Page 1


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

EDI­TO­RI­AL OF­FI­CE

Acta Orthopaedica De­part­ment of Or­tho­pe­dics Lund Uni­ver­si­ty Hos­pi­tal SE–221 85 Lund, Swe­den Tel +46 46-17 15 96 E-mail: acta.ort@med.lu.se Homepage: http://www.actaorthop.org

EDI­TOR

BO­ARD

An­ders Ryd­holm Lund, Swe­den

Olle Svensson Umeå, Sweden, President Lars B Eng­e­sæ­ter Bergen, Nor­way Pe­ter Frand­sen Oden­se, Den­mark Ebbe Stender Hansen År­hus, Den­mark Ee­ro Hir­ven­sa­lo Hel­sinki, Fin­land Aare Märtson Tartu, Estonia Bjarne Møller-Madsen Århus, Denmark Rudolf Pöll Amsterdam, The Net­her­lands Grétar Ottó Róbertsson Reykjavik, Iceland An­ders Ryd­holm Lund, Swe­den, Edi­tor

DE­PU­TY EDI­TOR

Pe­ter A Frand­sen Oden­se, Den­mark CO-EDI­TORS

Per Aspenberg Linköping, Swe­den Ivan Hvid År­hus, Den­mark Yrjö T Konttinen Hel­sin­ki, Fin­land Lars Nordsletten Oslo, Norway Urban Rydholm Lund, Swe­den Ol­le Svens­son Umeå, Swe­den Bart A Swi­er­stra Nijmegen, The Net­her­lands Rolf Ön­ner­fält Lund, Swe­den STA­TIS­TICS CON­SUL­TANT

Jo­nas Ran­stam Lund, Swe­den PRO­DUC­TI­ON MA­NA­GER

Kaj Knut­son Lund, Swe­den

DE­PU­TI­ES

Jan-Magnus Björkenheim Helsinki, Finland Stefán Carlsson Reykjavik, Iceland Jan Erik Madsen Oslo, Norway Rein Raie Tallinn, Estonia Kjeld Søballe År­hus, Den­mark Jan A N Ver­haar Rot­ter­dam, The Net­her­lands Kristina Åkesson Malmö, Sweden

Vol. 82, No. 1, 2011


Cover illustration: Otte Sköld (1894–1958), Sweden. Arna with a cat, 1916. Oil on canvas, 73 × 59 cm. Malmö Konstmuseum, Sweden

Man­u­scripts should be uploaded at http://www.manuscriptmanager.com/ao/ for further handling at: Acta Orthopaedica Editorial Office De­part­ment of Or­tho­pe­dics Lund Uni­ver­si­ty Hos­pi­tal SE–221 85 Lund, Swe­den Correspondence concerning copyright and permissions should be addressed to: Håkan Pårup, Informa Healthcare P.O. Box 3255, SE-103 65 Stockholm, Sweden Tel: +46 8 440 80 40. Fax: +46 8 440 80 50 E-mail: hakan.parup@informa.com Correspondence concerning advertising should be sent to: Daniel Wallén, Informa Healthcare P.O. Box 3255, SE-103 65 Stockholm, Sweden. Tel: +46-8-440 80 40. Fax: +46-8-587 662 39 E-mail: salesscandinavia@informa.com Cor­re­spon­dence con­cern­ing sub­scrip­tions, dis­tri­bu­tion and back issues should be addressed to: Informa Healthcare, PO Box 6397, Basingstoke, Hampshire RG24 8QS, UK Fax +44 (0) 1256 33 0245 E-mail: orders@tandf.co.uk Informa Healthcare/Taylor & Francis retains a three year back issue stock of journals. Older volumes are held by our official stockists: Periodicals Service Company http://www.periodicals.com/tandf. html, 11 Main Street, Germantown, NY 12526, USA to whom all orders and enquiries should be addressed. Tel +1 518 537 4700; Fax +1 518 537 5899; E-mail: psc@periodicals.com Cus­tom­ers in USA and Can­ada are invited to ­address: Taylor & Francis Group, Journals Customer Service, 325 Chestnut Street, 8th Floor, Philadelphia, PA 19106 Tel +1 (800) 354-1420 or +1 (215) 625-8900 Fax +1 (215) 625-2940 E-mail: info@taylorandfrancis.com Acta Ortho­paed­ica (ISSN 1745-3674) vol­ume 81 (2010) com­prises six issues pub­lished in Feb­ru­ary, April, June, August, Octo­ber and Decem­ber plus sup­ple­ments (ISSN 1745-3690). Each issue com­ prises approx. 100 pages. Sub­scrip­tion rate for vol­ume 80 is USD 595 includ­ing air speeded deliv­ery and sup­ple­ments. Acta Ortho­paed­ica is pub­lished by Informa Healthcare/Taylor & Francis, Oslo, Norway. Information about journals and books from Informa Healthcare/ Taylor & Francis is available on our homepage on the world wide web, URL-address: http://www.informaworld.com

Sub­scrip­tions in Japan should be ordered through Maru­zen Co., Ltd., 3-10 Nihon­ba­shi, 2-chome, ­Chuo-ku, Tokyo 103, Japan. Exclusive agent in India, Pakistan, Bangladesh, Sri Lanka and Nepal is Globe Publication Prt Ltd., B-13, 3rd Floor, A Block, Shopping Complex, Naraina ­Vihar, Ring Road, New Dehli – 110028, India. Acta Orthopaedica (USPS permit number 014270) is published bimonthly in February, April, June, August, October and December. The 2010 US institutional subscription price is USD 595. Periodicals postage paid at Jamaica, NY 11431, by US Mailing Agent Air Business Ltd, c/o Worldnet Shipping USA Inc., 149-35 117th Street, Jamaica, New York, NY11434. US Postmaster: Send address changes to Acta Orthopaedica (sORT), Air Business Ltd, c/o Worldnet Shipping USA Inc., 14935 117th Street, Jamaica, New York, NY11434. Indexed in ASCA (Auto­matic Sub­ject Cita­tion Alert); Bio­log­i­cal Abstracts; Chem­i­cal Abstracts; Cumu­la­tive Index to Nursing and Allied Health Literature (CINAHL); Cur­rent Advances in Eco­log­ i­cal and Envi­ron­men­tal Sci­ences; Cur­rent Con­tents/Clin­i­cal Med­ i­cine; Cur­rent Con­tents/Life Sci­ences; Devel­op­men­tal Med­i­cine and Child Neu­rol­ogy; Faxon Finder; Focus on Sports Sci­ence and Med­i­cine; Energy ­Research Abstracts; Excerpta Med­ica/EMBASE; Health Plan­ning and Ad­min­is­tra­tion; In­dex Med­i­cus/MED­LINE; Index to Den­tal Lit­er­a­ture; Index Vet­e­ri­narius; INIS Ato­min­dex; Research Alert; Med­i­cal Doc­u­men­ta­tion Ser­vice; Nu­cle­ar Sci­ence Ab­stracts (Ceased); Life Sci­enc­es Col­lec­tion; Sci­ence Cita­tion Index; Sci­Search; Sport­Search; Un­cov­er; Vet­e­ri­nary Bul­le­tin. © 2011 Informa Healthcare/Taylor & Francis on licence from the Board of Acta Ortho­paed­ica. All arti­cles pub­lished in Acta Ortho­ paed­ica are pro­tected by cop­y­right, which cov­ers the exclu­sive rights to repro­duce and dis­trib­ute the arti­cle. No mate­rial in this jour­nal may be repro­duced pho­to­graph­i­cally or stored on micro­ film, in elec­tronic data bases, video or com­pact disks etc. with­out prior writ­ten per­mis­sion from Informa Healthcare/Taylor & Francis. Spe­cial reg­u­la­tion for pho­to­cop­ies in the USA: Author­i­za­tion to pho­to­copy items for inter­nal or per­sonal use, or the inter­nal or per­sonal use of spe­cific cli­ents, is granted by Scan­di­na­vian Uni­ver­sity Press for librar­ies and other users reg­is­tered with the Cop­y­right Clear­ance Cen­ter (CCC) Trans­ac­tional Report­ ing Ser­vice, pro­vided the fee of USD 86 per arti­cle is paid to CCC, 222 Rosewood Drive, Danvers, MA 01923, USA 00016470/08/USD 86. This author­i­za­tion does not include copy­ing for gen­eral dis­tri­bu­tion, pro­mo­tion, new works, or resale. In these cases, spe­cific writ­ten per­mis­sion must be obtained from Informa Healthcare/Taylor & Francis. Printed in Sweden by Wallin & Dalholm, Lund.


Ac­ta Or­tho­pa­e­di­ca

ISSN 1745-3674

Vol. 82, No. 1, February 2011 Register study Increasing incidence of hip arthroplasty for primary osteoarthritis in 30- to 59-year-old patients: A population based study from the Finnish Arthroplasty Register Original papers Elevation of circulating HLA DR+ CD8+ T-cells and correlation with chromium and cobalt concentrations 6 years after metal-onmetal hip arthroplasty: A randomized trial A prospective randomized study comparing electrochemically deposited hydroxyapatite and plasma-sprayed hydroxyapatite on titanium stems: 55 hips followed for 2 years with RSA and DXA The effect of hospital volume on length of stay, re-admissions, and complications of total hip arthroplasty: A population-based register analysis of 72 hospitals and 30,266 replacements Infectiological, functional, and radiographic outcome after revision for prosthetic hip infection according to a strict algorithm: 22 one-stage and 50 two-stage revisions with a mean follow-up time of 5 (2–17) years Validation of a 3D CT method for measurement of linear wear of acetabular cups: A hip simulator study Idiopathic and secondary osteonecrosis of the femoral head show different thrombophilic changes and normal or higher levels of platelet growth factors The rate of screw misplacement in segmental pedicle screw fixation in adolescent idiopathic scoliosis: The effect of learning and cumulative experience Would loss to follow-up bias the outcome evaluation of patients operated for degenerative disorders of the lumbar spine? A study of responding and non-responding cohort participants from a clinical spine surgery registry Transposition of the apophysis of the greater trochanter for reconstruction of the femoral head after septic hip arthritis in children: 4 children followed for more than 15 years Range of motion and strength after surgery for brachial plexus birth palsy: 107 patients followed for 12-year Wrist function recovers more rapidly after volar locked plating than after external fixation but the outcomes are similar after 1 year: A randomized study of 63 patients with a dorsally displaced fracture of the distal radius Health-related quality of life (EQ-5D) before and after orthopedic surgery Development of simulated arthroscopic skills: A randomized trial of virtual-reality training of 21 orthopedic surgeons Surgery for skeletal metastases in lung cancer: Complications and survival in 98 patients Time course of skeletal muscle regeneration after severe trauma: Muscle function against the background of MRI and histological findings Case report Diaphyseal femoral fatigue fracture associated with bisphosphonate therapy – 3 more cases Technical note Pseudoarthrosis repair after failed metatarsophalangeal 1 arthrodesis: Turnaround inlay graft in 26 patients Management of severe tibial bony defects with double metal blocks in knee arthroplasty—a technical note involving 9 cases

1

E T Skyttä, J Leskinen, A Eskelinen, H Huhtala, and V Remes

6

N P Hailer, R A Blaheta, H Dahlstrand, and A Stark

13

B G Bøe, S M Röhrl, T Heier, F Snorrason, and L Nordsletten

20

K T Mäkelä, U Häkkinen, M Peltola, M Linna, H Kröger, and V Remes

27

F H R De Man, P Sendi, W Zimmerli, T B Maurer, P E Ochsner, and T Ilchmann

35

A Jedenmalm, F Nilsson, M E Noz, D D Green, U W Gedde, I C Clarke, A Stark, G Q Maguire Jr, M P Zeleznik, and H Olivecrona E Cenni, C Fotia, E Rustemi, K Yuasa, G Caltavuturo, A Giunti, and N Baldini

42 50

K Abul-Kasim and A Ohlin

56

T K Solberg, A Sørlie, K Sjaavik, Ø P Nygaard, and T Ingebrigtsen

64

P Benum

69 76

M O Kirjavainen, Y Nietosvaara, S M Rautakorpi, V M Remes, T H Pöyhiä, I J Helenius, and J I Peltonen M K T Wilcke, H Abbaszadegan, and P Y Adolphson

82

K-Å Jansson and F Granath

90

C Andersen, T N Winding, and M S Vesterby

96

R J Weiss and R Wedin

102

T Winkler, P von Roth, G Matziolis, Maria R Schumann, S Hahn, P Strube, G Stoltenburg-Didinger, C Perka, G N Duda, and S V Tohtz

112

K Osugi, S Miwa, S Marukawa, K Marukawa, Y Kawaguchi, and S Nakato

114

I M Takács and B A Swierstra

116

S-W Baek and C H Choi


Correspondence Failed internal fixation Overtreatment of cruciate ligament injuries

119 122

Information to authors

124

Supplementum S 342. Effects of cox inhibitors on bone and tendon healing: Experimental studies of fracture healing, tendon healing and tendon-tobone healing in the rat (S Dimmen)

P Mereddy versus J Zustin S Løken, A Årøen, L Engebretsen versus P Aspenberg versus L S Lohmander, R B Frobell


Thank you, reviewers! Acta Orthopaedica thanks all reviewers for their work with manuscripts during year 2010. We depend on your work to keep and increase the quality of Acta. Your effort is much appreciated!

Aamodt, Arild Adolfsson, Lars Agholme, Fredrik Andersen, Thomas Apostolou, Constantinos Arner, Marianne Aro, Hannu T Arthursson, Astvaldur Johann Ashford, Robert Urquhart Aspenberg, Per Atroshi, Isam Baas, Jorgen Bakker, Astrid Balke, Maurice Bar-On, Elhanan Bauer, Henrik Benoni, Göran Bielecki, Tomasz Billi, Fabrizio Bjørgul, Kristian Björkenheim, Jan-Magnus Bodén, Henrik Boehling, Ulrich Boyer, Patrick Bramer, Jos A.M. Bremander, Ann B Breusch, Steffen Brittberg, Mats Bruun Lauritzen, Jes Bruzzone, Matteo Buchhorn, Gottfried H. Buma, Pieter Bøe, Berte Grimsmo Börlin, Niclas Cafiso, Viviana Campbell, David G Carlsson, Åke Catelas, Isabelle Catterall, Anthony Cenni, Elisabetta Chantelau, Ernst Choong, Peter F M Crnalic, Sead D’Angelo, Fabio Dahl, Jon

Dahlin, Lars B Dalén, Tore Davidsen, Michael de Waal Malefijt, Maarten Deehan, David John Degreef, Ilse Delank, Karl-Stefan Delaunay, Christian P Deutch, Søren Rasmussen Dijkstra, Sander Ding, Ming Djerf, Krister Doets, Hendrik Cornelis Engellau, Jacob Englund, Martin Eriksson, Bengt I Eskelinen, Antti Esteban, Jaime Fahlgren, Anna Figved, Wender Flivik, Gunnar Foss, Olav A Frandsen, Peter Frihagen, Frede Fritzell, Peter Fuhrmann, Renee Furia, John Furnes, Ove Galkowska, Hanna Gardeniers, JWM Gill, Richie Harinderjit Singh Gjerdet, Nils Roar Gosens, Taco Graves, Stephen Grimer, Rob Grotle, Margreth Arna Hagert, Elisabet Hailer, Nils P Hallab, Nadim James Hammarberg, Henrik Hansen, Torben B Hansson, Tommy Harilainen, Arsi Harman, Melinda Hasegawa, Masahiro

Havelin, Leif Ivar Hazell, Michelle Hedin, Hanne Heesterbeek, Petra Henricson, Anders Henriksen, Marius Hernandez-Vaquero, Daniel Herrlin, Kristian Hindsø, Klaus Hirschmann, Michael T Hogendoorn, Pancras Hsu, Wei-Hsiu Husby, Torstein Husby, Otto Schnell Hvid, Ivan Hägglund, Gunnar Hänninen, Arno Ilchmann, Thomas Inoue, Nozomu Itoh, Yoshi Jakobsen, Thomas Janssen, Dennis Jansson, Karl-Åke Jensen, Steen Lund Johannesson, Anton Jost, Bernhard Juliebø, Vibeke Juszczyk, Mateusz Jäger, Marcus Jämsen, Esa Järvinen, Teppo LN Kambhampati, Srinivas Kaptein, Bart L Kent, Peter Michael Kjærsgaard-Andersen, Per Kobayashi, Naomi Konttinen, Yrjö T Kristensen, Morten Tange Kristiansen, Leif Pål Krukhaug, Yngvar Kurmis, Andrew Paul Kärrholm, Johan Landgraeber, Stefan Landin, Lennart Larsson, Sune


Lauge-Pedersen, Henrik Laursen, Mogens Berg Laursen, Jens Ole Leonhard, Ramseier Leunig, Michael Lie, Stein Atle Lind, Martin Little, David Lohmander, Stefan Louwerens, Jan Willem Lundborg, Göran Lysholm, Jack Løken, Sverre Maffulli, Nicola Malizos, Konstantinos N. Malmivaara, Antti Mayer-Wagner, Susanne McBride, Callum W Mittelmeier, Wolfram Moojen, Dirk Jan Frederik Morlock, Michael M Mosconi, Paola Mueller, Lutz Arne Murray, James R D Mäkelä, Keijo T Märtson, Aare Møller-Madsen, Bjarne Nelissen, Rob Nijhof, Marc Nilsson, Kjell G Nordsletten, Lars Norlin, Rolf Oh, Chang-Wug Overgaard, Søren Pajarinen, Jarkko Parker, Martyn Parsch, Klaus Parvizi, Javad Paulsen, Aksel Pedersen, Alma B.

Peltonen, Jari Perren, Stephan M Petersen, Morten Boye Picci, Piero Pinzur, Michael S Pitkänen, Mikko Poolman, Rudolf W Pratt, Nicole Rahbek, Ole Randsborg, Per-Henrik Ranstam, Jonas Remes, Ville Robertsson, Otto Rogmark, Cecilia Rosenberg, Andrew Ryd, Leif Räsänen, Pirjo Röhrl, Stephan Maximilian Savarino, Lucia Schimmel, Janneke Schreurs, B. Willem Seebauer, Ludwig Siggeirsdottir, Kristin Simon, Jean-Pierre Skripitz, Ralf Soinivaara, Tarja Sokka, Tuulikki Sollerman, Christer Spruit, Maarten Stark, André Steen, Harald Stefánsdóttir, Anna Steiger, Richard N de Sterner, Thomas Stumpe, Katrin D Suehara, Hiroshi Sundberg, Martin Svensson, Olle Swierstra, Bart A Søballe, Kjeld

Takakubo, Yuya Tarala, Maria Tengvall, Pentti Terjesen, Terje Theil, Jørn Tjernström, Björn Toksvig-Larsen, Sören Torfing, Trine Trovik, Clement Stig Tägil, Magnus W-Dahl, Annette Waddell, James P Walsh, William Robert Valstar, Edward Van der Heide, Huub J.L. Van Dijk, Niek van Osch, Gerjo Van Royen, Barend J van Tienen, Tony Wang, Jian-Sheng vanRiet, Roger Wedin, Rikard Weidenhielm, Lars Weiss, Stefan Verdonschot, Nico Verheyen, Cees CPM Wik, Tina Strømdal Wisbech Pedersen, Niels Wolf, Sebastian I von Schewelov, Thord Wykman, Anders You, Zongbing Zaikova, Olga Zimmerli, Werner Zoccali, Carmine Åstrand, Jörgen Öner, Cumhur Önnerfält, Rolf Önsten, Ingemar


Acta Orthopaedica 2011; 82 (1): 1–5

1

Increasing incidence of hip arthroplasty for primary osteoarthritis in 30- to 59-year-old patients A population based study from the Finnish Arthroplasty Register Eerik T Skyttä1,2, Leskinen Jarkko3, Eskelinen Antti 1, Heini Huhtala4, and Remes Ville3 1COXA Hospital for Joint Replacement, Tampere; 2Centre for Rheumatic Diseases, Department of Orthopaedics, Tampere University Hospital, Tampere; 3Department of Orthopedics, Peijas Hospital, Helsinki University Central Hospital, Helsinki; 4School of Public Health, University of Tampere, Tampere,

Finland Correspondence: eerik.skytta@fimnet.fi Submitted 10-04-24. Accepted 10-09-28

Background and purpose The use of hip arthroplasties is evidently increasing, but there are few published data on the incidence in young patients. Methods We used data on total and resurfacing hip arthroplasties (THAs and RHAs) from the Finnish Arthroplasty Register and population data from Statistics Finland to analyze the incidences of THA and RHA in patients aged 30–59 years in Finland, for the period 1980 through 2007. Results The combined incidences of THAs and RHAs among 30- to 59-year-old inhabitants increased from 9.5 per 105 inhabitants in 1980 to 61 per 105 inhabitants in 2007. Initially, the incidence of THA was higher in women than men, but since the mid90s the incidences were similar. The incidence increased in all age groups studied (30–39, 40–49, and 50–59 years) but the increase was 6-fold and 36-fold higher in the latter two groups than in the first. The incidence of THA was constant; the increased incidence of overall hip arthroplasty was due to the increasing number of RHAs performed. Interpretation We have found a steady increase in the incidence of hip arthroplasty in patients with primary hip osteoarthritis in Finland, with an accelerating trend in the past decade, due to an increase in the incidence of RHA. As the incidence of hip osteoarthritis has not increased, the indications for hip arthroplasty appear to have become broader. 

75% of THAs are performed on elderly patients for painful osteoarthritis (OA); in younger patients (under 50–60 years), the proportion of OA diminishes to 42–54%. Most patients are women (Lucht 2000, Furnes et al. 2001, Puolakka et al. 2001, Malchau et al. 2002). Several authors have reported an increasing incidence of treatment of OA with THA (Birrell et al. 1999, Ostendorf et al. 2002, Wells et al. 2002, Merx et al. 2003, Kurtz et al. 2005,

2007), but only a few authors have reported the incidences for younger patients separately (Birrell et al. 1999, Ingvarsson et al. 1999). Resurfacing hip arthroplasty (RHA) is an option marketed for younger patients, though its value is still uncertain (McGrory et al. 2009). We examined the changes in incidence of primary THA and RHA in young patients with OA in Finland between the years 1980 and 2007.

Patients and methods Inclusion criteria All THAs and RHAs performed in patients aged 30–59 years with primary hip osteoarthritis who had been entered in the Finnish Arthroplasty Register between 1980 and 2007 were included (Table). Data concerning the Finnish population during the study period were obtained from Statistics Finland; we used the end-of-year population as an approximation of the population each year. The population data were divided into cohorts by age (30–39, 40–49, and 50–59 years) and subgroups according to sex. Incidence of hip arthroplasty The general incidence is presented as the number of operations performed per 105 person years in Finnish inhabitants aged 30–59 years. Incidence by sex is presented as the number of operations performed per 105 person years in men or women aged 30–59 years. The incidences by age group are presented as the number of operations performed per 105 person years in Finnish inhabitants aged 30–39 years, 40–49 years, and 50–59 years. All incidence calculations were done separately for THAs and RHAs.

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2010.548029


2

Acta Orthopaedica 2011; 82 (1): 1–5

Hip arthroplasties in young patients (30–59 years of age) during 1980–2007, according to type of hip arthroplasty, sex, and age group Patients aged 30–59 years (primary OA) THA RHA Proportion of women (< 60 years) THA RHA Age groups of the study population 30–39 years of age THA RHA 40–49 years of age THA RHA 50–59 years of age THA RHA THA: total hip arthroplasty; RHA: resurfacing hip arthroplasty.

n

%

17,248 15,014 2,234 8,159 7,335 824 17,248 407 302 105 2,619 2,083 536 14,222 12,629 1,593

100 87 13 47 43 4.8 100 2.4 1.8 0.6 15 12 3.1 82 73 9.2

Statistics We analyzed the differences between every group for general incidence of THA and RHA, sex-specific incidence, and incidence by age group by calculating incidence rate ratio (IRR) using Poisson regression. For each comparison, the IRR for annual increase in incidence was calculated first and the effect of the group was then calculated. The increase in IRR is presented as a percentage (e.g. IRR = 1.048 is shown as 4.8%) with 95% CI. The differences were considered to be statistically significant if p-values were less than 0.05 in a two-tailed test. We used STATA 8.2 statistical software (StataCorp, College Station, TX).

Results Patient characteristics Between 1980 and 2007, 17,248 hip arthroplasties were performed for primary osteoarthritis in patients aged less than 60 years in Finland (Table). Of these operations, 8,159 (47%) were performed on women and 2,234 (13%) were RHAs. All RHAs were performed from 2001 through 2007. Median age at the time of surgery was 55 years (interquartile range: 51–58). General incidence There was a 6-fold increase in the incidence of hip arthroplasty over the whole study period: the incidence increased from 9.5 operations per 105 inhabitants in 1980 (in patients aged 30–59 years) to 61 in 2007 (Figure 1). The peak incidence was noted in 2006 (62 per 105) but the decrease in the number of RHAs reduced the general incidence despite the steady growth in the

Hip replacements / 100,000 70 60

All hip replacements Total hip replacement Resurfacing hip replacement

50 40 30 20 10 0 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007

Figure 1. General incidence of hip arthroplasty for primary osteoarthritis in young patients (30–59 years of age).

number of THAs. The IRR for annual increase in general incidence was 6.6% (CI: 6.3–6.8). The incidence of THA grew steadily from 9.5 to 39 per 105 during the period 1980–2007, with 858 THAs performed in 2007. The IRR for annual increase in incidence of THA was 4.5% (CI: 4.3–4.7). The incidence of RHA increased rapidly from 1.1 to 26 operations per 105 during 2001–2006 (Figure 1), but the incidence decreased to 22 in 2007, with 484 RHAs performed in that year. The IRR for annual increase in incidence of RHA was 37% (CI: 36–40). Incidence by sex In 1980, hip arthroplasties were performed 1.7 times more frequently in women than in men (11.9 vs. 7.1 per 105 inhabitants) (Figure 2A). However, regarding whether there was any difference between the sexes, the IRR for increase in incidence of hip arthroplasty was 9.1% (CI: 5.9–12.4) higher in men than in women (p < 0.001). Given the IRR for annual increase in general incidence of 6.6% (see above), hip arthroplasties were performed 1.2 times more frequently in men than in women (67 vs. 56 per 105) in 2007. There were no statistically significant differences in annual increase in the incidence of THA between men and women (IRR = 3.0%, CI: –1.0 to 5.9; p = 0.1) (Figure 2B). In contrast, the IRR for increase in incidence of RHA was 68% (CI: 54–83) higher in men than in women (p < 0.001) (Figure 2C). Incidence by age group In all age groups, there appeared to be an increase in incidence of all hip arthroplasties—THAs and RHAs (Figure 3A). In the 30–39-year age group, the IRR for annual increase in incidence of hip arthroplasty was 5.2% (CI: 5.0–5.4). Between


Acta Orthopaedica 2011; 82 (1): 1–5

3

Hip replacements / 100,000

Hip replacements / 100,000

Hip replacements / 100,000

70

70

70

60

Men Women

60

Men Women

60

50

50

50

40

40

40

30

30

30

20

20

20

10

10

10

0 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007

A

0 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007

B

Men Women

0 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007

C

Figure 2. A. Incidence of all hip arthroplasties for primary osteoarthritis in young patients (30–59 years of age), by sex. The increase in incidence during 1980–2007 was greater in men than in women (p < 0.001). B. Incidence of total hip arthroplasty (THA) for primary osteoarthritis in young patients (30–59 years of age), by sex. C. Incidence of resurfacing hip arthroplasty (RHA) for primary osteoarthritis in young patients (30–59 years of age), by sex. The increase in incidence during 2001–2007 was greater in men than in women (p < 0.001). Hip replacements / 100,000

Hip replacements / 100,000

Hip replacements / 100,000

150

150

150

140

50–59 years

130

40–49 years

120

30–39 years

140

50–59 years

130

40–49 years

120

30–39 years

140

50–59 years

130

40–49 years

120

110

110

110

100

100

100

90

90

90

80

80

80

70

70

70

60

60

60

50

50

50

40

40

40

30

30

30

20

20

20

10

10

10

0 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007

A

30–39 years

0 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007

0 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007

B

C

Figure 3. A. Incidence of all hip arthroplasties for primary osteoarthritis in young patients (30–59 years of age), by age group. B. Incidence of total hip arthroplasty (THA) for primary osteoarthritis in young patients (30–59 years of age), by age group. C. Incidence of resurfacing hip arthroplasty (RHA) for primary osteoarthritis in young patients (30–59 years of age), by age group.

groups, the difference in IRR for increase in incidence of hip arthroplasty was 6-fold (614%, CI: 594–634; p < 0.001), i.e. any given incidence in the 30–39-year group was 6-fold higher in the 40–49-year group, and it was 36-fold higher in the 50–59-year group. In THAs, the difference between groups was even more pronounced, with a 7-fold increase in IRR for increase in incidence of hip arthroplasty (690%, CI: 665–716; p < 0.001). The IRR for annual increase in incidence of THA was 3.2% (CI: 2.9–3.4). In RHAs, there was also a marked 3-fold increase in IRR between groups (320%, CI: 298–343, p < 0.001).

Discussion Our main finding was the steady annual increase in incidence of hip arthroplasty in young patients over the period 1980– 2007, with an accelerating trend over the past decade (Figure 1). The incidence of THA had a constant annual increase of 4.5%, and the acceleration appeared to be triggered by RHA quickly becoming popular since its inception in 2001. In Finland, the overall hip arthroplasty (THA and RHA) incidences have grown more quickly in men than in women. In the age groups 30–39, 40–49, and 50–59 years, the increase in incidence has been more pronounced in the latter 2 groups, the most substantial growth being in the last group.


