Acta Orthopaedica, Vol. 92, Issue 3, 2021

3/21 ACTA ORTHOPAEDICA

Medical

IMPROVE THE CHANCES

REDUCE RISK FOR INFECTION Reduction of infection risk* using dual antibiotic-loaded bone cement in high risk patients

in aseptic revision TKA * as reported in study results

10453

www.heraeus-medical.com

in fractured neck of femur

Vol. 92, No. 3, 2021 (pp. 249–370)

34 % 69 % 57 %

in primary hip & knee arthroplasty

Volume 92, Number 3, June 2021

COVER.indd 1

26-05-2021 22:08:39

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

EDITORIAL OFFICE

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

THE FOUNDATION BOARD OF THE NORDIC ORTHOPAEDIC FEDERATION AND ACTA ORTHOPAEDICA

EDITOR

Anders Rydholm Lund, Sweden 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 STATISTICAL EDITOR

Jonas Ranstam Lund, Sweden Philippe Wagner Västerås, Lund PRODUCTION MANAGER

Kaj Knutson Lund, Sweden

Vol. 92, No. 3, 2021

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 © 2021 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. 92, No. 3, June 2021

COVID-19, severe injuries Severely injured patients do not disappear in a pandemic: Incidence and characteristics of severe injuries during COVID-19 lockdown in Finland AI, diagnosis of hip complaints Machine learning algorithms trained with pre-hospital acquired history-taking data can accurately differentiate diagnoses in patients with hip complaints Shoulder arthroplasty, revisions The rate of 2nd revision for shoulder arthroplasty as analyzed by the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) Spine, reoperations Reoperations after decompression with or without fusion for L4–5 spinal stenosis with or without degenerative spondylolisthesis: a study of 6,532 patients in Swespine, the national Swedish spine register Hip What is the association between MRI and conventional radiography in measuring femoral head migration?

Direct superior approach versus posterolateral approach in total hip arthroplasty: a randomized controlled trial on early outcomes on gait, risk of fall, clinical and self-reported measurements Population-based 10-year cumulative revision risks after hip and knee arthroplasty for osteoarthritis to inform patients in clinical practice: a competing risk analysis from the Dutch Arthroplasty Register Isometric hip strength impairments in patients with hip dysplasia are improved but not normalized 1 year after periacetabular osteotomy: a cohort study of 82 patients Good clinical outcome for the majority of younger patients with hip fractures: a Swedish nationwide study on 905 patients younger than 50 years of age Good results at 2-year follow-up of a custom-made triflange acetabular component for large acetabular defects and pelvic discontinuity: a prospective case series of 50 hips International variation in distribution of ASA class in patients undergoing total hip arthroplasty and its influence on mortality: data from an international consortium of arthroplasty registries Compensation claims after hip arthroplasty surgery in Norway 2008–2018 Implant survival of 2,723 vitamin E-infused highly crosslinked polyethylene liners in total hip arthroplasty: data from the Finnish Arthroplasty Register Femur, knee Epidemiology and mortality of pelvic and femur fractures—a nationwide register study of 417,840 fractures in Sweden across 16 years: diverging trends for potentially lethal fractures Femoral lengthening might impair physical function and lead to structural changes in adjacent joints: 10 patients with 27 to 34 years’ follow-up T2 relaxation times of knee cartilage in 109 patients with knee pain and its association with disease characteristics Acetabular anteversion in children Development of acetabular anteversion in children with normal hips and those with developmental dysplasia of the hip: a crosssectional study using magnetic resonance imaging

249

A Riuttanen, V Ponkilainen, I Kuitunen, A Reito, J Sirola, and V M Mattila

254

M Siebelt, D Das, A van den Moosdijk, T Warren, P van der Putten, and W van der Weegen

258

D R J Gill, R S Page, S E Graves, S Rainbird, and A Hatton

264

A Joelson, F Nerelius, M Holy, and F G Sigmundsson

269

H-C Husum, M B Hellfritzsch, M Henriksen, K S Duch, M Gottliebsen, and O Rahbek M Ulivi, L Orlandini, J A Vitale, V Meroni, L Prandoni, L Mangiavini, N Rossi, and G M Peretti

274 280

M G J Gademan, L N van Steenbergen, S C Cannegieter, R G H H Nelissen, and P J Marang-van de Mheen

285

J S Jacobsen, S S Jakobsen, K Søballe, P Hölmich, and K Thorborg

292

O Thoors, C Mellner, and M Hedström

297

M Scharff-Baauw, M L van Hooff, G G Van Hellemondt, P C Jutte, S K Bulstra, and M Spruit

304

A J Silman, C Combescure, R J Ferguson, S E Graves, E W Paxton, C Frampton, O Furnes, A M Fenstad, G Hooper, A Garland, A Spekenbrink-Spooren, J M Wilkinson, K Mäkelä, A Lübbeke, and O Rolfson T F Aae, R B Jakobsen, I R K Bukholm, A M Fenstad, O Furnes, and P-H Randsborg M Hemmilä, I Laaksonen, M Matilainen, A Eskelinen, J Haapakoski, A-P Puhto, J Kettunen, K Pamilo, and K T Mäkelä

311 316

323

N Lundin, T T Huttunen, A Enocson, A I Marcano, Li Felländer-Tsai, and H E Berg

329

P A Bjørge, A-T Tveter, H Steen, R Gunderson, and J Horn

335

J Verschueren, S J Van Langeveld, J L Dragoo, S M A BiermaZeinstra, M Reijman, G E Gold, and E H G Oei

341

W Lu, L Li, L Zhang, Q Li, and E Wang

van Neck-Odelberg disease Invasive diagnostic and therapeutic measures are unnecessary in patients with symptomatic van Neck–Odelberg disease (ischiopubic synchondrosis): a retrospective single-center study of 21 patients with median follow-up of 5 years

347

K N Schneider, L P Lampe, G Gosheger, C Theil, M Masthoff, R Rödl, B Vogt, and D Andreou

Skeletal metastases Surgical treatment of skeletal metastases in proximal tibia: a multicenter case series of 74 patients

352

K Kilk, J Ehne, J D Stevenson, G Kask, J Nieminen, R Wedin, M C Parry, and M K Laitinen

Nutrition in old patients Impact of malnutrition and vitamin deficiency in geriatric patients undergoing orthopedic surgery

358

M Meyer, F Leiss, F Greimel, T Renkawitz, J Grifka, G Maderbacher, and M Weber

364

J Sattelberger, H Hillebrand, G Gosheger, A Laufer, A Frommer, S Appelbaum, A A-H Abood, M Gottliebsen, O Rahbek, B Moller-Madsen, R Roedl, and B Vogt

Tension-band experiment Comparison of histomorphometric and radiographic effects of growth guidance with tension-band devices (eight-Plate and FlexTack) in a pig model Information to authors (see http://www.actaorthop.org/)

Acta Orthopaedica 2021; 92 (3): 249–253

249

Severely injured patients do not disappear in a pandemic: Incidence and characteristics of severe injuries during COVID-19 lockdown in Finland Antti RIUTTANEN 1, Ville PONKILAINEN 2, Ilari KUITUNEN 3, Aleksi REITO 4, Joonas SIROLA 5, and Ville M MATTILA 6 1 Department

of Orthopedics, Tampere University, Faculty of Medicine and Health Technology and Tampere University Hospital, Tampere; 2 Department of Surgery, Central Finland Hospital, Jyväskylä; 3 Mikkeli Central Hospital, Mikkeli, University of Eastern Finland, School of Medicine, Kuopio; 4 Department of Orthopedics, Tampere University, Faculty of Medicine and Health Technology and Tampere University Hospital, Tampere; 5 Kuopio Musculoskeletal Research Unit (KMRU), University of Eastern Finland, Kuopio, University Hospital, Department of Orthopedics, Traumatology and Hand Surgery, Kuopio; 6 Department of Orthopedics, Tampere University, Faculty of Medicine and Health Technology and Tampere University Hospital, Tampere, Finland Correspondence: antti.riuttanen@pshp.fi Submitted 2020-11-09. Accepted 2021-01-11.

Background and purpose — COVID-19 lockdowns have resulted in noteworthy changes in trauma admissions.

We report and compare the incidence and characteristics of severe injuries (New Injury Severity Score [NISS] >15) during the COVID-19 lockdown in Finland with earlier years.

Methods — We retrospectively analyzed incidence rate, injury severity scores, injury patterns, and mechanisms of injury of all severely injured patients (NISS >15) in 4 Finnish hospitals (Tampere University Hospital, Kuopio University Hospital, Central Finland Hospital, Mikkeli Central Hospital) during the 11-week lockdown period (March 16–May 31, 2020) with comparison with a matching time period in earlier years (2016–2018). These 4 hospitals have a combined catchment area of 1,150,000 people or roughly one-fifth of the population of Finland.

Results — The incidence rate of severe injuries during the lockdown period was 4.9/105 inhabitants (95% CI 3.7–6.4).

The incidence rate of severe injuries during years 2016–2018 was 5.1/105 inhabitants (CI 3.9–6.5). We could not detect a significant incidence difference between the lockdown period and the 3 previous years (incidence rate difference –0.2 (CI –2.0 to 1.7). The proportion of traffic-related accidents was 55% during the lockdown period and 51% during previous years. There were no detectable differences in injury patterns. During the lockdown period, the mean age of patients was higher (53 years vs. 47 years, p = 0.03) and the rate of severely injured elderly patients (aged 70 or more) was higher (30% vs. 16%).

Interpretation — Despite heavy social restrictions, the incidence of severe injuries during the lockdown period was similar to previous years. Notably, a decline in road use and traffic volumes did not reduce the number of severe traffic accidents. Although our data is compatible with a decrease of 2.0 to an increase of 1.7 severely injured patients per 105 inhabitants, we conclude that severely injured patients do not disappear even during pandemic and stabile hospital resources are needed to treat these patients.

On March 16, 2020, the Finnish Government declared a state of emergency in response to the COVID-19 outbreak. All permanent residents were asked to minimize social contacts and to avoid non-essential travel and spending time in public places. Schools and educational institutions were closed down and face-to-face teaching was suspended. In addition, remote working was recommended where possible. Residents aged 70 years and older were told to stay in quarantine-like conditions at home. The capacity of healthcare and social welfare services was increased, and non-urgent activities were reduced (Government Communications Department 2020a). As the highest incidence of COVID-19 was observed in the Uusimaa region, which is the most populous area (population 1,700,000) in Finland, a decision was made to further restrict movement by means of an Emergency Powers Act (Government Communications Department 2020b). The police enforced a lockdown of the Uusimaa region (from March 27 to April 15, 2020),

© 2021 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 10.1080/17453674.2021.1881241

250

Acta Orthopaedica 2021; 92 (3): 249–253

investigates the incidence rate and characteristics of severe injuries (New Injury Severity Score [NISS] >15) during the lockdown period in Finland and how this rate compares with earlier years.

Monthly change (%) in traffic volume 10

0

Methods

–10

–20

–30

–40

CFH KUH MCH TAUH

January Febr. March April

May

June

July August

Monthly change in traffic volumes in 2020 compared with 2019 in the 4 hospital regions participating in this study. CFH = Central Finland Hospital region, KUH = Kuopio University Hospital region, MCH =Mikkeli Central Hospital region, TAUH = Tampere University Hospital region.

and restricted all non-essential traffic to and from the Uusimaa region. The capital city of Finland, Helsinki, is located in the Uusimaa region. All of the above steps combined resulted in a one-third decline in traffic volumes on the major roads across Finland (Finnish Transport Infrastructure Agency 2020).

After the widespread transmission of the disease in Europe, similar regional or national restrictions were set in place in various countries (Lau et al. 2020). The different lockdown procedures implemented have been reported to have had substantial effects on various types of trauma. An early study from New Zealand reported that lockdown resulted in a pronounced 43% reduction in all trauma admissions in a single trauma center (Christey et al. 2020). Similar findings were also reported from the UK, where trauma referrals decreased by nearly 50% (Park et al. 2020). As a result of decreased traffic volumes, the largest reduction was seen in road traffic injuries (Christey et al. 2020).

Social distancing and recommendations to avoid large crowds have also increased the amount of time spent at home. Consequently, the place of injury has moved from sportsgrounds to the home (Pellegrini et al. 2020). In France, for example, the rate of domestic hand injuries treated at university trauma hand units doubled, even though the overall rate of upper limb emergencies plummeted (Pinggera et al. 2020).

It is likely that the lockdown procedures imposed in Finland during the COVID-19 pandemic could have changed the incidence and characteristics of injuries in Finland. However, despite multiple reports indicating notable changes in various types of trauma, the information concerning severe injuries is limited. As treatment of both severe injuries and severe COVID-19 infection are resource-consuming, it is important to find out whether the lockdown procedures have had an effect on incidence of severe injuries. Therefore, this study

This study is a retrospective cohort study comparing severely injured patient during the lockdown period with severely injured patients treated in the years 2016–2018. The study was conducted in 4 Finnish hospitals (Tampere University Hospital [TAUH], Kuopio University Hospital [KUH], Central Finland Hospital [CFH], and Mikkeli Central Hospital [MCH]) serving a combined population base of approximately 1,150,000 inhabitants (Statistics Finland 2020a). The study population covers roughly one-fifth of the whole population of Finland.

To assure adequate coverage we selected 2 major university hospitals, 1 big central hospital, and 1 smaller central hospital.

All 4 hospitals operate as the primary trauma care provider in their hospital district providing immediate services in orthopedic surgery, anesthesiology, emergency medicine, radiology, internal medicine, plastic surgery, oral and maxillofacial surgery, pediatrics, and critical care. In addition, TAUH and KUH also serve as tertiary trauma care units providing immediate services in neurosurgery.

The study population consists of severely injured patients treated during the 11-week lockdown period (March 16–May 31, 2020). The data from patients treated in the TAUH region was extracted from TAUH’s Trauma Registry. All trauma patients treated at TAUH meeting the inclusion criteria: (NISS > 15, minimum Maximum Abbreviated Injury Score (MAIS) 3, treated at intensive care unit (ICU) or High Dependency Unit (HDU) have been enrolled prospectively into the TAUH’s Trauma Registry since 2015. Each year, approximately 150 to 200 severely injured patients are enrolled. As TAUH is the only hospital among the participating hospitals with a trauma registry, the data from the other hospitals was retrospectively collected from patient records and coded to match the TAUH Trauma registry. All trauma CTs in these 3 hospitals were retrospectively inspected and both the Injury Severity Score (ISS) and the NISS were calculated. The following inclusion criteria were used: NISS > 15, MAIS ≥ 3, admission through emergency room to further treatment at ICU or HDU. As the number and characteristics of severe injuries may vary between years, we calculated the reference incidence of severe injuries on the basis of a 3-year average (2016–2018) for all 4 hospitals. The reference incidence for the TAUH region was derived from prospectively collected data from TAUH’s Trauma Register. The reference population for the other 3 hospitals was collected retrospectively with matching inclusion criteria. The 11-week reference period was set out to match the lockdown period in Finland (March 16–May 31, 2020).

The incidence rate of severe injuries was calculated based on

Acta Orthopaedica 2021; 92 (3): 249–253

the population of each hospital area: TAUH (530,000 inhabitants), KUH (250,000), CFH (270,000), and MCH (100,000) (Statistics Finland 2020a).

Mechanisms of injury were extracted from TAUH’s Trauma Registry or from medical records (KUH, CFH, MCH). The causes of injuries were classified with the injury-related 10th revision of the International Classification of Diseases and Related Health Problems (ICD-10) as either traffic- (V01– V99) or non-traffic-related injuries (W00–X59), such as falls and assaults. Traffic-related injuries (V01–V99) included all motor vehicle collisions, such as crashes and driving off the road (V49), injuries caused to pedestrians (V01–V09), motorbike accidents (V28, V29) and injuries that occurred while riding a bicycle (V10–V19). Other traffic-related injuries include miscellaneous injuries, such as collisions with trains (V81) and all-terrain vehicle accidents (V86). Falls (W00– W19) were further divided into high (≥ 3 meters) and low falls (< 3 meters), to better distinguish high-energy falls from lowenergy falls. Assaults (X85–Y09) included all forms of physical abuse, including shootings and stabbings. Other accidents included injury mechanisms, such as crush injuries (W23) or injuries caused by falling items (W20), that did not fit in the previous classifications.

Injury patterns were based on the Abbreviated Injury Scale (AIS) (Committee on Medical Aspects of Automotive Safety 1971) with pelvic injuries classified as lower extremity injuries. Minor (AIS1) injuries, such as cuts and bruises, were excluded from the analysis concerning injury patterns because they are seldom collected systematically in medical records.

The severity of the injuries was classified according to AIS version 2015. Both the Injury Severity Score (ISS) (Osler et al. 1997) and the New Injury Severity Score (NISS) (Baker et al. 1974) were calculated.

Length of stay in ICU/HDU was obtained from TAUH’s Trauma Register or from the medical records of individual patients.

Statistics Statistical analysis was performed with R version 4.0.2 (R Foundation for Statistical Computing, Vienna, Austria). Age was reported using mean (SD) and ISS was reported using median (IQR). Welch’s t-test was used to compare means. The Mann–Whitney U test was used to compare ranks between groups. A chi-square test without Yates’ correction was used to compare the proportions of trauma mechanisms between different years. Incidences between time periods were compared using incidence rate difference. Confidence interval (CI) was determined at 95%, and therefore p-values < 0.05 were considered to be statistically significant. CIs for incidence rates were calculated using Poisson’s exact method.

Ethics, funding, and potential conflicts of interest Under Finnish legislation, the current study is exempt from the need to obtain ethical approval because of its retrospec-

251

Table 1. Patients treated during lockdown period (March 16 to May 31, 2020) and reference period (2016–2018) at 4 hospitals. Values are frequency (%) unless otherwise specified

Factor

2020 2016–2018 n = 56 n = 173 p-value

Age, mean (SD) 47 (21) 53 (19) 0.03 a Sex, male 43 (77) 132 (76) 0.9 b ASA 1–2 26 (46) 79 (46) ASA 3–4 30 (54) 94 (54) 0.9 b Injury severity scores, median (interquartile range) ISS 18 (9) 21 (10) 0.008 c NISS 24 (10 27 (13) 0.1 c Mechanism of injury 0.5 b Traffic accidents 31 (55) 88 (51)

Car 13 (23) 50 (29)

Motorbike 7 (13) 16 (9.3)

Bicycle 7 (13) 11 (6.4)

Pedestrian 2 (4) 5 (2.9)

Other 2 (4) 6 (6.8) Low fall (< 3 m) 13 (23) 40 (23) High fall (> 3 m) 8 (5) 28 (16) Self-inflicted (suspected) 2 (4) 0 (0) Assault (suspected) 2 (4) 7 (4.0) Blunt injury 54 (96) 166 (96) Penetrating 2 (4) 7 (4.0)

Gun shot 1 (2) 1 (0.6)

Stabbing 1 (2) 6 (3.5) Other 0 (0) 8 (4.6) Injury pattern (AIS, minimum 2) 0.9 b Head 32 (57) 106 (61) Face 7 (13) 20 (2) Thorax 25 (45) 93 (54) Abdomen 17 (30) 48 (28) Pelvis and extremities 19 (34) 67 (39) External (soft tissue) 3 (5) 5 (2.9) Length of stay in ICU, median (interquartile range) 2 (3) 2 (4) 0.6 c

AIS = Abbreviated Injury Score.

ICU = Intensive care unit. a Welch’s t-test. b Chi-square test. c Mann–Whitney U test.

tive nature, as stated by the Regional Ethics Committee of the Expert Responsibility area of Tampere University Hospital.

This study was financially supported partly by the Competitive State Research Financing of the Expert Responsibility area of Tampere University Hospital. The authors declare they have no competing interests.

Results During the lockdown period (March 16–May 31, 2020) 56 severely injured patients with matching inclusion criteria were treated at the 4 participating hospitals (Table 1), covering a catchment area of around 1.15 million inhabitants. Of these, 35 patients were treated at TAUH, 10 patients at KUH, 5 patients at CFH, and 6 patients at MCH. The patients were mostly men

252

Acta Orthopaedica 2021; 92 (3): 249–253

Table 2. Number of severely injured patients (n) and incidence rate of severe injuries during lockdown (March 16 to May 31, 2020) and reference period (2016–2018) treated at 4 hospitals

2020 2016–2018

Mean Incidence Mean Incidence Incidence rate

n population rate (95% CI) n population rate (95% CI) difference (95% CI)

Total 56 1,133,274.5 4.9 (3.7–6.4) 58 1,132,671.7 5.1 (3.9–6.5) –0.2 (–2.0 to 1.7) Pirkanmaa Hospital District 35 536,135 6.5 (4.5–9.1]) 38 531,050.5 7.2 (5.2–9.8) –0.7 (–3.8 to 2.5) Hospital District of Northern Savo 10 244,919 4.1 (2.0–7.5) 9 247,098.2 3.8 (1.9–6.9) 0.3 (–3.2 to 3.8) Central Finland Hospital District 5 252,696 2.0 (0.6–4.6) 8 252,614.3 3.0 (1.4–5.7) –1.1 (–3.8 to 1.7) Southern Savo Hospital District 6 99,524.5 6.0 (2.2–13) 2 101,908.7 2.3 (0.6–7.0) 3.7 (–1.9 to 9.4)

CI = confidence interval

(n = 43, 77%) with a mean age of 53 years (SD 19). During the matching reference period (March 16–May 31), the number of severely injured patients treated at the participating hospitals in the years 2016 to 2018 was 173, or an average of 58 patients per year (Table 1). Of these 173 patients, 115 were treated at TAUH, 28 at KUH, 23 at CFH, and 7 at MCH. The patients were predominantly male (n = 132, 76%) with a mean age of 53 years (SD 19).

Both during and before lockdown, roughly half of all injuries were traffic related (55% vs. 51%). The majority of the injuries (96%) were caused by a blunt mechanism (Table 1).

The injury patterns of those patients treated before and during the lockdown period are also shown in Table 1. The rate of thoracic injuries was lower (45% vs. 54%) during the lockdown period. Otherwise, there were no notable differences in injury patterns. During the lockdown period the median ISS was 18 (IQR 9) and median NISS 24 (IQR10).

The incidence rate of severe trauma during lockdown was 6.5/105 in the TAUH region, 4.1/105 in the KUH region, 2.0/105 in the CFH region, and 6.0/105 in the MCH region.

The overall observed incidence of severe injuries throughout the whole study population (1,150,000) was 4.9/105 (Table 2).

The average annual reference period incidences were 7.2/105 in the TAUH region, 3.7/105 in the KUH region, 2.9/105 in the CFH region, and 2.3/105 in the MCH region. The overall reference incidence of severe injuries throughout the whole study population was 5.1/105. The change in incidence between study points was –0.2/105.

Discussion Despite heavy social restrictions, the incidence of severe injuries during the lockdown period (March 16–May 31, 2020) was similar when compared with earlier years (2016–2018).

The incidence rate decreased 0.2/105; however, our data showed a decrease of 2.0 to increase of 1.7 severely injured patients per 105 inhabitants.

Several studies have been published on observed changes on traumatic injuries treated at ERs during lockdowns. However, most of these studies have concentrated on milder inju-

ries, which are usually more common and their trends easier to follow. Interestingly, a recent article has suggested that the incidence of or the need for intensive care for traumatic head injuries in the Uusimaa Hospital District in Finland was not affected by lockdown procedures (Luostarinen et al. 2020).

We are unaware of previous studies that have examined the incidence of severe injuries during the COVID-19 lockdown in a larger patient population.

Contrary to what we expected, we were not able to detect a statistically significant difference in the rate of traffic-related injuries even when there was a considerable reduction in traffic volumes on major roads. The absolute number of patients injured in traffic-related accidents was 31, which is close to the absolute annual average (29 patients per year). The high rate of traffic-related injuries can be explained by several factors. First, even in a lockdown period, people are forced to use motor vehicles for commuting. Second, as reported by the National Police Board of Finland, less traffic on less-crowded roads resulted in increased risky behavior, such as speeding and driving under the influence (DUI) (YLE News 2020). The third factor can be explained by our inclusion criteria; in order for a patient to suffer an injury serious enough to be included in the study, the accident must have had enough kinetic energy.

The injury patterns were similar before and during the lockdown period. During lockdown, the number of deaths due to road traffic injuries did not increase (Statistics Finland 2020b).

It is a well-known fact that the majority of trauma patients are usually working-age males. In Finland, one part of the national lockdown policy was a countrywide recommendation directly from the President of Finland to all citizens aged 70 years and over to stay at home in quarantine-like conditions (Government Communications Department 2020a). Thus, we expected to see some reduction in the age profile of injury victims. Contrary to what we expected, the average age of severely injured patients actually increased from a mean age of 47 years to 53 years. Furthermore, 30% of patients were over 70 years of age.

It is evident that a second “wave” of COVID-19 with a rapidly increasing number of infections is currently emerging in Europe (Dong et al. 2020). Various European countries, such as France, Italy, Spain, Germany, and Belgium, have already

Acta Orthopaedica 2021; 92 (3): 249–253

brought in new restrictions, which bear a resemblance to those in the earlier lockdowns (BBC News 2020). Our findings suggest that the need for intensive care for severely injured patients remained unchanged throughout the lockdown period.

In addition, both the length of stay in ICU and the injury severity scores were comparable before and during lockdown.

The treatment of severely injured patients often requires major hospital resources that include operation rooms and ICU, which in turn may have to be re-utilized in treating patients with severe COVID-19 infection. Since the COVID-19 pandemic has increased the use of these essential resources, it could be argued that success in the acute treatment of severely injured patients can been seen as a kind of benchmark of a well-functioning healthcare system (Publications Office of the European Union 2019).

The strength of our study lies in the collaboration among four major hospitals in 4 hospital districts. Together, these 4 hospital districts provide all major trauma care for roughly one-fifth (1,150,000) of the whole population of Finland (5,518,000). Therefore, it can be assumed that these results are likely generalizable to most of the country. Also, as these hospitals are the sole providers of major trauma care in their hospital districts, it is unlikely that any major trauma was treated outside of these hospitals during the lockdown period. With the exception of the Uusimaa region, the number of COVID19 patients treated at ICUs in Finland was fairly low (National Coordinating Office for Intensive Care 2020).

In conclusion, we did not detect a difference in the incidence of severe injuries during the lockdown period in Finland when compared with previous years. The one-third decline in traffic volumes did not reduce the number of severe traffic accidents. Even though there were small changes in patient demographics, severely injured patients do not seem to “disappear” during lockdown, even when strict restrictions are in place. Thus, stable hospital resources are needed to treat severely injured patients.

ARi conceived of the study, participated in data collection, and drafted the manuscript. VP participated in data collection, commented on the manuscript, and conducted data analysis. IK participated in data collection and commented on the manuscript. JS participated in data collection and commented on the manuscript. ARe participated in the study design and commented on the manuscript. VM participated in the design, coordination, and drafting of the manuscript.

The authors wish to thank trauma registry nurse Sonja Nieminen for her cooperation with this study.

Acta thanks Louis Riddez and Martin Gerdin Wärnberg for help with peer review of this study.

Baker S P, O’Neill B, Haddon W Jr, Long W B. ‘The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care’, J Trauma 1974; 14: 187-96.

253

BBC News. Covid: What are the lockdown rules in place across Europe?

2020. Available from: https://wwwbbccom/news/explainers-53640249.

Accessed October 29, 2020.

Christey, G, Amey J, Campbell A, Smith A. Variation in volumes and characteristics of trauma patients admitted to a level one trauma centre during national level 4 lockdown for COVID-19 in New Zealand. N Z Med J 2020; 133: 81-8.

Committee on Medical Aspects of Automotive Safety. Rating the severity of tissue damage, I: The Abbreviated Scale. JAMA 1971; 215: 277-80.

Dong E, Du H, Gardner L. An interactive web-based dashboard to track COVID-19 in real time. Lancet Infect Dis 2020; 20: 533-4.

Finnish Transport Infrastructure Agency. Road Statistics, open data. 2020.

Available from: https://vaylafi/vaylista/aineistot/avoindata.

Government Communications Department, Ministry of Education and Culture, Ministry of Social Affairs and Health. Government, in cooperation with the President of the Republic, declares a state of emergency in Finland over coronavirus outbreak. 2020a. Available from: https://valtioneuvostofi/ en/article/-/asset_publisher/10616/hallitus-totesi-suomen-olevan-poikkeusoloissa-koronavirustilanteen-vuoksi. Accessed June 8, 2020.

Government Communications Department, Ministry of the Interior.

Restrictions on movement to and from Uusimaa enter into force on 28 March 2020. 2020b. Available from: https://valtioneuvostofi/artikkeli/-/ asset_publisher/10616/liikkumisrajoitukset-uudellemaalle-voimaan28-maaliskuuta-2020-klo-00-00?_101_INSTANCE_LZ3RQQ4vvWXR_ languageId=en_US. Accessed June 8, 2020.

Lau H, V Khosrawipour, Kocbach P, Mikolajczyk A, Schubert J, Bania J, Khosrawipour T. The positive impact of lockdown in Wuhan on containing the COVID-19 outbreak in China. J Travel Med 2020; 27.

Luostarinen T, Virta J, Satopaa J, Backlund M, Kivisaari R, Korja M, Raj R. Intensive care of traumatic brain injury and aneurysmal subarachnoid hemorrhage in Helsinki during the Covid-19 pandemic. Acta Neurochir (Wien) 2020; 162: 2715-24.

National Coordinating Office for Intensive Care. Tehohoidon koordinoiva toimisto. Situation report., 2020. Available from: https://wwwpsshpfi/ documents/7796350/7841067/Tehohoidon+tilannekuva+-+Koordinoiva n+toimiston+viikkoraportti+2020_06_03pdf/5acf3a25-8fe0-44d2-80d1c0b9b4cdc378. Accessed October 26, 2020.

Osler T, Baker S P, Long W. A modification of the Injury Severity Score that both improves accuracy and simplifies scoring. J Trauma Acute Care Surg 1997; 43: 922-6.

Park C, Sugand K, Nathwani D, Bhattacharya R, Sarraf K M. Impact of the COVID-19 pandemic on orthopedic trauma workload in a London level 1 trauma center: the ‘golden month’. Acta Orthop 2020; 91(5): 556-61.

Pellegrini M, Roda M, Di Geronimo N, Lupardi E, Giannaccare G, Schiavi C. Changing trends of ocular trauma in the time of COVID-19 pandemic. Eye (Lond) 2020; 34(7): 1248-50.

Pinggera D, Klein B, Thome C, Grassner L. The influence of the COVID-19 pandemic on traumatic brain injuries in Tyrol: experiences from a state under lockdown. Eur J Trauma Emerg Surg 2020; Jul 22: 1-6. Online ahead of print.

Publications Office of the European Union. Tools and methodologies to assess the efficiency of health care serviced in Europe. 2019. Available from https://eceuropaeu/health/sites/health/files/systems_performance_assessment/docs/2019_efficiency_enpdf. Accessed October 21, 2020.

Statistics Finland. Population structure. 2020a. Available from: http://pxnet2statfi/PXWeb/pxweb/fi/StatFin/StatFin__vrm__vaerak/statfin_vaerak_ pxt_11s1px/.

Statistics Finland. Tieliikenneonnettomuustilasto [toukokuu (2020) Liitetaulukko 2 Tieliikenneonnettomuuksissa kuolleet ja loukkaantuneet. 2020b.

Available from: http://wwwstatfi/til/ton/2020/05/ton_2020_05_2020-0616_tau_002_fihtml Accessed October 20, 2020.

YLE News. Coronavirus directly linked to heavily increased speeding Koronalla on suora yhteys rajusti lisääntyneisiin törkeisiin ylinopeuksiin— ‘Tilanne on nuorten kannalta ihan hirveä’. 2020. Available from: https:// ylefi/uutiset/3-11353861. Accessed October 26, 2020.

254

Acta Orthopaedica 2021; 92 (3): 254–257

Machine learning algorithms trained with pre-hospital acquired historytaking data can accurately differentiate diagnoses in patients with hip complaints Michiel SIEBELT 1, Dirk DAS 1, Amber VAN DEN MOOSDIJK 1, Tristan WARREN 1, Peter VAN DER PUTTEN 2, and Walter VAN DER WEEGEN 1 1 Department of Orthopedic Surgery, St Anna Hospital, Geldrop; 2 Leiden Institute of Advanced Computer Science, Leiden University Leiden, The Netherlands Correspondence: w.vander.weegen@st-anna.nl Submitted 2020-06-22. Accepted 2021-01-11.

Background and purpose — Machine learning (ML) techniques are a form of artificial intelligence able to analyze big data. Analyzing the outcome of (digital) questionnaires, ML might recognize different patterns in answers that might relate to different types of pathology. With this study, we investigated the proof-of-principle of ML-based diagnosis in patients with hip complaints using a digital questionnaire and the Kellgren and Lawrence (KL) osteoarthritis score.

Patients and methods — 548 patients (> 55 years old) scheduled for consultation of hip complaints were asked to participate in this study and fill in an online questionnaire.

Our questionnaire consists of 27 questions related to general history-taking and validated patient-related outcome measures (Oxford Hip Score and a Numeric Rating Scale for pain). 336 fully completed questionnaires were related to their classified diagnosis (either hip osteoarthritis, bursitis or tendinitis, or other pathology). Different AI techniques were used to relate questionnaire outcome and hip diagnoses. Resulting area under the curve (AUC) and classification accuracy (CA) are reported to identify the best scoring AI model. The accuracy of different ML models was compared using questionnaire outcome with and without radiologic KL scores for degree of osteoarthritis.

Results — The most accurate ML model for diagnosis of patients with hip complaints was the Random Forest model (AUC 82%, 95% CI 0.78–0.86; CA 69%, CI 0.64–0.74) and most accurate analysis with addition of KL scores was with a Support Vector Machine model (AUC 89%, CI 0.86–0.92; CA 83%, CI 0.79–0.87).

Interpretation — Analysis of self-reported online questionnaires related to hip complaints can differentiate between basic hip pathologies. The addition of radiological scores for osteoarthritis further improves these outcomes.

Use of artificial intelligence (AI) techniques like data mining, machine learning (ML), and deep learning are now starting to erupt within healthcare, with first applications aimed at cancer diagnostics (Nguyen et al. 2018, Codari et al. 2019), cardiology (Nirschl et al. 2018) and image recognition in radiology (Wang et al. 2017, Fourcade and Khonsari 2019).

AI is also emerging within the field of orthopedic surgery (Duffield et al. 2017). Earlier work using AI in orthopedic studies showed the ability of ML to classify knee osteoarthritis (OA) subjects versus healthy patients. Based on kinematic data Kotti et al. (2017) achieved an accuracy of 73%. In comparison with that study, which collected its data in a laboratory setting, Dolatabadi et al. (2017) used kinematic data from more unobtrusive sensors and were also able to distinguish OA subjects from healthy patients. Other ML-related publications in orthopedics report on spine pathology detection, fracture detection, and bone and cartilage image segmentation (Ashinsky et al. 2015).

However, to our knowledge, no studies in orthopedics have developed ML algorithms for predicting a clinical diagnosis.

In this paper we used information from digital intake forms, which were completed online by our patients before initial consultation with an orthopedic surgeon. We sought to determine (1) the accuracy of different ML algorithms to predict a pre-hospital diagnosis in patients suffering from hip complaints based on history-taking questions only, and (2) how much radiographic imaging results contribute to accurately predicting a diagnosis in these patients.

Patients and methods For the development of an ML algorithm we designed a prospective cohort study that included patient data from a single hospital (St Anna Hospital, Geldrop, The Netherlands).

© 2021 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 10.1080/17453674.2021.1884408

Acta Orthopaedica 2021; 92 (3): 254–257

All patients aged > 55 years with hip complaints were eligible for inclusion. Immediately after contacting the hospital to schedule an appointment, all participating patients received our questionnaire by e-mail, which had a hyperlink embedded, leading to a secure online environment (Interactive Studios, Rosmalen, the Netherlands). Here, patients were able to answer all questions before initial consultation, which was usually within 1 to 2 weeks. For this purpose we used our online patient reported outcome measurement (PROM) system. This system is normally used to collect standardized PROMs before and after surgery to track orthopedic healthcare outcomes from a patient’s perspective. Within this online environment, 2 authors (MS and WvdW) created a new questionnaire for the purpose of this study, which was verified by a third author (DD).

This new questionnaire included standard history-taking questions for suspected hip pathology (i.e., location of pain, severity and duration of symptoms; an overview of all questions is presented in Supplementary data 1—complete questionnaire). These questions were combined with well-validated PROM questionnaires: the Oxford Hip Score (OHS) (Dawson et al. 1996, de Groot et al. 2007), and severity of pain measured with a Numeric Rating Scale (NRS) (Salaffi et al. 2004). Questionnaires of patients who responded to our digital intake form were checked. Incomplete questionnaires were excluded, except for missing answers in the Oxford Hip Score (OHS). As advocated for this specific hip score, a maximum of 2 missing items is allowed and can be dealt with by replacing missing scores with the average score of completed items (Dawson et al. 1996, Murray et al. 2007).

After history taking, physical examination and radiographic evaluation, all patients were informed of their diagnosis by their consulting orthopedic surgeon. We retrieved this diagnosis from the medical file and linked it to the questionnaire for that specific patient. This diagnosis was assigned to 1 of 3 categorical outcomes: (1) osteoarthritis (OA) of the hip; (2) bursitis or tendinitis around the hip; or defined as (3) other pathology. These 3 diagnoses were chosen since they represent a large portion of hip complaints. For this proof-ofprinciple we did not want to start with more diagnoses, since ML techniques will have more trouble differentiating between many possible outcome options and therefore would require larger numbers of patient-reported questionnaires.

This dataset was imported into Orange Workflow (version 3.22, Ljubljana, Slovenia), which is an open-source AI software system using different ML techniques. Using Orange, a data file was created to train and test the algorithms in a 10-fold stratified cross-validation loop. The 27 variables were ranked for their ability to differentiate between the 3 diagnosis groups by averaging the outcomes of multiple ranking techniques (Information Gain, Information Gain Ratio, Gini Decrease, X2, ReliefF, and Fast Correlation Based Filter [FCBF]). We trained and tested all ML models available in Orange Workflow (Constant, CN2 rule induces, k Nearest Neighbour [kNN], Tree, Random

255

Table 1. Selected hyperparameters for each evaluated algorithm Algorithm

Hyperparameter Value

SVM Epsilon 0.1

Cost (C) 1

Kernel RBF Decision tree Min. number of

instances in leaves 2

Do not split subsets smaller than 5

Max. tree depth 100 Logistic regression Regularization L2

Cost (C) 1 Neural network Hidden neurons 100

Activation ReLu

Solver Adam

Alpha 0.0001 KNN Number of neighbors 5

Metric Euclidean

Weight Uniform Random Forest Number of trees 10

Number of attributes

considered at each split 5

Max. tree depth 3

Do not split subsets smaller than 5

Forest, Support Vector Machine (SVM), Logistic Regression, Naïve Bayes, AdaBoost, and Neural Network). Orange does not have hyperparameter tuning capabilities so hyperparameters were selected by hand (for an overview of the hyperparameters see Table 1). Resulting area under the curve (AUC) and classification accuracy (CA) outcomes were used to identify the best scoring model (Duffield et al. 2017). 95% confidence intervals (CI) were calculated for AUC and AC with each ML model.

Each model was first trained and evaluated on the questionnaire data with all questions included. Next, we investigated the possibility of achieving similar performance with fewer questions included in the dataset. For this purpose we evaluated the performance of predictive models that were trained on data only including the top 5 ranking questions, and in a second experiment only including the top 10 ranking questions. To analyze the contribution of radiographic imaging results to the diagnosis process, we scored the pelvic radiograph for each included patient using the Kellgren–Lawrence (KL) scoring method (Kellgren and Lawrence 1957) and trained and tested the algorithms again with this score KL added to the full dataset. With inclusion of KL scores, the model was again retested with all questions of the questionnaire, with only the top 5 questions and using only the top 10 questions.

Ethics, registration, funding, and potential conflicts of interest This study was reviewed by the regional medical ethical committee and was considered to be exempt from full review (registration number N19.066) according to Dutch law. The study protocol was registered in the Dutch Trial Register (Trial registration number NL8229). No external funding was obtained.

The authors report no conflicts of interest.

256

Acta Orthopaedica 2021; 92 (3): 254–257

Patients invited to complete questionnaire n = 548 Excluded (n = 212): – did not respond to invitation, 31 – incomplete questionnaires, 70 – no diagnosis at first consultation, 96 – incorrect referral, 7 – unable to identify patient, 8 Final dataset including all patients with complete questionnaire and with diagnosis n = 336

Study flow.

Results Questionnaires of 517 participating patients were received, but after checking for completeness of answers 336 patients could be included in the study (Figure). The collection of this data resulted in a dataset with 283 observations of 27 variables from the questionnaire (see Supplementary material) and 1 target variable (diagnosis group). The distribution of the target variables is as follows. 191 (68%) patients were diagnosed with OA, 61 (22%) patients were diagnosed with bursitis or tendinitis around the hip, and 31 (11%) were diagnosed with other pathology. There is a clear imbalance in the distribution of the target variables with OA being the overrepresented class. The Random Forest algorithm with 20 folds trained on the full dataset (all answers to all questions included) resulted in the highest AUC (82%, CI 0.78–0.86) and CA (69%, CI 0.64–0.74). The 5 most differentiating questions were (in decreasing order of differentiating power): 1. OHS 4: Have you been able to put on a pair of socks, stockings, or tights?

2. Do you experience pain in the groin area?

3. Does your hip feel stiff during the first steps you take when walking?

4. Would you be willing to undergo surgery if needed?

5. OHS 7: Are you able to walk up and down the stairs?

All ML models were tested to see if the AUC and/or CA improved by leaving out possibly less important questions.

Using only the top 5 questions in the training set, logistic regression (10 folds) gave the highest results of 81% AUC (CI 0.77–0.85) and 73% CA (CI 0.67–0.77). When we selected the top 10 ranking questions, Neural Network (10 folds) gave the highest results with 74% AUC (CI 0.69–0.78) and 67% CA (CI 0.62–0.72) (Table 2). Adding the radiographic data to the dataset increased both the AUC and the CA of the ML models. The distribution of the KL scores accross the 3 diagnosis groups is presented in Table 3. Under this condition, SVM resulted in the highest AUC and CA scores of 89% (CI 0.86–0.92) and 83% (CI 0.79–0.87) respectively (Table 2). A full overview of all ML algorithms is described in Supplementary data.

Table 2. Artificial intelligence analysis using machine learning (ML) algorithms on pre-hospital-acquired patient history-taking form for patients aged > 55 years with hip complaints. Values are ML algorithm accuracy in percent

History-taking only KL score added Prediction Prediction Dataset AUC CA model AUC CA model All questions Top 5 questions only Top 10 questions only

82 69 RF 89 83 SVM 82 73 SVM 85 79 SVM 78 70 SVM 79 79 SVM

RF = Random Forest.

SVM = Standard Vector Machine.

Table 3. Distribution (%) of the KL scores accross the 3 diagnosis groups

Diagnosis

KL score 0 1 2 3 4

Bursitis/tendinitis Osteoarthritis Other

24 55 20 1 0 1 7 26 48 18 11 52 11 15 11

Discussion Computer algorithms which use patients’ answers to digital history-taking questions are capable of differentiating a hip complaint related diagnosis with fairly good accuracy (AUC 74% and CA 67%). Adding radiographic information results in even higher accuracy and improves AI performance (AUC 89% and CA 83%). Obviously, there are clear logistical problems that need to be solved in order to achieve integration of conventional radiological examination into pre-hospital AI analysis, but this study shows the proof-of-principle for ML techniques in orthopedics.

Our approach using ML may help improve patient care in many ways. With accurate prediction of diagnosis and related treatment, patients can be educated about their condition in advance of their hospital visit, which is easily managed by using a smartphone app. Such an app with supporting information may help patients to increase knowledge and understanding of underlying hip pathology related to their complaint.

Subsequently, patients might experience a more in-depth first consultation with the orthopedic surgeon during their hospital visit (Timmers et al. 2018). In order to test this hypothesis, we are currently enrolling a prospective randomized controlled trial using pre-hospital AI diagnosis and its effect on patient knowledge and satisfaction levels during hospital consultation.

Besides patient satisfaction, ML diagnosis may also increase outpatient clinic efficiency. First, patients who are more likely to have a diagnosis that is treated nonoperatively can be grouped together when outpatient clinic appointments

Acta Orthopaedica 2021; 92 (3): 254–257

are scheduled. Other supporting healthcare providers (i.e., physician assistants, physical therapists) can be scheduled to join these consultations and patients’ complaints may be dealt with by a multidisciplinary team. Second, patients who are more likely to be treated with surgery (e.g., hip arthroplasty) can also be grouped together and simultaneously planned for preoperative screening, reducing the number of visits needed to the hospital to a minimum.

This predictive analytic study has several limitations. Most importantly, our questionnaire is of course in need of validation in other hospitals. Next, we grouped multiple hip pathologies in 3 categorical groups. This does not cover clinical reality in which orthopedic surgeons make a much more detailed diagnosis. Since this study is a first exploration of ML applied to the clinical diagnostic process in orthopedic surgery and history-taking in particular, we consider our approach justifiable for now. Larger datasets should allow further explorations using more detailed diagnostic outcomes.

Furthermore, our resulting accuracy of 82% is high, but could be insufficient in daily clinical practice since it still results in approximately 2 out of 10 patients receiving an incorrect diagnosis. The most important consideration is related to the number of patients receiving a wrong prediction (either a false-positive or a false-negative prediction. However, these computer algorithms should not be considered a substitute for the diagnostic process, but rather an aid to educate patients pre-hospital and organize outpatient clinic logistics.

In conclusion, ML algorithms are capable of making a clinical diagnosis for selected patients who suffer from hip complaints using online questionnaires. This first study yields an accuracy of 82% using outcome from our digital questionnaire only, which improved to 89% in combination with radiological osteoarthritis scores. Consultation of patients with complaints using ML techniques can therefore be considered as a valuable tool to aid the orthopedic surgeon in many practical ways, but should not yet be considered as a substitution for human made diagnosis.

Supplementary data The complete questionnaire and overview of ML algorithms are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2021.1884408

MS and WvdW conceived and designed the study, which was conducted with the help of DD and AvdM. TW and PvdP supervised the data and the ML analysis. MS drafted the manuscript. All authors contributed to its revision.

257

Acta thanks Mats Ericson and Bermd Grimm for help with peer review of this study.

Ashinsky B G, Coletta C E, Bouhrara M, Lukas V A, Boyle J M, Reiter D A, Neu C P, Goldberg I G, Spencer R G. Machine learning classification of OARSI-scored human articular cartilage using magnetic resonance imaging. Osteoarthritis Cartilage 2015; 23: 1704-12.

Codari M, Schiaffino S, Sardanelli F, Trimboli R M. Artificial intelligence for breast MRI in 2008–2018: a systematic mapping review. Am J Roentgenol 2019; 212: 280-92.

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 Groot I B, Reijman M, Terwee C B, Bierma-Zeinstra S M A, Favejee M, Roos E M, Verhaar J A. Validation of the Dutch version of the Hip disability and Osteoarthritis Outcome score. Osteoarthritis Cartilage 2007; 15(1): 104-9.

Dolatabadi E, Taati B, Mihailidis A. An automated classification of pathological gait using unobtrusive sensing technology. EEE Trans Neural Syst Rehabil Eng 2017; 25: 2336-46.

Duffield S J, Ellis B M, Goodson N, Walker-Bone K, Conaghan P G, Margham T, Loftis T. The contribution of musculoskeletal disorders in multimorbidity: implications for practice and policy. Best Pract Res Clin Rheumatol 2017; 31: 129-44.

Fourcade A, Khonsari R H. Deep learning in medical image analysis: a third eye for doctors. J Stomatol Oral Maxillofac Surg 2019; 120: 279-88.

Kellgren J H, Lawrence J S. Radiological assessment of osteo-arthrosis. Ann Rheum Dis 1957; 16: 494-502.

Kotti M, Duffell L D, Faisal A A, McGregor A H. Detecting knee osteoarthritis and its discriminating parameters using random forests. Med Eng Phys 2017; 43: 19-29.

Murray D W, Fitzpatrick R, Rogers K, Pandit H, Beard D J, Carr A J, Dawson J. The us of the Oxford hip and knee scores. J Bone Joint Surg Br 2007; 89(8): 1010-14.

Nguyen A V, Blears E E, Ross E, Lall R R, Ortega-Barnett J. Machine learning applications for the differentiation of primary central nervous system lymphoma from glioblastoma on imaging: a systematic review and metaanalysis. Neurosurg Focus 2018; 45: E5.

Nirschl J J, Janowczyk A, Peyster E G, Frank R, Margulies K B, Feldman M D, Madabhushi A. A deep-learning classifier identifies patients with clinical heart failure using whole-slide images of H&E tissue. PLOS ONE 2018; 13: e0192726.

Salaffi F, Stancati A, Silvestri C A, Ciapetti A, Grassi W. Minimal clinically important changes in chronic musculoskeletal pain intensity measured on a numerical rating scale. Eur J Pain 2004; 8(4): 283-91 Timmers T, Janssen L, Pronk Y, van der Zwaard B C, Koëter S, van Oostveen D, de Boer S, Kremers K, Rutten S, Das D, van Geenen RC, Koenraadt K L, Kusters R, van der Weegen W. Assessing the efficacy of an educational smartphone or tablet app with subdivided and interactive content to increase patients’ medical knowledge: randomized controlled trial. JMIR Mhealth Uhealth 2018; 6: e10742.

Wang X, Peng Y, Lu L, Lu Z, Bagheri M, Summers R M. ChestX-ray8: hospital-scale chest X-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI; 2017. pp 3462-71.

258

Acta Orthopaedica 2021; 92 (3): 258–263

The rate of 2nd revision for shoulder arthroplasty as analyzed by the Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) David R J GILL 1, Richard S PAGE 2,3, Stephen E GRAVES 3, Sophia RAINBIRD 3, and Alesha HATTON 4 1 Orthopaedics

Central, Monash Avenue, Nedlands; 2 Barwon Centre of Orthopaedic Research and Education, Deakin University, Geelong; 3 Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR), Adelaide; 4 South Australia Health and Medical Research Institute (SAHMRI), Adelaide, South Australia, Australia Correspondence: dg@davidgillortho.com.au Submitted 2020-09-07. Accepted 2020-12-15.

Background and purpose — The increase in shoulder arthroplasty may lead to a burden of revision surgery. This study compared the rate of (2nd) revision following aseptic 1st revision shoulder arthroplasty, considering the type of primary, and the class and type of the revision.

Patients and methods — All aseptic 1st revisions of primary total reverse shoulder arthroplasty (rTSA group) and of primary total stemmed and stemless total shoulder arthroplasty (non-rTSA group) procedures reported to our national registry between April 2004 to December 2018 were included. The rate of 2nd revision was determined using Kaplan–Meier estimates and comparisons were made using Cox proportional hazards models.

Results — There was an increased risk of 2nd revision in the 1st month only for the rTSA group (n = 700) compared with the non-rTSA group (n = 991); hazard ratio (HR) = 4.8 (95% CI 2.2–9). The cumulative percentage of 2nd revisions (CPR) was 24% in the rTSA group and 20% in the non-rTSA group at 8 years. There was an increased risk of 2nd revision for the type (cup vs. head) HR = 2.2 (CI 1.2–4.2), but not class of revision for the rTSA group. Minor (> 3 months) vs. major class revision, and humeral revision vs. all other revision types were second revision risk factors for the nonrTSA group.

Interpretation — The CPR of revision shoulder arthroplasty was > 20% at 8 years and was influenced by the type of primary, the class, and the type of revision. The most common reasons for 2nd revision were instability/dislocation, loosening, and infection.

The volume of shoulder arthroplasty surgery is increasing, and with it the expectation of future revision surgery (Lübbeke et al. 2017, AOANJRR 2019). The survivorship for anatomic shoulder arthroplasty has been well documented (Singh et al.

2011, Page et al. 2014, 2018, Dillon et al. 2019). Previously, revising anatomic shoulder arthroplasty often necessitated bone grafting and re-cementing of the glenoid component.

This was associated with soft tissue failure and graft reabsorption (Scalise and Iannotti 2008). Reverse total shoulder arthroplasty (rTSA) offered an opportunity to solve some of these problems (Boileau et al. 2006). Early reports indicated a high complication rate, but satisfactory clinical outcomes (Melis et al. 2012).

The outcome of lower limb revision arthroplasty has demonstrated higher subsequent revision rates compared with known primary procedures (AOANJRR 2018). Similar analysis for shoulder arthroplasty has been more limited, concentrating on specific component revisions or technical descriptions (Cheung et al. 2008, Scalise and Iannotti 2008, Melis et al. 2012, Bonnevialle et al. 2013). These studies have been on small numbers and of short duration. A larger populationbased study enables new insights, with a larger number of revision arthroplasty procedures for analysis.

The Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) has a high enrolment rate. This allows the opportunity to accurately plot population-based shoulder arthroplasty outcomes based on a large volume of implants and across a broad range of diagnoses and surgeons.

This study determined the rate of (second) revision following aseptic first revision shoulder arthroplasty, taking into account the type of primary shoulder arthroplasty revised, and the class of revision undertaken.

© 2021 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 10.1080/17453674.2020.1871559

Acta Orthopaedica 2021; 92 (3): 258–263

Patients and methods The AOANJRR began data collection on September 1, 1999 and includes data on almost 100% of the hip and knee arthroplasty procedures performed in Australia since 2002. Data collection was expanded to include shoulder arthroplasty procedures in April 2004 and has documented 97.1% shoulder arthroplasty procedures Australia-wide since November 2007.

These data are externally validated against patient-level data provided by all Australian state and territory health departments. A sequential, multilevel matching process is used to identify any missing data, which are subsequently obtained by follow-up with the relevant hospital. Each month, in addition to internal validation and data quality checks, all primary procedures are linked to any subsequent revision involving the same patient, joint, and side. Data are also matched bi-annually to the Australian National Death Index data to identify patients who have died.

In this study, the 1st revisions of primary total reverse shoulder arthroplasty (rTSA group) and total stemmed and stemless (non-rTSA group) procedures performed between April 16, 2004 and December 31, 2018 were analyzed. Resurfacing arthroplasty and hemiarthroplasty were excluded. Revision was defined as removal/exchange or addition of a joint replacement implant. The type of shoulder arthroplasty was the description of the implant (e.g., rTSA or non-rTSA) and the type of revision was the specifically removed/exchanged component (e.g., glenoid articular insert). Revision was further categorized as minor or major. A minor revision involved an exchange of implant not fixed to bone. A major revision involved an exchange of a component with bone fixation on both the glenoid and humeral sides, whilst a partial major revision exchanged bone-fixed components on either the glenoid or humeral side exclusively. A humeral revision included either metaphyseal or both metaphyseal and diaphyseal components.

All 1st revisions for infection were excluded to remove the confounding effect on subsequent revisions. The outcome measure was time to 2nd revision for all diagnoses including infection to capture all subsequent revision endpoints.

Statistics Kaplan–Meier estimates of survivorship were used to report the time to second revision, with censoring at the time of death and closure of the dataset at the end of December 2018.

The unadjusted cumulative percentage revisions (CPRs), with 95% confidence intervals (CI), were calculated using unadjusted point-wise Greenwood estimates. The CPR is displayed until the number at risk for the group reaches 40, unless the initial number for the group is less than 100, in which case the cumulative percentage revision is reported until 10% of the initial number at risk remains. Age- and sexadjusted hazard ratios (HR) were calculated from Cox proportional hazard models to compare the rate of second revi-

259

sion between groups. The assumption of proportional hazards was checked analytically for each model. If the interaction between the predictor and the log of time was statistically significant in the standard Cox model, then a time-varying model was estimated. Time points were selected based on the greatest change in hazard, weighted by a function of events.

Time points were iteratively chosen until the assumption of proportionality was met and HRs were calculated for each selected time-period. For the current study, if no time-period was specified, the HR was calculated over the entire followup period. All tests were 2-tailed at 5% levels of significance.

Statistical analysis was performed using SAS software version 9.4 (SAS Institute Inc., Cary, NC, USA).

Ethics, funding, and potential conflicts of interest The AOANJRR is approved by the Australian Federal Government as a federal quality assurance activity under Section 124X of the Australian Federal Health Insurance Act 1973.

All investigations were conducted in accordance with the ethical principles of research (The Helsinki Declaration II).

The AOANJRR is funded by the Commonwealth of Australia Department of Health and Ageing. The data of the AOANJRR is the intellectual property of the Australian Orthopaedic Association. The authors declare no financial disclosures.

Results There were 1,845 1st revisions of primary total shoulder arthroplasty, with 154 excluded as septic 1st revisions during the study period. Of those remaining, 41% (n = 700) were in the rTSA group, and 59% (n = 991) were in the non-rTSA group. Among the rTSA group 96% (n = 258) remained reverse shoulder arthroplasty at revision and 83% (n = 710) of the non-rTSA group were converted to a reverse shoulder arthroplasty at revision. There were 429 minor revisions, 232 were partial major, and 39 were major in the rTSA group. In the non-rTSA group 128 were minor, 621 were partial major, and 242 were major. Amongst the rTSA group the revisions by type were 270 head/cup, head only, or cup only, 64 glenoid, 166 humeral, and 39 humeral/glenoid. In the non-rTSA group 123 humeral head/glenoid insert or humeral head only, 66 glenoid, 553 humeral, and 242 humeral/glenoid components were revised (Table 1).

For 1st revision shoulder arthroplasty, the mean age of patients in the rTSA group was 73 years (72 years for males and 73 years for females) and 52% of patients were female.

Patients in the non-rTSA group had a mean age of 67 years (65 years for males and 69 years for females) and 60% of patients were female (Table 1).

2nd revisions were undertaken in 18% (n = 129) of the rTSA group and in 16% (n = 157) of the non-rTSA group (Table 1).

In the rTSA group, the CPR at 1 year was 13% and at 8 years it was 24%. In the non-rTSA group, the 1-year CPR was 9% and

260

Acta Orthopaedica 2021; 92 (3): 258–263

Table 1. Characteristics of rTSA group and non-rTSA group (all diagnoses, excluding 1st revision for infection)

Variable

Non-rTSA group rTSA group 2nd revisions 1st revisions 2nd revisions 1st revisions n = 157 (15.8%) n = 991 n = 129 (18.4%) n = 700 n (% of 1 st rev.) n (%) n (% of 1 st rev.) n (%)

Table 2. Yearly cumulative percentage of second revision (CPR) of first revision groups rTSA and non-rTSA (all diagnoses, excluding 1st revision for infection)

CPR (CI) Non-rTSA after years group rTSA group

Female sex 594 (59.9) 367 (52.4) 1 9.4 (7.7–11) 13 (11–16) Age 2 14 (12–16) 16 (13–19) < 55 14 (18.7) 75 (7.6) 4 (19.0) 21 (3.0) 3 15 (13–18) 18 (15–21) 55–64 57 (20.2) 282 (28.5) 25 (27.2) 92 (13.1) 4 16 (14–19) 20 (17–24) 65–74 61 (15.0) 407 (41.1) 54 (19.4) 279 (39.9) 5 17 (14–20) 22 (19–26) ≥ 75 25 (11.0) 227 (22.9) 46 (14.9) 308 (44.0) 6 19 (16–22) 23 (19–27) a Primary diagnosis 7 19 (16–22) 23 (19–27) Fracture 1 (5.9) 17 (1.7) 28 (25.0) 112 (16.9) 8 20 (17–24) 24 (20–28) Osteoarthritis 142 (15.4) 923 (94.6) 54 (19.2) 281 (42.4) 9 20 (17–24) – Osteonecrosis 5 (31.3) 16 (1.6) 0 (0.0) 6 (0.9) Rheumatoid arthritis 2 (20.0) 10 (1.0) 3 (15.8) 19 (2.9) Rotator cuff arthropathy 3 (30.0) 10 (1.0) 36 (14.7) 245 (37.0) Class of 1st revision at 8 years it was 20% (Table 2). There was Minor 67 (52.3) 128 (12.9) 81 (18.9) 429 (61.3) a significantly higher rate of 2nd revisions Major partial 56 (9.0) 621 (62.7) 40 (17.2) 232 (33.1) Major total 34 (14.0) 242 (24.4) 8 (20.5) 39 (5.6) in the rTSA group for the 1st month only Type of 1st revision b (HR 4.8; CI 2.2–8.9) (Figure 1). After 1 c Head/insert revision 25 (78.1) 32 (3.3) 0 (0.0) 1 (0.2) month, there was no significant difference Head only revision c 41 (45.1) 91 (9.2) 13 (10.7) 121 (22.4) Cup only c 38 (25.7) 148 (27.5) in the risk of a 2nd revision.

Humeral/glenoid revision 34 (14.0) 242 (24.6) 8 (20.5) 39 (7.2) Analysis of the class of revision found Glenoid component revision 13 (19.7) 66 (6.7) 13 (20.3) 64 (11.9) no significant difference in the risk of a Humeral component revision 43 (7.8) 553 (56.2) 27 (16.3) 166 (30.8)

2nd revision for the rTSA group. However, for the non-rTSA group, there was an increased risk of a 2nd revision after 3 months for the minor class compared = humeral head with other classes of revision (p < 0.001) (Figure 1). The non-rTSA group also had an increased risk of a 2nd revision for total major compared with partial major classes of revision (HR 1.9; CI 1.2–2.9) (Figure 2).

The type of revision was a statistically significant risk of 2nd revision of cup only compared with head only for the rTSA group (entire period: HR 2.2; CI 1.2–4.2). Other types of revision were not 2nd revision risks. However, humeral head/glenoid insert (after 3 months), humeral head only, glenoid, and glenoid/humeral revision were risk factors for 2nd revision compared with humeral component revision for the non-rTSA group (Figure 3).

The distribution of patients from the rTSA group and the non-rTSA group at 2nd revision by 4 age categories is recorded in Table 1. There was an increased risk of a 2nd revision for patients aged 65–74 years in the rTSA group compared with the non-rTSA group (HR 1.5; CI 1.0–2.1). Patients aged 55–64 years had an increased risk of a 2nd revision in the first 2 weeks in the rTSA group compared with the non-rTSA group (HR 16; CI 2–134), with no significant difference after this time. There was no statistically significant difference in the risk of 2nd revision in any other age group category over the entire period of the study.

The most common diagnosis of primary shoulder arthroplasty that underwent aseptic 1st revision was osteoarthritis,

a Only the outcome of the 5 most common primary diagnoses have been included. b Only the outcome of the six most common types of 1st revision have been listed. c rTSA group head/insert = glenosphere/humeral insert, head only = glenosphere only,

cup only = humeral insert only. group head/insert = humeral head/glenoid insert, head only only, cup only = N/A.

c non-rTSA

Cumulative percent revision of 1st revision 40 1st revision of primary rTSA 1st revision of primary non-rTSA

35 30 25 20 15 10 5 0

Number at risk

0

1

2

3

4

5

6

7

8

9

10

Years since 1st revision

Follow-up years 0 1 2 3 4 5 6 7 8 9 10 11 12 1st revisions, n non-rTSA 991 846 673 541 435 328 237 148 90 56 26 5 1 rTSA 700 537 407 306 224 157 112 77 50 31 15 2 0

Figure 1. Cumulative percentage of 2nd revision of 1st revision groups rTSA and non-rTSA by type of primary (all diagnoses, excluding 1st revision for infection).

HR (CI)—adjusted for age and sex 1st revision of primary rTSA vs 1st revision of primary non-rTSA 0–1 month HR 4.8 (2.5–9.0) 1–6 months HR 1.1 (0.68–1.6) > 6 months HR 1.2 (0.82–1.6)

Acta Orthopaedica 2021; 92 (3): 258–263

261

Cumulative percent revision of 1st revision

Cumulative percent revision of 1st revision

70

100 Minor 1st revision of primary TSA Major partial 1st revision of primary TSA Major total 1st revision of primary TSA

60

1st revision: Head/insert Head only Humeral/glenoid Glenoid component Humeral component

90 80

50

70 60

40

50 30

40 30

20

20 10 0

10 0

1

2

3

4

5

6

7

8

9

0

10

Years since 1st revision

1

2

3

4

5

6

7

8

9

10

Years since 1st revision

Number at risk

Number at risk

Follow-up years 0 1 2 3 4 5 6 7 8 9 10 11 12 Non-rTSA group, n minor 128 90 66 53 41 31 22 17 10 5 3 1 0 major partial 621 562 471 399 327 247 180 111 68 45 19 4 1 major total 242 194 136 89 67 50 35 20 12 6 4 0 0

Figure 2. Cumulative percentage of second revision of first revision group non-rTSA by class of the first revision (all diagnoses, excluding first revision for infection).

HR (CI)—adjusted for age and sex Non-rTSA group minor revision vs. non-rTSA group major partial revision 0–3 months HR 2.1 (0.95–4.7) 3 months–2 years HR 9.7 (6.1–15) > 2 years HR 9.7 (4.6–20) Non-rTSA group minor revision vs. non-rTSA group major total revision 0–3 months HR 1.1 (0.50–2.6) 3 months–1.5 years HR 5.3 (3.1–9.0) 1.5 years–2 years HR 4.8 (1.8–13) > 2 years HR 5.2 (2.4–12), Non-rTSA group major total revision vs. non-rTSA group major partial revision Entire period HR 1.9 (1.2–2.9) Table 3. 2nd revision diagnosis of first revision groups rTSA and non-rTSA (all diagnoses excluding 1st revision for infection)

2nd revision diagnosis

0

Non-rTSA group 2nd rev. % of 1st n = 157 revisions n (%) n = 991

rTSA group 2nd rev. % of 1st n = 129 revisions n (%) n = 700

Instability/dislocation 57 (36.3) 5.8 70 (54.3) Loosening 31 (19.7) 3.1 20 (15.5) Infection 18 (11.5) 1.8 20 (15.5) Rotator cuff insufficiency 14 (8.9) 1.4 1 (0.8) Fracture 4 (2.5) 0.4 6 (4.7) Implant breakage glenoid insert 6 (3.8) 0.6 1 (0.8) Pain 5 (3.2) 0.5 1 (0.8) Dissociation 4 (2.5) 0.4 Implant breakage glenoid 2 (1.3) 0.2 Malposition 2 (1.3) 0.2 1 (0.8) Metal related pathology 2 (1.3) 0.2 Arthrofibrosis 1 (0.8) Implant breakage humeral 1 (0.6) 0.1 1 (0.8) Wear glenoid 1 (0.6) 0.1 Wear glenoid insert 1 (0.6) 0.1 Wear humeral cup 1 (0.6) 0.1 1 (0.8) Other 8 (5.1) 0.8 6 (4.7)

10.0 2.9 2.9 0.1 0.9 0.1 0.1 0.1 0.1 0.1 0.1 0.9

Follow-up years 0 1 2 3 4 5 6 7 8 9 10 11 12 Non-rTSA group, n humeral head/ glenoid insert 32 19 10 7 5 3 2 2 1 0 0 0 0 humeral head only 91 67 52 44 34 27 19 14 8 4 2 1 0 humeral/glenoid 242 194 136 89 67 50 35 20 12 6 4 0 0 glenoid comp.

66 58 49 46 39 28 21 16 12 10 7 1 1 humeral comp. 553 502 421 352 287 218 158 94 55 34 12 3 0

Figure 3. Cumulative percentage of second revision of first revision group non-rTSA by type of revision (all diagnoses, excluding revision for infection).

HR (CI)—adjusted for age and sex Humeral head/glenoid insert revision vs. humeral component revision 0–3 months HR = 3.2 (0.74–13) 3 months–1.5 years HR = 15 (7.7–29) > 1.5 years HR = 34 (16–72) Humeral head only revision vs. humeral component revision Entire period HR = 6.8 (4.4–10) Humeral/glenoid revision vs. humeral component revision Entire period HR = 2.2 (1.4–3.4) Glenoid component revision vs. humeral component revision Entire period HR = 2.5 (1.3–4.6)

with 54 of 281 in the rTSA group and 142 of 923 in the nonrTSA group undergoing a 2nd revision (Table 1). Primary shoulder arthroplasty undertaken for osteoarthritis had a higher risk of a 2nd revision in the rTSA group compared with the non-rTSA group (HR 1.5; 1.1–2.1). There was insufficient data to determine whether the diagnoses of fracture, osteonecrosis, rheumatoid arthritis, and rotator cuff arthropathy were risk factors for a 2nd revision.

The 3 most common reasons for 2nd revision were similar for both groups of 1st revision shoulder arthroplasty, but the percentages differed. The most common reason for 2nd revision was instability/dislocation and was highest for the rTSA group (54%) compared with the non-rTSA group (36%). There were equal numbers of revisions for loosening (16%) and infection (16%) for the rTSA group. For the non-rTSA group, 2nd revision for loosening was higher (20%) but for infection it was lower (12%) compared with the rTSA group. Collectively, the diagnoses of instability/dislocation, loosening, and infection accounted for 85% of 2nd revision procedures in the rTSA group, and 68% in the non-rTSA group (Table 3). During the 1st month after revision the most common reason for a 2nd

262

revision for both the rTSA group (80%) and the non-rTSA group (62%) was instability/dislocation.

Discussion The AOANJRR (2018) reported the CPR of primary total shoulder arthroplasty. By comparison, this study found the CPR of first revisions of primary rTSA and non-rTSA to be 2 to 3 times higher at 12 months, and then at year 8 greater than 20%. We are not aware of any large patient study that previously indicated such an increase in CPR for revision shoulder arthroplasty. Others have also reported the outcome of primary TSA and rTSA from the perspective of survivorship (Boileau et al. 2006, Singh et al. 2011, Craig et al. 2019). This study has indicated that, for both rTSA and non-rTSA, the survivorship of the primary was substantially higher than their comparable first revisions (AOANJRR 2018) despite our exclusion of septic 1st revisions. We suggest that studies such as ours serve as a baseline that future revision methods look to and improve upon.

Craig et al. (2019) examined revisions and reoperations in a population-based study of 58,054 primary shoulder arthroplasty procedures and reported similar demographics, but also a high number of non-implant reoperations for primary shoulder arthroplasty. Our study highlights the high 2nd revision rate of shoulder arthroplasty looking exclusively at implant exchange. Both studies confirm the increased operative burden that occurs with shoulder arthroplasty.

There was a statistically significant increase in risk of second revision in the 1st month after the revised primary rTSA over non-rTSA. Our study showed the most common reason for 2nd revision in that period was dislocation/instability. We did not identify specific early (under 1 month) risk factors of second revision on comparing the 2 cohort groups by class or type of revision. It has been previously reported that revision of rTSA is difficult, with a complication rate ranging from 11% to 36% (Sajadi et al. 2010, Austin et al. 2011, Kelly et al. 2012, Patel et al. 2012, Wagner et al. 2018). We confirm that, and the temporal nature of those difficulties. It is likely that the high conversion to reverse shoulder arthroplasty of primary non-rTSA led to the similar CPR results overall after 1 month in this study for the 2 cohort groups.

Our study indicates there would not be confidence in undertaking minor procedures of first revision non-rTSA. Dines et al.

