managed to survive for 20 years. Similarly to Fawzy et al. (2005), Migaud et al. (2004) treated 32 hips at baseline with grade III or higher on the De Mourgues and Patte scale (1978). These 32 severely osteoarthritic hips had a significant lower survival than the 24 lower graded hips, respectively 27% and 83% survival at 18 years. Hirose et al. (2011) analyzed 28 hips in 26 patients with an average age of 34 years. All had some amount of osteoarthritis but not one was graded as severe. 29 patients (51%) were lost to follow-up and were therefore not included in the analysis. All hips lasted to the 10-year mark, 93% lasted 20 years, and 71% lasted until 32 years’ follow-up. Hirose et al. (2011)

Acta Orthopaedica 2020; 91 (4): 383–389

387

All articles reported the number of conversions to THA but only in regard to their average follow-up. This ranged from 105 2% conversions in 10 years to 47% conversions in 16 years 60 (Table 2). Fawzy et al. (2005) and Migaud et al. (2004) stratiNormal 95 50 range fied their outcomes for the grade of preoperative osteoarthritis and Saito et al. (1986) for severe degenerative changes. All 85 Normal 40 range found a negative effect of preoperative advanced osteoarthritis 75 30 on the outcome of the shelf arthroplasty. In general, functional outcomes between studies were difficult 20 65 to compare because of heterogeneous clinical scoring methods 10 55 and patient characteristics (Table 2). Moreover, evaluation time Bartoníček Hirose points in relation to the surgery or the number of patients per 0 45 Fawzy Nishimatsu evaluation were often not reported. The average functional out–10 Saito 35 come improved postoperatively (Saito et al. 1986, Hamanishi et Tanaka –20 al. 1992, Hirose et al. 2011, Bartoníček et al. 2012, Tanaka et Preoperative Postoperative Preop Postop al. 2018) and this improvement lasted up to the final follow-up (Saito et al. 1986, Hamanishi et al. 1992, Hirose et al. 2011) Figure 3. The left panel) displays the average center-edge (CE) angle and the right panel the acetabular head index (AHI) pre(operative) even after 25 years of follow-up (Tanaka et al. 2018). and post(operative). Whiskers display postoperative center-edge angle Most studies documented radiological angles. Perioperaranges in relation to the healthy normal/target zone (green areas). a No range reported, 2 SD was taken as alternative. tive CE angles were documented in all studies and the AHI was measured in 5 studies (Saito et al. 1986, Nishimatsu et undertook additional survival analysis for clinical evaluation al. 2002, Berton et al. 2010, Hirose et al. 2011, Tanaka et al. and stage of joint space narrowing of 28 hips. The survival 2018). All studies that documented both preoperative and postwith joint space narrowing < stage 3 on the (0–4) scale of the operative values found a postoperative increase in average CE JOA as an endpoint was 79% at 10 years, 54% at 20 years, and angle and/or AHI (Figure 3). However, the range of surgical 21% at 32 years. Survival with a pain score of 20 (scale 0–40) correction achieved was not always within the target values as an endpoint was 100% at 10 years, 86% at 20 years, and (Figure 3). Both radiographic parameters and functional outcomes were documented in 4 manuscripts (Nishimatsu et al. 51% at 32 years. Tanaka et al. (2018) analyzed 35 hips in 32 patients with an 2002, Hirose et al. 2011, Bartoníček et al. 2012, Tanaka et al. average age of 31 years and no cases of advanced osteoarthri- 2018), yet no relation between radiographic scores and functis at the time of shelf arthroplasty. The hip survival with con- tion was reported. Rehabilitation and postoperative weightbearing was docuversion to THA as the endpoint was 91% at 25 years and 72% at 35 years. The survival with a Tönnis osteoarthritis score of mented in 6 studies with no clear consensus between the different studies (Saito et al. 1986, Hamanishi et al. 1992, Fawzy 3 or higher as the endpoint was 74% at 25 years’ follow-up. et al. 2005, Hirose et al. 2011, Bartoníček et al. 2012, Tanaka et al. 2018). NonTable 4. Reported complications of shelf procedure weightbearing walking started at 2 days to 6 weeks, partial weightbearing started at 6 Reference n (%) Complications to 8 weeks and full weightbearing started at 10 weeks to 6 months. Bartoníček et al. (2012) 5 (20) Paresthesia lateral femoral cutaneous nerve The complication rate and the back (disappeared over time) 2 (8) Too large a graft (limited external rotation of 1 hip) ground information on the complications Partial resorption of graft (still sufficient coverage) were reported by 4 articles. No major 1 (4) Extra screw fixation complications were encountered (Table 4). Non-displacement fracture of graft (after a fall) Center-edge angle (°)

Acetabular head index (%)

Fawzy et al. (2005) 10 (13) 4 (5) 3 (4) 2 (3) 1 (1) Migaud et al. (2004) 5 (9) 2 (4) 2 (4) Saito et al. (1986) 2 (7) 2 (7)

Meralgia paraesthetica Nonunion and graft breakage Superficial wound infection Bursa over metalwork (femoral osteotomy) Wound hematoma, knee stiffness after traction, flexion contracture, deep venous thrombosis, heterotopic ossification, pulmonary edema Non-unions Temporary peroneal palsies Sacroiliac pain Fracture of the base of the reflected outer cortex of the ilium Wrong shelf placement

Discussion The aim of this systematic review of the shelf arthroplasty was to describe longterm survival, the ability to correct hip dysplasia radiologically, complications, and surgical indications used. The shelf arthroplasty is considered a simple proce-

388

dure with a THA-free survival of up to 72% over a 35-year period, provided the right surgical indication is used. The THA-free survival of the shelf procedure reported in this review is comparable to those of the PAO while not being associated with major complications (Clohisy et al. 2009). However, different approaches of the PAO such as the adductor-sparing approaches could result in better recovery of the patient and fewer complications, yet long-term follow-up is still sparse (Murphy and Millis 1999). When evaluating the 5 out of 9 articles that undertook a Kaplan–Meier analysis as part of their survival analysis, the shelf procedure shows surprisingly high survival results (Figure 2). Especially so when noting that both Migaude et al. (2004) and Fawzy et al. (2005) had a high number of patients with severe preoperative osteoarthritis and Berton et al. (2010) had many cases with an existing labral tear. Both the advanced osteoarthritic and labral tear patients had significantly inferior results as compared with patients without osteoarthritis or labral tears. When fewer patients with advanced osteoarthritis were included, as in the studies of Hirose et al. (2011) and Tanaka et al. (2018), the THA-free survival percentage even reached 72% at 35 years of follow- up. These survival results are in line with a recent study by Holm et al. (2017), who reported very long shelf survival rates in children and adolescents. That study was not included in this systematic analysis because the average age of 56 patients (70 hips) was only 12 years (5–22), an average age that was too low for the inclusion criteria. Holm et al. (2017) reported a THA-free survival percentage of 100% at 20 years, 83% at 30 years, and up to 22% at 50 years. In a separate report from the same hospital, Terjesen (2018) made a subanalysis for the age group > 12 years (average age 16.1 years). The Kaplan–Meier analysis showed a survival of 100% at 20 years, 72% at 30 years, and 32% at 40 years of follow-up. However, because it concerned a sub-analysis many specifics were not given (e.g., number of patients, sex, average followup, combinations with other treatment, previous operations, preoperative osteoarthritis scale, clinical hip score, and lostto-follow up) and therefore the study was not included in this review. The shelf survival values resemble or are even better than PAO survival in the long term (Schramm et al. 2003, Hasegawa et al. 2014, Lerch et al. 2017). Nonetheless, the shelf arthroplasty is considered a salvage procedure, while the peri-acetabular osteotomy is considered to be joint-preserving surgery. Once again, this raised the question as to whether the shelf procedure should be reconsidered in the palette of treatment options for residual hip dysplasia. Klaue et al. (1993) noticed that a normal CE angle on a radiograph after a shelf arthroplasty is commonly an overestimation when compared with the true femoral coverage on a CT scan. Therefore, parameters such as the CE angle and the AHI might be overestimated. Nevertheless, new 3D planning and evaluation techniques can overcome difficulties in graft placement and improve the effectiveness of correcting the radiological dysplastic parameters in all dimensions (Figure

Acta Orthopaedica 2020; 91 (4): 383–389

3). However, it should be noted that the shelf arthroplasty does not change the hyaline cartilage but rather induces fibrocartilaginous metaplasia of the joint capsule to increase the amount of weight-bearing tissue. Evaluation of the literature shows substantial limitations. First, the level of evidence was low: 8 out of 9 articles were retrospective with level IV evidence and only Berton et al. (2010) was prospective with level III evidence (Table 1). Lowlevel evidence is common in orthopedics studies as different surgical techniques are often difficult to compare (Obremskey et al. 2005). The included studies used 6 different modifications of the shelf procedure and all had a different postoperative rehabilitation process. The effects of these differences on the outcome were not clear. Second, the investigated population could be considered a limitation as 5 out of 9 studies were completed in Japan, which has a population with a well-known higher incidence of hip dysplasia (Nakamura et al. 1989). Furthermore, far more women participated in the studies investigated, which could have influenced the results, but none of the included studies stratified for sex. Another limitation could be the search syntax. Additional unknown nomenclature for the shelf arthroplasty could have influenced the effectiveness of the search syntax. However, cross-referencing did not provide any additional articles, causing the impact of this aspect to be low, presumably. Lost to follow-up was not documented in Fawzy et al. (2005) and Nishimatsu et al. (2002). Therefore, selection bias could have occurred. Only 2 studies documented the number of patients who died before final follow-up. Berton et al. (2010) reported 2 “unrelated” deaths and Migaud et al. (2004) noted 2 deaths without further explanation. Another type of selection bias may arise from the lack of consensus on the correct indication for performing a shelf procedure. For example, studies that included patients with incongruency and advanced osteoarthritis showed lower survival of the shelf arthroplasty (Migaud et al. 2004, Fawzy et al. 2005). Saito et al. (1986), Berton et al. (2010) and Bartoníček et al. (2012) included only a few patients with severe osteoarthritis (8–22%), Nishimatsu et al. (2002), Migaud et al. (2004) and Fawzy et al. (2005) included roughly half of their patients with severe osteoarthritis (42–57%), while Hirose et al. (2011) and Tanaka et al. (2018) included no patients with severe osteoarthritis. Differences were also found in inclusion of aspheric hips (Migaud et al. 2004) or spheric hips (Bartoníček et al. 2012), younger patients (Saito et al. 1986, Hamanishi et al. 1992, Nishimatsu et al. 2002) or older patients (Berton et al. 2010) even up to their 6th decade (Bartoníček et al. 2012). An additional evident selection bias was introduced by Migaud et al. (2004) who considered shelf arthroplasty as salvage only in patients not eligible for a peri-acetabular osteotomy. Conclusion The shelf arthroplasty is competent in restoring radiographic hip parameters to normal levels, increases functional outcomes,

Acta Orthopaedica 2020; 91 (4): 383–389

and is not associated with major complications. When selected on minimal osteoarthritic changes, adolescent and adult hip dysplasia patients may benefit from the shelf procedure with satisfactory survival rates. Therefore, based on the findings in this review, the indications for shelf arthroplasty should more often be considered in the treatment of residual hip dysplasia, especially with regard to the difficult-to-perform peri-acetabular osteotomy surgery. Given the constant development of 3D-planning techniques, shelf placement can even be further optimized and therefore may increase its clinical effectiveness. Supplementary data Search strategies are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/ 17453674.2020.1747210

Conceptualization: KW, BW; methodology: KW, CD, AS; data curation: KW, CD, AS; validation: KW; RS, BW, HW; supervision: KW; RS, HW, RS; writing, review, and editing: KW, CD, AS, PS, BW, RS, HW. Acta thanks Michael Brian Millis and Terje Terjesen for help with peer review of this study. Bartoníček J, Vávra J, Chochola A. Bosworth hip shelf arthroplasty in adult dysplastic hips: ten to twenty-three year results. Int Orthop 2012; 36(12): 2425-31. Berton C, Bocquet D, Krantz N, Cotten A, Migaud H, Girard J. Shelf arthroplasties long-term outcome: influence of labral tears. A prospective study at a minimal 16 years’ follow-up. Orthop Traumatol Surg Res 2010; 96(7): 753-9. Bosworth D M, Fielding JW, Ishizuka T, Ege R. Hip-shelf operation in adults. J Bone Joint Surg 1961; 43(1): 93-106. Clohisy J C, Nunley R M, Curry M C, Schoenecker P L. Periacetabular osteotomy for the treatment of acetabular dysplasia associated with major aspherical femoral head deformities. J Bone Joint Surg–Ser A 2007; 89(7): 1417-23. Clohisy J C, Schutz A L, John L S, Schoenecker P L, Wright R W. Periacetabular osteotomy: a systematic literature review. Clin Orthop Relat Res 2009; 467(8): 2041-52. Dawson J, Fitzpatrick R, Carr A, Murray D. Questionnaire on the perceptions of patients about total hip replacement. J Bone Joint Surg Br 1996; 78(2): 185-90. De Mourgues G, Patte D. Résultats, après au moins 10 ans, des ostéotomies d’orientation du col du fémur dans les coxarthroses secondaires peu évoluées chez l’adulte: symposium. Rev Chir Orthop 1978; 64(7): 525-9. Fawzy E, Mandellos G, De Steiger R, McLardy-Smith P, Benson M K D, Murray D. Is there a place for shelf acetabuloplasty in the management of adult acetabular dysplasia? Bone Joint J 2005; 87(9): 1197-202. Hamanishi C, Tanaka S, Yamamuro T. The Spitzy shelf operation for the dysplastic hip retrospective 10 (5–25) year study of 124 cases. Acta Orthop Scand 1992; 63(3): 273-7. Harris W H. Traumatic arthritis of the hip after dislocation and acetabular fractures: treatment by Mold arthroplasty. An end-result study using a new method of result evaluation. J Bone Joint Surg Am 1969; 51(4): 737. Hasegawa Y, Iwase T, Kitamura S, Kawasaki M, Yamaguchi J. Eccentric rotational acetabular osteotomy for acetabular dysplasia and osteoarthritis: follow-up at a mean duration of twenty years. J Bone Joint Surg 2014; 96(23): 1975-82.

389

Hirose S, Otsuka H, Morishima T, Sato K. Long-term outcomes of shelf acetabuloplasty for developmental dysplasia of the hip in adults: a minimum 20-year follow-up study. J Orthop Sci 2011; 16(6): 698-703. Holm A G V, Reikerås O, Terjesen T. Long-term results of a modified Spitzy shelf operation for residual hip dysplasia and subluxation: a fifty-year follow-up study of fifty-six children and young adults. Int Orthop 2017; 41(2): 415-21. Klaue K, Sherman M, Perren S M, Wallin A, Looser C, Ganz R. Extra-articular augmentation for residual hip dysplasia: radiological assessment after Chiari osteotomies and shelf procedures. Bone Joint J 1993; 75(5): 750-4. König F. Osteoplastische Behandlung der congenitalen Hüftgelenkluxation. Verh Deutsch Ges Chir 1891; 20: 75-80. Lerch T D, Steppacher S D, Liechti E F, Tannast M, Siebenrock K A. Onethird of hips after periacetabular osteotomy survive 30 years with good clinical results, no progression of arthritis, or conversion to THA. Clin Orthop Relat Res 2017; 475(4): 1154-68. Merle d’Aubigné R. Cotation chiffrée de la fonction de la hanche. Rev Chir Orthopédique 1990; 76(6): 371-4. Migaud H, Chantelot C, Giraud F, Fontaine C, Duquennoy A. Long-term survivorship of hip shelf arthroplasty and Chiari osteotomy in adults. Clin Orthop Relat Res 2004; (418): 81-6. Mizuno S. Short Lectures in Orthopaedic Surgery, Illustrated. Tokyo: Ishiyaku Publishers1970. PP: 27-39. Murphy S B, Millis M B. Periacetabular osteotomy without abductor dissection using direct anterior exposure. Clin Orthop Relat Res 1999; (364): 92-8. Nakamura S, Ninomiya S, Nakamura T. Primary osteoarthritis of the hip joint in Japan. Clin Orthop Relat Res 1989; (241): 190-6. Nishimatsu H, Iida H, Kawanabe K, Tamura J, Nakamura T. The modified Spitzy shelf operation for patients with dysplasia of the hip. Bone Joint J 2002; 84(5): 647-52. Obremskey W T, Pappas N, Attallah-Wasif E, Tornetta III P, Bhandari M. Level of evidence in orthopaedic journals. J Bone Joint Surg 2005; 87(12): 2632-8. Roy-Camille R. Butée ostéoplastique armée de hanche. Technique permettant la marche au 21e jour. Press Med 1968; 76: 273-5. Saito S, Takaoka K, Ono K. Tectoplasty for painful dislocation or subluxation of the hip: long-term evaluation of a new acetabuloplasty. Bone Joint J 1986; 68(1): 55-60. Schramm M, Hohmann D, Radespiel-Troger M, Pitto R P. Treatment of the dysplastic acetabulum with Wagner spherical osteotomy: a study of patients followed for a minimum of twenty years. J Bone Joint Surg 2003; 85(5): 808-14. Spitzy H. Kunstliche Pfannendachbildung: Benutzung von Knochenbolzen zur temporären Fixation. Z Orthop Chir 1933; 58: 470-86. Takatori Y, Ito K, Sofue M, Hirota Y, Itoman M, Matsumoto T, Hamada Y, Shindo H, Yamada H, Yasunaga Y. Analysis of interobserver reliability for radiographic staging of coxarthrosis and indexes of acetabular dysplasia: a preliminary study. J Orthop Sci 2010; 15(1): 14-19. Takeda H, Kamogawa J, Sakayama K, Kamada K, Tanaka S, Yamamoto H. Evaluation of clinical prognosis and activities of daily living using functional independence measure in patients with hip fractures. J Orthop Sci 2006; 11(6): 584-91. Tanaka H, Chiba D, Mori Y, Kuwahara Y, Baba K, Yamada N, Fujii G, Itoi E. Long-term results of a modified Spitzy shelf operation for developmental dysplasia of the hip in adults and adolescents. Eur J Orthop Surg Traumatol 2018; 28(7): 1341-7. Tanaka S. Surface replacement of the hip joint. Clin Orthop Relat Res 1978; 134:75-9. Terjesen T. Current concepts review—residual hip dysplasia: is there a place for hip shelf operation? J Child Orthop 2018; 12: 358-63. Tönnis D, Heinecke A. Current concepts review—acetabular and femoral anteversion: relationship with osteoarthritis of the hip. J Bone Joint Surg 1999; 81(12): 1747-70.

390

Acta Orthopaedica 2020; 91 (4): 390–395

Effective radiation dose in radiostereometric analysis of the hip Ian F BLOM 1a, Lennard A KOSTER 2a, Bart TEN BRINKE 3, and Nina M C MATHIJSSEN 3 1 Department of Radiology, Reinier de Graaf Groep, Delft, The Netherlands; 2 Department of Orthopaedic Surgery, Leids Universitair Medisch Centrum, Leiden, The Netherlands; 3 Department of Orthopaedic Surgery, Reinier de Graaf Groep, Delft, The Netherlands a Shared first authorship Correspondence: i.blom@rdgg.nl Submitted 2019-09-28. Accepted 2020-04-02.

Background and purpose — Radiostereometric analysis (RSA) is the gold standard to study micromotion of joint replacements. RSA requires the acquisition of additional radiographs increasing the radiation dose of patients included in RSA studies. It is important to keep this dose as low as possible. Effective radiation dose (ED) measurements of RSA radiographs for different joints were done by Teeuwisse et al. some years ago using conventional radiology (CR); for total hip arthroplasty (THA), Teeuwisse et al. reported an ED of 0.150 milliSievert (mSv). With the modern digital radiography (DR) roentgen technique the ED is expected to be less. Material and methods — In this phantom study, simulating a standard patient, the ED for hip RSA radiographs is determined using DR under a variety of different roentgen techniques. The quality of the RSA radiographs was assessed for feasibility in migration analysis using a (semi-)automatic RSA analysis technique in RSA software. Results — A roentgen technique of 90 kV and 12.5 mAs with additional 0.2 copper (Cu) + 1 mm aluminum (Al) external tube filters results in an ED of 0.043 mSv and radiographs suitable for analysis in RSA software. Interpretation — The accumulated ED for a standard patient in a 2-year clinical hip RSA study with 5 follow-up moments and a double acquisition is below the acceptable threshold of 1.0 mSv provided by the EU radiation guideline for studies increasing knowledge for general health.

Since its introduction in 1970 by Göran Selvik (1989), radiostereometric analysis (RSA) has frequently been used to study micromotion of orthopedic implants (Kärrholm et al. 1994, Ryd et al. 1995, Valstar et al. 2002). An RSA study consists of several follow-up moments, each requiring 2 simultaneously taken radiographs, in addition to regular imaging. This results in an increased radiation dose for patients in an RSA study. As stated in the RSA ISO standard, precision of an RSA study needs to be determined with double examinations, adding another RSA radiograph (ISO 16087:2013). In several countries medical ethics committees do not permit the acquisition of double RSA examinations because of the added radiation dose (Valstar et al. 2005). In general, a combination of increasing kiloVoltage (kV) and decreasing milliAmpere-seconds (mAs) results in a decrease in radiation dose. However, decreasing the radiation dose results in a lower image quality (Bushong 1975, Fauber et al. 2011, Carroll 2014, Ma et al. 2014). Decreasing radiation dose, while the image quality remains acceptable for the purpose, is called the As Low As Reasonably Achievable (ALARA) principle (ICRP 1955). According to Teeuwisse et al. (1998) an RSA radiograph of the hip, using computed radiography (CR) roentgen detectors, has an effective radiation dose (ED) of 0.150 miliSievert (mSv) (Valstar 2001). However, most RSA studies do not provide the ED of the applied roentgen technique and thus the actual radiation dose remains unknown. Currently, digital radiography (DR) roentgen systems are becoming the standard that provides better image quality with similar or lower radiation dose (Bragdon et al. 2003, Ching et al. 2014). In addition to DR detectors, modern mobile X-ray tubes contain external tube filters, which can be applied to improve image quality.

Acta Orthopaedica 2020; 91 (4): 390–395

In order to assess the dose given to hip patients in an RSA study, we: 1. determined the ED of an RSA radiograph with standard roentgen settings in a hip phantom model using a DR roentgen system with and without external tube filters; and 2. determined the optimal roentgen settings in a hip phantom model using a DR roentgen system.

Material and methods RSA set-up 2 DR system roentgen tubes were used: 1 fixed ceiling tube (DigitalDiagnost, Philips, Best, the Netherlands) and a mobile tube (MobileDiagnost wDR, Philips, Best, the Netherlands). Both tubes were used in combination with a Philips large Skyplate detector of 35 × 43 cm (2,330 ×2,846 pixels, image resolution 165 dpi). Roentgen images were saved as lossless JPGs for analysis in model-based RSA. The tubes were positioned over a carbon calibration cage (Carbonbox number 20, Medis Specials BV, Leiden, the Netherlands) angulated at 20° to the vertical and the roentgen beams were collimated to fit the indicated areas, with sizes of 34 × 41 cm, on the calibration cage. The source–image distance (SID) for this study was 160 cm, which is similar to clinical RSA studies using this type of calibration cage. The object-image distance (OID) is 40 cm and therefore the source–object distance (SOD) is 120 cm. Effective radiation dose An Alderson phantom was used to simulate an adult pelvis including soft tissue. A Piranha-meter Multi 657 (RTI, Mölndal, Sweden), with backscatter protection, was positioned in the RSA set-up on top of the Alderson phantom to measure absorbed dose (AD) (Figure 1) (Bushong 1975, Carroll 2014). Based on the measured AD, the ED was calculated for each tube using PCXMC software including all pelvic tissue weighting factors of ICRP publication 103 (STUK, Helsinki, Finland version 2.0.1.4) (ICRP 1991, Veldkamp et al. 2012). The standard roentgen settings as indicated for RSA radiographs of the hip by Valstar were used to determine ED in this study (Valstar 2001). For the medio-lateral tube 73 kV, 25 mAs; latero-medial 90 kV, 12.5 mAs and no external tube filtration. RSA images were also made with the standard roentgen settings and different external tube filters (2 mm Al, 0.1 Cu + 1 mm Al, and 0.2 Cu + 1 mm Al). Optimal roentgen settings In this study we define optimal roentgen settings as the settings where ED is the lowest in combination with RSA images that are of such quality that they can be used for RSA analysis. In order to determine the ED for different roentgen settings, the same set-up was used as to determine the ED for standard settings, but now with a variety of different, but identical for both tubes, roentgen settings (Table 1).

391

Table 1. The different settings for tube voltage (kV), current exposure time product (mAs), and external tube filters, in all possible combinations, that were used to determine ED and image quality Tube voltage (kV): Exposure (mAs): External tube filter:

77, 85, 90 and 102 8, 12.5, 16 and 20 No filtration, 2 mm Al, 0.1 mm Cu + 1 mm Al, and 0.2 mm Cu + 1 mm Al

All the ED measurements, including standard roentgen settings with and without external tube filtering, were performed twice. The largest of the 2 calculated EDs for each combination of roentgen settings is reported. All RSA measurements were performed by 1 author (IB) and calculation of 10% of the RSA measurements, selected by a random number generator, was performed by a second author (LAK). In addition, another 10% of the RSA measurements was performed twice by IB (no differences were detected). To assess the image quality, RSA images were acquired with all the combinations of roentgen settings, including standard settings with and without external tube filtering. To mimic a standard adult patient a phantom was placed in a Perspex box with walls of 12 mm filled with 24.8 cm water (Figure 2) (Slade-Schaaphok 2016). Perspex has almost the same density as human tissue and therefore the backscatter of the Perspex is similar to backscatter of human tissue (SladeSchaaphok 2016). Furthermore, water is a good approximation of soft tissue of a human body (Sandborg 1990, SladeSchaaphok 2016). The phantom model consisted of an Allofit Acetabular Cup with a highly crosslinked polyethylene liner (titanium outershell, size 54 mm, Zimmer Biomet, Warsaw, IN, USA) surgically placed in a hemipelvic sawbone (Sawbones, Vashon Island, WA, USA). 18 tantalum markers (1.0 mm diameter) were attached to the acetabulum (Figure 3), part cranially of the cup, part in the ischium bone. In between the RSA acquisitions the phantom and tube positions were left unchanged. Acceptable image quality was defined as an image suitable for analysis with model-based RSA (version 4.2014, RSAcore, department of Orthopedic Surgery, LUMC, Leiden, the Netherlands). Default model-based RSA settings require at least 6 non-collinear fiducial markers and 4 control markers to be correctly detected in each X-ray image in order to analyze an RSA radiograph (Kaptein et al. 2003). Furthermore, contour difference should be below 0.2 for optimal pose estimation of the CAD model, as defined by a performed phantom study of the model (mean contour difference of 0.2). For migration calculation (translation and rotation) at least 3 3D bone markers are required which need to meet the ISO standard (mean error of rigid body matching < 0.35 and condition number < 150 ms-1) (ISO 16087:2013). For all different RSA radiographs the number of automatically detected calibration cage markers with Hough threshold of 16, the number of automatically detected markers attached

392

Figure 1. RSA set-up with Alderson phantom and Piranha meter (within the green circle) positioned on top of the phantom to measure the entrance dose. Source–image distance (SID) = 160 cm, object-image distance (OID) is 40 cm and source–object distance (SOD) is 120 cm. Tube angulation is 20° for both tubes.

Acta Orthopaedica 2020; 91 (4): 390–395

Figure 2. Sawbone of the hemipelvis placed in the Perspex box filled with 24.8 cm vertical water column.

to the pelvic sawbone, and the length of the automatically detected contour of the cup were scored. The longest contour detected within the region of contour detection around the acetabular cup projection, set by the analyst, was automatically selected by the software. Other contours that were considered by the analyst to be part of the acetabular cup contour were manually selected. Contours were not cut into smaller pieces. Funding and potential conflicts of interest The authors declare no conflicts of interest condition number.

Results Effective dose Images with standard roentgen settings with no external tube filtration, 0.1 Cu + 1 Al and 0.2 Cu + 1 Al external tube filtration, have acceptable image quality. The lowest calculated ED was 0.044 mSv, on standard settings with an external tube filter of 0.2 Cu + 1 Al (Table 2). The maximum calculated ED was 0.094 mSv with standard settings and without external tube filtration.

Figure 3. Phantom: sawbone of the hemipelvis, with the Allofit acetabular cup with a polyethylene liner (size 54 cm diameter) with tantalum markers attached around the acetabular bone.

Optimal roentgen settings Table 3 shows the results on ED and image quality of all tested combinations of roentgen settings (identical for both tubes). The lowest ED with acceptable image quality was 0.043 mSv with a roentgen technique of 90 kV, 12.5 mAs, and 0.2 Cu + 1 Al. The highest ED with acceptable image quality was 0.223 mSv with a roentgen technique of 102 kV, 25 mAs, and without external tube filtration. Between the lowest and highest ED with acceptable image quality is a difference of 0.18 mSv. Tables 4 and 5 (see Supplementary data) provide the results of the ED measurements and the different roentgen settings on the RSA parameters used to determine image quality for RSA analysis. Table 2. ED (mSv) with standard roentgen settings (medio-lateral 73 kV, 25 mAs; latero-medial 90 kV, 12.5 mAs, no external filtration) 0.1 mm Cu + 0.2 mm Cu + Item No filtration 2 mm Al 1 mm Al 1 mm Al ED (mSv) a Setting

0.094

0.072 a 0.061 0.044

that result in poor image quality; other settings resulted in acceptable image quality.

Acta Orthopaedica 2020; 91 (4): 390–395

393

Table 3. ED (mSv) with various, but identical for both tubes, roentgen settings Filtration/Voltage 0 mm Al 77 kV 85 kV 90 kV 102 kV 2 mm Al 77 kV 85 kV 90 kV 102 kV 0.1 mm Cu + 1 mm Al 77 kV 85 kV 90 kV 102 kV 0.2 mm Cu + 1 mm Al 77 kV 85 kV 90 kV 102 kV

8 mAs

12.5 mAs 16 mAs

25 mAs

0.033 a 0.052 a 0.066 0.104 a 0.044 a 0.069 0.088 0.138 0.051 a 0.080 a 0.103 a 0.159 0.071 a 0.111 0.143 0.223 0.026 a 0.040 a 0.051 a 0.081 0.035 a 0.054 a 0.069 0.108 a 0.041 a 0.064 a 0.081 0.128 0.057 a 0.089 0.114 0.179 0.022 a 0.034 a 0.044 0.069 0.030 a 0.047 a 0.060 0.094 0.036 a 0.056 a 0.071 0.112 0.051 a 0.079 0.102 a 0.159 0.016 a 0.025 a 0.032 a 0.051 a 0.023 a 0.036 a 0.046 a 0.072 a 0.027 a 0.043 0.055 0.087 0.040 a 0.063 0.080 a 0.125

a Setting

that result in poor image quality; other settings resulted in acceptable image quality.

Discussion In this study we evaluated the ED of an RSA radiograph with standard roentgen settings in a hip phantom model using DR roentgen systems. Furthermore, we determined the optimal roentgen settings to acquire RSA radiographs in a hip phantom model that are acceptable for analysis while minimizing the radiation dose. We hypothesized that with the introduction of DR roentgen systems, the ED of hip RSA acquisitions could be reduced compared with CR roentgen systems while maintaining acceptable image quality for RSA analysis. In this study the RSA set-up consists of a fixed roentgen tube and a mobile roentgen tube. Several RSA research sites use a high-end system with 2 fixed tubes. It is assumed that this can result in a lower ED than measured in our study, because of better internal tube filtration and more powerful generators to generate higher roentgen techniques. Based on the standard roentgen settings recommended by Valstar (2001) the ED of a single RSA radiograph of the hip using 2 DR systems was 0.094 mSv. This is lower compared with the reported ED of 0.150 mSv by Teeuwisse et al. (1998), who used the same settings in combination with CR systems. To our knowledge there is only 1 other publication that reports the ED of hip RSA radiographs in a phantom study, although image quality was not objectified in that study (Brodén et al. 2016). Furthermore, we used external tube filters to lower the ED, which has not been reported in any other RSA study so far. Our results show that the optimal roentgen setting, a combination of the standard roentgen settings recommended by

Valstar in combination with 0.2 Cu + 1 Al external tube filter, resulted in an ED of 0.043 mSv. However, this ED is based on imaging a standard patient. When patients have larger BMI, the roentgen technique should be adapted, resulting in a higher ED as larger kV and mAs are necessary. It is not expected that the roentgen technique resulting in the highest calculated ED (0.223 mSv) measured in this study is necessary for the THA patient with a larger BMI compared with the standard patient, in order to acquire images suitable for RSA analysis. CT-based RSA of the hip looks promising, with a calculated ED of 0.33 mSv for an experimental hip study (Brodén et al. 2016) and 0.2–2.3 mSv for a clinical hip study (Brodén et al. 2020). However, the range in radiation dose in this clinical study is quite wide. Though CT-based RSA could have advantages over RSA, further optimization of CT protocols is necessary to reduce the radiation dose in order to achieve acceptable radiation dose for patients in a long-term follow-up migration study. Clinical implications A standard RSA study typically consists of 6 RSA radiographs: 5 follow-up moments and a double examination. With optimal roentgen settings this results in a cumulative ED of approximately 0.26 mSv. RSA studies mostly fall into the category of increasing knowledge leading to a health benefit for the population, which is classified as a category IIa study. The acceptable cumulative ED for category IIa studies, according to EU guideline ‘Radiation Protection 99’, is 0.1–1.0 mSv (European Commission 1998). When adults over 50 years of age are participating in a category IIa study, the thresholds can be increased 5- to 10-fold, resulting in minimum thresholds of 0.5–5.0 mSv (ICRP 2007). Our results show that the cumulative ED for a standard patient in a hip RSA study is far below the upper threshold applicable for this kind of study when the optimal settings are used for DR roentgen systems. Even using the highest calculated ED in this study for a single RSA radiograph, the cumulative ED is approximately 1.4 mSv. Adjustment of roentgen settings for patients with a higher BMI compared with the standard patient is thus unlikely to result in an ED above the 5.0 mSv. Limitations Our study has several limitations. To determine the ED we used an Alderson phantom, which is based on a standard patient with a BMI of 25.4. Therefore, the assessment of image quality was also performed mimicking a standard patient. Arthroplasty patients in our general RSA hip studies have a BMI of 28.1. Due to more soft tissue around the hip joint more radiation is necessary to obtain an acceptable roentgen image and, hence, these patients receive a higher ED. However, even without considering the age factor, which is usually over 50, adding additional kV and mAs will not result in unacceptable ED. For each individual patient the radiology assistant might change the roentgen settings to acquire acceptable-quality images. For some individual patients this results in higher ED

394

compared with the results presented in this paper. It is important, however, to acquire acceptable-quality images suitable for analysis. In that perspective it is probably better to overcompensate the roentgen settings for patients with more soft tissue, than to be cautious with the radiation applied and then have to retake the RSA radiograph with higher settings due to poor image quality. The calculated ED is the sum of the ED from the PCXMC software for each tube using the measured AD. Though this is an overestimation of the real ED, this calculated ED is lower than the known values in the literature or calculated using DoseAreaProduct calculations. Another limitation of our study is the use of a hip phantom model. We used a hip phantom model because the roentgen settings, and therefore the ED, in a hip RSA study are higher compared with the settings in other joints in the extremities and because of the nature of the irradiated tissue. Though ED in the shoulder and the spine is probably higher, we opted to use the hip model as this is, together with the knee joint, the most frequently studied joint in RSA research. For spine and shoulder joints, the results of our study cannot be taken as an indication, but for other joints in the extremities we can confidently say that the ED for RSA acquisitions is below that of the hip. Normal DR of these joints has a lower ED than the hip (RIVM 2011). In this study we have used Philips X-ray DR systems in combination with Skyplate detectors, a uniplanar carbon cage and a metal backed acetabular cup. There are, however, many different combinations of roentgen systems, RSA cages, and prostheses available. All the aforementioned factors will likely have an effect on the necessary roentgen settings and the ED for the patient. However, we do not expect that the variability of these parameters will results in an increase in the ED above the acceptable threshold of 1.0 mSv for patients under 50 years of ages (ICRP 2007). Regarding image quality, we have used the default modelbased RSA software settings with automatic marker detection and labelling option active. Based on the experience with the hardware and the visual evaluation of the images by the RSA analysts, the Hough threshold for marker detection was decreased. This resulted in a combination of roentgen settings, image quality, and detection settings that can be used for analysis without manually adjusting marker projections and contours. Experienced RSA analysts will be able to use even poorer image quality; however, the analysis of these images will require more time and migration results might be more sensitive to image noise. The applied combination of roentgen settings can be optimized even further and could be made specific for each joint in the extremities and for different kind of prostheses or roentgen hardware used. We believe however, that the optimization for these possibilities will result in even lower ED and this will not have any implications for the use of RSA in standard RSA studies considering the applicable radiation threshold for research purposes.

Acta Orthopaedica 2020; 91 (4): 390–395

Conclusion The lowest calculated ED of an RSA radiograph with standard roentgen settings and an external tube filter of 0.2 Cu + 1 mm Al was 0.044 mSv. This is more than 0.1 mSv lower than the given 0.150 mSv as stated by Teeuwisse et al. (1998). With modern DR equipment, the roentgen technique for both tubes of 90 kV, 12.5 mAs, and 0.2 mm Cu + 1 mm Al gave the optimal result: an ED of 0.043 mSv and good image quality. The accumulated ED for a patient in a 2-year clinical hip RSA study with 5 FU moments and a double acquisition is below the acceptable threshold of 1.0 mSv provided by the EU radiation guideline for studies increasing knowledge for the general health of the population. The double examination in a regular RSA study is essential to determine the clinical precision of RSA (ISO 16087:2013). Though it is an additional RSA acquisition, this does not result in exceeding the threshold for ED of 1.0 mSv (ICRP 2007). As a result the additional radiation dose from the double examination does not have to be a reason for Medical Ethics Committees to prohibit the acquisition of the double examination (Valstar et al. 2005). Supplementary data Tables 4 and 5 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674. 2020.1767443 IB designed the study. IB, LAK, and NM performed the data collection. IB and LAK performed the data analysis. IB, LAK, and NM wrote the initial draft of the manuscript. IB, NM, LAK, and BtB critically reviewed the manuscript. Acta thanks Niels Egund and Johan Kärrholm for help with peer review of this study.

Bragdon C R, O’Keefe, M C, Harris, W H. Radiostereometric analysis (RSA) studies at Massachusetts General Hospital. Ortho J HMS 2003; 5. Brodén C, Olivecrona H, Maguire G Q Jr, Noz M E, Zeleznik M P, Skoldenberg O. Accuracy and precision of three-dimensional low dose CT compared to standard RSA in acetabular cups: an experimental study. Biomed Res Int 2016; 2016: 5909741. Brodén C, Sandberg O, Sköldenberg O, Stigbrand H, Hänni M, Giles J W, Emery R, Lazarinis S, Nyström A, Olivecrona H. Low-dose CT-based implant motion analysis is a precise tool for early migration measurements of hip cups: a clinical study of 24 patients. Acta Orthop 2020 Feb 14: 1-6. doi: 10.1080/17453674.2020.1725345. [Epub ahead of print] Bushong S C. Radiologic science for technologists, physics, biology and protection. 1st ed. Maryland Heights, MS: Mosby; 1975. ISBN 0801609151. Carroll Q B. Radiography in the digital age. 2nd ed. Springfield, IL: Charles C Thomas; 2014. ISBN 0398080968 . Ching W, Robinson J, McEntee M. Patient-based radiographic exposure factor selection: a systematic review. J Med Radiat Sci 2014; 61(3): 176-90. European Commission Directorate General Environment, Nuclear Safety and Civil Protection. Guidance on medical exposures in medical and biomedical research. Brussels: European Commission; 1998. Fauber T L, Cohen T F, Dempsey M C. High kilovoltage digital exposure techniques and patient dosimetry. Radiol Technol 2011; 82(6): 501-10.

Acta Orthopaedica 2020; 91 (4): 390–395

ICRP. Recommendations of the International Commission on Radiological Protection. Br J Radiol 1955; (Suppl. 6). ICRP. 1990 Recommendations of the International Commission on Radiological Protection. ICRP Publication 60. Ann ICRP 1991; 21(1-3). ICRP. The 2007 Recommendations of the International Commission on Radiological Protection. ICRP Publication 103. Ann ICRP 2007; 37 (2-4). Kaptein B L, Valstar E R, Stoel B C, Rozing P M, Reiber J H C. A new modelbased RSA method validated using CAD models and models from reversed engineering. J Biomech 2003; 36: 873-82. Kärrholm J, Borssén B, Löwenhielm G, Snorrason F. Does early micromotion of femoral stem prostheses matter? 4-7-year stereoradiographic follow-up of 84 cemented prostheses. J Bone Joint Surg Br 1994; 76-B(6): 912-17. Ma W, Hogg P, Norton S. Effects of kilovoltage, milliampere seconds, and focal spot size on image quality. Radiol Technol 2014; 85: 479-85. RIVM. Gemiddelde effectieve dosis voor verschillende typen röntgenverrichtingen https: //www.rivm.nl/medische-stralingstoepassingen/trends-enstand-van-zaken/diagnostiek/r-ntgen-exclusief-ct/gemiddelde-effectievedosis-voor-verschillende-typen-r-ntgenverrichtingen; 2011. 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-B(3): 377-83. Sandborg M. Comparison between lucite and water as a phantom material

395

in medical radiology. Progress in Nuclear Energy 1990; 24(1–3): 355-64. Selvik G. Roentgen stereophotogrammetry: a method for the study of the kinematics of the skeletal system. Acta Orthop Scand 1989; 60 (Suppl. 232): 1-51. Slade-Schaaphok R. De effecten van CuAl filter op de intreedosis, effectieve dosis en de beeldkwaliteit NVMBR. Digitaal Magazine “MBB’er in Beeld” 2016 (March): 1-4. Teeuwisse W, Berting R, Geleijns J. Stralenbelasting bij orthopedische radiologie. Gamma 1998; 8-9: 197-200. Valstar E. Digital roentgen stereophotogrammetry: development, validation, and clinical application. Thesis University of Leiden 2001. ISBN 90-9014397-1. Valstar E R, Nelissen R G H H, Reiber J H C, Rozing P M. The use of roentgen stereophotogrammetry to study micromotion of orthopaedic implants. ISPRS J Photogramm Remote Sens 2002; 56(5-6): 376-89. Valstar E R, Gill R, Ryd L, Flivik G, Borlin N, Kärrholm J Guidelines for standardization of radiostereometry (RSA) of implants Acta Orthop 2005; 76(4): 563-72. Veldkamp W J H, Becht A, Bouwman R W, Boer den A, Crompvoets-Jeukens C R L, Geertse T, Geleijns J, Holscher H, Hummel W A, Molen van der A J, Nievelstein R J, Stoop P, Schilham A, Streekstra G J, Velde van der E, Zweers D. Diagnostische referentieniveaus in Nederland NCS platform ‘stralingsbescherming in het zieknhuis’ Nederlandse Commissie voor Stralingsdosimetrie; 2012. https: //doi.org/10.25030/ncs-021

396

Acta Orthopaedica 2020; 91 (4): 396–400

No association between waiting time to surgery and mortality for healthier patients with hip fracture: a nationwide Swedish cohort of 59,675 patients Katarina GREVE 1,2, Karin MODIG 3, Mats TALBÄCK 3, Erzsébet BARTHA 1,2, and Margareta HEDSTRöM 1,4 1 Institution

of Clinical Sciences, Intervention and Technology (CLINTEC) Karolinska Institutet, Stockholm; 2 Perioperative Medicine and Intensive Care, Karolinska University Hospital Huddinge, Stockholm; 3 Institute of Environmental Medicine, Unit of Epidemiology, Karolinska Institutet, Stockholm; 4 Department of Orthopaedics, Karolinska University Hospital Huddinge, Stockholm, Sweden Correspondence: katarina.greve@sll.se Submitted 2019-09-27. Accepted 2020-03-16.

Background and purpose — Waiting time to surgery for patients with hip fractures and its potential association with mortality has been frequently studied with the hypothesis that longer waiting time is associated with adverse outcomes. However, despite numerous studies, there is no consensus regarding which time frames are appropriate, and whether some patients are more vulnerable to waiting than others. We explored the association between waiting time to surgery and short-term mortality and whether sex, age, surgical method, and comorbidity (ASA) modified this association. Patients and methods — This is a nationwide cohort study of 59,675 patients undergoing hip fracture surgery between January 1, 2013 and December 31, 2017 with a 4-month follow-up of mortality. Data were extracted from the Swedish Registry for Hip Fracture Patients and Treatment (RIKSHÖFT) and mortality was obtained from Statistics Sweden. Results — Unadjusted analyses revealed an association between waiting more than 24 hours for surgery and increased mortality, primarily for women. However, when stratifying for ASA grade, an association persisted only among patients with ASA 3 and 4. Furthermore, the absolute differences in mortality risk between those waiting less or longer than 24 hours were small. Age, fracture type, and surgical method did not modify the association between waiting time and mortality. Interpretation — This study suggests that there may be a need for new guidelines, which take into account the heterogeneity of the patient population.

Waiting time to surgery for patients with hip fractures has been studied with the hypothesis that longer waiting time is associated with adverse outcomes for those patients (Ryan et al. 2015, Morrisey et al. 2017, Hongisto et al. 2019). The underlying mechanism as to why prolonged waiting time to surgery would be detrimental is the longer immobilization with a following catabolism (Hedström et al. 2006) as well as the subsequent increased risk of complications. However, there is no consensus regarding what time frames are optimal, and what constitutes a “longer” waiting time varies widely in different studies (Lewis and Wadell 2016). In Sweden, the latest national guidelines prescribe that all patients should receive surgery within 24 hours (National Board of Health and Welfare 2003). Other countries have similar guidelines: the British National Clinical Guideline Centre (NICE) recommends surgery the same day or the day after hospital admission (NICE 2011). The American Academy of Orthopaedic Surgeons recommends surgery within 48 hours of hospital admission (AAOS 2014). One way to attempt to decrease waiting time to surgery is to institute “fast track care” for patients with hip fracture, often consisting of attempts at early recognition of the hip fracture and thereafter expedient admission to the hospital ward, sometimes bypassing the emergency room entirely (Larsson et al. 2016, Pollmann et al. 2019) It is not known how, and if, waiting longer than 24 hours for surgery was associated with increased mortality compared with waiting less than 24 hours for surgery for patients with hip fractures in Sweden in recent years. It is further possible that the inconsistent results in previous studies on the risks of adverse outcomes due to prolonged waiting time to surgery may be due to different population characteristics of the study

Acta Orthopaedica 2020; 91 (4): 396–400

Patients with non-pathological hip fractures registered in Rikshöft 2013–2017, aged ≥65 years, first appearance in database during the study period only n = 62,269 Excluded (n = 2,594): – missing data regarding date of surgery or surgical method, and patients treated conservatively, 213 – operated within 2 hours, 152 – operated after 7 days, 1,350 – missing data regarding ASA score, 824 – patients with ASA score ≥5, 55 Final study population n = 59,675

Figure 1. Flowchart of the study population.

subjects in different studies. Some patient groups may be more vulnerable to waiting than others, which calls for studies looking at subgroups separately. We explored the association between waiting time to surgery and the 4-month mortality risk in patients with a hip fracture in Sweden between 2013 and 2017, and whether sex, surgical method, age, and comorbidity modified this association.

Patients and methods This is a nationwide cohort study of patients operated for a hip fracture between January 1, 2013 and December 31, 2017. Patient data were extracted from the Swedish Registry for Hip Fracture Patients and Treatment (RIKSHÖFT), a register with estimated coverage of 80–86% for these years (National Board of Health and Welfare 2014, National Board of Health and Welfare 2018). Exclusion criteria were age < 65 years, pathological fractures (i.e., caused by malignancies, bone cysts, or Paget’s disease), waiting time less than 2 hours (assumed as erroneous reporting) or more than 7 days, ASA score ≥ 5. If an individual was registered more than once during the study period, the first fracture only was considered (Figure 1). Variables The exposure, waiting time to surgery, was measured as the time in hours that elapsed between arrival to hospital and start of surgery as registered in RIKSHÖFT. The outcome was time to death up to 4 months from the date of surgery. The date of death was obtained from Statistics Sweden (Ludvigsson et al. 2016). Comorbidity was measured through ASA physical status classification system (Dripps 1963). In our material, ASA classification was assessed preoperatively by local anesthesiologists as part of standard preoperative practice, and registered in RIKSHÖFT. Type of fracture was registered in RIKSHÖFT and adjusted for in the analysis in 2 categories, cervical (consisting of nondisplaced, displaced, and basicervical femoral neck fractures)

397

and non-cervical (2-part, multiple-part intertrochanteric and subtrochanteric fractures). Type of surgery was dichotomized into 2 groups: surgical method 1, which included procedures considered more invasive (intramedullary nail, hemiarthroplasty, or total hip replacement); and surgical method 2, less invasive procedures (screw, pin or nail, 2 screws, pins or nails, 3 or more screws, pins or nails, screw, pin or nail with side plate or other). Living independently before fracture was defined as patients residing in their own homes, with or without assistance from family and/or home care aids. Statistics First, descriptive statistics were produced for the study population. Associations between waiting time and mortality by different fracture types, surgical methods, and ASA score were plotted. Patients with ASA 1 and 2 were compounded into 1 group, and patients with ASA 3 and 4 into another. Next, the proportion of patients who died within 4 months was plotted by hours of waiting time, up to 72 hours. Finally, Cox proportional hazards regression models, using age as the underlying time scale, were used to assess the association between waiting more or less than 24 hours on the time to death, up to 4 months. Crude and adjusted Cox models were performed. Statistical analyses were conducted using STATA version 14.2 (Stata Corp LLC, College Station, TX, USA). Ethics, funding, data sharing plan, potential conflicts of interest The study was approved by the regional Ethics Committee of Stockholm Dnr 2017/1088-31 and 2018/84-32. The study was funded by the Kamprad Foundation for Entrepreneurship, Research and Charity, reference number 20190135, as well as by grants provided by Region Stockholm (ALF project). This study was based on sensitive individual-level data protected by the Swedish personal data act. Data can therefore only be shared after ethical approval and the consent of the principal investigator. The authors declare no conflicts of interest

Results Patient characteristics stratified for those undergoing surgery within and after 24 hours, as well as total patient characteristics, are presented in Table 1. 59,675 patients were operated on for a hip fracture and included in the study, of which 68% were women. The mean age was 83 (SD 8) years and the median waiting time to surgery was 20 hours. Overall 30-day mortality was 8%, and 4-month mortality was 16%. A majority of the patients, 70%, underwent surgery within 24 hours. 51% of the patients had a femoral neck fracture. There were no statistically significant differences between the groups who underwent surgery within or after 24 hours with respect to age, sex, fracture type, or surgical method. A larger fraction of the

398

Acta Orthopaedica 2020; 91 (4): 396–400

Table 1. Descriptive statistics of study population presented for waiting time to hip fracture surgery less or more than 24 hours (first fractures, 2013–2017). Values are number (%) unless otherwise specified Factor

≤ 24 hours

> 24 hours

Total

No. of patients Women Men Mean age (SD) ASA 1 ASA 2 ASA 3 ASA 4 Cervical fractures Non-cervical fractures Surgical method 1 a Surgical method 2 b Time until surgery c 30-day survival d 4-month survival d Living independently before hip fracture On anticoagulants e

41,569 (70) 28,615 (71) 12,954 (67) 83 (8) 1,845 (76) 15,278 (73) 21,798 (68) 2,648 (58) 20,544 (67) 21,025 (72) 25,501 (69) 16,068 (71) 16 (10–20) 93 (92–93) 85 (85–85)

18,106 (30) 11,837 (29) 6,269 (33) 83 (8) 571 (24) 5,539 (27) 10,115 (32) 1,881 (42) 10,087 (33) 8,019 (28) 11,629 (31) 6,477 (29) 32 (27–44) 91 (91–92) 82 (81–82)

59,675 40,452 19,223 83 (8) 2,416 20,817 31,913 4,529 30,631 29,044 37,130 22,545 20 (13–26) 92 (92–93) 84 (83–84)

29,468 (70) 4,687 (57)

12,923 (30) 3,543 (43)

42,391 8,230

a b

Intramedullary nail, hemiarthroplasty; total hip replacement. Two screws, pins, or nails; screw, pin, or nail with side plate; three or more screws, pins, or nails; others. c Median (25th and 75th percentiles) tme in hours. Calculated as time in minutes between arrival at hospital and surgery divided by 60 minutes. d Percentage survival with (95% confidence interval) e Information concerning anticoagulant use on admission (yes/no) was available for 46,311 patients, i.e., 78% of the study population.

Table 2. Adjusted Hazard ratios (95% CI) for the association between waiting more than 24 hours compared with surgery within 24 hours and 4-month mortality, stratified by sex HR adjusted for

Hazard ratios (95% CI) Men Women

Age Age and ASA Age, ASA, and type of fracture Age, ASA, type of fracture, and type of surgery

1.16 (1.08–1.24) 1.06 (0.99–1.13) 1.06 (0.99–1.13)

1.27 (1.20–1.34) 1.15 (1.09–1.22) 1.16 (1.09–1.22)

1.06 (1.00–1.14)

1.16 (1.09–1.22)

Table 3. Adjusted Hazard ratios (aHR) (95% CI) for the association between waiting more than 24 hours compared with surgery within 24 hours and 4-month mortality, stratified by ASA, all patients aHR (CI) a

aHR (CI) b

aHR (CI) c

1.17 (0.72–1.89) 1.05 (0.94–1.17) 1.13 (1.07–1.19) 1.17 (1.06–1.29)

1.27 (0.78–2.08) 1.05 (0.94–1.17) 1.13 (1.07–1.19) 1.17 (1.06–1.29)

1.28 (0.78–2.10) 1.05 (0.94–1.17) 1.13 (1.07–1.19) 1.17 (1.07–1.29)

ASA 1 ASA 2 ASA 3 ASA 4

a Adjusted b Adjusted c Adjusted

for age. for age and type of fracture. for age, type of fracture, and type of surgery.

Figure 2. Probability of death within 4 months by waiting time, stratified for ASA score and sex. The size of the dots is relative to the number of patients operated on at each point in time.

The association between waiting time and death was nonlinear. Men had higher 4-month mortality than women, and patients with ASA 3–4 had higher mortality than patients with ASA 1–2. However, the association between waiting time and mortality was different for the 2 ASA categories and for men vs. women. While the mortality was the same regardless of waiting time for ASA 1–2, there was an initial decline in mortality followed by an increase with longer waiting time for the patients with ASA 3–4, especially among women. Fully adjusted regression analyses confirmed the stronger association between waiting time to surgery and mortality among women compared with men (Table 2), and that the increased mortality remained only for patients with ASA 3 and 4 when stratifying by ASA (HR 1.1, 95% CI 1.1–1.2) and (HR 1.2, CI 1.1–1.3) (Table 3). Type of fracture and surgical method did not modify the association. When additionally stratifying the analyses for patients younger and older than 85 years we found that the association with waiting more than 24 hours for surgery remained only among the ASA 3 and 4 patients, regardless of age (Table 4).

healthier patients (ASA 1 and 2) underwent surgery within 24 hours, compared with the sicker patients (ASA 3 and 4). Overall, a slightly higher fraction of patients were still alive after 4 months in the group that waited less than 24 hours for surgery, 85% compared with 82%. 4-month mortality by waiting time stratified by ASA score and sex is presented in Figure 2.

Sensitivity analysis For a subset of the patients, 4,850 individuals, there was additional information regarding the time of the fracture (as opposed to arrival time at the hospital). For these patients we re-ran the analyses to see if the association between waiting time and mortality would change. The results were similar to those using time of arrival at hospital.

Acta Orthopaedica 2020; 91 (4): 396–400

Table 4. Hazard ratios (95% CI) for the association between waiting more than 24 hours compared with surgery within 24 hours and 4-month mortality, stratified by ASA, patients aged 65–85, and patients aged > 85 Group Age 65–85 Age > 85

HR (CI) for waiting time > 24 h ASA 1–2 ASA 3–4 ASA 1–2 ASA 3–4

1.02 (0.85–1.22) 1.11 (1.03–1.19) 1.09 (0.96–1.25) 1.22 (1.15–1.30)

Waiting time < 24 h is reference value 1.00

30-day mortality is an often-used endpoint in other studies. To facilitate comparisons, 30-day mortality was plotted by waiting time and stratified by ASA score and sex, with results similar to when plotting 4-month mortality (Figure 3, see Supplementary data).

Discussion Waiting for hip fracture surgery of more than 24 hours was associated with higher risk of death within 4 months but only for patients with ASA score 3 and 4, and primarily for women. Overall the associations between waiting time to surgery and mortality were rather weak, an absolute difference of a couple of percentage points, and OR equal to 1.6 to 1.2. Fracture type and surgical method did not affect the association between waiting time and mortality. Women in the ASA 3–4 category who underwent early surgery (within 4–10 hours) demonstrated an increased 4-month mortality compared with those undergoing surgery slightly later. This could conceivably be attributed to 2 factors: these women could be the most vulnerable patients and either could have benefited from more careful preoperative optimization, or they were selected for early surgery based on the presumed benefit of expedient surgical intervention. Notably, if the first hypothesis is true, the 24-hour “rule” may lead to inappropriately rushed perioperative management of the sickest women with hip fracture. There are no comprehensive data regarding reasons for “delay” of surgery in RIKSHÖFT. It is possible that the risk of mortality can differ depending on whether the delay was a consequence of medically related vs. administrative reasons. A previous study from Sweden, however, suggests that waiting time > 36 hours to surgery was detrimental to patients, at least for functional outcome, regardless of reason (Al-Ani et al. 2008). In the same study, administrative reasons were cited in two-thirds of the cases where the patients had waited more than 24 hours for surgery. One potential medical reason for delaying surgery is the need for reversal of anticoagulant medication. However, the proportion of the group with waiting time to surgery within 24 hours who were on anticoagulant

399

medication on hospital admission was 57%, compared with 43% in the group with longer waiting time, which makes it unlikely that treatment with anticoagulants was an important reason for delayed surgery in our material. The chosen outcome of this study, mortality, does not capture all consequences of the 24-hour rule. There may be benefits such as lower numbers of complications, but also drawbacks such as rescheduled or postponed surgery for patients not suffering from hip fracture. Considering this, other patientoriented outcome measures need to be studied. Comparison with previous studies What sets our material most apart from many other comparable studies is that waiting time to surgery overall was very short for the entire cohort, which means that the hazards of prolonged waiting time to surgery could be underestimated if compared directly with cohorts with longer waiting times. The heterogeneity of outcome criteria (in-hospital, 30-day, or 4-month mortality) of measures of comorbidities and of clinical settings makes comparisons difficult between our study and previous studies. Overall, our finding of a weak association between increased risk of death and longer waiting time is in line with previous reports (Pincus et al. 2017, Hongisto et al. 2019, Öztürk et al. 2019). However, there are studies that fail to confirm an association between waiting time to surgery and mortality (Majumdar et al. 2006). Consistent with several previous studies (Endo et al. 2005, Uzoigwe et al. 2013), men in our study had higher overall mortality rate than women. However, there was a stronger association for women in ASA 3–4 (compared with men) between waiting time to surgery and increased mortality, and this has not been previously reported to our knowledge. Contrary to our results, a Danish study (Öztürk et al. 2019), conducted in a clinical setting similar to ours, found a weak association between waiting time and 30-day mortality in “healthier” patients. This could be attributed to using different measures of comorbidity. The Öztürk study used the Charlson comorbidity index (CCI), while we used the ASA classification. Strengths and limitations This is, to our knowledge, the first study to explore the association between waiting time to hip fracture surgery and subsequent mortality in a nationwide study from a data source with high coverage and validity. In the subgroup analyses, potential confounders were considered and adjusted for whenever possible. In observational studies, it is not possible to conclude causal relationships between variables. It is possible that factors we consider confounders could really be mediators; this is a limitation that is difficult for us to completely avoid. On the other hand, we know that waiting time to surgery is often affected by system-related factors, which makes it less likely that waiting time to surgery should be affected by for example ASA score or fracture type.

400

Conclusion For patients with an ASA score of 3 or 4, there was a small increase in the risk of 4-month mortality for those who waited at least 24 hours for surgery. The association was stronger for women than for men, and for patients > 85. Fracture type and type of surgery had no impact on the association. Our findings give no support for the hypothesis that surgery within 24 hours reduces mortality risks of hip fracture patients with an ASA score of 1 or 2. Given these differences between men and women, and for patients with different ASA scores, our results suggest that a strict time limit applying to all patients may not be the best strategy. Supplementary data Figure 3 is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2020. 1754645

The study was conceived by MH, KM, EB, and KG. MT and KM performed the analyses. KG, MH, EB, and KM wrote the initial draft. All the authors contributed to the interpretation of the data and to revision of the manuscript. The authors would like to acknowledge and appreciate the help given to us from RIKSHÖFT. Acta thanks Aare Märtson and Jan A N Verhaar for help with peer review of this study.

AAOS. Management of hip fractures in the elderly, evidence based clinical practice guideline. Rosemont: American Academy of Orthopaedic Surgeons; 2014. https://www.aaos.org/cc_files/aaosorg/research/guidelines/ hipfxguideline.pdf Al-Ani A N, Samuelsson B, Tidermark J, Norling A, Ekström W, Cederholm T, Hedström M. Early operation on patients with a hip fracture improved the ability to return to independent living: a prospective study on 850 patients. J Bone Joint Surg Am 2008; 90(7): 1436-42. Dripps R D. New classification of physical status. Anesthesiol 1963; 24: 111. Endo Y, Aharonoff G B, Zuckerman J D, Egol K A, Koval K J. Gender differences in patients with hip fracture: a greater risk of morbidity and mortality in men. J Orthop Trauma 2005; 19(1): 29-35. Hedström M, Ljungqvist O, Cederholm T. Metabolism and catabolism in hip fracture patients: nutritional and anabolic intervention—a review. Acta Orthop 2006; 77(5): 741-7. Hongisto M T, Nuotio M S, Luukkaala T, Väistö O, Pihlajamäki H K. Delay to

Acta Orthopaedica 2020; 91 (4): 396–400

surgery of less than 12 hours is associated with improved short- and longterm survival in moderate to high-risk hip fracture patients. Geriatr Orthop Surg Rehabil 2019; 10: 1-7. Larsson G, Strömberg U, Rogmark C, Nilsdotter A. Prehospital fast-track care for patients with hip fracture: impact on time to surgery, hospital stay, post-operative complications and mortality a randomized, controlled trial. Injury 2016; 47(4): 881-6. Lewis P M, Waddell J P. When is the ideal time to operate on a patient with a fracture of the hip? Bone Joint J 2016; 98-B: 1573-81. Ludvigsson J F, Almqvist C, Bonamy A K, Ljung R, Michaëlsson K, Neovius M, Stephansson O, Ye W. Registers of the Swedish total population and their use in medical research. Eur J Epidemiol 2016; 31(2): 125-36. Majumdar S R, Beaupre L A, Johnston D W, Dick D A, Cinats J G, Jiang H X. Lack of association between mortality and timing of surgical fixation in elderly patients with hip fracture: results of a retrospective populationbased cohort study. Med Care 2006; 44(6): 552-9. Morrisey N, Iliopoulus E, Osmani A W, Newman K. Neck of femur fractures in the elderly: does every hour to surgery count? Injury 2017; 48(6): 1155-8. National Board of Health and Welfare, Socialstyrelsen [internet]. The National Patient Register [in English]. Available from: http://www.socialstyrelsen. se/register/halsodataregister/patientregistret/inenglish National Board of Health and Welfare, Socialstyrelsens riktlinjer för vård och behandling av höftfraktur [in Swedish]. Stockholm: Socialstyrelsen; 2003. National Board of Health and Welfare, Socialstyrelsen [internet]. Rapporteringen till nationella kvalitetsregister och hälsodataregistren. Available from: https://www.socialstyrelsen.se/globalassets/sharepoint-dokument/ artikelkatalog/statistik/2014-12-7.pdf [in Swedish]. Stockholm: Socialstyrelsen; 2014. National Board of Health and Welfare, Socialstyrelsen [internet]. Täckningsgrader 2018. Article number: 2018-12-55. Available from: www.socialstyrelsen.se/publikationer2018/2018-12-55 [in Swedish]. Stockholm: Socialstyrelsen; 2018. NICE. The management of hip fracture in adults. London: National Clinical Guideline Centre; 2011. https://www.nice.org.uk/guidance/cg124/evidence/full-guideline-183081997 Öztürk B, Johnsen S P, Röck N D, Pedersen L, Pedersen A B. Impact of comorbidity on the association between surgery delay and mortality in hip fracture patients: a Danish nationwide cohort study. Injury 2019; 50(2): 424-43. Pincus D, Ravi B, Wasserstein D, Huang A, Paterson J M, Nathens A B, Kreder H J, Jenkinson R J, Wodchis W P. association between wait time and 30-day mortality in adults undergoing hip fracture surgery. JAMA 2017; 318(20): 1994-2003. Pollmann C T, Røtterud J H, Gjertsen J E, Dahl F A, Lenvik O, Årøen A. Fast track hip fracture care and mortality: an observational study of 2230 patients. BMC Musculoskelet Disord 2019; 20: 248. RIKSHÖFT. http://rikshoft.se/ [in Swedish]. Ryan D J, Yoshihara H, Yoneoka D, Egol K A, Zuckerman JD. Delay in hip fracture surgery: an analysis of patient-specific and hospital-specific risk factors. J Orthop Trauma 2015; 29(8): 343-8. Uzoigwe C E, Burnand H G F, Cheesman C L, Aghedo D O, Faizi M, Middleton R G. Early and ultra-early surgery in hip fracture patients improves survival. Injury 2013; 44(6): 726-9.