4

Registry-based studies have certain limitations. The Finnish Arthroplasty Register covered 90% of hip implants in the year 1995, and after that the coverage increased to 98% (Mäkelä et al. 2008). This might explain a minor part of the increase shown in our study. Because of the accurate civil registry of Statistics Finland and the low level of illegal immigration in Finland, the population data can be assumed to correspond to the actual population. Nevertheless, limitations in the accuracy of the data and in conclusions drawn may result from inconsistencies or errors in the diagnostic coding entered into the registry (Ingvarsson et al. 1999, Pedersen et al. 2004) There have been few studies in which separate incidences have been reported for young patients. Kurtz et al. (2005) found a 50% increase in incidence in THA in the USA during 1990–2002, and, although not directly stated, an approximate annual increase in incidence of 4.5% in the youngest-studied group (45–64 years) can be calculated from their article, which coincides with our results. In a study from NHS hospitals in the UK in 1996 (Dixon et al. 2004), incidences for THA in women in the age groups < 39, 40–49, and 50–59 years were 2.1, 16, and 115 per 105 individuals. In men, the corresponding incidences were 1.4, 13, and 90 per 105. In our study, the incidence of THA in women was higher than in men only until the mid-1990s. Presently, the incidence of THA is similar in men and women, and the incidence of all hip arthroplasties is 20% higher in men (Figure 2). Another study using the same UK NHS hospital data from 1996 showed total incidence rates of 12 and 189 per 105 for the age groups 0–54 and 55–64 years (Birrell et al. 1999). The authors also projected the need for THA in the UK—but only for patients aged over 55 years— using estimated changes in demographic data until 2026. The projected overall increase in number of THAs was 24% for the period 1996–2006 in the youngest group of 55–64-yearold patients with OA. During the same time period in our study, the increase in number of hip arthroplasties in the oldest group (50–59 years) was 42% and 121% for THAs and all hip arthroplasties, respectively. However, any direct comparison between these studies is impossible, because the age groups were not the same, registration differed considerably, and demographic changes (immigration, birth rates, etc) can be divergent. We cautiously conclude that our increasing incidence rates support the direction of the projected changes in hip arthroplasty numbers in the study by Birrell et al. (1999). Wells et al. (2002) reported an increase in the incidence of THA from 48 to 84 per 105 in patients with primary osteoarthritis in the province of South Australia between 1988 and 1998. Between 1994 and 1998 in the total Australian population, the incidence rates increased from 51 to 61 per 105. The article does not provide actual incidence rates in numbers for different age groups but the annual increase in incidence of THA was 5% in all age groups (Wells et al. 2002). We have shown an obvious increase in the incidence of hip arthoplasty in young patients. The finding is apparent in both sexes and even in the youngest age groups. Demographic

Acta Orthopaedica 2011; 82 (1): 1–5

changes alone give no explanation to the changes in incidence. Furthermore, the prevalence of hip OA and hip symptoms has remained unchanged in Finland and other countries over earlier decades (Danielsson and Lindberg 1997, Heliövaara et al. 2007). Orthopedic surgeons are ultimately responsible for the increased rate, as they do the operations. Have we changed our approach due to pressure from patients, industry, or new clinical data? Our modern way of life possibly requires higher functional ability in both work and leisure. Younger patients may opt for elective operations at an earlier stage, with milder symptoms, than previously. Prosthesis manufacturers are eager to develop and market new concepts and designs, including novel THA bearings and RHA. This marketing takes place both indirectly through community- or internet-based advertising, or directly through their relationships with orthopedic surgeons. Good long-term results in terms of implant survivorship of THA in older age groups have encouraged surgeons to broaden the indications for THA to younger and more active patients. Whatever the underlying reasons for the observed increase in the incidence of hip arthroplasty, the advent of RHA has had a substantial impact on the incidence of hip arthroplasty. In conclusion, age- and sex-standardized incidences of THA and RHA in young patients with primary osteoarthritis have increased in Finland. This phenomenon has been especially strong during the current decade. There is no single explanatory factor for this growth. Some of the increase in incidence may be explained by hospital volume or geographic location of the patients. In general, the observed growth is so rapid and high, that given the lack of information on long-term outcome of hip arthroplasties in younger patients and the recent concerns about metal-on-metal hip arthroplasties (Langton et al. 2010), we should really be worried about a rapid increase in revision hip arthroplasties in the near future. Future long-term results of hip arthroplasty in younger age groups will reveal whether or not our current policy of widespread use of hip arthroplasty in younger patients is wise.

All authors designed the study, ETS gathered the data, ETS and HH analyzed the data and ensured its accuracy, and all authors wrote the initial drafts of the manuscript.

No competing interests declared.

Birrell F, Johnell O, Silman A. Projecting the need for hip replacement over the next three decades: Influence of changing demography and threshold for surgery. Ann Rheum Dis 1999; 58: 569-72. Danielsson L, Lindberg H. Prevalence of coxarthrosis in an urban population during four decades. Clin Orthop 1997; (342): 106-10. Dixon T, Shaw M, Ebrahim S, Dieppe P. Trends in hip and knee joint replacement: Socioeconomic inequalities and projections of need. Ann Rheum Dis 2004; 63: 825-30.


Acta Orthopaedica 2011; 82 (1): 1–5

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: 579-86. Heliövaara M, Impivaara O, Nykyri E, Riihimäki H. Changes in morbidity. In: Musculoskeletal disorders and diseases in Finland (ed. Kaila-Kangas L). Publications of the National Public Health Institute, Helsinki 2007; B 25: 60-1. Ingvarsson T, Hagglund G, Jonsson H, Jr, Lohmander L S. Incidence of total hip replacement for primary osteoarthrosis in Iceland 1982-1996. Acta Orthop Scand 1999; 70: 229-33. Kurtz S, Mowat F, Ong K, Chan N, Lau E, Halpern M. Prevalence of primary and revision total hip and knee arthroplasty in the United States from 1990 through 2002. J Bone Joint Surg (Am) 2005; 87: 1487-97. 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: 780-5. Langton D J, Jameson S S, Joyce T J, Hallab N J, Natu S, Nargol A V. Early failure of metal-on-metal bearings in hip resurfacing and large-diameter total hip replacement: A consequence of excess wear. J Bone Joint Surg (Br) 2010; 92: 38-46. Lucht U. The Danish hip arthroplasty register. Acta Orthop Scand 2000; 71: 433-9. Mäkelä K T, Eskelinen A, Pulkkinen P, Paavolainen P, Remes V. Total hip arthroplasty for primary osteoarthritis in patients fifty-five years of age or older. an analysis of the Finnish arthroplasty registry. J Bone Joint Surg (Am) 2008; 90: 2160-70.

5

Malchau H, Herberts P, Eisler T, Garellick G, Soderman P. The Swedish total hip replacement register. J Bone Joint Surg (Am) (Suppl 2) 2002; 84: 2-20. McGrory B, Barrack R L, Lachiewicz P F, Schmalzried T P, Yates A J. Modern metal-on-metal hip resurfacing - A technology overview. AAOS 2009. http://www.aaos.org/research/overviews/hipresurfacing.pdf Merx H, Dreinhofer K, Schrader P, Sturmer T, Puhl W, Gunther K P, Brenner H. International variation in hip replacement rates. Ann Rheum Dis 2003; 62: 222-6. Ostendorf M, Johnell O, Malchau H, Dhert W J, Schrijvers A J, Verbout A J. The epidemiology of total hip replacement in the Netherlands and Sweden: Present status and future needs. Acta Orthop Scand 2002; 73: 282-6. Pedersen A, Johnsen S, Overgaard S, Soballe K, Sorensen H T, Lucht U. Registration in the Danish hip arthroplasty registry: Completeness of total hip arthroplasties and positive predictive value of registered diagnosis and postoperative complications. Acta Orthop Scand 2004; 75: 434-41. Puolakka T J, Pajamaki K J, Halonen P J, Pulkkinen P O, Paavolainen P, Nevalainen J K. The Finnish arthroplasty register: Report of the hip register. Acta Orthop Scand 2001; 72: 433-41. Wells V M, Hearn T C, McCaul K A, Anderton S M, Wigg A E, Graves S E. Changing incidence of primary total hip arthroplasty and total knee arthroplasty for primary osteoarthritis. J Arthroplasty 2002; 17: 267-73.


6

Acta Orthopaedica 2011; 82 (1): 6–12

Elevation of circulating HLA DR+ CD8+ T-cells and correlation with chromium and cobalt concentrations 6 years after metal-on-metal hip arthroplasty A randomized trial Nils P Hailer,1 Roman A Blaheta2, Henrik Dahlstrand,3 and André Stark3 1Department of Orthopedics, Institute of Surgical Sciences, Uppsala University Hospital, Uppsala, Sweden; 2Department of Urology, Goethe University, Frankfurt am Main, Germany; 3Department of Molecular Medicine and Surgery, Section of Orthopedics, Karolinska Institute, Karolinska University Hospital, Stockholm, Sweden Correspondence: nils.hailer@surgsci.uu.se Submitted 10-03-15. Accepted 10-09-06

Background and purpose Following metal-on-metal hip arthroplasty (THA), immunological reactions including changes in lymphocyte populations, aseptic loosening, and lymphocytic pseudotumors occur. We hypothesized that changes in lymphocyte subpopulations would be associated with elevated metal ion concentrations. Methods A randomized trial involving 85 patients matched for age and sex and randomized to receiving metal-on-metal (n = 41) or metal-on-polyethylene total hip arthroplasty (n = 44) was conducted. 36 patients were eligible for follow-up after mean 7 (6–8) years. Concentrations of chromium and cobalt were analyzed by high-resolution inductively coupled plasma mass spectrometry. Leukocyte subpopulations and immunoglobulins in patient blood were measured using standard laboratory methods. Results Patients with a metal-on-metal hip had higher serum concentrations of chromium (1.05 vs. 0.36 µg/L; p < 0.001) and cobalt (0.86 vs. 0.24 µg/L; p < 0.001) than those with metal-onpolyethylene. The percentage of HLA DR+ CD8+ T-cells was higher in the metal-on-metal group (10.6 vs. 6.7%; p = 0.03) and correlated positively with chromium and cobalt concentrations in patient blood (Pearson’s correlation coefficient: 0.39, p = 0.02; 0.36, p = 0.03, respectively). The percentage of B-cells was lower in the metal-on-metal group (p = 0.01). The two groups were similar with respect to immunoglobulin concentrations and Harris hip scores, and there were no radiographic signs of loosening. Interpretation We conclude that immunological alterations appear to be associated with increased cobalt and chromium concentrations. It is tempting to speculate that HLA DR+ CD8+ T-cells are involved in the pathogenesis of allergic reactions, implant loosening, and lymphocytic pseudotumors. 

Immunological phenomena—both local and systemic— that are attributed to elevated metal ion concentrations have been described after modern metal-on-metal arthroplasty: 1. Lymphocyte-mediated inflammatory reactions occur in the vicinity of metal-on-metal articulations, and polyethyleneindependent osteolysis has been characterized histologically in such cases (Davies et al. 2005, Willert et al. 2005, Lazarinis et al. 2008). 2. The development of periprosthetic soft-tissue masses containing large numbers of lymphocytes has been identified as a cause of persistent pain, especially in females, after metal-on-metal hip resurfacing (Pandit et al. 2008, Toms et al. 2008). 3. At the systemic level, the induction of delayedtype hypersensitivity directed against metal ions has been observed after metal-on-metal THA (Hallab et al. 2004). 4. A decrease in the amount of circulating CD8+ T-cells has been described in patients with elevated metal ion levels subsequent to metal-on-metal THA, indicating further systemic immunological effects (Hart et al. 2006, 2009). We have recently published a study of patients who were randomized to receive either a metal-on-polyethylene or a metal-on-metal bearing with a 28-mm metal head (Dahlstrand et al. 2009). Clinical parameters, radiology results, and concentrations of chromium, cobalt, nickel, and manganese were followed, and we found elevation of all metal ions after 2 years in the metal-on-metal group. In the present study, we hypothesized that immunological changes can occur as a consequence of elevated metal ion concentrations in the medium term. Specifically, in the light of previously published findings, we expected changes in subsets of CD4+ or CD8+ lymphocytes, but no gross changes in other lymphocyte subpopulations or in serum immunoglobulins.

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2010.548028


Acta Orthopaedica 2011; 82 (1): 6â&#x20AC;&#x201C;12

7

Figure 1. CONSORT flow chart.

Enrollment Assessed for eligibility: n = 166

Excluded: n = 81 Did not meet inclusion criteria or declined to participate

Allocated to intervention M-PE: n = 44

Allocation

Received allocated intervention: n = 44 Did not receive allocated intervention: n = 0

Lost to follow-up: n = 5

Allocated to intervention M-M: n = 41 Received allocated intervention: n = 41 Did not receive allocated intervention: n = 0

Follow-up

Lost to follow-up: n = 5

4 dead, 1 too sick to attend

1 dead, 1 demented, 1 address unknown, 1 declined further participation, 1 too sick to attend

Discontinued intervention: n = 0

Discontinued intervention n = 0

Excluded from analysis: n = 22 Received additional metal implants prior to follow-up

Excluded from analysis: n = 17 Received additional metal implants prior to follow-up

Analysed minimum 72 months: n = 17

Analysis

Patients and methods Study design and population This prospective randomized study was performed in accordance with the ethical standards of the Helsinki declaration. Informed consent was obtained from all patients and the study was approved by the local ethics committee (no. 2006/958). The primary endpoints of the study were (1) the determination of concentrations of the heavy metal ions chromium, cobalt, nickel, and manganese in patient blood and (2) implant migration relative to surrounding bone, as determined by radiostereometry in 2 groups of patients that were randomized to receiving either a metal-on-metal bearing or a metalon-polyethylene bearing. The investigation of immunological parameters was added as a secondary endpoint for this study at a later stage. 166 patients, referred to the Department of Orthopedics, Karolinska Hospital, Sweden because of osteoarthritis of the hip, were eligible for study participation. Inclusion criteria were pain due to radiographically verified osteoarthritis and age between 40 and 75 years. Exclusion criteria were refusal to participate in the study, previous surgery with either osteosynthesis or joint replacement, weight over 105 kg, previous infection or surgery in the affected hip, local or general osteoporosis, intake of systemic cortical steroids for more

Analysed minimum 72 months: n = 19

than 3 months during the previous year, abuse of alcohol or drugs, and mental disorders including dementia. Strict application of inclusion and exclusion criteria left a cohort of 85 patients who were allocated to one of two groups according to the minimization method: 44 patients received a metal-onpolyethylene bearing and 41 patients received a metal-onmetal bearing (Figure 1). The groups were matched according to sex, smoking habits, body weight, and age. Neither the patients nor the authors involved were blinded as to the type of bearing; however, for determination of the Harris hip score, a physiotherapist blinded to the type of bearing collected the data. All patients were followed for 2 years, but 10 were subsequently lost to follow-up. Furthermore, 39 patients had to be excluded from the analysis presented here because they had received additional metal implants, rendering the measurement of metal ion concentrations meaningless. This left a study population of 36 patients, 19 in the metal-on-metal group and 17 in the metal-on-polyethylene group, who were followed for a mean of 82 (72â&#x20AC;&#x201C;97) months. Implants In all patients, a matte, tapered stainless steel stem (MS30; Sulzer, Winterthur, Switzerland) with a modular 28-mm head was combined with a low-profile cup. Two types of bearings were used: one consisted of a wrought high-carbon, cobalt-


8

chromium alloy head against a cobalt-chromium liner fitted in a polyethylene cup, making it a metal-on-metal bearing (Metasul; Sulzer). The other articulation was a similar cobaltchromium head (Protasul; Sulzer) against an all-polyethylene cup (Müller, Durasul; Sulzer). The alloy of the heads and liner consisted of a minimum of 50% cobalt, 26–30% chromium, 7% molybdenum, 1% nickel, and 1% manganese. The Metasul head is more polished than the Protasul, in order to minimize friction against the metal liner. Stem and cup were cemented with Palacos with gentamicin (Heraeus, Hanau, Germany) in all cases. All procedures were performed through a standard posterior approach by the same senior surgeon (AS) between October 1998 and January 2001. Clinical outcome measures and radiology The Harris hip score (HHS) was determined preoperatively and at follow-up visits 3, 6, 12, 24, and 72 months after surgery. In all patients, follow-up included anteroposterior and lateral radiographs of the hip. Analysis of ion concentrations in serum Samples of serum were collected from patients preoperatively and at 3, 6, 12, 24, and 72 months, in a standardized manner by a registered nurse using an 18-gauge intravenous cannula with the central stainless steel needle removed. The first 5 mL of blood drawn from each patient was discarded. The containers used were 10-mL polypropylene tubes with sodium heparin (Teklab Ltd., Durham, UK). One batch of cannulas and tubes was used throughout the study. All materials used to collect and store samples were chosen for their lack of metals investigated in this study. The whole blood samples were centrifuged at 2,000 rpm for 10 min; the serum was then transferred to acid-washed polypropylene storage tubes using disposable plastic pipettes and stored at –20°C until further analysis. All samples were subsequently sent to an accredited external laboratory specialized in trace metal analysis and analyzed for concentrations of cobalt and chromium by high-resolution inductively coupled plasma mass spectrometry using a commercially available device (Element; Finnigan-MAT, Bremen, Germany). The detection limit for chromium was 0.2 µg/L (4 nMol/L) and that for cobalt was 0.05 µg/L (0.8 nMol/L). To facilitate comparisons with previous literature, ion concentrations are presented as mean concentrations in µg/L with 95% confidence intervals (CIs). Immunological analysis Leukocyte subpopulations were investigated by flow cytometry on a Coulter EPICS-XL-MCL using commercially available antibodies (CD3, CD8, CD45, CD4: Tetracrome Beckman Coulter, Bromma, Sweden cat. no. 6607013; CD3, CD16/56, CD45, CD19: Tetracrome cat. no. 6607073 and Becton-Dickinson cat. no. 332779; HLA DR: Becton-Dickinson cat. no. 6604366; CD4: Becton-Dickinson cat. no. 345768; CD8:

Acta Orthopaedica 2011; 82 (1): 6–12

Dako cat. no. C7079). Results are given as numbers per nL, or as percentages. Immunoglobulins and IgG subclasses were analyzed by nephelometry (DADE Behrings BNll); subclasses were determined using a commercially available kit (The Binding Site Group Ltd., Birmingham, UK) and the results are expressed in g/L. Statistics Before the study, a power analysis indicated that 20 patients per group would be sufficient to detect differences in metal ion concentrations (our primary endpoint) of one standard deviation with a power of 80%, given a two-tailed p-value of 0.05. The investigation of immunological parameters was not planned at the onset of the study, but was deemed to be relevant at a later stage. Thus, the power estimation was not based on this secondary endpoint. In addition, a large number of patients had to be excluded from follow-up due to the implantation of a contralateral THA, which rendered analysis of metal ion concentrations meaningless. All variables were summarized using standard descriptive statistics such as frequencies, means, medians, and standard deviations. The distributions of metal ion concentrations were severely positively skewed; i.e., most subjects were found at the lower ends of concentration ranges, contrasted by outliers with high concentrations. By calculation of the natural logarithms of metal ion concentrations, they were transformed to normal distributions and metal ion concentrations were therefore subsequently analyzed using a linear model with cobalt or chromium concentrations after 6 years as the dependent variable, the type of bearing as the fixed factor, and preoperative cobalt or chromium concentrations as a covariate. Model residuals were normally distributed, and Cook’s distances were < 1. Most immunological parameters were normally distributed, but baseline data were not available for these parameters; thus, inter-group comparisons of immunological parameters were performed using the independent t-test. Correlation analysis of logarithmically transformed metal ion concentrations with immunological parameters was performed using Pearson’s correlation coefficient. The level of significance (two-tailed) was 0.05 in all analyses.

Results Clinical and radiographic outcome The two groups of patients did not differ with respect to age, sex, or BMI. Preoperatively, the HHS was 39 (CI: 32–46) in the metal-on-polyethylene group and 36 (CI: 31–41) in the metalon-metal group. At 6 years, the HHS was 94 (CI: 89–100) in the metal-on-polyethylene group and 90 (CI: 82–98) in the metal-on-metal group. The radiographs obtained at follow-up did not show signs of liner wear, periprosthetic osteolysis, or implant loosening in any patient.


Acta Orthopaedica 2011; 82 (1): 6–12

9

Mean Cr concentration (µg/L) and 95% CI 2.0

Mean Co concentration (µg/L) and 95% CI 1.2

Metal-on-polyethylene Metal-on-metal

Metal-on-polyethylene Metal-on-metal

1.0

1.5

0.8 1.0

0.6 0.4

0.5 0.2

0

Preoperatively

0

6 years postop.

Preoperatively

6 years postop.

Figure 2. Chromium concentrations in serum prior to and 6 years after hip arthroplasty. Mean and 95% confidence intervals of chromium concentrations (in µg/L).

Figure 3. Cobalt concentrations in serum prior to and 6 years after hip arthroplasty. Mean and 95% confidence intervals of cobalt concentrations (in µg/L).

Concentrations of chromium and cobalt were significantly elevated 6 years after metal-on-metal arthroplasty Preoperatively, the concentration of chromium was low and mostly below the detection limits in both groups (Figure 2). In the metal-on-polyethylene group, the mean chromium concentration was 0.31 µg/L (CI: 0.20–0.42), and it was 0.32 µg/L (CI: 0.22–0.42) in the metal-on-metal group. Postoperatively, the chromium concentrations increased substantially in the metal-on-metal group with a mean chromium concentration of 1.05 µg/L (CI: 0.53–1.6) after 6 years. In the metalon-polyethylene group, the mean chromium concentration postoperatively showed no increase; after 6 years it was 0.36 µg/L (CI: 0.15–0.58). The chromium concentrations measured in the metal-on-metal group were substantially higher after 6 years than those measured in the metal-on-polyethylene group (p < 0.001). Baseline chromium concentrations made a minor but statistically significant contribution to the linear model (p = 0.05). The concentration of cobalt was low preoperatively in both groups (Figure 3). The mean cobalt concentration was 0.16 µg/L (CI: 0.1–0.23) in the metal-on-polyethylene group and 0.09 µg/L (CI: 0.06–0.12) in the metal-on-metal group. 6 years after the index procedure, the mean concentration of cobalt showed a robust increase to 0.86 µg/L (CI: 0.49–1.22) in the metal-on-metal group. In the metal-on-polyethylene group, no relevant increase occurred (0.24 µg/L (CI: 0.12–0.36)). After 6 years, cobalt concentrations were higher in the metal-on-metal group than in the metal-on-polyethylene group (p < 0.001). Baseline cobalt concentrations made no statistically significant contribution to the linear model (p = 0.9).

Table 1. Cellular immunological parameters. Mean differences in numbers of leukocytes, lymphocytes, or lymphocyte subpopulations per nL, or percentages of lymphocytes, between the two bearing groups. Positive differences indicate higher values in the metal-on-metal group; negative differences indicate lower values in the metal-on-metal group

Number of HLA DR+ CD8+ T-cells The number of leukocytes per nL was similar in the 2 groups

Mean difference

Leukocytes Lymphocytes T-cells CD3 T-cells CD3 (%) T-cells CD4 T-cells CD4 (%) T-cells CD4 HLA DR+ T-cells CD8 T-cells CD8 (%) T-cells CD8 HLA DR+ CD4/CD8 ratio B-cells CD19 B-cells CD19 (%) NK-cells CD16/CD56 NK-cells CD16/CD56 (%) a

–0.4 0.1 0.0 0.8 0.0 –3.1 1.5 0.1 5.0 3.9 –1.9 0.0 –3.7 0.1 3.2

95% confidence interval –1.3–0.5 –0.4–0.5 –0.4–0.4 –4.6–6.2 –0.3–0.2 –9.5–3.3 –1.3–4.2 –0.1–0.3 –1.6–11.5 0.3–7.4 –4.8–1.1 –0.1–0.0 –6.5– -0.9 0.0–0.2 –1.8–8.3

p-value a 0.4 0.8 0.9 0.8 0.7 0.3 0.3 0.4 0.1 0.03 0.2 0.2 0.01 0.1 0.2

derived from independent t-test.

of bearing types: 6.8 in the metal-on-polyethylene group and 6.4 in the metal-on-metal group (Table 1). The mean number of lymphocytes per nL was 1.9 in both groups. The concentrations of different subsets of lymphocytes (CD3+ T-cells, CD19+ B-cells, and CD16+/CD56+ NK-cells) were also similar in the 2 groups. However, a lower proportion of B-cells was found in the metal-on-metal group (9.3% vs. 12.9% in the metal-on-polyethylene group; p = 0.01). T-cells were further differentiated into CD4+ T-(helper) cells and CD8+ T-(cytotoxic) cells, but there were no statistically significant differences between the two bearing groups. The


10

Acta Orthopaedica 2011; 82 (1): 6–12

Percentage CD8 HLA DR+ T-cells

Percentage CD8 HLA DR+ T-cells

20

20

15

15

10

10

5

5

0

0 -2

A

-1

0

1

2

Ln Cr

Ln Cr 2

1

0

-1

-2 -4

-3

-2

B

-1

0

1

2

Ln Co

-4

C

-3

-2

-1

0

1

2

Ln Co

Figure 4. Correlation between the percentage of HLA DR+ CD8+ T-cells and metal ion concentrations after 6 years. A. A positive correlation was found between the percentage of HLA DR+ CD8+ T-cells and logarithmically transformed chromium concentrations after 6 years (Pearson’s correlation coefficient: 0.39, p = 0.02). B. A positive correlation was also found between the percentage of HLA DR+ CD8+ T-cells and logarithmically transformed cobalt concentrations after 6 years (Pearson’s correlation coefficient: 0.36, p = 0.03). C. Logarithmically transformed cobalt and chromium concentrations were almost linearly correlated with each other (Pearson’s correlation coefficient: 0.85, p < 0.001).

mean CD4/CD8 ratio was 4.7 in the metal-on-polyethylene group and 2.8 in the metal-on-polyethylene group. There were 5.1% HLA DR+ CD4+ T-cells in the metal-on-polyethylene group, as compared to 6.5% in the metal-on-metal group. A statistically significant difference between the two bearing groups was, however, found in the population of CD8+ T-cells positive for the HLA DR-antigen: 6.7% CD8+ T-cells were HLA DR+ in the metal-on-polyethylene group, as compared to 10.6% in the metal-on-metal group (p = 0.03). A separate statistical analysis was done after stratifying the population for sex, but no significant differences between groups were found (data not shown). Correlation of HLA DR+ T-cells with chromium concentrations Correlation analysis showed that there was a positive correlation between the percentage of HLA DR+ CD8+ T-cells and logarithmically transformed chromium and cobalt concentrations after 6 years (Figure 4 and Table 2). We also found a negative correlation between the CD4/CD8 ratio and logarithmically transformed cobalt concentrations, but no other cellular immunological parameters showed any statistically significant correlation with the concentration of either ion. Logarithmically transformed cobalt and chromium concentrations after 6 years showed a strong positive correlation with each other (Figure 4, Table 2; supplementary data). No effects of elevated metal ion concentrations on the humoral immune response The concentrations of the immunoglobulins IgG (and subgroups 1–4), IgM, and IgA were similar in both groups of patients (Table 3; see supplementary data). Furthermore, there were no correlations between immunoglobulin concentrations

and logarithmically transformed cobalt or chromium concentrations (Table 4; see supplementary data).

