Acta Orthopaedica Vol. 90 Issue 5, October 2019 by Acta

5/19 ACTA ORTHOPAEDICA

Element of success in joint replacement

www.heraeus-medical.com

Vol. 90, No. , 2019 (pp. 411–506)

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

Volume 90, Number 5, October 2019

IORT_90_05_COVER.indd 1

16-08-2019 20:31:05

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

E DITORIAL O F FICE

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

EDITOR

THE FOUNDATION BOARD OF

Anders Rydholm Lund, Sweden

THE NORDIC O RTHOPAEDIC F EDERATION AND A CTA O RTHOPAEDICA

DEPUTY EDITOR

Peter A Frandsen Odense, Denmark CO-EDITORS

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

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

WEB EDITOR

Magnus Tägil Lund, Sweden S TATISTICAL EDITOR

Jonas Ranstam Lund, Sweden P RODUCTION MANAGER

Kaj Knutson Lund, Sweden

Vol. 90, No. 5, 2019

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

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

Annual Institutional Subscription, Volume 90, 2019

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

$1,218

£753

€976

The subscription fee purchases an online subscription. The price includes access to current content and back issues to January 1997 (if available). Printed copies of the journal are provided on request as a free supplementary service accompanying an online subscription. Supplements to the journal are also included in the subscription price. For more information, visit the journal’s website: http://www.tandfonline.com/IORT Manuscripts should be uploaded at http://www.manuscriptmanager.com/ao/ for further handling at: Acta Orthopaedica Editorial Office, Department of Orthopaedics, Lund University Hospital, SE-221 85 Lund, Sweden Correspondance concerning copyright and permissions should be sent to: Maria Montzka Håkan Pårup P.O. Box 3255, SE-103 65 Stockholm, Sweden, Tel: +46 (0)8 440 80 40. Fax: +46 (0)8 440 80 50. maria.montzka@informa.com E-mail: hakan.parup@informa.com. Ordering information: Please contact your local Customer Service Department to take out a subscription to the Journal: USA, Canada: Taylor & Francis, Inc., 530 Walnut Street, Suite 850, Philadelphia, PA 19106, USA. Tel: +1 800 354 1420; Fax: +1 215 207 0050. UK/ Europe/Rest of World: T&F Customer Services, Informa UK Ltd, Sheepen Place, Colchester, Essex, CO3 3LP, United Kingdom. Tel: +44 (0) 20 7017 5544; Fax: +44 (0) 20 7017 5198; Email: subscriptions@tandf.co.uk Dollar rates apply to all subscribers outside of Europe. Euro rates apply to all subscribers in Europe except the UK and Republic of Ireland. If you are unsure which applies, contact Customer Services. All subscriptions are payable in advance and all rates include postage. Journals are sent by air to the USA, Canada, Mexico, India, Japan and Australasia. Subscriptions are entered on an annual basis, i.e., January to December. Payment may be made by sterling check, US dollar check, euro check, international money order, National Giro, or credit card (Amex, Visa and Mastercard). Back issues: Taylor & Francis retains a two-year back issue stock of journals. Older volumes are held by our official stockists to whom all orders and enquiries should be addressed: Periodicals Service Company, 351 Fairview Ave., Suite 300, Hudson, New York 12534, USA. Tel: +1 518 537 4700; fax: +1 518 537 5899; e-mail: psc@ periodicals.com.

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

Copyright © 2019 The Author(s). Published by Taylor & Francis on behalf of the Nordic Orthopedic Federation. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Non-Commercial License (https://creativecommons.org/licenses/by-nc/3.0 . Informa UK Limited, trading as Taylor & Francis Group makes every effort to ensure the accuracy of all the information (the “Content”) contained in its publications. However, Informa UK Limited, trading as Taylor & Francis Group, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Informa UK Limited, trading as Taylor & Francis Group. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Informa UK Limited, trading as Taylor & Francis Group shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. Terms & Conditions of access and use can be found at http://www.tandfonline. com/page/terms-and-conditions Indexed/abstracted in: Allied and Complementary Medicine Library (Amed); ASCA (Automatic Subject Citation Alert); Biological Abstracts; Chemical Abstracts; Cumulative Index to Nursing and Allied Health Literature(CINAHL); Current Advances in Ecological and Environmental Sciences; Current Contents/Clinical Medicine; Current Contents/Life Sciences; Developmental Medicine and Child Neurology; Energy Research Abstracts; EMBASE/ Excerpta Medica; Faxon Finder; Focus On: Sports Science & Medicine; Health Planning and Administration; Index Medicus/MEDLINE; Index to Dental Literature; Index Veterinarius; INIS Atomindex; Medical Documentation Service; Nuclear Science Abstracts (Ceased); Periodicals Scanned and Abstracted. Life Sciences Collection; Research Alert; Science Citation Index; SciSearch; SportSearch; Uncover Veterinary Bulletin. Printed in England by Henry Ling

Acta Orthopaedica

ISSN 1745-3674

Vol. 90, No. 5, October 2019 Hip Contemporary posterior surgical approach in total hip replacement: still more reoperations due to dislocation compared with direct lateral approach? An observational study of the Swedish Hip Arthroplasty Register including 156,979 hips Registry study on failure incidence in 1,127 revised hip implants with stem trunnion re-use after 10 years of follow-up: limited influence of an adapter sleeve Uncemented or cemented revision stems? Analysis of 2,296 firsttime hip revision arthroplasties performed due to aseptic loosening, reported to the Swedish Hip Arthroplasty Register Reduced periprosthetic fracture rate when changing from a tapered polished stem to an anatomical stem for cemented hip arthroplasty: an observational prospective cohort study with a follow-up of 2 years No effect of delivery on total hip replacement survival: a nationwide register study in Finland Low complication rate after same-day total hip arthroplasty: a retrospective, single-center cohort study in 116 procedures Cementing of the hip arthroplasty stem increases load-to-failure force: a cadaveric study Sequence of 305,996 total hip and knee arthroplasties in patients undergoing operations on more than 1 joint Optimization of the empirical antibiotic choice during the treatment of acute prosthetic joint infections: a retrospective analysis of 91 patients Knee Uncemented monoblock trabecular metal posterior stabilized high-flex total knee arthroplasty: similar pattern of migration to the cruciate-retaining design — a prospective radiostereometric analysis (RSA) and clinical evaluation of 40 patients (49 knees) 60 years or younger with 9 years’ follow-up Primary constrained and hinged total knee arthroplasty: 2- and 5-year revision risk compared with unconstrained total knee arthroplasty: a report on 401 cases from the Norwegian Arthroplasty Register 1994–2017 Failure modes of patellofemoral arthroplasty—registries vs. clinical studies: a systematic review Bone remodeling of the proximal tibia after uncemented total knee arthroplasty: secondary endpoints analyzed from a randomized trial comparing monoblock and modular tibia trays—2 year follow-up of 53 cases Manipulation under anesthesia after primary knee arthroplasty in Sweden: incidence, patient characteristics and risk of revision Shoulder Increased use of total shoulder arthroplasty for osteoarthritis and improved patient-reported outcome in Denmark, 2006–2015: a nationwide cohort study from the Danish Shoulder Arthroplasty Registry Children, hip dislocation Primary surgery to prevent hip dislocation in children with cerebral palsy in Sweden: a minimum 5-year follow-up by the national surveillance program (CPUP) Case report Recurrent arthrocele and sterile sinus tract formation due to ceramic wear as a differential diagnosis of periprosthetic joint infection — a case report

411

O Skoogh, G Tsikandylakis, M Mohaddes, S Nemes, D Odin, P Grant, and O Rolfson

417

S Affatato, M Cosentino, F Castagnini, and B Bordini

421

Y Tyson, O Rolfson, J Kärrholm, N P Hailer, and M Mohaddes

427

J Mohammed, S Mukka, C-J Hedbeck, G Chammout, M Gordon, and O Sköldenberg

433

I Kuitunen, E T Skyttä, M Artama, H Huhtala, and A Eskelinen

439

M N Madsen, M L Kirkegaard, M Laursen, J R Larsen, M F Pedersen, B Skovgaard, T Prynø, and L R Mikkelsen A Klasan, M Bäumlein, C Bliemel, S E Putnis, T Neri, M D Schofer, and T J Heyse P Espinosa, R J Weiss, O Robertsson, and J Kärrholm

445 450 455

J H J van Erp, A C Heineken, R J A van Wensen, R W T M van Kempen, J G E Hendriks, M Wegdam-Blans, J M Fonville, and M C van der Steen

460

R Wojtowicz, A Henricson, K G Nilsson, and S Crnalic

467

M Badawy, A M Fenstad, and O Furnes

473

N B Bendixen, P W Eskelund, and A Odgaard

479

M Rathsach Andersen, N Winther, T Lind, H M Schrøder, and M M Petersen

484

H Thorsteinsson, M Hedström, O Robertsson, N Lundin, and A W-Dahl

489

J V Rasmussen, A Amundsen, A K B Sørensen, T W Klausen, J Jakobsen, S L Jensen, and B S Olsen

495

N Kiapekos, E Broström, G Hägglund, and P Åstrand

501

N M Jandl, T Rolvien, D Gätjen, A Jonitz-Heincke, A Springer, V Krenn, R Bader, and W Rüther

Correspondence An infrapatellar nerve block reduces knee pain in patients with chronic anterior knee pain after tibial nailing: a randomized, placebo-controlled trial in 34 patients Information to authors (see http://www.actaorthop.org/)

505

P Bhakta, H M R Karim, B Oâ&#x20AC;&#x2122;Brien, and M C Vassallo versus M S Leliveld, S J M Kamphuis, and M H J Verhofstad

Acta Orthopaedica 2019; 90 (5): 411–416

411

Contemporary posterior surgical approach in total hip replacement: still more reoperations due to dislocation compared with direct lateral approach? An observational study of the Swedish Hip Arthroplasty Register including 156,979 hips Oscar SKOOGH 1, Georgios TSIKANDYLAKIS 1–3, Maziar MOHADDES 1–3, Szilard NEMES 2,3, Daniel ODIN 2, Peter GRANT 1, and Ola ROLFSON 1–3 1 Department

of Orthopaedics, Sahlgrenska University Hospital; 2 Swedish Hip Arthroplasty Register; 3 Department of Orthopaedics, Institute of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg, Sweden Correspondence: ola.rolfson@vgregion.se Submitted 2018-12-03. Accepted 2019-04-11.

Background and purpose — The direct lateral approach (DLA) and the posterior approach (PA) are the most common surgical approaches in total hip replacement (THR) in Sweden. We investigated how the relationship between surgical approach and risk of reoperation due to dislocation has evolved over time. Patients and methods — Data were extracted from the Swedish Hip Arthroplasty Register from 1999 to 2014. We selected all THRs due to osteoarthritis with head sizes 28, 32, and 36 mm that were performed with either the DLA or the PA. Resurfacing prostheses were excluded. Kaplan– Meier curves for risk of reoperation due to dislocation and all-cause for the 2 surgical approaches were compared for 2 periods (1999–2006 and 2007–2014) up to 2 years postoperatively. We used Cox regression for sex, age, type of fixation, and head size to determine hazard ratios (HR) with DLA set as reference. Results — 156,979 THRs met the selection criteria. In 1999–2006, the PA was associated with increased risk of reoperation due to dislocation (HR 2.3, 95% CI 1.7–3.0) but there was no difference in the risk of all-cause reoperation (HR 1.1, CI 0.9–1.2). In 2007–2014 there was no statistically significant difference in the risk of reoperation due to dislocation (HR 1.2, CI 0.9–1.6) but the risk of all-cause reoperation was lower (HR 0.8, CI 0.7–0.9) for the PA. Interpretation — This study confirms historic reports on the increased risk of early reoperations due to dislocations using the PA compared with the DLA. However, in contemporary practice, the higher risk of reoperation due to dislocation associated with PA has declined, now being similar to that after DLA. We believe improved surgical technique for the PA may explain the results. Surprisingly, the PA was associated with lower risk of all-cause reoperation in 2007– 2014. This finding warrants further investigation.

The posterior approach (PA) and the direct lateral approach (DLA) are the 2 principal surgical approaches for total hip replacement (THR) in Sweden (Kärrholm et al. 2017). There is no consensus regarding the optimum surgical approach to be used in primary THR (Jolles and Bogoch 2006). A majority of the dislocations occurs within the first 2 years postoperatively and is 1 of the most common early complications following THR (Hailer et al. 2012, Gausden et al. 2018). Multiple risk factors for dislocation have been identified such as surgical approach, orientation of components, femoral head size, previous surgery, age, sex, BMI, indication for surgery, and comorbidities such as neurological disability and spinal disease (Bystrom et al. 2003, Berry et al. 2005, Patel et al. 2007, Hailer et al. 2012, Fessy et al. 2017, Seagrave et al. 2017, Gausden et al. 2018). The PA has been associated with higher dislocation rate in comparison with the DLA (Robinson et al. 1980, Woo and Morrey 1982, Demos et al. 2001, Masonis and Bourne 2002, Bystrom et al. 2003, Jolles and Bogoch 2006, Hailer et al. 2012, Lindgren et al. 2012). However, enhanced soft tissue repair and improved surgical technique for the PA lower dislocation rates (Pellicci et al. 1998, White et al. 2001, Soong et al. 2004, Suh et al. 2004, Kwon et al. 2006, Kim et al. 2008, Zhou et al. 2017). According to the Swedish Hip Arthroplasty Register (SHAR), the PA and the DLA are used in 95% to 99% of the primary THRs in Sweden. The DLA that has increased from 37% in 1999 to 48% in 2014 at the expense of the PA, which has decreased from 60% in 1999 to 51% in 2014 (Figure 1). To our knowledge there are no studies comparing trends for the risk of reoperation due to dislocation for the PA and the DLA over a long period of time. In the light of improvements in surgical technique, we investigated how the relationship between surgical approach and risk of reoperation due to dislocation has evolved over time.

© 2019 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group, on behalf of the Nordic Orthopedic Federation. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. DOI 10.1080/17453674.2019.1610269

412

Acta Orthopaedica 2019; 90 (5): 411–416

Distribution of approaches (%)

Table 1. Study demographics. Direct lateral and posterior approaches compared

100

1999–2006 2007–2014 Lateral Posterior Lateral Posterior n = 22,507 n = 37,691 p-value n = 44,933 n = 51,848 p-value

Posterior Lateral Other 2000 2002 2004 2006 2008 2010 2012 2014

Year of index operation

Figure 1. Distribution of posterior, direct lateral and other approaches among THRs performed in Sweden from 1999 to 2014.

Age, mean (SD) 68.7 (9.9) 69.6 (9.5) < 0.001 a 68.8 (9.7) 69.2 (9.5) < 0.001 a Female sex, n (%) 12,829 (57) 21,297 (57) 0.2 b 26,329 (59) 29,401 (57) < 0.001 b ASA-class, n (%) < 0.001 b I 9,994 (22) 10,160 (20) II 23,902 (53) 26,474 (51) III 5,599 (13) 7,069 (14) IV 143 (0.3) 168 (0.3) V 0 (0) 3 (0.0) Missing 5,295 (12) 7,974 (15) BMI, mean (SD) 27.3 (5.1) 27.5 (5.2) < 0.001 a Missing, n (%) 6,236 (14) 8,451 (16) a ANOVA, b Chi-squared

test

All THRs in Sweden 1999–2014 n = 226,254 Excluded (n = 69,275): – not osteoarthritis, 48,116 – resurfacing prosthesis, 2,258 – missing head size or not 28, 32, 36 mm, 10,838 – not posterior or lateral, 8,063 Study population n = 156,979

Figure 2. Patient selection flowchart. In order to reduce heterogeneity, the study population was defined according to preset selection criteria. Starting with all THRs in Sweden between 1999–2014 we applied the selection criteria to step-wise filter out relevant surgeries.

Patients and methods Since 1979, the SHAR has collected data from all units in Sweden performing THR. The completeness of primary registrations to the SHAR is 98% to 99%, and 93% for revisions (Kärrholm et al. 2017). Data we extracted from the Register included all patients who had primary THR due to osteoarthritis between 1999 and 2014. Resurfacing prostheses and head sizes other than 28, 32, and 36 mm were excluded. Only patients operated with PA or DLA were included (Figure 2). 156,979 hips met the selection criteria. SHAR started collecting data on ASA class and BMI in 2008 (Table 1). Outcome measures Primary outcome measure was first reoperation due to dislocation as reported to the register within 2 years following index surgery. Secondary outcome was all-cause reoperation within 2 years. A reoperation was defined as any further open surgery to the hip, regardless of implant components being removed, exchanged, added, or not. Thus, surgeries such as gluteus

maximus repair, tenotomy, and hip arthroscopy were included among reoperations. Statistics We used 1-way ANOVA test (with equal variance assumption) for continuous demographic variables (age and BMI) and chi-square test (with continuity correction) for categorical variables (sex, fixation, head size, ASA, and cause of reoperation). Kaplan–Meier curves for posterior and direct lateral surgical approaches were compared for 2 different time periods (1999–2006 and 2007–2014) until 2 years follow-up. We used data from both operations if patients were bilaterally operated during the study period; the violation of the assumption of independent observations was considered not to have any practical implications (Ranstam et al. 2011). Each hip was followed from primary THR to first reoperation. Hips were censored at death, reoperation, or at 2 years after primary surgery, whichever came first. For the dislocation analyses, all other first reoperations were censored. Cox regression analyses were used to compare hazard ratios (HR) with and without adjustments with 95% confidence intervals (CI). We adjusted for sex, age, fixation method (cemented, uncemented, hybrid, and reversed hybrid) and head size (except for 1999–2006 as 99% of hips had 28-mm head size). Proportional assumption was checked graphically. Divided by year of surgery and for PA and DLA separately, we calculated Kaplan–Meier survival estimates with reoperation due to dislocation at 2 years as endpoint. The uncertainty of the Kaplan–Meier estimates was indicated by CIs. Linear regression was used to determine whether the linear trend was statistically significant. A p-value of less than 0.05 indicated statistical significance. R software version 3.4.2 (R Core Team 2017) was used for all analyses with “survival” packages “A Package for Survival Analysis in S” and “ggplot2” (Therneau 2015, Wickham 2016).

Acta Orthopaedica 2019; 90 (5): 411–416

413

Figure 3. Kaplan Meier estimates for not being reoperated due to dislocation within 2 years for posterior and direct lateral surgical approaches during 1999–2006 and 2007–2014. Shaded area are 95% confidence intervals.

Ethics, funding, and potential conflicts of interest The Regional Ethical Review Board in Gothenburg approved the study (entry number 804-17). The study received grants from the Handlaren Hjalmar Svensson fund. Grants from the Swedish state under the agreement between the Swedish government and the county councils, the ALF-agreement (ALFGBG–522591) also contributed to the study. The authors have no conflicts of interest.

Results

Figure 4. Kaplan-Meier estimates for not being reoperated up to 2 years (all causes) for posterior and direct lateral surgical approaches during 1999–2006 and 2007–2014. Shaded area are 95% confidence intervals.

Table 2. Cox regression analyses were used to compare hazard ratio (HR) for reoperation due to dislocation and reoperation due to all causes within 2 years with and without adjustments for time periods 1999–2006 and 2007–2014 Reoperation cause

1999–2006 2007–2014 HR (95% CI) p-value HR (95% CI) p-value

Dislocation Unadjusted Adjusted Posterior approach (ref. direct lateral approach) Age Female (ref. male) Hybrid (ref. cemented) Reverse hybrid (ref. cemented) Uncemented (ref. cemented) 32-mm head (ref. 28-mm head) 36-mm head (ref. 28-mm head) All causes Unadjusted Adjusted Posterior approach (ref. direct lateral approach) Age Female (ref. male) Hybrid (ref. cemented) Reverse hybrid (ref. cemented) Uncemented (ref. cemented) 32-mm head (ref. 28-mm head) 36-mm head (ref. 28-mm head)

As indicated in the Kaplan–Meier curves, the risk of reoperation due to dislocation in 1999–2006 was statistically significantly higher for the PA almost directly postoperatively (Figure 3). PA was associated with higher risk of reoperation due to dislocation in 1999–2006 (HR 2.3, CI 1.7–3.0) but not in 2007–2014 (HR 1.2, CI 0.9–1.6) compared with the DLA (Table 2). All-cause reoperation Kaplan–Meier estimates for the 2 approaches were similar for the first period. For the period 2007–2014, however, the Kaplan– Meier curves indicated a higher risk of reoperation due to all causes for the DLA starting at approximately 1 year postoperatively (Figure 4). PA was associated with similar risk of all-cause reoperation in 1999–2006 (HR 1.1, CI 0.9–1.2) compared with the DLA (Table 2). In 2007–2014, PA was associated with statistically significantly lower risk of reoperation due to all causes (HR 0.8, CI 0.7–0.9) compared with the DLA (Table 2).

2.2 (1.7–2.9) < 0.001

1.1 (0.9–1.5)

2.3 (1.7–3.0) < 0.001 1.0 (1.0–1.0) < 0.001 0.8 (0.6–1.0) 0.03 1.1 (0.5–2.4) 0.8 1.6 (0.8–3.1) 0.2 1.8 (1.0–3.3) 0.07 N/A N/A

1.2 (0.9–1.6) 0.2 1.0 (1.0–1.0) 0.004 0.9 (0.7–1.1) 0.4 2.0 (0.9–4.6) 0.1 0.9 (0.6–1.5) 0.8 2.7 (1.9–3.9) < 0.001 0.7 (0.5–0.9) 0.01 0.6 (0.4–1.1) 0.09

1.0 (0.9–1.2)

0.8 (0.7–0.9) < 0.001

0.8

1.1 (0.9–1.2) 0.4 1.0 (1.0–1.0) 0.002 0.8 (0.7–0.9) < 0.001 1.1 (0.8–1.6) 0.6 1.9 (1.4–2.6) < 0.001 1.5 (1.1–2.0) 0.02 N/A N/A

0.8 (0.7–0.9) 1.0 (1.0–1.0) 0.7 (0.7–0.8) 1.0 (0.7–1.6) 1.5 (1.3–1.7) 1.9 (1.6–2.2) 1.0 (0.9–1.1) 1.1 (0.9–1.4)

0.3

< 0.001 < 0.001 < 0.001 1.0 < 0.001 < 0.001 0.9 0.5

Split by year for primary surgery, the trend analysis of Kaplan–Meier estimates for not being reoperated due to dislocation at 2 years demonstrated positive linear trends for both the DLA (p < 0.05) and the PA (p < 0.01) (Figure 5).

Discussion This study confirms historic reports on the increased risk of early reoperations due to dislocations using the PA compared

414

Acta Orthopaedica 2019; 90 (5): 411–416

Table 3. Distribution of method of fixation, head size and different causes for reoperation within 2 years for direct lateral and posterior approaches during 1999–2006 and 2007–2014. Chi-squared test was used. Values are frequency (%)

Dislocation survival at 2 years (%) 100

1999–2006 2007–2014 Lateral Posterior Lateral Posterior n = 22,507 n = 37,691 p-value n = 44,933 n = 51,848 p-value 98 Method of fixation < 0.001 < 0.001 Cemented 18,171 (81) 35,050 (93) 30,919 (69) 37,532 (72) Hybrid 1,219 (5.4) 869 (2.3) 609 (1.4) 933 (1.8) Reverse hybrid 1,002 (4.5) 908 (2.4) 5,975 (13) 6,272 (12) Uncemented 2,115 (9.4) 864 (2.3) 7,430 (17) 7,111 (14) Head size (mm) < 0.001 < 0.001 28 22,386 (99) 37,479 (99) 22,239 (50) 19,799 (38) 32 119 (0.5) 179 (0.5) 21,885 (49) 27,586 (53) 36 2 (0.0) 33 (0.1) 809 (1.8) 4,463 (8.6) Reoperations within 2 years < 0.001 < 0.001 Aseptic loosening 43 (0.2) 37 (0.1) 111 (0.2) 81 (0.2) Fracture 41 (0.2) 70 (0.2) 96 (0.2) 147 (0.3) Infection 124 (0.6) 180 (0.5) 486 (1.1) 441 (0.9) Dislocation 60 (0.3) 220 (0.6) 136 (0.3) 185 (0.4) Other 52 (0.2) 42 (0.1) 138 (0.3) 68 (0.1) Not reoperated 22,187 (99) 37,142 (99) 43,966 (98) 50,926 (98)

with the DLA in primary THR due to OA. However, in contemporary practice, the higher risk associated with PA had declined and did not entail a statistically significant increased risk of reoperation due to dislocation within 2 years from primary surgery compared with DLA. Surprisingly, the PA was associated with lower risk of reoperation due to all causes. Despite differences in head size, fixation type, and demography between groups, adjusting for confounders did not alter the results. In this nationwide observational study, the rate of reoperations due to dislocation within 2 years following THR for OA was 0.6% in 1999–2006 and 0.4% in 2007–2014 for the PA, and 0.3% in 1999–2006 and 0.3% in 2007–2014 for the DLA. In the meta-analyses by Kwon et al. (2006) regarding 11 papers between 1997 and 2004 the dislocation rate for the DLA was 0.4% while it was 1.0% for the PA. In the review by Masonis and Bourne (2002) of 14 studies between 1976 and 2001 involving 13,203 primary THRs the dislocation rate was estimated as 6 times higher for the PA (3%) than the DLA (0.5%). According to a study by Hailer et al. (2012) with data from the SHAR on 78,098 THRs in 61,743 patients performed between 2005 and 2010 there was a 1.3 times increased relative risk of revision due to dislocation for the PA compared with the DLA with mean follow-up of 2.7 (0–6) years. However, in our study PA was not associated with higher risk of reoperation due to dislocation in 2007–2014. There are important differences between ours and other studies when it comes to, e.g., selection criteria, time period, follow-up time, and use of components that have to be considered when comparing. We believe there has been an ongoing refinement of the surgical technique in THR over the years. For instance, in our study we found an extensive increase in head sizes larger than 28 mm in 2007–2014 compared with

Lateral Posterior 2000 2002 2004 2006 2008 2010 2012 2014

Year of index operation

Figure 5. Annual Kaplan Meier estimates (and 95% confidence intervals) for not being reoperated due to dislocation at 2 years after primary THR for posterior and direct lateral approach. Linear regression was used to investigate if the linear trend was statistically significant. Lateral, p < 0.05 and posterior, p < 0.001.

1999–2006 (Table 3). According to a study by Berry et al. (2005) including 21,047 primary THRs performed in a single institution between 1969 and 1999, the relative risk of dislocation was 1.3 for 28 mm heads compared with 32 mm heads. This is consistent with our study where the 32 mm heads were associated with a statistically significantly lower risk of reoperation due to dislocation up to 2 years (HR 0.7, CI 0.5–0.9) compared with 28 mm heads (Table 2). However, 36 mm heads were not associated with lower risk of reoperation due to dislocation (HR 0.6, CI 0.4–1.1) in our study (Table 2). Larger head size and implant use may certainly explain some of the overall improvement in reoperation rates within 2 years for the 2 approaches investigated. However, we believe it is unlikely that the use of larger head size accounts for all the improvement. In 1999–2006, more than 99% of the THRs in Sweden were performed with 28 mm heads for both DLA and PA. Divided by year of primary surgery, improvement trend in annual Kaplan–Meier estimates was much more pronounced for PA compared with DLA (Figure 5). The increased reoperation-free survival for PA was evident already in 1999–2006. In those years, more than 99% of THRs were operated with 28 mm heads. Hence, the improvement for the PA in 1999–2006 is not attributable to use of larger head sizes. As discussed below, improved surgical technique may contribute to the positive trend for PA. Lindgren et al. (2014) used the SHAR to study 42,233 patients undergoing primary THR for OA operated between 2002 and 2010 and found that the PA was associated with slightly better patient-reported outcomes compared with the DLA. Hence, we believe the possible difference in PROMs should be considered in the choice of THR approach in OA patients.

Acta Orthopaedica 2019; 90 (5): 411–416

To our knowledge there is no other study that has compared dislocation survival trends for the DLA and PA for such a long period of time. Our data from the SHAR showed statistically significant positive linear trends for both surgical approaches regarding the risk of being reoperated due to dislocation within 2 years after primary THR for OA in Sweden (Figure 5). This improvement over time was most apparent for the PA. Since its inception in 1979, Swedish orthopedic surgeons have been influenced by the reports of SHAR. Register findings discussed at internal meetings in the mid-2000s indicated an increased risk of revision for the PA and, as demonstrated here, the use of the DLA increased at the expense of the PA. Between 2001 and 2006, a large body of research, including clinical trials, meta-analyses, and literature reviews, suggested that improved surgical technique with soft tissue repair following the PA in primary THR would reduce the risk of dislocation (Masonis and Bourne 2002, Mahoney and Pellicci 2003, Soong et al. 2004, Suh et al. 2004, Kwon et al. 2006). However, Hailer et al. (2012) concluded that patients with femoral neck fracture or osteonecrosis of the femoral head were at higher risk of dislocation and raised the question as to whether patients belonging to risk groups should be operated using lateral approaches. Furthermore, in a study by Enocson et al. (2009) on 713 consecutive hips, the use of the anterolateral approach for THR in patients with femoral neck fractures was advocated. In the 2011 annual report, the SHAR reported specifically on the increased risk of revision due to dislocation for the PA compared with the DLA (Garellick et al. 2012). None of the surgical approaches were considered superior in adult patients undergoing THR for OA, which was consistent with the Cochrane review by Jolles and Bogoch (2006). However, the SHAR advocated the use of the DLA in patients with risk factors for dislocation (Garellick et al. 2012). These reports likely influenced surgeons’ awareness and provided evidence for improvements in the PA surgical technique. Strengths and limitations The SHAR has a high completeness on primary THR ranging from 98% to 99% and intentionally includes all reoperations and not only revisions. The SHAR has nationwide coverage, which makes the results generalizable. Hence, geographical differences are not likely to affect the results. The inclusion criteria contribute to a more homogeneous study population. The choice of surgical approach for the selected population has most likely been influenced by the local tradition at each hospital rather than patient-specific attributes, which in turn affect the risk of reoperation due to dislocation. The lack of data on BMI and ASA from 1999 to 2006 (given that the SHAR did not start the registration of those variables until 2008) means we were not able to adjust for these confounding factors. Another limitation pertains to the lack of information on the extent of soft tissue repair or the

415

orientation of components. Furthermore, we did not include information on prosthesis type or surgeons’ experience. The SHAR does not capture information on closed reductions. It is unlikely, however, that dislocations can be treated non-surgically at higher success rates with one or the other approach. To what extent the recommendations from the SHAR’s Annual Reports may have caused a selection bias, with complex OA cases with higher anticipated dislocation risk having been operated through a DLA, is uncertain. The skewed accumulation of bigger head sizes and more cemented THRs in the PA group may have favored this surgical approach regarding dislocation survival; however, the accumulation of fewer women and older patients may have disfavored it. The differences between the groups are statistically significant but seemingly small and adjusted for in the statistical analyses. Conclusion In this nationwide observational study, we demonstrate that the historic increased risk of reoperation due to dislocation within 2 years for the PA compared with the DLA has declined substantially in contemporary Swedish THR practice. We believe enhanced surgical technique for the PA, increased awareness of the historically higher dislocation risk of PA, or possibly selection bias may explain this finding. The PA was associated with lower risk of reoperation due to all causes in 2007–2014 compared with the DLA. These findings warrant further research.

OR and MM conceived and designed the study. DO and SN performed statistical analysis. OS drafted the manuscript. All authors interpreted the results, contributed to the discussion, and reviewed the manuscript. Acta thanks Stephan M Röhrl for help with peer review of this study.

Berry D J, von Knoch M, Schleck C D, Harmsen W S. Effect of femoral head diameter and operative approach on risk of dislocation after primary total hip arthroplasty. J Bone Joint Surg Am 2005; 87(11): 2456-63. Byström S, Espehaug B, Furnes O, Havelin L I. Femoral head size is a risk factor for total hip luxation: a study of 42,987 primary hip arthroplasties from the Norwegian Arthroplasty Register. Acta Orthop Scand 2003; 74(5): 514-24. Demos H A, Rorabeck C H, Bourne R B, MacDonald S J, McCalden R W. Instability in primary total hip arthroplasty with the direct lateral approach. Clin Orthop Relat Res 2001; (393): 168-80. Enocson A, Hedbeck C J, Tidermark J, Pettersson H, Ponzer S, Lapidus L J. Dislocation of total hip replacement in patients with fractures of the femoral neck. Acta Orthop 2009; 80(2): 184-9. Fessy M H, Putman S, Viste A, Isida R, Ramdane N, Ferreira A, et al. What are the risk factors for dislocation in primary total hip arthroplasty? A multicenter case-control study of 128 unstable and 438 stable hips. Orthop Traumatol Surg Res 2017; 103(5): 663-8. Garellick G, Kärrholm J, Rogmark C, Rolfson O, Herberts P. The Swedish Hip Arthroplasty Register, Annual Report 2011; 2012.

416

Gausden E B, Parhar H S, Popper J E, Sculco P K, Rush B N M. Risk factors for early dislocation following primary elective total hip arthroplasty. J Arthroplasty 2018; 33(5): 1567-71. Hailer N P, Weiss R J, Stark A, Kärrholm J. The risk of revision due to dislocation after total hip arthroplasty depends on surgical approach, femoral head size, sex, and primary diagnosis: an analysis of 78,098 operations in the Swedish Hip Arthroplasty Register. Acta Orthop 2012; 83(5): 442-8. Jolles B M, Bogoch E R. Posterior versus lateral surgical approach for total hip arthroplasty in adults with osteoarthritis. Cochrane Database of Systematic Reviews 2006; (3) CD003828. Kim Y S, Kwon S Y, Sun D H, Han S K, Maloney W J. Modified posterior approach to total hip arthroplasty to enhance joint stability. Clin Orthop Relat Res 2008; 466(2): 294-9. Kwon M S, Kuskowski M, Mulhall K J, Macaulay W, Brown T E, Saleh K J. Does surgical approach affect total hip arthroplasty dislocation rates? Clin Orthop Relat Res 2006; 447: 34-8. Kärrholm J, Lindahl H, Malchau H, Mohaddes M, Nemes S, Rogmark C, et al. The Swedish Hip Arthroplasty Register, Annual Report 2016; 2017. Lindgren V, Garellick G, Kärrholm J, Wretenberg P. The type of surgical approach influences the risk of revision in total hip arthroplasty: a study from the Swedish Hip Arthroplasty Register of 90,662 total hip replacements with 3 different cemented prostheses. Acta Orthop 2012; 83(6): 55965. Lindgren J V, Wretenberg P, Kärrholm J, Garellick G, Rolfson O. Patientreported outcome is influenced by surgical approach in total hip replacement: a study of the Swedish Hip Arthroplasty Register including 42,233 patients. Bone Joint J 2014; 96-b(5): 590-6. Mahoney C R, Pellicci P M. Complications in primary total hip arthroplasty: avoidance and management of dislocations. Instructional Course Lectures 2003; 52: 247-55. Masonis J L, Bourne R B. Surgical approach, abductor function, and total hip arthroplasty dislocation. Clin Orthop Relat Res 2002; (405): 46-53.

Acta Orthopaedica 2019; 90 (5): 411–416

Patel P D, Potts A, Froimson M I. The dislocating hip arthroplasty: prevention and treatment. J Arthroplasty 2007; 22(4 Suppl. 1): 86-90. Pellicci P M, Bostrom M, Poss R. Posterior approach to total hip replacement using enhanced posterior soft tissue repair. Clin Orthop Relat Res 1998; (355): 224-8. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria: R Core Team; 2017. Ranstam J, Kärrholm J, Pulkkinen P, Mäkelä K, Espehaug B, Pedersen AB, et al. Statistical analysis of arthroplasty data. II. Guidelines. Acta Orthop 2011; 82(3): 258-67. Robinson R P, Robinson H J, Jr., Salvati E A. Comparison of the transtrochanteric and posterior approaches for total hip replacement. Clin Orthop Relat Res 1980; (147): 143-7. Seagrave K G, Troelsen A, Malchau H, Husted H, Gromov K. Acetabular cup position and risk of dislocation in primary total hip arthroplasty. Acta Orthop 2017; 88(1): 10-17. Soong M, Rubash H E, Macaulay W. Dislocation after total hip arthroplasty. J Am Acad Orthop Surg 2004; 12(5): 314-21. Suh K T, Park B G, Choi Y J. A posterior approach to primary total hip arthroplasty with soft tissue repair. Clin Orthop Relat Res 2004; (418): 162-7. Therneau T. A package for survival analysis in S. version 2.38; 2015. White R E, Jr, Forness T J, Allman J K, Junick D W. Effect of posterior capsular repair on early dislocation in primary total hip replacement. Clin Orthop Relat Res 2001; (393): 163-7. Wickham H. ggplot2: Elegant graphics for data analysis. New York: SpringerVerlag; 2016. Woo R Y, Morrey B F. Dislocations after total hip arthroplasty. J Bone Joint Surg Am 1982; 64(9): 1295-306. Zhou Y, Cao S, Li L, Narava M, Fu Q, Qian Q. Is soft tissue repair a right choice to avoid early dislocation after THA in posterior approach? BMC Surg 2017; 17(1): 60.

Acta Orthopaedica 2019; 90 (5): 417–420

417

Registry study on failure incidence in 1,127 revised hip implants with stem trunnion re-use after 10 years of follow-up: limited influence of an adapter sleeve Saverio AFFATATO 1, Monica COSENTINO 1, Francesco CASTAGNINI 2, and Barbara BORDINI 1 1 Laboratorio di Tecnologia Medica, IRCCS—Istituto Ortopedico Rizzoli, Bologna, Italy; 2 Ortopedia-Traumatologia e Chirurgia protesica e dei reimpianti d’anca e di ginocchio, IRCCS—Istituto Ortopedico Rizzoli, Bologna, Italy Correspondence: affatato@tecno.ior.it Submitted 2018-12-20. Accepted 2019-04-12.

Background and purpose — Little is known about the role of retained trunnions in revision hip arthroplasties, i.e., when only the femoral head is substituted. Wear (fretting corrosion) and ceramic head fractures are 2 poorly understood concerns related to use, and the role of adapter sleeves has not been defined. In this registry study we assessed the influence of sleeve interposition on re-revision rates in revision hip arthroplasties with retained stems. Confounding factors (demographics, implant-related features) and failures were also analyzed. Patients and methods — We conducted a registry study on 1,127 revised implants (retained trunnion and head exchange). In 26% of implants an adapter sleeve was interposed; in 74% no adapter sleeve was implanted. Demographic and implant-related features were investigated including a descriptive analysis of failures. Results — The mean follow-up of revised implants with and without the use of an adapter sleeve was 3.3 and 5.1 years, respectively. The implant survival without an adapter sleeve was significantly higher, 98.4% (95% CI 96.9–99.8) vs. 95.2% (CI 93.2–96.6) with an adapter sleeve at 5 years. No re-revisions due to adverse local tissue reactions or ceramic head fractures were reported. In order to overcome the different distribution of head materials and head sizes in the two cohorts, only Delta balls were investigated. Interpretation — Adapter sleeve interposition had a minor influence on the revision rates. No adverse local tissue reactions or head fractures occurred.

Retained modular head–neck junctions in revision hip arthroplasty imply 2 main concerns: wear and ceramic head failure (Hannouche et al. 2010, Gührs et al. 2015, Higgs et al. 2016, Osman et al. 2016). Fretting and mechanically assisted crevice corrosion may occur at the head–neck junction in case of metal-on-metal contact, potentially leading to metallic ion release, adverse local tissue reactions, and implant failures may ensue (Osman et al. 2016, Koch et al. 2017, MacDonald et al. 2017). In particular, mixed metal couplings have been reported to increase corrosion and fretting (Koch et al. 2017). Concerning ceramic balls, trunnion wear may act as a stress enhancer, initiating cracks that, eventually, would lead to head fracture (Hannouche et al. 2003, 2010). Therefore, when the stem is retained and ceramic heads are implanted, producers advise surgeons to interpose a titanium adapter sleeve to improve contacts between the conical trunnion of the femoral stem and the conical bore of the femoral head (Koch et al. 2017). However, with the exception of a few in vitro studies concerning adapter sleeve use (Gührs et al. 2015, Koch et al. 2017, MacDonald et al. 2017, Berstock et al. 2018), little is known about the influence on implant survivorship of retained trunnions at revision, with or without the interposition of adapter sleeves. Clinical impacts of wear and ceramic head fractures due to trunnion re-use appeared negligible, regardless of the interposition of an adapter sleeve, with only anecdotal reports of head fractures and metallosis (Pulliam and Trousdale 1997, Hannouche et al. 2010, Koch et al. 2017). We assessed with registry data the influence of adapter sleeves on re-revision rates, and the role of demographics and implant-related features on survivorship.

418

Patients and methods The Register of Prosthetic Orthopedic Implants (RIPO) is a regional database (surveillance over 4,500,000 inhabitants) recording clinical conditions of patients, surgical procedures, implant characteristics, i.e., fixation, head size and material, and type of acetabular cup and stem (RIPO 2018, Bordini et al. 2019). Primary and revision hip and knee replacement surgeries are included. All hospital admissions of Emilia Romagna residents, even when occurring in other regions, are paid for by the region itself, and thus recorded in the registry. Conversely, patients undergoing arthroplasty procedures in Emilia Romagna but living outside the region were excluded from the analysis, as these patients had no stable connections with the regional health system (Bordini et al. 2019). In the RIPO we identified all cases with stem retention and head exchange during 2000–2016. Statistics Patients’ ages were compared using a t-test. Sex and implantrelated features were compared using chi-square analysis. The survival rate of patients was calculated and plotted according to the Kaplan–Meier method. The end-point was stem or neck revision, whereas revision of the cup/insert was not considered as a failure, since it did not involve stem issues. Implants were followed until the last date of observation (date of death or December 31, 2016). Statistical analyses were performed using SPSS 14.0, version 14.0.1 (SPSS Inc, Chicago, IL, USA) and JMP, version 12.0.1 (SAS Institute Inc, Cary, NC, USA, 1989–2007). Ethics, funding, and potential conflicts of interests Ethical approval was not necessary due to the features of registries and databases (data anonymization). No funds were received for this study. All authors declare that no potential conflict of interest exists.

Results 1,127 cases with revision implants were included in the study. 2 cohorts were identified, with and without interposition of an adapter sleeve. The demographic features of the 2 cohorts were similar (Table 1). The vast majority of stems were made of titanium alloy and had a 12/14 taper. The main differences in the 2 cohorts were related to head size (larger in the “adapter sleeve” cohort) and head material (more ceramic heads in the “adapter sleeve” cohort, more CrCo heads in the “no adapter sleeve” cohort) (p = 0.001). The implant survival curves of the 2 cohorts showed a statistically significant difference (p = 0.04), with revision implants with adapter sleeve interposition performing better (Figure 1).

Acta Orthopaedica 2019; 90 (5): 417–420

Table 1. Demographics and implant-related features of the 2 cohorts (with and without adapters). Values are frequency (%) unless otherwise specified Descriptive

Adapter

Number of revised implants Mean age at revision (range) Sex Female Male BMI group a Underweight Normal Overweight Obese Stems b CONUS Sulzer (Ti6Al7Nb) CBC Mathys (Ti6Al4V) ANCA FIT Cremascoli (Ti6Al4V) CLS Sulzer (Ti6Al7Nb) TAPERLOC Biomet (Ti6Al4V) SL PLUS Endoplus (Ti6Al7Nb) VERSYS FIBER METAL TAPER Zimmer (Ti6Al4V) CONUS Zimmer (Ti6Al7Nb) CORAIL Depuy (Ti6Al4V) ABGII Howmedica (TMZF) CFP Link (Ti) CLS Zimmer (Ti6Al7Nb) SPS Symbios (Ti6Al4V) Others (models with < 20 cases) Fixation Cementless Cemented Neck c Fixed Modular Head size < 28 mm 28 mm 32 mm 36 mm ≥ 38 mm Head material d Biolox Delta Biolox Forte Ceramys Inox Stainless Steel CrCo(Mo) Oxinium Number of failures (stem + neck) Number of failures (stem)

296 (26.3) 69 (25–89)

831 (73.7) 71 (29–92)

177 (60) 119 (40)

533 (64) 298 (36)

3 (1) 87 (36) 37 (15) 118 (48)

9 (1) 262 (39) 129 (19) 279 (41)

26 (9) 25 (8) 35 (12) 15 (5) 15 (5) 5 (2)

64 (8) 39 (5) 21 (2) 39 (5) 29 (3) 38 (5)

6 (2) 5 (2) 5 (2) 3 (1) 6 (2) 6 (2) 9 (3) 131 (45)

25 (3) 25 (3) 25 (3) 25 (3) 16 (2) 15 (2) 11 (1) 454 (55)

271 (92) 22 (8)

677 (82) 150 (18)

218 (74) 78 (26)

731 (88) 100 (12)

– 57 (19) 68 (23) 143 (48) 28 (10)

10 (1) 442 (56) 164 (21) 157 (19) 21 (3)

268 (91) 3 (1) – 16 (5) 9 (3) – 5 (1.7) 3 (1.0)

160 (20) 85 (11) 5 (1) 66 (8) 444 (56) 37 (4) 39 (4.7) 36 (4.3)

a Missing data: 203 (18% of the total). b Alloy of the cementless implant, all tapers

(V40) — missing data: 9 (1% of the total). c Missing data: 37 (3% of the total). d Missing data: 34 (3% of the total).

No adapter

are 12/14 except ABGII

The most frequent reason for re-revision was stem aseptic loosening in both cohorts (Table 2). Re-revisions due to sepsis occurred only in the “no adapter sleeve” cohort (1.0%). No ceramic head fracture occurred. No demographic or implant-related factor apparently influenced rerevision rates or reasons for re-revision (p = 0.4). Of the 44 cases, 18 were in overweight patients, and 6 were obese. 22 failures

Acta Orthopaedica 2019; 90 (5): 417–420

419

Re-revisionfree survival (%) – all heads

Re-revisionfree survival (%) – Delta heads

100

Years after index revision

Figure 1. Kaplan–Meier survival rates of the 2 cohorts (with and without adapters). At risk Adapter No adapter

Years after index revision 1 3 5

296 227 137 79 30 831 688 515 371 248

Table 2. Reasons for re-revision. Values are frequency (%) Reason

Adapter

No adapter

Stem aseptic loosening Modular neck breakage Septic loosening Total aseptic loosening Periprosthetic bone fracture Recurrent prosthesis dislocation Cup aseptic loosening Pain without loosening Unknown

2 (0.7) 2 (0.7) – – – – – – 1 (0.3)

9 (1.1) – 8 (1.0) 6 (0.7) 5 (0.6) 4 (0.5) 4 (0.5) 1 (0.1) 2 (0.2)

Total

5 (1.7)

39 (4.7)

were CrCo head implants (5% of the total CrCo head implants): all these cases were without an adapter sleeve and a 28 mm head was frequently implanted (in 16/22 cases). 8 failures involved BIOLOX-Delta heads (1.9%) (CeramTec GmbH, Plochingen, Germany). 9/88 BIOLOX-Forte heads (CeramTec GmbH, Plochingen, Germany) failed: all failures were implanted without an adapter sleeve; 5 failures involved 28 mm heads. Most of the failures occurred in cementless, titanium stems with 12/14 tapers (the most represented cases in both cohorts). However, this comparison is affected by a different distribution of 2 notable implant-related variables, head size and head material, which were strongly connected (as CrCo heads were frequently 28 mm). Thus, in order to control these confounding factors a further analysis was performed, involving only Delta heads. The “adapter sleeve” cohort (268 implants at a mean follow-up of 3.2 years [0–11]) achieved a lower survivorship than the “no adapter sleeve” cohort (60 implants at a

Years after index revision

Figure 2. Kaplan–Meier survival rates of the 2 cohorts (with and without adapters) involving only the Delta head. At risk Adapter No adapter

Years after index revision 1 3 5

268 204 121 66 22 160 127 89 45 20

mean follow-up of 3.6 years [0–9.4]), (Figure 2); a non-statistically significant difference was observed (p = 0.2). 4 failures occurred in both groups. Middle-aged, overweight/obese men were more commonly involved, regardless of the use of adapter sleeves (Table 3, see Supplementary data).

Discussion The purpose of this study was to investigate the role of adapters in revision hip arthroplasties with retained stems. We found that the use of adapter sleeves statistically significantly reduced the rate of re-revisions. However, no cases of re-revisions due to metallosis or head fractures were reported. Moreover, the 2 groups differed strongly in terms of head materials and head size. When only Delta heads were involved, the re-revision rates were similar between the “adapter sleeve” and “no adapter sleeve” cohorts. Thus, the role of adapter sleeves seems clinically negligible at mid-term follow-up. Indeed, the current pertinent literature involving clinical evaluations tends to support this finding. Hannouche et al. (2010) performed a retrospective investigation on the fracture risk of ceramic heads implanted on retained trunnions without the use of an adapter sleeve. Their investigation was limited to 61 revised hip implants, following the follow-up of alumina–alumina primary THAs. The authors observed no fractures and suggested replacing the ceramic head when the femoral stem was fully integrated. Kim et al. (2018) performed a prospective study to assess the prevalence of ceramic head fractures on retained trunnions without interposition of titanium adapter sleeves in 100 implants of the same material and design. They found no cases of ceramic

420

head fractures, ascribed to the relatively pristine trunnion surface—assessed before fixing the new head. Moreover, the authors suggested that a titanium adapter sleeve was not necessary with minimal fretting and/or corrosion scores. There were 2 notable findings in our study. The first was that without sleeve interposition high rates of failure occurred with 28 mm CrCo heads and with Forte heads. It is likely that mechanical (high rates of 28 mm head involvement) and tribological features (higher wear in CrCo heads) were responsible, rather than adapters, as no re-revisions were due to suspicious adapter sleeve failure. The second was that septic failures occurred only in the “no adapter sleeve” cohort, where there was a higher rate of CrCo heads with no adapter sleeve interposition, implanted on titanium stems. Metal-on-metal interfaces are more prone to infections, probably due to necrosis and immunomodulation induced by metal particle release (Bordini et al. 2019). Mixing metals may even worsen the local situation. Thus, taper wear as a generator of local metal ion could occur in these cases. However, these speculations require in vitro testing and large observational studies. Strengths of our study are mainly related to the large numbers of implants involved. Our study has a number of limitations. First, the study involved a large number of different materials, designs, and taper sleeve implant techniques based on surgeon personal criteria. Second, only limited followups were achieved and it was not possible to take account of the level of activity of the patients. Third, visual inspection of tapers and surrounding soft tissues was lacking as offset evaluation, due to the large number of stem designs. In addition, CrCo cohorts with and without adapter sleeves could not be compared, as there were only 9 CrCo head implants with adapter sleeve interposition. In summary, the interposition of an adapter sleeve seemed only to influence re-revision rates slightly. The 2 main concerns related to trunnion re-use, wear and ceramic fractures, did not occur in this registry study. Possible adverse local tissue reactions due to metal release could not be completely ruled out by the registry studies, as no visual inspection and histology were available. We found a high rate of 28 mm CrCo head implant failures when no adapter sleeves were used. Similarly, a very high rate of Forte head failures on retained, “uncovered” tapers occurred. These 2 conclusions are only merely speculative (no clear association with lack of adapter sleeve), but may be a matter for further investigation.

Acta Orthopaedica 2019; 90 (5): 417–420

Supplementary data Table 3 is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2019. 1618649

SA and BB performed the conception and design of the study. SA and FC wrote the main manuscript text. MC and BB performed the statistical analysis and interpretation of data. MC and BB prepared Figure 1. All authors reviewed the final version of the manuscript. Acta thanks Bart Bosker and Timothy Wrightfor help with peer review of this study.

Berstock J R, Whitehouse M R, Duncan C P. Trunnion corrosion: what surgeons need to know in 2018. Bone Joint J 2018; 100-B (1 Suppl. A): 44-9. Bordini B, Stea S, Castagnini F, Busanelli L, Giardina F, Toni A. The influence of bearing surfaces on periprosthetic hip infections: analysis of thirty nine thousand, two hundred and six cementless total hip arthroplasties. Int Orthop 2019; 43(1): 103-9. Gührs J, Krull A, Witt F, Morlock M M. The influence of stem taper re-use upon the failure load of ceramic heads. Med Eng Phys 2015; 37(6): 545-52. Hannouche D, Nich C, Bizot P, Meunier A, Nizard R, Sedel L. Fractures of ceramic bearings: history and present status. Clin Orthop Relat Res 2003; (417): 19-26. Hannouche D, Delambre J, Zadegan F, Sedel L, Nizard R. Is there a risk in placing a ceramic head on a previously implanted trunion? Clin Orthop Relat Res 2010; 468(12): 3322-7. Higgs G B, MacDonald D W, Gilbert J L, Rimnac C M, Kurtz S M, Implant Research Center Writing Committee. Does taper size have an effect on taper damage in retrieved metal-on-polyethylene total hip devices? J Arthroplasty 2016; 31(9 Suppl): 277-81. Kim Y-H, Park J-W, Kim J-S. Adapter sleeves are not needed to reduce the risk of fracture of a new ceramic head implanted on a well-fixed stem. Orthopedics 2018; 41(3): 158–63. Koch C N, Figgie M, Figgie M P, Elpers M E, Wright T M, Padgett D E. Ceramic bearings with titanium adapter sleeves implanted during revision hip arthroplasty show minimal fretting or corrosion: a retrieval analysis. HSS J 2017; 13(3): 241–7. MacDonald D W, Chen A F, Lee G C, Klein G R, Mont M A, Kurtz S M, Taper Corrosion Writing Committee, Cates H E, Kraay M J, Rimnac C M. Fretting and corrosion damage in taper adapter sleeves for ceramic heads: a retrieval study. J Arthroplasty 2017; 32(9): 2887-91. Osman K, Panagiotidou A P, Khan M, Blunn G, Haddad F S. Corrosion at the head–neck interface of current designs of modular femoral components: essential questions and answers relating to corrosion in modular head–neck junctions. Bone Joint J 2016; 98-B(5): 579-84. Pulliam I T, Trousdale R T. Fracture of a ceramic femoral head after a revision operation: a case report. J Bone Joint Surg Am 1997; 79(1): 118–21. RIPO. Report annuale 2016 Regione Emilia-Romagna; 2018.

Acta Orthopaedica 2019; 90 (5): 421–426

421

Uncemented or cemented revision stems? Analysis of 2,296 firsttime hip revision arthroplasties performed due to aseptic loosening, reported to the Swedish Hip Arthroplasty Register Yosef TYSON 1,2, Ola ROLFSON 2,3, Johan KÄRRHOLM 2,3, Nils P HAILER 1,2, and Maziar MOHADDES 2,3 1 Section

of Orthopedic Surgery, Department of Surgical Sciences, Uppsala University Hospital, Uppsala; 2 The Swedish Hip Arthroplasty Register, Gothenburg; 3 Department of Orthopedics, Institute of Clinical Sciences, Sahlgrenska Academy, Sahlgrenska University Hospital, Mölndal, Sweden Correspondence: yosef.tyson@surgsci.uu.se Submitted 2019-01-26. Accepted 2019-05-09.