Acta Orthopaedica 2020; 91 (4): 401–407

401

Is there a reduction in risk of revision when 36-mm heads instead of 32 mm are used in total hip arthroplasty for patients with proximal femur fractures? A matched analysis of 5,030 patients with a median of 2.5 years’ follow-up between 2006 and 2016 in the Nordic Arthroplasty Register Association Georgios TSIKANDYLAKIS 1–3, Johan N KÄRRHOLM 1–3, Geir HALLAN 4–5, Ove FURNES 4–5, Antti ESKELINEN 6–7, Keijo MÄKELÄ 7,8, Alma B PEDERSEN 9–10, Søren OVERGAARD 10–12, and Maziar MOHADDES 1–3 1 Department

of Orthopaedics, Institute of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg; 2 The Swedish Hip Arthroplasty Register, Gothenburg, Sweden; 3 Region Västra Götaland, Sahlgrenska University Hospital, Dept of Orthopaedics, Gothenburg, Sweden; 4 The Norwegian Arthroplasty Register, Department of Orthopedic Surgery, Haukeland University Hospital, Norway; 5 Department of Clinical Medicine, University of Bergen, Norway; 6 Coxa Hospital of Joint Replacement, Tampere Finland; 7 The Finnish Arthroplasty Register, Finland; 8 Department of Orthopaedics and Traumatology, Turku University Hospital, Finland; 9 Department of Clinical Epidemiology, Aarhus University Hospital, Denmark; 10 The Danish Hip Arthroplasty Register, Denmark; 11 Department of Orthopaedic Surgery and Traumatology, Odense University Hospital, Denmark; 12 Institute of Clinical Research, University of Southern Denmark Correspondence: georgios.tsikandylakis@vgregion.se Submitted 2019-12-08. Accepted 2020-03-22.

Background and purpose — 32-mm heads are widely used in total hip arthroplasty (THA) in Scandinavia, while the proportion of 36-mm heads is increasing as they are expected to increase THA stability. We investigated whether the use of 36-mm heads in THA after proximal femur fracture (PFF) is associated with a lower risk of revision compared with 32-mm heads. Patients and methods — We included 5,030 patients operated with THA due to PFF with 32- or 36-mm heads from the Nordic Arthroplasty Register Association database. Each patient with a 36-mm head was matched with a patient with a 32-mm head, using propensity score. The patients were operated between 2006 and 2016, with a metal or ceramic head on a polyethylene bearing. Cox proportional hazards models were fitted to estimate the unadjusted and adjusted hazard ratio (HR) with 95% confidence intervals (CI) for revision for any reason and revision due to dislocation for 36-mm heads compared with 32-mm heads. Results — 36-mm heads had an HR of 0.9 (CI 0.7–1.2) for revision for any reason and 0.8 (CI 0.5–1.3) for revision due to dislocation compared with 32-mm heads at a median follow-up of 2.5 years (interquartile range 1–4.4). Interpretation — We were not able to demonstrate any clinically relevant reduction of the risk of THA revision for any reason or due to dislocation when 36-mm heads were used versus 32-mm. Residual confounding due to lack of data on patient comorbidities and body mass index could bias our results.

During the past years total hip arthroplasty (THA) has become the preferred treatment option for displaced femoral neck fractures in even younger (55–64 years) patients (Rogmark et al. 2017). Previous studies have shown an increased risk of revision, especially due to dislocation, in patients receiving THA after proximal femur fracture (PFF) compared with patients operated due to primary osteoarthritis (OA) (Conroy et al. 2008, Hailer et al. 2012). The risk of THA dislocation in fracture patients varies widely from as low as 5% (Tabori-Jensen et al. 2019), especially when dual mobility cups (DMCs) are used, up to 6–18% (Burgers et al. 2012, Johansson 2014, Noticewala et al. 2018) with conventional cups. The risk of THA revision due to dislocation has been reported as even lower, ranging from 0.5 to 0.7% in national register studies (Conroy et al. 2008, Hailer et al. 2012), as not all unstable THAs are revised. According to the above-mentioned studies, increased age, male sex, the use of a posterior approach, and smaller head sizes are associated with increased risk of revision due to dislocation. To counteract the risk of dislocation, bigger head sizes have been used as they increase the impingement-free range of motion (Burroughs et al. 2005, Tsuda et al. 2016) and jumping distance of THA (Sariali et al. 2009). During the past years, the use of larger heads in THA has increased with 28-mm continuously declining and 32- and 36-mm increasing (Tsikandylakis et al. 2018b). However, register studies performed on patients with displaced femoral neck fracture (Jameson et al. 2012, Cebatorius et al. 2015) have not demonstrated any superiority of larger heads over smaller ones regarding risk of revision, especially due to dislocation. This effect has only been demonstrated in studies

402

Acta Orthopaedica 2020; 91 (4): 401–407

Annual distribution of head sizes (%) 100

THAs registered in NARA 1995–2016 n = 745,808 Excluded (n = 733,332): – hip diagnosis other than fracture, 638,253 – missing dta on hip diagnosis, 41,685 – head size < 32 or > 36 mm, 40,775 – missing data on head size, 5,520 – other bearings than MoP and CoP, 1,685 – conventional polyethylene cup, 4,672 – missing data on bearing material, 333 – THAs 1995–2005, 44 – cases of second operated hip, 120 – missing data on type of fixation, 117 – missing data on surgical approach, 128

36-mm heads 32-mm heads 2004

2008

2012

Year of surgery

2016

Figure 1. Use of 32- and 36-mm heads in MoXLPE and CoXLPE THA after proximal femur fracture in NARA database.

performed on a case mix of hip diagnoses that have reported an increased risk of revision due to dislocation when 28-mm or smaller heads are used compared with 32-mm or larger heads (Hailer et al. 2012, Kostensalo et al. 2013). Most of the above-mentioned register studies have used 28-mm heads as reference, which are rarely used nowadays (Tsikandylakis et al. 2018b). Patients receiving THA after PFF have a higher risk for revision than patients with OA and should preferably be studied separately, setting 32 mm as contemporary standard of reference. We therefore investigated if increasing head size from 32 to 36 mm is associated with a decreased risk of revision, especially due to dislocation, in patients with PFF in the Nordic Arthroplasty Register Association (NARA) database. We hypothesized that the risk is lower when 36-mm heads are used.

Patients and methods This study was designed as a propensity-matched cohort study within NARA, a collaboration among the national arthroplasty registries of Denmark, Finland, Norway, and Sweden (Havelin et al. 2011). We included patients operated with THA due to PFF, registered in the NARA database between January 1, 1995 and December 31, 2016. Patients operated with head sizes other than 32 or 36 mm, DMCs, and hip resurfacing were excluded. As metal on cross-linked polyethylene and ceramic on crosslinked polyethylene are the most common bearing types used in modern THA (Tsikandylakis et al. 2018b) we excluded all other bearing combinations. As 36-mm heads were not used in the Nordic countries until 2006 (Figure 1), we excluded all THAs performed before 2006. In patients with bilateral THA, the 2nd operated hip was excluded to fulfil the assumption of independent observations (Ranstam et al. 2011). The type of implant fixation included 4 categories: cemented, uncemented, hybrid, and reverse hybrid. The type of surgical approach is registered in NARA as either posterior or non-

THAs eligible for propensity score matching n = 12,476 Year of surgery: 2006–2016 Indication for surgery: hip fracture Bearing material: MoXLPE or CoXLPE Head size: 32 or 36 mm First operated side only in bilateral THAs No missing data on variables of interest Head size 32 mm, 8,957 (72%): – unmatched cases, 6,442 – matched cases, 2,515 (50%)

Head size 36 mm, 3,519 (28%): – unmatched cases, 1,004 – matched cases, 2,515 (50%)

Figure 2. Flowchart of the selection and matching process.

posterior because 1 of the national registries does not report further details on non-posterior approaches. Follow-up time, age, and year of surgery were handled as quantitative variables without grouping. Operations with any missing data on the above-mentioned variables were excluded (Figure 2). After exclusions, 12,476 patients remained, of whom 72% had received a 32-mm and 28% a 36-mm head. There were differences between the groups as patients operated with 36-mm heads were younger with a higher proportion of males, operated more recently with predominantly a posterior approach and uncemented implant fixation (Table 1). These imbalances may confound the risk of THA revision, especially due to dislocation (Hailer et al. 2012, Jameson et al. 2012, Cebatorius et al. 2015). To reduce bias due to confounding, patients from the 36-mm group were matched to patients in the 32-mm group with a 1:1 ratio, using propensity score (PS) (Kuss et al. 2016) based on patient age, sex, year of surgery, type of implant fixation, bearing, and surgical approach. We were able to match 2,515 patients with 36-mm heads to 2,515 patients with 32-mm heads using the PS (Table 2). In the matched sample, the differences in sex, age, year of surgery, and type of surgical approach decreased considerably. We evaluated the balance of the covariates between the 2 head size groups before and after matching using absolute standardized differences in means (ASDM). The highest ASDM after matching was 0.1 (Table 2, Figure 3, see Supplementary data), below the threshold of 0.15 that indicates significant imbalance between groups (Austin 2011). Mortality rates were high in both head size groups (18–20%) but did not differ between them either before (Table 1) or after (Table 2) PS matching. In the matched sample median followup was 2.5 years (interquartile range 1–4.4).

Acta Orthopaedica 2020; 91 (4): 401–407

403

Table 1. Descriptive statistics of study population before propensity score matching. Values number (%) unless otherwise specified

32-mm head (n = 8,957)

36-mm head (n = 3,519)

Follow-up, years b 2.4 (1.0–4.4) 2.3 (1.0–4.0) Mortality 1,702 (19) 621 (18) Age (standard deviation) 73 (10) 70 (11) Year of surgery b 2013 2014 (2011–2015) (2012–2015) Female sex 6,266 (70) 1,812 (52) Cemented THA 6,276 (70) 813 (23) Cementless THA 1,219 (14) 1,729 (50) Hybrid THA 430 (5) 885 (25) Reverse hybrid 1,032 (12) 92 (3) MoXLPE c 7,954 (89) 3,083 (88) CoXLPE d 1,003 (11) 436 (12) Posterior approach 3,912 (44) 2,599 (74)

ASDM a 0.1 0.04 0.2 0.1 0.4 0.6 0.7 0.5 0.6 0.04 0.04 0.7

a Absolute standardized difference in means. b Median (interquartile range). c Metal head on cross-linked polyethylene. d Ceramic head on cross-linked polyethylene.

Table 2. Descriptive statistics of study population after propensity score matching. Values number (%) unless otherwise specified

32-mm head (n = 2,515)

Follow-up, years b Mortality Age (standard deviation) Year of surgery b Female sex Cemented THA Cementless THA Hybrid THA Reverse hybrid MoXLPE c CoXLPE d Posterior approach

2.4 (0.9–4.4) 2.6 (1.1–4.3) 477 (19) 507 (20) 70 (11) 71 (11) 2013 2013 (2011–2015) (2011–2015) 1,570 (62) 1,453 (58) 823 (33) 813 (32) 1,148 (46) 1,068 (43) 428 (17) 542 (22) 116 (5) 92 (4) 2,108 (84) 2,152 (86) 407 (16) 363 (14) 1724 (69) 1705 (68)

a–d See

ASDM a 0.03 0.03 0.07 0.05 0.09 0.02 0.06 0.1 0.06 0.06 0.05 0.02

Table 1.

Approach

Patient age

The primary outcome of our study was the 1st THA revision for any reason and the secondary outcome was the 1st revision due to dislocation. Revision was defined as the exchange or removal of any of the hip prosthetic components. Follow-up time was defined as the time between primary surgery until 1st revision, death, emigration, or December 31, 2016, whichever came first.

36-mm head (n = 2,515)

Head material

Head size Adjusted Unadjusted Revision

Fixation

Sex Year of surgery

Statistics Descriptive statistics were performed in SPSS, Version 25 (IBM Corp, Armonk, NY, USA). Mean and standard deviation was used to describe age. Follow-up time was described with median and interquartile range (IQR). PS matching and survival analysis were performed in R software, Version 3.4.4 (R Foundation for Statistical Computing, Vienna, Austria). PS was calculated using the function “matchit” setting head size as the dependent variable and age, sex, year of surgery, type of fixation, type of bearing, and surgical approach as exploratory variables. The calliper was set to 0.15. Patients from the 36-mm group were matched to the 32-mm group using the nearest neighbor method at a 1:1 ratio. Unmatched patients were discarded from both groups. Kaplan–Meier survival curves for the whole observation time were drawn for each head size. The follow-up period was censored at 7 years because after that time point the number of patients at risk in the 36-mm group dropped below 100. After the 7th year of follow-up only 1 revision occurred in the 32-mm group and none in the 36-mm. The Mantel–Cox log rank test was used to compare the survival curves. Univariable Cox proportional hazards models were fitted to calculate the hazard ratio (HR) for head size with 95% confidence intervals (CI) for the period 0–7 years. Multivariable Cox proportional hazards models were also fitted adjusting for patient age, sex, year

Figure 4. A directed acyclic graph (DAG) was constructed under the following assumptions: 1) THA ‘revision’ is dependent on ‘head size’, ‘patient age’, ‘sex’, ‘year of surgery’, surgical ‘approach’, and type of THA ‘fixation’. Choice of ‘head material’ is not expected to affect ‘revision’ due to the short follow-up of the study. 2) Choice of ‘head size’ is dependent on ‘approach’, ‘year of surgery’, ‘sex’, and ‘patient age’ as surgeons operating on older patients through a posterior approach have presumably chosen a larger head in order to, hopefully, reduce the risk of dislocation. Male patients, operated more recently, have probably received a larger head due to their larger acetabulum and because the use of larger heads has become more popular with time. 3) ‘Fixation’ is dependent on ‘year of surgery’ and ‘age’ because patients operated more recently have probably received an uncemented THA, due to the popularization of this technique, and older patients have probably received a cemented THA due to their poorer bone quality. 4) ‘Head material’ is dependent on ‘head size’ and ‘patient age’ because surgeons have probably chosen ceramic over metal heads in younger patients and when choosing larger heads due to the presumed lower polyethylene wear. Provided that our assumptions are correct, adjusting for ‘patient age’, ‘sex’, ‘year of surgery’, and ‘approach’ in the multivariable Cox regression model should block all backdoor pathways (for variables available in our database) confounding the effect of ‘head size’ on ‘revision’.

of surgery, and surgical approach. Despite the minimal differences in these 4 covariates between the head size groups, we chose to put them in the models in order to block all presumed backdoor pathways (for variables available in NARA

404

Acta Orthopaedica 2020; 91 (4): 401–407

Table 4. Cox proportional hazards models with endpoint revision for any reason and due to dislocation Outcome Univariable model Multivariable model Head size HR (CI) HR (CI) Revision for any reason 32-mm 1 1 36-mm 0.9 (0.7–1.2) 0.9 (0.7–1.2) Revision due to dislocation 32-mm 1 1 36-mm 0.8 (0.5–1.3) 0.8 (0.5–1.3)

Figure 5. Kaplan–Meier survival function for THA with 32- and 36-mm heads with endpoint revision for any reason (left panel) and revision due to dislocation (right panel). Table 3. Kaplan–Meier survival estimates (%) at 1, 3, and 7 years for 32- and 36-mm heads with endpoint revision for any reason Follow-up 32-mm heads 1-year 3-year 7-year 36-mm heads 1-year 3-year 7-year

Patients Cumulative at risk revisions

Cumulative survival rate (CI)

1,816 1,032 223

88 108 119

95.8 (95.1–96.7) 94.4 (93.4–95.5) 92.8 (91.2–94.4)

1,922 1,099 140

87 102 111

95.9 (95.1–96.7) 95.2 (94.3–96.1) 93.7 (92.2–95.2)

CI = 95% confidence intervals.

database) that could confound the effect of head size on THA revision (Figure 4). 32-mm heads were the reference group. Shoenfeld residuals were used to ascertain the proportional hazards assumption. The level of statistical significance was set at alpha = 0.05. There might be some few patients whose revision has not been registered due to lower completeness for revision THA (80–95%) than for primary THA (95–98%) or patients revised in another country than the country of their primary operation. These patients could not be followed up and were considered unrevised. Ethics, funding, and potential conflicts of interests The study was approved by the Regional Ethical Review Board of Gothenburg on October 26, 2016 (reg. ID 85816). The manuscript was written according to the STROBE (Strengthening the Reporting of Observational studies in Epidemiology) guidelines. No funding specific to this study has been received. 7 co-authors (GT, JNK, GH, OF, KM, ABP, and SO) declare no conflict of interests relevant to this study. 1 co-author (AE) has given paid presentations for and received institutional support from commercial parties (ZimmerBiomet and DePuy Synthes) related indirectly to the subject of this study. 1 co-author (MM) has given paid presentation for

The multivariable model was adjusted for patient age, sex, year of surgery, and type of surgical approach. HR (CI) = Hazard ratio (95% confidence interval)

Table 5. Kaplan–Meier survival estimates (%) at 1, 3, and 7 years for 32- and 36-mm heads with endpoint revision due to dislocation Follow-up 32-mm heads 1-year 3-year 7-year 36-mm heads 1-year 3-year 7-year

Patients Cumulative at risk revisions

Cumulative survival rate (CI)

1,816 1,032 223

33 39 40

98.4 (97.9–98.9) 98.1 (97.6–98.7) 97.8 (97.0–98.7)

1,922 1,099 140

28 31 33

98.7 (98.2–99.2) 98.6 (98.2–99.1) 98.3 (97.6–99.0)

CI = 95% confidence intervals.

commercial parties (Zimmer-Biomet, Link) related indirectly to the subject of this study.

Results Up to 7-year follow-up, 119 (4.7 %) 1st-time revisions for any reason had occurred in the 32-mm group and 111 (4.4%) 1sttime revisions in the 36-mm group. The Kaplan–Meier survival, although slightly higher for 36-mm heads, did not differ statistically significantly between the 2 head sizes (plog-rank = 0.6). The 7-year survival rate was 92.8% (CI 91.2–94.4) for 32-mm and 93.7% (CI 92.2–95.2) for 36-mm heads (Figure 5, Table 3). Both the univariable and the multivariable Cox regression models (adjusting for age, sex, year of surgery, and surgical approach) showed HR estimates favoring 36-mm heads during the first 7 years after THA, but with CIs extending on both sides of 1 (Table 4). 73 1st-time revisions due to dislocation had occurred during the 1st 7 years of follow-up, of which 61 were done during the 1st year after surgery. 6 were done during the 2nd and 3 during the 3rd year after THA. There were 40 (1.6 %) revisions in the 32-mm and 33 (1.3 %) in the 36-mm group. With endpoint

Acta Orthopaedica 2020; 91 (4): 401–407

revision due to dislocation the Kaplan–Meier survival did not differ significantly between the 2 groups (plog-rank = 0.4). The 7-year survival rate for THAs with 32-mm heads was 97.8% (CI 97.0–98.7) and 98.3% (CI 97.6–99.0) for 36-mm heads (Figure 5, Table 5). In the univariable and the multivariable Cox regression model after adjusting for age, sex, year of surgery, and surgical approach, HR estimates were in favor for 36-mm heads, but with CIs including 1 (Table 4).

Discussion In this matched observational study in the NARA database we found that the choice between a 32- or 36-mm head in THA after PFF is not associated with any clinically relevant difference in the risk of revision for either any reason or due to dislocation. The was a trend favoring 36-mm heads, but in absolute numbers this difference corresponded to a decrease in revision rates of only 0.3%. In our opinion, a reduction in revision rates by at least 50% would be clinically relevant, which corresponds to a minimum risk decrease of 2.3% in revisions for any reason and 0.8% in revisions due to dislocations. Besides its small effect size, the difference between 32- and 36-mm heads in our sample is difficult to generalize at a population level due to the lack of statistical significance. Our study could be insufficiently powered to detect smaller risk differences. PS matching at a 1:1 ratio may have caused a considerable loss of statistical power; however, it was the only matching ratio that allowed an acceptable calliper below 0.2, due to the extent of heterogeneity in the unmatched sample. Moreover, our study has inherent limitations due to the lack of randomization. Unmeasured confounding due to factors unknown to us could have skewed the results. For example, the NARA database does not distinguish among femoral neck fracture, trochanteric, or pathologic PFFs. Most of the trochanteric fractures are treated with internal fixation whereas pathologic fractures might receive a THA. Such cases, although probably very few, might be subjected to increased risk of revision due to poor bone quality. BMI and patient comorbidities that increase the risk of THA revision or dislocation (Peters et al. 2020) are not registered in the NARA database. Surgeons may have used the largest available head size when operating on patients with comorbidities, high BMI, poor compliance, or poor bone quality, hoping to reduce the risk of dislocation. In this case, the accumulation of such patients in the 36-mm group may have disfavored THA survival due to bias by indication. We could not adjust for implant positioning either, since radiographs are not available in our database, but we do not expect the head size groups to differ in terms of prosthesis orientation. 36-mm heads can only be used for cups down to a certain diameter to allow for sufficient thickness of the polyethylene. This could be another source of confounding, provided that the risk for revision varies depending on cup size. Implant size is, however, not recorded in the NARA database

405

and could therefore not be accounted for during PS matching. However, analysis stratified for sex, based on the presumption that females in our study received a higher share of small cups, does not support that this factor had any decisive influence (Table 6, see Supplementary data). A non-posterior approach may include different approaches such as direct anterior, anterolateral, lateral, and transtrochanteric that are not specified in the NARA database. However, no major differences in the risk of THA revision have been reported among these surgical approaches (Berry et al. 2005, Mjaaland et al. 2017, Zijlstra et al. 2017). Patients lost to follow-up due to unregistered THA revisions or having their revision in another country were considered unrevised. These patients are not expected to be overrepresented in any of the head size groups and are therefore not considered a major source of bias. We used revision due to dislocation as an estimate of THA dislocation. However, surgeons might be more reluctant to revise an unstable THA with a larger head, which could favor the survival of 36-mm THA. The median follow-up in our patients (2.5 years) was not long enough to capture time-dependent complications related to head size such as polyethylene wear, osteolysis, and subsequent revision due to aseptic loosening or periprosthetic fracture. However, it was long enough to capture revisions due to early complications such as dislocation, which is the leading cause of 32- and 36-mm THA revision after PFF (Jobory et al. 2019). Finally, heterogeneity in the revision risk related to head size among the 4 national registries could affect the precision of the survival estimates in our sample. We therefore performed Cox regression analyses stratified for each country. There were still no major differences between 32- and 36-mm heads except for Finland, where the risk of revision due to dislocation was found to be lower for 36-mm heads (Table 7, see Supplementary data). This seems to be a result of the lower survival of 32 mm THA in the Finnish Register, compared with the other 3 registries (Figure 6, see Supplementary data). This observation should, however, be viewed cautiously since the Finnish Register contributed with significantly fewer 32-mm THAs than 36-mm (Table 8, see Supplementary data). Our results confirm 2 previous reports on patients with femoral neck fracture who received THA and where head size could not be identified as a risk factor for revision. In the study from the National Joint Registry (Jameson et al. 2012) head sizes were grouped as < 28, 28, 30 or 32, and ≥ 36 mm. Using 28 mm as reference no association between head size and risk of THA revision for any reason was found. Cementless fixation, on the other hand, was a risk factor for revision (HR 1.8, CI 1.1–3.1). Neither did the study from the Lithuanian register (Cebatorius et al. 2015) find any difference in risk of revision due to dislocation between 28- and 32-mm heads. The use of a posterior approach, however, was associated with 2.7 times (CI 1–5) higher risk of revision due to dislocation. Due to the nature of our study, we cannot make any unbiased estimation of the effect of surgical approach or type of fixation on the risk of THA revision. The absence of further

406

decrease in the revision risk with increased head size, in studies performed exclusively on patients with femoral neck fractures, could be explained by residual confounding or it could also be that any dislocation-preventing effect of using 36-mm heads over 32-mm is so small that it cannot be detected even in larger register studies like ours. Such an effect, if truly present, could probably only marginally address the increased risk of dislocation. Until now, only DMCs that can accommodate considerably larger heads have shown a decreased risk of revision due to dislocation compared with 32- and 36-mm heads in patients with THA after PFF (Jobory et al. 2019). When a higher risk of dislocation is anticipated, DMCs are probably a better option than 36-mm heads but their higher price needs to be considered. Other register studies, performed on either a case mix of hip diagnoses (Hailer et al. 2012, Kostensalo et al. 2013) or exclusively primary osteoarthritis (Zijlstra et al. 2017, Tsikandylakis et al. 2018a), have reported a decreased revision risk due to dislocation with increasing head size, at least up to 32 mm. Our study cannot be compared with such studies as the diagnosis of femoral neck fracture itself is a risk factor for THA revision. Conclusion Choosing a head size of 36 instead of 32 mm does not seem to be associated with any clinically important decrease in the risk of revision due to any reason and not even due to dislocation after THA in patients operated because of PFF. As THA revision due to dislocation is a rare complication, larger studies with better control of confounders, such as register-based randomized control trials, are needed to make sufficiently powered and unbiased estimations of differences in revision risks between head sizes. Supplementary data Figures 3 and 6, and Tables 6–8 are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674.2020.1752559 GT a–g, JK a–c, e–g, GH a, c, e–g, OF a, c, e–g, AE a, c, e–g, KM a, c, e–g, AP a, c, e–g, SO a, c, e–g, MM a, c, e–g a Substantial contributions to the conception or design of the work. b Substantial contributions to the acquisition and analysis of data for the work. c Substantial contributions to the interpretation of data for the work. d Drafting the work. e Revising the work critically for important intellectual content. f Final approval of the version to be published. g Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The authors would like to thank Emma Nauclér for her valuable input on the choice of appropriate statistical methodology, as well as Nordforsk for its previous financial support of the NARA project. Acta thanks Stephen Ellis Graves and Wierd P Zijlstra for help with peer review of this study.

Acta Orthopaedica 2020; 91 (4): 401–407

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. Berry D J, von Knoch M, Schleck C D, Harmsen W S. Effect of femoral head diameter and operative approach on risk of dislocation after primary total hip arthroplasty. J Bone Joint Surg Am 2005; 87(11): 2456-63 . Burgers P T P W, Van Geene A R, Van den Bekerom M P J, Van Lieshout E M M, Blom B, Aleem I S, Bhandari M, Poolman R W. Total hip arthroplasty versus hemiarthroplasty for displaced femoral neck fractures in the healthy elderly: a meta-analysis and systematic review of randomized trials. Int Orthop 2012; 36(8): 1549-60. Burroughs B R, Hallstrom B, Golladay G J, Hoeffel D, Harris W H. Range of motion and stability in total hip arthroplasty with 28-, 32-, 38-, and 44-mm femoral head sizes. J Arthroplasty 2005; 20(1): 11-19 . Cebatorius A, Robertsson O, Stucinskas J, Smailys A, Leonas L, Tarasevicius S. Choice of approach, but not femoral head size, affects revision rate due to dislocations in THA after femoral neck fracture: results from the Lithuanian Arthroplasty Register. Int Orthop 2015; 39(6): 1073-6. Conroy J L, Whitehouse S L, Graves S E, Pratt N L, Ryan P, Crawford R W. Risk factors for revision for early dislocation in total hip arthroplasty. J Arthroplasty 2008; 23(6): 867-72. Hailer N P, Weiss R J, Stark A, Karrholm J. The risk of revision due to dislocation after total hip arthroplasty depends on surgical approach, femoral head size, sex, and primary diagnosis: an analysis of 78,098 operations in the Swedish Hip Arthroplasty Register. Acta Orthop 2012; 83(5): 442-8. Havelin L I, Robertsson O, Fenstad A M, Overgaard S, Garellick G, Furnes O. A Scandinavian experience of register collaboration: the Nordic Arthroplasty Register Association (NARA). J Bone Joint Surg Am 2011; 93(Suppl. 3): 13-9. Jameson S S, Kyle J, Baker P N, Mason J, Deehan D J, McMurtry I A, Reed M R. Patient and implant survival following 4323 total hip replacements for acute femoral neck fracture: a retrospective cohort study using National Joint Registry data. J Bone Joint Surg Br 2012; 94(11): 1557-66. Jobory A, Karrholm J, Overgaard S, Becic Pedersen A, Hallan G, Gjertsen J E, Makela K, Rogmark C. Reduced revision risk for Dual-Mobility Cup in total hip replacement due to hip fracture: a matched-air analysis of 9,040 cases from the Nordic Arthroplasty Register Association (NARA). J Bone Joint Surg Am 2019; 101(14): 1278-85 . Johansson T. Internal fixation compared with total hip replacement for displaced femoral neck fractures: a minimum fifteen-year follow-up study of a previously reported randomized trial J Bone Joint Surg Am 2014; 96(6): e46. 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. Kuss O, Blettner M, Börgermann J. Propensity score: an alternative method of analyzing treatment effects. Deutsches Arzteblatt Int 2016; 113(35-36): 597-603. Mjaaland K E, Svenningsen S, Fenstad A M, Havelin L I, Furnes O, Nordsletten L. Implant survival after minimally invasive anterior or anterolateral vs conventional posterior or direct lateral approach: an analysis of 21,860 total hip arthroplasties from the Norwegian Arthroplasty Register (2008 to 2013). J Bone Joint Surg Am 2017; 99(10): 840-7. Noticewala M, Murtaugh T S, Danoff J, Cunn G J, Shah R P, Geller J. Has the risk of dislocation after total hip arthroplasty performed for displaced femoral neck fracture improved with modern implants? J Clin Orthop Trauma 2018; 9(4): 281-4. Peters R M, van Steenbergen L N, Stewart R E, Stevens M, Rijk P C, Bulstra S K, Zijlstra W P. Patient characteristics influence revision rate of total hip arthroplasty: American Society of Anesthesiologists score and body mass index were the strongest predictors for short-term revision after primary total hip arthroplasty. J Arthroplasty 2020; 35(1): 188-92. Ranstam J, Kärrholm J, Pulkkinen P, Mäkelä K, Espehaug B, Pedersen A B, Mehnert F, Furnes O, group, NARA Srudy Group. Statistical analysis of arthroplasty data II Guidelines. Acta Orthop 2011; 82(3): 258-67.

Acta Orthopaedica 2020; 91 (4): 401â&#x20AC;&#x201C;407

Rogmark C, MĂśller M, Sundfeldt M, Wolf O. The Swedish Fracture Register Annual Report 2016; 2017, https://stratum registercentrum se/Handlers/ ResourceManager ashx?ID=32491. Accessed December 6, 2018. Sariali E, Lazennec J Y, Khiami F, Catonne Y. Mathematical evaluation of jumping distance in total hip arthroplasty: influence of abduction angle, femoral head offset, and head diameter. Acta Orthop 2009; 80(3): 277-82. Tabori-Jensen S, Hansen T B, Stilling M. Low dislocation rate of Saturne((R))/ Avantage((R)) dual-mobility THA after displaced femoral neck fracture: a cohort study of 966 hips with a minimum 16-year follow-up. Arch Orthop Trauma Surg 2019; 139(5): 605-12. Tsikandylakis G, Karrholm J, Hailer N P, Eskelinen A, Makela K T, Hallan G, Furnes O N, Pedersen A B, Overgaard S, Mohaddes M. No increase in survival for 36-mm versus 32-mm femoral heads in metal-on-polyethylene

407

THA: a registry study. Clin Orthop Relat Res 2018a; 476(12): 2367-78. Tsikandylakis G, Mohaddes M, Cnudde P, Eskelinen A, Karrholm J, Rolfson O. Head size in primary total hip arthroplasty. EFORT Open Rev 2018b; 3(5): 225-31. Tsuda K, Haraguchi K, Koyanagi J, Takahashi S, Sugama R, Fujiwara K. A forty millimetre head significantly improves range of motion compared with a twenty eight millimetre head in total hip arthroplasty using a computed tomography-based navigation system. Int Orthop 2016; 40(10): 2031-9. Zijlstra W P, De Hartog B, Van Steenbergen L N, Scheurs B W, Nelissen R. Effect of femoral head size and surgical approach on risk of revision for dislocation after total hip arthroplasty. Acta Orthop 2017; 88(4): 395-401.

408

Acta Orthopaedica 2020; 91 (4): 408–413

Mortality and revision rate of cemented and uncemented hemiarthroplasty after hip fracture: an analysis of the Dutch Arthroplasty Register (LROI) Bouke J DUIJNISVELD 1, Koen L M KOENRAADT 2, Liza N VAN STEENBERGEN 3, and Stefan B T BOLDER 4 1 Department of Orthopaedic Surgery, Sint Maartenskliniek, Nijmegen; 2 Foundation for Orthopaedic Research, Care and Education, Amphia Hospital, Breda; 3 Dutch Arthroplasty Register, Landelijke Registratie Orthopedische Implantaten (LROI), ‘s Hertogenbosch; 4 Department of Orthopaedic Surgery, Amphia Hospital, Breda, The Netherlands Correspondence: bduijnisveld@gmail.com Submitted 2019-07-09. Accepted 2020-03-24.

Background and purpose — Femoral neck fractures are commonly treated with cemented or uncemented hemiarthroplasties (HA). We evaluated differences in mortality and revision rates in this fragile patient group. Patients and methods — From January 1, 2007 until December 31, 2016, 22,356 HA procedures from the Dutch Arthroplasty Register (LROI) were included. For each HA, follow-up until death, revision, or end of follow-up (December 31, 2016) was determined. The crude revision rate was determined by competing risk analysis. Multivariable Cox regression analyses were performed to evaluate the effect of fixation method (cemented vs. uncemented) on death or revision. Age, sex, BMI, Orthopaedic Data Evaluation Panel (ODEP) rating, ASA grade, surgical approach, and previous surgery were included as potential confounders. Results — 1-year mortality rates did not differ between cemented and uncemented HA. 9-year mortality rates were 53% (95% CI 52–54) in cemented HA compared to 56% (CI 54–58) in uncemented HA. Multivariable Cox regression analysis showed similar mortality between cemented and uncemented HA (HR 1.0, CI 0.96–1.1). A statistically significantly lower 9-year revision rate of 3.1% (CI 2.7–3.6) in cemented HA compared with 5.1% (CI 4.2–6.2) in the uncemented HA was found with a lower hazard ratio for revision in cemented compared with uncemented HA (HR 0.56, CI 0.47–0.67). Interpretation — Long-term mortality rates did not differ between patients with a cemented or uncemented HA after an acute femoral neck fracture. Revision rates were lower in cemented compared with uncemented HA.

The number of hemiarthroplasties (HA) after displaced femoral neck fracture increases as a result of global aging, and inferior results and high risk of reoperation after internal fixation. Although the literature on the decision to use cemented or uncemented HA may favor a cemented implant, both techniques are currently used. The use of bone cement is associated with bone cement implantation syndrome (BCIS) characterized by hypoxia, hypotension, loss of consciousness around the time of bone cementation, and intraoperative death (Olsen et al. 2014, Rutter et al. 2014). More intraoperative complications including intraoperative death were found in cemented HA in the Norwegian register (Gjertsen et al. 2012, Talsnes et al. 2013). However, no differences in mortality were found after 1 week (Costain et al. 2011, Yli-Kyyny et al. 2014). More studies including randomized controlled trials (Deangelis et al. 2012, Taylor et al. 2012) and registry studies (Costa et al. 2011, Ekman et al. 2019) did not show differences in mortality between cemented and uncemented HA. Randomized controlled trials (Taylor et al. 2012, Langslet et al. 2014, Inngul et al. 2015) and register studies (Gjertsen et al. 2012, Yli-Kyyny et al. 2014) have shown that the use of uncemented implants could result in a higher risk of periprosthetic fractures. A meta-analysis by Li et al. (2013) concluded that differences in several outcome parameters indicated cemented hemiarthroplasty to be superior to the uncemented counterpart. However, a serious flaw in this analysis is that several studies were included using an outdated stem like the Austin Moore (Sonne-Holm et al. 1982, Emery et al. 1991, Parker et al. 2010) and the experimental uncemented Thomson stem (Sadr and Arden 1977). The use of a prosthesis without Orthopaedic Data Evaluation Panel (ODEP) rating > 3A could influence outcome and is therefore discouraged (Grammatopoulos

Acta Orthopaedica 2020; 91 (4): 408–413

et al. 2015). A recent review by Rogmark and Leonardsson (2016) included 5 randomized studies comparing modern uncemented and cemented hemiarthroplasties. They found no differences in mortality, but more periprosthetic fractures in uncemented cases. We compared cemented and uncemented HA after an acute hip fracture with primary outcome mortality and revision rate. Data from the Dutch Arthroplasty Register (LROI) were used and the cohort of cemented HAs was compared with uncemented HAs, accounting for the ODEP rating and other confounders.

Patients and methods The Dutch Arthroplasty Register (LROI) is a nationwide population-based register that includes information on arthroplasties in the Netherlands since 2007. It covers 100% of Dutch hospitals and has a completeness of reporting of 70% for primary orthopedic HAs in 2013 to 88% in 2016 (Van Steenbergen et al. 2015). The LROI database contains information on patient, procedure, and prosthesis characteristics registered by registrars from each hospital. For each component a product number is registered to identify the characteristics of the prosthesis. Vital status of all patients was obtained actively on a regular basis from Vektis, the national insurance database on health care in the Netherlands, which records all deaths of Dutch citizens. The LROI requires the informed consent of patients and uses an opt-out system in this respect. For this study, we included all HA procedures in patients with an acute femoral neck fracture registered by orthopedic surgeons in the LROI from January 1, 2007 until December 31, 2016 (N = 22,351). All femoral stems were classified as ODEP rating 3A or other/no rating by checking the prosthesis in the ODEP database (http://www.odep.org.uk). Prostheses that were not available in the ODEP database were manually explored by the researchers. When these prostheses were not found as ODEP rating ≥ 3A, they were classified as “no ODEP 3A rating.” Other parameters such as age, sex, BMI, ASA classification, and previous surgery on the affected hip were used from the LROI database. BMI and smoking have only been available in the LROI since 2014. Closed reductions after a dislocation or incision and drainage for infection were not included in the LROI, as in these procedures no component exchange was performed. The median follow-up was 1.8 years (0–10). Statistics Kaplan–Meier survival analysis was performed to examine the survival rates of the patients over time. A multivariable Cox proportional hazard analysis was performed to examine the effect of fixation type (i.e., cemented vs. uncemented) on death after HA. Demographic variables such as age, sex, ASA classification, BMI, and smoking habit were included as covariates. Age and BMI with impossible values were excluded.

409

Survival time for revision was calculated as the time from primary HA to the first revision arthroplasty for any reason, death of the patient, or the end of the study follow-up (December 31, 2016). Cumulative crude incidence of revision was calculated using competing risk analysis where death was considered to be the competing event (Lacny et al. 2015, Wongworawat et al. 2015). Multivariable Cox proportional hazard analyses were performed to examine the effect of fixation type on revision. Demographic variables and possible risk factors including approach, ODEP rating (> 3A vs. other/no rating), and previous surgery on the affected hip were included as covariates. Furthermore, the reasons for revision were compared between cemented and uncemented HA. Finally, multivariable logistic regression analyses were performed on the revisions (n = 517) to examine independent risk factors for revision due to dislocation, infection, femoral loosening, periprosthetic fracture, or other reasons for revision. Demographic variables and possible risk factors were included as covariates in the multivariable logistic regression analysis. All confidence intervals (CI) are defined as 95%. For the CI, we assumed that the number of observed cases followed a Poisson distribution. Ethics, funding, and potential conflicts of interest Ethical approval was not required for this study. Our Foundation for Orthopedic Research, Care and Education (FORCE) receives money from Zimmer-Biomet, Stryker, and Mathys not directly related to this study. SB is a consultant for Stryker.

Results Patients and procedure characteristics are given in Table 1. Mortality 1-week mortality was 2.1% (CI 1.8–2.3) in the cemented HA group compared with 1.8% (CI 1.6–2.2) in the uncemented HA group. 1-month mortality was 6.0% (CI 5.6–6.4) in the cemented HA group compared with 5.4% (CI 4.9–6.0) in the uncemented HA group. 1-year mortality was 19.7% (CI 19.1–20.4) in the cemented HA group compared with 19.5% (CI 18.6–20.4) in the uncemented HA group. 9-year mortality rates were 53% (CI 52–54) in the cemented HA group compared with 56% (CI 54–58) in the uncemented HA group (Figure 1). Univariable Cox regression analysis showed no statistically significant difference in mortality between patients with a cemented and an uncemented HA (HR 0.99, CI 0.95–1.04). Multivariable Cox regression analysis showed an HR of 1.00 (CI 0.96–1.05) adjusted for age at surgery, sex, and ASA classification. In a subset from 2014 to 2016 that included also BMI and smoking, no statistically significant difference in mortality in cemented and uncemented HA was found (HR 1.06, CI 0.97–1.15).

410

Acta Orthopaedica 2020; 91 (4): 408–413

Table 1. Patient characteristics of cemented (n = 14,736) and uncemented (n = 7,615) hemiarthroplasties in the Netherlands 2007– 2016. Values are number (%) unless otherwise specified

Cumulative survival 1.0 Cemented Uncemented

0.9

Factor

Cemented Uncemented hemiarthroplasty hemiarthroplasty n = 14,736 n = 7,615

0.8 0.7

Mean age (SD) 82.5 (8.1) Sex Male 4,345 (30) Female 10,362 (70) Missing 29 (0.2) ASA classification I 334 (2) II 5,514 (37) III or IV 8,490 (58) Missing 398 (3) Previous surgery on affected joint Yes 147 (1) No 13,858 (94) Missing 575 (4) Smoking a Yes 526 (4) No 5,370 (36) Missing 8,840 (60) Surgical approach Posterolateral 7,639 (52) Direct lateral 5,149 (35) Anterolateral 1,646 (11) Anterior 172 (1) Other or missing 117 (1) ODEP 3A rating Yes 13,189 (90) No 1,091 (7) Missing 203 (3) Mean BMI (SD) a 24.3 (4.2)

82.7 (7.9) 2,271 (30) 5,317 (70) 27 (0.4) 172 (2) 2,730 (36) 4,540 (60) 173 (2) 71 (1) 6,995 (92) 480 (6) 193 (3) 2,603 (34) 4,819 (63) 4,098 (54) 2,330 (31) 884 (12) 230 (3) 60 (1) 5,078 (67) 2,334 (31) 456 (3) 23.6 (5.1)

BMI: body mass index, ASA: American Society of Anesthesiologists, ODEP: Orthopaedic Data Evaluation Panel. a Only available for the period 2014–2016.

0.6 0.5 0.4

Years after index operation

Figure 1. Cumulative mortality rate of uncemented (n = 7,615) and cemented (n = 14,736) hemiarthroplasties in the Netherlands 2007– 2016. Cumulative revision rate (%) 6 Cemented Uncemented

5 4 3 2 1 0

Years after index operation

Figure 2. Crude cumulative revision rate of uncemented (n = 7,615) and cemented (n = 14,736) hemiarthroplasties in the Netherlands 2007–2016.

Revision tion was lower in direct lateral approach (HR 0.37, CI 0.24– Competing risk analysis showed a lower crude revision rate at 0.56) and anterolateral approach (HR 0.32, CI 0.16–0.66) 1 month (0.5%, CI 0.4–0.7), 1 year (1.3%, CI 1.1–1.5), and 9 compared with posterolateral approach. years (3.1%, CI 2.7–3.6) in cemented HA compared with the Risk for revision because of infection was statistically sigrevision rate at 1 month (1.1%, CI 0.9–1.4), 1 year (2.5%, CI nificantly higher in patients with previous surgery (HR 4.0, CI 2.1–2.9), and 9 years (5.1%, CI 4.2–6.2) in uncemented HA 1.3–13). Femoral stem loosening was less often the reason for (Figure 2). Multivariable Cox regression revealed a lower hazard ratio for revision (HR 0.56, CI 0.47– Table 2. Reasons for revision of cemented and uncemented hemiarthroplasty 0.67) in cemented compared with uncemented HA, adjusted for confounders including sex, age, ASA Cemented Uncemented classification, approach, ODEP rating, and previous n = 14,726 n = 7,551 Hazard a surgery on the affected hip. These findings were perReason n (%) n (%) ratio (95% CI) p-value sistent after adjusting also for BMI and smoking in Dislocation 92 (0.6) 62 (0.8) 0.77 (0.55–1.08) 0.1 the subset from 2014 to 2016. Reasons for revision Dislocation, infection, femoral loosening, and periprosthetic fractures were the most common reason for revision of HA (Table 2). Multivariable logistic regression analysis showed that the risk for disloca-

Infection 55 (0.4) Loosening of stem 19 (0.1) Periprosthetic fracture 12 (0.1) Other 105 (0.7) a Hazard

19 (0.2) 40 (0.5) 77 (1.0) 78 (1.0)

1.48 (0.85–2.56) 0.2 0.21 (0.12–0.36) < 0.001 0.07 (0.04–0.13) < 0.001 0.66 (0.48–0.90) 0.009

ratios with 95% CI and p-values are shown for logistic regression analysis for fixation type, adjusted for age, sex, ASA classification, previous surgery, surgical approach, and ODEP rating.

Acta Orthopaedica 2020; 91 (4): 408–413

revision in the cemented HA group compared with the uncemented HA group (HR 0.21, CI 0.12–0.36) and more often in patients aged 60– 80 compared with those aged > 80 years (HR 2.2, CI 1.3–3.8). The risk for revision because of a periprosthetic fracture was less in cemented HA compared with cemented HA (HR 0.07, CI 0.04–0.13).

Discussion In this study with more than 22,000 hemiarthroplasties for acute femoral neck fracture from the Dutch Arthroplasty Register, we found comparable 9-year mortality rates between cemented and uncemented HA. The 9-year revision rate was lower in cemented HA compared with uncemented HA. Dislocation, infection, femoral stem loosening, and periprosthetic fractures were the most common reasons for revision. The register data show that in the Netherlands one-third of hemiarthroplasties are performed with an uncemented stem. One of the reasons for choosing an uncemented stem may be the assumed risk for BCIS when using a cemented stem in the fragile patient. We found a difference in early mortality rates in favor of the uncemented group. This might indicate the presence of BCIS, although this cannot be proven from a register study and could also be the result of selection bias. Our 1-month mortality rate was lower in the uncemented HA group. However, the difference was small and the 1-month mortality rate in the cemented HA was comparable to mortality rates of 5–8% previously found in the literature (Costain et al. 2011, Olsen et al. 2014). From these results, in patients with high risk for BCIS, an uncemented HA may be a good option to improve the earliest outcome. Olsen et al. (2014) showed an incidence of 21%, 5.1%, and 1.7% of BCIS grades 1, 2, and 3 respectively with a 1-month mortality of 9.3%, 35%, and 88% respectively. BCIS could also influence morbidity due to hypoxia and hypotension leading to a higher mortality during follow-up. An increased mortality in the cemented HA group was, however, not observed compared with the uncemented HA group at 1- and 9-year follow-up. This effect may be due to a higher revision rate in the uncemented group. Our 1-year mortality rate of 20% is comparable to results from the Norwegian and the Swedish Registry (Leonardsson et al. 2012, Gjertsen et al. 2017). Our 9-year mortality rates of 53% and 56% are in line with the mortality rate of 45% found in the Swedish Registry after 7 years’ follow-up (Jawad et al. 2019). A recent study of the Norway Registry showed a higher mortality of about 90% at 9 years’ follow-up (Kristensen et al. 2019). This could be due to patient selection, as Kristensen et al. selected only patients of 70 years of age or older, whereas in the current study and in the study of the Swedish Registry no age selection was performed. Our findings of a lower revision rate after cemented HA, compared with uncemented HA, is supported by other register

411

studies (Gjertsen et al. 2012, Jameson et al. 2013, Yli-Kyyny et al. 2014, Kristensen et al. 2019). We have no data on reoperations other than revision procedures in the Dutch register. Other studies showed a reoperation rate of 7–11% at 5–19 years’ follow-up (Parker et al. 2010, Viberg et al. 2013). These reoperation rates cannot be compared with our revision rates as reoperations do not always include a revision. Surprisingly, ODEP rating > 3A did not influence revision rate in our study. Although individual prosthesis brands could in theory influence revision rate, the data on individual prosthesis brands were not available for this study. Dislocation, infection, femoral stem loosening, and periprosthetic fractures were the most common reasons for revision. The posterolateral approach was an independent risk factor for dislocation as shown earlier (Leonardsson et al. 2012, Rogmark et al. 2014, Moerman et al. 2018). The posterolateral approach could be considered as surgical approach in HA because functional outcome including pain, walking without mobility aids, and patient-reported outcome measures have been shown to be in favor of the posterolateral approach (Kristensen et al. 2017, Hongisto et al. 2018). We could not measure the functional outcome. Mukka et al. (2017) did not find any differences in functional outcome between direct lateral and posterolateral approach. We found an uncemented HA as independent risk factor for femoral loosening and periprosthetic fracture as previously shown (Leonardsson et al. 2012, Rogmark et al. 2014, Moerman et al. 2018), which could also be influenced by a lower threshold to revise an uncemented stem when compared with a cemented implant. The strength of this population-based registry study is the large study population of over 22,000 patients with follow-up of up to 9 years and the inclusion of several potential confounders such as patient, procedure, and prosthesis characteristics, like type of fixation and ODEP rating. A limitation of our study is the observational nature of the data. Therefore, causal relationships cannot be identified. To minimize selection bias, confounders including age, sex, ASA classification, BMI, previous surgery, and smoking habit were added to the multivariable regression analysis. However, potential residual confounding like socioeconomic factors and alcohol consumption could still be present. Severely ill elderly patients could have received internal fixation or non-operative treatment instead of HA. Also, “young old” patients could receive HA rather than THA or internal fixation. Both regimes would lower the mortality after HA. Furthermore, the Dutch Arthroplasty Register does not allow for comparison of individual prosthesis brands. Reoperations like debridement of the wound or Vancouver B1 fracture fixation without prosthesis component replacement are not registered in the Dutch Arthroplasty Register. This may influence the view on fracture rates in our cohort. In addition, closed reduction for a dislocated hip and acetabular erosion as reason for revision are not registered in the LROI database and are therefore not included in this study. Because of these reoperations without

412

revision of the prosthesis, reoperation rate will be higher than revision rate. An uncemented HA may be considered for the patient with high risk for BCIS with short life expectancy. Regular use of an uncemented stem does not seem to offer benefit to patients. We found no evidence that cemented HA leads to higher mortality in the longer term. In summary, based on the outcome of this study and earlier findings in the literature, in which longterm mortality rates were similar between cemented and uncemented HA for displaced femoral neck fracture and revision rates were lower in cemented HA, we recommend the use of a cemented HA for patients with an acute femoral neck fracture.

Conception of the study: BD, KK, and SB. Data analysis: BD, KK, LS, and SB. Preparation and final approval of the manuscript: BD, KK, LS, and SB. Acta thanks Jan-Erik Gjertsen and Cecilia Rogmark for help with peer review of this study.

Costa M L, Griffin X L, Pendleton N, Pearson M, Parsons N. Does cementing the femoral component increase the risk of peri-operative mortality for patients having replacement surgery for a fracture of the neck of femur? Data from the National Hip Fracture Database. J Bone Jt Surg—Ser B 2011; 93(10): 1405-10. Costain D J, Whitehouse S L, Pratt N L, Graves S E, Ryan P, Crawford R W. Perioperative mortality after hemiarthroplasty related to fixation method. Acta Orthop 2011; 82(3): 275-81. Deangelis J P, Ademi A, Staff I, Lewis C G. Cemented versus uncemented hemiarthroplasty for displaced femoral neck fractures: a prospective randomized trial with early follow-up. J Orthop Trauma 2012; 26(3): 135-40. Ekman E, Laaksonen I, Isotalo K, Liukas A, Vahlberg T, Mäkelä K. Cementing does not increase the immediate postoperative risk of death after total hip arthroplasty or hemiarthroplasty: a hospital-based study of 10,677 patients. Acta Orthop 2019; 90(3): 270-4. Emery R J, Broughton N S, Desai K, Bulstrode C J, Thomas T L. Bipolar hemiarthroplasty for subcapital fracture of the femoral neck: a prospective randomised trial of cemented Thompson and uncemented Moore stems. J Bone Joint Surg Br 1991; 73(2): 322-4. Gjertsen J E, Lie S A, Vinje T, Engesæter L B, Hallan G, Matre K, Furnes O. More re-operations after uncemented than cemented hemiarthroplasty used in the treatment of displaced fractures of the femoral neck: an observational study of 11 116 hemiarthroplasties from a national register. J Bone Joint Surg Br 2012; 94(8): 1113-19. Gjertsen J E, Dybvik E, Furnes O, Fevang J M, Havelin L I, Matre K, Engesæter L B. Improved outcome after hip fracture surgery in Norway: 10-year results from the Norwegian Hip Fracture Register. Acta Orthop 2017; 88(5): 505-11. Grammatopoulos G, Wilson H A, Kendrick B J L, Pulford E C, Lippett J, Deakin M, Andrade A J, Kambouroglou G. Hemiarthroplasty using cemented or uncemented stems of proven design: a comparative study. Bone Joint J 2015; 97-B(1): 94-9. Hongisto M T, Nuotio M S, Luukkaala T, Väistö O, Pihlajamäki H K. Lateral and posterior approaches in hemiarthroplasty. Scand J Surg 2018; 107(3): 260-8. Inngul C, Blomfeldt R, Ponzer S, Enocson A. Cemented versus uncemented arthroplasty in patients with a displaced fracture of the femoral neck: a randomised controlled trial. Bone Joint J 2015; 97-B(11): 1475-80.

Acta Orthopaedica 2020; 91 (4): 408–413

Jameson S S, Jensen C D, Elson D W, Johnson A, Nachtsheim C, Rangan A, Muller S D, Reed M R. Cemented versus cementless hemiarthroplasty for intracapsular neck of femur fracture: a comparison of 60,848 matched patients using national data. Injury 2013; 44(6): 730-4. Jawad Z, Nemes S, Bülow E, Rogmark C, Cnudde P. Multi-state analysis of hemi- and total hip arthroplasty for hip fractures in the Swedish population: results from a Swedish national database study of 38,912 patients. Injury 2019; 50(2): 272-7. Kristensen T B, Vinje T, Havelin L I, Engesæter L B, Gjertsen J E. Posterior approach compared to direct lateral approach resulted in better patientreported outcome after hemiarthroplasty for femoral neck fracture. Acta Orthop 2017; 88(1): 29-34. Kristensen T B, Dybvik E, Kristoffersen M, Dale H, Engesæter L B, Furnes O, Gjertsen J-E. Cemented or uncemented hemiarthroplasty for femoral neck fracture? Data from the Norwegian Hip Fracture Register. Clin Orthop Relat Res 2019; Epub ahead of print. 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. Langslet E, Frihagen F, Opland V, Madsen J E, Nordsletten L, Figved W. Cemented versus uncemented hemiarthroplasty for displaced femoral neck fractures: 5-year followup of a randomized trial. Clin Orthop Relat Res 2014; 472(4): 1291-9. Leonardsson O, Kärrholm J, Åkesson K, Garellick G, Rogmark C. Higher risk of reoperation for bipolar and uncemented hemiarthroplasty. Acta Orthop 2012; 83(5): 459-66. Li T, Zhuang Q, Weng X, Zhou L, Bian Y. Cemented versus uncemented hemiarthroplasty for femoral neck fractures in elderly patients: a metaanalysis. PLoS One 2013; 8(7): e68903. Moerman S, Mathijssen N M C, Tuinebreijer W E, Vochteloo A J H, Nelissen R G H H. Hemiarthroplasty and total hip arthroplasty in 30,830 patients with hip fractures: data from the Dutch Arthroplasty Register on revision and risk factors for revision. Acta Orthop 2018; 89(5): 509-14. Mukka S, Knutsson B, Majeed A, Sayed-Noor A S. Reduced revision rate and maintained function after hip arthroplasty for femoral neck fractures after transition from posterolateral to direct lateral approach. Acta Orthop 2017; 88(6): 627-33. Olsen F, Kotyra M, Houltz E, Ricksten S E. Bone cement implantation syndrome in cemented hemiarthroplasty for femoral neck fracture: incidence, risk factors, and effect on outcome. Br J Anaesth 2014; 113(5): 800-6. Parker M I, Pryor G, Gurusamy K. Cemented versus uncemented hemiarthroplasty for intracapsular hip fractures: a randomised controlled trial in 400 patients. J Bone Joint Surg Br 2010; 92(1): 116-22. 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, Fenstad A M, Leonardsson O, Engesæter L B, Kärrholm J, Furnes O, Garellick G, Gjertsen J E. Posterior approach and uncemented stems increases the risk of reoperation after hemiarthroplasties in elderly hip fracture patients. Acta Orthop 2014; 85(1): 18-25. Rutter P D, Panesar S S, Darzi A, Donaldson L J. What is the risk of death or severe harm due to bone cement implantation syndrome among patients undergoing hip hemiarthroplasty for fractured neck of femur? A patient safety surveillance study. BMJ Open 2014; 4(6): e004853. Sadr B, Arden G P. A comparison of the stability of Proplast-coated and cemented Thompson prostheses in the treatment of subcapital femoral fractures. Injury 1977; 8(3): 234-7. Sonne-Holm S, Walter S, Jensen J S. Moore hemi-arthroplasty with and without bone cement in femoral neck fractures: a clinical controlled trial. Acta Orthop Scand 1982; 53(6): 953-6. Van Steenbergen L N, Denissen G A W, Spooren A, Van Rooden S M, Van Oosterhout F J, Morrenhof J W, Nelissen R G H H. More than 95% completeness of reported procedures in the population-based Dutch Arthroplasty Register. Acta Orthop 2015; 86(4): 498-505.

Acta Orthopaedica 2020; 91 (4): 408–413

Talsnes O, Vinje T, Gjertsen J E, Dahl O E, Engesæter L B, Baste V, Pripp A H, Reikerås O. Perioperative mortality in hip fracture patients treated with cemented and uncemented hemiprosthesis: a register study of 11,210 patients. Int Orthop 2013; 37(6): 1135-40. Taylor F, Wright M, Zhu M. Hemiarthroplasty of the hip with and without cement: a randomized clinical trial. J Bone Joint Surg Am 2012; 94(7): 577-83. Viberg B, Overgaard S, Lauritsen J, Ovesen O. Lower reoperation rate for cemented hemiarthroplasty than for uncemented hemiarthroplasty and

413

internal fixation following femoral neck fracture: 12- to 19-year follow-up of patients aged 75 years or more. Acta Orthop 2013; 84(3): 254-9. Wongworawat M D, Dobbs M B, Gebhardt M C, Gioe T J, Leopold S S, Manner P A, Rimnac C M, Porcher R. Editorial: Estimating survivorship in the face of competing risks. Clin Orthop Relat Res 2015; 473(4): 1173-6. Yli-Kyyny T, Sund R, Heinänen M, Venesmaa P, Kröger H. Cemented or uncemented hemiarthroplasty for the treatment of femoral neck fractures? Acta Orthop 2014; 85(1): 49-53.

414

Acta Orthopaedica 2020; 91 (4): 414–419

Increases in the rates of primary and revision knee replacement are reducing: a 15-year registry study across 3 continents Peter L LEWIS 1,4, Stephen E GRAVES 1, Otto ROBERTSSON 2,4, Martin SUNDBERG 2,4, Elizabeth W PAXTON 3, Heather A PRENTICE 3, and Annette W-DAHL 2,4 1 Australian Orthopaedic Association National Joint Replacement Registry, Adelaide, Australia; 2 Swedish Knee Arthroplasty Register, Lund, Sweden; 3 Surgical Outcomes and Analysis, Kaiser Permanente, San Diego, CA, USA; 4 Lund University, Faculty of Medicine, Clinical Sciences Lund, Department

Orthopedics, Lund, Sweden Correspondence: plewis@aoanjrr.org.au Submitted 2019-10-30. Accepted 2020-01-16.

Background and purpose — Rates of knee replacement (KR) are increasing worldwide. Based on population and practice changes, there are forecasts of a further exponential increase in primary knee replacement through to 2030, and a corresponding increase in revision knee replacement. We used registry data to document changes in KR over the past 15 years, comparing practice changes across Sweden, Australia, and the United States. This may improve accuracy of future predictions. Patients and methods — Aggregated data from the Swedish Knee Arthroplasty Register (SKAR), the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR), and the Kaiser Permanente Joint Replacement Registry (KPJRR) were used to compare surgical volume of primary and revision KR from 2003 to 2017. Incidence was calculated using population census statistics from Statistics Sweden and the Australian Bureau of Statistics, as well as the yearly active membership numbers from Kaiser Permanente. Further analysis of KR by age < 65 and ≥ 65 years was carried out. Results — All registries recorded an increase in primary and revision KR, with a greater increase seen in the KPJRR. The rate of increase slowed during the study period. In Sweden and Australia, there was a smaller increase in revision surgery compared with primary procedures. There was consistency in the mean age at surgery, with a steady small decrease in the proportion of women having primary KR. The incidence of KR in the younger age group remained low in all 3 registries, but the proportional increases were greater than those seen in the ≥ 65 years of age group. Interpretation — There has been a generalized deceleration in the rate of increase of primary and revision KR. While there are regional differences in KR incidence, and rates of change, the rate of increase does not seem to be as great as previously predicted.