Discussion The main findings of our study are that: (1) serum concentrations of chromium and cobalt were elevated 6 years after metal-on-metal THA; (2) the proportion of HLA DR+ CD8+ T-cells was higher in patients who underwent metal-on-metal THA; (3) there was a positive correlation between the proportion of HLA DR+ CD8+ T-cells and both chromium and cobalt concentrations; and (4) the percentage of B-cells was reduced in the metal-on-metal group. Increased proportion of HLA DR+ CD8+ T-cells after metal ion exposure Our finding that the percentage of HLA DR+ CD8+ T-cells was higher in patients with a metal-on-metal bearing has not been reported before. Against the background of the previously described reduction in the total population of CD8+ T-cells (Hart et al. 2006, 2009), a decrease in the number of HLA DR+ CD8+ T-cells could have been anticipated. However, in contrast to the data described by Hart et al., we found no reduction in the number of CD8+ T-cells in patients who underwent metal-on-metal THA. This discrepancy could be explained by differences between the various study settings. Our trial was randomized, the two groups of patients in our study were matched for age and sex, and there were no differences in BMI or time to follow-up between the two groups. In contrast, Hart et al. performed a retrospective review of a heterogeneous population; the groups investigated differed in sex distribution, age, BMI, and length of follow-up. The findings


Acta Orthopaedica 2011; 82 (1): 6–12

by Hart et al. are also contradicted by reports of an increased proportion of CD8+ T-cells and CD16+ NK cells in the peripheral blood of patients undergoing revision arthroplasty due to aseptic loosening (Case et al. 2000). Immunological reactions around arthroplasty implants Several reports have described predominantly lymphocytic infiltrates in the neo-capsule of revised metal-on-metal THAs, consisting mainly of CD3+ T-cells and CD20+ B-cells, but also containing considerable numbers of macrophages (Davies et al. 2005, Willert et al. 2005, Witzleb et al. 2007). This type of infiltrate appears to be rather specific for metal-on-metal bearings; it is quite distinct from infiltrates that can be seen around metal-on-polyethylene THAs or in situations involving conventional deep infection. In addition, expression of interferon-γ and the presence of activated T-cells have been described in the pseudomembrane surrounding implants (Weyand et al. 1998). An immunohistochemical analysis of loose orthopedic implants with a cobalt-chromium component revealed an elevated level of HLA DR+ macrophages and enrichment of CD4+ and CD8+ T-cells at the bone-implant interface (al Saffar and Revell 1994). It has also been reported that women with a metal-on-metal resurfacing arthroplasty are predisposed to development of periprosthetic pseudotumors with extensive connective tissue necrosis and mainly lymphocytic infiltration (Pandit et al. 2008). We therefore stratified our study population by sex but found no differences between the sexes. Thus, our findings cannot explain the observed predisposition of women regarding this phenomenon. Sensitization to heavy metal ions after THA Sensitization to various heavy metals after THA was already described after the use of first-generation metal-on-metal arthroplasties during the 1970s, but even today it “… remains uncertain whether the loosening causes the sensitization or vice versa” (Elves et al. 1975). Even with the use of modern metal-on-metal THA, immunological reactions to heavy metal ions are known to occur, as increased lymphocytic reactivity to both cobalt and nickel has been described in patients with metal-on-metal articulations (Hallab et al. 2004). Furthermore, the lymphocytes of patients reacting to chromium more readily release interferon-γ (Hallab et al. 2008). The mechanisms behind the immunological reactions after metal-on-metal THA are not yet understood. Macrophages at the prosthesis/tissue interface are considered to present haptenic metal ions in the context of self peptides (metal ionaltered self) as antigens in combination with MHC class II molecules, leading to T-cell priming. The mediation of metal sensitization via T-cells has been suggested by several authors: T-cell infiltrates at the implant interface suggest an association with type IV, cell-mediated sensitivity (Lalor et al. 1990), and CD3- and CD28-related signaling was observed in T-cells

11

after exposure to cobalt/chromium-particles in vitro (Ogunwale et al. 2009). There was a correlation between the percentage of HLA DR+ CD8+ T-cells in our study and the concentration of chromium and cobalt, although the correlation was weak. We therefore speculate that heavy metal ions released from the implant may be processed on HLA DR+ CD8+ T-cells—either like superantigens by forming and stabilizing bridges between the T-cell receptor and MHC class II molecules, or as neoantigens created by metal-protein interactions. We did not analyze cytokine release and cellular activation markers, however. Limitations of the study The analysis of immunological parameters was added as a secondary endpoint after patient recruitment had been completed, and was not included in the primary study protocol. The present study, based on only 36 individuals eligible for immunological analysis, was therefore open to type-II error. At the same time, small cohorts are vulnerable to outliers, and type-I errors can therefore also occur. Approximately equal numbers were lost to follow-up or were excluded from final analysis in both study groups; we therefore believe that there was no selection bias. Taken together, our findings must be regarded as being derived from a pilot study on a small cohort of patients. Conclusion Our study indicates that elevated chromium and cobalt concentrations after metal-on-metal THA correlate with immunological reactions, reflected by an elevation in HLA DR+ CD8+ T-cells and a decrease in the percentage of B-cells. We currently have no evidence that the elevation of this subtype of lymphocytes is the cause of allergic reactions, aseptic loosening, or periprosthetic pseudotumors that have been described after metal-on-metal arthroplasty. Future studies investigating a possible association between these phenomena must be performed on larger groups of patients.

The study was conceived by AS who also performed all surgeries and interpreted data. RAB was involved in data interpretation. HD performed the follow-up investigations. NPH performed the statistical analyses, interpreted data, and drafted the manuscript. All authors were involved in writing of the final version.

Financial support was received from the Sven Norén Bequest Fund, a charity foundation with no relation to the manufacturer of the implants used in this study.

Supplementary data Tables 2, 3 and 4 are available at our website (www.actaorthop.org), identification number 4050.


12

al Saffar N, Revell P A. Interleukin-1 production by activated macrophages surrounding loosened orthopaedic implants: a potential role in osteolysis. Br J Rheumatol 1994; 33 (4): 309-16. Case C P, Langkamer V G, Lock R J, Perry M J, Palmer M R, Kemp A J. Changes in the proportions of peripheral blood lymphocytes in patients with worn implants. J Bone Joint Surg (Br) 2000; 82 (5): 748-54. Dahlstrand H, Stark A, Anissian L, Hailer N P. Elevated serum concentrations of cobalt, chromium, nickel, and manganese after metal-on-metal alloarthroplasty of the hip: A prospective randomized study. J Arthroplasty 2009; 24 (6): 837-45. Davies A P, Willert H G, Campbell P A, Learmonth I D, Case C P. An unusual lymphocytic perivascular infiltration in tissues around contemporary metalon-metal joint replacements. J Bone Joint Surg (Am) 2005; 87 (1): 18-27. Elves M W, Wilson J N, Scales J T, Kemp H B. Incidence of metal sensitivity in patients with total joint replacements. Br Med J 1975; 4 (5993): 376-8. Hallab N J, Anderson S, Caicedo M, Skipor A, Campbell P, Jacobs J J. Immune responses correlate with serum-metal in metal-on-metal hip arthroplasty. J Arthroplasty (Suppl 3) 2004; 19 (8): 88-93. Hallab N J, Caicedo M, Finnegan A, Jacobs J J. Th1 type lymphocyte reactivity to metals in patients with total hip arthroplasty. J Orthop Surg 2008; 3: 6. Hart A J, Hester T, Sinclair K, Powell J J, Goodship A E, Pele L, Fersht N L, Skinner J. The association between metal ions from hip resurfacing and reduced T-cell counts. J Bone Joint Surg (Br) 2006; 88 (4): 449-54. Hart A J, Skinner J A, Winship P, Faria N, Kulinskaya E, Webster D, Muirhead-Allwood S, Aldam C H, Anwar H, Powell J J. Circulating levels of cobalt and chromium from metal-on-metal hip replacement are associated with CD8+ T-cell lymphopenia. J Bone Joint Surg (Br) 2009; 91 (6): 83542.

Acta Orthopaedica 2011; 82 (1): 6â&#x20AC;&#x201C;12

Lalor P A, Gray A B, Wright S, Railton G T, Freeman M A, Revell P A. Contact sensitivity to titanium in a hip prosthesis? Contact Dermatitis 1990; 23 (3): 193-4. Lazarinis S, Milbrink J, Hailer N P. Avascular necrosis and subsequent femoral neck fracture 3.5 years after hip resurfacing: a highly unusual late complication in the absence of risk factors-a case report. Acta Orthop 2008; 79 (6): 763-8. Ogunwale B, Schmidt-Ott A, Meek R M, Brewer J M. Investigating the immunologic effects of CoCr nanoparticles. Clin Orthop 2009; (467) (11): 3010-6. Pandit H, Glyn-Jones S, McLardy-Smith P, Gundle R, Whitwell D, Gibbons C L, Ostlere S, Athanasou N, Gill H S, Murray D W. Pseudotumours associated with metal-on-metal hip resurfacings. J Bone Joint Surg (Br) 2008; 90 (7): 847-51. Toms A P, Marshall T J, Cahir J, Darrah C, Nolan J, Donell S T, Barker T, Tucker J K. MRI of early symptomatic metal-on-metal total hip arthroplasty: a retrospective review of radiological findings in 20 hips. Clin Radiol 2008; 63 (1): 49-58. Weyand C M, Geisler A, Brack A, Bolander M E, Goronzy J J. Oligoclonal T-cell proliferation and interferon-gamma production in periprosthetic inflammation. Lab Invest 1998; 78 (6): 677-85. Willert H G, Buchhorn G H, Fayyazi A, Flury R, Windler M, Koster G, Lohmann C H. Metal-on-metal bearings and hypersensitivity in patients with artificial hip joints. A clinical and histomorphological study. J Bone Joint Surg (Am) 2005; 87 (1): 28-36. Witzleb W C, Hanisch U, Kolar N, Krummenauer F, Guenther K P. Neocapsule tissue reactions in metal-on-metal hip arthroplasty. Acta Orthop 2007; 78 (2): 211-20.


Acta Orthopaedica 2011; 82 (1): 13–19

13

A prospective randomized study comparing electrochemically deposited hydroxyapatite and plasma-sprayed hydroxyapatite on titanium stems 55 hips followed for 2 years with RSA and DXA Berte Grimsmo Bøe1,2, Stephan M Röhrl1, Tore Heier3, Finnur Snorrason1, and Lars Nordsletten1,2 1Department

of Orthopaedics, Oslo University Hospital, Ullevål; 2Faculty of Medicine, University of Oslo; and 3Department of Surgery, Diakonhjemmet Hospital, Oslo, Norway Correspondence BB: berte2@mac.com Submitted 10-03-08. Accepted 10-07-15

Background and purpose Plasma-sprayed hydroxyapatite (HA) is a successful coating for fixation of uncemented femoral stems. There may be alternative coatings with advantages in bone remodeling and transport of bone-active substances. We investigated whether an electrochemically deposited hydroxyapatite, Bonemaster (BM), might be a safe alternative in total hip arthroplasty. Our hypothesis was that the new coating would not be inferior to the conventional one. Patients and methods 50 patients (55 hips) were included. The stem was tapered and porous-coated proximally. On top of the porous coating was either HA or BM. Patients were evaluated postoperatively and after 3, 6, 12, and 24 months to measure fixation by radiostereometric analysis (RSA), bone mineral density by dual-energy X-ray absorptiometry (DXA), and conventional radiography. Clinical evaluation was performed with Harris hip score and Oxford hip score, both preoperatively and after 2 years. Results After 2 years, the stems had subsided 0.25 (HA) and 0.28 (BM) mm and there were no statistically significant differences between the groups in any direction, regarding both migration and rotation. The BM group retained significantly more bone than the HA group in Gruen zone 1 during the first 2 years. The Harris and Oxford hip scores were similar in both groups. Interpretation Electrochemically deposited hydroxyapatite on an uncemented stem does not appear to be inferior to plasmasprayed HA regarding clinical and radiological results, bone remodeling, and micromotion after 2 years follow-up. 

Aseptic loosening is the most frequent complication of total hip arthroplasty (THA) (Havelin et al. 2000). The long-term survival is thought to depend partly on bone loss or osteol-

ysis in the proximal femur after insertion. Plasma-sprayed hydroxyapatite (HA) coatings appear to give effective fixation in the femur (Hallan et al. 2007). Alternative coatings may, however, influence bone remodeling around the prosthesis and may function as a carrier of bone-active substances. Bonemaster (BM) (Bonemaster is a registered trademark of Biomet Europe) is an electrochemically deposited hydroxyapatite (EDHA) coating (Rößler et al. 2002). This technique makes it possible to add biological substrates such as antibiotics or adhesion peptides to the coating and still keep the coating very thin compared to plasma-sprayed HA. The thickness of a hydroxyapatite coating is a compromise between the mechanical properties and dissolution of the coating. A thinner coating minimizes the potential of particle shedding during insertion. Fewer particles in the joint mean less third-body wear and less periprosthetic osteolysis (Peters et al. 1992, Shanbhag et al. 1994, Campbell et al. 1995, ­McKellop et al. 1995). Thinner coatings also lower the risk of HA delamination and preserve the porosity of the underlying metallic coating of the implant. The irregular implant surface increases the surface area, providing a greater contact and ingrowth area (Sewing et al. 2002). EDHA, as in Bonemaster, forms a needle-like porous structure (Rößler et al. 2001) and enhances early-stage fixation between implant and bone (Ban et al. 1997). We designed a prospective randomized trial to compare conventional plasma-sprayed HA with electrochemically deposited HA after insertion of an uncemented femoral stem. This is the first clinical trial with the Bonemaster coating. We hypothesized that implants with Bonemaster would achieve the same degree of stability and bone remodeling, and the same clinical outcome as implants with traditionally plasma-sprayed HA.

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2010.548027


14

Figure 1. Taperloc stem coated with Bonemaster.

Patients and methods 50 patients (31 of whom were women; 55 hips) with noninflammatory end-stage osteoarthritis of the hip participated. Inclusion criteria were health condition expected to allow follow-up for 10 years and anatomy compatible with use of a standard implant. Exclusion criteria were infection, revision arthroplasty, marked bone loss, and severe morbidity. Mean age at the time of operation was 63 (27–81) years. From December 2003 through June 2005, patients underwent THA with the Taperloc uncemented stem (Figure 1), a 28-mm cobalt-chrome modular head and the SHP cemented cup. (The stems were manufactured by Biomet UK Healthcare Ltd.; all other components were from Biomet, Warsaw, IN). Both hips of 5 patients were included. Recruitment was by informed consent and the patients were on our waiting list for THA. The Norwegian Data Inspectorate and the regional ethical committee approved the study and it was carried out in line with the Helsinki declaration. The patients were randomized (with sealed envelopes) to a stem with either plasma-sprayed HA or Bonemaster. 31 hips were operated with BM-coated stems and 24 with plasmasprayed HA-coated stems. After a power analysis performed during the study, we ended the inclusion after recruiting 55 hips, leaving 45 envelopes unopened. 1 patient was excluded because of a periprosthetic fracture. 2 patients have subsequently been revised because of loose cups. These 2 patients have been followed with measurements of the stem after their revisions. 1 patient was reoperated after 5 weeks with soft tissue revision and change of femoral head because of infection, but was kept in the study. 4 orthopedic consultants in 2 hospitals operated the patients using the modified Hardinge approach.

Acta Orthopaedica 2011; 82 (1): 13–19

The Taperloc stem was manufactured from forged titanium alloy, Ti-6Al-4V. It had a tapered form and was porouscoated proximally. On top of the porous coating, the stem was coated with either plasma-sprayed HA or electrochemically deposited hydroxyapatite (BM). Plasma spraying of HA is a high-temperature process designed to deliver slightly molten Ca(PO4)2 granules of μm size onto metal surfaces. The process was first described by de Groot et al. (1987). The specifications for the implants in this study were according to the manufacturer: 50-micron thick HA coating (Ca/P ratio: 1.67), a mean surface roughness of 41 microns, a maximum roughness depth of 445 microns, and 62% crystallinity. The electrochemically deposited hydroxyapatite coating was performed in an electrolyte solution near physiological conditions (pH 6.4, 37°C), consisting of 1.67 mM CaCl2 and 1 mM NH4(H2PO4) in equal volumes with the implant polarized in cathode galvanostatic mode (–75 A/m2). The layer consisted of 70–72% crystalline HA with the balance being amorphous, and with a thickness of 5 μm and a Ca/P ratio of 2.0. The time taken to apply this form of HA coating is much slower than that of HA applied by plasma spray, and is typically 75 min per implant. Details of preparations, characteristics, and appearances of coatings are as described by Rößler et al. (2002) and Sewing et al. (2004). On the acetabular side, we used SHP—an all-poly gammairradiated (ArCom) cemented cup—inserted with Palacos (Schering-Plough) gentamycin-containing cement. Bone mineral density (BMD) was measured by experienced technicians using DXA. 3 different DXA machines were used (Prodigy and Expert; both from Lunar, Madison, WI—and Hologic QDR; Hologic Inc., Bedford, MA). The patient was placed supine on the scan table with a foot support to achieve standard rotation of the hip. Orthopedic software (Lunar version 1.2 and Hologic QDR version 12.3) was used to analyze periprosthetic BMD in 7 regions of interest (ROIs). The ROIs were based on the Gruen zones. The patients were measured within a few days postoperatively (mean 5.8 days) and after 3, 6, 12, and 24 months. 30 patients treated at “hospital A” were measured with a Lunar Expert densitometer until January 1, 2005 and later with a Hologic densitometer. 25 patients treated at “hospital B” were measured with a Lunar Prodigy densitometer. We calculated a transformation formula between Lunar Expert and Hologic values based on measurements from 5 of the patients included. These 5 patients were measured twice on both densitometers, and on the same day. Assuming linearity between the 2 machines, the best fit was found using the formula BMDLunar = 0.789 × (BMDHologic) + 0.2089. Because this transformation of values represents a bias in the patient group from hospital A, we performed statistical analysis on the total number of patients, and on the patients from hospital B separately. To calculate precision error (the coefficient of variation, CV%) of the 3 densitometers, 130 examinations were repeated on the same day, with repositioning between the scans (Table 1).


Acta Orthopaedica 2011; 82 (1): 13â&#x20AC;&#x201C;19

15

Table 1. Coefficient of variation (CV) in per cent for the 3 densiometers used in 7 Gruen zones Zone: Expert Hologic Prodigy

1

2

3

4

5

6

7

2.0 0.5 2.6

1.5 5.8 2.1

2.0 1.5 1.7

1.4 0.6 3.6

2.3 1.5 2.2

2.3 1.1 5.1

2.8 0.8 2.9

100 hips planned

45 envelopes not opened

55 hips randomized

31 BM

DXA

24 HA

RSA

DXA

1 without postoperative examination

RSA During the operation 7-8 tantalum markers of 1.0 mm were inserted into the proximal part of the femur. The manufacturer had attached 3 tantalum markers to the femoral stem, 1 to the shoulder, 1 to the neck and 1 to the tip. We also computed the position of the femoral head centre. Radiostereometric examinations were done at approximately 7 days, 3 months, 6 months, 1 year, and 2 years after the operation. We evaluated migration of the gravitational centre of the segment which was defined by the stem markers and the centre of the femoral head. Migration was measured along the cardinal axes. Stem rotations were measured as rotations of that segment around the same axes. Analyses were performed using UmRSA (Digital measurement 6.0 RSA Biomedical, UmeĂĽ, Sweden). In 4 cases the quality of the postoperative stereoradiographs did not allow a proper evaluation. 1 patient was lost to follow up, 2 did not meet and 1 patient was excluded from the 2 years analysis because of high condition number. 47 patients were to be analyzed at 2 years with mean error below 0.3 and condition number below 100 (Figure 2). 83 examinations were repeated the same day to calculate the precision of our measureRSA ments. RSA results are presented as mean 3 without values with standard error of mean (SEM).

postoperative examination

3 months

3 months

3 months

3 months

30 measured 1 did not attend

23 analyzed 2 not analyzed 5 did not attend

20 measured 1 lost to follow-up 3 did not attend

16 analyzed 1 lost to follow-up 2 not analyzed 2 did not attend

6 months

6 months

6 months

6 months

29 measured 2 did not attend

23 analyzed 1 not analyzed 6 did not attend

22 measured 1 lost to follow-up 1 did not attend

16 analyzed 1 lost to follow-up 2 not analyzed 2 did not attend

12 months

12 months

12 months

31 measured

26 analyzed 2 not analyzed 2 did not attend

22 measured 1 lost to follow-up 1 did not attend

24 months

24 months

24 months

30 measured 1 did not attend

27 analyzed 1 not analyzed 2 did not attend

22 measured 1 lost to follow-up 1 did not attend

12 months 19 analyzed 1 lost to follow-up 1 did not attend

24 months 20 analyzed 1 lost to follow-up

Figure 2. Flow chart of the patients with DXA and RSA measurements. The patient lost to follow-up had an HA-coated stem inserted at hospital A. 17 BM-coated stems were inserted at hospital A and 14 BM-coated stems were inserted at hospital B. 13 HAcoated stems were inserted at hospital A and 11 HA-coated stems were inserted at hospital B. Those patients who were not analyzed by RSA did not meet the criteria of a maximum condition number of 100 or mean error of less than 0.3.

Conventional radiography Anteroposterior and lateral examinations were done preoperatively, postoperatively, and after 3, 6, 12, and 24 months. 2 surgeons evaluated the radiographs. The parameters registered were implant position, radiolucency or lysis, calcar resorption, heterotopic bone formation, trabecular remodeling, cyst formation, pedestal formation, and visible migration. Radiolucent lines were considered to be present if they were > 1.0 mm and occupied more than 50% of the interface in each Gruen zone. Clinical evaluation Harris hip score (Harris 1969) and Oxford hip score (Dawson et al. 1996) were evaluated preoperatively and after 2 years. Statistics The statistical analysis of BMD and RSA results to compare Bonemaster to HA was done using non-inferiority testing comparing areas (SPSS for Mac version PASW 18.0). The BMD results were expressed as percentages (mean) with standard deviation (SD). RSA results were signed values (mean) with


16

Acta Orthopaedica 2011; 82 (1): 13–19

Table 2. Periprosthetic changes in bone mineral density around hydroxyapatite- (HA-) and Bonemaster- (BM-) coated Taperloc stems measured by dual-energy X-ray absorbtiometry (DXA). Results are given in percentage (standard deviation) of postoperative values after 3 months, 6 months, 1 year, and 2 years Gruen zone

n 3 months

n

6 months

n

1 year

n

2 year

1 Bonemaster HA 2 Bonemaster HA 3 Bonemaster HA 4 Bonemaster HA 5 Bonemaster HA 6 Bonemaster HA 7 Bonemaster HA

30 20 30 20 30 20 30 20 30 20 30 20 30 20

29 22 29 22 29 22 29 22 29 22 29 22 29 22

94 (25) 79 (13) 92 (8) 90 (8) 96 (6) 96 (5) 95 (6) 95 (4) 97 (7) 97 (6) 94 (8) 86 (23) 77 (14) 77 (11)

31 22 31 22 31 22 31 22 31 22 31 22 31 22

90 (26) 76 (15) 90 (10) 89 (9) 94 (8) 95 (7) 94 (5) 94 (6) 96 (9) 93 (10) 94 (9) 90 (10) 73 (15) 73 (12)

30 22 30 22 30 22 30 22 30 22 30 22 30 22

87 (19) 79 (18) 87 (14) 86 (10) 90 (12) 92 (9) 90 (9) 90 (7) 90 (11) 88 (12) 89 (11) 88 (9) 70 (16) 69 (13)

92 (19) 83 (14) 89 (10) 90 (10) 93 (7) 94 (5) 95 (5) 96 (5) 94 (7) 97 (5) 93 (9) 94 (7) 82 (13) 83 (9)

SEM. The DXA and RSA results were not normally distributed. The 2 groups were compared with the Mann-Whitney U-test. Changes in DXA, postoperatively to the 2-year followup, were analyzed with Wilcoxon signed rank test. A power analysis was not performed before the study started. Based on an estimated clinically important difference in BMD of 10% (SD 10), stem migration of 0.6 mm (SD 0.6), and stem rotation of 0.7 degrees (SD 0.7) between groups, the sample size calculation indicated 17 patients would be required in each group to achieve 80% power at the 0.05 significance level. Due to the risk of patient dropout, at least 24 patients were included in each group.

Results DXA After 2 years, there was bone loss compared to the postoperative values in both groups in all regions around the stem (p < 0.05), which was most pronounced in Gruen zones 1 and 7 (Table 2). With non-inferiority testing between the 2 groups, we had to reject the null hypothesis (that Taperloc would function equally well with Bonemaster and HA) for zone 1. Comparison of the areas under the graph showed a significant difference between HA and BM in zone 1 after 2 years (p = 0.01). The bone loss was less in the Bonemaster group. Because of the possible bias with the transformation formula used in hospital A, we also performed the analyses with the results from hospital B alone. We found the same as for the whole group: rejection of the null hypothesis in zone 1 (p = 0.01). For all other Gruen zones, the null hypothesis was retained. RSA The precision of our measurements was 0.11 mm for subsid-

ence and 0.66 degrees for retroversion. The migration pattern for both stems showed that they moved during the first 3 months after surgery and then stabilized (Figures 3 and 4). The mean (SD) subsidence for the center of the stem after 2 years was 0.28 (0.47) mm for BM and 0.25 (0.69) mm for HA (Figure 3). The stems with Bonemaster moved mean 0.46 (0.73) degrees in retroversion compared to 0.17 (0.83) degrees for the HA-coated stems (p = 0.2) (Figure 4). Both groups had retroversion that was lower than the precision in this direction (0.66). With non-inferiority testing, there were no significant differences between groups in any migration or rotation after 2 years. Clinical results Harris hip score increased from 55 (pain score 20) preoperatively to 95 (pain score 41) after 2 years in the Bonemaster group, which was almost similar to the increase from 52 (18) to 89 (38) in the HA group. Oxford hip score improved from 39 preoperatively to 16 after 2 years in the BM group and from 35 to 19 in the HA group. The differences after 2 years were not statistically significant. There were no radiolucent lines around the stems after 2 years. 1 patient was excluded because of a periprosthetic fracture 6 weeks after the operation. 2 patients in the BM group were revised after 10 and 13 months because of cup loosening. They were both revised because of pain, and we continued to follow the femoral component in the study. Another patient (in the BM group) had a postoperative infection and was revised with soft tissue debridement and irrigation. Treatment of the infection was successful and the patient continued in the study. RSA and BMD results were analyzed both with and without the reoperated patients included, with no significant changes in the results.


Acta Orthopaedica 2011; 82 (1): 13–19

17

Subsidence (mm)

Retroversion (°)

0

1.0

-0.1

0.9

-0.2

0.8

-0.3

0.7

-0.4

0.6

-0.5

0.5

-0.6

0.4

-0.7

0.3

-0.8

0.2 HA BM

-0.9 1.0

HA BM

3

0.1 6

12

24

Months postoperatively

Figure 3. Subsidence (0.25 (HA) and 0.28 (BM) mm in the first 2 years) after implantation of hydroxyapatite- (HA-) and Bonemaster(BM-) coated Taperloc stems, analyzed by radiostereometric analysis (RSA). Results are given in mm with standard error of the mean after 3 months, 6 months, 1 year, and 2 years.

Discussion Our study was started as a safety study of a new electrochemically deposited HA coating. The null hypothesis was that Taperloc would function equally well with Bonemaster and HA. Our results show that Bonemaster-coated stems do not function in an inferior way to stems coated with plasmasprayed HA after 2 years. The most recent publication for the Taperloc stem without any HA coating (McLaughlin and Lee 2008) describes good long-term results for aseptic loosening with 87% survival after 20 years. The Taperloc is therefore a good stem for safety studies of new technology. We already knew that HA coating of a porous surface is an excellent fixation method. Both Karrholm et al. (1994) and Soballe et al. (1993) showed that there is less subsidence with HA-coated stems than with porous-coated stems. The question is therefore “do we need Bonemaster?”. The total number of cementless procedures has increased in recent years. A benefit of a coating placed on the stem by galvanic electrolysis is that it can act as a carrier for other substrates. Infections remain a critical issue in total joint arthroplasty. Antibiotics added to the coating would probably lower the infection rate. Alt et al. (2006) demonstrated a lower infection rate in a combined gentamicin-HA group than in an ordinary HA group in a rabbit model. Osteolysis and bone loss may lead to loosening of the implant or periprosthetic fractures. We suggest that some of the proximal bone loss is due to the surgical trauma from cutting of the neck and preparing the proximal femur to receive the prosthesis. We found less reduction in bone density in the Bonemaster group than in the HA group in Gruen zone 1 during the first 2 years. We found no difference in zones 2, 6, and 7, which are also regions with BM coating. Perhaps the initial difference

0

3

6

12

24

Months postoperatively

Figure 4. Retroversion of hydroxyapatite- (HA-) and Bonemaster- (BM-) coated Taperloc stems after 2 years, analyzed by radiostereometric analysis. Results are in degrees with standard error of the mean.

between the 2 groups was detected in zone 1 because this is an area dominated by trabecular bone. Higher bone density values in the Bonemaster group may indicate a higher degree of bone turnover in trabecular bone in this group. An expected result of DXA measurements after implantation of a hip prosthesis is a marked bone loss initially, and then restoration of bone (Trevisan et al. 1997, Wixson et al. 1997, Karachalios et al. 2004). We have not seen any restoration of bone around the Taperloc stem during the first two years, but bone loss from 7% (zone 3) to 31% (zone 7) is acceptable compared to other stems. It remains to be seen whether the bone loss would continue. For prostheses with satisfactory results, the bone loss is often limited to the proximal zones. In patients with early aseptic loosening, a different pattern of bone remodeling with reduction along the entire stem has been found (Boden et al. 2004). Scott and Jaffe (1996) predicted that higher BMD indicates better ingrowth of bone to the implant. In that case, Bonemaster leads to faster bone ingrowth than HA. This was not reflected by better stability early on, however, as measured by RSA. In an animal study with mechanical pull-out testing, Yang et al. (2008) showed that roughened titanium implants had better initial fixation to bone when they were coated with electrochemically deposited HA than when they only had roughened titanium on the surface. Another experimental study has shown that plasma-sprayed HA accelerates the early-stage mineralization (< 7 days) of bone more than EDHA coating (Wang et al. 2006). However, EDHA appeared to have resulted in better mechanical integration between the coating and mineralized tissue. Plasmasprayed HA and EDHA were indistinguishable later (14 days) regarding the mineralized tissue ratio and microstructure they induced in vivo. Results for the patients measured on the same DXA machine during the entire study were the same as for the whole group.


18

We therefore consider the values to be reliable for comparison, although a change of DXA machine at one institution complicated calculation of bone remodeling. The amount of acceptable subsidence probably varies between different stem designs and fixation methods. Karrholm et al. (1994) showed that the amount of subsidence after 2 years was the best predictor of later revisions in cemented stems. The cut-off values for the probability of revision to exceed 50% and 95% were 1.2 mm and 2.6 mm of subsidence after 2 years. In a study with HP-Garche uncemented stems, Wykman et al. (1988) reported 0.6–3.9 mm of subsidence in 7 of 8 stems after 2 years. In a later report, the same group reported that 13 of 78 HP-Garche stems (17%) had to be revised in less than four years (Wykman et al. 1991). Wykman and Lundberg (Wykman and Lundberg 1992) presented an RSA study of 9 patients with porous-coated Taperloc stems. 3 stems had subsided 0.7–0.9 mm after 2 years, and the mean subsidence after 2 years was 0.44 mm. Compared to earlier studies our results thus indicate that concerning subsidence, HA coating of the Taperloc stem is beneficial. In our trial, the stem subsided in the three first months postoperatively and then stabilized. This migration pattern has also been shown with other HA-coated implants. Thien et al. (2007) published the same pattern of subsidence for 43 ABG stems. For the clinically proven Corail stem, Campbell et al. (2009) documented more subsidence (0.58 mm) than for Taperloc over the first 2 years, using RSA analysis. Corail had the same migration pattern, however, with stability after 6 months. As both Corail and Taperloc have good long-term clinical results (Hallan et al. 2007), it appears that subsidence during the first months and then stabilization within a year is typical for these designs of HA-coated uncemented implants. This might be explained by the ability of hydroxyapatite to close the gap between bone and implant (Overgaard et al. 1997). Long-term follow-up will be necessary to evaluate the effect of electrochemically applied HA on long-term fixation. Further studies are required to investigate whether electrochemically deposited HA combined with antibiotics might lower the infection rate. In conclusion, the Taperloc stem with Bonemaster does not appear to be inferior to the Taperloc stem with plasma-sprayed HA, concerning clinical and radiological results, bone remodeling, and micromotion—at least up to 2 years.