Background and purpose — Uncemented stems are increasingly used in revision hip arthroplasty, but only a few studies have analyzed the outcomes of uncemented and cemented revision stems in large cohorts of patients. We compared the results of uncemented and cemented revision stems. Patients and methods — 1,668 uncemented and 1,328 cemented revision stems used in first-time revisions due to aseptic loosening between 1999 and 2016 were identified in the Swedish Hip Arthroplasty Register. Kaplan–Meier analysis was used to investigate unadjusted implant survival with re-revision for any reason as the primary outcome. Hazard ratios (HR) for the risk of re-revision were calculated using a Cox regression model adjusted for sex, age, head size, concomitant cup revision, surgical approach at primary and at index revision surgery, and indication for primary total hip arthroplasty. Results — Unadjusted 10-year survival was 85% (95% CI 83–87) for uncemented and 88% (CI 86–90) for cemented revision stems. The adjusted HR for re-revision of uncemented revision stems during the first year after surgery was 1.3 (CI 1.0–1.6), from the second year the HR was 1.1 (CI 0.8–1.4). Uncemented stems were most often re-revised early due to infection and dislocation, whereas cemented stems were mostly re-revised later due to aseptic loosening. Interpretation — Both uncemented and cemented revision stems had satisfactory long-term survival but they differed in their modes of failure. Our conclusions are limited by the fact that femoral bone defect size could not be investigated within the setting of the current study.

Uncemented revision stems are increasingly used in Sweden (Garellick et al. 2015), with revision being defined as removal or exchange of 1 or all parts of the prosthesis. In a registerbased study, uncemented modular revision stems had a higher rate of re-revision compared with cemented revision stems (Weiss et al. 2011); however, mean follow-up was only 3.4 years for uncemented and 4.2 years for cemented revision stems. A retrospective study of 85 uncemented and 124 cemented revision stems (Hernigou et al. 2015) also observed increased risk of re-revision after use of uncemented fixation. In contrast, uncemented revision stems had a better survival in another retrospective study of 86 stems (Schmale et al. 2000), whereas Iorio et al. (2008) reported similar survival of uncemented and cemented revision stems, but the mortality was lower in patients operated with an uncemented stem. However, this study was also small, based on only 86 stems. Thus, there is no consensus on whether uncemented or cemented revision stems are the best choice in femoral revision surgery. Bone defect size and varus remodeling, patient age and comorbidity, surgeon training and skills are all factors that are described to influence surgeon choice of fixation method (Della Valle and Paprosky 2004, Hartman and Garvin 2011). The reasons for re-revision in either type of stem fixation are also poorly delineated, even though this has been investigated for acetabular revision, describing differences between fixation methods (Mohaddes et al. 2013, 2015, Laaksonen et al. 2017), suggesting there might be a difference for stems as well. Mortality is also seldom reported. Our primary aim was to investigate how the risk of re-revision differs between uncemented and cemented revision stems. Secondary aims were to study how modes of failure differ between uncemented and cemented revision stems, and how the mode of stem fixation influenced short-term mortality after hip revision arthroplasty.

Acta Orthopaedica 2019; 90 (5): 421–426

422

Patients and methods We designed a comparative cohort study on patients reported to the Swedish Hip Arthroplasty Register (SHAR), which is the oldest national THA registry in the world and has collected data on revisions from 1979 (Herberts et al. 1989). The reporting is voluntary but the completeness for primary THA is estimated to be 98%, and for revision THA 94% (Soderman et al. 2000). First-time stem revisions due to aseptic loosening between 1999 and 2016 were identified. The rationale for choosing 1999 as the 1st year was the introduction of a more detailed recording of the implants in the registry. To reduce heterogeneity within the studied cohort the following exclusion criteria were applied: stem types with fewer than 100 observations; cement-in-cement revisions; head sizes with fewer than 50 observations; and surgical approaches at either index or primary surgery with less than 50 observations (Figure 1). Only cemented revision stems with a length of 165 mm or longer were included. Because only distally anchored uncemented revision stems are used, all stem lengths were included (Table 1, see Supplementary data). If both hips were revised, only the 1st revised hip was included in the study. Patients were divided into 2 treatment groups, those who received uncemented (n = 1,668) and those who received cemented revision stems (n = 1,328) during the index procedure. Primary outcome was re-revision for any reason, re-revision being defined as any subsequent revision, including but not limited to isolated stem revisions. Secondary outcomes were re-revision due to aseptic loosening, deep infection, dislocation, fracture, and other, and 90-day mortality. Follow-up started at index surgery (1st revision), and ended at re-revision, death, emigration, or December 31, 2016, whichever came first. Statistics Unadjusted survival was estimated using Kaplan–Meier analysis. Cox multivariable regression models were fitted to calculate hazard ratios (HR) with 95% confidence intervals (CI) with adjustment for age at index revision surgery, sex, indication for primary total arthroplasty, head size, concomitant cup revision (defined as acetabular shell revision or cup in cup revisions, whereas isolated liner exchanges were not included), and surgical approach at both revision and primary surgery. Because unadjusted survival curves deviated considerably from the assumption of proportionality, the regression model was divided into 2 time periods, choosing the dividing line at the time point where the unadjusted survival curves become roughly parallel. Schoenfeld residuals were calculated in order to assess whether model assumptions were met. The 1st year after index surgery was thus chosen as the 1st period, and the 2nd period described the 2nd to 13th years after index surgery. Sensitivity analyses were performed by calculating the HR only for patients operated with cemented

First time stem revisions due to aseptic loosening in the Swedish Hip Arthroplasty register between 1999 and 2016 n = 3,645 Excluded (n = 649): – second hip in bilateral revisions, 140 – stem types with < 100 observations, 55 – head sizes with < 50 observations or missing, 65 – primary surgical approaches with < 50 observations or missing, 283 – index surgical approaches with < 50 observations or missing, 54 – other missing variables in the Cox regression, 17 – cement in cement revisions, 35 Included revisions n = 2,996

Figure 1. Flow chart of excluded stems.

stems at primary THA surgery (n = 2,815) in order to investigate whether fixation at primary surgery was associated with outcome, and by stratifying the study population into the following time periods for index revision surgery: 1999–2004 (n = 847), 2005–2010 (n = 1,176), and 2011–2016 (n = 973). The cohort was age stratified into 4 roughly equally large groups: age 67 and under (n = 750); age 68–73 (n = 703); age 74–79 (n = 793); and age > 79 (n = 750), followed by HR stratified for the different age groups. Descriptive statistics were used to study reasons for re-revision and early mortality. SPSS (IBM, version 22; IBM Corp, Armonk, NY, USA) and R Statistical Software (R Foundation for Statistical Computing, Vienna, Austria) were used for the calculations. Ethics, funding, and potential conflicts of interest Ethical approval was obtained from the Regional Ethics Review Board in Gothenburg, Sweden (decision 271-14). Financial support was received from the Health Care Committees in Region Uppsala and Region Västra Götaland. No competing interests were declared.

Results 2,996 patients were included in the study, of whom 1,668 received uncemented revision stems and 1,328 cemented revision stems at the index procedure. Mean follow-up time among patients with uncemented revision stems was 5.5 years (SD 4.0), and it was 7.5 years (SD 4.5) for those with cemented revision stems. The sex distribution was similar in the 2 groups, but the head size differed, with uncemented stems more often receiving larger head sizes (Table 2). Unadjusted 10-year survival was lower for uncemented compared with cemented revision stems (85%, 95% CI 83–87 versus 88%, CI 86–90) (Figure 2). The adjusted risk of rerevision for any reason was slightly higher for uncemented revision stems (HR 1.3, CI 1.0–1.6) during the 1st year after

Acta Orthopaedica 2019; 90 (5): 421–426

423

Table 2. Demographic data of patients operated with uncemented or cemented revision stems. Values are frequency (%) unless otherwise specified Factor

Uncemented Cemented n = 1,668 n = 1,328

Women 713 (43) Mean age at index surgery (SD) 72 (10) Mean follow-up time, years (SD) 5.5 (4.0) Mean time between primary and revision surgery, years (SD) 12.3 (6.0) Diagnosis at primary THA Osteoarthritis 1,321 (79) Fracture 90 (5.4) Inflammatory disease 115 (6.9) Condition after childhood disease 70 (4.2) Other 72 (4.3) Approach at index surgery Direct lateral 831 (50) Posterior 837 (50) Approach at primary THA Direct lateral 858 (51) Posterior 810 (49) Concomitant cup revision 1,258 (75) Head size, mm 22 90 (5.4) 28 841 (50) 32 570 (34) 36 167 (10)

580 (44) 74 (9) 7.5 (4.5) 12.4 (5.7) 1,057 (80) 86 (6.5) 95 (7.2) 39 (2.9) 51 (3.8) 554 (42) 774 (58) 506 (38) 822 (62) 1,328 (74) 112 (8.4) 870 (66) 271 (20) 75 (5.6)

Concomitant cup revision does not include only liner change.

Re-revisionfree survial (%) 100

70 Cemented stem Uncemented stem

9 10 11 12

Years from index revision

Figure 2. Unadjusted survival of uncemented and cemented revision stems.

surgery (Table 3, see Supplementary data). From the 2nd year, the adjusted relative risk of re-revision for any reason for uncemented compared with cemented revision stems was 1.1 (CI 0.8–1.4) (Table 4, see Supplementary data). In the sensitivity analyses, the adjusted risk of re-revision for any reason of joints that had been primarily operated with cemented stems, and the risk of re-revision during different time periods were not substantially different from the risk that was estimated in our main analyses (Table 5, see Supplementary data). In patients 67 years and younger, the risk of re-revision was simi-

Table 6. Risk of re-revision for uncemented revision stems, stratified by age groups at index revision surgery, with cemented stems as reference Patient age < 68 68–73 74–79 > 79

Year after index revision surgery 1st Between 2nd and 13th HR (CI) HR (CI) 1.0 (0.7–1.4) 1.0 (0.7–1.8) 1.4 (0.9–2.2) 2.4 (1.3–4.4)

1.0 (0.7–1.6) 0.6 (0.3–1.3) 1.5 (0.8–3.0) 0.9 (0.3–2.5)

HR (CI): adjusted hazard ratio and 95% confidence interval.

Table 7. Reasons for re-revision for uncemented and cemented revision stems. Values are frequency (%) Reasons for re-revision Aseptic loosening Deep infection Dislocation Fracture Other

Uncemented

Cemented

49 (25) 38 (20) 65 (34) 12 (6) 29 (15)

65 (52) 18 (14) 17 (14) 13 (10) 13 (10)

lar between uncemented and cemented revision stems (Table 6). Patients between 68 and 73 years also had a similar risk of early re-revision, whereas the risk of late re-revision tended to be lower for uncemented revision stems. In the group 74–79 years of age there was a slightly higher risk of re-revision for uncemented revision stems during both time periods. In patients older than 79 years, uncemented revision stems had a higher risk of early re-revision, but during the following period the risk was similar in the 2 groups. Taken together, this means that in patients up to 73 years of age uncemented stems perform as well as or better than cemented stems, whereas in patients 74 years or older, cemented stems perform as well as or better than uncemented stems. The distribution of reasons for re-revision differed between groups. Of the revision stems that were re-revised, more uncemented than cemented revision stems were re-revised due to dislocation and infection, but fewer of the uncemented revision stems were re-revised for aseptic loosening compared with cemented revision stems (Table 7). The 90-day mortality was similar for patients operated with uncemented revision stems compared with those operated with cemented revision stems (12 [1%] versus 15 [1%]). Further, we observed that patients who received concomitant cup revision during the index procedure had a lower adjusted risk of re-revision, both during the 1st year after surgery and between the 2nd and 13th year (HR 0.5, CI 0.4–0.7, and HR 0.5, CI 0.3–0.6) (Tables 3 and 4, see Supplementary data).

424

Discussion In this large cohort study analyzing 2,996 revisions, patients with uncemented revision stems had a higher overall risk of re-revision, especially among older patients. Uncemented stems were more often re-revised due to infection and dislocation, whereas cemented stems were more often re-revised due to aseptic loosening. The mortality was similar in both fixation groups. Uncemented revision stems are increasingly used, but few studies have compared the survival of uncemented and cemented fixation. There is a controversy in the literature, describing uncemented fixation as inferior (Weiss et al. 2011, Hernigou et al. 2015), similar (Iorio et al. 2008), or better (Schmale et al. 2000) than cemented fixation. Revision THA is associated with worse patient-reported outcomes than primary THA and inferior implant survival (Patil et al. 2008, Singh and Lewallen 2009, Ong et al. 2010, Singh and Lewallen 2013), and the revision burden is estimated to double by 2030 (Kurtz et al. 2007). In this study, the implant survival up to 13 years of uncemented and cemented revision stems was similar, with a higher re-revision rate for uncemented revision stems during the early postoperative period. This is in line with Weiss et al. (2011) who also describe higher early re-revision rates for uncemented revision stems. In our study, patients 68â&#x20AC;&#x201C;73 years of age appeared to benefit from the decreased risk of long-term re-revision of uncemented revision stems. One could speculate that younger patients would benefit from use of uncemented stem fixation in 1st-time revisions because of decreased risk of late aseptic loosening, but we could not document such an advantage. Older patients on the other hand, with a shorter life expectancy, might benefit from the decreased risk of early re-revision with cemented revision stems. The increased risk of re-revision for uncemented revision stems in older patients might be due to compromised bone stock among these patients, increasing the risk for distal migration in uncemented revision stems, but this needs further elucidation. The risk of periprosthetic fracture was however not increased in this subgroup. Uncemented revision stems were more often re-revised due to dislocation or infection. Similar findings have been reported from several national registries when analyzing the method of fixation in acetabular revisions (Mohaddes et al. 2013, 2015, Laaksonen et al. 2017). However, we recorded only the rate of re-revision, excluding dislocations treated without any surgical intervention, and also those treated with open reduction without removal or exchange of any implant parts. Several studies have reported dislocation rates between 3% and 19% for both uncemented and cemented revision stems (Hultmark et al. 2000, Haydon et al. 2004, Weiss et al. 2011, Pelt et al. 2014, te Stroet et al. 2015). In our study, larger femoral head sizes were more often used in uncemented revisions. Previous

Acta Orthopaedica 2019; 90 (5): 421â&#x20AC;&#x201C;426

studies describe that larger head sizes decrease the risk for dislocation (Kosashvili et al. 2011, Garbuz et al. 2012, Wetters et al. 2013). It could be speculated that the increased risk of dislocation for uncemented stems would have been even higher if smaller heads had been used. It might also be that that the threshold to re-revise the proximal part of a modular uncemented revision stem is lower than re-revising a cemented revision stem due to dislocation, as suggested by other authors (Weiss et al. 2011). Also, older patients with cemented stems might have lower demand than younger patients with uncemented stems, potentially increasing the threshold to revise such patients due to dislocation. Of all the 194 uncemented revision stems that were re-revised, 46 (24%) included repositioning or exchange of only the proximal part of the stem, with or without concomitant cup revision, which supports this explanation. It could also be that the uncemented revision stems migrate distally and rotate into retroversion more frequently than cemented stems, especially when bone quality is compromised (Paprosky et al. 1999, Lakstein et al. 2010, Hernigou et al. 2015). Such migration might make the joint less stable and facilitate dislocation. The lower rate of re-revision due to infection after the use of cemented revision stems could be related to use of antibiotic loaded cement, a phenomenon that is described for primary total hip arthroplasty (Parvizi et al. 2008, Voigt et al. 2016). The difference in infection rate may also be due to selection bias. Cemented revision stems are more often re-revised due to aseptic loosening. Aseptic loosening is a late complication and the cemented stems have a longer follow-up time, which could in part explain this observation. Further, we do not have sufficient information on whether bone impaction grafting was used, which could have an impact on aseptic loosening. One could speculate that the cementation in a femoral canal devoid of trabecular bone could facilitate aseptic loosening. Also, fewer cemented revision stems were re-revised due to dislocation or infection, leaving more stems at risk of aseptic loosening. The pattern of causespecific reason for re-revision of uncemented and cemented revision stems found by us has also been observed when comparing primary uncemented and cemented stems (Hailer et al. 2010, Gromov et al. 2015). Even though a larger proportion of uncemented revision stems received larger head sizes, they were more often re-revised due to dislocation. This seems to be in contradiction to previous studies where, in accordance with abundant literature, increasing head size was associated with a lower risk of revision due to dislocation (Hailer et al. 2012). We think that the increased re-revision rate might be explained by a lower threshold to exchange and lengthen or reposition the proximal part of an uncemented modular stem in cases with repeated dislocation, an intervention that is not easily available with well-fixed cemented non-modular stems (Paprosky et al. 1999, Lakstein et al. 2010, Weiss et al. 2011, Hernigou et al. 2015). The groups had similar 90-day mortality, suggesting that cemented fixation does not increase the risk of short-term

Acta Orthopaedica 2019; 90 (5): 421–426

mortality. This is not in line with the observations of Iorio et al. (2008), but in that study mortality was measured over the entire study period and the difference observed could be caused by selecting older patients for cemented fixation. According to our observations, reduced early mortality is not a sustainable argument to use uncemented fixation, even if prospective, and preferably randomized comparisons are needed to support our findings. Concomitant cup revision appeared to be associated with decreased risk of re-revision. This finding is difficult to interpret because this issue was not specifically addressed in our study, but it confirms previous reports from the SHAR (Karrholm et al. 2018). However, one could speculate that concomitant cup revision results in improved joint mechanics and might thus reduce the risk for dislocation (Bohm and Bischel 2004, Wetters et al. 2013). It has also been suggested that the primary THA cup is more likely to migrate and loosen than a revised cup (Mulliken et al. 1996, Bohm and Bischel 2004). On the other hand, the opposite situation has also been observed, namely that revision of the femoral component only is associated with lower risk of dislocation than if both components had been revised (Kosashvili et al. 2014). It is notable that three-quarters of procedures in both groups included a concomitant cup revision. We can only speculate on the reasons for this, but aseptic cup loosening, acetabular osteolysis, or simply wear of polyethylene cups or liners are frequent reasons for cup revisions during THA revision surgery. It is possible that Swedish surgeons have embraced findings from the SHAR indicating that concomitant cup revision lowers the rate of re-revision (Karrholm et al. 2018). Based on our findings, concomitant cup revision is advocated since it is associated with a reduced risk of re-revision both in the short and in the long term. Strengths and limitations Although limited, this is the largest cohort study analyzing implant survival following uncemented and cemented revision stem fixation, using data from a national register with excellent coverage and high completeness. This study has several limitations. First, there is no information regarding the presence or magnitude of any bone defects in either fixation group or the rationale for choosing either type of fixation, which is of course a major limitation. Hospitals tend to use either uncemented or cemented fixation (Figure 3, see Supplementary data); whether these choices are due to case mix or not is unclear, but it is unlikely that hospitals treat only patients with either large or small bone defects, and we believe that tradition might have an influence. It is also likely that surgeons in other parts of the world have different indications and usage patterns, which might decrease the generalizability of the findings outside of Sweden. In addition, we investigated only aseptic loosening as reason for rerevision, and it is possible that the findings would be different if other reasons, for example infection and periprosthetic frac-

425

ture, were included. We did not report on cement in cement revisions since this has already been analyzed by our team (Cnudde et al. 2017). Proximally fixed uncemented revision stems were not investigated due to the fact that they are only in limited use in Sweden. The second limitation is lack of a population-based control group as regards mortality and no data on perioperative mortality; however, the risk of death on the operating table is low (Sierra et al. 2009, Ginsel et al. 2014). Finally, as in all register-based studies, there might be residual confounding. In summary, the 8-year risk of re-revision for uncemented stem fixation at first-time revision due to aseptic loosening was similar to that of cemented revision stems, but the different fixation principles differ in their modes of failure. Overall, both stem types offer excellent results. Supplementary data Tables 1, 3, 4, 5, and Figure 3 are available as supplementary data in the online version of this article, http://dx.doi.org/ 10.1080/17453674.2019.1624336

All authors contributed to the design of the study, interpretation of the analysis, writing and revision of the manuscript. YT and MM conducted the statistical analysis. The authors would like to thank the Swedish orthopedic surgeons and administrators reporting to the Swedish Hip Arthroplasty Register. Acta thanks Richard N de Steiger and Wierd P Zijlstrafor help with peer review of this study.

Bohm P, Bischel O. The use of tapered stems for femoral revision surgery. Clin Orthop Relat Res 2004; (420): 148-59. Cnudde P H, Karrholm J, Rolfson O, Timperley A J, Mohaddes M. Cementin-cement revision of the femoral stem: analysis of 1179 first-time revisions in the Swedish Hip Arthroplasty Register. Bone Joint J 2017; 99-B(4 Suppl. B): 27-32. Della Valle C J, Paprosky W G. The femur in revision total hip arthroplasty evaluation and classification. Clin Orthop Relat Res 2004; (420): 55-62. Garbuz D S, Masri B A, Duncan C P, Greidanus N V, Bohm E R, Petrak M J, Della Valle C J, Gross A E. The Frank Stinchfield Award: Dislocation in revision THA: do large heads (36 and 40 mm) result in reduced dislocation rates in a randomized clinical trial? Clin Orthop Relat Res 2012; 470(2): 351-6. Garellick G, Karrholm J, Lindahl H, Malchau H, Rogmark C, Rolfson O. Svenska Höftprotesregistret Årsrapport 2014. Svenska Höftprotesregistret Årsrapport 2014. Gothenburg; 2015. Ginsel B L, Taher A, Ottley M C, Jayamaha S, Pulle C R, Crawford R W. Hospital mortality after arthroplasty using a cemented stem for displaced femoral neck fractures. J Orthop Surg (Hong Kong) 2014; 22(3): 279-81. Gromov K, Pedersen A B, Overgaard S, Gebuhr P, Malchau H, Troelsen A. Do rerevision rates differ after first-time revision of primary THA with a cemented and cementless femoral component? Clin Orthop Relat Res 2015; 473(11): 3391-8. Hailer N P, Garellick G, Karrholm J. Uncemented and cemented primary total hip arthroplasty in the Swedish Hip Arthroplasty Register. Acta Orthop 2010; 81 (1): 34-41.

426

Hailer N P, Weiss R J, Stark P, Karrholm J. The risk of revision due to dislocation after total hip arthroplasty depends on surgical approach, femoral head size, sex, and primary diagnosis: an analysis of 78,098 operations in the Swedish Hip Arthroplasty Register. Acta Orthop 2012; 83(5): 442-8. Hartman C W, Garvin K L. Femoral fixation in revision total hip arthroplasty. J Bone Joint Surg Am 2011; 93-A(24): 2311-22. Haydon C M, Mehin R, Burnett S, Rorabeck C H, Bourne R B, McCalden R W, MacDonald S J. Revision total hip arthroplasty with use of a cemented femoral component: results at a mean of ten years. J Bone Joint Surg Am 2004; 86-A(6): 1179-85. Herberts P, Ahnfelt L, Malchau H, Stromberg C, Andersson G B. Multicenter clinical trials and their value in assessing total joint arthroplasty. Clin Orthop Relat Res 1989; (249): 48-55. Hernigou P, Dupuys N, Delambre J, Guissou I, Poignard A, Allain J, Flouzat Lachaniette C H. Long, titanium, cemented stems decreased late periprosthetic fractures and revisions in patients with severe bone loss and previous revision. Int Orthop 2015; 39(4): 639-44. Hultmark P, Karrholm J, Stromberg C, Herberts P, Mose C H, Malchau H. Cemented first-time revisions of the femoral component: prospective 7 to 13 years’ follow-up using second-generation and third-generation technique. J Arthroplasty 2000; 15(5): 551-61. Iorio R, Healy W L, Presutti A H. A prospective outcomes analysis of femoral component fixation in revision total hip arthroplasty. J Arthroplasty 2008; 23(5): 662-9. Karrholm J, Mohaddes M, Odin D, Vinblad J, Rogmark C, Rolfson O. Svenska Höftprotesregistret Årsrapport 2017. Svenska Höftprotesregistret Årsrapport 2017. Gothenburg; 2018. 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. Kosashvili Y, Drexler M, Backstein D, Safir O, Lakstein D, Safir A, Chakravertty R, Dwyer T, Gross A. Dislocation after the first and multiple revision total hip arthroplasty: comparison between acetabulum-only, femur-only and both component revision hip arthroplasty. Can J Surg 2014; 57(2): E15-8. Kurtz S, Ong K, Lau E, Mowat F, Halpern M. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am 2007; 89(4): 780-5. Laaksonen I, Lorimer M, Gromov K, Rolfson O, Makela K T, Graves S E, Malchau H, Mohaddes M. Does the risk of rerevision vary between porous tantalum cups and other cementless designs after revision hip arthroplasty? Clin Orthop Relat Res 2017; 475(12): 3015-22. Lakstein D, Backstein D, Safir O, Kosashvili Y, Gross A E. Revision total hip arthroplasty with a porous-coated modular stem: 5 to 10 years follow-up. Clin Orthop Relat Res 2010; 468(5): 1310-15. Mohaddes M, Garellick G, Karrholm J. Method of fixation does not influence the overall risk of rerevision in first-time cup revisions. Clin Orthop Relat Res 2013; 471(12): 3922-31. Mohaddes M, Rolfson O, Karrholm J. Short-term survival of the trabecular metal cup is similar to that of standard cups used in acetabular revision surgery. Acta Orthop 2015; 86(1): 26-31.

Acta Orthopaedica 2019; 90 (5): 421–426

Mulliken B D, Rorabeck C H, Bourne R B. Uncemented revision total hip arthroplasty: a 4-to-6-year review. Clin Orthop Relat Res 1996; (325): 156-62. Ong K L, Lau E, Suggs J, Kurtz S M, Manley M T. Risk of subsequent revision after primary and revision total joint arthroplasty. Clin Orthop Relat Res 2010; 468(11): 3070-6. Paprosky W G, Greidanus N V, Antoniou J. Minimum 10-year-results of extensively porous-coated stems in revision hip arthroplasty. Clin Orthop Relat Res 1999; (369): 230-42. Parvizi J, Saleh K J, Ragland P S, Pour A E, Mont M A. Efficacy of antibioticimpregnated cement in total hip replacement. Acta Orthop 2008; 79(3): 335-41. Patil S, Garbuz D S, Greidanus N V, Masri B A, Duncan C P. Quality of life outcomes in revision vs primary total hip arthroplasty: a prospective cohort study. J Arthroplasty 2008; 23(4): 550-3. Pelt C E, Madsen W, Erickson J A, Gililland J M, Anderson M B, Peters C L. Revision total hip arthroplasty with a modular cementless femoral stem. J Arthroplasty 2014; 29(9): 1803-7. Schmale G A, Lachiewicz P F, Kelley S S. Early failure of revision total hip arthroplasty with cemented precoated femoral components: comparison with uncemented components at 2 to 8 years. J Arthroplasty 2000; 15(6): 718-29. Sierra R J, Timperley J A, Gie G A. Contemporary cementing technique and mortality during and after Exeter total hip arthroplasty. J Arthroplasty 2009; 24(3): 325-32. Singh J A, Lewallen D. Age, gender, obesity, and depression are associated with patient-related pain and function outcome after revision total hip arthroplasty. Clin Rheumatol 2009; 28(12): 1419-30. Singh J A, Lewallen D G. Operative diagnosis for revision total hip arthroplasty is associated with patient-reported outcomes (PROs). BMC Musculoskelet Disord 2013; 14: 210. Soderman P, Malchau H, Herberts P, Johnell O. Are the findings in the Swedish National Total Hip Arthroplasty Register valid? A comparison between the Swedish National Total Hip Arthroplasty Register, the National Discharge Register, and the National Death Register. J Arthroplasty 2000; 15(7): 884-9. te Stroet M A, Rijnen W H, Gardeniers J W, van Kampen A, Schreurs B W. The outcome of femoral component revision arthroplasty with impaction allograft bone grafting and a cemented polished Exeter stem: a prospective cohort study of 208 revision arthroplasties with a mean follow-up of ten years. Bone Joint J 2015; 97-B(6): 771-9. Voigt J, Mosier M, Darouiche R. Antibiotics and antiseptics for preventing infection in people receiving revision total hip and knee prostheses: a systematic review of randomized controlled trials. BMC Infect Dis 2016; 16(1): 749. Weiss R J, Stark A, Karrholm J. A modular cementless stem vs. cemented long-stem prostheses in revision surgery of the hip: a population-based study from the Swedish Hip Arthroplasty Register. Acta Orthop 2011; 82(2): 136-42. Wetters N G, Murray T G, Moric M, Sporer S M, Paprosky W G, Della Valle C J. Risk factors for dislocation after revision total hip arthroplasty. Clin Orthop Relat Res 2013; 471(2): 410-16.

Acta Orthopaedica 2019; 90 (5): 427–432

427

Reduced periprosthetic fracture rate when changing from a tapered polished stem to an anatomical stem for cemented hip arthroplasty: an observational prospective cohort study with a follow-up of 2 years Jabbar MOHAMMED 1, Sebastian MUKKA 1, Carl-Johan HEDBECK 2, Ghazi CHAMMOUT 2, Max GORDON 2, and Olof SKÖLDENBERG 2 1 Department of Surgical and Perioperative Sciences, Umeå University; 2 Department of Clinical Sciences, Division of Orthopedics, Karolinska Institutet at Danderyd Hospital, Stockholm, Sweden Correspondence: jabbar.mohammed@umu.se Submitted 2019-04-08. Accepted 2019-05-13.

Background and purpose — Straight collarless polished tapered stems have been linked to an increased risk for periprosthetic femur fractures in comparison with anatomically shaped stems, especially in elderly patients. Therefore, we evaluated the effect of an orthopedic department’s full transition from the use of a cemented collarless, polished, tapered stem to a cemented anatomic stem on the cumulative incidence of postoperative periprosthetic fracture (PPF). Patients and methods — This prospective singlecenter cohort study comprises a consecutive series of 1,077 patients who underwent a cemented hip arthroplasty using either a collarless polished tapered stem (PTS group, n = 543) or an anatomic stem (AS group, n = 534). We assessed the incidence of PPF 2 years postoperatively and used a Cox regression model adjusted for age, sex, ASA class, cognitive impairment, BMI, diagnosis, and surgical approach for outcome analysis. Results — Mean age at primary surgery was 82 years (49–102), 73% of the patients were female, and 75% underwent surgery for a femoral neck fracture. The PPF rate was lowered from 3.3% (n = 18) in the PTS group to 0.4% (n = 2) in the AS group. The overall complication rate was also lowered from 8.8% in the PTS group to 4.5% in the AS group. In the regression model only cognitive dysfunction (HR 3.8, 95% CI 1.4–10) and the type of stem (PTS vs AS, HR 0.1, CI 0.0–0.5) were correlated with outcome. Interpretation — For elderly patients with poor bone quality use of cemented anatomic stems leads to a substantial reduction in periprosthetic fracture rate without increasing other complications.

A severe complication of hip arthroplasty is the periprosthetic femoral fracture (PPF), which is associated with increased mortality (Bhattacharyya et al. 2007, Young et al. 2008). The surgical treatment of PPF is demanding, with high complication and reoperation rates (Lindahl et al. 2005, 2006a, 2006b). Previous studies have reported differences between type of implant and the risk for PPF (Lindahl et al. 2005, Franklin and Malchau 2007, Palan et al. 2016). The different designs of cemented femoral implants rely on different principles of fixation to the femur. The collarless polished tapered stems are designed to subside inside the cement mantle to achieve an even load-bearing and the matte composite-beam anatomical stems are designed to be fixed in the cement mantle. Straight collarless polished tapered stems have been linked to an increased risk for PPF in comparison with anatomically shaped stems, especially in elderly patients, with fracture as an indication for surgery (Lindahl et al. 2006b, Brodén et al. 2015, Mukka et al. 2016, Palan et al. 2016, Kristensen et al. 2018, Chatziagorou et al. 2019). Based on these results our institution changed our standard femoral implant in 2014 for all cemented arthroplasty surgeries. We hypothesized that the transition from a straight, polished, tapered stem to an anatomic matte composite-beam stem would reduce the incidence of PPF and reoperations.

Patients and methods Study setting This observational prospective cohort study was performed between 2012 and the beginning of 2018 (inclusion period 2012–2015) at the Orthopedic Department of Danderyd Hospital in Stockholm, Sweden. Danderyd Hospital is a university

428

hospital affiliated with the Karolinska Institute and provides medical care to a catchment area of approximately 500,000 inhabitants. Study subjects The study patients were identified from an ongoing prospective cohort study including a consecutive series of all hip arthroplasties performed at the Orthopedic Department of Danderyd Hospital. We included all patients operated between 2012 and 2015 with a cemented hip arthroplasty. During the first 2 years of the study (2012–2013) the control group (PTS) were recruited; in the case of bilateral surgeries during the inclusion period, only the 1st hip was included in the analysis. Surgery At our department, a cemented stem is used for both hemiarthroplasty (HA) and total hip arthroplasty (THA). For THA patients with degenerative joint disease and with a type A femur according to Dorr et al. (1993), an uncemented stem is often used. The choice of fixation method is ultimately up to the surgeon but the type of implants used is centralized and decided by the department. In 2014, a policy change was implemented whereby all cemented stems in HA and THA surgeries were changed from a polished tapered stem (PTS group) (CPT, Zimmer Inc., Warsaw, IN, USA) to an anatomic stem (AS group) (Lubinus SP2, Waldemar Link, Hamburg, Germany). In the fall of 2013, all surgeons were trained in the use of the new stem in seminars and with surgeries on saw-bones. Then, in January 2014, the polished, tapered stem was no longer available in the department and all cemented arthroplasty surgery was done using the new stem. The learning curve is thus included in the study. The operations were performed either by a consultant orthopedic surgeon or by a registrar with assistance from a consultant. A modular, size 32 mm cobalt-chrome femoral head and an uncemented or a cemented acetabular component were used for patients operated with THA and a unipolar head for patients operated with HA. Surgical approach According to surgeon preference, a standard posterolateral or a direct lateral approach was used. Based on a previous study from the same cohort of patients (Sköldenberg et al. 2010), the majority of patients with a femoral neck fracture are operated with a direct lateral approach whereas the posterolateral approach is preferred for patients with osteoarthritis. Peri- and postoperative prophylaxis Antibiotic-loaded bone cement was used for all patients (Palacos with gentamicin, Heraues Medical GmbH, Wehrheim, Germany). Prophylactic antibiotics were administered 30 minutes preoperatively and twice more over 24 h postoperatively. Low-molecular-weight heparin was administered for 10–30 days postoperatively.

Acta Orthopaedica 2019; 90 (5): 427–432

Postoperative care Patients were mobilized according to a standard physiotherapeutic program, and immediate full weight-bearing with the use of aids was encouraged. Patients who underwent surgery with a posterolateral approach were instructed to minimize flexion in combination with adduction and internal rotation for the first 3 months. Outcomes and data collection Using the unique Swedish personal ID number, we collected data prospectively throughout the study period through a combination of a search of our in-hospital surgical and medical databases and regular follow-up visits. A digital case report form was used throughout the study. We also used the Swedish Hip Arthroplasty Register to identify any reoperations performed outside our hospital, but no such case was found. All patients were followed up until 2 years after primary surgery or until death. The mean follow-up time was 20 months (median 24, 0–24 months) with no loss to follow-up. The guidelines of the STROBE (STRrengthening the Reporting of OBbservational studies in Epidemiology) statement were followed. Variables We collected data including stem type (PTS/AS) age, sex, cognitive dysfunction (no/yes, classified by the treating surgeon. Temporary confusion was not classified as cognitive dysfunction), ASA score, indication for surgery (primary osteoarthritis/other arthritic diseases (i.e. dysplasia, rheumatoid arthritis)/femoral neck fracture[FNF]/other fracture (i.e. trochanteric or acetabular), type of arthroplasty (THA/HA), surgical approach (posterolateral/direct lateral), and complications leading to reoperation including open surgery with revision of implants. Periprosthetic fractures were classified according to the Vancouver system (Brady et al. 2000) and the surgical treatment used in the reoperation (open reduction and internal fixation [ORIF]/stem revision). For patients with a PPF the radiographs were analyzed by OS, a senior consultant specialized in hip revision surgery. The clinical and radiographic outcomes for the patients with PPF were evaluated by a combination of a medical chart review and radiographic analysis at follow-up visits as in a previously published study at our institution (Brodén et al. 2015). The outcome was graded as: “good” in patients with a radiographically healed fracture and no or little walking impairment, “intermediate” in patients with a healed fracture but impaired walking ability, and “poor” in patients with an unhealed fracture and a severely impaired walking ability. Sample size Prior to the start of the study, a power analysis showed that a 5% significance level, and with 431 hips in each group, would give a power of 80% to detect a statistically significant difference in PPF rate with an assumed 3% fracture rate for the

Acta Orthopaedica 2019; 90 (5): 427–432

429

Table 1. Characteristics of subject. Values are frequency (%) unless otherwise specified Factor Sex Male Female Age (years) (mean, SD) ASA 1–2 3–4 Height (cm) Mean (SD) Missing Weight (kg) Mean (SD) Missing BMI Mean (SD) Missing Cognitive dysfunction Yes No Missing Indication for surgery Primary OA Femoral neck fracture Other arthritic Other fracture Type of hip arthroplasty) Total Hemi Surgical approach Direct lateral Posterolateral

AS group n = 534

PTS group n = 543

135 (25) 399 (75) 82 (8.0)

156 (29) 387 (71) 82 (8.4)

185 (35) 349 (65)

138 (25) 405 (75)

166 (9) 1 (0.2)

167 (9) 6 (1)

67 (15) 1 (0.2)

68 (13) 3 (0.6)

24 (4.5) 1 (0.2)

24 (4.1) 7 (1)

57 (11) 467 (87) 10 (2)

51 (9) 492 (91) 0 (0.0)

124 (23) 383 (72) 16 (3) 11 (2)

94 (17) 421 (78) 17 (3) 11 (2)

248 (46) 286 (54)

211 (39) 332 (61)

354 (66) 180 (34)

411 (76) 132 (24)

PTS group and a 0.5% rate for the AS group. Approximately 250–300 patients are operated yearly in the department with cemented hip arthroplasty and to achieve this sample we included all patients operated 2 years before, and 2 years after the change of implants. Statistics For analysis of the primary outcome, we used Cox proportional hazards with follow-up time defined as time to death, reoperation, or end of follow-up (max. 2 years after surgery). Our main outcome variable was the occurrence of a PPF during the study period and we adjusted for exposure variable (PTS/AS), age, sex, ASA category, cognitive impairment, BMI, whether the indication was fracture or not, and surgical approach. Results are presented as hazard ratios (HRs) with 95% confidence intervals (CI). The statistical analysis is based on the assumption that the studied observations are independent; therefore, no bilateral fractures were included. In patients with 2 fractures during the study period, only the 1st fracture was included. All continuous variables were left as continuous but checked for non-linearity using ANOVA. We investigated the proportional hazards assumption using Grambsch and Therneau

analysis of Schoenfeld residuals. All analyses were performed using R 3.5.2 (R Project for Statistical Computing, Vienna, Austria), using the rms package (v. 5.1-3) for survival modelling, knitr (v. 1.21) for reproducible research, ggplot2 for plots (v. 3.1.0) and Gmisc (v. 1.8) with Greg (v. 1.3) for table output. Ethics, funding, and potential conflicts of interest The study was conducted in accordance with the ethical principles of the Helsinki Declaration and was approved by the Ethics Committee of Karolinska Institutet (entry number dnr 2013/285-31/2). According to the ethical permission, individual consent was not needed from the patients in this observational cohort. The study was funded by the regional agreement on medical training and clinical research (ALF) between Stockholm County Council and Karolinska Institutet and by a research grant from LINK. None of the funding bodies had any input into the data collection, analysis, or conclusions from the study. The authors declare no competing interests.

Results Study subjects and descriptive data Of 2,007 hip arthroplasties performed at our institution during the inclusion period, 1,077 cemented arthroplasties in 1,077 patients were included in the final study cohort after exclusion of uncemented stems and 42 bilateral cases of cemented stems. The mean age at primary surgery was 82 years (49–102), 73% of the patients were female, and 75% underwent surgery for a femoral neck fracture. The baseline demographics were similar between the groups (Table 1). The 1- and 2-year mortality was 16% (n = 173) and 24% (n = 263), respectively with similar 2-year mortality between the groups (HR 1.1, CI 0.8–1.4). Outcome data 1.9% (n = 20) PPFs were identified during 2-year follow-up. The PPFs occurred at a median of 2 months (0.2–23) after primary surgery. The PPF rate was 3.3% (n = 18) in the PTS group and 0.4% (n = 2) in the AS group (Figure 1). The fracture rate was higher for patients operated due to fracture in comparison with those with degenerative hip disease, 2.2% (n = 18) vs. 0.8% (n = 2). It was also generally higher for men, ASA category 3–4, cognitive dysfunction, posterolateral approach, and the use of the CPT stem (Table 2). However, in the multivariable Cox proportional hazard regression only cognitive dysfunction (HR 3.8, CI 1.4–10) and the type of stem (PTS as denominator, HR 0.1, CI 0.0–0.5) were statistically significant. The overall complication rate including PPFs was 6.7% with 8.8% (n = 48) in the PTS group and 4.5% (n = 24) in the AS group (p = 0.004). In the PTS group, PPF was the most common complication (3.3%) followed by dislocation (2.6%) and periprosthetic joint infection (2.2%). In the AS group,

430

Acta Orthopaedica 2019; 90 (5): 427–432

Cumulative periprosthetic fracture rate (%) 8 Polished tapered stem Anatomic stem 6

Variable

0 0

0.5

1.5

Years after index operation

Figure 1. Cox regression of cumulative periprosthetic fracture rate after surgery adjusted for age, sex, cognitive dysfunction, BMI, indication for surgery, and surgical approach. Values are percentage at risk

Table 2. Cox proportional hazard regression crude and adjusted models Association with periprosthetic fracture presented as Hazard ratio (HR)

Years after index operation 0 0.5 1 1.5 2

Total

PPF n (%)

Crude Adjusted HR (95 %Cl) HR (95 %C)

Age, mean (SD) 82 (8) 1.0 (0.9–1.1) 1.0 (0.9–1.0) Sex Male 291 9 (3.1) 1.0 ref. 1.0 ref. Female 786 11 (1.4) 0.4 (0.2–1.0) 0.4 (0.1–0.9) ASA category 1–2 323 2 (0.6) 1.0 ref. 1.0 ref. 3–4 754 18 (2.4) 4.5 (1.0–19) 3.0 (0.7–14) Cognitive dysfunction No 959 14 (1.5) 1.0 ref. 1.0 ref. Yes 108 6 (5.6) 4.4 (1.7–12) 3.8 (1.4–10) BMI, mean (SD) 24 (4) 0.9 (0.8–1.0) 0.9 (0.8–1.0) Indication for surgery Degenerative hip 251 2 (0.8) 1.0 ref. 1.0 ref. Fracture 826 18 (2.2) 3.2 (0.7–14) 3.8 (0.6–24) Approach Direct lateral 765 15 (2.0) 1.0 ref. 1.0 ref. Posterolateral 312 5 (1.6) 0.7 (0.3–2.0) 3.2 (0.9–12) Group PTS 534 2 (0.4) 1.0 ref. 1.0 ref. AS 543 18 (3.3) 0.1 (0.0–0.5) 0.1 (0.0–0.5)

PTS 100 83 78 74 69 AS 100 85 81 77 67

periprosthetic joint infection was the most common reason for revision surgery (1.7%), followed by dislocation (1.1%) and PPF (0.4%). Periprosthetic fractures 70% of PPFs were Vancouver type B2 (n = 12) and type C fractures (n = 2); none of the hips had any radiographic sign of loosening of the stem or periprosthetic osteolysis before fracture. All were sustained through low-energy falls. 13 of the PPFs were treated with stem revision. 12 of 20 PPFs had a good outcome according to the previous definition (Table 3).

Discussion In this prospective, observational cohort study on an elderly cohort of patients comparing a collarless polished tapered stem with an anatomic stem, we have shown that it is possible to reduce the periprosthetic fracture rate dramatically for an orthopedic department by changing the standard implant used for hip arthroplasty patients. As we included all patients during the study period, this study therefore includes the whole department’s learning curve with the new stem. The main strength of our study is the prospective design with an isolated change of stem implant without a change in catchment area, surgeons, indication for surgery, or other routines at an orthopedic department. The completeness of data on the incidence of surgically treated PPF and the homogeneity of implant choice are other strengths.

Table 3. Periprosthetic fractures, surgical treatment and surgical outcome Factor

PTS AS

Vancouver classification A 2 1 B1 3 0 B2 12 0 B3 0 0 C 1 1 Surgical treatment: Open reduction and internal fixation 6 1 Stem revision 12 a 1 Surgical outcome Good 10 2 Intermediate 4 0 Poor 4 0 a 10 of 12 Vancouver B2 fractures were treated with stem revision.

The limitations of this study are the lack of randomization for stem type, leading to a risk of confounding variables in spite of the fact that we that used adjusted regression models. Based on our department’s indication for surgery, these results only apply and are limited to an elderly cohort of patients with degenerative hip disease and femoral neck fracture. The limited sample size and the short follow-up time are counterbalanced by the study design. There is still a need for prospective cohort studies due to risk of under-reporting reoperations to the Swedish hip arthroplasty registry of those PPFs treated with open reduction and internal fixation without change of

Acta Orthopaedica 2019; 90 (5): 427–432

the implant (Thien et al. 2014, Swedish hip arthroplasty registry. Annual report 2017). In concordance with previous studies from our department, we found a high incidence of early PPFs associated with the CPT stem (Brodén et al. 2015, Mukka et al. 2016). It seems that the CPT stem, designed to subside in the cement mantle with axial load, acts as a wedge breaking the femur after a direct hip contusion—a concept discussed by Sarvilinna et al. (2004). The tapered stem generates a stress riser, which in turn splits the femur into complex periprosthetic Vancouver B fractures. We found 12 of the 18 PPFs in the proximity of the CPT stem, generating Vancouver B2 fractures necessitating stem revision using longer cemented or uncemented implants. The standard length (130 mm) of the CPT stem is shorter than the most commonly used version in Sweden of the Lubinus SP2 stem (150 mm). Longer cemented stems anchored distally in harder diaphyseal bone might conceptually reduce this risk further. This is supported by mechanical studies showing that shorter stems have a decreased resistance to torque forces (Bishop et al. 2010, Morishima et al. 2014). However, the shorter 130 mm Lubinus SP2 stem has shown a good longterm outcome without any increased risk for PPF, in a younger population treated mainly for osteoarthritis (Prins et al. 2014). The anatomical shape of the Lubinus SP2 provides a prerequisite for a homogenic cement mantle around the prosthesis which, in turn, reduces the risk of contact between the tip of the prosthesis and cortical bone. Several studies have cited risk factors predisposing to PPF such as sex, advanced age, ASA 3 or 4, osteoporosis, surgical approach, and type of implant (Sarvilinna et al. 2004, Franklin and Malchau 2007, Cook et al. 2008, Jasvinder et al. 2013). We did not find any association between sex and the risk for PPF. Inngul and Enocson (2015) described an increased risk for PPF in men while other studies found a higher proportion of PPF among women (Franklin and Malchau 2007, Sheth et al. 2013). Age has frequently been proposed as a risk factor for PPF; elderly patients are at a higher risk for osteoporosis and frequent falls, which in turn predisposes to PPF (Franklin and Malchau 2007). Cook et al. (2008) described lower risk of fracture for patients below the age of 70 and the highest among those above 80 years of age. Rheumatoid arthritis is a risk factor for PPF, because of a decrease in bone mineral density (Haddad et al. 1999). We could not confirm ASA class 3 or higher as a risk factor for PPF (Singh et al. 2013). In concordance with our findings, Sarvilinna et al. (2004) and Jasvinder et al. (2013) did not find any association between BMI and PPFs after THR. It has been suggested that the surgical approach used could alter the rate of PPF. A higher risk has been proposed for the direct lateral approaches due to a higher incidence of anteroposterior malalignment in the sagittal view (Garellick et al. 1999, Lindahl et al. 2006a). A very recently published registry-based study by Chatziagorou et al. (2019) found an

431

increased risk associated with the posterior approach. The differences in implantation of the stems might affect the loading and behavior of the implant and thus alter the risk for PPF. The possible influence of surgical approach on the risk for PPF needs to be addressed in future studies. In summary, in this prospective cohort study of a change of stem implant at an orthopedic department, the use of an anatomic stem as compared with a collarless polished tapered stem resulted in a substantially lower rate of periprosthetic femoral fractures without increasing other complications such as dislocation. In hospitals where elderly patients with poor bone quality are operated, the switch to use of cemented anatomic stems is to be recommended. JM analyzed the data and wrote the manuscript. SM supervised JM, analyzed the data, and wrote the manuscript. GC and CJH operated patients and wrote the manuscript. MG analyzed data and reviewed the manuscript. OS initiated the study, collected data, supervised JM, operated patients, and wrote the manuscript. Acta thanks Anders Enocson and Geert Meermans for help with peer review of this study.

Bhattacharyya T, Chang D, Meigs J B, Estok D M, Malchau H. Mortality after periprosthetic fracture of the femur. J Bone Joint Surg Am 2007; 89(12): 2658-62. Bishop N E, Burton A, Maheson M, Morlock M M. Biomechanics of short hip endoprostheses: the risk of bone failure increases with decreasing implant size. Clin Biomech 2010; 25(7): 666-74. Brady O H, Garbuz D S, Masri B A, Duncan C P. The reliability and validity of the Vancouver classification of femoral fractures after hip replacement. J Arthroplasty 2000; 15(1): 59-62. Brodén C, Mukka S, Muren O, Eisler T, Boden H, Stark A, Skoldenberg O. High risk of early periprosthetic fractures after primary hip arthroplasty in elderly patients using a cemented, tapered, polished stem. Acta Orthop 2015; 86(2): 169-74. Chatziagorou G, Lindahl H, Kärrholm J. The design of the cemented stem influences the risk of Vancouver type B fractures, but not of type C: an analysis of 82,837 Lubinus SPII and Exeter Polished stems. Acta Orthop 2019; 90 (2): 135-42. Cook R E, Jenkins P J, Walmsley P J, Patton J T, Robinson C M. Risk factors for periprosthetic fractures of the hip: a survivorship analysis. Clin Orthop Relat Res 2008; 466(7): 1652-6. Franklin J, Malchau H. Risk factors for periprosthetic femoral fracture. Injury 2007; 38(6): 655-60. Dorr L D, Faugere M C, Mackel A M, Gruen T A, Bognar B, Malluche H H. Structural and cellular assessment of bone quality of proximal femur. Bone 1993; 14: 231–42. Garellick G, Malchau H, Herberts P. The Charnley versus the Spectron hip prosthesis: clinical evaluation of a randomized, prospective study of 2 different hip implants. J Arthroplasty 1999; 14(4): 407-13. Haddad F S, Masri B A, Garbuz D S, Duncan C P. The prevention of periprosthetic fractures in total hip and knee arthroplasty. Orthop Clin North Am 1999; 30(2): 191-207. Inngul C, Enocson A. Postoperative periprosthetic fractures in patients with an Exeter stem due to a femoral neck fracture: cumulative incidence and surgical outcome. Int Orthop 2015; 39(9): 1683-8. Jasvinder A. S, Matthew J, Scott H, David L. Are gender, comorbidity and obesity risk factors for postoperative periprosthetic fractures following primary total hip replacement? J Arthroplasty 2013; 28(1): 126–31.

432

Kristensen T B, Dybvik E, Furnes O, Engesæter L B, Gjertsen J E. More reoperations for periprosthetic fracture after cemented hemiarthroplasty with polished taper-slip stems than after anatomical and straight stems in the treatment of hip fractures. Bone Joint J 2018; 100B(12):1565-71. Lindahl H, Malchau H, Herberts P, Garellick G. Periprosthetic femoral fractures: classification and demographics of 1049 periprosthetic femoral fractures from the Swedish national hip arthroplasty register. J Arthroplasty 2005; 20(7): 857-65. Lindahl H, Garellick G, Regner H, Herberts P, Malchau H. Three hundred and twenty-one periprosthetic femoral fractures. J Bone Joint Surg Am 2006a; 88(6): 1215-22. Lindahl H, Malchau H, Oden A, Garellick G. Risk factors for failure after treatment of a periprosthetic fracture of the femur. J Bone Joint Surg Br 2006b; 88(1): 26-30. Morishima T, Ginsel B L, Choy G G, Wilson L J, Whitehouse S L, Crawford R W. Periprosthetic fracture torque for short versus standard cemented hip stems: an experimental in vitro study. J Arthroplasty 2014; 29(5): 1067-71. Mukka S, Mellner C, Knutsson B, Sayed-Noor A, Sköldenberg O. Substantially higher prevalence of postoperative periprosthetic fractures in octogenarians with hip fractures operated with a cemented, polished tapered stem rather than an anatomic stem. Acta Orthop 2016; 87(3): 257-61. Palan J, Smith M C, Gregg P, Mellon S, Kulkarni A, Tucker K, Blom A W, Murray D W, Pandit H. The influence of cemented femoral stem choice on the incidence of revision for periprosthetic fracture after primary total hip arthroplasty: an analysis of national joint registry data. Bone Joint J 2016; 98-b(10): 1347-54.

Acta Orthopaedica 2019; 90 (5): 427–432

Prins W, Meijer R, Kollen B J, Verheyen C C, Ettema H B. Excellent results with the cemented Lubinus SP II 130-mm femoral stem at 10 years of follow-up: 932 hips followed for 5-15 years. Acta Orthop 2014; 85(3): 276-9. Sarvilinna R, Huhtala H S, Sovelius R T, Halonen P J, Nevalainen J K, Pajamaki K J. Factors predisposing to periprosthetic fracture after hip arthroplasty: a case-control study. Acta Orthop Scand 2004; 75(1): 16-20. Sheth N P, Brown N M, Moric M, Berger R A. Della Valle C J. Operative treatment of early periprosthetic femur fractures following primary total hip arthroplasty. J Arthroplasty 2013; 28(2): 286–91. Singh J A, Jensen M R, Harmsen S W, Lewallen D G. Are gender, comorbidity, and obesity risk factors for postoperative periprosthetic fractures after primary total hip arthroplasty? J Arthroplasty 2013; 28(1): 126-31 e1-2. Swedish Hip Arthroplasty Registry. Annual report 2017. Available at https:// shpr.registercentrum.se/nyheter/annual-report-2017 (Accessed February 23, 2019). Sköldenberg O, Ekman A, Salemyr M, Boden H. Reduced dislocation rate after hip arthroplasty for femoral neck fractures when changing from posterolateral to anterolateral approach. Acta Orthop 2010; 81(5): 583-7. Thien T M, Chatziagorou G, Garellick G, Furnes O, Havelin L I, Makela K, Overgaard S, Pedersen A, Eskelinen A, Pulkkinen P, Karrholm J. Periprosthetic femoral fracture within two years after total hip replacement: analysis of 437,629 operations in the nordic arthroplasty register association database. J Bone Joint Surg Am 2014; 96(19): e167. Young S W, Walker C G, Pitto R P. Functional outcome of femoral periprosthetic fracture and revision hip arthroplasty: a matched-pair study from the New Zealand Registry. Acta Orthop 2008; 79(4): 483-8.