Knee replacement (KR) has a favorable survival rate with cumulative revision as low as 3% at 10 years (AOANJRR 2018, SKAR 2018) and this result appears to be improving with time as wear-related revisions become less common (Sharkey et al. 2014, Koh et al. 2017, Postler et al. 2018). Throughout the last decade, national joint replacement registries have recorded increasing yearly volumes of KR (AOANJRR 2018, NJR 2018, SKAR 2018). The reasons for this increase in procedure numbers are proposed to be increased surgeon and patient acceptance of KR (Hamilton et al. 2015), improved longevity (Patel et al. 2015), increasing incidence of osteoarthritis (OA), and use of KR in younger patients (Weinstein et al. 2013, Leyland et al. 2016, Karas et al. 2019). With increasing primary KR use it is predicted that the numbers of revision procedures will also rise (Kumar et al. 2015, Patel et al. 2015). Not only are more people receiving a KR, but some of the factors driving increased primary usage of KR also contribute to increased failure. These include longer lifeexpectancy, whereby patients with a KR have more time to be revised, and use in young and obese patients who place higher demands on their KR (Hamilton et al. 2015). Counterbalancing this trend, to a small extent, is improved prosthesis performance (Pitta et al. 2018). There is international variation in the use of KR (Kurtz et al. 2011). In a comparative study of 18 countries in 2008, Kurtz et al. (2011) found a range of 8.6 to 213 primary procedures /100,000 population, and a range of 0.2 to 28 revision procedures/100,000 population, but they could not determine if the observed variation related to healthcare systems, access to care, number and distribution of orthopedic surgeons, or the prevalence of joint disease. There are expectations of exponential increases for both primary and revision KR. However, predictions of revision KR in the year 2030 compared with 2005 levels vary widely, from a 75% increase in Taiwan to a 600% increase in the USA and a similar increase in the UK

Acta Orthopaedica 2020; 91 (4): 414–419

415

Table 1. Yearly totals of knee replacement (KR) procedures recorded in the SKAR, AOANJRR, and KPJRR KR type a

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Sweden Primary 8,832 9,195 9,797 10,691 10,527 11,004 12,841 12,848 12,845 13,411 13,361 13,145 12,924 14,053 14,964 Uni 982 892 928 916 728 712 693 689 594 536 494 465 648 984 1169 PF 11 16 21 9 12 17 37 31 52 43 56 58 65 52 48 Total 7,339 8,287 8,848 9,766 9,787 10,275 12,111 12,228 12,198 12,832 12,808 12,622 12,206 13,008 13,743 Revision 596 625 650 650 657 702 758 860 845 869 1002 959 936 934 945 All 9,428 9,820 10,447 11,341 11,184 11,706 13,599 13,708 13,690 14,280 14,363 14,104 13,860 14,987 15,909 Australia Primary 26,008 7,540 30,409 31,231 33,064 36,160 37,683 40,838 43,051 44,839 46,903 49,813 53,578 55,878 59,002 Uni 4,109 3,730 3,382 3,628 2,502 3,225 3,087 2,615 2,411 21,46 2,137 2,270 2,557 3,056 3,652 PF 151 180 174 181 195 232 229 268 247 225 246 234 248 307 298 Total 21,735 23,603 26,337 27,376 29,294 32,622 34,307 37,922 40,375 42,453 44,495 47,288 50,763 52,510 55,077 Revision 2,314 2,663 2,721 2,826 2,994 3,250 3,294 3,716 3,894 3,910 4,173 4,301 4,447 4,559 4,791 All 28,322 30,203 33,130 34,057 36,058 39,410 40,977 44,554 46,945 48,749 51,076 54,114 58,025 60,437 63,793 Kaiser Permanente Primary 4,271 5,824 7,050 8,255 9,283 10,234 10,806 12,904 13,495 14,084 15,445 17,796 18,324 20,093 20,672 Uni 144 234 210 212 200 330 448 420 371 439 522 631 602 563 579 PF 7 6 6 14 10 24 27 35 38 30 44 57 54 65 84 Total 4,120 5,584 6,834 8,029 9,073 9,880 10,331 12,449 13,086 13,616 14,879 17,109 17,669 19,465 20,009 Revision 274 363 456 556 627 773 766 850 981 1,021 1,091 1,173 1,267 1,305 1,309 All 4,545 6,187 7,506 8,810 9,910 11,007 11,572 13,754 14,476 15,106 16,536 18,969 19,592 21,398 21,981 a Uni

= unicompartmental; PF= patellofemoral Note: A small number of other primary knee replacement (unispacer, partial resurfacing, bicompartmental) are included in primary knee totals.

(Kurtz et al. 2007, Kumar et al. 2015, NJR 2018). A further study comparing 24 OECD countries’ KR utilization predicted a 400% increase by 2030 (Pabinger et al. 2015). There are other predictive models with a more conservative forecast for the United States (Inacio et al. 2017). We performed a multi-country comparison of KR, comparing the changing procedure volume and incidence of primary and revision KR using data from the Swedish Knee Arthroplasty Register (SKAR), the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR), and the Kaiser Permanente Joint Replacement Registry (KPJRR) over a 15 year period (2003–2017).

Patients and methods Data were obtained for the period January 1, 2003 until December 31, 2017 for KR procedures recorded in the SKAR, AOANJRR, and the KPJRR. Primary KR procedures were defined as all initial unicompartmental, patellofemoral, and total KR. Where replacements were bilateral, both knees were included. Revision KR included all revision procedures of a previous replacement (partial or total) where 1 or more components were added, removed, or exchanged, regardless of whether this was the 2nd or subsequent procedure in chronology. The capture rate of these registries exceeds 95% and loss to follow-up was less than 8% over the study period. Validation and quality control methods of these registries have been published previously (Paxton et al. 2010, Robertsson et al. 2014, AOANJRR 2018).

There were 1,133,079 KR included in this analysis. The SKAR contributed 199,020 KR (186,473 primary and 12,547 revision procedures), there were 732,521 KR from the AOANJRR (674,045 primary and 58,476 revision procedures), and 201,350 KR from the KPJRR (188,538 primary and 12,812 revision procedures) (Table 1). Statistics Aggregated data regarding type of procedure as well as patient age and sex were obtained. Population data were obtained from Statistics Sweden and the Australian Bureau of Statistics, as well as the yearly active membership numbers from Kaiser Permanente. Comparisons were made between countries for yearly procedure volume for both primary and revision KR, as well as yearly incidence per 100,000 population. Stratified analysis for ages < 65 and ≥ 65 years was also carried out. Inclusion of bilateral procedures and multiple revisions was thought to affect each country’s analysis similarly. Mean age and sex tables were compiled and the proportions by ages < 65 and ≥ 65 years for both primary and revision KR were included. Annual percentage change in procedure volume for both primary and revision KR was calculated and the mean for each of the 3 5-year time periods was derived, as described by Patel et al. (2015), to summarize trends in these procedures over time and across countries. Ethics, funding, and conflicts of interest Ethics approval covering the SKAR data use was approved by the Ethics Board of Lund University (LU20-02). The AOAN-

416

Acta Orthopaedica 2020; 91 (4): 414–419

increased from 596 in Sweden to 945, from 2,314 in Australia to 4,791, and from 274 in the KPJRR to 1,309 (Figure 1). Primary KR 60,000 6,000 volume increases were 79% in Sweden, 127% 50,000 5,000 in Australia, and 384% in the KPJRR. During the same time period, revision KR procedure 40,000 4,000 volume increases were 59% in Sweden, 107% 30,000 3,000 in Australia, and 378% in the KPJRR. The incidence of primary KR per 100,000 20,000 2,000 population over this same time span in Sweden 10,000 1,000 increased from 73 to 131 and revision KR incidence increased from 6.6 to 9.4, while in Aus0 0 2003 2005 2007 2009 2011 2013 2015 2017 2003 2005 2007 2009 2011 2013 2015 2017 tralia primary KR incidence rose from 132 to Figure 1. Yearly procedure volume of primary KR (left panel) and revision KR (right panel) 240 and revision KR incidence increased from recorded by the SKAR, AOANJRR, and KPJRR from 2003 to 2017. 11.7 to 19.5. In the KPJRR cohort primary KR incidence/105 insured increased from 52 to 187 Incidence of primary KR/10 Incidence of revision KR/10 and revision KR from 3.3 to 11.8 (Figure 2). By 300 30 Australia Australia this measure, primary KR incidence increased KPJRR KPJRR Sweden Sweden 250 25 from 2003 to 2017 by 79% in Sweden, 102% in Australia, and 258% in the KPJRR, while 200 20 over this same time revision KR incidence increased by 42% in Sweden, 63% in Austra150 15 lia, and 255% in the KPJRR. When stratified by age < 65 and ≥ 65 years, the annual inci100 10 dence/105 population for the younger group 50 5 remained less than 90 for primary KR and less than 8 for revision KR in all 3 registries, while 0 0 the older cohort from the KPJRR showed the 2003 2005 2007 2009 2011 2013 2015 2017 2003 2005 2007 2009 2011 2013 2015 2017 Figure 2. Yearly incidence of primary KR (left panel) and revision KR (right panel) per 105 largest increases (from 320 to 884 for primary population recorded by the SKAR, AOANJRR, and KPJRR from 2003 to 2017. KR and from 21 to 57 for revision KR) (Figure 3). While the incidence/105 in the younger age JRR is a declared Commonwealth of Australia Quality Assur- group remained low, the proportional change over the 15 years ance Activity under section 124X of the Health Insurance Act, in this group for primary KR was 76%, 141%, and 276% for 1973. All AOANJRR studies are conducted in accordance Sweden, Australia, and the KPJRR, respectively, while it was with ethical principles of research (Helsinki Declaration 35%, 58%, and 177% for the ≥ 65 years age group. Over the II). Approval for inclusion of data from the Kaiser Perman- same time period the increases for revision KR incidence for ente Joint Replacement Registry Institutional Review Board the < 65 years age group were 39%, 85%, and 277%, and for the ≥ 65 years age group 26%, 32%, and 171% in Sweden, approval (#5488) was granted on November 15, 2018. Australia, and the KPJRR, respectively. There was no funding. There are no conflicts of interest. When the mean change for each of the 3 5-year periods was calculated for primary and revision KR, all regions showed a deceleration in the increase. The exception is an increase in Results revision in Sweden between the periods 2003–2007 to 2008– Throughout the 15 years from 2003 to 2017, annual primary 2012 (Figure 4). KR procedure volume increased from 8,832 in 2003 to 14,964 During the study period, the mean age of primary and reviin 2017 in Sweden, from 26,008 to 59,002 in Australia, and sion KR patients remained stable in all countries (Table 2). from 4,271 to 20,672 in the KPJRR. The proportion of total The proportion of patients aged < 65 years for both primary KR rose in both Sweden and Australia from 83.1% and 83.6% and revision KR varied in a narrow range for each registry, to 91.8% and 93.3%, respectively, while the volume of uni- peaking in the years 2008–2012 and decreasing again in all compartmental KR reduced. This contrasts with the KPJRR, instances toward the end of the study period (Table 2). The which had a more constant proportion of total KR remain- proportion of women undergoing primary KR decreased in all ing around 96% for the entire period. In all 3 registries, the countries over this 15-year period. The proportion of women proportion of patellofemoral KR remained low (less than undergoing revision KR was lower than in primary KR in all 1%). Over the study period, revision KR procedure volume countries and showed little change with time (Table 2). Annual volume of primary KR

Annual volume of revision KR

70,000

7,000

Australia KPJRR Sweden

Acta Orthopaedica 2020; 91 (4): 414–419

417

Incidence by age of primary KR/105

Incidence by age of revision KR/105

1,200

120

1,000

Australia <65 KPJRR <65 Sweden <65

Australia ≥65 KPJRR ≥65 Sweden ≥65

100

800

600

400

200

Australia < 65 KPJRR < 65 Sweden < 65

Mean 5-yearly procedure volume change (%) 25

Australia ≥ 65 KPJRR ≥ 65 Sweden ≥ 65

2003–2007 2008-2012 2013–2017

0 2003

0 2017

Primary Revision Primary Revision Primary Revision KR Sweden KR Australia KR KPJRR

Figure 2. Yearly incidence of primary KR (left panel) and revision KR (right panel) by patient age < 65 and ≥ 65 years per 105 population recorded by the SKAR, AOANJRR, and KPJRR from 2003 to 2017.

Figure 4. Mean 5-yearly percentage increases in procedure volume in SKAR, AOANJRR, and KPJRR.

2005

2007

2009

2011

2013

2015

2017

2005

2007

2009

2011

2013

2015

Table 2. Yearly mean ages, percentage women, and proportion age < 65 years for primary and revision KR by registry Factor

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Sweden Primary KR mean age 69.6 69.5 69.6 69.3 69.3 68.9 69.0 68.8 68.7 68.6 68.6 68.8 68.6 68.9 68.9 women (%) 61.0 61.9 60.0 59.5 60.0 59.3 58.3 58.4 58.2 57.7 57.0 57.2 57.0 56.0 55.7 age < 65 (%) 31.3 31.1 31.0 33.1 32.9 34.3 34.4 34.1 34.7 34.7 33.9 32.9 33.9 33.7 33.2 Revision KR mean age 70.7 70.5 70.5 69.9 69.9 69.8 69.2 68.7 68.5 69.4 68.8 68.3 68.5 69.4 69.9 women (%) 58.2 61.0 59.2 61.5 57.5 54.0 61.1 56.6 56.1 55.5 55.5 51.9 54.2 54.6 53.2 age < 65 (%) 29.0 28.9 30.9 35.1 34.2 33.7 33.6 36.7 36.1 33.3 33.5 35.9 34.6 30.9 27.9 Australia Primary KR mean age 68.7 68.7 68.8 68.6 68.4 68.2 68.1 68.0 67.9 68.1 68.0 68.0 67.9 67.8 67.9 women (%) 56.3 56.8 57.2 56.6 56.9 56.5 56.5 56.3 55.9 56.1 56.4 55.7 55.4 55.5 54.8 age < 65 (%) 32.2 32.4 32.7 33.5 34.2 35.6 35.9 36.2 36.7 35.2 35.6 34.8 35.1 35.4 34.5 Revision KR mean age 69.9 68.8 69.5 69.0 69.1 68.7 68.9 68.4 68.6 68.5 68.7 68.6 68.2 68.9 68.9 women (%) 51.6 51.8 50.5 51.5 52.7 51.9 50.8 52.2 50.7 52.3 51.2 52.4 51.3 50.0 51.1 age < 65 (%) 28.7 32.5 29.7 32.4 33.0 34.4 33.9 35.5 33.8 34.2 32.8 34.6 31.3 31.3 30.3 Kaiser Permanente Primary KR mean age 67.8 67.8 67.8 67.7 67.1 67.1 67.0 67.3 67.3 67.3 67.3 67.3 67.3 67.4 67.5 women (%) 64.4 63.0 64.6 63.4 62.4 61.2 61.7 60.9 61.1 60.6 61.0 60.2 61.5 60.9 60.3 age < 65 (%) 35.8 37.3 36.8 37.2 40.5 41.5 41.3 39.9 40.7 38.8 38.4 38.4 38.0 36.9 36.9 Revision KR mean age 68.0 68.0 67.6 66.2 67.5 66.7 66.4 67.3 66.7 66.7 67.1 67.2 66.9 67.5 67.7 women (%) 48.9 54.5 53.3 55.5 53.9 52.8 58.9 55.5 53.0 57.9 53.4 56.9 56.9 53.8 54.8 age < 65 (%) 35.4 34.7 37.9 41.1 39.9 43.9 44.1 43.4 42.1 43.2 41.4 39.7 41.8 39.4 35.7

Discussion Through the last 15 years, primary and revision KR have increased in all 3 countries studied. Suggested reasons for this widespread change are the increase in the prevalence of knee OA, or increased recognition of the utility of KR by surgeons and the community (Weinstein et al. 2013, Hamilton et al. 2015). The growth in KR in the KPJRR was greater than that seen in the other 2 registries with no clear reason for this difference. This may indicate a previously unmet demand is

being filled in this population or be due to local market conditions in the USA. A previous population predictive study has suggested that the rising rate of KR is linked to increasing community obesity (Culliford et al. 2015). As population profiles may vary both between countries and over time, perhaps a better measure for comparison is incidence/105 population. Australia has a higher incidence of both primary and revision KR/105 but the incidence in the KPJRR is approaching that of Australia in primary KR. Incidence changes show a parallel increase in primary and revision KR in the KPJRR, while revision incidence growth in

418

both Sweden and Australia has been proportionately less than that of primaries. Incidence increases in the KPJRR cohort were less than the changes in procedure numbers, indicating a larger rise in the population with this insurance. There has been little change in the mean age of patients receiving either a primary or revision KR in all countries, and little variation in the proportions of KR for patients aged < 65 years. Increases in the use of KR in younger patients are therefore balanced by a comparable increase in the ≥ 65 years age group. This counters the suggestion that KR has been proportionally more frequently used for younger patients over this time (Karas et al. 2019). While the proportion of younger to older KR patients remained stable, the percentage increases in incidence/105 in the younger group were greater, a finding consistent with others (Weinstein et al. 2013, Pabinger et al. 2015). In all 3 registries over the study period, there is an increase in the proportion of males receiving a primary KR, and as there are proportionately more males requiring revision (Table 2) this trend may increase future revision rates. Variation in rates of KR among countries may be due to local economic concerns and health policy, differences in rates of OA, availability of pre-surgical treatments for OA, and access to KR, as well as surgeon availability and variation in thresholds for suitability for operative treatment. The higher incidence of revision KR in Australia compared with the other countries may simply mirror the higher incidence of primary surgery or be due to differences in surgical practice (such as the proportional use of patella resurfacing or cementless fixation) but could also be related to less restricted prosthesis choice in this country. Part of the reason for a smaller rate of increase in revision KR when compared with primary KR in Sweden and Australia may be due to the decrease in proportion of unicompartmental KR in these countries, as partial KR has more than 2.5 times the rate of revision of total KR at 10 years (AOANJRR 2018). This change may also reflect improved prosthesis performance during this time span, related to factors such as the introduction of more component sizing options or highly cross-linked polyethylene (de Steiger et al. 2015, Turnbull et al. 2016). Alternatively, the relative slowing of revision compared with primary KR may be due to the presence of a time lag between increased numbers of primaries and when they will require revision. When analysis was carried out by 5-year time periods the increase in both primary and revision surgery decelerated in all countries over the duration of this study, with the only exception being the increase in revision KR in Sweden in 2008– 2012 when compared with the earlier period. From our findings, we contend that previous studies predicting an exponential increase in primary and revision KR are incorrect and that a universal deceleration of the growth in primary KR has been experienced, with an even greater slowing in growth of revision KR being evident (Kurtz et al. 2007, Kumar et al. 2015, Patel et al. 2015). However, there is quite a large variation

Acta Orthopaedica 2020; 91 (4): 414–419

between countries, with the KPJRR cohort showing the greatest percentage increase in both of these procedure types, while Australia and Sweden have a lower growth rate and increasing disparity between primary and revision rates with time. The variations between countries seen in this study over this time period show that predictive models of future demands for 1 region may not translate to others. Our findings also imply that more conservative future estimates would potentially be more accurate, as suggested by Inacio et al. (2017). While there has been a slowing of the increase in the rate of KR in all 3 countries, our findings may not be generalizable to other countries, where different health systems are in place. A limitation of our study is its retrospective nature, which may have little bearing on future trends. In addition, the study is a simple overview of population changes with time, which can be influenced by many factors, and little or no detail as to the reasons for changes is revealed. This area could be the subject of further analysis. Caution should be used in extrapolating the findings of the cohort from the KPJRR as these may not be representative of the changes found elsewhere in the United States. In addition, revision incidence would be overestimated as it has been calculated irrespective of multiple surgeries for the same patient or knee. These methodological limitations are expected to affect each registry similarly and be consistent throughout the study period. There may also be other unknown influences, such as the introduction of new technologies or changing health policies and economics, which can affect each country differently, and these have not been examined in this study. Conclusion While there has been an increase in both primary and revision KR across all 3 countries during the past 15 years, the rate of increase has slowed. While there are regional differences in KR incidence, and also differences in rates of change, the rate of increase, particularly in Sweden and Australia, does not seem to be as great as previously predicted. Additionally, the rate of increase in revision KR in these 2 countries is less than the increase in primary KR.

PL: conception of study, statistical analysis, interpretation of data, and manuscript preparation; AWD, MS, OR, EP, HP, and SG: interpretation of data and manuscript preparation. Acta thanks David F Hamilton and Christof Pabinger for help with peer review of this study.

AOANJRR. Hip, knee and shoulder arthroplasty: 2018 Annual Report. Adelaide: AOA; 2018. p 1-444. Retrieved from https://aoanjrr.sahmri.com/ documents/10180/576950/Hip%2C%20Knee%20%26%20Shoulder%20 Arthroplasty. Accessed September 18, 2019. Culliford D, Maskell J, Judge A, Cooper C, Prieto-Alhambra D, Arden N K. 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. doi: 10.1016/j.joca.2014.12.022.

Acta Orthopaedica 2020; 91 (4): 414–419

de Steiger R N, Muratoglu O, Lorimer M, Cuthbert A R, Graves S E. Lower prosthesis-specific 10-year revision rate with crosslinked than with noncrosslinked polyethylene in primary total knee arthroplasty: 386,104 procedures from the Australian Orthopaedic Association National Joint Replacement Registry. Acta Orthop 2015; 86(6): 721-7. doi: 10.3109/17453674.2015.1065046. Hamilton D, Howie C, Burnett R, Simpson A, Patton J. Dealing with the predicted increase in demand for revision total knee arthroplasty: challenges, risks and opportunities. Bone Joint J 2015; 97(6): 723-8. doi: 10.1302/0301-620X.97B6.35185. Inacio M, Paxton E, Graves S, Namba R, Nemes S. Projected increase in total knee arthroplasty in the United States: an alternative projection model. Osteoarthritis Cartilage 2017; 25(11): 1797-803. doi: 10.1016/j. joca.2017.07.022. Karas V, Calkins T E, Bryan A J, Culvern C, Nam D, Berger R A, Rosenberg A G, Della Valle C J. Total knee arthroplasty in patients less than 50 years of age: results at a mean of 13 years. J Arthroplasty 2019; 34(10): 2392-7. doi: 10.1016/j.arth.2019.05.018. Koh C K, Zeng I, Ravi S, Zhu M, Vince K G, Young S W. Periprosthetic joint infection is the main cause of failure for modern knee arthroplasty: an analysis of 11,134 knees. Clin Orth Relat Res 2017; 475(9): 2194-201. doi: 10.1007/s11999-017-5396-4. Kumar A, Tsai W-C, Tan T-S, Kung P-T, Chiu L-T, Ku M-C. Temporal trends in primary and revision total knee and hip replacement in Taiwan. J Chin Med Assoc 2015; 78(9): 538-44. doi: 10.1016/j.jcma.2015.06.005. Kurtz S, Ong K, Lau E, Mowat F, Halpern M. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am 2007; 89(4): 780-5. doi: 10.2106/JBJS.F.00222. Kurtz S M, Ong K L, Lau E, Widmer M, Maravic M, Gómez-Barrena E, de Pina M d F, Manno V, Torre M, Walter W L. International survey of primary and revision total knee replacement. Int Orthop 2011; 35(12): 1783-9. doi: 10.1007/s00264-011-1235-5. Leyland K M, Judge A, Javaid M K, Diez-Perez A, Carr A, Cooper C, Arden N K, Prieto-Alhambra D. Obesity and the relative risk of knee replacement surgery in patients with knee osteoarthritis: a prospective cohort study. Arthritis Rheum 2016; 68(4): 817-25. doi: 10.1002/art.39486. NJR. 15th Annual Report 2018. Hertfordshire, UK: National Joint Registry for England, Wales, Northern Ireland and the Isle of Man; 2018.

419

p 1-220. Retrieved from https://www.hqip.org.uk/resource/nationaljoint-registry-15th- annual-report-2018/#.XY2f90YzaUk. Accessed September 18, 2019. Pabinger C, Lothaller H, Geissler A. Utilization rates of knee-arthroplasty in OECD countries. Osteoarthritis Cartilage 2015; 23(10): 1664-73. doi: 10.1016/j.joca.2015.05.008. Patel A, Pavlou G, Mújica-Mota R, Toms A. The epidemiology of revision total knee and hip arthroplasty in England and Wales: a comparative analysis with projections for the United States. A study using the National Joint Registry dataset. Bone Joint J 2015; 97(8): 1076-81. doi: 10.1302/0301620X.97B8.35170. Paxton E W, Inacio M C, Khatod M, Yue E J, Namba R S. Kaiser Permanente national total joint replacement registry: aligning operations with information technology. Clin Orth Relat Res 2010; 468(10): 2646-63. doi: 10.1007/ s11999-010-1463-9. Pitta M, Esposito C I, Li Z, Lee Y-y, Wright T M, Padgett D E. Failure after modern total knee arthroplasty: a prospective study of 18,065 knees. J Arthroplasty 2018; 33(2): 407-14. doi: 10.1016/j.arth.2017.09.041. Postler A, Lützner C, Beyer F, Tille E, Lützner J. Analysis of total knee arthroplasty revision causes. BMC Musculoskelet Disord 2018; 19(1): 55. doi: 10.1186/s12891-018-1977-y. Robertsson O, Ranstam J, Sundberg M, W-Dahl A, Lidgren L. The Swedish Knee Arthroplasty Register: a review. Bone Joint J 2014; 3(7): 217-22. doi: 10.1302/2046-3758.37.2000289. Sharkey P F, Lichstein P M, Shen C, Tokarski A T, Parvizi J. Why are total 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. SKAR. Swedish Knee Arthroplasty Register annual report 2018. Lund, Sweden: Lund University; 2018. p 1-104. Retrieved from http://www. myknee.se/pdf/SVK_2018_Eng_1.0.pdf. Accessed September 18, 2019. Turnbull N, Berend K, Ng V, Adams J, Crawford D, Lombardi J A. Increased femoral component size options improves manipulation rate in females and Knee Society clinical scores in males. Surg Technol Int 2016; 29:279-86. Weinstein A M, Rome B N, Reichmann W M, Collins J E, Burbine S A, Thornhill T S, Wright J, Katz J N, Losina E. Estimating the burden of total knee replacement in the United States. J Bone Joint Surg Am 2013; 95(5): 385. doi: 10.2106/JBJS.L.00206.

420

Acta Orthopaedica 2020; 91 (4): 420–425

A matched comparison of revision rates of cemented Oxford Unicompartmental Knee Replacements with Single and Twin Peg femoral components, based on data from the National Joint Registry for England, Wales, Northern Ireland and the Isle of Man Hasan R MOHAMMAD 1,2, Gulraj S MATHARU 1,2, Andrew JUDGE 1,2, and David W MURRAY 1 1 Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences 2 Musculoskeletal Research Unit, Bristol Medical School, University of Bristol, Bristol,

Correspondence: hasanmohammad@doctors.org.uk Submitted 2019-12-02. Accepted 2020-03-09.

Background and purpose — Registries report high revision rates after unicompartmental knee replacement (UKR) due, in part, to aseptic loosing. In an attempt to improve Oxford UKR femoral component fixation a new design was introduced with a Twin rather than a Single peg. We used the National Joint Registry (NJR) to compare the 5-year outcomes of the Single and Twin Peg cemented Oxford UKRs. Patients and methods — We performed a retrospective observational study using NJR data on propensity score matched Single and Twin Peg UKRs (matched for patient, implant and surgical factors). Data on 2,834 Single Peg and 2,834 Twin Peg were analyzed. Cumulative implant survival was calculated using the Kaplan–Meier method and comparisons between groups performed using Cox regression models. Results — In the matched cohort, the mean follow up for both Single and Twin Peg UKRs was 3.3 (SD 2) and 3.4 years (SD 2) respectively. The 5-year cumulative implant survival rates for Single Peg and Twin Peg were 94.8% (95% CI 93.6–95.8) and 96.2% (CI 95.1–97.1) respectively. Implant revision rates were statistically significantly lower in the Twin Peg (hazard ratio [HR)] = 0.74; p = 0.04). The revision rate for femoral component aseptic loosening decreased significantly (p = 0.03) from 0.4% (n = 11) with the Single Peg to 0.1% (n = 3) with the Twin Peg. The revision rate for pain decreased significantly (p = 0.01) from 0.8% (n = 23) with the Single Peg to 0.3% (n = 9) with the Twin Peg. No other reasons for revision had significant differences in revision rates. Interpretation — The revision rate for the cemented Twin Peg Oxford UKR was 26% less than the Single Peg Oxford UKR. This was mainly because the revision rates for femoral loosening and pain more than halved. This suggests that the Twin Peg component should be used in preference to the Single Peg design.

University of Oxford, Nuffield Orthopaedic Centre, Oxford, UK; UK

Unicompartmental knee replacement (UKR) for anteromedial knee arthritis has many advantages over total knee replacement (TKR) (Wilson et al. 2019). However, national registries suggest that UKR revision rates are several times higher than TKR, with aseptic loosening a leading cause (New Zealand Joint Registry 2016, National Joint Registry 2018). The most commonly used UKR is the Phase 3 Oxford UKR. The initial Phase 3 femoral component, like its predecessors, was spherical and cemented. It had a single peg that was thought to be helpful as it allowed the component to seat optimally (Figure 1). However, as 25–50% of aseptic loosening was femoral (Mohammad et al. 2018), it was felt that the introduction of a Twin Peg component that might improve fixation would be advantageous. During surgery a small hole is made in the femur anterior to the main peg hole to stabilize the femoral saw guide. A second peg, which would fit in the small hole, was therefore added, allowing the new component to be used with standard instrumentation. In order to support the peg, the spherical part of the component was extended about 15° further anteriorly. To accommodate this extension more bone is removed anteriorly, which decreases the risk of the bearing impinging This also allows the femoral component to be implanted in increased flexion.

Figure 1. The Oxford UKR with Single and Twin Peg femoral component.

Acta Orthopaedica 2020; 91 (4): 420–425

All primary Oxford UKRs implanted January 1st, 2009 to December 31st, 2017 n = 41,593 Excluded (n = 20,901): – complex primary surgery, 3 – lateral UKRs, 1,825 – hybrid UKRs, 516 – cementless UKRs, 12,715 – missing/inconsistent component details, symmetric bearings, old tibial sizes, and titanium niobium nitride implants, 5,842 UKRs available for matching n = 20,692 Number of matched UKRs n = 5,668

Figure 2. Flowchart of study selection criteria.

The Twin Peg cemented femoral component was introduced in 2003 but has only been widely used since 2009 (Figure 1) (White et al. 2015). A cementless version of the Twin Peg component was also introduced at a similar time. The Twin Peg cemented component is used with the same cemented tibial component and polyethylene bearing as the Single Peg component. We are not aware of any direct comparative clinical studies of Single Peg and Twin Peg cemented femoral components. The National Joint Registry for England, Wales, Northern Ireland and Isle of Man (NJR) is the largest replacement registry (National Joint Registry 2018). We used NJR data to compare the revision rate and mechanisms of implant failure, in particular femoral component aseptic loosening, following cemented medial Oxford UKRs using Single and Twin Peg femoral components.

Patients and methods We performed a retrospective observational study using the NJR database after NJR Research Sub-Committee approval (National Joint Registry 2018). Data collected by the NJR includes patient, implant, and surgical information. The database has excellent linkability to subsequent revision surgery and is also linked to the Office of National Statistics, which provides mortality data. Anonymized patient data were extracted from the NJR, which included all primary Oxford UKRs implanted between January 1, 2009 and December 31, 2017 (n = 41,593). After data cleaning there were 20,692 medial cemented Oxford UKRs (17,855 Single Peg and 2,837 Twin Peg) eligible for study inclusion (Figure 2). The study exposure was the peg design (Single vs. Twin Peg). Given the potential for factors other than peg design to affect the revision rate (Prempeh and Cherry 2008, Selby et

421

al. 2012, Judge et al. 2013, Elmallah et al. 2015, Lim et al. 2015, Bayliss et al. 2017, Hosaka et al. 2017, Picard et al. 2018, Mohammad et al. 2020b) we a priori matched the Single and Twin Peg UKRs for multiple known confounders using propensity scores. Logistic regression was used to generate a propensity score representing the probability that a patient received a Twin Peg UKR. These propensity scores were generated from patient demographics, surgical factors, and implant factors. Specifically, factors used for matching were: age, sex, primary diagnosis, unilateral/bilateral UKRs, ASA grade, chemical thromboprophylaxis, mechanical thromboprophylaxis, year of surgery, operating surgeon grade, surgeon caseload, surgical approach, operating technique and implant component sizes (Table 1). Surgeon caseload was defined as the average number of UKRs done per calendar year by the operating surgeon and stratified into low (< 10 cases/year), medium (10 to < 30 cases/year) and high volume (≥ 30 cases/ year) as described previously (Mohammad et al. 2020a). BMI was not used for matching given it had a significant proportion of missing data. However, our data demonstrate that BMI was well balanced between groups and our approach is similar to previous studies using NJR data (Matharu et al. 2017, Mohammad et al. 2020a). We matched using a 1:1 ratio on the logit of the propensity score with a 0.02-SD calliper width. We used greedy matching without replacement, which has superior performance for estimating treatment effects (Austin 2009). Standardized mean differences (SMDs) were examined both before and after matching to assess for any covariate imbalance between the different peg design groups, with SMDs of 10% or more considered suggestive of imbalance (Austin 2009). After matching, 5,668 UKRs (2,834 Single Peg and 2,834 Twin Peg UKRs) were included for analysis. Statistics Outcomes of interest were: (1) implant survival and (2) indications for revision surgery, particularly femoral component aseptic loosening. Cumulative survival was determined using the Kaplan–Meier method. The endpoint for implant survival was revision surgery (any addition, removal, or exchange of implant component). Implant survival was compared between the Single and Twin Peg groups, using Cox regression models, with the proportional hazards assumptions assessed and satisfied in all analyses. Additionally, to account for clustering within the matched cohort, a robust variance estimator was used in regression models. Univariable and adjusted models were also assessed. The adjusted models included covariates with residual imbalance after matching (SMD of 10% or more) (Austin 2009). A multi-level frailty model was tested in regression models to adjust for patient clustering within surgeons. The proportional chi-squared test with Yate’s correction or 2-sided Fisher’s exact test was used to compare the indications for revision surgery between groups. The latter was used only when either group had an expected frequency of under 5.

422

The NJR database allows for revisions for UKRs with any aseptic loosening to be analyzed and also for aseptic loosening by each component involved (e.g., femoral or tibial component). The primary analysis was of revision for aseptic femoral loosening. Aseptic tibial loosening and overall loosening rates were also analyzed as there were some cases of combined tibial and femoral loosening. All statistical analyses were performed using Stata (Version 15.1; StataCorp, College Station, TX, USA) except propensity score matching, which was performed using R (Version 3.4.0; R Foundation for Statistical Computing, Vienna, Austria). P-values of < 0.05 were considered significant, with 95% confidence intervals (CI) presented. Ethics, funding, and potential conflicts of interest This study was based entirely on existing patient records acquired during routine clinical care and thus did not require ethical approval. This project was fully approved by the NJR Research Sub Committee. Zimmer Biomet provided funding for the research but were not involved in the study.

Results The matched cohort included 5,668 Oxford UKRs, with 2,834 Single Peg UKRs and 2,834 Twin Peg UKRs. The mean age at surgery was 65 years (SD 10), with 51% of the cohort being female. The mean BMI was 30 (SD 5) with the primary indication for surgery being osteoarthritis in 99%. Patient, surgical, and implant characteristics became well balanced between the Single Peg and Twin Peg groups after propensity score matching (Table 1). The only covariate with residual imbalance was year of primary surgery, which when adjusted for in the regression models did not change the findings and is presented below.

Acta Orthopaedica 2020; 91 (4): 420–425

Table 1. Patient and surgical factors. Values are number (%) unless otherwise specified Unmatched cohort Matched cohort Single Peg Twin Peg Single Peg Twin Peg Covariate 17,855 (86) 2,837 (14) SMD 2,834 (50) 2,834 (50) SMD Sex 0.06 0.009 Female 8,454 (47) 1,431 (50) 1,441 (51) 1,428 (50) Male 9,401 (53) 1,406 (50) 1,393 (49) 1,406 (50) Age at surgery 0.02 0.01 mean (SD) 65 (10) 65 (10) 65 (10) 65 (10) 0.03 BMI, n 13,159 1,868 0.04 2,165 1,865 mean (SD) 30 (5) 30 (5) 30 (5) 30 (5) Primary diagnosis 0.04 0.02 Primary OA 17,676 (99) 2,797 (99) 2,799 (99) 2,794 (99) Other 179 (1) 40 (1) 35 (1) 40 (1) Bilateral UKRs 469 (3) 44 (2) 0.08 55 (2) 44 (2) 0.03 ASA grade 0.05 0.05 1 3,344 (19) 488 (17) 483 (17) 488 (17) 2 13,024 (73) 2,082 (73) 2,119 (75) 2,079 (73) 3 or above 1,487 (8) 267 (10) 232 (8) 267 (10) VTEP— chemical 0.4 0.06 LMWH (± other) 10,447 (59) 2,020 (71) 1,972 (70) 2,017 (71) Aspirin only 1,426 (8) 208 (7) 240 (8) 208 (7) Other 4,731 (26) 590 (21) 593 (21) 590 (21) None 1,251 (7) 19 (1) 29 (1) 19 (1) VTEP— mechanical 0.08 0.003 Any 17,430 (98) 2,800 (99) 2,798 (99) 2,797 (99) None 425 (2) 37 (1) 36 (1) 37 (1) Year of surgery 0.7 0.3 2009 2,416 (14) 133 (5) 74 (3) 133 (5) 2010 2,396 (13) 175 (6) 85 (3) 175 (6) 2011 2,371 (13) 178 (6) 167 (6) 178 (6) 2012 2,285 (13) 192 (7) 246 (9) 192 (7) 2013 2,296 (13) 276 (10) 366 (13) 276 (10) 2014 2,173 (12) 457 (16) 486 (17) 457 (16) 2015 1,633 (9) 475 (17) 520 (18) 475 (17) 2016 1,275 (7) 440 (15) 492 (17) 440 (15) 2017 1,010 (6) 511 (18) 398 (14) 508 (18) Surgeon grade 0.06 0.02 Consultant 16,462 (92) 2,656 (94) 2,668 (94) 2,653 (94) Other 1,393 (8) 181 (6) 166 (6) 181 (6) Surgeon caseload 0.3 0.04 < 10 cases/year 7,266 (41) 893 (31) 915 (32) 892 (31) 10–29 cases/year 7,731 (43) 1,104 (39) 1,131 (40) 1,103 (39) ≥ 30 cases/year 2,858 (16) 840 (30) 788 (28) 839 (30) Surgical approach 0.09 0.03 Medial parapatellar 16,283 (91) 2,513 (89) 2,536 (89) 2,511 (89) Other 1,572 (9) 324 (11) 298 (11) 323 (11) Minimally invasive surgery 0.2 0.01 No 9,171 (51) 1,691 (60) 1,675 (59) 1,689 (60) Yes 8,684 (49) 1,146 (40) 1,159 (41) 1,145 (40) Femoral component size 0.1 0.02 Extra small 43 (0.2) 6 (0.2) 4 (0.1) 6 (0.2) Small 4,076 (23) 796 (28) 788 (28) 793 (28) Medium 9,536 (53) 1,451 (51) 1,462 (52) 1,451 (51) Large 4,170 (23) 582 (21) 577 (20) 582 (21) Extra large 30 (0.2) 2 (0.1) 3 (0.1) 2 (0.1) Tibial component size 0.09 0.02 AA 83 (0.5) 11 (0.4) 12 (0.4) 11 (0.4) A 2,109 (12) 399 (14) 382 (14) 397 (14) B 4,151 (23) 607 (21) 614 (22) 607 (21) C 5,071 (28) 798 (28) 809 (29) 798 (28) D 4,120 (23) 636 (22) 642 (23) 636 (22) E 1,871 (11) 294 (10) 286 (10) 293 (10) F 450 (3) 92 (3) 89 (3) 92 (3) Bearing size 0.1 0.04 3 3,972 (22) 635 (22) 652 (23) 635 (22) 4 7,201 (40) 1,163 (41) 1,128 (40) 1,162 (41) 5 3,786 (21) 623 (22) 617 (22) 622 (22) 6 1,705 (10) 282 (10) 286 (10) 281 (10) 7 760 (4) 105 (4) 120 (4) 105 (4) 8 282 (2) 16 (0.6) 15 (0.5) 16 (0.6) 9 149 (0.8) 13 (0.5) 16 (0.6) 13 (0.5) Bone graft used 0.03 < 0.001 No 17,810 (100) 2,833 (100) 2,830 (100) 2,830 (100) Yes 45 (0.3) 4 (0.1) 4 (0.1) 4 (0.1) OA = osteoarthritis, SD = standard deviation, SMD = standardized mean difference, UKR = unicompartmental knee replacement, VTEP = venous thromboembolism prophylaxis.

Acta Orthopaedica 2020; 91 (4): 420–425

423

Table 2. Reasons for revision in matched cohort Reasons for revision

Figure 3. Kaplan–Meier implant survival rates for matched Single Peg (n = 2,834) and Twin Peg (n = 2,834) UKR implants up to 5 years.

Single Peg UKRs (n = 2,834) n (%) time (SD) a

Aseptic loosening (any component) Femoral component aseptic loosening c Tibial component aseptic loosening Lysis OA progression Pain c Other Dislocation subluxation revision Instability Component dissociation Malalignment Infection Periprosthetic fracture Wear Stiffness Implant fracture Patellar wear Tibial wear Incorrect sizing Patellar mal-tracking

18 (0.6) 11 (0.4) 10 (0.4) 3 (0.1) 26 (0.9) 23 (0.8) 19 (0.7) 8 (0.3) 6 (0.2) 4 (0.1) 8 (0.3) 7 (0.2) 3 (0.1) 3 (0.1) 2 (0.1) 1 (0) 0 (0) 0 (0) 0 (0) 0 (0)

2.0 (1.1) 1.8 (0.8) 2.1 (1.4) 1.8 (0.6) 2.6 (1.4) 2.1 (1.0) 2.7 (1.9) 1.5 (1.0) 1.9 (1.0) 2.1 (0.4) 1.5 (0.7) 1.2 (1.1) 2.3 (3.0) 3.4 (1.9) 2.2 (0.2) 2.0 N/A N/A N/A N/A

Twin Peg UKRs (n = 2,834) n (%) time (SD) a p-value b 14 (0.5) 3 (0.1) 12 (0.4) 4 (0.1) 25 (0.9) 9 (0.3) 14 (0.5) 4 (0.1) 7 (0.2) 3 (0.1) 2 (0.1) 5 (0.2) 3 (0.1) 4 (0.1) 5 (0.2) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)

2.9 (2.4) 3.3 (2.9) 3.1 (2.6) 3.3 (2.9) 3.1 (2.0) 2.6 (2.1) 2.7 (1.8) 0.9 (0.6) 3.2 (2.9) 2.9 (2.6) 2.1 (2.9) 1.1 (0.8) 1.0 (1.0) 5.7 (2.7) 2.1 (1.6) N/A N/A N/A N/A N/A

0.5 0.03 0.7 1.0 0.9 0.01 0.4 0.3 0.8 1.0 0.1 0.6 1.0 1.0 0.5 N/A – – – –

a b

Mean time in years (standard deviation) to revision indication. Comparisons between the revision indications were conducted using the chi-square test or Fisher’s exact test. The latter was used in cases where the expected frequencies were < 5 in either group. c Revision indications that were statistically significantly different in frequency between the groups. Abbreviations: OA = osteoarthritis, UKR = unicompartmental knee replacement.

In the matched cohort, the mean follow-up for both Single and Twin Peg UKRs was 3.3 (SD 2) and 3.4 years (SD 2) respectively. In total 176 knees underwent revision surgery: 102 (3.6%) Single Peg UKRs and 74 (2.6%) Twin Peg UKRs. The 5-year cumulative all-cause implant survival rates were 94.8% (CI 93.6–95.8) for Single Peg UKRs and 96.2% (CI 95.1–97.1) for Twin Peg UKRs (Figure 3). The difference in cumulative revision rates between Twin Peg and Single Peg UKRs was statistically significant (HR = 0.74, p = 0.04). The most common reasons for revision in the Single Peg group were osteoarthritis progression (n = 26, 0.9%), pain (n = 23, 0.8%), and aseptic loosening (n = 18, 0.6%) (Table 2). In the Twin Peg UKR group the most common reasons for revision were osteoarthritis progression (n = 25, 0.9%), aseptic loosening (n = 14, 0.5%), and pain (n = 9, 0.3%) (see Table 2). There was a statistically significant (p = 0.01) difference in the revision rate for pain between the Single Peg (n = 23, 0.8%) and the Twin Peg (n = 9, 0.3%). The revision rate for femoral component aseptic loosening was significantly lower (p = 0.03) in the Twin Peg group (n = 3, 0.1%) compared with the Single Peg group (n = 11, 0.4%). However, there was no statistically significant change in the revision rate for aseptic loosening overall (Twin Peg n = 14, 0.5%; Single Peg n = 18, 0.6%; p = 0.5) or tibial component loosening (Twin Peg n = 12, 0.4%; Single Peg n = 10, 0.4%; p = 0.7).

Discussion This is the first formal comparative observational clinical study of Single and Twin Peg cemented medial Oxford UKR femoral components. We found that the 5-year survival improved from 94.8% with the Single Peg to 96.2% with the Twin Peg and the overall revision rate decreased by 26% (p = 0.04). The main reason for this was that the revision rate for femoral component aseptic loosening (0.4% to 0.1%) and pain (0.8% to 0.3%) more than halved with the Twin Peg component. The Twin Peg was not associated with a significant increase in revision rate for any reason. This suggests that the Twin Peg femoral component is a safer component than the Single Peg. The Twin Peg femoral component was introduced primarily to decrease the rate of femoral loosening. Our study has shown that it has reduced the rate of femoral component loosening from 11/2,834 with the Single Peg to 3/2,834 with the Twin Peg. As the incidence of loosening is low there is some uncertainly about the magnitude of the decrease but it is approximately three-quarters and it is certainly not increased. This suggests that it has improved fixation and has achieved its design aim. The Twin Peg component was also associated with a halving in the revision rate for pain. There are various possible reasons for this. Surgeons are able to record more than 1 reason for revision, so some revisions for pain may have been cases of

424

painful aseptic loosening where surgeons recorded both femoral loosening and pain. As femoral loosening is usually obvious at revision it is possible, but unlikely, that cases of painful early femoral loosening were just recorded as pain. With the Twin Peg there is likely to be a reduction in the incidence of anterior bearing impingement on bone, which is a potent cause of pain. The distal femur is prepared with a mill and this removes just enough bone to accommodate the Single Peg femoral component. Additional bone has to be removed in front of the milled surface to accommodate the anterior part of the bearing in full extension. We suspect that surgeons occasionally forget to remove this anterior bone in the cases where there is a fixed flexion deformity due to posterior capsule shortening, so anterior bearing impingement tends not to occur and therefore cannot be seen. Postoperatively the fixed flexion deformity steadily corrects, and impingement and pain develop. The Twin Peg component has an anterior extension to support the additional peg, which therefore cannot be inserted if the anterior bone is not removed. As a result, with the Twin Peg component surgeons cannot forget to remove the anterior bone and pain due to anterior impingement is therefore less likely. There was no statistically significant difference in the revision rate between the Single and Twin Peg groups for any reason other than pain and femoral loosening. In particular there was no difference in the other common reasons for revision, arthritis progression and tibial loosening. More importantly there was no reason for revision that increased significantly. This suggests that the use of the 2-peg component has no downside, which is perhaps not surprising as the same operative technique, instrumentation, tibial component, and bearing are used with both Single and Twin Peg femoral components. Before our study there were no clinical studies comparing Single Peg and Twin Peg UKR designs, with the only direct comparative studies in the literature being cadaveric (Reiner et al. 2014, 2018). Reiner et al. (2018) found that that the pull-out force from cadaveric bone, as a surrogate for fixation, was substantially higher for the Twin Peg design when compared with the Single Peg design. In another study Reiner et al. (2014) observed a trend towards less subsidence in the Twin Peg design in cadaveric bone, although this did not reach statistical significance. White et al. (2015) reported the 5-year implant survival of Twin Peg UKRs as 98% but did not have a Single Peg UKR comparative arm and therefore compared their results with other Single Peg cohorts (Luscombe et al. 2007, Pandit et al. 2011) and found no differences in implant survival or in patient-reported outcome measures between the different peg groups. However, the study (White et al. 2015) was limited by a small sample size of 249 patients with only 5 revisions, thus not allowing detailed analysis of mechanisms of failure. The main limitation of this study is the short follow up of the Twin Peg component. Additionally, the work is based on Registry data and therefore the only outcome measure is revision. However, studies of Single (Luscombe et al. 2007,

Acta Orthopaedica 2020; 91 (4): 420–425

Pandit et al. 2011) and Twin Peg (White et al. 2015, Lum et al. 2016) cohorts appear to report equivalent functional outcomes. Additionally, revision reasons in the NJR are those recorded at the time of surgery even if these subsequently change. Registries can underreport revisions although there is no reason to believe this would differ between the groups. Another limitation is that, despite matching, there is potential residual confounding and matching can reduce generalizability. However, virtually all Twin Peg cases were matched, which improves generalizability. Following matching the only variable with appreciable imbalance was the year of primary surgery, which is important as technique and instrumentation improved with time. However, there were no differences in our findings when we adjusted the regression models for year of primary surgery. There was a substantial proportion of BMI data missing so we did not match on BMI. However, the BMI was well balanced between groups. In summary, this propensity-matched registry-based study found the risk of revision of the cemented Oxford UKR was 26% less with the Twin Peg femoral component compared with the Single Peg. This was primarily because the revision rates for femoral component loosening and pain more than halved. The Twin Peg was not associated with a significant increase in revision rate for any reason. This suggests that the cemented Twin Peg femoral component should be used instead of the Single Peg design.

The authors would like to thank the patients and staff of all the hospitals in England, Wales, Northern Ireland, and Isle of Man who have contributed data to the NJR. They are grateful to the Healthcare Quality Improvement Partnership, the NJR Research Sub-Committee, and staff at the NJR Centre for facilitating this work. The views expressed represent those of the authors and do not necessarily reflect those of the National Joint Registry Steering Committee or the Healthcare Quality Improvement Partnership who do not vouch for how the information is presented. Additionally, the authors would like to thank the University of Oxford for the Henni Mester Scholarship and the Royal College of Surgeons Research Fellowship who supplied HRM with funding to undertake this research. Andrew Judge was supported by the NIHR Biomedical Research Centre at the University Hospitals Bristol NHS Foundation Trust and the University of Bristol. HRM, GSM, AJ, and DWM designed the study. HRM analyzed the data with statistical support from AJ. HRM, GSM AJ, and DWM helped with data interpretation. HRM wrote the initial manuscript draft which was then revised by all authors. Acta thanks Reinoud W Brouwer and Bernhard Flatøy for help with peer review of this study.

Austin P C. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity–score matched samples. Stat Med 2009; 28(25): 3083-107. 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. 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.

Acta Orthopaedica 2020; 91 (4): 420â&#x20AC;&#x201C;425

Elmallah R D, Cherian J J, Robinson K, Harwin S F, Mont M A. The effect of comorbidities on outcomes following total knee arthroplasty. J Knee Surg 2015; 28(05): 411-16. Hosaka K, Saito S, Oyama T, Fujimaki H, Cho E, Ishigaki K, Tokuhashi Y. Union, knee alignment, and clinical outcomes of patients treated with autologous bone grafting for medial tibial defects in primary total knee arthroplasty. Orthopedics 2017; 40(4): e604-e8. Judge A, Arden N K, Batra R N, Thomas G, Beard D, Javaid M K, Cooper C, Murray D, group EPOS. The association of patient characteristics and surgical variables on symptoms of pain and function over 5 years following primary hip-replacement surgery: a prospective cohort study. BMJ Open 2013; 3(3): e002453. Lim J B T, Chi C H, Lo L E, Lo W T, Chia S-L, Yeo S J, Chin P L, Tay K J D, Lo N N. Gender difference in outcome after total knee replacement. J Orthop Surg 2015; 23(2): 194-7. Lum Z, Lombardi A, Hurst J, Morris M, Adams J, Berend K. Early outcomes of twin-peg mobile-bearing unicompartmental knee arthroplasty compared with primary total knee arthroplasty. Bone Joint J 2016; 98(10 Suppl. B): 28-33. Luscombe K, Lim J, Jones P, White S. Minimally invasive Oxford medial unicompartmental knee arthroplasty. Int Orthop 2007; 31(3): 321-4. Matharu G S, Judge A, Murray D W, Pandit H G. Trabecular metal acetabular components reduce the risk of revision following primary total hip arthroplasty: a propensity score matched study from the National Joint Registry for England and Wales. J Arthroplasty 2017; 33(2): 447-52. Mohammad H R, Strickland L, Hamilton T W, Murray D W. Long-term outcomes of over 8,000 medial Oxford Phase 3 Unicompartmental Kneesâ&#x20AC;&#x201D;a systematic review. Acta Orthop 2018; 89(1): 101-7. Mohammad H R, Matharu G S, Judge A, Murray D W. Comparison of the 10-year outcomes of cemented and cementless unicompartmental knee replacements: data from the National Joint Registry for England, Wales, Northern Ireland and the Isle of Man. Acta Orthop 2020a; 91(1): 76-81. Mohammad H R, Matharu G S, Judge A, Murray D W. The effect of surgeon caseload on the relative revision rate of cemented and cementless unicompartmental knee replacements: an analysis from the National Joint Reg-

425

istry for England, Wales, Northern Ireland and the Isle of Man. J Bone Joint Surg Am 2020b (January 31) [Epub ahead of print]. doi: 10.2106/ jbjs.19.01060.7. National Joint Registry. National Joint Registry 15th Annual Report. National joint registry for England, Wales, Northern Ireland and the Isle of Man; 2018 [accessed January 12, 2019]. New Zealand Joint Registry. Seventeen Year Report January 1999 to December 2015. New Zealand Joint Registry; 2016. Pandit H, Jenkins C, Gill H, Barker K, Dodd C, Murray D. Minimally invasive Oxford phase 3 unicompartmental knee replacement: results of 1000 cases. Bone Joint J 2011; 93(2): 198-204. Picard F, Deakin A, Balasubramanian N, Gregori A. Minimally invasive total knee replacement: techniques and results. Eur J Orthop Traumatol 2018; 28(5): 781-91. Prempeh E, Cherry R. Asa grading vs. mortality in elective orthopaedic procedures. In: Orthopaedic Proceedings. The British Editorial Society of Bone & Joint Surgery; 2008. p 536-. Reiner T, Jaeger S, Schwarze M, Klotz M, Beckmann N, Bitsch R. The stability of the femoral component in the Oxford unicompartmental knee replacement: a comparison of single and twin peg designs. Bone Joint J 2014; 96(7): 896-901. Reiner T, Schwarze M, Panzram B, Klotz M C, Bitsch R G, Jaeger S. The influence of the twin peg design on femoral interface temperature and maximum load to failure in cemented Oxford unicompartmental knee arthroplasty. Clin Biomech 2018; 55: 23-7. Selby R, Borah B J, McDonald H P, Henk H J, Crowther M, Wells P S. Impact of thromboprophylaxis guidelines on clinical outcomes following total hip and total knee replacement. Thromb Res 2012; 130(2): 166-72. White S H, Roberts S, Kuiper J H. The cemented twin-peg Oxford partial knee replacement survivorship: a cohort study. Knee 2015; 22(4): 333-7. Wilson H A, Middleton R, Abram S G, Smith S, Alvand A, Jackson W F, Bottomley N, Hopewell S, Price A J. Patient relevant outcomes of unicompartmental versus total knee replacement: systematic review and metaanalysis. BMJ 2019; 364: 1352.

426

Acta Orthopaedica 2020; 91 (4): 426–432

Higher risk of revision for partial knee replacements in low absolute volume hospitals: data from 18,134 partial knee replacements in the Dutch Arthroplasty Register Iris VAN OOST 1, Koen L M KOENRAADT 1, Liza N VAN STEENBERGEN 2, Stefan B T BOLDER 3, and Rutger C I VAN GEENEN 3 1 Foundation for Orthopedic Research, Care & Education, Amphia 3 Department of Orthopedic Surgery, Amphia Hospital, Breda, the

Correspondence: ivanoost@amphia.nl Submitted 2019-11-25. Accepted 2020-03-16.

Hospital, Breda; 2 Dutch Arthroplasty Register (LROI), ‘s-Hertogenbosch; Netherlands

Background and purpose — Partial knee replacement (PKR) survival rates vary a great deal among registries and cohort studies. These discrepancies can largely be attributed to inappropriate indications of the PKR and low thresholds for revision, but also to the PKR volume. This study used Dutch Arthroplasty Register data to analyze whether absolute PKR or proportional PKR hospital volume is associated with the risk of revision. Patients and methods — 18,134 PKRs were identified in the Dutch Arthroplasty Register from 2007 to 2016. For each year, hospitals were divided into 4 groups based on the quartiles for the absolute volume (< 22, 22–36, 36–58 and > 58 PKRs per year) and the proportional volume (< 8.5, 8.6–14.2, 14.3–25.8 and > 25.8% PKRs). Kaplan–Meier survival analysis was performed to determine survival rates. A multivariable Cox regression adjusted for age category, sex, ASA score, year of surgery, diagnosis, unicondylar side, and type of hospital was used to estimate hazard ratios (HR) for revision. Results and interpretation — Proportional PKR volume did not, but absolute PKR volume did influence the risk of revision. The adjusted HR for hospitals with an absolute volume of 22–36 PKRs per year was 1.04 (95% CI 0.91–1.20), 0.96 (CI 0.83–1.10) for the hospitals with 36–58 PKRs, and 0.74 (CI 0.62–0.89) for hospitals with more than 58 PKRs compared with hospitals that had fewer than 22 PKRs per year. So, patients treated with a PKR in a high absolute volume hospital have a lower risk of revision compared with those treated in a low absolute volume hospital.

One of the advantages of partial knee replacement (PKR) is that the native mechanics of the knee are largely preserved, whereas in total knee replacement (TKR) the anterior cruciate ligament is sacrificed and the mechanics change substantially (Laurencin et al. 1991). This might contribute to better a postoperative clinical outcome following PKR compared with TKR (Liddle et al. 2015). Also, a lower risk of complications has been reported for PKR (Liddle et al. 2014a, Beard et al. 2019). Furthermore, a recent randomized trial demonstrated similar Oxford Knee Scores, a higher perceived knee improvement, higher willingness to undergo the operation again, and better cost-effectiveness after PKR compared with TKR at 5 years’ follow-up (Beard et al. 2019). Nevertheless, since its introduction, the PKR has been a topic of debate due to the diversity in reported long-term survival rates. Multiple studies have reported survival rates over 94% at 10 years (Svärd and Price 2001, Pandit et al. 2011, Lisowski et al. 2011, Burnett et al. 2014). However, registries and low-volume PKR centers showed significantly lower survival rates at 5 and 10 years (Baker et al. 2012, Schroer et al. 2013, Badawy et al. 2017). The discrepancies in survival rates can largely be attributed to low absolute and proportional PKR volume (Liddle et al. 2015, Badawy et al. 2017). Badawy et al. showed that in the Nordic countries the most common annual absolute hospital volume was 1–3 PKRs per year, which could have resulted in the reported low survival rates. The use of PKR is generally accepted in the Netherlands. However, large variation in knee replacement volume exists between hospitals. Data from high-volume and low-volume hospitals are available in the Dutch Arthroplasty Register (LROI)(Van Steenbergen et al. 2015). The LROI database contains data concerning orthopedic joint implants in the Neth-

Acta Orthopaedica 2020; 91 (4): 426–432

erlands since 2007. This study investigates whether absolute PKR hospital volume and proportional PKR hospital volume are associated with higher risk of revision using populationbased national register data.

Patients and methods Study population The Dutch Arthroplasty Register database contains data on orthopedic joint implants in the Netherlands since 2007. The completeness of the LROI database is more than 95% for primary knee arthroplasty with a coverage of all hospitals in the Netherlands (Van Steenbergen et al. 2015). Data were extracted from the Dutch Arthroplasty Register database for all primary TKR and PKR procedures between 2007 and 2016, including bilateral procedures. 18,134 PKRs and 190,204 TKRs were registered in this period. For each patient, data were gathered regarding surgical characteristics (e.g., implantation of TKR or PKR, year of operation, type of hospital (university medical center, general hospital, independent treatment center), and anonymized hospital) and patient characteristics (e.g., age, sex, ASA score, diagnosis, and previous surgery on the affected knee). Data analyses Follow-up was defined as time between primary procedure until revision, death, or end date of follow-up (January 1, 2017). Revision was defined as every change (placement, replacement, or removal) of one or more components of the knee prosthesis. Per year, the absolute and proportional PKR volume was determined for each hospital. These volumes were determined by calculating the PKR volume per year (absolute PKR volume) and TKR volume of each hospital. Subsequently, the percentage of PKRs was calculated for the total number of knee arthroplasties (PKR/(TKR+PKR)*100; proportional PKR volume). The volumes across all years and hospitals were divided into 4 groups based on the quartiles. With respect to the absolute PKR volume, the PKRs were classified into < 22, 22–36, 36–58, or > 58 PKRs per year. For the proportional PKR volume, each PKR was classified into < 8.5, 8.6–14.2, 14.3–25.8, or > 25.8% PKRs per year. The median follow-up of PKRs was 4.0 years (IQR 2.2–6.9). Statistics Kaplan–Meier survival analyses were performed to determine survival rates at 4-year and 8-year follow-up. These survival analyses were performed for both absolute and proportional hospital volumes, with revision for any reason as endpoint. The follow-up started on the day of the primary PKR procedure and ended on the day of first revision, death, or the end of the follow-up period. Differences in patient and procedure characteristics of both absolute and proportional hospital volumes were assessed by Pearson’s chi-square test.

427

Univariable and multivariable Cox regression models were conducted. Spearman’s rho was calculated to estimate the correlation between the absolute and proportional volume of the hospitals. In case of a strong correlation (rho > 0.6) the absolute and proportional hospital volumes are analyzed in separate univariate and multivariable Cox regressions. Cox regression analyses were used to estimate hazard ratios (HRs) with a 95% confidence interval (CI) to investigate the association between absolute and proportional volume and the survival of the PKRs. In the multivariable Cox regression model the HRs were adjusted for age category, sex, ASA score, year of surgery, diagnosis, unicondylar side, and type of hospital. These factors can independently influence the risk of revision following PKR. Residual confounding was investigated with a sensitivity analysis. As a sensitivity analysis, Cox models were performed for every individual variable, for all patient-level variables together and also separately for the hospital level variable, type of hospital. HRs were presented relative to the lowest absolute and proportional volume group. For all covariates added to the model, the proportional hazards assumption was checked by inspecting log-minus-log curved and met. P-values < 0.05 were considered significant. For the 95% confidence intervals (CI) we assumed that the number of observed cases followed a Poisson distribution. The statistical package SPSS (version 25, IBM Corp, Armonk, NY, USA) was used for all statistical analyses.

Ethics, funding, and potential conflicts of interest As the study was based on registry data, ethical approval was not needed. This study received no funding and the authors declare no conflicts of interest regarding this study.