BB did the clinical follow-up, RSA and DXA analysis, wrote the manuscript and performed the statistical evaluation. LN designed the study, operated, and helped with the statistics and writing of the manuscript. SF designed the study with LN, operated, and followed many of the patients. TH included, followed, and operated most of the patients. SR supervised the RSA analysis and helped in writing the manuscript.

Biomet Europe provided financial support, but took no part in the organization of the study or in analysis of the results and writing of the manuscript.

Acta Orthopaedica 2011; 82 (1): 13–19

Alt V, Bitschnau A, Osterling J, Sewing A, Meyer C, Kraus R, et al. The effects of combined gentamicin-hydroxyapatite coating for cementless joint prostheses on the reduction of infection rates in a rabbit infection prophylaxis model. Biomaterials 2006; 27 (26): 4627-34. Ban S, Maruno S, Arimoto N, Harada A, Hasegawa J. Effect of electrochemically deposited apatite coating on bonding of bone to the HA-G-Ti composite and titanium. J Biomed Mater Res 1997; 36 (1) :9-15. Boden H, Adolphson P, Oberg M. Unstable versus stable uncemented femoral stems: a radiological study of periprosthetic bone changes in two types of uncemented stems with different concepts of fixation. Arch Orthop Trauma Surg 2004; 124 (6): 382-92. Campbell P, Ma S, Yeom B, McKellop H, Schmalzried T P, Amstutz H C. Isolation of predominantly submicron-sized UHMWPE wear particles from periprosthetic tissues. J Biomed Mater Res 1995; 29 (1): 127-31. Campbell D, Mercer G, Nilsson K G, Wells V, Field J R, Callary S A. Early migration characteristics of a hydroxyapatite-coated femoral stem: an RSA study. Int Orthop 2009; Epub ahead of print. Dawson J, Fitzpatrick R, Carr A, Murray D. Questionnaire on the perceptions of patients about total hip replacement. J Bone Joint Surg (Br) 1996; 78 (2): 185-90. de Groot K, Geesink R, Klein C P, Serekian P. Plasma sprayed coatings of hydroxylapatite. J Biomed Mater Res.1987; 21 (12): 1375-81. Gruen T A, McNeice G M, Amstutz H C. ”Modes of failure” of cemented stem-type femoral components: a radiographic analysis of loosening. Clin Orthop 1979; (141): 17-27. Hallan G, Lie S A, Furnes O, Engesaeter L B, Vollset S E, Havelin L I. Medium- and long-term performance of 11,516 uncemented primary femoral stems from the Norwegian arthroplasty register. J Bone Joint Surg (Br) 2007; 89 (12): 1574-80. Harris W H. Traumatic arthritis of the hip after dislocation and acetabular fractures: treatment by mold arthroplasty. An end-result study using a new method of result evaluation. J Bone Joint Surg (Am) 1969; 51 (4): 737-55. 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. Karachalios T, Tsatsaronis C, Efraimis G, Papadelis P, Lyritis G, Diakoumopoulos G. The long-term clinical relevance of calcar atrophy caused by stress shielding in total hip arthroplasty: a 10-year, prospective, randomized study. J Arthroplasty 2004; 19 (4): 469-75. Karrholm J, Borssen B, Lowenhielm G, Snorrason F. Does early micromotion of femoral stem prostheses matter? 4-7-year stereoradiographic follow-up of 84 cemented prostheses. J Bone Joint Surg (Br) 1994; 76 (6): 912-7. McKellop H A, Campbell P, Park S H, Schmalzried T P, Grigoris P, Amstutz H C, et al. The origin of submicron polyethylene wear debris in total hip arthroplasty. Clin Orthop 1995; (311): 3-20. McLaughlin J R, Lee K R. Total hip arthroplasty with an uncemented tapered femoral component. J Bone Joint Surg (Am) 2008; 90 (6): 1290-6. Overgaard S, Lind M, Rahbek O, Bunger C, Soballe K. Improved fixation of porous-coated versus grit-blasted surface texture of hydroxyapatite-coated implants in dogs. Acta Orthop Scand 1997; 68 (4): 337-43. Peters P C, Jr., Engh G A, Dwyer K A, Vinh T N. Osteolysis after total knee arthroplasty without cement. J Bone Joint Surg (Am) 1992; 74 (6): 864-76. Rößler S, Born R, Scharnweber D, Worch H, Sewing A, Dard M. Biomimetic coatings functionalized with adhesion peptides for dental implants. J Mater Sci Mater Med 2001; 12 (10-12): 871-7. Rößler S, Sewing A, Sto¨lzel M, Born R, Scharnweber D, Dard M, Worch H. Electrochemically assisted deposition of thin calcium phosphate coatings at near-physiological pH and temperature. J Bone Mineral Res 2002; 03 (64A): 655. Scott D F, Jaffe W L. Host-bone response to porous-coated cobalt-chrome and hydroxyapatite-coated titanium femoral components in hip arthroplasty. Dual-energy x-ray absorptiometry analysis of paired bilateral cases at 5 to 7 years. J Arthroplasty 1996; 11 (4): 429-37.


Acta Orthopaedica 2011; 82 (1): 13â&#x20AC;&#x201C;19

Sewing A, Roessler S, Scharnweber D, Worch H, Schliephake H. Cell culture studies with keratinocytes, fibroblast and osteoblast cell lines on biomimetic coatings on titanium. 17th European Conference on Biomaterials 11-14 September 2002; Barcelona, Spain 2002. Sewing A, Lakatos M, Scharnweber D, Roessler S, Born R, Dard M, Worch H. Influence of Ca/P ratio on electrochemical assisted deposition of hydroxyapatite on titanium. Key Engineering Materials 2004; 254-256: 419-22. Shanbhag A S, Jacobs J J, Glant T T, Gilbert J L, Black J, Galante J O. Composition and morphology of wear debris in failed uncemented total hip replacement. J Bone Joint Surg (Br) 1994; 76 (1): 60-7. Soballe K, Toksvig-Larsen S, Gelineck J, Fruensgaard S, Hansen E S, Ryd L, et al. Migration of hydroxyapatite coated femoral prostheses. A Roentgen Stereophotogrammetric study. J Bone Joint Surg (Br) 1993; 75 (5): 681-7. Thien T M, Ahnfelt L, Eriksson M, Stromberg C, Karrholm J. Immediate weight bearing after uncemented total hip arthroplasty with an anteverted stem: a prospective randomized comparison using radiostereometry. Acta Orthop Scand 2007; 78 (6): 730-8. Trevisan C, Bigoni M, Randelli G, Marinoni E C, Peretti G, Ortolani S. Periprosthetic bone density around fully hydroxyapatite coated femoral stem. Clin Orthop 1997; (340): 109-17.

19

Wang H, Eliaz N, Xiang Z, Hsu H P, Spector M, Hobbs L W. Early bone apposition in vivo on plasma-sprayed and electrochemically deposited hydroxyapatite coatings on titanium alloy. Biomaterials 2006; 27 (23): 4192-203. Wixson R L, Stulberg S D, Van Flandern G J, Puri L. Maintenance of proximal bone mass with an uncemented femoral stem analysis with dual-energy x-ray absorptiometry. J Arthroplasty 1997; 12 (4): 365-72. Wykman A, Lundberg A. Subsidence of porous coated noncemented femoral components in total hip arthroplasty. A roentgen stereophotogrammetric analysis. J Arthroplasty 1992; 7 (2): 197-200. Wykman A, Olsson E, Axdorph G, Goldie I. Total hip arthroplasty. A comparison between cemented and press-fit noncemented fixation. J Arthroplasty 1991; 6 (1): 19-29. Wykman A, Selvik G, Goldie I. Subsidence of the femoral component in the noncemented total hip. A roentgen stereophotogrammetric analysis. Acta Orthop Scand 1988; 59 (6): 635-7. Yang G L, He F M, Hu J A, Wang X X, Zhao S F. Biomechanical comparison of biomimetically and electrochemically deposited hydroxyapatite-coated porous titanium implants. J Oral Maxillofac Surg 2008; 68 (2): 420-7.


20

Acta Orthopaedica 2011; 82 (1): 20–26

The effect of hospital volume on length of stay, re-admissions, and complications of total hip arthroplasty A population-based register analysis of 72 hospitals and 30,266 replacements Keijo T Mäkelä1, Unto Häkkinen2, Mikko Peltola2, Miika Linna2, Heikki Kröger3, and Ville Remes4 1Department of Orthopaedics and Traumatology, Turku University Central Hospital; 2National Institute for Health and Welfare; 3Department of Orthopaedics and Traumatology, Kuopio University Hospital3; 4Department of Orthopaedics and Traumatology, Peijas Hospital, Helsinki University Central Hospital, Finland Correspondence: keijo.makela@tyks.fi Submitted 10-02-08. Accepted 10-07-17

Background and purpose Hospital volume has been suggested to be one of the best indicators of adverse orthopedic events in patients undergoing THR surgery. We therefore evaluated the effect of hospital volume on the length of stay, re-admissions, and complications of THR at the population level in Finland. Methods 30,266 THRs performed for primary osteoarthritis were identified from the Hospital Discharge Register. Hospitals were classified into 4 groups according to the number of THRs performed on an annual basis over the whole study period: 1–50 (group 1), 51–150 (group 2), 151–300 (group 3), and > 300 (group 4). Results In 2005, the length of the period of surgical treatment was 5.5 days in group 4 and 6.8 days in group 1 (the reference group). During the whole study period (1998–2005), the length of surgical treatment period was shorter in group 4 than in group 1 (p < 0.001). The odds ratio for dislocations (0.7, 95% CI: 0.6–0.9) was lower in group 3 than in group 1. Interpretation Hip replacements performed in high-volume hospitals reduce costs by shortening the length of stay, and they may reduce the dislocation rate. 

The association between hospital volume and results of total hip replacement (THR) has been investigated in several studies (Lavernia and Guzman 1995, Battaglia et al. 2006, Doro et al. 2006, Shervin et al. 2007). The surgeon volume and the hospital volume have been suggested to be the best indicators of adverse orthopedic events in patients undergoing THR surgery (Solomon et al. 2002). A systematic review of the literature found an association between high hospital volumes and low numbers of hip dislocations (Shervin et al. 2007). Lower provider volumes have been related to longer hospital stay after THR surgery (Lavernia and Guzman 1995, Doro et al. 2006, Judge et al. 2006).

We evaluated the effects of hospital volume on the length of stay, the number of re-admissions, and the number of complications of THRs in the whole population of Finland.

Patients and methods All public and private hospitals in Finland are obliged to report any surgical procedures requiring an overnight stay to the National Institute for Health and Welfare. The Hospital Discharge Register, which is maintained by the National Institute for Health and Welfare, was the main database we used. All other data were obtained from the following sources: the Benchmarking database compiled for the Hospital Benchmarking project for productivity in specialized care (Linna and Häkkinen 2008), the Social Insurance Institution database, and the Finnish Arthroplasty Register. The characteristics of the data derived from these registries are dealt with later. Every Finnish resident has a unique personal identification number, which can be used to combine data from different registers. The effect of hospital volume on the length of the surgical treatment period (referred to later as length of stay (LOS)), the length of uninterrupted institutional care (LUIC), and the rates of unscheduled re-admissions, reoperations, dislocations, and infections were evaluated. The surgical treatment period was defined as the period during which the THR was performed in the hospital, as shown by the Hospital Discharge Register. A surgical treatment period ended in discharge, in transfer to another facility, or in death of the patient. Uninterrupted institutional care was defined as the combination of the surgical treatment period and the immediate period of rehabilitation. Uninterrupted institutional care ended in either death or discharge of the patient, which also included the patient being transferred to another facility such as an old people’s home or

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2010.533930


Acta Orthopaedica 2011; 82 (1): 20–26

21

Table 1. Number of patients and annual number of hospitals in different hospital volume groups, with annual number of private hospitals in parentheses Group No. of patients 1998–2005

Annual no. of hospitals 1998

1999

2000

2001

2002

2003

2004

2005

1 2 3 4

2,759 11,591 8,134 7,782

15 (7) 29 (2) 2 (0) 1 (1)

18 (8) 32 (2) 9 (0) 2 (1)

23 (8) 26 (2) 9 (0) 4 (1)

25 (9) 26 (3) 8 (0) 5 (1)

23 (10) 28 (2) 10 (1) 3 (1)

19 (11) 22 (2) 9 (0) 7 (2)

19 (10) 21 (2) 10 (0) 6 (2)

18 (9) 23 (4) 8 (0) 9 (2)

Total

30,266

57 (10)

61 (11)

62 (11)

64 (13)

64 (14)

57 (15)

56 (14)

58 (15)

See Material and Methods for explanation of groups

another long-term care or social welfare institution. The maximum length of institutional care was limited to 60 days as a cutoff point in the calculations. The study period was from 1998 through 2005. Reoperations, closed and open reductions of dislocated hip prostheses, and infections of the THR were followed to the end of 2008. Hospital grouping Hospitals were classified into 4 groups according to the number of THRs (NOMESCO codes NFB30–NFB99) performed annually during the whole study period: 1–50 (lowvolume hospitals, group 1), 51–150 (medium-volume hospitals, group 2), 151–300 (high-volume hospitals, group 3), and > 300 (very-high-volume hospitals, group 4) (Table 1). The low-volume hospitals were used as a reference group. There were 72 hospitals involved. Hospitals that changed hospital volume group during the study period were also included. 39 hospitals did not change volume group (9,926 hips) and 33 hospitals changed volume group once (20,340 hips). None of the hospitals changed volume group more than once. Inclusion criteria The study population was formed by selecting patients from the Hospital Discharge Register according to the International Classification of Diseases (ICD-10), and the following were used as selection criteria: M16.0/M16.1 for primary osteoarthritis (OA) of the hip or M16.2/M16.3 for developmental dysplasia of the hip (DDH), associated with a code for primary THA performed over the 1998–2005 period. Patients with a diagnosis of developmental dysplasia of the hip (DDH) (ICD-10: M16.2/M16.3) were included in the study because there is variation in coding of mild DDH and primary OA. Codes for THR included NFB30 for cementless THR, NFB40 for hybrid THR, NFB50 for cemented THR, NFB60 for demanding THR, and NFB99 for other THR procedures such as hip resurfacing. Patients with congenital hip dislocation are classified under a different diagnosis code (Q65.0-Q65.9) and were not included in the study. The accuracy of diagnosis for primary OA was double-checked against the relevant data in

Table 2. Number (Na) and percentage of replacements in cohorts when analyzing length of stay and unscheduled re-admissions, and number (Nb) and percentage of replacements in cohorts when analyzing dislocations, re-operations, and infections Cohort

No. (Na) of hips (%)

No. (Nb) of hips (%)

1998 1999 2000 2001 2002 2003 2004 2005

2,863 (10) 3,083 (10) 3,360 (11) 3,554 (12) 3,877 (13) 4,336 (14) 4,113 (14) 5,080 (17)

1,906 (9) 2,025 (10) 2,178 (11) 2,389 (12) 2,639 (13) 3,056 (15) 2,963 (14) 3,658 (18)

Total

30,266 (100)

20,814 (100)

the Finnish Arthroplasty Register. Total hip replacements— not patients—were evaluated when considering the length of surgical treatment period, the length of institutional care, and unscheduled re-admissions. Thus, it was possible that 2 THRs were evaluated from the same patient. The total number of all the THRs included was 30,266. However, when considering reoperations, dislocations, and infections, only patients with unilateral THR implants were evaluated over the period 1988–2008 (n = 20,814) (Table 2). The reason for this is that the side of the operation (left/right) is not reliably coded in the Hospital Discharge Register. If the data for the side of the operation are unavailable, then it is not possible to compare left-side with right-side THR from the registry data with any certainty. Exclusion criteria THRs for secondary OA, reoperations, and revisions of an implant were excluded (Table 3, see supplementary data). The manifestation of the diagnosis of secondary hip OA was noted retrospectively from the beginning of 1988. A patient was excluded from the study if there was a diagnosis of secondary hip OA in the Hospital Discharge Register between the beginning of 1988 and the day of the operation. Patients


22

Acta Orthopaedica 2011; 82 (1): 20–26

who were eligible for reimbursement in the Social Insurance Institution database for the sequelae of transplantation, uremia requiring dialysis, rheumatoid arthritis, or connective tissue disease were also excluded from the study. Patients with a simultaneously performed hip and knee arthroplasty were also excluded.

price of 1 re-admission in a specialized care facility was used to calculate increase in the costs if all THRs had been performed in the group of hospitals with the highest amount of re-admissions at 2 weeks. In 2006, the price of 1 re-admission in a specialized care facility was 278 euros on average in the whole country (Hujanen et al. 2008).

Dislocations We defined a reduction of a dislocated THR in 2 ways. Firstly, a reduction of a dislocated THR was considered to have been performed if there was a notification in the Hospital Discharge Register that either an open or closed reduction of a dislocated total hip prosthesis had been performed (NFH30 or NFH32), associated with a diagnosis of an internal mechanical complication of endoprosthesis (ICD10: T84.0). Secondly, closed reduction may also have been performed in the emergency room under light sedation anesthesia. These patients are often discharged from the accident and emergency units after closed reduction without an overnight stay in the hospital. According to Finnish regulations, it is mandatory to compile statistics on diagnosis codes in the accident and emergency units but relevant operational codes are not recorded routinely. Thus, a closed reduction of a dislocated THR was considered to have been performed if the patient had an unscheduled re-admission with a diagnosis code of mechanical complication of endoprosthesis (ICD: T84.0) but without any actual hospital admission. We believe that most of these non-admission cases are true dislocations, not periprosthetic fractures or aseptic loosening of the implant.

Statistics A logistic regression model for data at the individual level was used for dichotomous dependent variables and a generalized linear model (gamma-distribution, log-link) was used for continuous dependent variables, with hospital volume (classified) as explanatory variable. To improve the comparability of the study data, the analyses were adjusted for confounding factors (Iezzoni 2003). In addition, 95% confidence intervals (CIs) were determined. The age of the patient (under 40 years, over 40 years (divided into 9 incremental groups each of 5-year intervals up to 85 years), and over 85 years), sex, any previous THR, and co-morbidities were also adjusted for. In the models of LOS and LUIC, operation year dummies were included. We also performed calculations using head size of the prostheses as an adjustment factor to eliminate the effect of head size on dislocation rate. Co-morbidities were determined using the diagnoses obtained from the Hospital Discharge Register from the beginning of 1987 to the date of operation. In addition, the Social Insurance Institution database for eligibility for reimbursement including the use and cost of drugs was used to adjust for co-morbidity (Table 4). The illnesses chosen for adjustment were such that they might have had an effect on the performance of THR, on length of stay in the hospital, or on the rate of complications. The length of follow-up was also identified as a confounding factor for adjustment of the rates of complications.

Infections A diagnosis code for deep prosthetic infection (ICD-10: T84.5) was used in the data search from the beginning of the surgical treatment period to the end of the follow-up. Data from the Finnish Arthroplasty Register were used to determine reoperations performed due to deep prosthetic infections. Unscheduled re-admission An unscheduled re-admission was recorded if a patient was re-admitted to hospital or had required medical attention in an outpatient department or an accident and emergency unit of any hospital in Finland. Such an unscheduled re-admission also had to occur within the first 14 and the first 42 days from the end of the surgical treatment. Costs The total cost of 1 day of treatment in a specialized care facility was used to calculate the potential savings if all THRs had been performed in the group of hospitals with the shortest LOS. The price of 1 day of treatment in hospitals performing THRs was on average 527 euros in Helsinki and Uusimaa district (HUS) during the period 2003–2005 (Peltola 2008). However, the mean cost of 1 day of treatment in hospitals performing THRs was not available for the whole country. The

Results Length of stay (LOS) and length of uninterrupted institutional care (LUIC) (Table 5) In 2005, LOS was 5.49 days in group 4, 6.65 days in group 3, 7.63 days in group 2, and 6.84 days in group 1 (the reference group). During the whole study period, LOS was longer in group 1 than in group 4 (p < 0.001). However, LOS was shorter in group 1 than in group 2 (p < 0.001). In 2005, LUIC was 9.91 days in group 4, 10.47 days in group 3, 10.59 days in group 2, and 10.27 days in group 1. During the whole study period, LUIC was shorter in group 4 than in group 1 (p < 0.001). Nonetheless, LUIC was longer in group 3 and in group 2 than in group 1 (p < 0.001 for both comparisons). Theoretically, if all THRs during the study period in Finland had been performed in the very–high-volume hospitals with the shortest length of stay, total LOS in the hospital would have decreased by 29,761 days (1.0 day per patient). Thus, the costs would have been reduced by 15,684,047 euros.


Acta Orthopaedica 2011; 82 (1): 20–26

Table 4. Related diseases used in the adjustment of the study population Hypertension (ICD-10: I10*-I15*, ICD-9: 40*, Social Insurance Institutions entitlement to reimbursement: 205, use and cost of drugs ATC: C03*, C07* (if there is no coronary disease or atrial fibrillation), C09A*, C09B*, C09C*, C09D*, C08*) Coronary disease (ICD-10: I20*-I25*, ICD-9: 410*-414*, Social Insurance Institutions entitlement to reimbursement: 206, 213, 280) Atrial fibrillation (ICD-10: I48*, ICD-9: 4273*, Social Insurance Institutions entitlement to reimbursement: 207, use and cost of drugs ATC: B01AA03) Heart insufficiency (ICD-10: I50*, ICD-9: 428*, Social Insurance Institutions entitlement to reimbursement: 201) Diabetes (ICD-10: E10*-E14*, ICD-9: 250*, Social Insurance Institutions entitlement to reimbursement: 103 use and cost of drugs ATC-DDD: A10A*, A10B*) Cancer (ICD-10: C00*-C99*, D00*-D09*, ICD-9: 140*-208*, Social Insurance Institutions entitlement to reimbursement: 115, 116, 117, 128, 130, 180, 184, 185, 189, 311, 312, 316, use and cost of drugs ATC L01* except L01BA01) COPD and asthma (ICD-10: J44*-J46*, ICD-9: 4912*, 496*, 493*, Social Insurance Institutions entitlement to reimbursement: 203, use and cost of drugs ATC: R03*) Depression (ICD-10: F32*-F34*, ICD-9: 2960*, 2961*, 2069*, use and cost of drugs ATC: N06A*) Parkinson’s disease (ICD-10: G20*, ICD-9: 332*, Social Insurance Institutions entitlement to reimbursement: 110, use and cost of drugs ATC: N04B*) Dementia (ICD-10: F00*-F03*, G30*, ICD-9: 290*, 3310*, Social Insurance Institutions entitlement to reimbursement: 307, use and cost of drugs ATC: N06D*) Kidney insufficiency (ICD-10: N18*, ICD-9: 585*, Social Insurance Institutions entitlement to reimbursement: 137) Mental disorders (ICD-10: F20*-F31*, ICD-9: 295*-298*, except 2960*, 2961*, Social Insurance Institutions entitlement to reimbursement: N05A* except N05AB04 and N05AB01 and there is no dementia)

If all THRs during the study period in Finland had been performed in the very–high-volume hospitals, this would have led to a total of 355 new re-admissions at 2 weeks. The increase in cost for these re-admissions would have been 98,690 euros. If we then subtract the increased costs due to a higher rate of re-admissions (98,690 euros) from the total savings in LOS (15,684,047 euros), the final savings during the follow-up time would have been 15,585,357 euros. Complications (Table 6) In unadjusted data, statistically there were significantly less dislocations in groups 3 and 4 than in group 1. There were also less reoperations in group 4 than in group 1. In group 1, there were more re-admissions both at 2 and 6 weeks than in group 1. However, in adjusted data there was only a trend of fewer re-admissions within 14 days in group 1 than in group 4. There were significantly more dislocations in group 1 than in group 3. There was no association between hospital volume and reoperation rates or infection rates when using adjusted data. We also performed the analyses using the ownership of the hospitals (private or public) as a confounding factor. Of all

23

Table 5. The average length of stay and the number of days saved if patients were treated in hospitals with the shortest length of stay Year Group LOS N 1998

Difference in LOS

1 10.06 233 0.35 2 10.52 1,395 0.81 3 9.05 1,145 -0.66 4 9.71 90 – In total 2,863 1999 1 9.80 227 0.59 2 10.01 1,664 0.8 3 8.85 953 -0.36 4 9.21 239 – In total 3,083 2000 1 9.37 388 1.50 2 9.81 1,452 1.94 3 8.82 932 0.95 4 7.87 588 – In total 3,360 2001 1 8.65 493 0.92 2 9.66 1,386 1.93 3 8.57 899 0.84 7.73 776 – 4 3,554 In total 2002 8.34 449 1.21 1 9.15 1,647 2.02 2 7.76 1,207 0.63 3 7.13 574 – 4 3,877 In total 2003 7.17 315 1.08 1 8.58 1,358 2.49 2 7.93 988 1.84 3 6.09 1,675 – 4 4,336 In total 2004 7.22 346 1.34 1 8.46 1,181 2.58 2 7.38 1,059 1.50 3 5.88 1,527 – 4 4,113 In total 2005 6.84 308 1.35 1 7.63 1,508 2.14 2 6.65 951 1.16 3 5.49 2,313 – 4 5,080 In total 1998–2005

Days saved 82 1,130 -756 – 456 134 1,331 -343 – 1,122 582 2,817 885 – 4,284 454 2,675 755 – 3,884 543 3,327 760 – 4,630 340 3,381 1,818 – 5,539 464 3047 1,589 – 5,100 416 3,227 1,103 – 4,746 29,761

LOS: annual mean length of stay (surgical treatment period) in days; N: number of patients; Difference in LOS: difference between the shortest length of stay and that of other hospital groups in days; Days saved: number of days saved if patients were operated in hospitals with the shortest length of stay (days saved = difference in LOS multiplied by N). In 1998 and 1999, the average LOS was shorter in group 3 than in group 4. This is why “Difference in LOS” and “Days saved” have negative values in the Table.

30, 266 hips studied, 4,150 hips were operated on in private hospitals. There was no significant difference between hospital volume results when the private/public dichotomy was included in the model. LOS was statistically significantly shorter in private hospitals than in public hospitals.


24

Acta Orthopaedica 2011; 82 (1): 20–26

Table 6a. Unadjusted odds ratios for unscheduled re-admissions within 14 and 42 days, and for dislocations, reoperations, and infections

Groups

Re-admsissions 14 days OR 95% WCI

Re-admsissions Dislocations Reoperations 42 days OR 95% WCI OR 95% WCI OR 95% WCI

OR

95% WCI

2 vs. 1 3 vs. 1 4 vs. 1

0.9 1.0 1.2

0.9 1.0 1.1

1.0 0.9 0.8

0.7–1.6 0.6–1.5 0.5–1.3

0.8–1.1 0.9–1.2 1.0–1.4

0.8–1.0 0.8–1.1 1.0–1.3

0.9 0.7 0.8

0.7–1.2 0.6–0.9 0.6–1.0

1.1 1.0 0.7

0.9–1.3 0.8–1.2 0.5–0.9

Infections

OR: odds ratio; 95% WCI: 95% Wald confidence interval. See Methods for explanation of groups.

Table 6b. Adjusted odds ratios for unscheduled re-admissions within 14 and 42 days, and for dislocations, reoperations, and infections

Groups

Re-admsissions 14 days OR 95% WCI

Re-admsissions Dislocations Reoperations 42 days OR 95% WCI OR 95% WCI OR 95% WCI

OR

95% WCI

2 vs. 1 3 vs. 1 4 vs. 1

0.9 1.0 1.2

0.9 0.9 1.1

1.0 0.9 0.8

0.6–1.6 0.6–1.4 0.5–1.4

0.8–1.1 0.9–1.2 1.0–1.4

0.8–1.0 0.8–1.1 1.0–1.2

0.9 0.7 1.1

0.7–1.1 0.6–0.9 0.9–1.4

1.0 0.9 0.9

0.8–1.3 0.8–1.2 0.7–1.1

Infections

OR: odds ratio; 95% WCI: 95% Wald confidence interval. See Methods for explanation of groups.