Acta Orthopaedica 2019; 90 (5): 433–438

433

No effect of delivery on total hip replacement survival: a nationwide register study in Finland Ilari KUITUNEN 1, Eerik T SKYTTÄ 2, Miia ARTAMA 3,4, Heini HUHTALA 3, and Antti ESKELINEN 2 1 Faculty of Medicine and Health Technologies, Tampere University, Tampere; 2 Coxa Hospital for Joint Replacement, and Faculty of Medicine and Health Technologies; 3 Faculty of Social Sciences, Tampere University, Tampere; 4 National Institute of Health and Welfare, Tampere, Finland Correspondence: ilari.kuitunen@tuni.fi Submitted 2019-03-29. Accepted 2019-05-21

Background and purpose — Previous small studies have suggested that delivery does not adversely affect the survivorship of total hip replacement (THR). We investigated whether delivery after primary THR affects hip implant survivorship in a large population-based study sample Patients and methods — In this register-based nationwide cohort study, all women aged 15–45 who underwent primary THR in Finland from 1987 to 2007 were included from the Finnish Arthroplasty Register. Data on deliveries were obtained from the medical birth register. After primary THR, 111 women (133 THRs) delivered and formed the delivery group. In the reference group, 1,878 women (2,343 THRs) had no deliveries. We used Kaplan–Meier analysis with 95% confidence intervals (CI) to study implant survivorship at 6 and 13 years, and Cox multiple regression to assess survival and hazard ratios (HRs), with revision for any reason as an endpoint with adjustment for age, rheumatoid arthritis, and stem and cup fixation. Results — 51 (38%) revisions were recorded in the delivery group and 645 (28%) revisions in the reference group. The 6-year implant survivorship was 91% (CI 85–96) in the delivery group and 88% (CI 87–90) in the reference group. The 13-year survival rates were 50% (CI 39–62) and 61% (CI 59–64). The adjusted HR for revision after delivery was 0.7 (CI 0.4–1.2) in ≤ 6.8 years’ follow-up and 1.1 (CI 0.8–1.6) in > 6.8 years’ follow-up. Interpretation — Based on the findings in this nationwide study of hip replacement in fertile-aged women, delivery does not seem to decrease THR implant survivorship; women should not be afraid of or avoid becoming pregnant after THR.

The most common indications for THR in very young patients aged under 30 years are rheumatoid arthritis (RA), avascular necrosis of the femoral head, and developmental dysplasia of the hip (Adelani et al. 2013). The incidence of primary THR among young patients (30 to 59 years old) has increased annually in Finland from 9.5 per 100,000 person years in 1980 to 61 per 100,000 in 2007 (Skyttä et al. 2011). In 2017, over 1,000 women aged under 55 underwent a primary THR operation in Finland (open access statistical report of the Finnish Arthroplasty Register 2018: National Institute of Health and Welfare 2018). Only a few studies with rather small sample sizes and local data have analyzed the effects of delivery and THR on each other. None of these studies have reported problems with deliveries after THR, and they indicate that THR does not majorly affect the mode of delivery (Monaghan et al. 1987, Boot et al. 2003, Meldrum et al. 2003, Yazici et al. 2003, Sierra et al. 2005, Stea et al. 2007, Smith et al. 2008) Further, THR survival is not decreased, and the delivery method does not affect THR survival (Meldrum et al. 2003, Sierra et al. 2005). However, women have reported concerns regarding vaginal delivery and fear of delivery positions harming the THR (Ostensen 1993, Meldrum et al. 2003, Stea et al. 2007). Very young patients seem to have worse clinical outcomes in terms of pain relief and function after THR, even though implant survival rates and radiological outcomes have improved (Adelani et al. 2013, Swarup et al. 2017). Clinical outcomes may be limited by systemic diseases, such as RA, that still comprise the majority of indications for THR in these very young patients. The survivorship of the THR is often shortened due to the loosening of cup or stem in very young patients, men, and patients with a higher BMI (Melloh et al. 2011). While in some studies underlying diseases have not negatively affected the survival of the hip prosthesis (Hannouche et al. 2016), dysplastic hips appear to have worse sur-

434

vival rates compared with non-dysplastic hips (Tsukanaka et al. 2016). Metal-on-metal (MoM) implants have worse survival rates compared with non-MoM implants and are since 2012 are no longer used in Finland due to common adverse local tissue reactions that have led to numerous revisions (Smith et al. 2012, Furnes et al. 2014, Varnum et al. 2015). Because THR implant survival is substantially lower in very young patients compared with older patients, THR should be considered as the treatment option of last resort for very young patients (Swarup et al. 2015, Hannouche et al. 2016). We evaluated whether delivery adversely affects the survivorship of THR in a nationwide register-based study sample.

Patients and methods Data for this nationwide register-based study were gathered from 3 different national registers. Information on all women aged 15 to 45 who underwent THR operation in Finland between 1987 and 2007 was obtained from the Finnish Arthroplasty Register (FAR). The register is maintained by the National Institute for Health and Welfare (THL), and it contains information on all orthopedic prostheses operated from 1980 in Finland. All the information in the FAR has been collected prospectively. The current (2017) completeness of the register is 95% for primary THR, and it matches well with data from the Finnish Hospital Discharge Register (open access statistical report of the Finnish Arthroplasty Register 2018: NIHW 2018). In the present study, the operation day of the primary THR was used as the starting point of the follow-up. Because we did not have information on primary THR operations before 1987, a revision THR as the first event in the FAR after January 1, 1987 was an exclusion criterion in the study. Women with bilateral prostheses were included, as earlier research has shown that this does not bias the results (Lie et al. 2004, Ranstam and Robertsson 2010). The endpoint for the followup was either revision, death, emigration, or December 31, 2007, whichever came first. The outcome was the revision of the hip for any reason. 2,012 women with 2,499 primary THRs were selected from the register. Of the THRs selected, 23 were excluded due to a lack of information on many key variables (Figure 1). Information on pregnancies and deliveries was gathered from the National Medical Birth Register (MBR) maintained by the THL. Pregnancies and deliveries from January 1, 1987 to December 31, 2007 were included in this study. The MBR contains information on all pregnancies of at least 22 gestational weeks ending in delivery and information on deliveries and newborns. MBR data match well with hospital discharge data and the coverage of the register has improved over the years. If there was no information in the MBR, the woman was not considered to have been pregnant. In the study, women who had given birth after THR formed the delivery

Acta Orthopaedica 2019; 90 (5): 433–438

Women aged 15–45 with primary THR 1987–2007 in the Finnish Arthroplasty Register n = 2,012 Excluded missing data n = 23

YES

Included n = 1,989 Delivery after THR

Delivery group 111 women 133 THRs 51 revisions

Reference group 1,878 women 2,343 THRs 645 revisions

Figure 1. Flow chart of study population and events of total hip replacement (THR) survival among fertile-aged (15 to 45) women having delivery compared with women not having delivery after THR.

group, and women without pregnancy after THR formed the reference group. The Register for Reimbursable Diseases is maintained by the Finnish Social Insurance Institution of Finland. It contains information on reimbursable chronic diseases. A medical statement written by a certified doctor is needed to gain reimbursement for chronic disease. Information on all the reimbursements for this study population was obtained. If there was no information available, women were considered as not having chronic diseases. In this study, RA was the most common diagnosis. Other chronic diseases were rare, but the following diseases were found: asthma, diabetes mellitus type 1, epilepsy, hypothyroidism, hypertension, inflammatory bowel disease, and major psychiatric disease. Statistics Categorized variables were compared by chi-square test between the groups and reported as proportions. Continuous variables were compared by their distribution. Normally distributed variables were compared by Student’s t-test and reported by means with standard deviations (SD). Non-normally distributed variables were compared by Mann–Whitney U-test and reported by medians with interquartile range. A p-value under 0.05 was considered statistically significant in all analyses. Kaplan-Meier survival analyses with 95% confidence intervals (CI) were performed to evaluate the survival of the hips in both the delivery group and the reference group. Survival rates were calculated for 6 years’ and 13 years’ follow-up. The follow-up was continued until 13 years when 20 THRs were still at risk (life table analysis) in the delivery group. The follow-up period was calculated from primary THR until revision THR or until the date the patient was censored at the end of the study (December 31, 2007), or date of emigration, or date of death. The Cox proportional hazards model was used to analyze the effect of potential confounders and count hazard ratios (HR). The adjustments

Acta Orthopaedica 2019; 90 (5): 433–438

435

Table 1. Background characteristics of the study population, types of hip prosthesis, and indications for revisions between the delivery group and the reference group. Values are frequency (%) unless otherwise specified Factor

Delivery group Reference group n = 133 n = 2,343

Age at primary THR a 29 (8) Follow-up period (years) a 9.1 (6) Rheumatoid arthritis 62 (47) Other chronic disease b 5 (4) Nulliparous at primary THR 78 (64) Indication for THR Inflammatory arthritis (RA + others) 62 (47) Primary osteoarthritis 12 (9) Secondary arthrosis 21 (16) DDH c 22 (17) Other 16 (12) Metal-on-metal bearing 16 (12) Type of primary THR fixation Uncemented 114 (86) Hybrid 7 (5) Inverse hybrid 0 (0) Cemented 12 (9) Revisions 51 (38) Revision indications Aseptic loosening 30 (59) Deep infection 1 (2) Periprosthetic fracture 0 (0) Dislocation 1 (2) Others 14 (27) Missing 5 (10)

40 (8) 8.0 (8) 774 (33) 208 (9) 778 (33) 731 (31) 532 (23) 363 (16) 493 (21) 224 (10) 390 (17) 1,859 (79) 237 (10) 1 (0) 245 (10) 645 (28) 318 (50) 11 (2) 12 (2) 30 (5) 193 (30) 81 (12)

a Median and interquartiles. b Includes: asthma, diabetes

mellitus type 1, epilepsy, hypothyroidism, hypertension, inflammatory bowel disease, major psychiatric disease. c DDH = developmental dysplasia of the hip.

used in the Cox proportional analysis were the following: age at the time of primary THR, RA, stem fixation, and cup fixation. Because the proportional hazards assumption was not met in the Cox model (crossing survival curves at 6.8 years), the followup was divided into 2 time periods, and a piecewise Cox proportional model was performed. The first follow-up period was the time before the crossing at 6.8 years, and the second period was from the crossing until the end of the follow-up (6.8–21.0 years). All the analyses were performed using SPSS statistical software version 25.0 (IBM Corp, Armonk, NY, USA). Ethics, registration, funding, and potential conflicts of interest In accordance with Finnish regulations, informed patient consent was not required as the women were not contacted. Our study protocol went through the ethical evaluation of the National Institute for Health and Welfare to gain access to register data, permission number: THL/599/5.05.00/2010. This study was funded by the Competitive Research funds of Pirkanmaa Hospital District, Tampere, Finland, representing governmental funding. The authors have no potential conflicts of interests to declare.

Table 2. Comparison of primary diagnoses and revision indications in the delivery group between women with at least 1 vaginal delivery after total hip replacement (THR) with women with only Cesarean sections after THR Factor

Vaginal delivery Cesarean section after THR after THR n = 53 n = 80

Indication for THR Inflammatory arthritis (RA + others) 19 43 Primary osteoarthritis 4 8 Secondary arthrosis 13 8 DDH 10 12 Other 7 9 Revisions 15 36 Revision indications: Aseptic loosening 10 20 Deep infection 1 0 Dislocation 0 1 Others 4 10 Missing 0 5 DDH = developmental dysplasia of the hip.

Results 1,989 women with 2,476 THRs were included in the study (Table 1). Of these, 111 (5.6%) women with 133 (5.4%) THRs had a delivery during the follow-up. The mean followup in the delivery group was 9.3 years (0–21), and the median age at the start of the follow-up was 29 years. In the reference group, 1,878 women with 2,343 THRs had no deliveries. The mean follow-up was 8.1 years (0–21), and the median age at the start of the follow-up was 40. RA was the most common indication for THR in both groups. It was, however, more prevalent in the delivery group (47%) than in the reference group (33%) (p = 0.001). Other chronic diseases were more common in the reference group. The distribution of THR fixation method or bearing type was similar between the groups. The delivery group had 51 revisions, and 30 (59%) of the revisions were performed due to aseptic loosening. In the reference group, 645 THRs were revised, and 318 (49%) revisions were performed due to aseptic loosening. The deliveries were analyzed and recorded per THR. 170 deliveries occurred during the follow-up (mean of 1.3 deliveries per THR). The maximum number of deliveries per patient during the follow-up was 5. Of the deliveries, 75 (44%) were vaginal and 95 (56%) Cesarean sections. 50 women with 53 THRs had at least 1 vaginal delivery after THR and 61 women with 80 THRs had only Cesarean sections after THR. The primary THR diagnoses and revision indications were similar in the vaginal delivery group and Cesarean section group (Table 2). At 6 years, the survival rate in the delivery group was 91% (CI 85–96) and in the reference group 88% (CI 87–90). At 13

436

Acta Orthopaedica 2019; 90 (5): 433–438

Revisionfree survival (%) 100 Delivery group Reference group 80

Table 3. Kaplan–Meier 6- and 13-year survival rates with 95% confidence intervals (CI) of primary total hip replacement of fertile-aged women aged 15 to 45 years at the time of THR No. of No. of Delivery hips revisions

K–M survivorship at 6 years No. at survival risk % (CI)

Yes No

100 91 (85–96) 22 50 (39–62) 1,411 88 (87–90) 456 62 (59–64)

133 2,343

51 645

K–M survivorship at 13 years No. at survival risk % (CI)

0 0

Years after index operation

Figure 2. Kaplan–Meier survival curves (with 95% confidence intervals) of primary total hip replacement (THR) among fertile-aged women aged 15 to 45 years at the time of THR having 1 or more deliveries after THR (delivery group) compared with no deliveries after THR (reference group).

years, the survival rate was 50% (CI 39–62) for the delivery group and 61% (CI 59–64) for the reference group, respectively (Figure 2, Table 3). During the first time period (0 to 6.8 years’ follow-up), the adjusted Cox regression model showed no statistically significant difference in the risk for revision between the delivery and the reference groups (adjusted HR 0.7, CI 0.4–1.2). During the later follow-up (6.8 to 21 years), there was still no difference in adjusted HR between the groups (HR 1.1, CI 0.8–1.6).

Discussion To our knowledge, our study is the first to assess THR implant survivorship in fertile-aged women in a large populationbased study sample. Based on our results, delivery does not seem to adversely affect hip implant survivorship after primary THR. Our results are in concordance with previous smaller studies. In their study, Sierra et al. (2005) reported that delivery after primary THR does not decrease the survival rate of the implant. They had the largest number of participants prior to our study. 343 women with 420 THR were contacted and 47 of those had pregnancy ending in delivery. However, the survival rates for 5-, 10- and 15-year follow-up periods were calculated for the whole cohort with no comparisons made between the delivery and non-delivery groups. Our 6-year survival rate in both groups was in line with these results. Meldrum et al. (2003) had 13 hips with deliveries in their study population and reported no adverse effects for THR. Yazici et al. (2003) reported 21 THR patients with deliveries and no decrease in the survival rate of the THR. All these studies were retrospective with alternative response rates (30–75%). McDowell and Lachiewicz (2001) reported 5 women with 7 uncemented

THRs having deliveries and, compared with matched referents, no differences between survival or hip functions were reported. Our study was the only one to report a slight but not statistically significant decrease in implant survival rate in the Kaplan–Meier analysis after delivery. Cesarean section (CS) rate was markedly increased in the delivery group compared with overall CS rate in Finland. There have been previous reports in which women with dysplastic hips have been discussed to have smaller pelvic diameters and therefore could tend to have CS (Sierra et al. 2005; Stea et al. 2007). Developmental dysplasia of the hip was an equally common indication for THR in women who only had Cesarean sections after THR as in those who delivered vaginally after THR in our study. Also, revision indications did not differ between them. The reason for the very high CS rate in the delivery group remains unknown. We can only speculate that the presence of THR may have affected the patients’ and/or the physicians’ choice of delivery. However, it did not have any effect on THR survival rates. Age was the only statistically significant variable that negatively affected THR implant survivorship. The delivery group’s median age at the start of the follow-up was 29 years compared with the reference group’s 40 years. Previously, only Sierra et al. (2005) have applied the Cox regression model to analyze implant survivorship after delivery. In their model, delivery seemed to decrease THR survivorship, but once age at the time of primary THR was taken as part of the model, no further differences between the delivery group and the reference group were obtained. Previous non-delivery-related THR survival studies have reported similar findings of weaker implant survivorship in younger patients (Dorr et al. 1994, Nam et al. 2016, Tsukanaka et al. 2016). In particular, very young patients under 30 years have been reported to have had decreased THR survivorship (Mohaddes et al. 2019) probably because of higher activity levels (Dorr et al. 1994, Adelani et al. 2013). Our survival rates were slightly lower compared with the recent study of Mohaddes et al. (2019), in which the 15-year THR survival rate for patients aged under 30 at the time of THR was 76%. In young patients (< 50 years or less), indications for THR differ in comparison with older patients (+50 years). In younger patients, inflammatory arthritis and developmental hip diseases are more common, and primary osteoarthritis is

Acta Orthopaedica 2019; 90 (5): 433–438

rare (Adelani et al. 2013). Developmental dysplasia of the hip decreases the survival of the hip prosthesis in young patients (Havelin et al. 2000, Tsukanaka et al. 2016). There have been controversial results regarding the survival of the THR in RA patients. Some studies have suggested decreased THR survivorship, more common radiographic findings indicating implant failure, poorer function, and increased mortality among patients with RA (Creighton et al. 1998, Havelin et al. 2000, Tang and Chiu 2001, Singh and Lewallen 2013, Goodman et al. 2014, Schrama et al. 2015). Inflammatory arthritis as primary diagnosis for THR may also increase revisions due to deep infections (Dale et al. 2012). Previous large national cohort studies, however, have shown no decrease in THR survival due to RA (Havelin et al. 2000 , Furnes et al. 2001, Eskelinen et al. 2006). Because of the high prevalence of RA among young patients, it was taken as part of the Cox model. In our model, RA did not decrease THR survival. Indeed, it seemed patients with RA had better results during the first follow-up period (< 6.8 years). A similar finding was seen in a previous THR and delivery study, where Serra et al. (2005) found no decrease in the survival of hips operated due to an RA diagnosis in their step-by-step Cox results. The main strength of our study is the register-based design. Previous THR survival studies after delivery have been retrospective cohorts with questionnaires. Our design eliminates possible recall bias and has better completeness because revision indications are also reported to the Finnish Arthroplasty Register. In addition, we had by far the largest study population with the longest follow-up, and our results are nationwide instead of from one hospital district catchment area. In addition, we were also able to combine information from several nationwide registers on patients’ long-term diseases and pregnancies. The first limitation of our study is the lack of patient-reported outcome measurements (PROMs), which forced our study to focus strictly on the survival of the implant. However, absence of PROM data does not affect our interpretation of the survival results. The second limitation was the study period. Our study period was from 1987 to 2007. Even though the implants used today differ greatly from those implanted 30 years ago, contemporary implant designs were used in the latter half of the study period, and this approach also enabled us to assess long-term implant survivorship in this rare cohort of patients. In conclusion, based on the findings in this nationwide study offering hip replacement to fertile females, delivery does not seem to decrease THR implant survivorship. Women should not be afraid of or avoid becoming pregnant after THR.

IK, MA, ES, HH, and AE had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. MA, ES, and AE contributed to the conception and design of the study. IK drafted the manuscript. AE supervised the study. HH, IK, and AE were responsible for the statistical analyses. All authors participated in the interpretation of data and in the critical revision of the manuscript.

437

The authors would like to thank Mr. Peter Heath for language editing of the manuscript. Acta thanks Anne Garland and Maziar Mohaddes for help with peer review of this study.

Adelani M A, Keeney J A, Palisch A, Fowler S A, Clohisy J C. Has total hip arthroplasty in patients 30 years or younger improved? A systematic review. Clin Orthop 2013; 471(8): 2595-601. Boot C L, Heyligers I C, Heins K F. Pregnancy and delivery after revised total hip replacement. Orthopedics 2003; 26(8): 813-14. Creighton M G, Callaghan J J, Olejniczak J P, Johnston R C. Total hip arthroplasty with cement in patients who have rheumatoid arthritis: a minimum ten-year follow-up study. J Bone Joint Surg Am 1998; 80(10): 1439-46. Dale H, Fenstad A M, Hallan G, Havelin L I, Furnes O, Overgaard S, et al. Increasing risk of prosthetic joint infection after total hip arthroplasty. Acta Orthop 2012; 83(5): 449-58. Dorr L D, Kane T J, Conaty J P. Long-term results of cemented total hip arthroplasty in patients 45 years old or younger: a 16-year follow-up study. J Arthroplasty 1994; 9(5): 453-6. Eskelinen A, Paavolainen P, Helenius I, Pulkkinen P, Remes V. Total hip arthroplasty for rheumatoid arthritis in younger patients: 2,557 replacements in the Finnish Arthroplasty Register followed for 0–24 years. Acta Orthop 2006; 77(6): 853-65. Furnes O, Lie S A, Espehaug B, Vollset S E, Engesaeter L B, Havelin L I. Hip disease and the prognosis of total hip replacements: a review of 53,698 primary total hip replacements reported to the Norwegian Arthroplasty Register 1987–99. J Bone Joint Surg Br 2001; 83(4): 579-86. Furnes O, Paxton E, Cafri G, Graves S, Bordini B, Comfort T, Rivas M C, Banerjee S, Sedrakyan A. Distributed analysis of hip implants using six national and regional registries: comparing metal-on-metal with metalon-highly cross-linked polyethylene bearings in cementless total hip arthroplasty in young patients. J Bone Joint Surg Am 2014; 96 Suppl 1: 25-33. Goodman S M, Ramsden-Stein D N, Huang W, Zhu R, Figgie M P, Alexiades M M, et al. Patients with rheumatoid arthritis are more likely to have pain and poor function after total hip replacements than patients with osteoarthritis. J Rheumatol 2014; 41(9): 1774-80. Hannouche D, Devriese F, Delambre J, Zadegan F, Tourabaly I, Sedel L, et al. Ceramic-on-ceramic THA implants in patients younger than 20 years. Clin Orthop Relat Res 2016; 474(2): 520-7. Havelin L I, Engesaeter L B, Espehaug B, Furnes O, Lie S A, Vollset S E. The Norwegian Arthroplasty Register: 11 years and 73,000 arthroplasties. Acta Orthop Scand 2000; 71(4): 337-53. Lie S A, Engesaeter L B, Havelin L I, Gjessing H K, Vollset S E. Dependency issues in survival analyses of 55,782 primary hip replacements from 47,355 patients. Stat Med 2004; 23(20): 3227-40. McDowell C, Lachiewicz P. Pregnancy after total hip arthroplasty. J Bone Joint Surg Am 2001; 83(10): 1490-4. Meldrum R, Feinberg J, Capello W, Detterline A. Clinical outcome and incidence of pregnancy after bipolar and total hip arthroplasty in young women. J Arthroplasty 2003; 18(7): 879-85. Melloh M, Eggli S, Busato A, Roder C. Predictors of early stem loosening after total hip arthroplasty: a case-control study. J Orthop Surg 2011; 19(3): 269-73. Mohaddes M, NaucléR E, Kärrholm J, Malchau H, Odin D, Rolfson O. Implant survival and patient-reported outcome following total hip arthroplasty in patients 30 years or younger: a matched cohort study of 1,008 patients in the Swedish Hip Arthroplasty Register. Acta Orthop 2019; 90(3): 249-52. Monaghan J, Lenehan P, Stronge J, Gallagher J. Pregnancy and vaginal delivery following bilateral hip replacement. Eur J Obstet Gynecol Reprod Biol 1987; 26(3): 261-4.

438

Nam D, Nunley R M, Berend M E, Berend K R, Lombardi A V, Barrack R L. Residual symptoms and function in young, active hip arthroplasty patients: comparable to normative controls? J Arthroplasty 2016; 31(7): 1492-7. National Institute of Health and Welfare. Open access statistical report of the Finnish Arthroplasty Register. 2018. Available at http:www.thl.fi/far. Ostensen M. [Hip prostheses in women of fertile age: consequences for sexuality and reproduction]. Tidsskr Nor Laegeforen 1993; 113(12): 1483-5. Ranstam J, Robertsson O. Statistical analysis of arthroplasty register data. Acta Orthop 2010; 81(1): 10-14. Schrama J C, Fenstad A M, Dale H, Havelin L, Hallan G, Overgaard S, et al. Increased risk of revision for infection in rheumatoid arthritis patients with total hip replacements. Acta Orthop 2015; 86(4): 469-76. Sierra R, Trousdale R, Cabanela M. Pregnancy and childbirth after total hip arthroplasty. J Bone Joint Surg Br 2005; 87(1): 21-4. Singh J A, Lewallen D G. Patients with osteoarthritis and avascular necrosis have better functional outcomes and those with avascular necrosis worse pain outcomes compared to rheumatoid arthritis after primary hip arthroplasty: a cohort study. BMC Med 2013; 11: 210. Skyttä E T, Leskinen J, Eskelinen A, Huhtala H, Remes V. Increasing incidence of hip arthroplasty for primary osteoarthritis in 30- to 59-year-old patients. Acta Orthop 2011; 82(1): 1-5. Smith M, Marcus P, Wurtz L. Orthopedic issues in pregnancy. Obstet Cynegol Surv 2008; 63(2): 103-11. Smith A J, Dieppe P, Vernon K, Porter M, Blom A W; National Joint Reg-

Acta Orthopaedica 2019; 90 (5): 433–438

istry of England and Wales. Failure rates of stemmed metal-on-metal hip replacements: analysis of data from the National Joint Registry of England and Wales. Lancet 2012; 379(9822): 1199-204. Stea S, Bordini B, De Clerico M, Traina F, Toni A. Safety of pregnancy and delivery after total hip arthroplasty. J Women’s Health 2007; 16(9): 1300-4. Swarup I, Lee Y, Christoph E I, Mandl L A, Goodman S M, Figgie M P. Implant survival and patient-reported outcomes after total hip arthroplasty in young patients with juvenile idiopathic arthritis. J Arthroplasty 2015; 30(3): 398-402. Swarup I, Shields M, Mayer E N, Hendow C J, Burket J C, Figgie M P. Outcomes after total hip arthroplasty in young patients with osteonecrosis of the hip. Hip Int 2017; 27(3): 286-92. Tang W M, Chiu K Y. Primary total hip arthroplasty in patients with rheumatoid arthritis. Int Orthop 2001; 25(1): 13-16. Tsukanaka M, Halvorsen V, Nordsletten L, Engesaeter I O, Engesaeter L B, Marie Fenstad A, et al. Implant survival and radiographic outcome of total hip replacement in patients less than 20 years old. Acta Orthop 2016; 87(5): 479-84. Varnum C, Pedersen A B, Mäkelä K, Eskelinen A, Havelin L I, Furnes O, Kärrholm J, Garellick G, Overgaard S. Increased risk of revision of cementless stemmed total hip arthroplasty with metal-on-metal bearings. Acta Orthop 2015; 86(4): 491-7. Yazici Y, Erkan D, Zuniga R, Bateman H, Salvati E A, Magid S K. Pregnancy outcomes following total hip arthroplasty: a preliminary study and review of the literature. Orthopedics 2003; 26(1): 75-6.

Acta Orthopaedica 2019; 90 (5): 439–444

439

Low complication rate after same-day total hip arthroplasty: a retrospective, single-center cohort study in 116 procedures Merete N MADSEN, Maria L KIRKEGAARD, Malene LAURSEN, Jens R LARSEN, Merete F PEDERSEN, Birgitte SKOVGAARD, Thomas PRYNØ, and Lone R MIKKELSEN

Elective Surgery Centre, Silkeborg Regional Hospital, Denmark Correspondence: Merete.Madsen@midt.rm.dk Submitted 2018-12-21. Accepted 2019-06-03.

Background and purpose — Length of hospital stay (LOS) following total hip arthroplasty (THA) has been markedly reduced. Recently, same-day THA (SD-THA) was introduced, and previous studies have indicated satisfactory safety. However, studies are heterogeneous and only a few report results on SD-THA when using a posterolateral surgical approach. Thus, our aim was to evaluate the feasibility of and complications after SD-THA when using a posterolateral approach. Patients and methods — Consecutive patients scheduled for SD-THA between October 2015 and June 2016 were included. Eligibility criteria for SD-THA were: primary THA, motivation for same-day procedure, age > 18 years, ASA I or II, and the presence of a support person who could remain with the patient for 24 hours after surgery. A posterolateral surgical approach was used. Data were collected retrospectively from hospital records and the Danish National Patient Registry. Outcome measures were: complications during admission, LOS, causes of prolonged admission, and prevalence and causes of readmission at 90 days’ follow-up. Results — 102 of 116 (88 %) patients scheduled for SD-THA were discharged on the day of surgery. The remaining 14 patients were discharged the following day. Primary causes of prolonged admission were: dizziness/nausea, pain, and wound seepage. 7 patients had an estimated blood loss above 400 mL, but all were discharged as planned. No major complications occurred during admission. At follow-up, 3 patients had been readmitted due to pneumonia, wound infection, and dislocation, respectively. Interpretation — The results indicate that SD-THA performed with a posterolateral approach is feasible and can be performed with a low complication rate in a selected group of patients.

Fast-track regimens have been shown to markedly reduce length of hospital stay (LOS) after total hip arthroplasty (THA) (Husted et al. 2012) without compromising morbidity and mortality (Husted 2012, Glassou et al. 2014). In recent years, patient discharge on the day of surgery (SD-THA) has been introduced, and newly published systematic reviews have concluded that outpatient arthroplasty (THA as well as TKA) may be a safe procedure in selected patients (Pollock et al. 2016, Kort et al. 2017, Hoffmann et al. 2018, LovettCarter et al. 2018). However, Pollock et al. (2016) also stated, that the majority of studies in the review were of poor quality. Additional studies, not included in the systematic reviews, support the findings of feasibility of SD-THA in selected patient groups (Parcells et al. 2016, Larsen et al. 2017, Klein et al. 2017, Berend et al. 2018, Kim et al. 2018, Toy et al. 2018) and to a lesser degree in unselected patients (Gromov et al. 2017). In addition, studies have reported satisfactory results regarding safety (Parcells et al. 2016, Basques et al. 2017, Courtney et al. 2017, Klein 2017, Larsen et al. 2017, Nelson et al. 2017, Berend 2018, Toy et al. 2018). Conversely, higher rates of postoperative complications, primarily medical complications, have been reported in 2 database studies comparing non-randomized groups of SD-THA patients and THA inpatients (Lovecchio et al. 2016, Otero et al. 2016). Despite the already mentioned promising results, further research is needed. Available studies are heterogenous, i.e., due to differences in health care set-up, patient selection criteria, surgical approach, and duration of follow-up. Studies reporting on the posterolateral surgical approach are few (Gromov et al. 2017, Larsen et al. 2017, Springer et al. 2017) and only with a short follow-up period of 6 weeks or less. Thus, we evaluated feasibility of and complications after SD-THA when using a posterolateral approach and with 90 days’ follow-up.

440

Acta Orthopaedica 2019; 90 (5): 439â&#x20AC;&#x201C;444

Primary THA during study period n = 669

Patients and methods Design The study was a retrospective cohort study conducted at a public hospital with free access and with patients being treated free of charge. Participants All patients scheduled for SD-THA between October 2015 and June 2016 were included in the study. Eligibility criteria were: primary THA, motivation for and acceptance of a sameday procedure, age above 18 years, and a health condition categorized as ASA class I or II. In addition, an adult support person, who was capable of providing emotional and physical support to the patient during the first 24 hours after surgery, had to be available. The clinical pathway is presented in Table 1, see Supplementary data. Data collection Data were retrospectively retrieved from hospital records and the Danish National Patient Registry (DNPR) (Lynge et al. 2011). Data were collected from admission until 90 days after discharge. The total number of patients receiving a THA during the study period was retrieved from the Business Intelligence Portal. No further data on patients following a standard fast-track procedure were collected. Outcome measures related to feasibility were discharge readiness (before 9 p.m.), LOS, and causes of prolonged admission. Causes of prolonged admission and readmission were extracted from clinical notes. Outcome measures related to complication rate were: complications during admission (major intraoperative blood loss [estimated > 400 mL], blood transfusion, fracture, hip dislocation, other), prevalence and causes of readmissions (readmission defined as a patient admitted to hospital ward and taking up a bed). Furthermore, data were collected regarding patient demographics (sex, age, marital status, employment status [defined as employed, job seeking, pensioner/retired, sick leave]), ASA class, previous lower limb arthroplasty (THA/TKA), primary diagnosis, duration of surgery, type of prosthesis, use of IV glucocorticosteroids, and pain intensity during admission measured in both rest and activity using the Numeric Range Scale (NRS). Statistics Data are presented using descriptive statistics. Categorical variables are described in numbers and proportions with 95% confidence interval (Cl). All continuous variables were nonnormally distributed and therefore reported in median with either interquartile range (IQR) or range or both. Readmission rate was reported for the group of patients in total and furthermore described in numbers for the group of patients discharged before and after 9 p.m., respectively.

Excluded: Scheduled for standard fast-track THA n = 553 Scheduled for same-day THA n = 116 Discharged at day of surgery n = 102 Readmission n=2

Discharged day 1 after surgery n = 14

No readmission at 90 days follow-up n = 113

Readmission n=1

Participant flow.

In an exploratory analysis, associations between patient characteristics (age, BMI, sex, ASA class) and prolonged admission were tested with the Wilcoxon rank sum test for continuous variables and by estimating relative risk with CI for categorical variables. Data entry and data analysis was performed in Epidata 3.1 (https://www.epidata.dk/) and IBM SPSS Statistics 24 (IBM Corp, Armonk, NY, USA), respectively. Supplemental analysis was performed in Stata 15. Ethics, funding, and conflicts of interest The study was performed in accordance with the Helsinki Declaration of 1975 (revised in 2013) and approved by the Danish Data Protection Agency (1-16-02-165-17) and the management at the Hospital. The Danish Health Data Board provided the requested data from the Danish National Patient Registry. According to Danish law, no further approvals were required. No grants were received. All authors declare no conflicts of interest.

Results During the study period, 669 primary THA procedures were performed and, of these, 116 (in 115 patients) were scheduled for SD-THA. Data were available for all patients at 90 daysâ&#x20AC;&#x2122; follow-up (Figure). Baseline patient characteristics are given in Table 2 and surgery characteristics, use of dexamethasone, and postoperative pain intensity are presented in Table 3. The NRS pain score did not exceed 3 at rest among 58% of the patients and 5 during activity among 80%. Feasibility From 669 elective primary THA procedures, 116 (17%) were scheduled for SD-THA. In 102 of 116 procedures (88%),

Acta Orthopaedica 2019; 90 (5): 439–444

Table 2. Baseline patient characteristics. Values are frequency unless otherwise stated

441

Table 3. Surgery characteristics, use of dexamethasone and postoperative pain intensity

Table 4. Reasons for prolonged admission

Variable (n = 115 a, unless stated otherwise)

Factor (n = 116, unless stated otherwise)

Reason (n = 14)

Male sex Age, median (range) Married/spouse (n = 114) BMI (n = 114), median (range) Employment status (n = 111) employed job seeking pensioner/retired sick leave ASA class (I / II) Previous THA or TKA Primary diagnosis (n = 114) hip arthrosis rheumatoid arthritis fracture sequelae other

Duration of surgery (n = 115), min. median (IQR) Type of prosthesis, n uncemented / hybrid Estimated blood loss (n = 92), mL median (IQR) Use of steroid (dexamethasone), n yes / no Pain (NRS a), median (IQR) [range] rest b activity (n = 90) c

Dizziness and nausea 5 Pain 3 Wound seepage 4 Personal matters spouse concerned 1 patient concerned about discharge 1

72 62 (29–83) 107 27 (19–38) 61 5 44 1 59 / 56 24 104 2 2 6

The number is 115, as 1 patient had 2 procedures performed during the study period. Only data from this patient’s first operation are included.

32 (27–41) 114 / 2 150 (100–225) 97 / 19

1.5 (0.1–2) [0–5] 3.0 (2–4) [0–6]

a b

Numeric Range Scale. NRS pain at rest was measured between 1 and 6 times after surgery. c NRS pain during activity was measured between 0 and 4 times after surgery.

Discussion

patients were discharged as planned. Median LOS was 10.5 hours (range: 6–12 hours) from end of surgery until discharge. The 14 patients with prolonged admission had a median LOS of 25 hours and all were discharged the day after surgery (Table 4). In an exploratory univariate analysis, it was found that female sex (RR 3.1 [CI 1.1–8.5]) and ASA class II compared with ASA class I (RR 3.8 [CI 1.1–13]) were significantly associated with higher risk of prolonged admission. Neither age (p = 0.9) nor BMI (p = 0.8) was found to be statistically significantly associated with prolonged admission. Complications Major bleeding occurred in 7 of 116 patients but all were discharged on the day of surgery as planned. Estimated blood losses in these 7 patients were between 450 mL and 1000 mL. There was no statistically significant difference (p = 0.5) between estimated blood loss among the patients with prolonged admission (median: 200 mL) compared with the patients discharged as planned (median: 150 mL). No hip dislocations, fractures, blood transfusions, or any other major complications occurred during admission. At 90 days’ follow-up 3 patients had been readmitted. The readmissions occurred 3, 14, and 17 days after surgery, respectively, and were due to pneumonia (1), hip dislocation (1), and wound infection (1). 2 of the patients readmitted had been discharged on day of surgery, while the patient readmitted due to dislocation was discharged the day after surgery due to wound seepage.

Our primary findings were a high proportion (88%) of patients being discharged on the day of surgery as planned, no occurrence of major complications and a 2.6 % readmission rate, indicating that SD-THA in this selected group of patients was feasible and safe. Feasibility In this study, 17% of all primary THA patients were eligible for and accepted SD-THA. Of these, 88% were successfully discharged on the day of surgery. Altogether, 102/669 (15%) from our total group of THA patients were discharged on the day of surgery. This result is in accordance with the findings of Gromov et al. (2017), who reported that in an unselected group of THA patients 15% could be discharged on the day of surgery. Even though it is reported to be conducted in unselected patients, nevertheless, Gromov et al. described that to be eligible for same-day surgery patients should be classified as ASA class < 3, operated as number 1 or 2 in the operating room, and have an adult person present at home for at least 24 hours after discharge. These criteria are similar to part of our patient selection criteria, thus making our results likely to be comparable. The number of primary THA patients discharged on the day of surgery from our hospital could possibly be higher than the reported 102 patients, as some patients might have been ready for discharge even though this was not planned in advance. This was the case in the randomized controlled trial by Goyal et al. (2017), in which 18 of 108 patients allocated to be inpatients (discharge the day after surgery) fulfilled the criteria for discharge and chose to leave hospital on the day of surgery. Furthermore, whether all 669 primary THA patients

442

in our study were actually screened for eligibility to be offered a same-day procedure was not registered. Thus, there may be a potential to increase the proportion of SD-THA patients. On the other hand, neither was it registered how many patients were offered but declined to be scheduled for SD-THA. We consider the result of 88% patients being discharged according to plan to be satisfactory, as it is comparable to the result of 85% from our pilot study (Larsen et al. 2017). Comparing our result with studies performed in a traditional hospital setting, only 2 more studies report the use of a posterolateral surgical approach similar to our procedure (Gromov et al. 2017, Springer et al. 2017). Springer et al. do not report the proportion of patients scheduled for SD-THA discharged as planned, while, as mentioned earlier, the result from Gromov et al. is in accordance with our result. In studies using other surgical approaches, the proportion of patients scheduled for SD-THA and discharged as planned varied between 76% and 100% (Berger et al. 2009, Dorr et al. 2010, Aynardi et al. 2014, den Hartog et al. 2015, Goyal et al. 2017). Furthermore, 4 studies, performed in ambulatory surgical centers, had 94% or more discharged on the day of surgery (Parcells et al. 2016, Klein et al. 2017, Berend et al. 2018, Toy et al. 2018), but the different set-up should be taken into account when comparing our result with these studies. Complications during admission Postoperative level of hemoglobin was not routinely tested but required per need based on the patientâ&#x20AC;&#x2122;s well-being and blood loss. A postoperative hemoglobin level lower than 4.3 mmol/L (69 g/L) and/or clinical signs of anemia (not responding on fluid therapy) would constitute transfusion requirement in non-ischemic heart disease patients (Danish Health and Medicines Authority 2015), but in this study none of the patients scheduled for SD-THA required a blood transfusion. This is less than in the total group of THA patients at our hospital, as the proportion of THA patients requiring a blood transfusion within 7 days after surgery was 3.0% in 2016 and 2.3% in 2015 (DHR Annual Report 2017). Our result supports the findings from an American database study, stating that blood transfusion was less for SD-THA than for THA inpatients (Nelson et al. 2017). However, other studies based on the same database have found no difference in transfusion rate (Lovecchio et al. 2016, Basques et al. 2017) or even a higher rate (Otero et al. 2016). Although comparison should be made with caution, we find our result satisfactory. 7 patients in our study had an estimated blood loss above 400 mL, but they were all feeling well (thus, no clinical signs of anemia) and were ready for discharge on the day of surgery. In 24 patients, the estimated blood loss was not registered. 5 of these patients were not discharged on the day of surgery but none of them was readmitted during follow-up. In summary, our results indicate that a blood loss above 400 mL did not have an essential influence on the patientâ&#x20AC;&#x2122;s discharge readiness. However, due to missing data, this finding should

Acta Orthopaedica 2019; 90 (5): 439â&#x20AC;&#x201C;444

be considered with caution. Introduced as a safety consideration, one of the criteria for discharge by Gromov et al. (2017) was an estimated blood loss < 500 ml. That criterion was not fulfilled in 27% of their patients, whereas in our study the corresponding number was only 5%, thus indicating a wellfunctioning blood-saving strategy in our set-up. We found no other major complications during admission. This result is comparable to our previous results from 2015, in which only one minor perioperative complication occurred (drilling fracture during surgery) (Larsen et al. 2017). Our results thereby support the findings in previous studies reporting few or no acute major complications (Berger et al. 2009, Aynardi et al. 2014, den Hartog et al. 2015, Klein et al. 2017, Springer et al. 2017, Berend et al. 2018, Toy et al. 2018). 14 patients had prolonged admission, among whom 12 were due to dizziness and nausea, pain, and wound seepage, with none of the causes being dominating. These causes are well known from other studies (Berger et al. 2009, Dorr et al. 2010, den Hartog et al. 2015, Goyal et al. 2017, Gromov et al. 2017, Springer et al. 2017, Berend et al. 2018, Kim et al. 2018). It should be considered that dexamethasone, which is supposed to reduce pain and nausea, was gradually employed as a standard procedure in late 2015. Thereby, patients operated at the beginning of the study period were less likely to have received it. Though not statistically significant, the proportion of patients with prolonged admission was higher among patients who did not receive dexamethasone (21%) compared with those who did (10%). Thus, it could be hypothesized that, given all eligible patients had received dexamethasone, the number of patients discharged as planned could have been higher. Readmissions Due to low complication rates, it requires a vast number of patients to truly compare complication rates after SD-THA with a population of standard fast-track THA. As an approximative alternative, we find it relevant to evaluate the readmission rate in our study with readmission rates reported in the Danish Hip Arthroplasty Register in 2016 (DHR Annual Report 2017). Although it is a problem to compare with populations, which both comprises part of our study population and who furthermore in general are older and less healthy than in the current study, it has some advantages. The register uses the same definition of readmission as in the current study and a 30-day follow-up, which is comparable to the current study, as all our readmissions occurred within the first 30 postoperative days. Our raw readmission rate (2.6%) did not exceed the raw estimates for the total primary THA populations, which were 5.9% in our local hospital setting and 9.7% in total in Denmark, respectively (DHR Annual Report 2017). The result is in accordance with comparative studies, in which similar readmission rates between SD-THA and THA inpatients were found (Lovecchio et al. 2016, Basques et al. 2017, Courtney et al. 2017, Goyal et al. 2017, Springer et al. 2017). Also, our

Acta Orthopaedica 2019; 90 (5): 439–444

result is similar to that of other observational studies (Berger et al. 2009, den Hartog et al. 2015, Dorr et al. 2010, Larsen et al. 2017, Toy et al. 2018) reporting between 0% (Larsen et al. 2017, Dorr et al. 2010) and 3.7 % (den Hartog et al. 2015). Study limitations and strengths The main limitation concerns selection in patient material. Our patients differed from the total THA population in terms of being younger, healthier, and with a higher proportion of male sex. Furthermore, no data were available on whether all eligible patients were offered SD-THA, or the reasons for patients not being scheduled for SD-THA. Thus, beside the predefined criteria, there is a potential risk of our patient population being further selected. For instance, cognitive status, personal resources, and distance from the patient’s home to the hospital are not part of the eligibility criteria, but may be accounted for when deciding whether to be scheduled for SD-THA. These factors could influence both the surgeon’s decision to offer SD-THA and also the patient’s willingness and motivation to accept the offer. However, as our result is in accordance with the result from Gromov et al. (2017), the risk of further selection might be minor. In addition, extension of the result to different settings as well as in other countries may be of limited value. Due to the retrospective design, we had areas with missing data (i.e., estimated blood loss), and other variables with data not being systematically registered (i.e., pain). Additionally, causes of prolonged admission were based on clinical notes and not registered in predefined categories, thus potentially increasing the risk of misclassification. However, as data were registered independently of research, the risk of differential misclassification is considered low. Furthermore, after discharge, minor complications not leading to readmission were not registered. This should be considered when comparing with other studies. Primarily, the strength of the study was the 100% follow-up regarding readmissions. Secondly, data reflect daily practice, which increases the validity of clinical effectiveness. In summary, 88% of patients scheduled for SD-THA, performed with a posterolateral approach, were discharged on the day of surgery. Primary causes of prolonged admission were dizziness/nausea, pain, and wound seepage. 6% of the patients had an estimated blood loss above 400 mL, but all were discharged as planned without any blood transfusions. No major complications occurred during admission. At 90 days’ follow-up, 3 patients had been readmitted due to pneumonia or wound infection or hip dislocation, respectively. Thus, we conclude that SD-THA is feasible and can be performed with a low complication rate in a selected group of patients. Supplementary data Table 1 is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674.2019. 1637631

443

All the authors contributed in the planning of the study. MNM, MLK, and MFP collected the data. MNM and MLK analyzed the data. MNM, MLK, LRM, ML, and JRL interpreted the results. MNM, MLK, and LRM drafted the manuscript, and it was revised by all the authors. Acta thanks Kirill Gromov for help with peer review of this study.

Aynardi M, Post Z, Ong A, Orozco F, Sukin D C. Outpatient surgery as a means of cost reduction in total hip arthroplasty: a case-control study. HSS J 2014; 10(3): 252-5. Basques B A, Tetreault M W, Della Valle C J. Same-day discharge compared with inpatient hospitalization following hip and knee arthroplasty. J Bone Joint Surg Am 2017; 99(23): 1969-77. Berend K R, Lombardi A V Jr, Berend M E, Adams J B, Morris M J. The outpatient total hip arthroplasty: a paradigm change. Bone Joint J 2018; 100-B(1 Suppl. A): 31-5. Berger R A, Sanders S A, Thill E S, Sporer S M, Della Valle C. Newer anesthesia and rehabilitation protocols enable outpatient hip replacement in selected patients. Clin Orthop Relat Res 2009; 467(6): 1424-30. Courtney P M, Boniello A J, Berger R A. Complications following outpatient total joint arthroplasty: an analysis of a national database. J Arthroplasty 2017; 32(5): 1426-30. Danish Health and Medicines Authority. Instruction on blood transfusion [in Danish: Vejledning om blodtransfusion]. Version: January 2015. Available from: https://stps.dk/da/nyheder/2015/~/media/8147748CEC6140FF9112 674B5EF1307B.ashx. Accessed January 28, 2019. den Hartog Y M, Mathijssen N M, Vehmeijer S B. Total hip arthroplasty in an outpatient setting in 27 selected patients. Acta Orthop 2015; 86(6): 667-70. DHR Annual Report 2017, Danish Hip Arthroplasty Register. Available from: http://danskhoftealloplastikregister.dk/en/publications/annual-reports/. Accessed November 27, 2018. Dorr L D, Thomas DJ, Zhu J, Dastane M, Chao L, Long WT. Outpatient total hip arthroplasty. J Arthroplasty 2010; 25(4): 501-6. Glassou E N, Pedersen A B, Hansen T B. Risk of re-admission, reoperation, and mortality within 90 days of total hip and knee arthroplasty in fast-track departments in Denmark from 2005 to 2011. Acta Orthop 2014; 85(5): 493-500. Goyal N, Chen A F, Padgett S E, Tan T L, Kheir M M, Hopper R H Jr, Hamilton W G, Hozack W J. Otto Aufranc Award: A multicenter, randomized study of outpatient versus inpatient total hip arthroplasty. Clin Orthop Relat Res 2017; 475(2): 364-72. Gromov K, Kjærsgaard-Andersen P, Revald P, Kehlet H, Husted H. Feasibility of outpatient total hip and knee arthroplasty in unselected patients. Acta Orthop 2017; 88(5): 516-21. Hoffmann J D, Kusnezov N A, Dunn J C, Zarkadis N J, Goodman G P, Berger R A. The shift to same-day outpatient joint arthroplasty: a systematic review. J Arthroplasty 2018; 33(4): 1265-74. Husted H. Fast-track hip and knee arthroplasty: clinical and organizational aspects. Acta Orthop Suppl 2012; 83(346): 1-39. Husted H, Jensen C M, Solgaard S, Kehlet H. Reduced length of stay following hip and knee arthroplasty in Denmark 2000–2009: from research to implementation. Arch Orthop Trauma Surg 2012; 132(1): 101-4. Kim K Y, Anoushiravani A A, Elbuluk A, Chen K, Davidovitch R, Schwarzkopf R. Primary total hip arthroplasty with same-day discharge: who failed and why. Orthopedics 2018; 41(1): 35-42. Klein G R, Posner J M, Levine H B, Hartzband M A. Same day total hip arthroplasty performed at an ambulatory surgical center: 90-day complication rate on 549 patients. J Arthroplasty 2017; 32(4): 1103-6. Kort N P, Bemelmans Y F L, van der Kuy P H M, Jansen J, Schotanus M G M. Patient selection criteria for outpatient joint arthroplasty. Knee Surg Sports Traumatol Arthrosc 2017; 25(9): 2668-75.

444

Larsen J R, Skovgaard B, Prynø T, Bendikas L, Mikkelsen L R, Laursen M, Høybye M T, Mikkelsen S, Jørgensen L B. Feasibility of day-case total hip arthroplasty: a single-centre observational study. Hip Int 2017; 27(1): 60-5. Lovecchio F, Alvi H, Sahota S, Beal M, Manning D. Is outpatient arthroplasty as safe as fast-track inpatient arthroplasty? A propensity score matched analysis. J Arthroplasty 2016; 31(9 Suppl.): 197-201. Lovett-Carter D, Sayeed Z, Abaab L, Pallekonda V, Mihalko W, Saleh K J. Impact of outpatient total joint replacement on postoperative outcomes. Orthop Clin North Am 2018; 49(1): 35-44. Lynge E, Sandegaard J L, Rebolj M. The Danish National Patient Register. Scand J Public Health 2011; 39: 30-3. Nelson S J, Webb M L, Lukasiewicz A M, Varthi A G, Samuel A M, Grauer J N. Is outpatient total hip arthroplasty safe? J Arthroplasty 2017; 32(5): 1439-42. Otero J E, Gholson J J, Pugely A J, Gao Y, Bedard N A, Callaghan J J. Length of hospitalization after joint arthroplasty: does early discharge

Acta Orthopaedica 2019; 90 (5): 439–444

affect complications and readmission rates? J Arthroplasty 2016; 31(12): 2714-25. Parcells B W, Giacobbe D, Macknet D, Smith A, Schottenfeld M, Harwood D A, Kayiaros S. Total joint arthroplasty in a stand-alone ambulatory surgical center: short-term outcomes. Orthopedics 2016; 39(4): 223-8. Pollock M, Somerville L, Firth A, Lanting B. Outpatient total hip arthroplasty, total knee arthroplasty, and unicompartmental knee arthroplasty: a systematic review of the literature. JBJS Rev 2016; 4(12). doi: 10.2106/JBJS. RVW.16.00002. Springer B D, Odum S M, Vegari D N, Mokris J G, Beaver W B Jr. Impact of inpatient versus outpatient total joint arthroplasty on 30-day hospital readmission rates and unplanned episodes of care. Orthop Clin North Am 2017; 48(1): 15-23. Toy P C, Fournier M N, Throckmorton T W, Mihalko W M. Low rates of adverse events following ambulatory outpatient total hip arthroplasty at a free-standing surgery center. J Arthroplasty 2018; 33(1): 46-50.

Acta Orthopaedica 2019; 90 (5): 445–449

445

Cementing of the hip arthroplasty stem increases load-to-failure force: a cadaveric study Antonio KLASAN 1, Martin BÄUMLEIN 1, Christopher BLIEMEL 1, Sven Edward PUTNIS 2, Thomas NERI 3, Markus Dietmar SCHOFER 4, and Thomas Jan HEYSE 4 1 University Hospital Marburg, Center for Orthopedics and Traumatology, Marburg, Germany; 2 Sydney Orthopaedic Research Institute, Chatswood, Australia; 3 University Hospital St. Etienne, Department of Orthopaedic Surgery, Saint-Priest-en-Jatez, France; 4 Orthomedic Frankfurt Offenbach,

Offenbach, Germany Correspondence: klasan.antonio@me.com Submitted 2019-03-23. Accepted 2019-06-10.

Background and purpose — To date, there is not a single clinical or mechanical study directly comparing a cemented and a cementless version of the same stem. We investigated the load-to-failure force of a cementless and a cemented version of a double tapered stem. Material and methods — 10 femurs from 5 human cadaveric specimens, mean age 74 years (68–79) were extracted. Bone mineral density (BMD) was measured using peripheral quantitative computed tomography. None of the specimens had a compromised quality (average T value 0.0, –1.0 to 1.4). Each specimen from a pair randomly received a cemented or a cementless version of the same stem. A material testing machine was used for lateral load-to-failure test of up to a maximal load of 5.0 kN. Results — Average load-to-failure of the cemented stem was 2.8 kN (2.3–3.2) and 2.2 kN (1.8–2.8) for the cementless stem (p = 0.002). The cemented version of the stem sustained a higher load than its cementless counterpart in all cases. Failure force was not statistically significantly correlated to BMD (p = 0.07). Interpretation — Implanting a cemented version of the stem increases the load-to-failure force by 25%.

It has been shown that cementless hip arthroplasty components have a higher incidence of periprosthetic femoral fractures (PFFs), with type 2 stems the most susceptible (Carli et al. 2017). The type 2 stem is calcar loading and the most commonly used (Khanuja et al. 2011). Cemented stems are primarily classified as loaded-taper or composite-beam (Scheerlinck and Casteleyn 2006) with the former having a reported higher PFF incidence rate (Carli et al. 2017). Mechanically, only loaded-taper designs have been investigated for load-to-failure. Ginsel et al. (2015) found that an Exeter stem (Stryker Orthopaedics, Mahwah, NJ, USA) with a larger cross-section tolerates more torque until failure, Morishima et al. (2014) found that an Exeter stem (Stryker) with increased length tolerates more torque and energy until failure. Erhardt et al. (2013) compared the double taper CPT stem (Zimmer, Warsaw, IN, USA) with the triple taper C-Stem (DePuy, Raynham, MA, USA) and found no difference in torque or energy, but only in fracture patterns. Clinically, both in electively implanted setting and in fracture treatment settings, cemented stems have a substantially lower incidence of PFF (Carli et al. 2017). Due to different classifications of cementless and cemented stem designs, different availability of each stem in different countries as well as the fact that not all stem designs have a cemented and a cementless version, we were unable to identify a mechanical comparison of a cemented and cementless version of the same stem (Carli et al. 2017). We investigated and quantified the direct load-to-failure force of a cementless and a cemented version of a double tapered stem. Based on current literature evidence, our hypothesis was that the cementless version would have lower load-to-failure.

446

Acta Orthopaedica 2019; 90 (5): 445–449

Figure 2. The mechanical setup.

mineral density (BMD) analysis was performed. Peripheral quantitative computed tomography (pQCT) measurements were used to analyze BMD. For the pQCT measurements, a Stratec XCT Research SA instrument was used (Stratec Medizintechnik GmbH, Pforzheim, Germany). Measurements of BMD were performed at the neck region after obtaining specimens from the cadavers.