Results 98 hospitals performed 190,204 TKRs and 18,134 PKRs between 2007 and 2016 in the Netherlands. The patient and procedure characteristics for the absolute as well as the proportional volume groups are shown in Tables 1 and 2. The median number of PKRs performed per hospital was 36 per year (IQR = 22–58). With regard to the proportional hospital volume, the median percentage of knee arthroplasties performed with a PKR was 14.2%. Spearman’s correlation test showed a strong correlation between the absolute and proportional hospital volumes (rho = 0.69). With regard to absolute volume, the 4-year survival was 90.9% for the lowest volume group (< 22 PKRs per year), 90.7% for the 22–36 PKRs per year group, 92.4% for the 36–58 PKRs per year group, and 93.5% for the highest absolute PKR volume hospitals (> 58 PKRs per year) (Table 3). The Kaplan–Meier estimated survival at 8 years’ follow-up had dropped to 86.7% for the lowest volume group (< 22 PKRs per year), 85.9% for the 22–36 PKRs per year group, 87.7% for the 36–58 PKRs per

428

Acta Orthopaedica 2020; 91 (4): 426–432

Table 1. Patient and procedure characteristics of 18,134 PKRs according to 4 absolute hospital volume groups from 2007 to 2016. Values are number (%) unless otherwise specified Factor

< 22 n = 4,734

Absolute hospital volume groups 22–36 36–58 > 58 n = 4,559 n = 4,448 n = 4,393

p-value

Age, mean (SD) 61 (9) 62 (9) 63 (9) 64 (9) < 0.001 Age group < 0.001 < 55 1,028 (21.8) 876 (19.2) 781 (17.6) 756 (17.2) 55–64 2,126 (45.0) 1,925 (42.3) 1,841 (41.4) 1,559 (35.5) 65–74 1,270 (26.9) 1,348 (26.9) 1,389 (31.2) 1,532 (34.9) ≥ 75 302 (6.4) 407 (8.9) 434 (9.8) 541 (12.3) Men (41) (42) (41) (41) > 0.05 ASA grade < 0.001 I 1,722 (36.4) 1,551 (34.0) 1,537 (34.6) 1,161 (26.4) II 2,436 (51.5) 2,430 (53.3) 2,532 (56.9) 2,674 (60.9) III–IV 256 (5.4) 261 (5.7) 231 (5.2) 419 (9.5) Diagnosis < 0.001 Osteoarthritis 4,566 (96.5) 4,423 (97.0) 4,353 (97.9) 4,314 (98.2) Other 93 (2.0) 79 (1.7) 63 (1.4) 68 (1.5) Unicondylar side < 0.001 Medial 4,061 (85.8) 4,077 (89.4) 4,038 (90.8) 3,955 (90.0) Lateral 80 (1.7) 111 (2.4) 124 (2.8) 176 (4.0) Year of surgery < 0.001 2007–2010 1,679 (35.4) 1,746 (38.4) 1,352 (30.4) 447 (10.2) 2011–2013 1,586 (33.5) 1,155 (25.3) 1,340 (29.8) 868 (19.8) 2014–2016 1,469 (31.0) 1,658 (36.3) 1,756 (39.5) 3,078 (70.1)

Table 2. Patient and procedure characteristics of the 4 proportional hospital volume groups from 2007 to 2016. Values are number (%) unless otherwise specified Factor

< 8.5% n = 4,535

Proportional hospital volume groups 8.5–14.2% 14.2–25.8% > 25.8% n = 4,576 n = 4,563 n = 4,460

p-values

Age, mean (SD) 61 (9) 63 (9) 63 (9) 63 (9) < 0.001 Age group < 0.001 < 55 1,027 (22.7) 831 (18.2) 772 (16.9) 811 (18.2) 55–64 2,025 (44.7) 1,854 (40.6) 1,870 (41.0) 1,702 (38.2) 65–74 1,202 (26.5) 1,440 (31.5) 1,458 (32.0) 1,439 (32.3) ≥ 75 276 (6.1) 442 (9.7) 462 (10.1) 504 (11.3) Men (42) (39) (42) (41) < 0.01 ASA grade < 0.001 I 1,576 (34.8) 1,428 (31.2) 1,297 (28.4) 1,670 (37.4) II 2,489 (54.9) 2,590 (56.6) 2,573 (56.4) 2,420 (54.3) III–IV 230 (5.1) 302 (6.6) 360 (7.9) 275 (6.2) Diagnosis < 0.05 Osteoarthritis 4,423 (97.5) 4,483 (98.0) 4,487 (98.3) 4,396 (98.6) Other 52 (1.1) 39 (0.9) 28 (0.6) 53 (1.2) Unicondylar side < 0.001 Medial 3,935 (86.8) 4,089 (89.4) 4,087 (89.6) 4,020 (90.1) Lateral 66 (1.5) 48 (1.0) 115 (2.5) 262 (5.9) Year of surgery < 0.001 2007–2010 1,421 (31.3) 1,542 (33.7) 1,224 (26.8) 1,037 (23.2) 2011–2013 1,531 (33.3) 1,111 (24.3) 1,284 (28.1) 1,041 (23.3) 2014–2016 1,601 (35.3) 1,923 (42.0) 2,055 (45.1) 2,382 (53.4)

year group, and 89.3% for the highest absolute PKR volume hospitals (> 58 PKRs per year) (Table 3). The univariate Cox regression showed a statistically significant difference between the highest absolute volume group (> 58 PKRs per year) and the lowest absolute volume group (< 22 PKRs per year) with an HR

of 0.74 (CI 0.63–0.86). Also, in the multivariable Cox regression model, the absolute hospital volume influenced the risk of revision (Table 4). The highest absolute volume group (> 58 PKRs per year) had a lower risk of revision compared with the lowest volume group (< 22 PKRs per year) with an adjusted HR of 0.74 (CI 0.62–0.89) (Figure 1 and Table 4). Other factors that influenced the risk of revision in the multivariable Cox regression for the absolute hospital volume groups were ASA classification, unicondylar side, and age (Table 4). ASA classification III–IV showed inferior results with an adjusted HR of 1.37 (CI 1.09–1.73) compared with ASA classification I. All age groups had a statistically significant reduced risk of revision compared with the lowest age group. The lateral PKRs showed a higher risk of revision compared with the medial PKRs with an HR of 1.32 (CI 1.02–1.71). With regard to proportional PKR volume, the Kaplan–Meier 4-year survival was 91.3% for the lowest volume group (< 8.5% PKRs per year), 91.8% for the 8.5–14.2% PKRs per year group, 91.4% for the 14.2–25.8% PKRs per year group, and 92.7% for the highest proportional PKR volume group (> 25.8% PKRs per year). The Kaplan–Meier estimated survival at 8 years’ follow-up had dropped to 86.6% for the lowest proportional volume group (< 8.5% PKRs per year), 87.2% for the 8.5–14.2% PKRs per year group, 87.4% for the 14.2–25.8% PKRs per year group, and 87.8% for the highest proportional PKR volume hospitals (> 25.8% PKRs per year) (Figure 2 and Table 5). The univariate Cox regression revealed that the proportional hospital volume groups did not influence the risk of revision. Also in the multivariable Cox regression model the proportional hospital volume groups did not influence the risk of revision. Other factors that influenced the risk of revision in the multivariable Cox regression for the proportional hospital volume groups were ASA classification, age, and year of surgery (Table 6).

Discussion In this Dutch arthroplasty register study 98 hospitals performed more than 190,000 TKRs and over 18,000 PKRs with a median follow-up of around 4.0 years for the PKRs. The

Acta Orthopaedica 2020; 91 (4): 426–432

429

Table 3. Results from the 4- and 8-year KM survival analysis on the absolute hospital volume Absolute hospital Revisions volume n (%) < 22 22–36 36–58 > 58

Deaths K–M-4 year K–M 8-year n (%) survival (95% CI) survival (95% CI)

507 (10.7) 150 (3.2) 90.9 (90.1–91.7) 482 (10.6) 144 (3.2) 90.7 (89.7–91.7) 390 (8.8) 169 (3.8) 92.4 (91.6–93.2) 252 (5.7) 97 (2.2) 93.5 (92.5–94.5)

86.7 (85.5–87.9) 85.9 (84.5–87.3) 87.7 (86.3–89.1) 89.3 (87.5–91.1)

Table 4. Multivariable Cox regression results of the absolute hospital volume groups adjusted for age category, sex, ASA, year of surgery, diagnosis, unicondylar side, and type of hospital Factor Absolute hospital volume < 22 22–36 36–58 > 58 Age < 55 55–64 65–74 ≥ 75 Sex Male Female ASA I II III–IV Year of surgery 2007–2010 2011–2013 2014–2016 Diagnosis Osteoarthritis Other Unicondylar side Medial Lateral Unknown Type of hospital General hospital University center Independent center

HR (95%CI)

p-value

1.0 (ref) 1.04 (0.91–1.20) 0.96 (0.83–1.10) 0.74 (0.62–0.89)

0.6 0.5 < 0.005

1.0 (ref) 0.68 (0.60–0.77) 0.53 (0.46–0.61) 0.39 (0.30–0.50)

< 0.001 < 0.001 < 0.001

1.0 (ref) 0.97 (0.87–1.08)

0.6

1.0 (ref) 1.10 (0.98–1.23) 1.37 (1.09–1.73)

0.1 0.01

1.0 (ref) 0.97 (0.88–1.14) 0.93 (0.80–1.08)

0.8 0.3

1.0 (ref) 1.10 (0.69–1.75)

0.7

1.0 (ref) 1.32 (1.02–1.71) 1.36 (0.88–2.10)

0.04 0.2

1.0 (ref) 0.87 (0.60–1.27) 1.10 (0.95–1.28)

0.5 0.2

main finding was that hospitals performing more than 58 PKR procedures per year demonstrated a lower risk of revision compared with hospitals that had less than 22 PKR procedures. In contrast, proportional PKR volume was not associated with the risk of revision. The association between absolute hospital PKR volume and survival found in our study is in line with findings of previous studies. However, these studies found annual hospital volumes of less than 12, 13, and 40 to be associated with a higher risk of revision, compared with 58 in our study (Baker et al. 2013;

Table 5. Results from the 4- and 8-year KM survival analysis on the proportional hospital volume Proportional hospital Revisions Deaths K–M-4 year K–M 8-year volume (%) n (%) n (%) survival (95% CI) survival (95% CI) < 8.5 457 (10.1) 138 (3.0) 8.5–14.2 420 (9.2) 144 (3.1) 14.2–25.8 406 (8.9) 144 (3.2) > 25.8 348 (7.8) 134 (3.0)

91.3 (90.5–92.1) 91.8 (91.0–92.6) 91.4 (90.4–92.4) 92.7 (91.9–93.5)

86.6 (85.2–88.0) 87.2 (85.8–88.6) 87.4 (86.0–88.8) 87.8 (86.4–89.2)

Table 6. Multivariable Cox regression results of the proportional hospital volume groups adjusted for age category, sex, ASA, year of surgery, diagnosis, unicondylar side, and type of hospital Factor

HR (95%CI)

Proportional hospital volume < 8.5% 1.0 (ref) 8.5–14.2% 1.03 (0.89–1.19) 14.2–25.8% 1.06 (0.92–1.24) > 25.8% 0.93 (0.79–1.11) Age < 55 1.0 (ref) 55–64 0.68 (0.60–0.77) 65–74 0.52 (0.45–0.61) ≥ 75 0.38 (0.30–0.49) Sex Male 1.0 (ref) Female 0.97 (0.87–1.08) ASA I 1.0 (ref) II 1.08 (0.96–1.21) III–IV 1.32 (1.04–1.66) Year of surgery 2007–2010 1.0 (ref) 2011–2013 0.97 (0.85–1.11) 2014–2016 0.86 (0.75–1.0) Diagnosis Osteoarthritis 1.0 (ref) Other 1.12 (0.70–1.78) Unicondylar side Medial 1.0 (ref) Lateral 1.31 (1.01–1.70) Unknown 1.36 (0.88–2.09) Type of hospital General hospital 1.0 (ref) University center 0.93 (0.64–1.35) Independent center 1.05 (0.89–1.23)

p-value

0.7 0.4 0.4 < 0.001 < 0.001 < 0.001 0.5 0.2 < 0.05 0.7 < 0.05 0.6 < 0.05 0.2 0.7 0.6

Badawy et al. 2014, 2017). The Cox regression analysis of Badawy et al. showed much lower HRs for the higher volume groups than our study, which indicates a stronger effect of hospital volume on risk of revision in their study (Badawy et al. 2014). In particular, for the group performing over 41 procedures annually the HR was 0.59, whereas in our study a HR of 0.74 was found for the group performing more than 58 procedures annually. This HR was also our only statistically significant HR, whereas in both studies of Badawy et al. all the volume groups had a significant lower HR compared with the

430

Acta Orthopaedica 2020; 91 (4): 426–432

between absolute surgeon PKR volume and survival (Baker et al. 2013, Liddle et al. 2016). < 22 procedures < 8.5% Hence, the possible positive effect of higher 22–35 procedure 8.5–14.1% 36–58 procedures 14.2–25.8% proportional surgeon volume is not reflected > 58 procedures > 25.8%s in higher proportional hospital volume. The 95 95 assumption of the presence of a similar proportional surgeon volume effect compared with Liddle et al. in our study might, however, 90 90 be doubted, especially because the absolute numbers of PKRs in their study per surgeon is very low with on average 2.8 PKRs per knee surgeon (Liddle et al. 2015). If we compare 85 85 0 2 4 6 8 0 2 4 6 8 this with the situation of the country of our Years after index operation Years after index operation study, the average number of PKRs is already Figure 1. Cox regression survival curve for Figure 2. Cox regression survival curve for 3.6 when the number of PKRs is divided by the the absolute hospital volume adjusted for the proportional hospital volume adjusted age category, sex, ASA score, year of sur- for age category, sex, ASA score, year of number of all practicing orthopedic surgeons. gery, diagnosis, unicondylar side, and type surgery, diagnosis, unicondylar side, and This includes also upper extremity surgeons. of hospital. type of hospital. These absolute numbers directly influence the proportional surgeon volumes and therefore lowest volume group (Badawy et al. 2014, 2017). This discrep- we expect the influence of proportional surgeon volume on ancy might be the result of the lower survival rates observed in prosthesis survival to be less obvious in the Netherlands comthe Nordic studies, indicating that the low-volume groups in pared with the UK. However, we were not able to clarify this. our study already perform quite well. In addition, in general Therefore, future studies are needed to see if it is only the prothe PKR volumes were much lower in the studies of Badawy portional surgeon or also the proportional hospital volume that et al., resulting in a lowest volume group with much lower influences the risk of revision and these studies should take PKR volumes compared with our study. Hence, although the the internal hospital referral policy into account. In addition to effect was smaller compared with data from other registries, the finding that absolute hospital volume is related to the risk possibly as a result of more PKR use in the Netherlands and of revision, our study also revealed that younger patients have higher survival rates in general, higher absolute PKR volume a higher risk of revision compared with older patients. Other still results in a lower risk of revision. studies have also shown a higher risk of revision in younger With regard to the proportional PKR volume, our study patients after PKR with similar findings in primary hip and revealed no association with revision rate. The effect of pro- total knee arthroplasty (Kuipers et al. 2010, Badawy et al. portional PKR volume on the risk of revision has been evalu- 2017, Bayliss et al. 2017). Data from other registers (AOANated in 1 previous study (Liddle et al. 2015), which did find JRR 2018, NZJR 2018, SKAR 2018) support these findings. an association between proportional surgeon volume and the As shown in the study of Bayliss et al., the majority of revisurvival of PKRs: the lowest relative risk of revision in PKRs sions occur after a long period of follow-up, suggesting that when 40–60% of the knee arthroplasties were performed with the only reason for an increased risk of revision in younger a PKR. The difference between our study and the study of patients results from the simple reason of longer follow-up in Liddle et al. (2015) might be explained by the fact that they this group. Another reason, possibly more of influence in PKR investigated the proportional volume on the level of the sur- than total knee and hip replacement, might be that PKRs are geon while this study focused on the proportional hospital performed for the wrong indication in the younger patients. volume. Unfortunately, we were not able to extract the data at This hypothesis is supported by the fact that in our study the the level of the surgeon so whether the proportional surgeon age was also significantly lower in the lower hospital volume volume actually influences the survival based on the data in groups and the fact that a study of the designer group did not our study remains unknown. Assuming that a similar propor- revealed an effect of age on the risk of revision (Kennedy et al. tional surgeon volume effect is present in the data of our study, 2018). Hence, both the longer period young patients depend the reason the same effect was not observed for proportional on their prosthesis but also the possibility of low age as an hospital volume might be explained by in-hospital referral of indication for PKR might result in a higher risk of revision in patients eligible for PKR to the more specialized PKR knee younger patients. surgeons. As a result, hospitals with 1 surgeon performing Data gathering and data completeness are necessary for all PKRs can be assigned to the same proportional hospital reliable and meaningful knee arthroplasty research. The volume group as hospitals with every surgeon performing a completeness of the LROI for knee arthroplasty is more than couple of PKRs. This while the situation in the former hospital 95% with a 100% coverage of hospitals (Van Steenbergen et is expected to be superior as a result of the positive association al. 2015). However, a limitation of the LROI is the lack of Revisionfree survival (%)

Revisionfree survival (%)

100

Acta Orthopaedica 2020; 91 (4): 426–432

historical information on individual surgeon volume. This lack of information on surgeon volume made it impossible to investigate the proportional surgeon volume and absolute surgeon volume. This is an issue mainly for the proportional volume question. However, the proportional hospital volume (as used in our study) may have advantages, because the overor underestimation as a result of in-hospital referrals of proportional surgeon volume (used by Liddle et al. 2015) does not play a role and the hospital volume addresses the entire patient population of a hospital and reflects how a hospital has incorporated PKR. Nevertheless, to get a complete overview of the effect of proportional volume on survival we suggest evaluating both surgeon and hospital volume for the evaluation of internal hospital referral policy (Liddle et al. 2015). Another limitation is that residual confounding may persist in our data despite the performed adjustments. We could only adjust for factors present in the LROI database and therefore residual confounding may persist due to unmeasured and also imperfectly measured variables. An example of an unmeasured variable is the operating surgeon grade, which influences the implant survival as found by Liddle et al. (2014b). After the sensitivity analysis we concluded that the multivariable Cox regression adjusted for age category, sex, ASA score, year of surgery, diagnosis, unicondylar side, and type of hospital was the best fitted model with the least residual confounding. Based on the knowledge available now regarding absolute and proportional hospital and surgeon PKR volume, we can state that the longevity of the prosthesis is positively associated with volume. Most studies focused on absolute volume (surgeon and/or hospital), resulting in several different cut-off levels ranging from 11–58 PKR per year for hospital volume and 10–13 PKR per year for surgeon volume with a clear tendency of even better survival rates with increasing volumes. With regard to the proportional volume this has only been proven on the level of the surgeon, with surgeons performing 40–60% of their knee arthroplasties with a PKR showing the best results. This might actually be a derivative of performing PKR for the proper indication, since Willis-Owen et al. (2009) demonstrated 47.6% to be eligible for PKR based on the wear pattern of the knee. Combining these numbers, it seems justified to perform PKR in hospitals with knee arthroplasty populations of 50–100 patients or larger, taking into account that each surgeon should perform at least 10–13 PKRs per year. In conclusion, data from the Dutch Arthroplasty Register confirm that the absolute hospital PKR volume should be high (in this study > 58 PKRs annually) to achieve the lowest risk of revision. Proportional hospital PKR volume did not show an effect on implant survival, indicating that incorporation of PKR in different hospital practices is possible as long as the absolute hospital PKR volume is high enough. The latter can be easily achieved since more than 40% of the arthroplasty population actually seems to be eligible for PKR. In addition, by using in-hospital referral the desired absolute and proportional surgeon levels can also be easily achieved.

431

IvO: study design, data analysis, and drafting of manuscript. KK: study design, data analysis, and revision of manuscript. LvS: data acquisition, statistics, and revision of manuscript. SB: revision of manuscript. RvG: interpretation of data, and revision of manuscript. Acta thanks Andrew Price for help with peer review of this study.

AOANJRR. Australian Orthopaedic Association National Joint Replacement Registry. Annual Report 2018. www.aaoanjrr.sahmri.com/annualreports-2018. Badawy M, Espehaug B, Indrekvam K, Havelin L I, Furnes, O. Higher revision risk for unicompartmental knee arthroplasty in low-volume hospitals. Acta Orthop 2014; 85(4): 342-7. Badawy M, Fenstad A M, Bartz-Johannessen C A, Indrekvam K, Havelin L I, Robertsson O, W-Dahln A, Eskelinen A, Mäkelä K, Pedersen A, Schrøder H M, Furnes O. Hospital volume and the risk of revision in Oxford unicompartmental knee arthroplasty in the Nordic countries: an observational study of 14,496 cases. BMC Musculoskelet Disord 2017; 18: 388. Baker P N, Jameson S S, Deehan D J, Gregg P J, Porter M, Tucker K. Midterm equivalent survival of medial and lateral unicondylar knee replacement: an analysis of data from a national joint registry. J Bone Joint Surg Br 2012; 94(12): 1641-9. Baker P, Jameson S, Critchley R, Reed M, Gregg P, Deehan D. Center and surgeon volume influence the revision rate following unicondylar knee replacement. J Bone Joint Surg Am 2013; 95(8): 702-9. 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: 1424-30. Beard D J, Davies L J, Cook J A, MacLennan G, Price A, Kent S, Hudson J, Carr A, Leal J, Campbell H, Fitzpatrick R, Arden N, Murray D, Campbell M K, for the TOPKAT study group. The clinical and cost-effectiveness of total versus partial knee replacement in patients with medial compartment osteoarthritis (TOPKAT): 5-year outcomes of a randomised controlled trial. Lancet 2019; 394: 746-56. Burnett R S, Nair R, Hall C A, Jacks D A, Pugh L, McAllister M M. Results of the Oxford Phase 3 mobile bearing medial unicompartmental knee arthroplasty from an independent center: 467 knees at a mean 6-year follow-up: analysis of predictors of failure. J Arthroplasty 2014; 29(9): 193-200. Kennedy J A, Matharu G S, Hamilton T W, Mellon S J, Murray D W. Age and outcomes of medial meniscal-bearing unicompartmental knee arthroplasty. J Arthroplasty 2018; 33(10): 3153-9. Kuipers B M, Kollen B J, Kaijser Bots P C, Burger B J, van Raay J J A M, Tulp N J A, Verheyen C C P M. Factors associated with reduced early survival in the Oxford phase III medial unicompartment knee replacement. Knee 2010; 17(1): 48-52. Laurencin C T, Zelicof S B, Scott R D, Ewald F C. Unicompartmental versus total knee arthroplasty in the same patient: a comparative study. Clin Orthop Relat Res 1991; (273): 151-6. Liddle A D, Judge A, Pandit H, Murray D W. Adverse outcomes after total and unicompartmental knee replacement in 101330 matched patients: a study of data from the National Joint Registry for England and Wales. Lancet 2014a; 384: 1437-45. Liddle A D, Judge A, Pandit H, Murray D W. Determinants of revision and functional outcome following unicompartmental knee replacement. Osteoarthritis Cartilage 2014b; 22(9): 1241-50. Liddle A D, Pandit H, Judge A, Murray D W. Optimal usage of unicompartmental knee arthroplasty: a study of 41 986 cases from the National Joint Registry for England and Wales. Bone Joint J 2015; 97-B: 1506-11. Liddle A D, Pandit H, Judge A, Murray D W. Effect of surgical caseload on revision rate following total and unicompartmental knee replacement. J Bone Joint Surg Am 2016; 98(1): 1-8.

432

Lisowski L A, van den Bekerom M P J, Pilot P, van Dijk N N, Lisowski A E. Oxford Phase 3 unicompartmental knee arthroplasty: medium-term results of a minimally invasive surgical procedure. Knee Surg Sports Traumatol Arthrosc 2011; 19(2): 277-84. NZJR. New Zealand Joint Registry. Annual Report 2018. www.nzoa.org.nz/ nz-joint-registry Pandit H, Jenkins C, Gill H S, Barker K F 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: 198-204. Schroer W C, Barnes C L, Diesfeld P, Lemarr A, Ingrassia R, Morton D J, Reedy M. The Oxford unicompartmental knee fails at a high rate in a highvolume knee practice. Clin Orthop Relat Res 2013; 471(11): 35339.

Acta Orthopaedica 2020; 91 (4): 426â&#x20AC;&#x201C;432

SKAR. The Swedish Knee Arthroplasty Register. Annual report 2018. www. myknee.se SvĂ¤rd U C G, Price A J. Oxford medial unicompartmental knee arthroplasty: a survival analysis of an independent series. J Bone Joint Surg Br 2001; 83(2): 191-4. Van Steenbergen L N, Denissen G A W, Spooren A, Van Rooden S M, Van Oosterhout F J, Morrenhof J W, Nelissen R G H H. More than 95% completeness of reported procedures in the population-based Dutch Arthroplasty Register. Acta Orthop 2015; 86(4): 498-505. Willis-Owen C A, Brust K, Alsop H, Miraldo M, Cobb J P. Unicondylar knee arthroplasty in the UK National Health Service: an analysis of candidacy, outcome and cost efficacy. Knee 2009; 16(6): 473-8.

Acta Orthopaedica 2020; 91 (4): 433–438

433

Why are patients still in hospital after fast-track, unilateral unicompartmental knee arthroplasty Christian Bredgaard JENSEN 1,2, Anders TROELSEN 1,2, Christian Skovgaard NIELSEN 1,2, Niels Kristian Stahl OTTE 1,2, Henrik HUSTED 2, and Kirill GROMOV 1,2 1 Department

of Orthopaedic Surgery, Copenhagen University Hospital Hvidovre; 2 Clinical Orthopaedic Research Hvidovre, Copenhagen University Hospital Hvidovre, Denmark Correspondence: chrisbredgaard@hotmail.com Submitted 2020-02-11. Accepted 2020-03-22.

Background and purpose — Previous studies have investigated risk factors related to prolonged length of stay following total knee arthroplasty (TKA), but little is known about specific factors resulting in continued hospitalization within the 1st postoperative days after unicompartmental knee arthroplasty (UKA). We investigated what specific factors prevent patients from being discharged on the day of surgery (DOS) and the first postoperative day (POD-1) following primary UKA in a fast-track setting. Patients and methods — We prospectively collected data on 100 consecutive and unselected medial UKA patients operated from December 2017 to May 2019. All patients were operated in a standardized fast-track setup with functional discharge criteria continuously evaluated from DOS and until discharge. Results — Median length of stay for the entire cohort was 1 day. 22% and 78% of all patients were discharged on DOS and POD-1, respectively. Lack of mobilization and pain separately delayed discharge in respectively 78% and 24% of patients on DOS. The main reasons for lack of mobilization were motor blockade (37%) and logistical factors (26%). For patients placed 1st or 2nd on the operating list, we estimate that the same-day discharge rate would increase to 55% and 40% respectively, assuming that pain and mobilization were successfully managed. Interpretation — One-fifth of unselected UKA patients operated in a standardized fast-track setup were discharged on DOS. Pain and lack of mobilization were the major reasons for continued hospitalization within the initial postoperative 24–48 hours. Strategies aimed at decreasing length of stay after UKA should strive to improve analgesia and postoperative mobilization.

The number of unicompartmental knee arthroplasties (UKAs) performed in patients suffering from osteoarthritis has steadily increased. UKA has the potential benefit of not only improving patient-reported outcomes, but also to reduce morbidity, complications, and cost (Liddle et al. 2014, Beard et al. 2019). In the United Kingdom, 9% of all primary knee arthroplasties performed in 2018 were UKAs while this number is as high as 20% in Denmark (Danish Knee Arthroplasty Register 2019, National Joint Registry for England 2019). UKA is effective and safe when performed in a fast-track setting and outpatient UKA in selected patients has been shown to be feasible and safe (Munk et al. 2012, Cross and Berger 2014, Bovonratwet et al. 2017, Kort et al. 2017). However, the number of patients actually being discharged on DOS that were scheduled for outpatient surgery differs between studies and ranges from 37% to 100% (Gondusky et al. 2014, Bradley et al. 2017, Jenkins et al. 2019, Rytter et al. 2019). Studies have shown an association between increased length of stay (LOS) and an increase in both complication and readmission rates (Otero et al. 2016). In order to reduce LOS and increase patient satisfaction, a focus on successfully managing well-defined discharge criteria in a multimodal approach is imperative (Husted et al. 2008, Cross and Berger 2014). In addition, decreased LOS and outpatient procedures are associated with financial benefits, which have further fueled interest in decreasing LOS and ensuring DOS discharged following UKA (Bradley et al. 2017). Finally, decreased LOS is also shown to increase patient satisfaction levels (Reilly et al. 2005, Richter and Diduch 2017). A study has been conducted to explore reasons for prolonged hospitalization in a fast-track setting following TKA (Husted et al. 2011). However, in spite of a growing number

434

of UKAs performed each year, no study explicitly exploring reasons for prolonged hospitalization beyond DOS following UKA in a fast-track setting has been published at present. Therefore, we investigated reasons for continued hospitalization beyond DOS following UKA in a fast-track setting.

Patients and methods 100 consecutive and unselected patients undergoing primary, unilateral, medial UKA in one institution between December 2017 and May 2019 were enrolled in this study. All patients were operated with cementless mobile bearing implants using microplasty instruments with a minimal invasive technique as described by the manufacturer. Patients were operated on by 4 surgeons, all using UKA in above 20% of all performed knee arthroplasties. Tourniquet was used during the entire surgery, set at 100 mmHg above the individual systolic pressure. Patients were intended to be operated using spinal anesthesia (SA) with 2 mL 0.5% hyperbaric bupivacaine, unless the patient specifically requested general anesthesia (GA) or GA was chosen by the anesthesiologist due to patient characteristics. If GA was chosen, remifentanil and propofol was used. All patients received local infiltration analgesia (LIA) with 200 mL 0.2% ropivacaine injected in the posterior capsule, periarticular tissues, and subcutaneous tissue. Standard 3-layer closure with tissue adhesive was performed and a compression bandage was applied to the limb (Andersen et al. 2008, Gromov et al. 2019). All patients received a singleshot intravenous injection of 125 mg methylprednisolone 30 minutes before the beginning of the surgery together with 2 g dicloxacillin. Pain medication included paracetamol 1 g and celecoxib 200 mg as single doses preoperatively, and paracetamol 1g x 4 and celecoxib 200 mg x 2 daily for 7 days postoperatively. No opioids were given routinely, and morphine 5 mg was given as rescue medication only. Upon completed surgery, patients operated with SA with ASA 1 and 2 were transferred directly to the patient ward, while patients with ASA > 2 and patients operated using GA were transferred to the post-anesthesia care unit (PACU), where they stayed until fulfilling modified Aldrete discharge criteria (Aasvang et al. 2017). Mobilization was attempted upon patients’ return to the ward as soon as motor function allowed. Physiotherapy was focused on reaching functional discharge criteria without any requirements for specific range of motion prior to discharge. Postoperative radiographs were taken on the DOS or the day after, if the patient was not discharged on DOS. All patients were evaluated continuously with regard to functional discharge criteria. For patients who were not discharged, the reason for not fulfilling the discharge criteria was recorded at 8 pm on the day of surgery, and at 2 pm on postoperative day 1 (POD-1) and postoperative day 2 (POD-2). Fulfillment of

Acta Orthopaedica 2020; 91 (4): 433–438

the following criteria was recorded and included: independent mobilization, postoperative nausea and vomiting (PONV), circulatory insufficiency, pain, wound issues, and urinary retention. If patients were not sufficiently mobilized, the following possible underlying reasons were recorded and registered as 1 or more of following: logistics, motor blockade, PONV, pain, muscle weakness, or dizziness. Logistics included lack of postoperative radiographs and limited access to physiotherapy due to organizational factors. Acceptable pain levels were < 5 VAS at rest and < 7 VAS during physical activity. Patients were regarded as circulatory stabile with a pulse < 90 and systolic blood pressure > 100. Wound issues covered bleeding from the surgical wound. Urinary retention was evaluated using a bladder scanner, with 800 mL being the cutoff for catheterization. Ethics, funding, and potential conflicts of interest Approval from the ethical committee was not required since this was purely an observational study. Data access was approved by the national data committee (HVH-2012-048). This research did not receive any financial support. The authors declare no conflicts of interest.

Results Patient demographics are given in Table 1. Among all patients, 22% were discharged on DOS, with 78% discharged on POD-1, and 98% on POD 2. Of the 22 patients (22%) discharged on DOS, 18 patients were ASA 1–2, and the remaining 4 patients were ASA 3–4. Of the 11 patients operated using GA, 3 were discharged on DOS, 5 on POD-1, and 3 on POD-2. When only looking at patients operated as #1 or #2 on the surgery schedule, 27% of patients were discharged on DOS and 80% and 99% of patients discharged on POD-1 and POD-2, respectively. Median LOS for the entire group was 1 day (range 0–3). Lack of mobilization (81%), pain (19%), and urinary retention (18%) were the main reasons for patients not meeting the discharge criteria on DOS (Table 2). When only looking at patients operated as #1 and #2, urinary retention was an issue in 11% of patients, with lack of mobilization (78%) and pain (24%) being the most important reasons for not being discharged. PONV, circulatory insufficiency, and wound issues were present in 1–6% of all patients (Table 2). Several reasons for patients not being sufficiently mobilized were recorded. In patients with mobilization issues still hospitalized on DOS the major reasons were motor blockade (44%), logistics (24%), and pain (19%) (Table 3). 1 patient operated using GA was registered as insufficiently mobilized due to motor blockade. When looking only at patients operated as #1 or #2 on the surgery schedule, the percentage of patients not mobilized sufficiently due to motor blockade decreases to 37%, while the

Acta Orthopaedica 2020; 91 (4): 433–438

435

Table 1. Patient demographics Factor

Sex Female 57 Male 43 Total 100 BMI Age ASA score 1–2 80 3–4 20 Anaesthesia General 11 Spinal 89

mean (range)

30 (21–53) 67 (39–93)

Table 2. Reasons for not being discharged displayed as count and percentage of total amount of patients still hospitalized on day of surgery (DOS), postopertive day 1 (POD-1) and postopertive day 2 (POD-2) Not discharged on DOS POD-1 POD-2 Reasons n = 78 n = 22 n=2 Lack of mobilization 63 (81) PONV 5 (6) Circulatory insufficiency 1 (1) Pain 15 (19) Wound issues 2 (3) Urinary retention 14 (18)

Table 4. Possible improved discharge percentage on the day of surgery if the individual discharge criteria was managed successfully. Not DOS discharge are the number of patients failing to meet only that specific discharge criteria

15 (68) 2 (9) 0 (0) 11 (50) 4 (18) 1 (5)

1 (50 2 (100) 0 (0) 0 (0) 1 (50) 0 (0)

Table 3. Reasons for lack of mobilization displayed as count and percentage of total amount of patients lacking mobilization on day of surgery (DOS), postopertive day 1 (POD-1) and postopertive day 2 (POD-2) Reasons

Not mobilized on DOS POD-1 POD-2 n = 63 n = 15 n=2

Logistics Motor blockade PONV Pain Muscle weakness Dizziness

15 (24) 28 (44) 4 (6) 12 (19) 3 (5) 1 (2)

1 (7) 0 (0) 0 (0) 9 (60) 2 (13) 3 (20)

0 (0) 0 (0) 1 (50) 0 (0) 0 (0) 0 (0)

percentage of patients not mobilized sufficiently due to pain increases to 26%. PONV, muscle weakness, and dizziness were infrequent reasons for lack of mobilization. Assuming the lack of mobilization could be managed successfully in patients operated as #1 or #2 on the surgery schedule, we estimate an increase in discharge percentage up to 55% on DOS. Sorted for the specific issues resulting in lack of mobilization, an additional 11 and 9 patients could possibly be discharged, when assuming successful management of motor blockade and logistics, respectively, and thus potentially increasing DOS discharge to 41% and 39%, respectively. Assuming pain could be managed successfully in patients

Not DOS Possible DOS discharge a discharge a, b n n (%)

Actual DOS discharge a Cause of not DOS discharge Lack of mobilization c PONV Circulatory insufficiency Pain Wound issues Urinary retention

–

20 (27)

21 4 0 10 0 1

41 (55) 24 (32) 20 (27) 30 (40) 20 (27) 21 (28)

a 1st and 2nd patient on program. b Possible DOS discharge is the sum

of Actual DOS discharge and Not DOS discharge c Patients not mobilized due to pain or PONV were excluded from lack of mobilization and included in pain and PONV, respectively.

operated as #1 or #2 on the surgery schedule, we estimate an increase in discharge percentage up to 40% on DOS (Table 4).

Discussion In this prospective single-center study, we investigated specific factors responsible for continued hospitalization following medial UKA in a fast-track setting. Reasons for not being discharged on the DOS were primarily lack of mobilization and pain. Primary reasons for lack of mobilization were motor-blockade, pain and logistics. We found that 22% of unselected patients were discharged on DOS. A study by Jenkins et al. (2019) reported 39% (n = 669) of consecutive unselected unilateral UKA patients to be discharged on DOS. The aim of that study was to investigate the effect of delaying knee flexion on different outcome measures such as LOS. Some important changes in the postoperative protocol in the study were delaying knee flexion and a physiotherapist working late shifts. Also, efforts were made to have UKA patients scheduled for surgery as #1 if possible. Jenkins et al. (2019) reported reduced muscle strength to be the most common reason for continued hospitalization as well as dizziness and nausea. A study by Bradley et al. (2017) reported a day of surgery discharge percentage of 85% (n=72). Patients in that study were included in the DOS discharge group only if they were cleared by a preoperative team after testing and optimizing their coexisting medical conditions. Another study found 85% of patients (n = 20) to be discharged on DOS following UKA, when including patients only if they had no severe cardiologic, pulmonary, internal disease, or fear of an outpatient procedure (Kort et al. 2017). The reason for these exclusion criteria was that such patients might need postoperative adjustment of medication resulting in a delayed

436

discharge. Excluded patients were operated in a rapid-recovery inpatient pathway. A study including 160 outpatient UKA patients reports 100% of patients to be discharged on DOS (Gondusky et al. 2014). All patients had to be cleared by their primary care physician; however, the exact medical reasons for exclusion are not fully described. Following discharge, patients were seen by a physiotherapist at home, starting from POD-1 and then 3 times a week for 3 weeks. We found a substantially lower number of UKA patients discharged on DOS compared with the studies mentioned above. In the case of Jenkins et al. (2019) the main reason for this difference might be that we have introduced no changes to our postoperative protocol in relation to this study, thus not increasing the focus on LOS as measurement of performance within the department. Also, no changes were made regarding available resources such as a physiotherapist working late or UKA patients being scheduled for early surgery slots. The main reason for the difference between our study and the other studies cited above is most likely patient selection, as these other studies apply specific inclusion criteria for outpatient patients, while we investigated all consecutive and unselected patients deemed eligible for UKA. This resulted in a higher mean ASA score in patients compared with the above-mentioned studies, and a very large range in both patient age and BMI. This large range in patient demographics in an unselected cohort was also reported by Jenkins et al. (2019). Furthermore, all patients in our study were discharged to their own homes with no additional care. The difference in the amount of postoperative care needed in unselected and selected patient groups could explain the difference in length of stay. While allocating patients to outpatient and inpatient settings would potentially allow for an optimized postoperative approach focused on patients with a high possibility of discharge on DOS, this was not the aim of our study. We included all patients with the purpose of identifying possible factors preventing early discharge regardless of specific patient characteristics. Pain was an issue in 24% of cases in our study, which is lower compared with a similar study investigating TKA conducted in the same surgical department in 2011, which found pain to be an issue in 53% of cases (Husted et al. 2011). This finding is expected as UKA is a less invasive procedure compared with TKA, resulting in less immediate postoperative pain. The multimodal opioid-sparing analgesic regime in our setting consists of a single-shot intravenous injection of 125 mg methylprednisolone as well as paracetamol 1g x 4 and celecoxib 200 mg x 2 daily for 7 days postoperatively with morphine used as rescue medication only. A preoperative injection of methylprednisolone has been shown to reduce pain at rest, pain during walking, and opioid consumption in the first 24 hours after UKA surgery (Rytter et al. 2017). Similar results, including a reduction in PONV and ondansetron use, have been shown in the first postoperative 48

Acta Orthopaedica 2020; 91 (4): 433â&#x20AC;&#x201C;438

hours following TKA surgery (Lunn et al. 2011). The use of methylprednisolone preoperatively is also part of a strategy to reduce PONV, which was present in only 6% of our patients still hospitalized on DOS. The use of midodrine has been suggested to decrease orthostatic hypotension and subsequently nausea and dizziness but was found to have limited effect (Jans et al. 2015). LIA is a part of our intraoperative procedure because it has been found to reduce postoperative pain following TKA surgery, and also when combined with a multimodal opioid-sparing analgesic regime (Andersen and Kehlet 2014, Seangleulur et al. 2016). In regard to UKA, a study found LIA to improve postoperative pain management and reduce opioid consumption and LOS (Essving et al. 2009). That study, however, did use a different composition of drugs compared with our LIA regime. Improved pain management could possibly be achieved using opioids or peripheral nerve blocks as part of the standard analgesic regime, but adverse effects such as an increased risk of falls upon initiation of opioid use as well as sedation, delirium, nausea, and urinary retention have been reported (Golladay et al. 2017, Seppala et al. 2018). Also, peripheral nerve blocks have limited additional value alongside LIA (Gudmundsdottir and Franklin 2017). Our postoperative strategy following fast-track knee surgery focuses on achieving sufficient and early mobilization, which could be impaired due to adverse effects of increased opioid consumption. We are therefore inclined to accept a higher level of pain in exchange for better postoperative mobilization. Reduction of pain is crucial, but it is not only desirable as a short-term goal regarding discharge as a recent study found control of early postoperative pain to be associated with improved 2-year functional outcome following UKA (Lakra et al. 2019). Among patients not discharged due to insufficient mobilization on DOS, a motor blockade was registered as the main reason in 44% of patients. Management of motor blockade in patients operated as #1 or #2 would allow for additional discharge of 11 patients on DOS. Interestingly, 1 patient operated using GA was reported to have motor blockade suggesting difficulty distinguishing between muscle weakness and â&#x20AC;&#x153;trueâ&#x20AC;? motor blockade. Therefore, the proportions of motor blockade and muscle weakness reported in our study might also be a product of difficulty distinguishing between the two. It is possible to speculate that GA might be better suited for UKA surgery with intended DOS discharge, though there is insufficient evidence to advocate for one over the other (Kehlet and Joshi 2019). Since logistical factors such as limited access to physiotherapy impacted 24% of insufficiently mobilized patients on DOS, organizational factors are to be further optimized. Physiotherapy is an important factor in short-term recovery since a loss of quadriceps function close to 80% is reported after knee arthroplasty (Bandholm and Kehlet 2012). Yet, due to

Acta Orthopaedica 2020; 91 (4): 433–438

limited literature no specific best practice has currently been determined. Even so, one aspect of acute postoperative rehabilitation has been determined: early initiation of physiotherapy after both UKA and TKA is associated with a reduction in LOS and overall cost, with no increase in adverse effects (Masaracchio et al. 2017, Henderson et al. 2018). However, the need for intense and early physiotherapy could be the reason why it is an organizational challenge. As previously mentioned, Jenkins et al. (2019) reported a potential reduction in LOS after UKA, due to a delay in knee flexion and physiotherapists working late shifts. However, that study does not include a control group. Our study has several limitations. Mainly, our results might have limited external validity since alternative setups regarding anesthesia, perioperative care, and treatment could impact reasons for continued hospitalization. However, our fasttrack setup is well described and our study investigates an unselected patient population, increasing the ability for other surgical centers to compare their results with ours. Also, all patients eligible for UKA surgery were considered eligible for DOS discharge. Additionally, both GA and SA were used, but no specific distinctions between the 2 groups were made, even though differences between the groups might be present. Conclusion One fifth of unselected UKA patients operated in a standardized fast-track setup were discharged on DOS. Pain and lack of mobilization were the major reasons for continued hospitalization within the initial postoperative 24–48 hours. To improve the number of patients discharged on the day of surgery, initiatives should focus on improving postoperative mobilization and postoperative pain management.

CBJ and KG had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: CBJ, AT, and KG. Acquisition, analysis, and interpretation of data: all authors. Drafting of the manuscript: CBJ and KG. Critical revision of the manuscript for important intellectual content: all authors. Acta thanks Karen Barker for help with peer review of this study.

Aasvang E K, Jorgensen C C, Laursen M B, Madsen J, Solgaard S, Kroigaard M, Kjaersgaard-Andersen P, Mandoe H, Hansen T B, Nielsen J U, Krarup N, Skott A E, Kehlet H. Safety aspects of postanesthesia care unit discharge without motor function assessment after spinal anesthesia: a randomized, multicenter, semiblinded, noninferiority, controlled trial. Anesthesiology 2017; 126: 1043-52. Andersen L O, Husted H, Otte K S, Kristensen B B, Kehlet H. A compression bandage improves local infiltration analgesia in total knee arthroplasty. Acta Orthop 2008; 79: 806-11. Andersen L O, Kehlet H. Analgesic efficacy of local infiltration analgesia in hip and knee arthroplasty: a systematic review. Br J Anaesth 2014; 113: 360-74. Bandholm T, Kehlet H. Physiotherapy exercise after fast-track total hip and knee arthroplasty: time for reconsideration? Arch Phys Med Rehabil 2012; 93: 1292-4.

437

Beard D J, Davies L J, Cook J A, MacLennan G, Price A, Kent S, Hudson J, Carr A, Lea, J, Campbell H, Fitzpatrick R, Arden N, Murray D, Campbell M K. The clinical and cost-effectiveness of total versus partial knee replacement in patients with medial compartment osteoarthritis (TOPKAT): 5-year outcomes of a randomised controlled trial. Lancet 2019; 394: 746-56. Bovonratwet P, Ondeck N T, Tyagi V, Nelson S J, Rubin L E, Grauer J N. Outpatient and inpatient unicompartmental knee arthroplasty procedures have similar short-term complication profiles. J Arthroplasty 2017; 32: 2935-40. Bradley B, Middleton S, Davis N, Williams M, Stocker M, Hockings M, Isaac D L. Discharge on the day of surgery following unicompartmental knee arthroplasty within the United Kingdom NHS. Bone Joint J 2017; 99-b: 788-92. Cross M B, Berger R. Feasibility and safety of performing outpatient unicompartmental knee arthroplasty. Int Orthop 2014; 38: 443-7. Danish Knee Arthroplasty Register. Annual Report; 2019. Essving P, Axelsson K, Kjellberg J, Wallgren O, Gupta A, Lundin A. Reduced hospital stay, morphine consumption, and pain intensity with local infiltration analgesia after unicompartmental knee arthroplasty. Acta Orthop 2009; 80: 213-19. Golladay G J, Balch K R, Dalury D F, Satpathy J, Jiranek W A. Oral multimodal analgesia for total joint arthroplasty. J Arthroplasty 2017; 32: S69s73. Gondusky J S, Choi L, Khalaf N, Patel J, Barnett S, Gorab R. Day of surgery discharge after unicompartmental knee arthroplasty: an effective perioperative pathway. J Arthroplasty 2014; 29: 516-19. Gromov K, Troelsen A, Raaschou S, Sandhold H, Nielsen C S, Kehlet H, Husted H. Tissue adhesive for wound closure reduces immediate postoperative wound dressing changes after primary TKA: a randomized controlled study in simultaneous bilateral TKA. Clin Orthop Relat Res 2019; Feb 6. doi: 10.1097/CORR.0000000000000637. [Epub ahead of print]. Gudmundsdottir S, Franklin J L. Continuous adductor canal block added to local infiltration analgesia (LIA) after total knee arthroplasty has no additional benefits on pain and ambulation on postoperative day 1 and 2 compared with LIA alone. Acta Orthop 2017; 88: 537-42. Henderson K G, Wallis J A, Snowdon D A. Active physiotherapy interventions following total knee arthroplasty in the hospital and inpatient rehabilitation settings: a systematic review and meta-analysis. Physiotherapy 2018; 104: 25-35. Husted H, Holm G, Jacobsen S. Predictors of length of stay and patient satisfaction after hip and knee replacement surgery: fast-track experience in 712 patients. Acta Orthop 2008; 79: 168-73. Husted H, Lunn T H, Troelsen A, Gaarn-Larsen L, Kristensen B B, Kehlet H. Why still in hospital after fast-track hip and knee arthroplasty? Acta Orthop 2011; 82: 679-84. Jans O, Mehlsen J, Kjaersgaard-Andersen P, Husted H, Solgaard S, Josiassen J, Lunn T H, Kehlet H. Oral midodrine hydrochloride for prevention of orthostatic hypotension during early mobilization after hip arthroplasty: a randomized, double-blind, placebo-controlled trial. Anesthesiology 2015; 123: 1292-1300. Jenkins C, Jackson W, Bottomley N, Price A, Murray D, Barker K. Introduction of an innovative day surgery pathway for unicompartmental knee replacement: no need for early knee flexion. Physiotherapy 2019; 105: 46-52. Kehlet H, Joshi G P. Anesthesia in enhanced recovery pathways for hip and knee arthroplasty: where is the evidence? Anesth Analg 2019; 128: e52. Kort N P, Bemelmans Y F L, Schotanus M G M. Outpatient surgery for unicompartmental knee arthroplasty is effective and safe. Knee Surg Sports Traumatol Arthrosc 2017; 25: 2659-67. Lakra A, Murtaugh T, Shah R P, Cooper H J, Geller J A. Early postoperative pain predicts 2-year functional outcomes following knee arthroplasty. J Knee Surg 2019; Jul 3. doi: 10.1055/s-0039-1692650. [Epub ahead of print] Liddle A D, Judge A, Pandit H, Murray D W. Adverse outcomes after total and unicompartmental knee replacement in 101,330 matched patients: a study of data from the National Joint Registry for England and Wales. Lancet 2014; 384: 1437-45.

438

Lunn T H, Kristensen B B, Andersen L O, Husted H, Otte K S, Gaarn-Larsen L, Kehlet H. Effect of high-dose preoperative methylprednisolone on pain and recovery after total knee arthroplasty: a randomized, placebo-controlled trial. Br J Anaesth 2011; 106: 230-8. Masaracchio M, Hanney W J, Liu X, Kolber M, Kirker K. Timing of rehabilitation on length of stay and cost in patients with hip or knee joint arthroplasty: a systematic review with meta-analysis. PLoS One 2017; 12: e0178295. 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: 41-5. National Joint Registry for England, Wales, Northern Ireland and the Isle of Man. National Joint Registryâ&#x20AC;&#x201D;16th Annual Report; 2019. Otero J E, Gholson J J, Pugely A J, Gao Y, Bedard N A, Callaghan J J. Length of hospitalization after joint arthroplasty: does early discharge affect complications and readmission rates? J Arthroplasty 2016; 31: 2714-25. Reilly K A, Beard D J, Barker K L, Dodd C A, Price A J, Murray D W. Efficacy of an accelerated recovery protocol for Oxford unicompartmental knee arthroplasty: a randomised controlled trial. Knee 2005; 12: 351-7.

Acta Orthopaedica 2020; 91 (4): 433â&#x20AC;&#x201C;438

Richter D L, Diduch D R. Cost comparison of outpatient versus inpatient unicompartmental knee arthroplasty. Orthop J Sports Med 2017; 5: 2325967117694352. Rytter S, Stilling M, Munk S, Hansen T B. Methylprednisolone reduces pain and decreases knee swelling in the first 24 h after fast-track unicompartmental knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2017; 25: 284-90. Rytter S, Jensen B G, Munk S, Hovsgaard S J, Hansen T B. A prospective study of day of surgery discharge in 368 consecutive patients with unicompartmental knee replacement. Dan Med J 2019; 66(9). pii: A5569. Seangleulur A, Vanasbodeekul P, Prapaitrakool S, Worathongchai S, Anothaisintawee T, McEvoy M, Vendittoli P A, Attia J, Thakkinstian A. The efficacy of local infiltration analgesia in the early postoperative period after total knee arthroplasty: a systematic review and meta-analysis. Eur J Anaesthesiol 2016; 33: 816-31. Seppala L J, van de Glind E M M, Daams J G, Ploegmakers K J, de Vries M, Wermelink A, van der Velde N. Fall-risk-increasing drugs: a systematic review and meta-analysis: III. Others. J Am Med Dir Assoc 2018; 19(4): 372.e1-372.e8.

Acta Orthopaedica 2020; 91 (4): 439–443

439

Socioeconomic position is associated with surgical treatment of open fractures of the lower limb: results from a Swedish populationbased study Yamin GRANBERG 1, Kalle T LUNDGREN 1,2, and Ebba K LINDQVIST 1,3 1 Department of Molecular Medicine and Surgery, Karolinska Institute, Stockholm; 2 Craniofacial 3 Reconstructive Plastic Surgery, Karolinska University Hospital, Stockholm, Sweden

Diseases, Karolinska University Hospital, Stockholm;

Correspondence: Yamin.Granberg@ki.se Submitted 2019-12-01. Accepted 2020-03-09.

Background and purpose — High-energy trauma to the lower limbs can result in open fractures, treated by reconstructive surgery or amputation. We examined whether socioeconomic position is associated with choice of primary treatment. Patients and methods — We performed a nationwide population-based study using the Swedish National Patient Register to identify all adult patients who between 1998 and 2013 underwent reconstruction or amputation after an open fracture below the knee. Information on socioeconomic position was collected from Statistics Sweden. Results — Of 275 individuals undergoing surgery after an open fracture below the knee during the study period, the 1st surgery was reconstructive in 58% of the patients and amputation in 42%. The chance of having an initial reconstruction was lower for women than for men (OR 0.5, 95% CI 0.3– 0.9), lower with age (OR 0.97, CI 0.96–0.99), and lower for individuals without employment compared with individuals in employment (OR 0.3, CI 0.2–0.5). Primary treatment was in women associated with family composition, whereas in men it was associated with level of education. Interpretation — Choice of primary treatment after open fracture in the lower limb is affected by socioeconomic position including sex, age, employment, family composition, level of education, and income.

In Sweden the prevalence of open tibia fractures is around 220 per year of which around one-third are classified as Gustilo– Anderson III (Weiss et al. 2008, Tampe et al. 2014). Similar incidences of open tibia fractures have been shown in studies on other populations (Court-Brown et al. 2012). Outcomes are poor for reconstructed and amputated patients alike, and in terms of function and pain do not necessarily differ between reconstruction and amputation (Bosse et al. 2002, Busse et al. 2007, Akula et al. 2011, Soni et al. 2012). Nearly half of patients treated for an open lower limb fracture will end up with a decreased range of motion, and little more than half of the patients are able to return to work (Busse et al. 2007, Soni et al. 2012, Barla et al. 2017). Reconstruction of the limb is easier for patients to accept, and may be preferred (Akula et al. 2011). Scoring systems such as the Ganga Hospital Open Injury Score (GHOISS) and the Mangled Extremity Severity Score (MESS) are available to guide the treating surgeon in the decision-making process, and account for the degree of tissue damage as well as other patient-related factors (Helfet et al. 1990, Rajasekaran et al. 2015). However, the utility of such scoring systems has been questioned (Ly et al. 2008, Loja et al. 2017). Long-term outcomes also appear to be affected by patient-related factors such as socioeconomic position and personal resources (MacKenzie et al. 2006, Driesman et al. 2017). Socioeconomic position, such as sex, level of education, income, family composition, and immigrant status, has in other healthcare areas been connected to incidence and outcome of disease (Woodward et al. 2015, Abdoli et al. 2017, Zommorodi et al. 2019). Furthermore, socioeconomic position, as determined by income and education, has been shown to affect the likelihood of undergoing operative treatment

440

Acta Orthopaedica 2020; 91 (4): 439–443

Table 2. Characteristics of the study population. Values are frequency (%) Factor

Amputation Reconstruction All n = 115 (42) n = 160 (58) n = 275 (100)

Sex Male 77 (67) 128 (80) 205 (75) Female 38 (33) 32 (20) 70 (25) Age category 16–19 3 (2.6) 11 (6.9) 14 (5.1) 20–39 18 (16) 43 (27) 61 (22) 40–59 25 (22) 53 (33) 78 (28) 60–79 28 (24) 36 (23) 64 (23) ≥ 80 20 (17) 5 (3.0) 25 (9.1) Missing 21 (18) 12 (7.5) 33 (12) Site of injury Upper tibia 13 (11) 17 (11) 30 (11) Shaft of tibia 49 (43) 78 (49) 127 (46) Lower tibia 21 (18) 35 (22) 56 (20) Multiple fractures 6 (5.2) 8 (5.0) 14 (5.1) Lower limb misc. 26 (23) 22 (14) 48 (18)

after a cruciate ligament injury in Sweden (Nordenvall et al. 2017). We examined whether determinants of socioeconomic position are associated with choice of primary treatment in patients with open fractures of the lower extremity.

Patients and methods We performed a population-based, nationwide register study using the Swedish National Patient Register (NPR) as well as register data from Statistics Sweden. All patients from the age of 16 and above who underwent amputation and/or reconstructive surgery for open fractures as a primary intervention in the lower extremity between January 1, 1998 and December 31, 2013 were identified through the NPR. From the same data source, information on type of injury and comorbidity by ICD-10 code, date and type of surgical procedure, hospital, county of habitation, age, and sex were collected (Table 1, see Supplementary data for a list of ICD codes). The identified individuals, being primarily treated with reconstructive surgery and/or amputation, were assumed to have extensive soft tissue damage and/or articular communication and/or contamination, therefore corresponding to Gustilo-Anderson grade III injuries. The dataset was cross-referenced by the register holder, the National Board of Health and Welfare (Socialstyrelsen), with the Statistics and Longitudinal Integration Database for Health Insurance and Labour Market Studies (Swedish acronym: LISA) from Statistics Sweden. From the LISA database information was retrieved on socioeconomic variables including disposable income (the sum of income, capital gain/loss, and tax), employment, level of education, and family composition, for all study subjects. The LISA database has a high level of completeness (Ludvigsson et al. 2019). Information

Table 3. Distribution of socioeconomic factors by primary treatment. Values are frequency (%) Factor

Amputation Reconstruction All n = 115 (42) n = 160 (58) n = 275

Disposable income a F1 (–11,527–7,963) 19 (17) 31 (19) F2 (8,036–10,700) 28 (24) 21 (13) F3 (10,718–14,082) 26 (23) 24 (15) F4 (14,182–19,327) 15 (13) 34 (21) F5 (19,355–56,991) 15 (13) 34 (21) Missing 12 (10) 16 (10) Employment status Employed 40 (35) 97 (61) Not employed 63 (55) 47 (29) Missing 12 (10) 16 (10) Family composition Married/cohabiting 28 (24) 48 (30) Single with child/ren 4 (3,5) 12 (7.5) Single without child/ren 64 (56) 74 (46) Living with parents 7 (6.1) 10 (6.3) Missing 12 (10) 16 (10) Level of education Low (≤ 9 years) 48 (42) 42 (26) Middle (10–12 years) 45 (39) 65 (41) High (> 12 years) 8 (7.0) 36 (23) Missing 14 (12) 17 (11) a

50 49 50 49 49 28 (10) 137 (50) 110 (40) 28 (10) 76 (28) 16 (5.8) 138 (50) 17 (6.2) 28 (10) 90 (33) 110 (40) 44 (16) 31 (11)

Yearly, in € (1 € = 11 Swedish Krona).

was collected from the year prior to the surgical treatment for the lower extremity injury. Statistics Categorical variables were described with frequencies and percentages, and continuous variables with means and standard deviations (SD). Age and disposable income were analyzed both as continuous variables and as categorical variables. The quintiles of the distribution of disposable income were used to divide disposable income into 5 equal parts, with the lowest 5th referred to as F1 and the highest as F5. From records of number of years in school, the levels of education were defined as low (< 10 years), middle (10–12 years), or high (> 12 years). County of habitation was created as a binary variable, divided into urban counties, with a university hospital and level 1 trauma center, and rural counties. In the analysis, F5 was used as a reference for disposable income and the highest educational level was used as reference for level of education. When investigating association between employment and choice of treatment we excluded patients over the age 65, the general age for retirement in Sweden, as well as patients younger than 19. Comparisons of binary or categorical variables were performed using the chi-square test or analysis of variance (ANOVA). Continuous variables were analyzed using Students t-test. A p-value smaller than 0.05 was considered statistically significant. Effect sizes were estimated using logistic regression and presented as odds ratios (OR) with 95% confidence intervals (CI). Statistical analyses were performed using Stata/IC 13 (StataCorp LP, College Station, TX, USA).

Acta Orthopaedica 2020; 91 (4): 439–443

441

Table 4. Associations of socioeconomic factors and reconstruction as primary treatment Factor

Men OR (95% CI)

Women OR (95% CI)

All OR (95% CI)

Sex Male N/A N/A Ref Female N/A N/A 0.5 (0.3–0.9) Age 16–19 Ref Ref Ref 20–39 0.8 (0.2–3.5) 48 (3.6–632) 0.7 (0.2–2.6) 40–59 0.8 (0.2–3.2) 18.7 (2.5–138) 0.6 (0.1–2.3) 60–79 0.6 (0.1–2.4) 8 (1.5–43) 0.4 (0.1–1.4) ≥ 80 years 0.3 (0.04–2.1) – 0.1 (0.01–0.3) Disposable income a F1 0.7 (0.3–1.7) 1.1 (0.2–8.0) 0.7 (0.3–1.7) F2 0.5 (0.18–1.3) 0.3 (0.04–1.5) 0.3 (0.1–0.8) F3 0.4 (0.2–1.0) 0.6 (0.1–3.5) 0.4 (0.2–0.9) F4 1.1 (0.4–3.0) 0.9 (0.1–6.1) 1 (0.4–2.4) F5 Ref Ref Ref Employment status Employed Ref Ref Ref Not employed 0.4 (0.2–0.8) 0.1 (0.04–0.5) 0.3 (0.2–0.5) Family composition Married/cohabiting 1.2 (0.56–2.1) 2.9 (0.8–10) 1.5 (0.84–2.6) Single with child/ren 0.8 (0.2–3.6) 20 (2.3–183) 2.6 (0.8–8.4) Single without child/ren Ref Ref Ref Living with parents 0.7 (0.3–2.1) – 1.2 (0.4–3.4) Level of education Low (≤ 9 years) 0.04 (0.005–0.3) 0.4 (0.1–1.7) 0.2 (0.1–0.5) Middle (10–12 years) 0.05 (0.007–0.4) 1.5 (0.4–5.5) 0.3 (0.1–0.8) High (> 12 years) Ref Ref Ref County of habitation Rural county 0.9 (0.5–1.6) 3.6 (1.2–11) 1.4 (0.8–2.2) Urban county Ref Ref Ref a

See Table 2.

Ethics, funding, and potential conflicts of interests The study was approved by the Ethical Review Board of Stockholm, 2015/1174-32. The authors received no funding for the study and have no conflicts of interest.

Results 275 individuals undergoing surgery after an open fracture in the lower limb during the study period were captured through NPR and included in the study (Table 2). The 1st surgery was reconstructive in 160 patients (58%) and amputation was performed in 115 (42%) patients. 75% were males and mean age at injury was 51 years (16–96). The most common injury was to the tibial shaft or the lower tibia. Details of socioeconomic factors are displayed in Table 3. In the entire cohort, the chance of having an initial reconstruction was lower for women than for men (OR 0.5, CI 0.3– 0.9), lower with higher age (OR 0.97, CI 0.96–0.99), lower for individuals without employment compared with individuals in employment (OR 0.3, CI 0.2–0.5), and lower for individuals with a low (OR 0.2, CI 0.1–0.5) or middle (OR 0.3, CI

0.1–0.8) level of education (Table 3). We found no statistically significant difference in choice of primary treatment related to lower disposable income (OR 1, CI 1–1) or to county of habitation (OR 1, CI 1–2). For men, the chance of having an initial reconstruction was lower with higher age (OR 0.99, CI 0.97–1.0), lower for individuals without employment compared with individuals in employment (OR 0.4, CI 0.2–0.8), and lower for individuals with a low (OR 0.1, CI 0.1–0.3) or middle (OR 0.1, CI 0.1–0.4) level of education (Table 4). For women, the chance of having an initial reconstruction was higher in 3 of the age groups (age 20–-79) compared with the youngest age group. The chance of an initial reconstruction was in women lower for individuals without employment compared to individuals in employment (OR 0.1, CI 0.1–0.5), higher for single women living with children (OR 20, 2–183) compared with single women without children, and higher in rural counties (OR 4, CI 1–11). In women there was no statistically significant association between level of education and primary treatment. After excluding individuals under 19 years of age (13 individuals) and over 65 years of age (60 individuals), the association between employment and chance of initial reconstruction persisted (p = 0.006).