Discussion Our findings indicate that specialization of hip replacements in high-volume hospitals should reduce costs by shortening LOS, and may reduce the dislocation rate. Validity of the data The reporting accuracy of the Hospital Discharge Register in Finland is high when considering surgical operations. In the late 1980s, at least 95% of operations were already being recorded in this register (Keskimäki and Aro 1991). The correlation between the Finnish Discharge Register and the Finnish Arthroplasty Register is high (Peltola 2008). The strength of our study is that it included operative data from both private and public hospitals. We adjusted for patient age, sex, surgery, and medical diagnosis. Adjustment calculations were also done for the head size of the prosthesis. However, not all factors associated with dislocation rate, such as surgical approach, could be adjusted for in the data from the current study. Our basic assumption was that performing more hip surgery correlates with skills. Change of hospital volume group would not interfere with our analyses if hospital volume is an important factor compared to other variables affecting the quality of hip surgery. Length of stay The LOS after THR varies considerably. There are reports

in the literature that THRs have actually been performed as day-case surgery (Berger et al. 2004), even though very early discharge has been cautioned against (Parvizi et al. 2007). Nevertheless, reducing LOS reduces the cost of care and permits increase in bed occupancy rates (Williams et al. 2005). In previous studies, longer LOSs after THR have been associated with lower provider volumes (Lavernia and Guzman 1995, Doro et al. 2006, Judge et. al. 2006). In our study, not only was a very high hospital volume associated with shorter surgical treatment periods but also with shorter LUIC. Large amounts of money can be saved if THRs are performed in very–high-volume hospitals. However, when the postoperative care is made more efficient, the easiest and cheapest days are removed, not the heaviest days including the first postoperative day and the discharge day. In the near future in Finland, a large number of people born in the late 1940s and early 1950s will retire. Because of the aging population and the decreasing number of nurses, more patients will have to be treated using the same resources as those used today—by optimizing efficiency. Unscheduled re-admissions Unscheduled re-admission rate is a national key performance indicator used by the UK Department of Health (Adeyemo and Radley 2007). A 28-day emergency re-admission rate has been used as a clinical indicator to compare surgical and orthopedic performance between trusts in England and in Scotland (Courtney et al. 2003). In a study by Cullen et al. (2006), 9%


Acta Orthopaedica 2011; 82 (1): 20–26

of patients were re-admitted within 28 days of discharge after THR. The main reasons for re-admission were thromboembolism, dislocations, and wound complications. Reducing the length of stay lowered the cost of care per patient and permitted increased bed occupancy, but the effect on emergency re-admission rates was equivocal (Williams et al. 2005). The odds ratio of emergency re-admission for primary hip replacement was 0.54 when the length of stay was 4–7 days and 0.55 when it was 8–14 days, considering the odds ratio 1.0, when the length of stay was 4 days or under (Williams et al. 2005). In a study by Judge et al. (2006), there was evidence of a higher re-admission rate for high-volume trusts. In our study, the rate of re-admissions in the low-volume hospitals was also lower than in all other hospital groups. When the length of stay is longer, early problems manifest in the hospital and are treated immediately in situ; thus, re-admissions are less likely to occur than in shorter-stay facilities. However, the reasons for re-admissions are most often not major surgical complications but minor wound problems, suspicion of venous thromboembolism, and medication problems. The costs of re-admissions were low compared to the costs of longer length of stay. Dislocations The dislocation rate during the first year after THR has been reported to range from less than 1% to almost 4% (Phillips et al. 2003, Khatod et al. 2006, Meek et al. 2006). Several factors are constantly reported to be statistically associated with THR dislocation rate. These include surgical diagnosis, femoral head size, patient age and sex, American Association of Anesthesiologists (ASA) score, cognitive dysfunction, surgical approach, surgeon volume, and hospital volume (Meek et al. 2006). It is not possible to influence all of these factors to reduce the number of dislocations. However, healthcare workers should try to manipulate those factors that can be influenced, such as surgeon volume, hospital volume, head size, and repair of soft tissues. In our study, the low-volume hospitals had higher dislocation rates than the high-volume hospitals and the very–highvolume hospitals. However, the dislocation rate for very-highvolume hospitals was not lower than that of low-volume hospitals when we analyzed adjusted data. The patients in large-volume units are probably somewhat younger than in low-volume hospitals. In small units, there are more typical osteoarthritic patients of advanced age who are at risk of having short-term complications. On the other hand, very–high-volume hospitals are mainly university hospitals with junior surgeons who perform replacements as part of their education. Furthermore, not all diseases— for example, obesity and alcoholism—nor the condition of bone and soft tissues can be adjusted, and it is likely that even after adjustment, there will be more difficult patients in university hospitals that have adverse influential effects on the dislocation rates observed. High dislocation rates in low-volume hospitals are an alarming finding that has also been reported previously (Shervin et al. 2007).

25

Reoperations Instability has been the most common reason for early reoperations, whereas osteolysis and loosening are the most common reasons for late revisions (Clohisy et al. 2004). Because the follow-up time in our study was short, we assumed that many reoperations were performed either because of infections, instability, or periprosthetic fractures. Early dislocations after THR are most often treated by closed reduction; these were analyzed separately. Open reduction and revision operations as a treatment for dislocated hip prostheses are uncommon. Unfortunately, the reason for revision is not directly available from the Finnish Arthroplasty Register, only the type of revision performed (for example exchange of the stem with the diagnosis code NFC20, but not if the revision was performed for periprosthetic fracture or for dislocation, both of which have the diagnosis code T84.0). We found less reoperations in the very-high-volume group than in the low-volume group when using unadjusted raw data. However, this difference did not reach statistical significance when using the adjusted data. In a study by Manley et al. (2008), the patients of lowvolume surgeons had a greater risk of arthroplasty revision at 6 months but no greater risk of revision at the time of longerterm follow-up. No associations between hospital volumes and the rates of revision of THA were found by Judge et al. (2006) and Manley et al. (2008). Infections In recent studies, the rate of deep infection after THA has varied between 0.6% and 0.9% (Gastmeier et al. 2005, Muilwijk et al. 2006, Phillips et al. 2006). Independent risk factors for surgical site infections after THR are patient age, surgical diagnosis, ASA score, and duration of operation (Ridgeway et al. 2005). In a systematic literature review, no association between hospital volumes and infection rates was found (Shervin et al. 2007). Nowadays, deep infections after THR are rare that the capability of population-based studies to determine significant differences between hospitals is limited. In our study, patient age, sex, surgical diagnosis, and medical diagnoses were adjusted for. Thus, it was not surprising that no association between infection rate and hospital volume was detected. Summary We found that LOS was shortest in the very-high-volume hospitals. The costs of more frequent re-admissions in the very– high-volume hospitals were low compared to the savings from short LOS. Specialization of high-volume hospitals regarding hip replacements should reduce costs by significantly shortening length of stay, and it may reduce the dislocation rate.

MP performed data analysis. KTM wrote the manuscript. All the authors contributed to the conception and design of the study, to critical analysis of the data, to interpretation of the findings, and to critical revision of the manuscript.


26

This study was supported by the Sigrid Juselius Foundation and by a research grant from Helsinki University Central Hospital (TYH7306).

No competing interests declared.

Supplementary data Table 3 is available at our website (www.actaorthop.org), identification number 3967. Adeyemo D, Radley S. Unplanned general surgical re-admissions - how many, which patients and why? Ann R Coll Surg Engl 2007; 89: 363-7. Battaglia T C, Mulhall K J, Brown T E, Saleh K J. Increased surgical volume is associated with lower THA dislocation rates. Clin Orthop 2006; (447): 28-33. Berger R A, Jacobs J J, Meneghini R M, Della Valle C, Paprosky W, Rosenberg A G. Rapid rehabilitation and recovery with minimally invasive total hip arthroplasty. Clin Orthop 2004; (429): 239-47. Clohisy J C, Calvert G, Tull F, McDonald D, Maloney W J. Reasons for revision hip surgery: a retrospective review. Clin Orthop 2004; (429): 188-92. Courtney E D, Ankrett S, McCollum P T. 28-Day emergency surgical readmission rates as a clinical indicator of performance. Ann R Coll Surg Engl 2003; 85: 75-8. Cullen C, Johnson D S, Cook G. Re-admission rates within 28 days of total hip replacement. Ann R Coll Surg Engl 2006; 88: 475-8. Doro C, Dimick J, Wainess R, Upchurch G, Urquhart A. Hospital volume and inpatient mortality outcomes of total hip arthroplasty in the United States. J Arthroplasty (Suppl) 2006; 21: 10-6. Gastmeier P, Sohr D, Brandt C, Eckmanns T, Behnke M, Ruden H. Reduction of orthopaedic wound infections in 21 hospitals. Arch Orthop Trauma Surg 2005; 125: 526-30. Hujanen T, Kapiainen S, Tuominen U, Pekurinen M. Terveydenhuollon yksikkökustannukset Suomessa vuonna 2006. Stakes, Työpapereita 3/2008. ICD-10, International Statistical Classification of Diseases and Related Health Problems 10th Revision Version for 2007 WHO. www.who.int/classifications/icd/icd10online/ (Date last accessed 25.10.2008) Iezzoni L I. Risk adjustment for measuring health care outcomes. Health Administration Press, Chicago. 2003. Judge A, Chard J, Learmonth I, Dieppe P. The effects of surgical volumes and training centre status on outcomes following total hip replacement: analysis of the Hospital Episode Statistics for England. J Public Health 2006; 28: 116-24.

Acta Orthopaedica 2011; 82 (1): 20–26

Keskimäki I, Aro S. Accuracy of data on diagnosis, procedures and accidents in the Finnish Hospital Discharge Register. Int J Health Sci 1991; 2: 15-21. Khatod M, Barber T, Paxton E, Namba R, Fithian D. An analysis of the risk of hip dislocation with a contemporary total joint registry. Clin Orthop 2006; (447): 19-23. Lavernia C J, Guzman J F. Relationship of surgical volume to short-term mortality, morbidity, and hospital charges in arthroplasty. J Arthroplasty 1995; 10: 133-40. Linna M, Häkkinen U. Benchmarking Finnish Hospitals. In: Evaluating hospital policy and performance: contributions from hospital policy and productivity research. Advances in health economics and health services research (eds Blank J, Valdmanis V), Elsevier, Oxford 2008; 18: 179-90. Manley M, Ong K, Lau E, Kurtz S M. Effect of volume on total hip arthroplasty revision rates in the United States Medicare population. J Bone Joint Surg (Am) 2008; 90: 2446-51. Meek R M, Allan D B, McPhillips G, Kerr L, Howie C R. Epidemiology of dislocation after total hip arthroplasty. Clin Orthop 2006; (447): 9-18. Muilwijk J, Walenkamp G H, Voss A, Wille J C, van den Hof S. Random effect modelling of patient-related risk factors in orthopaedic procedures: results from the Dutch nosocomial infection surveillance network ‘PREZIES’. J Hosp Infect 2006; 62: 319-26. NOMESCO Classification of Surgical Procedures http://www.nordclass. uu.se/verksam/Ncsp1_5. (Date last accessed 7.11.2008). Parvizi J, Mui A, Purtill J J, Sharkey P F, Hozack W J, Rothman RH. Total joint arthroplasty: When do fatal or near-fatal complications occur? J Bone Joint Surg (Am) 2007; 89: 27-32. Peltola M, personal communication 8.9.2008, The National Research and Development Centre for Welfare and Health in Finland. Phillips C B, Barrett J A, Losina E, Mahomed N N, Lingard E A, Guadagnoli E, et al. Incidence rates of dislocation, pulmonary embolism, and deep infection during the first six months after elective total hip replacement. J Bone Joint Surg (Am) 2003; 85: 20-6. 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: 943-8. Ridgeway S, Wilson J, Charlet A, Kafatos G, Pearson A, Coello R. Infection of the surgical site after arthroplasty of the hip. J Bone Joint Surg (Br) 2005; 87: 844-50. Shervin N, Rubash H E, Katz J N. Orthopaedic procedure volume and patient outcomes: a systematic literature review. Clin Orthop 2007; (457): 35-41. Solomon D H, Losina E, Baron J A, Fossel A H, Guadagnoli E, Lingard E A, et al. Contribution of hospital characteristics to the volume-outcome relationship: dislocation and infection following total hip replacement surgery. Arthritis Rheum 2002; 46: 2436-44. Williams S, Bottle A, Aylin P. Length of hospital stay and subsequent emergency readmission. BMJ 2005; 331: 371.


Acta Orthopaedica 2011; 82 (1): 27–34

27

Infectiological, functional, and radiographic outcome after revision for prosthetic hip infection according to a strict algorithm 22 one-stage and 50 two-stage revisions with a mean follow-up time of 5 (2–17) years F Harald R De Man1*, Parham Sendi2*, Werner Zimmerli2, Thomas B Maurer2, Peter E Ochsner2, and Thomas Ilchmann2 1Department of Orthopaedics, Sint Maartenskliniek, Nijmegen, the Netherlands; 2Clinic of Orthopaedic Surgery, Kantonsspital, and the Infectious Diseases Unit, Basel University Medical Clinic, Liestal, Switzerland * F Harald R De Man and Parham Sendi contributed equally to the study. Correspondence: thomas.ilchmann@ksli.ch Submitted 10-03-01. Accepted 10-09-17

Background and purpose Successful treatment of prosthetic hip joint infection (PI) means elimination of infection and restored hip function. However, functional outcome is rarely studied. We analyzed the outcome of the strict use of a treatment algorithm for PI. Patients and methods The study groups included 22 hips with 1-stage exchange for PI (group 1), 22 matched hips revised for aseptic loosening (controls), and 50 hips with 2-stage exchange (group 2). Relapse of infection, Harris hip score (HHS), limping, use of crutches, reoperations, complications, and radiographic changes were compared between the groups. Results There was 1 relapse of infection, which occurred in group 2. In group 1, the mean HHS was 84; 4 of 19 patients were limping and 2 required 2 crutches, which was similar to the control results. In group 2, scores were lower and complication rates higher. The use of a Burch-Schneider ring and the presence of a deficient trochanter impaired function. There were no differences in radiographic outcome between the groups. Interpretation With the algorithm used, infection can be cured with high reliability. With a 1-stage procedure, mobility is maintained. After 2-stage procedures, function was impaired due to there being more previous surgery and more serious infection. 

Infections associated with prosthetic joints cause significant morbidity and account for a substantial proportion of healthcare expenditure (Bozic and Ries 2005). The management of infection associated with prosthetic joints is poorly standardized because of the varied clinical presentations and the lack

of data from randomized, controlled trials. We have recently published a treatment algorithm that was developed at the Kantonsspital Liestal over the past 25 years (Zimmerli et al. 2004). Adherence to this treatment concept has shown a success rate of 85–100% in curing infection (Giulieri et al. 2004, Sendi et al. 2006). Successful treatment of prosthetic hip joint infection (PI) consists not only of eliminating the infection but also of restoring patient mobility, which is important for patient satisfaction (Britton et al. 1997). Even so, functional and radiographic outcome has been mainly investigated in aseptic revisions (Saleh et al. 2003). Ideally, the evaluation of any treatment algorithm for PI should be from a multidisciplinary perspective, including infectiological, radiographic, and functional outcome. Our surgical technique and choice of components are entirely dictated by the quality of soft tissue and bone. In our view, the functional outcome after 1-stage exchange for PI should therefore be similar to that for aseptic reasons. After 2-stage exchange, this outcome is expected to be worse because this intervention is commonly performed in patients with more severe infections, requiring a more complex surgical procedure. Moreover, in revision hip surgery and especially in cases with PI, patient and surgical factors (e.g. femoral osteotomy and components) are highly variable because of the different quality of soft tissue and bone in individual cases. Thus, when reporting on functional outcome in cases of PI, the presence of such factors should be acknowledged and their possible influence on outcome analyzed. Because elimination of the infection is a prerequisite for a good functional outcome, we first wanted to confirm that our

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2010.548025


28

treatment algorithm is associated with successful infectiological outcome as reported in previous studies (Giulieri et al. 2004, Sendi et al. 2006). Since these reports were published, 37 additional hips have been treated accordingly and could be included for analysis of infectiological outcome. Secondly, we wanted to determine whether functional and radiographic outcome, aseptic revision, and complication rate would be different (1) after 1-stage exchange for PI as compared to 1-stage aseptic revisions, and (2) after 1-stage exchange for PI as compared to 2-stage exchange. We hypothesized that all outcomes would be similar in 1-stage groups and lower after 2-stage exchange. Thirdly, we wanted to determine whether functional outcome was affected by various pre-defined surgeryrelated parameters.

Acta Orthopaedica 2011; 82 (1): 27–34

Clinical symptoms ≤3 weeks AND stable prosthesis AND no moderately or severely compromised soft-tissue AND no presence of difficult-to-treat microorganism

YES

Debridement and retention of the implant

NO Exchange of the implant

Moderately or severely compromised soft-tissue AND/OR presence of difficult-to-treat microorganism

NO

1-stage exchange

YES 2-stage exchange

Figure 1. The algorithm showing decision making for a 1-stage or 2-stage revision.

Patients and methods Population We retrospectively analyzed all patients who had been treated with exchange of a THA, because of PI, at our center between 1985 and 2004. We formed 3 groups. The “1-stage exchange due to PI” group (group 1), a matched aseptic control group (Control), and the “2-stage exchange” group (group 2) who were analyzed separately. Functional and radiographic parameters, rates of revision, and complications were compared between controls and group 1, and between group 1 and group 2. A transfemoral osteotomy, a deficient greater trochanter, a 2-stage revision, a history of multiple cup or stem revision, a BS-ring, or a Wagner stem were identified to possibly lead to a lower HHS, and/or more limping or use of support. In order to analyze whether or not these variables independently influenced functional outcome all hips with an event-free followup were pooled in one single study group and variables were analyzed separately. To establish the diagnosis of PI, the presence of a sinus tract or the growth of the same microorganism in at least 2 cultures, or inflammation consistent with infection on histopathological examination was required (Giulieri et al. 2004, Zappe et al. 2008). Hips with a PI were only included if treatment was strictly according to the algorithm and an exchange of the implant had been done (Zimmerli et al. 2004) (Figure 1). Antimicrobial compounds were selected as described previously (Zimmerli et al. 2004) and typically administered for 8–12 weeks. Hips were excluded when results were not available 2 years after the index operation, or if they were lost to follow-up or when documentation was poor. 91 hips were treated with a 1-stage exchange or a 2-stage exchange because of PI. 12 hips were considered not to have

been treated according to the algorithm and were excluded: in 2 hips the infection was detected after death, in 2 hips treatment of infection was declined, in 4 hips components remained in site despite the presence of severe soft tissue damage because the operation risk due to co-morbidities was inordinately high and in 4 hips the infection was unknown during revision and therefore treatment started too late. In group 1, 1 hip was excluded due to loss of 2 year follow-up and another hip because follow-up was not well documented, resulting in 22 selected hips (21 patients) for final inclusion. These patients were matched with 22 patients (controls) who were selected from 474 consecutive aseptic revision procedures. Matching was performed—in decreasing order of importance—for previous surgery on trochanter, number of revisions of the cup, number of revisions of the stem, type of implant, use of transfemoral osteotomy, Charnley score, duration of follow-up, age, and sex. In group 2, 5 hips were excluded all because of loss of 2 year follow-up resulting in 50 selected hips (48 patients) for final inclusion (Figure 2), of which 34 hips had severely damaged soft tissues with a sinus and/or abcess formation. 12 of the 22 hips in group 1 (11 patients) and 39 of the 50 hips in group 2 (37 patients) had been referred. The referred cases had had more surgery before referral and qualified less often for a 1-stage exchange compared to the non-referred patients. Staphylococci were the most commonly involved pathogens (40 hips) followed by Streptococcus spp. (11 hips) and gramnegative rods (8 hips). In 17 hips, the isolates were classified as difficult to treat, and included Pseudomonas aeruginosa, rifampin-resistant staphylococci, Enterococcus spp, MRSA, small-colony variants of S. aureus, and Abiotrophia adiacens, requiring a 2-stage exchange in these cases. In 10 hips the infection was polymicrobial, and in 3 hips no pathogen could be cultured.


Acta Orthopaedica 2011; 82 (1): 27–34

29

Patients were followed up after 6 weeks, 3 months, 6 months, 1 year, 2 years, and every 5 years thereafter. The mean clinical follow-up was 3.8 years (SD 2.2) for both group 1 and the control group, and 4.9 years (SD 3.6) for group 1-stage Not treated according 2-stage 2. 19 of the 22 patients in group 1, all 22 in the control group, 24 hips to algorithm 55 hips and 43 of the 50 hips in group 2 had an event-free survival for 12 hips ≥ 2 years and qualified for functional outcome analysis. Lost to 2-year follow-up Operative technique 1 hip 5 hips The index operation was defined as the reimplantation procedure done at our center. Bone defects on the latest preoperative Follow-up not well-documented radiographs were classified according to Paprosky (Paprosky et al. 1994, Valle and Paprosky 2003) (Table 1). Patients were 1 hip – operated in the supine position with a straight lateral approach. In 34 hips, this was combined with a transfemoral osteotomy 1-stage 2-stage (Wagner and Wagner 1999) (Figure 3). In the infected cases, (group 1) (group 2) after removal of the components a thorough debridement was 22 hips 50 hips performed—taking care to preserve bone and soft tissue. In all groups, the choice of a particular implant was dictated by Figure 2. Flow of hips included in the study. Exclusion of 19 of 91 hips the classification of bone defects irrespective of the presence that did not fulfill the inclusion criteria. of infection. For reconstruction of the acetabulum, an uncemented cup (SL-Müller), a reinTable 1. Patient characteristics prior to the index operation forcement ring (Müller), or a Burch-Schneider (BS) reinforcement ring (all Zimmer, Warsaw, Matched a IN) were used. Morselized autograft were added Control p-value Group 1 p-value Group 2 for small defects, and/or slices of allograft for larger defects. For reconstruction of the femur, Total hip arthroplasties (patients) 22 (22) 22 (21) 50 (48) Age, median 67 0.4 b 69 0.2 b 70 an uncemented titanium stem (Wagner SL) or range 53–84 48–88 40–88 a cemented stem (Müller straight stem; CDH, Male 15 0.2 c 10 0.1 c 29 d d Virtec) with gentamicin cement (Palacos) was Charnley classification 0.9 0.15 A 12 14 31 used (all Zimmer, Warsaw, IN). In 2-stage pro B1 1 1 0 cedures, reimplantation was performed after 3–5 B2 9 7 17 weeks (damaged soft tissue) or after 6–8 weeks C 0 0 2 Previous surgery performed 2 1.0 c 3 < 0.001 c 39 (difficult-to-treat bacteria). In group 2, a Burchon trochanter 2 1.0 d 2 0.4 d 17 Schneider (BS) reinforcement ring was needed on cup, for any reason 1 0.3 d 0 0.002 d 23 d d more often (Table 2). on stem, for any reason 2 1 2 0.01 25 PHJI treated in our center 1985–2004 91 hips

debridement(s) 0 1 d Girdlestone/spacer – Proprosky classification cup 0.5 d 1 13 2A 3 2B 2 2C 2 3A 2 3B 0 Proprosky classification stem 0.6 d I 0 II 17 IIIA 4 IIIB 1 IV 0 Deficient greater trochanter 1 1 c

0 < 0.001 d – f 0.004 d 15 1 2 2 1 0 f 0.2 d 1 16 3 1 0 1 0.05 c

a Control and group 1 were matched for all listed variables. b Student t-test c Chi-square test. d Mann-Whitney U-test. e 16 hips received ≥ 2 (range 2–10) debridements. f One preoperative radiograph was not available.

70 e 4/2 f

17 5 5 12 6 4 f

1 19 20 2 7 12

Infectiological, clinical, and radiographic evaluation The infection was classified as “cured” when there were no signs of infection 2 years or more after implantation (Giulieri et al. 2004). In case of death for an unrelated cause before the 2-year follow-up, infection was classified as “probably cured”. Relapse was defined as an infection with the same pathogen, and reinfection was defined as an infection with a different microorganism. Functional outcome parameters were the Harris hip score (HHS), the presence of limping, and the use of a support, and they were only recorded in patients with an event-free follow-up at 2 years (no revision and no death). This time point was chosen because after 2 years, results decline due to patient-related factors (Riede et al. 2007).


30

Acta Orthopaedica 2011; 82 (1): 27–34

Complications and reoperations were recorded for all the hips included. For radiographic follow-up, AP radiographs of pelvis and femur faux–profiles views were scrutinized for signs of loosening according to Gill et al. (1998) for the acetabulum and according to Harris (1982) for the femur. Migration and subsidence were measured (Callaghan et al. 1985). In case of osteotomy or bone lesions, images were screened for signs of non-union. All radiographic variables were obtained from the latest available radiograph. Endpoints for both clinical and radiographical follow-up were rerevision or death.

Figure 3. A 59-year-old woman with bilateral PI following Staphylococcus aureus sepsis, 2 years after bilateral 1-stage exchange. On the right side: reconstruction with a Müller reinforcement ring for a type-1 defect, and a long Wagner stem by transfemoral approach. On the left side: a reconstruction with a Burch-Schneider ring to bridge a type-2B defect. A cemented Virtec stem with transgluteal approach was implanted. At the 2-year follow-up, the HHS (bilateral) was 97, there was a slight limp, there were no radiographic signs of loosening, the transfemoral osteotomy was healed, and the infection was cured.

Table 2. Surgical procedures and specifics at the index operation Matched a Control Group 1 Group 2 Total hip arthroplasties (patients) 22 (22) p-value 22 (21) p-value 50 (48) No osteotomy 14 0.5 b 15 0.2 b 22 Transfemoral osteotomy 7 0.5 5 0.1 22 Greater trochanter osteotomy 1 1 1 0.5 1 Approach via nonunion in greater trochanter 0 0.9 1 0.2 5 – – – – 29 Spacer c Acetabular component SL uncemented cup 2 1 b 2 0.6 b 2 Müller reinforcement ring 16 1 16 0.08 24 Burch-Schneider reinforcement ring 4 1 4 0.03 23 c Femoral component 1 b 0.8 b uncemented 11 11 26 cemented 11 11 23 c Acetabular graft 0.2 d 0.9 d no graft 5 9 27 autograft, morselized 4 10 9 allograft, blocks > 2 cm or slices 2 3 10 auto- and allograft 1 0 3 a b c d

Control and group 1 were matched for all the variables listed. Chi-square test. In group 2, one patient died prior to reimplantation (see also complications, Table 4). Mann-Whitney U-test.

Statistics Statistical comparisons for patient characteristics, surgical specifics, and functional, radiographic, and microbiological outcomes were conducted with chi-square and Student’s t-test. For nonparametric data, the Mann-Whitney U test was used. For all 84 hips with an event-free follow-up, multiple univariate regression analysis (ANOVA) was performed to determine whether surgical parameters lead to lower functional outcome(s). Calculations were done with SPSS software version 15.0 and statistical significance was set at p < 0.05. Ethics Informed consent was obtained from all patients. The study design was approved by the local ethics committee of the hospital district in which the study was conducted (01-05-2006; WN 420).

Results Infectiological outcome In group 1, there was no relapse of infection but there was 1 case of reinfection with another pathogen. In group 2, there was 1 relapse of infection (Table 3). Functional and radiographic outcome, revisions, and complications Control vs. group 1. We did not find any statistically significant differences in the functional parameters HHS, limping, and


Acta Orthopaedica 2011; 82 (1): 27–34

31

use of support—or in radiographic parameters—between the 2 groups. Definitive loosening of Control p-value Group 1 p-value Group 2 the stem with subsequent revi Infectiological outcome 0.5 a sion occurred in 2 hips in group total hip arthroplasties 22 22 50 1 and in 1 hip in the control cured – – 19 46 group (Table 3). 1 cup in the probably cured – – 2 1 relapse of infection with same pathogen – – 0 1 control group showed signs of reinfection with different pathogen 0 0.2 b 1 0 definitive loosening but the hip unknown – – 0 2 score was 100 and revision was Functional outcome at 2-year follow-up total hip arthroplasty c 22 19 43 not performed. All osteotomies mean HHS (SD) 85 (16) 0.9 d 84 (17) 0.3 d 80 (18) and the pre-existing nonunion mean HHS with/without spacer – – – – 80/78 of the greater trochanter healed limping 0.5 a 0.5 a none or slight 19 15 30 successfully. 1 iatrogenic frac moderate or severe 3 4 13 ture of the greater trochanter that walking 0.2 a 0.5 a occurred in group 1 developed without support 15 9 18 with one cane/crutch 7 8 13 a nonunion. The total number with two canes/crutches 0 2 11 of postoperative complications unable to walk/use of wheelchair 0 0 1 and re-interventions was similar Radiographic outcome at last follow-up e total hip arthroplasties 22 20 f 47 f in both groups (Table 4). mean follow-up in years (SD) 4.8 (2.8) 0.05 d 3.3 (2.4) 0.5 d 4.0 (3.7) Group 1 vs. group 2. Compardefinitive loosening of stem (revised) 1 (1) 0.5 b 2 (2) 0.9 b 4 (2) ing both groups, all functional definitive loosening of cup (revised) 1 (0) 0.2 b 0 0.1 b 3 (1) stem subsidence > 5 mm (revised) 0 – 0 0.6 b 2 (0) outcomes were lower in group 2 but the differences were not a Mann-Whitney U-test. b Chi-square test. statistically significant (Table c Only hips with an event-free survival for ≥ 2 years were included in functional outcome analyses. 3). Also, radiographic results d Student t-test. were not significally differe Statistical analysis was performed with endpoints definitive component loosening and stem subsident between groups 1 and 2. In ence > 5 mm. f For 2 hips, in group 1 and group 2, radiographs were not available since the patients died prior to the group 2, 4 stems showed signs follow-up examinations. Also, postoperative radiographs were not available for another hip in group 2. of definitive loosening and 2 of those were revised; the other 2 stems showed subsidence of 4 mm and 28 mm, but they remained stable and patient co-morbidity did not allow Table 4. Number of complications associated with the index operation intervention. Also, 3 cups were definitively loose: 1 of them (a Müller reinforcement Control p-value Group 1 p-value Group 2 ring) was revised and the other 2 (BS-rings) did not cause symptoms (Table 3). 1 of Total hip arthroplasties 22 (22) 22 (21) 50 (48) Surgical complications the transfemoral osteotomies and the only fracture, greater trochanter 3 1 1 trochanteric osteotomy developed a nonfracture/fissure, proximal femur 1 3 2 union. 2 of the 5 pre-existing nonunions of Post-surgical complications requiring reoperation the greater trochanter persisted. All intra2 1 17 hematoma operative fractures of the proximal femur 0 0 1 wound infection healed uneventfully. Group 2 had a higher – – 1 pin tract infection requiring closed reduction incidence of (post-) surgical complications, – – 1 dislocation of spacer mainly due to more postoperative hema0 0 6 dislocation of hip tomas and dislocations, and had more re2 0.5 a 1 0.004 a 26 total reinterventions Total (post-) surgical complications 6 0.9 a 5 0.03 a 29 interventions (Table 4). Table 3. Infectiological, functional, and radiographic outcome, and number of revisions

Non-surgical complications during hospitalization thrombosis/emboli 0 0 2 miscellaneous 1 3 4 early death 0 0 2b a Mann-Whitney U-test. b Death due to heart failure and pneumonia, one month after the index operation.