(2006) reviewed the outcome of 78 shoulders that underwent revision. They reported component revision was superior to soft tissue revision surgery on the basis of clinical outcome. Their study concluded that glenoid revision had a lower rate of revision compared with other revision procedures. Our results suggest that humeral cup revision has an increased 2nd revision risk.

Whilst this may be confounded by diagnosis, there are other minor class revisions that are not 2nd revision risks, such as head revision. Given instability is a common reason for primary rTSA revision, we believe our study is consistent with the findings of

Acta Orthopaedica 2021; 92 (3): 258–263

Cheung et al. (2018). Cil et al. (2009) reviewed the outcome of 38 humeral revisions for aseptic loosening for total shoulder arthroplasty and hemiarthroplasty. This type of revision surgery gave reliable pain relief but at a high risk of intraoperative complication, with 89% survivorship achieved at 10 years.

There were no clear advantages between differing classes of 1st revision rTSA in our study. Boileau et al. (2013) reported 37 patients who underwent revision rTSA with 11 patients requiring 2nd revision surgery. Overall, 32 retained an rTSA at follow-up at 34 months. Black et al. (2015) reported on 16 patients who underwent component revision after rTSA with a mean follow-up at 5 years. The study reported 6 of the 16 patients underwent further surgery and 9 sustained major complications. There was an improvement in pain and functional scores in their series.

Little is known about the 2nd revision outcome of shoulder arthroplasty. Zumstein et al. (2011) reported a 2nd revision rate of 16% for reverse replacements in a meta-analysis. In the study by Black et al. (2015), of the 6 rTSA revised for instability 2 subsequently required resection arthroplasty and of the 7 revised for glenoid baseplate failure 1 underwent resection arthroplasty. Our rates of 2nd revision were consistent with the combined series by Zumstein et al. (2011).

The diagnosis of osteoarthritis (when controlling for age and sex) was a risk factor for a 2nd revision of shoulder arthroplasty in our study. Singh et al. (2011) examined date of revision for total shoulder arthroplasty for 2,588 shoulders from 1976 to 2008. They found that male sex and rotator cuff disease were independent risk factors for revision of total shoulder arthroplasty. We were only able to confirm younger age as a risk factor for a 2nd revision in the first 2 weeks after the primary shoulder arthroplasty was revised.

The indications for 2nd revisions of rTSA and non-rTSA reflect similar problems to those that occur with primary TSA and primary rTSA. Kelly et al. (2012) observed for reverse total shoulder failed arthroplasty that the indications for the procedure included cuff deficiency, glenoid bone loss, humeral bone loss, or instability of the primary arthroplasty. The Norwegian Registry study of rTSA revisions found infection and loosening to be the most common diagnoses at revision surgery (Fevang et al. 2009). Khan et al. (2009) reported that rotator cuff tears, especially in those with rheumatoid arthritis, were the most common reason for cemented total shoulder arthroplasty revisions.

There are limitations to this study. The differing methods of enrolment and reporting of statistical outcomes of joint registries limit the ability of comparison. The recorded diagnosis is based on a categorical hierarchy and only the primary diagnosis has been used in this study so there may be additional reasons for the revision contributing to revision. This study excluded all primary rTSA and non-rTSA revised for infection. The aseptic study population gave guidance regarding relative implant performance of the 2 cohort groups. Infection of primary shoulder arthroplasty may lead to more complex

Acta Orthopaedica 2021; 92 (3): 258–263

revisions, a staged response with an outcome more related to the infection than the device. To ensure capture of all revision endpoints, 2nd revision for infection was included. The exclusion of 1st revision for infection, and those 1st revisions subsequently diagnosed as infective managed by antibiotic suppression, may confound infection as reason for 2nd revision. We examined survivorship outcomes measured on a population basis without seeking to select out variations due to implant type, patient selection, surgeons, or geography.

This does not represent, necessarily, the optimum achievable results for any particular implant or procedure. This registry data did not include clinical recording such as VAS scores, strength, or range of motion. Similar survivorship may not necessarily imply similar outcomes for pain and function.

Conclusions A revised total shoulder arthroplasty has a greater than 20% risk of 2nd revision at 8 years. The type of primary was not a risk factor for a 2nd revision, except in the 1st month, where an rTSA was an increased risk for a 2nd revision compared with a non-rTSA. The risk of a 2nd revision was affected by type and class of the revision, patient age, and the primary diagnosis osteoarthritis when adjusted for sex. The most common reasons for undertaking a 2nd revision were instability/dislocation, loosening, and infection.

The authors would like to thank the AOANJRR staff, orthopedic surgeons, hospitals, and patients whose data made this work possible.

DRJG, RSP, and SEG were involved in study design. SR, AH, and DRJG supervised the study. SR and AH undertook the statistical assessment, and all authors were involved in the preparation and proofreading of the manuscript.

Acta thanks Jeppe Vejlgaard Rasmussen and Björn Salomonsson for help with peer review of this study.

AOANJRR. Hip, knee & shoulder arthroplasty annual report; 2019. Available from https://aoanjrr.sahmri.com/annual-reports-2019. Accessed 2 June 2019.

AOANJRR. Hip, knee & shoulder arthroplasty annual report; 2018. Available from https://aoanjrr.sahmri.com/annual-reports-2018. Accessed 2 June 2019.

Austin L, Zmistowski B, Chang E S, et al. Is reverse shoulder arthroplasty a reasonable alternative for revision arthroplasty? Clin Orthop Relat Res 2011; 469(9): 2531-7. doi: 10.1007/s11999-010-1685-x.

Black E M, Roberts S M, Siegel E, et al. Failure after reverse total shoulder arthroplasty: what is the success of component revision? J Shoulder Elbow Surg 2015; 24(12): 1908-14. doi: 10.1016/j.jse.2015.05.029.

Boileau P, Watkinson D, Hatzidakis A M, et al. Neer Award 2005: The Grammont reverse shoulder prosthesis: results in cuff tear arthritis, fracture sequelae, and revision arthroplasty. J Shoulder Elbow Surg 2006; 15(5): 527-40. doi: 10.1016/j.jse.2006.01.003.

Boileau P, Melis B, Duperron D, et al. Revision surgery of reverse shoulder arthroplasty. J Shoulder Elbow Surg 2013; 22(10): 1359-70. doi:10.1016/j. jse.2013.02.004 Bonnevialle N, Melis B, Neyton L, et al. Aseptic glenoid loosening or failure in total shoulder arthroplasty: revision with glenoid reimplantation. J Shoulder Elbow Surg 2013; 22(6): 745-51. doi: 10.1016/j.jse.2012.08.009.

263

Cheung E V, Sperling J W, Cofield R H. Revision shoulder arthroplasty for glenoid component loosening. J Shoulder Elbow Surg 2008; 17(3): 371-5. doi: 10.1016/j.jse.2007.09.003.

Cheung E V, Sarkissian E J, Sox-Harris A, et al. Instability after reverse total shoulder arthroplasty. J Shoulder Elbow Surg 2018; 27(11): 1946-52. doi: 10.1016/j.jse.2018.04.015.

Cil A, Veillette C, Sanchez-Sotelo J, et al. Revision of the humeral component for aseptic loosening in arthroplasty of the shoulder. J Bone Joint Surgery Br 2009; 91(1): 75-81. doi: 10.1302/0301-620X.91B1.21094.

Craig R S, Lane J C, Carr A J, et al. Serious adverse events and lifetime risk of reoperation after elective shoulder replacement: population based cohort study using hospital episode statistics for England. Br Med J 2019; 364: l298. doi: 10.1136/bmj.l298.

Dillon M T, Page R S, Graves S E, et al. Early revision in anatomic total shoulder arthroplasty in osteoarthritis: a cross-registry comparison. Shoulder Elbow 2019: 1758573219842168. doi: 10.1177/1758573219842168.

Dines J S, Fealy S, Strauss E J, et al. Outcomes analysis of revision total shoulder replacement. J Bone Joint Surg 2006; 88(7): 1494-500. doi: 10.2106/JBJS.D.02946.

Fevang B-T S, Lie S A, Havelin L I, et al. Risk factors for revision after shoulder arthroplasty: 1,825 shoulder arthroplasties from the Norwegian Arthroplasty Register. Acta Orthop 2009; 80(1): 83-91. doi: 10.1080/17453670902805098.

Kelly J D, Zhao J X, Hobgood E R, et al. Clinical results of revision shoulder arthroplasty using the reverse prosthesis. J Shoulder Elbow Surg 2012; 21(11): 1516-25. doi: 10.1016/j.jse.2011.11.021.

Khan A, Bunker T, Kitson J. Clinical and radiological follow-up of the Aequalis third-generation cemented total shoulder replacement: a minimum ten-year study. J Bone Joint Surg Br 2009; 91(12): 1594-600. doi: 10.1302/0301-620X.91B12.22139.

Lübbeke A, Rees J L, Barea C, et al. International variation in shoulder arthroplasty: incidence, indication, type of procedure, and outcomes evaluation in 9 countries. Acta Orthop 2017; 88(6): 592-9. doi: 10.1080/17453674.2017.1368884.

Melis B, Bonnevialle N, Neyton L, et al. Glenoid loosening and failure in anatomical total shoulder arthroplasty: is revision with a reverse shoulder arthroplasty a reliable option? J Shoulder Elbow Surg 2012; 21(3): 342-9. doi: 10.1016/j.jse.2011.05.021.

Page R S, Navarro R A, Salomonsson B. Establishing an international shoulder arthroplasty consortium. J Shoulder Elbow Surg 2014; 23(8): 1081-2. doi: 10.1016/j.jse.2014.04.001.

Page R S, Pai V, Eng K, et al. Cementless versus cemented glenoid components in conventional total shoulder joint arthroplasty: analysis from the Australian Orthopaedic Association National Joint Replacement Registry. J Shoulder Elbow Surg 2018; 27(10): 1859-65. doi: 10.1016/j.jse.2018.03.017.

Patel D N, Young B, Onyekwelu I, et al. Reverse total shoulder arthroplasty for failed shoulder arthroplasty. J Shoulder Elbow Surg 2012; 21(11): 1478-83. doi: 10.1016/j.jse.2011.11.004.

Sajadi K R, Kwon Y W, Zuckerman J D. Revision shoulder arthroplasty: an analysis of indications and outcomes. J Shoulder Elbow Surg 2010; 19(2): 308-13. doi: 10.1016/j.jse.2009.05.016.

Scalise J J, Iannotti J P. Bone grafting severe glenoid defects in revision shoulder arthroplasty. Clin Orthop Relat Res 2008; 466(1): 139-45. doi: 10.1007/s11999-007-0065-7.

Singh J A, Sperling J W, Cofield R H. Revision surgery following total shoulder arthroplasty: analysis of 2588 shoulders over three decades (1976 to 2008). J Bone Joint Surg Br 2011; 93(11): 1513-7. doi: 10.1302/0301620X.93B11.26938.

Wagner E, Hevesi M, Houdek M, et al. Can a reverse shoulder arthroplasty be used to revise a failed primary reverse shoulder arthroplasty? Revision reverse shoulder arthroplasty for failed reverse prosthesis. Bone Joint J 2018; 100(11): 1493-8. doi: 10.1302/0301-620X.100B11.BJJ-2018-0226.R2.

Zumstein M A, Pinedo M, Old J, et al. Problems, complications, reoperations, and revisions in reverse total shoulder arthroplasty: a systematic review. J Shoulder Elbow Surg 2011; 20(1): 146-57. doi: 10.1016/j. jse.2010.08.001.

264

Acta Orthopaedica 2021; 92 (3): 264–268

Reoperations after decompression with or without fusion for L4–5 spinal stenosis with or without degenerative spondylolisthesis: a study of 6,532 patients in Swespine, the national Swedish spine register Anders JOELSON, Fredrik NERELIUS, Marek HOLY, and Freyr Gauti SIGMUNDSSON

Department of Orthopedics, Örebro University School of Medical Sciences and Örebro University Hospital, Örebro, Sweden Correspondence: anders@joelson.se Submitted 2020-09-19. Accepted 2020-11-16.

Background and purpose — There are different opinions on how to surgically address lumbar spinal stenosis with concomitant degenerative spondylolisthesis (DS). We investigated reoperation rates at the index and adjacent levels after L4–5 fusion surgery in a large cohort of unselected patients registered in Swespine, the national Swedish spine register.

Patients and methods — 6,532 patients, who underwent surgery for L4–5 spinal stenosis with or without DS between 2007 and 2012, were followed up to 2017 to identify reoperations at the index and adjacent levels. The reoperation rates for decompression and fusion were compared with the reoperation rates for decompression only and for patients with or without DS. Patient-reported outcome data were collected preoperatively, and at 1 and 2 years after surgery and used to evaluate differences in outcome between index operations and reoperations.

Results — For spinal stenosis with DS, the reoperation rate at the index level was 3.0% for decompression and fusion and 6.0% for decompression only. At the adjacent level, the corresponding numbers were 9.7% and 4.2% respectively.

For spinal stenosis without DS, the reoperation rate at the index level was 3.7% for decompression and fusion and 6.2% after decompression only. At the adjacent level, the corresponding numbers were 8.1% and 3.8% respectively.

For the reoperations at the adjacent level, there was no difference in patient-reported outcome between extended fusion or decompression only.

Interpretation — Single-level lumbar fusion surgery is associated with an increased rate of reoperations at the adjacent level compared with decompression only. When reoperations at the index level are included there is no difference in reoperation rates between fusion and decompression only.

There is an ongoing controversy on how to surgically address lumbar spinal stenosis with and without concomitant degenerative spondylolisthesis (DS). The debate has focused on the durability of the index operation versus accelerated symptomatic degeneration at the adjacent segments. In this debate, durability has been defined as maintenance of clinical benefit without the need for additional intervention (Ghogawala et al.

2017). While decompression without fusion may increase the risk for early index-level reoperations for instability (Ghogawala et al. 2016, Urakawa et al. 2020), fusion surgery might increase the risk of adjacent segment disease (ASD) (Sears et al. 2011, Okuda et al. 2018).

The national Swedish spine register (Swespine) covers 90% of the spine units in Sweden and the follow-up rate is 75–80% (Strömqvist et al. 2013). The Swespine offers possibilities to examine outcome in a large dataset of patients operated on for isolated L4–5 disease, which is the most common clinical scenario within spine surgery. Large register studies may contribute to increased evidence in this area.

Using Swespine data, we investigated reoperation rates at the index and the adjacent levels after L4–5 decompression only or decompression and fusion for spinal stenosis with and without concomitant DS.

Patients and methods Study design National Register study with prospectively collected data.

Patients The Swespine is a national quality register administered by the Swedish Association of Spine Surgeons. Demographic information is registered and follow-up forms are mailed to patients with a prepaid return envelope. Participation is volun-

© 2021 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 10.1080/17453674.2021.1879505

Acta Orthopaedica 2021; 92 (3): 264–268

265

Table 1. Baseline demographics by procedure after L4–5 decompression and fusion or decompression only

Parameter Mean age (SD) BMI (SD) Women, n (%) Years of follow-up (SD)

Stenosis with DS Stenosis without DS Decompression Decompression Decompression Decompression and fusion only and fusion only n = 1,338 n = 597 p-value n = 481 n = 4,116 p-value 65 (9.1) 27 (4.5) 1,027 (77) 7.8 (1.6)

69 (9.9) 27 (4.1) 407 (68) 7.9 (1.8)

tary for both the patients and the participating clinics and can be withdrawn at any time (Strömqvist et al. 2013).

We identified 6,584 patients who underwent surgery for lumbar L4–5 spinal stenosis with or without DS between 2007 and 2012. 52 patients were subject to acute reoperations (e.g., due to postoperative hematomas) and were excluded. The remaining 6,532 patients were followed up between 2007 and 2017 to identify reoperations. The diagnosis spinal stenosis was reported by the operating surgeon. DS was defined as slip > 3 mm on preoperative radiographs. Only patients who underwent surgery at the L4–5 level were included in the study.

Reoperation rates In Swespine, index operations and reoperations are registered separately. In order to identify reoperations that were incorrectly classified as index operations, we scanned the register for patients who had multiple operations. Reoperations because of spinal stenosis with or without concomitant DS or disk herniation at the index level (L4–5), the 1st cranial adjacent level (L3–4), the 2nd cranial adjacent level (L2–3), and the 1st caudal adjacent level (L5–S1) were counted. The diagnoses spinal stenosis and disk herniation were reported by the operating surgeon. Only the 1st reoperation was counted.

Reoperations at the index level (L4–5) because of implant failure were counted separately.

Patient-reported outcome measures Patient-reported outcome measures (PROMs) were recorded preoperatively, and at 1 and 2 years after surgery. We evaluated pain, disability, and health-related quality of life. Numeric rating scales (NRS) were used to assess leg and back pain (range 0–10, 0 being the best). The Oswestry disability index version 2.0 (ODI) (Fairbank and Pynsent 2000) was used to assess disability (range 0–100, 0 being the best and 100 the worst). The EQ5D index (UK tariff) (EuroQol Group 1990, Dolan 1997) was used to assess health-related quality of life (range –0.59 to 1, –0.59 being the worst and 1 the best).

Statistics The results are presented as mean (SD). Student’s t-test for unpaired data was used to compare normally distributed data.

The Mann–Whitney U-test was used to compare unpaired

< 0.01 0.02 < 0.01 0.4

60 (10.6) 27 (4.1) 266 (55) 7.7 (1.6)

65 (11.4) 27 (4.1) 2,097 (51) 7.8 (1.7)

< 0.01 0.3 0.07 0.5

non-normally distributed data. The chi-square test was used to compare frequencies. A p-value of < 0.05 was considered significant, and 2-tailed tests were used. The 95% confidence interval (CI) for relative risks is calculated as described by Altman (1991).

Ethics, data sharing, funding, and potential conflicts of interest The study was approved by the regional ethical review board (registration number: 2020-01505). Data are available from the national Swedish spine register (Swespine) after approval by a Swedish regional ethical review board and approval by the Swespine board. There was no external source of funding for this study. The authors declare no conflicts of interest.

Results Baseline (Table 1) There were no statistically significant differences in BMI and mean duration of follow-up but there were some minor differences in age and sex distributions.

Reoperation rates (Figure 1, Table 2) For spinal stenosis with DS, the reoperation rate at the index level (L4–5) was 3.0% for decompression and fusion and 6.0% for decompression only. At the 1st cranial adjacent level (L3–4), the corresponding numbers were 9.7% and 4.2% respectively. For spinal stenosis without DS, the reoperation rate at the index level (L4–5) was 3.7% for decompression and fusion and 6.2% after decompression only. At the 1st cranial adjacent level (L3–4), the corresponding numbers were 8.1% and 3.8% respectively. There were in total 38 reoperations because of implant related problems (loosening 4, breakage 1, pain because of implant 24, and pseudarthrosis 9).

Risk factors for additional surgery (Table 3) There was a minor age difference but no statistically significant difference in BMI, sex, or fusion method for patients with DS who underwent additional surgery at the adjacent level after decompression and fusion compared with the patients that required no additional surgery at the adjacent level.

266

Acta Orthopaedica 2021; 92 (3): 264–268

Table 2. Results by procedure after L4–5 decompression and fusion or decompression only at the index level (L4–5), 1st cranial adjacent level (L3–4), 2nd cranial adjacent level (L2–3), and 1st caudal adjacent level (L5–S1)

Reoperations, n (%)

Stenosis with DS Stenosis without DS Decompression Decom- Decompression DecomCount – index level fused Count – adjacent level fused and fusion pression only and fusion pression15 only 15 n = 1,338 n = 597 RR (95% CI) n = 481 n = 4,116 RR (95% CI)

Index level excluding implant failure including implant failure Adjacent level 1st cranial (L3–4) 2nd cranial (L2–3) 1st caudal (L5–S1) Total

11 (0.8) 40 (3.0)

36 (6.0) 36 (6.0)

0.1 (0.07–0.3) 10 0.5 (0.3–0.8)

9 (1.9) 18 (3.7)

255 (6.2) 10 255 (6.2)

0.3 (0.2–0.6) 0.6 (0.4–1.0)

130 (9.7) 39 (2.9) 7 (0.5) 216 (16.1)

25 (4.2) 10 (1.7) 9 (1.5) 80 (13.4)

2.3 (1.5–3.5) 1.7 (0.9–3.5) 5 0.3 (0.1–0.9) 1.2 (0.9–1.5)

39 (8.1) 7 (1.5) 5 (1.0) 69 (14.3)

155 (3.8) 57 (1.4) 5 73 (1.8) 540 (13.1)

2.2 (1.5–3.0) 1.1 (0.5–2.3) 0.6 (0.2–1.4) 1.1 (0.9–1.4)

0

2

4

6

8

10

0

2

4

6

Count – index level fused

Count – adjacent level fused

Count – index level not fused

Count – adjacent level not fused

15

15

15

15

10

10

10

10

5

5

5

5

0

2

4

6

8

0

10

2

4

6

Count – index level not fused

Count – adjacent level not fused

10

10

8

10

0

2

4

6

8

10

0

2

4

6

8

10

8

10

Years to reoperation

Figure 1. Histogram of time to reoperation at the index level (L4–5) and adjacent level (L3–4) for patients with spinal stenosis and degenerative 15 15 spondylolisthesis after decompression and fusion or decompression only. ▲ indicates the mean values.

5

Table 3. Characteristics of patients with degenerative spondylolisthesis who underwent additional surgery at the adjacent level after L4–5 decompression and fusion compared 5 with patients who required no additional surgery at the adjacent level

0

2 4

Parameter

6

Reoperation at 1st adjacent level (L3–4) 8 10 0 2 4 6 8 10 Yes No Years to reoperation n = 130 n = 1,208 p-value

Mean age (SD) 63 (8.9) 65 (9.5) BMI (SD) 27 (4.0) 27 (4.6) Women, n (%) 103 (80) 924 (76.5) Fusion method, n (%) PLF 112 (86) 1101 (91.1) PLIF 8 (6) 47 (3.9) TLIF 10 (8) 60 (5.0)

< 0.01 0.8 0.5 0.2

Patient-reported outcome measures (Figures 2 and 3) The 1-year and 2-year response rates for PROMs were 73% and 65% respectively. For all PROMs, there was no statistically significant difference between decompression and fusion and decompression only.

Discussion Adjacent segment disease after lumbar spinal fusion surgery has been a topic of concern for many years. The randomized controlled trials (RCTs) of Försth et al. (2016) and Ghogawala et al. (2016) did little to settle the controversy as the results were divergent. We report reoperation rates at adjacent segments for the largest cohort of patients who underwent surgery for lumbar L4–5 spinal stenosis.

We found a statistically significant lower risk for reoperation at the index level in the fusion group compared with the decompression group. When the reoperations at index level were counted together with the reoperations at the adjacent level there were no statistically significant differences in reoperation rates between decompression and fusion and decompression only. Also, Försth et al. (2016) and Ghogawala et al.

(2016) found that reoperations at the index level were more common after decompression only while reoperations at adjacent levels were more common after fusion. Our results confirm the results of Radcliff et al. (2013) that lumbar fusion is not associated with increased reoperation rate compared with decompression only.

8

6

decompression only

8 ●

6

4

●

4 ●

●

●

2

●

2

Acta Orthopaedica 2021; 92 (3): 264–268 0

Year 0

Year 1

Year 2

0

267

Year 0

Year 1

Year 2

Mean ODI

Mean EQ5D index

Mean NRS leg

Mean NRS back

100

1.0

10

10

80

0.8

8

8

Mean NRS leg

Mean NRS back

10 60 8 40

10 0.6 ●

8 0.4

●

●

6 20 4 0

●

Year 0

Index operation with ● decompression and fusion decompression only

Year 1

●

● ●

● ● Reoperation at adjacent level with ● decompression and fusion decompression only

4 0

●

6

4

●

Year 1

Year 2

0 Year 0

●

●

2 ●

●

4 ●

6 0.2

●

Year 2

6

●

2

Year 0

Year 1

Year 2

0

Year 0

Year 1

Year 2

2 2 Figure 2. Patient-reported outcome measures for the index operations (L4–5) and degenerative spondyloMeanfor ODIpatients with L4–5 spinal stenosis Mean EQ5D index listhesis after decompression and fusion or decompression only. Non-responders at the 1- and 2-year follow-up 1.0 were not excluded. The error bars 100 0 represent 95% confidence intervals. 0 Year 0

Year 1

Year 2

Year 0

Year 1

Year 2

80

Mean ODI

Mean EQ5D index

100

1.0

10 60

80

0.8

8 40

0.8

Mean NRS leg

Mean NRS back 10 0.6

●

8 0.4

●

60 ●

40

● ●

●

0

Year 0

Year 1

Year 2

0

●

●

4 0

0.4

0.2

20

●

6 20

0.6

Year 1

Year 1

Year 2

0

● ●

6 0.2 ●

●

Year 0

Year 1

Year 2

Year 0

Year 1

Year 2

●

Year 2

4 0 2

2

●

Year 0

Year 0

●

● ● Reoperation at adjacent level with ● decompression and fusion decompression only

Year 0

Year 1

Year 2

0

Mean ODI Mean EQ5D index Figure 3. Patient-reported outcome measures for the reoperations at the 1st adjacent level (L3–4) with decompression and extended fusion (n = 41) or decompression only (n = 79) for patients who underwent decompression and fusion for L4–5 spinal stenosis 1.0 with degenerative spondylolis100 thesis. Non-responders at the 1- and 2-year follow-up were not excluded. The error bars represent 95% confidence intervals.

80

The overall reoperation rates reported by Sears et al. (2011) and Radcliff et al. (2013) were similar to ours. The RCT of Försth et al. (2016), however, reported higher reoperation rates but that study included multilevel procedures. Sears et al. (2011) found a correlation between the number of levels fused and reoperation rate, which might explain our somewhat lower reoperation rate. Moreover, we excluded 52 acute reoperations, which reduce our reoperation rate. Our lower reoperation rate may also be related to coverage issues of register data.

Our data suggests that the times to reoperation were shorter for reoperations at the index level compared with reoperations at the adjacent levels (Figure 1). A possible explanation is that reoperations at the index level may have several causes, e.g., insufficient decompression, instability, or implant failure, that require more urgent intervention while ASD is a more slowly progressing condition.

The reoperation rate at the 1st caudal level (L5–S1) was low and we agree with Maragkos et al. (2020) that surgeons should refrain from prophylactic procedures at the L5–S1 level when considering a posterior L4–L5 fusion.

For spinal stenosis with DS, the reoperation rate at the 1st adjacent level was 4% for decompression only (Table 2). We agree with Bydon et al. (2016) that this implies that increased rigidity across previously mobile segments introduced by

0.8

spinal fusion is not the only contributing factor in the develop0.6 60 ment of ASD. ● ● ● surgery for The improvements in PROMs after revision ● ● 0.4 40 ASD were similar to the minimal important change values reported

by Parai et al. (2020)0.2(Figure 3). Also Park et al.

20 ● (2004) reported that surgery for ASD had relatively modest 0 0 outcomes.

This is an important finding in the context of patient Year 0 Year 1 Year 2 Year 0 Year 1 Year 2 information and shared decision-making. Patients should be informed that the results of a reoperation might be inferior in comparison with the index operation.

We found no statistically significant difference in sex or BMI for patients who required additional surgery at the adjacent level compared with the patients who required no additional surgery. Several authors have reported that sex is not associated with the development of ASD (Lee et al. 2014, Heo et al. 2015, Ou et al. 2015). Contrary to our findings, Ou et al.

(2015) found an association between BMI and ASD development. That study included only 13 cases of ASD and used MRI and clinical findings (i.e., not reoperations) to define ASD. Sears et al. (2011), Lee et al. (2014) and Heo et al. (2015) found that increasing age was a risk factor for ASD requiring further surgery. In contrast, Radcliff et al. (2013) found no association between age and reoperation rate. Although statistically significant, the difference in mean age between the groups of our study was small. If age is an important factor in

268

the development of ASD, it might be more correct to assess biological age or frailty rather than chronological age. Furthermore, a possible confounding factor when studying age and surgery is that younger patients may be more likely to be selected for additional surgery than older patients. This might be one part of the explanation why the patients in the reoperation group of our study are younger than the patients who did not require additional surgery.

We could not find any statistically significant association between type of fusion and additional operations at the adjacent level. In contrast, Lee et al. (2014) and Heo et al. (2015) found associations between fusion types and development of ASD, but their findings were inconsistent. Lee et al. (2014) found that interbody fusion showed a higher incidence of ASD requiring surgery than posterolateral fusion while Heo et al.

(2015) reported that posterolateral fusion showed a lower survival rate compared with interbody fusion. The importance of type of surgery on the development of ASD remains to be established.

Our study has several limitations. We recognize the limitations of a register study, e.g., different implants, surgical techniques, postoperative regimen, and selection bias. In this register study, no radiographs were available for analysis. Furthermore, higher response rates for the PROMs would have strengthened our results but studies following non-responders have revealed similar results in terms of the PROMs (Solberg et al. 2011). Our primary outcome variable, the reoperation rate, was of course not affected by the missing PROMs. Moreover, to present our data in a simple and transparent way we did not adjust the reoperation rates and the PROMs for differences in baseline data by using advanced statistical methods, e.g., regression analysis or propensity score matching. Given the diverging results of previous studies (Sears et al. 2011, Radcliff et al. 2013) concerning age as a confounding factor in relation to ASD, we find unlikely that the minor age differences in our groups have any substantial impact on the results.

In conclusion, single-level lumbar fusion surgery at the L4–5 level is associated with an increased rate of reoperations at the proximal adjacent level compared with decompression only. However, when reoperations on the index level are included there is no difference in reoperation rates between fusion and decompression only. Addressing the index segment for durable outcome should be the primary objective of surgery. This can be achieved with decompression or decompression and fusion with similar risk for additional surgery in selected cases.

Study design: AJ, FGS. Analysis of data: AJ. Interpretation of data: AJ, FN, MH, FGS. Drafting the manuscript: AJ, FGS. Critically revising the manuscript: FN, MH, FGS. Acta thanks Christian Hellum and Anne Versteeg for help with peer review of this study.

Acta Orthopaedica 2021; 92 (3): 264–268

Altman D G. Practical statistics for medical research. London: Chapman and Hall; 1991.

Bydon M, Macki M, De la Garza-Ramos R, McGovern K, Sciubba D M, Wolinsky J P, Witham T F, Gokaslan Z L, Bydon A. Incidence of adjacent segment disease requiring reoperation after lumbar laminectomy without fusion: a study of 398 patients. Neurosurgery 2016; 78(2): 192-9.

Dolan P. Modeling valuations for EuroQol health states. Med Care 1997; 35(11): 1095-108.

EuroQol Group. EuroQol: a new facility for the measurement of healthrelated quality of life. Health Policy 1990; 16(3): 199-208.

Fairbank J C, Pynsent P B. The Oswestry disability index. Spine (Phila Pa 1976) 2000; 25(22): 2940-52.

Försth P, Ólafsson G, Carlsson T, Frost A, Borgström F, Fritzell P, Öhagen P, Michaëlsson K, Sandén B. A randomized, controlled trial of fusion surgery for lumbar spinal stenosis. N Engl J Med 2016; 374(15): 1413-23.

Ghogawala Z, Dziura J, Butler W E, Dai F, Terrin N, Magge S N, Coumans J V C E, Harrington J F, Amin-Hanjani S, Schwartz J S, Sonntag V K H, Barker F G 2nd, Benzel E C. Laminectomy plus fusion versus laminectomy alone for lumbar spondylolisthesis. N Engl J Med 2016; 374(15): 1424-34.

Ghogawala Z, Resnick D K, Glassman S D, Dziura J, Shaffrey C I, Mummaneni P V. Randomized controlled trials for degenerative lumbar spondylolisthesis: which patients benefit from lumbar fusion? J Neurosurg Spine 2017; 26(2): 260-6.

Heo Y, Park J H, Seong H Y, Lee Y S, Jeon S R, Rhim S C, Roh S W. Symptomatic adjacent segment degeneration at the L3–4 level after fusion surgery at the L4–5 level: evaluation of the risk factors and 10-year incidence.

Eur Spine J 2015; 24(11): 2474-80.

Lee J C, Kim Y, Soh J W, Shin B J. Risk factors of adjacent segment disease requiring surgery after lumbar spinal fusion: comparison of posterior lumbar interbody fusion and posterolateral fusion. Spine (Phila Pa 1976) 2014; 39(5): E339-45.

Maragkos G A, Atesok K, Papavassiliou E. Prognostic factors for adjacent segment disease after L4-L5 lumbar fusion. Neurosurgery 2020; 86(6): 835-42.

Okuda S, Nagamoto Y, Matsumoto T, Sugiura T, Takahashi Y, Iwasaki M. Adjacent segment disease after single segment posterior lumbar interbody fusion for degenerative spondylolisthesis: minimum 10 years follow-up.

Spine (Phila Pa 1976) 2018; 43(23): E1384-E1388.

Ou C Y, Lee T C, Lee T H, Huang Y H. Impact of body mass index on adjacent segment disease after lumbar fusion for degenerative spine disease. Neurosurgery 2015; 76(4): 396-401.

Parai C, Hägg O, Lind B, Brisby H. ISSLS prize in clinical science 2020: The reliability and interpretability of score change in lumbar spine research.

Eur Spine J 2020; 29(4): 663-9.

Park P, Garton H J, Gala V C, Hoff J T, McGillicuddy J E. Adjacent segment disease after lumbar or lumbosacral fusion: Review of the literature. Spine (Phila Pa 1976) 2004; 29(17): 1938-44.

Radcliff K, Curry P, Hilibrand A, Kepler C, Lurie J, Zhao W, Albert T J, Weinstein J. Risk for adjacent segment and same segment reoperation after surgery for lumbar stenosis: a subgroup analysis of the Spine Patient Outcomes Research Trial (SPORT). Spine (Phila Pa 1976) 2013; 38(7): 531-9.

Sears W R, Sergides I G, Kazemi N, Smith M, White G J, Osburg B. Incidence and prevalence of surgery at segments adjacent to a previous posterior lumbar arthrodesis. Spine J 2011; 11(1): 11-20.

Solberg T K, Sørlie A, Sjaavik K, Nygaard Ø P, Ingebrigtsen T. Would loss to follow-up bias the outcome evaluation of patients operated for degenerative disorders of the lumbar spine? Acta Orthop 2011; 82(1): 56-63.

Strömqvist B, Fritzell P, Hägg O, Jönsson B, Sandén B, Swedish Society of Spinal Surgeons. Swespine: the Swedish spine register: the 2012 report.

Eur Spine J 2013; 22(4): 953-74.

Urakawa H, Jones T, Samuel A, Vaishnav A S, Othman Y, Virk S, Katsuura Y, Iyer S, McAnany S, Albert T, Gang C H, Qureshi S A. The necessity and risk factors of subsequent fusion after decompression alone for lumbar spinal stenosis with lumbar spondylolisthesis: 5 years follow-up in 2 different large populations. Spine J 2020; 20(10): 1566-72.

Acta Orthopaedica 2021; 92 (3): 269–273

269

What is the association between MRI and conventional radiography in measuring femoral head migration?

Hans-Christen HUSUM 1,4, Michel Bach HELLFRITZSCH 2,4, Mads HENRIKSEN 2,4, Kirsten Skjærbæk DUCH 5, Martin GOTTLIEBSEN 3,4, and Ole RAHBEK 1,4 1

Interdisciplinary Orthopaedics, Aalborg University Hospital, Aalborg; 2 Department of Radiology, Aarhus University Hospital, Aarhus; 3 Department of Orthopedics, Aarhus University Hospital, Aarhus; 4 Danish Pediatric Orthopaedic Research; 5 Unit of Epidemiology and Biostatistics, Aalborg University Hospital, Aalborg, Denmark Correspondence: h.husum@rn.dk Submitted 2020-06-02. Accepted 2020-11-15.

Background and purpose — Pelvic radiographs are traditionally used for assessing femoral head migration in residual acetabular dysplasia (RAD). Knowledge of the heightened importance of cartilaginous structures in this condition has led to increased use of MRI in assessing both osseous and cartilaginous structures of the pediatric hip. Therefore, we assessed the relationship between migration percentages (MP) found on MRI and conventional radiographs. Second, we analyzed the reliability of MP in MRI and radiographs.

Patients and methods — We retrospectively identified 16 patients (mean age 5 years [2–8], 14 girls), examined for RAD during a period of 2½ years. 4 raters performed blinded repeated measurements of osseous migration percentage (MP) and cartilaginous migration percentage (CMP) in MRI and radiographs. Pelvic rotation and tilt indices were measured in radiographs. Bland–Altman (B–A) plots and intraclass correlation coefficients (ICC) were calculated for agreement and reliability.

Results — B–A plots for MPR and MPMRI produced a mean difference of 6.4 with limits of agreement –11 to 24, with higher disagreements at low average MP values. Mean MPR differed from mean MPMRI (17% versus 23%, p < 0.001). MPR had the best interrater reliability with an ICC of 0.92 (0.86–0.96), compared with MPMRI and CMP with ICC values of 0.61 (0.45–0.70) and 0.52 (0.26–0.69), respectively. Intrarater reliability for MPR, MPMRI and CMP all had ICC values above 0.75 and did not differ statistically significantly. Differences inMPMRI and MPR showed no correlation to pelvic rotation index, pelvic tilt index, or interval between radiograph and MRI exams.

Interpretation — Pelvic radiographs underestimated MP when compared with pelvic MRI. We propose CMP as a new imaging measurement, and conclude that it has good intrarater reliability but moderate interrater reliability. Measurement of MP in radiographs and MRI had mediocre to excellent reliability.

MRI scans have been used in the diagnosis and prognostication of developmental dysplasia of the hip (DDH) for the last 30 years (Bos et al. 1988). MRI scans permit the examiner complete control over orientation of the examined pelvis, allowing for more accurate measurements and visualization of non-bony structures. However, conventional pelvic radiographs are still the preferred method of examination for children over the age of 6 months due to financial considerations and challenges in scanning children, such as anxiety of the child and the need for sedation.

Residual acetabular dysplasia (RAD) occurs in 3.5–17% of treated cases of DDH (Tucci et al. 1991, Alexiev et al.

2006) and is a known risk factor for secondary osteoarthritis (Malvitz and Weinstein 1994). Indications for corrective surgery for this condition remain controversial, as seen in the tendency of surgeons to undertreat RAD patients who need surgery, rather than overtreat those who do not (Ömeroǧlu et al. 2012).

Many radiographic measurements have been proposed to indicate the severity and prognosis of DDH and many are used when deciding which patients should receive corrective surgery. The most commonly used are the osseous acetabular index (OAI) and acetabular head coverage (Ömeroǧlu et al.

2012). None of these measurements, on their own or in combination, have been shown to predict DDH prognosis accurately in all cases, and are therefore most commonly used in various combinations at the discretion of the surgeon. Acetabular head coverage, both cartilaginous and osseous, is of importance for the stability of the hip joint (Bos et al. 1991, Domenech et al.

2001) and is commonly estimated by the osseous migration percentage (MP), first proposed by Reimers (Reimers 1980), and has been shown to be predictive of later osteoarthritis (Terjesen 2011). MP estimates the percentage of the osseous femoral head that is covered by the osseous acetabulum. In this study we emulated the method developed by Reimers, in measuring the percentage of the cartilaginous femoral head

© 2021 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 10.1080/17453674.2020.1864124

270

Acta Orthopaedica 2021; 92 (3): 269–273

covered by the cartilaginous acetabulum, and propose the name: cartilaginous migration percentage (CMP).

Currently, there are no studies comparing reliability, agreement, and correlation of MP in radiographs and MRI or between MP and CMP. We compared agreement and correlation of MP measurements in pelvic radiographs and coronal MRI sequences. We assessed values and inter- and intrarater reliability when measuring CMP on pelvic MRI, and evaluated inter- and intrarater reliability when measuring MP on pelvic radiographs and MRI in hips evaluated for RAD corrective surgery.

Patients and methods This was a retrospective cohort study based on pelvic MRI and radiographs of a consecutive series of children examined for RAD at the Departments of Orthopedics and Radiology, Aarhus University Hospital (AUH), Aarhus, Denmark, during a 2½-year period from September 2016 to April 2019. Reporting follows STROBE and GRRAS statements. We included all children examined for RAD who had pelvic MRI scans done as a supplement to their previous pelvic radiographs.

Exclusion criteria were: unacceptable MRI (T1 sequence not obtained, movement artifacts, relevant structures for performing each measurement not visualized) or unacceptable radiograph (relevant structures for performing each measurement not visualized due to gonadal shielding).

The examiner group consisted of 2 senior pediatric orthopedic surgeons (OR and MG) and 2 senior musculoskeletal radiologists (MBH and MH). Each examiner had at least 7 years of experience in interpreting pediatric hip radiographs.

MBH had 10 years of pediatric musculoskeletal MRI experience, OR and MG each had 5 years of experience and, with only a few months’ worth, MH had the least pediatric MRI experience.

We performed measurements on T1-weighted pelvic MRI scans and the pelvic radiographs that led to the MRI scans.

All MRI scans were performed at AUH where 5 different scanners were used with similar settings (Philips Medical Systems, Best, Netherlands: Achieva dStream 3.0T, Ingenia 1.5T. GE Medical Systems, Milwaukee, USA: Optima MR 450w 1.5T. Siemens Medical System, Germany: Skyra 3T, Avanto fit 1.5T). Scans were archived and viewed using Picture Archiving and Communication Software (PACS) at AUH (Impax, client 6.5 AGFA Healthcare N.V., Mortsel, Belgium).

Coronal T1-weighted spin echo images were acquired. To determine the imaging sections, a transverse scout view of the acetabular region and symmetrical coronal sections was obtained. The slice thickness varied between 3 and 4 mm and the most central section was chosen. All scans were performed with the parents present and without sedation of the child. The child was placed in a supine position and had a body array coil placed anteriorly and posteriorly to the hip joint. Measure-

Figure 1. Coronal T1-weighted magnetic resonance imaging scan of pediatric hip included in this study. LOF: lateral edge of osseous femoral head, MOF: medial edge of osseous femoral head, LCF: lateral edge of cartilaginous femoral head, MCF: medial edge of cartilaginous femoral head, H: Hilgenreiner’s line, P: Perkin’s line, and CAR: vertical line through lateral edge of cartilaginous acetabular roof perpendicular to Hilgenreiner’s line. MPMRI = B/A×100%, CMP = C/A×100%

ments performed on the MRI scans and AP pelvic radiographs reported in this study were: Hilgenreiner’s line, Perkins’ line, MP and CMP (Hilgenreiner 1925, Perkins 1928, Reimers 1980).

To calculate CMP we measured the distance between the medial and lateral sides of the cartilaginous edge of the femoral head (A) as well as the distance between a vertical line through the most lateral aspect of the cartilaginous acetabular roof perpendicular to Hilgenreiner’s line and the lateral side of the cartilaginous edge of the femoral head (C). CMP was then calculated as: CMP = C/A x 100% (Figure 1).

A pre-study workshop was held to reach consensus on how measurements should be performed. 1 MRI scan was used for the workshop and was also included in the study. Measurements for the study began in the following weeks at the discretion of each rater. MRI sequences and image selection were noted; all data was stored in an encrypted standardized data sheet by each investigator. After a minimum of 1 week after initial rating, measurements were repeated in the same order by each investigator and stored in a secondary encrypted data sheet. Measurements were performed once before saving and raters were instructed not to edit saved measurements.

Pelvic tilt index and pelvic rotation index were based on measurements on pelvic radiographs made by 1 senior musculoskeletal radiologist (MBH) and carried out according to the specifications of Ball and Kommenda (1968) and Tönnis (1976).

Each investigator was blinded to other investigators’ measurements, their own previous measurements and all information on the patient except pelvic radiographs and MRI scans

Acta Orthopaedica 2021; 92 (3): 269–273

Table 1. Demographics of included patients Factor n Mean (SD) [range]

Time between radiographic and MRI examination 14 133 (72) [16–234] Age at radiography (years) 15 5.3 (1.6) [2.1–7.9] Age at MRI (years) 15 5.6 (1.5) [2.6–8.2] Radiographic pelvic tilt and rotation indices Pelvic tilt index 15 0.73 (0.17) [0.47–1.03] Pelvic rotation index 14 1.0 (0.14) [0.72–1.25] Female sex 14/16 MRI scanner Philips: Achieva dStream 3T 4 Philips: Ingenia 1.5T 2 GE: Optima MR450w 1.5T 3 Siemens: Skyra 3T 4 Siemens: Avanto Fit 1.5T 2 Previous treatment for DDH Hip brace (Denis Browne) 2 Closed reduction and hip spica 4 None 8 Unknown 2

to minimize information bias. All measurements and ratings were performed independently; raters were instructed not to communicate results in any way and were aware that they would be compared with each other.

Statistics When evaluating the mean for each of the 3 measurement methods, both right and left hip were included for all children.

Variation across individual and measured side were accounted for using a linear nested mixed model with patient ID and side as random effect, and measurement method as fixed effect, thus assuming the 4 raters were independent according to the study design. Model control was performed by investigation of the qq-plot of the model residuals.

Considerations for bilaterality were made in the ICC calculations by using 1,000 bootstrap samples for both interrater and intrarater ICC. Due to missing observations each bootstrap sample representing the total population varied in size from 13 to 16 patients. All ICC calculations were made using a 2-way, mixed-effect, single-rater ICC with absolute agreement and presented with the bootstrap mean and crude 95% confidence interval (CI). Interrater ICC was calculated using the first round of measurements with each bootstrap sample containing either all-right or all-left measurements for a specific patient. Intrarater ICC was calculated across rounds using a fixed randomly chosen rater in each bootstrap sample and a randomly chosen side for each patient.

ICC values were interpreted according to general guidelines where a value of 0 or less represents no reliability, 0.75 represents good reliability, and 1 represents complete reliability (Portney and Watkins 2012).

Spearman’s correlation coefficient was calculated among the absolute differences between radiographs and MRI and

271

Table 2. Mean values (%) of bilateral migration measurements made in 1 round by all raters Migration percentage Osseous, radiographic (MPR) Osseous, MRI (MPMRI) Cartilaginous, MRI (CMP)

Mean (95% CI) 17 (14–20) 23 (20–26) 19 (16–22)

pelvic tilt, pelvic rotation, and interval. Scatter plots were investigated for any systematic association not indicated by the p-values. The plots were constructed using the absolute difference of both left and right side for all individuals measured by 1 rater (MBH) in 1 round of rating.

A B–A plot for agreement between MPR and MPMRI was constructed under the assumption that the differences were independent, as measurements were made and subtracted within each patient.

No sample size calculation was made; the number of patients considered for surgery for RAD at AUH during the study period determined the sample size. Statistical analyses were performed using Stata version 16.1 (StataCorp, College Station, TX, USA).

Ethics, funding, and potential conflicts of interest Ethical approval was not required in accordance with the guidelines of the Danish National committee on health research ethics for non-interventional studies. No external funding was obtained for this study. No conflict of interest was declared.

Results 16 children (14 girls) were identified. 1 had missing pelvic radiographs and 1 child had incomplete MRI data, totaling pelvic radiographs and MRI scans of 30 hips. 8 hips were scanned in 3.0T scanners and 7 were scanned in 1.5T scanners, mean age at radiographic examination was 5 years (2–8), and mean interval between radiograph and MRI was 133 days (16–234). The ethnicity of included patients was: Caucasian (n = 14), Turkish (n = 1), and African (n = 1). 4 patients had musculoskeletal disorders which were: bilateral coxae vara, Calvé–Legg–Perthes disease with secondary acetabular dysplastic changes, and unilateral proximal femoral focal deficiency (Table 1).

Mean values (CI) for the first round of rating across all raters were: MPR 17 (14–20), MPMRI 23 (20–26), and CMP 19 (16–22). The mean value of CMP did not differ statistically significantly from the mean of MPR and MPMRI (Table 2).

The mean difference between MPMRI and MPR was 6.4 (–15 to 38). The mean absolute difference between MPMRI and MPR showed no statistically significant correlation to pelvic rotation index, pelvic tilt index, or interval between examina-

272

Acta Orthopaedica 2021; 92 (3): 269–273

Table 3. ICC values for inter- and intrarater reliability with 95% bootstrap confidence intervals, bootstrap samples 1,000

MPMRI – MPR (%) 40

Migration percentage

30

Interrater reliability Osseous, radiographic (MPR) Osseous, MRI (MPMRI) Cartilaginous, MRI (CMP) Intrarater reliability Osseous, radiographic (MPR) Osseous, MRI (MPMRI) Cartilaginous, MRI (CMP)

20

10

0

–10

–20 0

10

20

30

40

Mean of MPMRI and MPR (%)

Figure 2. Bland–Altman plot, agreement of MPR and MPMRI measurements made by all raters in 1 round. Mean difference 6.4, limits of agreement –11 to 24. MP = osseous migration percentage.

tions (Spearman’s rho 0.29, 0.35, and 0.44 respectively). Scatter plots revealed no systematic correlation.

The B–A plot for MPR and MPMRI produced a mean difference of 6.4 with limits of agreement –11 to 24, with higher disagreements at low average MP values (Figure 2).

The ICC value for MPR interrater reliability was statistically significantly higher at 0.92 (0.86–0.96) when compared with MPMRI and CMP, with ICC values of 0.61 (0.45–0.70) and 0.52 (0.26–0.69) respectively. No statistically significant difference was found in intrarater reliability ICC for MPR, MPMRI, and CMP with ICC values of 0.94, 0.79, and 0.82 respectively (Table 3).

Discussion In this first study, comparing agreement, reliability, and correlation of MP in radiographs and MRI in 16 children examined for RAD, we found disagreement in MP between MRI and radiograph modalities at low values, interestingly with no correlation to pelvic orientation or interval between radiographs and MRI. Intrarater reliability in radiographs and MRI modalities across 4 independent raters was excellent and interrater reliability for the novel CMP measurement was comparable to MPMRI, but inferior to radiographic MP. Limitations This study had some limitations. To limit selection bias, we consecutively included all patients considered for surgical intervention, at the same institution, during a period of 2½ years. However, our sample size for this rare condition was small, which translates to wide confidence intervals in our calculated means for CMP.

ICC (95% CI) 0.92 (0.86–0.96) 0.61 (0.47–0.70) 0.52 (0.26–0.69) 0.94 (0.81–0.99) 0.79 (0.50–0.93) 0.82 (0.41–0.97)

Acceptable ranges for pelvic tilt index and pelvic rotation index have been reported to be between 0.9–1.4 and 0.7–1.5 respectively (Yang et al. 2020). These values were exceeded for pelvic tilt index in over 80% of observations included in this study, whereas values of pelvic rotation where within acceptable limits. This reflects the clinical reality, but it has been shown to affect the accuracy of measurements on pelvic radiographs (Hamano et al. 2019), and in extension it could affect the agreement between MPR and MPMRI. However, we found no evidence for a correlation deviating from 0 between MPR–MPMRI and pelvic rotation or pelvic tilt.

We used 5 different MRI scanners in this study. Using higher strength scanners in musculoskeletal imaging reportedly improves image quality due to higher signal-to-noise ratios, which in turn could affect the correlation analysis of this study. However, large reviews have found insufficient evidence to link MRI scanners’ technical specifications to clinically meaningful outcomes (Wood et al. 2011). Pelvic radiographs and MRI scans were performed with a mean interval of 133 days (16–234). During this time the morphology of the examined hip could potentially have changed but we found no statistically significant correlation between the absolute mean difference, MPR–MPMRI, and the interval between radiographs and scans.

Prior to measurements, each rater participated in a workshop aimed at reaching a consensus in measurements. This was necessary as CMP is a novel measurement but reduces the independence of raters and could affect the interrater reliability and external validity of the study results.

Interpretation In 41 out of 235 observations, MPR was measured at values of 0% and 1%, whereas the corresponding MPMRI values ranged across the entire span of measurements (0–39%). This could indicate a lack of information when evaluating the pediatric hip on pelvic radiographs and shows a trend toward underestimation of MP compared with MRI. The high level of intrarater reliability and low level of interrater reliability found in MRI measurements in this study underlines the need for proper standardized measurement guidelines.

Acta Orthopaedica 2021; 92 (3): 269–273

It has long been known that the cartilaginous structures of the hip play an important part in the stability of the pediatric hip joint in DDH (Bos et al. 1988), and attempts have been made to quantify the cartilaginous acetabular head coverage.

Using cartilage-optimized computed tomography (CT) scans, Lin et al. (1997) found a mean difference in coverage percentages of 17%, but this method relied on fitting anatomical structures to best-fit computer-generated geometric shapes, as CT is not optimized for visualizing cartilaginous structures.

Domenech et al. (2001) proposed the acetabular head index (AHI) defined as the percentage ratio of the width of the cartilaginous femoral head covered by the cartilaginous acetabulum over the total width of the head, measured in both sagittal and coronal planes on MRI scans. The use of this parameter is not widespread, and no studies have been published on the reliability or agreement of this measurement.

Utilizing MRI, we emulated the commonly used measurement of MP developed by Reimers (1980), which is a proven prognosticator for the long-term risk for osteoarthritis (Terjesen 2011). CMP derives its measurements from the cartilaginous edge of the roof of the acetabulum, but how this measurement relates to the stability of the hip and the prognosis of DDH is currently unknown. Other measurements based on the cartilaginous edge of the acetabulum have been proposed, most notably the cartilaginous acetabular index (CAI), which is increased in DDH (Li et al. 2012) and has been proposed as a guiding measurement in the selection of patients for surgical treatment in borderline RAD cases (Merckaert et al. 2019).

This could mean a future role for CAI as predictor of the developmental potential of the hip and may show promise for measurements sharing the same anatomical landmarks such as the CMP. Generalizability Raters were aware that their measurements would be compared with those of their fellow raters. This awareness is subject to the Hawthorne effect, the change in subjects’ behavior due to their awareness of being observed (Chen et al. 2015), which could affect the external validity of the measurements.

Conclusion We compared MP measurements on MRI and pelvic radiographs and found significant disagreement, especially at low MP values. Intrarater reliability for MP and CMP across radiograph and MRI modalities was good to excellent, while interrater reliability for MRI measurements was poor. We have established the CMP as a novel MRI measurement for assessment of the pediatric hip with reliability comparable to existing MRI measurement techniques, although still inferior to the reliability of pelvic radiographs.

Future investigations into the role of CMP as a marker for the prognosis of RAD, and standardized guidelines for performing this measurement, are needed before it can be utilized in a clinical setting.

273

Conceptualization, HCH, MBH, MG, MH, and OR. Methodology, MBH, KD, MG, MH, and OR. Investigation, MBH, MG, MH, and OR. Writing— original draft: HCH. Writing—review and editing: HCH and OR. Supervision: OR. Acta thanks Gunnar Hägglund and Terje Terjesen for help with peer review of this study.

Alexiev V A, Harcke H T, Kumar S J. Residual dysplasia after successful Pavlik harness treatment: early ultrasound predictors. J Pediatr Orthop 2006; 26(1): 16-23.

Ball F, Kommenda K. [Sources of error in the roentgen evaluation of the hip in infancy]. Ann Radiol (Paris) 1968; 11(5): 298-303.

Bos C F, Bloem J L, Obermann W R, Rozing P M. Magnetic resonance imaging in congenital dislocation of the hip. J Bone Joint Surg Br 1988; 70(2): 174-8.

Bos C F, Bloem J L, Verbout A J. Magnetic resonance imaging in acetabular residual dysplasia. Clin Orthop Relat Res 1991; (265): 207-17.

Chen L F, Vander Weg M W, Hofmann D A, Reisinger H S. The Hawthorne effect in infection prevention and epidemiology. Infect Control Hosp Epidemiol 2015; 36(12): 1444-50.

Domenech B, Baunin C, Sales de Gauzy J, Cahuzac J P, Guitard J, Puget C, Leclet H, Railhac J J. Imaging of the hip in healthy children in MRI: evaluation of the acetabular coverage. J Radiol 2001; 82(12 Pt 1): 1711-8.

Hamano D, Yoshida K, Higuchi C, Otsuki D, Yoshikawa H, Sugamoto K. Evaluation of errors in measurements of infantile hip radiograph using digitally reconstructed radiograph from three-dimensional MRI. J Orthop 2019; 16(3): 302-6.

Hilgenreiner W H. The early diagnosis and early treatment of congenital dislocation of the hip. Med Klin 1925; 21: 1385-9, 1425-9.

Li L Y, Zhang L J, Li Q W, Zhao Q, Jia J Y, Huang T. Development of the osseous and cartilaginous acetabular index in normal children and those with developmental dysplasia of the hip: a cross-sectional study using MRI. J Bone Joint Surg 2012; 94.B(12): 1625-31.

Lin C J, Romanus B, Sutherland D H, Kaufman K, Campbell K, Wenger D R. Three-dimensional characteristics of cartilaginous and bony components of dysplastic hips in children: three-dimensional computed tomography quantitative analysis. J Pediatr Orthop 1997; 17(2): 152-7.

Malvitz T A, Weinstein S L. Closed reduction for congenital dysplasia of the hip. J Bone Joint Surg 1994; 76-A(12): 1777-92.

Merckaert S R, Pierzchala K, Bregou A, Zambelli P Y. Residual hip dysplasia in children: osseous and cartilaginous acetabular angles to guide further treatment-a pilot study. J Orthop Surg Res 2019; 14(1): 379. Ömeroǧlu H, Aǧuş H, Biçimoǧlu A, Tümer Y. Evaluation of experienced surgeons’ decisions regarding the need for secondary surgery in developmental dysplasia of the hip. J Pediatr Orthop 2012; 32(1): 58-63.

Perkins G. Signs by which to diagnose congenital dislocation of the hip.

Lancet 1928; (274): 3-5.

Portney L G, Watkins M P. Foundation clinical research. Foreign Aff 2012; 91(5): 1689-99.

Reimers J. The stability of the hip in children. Acta Orthop Scand 1980; S184: 1-97.

Terjesen T. Residual hip dysplasia as a risk factor for osteoarthritis in 45 years follow-up of late-detected hip dislocation. J Child Orthop 2011; 5(6): 425-31.

Tönnis D. Normal values of the hip joint for the evaluation of X-rays in children and adults. Clin Orthop Relat Res 1976; (119): 39-47.

Tucci J J, Kumar S J, Guille J T, Rubbo E R. Late acetabular dysplasia following early successful Pavlik harness treatment of congenital dislocation of the hip. J Pediatr Orthop 1991; 11(4): 502-5.

Wood R, Basset K, Foerster V, Spry C, Tong L. Tesla magnetic resonance imaging scanners compared with 3.0 Tesla magnetic resonance imaging scanners: systematic review of clinical effectiveness. [Internet]. Ottawa Can Agency Drugs Technol Heal; 2011.

Yang Y, Porter D, Zhao L, Zhao X, Yang X, Chen S. How to judge pelvic malposition when assessing acetabular index in children: three simple parameters can determine acceptability. J Orthop Surg Res 2020; 9:1-9.

274

Acta Orthopaedica 2021; 92 (3): 274–279

Direct superior approach versus posterolateral approach in total hip arthroplasty: a randomized controlled trial on early outcomes on gait, risk of fall, clinical and self-reported measurements Michele ULIVI 1, Luca ORLANDINI 1, Jacopo A VITALE 1, Valentina MERONI 1, Lorenzo PRANDONI 2, Laura MANGIAVINI 1,3, Nicolò ROSSI 2, and Giuseppe M PERETTI 1,3 1 IRCCS

Istituto Ortopedico Galeazzi, Milan; 2 Residency Program in Orthopedics and Traumatology, University of Milan, Milan; 3 Department of Biomedical Sciences for Health, University of Milan, Milan, Italy Correspondence: jacopo.vitale@grupposandonato.it Submitted 2020-09-10. Accepted 2020-11-24.

Background and purpose — Several surgical approaches are used in primary total hip arthroplasty (THA). In this randomized controlled trial we compared gait, risk of fall, self-reported and clinical measurements between subjects after direct superior approach (DSA) versus posterolateral approach (PL) for THA. Patients and methods — Participants with DSA (n = 22; age 74 [SD 8.9]) and PL (n = 23; age 72 [7.7]) underwent gait analysis, risk of fall assessment and Timed Up and Go Test (TUG) before (PRE), 1 month (T1) and 3 months after (T3) surgery. Data on bleeding and surgical time was collected.

Results — DSA resulted in longer surgical times (90 [14] vs. 77 [20] min) but lower blood loss (149 [66] vs. 225 [125] mL) than PL. DSA had lower risk of fall at T3 compared with T1 and higher TUG scores at T3 compared with T1 and PRE. PL improved balance at T3 compared with T1 and PRE. Spatiotemporal gait parameters improved over time for both DSA and PL with no inter-group differences, whereas DSA, regarding hip rotation range of motion, showed lower values at T3 and T1 compared with PRE and, furthermore, this group had lower values at T1 and T3 compared with PL. All foregoing comparisons are statistically signficant (p < 0.05) Interpretation — DSA showed longer surgical time and lower blood loss compared with PL and early improvements in TUG, spatiotemporal, and kinematic gait parameters, highlighting rapid muscle strength recovery.

The lateral and posterolateral surgical approaches to the hip, either performed traditionally or with a mini invasive approach, foresee incision of the iliotibial band and imply the incision of the abductor muscles or the external rotator muscles (Murphy and Millis 1999). While the impact of the damage of these muscles on patients’ outcome is well known, the effect of surgical technique on the stabilization system, such as the fascia lata and the entire fascial system, has rarely been studied. It is known that a distortion within the fascial system leads to a loss of function and changes in the connecting structure of the locomotor system, as the fascial system is the key for stability and sensorimotor function (Huijing 2012, Vitale et al. 2019).

The minimally invasive direct superior approach (DSA) combines minimal invasiveness and the advantage of preserving the fascia lata and the abductor muscles. DSA was first described by Stephen Murphy: it consists of a blunt dissection of the gluteus maximus muscle and a superior capsulotomy to reach the femoral neck and requires specialized instrumentation to preserve the tendons of the extrarotator muscles and to easily reach the femoral neck (Murphy and Millis 1999).

This randomized controlled trial (RCT) investigates whether a modified DSA with avoidance of sectioning the iliotibial band (ITB) can elicit better early outcomes in gait and risk of fall.

Total hip arthroplasty (THA) was performed with the aid of dedicated and modified instrumentation. We hypothesized that DSA, with the preservation of the ITB, would lead to a faster recovery compared with the posterolateral approach (PL).

© 2021 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 10.1080/17453674.2020.1865633

Acta Orthopaedica 2021; 92 (3): 274–279

ENROLLMENT

Assessed for eligibility n = 50 Randomized n = 50

ALLOCATION

Allocated to DSA intervention (n = 25) Received allocated intervention (n = 25)

Allocated to PL intervention (n = 25) Received allocated intervention (n = 25)

FOLLOW-UP

Lost to follow-up (n = 3): – anterior traumatic dislocation, 2 – periprosthetic fracture, 1

Lost to follow-up (n = 1): – ischemic stroke, 1

ANALYSIS

Analyzed (n = 22)

Analyzed (n = 23)

Excluded from analysis (n = 0)

Excluded from analysis (n = 1): – incomplete data due to technical problems during data acqusition

Figure 1. CONSORT 2010 flow diagram of steps involved in the screening and enrollment of the DSA and PL groups. PL, posterolateral mini approach; DSA, direct superior approach.

Patients and methods Participants and design Based on the literature, we considered it clinically significant to observe a mean effect size of 0.9 in the hip rotation range of motion (ROM) between the 2 groups. Therefore, considering an α level with p = 0.05 and a power of 90%, 22 subjects would be necessary in each of the 2 groups to detect a statistically significant difference in hip rotation ROM. To prevent possible dropout of subjects during the study (estimated to 10–15%), the sample size was increased to a total of 50 subjects. The inclusion criteria were presence of noninflammatory degenerative joint disease, rheumatoid arthritis, age between 60 and 75, BMI between 18 and 30 and absence of contralateral THA (Figure 1). The subjects were randomly allocated to one of the 2 treatment groups with a computer-generated 1:1 randomization list.

Surgical procedure Between April 2017 and December 2018 1 senior orthopedic surgeon (MU), experienced in the PL approach, performed all surgeries. All patients were positioned on the lateral decubitus of the contralateral side. In the PL group a mini standard approach was used. The total incision of the fascia lata accounted for approximately 3–5 cm proximally and distally to the tip of the greater trochanter. In the DSA group, no iliotibial band section was performed; however, splitting of the gluteus maximus muscle, short external rotator preservation with selective division of the piriformis tendon and a posterior capsulotomy were performed. Implants for both DSA and PL were the Accolade II femoral stem (Stryker, Michigan, USA) and Trident cup with poly insert (Stryker, Michigan, USA).

Surgical time, blood loss, pre- and post-surgery hemoglobin levels, and adverse outcomes were recorded. Radiographs

275

were obtained accordingly to routine procedures preoperatively, postoperatively, and at 3 months and 6 months.

Self-reported and clinical outcome measures All patients underwent preoperatively (PRE), at 1 week (W1), 2 weeks (W2), 3 weeks (W3), 1 month (T1), 3 months (T3) and 6 months (T6) after surgery the following clinical evaluations: the Hip Disability and Osteoarthritis Outcome Score (HOOS) (Klässbo et al. 2003), subjective ratings of pain by VAS, the Harris Hip Score (HHS), and the SF-12. Evaluation of all scales at T6 was conducted by the same investigators (VM, LP and NR).

Gait analysis Patients underwent clinical gait analysis before (PRE), 1 month (T1), and 3 months after (T3) surgery in the Motion Analysis Laboratory of our institute. For the gait analysis, a Helen Hayes marker set of 22 retro-reflective passive markers was used and a Davis biomechanical model was applied during data acquisition and processing (Davis et al. 1991).

Patients were asked to walk as best as they could at a selfselected speed without walking aids along a 13-meter walkway at least 6 times. An optoelectronic system (SMART-D, BTS Bioengineering, Milan, Italy) with 8 infrared cameras (sampling rate 100 Hz) was used for spatiotemporal and kinematic data acquisition. Mark trajectories were recorded, reconstructed, and processed by SMART-D Analyzer software (BTS Bioengineering, Quincy, MA, USA). The gait parameters were: (1) spatiotemporal variables: stance phase (percentage), swing phase (percentage) step length (meters), stride length (meters), gait speed (m/s), and gait cadence (steps/ minute); stance and swing were normalized as a percentage of the gait cycle; (2) kinematic parameters (in degrees): hip flexion–extension ROM, hip abduction–adduction ROM, hip rotation ROM, hip obliquity ROM. Risk of fall assessment and Timed Up and Go test The risk of fall was evaluated before (PRE), 1 month (T1), and 3 months after (T3) surgery with the OAK system (Khymeia, Padova, Italy). For this purpose, the device provides an automated version of the Brief-BESTest (Padgett et al. 2012).

It yields a point-score from 0 to 24 and it has been shown that the relative optimal cutoff point was a 16 point score out of 24; a point score between 17 and 24 classifies a subject as low risk who would otherwise be classified as being at medium/high risk (Castellini et al. 2019). In addition, the Timed Up and Go test (TUG) (Podsiadlo and Richardson 1991) was performed at PRE, T1, and T3.

Statistics Baseline characteristics Unpaired Student’s t-tests, or non-parametric Mann–Whitney rank tests when needed, were used to test the differences between groups. Fisher’s exact test was used to evaluate the

276

Acta Orthopaedica 2021; 92 (3): 274–279

Table 1. Baseline characteristics of DSA and PL. Data are mean (SD) Approach

direct superior posterolateral Factor (n = 22) (n = 23) Age (years) 74 (8.9) Height (m) 173 (5.2) Weight (kg) 69 (10) BMI 23 (2.8) Sex (Male:Female) 7:15 Gait speed at PRE (m/s) 0.60 (0.25) Surgical time (minutes) 90 (14) Bleeding (mL) 149 (66) Hemoglobin (g/dL) preoperative 14 (1.2) postoperative 12 (1.2)

72 (7.7) 174 (6.4) 72 (11) 24 (2.0) 10:13 0.55 (0.22) 77 (20) b 225 (125) a 14 (0.79) 12 (1.0)

a p = 0.04; b p = 0.002 (Mann–Whitney rank test due to non-normally distributed data).

differences in the frequency distribution of females and males in the 2 study groups while intragroup (pre- vs. post-surgery) and intergroup (DSA vs. PL) differences in hemoglobin levels were checked using 2-way analysis of variance (ANOVA) with Bonferroni’s multiple comparisons test. P < 0.05 was considered statistically significant (Table 1).

Gait and risk of fall Gait parameters, OAK, and TUG data were acquired at 3 time points, PRE, T1, and T3 (before, 1 month, and 3 months after surgery), for each group (DSA and PL). The normal distribution of each parameter was then checked with the Shapiro– Wilk test. The non-normal distributed variables were tested using non-parametric methods. Mixed ANOVA was applied to each gait parameter, OAK and TUG data. 1st, we checked the presence of an interaction between the 2 factors, within-subjects factor (time) and between-subjects factor (groups). 2nd, we evaluated the simple main effects of group and time. The time effect was assessed using 2 separate one-way repeatedmeasure ANOVAs (1 for DSA and 1 for PL) followed by the Tukey–Kramer post-hoc test for differences in each parameter between PRE, T1, and T3. The effect of group was determined by evaluating the differences in each parameter between the DSA and PL with 3 separate unpaired Student’s t-tests (one for PRE, T1, and T3). Significance was set at p < 0.05.

Self-reported clinical measures HOOS, VAS, HHS, and SF-12 data were acquired at 7 time points (PRE, W1, W2, W3, T1, T3, T6). The normal distribution of each measure was checked with the Shapiro–Wilk test.

Mixed ANOVA was applied to all data and the non-normal distributed variables were tested using non-parametric methods. Significance was set at p < 0.05. All statistical analysis was performed using GraphPad Prism version 6.00 (GraphPad Software, San Diego, CA, USA).

OAK score

TUG score

24

3

20 16

2

12

p < 0.01

p < 0.05

8 4 0

1

p < 0.05

p < 0.01

p < 0.05

DSA PL PRE

T1

T3

0

DSA PL PRE

T1

T3

Figure 2. Mean (dot) and standard deviation (whiskers) of OAK and TUG for DSA group (n = 22) and PL group (n = 23) before (PRE), 1 month (T1), and 3 months (T3) after surgery. Dashed line indicates the 16 cutoff point score for OAK (i.e., a point score below 16 indicates a medium/high risk of fall).

Ethics, registration, funding, and potential conflicts of interest The study was supported by the Italian Ministry of Health (Ricerca Corrente) and was approved by the Ethical Committee of Vita-Salute San Raffaele University (Milan, Italy; registration number: 129/INT/2016) in compliance with current national and international laws and regulations governing the use of human subjects (Declaration of Helsinki II). Written informed consent was obtained from all patients. The study was registered at clinicaltrials.gov (NCT04358250). This study has received a liberal grant from Stryker, which has not been involved at any point neither in the conception of study design, collection, analysis, interpretation of the data nor in writing of this manuscript. LO is a paid part-time employee of a manufacturer (Smith&Nephew) while MU, JV, VM, LP, LM, NR and GP declared that they have no conflict of interest.