Discussion In this nationwide, population-based study, we found that choice of primary treatment of open fracture in the lower limb is associated with socioeconomic position. We found a sex bias favoring men in initial reconstruction rates. Furthermore, we found that the chance of initial reconstruction was associated with family composition for women, but with level of education for men. Long-term outcomes after an open fracture of the lower limb with extensive soft tissue damage appear to be affected not only by the type of injury and treatment factors, but also by patient-related factors including socioeconomic position (MacKenzie et al. 2006). Our results suggest that socioeconomic position may also influence choice of primary treatment, thus potentially contributing 2-fold to long-term outcomes. Socioeconomic position has in other healthcare areas been connected to incidence of disease as well as to outcomes; however, whether socioeconomic position affects choice of treatment in open lower limb fractures has not previously been investigated. This is to our knowledge the 1st study to dem-

442

onstrate an association between treatment and patient factors such as sex, age, level of education, income, and family composition. As Swedish residents are entitled to the same quality of health care, medically unmotivated differences in treatment strategy should be explored further. It is important to examine the precise mechanisms by which the socioeconomic position influences the decision to undertake primary reconstruction or to amputate. Choice of primary treatment of open lower limb fractures is a complex decision, and many factors are taken into account, not least patient compliance with postoperative restrictions and recommendations. Patient compliance in treatment, for example in secondary prevention of cardiovascular disease, has been shown to be lower in those with a low socioeconomic position (Wallach-Kildemoes et al. 2013). The reasons for this are unknown, but it is plausible that factors such as alcohol or substance abuse, closely associated with socioeconomic position, contribute to clinical treatment decisions as well as to patient involvement in those decisions. Since alcohol or substance abuse was very rarely recorded in the register used for comorbidities, we could not control for these factors. We demonstrate a sex bias favoring men in initial reconstruction rates. Furthermore, we show that the chance of initial reconstruction is associated with family composition for women, and with level of education for men. These findings could possibly be explained by the residual confounding factors mentioned above. That patients with a higher level of education are more likely to get offered initial reconstructive surgery is not perhaps surprising; however, it is a notable finding that this association is seen only in men and not in women in our cohort. This could be due to a treatment selection sex bias, but we cannot fully explain this discrepancy from the data at hand. With the use of the registries available in Sweden it is possible to get a high inclusion rate and a large amount of cases, despite the injuries being rare. A potential challenge when working with register data that covers a long time span is that the definition of the parameters might change over the years and not be comparable. Therefore, we have tried to limit the parameters in the study to those that have kept the same or similar definition during the full study period. As expected, there was a higher proportion of reconstructed patients in the younger age groups. It is likely that comorbidity increases with age, and that this affects treatment decisions. We collected data for all ICD codes registered during the hospital stay, but adjustment for comorbidities such as associated trauma, cardiovascular disease, or substance abuse was impossible due to very few such registrations. However, at least 1 prior study including severely injured limbs has shown that neither severity of injury nor associated injuries affect functional outcome (Bosse et al. 2002). Complex lower limb injuries should be referred to level 1 trauma centers managed in multidisciplinary collaboration (Naique et al. 2006, Sommar et al. 2015). In Sweden, most

Acta Orthopaedica 2020; 91 (4): 439â&#x20AC;&#x201C;443

local or regional hospitals have some collaboration with a university hospital with access to both plastic and orthopedic surgeons. This is probably reflected by our finding that, in the entire cohort, there was no difference in the choice of primary treatment by county of habitation. That women in rural counties were more likely to be reconstructed than women in urban counties is more difficult to explain. No such association was found for men, and it is possible that family composition is a confounder in this relationship in the female subgroup. The strengths in our study lie in the population-based study design, and in the register-based methodology, excluding the possibility of recall bias or physician selection bias. Limitations include some missing data in a few of the variables, and the retrospective study design. Missing data in the age variable are likely due to patient age not always being noted in the electronic medical records on emergency admissions of unidentified trauma patients, whereas missing data in for example disposable income is likely age-related. We did not have access to individual medical records and, as previously mentioned, residual confounding could explain some of the associations detected. We were not able to ascertain unilateral or bilateral injuries, or even the individual grades of injury. However, since all individuals received treatment with either amputation or reconstructive surgery, we deem it reasonable to assume that a majority of them had injuries corresponding to Gustilo-Anderson grade III. Orthopedic and reconstructive plastic surgeons should be alert to the risk of undue influence in treatment selection from patient socioeconomic factors, and treatment guidelines are probably helpful tools to guide clinicians towards non-biased management. Supplementary data Table 1 is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2020. 1751418 YG, KL and EL designed the study. YG and EL collected the data. EL performed the statistical analysis. YG, KL and EL wrote the report. Acta thanks Charles Court-Brown and Micha Holla for help withÂ peer review of this study.

Abdoli G, Bottai M, Sandelin K, Moradi T. Breast cancer diagnosis and mortality by tumor stage and migration background in a nationwide cohort study in Sweden. Breast 2017; 31: 57-65. doi: 10.1016/j.breast.2016.10.004. Akula M, Gella S, Shaw C J, McShane P, Mohsen A M. A meta-analysis of amputation versus limb salvage in mangled lower limb injuries: the patient perspective. Injury 2011; 42(11): 1194-7. doi: 10.1016/j. injury.2010.05.003. Barla M, Gavanier B, Mangin M, Parot J, Bauer C, Mainard D. Is amputation a viable treatment option in lower extremity trauma? Orthop Traumatol Surg Res 2017; 103(6): 971-5. doi: 10.1016/j.otsr.2017.05.022. Bosse M J, MacKenzie E J, Kellam J F, Burgess A R, Webb L X, Swiontkowski M F, Sanders R W, Jones A L, McAndrew M P, Patterson B M,

Acta Orthopaedica 2020; 91 (4): 439–443

McCarthy M L, Travison T G, Castillo R C. An analysis of outcomes of reconstruction or amputation after leg-threatening injuries. N Engl J Med 2002; 347(24): 1924-31. doi: 10.1056/NEJMoa012604. Busse J W, Jacobs C L, Swiontkowski M F, Bosse M J, Bhandari M, Evidence-Based Orthopaedic Trauma Working G. Complex limb salvage or early amputation for severe lower-limb injury: a meta-analysis of observational studies. J Orthop Trauma 2007; 21(1): 70-6. doi: 10.1097/ BOT.0b013e31802cbc43. Court-Brown C M, Bugler K E, Clement N D, Duckworth A D, McQueen M M. The epidemiology of open fractures in adults: a 15-year review. Injury 2012; 43(6): 891-7. doi: 10.1016/j.injury.2011.12.007. Driesman A, Fisher N, Konda S R, Pean C A, Leucht P, Egol K A. Racial disparities in outcomes of operatively treated lower extremity fractures. Arch Orthop Trauma Surg 2017; 137(10): 1335-40. doi: 10.1007/s00402-017-2766-z. Helfet D L, Howey T, Sanders R, Johansen K. Limb salvage versus amputation: preliminary results of the Mangled Extremity Severity Score. Clin Orthop Relat Res 1990; (256): 80-6. Loja M N, Sammann A, DuBose J, Li C S, Liu Y, Savage S, Scalea T, Holcomb J B, Rasmussen T E, Knudson M M. The mangled extremity score and amputation: time for a revision. J Trauma Acute Care Surg 2017; 82(3): 518-23. doi: 10.1097/ta.0000000000001339. Ludvigsson J F, Svedberg P, Olén O, Bruze G, Neovius M. The longitudinal integrated database for health insurance and labour market studies (LISA) and its use in medical research. Eur J Epidemiol 2019; 34(4): 423-37. doi:10.1007/s10654-019-00511-8. Ly T V, Travison T G, Castillo R C, Bosse M J, MacKenzie E J, Group L S. Ability of lower-extremity injury severity scores to predict functional outcome after limb salvage. J Bone Joint Surg Am 2008; 90(8): 1738-43. doi: 10.2106/jbjs.g.00136. MacKenzie E J, Bosse M J. Factors influencing outcome following limbthreatening lower limb trauma: lessons learned from the Lower Extremity Assessment Project (LEAP). J Am Acad Orthop Surg 2006; 14(10 Spec. No.): S205-10. doi: 10.5435/00124635-200600001-00044. Naique S B, Pearse M, Nanchahal J. Management of severe open tibial fractures: the need for combined orthopaedic and plastic surgical treatment in specialist centres. J Bone Joint Surg Br 2006; 88(3): 351-7. doi: 10.1302/0301-620x.88b3.17120.

443

Nordenvall R, Marcano A I, Adami J, Palme M, Mattila V M, Bahmanyar S, Felländer-Tsai L. The effect of socioeconomic status on the choice of treatment for patients with cruciate ligament injuries in the knee: a population-based cohort study. Am J Sports Med 2017; 45(3): 535-40. doi: 10.1177/0363546516672651. Rajasekaran S, Sabapathy S R, Dheenadhayalan J, Sundararajan S R, Venkatramani H, Devendra A, Ramesh P, Srikanth K P. Ganga hospital open injury score in management of open injuries. Eur J Trauma Emerg Surg 2015; 41(1): 3-15. doi: 10.1007/s00068-014-0465-9. Sommar P, Granberg Y, Halle M, Skogh A C, Lundgren K T, Jansson K A. Effects of a formalized collaboration between plastic and orthopedic surgeons in severe extremity trauma patients; a retrospective study. J Trauma Manag Outcomes 2015; 9: 3. doi: 10.1186/s13032-015-00234. Soni A, Tzafetta K, Knight S, Giannoudis P V. Gustilo IIIC fractures in the lower limb: our 15-year experience. J Bone Joint Surg Br 2012; 94(5): 698703. doi: 10.1302/0301-620x.94b5.27948. Tampe U, Weiss R J, Stark B, Sommar P, Al Dabbagh Z, Jansson K A. Lower extremity soft tissue reconstruction and amputation rates in patients with open tibial fractures in Sweden during 1998–2010. BMC Surg 2014; 14: 80. doi: 10.1186/1471-2482-14-80. Wallach-Kildemoes H, Andersen M, Diderichsen F, Lange T. Adherence to preventive statin therapy according to socioeconomic position. Eur J Clin Pharmacol 2013; 69(8): 1553-63. doi: 10.1007/s00228-013-1488-6. Weiss R J, Montgomery S M, Ehlin A, Al Dabbagh Z, Stark A, Jansson K A. Decreasing incidence of tibial shaft fractures between 1998 and 2004: information based on 10,627 Swedish inpatients. Acta Orthop 2008; 79(4): 526-33. doi: 10.1080/17453670710015535. Woodward M, Peters S A, Batty G D, Ueshima H, Woo J, Giles G G, Barzi F, Ho S C, Huxley R R, Arima H, Fang X, Dobson A, Lam T H, Vathesatogkit P, Asia Pacific Cohort Studies C. Socioeconomic status in relation to cardiovascular disease and cause-specific mortality: a comparison of Asian and Australasian populations in a pooled analysis. BMJ Open 2015; 5(3): e006408. doi: 10.1136/bmjopen-2014-006408. Zommorodi S, Bottai M, Hultgren R. Sex differences in repair rates and outcomes of patients with ruptured abdominal aortic aneurysm. Br J Surg 2019; 106(11): 1480-7. doi: 10.1002/bjs.11258.

444

Acta Orthopaedica 2020; 91 (4): 444–449

12-year survival analysis of 322 Hintegra total ankle arthroplasties from an independent center Mina Jane ZAFAR 1,2, Thomas KALLEMOSE 3, Mostafa BENYAHIA 1, Lars Bo EBSKOV 1, and Jeannette Østergaard PENNY 1,4 1 Department of Orthopedic Surgery, Hvidovre University Hospital, Denmark; 2 Department of Orthopedics, Sahlgrenska University Hospital, Gothenburg, Sweden; 3 Clinical Orthopedic Research Centre, Hvidovre Hospital, Copenhagen University Hospital, Denmark; 4 Department of Orthopedic Surgery.

University Hospital Zealand, Koege, Denmark Correspondence: mina@tpu.se Submitted 2019-07-31. Accepted 2020-03-03.

Background and purpose — Total ankle arthroplasties (TAAs) have larger revision rates than hip and knee implants. We examined the survival rates of our primary TAAs, and what different factors, including the cause of arthritis, affect the success and/or revision rate. Patients and methods — From 2004 to 2016, 322 primary Hintegra TAAs were implanted: the 2nd generation implant from 2004 until mid-2007 and the 3rd generation from late 2007 to 2016. A Cox proportional hazards model evaluated sex, age, primary diagnosis, and implant generation, pre- and postoperative angles and implant position as risk factors for revision. Results — 60 implants (19%) were revised, the majority (n = 34) due to loosening. The 5-year survival rate (95% CI) was 75% (69–82) and the 10-year survival rate was 68% (60–77). There was a reduced risk of revision, per degree of increased postoperative medial distal tibial angle at 0.84 (0.72–0.98) and preoperative talus angle at 0.95 (0.90–1.00), indicating that varus ankles may have a larger revision rate. Generation of implant, sex, primary diagnosis, and most preand postoperative radiological angles did not statistically affect revision risk. Interpretation — Our revision rates are slightly above registry rates and well above those of the developer. Most were revised due to loosening; no difference was demonstrated with the 2 generations of implant used. Learning curve and a low threshold for revision could explain the high revision rate.

Arthritis in the ankle often develops earlier than in the hip or knee, and 70% have a traumatic etiology (Saltzman et al. 2005, Brown et al. 2006). Total ankle arthroplasty (TAA) can be indicated for severe arthritis in the ankle joint, but the anatomical preconditions, like a small surface area and high stress from compression and torque (Bouguecha et al. 2011, Kakkar and Siddique 2011), makes it less durable than hip and knee prosthetics. The Hintegra TAA, a 3-component mobile bearing, uncemented implant (Hintermann et al. 2004) is widely used and results from the development center demonstrate survival rates of 94% and 84% after 5 and 10 years’ follow-up (Barg et al. 2013). This is considerably more than the survival rates from national registries. Labek et al. (2011) demonstrated that development centers report only half of the revision rate that can be found in the few existing national registers. In a systematic review of primary Agility total ankle arthroplasty (DePuy Synthes Orthopedics, Warsaw, IN, USA), the author (Roukis 2012) found that the incidence of complications increased from 7% to 12%, in studies where the inventor was excluded. Similar results were found by Prissel and Roukis (2013), who found an increased incidence of complications from 6% to 13% in studies where the inventor or faculty consultants were excluded. These studies indicated the risk of selection (inventor) and publication (conflict of interest) bias. Planning and surgical technique, including significant experience, are mandatory for a successful outcome. The better result from development centers may reflect, besides the above-mentioned bias, that there is a long learning curve and that the indication for revision surgery varies. We examined the survival rates of primary Hintegra TAAs performed at Hvidovre Hospital, with revision rate as outcome. We report primary diagnosis for primary TAA and examine whether sex, generation of the implant, preoperative angles and implant position affect the revision rate.

Acta Orthopaedica 2020; 91 (4): 444–449

445

Patients and methods The study is retrospective, and all operations were carried out at a specialized foot and ankle department, where the 2nd generation Hintegra TAA has been implanted since 2004 in lowdemand elderly patients with severe arthritis. The Hintegra TAA (Integra Life Sciences, Newdeal SA, Lyon, France) is an unconstrained, 3-component system that provides inversion–eversion stability. The Hintegra TAA includes 2 metallic components and an ultra-high-density polyethylene mobile bearing. The non-articulating surfaces have a porous coating with 20% porosity and are covered by cobalt-chromium and double hydroxyapatite coating (2nd generation Hintegra) or titanium fluid and hydroxyapatite (3rd generation Hintegra). Our institution changed from 2nd to 3rd generation mid-2007. The patients were operated using a standard anterior approach, in accordance with the manufacturer’s instructions with relevant cutting guides and sizing regulations. The goal was neutral alignment in the AP view without collision laterally and the peak of the talus in side-view level between 40% and 50% from the frontal tibial edge. The operations were carried out using a tourniquet, and 1.5 g of cefuroxime was administered preoperatively. The foot was immobilized in a circular cast from the 2nd postoperative day for 3 weeks. The cast was converted to a removable boot (Don Joy type) for another 3 weeks, with weight-bearing as tolerated, followed by home physiotherapy if needed. Dalteparin 5000 I.E. s.c. was administered postoperatively for 3 to 5 days (this was prolonged in patients with risk factors). Data collection The in-house operation booking system (Orbit, Evry Healthcare Systems AB, Kristianstad, Sweden) and the hospital’s local registration system were searched for primary and secondary insertion of Hintegra TAA, using the designated codes for ankle implants between 2004 and 2016. This search yielded 322 surgeries from 2004 to 2016 using primary Hintegra TAA. All had a patient record review, and an assessment of the radiographs. Revisions were noted, as were the reasons for revision and the revision type. The revisions were defined according to Henricson et al. (2011a): “A revision of the TAA is defined as removal or exchange of one or more of the prosthetic components with the exception of incidental exchange of the polyethylene insert.” For this study, only the 1st revision was included. We did not record minor wound complications etc. that did not lead to additional surgery. Aseptic loosening was defined as the failure of the bond, between an implant and bone, in the absence of infection, defined on radiographic findings of radiolucent lines around the implant and/or peroperatively, as lack of bony ingrowth. Primary diagnosis (post-traumatic, primary osteoarthritis [OA], rheumatoid arthritis [RA], and other), sex, age at sur-

Figure 1. The center of the tibial plateau was determined by drawing a circle within the medial and lateral cortex. A second circle fit inside the distal tibia between the medial and lateral cortex and touched the plafond distally. The mechanical axis goes through both the center of the distal tibia and the center of the talus. A line marking the tibial plateau/ distal tibial component intersected the mechanical axis for the medial distal tibial (MDTA; small arch) angle. The medial talus (large arch) angle was measured from a transecting line, tracing the superior talus/ talar component. An angle above 90° is a valgus angle and below 90° is a varus angle.

gery, time since the surgery, and the generation of the prosthesis were registered. Radiology The preoperative radiographs were digital and varying in length. For the preoperative images, images within 6 months prior to the surgery were accepted. We attempted to use the most recent and aimed to use a mortise view (mortise view was missing in a few cases where a straight AP was used). The postoperative radiographs were not strictly standardized but contained a mixture of weight-bearing and nonweight-bearing images. The postoperative implant position was accepted in images up to 4 months postoperatively. The valgus/varus position pre- and postoperatively was measured on short images using medial distal tibial angle (MDTA) and medial talus angles (Figure 1) according to Barg et al. (2012). The horizontal line of the anterior distal tibial angle (ADTA; Figure 2) was defined by the most distal anterior and posterior bony points preoperatively, and by the distal tibial component postoperatively. Statistics Survival rates by Kaplan–Meier curves were stratified by the generation of the implant. The analysis of risk factors for revision rate used the Cox proportional hazards model, with revision as outcome. The variables of interest were: time since surgery, age at surgery, sex, placement of the TAA (MDTA and ADTA as well as medial talus angle), preoperative alignment of the ankle, generation of the TAA, and primary diag-

446

Acta Orthopaedica 2020; 91 (4): 444–449

Primary TAA 2004–2016 n = 565 Excluded Not Hintegra or miscoded n = 243 Hintegra (n = 322): – men, 172 – women, 150

Figure 2. Markings to calculate the anterior distal tibial angle tibial plateau/distal tibial component (ADTA; green arch). For tibial axis, see Figure 1.

Generation 2 n = 47

Generation 3 n = 273

Revised (n = 22): – new Hintegra, 9 – fusion, 11 – amputation, 2

Revised (n = 36): – new Hintegra, 17 – fusion, 17 – amputation, 1 – change of meniscus, 1

Unknown generation n=2 Revised (n = 2): – fusion, 2

Figure 3. Flow chart.

Table 1. 322 total ankle arthroplasties (TTAs) distributed among 8 different surgeons and subsequent revision surgeries, 13 unknowns Surgeon 1 2 3 4 5 6 7 8 Unknown Years active 2008–2014 2005–2010 2014–2016 2006–2016 2010–2016 2004–2011 2007–2016 2012 2003–2007 Primary TAAs 21 11 14 117 67 38 40 1 13 Revision, n 1 4 0 11 11 23 3 1 6

nosis (post-traumatic, primary OA, RA, and other). The model included a combination of the generation of the TAA and the time the specific generation had been in use, thereby allowing the effect of time within each prosthesis type. Hazard ratios and 95% confidence intervals (CI) were determined. P-values below 0.05 were considered statistically significant. The analysis was carried out in R 3.0.2 (R Foundation for Statistical Computing, Vienna, Austria). Ethics, funding, data sharing, and potential conflicts of interest As the study was based on registry data, ethical approval is not needed according to Danish law. This research did not receive grants from any funding agency. The data are available from the corresponding author. We have no conflict of interest to declare.

Results The search identified 322 Hintegra TAA (Figure 3). 47 were 2nd generation and 273 were 3rd generation. 2 were unknown generation but they were implanted around the time we changed from 2nd to 3rd generation. 150 patients were females and 172 males. 13 were bilateral. Mean age was 60 years (24–81).

60% of the 322 primary implants had posttraumatic OA, 31% primary OA, 9% RA, and < 1% other, which had no statistically significant impact on revision risk. 60 implants were revised. The reasons for revision were 34 cases with suspected aseptic loosening, aseptic loosening with or without cyst formation as primary indication (8 of these cases with radiographic cyst formation/mechanical loosening), 7 cases of infection, 6 cases of malalignment, 3 cases of persistent pain, 3 cases of fracture, and 7 cases of other reasons. 22 of 47 2nd-generation Hintegra were revised: 9 to new Hintegra (5 partly and 4 fully revised), 11 were revised to fusions, and 2 were amputated due to problems with wound healing and chronic nerve pain. 36 of 274 3rd-generation were revised to 17 new Hintegra (2 fully revised, 15 partly revised), 17 to fusion, 1 was amputated for unknown reasons, and 1 had a change of a fractured meniscus. 2 unknown generation implants were both revised to fusion. Additional surgery after the primary Hintegra implant was recorded in 40 cases but was not registered as revision as the implants were left intact. The procedures included cheilectomy/osteophyte removal (n = 4), decompression medial or lateral (n = 8), bone grafting of cyst (n = 3), ligament reconstruction (n = 2), fracture repair (n = 2), osteotomy (n = 6), wound infection (n = 5), a mixture of these above (n = 6), and other (n = 9).

Acta Orthopaedica 2020; 91 (4): 444–449

447

Table 2. The Cox proportional hazard multivariate analysis of risk for revision Factor

Figure 4. Kaplan–Meier plot of the survival rates of generation 2 and generation 3.

The operations were performed by 8 different surgeons and 13 were operated by unknown surgeons (Table 1). The survival at 5 and 10 years was 0.75 (CI 0.69–0.82) and 0.68 (CI 0.60–0.77) respectively. For the 2nd generation, the 5- and 10-year survival was 0.67 (CI 0.54–0.82) and 0.60 (CI 0.47–0.76) and for the 3rd generation the 5- and 10-year survival was 0.79 (CI 0.73–0.87) and 0.78 (CI 0.71–0.86) respectively (Figure 4). In the Kaplan–Meier plot, the survival rate of generation 2 was slightly below that of generation 3. The univariate analysis found a non-statistically significant, lower expected risk of revision of the 3rd generation than the 2nd of 0.60 (0.33–1.08). The multivariate analysis found a similar, non-statistically significant effect size of 0.56 (0.23–1.3). The multivariate Cox proportional hazards model (Table 2) found a reduced risk of revision for each increased degree of the postoperatively MDTA, 0.84 (0.72–0.99), suggesting that a valgus position of the implant is safer than a varus position. Similarly, the small effect of reduced risk of revision of 0.95 (0.90–1.00) per degree of increased preoperative medial talus angle suggests that patients with pre-existing varus before TAA surgery had revision more frequently than valgus ankles. No risk was found for sex, age, the primary diagnosis, and the remaining pre- and postoperative angles in the multivariate analysis but increasing age at surgery was associated with a lower risk of revision in the univariate analyses, 0.97 (CI 0.94–0.99). The results were similar by restricting them to revisions due to osteolysis/aseptic loosening and malalignment alone.

Discussion The primary aim of our study was to analyze the survival rates of primary Hintegra TAA. With a total survival rate of 75% after 5 years and 60% after 10 years, our results are inferior compared with the results reported by Barg et al. (2013) where

Hazard ratio (95% CI)

Male sex 1.10 (0.50–2.4) Age 0.98 (0.95–1.0) Generation 3 0.56 (0.23–1.3) Post-trauma 1.23 (0.48–3.2) Post-infection 1.26 (0.10–22) Rheumatoid arthritis 1.26 (0.31–5.1) Angle pre-Hintegra (per 1°) MDTA 1.05 (0.95–1.2) Medial talus angle 0.95 (0.90–1.0) ADTA 1.01 (0.94–1.1) Angle post-Hintegra (per 1°) MDTA 0.84 (0.72–0.98) Medial talus angle 1.02 (0.90–1.2) ADTA 1.04 (0.93–1.2) The Cox proportional hazard multivariate analysis shows changes in risk for revision when the variable is increased by 1 unit. When the MDTA and medial talus variable are increased by 1° it means that the ankle goes towards a valgus position. Whether it results in an increase or reduction of the risk depends on the hazard ratio size; if this is less than 1, there is a reduced risk and if it is more than 1, it is an increased risk. MDTA = medial distal tibial angle, ADTA = anterior distal tibial angle.

overall survival rates of 634 patients were 94% and 84% after 5 and 10 years, respectively. The Swedish Ankle Registry annual report for 2014 reports a 1.9 increased risk of revision compared with the development center, but the numbers in the register are very low and the difference does not reach statistical significance. Willegger et al. (2013) found that 10 of 16 Hintegra TAA implants had survived after 4 years and suggests that publications by implant inventors show a tendency towards superior results. Our 3rd-generation Hintegra univariate risk of revision was 0.60 (0.33–1.08), compared with 2nd-generation Hintegra. The difference is not statistically significant but could reflect a type II error since the effect size (0.60 vs. 0.56) remained relatively unchanged. However, as the majority of the surgeons had their learning curve before the introduction of the 3rd-generation Hintegra the effect size could likely reflect learning curve bias rather than type II error of the design difference. This is supported by Roukis et al. (2016), who pointed out that design may not be so important as previously believed, as much as the weighted mean survival after the 1st-generation TAA prostheses was 0.76 at 10 years, for 2nd generation 0.83 and for 3rd generation 0.83. In our study there was a small number of operations per surgeon (6 of 8 surgeons had 40 TAAs or below) compared with the developer center. It is well established that TAA surgery involves a significant learning curve (Haskell and Mann 2004). This may explain the higher revision rate found in this study. Yang et al. (2019) reported, in a large series but single surgeon study, good survival rates of 92% at a mean 6.4 years. The 5-year survival rate of TAAs in registries is between 93% and 78% (Bartel 2015). The Australian Registry (2019)

448

with 442 primary Hintegra implants reports a 5-year cumulative revision rate of 11%, in line with the other implants in the register. The 5-year cumulative revision rate of all TAAs in the UK is 7%, but the British Orthopaedic Association itself suspects underreporting of revisions to the registry (UK National Joint Registry 2019). The Hintegra implants in the New Zealand register are limited and with very wide 95% confidence intervals for revision. Both the NZ and the Australian Registry use revision per component year. In registries with many “young” implants this will reflect the early revisions due to infection etc. but may not capture many of the aseptic loosenings that occur later and may give a different profile than, e.g., the Swedish registry (NZJR 2013). The revision rates in our study are not only considerably higher than those of the developer but also above those of the registries. Our survival rates are lower than the weighted mean survival after 1st-generation TAA. The registries, however, are also flawed. The UK suspects a completion rate below 90% but has no data to validate this yet. The Finnish rate is just above 90% (Skyttä and Koivu 2010), and most others are not disclosed. The French registry registers below 80% (Besse et al. 2010) and misses many from small-volume centers, and the risk of underreporting of revision is high in the registry data. The strength of our study is the unedited volume of a single center with multiple surgeons. In our study, the indications for revision in more than half of cases were suspected aseptic loosening because of diffuse loosening or cyst formation. Approximately half were revised to a new TAA and half to a fusion. The prosthetic design may not be the most important factor influencing long-term survival. Proper patient selection and optimal prosthetic implantation including a perfect balanced ankle are probably more important, as regards the impact on the survival rate, than the indication for revision. The high revision rate may reflect that our center tends to have an aggressive revision approach, if we suspect aseptic loosening and/or cyst formation. We believe it facilitates revision surgery without the need for more substantial bone grafting and often prevents fusion. As a referral center, we treat patients from other centers, where large cyst formation and possible loosening are typically accepted if the patient does not have significant pain. It is our experience that this pending strategy often leads to complex fusions with large structural allografts. The problem, when comparing survival and revision rates, is that the criteria for revision are not formalized, nor are the different ways of reporting the function of the patients. We found patients with high talar angle preoperatively and high MDTA postoperatively to have fewer revisions. This is in line with the findings of Henricson and Ågren (2007) and further discussed by Coetzee (2008). However, Lee et al. (2018) showed in their comparison of the survival rate on their 144 Hintegra patients that the long-term outcome was equally as

Acta Orthopaedica 2020; 91 (4): 444–449

good regardless of the preoperative alignment (up to 20°) as long as the postoperative alignment was neutral. Considering the non-standardized and retrospective radiographic set-up available in our study these results should be evaluated with care; nonetheless, optimal alignment of the prosthesis should logically reduce the possible factors that may lead to aseptic loosening and pain. The univariate analysis suggested a reduction of revision risk with increasing age, in line with the findings of Henricson et al. (2011b), but the effect was not statistically significant in the multivariate analysis. It does, however, mirror the results from the Australian (2019) and New Zealand Ankle Registry (NZJR 2018) and probably reflects the lower demands for function, as well as a disinclination to perform surgery on the elderly with their associated somatic risks (NZJR 2013). However, a short-term study only found slightly lower and nonsignificant functional levels and slightly more comorbidities for the elderly (Demetracopoulos et al. 2015). Johnson-Lynn et al. (2018) did not show any definite association between age and revision, and discussed that this was likely due to the small total number of revisions at their 5-year follow-up of 106 patients. The strength of our study is the unedited volume of a single center. It is the largest independent Hintegra patient group, with a 10-year survival analysis reported. The drawback of the study is the retrospective design. The assumed primary diagnosis was not always present in the patient record and the exact cause of revision is, in some cases, not precisely described. Cysts are known to affect the revision rate (Labek et al. 2011). It was not our impression that cyst formation was prominent but, due to a new IT provider, the old images are not available for confirmation of this. We are also limited retrospectively in having only plain radiographs and not CT images. Hanna et al. (2007) showed that radiographs are not sufficient to diagnose cyst formation or osteolysis. Deleu et al. (2014) indicated 24 cases of osteolysis in 50 patients with the Hintegra prosthesis. This number, larger than ours, can most likely be explained by our limitation in radiographic imaging. We do agree with Yang et al. (2019), who had 31 cases of peri-prosthetic osteolysis in their study of 242 TAA, that the prevention of peri-prosthetic osteolysis should be a priority for TAA. Another drawback is that the postoperative radiographs were not standardized. Only a few radiographs were AP rather than mortise view. We have been unable to find literature concerning the effect of axial rotation on the distal tibial angle but, with only these few exceptions from standard view, we find that our results are valid and further supported by being in agreement with previous findings (Henricson and Ågren 2007, Coetzee 2008). In conclusion, we found revision rates slightly above the registries, but well above those of the development center. More than half were revised due to suspected loosening with aseptic loosening or cysts. No difference was demonstrated

Acta Orthopaedica 2020; 91 (4): 444–449

between the 2 generations of implant used. The high number of surgeons with a low number of operations (learning-curve surgeons) and an aggressive approach to treating aseptic loosening and suspected aseptic loosening might explain the inferior survival rate in this study.

All the authors made a substantial contribution to the conception of the study. MZ, LBE, MB and JØP conceived and planned the study. TK performed the statistical analysis. All authors discussed the results and critically revised the manuscript, which was drafted by MZ. Acta thanks Anders Henricson and Helka Koivu for help with peer review of this study.

Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR). Hip, Knee & Shoulder Arthroplasty: 2019 Annual Report. Adelaide: AOA, 2019. Barg A, Harris M D, Henninger H B, Amendola R L, Saltzman C L, Hintermann B. Medial distal tibial angle: comparison between weightbearing mortise view and hindfoot alignment view. Foot Ankle Int 2012; 33: 65561. Barg A, Zwicky L, Knupp M, Henninger H B. HINTEGRA total ankle replacement: survivorship analysis in 684 patients. J Bone Joint Surg Am 2013; 95: 1175-83. Bartel A F. Total ankle replacement survival rates based on Kaplan–Meier survival analysis of National Joint Registry data. Clin Pod Med Surg 2015 32(4): 483-94. Besse J L, Colombier J A, Asencio J, Bonnin M, Gaudot F, Jarde O. Total ankle arthroplasty in France. Orthop Traumatol Surg Res 2010; 96(3]): 291-303. Bouguecha A, Weigel N, Behrens B A, Stukenborg-Colsman C, Waizy H. Numerical simulation of strain-adaptive bone remodelling in the ankle joint. Biomed Eng Online 2011; 10: 58. Brown T D, Johnston R C, Saltzman C L, Marsh J L, Buckwalter J A. Posttraumatic osteoarthritis: a first estimate of incidence, prevalence, and burden of disease. J Orthop Trauma 2006; 20(10): 739-44. Coetzee J C. Management of varus or valgus ankle deformity with ankle replacement. Foot Ankle Clin 2008; 13(3): 509-20. Deleu P A, Devos Bevernage B, Gombault V, Maldague P, Leemrijse T. Intermediate-term results of mobile-bearing total ankle replacement. Foot Ankle Int 2014 (36)5: 518-30. Demetracopoulos C A, Adams S B Jr, Queen R M, DeOrio J K, Nunley J A 2nd, Easley M E. Effect of age on outcomes in total ankle arthroplasty. Foot Ankle Int 2015; 36(8): 871-80. Hanna R S, Haddad S L, Lazarus M L. Evaluation of periprosthetic lucency after total ankle arthroplasty: helical CT versus conventional radiography. Foot Ankle Int 2007; 28 (8): 921-6. Haskell A, Mann R A. Perioperative complication rate of total ankle replacement is reduced by surgeon experience. Foot Ankle Int 2004; 25(5): 283-9.

449

Henricson, A, Ågren P-H. Secondary surgery after total ankle replacement: the influence of preoperative hindfoot alignment. Foot Ankle Surg 2007; (13): 41-4. Henricson A, Carlsson A, Rydholm U. What is a revision of total ankle replacement? Foot Ankle Surg 2011a; 17(3): 99-102. Henricson A, Nilsson J A, Carlsson A. 10-year survival of total ankle arthroplasties: a report on 780 cases from the Swedish Ankle Register. Acta Orthop 2011b; 82(6): 655-9. Hintermann B, Valderrabano V, Dereymaeker G D W. The HINTEGRA ankle: rationale and short-term results of 122 consecutive ankles. Clin Orthop Relat Res 2004; (424): 57-68. Johnson-Lynn S E, Ramaskandhan J, Siddique M S. The effect of patient age and diagnosis on the 5-year outcomes of mobile-bearing total ankle replacement. Foot (Edinb) 2018; 36: 1. Kakkar R, Siddique M S. Stresses in the ankle joint and total ankle replacement design. Foot Ankle Surg 2011; 17(2): 58-63. Labek G, Klaus H, Schlichtherle R, Williams A, Agreiter M. Revision rates after total ankle arthroplasty in sample-based clinical studies and national registries. Foot Ankle Int 2011; 32(8): 740-5. Lee G W, Wang S H, Lee K B. Comparison of intermediate to long-term outcomes of total ankle arthroplasty in ankles with preoperative varus, valgus, and neutral alignment. J Bone Joint Surg Am 2018; 100(10): 835-42. NZJR 2013. The New Zealand Joint Registry. Fifteen year report, January 1999 to December 2013. https://nzoa.org.nz/nzoa-joint-registry. (date last accessed 3 March 2019) NZJR 2018. The New Zealand Joint Registry. Twenty year report, January 1999 to December 2018. https://nzoa.org.nz/nzoa-joint-registry. (date last accessed 10 February 2020) Prissel M, Roukis T. Incidence of revision after primary implantation of the Scandinavian Total ankle replacement system: a systematic review. Clin Podiatr Med Surg 2013; 30(2): 237-50. Roukis T. Incidence of revision after primary implantation of the AgilityTM total ankle replacement system: a systematic review. J Foot Ankle Surg 2012; 51(2): 198-204. Roukis T, Berlet G, Bibbo C, Hyer C, Penner M, Wünschel M. Survivorship of first-, second-, and third-generation total ankle replacement systems. In: Primary and revision total ankle replacement. New York: Springer; 2016. Saltzman C L, Salamon M L, Blanchard G M, Huff T, Hayes A, Buckwalter J A. Epidemiology of ankle arthritis: report of a consecutive series of 639 patients from a tertiary orthopaedic center. Iowa Orthop J 2005; 25: 44-6. Skyttä E T, Koivu H E A. Total ankle replacement: a population-based study of 515 cases from the Finnish Arthroplasty Register. Acta Orthop 2010; 81(1): 114-18. The Swedish Ankle Registry Annual Report for 2014. www.swedankle.se (date last accessed 5 February 2020) UK National Joint Registry; 2019. NJR annual reports. http://www.njrreports. org.uk/ (date last accessed 10 February 2020) Willegger M, Trnka H J, Schuh R. The HINTEGRA® ankle arthroplasty: intermediate term results of 16 consecutive ankles and a review on the current literature. Clin Res Foot Ankle 2013; 2: 1. Yang H-Y, Wang S-H, Lee K-B. The HINTEGRA total ankle arthroplasty. Bone Joint J 2019; 101-B: 695-701.

450

Acta Orthopaedica 2020; 91 (4): 450–456

Effectiveness of hallux valgus surgery on patient quality of life: a systematic review and meta-analysis Luis Enrique HERNÁNDEZ-CASTILLEJO 1, Vicente MARTÍNEZ VIZCAÍNO 1,2, Miriam GARRIDO-MIGUEL 1, Iván CAVERO-REDONDO 1,3, Diana P POZUELO-CARRASCOSA 1, and Celia ÁLVAREZ-BUENO 1,3 1 Health and Social Research Center, Universidad de Castilla La Mancha, Cuenca, Spain; 2 Facultad de Ciencias de la Salud, Universidad Autónoma de Chile, Talca, Chile; 3 Universidad Politécnica y Artística del Paraguay, Asunción, Paraguay Correspondence: vicente.martinez@uclm.es Submitted 2019-11-18. Accepted 2020-04-08.

Background and purpose — The quality of life (QoL) of patients with hallux valgus (HV) usually improves postoperatively. Evidence regarding the effect of HV surgery on different domains of patient QoL remains inconclusive. This systematic review and meta-analysis estimates the effect of HV surgery on patient QoL through distinguishing effects on physical domains (comprising physical function and body pain domains) using the EuroQol-5D, short form (SF) health survey-12, and SF-36 QoL scales and a visual analogue scale (VAS) score and mental and social domains using QoL scales. Patients and methods — MEDLINE, EMBASE, Cochrane Library, and Web of Science databases were systematically searched from inception to March 2019 for studies on the effect of HV surgery on patient QoL. A standardized mean difference score was calculated for each specific QoL domain (mental, social, pain, physical, and VAS) using Cohen’s d index. The pooled effect size (ES) was estimated using a random-effects model based on the DerSimonian and Laird method. Results — From 12 published studies selected, the estimated pooled ES for QoL was 1.01 (95% confidence interval [CI] 0.52–1.51; I2 = 87%) for body pain and 0.43 (CI 0.31–0.55, I2 = 35%) for physical function. Regarding the composite mental and social domains of QoL, the pooled ES estimates were 0.24 (CI 0.00–0.47, I2 = 80%) and 0.42 (CI 0.21–0.63, I2 = 6.4%), respectively. The pooled difference in means for the VAS score was –4.1 ( CI –4.5 to –3.6, I2 = 90%). Interpretation — Our data showed that HV surgery decreased patients’ perceptions regarding pain. Furthermore, the data confirmed that HV surgery increased patients’ QoL, particularly concerning physical and social domains.

More than 200 different surgical procedures have been developed for treating hallux valgus (HV) (Myerson 2000, Magnan et al. 2005, Easley and Trnka 2007). Evidence supporting these differing surgical approaches for HV remains inconclusive; therefore, patient-reported outcome measures (PROMs) could be decisive for favoring one approach over another among these numerous surgical alternatives. PROMs are typically classified as pain scales, general scales, and region-specific outcomes. For region-specific PROMs, the Manchester-Oxford Foot Questionnaire (MOXFQ), the Foot and Ankle Outcome Score (FAOS), the Self-reported Foot and Ankle Score (SEFAS) (Schrier et al. 2015), and the American Orthopaedic Foot and Ankle Society (AOFAS) score (Hunt and Hurwit 2013, Arbab et al. 2019, Nilsdotter et al. 2019) have been developed and validated. Generic PROMs, including scales such as the EuroQol-5D (EQ-5D), the Short Form12 Health Survey (SF-12), and the SF-36, have also been used. While surgical treatment for HV is more effective than nonoperative or no treatment (Klugarova et al. 2017), there are no substantial differences among surgical alternatives in terms of pain or other region-specific outcomes such as recurrence rates or nerve injury. However, there are few studies reporting results on HV surgery in terms of quality of life (QoL), and the effectiveness of HV surgery on different domains in terms of patient QoL remains inconclusive. This systematic review and meta-analysis aimed to estimate the effect of HV surgery on patient QoL through distinguishing physical domains (including physical function and body pain domains) using the QoL scales EQ-5D, SF-12, and SF-36, and visual analogue scale (VAS) scores, and mental and social domains using QoL scales.

Acta Orthopaedica 2020; 91 (4): 450–456

Patients and methods This systematic review and meta-analysis was undertaken following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (Moher et al. 2010), Statement, and guided by the Cochrane Collaboration Handbook (Higgins and Green 2008). Search strategy The electronic databases MEDLINE (via PubMed), EMBASE, Cochrane Library, and Web of Science were systematically searched to identify relevant publications. The systematic search included the following keywords: “hallux valgus”, foot, ankle, metatarsal, surgery, osteotomy, “quality of life”, SF-12, FAOS, SF-36, SEFAS, VAS, EQ-5D, MOXFQ, AOFAS, LLFI, NRS, and SAFE-Q. The search was conducted from database inception to March 2019. The complete search strategy for MEDLINE is reported in Table 1 (see Supplementary data). Eligibility Studies on the effect of HV surgery on patient QoL measured through QoL scales other than the AOFAS were included. The inclusion criteria were as follows: (i) patients > 16 years old; (ii) open (i.e., Chevron, Scarf, Kramer, or Bösch) or minimally invasive surgery procedures; (iii) QoL reports on physical (including physical function and body pain domains of QoL scales, and a VAS score), mental and social domains (including social function, emotional role, mental health, vitality, and general health) using QoL scales other than the AOFAS; (iv) randomized controlled trials (RCTs), non-randomized experimental studies, and single-arm pre–post studies, and; (v) reports written in English or Spanish. We excluded studies reporting data on foot or ankle pathologies other than HV, as well as patients who had undergone HV revision surgery. Data extraction and risk of bias assessment The following data were extracted from the included studies: (1) author, (2) year of publication, (3) country of study, (4) number of participants according to sex (in the control and intervention groups), (5) mean age, (6) QoL scale and domains reported, (7) type of intervention, and; (8) end-point measures (in months). After concealing information regarding authors, affiliations, date, and source of each manuscript, 2 investigators (CA-B and LH-C) independently assessed the risk of bias of included studies, and inconsistencies were resolved through consensus or through consulting a third researcher (VM-V). The Quality Assessment Tool for Observational Cohort and Cross-sectional Studies from the National Heart, Lung, and Blood Institute (NIH) (National Heart, Lung, and Blood Institute, 2019) was used to assess the risk of bias of the pre–post studies. This tool evaluates 7 domains including selection bias,

451

study design, confounders, blinding, data collection method, withdrawals, and dropouts. Each domain could be considered as strong, moderate, or weak (Table 3, Supplementary data). The Cochrane Risk of Bias tool for randomized trials (RoB 2) (a revised tool to assess risk of bias in randomized trials) was used to assess the risk of bias in the RCTs. This tool evaluates 6 domains: randomization, assignment to intervention, adherence to intervention, missing outcome data, measurement of the outcome, and selection of the reported results. Each domain could be considered to have low bias concerns, some bias concerns, or a high risk of bias (Table 4, Supplementary data). Statistics A standardized mean difference score was calculated for each specific QoL reported, using Cohen’s d index (Cohen 1977), in which positive ES values indicated higher outcome scores in favor of the intervention group. Cohen’s d values were interpreted as follows: 0.2 (considered a weak effect), 0.5 (considered a moderate effect), 0.8 (considered a strong effect), and > 1.0 (considered a very strong effect). Pooled ES estimates for pre–post interventions and their 95% confidence intervals (CIs) were estimated using a random-effects model based on the DerSimonian and Laird method (DerSimonian and Kacker 2007) through distinguishing physical domains (including body pain and physical function) using QoL scales, and mental and social domains. For the VAS score, the pooled difference in means for pre–post interventions and their CIs were calculated, in which negative values indicated better VAS scores in favor of the intervention group. Heterogeneity across studies was assessed using the I2 statistic (Higgins and Thompson 2002), with values of < 30%, ≤ 30% to < 60%, ≤ 60% to 85%, and > 85% considered as indicating not important, moderate, substantial, and considerable heterogeneity, respectively (Higgins and Green 2008). Some considerations should be noted. First, only studies providing complete data for pre- and post-intervention measurements were included in the meta-analysis. Second, when studies provided 2 measures for the same domain, both measures were pooled to calculate the mean ES. Third, data from different cohorts were considered as independent samples. Finally, when studies reported data for several follow-up points, only the data in relation to the longest follow-up were considered. The influence of each study in the pooled ES estimates was examined using sensitivity analyses. Additionally, metaregression analyses were conducted to assess the influence of mean age and the percentage of females on the magnitude of the effect of HV surgery on QoL domains. Egger’s regression asymmetry test (Sterne et al. 2001, Tanner-Smith et al. 2019) was used to assess publication bias. The significance value of the pooled ES was estimated based on a 95% CI. Statistical analyses were performed using STATA SE software, version 15 (StataCorp, College Station, TX, USA).

452

Acta Orthopaedica 2020; 91 (4): 450–456

Table 2. Characteristics of the studies included in this systematic review and meta-analysis Reference (country)

Sample size Mean women men age Scales

Preoperative mean (95% CI or SD)

Postoperative mean (95% CI or SD)

MOXFQ: 74 (30–100) • Walking: 75 (29–100) • Pain: 79 (35–100) • Social: 67 (25–100)

MOXFQ: 13 (0–61) • Walking: 12 (0–66) • Pain: 16 (0–69) • Social: 11 (0-69)

14–60

• VAS: 5 (4–7) • PCS: 50 (39–54) • MCS: 53 (46–60) • VAS: 6 (4–8) • PCS: 45 (35–52) • MCS: 49 (42–61)

• VAS: 0 (0–4) • PCS: 51 (43–54) • MCS: 56 (49–60) • VAS: 0 (0–3) • PCS: 53 (46–59) • MCS: 48 (42–52)

6–24

• VAS: 5 (4–5) • PCS: 55 (53–57) • MCS: 55 (53–56) • VAS: 5 (5–6) • PCS: 49 (44–54) • MCS: 53 (50–56)

• VAS: 1 (0–1) • PCS: 85 (84–87) • MCS: 55(54–56) • VAS: 1 (0–2) • PCS: 84 (79–88) • MCS: 54 (51–57)

Choi et al. 2013 (USA) 48 3 59 VAS SF-36

VAS: 5.8 (1.9) SF-36: • Physical: 46 (8.9) • Mental: 55 (6.8)

VAS: 1.1 (1.4) SF-36: • Physical: 52 (7.3) • Mental: 55 (6.9)

Dawson et al. 2006 (UK) 95 5 50 MOXFQ SF-36

MOXFQ: • Foot pain: 53 (SD 21) • Walking: 45 (SD 25) • Social: 47 (SD 23) SF 36: • Pain: 62 (SD 24) • Physical: 75 (SD 23) • Role P: 75 (SD 27) • Mental: 71 (SD 17) • Role M: 83 (SD 23) • Vitality: 57 (SD 20) • Social: 78 (SD 23) • Health: 76 (SD 19)

MOXFQ: • Foot pain: 20 (SD 21) • Walking: 16 (SD 23) • Social: 12 (SD 19) SF 36: • Pain: 77 (SD 21) • Physical: 85 (SD 19) • Role P: 86 (SD 25) • Mental: 78 (SD 16) • Role M: 91 (SD 18) • Vitality: 63 (SD 18) • Social: 85 (SD 21) • Health: 80 (SD 17)

Hogea et al. 2017 (Romanian) 35 21 44.4 VAS EQ5-D Kaufmann et al. 2018 (Austria) Open 19 3 44 VAS Percutaneous 21 4 52

VAS: 59 (SD 31) EQ5-D: • Anxiety: 1.9 (SD 0.65) • Usual activities: 2.9 (SD 0.63) • Self-care: 1.9 (SD 0.82) • Mobility: 2.6 (SD 0.33) • Pain: 2.9 (SD 0.94)

VAS: 20 (SD 23) 24–60 EQ5-D: • Anxiety: 1.0 (SD 0.40) • Usual activities: 2.9 (SD 0.55) • Self-care: 1.9 (SD 0.62) • Mobility: 1.5 (SD 0.32) • Pain: 1.7 (SD 0.65)

VAS: 6 VAS: 5

VAS: 0 1.5–9 VAS: 1

VAS: 4.9 (SD 2.6) SF-36: • Physical: 82 (SD 19) • Mental: 86 (SD 15) VAS: 4.0 (SD 2.9) SF-36: • Physical: 76 (SD 22) • Mental: 79 (SD 18)

VAS: 0.4 (SD 1.5) 6–24 SF-36: • Physical: 83 (SD 20) • Mental: 86 (SD 15) VAS: 0.7 (SD 1.9) SF-36: • Physical: 83 (SD 22) • Mental: 85 (SD 15)

VAS: 6.9 (SD 1.7) VAS: 7.1 (SD 1.5)

VAS: 0.5 (SD 1.1) VAS: 0.3 (SD 0.9)

VAS: 5 (4–6) VAS: 5 (4–6)

VAS: 2 (1–3) VAS: 2 (1–2)

Al Nammari et al. 2015 (UK) 43 4 56 MOXFQ Chen et al. 2016 (Singapore) Mild residual pain 65 5 52 VAS PCS MCS Several residual pain 19 1 Chen et al. 2015 (Singapore) Control 375 28 51 VAS PCS MCS Obese 44 5 55

Lai et al. 2017 (Singapore) Open 52 6 54 VAS SF-36 Percutaneous 25 4 Lee et al. 2017 (Australia) Open 22 3 53.4 VAS Percutaneous 23 2 52.6 Milczarek et al. 2017 (Poland) Normal BMI High BMI

71 (W+M) 62 (W+M)

52 VAS 61

Follow up (months)

12–24 12

Acta Orthopaedica 2020; 91 (4): 450–456

453

Table 2. Continued Reference (country)

Sample size Mean women men age Scales

Preoperative mean (95% CI or SD)

Postoperative mean (95% CI or SD)

Follow up (months)

Niki et al. 2017 (Japa)n 92 8 62 SF-36 SAFE-Q

SF-36: • Pain: 53 (SD 23) • Physical: 70 (SD 23) • Role P: 70 (SD 28) • Mental: 67 (SD 20) • Role M: 67(SD 20) • Vitality: 56 (SD 19) • Social: 76 (SD 25) • Health: 57 (SD 20) SAFE-Q: • Pain: 59 (SD 24) • Physical: 71 (SD 25) • Social: 68 (SD 30) • Shoe-related: 37 (SD 20) • General health: 64 (SD 27)

SF-36: 9–12 • Pain: 74 (SD 21) • Physical: 80 (SD 20) • Role P: 85 (SD 18) • Mental: 71 (SD 19) • Role M: 87 (SD 18) • Vitality: 64 (SD 17) • Social: 85 (SD 21) • Health: 61 (SD 18) SAFE-Q: • Pain: 87 (SD 13) • Physical: 87 (SD 16) • Social: 90 (SD 16) • Shoe-related: 71 (SD 23) • General health: 90 (SD 15)

Saro et al. 2007 (Sweden) 94 6 48 SF-36

SF-36: • Pain: 58 (SD 22) • Physical: 83 (SD 17) • Role P: 77 (SD 37) • Mental: 77 (SD 18) • Role M: 81 (SD 34) • Vitality: 62 (SD 23) • Social: 83 (22) • Health: 77 (SD 20)

SF-36: • Pain: 75 (SD 24) • Physical: 86 (SD 19) • Role P: 86 (SD 31) • Mental: 85 (SD 16) • Role M: 91 (SD 26) • Vitality: 71 (SD 22) • Social: 89 (SD 21) • Health: 85 (SD 16)

Abbreviations: W: Women; M: Men, VAS: visual analogical scale, SF: Short Form-36 Health Survey, MOXFQ: Manchester-Oxford Foot Questionnaire, PCS: Physical Component Score, MCS: Mental Component, Score, SD: Standard deviation. EQ5-D: EuroQol-5D, BMI: Body-mass index, SAFE-Q: Self-administered foot evaluation Questionnaire

Ethics, funding, and potential conflicts of interest The protocol for this systematic review and meta-analysis was registered on PROSPERO (Registration number: CRD42019121120). There is no funding. There are no potential conflicts of interest.

Results Systematic review A literature search retrieved 3,924 studies and, after exclusion of non-relevant studies, 47 full-text studies were assessed for eligibility. Of these, 12 studies were included in this systematic review and meta-analysis (Figure 1, Supplementary data). Table 2 presents data from the included studies, which had been published between 2006 and 2018, and included 1 RCT and 10 single-arm pre–post studies. These studies had been conducted in 9 countries (the United Kingdom, Singapore, the United States, Romania, Austria, Australia, Poland, Japan, and Sweden), and comprised 1,313 patients who had undergone HV surgery, and whose ages ranged from 44 to 62 years. Endpoint measurements were obtained from 4 weeks to 5 years. The evaluation of the clinical outcomes was performed using the following tools: MOXFQ (Dawson et al. 2006, Al-Nammari et al. 2015), VAS (Dawson et al. 2006, Choi et al. 2013, Chen et

al. 2015, 2016, Hogea et al. 2017, Lee et al. 2017, Milczarek et al. 2017, Lai et al. 2018, Kaufmann et al. 2019), the SF-36 (Saro et al. 2007, Choi et al. 2013), EuroQol-5D (EQ5-D) (Kaufmann et al. 2019), and the SAFE-Q (Niki et al. 2017). Risk of bias According to the Quality Assessment Tool for Observational Cohort and Cross-sectional Studies, all pre–post studies included in the meta-analysis were considered as having a moderate risk of bias. When individual domains using the tool were assessed, all the studies included information relating to the representativeness of the sample and the description of the intervention; however, all the studies were limited in terms of sample eligibility and blinding (Table 3, Supplementary data). The study, analyzed using a RoB 2.0 tool (Cochrane Collaboration n.d.) for randomized trials, was categorized as having a high risk of bias (Table 4, Supplementary data). Meta-analysis The pooled ES estimates for the QoL physical domain scores were 1.01 (CI 0.52–1.51; I2 = 87%) for body pain and 0.43 (CI 0.31–0.55, I2 = 35%) for physical function (Figure 2). Concerning the mental and social domains of QoL, the pooled ES estimates were 0.24 (CI 0.00–0.47, I2 = 80%) and 0.42 (CI 0.21–0.63, I2 = 6.4%), respectively (Figure 2).

454

Acta Orthopaedica 2020; 91 (4): 450–456

Reference

Effect size (95% CI)

Prediction Weight interval %

BODY PAIN Al Nammari et al. 2015 Dawson et al. 2006 Niki et al. 2017 Saro et al. 2007 Subtotal (I2 = 87.1%, p = 0.000)

0.5 (0.1 to 0.9) 1.6 (1.3 to 1.9) 1.2 (0.8 to 1.8) 0.7 (0.5 to 1.0) 1.0 (0.5 to 1.5)

(0.3 to 0.8) (1.3 to 1.9) (0.9 to 1.5) (0.4 to 1.0) (0.5 to 1.5)

24.4 26.3 22.9 26.3 100

PHYSICAL FUNCTION Chen et al. 2015 (Control) Chen et al. 2015 (Obese) Chen et al. 2016 (Mild residual pain) Chen et al. 2016 (Severe residual pain) Choi et al. 2013 Dawson et al. 2006 Lai et al. 2017 (Open) Lai et al. 2017 (Percutaneous) Niki et al. 2017 Saro et al. 2007 Subtotal (I2 = 34.8%, p = 0.1)

0.4 (0.3 to 0.5) 0.6 (0.2 to 1.0) 0.4 (0.1 to 0.7) 0.6 (0.3 to 1.0) 0.7 (0.3 to 1.1) 0.5 (0.2 to 0.8) 0.0 (–0.4 to 0.4) 0.3 (–0.2 to 0.8) 0.6 (0.3 to 0.9) 0.2 (–0.1 to 0.5) 0.4 (0.3 to 0.6)

(0.1 to 0.7) (0.3 to 0.9) (0.1 to 0.7) (0.5 to 0.7) (0.5 to 1.0) (0.2 to 0.8) (–0.7 to 0.3) (0.0 to 0.6) (0.3 to 0.9) (–0.1 to 0.5) (0.3 to 0.6)

21.9 7.1 9.2 9.0 7.2 12.2 8.1 4.7 9.8 11.1 100

MENTAL DOMAIN Chen et al. 2015 (Control) Chen et al. 2015 (Obese) Chen et al. 2016 (Mild residual pain) Chen et al. 2016 (Severe residual pain) Choi et al. 2013 Dawson et al. 2006 Lai et al. 2017 (Open) Lai et al. 2017 (Percutaneous) Niki et al. 2017 Saro et al. 2007 Subtotal (I2 = 79.5%, p = 0.000)

0.0 (–0.1 to 0.1) 0.1 (–0.3 to 0.5) 2.0 (1.2 to 2.7) –0.6 (–1.3 to 0.0) –0.0 (–0.4 to 0.4) 0.4 (0.2 to 0.7) 0.0 (–0.4 to 0.4) 0.4 (–0.1 to 0.9) 0.2 (–0.1 to 0.5) 0.5 (0.2 to 0.8) 0.2 (0.0 to 0.5)

(–0.3 to 0.3) 13.4 (–0.2 to 0.4) 10.0 (1.3 to 2.6) 5.7 (–0.9 to –0.4) 6.9 (–0.3 to 0.2) 10.3 (0.2 to 0.7) 12.1 (–0.3 to 0.3) 10.5 (0.1 to 0.7) 8.3 (–0.1 to 0.5) 11.3 (0.2 to 0.8) 11.6 (–0.0 to 0.5) 100

and Choi et al. studies (2013; Table 5, Supplementary data). Additionally, meta-regressions showed that only the percentage of females affected the association between HV surgery and the VAS score (Table 6), Supplementary data. Publication bias Evidence of publication bias was identified using funnel plot asymmetry and Egger’s test for the effect of HV surgery on the VAS score (p = 0.02) (Figure 4, Supplementary data).

Discussion

This systematic review and meta-analysis aimed to estimate the effect of HV surgery on patient QoL through distinguishing the effects on the physical domain and VAS SOCIAL DOMAIN scores, and mental and social domains, Al Nammari et al. 2015 0.4 (0.0 to 0.8) (0.2 to 0.7) 26.1 using QoL scales. Our findings showed Dawson et al. 2006 0.7 (–1.2 to 2.6) (0.4 to 1.0) 1.2 Niki et al. 2017 0.7 (0.3 to 1.1) (0.4 to 1.0) 24.9 that HV surgery resulted in decreased Saro et al. 2007 0.3 (–0.0 to 0.6) (0.2 to 0.8) 47.8 body pain and improved physical funcSubtotal (I = 6.4%, p = 0.4) 0.4 (0.2 to 0.6) (0.3 to 0.7) 100 tion, and improved the social domain of –2 –1 0 1 2 3 patient QoL, although it did not modify Figure 2. Forest plot for ES of the body pain and physical function components, and mental and the mental QoL domain score. social domains. Most QoL scales include patients’ perceptions of different domains of physical Mean Prediction Weight Reference difference (95% CI) interval % health status. Although the validity and Chen et al. 2015 (Control) –4.0 (–4.7 to –3.3) (–0.4 to –2.6) 8.0 reliability of the VAS and QoL scales used Chen et al. 2015 (Obese) –4.0 (–5.1 to –2.9) (–4.8 to –3.2) 6.1 in the included studies have been conChen et al. 2016 (Mild residual pain) –4.3 (–4.6 to –3.9) (–4.6 to –4.0) 9.4 Chen et al. 2016 (Severe residual pain) –5.3 (–5.9 to –4.6) (–5.6 to –5.0) 8.2 firmed (Schrier et al. 2015), the numerous Choi et al. 2013 –5.1 (–5.3 to –4.1) (–5.5 to –4.6) 8.3 Hogea et al. 2017 –2.6 (–3.4 to –1.7) (–3.2 to –1.9) 7.5 surgical approaches and the small number Kaufmann et al. 2018 (Percutaneous) –4.0 (–5.0 to –3.0) (–4.5 to –3.5) 6.5 of studies meeting the inclusion criteria Kaufmann et al. 2018 (Open) –6.0 (–7.2 to –4.8) (–6.6 to –5.4) 5.6 Lai et al. 2017 (Open) –4.5 (–5.3 to –3.7) (–5.1 to –3.9) 7.7 could have affected our results. FurtherLai et al. 2017 (Percutaneous) –3.3 (–4.6 to –2.0) (–4.0 to –2.6) 5.6 more, because of the scarcity of studies, Lee et al. 2017 (Percutaneous Chevron/Akin) –4.6 (–6.5 to –2.7) (–5.5 to –3.7) 3.6 Lee et al. 2017 (Open Scarf/Akin) –3.8 (–5.6 to –2.1) (–4.7 to –2.9) 3.9 we could not undertake stratified analyses Milczarek et al. 2017 (Normal BMI) –3.0 (–3.2 to –2.8) (–3.2 to –2.8) 9.8 involving differing surgical techniques Milczarek et al. 2017 (High BMI) –3.3 (–3.4 to –3.1) (–3.4 to –3.1) 9.9 Overall (I = 90.1%, p = 0.000) –4.1 (–4.5 to –3.6) (–4.6 to –3.6) 100 to examine subsequent differences in the –7.5 –7 –6 –5 –4 –3 –2 –1 0 effects on QoL and pain. However, our findings show that HV surgery had a posiFigure 3. Forest plot for the difference in means of the visual analogue scale (VAS). tive effect on the physical domain, regardless of the scale used to measure it. Finally, the pooled difference in means for the VAS score Anxiety and depression are common disorders that, simiwas –4.1 (CI –4.5 to –3.6, I2 = 90%) (Figure 3). lar to HV, have been reported to be more prevalent in women (Shakked et al. 2018). Moreover, because those with anxiety Sensitivity analysis and meta-regressions and/or depression have a greater percentage of severe deforWhen the effect of individual studies was examined by remov- mities than those unaffected by these mental disorders, they ing studies from the analysis one at a time, only the estimate score lower in the mental domain of QoL at baseline (Cody for the mental domain of QoL was modified after removing et al. 2018, Shakked et al. 2018). Our findings, which are the samples from the Chen et al. (2016; severe residual pain) in line with previous studies (Dawson et al. 2006, Chen et 2

Acta Orthopaedica 2020; 91 (4): 450–456

al. 2016), suggest no improvement after HV surgery in the mental domain of QoL. Because most included studies did not report the percentage of participants with a diagnosis of these mental disorders, it is not possible to know whether anxiety and/or depressive disorders explain this lack of improvement in the mental health domain of QoL following HV surgery. It is important that HV surgery be approached in a multidisciplinary and multidimensional manner (Hogea et al. 2017), especially for patients with comorbidities and high risks concerning mental health, such as depression, or general health issues. Our assessment of multiple domains of QoL scales aimed to objectively measure the effectiveness of HV surgery on patient daily life. Each domain included in these scales (physical, mental, and social) is related to QoL and these are closely related to each other. To ensure an appropriate patient evaluation and follow-up after HV surgery, besides the AOFAS questionnaire commonly used by orthopedic surgeons, the use of different QoL scales is encouraged (Thordarson et al. 2005, Fraissler et al. 2016). It is not advisable to use a single instrument to collect quality orthopedic data as selection is dependent on the population being examined and the questions being asked, and all PROMs should be appropriately referenced in the studies. We suggest that for patient-reported evaluation of HV surgery, one of the region-specific validated PROMs (such as the MOXFQ, FAOS, and SEFAS) would be a good option and should be used in combination with a generic PROMs tool measuring QoL (such as EQ-5D, SF-12, and SF-36) (Schrier et al. 2015, Kitaoka et al. 2018, Arbab et al. 2019, Nilsdotter et al. 2019). A recent systematic review (Barg et al. 2018), which reviewed 229 articles that included 16,237 surgeries, found unfavorable outcomes of surgical treatment for HV deformity and highlighted the limited quality of the evidence provided by the published studies involving predominantly retrospective case series, and the lack of sufficient studies to conduct analyses according to type of surgery. Moreover, although less-invasive procedures for HV have generated great enthusiasm, another systematic review (Bia et al. 2018), which aimed to synthesize the clinical evidence for percutaneous HV surgery, reported inconclusive results. Our meta-regression analyses using mean age and percentage of females showed no changes in the pooled ES. Previous findings in this regard have been inconclusive (Chou et al. 2008). While some authors state that age is not a significant predictor for pain scores (Chou et al. 2008, Fernández et al. 2017), others state that age is related to some subscale scores (Niki et al. 2013). Furthermore, although the prevalence of HV is higher among females, the predictive effect of HV surgery does not appear to depend on patients’ sex, age, or severity of HV (Chou et al. 2008, Fraissler et al. 2016). Our study had some limitations. First, the study had limitations common to systematic reviews and meta-analyses such as selection bias and limited availability of complete information

455

from study reports. Second, although there was no evidence of publication bias from Egger’s test in most of the domains analyzed, results from unpublished studies could have modified the results of our meta-analysis. Third, several factors could have influenced both clinician- and patient-related outcomes, such as the type of surgical approach, surgeons’ skills, and comorbidities (Cöster et al. 2014), and these were not controlled for in the analyses owing to the scarcity of information in the included studies. Fourth, we were unable to establish cause–effect inferences due to the nature of the observational studies selected. Finally, language selection bias could not be ruled out since studies published only in English and Spanish were included. In conclusion, this systematic review and meta-analysis provided a synthesis of the evidence that HV surgery decreased patient perception of pain. Furthermore, our data showed that HV surgery increased patient QoL, especially in the physical and social domains. This study highlights the need to include PROM measures, such as pain or QoL, through region-specific validated scales prior to and post-HV surgery. Supplementary data Tables 1 and 3–6 and Figures 1 and 4 are available as supplementary data in the online version of this article, http://dx.doi. org/10.1080/17453674.2020.1764193

LEH: conception of study, interpretation of data, and manuscript preparation. MG, IC, DP: interpretation of data and manuscript preparation. VM, CA: statistical analyses, interpretation of data, and manuscript preparation. Acta thanks Maria C Cöster and lka Kamrad help with peer review of this study.