Surgery-related parameters and functional outcome The use of a transfemoral osteotomy was not associated with lower functional outcome (p = 0.1 for HHS; p = 0.2 for limping;


32

p = 0.9 for use of support). A deficient greater trochanter (p = 0.01 for limping; p = 0.05 for use of support), the use of a BS-ring (p = 0.03 for limping), a trochanteric osteotomy (p = 0.05 for use of support), and a history of multiple cup revisions (p = 0.06 for use of support) were associated with lower functional outcome.

Discussion A commonly proposed treatment strategy for PI is a 2-stage procedure with up to a 6-month interval between surgeries (McDonald et al. 1989, Berry et al. 1991, Colyer and Capello 1994, Lieberman et al. 1994, Nestor et al. 1994, Garvin and Hanssen 1995, Lai et al. 1996, Wang and Chen 1997). However, the use of a 1-stage exchange provides several advantages such as lower perioperative morbidity, a shorter hospital stay, lower costs, and earlier rehabilitation (Buchholz et al. 1981, Wroblewski 1986, Raut et al. 1994, 1995, 1996, Ure et al. 1998, Bozic and Ries 2005). Yet, the results of some reports indicate the possibility of a higher relapse rate (Raut et al. 1994, 1995). This implies that only selected patients may benefit from a 1-stage procedure. With this line of reasoning, we have established an algorithm that selects patients for either a 1-stage exchange or a 2-stage exchange according to welldefined criteria (Zimmerli et al. 2004). In agreement with our previous studies (Giulieri et al. 2004, Sendi et al. 2006), we found a very low relapse rate of infection in the patients in the present study. In the study, we focused on functional outcome. This study had several limitations. First, the number of patients was small. On the other hand, our analysis included only hips that were treated according to a strict protocol with a multidisciplinary approach, and there was almost no relapse. Inclusion of the simultaneous analysis of functional, radiographic, and infectiological parameters is crucial when judging the outcome of a treatment algorithm for PI. Also, we applied strict methodological criteria to functional outcome evaluation by only including hips that had an event-free follow-up of at least 2 years. Given these study constraints, it is conceivable that the sizes of the groups were relatively small. We cannot exclude the possibility that the statistically insignificant differences we found between infectious and non-infectious hips might have become significant in a larger series. However, in our view it is not feasible to perform a study with the high numbers of hips required from a statistical point of view that would also include a standardized treatment and a multidisciplinary follow-up as performed in our study. We believe that our analysis was conducted on a solid study population and reflects clinical reality. Secondly, the preoperative functional scores were not available for the hips included in the study. This partially limits the interpretation of the 2-year HHS. However, many patients had had previous surgery and complaints due to infection; thus, such preoperative scores would not allow any reliable comparison with the postoperative state.

Acta Orthopaedica 2011; 82 (1): 27–34

Thirdly, we used several different implants. As in any revision surgery, the variability in quality of bone is often substantial, requiring an individual choice of implant; our algorithm was not bound to any preferred reconstruction technique. Finally, functional outcome was based on only 3 variables (i.e. HHS, limping, and walking support). We focused on simple variables, however, to allow comparison between groups, and between our results and those reported in the literature. HHS was almost identical in the 1-stage group and in the control group (mean HHS of 84 and 85). Moreover, the values were higher when compared to a report of 31 one-stage procedures for infection, in which the mean HHS was 75 after 3.5 years (Tsukayama et al. 1996). In a recent meta-analysis of 24 studies analyzing mainly aseptic revisions, the mean HHS was 82 after a mean follow-up of 4 years (Saleh et al. 2003). The high functional scores after 1-stage exchange can be partly explained by the design of the algorithm, because it selects less complex cases for such procedures. This means that the algorithm selects more complex cases for a 2-stage exchange. These cases had moderately or severely damaged soft tissue (i.e. sinus tract, multiple previous surgeries), or involved a difficult-to-treat microorganism. Since a 2-stage exchange is only performed in cases of infection, there was no matched control group available. Even so, the functional scores of the 2-stage group were good (mean HHS of 80), although they were lower than those of the 1-stage group (mean HHS of 84). A difference of 4 points may be clinically important (Hoeksma et al. 2003) and such a discrepancy is not surprising considering the preoperative conditions—including multiple surgeries, more serious infections, limited bone stock, and deficiency of the greater trochanter. Consequently, the surgical procedure becomes more complex. The Harris hip scores in our study (80–85) are similar to those in the literature, where the scores reported have ranged from 72 to 91 (Lai et al. 1996, Haddad et al. 2000). A transfemoral osteotomy, an important feature of our surgical technique, was not associated with poor functional outcome. With this technique (Wagner and Wagner 1999), safe removal of the components and a thorough debridement can be achieved. None of the osteotomies were associated with a periprosthetic fracture; the only relapse of infection occurred in a case of osteotomy but appeared not to be related to this technique, and another developed a nonunion. An important aspect of our treatment strategy is that we strive to preserve bone and soft tissue as much as possible in order to optimize the functional result. This contrasts with other surgical regimens where infected tissue is radically debrided or even resected (Friesecke and Wodtke 2008, Wodtke and Lohr 2008). In our study, infections were eliminated in almost all cases without the need for aggressive removal of bone and soft tissue. We found that a trochanter deficiency and the use of a BS-ring were associated with an increased risk of limping. Whereas the occurrence of limping may be explained by the presence of a deficient trochanter with impairment of the


Acta Orthopaedica 2011; 82 (1): 27–34

abductor apparatus, this appears to be less evident with the use of a BS-ring. It has been proposed that the exposure of the ileum may damage the anterior part of the gluteus medius muscle or the superior gluteal nerve (Possai et al. 1996, Perez et al. 2004, Ikeuchi et al. 2006). However, this finding may also be due to a confounding factor, considering that a BSring was implanted in 6 of 14 patients with a trochanter deficiency. The uncemented Wagner SL stem was associated with a good functional outcome and in contrast to previous reports (Hartwig et al. 1996, Kolstad et al. 1996, Bircher et al. 2001), stem subsidence was uncommon. The incidence of loosening of acetabular and femoral components was low and similar to that for identical implants used in aseptic revision surgery with equal length of follow-up (Haentjens et al. 1986, Korovessis et al. 1992, Peters et al. 1995, Ilchmann et al. 2006). Considering the criteria in our algorithm (including the causative pathogen), we found that some of the referred cases should have been treated with a 1-stage exchange or a 2-stage exchange a priori, instead of having inadequate trials of debridement with implant retention or 1-stage exchange, respectively. Initial treatment failure with prosthetic infections increases the probability of a requirement for a more complex procedure (2-stage exchange) with a greater burden for the patients, and might therefore end in an inferior functional outcome. Our findings support the idea of using well-defined criteria in order to select an optimal surgical strategy for patients with prosthetic infections. In our experience, the algorithm used here provides such criteria, and is therefore helpful in the surgical decision making. In this way, relapse of infection and impairment of the functional outcome can be avoided.

FHRDEM: study design, patient follow-up, preparation of data and manuscript writing. PS: infection treatment and manuscript writing. WZ: infection treatment and study design. ThBM: surgical treatment and patient follow-up. PEO: surgical treatment, patient follow-up and study design. ThI: patient follow-up, study design and manuscript writing.

We thank Peter Graber and Martin Lüem for their outstanding contributions to this study, Petra Heesterbeek and Patsy Anderson for statistical help, and Susanna Häfliger for secretarial assistance.

Berry D J, Chandler H P, Reilly D T. The use of bone allografts in two-stage reconstruction after failure of hip replacements due to infection. J Bone Joint Surg (Am) 1991; 73: 1460-8. Bircher H P, Riede U, Luem M, Ochsner P E. The value of the Wagner SL revision prosthesis for bridging large femoral defects. Orthopade 2001; 30: 294-303. Bozic K J, Ries M D. The impact of infection after total hip arthroplasty on hospital and surgeon resource utilization. J Bone Joint Surg (Am) 2005; 87: 1746-51. Britton A R, Murray D W, Bulstrode C J, McPherson K, Denham R A. Pain levels after total hip replacement: their use as endpoints for survival analysis. J Bone Joint Surg (Br) 1997; 79: 93-8.

33

Buchholz H W, Elson R A, Engelbrecht E, Lodenkamper H, Rottger J, Siegel A. Management of deep infection of total hip replacement. J Bone Joint Surg (Br) 1981; 63: 342-53. Callaghan J J, Salvati E A, Pellicci P M, Wilson P D, Jr., Ranawat C S. Results of revision for mechanical failure after cemented total hip replacement, 1979 to 1982. A two to five-year follow-up. J Bone Joint Surg (Am) 1985; 67: 1074-85. Colyer R A, Capello W N. Surgical treatment of the infected hip implant. Two-stage reimplantation with a one-month interval. Clin Orthop 1994; (298): 75-9. Friesecke C, Wodtke J. Management of periprosthetic infection. Chirurg 2008; 79: 777-92. Garvin K L, Hanssen A D. Infection after total hip arthroplasty. Past, present, and future. J Bone Joint Surg (Am) 1995; 77: 1576-88. Gill T J, Sledge J B, Muller M E. Total hip arthroplasty with use of an acetabular reinforcement ring in patients who have congenital dysplasia of the hip. Results at five to fifteen years. J Bone Joint Surg (Am) 1998; 80: 969-79. Giulieri S G, Graber P, Ochsner P E, Zimmerli W. Management of infection associated with total hip arthroplasty according to a treatment algorithm. Infection 2004; 32: 222-8. Haddad F S, Muirhead-Allwood S K, Manktelow A R, Bacarese-Hamilton I. Two-stage uncemented revision hip arthroplasty for infection. J Bone Joint Surg (Br) 2000; 82: 689-94. Haentjens P, Handelberg F, Casteleyn P P, Opdecam P. The Muller acetabular support ring. A preliminary review of indications and clinical results. Int Orthop 1986; 10: 223-30. Harris W H, McCarthy J C, Jr., O’Neill D A. Femoral component loosening using contemporary techniques of femoral cement fixation. J Bone Joint Surg (Am) 1982; 64: 1063-7. Hartwig C H, Bohm P, Czech U, Reize P, Kusswetter W. The Wagner revision stem in alloarthroplasty of the hip. Arch Orthop Trauma Surg 1996; 115: 5-9. Hoeksma H L, van den Ende C H, Ronday H K, Heering A, Breedveld F C. Comparison of the responsiveness of the Harris Hip Score with generic measures for hip function in osteoarthritis of the hip. Ann Rheum Dis 2003; 62: 935-8. Ikeuchi M, Kawakami T, Yamanaka N, Okanoue Y, Tani T. Safe zone for the superior gluteal nerve in the transgluteal approach to the dysplastic hip: intraoperative evaluation using a nerve stimulator. Acta Orthop 2006; 77: 603-6. Ilchmann T, Gelzer J P, Winter E, Weise K. Acetabular reconstruction with the Burch-Schneider ring: an EBRA analysis of 40 cup revisions. Acta Orthop 2006; 77: 79-86. Kolstad K, Adalberth G, Mallmin H, Milbrink J, Sahlstedt B. The Wagner revision stem for severe osteolysis. 31 hips followed for 1.5-5 years. Acta Orthop Scand 1996; 67: 541-4. Korovessis P, Spastris P, Sdougos G, Salonikides P, Christodoulou G, Katsoudas G. Acetabular roof reinforcement rings. Clin Orthop 1992; (283): 149-55. Lai K A, Shen W J, Yang C Y, Lin R M, Lin C J, Jou I M. Two-stage cementless revision THR after infection. 5 recurrences in 40 cases followed 2.5-7 years. Acta Orthop Scand 1996; 67: 325-8. Lieberman J R, Callaway G H, Salvati E A, Pellicci P M, Brause B D. Treatment of the infected total hip arthroplasty with a two-stage reimplantation protocol. Clin Orthop 1994; (301): 205-12. McDonald D J, Fitzgerald R H, Jr., Ilstrup D M. Two-stage reconstruction of a total hip arthroplasty because of infection. J Bone Joint Surg (Am) 1989; 71: 828-34. Nestor B J, Hanssen A D, Ferrer-Gonzalez R, Fitzgerald R H, Jr. The use of porous prostheses in delayed reconstruction of total hip replacements that have failed because of infection. J Bone Joint Surg (Am) 1994; 76: 349-59. Paprosky W G, Perona P G, Lawrence J M. Acetabular defect classification and surgical reconstruction in revision arthroplasty. A 6-year follow-up evaluation. J Arthroplasty 1994; 9: 33-44.


34

Perez M M, Llusa M, Ortiz J C, Lorente M, Lopez I, Lazaro A, Perez A, Gotzens V. Superior gluteal nerve: safe area in hip surgery. Surg Radiol Anat 2004; 26: 225-9. Peters C L, Curtain M, Samuelson K M. Acetabular revision with the BurchSchnieder antiprotrusio cage and cancellous allograft bone. J Arthroplasty 1995; 10: 307-12. Possai K W, Dorr L D, McPherson E J. Metal ring supports for deficient acetabular bone in total hip replacement. Instr Course Lect 1996; 45: 161-9. Raut V V, Siney P D, Wroblewski B M. One-stage revision of infected total hip replacements with discharging sinuses. J Bone Joint Surg (Br) 1994; 76: 721-4. Raut V V, Siney P D, Wroblewski B M. One-stage revision of total hip arthroplasty for deep infection. Long-term followup. Clin Orthop 1995; (321): 202-7. Raut V V, Orth M S, Orth M C, Siney P D, Wroblewski B M. One stage revision arthroplasty of the hip for deep gram negative infection. Int Orthop 1996; 20: 12-4. Riede U, Luem M, Ilchmann T, Eucker M, Ochsner P E. The M.E Muller straight stem prosthesis: 15 year follow-up. Survivorship and clinical results. Arch Orthop Trauma Surg 2007; 127: 587-92. Saleh K J, Celebrezze M, Kassim R, Dykes D C, Gioe T J, Callaghan J J, Salvati E A. Functional outcome after revision hip arthroplasty: a metaanalysis. Clin Orthop 2003; (416): 254-64. Sendi P, Rohrbach M, Graber P, Frei R, Ochsner P E, Zimmerli W. Staphylococcus aureus small colony variants in prosthetic joint infection. Clin Infect Dis 2006; 43: 961-7.

Acta Orthopaedica 2011; 82 (1): 27â&#x20AC;&#x201C;34

Tsukayama D T, Estrada R, Gustilo R B. Infection after total hip arthroplasty. A study of the treatment of one hundred and six infections. J Bone Joint Surg (Am) 1996; 78: 512-23. Ure K J, Amstutz H C, Nasser S, Schmalzried T P. Direct-exchange arthroplasty for the treatment of infection after total hip replacement. An average ten-year follow-up. J Bone Joint Surg (Am) 1998; 80: 961-8. Valle C J, Paprosky W G. Classification and an algorithmic approach to the reconstruction of femoral deficiency in revision total hip arthroplasty. J Bone Joint Surg (Am) (Suppl 4) 2003; 85: 1-6. Wagner M, Wagner H. Der transfemorale Zugang zur Revision von Huftendoprothesen. Oper Orthop Traumatol 1999; 11: 278-95. Wang J W, Chen C E. Reimplantation of infected hip arthroplasties using bone allografts. Clin Orthop 1997; (335): 202-10. Wodtke J, Lohr J F. The infected implant. Orthopade 2008; 37: 257-67. Wroblewski B M. One-stage revision of infected cemented total hip arthroplasty. Clin Orthop 1986; (211): 103-7. Zappe B, Graf S, Ochsner P E, Zimmerli W, Sendi P. Propionibacterium spp. in prosthetic joint infections: a diagnostic challenge. Arch Orthop Trauma Surg 2008; 128: 1039-46. Zimmerli W, Trampuz A, Ochsner P E. Prosthetic-joint infections. N Engl J Med 2004; 351: 1645-54.


Acta Orthopaedica 2011; 82 (1): 35–41

35

Validation of a 3D CT method for measurement of linear wear of acetabular cups A hip simulator study Anneli Jedenmalm1,2, Fritjof Nilsson2, Marilyn E Noz3, Douglas D Green4, Ulf W Gedde2, Ian C Clarke4, Andreas Stark5, Gerald Q Maguire Jr6, Michael P Zeleznik7, and Henrik Olivecrona5 1Department of Clinical Sciences, Biomaterial- and Biomechanical Laboratory, Lund University, Lund, Sweden; 2Department of Fiber and Polymer Technology, School of Chemical Science and Engineering, Royal Institute of Technology, Stockholm, Sweden; 3Department of Radiology, New York University School of Medicine, New York, NY, USA; 4Orthopaedic Research Center, Loma Linda University, Loma Linda, CA, USA; 5Department of Molecular Medicine and Surgery, Karolinska Institute, Stockholm, Sweden; 6School of Information and Communication Technology, Royal Institute of Technology, Stockholm, Sweden; 7Saya Systems Inc., Salt Lake City, UT, USA Correspondence AJ: annelije@kth.se Submitted 10-04-13. Accepted 10-08-16

Background We evaluated the accuracy and repeatability of a 3D method for polyethylene acetabular cup wear measurements using computed tomography (CT). We propose that the method be used for clinical in vivo assessment of wear in acetabular cups. Material and methods Ultra-high molecular weight polyethylene cups with a titanium mesh molded on the outside were subjected to wear using a hip simulator. Before and after wear, they were (1) imaged with a CT scanner using a phantom model device, (2) measured using a coordinate measurement machine (CMM), and (3) weighed. CMM was used as the reference method for measurement of femoral head penetration into the cup and for comparison with CT, and gravimetric measurements were used as a reference for both CT and CMM. Femoral head penetration and wear vector angle were studied. The head diameters were also measured with both CMM and CT. The repeatability of the method proposed was evaluated with two repeated measurements using different positions of the phantom in the CT scanner. Results The accuracy of the 3D CT method for evaluation of linear wear was 0.51 mm and the repeatability was 0.39 mm. Repeatability for wear vector angle was 17°. Interpretation This study of metal-meshed hip-simulated acetabular cups shows that CT has the capacity for reliable measurement of linear wear of acetabular cups at a clinically relevant level of accuracy. 

Debris due to polyethylene wear may trigger aseptic loosening in total hip arthroplasty (THA) (Santavirta et al. 1995, Mohanty 1996). With the methods currently available, it is difficult to determine small amounts of wear in vivo without using invasive methods such as radiostereometric analysis

(RSA) (Schewelov et al. 2004). Conventional radiography is the most common routine clinical method, but it yields only 2D results with low accuracy—of about 4 mm (Clarke et al. 1976). Several 3D reconstruction methods exist, such as AP radiographs, but these methods are still too complicated to be reliable in a routine clinical setting and the accuracy tends to be lower in a clinical situation than under laboratory conditions (Clarke et al. 1976). Current multislice CT scanners, which offer accurate spatial volume resolution in both 2D and 3D without substantial distortion, are non-invasive and fast. Metal artifacts from the implant are suitably suppressed by software algorithms from the CT manufacturers. We have previously shown that CT can also be used for evaluation of acetabular cup position and migration (Olivecrona et al. 2002, 2003c, 2004). A retrieval study showed that CT can also be used for evaluation of 3D penetration of the femoral head into metal-backed acetabular cups with an accuracy of 1 mm (Olivecrona et al. 2005). Based on that study, we developed a new approach for wear assessment using CT and it was shown to achieve an accuracy of 0.6 mm and a repeatability of 0.4 mm (Olivecrona et al. 2005). This new approach relies on placing several landmarks in the 3D CT volume on the surfaces of the head and cup. Spheres are then fitted to these data points, and from these spheres the femoral head penetration can be calculated. As part of this study, the method described was developed futher—mainly with improved software that allowed many more landmark points to be placed on the 3D surfaces, and the possibility of comparing the same implant at 2 time points (e.g. pre-wear and post-wear). A model study was performed with a hip simulator in order to estimate the accuracy and reproducibility of the method proposed. Both gravimetric results and

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2011.552777


36

Acta Orthopaedica 2011; 82 (1): 35–41

after 0, 0.4, 1.15, 1.65, 2.5, and 2.8 million cycles (Mc). 4 cups (2 unsterile and 2 sterilized) were used as soak controls in serum. Wear rate was determined by weighing all cups, and it was corrected for soak. Briefly, the cups were ultrasonically cleaned, blow-dried with nitrogen gas, vacuum dried for 30 min, and then weighed 4 times in rotation with the microbalance at room temperature. Wear volume was calculated by dividing the gravimetric data by the density of the cups (935 kg/m3). Figure 1. Polyethylene acetabular cup with titanium mesh molded on the outside.

measurement results from a coordinate measurement machine (CMM) were used as reference methods. We used a different cup design with a metal mesh molded into the polyethylene outer surface, in order to avoid back-side wear. Due to an unanticipated upgrade of the CT software, we also simulated the clinical situation where a CT scan is done postoperatively and then at a follow-up several years later, by using a different CT scan protocol before and after wear.

Material and methods We studied 12 commercial ultra-high molecular weight (GUR1020 UHMWPE) cups (Zimmer) with a molded titanium mesh on the outside (Figure 1). According to the manufacturer, the inner diameter was 28 mm, whereas the outer diameter ranged from 48 to 64 mm. The femoral heads were Durasul cobalt-chromium (CoCr) heads (Zimmer). They were manually roughened with P320-grit SiC paper, producing circular multidirectional wear tracks before the tests in order to increase the wear rate. The clearance between head and cup diameter was measured with a coordinate measurement machine (CMM) (Wenzel LH44; Wenzel, Wiesthal, Germany; accuracy in 3D: (3 + L/300) µm; maximum measured deviation at calibration: 2.4 µm) and was estimated to be 0.51 mm (SD 0.01). All cups were measured using the CMM, a microbalance (resolution: ± 0.01 mg; accuracy: ± 0.1 mg; Sartorius AG, Göttingen, Germany) and CT scanner (LightSpeed QX/I; General Electric Medical Systems, Milwaukee, WI) before and after wear testing in a hip simulator (Shore Western, Monrovia, CA). Hip simulator wear testing with gravimetric measurement All cups were presoaked in distilled water for 4 weeks before wear testing, in order to stabilize the soak rate. 8 cups were wear tested together with femoral heads in a hip simulator, with bovine serum as lubricant (a 10% solution in order to increase the wear rate (Wang et al. 1998, 2004)) until a total wear of at least 200 mg was reached. Weighing was performed

Wear testing with measurement by CT examination A model made of polymethyl methacrylate was used to fix the implant in an anatomical position at a 45° angle to the sagittal plane. A titanium stem (attached to this model) was used to fix the femoral head in position in the implant. Pen marks were then used to allow recreation of this relative orientation of the implant and femoral head at different measurement times. 2 CT examinations of each implant were done. The data volumes were acquired with 1.25 mm collimation and a pitch of 3 (i.e. 0.75 mm/rotation) at 120 kV. The current was 40 mA at the measurement before wear, and then changed to 100 mA at measurement after wear, due to an unanticipated CT software upgrade that can well happen in the clinical situation with hospital-based CT units—since it is common to change acquisition parameters and upgrade CT software on an annual basis. Slices were reconstructed at 1.25-mm increments. The implants were scanned in approximately the same positions as when implanted in a patient. To mimic different patient positions, the position of the model in the CT was altered between the scans. The changes in position of the model were larger than would be expected with patients. The height of the table was 81–83 cm, and it was not changed from scan to scan or during a scan. CT image analysis The CT method developed had 2 distinct parts: (1) image analysis tools for obtaining CT coordinate data, and (2) locally developed software for determining linear wear and wear angles from the coordinate data. Image analysis of the CT volumes was performed using a validated 3D volume fusion tool (Maguire et al. 1991, Noz et al. 2001, Gorniak et al. 2003). The first step, for each CT volume, was to designate 3D coordinate points to define (1) the outer surface of the metal shell of the cup, (2) the femoral head surface, and (3) the 2 cup opening faces. Using 3D isosurfaces and a 3D point selection mode, evenly distributed coordinate points were designated that corresponded to the outer surface of the cup (1,000–4,000 points) and the femoral head surface (2,500–5,000 points) (Figure 2A). In order to filter out metal artifacts, all points positioned more than 0.5 mm (i.e. ca. 1 SD) from a spherical surface were eliminated. Using the same strategy, about 200–800 coordinate points on the cup opening faces were also designated and points positioned more than 0.5 mm from a flat surface were eliminated.


Acta Orthopaedica 2011; 82 (1): 35–41

37

Figure 2. A. Digitized points from CT scan image of head and cup. Values are in mm.

Figure 2. B. Example of CMM measurement. Digitized points on cup inner surface and planes, center points before and after wear (square hollow symbols), and coordinate axes (blue lines). Values are in mm.

The second step was to register the volumes for each pair of CT volumes (pre-wear and post-wear) into a single coordinate system, thus allowing comparison of femoral head and cup positions. The software for calculating wear properties from CT data was based on ideas previously reported (Olivecrona et al. 2003a, b, Jedenmalm et al. 2008), but used new algorithms giving substantially improved numerical accuracy and with several additional utility functions: for instance, algorithms for calculating wear vector angles. The computer program was implemented in Matlab and ComsolScript. Using non-linear optimization methods, planes were fitted to the CT points on the cup opening faces and spheres were fitted to the points on (1) the cup surface, and (2) the femoral head surface. This was done both before and after wear. Local coordinate frames for each were calculated from the normal axes of the plane of the cup opening and the midpoints of the cup surface. Coordinate transformations were then used to transfer all data to a common global coordinate frame, as defined by the head surface sphere, such that the head positions in the worn and unworn cup could be compared. Linear wear was defined as the distance between midpoints of the head sphere in worn and unworn systems. The wear angles (θ and ϕ) were defined as the angles of a standard spherical coordinate system centered in the unworn head.

method using 8,000 evenly distributed digitizing points; the geometry of the cup opening face planes was digitized using about 30 points (Figure 2B) and that of the femoral head surfaces using 200 points. The locally developed software for calculating the maximum linear penetration depth, and wear direction, resembled its CT equivalent. The major differences were twofold: (a) instead of calculating the local coordinate axes with the metal part of the cup opening face planes and the cup outer surface, the plastic part of the cup opening face planes and the ultimately unworn (left) side of the inner part of the cup were used; (b) instead of defining the linear wear as movement of the head due to wear, we defined it in terms of the change in the inner, ultimately worn (right) hemisphere of the cup—specifically as the movement of the midpoint of a sphere fitted to that inner hemisphere. The asymmetry in wear was due to the way in which wear is induced by the hip simulator.

Wear testing with measurement by coordinate measurement machine (CMM) The CMM procedure, just like the CT methodology, consisted of (1) tools for obtaining geometric data, and (2) locally developed software for analyzing the data. However, since it is well established that geometric data obtained by CMM measurements are reliable, the CMM procedure was used as the reference method, together with the gravimetric measurements. The fundamental principles of CMM have been described previously (Olivecrona et al. 2005). In this particular study, the inside of the polyethylene cups was examined with the CMM

Statistical evaluation of errors Data were tested for normality using Bland-Altman plots, and the graphical evaluation with line of equality and scatter plots. Confidence intervals were calculated using the Student’s t-distribution. Accuracy, repeatability, and bias of the CT method were calculated with the 95% confidence level according to ISO definitions (ISO 3534: 3,11 & 3,14) (ISO 1998). Precision is defined as the closeness of agreement between independent test results obtained under stipulated conditions. For repeatability, as a way to express precision we used the ISO standards, where the method under evaluation is the combined error of acquisition of data with a CT unit and image postprocessing (ISO 3534: 3,14) (ISO 1998). Accuracy is defined as the closeness of agreement of a test result (CT in this study) and the accepted reference value (the CMM and gravimetric results). The term accuracy when applied to a set of test results refers to a combination of systematic errors (bias) and random errors (ISO 3534: 3,11) (ISO 1998).


38

Acta Orthopaedica 2011; 82 (1): 35–41

Results An example of the 3D geometry of the CMM measurement and results, with angles ϕ and θ and wear vector, is given in Figure 3.

Figure 3. Representative 3D plot with the coordinate axes of linear wear surface distribution and the angles ϕ and θ to the wear vector in blue. [Authors: this sentence is difficult to follow] The color bar represents the magnitude of wear in mm.

We tested the method for systematic errors by computing an interval estimate of the bias (the mean error and the 95% confidence interval for the mean). Accuracy and repeatability were calculated at the 95% confidence level as described by Ranstam et. al. (2000).

Test for systematic differences If the differences obey the normal distribution, then 95% of the differences will lie within the 95% limits of agreement; that is, the mean difference ± 1.96 times the standard deviation. Most of the data points were within these limits. The difference plot for femoral head measurement (Figure 4A) revealed no systematic difference between CT and CMM. The plot for linear wear showed a slight underestimation of the wear (Figure 4B). The plot for angle ϕ (Figure 4C) shows a negative mean difference value that suggests an underestimation with CT, whereas the plot for angle θ (Figure 4D) shows the opposite behavior with a positive mean difference, suggesting an overestimation with CT. Accuracy and repeatability along with an interval estimate of bias for the experiments are summarized in Table 1. No significant systematic differences were found between repeated CT scans. The landmark points showed a normal distribution.

a

b

c

d

Figure 4. Difference against mean for CT and CMM measurements a) of head diameter b) linear wear c) Angle ϕ d) Angle θ.