Figure 1. Cemented and cementless Polarstem with the original instruments.

Material and methods Specimen preparation The femurs were donated by the authors’ anatomical institute. The specimens originated from 4 male and 1 female adults with an average age of 74 years (68–79). After obtaining the paired femurs they were embalmed with a solution consisting of 96% ethanol and 2% formaldehyde. During perfusion, approximately 15 L of the solution was passed through the femoral artery. All specimens were thawed at thermostat temperature of 19°C and the implantation and testing were performed at that same ambient temperature. To exclude damage related to preexisting fractures or tumors, all specimens were clinically and radiographically examined for integrity. The surrounding soft tissue was stripped from the specimens. Bones were then wrapped in moist towels using the aforementioned embalming solution and stored in a cooling chamber at 4°C to avoid drying artifacts until testing, which occurred after 6–12 months. Bone mineral density assessment In order to exclude osteoporotic specimens and to do a comparison of load-to-failure and bone density, a bone

Femoral stem tested The implants we compared were the cementless Polarstem and the cemented Polarstem (Smith & Nephew, Baar, Switzerland). Cementless Polarstem is a titanium alloy (Ti-6Al-4V ISO 5832-3) double tapered femoral stem with 180 µm of Ti-plasma spray combined with 50 µm of hydroxyapatite coating with fixation occurring on the calcar and metaphysis. Cemented Polarstem is a stainless-steel ISO 5832-9 double tapered femoral stem. We used standard stems with CCD 135°, with each specimen in a pair randomly receiving either a cemented or a cementless stem (Figure 1). The implantation was performed according to the manufacturer’s operating manual and using the original instruments. The trials were implanted until a secure pressfit was obtained. The trials were controlled radiologically for size and fracture. For the cemented version, we used 40 mL of Palacos R+G (Heraeus, Hanau, Germany) in a thirdgeneration technique, with a Buck cement restrictor (Smith & Nephew, Baar, Switzerland). A minimum of 20 minutes was allowed for the cement to set. A polyethylene cup with an inner diameter of 32 mm was used as the acetabulum (Reflection Smith & Nephew, Baar, Switzerland). It was fixed with cement and screws in 45° inclination and 10° of anteversion. A 32 mm ceramic head (Biolox, Ceramtec, Plochingen, Germany) was implanted on the femoral component. The distal femoral fixation was placed at 40 cm distal from the resection using a screw clamp to prevent axial rotation of the specimen. Proximally, a joint was created by inserting the ceramic head into the created acetabulum. The femoral mechanical axis was set parallel to the ground (Figure 2).

Acta Orthopaedica 2019; 90 (5): 445–449

447

Figure 3. (a) Radiograph of the fracture of the femur with a cementless stem, anteroposterior view. (b) Radiograph of the fracture of the femur with a cemented stem, anteroposterior view. (c) Radiograph of the fracture of the femur with a cemented stem, lateral view.

Load to failure assessment Each specimen was tested with load-to-failure on an Instron 5566 universal testing machine (Instron Corp., Darmstadt, Germany) (Figure 2). A 3 cm diameter cylinder was attached to the testing machine and used to apply compression force axially. The test sequence started at 5 N force with the cylinder positioned directly over the greater trochanter. The load was continuously raised at a velocity of 3 N/s. Criteria for discontinuation of testing were premature rotation of the femur (fixation failure) or occurrence of a fracture (final result). Power analysis and statistics First, a trial run with 4th-generation composite femurs (Sawbones, Pacific Research Laboratories, Inc., Vashon, WA, USA) was performed in the same manner as described above for the purpose of power analysis. Based on the differences in forces observed, to reach an alpha of 0.05 and a power of 0.8 the number of cadaveric pairs necessary was 5. The data were collected at 100 ms intervals using the instrument-specific Bluehill Software (Instron, Norwood, MA, USA). Load at failure (kN), and time (s) were recorded. The difference in force between implants was statistically analyzed using a paired t-test and the correlation of force and BMD was analyzed using Pearson’s correlation. Statistical significance was set at p < 0.05. Ethics, funding and potential conflicts of interest All donors provided written consent by their own free will for the use of their body for research purposes. The study was approved by our institution’s ethics board (164/17, 02 November 2017). This study received no external funding. AK has received research support from Implantcast. MS has been paid for presentations by DePuy/Synthes and Smith

& Nephew. TH has been paid for presentations for Smith & Nephew, Zimmer Biomet, and Implantcast. He has received research support from Smith & Nephew, Zimmer Biomet, and Implantcast. He is a consultant to Smith & Nephew. Other authors have no conflicts of interest to declare.

Results Bone mineral density Bone mineral density of the tested femurs was 0.9 g/cm2 (0.8–1.1). There was no statistically significant difference in BMD between the specimens receiving cemented and cementless components (p = 0.8). None of the specimens were of compromised bone quality compared with the corresponding reference value (T-value range –1.0 to 1.4). Implants and load-to-failure Implant sizes were the same in each pair with 1 pair size 1, 2 size 3, 1 size 4, and 1 size 5. A fracture was produced consistently in all specimens. Average load-to-failure of the cemented stem was 2.8 kN (2.3–3.2). Average load-to-failure of the cementless stem was 2.2 kN (1.8–2.8). Cementless stems suffered a fracture of the medial wall on the level of the fracture, extending distally. Cemented stems fractured primarily on the greater trochanter, extending onto the lateral wall, which ultimately caused a dislocation of the stem (Figure 3). The cemented stems sustained a higher load than the cementless stems in all pairs. The mean estimated difference in force between the 2 stems was 0.6 kN (95% CI 0.3–0.8), which corresponds to a force produced by 57 kg. This difference was statistically significant (p = 0.002). There was no statistically significant correlation between fracture force and BMD (p = 0.07).

448

Discussion This study demonstrates on average 25% increased loadto-failure force of a cemented version compared with a cementless version of the same femoral stem. There are a number of studies comparing PFF rates in cemented and cementless stems after both elective total arthroplasty and hemiarthroplasty due to femoral neck fracture, but of different designs. After elective arthroplasty, an increased PFF incidence in cementless compared with cemented stems, with a 3- to 14-fold difference in incidence, was found in a number of studies (Cook et al. 2008, Sheth et al. 2013, Thien et al. 2014, Abdel et al. 2016). The highest reported incidence is nevertheless only 0.45% after 2 years (Thien et al. 2014) and 3.5% after 20 years (Abdel et al. 2016) for uncemented stems. After hemiarthroplasty due to femoral neck fracture, the differences are even more extreme and – for some cementless stems – concerning. The differences occur mainly due to poorer bone quality observed in hemiarthroplasty patients (Langslet et al. 2014). In their respective studies, Langslet et al. (2014), Inngul et al. (2015), Phillips et al. (2013), Foster et al. (2005) all uniformly report a higher incidence of PFF in cementless stems, with incidence of PFF with a cementless stem as high as 12% after 12 months (Inngul et al. 2015). Even young patients have been reported to have excellent outcomes after cemented stems (Kiran et al. 2018). Mechanically, it has been shown that larger stems can improve primary stability of both cemented (Ginsel et al. 2015) and cementless (Fottner et al. 2017) stems due to a better bone load. Increasing the length of the cemented stem has also been shown to increase primary stability (Morishima et al. 2014). Both of these aspects increase the surface contact area between the implant and the bone and have been utilized in designing the stem used in this study (Klasan et al. 2018), when compared with the implant it was based on, the Corail stem (DePuy, Raynham, MA, USA). These 2 stems have not been tested against each other. Furthermore, Carli et al. (2017) report a lack of comparison between a cemented and a cementless version of the same stem with regards to PFF rate or load-to-failure. Data suggest that bone mineral density around the stem decreases more with a cemented stem (Li et al. 2007) than with a cementless stem (Flatøy et al. 2016, Aro et al. 2018), without a correlation to subsidence. Again, a direct comparison like that in this study is yet to be performed. In clinical settings, the evidence suggests a higher risk of PFF (Hailer et al. 2010, Sidler-Maier and Waddell 2015, Carli et al. 2017) and a higher revision rate (McMinn et al. 2012) for cementless stems. Yet, the evidence suggests cementless stems are used more often (Lehil and Bozic 2014). According to some researchers, a clear consensus on when to use cemented stems is missing (Moskal et al. 2016).

Acta Orthopaedica 2019; 90 (5): 445–449

This lack of consensus and data was one of the reasons to perform this study. Due to a higher risk of revision in older patients with cementless stems (Jämsen et al. 2014), studies set different thresholds for “older age” patients (Moskal et al. 2016), which affects the results depending on the study population. Since the definition of “elderly” is also unclear in the orthopedic literature (Sabharwal et al. 2015), these thresholds are even more difficult to define. There is also an expectation of older patients having lower bone quality (Wright et al. 2014) where a combination of a higher risk of falling, soft bone, and slower osseointegration make the surgeon incline toward cementing the stem. This provides immediate osseointegration and protects the implant at the same time, as was shown in our study. In this study, where non-osteoporotic bones were used, the difference in forceto-failure corresponds to 80% of the weight of an average European adult (Walpole et al. 2012). Stem design has also been shown to influence the PFF and revision rate for both cemented (Palan et al. 2016, Kazi et al. 2019) and cementless stems (Carlson et al. 2017). Also, cemented and cementless versions of the same stem do not exist as often or do not have worldwide distribution. For instance, the cemented version of the stem used in our study is not available in the United States despite showing excellent clinical results (Klasan et al. 2018). Further studies with other stems available in both versions in a clinical and a mechanical setting are needed to provide this information. Several limitations need to be noted for our study. Cemented fixation occurs within 10 minutes at room temperature, whereas cementless stems were press-fitted. The direct implications of our study can therefore only be observed in the postoperative phase, prior to bone in-growth in vivo; this, however, remains a clinically relevant period with registry data showing a PFF rate of 2.1% ≤ 90 days postoperatively (Lindberg-Larsen et al. 2017). Second, unidirectional compression force was used to produce a fracture, whereas fractures are always a result of a combination of forces. Our study also does not account for soft tissue contributions due to stripping of soft tissue and it therefore cannot precisely determine the mechanical effect of the implant in a patient where soft tissue is present. Finally, we used embalmed and not fresh frozen specimens. It has been shown, however, that the mechanical characteristics of fresh frozen specimens and embalmed cadaveric specimens are similar (Topp et al. 2012). A polyethylene-ceramic bearing has been used, as this is the standard bearing used in our institution. Even though the revision rate for this bearing is lower (Peters et al. 2018), we do not believe it affected the outcome of this particular study. In summary, implanting a cemented version of the stem increases the load-to-failure force by 25%. We recommend using a cemented stem in older patients and patients who are at risk of falling.

Acta Orthopaedica 2019; 90 (5): 445–449

AK and MB conducted the experiments. AK, SP, and TN wrote the manuscript. CB and TH devised the study and reviewed the manuscript together with MS, who did the statistical analysis. The authors would like to thank Dr Jürgen Paletta for his help in the biomechanical laboratory. Acta thanks Stephan Maximilian Röhrl for help with peer review of this study.

Abdel M P, Watts C D, Houdek M T, Lewallen D G, Berry D J. Epidemiology of periprosthetic fracture of the femur in 32 644 primary total hip arthroplasties: a 40-year experience. Bone Joint J 2016; 98-B(4): 461-7. Aro E, Alm J J, Moritz N, Mattila K, Aro H T. Good stability of a cementless, anatomically designed femoral stem in aging women: a 9-year RSA study of 32 patients. Acta Orthop 2018; 89(5): 490–5. Carli A V, Negus J J, Haddad F S. Periprosthetic femoral fractures and trying to avoid them: what is the contribution of femoral component design to the increased risk of periprosthetic femoral fracture? Bone Joint J 2017; 99-B(1 Suppl. A): 50-9. Carlson S W, Liu S S, Callaghan J J. Not all cementless femoral stems are created equal but the results may be comparable. Bone Joint J 2017; 99-B(1 Suppl. A): 14-17. Cook R E, Jenkins PJ, Walmsley P J, Patton J T, Robinson C M. Risk factors for Periprosthetic Fractures of the Hip: A Survivorship Analysis. Clin Orthop Relat Res 2008; 466(7): 1652–6. Erhardt J B, Khoo P P, Stoffel K K, Yates P J. Periprosthetic fractures around polished collarless cemented stems: the effect of stem design on fracture pattern. Hip Int 2013; 23(5): 459-64. Flatøy B, Röhrl S M, Bøe B, Nordsletten L. No medium-term advantage of electrochemical deposition of hydroxyapatite in cementless femoral stems. 5-year RSA and DXA results from a randomized controlled trial. Acta Orthop 2016; 87(1): 42–7 Foster A P, Thompson N W, Wong J, Charlwood A P. Periprosthetic femoral fracture: a comparison between cemented and uncemented hemiarthroplasties. Injury 2005; 36(3): 424-9. Fottner A, Woiczinski M, Kistler M, Schröder C, Schmidutz T F, Jansson V, Schmidutz F. Influence of undersized cementless hip stems on primary stability and strain distribution. Arch Orthop Trauma Surg 2017; 137(10): 1435-41. Ginsel B L, Morishima T, Wilson L J, Whitehouse S L, Crawford R W. Can larger-bodied cemented femoral components reduce periprosthetic fractures? A biomechanical study. Arch Orthop Trauma Surg 2015; 135(4): 517-22. Hailer N P, Garellick G, Kärrholm J. Uncemented and cemented primary total hip arthroplasty in the Swedish Hip Arthroplasty Register. Acta Orthop 2010; 81(1): 34-41. Inngul C, Blomfeldt R, Ponzer S, Enocson A. Cemented versus uncemented arthroplasty in patients with a displaced fracture of the femoral neck: a randomised controlled trial. Bone Joint J 2015; 97-B(11): 1475-80. Jämsen E, Eskelinen A, Peltola M, Mäkelä K. High early failure rate after cementless hip replacement in the octogenarian. Clin Orthop Relat Res 2014; 472(9): 2779-89. Kazi H A, Whitehouse S L, Howell J R, Timperley A J. Not all cemented hips are the same: a register-based (NJR) comparison of taper-slip and composite beam femoral stems. Acta Orthop 2019; 90(3): 214–9. Khanuja H S, Vakil J J, Goddard M S, Mont M A. Cementless femoral fixation in total hip arthroplasty. J Bone Joint Surg Am 2011; 93(5): 500-9.

449

Kiran M, Johnston L R, Sripada S, Mcleod G G, Jariwala A C. Cemented total hip replacement in patients under 55 years. Acta Orthop 2018;8 9(2): 152–5. Klasan A, Sen A, Dworschak P, El-Zayat B F, Ruchholtz S, Schuettler K F, Schmitt J, Heyse T J. Ten-year follow-up of a cemented tapered stem. Arch Orthop Trauma Surg 2018; 138(9): 1317-22. Langslet E, Frihagen F, Opland V, Madsen J E, Nordsletten L, Figved W. Cemented versus uncemented hemiarthroplasty for displaced femoral neck fractures: 5-year followup of a randomized trial. Clin Orthop Relat Res 2014; 472(4): 1291-9. Lehil M S, Bozic K J. Trends in total hip arthroplasty implant utilization in the United States. J Arthroplasty 2014; 29(10): 1915-8. Li M G, Rohrl S M, Wood D J, Nivbrant B. Periprosthetic changes in bone mineral density in 5 stem designs 5 years after cemented total hip arthroplasty. No relation to stem migration. J Arthroplasty 2007; 22(5): 689–91. Lindberg-Larsen M, Jørgensen C C, Solgaard S, Kjersgaard A G, Kehlet H, on behalf of the LFC for F-T, Group KRC. Increased risk of intraoperative and early postoperative periprosthetic femoral fracture with uncemented stems. Acta Orthop 2017; 88(4): 390-4. McMinn D J W, Snell K I E, Daniel J, Treacy R B C, Pynsent P B, Riley R D. Mortality and implant revision rates of hip arthroplasty in patients with osteoarthritis: registry based cohort study. BMJ 2012; 344: e3319. Morishima T, Ginsel B L, Choy G G, Wilson L J, Whitehouse S L, Crawford R W. Periprosthetic fracture torque for short versus standard cemented hip stems: an experimental in vitro study. J Arthroplasty 2014; 29(5): 1067-71. Moskal J T, Capps S G, Scanelli J A. Still no single gold standard for using cementless femoral stems routinely in total hip arthroplasty. Arthroplasty Today 2016; 2(4): 211-18. Palan J, Smith M C, Gregg P, Mellon S, Kulkarni A, Tucker K, Blom A W, Murray D W, Pandit H. The influence of cemented femoral stem choice on the incidence of revision for periprosthetic fracture after primary total hip arthroplasty: an analysis of national joint registry data. Bone Joint J 2016; 98-B(10): 1347-54. Peters R M, Van Steenbergen L N, Stevens M, Rijk P C, Bulstra S K, Zijlstra W P. The effect of bearing type on the outcome of total hip arthroplasty. Acta Orthop 2018; 89(2): 163–9. Phillips J R A, Moran C G, Manktelow A R J. Periprosthetic fractures around hip hemiarthroplasty performed for hip fracture. Injury 2013; 44(6): 757-62. Sabharwal S, Wilson H, Reilly P, Gupte C M. Heterogeneity of the definition of elderly age in current orthopaedic research. Springerplus. 2015; 4. Scheerlinck T, Casteleyn P-P. The design features of cemented femoral hip implants. J Bone Joint Surg Br 2006; 88(11): 1409-18. Sheth N P, Brown N M, Moric M, Berger R A, Della Valle C J. Operative treatment of early peri-prosthetic femur fractures following primary total hip arthroplasty. J Arthroplasty 2013; 28(2): 286-91. Sidler-Maier C C, Waddell J P. Incidence and predisposing factors of periprosthetic proximal femoral fractures: a literature review. Int Orthop 2015; 39(9): 1673-82. Thien T M, Chatziagorou G, Garellick G, Furnes O, Havelin L I, Mäkelä K, Overgaard S, Pedersen A, Eskelinen A, Pulkkinen P, Kärrholm J. Periprosthetic femoral fracture within two years after total hip replacement: analysis of 437,629 operations in the Nordic Arthroplasty Register Association database. J Bone Joint Surg Am 2014; 96(19): e167. Topp T, Müller T, Huss S, Kann P H, Weihe E, Ruchholtz S, Zettl R P. Embalmed and fresh frozen human bones in orthopedic cadaveric studies: which bone is authentic and feasible? Acta Orthop 2012; 83(5): 543-7. Walpole SC, Prieto-Merino D, Edwards P, Cleland J, Stevens G, Roberts I. The weight of nations: an estimation of adult human biomass. BMC Public Health 2012; 12: 439. Wright N C, Looker AC, Saag K G, Curtis J R, Delzell E S, Randall S, Dawson-Hughes B. The recent prevalence of osteoporosis and low bone mass in the United States based on bone mineral density at the femoral neck or lumbar spine. J Bone Miner Res 2014; 29(11): 2520–6.

450

Acta Orthopaedica 2019; 90 (5): 450–454

Sequence of 305,996 total hip and knee arthroplasties in patients undergoing operations on more than 1 joint Peter ESPINOSA 1, Rüdiger J WEISS 1, Otto ROBERTSSON 2,3, and Johan KÄRRHOLM 4,5 1 Department of Molecular Medicine and Surgery, Section of Orthopaedics and Sports Medicine, Karolinska Institutet, Karolinska University Hospital, Stockholm; 2 Swedish Knee Arthroplasty Register, Lund; 3 Faculty of Medicine, Department of Clinical Sciences Lund, Orthopedics, Lund University, Lund; 4 Swedish Hip Arthroplasty Register, Gothenburg; 5 Department of Orthopaedics, Institute of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden Correspondence: peter.espinosa@sll.se Submitted 2018-11-28. Accepted 2019-05-13.

Background and purpose — Patient-specific data on multiple total arthroplasties (TA) of the lower limbs due to osteoarthritis (OA) are limited. We investigated the sequence of surgical procedures and risk factors for additional surgery in such patients. Patients and methods — 305,996 patients operated with a TA of the hip and/or knee due to OA were extracted from the Swedish National Hip (SHAR) and the Swedish Knee Arthroplasty Register (SKAR). 177,834 total hip arthroplasty (THA, 56% women, mean age 69 years) and 128,162 total knee arthroplasty (TKA, 60% women, mean age 69 years) procedures constituted the index operations. The mean, median, and maximum follow-up was 8, 6, and 23 years. Multivariable Cox regression analysis was used and Kaplan–Meier survival curves were constructed. Results — Right-sided primary TA (34%) was most frequent. Subsequent surgery was most frequent after primary left-sided TKA (33%). The time interval to a second TA procedure was 3.1 (SD 3.2) years after TKA and 4.0 (SD 3.9) years after THA. After the index TA the probability of no subsequent surgery amounted to 64% (SD 0.3) for THA and 58% (SD 0.4) for TKA over 20 years. Lower age, female sex, left side, and TKA at index operation were associated with a higher probability for subsequent TA. Interpretation — Delineation of factors that influence risk and the size of the risk for subsequent TA in 1 of the 3 major remaining joints is of value for clinicians and healthcare providers in the decision-making process for future resource allocation.

The progression of osteoarthritis (OA) is still unclear, particularly in patients with multiple severe OA in the large weightbearing joints leading to total hip (THA) and total knee arthroplasty (TKA). Previous studies suggest an association between the side of the first TA and the following THA or TKA (Shakoor et al. 2002, Gillam et al. 2013, Shao et al. 2013, Sanders et al. 2017). There are also a few other studies reporting on the outcome of patients with multiple TA (Papanikolaou et al. 2000, Mulhall et al. 2008, Hui et al. 2012). Some of these studies included relatively small patient numbers (< 100 patients) and were published more than a decade ago (Papanikolaou et al. 2000, Justen et al. 2003, Mulhall et al. 2008). We are only aware of a limited number of articles that have been population-based or published in the last decade (Gillam et al. 2012, 2013, Shao et al. 2013, Maradit Kremers et al. 2015, Sanders et al. 2017). Projections in Sweden show a steady increase to approximately 20,000 THA and 22,000 TKA performed in year 2030 which make TA one of the most common elective procedures (Nemes et al. 2014, 2015). The projected increase in patients with multiple TA and the healthcare resources spent on this group of patients (Nemes et al. 2014, 2015) led us to investigate the sequence of surgical procedures and risk factors for additional surgery in patients with multiple TA of the lower limbs due to OA.

Patients and methods The Swedish Hip Arthroplasty Register (SHAR) is a national quality register that collects data on primary THA performed in Sweden since 1979. Revisions and reoperations are regis-

Acta Orthopaedica 2019; 90 (5): 450–454

451

Cumulative risk of further TA (%)

Patients with primary TA registered in SHAR/SKAR 1992–2014 n = 377,044

50 Left TKR Right TKR Left THR Right THR

Excluded (n = 71,048): – other diagnosis than osteoarthritis, 63,741 – same day multiple arthroplasty, 5,441 – missing data, 1,866

Single TA due to primary osteoarthritis n = 305,996

Figure 1. Flowchart showing patient selection. TA = total hip or knee arthroplasty, SHAR = Swedish Hip Arthroplasty Register, SKAR = Swedish Knee Arthroplasty Register.

0 0

tered on an individual level. The completeness of the SHAR is around 98% (SHAR 2017). Since 1975 all primary and revision TKA performed in Sweden are gathered in the Swedish Knee Arthroplasty Register (SKAR). This register has national coverage and 97% completeness (SKAR 2018). Both registers record individual patient data such as age, sex, the diagnosis leading to surgery, side of the surgical procedure, surgical technique, and type of implant used. The diagnosis is reported to the SHAR and SKAR as found in patients’ medical records. No validation was performed for this study. SKAR started gathering detailed patient information on an individual level in 1975 and SHAR in 1992. At the time of the linkage of SHAR and SKAR for this study in 2016, data until 2014 were available. By linking the registers, we could extract all patients operated with a THA and/or TKA during 1992–2014. This generated a data-set of 377,044 patients. After exclusions a final study cohort of 305,996 patients was formed (Figure 1). Statistics Demographic data are presented as means, standard deviations (SD), and ranges. We used Cox multiple proportional hazards regression analysis to evaluate factors with possible influence on the risk for subsequent TA surgery after initial TA. Hazard ratios (HR) with 95% confidence intervals (CI) are presented, both unadjusted and adjusted. Variables included in the Cox model were age, sex, side, type of TA, and time period for the index operation (1992–1998, 1999–2006 and 2007–2014). Subdivision of the time period was arbitrary. Proportionality was tested by plotting survival curves for the different variables individually and log-minus-log plots. Kaplan–Meier survival curves were constructed to evaluate the risk of a second TA over time. Patients were censored at their second TA, the end of the study period (December 31, 2014) or at death. Survival probabilities and 95% CIs are presented. Patients undergoing a 3rd and 4th single TA were comparatively few and are only presented as percentages. All analyses were performed using the PASW statistics package version 25 (SPSS Inc, Chicago, IL, USA).

Years after index operation

Figure 2. Cumulative (Kaplan–Meier) risk of further total joint arthroplasty (TA) after primary operation of right or left total hip arthroplasty (THA) or total knee arthroplasty (TKA).

Ethics, funding, and potential conflicts of interest Ethical approval was granted by the Regional Ethical Review Board in Gothenburg (approval number: 2017-01-26 dnr: 1034-16.). PE has been supported by grant from LINK Sweden AB. No competing interest declared.

Results Study population 177,834 patients underwent THA (56% women, n = 99,409) at a mean age of 69 (15–99) years. 128,162 patients underwent TKA (60% women, n = 76,641) at a mean age of 69 (12–97) years. The mean, median, and maximum follow-up was 8, 6, and 23 years, respectively. The data resulted in 40 different possible combinations of TA surgery sequences (Table 1, see Supplementary data). Distribution of patients with TA Of all patients with a TA, 34% had a right-sided THA, 25% had a left-sided THA, 22% a right-sided TKA, and 19% a left-sided TKA as the first (index) surgical procedure. The sequence between index and subsequent TA are presented in Table 2. Table 3 (see Supplementary data) presents 3rd and 4th TAs after the index TA. The “survival” based on risk to be operated in 1 of the 3 remaining joints for patients alive during the study period was highest after right-sided index THA (65% at 22 years) followed by left-sided index THA (62%), index right (60%), and index left TKA (57%) (Figure 2). Cox regression analysis The Cox regression analysis revealed that lower age (HR 1.028; CI 1.027–1.029), female sex (HR 1.14; CI 1.12–1.15),

452

Acta Orthopaedica 2019; 90 (5): 450–454

Table 2. Numbers at index operation and number of patients (%) who underwent subsequent joint replacements of the hip and/or the knee

Discussion

We found that patients had a slightly increased risk of at least 1 subsequent TA if they were younger, female, or if they THA right 102,705 76,618 (75) – 21,568 (21) 1,643 (2) 2,876 (3) had a left-sided index TKA. The probTHA left 75,129 53,868 (72) 18,106 (24) – 2,329 (3) 826 (1) ability to be operated in more than 1 joint TKA right 68,836 48,248 (70) 1,674 (2) 1,267 (2) – 17,647 (26) TKA left 59,326 40,069 (68) 1,594 (3) 988 (2) 16,675 (28) – increased during the study period. With the use of TA as a proxy for severe OA, we could confirm that the same joint on the contralateral side as the initial operation had the highest risk for further TA surgery. If the second Table 4. Cox regression analysis: risk factors for further total joint arthroplasty (TA) of the hip and operation after a THA involved the knee, this second operaknee after the index TA during 1992–2014 tion was also most frequently localized to the contralateral side. Following initial TKA, the second operation, if a THA, Factor Hazard ratio (95% CI) was most commonly right-sided, disregarding localization of side for the index operation. Unadjusted Gillam et al. (2013) found that more women than men Low age 1.028 (1.027–1.029) Male sex 1.0 (ref.) had an index TA in the Norwegian population. This was also Female sex 1.08 (1.06–1.09) observed in our study. A higher proportion of subsequent Right side 1.0 (ref.) right-sided rather than left-sided THA after an index TKA Left side 1.14 (1.12–1.15) THA 1.0 was found in the Norwegian population, as in our study. This TKA 1.30 (1.28–1.32) similarity is not unexpected since Norway and Sweden share Year of surgery a border and the populations at large have a common genetic 1992–1998 1.0 (ref.) 1999–2006 1.28 (1.26–1.30) and cultural background. The higher probability for subse2007–2014 1.43 (1.40–1.46) quent right-sided THA after index TKA in our study does not, a Adjusted however, completely concur with this previous study partly Low age 1.028 (1.027–1.029) Male sex 1.0 (ref.) based on the Norwegian population. In that study based on Female sex 1.14 (1.12–1.15) both the Australian and Norwegian National Joint Registries, Right side 1.0 (ref.) Gillam et al. (2013) found that patients primarily operated Left side 1.12 (1.10–1.13) THA 1.0 (ref.) with TKA generally had a higher risk of receiving subsequent TKA 1.33 (1.31–1.35) THA on the contralateral side. Year of surgery Previous studies suggested that the progression of OA and 1992–1998 1.0 (ref.) 1999–2006 1.23 (1.21–1.25) TA surgery does not represent random progress (Shakoor et 2007–2014 1.31 (1.28–1.33) al. 2002, Gillam et al. 2013). We found a similar pattern as a Adjustments for the following variables: age, sex, previously observed with a higher share of patients who had side, type of the TA and time period for the index their subsequent TA in the cognate contralateral joint (Shaoperation. koor et al. 2002, Gillam et al. 2013, Shao et al. 2013, Sanders et al. 2017). We could also confirm the observation that patients who received a first-time THA and continued with a TKA at index operation (HR 1.33; CI 1.31–1.35), and left- subsequent TKA were more frequently operated in the contrasided index operation (HR 1.12; CI 1.10–1.13) were asso- lateral knee (Shakoor et al. 2002, Gillam et al. 2013, Sanders ciated with an increased risk for at least 1 additional joint et al. 2017). replacement of the hip or knee. The risk of undergoing at Sanders et al. (2017) described that younger age was a risk least 1 further TA was increased for each observation period factor for subsequent THA after an index THA but not after an (1999–2006: HR 1.23, CI 1.21–1.25; 2007–2014: HR 1.31, index TKA. This compares to a certain extent with our results CI 1.28–1.33) (Table 4). where younger age was a risk factor for subsequent TA, but we did not perform separate analyses of index THA and TKA Time interval between index and subsequent TA operations. Concerning the time interval between an index and The mean time between surgery with an index THA and any a second TA, Shao et al. (2013) showed faster progress after further TA of the remaining 3 joints was 4 years (SD 4, range an index TKA as observed in our study. The degenerative joint disease in combination with pain may 0–23). The mean time after index TKA was shorter than after cause changes in the external muscular forces, the joint load, index THA (3 years, SD 3, range 0–22). First TA (index) n

No further primary TA

THA right

Second primary TA THA left TKA right

TKA left

Acta Orthopaedica 2019; 90 (5): 450–454

and the walking pattern, which in a complex way may influence the progression of the disease in other non-operated joints both before and after surgery (Shakoor et al. 2002). An indicator for this might be the progression from index TA to a subsequent TA of the same joint on the contralateral side. This may be explained by changes in weight load before and after surgery putting more stress on the remaining joints. Elimination of pain from 1 joint after insertion of a TA may also change the threshold for experience of pain from other joints with OA. Localization and progression of the disease may also be related to the patient phenotype, suggesting that the pattern of OA evolution may vary depending on genetic background (Nelson et al. 2013). Some authors reported that willingness to undergo TA surgery may also vary between different patient groups and cultural backgrounds (Allen et al. 2014, Krupic et al. 2014). We found an increased risk of multiple surgery during the later study period. The reason for this is not known, but increasing life expectancy, subtle changes of indication over time, increasing demands on an active lifestyle, and increased accessibility to health care may be possible explanations. Our study has several limitations. Since this is a registerbased study of TA surgery, only analyses of the sequence of severe OA requiring surgery could be studied, excluding all cases with OA not subjected to arthroplasty surgery. No radiographic analysis determining the severity of OA was performed. However, previous studies validated the use of TA as an appropriate indicator for severe OA (Shakoor et al. 2002, Dougados 2004, Lohmander et al. 2009). We included all patients operated with a primary THA and TKA registered in the SHAR and SKAR during 1992–2014. SHAR did not record data on patient identification before 1992 and therefore we cannot account for THA operations before 1992. SKAR has recorded data on patient identification since 1975 but for this study we did not exclude patients who received a TKA before 1992. Thus, some patients might have received a TA before 1992 although this subgroup of patients can be expected to be small due to a more conservative attitude towards TA surgery in Sweden during the years before 1992. A total of 1,866 incomplete surgical procedures were discarded. These patients represented only 0.6% of the total number but could nonetheless have an impact on the results and especially for patients with 3 and 4 TAs with comparatively few recorded procedures. Moreover, reoperations and revisions were not included in our study. Most probably, the occurrence of any type of reoperation may influence the willingness to undergo further surgery of other joints, though to what extent is unknown. Another factor to consider is surgeon’s preference and any bias associated with choice of 1st joint to operate in patients who at initial visit present with multiple joint severe OA. The diagnosis is reported to the SHAR and SKAR as found in patient medical records. No other validation was performed. There is always a risk of misclassification of the primary diagnosis, and especially in very young patients with joint disease classified as primary OA.

453

A strength is our large cohort of OA patients undergoing multiple TA of the hip and knee with a national coverage of almost 100%. This represents the first study where the SHAR and SKAR were merged on patient identification. To further explore the pattern of OA progression between different joints, studies including radiographic examination and non-operated cases would be necessary. In summary, during an observation period of 20 years, we found that after TA due to primary OA the probability of no further TA in the remaining major joints of the lower extremity was about 60%, slightly higher after THA than after TKA. Lower age, female sex, left side, and TKA at index operation were associated with a higher probability for subsequent TA. If performed, the time interval to a second TA was almost 1 year shorter after index TKA than after index THA. Supplementary data Tables 1 and 3 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674. 2019.1638177

PE and JK: Conception and design of the study, analysis and interpretation of data, drafting the article and revision. RJW: Revision of article. OR: Analysis and interpretation of data, drafting the article and revision. Acta thanks Marianne Westberg for help with peer review of this study.

Allen K D, Golightly Y M, Callahan L F, Helmick C G, Ibrahim S A, Kwoh C K, Renner J B, Jordan J M. Race and sex differences in willingness to undergo total joint replacement: the Johnston County Osteoarthritis Project. Arthritis Care Res (Hoboken) 2014; 66(8): 1193-202. doi: 10.1002/ acr.22295. Dougados M. Monitoring osteoarthritis progression and therapy. Osteoarthritis Cartilage 2004; 12:55-60. doi: 10.1016/j.joca.2003.09.009. Gillam M H, Ryan P, Salter A, Graves S E. Multi-state models and arthroplasty histories after unilateral total hip arthroplasties: introducing the Summary Notation for Arthroplasty Histories. Acta Orthop 2012; 83(3): 220-6. doi: 10.3109/17453674.2012.684140. Gillam M H, Lie S A, Salter A, Furnes O, Graves S E, Havelin L I, Ryan P. The progression of end-stage osteoarthritis: analysis of data from the Australian and Norwegian joint replacement registries using a multistate model. Osteoarthritis Cartilage 2013; 21(3): 405-12. doi: 10.1016/j. joca.2012.12.008. Hui C, Ben-Lulu O, Rendon J S, Soever L, Gross A E, Backstein D. Clinical and patient-reported outcomes of patients with four major lower extremity arthroplasties. J Arthroplasty 2012; 27(4): 507-13. doi: 10.1016/j. arth.2011.06.014. Justen H P, Glennemeier A, Kisslinger E, Wessinghage D. Multiple joint replacement of the lower limbs in rheumatoid arthritis. Zeitschrift fur Rheumatologie 2003; 62(2): 161-7. doi: 10.1007/s00393-003-0459-1. Krupic F, Garellick G, Gordon M, Karrholm J. Different patient-reported outcomes in immigrants and patients born in Sweden: 18,791 patients with 1 year follow-up in the Swedish Hip Arthroplasty Registry. Acta Orthop 2014; 85(3): 221-8. doi: 10.3109/17453674.2014.919556. Lohmander L S, Gerhardsson de Verdier M, Rollof J, Nilsson P M, Engstrom G. Incidence of severe knee and hip osteoarthritis in relation to different measures of body mass: a population-based prospective cohort study. Ann Rheum Dis 2009; 68(4): 490-6. doi: 10.1136/ard.2008.089748.

454

Maradit Kremers H, Larson D R, Crowson C S, Kremers W K, Washington R E, Steiner C A, Jiranek W A, Berry D J. Prevalence of total hip and knee replacement in the United States. J Bone Joint Surg Am 2015; 97(17): 1386-97. doi: 10.2106/JBJS.N.01141. Mulhall K J, Saleh K J, Thompson C A, Severson E P, Palmer D H. Results of bilateral combined hip and knee arthroplasty in very young patients with juvenile rheumatoid arthritis. Arch Orthop Trauma Surg 2008; 128(3): 24954. doi: 10.1007/s00402-007-0450-4. Nelson A E, Golightly Y M, Renner J B, Schwartz T A, Kraus V B, Helmick C G, Jordan J M. Brief report: differences in multijoint symptomatic osteoarthritis phenotypes by race and sex: the Johnston County Osteoarthritis Project. Arthritis Rheum 2013; 65(2): 373-7. doi: 10.1002/art.37775. Nemes S, Gordon M, Rogmark C, Rolfson O. Projections of total hip replacement in Sweden from 2013 to 2030. Acta Orthop 2014; 85(3): 238-43. doi: 10.3109/17453674.2014.913224. Nemes S, Rolfson O, A W D, Garellick G, Sundberg M, Karrholm J, Robertsson O. Historical view and future demand for knee arthroplasty in Sweden. Acta Orthop 2015; 86(4): 426-31. doi: 10.3109/17453674. 2015.1034608.

Acta Orthopaedica 2019; 90 (5): 450â&#x20AC;&#x201C;454

Papanikolaou A, Droulias K, Nikolaides A, Polyzoides A J. Results of a single total knee prosthesis compared with multiple joint replacement in the lower limb. Int Orthop 2000; 24(2): 80-2. Sanders T L, Maradit Kremers H, Schleck C D, Larson D R, Berry D J. Subsequent total joint arthroplasty after primary total knee or hip arthroplasty: a 40-year population-based study. J Bone Joint Surg Am 2017; 99(5): 396401. doi: 10.2106/JBJS.16.00499. Shakoor N, Block J A, Shott S, Case J P. Nonrandom evolution of end-stage osteoarthritis of the lower limbs. Arthritis Rheum 2002; 46(12): 3185-9. doi: 10.1002/art.10649. Shao Y, Zhang C, Charron K D, Macdonald S J, McCalden R W, Bourne R B. The fate of the remaining knee(s) or hip(s) in osteoarthritic patients undergoing a primary TKA or THA. J Arthroplasty 2013; 28(10): 1842-5. doi: 10.1016/j.arth.2012.10.008. SHAR. Swedish Hip Arthroplasty Register, Annual Report 2017. Available at https://shprregistercentrumse/shar-in-english/annual-reports/p/rkeyyeElz SKAR. Swedish Knee Arthroplasty Register, Annual Report 2018. Available at http://www.myknee.se/en/publications/annual-reports

Acta Orthopaedica 2019; 90 (5): 455–459

455

Optimization of the empirical antibiotic choice during the treatment of acute prosthetic joint infections: a retrospective analysis of 91 patients Joost H J VAN ERP 1, Adriaan C HEINEKEN 1, Remco J A VAN WENSEN 1, Robin W T M VAN KEMPEN 1, Johannes G E HENDRIKS 2, Marjolijn WEGDAM-BLANS 3, Judith M FONVILLE 3, and M C (Marieke) VAN DER STEEN 1,2 1 Department of Orthopaedic Surgery, Catharina Hospital Eindhoven, Eindhoven; 2 Orthopaedic 3 Laboratory of Medical Microbiology, Stichting PAMM, Veldhoven, The Netherlands

Center Máxima, Máxima Medical Center, Eindhoven;

Correspondence: J.Fonville@pamm.nl Submitted 2018-11-14. Accepted 2019-04-29.

Background and purpose — The preferred treatment of an acute prosthetic joint infection (PJI) is debridement, antibiotics, irrigation and retention of the prosthesis (DAIR). The antibiotic treatment consists of an empirical and targeted phase. In the empirical phase, intravenous antibiotics are started after surgery before micro-organisms are determined in microbiological cultures. Which empirical antibiotic is used differs between hospitals, partly reflecting geographic differences in susceptibility spectrums. We investigated whether flucloxacillin should remain the antibiotic of choice in our hospital for empiric treatment of acute PJI with DAIR. Patients and methods — We retrospectively analyzed 91 patients treated for PJI with DAIR between 2012 and 2016. The susceptibility of micro-organisms was determined in multiple cultures of periprosthetic tissue and synovial fluid for 3 antibiotics: amoxicillin/clavulanic acid, cefazolin, and flucloxacillin. Results — Positive microbiological cultures from 68 patients were analyzed. Staphylococcus aureus was the predominant pathogen, cultured in half of the patients. In onethird of patients more than 1 micro-organism was found. On a patient level, the data showed that 65% were responsive to flucloxacillin, 76% to amoxicillin/clavulanic acid, and 79% to cefazolin. Interpretation — Flucloxacillin appeared to be a suboptimal choice in our patient population treated with DAIR. We therefore changed our practice to cefazolin as the preferred antibiotic in the empirical treatment of acute PJI with DAIR.

A prosthetic joint infection (PJI) typically develops in 1 of 3 ways: through perioperative colonization of the implant, hematogenous seeding caused by a bacteremia, or spread from an infection of the surrounding tissue (Widmer 2001). Furthermore, PJI can be classified in 3 time categories. Early infections occur within 3 months after implantation. Delayed PJI appears 3–24 months after implantation and late PJI after 24 months (Zimmerli et al. 2004). Early and hematogenous PJIs are classified as acute infections, which often have an acute onset and are caused by virulent micro-organisms (Zimmerli et al. 2004). The recommended treatment of an acute PJI is drainage, antibiotics, irrigation, and retention of the prosthesis (DAIR) (Zimmerli et al. 2004). DAIR, for hip and knee prostheses, has a success rate of approximately 70% (Kuiper et al. 2014). In the empirical phase intravenous antibiotics are started blind after surgery until the causative micro-organisms are determined in microbiological cultures. The importance of tailored antibiotic treatment during the targeted phase is well known (Argenson et al. 2019). However, far less literature is available on which antibiotic to use in the empirical phase. The most frequently cultured micro-organisms in PJI are coagulasenegative staphylococci and Staphylococcus aureus (Phillips et al. 2006, Stefánsdóttir et al. 2009, de Vries et al. 2016). Other commonly found micro-organisms are streptococci, gram-negative bacilli, enterococci, and anaerobes (Segawa et al. 1999, Steckelberg and Osmon 2000). In approximately 46% of acute PJI multiple pathogens are found in a patient (de Vries et al. 2016). Reflecting local policy, micro-organism prevalence, and resistance patterns, the antibiotics used in the empirical

456

phase differ between countries and hospitals (Kuiper et al. 2016). Moran et al. (2007) (in the United Kingdom) advised a combination of vancomycin and carbapenem, both broadspectrum antibiotics, as empirical antibiotic regime during DAIR. The population in the UK consisted of a relatively high number of patients with methicillin-resistant Staphylococcus aureus (MRSA) or cephalosporin- and beta-lactam resistance. Sousa et al. (2010) (in Portugal) found similar results and recommended the same regime. Since MRSA prevalence in the Netherlands is much lower than in those countries (van Cleef et al. 2013), double therapy with broadspectrum antibiotics might not be in line with good antibiotic stewardship (Tiemersma et al. 2012). In our orthopedic infection center, flucloxacillin has been the antibiotic of first choice in the empirical phase, because of its high effectiveness against common pathogens like Staphylococcus aureus and coagulase-negative staphylococci. Ideally, the empiric antibiotic treats as many patients as possible, has limited side effects and no restricted usage. Optimal empiric antibiotic therapy contributes to an effective treatment of PJI and will result in better clinical outcomes regarding retention of the prosthesis, complications, and morbidity. We determined which antibiotic should be used in the empirical phase of the treatment of acute PJI with DAIR, based on analyses of local culturing results on PJI samples. We focus on 3 commonly used antibiotics, namely amoxicillin/clavulanic acid, cefazolin, and flucloxacillin.

Patients and methods Using electronic medical records, we retrospectively identified eligible patients from the Catharina Hospital in Eindhoven based on Dutch diagnosis treatment codes for irrigation of the knee or hip joint (CoTG 038640 or 038540, respectively). Patients who were diagnosed with a PJI and underwent DAIR are registered under these codes. We included patients who were diagnosed with a PJI and had a DAIR performed between 2012 and 2016. Additional inclusion criteria entailed a joint prosthesis in situ and the suspicion of PJI within 90 days of index surgery or in case of a hematogenous infection within 3 weeks after onset of PJI signs. Furthermore, only patients from whom periprosthetic tissue was obtained during surgery and sent for microbiological analysis were included. Patients with negative cultures were excluded from the analyses. Patients who underwent DAIR between 2014 and 2016 in the Máxima Medical Center, which was collaborating with the same microbiological laboratory, were included following the same criteria. Prevention and treatment and of PJI To prevent PJI, several recommended measures have been taken in both centers. We swab preoperatively for nasal carriage of S. aureus. When the results are positive for S. aureus,

Acta Orthopaedica 2019; 90 (5): 455–459

patients use chlorhexidine scrub and Bactroban (mupirocin). Cefazolin is used as prophylactic antibiotic, during primary and revision arthroplasty. All DAIR were performed according to the regional treatment protocol for PJI. Part of the DAIR treatment is taking multiple cultures of periprosthetic tissue and synovial fluid before perioperative prophylaxis is administered (Zimmerli et al. 2004). For collection of the cultures, separate clean instruments were used. Typically, 5 cultures per patient were obtained. After debridement, interchangeable parts are replaced and the joint is flushed with at least 3L of saline, using pulsed lavage. Microbiological analyses Following local protocol, samples were incubated on aerobic and anaerobic agar plates and plates were examined for bacterial growth after 2, 7, and 14 days. In the case of bacterial growth, all colony-forming units were determined with MALDI. After determination of the micro-organism(s), antimicrobial resistance was measured using VITEK (BioMérieux, Marcy-l’Étoile, France). For each patient, multiple samples were sent for culturing. If a micro-organism was found in only 1 of the samples, it was excluded from analysis as it was considered contamination. An exception was made for Staphylococcus aureus, as infection with this organism could possibly have such grave consequences that the risk of missing this infectious agent theoretically outweighed the risk of treating contamination. After routine laboratory antibiotic susceptibility testing of the micro-organisms as described above, the measured susceptibility patterns were supplemented with known intrinsic and derived resistance information as described by EUCAST (Leclercq et al. 2013). The resistance pattern for cefazolin was equated to the resistance of cefuroxime. Resistance was further inferred from related antibiotics and literature studies as follows. Corynebacterium, Finegoldia, Granulicatella, Peptoniphilus, Cutibacterium acnes, and streptococcal species were set to be sensitive to cefuroxime when sensitive to penicillin. Streptococci and Granulicatella species were set to be sensitive to amoxicillin/clavulanic acid when sensitive to penicillin, whilst Enterococcus faecium was set to be resistant to amoxicillin/clavulanic acid. For some combinations of micro-organism and antibiotic, the resistance pattern could not be inferred from known patterns or literature studies. Susceptibility was then set as unknown. Analyses The measured and inferred susceptibility patterns were used to determine susceptibility to the different antibiotics, on the levels of both micro-organism and patient. A patient was considered responsive if all cultured micro-organisms were reported as sensitive to the antibiotic. For statistical analysis the McNemar test was used to compare the sensitivity of the evaluated antibiotics cefazolin and amoxicillin/clavulanic acid against flucloxacillin.

Acta Orthopaedica 2019; 90 (5): 455–459

457

Table 1. Patients’ characteristics. Values are number unless otherwise specified

amoxicillin/ clavulanic acid

cefazolin

flucloxacillin

Coagulase negative staphylocci

Factor Male sex Age at time of DAIR, mean (SD) ASA classification 1 2 3 4 Unknown Type of index arthroplasty Total knee arthroplasty Total hip arthroplasty Hemi-arthroplasty Revision total hip arthroplasty Type of acute PJI Early Hematogenous Months after index surgery, mean (SD) Early Hematogenous

N = 91 54 73 (10) 4 52 31 1 3 37 33 13 8

Corynebacterium spp Enterobacter spp Enterococcus faecalis Enterococcus faecium Escherichia coli Hemolytic streptococci (not group A/B) Hemolytic streptococci group B Klebsiella spp Other Pseudomonas aeruginosa

Staphylococcus aureus

75 16 Viridans streptococci

1 (18) 65 (53)

DAIR: debridement, antibiotics, irrigation, and retention of the prosthesis; ASA: American Society of Anesthesiologists’ classification of Physical Health

Ethics, funding, and potential conflicts of interest The Medical Research Ethics Committees United declared that this study did not meet the criteria as stated by the Medical Research Involving Human Subjects Acts (WMO) and the local committee approved this retrospective cohort study (nWMO-2017.51). No funding was received for this study. There are no potential conflicts of interest.

Results Patients We analyzed available data for the 91 patients who presented themselves with a suspicion of PJI and subsequently underwent DAIR (Table 1). A median of 4 (1–8) cultures were acquired per patient. In 14 patients no micro-organisms could be cultured, despite clinical signs of PJI. Another 9 patients were excluded because only 1 positive culture was found, which was considered contamination (except when Staphylococcus aureus was found (n = 4). Culturing results of the remaining 68 patients were included in the analyses. Microbiological analyses In 43 of the 68 patients who had positive microbiological cultures, only a single micro-organism was found, while in 25 cases of acute PJI multiple micro-organisms were found in 1 patient (there were 18 patients with 2 organisms; 5 patients with 3 organisms, 1 patient with 4 organisms, and 1 patient with 5 organisms). 31 different micro-organisms were deter-

Figure 1. Sensitivity of the most common groups of micro-organisms. Sensitive (S) is displayed in blue, resistant (R) in red, and unknown in grey. The blocks’ height represents the prevalence of the micro-organism. The coagulase-negative staphylococci exclude Staphylococcus aureus; the group “other” consists of Finegoldia magna, Granulicatella adiacens, Mycoplasma hominis, Peptoniphilus harei, Acinetobacter genomospecies, and Cutibacterium acnes.

mined, 16 of which were found in only a single patient. Microorganisms were grouped in commonly used relevant biological categories. Categories that were found only in a single patient were collated in the group ‘other’, which consists of Finegoldia magna, Granulicatella adiacens, Mycoplasma hominis, Peptoniphilus harei, Acinetobacter genomospecies, and Cutibacterium acnes. The most commonly found group of micro-organisms was Staphylococcus aureus (34 patients). Other frequently cultured micro-organisms were other coagulase-negative staphylococci, found in 13 patients, and nongroup A/B hemolytic streptococci, determined in 11 patients. In none of the patients were multi-resistant organisms found. Antibiotic susceptibility Micro-organisms For each cultured micro-organism, the sensitivity to the 3 evaluated empiric antibiotic options was determined based on measured and inferred resistance patterns, for each culture in each patient separately (Figure 1). Effectiveness for each patient Patients can be infected by more than a single micro-organism. For treatment choice it is therefore important to evaluate how many patients would have been treated effectively with each of the different antibiotics in the empirical phase of PJI infection with DAIR. For some micro-organisms, the sensitivity to antibiotics was not measured or inferred. We evaluated patients with such micro-organisms as either sensitive or resistant in 2 separate

458

Acta Orthopaedica 2019; 90 (5): 455–459

Table 2. Sensitivity of used antibiotics in percentages with 95% confidence interval based on the binomial distribution Antibiotic

%patients sensitive (unknown = resistant) (unknown = sensitive)

Amoxicillin/clavulanic acid Cefazolin Flucloxacillin

77 (65–86) 79 (68–88) 65 (52–76)

84 (73–92) 81 (70–89) 72 (60–82)

analyses. In both scenarios, a substantially lower sensitivity for flucloxacillin, the currently used empiric antibiotic in our clinic, was seen (Table 2). Conservatively interpreting unknown as resistant, we conclude that patients would have responded better to cefazolin (p = 0.002) or amoxicillin/clavulanic acid (p = 0.008) than the currently used flucloxacillin.

Discussion We gained improved insight into the prevalence of different micro-organisms with their resistance pattern to tailor our treatment protocol. The optimal treatment of an acute PJI is DAIR, in which an empiric antibiotic effective against the most commonly found micro-organisms is essential. The variety of micro-organisms and susceptibility patterns in our population, and as a result which antibiotic would perform best, were unknown prior to this study. Our results show that in one-third of the patients, multiple micro-organisms were found in each patient, which emphasizes the importance of a broad-spectrum antibiotic (Moran et al. 2007). Recent studies confirmed that a higher effectiveness of antibiotics is related to fewer failures and better results in long-term follow-up (Puhto et al. 2015). Comparing 3 antibiotics used in the Netherlands for empiric therapy, we conclude that patients would be more sensitive to treatment with cefazolin and amoxicillin/clavulanic acid than flucloxacillin. Prior to this study, the preferred empiric antibiotic for PJI in our center was flucloxacillin. The results of this study precipitated a change in protocol to cefazolin as antibiotic in the empirical phase. Another option was amoxicillin/ clavulanic acid, the effectiveness of which is comparable to cefazolin. There were 2 reasons to prefer cefazolin to amoxicillin/clavulanic acid. First, cefazolin has few side effects and is widely used as a prophylactic for surgery (Bratzler et al. 2013). Second, allergic reactions are thought to occur more frequently to amoxicillin/clavulanic acid than to cefazolin. In our study, no patients had reported allergies to cefazolin, while eight patients were allergic to amoxicillin/clavulanic acid. The prevalence of the micro-organisms found in our study, with a high number of patients infected with Staphylococcus aureus or coagulase negative staphylococci, is in line with other reports (Phillips et al. 2006, Stefánsdóttir et al. 2009, de Vries et al. 2016). However, recent studies, performed

in northern China and the USA, demonstrated geographical differences in susceptibility spectrums (Ravi et al. 2016, Li et al. 2018). These studies investigated similar groups of patients, but reported a completely different prevalence of micro-organisms and, with this, other empiric antibiotics were recommended. Also compared with Moran et al. (2007) and Sousa et al. (2010), we found less resistant microorganisms, which enables us to use cefazolin instead of glycopeptides and carbapenem. The latter 2 antibiotics are usually reserved for multidrug-resistant infections and have more side effects, and are therefore less eligible as empiric antibiotics. Furthermore, inappropriate use could result in more resistance. Our findings are in line with a study by Schindler et al. (2013), who were dissuaded from the use of a glycopeptides and carbapenem in the empirical phase because no more failures were seen using only a penicillin or cephalosporin. However, only half of the patients in that study had an infected prosthesis; the other included patients had infected osteosyntheses, such as plates or nails (Schindler et al. 2013). The geographical differences in susceptibility spectrums highlights the importance of locally tailored antibiotic treatment protocols. Our study is the first that evaluated the efficiency of empiric antibiotics in PJI in the Netherlands. The inclusion of patients of 2 medical centers might have affected the cultured micro-organisms. However, we think this bias is limited, since the hospitals are located in the same city, perform comparable surgeries, orthopedic surgeons work in cooperation, follow the same treatment protocol, and treat a similar population. We included both early postoperative and hematogenous PJIs in our definition of acute PJI patients, which could potentially create a bias. However, we expect that the effect of this bias is limited as both groups were treated with DAIR, and are typically the result of similar micro-organisms. In our centers these patients are subjected to the same PJI treatment protocol. In future studies, subgroup analysis of microbiological cultures from early postoperative versus hematogenous PJI could be conducted; this was not possible in our patient group, because there was only a small group of patients with hematogenous PJI. In summary, in this study we characterized which microorganisms caused acute PJI and analyzed their pattern of resistance. With this information we compared the efficiency of three commonly used antibiotics when treating PJI, namely amoxicillin/clavulanic acid, cefazolin, and flucloxacillin. Flucloxacillin was the local empiric antibiotic of choice but proved to be a suboptimal choice in our patient population treated with DAIR. We have changed our practice to the use of cefazolin as the preferred antibiotic in the empirical treatment of acute PJI with DAIR. Optimizing antibiotic therapy could potentially contribute to more effective treatment of PJI and hence logically result in better clinical outcomes and less morbidity. Since susceptibility spectrums differ geographically, it is recommended that centers study their local data to evaluate which antibiotic will optimally treat their PJI patients.