Arbab D, Kuhlmann K, Schnurr C, Lüring C, König D, Bouillon B. Comparison of the Manchester-Oxford Foot Questionnaire (MOXFQ) and the SelfReported Foot and Ankle Outcome Score (SEFAS) in patients with foot or ankle surgery. Foot Ankle Surg 2019; 25(3): 361-5. Al-Nammari S S, Christofi T, Clark C. Double first metatarsal and Akin osteotomy for severe hallux valgus. Foot Ankle Int 2015; 36(10): 1215-22. Barg A, Harmer J R, Presson A P, Zhang C, Lackey M, Saltzman C L. Unfavorable outcomes following surgical treatment of hallux valgus deformity: a systematic literature review. J Bone Joint Surg Am 2018; 100(18): 156373. Bia A, Guerra-Pinto F, Pereira B S, Corte-Real N, Oliva X M. Percutaneous osteotomies in hallux valgus: a systematic review. J Foot Ankle Surg 2018; 57(1): 123-30. Chen J Y, Lee M J, Rikhraj K, Parmar S, Chong H C, Yew A K, Koo K O, Singh Rikhraj I. Effect of obesity on outcome of hallux valgus surgery. Foot Ankle Int 2015; 36(9): 1078-83. Chen J Y, Ang B F, Jiang L, Yeo N E, Koo K, Singh Rikhraj I. Pain resolution after hallux valgus surgery. Foot Ankle Int 2016; 37(10): 1071-1075. Choi J H, Zide J R, Coleman S C, Brodsky J W. Prospective study of the treatment of adult primary hallux valgus with Scarf osteotomy and soft tissue realignment. Foot Ankle Int 2013; 34(5): 684-90.

456

Chou L B, Wagner D, Witten D M, Martinez-Diaz G J, Brook N S, Toussaint M, Carroll I R. Postoperative pain following foot and ankle surgery: a prospective study. Foot Ankle Int 2008; 29(11): 1063-8. Cochrane Collaboration. A revised tool to assess risk of bias in randomized trials (RoB 2). Available online from https://methods.cochrane.org/riskbias-20-tool. Accessed January 20, 2019. Cody E A, Do H T, Koltsov J C B, Mancuso C A, Ellis S J. Influence of diagnosis and other factors on patients’ expectations of foot and ankle surgery. Foot Ankle Int 2018; 39(6): 641-8. Cohen J. Statistical power analysis for the behavioral sciences. Rev. ed. New York: Academic Press; 1977. Cöster M C, Rosengren B E, Bremander A, Brudin L, Karlsson M K. Comparison of the Self-reported Foot and Ankle Score (SEFAS) and the American Orthopedic Foot and Ankle Society Score (AOFAS). Foot Ankle Int 2014; 35(10): 1031-6. Dawson J, Coffey J, Doll H, Lavis G, Cooke P, Herron M, Jenkinson C. A patient-based questionnaire to assess outcomes of foot surgery: validation in the context of surgery for hallux valgus. Qual Life Res 2006; 15(7): 1211-22. DerSimonian R, Kacker R. Random-effects model for meta-analysis of clinical trials: an update. Contemp Clin Trials 2007; 28(2): 105–14. Easley M E, Trnka H J. Current concepts review: Hallux valgus part II: operative treatment. Foot Ankle Int 2007; 28(6): 748-58. Fernández O, Sánchez J, Maderuelo J A, Benito E, Ramos L R. Resultados del tratamiento quirúrgico del hallux valgus. Análisis de su influencia sobre la calidad de vida de los pacientes. Rev Pie Tobillo 2017; 31(1): 22-33. Fraissler L, Konrads C, Hoberg M, Rudert M, Walcher M. Treatment of hallux valgus deformity. EFORT Open Rev 2016; 1(8): 295–302.38. Higgins J P, Green S, editors. Cochrane handbook for systematic reviews of interventions. Chichester, UK: Wiley-Blackwell; 2008. Higgins J P T, Thompson S G. Quantifying heterogeneity in a meta-analysis. Stat Med 2002; 2 (11): 1539-58. Hogea L M, Hogea B G, Nussbaum L A, Grigoraş M L, Andor B C, Levai C M, Bredicean A C. Health-related quality of life in patients with hallux valgus. Rom J Morphol Embryol 2017; 58(1): 175-9. Hunt K J, Hurwit D. Use of patient-reported outcome measures in foot and ankle research. J Bone Joint Surg Am 2013; 95(16): e1181-89. Kaufmann G, Dammerer D, Heyenbrock F, Braito M, Moertlbauer L, Liebensteiner M. Minimally invasive versus open Chevron osteotomy for hallux valgus correction: a randomized controlled trial. Int Orthop 2019; 43(2): 343-50. Klugarova J, Hood V, Bath-Hextall F, Klugar M, Mareckova J, Kelnarova Z. Effectiveness of surgery for adults with hallux valgus deformity: a systematic review. JBI Database System Rev Implement Rep 2017; 15(6): 1671-1710. Kitaoka H B, Meeker J E, Phisitkul P, Adams S B Jr, Kaplan J R, Wagner E. AOFAS position statement regarding patient-reported outcome measures. Foot Ankle Int 2018; 39(12): 1389-93.

Acta Orthopaedica 2020; 91 (4): 450–456

Lai M C, Rikhraj I S, Woo Y L, Yeo W, Ng Y C S, Koo K. Clinical and radiological outcomes comparing percutaneous Chevron-Akin osteotomies vs open Scarf-Akin osteotomies for hallux valgus. Foot Ankle Int 2018; 39(3): 311-17. Lee M, Walsh J, Smith M M, Ling J, Wines A, Lam P. Hallux valgus correction comparing percutaneous Chevron/Akin (PECA) and open Scarf/Akin osteotomies. Foot Ankle Int 2017; 38(8): 838-46. Magnan B, Pezzè L, Rossi N, Bartolozzi P. Percutaneous distal metatarsal osteotomy for correction of hallux valgus. J Bone Joint Surg Am 2005; 87: 1191–9. Milczarek M A, Milczarek J J, Tomasik B, Łaganowski P, Nowak K I, Domżalski M. Being overweight has limited effect on SCARF osteotomy outcome for hallux valgus correction. Int Orthop 2017; 41(4): 765-72. Moher D, Liberati A, Tetzlaff J, Altman D G; PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Int J Surg 2010; 8: 336–41. Myerson M S. Hallux valgus. In: Meyerson MS, editor. Foot and ankle disorders. Philadelphia: W B Saunders; 2000. p. 213–89. National Heart, Lung and Blood Institute. Quality assessment tool for observational cohort and cross-sectional studies. Available online from https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools. Accessed January 21, 2019. Niki H, Tatsunami S, Haraguchi N, Aoki T, Okuda R, Suda Y, Takao M, Tanaka Y. Validity and reliability of a self-administered foot evaluation questionnaire (SAFE-Q). J Orthop Sci 2013; 18(2): 298-320. Niki H, Haraguchi N, Aoki T, Ikezawa H, Ouchi K, Okuda R, Kakihana M, Shima H, Suda Y, Takao M, Tanaka Y, Watanabe K, Tatsunami S. Responsiveness of the Self-Administered Foot Evaluation Questionnaire (SAFEQ) in patients with hallux valgus. J Orthop Sci 2017; 22(4): 737-42. Nilsdotter A K, Cöster M E, Bremander A, Cöster M C. Patient-reported outcome after hallux valgus surgery: a two year follow up. Foot Ankle Surg 2019; 25(4): 478-81. Saro C, Jensen I, Lindgren U, Felländer-Tsai L. Quality-of-life outcome after hallux valgus surgery. Qual Life Res 2007; 16(5): 731-8. Schrier J C, Palmen L N, Verheyen C C, Jansen J, Koëter S. Patient-reported outcome measures in hallux valgus surgery: a review of literature. Foot Ankle Surg 2015; 21(1): 11-15. Shakked R, McDonald E, Sutton R, Lynch M K, Nicholson K, Raikin S M. Influence of depressive symptoms on hallux valgus surgical outcomes. Foot Ankle Int 2018; 39(7): 795-800. Sterne J A, Egger M, Smith G D. Investigating and dealing with publication and other biases in meta-analysis. BMJ 2001; 323: 101-5. Tanner-Smith E E, Egger M, Higgins J. Systematic reviews in health research. Res Synth Methods 2019; 10(3): 310-11. Thordarson D, Ebramzadeh E, Moorthy M, Lee J, Rudicel S. Correlation of hallux valgus surgical outcome with AOFAS forefoot score and radiological parameters. Foot Ankle Int 2005; 26(2): 122-7.

Acta Orthopaedica 2020; 91 (4): 457–463

457

Fate of patients with slipped capital femoral epiphysis (SCFE) in later life: risk of obesity, hypothyroidism, and death in 2,564 patients with SCFE compared with 25,638 controls Yasmin D HAILER

Section of Orthopedics, Department of Surgical Sciences, Uppsala University, Sweden Correspondence: Yasmin.hailer@surgsci.uu.se Submitted 2019-08-08. Accepted 2020-03-09.

Background and purpose — Associations between obesity and slipped capital femoral epiphysis (SCFE) during adolescence are described; however, few studies report on the lifetime risk of obesity in patients with SCFE. In addition, with the obesity epidemic in children and adolescents, an increasing incidence of SCFE might be expected. An association of SCFE with hypothyroidism seems ambiguous, and the association between SCFE and depression and all-cause mortality has not yet been evaluated. This study investigates the associations of SCFE with obesity, hypothyroidism, depression, and mortality, and putative changes in the yearly incidence of SCFE. Patients and methods — 2,564 patients diagnosed with SCFE at age 5–16 diagnosed between 1964 and 2011 were identified in the Swedish Patient Register. These were matched for age, sex, and residency with unexposed control individuals. Cox regression models were fitted to estimate the risk of obesity, hypothyroidism, depression, and death, in exposed compared with unexposed individuals. Results — The risk of obesity (HR 9, 95% CI 7–11) and hypothyroidism (HR 3, CI 2–4) was higher in SCFE patients compared with controls. There was no increase in the risk of developing depression (HR 1, CI 1–1.3) in SCFE patients. In contrast, all-cause mortality was higher in SCFE patients than in controls (HR 2, CI 1–2). The incidence of SCFE did not increase over the past decades. Interpretation — Patients with SCFE have a higher lifetime risk of obesity and hypothyroidism and a higher risk of all-cause mortality compared with individuals without SCFE. These findings highlight the lifetime comorbidity burden of patients who develop SCFE in childhood, and increased surveillance of patients with a history of SCFE may be warranted. The incidence of SCFE did not increase over the last decades despite increasing obesity rates.

Slipped capital femoral epiphysis (SCFE) occurs commonly in overweight children and adolescents. The etiology of the disease is still unknown but several studies have concluded that overweight and obesity are catalyzing factors, either by overloading the growth plate (Fishkin et al. 2006) or as an endocrine condition diminishing the stability of the growth plate. The latter would explain the age-dependent relationship between obesity and SCFE onset, where obese children are found to suffer from SCFE at a younger age compared with children of age- and length-adequate weight (Perry et al. 2018). Wensaas et al. (2011) investigated the long-term outcome of 66 patients with a history of SCFE and found that one-third were overweight or obese in adulthood. However, the risk of developing obesity in SCFE patients, not only in childhood but during later life, is still unknown. In contrast, presuming obesity as a causal factor, one would expect higher incidences of SCFE due to epidemic obesity rates in children and adolescents (Murray and Wilson 2008). However, comparisons of incidence rates are often difficult because the calculations are based on different age groups and changes in incidence rates over the past decades have not yet been reported. Inconsistent findings concerning the association between SCFE and hypothyroidism have emerged. Some authors found no association between SCFE and hypothyroidism (Brenkel et al. 1988), whereas others found an association of the 2 diseases (Kadowaki et al. 2017). Congenital hypothyroidism is part of the screening program of newborns in Sweden (National Board of Health and Welfare 2018) but acquired hypothyroidism is often underdiagnosed in children and adolescents (Ghaemi et al. 2015). To my knowledge, there is no study investigating the lifetime risk of hypothyroidism in patients with a history of SCFE. Studies focusing on the long-term outcome after SCFE (Wensaas et al. 2011, Castaneda et al. 2013, Wiemann and

458

Herrera-Soto 2013, de Poorter et al. 2016) attest that SCFE is not only a childhood hip disease: In some patients SCFE transforms into a chronic disease by creating hip joint impingement (Lerch et al. 2019) or premature osteoarthritis, or both (Goodman et al. 1997). It is known that patients with chronic diseases are at greater risk of developing depression (Moussavi et al. 2007, Podeszwa et al. 2015) and die earlier (Ng et al. 2007, Hailer and Nilsson 2014) compared with the general population. Thus, the lifetime burden of obesity, hypothyroidism, and depression in patients exposed to SCFE remains unclear and leads to the following questions: (1) Do patients with SCFE have an increased lifetime risk of obesity and hypothyroidism? (2) Has the average incidence of SCFE in Sweden changed over the past few decades? (3) Is SCFE associated with a higher risk of depression and a higher risk of all-cause mortality?

Patients and methods Study design and data sources To conduct this nationwide, population-based cohort study the Swedish Patient Register was used to identify all patients with a diagnosis of SCFE (ICD-7: 732.03, ICD-8: 722.10, ICD-9: 732.2, ICD-10: M93.0) from 1964 (when the register was established) until 2011. Through the Swedish Total Population Register, up to 10 controls without SCFE were matched, using the matching criteria date of birth, sex, region of residence, and being alive at the time of SCFE diagnosis, to compare the risk of obesity (ICD-7: 287.09, ICD-8: 277.99, ICD-9: 278 and 783.6, ICD-10: E66.0), hypothyroidism (ICD-7: 250–252, ICD-8: 240–242, ICD-9: 240–244, ICD-10: E00–E03) and depression (ICD-7: 790.29, ICD-8: 790.20, ICD-9: 308 and 311, ICD-10: F92.0 and F32-38). The Swedish Patient Register provides information on diagnosis codes and dates of admissions and discharge for all individuals in Sweden. Whenever hospital care (in- or outpatient) is given, it is mandatory for all public and private hospitals to deliver information on dates of admission and discharge, registered diagnoses (categorized by the International Classification of Diseases [ICD]), and applied treatments to the Swedish Patient Register together with the unique personal identification number of each individual. Primary health care institutions do not report to the Swedish Patient Register. However, whenever the SCFE diagnosis is apparent on radiographs the patient is referred to hospital care (in- or outpatient). The control group was identified through the Swedish Population Register based on the matching criterion. Statistics Sweden, a government agency, provided information on population size and changes, such as the number of births, deaths, and immigration and emigration at time of interest.

Acta Orthopaedica 2020; 91 (4): 457–463

Number of patients 600

400

200

0 5

Age at SCFE diagnosis

Figure 1. Age at onset of SCFE diagnosis with a peak between 10 and 15 years of age.

Study population The study population was followed from 1964 until the diagnosis of interest (obesity, hypothyroidism, depression), death, emigration, or December 31, 2013, whichever occurred first. The endpoints were obesity (based on the World Health Organization [WHO] criterion of body mass index (BMI) ≥ 30), hypothyroidism, depression, and death from all causes. Any given individual could experience multiple endpoints when diagnosed with more than one diagnosis of interest. The initial database contained 2,989 patients diagnosed with SCFE and 29,876 controls. 11 SCFE patients were excluded because no matching controls were found. In addition, patients whose age at SCFE diagnosis was < 5 years or ≥ 17 years, those with diagnosed developmental dysplasia (7 patients) of the hip, and those with Legg–Calvé–Perthes disease (38 patients) were excluded together with their controls. Characteristics of the study population The final study population consisted of 2,564 SCFE patients and 25,638 matched controls; there were 11,253 (40%) females, and the mean follow-up time was 34 years (5–66). Median year of birth was 1980 (range 1948–2004), and the mean age at SCFE onset was 12.7 years (Figure 1). Statistics Kaplan–Meier analysis was used to calculate cumulative unadjusted survival functions, with each of comorbidity or death as separate endpoints. Cox proportional hazard models were fitted to estimate the hazard ratio (HR) of developing obesity, hypothyroidism, depression, and death in patients with SCFE compared with unexposed individuals, with or without adjustment for birth year and sex. The assumption of proportional hazards was verified by visual inspection of unadjusted cumulative survival function plots. Estimation uncertainty was assessed by calculating 95% confidence intervals (CIs). Stratified analyses were performed by sex and by the differ-

Acta Orthopaedica 2020; 91 (4): 457–463

Table 1. Prevalence of obesity, hypothyroidism, depression, or death in patients with SCFE and in controls. Values are number (%) Comorbidities Obesity Hypothyroidism Depression Death

459

Table 2. Hazard risks (HR) with 95% confidence intervals (CI) of comorbidities and allcause mortality in patients with SCFE and in controls

Control n = 25,638

SCFE n = 2,564

Comorbidities HR (CI)

129 (0.5) 166 (0.7) 998 (3.9) 329 (1.3)

109 (4.3) 46 (1.8) 101 (3.9) 52 (2.0)

Obesity Hypothyroidism Depression Death a

aHR (CI) a

8.7 (6.7–11.2) 2.8 (2.0–3.9) 1.0 (0.8–1.3) 1.6 (1.2–2.1)

Adjustment for birth year and sex.

ent ICD coding periods described earlier (data not shown). The annual incidence density for SCFE between 1977 and 2011 was calculated based on the population data provided by Statistic Sweden for each year of interest. We calculated the incidence density for SCFE by dividing the number of incident cases per year versus the number of person-years at risk for each corresponding year. We approximated the number of person-years at risk by calculating the average of the population size at risk (children between age 5 and 16 years resident in Sweden) at the start and end of each year of interest. To calculate the 95% CI for the incidence density, the standard deviation (SD) is estimated as sqrt((p × (1–p))/N) and the 95% CI is estimated as p ±1.96 × SD. In order to make this study comparable to the study group of Herngren et al. (2017) a sub-group analysis for SCFE patients with a narrower and more typical age frame for SCFE was performed. Only patients with SCFE diagnosis at age 9 and 15 years between 2000 and 2006 were analyzed. All statistical analyses were performed using R statistical software (Version 3.3.3; R Foundation for Statistical Computing, Vienna, Austria), including the rms, magrittr, Gmisc, ggplot2, Formula and MASS packages. Ethics, funding, data sharing, and potential conflicts of interest This study was approved by the Ethics Research Committee in Uppsala, Sweden (registration number 2012/065, date of issue March 21, 2012). This research was not supported by grants from any funding agency in the public, commercial or not-for-profit sectors. The dataset that is necessary to replicate main findings can be obtained from the author upon reasonable request. I have no conflicts of interest to declare.

Results Lifetime risk of comorbidities and death in SCFE patients compared with unexposed individuals The risk of developing obesity was higher in patients with SCFE than in controls (HR 9, CI 7–11), as was the risk of developing hypothyroidism (HR 3, CI 2–4). The risk of devel-

Table 3. Causes of death in patients with SCFE and in controls. Values are number (%) Cause of death Vascular Endocrine Injury Neuropsychiatric Cancer Suicide

Control SCFE n = 25,638 n = 2,564 38 (0.2) 5 (0.02) 70 (0.3) 15 (0.1) 66 (0.3) 36 (0.1)

14 (0.6) 5 (0.2) 8 (0.3) 3 (0.1) 4 (0.2) 4 (0.2)

Incidence density with 95% CI per 100,000 10

0 1980

1985

1990

1995

Year

2000

2005

2010

Figure 2. Incidence density of SCFE diagnoses per 100,000 children aged 5 to 16 years in Sweden from 1977 to 2011.

oping obesity was higher in male (HR 11, CI 8–16) than in female SCFE patients (HR 7, CI 5–10), and a similar pattern was seen for the risk of hypothyroidism, which was higher in male (HR 5, CI 3–8) than in female SCFE patients (HR 2, CI 2–4; Table 1). There was no statistically significant difference in the risk of depression between the SCFE patients and controls (HR 1, CI 1–1.3). During the observation period, 381 (1.4%) individuals died. There were 52 deaths in the SCFE group with the youngest dying at the age of 9 years and the oldest at the age of 53 years. 329 were dead in the control group with the youngest dying at the age of 8 months and the oldest dying at the age of 63 years. The all-cause mortality risk was higher in SCFE patients than in controls (HR 2, CI 1–2). Adjustment for birth year and sex hardly changed the estimates (Table 2). The causes of death are given in Table 3. Incidence density of SCFE in Sweden over the past decades The annual incidence density of SCFE varied from 3.4/100,000 to 7.8/100,000 in children aged 5 to 16 years between 1977 and 2011 (Figure 2). In order to make these data comparable with those calculated by Herngren et al. (2017) a subgroup

460

Acta Orthopaedica 2020; 91 (4): 457–463

Age at obesity

Age at hypothyroidism

Control

SCFE

Control

Age at depression

Age at death

Control

SCFE

Control

SCFE

Figure 3. Age at onset of obesity, hypothyroidism, depression, or death of SCFE patients. Age at onset of the diagnosis of obesity and hypothyroidism is younger in patients with SCFE than in the control group (obesity: p-value < 0.05; hypothyroidism: p-value < 0.001). No differences were found for age at diagnosis of depression or death between the two groups. The boxes represent the interquartile range and the red line median. The whisker represents the age range at diagnosis or death.The dot represents an outlier.

analysis of children aged 9 to 15 years between 2000 and 2006 was performed. In this 7-year-period 529 children were diagnosed with SCFE. The average annual number of children at risk was 833,632. Temporal sequence of SCFE and comorbidities In most patients the comorbidity diagnosis was registered after the SCFE diagnosis. Only 20 patients were diagnosed with obesity (1 to 7 years) before being diagnosed with SCFE, and 5 were diagnosed with hypothyroidism (a few months to 7 years) before being diagnosed with SCFE. None had a diagnosis of depression before the SCFE diagnosis. The mean and median age at diagnosis for the comorbidities obesity and hypothyroidism was lower in patients with SCFE than in the control group (Figure 3). A sensitivity analysis involving only patients and controls who were registered from 2001 when both in- and outpatient hospital care were registered revealed that the adjusted risks for obesity (HR 6, CI 3–10) and hypothyroidism (HR 3, CI 2–8) remained higher in SCFE patients than in controls. In this subgroup, the risk of developing depression was similar for SCFE patients and controls.

Discussion SCFE is one of the most common hip disorders in children that can lead to lifelong impairment of hip function. Associa-

tions between SCFE and obesity and hypothyroidism have been described but many studies used a retrospective design and small cohorts (Bhatia et al. 2006, Ucpunar et al. 2018). In the past years reports from national databases have increased (Murray and Wilson 2008, Herngren et al. 2017, Perry et al. 2017, Ravinsky et al. 2019), making it possible to analyze epidemiological characteristics of different disease conditions. However, studies emerging from these databases lack a control group or have only appeared in the past decade, making it difficult to make observations before and after exposure or to analyze changes of incidences over time. Moreover, I know of no population-based study describing the lifetime comorbidities of patients with SCFE compared with those without SCFE that focuses on obesity, hypothyroidism, depression, and mortality. This study finds that patients with a history of SCFE have a 9-fold higher risk of being diagnosed with obesity than their age- and sex-matched controls, with a higher risk for male patients. Over time, the prevalence of obesity in children has increased in both sexes (Eriksson et al. 2018). Perry et al. (2018) found a strong association of BMI and SCFE, with a higher risk for SCFE in children with obesity and high BMI compared with children of normal BMI. The concerns regarding the rising incidence of SCFE relative to the increasing prevalence of obesity in children (Neovius et al. 2006, Nguyen et al. 2011) are not supported by this study. As a matter of fact, the incidence rate has some variation since 1978, with peaks in 1981 and 2002, but it then remains roughly the same until 2011. In comparison, Herngren et al. (2017) investigated the incidence of SCFE in children between 9 and 15 years old for the period 2007 to 2013 and identified 379 children with SCFE. The average annual number of children 9–15 years old in Sweden between 2007 and 2013 was 726,304. We analyzed a comparable preceding 7-year-period from 2000 to 2006 with an average annual number of children at risk of 833,632 and identified 529 children with SCFE. Thus, the incidences in these two 7-years periods were similar. Interestingly, a 20-year periodicity of peak incidences has also been described for the period from 1900 to 1970 (Hagglund et al. 1984). Nevertheless, the underlying causes of these longterm fluctuations are unclear. The inconsistency between rising obesity rates in children, on the one hand, and stable rates for SCFE, on the other, together with the elevated risk of developing obesity in later life found in this study indicates that not only obesity per se but other factors associated with obesity may play a role in the etiology of SCFE. Halverson et al. (2017) reported elevated serum leptin levels in patients with SCFE, regardless of their BMI, suggesting a shared risk factor for obesity and SCFE. Leptin is a cytokine-like hormone with proinflammatory properties known to be associated with autoimmune disorders, infections, and endocrine and metabolic diseases (Procaccini et al. 2015). Higher leptin levels with an effect on chondrocytes via interleukin (IL)-1 regulation were found in the serum and

Acta Orthopaedica 2020; 91 (4): 457–463

synovial fluid of patients with osteoarthritis and rheumatoid arthritis (Yan et al. 2018). However, an effect on the growth plate, if any, has not yet been investigated. Another pathway is the link between high leptin levels and thyroiditis, the most common cause of hypothyroidism (Procaccini et al. 2015) and the association of hypothyroidism with SCFE (Marquez et al. 2014, Uday et al. 2014, Kadowaki 2017). Case series and smaller retrospective studies in patients with SCFE (Loder et al. 1995) have demonstrated abnormal levels of thyroid hormone typical for hypothyroidism, but this study is the first to show a 3-fold higher risk of hypothyroidism in patients with SCFE compared with their matched controls in a populationbased setting. The risk of hypothyroidism was almost 5-fold higher in male patients with SCFE and 2-fold higher in female patients with SCFE compared with their respective controls. The prevalence rate of hypothyroidism in Europe is between 0.2% and 5%, being higher in females (Chaker et al. 2017) with an estimated prevalence of undiagnosed hypothyroidism of 6% in females and 3% in males (Garmendia Madariaga et al. 2014). Although SCFE is a childhood hip disease it can lead to hip pain and premature osteoarthritis of the hip in adulthood (Goodman et al. 1997), especially in patients who suffered severe slips, complications, or both (e.g., impaired range of motion of the hip joint, avascular necrosis of the femoral head or chondrolysis) (Tosounidis et al. 2010). The association between osteoarthritis and depression is well documented (Moussavi et al. 2007, Veronese et al. 2017) and the psychological burden of a chronic disease is not a novel concept (Moussavi et al. 2007). However, this study did not reveal a strong association of SCFE with depression, which could be explained by the assumption that pediatric patients cope better than adults with chronic diseases. Our cohort includes both mild and severe slips and severe slips presumably pose a higher risk of developing early osteoarthritis (Novais and Millis 2012). It is reasonable to suppose that the risk of depression might be higher in those who have severe slips and who later develop osteoarthritis than in those with mild slips. This is the first study investigating the all-cause mortality in patients with SCFE. The 2-fold higher mortality risk together with an overrepresentation of vascular and endocrine causes of death in SCFE patients could be explained by the higher prevalence and risk of obesity and hypothyroidism. Flegal et al. (2013) found a higher risk of all-cause mortality in patients with severe obesity (BMI > 35). However, a lower mortality risk was found in individuals with overweight (BMI 25–< 30) or mild obesity (BMI 30–< 35). Hypothyroidism (even subclinical) is also associated with higher all-cause mortality (Huang et al. 2018). Another explanation for increased mortality risk relates to socioeconomic factors. Perry et al. (2017) found higher incidences of SCFE in individuals living in deprived areas. In addition, an association has been noted between lower socioeconomic status and higher mortality risk (Foster et al. 2018).

461

In most patients the comorbidity was diagnosed and registered after the diagnosis of SCFE. This could either be because the main focus was on the hip and not on the comorbidity or because of the fact that the comorbidity was not apparent at the time of the SCFE diagnosis. In Sweden, SCFE patients are treated at orthopedic departments and blood test for thyroid hormones is not routinely taken in patients with orthopedic diseases. Concerning obesity, the diagnosis might have been apparent at the same time at the SCFE diagnosis but not registered. This study has a number of limitations. Unfortunately, I had no information on the socioeconomic background of our cohort. The study covers a long time period with parallel changes in ICD systems, and, additionally, changes in the coding practice may have influenced the results. This problem was addressed by analyzing the ICD periods separately, indicating no major differences in outcomes (data not shown). Furthermore, diagnoses that were managed in outpatient settings prior to 2001 have been missed, but this applies for both the patients and the control group, which makes a profound influence on the estimates attained unlikely. A separate analysis including only patients and control individuals diagnosed after 2001 was performed, but this resulted in no major differences in the attained risk estimates. Patients with diagnoses that can be treated in primary healthcare are not included in the Swedish Patient Register, which explains the low prevalence of obesity and depression in this study. The estimated prevalence of obesity in Sweden in 2011 was 14% in men and 13% in women aged 16–84 years (WHO 2013). Another drawback is that the Swedish Patient Register does not provide information on laterality or bilateral affection during the observation period. Bhatia et al. (2006) showed that patients bilaterally affected by SCFE had higher BMI than patients affected unilaterally. The prevalence of hypothyroidism and other endocrine disorders seems to be related to bilateral slips (Wells et al. 1993). Because patients with a history of SCFE inevitably have more frequent contacts with the healthcare system detection bias might be an issue. Nevertheless, in most patients there was a considerable time interval between the diagnosis of SCFE and the diagnosis of the comorbidity, which makes this scenario less likely. In addition, SCFE and its consequences, such as impingement of the hip joint or premature osteoarthritis, are mainly treated by orthopedic surgeons, whereas endocrine and metabolic diseases are often diagnosed by other specialists, which makes detection bias as an issue less likely. The higher risk for all-cause mortality in patients with a history of SCFE might also serve as an argument against a large impact of detection bias. Conclusions For patients with SCFE and physicians it might be important to be aware of the higher lifetime risk of developing obesity, enabling preventive measures. Despite a rise in the prevalence of obesity in children and adolescents, the incidence of SCFE

462

did not increase in Sweden over time. Patients with SCFE had a higher risk of hypothyroidism and a higher risk of all-cause mortality compared with a control cohort without SCFE. However, a cost–benefit analysis of screening for hypothyroidism needs to be performed before giving recommendations. The psychological burden of SCFE, as expressed in the risk of developing depression, was not confirmed in this study. However, the total burden of SCFE should be investigated more fully using patient-reported outcome measures.

Thanks are offered to Niclas Eriksson, Biostatistician, Uppsala Clinical Research Center, Uppsala, Sweden. Acta thanks Gunnar Hägglund and Anders Wensaas for help with peer review of this study.

Bhatia N N, Pirpiris M, Otsuka NY. Body mass index in patients with slipped capital femoral epiphysis. J Pediatr Orthop 2006; 26(2): 197-9. doi: 10.1097/01.bpo.0000218526.36362.3f. Brenkel I J, Dias J J, Iqbal S J, Gregg P J. Thyroid hormone levels in patients with slipped capital femoral epiphysis. J Pediatr Orthop 1988; 8(1): 22-5. Castaneda P, Ponce C, Villareal G, Vidal C. The natural history of osteoarthritis after a slipped capital femoral epiphysis/the pistol grip deformity. J Pediatr Orthop 2013; 33(Suppl. 1): S76-82. doi: 10.1097/BPO.0b013e318277174c. Chaker L, Bianco A C, Jonklaas J, Peeters R P. Hypothyroidism. Lancet 2017; 390(10101): 1550-62. doi: 10.1016/S0140-6736(17)30703-1. de Poorter J J, Beunder T J, Gareb B, Oostenbroek H J, Bessems GbH, van der Lugt J C, Maathuis P G, van der Sande M A. Long-term outcomes of slipped capital femoral epiphysis treated with in situ pinning. J Child Orthop 2016; 10(5): 371-9. doi: 10.1007/s11832-016-0759-z. Eriksson M, Lingfors H, Golsater M. Trends in prevalence of thinness, overweight and obesity among Swedish children and adolescents between 2004 and 2015. Acta Paediatr 2018; 107(10): 1818-25. doi: 10.1111/apa.14356. Fishkin Z, Armstrong D G, Shah H, Patra A, Mihalko W M. Proximal femoral physis shear in slipped capital femoral epiphysis: a finite element study. J Pediatr Orthop 2006; 26(3): 291-4. doi: 10.1097/01. bpo.0000217730.39288.09. Flegal K M, Kit B K, Orpana H, Graubard B I. Association of all-cause mortality with overweight and obesity using standard body mass index categories: a systematic review and meta-analysis. JAMA 2013; 309(1): 71-82. doi: 10.1001/jama.2012.113905. Foster H M E, Celis-Morales C A, Nicholl B I, Petermann-Rocha F, Pell J P, Gill J M R, O’Donnell C A, Mair F S. The effect of socioeconomic deprivation on the association between an extended measurement of unhealthy lifestyle factors and health outcomes: a prospective analysis of the UK Biobank cohort. Lancet Public Health 2018; 3(12): e576-e85. doi: 10.1016/ s2468-2667(18)30200-7. Garmendia Madariaga A, Santos Palacios S, Guillen-Grima F, Galofre J C. The incidence and prevalence of thyroid dysfunction in Europe: a metaanalysis. J Clin Endocrinol Metab 2014; 99(3): 923-31. doi: 10.1210/ jc.2013-2409. Ghaemi N, Bagheri S, Elmi S, Mohammadzade Rezaee S, Elmi S, Erfani Sayyar R. Delayed Diagnosis of hypothyroidism in children: report of 3 cases. Iranian Red Crescent Med J 2015; 17(11): e20306. doi: 10.5812/ ircmj.20306. Goodman D A, Feighan J E, Smith A D, Latimer B, Buly R L, Cooperman D R. Subclinical slipped capital femoral epiphysis: relationship to osteoarthrosis of the hip. J Bone Joint Surg Am 1997; 79(10): 1489-97.

Acta Orthopaedica 2020; 91 (4): 457–463

Hagglund G, Hansson L I, Ordeberg G. Epidemiology of slipped capital femoral epiphysis in southern Sweden. Clin Orthop Relat Res 1984; (191): 82-94. Hailer Y D, Nilsson O. Legg–Calve–Perthes disease and the risk of ADHD, depression, and mortality. Acta Orthop 2014; 85(5): 501-5. doi: 10.3109/17453674.2014.939015. Halverson S J, Warhoover T, Mencio G A, Lovejoy S A, Martus J E, Schoenecker J G. Leptin elevation as a risk factor for slipped capital femoral epiphysis independent of obesity status. J Bone Joint Surg Am 2017; 99(10): 865-72. doi: 10.2106/jbjs.16.00718. Herngren B, Stenmarker M, Vavruch L, Hagglund G. Slipped capital femoral epiphysis: a population-based study. BMC Musculoskelet Disord 2017; 18(1): 304. doi: 10.1186/s12891-017-1665-3. Huang H K, Wang J H, Kao S L. Association of hypothyroidism with allcause mortality: a cohort study in an older adult population. J Clin Endocrinol Metab 2018; 103(9): 3310-18. doi: 10.1210/jc.2018-00408. Kadowaki S, Hori T, Matsumoto H, Kanda K, Ozeki M, Shirakami Y, Kawamoto N, Ohnishi H, Fukao T. Prepubertal onset of slipped capital femoral epiphysis associated with hypothyroidism: a case report and literature review. BMC Endocrine Disorders 2017; 17(1): 59. doi: 10.1186/s12902017-0210-6. Lerch T D, Novais E N, Schmaranzer F, Ziebarth K, Steppacher S D, Tannast M, Siebenrock K A. What is the prevalence of CAM deformity after prophylactic pinning of the contralateral asymptomatic hip in unilateral slipped capital femoral epiphysis? A 10-year minimum followup study. Clin Orthop Relat Res 2019; 477(5): 1111-22. doi: 10.1097/corr.0000000000000592. Loder R T, Wittenberg B, DeSilva G. Slipped capital femoral epiphysis associated with endocrine disorders. J Pediatr Orthop 1995; 15(3): 349-56. Marquez D, Harb E, Vilchis H. Slipped capital femoral epiphysis and hypothyroidism in a young adult: a case report. J Med Case Rep 2014; 8: 336. doi: 10.1186/1752-1947-8-336. Moussavi S, Chatterji S, Verdes E, Tandon A, Patel V, Ustun B. Depression, chronic diseases, and decrements in health: results from the World Health Surveys. Lancet 2007; 370(9590): 851-8. doi: 10.1016/S01406736(07)61415-9. Murray A W, Wilson N I. Changing incidence of slipped capital femoral epiphysis: a relationship with obesity? J Bone Joint Surg Br 2008; 90(1): 92-4. doi: 10.1302/0301-620x.90b1.19502. National Board of Health and Welfare. National guidelines for screening of congenital diseases in newborns [Screening för 24 medfödda sjukdomar hos nyfödda barn]; 2018. Report: 2019-7-6223, www.socialstyrelsen.se/ regler-och-riktlinjer/nationella-screeningprogram. Neovius M, Janson A, Rossner S. Prevalence of obesity in Sweden. Obes Rev 2006; 7(1): 1-3. doi: 10.1111/j.1467-789x.2006.00190.x. Ng T P, Niti M, Tan W C, Cao Z, Ong K C, Eng P. Depressive symptoms and chronic obstructive pulmonary disease: effect on mortality, hospital readmission, symptom burden, functional status, and quality of life. Arch Intern Med 2007; 167(1): 60-7. doi: 10.1001/archinte.167.1.60. Nguyen A R, Ling J, Gomes B, Antoniou G, Sutherland L M, Cundy P J. Slipped capital femoral epiphysis: rising rates with obesity and aboriginality in South Australia. J Bone Joint Surg Br 2011; 93(10): 1416-23. doi: 10.1302/0301-620x.93b10.26852. Novais E N, Millis M B. Slipped capital femoral epiphysis: prevalence, pathogenesis, and natural history. Clin Orthop Relat Res 2012; 470(12): 3432-8. doi: 10.1007/s11999-012-2452-y. Perry D C, Metcalfe D, Costa M L, Van Staa T. A nationwide cohort study of slipped capital femoral epiphysis. Arch Dis Child 2017; 102(12): 1132-6. doi: 10.1136/archdischild-2016-312328. Perry D C, Metcalfe D, Lane S, Turner S. Childhood obesity and slipped capital femoral epiphysis. Pediatrics 2018; 142(5). doi: 10.1542/peds.2018-1067. Podeszwa D A, Richard H M, Nguyen D C, De La Rocha A, Shapiro E L. Preoperative psychological findings in adolescents undergoing hip preservation surgery. J Pediatr Orthop 2015; 35(3): 253-7. doi: 10.1097/ bpo.0000000000000243.

Acta Orthopaedica 2020; 91 (4): 457–463

Procaccini C, Pucino V, Mantzoros C S, Matarese G. Leptin in autoimmune diseases. Metabolism 2015; 64(1): 92-104. doi: 10.1016/j. metabol.2014.10.014. Ravinsky R, Rofaiel J, Escott B G, Lim Z, Ravi B, Howard A. Epidemiology of slipped capital femoral epiphysis in Ontario, Canada. J Pediatr Orthop 2019; 39(3): e165-e7. doi: 10.1097/BPO.0000000000001254. Tosounidis T, Stengel D, Kontakis G, Scott B, Templeton P, Giannoudis P V. Prognostic significance of stability in slipped upper femoral epiphysis: a systematic review and meta-analysis. J Pediatr 2010; 157(4): 674-80, 80.e1. doi: 10.1016/j.jpeds.2010.04.018. Ucpunar H, Camurcu I Y, Duman S, Ucpunar E, Sofu H, Bayhan A I. Obesityrelated metabolic and endocrine disorders diagnosed during postoperative follow-up of slipped capital femoral epiphysis. Acta Orthop 2018; 89(3): 314-19. doi: 10.1080/17453674.2018.1445167. Uday S, Scott B, Alvi S. Hashimoto’s hypothyroidism presenting with SUFE (slipped upper femoral epiphysis). BMJ Case Reports 2014; 2014. doi: 10.1136/bcr-2013-203095.

463

Veronese N, Stubbs B, Solmi M, Smith T O, Noale M, Cooper C, Maggi S. Association between lower limb osteoarthritis and incidence of depressive symptoms: data from the osteoarthritis initiative. Age Ageing 2017; 46(3): 470-6. doi: 10.1093/ageing/afw216. Wells D, King J D, Roe T F, Kaufman F R. Review of slipped capital femoral epiphysis associated with endocrine disease. J Pediatr Orthop 1993; 13(5): 610-14. Wensaas A, Svenningsen S, Terjesen T. Long-term outcome of slipped capital femoral epiphysis: a 38-year follow-up of 66 patients. J Child Orthop 2011; 5(2): 75-82. doi: 10.1007/s11832-010-0308-0. WHO. Nutrition, Physical activity and obesity—Monitoring and surveillance—Overweight and obesity in three age groups: Sweden; 2013. Wiemann J M 4th, Herrera-Soto J A. Can we alter the natural history of osteoarthritis after SCFE with early realignment? J Pediatr Orthop 2013; 33(Suppl. 1): S83-7. doi: 10.1097/BPO.0b013e318277167b. Yan M, Zhang J, Yang H, Sun Y. The role of leptin in osteoarthritis. Medicine (Baltimore) 2018; 97(14): e0257. doi: 10.1097/MD.0000000000010257.

464

Acta Orthopaedica 2020; 91 (4): 464–470

SwedeAmp—the Swedish Amputation and Prosthetics Registry: 8-year data on 5762 patients with lower limb amputation show sex differences in amputation level and in patient-reported outcome Ilka KAMRAD 1, Bengt SÖDERBERG 2, Hedvig ÖRNEHOLM 1, and Kerstin HAGBERG 3 1 Departments

of Orthopedics and Clinical Sciences, Lund University and Skåne University Hospital Malmö; 2 Department of Orthopedics, Skåne University Hospital Lund; 3 Centre for Advanced Reconstruction of Extremities and Department for Prosthetics and Orthotics, Sahlgrenska University Hospital, Gothenburg, and Institute of Clinical Sciences, Department of Orthopedics, Sahlgrenska Academy, Gothenburg, Sweden Correspondence: ilka.kamrad@med.lu.se Submitted 2019-06-10. Accepted 2020-03-25.

Background and purpose — For want of national guidelines for lower limb amputation (LLA) the quality registry SwedeAmp was started in 2011 to increase knowledge around LLA and prosthetic rehabilitation. We now present data from the first 8 years of registration. Patients and methods — We present descriptive data from the first 8 years (2011–2018) of registration. Patientreported outcome was collected at baseline and at follow-up 12 and 24 months after surgery for patients with prosthetic supply and included generic (EQ-5D-5L) and amputee-specific (e.g., LCI-5L and Prosthetic Use Score) measures. Sex differences were investigated. Results — As at December 31, 2018, 5,762 patients, 7,776 amputations, 2,658 prosthetic supplies, 1,848 baselines, and 2,006 follow-ups were registered. 61% of the patients were male, and mean age by the time of the first registered amputation was 74 years (SD 14). Women were older, more frequently had vascular disease without diabetes and more often underwent amputation at a higher level compared with men (p < 0.001). Time from amputation to fitting of first individual prosthesis was median 69 days (6–500) after transtibial amputation (TTA) and 97 days (19–484) after transfemoral amputation (TFA). The outcomes were lower after TFA than after TTA. Interpretation — SwedeAmp shows sex differences concerning amputation level, diagnosis, and age, leading to the conclusion that women have worse preconditions for successful prosthetic mobility after LLA. With increasing coverage, SwedeAmp can provide deeper knowledge with regard to patients undergoing LLA in Sweden.

Lower limb amputation (LLA) is often discussed from specific points of view such as amputation incidence (Johannesson et al. 2008, Buckley et al. 2012, Fortington et al. 2013b, Jones et al. 2013), mortality (Fortington et al. 2013a, Jones et al. 2013), prosthetic prescription, mobility, and patient-reported outcome (Raichle et al. 2008, Norvell et al. 2011, Davie-Smith et al. 2017a). Moreover, most of those studies include only patients from one hospital or region. The diversity of data and the difficulty of comparing results have been discussed repeatedly (Ephraim et al. 2003, Sinha and Van Den Heuvel 2011, Fortington et al. 2012, Samuelsson et al. 2012, van Netten et al. 2016, Davie-Smith et al. 2017a, b). According to the Swedish National Board of Health and Welfare, the incidence of major (transtibial level or more proximal) LLA in Sweden, including revisions and re-amputations, has for many decades been between 33 and 39/100,000 inhabitants, with yearly and regional variations from 9–107/100,000. Amputation level, pre- and postoperative care, rehabilitation, and prosthetic supply differ between regions and hospitals. With the intention to provide equal and best possible care for patients with LLA, the Swedish Amputation and Prosthetics Registry for the lower extremity (SwedeAmp) was founded in 2011. The aim of SwedeAmp is to evaluate the entire medical process regarding LLA in Sweden (Figure 1). We present descriptive data on the first 8 years of registration, evaluate the outcome at 12- and 24-months’ follow-up, and investigate possible sex differences with a focus on major amputations.

Acta Orthopaedica 2020; 91 (4): 464–470

465

Number of patients and surgical procedures 9,000 Patients Surgical procedures

8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000

Figure 1. The complex process around lower limb amputation.

Patients and methods SwedeAmp includes patient-based data regarding amputation surgery, prosthetic supply, the patient’s situation, and mobility before amputation and outcome 6, 12, and 24 months after the amputation. Data are registered in 6 different forms (Table 1, see Supplementary data). Forms 1 and 2 include all levels of LLA from partial toe amputation to hemipelvectomy, Forms 3–6 solely major amputations at or proximal to transtibial amputation (TTA) level. The registry uses 3 definitions for surgical procedures: primary amputation, re-amputation, and revision. A revision is defined as a soft tissue revision and/or boneshortening procedure performed within the same amputation level. Re-amputation is defined as a second procedure on an unhealed residual limb leading to a higher classified amputation level, e.g., from TTA to knee disarticulation (KD) or transfemoral amputation (TFA). Primary amputation is defined as amputation not matching the criteria for revision or re-amputation. Amputation on a higher classified level after previous healed amputation is considered to be a new primary amputation. Furthermore, bilateral amputation is defined in the registry as amputation at/or proximal to the tarsometatarsal level on both sides, performed simultaneously or at different times. To estimate the survival of the patients registered, we used dates of death from the Swedish National Board of Health and Welfare. For the evaluation of functionality and quality of life, several validated tests and scores are performed at baseline and/or follow-up. The Locomotor Capability Index-5Level (LCI-5L) measures self-reported mobility with a prosthesis (Franchignoni et al. 2004, Larsson et al. 2009). The LCI-5L basic score (values 0–28) and advanced score (0–28) are reported separately and the sum of the 2 scores results in the LCI-5L total score (0–best possible 56). LCI-5L can be used for mobility assessment prior to amputation simply by removing the word prosthesis. The Prosthetic Use Score combines the number of

2012

2013

2014

2015

2016

2017

2018

Figure 2. Development of the SwedeAmp registry showing the total registrations in number of patients and surgical procedures from 2012 to 2018.

days/week and the number of hours/day the prosthesis is used. A score of 0 indicates that the prosthesis is not used any day/ week while a score of 100 indicates wear of prosthesis 7 days/ week and more than 15 hours/day (Hagberg et al. 2004). The EQ-5D-5L index estimates the patient’s general health with scores between -0.594 and 1 (full health) (www.euroqol.com). The Timed-Up-and-Go Test (TUG) (Schoppen et al. 2003) assesses functionality and falling risk. SwedeAmp aims to involve all key professions within the multidisciplinary team. The medical and surgical data are preferably registered by a surgeon, the prosthetic supply by a certified prosthetist and orthotist (CPO), the baseline and follow-up data by a rehabilitation therapist, and the gait data by a CPO or a physiotherapist. SwedeAmp has not yet gained full coverage in Sweden. Figure 2 shows the annual total registrations from 2012 to 2018. In 2018, 11 of the 21 Swedish regions were registered in SwedeAmp. Among these, the coverage for the most common level of amputation, TTA, was 62% compared with data from the National Board of Health and Welfare in 2017 (www.socialstyrelsen.se). Statistics Data are presented as numbers (n) and %. For continuous data means (SD) are presented and for ordinal data median (md) and min–max values. 95% confidence interval (CI) was reported for mean age at the time of the first registered amputation. To estimate statistically significant differences between groups, chi-square and Mann–Whitney U tests were performed. A p-value of < 0.05 was considered statistically significant. Ethics, funding, and potential conflicts of interest In accordance with the rules for Swedish national quality registries, patients are informed of registrations in quality registries and have the possibility to decline participation at any time, but no signed consent is needed. This report is based on descrip-

466

Acta Orthopaedica 2020; 91 (4): 464–470

Distribution (%) of age groups and sex

Distribution (%) of amputation levels

Distribution (%) of walking aids at home Distribution (%) of walking aids outdoors

100

Women Men

No aid 1 crutch/stick 2 crutches/sticks Wheeler Not walking

30 25 20 15 10

5 0

< 20 20–49 50–59 60–69 70–79 80–89 ≥ 90

Age

Figure 3. Age and sex at the time of the first registered amputation.

Transtibial amputation Knee disarticulation Transfemoral amputation Hip disarticulation or hemipelvectomy

Women (n = 2,442)

Men (n = 3,393)

Figure 4. Differences in major amputation levels for men and women (p < 0 .001).

tive data from the open source SwedeAmp report 2018 (www. swedeamp.com) and did not require ethical approval. Funding was received from ALF Skåne and FoU Skåne and the Swedish Government research grant. There are no conflicts of interest.

Results Patients As at December 31, 2018, 5,762 patients, 7,776 amputations, 2,658 prosthetic supplies, 1,848 baseline, and 2,006 followup registrations were registered in SwedeAmp. 61% of the patients were male. Mean age at the time of the first registered amputation in our sample was 74 years (SD 14); women were older (78, SD 14, CI 77–79) than men (72, SD 14, CI 72–73). 43% of the patients were 80 years or older by the time of the primary amputation (Figure 3). The mortality rate of the registered patients was 19% within 6 months and 24% within the 1st year after the last registered amputation. The 1-year mortality rate after TFA was 40%, after KD 38%, and after TTA 24%. In 85% of the patients with a registered primary diagnosis, amputation was due to diabetes and/or vascular disease (Table 2). Amputation due to vascular disease without diabetes was reported in 45% of the female patients and in 32% of the male patients (p < 0.0001). 93% of the patients had at least 1 comorbidity, of which the most common were heart disease, lung disease, neurological disease, stroke, or dementia. Smoking habits were registered for 2,315 patients. At the time of amputation, 39% were non- smokers, 35% previous smokers (not smoked within the last year), 24% current smokers, and 2% were consuming other nicotine products. Amputation data 89% of the registrations were unilateral amputations and 80% were primary ones. 14% were re-amputations to a more proximal level and 6% revisions at the same level. TTA was most

TTA (n = 408)

KD (n = 39)

TFA (n = 110)

TTA (n = 408)

KD (n = 39)

TFA (n = 110)

Figure 5. Use of walking aids when walking with the prosthesis at home (left panel) and outdoors (right panel) at 12-months’ followup after unilateral amputation. Table 2. Underlying diagnosis leading to lower limb amputation and sex differences. Values are number (%) Underlying diagnosis

Total n = 5,544

Women n = 2,194

Men n = 3,350

Diabetes with or without vascular disease 2,475 (45) 779 (36) 1,696 (51) Vascular disease without diabetes 1,909 (34) 955 (44) 954 (28) All other diagnoses 1,160 460 (20) 700 (21) Infection not related to diabetes or vascular disease 207 (4) Trauma 215 (4) Other (e.g., tumor, congenital or acquired deformity) 367 (7) Diagnosis unknown or not registered 371 (7) There were statistically significant differences between men and women when underlying diagnosis leading to amputation was grouped as following: diabetes with/without vascular disease, vascular disease without diabetes, and all other diagnoses, p < 0.0001.

common (47%) followed by TFA (26%), minor amputation (partial foot amputation distally to ankle level) (20%), KD (7%), hip disarticulation, or hemipelvectomy (< 1%). Figure 4 shows sex differences regarding major amputation levels. 10% of the patients with primary TTA underwent re-amputation to a more proximal level. The most frequently registered surgical technique for TTA was sagittal flaps (72%) followed by anterior/posterior flaps (14%), long posterior flaps (9%), and skew flaps (4%). Regional differences were seen when considering the use of sagittal flaps, ranging from 33% to 85%. Primary skin closure was performed with sutures in 67% of our cases, with staples in 21%, and open treatment was registered in 2%. In 10% of cases, negative pressure wound therapy was applied addition-

Acta Orthopaedica 2020; 91 (4): 464–470

467

Table 3. Prosthetic Use score and LCI-5L score at 12- and 24-months’ follow-up for patients after unilateral transtibial amputation and transfemoral amputation depending on underlying diagnosis DIagnosis

Prosthetic Use score a 12 months 24 months n mean (SD) n mean (SD)

Transtibial amputation Diabetes and/or vascular disease 407 43 (32) 193 55 (33) All other diagnoses c 83 51 (33) 51 60 (36) Transfemoral amputation Diabetes and/or vascular disease 74 16 (21) 35 16 (23) All other diagnoses c 49 28 (32) 33 34 (35) a Prosthetic Use score: score 0–100 (= best possible). b Locomotor Capability Index–5L: score 0–56 (= best possible). c Includes, e.g., trauma, tumor, infection, and not-specified diagnoses.

ally. Postoperative residual limb care after TTA included in 95% of the cases a rigid dressing followed by compression treatment with a silicone liner, sometimes combined with an elastic stump shrinker. Liner therapy was in 79% of cases started within 3 weeks postoperatively. Of the 618 patients registered with bilateral amputations, 68% had TTA at least on 1 side. Information on antibiotic prophylaxis has been added recently, and of the registered 1,385 cases, no antibiotics were given in 2%, 25% received peroperative prophylaxis, 6% postoperative, and 67% per- and postoperative treatment. Prosthetic supply and self-reported prosthetic use Postoperatively, 55% of patients with TTA, 25% with KD, and 21% with TFA were assessed as potential users of a functional prosthesis. Of the 2,652 registered prosthetic supplies, 79% were TTA prostheses. The most common type of TTA prostheses included a liner, had vacuum suspension (71%), and an energy-storing foot (79%). TFA prostheses included a large variation of different prosthetic knee components, among which 40% were more advanced knee components such as pneumatic, hydraulic, and/or microprocessor-controlled knee units. Time from final-level amputation to fitting of the first individual TTA prosthesis was md 69 days (6–500, n = 837) and for TFA 97 days (19–484, n = 158), showing for TTA a decrease over time from 79 days during the first years of registration (2011–2013) to 56 days in 2017–2018. Time from surgery to start of TTA prosthetic rehabilitation was md 82 days (5–484, n = 766) and 112 days (19–490, n = 165) after TFA. Even here a decrease in time to TTA prosthetic rehabilitation could be stated, from md 87 days during 2011–2013 to 64 days in 2017–2018. Table 3 illustrates the Prosthetic Use Score for patients after unilateral TTA and TFA at 12- and 24-months’ follow-up, respectively. In patients with unilateral TTA due to diabetes and/or vascular disease the mean time to perform the TUG test at 12 months was 26 seconds (SD 19, n = 159) and at 24 months 24 seconds (SD 17, n = 74). In patients with unilateral

LCI-5L Total score b 12 months 24 months n median (range) n median (range) 458 33 (0–56) 77 43 (3–56)

149 37 (0–56) 44 45 (1–56)

63 16 (0–56) 46 34 (1–56)

25 22 (0–52) 22 43 (8–56)

TTA due to other diagnoses the TUG was 17 seconds (SD 9, n = 54) and 16 seconds (SD 11, n = 23) at 12 and 24 months, respectively. Baseline and follow-up patient-reported outcome LCI-5L total score prior to amputation was md 43 (0–56, n = 1505), and lower for women (md 36, n = 555) than for men (md 47, n = 950) (p < 0.001). At follow-up LCI-5L scores decreased (Table 3), with lower LCI advanced scores involving more demanding abilities such as walking while carrying an object, walking on stairs without a handrail, or getting up from the floor. After both TTA and TFA, the LCI-5L scores were higher in patients with amputations due to reasons other than diabetes and/or vascular disease. At follow-up, 81% of the registered patients had returned to the same kind of accommodation as before amputation. About half of our patients used walking aids prior to amputation, and around one-third additionally used a wheelchair. At 12 months’ follow-up most patients used some kind of walking aid together with the prosthesis both at home and outdoors (Figure 5). In addition, > 80% used a wheelchair. Patients with TFA more often than TTA patients reported not walking at all. At 12- and 24-months’ follow-up, residual limb pain was experienced by 44% and 48% of our patients, respectively, and phantom limb pain by 73% and 69%. The EQ-5D-5L index in our sample at 12 months was mean 0.57 (SD 0.3, n = 188) and at 24 months 0.56 (SD 0.32, n = 113) in patients with unilateral TTA, and 0.47 (SD 0.36, n = 59) and 0.51 (SD 0.37, n = 43) respectively in patients with unilateral KD or TFA.

Discussion To our knowledge, SwedeAmp is the first national quality registry that provides nationwide data on patients undergoing LLA including surgical facts, rehabilitation outcome, and details on prosthetic supply.

468

Our data confirm LLA patients in the industrialized world to be old and fragile (Dillingham et al. 2002) with high morbidity and mortality rates. Mobility has previously been reported to be related to satisfaction and quality of life for patients after major LLA (Sinha et al. 2011). Our data showed that the use of walking aids and wheelchair were mandatory for the majority of our patients 1 year after amputation. Moreover, depending on amputation level, not walking outdoors was reported among 27–48% of patients, which increases isolation and dependency. In SwedeAmp most patients underwent unilateral TTA as a result of diabetes and/or vascular disease. This group showed an increased falling risk as the mean TUG time was beyond the 19 seconds Dite et al. (2007) reported to indicate a risk of falling for patients with LLA. Our results emphasize the importance of adequate prosthetic rehabilitation and provision of other assistive devices for patients with major LLA to regain mobility. SwedeAmp reports females to be older than males by the time of the first registered amputation. Moreover, women were diagnosed more often with vascular disease without diabetes and, probably as a consequence, underwent TFA more often compared with men. This is in line with previous literature (Singh et al. 2008, Davie-Smith et al. 2017b). A possible cause might be the protective effect of estrogen against atherosclerosis before menopause (Vavra and Kibbe 2009, Boese et. al 2017) and thus women may develop vascular disease and related complications later in life. In addition to sex differences, TFA patients scored lower than TTA patients in LCI-5L, Prosthetic Use Score, and EQ-5D-5L. In conclusion, according to our patient sample, female patients in Sweden seem to have worse preconditions for regaining mobility, independence, and general health due to the higher incidence of older age at the time of amputation in combination with the loss of the knee. On the other hand, hypothetically, women might have had a healthier life for longer than men previous to the amputation. The optimal amputation level is difficult to define. With regard to lower mobility scores and limited use of prostheses after KD/TFA compared with TTA, the surgical aim should be to save the knee (Sansam et al. 2009). On the other hand, vascular impairment is usually worse below the knee level, resulting in an increased risk of revision or re-amputation after TTA compared with TFA. In this sample 10% of the primary TTA patients underwent re-amputation to a more proximal level. Moxey et al. (2010) reported that only 3 of 10 regions in England managed to achieve a TTA/TFA rate greater than 1, a figure which the authors stated to be a quality mark of amputation care. Residual limb pain and phantom limb pain are common after LLA with incidences up to 70% (Ehde et al. 2000, Morgan et al. 2017). SwedeAmp reports that over 40% of patients experience residual limb pain and about 70% phantom limb pain at least to some degree, without reduction over time. Optimal pain control and surgical technique should be sought to prevent long-lasting pain problems.

Acta Orthopaedica 2020; 91 (4): 464–470

In Sweden the use of a postoperative liner for residual limb compression is standard in the postoperative care after TTA (Johannesson et al. 2004). Liner compression was commonly started within 3 weeks after surgery. Regardless, median time from TTA to first fitting of prosthesis was 10 weeks, and a delay of a further 2 weeks was seen until prosthetic rehabilitation started. The cause for this delay cannot be identified by our data. However, the time to the first prosthesis in our material is clearly shorter than the mean of 145 days reported from the United States (Resnik and Borgia 2015). A positive trend could be noted in the registry with decreasing numbers of days to first TTA prosthesis, from md 79 days during the first years of registration (year 2011–2013) to md 56 days (year 2017–2018). With regard to the time from TTA to first fitting of individual prosthesis, the shortest time record (6 days) may be explained by the use of methods involving a laminated socket being produced directly on the residual limb, allowing the start of prosthetic use even before wound healing. However, until now, specific registration of methods involving a socket being directly fitted to the residual limb has not been included in the registry and therefore it cannot be excluded that this short time interval was due to misinput. EQ-5D-5L is a frequently used index to estimate a patient’s general health; however, it is sparsely used in LLA research. The average EQ-5D-5L index at 12- and 24-months’ follow-up was between 0.47 and 0.56. For comparison, patients undergoing orthopedic surgery in general increase from a preoperative mean of 0.54 to a postoperative mean of 0.72 (Jansson and Granath 2011) and patients with acute coronary syndromes score 0.82 one year after treatment (Gencer et al. 2016). The low scores of amputees show the general morbidity of these patients, and the mortality rate of our patients of 24% within the first year after amputation supports this. Limitations This study is a retrospective registry study and the available data were limited. Interesting facts such as decision-making on the amputation level cannot be detected by the registry. SwedeAmp has not yet gained full coverage and, thus far, our data cannot be taken as representative for Sweden. In some regions, registration of surgical data was done at rehabilitation units with the consequence that only patients who have proceeded to prosthetic rehabilitation are registered from those regions. Thus, surgical data should be interpreted with caution and registration of minor amputations, re-amputations, and revisions is probably underreported. Moreover, mortality and outcome can be considered to represent the best possible indications, as patients not attending prosthetic rehabilitation are so far underrepresented in the registry. Therefore, we cannot make any statement on how patients never reaching prosthetic supply rate their quality of life. Our study only sparsely presents the separate outcome for patients with amputations due to diagnoses other than diabetes and/or vascular disease or for patients with bilateral amputations, due to small numbers in

Acta Orthopaedica 2020; 91 (4): 464–470

these sub-groups to date. SwedeAmp is continuously being adapted and several variables have been added more recently, e.g., several surgical details such as antibiotic treatment and the prosthetic comfort score (Hanspal et al. 2003). Not all variables are mandatory, leading to different numbers depending on the variable. Conclusion We found worse functional outcome after TFA compared with TTA. Female patients were older by the time of amputation and amputation was performed at a higher level. Time from amputation to prosthetic supply and training has decreased during the most recent years. The results give a general insight into the patient group (dominated by the frail elderly) and the outcomes after major amputation. With increasing coverage, SwedeAmp may provide deeper knowledge with regard to patients undergoing LLA in Sweden and help to identify associations between the patient’s preoperative preconditions, surgical facts, the prosthetic supply, and the postoperative outcome. Supplementary data Table 1 is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2020. 1756101

IK, BS, HÖ, and KH designed the study; IK and KH collected and interpreted the data; IK and KH wrote the first version and all contributing authors together reviewed the manuscript in its final form. The authors disclosed receipt of the following financial support: ALF Skåne and FoU Skåne (IK) and a Swedish Government research grant (ALFGBG773581) (KH). In addition, the authors wish to express their gratitude to Dr Jan Larsson, formerly on the SwedeAmp Board, for valuable assistance in preparing this paper. Acta thanks Hans E Berg and Klaus Kjær Petersen for help with peer review of this study.