Acta Orthopaedica 2011; 82 (1): 35–41

39

Table 1. Repeatability, accuracy, and interval estimate of bias at 95% confidence level for the proposed CT method

Repeatability Accuracy (SD) (SD)

Linear wear, mm Angle θ, degrees Angle ϕ, degrees Head diameter, mm

0.39 (0.19) 17 (179) 16 (26) 0.11 (0.06)

0.51 (0.14) 106 (144) 50 (18) 0.20 (0.04)

Interval estimate of bias –0.12 ± 0.10 89 ± 40.9 –33 ± 5.2 0.09 ± 0.03

Table 2. Mean (SD) linear wear (mm) results from CT and CMM measurements Cup no. 1 2 3 4 5 6 7 8 9 10 11 12 All cups

CT

CMM

Difference

0.65 (0.14) 0.89 (0.15) 0.81 (0.11) 0.20 (0.16) 2.22 (0.12) 2.18 (0.13) 0.01 (0.14) 0.17 (0.07) 1.00 (0.12) –0.01 (0.10) 0.46 (0.12) 1.04 (0.14)

1.55 (0.21) 1.23 (0.06) 1.15 (0.16) 0.08 (0.14) 2.54 (0.27) 1.69 (0.25) 0.02 (0.23) 0.02 (0.24) 1.26 (0.09) 0.05 (0.08) 0.60 (0.08) 0.91 (0.13)

–0.90 –0.34 –0.34 0.12 –0.32 0.49 –0.00 0.15 –0.26 –0.06 –0.14 0.13

0.80 (0.13)

0.92 (0.16)

–0.12

Analysis of wear Linear wear of the 12 cups assessed with CMM varied between 0 and 2.54 mm, and the mean difference between CT and CMM measurements of linear wear was –0.12 (95% CI: –0.02 to –0.23) mm (Table 2). Volumetric wear of the 12 cups assessed by weighing varied between –5 and 936 mm3. Comparison of the ranking order (lowest to highest wear) showed excellent agreement between gravimetric results and CMM. In the same comparison, CT also agreed well with the gravimetric results except for cups no. 1 and 6. Analysis of creep From CMM measurements, the creep could be estimated from the soak controls (cups no. 4, 7, 8, and 10). The average creep was estimated to be about 0.05 mm. Analysis of the head and cup diameters The head diameters varied between 27.93 and 27.99 mm (SD 0.02 mm) according to CMM. The mean difference in head diameter before wear between CT and CMM measurements was 0.09 (95% CI: 0.06 to 0.12) mm, and after wear it was –0.09 (95% CI: –0.13 to –0.6) mm. For both head diameter and cup diameter, no systematic measurement differences were found due to different position-

Table 3. Measurements (by CT) of outer diameter of cups, and of femoral head diameter Object Cup unworn Cup worn Head unworn Head worn

Mean diameter) in mm (SD

Mean SD of repeated measurements

54.37 (5.14) 54.77 (5.12) 28.04 (0.06) 27.86 (0.05)

0.034 0.030 0.041 0.045

Table 4. Mean standard deviation (SD) of points fitted to the reference planes Object Mean fit SD Plane 1 unworn Plane 1 worn Plane 2 unworn Plane 2 worn

0.195 (0.106) 0.149 (0.060) 0.190 (0.028) 0.164 (0.053)

Mean SD of repeated measurements 0.058 (0.075) 0.038 (0.038) 0.020 (0.019) 0.030 (0.019)

Values in parentheses are SD of mean SD.

ing or repeated measurement of the same object. The average difference was 0.001 mm (SD 0.01) and the maximum difference for the same object was 0.15 mm. The outer diameter of the cup could not be accurately measured with direct measurement since the metal mesh provided an uneven surface, which made automatic measurement with CMM difficult due to the sensitivity of the CMM probe—and it would be too time-consuming to do it manually. However, the outer diameter and the repeatability of the measurements were measured with CT before and after wear (Table 3). The mean diameter was based on all 48 measurements ± the standard deviation. Repeatability was calculated by taking the standard deviation of the 4 measurements of each prostheses and taking the mean of those twelve values. The outer diameter stated by the manufacturer varied between 46 mm and 64 mm, but no bias due to size could be found. The average difference in outer diameter between worn and unworn cups was 0.40 ± 0.094 mm, whereas for femoral heads the average difference was –0.18 ± 0.068 mm. Since the positions of the cup centers are used as reference points in calculating the wear, it is reasonable to expect a bias of linear wear values of about half the cup diameter, i.e. approximately 0.2 mm. Analysis of fitting of the planes The average standard deviations of points fitted to the planes are tabulated in Table 4. The comparatively large standard deviation of the repeatability was mainly due to one outlier of each plane in cups 1 and 6 (Table 2). If this was removed, the standard deviation of the repeatability was approximately halved.


40

Discussion We found that it is possible to use CT for wear measurement of acetabular cups. The accuracy was 0.51 mm for linear wear detection. In our previous study on explants (Olivecrona et al. 2005), the accuracy was 0.55 mm. Theoretically, though, we should have clearly improved our method due to the increased number of landmark points and better algorithms. However, compared to the previous study, for estimation of head diameter the accuracy improved from 0.8 mm to 0.2 mm and repeatability improved from 0.3 mm to 0.1 mm, respectively. One explanation for the only small improvement in linear wear estimation is that even though the head diameter estimation was improved, this cup design was more difficult to measure in our image analysis tool due to the metal mesh and the nonspherical outer surface. We also found large deviations in estimation of outer cup diameter, which should be the same before and after exposure to wear. This can be explained by the change in CT acquisition protocol between the “before” and “after” CT scans. This would also be likely to contribute to the lower accuracy and repeatability values. The linear wear data showed 2 outliers: cup no. 1 and cup no. 6. If these were excluded, the accuracy of the proposed method would be improved by 0.01 mm. In vivo wear rate values for conventional acetabular cups range from 0.1 mm/year to 0.2 mm/year, and total wear at revision is about 1–3.5 mm (Kabo et al. 1993). Acetabular cups with wear rates below 0.1 mm/year can have 90% survivorship after 25 years, whereas none of the cups with wear rates above 0.2 mm/year survive that long (Sochart 1999). This would mean that 3 years after implantation, CT can be used to detect wear in conventional acetabular cups, since anticipated wear would be 0.3–0.6 mm, which CT can indeed dete Using digitized AP radiographs combined with dedicated computer software, the accuracy was 1.3 mm in a model study (Schewelov et al. 2004), while a study resembling clinical routine has reported errors (of 2 SD) ranging from –1.8 mm to +1.2 mm for 1 mm of wear and from – 4.4 to +0.8 mm for 4 mm of wear after manual assessment of wear by comparing two AP radiographs (Clarke et al. 1976). A drawback of all 2D methods is the risk that the wear vector might be out of the plane of the radiograph, and that the magnitude of the error depends on the degree of anteversion of the cup (Sychterz et al. 1997). The most accurate method today is considered to be radiostereometric analysis (RSA) (Kärrholm et al. 1997). In one study, the accuracy of the RSA digital measurements was 0.42 mm with a mean measurement error of 0.01 mm (Schewelov et al. 2004). However, such methods are not normally available in clinical practice. Our study was limited by the small number of subjects; therefore, the data should be interpreted with caution. The accuracy of the angles was low due to the difficulty in measurement of the reference plane, since this particular cup design has uneven edges due to the metal mesh. However, in

Acta Orthopaedica 2011; 82 (1): 35–41

a clinical situation this would not be a problem because the direction of wear could be measured relative to the pelvis. It is possible to underestimate the wear if the head of the femoral component is not placed into the most worn part of the polyethylene liner during the CT examination. Also, even with modern software, metal artefacts might still complicate interpretation of the image. It is also common that there is a clearance of about 0–0.5 mm between the head and cup diameters, even though it is not stated by the manufacturer (Lewis et al. 2003). This would indicate a possible initial displacement of the head inside the polyethylene liner. In our study, the clearance was measured to be 0.5 mm. Since we measured both before and after wear and calculated the difference, this clearance was taken into account for in this study. In radiographic techniques including CT, the wear is estimated as the combination of creep and wear. Creep in polyethylene acetabular cups is commonly around 0.1 mm (Isaac et al. 1996), and this is also comparable to our finding with the CMM measurements of the soak controls. In conclusion, we found that CT has the potential for reliable measurement of linear wear of acetabular cups at a clinically relevant level of accuracy. With appropriate development and automation, this method may help in identifying patients with increased risk of aseptic loosening.

AJ: study design, hip simulator, CT scan, data evaluation, data analysis and statistics, literature search, and manuscript preparation. FN: algorithm development and data programming in Matlab and Comsol, data analysis, and manuscript preparation. MEN, GQMJr and MPZ: CT image analysis, 3D software programming, and manuscript preparation. DDG: hip simulator. UWG and AS: supervisors. ICC: hip simulator supervisor. HO: study design, CT scan, CT image analysis, and manuscript preparation.

We are grateful to A. Témun for his assistance with CMM in this work. We thank the Royal Institute of Technology, Sweden, for financial support, and Scholarship Blanceflor Boncompagni-Ludovisi, née Bildt, and the Lorenz Carlsson scholarship for travel expenses between Stockholm and Loma Linda.

No competing interests declared.

Clarke I C, Black K, Rennie C, et al. Can wear in total hip arthroplasties be assessed from radiographs? Clin Orthop 1976; (121): 126-42. Gorniak R J T, Kramer E L, Maguire G Q J, et al. Evaluation of a semi-automatic 3D fusion technique applied to molecular imaging and MRI brain/ frame volume data sets. J Med Syst 2003; 27 (2): 141-56. Isaac G H, Dowson D, Wroblewski B M. An investigation into the origins of time-dependent variation in penetration rates with Charnley acetabular cups-wear, creep or degradation. Proc Inst Mech Eng 1996; 210: 209-16. ISO. Accuracy (trueness and precision) of measurement. International standard ISO 1998; 5725-1:1994 (Switzerland). Jedenmalm A, Noz M E, Olivecrona H, et al. A new approach for assessment of wear in metal-backed acetabular cups using computed tomography: a phantom study with retrievals. Acta Orthop 2008; 79 (2): 218-24.


Acta Orthopaedica 2011; 82 (1): 35–41

Kabo J M, Gebhard J S, Loren G, et al. In vivo wear of polyethylene acetabular components. J Bone Joint Surg (Br) 1993; 75 (2): 254-8. Kärrholm J, Herberts P, Hultmark P, et al. Radiostereometry of Hip Prostheses: Review of Methodology and Clinical Results. Clin Orthop 1997; (344): 94-110. Lewis G, Fencl R M, Caroll M, et al. The relative influence of five variables on the in vitro wear rate of uncrosslinked UHMWPE acetabular liners. Biomaterials 2003; 24: 1925-35. Maguire G Q J, Noz M E, Rusinek H, et al. Graphics applied to image registration. IEEE Computer Graphics Appl 1991; 11: 20-9. Mohanty M. Cellular basis for failure of joint prosthesis. Bio Med Mater Eng 1996; 6: 165-72. Noz M E, Maguire G Q J, Zeleznik M P, et al. A versatile functional-anatomic image fusion method for volume data sets. J Med Syst 2001; 25: 297-307. Olivecrona L, Crafoord J, Olivecrona H, et al. Acetabular component migration in total hip arthroplasty using CT and a semi-automated program for volume merging. Acta Radiologica 2002; 43: 517-27. Olivecrona H, Olivecrona L, Weidenhielm L, et al. Stability of acetabular axis after total hip arthroplasty. Repeatability using CT and a semiautomated program for volume fusion. Acta Radiol 2003a; 44: 653-61. Olivecrona H, Weidenhielm L, Olivecrona L, et al. Spatial component position in total hip arthroplasty. Accuracy and repeatability with a new CT method. Acta Radiol 2003b; 44: 84-91. Olivecrona L, Olivecrona H, Weidenhielm L, et al. Model studies on acetabular component migration in total hip arthroplasty using CT and a semiautomated program for volume merging. Acta Radiologica 2003c; 44: 419-29.

41

Olivecrona H, Weidenhielm L, Olivecrona L, et al. A new CT method to measure cup orientation after Total Hip Arthroplasty. A study on 10 patients. Acta Orthop Scand 2004; 75: 253-60. Olivecrona L, Jedenmalm A, Aspelin P, et al. Assessing wear of the Acetabular cup using computed tomography: an ex vivo study. Acta Radiologica 2005; 46: 852-7. Ranstam J, Ryd L, Onsten I. Accurate accuracy assessment. Review of basic principles. Acta Orthop Scand 2000; 71 (1): 106-8. Santavirta S, Anttila A, Aspenberg P, et al. Tekonivelen biokompatibiliteettitutkimus. Finn J Ortop Pharmatol 1995; 18: 356. Schewelov T, Sanzén L, Börlin N, et al. Accuracy of radiographic and radiostereometric wear measurement of different hip prostheses: an experimental study. Acta Orthop Scand 2004; 75 (6): 691-700. Sochart D H. Relationship of acetabular wear to osteolysis and loosening in total hip arthroplasty. Clin Orthop 1999; (363): 135-50. Sychterz C J, Engh C A, Shah N, et al. Radiographic evaluation of penetration by the femoral head into the polyethylene liner over time. J Bone Joint Surg (Am) 1997; 79: 1040-6. Wang A, Polineni VK, Essner A, et al. Role of proteins and hyaluronic acid in the lubrication and wear of UHMWPE acetabular cups. In 24th Annual Meeting of the Society for Biomaterials, April 22-26. San Diego, CA, USA; 1998: 218. Wang A, Essner A, Schmidig G. The effects of lubricant composition on in vitro wear testing of polymeric acetabular components. J Biomed Mater Res Part B: Appl Biomat 2004; 68B (1): 45-52.


42

Acta Orthopaedica 2011; 82 (1): 42–49

Idiopathic and secondary osteonecrosis of the femoral head show different thrombophilic changes and normal or higher levels of platelet growth factors Elisabetta Cenni1, Caterina Fotia1, Enis Rustemi2, Kimitachi Yuasa3, Giuseppe Caltavuturo4, Armando Giunti1,2, and Nicola Baldini1,2 1Laboratory for Orthopaedic Pathophysiology and Regenerative Medicine, Rizzoli Orthopaedic Institute, Bologna; 2Department of Human Anatomy and Musculoskeletal Pathophysiology, University of Bologna, Bologna, Italy; 3Department of Orthopaedic Surgery, Mie University Graduate School of Medicine, Tsu City, Mie, Japan; 4Immunohaematology and Transfusion Service, Maggiore Hospital, A.U.S.L. Bologna, Bologna, Italy Correspondence EC: ecenni70@gmail.com Submitted 10-04-01. Accepted 10-11-01

Background and purpose Thrombophilia represents a risk factor both for idiopathic and secondary osteonecrosis (ON). We evaluated whether clotting changes in idiopathic ON were different from corticosteroid-associated ON. As platelet-rich plasma has been proposed as an adjuvant in surgery, we also assessed whether platelet and serum growth factors were similar to those in healthy subjects. Methods 18 patients with idiopathic ON and 18 with corticosteroid-associated ON were compared with 44 controls for acquired and inherited thrombophilia. Platelet factor 4 (PF4), transforming growth factor-β1, platelet-derived growth factor-BB (PDGFBB), and vascular endothelial growth factor were assayed in the supernatants of thrombin-activated platelets, in platelet lysates, and in serum from 14 ON patients and 10 controls. Results Idiopathic ON patients had higher plasminogen levels (median 118%) than controls (101%) (p = 0.02). Those with corticosteroid-associated ON had significantly higher D-dimer (333 ng/mL) and lower protein C levels (129%) than controls (164 ng/mL, p = 0.004; 160%, p = 0.02). The frequency of inherited thrombophilia was not different from the controls. No statistically significant differences were found between idiopathic and corticosteroid-associated ON. 20 of the 36 ON patients were smokers. (The controls were selected from smokers because nicotine favors hypercoagulability). ON patients had significantly higher serum PF4 levels (7,383 IU/mL) and PDGF-BB levels (3.1 ng/mL) than controls (4,697 IU/mL, p = 0.005; 2.2 ng/mL, p = 0.02). Interpretation Acquired hypercoagulability was common in both ON types, but the specific changes varied. The release of GF from platelets was not affected, providing a biological basis for platelet-rich plasma being used as an adjuvant in surgical treatment. 

Osteonecrosis (ON) of the femoral head may lead to joint collapse and arthritis. In the late stages of the disease, when pain, stiffness, and disability cannot be controlled by non-surgical means, the treatment is total hip replacement. At the early stages, when less invasive treatments to preserve the femoral head such as core decompression, osteotomy, or vascularized bone grafting—or even non-invasive therapy, such as pharmacological measures, electrical stimulation, shock waves, and electromagnetic fields (Mont et al. 2007)—can prevent the affected bone from collapsing, the disease can be diagnosed only with magnetic resonance imaging. A history of one or more risk factors strengthen the suspicion of ON. It is commonly postulated that ON is caused by reduced blood flow to the bone (Assouline-Dayan et al. 2002). Vascular occlusion and ischemia may depend on different underlying conditions such as trauma, radiation, chemotherapy, caisson disease, alcoholism, or high-dose corticosteroid therapy. However, predisposing conditions are not always evident; about 20% of cases appear to be idiopathic in origin without any associated cause (Min et al. 2008). It has been hypothesized that inherited or acquired thrombophilia and hypofibrinolysis may be risk factors for idiophatic ON (Glueck et al. 2003). However, patients with ON seldom relate having a familial or personal history of thromboembolism and the results of the specific alterations in the clotting system or of fibrinolysis are contrasting (Mehsen et al. 2009). Moreover, research has been focused on idiopathic ON without considering that thrombophilia also represents an additional risk factor for secondary ON. On the basis of the high concentration of growth factors (GFs) released by a-granules, platelet-rich plasma (PRP) has been proposed as an adjuvant to improve angiogenesis and osteogenesis in the surgical treatment of ON (Yokota et

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2011.555368


Acta Orthopaedica 2011; 82 (1): 42–49

43

Table 1. Profiles of 18 patients with idiopathic ON, 18 patients with corticosteroid-associated ON, and 44 healthy controls. For qualitative variables, the number of patients is reported Variable Sex male female Age (years) mean (SD) median (range) Bilaterality Smoking habit no. of smokers no. of non-smokers no. of unknown

Idiopathic ON

Corticosteroid- associated ON

Healthy controls

15 3

12 6

33 11

45 (12) 41 (27–74) 9

41 (10) 42 (25–64) 9

45 (10) 44 (22–64) –

13 3 2

7 6 5

42 1 1

al. 2008, Marx 2009). However, nothing is known about the levels of platelet GFs in ON. We evaluated whether patients affected by idiopathic ON had different plasmatic and genetic abnormalities of thrombophilia and fibrinolysis from patients with corticosteroidassociated ON. We also assessed whether the platelet levels of osteogenic and angiogenic GFs were similar to those of healthy subjects.

Patients and methods The study protocol was approved by the institutional ethical committee on human research and was performed according to the Helsinki Declaration of 1975, as revised in 2000. The study included patients with idiopathic and corticosteroidassociated ON of the femoral head who had been diagnosed by clinical and radiological criteria and were hospitalized for hip surgery. Before surgery, an informed consent document was signed, the familial and clinical history was registered, and a sample of blood was taken. The patients had not taken heparin, oral anticoagulants, aspirin, or other platelet antiaggregants in the 2 weeks preceding blood collection. Patients with platelet disorders or platelet number less than 1 × 105/µL were excluded. 18 patients had idiopathic ON and in 18 individuals the disease developed after corticosteroid treatment. Contrary to other studies, we did not find any case of alcohol-related ON. The consumption of alcohol was occasional or limited to half or one glass of wine per meal. There were no statistically significant differences in age and sex between idiopathic and corticosteroid-associated ON. The patients were compared with 44 healthy individuals who were similar to the patients regarding age and sex, and who had no clinical disorder or history of thrombosis (Table 1). A familial history of thrombosis was seldom present: there were 3 such cases with idiopathic ON and 1 such case with corticosteroid-associated

Table 2. Stages of osteonecrosis (ON) in 18 patients with idiopathic ON and 18 patients with corticosteroid-associated ON, according to the radiographic criteria of the Japanese Specific Disease Investigation Committee under the auspices of the Japanese Ministry of Health, Labour and Welfare (Sugano et al. 2002). The number of patients for each stage is reported. In 1 patient with idiopathic ON, the radiographs did not show any lesions and the clinical diagnosis was confirmed by magnetic resonance imaging

1

2

Idiopathic ON Corticosteroid-associated ON

0 0

2 2

Stage 3A 7 6

3B

4

5 5

3 5

ON. None of the individuals had any history of deep venous thrombosis, embolism, or myocardial infarction. Risk factors for thrombosis were seldom present: arterial hypertension (3 patients with idiopathic ON and 3 patients with corticosteroidassociated ON) and obesity (Lijnen 2009) (1 case with idiopathic ON and 2 patients with corticosteroid-associated ON). As smoking was common in patients with either idiopathic or corticosteroid-associated ON, the healthy controls were selected from a population of smokers to avoid bias based on hypercoagulability (that is favored by nicotine (Benowitz et al. 1993)) and to make a better comparison with ON patients. Staging was performed according to the radiographic criteria of the working group of the Japanese Specific Disease Investigation Committee under the auspices of the Japanese Ministry of Health, Labour and Welfare, to establish criteria for diagnosis and management of idiopathic ON of the femoral head (Sugano et al. 2002) (Table 2). There were no statistically significant differences in stage between the ON types, as analyzed with the Chi-square test. In a group of 14 ON patients (mean age 35 (19–64) years, 13 males) selected on the basis of surgery, platelet factor 4 (PF4) and GF levels in platelets and serum were compared with those in 10 healthy controls. 12 patients were affected by idiopathic ON and 2 patients by secondary ON (1 by posttraumatic ON and 1 by corticosteroid-associated ON). Age and sex were similar in the controls and patients. Blood collection and handling For the clotting tests and for PF4 and GF assays, the blood was collected in 0.129 M sodium citrate. Within 2 h of collection, the samples for the clotting tests were centrifuged at 600 g for 10 min. The plasma was frozen in aliquots at –70°C. The samples for the assays of PF4 and GF were centrifuged at 180 g for 5 min in order to separate PRP from erythrocytes and leukocytes. Then the PRP was transferred to a clean tube and centrifuged at 600 g for 15 min. The platelets were pelleted at the bottom of the tube; platelet-poor plasma (PPP) in the upper phase was carefully transferred to a clean tube. Platelet number was adjusted to 1 × 106/µL with PPP. Then the PRP was divided into 2 aliquots. In the first aliquot, platelets were


44

Acta Orthopaedica 2011; 82 (1): 42–49

lysed by 2 cycles of freeze-thawing at –80°C. In the second aliquot, the platelet release reaction was induced by adding 5 U/mL bovine thrombin (Sigma-Aldrich Corp., St. Louis, MI) for 30 min at room temperature, the samples were centrifuged, and the supernatants (from platelet release) were stored at –80°C prior to assay. PF4 and GF were also determined in serum, which was stored at –80°C after centrifugation at 600 g for 45 min. Blood for genetic analysis was drawn in tubes containing ethylenediaminetetraacetic acid and stored at –20°C.

PDGF-BB: R&D Systems Inc., Minneapolis, MN; DuoSet Human VEGF: R&D Systems). Before assay of TGF-β1, the samples were activated with 1N HCl to pH 3.0 or lower for 15 min and then neutralized with 1N NaOH to pH 7.6.

Antithrombin III Antithrombin III was determined with a chromogenic assay (HemosIL Antithrombin, Instrumentation Laboratory, Lexington, MA). The percentage of antithrombin III activity in the samples was extrapolated from a standard curve of scalar dilutions of Calibration Plasma (Instrumentation Laboratory) according to the linear regression function.

Factor V Leiden mutation The single point mutation (G to A at position 1691) of the factor V gene (factor V Leiden) was detected using real-time PCR (Factor V Leiden Kit; Roche Applied Science, Monza, Italy) on the LightCycler 1.2 instrument (Roche Applied Science). The amplification program consisted of a first denaturation at 95°C for 30 seconds and 45 cycles of amplification (annealing at 55°C for 10 seconds and extension at 72°C for 5 seconds).

Protein C Protein C activity was determined with a clotting test (HemosIL ProClot APTT-SP, Instrumentation Laboratory) based on the prolongation of APTT in the presence of activated protein C, generated by an activator derived from snake venom. The percentage of protein C activity was extrapolated by a standard curve of scalar dilutions of Calibration Plasma according to the linear regression function. D-dimer D-dimer was measured by enzyme immunoassay (Asserachrom D-DI; Diagnostica Stago, Asnieres, France). The standard curve was plotted using the point-to-point method. Plasminogen Plasminogen was determined with a chromogenic assay (HemosIL Plasminogen, Instrumentation Laboratory). The percentage of activity of plasminogen in the samples was extrapolated from a standard curve of scalar dilutions of Calibration Plasma according to the linear regression function. Plasminogen activator inhibitor-1 PAI-1 was detected in citrated plasma by enzyme immunoassay (AssayMax Human Plasminogen Activator Inhibitor-1 (PAI-1) ELISA Kit; Assaypro, St. Charles, MO). The standard curve was plotted by regression analysis using log-log curve fit. Platelet factor 4 and growth factor assays PF4, transforming growth factor-β1 (TGF-β1), plateletderived growth factor-BB (PDGF-BB), and vascular endothelial growth factor-A (VEGF-A) were determined in platelet releasates, lysates, and in serum with enzyme immunoassays according to the directions of the manufacturers (Asserachrom PF4: Stago, Asnieres, France; TGF-β1 Emax ImmunoAssay System: Promega Corp., Madison, WI; Quantikine Human

DNA purification Total DNA was purified from whole blood (QIAamp DNA Mini Kit; Qiagen GmbH, Hilden, Germany) and stored at –20°C.

Prothrombin gene mutation The prothrombin gene mutation 20210A was detected using real-time PCR (Factor II (Prothrombin) G20210A Kit; Roche Applied Science) on the LightCycler 1.2 instrument. The amplification program consisted of a first denaturation at 95°C for 30 seconds and 45 cycles of amplification (annealing at 55°C for 10 seconds and extension at 72°C for 5 seconds). PAI-1 gene 4G/5G polymorphism The PAI-1 gene 4G/5G polymorphism was detected using real-time PCR (huPlasminogen activator inhibitor 4/5G: TibMolbiol, Genoa, Italy; LC DNA Master Hybridization Probes: Roche Applied Science) on the LightCycler 1.2 instrument. Five μL of the DNA template was added to a master mix composed of 2.2 μL 4 mM MgCl2, 2 μL LC DNA Master Hybridization Probes and 1 μL of each of the primers huPAI 4/5G, in a total volume of 20 μL. The amplification program consisted of a denaturation step at 95°C for 60 seconds followed by 37 cycles of amplification (denaturation at 95°C for 1 seconds, annealing at 64°C for 10 seconds, and extension at 72°C for 13 seconds). Statistics Statistical evaluations were performed using StatView for Windows version 5.0.1 (SAS Institute, Cary, NC). The statistical power was calculated with Simple Interactive Statistical Analysis (SISA) (http://www.quantitativeskills.com/sisa/). Quantitative results are reported as arithmetic mean, standard deviation (SD), median, and minimum-maximum range. Because normal distribution and homogeneity of variance were not verified by Levene’s test, the differences among the groups were evaluated with Kruskal-Wallis test, and the differences between ON groups, or each ON group and the negative control, were evaluated with the Mann-Whitney U-test.


Acta Orthopaedica 2011; 82 (1): 42–49

45

Table 3. Protein C and antithrombin III activity in 18 patients with idiopathic ON, 18 patients with corticosteroid-associated ON, and 44 healthy controls Idiopathic ON mean (SD) median minimum–maximum range Corticosteroid-associated ON mean (SD) median minimum–maximum range Healthy controls mean (SD) median minimum-–maximum range a

Protein C (%)

Antithrombin III (%)

174 (50) 163 95–273

110 (22) 115 72–145

138 (42) a 129 85–243

103 (17) 107 68–125

167 (35) 160 99–237

102 (18) 98 80–146

p = 0.02 when compared with healthy controls.

The differences between the concentrations of GFs in serum and in platelet releasates or lysates were evaluated with the Wilcoxon signed rank test. The differences were considered significant at p-values < 0.05.

Results Coagulation inhibitors In corticosteroid-associated ON, protein C activity was statistically significantly lower; however, in the idiopathic type the results were similar to the healthy controls. There were no statistically significant differences between idiopathic and corticosteroid-associated ON. Only non-relevant differences were demonstrated for antithrombin III levels, either between the 2 groups of patients or between each group and the negative control (Table 3). Hypercoagulability The patients with corticosteroid-associated ON showed statistically significantly higher D-dimer concentrations than the healthy controls. In idiopathic ON, a trend towards high levels of D-dimer was observed but this was not statistically significant. No statistically significant differences were found between idiopathic and corticosteroid-associated ON (Table 4). Fibrinolysis The patients with idiopathic ON had statistically significantly higher plasminogen activity than the controls. The patients with corticosteroid-associated ON had plasminogen concentrations similar to the healthy controls. No statistically significant differences were found between the 2 types of ON. No significant differences in plasma concentration of PAI-1 were seen, either between the patients and the controls or when comparing the different types of ON (Table 4).