Results 45 patients were included in the final analysis, 22 for DSA and 23 for PL (Table 1). The number of major adverse events was higher in DSA (3/25) than in PL (1/25). 1 periprosthetic fracture and 2 anterior dislocations due to falls occurred in DSA and 1 ischemic stroke occurred in the PL group. No other adverse events were observed. The groups were similar regarding height, weight, BMI, age, and sex distribution, whereas DSA had longer surgical times (90 [SD14] minutes vs. 77 [20] minutes; p = 0.002) but lower blood loss (148 [66] mL vs. 225 [125] mL; p = 0.03) than PL. Risk of fall and TUG DSA had higher OAK values at T3 compared with T1 (p = 0.03) and higher TUG scores at T3 compared with T1 (p = 0.009) and PRE (p = 0.009). Furthermore, PL registered greater OAK

Acta Orthopaedica 2021; 92 (3): 274–279

277

Stance phase (%) – surgery

Stance phase (%) – no surgery

Step length (m) – surgery

Step length (m) – no surgery

100

100

1.0

1.0 p < 0.001

p < 0.001

80

80

0.8

p < 0.05

60

60

0.6

40

40

0.4

20 0

DSA PL PRE

T1

T3

20

DSA PL

0

PRE

T1

T3

p < 0.001

0.2 0

PRE

T1

Cadence (steps/minute)

Gait speed (m/s) 1.5

1.6

0.8

0

DSA PL PRE

T1

T3

0

p < 0.01

T1

DSA PL PRE

T1

T3

p < 0.001

0.5 DSA PL

PRE

0

p < 0.01

1.0

50

0.2

p < 0.001

100

1.2

0.4

T3

p < 0.01

p < 0.001

p < 0.001

0.6

DSA PL

150 p < 0.001

p < 0.01

0.4

Stride length (m) p < 0.01

0.8

p < 0.001

2.0

p < 0.001

p < 0.05

T3

scores at T3 compared with T1 (p = 0.02) and PRE (p = 0.04) while TUG did not show any statistically significant difference (Figure 2 and Table 2, see Supplementary data).

Spatiotemporal variables Stance phase and swing phase of the operated limb did not show any statistically significant intra- and inter-group difference whereas both DSA (p = 0.04) and PL (p < 0.001) had lower swing phase and stance phase values in T3 compared with T1. DSA increased step length (+0.05 m for both legs; p = 0.001 and p < 0.001 for operated and non-operated leg respectively) and stride length (+0.10 m for both legs; p = 0.007 and p = 0.001 for operated and non-operated leg respectively) in T3 compared with T1 and, in addition, PL significantly increased from PRE and T1 to T3 for both stride and step length (Figure 3 and Table 2, see Supplementary data).

DSA increase gait cadence from 94 [2] steps/minute to 101 [13] steps/minute (p < 0.001) and gait speed from 0.56 [0.23] m/s to 0.73 [0.23] m/s (p < 0.001) from T1 to T3.

Kinematic variables PL increased hip flexion–extension ROM of the operated leg from PRE (p < 0.001) and T1 (p < 0.001) to T3 while DSA did not show any statistically significant difference. Regarding hip abduction–adduction ROM, DSA registered higher values in T3 compared with PRE (p = 0.03) for the surgical leg, while PL showed an increased from T1 to T3 (p = 0.003) for the healthy leg (Figure 4 and Table 3, see Supplementary data).

Hip rotation ROM for the healthy leg did not show any statis-

0

DSA PL PRE

T1

T3

Figure 3. Mean (dot) and standard deviation (whiskers) of stance phase, swing phase, step length, stride length, cadence, and velocity for DSA group (n = 22) and PL group (n = 23) before (PRE), 1 month (T1), and 3 months (T3) after surgery.

tically significant inter- or intra-group difference while DSA for the operated limb, showed lower values at T3 (p = 0.01) and T1 (p = 0.005) compared with PRE and, furthermore, had lower values at T1 (p = 0.04) and T3 (p = 0.04) compared with PL (Figure 5). Finally, PL did not show significant differences in hip obliquity ROM while DSA revealed a significant increase only at T1 compared with PRE for both the operated (p = 0.007) and non-operated (p = 0.04) lower limb.

Self-reported clinical measures HOOS, HHS, SF-12, and VAS scores improved from PRE to T6 (p < 0.001), while no statistically significant differences between groups were detected. All self-reported measures for both DSA and PL showed a constant improvement from baseline to each evaluation time with a plateau observed at T3.

Discussion To our knowledge, this is the first RCT assessing the shortterm impact of DSA and PL approaches in total hip arthroplasty. DSA foresees the complete sparing of the fascia lata, whereas in PL this anatomical structure is incised and laterally sutured. We found that DSA had longer surgical time but lower blood loss and, in addition, DSA showed better TUG at T3 while PL did not display any difference over time. Overall, we found similar results between DSA and PL in risk of fall, spatiotemporal, and kinematic gait variables; only hip rotation ROM displayed inter-group differences

278

Acta Orthopaedica 2021; 92 (3): 274–279

Flexion–extension (°) – surgery

Flexion–extension (°) – no surgery

Adduction–abduction (°) – surgery

Adduction–abduction (°) – no surgery

80

80

20

20

p < 0.05

DSA PL p < 0.001

60

p < 0.01

DSA PL

p < 0.05

DSA PL

p < 0.05

60

15

40

40

10

10

20

20

5

5

p < 0.05

p < 0.05 p < 0.05

15

p < 0.001

0

PRE

T1

T3

0

DSA PL PRE

T1

T3

Obliquity ROM (°) – surgery

Obliquity ROM (°) – no surgery

12

12

p < 0.01

10

DSA PL

PRE

6

6

4

4

2

2

PRE

T1

0

T3

T1

T3

Rotation ROM (°) – surgery DSA PL

30

p < 0.05

0

PRE

T1

T3

Rotation ROM (°) – no surgery DSA PL

DSA PL

30

p < 0.01

8

8

0

p < 0.05

10

0

p < 0.05

PRE

T1

T3

p < 0.05

20

20

10

10

0

PRE

T1

T3

0

PRE

T1

T3

Figure 4. Median (black line), first and third quartiles (box), and minimum and maximum (whiskers) of hip extension–flexion ROM, hip abduction–adduction ROM, hip obliquity ROM, and hip rotation ROM, for DSA group (n = 22) and PL group (n = 23) before (PRE), 1 month (T1) and 3 months (T3) after surgery.

Internal rotation (°) – surgery

External rotation (°) – surgery

40

60

p < 0.01

DSA PL

20

40

0

20

–20

0 p < 0.01

–40

PRE

T1

DSA PL T3

–20

PRE

T1

T3

Figure 5. Median (black line), first and third quartiles (box), and minimum and maximum (whiskers) of hip internal and external rotation values for DSA group (n = 22) and PL group (n = 23) before (PRE), 1 month (T1) and 3 months (T3) after surgery.

with lower values at T1 and T3 for DSA compared with PL. Results of HOOS, HHS, SF-12, and VAS were similar between DSA and PL. Thus, our initial hypothesis was partially confirmed.

In a previous study, the consequences of dissection and suturing the fascia lata were studied using MRI and ultrasonography and it was reported that the fascia has strong relationships with the underlying musculature. It appears that an intact fascia represents a vital component for the normal function of thigh muscles and knee control in bipedal locomotion (Huijing 2012). As a consequence, our primary aim was to

evaluate 3D movement differences during walking (i.e., gait analysis) between the 2 study groups. The risk of fall, evaluated by the OAK device, did not show any statistically significant inter-group difference but a significant improvement for both DSA and PL was detected at T3 compared with earlier assessments. TUG was significantly different from preoperative values to T1 (p = 0.009) and T3 (p = 0.009) only for DSA but not for PL (Figure 2). However, the inter-group difference was not statistically significant. All spatiotemporal parameters significantly improved in both groups. Gait cadence and speed had a highly significant improvement from T1 to T3, which was more pronounced in the DSA group. The analysis of the kinematic parameters deserves particular attention and accurate interpretation. While ROM improvement for hip flexion/ extension was significant in PL from PRE to T3 and from T1 to T3, results on hip abduction/adduction ROM showed a different pattern with a significant difference from PRE to T3 only in the DSA group. However, the inter-group differences for these 2 kinematic parameters were not significant.

It is noteworthy that hip rotation ROM is the only kinematic parameter that showed a statistically significant difference between groups: DSA registered a significant reduction from PRE to T1 and T3; in addition, DSA had lower values at T1 (p = 0.04) and T3 (p = 0.04) compared with PL. ROM value for hip rotation, expressed in degrees, is the result of the difference between extra-rotation (max) and intra-rotation (min); therefore, the lower ROM value in T1 and T3 reported in the DSA

Acta Orthopaedica 2021; 92 (3): 274–279

group should be interpreted as a difference in the capacity of intra-rotating the hip in DSA compared with PL. The clinical interpretation of this kinematic parameter may indicate that the DSA approach, with consequent surgical sparing of the fascia lata, may result in higher hip stability at the immediate postoperative time points. This result is corroborated by the higher abduction/adduction ROM and the concomitant decrease of the hip rotation ROM, which is mainly ascribed to an improved intra-rotation component as reported in the literature (Ewen et al. 2012, Behery and Foucher 2014, Zeni et al. 2018, Yoo et al. 2019). Previous studies also report that kinematic data on healthy subjects indicates that during the abduction phase, corresponding to gait start, the hip is intrarotated, supporting the importance of this kinematic parameter (Behery and Foucher 2014, Kolk et al. 2014).

The clinical results (HHS) as well as self-reported outcome measures (SF12, HOOS, VAS) were similar between the 2 groups. Surgical time was prolonged in the DSA group (90 vs. 77 minutes), possibly attributable to a prolonged learning curve on the new surgical access and the use of dedicated instruments for the DSA approach. We also found a reduction of intra- and perioperative blood loos in DSA compared with PL (149 vs. 216 mL), which we consider to be clinically important.

The number of serious adverse events was higher in DSA (3/25) compared with PL (1/25). In particular, 1 periprosthetic fracture of the greater trochanter and 2 anterior traumatic dislocations appeared in the DSA group. The periprosthetic fracture was treated with open reduction and internal fixation whereas, as concerns the 2 dislocations, cup inclination and anteversion values before reduction were 28° and 31° respectively, which may explain the instability. Both patients were treated with closed reduction and had no further consequences. The reported serious adverse events in DSA could be ascribed to the learning process for the new surgical approach, which implies reduced visualization of the proximal femur and acetabulum.

Furthermore, 1 ischemic stroke occurred in the PL group.

The main strength of this study is the RCT design, which, together with appropriate statistically pre-determined sample size calculation, guarantees robustness of the findings. However, a possible limitation is the relatively short-term followup. Nevertheless, the present study focused on the short-term follow-up of patients (1–3 months) to investigate the potential effects on early functional recovery of this type of minimally invasive fascia lata sparing surgery.

In summary, we demonstrated that the novel direct superior approach is associated with a decrease in intra- and perioperative blood loss and with early improvements in TUG, spatiotemporal, and kinematic parameters highlighting rapid muscle strength recovery. Surgical time was longer in the DSA group.

279

Clinical and self-reported functional scores did not differ between the 2 treatments.

Supplementary data Table 2 and 3 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2020.

1865633.

MU, LO, and GP designed and coordinated the study. JAV, VM, NR, LP, and LM collected the data and drafted the manuscript. LO helped to draft the manuscript and JAV calculated the statistics. All authors contributed to the interpretation of the data and results and to the preparation of the manuscript.

Acta thanks Scott Crawford for help with peer review of this study.

Behery O A, Foucher K C. Are Harris Hip Scores and gait mechanics related before and after THA? Clin Orthop Related Res 2014; 472(11): 3452-61.

Castellini G, Gianola S, Stucovitz E, Tramacere I, Banfi G, Moja L. Diagnostic test accuracy of an automated device as a screening tool for fall risk assessment in community-residing elderly: a STARD compliant study.

Medicine (Baltimore) 2019; 98(39): e17105 Davis R B, Õunpuu S, Tyburski D, Gage J R. A gait analysis data collection and reduction technique. Hum Movement Sci 1991; 10(5): 575-87.

Ewen A M, Stewart S, St Clair Gibson A, Kashyap S N, Caplan N. Postoperative gait analysis in total hip replacement patients-a review of current literature and meta-analysis. Gait and Posture 2012; 36(1): 1-6.

Huijing P A. Fascia: clinical and fundamental scientific research. In Fascia: the tensional network of the human body. Amsterdam: Elsevier; 2012. p.

481-82.

Klässbo M, Larsson E, Mannevik E. Hip Disability and Osteoarthritis Outcome Score: an extension of the Western Ontario and McMaster Universities Osteoarthritis Index. Scand J Rheumatol 2003; 32(1): 46-51.

Kolk S, Minten M J M, van Bon G E A, Rijnen W H, Geurts A C H, Verdonschot N, Weerdesteyn V. Gait and gait-related activities of daily living after total hip arthroplasty: a systematic review. Clin Biomech 2014; 29(6): 705-18.

Murphy S B, Millis M B. Periacetabular Osteotomy without abductor dissection using direct anterior exposure. In Clinical orthopaedics and related research. Philadelphia: Lippincott Williams & Wilkins; 1999. pp: 92-8.

Padgett P K, Jacobs J V, Kasser S L. Is the BESTest at its best? A suggested brief version based on interrater reliability, validity, internal consistency, and theoretical construct. Phys Ther 2012; 92(9): 1197-1207.

Podsiadlo D, Richardson S. The Timed ‘Up & Go’: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc 1991; 39(2): 142-8.

Vitale J A, Castellini G, Gianola S, Stucovitz E, Banfi G. Analysis of the Christiania Stop in professional roller hockey players with and without previous groin pain: a prospective case series study. Sport Sciences for Health 2019; 15(3).

Yoo J-I, Cha Y-H, Kim K-J, Kim H-Y, Choy W-S, Hwang S-C. Gait analysis after total hip arthroplasty using direct anterior approach versus anterolateral approach: a systematic review and meta-analysis. BMC Musculoskelet Disord 2019; 20(1): 63.

Zeni J, Madara K, Witmer H, Gerhardt R, Rubano J. The effect of surgical approach on gait mechanics after total hip arthroplasty. J Electromyography Kinesiology 2018; 38: 28-33.

280

Acta Orthopaedica 2021; 92 (3): 280–284

Population-based 10-year cumulative revision risks after hip and knee arthroplasty for osteoarthritis to inform patients in clinical practice: a competing risk analysis from the Dutch Arthroplasty Register Maaike G J GADEMAN 1,2, Liza N VAN STEENBERGEN 3, Suzanne C CANNEGIETER 2, Rob G H H NELISSEN 1, and Perla J MARANG-VAN DE MHEEN 4 1 Department of Orthopaedics, Leiden University Medical Center, Leiden; 2 Department of Clinical Epidemiology, Leiden University Medical Center, Leiden; 3 Dutch Arthroplasty Register, ‘s Hertogenbosch; 4 Department of Biomedical Data Sciences, Medical Decision Making, Leiden University Medical Center,

Leiden, The Netherlands Correspondence: m.g.j.gademan@lumc.nl Submitted 2019-11-13. Accepted 2020-11-30.

Background and purpose — A lifetime perspective on revision risks is needed for optimal timing of arthroplasty in osteoarthritis (OA) patients, weighing the benefit of total hip arthroplasty/total knee arthroplasty (THA/TKA) against the risk of revision, after which outcomes are less favorable.

Therefore, we provide population-based 10-year cumulative revision risks stratified by joint, sex, fixation type, and age.

Patients and methods — Data from the Dutch Arthroplasty Register (LROI) was used. Primary THAs and TKAs for OA between 2007 and 2018 were included, except metalon-metal prostheses or hybrid/reversed hybrid fixation. Revision surgery was defined as any change of 1 or more prosthesis components. The 10-year cumulative revision risks were calculated stratified by joint, age, sex, at primary arthroplasty, and fixation type (cemented/uncemented), taking into account mortality as a competing risk. We estimated the percentage of potentially avoidable revisions assuming all OA patients aged < 75 received primary THA/TKA 5 years later while keeping age-specific 10-year revision risks constant.

Results — 214,638 primary THAs and 211,099 TKAs were included, of which 31% of THAs and 95% of TKAs were cemented. The 10-year cumulative revision risk varied between 1.6% and 13%, with higher risks in younger age categories. Delaying prosthesis placement by 5 years could potentially avoid 23 (3%) THA and 162 (17%) TKA revisions.

Interpretation — Cumulative 10- year revision risk varied considerably by age in both fixation groups, which may be communicated to patients and used to guide timing of surgery.

In Western countries about 10–23% of women and 6–15% of men receive a total knee arthroplasty (TKA) during their life.

For total hip arthroplasty (THA) these numbers are 12–16% of women and 8–11% of men (Ackerman et al. 2017a, b).

Although arthroplasty is an effective intervention, the optimal timing of arthroplasty is crucial given the long-term survival of the prosthesis is still limited.

To achieve optimal timing of primary THA/TKA, the benefit of surgery has to be weighed against the risk for revision surgery, which has less favorable outcomes (Petersen et al. 2015).

Hence, when younger patients have lower revision risks than elderly patients, one might consider delaying surgery to optimize outcome from a lifetime perspective.

Valid prediction models providing individualized lifetime revision risks may help guide decision-making on optimal timing, but such models are rare (Prokopetz et al. 2012, Paxton et al. 2015). Valid revision risks can also be provided by simply calculating these risks in the population of interest.

For example, Bayliss et al. (2017) modelled lifetime revision risks after TKA/THA with Clinical Practice Research Datalink data. However, they did not include implant (fixation) type or indication for surgery, while both may impact revision risks. Furthermore, most arthroplasty registries published cumulative revision risks, without taking the competing risk of dying into account. Also, stratification of revision risks into more than one/two subgroups was not performed, whereas this is important when providing personalized information.

We therefore provide 10-year cumulative revision risks for OA patients stratified by joint, sex, age, and fixation type, using data from the Dutch Arthroplasty Register and taking into account the competing risk of dying. We also estimate the number of potentially avoided arthroplasties by delaying TKA/THA by 5 years. Furthermore, we project our numbers onto the expected Dutch population in 2025/2035.

© 2021 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 10.1080/17453674.2021.1876998

Acta Orthopaedica 2021; 92 (3): 280–284

Total hip arthroplasties registered in the Dutch Arthroplasty Register 2007–2018 n = 243,529 Excluded (n = 28,891): – hybrid arthroplasty, 21,717 – metal on metal articulation, 5,518 – missing/unknown fixation, 1,430 – missing information on sex, 226 Study population n = 214,638

281

Total knee arthroplasties registered in the Dutch Arthroplasty Register 2007–2018 n = 224,923 Excluded (n = 13,824): – hybrid/unknown fixation, 13,644 – missing/incorrect data, 180 Study population n = 211,099

Flow chart of study population, total hip arthroplasty (left panel) and total knee arthroplasty (right panel).

Patients and methods Study design This is a population-based cohort study.

Data sources The Dutch Arthroplasty Register (LROI) is a nationwide population-based registry that includes arthroplasties implanted in the Netherlands since 2007. The LROI was initiated by the Netherlands Orthopaedic Association (NOV), and has a completeness of reporting of over 95% for primary THA and TKA and over 88% of hip and knee revision arthroplasties up to 2013 (van Steenbergen et al. 2015), further increasing to 98% in 2017 (www.lroi-report.nl).

Statistics Netherlands Statline is the electronic databank of Statistics Netherlands, which publishes statistical information regarding various aspects of the Dutch population. Data from Statline (http:// statline.cbs.nl) was used to project the number of revisions to the expected Dutch population by age and sex in 2025 and 2035. In this way we could quantify the impact on revision surgery of delaying primary arthroplasty by 5 years.

Study population and definitions All primary TKAs or THAs for OA from the LROI in the period 2007–2018 were included. Metal-on-metal THAs were excluded (n = 5,518, 2%), as these are not used anymore due to the high failure rates. Only prostheses with cemented or uncemented fixation were included and patients with missing data were excluded (Figure). For TKA and THA median length of follow-up was respectively 4.2 years (IQR 4.9) and 4.3 years (IQR 5.1), with a maximum of 12 years in both groups. Age at primary surgery, sex, fixation type (uncemented or cemented), and time between primary surgery and revision, death, or end of follow-up as well as status (revision, death, or alive without revision) were extracted from the LROI database. Revision surgery was defined as any change (insertion, replacement, and/or removal) of one or more components of the prosthesis.

Statistics Baseline characteristics stratified by joint were summarized using mean (SD) for continuous outcomes or number with percentage for categorical outcomes. All confidence intervals (CI) given represent the 95% confidence interval.

Survival time of the implant was calculated as the time from primary THA or TKA to first revision arthroplasty for any reason, death of the patient, or January 1, 2019. Cumulative revision risk within 10 years was calculated using competing risk analysis, where death was considered a competing risk (Lacny et al. 2015, Wongworawat et al. 2015). These cumulative revision risks were calculated stratified by joint, age at primary arthroplasty, sex, and cemented/uncemented fixation of the prosthesis. Cumulative risks within 10 years were not given if at 10 year less than 20 patients were at risk. Age at primary arthroplasty was categorized into to the following predetermined groups: 50–54, 55–59, 60–64, 65–69, 70–74, 75–79, 80–84 and 85–90 years.

In addition, we estimated the percentage of potentially avoided revisions by assuming that all patients under 75 years of age received their primary arthroplasty 5 years later while keeping the age-specific revision risks constant. This age was chosen in the context of the remaining life expectancy at older ages, where assuming a delay of 5 years for an 80-year-old seemed unrealistic, as well as the increasing operative risks at these older ages.

Finally, to quantify the impact of age at primary THA/ TKA on OA revision surgery in the near future we projected our numbers (primary arthroplasties and revision risks) onto the expected Dutch population in 2025 and 2035. We calculated the expected relative increase in the Dutch population by dividing the expected Dutch population, estimated by Statistics Netherlands (CBS) in the different age categories in 2025 and 2035, by the existing population of 2010 as reported

by CBS. We then multiplied the yearly number of primary arthroplasties (from the period 2007–2018) within each age category by these age-specific relative population increases to calculate the expected total number of primary arthroplasties for each separate category (stratified for age and fixation type) in 2025 and 2035. By applying the estimated revision risks to

282

Acta Orthopaedica 2021; 92 (3): 280–284

Table 1. Baseline characteristics

Factor

Table 4. Avoided revision surgeries in the Netherlands by delaying primary arthroplasty by 5 years

Total knee arthroplasty Total hip arthroplasty n = 211,099 n = 214,638

Age, mean (SD) Male sex, n (%) Fixation, n (%) Uncemented Cemented

69.2 (8.6) 73,070 (35)

70.4 (8.6) 70,723 (33)

10,598 (5) 200,501 (95)

148,362 (69) 66,123 (31)

Table 2. Cumulative revision percentages (CR) within 10 years for osteoarthritis patients with a primary total knee arthroplasty (TKA)

Avoided revision surgeries, n Type 2010 2025 2035 Total knee arthroplasty Total hip arthroplasty

162 23

203 26

198 20

Table 3. Cumulative revision percentages (CR) within 10 years for osteoarthritis patients with a primary total hip arthroplasty (THA)

TKA Male Female THA Male Female

Cemented Uncemented Cemented Uncemented

Cemented Uncemented Cemented Uncemented Age CR (95% CI) CR (95% CI) CR (95% CI) CR(95% CI) Age CR (95% CI) CR (95% CI) CR (95% CI) CR(95% CI) 50–54 55–59 60–64 65–69 70–74 75–79 80–84 85–90

11 (9.3–13) 9.4 (8.5–11) 6.4 (5.7–7.1) 5.4 (4.9–6.0) 4.8 (4.3–5.4) 3.4 (3.0–3.9) 2.5 (2.1–3.1) 1.6 (1.0–2.6)

13 (8.0–21) 8.9 (6.2–13) 7.6 (5.6–11) 4.8 (3.1–7.5) 4.4 (2.8–7.0) 3.3 (2.0–5.6) 3.2 (1.6–6.4) – a

11 (10–13) 8.9 (8.1–9.8) 6.8 (6.3–7.4) 5.4 (5.1–5.9) 4.8 (4.5–5.2) 3.6 (3.3–3.9) 2.3 (2.1–2.6) 1.7 (1.3–2.2)

10 (7.1–15) 7.7 (5.6–11) 6.6 (4.9–8.7) 7.8 (6.1–9.9) 5.4 (3.9–7.3) 3.4 (2.4–5.0) 3.0 (1.9–4.7) 2.0 (0.9–4.9)

a The

subgroup of uncemented TKAs in males 85–90 years was too small to calculate valid 10-year cumulative risk percentages.

the expected number of primary arthroplasties and comparing these with the current number of revisions, the total amount of avoided revisions in 2025 and 2035 could be estimated.

Ethics, funding, and potential conflicts of interest The medical ethical committee of the Leiden University Medical Centre considered the study not to subject to the Medical Research Involving Human Subjects Act (WMO) (protocol number G18.037).

The study was funded by a grant from the Dutch Arthritis Foundation (ARGON, BP12-3-401). This foundation did not play a role in the study’s design, conduct or reporting.

The authors have no conflicts of interest to declare.

Results 211,099 primary TKAs and 214,638 primary THAs were included, of which 95% and 31% were cemented. Most prostheses were placed in women (Table 1). The absolute numbers of arthroplasties, revision surgeries, and deaths per age category are included in Appendix 1.

In TKAs, there was a strong age gradient; in both fixation groups the 10-year cumulative revision risk, with arthroplasties placed at a younger age, having a higher cumulative risk (Table 2). The lowest cumulative risks were found in cemented TKAs that were placed in males aged 85–90 years (cumula-

50–54 – a 55–59 4.3 (2.6–7.1) 60–64 3.5 (2.4–5.2) 65–69 5.4 (4.3–6.8) 70–74 4.2 (3.5–5.1) 75–79 4.1 (3.3–5.0) 80–84 2.8 (2.3–3.5) 85–90 1.7 (1.1–2.7)

5.2 (4.2–6.6) 5.5 (4.6–6.5) 4.6 (4.0–5.2) 4.4 (3.8–5.0) 4.3 (3.7–4.8) 4.1 (3.5–4.7) 3.5 (2.8–4.3) 3.2 (2.2–4.6)

5.3 (2.9–9.6) 7.3 (4.6–11.5) 4.7 (3.6–6.1) 3.7 (3.0–4.5) 3.1 (2.7–3.6) 3.2 (2.8–3.6) 2.2 (1.9–2.6) 2.1 (1.6–2.6)

6.0 (4.8–7.4) 5.1 (4.3–6.0) 4.2 (3.7–4.8) 3.6 (3.3–4.1) 4.2 (3.7–4.6) 4.2 (3.7–4.7) 3.3 (2.8–3.8) 3.5 (2.8–4.5)

a The subgroup of cemented THAs in males 50–54 years was too small to calculate valid 10-year cumulative risk percentages.

tive revision risk 1.6%, 95% CI 0.96–2.6) (Table 2). The highest cumulative risks were found in uncemented TKAs placed in males aged 50–54 years (cumulative revision risk 13%, CI 8–21). For THA, a similar age gradient was found as for TKA, with higher 10-year cumulative risks for prostheses that were placed in younger patients, although less pronounced (Table 3).

10-year risk was higher for TKA than for THA (Tables 2 and 3).

We estimated that by delaying primary TKA and THA in OA patients for 5 years, 162 TKA revision (17%) and 23 THA revision surgeries (3%) could be avoided. Using the expected Dutch population in 2025 rather than the population from 2010, delaying primary TKA and THA in OA patients for 5 years, 203 (16%) TKA revision and 26 (3%) THA revision surgeries could be avoided. Projecting to the expected Dutch population in 2035 shows a slight decrease in these numbers because of the change in age distribution; in 2035 the Dutch population will consist of more elderly people compared with the years before (Table 4).

Discussion In this nationwide population-based registry study we estimated the 10-year cumulative revision risks stratified by joint, age, sex, and fixation of prosthesis for OA to provide a simple tool to help to estimate the revision risk when considering arthroplasty.

Acta Orthopaedica 2021; 92 (3): 280–284

The 10-year cumulative revision risks varied between 1.6% (male cemented TKA patients aged 85–90 years) and 13% (male uncemented TKA patients aged 50–54 years). The cumulative 10-year revision risks decreased by age irrespective of sex and fixation type. The age gradient was less pronounced in THA than in TKA patients. Delaying primary TKA and THA surgery by 5 years in patients under 75 years of age was estimated to avoid 3% of THA revisions and 17% of TKA revisions. Elderly patients may have a higher revision risk in the first year after arthroplasty, especially when they are frail and have various comorbidities (Johnson et al. 2019, Peters et al. 2019). However, we found that the long-term cumulative revision risks of arthroplasties placed in elderly patients were lower in all our categories than in younger patients. This finding is in accordance with previous findings (Julin et al.

2010, Wainwright et al. 2011, Carr et al. 2012, Ackerman et al. 2017a, Bayliss et al. 2017, SKAR 2017) although previous studies often did not take into account the competing risk of dying.

The choice of taking into account competing risks has been debated for several years within arthroplasty register societies and depends on the perspective taken (Van Der Pas et al. 2018).

When the competing risk of dying is not taken into account this answers the question “What would happen if the competing event could be prevented [from occurring], creating an imaginary world in which an individual remains at risk of failure from the event of interest” (Putter et al. 2007), in this case the risk of revision if there is no mortality, which is appropriate when considering the perspective on which implant would have the best longevity or for etiological questions (Sayers et al. 2018, Van Der Pas et al. 2018). However, one can argue as to whether this is appropriate when communicating absolute revision risks to patients, as then the risk of death is also of interest and should be included in the estimates (Koller et al.

2012, Lacny et al. 2015, Wongworawat et al. 2015, Ranstam and Robertsson 2017, Sayers et al. 2018). When including competing risks, a different question is answered: “What is the absolute risk of revision surgery as observed in practice?”

For the latter, the mortality pattern of the underlying population is also taken into account, which might be important particularly in older patients where mortality risks are higher.

Because individuals who die before experiencing a revision are censored in estimates that do not take competing risks into account while they are included in competing risks analyses, revision risks are lower in practice when taking into account death as a competing risk. We have chosen to use competing risk analysis as our aim was to present the “real life” observed 10-year revision risks in practice. Here the underlying mortality patterns of the different age groups are an essential underlying process (Koller et al. 2012, Sayers et al. 2018).

For optimal timing of primary arthroplasty, it seems beneficial to postpone the time to arthroplasty to decrease the risk of revision surgery and thereby optimize the overall outcome across the entire life course, given that outcomes after

283

revision surgery are often worse than after primary surgery.

However, only revision surgery was taken into account as an outcome in this study, which implicitly also takes into account any underlying patient and surgeon preference to perform the revision, whereas other relevant outcomes such as patientreported outcomes were not included. If delaying the primary surgery means that patients experience decreased functioning or increased pain during these years, then we may need to reconsider whether lifetime outcomes are in fact better. In the last decade, registers and cohort studies have started the assessment of patient-reported outcomes, but these are usually assessed only in the first year after primary arthroplasty and long-term outcomes as well as patient-reported outcomes after revision surgery are scarce. Such PROMs data would be valuable to compare long-terms PROMs after primary surgery with PROMs after revision surgery to inform patients on the extent of benefit they will attain, and thus be part of the decision-making on weighing risks and benefits. However, caution is required as, with time, other factors like comorbidities or ageing may also affect the PROMs and should not be attributed to surgery long ago. Moreover, we showed that delaying primary TKA and THA surgery by 5 years in the groups aged younger than 75 reduced the number of revision surgeries and thereby potentially their associated costs. However, one should bear in mind that these patients will likely need other treatment instead. For instance, patients could be offered physical therapy and additional pain medication to cope with their OA complaints. Postponing surgery may also lead to costs due to loss of productivity in patients who are still of working age.

Our study should be interpreted with its strengths and limitations in mind. One of its strengths is that our study is based on population-based data from the Dutch Arthroplasty Register, a nationwide registry that contains over 95% of primary hip and knee arthroplasties since 2010 in the Netherlands (van Steenbergen et al. 2015). The registry-based nature of the data implies that we have information only on revision surgery that was registered. Nevertheless, we consider this information bias regarding surgery was no big issue here as the completeness of TKA and THA revision arthroplasties in the LROI has been over 85% since 2012 and reached 98% in 2017 (www. lroi-report.nl).

A limitation is that we only took revision surgery into account and did not include other patient outcomes. For some patients it may not be possible to delay surgery for 5 years as functional complaints and pain may become too disabling, so that we will have overestimated the number of avoided revisions. Moreover, data are surgeon-reported revision risks that represent daily clinical practice. Hence, as indication criteria for revision surgery are not clearly defined and may vary between different surgeons, this could mean that in 2 similar patients 1 will receive a revision whereas the other will not.

This is reinforced further because treatment preferences may vary between patients as well as by sex and age (Mota et al.

2012). In certain cases the physical condition of a patient will

284

not allow revision surgery although it is indicated, which is likely to occur more often in elderly patients. As such, the patients receiving revision surgery will not include all patients in need of revision surgery. Also, we presented only the cumulative 10-year revision risks and risks over an even longer period (e.g., 20 years) may be substantially higher, especially in the younger age groups. For instance, Bayliss et al. (2017) found, in a study in which they predicted lifetime revision risks, that men who had their initial primary TKA surgery between the age of 50 and 54 years had a lifetime revision risk of 35%, considerably higher than the 10-year risks we found, but without taking competing mortality risks into account and this is therefore likely overestimated. Our nationwide population-based study has the advantage of including far more age groups as well as specific fixation groups, to make cumulative revision risks ready to be used by patients and surgeons in daily practice to improve decision-making regarding timing of primary surgery.

In conclusion, in this nationwide study we found that in both TKAs and THAs the cumulative 10-year revision risk percentages varied considerably by age, irrespective of sex and fixation of the prosthesis, but with a stronger age gradient for TKAs. By delaying the primary arthroplasty, revision procedures might be avoided, resulting in substantial revision reductions.

Supplementary data Appendix is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2021.

1876998

MG was involved in the conception and the design of the study, drafted the manuscript, and performed the statistical analysis. LS checked the statistical analysis and contributed to the revision of the manuscript. RN, SC, and PM contributed to the conception and design of the study, and revision of the manuscript.

Acta thanks Sten Rasmussen and Annette W-Dahl for help with peer review of this study.

Ackerman I N, Bohensky M A, de Steiger R, et al. Substantial rise in the lifetime risk of primary total knee replacement surgery for osteoarthritis from 2003 to 2013: an international, population-level analysis. Osteoarthritis Cartilage 2017a; 25(4): 455-61. doi: 10.1016/j. joca.2016.11.005.

Ackerman I N, Bohensky M A, de Steiger R, et al. Lifetime risk of primary total hip replacement surgery for osteoarthritis from 2003 to 2013: a multinational analysis using national registry data. Arthritis Care Res 2017b; 69(11): 1659-67. doi: 10.1002/acr.23197.

Bayliss L E, Culliford D, Monk A P, et al. The effect of patient age at intervention on risk of implant revision after total replacement of the hip or knee:

Acta Orthopaedica 2021; 92 (3): 280–284

a population-based cohort study. Lancet 2017; 389(10077): 1424-30. doi: 10.1016/S0140-6736(17)30059-4.

Carr A J, Robertsson O, Graves S, et al. Knee replacement. Lancet 2012; 379(9823): 1331-40. doi: 10.1016/S0140-6736(11)60752-6.

Johnson R L, Abdel M P, Frank R D, et al. Impact of frailty on outcomes after primary and revision total hip arthroplasty. J Arthroplasty 2019; 34(1): 56-64 e5. doi: 10.1016/j.arth.2018.09.078.

Julin J, Jamsen E, Puolakka T, et al. Younger age increases the risk of early prosthesis failure following primary total knee replacement for osteoarthritis: a follow-up study of 32,019 total knee replacements in the Finnish Arthroplasty Register. Acta Orthop 2010; 81(4): 413-19. doi: 10.3109/17453674.2010.501747.

Koller M T, Raatz H, Steyerberg E W, et al. Competing risks and the clinical community: irrelevance or ignorance? Stat Med 2012; 31(11-12): 1089-97. doi: 10.1002/sim.4384.

Lacny S, Wilson T, Clement F, et al. Kaplan–Meier survival analysis overestimates the risk of revision arthroplasty: a meta-analysis. Clin Orthop Relat Res 2015; 473(11): 3431-42. doi: 10.1007/s11999-015-4235-8.

Mota R E, Tarricone R, Ciani O, et al. Determinants of demand for total hip and knee arthroplasty: a systematic literature review. BMC Health Serv Res 2012; 12: 225. doi: 10.1186/1472-6963-12-225.

Paxton E W, Inacio M C, Khatod M, et al. Risk calculators predict failures of knee and hip arthroplasties: findings from a large health maintenance organization. Clin Orthop Relat Res 2015; 473(12): 3965-73. doi: 10.1007/ s11999-015-4506-4.

Peters R M, van Steenbergen L N, Stewart R E, et al. 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 2019. doi: 10.1016/j.arth.2019.08.024.

Petersen K K, Simonsen O, Laursen M B, Nielsen T A, et al. Chronic postoperative pain after primary and revision total knee arthroplasty. Clin J Pain 2015; 31(1): 1-6. doi: 10.1097/AJP.0000000000000146.

Prokopetz J J, Losina E, Bliss R L, et al. Risk factors for revision of primary total hip arthroplasty: a systematic review. BMC Musculoskeletal Disord 2012; 13: 251. doi: 10.1186/1471-2474-13-251.

Putter H, Fiocco M, Geskus R B. Tutorial in biostatistics: competing risks and multi-state models. Stat Med 2007; 26(11): 2389-430. doi: 10.1002/ sim.2712.

Ranstam J, Robertsson O. The Cox model is better than the Fine and Gray model when estimating relative revision risks from arthroplasty register data.

Acta Orthop 2017; 88(6): 578-80. doi: 10.1080/17453674.2017.1361130.

SKAR. The Swedish Knee Arthroplasty Register Annual Report; 2017. https://www.myknee.se/pdf/SVK_2017_1.2.pd Sayers A, Evans J T, Whitehouse M R, et al. Are competing risks models appropriate to describe implant failure? Acta Orthop 2018; 89(3): 256-8. doi: 10.1080/17453674.2018.1444876.

Van Der Pas S, Nelissen R, Fiocco M. Different competing risks models for different questions may give similar results in arthroplasty registers in the presence of few events. Acta Orthop 2018; 89(2): 145-51. doi: 10.1080/17453674.2018.1427314. van Steenbergen L N, Denissen G A, Spooren A, et al. More than 95% completeness of reported procedures in the population-based Dutch Arthroplasty Register. Acta Orthop 2015; 86(4): 498-505. doi: 10.3109/17453674.2015.1028307.

Wainwright C, Theis J C, Garneti N, et al. Age at hip or knee joint replacement surgery predicts likelihood of revision surgery. J Bone Joint Surg Br 2011; 93(10): 1411-5. doi: 10.1302/0301-620X.93B10.27100.

Wongworawat M D, Dobbs M B, Gebhardt M C, et al. Editorial: Estimating survivorship in the face of competing risks. Clin Orthop Relat Res 2015; 473(4): 1173-6. doi: 10.1007/s11999-015-4182-4.

Acta Orthopaedica 2021; 92 (3): 285–291

285

Isometric hip strength impairments in patients with hip dysplasia are improved but not normalized 1 year after periacetabular osteotomy: a cohort study of 82 patients Julie Sandell JACOBSEN 1,2, Stig Storgaard JAKOBSEN 3, Kjeld SØBALLE 3,4, Per HÖLMICH 5, and Kristian THORBORG 5,6 1 Research

Centre for Health and Welfare Technology, Programme for Rehabilitation, VIA University College, Aarhus; 2 Research Unit for General Practice in Aarhus, Aarhus; 3 Department of Orthopaedic Surgery, Aarhus University Hospital, Aarhus; 4 Department of Clinical Medicine, Aarhus University, Aarhus; 5 Sports Orthopaedic Research Center-Copenhagen (SORC-C), Department of Orthopaedic Surgery, Copenhagen University Hospital, Hvidovre; 6 Physical Medicine and Rehabilitation Research-Copenhagen (PMR-C), Department of Physical and Occupational Therapy, Copenhagen University Hospital, Hvidovre, Denmark Correspondence: jsaj@via.dk Submitted 2020-08-21. Accepted 2020-12-04

Background and purpose — In patients with hip dysplasia, knowledge of hip muscle strength after periacetabular osteotomy is lacking. We investigated isometric hip muscle strength in patients with hip dysplasia, before and 1 year after periacetabular osteotomy, and compared this with healthy volunteers. Furthermore, we investigated whether pre- to post-surgical changes in self-reported pain and sporting function were associated with changes in isometric hip muscle strength.

Patients and methods — Isometric hip muscle strength was assessed twice in 82 patients (11 men) with a mean age of 30 (SD 9) years, before and 1 year after surgery, and once in 50 healthy volunteers. Isometric hip muscle strength was assessed with a hand-held dynamometer. Copenhagen Hip and Groin Outcome Score was used to measure self-reported outcome.

Results — Despite 1-year improvements in isometric hip flexion (0.1 Nm/kg; 95% CI 0.06–0.2) and abduction (0.1 Nm/kg; CI 0.02–0.2), the patients’ muscle strength was 13–34% lower than the strength of the healthy volunteers both pre- and post-surgery (p < 0.01). Moreover, changes in self-reported pain were associated with changes in hip flexion (13 points per Nm/kg; CI 1–26) and abduction (14 points per Nm/kg; CI 3–25), while changes in self-reported sporting function were associated with changes in hip extension (9 points per Nm/kg; CI 1–18).

Interpretation — Isometric hip muscle strength is impaired in symptomatic dysplastic hips measured before periacetabular osteotomy. 1 year after surgery, isometric hip flexion and abduction strength had improved but muscle strength did not reach that of healthy volunteers.

Hip dysplasia, with a prevalence of 5%, can be asymptomatic but in some cases, pain presents in early adulthood, most commonly among young women aged 24–35 years ((Jacobsen and Sonne-Holm 2005, Nunley et al. 2011, Jacobsen et al. 2018b) These patients are at risk of developing osteoarthritis at an early age (Murphy et al. 1995, Wyles et al. 2017).

The periacetabular osteotomy (PAO) is the preferred treatment for symptomatic patients with hip dysplasia in Western Europe, North America, and Australia (Lerch et al. 2017).

Numerous studies have shown large improvements in selfreported pain, sporting function, and quality of life after PAO (Clohisy et al. 2017, Wasko et al. 2019). Nevertheless, little is known about what patients can expect with regard to physical capacity. In a study of 41 patients undergoing PAO (Mechlenburg et al. 2018), patients improved their leg extension power at 1-year follow-up to no differences between the operated and contralateral leg. Contrary to these findings, no changes in isometric hip flexor and abductor strength were found 2 years after PAO in a study of 22 patients, and their hip muscle strength was impaired compared to the strength of 29 healthy volunteers (Sucato et al. 2015). It is noteworthy, in both studies, that physical capacity was investigated in small populations and with conflicting results. Physical capacity should better be investigated prospectively in larger populations to gain knowledge on what patients can expect 1 year after PAO. We investigated isometric hip muscle strength in patients with hip dysplasia, before and 1 year after periacetabular osteotomy, and compared this with healthy volunteers. Furthermore, we investigated whether pre- to post-surgical changes in self-reported pain and sporting function were associated with changes in isometric hip muscle strength.

© 2021 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 10.1080/17453674.2020.1864911

286

Patients and methods Study design This is a prospective cohort study with 1-year follow-up, comparing isometric hip muscle strength between patients with hip dysplasia and healthy volunteers. The study is part of a prospective investigation on the same patient population, reporting pain and ultrasonographic abnormalities in hip muscles and tendons and physical activity level before and after PAO (Jacobsen et al. 2018a, 2018b, 2018c, 2019).

Setting From May 2014 to August 2015, we prospectively recruited patients with bilateral or unilateral hip dysplasia from the Department of Orthopedics at Aarhus University Hospital in Denmark (Jacobsen et al. 2018b). Later, from February 2019 to May 2019, we recruited healthy volunteers through personal networks, through invitations in social media, and from local private companies and public institutions (e.g., university hospitals, universities, and university colleges). We measured participant characteristics and outcomes at a clinical examination scheduled twice in the patients, before and 1 year after PAO, and once in the healthy volunteers.

Participants Patients with hip dysplasia were eligible for inclusion in this study if they had groin pain for at least 3 months, if Wiberg’s center–edge (CE) angle was < 25 degrees and if they were scheduled to undergo PAO. Patients with known comorbidities or a history of previous trauma or surgical interventions affecting the hip were excluded (Jacobsen et al. 2018b). The healthy volunteers were eligible if they were of same age and sex as the patients (not matched 1:1). Specifically, they were included if they were 18–49 years of age (i.e., equal to the patients), had a BMI of 18–26 (i.e., equal to the patients), and had no pain in back, hip/groin, knee, or ankle joint. We excluded healthy volunteers diagnosed with known hip dysplasia, healthy volunteers who performed sport at elite level and healthy volunteers having > 1 of the above listed exclusion criteria of the patients.

Participant characteristics Characteristics of the participants including age, sex, preferred sports, time in preferred sports, and time in general physical activity were recorded through standardized questions. Back pain intensity was measured with the Oswestry Disability Index (Fairbank et al. 1980). Pain at rest was measured on a numerical pain rating scale (NPRS) from no pain to unbearable pain (0–10). Hip-related pain was assessed with the Flexion/Adduction/Internal Rotation (FADIR) test and the Flexion/Abduction/External Rotation test (FABER) test (Troelsen et al. 2009). Occurrence of internal snapping hip was assessed with a standardized clinical test (i.e., moving from

Acta Orthopaedica 2021; 92 (3): 285–291

FABER position into extension, adduction, and internal rotation) (Tibor and Sekiya 2008), while back pain was assessed with posterior to anterior spring testing over the lumbar spinous processes (Schneider et al. 2008). SSJ measured the CE angle (Wiberg 1939), the Tönnis acetabular index angle, and the Tönnis osteoarthritis grade using standardized standing anteroposterior radiographs, while height and weight were measured and used to calculate BMI. Periacetabular osteotomy After baseline examination, all patients underwent the minimally invasive approach for PAO performed by SSJ and KS (Jacobsen et al. 2019). The surgical procedure has been described previously (Troelsen et al. 2008). However, in short, the acetabulum was reoriented through 3 separate osteotomies aiming to improve the coverage of the femoral head.

Post-surgery, the patients received in-hospital standardized rehabilitation including active range-of-motion exercises in lying and standing, and stair and gait training with crutches.

The patients were discharged after approximately 2 days. For the first 6–8 weeks, the patients were allowed only partial weight-bearing with a maximum load of 30 kg. After discharge, the patients followed individualized physiotherapyled rehabilitation for 2–4 months, including 2 weekly training sessions in groups.

Outcomes Muscle strength GHM and JSJ assessed isometric hip muscle strength in the index limb with a handheld dynamometer (PowerTrack II Commander, JTECH Medical, Salt Lake City, UT, USA). The examiners used a standardized reliable dynamometer technique previously published by Thorborg et al. (2010). Isometric hip muscle strength was assessed with a make test in sitting position for flexion, in supine position for abduction and adduction, and in prone position for extension. The order of the individual strength assessments was randomized in order to avoid systematic bias.

The participants were instructed to stabilize themselves by holding on to the side of the examination table. They were instructed to practice 2 sub-maximal contractions, the 1st into the examiner’s hand and the 2nd into the dynamometer.

During assessment, the participants exerted a 5-second maximum voluntary contraction against the dynamometer. We normalized all strength values to moment arms and weight and reported strength values in Nm/kg bodyweight. After each muscle strength assessment (i.e., hip flexion, extension, abduction, and adduction) patients verbally rated pain during maximum muscle contraction on a NPRS. The inter-rater reliability of the muscle strength assessments was investigated by Jacobsen et al. (2018b). The intraclass correlation coefficient was > 0.70 for all muscle strength assessments and the standard error of measurement ranged between 9.5% and 14%.

Acta Orthopaedica 2021; 92 (3): 285–291

287

Self-reported outcome Eligible volunteers Eligible patients n = 71 n = 135 Self-reported outcome was measured with the Copenhagen Hip and Groin Outcome Score Declined to participate Declined to participate n=6 n = 19 (HAGOS) (Thorborg et al. 2011a). HAGOS is designed to measure self-reported outcome in young Excluded Excluded n = 14 n = 16 to middle-aged patients with hip and/or groin pain and measures self-reported outcome from 0 to 100 Controls informed consent Patients informed consent n = 51 n = 100 points on 6 separate subscales. These subscales are Excluded after test pain, symptoms, physical function in daily living Lost to follow-up (n = 18): Knee pain during test – postponed surgery, 7 (ADL), physical function in sports and recreation n=1 – time and transport, 4 – serious disease unrelatred to PAO, 3 (sporting function), participation in physical activ– injuries unrelated to PAO, 2 ity (participation), and quality of life. HAGOS is – non-union of pubic bone, 1 – emigrated, 1 a reliable, valid, and responsive outcome measure, Healthy volunteers associated with correlation coefficients of 0.2–0.7 Patients for analyses for analyses n = 82 n = 50 across sub-items when correlated to relevant constructs. The measurement error ranges from 1 to 5 Figure 1. Flow of patients and healthy volunteers.: PAO = periacetabular osteotomy. points across sub-items at the group level (Thorborg et al. 2011a, Kemp et al. 2013, Thomeé et al.

2014). Besides HAGOS symptoms and ADL, the sub-items Ethics, registration, funding, and potential conflicts of have a high responsiveness, reported as effect sizes of 1.1–1.9 interest (Thorborg et al. 2011a, Thomeé et al. 2014).

Ethical approval was obtained from the Central Denmark Region Committee on Biomedical Research Ethics (patients: Sample size considerations 5/2014 and volunteers: 252/2018). The Danish Data ProtecAs mentioned earlier, this study is part of a prospective inves- tion Agency authorized patient data handling (1-16-02-47tigation on the same patient population (Jacobsen et al. 2018a, 14), and the study protocol was registered at ClinicalTrials.

2018b, 2018c, 2019). Thus, we did not perform an ordinary gov (20140401PAO). The study was performed in accordance sample size calculation for this study as the numbers of with the Code of Ethics of the World Medical Association patients were fixed when planning this analysis. and the Danish Code of Conduct for Research Integrity. All patients and healthy volunteers gave informed consent to Statistics participate. The authors declare that they have no potential Normally distributed continuous data were reported as means conflicts of interest. This study was kindly supported by the (SD), otherwise reported as medians with either ranges or Danish Rheumatism Association (Grant number A3280), the interquartile ranges (IQR). Categorical data were reported Aase and Ejnar Danielsen Fund (Grant number 10-000761/ as numbers. Differences in hip muscle strength (i.e., flexion, LPJ), and the Fund of Family Kjaersgaard, Sunds (Grant extension, abduction, and adduction) and self-reported out- number 6041401). come (6 HAGOS subscales) between groups and changes from baseline to 1-year follow-up in the patients were analyzed with a mixed-effect model with patients as random factors, and time and group (i.e., patients or healthy volunteers) as Results fixed factors. Model assumptions were based on inspection of 135 patients were assessed for eligibility (Figure 1). 19 plots of standardized residuals versus fitted values and quan- patients declined to participate and 16 patients were excluded tile–quantile (QQ) plots of the standardized residuals. Differ- (Jacobsen et al. 2018b). The patients who declined to particiences were reported as means with 95% confidence intervals pate were, on average, somewhat younger (25 years; SD 9) (CI). Simple linear regression analyses were performed with compared with study participants completing 1-year followhip strength of each muscle group as independent variables up but included a similar percentage of men (10%). Finally, (i.e., flexion, extension, abduction, and adduction), and the 100 consecutive patients fulfilled the in- and exclusion criteria HAGOS subscales covering pain and sporting function as and gave informed consent to participate. Similarly, of 71 elithe dependent variables. Crude β coefficients were reported gible healthy volunteers, 50 volunteers gave informed consent with CI, and the model assumptions of the linear regression to participate. analyses were based on inspection of scatter plots, QQ plots of At 1-year follow-up, 18 patients were lost to follow-up, givthe standardized residuals and plots of standardized residuals ingn 82 patients for the follow-up analyses. The patients lost versus fitted values. The level of significance was 0.05 and the to follow-up were comparable to patients completing 1-year STATA 14.2 (StataCorp, College Station, TX, USA) software follow-up regarding age, BMI, physical activity, NRPS, CE, package was used for the data analysis. and AI angle (data not shown). The median follow-up time

288

Acta Orthopaedica 2021; 92 (3): 285–291

Table 1. Baseline characteristics of patients with hip dysplasia and healthy volunteers. Values are count unless otherwise specified

Characteristics

Patients with hip dysplasia (n = 100)

Mean age (SD) 30 (9) Mean BMI (SD) 23 (3) Men 17 Preferred sports, h/week (IQR) 0 (2) Preferred sports Fitness 28 Running 19 Team sports 13 Gymnastics 7 Horseback riding 6 Racket sports 2 Swimming 2 Dancing 2 Other a 18 No preferred sports 3 General physical activity h/week < 2.5 h/week 13 2.5 to < 5 h/week 19 5 to < 10 h/week 42 ≥ 10 h/week 26 Back pain intensity No 31 Very mild 23 Moderate 26 Fairly severe 14 Very severe 5 Worst imaginable 1 Posterior to anterior spring testing (SP) Hip pain 12 Back pain 35 No pain 53 Positive FADIR test 83 Positive FABER test 74 Positive internal snapping hip test 30 Bilateral affection 89 NRS pain (range) 3 (2–5) Centre-edge angle (°; SD) 17 (5) Tönnis acetabular index angle (°; SD) 14 (5) Osteoarthritis grade 0/1 97/3

Isometric hip muscle strength (Nm/kg) 6 Healthy volunteers Patients before PAO Patients 1 year after PAO

5

Healthy volunteers (n = 50)

4

31 (9) 23 (3) 8 4 (10)

3

2

17 10 12 3 0 2 0 0 6 0

1

0

Flexion

Extension Abduction Adduction

Figure 2. Median isometric hip muscle strength in patients with hip dysplasia and in healthy volunteers in Nm/kg; box represents 25th and 75th percentiles and error bars represent 10th and 90th percentiles.

PAO = periacetabular osteotomy.

was 1.0 year (0.8–1.6). The patients and healthy volunteers were comparable in age, sex, and BMI (Table 1). However, the patients were less physically active and reported more back pain compared with the healthy volunteers.

3 7 30 10 41 7 2 0 0 0

Muscle strength The patients improved their hip muscle strength statistically significantly from before to 1 year after PAO in hip flexion and abduction, whereas muscle strength hip extension and adduction were unchanged (Table 2). Moreover, both pre- and 1-year post-surgery, patients had 13–34% lower hip muscle strength compared with the healthy volunteers, covering all muscle groups (Figure 2, Table 3).

The median pre-surgical pain intensity levels during each muscle strength assessment were 3 (0–10) in hip flexion, 2 (0–10) in hip extension, 4 (0–10) in hip abduction, and 2 (0–10) in hip adduction. The median post-surgical levels were 1 (0–8) in hip flexion, 0 (0–8) in hip extension, 1 (0–8) in hip abduction, and 1 (0–10) in hip adduction.

0 6 44 3 2 0 – – – – –

a Covers

different combat and self-defense sports, bicycling and hiking.

IQR = interquartile range, SP = spinous processes, FADIR = flexion/adduction/internal rotation, FABER = flexion/abduction/external rotation, NRS = numerical rating scale.

Self-reported outcome From before to 1 year after PAO, all HAGOS subscales improved statistically significantly. However, the patients

Table 2. Mean changes in isometric hip muscle strength in Nm/kg Hip strength

Volunteers n mean (SD)

Flexion Extension Abduction Adduction

50 50 50 50

a Statistically

1.8 (0.34) 2.5 (0.63) 1.5 (0.37) 1.5 (0.46)

n

Before PAO mean (SD)

100 97 99 98

1.2 (0.40) a 1.8 (0.68) a 1.2 (0.43) a 1.1 (0.44) a

1 year after PAO n mean (SD) 80 80 81 80

1.3 (0.32) a 1.9 (0.59) a 1.3 (0.44) a 1.1 (0.37) a

HD change HD change mean (95% CI) (%) p-value 0.13 (0.06 to 0.20) 0.09 (–0.05 to 0.23) 0.10 (0.02 to 0.18) –0.03 (–0.11 to 0.06)

significantly lower than the values of the healthy volunteers (p < 0.01).

PAO = periacetabular osteotomy.

11 5 8 –3

< 0.001 0.2 0.02 0.5

Acta Orthopaedica 2021; 92 (3): 285–291

289

Table 3. Mean differences in muscle strength in Nm/kg between 100 patients and 50 volunteers Hip strength Flexion Extension Abduction Adduction

Before PAO Mean diff. (95% CI) % diff. p-value 0.60 (0.48–0.73) 0.66 (0.44–0.87) 0.30 (0.16–0.44) 0.41 (0.27–0.56)

34 27 20 27

< 0.001 < 0.001 < 0.001 < 0.001

1 year after PAO Mean diff. (95% CI) % diff. p-value 0.47 (0.34–0.60) 0.56 (0.34–0.79) 0.20 (0.05–0.35) 0.44 (0.29–0.59)

27 23 13 29

< 0.001 < 0.001 0.008 < 0.001

HAGOS score 100

80

60

40

PAO = periacetabular osteotomy.

20

reported lower self-reported outcome both pre- and 1-year post-surgery compared with the healthy volunteers (Figure 3).

Associations between HAGOS and muscle strength Changes in HAGOS pain and sporting function were positively associated with changes in hip muscle strength (Table 4), where an increase of 1 Nm/kg in hip flexion, extension, and abduction was associated with 9–14-point higher HAGOS score in pain and sporting function.

Discussion We found that patients with hip dysplasia improved their isometric hip muscle strength in hip flexion and abduction by 8–11%. Compared with healthy volunteers, the hip muscle strength of the patients was impaired both pre- and postsurgery, and the pre- to post-surgical changes in hip muscle strength were associated with self-reported outcome.

In a previous study, isometric hip muscle strength was investigated in 22 patients with hip dysplasia undergoing the abductor-sparing approach for PAO (Ezoe et al. 2006). Compared with our results, the pre-surgical hip muscle strength improved somewhat more (0.6–0.8 Nm/kg to 0.8–0.9 Nm/kg).

Moreover, similar to our results, the pre-surgical isometric hip muscle strength was 25–46% lower in all muscle groups than the hip muscle strength of 24 healthy volunteers, and the 12-months muscle strength was 27–31% lower. In another study, isometric muscle strength was investigated pre-, 1-, and 2-year post-surgery in 22 hips undergoing PAO (Sucato et al.

2015). Contraty to our findings, the authors found no pre- to

Healthy volunteers Patients 1 year after PAO Patients before PAO

0 Pain Symptoms

ADL

Sport/rec

PA

QoL

HAGOS subscales

Figure 3. Profile of Copenhagen Hip and Groin Outcome Score (HAGOS) in patients and healthy volunteers in HAGOS points (0–100); error bars represent 95% confidence intervals. PAO = periacetabular osteotomy, ADL = physical function in daily living, Sport/rec = physical function in sports and recreation, PA = participation in physical activity, QoL = quality of life.

2-year post-surgical changes in hip strength and only small insignificant changes from 1 to 2 years. Compared with the 29 healthy volunteers, the hip muscle strength was impaired both pre- and post-surgery. In summary, the results of the previous studies represent divergent findings. However, the previous studies do indicate that hip muscle strength is impaired in patients with hip dysplasia, and that small to moderate changes can be expected 1 year after PAO. No previous studies have investigated associations between changes in self-reported outcome and hip muscle strength.

However, correlations between HAGOS and performancebased outcomes have been investigated in 32 patients with hip dysplasia (Jacobsen et al. 2013). Statistically significant correlations were reported between self-reported pain and sporting function and kinetic gait variables (i.e., hip flexor joint moment measured with a motion-capture system in walking and running). Compared with our results, the correlations were somewhat smaller. Nevertheless, we found that an increase of 1 Nm/kg in hip muscle strength was associated with 9–14 higher HAGOS points, indicating clinically relevant associations. However, previous studies have reported hip muscle strength improvements of only 9–14% after 8 weeks

Table 4. Associations of change in HAGOS to change in muscle strength in HAGOS points/Nm/kg. Values are crude β coefficients with 95% confidence interval (CI) for associations between HAGOS and hip muscle strength measured before and 1 year after periacetabular osteotomy

Outcomes HAGOS pain HAGOS sport

Hip flexion (n = 80) β (CI) p-value 13.4 (0.5 to 26.3) 13.2 (-4.1 – 30.5)

0.04 0.1

Hip extension (n = 78) β (CI) p-value 6.3 (–0.5 to 13.1) 0.07 9.3 (0.6 to 17.9) 0.04

Hip abduction (n = 81) β (CI) p-value 14.0 (3.3 to 24.7) 11.6 (–2.9 to 26.1)

0.01 0.1

HAGOS = Copenhagen Hip and Groin Outcome Score, sport = physical function in sport and recreation.

Hip adduction (n = 79) β (CI) p-value 7.6 (–3.3 to 18.5) 0.9 (–13.5 to 15.4)

0.2 0.9

290

of progressive resistance training (Mortensen et al. 2018), 3 months of hip abductor strengthening (Kuroda et al. 2013), and 6 weeks of task-specific training and progressive hip strengthening (Harris-Hayes et al. 2016). Furthermore, based on our results, a hip muscle strength improvement of 1 Nm/kg would correspond to an improvement of 56%, which is much higher than the higher limits of our and previous reported CIs on improved hip muscle strength, indicating that pre- to postsurgical changes in hip muscle strength were only to a lesser extent associated with changes in patient-reported outcome.

Pre- to post-surgical changes in hip muscle strength may occur due to multiple factors, covering biomechanical improvements, pain reduction, physical rehabilitation etc. 1st, in hip dysplasia, the hip joint is lateralized (Harris et al. 2017), implying that the hip abductors have to generate higher medially directed forces to produce normal movement (Skalshøi et al. 2015, Harris et al. 2017). Therefore, improving the biomechanical conditions with the PAO could possible also change the length tension relationships, which could positively change hip muscle strength. 2nd, pain intensity improved from pre-surgical levels of 2–4 to post-surgical levels of 0–1 using an NPRS, possibly improving the patients’ ability to perform maximum voluntary contractions as previously reported (Thorborg et al. 2011b). 3rd, the patients followed individualized physiotherapy-led rehabilitation for 2–4 months, possibly associated with improved hip muscle strength. On the other hand, self-reported time in preferred physical activities was much lower in the patients compared with the healthy volunteers (Table 1), and the patients were less physical active compared to the healthy volunteers. The lower self-reported physical activity level of the patients may explain why the patients had larger strength deficits compared with the healthy volunteers and why only small improvements in hip muscle strength were found 1-year post-surgery. However, despite biomechanical improvements, pain reduction, and physical rehabilitation, the patients had impaired hip muscle strength 1-year post-surgery, which most likely also compromises function in strenuous activities such as running, stair climbing, jumping, and standing on 1 leg (Graven-Nielsen and Arendt-Nielsen 2008). The direct pathway from impaired muscle strength to reduced ability to perform everyday activities has not yet been established. However, our results indicate an association between hip muscle strength and self-reported hip function in patients rehabilitating from PAO surgery. Moreover, previous studies have suggested that low muscle strength is associated with pathological hip joint biomechanics and reduced dynamic postural control, and that these two are associated with hip osteoarthritis and lower extremity injury (Murphy et al. 2016, Wilson et al. 2018). Therefore, our results may have an implication for future patients as surgeons have the possibility to inform patients about the level of hip muscle strength that can be expected 1 year after PAO and to inform patients of possible consequences. Moreover, the low hip muscle strength 1-year post-surgery implies that future intervention studies

Acta Orthopaedica 2021; 92 (3): 285–291

should focus on improving hip muscle strength after PAO. To our knowledge, no studies have investigated the effect of different physical rehabilitation programs for patients with hip dysplasia.

Methodological considerations No attempt was made to monitor the post-surgical physical rehabilitation; our aim was to investigate pre- and postsurgical hip muscle strength in a setting comparable to usual care. Therefore, the observed post-surgical muscle strength may have been impacted by surgical biomechanical improvement, improved pain, the physical rehabilitation, or some combination of these factors. Furthermore, differences in age and sex could potentially affect estimates of differences between patients and healthy volunteers, as differences in age and gender are associated with muscle strength. However, we managed to control this by recruiting healthy volunteers of the same age and sex as the patients. Hence, we considered it unnecessary to control for this in our analyses. Moreover, the younger age among the patients who declined to participate could not bias our findings as non-participation before baseline test affects only generalizability. Finally, hip muscle strength was not assessed later than 1 year after PAO, and we cannot rule out that the hip muscle strength may continue to improve further. However, Sucato et al. (2015) investigated hip muscle strength pre-, 1, and 2 years post-surgery and found no changes in hip muscle strength post-surgery compared with pre-surgery and only small insignificant differences from 1 to 2 years post-surgery, indicating that the hip muscle strength has plateaued already after 1 year. Of note is that, despite the younger age among the patients who declined to participate, the generalizability of our findings is considered high and a strength of this study as the rate of eligible patients was similar to the general flow of patients at our institution.

Conclusion Isometric hip muscle strength is impaired in patients with symptomatic dysplastic hips measured before PAO. 1 year after surgery, isometric hip flexion and abduction strength had improved but muscle strength did not reach that of healthy volunteers.

All authors took part in the planning of this study. JSJ drafted the manuscript, and all co-authors critically reviewed and edited the manuscript.

The authors would like to thank physiotherapist Charlotte Møller Sørensen and physiotherapist Gitte Hjørnholm Madsen for their invaluable assistance carrying out the clinical examinations. Furthermore, they would like to thank physiotherapist Anne Lida Schiøler Lindhardt, physiotherapist Louise Venborg Eriksen, and physiotherapist Daniel Vestergaard Kristensen for assisting with data collection.

Acta thanks Joanne L Kemp and Veronika Kralj-Iglic for help with peer review of this study.

Acta Orthopaedica 2021; 92 (3): 285–291

Clohisy J C, Ackerman J, Baca G, Baty J, Beaule P E, Kim Y J, Millis M B, Podeszwa D A, Schoenecker P L, Sierra R J, Sink E L, Sucato D J, Trousdale R T, Zaltz I. Patient-reported outcomes of periacetabular osteotomy from the prospective ANCHOR cohort study. J Bone Joint Surg Am 2017; 99(1): 33-41.

Ezoe M, Naito M, Asayama I. Muscle strength improves after abductor-sparing periacetabular osteotomy. Clin Orthop Relat Res 2006; 444: 161-8.

Fairbank J C, Couper J, Davies J B, O’Brien J P. The Oswestry low back pain disability questionnaire. Physiotherapy 1980; 66(8): 271-3.

Graven-Nielsen T, Arendt-Nielsen L. Impact of clinical and experimental pain on muscle strength and activity. Curr Rheumatol Rep 2008; 10(6): 475-81.

Harris-Hayes M, Czuppon S, Van Dillen L R, Steger-May K, Sahrmann S, Schootman M, Salsich G B, Clohisy J C, Mueller M J. Movement-pattern training to improve function in people with chronic hip joint pain: a feasibility randomized clinical trial. J Orthop Sports Phys Ther 2016; 46(6): 452-61.

Harris M D, MacWilliams B A, Bo Foreman K, Peters C L, Weiss J A, Anderson A E. Higher medially-directed joint reaction forces are a characteristic of dysplastic hips: a comparative study using subject-specific musculoskeletal models. J Biomech 2017; 54:80-7.

Jacobsen S, Sonne-Holm S. Hip dysplasia: a significant risk factor for the development of hip osteoarthritis. A cross-sectional survey. Rheumatology 2005; 44(2): 211-8.

Jacobsen J S, Nielsen D B, Sørensen H, Søballe K, Mechlenburg I. Changes in walking and running in patients with hip dysplasia. Acta Orthop 2013; 84(3): 265-70.

Jacobsen J S, Bolvig L, Hölmich P, Thorborg K, Jakobsen S S, Søballe K, Mechlenburg I. Muscle-tendon-related abnormalities detected by ultrasonography are common in symptomatic hip dysplasia. Arch Orthop Trauma Surg 2018a; 138(8): 1059-67.

Jacobsen J S, Hölmich P, Thorborg K, Bolvig L, Jakobsen S S, Søballe K, Mechlenburg I. Muscle-tendon-related pain in 100 patients with hip dysplasia: prevalence and associations with self-reported hip disability and muscle strength. J Hip Preserv Surg 2018b; 5(1): 39-46.

Jacobsen J S, Thorborg K, Hölmich P, Bolvig L, Storgaard Jakobsen S, Søballe K, Mechlenburg I. Does the physical activity profile change in patients with hip dysplasia from before to 1 year after periacetabular osteotomy? Acta Orthop 2018c; 89(6): 622-7.

Jacobsen J S, Søballe K, Thorborg K, Bolvig L, Storgaard Jakobsen S, Hölmich P, Mechlenburg I. Patient-reported outcome and muscle-tendon pain after periacetabular osteotomy are related: 1-year follow-up in 82 patients with hip dysplasia. Acta Orthop 2019; 90(1): 40-5.

Kemp J L, Collins N J, Roos E M, Crossley K M. Psychometric properties of patient-reported outcome measures for hip arthroscopic surgery. Am J Sports Med 2013; 41(9): 2065-73.

Kuroda D, Maeyama A, Naito M, Moriyama S, Yoshimura I, Nakamura Y, Kiyama T. Dynamic hip stability, strength and pain before and after hip abductor strengthening exercises for patients with dysplastic hips. Isokinet Exerc Sci 2013; 21(2): 95-100.

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.

Mechlenburg I, Jørgensen P B, Stentz-Olesen K, Tjur M, Grimm B, Soballe K. Leg power, pelvic movement and physical activity after periacetabular osteotomy: a prospective cohort study. Acta Orthop Belg 2018; 84(2): 163-71.

Mortensen L, Schultz J, Elsner A, Jakobsen S, Søballe K, Jacobsen J, Kierkegaard S, Dalgas U, Mechlenburg I. Progressive resistance training in patients with hip dysplasia: a feasibility study. J Rehabil Med 2018; 50(8): 751-8.

291

Murphy S B, Ganz R, Müller M E. The prognosis in untreated dysplasia of the hip: a study of radiographic factors that predict the outcome. J Bone Joint Surg Am 1995; 77(7): 985-9.

Murphy N J, Eyles J P, Hunter D J. Hip osteoarthritis: etiopathogenesis and implications for management. Adv Ther Springer Healthcare; 2016. p.

1921-46.

Nunley R M, Prather H, Hunt D, Schoenecker P L, Clohisy J C. Clinical presentation of symptomatic acetabular dysplasia in skeletally mature patients.

J Bone Joint Surg Am 2011; 93(Suppl. 2): 17-21.

Schneider M, Erhard R, Brach J, Tellin W, Imbarlina F, Delitto A. Spinal palpation for lumbar segmental mobility and pain provocation: an interexaminer reliability study. J Manipulative Physiol Ther 2008; 31(6): 465-73.