Boese A C, Kim S C, Yin K J, Lee J P, Hamblin M H. Sex differences in vascular physiology and pathophysiology: estrogen and androgen signaling in health and disease. Am J Physiol Heart Circ Physiol 2017; 313(3): H524-H45. Buckley C M, O’Farrell A, Canavan R J, Lynch A D, De La Harpe D V, Bradley C P, et al. Trends in the incidence of lower extremity amputations in people with and without diabetes over a five-year period in the Republic of Ireland. PLoS One 2012; 7(7): e41492. Davie-Smith F, Coulter E, Kennon B, Wyke S, Paul L. Factors influencing quality of life following lower limb amputation for peripheral arterial occlusive disease: a systematic review of the literature. Prosthet Orthot Int 2017a; 41(6): 537-47. Davie-Smith F, Paul L, Nicholls N, Stuart W P, Kennon B. The impact of gender, level of amputation and diabetes on prosthetic fit rates following major lower extremity amputation. Prosthet Orthot Int 2017b; 41(1): 19-25. Dillingham T R, Pezzin L E, MacKenzie E J. Limb amputation and limb deficiency: epidemiology and recent trends in the United States. South Med J 2002; 95(8): 875-83.

469

Dite W, Connor H J, Curtis H C. Clinical identification of multiple fall risk early after unilateral transtibial amputation. Arch Phys Med Rehabil 2007; 88(1): 109-14. Ehde D M, Czerniecki J M, Smith D G, Campbell K M, Edwards W T, Jensen M P, et al. Chronic phantom sensations, phantom pain, residual limb pain, and other regional pain after lower limb amputation. Arch Phys Med Rehabil 2000; 81(8): 1039-44. Ephraim P L, Dillingham T R, Sector M, Pezzin L E, Mackenzie E J. Epidemiology of limb loss and congenital limb deficiency: a review of the literature. Arch Phys Med Rehabil 2003; 84(5): 747-61. Fortington L V, Rommers G M, Geertzen J H, Postema K, Dijkstra P U. Mobility in elderly people with a lower limb amputation: a systematic review. J Am Med Dir Assoc 2012; 13(4): 319-25. Fortington L V, Geertzen J H, van Netten J J, Postema K, Rommers G M, Dijkstra P U. Short and long term mortality rates after a lower limb amputation. Eur J Vasc Endovasc Surg 2013a; 46(1): 124-31. Fortington L V, Rommers G M, Postema K, van Netten J J, Geertzen J H, Dijkstra P U. Lower limb amputation in Northern Netherlands: unchanged incidence from 1991–1992 to 2003–2004. Prosthet Orthot Int 2013b; 37(4): 305-10. Franchignoni F, Orlandini D, Ferriero G, Moscato T A. Reliability, validity, and responsiveness of the locomotor capabilities index in adults with lower-limb amputation undergoing prosthetic training. Arch Phys Med Rehabil 2004; 85(5): 743-8. Gencer B, Rodondi N, Auer R, Nanchen D, Raber L, Klingenberg R, et al. Health utility indexes in patients with acute coronary syndromes. Open Heart 2016; 3(1): e000419. Hagberg K, Branemark R, Hagg O. Questionnaire for persons with a Transfemoral Amputation (Q-TFA): initial validity and reliability of a new outcome measure. J Rehabil Res Dev 2004; 41(5): 695-706. Hanspal R S, Fisher K, Nieveen R. Prosthetic socket fit comfort score. Disabil Rehabil 2003; 25(22): 1278-80. Jansson K A, Granath F. Health-related quality of life (EQ-5D) before and after orthopedic surgery. Acta Orthop 2011; 82(1): 82-9. Johannesson A, Larsson G U, Oberg T. From major amputation to prosthetic outcome: a prospective study of 190 patients in a defined population. Prosthet Orthot Int 2004; 28(1): 9-21. Johannesson A, Larsson G U, Ramstrand N, Turkiewicz A, Wirehn A B, Atroshi I. Incidence of lower limb amputation in the diabetic and nondiabetic general population: a 10-year population-based cohort study of initial unilateral, contralateral and re-amputations. Diabetes Care 2008; epub Nov 10. Jones W S, Patel M R, Dai D, Vemulapalli S, Subherwal S, Stafford J, et al. High mortality risks after major lower extremity amputation in Medicare patients with peripheral artery disease. Am Heart J 2013; 165(5): 809-15, 15 e1. Larsson B, Johannesson A, Andersson I H, Atroshi I. The Locomotor Capabilities Index; validity and reliability of the Swedish version in adults with lower limb amputation. Health Qual Life Outcomes 2009; 7:44. Morgan S J, Friedly J L, Amtmann D, Salem R, Hafner B J. Cross-sectional assessment of factors related to pain intensity and pain interference in lower limb prosthesis users. Arch Phys Med Rehabil 2017; 98(1): 105-13. Moxey P W, Hofman D, Hinchliffe R J, Jones K, Thompson M M, Holt P J. Epidemiological study of lower limb amputation in England between 2003 and 2008. Br J Surg 2010; 97(9): 1348-53. Norvell D C, Turner A P, Williams R M, Hakimi K N, Czerniecki J M. Defining successful mobility after lower extremity amputation for complications of peripheral vascular disease and diabetes. J Vasc Surg 2011; 54(2): 41219. Raichle K A, Hanley M A, Molton I, Kadel N J, Campbell K, Phelps E, et al. Prosthesis use in persons with lower- and upper-limb amputation. J Rehabil Res Dev 2008; 45(7): 961-72. Resnik L, Borgia M. Predicting prosthetic prescription after major lower-limb amputation. J Rehabil Res Dev 2015; 52(6): 641-52. Samuelsson K A, Toytari O, Salminen A L, Brandt A. Effects of lower limb prosthesis on activity, participation, and quality of life: a systematic review. Prosthet Orthot Int 2012; 36(2): 145-58.

470

Sansam K, Neumann V, O’Connor R, Bhakta B. Predicting walking ability following lower limb amputation: a systematic review of the literature. J Rehabil Med 2009; 41(8): 593-603. Schoppen T, Boonstra A, Groothoff J W, de Vries J, Goeken L N, Eisma W H. Physical, mental, and social predictors of functional outcome in unilateral lower-limb amputees. Arch Phys Med Rehabil 2003; 84(6): 803-11. Singh R, Hunter J, Philip A, Tyson S. Gender differences in amputation outcome. Disabil Rehabil 2008; 30(2): 122-5. Sinha R, Van Den Heuvel W J. A systematic literature review of quality of life

Acta Orthopaedica 2020; 91 (4): 464–470

in lower limb amputees. Disabil Rehabil 2011; 33(11): 883-99. Sinha R, van den Heuvel W J, Arokiasamy P. Factors affecting quality of life in lower limb amputees. Prosthet Orthot Int 2011; 35(1): 90-6. van Netten J J, Fortington L V, Hinchliffe R J, Hijmans J M. Early post-operative mortality after major lower limb amputation: a systematic review of population and regional based studies. Eur J Vasc Endovasc Surg 2016; 51(2): 248-57. Vavra A K, Kibbe M R. Women and peripheral arterial disease. Women’s Health 2009; 5(6): 669-83.

Acta Orthopaedica 2020; 91 (4): 471–478

471

Can MRI differentiate between atypical cartilaginous tumors and high-grade chondrosarcoma? A systematic review Claudia DECKERS 1, Maarten J STEYVERS 2, Gerjon HANNINK 3, H W Bart SCHREUDER 1, Jacky W J DE ROOY 2, and Ingrid C M VAN DER GEEST 1 1 Department 3 Department

of Orthopedics, Radboud University Medical Center, Nijmegen; 2 Department of Radiology, Radboud University Medical Center, Nijmegen; of Operating Rooms, Radboud University Medical Center, Nijmegen, The Netherlands Correspondence: claudia.deckers@radboudumc.nl Submitted 2019-10-01. Accepted 2020-03-31.

Background and purpose — Adequate staging of chondroid tumors at diagnosis is important as it determines both treatment and outcome. This systematic review provides an overview of MRI criteria used to differentiate between atypical cartilaginous tumors (ACT) and high-grade chondrosarcoma (HGCS). Patients and methods — For this systematic review PubMed and Embase were searched, from inception of the databases to July 12, 2018. All original articles describing MRI characteristics of pathologically proven primary central chondrosarcoma and ACT were included. A quality appraisal of the included papers was performed. Data on MRI characteristics and histological grade were extracted by 2 reviewers. Meta-analysis was performed if possible. The study is registered with PROSPERO, CRD42018067959. Results — Our search identified 2,132 unique records, of which 14 studies were included. 239 ACT and 140 HGCS were identified. The quality assessment showed great variability in consensus criteria used for both pathologic and radiologic diagnosis. Due to substantial heterogeneity we refrained from pooling the results in a meta-analysis and reported non-statistical syntheses. Loss of entrapped fatty marrow, cortical breakthrough, and extraosseous soft tissue expansion appeared to be present more often in HGCS compared with ACT. Interpretation — This systematic review provides an overview of MRI characteristics used to differentiate between ACT and HGCS. Future studies are needed to develop and assess more reliable imaging methods and/or features to differentiate ACT from HGCS.

The incidence of chondrosarcoma of bone appears to have been increasing during the last decade and is now reported to be the most common primary malignant bone tumor in several countries (Thorkildsen et al. 2018, van Praag et al. 2018). Conventional chondrosarcoma is the most common subtype of chondrosarcoma. Other subtypes of chondrosarcoma (e.g., juxtacortical, mesenchymal, or secondary chondrosarcoma) are rare and show different radiologic appearance and clinical behavior (Bindiganavile et al. 2015). Conventional chondrosarcoma is classified into the histological grades 1 (currently known as atypical cartilaginous tumor [ACT]), 2, and 3. The metastatic potential, and therefore the disease-specific survival, correlates with the histological grade (Fletcher et al. 2013, Laitinen et al. 2018, Thorkildsen et al. 2018). ACTs rarely metastasize and are therefore reclassified as an intermediate type of tumor, not a malignancy (Fletcher et al. 2013). Due to the increase in patients undergoing MRI examinations for joint-related complaints, the incidental detection of ACT has increased substantially (van Praag et al. 2018). With the increasing incidence of ACT, clear radiologic criteria to differentiate ACT from high-grade chondrosarcoma (i.e., grades 2 and 3) become more and more important. Adequate staging of chondroid tumors at diagnosis is important as it determines both treatment and prognosis. High-grade chondrosarcomas behave aggressively. Between 10% and 30% of grade 2 and about 70% of grade 3 chondrosarcomas metastasize (Evans et al. 1977). Hence, high-grade chondrosarcoma (HGCS) requires wide en bloc resection with free surgical margins. In contrast, ACTs are intermediate tumors and can be treated either with intralesional curettage and local adjuvant or nonoperatively with regular follow-up when located in the long bones (Deckers et al. 2016).

472

Due to the heterogenous composition of chondroid tumors, diagnostic biopsy is unreliable in assessing the genuine histological grade and malignant potential of chondrosarcomas (Laitinen et al. 2018). Therefore, physicians need to rely on imaging and clinical findings (e.g., pain is more common in HGCS) to differentiate ACT from HGCS. Imaging evaluation of cartilaginous and other bone tumors is generally based on multimodal assessment including at least conventional radiography and MRI (Nascimento et al. 2014). During the most recent decades research has focused mainly on differentiating enchondroma from chondrosarcoma (Choi et al. 2013, Douis et al. 2014, Crim et al. 2015, Lisson et al. 2018). New insights have shown that both enchondroma and ACT located in the long bones can be observed without treatment (Deckers et al. 2016, Sampath Kumar et al. 2016, Chung et al. 2018). These insights make the differentiation between ACT and HGCS clinically relevant. Currently, literature on differentiating ACT from HGCS is sparse and clear radiologic criteria are lacking. Therefore, we performed a systematic review to provide an overview of MRI characteristics used to date to differentiate between ACT and HGCS. â&#x20AC;&#x192;

Methods The aim of this systematic review is to provide an overview of MRI characteristics used to differentiate between atypical cartilaginous tumors (ACT) and high-grade chondrosarcoma (HGCS). The inclusion criteria and method of analysis were specified in advance and documented in a PROSPERO protocol (CRD42018067959). This study was conducted and reported according to PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) and MOOSE guidelines. Search strategy and selection of studies The search strategy, composed of 3 elements (histology, MRI, and chondrosarcoma), was developed in collaboration with information specialists from the medical library of the Radboud University Medical Center Nijmegen, the Netherlands. The detailed search strategy can be found in Table 1 (see Supplementary data). No limits (e.g., language or publication date) were used. The search strategy was carried out in Pubmed and Embase (last search performed July 12, 2018). Additionally, reference lists of the included studies and of relevant reviews were screened for potentially relevant papers. After removal of duplicates, all unique records were imported into EROS (Early Review Organizing Software, Buenos Aires, Argentina) to allocate references randomly to 2 independent reviewers (CD, MS) responsible for screening and selection. Discrepancies were solved by discussion. During the first screening phase, original studies (i.e., no case reports, conference proceedings, systematic reviews) were included if they mentioned the combination of chon-

Acta Orthopaedica 2020; 91 (4): 471â&#x20AC;&#x201C;478

drosarcoma, histology/pathology, and imaging in title and/ or abstract. If not enough information was provided to make a valid judgment, the full text was evaluated. Full-text versions of all selected studies were screened and included if they met the pre-specified eligibility criteria: (1) preoperative MRI grading; (2) histopathological grading; (3) presence of MRI characteristics per chondrosarcoma grade; (4) primary central chondrosarcoma of bone; (5) adult patients. Types other than primary central chondrosarcoma of bone (e.g., juxtacortical, mesenchymal, or secondary chondrosarcoma) were excluded as these different types of tumor show different radiologic appearance and clinical behavior (Bindiganavile et al. 2015). Data extraction 2 independent reviewers (CD, MS) performed data extraction from each included study in a pre-piloted form. Information was extracted related to: study design, studied population, tumor location and size, tumor grade based on postoperative histology/pathology, pathology criteria used for diagnosis, type of MRI used, and MRI characteristics described per grade of chondrosarcoma (e.g., cortical breakthrough, soft tissue expansion). If studies included other types of chondrosarcoma (e.g., juxtacortical, mesenchymal, or secondary chondrosarcoma), only data related to central ACT and high-grade chondrosarcoma were extracted. If outcome data were presented incompletely, we tried to contact the authors to obtain the original data. A reminder was sent to those who did not reply within 2 weeks. When attempts to obtain original data failed, the article was excluded. According to the WHO classification, ACTs (i.e., chondrosarcoma grade 1) were categorized as low-grade chondrosarcoma (LGCS). Grade 2, grade 3 and dedifferentiated chondrosarcomas were categorized as HGCS (Fletcher et al. 2013). Quality appraisal The quality of the included studies was assessed using STROBE for the assessment of observational studies (Table 2, see Supplementary data). We are aware of the fact that the authors of STROBE did not develop their tool for methodological quality assessment. However, due to the lack of validated and accepted tools for such assessments of observational studies, STROBE is often used for this purpose (da Costa et al. 2011). In accordance with other studies, only 10 of the 22 items of the STROBE checklist were used for methodological assessment (da Costa et al. 2011, Shemesh et al. 2017). The other 12 of the 22 items were found not to contribute to the methodological assessment. In addition, we analyzed the quality of histopathology and MRI assessments. We checked whether there was (1) a description of the criteria used for diagnosis, (2) cited reference to consensus criteria used for diagnosis, and (3) if the diagnosis was established by an experienced musculoskeletal

Acta Orthopaedica 2020; 91 (4): 471–478

pathologist/radiologist (Shemesh et al. 2017). In addition, we added whether the pathologist and/or radiologist was blinded. If the level of experience of the pathologist/radiologist was not specified in the article, the authors were contacted. 2 reviewers (CD, MS) independently scored each item as: well described (+), partly described (±), or poorly/not described (–). Discrepancies were solved by discussion. No overall score was calculated, as we felt different study characteristics that are related to study quality cannot be judged as if they are of equal importance or interchangeable (Ioannidis 2011). Data analysis Heterogeneity was assessed by visual inspection of forest plots and quantified using the I2 and t2. The latter were calculated even when the judgment was made that calculating a pooled estimate was not justifiable (Higgins et al. 2003). Before undertaking a meta-analysis, we first checked whether the studies were similar enough to justify combining their results. If the features of studies were deemed not sufficiently similar to combine in a meta-analysis, we displayed the results of included studies in a forest plot but suppressed the summary estimate (Faber et al. 2016, Mueller et al. 2018, Reeves et al. 2019). If possible, pooled estimates of proportions with their corresponding 95% confidence intervals were calculated using the logit transformation using inverse-variance weighting within a random effects model framework. Between-study variance was quantified using the t2 statistic, estimated using the Sidik-Jonkman estimator. Data were analyzed using R version 3.4.3 (R Foundation for Statistical Computing, Vienna, Austria) using the meta package. Publication bias was assessed only if more than 10 studies were included in the meta-analysis. Data for different MR modalities (conventional MRI, diffusion-weighted imaging, dynamic contrast enhancement, and quantitative texture analysis) were reported separately, as these outcome measures were found not to be comparable to pool. MRI signal intensity, such as high signal on T1, can be related to several histopathological findings (e.g., hemorrhage, entrapped fat) and therefore does not necessarily indicate grade of chondrosarcoma. Therefore, we have chosen to exclude these MRI characteristics from our analysis. Funding and potential conflicts of interest There was no funding source for this study. None of the authors reported any conflict of interest.

Results Conducting our search strategy in PubMed and Embase retrieved 2,132 unique records. 5 additional relevant articles were found via cross-referencing. 2,123 articles were excluded because they did not meet our eligibility criteria (Figure 1).

473

Records identified through database searching n = 2,522

Additional records identified through other sources n=5

Records after duplicates removed n = 2,137 Records excluded on title and abstract n = 2,022 Full-text articles assessed for eligibility n = 115 Full-text records excluded (n = 101): – no original study, 26 – no preoperative MRI, 12 – no conventional chondrosarcoma, 9 – no histology, 2 – no MRI criteria per chondrosarcoma grade, 47 – no adult patients, 5 Studies included in qualitative synthesis n = 14

Figure 1. PRISMA flow diagram.

Errani et al. (2017) provided additional data on request. Consequently 14 articles were included in our systematic review (Table 3). 239 ACT and 140 HGCS were included in this systematic review. The following conventional MRI characteristics were reported by the included studies and analyzed: entrapped fat, perilesional bone marrow edema, internal lobular architecture, lobular outer margin, bone expansion, cortical thickening, scalloping, cortical breakthrough, periosteal edema, soft tissue edema, extra-osseous soft tissue expansion, ring and arc enhancement, solid enhancement, and central non-enhancing region. Due to substantial heterogeneity (I2 50–90%) and insufficient information to further investigate this heterogeneity we decided to refrain from pooling the results and only provide non-statistical syntheses. The reported presence of conventional MRI characteristics in both ACT and HGCS is displayed in separate forest plots but we suppressed the summary estimates (Figure 2). Only the most commonly reported MRI characteristics are shown in Figure 2; all other MRI characteristics can be found in Figure 3 (see Supplementary data). Both Kang et al. (2016) and Douis et al. (2018) compared maximum tumor size between ACT and HGCS. Kang et al. found a significant difference in tumor length between ACT (3.0 cm, SD 0.7 cm) and HGCS (7.4 cm, SD 2.7 cm), whereas Douis et al. did not find a difference in tumor length between ACT (11 cm, range 2.1–26 cm) and HGCS (13 cm, range 4.3–30 cm). 3 DWI studies were included describing apparent diffusion coefficient (ADC). Douis et al. (2015) found no statistically significant difference in both mean apparent diffusion coefficient (ADC) and minimum ADC between ACT and highgrade chondrosarcoma. Welzel et al. (2018) found in their subgroup analyses that chondrosarcoma grade 1 had statistically significantly higher,

474

Acta Orthopaedica 2020; 91 (4): 471–478

Table 3. Study characteristics Study

Study setting Patients (n) Tumor location (n)

MRI field strength

Intravenous contrast

Conventional MRI Crim et al. 2015 Retrospective 12 CS 1 Humerus (5), radius (1), NR + femur (4), fibula (2) Douis et al. 2014 Retrospective 28 ACT Humerus (58), femur (98), NR – 79 CS 1 tibia (24) 36 CS 2 13 CS 3 23 Dediff a b Douis et al. 2018 Retrospective 15 CS 1 Humerus (10), femur (9), 3T + 1980–2016 3 CS 2 tibia (3), fibula (1) 1 CS 3 4 Dediff a Errani et al. 2017 Retrospective 17 ACT Humerus (5), femur (9), 1.5T NR 1986–2015 tibia (3) Fayad et al. 2015 Retrospective 6 CS 2 Hands and feet (7) 1.5T + 1991–2014 1 CS 3 Kang et al. 2016 Retrospective 6 CS 1 Para-acetabular (21) 1.5T + 1993–2016 15 HGCS Liu et al. 2017 Retrospective 17 Dediff a NR 3T + 2008–2015 MacSweeney et al. Retrospective 8 Dediff a Humerus (2), femur (6) 1.0 or 1.5T + 2003 1995–2005 Yoo et al. 2009 Retrospective 28 LG Humerus (16), scapula (1), 1.0T or 1.5T + 1999–2008 14 HG pelvic bone (9), femur (15), fibula (1) Yoshimura et al. Retrospective 6 CS 1 Humerus (4), ulna (1), NR + 2013 1996–2011 10 CS 2 phalange (2), femur (7), 1 CS 3 tibia (1), calcaneus (1), rib (1) Diffusion weighted imaging Douis et al. 2015 Retrospective 5 ACT Humerus (19), rib (2), 3T – 2012–2013 15 CS 1 hand (3), spine (1), 3 CS 2 pelvis (5), femur (17), 2 CS 3 tibia (5) d 3 Dediff a Müller et al. 2016 Retrospective 8 CS 1 Skull base NR NR 2007–2012 Welzel et al. 2018 Retrospective 24 CS 1 Skull base 3T + 2009–2014 10 CS 2 1 CS 3 Dynamic contrast-enhanced MRI Douis et al. 2018 Retrospective 15 CS 1 Humerus (10), femur (9), 3T + 1980–2016 3 CS 2 tibia (3), fibula (1) 1 CS 3 4 Dediff a Quantitative texture analysis Lisson et al. Retrospective 11 CS1 NR 1.5 & 3T + 2018 NR = not reported. Dedifferentiated chondrosarcoma. b Study mentioned twice as different imaging modalities are used in the same study. c MRI characteristic not analyzed in our systematic review. d 24 enchondroma tumors are included in description of tumor location. a

MRI characteristics assessed

Length, deep endosteal scalloping, cortical breakthrough, soft tissue mass, gadolinium enhancement Bone marrow edema, soft tissue edema, bone expansion, cortical thickening, cortical destruction, active periostitis, soft tissue mass, tumor length Tumor length, endosteal scalloping, bone marrow edema, soft tissue edema, cortical destruction, periosteal reaction, bone expansion, macroscopic fat, calcification, soft tissue mass, hemorrhage Scalloping, soft tissue mass T1 signal c, T1 heterogeneity c, T2 hyperintense c, T2 heterogeneity c, bone marrow edema, soft tissue edema, gadolinium enhancement, soft tissue mass Length, high signal foci on T1 c, high signal on T1–T2-STIR c, soft-tissue mass, peritumoral edema, lobular border, acetabular cartilage destruction c, diffuse signal changes in acetabulum c, mass inside hip joint c, femoral head involvement c Patterns of bone destruction, periosteal reaction, matrix mineralization, soft tissue mass, enhancement pattern, signal intensity Soft tissue extension T1 signal c, entrapped fat within the tumor, lobular architecture preservation, cortical destruction, soft tissue mass, gadolinium enhancement Entrapped fat within the tumor, lobular architecture, ring and arc enhancement, T1 signal c, soft tissue mass, gadolinium enhancement Apparent diffusion coefficient

Apparent diffusion coefficient Apparent diffusion coefficient

Dynamic contrast-enhanced (DCE) MRI parameters; angle of DCE-MRI curve, absolute enhancement and relative enhancement Quantitative texture analysis to assess tumor heterogeneity

Acta Orthopaedica 2020; 91 (4): 471–478

Study ACT Yoshimura 2013 Yoo 2009 Douis 2018

475

Events

Total

Entrapped fat

Proportion (95% CI)

4 26 9

6 28 15

0.67 (0.22–0.96) 0.93 (0.76–0.99) 0.60 (0.32–0.84)

1 1 2

11 14 8

0.09 (0.00–0.41) 0.07 (0.00–0.34) 0.25 (0.03–0.65)

l 2 = 66%, τ2 = 0.9551

HGCS Yoshimura 2013 Yoo 2009 Douis 2018 l 2 = 0%, τ2 = 0

a Study ACT Douis 2014 Douis 2018 Kang 2016

0.0

0.2

0.4

0.6

0.8

1.0

Events

Total

Bone marrow edema

Proportion (95% CI)

12 7 6

107 15 6

0.11 (0.06–0.19) 0.47 (0.21–0.73) 1.00 (0.54–1.00)

26 6 15

72 8 15

0.36 (0.25–0.48) 0.75 (0.35–0.97) 1.00 (0.78–1.00)

l 2 = 89%, τ2 = 2.5630

HGCS Douis 2014 Douis 2018 Kang 2016 l 2 = 82%, τ2 = 2.4521

b Study ACT Douis 2014 Douis 2018 Kang 2016 Crim 2011

0.0

0.2

0.4

0.6

0.8

1.0

Events

Total

Cortical breakthrough

Proportion (95% CI)

1 5 0 3

107 15 6 12

0.01 (0.00–0.05) 0.33 (0.12–0.62) 0.00 (0.00–0.46) 0.25 (0.05–0.57)

40 8 15

72 8 15

0.56 (0.43–0.67) 1.00 (0.63–1.00) 1.00 (0.78–1.00)

l 2 = 77%, τ2 = 2.1821

HGCS Douis 2014 Douis 2018 Kang 2016 l 2 = 74%, τ2 = 3.1500

c Study ACT Yoshimura 2013 Yoo 2009 Douis 2014 Douis 2018 Kang 2016 Crim 2011 Errani 2017

0.0

Events

0.2

0.4

0.6

0.8

1.0

Total Extra-osseous soft tissue expansion Proportion (95% CI)

1 1 3 6 0 3 4

6 28 107 15 6 12 17

0.17 (0.00–0.64) 0.04 (0.00–0.18) 0.03 (0.01–0.08) 0.40 (0.16–0.68) 0.00 (0.00–0.46) 0.25 (0.05–0.57) 0.24 (0.07–0.50)

8 11 39 8 15 7 8 16

11 14 72 8 15 7 8 17

0.73 (0.39–0.94) 0.79 (0.49–0.95) 0.54 (0.42–0.66) 1.00 (0.63–1.00) 1.00 (0.78–1.00) 1.00 (0.59–1.00) 1.00 (0.63–1.00) 0.94 (0.71–1.00)

l 2 = 71%, τ2 = 1.2633

HGCS Yoshimura 2013 Yoo 2009 Douis 2014 Douis 2018 Kang 2016 Fayad 2015 Macsweeney 2003 Liu 2017

Discussion

l 2 = 65%, τ2 = 1.0629

d Study ACT Yoshimura 2013 Yoo 2009

0.0

0.2

0.4

0.6

0.8

1.0

Events

Total

Ring and arc enhancement

Proportion (95% CI)

4 28

6 28

0.67 (0.22–0.96) 1.00 (0.88–1.00)

0 14 3 8

11 14 6 12

0.00 (0.00–0.28) 1.00 (0.77–1.00) 0.50 (0.12–0.88) 0.67 (0.35–0.90)

l 2 = 75%, τ2 = 4.2184

HGCS Yoshimura 2013 Yoo 2009 Fayad 2015 Liu 2017 l 2 = 72%, τ2 = 2.4000

mean, minimum, maximum, and normalized ADC values than grade 2 chondrosarcoma in the skull base. Müller et al. (2016) measured the following ADC values in 8 chondrosarcoma grade 1 tumors of the skull base: mean ADC 2017 (SD 140) × 10–6 mm2/s. No ADC values of highgrade chondrosarcoma were measured. Only 1 study was found that described dynamic contrastenhanced (DCE) MRI parameters. Douis et al. (2018) found no statistically significant difference for the various DCE-MRI parameters (angle of the DCEMRI curve, absolute enhancement, and relative enhancement on DCE MRI) between LGCS and HGCS. Lisson et al. (2018) performed an MRI-based 3D texture analysis in which they compared enchondroma with lowgrade chondrosarcoma. No comparison with high-grade chondrosarcoma was made. The most promising texture parameters for differentiation were, among others, kurtosis (the magnitude of pixel distribution) in the contrast-enhanced T1-weighted images and entropy in non-contrast T1-weighted images. The quality appraisal of diagnosis is presented in Table 4 (see Supplementary data). The individual scored items on the STROBE checklist of each study can be found in the Supplementary data. Our assessment of the reporting quality shows great variability in consensus criteria used for diagnosis for both pathologic and radiologic diagnosis. Only in 7 of 14 studies did an experienced pathologist in musculoskeletal oncology perform pathologic assessment. In the other 7 studies level of expertise was not mentioned. In 10 of 14 studies MRI assessment was performed by experienced musculoskeletal radiologists.

0.0

0.2

0.4

0.6

0.8

1.0

Figure 2. Forest plots of proportions of the reported presence of (a) entrapped fat, (b) bone marrow edema, (c) cortical breakthrough, (d) extra-osseous soft tissue expansion, and (e) ring and arc enhancement on conventional MRI in atypical cartilaginous tumors (ACT) and high-grade chondrosarcoma (HGCS).

Correct diagnosis of chondrosarcoma grade is crucial in determining both treatment and prognosis. Therefore, we performed a systematic review to provide an overview of MRI characteristics used to differentiate between ACT and highgrade chondrosarcoma. Although we did not pool the overall results due to the considerable amount of heterogeneity, it appears that, compared with ACT, high-grade chondrosarcoma may present more often with the following MRI characteristics: loss of entrapped fatty marrow, cortical breakthrough, and extraosseous soft tissue expansion. These MRI findings are in line with the histopathological findings described by several authors (Brien et al. 1997, Yoo et al. 2009, Logie et al. 2013). In cartilaginous tumors production of chondroid matrix results in the typical lobulated growth pattern and the so-called ring and arc appearance (Logie et al. 2013). In HGCS these typical chondroid features become lost due to poor differentiation of cells. Chondrosarcoma cells actively infiltrate between

476

individual fat cells, compressing and eventually replacing them (Brien et al. 1997). Absence of areas of entrapped fat is therefore highly indicative of HGCS. In addition, invasion of Haversian systems leads to periosteal reaction. Eventually there is destruction of the cortex and invasion of soft tissue (Brien et al. 1997). Yoo et al. (2009) found that on gross pathological evaluation, a central non-enhancing region corresponded to an area of hemorrhagic cyst, necrosis, and/or yellow-brown soft tissue mass reflecting a myxoid change, all characteristics of malignant tumors. Due to the heterogeneity of cartilage tumors, areas of ACT can be seen in HGCS lesions. Therefore, the presence of MRI characteristics indicating ACT must be viewed in context and clinical findings must be taken into account. In addition, single MRI characteristics alone cannot differentiate between ACT and HGCS. The assessment of the clinical relevance of our findings is not straightforward. Heterogeneity was substantial (I2 50â&#x20AC;&#x201C;90%) in the majority of the analyses. Due to the considerable heterogeneity we decided not to perform a metaanalysis. Heterogeneity may be explained by either clinical and/or methodological diversity between included studies. Included studies showed great variability in tumor location within and between studies. Different bones (e.g., phalanges, femur) as well as types of bone (e.g., flat, long bones) were included in most studies, which might show different clinical behavior and radiologic appearance (Bindiganavile et al. 2015). We were unable to perform a sensitivity analysis on tumor location. In addition, heterogeneity might be caused by poor reliability between radiologists. The SLICED study group showed poor to slight reliability between radiologists for the subgroup of outcome-determined high-risk patients (SLICED Study Group 2007). However, the imaging modalities available for radiologists varied and different criteria were used. In those cases where MRI scans were available the reliability increased substantially. Zamora et al. (2017) showed fair interobserver agreement between orthopedic oncologists for diagnosis and grading of cartilaginous neoplasms. Nevertheless, no evaluator proposed observation or follow-up for lesions considered to be a malignant neoplasm. Limitations To reduce bias we excluded tumors other than primary central chondrosarcoma from our systematic review. Several studies were excluded as they included, e.g., secondary or periosteal chondrosarcoma as well and we were not able to extract data on the primary central chondrosarcoma (Varma et al. 1992, Geirnaerdt et al. 1993, De Beuckeleer et al. 1995, Geirnaerdt et al. 2000, Fritz et al. 2018). Excluding studies to reduce bias resulted in a limited number of tumors being included in this systematic review. Several studies have shown that both radiological and histopathological diagnosis of chondrosarcoma is subject to low

Acta Orthopaedica 2020; 91 (4): 471â&#x20AC;&#x201C;478

reproducibility, which may be caused by difficult and ambiguous definitions (SLICED Study Group 2007, Zamora et al. 2017). Different terminology has been used in chondrosarcoma literature during the past years, for example CLUMP (cartilaginous lesion of unknown malignant potential), borderline chondrosarcoma, or grade 0.5 CS, compromising comparability of studies. As can be seen in Table 4, several different grading methods have been used to assess the level of malignancy of chondrosarcoma. In addition, other imaging methods used (e.g., radiographs, CT) could have influenced the radiologist during MRI interpretation. Only Crim et al. (2015) and Fayad et al. (2015) stated that both radiographs and MRI were available for the radiologist. Other articles included did not report information on other imaging methods used but this could have been the case as combining different imaging methods is common practice. Possible interreader variability of chondrosarcoma grading may have resulted in misclassification bias in our systematic review. We would recommend a standardized grading method and terminology for chondroid tumors to improve comparability between studies and decrease the amount of bias. Third, we are aware of the fact that the authors of STROBE did not develop their tool for methodological quality assessment. Due to the lack of validated and accepted tools for such assessments for observational studies, STROBE is often used for this purpose (da Costa et al. 2011). We have used relevant items of the STROBE tool to give an overview of the methodology through the included papers. As shown by Mueller et al. (2018) there is considerable disagreement on how systematic reviews of observational studies should be done. We agree that there is a need for a comprehensive source of methodological guidance, in particular for quality assessment of observational studies. This systematic review provides an overview of currently used MRI characteristics. Future studies are needed to develop and assess a reliable method for differentiating chondrosarcoma based on radiologic and clinical findings. Reliability could be increased by protocol-driven image acquisition for cartilaginous lesions and an easy to use grading system that could be reliably quantified. From this systematic review it appears that MRI may possibly be helpful to differentiate ACT from HGCS. Extraosseous soft tissue expansion and cortical breakthrough appear to be present more often in HGCS and entrapped fat presents more often in ACT. As a correct differentiation of ACT and HGCS is important, we recommend future studies to develop and assess more reliable imaging methods and/or features to differentiate ACT from HGCS. Supplementary data Tables 1, 2, and 4 and Figure 3 are available as supplementary data in the online version of this article, http://dx.doi.org/10. 1080/17453674.2020.1763717

Acta Orthopaedica 2020; 91 (4): 471–478

All authors contributed to the creation of the content, and the writing and revising of this manuscript. CD and MS both contributed to the literature search and screening for eligible articles. Acta thanks Clement Trovik and Nils Vetti for help with peer review of this study.

Aoki J, Sone S, Fujioka F, Terayama K, Ishii K, Karakida O, Imai S, Sakai F, Imai Y. MR of enchondroma and chondrosarcoma: rings and arcs of GdDTPA enhancement. J Comput Assist Tomogr 1991; 15(6): 1011-6. Bindiganavile S, Han I, Yun J Y, Kim H-S. Long-term outcome of chondrosarcoma: a single institutional experience. Cancer Res Treat 2015; 47(4): 897. Brien E W, Mirra J M, Kerr R. Benign and malignant cartilage tumors of bone and joint: their anatomic and theoretical basis with an emphasis on radiology, pathology and clinical biology. Skeletal Radiol 1997; 26(6): 325-53. Choi B-B, Jee W-H, Sunwoo H-J, Cho J-H, Kim J-Y, Chun K-A, Hong S-J, Chung H W, Sung M-S, Lee Y-S. MR differentiation of low-grade chondrosarcoma from enchondroma. Clin Imaging 2013; 37(3): 542-7. Chung B M, Hong S H, Yoo H J, Choi J-Y, Chae H-D, Kim D H. Magnetic resonance imaging follow-up of chondroid tumors: regression vs. progression. Skeletal Radiol 2018; 47(6): 755-61. Crim J, Schmidt R, Layfield L, Hanrahan C, Manaster B J. Can imaging criteria distinguish enchondroma from grade 1 chondrosarcoma? Eur J Radiol 2015; 84(11): 2222-30. da Costa B R, Cevallos M, Altman D G, Rutjes A W, Egger M. Uses and misuses of the STROBE statement: bibliographic study. BMJ open 2011:2010000048. Dahlin D C, Henderson E D. Chondrosarcoma, a surgical and pathological problem: review of 212 cases. J Bone Joint Surg 1956; 38(5): 1025-125. Dahlin D C, Beabout J W. Dedifferentiation of low-grade chondrosarcomas. Cancer 1971; 28(2): 461-6. De Beuckeleer L H, De Schepper A M, Ramon F, Somville J. Magnetic resonance imaging of cartilaginous tumors: a retrospective study of 79 patients. Eur J Radiol 1995; 21(1): 34-40. De Beuckeleer L H, De Schepper A, Ramon F. Magnetic resonance imaging of cartilaginous tumors: is it useful or necessary? Skeletal Radiol 1996; 25(2): 137-41. Deckers C, Schreuder B H, Hannink G, de Rooy J W, van der Geest I C. Radiologic follow-up of untreated enchondroma and atypical cartilaginous tumors in the long bones. J Surg Oncol 2016; 114(8): 987-91. Dorfman H D, Czerniak B, Bone tumors, in: H.D. Dorfman, B. Czerniak (Eds.), Benign Cartilage Lesions, Mosby, St. Louis, 1998, p. 263. Douis H, Singh L, Saifuddin A. MRI differentiation of low-grade from highgrade appendicular chondrosarcoma. Eur Radiol 2014; 24(1): 232-40. Douis H, Jeys L, Grimer R, Vaiyapuri S, Davies A. Is there a role for diffusion-weighted MRI (DWI) in the diagnosis of central cartilage tumors? Skeletal Radiol 2015; 44(7): 963-9. Douis H, Parry M, Vaiyapuri S, Davies A. What are the differentiating clinical and MRI-features of enchondromas from low-grade chondrosarcomas? Eur Radiol 2018; 28(1): 398-409. Enneking W F. A system of staging musculoskeletal neoplasms. Clin Orthop Rel Res 1986; (204):9-24. Errani C, Tsukamoto S, Ciani G, Akahane M, Cevolani L, Tanzi P, Kido A, Honoki K, Tanaka Y, Donati D M. Risk factors for local recurrence from atypical cartilaginous tumour and enchondroma of the long bones. Eur J Orthop Surg Traumatol 2017; 27(6): 805-11. Evans H L, Ayala A G, Romsdahl M M. Prognostic factors in chondrosarcoma of bone: a clinicopathologic analysis with emphasis on histologic grading. Cancer 1977; 40(2): 818-31. Faber T, Ravaud P, Riveros C, Perrodeau E, Dechartres A. Meta-analyses including non-randomized studies of therapeutic interventions: a methodological review. BMC Med Res Methodol 2016; 16(1): 35.

477

Fayad L M, Ahlawat S, Khan M S, McCarthy E. Chondrosarcomas of the hands and feet: a case series and systematic review of the literature. Eur J Radiol 2015; 84(10): 2004-12. Fletcher C D M, Bridge J A, Hogendoorn P, Mertens F. WHO classification of tumours of soft tissue and bone. Lyon, France: IARC Press; 2013. Fritz B, Müller D A, Sutter R, Wurnig M C, Wagner M W, Pfirrmann C W, Fischer M A. Magnetic Resonance Imaging–Based Grading of Cartilaginous Bone Tumors: Added Value of Quantitative Texture Analysis. Invest Radiol 2018; 53(11): 663-72. Geirnaerdt M, Bloem J, Eulderink F, Hogendoorn P, Taminiau A. Cartilaginous tumors: correlation of gadolinium-enhanced MR imaging and histopathologic findings. Radiology 1993; 186(3): 813-7. Geirnaerdt M, Hermans J, Bloem J L, Kroon H M, Pope T, Taminiau A, Hogendoorn P. Usefulness of radiography in differentiating enchondroma from central grade 1 chondrosarcoma. AJR Am J Roentgenol 1997; 169(4): 1097-104. Geirnaerdt M J, Hogendoorn P C, Bloem J L, Taminiau A H, van der Woude H-J. Cartilaginous tumors: fast contrast-enhanced MR imaging. Radiology 2000; 214(2): 539-46. Henderson E D, Dahlin D C. Chondrosarcoma of bone—a study of two hundred and eighty-eight cases. J Bone Joint Surg 1963; 45(7): 1450-8. Higgins J P, Thompson S G, Deeks J J, Altman D G. Measuring inconsistency in meta-analyses. BMJ 2003; 327(7414): 557-60. Hogendoorn P C W, Bovee J, Nielsen G P. World health organizationclassification of tumours of soft tissue and bone. In: Fletcher C D M, Bridge J A, Hogendoorn P C W, Mertens F (Eds). IARCPress, Lyon 2013, pp 264-8. Inwards C Y, Unni K K. Classification and grading of bone sarcomas. Hematol Oncol Clin North Am 1995; 9(3): 545-70. Inwards C H P. In: Fletcher C D M, Bridge J A, Hogendoorn P C W, Mertens F (Eds.) World Health Organization Classification of Tumours of Soft Tissue and Bone. IARC Press, Lyon 2013, pp 269–270. Ioannidis J P. Commentary: adjusting for bias: a user’s guide to performing plastic surgery on meta-analyses of observational studies. Int J Epidemiol 2011; 40(3): 777-9. Kang Y, Yuan W, Ding X, Wang G. Chondrosarcoma of the para-acetabulum: correlation of imaging features with histopathological grade. Radiol Med 2016; 121(12): 897-904. Laitinen M, Stevenson J, Parry M, Sumathi V, Grimer R, Jeys L. The role of grade in local recurrence and the disease-specific survival in chondrosarcomas. Bone Joint J 2018; 100(5): 662-6. Lisson C S, Lisson C G, Flosdorf K, Mayer-Steinacker R, Schultheiss M, von Baer A, Barth T F, Beer A J, Baumhauer M, Meier R. Diagnostic value of MRI-based 3D texture analysis for tissue characterisation and discrimination of low-grade chondrosarcoma from enchondroma: a pilot study. Eur Radiol 2018; 28(2): 468-77. Liu C, Xi Y, Li M, Jiao Q, Zhang H, Yang Q, Yao W. Dedifferentiated chondrosarcoma: radiological features, prognostic factors and survival statistics in 23 patients. PloS One 2017; 12(3): e0173665. Logie C I, Walker E A, Forsberg J A, Potter B K, Murphey M D. Chondrosarcoma: a diagnostic imager’s guide to decision making and patient management. In: Seminars in Musculoskeletal Radiology. New York: Thieme Medical Publishers; 2013. p 101-15. MacSweeney F, Darby A, Saifuddin A. Dedifferentiated chondrosarcoma of the appendicular skeleton: MRI-pathological correlation. Skeletal Radio 2003; 32(12): 671-8. Mercuri M, Campanacci L. Dedifferentiated chondrosarcoma. Skeletal Radiol 1995; 24(6): 409-16. Mirra M, Gold R, Downs J, Eckardt J J. A new histologic approach to the differentiation of enchondroma and chondrosarcoma of the bones. A clinicopathologic analysis of 51 cases. Clin Otyhop rel Res 1985; (201): 214-37. Mirra J M, Picci P, Gold R H. Bone Tumors: Clinical, Radiologic, andPathologic Correlations, Lea & Febiger, Philadelphia, 1989. Mueller M, D’Addario M, Egger M, Cevallos M, Dekkers O, Mugglin C, Scott P. Methods to systematically review and meta-analyse observational studies: a systematic scoping review of recommendations. BMC Med Res Methodol 2018; 18(1): 44.

478

Müller U, Kubik-Huch R A, Ares C, Hug E B, Löw R, Valavanis A, Ahlhelm F J. Is there a role for conventional MRI and MR diffusion-weighted imaging for distinction of skull base chordoma and chondrosarcoma? Acta Radiol 2016; 57(2): 225-32. Murphey M D, Flemming D J, Boyea S R, Bojescul J A, Sweet DE , Temple H T. Enchondroma versus chondrosarcoma in the appendicular skeleton: differentiating features. Radiographics 1998; 18(5): 1213-37. Murphey M D, Walker E A, Wilson A J, Kransdorf M J, Temple H T, Gannon F H. From the archives of the AFIP: imaging of primary chondrosarcoma: radiologic-pathologic correlation. Radiographics 2003; 23(5): 1245-78. Nascimento D, Suchard G, Hatem M, de Abreu A. The role of magnetic resonance imaging in the evaluation of bone tumours and tumour-like lesions. Insights Imaging 2014; 5(4): 419-40. Pritchard D J, Lunke R J, Taylor W F, Dahlin D C, Medley B E. Chondrosarcoma: a clinicopathologic and statistical analysis. Cancer 1980; 45(1): 149-57. Reeves B C, Deeks J J, Higgins J P, Shea B, Tugwell P, Wells G A, Group CNRSoIM. Including non-randomized studies on intervention effects. Cochrane Handbook for Systematic Reviews of Interventions 2019: 595620. Rozeman L, Cleton-Jansen A, Hogendoorn P. Pathology of primary malignant bone and cartilage tumours. Int Orthop 2006; 30(6): 437-44. Sampath Kumar V, Tyrrell P, Singh J, Gregory J, Cribb G, Cool P. Surveillance of intramedullary cartilage tumours in long bones. Bone Joint J 2016; 98(11): 1542-7. Sanerkin N. The diagnosis and grading of chondrosarcoma of bone a combined cytologic and histologic approach. Cancer 1980; 45(3): 582-94. Shemesh S, Acevedo-Nieves J, Pretell-Mazzini J. Treatment strategies for central low-grade chondrosarcoma of long bones: a systematic review of the literature and meta-analysis. Musculoskelet Surg 2017: 1-15. SLICED Study Group. Reliability of histopathologic and radiologic grading of cartilaginous neoplasms in long bones. J Bone Joint Surg Am 2007; 89(10): 2113-23.

Acta Orthopaedica 2020; 91 (4): 471–478

Thorkildsen J, Taksdal I, Bjerkehagen B, Haugland H K, Børge Johannesen T, Viset T, Norum O-J, Bruland Ø, Zaikova O. Chondrosarcoma in Norway 1990–2013: an epidemiological and prognostic observational study of a complete national cohort. Acta Oncol 2018: 1-10. van Praag V, Rueten-Budde A, Ho V, Dijkstra P, van der Geest I C, Bramer J A, Schaap G R, Jutte P C, Schreuder H B, Ploegmakers J. Incidence, outcomes and prognostic factors during 25 years of treatment of chondrosarcomas. Surg Oncol 2018; 27(3): 402-8. Varma D, Ayala A, Carrasco C, Guo S, Kumar R, Edeiken J. Chondrosarcoma: MR imaging with pathologic correlation. Radiographics 1992; 12(4): 687-704. Walden M J, Murphey M D, Vidal J A. Incidental enchondromas of the knee. AJR Am J Roentgenol 2008; 190(6):1 611-5. Welkerling H, Werner M, Delling G. Histologic grading of chondrosarcoma. A qualitative and quantitative analysis of 74 cases of the Hamburg bone tumor register. Pathologe 1996; 17(1): 18-25. Welkerling H, Kratz S, Ewerbeck V, Delling G. A reproducible and simple grading system for classical chondrosarcomas. Analysis of 35 chondrosarcomas and 16 enchondromas with emphasis on recurrence rate and radiological and clinical data. Virchows Arch 2003; 443(6): 725-33. Welzel T, Meyerhof E, Uhl M, Huang K, von Deimling A, Herfarth K, Debus J. Diagnostic accuracy of DW MR imaging in the differentiation of chordomas and chondrosarcomas of the skull base: a 3.0-T MRI study of 105 cases. Eur J Radiol 2018; 105: 119-24. Yoo H J, Hong S H, Choi J-Y, Moon K C, Kim H-S, Choi J-A, Kang H S. Differentiating high-grade from low-grade chondrosarcoma with MR imaging. Eur Radiol 2009; 19(12): 3008-14. Yoshimura Y, Isobe K-i, Arai H, Aoki K, Kito M, Kato H. Preoperative radiographic and histopathologic evaluation of central chondrosarcoma. Arch Orthop Trauma Surg 2013; 133(9): 1225-31. Zamora T, Urrutia J, Schweitzer D, Amenabar P P, Botello E. Do orthopaedic oncologists agree on the diagnosis and treatment of cartilage tumors of the appendicular skeleton? Clin Orthop Relat Res 2017; 475(9): 2176-86.

Acta Orthopaedica 2020; 91 (4): 479–484

479

Assessing variability and uncertainty in orthopedic randomized controlled trials Lauri RAITTIO 1 and Aleksi REITO 2 1 Tampere University, Faculty of Medicine and Health Technology, Tampere; 2 Department of Orthopaedics and Traumatology, Tampere University Hospital, Tampere, Finland Correspondence: lauri.raittio@tuni.fi Submitted 2020-02-16. Accepted 2020-03-27.

Background and purpose — Low statistical power remains endemic in clinical medicine including orthopedics and manifests as high uncertainty and wide confidence intervals (CI). We evaluated the reporting and correspondence between power calculation and observed data on key parameters of variability and uncertainty in orthopedic randomized controlled trials (RCTs). Material and methods — RCTs with 1:1 allocation published in 8 major orthopedic journals between 2016 and 2017 with one continuous primary outcome were included in the review. The components of power calculation and observed standard deviation (SD), mean difference (MD), and confidence interval (CI) of MD between groups were assessed for primary outcome. Results — 160 RCTs were included, of which 93 (58%) and 138 (86%) studies reported the estimated SD and MD in the power calculation, respectively. The median ratio of the estimated SD and SDs observed in the data was 1.0 (IQR –0.76 to 1.32) for 69 (43%) studies. Only 31 of 138 studies reported the CI of MD in primary outcome. In 42% of the negative studies, the estimated MD was included in the CI of the observed MD. Interpretation — The key parameters of data variability, both in power analyses and in final study results, were poorly reported. Low power in orthopedics may result from too high an estimated effect size due to an overoptimistic estimate of MD between study groups. In almost half of the studies, overlap of the CI of the observed MD and estimated MD suggested that the reported results of these studies were inconclusive.

Adequate statistical power is the cornerstone of reproducible and high-quality clinical research. High statistical power is needed to increase the likelihood that a study will detect an effect when there is an effect to be detected. According to the CONSORT statement (Schulz et al. 2010), power calculations are based on the estimated mean difference (MDest) between compared groups, the estimated standard deviation (SDest) or variability of the outcome at a particular point in time, and the chosen level of error, namely, type 1 and 2 errors. A complement of type 2 error is statistical power. Despite the increasing use of power calculations, low power among RCTs to find small and medium effect sizes still remains endemic in clinical medicine, including orthopedics (Button et al. 2013, Abdullah et al. 2015, Sabharwal et al. 2015, Szucs and Ioannidis 2017, Reito et al. 2020). In studies using a priori power analysis, low power may arise from overestimated mean difference (MD), from underestimated standard deviation (SD), or from both (Vickers 2003, Cook et al. 2018). In many orthopedic RCTs, a patient-level minimal clinically important difference (MCID) is currently the basis of the group-level MDest used in power calculations. Usually, the MDest used in power calculations represents the clinically relevant difference valued by the investigators (Ostelo et al. 2008, de Vet and Terwee 2010, Angst et al. 2017, Jayadevappa et al. 2017, Dabija and Jain 2019). In this study we use the terms “MDest” and “MCID” interchangeably. Small sample sizes will yield high uncertainty of the outcome variable, which may, in turn, manifest as wide confidence intervals (CIs) (Anderson 2019). The mainstay in the interpretation of negative trials is to declare no statistically significant difference or “no difference” between the study groups if the CI of MD (CIMD) between groups includes equivalence in means, i.e., zero difference. A more appropriate interpretation would be to interpret the CIMD to see which

480

values for group difference are excluded by the data based on the chosen confidence level (Gelman and Greenland 2019). In this systematic review, we investigated (1) the reporting of the key parameters of variability and uncertainty; (2) the correspondence of the SDest of the primary outcome used in the power analysis to that actually observed in the study population; (3) the overlap of the MDest between groups to the CIMD in the primary outcome between study groups, and (4) the difference in sample size and estimated effect size in studies with and without overlap in MDest and CIMD in orthopedic RCTs published in 8 journals in the years 2016 and 2017.

Material and methods Study selection We reviewed 8 journals focused on clinical orthopedic research, namely the Journal of Bone and Joint Surgery; Clinical Orthopaedics and Related Research; the Bone and Joint Journal; the American Journal of Sports Medicine; Arthroscopy; the Journal of Arthroplasty; Knee Surgery, Sports Traumatology, Arthroscopy; and Acta Orthopaedica. The electronic table of contents from the 2016 and 2017 volumes of each of the 8 journals were searched issue by issue in chronological order to identify any RCTs. All studies that claimed to be a 1:1 RCT were included in the analysis. Data extraction All selected studies were examined in detail. The use of power analysis and the type of primary outcome (continuous, binary, noninferiority, other) used in the studies was recorded. We used the primary outcome and the power outcome in this study interchangeably. If continuous primary outcome was used in the power analyses, we recorded the MDest and SDest used to derive the sample size estimate. The number of patients available, means, and estimate of variability (SD or standard error, SE) for both study groups (i.e., SD1, SD2) at the pre-specified or at the latest follow-up time point when the results were reported was recorded. If these were not reported, we assessed whether the authors had reported CIs for the primary outcome in the study groups (CI1 and CI2). In cases where the SDs of primary outcome (SD1 and SD2) for the study groups were not reported, they were calculated from the SEs (SE1 and SE2) or CIs (CI1 and CI2) if reported. For all studies where the SDs of primary outcome for the study group were reported or calculated, we also calculated the pooled SD (SDpooled) of the primary outcome in the study participants. This was calculated as described in the Cochrane handbook (Higgins and Green 2011). Assuming sample sizes were reported, SDpooled was calculated from the CIMD if the SDs for the study groups were not available. Finally, we assessed whether the authors had reported the CI for the MD between the groups (CIMD). However, if the observed CIMD was not found, it was calculated from the

Acta Orthopaedica 2020; 91 (4): 479–484

Superiority cannot be shown but clinically relevant difference can be excluded Superiority cannot be shown but clinically relevant difference favoring treament B can be excluded while similar difference favoring treatment A cannot Superiority cannot be shown but clinically relevant difference favoring treament A can be excluded while similar difference favoring treatment B cannot Superiority or clinically relevant difference favoring either treatment cannot be shown meaning study result is inconclusive

–MCID 0 +MCID Favors treatment A Favors treatment B Group difference

Figure 1. Interpretation of studies in which similarity between study groups cannot be rejected, i.e., “negative studies.”

SDpooled, assuming the authors had reported the sample size and the study group means for the primary outcome. Data assessment The ratio of the observed and estimated SDs (SD1/SDest, SD2/ SDest, and SDpooled/SDest) was calculated for each study. The median, inter-quartile range (IQR), and geometric mean (SD) values for these ratios were reported. For each study, the overlap of MDest with regard to the upper and lower boundary of CIMD was investigated. This was basically a unidirectional analysis, i.e., we checked whether the higher of the absolute values of the upper and lower limit of CIMD was smaller or higher than MDest used in power calculation of the study. Of the “negative” RCTs, i.e., those studies that reported statistically not-significant results, the proportion of studies with and without this bidirectional overlap was reported. In other words, we checked whether the lower limit of CIMD was higher or lower than the negative value of MDest or the upper limit of CIMD was higher or lower than the positive value of MDest. In 3 studies CIMD was not reported but authors declared significant or nonsignificant results referring to some p-value, and our calculation showed marginal compatibility of data with zero effect size, e.g. (CI –0.16 to 1.72). In these 3 studies, we classified the results to positive and negative groups using the classification of the authors. In the optimal situation, for “negative” studies, CIMD excludes both negative and positive MDest (Figure 1). The estimated effect size in each study was calculated by dividing MDest by SDest and the mean estimated effect sizes and sample sizes were compared between studies with and without overlap between MDest and CIMD. In addition, sample sizes were compared using the Mann–Whitney U-test.

Results Of the 254 RCTs identified in our study, 209 studies (82%) employed a priori power analysis. The primary outcome was

Acta Orthopaedica 2020; 91 (4): 479–484

481

Table 1. Reporting of power calculations and study results among 160 orthopedic RCTs

RCTs with 1:1 parallel design n = 254 Excluded RCTs without power calculation n = 45

Factor

Power analysis 160 Estimated SD reported 93 (58) Estimated MD reported 138 (86) Observed variability of primary outcome measure SDs of primary outcome for study groups 89 (57) SEs for study groups 4 (3) CIs for study groups 19 (12) CI for mean difference 3 (2) Data on mean difference between groups Reported CI for mean difference 37 (23) Calculated CI for mean difference 75 (47)

RCTs with power calculation n = 209 Excluded Other than one continuous primary outcome n = 49 RCTs with power calculation for continuous primary outcome n = 160

SDest missing from power calculation n = 67

SDest as part of power calculation n = 93

SD/SE/CI of primary outcome not accessible n = 22 SD (n = 45): – reported, 37 – calculated, 8

MDest as part of power calculation n = 138

SD/SE/CI of primary outcome not accessible n = 26

SD (n = 67): – reported, 52 – calculated, 15

MDest missing from power calculation n = 22

CI of MD of primary outcome not accessible n = 39

CI for MD (n = 99): – reported, 31 – calculated, 68

n (%)

CI of MD of primary outcome not accessible n=9

CI for MD (n = 13): – reported, 6 – calculated, 7

Figure 2. Flow chart of study selection.

binary in 26 (12%), several in 6 (3%), not reported in 2 (1%), and generic in 7 (3%) of these studies. 8 studies (4%) had a noninferiority study setting. In total, 160 (77%) studies had a continuous primary outcome and were included in the analysis (Figure 2). Reporting of the key parameters of variability and uncertainty SDest as a part of the power calculation was reported in 93 (58%) studies (Figure 2). Observed SDs (SD1 and SD2) of the outcome in the study groups were reported in 89 (57%) studies (Table 1). The rate of reporting SD1 and SD2 was comparable whether or not power calculation consisted of SDest (Figure 2). Both estimated and observed SDs were reported in 52 (33%) studies. In addition, the observed SDs in the study groups were calculated from the observed SEs or CIs in 15 (14%) of the studies that also reported SDest (Figure 2). A quarter (26/93) of the studies did not report any variability parameter of primary outcome data when SDest was presented (Figure 2). The MDest of the primary outcome in the power calculation was found in 138 studies (86%). Of these, 31 (19%) reported CIMD and in a further 68 (49%) studies they could be calculated from the means and pooled SD, resulting in a total of 99 (72%) studies.

Correspondence of estimated and observed variability The pooled SD (SDpooled) for the primary outcome variable was calculated for 62 studies in which SDest was also available. The median value for the ratio of pooled observed SD to estimated SD (SDpooled/SDest) was 1.0 (IQR: 0.76–1.32). The geometric mean value of SDpooled/SDest was 1.01 (Table 2).

Overlap of the estimated difference and confidence interval of mean difference between groups In those studies that had CIMD available, 66 had reported a negative outcome (statistically not-significant finding). Of these, the MDest did not belong to the observed CIMD between groups for the primary outcome at the last or pre-specified follow-up time point in 38 (58%) studies. In other words, 42% of the negative studies could not exclude a clinically meaningful mean difference sized MDest between groups. Figure 3 illustrates the CIMD of these negative studies corresponding to the positive and negative values of MDest chosen in the power calculations (66 studies). Table 2. Correspondence of estimated and observed variability in the primary outcome, pooled, and in the two study groups, respectively Geometric Measure Median (IQR) mean (SD) Ratio of observed and estimated pooled SD Ratio of SD1/SDest Ratio of SD2/SDest

1.00 (0.76–1.32) 1.03 (0.73–1.43) 0.96 (0.74–1.20)

SD = standard deviation of primary outcome.