Acta Orthopaedica 2019; 90 (5): 455–459

AHC, JHJE: Data collection, analysis, and interpretation. Drafting the article. RJAW, RWTMK: Conception or design of the work. Data interpretation. Critical revision of the article. JGEH: Data interpretation. Critical revision of the article. MW: Conception or design of the work. Data collection and interpretation. Critical revision of the article. JMF, MCS: Conception or design of the work. Data collection, analysis, and interpretation. Drafting the article. Acta thanks Jon Goosen and Tina Strømdal Wik for help with peer review of this study.

Argenson J N, Arndt M, Babis G, Battenberg A, Budhiparama N, Catani F, et al. Hip and knee section, treatment, debridement and retention of implant: Proceedings of International Consensus on Orthopedic Infections. J Arthroplasty 2019; 34(2): S399–419. Bratzler D W, Dellinger E P, Olsen K M, Perl T M, Auwaerter P G, Bolon M K et al. Clinical practice guidelines for antimicrobial prophylaxis in surgery. Am J Heal Pharm 2013; 70(3): 195-283. de Vries L, van der Weegen W, Neve W, Das H, Ridwan B, Steens J. The effectiveness of debridement, antibiotics and irrigation for periprosthetic joint infections after primary hip and knee arthroplasty: a 15 years retrospective study in two community hospitals in the Netherlands. J Bone Joint Infect 2016; 1: 20-4. Kuiper J W P, Willink R T, Moojen D J F, van den Bekerom M P, Colen S. Treatment of acute periprosthetic infections with prosthesis retention: review of current concepts. World J Orthop 2014; 5(5): 667. Kuiper J W P, Vos S C J, Burger B J, Colen S. Variety in diagnosis and treatment of periprosthetic joint infections in Belgium and the Netherlands. Acta Orthop Belg 2016; 82(2): 149-60. Leclercq R, Cantón R, Brown D F J, Giske C G, Heisig P, Macgowan A P, et al. EUCAST expert rules in antimicrobial susceptibility testing. Clin Microbiol Infect 2013; 19(2): 141-60. Li Z L, Hou Y F, Zhang B Q, Chen Y F, Wang Q, Wang K, Chen Z Y, Li X W, Lin J H. Identifying common pathogens in periprosthetic joint infection and testing drug-resistance rate for different antibiotics: a prospective, single center study in Beijing. Orthop Surg 2018; 10(3): 235-40. Moran E, Masters S, Berendt A R, McLardy-Smith P, Byren I, Atkins B L. Guiding empirical antibiotic therapy in orthopaedics: The microbiology of

459

prosthetic joint infection managed by debridement, irrigation and prosthesis retention. J Infect 2007; 55(1): 1-7. Phillips J E, Crane T P, Noy M, Elliott T S, Grimer R J. The incidence of deep prosthetic infections in a specialist orthopaedic hospital: a 15-year prospective survey. J Bone Joint Surg Br 2006; 88-B(7): 943-8. Puhto A P, Puhto T, Niinimäki T, Ohtonen P, Leppilahti J, Syrjälä H. Predictors of treatment outcome in prosthetic joint infections treated with prosthesis retention. Int Orthop 2015; 39(9): 1785-91. Ravi S, Zhu M, Luey C, Young S W. Antibiotic resistance in early periprosthetic joint infection. ANZ J Surg 2016; 86(12): 1014-18. Schindler M, Gamulin A, Belaieff W, Francescato M, Bonvin A, Graf V et al. No need for broad-spectrum empirical antibiotic coverage after surgical drainage of orthopaedic implant infections. Int Orthop 2013; 37(10): 2025-30. Sousa R, Pereira A, Massada M, Vieira Da Silva M, Lemos R, Costa E Castro J. Empirical antibiotic therapy in prosthetic joint infections. Acta Orthop Belg 2010; 76(2): 254-9. Segawa H, Tsukayama D T, Kyle R F, Becker D, Gustilo R B. Infection after total knee arthroplasty: a retrospective study of the treatment of eighty-one infections. J Bone Joint Surg Ser A 1999; 81(10): 1434-45. Steckelberg J M, Osmon D R. Prosthetic joint infections associated with indwelling medical devices, 3rd ed. Washington, DC: American Society of Microbiology 2000; p 173-209. Stefánsdóttir A, Johansson D, Knutson K, Lidgren L, Robertsson O. Microbiology of the infected knee arthroplasty: report from the Swedish Knee Arthroplasty Register on 426 surgically revised cases. Scand J Infect Dis 2009; 41(11-12): 831-40. Tiemersma E W, Bronzwaer S L A M, Lyytikäinen O, Degener J E, Schrijnemakers P, Bruinsma N, et al. Methicillin-resistant Staphylococcus aureus in Europe, 1999-2002. Emerg Infect Dis 2012; 10(9): 1627-34. van Cleef B A G L, van Benthem B H B, Haenen A P J, Bosch T, Monen J, Kluytmans J A J W. Low incidence of livestock-associated methicillinresistant Staphylococcus aureus bacteraemia in the Netherlands in 2009. PLoS One 2013; 8(8): e73096. Widmer A F. New developments in diagnosis and treatment of infection in orthopedic implants. Clin Infect Dis 2001; 33(s2): S94-106. Zimmerli W, Trampuz A, Ochsner P E. Prosthetic-joint infections. N Engl J Med 2004; 351(16): 1645-54.

460

Acta Orthopaedica 2019; 90 (5): 460–466

Uncemented monoblock trabecular metal posterior stabilized high-flex total knee arthroplasty: similar pattern of migration to the cruciate-retaining design — a prospective radiostereometric analysis (RSA) and clinical evaluation of 40 patients (49 knees) 60 years or younger with 9 years’ follow-up Radoslaw WOJTOWICZ 1, Anders HENRICSON 2, Kjell G NILSSON 1, and Sead CRNALIC 1 1 Department of Surgical and Perioperative Sciences, Umeå University, Umeå, Sweden; 2 Department of Orthopaedics, Falu Hospital, Falun, Sweden Correspondence: kjell.g.nilsson@umu.se Submitted 2019-01-16. Accepted 2019-04-14.

Background and purpose — Uncemented monoblock cruciate retaining (CR) trabecular metal (TM) tibial components in total knee arthroplasty (TKA) work well in the long-term perspective in patients ≤ 60 years. Younger persons expect nearly normal knee flexion after TKA, but CR implants generally achieve less knee flexion compared with posterior stabilized (PS) implants. Cemented PS implants have higher revision rate than CR implants. Can an uncemented monoblock PS TM implant be used safely in younger patients? Patients and methods — 40 patients (49 knees) age ≤ 60 years with primary (20 knees) or posttraumatic osteoarthritis (OA) were operated with a high-flex TKA using an uncemented monoblock PS TM tibial component. Knees were evaluated with radiostereometric analysis (RSA) a mean 3 days (1–5) postoperatively, and thereafter at 6 weeks, 3 months, 1, 2, 5, and 9 years. Clinical outcome was measured with patient-related outcome measures (PROMs). Results — The implants showed a pattern of migration with initial large migration followed by early stabilization lasting up to 9 years, a pattern known to be compatible with good long-term results. Clinical and radiological outcome was excellent with 38 of the 40 patients being satisfied or very satisfied with the procedure and bone apposition to the entire implant surface in 46 of 49 knees. Mean knee flexion was 130°. 1 knee was revised at 3 months due to medial tibial condyle collapse. Interpretation — The uncemented monoblock PS TM implant works well in younger persons operated with TKA due to primary or secondary OA.

Normal knee flexion after TKA is often expected by patients, not least by younger patients. Posterior stabilized implants (PS) have consistently been shown to lead to greater flexion compared with cruciate retaining (CR) implants (Jacobs et al. 2005, Bercik et al. 2013, Li et al. 2014, Jiang et al. 2016). However, the revision rates for PS implants have been found to be higher compared with CR implants (Comfort et al. 2014, Vertullo et al. 2017), perhaps due to the higher constraint of the PS articulation leading to larger forces at the implant–bone interface (Catani et al. 2004). Uncemented trabecular metal (TM) tibial implants have shown excellent clinical results in younger patients in the medium-term perspective (Kamath et al. 2011, FernandezFairen et al. 2013). This may be due to inherent properties of the trabecular metal, seemingly enhancing biologic fixation (Rahbek et al. 2005, Bullens et al. 2010, Sagomonyants et al. 2011, Sambaziotis et al. 2012). In a radiostereometric (RSA) study of younger persons with 10-years’ follow-up, the CR uncemented monoblock TM implant displayed no loosening and a pattern of migration with early stabilization lasting up to 10 years, indicating good long-term performance as regards fixation (Henricson and Nilsson 2016). However, the mean postoperative knee flexion in that study was only 110°. Posterior stabilized TM implants, thus, would be an attractive solution for the younger patient. However, there are few clinical long-term follow-up studies of such components in the literature, and to our knowledge no long-term RSA studies. Most RSA studies have focused on the magnitude of migration during the early postoperative years. Pijls et al. (2012b, 2018), in 2 systematic reviews of mostly cemented implants found that the magnitude of migration at 6 months and 1 year was associated with late revision. However, the pattern of migration over time is also important in determining long-

Acta Orthopaedica 2019; 90 (5): 460–466

All TKAs operated between January 2007 and November 2008 n = 216 (206 patients) Excluded (n = 167): – older than 60 years, 139 (138 patients) – declined to participate, 28 (28 patients) Included in the study n = 49 (40 patients) RSA at 6 weeks: 48 3 months: 48 1 year: 48 2 years: 47 5 years: 45 9 years: 45

Excluded: 6 weeks: missed RSA, 1 3 months: missed RSA, 1 1 year: was revised, 1 2 years: moved abroad, 1 5 years: moved abroad, 1 – ” – : died, 1

461

Table 1. Preoperative clinical data Factor Mean Median Range Age (years) 55 56 40–60 Weight (kg) 88 85 57–120 Length (cm) 171 169 153–198 BMI 30 30 21–43 Extension lag (°) 13 15 0–30 Flexion (°) 117 125 80–135 Range of motion (°) 104 110 60–135 Knee Society (KS) score a 17 14 0–48 KS function score a 48 50 25–55 Knee alignment (°) b 176 174 163–196 a Knee Society knee and Function score (max. 100). b Hip Knee Ankle angle (varus < 180°, valgus > 180°).

Figure 1. Flowchart of the patients. TKA = total knee arthroplasty, RSA = radiostereometric analysis.

term performance (Wilson et al. 2012, Henricson and Nilsson 2016, Molt et al. 2016, van Hamersveld et al. 2017, 2018). We determined the migration, clinical, and radiological outcome of the NexGen LPS-Flex TM Monoblock (Zimmer, Warsaw, IN, USA) tibial component up to 9 years. The primary objective was to compare the pattern of migration of the PS implant with the CR variety of the same implant, results of which have been presented recently (Henricson and Nilsson 2016). Secondary objectives were to assess clinical results as measured by patient and clinician reported outcomes measures, postoperative knee motion, and radiological result.

Patients and methods The patients were recruited from the waiting list for knee arthroplasty at the Department of Orthopedics, Umeå University Hospital. Inclusion criteria were primary or secondary osteoarthritis (OA) with symptoms warranting knee arthroplasty, age 60 years or younger, and willingness and ability to participate in the study (Figure 1, Table 1). 29 knees were operated due to posttraumatic OA; 9 knees had had an anterior cruciate ligament lesion (4 operated, 5 non-operated), 10 meniscal surgery, 3 because of previous tibial osteotomy, 2 recurrent patellar luxation, 2 cartilage transplantation, and 1 tibial condyle fracture. 1 patient (2 knees) had OA secondary to psoriatic arthropathy. The operations were performed under tourniquet by KGN. Ligaments were balanced as needed and the distal femoral and proximal tibial cuts were made using intramedullary guides. It was aimed to insert the tibial component at a right angle to the mechanical axis in the frontal plane and with 6–7° posterior tilt. All femoral components were cemented using Palacos with gentamicin bone cement (Heraeus Kulzer GmbH, Wehrheim, Germany). In 2 knees an all polyethylene patellar component was inserted. In 9 patients (6 women) both knees

were operated, with a mean 6 months (5–16) between procedures. All patients received a PS Monoblock TM tibial component and a NexGen LPS-Flex femoral component (Zimmer, Warsaw, IN, USA). Postoperatively the patients were allowed immediate weight-bearing. For the RSA analysis, 6 tantalum markers were inserted into the polyethylene of the monoblock tibial component in standardized positions, and another 9 markers were spread out in the proximal tibial metaphysis. The initial postoperative RSA examination was performed a mean 3 days (1–5) after the operation. Subsequent examinations were performed at 6 weeks, at 3, 12, and 24 months, and at 5 and 9 years postoperatively. The patients were examined lying supine with the knee placed within a calibration cage (Cage 10, RSA Biomedical, Umeå, Sweden). At the initial postoperative examination the knee was positioned with its anatomical axes parallel to the cardinal axes of the calibration cage. Analysis of implant migration was performed using UmRSA software (version 6.0; RSA Biomedical, Umeå, Sweden). The relative movements (rotations and translations) of the tibial component were measured in relation to the markers in the proximal tibial metaphysis as the fixed reference segment. The rotations were expressed about the transverse, longitudinal, and sagittal axes. Translations of the tibial tray were measured at 5 standardized positions (Nilsson et al. 1991) at the periphery of the tray. In each implant, the largest negative value for translation along the vertical (y) axis was called maximum subsidence, and the largest positive y-translation was called lift-off. Maximum total point motion (MTPM, Ryd 1986) was defined as the length of the 3-D translation vector of the standardized point at the tibial tray periphery that moved the most. The change in MTPM between 1 and 2 years (Ryd et al. 1995), between 2 and 5 years (Wilson et al. 2012), and between 5 and 9 years was calculated. The repeatability of the RSA measurements was calculated using double examinations performed postoperatively, and after 1, 2, and 9 years as described by Ranstam et al. (2000). Statistically significant rotations at the 95% significance level

462

Acta Orthopaedica 2019; 90 (5): 460–466

Maximum migration along the y-axis (mm)

MTPM (mm) 1.2

0.0

1.0

–0.2

0.8

–0.4

0.6

–0.6

0.4

–0.8

Absolute rotation around the x-axis (°) 0.8

PS CR 0.6

0.4

0.2 0.2

–1.0

PS CR 0

100

120

Months after index operation

Figure 2. Maximum migration (MTPM) for the NexGen Trabecular Metal posterior stabilized (PS) monoblock tibial component (red). For comparison values for the cruciate retaining (CR) variety of the same implant (blue) (Henricson and Nilsson 2016). Values are mean (95% confidence interval).

–1.2

PS CR 0

100

120

Months after index operation

Figure 3. Maximum subsidence (negative y-axis translation) for the NexGen Trabecular Metal posterior stabilized (PS) monoblock tibial component (red). For legend, see Figure 2.

were > 0.23° (transverse), > 0.21° (longitudinal), and > 0.21° (sagittal). Corresponding value for translations was > 0.10 mm. Pre- and postoperative knee alignment was measured as the hip–knee–ankle (HKA) angle, with varus being < 180° and valgus > 180°. Alignment of the tibial component in relation to the tibia was measured as described by Nilsson et al. (1991). Clinical evaluation was performed using the Knee Society knee score at all follow-ups. At the 9-year follow-up the Knee Injury and Osteoarthritis Outcome Score (KOOS) and Forgotten Joint Score (FJS) (Behrend et al. 2012) were also obtained. In addition, at 9 years the patients were asked to grade their satisfaction of the operation as very satisfied, satisfied, somewhat dissatisfied, or dissatisfied. The implant–bone interface was analyzed at the 9-year followup according to Hayakawa et al. (2014). Type A is defined as new bone formation at the base of the tibial tray and around the pegs, as well as longitudinal trabecular thickening at the distal ends of the pegs. In Type B, there is only longitudinal trabecular thickening at the distal ends of the pegs. Type C is defined as no new bone changes or even existence of radiolucencies. Not all knees could be analyzed with RSA at all follow-ups (Figure 1). Thus, at 6 weeks and 3 months 1 knee each missed RSA. 1 patient was revised after 3 months, 2 moved abroad, and 1 died of cancer. Statistics The main purpose of the study was to analyze the magnitude and pattern of migration over time. Therefore, we analyzed absolute values of migration parameters for which both negative and positive values were possible. Since the absolute values of rotation, MTPM, and maximum subsidence/lift-off were not normally distributed (according to Shapiro–Wilk test for normality), the data were log transformed to achieve nor-

0.0

100

120

Months after index operation

Figure 4. Rotation around the transverse (x-) axis of the knee (absolute values). For legend, see Figure 2. The majority of the TM PS implants rotated into posterior tilt.

mality. Then mean values and 95% confidence intervals (CI) could be calculated and thereafter re-transformed back to the original scale. This was done in order to facilitate comparison with data on the monoblock TM CR implant published recently (Henricson and Nilsson 2016). In this calculation only data for unilaterally operated knees and the first knee in bilaterally operated patients were used. For information on individual pattern of migration, data for all knees were displayed in 1 graph (Figure 7). The clinical results (KOOS, FJS, Knee Society scores) were reported only for the first knee in bilaterally operated persons. Ethics, funding, and potential conflicts of interest The study was approved by the Ethics Committee of Umeå University (entry no. 07-135M). All patients who accepted participation signed an informed consent. The study was not registered at a public registry because the study started in 2007 when registration was still in its early stages. The study was supported by institutional grants from Zimmer-Biomet and Umeå University, Umeå, Sweden. Except for sponsoring the RSA analysis, Zimmer-Biomet had no influence on study planning or implementation. KGN is engaged at ZimmerBiomet speaker’s bureau, whereas the other authors declare no conflict of interest.

Results Implant migration For all migration parameters the pattern of migration was similar, with most of the migration occurring within the initial 3 months and thereafter stabilization over time (Figures 2–6). Compared with the monoblock CR TM implants, the

Acta Orthopaedica 2019; 90 (5): 460–466

463

Absolute rotation around the y-axis (°)

Absolute rotation around the z-axis (°)

Individual MTPMs for the PS implants (mm)

0.8

PS CR

Revised 8

0.6

0.4

0.2

0.2 2

0.0

100

120

0.0

Months after index operation

Figure 5. Rotation around the longitudinal (y-) axis of the knee (absolute values). For legend, see Figure 2. The majority of the TM PS implants rotated externally.

100

120

Months after index operation

Figure 6. Rotation around the sagittal (z-) axis of the knee (absolute values). For legend, see Figure 2. Rotation into varus or valgus was evenly distributed among the TM PS implants.

Table 2. Clinical and radiological data at the 9-year follow-up Factor Mean SD Median Range Knee Osteoarthritis Outcome Score (KOOS) a Symptoms 85 13 86 64–100 Pain 87 15 92 42–100 Function, daily living 86 16 91 35–100 Function, sports and recreational activities 53 31 50 5–100 Quality of life 72 25 81 31–100 Forgotten Joint Score (FJS) a 65 26 73 23–100 Knee Society Knee Score a 97 3.9 100 80–100 Knee Society Pain Score b 50 1.3 50 40–50 Knee Society Function Score a 98 9.2 100 60–100 Extension lag (°) 2 2.2 0 0–15 Flexion (°) 131 10 130 120–145 Knee alignment postop (°) c 179 3.0 179 172–185 Tibial component alignment frontal plane (°) d 89 2.0 89 83–93 posterior slope (°) 6 2.3 6 2–10 a Maximum 100. b 50 = no pain c Hip–knee–ankle angle, < 180° = varus, > 180° = valgus alignment. d < 90° = varus, > 90° = valgus alignment.

time for stabilization was somewhat longer. The majority of the implants displayed rotation into posterior tilt and external rotation, whereas rotation into varus or valgus was evenly distributed. Rotation around the transverse and sagittal axes was associated with a mean maximum subsidence of the periphery (usually posteriorly) of a mean 0.63 (CI 0.41–0.87) mm (Figure 3). Maximum lift-off of the edge of the tibial tray was small and occurred only occasionally, mostly anteriorly. Between 1 and 2 years one implant displayed MTPM > 0.2 mm, and between 2 and 5 years 4 implants displayed MTPM > 0.3 mm. Between 5 and 9 years all implants except one were stable. Individual MTPM values are shown in Figure 7.

100

120

Months after index operation

Figure 7. Individual MTPM values for the TM PS implants.

Radiographic findings The postoperative alignment of the knee and the tibial component for all implants is given in Table 2. The majority of the knees (n = 37) were aligned 180° ± 3°. In the frontal plane the tibial trays were inserted in slight varus (mean 89° [SD 2.0°]). The mean posterior slope was 6° (SD 2.0°). At 9 years all but 2 knees displayed Hayakawa Type A implant–bone interface, 1 knee type B, and 1 knee type C. Clinical findings The results for KOOS, FJS, and Knee Society scores at 9 years are listed in Table 2. Of the 36 patients examined, 30 were very satisfied with their knee operation, 4 were satisfied, 1 was somewhat dissatisfied, and 1 was dissatisfied (see also below). Mean knee flexion was 130° and only 1 patient had a slight flexion contracture. Complications 1 patient (female, 55 years) with bilateral operations staged 6 months apart had her second (right) operated knee revised 3 months postoperatively. Postoperative knee alignment was 3° varus and tibial component alignment 4° varus. The tibial implant subsided 9 mm medially within the first weeks postoperatively, resulting in a severe varus malalignment. There were no signs of infection. At revision 3 months after the index operation, the tibial implant was firmly fixed to bone and had to be cut out with saw and chisels. Bone underneath and adjacent to the implant showed signs of bone necrosis on microscopic analysis. A stemmed revision tibial component was inserted. The first (left) knee operated on this patient had postoperative HKA angle 180° and tibial component alignment of 1° varus, and functioned very well during the follow-up. 2 patients developed early (within 2 weeks postoperatively) wound complications and peri-prosthetic joint infection, which required early debridement, antibiotics, irrigation, and

464

implant retention (DAIR). Both infections healed. 1 of the patients moved abroad and stated by telephone enquiry 5 and 9 years postoperatively that the knee prosthesis was still in place and functioned well. The other patient scored “very satisfied” at the 9-year follow-up.

Discussion This study shows that a TM posterior-stabilized monoblock tibial component together with a high-flex femoral component in patients 60 years or younger displays a benign pattern of migration up to 9 years, i.e., initial migration followed by early stabilization lasting up to 9 years. Mean knee flexion was 130° and the radiological and clinical results were good with 34 of the 36 patients being very satisfied or satisfied with the procedure at the 9-year follow-up. Most RSA studies have focused on the magnitude of migration during the early postoperative years. Pijls et al. (2012b, 2018), in 2 systematic reviews of mostly cemented implants, found that the magnitude of migration at 6 months and 1 year was associated with rate of late revision. However, several RSA studies in recent years have stressed that the pattern of migration over time is also of importance in evaluating quality of implant fixation (Pijls et al. 2012a, Henricson and Nilsson 2016, Molt et al. 2016, van Hamersveld et al. 2017, 2018). Indeed, Pijls et al. (2018) in their study, albeit advocating only short RSA follow-up to determine long-term performance, stressed that “a particular migration pattern may be normal for one TKA design or fixation, but pathological for another,” indicating a need for longer RSA follow-up. The migratory pattern for the uncemented monoblock PS TM implant was similar to what has been found for other types of uncemented tibial components (Nilsson et al. 2006, Henricson and Nilsson 2016), with almost all migration occurring during the early postoperative months followed by stabilization. Bellemans (1999) in an RSA study on sheep found that implant stability was compatible with bony ingrowth. The pattern of migration and the radiological findings in the present study thus make it probable that bony ingrowth did occur. Indeed, in the retrieval study of Hanzlik et al. (2013) bony ingrowth was found in all retrieved TM tibial components. The pattern of migration for the monoblock TM PS implant was similar to its CR counterpart. In registry-based comparisons of cemented PS and CR implants, revision rates have been found to be larger for the PS design (Comfort et al. 2014, Vertullo et al. 2017). This difference in favor of the CR implant has been suggested to be caused by selection bias (Ritter et al. 2014). This statement was, however, refuted by Vertullo et al. (2017), who, by eliminating the selection bias in the analysis, showed cemented PS implants to have larger revision rate than CR implants up to 13 years’ follow-up. This implies that the inferior fixation of the cemented PS implants may be due to the cement mantle not being strong enough to

Acta Orthopaedica 2019; 90 (5): 460–466

resist the increased load to the implant–bone interface caused by the larger constraint in the PS implants (Catani et al. 2004). We did not see the potential negative effect of the increased constraint of the PS implant on fixation. The increased forces at the interface in the PS design did not translate into larger migration or lack of stabilization over time. Also, in all but 2 knees there were no radiolucent lines at the implant–bone interface, but rather new bone formation at the base of the tibial tray and around the pegs. This finding may reflect some specific features of how TM interacts with bone. TM has a very high coefficient of friction (Zhang et al. 1999), a porosity, pore size, and modulus of elasticity resembling cancellous bone (Zardiackas et al. 2001, Levine et al. 2006), and ability to stimulate attachment of, and mineralization by, osteoblasts (Sagomonyants et al. 2011), all of which are factors that favors fixation. The resemblance to cancellous bone also makes TM less likely to induce stress shielding (Minoda et al. 2013). The positive results of the monoblock TM implant may, however, not necessarily be valid for modular TM implants, where the tibial tray is a combination of a titanium shell and TM coating, a construct that is less elastic (Zandee van Rilland et al. 2015, Behery et al. 2017). Male sex, posttraumatic OA, uncemented fixation, and low patient age are all considered risk factors for inferior results after knee arthroplasty (Scott et al. 2015, Watters et al. 2017, Putman et al. 2018). In our study there were equally as many men as women, the mean age was 55 years, all tibial implants were uncemented and PS, and more than half had posttraumatic OA. The low incidence of complications and benign pattern of migration we found indicate that the monoblock TM PS implant may be able to counteract the negative effects of these factors. There are few articles published on younger persons operated with TKA using KOOS and FJS as patient-reported outcome measures (PROMs). Parratte et al. (2015), in a study comparing total knee arthroplasty with bicondylar knee arthroplasty (medial unicondylar plus patello-femoral implant) in patients with a mean age of 61 and minimum 2-year follow-up, found KOOS results similar to the present study for the total knee arthroplasties. The results for the bicondylar implants were, however, higher. Also, the results for FJS in the total knee arthroplasty group were on a par with our study. Mean knee flexion in our study was 130° which is much higher than the mean of 110° that was found for the TM CR implants in the previous study (Henricson and Nilsson 2016). Also, extension lag was seldom found. These findings corroborate the results from other studies of better knee flexion in PS designs (Li et al. 2014, Jiang et al. 2016), without the expense of inferior fixation. 2 patients sustained early postoperative infections that were treated with debridement, antibiotics, and implant retention (DAIR); however, due to the monoblock nature of the implant no modular parts could be exchanged. Both infections healed with no signs of recurrent infection during the follow-up. It has been suggested that tantalum may have antimicrobial

Acta Orthopaedica 2019; 90 (5): 460–466

properties (Schildhauer et al. 2006), but whether this was a contributing factor to the positive outcome in the present study cannot be determined. There was 1 reoperation of the right knee in a patient who had had the left knee operated 6 months earlier. The patient could not recall a trauma and there were no signs of infection. At revision the implant was firmly fixed to bone, but the bone beneath the tray was necrotic. The reason for this complication is unclear. Similar sized implants were used in both knees. However, in the left knee the knee as well as the tibial component was nearly normally aligned, whereas there was varus alignment of both the knee and tibial component in the right knee. It may as well be that this combined varus malalignment led to medial condyle overload, bone necrosis, and implant subsidence. The strengths of the study are the long follow-up using RSA on comparatively many knees. Weaknesses are loss of RSA follow-up for various reasons of 4 patients; however, clinical results are lacking for only 1. In summary, this study shows that the uncemented mono block TM PS tibial component in patients 60 years or younger yields high knee flexion and displays a pattern of migration up to 9 years indicating a good long-term prognosis as regards fixation and clinical result.

KGN initiated the study, performed the surgery, wrote and edited the manuscript. RW, AH, and SC wrote and edited the manuscript. Acta thanks Stephan Maximilian Röhrl and Sören Toksvig-Larsen for help with peer review of this study.

Behery O A, Kearns S M, Rabonowitz J M, Levine B R. Cementless vs cemented tibial fixation in primary total knee arthroplasty. J Arthroplasty 2017; 32: 1510-15. Behrend H, Giesinger K, Giesinger J M, Kuster M S. The “forgotten joint” as the ultimate goal in joint arthroplasty: validation of a new patient-reported outcome measure. J Arthroplasty 2012; 27: 430-6. Bellemans J. Osseointegration in porous coated knee arthroplasty: the influence of component coating type in sheep. Acta Orthop Scand 1999; (Suppl. 288): 1-35. Bercik M J, Joshi A, Parvizi J. Posterior cruciate-retaining versus posteriorstabilized total knee arthroplasty – a meta-analysis. J Arthroplasty 2013; 28: 439-44. Bullens H J, Screader H W B, de Waal Malefijt M C, Verdonschot N, Buma P. The presence of periosteum is essential for the healing of large diaphyseal segmental defects reconstructed with trabecular metal: a study in the femur of goats. J Biomed Mater Res B Appl Biomater 2010; 92: 24-31. Catani F, Leardini A, Ensini A, Cucca G, Bragonzoni L, Toksvig-Larsen S, Giannini S. The stability of the cemented tibial component of total knee arthroplasty: posterior cruciate-retaining versus posterior-stabilized design. J Arthroplasty 2004; 19: 775-82. Comfort T, Baste V, Froufe M A, Namba R, Bordini B, Robertsson O, Cafri G, Paxton E, Sedrakyan A, Graves S. International comparative evaluation of fixed-bearing non-posterior-stabilized and posterior-stabilized total knee replacements. J Bone Joint Surg Am 2014; 96 (Suppl. (E)): 65-72.

465

Fernandez-Fairen M, Hernández-Vaquero D, Murcia A, Llopis R. Trabecular metal in total knee arthroplasty associated with higher knee scores: a randomized controlled trial. Clin Orthop Relat Res 2013; 471: 3543-53. Hanzlik J A, Day J S, et al. Bone ingrowth in well-fixed retrieved porous tantalum implants. J Arthroplasty 2013; 28: 922-7. Hayakawa K, Date H, Tsujimura S, Nojiri S, Yamada H, Nakagawa K. Midterm results of total knee arthroplasty with a porous tantalum monoblock tibial component. The Knee 2014; 21: 199-203. Henricson A, Nilsson K G. Trabecular metal tibial knee component still stable at 10 years: an RSA study on 33 patients less than 60 years of age. Acta Orthop 2016; 87: 504-10. Jacobs W C H, Clement D, Wymenga A. Retention versus removal of the posterior cruciate ligament in total knee replacement. A systematic literature review within the Cochrane framework. Acta Orthop 2005; 76: 757-68. Jiang C, Liu Z, Wang Y, Bian Y, Feng B, Weng X. Posterior cruciate ligament retention versus posterior stabilization for total knee arthroplasty: a metaanalysis. PLoS ONE 2016; 11(1): e0147865. Kamath A F, Lee G-C, Sheth N P, Nelson C L, Garino J P, Israelite C L. Prospective results of tantalum monoblock tibia in total knee arthroplasty. J Arthroplasty 2011; 26(8): 1390-4. Levine B R, Sporer S, Poggie R A, Della Valle C J, Jacobs J J. Experimental and clinical performance of porous tantalum in orthopedic surgery. Biomaterials 2006; 7: 4671-81. Li N, Tan Y, Deng Y, Chen L. Posterior cruciate-retaining versus posterior stabilized total knee arthroplasty: a meta-analysis of randomized controlled trials. Knee Surg Sports Traumatol Arthrosc 2014; 22: 556-64. Minoda Y, Kobayashi A, Ikebuchi M, Iwaki H, Inori F, Nakamura H. Porous tantalum tibial component prevents periprosthetic loss of bone mineral density after total knee arthroplasty for five years: a matched cohort study. J Arthroplasty 2013; 28: 1760-4. Molt M, Ryd L, Toksvig-Larsen S. A randomized RSA study concentrating especially on continuous migration. Acta Orthop 2016; 87: 262-7. Nilsson K G, Kärrhom J, Ekelund L, Magnusson P. Evaluation of micromotion in cemented vs. uncemented knee arthroplasty in osteoarthritis and rheumatoid arthritis: randomized study using roentgen stereophotogrammetric analysis. J Arthroplasty 1991; 6: 265-78. Nilsson K G, Henricson A, Norgren B, Dalén T. Uncemented HA-coated implant is the optimum fixation for TKA in the young patient. Clin Orthop Rel Res 2006; 448: 129-38. Parratte S, Ollivier M, Opsomer G, Lunebourg A, Argenson J N, Thienpont E. Is knee function better with contemporary modular bicompartmental arthroplasty compared to total knee arthroplasty? Short-term outcomes of a prospective matched study including 68 cases. Orthop Traum Surg Res 2015; 101: 547-52. Pijls B G, Valstar E R, Kaptein B L, Fiocco M, Nelissen R G H H. The beneficial effect of hydroxyapatite lasts: a randomized radiostereometric trial comparing hydroxyapatite-coated, uncoated, and cemented tibial components for up to 16 years. Acta Orthop 2012a; 83: 135-41. Pijls B G, Valstar E R, Nouta K A, Plevier J W, Fiocco M, Middeldorp S, Nelissen R G H H. Early migration of tibial components is associated with late revision; a systemic review and meta-analysis of 21 000 knee arthroplasties. Acta Orthop 2012b; 83: 614-24. Pijls B G, Plevier J W M, Nelissen R G H H. RSA migration of total knee replacements: a systematic review and meta-analysis. Acta Orthop 2018; 89: 320-8. Putman S, Argenson J N, Bonnevialle P, Ehlinger M, Vie P, Leclercq S, Bizot P, Lustig S, Parratte S, Ramdane N, Comar N. Ten-year survival and complications of total knee arthroplasty for osteoarthritis secondary to trauma or surgery: a French multicenter study of 263 patients. Orthop Trauma Surg Res 2018; 104: 161-4. Rahbek O, Kold S, Zippor B, Overgaard S, Soballe K. Particle migration and gap healing around trabecular metal implants. Int Orthop 2005; 29: 368-74. Ranstam J, Ryd L, Önsten I. Accurate accuracy measurement: review of basic principles. Acta Orthop Scand 2000; 71: 106-8.

466

Ritter M A, Davis K E, Farris A, Keating E M, Faris P M. The surgeon’s role in relative success of PCL-retaining and PCL-substituting total knee arthroplasty. HSS J 2014; 10: 107-15. Ryd L. Micromotion in knee arthroplasty: a roentgen stereophotogrammetric analysis of tibial component fixation. Act Orthop Scand 1986; (Suppl. 220): 1-80. Ryd L, Albrektsson B E, Carlsson L, Dansgard F, Herberts P, Lindstrand A, Regner L, Toksvig-Larsen S. Roentgen stereophotogrammetric analysis as a predictor of mechanical loosening of knee prostheses. J Bone Joint Surg Br 1995; 77: 377-83. Sagomonyants K B, Hakim-Zargar M, Jhaveri A, Aronow A S, Gronowicz G. Porous tantalum stimulates the proliferation and osteogenesis of osteoblasts from elderly female patients. J Orthop Res 2011; 29: 609-16. Sambaziotis C, Lovy A J, Koller K E, Bloebaum R D, Hirsh D M, Kim S J. Histologic retrieval analysis of a porous tantalum metal implant in an infected primary total knee arthroplasty. J Arthroplasty 2012; 27: 1413e5-1413e9. Schildhauer T A, Ribie B, Muhr G, Köller M. Bacterial adherence to tantalum versus commonly used orthopedic implant materials. J Orthop Trauma 2006; 20: 476-84. Scott C E H, Davidson E, MacDonald D J, White T O, Keating J F. Total knee arthroplasty following tibial plateau fracture: a matched cohort study. Bone Joint J 2015; 97-B: 532-8. van Hamersveld K T, Marang-van de Mheen P J, Tsonaka R, Valstar E R, Toksvig-Larsen S. Fixation and clinical outcome of uncemented periapatite-coated versus cemented total knee arthroplasty: five-year follow-up

Acta Orthopaedica 2019; 90 (5): 460–466

of a randomized controlled trial using radiostereometric analysis (RSA). Bone Joint J 2017; 99-B: 1467-76. van Hamersveld K T, Marang-van de Mhen P J, Nelissen R G H H, ToksvigLarsen S. Peri-apatite coating decreases uncemented tibial component migration: long-term RSA results of a randomized controlled trial and limitations of short-term results. Acta Orthop 2018; 89: 425-30. Vertullo C J, Lewis P L, Lorimer M, Graves S E. The effect of long-term survivorship of surgeon preference for posterior-stabilized or minimally stabilized total knee replacement. J Bone Joint Surg Am 2017; 99: 1129-39. Watters T S, Zhen Y, Martin J R, Levy D L, Jennings J M, Dennis D A. Total knee arthroplasty after anterior cruciate ligament reconstruction: not just a routine primary arthroplasty. J Bone Joint Surg Am 2017; 99: 185-9. Wilson D A J, Richardson G, Hennigar A W, Dunbar M J. Continued stabilization of trabecular metal tibial monoblock total knee arthroplasty components at 5 years: measured with radiostereometric analysis. Acta Orthop 2012; 83: 36-40. Zandee van Rilland E, Varcadipane J, Geling O, Murai Kuba M, Nakasone C. A minimum 2-year follow-up using modular trabecular metal tibial components in total knee arthroplasty. Reconstructive Rev 2015; 5(3): 23-8. Zardiackas L D, Parsell D E, Dillon L D, Mitchell D W, Nunnery L A, Poggie R. Structure, metallurgy, and mechanical properties of a porous tantalum foam. J Biomed Mater Res 2001; 58: 180-7. Zhang Y, Ahn P B, Fitzpatrick D C, Heiner A D, Poggie R A, Brown T D.Interfacial frictional behaviour: cancellous bone, cortical bone, and a novel porous tantalum biomaterial. J Musculoskel Res 1999; 3:245-251.

Acta Orthopaedica 2019; 90 (5): 467–472

467

Primary constrained and hinged total knee arthroplasty: 2- and 5-year revision risk compared with unconstrained total knee arthroplasty: a report on 401 cases from the Norwegian Arthroplasty Register 1994–2017 Mona BADAWY 1, Anne Marie FENSTAD 2, and Ove FURNES 2,3,4 1 Coastal

Hospital in Hagavik, Department of Orthopaedic Surgery, Haukeland University Hospital; 2 Norwegian Arthroplasty Register, Department of Orthopaedic Surgery, Haukeland University Hospital; 3 Department of Orthopaedic Surgery, Haukeland University Hospital; 4 Department of Clinical Medicine, Faculty of Medicine, University of Bergen, 5021 Bergen, Norway Correspondence: mona.badawy@helse-bergen.no Submitted 2019-01-14. Accepted 2019-05-14.

Background and purpose — The number of primary, highly constrained knee arthroplasty implants has increased with a theoretically increased risk of early failure. Therefore we analyzed the risk of all revision following total knee arthroplasty (TKA) in patients receiving a hinged or condylar constrained knee (CCK) compared with a conventional unconstrained TKA. Patients and methods — The analyses included 401 primary highly constrained or hinged implants from 1994 to 2017. Kaplan–Meier survival curves were used to evaluate time to first revision with a maximum follow-up of 20 years. Cox regression was used to calculate hazard ratio (HR) comparing condylar constrained knee (CCK), hinged, and unconstrained TKA. Results — Kaplan–Meier estimated prosthesis survival after 2 years was 94.8% (95% CI 91.4–98.2) and 93.5% after 5 years for the primary CCK and 91.0% (CI 86.6–95.4) after 2 years and 85.5% after 5 years for the primary hinged TKA. Adjusted for sex, age groups, diagnosis, time period, previous surgery, and surgery time HR was 1.4 (CI 0.8–2.3) for the CCK and 2.4 (CI 1.6–3.7) for the hinged implants. The most common cause of revision in hinged implants was infection: 14 of 22 revisions. When excluding infection as revision cause, there were no differences in survival between the implant types. Estimated survival excluding infection revisions at 5 years was 96% for unconstrained, CCK, and hinged primary TKA implants. Interpretation — Primary rotating hinge total knee arthroplasty had a higher risk of revision compared with conventional TKA after 2 and 5 years’ follow-up. Infection was the most common cause of revision. When excluding infection revisions from the survival analysis, hinged and CCK implants had similar performance to unconstrained TKA.

Selective use of varus and valgus constrained or rotatinghinge implants in primary total knee arthroplasty is necessary in complex cases with severe ligament laxity, bone loss or deformity. The current awareness of planning for the right type of implant constraint has increased its use beyond salvage revision indication (National Joint Registry [NJR] 2017, Norwegian Arthroplasty Register [NAR] 2018). Instability is a major cause of revision in conventional total knee arthroplasty (Parratte and Pagnano 2008, Dyrhovden et al. 2017), particularly in younger patients (Victor 2017). Secondary osteoarthritis due to previous trauma and surgery is a major cause for knee replacement in the younger population. In these cases, constrained implants might be necessary to achieve a stable knee. The need for aggressive ligament releases in patients with major deformities or contractures may also require implants such as constrained and rotatinghinge designs (Westrich et al. 2000, Yang et al. 2012, Ghosh et al. 2016). Total knee arthroplasty surgery is predicted to increase worldwide and concurrently there is an increase in the use of primary constrained implants because of the awareness of the importance of obtaining primary stability. There are studies from single institutions (Petrou et al. 2004, Gehrke et al. 2014, Farid et al. 2015, Cottino et al. 2017), but the rates of failure leading to revision surgery have previously only been evaluated in 1 registry study to our knowledge (Baker et al. 2014). Due to the limited use of these implants and the heterogeneity of the studies, it is difficult to obtain certain conclusions regarding the long-term results. We analyzed the survival and revision causes in a large cohort of these complex primary total knee implants to provide objective evidence.

468

Acta Orthopaedica 2019; 90 (5): 467–472

Yearly number

Primary total knee arthroplasty in the Norwegian Arthroplasty Register 1994–2017 n = 71,969

45 40

CCK Hinged

Primary hinged knee arthroplasty

Primary condylar constrained knee arthroplasty

Unconstrained total knee arthroplasty

Eligible: 237 Excluded: 40 Included: 197

Eligible: 206 Excluded: 2 Included: 204

Eligible: 71,526 Excluded: 11 Included: 71,515

30 25 20 15

Figure 1. 401 cases of constrained or hinged implants were included from the Norwegian Arthroplasty Register from 1994 to 2017. 53 cases were excluded due to oncological indication for surgery. digits are number of implants.

10 5 0

Patients and method Data collection The Norwegian Arthroplasty Register has collected data on knee arthroplasty surgery since 1994 (Furnes et al. 2002). Type of procedure such as primary or revision, unconstrained or constrained/hinged, the use of stems and augments in addition to implant brand and indication for surgery is recorded. Using this information, condylar constrained knee implants and hinged knee replacements used for primary total knee arthroplasty submitted to the NAR from January 1994 until December 31, 2017 were identified. We used the stabilization of the polyethylene as definition of constraint. The implant library of the Norwegian Arthroplasty Register has listed the polyethylene inserts as minimally stabilized (CR, dished), posterior stabilized (S), rotating platform, constrained condylar (CCK), and hinged (only rotating hinged had been used) verified by catalogue numbers of the implants. Minimally stabilized and posterior stabilized rotating platform and fixed bearing implants were defined as unconstrained TKA. Patient and surgery information were available for analyses in 401 cases of constrained or hinged implants (Figure 1). Revision of the TKA was defined in the registry as exchange, complete or partial removal, or addition of implant component(s), and information on the indications for revision was obtained as well. The status of each knee replacement was assessed as revised, unrevised, or death of the patient. Information on death was retrieved from the National Population Register. Revision was linked to the primary procedure using the unique national identification number of the patient. Implant survival at 10 years was determined using revision for any reason as primary endpoint. Secondary endpoint was revisions excluding infections, analyzing aseptic reasons for revision separately. Since the percentage of hinged and CCK implants increased from 2005 and the majority were used in this time period, we did a sensitivity analysis including only ASA 1 and ASA 2 patients from 2005 to 2017 using the same adjustments (minus ASA classification) in the Cox model as mentioned in the statistical analyses.

1994 1997 2000 2003 2006 2009 2012 2015

Figure 2. Increasing usage of primary CCK and hinged TKA in the from 1994 to 2017 in the Norwegian Arthroplasty Register.

Statistics The Kaplan–Meier method was used for estimation of survival probabilities for the implants, with 95% confidence intervals (CI) with a maximum follow-up after 10 years. Cox regression analysis was used to calculate hazard ratios (HR) to estimate the survival rates adjusted for sex, diagnosis, age groups, time period, surgery time, previous surgery, perioperative complications, and ASA classification. These are presented with CIs relative to the conventional unconstrained TKA. Proportional hazard assumptions of the Cox regression model were assessed by tests and inspection of Schoenfeld residuals (Ranstam and Robertsson 2010). P-values less than 0.05 were considered statistically significant and all tests were 2-sided. Ethics, funding, and potential conflicts of interest The Norwegian Arthroplasty Register has permission from the Norwegian Data Inspectorate to collect patient data based on written consent from the patient (ref 24.1.2017: 16/01622-3/ CDG). The authors received no specific funding for this work. No conflicts of interest were declared.

Results From 1994 to 2017 primary hinged TKA was used in 197 primary cases and 22 cases (11%) were revised. Figure 2 demonstrates the increased use of primary hinged TKA and CCK since 1994. From 2005 there was an increased use until today, with 32 hinged and 42 CCK primary TKA in 2017. Regarding patient and procedure characteristics there were fewer male patients receiving constrained implants than conventional TKA (28% vs. 36%) (Table 1). Hinged TKA and CCK were more commonly used in young patients (< 50 years) and in the oldest patients (> 80 years) (Table 1). The Kaplan–Meier 2- and 5-year survival free of all revisions for primary hinged TKA was 91.0%. The CCK 2- and

Acta Orthopaedica 2019; 90 (5): 467–472

469

compared with conventional TKA, which had a 10-year survival of 93.7 %. The adjusted hazard ratio (HR) for primary Hinged CCK Unconstrained a p-value b hinged TKA as compared with conventional TKA was 2.4. HR for primary CCK versus conventional Primary procedures 197 204 71,515 TKA was 1.4 (Table 2). Revisions 22 (11) 14 (7) 3,565 (5) < 0.001 Male sex 55 (28) 58 (28) 25,423 (36) 0.008 Primary osteoarthritis was the diagnosis leadAge, years mean/median 67/70 67/68 69/70 < 0.001 ing to surgery for 83% of patients in unconstrained range 22–90 25–95 16–101 TKA compared with only 33% receiving hinged Age groups (%) < 0.001 < 50 14 11 4 TKA. Post-fracture osteoarthritis and post-ligament 50–60 15 17 14 injury osteoarthritis were dominating causes for 60–70 23 28 32 surgery in the hinged (33%) and CCK (30%) TKA 70–80 32 27 38 > 80 16 17 12 groups. Post-infection osteoarthritis was the preopDiagnosis c < 0.001 erative diagnosis in 5% of hinged TKA cases versus Primary OA 64 (33) 109 (53) 59,023 (83) 0.3% in conventional TKA (Table 1). The anterior Inflammatory arthritis 15 (8) 15 (7) 4,347 (6) Post-fracture arthritis 28 (14) 22 (11) 1,926 (3) cruciate ligament was reported to be deficient prePost ligament injury 37 (19) 38 (19) 5,161 (7) operatively in 56% of hinged TKA cases, 47% in Post infection 9 (5) 5 (2.5) 232 (0.3) CCK cases, and 19% in unconstrained TKA. The Instability 13 (6) 1 (0.5) 49 (0.1) Neuro orthop. sequelae 10 (5) 2 (1) 25 (0.1) posterior cruciate ligament was deficient preoperaOther 20 (10) 12 (6) 627 (0.5) tively in 32% of hinged cases, 24% in CCK cases, Time period < 0.001 and 2% in unconstrained TKA cases (NAR 2018). 1994–2007 15 (8) 21 (10) 25,942 (36) 2008–2017 182 (92) 183 (90) 45,573 (64) Patients with no previous surgery to the knee were Surgery time c (minutes) < 0.001 62% in the hinged and CCK group versus 71% in the median (IQR) 150 (55) 145 (55) 90 (35) conventional TKA group. 2% in the conventional range 85–420 80–360 31–654 ASA classification d < 0.001 TKA group were previously surgically treated for ASA 1 16 (9) 14 (8) 7,313 (14) fracture near the joint, whereas 14% and 17% had ASA 2 94 (51) 107 (58) 35,066 (66) previous fracture treatment in the hinged and CCK ASA 3+ 74 (40) 64 (33) 10,587 (20) Previous surgery < 0.001 group respectively. ASA 3+ patients were registered None 122 (62) 127 (62) 50,788 (71) in 40% of the hinged knee patient group whereas Fracture 34 (17) 29 (14) 1,402 (2) only 20% were ASA 3+ in the conventional TKA Ligament 20 (10) 19 (9) 9,005 (13) Osteotomy 4 (2) 11 (5) 2,393 (39 group. Median surgery time in the hinged and CCK Other 17 (9) 18 (9) 7,927 (11) group was 150 and 145 minutes respectively, versus c Perioperative complications 7 (3.6) 17 (8.5) 1,331 (1.9) < 0.001 90 minutes for conventional TKA. Patellar component e 34 (17) 40 (20) 5,235 (7) < 0.001 Stems femur/tibia, n f 155/150 167/164 527/4,290 7 (4%) and 17 (9%) cases of perioperative comAugments femur/tibia, n g 16/18 24/16 785/763 plications were reported in the hinged and CCK a Unconstrained TKA were procedures with cruciate retaining or posterior stabigroup respectively, versus 1,331 (1.9%) in convenlized, mobile or fixed bearing TKA. tional TKA (Table 1). The reported types of perib 2-sided t-test for continuous variables. Chi-square test for categorical variables. operative complications are demonstrated in Table Independent sample median test = non-parametric. c Diagnosis missing n = 126, surgery time missing n = 1,813, and perioperative 3 (see Supplementary data). There was a high percomplications missing n = 1,275. centage of fractures and tendon ruptures perioperad From 2005, n = 53,335 tively as compared with unconstrained TKA. e TKA with patellar component f 18% of information regarding use of stems was missing for hinged implants, There were no differences in HR compar15% missing for CCK, and 42% missing for unconstrained TKA in the registry ing female with male patients (reference) (HR = data. 1.0 (CI 0.1–1.0). Young patient (< 50 years) had g 58% of information regarding use of augments was missing for hinged a higher risk of revision HR = 2.2 (CI 1.9–2.5), implants, 56% missing for CCK, and 43% missing for unconstrained TKA in the registry data. and older patients (> 80 years) had a lower risk of Primary TKA with tumor indication were removed from the data material (n = 40 revision HR = 0.6 (CI 0.5–0.6) using age 60–70 as hinged, n = 2 CCK, n = 11 unconstrained). reference. Post-infection osteoarthritis had a higher risk of revision, HR = 1.8 (CI 1.3–2.6) (not shown in table). 5-year survival was 94.8%, comparable to 5-year survival of In primary hinged TKA, infection was the dominant reason unconstrained TKA, which was 95.3%. Hinged primary TKA for revision. 16 cases were revised due to deep infection. 5 had a shorter follow-up with the last revision at 6 years (Table revisions of CCK were due to deep infection (Tables 4 and 5, 2). The 10-year survival for primary CCK declined to 78.9% see Supplementary data). Table 1. Demographic data (n = 71,916) by TKA implant type from 1994 to 2017. Values are frequency (%) unless otherwise specified

470

Acta Orthopaedica 2019; 90 (5): 467–472

Table 2a. Kaplan–Meier survival a free of all cause revision at 2, 5, and 10 years postoperatively

No. of No. of patients revisions (%)

No. of deaths (%)

Unconstrained CCK Hinged

71,515 204 197

15,669 (22) 31 (15) 19 (10)

3,565 (5) 14 (7) 22 (11)

No. at K–M 2-year risk survival (CI) 57,585 128 115

No. of K–M 5-year risk survival (CI)

97 (97–98) 95 (91–98) 91 (87–95)

39,936 57 58

95 (95–96) 94 (89–98) 86 (79–92)

No. at K–M 10-year risk survival (CI) 17,379 8 8

94 (94–94) 79 (62–96) 84 (77–91)

Kaplan–Meier survival (%) with 95% CI in parentheses for all-cause reoperation for the entire follow-up period.

Table 2b. Unadjusted and adjusted hazard ratio a Unconstrained CCK Hinged

Unadjusted HR

Adjusted HR

1 (ref.) 2.0 (1.2–3.3) 3.3 (2.2–4.9)

1 (ref.) 1.4 (0.8–2.3) 2.4 (1.6–3.7)

Hazard ratio is shown unadjusted and adjusted for sex, age groups, diagnosis, time period, surgery time, previous surgery to the knee, perioperative complications, and ASA classification (registered since 2005 in the register).

Arthroplasty survival (%)

100

25 Unconstrained Condylar constrained Hinged

Unconstrained Condylar constrained Hinged

Number at risk (Figure 3)

Years from surgery

0 0

Years from surgery

Figure 3. Kaplan–Meier survival curve with revision for any reason (a) and excluding infection revision (b) from 1994 to 2017 for TKA by implant type; blue = primary hinged knee replacement, red = primary condylar constrained knee replacement, black = unconstrained total knee replacement.

Year 0 1 2 3 4 5 6 7 8 9 10 Hinged 197 154 116 86 67 59 47 30 19 15 9 CCK 204 153 128 108 82 57 37 27 19 10 8 Unconstrained 71,515 64,503 57,585 51,179 45,282 39,937 34,792 30,132 25,584 21,166 17,380

Figure 3a shows the inferior implant survival of primary hinged TKA as compared with unconstrained TKA with revision for any reason as endpoint. Figure 3b demonstrates all revisions excluding infection revisions, showing the similarity in survival comparing the 3 implant types when the most common cause of revision was removed from the data. Kaplan–Meier estimated survival for aseptic revisions at 2 years was 98.1% (CI 97.9–98.3) for unconstrained TKA, 97.6% (CI 95.2–100) for CCK, and 96.7% (CI 93.7–99.7) for hinged TKA. K–M 5-year survival for aseptic revisions was 96% for all 3 groups for TKA. In the sensitivity analysis including the latest time period from 2005 to 2017 including only ASA 1 and ASA 2 patients, we found similar results for all revisions as in the complete analysis. HR for CCK as compared with unconstrained TKA as reference was 1.5 (CI 0.7–3.1), whereas for hinged implants HR was 2.3 (CI 1.2–4.2).