Skalshøi O, Iversen C H, Nielsen D B, Jacobsen J, Mechlenburg I, Søballe K, Sørensen H. Walking patterns and hip contact forces in patients with hip dysplasia. Gait Posture 2015; 42(4): 529-33.

Sucato D J, Tulchin-Francis K, de La Rocha A, Kulkarni V, Podeszwa D A. Improved functional outcome with no decrease in hip strength 2 years following Ganz periacetabular osteotomies for adolescent hip dysplasia. J Pediatr Orthop B 2015; 24(2): 99-105.

Thomeé R, Jónasson P, Thorborg K, Sansone M, Ahldén M, Thomeé C, Karlsson J, Baranto A. Cross-cultural adaptation to Swedish and validation of the Copenhagen Hip and Groin Outcome Score (HAGOS) for pain, symptoms and physical function in patients with hip and groin disability due to femoro-acetabular impingement. Knee Surgery, Sport Traumatol Arthrosc 2014; 22(4): 835-42.

Thorborg K, Petersen J, Magnusson SP, Hölmich P. Clinical assessment of hip strength using a hand-held dynamometer is reliable. Scand J Med Sci Sports 2010; 20(3): 493-501.

Thorborg K, Hölmich P, Christensen R, Petersen J, Roos E M. The Copenhagen Hip and Groin Outcome Score (HAGOS): development and validation according to the COSMIN checklist. Br J Sports Med 2011a; 45(6): 478-91.

Thorborg K, Serner A, Petersen J, Moller Madsen T, Magnusson P, Hölmich P. Hip adduction and abduction strength profiles in elite soccer players: implications for clinical evaluation of hip adductor muscle recovery after injury. Am J Sports Med 2011b; 39(1): 121-6.

Tibor L M, Sekiya J K. Differential diagnosis of pain around the hip joint.

Arthrosc J Arthrosc Relat Surg 2008; 24(12): 1407-21.

Troelsen A, Elmengaard B, Søballe K. A new minimally invasive transsartorial approach for periacetabular osteotomy. J Bone Joint Surg Am 2008; 90(3): 493-8.

Troelsen A, Mechlenburg I, Gelineck J, Bolvig L, Jacobsen S, Søballe K. What is the role of clinical tests and ultrasound in acetabular labral tear diagnostics? Acta Orthop 2009; 80(3): 314-8.

Wasko M K, Yanik E L, Pascual-Garrido C, Clohisy JC. Psychometric properties of patient-reported outcome measures for periacetabular osteotomy. J Bone Joint Surg 2019; 101(6): e21.

Wiberg G. Studies on dysplastic acetabula and congenital subluxation of the hip joint with special references to the complication of osteoarthritis. Acta Chir Scand Suppl 1939; 58: 1-132.

Wilson B R, Robertson K E, Burnham J M, Yonz M C, Ireland M L, Noehren B. The relationship between hip strength and the Y balance test. J Sport Rehabil 2018; 27(5): 445-50.

Wyles C C, Heidenreich M J, Jeng J, Larson D R, Trousdale R T, Sierra R J. The John Charnley Award: Redefining the natural history of osteoarthritis in patients with hip dysplasia and impingement. Clin Orthop Relat Res 2017; 475(2): 336-50.

292

Acta Orthopaedica 2021; 92 (3): 292–296

Good clinical outcome for the majority of younger patients with hip fractures: a Swedish nationwide study on 905 patients younger than 50 years of age Oscar THOORS 1, Carl MELLNER 2, Margareta HEDSTRÖM 1,3 1 Department of Clinical Science, Intervention and Technology (CLINTEC), Karolinska Hospital; 3 Department of Orthopedics, Karolinska Hospital Stockholm, Sweden

Institutet, Stockholm; 2 Department of Orthopedics, Eskilstuna

Correspondence: oscar.thoors@stud.ki.se Submitted 2020-08-25. Accepted 2020-12-14.

Background and purpose — Studies regarding hip fractures in young patients are rare since the patient population is small. We assessed clinical outcomes 4 months after hip fracture in patients < 50 years of age and whether there were differences between sexes and different age groups.

Patients and methods — We included adult patients < 50 years with a hip fracture between January 1, 2014 and December 31, 2018. Baseline data were extracted from the Swedish Registry for Hip Fracture Patients and Treatment (RIKSHÖFT) and mortality data was obtained from Statistics Sweden. The outcome variables were change of walking ability, pain in fractured hip, use of analgesics, living conditions, and mortality rate at 4 months.

Results — Of the 905 patients included, 72% were men and femoral neck fractures were most common (58%).

4 months after surgery, 23% used a walking aid and 7% reported severe pain. Women reported slightly more pain and higher usage of analgesics. Patients aged 40–49 reported higher usage of analgesics than patients aged 15–39, although the latter group reported more pain. Nearly all of those who lived independently before fracture did so at 4 months. The mortality rate was < 1%.

Interpretation — Most patients did not use any walking aid and few had severe pain at 4 months. Furthermore, a hip fracture is not a life-threatening event in a patient < 50 years.

The living conditions did not change for those who lived independently before the fracture.

Numerous patients suffer from disability after a hip fracture and the 4-month mortality rate has recently been reported to be as high as 16% in patients older than 65 years of age (Greve et al. 2020). However, patients with hip fractures do not form a uniform entity. Of all patients with hip fractures, 2–11% are below 50 years of age (Rogmark et al. 2018). Consequently, most studies merely consider clinical outcome and mortality in the elderly and the generalizability to patients less than 50 years is therefore limited. The few previous studies have shown that the outcome of non-elderly patients is not as gruesome as in the elderly, but some report that the outcome in young patients is rather poor; however, in those studies the age limit for the non-elderly was set at less than 60 or 65 years of age and not 50 (Dargan et al. 2016, Ekegren et al. 2016, Rogmark et al. 2018). In this study we describe the < 50 years of age group and assess clinical outcomes 4 months after surgery and compare clinical outcome between sexes and between different age-groups.

Patient and methods Study design This nationwide cohort was based on data from all patients between 15 and 49 years of age, who had been operated on for a hip fracture between January 1, 2014 and December 31, 2018. All data was prospectively registered in the Swedish Registry for Hip Fracture Patients and Treatment (RIKS­ HÖFT).

Source of data and terminology RIKSHÖFT has registered hip fractures patients (> 15 years of age) in Sweden since 1988 and has an estimated coverage of 80–90% for the years studied (Meyer et al. 2020). The date of death was obtained through record linkage with the National

© 2021 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 10.1080/17453674.2021.1876996

Acta Orthopaedica 2021; 92 (3): 292–296

293

Patients younger than 50 years with hip fracture from RIKSHÖFT 2014–2018 n = 932 Excluded Pathological fracture n = 27 Patients included n = 905 Analysis Pain in fractured hip

Analysis Living conditions

Excluded No response to questionnaire 4 months after surgery n = 569

Analyzed n = 336

Analysis Use of pain drugs

Excluded No response to questionnaire, lives not independently before fracture n = 591 Analyzed n = 314

Analysis Use of walking aid

Excluded No response to questionnaire n = 556

Analyzed n = 349

Excluded No response to questionnaire 4 months after surgery, used walking aid before fracure n = 595 Analyzed n = 310

Patients included in the study.

Death Register, Statistics Sweden. Baseline data on all patients included age, sex, waiting time until surgery (hours between arrival at hospital and start of surgery), cognitive function, divided into 3 categories: no cognitive dysfunction; signs of confusion; a diagnosis of dementia. Based on RIKSHÖFT, fracture types were grouped into non-displaced, displaced femoral neck fractures, and basicervical fractures merged to femoral neck fractures (FNF) and into trochanteric and subtrochanteric fractures. Surgical methods were registered as 2 or more screws, sliding hip screw, intramedullary nail, total hip arthroplasty, hemiarthroplasty, or nonoperative treatment. Use of a walking aid was categorized as: no use of walking aid, 1 crutch, 2 crutches, walker, wheelchair. Coming from was categorized as: own home, group/service housing, full-service unit, rehabilitation clinic, emergency hospital, or other. Comorbidity was measured through ASA classification, which was assessed preoperatively by the local anesthesiologist or the local orthopedist on call as part of standard preoperative practice.

All patients either received a questionnaire from the register or were called by phone 4 months after the operation. The following variables were included: use of walking aid or not, living independently or not, pain in fractured hip, and use of analgesics because of the hip fracture. Living independently was defined as patients residing in their own homes, with or without assistance from home care aids. In RIKSHÖFT, pain is assessed in 6 different categories, but was in this study merged into 3: no/transient pain, mild/intermittent pain, and severe/continuous pain. The kind of analgesics used were not specified—simply a yes or no answer.

Patients were divided into subgroups: those aged 15–39 and 40–49, and sexes. Patients who used a walking aid before the fracture were excluded in the analysis of use of walking aid 4 months after surgery. Only patients living at home before the fracture were analyzed regarding their living conditions 4 months after surgery (Figure).

Statistics Descriptive data was presented with means (SD), percentages, and range. Non-normally distributed independent data were tested for differences with a Mann–Whitney–Wilcoxon test.

Contingency tables were used for categorical data and tested for differences using the chi-square test. A p-value < 0.05 was considered statistically significant. Statistical calculations were performed using IBM® SPSS Statistics® for Windows version 25.0 (IBM Corp, Armonk, NY, USA).

Ethics, data sharing, funding, and potential conflict of interests The study was conducted in accordance with the ethical principles of the Helsinki Declaration and approved by the Regional Ethics Committee of Stockholm (DNr: 2017/1088-31). 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 study was supported by grants provided by Region Stockholm (ALF project). The authors declare no conflicts of interest.

Results 905 patients were included in the study (Figure) and represented 1.2% of all hip fractures registered during these years in RIKSHÖFT. The median age was 42 years (15–49) (Table 1) and 72% were men.

Patient and descriptive data The majority (89%) of all patients lived independently before the fracture (Table 1). Women had a lower proportion of ASA-1 compared with men. 12% used a walking aid before the fracture, women more than men (18% and 11%, respectively).

294

Acta Orthopaedica 2021; 92 (3): 292–296

Table 1. Baseline data for all patients younger than 50 years with a hip fracture. Values are count (%) unless otherwise specified Factor

All patients

Men

Women p-value a

Overall sample 905 652 (72) 253 (28) Age, median 42 42 43 range 15–49 15–49 18–49 Age groups 0.5 15–39 360 (40) 264 (40) 96 (38) 40–49 545 (60) 388 (60) 157 (62) Fracture-type 0.07 Cervical (FNF) 528 (58) 365 (56) 163 (64) Trochanteric 234 (26) 179 (27) 55 (22) Subtrochanteric 143 (16) 108 (17) 35 (14) ASA score b < 0.001 1 446 (50) 347 (54) 99 (40) 2 293 (33) 199 (31) 94 (38) 3 135 (15) 91 (14) 44 (18) 4 16 (2) 6 (1) 10 (4) Mental status c 0.4 No cognitive

dysfunction 631 (95) 461 (96) 170 (93) Signs of confusion 29 (4.5) 18 (4.7) 11 (6) Diagnosed with

dementia 4 (0.5) 2 (0.3) 2 (1) Walking-aid 0.04 No use of walking aid 795 (88) 585 (89) 210 (82) 1 crutch 13 (1.5) 22 (3.5) 13 (5) 2 crutches 14 (2) 10 (1.5) 9 (3.5) Walking with walker 27 (3) 19 (3) 14 (6) Wheelchair 48 (5.5) 19 (3) 9 (3.5) Coming from 0.3 Own home 801 (89) 582 (89) 219 (87) Group/service housing 34 (3.8) 19 (3) 15 (6) Full-service unit 16 (2) 12 (2) 4 (1) Rehabilitation clinic 1 (0.1) 1 (0.2) 0 Emergency hospital 41 (4) 28 (4) 13 (5) Other 12 (1.1) 10 (1) 2 (1) a p-values

were calculated using Mann–Whitney U-test for variables on continuous or ordinal scale. Chi-square test for categorical variables. b Missing = 15. c Missing = 241.

The most common type of fracture overall was an FNF (58%) (Table 1). Subtrochanteric fractures were more common in the younger age group compared with the older age group (21% vs. 12%) and the FNF were less common in the younger age group compared with the older age group (55 vs. 61%) (Table 2). 45% were treated with 2 screws or more, 25% were treated with intramedullary nail, 26% were treated with sliding hip screw, 3% were treated with total or hemi­arthroplasty, and 1% received nonoperative treatment.

Clinical outcomes at 4-month follow-up 310 patients were analyzed regarding the use of a walking aid and 336 patients regarding pain in the hip at 4-month followup (Figure, Table 3). Of all patients walking without walking aids before fracture (88%), 77% did not use any walking aid 4 months after surgery. Severe pain was present in 8% of the women compared with 6% in men. Women also used more

Table 2. Baseline data on all patients with a hip fracture aged 15–39 and 40–49 years of age. Values are count (%) Factor

15–39

40–49 p-value a

Sex Men 264 (73) 388 (71) Women 96 (27) 157 (29) Fracture type Cervical (FNF) 198 (55) 330 (61) Trochanteric 86 (24) 148 (27) Subtrochanteric 76 (21) 67 (12) ASA score 1 206 (59) 240 (45) 2 102 (29) 191 (35) 3 39 (11) 96 (18) 4 3 (1) 13 (2) Mental status No cognitive dysfunction 231 (96) 400 (95) Signs of confusion 10 (4) 19 (4) Diagnosed with dementia 0 4 (1) Walking aid No use of walking aid 332 (93) 463 (86) 1 crutch 3 (0.8) 10 (1.8) 2 crutches 2 (0.5) 12 (2.2) Walking with walker 3 (0.8) 24 (4.5) Wheelchair 17 (4.9) 31 (5.5) Coming from Own home 323 (90) 478 (88) Group/service housing 9 (2.5) 25 (4.5) Full-service unit 4 (1) 12 (2) Rehabilitation clinic 0 1 (0.2) Emergency hospital 19 (5) 22 (4) Other 5 (1.5) 7 (1.3) a p-values

0.5 0.002

< 0.001

0.8

0.003

0.4

were calculated using Chi-square test

Table 3. Clinical outcomes 4 months after surgery in hip fracture patients younger than 50 years of age, living independently, and walking without a walking device. Values are count (%) Factor

All

Men

Women p-value a

Walking-aid b 0.2 No use of walking aid 240 (77) 167 (80) 73 (73) Use of walking aid 70 (23) 43 (20) 27 (27) Pain 0.2 No/transient pain 101 (30) 75 (33) 26 (23) Mild/intermittent pain 212 (63) 136 (61) 76 (69) Severe/substantial pain 23 (7) 14 (6) 9 (8) Analgesics because of fracture 0.007 Yes 68 (20) 36 (15) 32 (28) No 281 (80) 197 (85) 84 (72) Living conditions c 0.1 Lives independently 309 (98) 206 (98) 103 (100) Lives not independently 5 (2) 5 (2) 0 (0) Deaths 6 2 4 a p-values were calculated using Chi-square test. b Patients who walked without walking aid before the fracture. c Patients who lived independently before the fracture.

analgesics because of hip pain (28%) compared with men (15%) (Table 3). Of those who lived independently in their own home before the fracture (89%), 98% still lived indepen-

Acta Orthopaedica 2021; 92 (3): 292–296

295

Table 4. Differences between age groups in clinical outcomes 4 months after surgery, use of walking aid and living conditions.

Values are count (%) Factor

All

15–39

40–49 p-value a

Walking aid b 0.09 No use of walking aid 240 (77) 95 (83) 145 (74) Use of walking aid 70 (23) 20 (17) 50 (26) Pain 0.07 No/transient pain 101 (30) 25 (22) 76 (34) Mild/intermittent pain 212 (63) 80 (70) 132 (60) Severe/substantial pain 23 (7) 9 (8) 14 (6) Analgesics because of fracture 0.04 Yes 68 (20) 18 (9) 50 (15) No 281 (80) 179 (91) 281 (85) Living conditions c 0.7 Lives independently 309 (99) 105 (99) 204 (99) Lives not independently 5 (2) 1 (1) 4 (1) Deaths 6 2 4 a-c See

Table 3

dently at 4 months. There was no statistically significant difference between the 2 age groups regarding use of a walking aid. In those aged 15–39, some pain was present in a higher degree compared with patients aged 40–49 (78% and 66%, respectively). Patients aged 40–49 used more analgesics than their younger counterparts (15% versus 9%) (Table 4).

Mortality at 4 months The mortality rate was 0.7%: 2 men (aged 44, 45), and 4 women (aged 31, 39, 2 aged 49) (Tables 3 and 4).

Non-responders A non-response analysis of the differences in baseline data between patients with outcome data and those without showed a higher proportion of men and those aged 15–39 in the latter group (Table 5, see Supplementary data).

Discussion This register-based study identified that 1.2% of all patients with a hip fracture were younger than 50 years of age. The majority did not use any walking devices at 4 months postoperatively, 7% reported severe hip pain, and 20% used analgesics. The living conditions did not change considerably for patients who lived independently before the fracture and less than 1% were deceased 4 months after surgery.

A vast majority of this young age group with hip fracture were men, in accordance with other studies (Al-Ani et al.

2013, Lin et al. 2014, Mattisson et al. 2018), and with fractures in general (Farr et al. 2017). FNF was the most common fracture type, nearly 60%, in contrast to the known even distribution of fracture types in the elderly (RIKSHÖFT annual report 2019). Studies on fracture types in younger patients are scarce, but 2 earlier studies from Scotland and Taiwan also

reported FNF being most common among the youngest (Robinson et al. 1995, Wang et al. 2017).

23% used some walking aids 4 months postoperatively as compared with a previous study on patients aged 65 or younger where the numbers were similar but after 12 months (Dargan et al. 2016). 7% reported severe pain in their fractured hip at 4-month follow-up, slightly more women, and women also used more analgesics. The reason for this difference is unclear; it may be due to sex differences in pain expression (Skogö Nyvang et al. 2019) or coping strategies between sexes (Racine et al. 2012). A higher proportion of the youngest patients aged 15–39 reported more pain in their fractured hip than elderly patients. We found only 1 study that assessed hip pain as a clinical outcome after hip fracture in patients aged 21–56 (Jain et al. 2004). But comparisons with this study are difficult, since they reported that only 6 out of 23 patients had pain 6 months after their hip fracture. The assessment of pain is often included in different evaluation scores, such as the Harris Hip Score, Arnold Evaluation Score and Merle D’aubergine scoring system, without details regarding pain level, thus also making comparisons difficult with those studies (Sprague et al. 2015, Rogmark et al. 2018).

Whether trauma mechanism influenced and explained postoperative pain 4 months after the fracture is unclear; we do not have any data regarding trauma mechanism in this study.

However, 1 recent study showed that the majority of patients aged 20–49 with a hip fracture actually suffered from a lowenergy trauma (fall from standing height or less) (Al-Ani et al. 2013), similar to a Swedish study on younger men with distal radius fractures (Egund et al. 2016). The trauma mechanism behind a hip fracture in the youngest might have changed since 1982, when Zetterberg et al. (1982) reported that highenergy trauma was the leading trauma mechanism In our study, less than 1% of 905 patients were deceased at 4 months, to be compared with 16% in patients > 65 years with a hip fracture (Söderqvist et al. 2009, Greve et al. 2020).

Other studies have shown a similar low mortality rate after a hip fracture in young patients (Robinson et al. 1995, Lin et al.

2014).

Living conditions are considered to be a good measure in overall recovery after a hip fracture and have a positive impact on the quality of life (Boelhouwer 2002). In this study, we found that 98% of those living independently before fracture had return to their own living arrangements 4 months after surgery, contrary to the situation for many elderly patients (RIKSHÖFT annual report 2019).

Strength and limitations One strength is the prospective nationwide design based on RIKSHÖFT, including a larger number of patients. Another is the many variables included in the baseline data, which provided a broad picture of the young patient with a hip fracture.

The weakness of this study is the low response rate to the questionnaire regarding clinical outcomes. However, other

296

Swedish national registries have a similar low response rate to questionnaires (Swedish National Knee Ligament Registry, 2018, Swedish Fracture Register, 2019). The low response rate may have introduced a selection bias, thus reducing the external validity of this study, although a non-response analysis showed that the only differences were that men and patients aged 15–39 had a lower response rate (Table 5, see Supplementary data).

In conclusion, most young patients with a hip fracture had good walking function at 4 months and few reported severe pain. The mortality rate was low and living conditions for those living independently before the fracture did not change substantially.

Supplementary data Table 5 is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2021.

1876996

Study design: MH, OT. Data acquisition: MH, OT. Analysis: MH, OT, CM. Interpretation of the data: MH, OT, CM. Drafting and revision of the manuscript: OT, MH, CM. The authors would like to acknowledge and appreciate the help given to them from RIKSHÖFT. Acta thanks Charles Court-Brown and Jan-Erik Gjertsen for help with peer review of this study.

Al-Ani A N, Neander G, Samuelsson B, et al. Risk factors for osteoporosis are common in young and middle-aged patients with femoral neck fractures regardless of trauma mechanism. Acta Orthop 2013; 84(1): 54-9. doi: 10.3109/17453674.2013.765639.

Boelhouwer J. Quality of life and living conditions in the Netherlands. Social Indicators Research 2002; 58(1): 113-38. doi: 10.1023/a:1015779732321.

Dargan D P, Callachand F, Diamond O J, et al. Three-year outcomes of intracapsular femoral neck fractures fixed with sliding hip screws in adults aged under sixty-five years. Injury 2016; 47(11): 2495-500. doi: 10.1016/j. injury.2016.09.013.

Egund L, McGuigan F, Önnby K, et al. High prevalence of osteoporosis in men with distal radius fracture: a cross-sectional study of 233 men. Calcif Tissue Int 2016; 99(3): 250-8. doi: 10.1007/s00223-016-0142-6.

Ekegren C L, Edwards E R, Page R, et al. Twelve-month mortality and functional outcomes in hip fracture patients under 65 years of age. Injury 2016; 47(10): 2182-8. doi: 10.1016/j.injury.2016.05.033.

Acta Orthopaedica 2021; 92 (3): 292–296

Farr J N, Melton L J 3rd, Achenbach S J, et al. Fracture incidence and characteristics in young adults aged 18 to 49 years: a population-based study. J Bone Miner Res 2017; 32(12): 2347-54. doi: 10.1002/jbmr.3228.

Greve K, Modig K, Talbäck M, et al. No association between waiting time to surgery and mortality for healthier patients with hip fracture: a nationwide Swedish cohort of 59,675 patients. Acta Orthop 2020: 1-6. doi: 10.1080/17453674.2020.1754645.

Jain P, Maini L, Mishra P, et al. Cephalomedullary interlocked nail for ipsilateral hip and femoral shaft fractures. Injury 2004; 35(10): 1031-8. doi: 10.1016/j.injury.2003.09.039.

Lin J C-F, Wu C-C, Lo C, et al. Mortality and complications of hip fracture in young adults: a nationwide population-based cohort study. BMC Musculoskelet Disord 2014; 15(1): 362. doi: 10.1186/1471-2474-15-362.

Mattisson L, Bojan A, Enocson A. Epidemiology, treatment and mortality of trochanteric and subtrochanteric hip fractures: data from the Swedish fracture register. BMC Musculoskelet Disord 2018; 19(1): 369. doi: 10.1186/ s12891-018-2276-3.

Meyer A C, Hedström M, Modig K. The Swedish Hip Fracture Register and National Patient Register were valuable for research on hip fractures: comparison of two registers. J Clin Epidemiol 2020; 125: 91-9. doi: 10.1016/j. jclinepi.2020.06.003.

Racine M, Tousignant-Laflamme Y, Kloda L A, et al. A systematic literature review of 10 years of research on sex/gender and pain perception—part 2: do biopsychosocial factors alter pain sensitivity differently in women and men? Pain 2012; 153(3): 619-35. doi: 10.1016/j.pain.2011.11.026.

RIKSHÖFT. Available at http://rikshoft.se/ [in Swedish]; 2019.

Robinson C M, Court-Brown C M, McQueen M M, et al. Hip fractures in adults younger than 50 years of age: epidemiology and results. Clin Orthop Relat Res 1995; (312): 238-46.

Rogmark C, Kristensen M T, Viberg B, et al. Hip fractures in the non-elderly: who, why and whither? Injury 2018; 49(8): 1445-50. doi: 10.1016/j. injury.2018.06.028.

Skogö Nyvang J, Naili J E, Iversen M D, et al. Younger age is associated with greater pain expression among patients with knee or hip osteoarthritis scheduled for a joint arthroplasty. BMC Musculoskelet Disord 2019; 20(1): 365. doi: 10.1186/s12891-019-2740-8.

Sprague S, Slobogean G P, Scott T, et al. Young femoral neck fractures: are we measuring outcomes that matter? Injury 2015; 46(3): 507-14. doi: 10.1016/j.injury.2014.11.020.

Söderqvist A, Ekström W, Ponzer S, et al., Stockholm Hip Fracture G. Prediction of mortality in elderly patients with hip fractures: a two-year prospective study of 1,944 patients. Gerontology 2009; 55(5): 496-504. doi: 10.1159/000230587.

Swedish Fracture Register. Available at www.frakturregistret.se; 2019.

Swedish National Knee Ligament Registry, Available at www.aclregister.nu; 2018.

Wang M T, Yao S H, Wong P, et al. Hip fractures in young adults: a retrospective cross-sectional study of characteristics, injury mechanism, risk factors, complications and follow-up. Arch Osteoporos 2017; 12(1): 46. doi: 10.1007/s11657-017-0339-y.

Zetterberg C H, Irstam L, Andersson G B. Femoral neck fractures in young adults. Acta Orthop Scand 1982; 53(3): 427-35. doi: 10.3109/ 17453678208992237.

Acta Orthopaedica 2021; 92 (3): 297–303

297

Good results at 2-year follow-up of a custom-made triflange acetabular component for large acetabular defects and pelvic discontinuity: a prospective case series of 50 hips Marieke SCHARFF-BAAUW 1–3, Miranda L VAN HOOFF 1, Gijs G VAN HELLEMONDT 1, Paul C JUTTE 2, Sjoerd K BULSTRA 2, and Maarten SPRUIT 1 1 Orthopaedic

Department, Sint Maartenskliniek, Nijmegen; 2 Orthopaedic Department, University Medical Centre Groningen; 3 Orthopaedic Department, Medisch Centrum Leeuwarden, The Netherlands Correspondence: mbaauw@gmail.com Submitted 2020-09-24. Accepted 2020-12-15.

Background and purpose — Custom triflange acetabular components (CTACs) are suggested as good solutions for large acetabular defects in revision total hip arthroplasty.

However, high complication rates have been reported and most studies are of limited quality. This prospective study evaluates the performance of a CTAC in patients with large acetabular defects including pelvic discontinuity.

Patients and methods — Prospectively collected data of 49 consecutive patients (50 hips), who underwent an acetabular revision with a CTAC were analyzed. Follow-up (FU) was 2 years. The median age of the patients was 68 years (41–89) and 41 were women. Primary outcomes were re-revision of the CTAC and differences between the modified Oxford Hip Score (mOHS) preoperatively and at 2-year follow-up. Secondary outcomes included several patientreported outcomes (PROMs), radiological results, complications, and a comparison between hips with and without pelvic discontinuity (PD).

Results — 1 patient (1 hip) was lost to the 2-year FU. No CTAC needed re-revision. The preoperative and 2-year FU mOHS were available in 40 hips and improved statistically significantly. All of the other secondary outcomes improved over time. 5 hips (of 45 with radiological 2-year FU) had loosening of screws. 8 hips had complications, including 3 persistent wound leakage, 3 pelvic fractures, and 1 dislocation. The mOHS and complication rate were similar in hips with and without PD. Interpretation — Reconstruction of large acetabular defects with and without PD with this CTAC showed good improvement in patient-reported daily functioning, high patient-reported satisfaction, few complications, and no rerevisions at 2-year FU.

Acetabular revision is challenging when facing severe host bone loss and poor remaining bone quality. Pelvic discontinuity (PD) increases the difficulty of reconstructing such defects.

Custom triflange acetabular components (CTATCs) have been repeatedly suggested as good solutions to deal with large acetabular defects, even when PD is present (Sheth et al.

2013, Baauw et al. 2016, De Martino et al. 2019, Szczepanski et al. 2019, Volpin et al. 2019, Chiarlone et al. 2020, Malahias et al. 2020). A proposed advantage is the ability to customize and individualize the implant to the defect in each individual case (Berasi et al. 2015). As such, an immediately stable initial implant fixation might be accomplished. This might be due to restoring anatomical dimensions and re-distributing load anatomically, choosing the optimal center of rotation, and supporting host bone contact and osseointegration. We feel that good design of the CTAC prior to surgery, trying to achieve implant support and fixation to the best host bone quality, is important as the implant cannot be modified intraoperatively.

A disadvantage of the use of CTACs is the reported high complication rate in terms of reoperation, infection, nerve damage, and especially dislocation (Volpin et al. 2019, Chiarlone et al.

2020, Malahias et al. 2020). However, these higher rates may relate to the difficulty of revisions and severity of the acetabular bone defects encountered when using CTACs (De Martino et al. 2019, Volpin et al. 2019). As might be expected, the risk of postoperative hip dislocation is increased in these complex cases with multiple previous surgeries, extensile approaches, pre-existent leg-length discrepancies, and frequently abductor weakness (De Martino et al. 2019). An option to reduce dislocation in revision total hip arthroplasty (THA) is by using a dual mobility design (Faldini et al. 2018) and its imple-

© 2021 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 10.1080/17453674.2021.1885254

298

A

Acta Orthopaedica 2021; 92 (3): 297–303

B

Figure 1. Planning of case 17 with (A) the ultimate acetabular bone defect after subtracting all parts of the existing reconstruction and (B) the expected postoperative situation with the complete construct.

Figure 2. Dual mobility cup cemented into the custom-made implant.

A = (place of) dual mobility cup. B = triflange cage. C = porous metal augment.

mentation has been recommended in acetabular revision with CTACs (De Martino et al. 2019, Malahias et al. 2020).

The use of CTACs remains controversial as many studies that evaluate the performance of these implants are retrospective small case series and as such of limited quality. There is a need for prospective studies with consistent reporting of clinical, radiological, and patient-reported outcomes.

This prospective single-center study evaluates the revision rate, patient-reported outcomes, complications, and postoperative radiographs in a consecutive series of patients with large acetabular defects treated with a CTAC in which either a dual mobility cup or a constrained liner was cemented.

mining optimal anteversion, inclination, and center of rotation of the implant. Based on this information and feedback a porous metal augment and a triflange cage, with flanges on ilium, ischium, and pubis, were designed as a monoblock, with screw fixation planned into the best host bone quality (Figure 1). All patients were operated on by an orthopedic surgeon and either another orthopedic surgeon, a fellow, or a final-year resident. A posterolateral approach was used in all patients and surgeons had a printed hemi-pelvis, trial implants, and drill guides at their disposal during surgery. Allograft was used in case of voids and/or cavitary defects between host bone and implant. Taking into account the quality of the host bone, the implant was fixed with pre-planned trajectory screws using the patient-specific drill guides. Within the implant either a dual mobility cup (48 hips) or, in the case of abductor deficiency, a constrained liner (2 hips) was cemented in the same orientation as the implant (Figure 2). Further details concerning the acetabular defect analyses and the surgical technique have previously been described (Baauw et al. 2015, 2017). Postoperatively, patients were allowed 50% weight-bearing on the operated leg for the first 6 weeks. Systemic antibiotics were routinely used perioperatively and until results of intraoperative cultures were known and low-molecular-weight heparin (LMWH) was administered in the first 6 weeks postoperatively.

Patients and methods Prospectively collected data (questionnaires) of 49 consecutive patients (50 hips) was extracted and anonymized from the institution’s THA revision database. Inclusion criteria were an acetabular revision with a custom-made acetabular revision system (Materialise, Leuven, Belgium) and a minimum of 2 years’ follow-up. The study complied with the STROBE guidelines (von Elm et al. 2008).

The indication for the CTAC was the presence of a Paprosky type 3B acetabular defect (Paprosky et al. 1994) with or without PD in a patient for whom other options with off-the-shelf implants were not thought feasible.

Surgery Patients were operated on between February 2013 and September 2017. A preoperative CT scan was performed for defect analyses and reconstruction planning. The surgeons gave feedback on the defect analyses and the implant orientation, deter-

Patients Of the 49 included patients (50 hips), 41 were women. At the time of the hip revision surgery the median (range) age of the patients was 68 years (41–89) and their median (range) BMI was 27 (19–44). The ASA classification was 2 in most patients (30/50). The primary diagnosis was osteoarthritis (OA) in 26 patients, 41 were revised due to aseptic loosening, and the median (range) number of previous revisions was 2 (1–9).

Acta Orthopaedica 2021; 92 (3): 297–303

Based on preoperative analysis pelvic discontinuity (PD) was found in 16 hips. In 11 hips the stem was revised at the same time and bone graft was used in 32 hips. 2 patients (case 21 and 48) received a constrained liner instead of a dual mobility because of hip abductor deficiency. The median (range) time that patients stayed in hospital was 8 (4–28) days (Table 1, see Supplementary data).

Primary and secondary outcomes Our primary outcomes were re-revision of the CTAC at 2-year FU and the change in daily functioning as experienced by patients. To measure daily functioning the patient-reported modified Oxford Hip Score (mOHS) was used (Gosens et al.

2005). The preoperative mOHS (70–14) was compared with the mOHS at 2-year FU and its clinical relevance was analyzed. At 2-year FU we also looked at the mean mOHS of all available patients, including those who did not complete the mOHS preoperatively.

Secondary clinical outcomes included a comparison between preoperative and 2-year FU values of the EuroQol 5 dimensions 3 level (EQ5D-3L) utility (-0.329–1), the EQ5D3L numeric rating scale (NRS) from 0–100 (EuroQol group 1990), and the visual analogue scale (VAS) for pain at rest and during activities (0–100). At 2-year FU the following additional clinical outcomes were measured: satisfaction with surgical result using VAS (0–100) and several core questions, which could be answered “yes” or “no.”

Complications were registered during admission and until 2-year FU and all types of complications were registered.

Anteroposterior (AP) radiographs were taken at 1-year FU and 2-year FU. These were reviewed by MSB and MS for: notable breakage of the component, screw loosening (defined by radiolucency around the screws) or breakage, and bony fractures.

Finally, to explore and indicate the potential influence of PD, the re-revision rate, mOHS, and the complications in cases with PD were compared with cases without PD. Statistics The primary outcome, the mOHS, was descriptively summarized, using medians and ranges, and non-parametrically tested with the Wilcoxon signed-rank test to evaluate clinical performance preoperatively versus the performance at 2-year FU. Clinical relevance of the change in mOHS was assessed using a distribution-based approach. This was calculated by taking 0.5 SD of the mean difference between the preoperative scores and the scores at 2-year FU. To further substantiate clinical relevance, the effect size was determined using Cohen’s d, which is calculated by dividing the difference in scores from preoperative to 2-year FU by the SD of the preoperative scores (Norman et al. 2003, Copay et al. 2007). An effect size of 0.2 was considered small, 0.5 moderate, and 0.8 large (Cohen 1992). The secondary clinical outcome data was descriptively summarized using medians and ranges.

299

Missing cases for the primary outcome, the mOHS, were compared to complete cases on baseline characteristics (age, sex, BMI, primary diagnosis, number of previous revisions, stem revision, and use of bone graft and presence of PD) using the Wilcoxon signed-rank test for continuous data and Fisher’s exact test for categorical data. Statistical analyses were performed using STATA (version 13.1 for Windows; StataCorp, College Station, TX, USA). Statistical significance was defined as p < 0.05.

Ethics, funding, and potential conflict of interests Ethical approval from the Institutional review board was not required, as the Dutch Act on Medical Research involving Human Subjects does not apply to screening questionnaires that are part of routine clinical practice. For this study, patient data were obtained as a part of routine outcome monitoring for use in daily practice. All data were anonymized and identified for analyses and report.

Personal fees were received for faculty work from Materialise by MSB, GGvH, and MS, from Smith & Nephew by GGvH, from Zimmer Biomet by GGvH, and from DePuy Synthes by MS. SKB is the president of the Dutch Orthopedic Society and MS is chairman of the AOTK Spine.

Results Primary outcomes 1 patient (1 hip) was lost to the 2-year FU (case 49) and did not respond to questionnaires or follow-up appointments due to her comorbidities. None of the remaining 49 CTACs needed re-revision at 2-year FU. The mOHS was missing in 7 cases at preoperative assessment (cases 10, 18, 24, 37, 38, 39, 50) and in 3 cases at 2-year FU (cases 21, 25, 49). In the remaining 39 patients (40 hips) with complete mOHS a statistically significant improvement was shown from 51 (24–67) to 28.5 (14–56) at the 2-year FU. The clinically relevant difference (0.5 SD) was 5 points and present in 37 out of 40 patients with complete mOHS. The effect size was large (d = 1.6). The mOHS of all available patients (n = 47) at 2-year FU, irrespective of (in-)complete baseline mOHS, was 29 (14–56).

Patients who had incomplete data for the mOHS differed statistically significantly from patients with complete data with regard to the number of previous revisions: 3.5 (1–9) previous revisions in patients with incomplete mOHS and 2 (1–9) in patients with complete mOHS. No other significant differences in baseline characteristics were shown between complete and incomplete cases,.

Patient-reported clinical results Our secondary outcome measures on EQ5D-3L utility, EQ5D3L NRS, VASrest, and VASactivity improved between baseline and 2-year FU (Table 2). For these values we had 41/400 (10%) missing values.

300

Acta Orthopaedica 2021; 92 (3): 297–303

Table 2. Patient-reported outcomes in medians (ranges)

Preoperative n score

2-year FU n score

EQ5D-3L utility EQ5D-3L NRS VASrest VASactivity

44 43 45 45

47 0.77 (–0.20 to 1) 44 70 (40–100) 46 2 (0–100) 46 11.5 (0–100)

0.23 (– 0.13 to 0.89) 50 (7–100) 31 (0–100) 78 (0–100)

EQ5D-3L, EuroQol 5 dimensions 3 level, range –0.329 to 1.

NRS, numeric rating scale, range 0–100.

VAS, visual analog scale, range 0–100.

A

Table 3. Core questions at 2-year follow-up Core question (n = 47)

Yes

Has the operation improved the mobility or function of the hip?

Has the pain in/around the hip lessened since the operation?

Are you satisfied with the results of the operation?

Would you recommend the operation to a family member or friend?

38 45 42 47

Complications In 49 cases the complication registration was complete. 8 cases had complications. Of these, 3 cases had re-explorations for persistent wound discharge (cases 6, 16, and 28), with collection of intraoperative cultures. In 1 of these cases (case 16) cultures were found to be positive, which was treated with 3 months of antibiotics. During the re-exploration of this same case 3 loose ischium screws and 1 loose pubis screw were exchanged.

In 3 other cases a fracture of the pelvis (cases 2, 27, and 45) occurred, 2 postoperatively and 1 stress fracture after 6 months. These 3 cases were treated conservatively. The stress fracture evolved into a pseudoarthrosis; the other 2 fractures healed. At 3 weeks postoperatively, in another case a hip dislocated (case 3), which was treated conservatively with closed reduction and a brace and the hip did not dislocate again at the 2-year FU. This patient had ischiatic nerve irritation due to the dislocation. In the 8th case with complications, a general complication occurred, which involved a cerebrovascular accident directly postoperatively (case 26).

The rates of mOHS and complications were similar in patients with and without PD (Table 4).

B

Figure 3. Case 17 (A) preoperatively and (B) at 2-year follow-up.

Satisfaction with the surgical result was reported in 45 cases and was 96 (0–100). The results of the core questions are described in Table 3.

Radiological results AP radiographs were available of 49 hips at 1-year FU and of 45 hips at 2-year FU (Figure 3). 5 hips had loosening of screws at 1-year FU with no signs of progression at 2-year FU (cases 10, 31, 32, 38, and 42). In all of these patients screw loosening was found in 1 or more ischium screws and in one of these hips there was also screw loosening of a pubis screw (case 10) (Figure 4). The missing 4 hips at 2-year FU (case 16, 41, 43, and 50) did not show any complications at 1-year FU.

Table 4. Clinical and patient-reported outcomes of hips with and without PD

mOHS preoperative mean mOHS postoperative Overall clinical complication rate Dislocation rate

No PD (n = 33)

PD (n = 16)

52 (24–69) 28 (14–48) 5 1

53 (25–60) 32 (17–56) 3 0

PD, pelvic discontinuity; mOHS, modified Oxford Hip Score.

Acta Orthopaedica 2021; 92 (3): 297–303

A

301

B

C

Figure 4. Case 10 (A) preoperatively, (B) at 1-year follow-up, and (C) at 2-year follow-up.

Discussion To our knowledge this is the 1st prospective study on a large group of patients with this particular custom-made implant and in which pre- and postoperative patient-reported clinical outcome scores are compared (Colen et al. 2013, Baauw et al.

2017, Myncke et al. 2017). This study is the 2nd prospective case series, deBoer et al. (2007) being the 1st on the results of any CTAC for large acetabular defects. Furthermore, patient satisfaction is evaluated in more detail compared with most studies on CTACs and it is the 1st that reports on the clinical relevance of the improvement in patient-reported functioning over time (De Martino et al. 2019, Chiarlone et al. 2020).

In this study, all of the clinical patient-reported outcome scores improved over time, which is consistent with other studies on CTACs (De Martino et al. 2019, Chiarlone et al.

2020). The improvement in the mOHS between preoperatively and 2-year FU was also found to be clinically relevant.

When comparing our study with 2 recent review articles on CTACs, the revision rate, overall reoperation rate, and the complication rate were lower in our study (De Martino et al.

2019, Chiarlone et al. 2020). In particular, our low dislocation rate (1/49) is notable. Risks of a high dislocation rate in revision THA include multiple previous hip revisions (Kosashvili et al. 2011), abductor muscle deficiency, and severe acetabular bone loss (Faldini et al. 2018), all of which are often present in hips that are managed with a CTAC, the current study included. Another risk factor is the revision of only 1 com-

ponent (Faldini et al. 2018), which was the case in 39/50 of the hip revisions in the current study. We believe that the low number of dislocations in our study is related to the preoperative planning of implant anteversion, with the use of either a dual mobility design or a constrained liner, in the case of abductor deficiency, in all of our cases (Faldini et al. 2018).

This assumption is supported by 2 other studies on CTACs that reported no dislocations and either used a dual mobility cup in all cases (Colen et al. 2013) or a constrained liner in most of their cases (Berasi et al. 2015). To our knowledge, only 2 other studies have measured the accuracy of the placement of their custom-made implant (Weber et al. 2019, ­Zampelis and Flivik 2020). Both of them found similar good placement accuracy, as we have previously found (Baauw et al. 2016), and had 1 and 0 dislocations in 11 and 10 patients, highlighting the importance of accurate placement to diminish the dislocation rate.

Another notable finding in our study is the low deep infection rate, 1 of 49. Known risk factors for deep infections after total hip arthroplasty include an ASA score of 3 or higher, a longer duration of surgery (Urquhart et al. 2010) and a higher number of previous revisions (Kosashvili et al. 2011). In our patients the median (range) previous revisions were 2 (1–9) and 6 patients had an ASA classification of 3. However, the 1 patient with a deep infection (case 16) had an ASA classification of 2 and had 2 previous revisions. We did not report on the surgical time, but we assume this was relatively short compared with other hip revision surgeries because all operations were performed by 2 orthopedic surgeons and because

302

of the precise preoperative planning. Other factors that might explain our low infection rate are the following measurements that are routinely done in all THA revisions in our clinic: preoperative infection workup with lab work and intra-articular aspiration, the routine use of antibiotics perioperatively for at least 24 hours, intraoperative betadine lavage and irrigation, and finally meticulous wound closure and low-suction wound dressing in patients with a BMI of over 30.

When comparing revisions with PD and without PD we found similar results. This is in line with findings of 2 recent review articles on the treatment of PD that have found CTACs to be a viable treatment option (Szczepanski et al. 2019, Malahias et al. 2020). In our study there were no mechanical failures and no dislocations and the overall complication rate was 3 out of 16 in cases with PD. These results are favorable, not only compared with other studies on CTACs for PD but also when compared with other treatment options for PD, including cup-cages, anti-protrusion-cages, acetabular shells with plates, and pelvic distraction techniques (Szczepanski et al.

2019, Malahias et al. 2020).

There are some limitations in this study. 1st, the relatively short FU of 24 months. The average FU was found to be 5 years (range 1–18) in previous studies on CTACs (De Martino et al. 2019, Chiarlone et al. 2020). We will continue to follow up our patients. Another limitation is the fact that we cannot comment on the migration of the implant, which is difficult to determine for this particular implant on conventional radiographs. Recently, Zampelis and Flivik (2020) have determined the migration of a similar implant, same cage but without an augment, at 1-year follow-up using CT scans. They found small measured migration values of less than 1 degree or 1 mm. To determine the secondary stability of these implants in the long run new CT-based migration research will be necessary.

In conclusion, this CTAC used in large acetabular defects with and without PD demonstrates a relevant improvement in patient-reported daily functioning, high patient-reported satisfaction, few complications and no re-revisions at 2-year FU. Supplementary data Table 1 is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2021.

1885254

MSB: Study design, data collection, data analyzing, and writing draft and final versions of the paper. MLvH: Study design, statistical analysis, proofreading draft, and final versions of the paper. GGvH: Performed surgeries and proofreading of the final paper. PCJ: Proofreading of the final paper.

SKB: Study design and proofreading of the final version of the paper. MS: Study design, performed surgeries, and proofreading draft and final versions of the paper.

Acta thanks Harald Brismar and Gunnar Flivik for help with peer review of this study.

Acta Orthopaedica 2021; 92 (3): 297–303

Baauw M, van Hellemondt G G, van Hooff M L, Spruit M. The accuracy of positioning of a custom-made implant within a large acetabular defect at revision arthroplasty of the hip. Bone Joint J 2015; 97-B: 780-5.

Baauw M, van Hooff M L, Spruit M. Current construct options for revision of large acetabular defects: A systematic review. J Bone Joint Surg Rev 2016; Nov 8: 4.

Baauw M, van Hellemondt G G, Spruit M. A custom-made acetabular implant for Paprosky type 3 defects. Orthopedics 2017; 40: e195-e198.

Berasi C C 4th, Berend K R, Adams J B, Ruh E L, Lombardi A V Jr. Are custom triflange acetabular components effective for reconstruction of catastrophic bone loss? Clin Orthop Relat Res 2015; 473(2): 528-35.

Chiarlone F, Zanirato A, Cavagnaro L, Alessio-Mazzola M, Felli L, Burastero G. Acetabular custom-made implants for severe acetabular bone defect in revision total hip arthroplasty: a systematic review of the literature. Arch Orthop Trauma Surg 2020; 140: 415-24.

Cohen J. A power primer. Psychol Bull 1992; 112: 155-9.

Colen S, Harake R, De Haan J, Mullier M. A modified custom-made triflange acetabular reconstruction ring (MCTARR) for revision hip arthroplasty with severe acetabular defects. Acta Orthop Belg 2013; 79(1): 71-5.

Copay A G, Subach B R, Glassman S D, Polly D W Jr, Schuler T C. Understanding the minimum clinically important difference: a review of concepts and methods. Spine J 2007; 7: 541-6.

DeBoer D K, Christie M J, Brinson M F, Morrison J C. Revision total hip arthroplasty for pelvic discontinuity. J Bone Joint Surg [Am] 2007; 89-A: 835-40.

De Martino I, Strigelli V, Cacciola G, Gu A, Bostrom M P, Sculco P K. Survivorship and clinical outcomes of custom triflange acetabular components in revision total hip arthroplasty: a systematic review. J Arthroplasty 2019; 34: 2511-18.

EuroQol Group. EuroQol: a new facility for the measurement of healthrelated quality of life. Health Policy 1990; 16(3): 199-208.

Faldini C, Stefanini N, Fenga D, Neonakis E M, Perna F, Mazzotti A, Pilla F, Triantafyllopoulos I K, Traina F. How to prevent dislocation after revision total hip arthroplasty: a systematic review of the risk factors and a focus on treatment options. J Orthop Traumatol 2018; 19: 17.

Gosens T, Hoefnagels N H, de Vet R C, Dhert W J, van Langelaan EJ, Bulstra S K, Geesink R G. The “Oxford Hip Score”: the translation and validation of a questionnaire into Dutch to evaluate the results of total hip arthroplasty. Acta Orthop 2005; 76(2): 204-11.

Kosashvili Y, Backstein D, Safir O, Lakstein D, Gross A E. Dislocation and infection after revision total hip arthroplasty: comparison between the first and multiply revised total hip arthroplasty. J Arthroplasty 2011; 26(8): 1170-5.

Malahias M A, Ma Q L, Gu A, Ward S E, Alexiades M M, Sculco P K. Outcomes of acetabular reconstructions for the management of chronic pelvic discontinuity: a systematic review. J Arthroplasty 2020; 35: 1145-53.e2.

Myncke I, van Schaik D, Scheerlinck T. Custom-made triflanged acetabular components in the treatment of major acetabular defects: short-term results and clinical experience. Acta Orthop Belg 2017; 83(3): 341-50.

Norman G R, Sloan J A, Wyrwich K W. Interpretation of changes in healthrelated quality of life: the remarkable universality of half a standard deviation. Med Care 2003; 41: 582-92.

Paprosky W G, Perona P G, Lawrence J M. Acetabular defect classification and surgical reconstruction in revision arthroplasty: a 6-year follow-up evaluation. J Arthroplasty 1994; 9: 33-44.

Sheth N P, Nelson C L, Springer B D, Fehring T K, Paprosky W G. Acetabular bone loss in revision total hip arthroplasty: evaluation and management. J Am Acad Orthop Surg 2013; 21: 128-39.

Szczepanski J R, Perriman D M, Smith P N. Surgical treatment of pelvic discontinuity: a systematic review and meta-analysis. J Bone Joint Surg Rev 2019; 7: e4.

Urquhart D M, Hanna F S, Brennan S L, Wluka A E, Leder K, Cameron P A, Graves S E, Cicuttini F M. Incidence and risk factors for deep surgical site infection after primary total hip arthroplasty: a systematic review. J Arthroplasty 2010; 25(8): 1216-22.e1-3.

Acta Orthopaedica 2021; 92 (3): 297–303

Volpin A, Konan S, Biz C, Tansey R J, Haddad F S. Reconstruction of failed acetabular component in the presence of severe acetabular bone loss: a systematic review. Musculoskelet Surg 2019; 103:1-13. von Elm E, Altman D G, Egger M, Pocock S J, Gøtzsche P C, Vandenbroucke J P; STROBE Initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE)statement: guidelines for reporting observational studies. J Clin Epidemiol 2008; 61(4): 344-9.

303

Weber M, Witzmann L, Wieding J, Grifka J, Renkawitz T, Craiovan B. Customized implants for acetabular Paprosky III defects may be positioned with high accuracy in revision hip arthroplasty Int Orthop 2019; 43(10): 2235-43.

Zampelis V, Flivik G. Custom-made 3D-printed cup-cage implants for complex acetabular revisions: evaluation of pre-planned versus achieved positioning and 1-year migration data in 10 patients. Acta Orthop 2020 Sep; 15; 1-6. Online ahead of print.

304

Acta Orthopaedica 2021; 92 (3): 304–310

International variation in distribution of ASA class in patients undergoing total hip arthroplasty and its influence on mortality: data from an international consortium of arthroplasty registries Alan J SILMAN 1, Christophe COMBESCURE 2, Rory J FERGUSON 1, Stephen E GRAVES 3, Elizabeth W PAXTON 4, Chris FRAMPTON 5, Ove FURNES 6,7, Anne Marie FENSTAD 6, Gary HOOPER 8, Anne GARLAND 9, Anneke SPEKENBRINK-SPOOREN 10, J Mark WILKINSON 11,12, Keijo MÄKELÄ 13, Anne LÜBBEKE 1,14, and Ola ROLFSON 15,16 1 Nuffield

Department of Orthopaedics Rheumatology and Musculoskeletal Sciences, University of Oxford, UK; 2 Division of Clinical Epidemiology, Geneva University Hospitals, Switzerland; 3 Australian Orthopaedic Association National Joint Replacement Registry, Australia; 4 Surgical Outcomes and Analysis, Kaiser Permanente, San Diego, USA; 5 Department of Medicine, University of Otago, Christchurch, New Zealand; 6 The Norwegian Arthroplasty Register, Department of Orthopedic Surgery, Haukeland University Hospital, Bergen, Norway; 7 Department of Clinical Medicine, University of Bergen, Norway; 8 Orthopaedic Surgery and Musculoskeletal Medicine, University of Otago, Christchurch, New Zealand; 9 Department of Orthopaedics, Visby lasarett Institute of Surgical Scienses, Uppsala Universitet, Uppsala, Sweden ; 10 Dutch Arthroplasty Register, ’s-Hertogenbosch, The Netherlands; 11 Department of Oncology and Metabolism, University of Sheffield, Sheffield, UK; 12 National Joint Registry for England, Wales, Northern Ireland, and Isle of Man, London, UK; 13 Department of Orthopaedics and Traumatology, Turku University Hospital and University of Turku, Turku, Finland; 14 Division of Orthopaedics and Trauma Surgery, Geneva University Hospitals, Switzerland; 15 Department of Orthopaedics, Institute of Clinical Sciences Sahlgrenska Academy, University of Gothenburg, Sweden; 16 The Swedish Hip Arthroplasty Register, Gothenburg, Sweden Correspondence: alan.silman@gtc.ox.ac.uk Submitted 2020-05-07. Accepted 2020-12-20.

Background and purpose — A challenge comparing outcomes from total hip arthroplasty between countries is variation in preoperative characteristics, particularly comorbidity. Therefore, we investigated between-country variation in comorbidity in patients based on ASA class distribution, and determined any variation of ASA class to mortality risk between countries.

Patients and methods — All arthroplasty registries collecting ASA class and mortality data in patients with elective primary THAs performed 2012–2016 were identified.

Survival analyses of the influence of ASA class on 1-year mortality were performed by individual registries, followed by meta-analysis of aggregated data.

Results — 6 national registries and 1 US healthcare organization registry with 418,916 THAs were included. There was substantial variation in the proportion of ASA class III/ IV, ranging from 14% in the Netherlands to 39% in Finland.

Overall, 1-year mortality was 0.93% (95% CI 0.87–1.01) and increased from 0.2% in ASA class I to 8.9% in class IV. The association between ASA class and mortality measured by hazard ratios (HR) was strong in all registries even after adjustment for age and sex, which reduced them by half in all registries. Combined adjusted HRs were 2.0, 6.1, and 22 for ASA class II–IV vs. I, respectively. Associations were moderately heterogeneous across registries.

Interpretation — We observed large variation in ASA class distribution between registries, possibly explained by differences in background morbidity and/or international variation in access to surgery. The similar, strong mortality trends by ASA class between countries enhance the relevance of its use as an indicator of comorbidity in international registry studies.

In recent years, there has been a growing interest in comparing outcomes of total hip arthroplasty (THA) between arthroplasty registries, including rates of revisions and complications, and patient-reported benefits of surgery (Paxton et al. 2011, McGrory et al. 2016, Hughes et al. 2017, Springer et al. 2017, Paxton et al. 2018). Comparing aggregate-level registry data internationally allows examination of variation in practice and outcomes due to differences in implant use, populations, and healthcare system. However, these evaluations also have limitations. Registry populations may substantially differ in patients’ preoperative characteristics and therefore may not be directly comparable when assessing outcome. As an example, comorbidity is an important predictor of outcomes of THA, including perioperative mortality and severe complications (Weaver et al. 2003, Rauh

© 2021 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 10.1080/17453674.2021.1892267

Acta Orthopaedica 2021; 92 (3): 304–310

and Krackow 2004), patient-reported benefits (Judge et al.

2013, Greene et al. 2015) and the need for revision surgery (Hooper et al. 2012, Prokopetz et al. 2012). It is possible that there are differences in comorbidity level between patients undergoing surgery in different countries because of differences in population health (e.g. burden of cardiovascular disease), in health systems, and in how the former may influence access to surgery. There are, however, few published data on population dissimilarities in pre-existing comorbidity (Franklin et al. 2017). Comparisons of outcome might therefore require controlling for the population differences in the statistical analysis by stratifying or adjusting for such patient characteristics. This requires a consistent approach to the definition and the measurement of a possible confounder such as comorbidity.

Comorbidity is a multi-dimensional phenomenon that reflects the overall health status of a patient. It is strongly associated with mortality in patients undergoing THA and is an ideal candidate for adjustment in registry analyses of mortality. Several methods to measure comorbidity exist.

The scoring systems vary in the type and detail of information they require (Bjorgul et al. 2010, Inacio et al. 2015).

The most widely collected comorbidity system by arthroplasty registries is the ASA classification system (Lübbeke et al. 2018). The simplicity of the score underpins its widespread use, although several studies have shown variability among anesthesiologists in assigning ASA score (Ranta et al. 1997, Mak et al. 2002, Riley et al. 2014, Sankar et al.

2014). Differences in ASA class distribution and its association with mortality may arise from underlying population health variation such as obesity and cardiovascular disease prevalence, and differences in access to healthcare/surgery.

Finally, differences in registry populations (e.g., age and sex), independent of comorbidity, are susceptible to modification of the distribution of ASA class and its association with mortality.

Our objectives are therefore (i) to investigate the extent of variation in the distribution of ASA class in patients undergoing THA between arthroplasty registries internationally; (ii) to explore how far any variation identified is related to other routinely collected demographic data, specifically age and sex; and (iii) to investigate the consistency between ASA class and death within the first year after surgery between the registries studied.

Patients and methods Design We conducted an analysis of aggregated data prospectively collected from participating arthroplasty registries. The distribution of elective primary THAs by ASA class was compared between registries. The influence of ASA class on 1-year mortality after elective primary THA was investigated.

305

Patients and data sources Arthroplasty registries were eligible to take part in this study if they were full members of the International Society of Arthroplasty Registries and collected data on ASA class of patients.

To ensure comparability between registries, given the variable start date on which registries collected ASA class, we restricted inclusion to cases that were elective primary THAs performed within the period January 1, 2012 to December 31, 2016. THAs for which the indication was trauma or malignancy were excluded. Only the first THA in each patient was included. For included cases, the following patient characteristics were extracted: age, sex, BMI, diagnosis (primary or secondary osteoarthritis [OA]), and ASA class. Data on death from any cause within 1-year of THA was obtained.

Statistics Patient -level analysis was performed by the individual registries following a standardized protocol. The aggregated data from each registry was subsequently analyzed centrally.

Individual registries analysis All registries described the distribution of baseline characteristics using frequencies and proportions. Registries calculated the cumulative incidence of mortality at 1 year after index THA, both overall and by ASA class, using Kaplan–Meier survival estimates with 95% confidence intervals (CIs). Patients were censored at loss to follow-up or end of 1 year follow-up.

Registries then investigated the association between ASA class and risk of 1 year mortality with Cox proportional hazards models (presented as hazard ratios [HRs] with CIs) with ASA class I defined as the referent category. The proportionality of hazards was checked visually by plotting log–log of survival against time. Regression coefficients of the Cox models were reported with their variance–covariance matrix. The univariable model was accompanied by a multivariable model adjusting for age and sex. We did not anticipate a non-linear effect of age or an interaction between age and sex, but as a test of these assumptions we did repeat the adjusted analyses, to allow for these possibilities in the 3 largest registries: Sweden, the Netherlands, and Australia. Complete case analysis was used for adjusted models.

Aggregate analysis across registers Kaplan–Meier tables submitted by each participating registry were combined to create a summary life table of mortality up to 1 year. For this purpose, effective numbers of at-risk patients and estimates of mortality at intervals during follow-up were collected from each registry. The conditional mortality estimates from each registry were derived and combined using the DerSimonian model with random effects (Combescure et al.

2014) The summary mortality estimates were obtained by the product-limit of the conditional mortality estimates.

The regression coefficients of the Cox models from the individual registries’ analysis, both univariate and multivari-

306

Acta Orthopaedica 2021; 92 (3): 304–310

Table 1. Patient demographics by registry. Values are count (%) unless otherwise specified

Factor Australia Finland

Kaiser Permanente

Netherlands

New Zealand

Norway

Sweden

Age category < 55 15,306 (13) 2,726 (12) 13,176 (30) 10,866 (9.2) 6,304 (16) 3,750 (13) 7,025 (11) 55–64 27,580 (24) 5,356 (24) 15,485 (35) 24,960 (21) 10,510 (27) 6,930 (23) 14,273 (22) 65–74 39,978 (35) 8,214 (37) 7,704 (18) 44,703 (38) 13,054 (34) 11,008 (37) 25,228 (38) 75–84 25,907 (23) 5,318 (24) 6,018 (14) 32,657 (28) 7,429 (19) 6,877 (23) 16,447 (25) ≥ 85 5,610 (4.9) 801 (3.6) 1,355 (3.1) 5,150 (4.4) 1,165 (3.0) 1,274 (4.3) 2,742 (4.2) Missing 0 0 33 155 0 0 0 Age, mean (SD) 67 (12) 67 (11) 66 (11) 69 (12) 67 (11) 68 (11) 68 (11) BMI category < 18.5 475 (0.9) 118 (0.6) 351 (0.8) 603 (0.8) 222 (0.8) ND 491 (0.8) 18.5–24.9 11,384 (22) 5,093 (27) 9,370 (22.3) 24,730 (31) 5,974 (21) ND 19,931 (31) 25–29.9 19,296 (37) 7,983 (42) 15,318 (36.5) 34,449 (43) 10,748 (38) ND 27,664 (43) 30–34.9 12,812 (25) 4,353 (23) 10,762 (25.6) 14,920 (19) 6,981 (25) ND 12,275 (19) 35–39.9 5,350 (10) 1,379 (7.2) 4,893 (11.6) 3,906 (4.9) 3,181 (11) ND 3,173 (4.9) ≥ 40 2,636 (5.1) 328 (1.7) 1,327 (3.2) 1,000 (1.3) 1,054 (3.7) ND 570 (0.9) Missing 62,428 3,161 1,750 38,883 10,302 ND 1,611 BMI, mean (SD) 29.4 (6.2) 28.2 (4.8) 29.1 (5.6) 27.4 (4.5) 29.0 (5.6) ND 27.3 (4.4) Sex Women 60,923 (53) 12,724 (57) 25,148 (57) 77,566 (66) 20,345 (53) 18,961 (64) 37,148 (57) Missing 0 10 30 248 0 0 0 Diagnosis (primary/secondary OA) Primary OA 107,480 (94) 19,336 (90) 41,599 (95) 107,230 (91) 36,080 (94) 24,075 (81) 60,466 (92) Missing 0 1,020 0 0 0 87 0 ND, no data

ate, were combined also using the DerSimonian model with random effects for multivariate analyses (Jackson et al. 2010). With DerSimonian and Laird’s model, the logarithms of HRs are combined across studies by calculating a weighted average. The weight of a study depends on the precision of the estimated HR: the higher the precision, the higher the weight of the study. The advantage is that studies with a larger sample size tend to have a larger weight in the meta-analysis. In addition, the between-studies variability is accounted for in the calculation of the weights.

The extension of DerSimonian and Laird’s model for multivariate analyses has been used to account for the correlation between the combined HRs. Cochran Q tests and I2 statistics were used to assess the heterogeneity across registries as described previously (Higgins et al. 2003). I2 values of 25%, 50%, and 75% indicate low, moderate, and high levels of heterogeneity, respectively.

Statistical analyses were performed with R 4.0.2 for Windows (R Foundation for Statistical Computing, Vienna, Austria).

The R package mvmeta (v1.0.3) (Gasparrini et al. 2012) was used to combine logarithm of hazard ratios and the R package MetaSurv (v0.3) (Combescure et al. 2014) to combine mortality curves. The 2-sided statistical threshold for significance was 0.05.

Ethics, funding, and potential conflicts of interest As this was a study of anonymized aggregated data, with no individual patient data passed to the researchers, there were no ethical issues and consent was not necessary. All data col-

lection and analysis was funded by the core funding to the authors’ institutions and organizations.

The authors have no conflicts of interest to declare

Results International variation 7 registries were included in the study. 6 were national and population-based (Australia, Finland, Netherlands, New Zealand, Norway, and Sweden) and 1 was from a healthcare organization (Kaiser Permanente, USA).

The baseline demographic data varied by registry (Table 1).

The proportion of women ranged from 53% in Australia to 66% in the Netherlands. Only 9% of patients were aged under 55 years in the Netherlands compared with 30% in Kaiser Permanente. The proportion of obese patients (BMI ≥ 30) ranged from 25% in the Netherlands and Sweden to 40% in Australia, Kaiser Permanente, and New Zealand. The proportion of THAs performed for primary OA ranged from 81% in Norway to 95% in Kaiser Permanente.

Substantial variation in the proportion in each ASA class was observed (Table 2). The Netherlands and Sweden had the lowest proportions of ASA class III–IV (14% and 17% respectively), while Australia, Finland, and Kaiser Permanente had twice those proportions (34%, 39%, and 35% respectively).

Over all registries, the percentage of patients aged under 55 years decreased with ASA class (from 27% in ASA class I to 4.1% in ASA class IV) and the percentage of patients 85 years

Acta Orthopaedica 2021; 92 (3): 304–310

307

Table 2. ASA class by registry. Values are count (%)

Factor Australia Finland

Kaiser Permanente

Netherlands

New Zealand

Norway

Sweden

Total

ASA class I 11,092 (10) 2,773 (13) 1,390 (3.3) 23,750 (20) 5,972 (16) 4,586 (16) 15,116 (23) 64,679 (15) ASA class II 57,616 (55) 10,515 (48) 26,634 (62) 77,245 (66) 23,039 (61) 19,319 (65) 38,483 (60) 252,851 (60) ASA class III 33,969 (33) 8,076 (37) 14,212 (33) 16,586 (14) 8,718 (23) 5,600 (19) 10,793 (17) 97,954 (23) ASA class IV 1,676 (1.6) 339 (1.6) 499 (1.2) 286 (0.2) 238 (0.6) 111 (0.4) 283 (0.4) 3,432 (0.8) Any ASA class 104,353 21,703 42,735 117,867 37,967 29,616 64,675 418,916 Missing 10,028 712 1,036 624 495 223 1,040 14,158 Total 114,381 22,415 43,771 118,491 38,462 29,839 65,715 433,074

Age distribution (%) – ASA class I

ASA class IV

ASA class III

ASA class II

Australia Finland Kaiser Permanente Netherlands Norway Sweden 0

25

50

75

100 0

25

50 < 55

75

100 0

55–46

65–74

25

50

75–84

75

100 0

25

50

75

100

≥ 85

Distribution of age (years) in registries by ASA class.

or older increased (from 0.7% in ASA class I to 17% in ASA class IV) (Figure). Although these broad age patterns were observed in all registries, there were substantial differences between registries in the actual proportions of patients within the same ASA class. The percentage of patients aged under 55 years in ASA class I ranged from 22% (Netherlands) to 43% (Finland), and the percentage of patients over 85 years in ASA class IV ranged from 10% (Kaiser Permanente) to 20% (Australia).

varied. Whereas mortality rates for classes I and II were almost identical between registries, there was modest variation between registries in those in classes III ranging from 1.3% to 3.1%. Variation was most extensive in class IV (range 4.5% to 16%).

Unadjusted and age- and sex-adjusted HRs for the association between ASA class and 1-year mortality are given in Table 4. The pooled unadjusted HR confirmed an increasing 1-year mortality with increasing ASA class. This rose from an HR of 3.2 (CI 2.3–4.3) when comparing ASA class II with I, to a substantially higher HR of 59 (38–95) when comparing ASA class IV with I. There was moderate heterogeneity in the individual registries (I2 around 50%), with lowest HRs in Kaiser Permanente and the Netherlands. As an example, the unadjusted HRs in Netherlands were half those in Sweden and Finland.

ASA and mortality within 1 year after THA Across all registries combined, the overall mortality was 0.93% (CI 0.87–1.01). This rose with increasing ASA class, from 0.18, 0.52, 2.2, to 8.9%, respectively from class I to class IV. Although this trend was observed in all individual registries (Table 3), ASA class-specific 1-year mortality

Table 3. 1-year mortality in percentage (95% confidence interval) by registry and 1-year mortality combined across registries

ASA I ASA II Australia Finland Kaiser Permanente Netherlands New Zealand Norway Sweden Combined

0.10 (0.04–0.16) 0.12 (0.00–0.26) 0.25 (0.00–0.54) 0.30 (0.23–0.37) 0.15 (0.05–0.25) 0.15 (0.03–0.28) 0.16 (0.10–0.23) 0.18 (0.12–0.25)

1-year mortality (%; CI) ASA III

0.46 (0.40–0.52) 0.40 (0.27–0.53) 0.37 (0.29–0.45) 0.70 (0.64–0.76) 0.54 (0.45–0.64) 0.46 (0.35–0.56) 0.71 (0.63–0.80) 0.52 (0.43–0.64)

1.7 (1.5–1.8) 1.9 (1.6–2.2) 1.3 (1.1–1.5) 2.7 (2.5–3.0) 2.6 (2.3–3.0) 3.1 (2.6–3.6) 2.5 (2.2–2.8) 2.2 (1.8–2.7)

ASA IV 6.8 (5.6–8.0) 7.6 (4.7–11) 4.5 (2.6–6.5) 9.4 (5.9–13) 13 (9.0–18) 16 (8.0–23) 8.3 (5.0–12) 8.9 (6.7–12)

All 0.92 (0.87–0.98) 1.02 (0.89–1.2) 0.73 (0.64–0.81) 0.92 (0.87–0.98) 1.04 (0.94–1.1) 0.96 (0.84–1.1) 0.93 (0.86–1.1) 0.93 (0.87–1.0)

308

Acta Orthopaedica 2021; 92 (3): 304–310

Table 4. Meta-analysis of unadjusted and age- and sex-adjusted hazard ratios for the association between ASA classes and 1-year mortality

ASA I as reference Registries Australia Finland Kaiser Permanente Netherlands New Zealand Norway Sweden Pooled HR P-value a Heterogeneity Q Cochran Cochran test (p) I2 (%) a

ASA II

Unadjusted HR (CI) ASA III

ASA IV

4.5 (2.5–8.2) 17 (9.0–30) 69 (38–129) 3.3 (1.0–11) 16 (5.0–50) 69 (21–229) 1.5 (0.5–4.7) 5.3 (1.7–17) 19 (6.0–64) 2.2 (1.7–2.8) 8.8 (6.9–11) 32 (21–50) 3.6 (1.8–7.1) 18 (9.0–34) 97 (47–204) 3.1 (1.3–7.1) 21 (10–48) 120 (47–309) 4.4 (2.9–6.6) 16 (10–24) 54 (30–95) 3.2 (2.3–4.3) 14 (10–19) 59 (38–93) < 0.001 < 0.001 < 0.001

11 14 14 0.06 0.03 0.03 50 57 58

ASA II

Adjusted HR (CI) ASA III

2.7 (1.4–4.9) 6.9 (3.8–13) 2.0 (0.6–6.4) 6.8 (2.1–22) 1.0 (0.3–3.2) 2.7 (0.9–8.4) 1.4 (1.1–1.8) 4.2 (3.2–5.4) 2.3 (1.2–4.5) 8.0 (4.0–16) 1.7 (0.7–3.9) 7.7 (3.3–18) 3.0 (2.0–4.6) 8.6 (5.6–13) 2.0 (1.4–2.7) 6.1 (4.4–8.5) < 0.001 < 0.001

12 13 0.06 0.05 51 53

ASA IV 22 (12–41) 24 (7.0–84) 7 (2.0–25) 14 (9.0–22) 34 (16–74) 34 (13–91) 28 (16–50) 22 (15–32) < 0.001 10 0.11 42

p-value for testing the null hypothesis that the pooled HR equals 1.