1.01 (1.62) 1.00 (1.76) 0.96 (1.74)

482

Acta Orthopaedica 2020; 91 (4): 479–484

Discussion The rationale for our study was to investigate the etiology of the suggested low power in orthopedic RCTs and the subsequent consequences, namely, are unreasonably small variability estimates to blame for low power and is the uncertainty of the primary outcome measured affected by small sample sizes? Assuming that average statistical power was low among orthopedic RCTs, we hypothesized that there would be poor correspondence in the estimated and observed variability of the primary outcome. In addition, the overlap of MDest and 95% CIMD was investigated in the primary outcome. Therefore, we addressed the correspondence of the estimated variability of the primary outcome in the study population to that actually observed. Also, we compared the MDest of the primary outcome with the observed CIMD for the between group difference among orthopedic RCTs recently published in 8 major Figure 3. Relative confidence intervals for 66 negative studies, i.e., not scientific journals. We found good correspondence between reporting a difference, which reported CIMD or in which CIMD was calcuthe estimated and observed SDs based on median values. It is lated. Blue interval limits highlight studies in which MDest in both directions could be excluded, whereas red indicates studies in which MDest matter of great concern that in almost half of the RCTs there could not be excluded, thus indicating inconclusive studies. were major deficits in the reporting of the main outcome variables and that a clinically relevant difference between groups could not be excluded based on the CI of the mean difference in primary outcome variable. Sample size and estimated effect size in studies with Orthopedic researchers have widely incorporated the power and without overlap in the estimated difference and calculation in current studies and the power calculation was confidence interval of mean difference performed in 160 studies using 1 continuous primary outThe median estimated effect size in the power calculation come. However, SDest was reported in only 58% of studies, was 0.75 (IQR 0.60–1.0) for those studies that presented both whereas MDest was reported in 86% of all studies reporting SDest and MDest in the power calculation (Table 3). The mean power calculation. The SDest and the MDest of the primary estimated effect size was greater (0.84 versus 0.79, MD = outcome are mandatory in power calculation and are of the 0.05, 95% CI –0.7 to 0.26) in negative studies in which the utmost importance in evaluating the reasonableness of the estimated MDest was not included in the CIMD. The median sample size. Moreover, a quarter (28%) of the studies that sample size in negative studies in which MD was included in presented SDest and MDest, in accordance with CONSORT the CIMD was 53 (IQR 43–62). In studies where MDest was guidelines for power calculation did not report the observed not included, the median sample size was 86 (IQR 63–115). SD, SE, or CI of the primary outcome in the study groups in The groups had a difference in ranks when sample sizes were numerical format. The uncertainty of the investigated effect compared (p = 0.01). of the intervention in the RCT can be addressed only by these measures of variability of the mean difference. These missing values of the reported power calculations are in line with the situaTable 3. Median effect sizes in power calculation in all studies and studies divided by tion a decade ago in the major clinical medithe belonging of the mean difference (MD) estimate to the observed confidence interval (CI) of difference in means cine journals (Charles et al. 2009). Finally, it was also a concern to find that only one-fifth of studies reported the confidence intervals Measure n/N (%) Median (IQR) Mean (SD) for mean difference value, which is in line Effect size in power calculation of all with the situation in orthopedics a decade studies that estimated SD and MD 93/160 (58) 0.75 (0.60–1.0) 0.79 (0.30) ago (Vavken et al. 2009). In half the studAmong “negative” studies, MD estimate ies the CIMD could be calculated, but readers did not belong in the CI and corresponding estimated effect size 38/66 (58) 0.67 (0.56–0.92) 0.79 (0.39) cannot be expected to perform such a calcuAmong “negative” studies, MD estimate lation. Similar issues in reporting of results did belong in the CI and corresponding in medical literature were noticed by Altman estimated effect size 28/66 (42) 0.83 (0.60–1.08) 0.84 (0.32) (1980b), who concluded that bad scientific

Acta Orthopaedica 2020; 91 (4): 479–484

experiments are unethical, from which poor statistical methods and reporting of results are not detached (Altman 1980a). Moreover, praise for reporting CIs instead of solely relying on p-values was expressed long ago by the International Committee of Medical Journal Editors (1988). Among the studies included in this review, there was good correspondence based on median ratio values between estimated SD and observed SD in the primary outcome, which is contrary to a non-orthopedic review in which the estimated SDs tended to be smaller compared with that actually observed in the study population (Vickers 2003). After all, the estimates of SD and MD in the power calculation are estimates of unknown parameters. Simulation studies show that only a small amount of knowledge of the SD in the population is collected after a total of 70 patients using continuous variables (Teare et al. 2014) and most orthopedic trials include at least this number of patients, and thus yield well-established SD estimates of the population to be used in power analyses. Inferences made solely on the p-value or nominal significance should be treated with skepticism (Altman 2005). Instead, the CIMD can be used to convey important information about plausible effect sizes, especially in the case of negative trials. In an optimal situation, the MDest in power calculation or another estimate of clinically relevant MD size of difference does not overlap with the CIMD of the primary outcome measure when interpreting negative trials. In almost half of negative studies, the MDest did belong to the observed CIMD. Thus, a clinically relevant difference or MCID was excludable with a 0.95 confidence in less than half of the negative studies. A universal misinterpretation of a finding without nominal statistical significance is to declare that there is no difference between groups, suggesting an equivalence. Failure to reject the null hypothesis does not indicate the groups are equal (Altman and Bland 1995) and it should be stated that superiority cannot be established, and results should be interpreted based on the CIMD. The CIMD shows which values can be rejected at the chosen error level. If the MCID is included in the CIMD, little can be interpreted from the study because the result is inconclusive. High estimated effect size yields a lower sample size and eventually a point estimate with high uncertainty, i.e., wide CIMD. This is the major problem in orthopedic science because wide CIs give very poor inference chances for our studies. It is of course important to remember that increasing sample size always has implications in ethical, pragmatic, and financial aspects. However, little is known about why sample sizes remain low in orthopedic studies. Based on our results, we postulate that since there was good correspondence with SDs, we assume that high effect size estimates are partly due to optimistically high estimates of MD in power calculations. However, it should be noted that greater sample size yields narrower CIMD, but, holding constant the alpha and the beta error levels in power calculation, the MDest would be smaller, and results would still often include MDest size of difference in the CIMD. The distribution of effect sizes in over 11,000

483

meta-analyses and their respective RCTs in the Cochrane database shows that almost all effect sizes are small or moderate in size (Lamberink et al. 2018). If the sample size were to be larger than currently seen, it would be able to exclude not only spuriously high estimates of MD, but more realistic ones. Due to these aforementioned issues of the targeted estimate of MD used in power calculations, it has been proposed to view MDest as a context-specific difference “one would not like to miss” (Senn 2014) or to use estimated width of CI for primary outcome instead of MDest as a basis for sample size calculation (Rothman and Greenland 2018). We acknowledge that this review assessed RCT articles published in 8 orthopedic journals, which may not be a representative sample of the whole orthopedic literature. Also, only RCTs allocated in 2 arms with 1 continuous primary outcome and reported power calculation were included. In addition, due to deficiencies in reported parameters of variability (SD) and uncertainty (CIMD), we were able to compare in only a limited number of studies the estimated and the observed values of SD and MD. Conclusion Power calculations were used in most of the RCTs, but most of the studies lacked some of the essential components required by the CONSORT statement and the results required to replicate the analysis. The key parameters of data variability were also poorly reported. Low power is likely to prevail in orthopedics, but we observed good correspondence between the estimated and the observed SD of the study data among recent orthopedic RCTs. Hence, we postulate that low power is not fully responsible for the unreasonably small variability estimates in primary outcome measures. In fewer than half of the studies, the estimated MD overlapped with the CIMD in primary outcome, indicating that the conclusions based on these studies are very limited. An increase in power and sample size would yield lower uncertainty of effect size and serve to mitigate this issue. Further studies are needed to investigate the interpretation of negative studies in orthopedics. Funding, data sharing, and potential conflicts of interests This study had no funding. The data assessment table (as an csv file) can be obtained from the authors. The authors declare that they have no competing interests. Supplementary data The statistical codes are are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/ 17453674.2020.1755932

Abdullah L, Davis D E, Fabricant P D, Baldwin K, Namdari S. Is there truly “no significant difference” underpowered randomized controlled trials in the orthopaedic literature. J Bone Joint Surg Am 2015; 97(24): 2068-73.

484

Altman D G. Statistics and ethics in medical research: misuse of statistics is unethical. BMJ 1980a; 281(6249): 1182-4. Altman D G. Statistics and ethics in medical research, VI: Presentation of results. BMJ 1980b; 281(6254): 1542-4. Altman D G. Why we need confidence intervals. World J Surg 2005; 29(5): 554-6. Altman D G, Bland J M. Absence of evidence is not evidence of absence. BMJ 1995; 311(7003): 485. Anderson A A. Assessing statistical results: magnitude, precision, and model uncertainty. Am Stat 2019; 73(Supl.1): 118-21. Angst F, Aeschlimann A, Angst J. The minimal clinically important difference raised the significance of outcome effects above the statistical level, with methodological implications for future studies. J Clin Epidemiol 2017; 82: 128-36. Button K S, Ioannidis J P A, Mokrysz C, Nosek B A, Flint J, Robinson E S J, Munafò M R. Power failure: why small sample size undermines the reliability of neuroscience. Nat Rev Neurosci 2013; 14(5): 365-76. Charles P, Giraudeau B, Dechartres A, Baron G, Ravaud P. Reporting of sample size calculation in randomised controlled trials: review. BMJ 2009; 338: b1732-b1732. Cook J A, Julious S A, Sones W, Hampson L V, Hewitt C, Berlin J A, Ashby D, Emsley R, Fergusson D A, Walters S J, Wilson E C F, Maclennan G, Stallard N, Rothwell J C, Bland M, Brown L, Ramsay C R, Cook A, Armstrong D, Altman D, Vale L D. DELTA2 guidance on choosing the target difference and undertaking and reporting the sample size calculation for a randomised controlled trial. Trials 2018; 19(1): 606. Dabija D I, Jain N B. Minimal clinically important difference of shoulder outcome measures and diagnoses. Am J Phys Med Rehabil 2019; 98(8) 671-6. de Vet H C W, Terwee C B. The minimal detectable change should not replace the minimal important difference. J Clin Epidemiol 2010; 63(7): 804-5. Gelman A, Greenland S. Are confidence intervals better termed “uncertainty intervals”. BMJ 2019; 366: l5381. Higgins J P T, Green S. Cochrane handbook for systematic reviews of interventions. Vol. 4. New York: Wiley; 2011. International Committee of Medical Journal Editors. I. Uniform requirements for manuscripts submitted to biomedical journals. Ann Intern Med; 1988; 108(2): 258.

Acta Orthopaedica 2020; 91 (4): 479–484

Jayadevappa R, Cook R, Chhatre S. Minimal important difference to infer changes in health-related quality of life-a systematic review. J Clin Epidemiol 2017; 89: 188-98. Lamberink H J, Otte W M, Sinke M R T, Lakens D, Glasziou P P, Tijdink J K, Vinkers C H. Statistical power of clinical trials increased while effect size remained stable: an empirical analysis of 136,212 clinical trials between 1975 and 2014. J Clin Epidemiol 2018; 102: 123-8. Ostelo R W J G, Deyo R A, Stratford P, Waddell G, Croft P, Von Korff M, Bouter L M, de Vet H C. Interpreting change scores for pain and functional status in low back pain: towards international consensus regarding minimal important change. Spine 2008; 33(1): 90-4. Reito A, Raittio L, Helminen O. Revisiting the sample size and statistical power of randomized controlled trials in orthopaedics after 2 decades. JBJS Rev 2020; 8(2): e0079. Rothman K J, Greenland S. Planning study size based on precision rather than power. Epidemiology 2018; 29(5): 599-603. Sabharwal S, Patel N, Holloway I, Athanasiou T. Sample size calculations in orthopaedics randomised controlled trials: revisiting research practices. Acta Orthop Belg 2015; 81(1): 115-22. Schulz K F, Altman D G, Moher D. CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials. BMC Med; 2010; 8(1): 18. Senn S. Delta Force: To what extent is clinical relevance relevant? 2014. https://errorstatistics.com/2014/03/17/stephen-senn-on-how-to-interpretdiscrepancies-against-which-a-test-has-high-power-guest-post/ (Accessed February 14, 2020 Szucs D, Ioannidis J P A. Empirical assessment of published effect sizes and power in the recent cognitive neuroscience and psychology literature. PLoS Biol 2017; 15(3): e2000797. Teare M D, Dimairo M, Shephard N, Hayman A, Whitehead A, Walters S J. Sample size requirements to estimate key design parameters from external pilot randomised controlled trials: a simulation study. Trials 2014; 15(1): 264. Vavken P, Heinrich K M, Koppelhuber C, Rois S, Dorotka R. The use of confidence intervals in reporting orthopaedic research findings. Clin Orthop Relat Res 2009; 467(12): 3334-9 Vickers A J. Underpowering in randomized trials reporting a sample size calculation. J Clin Epidemiol 2003; 56(8): 717-20.

Acta Orthopaedica 2020; 91 (4): 485–488

485

External iliac artery injury following total hip arthroplasty via the direct anterior approach—a case report Ellen BURLAGE 1, Jasper G GERBERS 1, Bob R H GEELKERKEN 2,3, and Wiebe C VERRA 1 1 Department of Orthopedic Surgery, Medisch Spectrum Twente, Enschede; 2 Multimodality Medical Imaging M3i Group, Faculty of Science and Technology, Technical Medical Centre, University of Twente, Enschede; 3 Department of Vascular Surgery, Medisch Spectrum Twente, Enschede, the Netherlands Correspondence: w.verra@mst.nl Submitted 2019-10-02. Accepted 2020-01-29.

An 82-year old man attended our outpatient clinic with symptoms of osteoarthritis of the right hip. Radiographic examination confirmed this diagnosis. The patient’s medical history included: atrial fibrillation, for which a coumarin derivate was started in 2009, idiopathic thrombocytopenia, with platelet counts of circa 60×109 (normal 150–400×109), and prostate carcinoma. After transurethral prostate resection in 2016, at the request of the patient, anticoagulant treatment was replaced by ASA (Ascal). Moreover, despite suffering from atrial fibrillation and a low ventricular rate, the patient had chosen not to receive a pacemaker. Ascal was discontinued 7 days before the planned total hip arthroplasty (THA). Therapy to increase platelet function was not administered preoperatively because of low risk of bleeding associated with platelet counts > 30×109 (Yang and Zhong 2000). Surgical management After a period of nonoperative treatment the patient was scheduled for uncemented THA via direct anterior approach (DAA) using spinal anesthesia (Siguier et al. 2004). After incision, the fascia of the tensor fascia lata was incised. The tensor fascia lata and gluteus medius muscles were retracted laterally with a Hohmann retractor (Figure 1), and the sartorius and rectus muscles were retracted medially with a blunt Hohmann retractor in order to expose the anterior hip capsule. The third, pointed, Hohmann retractor was placed under the rectus tendon just at the bony border of the acetabular rim (Figure 2). The retractor aimed in the direction of the contralateral kidney and was fixed using a device that statically holds both the second and third retractor. After opening the anterior capsule, the osteotomy of the femoral neck was performed. There was minimal blood loss, by suction, 50 mL. The labrum was excised and the acetabulum reamed. In the meantime, the patient had become hemodynamically unstable with hypotension, tachycardia and no response to vasopressors and intravenous fluids. In view of the poor circulatory situation, the patient was intubated and cardiopulmonary resuscitation was started.

Figure 1. Type of Hohmann retractor used during the procedure.

Figure 2. Hohmann retractors placed on a pelvis phantom.

The surgery was interrupted immediately. The wound quickly closed with only an implanted acetabular cup without the use of drainage systems. A transthoracic diac duplex ultrasound was performed in the operating

was and carthe-

486

Acta Orthopaedica 2020; 91 (4): 485â&#x20AC;&#x201C;488

retroperitoneal approach and a small laparotomy. The distal aorta was controlled, and thereafter the patient was stabilized. Downstream exploration was performed and a nearby circular defect of at least half of the circumference of the external iliac artery was found. The external iliac artery was re-anastomosed end-to-end with Prolene 5-0. The patient was hemodynamically stable and transferred to the intensive care unit. A pacemaker was implanted due to a high-grade atrioventricular block. After 5 days it was considered to be sufficiently safe to finish the THA by implanting the stem of the prothesis. The patient was transferred to the nursing ward within 10 days and discharged to the rehabilitation center on the 16th postoperative day without further complications.

Discussion

Figure 3. Segmentation of the contrast-enhanced vascular CT. Window ranges corresponding to intravascular contrast are colored red. The implanted metal backed cup is colored gray. The arrow is placed on the site of extravasation at the external iliac artery. Of note is the distance between the acetabular rim and the extravasation site.

ater. There was poor contraction of the anterior wall of the heart. With a high likelihood of acute coronary syndrome, the patient was immediately taken to the cardiac catheterization room. Percutaneous coronary intervention via the right femoral artery showed diffuse coronary artery disease but no significant occlusions. Simultaneously the hematology and biochemistry markers were assessed; the hemoglobin was 3.2 mmol/L (normal 8.5â&#x20AC;&#x201C;11.0) and the hematocrit was 0.15 (normal 0.40â&#x20AC;&#x201C;0.50). Troponin-T was only mildly elevated at 23 ng/L (normal < 14). Under the suspicion of persisting hemorrhagic shock a multiphase, multislice abdominal aorta-iliac and femoral contrast-enhanced CT scan (ceCTa) was performed. The ceCTa revealed a large retroperitoneal hematoma on the right, with extravasation of contrast in the arterial phase, likely from the distal external iliac artery. Beneath, an asymptomatic left common iliac artery aneurysm with a diameter of 3.5 cm was demonstrated (Figure 3). The patient was brought back to the operating theater for emergency vascular repair. Due to the uncertainty of the location of the bleeding and the severe hemodynamic instability of the patient an endovascular approach was considered to be less appropriate. The distal aorta and the right common, external, and internal iliac arteries were dissected by a right-sided

A hemodynamically unstable patient due to external iliac artery injury during a DAA THA has, as far as we know, not yet been described in the literature. 2 recent case reports described common femoral artery lesions, closely related to external iliac artery injury, following THA through DAA (Marongiu et al. 2019, Mortazavi et al. 2019). In recent years THA via the DAA has gained popularity (Siquier et al. 2004, Wang et al. 2018). This popularity of THA procedures being performed via an anterior approach makes awareness of the occurrence of this life-threatening vascular complication important. Severe vascular injury during total hip arthroplasty is a rare complication estimated at between 0.16% and 0.25% (Nachbur et al. 1979). More recent studies report an incidence of 0.04% in primary THA with an increase to 0.19% in revision arthroplasty (Abularrage et al. 2008). In general, regardless of the surgical approach, injuries have been reported in all the main vessels around the hip, the common femoral artery being the most reported damaged vessel and the external iliac artery thereafter (Shoenfeld et al. 1990, Lazarides et al. 1991). They are at risk because of their anatomical location (Bach et al. 2002, Kawasaki 2012). At the level of the anterior inferior iliac spine the external iliac vessels lie only 7 millimeters from the bone. In some cases they lie directly on the osseous surface as they leave the cavity of the pelvis (Rue et al. 2004, Kawasaki 2012). Rue et al. (2004) described 2 main groups of vascular injuries during THA surgery: direct and indirect injuries. Direct damage may occur by arterial transection due to a misplaced retractor or by excessive reaming and by arterial penetration of a screw during cup fixation. Longitudinal vascular laceration may cause intraoperative bleeding and a decline in blood pressure. Because the bleeding of puncture injuries results in a slow and small amount of bleeding, it is likely this will not be observed during surgery. The bleeding will form a false aneurysm presenting as a hematoma or pulsatile mass. The patient can complain of hip pain due to pressure or ischemic symptoms caused by impaired blood flow (Proschek et al. 2006).

Acta Orthopaedica 2020; 91 (4): 485â&#x20AC;&#x201C;488

Indirect damage can be caused by compression, stretching, or tearing of a vessel or by excessive heating by the bone cement. Secondary formation of a thrombus or presence of an intimal flap can lead to hypoperfusion and ischemia of the distal leg. Immediately after surgery, the dorsalis pedis arterial pulse can be absent. It is also possible that ischemic pain due to hypoperfusion and the absence of pulsations will not appear until a few hours postoperatively (Mortazavi et al. 2019). Vascular damage in THA via an anterior approach has been described in 2 recent case reports (Marongiu et al. 2019, Mortazavi et al. 2019). 2 patients had absent distal pulsations immediately after surgery. The 3rd patient had delayed presentation of leg pain and a deficit in the distal arterial pulses. All 3 patients had arterial intimal damage with thrombus formation. Marongiu et al. (2019) reported damage of the common femoral artery caused by misplacement of the anterior retractor. The patient became symptomatic with dropped hemoglobin levels, hematoma, and groin pain a few days postoperatively. In both case reports the femoral artery was involved rather than the external iliac artery as in our case. A systematic review, with 11,810 DAA procedures included, reported 920 complications (fractures, infection, nerve injury, wound complications, dislocation, and revision) (Lee and Marconi 2015). None of these complications had a vascular cause. Alshameeri et al. (2015) performed a systematic review of vascular injuries in association with THA. They identified 124 vascular injuries during the last 22 years, irrespective of the surgical approach. In none of the identified cases was a DAA used. THA via the lateral and posterolateral approach and their association with vascular injuries has been well described in the literature. Shoenfeld et al. (1990) identified 63 cases via the lateral approach and found the external iliac artery to have the highest injury rate with 36 injuries. Injuries of the external iliac artery consisted of 11 pseudoaneurysms and 17 thromboembolic complications. For the remaining 8 external iliac artery injuries the type of injury was not specified. The causes of the vascular injuries were cement related (one-third), misplacement of a retractor (one-third) or excessive traction on the vessel (one-tenth). Emergent vascular intervention at the time of the THA was necessary in 27 cases. In half of these cases the external iliac artery was involved. The causes of the external iliac artery injury needing emergency intervention were not specified. The importance in prevention of vascular complications in THA is clear. Alshameeri et al. (2015) reported permanent disability due to ischemia (8%), amputation (2%), and mortality (7%) in 124 vascular injuries as significant consequences of this severe complication. Injury due to retractor misplacement is often described as a cause of vascular damage. It is paramount that patients should be placed in the correct position and that they are handled considering the anatomical location and integrity of the surrounding vascular structures. The neurovascular bundle that contains the external iliac artery travels along the iliopsoas muscle. The iliopsoas muscle bulk will

487

protect the neurovascular bundle against injury (Sullivan et al. 2019). Therefore, the retractor tips should be placed directly on the bone, and the iliopsoas muscle should not be interposed between the retractor and bone (Rue et al. 2004). Speculation remains as to what caused the active bleeding in our case. The frequently mentioned causes of direct damage to the external iliac artery seem unlikely. Reaming of the acetabulum occurred without complications and thereafter there were no signs of significant blood loss. Neither drills nor screws were necessary to fix the acetabular component. Because the iliac vessels run along the iliacus muscles the retractor has to cross the iliopsoas muscle to cause direct damage to the external iliac artery (Kawasaki 2012, Sullivan et al. 2019). In our case we had no indication during surgery that the Hohmann retractors were located out of the correct anatomical location and we do not believe they were inserted too deep or medially. Considering the type of damage, a nearby circular defect of the artery and repair by end-to-end anastomosis could argue for being a direct cause. The force that esd applied to the external iliac artery could be an indirect cause for the bleeding. Excessive limb manipulation to enact joint dislocation, relocation, and traction by retractors can exert longitudinal stress to the iliac vessels. The external iliac artery is at less risk of tearing due to its thicker intima and flexibility. The presence of atherosclerotic plaques increases the risk of intimal dissection resulting in thrombosis and distal ischemia (Shoenfeld et al. 1990). In our case the hemorrhagic shock occurred as a result of bleeding from the external iliac artery. Therefore the characteristics of this case seem to be less suitable for the pathophysiology of indirect causes of vascular damage. In conclusion, vascular injury during THA is a rare complication. This complication during DAA has not been described in literature before. Surgeons should be mindful of the fact that injury to the external iliac artery can occur during THA via the DAA, by either direct or indirect means. Last of all it is important to remember that hemorrhagic shock in peracute hemodynamically unstable patients cannot be excluded if there are no signs of significant blood loss in the surgical field. Ethics and potential conflicts of interest Informed consent was given by the patient. The authors declare no conflicts of interest. Supplementary data A video of the contrast-enhanced vascular CT is available at: https://jepper.stackstorage.com/s/GI6wgbB6pP5NBj3

EB undertook the research and wrote the manuscript. JG, RG, and WV revised the manuscript. WV and RG performed the surgery. Acta thanks Harald Brismar and Ola Rolfson for help withÂ peer review of this study.

488

Abularrage C J, Weiswasser J M, Dezee K J, Slidell M B, Henderson W G, Sidawy A N. Predictors of lower extremity arterial injury after total knee or total hip arthroplasty. J Vasc Surg 2008; 47(4): 803-7. Alshameeri Z, Varty K, Khanduja V, Bajekal R. Iatrogenic vascular injuries during arthroplasty of the hip. Bone Joint J 2015; 97B(11): 1447-55. Bach C M, Steingruber I E, Ogon M, Maurer H, Nogler M, Wimmer C. Intrapelvic complications after total hip arthroplasty failure. Am J Surg 2002; 183(1): 75-9. Kawasaki Y. Location of intrapelvic vessels around the acetabulum assessed by three-dimensional computed tomographic angiography: prevention of vascular-related complications in total hip arthroplasty. J Orthop Sci 2012; 17(4): 397. Lazarides M K, Arvanitis D P, Dayantas J N. Iatrogenic arterial trauma associated with hip joint surgery: an overview. Eur J Vasc Surg 1991; 5(5): 549-56. Lee G C, Marconi D. Complications following direct anterior hip procedures: costs to both patients and surgeons. J Arthroplasty 2015; 30(9): 98-101. Marongiu G, Rigotti S, Campacci A, Zorzi C. Acute common femoral artery lesion after direct anterior approach for THA: a case report and literature review. J Orthop Sci. 2019; 24(2): 382-4. Mortazavi S M J, Kazemi M, Noaparast M. Femoral artery intimal injury following total hip arthroplasty through the direct anterior approach: a rare but potential complication. Arthroplasty Today 2019; 5(3): 288-91.

Acta Orthopaedica 2020; 91 (4): 485â&#x20AC;&#x201C;488

Nachbur B, Meyer R P, Verkkala K. The mechanisms of severe arterial injury in surgery of the hip joint. Clin Orthop Relat Res 1979; (141): 122-33. Proschek D, Proschek P, Hochmuth K, et al. False aneurysm of the left femoral artery and thrombosis of the left femoral vein after total hip arthroplasty. Arch Orthop Trauma Surg 2006; 126(7): 493. Rue J P, Inoue N, Mont M A. Current overview of neurovascular structures in hip arthroplasty: anatomy, preoperative evaluation, approaches, and operative techniques to avoid complications. Orthop 2004; 27: 73-81. Shoenfeld N A, Stuchin S A, Pearl R, Haveson S. The management of vascular injuries associated with total hip arthroplasty. J Vasc Surg 1990; 11(4): 549-55. Siguier T, Siguier M, Brumpt B. Mini-incision anterior approach does not increase dislocation rate. Clin Orthop Relat Res 2004; (426): 164-73. Sullivan C W, Banerjee S, Desai K, Smith M, Roberts J T. Safe zones for anterior acetabular retractor placement in direct anterior total hip arthroplasty: a cadaveric study. J Am Acad Orthop Surg 2019; 27(1): e969-e976. doi: 10.5435/JAAOS-D-18-00712. Wang Z, Hou J Z, Wu C H, Zhou Y J, Gu X M, Wang H H, et al. A systematic review and meta-analysis of direct anterior approach versus posterior approach in total hip arthroplasty. J Orthop Surg Res 2018; 13: 229. Yang R, Zhong C H. Pathogenesis and management of chronic idiopathic thrombocytopenic purpura: an update. Int J Hematol 2000; 71: 8-24.

Acta Orthopaedica 2020; 91 (4):489–496

489

New 3-dimensional implant application as an alternative to allograft in limb salvage surgery: a technical note on 10 cases Jong Woong PARK 1,2, Hyun Guy KANG 1,2, June Hyuk KIM 1, and Han-Soo KIM 3 1 Orthopaedic Oncology Clinic, National Cancer Center, Goyang; 2 Division of Convergence 3 Department of Orthopaedic Surgery, Seoul National University Hospital, Seoul, Korea

Technology, National Cancer Center, Goyang;

Correspondence: (HGK) ostumor@ncc.re.kr Submitted 2020-02-03. Accepted 2020-03-29.

Recently, there have been attempts to reconstruct bone defects using 3-dimensional (3D)-printed implants (Imanishi and Choong 2015, Wong et al. 2015, Liang et al. 2017, Wei et al. 2017, Park et al. 2018a, 2018b, Angelini et al. 2019). A 3D-printed, titanium alloy implant with an appropriate pore structure is biocompatible and personalizable in terms of the surgical location and extent (Wu et al. 2013, Guyer et al. 2016, Lee et al. 2016, Mumith et al. 2017, Li et al. 2018, McGilvray et al. 2018, Park et al. 2020). 3D-printed titanium alloy implants are often used alone to fill bone defects, despite the lack of clinical evidence to support their use. The following major limitations of the 3D-printed titanium alloy implant have been identified: (1) mechanical safety of the 3D implant is not guaranteed, especially with regard to fatigue strength; (2) the maximum printable size is limited by the metal 3D-printer, and is usually an approximate length of 20 cm; and (3) 3D-printing using composite materials is technically difficult; accordingly, a single titanium alloy is often used for implant fabrication. For example, the titanium alloy Ti6Al4V is an ideal material in terms of reducing the stress shield effect and improving biocompatibility, but has weak wear-resistance. Thus, it is difficult to fabricate an implant (including a joint) using a single titanium alloy material. To take advantage of 3D-printed titanium alloy implants and overcome the aforementioned disadvantages, the 3D-printed implant may be combined with conventional orthopedic surgical instruments, such as the intramedullary nail, artificial arthroplasty implant, and tumor prosthesis. In other words, 3D-printed titanium alloy implants provide biocompatibility and personalized size-matched filling for bone defects, while orthopedic instruments give mechanical strength and durable joint function. This technical note explores experiences with 3D-printed implants and a prosthesis composite (3DiPC) approach in various surgical contexts.

Patients 10 patients who underwent surgeries combining 3D-printed titanium implants and conventional orthopedic surgical instruments (i.e., the 3DiPC procedure; 7 pelvic and 3 long bones) were studied. 7 surgeries were performed for oncological reasons, including osteosarcoma (n = 2), Ewing sarcoma (n = 2), undifferentiated pleomorphic sarcoma of the bone (n = 2), and bone metastasis from renal cell carcinoma (n = 1). 3 surgeries were revision cases and included massive bone defects due to previous oncological surgeries (n = 2) and a car accident (n = 1). The 3DiPC procedures combined 3D-printed implants and conventional total hip arthroplasty (THA) implants (n = 6), a modular tumor prosthesis (n = 2), and intramedullary nails (n = 2) (Tables 1 and 2). All patients were followed postoperatively using the standard schedule and strategy for conventional limb salvage surgery. For all patients, except 1 patient who underwent surgery using a 3D-printed reinforcement cage, bone cutting was performed with a 3D-printed surgical guide. All patients requiring surgery for an oncological diagnosis had negative bone margins with preoperatively planned distances. The 3D-printed implants fit perfectly into the bone defects created by multiplanar bone cutting. Depending on the surgical location and scale, the mean bone tumor resection times were 88 (65–102) and 223 (174– 280) minutes for long bones and the pelvis, respectively. For patients who underwent pelvic reconstruction, independent gait without moderate-to-severe pain was achieved in 6 weeks. Notably, 2 patients who underwent chemotherapy after limb salvage surgery showed delayed rehabilitation. For 2 patients who underwent femoral reconstructive surgery, both cases utilized an intramedullary nail for stability and thus immediate weight-bearing activity was allowed. Finally, a patient who underwent humeral limb salvage surgery showed limited shoulder motion, but all functions below the elbow

490

Acta Orthopaedica 2020; 91 (4):489–496

Table 1. Patient demographics Patient no. Age/sex Location Side Diagnosis Tumor presentation 1 52/F Pelvis R UPS of the bone 2 47/M Pelvis L Chondrosarcoma 3 52/M Pelvis R Osteosarcoma 4 50/F Pelvis R Breast cancer 5 28/M Pelvis L Ewing sarcoma 6 47/F Pelvis R Chondrosarcoma 7 47/F Pelvis L Major trauma 8 54/F Femur R UPS of the bone 9 68/F Femur L Renal cell carcinoma 10 20/M Humerus R Osteosarcoma

Cause of surgery

Recurred Primary Primary NA (mechanical failure) Primary NA (mechanical failure) NA (mechanical failure) Primary Primary Primary

Follow-up Oncologic (months) status

Oncologic Oncologic Oncologic Nononcologic Oncologic Nononcologic Nononcologic Oncologic Oncologic Oncologic

21 DOD 33 NED 24 AWD 21 NA 12 DOD 10 NA 10 NA 9 NED 7 NED 16 NED

Abbreviations: AWD, alive with disease; DOD, dead of disease; NA, not applicable; NED, no evidence of disease; UPS, undifferentiated pleomorphic sarcoma.

Table 2. Details of limb salvage surgery Surgery time (min) Patient Recon- no. Reason for 3DiP Resection struction

3D-printed implant Combined conventional implant (X×Y×Z mm3) Weight (g)

1 No commercial implant 224 220 108×82×126 573 2 No commercial implant 280 81 105×135×178 649 3 No commercial implant 214 113 117×88×170 352 4 No commercial implant NA 30 47×52×55 21 5 No commercial implant 174 102 106×77×185 452 6 No commercial implant NA 92 124×112×168 472 7 No commercial implant NA 104 98×98×129 192 8 Saving adjacent joint 102 104 52×40×158 350 9 Saving adjacent joint 98 81 35×34×165 25 10 Saving adjacent joint 65 108 35×17×200 207

Modular tumor prosthesis: proximal femur and acetabular cup Conventional THA Conventional THA Conventional THA Conventional THA Conventional THA Conventional THA Retrograde intramedullary nail Intramedullary nail Modular tumor prosthesis: proximal humerus

Abbreviations: 3DiPC, 3D-printed implant and prosthesis composite; NA, not applicable; THA, total hip arthroplasty.

were preserved. There were no complications related to the 3DiPC surgery, including infection and mechanical failure in short- and mid-term follow-up (range, 7–33 months).

Surgical technique (Table 2) Implant design and fabrication The design process was coordinated for customized 3D-printed implants and surgical bone-cutting guides through close communication between orthopedic oncologists and engineers. Computerized tomography (CT) and magnetic resonance imaging (MRI) scans with a thin section thickness of 1–2 mm were used in the design process. All medical images were stored in the Digital Imaging and Communications in Medicine format. A graphical 3D model was created, and a mirror technique and virtual resection were performed using MIMICS (Interactive Medical Image Control System; Materialise; Leuven, Belgium). The implants had both lattice and solid structures in order to enhance bone ingrowth and to sup-

port mechanical strength. The dode-thin mesh structure was applied as a lattice structure using Magics 22 (Materialise; Leuven, Belgium). After 3D-printed implant fabrication, a polishing process was completed to prevent abrasion or adhesion to a major neurovascular bundle. The cutting guide was fabricated using a PolyJet-type 3D printer (OBJET30 Prime, Stratasys, Eden Prairie, MN, USA) with MED610 (Stratasys, Eden Prairie, MN, USA), a biocompatible resin. MED610 is a rigid, almost colorless material and was approved as a United States Pharmacopeia Class VI plastic because of its cytotoxicity, genotoxicity, and delayed hypersensitivity. For the bone-cutting guide, the maximum build size was 294×192×148.6 mm3 with an accuracy of 0.1 mm and a minimum layer thickness of 16 microns. For implant fabrication, 3D printing was performed with medicalgrade titanium (Ti6Al4V-ELI Per ASTM 136) using a powder-based electron beam melting (EBM) 3D printer (ARCAM A1, Arcam AB, Mölndal, Sweden). For the titanium metal implant, the maximum build size was 200×200×180 mm3 with an accuracy of 0.2 mm. The MEDYSSEY Company (Jecheon,

Acta Orthopaedica 2020; 91 (4):489â&#x20AC;&#x201C;496

491

Figure 1. Pelvic reconstruction. Images of patient #5. (A) A preoperative gadolinium enhanced T1-weighted MR image showing Ewing sarcoma arising from the left acetabulum. (B) Graphical designs of the 3D-printed bone tumor resection guide (upper row) and implant (lower row). (C) Resected bone tumor as planned. Photographs of the (D) 3D-printed implant and THA cup before (left) and after conjugation (right). (E) Intraoperative photograph, (F) postoperative CT reconstruction image, and (G) plain radiograph showing pelvic reconstruction.

Korea) fabricated the implant and surgical guide, and certified a custom-made implant from the Ministry of Food and Drug Safety. The 3D-printed implant may be combined with conventional orthopedic surgical instruments, such as the intramedullary nail, artificial arthroplasty implant, and tumor prosthesis. Bone cement was used to assemble the 3D-printed implant and a conventional prosthesis to create a 3DiPC. This cementation procedure was identical to the protocol used for an allograft-prosthesis composite (APC) without expecting bone ingrowth between a structural allograft, or a recycled autograft and prosthesis. The interface between the host bone and the 3D-printed implant was saved from cementation so as not to

interfere with bone ingrowth into the implant. Matched screw holes in the 3D-printed implant and the conventional prosthesis (e.g., a cup for total hip arthroplasty) helped to ensure stability. Pelvis For pelvic reconstruction, conventional THA was performed using a 3D-printed pelvic implant generated via a metal-onmetal cementation technique. In other words, the 3D-printed metal implant provided an acetabular socket with a proper inclination and anteversion, and the 3D-printed implant and the THA cup were assembled using bone cement and screws through matched holes in both implants. The 3D-printed

492

Acta Orthopaedica 2020; 91 (4):489â&#x20AC;&#x201C;496

Figure 2. Femoral reconstruction. Images of patient #8. (A) Preoperative plain radiographs and (B) a T2-weighted MR image showing undifferentiated pleomorphic sarcoma of the bone arising from the distal femoral shaft. (C) Resected bone tumor as planned. (D) Graphical designs of the 3D-printed implant and (E) photograph of the 3D-printed implant and intramedullary nail to be used. (F) Intraoperative photograph showing cement injection through a premade hole. (G) Intraoperative photograph, (H) postoperative CT reconstruction image, and (I) teleradiogram showing femoral reconstruction.

pelvic implants were individually fabricated to fit each patient and varied from a simple reinforcement cage to a megaprosthesis. The 3D-printed implant was mainly fixed by screws on the plates, which were fabricated integrally with the implant. The implant surfaces in contact with the host bone had a uniform lattice structure to enhance bone ingrowth. In most cases, bone-cutting guides were utilized to achieve a safe bone margin from the tumor and to fit the implant to the bone defect. The mean surgery time for pelvic reconstruction with the hip joint was 106 (30â&#x20AC;&#x201C;220) minutes (Table 2; Figure 1). Femur diaphysis For femoral reconstruction, an intramedullary nail was utilized to provide mechanical strength. The main role of the 3D-printed implant was bone-to-implant integration. To penetrate the 3D implant by an intramedullary nail, the implant had a tunnel inside, mimicking the bone marrow space. The tunnel for the intramedullary nail needed to match the nail contour reflected the bowing of the femur and had a diameter slightly

larger than that of the nail. For the 1st patient to undergo femoral surgery (patient #8), the implant had a solid core structure coated with a lattice structure. The implant had a cuff circumference with a 1-cm depth for host bone insertion and a short plate with screw holes for additional rotational stability between the implant and host bone. To fill the gap between the 3D implant and the retrograde femoral nail, small side holes were made for bone cement injection. Although the retrograde femoral nail had a relatively straight shape, the reconstruction of a segmental defect with a single block of the 3D implant and a penetrating intramedullary nail was difficult (Figure 2). Based on this experience, for the 2nd patient with femoral metastasis (patient #9), the 3D-printed implants to be used with intramedullary nails were generated with a full lattice (rather than solid) structured body. This implant reconstructed the cortical bone only and provided scaffolding for bone ingrowth, and mechanical stability was achieved by an intramedullary nail and bone cement. The cortical implant was made in 2 pieces and wrapped around the host bone junction in a telescopic manner. This approach to implant utilization

Acta Orthopaedica 2020; 91 (4):489â&#x20AC;&#x201C;496

493

Figure 3. Femoral reconstruction. Images of patient #9. (A) Preoperative plain radiographs and a T2-weighted MR image showing metastatic renal cell carcinoma arising from the femoral shaft. (B) Graphical designs and (C) photograph of the 3D-printed implant. (D) Resected bone tumor as planned. (E) Intraoperative photograph and (F) postoperative plain radiograph showing femoral reconstruction. (G) Follow-up plain radiograph at 3 months postoperatively showing callus formation at both proximal and distal junctions.

enabled a simple intraoperative change in the cutting length according to the bone margin status within a few centimeters (Figure 3). Humerus Patient #10 had osteosarcoma in most of the humerus, except for a short segment above the elbow joint, and underwent limb salvage surgery using 3DiPC. This approach comprised a conventional proximal tumor prosthesis and a distal 3D-printed implant to preserve the elbow joint. Since the humerus is a non-weight-bearing bone and allows for shortening if necessary, the use of a 3D implant in combination with a conventional modular tumor prosthesis was a better surgical option than an intramedullary nail (Figure 4).

Discussion The 3D-printed titanium alloy implant is a personalized and biocompatible surgical option for massive bone defect reconstruction. However, this new type of implant still raises some concerns that need to be addressed. 1st, the long-term mechan-

ical strength of an implant fabricated by 3D printing is not guaranteed. 2nd, a material containing a single titanium alloy theoretically would not have durable wear resistance for joint reconstruction, and 3D-printing using 2 or more materials is technically difficult at this time. 3rd, there is a printing size limitation (~20 cm length). Therefore, it is necessary to use conventional internal fixation devices, arthroplasty implants, or modular tumor prostheses with plenty of clinical experience to supplement 3D-printed implants during limb salvage surgery. Historically, APC has been used to reconstruct large bone defects (Mankin et al. 1996, Wunder et al. 2001, Jeon et al. 2007, 2014). Although bone stock restoration is one of the most important advantages to using APC, the effect may not be significant (Wilke et al. 2019). The 3D-printed titanium alloy implants with appropriate internal pores not only have a bone conduction effect, but also avoid problems related to osteolysis; thus, these implants may be mechanically stronger and longer lasting than structural allografts. Therefore, 3D-printed implants with adequate internal pore structures could be considerable alternatives to allografts using the APC technique.

494

Acta Orthopaedica 2020; 91 (4):489â&#x20AC;&#x201C;496

limb salvage surgery for the distal femur, the 3D-printed implant consisted of a fullmesh body type with minimal mechanical strength and weighed only 25 g (Table 2). A lattice-structured body is a unique advantage of a product fabricated by 3D-printing technology. Although a latticestructured body is mechanically weaker than a solid body, it provides a scaffold to enhance bone conduction and prevents the stress-shielding effect on the host bone. In addition, this structure also reduces metal-induced artifacts, thus enabling the magnetic resonance (MR) surveillance of postoperative tumor recurrences. Titanium is known to cause fewer artifacts on MR images relative to other metals due to its lower susceptibility (Hargreaves et al. 2011, Dillenseger et al. 2016). In previous literature, the metal artifacts caused by 3D-printed titanium alloy implants (Ti6Al4V) were not severe. To our knowledge, no previous studies have shown that the lattice structure reduces metal-induced artifacts in MR images. However, the reducing effect of the lattice structure has been observed. Specifically, a solid body coated with a few-millimeters-thick lattice structure yielded better quality postoperative MRI images of tissues around the implant than a pure solid structure without a lattice structure (Figure 5). This technical note has some limitations. The small number of patients and singleinstitution design may limit the generalizability of the study results. The short followFigure 4. Humeral reconstruction. Images of patient #10. (A) A preoperative plain radiograph and (B) a T2-weighted MR image showing osteosarcoma in most of the humerus, except the up period made it difficult to ascertain the elbow joint. (C) Graphical designs of the 3D-printed implant. (D) Resected bone tumor as local recurrence rate after wide excision and planned. (E) Photograph of the 3D-printed implant and tumor prosthesis to be used. Intraop- implant longevity. Long-term follow-ups erative photograph showing (F) proximal and (G) distal parts of the reconstruction. (H) A postinvolving clinical and biomechanical data operative plain radiograph and (I) CT reconstruction image showing humeral reconstruction. subjected to dynamic finite element analyses and experimentation are needed to clarify the mechanical propThe weight of 3D-printed implants varies depending on the erties of 3D-printed implants. One major advantage of using bone defect size and mechanical strength requirements. In 3DiPC rather than APC may be reduced surgical time due to the pelvic area, a large bone defect including the acetabulum omitting allograft carving and easy fixation by the preoperarequired a large-sized implant with great mechanical strength; tively fabricated fixation part of the implant. However, proper thus, the mean weight was 500 (352â&#x20AC;&#x201C;649) g. However, in comparison while controlling for confounding factors related to the pelvic area, even a relatively small-sized reinforcement surgery time has not been done. cage-type implant weighed between 21 and 192 g. In the case In conclusion, 3D-printed implants provide another surrequiring 3DiPC with an intramedullary nail, the mechanical gical option involving the 3DiPC approach. This approach strength was reinforced by an intramedullary nail before bone could resolve some concerns regarding the use of a new type incorporation into the 3D-printed implant, and the 3D-printed of 3D-printed implant, such as possible mechanical weakness, implant mainly acted as a scaffold for the fusion of the bone lack of fatigue strength, weak wear resistance, and limitation and soft tissue. Therefore, for the last patient who underwent of the maximal printable size.

Acta Orthopaedica 2020; 91 (4):489–496

495

Figure 5. Metal artifacts in postoperative MR images. Metal artifacts around the titanium alloy (Ti6Al4V) megaprosthesis were not severe on postoperative MRI, and T2-weighted MR images (A, B) were clearer than enhanced T1-weighted MR images (C, D). Axial MR images were presented at the proximal solid cuff (B-1, D-1), proximal shaft with full lattice coating (B-2, D-2), distal shaft with anteromedial half-lattice coating (B-3, D-3), and distal solid cuff (B-4, D-4). (E) A photograph showing reference lines for axial images.

Ethics, funding, and potential conflicts of interest The study was conducted after approval was obtained from the Institutional Review Board (NCC2017-0129, 2017-06-10). All patients provided written informed consent prior to undergoing surgery after a thorough explanation of the surgical options. This research was supported by the Industrial Strategic Technology Development Program (‘P0008805’), funded by the Ministry of Trade, Industry & Energy (MOTIE, Korea). All authors declare that there are no conflicts of interest.

Acta thanks Harald Brismar, Per Kjærsgaard-Andersen, and Pietro Ruggieri for help with peer reviewof this study.

Angelini A, Trovarelli G, Berizzi A, Pala E, Breda A, Ruggieri P. Three-dimensionprinted custom-made prosthetic reconstructions: from revision surgery to oncologic reconstructions. Int Orthop 2019; 43(1): 123-32. Dillenseger J P, Moliere S, Choquet P, Goetz C, Ehlinger M, Bierry G. An illustrative review to understand and manage metalinduced artifacts in musculoskeletal MRI: a primer and updates. Skeletal Radiol 2016; 45(5): 677-88. Guyer R D, Abitbol J J, Ohnmeiss D D, Yao C. Evaluating osseointegration into a deeply porous titanium scaffold: a biomechanical comparison with PEEK and allograft. Spine 2016; 41(19): E1146-50. Hargreaves B A, Worters P W, Pauly K B, Pauly J M, Koch K M, Gold G E. Metalinduced artifacts in MRI. AJR Am J Roentgenol 2011; 197(3): 547-55. Imanishi J, Choong P F. Three-dimensional printed calcaneal prosthesis following total calcanectomy. Int J Surg Case Rep 2015; 10:83-7. Jeon D G, Kim M S, Cho W H, Song W S, Lee S Y. Pasteurized autograft for intercalary reconstruction: an alternative to allograft. Clin Orthop Relat Res 2007; 456: 203-10. Jeon D G, Song W S, Cho W H, Kong C B, Cho S H, Choi S W, Lee S Y. Overlapping allograft for primary or salvage bone tumor reconstruction. J Surg Oncol 2014; 110(4): 366-71. Lee Y H, Chung C J, Wang C W, Peng Y T, Chang C H, Chen C H, Chen Y N, Li C T. Computational comparison of three posterior lumbar interbody fusion techniques by using porous titanium interbody cages with 50% porosity. Comput Biol Med 2016; 71: 35-45. Li S, Li X, Hou W, Nune K C, Misra R D K, Correa-Rodriguez V L, Guo Z, Hao Y, Yang R, Murr L E. Fabrication of open-cellular (porous) titanium alloy implants: osseointegration, vascularization and preliminary human trials. Sci China Mater 2018; 61(4): 525-36.

Liang H, Ji T, Zhang Y, Wang Y, Guo W. Reconstruction with 3D-printed pelvic endoprostheses after resection of a pelvic tumour. Bone Joint J 2017; 99-B(2): 267-75. Mankin H J, Gebhardt M C, Jennings L C, Springfield D S, Tomford W W. Long-term results of allograft replacement in the management of bone tumors. Clin Orthop Relat Res 1996; (324): 86-97. McGilvray K C, Easley J, Seim H B, Regan D, Berven S H, Hsu W K, Mroz T E, Puttlitz C M. Bony ingrowth potential of 3D-printed porous titanium alloy: a direct comparison of interbody cage materials in an in vivo ovine lumbar fusion model. Spine J 2018; 18(7): 1250-60. Mumith A, Coathup M, Chimutengwende-Gordon M, Aston W, Briggs T, Blunn G. Augmenting the osseointegration of endoprostheses using lasersintered porous collars: an in vivo study. Bone Joint J 2017; 99-B(2): 27682. Park J W, Kang H G, Lim K M, Kim J H, Kim H S. Three-dimensionally printed personalized implant design and reconstructive surgery for a bone tumor of the calcaneus: a case report. JBJS Case Connect 2018a; 8(2): e25.

496

Park J W, Kang H G, Lim K M, Park D W, Kim J H, Kim H S. Bone tumor resection guide using three-dimensional printing for limb salvage surgery. J Surg Oncol 2018b; 118(6): 898-905. Park J W, Song C A, Kang H G, Kim J H, Lim K M, Kim H S. Integration of a three-dimensional-printed titanium implant in human tissues: case study. Appl Sci 2020; 10(2): 553. Wei R, Guo W, Ji T, Zhang Y, Liang H. One-step reconstruction with a 3D-printed, custom-made prosthesis after total en bloc sacrectomy: a technical note. Eur Spine J 2017; 26(7): 1902-9. Wilke B K, Houdek M T, Rose P S, Sim F H. Proximal femoral allograftprosthetic composites: do they really restore bone? A retrospective review of

Acta Orthopaedica 2020; 91 (4):489â&#x20AC;&#x201C;496

revision allograft-prosthetic composites. J Arthroplasty 2019; 34(2): 346-51. Wong K C, Kumta S M, Geel N V, Demol J. One-step reconstruction with a 3D-printed, biomechanically evaluated custom implant after complex pelvic tumor resection. Comput Aided Surg 2015; 20(1): 14-23. Wu S H, Li Y, Zhang Y Q, Li X K, Yuan C F, Hao Y L, Zhang Z Y, Guo Z. Porous titanium-6 aluminum-4 vanadium cage has better osseointegration and less micromotion than a poly-ether-ether-ketone cage in sheep vertebral fusion. Artif Organs 2013; 37(12): E191-201. Wunder J S, Leitch K, Griffin A M, Davis A M, Bell R S. Comparison of two methods of reconstruction for primary malignant tumors at the knee: a sequential cohort study. J Surg Oncol 2001; 77(2): 89-99.

Acta Orthopaedica 2020; 91 (4): 497–499

497

Correspondence

RSA of the Symax hip stem Sir,—We read with a great interest the recent article by Kruijntjens et al. (2020). The investigators performed a 2-year model-based radiostereometric analysis (RSA) of the uncemented Symax femoral stem. The article reported no previous RSA studies on the Symax stem but we have executed a randomized double-blind, placebo-controlled trial (RCT) on the primary stability of the Symax stem in 49 postmenopausal women (Aro et al. 2018). The trial included an extended RSA follow-up for 3–5 years and the follow-up of implant survival for 8–10 years. We want to highlight some methodological differences between the 2 studies which make comparisons of interest. The results of the 2 studies complement each other. The stem design lead to early stabilization (within 4–12 weeks) in both studies. In our RCT, the stem migration did not respond to antiresorptive therapy. All stems, independent of the amount of initial migration, osseointegrated radiographically. No revision arthroplasty was performed. Due to the low rate of clinical failure (< 2% at 9 years), no meaningful analysis of an association between early stem migration and implant survival could be carried out. In this respect, the Symax stem resembled the outcome of 7 uncemented femoral stems recently analyzed in a meta-analysis (van der Voort et al. 2015). Based on the literature, Kruijntjens et al. concluded that there is a substantial variability in the amount of initial subsidence between stem designs. However, the comparison of different studies is challenging because any variation of stem migration may reflect more the heterogeneity of the skeletal status of study populations than the characteristics of tested femoral stem designs. The study of Kruijntjens et al. included both sexes with a mean age of 59 (30–70) years. Osteoporosis was an exclusion criterion but the measurement of local and systemic BMDs was not reported. The mean subsidence of the stem was minimal (y-translation –1.0 mm, 95% confidence interval [CI] –3.4 to 1.4). The mean stem rotation (retroversion) was 2.4° (CI –2.2 to 7.0). In our RCT, only subjects with normal BMD had minimal stem subsidence (0.7 mm, CI 0.2– 1.2) and rotation into retroversion (0.8°, CI 0.3–1.4). On the contrary, osteopenic and osteoporotic subjects exhibited more stem subsidence and rotation during the first 12 weeks after surgery. The primary stability of uncemented femoral stems is sensitive to adequate bone stock (Nazari-Farsani et al. 2020). It is reasonable that all RSA arthroplasty studies have a pre-

operative evaluation of local and systemic BMD, if a study protocol accepts recruitment of subjects (like postmenopausal women) at a known risk of low systemic BMD. Our RCT was performed in collaboration with the implant manufacturer, facilitating the standard marker-based RSA. Kruijntjens et al. applied model-based RSA (Kaptein et al. 2006), with experts of this method as co-investigators. Indeed, model-based RSA is highly tempting for clinical trials. The results of marker-based and model-based RSA show high agreement (Nazari-Farsani et al. 2016). Looking at the modelbased RSA data of Kruijntjens et al. there was a considerable variation (CI –1.2° to 1.8°) in double examinations of y-axis rotation. The stem rotation to retroversion also had variation (CI –2.2 to 7.0). It would be great to get a comment of the investigators. Was the variation due to a actual inter-individual difference of stem rotation or only due to inherent challenges of model-based RSA in measurement of stem rotation? Finally, Kruijntjens et al. performed the baseline RSA prior to loading of the operated hip, during the first day after surgery. They suggested a similar approach for all RSA studies. The suggestion was made without performing a comparison of different imaging and rehabilitation protocols. The current recommendation (ISO 16087:2013) is to schedule baseline RSA measurements within 5 days postoperatively, preferably before weight-bearing. 2 published studies have performed the baseline RSA imaging when the patients still were anesthetized. Interestingly, these studies showed no migration of uncemented femoral stems (Ström et al. 2007) and acetabular cups (Wolf et al. 2010) during the first week after surgery. The RCT of Ström et al. even compared the effect of different weight-bearing regimen on stem migration. The degree of early weight-bearing (unrestricted versus partial weightbearing) did not change the migration pattern. The initial stem migration does not seem to start with the first steps of postoperative weight-bearing but progressively only after 1 week. Thus, the current recommendation for timing the baseline RSA may be still appropriate. Hannu T Aro and Sanaz Nazari-Farsani Department of Orthopaedic Surgery and Traumatology, University of Turku and Turku University Hospital, Turku, Finland Email: hannu.aro@utu.fi

498

Sir,—We would like to comment on the remarks of Aro and Sanaz Nazari-Farsani in their recent letter to the editor concerning our recent article in the Acta Orthopaedica ‘Early stabilization of the uncemented Symax hip stem in a 2-year RSA study’ (Kruijntjens et al. 2020). We apologize to have missed the study of Aro et al. (2018) and not having mentioned it in our references. The explanation is simple, this study is hard to find when searching on RSA and Symax; there were no search results on PubMed, neither in Clinical Trials, simply because the brand name of the hip was not mentioned in the title nor in the abstract of the paper by Aro et al. Surprisingly however, relevant publications from our group regarding the Symax stem were not referred to in the article of Aro et al. (Kruijntjens et al. 2018; ten Broeke et al. 2012). As the excellent paper by Aro et al. should have been discussed in our paper, we appreciate the opportunity now given to have this discussion. Aro and Sanaz Nazari-Farsani state in their comment that the results of the 2 studies complement each other. In his letter he mentions that the stem design lead to early stabilization (within 4–12 weeks) in both studies. In our view there is however an important difference between our studies. We do not agree that Aro et al. showed stabilization within 4 weeks, being impossible when the first follow-up is at 3 months. The point of our article is that by performing early RSA (at day 1 postoperatively, and 1, 3, 6, 12, and 24 months postoperatively), we could detect a much earlier stabilization of the stem at 4 weeks. Our findings correspond better to the histomorphometric results seen in our earlier study, showing very early osseointegration which we attribute to the combination of the fit and fill characteristics of the stem geometry in combination with the highly bioactive properties of the Bonit-HA coating (ten Broeke et al. 2011). From Aro’s study it can only be concluded that stabilization was reached at 3 months. Nevertheless, both the study by Aro and our study do not report early implant failures, despite different levels of migration before stabilization, implying that there is obviously a safe range for migration of this stem before it becomes at risk for early loosening. This was also confirmed by our international Symax study (Kruijntjens et al. 2018) as well as in the Danish register study by Edwards et al. (2018). All aforementioned studies show that early osseointegration and (good) survival of the stem is not negatively affected by osteopenic / osteoporotic bone conditions. In contrast with the statement of Aro and Sanaz NazariFarsani we did not exclude patients with osteoporosis, but patients taking medication that may influence bone metabolism, in order not to introduce a potentially confounding factor. We are aware, from earlier literature, that antiresorptive medication as well as vitamin D with calcium influence bone mineral density (Venesmaa et al. 2001, Sköldenberg

Acta Orthopaedica 2020; 91 (4): 497–499

et al. 2011), but have probably no clinically relevant influence on migration of uncemented hip stems (Aro et al. 2018), let alone definitive stabilization. Therefore, differentiating between groups with or without osteoporosis would not contribute in answering our research question, which focused on the potential of this stem, with its particular geometry and coating characteristics, for early stabilization across different patient groups. Already from an earlier DEXA-study from our group on bone remodelling around the Symax stem (ten Broeke et al. 2012), it was clear that at 1-year follow-up all stems showed radiological evidence of stable bone ingrowth, independent of BMD at t0. The conclusion of the article of Aro et al. also was that stem migration was not influenced by the use of zoledronic acid, although zoledronic acid treated patients maintained periprosthetic BMD better than the control group. This conclusion was also drawn in a study by Aro et al. using denosumab (Aro et al. 2019). In other words, one may question if there is a reason anyhow to differentiate between initial BMD for choosing a particular hip implant. A further remark by Aro and Sanaz Nazari-Farsani was on the wide range of the stem (Y-axis) rotations both in the double examinations and in the follow-up measurements. It is well known and accepted that model-based RSA is less precise compared to marker-based RSA. Still model based RSA has other advantages and is suitable in clinical studies as was also demonstrated in a study from Aro’s group (Nazari-Farsani et al. 2016). Prins et al. (2008) demonstrated that the precision of the Elementary Geometrical Shape (EGS) model-based approach (used in our study) was found to be acceptable for use in a clinical study. It is important to realize that by increasing the number of patients, the lower precision of model-based RSA can be mended, as the accuracy (the bias) is as low as for marker-based RSA. During the settling of the stem in the initial month postoperatively, the stem also rotates into retroversion. The difference in CI between the double examinations and clinical data show that the range in rotation in our study, is mainly the result of the variation in actual rotation between patients, in combination with the variation introduced by the slightly less precision caused by using a model-based RSA approach. But most important, as stated in our paper, the initial level of migration of the uncemented stem is less relevant compared to the stabilisation of the stem in the period after initial settling to predict long term survival of the stem. Finally, we completely agree that the current recommendation for timing the baseline RSA is still appropriate, as our recommendation is exactly the same as stated in the ISO standard (ISO 16087:2013) on RSA. Especially for uncemented hip stems, we do recommend to add an extra early follow-up moment at 1 month in order to have more exact data on the early post-operative migration patterns.

Acta Orthopaedica 2020; 91 (4): 497–499

Dennis S M G Kruijntjens, Lennard Koster, Bart L Kaptein, Jacobus J C Arts, and René H M ten Broeke Departments of Orthopaedic Surgery, Research School Caphri, Maastricht University Medical Centre, Maastricht; Department of Orthopaedic Surgery, RSAcore, Leiden University Medical Centre, Leiden, the Netherlands Email: d.kruijntjens@gmail.com

Aro E, Moritz N, Mattila K, Aro H T. A long-lasting bisphosphonate partially protects periprosthetic bone, but does not enhance initial stability of uncemented femoral stems: A randomized placebo-controlled trial of women undergoing total hip arthroplasty. J Biomech 2018; 75: 35-45. Aro H T, Nazari-Farsani S, Vuopio M, Löyttyniemi E, Mattila K. Effect of Denusomab on femoral periprosthetic BMD and early femoral stem subsidence in postmenopausal women undergoing cementless total hip arthroplasty. JBMR Plus 2019; 3: e10217 Kaptein B L, Valstar E R, Spoor CW , Stoel B C, Rozing P M. Model-based RSA of a femoral hip stem using surface and geometrical shape models. Clin Orthop Relat Res 2006; 448: 92-7. Edwards N M, Varnum C, Kjærsgaard-Andersen P. Up to 10-year follow-up of the Symax stem in THA: a Danish single-centre study. Hip Int 2018; 28: 375-81. Kruijntjens D S, Kjaersgaard-Andersen P, Revald P, Leonhardt J S, Arts J J, ten Broeke R H. 5-year clinical and radiographic follow-up of the uncemented Symax hip stem in an international study. J Orthop Surg Res 2018; 13: 191. Kruijntjens D S M G, Koster L, Kaptein B L, Jutten L M C, Arts J J, Ten Broeke R H M. Early stabilization of the uncemented Symax hip stem in a 2-year RSA study. Acta Orthop 2020; 91(2): 159-64. Nazari-Farsani S, Finnilä S, Moritz N, Mattila K, Alm J J, Aro H T. Is modelbased radiostereometric analysis suitable for clinical trials of a cement-

499

less tapered wedge femoral stem? Clin Orthop Relat Res 2016; 474(10): 2246-53. Nazari-Farsani S, Vuopio M E, Aro H T. Bone mineral density and corticalbone thickness of the distal radius predict femoral stem subsidence in postmenopausal women. J Arthroplasty 2020; (in press). Prins A H, Kaptein B L, Stoel B C, Nelissen R G, Reiber J H, Valstar E R. Handling modular hip implants in model-based RSA: Combined stemhead models. J Biomech 2008; 41: 2912-7. Sköldenberg O G, Salemyr M O, Bodén H S, Ahl T E, Adolphson PY . The effect of weekly risedronate on periprosthetic bone resorption following total hip arthroplasty: a randomized, double-blind, placebo-controlled trial. J Bone Joint Surg (Am) 2011; 93: 1857–64. Ström H, Nilsson O, Milbrink J, Mallmin H, Larsson S. Early migration pattern of the uncemented CLS stem in total hip arthroplasties. Clin Orthop Relat Res 2007; 454: 127-32. ten Broeke R H M, Alves A, Baumann A, Arts J JC , Geesink R G T. Bone reaction to a biomimetic third-generation hydroxyapatite coating and new surface treatment for the Symax hip stem. J Bone Joint Surg (Br) 2011; 93-B: 760-8. ten Broeke R H, Hendrickx R P, Leffers P, Jutten L M, Geesink R G. Randomised trial comparing bone remodelling around two uncemented stems using modified Gruen zones. Hip Int 2012; 22: 41-9. van der Voort P, Pijls B G, Nieuwenhuijse M J, Jasper J, Fiocco M, Plevier J W M, Middeldorp S, Valstar E R, Nelissen R G H H. Early subsidence of shape-closed hip arthroplasty stems is associated with late revision. Acta Orthop 2015; 86(5): 575-85. Venesmaa P K, Kröger H P, Miettinen H J, Jurvelin J S, Suomalainen O T, Alhava E M. Alendronate reduces periprosthetic bone loss after uncemented primary total hip arthroplasty: a prospective randomized study. J Bone Miner Res 2001; 16: 2126–31. Wolf O, Milbrink J, Larsson S, Mattsson P, Mallmin H. The optimal timing of baseline radiostereometric analysis of uncemented press fit cups. Scand J Surg 2010; 99(4): 244-9.

4/20 ACTA ORTHOPAEDICA

LONGER IMPLANT SURVIVAL. WITH THE RIGHT BONE CEMENT.

23%

lower revision risk* with PALACOSÂ® R+G compared to other bone cements

* Calculated difference of cumulated revision rates in hip arthroplasty at 13 years of implantation

www.heraeus-medical.com

10034

Vol. 91, No. 4, 2020 (pp. 365â&#x20AC;&#x201C;499)

The element of success in joint replacement

Volume 91, Number 4, August 2020

2006_19476_AZ_PALACOS_EOS_2.0_Acta_Orthop_Europe_215x280_EN.indd 1 TF-IORT200094.indd 1

14.07.20 13:39

26-08-2020 14:56:40