Discussion The principal findings in this study were higher reoperation and revision rates in patients undergoing complex primary total knee replacements as compared with conventional total knee arthroplasties. Most commonly, revisions were caused by infection. The adjusted relative risk of all-cause revision at 10 years was > 2 times higher in patients receiving a hinged implant compared with patients receiving an unconstrained conventional implant at the time of the index surgical procedure. The condylar constrained knee implants had good survivorship, comparable to unconstrained TKA. The mid-term survival of primary hinges in this cohort at 5 years was statistically significantly inferior to unconstrained implants. However, a separate analysis of aseptic revisions showed similar survival after 5 years for all implant types.

Acta Orthopaedica 2019; 90 (5): 467â&#x20AC;&#x201C;472

The assumption that revision implants used in the primary setting would perform as well as standard implants was not found in this study, probably due to patient selection and characteristics. The differences in results comparing unadjusted and adjusted analyses indicates that the higher risk of revision is attributable to factors other than implant design. Patients undergoing primary hinged and constrained implants typically had secondary osteoarthritis caused by previous fractures, ligament injuries, and infections; only one-third had primary osteoarthritis as indication for surgery. There were more young patients (< 50 years) and more old patients (> 80 years) as well as more ASA 3+ patients in the hinged and constrained cohort compared with patients receiving routine primary knee arthroplasty. Surgery time was also prolonged in these more complex knee implants, which can be explained by both the complexity of the implant itself and the abnormality of the preoperative deformity and instability. The increased risk of infection could be explained by the high amount of ASA 3+ patients, combined with prolonged surgery time with more soft tissue exposure and trauma (Badawy et al. 2017). We found that post-infection osteoarthritis had a higher risk of revision. This is supported by the proceedings of international consensus on orthopedic infections (Aalirezaie et al. 2019). A registry study from Finland also found increased risk of infection in patients with hinged or constrained implants (JĂ¤msen et al. 2009). The results of our study should, however, not discourage surgeons from using constrained implants when considered necessary to achieve a stable knee. The high risk of infection should lead to prevention measures both pre- and perioperatively, optimizing the patient preoperatively and performing atraumatic surgery to avoid hematoma formation due to poor soft tissue handling and avoiding perioperative complications such as fractures and ligament injuries. We found a higher risk of perioperative complications in the hinged and constrained knee implant groups. Our study supports the findings of previous series where the most common reason for revision in complex primary knee arthroplasty was infection (Petrou 2004, Yang et al. 2012, Baker et al. 2014, Cholewinski et al. 2015, Martin et al. 2016, Cottino 2017, Siqueira et al. 2017). A study from the NJR (Baker et al. 2014) found that the implant survival for hinged knee replacements was comparable to conventional knee replacements, in contrast to our study, whereas others found the risk of revision to be higher in constrained primary implants (Moussa et al. 2017). The NJR study reported a high number of older patients (mean age 72) with primary osteoarthritis (70%) with hinged TKA, contrary to our study with only 33% reported primary OA. Some studies report low rates of aseptic loosening in rotating hinge TKA, but fail to offer conclusions regarding cases of infection revisions (Westrich et al. 2000, Cottino et al. 2017). Comparable to our study, Cottino et al. (2017 )and Rai et al. (2018) found a high number of patients with intraoperative complications such as fractures and ligament injuries.

471

First-generation rotating hinge total knee designs were associated with a high failure rate (Rand et al. 1987, Barrack 2001). More recent designs have improved the rotating hinge mechanism, the patellofemoral tracking, metaphyseal sleeves, and cones, and improved articulation between the mobile-bearing element and the tibial component (Barrack 2001, Deehan et al. 2008, Smith et al. 2013). There seems to be a high satisfaction rate among primary constrained and hinged knee arthroplasty patients. PROMs data in the NJR study by Baker et al. 2014 demonstrated improvements in function and general health outcomes following surgery. CCK has potential disadvantages due to increased constraint and is thought to have a higher risk of aseptic loosening due to the tight fit of the insert post in the femoral component. Rai et al. (2018) found similar implant survival (95%) to our study and with a high complication rate for CCK. CCK used by Sabatini et al. (2017) showed good functional results in their series of 28 patients. Cholewinski et al. 2015 also demonstrated a high risk of infection revisions in CCK and, similar to our study, revision for reasons other than infection was close to values of unconstrained implants. They discussed the possible decreasing level of constraint over time due to polyethylene creep at the tibial post, thus mechanical long-term complications such as loosening did not occur. There are a number of limitations to interpretation of our data. The number of hinged and constrained implants accounted for a small number of the total amount of primary knee arthroplasty in Norway. Thus, the generalizability of the results is limited. We did not have PROMs data, so revision for any reason was the endpoint for our results. Even in the NJR study (Baker et al. 2014), where PROMs exists as a source of information in the national register, the information was available only for a small proportion of patients. There is also a heterogeneity in the cohort, since many patients had previous and varying surgical interventions. Even if we had made adjustments for this in the analysis, there could be a selection bias. The infection revisions are mostly soft tissue debridement with exchange of polyethylene insert, indicating early infection and revision. However, we have no information on whether there actually was bacterial growth in specimens after surgery, verifying the diagnosis. This should not, however, influence the comparison between implant types (Gundtoft et al. 2015). A higher threshold to revise such complex implants could lead to falsely high survival rates due to the complexity of revision. However, regarding infection revisions, there could be a lower threshold to do soft tissue revisions due to wound drainage at an earlier stage than with unconstrained TKA. In summary, there has been an increased use of primary hinged and condylar constrained total knee arthroplasty in the last decade. These implants should always be considered in complex cases to achieve a stable knee. Rates of septic failure are higher than for conventional total knee arthroplasty, probably caused by a higher number of patients with comorbidi-

472

ties, and also due to mechanical and soft tissue challenges in addition to the size and complexity of the implant, increasing the surgical duration and risk of haematoma formation in these complex cases. When excluding infection revisions from the survival curve, hinged and CCK implants showed similar performance to unconstrained TKA. The patients’ general risk of infection should be optimized, as should the surgeons’ skills regarding correct indication and use of these more complicated implants to ensure results comparable to conventional TKA. Supplementary data Tables 3–5 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674. 2019.1627638

The study was conceived by MB, AMF, and ONF, MB wrote the initial draft, AMF performed the analyses, all authors contributed to the interpretation of the data and to revision of the manuscript. Acta thanks Esa Jämsen and Maziar Mohaddes for help with peer review of this study.

Aalirezaie A, Arumugam S S, Austin M, Bozinovski Z, Cichos K H, Fillingham Y, Ghanem E, Greenky M, Huang W, Jenny J-Y, Lee G-C, Manrique J, Manzary M, Oshkukov S, Patel N K, Reyes F, Spangehl M, Vahedi H, Voloshin V. Hip and knee section, prevention, risk mitigation: Proceedings of International Consensus on Orthopedic Infections. J Arthroplasty 2019; 34(2S): S271-8. Badawy M, Espehaug B, Fenstad A M, Indrekvam K, Dale H, Havelin L I, Furnes O. Patient and surgical factors affecting procedure duration and revision risk due to deep infection in primary total knee arthroplasty. BMC Musculoskeletal Disord 2017; 18(1): 544. Baker P, Critchley R, Gray A, Jameson S, Gregg P, Port A, Deehan D. Midterm survival following primary hinged total knee replacement is good irrespective of the indication for surgery. KSSTA 2014; 22(3): 599-608. Barrack R L. Evolution of the rotating hinge for complex total knee arthroplasty. Clin Orthop Rel Res 2001; (392): 292-9. Cholewinski P, Putman S, Vasseur L, Migaud H, Duhamel A, Behal H, Pasquier G. Long-term outcomes of primary constrained condylar knee arthroplasty. Orthop Traumatol Surg Res 2015; 101(4): 449-54. Cottino U, Abdel M P, Perry K I, Mara K C, Lewallen D G, Hanssen A D. Long-term results after total knee arthroplasty with contemporary rotatinghinge prostheses. J Bone Joint Surg Am 2017; 99(4): 324-30. Deehan D J, Murray J, Birdsall P D, Holland J P, Pinder I M. The role of the rotating hinge prosthesis in the salvage arthroplasty setting. J Arthroplasty 2008; 23(5): 683-8. Dyrhovden G S, Lygre S H L, Badawy M, Gøthesen Ø, Furnes O. Have the causes of revision for total and unicompartmental knee arthroplasties changed during the past two decades? Clin Orthop Rel Res 2017; 475(7): 1874-86. Farid Y R, Thakral R, Finn H A. Intermediate-term results of 142 singledesign, rotating-hinge implants: frequent complications may not preclude salvage of severely affected knees. J Arthroplasty 2015; 30(12): 2173-80.

Acta Orthopaedica 2019; 90 (5): 467–472

Furnes O, Espehaug B, Lie S A, Vollset S E, Engesaeter L B, Havelin L I. Early failures among 7,174 primary total knee replacements: a follow-up study from the Norwegian Arthroplasty Register 1994–2000. Acta Orthop Scand 2002; 73(2): 117-29. Gehrke T, Kendoff D, Haasper C. The role of hinges in primary total knee replacement. Bone Joint J 2014; 96-b(11 Suppl. A): 93-5. Ghosh K M, Manning W A, Blain A P, Rushton S P, Longstaff L M, Amis A A, Deehan D J. Influence of increasing construct constraint in the presence of posterolateral deficiency at knee replacement: a biomechanical study. J Orthop Res 2016; 34(3): 427-34. Gundtoft P H, Overgaard S, Schonheyder H C, Moller J K, KjaersgaardAndersen P, Pedersen A B. The “true” incidence of surgically treated deep prosthetic joint infection after 32,896 primary total hip arthroplasties: a prospective cohort study. Acta Orthop 2015; 86(3): 326-34. Jämsen E, Huhtala H, Puolakka T, Moilanen T. Risk factors for infection after knee arthroplasty: a register-based analysis of 43,149 cases. J Bone Joint Surg Am 2009; 91(1): 38-47. Martin J R, Beahrs T R, Stuhlman C R, Trousdale R T. Complex primary total knee arthroplasty: long-term outcomes. J Bone Joint Surg Am 2016; 98(17): 1459-70. Moussa M E, Lee Y Y, Patel A R, Westrich G H. Clinical outcomes following the use of constrained condylar knees in primary total knee arthroplasty. J Arthroplasty 2017; 32(6): 1869-73. National Joint Registry (NJR). 14th Annual Report 2017,. Available at http:// www.njrcentre.org.uk.pdf. Norwegian Arthroplasty Register. Report 2018. Available at http:// nrlweb. ihelse.net.pdf. Parratte S, Pagnano M W. Instability after total knee arthroplasty. J Bone Joint Surg 2008; 90(1): 184-94. Petrou G, Petrou H, Tilkeridis C, Stavrakis T, Kapetsis T, Kremmidas N, Gavras M. Medium-term results with a primary cemented rotating-hinge total knee replacement. A 7- to 15-year follow-up. J Bone Joint Surg Br 2004; 86(6): 813-7. Rai S, Liu X, Feng X, Rai B, Tamang N, Wang J, Ye S, Yang S. Primary total knee arthroplasty using constrained condylar knee design for severe deformity and stiffness of knee secondary to post-traumatic arthritis. J Orthop Surg Res 2018; 13(1): 67. Rand J A, Chao E Y, Stauffer R N. Kinematic rotating-hinge total knee arthroplasty. J Bone Joint Surg Am 1987; 69(4): 489-97. Ranstam J, Robertsson O. Statistical analysis of arthroplasty register data. Acta Orthop 2010; 81(1): 10-4. Sabatini L, Risitano S, Rissolio L, Bonani A, Atzori F, Masse A. Condylar constrained system in primary total knee replacement: our experience and literature review. Ann Transl Med 2017; 5(6): 135. Siqueira M B P, Jacob P, McLaughlin J, Klika A K, Molloy R, Higuera C A, Barsoum W K. The varus–valgus constrained knee implant: survivorship and outcomes. J Knee Surg 2017; 30(5): 484-92. Smith T H, Gad B V, Klika A K, Styron J F, Joyce T A, Barsoum W K. Comparison of mechanical and nonmechanical failure rates associated with rotating hinged total knee arthroplasty in nontumor patients. J Arthroplasty 2013; 28(1): 62-7.e1. Victor J. Optimising position and stability in total knee arthroplasty. EFORT Open Reviews 2017; 2(5): 215-20. Westrich G H, Mollano A V, Sculco T P, Buly R L, Laskin R S, Windsor R. Rotating hinge total knee arthroplasty in severely affected knees. Clin Orthop 2000(379): 195-208. Yang J H, Yoon J R, Oh C H, Kim T S. Primary total knee arthroplasty using rotating-hinge prosthesis in severely affected knees. KSSTA 2012; 20(3): 517-23.

Acta Orthopaedica 2019; 90 (5): 473–478

473

Failure modes of patellofemoral arthroplasty—registries vs. clinical studies: a systematic review Nikolaj B BENDIXEN, Peter W ESKELUND, and Anders ODGAARD

Department of Orthopedic Surgery, Copenhagen University Hospital Herlev-Gentofte, Hellerup, Denmark Correspondence: birgerbendixen@live.dk Submitted 2019-01-29. Accepted 2019-05-15.

Background and purpose — Patellofemoral arthroplasty (PFA) has been debated since early studies showed poor implant survival. Recent studies show better results. This review reports failure modes for PFA and investigates differences in data reported from registries and clinical studies. Additionally, we report differences in failure modes among implant designs. Methods — A systematic search was performed in September 2018. All studies and registers describing failure modes of PFA were included and implant design was noted for each revision. Results — This review includes 1,299 revisions of a primary PFA reported in 47 clinical studies and 3 registers. The failure modes were: 42% OA progression, 16% pain, 13% aseptic loosening, 12% surgical error, 4% wear, 2% infection, 2% broken patellar component, 1% stiffness, 1% fracture, and 7% other. The data from registries and cohort studies differed statistically significantly in 7 out of 12 failure modes. Significant differences were found in several failure modes among implant designs. Interpretation — OA progression is the most common failure mode of PFA. There are significant differences in data on failure modes between registers and protocolled studies, notably for surgical error. The implant design significantly influences several of the failure modes. In conclusion, indication, surgical technique, and implant design are important for a successful PFA, and register-based failure modes should be interpreted with caution.

Isolated patellofemoral osteoarthritis (PF-OA) predominantly affects women, the elderly, and young adults with dysplasia or former trauma. 11% of male patients and 24% of female patients with symptomatic knees have PF-OA (McAlindon et al. 1992) and could therefore be candidates for patellofemoral arthroplasty (PFA). Controversy concerning PFA has led surgeons to choose total knee arthroplasty (TKA) in patients with PF-OA. Recent studies on PFA show good results, however, with high survival rates and satisfying patient-reported outcomes (Goh et al. 2015, Halai et al. 2016, Konan and Haddad 2016, Osarumwense et al. 2017, Odgaard et al. 2018), suggesting that PFA may be a good choice in suitably selected patients. A rise in PFA operations has been seen in Denmark from 0.6% of primary arthroplasties in 2012, to 1.3% in 2016, but the reported prevalence of PF-OA (McAlindon et al. 1992) suggests that PFA is under-used. Register studies show poor survival of PFA compared with TKA where the 2-year survival is 91% and 97% and 10-year survival is 73% and 93%, respectively (Odgaard et al. 2017). Because of this controversy, we performed a systemic review with the following purposes: (1) to investigate the reasons for revisions of PFA, (2) to highlight possible differences in data from registries and protocolled studies, and (3) describe differences in failure modes among implant designs. Van der List et al. (2017) published a similar review. In our review we present newer results, a higher number of revisions (1,299 vs. 938), and, as van der List et al. suggested, we investigate the impact of implant design on failure modes. Furthermore, we focus on the differences between data from registries and protocolled studies.

474

Methods Search strategy A systematic search was performed in September 2018. Electronic databases (PubMed, Embase, and Cochrane Library) were searched with the terms “arthroplasty AND (patellofemoral OR PF OR PFA OR PFR) AND (outcome OR functional outcome OR scores OR results OR revision OR revision rate OR reoperation OR treatment failure OR prosthesis failure OR failure OR failure rate OR survivorship OR survival).” 16 additional articles included in the protocol for the RCT study of Odgaard et al. (2018) were used for a cited reference search (Arciero and Toomey 1988, Grelsamer 1990, Argenson et al. 1995, Krajca-Radcliffe and Coker 1996, Tauro et al. 2001, De Winter et al. 2001, Smith et al. 2002, Kooijman et al. 2003, Merchant 2004, Board et al. 2004, Ackroyd and Chir 2005, Argenson et al. 2005, Blazina et al. 2005, Cartier et al. 2005, Ackroyd et al. 2007, Merchant et al. 2017). Articles were selected by relevance and had to be registered on Web of Science. NBB and PWE scanned titles and abstracts for relevance, and full texts were evaluated against the eligibility criteria. Included articles were checked for eligibility by AO. Studies found during the search were screened for any further studies containing relevant data. Final consensus on inclusion was reached between the authors. Finally, registries from Scandinavia, England/Wales, The Netherlands, New Zealand and Australia were screened for information, last accessed on December 17, 2018. The review was conducted in accordance with PRISMA guidelines and is registered on PROSPERO (CRD42018115774). Inclusion criteria All studies reporting reasons for revision of PFA were included. A revision was defined as surgery where single or multiple components of the implant were removed, replaced, or supplemented. Studies in languages other than English were included if data was extractable by the authors. Exclusion criteria Studies using the same cohort or database as in an already included study were excluded to eliminate the risk of patients being counted repeatedly. Data collection Parameters included: authors, year published, year cohort started/ended, mean years of follow-up, number of arthroplasties, number of revisions, type of implant, and reasons for revision. 26 different failure modes were identified. These were categorized into 12 groups (Table 1, see Supplementary data).

Acta Orthopaedica 2019; 90 (5): 473–478

Records identified through database searching n = 2,522

Additional records identified through other sources n = 17

Records after duplicates removed n = 1,344 Records excluded after screening n = 1,217 Full-text articles assessed for eligibility n = 127 Full-text articles excluded (n = 77): – no description of failure modes, 58 – same cohort as another study/register, 15 – no quantitative data on failure modes, 4 Studies incuded in qualitative synthesis n = 50

Flowchart.

Quality of studies All studies were assessed for quality by NBB and PWE according to Grading quality of evidence and strength of recommendations (GRADE). Statistics A pooled analysis was performed for each failure mode. Final groups of failure modes were presented in percentages. A chi-square test was performed in accordance to outcomes: (1) Failure modes in cohorts vs. registries and (2) Failure modes between each implant design vs. all others. In samples where expected frequencies were lower than five, Fisher’s exact test was used. P-values < 0.05 were considered significant. All statistical analyses were performed in Excel 2016 (Microsoft Corp., Redmond, WA, USA). Funding and potential conflicts of interest Institutional support was received from Stryker European Operations BV for studies on patellofemoral implants. Stryker had no influence on data accumulation, analysis or interpretation.

Results (Tables 2–8, see Supplementary data) Search results The primary search identified 2,522 studies and 17 additional studies were located by the cited reference search and by screening of references. 1,344 unique articles were found, and 1,217 articles were excluded during the screening of titles and abstracts. 127 articles were assessed for eligibility, and 50 articles (3 registers and 47 clinical studies) were included (Figure). Only the New Zealand, Australian, and English registry had relevant data according to the inclusion criteria (Table 2)

Acta Orthopaedica 2019; 90 (5): 473–478

(Muir et al. 1999, Baker et al. 2012, Australian Orthopaedic Association 2017) Quality of studies The search revealed 1 level I RCT study, 31 level II prospective cohort studies, 14 level III–IV observational studies, and 1 level VI case study (Table 3). The studies were published between 1996 and 2018. Heterogeneity existed in the followup periods and types of implants used. The quality of evidence was, according to GRADE guidelines, low.

Failure modes of the patellofemoral implant 47 clinical studies and 3 registers containing 1,299 revisions of PFAs were found. The most common failure modes were OA progression (42%), pain (16%), aseptic loosening (13%), and surgical error (12%). Failure modes varied among registries (Table 2). Registries vs. clinical studies Statistically significant differences were found between registers and clinical studies in 7 out of 12 failure modes (Table 4). Pain, aseptic loosening, operational flaws, and broken patellar component were all found to be highly significant (p < 0.001). Additionally, significant differences were observed with progression of OA and infection (p = 0.009 and p = 0.04, respectively). Pain and aseptic loosening were more often reported as indication for revision in registers (20% vs. 10% and 17% vs. 7%, respectively). In clinical studies, progression of OA and surgical errors were more dominant than in registers (48% vs. 38% and 21% vs. 7%, respectively) (Table 4). Failure modes and implant design 40 studies reported data on specific implants (Table 5). 7 studies did not relate cause for revision and implant design and were excluded in this outcome. 64 revisions were excluded from clinical studies, and all 807 revisions from registers were excluded in the same manner. 428 revisions were included (Table 5) and chi-square and Fisher’s tests were performed on these. 5 implant designs (Avon, Richards, Lubinus, Autocentric, and LCS) had more than 30 documented revisions, and these were chosen for further analysis (Tables 5 and 6). The Avon implant was revised more often (p = 0.02) for OA progression with relative risk (RR) = 1.4, and less for surgical errors (p = 0.001) with RR = 0.2. The Richards design was, with RR = 0.8, revised less for pain (p = 0.007). The Lubinus design was more often revised because of surgical errors (p < 0.001) with RR = 3.9 and wear (p = 0.001), RR = 4.3, and less often for OA progression (p = 0.005) RR = 0.5, pain (p = 0.03) RR = 0.2, and none for aseptic loosening (p = 0.008). The Autocentric implant had a higher incidence of aseptic loosening (p < 0.001), RR = 4.9, infection (p = 0.007), RR = 14 and was the only design with stiffness as a failure mode (p < 0.001), while 0 was revised because of pain (p = 0.04). The

475

LCS implant had a greater revision rate for pain (p < 0.001) RR = 5.5 and surgical errors (p = 0.002), while progression of OA (p < 0.001) was seen less frequently with RR = 0.2.

Discussion Progression of osteoarthritis as the main failure mode This review found the most frequently reported reason for revision of PFA to be OA progression with a proportion of 42% (48% in studies and 38% in registries). Because some revised knees in this group may have been reported with the nonspecific failure mode of pain, this could even be an underestimation. Progression of OA can be viewed in 2 ways: (1) as a victory for the implant in long-term follow-up studies since it survived implant-dependent failure modes, and (2) as poor patient selection in studies with short-term follow-up. When deciding between PFA and TKA, strict patient selection is suggested (Mohammed et al. 2008, Goh et al. 2015, Odgaard et al. 2017). Patients having a PFA for primary OA may have a higher incidence of progression of TF-OA compared with patients having surgery for dysplasia or previous fracture (Smith et al. 2002, Argenson et al. 2005). Progression of TF-OA has been noted mainly in patients with malalignment compared with a neutral hip–knee–ankle angle (Cartier et al. 2005). Studies including patients with TF-OA (Leadbetter et al. 2009, Williams et al. 2013, Dahm et al. 2014) reported a higher incidence (75%) of patients revised because of OA progression. These findings support the importance of strict patient selection. To identify patients with isolated PF-OA, Cartier et al. (2005) claimed that medial knee pain ascending stairs always indicates PF-OA, and pain descending stairs is of TF-OA origin. Bone-on-bone on tangential (skyline) radiographs is the deciding radiographic proof of PF-OA (Odgaard et al. 2017, Cuthbert et al. 2018). We encourage thorough clinical and imaging investigations preoperatively, and stress radiography, specialized radiographic projections, and MRI should be considered to ensure the absence of TF-OA. Odgaard et al. (2018) published early results of a doubleblinded RCT comparing PFA and TKA for isolated PF-OA. They concluded that patients treated with PFA obtained better knee function, better satisfaction, and larger knee-related quality of life compared with TKA within the first 2 years. The study found similar short-term survival rates for the 2 implants. Cartier et al. (2005) reported no complications in getting back to daily activities and sports after PFA surgery. Additionally, studies show that revision of PFA to TKA does not present any particular difficulties (van Jonbergen et al. 2009, Parratte et al. 2015). Therefore, it could be argued that PFA should have a place in the treatment of knee OA as a temporary implant towards TKA (Argenson et al. 2005, Leadbetter et al. 2009). Using the implant as a stepping stone could give the patient some years with a higher knee-related quality of life without compromising the results of a future TKA.

476

However, a recent study on a cohort from the Australian registry shows that the risk of re-revision of a secondary TKA from a PFA is larger than the risk of a revision of a primary TKA (Parker et al. 2019), but the extent of confounding in register studies is unknown. Only longitudinal protocolled studies with time-weighted outcomes, e.g., area under the curve for PRO data, can determine the better strategy. In line with results published by van der List et al. (2017) we found that progression of OA is the key failure mode for PFA. Our data, however, show a lower rate of OA progression (42% vs. 49%) and a higher rate of pain (16% vs. 6%), which indicates that the 2 populations differ. This review consists of newer data and a larger population. Registries vs. clinical studies In 7 of 12 failure modes, we found a statistically significant difference between register and cohort studies (Table 4). The difference may be explained by the way indications for revisions are recorded. Where registries use fixed modes of failure, studies tend to record more specific indications. Register data (Muir et al. 1999, Baker et al. 2012, Australian Orthopaedic Association 2017) are almost devoid of surgical errors and have a high rate of pain as indication, suggesting that surgeons have a preference for reporting the nonspecific symptom rather than surgical errors. It may reasonably be assumed that surgeons have a higher tendency to scrutinize indications for patients included in a clinical study. The variation in failure modes among registers suggests that registration practice or indications vary. Cohort studies (e.g. van Jonbergen et al. 2010) investigate a specific implant, its survival, and its failure modes. They report a high rate of OA progression, surgical error, and wear. No patients were revised by the nonspecific indication pain, suggesting that focus on diagnostics could be different in cohort studies vs. the daily registration for registers. Van der List et al. (2017) reported OA progression as the most common failure mode with more than 5 years’ follow-up. Most clinical studies included in our review have a short follow-up, and therefore include a large proportion of early revisions. The included registries date back to 1999–2003 and have no defined end follow-up. Therefore, it would be expected that registries would contain a larger proportion of revisions with more than 5 years’ follow-up and therefore a higher proportion of OA progression. The conflict between this expectation and actual register data supports the argument that revisions are reported with the less specific indication of pain. The year of operation could influence the cause for revision. Our cohorts report data back to 1972 (Argenson et al. 2005), and the earliest registries start in 1999 (Muir et al. 1999, Australian Orthopaedic Association 2017). The high rate of surgical errors in the studies could be caused by early, less successful implant designs. An example is the Lubinus implant where Tauro et al. (2001), in a cohort from 1989–1995, reported a 71% failure by maltracking and called the implant “Unforgiving.” They even converted the trochlear component later in

Acta Orthopaedica 2019; 90 (5): 473–478

their study, using the left component on the right knee and vice versa. The review by van der List et al. (2017) reports the same incoherence between registries and studies. The only difference in data is infection, which in our review is more often reported from registries. Failure modes and implant design OA progression may be a success criterion for the implant design or suggest poor patient selection. A high percentage of revisions for OA progression in cohort studies with a long follow-up period (> 10 years) suggests a satisfactory design. This was observed in the study by Kooijman et al. (2003), where the Richards II design has a 75% failure by OA progression, and in the cohort from Nicol et al. (2006) who reported an 80% failure by this mode with the Avon design. With a 56% rate of surgical error as failure mode (Table 6), the Lubinus design has met with controversies. More authors have questioned the anatomical design of the femoral component and agree that a non-anatomical design such as Avon or Richards is preferable (Tauro et al. 2001, Board et al. 2004, Cartier et al. 2005). Tauro et al. argue that the anatomical and asymmetrical design of the femoral groove in the component impairs the surgeon’s ability to correct an existing anatomical dysplasia with maltracking. The results of LCS and Autocentric (Table 6) may indicate constructional flaws. The mobile bearing LCS design has a metal-backed patellar component (Charalambous et al. 2011). Yadav et al. (2012) argue that rotational freedom of the polyethylene could lead to locking in the trochlear groove and may cause mechanical problems and wear. Charalambous et al. (2011) found overgrowth of soft tissue in several of their revisions, resulting in a blocking of the rotation of the polyethylene, which may have caused the high frequency of pain in their study. Furthermore, extensive metallosis was found in 3 of 17 revised knees, likely caused by the metal-backed patella. Because of non-satisfactory results, Australia stopped using the LCS implant in 2009 (Australian Orthopaedic Association 2017). In studies concerning the Autocentric implant, aseptic loosening was the dominant failure mode (Table 6) (De Cloedt et al. 1999, Argenson et al. 2005), which may give the impression of design problems. Limitations and bias This study has several limitations, mainly the problem of inclusion. We strove to include as many revisions as possible, even a revision from a biased case study reporting only a single failure mode. On these grounds we cannot exclude confounders and bias in general. To ensure that data of a cohort are only used once, we had to exclude some studies knowing that we lost data. It was not possible to detect duplicates between registries and studies, so this is also a bias in our study. This review’s first outcome has similarity to the article published by van der List et al. (2017), but some inconsistencies

Acta Orthopaedica 2019; 90 (5): 473–478

were found with inclusion of patients and extraction of data in their paper. An example is data from Kooijman et al. (2003) where data on 19 revisions were extracted, but only 10 of these fulfilled their definition of revision, which was conversion to a TKA. Conclusion Progression of TF-OA is the most common failure mode of PFA. The data from registries and cohort studies differ significantly in several modes of failure, notably for surgical error. The implant design has a significant influence on several of the failure modes. In conclusion, indication, surgical technique, and implant design are important for a successful PFA. Supplementary data Tables 1–8 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674. 2019.1634865

The article and search were made by NBB and PWE and were supervised and edited by AO. Acta thanks Kirill Gromov and Frank Madsen for help with peer review of this study

Ackroyd C E, Chir B. Development and early results of a new patellofemoral arthroplasty. Clin Orthop Relat Res 2005; (436): 7–13. Ackroyd C E, Newman J H, Evans R, Eldridge J D J, Joslin C C. The Avon patellofemoral arthroplasty: five-year survivorship and functional results. J Bone Joint Surg Br 2007; 89-B(3): 310-15. Ahearn N, Metcalfe A J, Hassaballa M A, Porteous A J, Robinson J R, Murray J R, Newman J H. The Journey patellofemoral joint arthroplasty: a minimum 5 year follow-up study. Knee 2016; 23(5): 900-4. Akhbari P, Malak T, Dawson-Bowling S, East D, Miles K, Butler-Manuel P A. The Avon patellofemoral joint replacement: mid-term prospective results from an independent centre. Clin Orthop Surg 2015; 7(2): 171-6.

477

Blazina M E, Fox J M, Del Pizzo W, Broukhim B, Ivey F M. Patellofemoral replacement 1979. Clin Orthop Relat Res 2005; (436): 3-6. Board T N, Mahmood A, Ryan W G, Banks A J. The Lubinus patellofemoral arthroplasty: a series of 17 cases. Arch Orthop Trauma Surg 2004; 124(5): 285-7. Butler J E, Shannon R. Patellofemoral arthroplasty with a custom-fit femoral prosthesis. Orthopedics 2009; 32(2): 81. Cartier P, Sanouiller J L, Khefacha A. Long-term results with the first patellofemoral prosthesis. Clin Orthop Relat Res 2005; (436): 47-54. Charalambous C P, Abiddin Z, Mills S P, Rogers S, Sutton P, Parkinson R. The low contact stress patellofemoral replacement: high early failure rate. Bone Joint J 2011; 93-B(4): 484-9. Christ A B, Baral E, Koch C, Shubin Stein B E, Gonzalez Della Valle A, Strickland S M. Patellofemoral arthroplasty conversion to total knee arthroplasty: retrieval analysis and clinical correlation. Knee 2017; 24(5): 1233-9. Cuthbert R, Tibrewal S, Tibrewal S B. Patellofemoral arthroplasty: current concepts. J Clin Orthop Trauma 2018; 9(1): 24-8. Dahm D L, Kalisvaart M M, Stuart M J, Slettedahl S W. Patellofemoral arthroplasty: outcomes and factors associated with early progression of tibiofemoral arthritis. Knee Surg Sport Traumatol Arthrosc 2014; 22(10): 2554-9. Davies A P. High early revision rate with the FPV patello-femoral unicompartmental arthroplasty. Knee 2013; 20(6): 482-4. De Cloedt P, Legaye J, Lokietek W. [Femoro-patellar prosthesis. A retrospective study of 45 consecutive cases with a follow-up of 3-12 years]. Acta Orthop Belg 1999; 65(2): 170-5. De Winter W E A E, Feith R, Van Loon C J M. The Richards type II patellofemoral arthroplasty: 26 cases followed for 1-20 years. Acta Orthop Scand 2001; 72(5): 487-90. Gadeyne S, Besse J-L, Galand-Desme S, Lerat J-L, Moyen B. [Results of selfcentering patellofemoral prosthesis: a retrospective study of 57 implants]. Rev Chir Orthop Reparatrice Appar Mot 2008; 94(3): 228-40. Goh G S H, Liow M H L, Tay D K J, Lo N N, Yeo S J. Four-Year follow up outcome study of patellofemoral arthroplasty at a single institution. J Arthroplasty 2015; 30(6): 959-63. Grelsamer R. Patellofemoral arthroplasty: 2–12-year follow-up study. J Arthroplasty 1990; 5(1): 49-55. Halai M, Ker A, Anthony I, Holt G, Jones B, Blyth M. Femoro Patella Vialla patellofemoral arthroplasty: an independent assessment of outcomes at minimum 2-year follow-up. World J Orthop 2016; 7(8): 487. Hendrix M R G, Ackroyd C E, Lonner J H. Revision patellofemoral arthroplasty: three- to seven-year follow-up. J Arthroplasty 2008; 23(7): 977-83. Hofmann A A, McCandless J B, Shaeffer J F, Magee T H. Patellofemoral replacement: the third compartment. Bone Joint J 2013; 95 B(11 Suppl A): 124-8.

Al-Hadithy N, Patel R, Navadgi B, Deo S, Hollinghurst D, Satish V. Midterm results of the FPV patellofemoral joint replacement. Knee 2014; 21(1): 138-41.

Hoogervorst P, de Jong R J, Hannink G, van Kampen A. A 21% conversion rate to total knee arthroplasty of a first-generation patellofemoral prosthesis at a mean follow-up of 9.7 years. Int Orthop 2015; 39(9): 1857-64.

Arciero R A, Toomey H E. Patellofemoral arthroplasty: a three- to nine-year follow-up study. Clin Orthop Relat Res 1988; (236): 60-71.

Hutt J, Dodd M, Bourke H, Bell J. Outcomes of total knee replacement after patellofemoral arthroplasty. J Knee Surg 2012; 26(4): 219-24.

Argenson J N, Guillaume J M, Aubaniac J. Is there a place for patellofemoral arthroplasty? Clin Orthop Relat Res 1995; (321): 162-7.

Jørgensen P S, Konradsen L A G, Mati W B, Tørholm C. [Treatment of patellofemoral arthritis with patello-femoral arthroplasties]. Ugeskr Laeger 2007; 169(23): 2201-4.

Argenson J N A, Flecher X, Parratte S, Aubaniac J M. Patellofemoral arthroplasty: an update. Clin Orthop Relat Res 2005 (440): 50-3. Arumilli B R B, Ng A B Y, Ellis D J, Hirst P. Unusual mechanical complications of unicompartmental low contact stress mobile bearing patellofemoral arthroplasty: a cause for concern? Knee 2010; 17(5): 362-4. Australian Orthopaedic Association. Australian Orthopaedic Association National Joint Replacement Registry; Hip, Knee & Shoulder Arthroplasty: Annual Report 2017. Natl Joint Replace Regist; 2017. p. 1-380. Baker P N, Refaie R, Gregg P, Deehan D. Revision following patello-femoral arthoplasty. Knee Surg Sport Traumatol Arthrosc 2012; 2 (10): 2047-53. Benazzo F, Rossi S M P, Ghiara M. Partial knee arthroplasty: patellofemoral arthroplasty and combined unicompartmental and patellofemoral arthroplasty implants—general considerations and indications, technique and clinical experience. Knee 2014; 21(S1): S43-6.

Konan S, Haddad F S. Midterm outcome of avon patellofemoral arthroplasty for posttraumatic unicompartmental osteoarthritis. J Arthroplasty 2016; 31(12): 2657-9. Kooijman H J, Driessen A P P M, van Horn J R. Long-term results of patellofemoral arthroplasty. J Bone Joint Surg Ser B 2003; 85(6): 836-40. Krajca-Radcliffe J B, Coker T P. Patellofemoral arthroplasty: a 2- to 18-year followup study. Clin Orthop Relat Res 1996; (330): 143-51. Leadbetter W B, Kolisek F R, Levitt R L, Brooker A F, Zietz P, Marker D R, Bonutti P M, Mont M A. Patellofemoral arthroplasty: a multi-centre study with minimum 2-year follow-up. Int Orthop 2009; 33(6): 1597-601. Lonner J H, Jasko J G, Booth R E. Revision of a failed patellofemoral arthroplasty to a total knee arthroplasty. J Bone Joint Surg Ser A 2006; 88(11): 2337-42.

478

McAlindon T E, Snow S, Cooper C, Dieppe P A. Radiographic patterns of osteoarthritis of the knee joint in the community: the importance of the patellofemoral joint. Ann Rheum Dis 1992; 51(7): 844-9. Merchant A C. Early results with a total patellofemoral joint replacement arthroplasty prosthesis. J Arthroplasty 2004; 19(7): 829-36. Merchant A C, Fulkerson J P, Leadbetter W. The diagnosis and initial treatment of patellofemoral disorders. Am J Orthop 2017; 46 (2): 68-75. Mertl P, Van F T, Bonhomme P, Vives P. [Femoropatellar osteoarthritis treated by prosthesis. Retrospective study of 50 implants]. Rev Chir Orthop Reparatrice Appar Mot 1997; 83(8): 712-8. Metcalfe A J, Ahearn N, Hassaballa M A, Parsons N, Ackroyd C E, Murray J R, Robinson J R, Eldridge J D, Porteous A J. The Avon patellofemoral joint arthroplasty. Bone Joint J 2018; 100-B(9): 1162-7. Middleton S W F, Toms A D, Schranz P J, Mandalia V I. Mid-term survivorship and clinical outcomes of the Avon patellofemoral joint replacement. Knee 2018; 25(2): 323-8.

Acta Orthopaedica 2019; 90 (5): 473–478

Osarumwense D, Syed F, Nzeako O, Akilapa S, Zubair O, Waite J. Patellofemoral joint arthroplasty: early results and functional outcome of the Zimmer Gender Solutions Patello-Femoral Joint System. Clin Orthop Surg 2017; 9(3): 295. Parker D, Graves S E, Myers P, Cuthbert A, Lewis P L. what is the risk of repeat revision when patellofemoral replacement is revised to TKA? An analysis of 482 cases from a large national arthroplasty registry. Clin Orthop Relat Res 2019; 477(6): 1402-10. Parratte S, Lunebourg A, Ollivier M, Abdel M P, Argenson J N A. Are revisions of patellofemoral arthroplasties more like primary or revision TKAs. Clin Orthop Relat Res 2015; 473(1): 213-19. Philippe H, Caton J. Design, operative technique and ten-year results of the HermesTM patellofemoral arthroplasty. Int Orthop 2014; 38(2): 437-42. Sarda P, Maheswaran S, Shetty A. Medium term results of Avon patellofemoral joint replacement. Indian J Orthop 2011; 45(5): 439.

Mohammed R, Jimulia T, Durve K, Bansal M, Green M, Learmonth D. Medium-term results of patellofemoral joint arthroplasty. Acta Orthop Belg 2008; 74(4): 472-7.

Smith A M, Peckett W R C, Butler-Manuel P A, Venu K M, D’Arcy J C. Treatment of patello-femoral arthritis using the Lubinus patello-femoral arthroplasty: a retrospective review. Knee 2002; 9(1): 27-30.

Mont M A, Johnson A J, Naziri Q, Kolisek F R, Leadbetter W B. Patellofemoral arthroplasty: 7-year mean follow-up. J Arthroplasty 2012; 27(3): 358-61.

Starks I, Roberts S, White S H. The Avon patellofemoral joint replacement: independent assessment of early functional outcomes. J Bone Joint Surg Br 2009; 91-B(12): 1579-82.

Morris M J, Lombardi A V, Berend K R, Hurst J M, Adams J B. Clinical results of patellofemoral arthroplasty. J Arthroplasty 2013; 28(9 Suppl.): 199-201.

Tauro B, Ackroyd C E, Newman J H, Shah N A. The Lubinus patellofemoral arthroplasty: a five- to ten-year prospective study. J Bone Joint Surg Br 2001; 83(5): 696-701.

Muir D, Surgeon O, Oakley A, Griffin H, Pettett A, Hobbs T, Coordinator R, Frampton C, Statistician R, Rothwell A, Supervisor R, James S, Hip T, Khalid A, Shoulder M. The New Zealand Joint Registry Annual Report Editorial Committee 1999 (January 1999).

van der List J P, Chawla H, Villa J C, Pearle A D. Why do patellofemoral arthroplasties fail today? A systematic review. Knee 2017; 24(1): 2-8.

Nicol S G, Loveridge J M, Weale A E, Ackroyd C E, Newman J H. Arthritis progression after patellofemoral joint replacement. Knee 2006; 13(4): 290-5. Odgaard A, Emmeluth C, Schrøder H, Kappel A, Madsen F, Troelsen A, Christensen H C, Kyndesen S. Dansk Knaealloplastikregister Årsrapport 2017 (December 2016): 107. Odgaard A, Madsen F, Kristensen P W, Kappel A, Fabrin J. The Mark Coventry Award: Patellofemoral arthroplasty results in better range of movement and early patient-reported outcomes than TKA. Clin Orthop Relat Res 2018; 476(1): 87-100. Odumenya M, Costa M L, Parsons N, Achten J, Dhillon M, Krikler S J. The Avon patellofemoral joint replacement: five-year results from an independent centre. J Bone Joint Surg Br 2010; 92-B(1): 56-60.

van Jonbergen H-P W, Werkman D M, van Kampen A. Conversion of patellofemoral arthroplasty to total knee arthroplasty. Acta Orthop 2009; 80(1): 62-6. van Jonbergen H P W, Werkman D M, Barnaart L F, van Kampen A. Longterm outcomes of patellofemoral arthroplasty. J Arthroplasty 2010; 25(7): 1066-71. Willekens P, Victor J, Verbruggen D, Kerckhove M Vande, Van Der Straeten C. Outcome of patellofemoral arthroplasty, determinants for success. Acta Orthop Belg 2015; 81(4): 759-67. Williams D P, Pandit H G, Athanasou N A, Murray D W, Gibbons C L M H. Early revisions of the Femoro-Patella Vialla joint replacement. Bone Joint J 2013; 95B(6): 793-7. Yadav B, Shaw D, Radcliffe G, Dachepalli S, Kluge W. Mobile-bearing, congruent patellofemoral prosthesis: short-term results. J Orthop Surg 2012; 20(3): 348-52.

Acta Orthopaedica 2019; 90 (5): 479–483

479

Bone remodeling of the proximal tibia after uncemented total knee arthroplasty: secondary endpoints analyzed from a randomized trial comparing monoblock and modular tibia trays—2 year follow-up of 53 cases Mikkel RATHSACH ANDERSEN 1,2, Nikolaj WINTHER 1, Thomas LIND 2, Henrik M SCHRØDER 1, and Michael Mørk PETERSEN 1 1 Department

of Orthopedics, Rigshospitalet, University of Copenhagen, Copenhagen; 2 Department of Orthopedics, Herlev Gentofte Hospital, Copenhagen, Denmark Correspondence: mikkelra1@hotmail.com Submitted 2018-08-21. Accepted 2019-05-17.

Background and purpose — Bone remodeling as a response to bone trauma, postoperative immobilization, and device-related bone reactions can lead to loss of bone stock and increase the risk of periprosthetic fracture and aseptic loosening. This study investigates the adaptive bone remodeling of the proximal tibia after uncemented total knee arthroplasty (TKA). Patients and methods — We performed a 2-year follow up of 53 patients (mean age 62 (38–70) years, 27 of whom were men, who received an uncemented TKA in a randomized controlled trial with bone mineral density (BMD) as secondary endpoint. Patients were randomized to 2 groups of either monoblock (A) or modular (B) polyethylene design. The TKAs were performed using the uncemented Zimmer Nexgen trabecular metal. Measurements of BMD were done postoperatively and after 3, 6, 12, and 24 months. BMD was measured in 3 regions of interest (ROI). Results and interpretation — In group A statistically significant changes in BMD were seen after 24 months in both the medial and lateral ROI. BMD decreased medially by 15% (p = 0.004) and laterally by 13% (p = 0.01). In group B the BMD changes were limited and after 24 months returned to the preoperative values. The differences in BMD change between groups were statistically significant in both the medial (p = 0.03) and lateral (p = 0.02) ROI. In the distal ROI we found no significant change in BMD in either group. A significantly different bone remodeling pattern of the proximal tibia was seen in the 2 groups with a higher degree of bone loss in the knees that received the monoblock polyethylene design, indicating that the flexible monoblock implant design, previously shown to improve fixation, does not decrease the bone loss of the proximal tibia.

Bone remodeling after joint replacement surgery is a wellknown consequence of bone trauma, immobilization, and redistribution of the mechanical loading after joint arthroplasty surgery (Bohr and Lund 1987, Järvinen and Kannus 1997, Soininvaara et al. 2004b, Andersen et al. 2018). After total knee arthroplasty (TKA) bone remodeling occurs in both the proximal tibia and distal femur. Most studies have found TKA to result in a decrease in bone mineral density (BMD) of the tibia and femur (Bohr and Lund 1987, Levitz et al. 1995, Petersen et al. 1995, 1996b, Soininvaara et al. 2004c). The long-term fixation of uncemented tibial components relies on bone ingrowth. The decrease in BMD of the bone is clinically important as BMD is directly related to the breaking strength of the bone (Hansson et al. 1980, Hvid et al. 1985, Petersen et al. 1996a), hence increasing the risk for complications such as periprosthetic fractures; also, the bone loss complicates revisions of the tibial components, and most importantly is related to a high degree of migration and possibly aseptic loosening (Andersen et al. 2017). In this study we investigated the bone remodeling of the proximal tibia after implantation of the Trabecular Metal Technology (TMT) Zimmer Nexgen (Zimmer Biomet, Warsaw, IN, USA), which has a porous tantalum surface. Porous implant surfaces enhance bone ingrowth at the bone–implant interface (Bobyn et al. 1980, 1999, Engh et al. 1987, 1995). In the monoblock implant design the polyethylene is compression-molded directly onto the trabecular metal back. This design in theory eliminates backside wear of the polyethylene. Besides eliminating backside wear the monoblock design also gives the component mechanical properties very similar to those of cancellous bone in terms of low stiffness and high resistance to compressive stress. The modular tibial compo-

480

nent is designed with a titanium plate molded on top of the tantalum trabecular metal in order to create a locking mechanism for the polyethylene (Andersen et al. 2016). This allows the surgeon to change only the polyethylene but can lead to backside wear of the polyethylene. Molding a titanium plate on top of the trabecular metal back, however, makes the modular component stiffer than the monoblock component. These differences in mechanical properties and polyethylene wear could affect the bone remodeling of the tibia (Engh et al. 1987, Bobyn et al. 1999). In this this study we quantify the adaptive bone remodeling of the proximal tibia after uncemented TKA using dual X-ray absorptiometry (DEXA) in a prospective randomized setting comparing monoblock versus modular tibia component design. The bone remodeling of the tibia was a secondary endpoint in a previously published RSA study designed to investigate the migration of uncemented tibia components (Andersen et al. 2016). The randomized RSA study found lower migration of the monoblock group compared with the modular group; however, both implants followed the expected migration pattern of cementless implants, that is high initial migration followed by stabilization from 6 to 24 months postoperatively.

Patients and methods Patients and implants 75 patients scheduled for TKA surgery with a cruciate-retaining TKA because of osteoarthritis were included in a prospective randomized clinical trial with 2 years of follow-up. All patients included were under 70 years of age at the date of the operation, and suffered no bone-related diseases other than osteoarthritis. The patients were randomized to receive the monoblock or modular polyethylene design version of the Cruciate Retaining Trabecular Metal Technology Nexgen tibial component (Andersen et al. 2016). The femur components used in all patients were the cruciate-retaining, uncemented titanium Zimmer Nexgen Flex, and all patients had a cemented Nexgen all poly patella component. The cohort of 53 patients, 26 with monoblock, 27 with modular, included and followed in this study are identical to those of the RSA study previously published and a flowchart and demographics table are included in that publication (Andersen et al. 2016). The 2 groups of patients were comparable in all preoperative demographics, and substantial and similar improvements in clinical results were seen in both groups. DEXA scans The DEXA scans were done using a Norland XR-46 bone densitometer (Norland Corp., Fort Atkinson, WI, USA). We used customized software with an adjustable threshold for metal exclusion allowing BMD measurements of the bone adjacent to the implants. Scan speed was set at 45 mm/sec, and a pixel size of 0.5 × 0.5 mm. The proximal tibias were scanned in

Acta Orthopaedica 2019; 90 (5): 479–483

Figure 1. DEXA scan of the tibia in the anterior–posterior projection with the three regions of interest. ROI 1: Periprosthetic tibial bone from the center of prosthesis to medial edge of the proximal tibia, length 4 cm. ROI 2: Periprosthetic tibial bone from the center of the prosthesis to the lateral edge of the tibia, length 4 cm. ROI 3: Full width of the tibia distal to ROI 1 and RO 2, length 2 cm.

the coronal plane with patients in a standardized supine position, with the knee extended and lower limb slightly internally rotated to avoid tibia–fibula overlay. We also measured BMD of the distal tibia and fibula 1 cm above the ankle joint in all patients in the operated and non-operated limbs. All DEXA scans were performed by an experienced laboratory technician. Patients were scanned postoperatively within 1 week and with follow-ups after 3, 6, 12, and 24 months. On the computerized scan images, we created 3 regions of interest (ROI) in which we measured the periprosthetic BMD changes over time (Figure 1). ROI 1 is a 4 cm long region ranging in width from the medial side of the tibia to the center of the tibia component. ROI 2 is a 4 cm long region stretching from the lateral side of the tibia to the center of the tibia component. ROI 3 is a 2 cm long region located distally to ROI 1 and ROI 2 stretching the entire width of the tibia (Figure 1). DEXA scans were also performed at the ankles, where we created ROI 4 located 1 cm above the ankle joint including both the tibia and the fibula with a length of 2 cm. The ankle bone mineral is reported as bone mineral content (BMC) in grams. We chose BMC because it is more accurate and reproducible than measuring BMD at the ankles due to the patient variation in tibia–fibula overlay (Figure 1). This problem is not relevant in the proximal tibia measurements. Precision error of the tibia BMD measurements was calculated by double DEXA scans of 23 knees and expressed as the mean coefficient of variation (CV). Patients were asked to step off the DEXA scanner, wait for 5 minutes and were then replaced in the standardized supine position and the DEXA scan repeated. Statistics The minimal relevant difference in BMD changes for this study was set at 7.5%. A previous study of BMD changes in

Acta Orthopaedica 2019; 90 (5): 479–483

481

BMD ROI 1 (g/cm²)

BMD ROI 2 (g/cm²)

BMD ROI 3 (g/cm²)

1.3

1.2

1.1

1.0

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

Monoblock Modular Total

0.2 0.1 0.0

0.0 0

Months after index operation

0.0 0

Months after index operation

Figure 2. Change in bone mineral density (BMD [SE]) in ROI 1, ROI 2, and ROI 3 of the proximal tibia.

the proximal tibia after uncemented TKA has found an SD of 7.5% (Petersen et al. 1995). Calculating the sample size (type I error = 5% and type II error = 15%) we found a sample size of 20 in each group. Since the patients included in this study also participated in an RSA study (Andersen et al. 2016) requiring 30 patients per group, we also included 30 patients per group for this study, leaving room for dropouts during the 2-year follow-up period. The BMD data in both groups could be considered normally distributed. We used the unpaired t-test to test for differences in BMD changes over time between the groups after 24 months, and repeated measures ANOVA for BMD change over time within the groups. Statistical analyses were done using the SPSS 21.0 software (IBM Corp., Armonk NY, USA). P-values below 0.05 were considered significant. Ethics, registration, funding, and potential conflicts of interest The study was approved by the Scientific Ethical Committee of Copenhagen (H-1-2012-033), and conducted in accordance with the Helsinki declaration with informed consent obtained (after written and oral information) from all study participants prior to inclusion in the study. The manuscript was written in accordance with the CONSORT guidelines for randomized trials. The study was approved by the Danish Data Protection Agency (ID 01766, GEH-2012-027), and the study was registered at ClinicalTrials.gov (NCT01637051) prior to study start. The study received financial support Zimmer Inc. (Warsaw, IN, USA) and from Gentofte Hospital (research grant). No competing interests were declared.

Results A table with the clinical outcome scores, and preoperative parameters including preoperative knee alignment, has previously been published (Andersen et al. 2016).

The precision expressed as mean (SD) CV of the tibia BMD measurements was 2.9% (2.5%), 2.1% (1.9), and 0.2% (0.2) for respectively ROI 1, ROI 2, and ROI 3. In the medial region (ROI 1) we found an increase in BMD in the modular group of 3.7% during the first 3 months, whereas in the monoblock group BMD decreased by 1.3%. After the first 3 months the BMD in ROI 1 of both groups decreased and did so for the rest of the 24 months’ follow-up period. In the monoblock group the BMD decreased by 4.7%, 9.4%, and 15% in ROI 1 after 6, 12, and 24 months, respectively (Table, see Supplementary data, Figure 2). In the modular group the BMD decreased slowly from 3 to 24 months, when it was close to the postoperative level. When comparing the changes in BMD between the two groups in ROI 1 during the 24 months of follow-up, we found a statistically significant larger BMD decrease in the monoblock group at 24 months of follow-up (p = 0.03). In the lateral region (ROI 2) we also found a BMD decrease in the monoblock group over the 24 months’ follow-up period. After 24 months, the BMD of the monoblock group had decreased to 13% below the postoperative value, while in the modular group only limited insignificant changes were seen at 24 months of follow-up, where BMD was almost on the same level as immediately after the operation. However, as in ROI 1, we found a 6-month increase in the modular group BMD of ROI 2, followed by a slow decrease towards 24 months. When comparing the changes in BMD between the groups in ROI 2 at the 24 months’ follow-up, we found a statistically significant (p = 0.02) difference. In the distal region (ROI 3) the BMD changes in the 2 groups were very limited and statistically insignificant. Furthermore, we found no significant differences between the 2 study groups. The BMC of the ankles was measured in both the operated and non-operated limbs. In general, we found only small nonsignificant changes in BMC of the ankles. There was no significant difference between the two groups regarding BMD changes of the ankle at any time during the follow-up.