After age and sex adjustment the HRs were lowered by half: both the pooled as well as registry-level HRs. However, not all of the effect of ASA on mortality could be captured by this adjustment: the increases in the pooled HRs with increasing ASA classes were attenuated but there were still 2.0-fold, 6.1-fold, and 22-fold (CI as shown) increases in mortality in ASA classes II, III, and IV respectively compared with class I. Although all registries showed this trend, there was moderate heterogeneity across registries, and I2 statistics were around 50% with flatter rises in Kaiser Permanente and in the Netherlands compared with the other registries. As mentioned in the methods, we had assumed the effect of age would be linear.

To test this assumption, we repeated the analysis in the 3 largest registries (Australia, Sweden, and Netherlands) introducing a non-linear effect of age and an age–sex interaction term.

Results (not shown) did not modify sensitively the adjusted HRs for ASA classes. As these 3 registries represent around 70% of the patients included in our study we have no reason to suspect that these findings are not generalizable to the other registries.

Discussion First, we have shown there was a large variation in the distribution of ASA class in patients undergoing THA. Second, given the demographic differences between registries, there were also differences in age distribution of the registry populations within ASA classes. The third conclusion relates to the association between ASA classes and 1-year mortality. Across all registries, worsening ASA class was associated with greater 1-year mortality but the magnitude of the unadjusted relationship differed between registries. Age and sex adjustment was only able to capture about half to two-thirds of the

impact of ASA class on mortality, though this was consistent.

After adjustment a moderate between-registry heterogeneity between ASA classes and mortality remained.

There are a number of methodological issues to consider in interpreting these findings, some of which could lead to bias in the results. The underlying aim of the study was to investigate the extent of variation in comorbidity of patients undergoing THA in different countries. Because ASA class is widely accepted as a useful guide to postoperative mortality and complications, and it is the only measure routinely collected by the registries, it was used as the proxy for comorbidity (Rauh and Krackow 2004, Hackett et al. 2015, Visser et al. 2015). Allocation of ASA class in clinical practice is subject to inter-rater variation, and as such to random error (Sankar et al. 2014).

This adds noise to comparisons between populations, making it more difficult to detect true underlying differences. Despite this potential for underestimation of variation, substantial differences in ASA class were observed between the populations covered by these registries.

There are likely to be between-country differences in scoring, and specifically issues related to local rules about the availability of surgery or “upcoding”: at different times and in different jurisdictions, there may have been advantages or disadvantages for healthcare providers to under- or over-estimate the ASA class to support their service. These issues might lead to inaccuracies in ASA class allocation, the extent or direction of which is unknown.

There are other methodological issues to consider in interpreting these findings. First, there was a modest amount of missing data on ASA (< 2% overall), but with completeness at 90% or greater in the registries studied it is unlikely to be important. There were also minor differences in the number of years with available data in the different registries, but not of sufficient magnitude to be concerned about secular changes.

Acta Orthopaedica 2021; 92 (3): 304–310

Third, we used meta-analytic approaches, which are commonly used in clinical research to combine studies, to analyze survival data pooled across registries. The advantage of these approaches is that they can detect heterogeneity between studies. Therefore, they were appropriate to investigate the consistency of the association between ASA classes and mortality over registries.

ASA class has substantial international acceptance as a useful measure of morbidity to identify those with the greatest hazards following surgery. Thus, it is important to consider the possible explanations for the substantial differences in ASA distribution. In addition to scoring differences there are 2 other principle groups of reasons, which are (i) underlying population differences in general health covering those factors that would be reflected in the ASA class and (ii) differences in the healthcare systems that either encourage or discourage surgery in those with greater underlying health problems.

There is no published data on ASA class in the general populations in these countries, given that the tool is used only in patients selected for surgery. It is therefore relevant to consider other data to suggest that there are country differences in general health. The higher ASA classes particularly focus on cardiovascular disease (CVD). The Global Burden of Disease initiative publishes data on CVD deaths by country (Institute for Health Metrics and Evaluation 2017). The latter rate does vary between the countries covered and appears to be broadly related to the data we observed on ASA class. Thus, from the 2 countries with the highest and the lowest ASA class III proportion, Finland and the Netherlands, Finland has the highest CVD mortality and the Netherlands the lowest of the countries included.

In addition, there may be differences in healthcare coverage.

The Kaiser Permanente registry is a healthcare organization registry and an exception in that the other registries are all national in their scope and during the period of the data collection > 90% complete in their population coverage. Kaiser Permanente showed the lowest proportion of ASA class I, the lowest 1-year mortality in ASA classes III and IV, and the lowest age- and sex-adjusted HRs. The main difference in the data from Kaiser Permanente is in the (lesser) effect of ASA class on mortality. This might suggest that being in ASA class III in Kaiser Permanente represents a healthier cohort than being in the same class in populations with universal healthcare coverage as Kaiser Permanente only covers the subgroup that enrolled in its healthcare plan (Wilper et al. 2009). However, the ASA distribution in KP is consistent with other US reports. A further potential concern with Kaiser Permanente is that, not being population based, attrition could be higher than in the national registers. However, Kaiser has a lower attrition rate than might be expected. Unlike most US healthcare systems, Kaiser has a 100% capture rate of patients who have an arthroplasty and a very low rate of members who leave the system, which are tracked through their membership enrolment. Over 19 years, only 8% of all the patients in KP’s registry have left the system. Other, more subtle factors can

309

influence mortality risk, which cannot be easily captured in an analysis such as this (Woolhandler and Himmelstein 2017).

There are also a few potential confounders to consider. We specifically did not adjust for BMI, given the role of obesity in ASA assignment (Mak et al. 2002). Adjustment for BMI could have masked the effect of comorbidity that was the underlying aim of the study. There may be other unknown confounders that could have been adjusted for, such as socioeconomic status. However, the relationship between socioeconomic status (SES) and ASA grade is complex. There are clear associations between SES and multimorbidity (Barnett et al. 2012) and it is likely, because of this link, that, after adjustment for SES, differences between countries in ASA may be attenuated.

In conclusion, there are substantial differences in ASA class distribution between the national registries, for which the most plausible explanation is between-country differences in the underlying health status and healthcare access, as well as in scoring. The similar and strong mortality trends by ASA class between countries enhances the relevance of the use of ASA class as an indicator of comorbidity in international registry studies.

All authors contributed to the design of the study, the interpretation of the results, and the content of the final manuscript. AJS and CC made equal contributions as first authors. SG, KM, EP, GC, CF, GH, OF, AF, AG, AL, ASS, MW, and OR analyzed their local data and contributed such analyses for the aggregated data. CC was responsible for the design and conduct of the statistical analysis of the aggregated data. AS, RF, AL, and OR initiated the design and protocol for the study, oversaw the data collection, and with CC jointly drafted the manuscript.

Acta thanks Max Gordon and Alexander David Liddle for help with peer review of this study.

Barnett K, Mercer S W, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross sectional study. Lancet 2012; 38 (9836): 37-43.

Bjorgul K, Novicoff W M, Saleh K J. Evaluating comorbidities in total hip and knee arthroplasty: available instruments. J Orthop Traumatol 2010; 11: 203-9. doi: 10.1007/s10195-010-0115-x Combescure C, Foucher Y, Jackson D. Meta-analysis of single-arm survival studies: a distribution-free approach for estimating summary survival curves with random effects. Stat Med 2014; 33: 2521-37. doi: 10.1002/ sim.6111 Franklin P D, Miozzari H, Christofilopoulos P, Hoffmeyer P, Ayers D C, Lübbeke A. Important patient characteristics differ prior to total knee arthroplasty and total hip arthroplasty between Switzerland and the United States. BMC Musculoskelet Disord 2017; 18: 14. doi: 10.1186/s12891016-1372-5.

Gasparrini A, Armstrong B, Kenward M G. Multivariate meta-analysis for non-linear and other multi-parameter associations. Stat Med 2012; 31(29): 3821-39. doi: 10.1002/sim.5471.

Greene M E, Rolfson O, Gordon M, Garellick G, Nemes S. Standard comorbidity measures do not predict patient-reported outcomes 1 year after total hip arthroplasty. Clin Orthop Relat Res 2015; 473: 3370-9. doi: 10.1007/ s11999-015-4195-z Hackett N J, De Oliveira G S, Jain U K, Kim J Y S. ASA class is a reliable independent predictor of medical complications and mortality following surgery. Int J Surg 2015; 18:184-90. doi: 10.1016/j.ijsu.2015.04.079

310

Higgins J P T, Thompson S G, Deeks J J, Altman D G. Measuring inconsistency in meta-analyses. BMJ 2003; 327(7414): 557-60.

Hooper G J, Rothwell A G, Hooper N M, Frampton C. The relationship between the American Society of Anesthesiologists physical rating and outcome following total hip and knee arthroplasty: an analysis of the New Zealand Joint Registry. J Bone Joint Surg Am 2012; 94: 1065-70. doi: 10.2106/JBJS.J.01681 Hughes R E, Batra A, Hallstrom B R. Arthroplasty registries around the world: valuable sources of hip implant revision risk data. Curr Rev Musculoskelet Med 2017; 10: 240-52. doi: 10.1007/s12178-017-9408-5 Inacio M C S, Pratt N L, Roughead E E, Graves S E. Comparing co-morbidities in total joint arthroplasty patients using the RxRisk-V, Elixhauser, and Charlson Measures: a cross-sectional evaluation. BMC Musculoskelet Disord 2015; 16: 385. doi: 10.1186/s12891-015-0835-4 Institute for Health Metrics and Evaluation. Global Burden of Disease. 2017.

Available from: http://www.healthdata.org/gbd (accessed March 19, 2019).

Jackson D, White I R, Thompson S G. Extending DerSimonian and Laird’s methodology to perform multivariate random effects meta-analyses. Stat Med 2010; 29: 1282-97. doi: 10.1002/sim.3602 Judge A, Arden N K, Batra R N, Thomas G, Beard D, Javaid M K, Cooper C, Murray D. 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: e002453.

Lübbeke A, Silman A J, Barea C, Prieto-Alhambra D, Carr A J. Mapping existing hip and knee replacement registries in Europe. Health Policy (New York) 2018; 122: 548-57. doi: 10.1016/j.healthpol.2018.03.010 Mak P H K, Campbell R C H, Irwin M G, American Society of Anesthesiologists.

The ASA Physical Status Classification: inter-observer consistency. American Society of Anesthesiologists. Anaesth Intensive Care 2002; 30: 633-40.

McGrory B J, Etkin C D, Lewallen D G. Comparing contemporary revision burden among hip and knee joint replacement registries. Arthroplasty Today 2016; 2: 83-6. doi: 10.1016/j.artd. 2016.04.003 Paxton E W, Furnes O, Namba R S, Inacio M C S, Fenstad A M, LHavelin L I. Comparison of the Norwegian Knee Arthroplasty Register and a United States Arthroplasty Registry. J Bone Joint Surg Am 2011; 93: 20-30. doi: 10.2106/JBJS.K.01045

Acta Orthopaedica 2021; 92 (3): 304–310

Paxton E W, Mohaddes M, Laaksonen I, Lorimer M, Graves S E, Malchau H, Namba R S, Kärrholm J, Rolfson O, Cafri G. Meta-analysis of individual registry results enhances international registry collaboration. Acta Orthop 2018; 89: 369-73. doi: 10.1080/17453674.2018.1454383 Prokopetz J J, Losina E, Bliss R L, Wright J, Baron J A, Katz J N. Risk factors for revision of primary total hip arthroplasty: a systematic review. BMC Musculoskelet Disord 2012; 13: 251. doi: 10.1186/1471-2474-13-251 Ranta S, Hynynen M, Tammisto T. A survey of the ASA physical status classification: significant variation in allocation among Finnish anaesthesiologists. Acta Anaesthesiol Scand 1997; 41: 629-32. doi: 10.1111/j.13996576.1997.tb04755.x.

Rauh M A, Krackow K A. In-hospital deaths following elective total joint arthroplasty. Orthopedics 2004; 27: 407-11.

Riley R H, Holman C D J, Fletcher D R. Inter-rater reliability of the ASA Physical Status Classification in a aample of anaesthetists in Western Australia. Anaesth Intensive Care 2014; 42: 614-8. doi: 10.1177/0310057X1404200511 Sankar A, Johnson S R, Beattie W S, Tait G, Wijeysundera D N. Reliability of the American Society of Anesthesiologists physical status scale in clinical practice. Br J Anaesth 2014; 113: 424-32. doi: 10.1093/bja/aeu100 Springer B D, Cahue S, Etkin C D, Lewallen D G, McGrory B J. Infection burden in total hip and knee arthroplasties: an international registrybased perspective. Arthroplasty Today 2017; 3: 137-40. doi: 10.1016/j. artd.2017.05.003 Visser A, Geboers B, Gouma D J, Goslings J C, Ubbink D T. Predictors of surgical complications: a systematic review. Surgery 2015; 158: 58-65. doi: 10.1016/j.surg.2015.01.012 Weaver F, Hynes D, Hopkinson W, Wixson R, Khuri S, Daley J, Henderson W G. Preoperative risks and outcomes of hip and knee arthroplasty in the Veterans Health Administration. J Arthroplasty 2003; 18: 693-708.

Wilper A P, Woolhandler S, Lasser K E, McCormick D, Bor D H, Himmelstein D U. Health insurance and mortality in US adults. Am J Public Health 2009; 99: 2289-95. doi: 10.2105/AJPH.2008.157685 Woolhandler S, Himmelstein D U. The relationship of health insurance and mortality: is lack of insurance deadly? Ann Intern Med 2017; 167: 424-31. doi: 10.7326/M17-1403

Acta Orthopaedica 2021; 92 (3): 311–315

311

Compensation claims after hip arthroplasty surgery in Norway 2008–2018 Tommy Frøseth AAE 1, Rune Bruhn JAKOBSEN 2,3, Ida Rashida Khan BUKHOLM 4, Anne Marie FENSTAD 5, Ove FURNES 5,6, and Per-Henrik RANDSBORG 2,7 1 Department

of Orthopaedic Surgery, Health Møre and Romsdal HF, Kristiansund Hospital, Kristiansund; 2 Department of Orthopaedic Surgery, Akershus University Hospital, Lørenskog; 3 Department of Health Management and Health Economics, Institute of Health and Society, The Medical Faculty, University of Oslo; 4 Norwegian System of Compensation to Patients, Oslo; 5 The Norwegian Arthroplasty Register, Department of Orthopaedic Surgery, Haukeland University Hospital, Bergen; 6 Department of Clinical Medicine, Faculty of Medicine, University of Bergen, Bergen, Norway; 7 Sports Medicine Institute, Hospital for Special Surgery, New York, USA. Correspondence: tommy.aae@gmail.com Submitted 2020-09-24. Accepted 2020-12-21.

Background and purpose — Orthopedic surgery is one of the specialties with most compensation claims, therefore we assessed the most common reasons for complaints following total hip arthroplasty (THA) reported to the Norwegian System of Patient Injury Compensation (NPE) and viewed these complaints in light of the data from the Norwegian Arthroplasty Register (NAR).

Patients and methods — We collected data from NPE and NAR for the study period (2008–2018), including age, sex, and type of complaint, and reason for accepted claims from NPE, and the number of arthroplasty surgeries from NAR. The institutions were grouped by quartiles into quarters according to annual procedure volume, and the effect of hospital procedure volume on the risk for accepted claim was estimated.

Results — 70,327 THAs were reported to NAR. NPE handled 1,350 claims, corresponding to 1.9% of all reported THAs. 595 (44%) claims were accepted, representing 0.8% of all THAs. Hospital-acquired infection was the most common reason for accepted claims (34%), followed by wrong implant position in 11% of patients. Low annual volume institutions (less than 93 THAs per year) had a statistically significant 1.6 times higher proportion of accepted claims compared with higher volume institutions.

Interpretation — The 0.8% risk of accepted claims following THAs is 1.6 times higher for patients treated in lowvolume institutions, which should consider increasing the volume of THAs or referring these patients to higher volume institutions.

In Norway, compensation claims are handled by the Norwegian System of Patient Injury Compensation (NPE) and not by the judiciary system. If a patient in Norway suffers a complication due to a treatment error, within either the public or private healthcare sector, the patient can file a free-of-charge compensation claim to NPE. For claims to be accepted, 3 criteria must be met. 1st, the injury must have occurred during medical treatment (examination, diagnosis, or treatment/lack of treatment) or during follow-up, and the treatment must be deemed substandard or erroneous based on current treatment guidelines. 2nd, the injury must have led to financial loss (currently set at €1,000) or to a persistent medical impairment of minimum 15%. Lastly, the claim must be filed within 3 years after the patient became aware that the injury was likely a treatment error. There is 1 exception clause to these criteria: If the injury is rare and severe, claims may be accepted even when no treatment error has been identified. The amount of compensation is being reviewed on an individual basis and calculated to cover the patient’s loss of income and increased medical expenses due to the treatment injury.

Orthopedic surgery is one of the specialties with most compensation claims following medical treatment (Jena et al.

2011). Previous studies on compensation claims after THAs have been limited by methodological inadequacies, such as short study period or limited sample size with claims ranging from 40 to just above 300 (Bhutta et al. 2011, Bokshan et al. 2017, Novi et al. 2020). We evaluated claims following both primary and revision THAs filed at the NPE from 2008 to 2018 and compared these findings with data from NAR, with a focus on institutional procedure volume.

© 2021 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 10.1080/17453674.2021.1872901

312

Patients and methods Patients All patients of any age who filed a claim with NPE following primary or revision THA from 2008 to 2018 were included.

Patients who underwent primary hemiarthroplasty or THA for a femoral neck fracture were excluded.

Methods The Norwegian Arthroplasty Register (NAR) was founded in 1987, with aims to monitor the safety and epidemiology of total joint arthroplasties (Havelin et al. 2000). Annually, approximately 98% of primary THAs and 93% of revision THAs are reported to NAR (NAR 2020).

Data from NAR was collected for the study period (2008 through 2018). The data was stratified by the number of arthroplasty procedures performed every year per institution.

The institutions were then divided by quartiles into quarters according to annual procedure volume groups: Quarter 1 (Q1) consisted of 7 institutions with an annual volume of less than 93 hip arthroplasties. Quarter 2 (Q2) included 8 institutions that performed 93–263 procedures per year. The third quarter (Q3) comprised 8 institutions with an annual surgical volume of 264–466 hips, and, finally, the highest volume quarter (Q4) included 7 institutions that perform more than 466 hip arthroplasties yearly.

All claims filed at the NPE following THA in the study period were collected, both primary THA and revision THA. The data was stratified by institution, the patient’s age and sex, type of complication, any reoperations, and any fatalities.

The reasons for the claims were recorded, together with the decision made by NPE (accepted or rejected claims). When evaluating the outcome of claims on institutional volume, the outcome of interest was the proportion of procedures resulting in an accepted claim, with the individual institution as the analysis unit.

Statistics Continuous variables were described by mean (SD) or median (range) while categorical data were presented in frequencies.

Groups were compared using the chi-square test. The institutions by procedure volume were compared using ANOVA after asserting conditions were met, and p-values adjusted for multiple testing by Tukey’s comparison test. Between-quarter associations were quantified by odds ratio. A p-value < 0.05 was considered statistically significant. The data was analyzed using IBM SPSS version 26 (IBM Corp, Armonk, NY, USA).

Ethics, funding, and conflicts of interest The Regional Ethical Committee deemed approval not necessary as all data is based on already anonymized records (REK 15.10.10). The study was funded by research grants from the Norwegian Research Council (the Norwegian Cartilage Proj-

Acta Orthopaedica 2021; 92 (3): 311–315

Table 1. Demography of hip arthroplasty procedures reported to the Norwegian Arthroplasty Registry (NAR) and claims due to treatment injuries following hip arthroplasties filed with the Norwegian System of Patient Injury Compensation (NPE) during 2008–2018

Hip

procedures Compensation Accepted Rejected

reported claims filed claims claims

to NAR to NPE n = 595 n = 755 Factor n = 70,327 n = 1,350 (44%) (56%) Age, mean (SD) 68 (11) range 11–100 Females, n (%) 45,572 (65)

64 (11) 20–89 899 (67)

64 (11) 21–89 348 (59)

64 (12) 20–89 551 (73)

SD, standard deviation.

ect, grant number 2015107). The authors declare no conflict of interests.

Results During the study period from 2008 to 2018 (11 years) 70,327 hip arthroplasties were reported to NAR, of which 86% were primary procedures and 14% were revision THAs. During this period, NPE received 53,000 claims, of which 31% were related to orthopedic surgery and 1,350 claims were filed following hip arthroplasties, representing 1.9% of all hip procedures reported to NAR. Patients filing a claim with NPE were younger than the average age of patients reported to NAR (64 years, SD 12) compared with 68 (SD 11) years, p < 0.001).

NAR received reports on 65% of women, which was comparable to the 67% of claims that were put forward by women (p = 0.2) (Table 1).

595 (44%) of 1,350 claims were accepted, representing 0.8% of all hip arthroplasties reported to NAR in the study period.

549 claims were accepted following primary procedures (0.9% of all primary hip arthroplasties) and 46 claims were accepted after revision THA (0.5% of all revisions) (p < 0.001).

Hospital-acquired infection was the most common reason for accepted claims (34%), followed by wrong implant position (11%) (Table 2). There was a decline in claims due to abductor deficiency towards the end of the period, with 67 claims (10 granted) between 2011 and 2014 compared with 20 claims (4 granted) during 2015–2018 (p = 0.04).

17 of 23 claims from private hospitals were accepted, compared with 578 of 1,327 claims from public hospitals (p = 0.004).

5 claims involving fatalities were filed, all occurring in male patients, and all 5 claims were accepted. 3 patients aged 63, 66, and 88 succumbed to complications related to surgical site infection following a primary THA. A 67-year-old patient with known diabetes mellitus underwent revision surgery and died due to hypoglycemia as blood glucose was not followed up according to guidelines. A 52-year-old patient suffered a

Acta Orthopaedica 2021; 92 (3): 311–315

Table 2. Reasons for claims (n = 595) accepted by the Norwegian System of Patient Injury Compensation for treatment injuries following total hip arthroplasty during 2008–2018

313

Table 3. Likelihood of accepted claims from the Norwegian System of Patient Injury Compensation during 2008–2018 by annual procedure volume divided by quartiles into quarters

Proportion of accepted claims (%) p = 0.04

2.5

p = 0.3 p = 0.05

Reason for accepted claims

n (%)

Hospital-acquired infection Malposition of implant Treatment failure Anisomelia Nerve injury Aseptic loosening a Abductor deficiency Exception clause Perioperative fracture Technical error Delayed treatment Wrong indication Pain Component failure Delayed diagnosis Artery injury Lack of information Anesthesia

201 (34) 67 (11) 50 (8.4) 50 (8.5) 45 (7.6) 35 (5.9) 33 (5.5) 30 (5.0) 21 (3.5) 17 (2.9) 16 (2.7) 11 (1.8) 7 (1.2) 4 (0.7) 3 (0.5) 2 (0.3) 2 (0.3) 1 (0.2)

Quarter (Q)

Odds ratio (95% CI)

2.0 p = 1.0

a

Q1 vs. all other Q1 vs. Q2 Q1 vs. Q3 Q1 vs. Q4 Q2 vs. Q3 Q2 vs. Q4 Q3 vs. Q4

1.6 (1.1–2.5) 1.8 (1.1–3.0) 1.4 (0.9–2.2) 1.7 (1.1–2.7) 0.8 (0.6–1.0) 0.9 (0.7–1.2) 1.2 (1.0–1.5)

Q1, Quarter 1 (< 93 annual procedures); Q2, Quarter 2 (93–263 annual procedures); Q3, Quarter 3 (264–466 annual procedures); Q4, Quarter 4 (> 466 annual procedures).

CI, confidence interval.

Within 3 years.

pulmonary embolus following a primary THA despite prophylactic administration of low-molecular-weight heparin. No treatment error was identified in that case, but compensation was granted based on the exception clause.

At the time of writing [August, 2020], 90% of accepted claims have had the compensation calculated, amounting to €15.4 million that has been paid out in compensation following hip arthroplasties performed during the period 2008–2018, with average compensation of €26,000.

Institutional procedure volume Institutions with the lowest annual volume (< 93, Q1) had a statistically significant higher fraction of accepted claims per procedure compared with higher volume institutions (Figure).

The odds ratio for a claim to be accepted following a hip arthroplasty performed at a low-volume institution was 1.6 (95% CI 1.1–2.5) compared with higher volume institutions (Table 3).

Discussion The main finding of the present study is that approximately 1% of all patients undergoing THA between 2008 and 2018 were granted compensation by NPE. The probability of being granted compensation claims due to a treatment error is 1.6 times more likely if the procedure was performed in a lowvolume institution compared with institutions with higher volumes of THAs. We found a 1.6 times increased risk of suffering a treatment error that led to accepted claims by NPE if the THA was performed in the lowest volume institutions compared with all other institutions. However, the proportion

1.5 p = 0.8

p = 0.9

1.0

0.5

0

Q1

Q2

Q3

Q4

Procedure volume quarters

Proportion of accepted claims by number of surgeries stratified by annual hospital procedure volume. The 4 categories represent quarters, see Table 3. P-values derived from ANOVA adjusted with Tukey’s comparison test.

of accepted claims was not statistically significantly different between the lowest volume institutions and the 2nd highest volume institutions. The 2nd highest quarter consists of 8 institutions, including 4 large university clinics in Norway.

These institutions typically treat high-risk patients and complex primary and revision cases, whereas the highest volume institutions consist of specialized “production” institutions that treat a high volume of relatively low-risk patients. We believe this might explain why Q3 has a somewhat higher number of accepted claims, despite relatively large annual procedure volumes.

Low institutional procedure volumes increase the risk of adverse outcomes following elective hip arthroplasty and revision surgery (Glassou et al. 2016, Mufarrih et al. 2019).

Our study confirms that low-volume institutions also have a higher ratio of accepted claims compared with institutions with higher volumes of THAs. However, the risk of ending up with compensation after a hip arthroplasty is only moderately elevated for the lowest volume institutions, with an odds ratio of 1.6 and a 95% confidence interval approaching 1.0. In contrast, in a similar study on compensation claims following total knee arthroplasty (TKA), we found a 3-fold increased risk of accepted claims in the lowest volume institutions (Randsborg et al. 2020). The effect of institutional volume on accepted claims seems to be much more pronounced for TKAs than for THAs as this is also found in a Finnish report (Järvelin et al. 2012). This is not surprising, since hospital volume has a greater effect on adverse outcome following knee arthroplasties compared with hip arthroplasties (Katz et al. 2004, Shervin et al. 2007). THA is indisputable a more successful procedure with higher patient satisfaction and less complications compared to TKA (Bourne et al. 2010), which may explain why institutional volume seems to have less impact on

314

compensation after hip arthroplasties than after knee arthroplasties.

The dominant reason for accepted claims was hospitalacquired infection, accounting for one-third of accepted claims. Hospital-acquired infection of an arthroplasty leads to prolonged hospitalization, and increased morbidity and mortality, with extensive costs to the society (Senard et al. 2019).

This is a reminder that all parties involved in arthroplasty surgery should strive to reduce the risk of infection.

The exception clause grants compensation to patients who suffer from infection following joint replacement, even when no treatment error has been identified. A similar decision has been made for early (within 3 years) aseptic loosening. It is a pragmatic policy in a no-blame compensation system. However, not all infections will automatically lead to compensation. All claims are reviewed independently. Patients with comorbidity, high infection risk, and poor compliance may not be granted compensation if infection occur.

Other leading causes for accepted claims were wrong implant position, treatment failure (no further explanation was given), anisomelia, and nerve injury.

Abductor deficiency was registered in 6% of accepted claims.

This is likely related to the surgical technique and the direct lateral approach to the hip joint (Amlie et al. 2014, Winther et al. 2016). This approach has decreased substantially in comparison with posterior and anterior approaches in Norway in recent years and is now performed in less than 5% of all THAs (NAR 2020). We believe this explains the decline in claims due to abductor deficiency registered at the end of the study period.

Pain was a common reason for claim (Table 1). NPE registered 215 claims due to pain, although only 7 (3%) were accepted for compensation; pain alone does not serve as a cause for compensation, which is in line with previous reports from NPE (Clementsen et al. 2018, Randsborg et al. 2018, Aae et al. 2020).

Khan et al. (2020) assessed compensation claims based on the Danish arthroplasty register and found that 2.5% of all THAs filed a claim, which is somewhat higher than our findings of 1.9%. They reported nerve damage and insufficient or incorrect treatment as the leading causes for accepted claims.

Half of the claims were accepted, which is similar to the 44% of claims accepted in our study. In contrast to our study, they did not include revision surgery. Our study also adds knowledge on accepted claims based on institutional volumes of THA, a topic that has received little attention in the literature.

Bhutta et al. (2011) reviewed compensation claims after hip and knee arthroplasties over a 5-year period in the United Kingdom. THAs due to trauma and revision surgery were excluded. They identified 271 claims that had reached a conclusion, where 109 (40%) resulted in payouts. Nerve injury, surgeon error, and pain were the 3 most common causes for claims. Our material contains 5 times as many claims as Bhutta et al. and includes revision surgery. Albeit nerve injury and surgeon error are common reasons for accepted claims

Acta Orthopaedica 2021; 92 (3): 311–315

in both studies, infection was far more common in our study.

This discrepancy is likely caused by the exception clause in the Scandinavian compensation systems that grants compensation after infection even if no treatment error is identified.

This is supported by a study on patient claims following prosthetic hip infections in Sweden, which found that 329 of 441 (75%) of claims were accepted (Kasina et al. 2018), which is the same as the 209 of 275 (76%) of infections that were accepted in our study.

2 studies from the United States analyzed malpractice lawsuits after THA and both identified substantially fewer claims than our study, which may relate to the systems (Bokshan et al. 2017, Samuel et al. 2019). In Norway, filing a complaint for compensation is free of charge and easily done online.

Additionally, the system is a no-fault system with claims not directed at individual surgeons or institutions, and surgeons are required by law in case of complications to inform patients how to apply for compensation. In the United States, filing a complaint will normally require legal representation, which might constitute a higher threshold for filing a claim.

Some treatment errors found in our study are avoidable or at least modifiable with adjustments in medical practice. This is coherent with findings from Sweden, which stated that 49% of THA patients suffered an adverse advent that could have been prevented (Magnéli et al. 2020). Delayed diagnosis or treatment and erroneous indications have previously been identified as avoidable causes for compensation (Clementsen et al. 2018, Randsborg et al. 2018, Aae et al. 2020). During the study period, all hospitals in Norway implemented the use of the safe surgery protocol initiated by the World Health Organization (WHO 2020). No cases of wrong-sided surgery were identified in our study and this may relate to the implementation of these guidelines.

1 accepted claim was not due to the surgery itself, but to anesthesia and is a reminder that arthroplasty carries risks not directly related to the surgery.

This study has limitations. Any mislabeling of procedures may lead to some patients not being included in the study.

NPE does not cover all possible complications following THA, and it is likely that some patients have experienced a complication due to a treatment error that would lead to compensation, but never filed a claim. These factors may induce biases to the database used in this study.

The data originates from 1 country using the principle of no-blame, which may reduce the generalizability of the study.

However, the purpose of this study is to evaluate compensation claims to identify areas for potential improvement in our patient care, which is likely of universal interest.

In conclusion, the main reasons for compensation were a hospital-acquired infection and malposition of the implant. The findings suggest that small-volume institutions should consider increasing the volume of THAs or referring these patients to higher volume institutions. Several complications identified are reducible with adjustments to current clinical practice.

Acta Orthopaedica 2021; 92 (3): 311–315

TFA: original idea, interpretation of data, drafting of manuscript. IRKB: data acquisition, revision of manuscript. AMF: data acquisition, interpretation of data, revision of manuscript. OF: interpretation of data, revision of manuscript. RBJ: study design, collection and interpretation of data, revision of manuscript. PHR: original idea, study design, collection and interpretation of data, revision of manuscript.

The authors would like to thank Myrthle Slettvåg Hoel, orthopedic research nurse and Trine Sandblost, librarian at Health Møre and Romsdal Hospital Trust, Kristiansund Hospital for secretarial support.

Acta thanks Pieter K Bos and Kim Lyngby Mikkelsen for help with peer review of this study.

Aae T F, Lian Ø B, Årøen A, et al. Compensation claims after knee cartilage surgery is rare: a registry-based study from Scandinavia from 2010 to 2015.

BMC Musculoskelet Disord 2020; 21(1): 287. doi: 10.1186/s12891-02003311-4.

Amlie E, Havelin L I, Furnes O, et al. Worse patient-reported outcome after lateral approach than after anterior and posterolateral approach in primary hip arthroplasty: a cross-sectional questionnaire study of 1,476 patients 1–3 years after surgery. Acta Orthop 2014; 85(5): 463-9. doi: 10.3109/17453674.2014.934183.

Bhutta M A, Arshad M S, Hassan S, et al. Trends in joint arthroplasty litigation over five years: the British experience. Ann R Coll Surg Engl 2011; 93(6): 460-4. doi: 10.1308/003588411X587226.

Bokshan S L, Ruttiman R J, DePasse J M, et al. reported litigation associated with primary hip and knee arthroplasty. J Arthroplasty 2017; 32(12): 3573-7.e1. doi: 10.1016/j.arth.2017.07.001.

Bourne R B, Chesworth B, Davis A, et al. Comparing patient outcomes after THA and TKA: is there a difference? Clin Orthop Relat Res 2010; 468(2): 542-6. doi: 10.1007/s11999-009-1046-9.

Clementsen S H, Hammer O L, Engebretsen E, et al. Compensation after distal radial fractures. a review of 800 claims to the Norwegian System of Patient Injury Compensation 2000–2013. Open Orthop J 2018; 12:419-26. doi: 10.2174/1874325001812010419.

Glassou E N, Hansen T B, Mäkelä K, et al. Association between hospital procedure volume and risk of revision after total hip arthroplasty: a population-based study within the Nordic Arthroplasty Register Association database. Osteoarthritis Cartilage 2016; 24(3): 419-26. doi: 10.1016/j. joca.2015.09.014.

Havelin L I, Engesaeter L B, Espehaug B, et al. The Norwegian Arthroplasty Register: 11 years and 73,000 arthroplasties. Acta Orthop Scand 2000; 71(4): 337-53. doi: 10.1080/000164700317393321.

Jena A B, Seabury S, Lakdawalla D, et al. Malpractice risk according to physician specialty. N Engl J Med 2011; 365(7): 629-36. doi: 10.1056/ NEJMsa1012370.

315

Järvelin J, Häkkinen U, Rosenqvist G, et al. Factors predisposing to claims and compensations for patient injuries following total hip and knee arthroplasty.

Acta Orthop 2012; 83(2): 190-6. doi: 10.3109/17453674.2012.672089.

Kasina P, Enocson A, Lindgren V, et al. Patient claims in prosthetic hip infections: a comparison of nationwide incidence in Sweden and patient insurance data. Acta Orthop 2018; 89(4): 394-8. doi: 10.1080/17453674.2018.1477708.

Katz J N, Barrett J, Mahomed N N, et al. Association between hospital and surgeon procedure volume and the outcomes of total knee replacement. J Bone Joint Surg 2004; 86(9): 1909-16.

Khan N, Petersen M M, Mikkelsen K L, et al. No-fault compensation from the patient compensation association in Denmark after primary total hip replacement in Danish hospitals 2005–2017. J Arthroplasty 2020; 35(7): 1784-91. doi: 10.1016/j.arth.2020.02.042.

Magnéli M, Unbeck M, Samuelsson B, et al. Only 8% of major preventable adverse events after hip arthroplasty are filed as claims: a Swedish multicenter cohort study on 1,998 patients. Acta Orthop 2020; 91(1): 20-5. doi: 10.1080/17453674.2019.1677382.

Mufarrih S H, Ghani M O A, Martins R S, et al. Effect of hospital volume on outcomes of total hip arthroplasty: a systematic review and meta-analysis. J Orthop Surg Res 2019; 14(1): 468. doi: 10.1186/s13018-019-1531-0.

NAR. Norwegian Arthroplasty Register. Annual Report 2020. http://nrlweb. ihelse.net/; 2020.

Novi M, Vanni C, Parchi P, et al. Claims in total hip arthroplasty: analysis of the instigating factors, costs and possible solution. Musculoskelet Surg 2020; 140(1): 43-8. doi: 10.1007/s12306-019-00590-6.

Randsborg P H, Bukholm I R K, Jakobsen R B. Compensation after treatment for anterior cruciate ligament injuries: a review of compensation claims in Norway from 2005 to 2015. Knee Surg Sports Traumatol Arthrosc 2018; 26(2): 628-33. doi: 10.1007/s00167-017-4809-y.

Randsborg P-H, Aae T F, Bukholm I R K, et al. Compensation claims after knee arthroplasty surgery in Norway 2008–2018. Acta Orthop 2020; 91(x): 1-5. Epub ahed of print. doi: 10.1080/17453674.2020.1871187 Samuel L T, Sultan A A, Rabin J M, et al. Medical malpractice litigation following primary total joint arthroplasty: a comprehensive, nationwide analysis of the past decade. J Arthroplasty 2019; 34(7s): S102-s7. doi: 10.1016/j.arth.2019.02.066.

Senard O, Houselstein T, Crémieux A C. Reasons for litigation in arthroplasty infections and lessons learned. J Bone Joint Surg Am 2019; 101(20): 180611. doi: 10.2106/jbjs.19.00101.

Shervin N, Rubash H E, Katz J N. Orthopaedic procedure volume and patient outcomes: a systematic literature review. Clin Orthop Relat Res 2007; 457: 35-41. doi: 10.1097/BLO.0b013e3180375514.

WHO. Safe surgery checklist. https://www.who.int/patientsafety/safesurgery/ checklist/en/; 2020.

Winther S B, Husby V S, Foss O A, et al. Muscular strength after total hip arthroplasty: a prospective comparison of 3 surgical approaches. Acta Orthop 2016; 87(1): 22-8. doi: 10.3109/17453674.2015.1068032.

316

Acta Orthopaedica 2021; 92 (3): 316–322

Implant survival of 2,723 vitamin E-infused highly crosslinked polyethylene liners in total hip arthroplasty: data from the Finnish Arthroplasty Register Matias HEMMILÄ 1, Inari LAAKSONEN 1, Markus MATILAINEN 2, Antti ESKELINEN 3, Jaason HAAPAKOSKI 4, Ari-Pekka PUHTO 5, Jukka KETTUNEN 6, Konsta PAMILO 3, and Keijo T MÄKELÄ 1 1 Department

of Orthopaedic Surgery, University of Turku and Turku University Hospital, Turku; 2 Turku PET Centre, University of Turku and Turku University Hospital, Turku; 3 Coxa Hospital for Joint Replacement and Faculty of Medicine and Health Technologies, Tampere University, Tampere; 4 National Institute for Health and Welfare, Helsinki; 5 Division of Operative Care, Department of Orthopaedic and Trauma Surgery, Oulu University Hospital, Oulu; 6 Department of Orthopaedics and Traumatology, Kuopio University Hospital, Kuopio, Finland Correspondence: matias.hemmila@utu.fi Submitted 2020-09-22. Accepted 2020-12-29.

Background and purpose — The use of crosslinked polyethylene in total hip arthroplasty (THA) has decreased wear remarkably. It has been suggested that the antioxidative effects of vitamin E may enhance the wear properties of polyethylene even further. This study evaluates revision rates between vitamin E-infused polyethylene liners (E1 and E-poly, ZimmerBiomet, Warsaw, IN, USA) versus moderately crosslinked polyethylene (ModXLPE) liners from the same manufacturer used in primary THA. Patients and methods — We conducted a study based on data from the Finnish Arthroplasty Register. The study group consisted of 2,723 THAs with a vitamin E-infused liner and a reference group of 2,707 THAs with a moderately crosslinked polyethylene liner. Survivorship, revision risk, and re-revision causes were compared between groups.

Results — The 7-year survival of the vitamin E-infused polyethylene liner group and of the reference group with revision for any reason as the endpoint was comparable (94% [95% CI 92.9–94.9] and 93% [CI 91.9–93.9], respectively). The adjusted hazard ratio (HR) for any revision was similar between the groups (0.7 [CI 0.4–1.1]). When revision for aseptic loosening was studied as the endpoint, the survival for the study group was 99% (CI 98.6–99.4) and for the reference group 99% (CI 98.7–99.5), and the risk of revision was comparable between the study groups (HR 1.3 [CI 0.7–2.5]).

Interpretation — After an observation period of 7 years vitamin E-infused liners shows results equal to results obtained with crosslinked polyethylene liners.

Highly crosslinked polyethylene (HXLPE) was introduced in the late 1990s to decrease polyethylene wear and periprosthetic osteolysis and to increase the long-term survivorship of THA (Bragdon et al. 2013). HXLPE has shown lower wear rates in vitro (McKellop et al. 2000) and in vivo (Bragdon et al. 2013) compared with conventional non-crosslinked ultrahigh-molecular-weight polyethylene (UHMWPE). However, as a downside, free radicals are released and oxidation is exacerbated, which may induce wear (Kurtz 2009). One potential solution to further decrease the number of free radicals in the liner material is to add vitamin E to HXLPE, which increases the resistance of polyethylene against these oxidative processes by stabilizing the material (Oral et al. 2006a, 2006b).

Since vitamin E-infused highly crosslinked polyethylene (VEPE) is a quite recent invention, there are only short- and mid-term data available on its efficacy and safety. Significantly lower femoral head penetration rates have been reported for VEPE liners by many authors in randomized controlled trials (RCTs) with radiostereometric analysis (RSA) compared with HXLPE liners (Nebergall et al. 2017, Scemama et al. 2017, Shareghi et al. 2017, Galea et al. 2018, Rochcongar et al.

2018, Sköldenberg et al. 2019). However, there is still a lack of real-world data from arthroplasty registers on the survival of VEPE liners.

We used the Finnish Arthroplasty Register (FAR) to assess implant survival of vitamin E-infused HXLPE liners (E1, E-poly, ZimmerBiomet, Warsaw, IN, USA) compared with moderately crosslinked polyethylene (ModXLPE) liners from the same manufacturer for revision for any reason. We further compared the study groups for those revisions performed for aseptic loosening, osteolysis, or polyethylene wear. We hypothesize that there is no statistically significant difference in outcome between the study groups.

© 2021 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 10.1080/17453674.2021.1879513

Acta Orthopaedica 2021; 92 (3): 316–322

317

Table 2. Demographic data of study population, number (%), unless stated otherwise

Uncemented cups in primary THA reported to the Finnish Arthroplasty Register between January 2000 and December 2017 in patients aged 18–100 years n = 133,473

Data

Excluded (n = 128,043): – head size other than 28/32/36 mm and head material other than metal or ceramic, 30,079 – other than selected cup models, 93,990 – dual mobility, constrained liners, and metal liners, 1,199 – other than selected stems (see Table 2), 2,489 – other than E1 orE-poly and missing liner description, 286 Study selection (n = 5,430): – E-poly and E1 liners, 2,723 – control group liners, 2,707

Figure 1. Flowchart of study selection.

Table 1. Components used, number (%)

Cup designs Biomex a Exceed a G7 a Regenerex a Vision Ringloc a Stem designs Bi-Metric a CDH a Echo a Reach a Taperloc a a

VEPE group

Reference group

0 (0) 506 (19) 501 (18) 741 (27) 975 (36)

56 (2) 312 (12) 211 (8) 6 (0.2) 2,122 (78)

1,143 (42) 3 (0.1) 1,357 (50) 18 (1) 202 (7)

2,459 (91) 20 (0.7) 174 (6) 4 (0.2) 50 (2)

ZimmerBiomet, Warsaw, IN, USA.

Patients and methods Our study is based on data from the Finnish arthroplasty register (FAR), which has collected information on arthroplasties since 1980 (Paavolainen et al. 1991). The register acts nationwide, with data completeness exceeding 95% for primary THA and 81% for revision THA (FAR). The register is administered by the Finnish National Institute for Health and Welfare. Orthopedic units are obligated to provide information essential for maintenance of the register. Dates of death are obtained from the Population Information System maintained by the Population Register Center. The data content of the FAR was scrutinized and revised in May 2014. The updated data now includes more detailed information such as more precise reasons for revisions. Prior to the 2014 update, revisions for liner wear were recorded as performed “for other reason.”

Study population Between January 2000 and December 2017, 133,488 primary THAs were reported to the FAR. We included in the study group any THAs in which the vitamin E-infused HXLPE

VEPE group

Reference group

Mean age, years (SD) 67 (10) 64 (9) BMI (SD) 29 (5) 28 (5) Male sex 1,341 (49) 1,357 (50) Diagnosis Primary osteoarthritis 2,328 (86) 2,274 (84) Rheumatoid arthritis 59 (2) 83 (3) Other a 336 (12) 350 (13) Femoral head size 28 4 (0.2) 2,229 (82) 32 321 (12) 284 (11) 36 2,398 (88) 194 (7) Femoral head material Ceramic 822 (30) 220 (8) Metal 1,901 (70) 2,487 (92) Status at end of follow up b Not revised 2,571 (94) 2,348 (87) Revised 152 (6) 359 (13) Operation year 2000–2008 6 (0.2) 2,376 (88) 2009–2017 2,717(99.8) 331 (12) a Fractures, avascular necrosis, osteoarthritis due to hip dysplasia, tumors, congenital hip dislocation, Mb Legg–Calve–Perthes, femoral head epiphysiolysis. b Excluding death.

(E1, E-poly) liner option was used with 1 of 5 uncemented acetabular components from the same manufacturer (ZimmerBiomet): Biomex, Exceed, G7, Regenerex, and Vision Ringloc (Table 1). The reference group consisted of THAs used with ModXLPE liners from the same manufacturer (mostly ArCom) with the same cup designs. Exclusion criteria were head size other than 28 mm, 32 mm, or 36 mm; dual mobility acetabular device; metal or ceramic liner; or constrained liner. Only THAs with uncemented stems were included in the study. In 5,430 THAs the study inclusion criteria were fulfilled, and femoral head material information was available (2,723 E-poly or E1 liner THAs). The study group patients were operated on between January 1, 2008 and December 31, 2017, with a mean follow-up time of 5.0 years (0–9.7). The reference group patients were operated on between January 1, 2000 and December 31, 2017, with a mean follow-up time of 11.0 years (0–18.5). The number of patients with bilateral hip prostheses was 410, of whom 85 had both hips done simultaneously. Mortality during the study period was 10% in the VEPE group and 27% in the control group. Differences in mortality between the groups are explained by the difference in follow-up time (Table 2).

Surgery In both groups, the most common cup design was Vision RingLoc (36% in the VEPE group, 78% in the control group).

The most frequently used stems in the study population were Echo in the VEPE group (50% of all VEPE THAs) and Bi-

318

Acta Orthopaedica 2021; 92 (3): 316–322

K–M revisionfree survival

K–M revisionfree survival (specified revisions)

1.00

1.00 Control group VEPE group

0.95

Control group VEPE group 0.98

0.90 0.96 0.85 0.94 0.80 0.92 0.75 0.90

0.70

0.65

0

5

10

15

20

0.88

Years postoperatively

0

5

10

15

20

Years postoperatively

Figure 2. Kaplan–Meier survival for VEPE group and Reference group with revision for any reason as the endpoint. 95% CI levels presented in blue and red.

Figure 3. Kaplan–Meier survival for VEPE group and Reference group with revision for osteolysis, liner wear, liner breakage, loosening of the cup, and other reason as an endpoint. 95% CI levels presented in blue and red.

Metric in the control group (91% of all THAs in the control group) (Table 1). A ceramic head was used in 30% of cases in the VEPE group compared with 8% in the reference group. In the VEPE group 36 mm femoral heads were used in 88% of cases, while 28 mm was the most commonly used head size in the control group (82%).

group was stratified. Head size was excluded from this model because of large differences in head sizes between the groups.

If the proportional hazards assumption for a variable was not fulfilled in the Cox model, the model was stratified by it instead. Stratification in Cox models means that the hazard functions can be estimated for all level combinations of the stratified variables, and the hazard ratios for the other variables (those that meet the proportional hazard assumption) are then optimized for all these hazard functions. Without stratification we would assume that the hazards were the same for all levels of such variables. The results of the Cox regression analysis are presented with the hazard ratio (HR) and CI. All analyses were performed using SAS software (Version 9.4; ASA Institute, Cary, NC. USA).

Statistics The primary outcome was revision for any reason and the secondary outcome was revision for loosening of the cup, osteolysis, liner wear, or liner breakage. Prior to the register update in 2014, revisions performed for osteolysis and wear were coded as performed for “other reason”; therefore, revisions performed for “other reason” prior to May 2014 are included in the analyses for secondary outcome. Patients were excluded for any other event than the outcome, or at the end of follow-up. Kaplan–Meier survival estimates with 95% confidence interval (CI) were calculated for both groups at 1, 3, 5, and 7 years for any reason for revision and for loosening of the cup, osteolysis, liner wear, or liner breakage. The survival curves were compared using the log-rank test. Revision was described as a change or removal of at least one component.

We adjusted the estimated revision risks in the Cox multiple regression model by sex, operated side, and femoral head material (ceramic, metal). Femoral head size (28, 32, 36 mm), age group (18–55, 56–65, 66–75, > 75 years), and preoperative diagnosis (primary osteoarthritis, rheumatoid arthritis, other) were stratified. None of these variables were considered to be along a causal pathway from exposure to outcome but were considered as confounders. The second analysis was performed for loosening of the cup, osteolysis, liner wear, or liner breakage as the endpoint. Side, femoral head material, sex, and diagnosis were adjusted for in the Cox model, and age

Ethics, funding, and potential conflicts of interest Ethical approval was from the Finnish National Institute for Health and Welfare (June 13, 2017, Dnor THL/926/5.05.

00/2017). This research received funding from the Finnish Government Research Grant. The authors declare no conflicts of interest.

Results Revision for any reason The 7-year survivorship with revision for any reason as endpoint was similar between the groups: 94.0% (CI 92.9–94.9) for the VEPE group and 93.0% (CI 91.9–93.9) for the reference group (Figure 2, Table 3, see Supplementary data). In the Cox regression analysis, the risk of revision in the VEPE group was lower, but the result was not statistically significant (HR 0.7 [CI 0.4–1.1]) (Table 4).

Acta Orthopaedica 2021; 92 (3): 316–322

319

Table 4. Revision risk according to Cox regression model with all revisions as endpoint

Group

HR (95% CI)

VEPE group vs. Reference group 0.69 (0.44–1.1) Adjusting variables Left vs. right side 0.98 (0.82–1.2) Female vs. male 0.99 (0.83–1.2) Ceramic vs. metal head 1.2 (0.90–1.5)

p-value 0.09 0.8 0.9 0.3

Adjusting variables stratified by head size, age group, and diagnosis.

HR = hazard ratio. CI = confidence interval.

Table 6. Revision risk according to Cox regression model with revision for osteolysis, liner wear, liner breakage, loosening of the cup (and other reason before May 15, 2014) as endpoint

Group

HR (95% CI)

VEPE group vs. Reference group 1.3 (0.71–2.5) Adjusting variables Female vs. male 1.0 (0.71–1.5) Other diagnoses vs. RA 0.96 (0.35–2.6) Primary OA vs. RA 0.85 (0.34–2.1) Ceramic vs. metal head 1.1 (0.70–1.8)

p-value

Table 7. Indication for revision prior to data content revision (May 15, 2014) of Finnish Register, number (%) Main reason for revision a VEPE group

Aseptic loosening Cup and stem 0 (0) Cup 9 (10) Stem 2 (2) Infection 10 (11) Dislocation 27 (30) Component malposition 7 (8) Fracture 24 (26) Component breakage 0 Other 12 (13) a

Reference group 2 (1) 10 (4) 9 (4) 17 (7) 103 (46) 22 (10) 23 (10) 2 (1) 41 (18)

No data available concerning indication for revision from 36 revisions.

Table 8. Indication for revision after new indications for revision were added at the data content revision: data starting from May 15, 2014, number (%)

0.4 0.9 0.9 0.7 0.6

Adjusting variables are stratified by age group and side.

HR = hazard ratio. CI = confidence interval, OA = osteoarthritis, RA = rheumatoid arthritis.

Revision for aseptic loosening of the cup, osteolysis, liner wear, or liner breakage The 7-year survivorship with revision for aseptic loosening of the cup, osteolysis, liner wear, or liner breakage as endpoint was equal between the groups: (VEPE group 99.1% [CI 98.6–99.4]); reference group 99.2% (CI 98.7–99.5) (Figure 3, Table 5, see Supplementary data). The risk of revision in the VEPE group was the same as in the reference group (HR 1.3 [CI 0.7–2.5]) (Table 6).

Reasons for revision The most frequent reason for revision before the register update (May 2014) were dislocation (27%), periprosthetic fracture (24%), and infection (11%) in the VEPE group, and dislocation (46%), other reason (18%), and component malposition (10%), as well as periprosthetic fracture (10%) in the reference group (Table 7). After the register update the most frequent reason for revision was dislocation (33%), followed by infection (21%), and periprosthetic femoral fracture (14%) in the VEPE group, and dislocation (23%), liner wear (22%), and periprosthetic femur fracture (17%) in the reference group (Table 8). Liner breakage was observed in 2 patients in the VEPE group and 3 patients in the reference group in the scrutinized register data (5% VEPE group, 3% reference group).

Main reason for revision a

VEPE group

Reference group

Aseptic loosening Cup 1 (2) 3 (3) Stem 3 (7) 2 (2) Osteolysis Cup 0 11 (12) Stem 0 3 (3) Liner wear 0 20 (22) Component breakage Cup 0 1 (1) Liner 2 (5) 3 (3) Modular neck 0 0 Infection 9 (21) 5 (6) Dislocation 14 (33) 21 (23) Component malposition Cup 2 (5) 4 (4) Periprosthetic fracture Acetabulum 0 1 (1) Femur 6 (14) 15 (17) Unexplained pain 1 (2) 0 Leg length discrepancy repair 2 (5) 0 Other 3 (6) 1 (1) a

No data available concerning indication for revision from 27 revisions.

Discussion We found that VEPE liners perform comparably to ModXLPE liners at mid-term follow-up. The risk of revision in the VEPE group was lower when revision for any reason was the end point but the result was not statistically significant (HR 0.7 [CI 0.4–1.1]). This is one of the largest studies of VEPE liners with a mean follow-up of 5 years. Our findings support the assumption that VEPE liners are durable and safe; however, further studies with longer follow-up are needed to assess the long-term survival and possible benefits of this material (Sillesen et al. 2016, Nebergall et al. 2017, Galea et al. 2019).

320

Charnley first introduced ultrahigh molecular weight polyethylene (UHMWPE) in 1960 as the bearing material for the recently developed THA. The sterilization of conventional polyethylene (PE) was performed by gamma irradiation in air.

The benefit of this process is molecular crosslinking, which provides improved wear resistance. On the downside, this process produces free radicals that decrease resistance and cause degradation and thus increase polyethylene wear (McKellop et al.

2000). PE debris may induce periprosthetic osteolysis through the release of cytokines and proteolytic enzymes and thus wear of bearing surfaces is thought to be the main limiting factor of long-term survival of THA (Merola and Affatato 2019).

HXLPE was introduced in the late 1990s to decrease polyethylene wear and osteolysis (Bragdon et al. 2013). Today, it is considered the gold standard for acetabular liners in THA (Oral et al. 2007). The number of free radicals formed in the crosslinking procedure can be reduced by heating the material above its melt temperature or annealing below its melt temperature after crosslinking (Baker et al. 2003). However, the processes do not eliminate all free radicals (Currier et al. 2007, Kurtz 2009). VEPE liners were developed to further improve the oxidative stability of radiated XLPE. The added vitamin E increases the resistance of polyethylene against oxidative processes by stabilizing the material (Oral et al. 2006a, 2006b).

VEPE has shown excellent wear characteristics and resistance to oxidative stress in laboratory conditions (Oral et al. 2006a).

There are generally 2 methods of adding vitamin E to crosslinked UHMWPE: blending the vitamin E with the UHMWPE before irradiation and crosslinking or infusing it after crosslinking (Rowell et al. 2011). A higher vitamin E concentration can be achieved by infusing it into the HXLPE (Rowell et al.

2011), but the clinical effect is unclear (Kurtz et al. 2009).

Liner wear is often assessed by measuring the penetration of the femoral head into the liner with RSA. However, the real penetration rate comprises not only the true loss of PE but also creep deformation of the liner. Several RCTs have been performed comparing femoral head penetration rates of VEPE and ModXLPE liners using RSA. Some authors have reported lower penetration rates in VEPE patients at short- to mediumterm follow-up, although wear rates have been low in both groups (Salemyr et al. 2015, Scemama et al. 2017, Shareghi et al. 2017, Galea et al. 2018, Rochcongar et al. 2018). Almost as many authors have reported equal penetration rates at medium-term follow-up (Nebergall et al. 2017, Galea et al.

2019, Busch et al. 2020). Lindalen et al. (2019) compared 32 mm versus 36 mm ceramic femoral heads with VEPE liners and did not find any differences in wear rates in RSA measurements at 6-year follow-up. The wear rates have been low in both groups and well below the reported osteolytic threshold of 0.1 mm/year (Dumbleton et al. 2002); therefore, the measured statistically lower penetration rates might not be clinically relevant, and longer follow-up is needed.

The Australian Orthopaedic Association National Joint Replacement Registry (AOANJRR) has reported a similar

Acta Orthopaedica 2021; 92 (3): 316–322

revision risk for THAs with antioxidant inserts compared with ModXLPE inserts (AOANJRR 2019). The 8-year Kaplan– Meier estimates (cumulative percent revision) of 6,046 conventional THAs using the Ringloc cup with XLPE or VEPE liners were 2.5 for both groups (2.5 [CI 2.0–3.2]) and 2.5 [CI 1.9–3.1], respectively). The 4-year Kaplan–Meier estimate for 2,729 THAs using the G7 cup with VEPE liner was inferior to that for the XLPE liner (1.8 [CI 1.3–2.5]) and 2.9 [CI 1.3–6.3]), respectively), although the number of XLPE liners was limited. Recently published work based on the National Joint Registry (NJR) found that VEPE liners, HXLPE liners (radiation ≥ 5 mrad), and liners heated above the melting point were associated with best survival in a cohort of 292,920 primary THAs. For VEPE liners, the 8-year cumulative incidence function of revision due to aseptic loosening was 0.3 and due to reasons other than aseptic loosening 1.7 (values estimated from the figure). However, the follow-up time of 11,926 VEPE liners was relatively short (3.3 years) (Davis et al. 2020). A multinational collaboration study of 977 patients reported equal performance between the VEPE liner and ModXLPE liner at 3-year follow-up, and no early in-vivo adverse effects were observed (Sillesen et al. 2016).

Our findings support these earlier findings. Our study design was to compare the same cup brands from the same manufacturer with either a VEPE or ModXLPE liner. We think this is an optimal study setting to compare differences between these liner materials as cup designs do not bias the results. The reference group consist of ModXLPE liners whereas VEPE liners are made of HXLPE. The amount of cross-linking and thus wear resistance increases with increasing radiation dose (McKellop et al. 1999), but higher doses are also associated with a decrease in tensile and fracture toughness (Gomoll et al. 2002). All in all, a recent large register study of 292,920 primary THAs did not find any difference in the survival of moderately and highly irradiated liners at maximum follow-up of 14 years (Davis et al. 2020).

Prior to the Finnish register revision of 2014, liner wear was not recorded separately but as “other reason,” which may cause minor bias. The proportion of revisions performed for loosening and wear in our study is in line with other registers (AOANJRR 2019). There were 2 revisions performed due to liner breakage in the VEPE group versus 3 revisions in the reference group after 2014, accounting for 5% of revisions in the VEPE group and 3% in the reference group. Reports of VEPE liner breakage in the literature are rare (Bates and Mauerhan 2015, Brazier and Mesko 2018), and current data support the previous findings. Concerns over safety issues have not been raised in several previous studies, and our results are in agreement with this (Jarrett et al. 2010, Gigante et al. 2015, Sillesen et al. 2015, Davis et al. 2020).

The primary strength of this nationwide study is the large population-based setup with a mean 5-year follow-up time for the VEPE group. A limitation of the study is that we were not able to assess radiographs to evaluate wear. Further, we were

Acta Orthopaedica 2021; 92 (3): 316–322

not able to assess patient comorbidities (e.g., cardiovascular diseases, psychiatric disorders, and cancer) which could affect revision rates. Revision operation was also the only outcome we were able to assess, as FAR data contents do not include patient-reported outcome measures. The study groups were operated on in somewhat different time eras. However, at the same time, this is also a strength of our study, as we wanted in particular to assess two generations of liner materials from the same manufacturer using the same acetabular components.

Femoral head size increased so substantially during the study period that we were not able to use it as a variable in the Cox model with osteolysis and wear as the endpoint (wide confidence intervals). The portion of ceramic heads between the groups was somewhat different (30% VEPE group versus 8% reference group), but a recent study found similar wear rates between metal and ceramic heads (Gaudiani et al. 2018).

Despite these weaknesses of our study, we do not feel that our message is undermined and VEPE liners are a safe option with good medium-term results.

In conclusion, after an observation period of 7 years, vitamin E-infused liners show results equal to results obtained with crosslinked polyethylene liners and results are in line with previous findings. Longer follow-up is needed to assess the potential advantages, if any, of VEPE liners in the long term.

Supplementary data Tables 3 and 5 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.

2021.1879513

MH, IL, KM: planning the study, analysis of the data, and writing the manuscript; MaM and JH: calculating the statistics and revision of the manuscript; AE, A-PP, JK, KP: revision of the manuscript.

Acta thanks Rob Nelissen and Leif Ryd for help with peer review of this study.

AOANJRR. Hip, knee & shoulder arthroplasty: 2019 Annual Report; 2019.

Baker D A, Bellare A, Pruitt L. The effects of degree of crosslinking on the fatigue crack initiation and propagation resistance of orthopedic-grade polyethylene. J Biomed Mater Res 2003; 66(1):1 46-54.

Bates M D, Mauerhan D R. Early fracture of a vitamin-E-infused, highly cross-linked polyethylene liner after total hip arthroplasty. JBJS Case Connect 2015; 5(3): e65.

Bragdon C R, Doerner M, Martell J, et al. The 2012 John Charnley Award: Clinical multicenter studies of the wear performance of highly crosslinked remelted polyethylene in THA. Clin Orthop Relat Res 2013; 471(2): 393-402.

Brazier B G, Mesko J W. Superior rim fracture of a vitamin E-infused highly cross-linked polyethylene (HXLPE) liner leading to total hip arthroplasty revision. Arthroplast Today 2018; 4(3): 287-90.

Busch A, Jäger M, Klebingat S, et al. Vitamin E-blended highly cross-linked polyethylene liners in total hip arthroplasty: a randomized, multicenter trial using virtual CAD-based wear analysis at 5-year follow-up. Arch Orthop Trauma Surg 2020; 140(12): 1859-66.

Currier B H, Currier J H, Mayor M B, et al. Evaluation of oxidation and fatigue damage of retrieved crossfire polyethylene acetabular cups. J Bone Joint Surg Am 2007; 89: 2023-9.

321

Davis E T, Pagkalos J, Kopjar B. Polyethylene manufacturing characteristics have a major effect on the risk of revision surgery in cementless and hybrid total hip arthroplasties. Bone Joint J 2020; 102-B(1): 90-101.

Dumbleton J H, Manley M T, Edidin A A. A literature review of the association between wear rate and osteolysis in total hip arthroplasty. J Arthroplasty 2002; 17(5): 649-61.

FAR. Finnish Arthroplasty Register. Available from: www.thl.fi/far.

Galea V P, Connelly J W, Shareghi B, et al. Evaluation of in vivo wear of vitamin E-diffused highly crosslinked polyethylene at five years: a multicentre radiostereometric analysis study. Bone Joint J 2018; 100-B(12): 1592-9.

Galea V P, Rojanasopondist P, Laursen M, et al. Evaluation of Vitamin E-diffused highly crosslinked polyethylene wear and porous titanium-coated shell stability: a seven-year randomized control trial using radiostereometric analysis. Bone Joint J 2019; 101-B(7): 760-7.

Gaudiani M A, White P B, Ghazi N, et al. Wear rates with large metal and ceramic heads on a second generation highly cross-linked polyethylene at mean 6-year follow-up. J Arthroplasty 2018; 33(2): 590-4.

Gigante A, Bottegoni C, Ragone V, et al. Effectiveness of vitamin-e-doped polyethylene in joint replacement: a literature review. J Funct Biomater 2015; 6(3): 889-900.

Gomoll A, Wanich T, Bellare A. J-integral fracture toughness and tearing modulus measurement of radiation cross-linked UHMWPE. J Orthop Res 2002; 20(6): 1152-6.

Jarrett B T, Cofske J, Rosenberg A E, et al. In vivo biological response to vitamin E and vitamin-E-doped polyethylene. J Bone Joint Surg Am 2010; 92(16): 2672-81.

Kurtz S M. UHMWPE biomaterials handbook. Amsterdam: Elsevier; 2009.

Kurtz S M, Dumbleton J, Siskey R S, et al. Trace concentrations of vitamin E protect radiation crosslinked UHMWPE from oxidative degradation. J Biomed Mater Res - Part A 2009; 90: 549-563.

Lindalen E, Thoen P S, Nordsletten L, et al. Low wear rate at 6-year follow-up of vitamin E-infused cross-linked polyethylene: a randomised trial using 32- and 36-mm heads. HIP Int 2019; 29(4): 355-62.

McKellop H, Shen F W, Lu B, et al. Development of an extremely wear-resistant ultra high molecular weight polyethylene for total hip replacements. J Orthop Res 1999; 17(2): 157-67.

McKellop H, Shen F W, Lu B, et al. Effect of sterilization method and other modifications on the wear resistance of acetabular cups made of ultra-high molecular weight polyethylene: a hip-simulator study. J Bone Joint Surg Am 2000; 82-A(12): 1708-25.

Merola M, Affatato S. Materials for hip prostheses: a review of wear and loading considerations. Materials (Basel) 2019; 12(3): 495.

Nebergall A K, Greene M E, Laursen M B, et al. Vitamin E diffused highly cross-linked polyethylene in total hip arthroplasty at five years. Bone Joint J 2017; 99-B: 577-84.

Oral E, Christensen S D, Malhi A S, et al. Wear resistance and mechanical properties of highly cross-linked, ultrahigh-molecular weight polyethylene doped with vitamin E. J Arthroplasty 2006a; 21:580-591.

Oral E, Rowell S L, Muratoglu O K. The effect of α-tocopherol on the oxidation and free radical decay in irradiated UHMWPE. Biomaterials 2006b; 27(32): 5580-7.

Oral E, Wannomae K K, Rowell S L, et al. Diffusion of vitamin E in ultrahigh molecular weight polyethylene. Biomaterials 2007; 28(35): 522537.

Paavolainen P, Hämäläinen M, Mustonen H, et al. Registration of arthroplasties in Finland. Acta Orthop 1991; 241: 27-30.

Rochcongar G, Buia G, Bourroux E, et al. Creep and wear in Vitamin E-infused highly cross-linked polyethylene cups for total hip arthroplasty: a prospective randomized controlled trial. J Bone Joint Surg Am 2018; 100(2): 107-14.

Rowell S L, Oral E, Muratoglu O K. Comparative oxidative stability of α-tocopherol blended and diffused UHMWPEs at 3 years of real-time aging. J Orthop Res 2011; (29): 773-80.

Salemyr M, Muren O, Ahl T, et al. Vitamin-E diffused highly cross-linked polyethylene liner compared to standard liners in total hip arthroplasty: a randomized, controlled trial. Int Orthop 2015; 39(8): 1499-505.

322

Scemama C, Anract P, Dumaine V, et al. Does vitamin E-blended polyethylene reduce wear in primary total hip arthroplasty: a blinded randomised clinical trial. Int Orthop 2017; 41: 1113-18.

Shareghi B, Johanson P E, Kärrholm J. Wear of vitamin e-infused highly crosslinked polyethylene at five years. J Bone Joint Surg Am 2017; 99: 1447-52.

Sillesen N H, Greene M E, Nebergall A K, et al. Three year RSA evaluation of vitamin e diffused highly cross-linked polyethylene liners and cup stability.

J Arthroplasty 2015; 30(7): 1260-4.

Acta Orthopaedica 2021; 92 (3): 316–322

Sillesen N H, Greene M E, Nebergall A K, et al. 3-year follow-up of a longterm registry-based multicentre study on vitamin E diffused polyethylene in total hip replacement. Hip Int 2016; 26(1): 97-103.

Sköldenberg O G, Rysinska A D, Chammout G, et al. A randomized double-blind noninferiority trial, evaluating migration of a cemented vitamin E-stabilized highly crosslinked component compared with a standard polyethylene component in reverse hybrid total hip arthroplasty. Bone Joint J 2019; 101-B(10): 1192-8.

Acta Orthopaedica 2021; 92 (3): 323–328

323

Epidemiology and mortality of pelvic and femur fractures—a nationwide register study of 417,840 fractures in Sweden across 16 years: diverging trends for potentially lethal fractures Natalie LUNDIN 1, Tuomas T HUTTUNEN 2–4, Anders ENOCSON 1, Alejandro I MARCANO 2, Li FELLÄNDER-TSAI 2, and Hans E BERG 2 1 Department

of Molecular Medicine and Surgery, Karolinska Institute, Karolinska University Hospital, Stockholm, Sweden; 2 Division of Orthopedics and Biotechnology, Department of Clinical Science, Intervention and Technology, Karolinska Institute, Karolinska University Hospital, Stockholm, Sweden; 3 Faculty of Medicine and Health Technology, Tampere University, Tampere University Hospital, Tampere, Finland; 4 Department of Emergency, Anesthesia and Pain Medicine, Tampere University Hospital, Tampere, Finland Correspondence: natalie.lundin@sll.se Submitted 2020-10-01. Accepted 2020-12-29.

Background and purpose — Fractures of the pelvis and femur are serious and potentially lethal injuries affecting primarily older, but also younger individuals. Long-term trends on incidence rates and mortality might diverge for these fractures, and few studies compare trends within a complete adult population. We investigated and compared incidence and mortality rates of pelvic, hip, femur shaft, and distal femur fractures in the Swedish adult population.

Patients and methods — We analyzed data on all adult patients ≥ 18 years in Sweden with a pelvic, hip, femur shaft, or distal femur fracture, through the Swedish National Patient Register. The studied variables were fracture type, age, sex, and 1-year mortality.

Results — While incidence rates for hip fracture decreased by 18% (from 280 to 229 per 105 person-years) from 2001 to 2016, incidence rates for pelvic fracture increased by 25% (from 64 to 80 per 105 person-years).

Incidence rates for femur shaft and distal femur fracture remained stable at rates of 15 and 13 per 105 person-years respectively. 1-year mortality after hip fracture was 25%, i.e., higher than for pelvic, femur shaft, and distal femur fracture where mortality rates were 20–21%. Females had an almost 30% lower risk of death within 1 year after hip fracture compared with males.

Interpretation — Trends on fracture incidence for pelvic and femur fractures diverged considerably in Sweden between 2001 and 2016. While incidence rates for femur fractures (hip, femur shaft, and distal femur) decreased or remained constant during the studied years, pelvic fracture incidence increased. Mortality rates were different between the fractures, with the highest mortality among patients with hip fracture.

Pelvic and femur fractures are potentially lethal to both young and elderly patients (Deakin et al. 2007). The younger multitraumatized patient risks fatal bleeding or other simultaneous mortal injuries after high-energy trauma (Enninghorst et al.

2013). Frail elderly patients exhibit high mortality during the first months after simple falls (Reito et al. 2019). While proximal femur (hip) fractures among the elderly are well studied with respect to incidence and mortality, pelvic and non-hip femur fractures are less well described, and comparisons within a complete population are lacking.

Hip fracture incidence has after many years of steady increase actually stabilized in several Western populations, and even decreased during the last decades, as especially evident in Scandinavia (Cooper et al. 2011, Rosengren et al.

2017, Kannus et al. 2018). Pelvic fractures seem instead to maintain an increasing incidence (Kannus et al. 2015, Melhem et al. 2020). While less frequent than hip and pelvic fractures, it has been suggested that shaft and distal femur fractures are increasing (Ng et al. 2012).

1-year mortality after hip fracture has globally been described to be between 18% and 27%, trending downwards (Downey et al. 2019). Mortality data on pelvic and distal femur fractures points at similar or higher levels, while data on femur shaft fractures is scarce (Streubel et al. 2011, Moloney et al. 2016, Reito et al. 2019). Little has been published regarding mortality for pelvic and femur fractures within whole populations, and to our knowledge no study has compared the incidence and mortality rates within a complete national population.

We investigated and compared the incidence and mortality rates of pelvic, hip, femur shaft, and distal femur fractures in the Swedish adult population over time, including age and sex distribution.

© 2021 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 10.1080/17453674.2021.1878329

324

Acta Orthopaedica 2021; 92 (3): 323–328

Table 1. Number of pelvic and femur fractures in patients ≥ 18 years in Sweden between 2001 and 2016 Year 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Total

Pelvic

Hip

Femur shaft Distal femur

Total

4,472 19,549 1,104 903 26,028 4,493 18,930 1,111 901 25,435 4,287 19,029 1,035 883 25,234 4,482 18,987 1,111 850 25,430 4,680 18,649 1,080 910 25,319 5,171 18,979 1,131 925 26,206 5,401 18,611 1,152 983 26,147 5,479 19,088 1,176 972 26,715 5,592 18,550 1,200 1,005 26,347 6,021 18,745 1,209 947 26,922 6,046 18,665 1,248 993 26,952 5,977 17,926 1,060 895 25,858 6,203 18,128 1,098 943 26,372 6,117 17,635 1,056 890 25,698 6,554 17,922 1,192 966 26,634 6,333 18,098 1,160 952 26,543 87,308 297,491 18,123 14,918 417,840

Patients and methods We used the Swedish National Patient Register (NPR) to find data on all healthcare visits with relevant diagnoses between 2001 and 2016. The NPR was established by the Swedish National Board of Health and Welfare in 1964 and has a high coverage of both in- and outpatients (Ludvigsson et al. 2011).

The register is based on admission/discharge from caregivers, and includes data on personal identity number, age, sex, diagnoses and surgical procedures. Diagnostic codes according to ICD-10 are used by the register.

We performed a search for pelvic fractures, including acetabulum (S32.1, S32.3, S32.4, S32.5, S32.7, S32.8), hip fractures (S72.0, S72.1, S72.2), femur shaft fractures (S72.3), and distal femur fractures (S72.4) from 2001–2016. We included all persons ≥ 18 years with a valid personal identity number and with any of the above fracture codes. Collected variables included age, sex, and diagnosis. Incidence rate was calculated as person-time incidence rate and was expressed as rate per 105 person-years. Statistics regarding the Swedish population were found via the open access register from Statistics Sweden (www.scb.se), and the reported population on July 1st each year was used as a representation for the whole year.

We used the Swedish Cause of Death Register to investigate 1-year mortality in the fracture cohorts. The register contains all deaths registered in Sweden. The national registration number, a unique identifier assigned to all Swedish citizens, allows linkage of data between registers.

The 1st admission for a fracture was regarded as the incident case, and subsequent visits were counted anew if the visit contained another fracture code or was encountered beyond 12 months from the 1st visit. This means patients could be

Table 2. Baseline patient characteristics for pelvic, hip, femur shaft, and distal femur fracture

Factor

Pelvic Hip Femur shaft Distal femur n = 87,308 n = 297,491 n = 18,123 n = 14,918

Sex, n (%) Male 25,495 (29) 95,295 (32) 7,304 (40) 3,994 (27) Female 61,813 (71) 202,196 (68) 10,819 (60) 10,924 (73) Mean age (SD) Total 75 (18) 80 (11) 68 (23) 71 (20) Male 68 (21) 78 (13) 57 (25) 56 (22) Female 78 (16) 82 (10) 76 (18) 77 (16)

included more than once if another fracture type occurred at any time, or the same fracture type occurred after a time frame of 1 year. Patients with concomitant fractures at the same time were included in several fracture groups according to fracture.

Statistics Data was extracted from a pseudonymized SAS database (SAS Institute, Cary, NC, USA) and statistical analysis was done using R version 4.0.0 (R Centre for Statistical Computing, Vienna, Austria). Data was stored in accordance with current GDPR regulations. Statistical testing of differences in mean age was done using a 1-way ANOVA test. The average change in the number of fractures was estimated using Poisson regression.

Ethics, funding, and potential conflicts of interest The study was approved by the regional ethics committee (reference numbers: 2013/581-31/5 and 2016-2251-32) and was performed according to the standards of the 1964 Declaration of Helsinki. Funding was received from the Regional Agreement on Medical Training and Clinical Research between Stockholm County Council and the Karolinska Institute (ALF). None of the authors report any conflict of interest.

Results The total number of pelvic and femur fractures during the study period was 417,840. The 71% hip fractures represented the largest fracture group in this material, followed by the pelvic fractures at 21%. The femur shaft and distal femur fractures represented 4.3% and 3.6% respectively of all fractures (Table 1). The Swedish adult population ≥ 18 years increased by 14% during the studied years (Statistics Sweden). Mean age was between 68 and 80 years in the fracture cohorts with lowest mean age among femur shaft fractures (68 years, SD 23), and highest among hip fractures (80 years, SD 11). Differences in mean age were statistically significant (p < 0.001).

Sex distribution showed a female dominance among all fracture cohorts (60–73%), with lowest proportion of females among femur shaft fractures (Table 2).

Acta Orthopaedica 2021; 92 (3): 323–328

325

Table 3. Mean incidence rate per 105 person-years in 2001–2002 and 2015–2016 for adults aged 18–49 years

Men Women Fracture site 2001–2002 2015–2016 2001–2002 2015–2016 Pelvic Hip Femur shaft Distal femur

16 16 11 17 13 12 6.5 5.8 12 6.6 4.4 2.6 4.6 4.2 2.9 2.3

Incidence per 10 5 person-years pelvic fractures

140 120

Incidence per 10 5 person-years hip fractures

100

400 350

250

60

200

0 2001 2003 2005 2007 2009 2011 2013 2015

Hip

2.2 1.3 1.8 25 20 21

2.0 3.5 2.5 32 22 25

25 Women All Men

20

Women All Men

Femur shaft Distal femur 1.6 2.2 1.8 23 21 21

1.4 1.9 1.6 16 21 20

Incidence per 10 5 person-years distal femur fractures

25

20

Women All Men

15

15

10

10

5

5

0 2001 2003 2005 2007 2009 2011 2013 2015

0 2001 2003 2005 2007 2009 2011 2013 2015

150 100

20

Males 18–49 Females 18–49 Total 18–49 Males ≥ 50 Females ≥ 50 Total ≥ 50

Pelvic

300

80

40

Sex and age

Incidence per 10 5 person-years femur shaft fractures

450 Women All Men

Table 4. 1-year mortality in percentage for pelvic, hip, femur shaft, and distal femur fractures in patients ≥ 18 years in Sweden between 2001 and 2016

50 0 2001 2003 2005 2007 2009 2011 2013 2015

Figure 1. Yearly total incidence rate of all pelvic fractures, hip fractures, femur shaft fractures, and distal femur fractures in patients ≥ 18 years in Sweden 2001–2016.

Pelvic fracture incidence The incidence of pelvic fractures rose by 25% during the study period, from 64 to 80 per 105 person-years (Figure 1).

The average annual increase in the number of fractures was 2.9%. The increase in incidence of pelvic fractures was seen in both males and females, and mainly in the oldest population (Figures 1 and 2). The incidence of pelvic fractures for younger patients (aged 18–49) was low compared with the older age groups, between 11 and 17 per 105 person years.

The incidence of younger females aged 18–49 increased during the study period to the same level as that of the males (Table 3).

Hip fracture incidence The incidence of hip fractures decreased by 18% from 280 to 229 per 105 person-years (Figure 1). The average annual decrease in the number of fractures was 0.5%. The decrease in incidence was mainly due to a gradual decrease in female incidence from 389 to 299 per 105 person-years. The incidence of hip fractures in males was fairly steady and decreased only slightly from 168 to 158 per 105 person-years in 2001–2016 (Figure 1). Hip fracture incidence increased markedly with age (Figure 2). Hip fractures were uncommon in young adults (18–49 years), at between 5.8 and 13 per 105 person-years, nevertheless about twice as frequent in males compared with females but showing no evident changes over the study period (Table 3).

Femur shaft fracture incidence The overall incidence rate was fairly stable at around 15 per 105 person-years during the studied years (Figure 1). The average annual increase in the number of fractures was 0.3%.

Incidence was highest among people aged ≥ 80 years (Figure 2). Femur shaft fractures were more than twice as common in younger males as in females (Table 3). A decrease was seen among males aged 18–49, where the incidence rate was reduced by almost half during the study period (Table 3).

Distal femur fracture incidence The overall incidence rate remained stable at around 13 per 105 person-years during the study period (Figure 1). The average annual increase in number of fractures was 0.4%. Distal femur fracture incidence increased markedly with age, especially in females (Figure 2). Females aged ≥ 80 years had an almost 5-fold incidence of distal femur fracture compared with males in the same age group (Figure 2d), which corresponds to the largest sex difference among the studied fractures. Distal femur fractures among young individuals were rare, at ≤ 5 fractures per 105 person-years (Table 3).

Mortality Unadjusted mortality within 1 year after pelvic, hip, femur shaft, or distal femur fracture was low among the younger population aged 18–49, between 1.3% and 3.5%. (Figure 3).

In the population ≥ 50 years, 1-year mortality was highest

326

Acta Orthopaedica 2021; 92 (3): 323–328

Incidence per 10 5 person-years pelvic fractures 900 Women 2001–2002 Women 2015–2016 Men 2001–2002 Men 2015–2016

800 700

3,000

Women 2001–2002 Women 2015–2016 Men 2001–2002 Men 2015–2016

120

2,000

500

Incidence per 10 5 person-years femur shaft fractures

140

2,500

600

Women 2001–2002 Women 2015–2016 Men 2001–2002 Men 2015–2016

Incidence per 10 5 person-years distal femur fractures

140 120

100

100

80

80

60

60

40

40

20

20

Women 2001–2002 Women 2015–2016 Men 2001–2002 Men 2015–2016

1,500

400

1,000

300 200

500

100 0

Incidence per 10 5 person-years hip fractures

0

18–29 30–39 40–49 50–59 60–69 70–79 80–

18–29 30–39 40–49 50–59 60–69 70–79 80–

Age groups

Age groups

0

18–29 30–39 40–49 50–59 60–69 70–79 80–

Age groups

0

18–29 30–39 40–49 50–59 60–69 70–79 80–

Age groups

Figure 2. Mean incidence rate of all pelvic fractures, hip fractures, femur shaft fractures, and distal femur fractures in patients ≥ 18 years in Sweden 2001–2002 and 2015–2016 per age group.

1-year mortality (%) – age 18–49

1-year mortality (%) – age ≥ 50

4.0

35

3.5

Men Women Total

30

3.0

Men Women Total

25

2.5

20

2.0 15

1.5

10

1.0

5

0.5 0

Pelvic

Hip

Femur shaft Distal femur

Fracture groups

0

Pelvic

Hip

Femur shaft Distal femur

Fracture groups

Figure 3. 1-year mortality after pelvic, hip, femur shaft, and distal femur fractures in patients aged 18–49 and ≥ 50 years in Sweden between 2001 and 2016.

after hip fracture at 25%. 1-year mortality for pelvic fracture was 21% and for femur shaft fracture and distal femur fracture this was 21% and 20% respectively among adults 50 years and older (Table 4). Sex differences in mortality in the older population were mainly seen among hip fracture patients (Figure 3). Males displayed both the highest (32% for hip fractures) and the lowest (16% for distal femur fractures) 1-year mortality within the cohort.

Discussion Our main finding was the diverging incidence trends between pelvic and hip fractures. While hip fracture incidence decreased by 18%, the incidence of pelvic fractures increased by 25% in the Swedish adult population between 2001 and 2016. Femur shaft and distal femur fractures showed marginal overall changes. Pelvic, hip, and other femur fractures remained rare in the young and middle-aged population.

1-year mortality was highest among hip fractures (25%) while similar for the other fracture groups (20–21%); however, sex differences were considerable.

Incidence Our hip fracture data was consistent with previously reported trends showing decreasing incidence rates from around the year 2000 in several European countries (Lucas et al. 2017).

Kannus et al. (2018) reported a declining incidence rate of hip fractures after 1997 in individuals ≥ 50 years, in the entire Finnish population. We found a similar decline, mainly due to decreasing incidence rates among females ≥ 70 years, rendering a decline in the total number of Swedish hip fractures of 7.4%, despite a 14% increase in population numbers between 2001 and 2016.

Pelvic fracture incidence increased markedly during the same years due to steady increases among both older males and females. Accordingly, pelvic fracture was the only fracture type to show a pronounced increase in absolute numbers during the study period (from 4,472 to 6,333; 42%). Other studies have also reported increasing incidences of pelvic fractures (Kannus et al. 2015, Melhem et al. 2020), but numbers are conflicting. In the French population between 2006 and 2016, a 69% increase in the total number of pelvic and acetabular fractures was reported (Melhem et al. 2020). The French study reported lower overall incidence rates within the entire population, including children and adolescents < 18 years.

Studies on incidence rates in large populations of patients with femur shaft fractures are somewhat limited. Our incidence rate of 15 per 105 person-years in 2016 was similar to previously reported numbers (Court-Brown and Caesar 2006, Weiss et al. 2009, Enninghorst et al. 2013). Earlier studies suggest a dominance in young males sustaining femur shaft fractures, and a bimodal distribution with regards to age and sex (Court-Brown and Cesar 2006, Enninghorst et al. 2013).

We found both male and female femur shaft fractures to be more frequent in the older population. The distribution was bimodal during the first study years, but this faded with time due to fewer young males sustaining the fracture. Comparisons between studies are complicated by the differences in age of studied populations, where some studies included children and consequently found a considerably lower mean age

Acta Orthopaedica 2021; 92 (3): 323–328

than in this study, and with a marked incidence peak among younger men (Weiss et al. 2009, Ng et al. 2012, Enninghorst et al. 2013).

The distal femur fracture (incidence rate of 13 per 105 person-years), overall as common as the shaft fracture, showed all signs of an osteoporotic fracture regarding age and sex distribution, with incidence increasing in women older than 60, and a steep rise with age. Interestingly, male fracture incidences were maintained as low throughout life, and were less than 20% of the female incidence even among the very old. These results confirm earlier reported trends of the distal femur fracture (Elsoe et al. 2018).

As expected, the vast majority of all fractures occurred in older individuals and 68% of the patients were female, in agreement with recent studies on the epidemiology of pelvic and hip fractures (Cauley et al. 2014, Kannus et al. 2015, 2018), and underpinned by the fact that hip and pelvic fractures together constituted 92% (71% and 21% respectively) of all counted fractures in our material. The 50% or larger incidence rate of females compared with males ≥ 70 years encountering a pelvic or femur fracture found in our study emphasizes the burden of osteoporosis in the female population.

Pelvic and femur fractures in the younger population, aged 18–49, were rather uncommon, with an incidence of 2.3– 17/105 person-years, with the distal femur fracture being especially uncommon (2.3–4.6/105 person-years). Our numbers confirm previously published rates for younger adults (Farr et al. 2017). While adding only marginally to total fracture numbers, relative changes are still relevant to each patient group.

The most prominent change among younger adults was the fall in femur shaft fracture incidence among males by almost half. Pelvic fractures among young females, conversely, showed a 50% increase in incidence that resulted in an equal rate between sexes. These changes are not easily explained and merit further investigation.

Mortality Mortality among patients ≥ 50 years after hip fracture was higher (25%) than for pelvic, femur shaft, and distal femur fracture (20–21%). Females exhibited an almost 30% lower risk of death within 1 year after hip fracture compared with males. Hip fracture mortality in this study confirmed previous findings with regards to both the overall 1-year risk of death, and sex differences (Abrahamsen et al. 2009, Downey et al.

2019).

Mortality within the 1st year after pelvic fracture has with few exceptions been investigated among inpatients only. A Dutch study of pelvic fractures among inpatients ≥ 65 years found a 1-year mortality of 27% (Banierink et al. 2019), substantially higher than the 21% found in our study. A German study from 2017, including both in- and outpatients, found a 1-year mortality of 21% (Andrich et al. 2017), like our results.

Reports on long-term mortality after femur shaft fracture are scarce, especially among older patients. A recent Swedish

327

study investigated mortality in femur shaft fractures among patients ≥ 65 years and found a 1-year mortality of 21%, as in our study (Wolf et al. 2020). A German study investigating high-energy femur shaft fractures among young and middleaged adults reported an in-hospital mortality of 10% (Kobbe et al. 2013). These numbers are considerably higher than the 1.8% found in our study for adults aged 18–49 years.

Our 20% 1-year mortality after distal femur fracture in adults ≥ 50 years was somewhat lower than previously reported (25%; Streubel et al. 2011, Moloney et al. 2016). However, those studies investigated only surgically treated patients (≥ 60 years), which could explain their higher mortality rates.

Earlier comparisons of mortality rates between hip and nonhip femur fractures either found similar rates (Streubel et al.

2011) or higher mortality rates for hip fractures, as found in our study (Deakin et al. 2007, Abrahamsen et al. 2009, Wolf et al. 2020). A Finnish study compared 30- and 90-day, mortality rates between patients with hip and pelvic fracture and found no differences (Reito et al. 2019). It can only be speculated as to whether the long-term clinical development affects differently the 1-year compared with 30- and 90-day mortality.

Additionally, the treatments of pelvic and femur fractures are highly different and might confound comparisons. Fractures surgically treated, like the vast majority of hip fractures, are of course at more overt medical risk. However, the 1-year mortality reflects the final outcome in these typically old and fragile patients, with a female dominance. In order to unveil the details of the striking high male mortality after hip fracture, or the lower mortality in males after distal femur fracture, future studies, stratified for age, sex, and also surgical procedures, are suggested.

Strengths and limitations The major strength of this study is the large number of included patients within a complete national cohort that allowed for detailed and unbiased comparisons between the different fractures. Moreover, the included patient population was unselected and thus any regional, ethnic, or socioeconomic bias was minimized. The main limitation is potential misclassification in the registry data. Still, the NPR data has been reported to be of high quality and any misclassification should most likely be accidental and therefore any correlation biased “toward the null.” Another limitation concerns lack of information on comorbidities and cause of death, which could not be retrieved from the NPR register. The absence of this data might influence the comparability of the patient cohorts.

The stratification in age and sex intervals of the whole population in each year allowed the unique comparison between different fracture sites among the very same individuals in Sweden for that year. Regarding the demographics of our population, we found the total male–female ratio to be stable throughout the study years, and the different age groups constituted a similar proportion of the entire population, with changes smaller than 3% increase or decrease across the 16

328

years. Moreover, incidence calculations by each year used the exact number of Swedish males or females in that year for comparison.

Conclusion Incidence of hip fractures decreased while incidence of pelvic fractures increased in the Swedish adult population between 2001 and 2016. Mortality within 1 year after fracture was higher for hip fracture patients compared with patients with fracture of the pelvis, femur shaft, or distal femur.

HB, AE, and NL designed the study, analyzed the data, and wrote the manuscript. TH and AIM extracted the relevant data along with aiding in writing the manuscript. LFT was responsible for the database and the handling of the data along with aiding in writing the manuscript.

Acta thanks Charles Court-Brown and Aleksi Reito for help with peer review of this study.

Abrahamsen B, van Staa T, Ariely R, et al. Excess mortality following hip fracture: a systematic epidemiological review. Osteoporos Int 2009; 20(10): 1633-50.

Andrich S, Haastert B, Neuhaus E, et al. Excess mortality after pelvic fractures among older people. J Bone Miner Res 2017; 32(9): 1789-1801.

Banierink H, Ten Duis K, de Vries R, et al. Pelvic ring injury in the elderly: fragile patients with substantial mortality rates and long-term physical impairment. PLoS One 2019; 14(5): e0216809.

Cauley J A, Chalhoub D, Kassem A M, et al. Geographic and ethnic disparities in osteoporotic fractures. Nat Rev Endocrinol 2014; 10(6): 338-51.

Cooper C, Cole Z A, Holroyd C R, et al. Secular trends in the incidence of hip and other osteoporotic fractures. Osteoporos Int 2011; 22(5): 1277-88.

Court-Brown C, Caesar B. Epidemiology of adult fractures: a review. Injury 2006; 37: 691-7.

Deakin D E, Boulton C, Moran C G. Mortality and causes of death among patients with isolated limb and pelvic fractures. Injury 2007; 38(3): 312-17.

Downey C, Kelly M, Quinlan J F. Changing trends in the mortality rate at 1-year post hip fracture: a systematic review. World J Orthop 2019; 10(3): 166-75.

Elsoe R, Ceccotti A A, Larsen P. Population-based epidemiology and incidence of distal femur fractures. Int Orthop 2018; 42(1): 191-6.

Acta Orthopaedica 2021; 92 (3): 323–328

Enninghorst N, McDougall D, Evans J A, et al. Population-based epidemiology of femur shaft fractures. J Trauma Acute Care Surg 2013; 74(6): 1516-20.

Farr J N, Melton L J 3rd, Achenbach S J, et al. Fracture incidence and characteristics in young adults aged 18 to 49 years: a population-based study. J Bone Miner Res 2017; 32(12): 2347-54.

Kannus P, Parkkari J, Niemi S, et al. Low-trauma pelvic fractures in elderly Finns in 1970–2013. Calcif Tissue Int 2015; 97(6): 577-80.

Kannus P, Niemi S, Parkkari J, et al. Continuously declining incidence of hip fracture in Finland: analysis of nationwide database in 1970–2016. Arch Gerontol Geriatr 2018; 77: 64-7.

Kobbe P, Micansky F, Lichte P, et al. Increased morbidity and mortality after bilateral femoral shaft fractures: myth or reality in the era of damage control? Injury 2013; 44(2): 221-5.

Lucas R, Martins A, Severo M, et al. Quantitative modelling of hip fracture trends in 14 European countries: testing variations of a shared reversal over time. Sci Rep 2017; 7(1): 3754.

Ludvigsson J F, Andersson E, Ekbom A, et al. External review and validation of the Swedish national inpatient register. BMC Public Health 2011; 11: 450.

Melhem E, Riouallon G, Habboubi K, et al. Epidemiology of pelvic and acetabular fractures in France. Orthop Traumatol Surg Res 2020; Feb 1. pii: S1877-0568(20)30001-3.

Moloney GB, Pan T, Van Eck C F, et al. Geriatric distal femur fracture: are we underestimating the rate of local and systemic complications? Injury 2016; 47(8): 1732-6.

Ng A C, Drake M T, Clarke B L, et al. Trends in subtrochanteric, diaphyseal, and distal femur fractures, 1984–2007. Osteoporos Int 2012; 23(6): 1721-6.

Reito A, Kuoppala M, Pajulammi H, et al. Mortality and comorbidity after non-operatively managed, low-energy pelvic fracture in patients over age 70: a comparison with an age-matched femoral neck fracture cohort and general population. BMC Geriatr 2019; 19(1): 315.

Rosengren B E, Björk J, Cooper C, et al. Recent hip fracture trends in Sweden and Denmark with age-period-cohort effects. Osteoporos Int 2017; 28(1): 139-49.

Statistics Sweden, Available from: www.scb.se Streubel P N, Ricci W M, Wong A, et al. Mortality after distal femur fractures in elderly patients. Clin Orthop Relat Res 2011; 469(4): 1188-96.

Weiss R, Montgomery S, Al Dabbagh Z, et al. National data of 6409 Swedish inpatients with femoral shaft fractures: stable incidence between 1998 and 2004. Injury 40 (2009) 304-8.

Wolf O, Mukka S, Ekelund J, et al. How deadly is a fracture distal to the hip in the elderly? An observational cohort study of 11,799 femoral fractures in the Swedish Fracture Register. Acta Orthop 2020; 91: 1-7. Online ahead of print.

Acta Orthopaedica 2021; 92 (3): 329–334

329

Femoral lengthening might impair physical function and lead to structural changes in adjacent joints: 10 patients with 27 to 34 years’ follow-up Patrick A BJØRGE 1, Anne-Therese TVETER 2, Harald STEEN 3, Ragnhild GUNDERSON 4, and Joachim HORN 5,6 1 Department of Physiotherapy, Oslo University Hospital, Rikshospitalet, Oslo; 2 National Resource Center for Rehabilitation in Rheumatology, Diakonhjemmet Hospital, Oslo; 3 Biomechanics Laboratory, Oslo University Hospital, Rikshospitalet, Oslo; 4 Department of Radiology, Oslo University Hospital, Rikshospitalet, Oslo; 5 Department of Children´s Orthopaedics and Reconstructive Surgery, Oslo University Hospital, Rikshospitalet, Oslo; 6 Institute of Clinical Medicine, Faculty of Medicine, University of Oslo, Oslo, Norway.

Correspondence: patbjo@ous-hf.no Submitted 2020-10-07. Accepted 2020-11-25.

Background and purpose — Literature describing longterm functional outcome and osteoarthritis (OA) in adjacent joints after femoral lengthening is rare. We evaluated physical function and the presence of radiographic OA in adjacent joints in 10 patients ≥ 27 years after femoral lengthening.

Patients and methods — We conducted a cross-sectional study of 10 patients treated by unilateral femoral lengthening. Follow-up was between 27 and 34 years. Physical function was evaluated by the 30-second sit-to-stand (30sSTS) and a stair test and was compared with reference values. 4 single-legged hop tests were used to assess difference in physical function between the lengthened and contralateral limb. Radiographic OA was evaluated by joint space width (JSW) and Kellgren and Lawrence (KL) classification.

Results — The patients scored worse compared with reference values on the 30sSTS and stair test, and worse on the lengthened limb on the single- and triple-hop test. Radiographic OA was found in the hip or knee in the lengthened limb in 3 of 10 patients based on JSW and 4 of 10 based on KL. No radiographic OA was found in unlengthened limbs.

Interpretation — Our results showed impaired physical function both in general and of the lengthened limb. Additionally, we found a possible association between femoral lengthening and radiographic OA in adjacent joints in the long term. However, the sample size of the current study is small.

Limb lengthening by the callotasis technique is a well-established method for treatment of leg length discrepancy (LLD).

However, literature describing the long-term functional outcome and eventual late side effects such as osteoarthritis (OA) in adjacent joints is rare. Previous studies have investigated whether femoral lengthening impacts muscle strength in the lengthened limb, but results are inconsistent (Bhave et al.

2017, Krieg et al. 2018). Concerns have been raised regarding whether limb lengthening might lead to OA in adjacent joints (Herring 2008, Sneppen et al. 2014). However, to our knowledge there is only 1 published article describing articular damage after femoral lengthening by the callotasis technique.

In this animal study, Stanitski (1994) found cartilage injury in the knees of all included canines after completing 30% distraction of initial femoral length. Even though this animal research showed knee OA after limb lengthening this is yet to be studied in humans. We evaluated physical function and the presence of radiographic signs of OA in the adjacent hip and knee joints ≥ 27 years after unilateral femoral lengthening by the callotasis technique. We hypothesized that femoral lengthening was associated with reduced physical function and OA in adjacent joints.

Patients and methods Design and participants We conducted a cross-sectional study of 10 patients treated with unilateral femoral lengthening between 1985 and 1992.

Patients were enrolled into the study during the period March–October 2019. Inclusion criteria were isolated unilateral femoral lengthening by the callotasis technique and a minimum of 15 years’ follow-up after completed lengthen-

© 2021 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 10.1080/17453674.2020.1866864

330

Acta Orthopaedica 2021; 92 (3): 329–334

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

Demographic data All patients answered a set of sociodemographic questions, including age, sex, employment status, relationship status, and education level. Bodyweight and height and BMI were recorded. LLD (mm) was measured by long standing radiographs. Based on information from the medical journals, LLD was classified as either congenital, developmental, or acquired as described by Castelein and Docquier (2016).

Information from the medical journals was collected to describe age at surgery, preoperative LLD (mm), preoperative alignment, the amount of lengthening (mm), and to calculate the time with fixator mounted (days), and the time from surgery to the last follow-up assessment (years). The consolidation index was expressed as the external fixator index (days with the fixator mounted per cm lengthened) (Brewster et al.

2010). Complications related to the lengthening procedure were described based on information from the medical journals. At assessment physical activity level was measured by the International Physical Activity Questionnaire short form (IPAQ short) and categorized into low, moderate, and high physical activity level according to the guidelines for the IPAQ short (IPAQ Group 2005).

ing—the time considered necessary to detect radiographic secondary OA changes (Schouten et al. 1992). All 50 patients who had undergone isolated femoral lengthening by the callotasis technique between 1977 and 2003 at Oslo University Hospital were identified based on protocols from the surgical department. Patients with inserted knee or hip prosthesis or any condition besides shortening that could lead to alteration of function or joint cartilage were excluded from the study.

These conditions included: all cases of congenital limb shortening with concurrent axis deviation, malrotation or affection of cruciate ligaments or hip, knee, or foot deformity (proximal focal femoral deficiency/congenital femoral deficiency, fibula hemimelia, tibia hemimelia, pes equino varus). Furthermore, cases of acquired shortening with any condition that could have altered function or joint cartilage in the long term were excluded. These conditions included: hip dysplasia, Perthes disease, epiphysiolysis capitis femoris, posttraumatic shortening with joint involvement, and post-infectious shortening after septic arthritis. Patients undergoing lengthening using the Wagner method (from 1977 to 1985) were also excluded, as this method is not based on the principles of the callotasis technique. The inclusion criteria left us with cases of pure shortening due to either idiopathic leg length discrepancy, posttraumatic shortening without joint involvement either at the time of injury or at follow-up, or congenital hypoplasia or hyperplasia without axis deviation, or any pathology at the hip, knee, or foot. Following our strict inclusion criteria, 16 patients out of 50 were found to be eligible for inclusion of whom 10 consented to participate in the study (Figure 1 and Table 1, see Supplementary data).

Functional tests and measurements All patients were summoned for assessment at Oslo University Hospital. A physiotherapist (PB) did the measurements in all functional procedures.

To provide a measure of lower extremity functional strength, a 30-second sit-to-stand test (30sSTS) was conducted as described by Jones et al. (1999). To assess functional aerobic capacity, a revised stair test was performed as described by Tveter et al. (2014a), measuring submaximal cardiopulmonary endurance. Both tests are shown to be valid and reliable in patients with various musculoskeletal conditions (Tveter et al. 2014b).

To compare physical function between the lengthened and unlengthened limb, 4 single-legged hop tests were performed as described by Noyes et al. (1991) and Barber et al. (1990).

The lengthened and unlengthened limb were tested twice, always starting with the unlengthened limb. The mean of 2 tests on each limb was used in the analyses. Limb symmetry index (LSI) was calculated for the single-, triple-, and crossover hop test by dividing the mean of the lengthened limb by the mean of the unlengthened limb and multiplying the result by 100. For the timed hop test a low value (time spent) represents better performance, unlike the other 3 tests where a high value (cm hopped) is best. Hence the inverse ratio was used by dividing the mean of the unlengthened limb by the mean of the lengthened limb and the result multiplied by 100. An index of ≥ 85% has been described as normal function regardless of sports activity level, sex, and dominant side (Barber et al. 1990). The hop tests are shown to be valid and reliable in patients following anterior cruciate ligament reconstruction (Reid et al. 2007).

Patients operated with isolated unilateral femoral lengthening by callotasis technique between 1977 and 2003 n = 50 Did not meet inclusion criteria after medical journal review n = 34 Eligible n = 16 Excluded: Unknown address n=1 Invited n = 15 Excluded (n = 5): – did not meet inclusion criteria, 1 – withdrawn consent, 1 – did not answer, 3 Included in analysis n = 10

Acta Orthopaedica 2021; 92 (3): 329–334

A

331

B

Figure 2. Patient number 5 in the supplementary data table. (A) The right knee with tibiofemoral and femoropatellar Kellgren and Lawrence (KL) grade 2/osteophyte. (B) The left knee with tibiofemoral KL grade 1 and femoropatellar KL grade 0.

Radiography Anteroposterior radiographs of the pelvis were taken with the patients in supine position. Film-to-focus distance was 130 cm, and all radiographs were centered 3 cm above the pubic symphysis and included the pelvis with both hips (Terjesen and Gunderson 2012). For the knees we obtained standing anteroposterior and lateral radiographs. The Syna-Flexer frame (Synarc Inc, Newark, CA, USA) was used for standardized fixed flexion positioning of the anteroposterior knee (20º flexion and 5º external foot rotation). This frame has been validated for joint space width (JSW) measurement (Kothari et al.

2004). Radiographic OA was defined as structural changes in hip and knee joints. For the hips and tibiofemoral joints OA was evaluated by measuring minimum JSW in mm and by Kellgren and Lawrence classification (KL) (Kellgren and Lawrence 1957). For the femoropatellar joint only KL was used as lateral knee radiographs are unsuited for JSW measurement.

Minimum JSW was registered as the narrowest part of the upper, weight-bearing part of the hip joint and the smallest JSW of the anteroposterior knee. JSW of < 2.0 mm and KL grade 2–4 were defined as radiographic OA. For KL grading of the knees we used the precision proposed by Felson et al. (2011) dividing KL2 into KL2/osteophyte when only definite osteophyte was present, and KL2, defined as definite osteophyte and possible narrowing of joint space (Figure 2). We reported the latter only as OA. As axis deviation might occur during limb lengthening and at the same time is considered a risk factor for development of knee OA (Brouwer et al. 2007), anteroposterior and lateral long standing radiographs were taken and used to describe alignment by the zones of the mechanical axis (±1 to ±3) as described by Stevens et al. (1999). Negative zones 2–3 were classified as varus, positive zones 2–3 as valgus, and zone ±1 as normal alignment. These radiographs were taken

with patella pointed forwards to eliminate rotation of the lower extremities as a source of error (Paley 2005).

Statistics Data were analyzed using SPSS Statistics version 26 (IBM Corp, Armonk, NY, USA). Results were presented as median (range) if continuous and frequency (%) if categorical. The paired Wilcoxon signed ranks test was used to analyze differences in performance between the patients and sex- and agematched reference values (Tveter 2014a) for the 30sSTS and stair tests, and to analyze differences between the lengthened and unlengthened limb concerning the hop tests. Radiographic OA was presented by descriptive statistics.

Ethics, registration, funding, and potential conflicts of interests Ethical approval was granted by the Regional Committee for Medical and Health Research Ethics in Norway (REK South East B 2018/416, date of issue April 23, 2018) and the study is registered in ClinicalTrials.gov (NCT03966573). The patients gave their written consent before participation. This work was supported by Sophies Minde Ortopedi AS, grant number 07/2018 and 06/2019. The authors have no conflicts of interest to declare.

Results 10 patients (7 females) were included in the study. The median time after completed lengthening was over 30 years; 9 out of 10 were in their 40s at assessment and the median lengthening was just below 40 mm (Table 2). Based on data calculated from the IPAQ short, 6 of 10 patients had low, 2 moderate,

332

Acta Orthopaedica 2021; 92 (3): 329–334

Table 2. Demographic data for patients treated by unilateral femoral lengthening with callotasis (N = 10). Values are median (range) unless otherwise indicated Characteristic Value Sex, female/male Age at surgery, years Age, years Years between surgery and assessment Height, cm Weight, kg BMI Working, n Living in partnership, n Higher education (college/university), n LLD preoperative, mm Present LLD, mm Lengthened, mm Days with fixator mounted External fixator index, days/cm Etiology, n Congenital LLD

Hypoplasia

Hemihyperplasia Developmental LLD

Idiopathic Acquired LLD

Post traumatic

Sequela osteomyelitis

7/3 14 (13–27) 46 (44–54) 30.5 (27–34) 165 (157–176) 75 (58–94) 28 (22–35) 7 8 6 38.5 (30–55) 6 (–10 to 40) a 38.5 (30–57) 215 (135–374) 52 (39–125)

Table 3. Alignment graded after Stevens et al. (1999) for patients treated by unilateral femoral lengthening with callotasis (N = 10) Alignment grade

2 3 1

LLD = limb length discrepancy. a Minus sign indicates overcorrection.

and 2 had high physical activity level at follow-up. Patients had been treated with unilateral femoral lengthening based on the callotasis technique by use of an Orthofix monolateral fixator. Preoperative long standing radiographs were not available; however, according to the medical journals 1 patient had pre-existing malalignment described as moderate knee varus, which was not operatively addressed. At assessment, a certain degree of frontal plane malalignment was found in several limbs both on the lengthened and unlengthened side (Table 3).

4 complications occurred related to the lengthening, including 2 fractures of the regenerate that were resolved by the end of treatment (1 nonoperatively treated and 1 that required a secondary procedure), 1 delayed consolidation that required bone grafting to consolidate followed by a fracture through the regenerate that resulted in in a knee varus zone -2, and 1 axis deviation allowed to heal in valgus zone +2.

Physical function The patients scored worse on both the 30sSTS (p = 0.008) and the stair test (p = 0.007) compared with reference values (Table 4). For the stair test, the median pulse was 146 (107– 173) bpm and BORG ratings of perceived exertion 14 (12–16) immediately after the test.

The patients performed worse on the lengthened limb in all 4 hop tests (Table 5). However, the results were only statistically significant for the single (p = 0.005) and triple hop for distance (p = 0.007). 5 of 10 patients had impaired physical

Unlengthened limb

7 1 2

8 0 2

Normal (±1) Valgus (+2) Varus (–2)

Table 4. Difference from sex- and age-matched reference values for 30 seconds sit-to-stand test (30sSTS) and stair test. Values are median (range)

Clinical field tests Score 30sSTS, no. of repetitions Stair test, s a

2 2

Lengthened limb

Difference from reference values p-value a

14 (12–25) –11.5 (–17 to 0) 48 (34–56) –13 (–25 to 1)

0.008 0.007

Wilcoxon signed ranks test.

Table 5. Difference between the lengthened and unlengthened limb for the 4 single-legged hop tests. Values are median (range)

Limb Single-legged Lengthened Unlengthened symmetry hop tests limb limb p-value a index (%) b Single, m 0.43 (0–1.0) 0.76 (0.48–1.1) Triple, m 1.8 (0–3.5) 3.1 (2.1–3.7) Timed, s 3.5 (0–7) 3.9 (3–5) Cross-over, m 1.5 (0–3.2) 2.4 (1.5–2.9) a Wilcoxon signed ranks test. b Limb symmetry index is the

unlengthened limb.

0.005 0.007 0.8 0.07

73 (0–100) 74 (0–101) 81 (0–112) 79 (0–114)

median % of the performance on the

function (LSI < 85%) of the lengthened compared with the unlengthened limb on the single, cross-over, and timed hop test, and 6 of 10 on the triple hop test.

Osteoarthritis We found radiographic OA in the hip or knee in the lengthened limb in 3 out of 10 patients (1 hip, 2 knees) when based on JSW, and in 4 of 10 patients (2 hips, 2 knees) when based on KL. Those 2 patients with knee OA based on KL had both tibiofemoral and femoropatellar OA. The 3 patients identified

by JSW also fulfilled the criteria based on KL. No patients had radiographic OA in both the hip and the knee in the lengthened limb, and no radiographic OA was found in the hip or the knee in unlengthened limbs regardless of the evaluation method. The aforementioned patient with preexisting moderate knee varus according to the medical journals showed a medial axis deviation corresponding to zone –2 on the long standing radiographs that were obtained for the current study.

However, no signs of OA were found in this patient’s hips or knees at assessment.

Acta Orthopaedica 2021; 92 (3): 329–334

Discussion Our results indicate that femoral lengthening may impair physical function in general, and/or physical function of the lengthened limb, and possibly lead to signs of radiographic OA in adjacent joints in the long term.

The results from the 30sSTS and stair test indicate a difference between the patients and reference values that extends beyond the measurement error as reported in patients with various musculoskeletal conditions (Tveter 2014b). Comparison of physical function between patients treated by femoral lengthening and age- and sex-matched reference values from a normal population has to our knowledge not previously been performed. Hence, our results are important as they indicate that femoral lengthening might lead to reduced general physical function. The patients were more sedentary and had higher BMI than the reference material, which on one hand could be a consequence of the lengthening procedure. On the other hand, we cannot rule out that the sedentary lifestyle could be random and have led to the reduced physical function without association with the lengthening procedure.

Even though the results are not unambiguous, 2 of 4 hop tests showed a statistically significant difference between the limbs and half of the patients had impaired physical function of the lengthened compared with the unlengthened limb at assessment. Our results are important as the literature is inconsistent in whether femoral lengthening impacts physical function of the lengthened limb or not (Bhave et al. 2017, Krieg et al. 2018). On one hand, we cannot rule out that the difference between the limbs was already present before the lengthening procedure as we do not have preoperative measurements. Preexisting differences have previously been described by Krieg et al. (2018), who found that the shorter limb was weaker than the longer both before and 2 years after femoral lengthening. On the other hand, the patients in the study by Krieg et al. (2018) could have had etiologies associated with impaired physical function of the lengthened limb, a weakness accounted for by the strict inclusion criteria in our study. Thus, the fact that we found impaired physical function of the lengthened limb in addition to the significant difference between the limbs in 2 of 4 hop tests strengthens the assumption of reduced physical function associated with the lengthening procedure. Our study adds to the literature suggesting that femoral lengthening might be associated with impaired physical function of the lengthened limb.

Furthermore, our results indicate a possible association between femoral lengthening and radiographic OA. The association between femoral lengthening and radiographic OA has to our knowledge not previously been described in humans. Our findings are in line with both animal research (Stanitski 1994) and assumptions in textbooks describing limb lengthening procedures (Herring 2008, Sneppen et al. 2014).

However, we must acknowledge that the presence of radio-

333

graphic OA in our sample could be random and explained as “natural history,” as all patients were in an age group at risk of developing OA (Sakellariou et al. 2017). In addition, we have to make reservations for the results in 1 of the patients with radiographic knee OA because of varus alignment outside the normal ranges in the lengthened limb at assessment, a known risk factor for development of knee OA (Brouwer et al. 2007). However, we believe that the absence of radiographic OA in unlengthened limbs despite the literature suggesting an association between LLD and hip OA in the longer limb (Gofton and Trueman 1971), in addition to the fact that 2 patients had varus alignment outside the normal ranges in the unlengthened limb at assessment, indicate a possible association between the lengthening procedure and development of radiographic OA in the long term.

The major limitation to our study was the small number of patients included, a consequence of our strict inclusion criteria. The patients were identified from surgical protocols containing all patients who had undergone limb lengthening at Oslo University Hospital since the first procedure in 1977, making it unlikely that any eligible patients were missing.

However, restricting inclusion to patients without etiology of LLD, which may be associated with reduced physical function and OA, made it more likely that our results could be associated with the lengthening procedure itself. A further limitation was that neither the physiotherapist nor the radiologist was blinded in terms of which limb had undergone a lengthening procedure, as this could be revealed by scars after the fixator and in some cases by persisting radiological bony features due to the lengthening.

In conclusion, our results showed impaired physical function both in general and of the lengthened compared with the unlengthened limb in 10 patients treated with unilateral femoral lengthening. Additionally, our results indicate a possible association between femoral lengthening and radiographic OA in adjacent joints in the long term. When planning femoral lengthening procedures patients must be informed of the fact that we cannot rule out a possible risk of impaired physical function and development of OA in adjacent joints in the long term. However, further research is needed, and we acknowledge that, based on the small sample, we must read our results with caution without the ability to generalize.

Supplementary data Table 1 (general table) is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17 453674.2020.1866864

PB, JH, and HS initiated the study and selected the patients based on the inclusion criteria. PB did the measurement in all functional procedures at assessment and wrote the manuscript. RG did all radiographic measurements. PB and ATT did the analyses and interpretation of data. JH, RG, ATT, and HS revised the manuscript.

334

The authors are grateful to the participating patients for their collaboration.

Acta thanks Søren Kold and Björn Harald Tjernström for help with peer review of this study.

Barber S, Noyes F R, Mangine R E, McCloskey J, Hartman W. Quantitative assessment of functional limitations in normal and anterior cruciate ligament-deficient knees. Clin Orthop Relat Res 1990; (255): 204-14.

Bhave A, Shabtai L, Woelber E, Apelyan A, Paley D, Herzenberg J E. Muscle strength and knee range of motion after femoral lengthening: 2- to 5-year follow-up. Acta Orthop 2017; 88(2): 179-84.

Brewster M B S, Mauffrey C, Lewis A C, Hull P. Lower limb lengthening: is there a difference in the lengthening index and infection rates of lengthening with external fixators, external fixators with intramedullary nails or intramedullary nailing alone? A systematic review of the literature. Eur J Orthop Surg Traumatol 2010; 20(2): 103-8.

Brouwer G, Tol A V, Bergink A, Belo J, Bernsen R, Reijman M, Pols H, Bierma–Zeinstra S. Association between valgus and varus alignment and the development and progression of radiographic osteoarthritis of the knee.

Arthritis Rheumatol 2007; 56(4): 1204-11.

Castelein S, Docquier P-L. Complications associated with bone lengthening of the lower limb by callotasis. Acta Orthop Belg 2016; 82(4): 806-13.

Felson D T, Niu J, Guermazi A, Sack B, Aliabadi P. Defining radiographic incidence and progression of knee osteoarthritis: suggested modifications of the Kellgren and Lawrence scale. Ann Rheum Dis 2011; 70(11): 1884-6.

Gofton J, Trueman G. Studies in osteoarthritis of the hip, Part II: Osteoarthritis of the hip and leg-length disparity. Can Med Assoc J 1971; 104(9): 791.

Herring J A. Tachdjian’s pediatric orthopaedics. 4th ed. Philadelphia: Elsevier; 2008.

IPAQ Group. Guidelines for the data processing and analysis of the “International Physical Activity Questionnaire”; 2005.

Jones C J, Rikli R E, Beam W C. A 30-s chair–stand test as a measure of lower body strength in community-residing older adults. Res Q Exerc Sport 1999; 70(2): 113-19.

Kellgren J, Lawrence J. Radiological assessment of osteo-arthrosis. Ann Rheum Dis 1957; 16(4): 494.

Acta Orthopaedica 2021; 92 (3): 329–334

Kothari M, Guermazi A, von Ingersleben G, Miaux Y, Sieffert M, Block J E, Stevens R, Peterfy C G. Fixed-flexion radiography of the knee provides reproducible joint space width measurements in osteoarthritis. Eur Radiol 2004; 14(9): 1568-73.

Krieg A H, Gehmert S, Neeser O L, Kaelin X, Speth B M. Gain of length— loss of strength? Alteration in muscle strength after femoral leg lengthening in young patients: a prospective longitudinal observational study. J Pediatr Orthop B 2018; 27(5): 399-403.

Noyes F R, Barber S D, Mangine R E. Abnormal lower limb symmetry determined by function hop tests after anterior cruciate ligament rupture. Am J Sports Med 1991; 19(5): 513-18.

Paley D. Principles of deformity correction 1st ed. New York: Springer; 2005.

Reid A, Birmingham T B, Stratford P W, Alcock G K, Giffin J R. Hop testing provides a reliable and valid outcome measure during rehabilitation after anterior cruciate ligament reconstruction. Phys Ther 2007; 87(3): 337-49.

Sakellariou G, Conaghan P G, Zhang W, Bijlsma J W, Boyesen P, D’Agostino M A, Doherty M, Fodor D, Kloppenburg M, Miese F. EULAR recommendations for the use of imaging in the clinical management of peripheral joint osteoarthritis. Ann Rheum Dis 2017; 76(9): 1484-94.

Schouten J, Van den Ouweland F, Valkenburg H. A 12 year follow up study in the general population on prognostic factors of cartilage loss in osteoarthritis of the knee. Ann Rheum Dis 1992; 51(8): 932-7.

Sneppen O, Bünger C, Hvid I, Søballe K. Ortopædisk kirurgi. Copenhagen: FADL’s Forlag; 2014.

Stanitski D F. The effect of limb lengthening on articular cartilage: an experimental study. Clin Orthop Relat Res 1994; (301): 68-72.

Stevens P M, Maguire M, Dales M, Robins A. Physeal stapling for idiopathic genu valgum. J Pediatr Orthop 1999; 19(5): 645.

Terjesen T, Gunderson R B. Radiographic evaluation of osteoarthritis of the hip: an inter-observer study of 61 hips treated for late-detected developmental hip dislocation. Acta Orthop 2012; 83(2): 185-9.

Tveter A T, Dagfinrud H, Moseng T, Holm I. Health-related physical fitness measures: reference values and reference equations for use in clinical practice. Arch Phys Med Rehabil 2014a; 95(7): 1366-73.

Tveter A T, Dagfinrud H, Moseng T, Holm I. Measuring health-related physical fitness in physiotherapy practice: reliability, validity, and feasibility of clinical field tests and a patient-reported measure. J Orthop Sports Phys Ther 2014b; 44(3): 206-16.

Acta Orthopaedica 2021; 92 (3): 335–340

335

T2 relaxation times of knee cartilage in 109 patients with knee pain and its association with disease characteristics Joost VERSCHUEREN 1,2 a, Stephan J VAN LANGEVELD 1 a, Jason L DRAGOO 3, Sita M A BIERMA-ZEINSTRA 1,4, Max REIJMAN 1, Garry E GOLD 5–7 b, and Edwin H G OEI 2 b 1 Department

of Orthopedic Surgery, Erasmus MC University Medical Center Rotterdam, The Netherlands; 2 Department of Radiology & Nuclear Medicine, Erasmus MC University Medical Center Rotterdam, The Netherlands; 3 Department of Orthopedic Surgery, University of Colorado, Denver, CO, USA; 4 Department of General Practice, Erasmus MC University Medical Center Rotterdam, The Netherlands; 5 Department of Radiology, Stanford University Medical Center, CA, USA; 6 Department of Bioengineering, Stanford University Medical Center, CA, USA; 7 Department of Orthopedic Surgery, Stanford University Medical Center, CA, USA a Shared first authorship. b Shared co-senior authorship.

Correspondence: e.oei@erasmusmc.nl Submitted 2020-06-11. Accepted 2021-01-04.

Background and purpose — Quantitative T2 mapping MRI of cartilage has proven value for the assessment of early osteoarthritis changes in research. We evaluated knee cartilage T2 relaxation times in a clinical population with knee complaints and its association with patients and disease characteristics and clinical symptoms.

Patients and methods — In this cross-sectional study, T2 mapping knee scans of 109 patients with knee pain who were referred for an MRI by an orthopedic surgeon were collected. T2 relaxation times were calculated in 6 femoral and tibial regions of interest of full-thickness tibiofemoral cartilage. Its associations with age, sex, BMI, duration of complaints, disease onset (acute/chronic), and clinical symptoms were assessed with multivariate regression analysis. Subgroups were created of patients with abnormalities expected to cause predominantly medial or lateral tibiofemoral cartilage changes.

Results — T2 relaxation times increased statistically significantly with higher age and BMI. In patients with expected medial cartilage damage, the medial femoral T2 values were significantly higher than the lateral; in patients with expected lateral cartilage damage the lateral tibial T2 values were significantly higher. A traumatic onset of knee complaints was associated with an acute elevation. No significant association was found with clinical symptoms.

Interpretation — Our study demonstrates age, BMI, and type of injury-dependent T2 relaxation times and emphasizes the importance of acknowledging these variations when performing T2 mapping in a clinical population.

Knee osteoarthritis (OA) is currently mainly diagnosed on clinical presentation (Hunter and Bierma-Zeinstra 2019). Conventional radiography depicts morphological articular cartilage changes indirectly and is insensitive to both early-stage OA and subtle progression of the disease (Guermazi et al. 2011).

MRI is able to visualize articular cartilage directly and is therefore more sensitive to osteoarthritic changes (Chan et al. 1991).

But, similar to conventional radiography, conventional MRI relies primarily on the identification of morphological changes in damaged knee cartilage and is also limited to depicting relatively advanced signs of degeneration (McCauley et al. 2001).

In the last 2 decades, innovative quantitative methods of MRI have been developed that have the potential to measure articular cartilage degeneration prior to morphological cartilage damage and, thus, might be able to identify cartilage at risk of developing irreversible cartilage damage (Matzat et al. 2013).

A well-validated and quantitative MRI technique, transverse relaxation time (T2) mapping, is regarded as the best technique to determine the hydration content, collagen fiber orientation, and collagen network integrity in articular cartilage (Oei et al.

2014). These cartilage properties are known to be altered in the initial stages of OA development (Setton et al. 1999). T2 mapping is expressed in T2 relaxation times, which tend to increase with more advanced stages of cartilage damage (Dunn et al. 2004). The technique is widely used in scientific studies such as the Osteoarthritis Initiative (OAI) (Joseph et al. 2015).

However, as current T2 mapping data is gathered in research settings with clear inclusion criteria based on age, sex, type of knee disorder, and OA stage, these results cannot directly be generalized to clinical practice (Joseph et al. 2015). Therefore, we assessed the association of T2 relaxation times of knee articular cartilage with patient and disease characteristics and

© 2021 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 10.1080/17453674.2021.1882131

336

Acta Orthopaedica 2021; 92 (3): 335–340

lateral compartment, respectively) of a sagittal T2 weighted sequence, which was part of the routine MR protocol (Figure 1). We used this sequence for segmentation because of better contrast between the cartilage and the surrounding tissue.

The T2 mapping scan was subsequently registered to the T2 weighted scan using rigid registration to calculate T2 relaxation times in the segmented masks. The masks were further divided into a femoral weight-bearing, tibial weight-bearing, and femoral posterior region of interest (ROI) for both the medial and lateral knee compartment. The outer perimeters of the menisci demarcated the weight-bearing ROIs of the femur and tibia. The posterior ROIs contained the femoral cartilage behind the posterior border of the menisci. The 6 ROIs were also combined to calculate an average tibiofemoral T2 relaxation time for each knee.

Figure 1. Cartilage segmentation on a T2-weighted image of the lateral compartment. Red area displays the femoral and tibial cartilage; white boxes represent the ROIs. Abbreviations: Fem_wb: weight-bearing femoral condyle; Fem_post: posterior femoral condyle; Tib_wb: weightbearing tibial plateau.

clinical symptoms in an unselected routine clinical population of patients with knee complaints.

Patients and methods In a period of 18 months, all patients with complaints of knee pain referred for MRI of the knee by an orthopedic surgeon (JLD) from Stanford University Medical Center were eligible for the study.

Image acquisition The patients were scanned on a 3.0 Tesla (T) MRI scanner (MR 750, GE Healthcare, Milwaukee, WI, USA) with a flexible 16-channel receive-only coil (NeoCoil, Pewaukee, WI, USA). The patient’s knee was fixed with a leg holder in slight flexion to position the coil and reduce motion artifacts. In addition to a routine clinical knee MR protocol used by the radiologist to assess structural changes in the knee, a 3D fast spin echo T2 mapping sequence was added to the protocol during the trial period. This sequence with variable refocusing flip angle schedules uses T2 magnetization preparation followed by pseudo steady-state 3D FSE acquisition (Chen and Hat 2011, Matzat et al. 2015). The main T2 mapping sequence parameters were: 5 echo times (6, 12, 25, 38, 64 ms); 3 mm slice thickness; an in-plane resolution of 0.5 x 0.8 mm; and a scan time of approximately 6 minutes.

Image analysis The T2 mapping images were analyzed using an in-house developed MATLAB software tool (Bron et al. 2013). Fullthickness tibiofemoral cartilage masks were segmented on 6 slices (3 central slices of the medial and 3 central slices of the

Patient and disease analysis Patient characteristics (age, sex, and BMI), disease characteristics (diagnosis, duration of complaints, and onset of disease), and clinical symptoms were retrospectively collected through the electronic patient record. Diagnosis was based on the surgical report (when available), clinical report, and MRI report.

The surgical report was considered the reference in case of discrepancies between the reports. The duration of complaints, defined as the period between the onset of knee pain and the date of the MRI, was divided into acute (< 1 month), subacute (1–6 months), and chronic (> 6 months). The onset of disease was specified as traumatic versus non-traumatic. To assess clinical symptoms, the Knee Injury and Osteoarthritis Outcome Score (KOOS) questionnaire was recorded for patients on their first visit to the Outpatient Clinic (Roos et al.

1998). In addition to the KOOS subscale (“symptoms’, “pain,” “activities of daily living,” “sport and recreation,” “quality of life”) scores, all 42 items of the KOOS were dichotomized into absence versus presence of knee complaints. When patients scored zero (i.e., no complaints), the complaint was considered absent, while a score of 1 to 4 indicated presence of the complaint. The KOOS questionnaire was disregarded when it was filled in more than 6 months before the MRI. Statistics Statistical analysis was performed using SPSS (IBM SPSS Statistics for Windows, Version 21.0; IBM Corp, Armonk, NY, USA). Associations between T2 relaxation times and patient characteristics, disease characteristics, and clinical symptoms were tested using linear regression models. T2 relaxation times were used as dependent variable and patient characteristics, disease characteristics, and clinical symptoms as independent variables. We performed both univariate and multivariate analyses. Subgroups were created of patients with abnormalities expected to cause predominantly isolated medial (medial meniscal tear, medial bone marrow edema, or medial focal cartilage/osteochondral damage/degeneration) or lateral tibiofemoral cartilage changes (lateral meniscal tear, lateral bone marrow

Acta Orthopaedica 2021; 92 (3): 335–340

Table 1. Population characteristics

Patient characteristics (n = 109) Male, n (%) 62 (57) Age, years (SD) 41 (14)

(range) (16–77) BMI (SD) 26 (5) Disease characteristics (n =109) Knee disorder causing medial tibiofemoral cartilage changes, n a 35

Medial meniscus injury 26

Medial bone marrow edema 6

Medial focal cartilage/osteochondral damage 9

Medial cartilage degeneration 4 Knee disorder causing lateral tibiofemoral cartilage changes, n a 21

Lateral meniscus injury 17

Lateral bone marrow edema 2

Lateral focal cartilage/osteochondral damage 5

Lateral cartilage degeneration 4 Duration of complaints, n (%) < 1 month 18 (17) 1–6 months 22 (20) > 6 months 69 (63) Onset of disease, n (%) Traumatic 47 (43) Clinical symptoms (n = 47) KOOS subscales, score (0–100) (SD)

Symptoms 63 (18)

Pain 45 (22)

Activities of daily living 66 (16)

Sports 74 (20)

Quality of life 35 (20) T2 relaxation times (n = 109), ms (SD) Femoral and tibial cartilage 40 (3) Weightbearing femoral condyle medial 41 (6) Posterior femoral condyle medial 38 (4) Weightbearing tibial plateau medial 40 (5) Weightbearing femoral condyle lateral 40 (5) Posterior femoral condyle lateral 37 (5) Weight bearing tibial plateau lateral 41 (6) a

Patients can have more than 1 diagnosis.

edema, or lateral focal cartilage/osteochondral damage/degeneration) (Su et al. 2013, Crema et al. 2014). When patients had abnormalities in both compartments of the knee, they were not included in the subgroups. Differences between the medial and lateral ROIs were tested with a paired t-test. Multiple imputations analysis was used for missing data. A p-value of less than 0.05 was considered statistically significant.

Ethics, funding, data sharing, and potential conflicts of interest The study was approved by the institutional review board of Stanford University Medical Center (protocol number 26840).

Informed consent was obtained from all participants. 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 authors upon reasonable request. GEG and EHGO receive research support from GE Healthcare. The study was performed during the visiting professorship of EHGO at Stan-

337

ford University Medical Center, which was partially funded by the Dutch Arthritis Foundation.

Results 146 patients met the inclusion criteria of whom 109 were eligible for further analyses (Table 1). Main reasons for exclusion were no T2 mapping scan undertaken or insufficient quality of this scan due to metal and movement artifacts, which occurred relatively frequently because a surface coil was used instead of a dedicated knee coil. In 8 patients both knees were scanned. The most troublesome knee was included for analysis. The KOOS questionnaire was available for 55 subjects, as not all participants filled in the questionnaire at their first visit to the orthopedic surgeon. 8 questionnaires were disregarded because of the time interval with the MRI scan. No statistically significant differences were found in patient and disease characteristics between the patients with and without a KOOS questionnaire (data not reported).

Patient characteristics Data on BMI was missing for 3 patients. In the multivariate analysis with age, sex, and BMI as independent variables, age showed a statistically significant association with T2 relaxation times in all medial ROIs and the lateral weight-bearing tibial ROI, as well as the total tibiofemoral cartilage (Table 2). Increasing T2 relaxation times were seen with higher age.

BMI showed a significant association with the total tibiofemoral cartilage. In the ROI analyses only, a significant association was seen in the lateral weight-bearing tibial cartilage. Sex did not seem to have an effect on T2 relaxation times. Figure 2 shows the scatter plots of age and BMI, respectively, with T2 relaxation times of the total tibiofemoral cartilage with the corresponding trend lines based on the (univariate) Pearson correlation coefficients.

Disease characteristics We identified 35 patients with abnormalities that are the likely cause of medial cartilage damage. The medial femoral ROIs showed statistically significantly higher T2 relaxation times compared with the lateral femoral ROIs (Table 3). 21 patients were expected to have predominantly lateral cartilage damage.

Statistically significantly higher T2 values were seen only in the lateral weight-bearing tibial ROI. A trend towards decreased cartilage T2 values with an increase in duration of complaints was observed (Figure 3).

However, this association was not statistically significant when the analysis was adjusted for age, sex, and BMI. In the case of a traumatic onset of knee pain, T2 relaxation times were the highest in patients with the shortest time between the onset and MRI acquisition. There was a gradual decline in T2 relaxation times between the MRIs undertaken in < 1 month, 1–6 months, and > 6 months after a traumatic onset. In

Acta 2021; 92 92 (3): (3): 335–340 335–340 Acta Orthopaedica Orthopaedica 2021;

338 338

Table 2. Multivariate linear regression of patient characteristics on total cartilage T2 values Age β (95% CI) Medial Femur weightbearing Femur posterior Tibia weightbearing Lateral Femur weightbearing Femur posterior Tibia weightbearing Total

BMI β (95% CI)

p-value

0.34 (0.16 to 0.53) 0.09 (0.18 to 0.54) 0.26 (0.07 to 0.45)

< 0.01 < 0.01 0.01

p-value

–0.02 (–0.20 to 0.16) 0.07 (–0.11 to 0.25) 0.07 (–0.11 to 0.26)

0.16 (–0.03 to 0.35) 0.1 0.00 (–0.20 to 0.19) 0.97 0.20 (0.02 to 0.38) 0.03 0.33 (0.16 to 0.51) < 0.01

0.8 0.4 0.4

0.14 (–0.05 to 0.33) 0.2 0.19 (–0.06 to 0.43) 0.1 0.25 (0.07 to 0.44) < 0.01 0.20 (0.02 to 0.38) 0.03

β (95% CI)

Sex

p-value

0.02 (–0.17 to 0.20) 0.01 (–0.17 to 0.19) 0.01 (–0.17 to 0.20)

0.9 1.0 0.9

–0.41 (–0.23 to 0.15) 0.04 (–0.15 to 0.23) 0.10 (–0.08 to 0.28) 0.08 (–0.09 to 0.26)

0.7 0.7 0.3 0.4

Calculated coefficients are the standardized coefficients (β) with corresponding p-value and 95% confidence interval. In this model, the independent variables were responsible for 19% of the variance in T2 relaxation times (R2 = 0.19) and no multicollinearity was detected.

Table 3. Subgroups of patients with unicompartmental cartilage damage Medial Lateral Mean T2 (SD) Mean T2 (SD)

Patients with

Medial cartilage damage (n = 35) Femur weight-bearing Femur posterior Tibia weight-bearing Lateral cartilage damage (n = 21) Femur weight-bearing Femur posterior Tibia weight-bearing

42 (9) 37 (6 40 (4)

39 (4) 36 (4) 40 (4)

0.05 0.01 0.5

41 (5) 37 (2) 39 (3)

39 (4 37 (4) 42 (5)

0.2 1.0 0.02

Mean T2 relaxation time (ms)

Mean T2 relaxation time (ms)

50

50

45

45

40

40

35

35

40

60

80

Age

30

15

20

25

30

All cases Non-traumatic onset Traumatic onset

40

p-value

patients with a non-traumatic onset of knee pain, T2 relaxation times appeared to be stable between the time points (Figure 3). These trends were seen for both the total tibiofemoral cartilage and the specific ROIs.

20

Mean T2 relaxation time (ms) 50

45

T2 values in milliseconds. Tested with paired sample t-test.

SD: standard deviation.

30

Mean T2 relaxation time (ms) 50

35

40

30 40 20 35

30

10

<1

1–6

>6

Months of disease

0

All cases Non-traumatic onset Traumatic onset

<1

Figure 3. Total tibiofemoral cartilage T2 values with 95% confidence interval for duration of disease for all cases and divided in non-traumatic and traumatic onset groups classified as acute (n = 18 [7 and 11]), subacute (n = 22 [13 and 9]), and chronic (n = 69 [42 and 27]).

Effect of duration on total cartilage T2 values for all cases was β = 0.31 (p = 0.4), for non-traumatic onset β = 0.06 (p = 0.6), and for traumatic onset β = –0.30 (p = 0.04) calculated by multiple linear regression analyses with sex, age, and BMI as covariates.

45

BMI

Figure 2. Scatter plots of age and mean T2 (left graph) and BMI and mean T2 (right graph) with corresponding trend lines (age: R2 = 0.15, and BMI: R2 = 0.068). Each circle represents the total tibiofemoral cartilage T2 value of 1 patient.

Clinical symptoms Mean KOOS values and standard deviations per subscale are displayed in Table 1. Univariate analyses showed a statistically significant association between clinical symptoms and total tibiofemoral T2 relaxation times for 2 of the 5 KOOS subscales (Pain: p = 0.02; Activities of daily living: p = 0.02). A lower score, i.e., more complaints, on the KOOS questionnaire was associated with elevated T2 relaxation times. When correcting for age, BMI, and sex, none of the associations remained significant. The item-specific analysis of the KOOS questionnaire revealed that, after adjusting for age, sex, and BMI, only “difficulties with descending stairs” was statistically significantly associated with elevated total tibiofemoral T2 relaxation times. Multivariate ROI-specific analysis did not show statistically significant associations with the different KOOS subscales either.

1–6

>6

Months of d

Acta Orthopaedica 2021; 92 (3): 335–340

Discussion

time (ms)

et

1–6

>6

Months of disease

In this study, we assessed the association of T2 relaxation times of the tibiofemoral knee cartilage with patient and disease characteristics and clinical symptoms in an unselected clinical population of 109 patients. A positive statistically significant association was observed between T2 relaxation times and age and BMI, while sex did not have an effect on T2 relaxation times.

Age seemed to have an overall effect on T2 values as increasing T2 values with increasing age were seen in most ROIs. Increasing T2 relaxation times with aging and higher BMI have previously been described in patients over 45 years old (Joseph et al. 2015). Furthermore, Mosher et al. found increasing T2 relaxation times in asymptomatic woman older than 45 compared with below 45 years (Mosher et al. 2004a). Our data shows these associations are seen in the whole adult range of age. BMI showed a trend towards increasing T2 values with increasing BMI, but a significant association was seen only in the lateral tibial weight-bearing cartilage. This is in contradiction to the findings of a recent paper that found an association between obesity and the risk of developing medial tibiofemoral OA (Wei et al. 2019). The range of T2 values in our study was between 35 and 50 ms, as can be seen in the scattorplots, which is in line with previously reported values (Oei et al. 2014). The increase in T2 relaxation time per unit of age or BMI was small, but this is what can be expected considering a difference of only 15 ms between the highest and lowest values. As most studies using T2 mapping focus on more advanced disease in selected patient groups, it is not surprising that larger differences in T2 values between damaged and healthy cartilage are found. We found no effect of sex on T2 relaxation times for both the total population and the age-dependent subgroups. A previously performed study looking at the influence of sex on T2 relaxation times also did not find such effect (Mosher et al.

2004b), but that study was based on a small and young population aged between 22 and 29 years. Other previous research showed only a weak association between T2 relaxation times and sex in the OAI population (age 45–65) without signs of radiographic OA (Joseph et al. 2015).

Differences in T2 values between medial and lateral compartments were found in patients with unicompartmental abnormalities. Previous studies with strict inclusion criteria already showed increasing T2 relaxation time in the medial knee compartment in patients with meniscal tears and in the lateral knee compartment in patients with anterior cruciate ligament injuries (Friedrich et al. 2009, Potter et al. 2012).

Our study confirms this effect in a heterogeneous population.

We found statistically significantly higher T2 values in the medial femoral ROIs in patients with abnormalities expected to cause predominantly medial cartilage changes. In patients with suspected isolated lateral cartilage changes, a statistically significant difference was found only in the tibial ROIs. Just like the correlations of age and BMI with T2 relaxation times,

339

it is remarkable that higher medial femoral cartilage T2 values were associated with increasing age and medial abnormalities and higher lateral tibial values were associated with increasing BMI and lateral abnormalities. It would be interesting to assess the influence of mechanical leg axis on these findings, but as long leg radiographs were not available, it was not possible to answer this question.

Duration of complaints could potentially lead to transient variation of T2 relaxation times within patients as evidence is provided that the integrity of the cartilage collagen network is compromised soon after joint injury (Lohmander et al. 2003).

Our study revealed higher T2 relaxation times in patients who had an interval of less than 1 month between trauma and MRI compared with patients with an interval longer than 6 months.