482

Discussion To our knowledge, there are no previous studies of bone remodeling of the proximal tibia after uncemented TKA with the modular trabecular metal tibia component. Also, there are no previous studies of tibia bone remodeling comparing the modular versus monoblock design in uncemented TKA. A limitation of this study is that the bone mineral data were collected as a secondary endpoint to the primary RSA study. The study is therefore considered exploratory, investigating whether the difference in mechanical properties and polyethylene wear of the monoblock and modular tibia tray designs has an effect on the bone remodeling of the proximal tibia. The BMD change in our study was not equally distributed in the 2 groups. The monoblock group decreased 15% in ROI 1, 13% in ROI 2, and 3.5% in ROI 3. In contrast, the BMD of the modular group was almost identical to the postoperative value after 2 years. Previous studies of bone remodeling of the proximal tibia after TKA with uncemented tibial components have found quite varying results. Petersen et al. (1995) found a decrease of 22% after 3 years. Levitz et al. (1995) found a 5% decrease per year (36% at 8 years), and Regnér et al. (1999) 26% at 5 years. Conversely Petersen et al. (1996a) found an increase of 6.1% in the lateral tibia after two years. Winther et al. (2016) also reported a small 24 months’ increase in BMD after uncemented TKA in the periprosthetic tibia-bone regions of 1.3%– 5.5% in 57 subjects. Minoda et al. (2013) performed a study using the same uncemented monoblock component as in the present study and compared it with a cemented tibia implant. They defined ROIs that were similar to ours. They found BMD changes of –41%, –12%, and –4% in the regions corresponding to our ROI 1, ROI 2, and ROI 3 respectively after 5 years. To our knowledge there are no previous bone remodeling studies of the modular component. The BMD of proximal ROI1 and RIO2 of the modular group increased during the first 6 months. A possible explanation for this could be the result of a larger average realignment correction in the modular group. Tables presenting preoperative demographics and clinical parameters including preand postoperative alignment were presented in our previous publication (Andersen et al. 2016). The altered weight transfer after realignment to a more physiologic alignment with altered mechanical loading of the proximal tibia postoperatively could explain the initial increase in ROI 1 and ROI 2. This temporary increase as an effect of realignment in TKA has previously been reported (Bohr and Lund 1987, Soininvaara et al. 2004b). Patients with knee arthrosis are known to have a higher BMD of the proximal tibia than in healthy knees and part of the postoperative bone loss could represent a return to normal (Madsen et al. 1994). This could also explain part of the difference between our groups as the postopera-

Acta Orthopaedica 2019; 90 (5): 479–483

tive BMD of ROI 1 of the monoblock group was 8% above the modular group (0.93 g/cm2 versus 0.87 g/cm2) and 5% above the modular group in ROI 2 (0.98 g/cm2 versus 0.95 g/ cm2). In ROI 3 the 2 groups had identical average postoperative BMD values of 1.3 g/cm2. The result that stands out most when comparing our results with previously reported results is the lack of bone loss in the modular group during the 2-year follow-up. We believe that the best explanation for this is the above-mentioned lower average starting point BMD and the larger shift in mechanical loading due to realignment, rather than mechanical properties specific to the modular design; if the follow-up period had been longer, the BMD graphs for the modular group probably would continue to decline at a rate similar to that of the monoblock group. The difference in BMD change between the groups did not reflect different knee function of the subjects, and our previous publication (Andersen et al. 2016) documented a good clinical outcome in both groups concerning Knee Society Scores and life quality and we found no statistically significant difference between the groups for these parameters. We also performed BMD measurements of the distal tibia at the ankles in all patients to ensure that the BMD change in the proximal tibia was a true local response to the operation and tibial component implantation and not an overall decrease in skeletal BMD or a result of immobilization of the operated limb. The BMD changes of the ankles were statistically insignificant and smaller than BMD changes of the proximal tibia, indicating that the significant BMD change of the ROIs in the proximal tibia was largely a true local response to the TKA surgery. However, they tended to follow the same pattern with a decrease in BMD in the monoblock group of both the operated and non-operated limbs, whereas in the modular group the BMD increased during the follow-up period (data not shown). The initial BMD increase in the modular group was followed by a slow decrease in BMD over the next 18 months with a slope similar to that of the monoblock group, indicating that a longer follow-up period would also have resulted in negative BMD changes in the modular group. The BMD of all the 3 ROIs in all subjects in this study decreased at a rate of 3.4% per year. This result corresponds relatively well to the 26% 5-year decrease in the study by Regnér et al. (1999), and the 40% 8-year decrease seen in the study by Levitz et al. (1995). From the RSA study of these patients we knew that the flexible monoblock component migrated statistically significantly less than the rigid modular component, which led us to hypothesize that we would find less bone loss in the monoblock group. We attributed the lower migration of the monoblock component to better mechanical properties and reduced backside wear and expected the superior fixation to result in a lower degree of bone loss. However, our findings indicate that this hypothesis was false. It is likely that other parameters than the degree of implant fixation have a greater effect on bone remodeling of the tibia; an example of such factors could be weight-transfer shifts after knee realignment and differences

Acta Orthopaedica 2019; 90 (5): 479–483

in preoperative level of BMD. Studies of distal femur bone remodeling after TKA support this notion, as the posterior shift in weight transfer and the absence of patellar pressure after TKA are believed to explain the large difference in bone loss anteriorly and posteriorly in the distal femur (Soininvaara et al. 2004c), Andersen et al. 2018). Though we did not find the expected relation between implant design and bone remodeling we do consider investigating such relations important becauses the bone loss of the tibia is clinically important as BMD is directly related to the breaking strength of the bone and complicates revision surgery (Hansson et al. 1980, Hvid et al. 1985, Petersen et al. 1996a). We considered a minimal difference of 7.5% BMD change between the groups to be clinically significant. We conclude that we did find a clinically relevant difference in the amount of bone loss in ROI 1 and ROI 2 over the 2-year follow-up, when compared with an expected bone loss of 1–2% in the general population during a 2-year period, but the difference probably should not be attributed to the mechanical properties of the tibial implants as hypothesized. A TKA is expected to last at least a 15-year period during which the BMD of the proximal tibia could be expected to decrease by approximately 50–80% assuming the BMD loss continues over longer periods at the reported rates. Such loss of tibia bone quality could result in serious complications such as implant loosening and periprosthetic fractures, and provide surgical difficulties in TKA revisions. As TKA patients become younger, poor tibia bone quality could represent an increasing problem in future revision arthroplasty surgery. Supplementary data The Table is available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/17453674. 2019.1637178 MRA: Study conception and design, acquisition of data, analysis and interpretation of data, drafting of manuscript, critical revision. NW: Study conception and design. TL: Study conception and design, acquisition of data. HMS: Study conception and design, acquisition of data. MMP: Study conception and design, analysis and interpretation of data, critical revision. The authors would like to thank Karen Elisabeth Sønderlev for technical support performing DEXA scans. Acta thanks Harald Brismar and Rüdiger J Weiss for help with peer review of this study.

Andersen M R, Winther N, Lind T, SchrøDer H, Flivik G, Petersen M M. Monoblock versus modular polyethylene insert in uncemented total knee arthroplasty. Acta Orthop 2016; 87(6): 607-14. Andersen M R, Winther N S, Lind T, Schrøder H M, Flivik G, Petersen M M. Low preoperative BMD is related to high migration of tibia components in uncemented TKA-92 patients in a combined DEXA and RSA study with 2-year follow-up. J Arthroplasty 2017; 32(7): 2141-6.

483

Andersen M R, Winther N S, Lind T, Schrøder H M, Mørk Petersen M. Bone remodeling of the distal femur after uncemented total knee arthroplasty-: a 2-year prospective DXA study. J Clin Densitom 2018; 21(2): 236-43. Bobyn J D, Pilliar R M, Cameron H U, Weatherly G C, Kent G M. The effect of porous surface configuration on the tensile strength of fixation of implants by bone ingrowth. Clin Orthop Rel Res 1980; (149): 291-8. Bobyn J D, Stackpool G J, Hacking S A, Tanzer M, Krygier J J. Characteristics of bone ingrowth and interface mechanics of a new porous tantalum biomaterial. J Bone Joint Surg Br 1999; 81(5): 907-14. Bohr H H, Lund B. Bone mineral density of the proximal tibia following uncemented arthroplasty. J Arthroplasty 1987; 2(4): 309-12. Engh C A, Bobyn J D, Glassman A H. Porous-coated hip replacement: the factors governing bone ingrowth, stress shielding, and clinical results. J Bone Joint Surg Br 1987; 69(1): 45-55. Engh C A, Hooten J P, Zettl-Schaffer K F, Ghaffarpour M, McGovern T F, Bobyn J D. Evaluation of bone ingrowth in proximally and extensively porous-coated anatomic medullary locking prostheses retrieved at autopsy. J Bone Joint Surg Am 1995; 77(6): 903-10. Hansson T, Roos B, Nachemson A. The bone mineral content and ultimate compressive strength of lumbar vertebrae. Spine 1980; 5(1): 46-55. Hvid I, Jensen N C, Bünger C, Sølund K, Djurhuus J C. Bone mineral assay: its relation to the mechanical strength of cancellous bone. Eng Med 1985; 14(2): 79-83. Järvinen M, Kannus P. Injury of an extremity as a risk factor for the development of osteoporosis. J Bone Joint Surg Am 1997; 79(2): 263-76. Levitz C L, Lotke P A, Karp J S. Long-term changes in bone mineral density following total knee replacement. Clin Orthop Rel Res 1995 (321): 68-72. Madsen O R, Schaadt O, Bliddal H, Egsmose C, Sylvest J. Bone mineral distribution of the proximal tibia in gonarthrosis assessed in vivo by photon absorption. Osteoarthritis Cartilage 1994; 2(2): 141-7. Minoda Y, Kobayashi A, Ikebuchi M, Iwaki H, Inori F, Nakamura H. Porous tantalum tibial component prevents periprosthetic loss of bone mineral density after total knee arthroplasty for five years: a matched cohort study. J Arthroplasty 2013; 28(10): 1760-4. Petersen M M, Nielsen P T, Lauritzen J B, Lund B. Changes in bone mineral density of the proximal tibia after uncemented total knee arthroplasty: a 3-year follow-up of 25 knees. Acta Orthop Scand 1995; 66(6): 513-6. Petersen M M, Jensen N C, Gehrchen P M, Nielsen P K, Nielsen P T. The relation between trabecular bone strength and bone mineral density assessed by dual photon and dual energy X-ray absorptiometry in the proximal tibia. Calcif Tissue Int 1996a; 59(4): 311-4. Petersen M M, Lauritzen J B, Pedersen J G, Lund B. Decreased bone density of the distal femur after uncemented knee arthroplasty: a 1-year follow-up of 29 knees. Acta Orthop Scand 1996b; 67(4): 339-44. Regnér L R, Carlsson L V, Kärrholm J N, Hansson T H, Herberts P G, Swanpalmer J. Bone mineral and migratory patterns in uncemented total knee arthroplasties: a randomized 5-year follow-up study of 38 knees. Acta Orthop Scand 1999; 70(6): 603-8. Soininvaara T A, Miettinen H J A, Jurvelin J S, Suomalainen O T, Alhava E M, Kröger H P. Periprosthetic tibial bone mineral density changes after total knee arthroplasty: one-year follow-up study of 69 patients. Acta Orthop Scand 2004a; 75(5): 600-5. Soininvaara T A, Miettinen H J A, Jurvelin J S, Suomalainen O T, Alhava E M, Kröger H P J. Periprosthetic femoral bone loss after total knee arthroplasty: 1-year follow-up study of 69 patients. The Knee 2004b; 11(4): 297-302. Soininvaara T A, Miettinen H J, Jurvelin J S, Alhava E M, Kröger H P. Bone mineral density in the proximal femur and contralateral knee after total knee arthroplasty. J Clin Densitom 2004c;7(4): 424-31. Winther N, Jensen C, Petersen M, Lind T, Schrøder H, Petersen M. Changes in bone mineral density of the proximal tibia after uncemented total knee arthroplasty. A prospective randomized study. Int Orthop 2016; 40(2): 285-94.

484

Acta Orthopaedica 2019; 90 (5): 484–488

Manipulation under anesthesia after primary knee arthroplasty in Sweden: incidence, patient characteristics and risk of revision Hunbogi THORSTEINSSON 1, Margareta HEDSTRÖM 3,4, Otto ROBERTSSON 1,2, Natalie LUNDIN 3,4, and Annette W-DAHL 1,2 1 Department

of Orthopedics, Skane University Hospital, Lund; 2 The Swedish Knee Arthroplasty Register; 3 Department of Orthopedics, Karolinska University Hospital, Huddinge; 4 Department of Clinical Science, Intervention and Technology, Division of Orthopaedics and Biotechnology, Karolinska Institutet, Stockholm, Sweden Correspondence: hunbogi1@gmail.com Submitted 2019-03-01. Accepted 2019-06-01.

Background and purpose — The incidence of manipulation under anesthesia (MUA) after knee arthroplasty surgery has been reported to vary between 0.5% and 10%. We evaluated the incidence of MUA after primary knee arthroplasty in Sweden, the demographics of the patients and the risk of revision. Patients and methods — Between 2009 and 2013, 64,840 primary total and unicompartmental knee arthroplasties (TKA and UKA) were registered in the Swedish Knee Arthroplasty Register (SKAR). MUAs performed between 2009 and 2014 were identified through the in- and outpatient registers of the Swedish National Board of Health and Welfare. Pertinent data were verified through medical records and patient demographics and revisions were obtained from the SKAR. Results — 1,258 MUAs were identified. Of these, 1,078 were 1st-time MUAs, performed within 1 year after the primary knee arthroplasty. The incidence of MUA was 1.7% and the incidence varied between hospitals from 0% to 5%. The majority were performed after TKA (98%), in younger patients (65% < 65 years), women (64%), and relatively healthy persons (88% had ASA ≤ 2). The cumulative risk of revision at 10 years was 10% (95% CI 8.6–12), similar for men and women. Interpretation — In Sweden, MUA is a rather uncommon measure after knee arthroplasty, especially after UKA. The CRR at 10 years was doubled compared to the general knee arthroplasty population. The frequency of the procedure varies between hospitals but in general it is performed more frequently in healthier and younger patients.

Joint stiffness following knee arthroplasty is a disabling complication. One treatment option is manipulation under anesthesia (MUA). However, the literature describes no clear definition/consensus of stiffness or the indications for MUA. Numerous potential risk factors have been reported for insufficient knee range of motion (ROM) after knee arthroplasty, among others younger age (Issa et al. 2015, Werner et al. 2015, Plate et al. 2016, Newman et al. 2018), female sex (Gadinsky et al. 2011, Werner et al. 2015), ethnicity (Issa et al. 2015), high BMI (Gadinsky et al. 2011), smoking (Werner et al. 2015, Issa et al. 2015, Newman et al. 2018), comorbidities such as diabetes (Plate et al. 2016), warfarin treatment (Desai et al. 2014), history of previous knee surgery (Plate et al. 2016, Newman et al. 2018), and limited preoperative ROM (Ritter et al. 2003, Kim et al. 2004, Keating et al. 2007, Issa et al. 2015, Newman et al. 2018). The incidence of MUA after knee arthroplasty surgery has been reported to range from 0.5% to almost 10% (Table 1, see Supplementary data). Most were single-center studies performed in the United States with relatively few patients. 2 large US studies (Pfefferle et al. 2014, Werner et al. 2015) were based on the PearlDiver database (publicly available including private payers and Medicare data) and the Explorys database (commercially available including electronic healthcare data) but these showed different MUA incidence (1.5% and 4.3% respectively). A Finnish study reported on the incidence of MUA in 1 hospital during the first 6 months after the primary knee arthroplasty surgery, before and after implementing fast-track (Pamilo et al. 2018). They found a similar incidence before (2009–2010) and after (2012–2013) fast-track (6%). A Danish study on 359 TKAs was also per-

Acta Orthopaedica 2019; 90 (5): 484–488

485

formed in a fast-track hospital and found the same incidence of MUA (Wied et al. 2015). The risk of revision after MUA has been sparsely studied. Werner et al. (2015) found that patients who required MUA after TKA had an increased risk of revision while Pierce et al. (2017) did not find any increased risk in a matched casecontrol study including 138 patients. We evaluated the incidence of 1st-time MUA performed within 1 year after primary total knee arthroplasty (TKA) and unicompartmental knee arthroplasty (UKA) surgery in Sweden, describe the demographics of the patients and the risk of revision.

Total number of MUAs 2009–2014 n = 1,544 Excluded (n = 286): – missing records, 40 – duplicates, 41 – MUA after revision/reoperation, 34 – MUA/primary outside study period, 21 – other, 23 – arthroscopic adhesiotomy, 49 – open adhesiotomy, 78 MUA n = 1,258 Excluded reapted MUA (n = 108): – 2 times, 90 – 3 times, 17 – 4 times, 1

Patients and methods We requested information on the 64,840 primary TKA (n = 61,835, 95.4%) and UKA (n = 3,005, 4.6%) that were registered in the Swedish Knee Arthroplasty Register (SKAR) between 2009 and 2013 from the Swedish National Board of Health and Welfare patient register (PAS). The SKAR has registered knee arthroplasties since 1975 and has a high completeness and correctness of data (SKAR 2018). The PAS contains in- and outpatient care, admission date, discharge date, surgical code (NOMESCO), diagnosis code (ICD-10), and operating hospital. Based on the patient’s personal identification number (that includes information on date of birth and sex), contained in both registers, we requested information on knee arthroplasties registered in the PAS 2009–2014 with ICD-10 codes for joint stiffness (M24.5, M24.6, M25.6) together with the NOMESCO code for MUA (NGT19) or the codes for percutaneous, arthroscopic and open adhesiotomy (NGH30, NGH31, NGH32). After receiving information from the PAS, the hospitals where the manipulations were performed were requested to provide medical records related to the manipulation in order to verify the side, diagnosis, NOMESCO code, surgical date, length of stay (LOS), and comorbidities. Further we had the intention to gather information from the records on the ROM before, during, and after the manipulation as well as the care after the manipulation. Information on patient characteristics, age, sex, BMI, the ASA class, and history of prior knee surgery were obtained from the SKAR. As the purpose of the study was to evaluate the incidence of MUA, we excluded percutaneous, arthroscopic, and open adhesiotomies, MUAs after revisions and reoperations, MUAs or primary knee arthroplasties performed outside the study period, duplicates, and those not verified as MUA. Further we included only the 1st-time MUAs performed within 1 year after the primary knee arthroplasty. The patients were followed-up until December 31, 2018. We use descriptive statistics and present the data in numbers and proportions.

First-time MUA n = 1,150 Excluded: MUA > 1 year after primary operation n = 72 First-time MUA < 1 year after primary operation n = 1,150

Figure 1. Flowchart of the study population.

Statistics Cumulative revision rate (CRR) curves were produced using the life table method with monthly intervals with 95% confidence intervals (CIs) calculated with the Wilson quadratic equation using the Greenwood and Peto effective sample size estimates (Dorey et al. 1993). When comparing the risk of age (continuous variable), sex, and MUA performed before or after 8 weeks from the primary knee arthroplasty, Cox regression was used to calculate relative risk estimates (RR) with CI. The reason for revision was presented as numbers and proportions. Ethics, funding, and potential conflicts of interests The study was approved by the regional Ethics Committee of Stockholm (2015/978-31), and was performed in accordance with the Declaration of Helsinki. The study was not financed by any external funding. The authors declare no conflicts of interest.

Results We identified 1,258 MUAs of which 1,150 were 1st-time MUAs with 1,078 MUAs being performed within 1 year of the primary knee arthroplasty. All the hospitals (n = 75) responded to our request for medical records but in 40 cases (2.6%) a record could not be found (Figure 1).

486

Acta Orthopaedica 2019; 90 (5): 484–488

Table 2. Characteristics of MUA patients and the general knee arthroplasty (TKA and UKA) population 2009–2013 in Sweden Factor

MUA (n = 1,154)

SKAR (n = 64,840)

Sex female, n (%) 729 (63) 37,490 (58) Age, mean (SD) 61 (9) 69 (9) ASA, n (%) a 1,136 63,440 I 336 (30) 12,345 (19) II 671 (59) 41,010 (65) III–IV 129 (11) 10,085 (16) BMI a 1,133 63,347 mean (SD) 29 (5) 29 (5) OA n (%) 1,091 (95) 62,042 (96) Prior knee surgery a 1,125 62,934 n (%) 406 (36) 12,454 (20) a Number

Incidence (%) of MUA within 1 year 6 5 4 3 2 1 0

Number of hospital Figure 3. Incidence of MUA/hospital.

of cases with data

The incidence of 1st-time MUA within 1 year after knee arthroplasty was 1.7% and was similar through the years (Figure 2, see Supplemenatry data). The incidence of MUA varied between hospitals from 0% to 5% (Figure 3). Of the TKAs, 1.7% (n = 1,061) underwent MUA and among the UKAs 0.6% (n = 17). Of the 1,078 MUAs, 60% were performed within 3 months after the primary knee arthroplasty. The vast majority (n = 1,011, 94%) of the MUA patients were treated as inpatients and the median LOS was 2 days (0–20) with 95% of the patients staying 0–6 days. As compared with the general knee arthroplasty population in Sweden, the MUA patients were younger (65% < 65 years), more often women, somewhat healthier and more often had a history of prior knee surgery. The BMI and diagnoses were similar (Table 2). 5.5% of the MUA patients were diagnosed with diabetes mellitus and 5.5% had warfarin treatment. Corresponding figures for the general arthroplasty population are not available. Among the 1,078 MUAs there were 109 revisions. The CRR at 10 years was 10% (CI 8.6–12). We found no statistically significant difference in risk of revision depending on sex (men hazard ratio [HR] 0.9 [CI 0.6–1.4]), age (HR 1 [CI 0.98–1]), or if the MUA was performed before or after 8 weeks following the knee arthroplasty surgery (< 8 weeks HR 1 [CI 0.6–1.6]). Femoro-patellar problems were the most common reason for revision (26%) followed by loosening and stiffness (Table 3, see Supplementary data).

Discussion To our knowledge, this is the first nationwide study showing the incidence of MUA after primary knee arthroplasty surgery. All the hospitals answered our requests for medical records but 3% of the records could not be found and may have been

misclassified. All privately run hospitals were represented in the PAS so we feel confident that the absolute majority of procedures were captured. However, unfortunately the medical records proved to be insufficient to evaluate the exact ROM before, during, and after MUA, regarding the change in ROM achieved during or after the MUA, or whether the patients were satisfied with the results. We found the incidence of a 1st-time MUA within 1 year of knee arthroplasty surgery to be 1.7%, i.e., a rather uncommon procedure. Our incidence may be regarded to be low as compared with what has most commonly been reported (see Supplementary data). The 2 Nordic studies (Wied et al. 2015, Pamilo et al. 2018) included patients operated with TKA during the same time period as the patients in our study, but in fast-track hospitals. These studies showed an incidence of MUA of almost 6%, which was comparable to the hospital in Sweden with the highest incidence (5%). The SKAR has no information on whether the hospitals consider themselves as fast-track hospitals or not. On the other hand, we could not see a difference in incidence between government and private-run or high- and low-volume hospitals. However, the incidence varied between hospitals in Sweden in a similar way to the incidences from different hospitals in the literature (Table 1). This may reflect the highlighted lack of clear indications for MUA after knee arthroplasty surgery. Several variables have been suggested as potential risk factors for stiffness requiring MUA, but little consensus exists (Kornuijt et al. 2018). The Swedish MUA population was more often women, younger, and somewhat healthier but had a higher proportion of previous knee surgery compared twith the general knee arthroplasty population. Diabetes and warfarin treatment have been suggested as potential risk factors (Issa et al. 2014b, Pfefferle et al. 2014) for joint stiffness but that information is not available in the SKAR and we do not know if they are over-represented in the Swedish MUA population.

Acta Orthopaedica 2019; 90 (5): 484–488

The optimal timing of MUA is unknown. Early intervention has been suggested to be favorable (Bawa et al. 2013, Issa et al. 2014a, Desai et al. 2014, Ferrel et al. 2015, Vanlommel et al. 2017, Newman et al. 2018) while others found no difference between early and late intervention (Ipach et al. 2011, Yeoh et al. 2012). However, early intervention has been reported to vary from ≤ 6 to 20 weeks in the above-mentioned studies. Werner et al. (2015) found that the risk of revision was less common in patients who underwent MUA within 8 weeks after the primary TKA (94/2,465 [3.8%]) as compared with patients who underwent MUA between 8 weeks and 3 months after (99/1,870 [5.3%]), p = 0.02. However, we found no statistically significant difference in risk of revision depending on whether the MUA was performed ≤ 8 or > 8 weeks. The reasons may be the difference in the number of MUAs in the studies and that our time limit for “late” MUAs was not 3 months but 1 year. Further, we found that the MUA patients had approximately double the 10-year CRR of the general knee arthroplasty population in Sweden (SKAR 2018). This is in line with the findings of Werner et al. (2015), who found that patients who required MUA after the primary TKA had increased risk of revision (4.8%, within 7 years) as compared with those not requiring MUA (2%) (OR 2.4, CI 2.1–2.8). Of the 109 revisions, 18 were due to stiffness and the rest for other reasons, and may not have had anything to do with the stiffness/MUA. We feel confident with the reasons for revision as we routinely read all the surgical records and the discharge letters at the register. We found that femoro-patellar problems comprised the most common reason for revision in our MUA cohort (26%) with infection (12%) being the 4th most common. Werner et al. (2015) reported only on the frequency of infections as reason for revision and found them to account for 16%. The relatively large variation in the incidence between hospitals in Sweden may indicate that factors other than known risk factors such as sex, age, health, and postoperative ROM influence the decision to perform MUA. Rather, the decision may to a larger extent be affected by the patient’s expectations and motivation as well as the surgeon’s expectations and willingness to perform MUA and not least the available resources in the hospitals concerned. In summary, in Sweden, MUA is a rather uncommon measure after knee arthroplasty, especially after UKA, and has double the CRR at 10 years as compared with the general knee arthroplasty population. The frequency of the procedure varies between hospitals but in general MUA is performed more frequently in healthier and younger patients. Supplementary data Tables 1 and 3 and Figure 2 are available as supplementary data in the online version of this article, http://dx.doi.org/10.1080/ 17453674.2019.1637177

487

The study was conceived by MH, OR, and AWD. HT, NL, MH, and AWD obtained and compiled the information from the hospitals. OR and AWD performed the analyses. HT and AWD wrote the initial draft. All the authors contributed to the interpretation of the data and to revision of the manuscript. The authors would like to thank the hospitals for their contribution to the study. Acta thanks Walter van der Weegen and other anonymous reviewers for help with peer review of this study.

Bawa H S, Wera G D, Kraay M J, Marcus R E, Goldberg V M. Predictors of range of motion in patients undergoing manipulation after TKA. Clin Orthop Relat Res 2013; 471(1): 258-63. Desai A S, Karmegam A, Dramis A, Board N, Raut V. Manipulation for stiffness following total knee arthroplasty: when and how often to do it? Eur J Orthop Surg Traumatol. 2014; 24(7): 1291-5. Dorey F, Nasser S, Amstutz H. The need for confidence intervals in the presentation of orthopaedic data. J Bone Joint Surg Am 1993; 75(12): 1844-52. Esler C N, Lock K, Harper W M, Gregg P J. Manipulation of total knee replacements. Is the flexion gained retained? J Bone Joint Surg Br 1999; 81(1): 27-9. Ferrel J R, Davis R L 2nd, Agha OA, Politi J R. Repeat manipulation under anesthesia for persistent stiffness after total knee arthroplasty achieves functional range of motion. Surg Technol Int 2015; 26: 256-60. Gadinsky N E, Ehrhardt J K, Urband C, Westrich G H. Effect of body mass index on range of motion and manipulation after total knee arthroplasty. J Arthroplasty 2011; 26(8): 1194-7. Ipach I, Mittag F, Lahrmann J, Kunze B, Kluba T. Arthrofibrosis after TKA: influence factors on the absolute flexion and gain in flexion after manipulation under anaesthesia. BMC Musculoskelet Disord 2011; 12: 184. Issa K, Banerjee S, Kester M A, Khanuja H S, Delanois R E, Mont M A. The effect of timing of manipulation under anesthesia to improve range of motion and functional outcomes following total knee arthroplasty. J Bone Joint Surg Am 2014a; 96(16): 1349-57. Issa K, Kapadia B H, Kester M, Khanuja H S, Delanois R E, Mont M A. Clinical, objective, and functional outcomes of manipulation under anesthesia to treat knee stiffness following total knee arthroplasty. J Arthroplasty 2014b; 29(3): 548-52. Issa K, Rifai A, Boylan MR, Pourtaheri S, McInerney V K, Mont M A. Do various factors affect the frequency of manipulation under anesthesia after primary total knee arthroplasty? Clin Orthop Relat Res 2015; 473(1): 143-7. Keating E M, Ritter M A, Harty L D, Haas G, Meding J B, Faris P M, Berend M E. Manipulation after total knee arthroplasty. J Bone Joint Surg Am 2007; 89(2): 282-6. Kelly M P, Prentice H A, Wang W, Fasig B H, Sheth D S, Paxton E W. Reasons for ninety-day emergency visits and readmissions after elective total joint arthroplasty: results from a US integrated healthcare system. J Arthroplasty 2018; 33(7): 2075-81. Kim J, Nelson C L, Lotke P A. Stiffness after total knee arthroplasty: prevalence of the complication and outcomes of revision. J Bone Joint Surg Am 2004; 86-A(7): 1479-84. Kornuijt A, Das D, Sijbesma T, de Vries L, van der Weegen W. Manipulation under anesthesia following total knee arthroplasty: a comprehensive review of literature. Musculoskelet Surg 2018; 102(3): 223-30. Namba R S, Inacio M. Early and late manipulation improve flexion after total knee arthroplasty. J Arthroplasty 2007; 22(6 Suppl. 2): 58-61. Newman E T, Herschmiller T A, Attarian D E, Vail T P, Bolognesi M P, Wellman S S. Risk factors, outcomes, and timing of manipulation under anesthesia after total knee arthroplasty. J Arthroplasty 2018; 33(1): 245-9.

488

Pagoti R, O’Brien S, Blaney J, Doran E, Beverland D. Knee manipulation for reduced flexion after total knee arthroplasty: is timing critical? J Clin Orthop Trauma 2018; 9(4): 295-9. Pamilo K J, Torkki P, Peltola M, Pesola M, Remes V, Paloneva J. Fast-tracking for total knee replacement reduces use of institutional care without compromising quality. Acta Orthop 2018; 89(2): 184-9. Pfefferle K J, Shemory S T, Dilisio M F, Fening S D, Gradisar I M. Risk factors for manipulation after total knee arthroplasty: a pooled electronic health record database study. J Arthroplasty 2014; 29(10): 2036-8. Pierce T P, Issa K, Festa A, Scillia A J, McInerney V K, Mont M A. Does manipulation under anesthesia increase the risk of revision total knee arthroplasty? A matched case control study. J Knee Surgery 2017; 30(7): 730-3. Plate J F, Wohler A D, Brown M L, Sun D, Fino N F, Lang J E. Factors associated with range of motion recovery following manipulation under anesthesia. Surg Technol Int 2016; 29: 295-301. Ritter M A, Harty L D, Davis K E, Meding J B, Berend M E. Predicting range of motion after total knee arthroplasty: clustering, log-linear regression, and regression tree analysis. J Bone Joint Surg Am 2003; 85-A(7): 1278–85.

Acta Orthopaedica 2019; 90 (5): 484–488

SKAR. The Swedish Knee Arthroplasty Register Annual Report 2018. 2018; ISBN 978-91-88017-19-2. Vanlommel L, Luyckx T, Vercruysse G, Bellemans J, Vandenneucker H. Predictors of outcome after manipulation under anaesthesia in patients with a stiff total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2017; 25(11): 3637-43. Werner B C, Carr J B, Wiggins J C, Gwathmey F W, Browne J A. Manipulation under anesthesia after total knee arthroplasty is associated with an increased incidence of subsequent revision surgery. J Arthroplasty 2015; 30(9 Suppl.): 72-5. Wied C, Thomsen M G, Kallemose T, Myhrmann L, Jensen L S, Husted H, Troelsen A. The risk of manipulation under anesthesia due to unsatisfactory knee flexion after fast-track total knee arthroplasty. Knee 2015; 22(5): 419-23. Yeoh D, Nicolaou N, Goddard R, Willmott H, Miles K, East D, Hinves B, Shepperd J, Butler-Manuel A. Manipulation under anaesthesia post total knee replacement: long term follow up. Knee 2012; 19(4): 329-31. Yoo J H, Oh J C, Oh H C, Park S H. Manipulation under anesthesia for stiffness after total knee arthroplasty. Knee Surg Relat Res 2015; 27(4): 233-9.

Acta Orthopaedica 2019; 90 (5): 489–494

489

Increased use of total shoulder arthroplasty for osteoarthritis and improved patient-reported outcome in Denmark, 2006–2015: a nationwide cohort study from the Danish Shoulder Arthroplasty Registry Jeppe V RASMUSSEN 1,2, Alexander AMUNDSEN 1, Anne Kathrine B SØRENSEN 1, Tobias W KLAUSEN 3, John JAKOBSEN 4, Steen L JENSEN 4,5, and Bo S OLSEN 1,2 1 Department

of Orthopaedic Surgery, Herlev Hospital, Herlev; 2 Department of Clinical Medicine, University of Copenhagen; 3 Department of Hematology, Herlev Hospital, Herlev; 4 Department of Orthopaedic Surgery, Aalborg University Hospital, Aalborg; 5 Department of Clinical Medicine, Aalborg University, Denmark Correspondence: jevera01@heh.regionh Submitted 2019-01-20. Accepted 2019-06-03.

Background and purpose — Osteoarthritis has become the most common indication for shoulder arthroplasty in Denmark, and the treatment strategies have changed towards the use of anatomical total shoulder arthroplasty and reverse shoulder arthroplasty. We investigated whether changes in the use of arthroplasty types have changed the overall patient-reported outcome from 2006 to 2015. Patients and methods — We included 2,867 shoulder arthroplasties performed for osteoarthritis between 2006 and 2015 and reported to the Danish Shoulder Arthroplasty Registry. The Western Ontario Osteoarthritis of the Shoulder (WOOS) index at 1 year was used as patient-reported outcome. The raw score was converted to a percentage of a maximum score. General linear models were used to analyze differences in WOOS. Results — The proportion of anatomical total shoulder arthroplasty and reverse shoulder arthroplasty increased from 3% and 7% in 2006 to 53% and 27% in 2015. The mean WOOS score was 70 (SD 26) after resurfacing hemiarthroplasties (n = 1,258), 68 (SD 26) after stemmed hemiarthroplasty (n = 500), 82 (SD 23) after anatomical total shoulder arthroplasties (n = 815), and 74 (SD 23) after reverse shoulder arthroplasties (n = 213). During the study period, the overall WOOS score increased with 18 (95% CI 12–22) in the univariate model and 10 (CI 5–15) in the multivariable model, and the WOOS scores for anatomical total shoulder arthroplasty increased by 14 (CI 5–23). Interpretation — We found an increased WOOS score from 2006 to 2015, which was primarily related to a higher proportion of anatomical total shoulder arthroplasty and reverse shoulder arthroplasty towards the end of the study period, and to improved outcome of anatomical total shoulder arthroplasty.

The Danish Shoulder Arthroplasty Registry (DSR) was started in 2004 (Rasmussen et al. 2012). Each year a report is published, which includes results and recommendations. Furthermore, peer-review publications from the registry are described in the annual reports and the results are presented at annual meetings of the Danish Society of Shoulder and Elbow Surgery and at the Danish Society of Orthopedic Surgery. Because of limited information from randomized trials and large observational studies, data from the DSR are an important source of information regarding patient-reported outcome, arthroplasty survival rates, and reasons for revision of different arthroplasty types. Osteoarthritis has become the most common indication for shoulder arthroplasty in Denmark. Furthermore, the treatment strategies have changed towards the use of anatomical total shoulder arthroplasty and reverse shoulder arthroplasty. We hypothesized that changes in the use of arthroplasty types have improved the overall patient-reported outcome of shoulder arthroplasty for osteoarthritis. We studied changes in the use of arthroplasty types for osteoarthritis in Denmark from 2006 to 2015, the patientreported outcome of different arthroplasty types, and whether changes in the use of arthroplasty types have changed the overall patient-reported outcome from 2006 to 2015.

Patient and methods Source of data Data were obtained from the DSR. All Danish hospitals and private clinics report patient and surgical data at the time of operation. Every year the completeness of reporting is calculated by comparing data from the DSR with data from the

490

National Patient Register—an administrative database used by the Danish healthcare authorities to reimburse expenses for any hospital treatment including shoulder arthroplasty. In its first 2 years of existence, the completeness of reporting to the registry was low, and therefore 2004–2005 is now regarded as a trial period, and the results usually not included in research. Reporting to the registry became mandatory in 2006. Since 2007, the completeness has been above 90% and the completeness for this report was 93% for both primary and revision arthroplasties. A patient-specific identification number given to all Danish citizens at the time of birth or immigration was used by the registry to link a revision to the primary procedure. The identification number was also used when information concerning death or emigration was obtained from the National Registry of Persons. Inclusion criteria We included all shoulder arthroplasties performed for osteoarthritis between 2006 and 2015 that had been reported to the DSR. In that period it was possible to report more than 1 indication, so for this study we used a hierarchy including the following indications ranked in descending order: acute fracture; fracture sequelae (nonunion; malunion; previous osteosynthesis; fractures reported together with osteoarthritis or humeral head necrosis); inflammatory arthritis; rotator cuff arthropathy; and osteoarthritis. If more than 1 indication was reported only that with the highest rank in the hierarchy was recorded. Patient-reported outcome The Western Ontario Osteoarthritis of the Shoulder (WOOS) index was not assessed preoperatively but assessed by a postal survey at 1 year. For economic and logistical reasons, the survey was only sent to the patients once. In case of revision, death, or emigration within the first year, the WOOS score cannot be obtained. WOOS is a disease-specific questionnaire that measures the quality of life of patients with glenohumeral osteoarthritis (Lo et al. 2001). The total score ranges from 0 to 1,900, with 1,900 being the worst. For simplicity of presentation, the total score is converted to a percentage of the maximum score, with 100 being the best. The Danish version of WOOS has been translated and cross-culturally adapted (Rasmussen et al. 2013). The minimal clinically important difference on the WOOS has never been validated. For the present study we used 190 (i.e. 10% of a maximum score), which was extrapolated from studies validating the minimal clinically important difference of other shoulder-specific outcome measures. Statistics WOOS scores were described using mean value and SD. A general linear model was used to analyze WOOS scores. Age (< 55 years, 55–75 years, > 75 years), sex, arthroplasty type, and year of surgery were included in the analyses. Interaction

Acta Orthopaedica 2019; 90 (5): 489–494

Table 1. Anatomical total shoulder arthroplasty (TSA), resurfacing hemiarthroplasty (RHA), stemmed hemiarthroplasty (SHA), and reverse shoulder arthroplasty (RSA), number (%) Sex Women Men Age ≥ 75 56–74 ≤ 55

TSA

RHA

SHA

RSA

543 (32) 271 (25)

691 (41) 567 (52)

297 (18) 203 (18)

156 (9) 57 (5)

175 (29) 590 (33) 50 (14)

194 (32) 811 (45) 253 (69)

132 (22) 306 (17) 62 (17)

102 (17) 107 (6) 4 (1)

The table does not include arthroplasties that were recorded as others (n = 68) or with a missing arthroplasty type (n = 13).

between year of surgery and arthroplasty type was used to study change during the studied period for each arthroplasty type. Estimates were given with 95% confidence intervals (CI). Due to the large sample size we considered it appropriate to use a parametric model. Linearity of year of surgery was investigated using a smoothing spline function. Patients with bilateral arthroplasty (n = 249) were included in the analysis as if the arthroplasties were independent. The analyses were performed using SPSS version 22.0 (IBM SPSS Statistics for Windows, Version 22.0. IBM Corp, Armonk, NY, USA). The level of significance was set at p < 0.05 and all p-values were 2-tailed. Ethics, funding, and potential conflicts of interest According to the regulations in Denmark, this study did not need permission from the National Committee on Health Research Ethics. No funding was received. There were no conflicts of interest to be declared related to this study.

Results Study population 2,867 primary arthroplasties for osteoarthritis were reported to the registry during the study period. There were 1,732 (60%) women. Mean age was 67 (SD 10) years. 13% patients were aged 55 years or younger, 65% patients were between 56 and 74 years, and 22% patients were 75 years or older. There were 1,258 resurfacing hemiarthroplasties, 500 stemmed hemiarthroplasties, 815 anatomical total shoulder arthroplasties, and 213 reverse shoulder arthroplasties. 68 arthroplasties were recorded as “others,” which included 21 stemless hemiarthroplasties and 47 stemless total shoulder arthroplasties. 13 cases were recorded with a missing arthroplasty type (Table 1). 249 patients had bilateral arthroplasties. Distribution of sex, age groups, and arthroplasty types over time The number of shoulder arthroplasties for osteoarthritis increased from 141 in 2006 to 393 in 2015 (Figure 1). In the

Acta Orthopaedica 2019; 90 (5): 489–494

491

Annual number of shoulder arthroplasties

Annual distribution (%)

Annual age distribution (%)

400

100

300

60 200

100

stemmed hemiarthroplasty resurfacing hemiarthroplasty reverse arthroplasty anatomic total arthroplasty

20 ≤ 55 years 56–74 years ≥ 75 years

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Year

Figure 1. Annual number of shoulder arthroplasties for osteoarthritis from 2006 to 2015.

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Year

Figure 2. The proportion of stemmed hemiarthroplasty (blue), resurfacing hemiarthroplasty (green), anatomical total shoulder arthroplasty (purple), and reverse shoulder arthroplasty (grey) from 2006 to 2015.

same period, the Danish population increased from 5,427,459 to 5,659,715 inhabitants. Thus, the number of arthroplasties increased by 179% whereas the population increased by 4%. 77% of all primary arthroplasties for osteoarthritis in 2006 were resurfacing hemiarthroplasties. Since then the proportion decreased to 3% in 2015. In the same period, the proportion of anatomical total shoulder arthroplasty and reverse shoulder arthroplasty increased from 3% to 53% and 7% to 27% respectively (Figure 2). The proportion of patients who were 55 years or younger decreased during the study period (Figure 3) Patient-reported outcome 51 (1.8%) patients died and 65 (2.3%) patients were revised within 1 year. Thus, the WOOS questionnaire was sent to 2,751 patients, of whom 69% returned a complete questionnaire. 6% of patients returned an incomplete questionnaire and 26% patients did not respond. The mean WOOS score was 59 (27) for patients who were 55 years or younger, 76 (24) for patients between 56 and 74 years, and 75 (25) for patients who were 75 years or older. The mean WOOS score was 73 (26) for women and 75 (24) for men. The mean WOOS score was 70 (26) after resurfacing hemiarthroplasties, 68 (26) after stemmed hemiarthroplasty, 82 (23) after anatomical total shoulder arthroplasties, and 74 (23) after reverse shoulder arthroplasties. In both the univariate and multivariable analysis, the outcome of anatomical total shoulder arthroplasty was better than that of any other arthroplasty type (Table 2). In the univariate analysis the overall WOOS increased by a mean of 1.8 (CI 1.2–2.2) each year or 18 (CI 12–22) from 2006 to 2015 (Table 1). To investigate whether changes in demographics had influenced the change in WOOS from 2006 to 2015 we included sex and age category in a multivariable analysis where the overall WOOS increased by a mean of

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Year

Figure 3. Proportion of patients who were 55 years or younger (dark green), between 56 and 74 years (light green), and 75 years or older (grey) from 2006 to 2015.

Table 2. Linear regression model with difference, 95% confidence interval (CI), and with WOOS at 1 year as the dependent variable Factor

Univariate model (95% CI)

Multivariable model (95% CI)

Year of operation (change per year) Arthroplasty type a TSA RHA SHA RSA Sex Women Men Age ≥ 75 56–74 ≤ 55

1.8 (1.2 to 2.2)

1.0 (0.5 to 1.5)

Reference –12 (–9.4 to –15) –14 (–10 to –17) –7.2 (–2.5 to –12)

Reference –6.6 (–3.5 to –9.8) –11 (–7.8 to –15) –7.0 (–2.4 to –12)

Reference 2.4 (0.1–4.8)

Reference 5.6 (3.3 to 8.0)

Reference 1.1 (–1.7 to 3.9) –16 (–12 to –20)

Reference 0.6 (–3.5 to 2.2) –18 (–13 to –23)

a For

abbreviations, see Table 1

17 (CI 13–21). This indicates that changes in demographics did not influence the overall WOOS score during the studied period. We looked for interaction between year of surgery and arthroplasty type and found improved WOOS scores for anatomical total shoulder arthroplasty (mean change: 14, CI 5–23), and resurfacing hemiarthroplasty (mean change: 12, CI 5–19), but not for reverse shoulder arthroplasty (mean change: 11, CI –7–29) and stemmed hemiarthroplasty (mean change: –2.3, CI –14–10). Sex and age category were included in the model. In a multivariable analysis, which included sex, age category, year of surgery, and arthroplasty type, the overall WOOS increased by a mean of 1.0 (CI 0.5–1.5) each year or

492

10 (CI 5â&#x20AC;&#x201C;15) during the study period (Table 1). This indicate that the improvement in WOOS was influenced by changes in the distribution of arthroplasty types. â&#x20AC;&#x192;

Discussion In this nationwide cohort study of patients with shoulder arthroplasty for osteoarthritis we found improved patientreported outcome from 2006 to 2015. During the same period, there was an increased use particularly of anatomical total shoulder arthroplasty, but also of reverse shoulder arthroplasty. The outcome of anatomical total shoulder arthroplasty was superior to any other arthroplasty type and increased during the study period. Changes in the use of arthroplasty types Information concerning the outcome of stemmed hemiarthroplasty and total shoulder arthroplasty was sparse until 2006. The first randomized trial was published in 2000 and included 27 total shoulder arthroplasties and 24 stemmed hemiarthroplasties, with mean ASES scores of 77 and 65 (Gartsman et al. 2000), respectively. Another randomized trial, which was published in 2005, included 20 total shoulder arthroplasties and 21 stemmed hemiarthroplasties. The mean WOOS score was 91 after total shoulder arthroplasty and 82 after hemiarthroplasty (Lo et al. 2005). Both studies were underpowered, and the differences were not statistically significant. A large observational study published in 2003 included 601 total shoulder arthroplasties and 89 stemmed hemiarthroplasties. With a mean follow-up of 3.5 years the mean Constant scores were 70 and 64, respectively. The difference was not statistically significant (Edwards et al. 2003). At that time, surgeons had concerns about rotator cuff problems and glenoid loosening after total shoulder arthroplasty (Bohsali et al. 2006). Thus, the inconclusive outcomes and the risk of glenoid loosening may explain why there were less than 20 total shoulder arthroplasties reported to the DSR each year in the beginning of the study period. In 2004, Levy and Copeland (2004) published their results of resurfacing arthroplasty for osteoarthritis. They included 39 total resurfacing arthroplasties and 30 resurfacing hemiarthroplasties. The postoperative Constant scores were 62 and 58, respectively. To avoid late complications with glenoid loosening the authors recommended the resurfacing hemiarthroplasty for osteoarthritis except in patients with nonconcentric or saddle-shaped erosion of the glenoid. They concluded that the results were at least equal to those of stemmed shoulder arthroplasty and that the resurfacing arthroplasty had the advantage of a bone-preserving design, short operation time, and an easy revision, should the need for revision arthroplasty arise. A systematic review from 2009 supported their results and concluded that resurfacing hemiarthroplasty is a viable option for shoulder replacement, especially in young patients (Burgess et al. 2009).

Acta Orthopaedica 2019; 90 (5): 489â&#x20AC;&#x201C;494

It is, of course, speculative but these studies are probably the main reason for the high proportion of resurfacing hemiarthroplasties in the first half of our study period. The results of arthroplasty types were published in the annual reports from the DSR, presented at the annual meeting of the Danish Society for Shoulder and Elbow surgery in 2011, and later published. The results showed unpredictable patient-reported outcomes of the resurfacing hemiarthroplasty with a high proportion of disappointing results and a high rate of revision, especially in young patients (Rasmussen et al. 2014b), and the patient-reported outcome was inferior to that of total shoulder arthroplasty (Rasmussen et al. 2014a). Furthermore, data from the registry showed poor patient-reported outcomes of revision arthroplasty after failed resurfacing hemiarthroplasty, belying the hypothesis of an easy revision (Rasmussen et al. 2016). A Cochrane review from 2010 included the 2 randomized trials by Gartsman et al. (2000) and Lo et al. (2005) with 88 arthroplasties and found a statistically significant superior ASES score after anatomical total shoulder arthroplasty compared with stemmed hemiarthroplasty (Singh et al. 2011). In 2011 the American Academy of Orthopedic Surgeons (AAOS) published its guidelines regarding the treatment of glenohumeral osteoarthritis and recommended anatomical total shoulder arthroplasty in patients with an intact rotator cuff (Izquierdo et al. 2011). Furthermore, in publications from 2012 and 2013 the results of the resurfacing hemiarthroplasty were questioned by the authors of small case series from independent centers (Al-Hadithy et al. 2012, Mechlenburg et al. 2013, Smith et al. 2013) and by data from the Norwegian Arthroplasty Register (Fevang et al. 2013). The reason for the changed surgical practice in Denmark is speculative. It may be related to international trends for using total shoulder arthroplasty and reverse shoulder arthroplasty for osteoarthritis. This trend is, however, not supported by randomized trials or other strong evidence and we believed that data and publications from the DSR and from the national arthroplasty registries in Australia, Norway, Sweden, and New Zealand have contributed to the changed practice among Danish surgeons. Thus, our results indicate that a national registry can improve the outcome of shoulder arthroplasty by monitoring and reporting the outcome of arthroplasty types with inferior outcomes. Outcome of arthroplasty types The results of arthroplasty types may not be directly comparable. The reverse shoulder arthroplasty has probably been used in patients with rotator cuff pathology or posterior subluxation. The excellent results of anatomical total shoulder arthroplasty might not have been the same if it had been used in the same patients. There are no existing randomized trials comparing the anatomical total shoulder arthroplasty and the reverse shoulder arthroplasty in patients with osteoarthritis and intact rotator cuff function. In our study, the outcome of

Acta Orthopaedica 2019; 90 (5): 489–494

anatomical total shoulder arthroplasty was better than reverse shoulder arthroplasty. This indicates that the anatomical total shoulder arthroplasty should be preferred in patients with an intact rotator cuff and no posterior subluxation, but the findings need to be confirmed by a randomized trial. The results of the anatomical total shoulder arthroplasty were superior to stemmed hemiarthroplasty and resurfacing hemiarthroplasty. This support the guidelines from the AAOS (Izquierdo et al. 2011) recommending anatomical total shoulder arthroplasty in patients with glenohumeral osteoarthritis and intact rotator cuff function. However, it is important to stress the risk of selection bias. Based on the existing literature (Izquierdo et al. 2011) and our results we consider the anatomical total shoulder arthroplasty as the gold standard in patients with an intact and well-functioning rotator cuff and without severe posterior glenoid wear. The reverse shoulder arthroplasty is reserved for patients with symptomatic rotator cuff or posterior subluxation, and the resurfacing hemiarthroplasty should be avoided. The stemmed hemiarthroplasty can be indicated in cases with severe glenoid bone loss where fixation of a glenoid component is not possible. Improved outcome from 2006 to 2015 Very few national arthroplasty registries have collected patient-reported outcome over a long period of time. So, to our knowledge this is the first study to report nationwide changes in patient-reported outcome after shoulder arthroplasty. We found a statistically and clinically significant improvement in WOOS of 18 from 2006 to 2015. The improvement was related to changes in the distribution of arthroplasty types during the studied period with a higher proportion in particular of anatomical total shoulder arthroplasty towards the end of the studied period. Changes in the distribution of sex and age group during the studied period had little influence on the improvement. We also found improved WOOS scores for anatomical total shoulder arthroplasty and resurfacing hemiarthroplasty during the studied period. Thus, the improvement in WOOS from 2006 to 2015 was primarily related to changes in the distribution of arthroplasty types and improved outcome of anatomical total shoulder arthroplasty and resurfacing hemiarthroplasty. The reason for the improved outcome of anatomical total shoulder arthroplasty is speculative and cannot be deducted from our study. The reason for the improved outcome of resurfacing hemiarthroplasty is most likely related to the increased use of reverse shoulder arthroplasty and focus on the preoperative rotator cuff function. Better treatment selection, where the reverse shoulder arthroplasty is used in patients with rotator cuff insufficiency, has properly reduced the number of resurfacing hemiarthroplasties that fail because of rotator cuff insufficiency, and thereby improved the outcome of the few resurfacing hemiarthroplasties that were used towards the end of the studied period. It is important to stress that there might

493

have been other reasons for the improved outcome which are not accounted for in the analysis. Limitations The indications for surgery and for a specific arthroplasty type were not clearly defined. Thus, the risk of selection bias is important to keep in mind when the results are interpreted. Differences in preoperative WOOS scores between the arthroplasty types or between the year of operation might have influenced the comparison at 1 year. Not all patients returned a complete WOOS questionnaire, and any systematic differences in demographics or WOOS score between responders and non-responders would have influenced the results and interpretation. Finally, incorrect reporting may diminish the accuracy and reliability of the data. Conclusions The patient-reported outcome of shoulder arthroplasty for osteoarthritis improved from 2006 to 2015. This may be related to different factors: improved outcome of anatomical total shoulder arthroplasty; the increased use of total shoulder arthroplasty towards the end of the study period; and better treatment selection, including the use of reverse shoulder arthroplasty in patients with poor rotator cuff function. The reason for the increased use of total shoulder arthroplasty is unknown but may be related to surgeons’ awareness of clinical results through annual reports from the DSR and other national arthroplasty registries. We recommend continued nationwide surveillance regarding the use of shoulder arthroplasty types.

All authors participated in the conception and design of the study, and in interpretation of the results. AA, SLJ, and JJ collected data. JVR and TWK performed the statistical analysis. AKS, SLJ, BSO, and JVR participated in the preparation of the manuscript. The authors would like to thank the orthopedic surgeons in Denmark for data reporting. Acta thanks Randi Hole and Richard Page for help with peer review of this study.

Al-Hadithy N, Domos P, Sewell M D, Naleem A, Papanna M C, Pandit R. Cementless surface replacement arthroplasty of the shoulder for osteoarthritis: results of fifty Mark III Copeland prosthesis from an independent center with four-year mean follow-up. J Shoulder Elbow Surg 2012; 21(12): 1776-81. Bohsali K I, Wirth M A, Rockwood C A, Jr. Complications of total shoulder arthroplasty. J Bone Joint Surg Am 2006; 88(10): 2279-92. Burgess D L, McGrath M S, Bonutti P M, Marker D R, Delanois R E, Mont M A. Shoulder resurfacing. J Bone Joint Surg Am 2009; 91(5): 1228-38. Edwards T B, Kadakia N R, Boulahia A, Kempf J F, Boileau P, Nemoz C, Walch G. A comparison of hemiarthroplasty and total shoulder arthroplasty in the treatment of primary glenohumeral osteoarthritis: results of a multicenter study. J Shoulder Elbow Surg 2003; 12(3): 207-13.