However, since we did not perform follow-up measurements of the same patient, no conclusions regarding the trend over time of T2 relaxation time following trauma can be made based on our data. Nonetheless, it is worth noting that in the case of non-traumatic knee pain the duration of complaints did not cause variation in T2 relaxation times.

As far as we know, no imaging modality has shown a good correlation with clinical symptoms of knee injury and osteoarthritis in an unselected routine clinical population. In our study, significant associations were found between T2 values and two domains of the KOOS questionnaire in the univariate analysis. However, this finding was not sustained when corrected for age, sex, and BMI, with age being the predominant covariate. When looking at the item-specific analysis, we found only “any difficulty with descending stairs” to be correlated with T2 relaxation times after correction. Although a large set of symptoms was tested, and based on repeated testing coincidental findings are possible, previous studies also reported difficulties with climbing stairs to be a sensitive and prodromal symptom in osteoarthritis (Case et al. 2015, Landsmeer et al. 2019). The wide range in age and the known increase in knee complaints with age might be responsible for the absence of further associations between T2 relaxation times and clinical symptoms in our study (Paradowski et al. 2006).

Our study has several limitations. By using a clinical orthopedic population, we included patients with a wide range in age, BMI, diagnoses, and clinical symptoms. The combination of this heterogeneity and limited sample size could explain the absence of clear associations between T2 values and disease characteristics and clinical symptoms in our study. A second limitation is that we had a valid KOOS questionnaire available for only half of the patients. It was common practice at the Orthopedic Outpatient Department to ask patients to fill in the questionnaire. Unfortunately, this was not strictly controlled.

We are aware that previous studies have shown T2 differences between superficial and deep cartilage layers (Mosher et al. 2004b, Bengtsson Moström et al. 2015). However, as our T2 mapping sequence is a 3D sequence with coverage of the whole knee, we considered the spatial resolution not good enough to perform these subregional analyses. Finally, we

340

realize the magic angle effect could influence T2 values. However, as all patients were positioned in a standardized fashion, the effect would be similar for all patients. Together with the type of analyses we performed, we do not think the magic angle effect substantially influenced our results.

To date, the application of T2 mapping is primarily in clinical research with patient groups based on well-defined inclusion criteria. In contrast to the success of T2 mapping in research trials like the OAI, the poor associations of T2 mapping with patient and disease characteristics observed in our study illustrate the difficulties of implementing such a quantitative MR technique in a routine clinical population. In conclusion, our results emphasize the importance of acknowledging patient and disease characteristics when performing T2 mapping in a clinical population.

JV, SJvL, GEG, JLD, SMABZ, MR, and EHGO contributed to the design of the work. JV, SJvL, JLD, GEG, and EHGO contributed to the acquisition of the data. JV, SJvL, GEG, JLD, SMABZ, MR, and EHGO contributed to the analysis and interpretation of the data. All authors contributed to drafting the work or revising the content critically and all authors have approved the final version. JV and EHGO have 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.

Acta thanks Carl Johan Tiderius and other anonymous reviewers for help with peer review of this study.

Bengtsson Moström E, Lammentausta E, Finnbogason T, Weidenhielm L, Janarv P M, Tiderius C J. Pre- and postcontrast T1 and T2 mapping of patellar cartilage in young adults with recurrent patellar dislocation. Magn Reson Med 2015; 74(5): 1363-9. doi: 10.1002/mrm.25511.

Bron E E, van Tiel J, Smit H, Poot D H, Niessen W J, Krestin G P, Weinans H, Oei E H, Kotek G, Klein S. Image registration improves human knee cartilage T1 mapping with delayed gadolinium-enhanced MRI of cartilage (dGEMRIC).

Eur Radiol 2013; 23(1): 246-52. doi: 10.1007/s00330-012-2590-3.

Case R, Thomas E, Clarke E, Peat G. Prodromal symptoms in knee osteoarthritis: a nested case-control study using data from the Osteoarthritis Initiative. Osteoarthritis Cartilage 2015; 23(7): 1083-9. doi: 10.1016/j. joca.2014.12.026.

Chan W P, Lang P, Stevens M P, Sack K, Majumdar S, Stoller D W, Basch C, Genant H K. Osteoarthritis of the knee: comparison of radiography, CT, and MR imaging to assess extent and severity. AJR Am J Roentgenol 1991; 157(4): 799-806. doi: 10.2214/ajr.157.4.1892040.

Chen W, Han E T. 3D quantitative imaging of T1rho and T2 (Abstract). Proc Annu Meet ISMRM 2011; 19: 231.

Crema M D, Hunter DJ, Burstein D, Roemer F W, Li L, Krishnan N, Marra M D, Hellio Le-Graverand M P, Guermazi A. Delayed gadolinium-enhanced magnetic resonance imaging of medial tibiofemoral cartilage and its relationship with meniscal pathology: a longitudinal study using 3.0T magnetic resonance imaging. Arthritis Rheumatol 2014; 66(6): 1517-24. doi: 10.1002/art.38518.

Dunn T C, Lu Y, Jin H, Ries M D, Majumdar S. T2 relaxation time of cartilage at MR imaging: comparison with severity of knee osteoarthritis. Radiology 2004; 232(2): 592-8. doi: 10.1148/radiol.2322030976.

Friedrich K M, Shepard T, de Oliveira V S, Wang L, Babb J S, Schweitzer M, Regatte R. T2 measurements of cartilage in osteoarthritis patients with meniscal tears. AJR Am J Roentgenol 2009; 193(5): W411-15. doi: 10.2214/AJR.08.2256.

Acta Orthopaedica 2021; 92 (3): 335–340

Guermazi A, Roemer F W, Burstein D, Hayashi D. Why radiography should no longer be considered a surrogate outcome measure for longitudinal assessment of cartilage in knee osteoarthritis. Arthritis Res Ther 2011; 13(6): 247. doi: 10.1186/ar3488.

Hunter D J, Bierma-Zeinstra S. Osteoarthritis. Lancet 2019; 393(10182): 1745-59. doi: 10.1016/S0140-6736(19)30417-9.

Joseph G B, McCulloch C E, Nevitt M C, Heilmeier U, Nardo L, Lynch J A, Liu F, Baum T, Link T M. A reference database of cartilage 3 T MRI T2 values in knees without diagnostic evidence of cartilage degeneration: data from the osteoarthritis initiative. Osteoarthritis Cartilage 2015; 23(6): 897905. doi: 10.1016/j.joca.2015.02.006.

Landsmeer M L A, Runhaar J, van Middelkoop M, Oei E H G, Schiphof D, Bindels P J E, Bierma-Zeinstra S M A. Predicting knee pain and knee osteoarthritis among overweight women. J Am Board Fam Med 2019; 32(4): 575-84. doi: 10.3122/jabfm.2019.04.180302.

Lohmander L S, Atley L M, Pietka T A, Eyre D R. The release of crosslinked peptides from type II collagen into human synovial fluid is increased soon after joint injury and in osteoarthritis. Arthritis Rheum 2003; 48(11): 31309. doi: 10.1002/art.11326.

Matzat S J, van Tiel J, Gold G E, Oei E H. Quantitative MRI techniques of cartilage composition. Quant Imaging Med Surg 2013; 3(3): 162-74. doi: 10.3978/j.issn.2223-4292.2013.06.04.

Matzat S J, McWalter E J, Kogan F, Chen W, Gold G E. T2 Relaxation time quantitation differs between pulse sequences in articular cartilage. J Magn Reson Imaging 2015; 42(1): 105-13. doi: 10.1002/jmri.24757.

McCauley T R, Recht M P, Disler D G. Clinical imaging of articular cartilage in the knee. Semin Musculoskelet Radiol 2001; 5(4): 293-304. doi: 10.1055/s-2001-19040.

Mosher T J, Collins C M, Smith H E, Moser L E, Sivarajah R T, Dardzinski B J, Smith M B. Effect of gender on in vivo cartilage magnetic resonance imaging T2 mapping. J Magn Reson Imaging 2004a; 19(3): 323-8. doi: 10.1002/jmri.20013.

Mosher T J, Liu Y, Yang Q X, Yao J, Smith R, Dardzinski B J, Smith M B. Age dependency of cartilage magnetic resonance imaging T2 relaxation times in asymptomatic women. Arthritis Rheum 2004b; 50(9): 2820-8. doi: 10.1002/art.20473.

Oei E H, van Tiel J, Robinson W H, Gold G E. Quantitative radiologic imaging techniques for articular cartilage composition: toward early diagnosis and development of disease-modifying therapeutics for osteoarthritis. Arthritis Care Res (Hoboken) 2014; 66(8): 1129-41. doi: 10.1002/acr.22316.

Paradowski P T, Bergman S, Sundén-Lundius A, Lohmander L S, Roos E M. Knee complaints vary with age and gender in the adult population: population-based reference data for the Knee injury and Osteoarthritis Outcome Score (KOOS). BMC Musculoskelet Disord 2006; 7:38. doi: 10.1186/1471-2474-7-38.

Potter H G, Jain S K, Ma Y, Black B R, Fung S, Lyman S. Cartilage injury after acute, isolated anterior cruciate ligament tear: immediate and longitudinal effect with clinical/MRI follow-up. Am J Sports Med 2012; 40(2): 276-85. doi: 10.1177/0363546511423380.

Roos E M, Roos H P, Lohmander L S, Ekdahl C, Beynnon B D. Knee Injury and Osteoarthritis Outcome Score (KOOS): development of a self-administered outcome measure. J Orthop Sports Phys Ther 1998; 28(2): 88-96. doi: 10.2519/jospt.1998.28.2.88.

Setton L A, Elliott D M, Mow V C. Altered mechanics of cartilage with osteoarthritis: human osteoarthritis and an experimental model of joint degeneration. Osteoarthritis Cartilage 1999; 7(1): 2-14. doi: 10.1053/ joca.1998.0170.

Su F, Hilton J F, Nardo L, Wu S, Liang F, Link T M, Ma C B, Li X. Cartilage morphology and T1ρ and T2 quantification in ACL-reconstructed knees: a 2-year follow-up. Osteoarthritis Cartilage 2013; 21(8): 1058-67. doi: 10.1016/j.joca.2013.05.010.

Wei J, Gross D, Lane N E, Lu N, Wang M, Zeng C, Yang T, Lei G, Choi H K, Zhang Y. Risk factor heterogeneity for medial and lateral compartment knee osteoarthritis: analysis of two prospective cohorts. Osteoarthritis Cartilage 2019; 27(4): 603-10. doi: 10.1016/j.joca.2018.12.013.

Acta Orthopaedica 2021; 92 (3): 341–346

341

Development of acetabular anteversion in children with normal hips and those with developmental dysplasia of the hip: a cross-sectional study using magnetic resonance imaging Wei LU, Lianyong LI, Lijun ZHANG, Qiwei LI, and Enbo WANG

Department of Pediatric Orthopedics, Shengjing Hospital of China Medical University, Shenyang, Liaoning Province, China Correspondence: LL: loyo_ldy@163.com Submitted 2020-06-13. Accepted 2020-11-16.

Background and purpose — Acetabular anteversion (AA) is related to hip function. Most previous studies were based on radiographic investigations that determine osseous acetabular anteversion (OAA). But children’s acetabulum is mostly composed of cartilage; the cartilaginous acetabular anteversion (CAA) represents the real anteversion of the acetabulum. We measured OAA and CAA in children of various ages using MRI, and compared the developmental patterns between children with normal hips and those with developmental dysplasia of the hip (DDH).

Patients and methods — The OAA and CAA were measured on MRI cross-sections of the hips in 293 children with normal hips (average age 8 years), and in 196 children with DDH (average age 34 months). Developmental patterns of OAA and CAA in children with normal hips were determined through age-based cross-sectional analysis. Differences in OAA and CAA between children with normal hips and those with DDH were compared.

Results — Normal OAA increased from mean 8.7° (SD 3.2) to 12° (3.0) during the first 2 years of life and remained unchanged until 9 years of age. From 9 to 16 years, the OAA showed a minimal increase of 2°–3°. The normal CAA increased rapidly from a mean of 12° (3.1) to 15° (2.7) within the first 2 years of life, and remained constant at 15° (SD 3.4) until 16 years of age. The age-matched average OAA in the normal and DDH cases was 11° (3.2) and 15° (3.0), respectively (p < 0.001). The age-matched average CAA in normal and DDH cases was 17° (4.2) and 23° (4.5), respectively (p < 0.001). Similarly, there was a significant difference in OAA and CAA between the uninvolved hips in unilateral DDH and normal cases (p < 0.001).

Interpretation — The CAA was fully formed at birth in normal children, and remained unchanged until adulthood, whereas the OAA increased with age. The OAA and CAA were both over-anteverted in DDH children. MRI evaluation is of importance in children during skeletal development when planning hip surgery.

Acetabular anteversion (AA) is the direction of the acetabular opening relative to the axial planes. This is among the important quantitative indicators to describe acetabular tilt and is related to hip joint function. Proper AA facilitates the inclusion of the femoral head. The related position between the anterior and posterior walls of the acetabulum determines the normal angle of AA. At the same time, AA is also an important reference index for various pelvic rotational osteotomy or hip replacement procedures. Traditionally, AA is measured on 3D computed tomography (CT) (Li et al. 2009), which represents the osseous acetabular anteversion (OAA). However, children’s acetabulum is mostly composed of cartilage; the cartilaginous acetabular anteversion (CAA) determined by the cartilaginous anterior and posterior walls is the true AA, but cannot be measured by CT scan. The normal value of CAA in childhood is rarely reported.

Developmental dysplasia of the hip (DDH) includes delayed cartilage ossification or acetabulum margin dysplasia. Many previous studies have reported that over-anteversion of the acetabulum is one of the morphological abnormalities in children with DDH (Albinana et al. 2005, Jia et al. 2011). However, most previous studies were based on radiographic investigation, which therefore determined the OAA. The changes in CAA in DDH have never been established. Recognizing the true AA is an important aspect of the morphological evaluation of DDH and an important reference index for the surgical correction of acetabulum direction; thus, the evaluation of CAA is of clinical significance, especially in children during skeletal development.

Owing to the advantage of noninvasively distinguishing the osseous rim from the cartilaginous acetabulum in the immature hip, magnetic resonance imaging (MRI) has been widely used to evaluate the pathological abnormity of DDH (Mootha et al. 2010, Krasny et al. 1991, Falliner et al. 2002, Douira-Khomsi et al. 2010, Goronzy et al. 2019). In this retrospective cross-sectional MRI study we measured OAA and CAA in children of various ages using MRI, analyzed their

© 2021 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 10.1080/17453674.2020.1866928

342

Figure 1. In the coronal position, the yellow line is the line passing through the center of the acetabulum.

The corresponding axial section through the line is shown in Figure 2.

Acta Orthopaedica 2021; 92 (3): 341–346

Figure 2. Measurement of osseous acetabular anteversion (OAA) and cartilaginous acetabular anteversion (CAA). In the T1-weighted axial magnetic resonance image (MRI), line A is tangential to the most posterior points of both ischia. Line B is drawn perpendicular to line A. a. Line C is tangential to the outermost osseous anterior and posterior walls of the acetabulum and the angle between lines B and C is defined as the OAA. b. Line D is tangential to the outermost cartilaginous anterior and posterior walls of the acetabulum and the angle between lines B and D is defined as the CAA.

developmental patterns, and compared the difference in OAA and CAA between children with normal hips and those with DDH.

Patients and methods Samples and measurements We performed a retrospective cross-sectional MRI study of children under 16 years of age who had underwent MRI (including the hip joint) from January 2008 to January 2018 in the authors’ institution. By using our Picture Archiving and Communication System (PACS), 2 pediatric orthopedics specialists (LYL and LJZ) read the pelvic MRI data. Patients with non-standard MRI examination or with other diseases of the hip such as DDH, femoral head necrosis, septic arthritis, or other syndromic disease of the hip were excluded. 293 consecutive children (586 hips) who underwent MRI examination because of non-neuromuscular and non-skeletal disease were included in the study. There were 147 boys and 146 girls with a mean age of 8.0 years (1 month to 16 years). All of the 586 hips were clinically normal, but with a diagnosis of lower limb pain, benign abdominal tumors, lymphangioma, or hemangioma in the cutaneous or subcutaneous tissues around the pelvis as the reason for the MRI examination.

Patients with DDH who were treated in Department of Pediatric Orthopedics of Shengjing Hospital from January 2008 to January 2018 were also included in the study. All patients were aged under 7 and had not been treated before. Patients with neuromuscular and syndromic DDH were excluded. 196 consecutive patients were reviewed in the study, including 159 girls and 37 boys (241 affected hips). Among these 196 children with DDH, 45 had bilateral involvement and 151 had unilateral involvement, including 95 and 56 patients with left and right hip involvement, respectively. The average age at reduction was 34 months (6–84 months). Patients with poor quality MRI scans and neuromuscular and pathological DDH were excluded. According to the International Hip Dysplasia

Institute (IHDI) classification (Narayanan et al. 2015), 30 hips were classified as grade I, 33 hips as grade II, 65 hips as grade III, and 113 hips as grade IV. All of the hips in children with normal hips and those with DDH were examined with standardized MR scans. The MR scans were performed using the 3.0 T Philips Medical System (Philips Achieva, Best, The Netherlands), including the pelvis and proximal femur with axial, sagittal, and coronal plane sequences. The patients were placed supine, with legs in a neutral position, and with a body array coil placed anterior and posterior to the pelvis. The MR scan was carried out under sedation before the MR examination in children aged < 4 years. The sequences included T1- and T2-weighted images obtained in the axial and coronal planes using 3-mm slice thickness and 0-mm interslice gap. All sequences used a TR of 4500 ms and TE of 120 ms in T2-weighted fast spin-echo; TR of 450 mms and TE of 12 ms in T1-weighted spin-echo; and matrix of 512×512.

Using the MR scans, the OAA and CAA were measured by using the PACS. OAA and CAA were measured on the axial section of the hip located through the center of the acetabulum on the coronal section (Figure 1). On the axial section, line A was tangential to the most posterior points of both ischia. Line B was drawn perpendicular to line A. Line C was tangential to the outermost osseous anterior and posterior walls of the acetabulum. Line D was tangential to the outmost cartilaginous anterior and posterior walls of the acetabulum. The angle between lines B and C was defined as OAA, and that between lines B and D was defined as CAA (Figure 2).

The OAA and CAA were measured (30 hips) 3 times independently, identified by initials only, to determine interobserver agreement; to evaluate intraobserver variation, the measurements were repeated 4 weeks later by initials.

Statistics Statistical analysis was performed using SPSS version 20.0 software (IBM Corp, Armonk, NY, USA). To evaluate the relationship between OAA and CAA, we used paired t-tests

Acta Orthopaedica 2021; 92 (3): 341–346

343

Table 1. Assessment of intra- and interobserver agreements of osseous acetabular anteversion (OAA) and cartilaginous acetabular anteversion (CAA)

Acetabular anteversion (°) 30

25

OAA CCA Observers ICC (95% CI) ICC (95% CI) WL–WL WL–LYL WL–LJZ LYL–LJZ

0.997 (0.996–0.998) 0.882 (0.858–0.911) 0.990 (0.987–0.992) 0.898 (0.875–0.912)

20

0.976 (0.973–0.982) 0.886 (0.864–0.911) 0.873 (0.849–0.892) 0.901 (0.882–0.923)

15

10

ICC, intraclass correlation coefficient; CI, confidence interval.

P-value for each observer pair < 0.01

5

Normal cartilaginous Normal osseous Dysplastic cartilaginous Dysplastic osseous

0

Table 2. Values of osseous (OAA) and cartilaginous acetabular anteversion (CAA) corresponding to age in normal children

Age 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Hips n 58 48 40 16 36 44 28 48 36 38 32 40 20 30 34 38

OAA (°) Mean (SD) [95% Cl]

CAA (°) Mean (SD) [95% Cl]

8.7 (3.2) [7.9–9.5] 11.8 (3.0) [11.0–12.7] 12.0 (1.7) [11.4–12.5] 12.0 (3.1) [10.4–13.7] 11.9 (2.9) [11.0–12.9] 12.1 (2.6) [11.3–12.8] 12.7 (3.4) [11.4–14.0] 12.0 (3.1) [11.1–12.9] 12.2 (3.0) [11.2–13.2] 12.7 (3.4) [11.6–13.8] 12.9 (3.4) [11.7–14.1] 13.3 (4.1) [11.9–14.6] 14.4 (6.3) [11.4–17.3] 14.7 (3.2) [13.5–15.9] 14.5 (4.2) [13.0–15.9] 14.4 (3.6) [13.2–15.5]

12.3 (3.1) [11.5–13.1] 15.1 (2.7) [14.4–16.0] 15.5 (2.4) [14.7–16.3] 15.1 (2.6) [13.7–16.5] 15.0 (2.7) [14.1–16.0] 15.2 (2.6) [14.4–16.0] 15.1 (3.2) [13.8–16.3] 15.2 (2.8) [14.4–16.0] 15.5 (3.0) [14.5–16.5] 15.2 (3.1) [14.1–16.2] 14.7 (3.3) [13.5–15.9] 15.4 (3.6) [14.3–16.6] 16.7 (6.2) [13.8–19.6] 16.2 (3.2) [15.0–17.4] 16.6 (4.3) [15.1–18.1] 15.3 (3.7) [14.1–16.5]

CI, confidence interval; SD, standard deviation.

P-value for OAA versus CAA in each age group < 0.01

and Pearson correlation analysis. Student’s t-test was used to compare the difference in OAA and CAA between the normal and DDH cases. Differences in OAA and CAA among different age and IHDI classification groups were assessed with 1-way analysis of variance. A p-value < 0.05 was considered statistically significant.

Interobserver agreements between 3 sets of measurements of three observers and intraobserver agreement between 2 sets of measurements of an observer were analyzed using Pearson’s correlation coefficient and the intraclass correlation coefficient (ICC). According to the usual recommendations, the concordance was examined as follows: < 0.2 (bad), 0.2–0.4 (low), 0.4–0.6 (moderate), 0.6–0.8 (good), and > 0.8 (excellent).

Ethics, funding, and potential conflicts of interest This study was approved by the Institutional Ethical Committee. The parents or guardians of all study participants provided informed consent. The authors benefited from a funding

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Age

Figure 3. Mean trend lines with 95% CI of all measurements for osseous and cartilaginous acetabular anteversion versus age of children with normal hips and those with developmental dysplasia of the hip.

program, the National Natural Science Foundation of China (81772296), and have no conflicts of interest to declare.

Results The intra- and interobserver agreements with 95% confidence intervals (CIs) were assessed by measuring the OAA and CAA repeatedly. For all measurements, the ICCs were > 0.8, indicating excellent agreement (Table 1).

The mean OAAs and CAAs according to age in the 586 normal hips are summarized in Table 2. The normal OAA had a mean of 8.7° at 1 year of age, which increased progressively during the first 2 years of life, reaching a mean of 12° at 2 years of age. Then, it remained constant until the age of 9 years. At 9–16 years of age, it increased slightly by approximately 2° (Figure 3). Notably, the normal CAA had a different pattern of development from the OAA. Within the first 2 years of life, the CAA increased rapidly from a mean of 12° to 15°, and then stayed at a constant mean level of 15° until the age of 16 years (Figure 3). In addition, acetabular retroversion was found in none of the hips.

There was a positive correlation between OAA and CAA in the normal hips (r = 0.92, p < 0.001). For the OAA in the normal hips, there was a statistically significant difference between the girls and boys (Figure 4). From 0–9 years of age, the development trend of OAA between the sexes was basically the same. After the age of 9 years, OAA showed a more obvious increase in girls compared with boys until 16 years of age. The relationship of CAA between the sexes was similar to that of OAA (Figure 4).

The average OAA in normal and DDH cases was 11° and 15°, respectively (p < 0.001). The average CAA in normal and DDH cases was 17° and 23°, respectively (p < 0.001).

Both OAA and CAA in children with DDH were larger than

344

Acta Orthopaedica 2021; 92 (3): 341–346

Table 3. Values of osseous (OAA) and cartilaginous acetabular anteversion (CAA) corresponding to age in children with developmental dysplasia of the hip

Acetabular anteversion (°) 30

25

Female cartilaginous Male cartilaginous Female osseous Male osseous

20

15

10

5

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Age

Hips n

OAA (°) Mean (SD) [95% Cl]

CAA (°) Mean (SD) [95% Cl]

1 2 3 4 5 6 7

13 121 50 19 14 14 10

13.7 (4.2) [11.2–16.3] 16.6 (3.9) [15.9–17.3] 17.8 (4.1) [16.6–19.0] 19.8 (4.4) [17.7–22.0] 18.9 (4.3) [16.4–21.4] 17.0 (4.7) [14.3–19.7] 17.9 (3.0) [15.8–20.0]

16.8 (5.0) [13.8–19.8] 22.1 (4.3) [21.4–22.9] 23.7 (3.7) [22.7–24.8] 24.9 (5.1) [22.4–27.3] 23.5 (4.2) [21.0–25.9] 23.8 (4.6) [21.1–26.4] 24.1 (2.5) [22.3–26.0]

CI, confidence interval; SD, standard deviation.

P-value for OAA versus CAA in each age group < 0.01

Age

Figure 4. Osseous acetabular anteversion and cartilaginous acetabular anteversion between the sexes.

those in normal children, when the age of the individuals was matched (p < 0.001, Table 3, Figure 3). In DDH cases, the OAA in the uninvolved hip in unilateral DDH cases showed a mean of 14° and that of the age-matched children with normal hips showed a mean of 11° (p < 0.001). The CAA in the uninvolved hip in unilateral DDH cases showed a mean of 18°, and that of the age-matched children with normal hips showed a mean of 15° (p < 0.001), indicating that both OAA and CAA in the uninvolved hips were also over-anteverted.

For the DDH cases, OAA was statistically significantly less than CAA, regardless of IHDI types (Table 4). OAA and CAA were similar among the IHDI types, except for the minimal difference (approximately 2°–3°) between type I and IV (p = 0.02), type II and IV (p < 0.001) for OAA; and between type III and IV (p = 0.01), type II and IV (p < 0.001) for CAA.

Discussion Although MRI is expensive and sometimes requires sedation, it is still widely used in the examination of immature hips because it does not expose children to radiation and can provide multiplanar, high-quality images that clearly show the boundaries between osseous and cartilaginous structures.

A large number of anatomical and MRI studies have shown that MRI can accurately show various morphological changes of the developing hip (Goronzy et al. 2019), but few studies have investigated whether MRI is reliable in measuring the OAA and CAA during development. We found excellent consistency between observers, indicating that the repeatability of MRI measurement is high.

The acetabulum is a ball and socket joint consisting of the anterior pubis, superior ilium, and posterior ischium. In a normal newborn, the acetabulum is a cartilage complex composed of acetabular cartilage and Y-shaped cartilage, and the development of the acetabulum is mainly characterized by

Table 4. Mean (SD) with 95% confidence intervals (CI) of osseous (OAA) and cartilaginous acetabular anteversion (CAA) in children with developmental dysplasia of the hip according to the IHDI classification IHDI types

Hips n

OAA (°) Mean (SD) [95% Cl]

CAA (°) Mean (SD) [95% Cl]

I II III IV

30 33 65 113

16.0 (3.5) [14.7–17.4] 15.3 (3.5) [14.1–16.6] 17.0 (4.0) [16.0–17.9] 18.1 (4.4) [17.2–18.9]

22.4 (4.2) [20.8–23.9] 20.8 (4.5) [19.2–22.4] 22.0 (4.5) [20.8–23.1] 23.7 (4.4) [22.9–24.5]

IHDI, International Hip Dysplasia Institute.

P-value for OAA versus CAA in each IHDI group < 0.01 There were no significant differences between any 2 means for OAA and CAA, except for the comparison between type I and IV (p = 0.02), type II and IV (p < 0.001) for OAA; and between type III and IV (p = 0.01), type II and IV (p < 0.001) for CAA.

endochondral ossification after birth. Li et al. (2016) measured the OAA and CAA of 180 children with normal hips in different age groups from 6 months to 16 years by using MRI, and found that OAA and CAA were both relatively constant among the different age groups. They speculated that this may be due to the compressive stress of the femoral head on the acetabulum being uniform in the normal hip joint, which tends to balance the growth of the ilium, ischia, and pubis, thereby keeping the acetabulum in a stable axial opening direction. Our study results differed from those of Li et al., which may be due to the difference in the age grouping method used or the small sample size of age subgroups in the study by Li et al.

In our cross-section investigation, there was a rapid increase in OAA and CAA in the first 2 years after birth, and stability is maintained from 2 to 9 years of age, showing a similar growth pattern in development of the osseous and cartilaginous acetabular index in children with normal hips as observed by Li et al. (2012). Before adolescence OAA is constant, being approximately 3° less than CAA until 9 years of age, suggesting that ossification in the anterior and posterior walls of the

Acta Orthopaedica 2021; 92 (3): 341–346

acetabulum has an identical velocity to the development of the cartilaginous acetabular wall. The second developmental stage of OAA is between ages 9 and 16 years, with an increase of 2°–3°. This change is mainly attributed to the fact that ossification of the acetabular posterior wall is faster than that of the anterior wall. Albers et al. (2017) suggested that the OAA increase is related to the appearance of the secondary ossification center in the anterior and posterior walls of the acetabulum that begins at 9 years of age.

Unlike OAA, CAA is almost constant at the level of 15° (SD 3.4) throughout the childhood and adolescence period.

This indicates that the cartilaginous acetabular anterior and posterior walls are fully formed at birth. This finding is supported by Johnson et al. (1989), who investigated the anatomical structure of the hip in infants by MRI. However, the postnatal development of cartilaginous acetabular anterior and posterior margins has not been reported previously. Our cross-sectional study was age categorized, with detailed mean and standard deviation from the age of 1 to 16 years, and provides the normal standards for the development of cartilaginous acetabular anterior and posterior walls in children. After infancy, a hip with a CAA exceeding 21° (> 2 SDs from the mean) should be considered as cartilaginous over-anteversion.

According to our findings, although OAA and CAA had a significant positive correlation, the anteverted level was different between them. In children during skeletal development, the cartilaginous acetabular edges determine the true AA. Therefore, CAA cannot be replaced by OAA. In addition, we found that before the age of 9 years, OAA development in boys and girls is basically the same, whereas at the age of 9–16 years, OAA in girls is larger than that of boys. The development of CAA between sexes showed a similar difference to that of OAA. Previous studies also support this conclusion (Rubalcava et al. 2012, Jiang et al. 2015b).

However, the reasons for this phenomenon are not clear; the authors speculate that it may be related to the differences in pelvic morphology between men and women.

Nevertheless, there was also a statistically significant correlation between OAA and CAA in the DDH cases, indicating that both OAA and CAA can reflect the degree of excessive anteversion of the acetabulum. However, due to the large individual differences in acetabulum pathology, especially the difference in cartilage acetabulum edge, it was unreasonable to evaluate CAA using OAA directly. We believe that cartilaginous anteversion of the acetabulum is more representative of the final state of acetabulum development than osseous anteversion of the acetabulum (Duffy et al. 2002, Zamzam et al. 2008). It has been reported that the rotational degree of the acetabulum in Salter pelvic osteotomy can be determined according to the degree of AA shown in 3D CT; thus, it may be cursory (Jiang et al. 2015a). We found that the true degree of AA is determined by the degree of CAA, which is often larger than the degree of OAA and should be taken into account in the intraoperative estimation of pelvic rotation. Excessive ace-

345

tabular anteversion or retroversion can lead to excessive containment, and then lead to impingement between the femoral head and acetabulum, which is an important cause of osteoarthritis (Kim et al. 1999). Owing to the limitation in using intraoperative MRI, it is difficult to evaluate the anteversion of the acetabulum. We suggest that the difference between OAA and CAA should be evaluated preoperatively by MRI, and the true AA (CAA) can be calculated by measuring the OAA using radiological methods. Additionally, one can directly observe the position of the anterior and posterior acetabular edges by intraoperative arthrography according to Forlin’s method (Forlin et al. 1992) to accurately evaluate the anteversion degree of the cartilaginous acetabular edge.

Mootha et al. (2010) used MRI to assess acetabular and femoral anteversion in 45 children with DDH in children aged 12 to 48 months and found that acetabular anteversion was increased in the dislocated group compared with the normal group. In this study, both OAA and CAA demonstrated an excessive anteversion in DDH cases, which is similar to results reported in previous studies (Jacobsen et al. 2006, Kobayashi et al. 2010). Moreover, no difference or only a little variation (approximately 2°–3°) for OAA and CAA was found among the different types of IHDI, indicating that the developmental defect of the anterior margin was a common physiological phenomenon after dislocation, which might be related to the loss of mutual stimulation between the femoral head and the acetabulum. Understanding these physiological characteristics is helpful to guide evaluation of the pathological morphology and surgical correction of DDH. For uninvolved hips in unilateral DDH cases, both OAA and CAA were increased compared with those of the normal hips when both groups were age matched. It is suggested that both the osseous and cartilaginous acetabulum in the uninvolved hip also have mild over-anteversion; thus, one should be discreet when using the unaffected hip of unilateral DDH as a control. This phenomenon has a similar performance in the observation of other DDH parameters (Li et al. 2012). This is also supported by the study of Song et al. (2008), who suggested that 40% of uninvolved hips in unilateral DDH patients have mild dysplasia.

Limitations There are limitations in our study. Although patients with a non-hip disorder were included, they were not healthy children and had diseases pertaining to lower limb pain, benign abdominal tumors, lymphangioma, or hemangioma in the cutaneous or subcutaneous tissues around the pelvis. In addition, with the increasing degree of dislocation in DDH patients, it could be difficult to locate the center of the acetabulum accurately, which may influence the results of measurements.

Conclusion The growth patterns of the osseous and cartilaginous parts in the acetabulum are different as measured by MRI. The CAA

346

is fully formed after birth and does not change substantially with age until adulthood. However, the OAA increases rapidly within the first 2 years postnatally, and then remains unchanged until 9 years of age; then, from 9 to 16 years of age, it increases slightly by approximately 2°. Both the osseous and cartilage acetabulum of children with DDH have excessive anteversion. The uninvolved hips of unilateral DDH cases are also over-anteverted compared with normal hips.

CAA represents the true AA. Preoperative evaluation of CAA in DDH patients by MRI is of importance for accurate correction of acetabular orientation.

WL: data collection and analysis, manuscript preparation. LJZ: data analysis, manuscript revision. QWL: data analysis, manuscript revision. EBW: data collection and analysis. LYL: project administration, manuscript revision, fund acquisition.

Acta thanks Federico Canavese and Klaus Dieter Parsch for help with peer review of this study.

Albers C E, Schwarz A, Hanke M S, Kienle K P, Werlen S, Siebenrock K A. Acetabular version increases after closure of the triradiate cartilage complex. Clin Orthop Relat Res 2017; 475: 983-94.

Albinana J, Dolan L A, Spratt K F, Morcuende J, Meyer M D, Weinstein S L. Acetabular dysplasia after treatment for developmental dysplasia of the hip: implications for secondary procedures. J Bone Joint Surg Br 2005; 86: 876-86.

Douira-Khomsi W, Smida M, Louati H, Hassine L B, Bouchoucha S, Saied W, Ladeb M F, Ghachem M B, Bellagha I. Magnetic resonance evaluation of acetabular residual dysplasia in developmental dysplasia of the hip: a preliminary study of 27 patients. J Pediatr Orthop 2010; 30: 37-43.

Duffy C M, Taylor F N, Coleman L, Graham H K, Nattrass G R. Magnetic resonance imaging evaluation of surgical management in developmental dysplasia of the hip in childhood. J Pediatr Orthop 2002; 22: 92-100.

Falliner A, Muhle C, Brossmann J. Acetabular inclination and anteversion in infants using 3D MR imaging. Acta Radiol 2002; 43: 221-4.

Forlin E, Choi I H, Guille J T, Bowen J R, Glutting J. Prognostic factors in congenital dislocation of the hip treated with closed reduction: the importance of arthrographic evaluation. J Bone Joint Surg Am 1992; 74: 1140-52.

Goronzy J, Blum S, Hartmann A, Plodeck V, Franken L, Günther K P, Thielemann F. Is MRI an adequate replacement for CT scans in the three-dimensional assessment of acetabular morphology? Acta Radiol 2019; 60: 726-34.

Jacobsen S, Rømer L, Søballe K. The other hip in unilateral hip dysplasia.

Clin Orthop Relat Res 2006; 446: 239-46.

Jia J Y, Li L Y, Zhang L J, Zhao Q, Wang E B, Li Q W. Can excessive lateral rotation of the ischium result in increased acetabular anteversion? A 3D-CT

Acta Orthopaedica 2021; 92 (3): 341–346

quantitative analysis of acetabular anteversion in children with unilateral developmental dysplasia of the hip. J Pediatr Orthop 2011; 31: 864-9.

Jiang J, Ren S, Liu M. Impact of Salter innominate osteotomy on acetabular morphology and direction in developmental dislocation of the hip by threedimensional computer tomography. Zhongguo Xiu Fu Chong Jian Wai Ke Za Zhi 2015a; 29: 1332-6.

Jiang N, Peng L, Al-Qwbani M, Xie G P, Yang Q M, Chai Y, Zhang Q, Yu B. Femoral version, neck-shaft angle, and acetabular anteversion in Chinese Han population: a retrospective analysis of 466 healthy adults. Medicine 2015b; 94: e891.

Johnson N D, Wood B P, Noh K S, Jackman K V, Katzberg R W, Westesson P L. MR imaging anatomy of the infant hip. AJR Am J Roentgenol 1989; 153: 127-33.

Kim S S, Frick S L, Wenger D R. Anteversion of the acetabulum in developmental dysplasia of the hip: analysis with computed tomography. J Pediatr Orthop 1999; 19: 438-42.

Kobayashi D, Satsuma S, Kuroda R, Kurosaka M. Acetabular development in the contralateral hip in patients with unilateral developmental dysplasia of the hip. J Bone Joint Surg Am 2010; 92: 1390-7.

Krasny R, Preacher A, Botschek A, Lmker R, Casser H R, Adam G. MRanatomy of infants hip: comparison to anatomical preparations. Pediatr Radiol 1991; 21: 211-15.

Li L Y, Zhang L J, Zhao Q, Wang E B. Measurement of acetabular anteversion in developmental dysplasia of the hip in children by two- and three-dimensional computed tomography. J Int Med Res 2009; 37: 567-75.

Li L Y, Zhang L J, Li Q W, Zhao Q, Jia J Y, Huang T. Development of the osseous and cartilaginous acetabular index in normal children and those with developmental dysplasia of the hip: a cross-sectional study using MRI. J Bone Joint Surg Br 2012; 94: 1625-31.

Li Y Q, Liu Y Z, Zhou Q H, Chen W D, Li J C, Yu L J, Xu H W, Xie D H. Magnetic resonance imaging evaluation of acetabular orientation in normal Chinese children. Medicine 2016; 95: e4878.

Mootha A K, Saini R, Dhillon M S, Aggarwal S, Kumar V, Tripathy S K. MRI evaluation of femoral and acetabular anteversion in developmental dysplasia of the hip: a study in an early walking age group. Acta Orthop Belg 2010; 76: 174-80.

Narayanan U, Mulpuri K, Sankar W N, Clarke N M, Hosalkar H, Price C T, International Hip Dysplasia Institute. Reliability of a new radiographic classification for developmental dysplasia of the hip. J Pediatr Orthop 2015; 35: 478-84.

Rubalcava J, Gómez-García F, Ríosreina J L. [Acetabular anteversion angle of the hip in the Mexican adult population measured with computed tomography]. Acta Ortop Mex 2012; 26: 155-61.

Song F S, Mccarthy J J, MacEwen G D, Fuchs K E, Dulka S E. The incidence of occult dysplasia of the contralateral hip in children with unilateral hip dysplasia. J Pediatr Orthop 2008; 28: 173-6.

Zamzam M M, Kremli M K, Khoshhal K I, Abak A A, Bakarman K A, Alsiddiky A M, Alzain K O. Acetabular cartilaginous angle: a new method for predicting acetabular development in developmental dysplasia of the hip in children between 2 and 18 months of age. J Pediatr Orthop 2008; 28: 518-23.

Acta Orthopaedica 2021; 92 (3): 347–351

347

Invasive diagnostic and therapeutic measures are unnecessary in patients with symptomatic van Neck–Odelberg disease (ischiopubic synchondrosis): a retrospective single-center study of 21 patients with median follow-up of 5 years Kristian Nikolaus SCHNEIDER 1, Lukas Peter LAMPE 1, Georg GOSHEGER 1, Christoph THEIL 1, Max MASTHOFF 2, Robert RÖDL 1, Björn VOGT 1, and Dimosthenis ANDREOU 1 1 Department

of Orthopedics and Tumor Orthopedics; University Hospital of Münster, Münster; 2 Department of Radiology; University Hospital of Münster, Münster, Germany Correspondence: kristian.schneider@ukmuenster.de Submitted 2020-09-03. Accepted 2020-12-30.

Background and purpose — Van Neck–Odelberg disease (VND) is a self-limiting skeletal phenomenon characterized by a symptomatic or asymptomatic uni- or bilateral overgrowth of the pre-pubescent ischiopubic synchondrosis.

It is frequently misinterpreted as a neoplastic, traumatic, or infectious process, often resulting in excessive diagnostic and therapeutic measures. This study assessed the demographic, clinical, and radiographic features of the condition and analyzed diagnostic and therapeutic pathways in a large single-center cohort.

Patients and methods — We retrospectively analyzed 21 consecutive patients (13 male) with a median age of 10 years (IQR 8–13) and a median follow-up of 5 years (IQR 42–94 months), who were diagnosed at our department between 1995 and 2019.

Results — VND was unilateral in 17 cases and bilateral in 4 cases. Initial referral diagnoses included suspected primary bone tumor (n = 9), fracture (n = 3), osteomyelitis (n = 2), and metastasis (n = 1). The referral diagnosis was more likely to be VND in asymptomatic than symptomatic patients (4/6 vs. 2/15). More MRI scans were performed in unilateral than bilateral VND (median 2 vs. 0). All 15 symptomatic patients underwent nonoperative treatment and reported a resolution of symptoms and return to physical activity after a median time of 5 months (IQR 0–6).

Interpretation — By understanding the physiological course of VND during skeletal maturation, unnecessary diagnostic and therapeutic measures can be avoided and uncertainty and anxiety amongst affected patients, their families, and treating physicians can be minimized.

Van Neck–Odelberg disease (VND) is a self-limiting skeletal phenomenon characterized by an asymptomatic or symptomatic uni- or bilateral overgrowth of the pre-pubescent junction between the inferior pubic ramus and ischium, which can be seen on radiographs during skeletal maturation (Figure 1) (Herneth et al. 2004, Wait et al. 2011, Mixa et al. 2017). The condition was first described in the 1920s by Odelberg and Van Neck, who classified it as a “disease” (Odelberg 1923, van Neck 1924). Today, however, VND is considered a physiological normal variant of the ischiopubic synchondrosis (IPS) that is usually obliterated between late childhood and early adolescence by bony fusion or synostosis (Herneth et al.

2000, Mixa et al. 2017).

Figure 1. Typical enlargement of the IPS in a 14-year-old female patient with right-sided VND. © 2021 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 10.1080/17453674.2021.1882237

348

Acta Orthopaedica 2021; 92 (3): 347–351

Suspected VND on conventional radiographs Bilateral

Unilateral Asymptomatic

Asymptomatic/ symptomatic

Symptomatic MRI scan Suspicious for VND

Nonoperative treatment for 6 weeks and clinical follow-up Asymptomatic

Symptomatic

Suspicious for differential diagnoses Careful evaluation of CT scan / biopsy

Radiographic follow-up

Figure 2. CT scan showing enlargement of the left IPS in a 16-year-old female patient with VND.

Despite VND being considered a normal variant, unilateral radiographic changes of the IPS are frequently misinterpreted as neoplastic, traumatic, or infectious processes (Herneth et al.

2000, Wait et al. 2011). Particularly in symptomatic patients, this often results in excessive, unnecessary, invasive, and costly diagnostic measures causing uncertainty and anxiety amongst patients and their families, as well as treating physicians (Herneth et al. 2000, Wait et al. 2011).

There are several case reports on VND in the literature; however, there are only 1 small series of 10 cases and 1 systematic review of 29 patients available, providing only limited data on demographics, possible diagnostic and therapeutic pathways, and patient outcome (Wait et al. 2011, Mixa et al. 2017). We therefore conducted this study to assess demographic, clinical, and radiographic features of VND, analyze the course of the disease, and evaluate the functional follow-up in a large single-center cohort. Based on these data, we additionally developed a standardized algorithm to help minimize the use of unnecessary diagnostic measures and to simplify diagnosis.

Patients and methods Data on patients who were diagnosed with VND at our department between 1995 and 2019 were retrieved from our hospital information system. 21 patients (13 male) were identified and included in the study. Pertinent data regarding diagnosis, clinical and radiographic features, as well as patient treatment and follow-up were retrospectively obtained from patients’ records.

All patients were consulted by specialized tumor or pediatric orthopedic surgeons. Patients presenting with only an MRI underwent additional radiographic imaging to confirm diagnosis, whereas patients presenting with radiographs and/or computed tomography (CT) scans did not undergo additional MRI if the conventional radiographic findings were typical of VND (Figure 2). Symptomatic patients were treated with analgesics and physical therapy, and were advised to refrain from sports and painful physical activities for 6 weeks. Symp-

No progression

Progression

Figure 3. Standardized algorithm in our institute for diagnosis and treatment of VND.

tomatic patients underwent a clinical and radiographic followup examination after 3–4 months to exclude a relevant growth.

Further consultations were recommended only in patients with persisting or new symptoms (Figure 3).

Statistics Medians with ranges were calculated for non-normally distributed data. Contingency tables were analyzed using the chisquare test. Non-parametric analyses were performed with the Mann–Whitney U-test. The distribution of dichotomous variables was evaluated with a 1-sample binominal test. All p-values were 2-sided; a p-value < 0.05 was considered significant. Statistical calculations were performed with SPSS Version 25.0 (IBM Corp, Armonk, NY, USA).

Ethics, funding, and potential conflicts of interest The study was approved by the local ethics committee (EthikKommission der Ärztekammer Westfalen-Lippe, No. 2019553-f-S) and performed in accordance with the Declaration of Helsinki. We acknowledge support from the Open Access Publication Fund of the University of Münster. The authors have no conflicts of interest to report.

Results The median age at diagnosis in our cohort was 10 years (IQR 8–13), and similar between male and female patients (median 9 vs. 11 years). The median follow-up was 5 years (IQR 42–94 months). 17 patients presented with a uni- and 4 patients with a bilateral VND. 15 patients presented with local complaints, consisting of groin pain (n = 10), gluteal pain (n = 2), upper leg pain (n = 2), and hip pain (n = 1). VND was an incidental finding in the remaining 6 patients who underwent imaging due to Legg–Calvé–Perthes disease (n = 2), and 1 case each for coxitis fugax, staging for Ewing’s sarcoma of the right ulna, and persistent contralateral knee pain and trauma. Initial

Acta Orthopaedica 2021; 92 (3): 347–351

A

349

B

Figure 4. Bilateral VND in a 10-year-old female patient with right-sided Legg–Calvé–Perthes disease (A). Complete dissolution of the bilateral VND at 4-year follow-up (B).

referral diagnoses by local doctors to the outpatient clinic of our tertiary hospital included suspected primary bone tumors (n = 9), fractures (n = 3), osteomyelitis (n = 2), and bone metastasis (n = 1). In the remaining 6 patients the suspected diagnosis was VND. The suspected diagnosis by the referring physicians was more likely to be VND in asymptomatic than in symptomatic patients (4/6 vs. 2/15, p = 0.02).

55 conventional radiographs, 37 MRI scans, and 9 CT scans were performed in our cohort. The 2 patients with Legg– Calvé–Perthes disease, who had bilateral VND, underwent 8 radiographic examinations each to monitor the development of the disease and the containment of the femoral head. Among the remaining patients, we found similar median numbers of conventional radiographs of patients with unilateral and bilateral VND (median 2 vs. 1). On the other hand, we observed that patients with bilateral VND underwent fewer MRI scans compared with patients with unilateral VND (median 0 vs. 2, p = 0.01).

The entire radiographic dissolution of VND was observed in 4 patients, all of which were asymptomatic. In the 1st patient the radiographic dissolution was observed at a 6-month radiographic follow-up examination with a local physician, whilst dissolution in the 2nd patient was observed 3 years following the initial diagnosis as part of the routine oncological followup of a Ewing’s sarcoma. The remaining 2 patients both had Legg–Calvé–Perthes disease and bilateral VND. Whilst we observed a synchronous dissolution in 1 patient 4 years after initial diagnosis (Figure 4), the dissolution in the 2nd patient was asynchronous, 2.5 years and 3 years after initial diagnosis, respectively.

A diagnostic biopsy was performed in 2 patients. 1 of them underwent staging for a Ewing’s sarcoma of the right ulna with a fluorodeoxyglucose positron emission tomography/ computed tomography (FDG-PET/CT) scan, which showed increased FDG uptake in the left lower pelvis. Despite the fact that conventional radiographs were consistent with the diagnosis of VND, a biopsy was recommended by the treating pediatric oncologists to exclude metastatic disease, given the impact this would have on patient treatment and progno-

Figure 5. T2-weighted MRI scan with a left-sided VND suspicious of osteomyelitis in a 7-year-old male patient.

sis. The second patient’s history, laboratory results, and MRI findings were suspicious of osteomyelitis, which could be ruled out in the microbiological and histological exams of the biopsy specimen (Figure 5). While we found regular cortical bone with normal osteoblastic and osteoclastic activity in the first patient, histopathological findings of the second patient showed hyaline cartilage and lamellar bone.

All 15 patients with local complaints reported a resolution of symptoms and return to physical activity after the median follow-up time of 5 months.

Discussion In 1923, Odelberg first described “destructive alterations of os ischii” in the adolescent and ruled out trauma, tumor, tuberculosis, and lues as a cause, concluding that “some form of non-specific chronic inflammation” might cause the condition (Odelberg 1923). In 1924, van Neck termed the radiographic observations “osteochondritis of the pubis” (van Neck 1924).

Since then, several authors have shown that radiographic changes of the IPS can be part of the normal fusion process during skeletal maturation (Byers 1963, Neitzschman 1997, Ceroni et al. 2004). While in conventional radiographs VND

350

Acta Orthopaedica 2021; 92 (3): 347–351

Figure 6. VND in a 22-year-old female patient.

typically presents with a characteristic enlargement or overgrowth of the IPS (Figure 1), hyperintense (T2-weighted) or hypointense (T1-weighted) signals around the IPS are found on MRI (Mixa et al. 2017). However, despite the improved understanding of this physiological process, there is a lack of clinical studies describing the course of VND, and radiographic alterations of the IPS still cause uncertainty among treating physicians.

Our study was designed to address this deficit and found that: (1) the median age at diagnosis of VND is 10 years.

(2) Initial diagnosis of VND is more likely to be accurate in asymptomatic than symptomatic patients. (3) More MRI scans are performed in unilateral than in bilateral VND. (4) Nonoperative treatment can quickly lead to a resolution of symptoms and return to full physical activity appears to be possible after a median time of 4 months.

The fusion of the IPS has been shown to be strongly agerelated and usually occurs in girls between the age of 4 and 9 years and in boys between the age of 7 and 13 years (Gregory et al. 2019). Although our findings are in general similar, we found a VND in a 22-year-old female patient with a 5-month history of left-sided groin pain and the referral diagnosis of a primary bone tumor (Figure 6). Although the VND in this case was most likely just an incidental finding, this case suggests that not all patients experience complete radiographic resolution of VND after skeletal maturation. Likewise, Morse and Lin (2016) have previously reported a VND case in a 17-year-old female patient with radiographic progression around the IPS during 4-month follow-up leading to a biopsy that confirmed the benign entity. We therefore conclude that physicians should be aware of VND as a possible differential diagnosis in adolescents and young adults, rather than just in prepubertal children.

The synostosis of the IPS can occur simultaneously or asynchronously and asymmetrically and lead to bilateral or uni-

lateral VND, respectively (Cawley et al. 1983). As a bilateral symmetrical appearance is rarely mistaken for pathology, the diagnosis of VND can be especially challenging in unilateral cases (Herneth et al. 2000). This was also the case in our analysis, where all 4 bilateral but only 2 of 17 unilateral cases were initially diagnosed correctly. The most frequent referral diagnosis in our cohort was a suspected primary bone tumor, which is in line with previous reports (Herneth et al. 2000, Mixa et al. 2017). Another possible frequent differential of VND is ischiopubic osteomyelitis (Wait et al. 2011). Wait et al. (2011) have argued that MRI may help to differentiate between the two entities, with ischiopubic osteomyelitis showing typical myositis, abscess, or fluid collections surrounding the IPS whereas VND appears with a characteristic focal area of marrow edema. However, we performed a biopsy in a 7-yearold boy in our cohort with typical laboratory and suggestive MRI findings of osteomyelitis, but the microbiological results were negative and histopathological findings were consistent with VND, where typical histopathology findings include (hyaline) cartilage and lamellar bone (Mixa et al. 2017). The patient underwent nonoperative, non-antibiotic treatment and was asymptomatic after 6 months, highlighting the possible difficulties in diagnosing VND. Another challenging aspect is the diagnosis in patients who suffer from a primary malignant bone tumor, like our patient with Ewing’s sarcoma. A biopsy may sometimes be inevitable to rule out possible bone metastasis, which would impact the further oncological treatment.

Similarly, Drubach et al. (2006) have described the case of a 10-year-old girl with a renal cell carcinoma and high uptake around a lucent expansile lesion at the right ischiopubic junction in an FDG PET/CT scan. In a previous scan 3 months earlier the IPS appeared closed and had no pathological uptake, so that the authors performed a biopsy expecting to find metastatic disease, but histopathological analysis confirmed the diagnosis of VND (Drubach et al. 2006).

Patients with VND are frequently subject to regular imaging with ionizing radiation (Cawley et al. 1983, Drubach et al.

2006, Macarini et al. 2011). In a systematic review, Wait et al.

(2011) reported 23 radiographs, 17 MRI scans, 4 CT scans, 3 PET/CT scans, and 2 bone scans performed for diagnosis in 29 patients. All of the CT, PET/CT, and bone scans in their study were performed in patients with unilateral VND, a finding we were able to confirm in our cohort (Wait et al. 2011).

We also found a higher number of MRI scans being performed in unilateral VND, emphasizing the challenges of diagnosis in unilateral compared with bilateral cases.

As VND is benign self-limiting condition, nonoperative treatment of symptomatic patients is recommended and should be accompanied by clinical and, if necessary, radiographic follow-up examinations (Mixa et al. 2017). However, there are also occasional reports on the surgical treatment of VND (Byers 1963, Oliveira 2010, Wait et al. 2011). Byers (1963) described an excision of the IPS in an 11-year-old boy due to “the severity of pain and uncertainty of the diagno-

Acta Orthopaedica 2021; 92 (3): 347–351

sis” and reported postoperative relief of pain, while Oliveira (2010) performed surgical curettage after failed nonoperative treatment in an 8-year-old boy. Wait et al. (2011) have outlined the lack of long-term follow-up data in VND. Given the results of our study, the resolution of pain in symptomatic patients appears to be very likely under nonoperative treatment, usually within a few months, but symptoms can also last for up to 1 year. A recurrence or possible sequelae was not observed in any of our patients. We therefore strongly recommend avoiding surgery in these patients.

Based on our findings we developed a standardized algorithm for the diagnosis and treatment of VND, in order to reduce uncertainty in treating physicians and avoid unnecessary ionizing radiation imaging in the affected patients (Figure 3).

In conclusion, despite advances in radiographic modalities, VND is still a condition that many physicians find difficult to diagnose. An understanding of the physiological processes of IPS fusion during skeletal maturation is essential and the standardized algorithm we developed may contribute to avoiding unnecessary diagnostic and therapeutic measures and minimizing uncertainty and anxiety amongst affected patients, their families, and treating physicians. Nonoperative treatment appears to be adequate for symptomatic patients and relief of symptoms is usually achieved within a few months, although cases with persisting pain for up to 1 year are possible.

KNS, LPL, GG, CT, MM, RR, BV, and DA designed the study and collected the data. KNS, CT, LPL, and DA were responsible for data management, data analysis, and preparation of figures. KNS and DA wrote the manuscript. KNS, MM, RR, BV, GG, and DA helped with data analysis and with editing of the manuscript.

Acta thanks Peter Holmberg Jørgensen and other anonymous reviewers for help with peer review of this study.

351

Byers P D. Ischio-pubic ‘osteochondritis’: a report of a case and a review. J Bone Joint Surg Br 1963; 45(4): 694-702.

Cawley K A, Dvorak A D, Wilmot M D. Normal anatomic variant: scintigraphy of the ischiopubic synchondrosis. J Nucl Med 1983; 24(1): 14-16.

Ceroni D, De Coulon G, Regusci M, Kaelin A. Gorham–Stout disease of costo-vertebral localization: radiographic, scintigraphic, computed tomography, and magnetic resonance imaging findings. Acta Radiol 2004; 45(4): 464-8.

Drubach L A, Voss S D, Kourmouzi V, Connolly L P. The ischiopubic synchondrosis: changing appearance on PET/CT as a mimic of disease. Clin Nucl Med 2006; 31(7): 414-7.

Gregory L S, Jones L V, Amorosi N M. Assessing the fusion of the ischiopubic synchondrosis using predictive modeling. Clin Anat 2019; 32(6): 851-9.

Herneth A M, Trattnig S, Bader T R, Ba-Ssalamah A, Ponhold W, WandlVergesslich K, et al. MR imaging of the ischiopubic synchondrosis. Magn Reson Imaging 2000; 18(5): 519-24 Herneth A M, Philipp M O, Pretterklieber M L, Balassy C, Winkelbauer F W, Beaulieu C F. Asymmetric closure of ischiopubic synchondrosis in pediatric patients: correlation with foot dominance. Am J Roentgenol 2004; 182(2): 361-5.

Macarini L, Lallo T, Milillo P, Muscarella S, Vinci R, Stoppino L P. Case report: Multimodality imaging of van Neck–Odelberg disease. Indian J Radiol Imaging 2011; 21(2): 107-10.

Mixa P J, Segreto F A, Luigi-Martinez H, Diebo B G, Naziri Q, Kolla S, et al. van Neck–Odelberg disease: a 3.5-year follow-up case report and systematic review. Surg Technol Int 2017; 31: 365-73.

Morse L J, Lin P P. Groin and medial thigh pain in a 17-year-old girl. Clin Orthop Relat Res 2016; 474(2): 594-9.

Neitzschman H R. Radiology case of the month. Hip trauma: normal physiologic asymmetric closure of the ischiopubic synchondroses. J La State Med Soc 1997; 149(6): 186-8.

Odelberg A. Some cases of destruction in the ischium of doubtful etiology.

Acta Chirurgica Scandinavica 1923; (56): 273-84.

Oliveira F. Differential diagnosis in painful ischiopubic synchondrosis (IPS): a case report. Iowa Orthop J 2010; 30: 195-200. van Neck M. Osteochondrite du pubis. Archives franco-belges de Chirurgie 1924; (27): 238-41.

Wait A, Gaskill T, Sarwar Z, Busch M. Van neck disease: osteochondrosis of the ischiopubic synchondrosis. J Pediatr Orthop 2011; 31(5): 520-4.

352

Acta Orthopaedica 2021; 92 (3): 352–357

Surgical treatment of skeletal metastases in proximal tibia: a multicenter case series of 74 patients Kaarel KILK 1,2, Jessica EHNE 3, Jonathan D STEVENSON 4,5, Gilber KASK 1,2, Jyrki NIEMINEN 6, Rikard WEDIN 3, Michael C PARRY 4,5, and Minna K LAITINEN 1 1 Department

of Orthopaedics, Helsinki University Hospital and University of Helsinki, Helsinki, Finland; 2 Department of Orthopaedics, Tampere University Hospital, Tampere, Finland; 3 Department of Reconstructive Orthopaedics, Karolinska University Hospital, Stockholm, Sweden; 4 Department of Orthopaedics, Royal Orthopaedic Hospital, Birmingham, UK; 5 Aston University Medical School, Aston University, Birmingham, UK; 6 Coxa, Hospital for Joint Replacement, Tampere, Finland Correspondence: kaarel.kilk@pshp.fi Submitted 2020-09-08. Accepted 2020-11-23.

Background and purpose — The proximal tibia is a rare site for metastatic bone disease and is a challenging anatomical site to manage due to the proximity to the knee joint and poor soft tissue envelope. We investigated implant survival and complications of different surgical strategies in the treatment of proximal tibia pathological fractures.

Patients and methods — The study comprised a 4 medical center, retrospective analysis of 74 patients surgically treated for metastases of the proximal tibia. Patient records were reviewed to identify outcome, incidence, and type of complications as well as contributing factors.

Results — Reconstruction techniques comprised cementaugmented osteosynthesis (n = 33), tumor prosthesis (n = 31), and total knee arthroplasty with long cemented stems (n = 10). Overall implant survival was 88% at 6 months and 1 year, and 67% at 3 years. After stratification by technique, the implant survival was 82% and 71% at 1 and 3 years with tumor prosthesis, 100% at 1 and 3 years with total knee arthroplasty, and 91% at 1 year and 47% at 3 years with osteosynthesis. Preoperative radiotherapy decreased implant survival. Complications were observed in 19/74 patients.

Treatment complications led to amputation in 5 patients.

Interpretation — In this study, the best results were seen with both types of prothesis reconstructions, with good implant survival, when compared with treatment with osteosynthesis. However, patients treated with tumor prosthesis showed an increased incidence of postoperative infection, which resulted in poor implant survival. Osteosynthesis with cement is a good alternative for patients with short expected survival whereas endoprosthetic replacement achieved good medium-term results.

The most common site for skeletal metastases requiring surgical intervention is the proximal femur, which accounts for approximately 65% of all cases. In contrast, the tibia accounts for only 3% of pathological fractures requiring surgery, mostly commonly in the proximal third (Ratasvuori et al. 2013). Due to the proximity of the knee joint and the poor soft tissue envelope in the proximal tibia, the management of metastatic deposits and pathological fractures in this region can be challenging.

Given the scarcity of this metastasis location, there is a paucity of evidence relating to the outcomes of surgical treatment of pathological fractures of the proximal tibia. In this retrospective case series we assessed the advantages and disadvantages of different surgical reconstructions, looking in particular at implant survival, the incidence of complications, and the possible factors that may affect these outcomes.

Patients and methods The study comprised a retrospective analysis of all patients treated for a complete or pending pathological fracture arising within the proximal tibia treated at 1 of 4 international collaborative hospitals: Helsinki University Hospital, Helsinki, Finland;, Coxa Hospital for Joint replacement, Tampere, Finland; Karolinska University Hospital, Stockholm, Sweden; and Royal Orthopaedic Hospital, Birmingham, UK. The study population comprised 74 patients treated between 2000 and 2018. All patients over 18 years of age with histologically confirmed metastatic bone disease of any primary malignancy, including multiple myeloma and lymphoma, were included.

The decision to undergo surgical intervention was discussed

© 2021 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 10.1080/17453674.2020.1866242

Acta Orthopaedica 2021; 92 (3): 352–357

Metastatic load

Diagnosis

Sex

Expected survival

Radiotherapy

353

Age

Preoperative activity Implant survival

Surgical technique

Implant

Figure 1. Causal pathways in directed acyclic graphs in the variable selection. Exposure of interest = implant survival, outcome = implant survival, suggested covariates, sex, age, diagnosis, metastatic load, radiotherapy.

at a multidisciplinary team conference at each of the 4 centres and was made following discussion between the operating surgeon and the patient. Preoperative radiological assessment comprised plain radiographs in all cases, and in selected cases with computed tomography (CT) or magnetic resonance imaging (MRI). Systemic staging comprised CT scan of the chest, abdomen, and pelvis, and whole-body skeletal imaging in the form of radiolabelled technicium bone scan.

Data was extracted from prospectively maintained institutional databases as well as medical records. Radiological assessment of relevant imaging was undertaken to assess eligibility. Patient- and reconstruction-related outcome measures were recorded. A causal-directed acyclic graph (DAG) was used to investigate confounding factors (Figure 1).

The surgical treatment methods for impending or pathological fracture were stratified into 1 of 3 possible reconstructions: tumor prosthesis, total knee arthroplasty (TKA) with long cemented stems, or osteosynthesis using plate and cement (Figure 2).

Postoperative complications, including mechanical complications, were defined according to the classification by Henderson et al. (2014) (Table 1). Additionally, complications were defined as minor and major. Major complications were defined as those that required further surgical intervention.

Minor complications were defined as those that did not require further surgical intervention.

Statistics Patient and implant survival rates were assessed using the Kaplan–Meier methods with 95% confidence intervals (CI).

Between-group comparisons were performed using the logrank test. Patient follow-up time was calculated from the date of surgery to the most recent follow-up date or the date of death. Implant survival was calculated from the date of surgery to revision surgery due to any cause. Continuous variables are reported as medians. The chi-squared test or Fisher’s exact test was used to compare variables between groups, and the Mann–Whitney U-test test for medians between groups.

Subdistribution hazard ratio (SHR) of the role of factors affecting implant survival was calculated using competing risk analysis, where death was considered as a competing event.

Statistical analyses were performed using SPSS Statistics 23.0

A

B

C

Figure 2. Surgical treatment methods: tumor prosthesis (A), total knee arthroplasty with long cemented stems (B), osteosynthesis using plate and cement (C).

(IBM Corp, Armonk, NY, USA) but competing risk analysis was performed using STATA 16 (StataCorp, College Station, TX, USA). A p-value < 0.05 was considered significant.

Ethics, funding, and potential conflicts of interest This retrospective study was approved by the local chairs of the audit department. This research study received a grant from the state funding for university-level health research.

The funder had no role in study design, data collection, and analysis, decision to publish, or preparation of the manuscript.

No competing interests are declared.

Results The study population comprised 74 patients (45 men) with a median age at the time of surgery of 64 years (18–86). The median follow-up period was 12 months (0–210) and at final follow-up, 26 patients were alive. The most common primary malignancy was renal cell carcinoma (RCC, n = 29), followed by melanoma (8), colon cancer (6), breast cancer, sarcoma, lung cancer, and myeloma (5 cases each). Of the 74 patients, 64 patients had an impending fracture and 10 patients had a complete fracture (Table 1). The overall mortality was 64% during the period of follow up. Overall patient survival after 6 months was 74% (61–83), at 1 year 58% (46–70), and at 3 years, 33% (21–45). Statistically significant factors negatively associated with survival were the incidence of a major complication (p = 0.04) and the presence of an actual pathological fracture (p = 0.02).

354

Acta Orthopaedica 2021; 92 (3): 352–357

Table 1. Patient characteristics (N = 74).

Values are count unless otherwise specified Male/Female 55/29 Impending fracture 64 Preoperative radiotherapy 12 Primary tumor Renal cell carcinoma 29 Melanoma 8 Colon carcinoma 6 Non-small-cell lung carcinoma 5 Sarcoma 5 Myeloma 5 Breast carcinoma 5 Prostate carcinoma 3 Esophagus 2 Bladder 2 Lymphoma 2 Retinoblastoma 1 Squamocellular 1 Median age, years a 64 (18–86) Mean follow-up, months a 12 (0–210) Mean size, cma 6.3 (3–16) Operative method Tumor prosthesis 31 Total knee arthroplasty with

long cemented stems 10 Osteosynthesis and cement 33 Complications 19 Complications according to Henderson’s classification 18 Type 1 2 Type 2 0 Type 3 2 Type 4 8 Type 5 6 Revision surgery 13 a

Range in parenthesis

Table 2. Complications

Osteo- Permanent Fatal synthesis Tumor nerve pulmonary Type Infection problem progression palsy embolism Tumor prosthesis TKA with long cemented stems Osteosynthesis with cement

8/11 0 0

0 0 2/8

1/11 0 5/8

2/11 0 0

0 0 1/8

Table 3. Treatment of complications

2-stage Conversion to Re-osteoType DAIR revision Amputation prosthesis synthesis Tumor prosthesis TKA with long cemented stems Osteosynthesis with cement

Implant survival (%)

1/7 0 0

2/7 0 0

4/7 0 1/6

0 0 3/6

Implant survival (%)

Implant survival (%)

100

100

100

90

90

90

80

80

80

70

70

70

60

60

60

50

50

40 0

25

50

75

100

125

Months from index operation

Figure 3. Implant survival.

Implant survival The chosen reconstruction method is given in Table 1. Overall implant survival, regardless of reconstruction technique, was 88% (80–97) after 6 months and 1 year, and 67% (50–83) at 3 and 5 years (Figure 3). After stratifying by operative method, implant survival for the tumor prosthesis group was 93% (84– 100) at 6 months, 82% (67–96) at 1 year, and 71% (52–90) at 3 and 5 years. In the osteosynthesis group, the implant survival was 91% (79–100) at 6 months and 1 year, and 47% (15–79) at 3 and 5 years. In the total knee arthroplasty (TKA) with long cemented stem group, implant survival was 100% after 6 months, 1, 3, and 5 years. The effect of preoperative radiotherapy and reconstruction technique on implant survival was analysed using a competing risk model. There was worse implant survival for the osteosynthesis group, but without statistical significance (p = 0.2) (Figure 4). Preoperative radiotherapy had a significantly negative effect on implant survival compared with no preoperative radiotherapy in the tumor prosthesis and osteosynthesis groups (p = 0.004) (Figure 5).

No other factors, identified by DAG, had a statistically significant effect on implant survival.

TKA with long cemented stems Tumor prosthesis Osteosynthesis 25

50

75

100

No preoperative radiotherapy

Preoperative radiotherapy

50 40

40 0

0 0 2/6

125

0

25

50

75

100

125

Months from index operation

Months from index operation

Figure 4. Implant survival stratified by surgical method in a competing risk model.

Figure 5. Implant survival stratified by radiotherapy in a competing risk model.

Complications Postoperative complications were seen in 19/74 patients.

The most common was deep wound infection (8), followed by tumor progression (6), aseptic loosening or osteosynthesis failure of reconstruction/osteosynthesis (2), peroneal nerve paralysis (2), and fatal pulmonary embolus (1).

13/19 complications necessitated revision surgery (Tables 2 and 3). 5 patients required subsequent amputation due to prosthetic joint infection (3) or tumor progression (2). In 3 cases, further revision surgery was required to treat a periprosthetic joint infection, either by 2-stage revision in 2 cases, or by debridement, antibiotics, and implant retention (DAIR) in 1 case. In 4 cases treated by osteosynthesis, the fixation failed, which required revision surgery in 2 cases. In the other 2 cases, nonoperative management was advocated due to health deterioration. 3 patients treated with osteosynthesis suffered tumor progression and underwent revision surgery to a tumor prosthesis.

6/7 patients with postoperative periprosthetic joint infection (PJI) underwent revision surgery. 3/6 patients had received preoperative radiotherapy. All infections with preoperative

Acta Orthopaedica 2021; 92 (3): 352–357

355

Table 4. Complications and radiotherapy Osteo- Permanent Fatal

synthesis Tumor nerve pulmonary Type n Infection problem progression palsy embolism Tumor prosthesis (n = 31)

Preoperative radiotherapy 5 2 0 1

No preoperative radiotherapy 26 6 0 0 TKA with long cemented stems (n = 10)

Preoperative radiotherapy 0 – – –

No preoperative radiotherapy 10 0 0 0 Osteosynthesis with cement (n = 33)

Preoperative radiotherapy 7 0 1 1

No preoperative radiotherapy 26 0 1 1

radiotherapy occurred in the tumor prosthesis group. There were no infections in the osteosynthesis group (33 patients), despite a comparable incidence of preoperative radiotherapy.

None of the 10 patients in TKA group underwent preoperative radiotherapy, and no postoperative infections requiring surgical intervention were seen. The outcome of patients who had received preoperative radiotherapy are summarized in Table 4. The indication for amputation was infection (3) or tumor progression/recurrence (2). 4/5 patients requiring amputation underwent pre- or postoperative radiotherapy.

Discussion The literature describing the outcomes of proximal tibial pathological fractures is limited (Smolle et al. 2019), but what is known is that treatment complications are common for tumors in this location (Mavrogenis et al. 2013). The optimum method for reconstruction lacks consensus.

Reconstruction of the proximal tibia using a tumor endoprosthesis is associated with a high rate of complication and failure (Smolle et al. 2019). Our study demonstrates that this is also true for patients treated with tumor prosthesis for MBD in the proximal tibia. Moreover, the complications were more severe when compared with patients treated with the other surgical methods studied. In particular, the risk of amputation due to infection was highest in patients treated with tumor prosthesis, and infection occurred in the early stages after the operation, consistent with reported findings for proximal tibial replacement following primary malignant bone tumor resection, where infection rates ranged between 6% and 44% (Flint et al. 2006, Myers et al. 2007, Wu et al. 2008, Schwartz et al.

2010, Mavrogenis et al. 2013, Muller et al. 2016). Amputation is the most devastating complication following proximal tibia reconstruction surgery. In our study the rate of amputation after tumor prosthesis infection was 5/8, which is higher than previously reported in primary bone tumor studies with evidently younger patients, where amputation rates due to infection were 5/12 in the study by Tsagozis et al. (2018) and 8/27

0 0 2 0 – – 0 0 0 0 0 1

in the study by Mavrogenis et al. (2013). The infection rate increases for the lifetime of the prosthesis due to the need for revision and servicing procedures, and can be as high as 87% (Ilyas et al. 2001, Grimer et al. 2002, Plotz et al. 2002, Jeys et al. 2005, Mavrogenis et al. 2011, Mavrogenis et al. 2015). The need for further revision procedures for proximal tibial tumor prosthesis was highlighted by Theil et al., who described the need for further procedures, including the management of infection, in 115/234 patients following a revision procedure (Theil et al. 2019).

The short-term implant survival in the osteosynthesis group was similar to that seen in the endoprosthesis group. Several patients in this group developed complications requiring revision due to infection, failure of fixation, or tumor progression.

The need for revision surgery in a plate-osteosynthesis group increases approximately 1 year after primary surgery. Moreover, due to comorbidities, some of these revision procedures were not undertaken. This may represent a selection bias for this method of reconstruction being used in patients in whom survival is considered to be short (less than a year). Due to the intralesional nature of this procedure, adjuvant local or systemic treatment, aimed at reducing the risk of local recurrence or tumor progression, should be considered.

We found the best results, in terms of implant survival or the need for revision surgery, in the cemented long-stemmed TKA group. However, it should be noted that the follow-up time in this patient group was the shortest. As was demonstrated in the other groups, as follow-up increases, the risk of tumor progression increases and thus the need for further surgical intervention. This is reflected by the implant survival rate when studied with competing risk analysis. However, in the short term at least, the incidence of PJI in the long-stem TKA group appears not to present the same challenge as seen in the tumor prosthesis group.

The increased risk of infection when undertaking procedures around the proximal tibia is well established and is at least in part due to the poor soft tissue envelope and the need for extensive dissection to mobilize the proximal tibia. The introduction of gastrocnemius flaps to improve the soft tissue envelope has

356

resulted in a reduction in infections (Myers et al. 2007), but postoperative infection at this site remains higher than for other anatomical locations (Jeys et al. 2005). A gastrocnemius flap was used in all tumor prosthesis reconstructions in our study but, in spite of this, the infection rate remained higher when compared with that seen following resection and reconstruction of a primary malignant tumor of bone. In comparison with the study by Mavrogenis et al. (2013), where complications following proximal tibial replacement were seen in 56 of 225 patients, and the infection rate was 27/225, we found complications in 17/74 patients, with an incidence of infection of 7/74.

It is worthy of note that the median age of patients in our study was 64 years compared with 27 years in the study by Mavrogenis et al., which concerned primary malignant bone tumors.

The role of radiotherapy in increasing complications is known (Theil et al. 2019). In our study, radiotherapy prior to tumor prosthesis further increased the risk of infection, despite the addition of a gastrocnemius flap. Radiotherapy leading to PJI often necessitates amputation to eradicate the infection (Mavrogenesis et al. 2011). This must be borne in mind when planning the method of reconstruction to address proximal tibial MBD. Higher age, poor general condition, and suppressed wound healing, for example by antiangiogenetic drugs, further increases the risk of delayed wound healing and subsequent infection (Carroll et al. 2014).

This study does have its limitations. This study is retrospective, with the inherent limitations. However, to our knowledge this is the first study focusing on the surgical management of proximal tibia metastases and, despite the limited numbers, represents the largest cohort to date. The small numbers presented will undoubtedly result in selection bias towards the mechanism of reconstruction. The study does not include patient-related quality of life or functional outcome. Thus a less interventional procedure may carry a better functional outcome but at the risk of tumor progression and the need for later surgical interventions. In addition, it was not possible to accurately assess patient comorbidity status, which may have contributed to the incidence of postoperative complications.

Clearly, this must be considered when planning the method of reconstruction in the treatment of MBD in the proximal tibia.

In conclusion, this study highlights the importance of expected patient survival when considering the method of reconstruction of the proximal tibia. For patients with a limited prognosis estimated at 6 months, a plate osteosynthesis with supplemented cement should be considered. For patients with an expected survival in excess of 12 months, a more robust reconstruction should be considered. A cemented long-stem knee replacement provides a durable reconstruction without significant complications, at least in the short term. Tumor prosthesis require prolonged postoperative immobilization due to the need for reconstruction of the extensor mechanism using a gastrocnemius flap, and are associated with a higher incidence of postoperative complications when compared with long-stem cemented knee replacement.

Acta Orthopaedica 2021; 92 (3): 352–357

Preoperative radiotherapy may increase the risk of complications, particularly infection, which in the presence of an endoprosthesis may require amputation.

MKL and KK concepted ad designed the study. Data extraction was performed by MKL, KK, GK, JE, RW and MCP. Data analysis and interpretation was performed by MKL and KK. KK, MKL and GK were major contributors in writing the article. Article drafting and revising were performed

by MKL, MCP, RW, JN, JDS. The final version was approved by all authors.

Acta thanks M A J van de Sande and Joachim Thorkildsen for help with peer review of this study

Carroll K, Dowsey M, Choong P, Peel T. Risk factors for superficial wound complications in hip and knee arthroplasty. Clin Microbiology Infect Dis 2014; 20(2): 130-5.

Flint M N, Griffin A M, Bell R S, Ferguson P C, Wunder J S. Aseptic loosening is uncommon with uncemented proximal tibia tumor prostheses. Clin Orthop Relat Res 2006; 450: 52-9.

Grimer R J, Belthur M, Chandrasekar C, Carter S R, Tillman R M. Two-stage revision for infected endoprostheses used in tumor surgery. Clin Orthop Relat Res 2002; (395): 193-203.

Henderson E R, O’Connor M I, Ruggieri P, Windhager R, Funovics P T, Gibbons C L, Guo W, Hornicek F J, Temple H T, Letson G D. Classificantion of failure of limb salvage reconstructive surgery for bone tumours. Bone Joint J 2014 (96-B): 1436-40.

Ilyas I, Kurar A, Moreau P G, Younge D A. Modular megaprosthesis for distal femoral tumors. Int Orthop 2001; 25(6): 375-7.

Jeys L M, Grimer R J, Carter S R, Tillman R M. Periprosthetic infection in patients treated for an orthopaedic oncological condition. J Bone Joint Surg Am 2005; 87(4): 842-9.

Mavrogenis A F, Papagelopoulos P J, Coll-Mesa L, Pala E, Guerra G, Ruggieri P. Infected tumor prostheses. Orthopedics 2011; 34(12): 991-8; quiz 9-1000.

Mavrogenis A F, Pala E, Angelini A, Ferraro A, Ruggieri P. Proximal tibial resections and reconstructions: clinical outcome of 225 patients. J Surg Oncol 2013; 107(4): 335-42.

Mavrogenis A F, Pala E, Angelini A, Calabro T, Romagnoli C, Romantini M, Drago G, Ruggieri P. Infected prostheses after lower-extremity bone tumor resection: clinical outcomes of 100 patients. Surg Infect 2015; 16(3): 267-75.

Muller D A, Beltrami G, Scoccianti G, Cuomo P, Capanna R. Allograft-prosthetic composite versus megaprosthesis in the proximal tibia: what works best? Injury 2016; 47(Suppl. 4): S124-S30.

Myers G J, Abudu A T, Carter S R, Tillman R M, Grimer R J. The longterm results of endoprosthetic replacement of the proximal tibia for bone tumours. J Bone Joint Surg Br 2007; 89(12): 1632-7.

Plotz W, Rechl H, Burgkart R, Messmer C, Schelter R, Hipp E, Gradinger R. Limb salvage with tumor endoprostheses for malignant tumors of the knee.

Clin Orthop Relat Res 2002 (405): 207-15.

Ratasvuori M, Wedin R, Keller J, Nottrott M, Zaikova O, Bergh P, Kalen A, Nilsson J, Jonsson H, Laitinen M. Insight opinion to surgically treated metastatic bone disease: Scandinavian Sarcoma Group Skeletal Metastasis Registry report of 1195 operated skeletal metastasis. Surg Oncol 2013; 22(2): 132-8.

Schwartz A J, Kabo J M, Eilber F C, Eilber F R, Eckardt J J. Cemented endoprosthetic reconstruction of the proximal tibia: how long do they last? Clin Orthop Relat Res 2010; 468(11): 2875-84.

Smolle M A, Andreou D, Tunn P U, Leithner A. Advances in tumour endoprostheses: a systematic review. EFORT Open Rev 2019; 4(7): 445-59.

Theil C, Roder J, Gosheger G, Deventer N, Dieckmann R, Schorn D, Hardes J, Andreou D. What is the likelihood that tumor endoprostheses will experience a second complication after first revision in patients with primary

Acta Orthopaedica 2021; 92 (3): 352–357

malignant bone tumors and what are potential risk factors? Clin Orthop Relat Res 2019; 477(12):2705-14.

Tsagozis P, Parry M, Grimer R. High complication rate after extendible endoprosthetic replacement of the proximal tibia: a retrospective study of 42

357

consecutive children. Acta Orthop 2018; 89(6): 678-82.

Wu C C, Henshaw R M, Pritsch T, Squires M H, Malawer M M. Implant design and resection length affect cemented endoprosthesis survival in proximal tibial reconstruction. J Arthroplasty 2008; 23(6): 886-93.

358

Acta Orthopaedica 2021; 92 (3): 358–363

Impact of malnutrition and vitamin deficiency in geriatric patients undergoing orthopedic surgery Matthias MEYER 1, Franziska LEISS 1, Felix GREIMEL 1, Tobias RENKAWITZ 2, Joachim GRIFKA 1, Günther MADERBACHER 1, and Markus WEBER 1 1 Department

of Orthopedic Surgery, Regensburg University Hospital, Bad Abbach; 2 Heidelberg University Orthopedic Hospital, Heidelberg, Germany This work was performed at Regensburg University Hospital, Department of Orthopedic Surgery, Bad Abbach, Germany.

Correspondence: matthias.meyer@ukr.de Submitted 2020-10-03. Accepted 2021-01-11

Background and purpose — There is growing evidence that hypoproteinemia is an important risk factor for adverse events after surgery. Less is known about the impact of vitamin deficiency on postoperative outcome. Therefore we evaluated the prevalence and impact of malnutrition and vitamin deficiency in geriatric patients undergoing elective orthopedic surgery.

Patients and methods — In a retrospective analysis of 599 geriatric patients who had undergone elective orthopedic surgery in 2018 and 2019, hypoproteinemia, and deficiency of vitamin D, vitamin B12, and folate were assessed.

Reoperation rates, readmission rates, complication rates, and transfusion rates were compared between malnourished patients and patients with normal parameters. Multivariable logistic regression models were used to assess the relationship between malnutrition and postoperative adverse events, controlling for confounding factors such as age, sex, diabetes mellitus, and frailty.

Results — Patients with malnutrition showed a higher rate of reoperation (13% vs. 5.5%; p = 0.01) and exhibited more wound-healing disorders (7.4% vs. 1.3%, p = 0.001) as well as Clavien–Dindo IV° complications (7.4% vs. 2.4%; p = 0.03). Deficiency of vitamin D led to a higher rate of falls (8.4% vs. 2.9%, p = 0.006). Deficiency of vitamin B12 and folate did not affect postoperative adverse events. Although correlated to frailty (p = 0.004), multivariable regression analysis identified malnutrition as independent risk factor for reoperation (OR 2.6, 95% CI 1.1–6.2) and wound healing disorders (OR 7.1, CI 1.9–26).

Interpretation — Malnutrition is common among geriatric patients undergoing elective orthopedic surgery and represents an independent risk factor for postoperative adverse events.

Malnutrition is an important risk factor for postoperative complications in orthopedic surgery (Bohl et al. 2016, Kamath et al. 2016). Previous studies found up to 50% of patients to be at risk of malnutrition (Cross et al. 2014). As malnutrition also increases healthcare costs for orthopedic surgery, public and private payers’ growing focus on bundled payment and Value Based Payment models may promote thorough preoperative nutritional screening for economic aspects also (Bala et al.

2020).

Malnutrition is considered as a modifiable risk factor, which is defined as a medical condition that can be altered positively prior to surgery. A recently published prospective study showed that a preoperative nutritional intervention led to improved outcome in malnourished patients undergoing arthroplasty (Schroer et al. 2019). Researching into malnutrition, previous studies mainly focused on the impact of hypoproteinemia on outcome after orthopedic surgery (Cross et al.

2014, Bohl et al. 2016, Kamath et al. 2016). In contrast, less is known about the effects of vitamins and minerals. Epidemiological studies revealed a high prevalence of vitamin D, vitamin B12, and folate deficiency, especially in the elderly (Laird et al. 2018, Sempos et al. 2018). The role of vitamin D deficiency in the context of orthopedic surgery is debated (Maier et al. 2016, Shin et al. 2017, Hegde et al. 2018).

We evaluated the prevalence of malnutrition in a cohort of 599 geriatric patients undergoing elective orthopedic surgery at a high-volume university center. Furthermore, we investigated whether malnutrition, in the form of hypoproteinemia, and deficiency of certain vitamins affected the rate of adverse events after operation.

© 2021 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 10.1080/17453674.2021.1882092

Acta Orthopaedica 2021; 92 (3): 358–363

Patients and methods

359

ng/L was defined as insufficiency (Hunt et al. 2014). Folate deficiency was defined as serum folate level < 2 ng/mL. A serum folate level between 2 ng/mL and 4 ng/mL was defined as insufficiency (Green and Datta Mitra 2017).

Study design and study population This retrospective analysis was based on a database derived from the department’s joint registry and the hospital information system.

As part of the establishment of an orthogeriatric department, all geriatric patients undergoing elective orthopedic surgery had serum levels for total protein, vitamin D, vitamin B12, and folate measured as part of their preoperative blood investigations. Geriatric patients were defined as aged above 65 years with typical geriatric comorbidity or aged above 80 years (Sieber 2007). All patients with complete preoperative laboratory findings and postoperative medical records were included. All operations took place at the Department of Orthopedic Surgery of the University Hospital Regensburg, Germany between January 2018 and December 2019.

Reoperation within 90 days after surgery, readmission within 90 days, complications, and transfusion were defined as endpoints of the study. Complications were categorized into surgical (wound healing disorder, iatrogenic fracture, mechanical complications), internal (myocardial infarction, acute heart failure, cardiac arrhythmias, pneumonia, renal failure) and other complications (deep vein thrombosis, pulmonary embolism, fall, delirium). Furthermore, complications were categorized according to the Clavien–Dindo classification (Dindo et al. 2004). This classification system ranks complications into 5 grades, based on the therapy used for correction. Any deviation from the normal postoperative course without the need for pharmacological treatment or surgical, endoscopic, and radiological intervention represents a Grade I complication. Grade II complications require specific pharmacological treatment, whereas Grade III complications result in surgical, endoscopic, or radiological intervention. Grade IV complications are defined as life-threatening events requiring intensive care management. Grade V represents the death of a patient.

Data collection Laboratory findings were extracted from the hospital information system (ORBIS; Agfa Healthcare, Mortsel, Belgium).

Further available data from our clinical information system were age, sex, length of stay, operative procedure, transfusion, transfer to intensive care unit, reoperation, readmission, and complications, as well as principal and secondary diagnoses at the time of hospitalization including corresponding ICD-10 codes. Diagnostic codes had been entered by professional clinical coders and were double-checked by physicians using information gathered from patients’ medical records.

Definitions of malnutrition and vitamin deficiencies Malnutrition was defined as total serum protein < 6.0 g/dL (Zhang et al. 2017). In a meta-analysis of biomarkers associated with malnutrition, total serum protein was found to be a useful marker of adult malnutrition (Zhang et al. 2017).

Although albumin is the most often used marker for diagnosis of malnutrition, total serum protein performs equally well and is considered less sensitive to acute stressors, which may be useful in perioperative settings (Zhang et al. 2017).

According to the recommendations of the First International Conference on Controversies in Vitamin D (Pisa, Italy, 2017), vitamin D deficiency was defined as serum 25-OH-D level < 20 ng/mL (Sempos et al. 2018). According to Shin et al.

(2017) a serum 25-OH-D level < 12 ng/mL was defined as severe deficiency. A serum cobalamin level < 200 ng/L was defined as Vitamin B12 deficiency, whereas a level of 200–300

Statistics Continuous data are presented as mean (SD). Group comparisons were performed by 2-sided t-tests. Absolute and relative frequencies were given for categorical data and compared between groups by chi-square tests. All hypotheses in the study were tested on 5% significance level.

Multivariable logistic regression analyses were performed to assess whether malnutrition or deficiency of vitamins is a significant predictor of complications and prolonged length of stay while controlling for other variables known to be associated with adverse surgical outcomes such as age, sex, and frailty (Weber et al. 2018, Meyer et al. 2020). Adjustment of covariates was performed based on considerations regarding cause–effect. The 6-step approach was used to prevent adjustment bias (Shrier and Platt 2008). The assumed cause–effect relation is shown in the appendix (Figure 1, see Supplemen-

Assessment of frailty The Hospital Frailty Risk Score (HFRS) was developed in order to provide hospitals with a frailty screening tool derived from routinely collected administrative data. As part of a complex statistical analysis of a geriatric patient cohort, Gilbert et al. (2018) could identify 109 ICD-10 codes characteristic of frailty. Dependent on how strong each ICD-10 code correlated with frailty, different points were awarded to each code and summed to a maximum possible score of 173 points (Gilbert et al. 2018). In a large validation cohort, the HFRS showed fair agreement with common frailty scales (i.e., Fried Phenotype, Rockwood’s Frailty index) and could identify patients at risk of higher 30-day mortality, prolonged length of stay, and readmission (Gilbert et al. 2018). In a previous study the authors found that the HFRS also predicts adverse events in primary total hip and knee arthroplasty (Meyer et al. 2020).

The HFRS was calculated retrospectively for each patient based on the available ICD-10 codes that were entered for the time of admission. According to literature frailty was defined as HFRS above or equal to 5 (Gilbert et al. 2018).

360

Acta Orthopaedica 2021; 92 (3): 358–363

Table 1. Characteristics of the study group (n = 599), frequencies of nutrient deficiencies, and distribution of orthopedic procedure in the study group. Values are frequency (%) unless otherwise specified Factor Age (SD) Sex (women) Comorbidities Diabetes mellitus Frailty Nutrient deficiencies Malnutrition Vitamin D deficiency Severe Vitamin D deficiency Vitamin B12 insufficiency Vitamin B12 deficiency Folate insufficiency Folate deficiency Orthopedic procedure Total hip arthroplasty Total knee arthroplasty Spine surgery Foot and ankle surgery Revision total knee arthroplasty Revision total hip arthroplasty Knee surgery Shoulder and elbow surgery Hand surgery Total shoulder arthroplasty Pelvic surgery

Frequency (%)

Frequency (%)

20

20

15

15

10

10

5

5

n (%) / mean (SD) 77 (5) 387 (65) 101 (15) 256 (43) 68 (11) 174 (29) 119 (20) 118 (20) 28 (4.7) 60 (10) 6 (1.0) 219 (37) 171 (29) 72 (12) 45 (7.5) 24 (4.0) 23 (3.8) 19 (3.2) 11 (1.8) 7 (1.2) 4 (0.7) 4 (0.7)

tary data). IBM SPSS Statistics 22 (IBM Corp, Armonk, NY, USA) was used for analysis.

Ethics, funding, data sharing, and potential conflicts of interest The study was approved by the Ethics Committee of the University Hospital Regensburg, Germany (20-1821-104). For this study no funding was received. The data that support the findings of this study is given within the article. Further data can be requested from the corresponding author. The authors have no conflicts of interest to declare that are relevant to the content of this article.

Results 599 geriatric patients underwent orthopedic surgery during the study period (Table 1 and Figure 2). The rate of malnutrition was 11% (68/599). Patients with insufficiency or deficiency of vitamin B12 and folate, respectively, were pooled for further statistical analysis (Table 1).

The malnourished cohort showed a higher reoperation rate (13% vs. 5.5%; p = 0.01) and exhibited more Clavien–Dindo IV° complications (7.4% vs. 2.4%; p = 0.03) than the cohort with normal parameters (Figure 3). Malnourished patients also had higher rates for readmission (8.8% vs. 5.5%; p = 0.3)

0

a

0

2

4

6

8

10

Total protein (g/dL)

0

0

b

20

40

60

Vitamin D (ng/mL)

Frequency (%)

Frequency (%)

30

20

15 20 10 10 5

0

c

0

200 400 600 800 1000 12001400

Vitamin B12 (ng/L)

0

0

10

20

30

Folate (ng/mL)

d

Figure 2. Distribution of total protein serum levels (a), vitamin D serum levels (b), vitamin B12 serum levels (c), and folate serum levels (d) in the study group.

Frequency (%) 20 Reoperation Wound healing disorder Clavien Dindo IV° complication

16

12

8

4

0

0–5

5–6

6–6.5

6.5–7

Total protein (g/dL)

Figure 3. In patients undergoing orthopedic surgery, the frequency of reoperation, wound healing disorders, and Clavien–Dindo IV° complications decreased as total protein serum levels increased.

as well as transfusion (8% vs. 4%; p = 0.09) and exhibited more surgical (16% vs. 9.0%; p = 0.06), other (12% vs. 7.6%; p = 0.2) and internal complications (7.4% vs. 3.6%; p = 0.1) than normally nourished patients, but the increase was not statistically significant. However, as a part of surgical complications, the frequency of wound-healing disorders was significantly higher in patients with malnutrition (7.4% vs. 1.3%; p = 0.001; Table 2 and Figure 3).

Acta Orthopaedica 2021; 92 (3): 358–363

361

Table 2. Adverse events after orthopedic surgery according to nutritional status. Values are frequency (%)

Adverse events Reoperation Readmission Surgical complications Wound healing disorder Iatrogenic fracture Mechanical complication Internal complications Other complications Clavien–Dindo IV° Transfusion

Total protein < 6.0 g/dL ≥ 6.0 g/dL n = 68 n = 531 9 (13) 6 (9) 11 (16) 5 (7) 3 (4) 3 (4) 5 (7) 8 (12) 5 (7) 6 (9)

29 (5.5) 29 (5.5) 48 (9.0) 7 (1.3) 7 (1.3) 34 (6.4) 19 (3.6) 40 (7.5) 13 (2.4) 22 (4.1)

p-value 0.01 0.3 0.06 0.001 0.06 0.5 0.1 0.2 0.03 0.09

Table 3. Adverse events after orthopedic surgery according to vitamin D status. Values are frequency (%)

Adverse events Reoperation Readmission Surgical complications Internal complications Other complications Deep vein thrombosis Pulmonary embolism Fall Delirium Clavien–Dindo IV° Transfusion

Vitamin D < 12 ng/mL ≥ 12 ng/mLl n = 119 n = 480 p-value 11 (9) 9 (8) 14 (12) 6 (5) 16 (13) 1 (1) 1 (1) 10 (8) 7 (6) 6 (5) 5 (4)

27 (5.6) 26 (5.4) 45 (9.4) 18 (3.8) 32 (6.7) 2 (0.4) 1 (0.2) 14 (2.9) 15 (3.1) 12 (2.5) 23 (4.8)

0.1 0.1 0.4 0.5 0.02 0.6 0.3 0.006 0.2 0.1 0.8

Patients with severe vitamin D deficiency exhibited more other complications (13% vs. 6.7%; p = 0.02) than patients with higher serum levels. This was due to a higher rate of falls (8.4% vs. 2.9%; p = 0.006) in the severely vitamin D deficient cohort. Severely vitamin D deficient patients also showed a higher rate of delirium (5.9% vs. 3.1%; p = 0.2; Table 3) than patients with higher serum levels, but the increase was statistically not significant.

Table 5. Multivariable analysis for total effect of malnutrition on reoperations, wound healing disorders, falls, and Clavien–Dindo IV° complications after orthopedic surgery Variable

OR (95% CI)

p-value

Reoperation Malnutrition 2.6 (1.1–6.2) 0.04 Severe Vitamin D deficiency 1.4 (0.62–3.0) 0.5 Folate deficiency 0.84 (0.29–2.4) 0.8 Vitamin B12 deficiency 0.58 (0.24–1.4) 0.2 Diabetes mellitus 1.5 (0.67–3.4) 0.3 Frailty 1.1 (1.0–1.2) 0.05 Age 0.99 (0.93–1.1) 0.8 Sex (male) 1.8 (0.89–3.5) 0.1 Wound healing disorder Malnutrition 7.1 (1.9–26) 0.003 Severe Vitamin D deficiency 1.2 (0.31–4.7) 0.8 Folate deficiency 0.31 (0.03–2.9) 0.3 Vitamin B12 deficiency 0.19 (0.02–1.6) 0.1 Diabetes mellitus 3.2 (0.87–12) 0.08 Frailty 1.1 (0.95–1.2) 0.2 Age 1.0 (0.89–1.1) 1.0 Sex (male) 0.73 (0.19–2.9) 0.7 Falls Malnutrition 0.58 (0.15–2.3) 0.4 Severe Vitamin D deficiency 3.1 (1.3–7.6) 0.01 Folate deficiency 1.1 (0.34–3.8) 0.8 Vitamin B12 deficiency 0.92 (0.32–2.7) 0.9 Diabetes mellitus 0.34 (0.08–1.5) 0.2 Frailty 1.1 (1.0–1.2) 0.009 Age 1.0 (0.94–1.1) 0.5 Sex (male) 0.48 (0.17–1.4) 0.2 Clavien–Dindo IV° complication Malnutrition 2.7 (0.80–9.0) 0.1 Severe Vitamin D deficiency 1.5 (0.48–4.4) 0.5 Folate deficiency 1.8 (0.49–6.3) 0.4 Vitamin B12 deficiency 0.46 (0.13–1.7) 0.2 Diabetes mellitus 2.8 (0.97–7.8) 0.06 Frailty 1.0 (0.91–1.2) 0.6 Age 1.0 (0.93–1.1) 0.6 Sex (male) 1.8 (0.67–4.8) 0.2 OR = odds ratio. CI = confidence interval. For assumed cause– effect relation please see Supplementary data.

Insufficiency or deficiency of Vitamin B12 or folate, respectively, neither affected reoperation rate nor rates of readmission, complications, and transfusion (Table 4).

The proportions of patients with frailty (59% vs. 40%; p = 0.004) Table 4. Adverse events after orthopedic surgery according to vitamin B12 and folate status. was higher in the malnourished Values are frequency (%) cohort. Despite its correlation to frailty, multivariable regres Vitamin B12 Folate sion analysis identified malnutri < 300 ng/mL ≥ 300 ng/mL < 4 ng/mL ≥ 4 ng/mL tion as an independent risk factor Adverse events n = 146 n = 453 p-value n = 66 n = 533 p-value for reoperation (OR 2.6; 95% CI Reoperation 7 (5) 31 (6.8) 0.4 5 (8) 33 (6.2) 0.6 1.1–6.2) and wound healing disorReadmission 9 (6) 26 (5.7) 0.8 5 (8) 30 (5.6) 0.5 ders (OR 7.1; CI 1.9–26). MoreSurgical complications 15 (10) 44 (9.7) 0.8 6 (9) 53 (9.9) 0.8 over, severe vitamin D deficiency Internal complications 7 (5) 17 (3.8) 0.6 2 (3) 22 (4.1) 0.7 Other complications 11 (8) 37 (8.2) 0.8 6 (9) 42 (7.9) 0.7 was identified as independent risk Clavien–Dindo IV° 3 (2) 15 (3.3) 0.4 4 (6) 14 (2.6) 0.1 factor for falls (OR 3.1; CI 1.3– Transfusion 6 (4) 22 (4.9) 0.7 6 (9) 22 (4.1) 0.08 7.6; Table 5).

362

Discussion The proportion of malnourished patients in our cohort was 11%. Previous studies found 4–50% of patients undergoing elective orthopedic surgery to be malnourished (Cross et al.

2014, Bohl et al. 2016, Kamath et al. 2016, Schroer et al.

2019). The majority of these studies define malnutrition as a serum albumin < 3.5 g/dL. In the context of malnutrition, serum albumin serves as standard serological biomarker. In their review of blood biomarkers associated with malnutrition, Zhang et al. (2017) promoted the use of total protein, as it is considered less sensitive to acute disease stress. As a consequence, total protein might perform better for diagnosis of malnutrition in geriatric patients, who are prone to chronic inflammation.

Regarding vitamins, we found vitamin D deficiency in 49% of patients, with 20% of patients being severely vitamin D deficient. In a retrospective analysis of over 1,000 patients aged over 70 years, who underwent hip or knee arthroplasty, Maier et al. (2016) found vitamin D deficiency in 60% of cases. In contrast Hegde et al. (2018) reported vitamin D deficiency in only 13% of patients undergoing knee arthroplasty.

Beyond selection bias, the variable proportion of vitamin D deficiency in patients undergoing joint replacement is likely due to endemic reasons as vitamin D level is affected by sun exposure, dietary supplementation, and genetic differences (Tran et al. 2017).

Insufficiency or deficiency of vitamin B12 and folate was found in 24% and 11% of patients, respectively. To the best of our knowledge, this is the first study to evaluate deficiency of vitamin B12 and folate in the context of orthopedic surgery. In a population-based cross-sectional analysis of 3,511 people aged 65 years or older, Clarke et al. (2004) found 5% of people aged 65–74 years and 10% of people aged 75 years or older to be deficient in vitamin B12 or folate.

We found a statistically significant correlation between malnutrition and postoperative adverse events. Malnourished patients were 6 times more likely to suffer from wound-healing disorders and showed a 3-fold increased rate of complications requiring ICU management. The risk of reoperation and transfusion was more than 2-fold increased for patients with malnutrition. In a retrospective analysis, Huang et al. (2013) also found a 4-fold increased complication rate for malnourished patients undergoing joint replacement surgery. Bohl et al. (2016) reported 2- to 3-fold increased rates for adverse events, such as surgical site infection, pneumonia, and readmission in patients after total joint arthroplasty. In a prospective study, the risk of unplanned ICU admission (corresponding to Clavien–Dindo IV° complications) was found to be 4 times higher for malnourished patients (Kamath et al. 2016).

Taken together, the findings of our study are consistent with existing literature and emphasize the importance of malnutrition as a risk factor for adverse events after orthopedic surgery.

Acta Orthopaedica 2021; 92 (3): 358–363

Severe vitamin D deficiency led to a 2-fold increase in falls in our study. Furthermore, multivariable analysis identified severe vitamin D deficiency as an independent risk factor for falls in patients undergoing orthopedic surgery. The role of vitamin D in fall prevention is discussed controversially among experts. In a meta-analysis of randomized controlled trials, vitamin D supplementation reduced the risk of falling by 19% (Bischoff-Ferrari et al. 2009). Conversely, in a sequential trial analysis with a risk reduction threshold of 15% conducted

by Bolland et al. (2014), the effect on falls lay within the futility boundary. Although statistically not significant, we found an almost 2-fold increased risk of postoperative delirium in patients with severe vitamin D deficiency. Interestingly, recent studies found evidence for a protective role of vitamin D in the occurrence of hospital-acquired delirium (Bowman et al.

2019). In our study, vitamin D deficiency had no impact on rates of reoperation and readmission or surgical and internal complications. In contrast, Hegde et al. (2018) found a correlation between preoperative vitamin D deficiency and reoperation rate as well as rates of different complications (i.e., thrombosis, myocardial infarction, cerebrovascular accident) after total knee arthroplasty. However, the possible coincidence of vitamin deficiency and other variables affecting postoperative outcome (i.e., hypoproteinemia, diabetes mellitus, frailty) was not taken into account in the latter study. Taken together, the results of our study provide little evidence that severe vitamin D deficiency is a risk factor for adverse events after orthopedic surgery. However, the role of vitamin D in the context of orthopedic surgery remains the subject of scientific discourse (Maier et al. 2016, Shin et al. 2017, Hegde et al. 2018).

To our knowledge this is the first study to evaluate a possible correlation between deficiency of vitamin B12 and folate and adverse events after orthopedic surgery. In contrast to hypoproteinemia and severe vitamin D deficiency, deficiency of vitamin B12 and folate had no impact on the risk of adverse events after orthopedic surgery.

This study has several limitations that are consistent with most database studies. Data acquisition was limited to the data available from the hospital information system. Due to the retrospective study design, results are susceptible to selection bias. Furthermore, other parameters with possible influence on postoperative outcome, such as BMI and psychosocial aspects, could not be captured. Despite these limitations, our study demonstrates the relevance of malnutrition in geriatric patients undergoing elective orthopedic surgery. Furthermore, to our knowledge this is the first study to research the effects of vitamin B12 and folate deficiency in orthopedic surgery.

Due to the high rate of malnutrition and its negative effects on postoperative outcome, a systematic screening of geriatric patients undergoing elective orthopedic surgery should be applied. As malnutrition is considered as a modifiable risk factor (Schroer et al. 2019), future research should focus on the effects of preoperative nutritional intervention in a prospective randomized setting.

Acta Orthopaedica 2021; 92 (3): 358–363

In conclusion, malnutrition and vitamin deficiency are common among geriatric patients undergoing elective orthopedic surgery. As malnutrition and severe vitamin D deficiency represent independent risk factors for postoperative complications, a systematic screening of at-risk patients should be applied.

Supplementary data Figure 1 is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2021.

1882092

GM, MW and MM originated the idea for the study and led on its design.

MW, TR and JG supervised the project. MW, MM, FL, FG and GM participated in the design of the study and were responsible for data acquisition.

GM, TR, MW and MM contributed to analysis and interpretation of data.

MW provided statistical consultation. MM drafted the manuscript. TR, GM, JG and MW revised the manuscript critically for important intellectual content. All authors read and approved the final version of the manuscript.

Acta thanks Katre Maasalu and Petri Virolainen for help with peer review of this study.

Bala A, Ivanov D V, Huddleston J I, Goodman S B, Maloney W J, Amanatullah D F. The cost of malnutrition in total joint arthroplasty. J Arthroplasty 2020; 35(4): 926-32 .

Bischoff-Ferrari H A, Dawson-Hughes B, Staehelin H B, Orav J E, Stuck A E, Theiler R, Wong J B, Egli A, Kiel D P, Henschkowski J. Fall prevention with supplemental and active forms of vitamin D: a meta-analysis of randomised controlled trials. BMJ 2009; 339: b3692.

Bohl D D, Shen M R, Kayupov E, Della Valle C J. Hypoalbuminemia independently predicts surgical site infection, pneumonia, length of stay, and readmission after total joint arthroplasty. J Arthroplasty 2016; 31(1): 15-21.

Bolland M J, Grey A, Gamble G D, Reid I R. Vitamin D supplementation and falls: a trial sequential meta-analysis. Lancet Diabetes Endocrinol 2014; 2(7): 573-80.

Bowman K, Jones L, Pilling L C, Delgado J, Kuchel G A, Ferrucci L, Fortinsky R H, Melzer D. Vitamin D levels and risk of delirium: a mendelian randomization study in the UK Biobank. Neurology 2019; 92(12): e1387-94.

Clarke R, Grimley Evans J, Schneede J, Nexo E, Bates C, Fletcher A, Prentice A, Johnston C, Ueland P M, Refsum H, Sherliker P, Birks J, Whitlock G, Breeze E, Scott J M. Vitamin B12 and folate deficiency in later life. Age Ageing 2004; 33(1): 34-41.

Cross M B, Yi P H, Thomas C F, Garcia J, Della Valle C J. Evaluation of malnutrition in orthopaedic surgery. J Am Acad Orthop Surg 2014; 22(3): 193-9.

Dindo D, Demartines N, Clavien P-A. Classification of surgical complica Dindo tions: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004; 240(2): 205-13.

363

Gilbert T, Neuburger J, Kraindler J, Keeble E, Smith P, Ariti C, Arora S, Street A, Parker S, Roberts H C, Bardsley M, Conroy S. Development and validation of a Hospital Frailty Risk Score focusing on older people in acute care settings using electronic hospital records: an observational study. Lancet 2018; 391(10132): 1775-82.

Green R, Datta Mitra A. Megaloblastic anemias: nutritional and other causes.

Med Clin North Am 2017; 101(2): 297-317.

Hegde V, Arshi A, Wang C, Buser Z, Wang J C, Jensen A R, Adams J S, Zeegen E N, Bernthal N M. Preoperative Vitamin D deficiency is associated with higher postoperative complication rates in total knee arthroplasty.

Orthopedics 2018; 41(4): e489-95.

Huang R, Greenky M, Kerr G J, Austin M S, Parvizi J. The effect of malnutrition on patients undergoing elective joint arthroplasty. J Arthroplasty 2013; 28(8): 21-4.

Hunt A, Harrington D, Robinson S. Vitamin B12 deficiency. BMJ 2014; 349:g5226.

Kamath A F, McAuliffe C L, Kosseim L M, Pio F, Hume E. Malnutrition in joint arthroplasty: prospective study indicates risk of unplanned ICU admission. Arch Bone Jt Surg 2016; 4(2): 128-31.

Laird E J, O’Halloran A M, Carey D, O’Connor D, Kenny R A, Molloy A M. Voluntary fortification is ineffective to maintain the vitamin B 12 and folate status of older Irish adults: evidence from the Irish Longitudinal Study on Ageing (TILDA). Br J Nutr 2018; 120(1): 111-20.

Maier G S, Maus U, Lazovic D, Horas K, Roth K E, Kurth A A. Is there an association between low serum 25-OH-D levels and the length of hospital stay in orthopaedic patients after arthroplasty? J Orthop Traumatol 2016; 17(4): 297-302.

Meyer M, Parik L, Leiß F, Renkawitz T, Grifka J, Weber M. Hospital Frailty Risk Score predicts adverse events in primary total hip and knee arthroplasty. J Arthroplasty 2020; 5(12): 3498-504 Schroer W C, LeMarr A R, Mills K, Childress A L, Morton D J, Reedy M E. 2019 Chitranjan S. Ranawat Award: Elective joint arthroplasty outcomes improve in malnourished patients with nutritional intervention: a prospective population analysis demonstrates a modifiable risk factor. Bone Joint J 2019; 101-B(7_Supple_C): 17-21.

Sempos C T, Heijboer A C, Bikle D D, Bollerslev J, Bouillon R, Brannon P M, DeLuca H F, Jones G, Munns C F, Bilezikian J P, Giustina A, Binkley N. Vitamin D assays and the definition of hypovitaminosis D: results from the First International Conference on Controversies in Vitamin D. Br J Clin Pharmacol 2018; 84(10): 2194-207.

Shin K-Y, Park K K, Moon S-H, Yang I H, Choi H-J, Lee W-S. Vitamin D deficiency adversely affects early post-operative functional outcomes after total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2017; 25(11): 3424-30.

Shrier I, Platt R W. Reducing bias through directed acyclic graphs. BMC Med Res Methodol 2008; 8:70.

Sieber C C. [The elderly patient—who is that?]. Internist 2007; 48(11): 1190, 1192-4.

Tran E Y, Uhl R L, Rosenbaum A J. Vitamin D in orthopaedics. JBJS Rev 2017; 5(8): e1.

Weber M, Craiovan B, Woerner M L, Schwarz T, Grifka J, Renkawitz T F. Predictors of outcome after primary total joint replacement. J Arthroplasty 2018; 33(2): 431-5.

Zhang Z, Pereira S, Luo M, Matheson E. Evaluation of blood biomarkers associated with risk of malnutrition in older adults: a systematic review and meta-analysis. Nutrients 2017; 9(8): 829.

364

Acta Orthopaedica 2021; 92 (3): 364–370

Comparison of histomorphometric and radiographic effects of growth guidance with tension-band devices (eight-Plate and FlexTack) in a pig model Julia SATTELBERGER 1, Hauke HILLEBRAND 1, Georg GOSHEGER 2, Andrea LAUFER 1, Adrien FROMMER 1, Sebastian APPELBAUM 3, Ahmed Abdul-Hussein ABOOD 4, Martin GOTTLIEBSEN 4, Ole RAHBEK 4, Bjarne MOLLER-MADSEN 4, Robert ROEDL 1, and Bjoern VOGT 1 1 Department

of Pediatric Orthopaedics, Deformity Reconstruction and Foot Surgery, University Hospital Muenster, Germany; 2 General Orthopaedics and Tumour Orthopaedics, University Hospital of Muenster, Germany; 3 Department of Research Methodology and Statistics in Psychology, University of Witten/Herdecke, Germany; 4 Danish Paediatric Orthopaedic Research, University Hospital Aarhus, Denmark Correspondence: andrea.laufer@ukmuenster.de Submitted Accepted

Background and purpose — Temporary hemiepiphysiodesis for growth modulation in skeletally immature patients is a long-known technique. Recently the use of tension-band devices has become popular. This study compares 2 tensionband implants (eight-Plate and FlexTack) regarding their effects on the growth plate.

Animals and methods — 12 pigs in 2 equally sized groups (A and B) were investigated. The right proximal medial tibia was treated with either eight-Plate or FlexTack.

The left tibia of the same pig was treated with the opposite implant. After 9 weeks all implants were removed. Animals in group B were then hosted for another 5 weeks. Histomorphometric analysis of the growth plate was carried out after 9 and 14 weeks, respectively. Radiographs were taken at implantation, removal, and after 14 weeks.

Results — Both tension-band devices achieved a statistically significant and clinically relevant growth inhibition, whereas the effect appeared to be more distinct after the use of FlexTack. Implant-related complications or physeal damage was not observed. After implant removal, rebound phenomenon was radiologically observed in all cases. The growth plates treated with eight-Plate showed a paradox reversal of the zonal distributions, with an increase of the proliferative zones at the previously arrested medial aspect of the physis and a decrease laterally.

Interpretation — Both eight-Plate and FlexTack proved to be appropriate devices for growth-guiding treatment.

The radiographic evaluation showed a change in angular axes after treatment with each implant, while the correction appeared to be faster with FlexTack. The paradox cartilaginous reaction observed after removal of the eight-Plate might be a histopathological correlate for rebound phenomenon.

To achieve realignment of angular deformities in skeletally immature patients, remaining bone growth can be used for growth modulation procedures to avoid extensive surgical interventions (Stevens 2007). Temporary hemiepiphysiodesis (THE) aims to mechanically inhibit growth on one side of the physis through a bridging implant. The procedure has to be performed before skeletal maturity, to maintain sufficient potential for correction while also reducing the risk of relapse of the deformity after implant removal (rebound phenomenon). Blount in 1949 described a stapling technique for THE, which provided good results in the correction of angular deformities but was later linked to complications like implant failure and physeal damage (Blount and Clarke 1949, Kanellopoulos et al. 2011, Stevens 2007). In 2007 Stevens introduced a non-locking 2-hole plate (eight-Plate [EP], Orthofix Medical Inc, Lewisville, TX, USA) based on the tension-band principle (Stevens 2007). Even though treatment with the EP seems to have an overall decreased complication rate, screw breakage is still reported rather frequently (Schroerlucke et al. 2009, Burghardt et al. 2010, 2011, Scott 2012, Vogt et al.

2016). This led to the development of a flexible staple (FlexTack [FT], Merete GmbH, Berlin, Germany). Both implants, EP and FT, achieve a tension-band effect through an extraphyseally located fulcrum of correction supposedly leading to reduced compression on the physis, thus decreasing the risk of premature closure of the physis (Vogt et al. 2016). Different from the EP, the FT is a 1-piece implant with a flexible midzone that provides dynamic bending under bone growth force.

We examined the physeal response to THE with these 2 tension-band devices and whether there is a histomorphological correlate for the extent of the blockage between FT and EP. Additionally, we sought a possible histomorphological expla-

© 2021 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 10.1080/17453674.2021.1873603

Acta Orthopaedica 2021; 92 (3): 364–370

Coronal plane

365

Sagittal plane

Figure 1. The medial proximal tibia was treated with either FT or EP, while the contralateral tibia received the opposite treatment (in this example right tibia: EP; left tibia: FT).

Implantation

Figure 3. The animals received one centrally placed either EP (upper row) or FT (lower row) on each arrested site. Both implants were inserted in the same animal. Fluoroscopic images show the anteroposterior and lateral plane of the proximal tibia immediately after implantation and before implant removal after nine weeks of bone growth.

Implant removal

Group A n=6

9 weeks of housing

Group B n=6

9 weeks of housing 5 weeks of housing

Histological and radiological analysis

Histological and radiological analysis

Figure 2. Chronological overview of the two study groups (group A: overall study time 9 weeks; group B: overall study time 14 weeks).

nation for the excessive unilateral bone growth frequently occurring after implant removal, which might be linked to the incidence of rebound phenomenon.

Animals and methods Study design, animal model, and sample size A randomized paired setup was chosen and a total time period of 14 weeks with histomorphometric analysis after 9 and 14 weeks was set. In order to achieve valid results, the smallest number of animals needed was determined through a statistical power analysis. 12 skeletally immature domestic female pigs were assigned into 2 groups (n = 6). The animal itself served as its control. The average weight of the pigs was 28 (24–33) kg at the time of first surgery. In both groups, each animal received an implantation of either EP or FT on the medial proximal tibia of the right hind leg. The opposite treatment was then performed on the medial proximal tibia of the left hind leg, so both implants were inserted in the same animal (Figure 1). The chosen implant size depended on the size of the proximal tibia.

After 9 weeks of housing under the same conditions, implants were removed in both groups. In group A, the histomorphological analysis was performed immediately after implant removal. In group B, animals were housed for another 5 weeks to investigate delayed reactions after physeal release.

Thus, the histomorphological analysis in group B was performed after 14 weeks (Figure 2).

Implants 1 eight-Plate (titanium alloy; 12 mm plate and two 24 mm fully threaded cannulated non-locking screws) and 1 FlexTack (titanium alloy [TiA16V4 ELI]; crossbar 25 mm, epiphysial leg 25 mm, metaphyseal leg 23 mm) were implanted in each pig. The EP is fixated with 2 screws that can be positioned in varying angles and diverge up to 30° during growth. The FT is a 1-piece implant with cannulated legs. Both devices are fixated extraperiosteally to avoid the development of a bone bridge, which might lead to untimely growth arrest. In both implants, the implantation technique is minimally invasive and K-wire guided.

Surgical technique General anesthesia was maintained with intravenous propofol (10 mg/kg/h) and fentanyl (25 µg/kg/h) for all surgical interventions on living and ventilated animals. Protocols established by the research group provided the basis for anesthetic techniques, postoperative assessment, and pain care management (Gottliebsen et al. 2013a, Hillebrand et al. 2020).

A longitudinal incision at the level of the physis was made, which allowed the extra-periosteal insertion of the devices under fluoroscopic control. 1 centrally placed implant (EP or FT) was positioned in the epiphysis and metaphysis. Digital fluoroscopic images were obtained for radiological analysis (Figure 3). After closing the wound, local anesthetics were applied.

After completing the last study procedure, each animal received a lethal dose of pentobarbital (0.5 mL/kg) through intravenous access. Implants were extracted, following the manufacturer’s guidelines. In the case of FT, first the cannulated staple shanks were separated from the bone with a guided U-chisel. Then the extraction instrument was fastened to the FT and removed with the help of a sliding hammer.

After implant removal, both tibiae were harvested and immediately stored at –25°C until further processing.

366

Acta Orthopaedica 2021; 92 (3): 364–370

Figure 4. Schematic drawing of standardized bone sample extraction.

Figure 6. Digital histological images in HE staining at 100x magnification showing zone of reserve (R), proliferative zone (P), and hypertrophy zone (H).

Figure 5. Example of medial bone sample after embedding in MMA.

Histological preparation All samples were obtained following the same method. 2 lines (dsag and dcor) connecting the outer edges of the tibial plateau were drawn and their point of intersection was determined in the transverse plane. With reference to this point and same distance from each other, 2 square samples (medial and lateral) were taken in a line parallel to the sagittal tibia axis using a sample preparation saw. Each sample was sized 3 cm x 1 cm x 1 cm and consisted of the area of interest with joint cartilage, epiphyseal bone, physeal cartilage, and metaphyseal bone (Figure 4).

Before embedding in cold methyl methacrylate, the removed bone samples were dehydrated in graded ethanol (70–100%).

From the outer rim of each biopsy, 10 μm-thick coronal sections were cut on a microtome (Polycut E, Reichert-Jung/ Leica Microsystems, Wetzlar, Germany). 6 sections were removed from one level and stained with HE (Figure 5).

Histomorphometric analysis The growth plate is divided into 3 different fractions of chondrocyte layers from epiphysis to diaphysis: zone of reserve (R), zone of proliferation (P), and zone of maturation and hypertrophy (H) (Figure 6) (Nuttall et al. 1999, Byers et al.

2000). Chondrocytes are located in the reverse zone at the epiphyseal end. In the proliferative zone chondrocytes of constant size start aligning in columns, before enlarging in the hypertrophic zone. In addition to these 3 regions, areas with unorganized cartilage tissue (loss of normal columnar appear-

ance of the cell distribution) can be found (Nuttall et al. 1999, Byers et al. 2000, Gottliebsen et al. 2013b).

Normal proportions of the cartilage zones and physiological joint architecture were defined by performing sham surgery on 1 animal in each study group. In these cases the implants were inserted and removed again before closure of the skin.

A microscope (BX50, Olympus, Shinjuku, Japan) was attached to a computer to transmit the microscope field to the computer monitor via video camera. Using software for stereology (newCAST, Visiopharm, Hoersholm, Denmark), on each section, all visible cartilage was outlined as region of interest (ROI) and highlighted. Point counting was used to define the fractions of chondrocyte layers (Gundersen et al.

1988).

Radiographic evaluation Anteroposterior (AP) and lateral radiographs of the knee joint were taken immediately before implantation and prior to implant removal after 9 weeks in both group A and B. In group B additionally radiographs were taken after 14 weeks, before the histomorphological samples were harvested.

To investigate changes in the tibial angulation, the medial proximal tibial angle (MPTA) and the articular line-diaphysis angle (ALDA) were analyzed. The MPTA is defined as the medial angle between the mechanical axis of the tibia and the tangent along the articular surface of the tibial plateau (joint line). The ALDA is measured between the anatomical axis of the tibia and the joint line, based on the description of Aykut et al. (2005).

Loss of correction was investigated in group B and defined as the difference in MPTA and ALDA at the time of implant removal (9 weeks after implantation) and 5 weeks after implant removal.

The effectiveness of the respective implant was determined through the device angle (DA; angle between the epiphyseal and metaphyseal screw or shank, respectively), which was

Acta Orthopaedica 2021; 92 (3): 364–370

367

Zonal distribution (%) 100

b

b

a

a

c

c

a

a

a

a

80

60

d 40

a

a

d

e

e

20 Zone of hypertrophy Zone of reserve

0

M L eight-Plate

M L FlexTack

Zone of proliferation Unorganized cartilage

M L eight-Plate

M L FlexTack

Removal (9 weeks) Follow-up (14 weeks)

Figure 7. MPTA, ALDA, and DA immediately after implantation of EP (upper row) and FT (lower row).

measured on the radiographs taken at implantation and after 9 weeks, prior to implant removal (Figure 7).

Figure 8. Left side: Zonal distributions after 9 weeks of THE with EP and FT. Right side: Zonal distributions 5 weeks after implant removal.

The red horizontal lines show the physiological zone distribution of the porcine growth plate, determined through sham surgery. Significance to the same zone in the same implant (a–c), and between implants (d) in the same period as well as significance to the same zone and implant between the periods (e) is shown: a p < 0.05, b p < 0.01, c p < 0.001, d p < 0.05, e p < 0.05.

Reproducibility The reliability of the histomorphometric analysis has previously been tested by the research group estimating the coefficients of variation (CV) for determination of the zonal distribution of the growth plate. The CV values were (very) good for all zones, ranging from 7% to 11% (Gottliebsen et al. 2013b).

The reliability of the radiographic measurements was tested by 2 raters (JS, BV). Rater 1 performed all the measurements just as 56 double-measurements, and rater 2 performed an additional 56 measurements. This correlates with a doublemeasurement rate of 33.3%. Intraclass correlation coefficients (ICC) were used to determine the inter-rater reliability. A 2-way mixed effects model using absolute agreement definition was used. The estimated ICC values were excellent for all angles, ranging from 0.927, 95% confidence interval (CI) (0.808– 0.972) to 0.996 (CI 0.989–0.999), p < 0.001. Regardless of the measured angle, the overall ICC value for all 56 measurements was 0.999 (CI 0.998–0.999), p < 0.001. The analysis was performed with SPSS 25 (IBM Corp, Armonk, NY, USA).

observed for 14 weeks a 2-sample Welch’s t-test was used. A p-value < 0.05 was considered to be statistically significant.

Analysis was performed using R language (R version 3.6.2 [2019-12-12]) (R Core Team 2019).

Power analysis and statistics EP and FT were implanted on the medial side of the leg. Compared with the lateral, a reduction of the growth plate was expected. A mean difference of 125 units and a standard deviation (SD) of 85 units, a significance level of α = 5% and a test power (1–β) of at least 80% results was assumed. As a result, 6 animals were needed to detect a statistically significant difference by a t-test for paired samples.

Measurements on the same animals were compared using a paired t-test. For the comparisons of individuals between those who were observed for 9 weeks and those who were

All 12 animals tolerated the interventions well and did not show either infections or any abnormalities in eating or movement behavior.

Ethics, funding, and potential conflicts of interest The Danish Animal Research Guidelines and the European Directive 2010/63/EU for animal experiments formed the basis for the experimental protocol (Danish Animal Research Guidelines, European Directive). The Danish Animal Experiment Committee agreed with the research proposal (file number 2015-15-0201-00761).

This study was fully financed by the research funds of the University Hospital of Muenster, Germany. RR is a paid consultant of Merete Medical GmbH. All other authors declare no conflicts of interest.

Results

Histomorphometric results In comparison with the physiological structure of the growth plate, the zonal distribution of the lateral physis remained unchanged after 9 weeks of THE with either EP or FT. Comparing the medial with the lateral part of the physis after 9 weeks, medially a reduction of the proliferative zone was detected after treatment with EP (p = 0.04), while the reserve

368

Acta Orthopaedica 2021; 92 (3): 364–370

MPTA

(°)

ALDA

100

and DA

(°) 50

e

Angular change (°) 30 MPTA ALDA DA

d d

80

40

60

30

40

20

20

10

0 c

20 a a a

b

a a

b

a

b

a

10 –10

a a a

0

0 eight-Plate FlexTack eight-Plate FlexTack eight-Plate FlexTack

Implantation (0 weeks)

Removal (9 weeks)

–5

Follow-up (14 weeks)

Figure 9. Both EP and FT achieved a significant reduction of MPTA and a significant increase of ALDA and DA after 9 weeks of THE. After 5 weeks both implants showed a significant loss of correction of MPTA and ALDA. The overall reduction of MPTA and increase of ALDA after 14 weeks were not significant. Significance to the same angle in the same implant during implantation (a), removal (b–c), and between implants (d–e) is shown: a p < 0.001, b p < 0.01, c p < 0.001, d p < 0.01, e p < 0.001.

zone increased (p = 0.01). THE with FT led to a reduction in both proliferative (p = 0.03) and hypertrophic zone (p < 0.001), while the reserve zone also increased (p < 0.001). THE with EP led to a greater decrease of the proliferative zone on the medial side of the physis than THE with FT (p = 0.02). In total, EP and FT achieved an equal reduction of the proliferative and hypertrophic zone (Figure 8).

5 weeks after removal of the EP, the proliferative zone of the medial physis showed an increase (p = 0.02), while laterally a reduction of the proliferative zone was observed (reversal of distributions). The proportion of the proliferative zone medially was greater than laterally (p = 0.05). The distribution of the hypertrophic zone remained mostly unchanged (Figure 8).

After removal of the FT, the medial physis showed only slight changes of the proliferative zone with increase of the hypertrophic zone, while in the lateral growth plate a marginal reduction of proliferative and hypertrophic zones was observed. The zonal distributions of the proliferative and hypertrophy zone were almost equal medially and laterally, while the proportion of the reserve zone medially was greater (p = 0.04) (Figure 8).

Radiological results After 9 weeks of THE, both implants, EP and FT, showed a reduction of the MPTA (p < 0.001) (Figure 9). The FT overall achieved a greater decrease than the EP (p = 0.02), with an average reduction of the MPTA of –11° using the FT and –8° using the EP (Figure 10). The ALDA increased after treatment with both EP and FT (p < 0.001) (Figure 9). The FT achieved a greater increase of the ALDA than the EP (p = 0.02), with an increase of the ALDA of an average +13° using the FT and

–20

eight-Plate FlexTack eight-Plate FlexTack eight-Plate FlexTack

Implantation to removal

Removal to follow-up

Implantation to follow-up

Figure 10. After 9 weeks of THE, the FT achieved a significantly greater decrease of the MPTA and a significantly greater increase of the ALDA and DA. 5 weeks after implant removal, the increase of the MPTA and decrease of the ALDA showed no significant differences between the 2 implants. The overall reduction of MPTA and increase of ALDA at the time of the last follow-up after 14 weeks also proved not to be significantly different between EP and FT. Significance to the same angle in the same period between the implants (a–b) is shown: a p < 0.05, b p < 0.001.

Figure 11. Decrease of the MPTA and increase of ALDA and DA after 9 weeks of THE with EP (upper row) and FT (lower row).

+10° using the EP (Figure 10). EP as well as FT showed an increase of the DA within 9 weeks of THE (p < 0.001) (Figures 9 and 11). The FT showed a greater change of the DA than the EP (p < 0.001), with an average divergence of +22° in the FT and +18° in the EP (Figure 10).

5 weeks after implant removal, both implants showed a loss of correction with a statistically significant increase of the MPTA and a significant decrease of the ALDA (Figure 9). No significant differences in the amount of loss of correction were

Acta Orthopaedica 2021; 92 (3): 364–370

observed between FT and EP (Figure 10). The average loss of correction of the MPTA was +7° using the EP and +9° using the FT. The average loss of correction of the ALDA was –7° using the EP and –10° using the FT (Figure 10).

At the final follow-up after 14 weeks, the EP showed an overall reduction of the MPTA of on average –2° and an overall increase of the ALDA of on average +3° (Figure 10). The FT showed an overall decrease of the MPTA of on average –4° and an increase of the ALDA of on average +5° (Figure 10).

The overall reduction of MPTA and increase of ALDA after 14 weeks was not significant in either implant (Figures 9 and 10).

Discussion Even though the EP has become a widely popular device for growth guiding treatment, hardware failure—in particular breakage of the metaphyseal screw—is a rather frequent complication (Schroerlucke et al. 2009, Burghardt et al. 2010, Burghardt et al. 2011, Scott 2012, Vogt et al. 2016). A possible explanation for screw failure might be the maximal divergence of the screws in the correction of severe angular deformities. Under sustained growth the plate consequently bends in the other direction, which might lead to blockage of the screws within the plate and subsequent screw breakage. Furthermore, Burghardt et al. stated that the EP has a delayed onset of correction in comparison with fixed angle staples (Burghardt et al. 2011). As the screws of the EP initially show a mechanical slackness within the 2-hole plate, correction is induced only after a certain degree of angulation and sufficient bone purchase is achieved (Burghardt et al. 2011). Eltayeby et al. (2019), on the other hand, did not detect a significant correlation between the initial screw angle and the average rate of correction; they stated that a wider initial screw angle did not result in a faster correction rate. Vogt et al. (2016) reported an earlier correction start and faster correction speed of the FT in comparison with the EP. They attributed this to the 1-piece design of the FT, which presumably leads to immediate correction commencement and prevents breakage of the implant.

In our study, neither EP nor FT showed implant failures.

The radiological evaluation revealed an adequate alteration of the angular axes with both devices. However, the FT achieved a faster correction than the EP after 9 weeks of THE. The higher mean DA in THE with FT was related to greater angular correction. In group A, both implants showed a statistically significant reduction in the proportion of the proliferative zone of the arrested medial physis in comparison with the lateral physis and with the physiological zonal distribution. The FT also achieved a significant reduction of the hypertrophy zone.

The reduction in the proportions of hypertrophic and proliferative zones of the medial physis seemed to be more pronounced after THE with FT, which might be a histomorphological correlation to the radiological findings.

369

Deformity recurrence occurring after implant removal before skeletal maturity is a common problem in growth modulation treatment (Stevens et al. 1999, Burghardt and Herzenberg 2010, Farr et al. 2018, Leveille et al. 2019). The likelihood of relapse of the deformity is reportedly higher if THE is performed at a young age, in the correction of secondary axis deviations or an angular axis deviation over 20°, and in the case of a fast correction rate (Stevens and Klatt 2008, Park et al. 2016, Farr et al. 2018, Leveille et al. 2019). Nonetheless, the occurrence of rebound phenomenon shows a very variable incidence, as the physeal response after release of the epiphysiodesis is hardly predictable (Aykut et al. 2005, Leveille et al.

2019). Widening of the formerly arrested part of the physis after release of the physeal compression has been described by several authors, and has been linked to accelerated unilateral growth with consecutive relapse of the initial deformity (Gottliebsen et al. 2013b), Corominas-Frances et al. 2015, Ding et al. 2018).

In our study, 5 weeks after implant removal restoration of the physiological cartilaginous zone distributions appeared to be slightly prolonged after THE with FT. This might indicate a minor risk for the occurrence of a rebound phenomenon. In comparison, after THE with EP the return to the physiological proportions of the medial physis appeared to take place more rapidly, with increase in the proportion of the proliferative cartilage in the formerly arrested physeal zone. The lateral, previously not blocked side of the growth plate showed a paradox reduction of the proliferative zone. This might be a possible histological correlate for the occurrence of rebound phenomenon. Radiologically, however, a rebound—or rather return to the physiological limb alignment—was observed in all cases and was similar between EP and FT. This study has several limitations. Apart from the small cohort and short period of observation, the results of THE in a porcine model may differ from those of THE in skeletally immature children. Furthermore, as previously described by Corominas-Frances et al. (2015), THE was performed in tibiae initially presenting physiological axes, in which an angular deformity was produced. In our study, what was interpreted as a rebound deformity may in fact have been bone remodeling to regain physiological limb alignment. Additionally, it is unclear whether the fact that the contralateral leg of the same pig has been used as a control may have influenced the results.

Conclusion Both tension band devices investigated achieved THE in the porcine model. A substantial inhibition of the growth plate was obtained in all cases without causing physeal damage.

Neither EP nor FT showed any implant-related complications.

The radiographic evaluation revealed a substantial change of angular axes and DA in THE with both implants, while the correction appeared to be faster with FT. In all cases a rebound was observed 5 weeks after implant removal. Histomorphologically, the growth inhibition of the medial physis

370

seemed to be slightly more pronounced after treatment with FT, while showing prolonged restoration of the physiological zonal distributions after implant removal. On the other hand, after removal of the EP a paradox cartilaginous reaction was observed. These observations could be possible histomorphological correlates for the rebound phenomenon.

JS, HH, AL, and BV contributed to acquisition, analysis, and interpretation of data, drafting of the paper, and approval of the submitted version. BV, GG, BMM, SA, RR, MG, AF, and AL contributed to research design, interpretation of data, and revising the paper and approval of the submitted and final versions. JS, HH, AAA, OR, and BV carried out the experiment and contributed to acquisition and analysis of data.

The authors would like to thank the Department of Paediatric Orthopaedics at the University of Aarhus for the opportunity to conduct this project at their facilities. Furthermore, they would like to thank Anette Baatrup for the opportunity to perform the histomorphometric procedures at her research laboratory and for her support on this project. Thanks are also due to Univ.-Prof. Dr. Thomas Ostermann for his expertise in statistics as Head of the Department of Research Methodology and Statistics in Psychology at Witten/Herdecke University. The authors would also like to thank Orthofix and Merete for providing instruments and implants.

Acta thanks Henrik Lauge-Pedersen and Richard Marsell for help with peer review of this study.

Aykut U S, Yazici M, Kandemir U, et al. The effect of temporary hemiepiphyseal stapling on the growth plate: a radiologic and immunohistochemical study in rabbits. J Pediatr Orthop 2005; 25(3): 336-41.

Blount W P, Clarke G R. Control of bone growth by epiphyseal stapling: a preliminary report. J Bone Joint Surg Am 1949; 31A(3): 464-78.

Burghardt R D, Herzenberg J E. Temporary hemiepiphysiodesis with the eight-Plate for angular deformities: mid-term results. J Orthop Sci 2010; 15(5): 699-704.

Burghardt R D, Kanellopoulos A D, Herzenberg J E. Hemiepiphyseal arrest in a porcine model. J Pediatr Orthop 2011; 31(4): e25-9.

Burghardt R D, Specht S C, Herzenberg J E. Mechanical failures of eightplate-guided growth system for temporary hemiepiphysiodesis. J Pediatr Orthop 2010; 30(6): 594-7.

Byers S, Moore A J, Byard R W, et al. Quantitative histomorphometric analysis of the human growth plate from birth to adolescence. Bone 2000; 27(4): 495-501.

Corominas-Frances L, Sanpera I, Saus-Sarrias C, et al. Rebound growth after hemiepiphysiodesis: an animal-based experimental study of incidence and chronology. Bone Joint J 2015; 97-B(6): 862-8.

Danish Animal Research Guidelines. Available from: https://www.retsinformation.dk/Forms/R0710.aspx?id = 145380.

Acta Orthopaedica 2021; 92 (3): 364–370

Ding J, He J, Zhang Z Q, et al. Effect of hemiepiphysiodesis on the growth plate: the histopathological changes and mechanism exploration of recurrence in mini pig model. Biomed Res Int 2018; 2018: 6348171.

Eltayeby H H, Iobst C A, Herzenberg J E. Hemiepiphysiodesis using tension band plates: does the initial screw angle influence the rate of correction? J Child Orthop 2019; 13(1): 62-6.

European Directive. 2010/63/EU for animal experiments. Available from: http://eara.eu/en/animal-research/eu-animal-research-law-directive-2010-63/. 2010/63/EU for animal experiments.

Farr S, Alrabai H M, Meizer E, et al. Rebound of frontal plane malalignment after tension band plating. J Pediatr Orthop 2018; 38(7): 365-9.

Gottliebsen M, Moller-Madsen B, Stodkilde-Jorgensen H, et al. Controlled longitudinal bone growth by temporary tension band plating: an experimental study. Bone Joint J 2013a; 95-B(6): 855-60.

Gottliebsen M, Rahbek O, Poulsen H D, et al. Similar growth plate morphology in stapling and tension band plating hemiepiphysiodesis: a porcine experimental histomorphometric study. J Orthop Res 2013b; 31(4): 574-9.

Gundersen H J, Bagger P, Bendtsen T F, et al. The new stereological tools: dissector, fractionator, nucleator and point sampled intercepts and their use in pathological research and diagnosis. APMIS 1988; 96(10): 857-81.

Hillebrand H, Sattelberger J, Gosheger G, et al. Comparison of temporary epiphysiodesis with RigidTacks and Blount-Staples in a porcine animal model using magnetic resonance imaging. J Orthop Res 2020; 38(5): 946-53.

Kanellopoulos A D, Mavrogenis A F, Dovris D, et al. Temporary hemiepiphysiodesis with Blount staples and eight-Plates in pigs. Orthopedics 2011; 34(4).

Leveille L A, Razi O, Johnston C E. Rebound deformity after growth modulation in patients with coronal plane angular deformities about the knee: who gets it and how much? J Pediatr Orthop 2019; 39(7): 353-8.

Nuttall J D, Brumfield L K, Fazzalari N L, et al. Histomorphometric analysis of the tibial growth plate in a feline model of mucopolysaccharidosis type VI. Calcif Tissue Int 1999; 65(1): 47-52.

Park S S, Kang S, Kim J Y. Prediction of rebound phenomenon after removal of hemiepiphyseal staples in patients with idiopathic genu valgum deformity. Bone Joint J 2016; 98-B(9): 1270-5.

R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria; 2019. Available from: https://www.R-project.org/.

Schroerlucke S, Bertrand S, Clapp J, et al. Failure of Orthofix eight-Plate for the treatment of Blount disease. J Pediatr Orthop 2009; 29(1): 57-60.

Scott A C. Treatment of infantile Blount disease with lateral tension band plating. J Pediatr Orthop 2012; 32(1): 29-34.

Stevens P M. Guided growth for angular correction: a preliminary series using a tension band plate. J Pediatr Orthop 2007; 27(3): 253-9.

Stevens P M, Klatt J B. Guided growth for pathological physes: radiographic improvement during realignment. J Pediatr Orthop 2008; 28(6): 632-9.

Stevens P M, Maguire M, Dales M D, et al. Physeal stapling for idiopathic genu valgum. J Pediatr Orthop 1999; 19(5): 645-9.

Vogt B, Kleine-König M-T, Gosheger G, et al. FlexTackTM and RigidTackTM: new devices for correction of angular deformities and leg length discrepancies by temporary epiphysiodesis. J Child Orthop 2016; 10: 17-8.

3/21 ACTA ORTHOPAEDICA

Medical

IMPROVE THE CHANCES

REDUCE RISK FOR INFECTION Reduction of infection risk* using dual antibiotic-loaded bone cement in high risk patients

in aseptic revision TKA * as reported in study results

10453

www.heraeus-medical.com

in fractured neck of femur

Vol. 92, No. 3, 2021 (pp. 249–370)

34 % 69 % 57 %

in primary hip & knee arthroplasty

Volume 92, Number 3, June 2021

COVER.indd 1

26-05-2021 22:08:39