494

Fevang B T, Lygre S H, Bertelsen G, Skredderstuen A, Havelin L I, Furnes O. Pain and function in eight hundred and fifty nine patients comparing shoulder hemiprostheses, resurfacing prostheses, reversed total and conventional total prostheses. Int Orthop 2013; 37(1): 59-66. Gartsman G M, Roddey T S, Hammerman S M. Shoulder arthroplasty with or without resurfacing of the glenoid in patients who have osteoarthritis. J Bone Joint Surg Am 2000; 82(1): 26-34. Izquierdo R, Voloshin I, Edwards S, Freehill M Q, Stanwood W, Wiater J M, Watters W C, III, Goldberg M J, Keith M, Turkelson C M, Wies J L, Anderson S, Boyer K, Raymond L, Sluka P, Hitchcock K. American Academy of Orthopedic Surgeons clinical practice guideline on: the treatment of glenohumeral joint osteoarthritis. J Bone Joint Surg Am 2011; 93(2): 203-5. Levy O, Copeland S A. Cementless surface replacement arthroplasty (Copeland CSRA) for osteoarthritis of the shoulder. J Shoulder Elbow Surg 2004; 13(3): 266-71. Lo I K, Griffin S, Kirkley A. The development of a disease-specific quality of life measurement tool for osteoarthritis of the shoulder: the Western Ontario Osteoarthritis of the Shoulder (WOOS) index. Osteoarthritis Cartilage 2001; 9(8): 771-8. Lo I K, Litchfield R B, Griffin S, Faber K, Patterson S D, Kirkley A. Qualityof-life outcome following hemiarthroplasty or total shoulder arthroplasty in patients with osteoarthritis: a prospective, randomized trial. J Bone Joint Surg Am 2005; 87(10): 2178-85. Mechlenburg I, Amstrup A, Klebe T, Jacobsen S S, Teichert G, Stilling M. The Copeland resurfacing humeral head implant does not restore humeral

Acta Orthopaedica 2019; 90 (5): 489â&#x20AC;&#x201C;494

head anatomy: a retrospective study. Arch Orthop Trauma Surg 2013; 133(5): 615-19. Rasmussen J V, Jakobsen J, Brorson S, Olsen B S. The Danish Shoulder Arthroplasty Registry: clinical outcome and short-term survival of 2,137 primary shoulder replacements. Acta Orthop 2012; 83(2): 171-3. Rasmussen J V, Jakobsen J, Olsen B S, Brorson S. Translation and validation of the Western Ontario Osteoarthritis of the Shoulder (WOOS) index: the Danish version. Patient Relat Outcome Meas 2013; 4: 49-54. Rasmussen J V, Polk A, Brorson S, Sorensen A K, Olsen B S. Patient-reported outcome and risk of revision after shoulder replacement for osteoarthritis. 1,209 cases from the Danish Shoulder Arthroplasty Registry, 2006â&#x20AC;&#x201C;2010. Acta Orthop 2014a; 85(2): 117-22. Rasmussen J V, Polk A, Sorensen A K, Olsen B S, Brorson S. Outcome, revision rate and indication for revision following resurfacing hemiarthroplasty for osteoarthritis of the shoulder: 837 operations reported to the Danish Shoulder Arthroplasty Registry. Bone Joint J 2014b; 96-B(4): 519-25. Rasmussen J V, Olsen B S, Al-Hamdani A, Brorson S. Outcome of revision shoulder arthroplasty after resurfacing hemiarthroplasty in patients with glenohumeral osteoarthritis. J Bone Joint Surg Am 2016; 98(19): 1631-7. Singh J A, Sperling J, Buchbinder R, McMaken K. Surgery for shoulder osteoarthritis: a Cochrane systematic review. J Rheumatol 2011; 38(4): 598-605. Smith T, Gettmann A, Wellmann M, Pastor F, Struck M. Humeral surface replacement for osteoarthritis. Acta Orthop 2013; 84(5): 468-72.

Acta Orthopaedica 2019; 90 (5): 495–500

495

Primary surgery to prevent hip dislocation in children with cerebral palsy in Sweden: a minimum 5-year follow-up by the national surveillance program (CPUP) Nikolaos KIAPEKOS 1,2, Eva BROSTRÖM 1,2, Gunnar HÄGGLUND 3, and Per ÅSTRAND 1,2 1 Department

of Women’s and Children’s Health, Karolinska Institute, Karolinska University Hospital, Stockholm; 2 Department of Pediatric Orthopedics, Astrid Lindgren’s Children Hospital, Karolinska University Hospital, Stockholm; 3 Lund University, Department of Clinical Sciences, Lund, Orthopedics, Lund, Sweden Correspondence: nikolaos.kiapekos@sll.se Submitted 2018-01-10. Accepted 2019-03-14.

Background and purpose — Children with cerebral palsy (CP) have an increased risk of hip dislocation. Outcome studies after surgery to prevent hip dislocation in children with CP are usually retrospective series from single tertiary referral centers. According to the national CP surveillance program in Sweden (CPUP), hip surgery should preferably be performed at an early age to prevent hip dislocation. Preventive operations are performed in 12 different Swedish hospitals. We compared the outcomes between soft tissue release and femoral osteotomy in children with CP treated in these hospitals. Patients and methods — 186 children with CP underwent either adductor–iliopsoas tenotomy (APT) or femoral osteotomy (FO) as the primary, preventive surgery because of hip displacement. They were followed for a minimum of 5 years (mean 8 years) regarding revision surgery and hip migration. A good outcome was defined as the absence of revision surgery and a migration percentage (MP) < 50% at the latest follow-up. Logistic and Cox regression analysis were used to investigate the influence of age, sex, preoperative MP, Gross Motor Function Classification System (GMFCS) level, and CP subtype. Results — APT was performed in 129 (69%) children. After 5 years, the reoperation rate was 43%, and 2 children (2%) had an MP > 50%. For the 57 children who underwent FO, the corresponding figures were 39% and 9%. Of the potential risk factors studied, the outcome was statistically significantly associated with preoperative MP only in children who underwent APT, but not in those who underwent FO. None of the other factors were significantly associated with the outcome in the 2 procedure groups. Interpretation — Reoperation rates after preventive surgery are high and indicate the importance of continued postoperative follow-up. Age, sex, GMFCS level, and CP subtype did not influence the outcome significantly.

In children with cerebral palsy (CP), the risk of hip displacement increases with decreasing gross motor function and may approach 90% in the most-affected children (Morton et al. 2006, Soo et al. 2006, Hägglund et al. 2007). Surveillance programs can be effective in identifying children with an increased risk of hip dislocation (Gordon and Simkiss 2006, Hägglund et al. 2014), but there are no current recommendations for the timing of a specific type of surgical procedure (Stott and Piedrahita 2004, Bouwhuis et al. 2015). In Sweden, the national surveillance program for children with CP (CPUP) includes almost all children (> 95%) with CP born in 2002 or later. In the CPUP, children with CP undergo regular radiographic examinations to detect early hip displacement, and preventive surgery is recommended when the migration percentage (MP) exceeds 40%. However, the CPUP makes no specific recommendations about when adductor– iliopsoas tenotomy (APT) or femoral osteotomy (FO) should be performed. Previous studies have shown that soft tissue release (e.g., APT) is more effective in children capable of walking and when the MP is < 50% (Turker and Lee 2000, Presedo et al. 2005, Shore et al. 2012, Terjesen 2017). Soft tissue release is commonly used in younger children and its use implies shorter operation time with fewer complications as well as a shorter and easier postoperative rehabilitation. Osseous procedures are usually used in older children and in children with more severe displacement (MP > 50%), but they are associated with longer rehabilitation and more complications (Bouwhuis et al. 2015). Generally, outcome studies after hip surgery in children with CP have been reported by tertiary referral centers. A recent report noted that the surgeon’s experience may have a major influence on the outcome after FO (Shore et al. 2015). In Sweden, surgical procedures to prevent hip dislocation in children with CP are performed in 7 university hospitals and 5 regional hospitals, all of which were included in this study.

496

The aim of this population-based study was to compare the outcomes between 2 different surgical procedures (APT or FO) used for prevention of hip dislocation. Outcome was assessed by reoperation rates and extent of hip migration. Logistic as well as Cox regression analysis was used to investigate the influence of age at the time of surgery, sex, preoperative MP, Gross Motor Function Classification System (GMFCS) level, and CP subtype.

Patients and methods This was a population-based register study of children with CP who were followed in the Swedish national quality register CPUP. It is estimated that the register includes more than 95% of the children with CP in Sweden (Westbom et al. 2007). At the end of the treatment period in this study (2012), 2,948 children up to 18 years of age were registered in the CPUP, and 1,238 of them were classified as GMFCS levels III–V. At that time, 13 children (8 girls) had an MP = 100% but were considered too weak for surgery (CPUP annual report 2013, ISBN 978-91-980722-0-4). The CPUP includes continuous standardized follow-up of gross and fine motor function, clinical findings and treatment. The CP diagnosis and subtype are reported by the child’s neuropediatrician. The dominant symptom is also classified by the child’s physiotherapist as spastic, dyskinetic, or mixed. 4 children classified with mixed symptoms and 1 child with ataxia were omitted from the statistical analyses based on CP subtype. The physiotherapist also classifies each child’s gross motor function according to the GMFCS (Palisano et al. 1997). A local coordinator is responsible for reporting the orthopedic surgical procedures to CPUP, using a recording form specifying the different types of procedures. The types of pre- and postoperative treatment were not investigated in this study. Anteroposterior radiographs of the hips in the supine position are acquired at the child’s inclusion in the register and then at least once a year until 8 years of age in children with GMFCS III–V. After that, the follow-up is continued on an individual basis. The MP is calculated according to Reimers’ index (1980). This study included all children in the CPUP register operated on primarily for hip displacement with either bilateral APT or uni- or bilateral FO with or without concurrent pelvic osteotomy (PO), with a minimum of 5 years of follow-up, i.e., children treated during the period 1995–2012. The criteria for failure were either a second hip operation (on any side) or MP > 50% at the latest radiological followup. The treatment of failure was further divided into 3 categories: (1) reoperation using soft tissue hip procedures (APT or adductor tenotomies), (2) reoperation using bony procedures (FO with or without PO), and (3) no reoperation but an MP > 50% in at least 1 hip. For children who were operated on with unilateral FO the side of the reoperation (ipsilateral or contra-

Acta Orthopaedica 2019; 90 (5): 495–500

lateral) was also specified. Removal of osteosynthesis material was not considered as revision surgery. In the statistical analysis, the influence of surgical volume on the outcome after FO was studied by identifying hospitals with a “high surgical volume unit” which was defined as the 3 larger university clinics that treated 37 of the 57 children who underwent FO. The inclusion criteria were: (1) confirmed CP diagnosis, (2) bilateral APT or uni- or bilateral FO as the index surgery, (3) MP >30% in at least 1 hip preoperatively before the first operation, and (4) a minimum 5 years of follow-up after the index surgery. Children undergoing unilateral (2 children) or bilateral adductor surgery without iliopsoas lengthening or tenotomy (7 children) were thus excluded. The rationale for including only children who underwent bilateral APT was to diminish variations in the surgical “dosage” (e.g., percutaneous adductor tenotomies). A total of 192 children fulfilled the inclusion criteria. 2 children moved abroad and 4 children died before the 5-year follow-up, and were therefore excluded. The remaining 186 children (79 girls) were followed for a minimum of 5 years and had a mean follow-up time of 8 (5–20) years. Statistics Continuous data are presented as mean (SD), and categorical data are presented as frequency count and percentage. When comparing the groups of children undergoing APT or FO, continuous data were analyzed using Student’s t-test or the Mann–Whitney U-test. The chi-square test was used to analyze categorical data. A p-value < 0.05 was considered to be significant. IBM SPSS Statistics (version 23; IBM Corp, Armonk, NY, USA) was used for these analyses. Logistic regression analysis models were constructed with failure as the outcome and age at the time of surgery, sex, preoperative MP, GMFCS level, and CP subtype as covariates. Preoperative MP was scaled in 5% increments. Hazard ratios using Cox regression analysis were also calculated using the same predictors. In 1 logistic regression model of the entire material, the type of surgery (APT vs. FO) was included as a covariate. In another logistic regression model to study the outcome after FO, the covariate “high surgical volume unit” was also included. These analyses were performed using STATA 13 software (Stata Corp, College Station, TX, USA). Kaplan– Meier curves are presented to show the cumulative proportion of failures over the follow-up period. Ethics, funding, and conflicts of interest Ethical approval was obtained from the Regional Ethical Review Board in Lund, Sweden (LU-443-99). The support of Stiftelsen Promobilia, Norrbacka-Eugeniastiftelsen, Josef och Linnéa Carlssons stiftelse, Stiftelsen för bistånd åt rörelsehindrade i Skåne, and Riksförbundet för Rörelsehindrade Barn och Ungdomar is gratefully acknowledged. The authors declare no conflicts of interest.

Acta Orthopaedica 2019; 90 (5): 495–500

Table 1. Characteristics of the children operated on with adductor–iliopsoas tenotomy (APT) or femoral osteotomy (FO). Values are frequency (%) unless otherwise specified Factor

APT FO n = 129 n = 57 p-value

Mean age (SD) at primary operation 4.9 (2.2) 5.6 (2.3) 0.03 at second operation 6.8 (2.6) 8.7 (2.5) 0.003 Girls 56 (43) 23 (40) 0.8 CP subtype 0.4 spastic 79 (61) 38 (67) dyskinetic 47 (36) 17 (30) GMFCS 0.3 I + II + III 13 (10) 7 (12) IV 37 (29) 10 (18) V 79 (61) 40 (70) Mean preoperative MP (SD) a 47 (12) 58 (16) 0.001 Reoperations at 5 years total 56 (43) 22 (39) 0.5 soft tissue reoperations 1 (1) 5 (9) 0.005 bony reoperations 55 (43) 17 (30) 0.1 MP > 50% at 5 years 2 (2) 5 (9) 0.2 Failures at 5 years 58 (45) 27 (47) 0.8 Failures at 5–20 years 62 (48) 30 (53) 0.6 a MP

worst hip GMFCS = Gross Motor Function Classification System level MP = migration percentage

Results Of the 186 children, 129 (69%) underwent bilateral APT and 57 (31%) FO (34 unilateral, 23 bilateral, with or without simultaneous PO) as their primary procedure. Almost 90% of the children operated on were in GMFCS level IV or V (Table 1). To allow for a meaningful statistical analysis, the small numbers of children in GMFCS levels I–III were pooled. The children who were operated on with APT were younger than those who were operated on with FO as the index operation (p = 0.03) as well as at the second surgery, and they had a lower mean preoperative MP in the worst hip. 6 (3%) children had APT as the second surgery. There were more soft tissue reoperations after FO than after APT, but GMFCS distribution, sex, CP subtype, other reoperation rates, and failure rates did not differ significantly between groups (Table 1). A logistic regression model was created in which age, sex, CP subtype, GMFCS level, and MP were adjusted, and the type of operation (APT or FO) was included as a covariate. APT was selected as the reference. No statistically significant association was found between the type of operation and failure (odds ratio = 0.6, 95% CI = 0.3–1.5). The only statistically significant predictor found was the preoperative MP (odds ratio = 1.3, 95% CI = 1.1–1.6). Of the 129 children who were operated on with an APT, 56 (43%) were reoperated on. All of these 56 children were operated on with FO (in combination with PO in 15 children) within 5 years, and 2 children (2%) had an MP > 50% at the

497

Table 2. Statistical analyses of potential risk factors for failure after adductor–psoas tenotomy Variable

Odds ratio (95% CI)

Hazard ratio (95% CI)

Age at surgery 0.9 (0.8–1.1) 0.9 (0.8–1.0) Boy Reference Reference Girl 0.6 (0.3–1.4) 0.7 (0.4–1.1) CP subtype, spastic Reference Reference dyskinetic 1.5 (0.6–3.5) 1.1 (0.6–1.9) GMFCS I + II + III Reference Reference IV 2.1 (0.4–10.4) 1.7 (0.5–6.1) V 2.7 (0.6–12.4) 2.1 (0.6–7) MP, 5% increase 1.5 (1.1–1.8) 1.2 (1.1–1.3) For abbreviations, see Table 1.

5-year follow-up. In 1 child (1%), a repeat APT was performed 6 months before the FO. The other 55 children (43%) were reoperated on with FO with or without simultaneous PO. The revision surgery was performed on average 2.3 years (0.5–7) after the index operation. 1 child received a reoperation after more than 5 years. The highest preoperative MP that improved after APT and was not associated with the need for a reoperation was 56%. In the logistic regression analysis, only preoperative MP was statistically significant and the other factors studied (age at the time of surgery, sex, GMFCS level, and CP subtype) were not significant predictors of failure (Table 2). In the Cox regression analysis of the hazard ratios, only preoperative MP was a statistically significant predictor of failure (Table 2). In a separate logistic regression model in which children with GMFCS IV and V were pooled, the preoperative MP was the only statistically significant predictor for failure (data not shown). Kaplan–Meier survivorship plots showed a higher failure rate for children in GMFCS IV–V as compared with those with GMFCS I–III (Figure 1), although the difference was not statistically significant. For children treated with APT, the outcomes of 50% of the children in GMFCS V were classified as a failure before the children were 5 years of age (Figure 1). Of the 57 children who had an FO (in 80 hips) as their primary surgery, 22 children (24 hips, 30%) were reoperated on within 5 years and 2 children after 5 years. The distributions of soft tissue and bony reoperations are given in Table 1. Revision surgery was performed on average 3.6 years (0.8–7) after the index surgery. At the 5-year follow-up, 5 children (9%) had an MP > 50% (Table 1), 23 children (40%) were operated on with a bilateral FO on the first occasion, and 15 children (26%) were operated on with an FO combined with PO (Table 3). The outcome did not differ significantly between groups. Of the 24 children who received reoperations after FO without PO as primary surgery, 9 were operated on with FO only, 9 with FO and PO, 1 with PO only, and 5 with APT as the second surgery. 9 of the 34 children with unilateral FO as their index operation had an ipsilateral reoperation, and 3 children had revision surgery on the contralateral side. 1 child devel-

498

Acta Orthopaedica 2019; 90 (5): 495–500

K–M cumulative failure rate

1.0

1.0 I–III IV–V

0.8

0.6

0.4

0.2

0.0 0

0.0

Years after index operation

I–III IV–V

Years after index operation

Number at risk Year

IV V

Number at risk

0 2 4 6 8 10

Year

11 10 10 6 2 0 111 91 64 41 27 13

IV V

0 2 4 6 8 10 10 10 9 6 1 1 36 31 23 12 2 0

Figure 1. Kaplan–Meier curves showing the proportion of failures with time after the index operation, bilateral adductor–psoas tenotomy, grouped according to the Gross Motor Function Classification System (GMFCS). GMFCS I–III and GMFCS IV–V are pooled for comparison of children who rely on a wheelchair for transport (GMFCS IV–V) with children more capable of walking (GMFCS I–III). Shaded areas represent 95% confidence intervals.

Figure 2. Kaplan–Meier curves showing the proportion of failures with time after the index operation, femoral osteotomy, for all patients. Gross Motor Function Classification System (GMFCS) levels I–III are not shown because of the small numbers of patients. Shaded areas represent 95% confidence intervals. The vertical end of the red curve indicates that the last patient followed was reoperated.

Table 3. Distribution of the 57 children operated on with femoral osteotomy

Table 4. Statistical analyses of potential risk factors for failure after femoral osteotomy

Factor

n Reoperations MP > 50% p-value

Unilateral surgery Bilateral surgery

34 23

17 7

3 2

0.6

Femoral osteotomy (FO) 42 FO + pelvic osteotomy 15

17 7

3 2

0.8

MP = migration percentage

oped contralateral MP > 50%. The latter 4 children were not classified as failures, when the failure rate was calculated per hip and not per patient. Of the 15 children who had combined FO and PO, 4 were operated on with an APT as their reoperation. Similar to the analysis of the outcome of APT, both logistic and Cox regression analyses were performed for the outcome of FO. In this model, we also included as a covariate “high surgical volume unit,” which represented the 3 university clinics that contributed 37 of the 57 patients who had FO as the index surgery. The confidence intervals are large because of the small sample size (Table 4). The Kaplan–Meier plot of the outcome for children undergoing FO showed a tendency for increasing differences in

Variable

Odds ratio (95% CI)

Hazard ratio (95% CI)

Age at surgery 0.6 (0.4–1.0) 1 (0.8–1.3) Boys Reference Reference Girls 0.1 (0.02–1.0) 0.3 (0.1–1.0) CP subtype, spastic Reference Reference dyskinetic 10.3 (1.1–92) 2.6 (0.9–7.6) GMFCS I + II + III Reference Reference IV–V 8.0 (0.8–79) 2.6 (0.5–14) MP, 5% increase 1.3 (0.9–1.7) 1.1 (0.9–1.3) High surgical volume unit 0.1 (0.01–0.7) 0.6 (0.2–2.1) For abbreviations, see Table 1.

reoperation rates between children with GMFCS levels IV or V with increasing follow-up time, but the number of patients followed for more than 5 years was very low (Figure 2).

Discussion Our study is based on data from the CPUP Swedish national quality register, which includes > 95% of all children with CP (Westbom et al. 2007). To our knowledge, this is the first

Acta Orthopaedica 2019; 90 (5): 495–500

national-level, population-based study describing the outcomes after primary hip surgery (APT or FO) to prevent hip dislocation in children with CP. We found similar failure rates for both methods (45% and 47%, respectively, Table 1) after 5 years of follow-up. Generally, the outcome after surgery for hip displacement in children with CP is influenced by patient selection, duration of follow-up, type(s) of procedure, and varying definitions of success and failure. A shorter follow-up is usually associated with higher success rates. Comparing outcome after FO is complicated by the facts that the reoperation or success rate is often reported per hip and not per patient, and the percentages of hips/patients who undergo PO and FO at the same time vary between studies. It seems likely that the percentage of patients who undergo bilateral FO may also influence the success rate per hip. Choosing contralateral surgery more often for reasons of symmetry in seating and standing rather than increased hip migration may increase the success rate per hip. APT was performed as the primary procedure in 69% of the children, and as the second procedure in 3%. To facilitate comparisons, we followed the method of Shore et al. (2012), who defined failure as either a reoperation or an MP > 50%. Compared with other studies on children in GMFCS IV–V with sufficiently long follow-up, our failure rate of 48% is higher than the approximately 40% reported by Presedo et al. (2005) and Terjesen (2017), but lower than the 58% reported by Turker and Lee (2000) and 75% by Shore et al. (2012). In our study, the only statistically significant outcome predictor after APT was the preoperative MP, which is consistent with several previous publications (Onimus et al. 1991, Miller et al. 1997, Turker and Lee 2000, Stott and Piedrahita 2004, Terjesen et al. 2005, Terjesen 2017), but contrasts with the study by Shore et al. (2012), who found that GMFCS level was the strongest predictor. However, in that study, the children in GMFCS III, IV, and V were compared with a larger group of children in GMFCS II, who had a very high success rate after APT (94%). The single highest preoperative MP that did not result in failure after APT was 56%. The logistic and Cox regression analyses showed that an increase in preoperative MP was significantly associated with an increased risk for failure after APT. Even though we did not try to identify a specific threshold MP value for increased risk of failure, we found no reason to change to the commonly recommended threshold value of MP < 50% for considering APT. Recent reviews indicate that combined FO and PO procedures may have a lower failure rate than FO only (Bouwhuis et al. 2015, El-Sobky et al. 2018). However, in studies that included FO and PO only, the failure rates varied widely from 2% (Rutz et al. 2015) to 47% (Shea et al. 1997). In studies of FO with > 10 years of follow-up, the failure rates range from 20% (Oh et al. 2007) to 44 % (Canavese et al. 2010) of hips. The percentage of hips treated with bilateral FO in the material may influence the failure rate per hip. For example,

499

Oh et al. (2007) reported a 20% reoperation rate in 61 hips in 31 children, whereas Canavese et al. (2010) found a 44% reoperation rate after unilateral FO only. In our study, 80 hips were operated on in 57 children, and the reoperation rate per hip was 30% at 5 years. Using the same criteria for failure per patient as for APT above, Shore et al. (2015) reported a failure rate of 37%, which is lower than the 53% in our study. In their study of 567 hips in 320 patients undergoing FO, the strongest predictors were age, GMFCS level, and surgeon volume. The higher failure rate in our study may be explained by the higher percentage of children in GMFCS IV–V (88% vs. 64%) and the lower mean age (5.6 vs. 8.2 years). Shore et al. speculated that their lower failure rates may reflect the high percentage of children who were operated on with combined FO and PO (50%) in their material, which was 26% in our study. The analysis of the association between FO and outcome predictors was limited by the small patient sample and large confidence intervals (Table 4). In 1 logistic regression model, we studied the effect of surgeon volume on the outcome by including “high surgical volume unit” as a covariate. Although there was a tendency for lower failure rates in these units, the large confidence intervals do not allow us to draw conclusions. The literature has few studies similar to ours that compared outcome after soft tissue procedures and skeletal reconstruction to prevent hip dislocation in children with CP. However, in another study of a younger group than ours (mean age at the time surgery of 3.9 years, mean follow-up 5 years), the reoperation rate per patient was 77% after soft tissue release and 74% after FO (Schmale et al. 2006). We found no statistically significant association between the type of operation (APT or FO) and failure rates after adjusting for age and MP. The failure rates after FO are within the ranges reported in similar studies, and it is possible that these numbers may improve by centralizing FO surgery to only a few hospitals. It seems appropriate to continue to regard APT as a valuable surgical option for young children with CP with early hip displacement. Preferably, the operation should be performed while the MP is < 50%. For APT as well as FO, the parents should be informed that there is a substantial risk of the need for reoperation within the following few years. In conclusion, surgical prevention of hip dislocation in children with CP is associated with a high reoperation rate, both after soft tissue release and after reconstructive osteotomy. Continued postoperative hip surveillance is important.

All authors designed the study. NK collected the data and wrote the first draft, which was then improved and revised by all authors. The authors would like to thank Tomasz Czuba and Elisabet Berg for statistical consultations. Acta thanks Tanja Kraus and Terje Terjesen for help with peer review of this study.

500

Bouwhuis C B, van der Heijden-Maessen H C, Boldingh E J, Bos C F, Lankhorst G J. Effectiveness of preventive and corrective surgical intervention on hip disorders in severe cerebral palsy: a systematic review. Disabil Rehabil 2015; 37(2): 97-105. Canavese F, Emara K, Sembrano J N, Bialik V, Aiona M D, Sussman M D. Varus derotation osteotomy for the treatment of hip subluxation and dislocation in GMFCS level III to V patients with unilateral hip involvement: follow-up at skeletal maturity. J Pediatr Orthop 2010; 30(4): 357-64. El-Sobky T A, Fayyad T A, Kotb A M, Kaldas B. Bony reconstruction of hip in cerebral palsy children Gross Motor Function Classification System levels III to V: a systematic review. J Pediatr Orthop B 2018; 27(3): 221-30. Gordon G S, Simkiss D E. A systematic review of the evidence for hip surveillance in children with cerebral palsy. J Bone Joint Surg Br 2006; 88(11): 1492-6. Hägglund G, Alriksson-Schmidt A, Lauge-Pedersen H, Rodby-Bousquet E, Wagner P, Westbom L. Prevention of dislocation of the hip in children with cerebral palsy: 20-year results of a population-based prevention programme. Bone Joint J 2014; 96-b(11): 1546-52. Hägglund G, Lauge-Pedersen H, Wagner P. Characteristics of children with hip displacement in cerebral palsy. BMC Musculoskelet Disord 2007; 8: 101. Miller F, Cardoso Dias R, Dabney K W, Lipton G E, Triana M. Soft-tissue release for spastic hip subluxation in cerebral palsy. J Pediatr Orthop 1997; 17(5): 571-84. Morton R E, Scott B, McClelland V, Henry A. Dislocation of the hips in children with bilateral spastic cerebral palsy, 1985–2000. Dev Med Child Neurol 2006; 48(7): 555-8. Oh C W, Presedo A, Dabney K W, Miller F. Factors affecting femoral varus osteotomy in cerebral palsy: a long-term result over 10 years. J Pediatr Orthop B 2007; 16(1): 23-30. Onimus M, Allamel G, Manzone P, Laurain J M. Prevention of hip dislocation in cerebral palsy by early psoas and adductors tenotomies. J Pediatr Orthop 1991; 11(4): 432-5. Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi B. Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dev Med Child Neurol 1997; 39(4): 214-23. Presedo A, Oh C W, Dabney K W, Miller F. Soft-tissue releases to treat spastic hip subluxation in children with cerebral palsy. J Bone Joint Surg Am 2005; 87(4): 832-41.

Acta Orthopaedica 2019; 90 (5): 495–500

Reimers J. The stability of the hip in children. A radiological study of the results of muscle surgery in cerebral palsy. Acta Orthop Scand 1980; 184(Suppl.): 1-100. Rutz E, Vavken P, Camathias C, Haase C, Junemann S, Brunner R. Long-term results and outcome predictors in one-stage hip reconstruction in children with cerebral palsy. J Bone Joint Surg Am 2015; 97(6): 500-6. Schmale G A, Eilert R E, Chang F, Seidel K. High reoperation rates after early treatment of the subluxating hip in children with spastic cerebral palsy. J Pediatr Orthop 2006; 26(5): 617-23. Shea K G, Coleman S S, Carroll K, Stevens P, Van Boerum D H. Pemberton pericapsular osteotomy to treat a dysplastic hip in cerebral palsy. J Bone Joint Surg Am 1997; 79(9): 1342-51. Shore B J, Yu X, Desai S, Selber P, Wolfe R, Graham H K. Adductor surgery to prevent hip displacement in children with cerebral palsy: the predictive role of the Gross Motor Function Classification System. J Bone Joint Surg Am 2012; 94(4): 326-34. Shore B J, Zurakowski D, Dufreny C, Powell D, Matheney T H, Snyder B D. Proximal femoral varus derotation osteotomy in children with cerebral palsy: the effect of age, Gross Motor Function Classification System level, and surgeon volume on surgical success. J Bone Joint Surg Am 2015; 97(24): 2024-31. Soo B, Howard J J, Boyd R N, Reid S M, Lanigan A, Wolfe R, et al. Hip displacement in cerebral palsy. J Bone Joint Surg Am 2006; 88(1): 121-9. Stott N S, Piedrahita L. Effects of surgical adductor releases for hip subluxation in cerebral palsy: an AACPDM evidence report. Dev Med Child Neurol 2004; 46(9): 628-45. Terjesen T. To what extent can soft-tissue releases improve hip displacement in cerebral palsy? Acta Orthop 2017; 88(6): 695-700. Terjesen T, Lie G D, Hyldmo A A, Knaus A. Adductor tenotomy in spastic cerebral palsy: a long-term follow-up study of 78 patients. Acta Orthop 2005; 76(1): 128-37. Turker R J, Lee R. Adductor tenotomies in children with quadriplegic cerebral palsy: longer term follow-up. J Pediatr Orthop 2000; 20(3): 370-4. Westbom L, Hägglund G, Nordmark E. Cerebral palsy in a total population of 4–11 year olds in southern Sweden: prevalence and distribution according to different CP classification systems. BMC Pediatr 2007; 7: 41.

Acta Orthopaedica 2019; 90 (5): 501–504

501

Recurrent arthrocele and sterile sinus tract formation due to ceramic wear as a differential diagnosis of periprosthetic joint infection — a case report Nico Maximilian JANDL 1–3, Tim ROLVIEN 1–3, Daniel GÄTJEN 1,2, Anika JONITZ-HEINCKE 4, Armin SPRINGER 5, Veit KRENN 6, Rainer BADER 4, and Wolfgang RÜTHER 1,2 1 Department 3 Department

of Orthopedics, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; 2 Orthopedic University Hospital Bad Bramstedt; of Osteology and Biomechanics, University Medical Center Hamburg-Eppendorf, Hamburg; 4 Department of Orthopedics, Biomechanics and Implant Technology Research Laboratory, Rostock University Medical Center, Rostock; 5 Medical Biology and Electron Microscopy Center, Rostock University Medical Center, Rostock; 6 MVZ-Zentrum für Histologie, Zytologie und Molekulare Diagnostik, Trier, Germany Correspondence: n.jandl@uke.de Submitted 2018-12-23. Accepted 2019-04-08.

A 63-year-old female patient with total hip arthroplasty (THA) presented at our clinic with a massive swelling of the right hip joint. 13 years ago, cementless THA with an alumina ceramic-on-ceramic bearing (Biolox forte, CeramTec GmbH, Plochingen, Germany) had been performed due to advanced osteoarthritis. In the first years after THA, the patient had been symptomfree. The patient then complained of a swelling when sitting. MRI and CT showed a fluid-filled tumor of the right hip joint expanding into the gluteal muscle. Laboratory infection parameters were normal (C-reactive protein: not detectable; leucocytes: 7.1 G/L). The patient was twice bursectomized elsewhere 12 years after the THA and histopathological examination pointed to a granulomatous disease. Various differential diagnoses such as rheumatoid arthritis, sarcoidosis, tuberculosis and rare causes such as brucellosis, toxoplasmosis, echinococcosis and mycosis as well as an adverse local tissue reaction (ALTR) due to abrasive wear particles were considered. As further diagnostic procedures including QuantiFERON test and chest radiography showed no signs of tuberculosis and sarcoidosis, we performed an open synovial biopsy 10 months later to search for abrasive wear particles and to exclude periprosthetic joint infection (PJI). The dorsal hip joint capsule and the arthrocele were completely resected. Histology showed granulomas of the foreign body type and very sparse birefringent wear particles. In 5 of 8 tissue samples, the histological criteria for PJI according to Krenn et al. (20179) were not met. The microbiological culturing of tissue samples and synovial fluid over 14 days remained sterile. As the arthrocele recurred, revision THA was performed another 2 months later with suspicion of ALTR to wear particles and the acetabular cup, which was firmly integrated into the bone, was exchanged (Allofit IT, Zimmer, Warsaw, IN, USA). No macroscopic abnormalities of the head–neck junction or the connection between ceramic liner and acetabular cup were observed. A new aluminum-zirconium composite ceramic head and acetabular liner (Biolox delta, CeramTec

GmbH) was implanted. Particle analysis of tissue samples by CytoViva dark-field microscopy (CytoViva Inc, Auburn, AL, USA) indicated intracellular particles (Figure 1). 6 months later, the patient presented with a reddened distal wound margin but no infection parameters. A large arthrocele was palpable, which finally emptied itself via a sinus tract, i.e., a major criterion for PJI according to the Musculoskeletal Infection Society (MSIS, Parvizi et al. 2011) was fulfilled. During 2-step revision surgery with total explantation of all components and subsequent spacer implantation, brownishreddish fluid with semolina-like admixtures, but no purulence, was found periarticularly. Microbiology remained negative once again, but calculated antibiotic treatment was initiated. According to the histopathological synovial-like interface membrane (SLIM) consensus classification, the tissue morphology was between an infectious type (SLIM type II) and an adverse reaction (SLIM type VI) (Krenn et al. 2014, 2016, Perino et al. 2018), evaluated by 2 independent pathologists (Figure 2).

Figure 1. CytoViva dark-field microscopy (100×; CytoViva Inc, Auburn, AL, USA) of synovial tissue showed intracellular particles shining pink. The synovial tissue had been taken before the acetabular cup as well as the ceramic head and liner were removed.

502

Acta Orthopaedica 2019; 90 (5): 501–504

Figure 2. Histological view of the synovial tissue biopsies from the interface (synovia-like interface membrane, SLIM) of the THA revision 1 year after acetabular cup, ceramic head, and liner exchange. Grocott staining (100×) showing a pseudocyst with no evidence of mycosis (A). HE staining (12.5×) showing a fibrous cyst (B). HE staining (220×) showing a fibrinoid necrosis with focal palisading of fibroblasts and accumulation of macrophages (C).

As the wound secretion did not improve in 1 month, a Girdlestone procedure was done. All tissue and synovial samples remained sterile again and further special analysis concerning echinococcosis, toxoplasmosis, tuberculosis (microscopic analysis, cultures with 8 weeks of incubation and mycobacterial PCR), brucellosis and mycosis as well as a non-routine pan-bacterial PCR (bacterial and mycotic DNA) also remained negative. Apart from Figure 3. Calculation of the wear volume of the retrieved primary implanted ceramic head (A). unchanged histopathological findings no A wear volume of 9.4 mm³ was determined (B). other MSIS minor criteria were fulfilled, i.e., PJI was unlikely and antibiotic therapy was terminated as wound secretion did not improve postoperatively. MRI showed retained fluid draining via a sinus tract that was in direct connection with Discussion the hip joint. An autoimmune reaction to wear particles was We initially assumed PJI due to the occurrence of a sinus our remaining explanation for the recurrent arthrocele for- tract as a typical sign of an infection after the primary hip cup mation. Since serum metal ion concentrations were in the exchange. Different classification systems for diagnosing PJI reference range (chromium < 0.5 µg/L, cobalt < 0.9 µg/L) all include a sinus tract communicating with the endoprostheand metallosis was not found either macroscopically or sis as a sufficient criterion. When considering all our findings, microscopically, ceramic wear was assumed to explain the PJI seems very unlikely due to the extensive and repeated sparse particles found histologically. Under high-dose cor- negative diagnostic results, but there is a residual risk of infectisone therapy (30 mg prednisolone p.o. for 1 week, week tion, since culture-negative PJI is reported to range between 2: 25 mg, weeks 3–6: 20 mg, week 7: 15 mg, week 8: 10 7% and 42% (Reisener et al. 2018). However, the fact that mg, weeks 9–10: 5 mg) the clinical condition of the patient terminating the antibiotic therapy did not lead to increasing improved dramatically. infection parameters and that temporary cortisone therapy After 5 months of slow-reducing cortisone therapy the induced a dramatical clinical improvement speak against this. wound healed completely, and a total hip reimplantation with Another special feature of this case was the repeated hisa ceramic-on-polyethylene bearing was conducted. In the tological finding of a granulomatous inflammation. Other clinical follow-up examination 1.5 years later there was no non-infectious causes of granulomas such as rheumatoid evidence of a recurrent arthrocele, and the patient was mobile arthritis and sarcoidosis were excluded by laboratory analyat full load without pain. ses. According to Krenn et al. (2014, 2016), the morphology Retrospectively, the ceramic wear volume of the primary was between an adverse reaction (SLIM type VI) and an infecimplanted ceramic head was calculated with CAD (Figure 3, tion (SLIM type II). Even 2 experienced pathologists could see also Supplementary data). A massive wear volume of 9.4 not make a conclusive, definite histological diagnosis, which mm³ was determined. emphasizes the peculiarity of this case. Granuloma formation

Acta Orthopaedica 2019; 90 (5): 501–504

503

Untreated control

Particle exposure

0.001 mg/mL

0.01 mg/mL

Figure 4. Microscopic images of the patient’s isolated peripheral blood mononuclear cells (PBMCs). (A) Before exposure of PBMCs to alumina ceramic particles. (B) After exposure of PBMCs to alumina ceramic particles, cell differentiation and adherence increased. However, cell differentiation and adherence were slower than in healthy controls (results not shown) as recently published (Klinder et al. 2018). (C) Decreased cell adherence and cell differentiation of PBMCs after increasing the alumina ceramic particle concentration.

and necrosis made SLIM classification difficult. Against the background of repeated negative microbiology with histologically proven macrophage accumulation and mild lymphocytosis, SLIM type VI appears more likely. While light microscopy revealed only sparse wear particles, further non-routine histological examination showed intracellular particles and the wear calculation of the retrievals showed massive wear of the ceramic head. Therefore, we assume that the main trigger for the recurrent arthrocele formation was ceramic wear. It is surprising that despite massive ceramic wear only sparse particles were found histologically. However, wear-induced ceramic particles are difficult to detect by conventional histology, if at all, as the size ranges from 20 to 100 nm and only in some cases up to several micrometers (Perino et al. 2018). The key question is whether the formation of the recurrent arthrocele was due to an adverse reaction to metal or ceramic abrasion particles. ALTR or cyst formation is reported in metal-on-metal bearings (Mabilleau et al. 2008) or due to corrosion at modular taper junctions (Gill et al. 2012). Only in a few cases was ALTR also reported in ceramic-on-polyethylene bearings, which was attributed to metal abrasion particles from the taper (Bonnaig et al. 2011) or ceramic wear particle induced abrasion of the coating of the prosthesis stem (Bohler et al. 2000). We cannot rule out that arthrocele formation was caused by an ALTR on metal abrasion particles in our case and a limitation is that we did not examine the neck taper microscopically to exclude taper abrasion. However, no macroscopic or microscopic metallosis was found and the patient’s metal ion levels were normal. Pseudotumors in patients with metal-on-metal bearings are reported to show significantly higher serum metal ion levels than those without pseudotumor formation (Kwon et al. 2011). We therefore consider it highly probable that ceramic wear led to the development of ALTR with arthrocele formation. In vitro studies concerning the extent of a biological response of ceramic wear particles are inconsistent, but alumina particles

in their clinically relevant nanometer size are supposed to have a limited impact on cell viability and cytokine production (Petit et al. 2002, Gibon et al. 2017). Alumina particles in the clinically relevant nanometer range are reported to show a cytotoxic effect on human histiocytes, although this effect was weaker than exposition with cobalt-chromium particles (Germain et al. 2003). Furthermore, macrophages showed increased phagocytosis of ceramic particles up to 2 µm in size when raising the particle concentration and macrophage viability decreased with particle size and concentration if greater than 2 µm, whereas smaller particles with 0.6 µm were only cytotoxic at higher concentrations (Catelas et al. 1998). There are few studies that have investigated the effects of ceramics on periprosthetic tissue. In vivo experiments on rats found a moderate non-specific granulomatous response of the synovial membrane after injection of ceramic particles with a size below 1 µm into the knee joint (Roualdes et al. 2010). The histological examination of our patient’s samples also revealed granulomatous changes in the (neo-)synovial tissue. In line with our assumption, a recent case reported ceramic wear induced pseudotumor formation several years after THA with a ceramic-on-ceramic articulation (Campbell et al. 2017). Other studies revealed that ceramic wear particles mainly caused a response of macrophages in biopsies of revised hip protheses with a ceramic-on-ceramic bearing (Mochida et al. 2001). Pronounced periprosthetic tissue fibrosis was found in 9 patients with ceramic-on-ceramic articulation after total hip revision surgery compared with other bearings. The effects of ceramics were further examined by in vitro experiments, in which fibroblasts and peripheral blood mononuclear cells showed an inflammatory response when incubated on ceramic surfaces, especially on alumina-toughened zirconia (Bertrand et al. 2018). Taken together, these studies illustrate that ceramic wear may be biologically active. The appearance of a sinus tract is not a sign of ALTR or has not yet been described as such. In order to examine the inflam-

504

matory response of the patient to ceramic particle exposure, we performed further analysis of the patient’s peripheral blood mononuclear cells (PBMC). In comparison with the results of healthy subjects that were recently published (Klinder et al. 2018) the patient’s isolated PBMCs showed a lower cell adherence and differentiation after exposure to alumina ceramic particles. This effect was even more pronounced with increasing particle concentration (Figure 4). The formation of a sterile sinus tract may thus be explained by an altered local immunological reaction of PBMCs to ceramic wear particles mimicking PJI. In conclusion, ceramic wear should be considered as a differential diagnosis for recurrent arthrocele and sinus tract formation with negative microbiological cultures. Conflicts of interest All authors declare that they have no conflict of interest concerning this article and report no financial support. All procedures performed in this case report involving the patient were in accordance with the ethical standards of the local ethics committee. Informed consent was obtained from the patient.

Acta Orthopaedica 2019; 90 (5): 501–504

Germain M A, Hatton A, Williams S, Matthews J B, Stone M H, Fisher J, Ingham E. Comparison of the cytotoxicity of clinically relevant cobaltchromium and alumina ceramic wear particles in vitro. Biomaterials 2003; 24(3): 469-79. Gibon E, Cordova L A, Lu L, Lin T H, Yao Z, Hamadouche M, Goodman S B. The biological response to orthopedic implants for joint replacement, II: Polyethylene, ceramics, PMMA, and the foreign body reaction. J Biomed Mater Res B Appl Biomater 2017; 105(6): 1685-91. Gill I P, Webb J, Sloan K, Beaver R J. Corrosion at the neck–stem junction as a cause of metal ion release and pseudotumour formation. J Bone Joint Surg Br 2012; 94(7): 895-900. Klinder A, Seyfarth A, Hansmann D, Bader R, Jonitz-Heincke A. Inflammatory response of human peripheral blood mononuclear cells and osteoblasts incubated with metallic and ceramic submicron particles. Front Immunol 2018; 9: 831. Krenn V, Morawietz L, Perino G, Kienapfel H, Ascherl R, Hassenpflug G J, Thomsen M, Thomas P, Huber M, Kendoff D, Baumhoer D, Krukemeyer M G, Natu S, Boettner F, Zustin J, Kolbel B, Ruther W, Kretzer J P, Tiemann A, Trampuz A, Frommelt L, Tichilow R, Soder S, Muller S, Parvizi J, Illgner U, Gehrke T. Revised histopathological consensus classification of joint implant related pathology. Pathol Res Pract 2014; 210(12): 779-86. Krenn V, Perino G, Krenn V T, Wienert S, Saberi D, Hugle T, Hopf F, Huber M. [Histopathological diagnostic work-up of joint endoprosthesis-associated pathologies]. Hautarzt 2016; 67(5): 365-72.

Supplementary data The wear volume determination method is available as supplementary data in the online version of this article, http://dx.doi. org/10.1080/17453674.2019.1616997

Krenn V T, Liebisch M, Kölbel B, Renz N, Gehrke T, Huber M, Krukemeyer M G, Trampuz A, Resch H, Krenn V. CD15 focus score: Infection diagnosis and stratification into low-virulence and high-virulence microbial pathogens in periprosthetic joint infection. Pathol Res Pract 2017; 213(5): 541-7.

The authors gratefully thank Dr Paul Johan Hol (University of Bergen, Norway) for performing CytoViva darkfield microscopy.

Mabilleau G, Kwon Y M, Pandit H, Murray D W, Sabokbar A. Metal-onmetal hip resurfacing arthroplasty: a review of periprosthetic biological reactions. Acta Orthop 2008; 79(6): 734-47.

Study conduct: AJH, AS, VK, WR; data collection: NMJ, TR, DG, VK, WR; data analysis: AJH, AS, VK; drafting manuscript: NMJ, TR, AJH, VK, RB, WR. Acta thanks Kaj Knutson for help with peer review of this study.

Bertrand J, Delfosse D, Mai V, Awiszus F, Harnisch K, Lohmann C H. Ceramic prosthesis surfaces induce an inflammatory cell response and fibrotic tissue changes. Bone Joint J 2018; 100-B(7): 882-90. Bohler M, Mochida Y, Bauer T W, Plenk H, Jr, Salzer M. Wear debris from two different alumina-on-alumina total hip arthroplasties. J Bone Joint Surg Br 2000; 82(6): 901-9. Bonnaig N S, Freiberg R A, Freiberg A A. Total hip arthroplasty with ceramic-onceramic bearing failure from third-body wear. Orthopedics 2011; 34(2): 132. Campbell J, Rajaee S, Brien E, Paiement G D. Inflammatory pseudotumor after ceramic-on-ceramic total hip arthroplasty. Arthroplasty Today 2017; 3(2): 83-7. Catelas I, Huk O L, Petit A, Zukor D J, Marchand R, Yahia L. Flow cytometric analysis of macrophage response to ceramic and polyethylene particles: effects of size, concentration, and composition. J Biomed Mater Res 1998; 41(4): 600-7.

Kwon Y M, Ostlere S J, McLardy-Smith P, Athanasou N A, Gill H S, Murray D W. “Asymptomatic” pseudotumors after metal-on-metal hip resurfacing arthroplasty: prevalence and metal ion study. J Arthroplasty 2011; 26(4): 511-18.

Mochida Y, Boehler M, Salzer M, Bauer TW. Debris from failed ceramicon-ceramic and ceramic-on-polyethylene hip prostheses. Clin Orthop Relat Res 2001; (389): 113-25. Parvizi J, Zmistowski B, Berbari E F, Bauer T W, Springer B D, Della Valle C J, Garvin K L, Mont M A, Wongworawat M D, Zalavras C G. New definition for periprosthetic joint infection: from the Workgroup of the Musculoskeletal Infection Society. Clin Orthop Relat Res 2011; 469(11): 2992-4. Perino G, Sunitsch S, Huber M, Ramirez D, Gallo J, Vaculova J, Natu S, Kretzer J P, Muller S, Thomas P, Thomsen M, Krukemeyer M G, Resch H, Hugle T, Waldstein W, Boettner F, Gehrke T, Sesselmann S, Ruther W, Xia Z, Purdue E, Krenn V. Diagnostic guidelines for the histological particle algorithm in the periprosthetic neo-synovial tissue. BMC Clin Pathol 2018; 18:7. Petit A, Catelas I, Antoniou J, Zukor D J, Huk O L. Differential apoptotic response of J774 macrophages to alumina and ultra-high-molecular-weight polyethylene particles. J Orthop Res 2002; 20(1): 9-15. Reisener M, Perka C. Do culture-negative periprosthetic joint infections have a worse outcome than culture-positive periprosthetic joint infections? A systematic review and meta-analysis. Biomed Res Int 2018; 2018: 6278012. Roualdes O, Duclos M E, Gutknecht D, Frappart L, Chevalier J, Hartmann D J. In vitro and in vivo evaluation of an alumina-zirconia composite for arthroplasty applications. Biomaterials 2010; 31(8): 2043-54.

Acta Orthopaedica 2019; 90 (5): 505–506

505

Correspondence

An infrapatellar nerve block reduces knee pain in patients with chronic anterior knee pain after tibial nailing: a randomized, placebo-controlled trial in 34 patients Sir,—We congratulate Leliveld and colleagues on their report (2019) of a randomized controlled trial of lidocaine as a therapeutic intervention for post-operative chronic pain following tibial nailing, and of highlighting this problematic complication. Pain that persists years after surgery is clearly difficult to treat, and the possibility that a drug as inexpensive, and readily available, as lidocaine may attenuate it is in many respects an exciting one. We wish to discuss some of the details of their study and the interpretation of their observations. The sensory innervation of the knee is complex and variable (Bademkiran et al. 2007, Kerver et al. 2013). The surgical technique described, where, via an infrapatellar longitudinal incision, the medial aspect of the patellar ligament is accessed, may of course damage the infrapatellar genicular branch of the saphenous nerve. However, the dermatomal distribution of that nerve is largely the inferomedial aspect of the anterior knee. It would not fully account for the patterns of pain seen clinically – that is pain that covers the entire anterior knee, which is supplied by sciatic, tibial and common peroneal nerve components. Also while the authors hypothesise that nerve injury was caused by incision at the inferomedical aspect of the knee joint, they identified sensory abnormalities of the inferolateral side. These aspects of the report appear to be rather at odds with each other. Given the complexity of the sensory innervation of the anterior knee, it is hard to view pain on kneeling as being necessarily attributable to infrapatellar nerve injury (Bademkiran et al. 2007, Kerver et al. 2013). Observations of classic findings of neuropathic pain, such as allodynia or hyperalgesia in the infrapatellar nerve distribution, were not featured in the report (Freeman et al. 2019). Direct nerve injury may be a component of course, but we suggest that the evidence presented does not describe a patient group with a homogenous clinical pain syndrome. The responses to therapy within the group vary widely, in keeping with this suggestion. Ultimately what is described may be better seen as a diagnostic test to identify those who would benefit from longer term, or definitive, nerve blockade. The therapeutic value of local anaesthetic infiltration with a short-acting drug for pain persisting years after surgery is clearly limited, and indeed the observation that pain on kneeling is reduced by injecting the knee with that class of drug is itself unsurprising. It would presumably create numbness over some of the anterior aspect

of the knee, reducing such pain. However this may be a useful method to identify those who can benefit meaningfully from longer term interventions. In that sense, rather than excluding people from a meaningful treatment prematurely, perhaps the injection might be optimally done with ultrasound guidance rather than a landmark technique. Anatomic landmarks can of course be distorted by arthritis, and in particular, years after surgery, this might be significant. Ultrasound can facilitate more accurate injection and thus a more reliable diagnostic test might ensue, identifying all who would benefit (Orduña Valls et al. 2017) We would thank the investigators for raising this issue, and hope that such a simple test may benefit patients with longterm pain that may be treatable with more meaningful, long lasting interventions. It is, finally, worth considering if the test could be performed earlier. While they limited their study to those more than 6 months after surgery, which is often used as a definition of chronic pain, the median times reported were well over 5 years post-operatively. We would hope that this interesting study might ultimately enable more rapid identification of patients who can benefit, and thus reduce such experiences of long-term pain. Pradipta Bhakta, Habib Md Reazaul Karim, Brian O’Brien, and Michelle Claudio Vassallo email: bhaktadr@hotmail.com

Sir,—We thank Bhakta and colleagues for their comments on our recently published article. The infrapatellar branch of the saphenous nerve provides sensation over the anterolateral aspect of the knee and lower leg. The course of this nerve is highly variable. It crosses the patellar tendon as one or multiple branches (Kerver et al. 2013, Henry et al. 2017). The skin incision used for insertion of the tibial nail is commonly longitudinal and the nail is inserted through or just medial to the patellar tendon. Injury to infrapatellar branch at this level will cause loss of sensation at the anterolateral aspect of the knee and lower leg. Any incision (for other purposes) made more medial and cranial will cause loss of sensation of the medial aspect of the knee and calf.

506

Persisting anterior knee pain following tibial nailing is indeed a problem that is difficult to treat. The hypothesis of our randomized controlled study was that injury of the infrapatellar nerve contributes to this problem. Anterior knee pain is often multifactorial and infrapatellar nerve injury is one of the causes that has presumably been underestimated. The effect of subcutaneous injection with a local anaesthetic as studied in our trial was never intended as a therapeutic intervention, as we e.g. did not perform a long term follow of the participants. An infrapatellar nerve block with lidocaine can be used as a diagnostic tool to filter those patients who would potentially benefit from denervation of the infrapatellar nerve. Future studies should also focus on avoiding this complication. Mandala S Leliveld, Saskia J M Kamphuis, and Michael H J Verhofstad email: m.leliveld@erasmusmc.nl

Acta Orthopaedica 2019; 90 (5): 505–506

Bademkiran F, Obay B, Aydogdu I, Ertekin C. Sensory conduction study of the infrapatellar branch of the saphenous nerve. Muscle Nerve 2007; 35(2): 224-7. Freeman R, Edwards R, Baron R, Bruehl S, Cruccu G, Dworkin RH, Haroutounian S. AAPT diagnostic criteria for peripheral neuropathic pain: focal and segmental disorders. J Pain 2019; 20(4):3 69-93. Henry B M, Tomaszewski K A, Pękala P A, Ramakrishnan P K, Taterra D, Saganiak K, Mizia E, Walocha J A. The variable emergence of the infrapatellar branch of the saphenous nerve. J Knee Surg 2017; 30(6): 585-93. doi: 10.1055/s-0036-1593870. Kerver A L, Leliveld M S, den Hartog D, Verhofstad M H, Kleinrensink G J. The surgical anatomy of the infrapatellar branch of the saphenous nerve in relation to incisions for anteromedial knee surgery. J Bone Joint Surg Am 2013; 95(23): 2119-25. Leliveld M S, Kamphuis S J M, Verhofstad M H J. An infrapatellar nerve block reduces knee pain in patients with chronic anteriao knee pain after tibial nailing: a randomized, placebo-controlled trial in 34 patients. Acta Orthop 2019: 1-9. Doi: 10.1080/17453674.2019.1613808. [Epub ahead of print]. Orduña Valls J M, Vallejo R, López Pais P, Soto E, Torres Rodríguez D, Cedeño DL, et al. Anatomic and ultrasonographic evaluation of the knee sensory innervation: a cadaveric study to determine anatomic targets in the treatment of chronic knee pain. Reg Anesth Pain Med 2017; 42(1): 90-8.

5/19 ACTA ORTHOPAEDICA

Element of success in joint replacement

www.heraeus-medical.com

Vol. 90, No. , 2019 (pp. 411–506)

Volume 90, Number 5, October 2019

IORT_90_05_COVER.indd 1

16-08-2019 20:31:05