Table 4. Plasma levels of D-dimer, plasminogen, and PAI-1 in 18 patients with idiopathic ON, 18 patients with corticosteroid-associated ON, and 44 healthy controls

D-dimer (ng/mL)

Plasminogen PAI-1 (%) (ng/mL)

Idiopathic ON mean (SD) 241 (132) median 173 range 86–553 Corticosteroid-associated ON mean (SD) 384 (274) a median 333 range 85–1000 Healthy controls mean (SD) 173 (72) median 164 range 65–338 a b

115 (18) b 118 72–149

4.6 (2.6) 4.7 1.2–10.5

104 (15) 101 84–129

3.6 (1.9) 2.9 1.5–7.1

104 (16) 101 65–129

3.4 (1.8) 3.2 0.6–8

p = 0.004 vs. healthy controls p = 0.02 vs. healthy controls

Table 5. Coagulation and fibrinolysis in smokers (13 with idiopathic ON, 7 with corticosteroid-associated ON, and 42 healthy controls) Protein C (%) mean (SD) median range Antithrombin III (%) mean (SD) median range D-dimer (ng/mL) mean (SD) median range Plasminogen (%) mean (SD) median range PAI-1 (ng/mL) mean (SD) median range a b

Idiopathic ON

Corticosteroid- associated ON

Healthy controls

178 (57) 163 95–273

152 (52) 148 97–243

165 (39) 158 99–237

107 (22) 107 72–134

104 (19) 112.7 68–117

103 (19) 99 80–146

227 (118) a 173 145–553

239 (131) 235 85–419

172 (73) 158 65–338

116 (20) 118 72–149

108 (18) 105 84–129

104 (16) 102 65–129

5.3 (2.5) b 4.8 1.8–10.5

3.8 (2.3) 2.6 2–7

3.4 (1.8) 3.2 0.6–8

p = 0.01 when compared with healthy controls; p = 0.05 when compared with healthy controls.

Hypercoagulability and fibrinolysis in smokers In smokers, D-dimer levels were significantly different in idiopathic ON, in corticosteroid-associated ON, and in healthy controls (Kruskall-Wallis test: p = 0.02), and they were increased in ON (Mann-Whitney U test: p = 0.006). In particular, smokers with idiopathic ON showed statistically significantly higher D-dimer levels than controls. In these patients, a trend was also evident regarding higher plasma PAI-1 levels than controls (Table 5). No statistically significant differences


46

Acta Orthopaedica 2011; 82 (1): 42–49

Table 6. Platelet factor 4 (PF4) and growth factors in releasates, platelet lysates, platelet-poor plasma (PPP), and serum in 14 patients with osteonecrosis (ON) and in 10 controls (Ctr). Releasates and lysates were obtained from platelet concentrates, after adjustment of platelet number to 1 × 106 / mL Sample Releasates mean (SD) median range Lysates mean (SD) median range Serum mean (SD) median range a

ON

PF4 (IU/mL) Control

TGF-β1 (ng/mL) ON Control

7,661 (2,648) 8,265 4,568–10,000

6,856 (2,171) 6,939 3,269–10,000

55 (76) 21 2–200

28 (35) 16 2–105

8,621 (3,231) 10,000 2,031–10,000

9,042 (1,670) 10,000 4,717–10,000

146 (85) 200 33–200

99 (68) 69 26–196

7,405 (985) a 7,383 5,777–8,553

5,142 (1240) 4,697 3,576–7,165

38 (18) 50 14.6–50

26 (15) 29 2.5–50

p = 0.005 when compared with controls;

b

PDGF-BB (ng/mL) ON Control 8.1 (3.5) 8.1 2.9–13.7

6.4 (2.6) 6.5 2.5–11.1

VEGF-A (pg/mL) ON Control 165 (83) 183 9–235

105 (88) 105 9–230

11.9 (7.4) 10.8 (6.3) 11.1 9.2 2.7–26.8 4–26

206 (180) 152 9–570

172 (134) 138 9–402

4.2 (3.6) b 2.3 (0.5) 3.1 2.2 1.1–13.5 1.3–3.1

215 (210) 155 9–628

182 (134) 160 9–493

p = 0.02 when compared with controls.

in all parameters examined were found between smokers with corticosteroid-associated ON and those with idiopathic ON or controls. Platelet factor 4 and growth factor assays A high degree of variability was seen in platelet and serum levels of PF4 and GFs, both in patients and in healthy controls. The levels of GFs in releasates were higher than in serum; in particular, PDGF-BB was 2.7- (2.1-) fold higher, TGF-β1 was 1.7- (2.3-) fold higher and VEGF was 1.5- (0.3-) fold higher. The lysate:serum ratio was higher than the releasate:serum ratio: for PDGF-BB it was 3.4- (1.8-) fold higher, for TGF-β1 it was 4.8- (4.5-) fold higher and for VEGF it was 1.3- (0.5-) fold higher. Even though the GF levels in releasates and lysates from ON patients were higher than those from healthy controls, no statistically significant differences were found. Instead, the serum concentrations of PF4 and PDGF-BB in ON patients were statistically significantly higher than in the controls (Table 6). Heritable thrombophilia and hypofibrinolysis Homozygosity or heterozygosity for factor V Leiden was present in only 2 patients affected by corticosteroid-associated ON. Factor V Leiden was not detected in any patients with idiopathic ON or in any controls. Homozygosity for the prothrombin gene mutation was not found in any patients or controls; heterozygosity was present in only 2 patients with idiopathic ON and in 1 control. The genotypes 4G/5G and 4G/4G of the PAI-1 gene were demonstrated in 7 and 6 of 18 patients with idiopathic ON, and in 9 and 4 of 18 patients with corticosteroid-associated ON. The frequency of 4G/4G homozygosity was also high in healthy controls (19/44), while 16 of the 44 control subjects were heterozygous.

Discussion We hypothesized that the alterations in the clotting system or in fibrinolysis in idiopathic ON may be different from those in corticosteroid-associated ON. The secondary hypothesis was that the levels of GFs in ON platelets are similar to or higher than in those of healthy subjects, thus supporting the use of PRP as an adjuvant for the treatment of ON. Although the idiopathic and corticosteroid-associated ON groups were homogenous regarding age, sex, and stage, one limitation of the study may be the low number of cases. The 2 types of ON showed different alterations in the clotting tests. In particular, higher levels of plasminogen were detected in idiopathic ON while lower protein C activity and higher concentrations of D-dimer were found in corticosteroid-associated ON. The high plasminogen levels in idiopathic ON may have several explanations. Some degree of synovial inflammation has been shown in ON and it was hypothesized to play a role in the disease progression (Rabquer et al. 2009). Plasminogen, which increases in phlogosis, could be a systemic marker of the inflammatory state present in ON, such as erythrosedimentation rate, which has been found to be elevated in 78% of ON patients (Séguin et al. 2008). Also, the trend toward higher levels of PAI-1—which was particularly evident in smokers with idiopathic ON—may have contributed to the increase in plasminogen. In fact, PAI-1 inhibits tissue plasminogen activator and therefore generation of plasmin from plasminogen. This trend of increased PAI-1 levels was only partially explained by homozygosity for the 4G/4G polymorphism of the PAI-1 gene, which leads to higher transcription of the gene and therefore higher production of PAI-1 protein (Asano et al. 2004). Notably, in our patients the frequency of homozygosity for the 4G/4G polymorphism was lower than in the healthy controls, and was also lower than the 41% reported by Glueck


Acta Orthopaedica 2011; 82 (1): 42–49

et al. (1999) for ON. Inflammation (Aso 2007, Kruithof 2008) or endothelial activation (Salgado et al. 1994) may be another possible explanation for the increase in PAI-1. Endothelial activation, which has been reported in ON (Séguin et al. 2008), could have caused the increase in PAI-1 in our patients. One possible limitation of our data was the low sample number. The statistical power analysis based on idiopathic ON-control differences in plasminogen revealed a power of 62% for α = 0.05. Thus, these results should be considered preliminary and they require confirmation. Normal or low levels of plasminogen have been reported by other authors in both ON and Perthes’ disease, which has a similar pathogenesis. Lee et al. (2003) did not show any variation between non-traumatic ON patients and controls. In Perthes’ disease, reduced plasminogen was reported by Pósán et al. (2003), but normal levels were reported by Koo et al. (2002). These discordant results may depend on ethnic differences, as hypothesized by Lee et al. (2003). On the other hand, contrasting data have been reported in the literature for other clotting or fibrinolytic alterations in ON, and in Perthes’ disease. Probably different thrombophilic changes, but all drivers to hypercoagulability, may be involved in ON pathogenesis among different ethnic groups or even among different individuals. For all the parameters examined, no statistically significant differences between idiopathic and corticosteroid-associated ON were found. Corticosteroid-associated ON showed peculiar alterations, consisting of a decrease in protein C and in an increase in D-dimer relative to healthy controls. A possible limitation of the finding of a reduction in protein C in corticosteroid-associated ON was the fact that the statistical power was 72% for α = 0.05; thus, our findings require to be confirmed by further studies. Other authors have found reduced protein C activity in non-traumatic ON (Zalavras et al. 2000). In Perthes’ disease, discordant results have been reported for protein C: some authors have not found any changes (Almeida Matos 2008) while others have found reduced levels (Mehta et al. 2006). Instead, the statistical power analysis based on difference between corticosteroid-associated ON and controls regarding D-dimer levels revealed a power of 95% for α = 0.01. An increase in D-dimer was reported in 25% and 21% of patients with idiopathic or secondary ON, respectively, as a consequence of the hypercoagulability that also induced ON (Pósán et al. 2003). Corticosteroid treatment did not affect D-dimer levels in healthy volunteers (Brotman et al. 2006), but induced thrombosis and ON in rabbits, when associated with a low dose of lipopolysaccharide (Wu et al. 2008). Even though glucocorticoids may increase PAI-1 gene transcription (Halleux et al. 1999) and plasma levels (Kerachian et al. 2009) in corticosteroid-associated ON, PAI-1 concentration was similar to that in the healthy controls, probably because the patients were not taking corticosteroids when they were enrolled. We found a low frequency of inherited thrombophilia both

47

in idiopathic ON and corticosteroid-associated ON. The frequencies of factor V Leiden and prothrombin gene mutations were similar to those in the controls, contrary to what has been asserted by some authors who have hypothesized that idiopathic ON might be favored by factor V Leiden or prothrombin gene mutation (Björkman et al. 2004). Interestingly, 55% of the ON patients were smokers, as compared to a rate of 23% in the general population (Activities for the prevention of smoking—2009 report, Ministry of Health, Italy, http://www.salute.gov.it/imgs/C_17_pubblicazioni_1161_ulteriorallegiati_ulterioreallegato_0_alleg.pdf). Other researchers have found a strong association between tobacco consumption and risk of ON. Mehsen et al. (2009) found a frequency of 69% smokers in ON patients and a frequency of 38% smokers in controls. Sakaguchi et al. (2010) observed that more patients who developed ON after corticosteroid therapy had a smoking habit (68%) than controls (50%). Also, in Perthes’ disease an association between the femoral lesion and smoking by the mother during pregnancy or passive smoking after birth has been demonstrated (Glueck et al. 1998). Nicotine has been reported to cause vasoconstriction (Winniford et al. 1987) and to induce bovine and human endothelial cells to produce PAI-1 (Kuo et al. 1989, Zidovetzki et al. 1999). Increased plasma levels of D-dimer (Wannamethee et al. 2005) and PAI-1 (Tapson 2005) have been found in healthy smokers. In our study, D-dimer levels were higher in smokers with ON than in smokers in the control group. This result was more evident in smokers with idiopathic ON. In these patients, a trend of higher plasma levels of PAI-1 was also found. Thus, we could hypothesize that smoking is only one of several risk factors for idiopathic ON. Other causes may induce hypercoagulability, which is important for ON pathogenesis. In smokers with corticosteroid-associated ON, D-dimer levels were not statistically significantly different from those in the controls, which is contrary to what was observed when we considered smokers and non-smokers together. In this group, other causes, independent of smoking and present also in non-smokers, could have induced the increase in D-dimer. Mehsen et al. (2009) did not consider the laboratory tests of ON patients and controls separately according to their smoking habits. Sakaguchi et al. (2010) did not perform any laboratory test because they considered only anamnestic risk factors for corticosteroid-associated ON. On the basis of the above considerations, smoking may be considered to be more of a possible risk factor when associated with other causes than an etiological factor of ON, through its proinflammatory, procoagulant, antifibrinolytic, and antiendothelial effects (Yanbaeva et al. 2007). In order to establish a scientific basis for the use of plateletrich plasma in ON, we determined the levels of some osteogenic and angiogenic GFs in platelet lysates and releasates. The levels of platelet GFs are known in healthy blood donors (Zimmermann et al. 2003), but the effects of bone diseases


48

such as ON on the concentrations are unknown. As GF levels in plasma originate from different cell types, we determined also PF4, which is specific for platelet a-granules (Kaplan 1978). In releasates, PDGF-BB, TGF-β1, and VEGF originated from platelet activation induced by thrombin and their levels were lower than in lysates, which contained the total amount present in platelets. Some of the GFs were entrapped in the fibrin network and were slowly released during clot retraction. The levels of GFs in ON were similar to or higher than in healthy controls. These results support the use of PRP as an adjuvant in ON treatment. The higher concentrations of PF4 and PDGF-BB in the serum of ON patients may be due to the increased platelet activation associated with the thrombophilic state (Pósán et al. 2003). In conclusion, we did not detect statistically significant differences in clotting tests between idiopathic and corticosteroid-associated ON. In comparison with healthy controls, idiopathic ON showed higher plasminogen activity; corticosteroid-associated ON had lower protein C levels and higher D-dimer levels. However, these results require to be confirmed by further studies on a larger number of patients, particularly the data regarding plasminogen and protein C. In both groups of patients, the frequency of heritable thrombophilia and fibrinolysis was similar to that for the controls. Thus, the results of this preliminary study support the idea that both idiopathic and corticosteroid-associated ON are favored by acquired hypercoagulability, but at a subclinical level. The high levels of PF4 in releasates showed that platelet release reaction is not affected in ON. The concentrations of GFs in releasates and lysates from ON patients were similar to those in healthy subjects. This result provides a biological basis for the use of PRP in ON treatment.

EC planned the study, evaluated the results and wrote the paper. CF performed the clotting and genetic tests. ER enrolled the patients. KY evaluated the radiographs. GC enrolled the controls. AG supervised the research. NB evaluated the results and also supervised the research.

The authors thank Dr Lucy Scioscia for help in editing the English. Financial support was received from the Rizzoli Orthopaedic Institute, “Ricerca corrente”, the “Fondazione del Monte di Bologna e Ravenna”, and EmiliaRomagna District “Progetto di Ricerca Regione-Università: Regenerative Medicine in Osteo-articular Diseases”.

No competing interests declared.

Almeida Matos M. The role of protein C deficiency in the etiology of Perthes disease. Ortop Traumatol Rehabil 2008; 10 (3): 274-8.

Acta Orthopaedica 2011; 82 (1): 42–49

Asano T, Takahashi K A, Fujioka M, Inoue S, Ueshima K, Hirata T, Okamoto M, Satomi Y, Nishino H, Tanaka T, Hirota Y, Kubo T. Relationship between postrenal transplant osteonecrosis of the femoral head and gene polymorphisms related to the coagulation and fibrinolytic systems in Japanese subjects. Transplantation 2004; 77 (2): 220-5. Aso Y. Plasminogen activator inhibitor (PAI)-1 in vascular inflammation and thrombosis. Front Biosci 2007; 12: 2957-66. Assouline-Dayan Y, Chang C, Greenspan A, Shoenfeld Y, Gershwin M E. Pathogenesis and natural history of osteonecrosis. Semin Arthritis Rheum 2002; 32 (2): 94-124. Benowitz N L, Fitzgerald G A, Wilson M, Zhang Q. Nicotine effects on eicosanoid formation and hemostatic function—comparison of transdermal nicotine and cigarette-smoking. J Am Coll Cardiol 1993; 22 (4): 1159-67. Björkman A, Svensson P J, Hillarp A, Burtscher I M, Rünow A, Benoni G. Factor V Leiden and prothrombin gene mutation: risk factors for osteonecrosis of the femoral head in adults. Clin Orthop 2004; (425): 168-72. Brotman D J, Girod J P, Posch A, Jani J T, Patel J V, Gupta M, Lip G Y, Reddy S, Kickler T S. Effects of short-term glucocorticoids on hemostatic factors in healthy volunteers. Thromb Res 2006; 118 (2): 247-52. Glueck C J, Freiberg R A, Crawford A, Gruppo R, Roy D, Tracy T, SieveSmith L, Wang P. Secondhand smoke, hypofibrinolysis, and Legg-Perthes disease. Clin Orthop 1998; (352): 159-67. Glueck C J, Fontaine R N, Gruppo R, Stroop D, Sieve-Smith L, Tracy T, Wang P. The plasminogen activator inhibitor-1 gene, hypofibrinolysis, and osteonecrosis. Clin Orthop 1999; (366): 133-46. Glueck C J, Freiberg R A, Wang P. Role of thrombosis in osteonecrosis. Curr Hematol Rep 2003; 2 (5): 417-22. Halleux C M, Declerck P J, Tran S L, Detry R, Brichard S M. Hormonal control of plasminogen activator inhibitor-1 gene expression and production in human adipose tissue: stimulation by glucocorticoids and inhibition by catecholamines. J Clin Endocrinol Metab 1999; 84 (11): 4097-105. Kaplan K L. Proteins secreted by platelets: significance in detecting thrombosis. Adv Exp Med Biol 1978; 102: 105-19. Kerachian M A, Séguin C, Harvey E J. Glucocorticoids in osteonecrosis of the femoral head: a new understanding of the mechanisms of action. J Steroid Biochem Mol Biol 2009; 114 (3-5): 121-8. Koo K H, Song H R, Ha Y C, Kim J R, Kim S J, Kim KI, Chang KC, Ahn I O, Cho S H. Role of thrombotic and fibrinolytic disorders in the etiology of Perthes’ disease. Clin Orthop 2002; (399): 162-7. Kruithof E K. Regulation of plasminogen activator inhibitor type 1 gene expression by inflammatory mediators and statins. Thromb Haemost 2008; 100 (6): 969-75. Kuo B S, Dryjski M, Bjornsson T D. Influence of nicotine and cotinine on the expression of plasminogen activator activity in bovine aortic endothelial cells. Thromb Haemost 1989; 61 (1): 70-6. Lee J S, Koo K H, Ha Y C, Koh K K, Kim S J, Kim J R, Song H R, Cho S H. Role of thrombotic and fibrinolytic disorders in osteonecrosis of the femoral head. Clin Orthop 2003; (417): 270-6. Lijnen H R. Obesity and vascular risk. Hamostaseologie 2009; 29 (1): 44-5. Marx R E. Reconstruction of defects caused by bisphosphonate-induced osteonecrosis of the jaws. J Oral Maxillofac Surg (5 Suppl) 2009; 67: 10719. Mehsen N, Barnetche T, Redonnet-Vernhet I, Guérin V, Bentaberry F, Gonnet-Gracia C, Schaeverbeke T. Coagulopathies frequency in aseptic osteonecrosis patients. Joint Bone Spine 2009; 76 (2): 166-9. Mehta J S, Conybeare M E, Hinves B L, Winter J B. Protein C levels in patients with Legg-Calve-Perthes disease: is it a true deficiency? J Pediatr Orthop 2006; 26 (2): 200-3. Min B W, Song K S, Cho C H, Lee S M, Lee K J. Untreated asymptomatic hips in patients with osteonecrosis of the femoral head. Clin Orthop 2008; (466) (5): 1087-92. Mont M A, Jones L C, Seyler T M, Marulanda G A, Saleh K J, Delanois R E. New treatment approaches for osteonecrosis of the femoral head: an overview. Instr Course Lect 2007; 56: 197-212.


Acta Orthopaedica 2011; 82 (1): 42–49

Pósán E, Szepesi K, Gáspár L, Csernátony Z, Hársfalvi J, Ajzner E, Tóth A, Udvardy M. Thrombotic and fibrinolytic alterations in the aseptic necrosis of femoral head. Blood Coagul Fibrinolysis 2003; 14 (3): 243-8. Rabquer B J, Tan G J, Shaheen P J, Haines G K 3rd, Urquhart A G, Koch A E. Synovial inflammation in patients with osteonecrosis of the femoral head. Clin Transl Sci 2009; 2 (4): 273-8. Sakaguchi M, Tanaka T, Fukushima W, Kubo T, Hirota Y. Impact of oral corticosteroid use for idiopathic osteonecrosis of the femoral head: a nationwide multicenter case-control study in Japan. J Orthop Sci 2010; 15 (2): 185-91. Salgado A, Bóveda J L, Monasterio J, Segura R M, Mourelle M, GómezJiménez J, Peracaula R. Inflammatory mediators and their influence on haemostasis. Haemostasis 1994; 24 (2): 132-8. Séguin C, Kassis J, Busque L, Bestawros A, Theodoropoulos J, Alonso M L, Harvey E J. Non-traumatic necrosis of bone (osteonecrosis) is associated with endothelial cell activation but not thrombophilia. Rheumatology (Oxford) 2008; 47 (8): 1151-5. Sugano N, Atsumi T, Ohzono K, Kubo T, Hotokebuchi T, Takaoka K. The 2001 revised criteria for diagnosis, classification, and staging of idiopathic osteonecrosis of the femoral head. J Orthop Sci 2002; 7 (5): 601-5. Tapson V F. The role of smoking in coagulation and thromboembolism in chronic obstructive pulmonary disease. Proc Am Thorac Soc 2005; 2 (1): 71-7. Wannamethee S G, Lowe G D, Shaper A G, Rumley A, Lennon L, Whincup P H. Associations between cigarette smoking, pipe/cigar smoking, and smoking cessation, and haemostatic and inflammatory markers for cardiovascular disease. Eur Heart J 2005; 26 (17): 1765-73.

49

Winniford M D, Jansen D E, Reynolds G A, Apprill P, Black W H, Hillis L D. Cigarette smoking-induced coronary vasoconstriction in atherosclerotic coronary artery disease and prevention by calcium antagonists and nitroglycerin. Am J Cardiol 1987; 59 (4): 203-7. Wu X, Yang S, Duan D, Zhang Y, Wang J. Experimental osteonecrosis induced by a combination of low-dose lipopolysaccharide and high-dose methylprednisolone in rabbits. Joint Bone Spine 2008; 75 (5): 573-8. Yanbaeva D G, Dentener M A, Creutzberg E C, Wesseling G, Wouters E F. Systemic effects of smoking. Chest 2007; 131 (5): 1557-66. Yokota K, Ishida O, Sunagawa T, Suzuki O, Nakamae A, Ochi M. Platelet-rich plasma accelerated surgical angiogenesis in vascular-implanted necrotic bone: an experimental study in rabbits. Acta Orthop 2008; 79 (1): 106-10. Zalavras C, Dailiana Z, Elisaf M, Bairaktari E, Vlachogiannopoulos P, Katsaraki A, Malizos K N. Potential aetiological factors concerning the development of osteonecrosis of the femoral head. Eur J Clin Invest 2000; 30 (3): 215-21. Zidovetzki R, Chen P, Fisher M, Hofman F M, Faraci F M. Nicotine increases plasminogen activator inhibitor-1 production by human brain endothelial cells via protein kinase C-associated pathway. Stroke 1999; 30 (3): 651-5. Zimmermann R, Arnold D, Strasser E, Ringwald J, Schlegel A, Wiltfang J, Eckstein R. Sample preparation technique and white cell content influence the detectable levels of growth factors in platelet concentrates. Vox Sang 2003; 85 (4): 283-9.


50

Acta Orthopaedica 2011; 82 (1): 50–55

The rate of screw misplacement in segmental pedicle screw fixation in adolescent idiopathic scoliosis The effect of learning and cumulative experience Kasim Abul-Kasim1 and Acke Ohlin2 1Division of Neuroradiology, Lund University, Diagnostic Centre for Imaging and Functional Medicine, 2Department of Orthopaedic Surgery, Lund University, Skåne University Hospital, Malmö, Sweden

Skåne University Hospital;

Correspondence: kasim.abul-kasim@med.lu.se Submitted 10-06-21. Accepted 10-09-17

Background and purpose There are no reports in the literature on the influence of learning on the pedicle screw insertion. We studied the effect of learning on the rate of screw misplacement in patients with adolescent idiopathic scoliosis treated with segmental pedicle screw fixation. Method We retrospectively evaluated low-dose spine computed tomography of 116 consecutive patients (aged 16 (12–24) years, 94 females) who were operated during 4 periods over 2005–2009 (group 1: patients operated autumn 2005–2006; group 2: 2007; group 3: 2008; and group 4: 2009). 5 types of misplacement were recorded: medial cortical perforation, lateral cortical perforation, anterior cortical perforation of the vertebral body, endplate perforation, and perforation of the neural foramen. Reslts 2,201 pedicle screws were evaluated, with an average of 19 screws per patient. The rate of screw misplacement for the whole study was 14%. The rate of lateral and medial cortical perforation was 7% and 5%. There was an inverse correlation between the occurrence of misplacement and the patient number, i.e. the date of operation (r = –0.35; p < 0.001). The skillfulness of screw insertion improved with reduction of the rate of screw misplacement from 20% in 2005–2006 to 11% in 2009, with a breakpoint at the end of the first study period (34 patients). Interpretation We found a substantial learning curve; cumulative experience may have contributed to continued reduction of misplacement rate. 

Since its introduction by Suk et al. (1995), segmental pedicle screw fixation in adolescent idiopathic scoliosis (AIS) has undergone continuous development. The use of pedicle screws in the thoracic spine is not without risk, however, especially for neurovascular compromise. The rate of neurovascular complications associated with pedicle screw misplacement varies from 0% to 1.3% (Liljenqvist et al. 1997, Belmont et

al. 2001 Suk et al. 2001, Coe et al. 2006, Upendra et al. 2008). Evolution of the operative techniques since 1995 (Suk et al. 1995, Kim and Lenke 2005, Kuklo et al. 2005, Lonner et al. 2006), our better understanding of the 3-dimensional nature of adolescent idiopathic scoliosis (AIS), and probably the advances in multidetector computed tomography (CT) technology—with the availability of modern CTs that enable dose reduction—have contributed to increasing use of the “pedicle screw-only construct” in the surgical correction of scoliosis. To our knowledge, there have been no reports in the literature on the learning curve regarding accuracy of screw placement. The rate of misplacement of thoracic pedicle screws, evaluated with CT, varies from about 6% (Di Silvestre et al. 2007) to 50% (Upendra et al. 2008). In all published reports on the accuracy of pedicle screw placement, the figures presented have shown the overall rate of misplacement while possible improvements in the skill of screw insertion over time have not been studied. We assessed whether a learning curve exists regarding the accuracy of screw placement in patients with AIS who are treated with segmental pedicle screw fixation.

Patients and methods Patient data The Suk technique was introduced at the Orthopaedics Department of our hospital in the autumn of 2005, after a study visit to Seoul by the senior author (AO). All 116 patients with AIS who underwent posterior corrective surgery with segmental pedicle screw fixation using titanium “all pedicle screw construct” between autumn 2005 and December 2009 were included in the study. The patients were categorized into 4 groups according to the date of operation as follows; group 1: 34 patients operated on from autumn 2005 through 2006;

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2010.548032


Acta Orthopaedica 2011; 82 (1): 50–55

51

group 2: 21 patients operated on in 2007; group 3: 25 patients operated on in 2008; and group 4: 36 patients operated on in 2009. As the number of patients operated on in the autumn of 2005 was only 7, these were included in the first period with patients operated on in 2006. The mean age of the patients was 16 (12– 24) years and 94 were females. As the primary aim of our study was to evaluate the effect of the learning curve on the rate of accurate screw placement and not the complications of screw misplacement, the study period was limited to the first postoperative clinical follow-up at 8 weeks. The following data were collected from the clinical and the operative records: age, sex, diagnosis, Figure 1. Axial (A–D), coronal (E) and sagittal (F) images obtained by low-dose spine CT from 6 patients. A. Normally-placed screw through the right pedicle of T7. B. Screw with total lateral cortical intraoperative events, postoperative perforation (LCP grade 2) at the level of T9. C. Screw with total medial cortical perforation (MCP events, and the clinical status at 8 grade 2) at the level of L3. D. Pedicle screw with anterior cortical perforation (ACP) with screw tip in weeks. At the latter visit, all patients close proximity to the posterior-lateral limit of aorta (arrow). This screw was removed a few months later. E. Screw tip perforating the upper endplate of L4 (arrow) on the left side (EPP). F. Screw passwere asked whether they had any ing through the lower boundary of the neural foramen (FP) below L3 (arrow). history of neurological symptoms related to medullary or nerve root compromise or evidence of myelopathy. Increased tendon mSv (Abul-Kasim et al. 2009a). The slice collimation of 0.75 reflexes, abnormal clonus, and Babinski sign were specifically mm allowed us to obtain 1-mm and 3-mm thick reformatted sought for. The radiological data gathered from the pre- and axial images with soft-tissue algorithm and skeletal algorithm postoperative standing radiographs and CTs were: Cobb angle respectively, and 2-mm thick coronal and sagittal reformatted before and after surgery, Lenke classification, the number of images. Soft-tissue algorithm was used to allow reduction of screws in the operated constructs, and the positions of screws. streak artifacts from the hardware. The level and the side of every individual screw as well as the relation to the scoliotic concavity and convexity were also Grading of screw misplacement recorded. The relationship of screws to the pedicle, vertebral The classification of the pedicle screw misplacement was body, scoliotic apex, concavity/convexity of the curve, and the performed according to the recently established and reported surrounding structures e.g. aorta, pleura, and the pedicle rib grading system based on whether the violation of pedicular unit (PRU) were recorded. cortex was total or partial rather than on measuring the degree The use of low-dose CT in the work-up of AIS was approved of misplacement in mm (Abul-Kasim et al. 2009b, 2010). 5 by the regional radiation protection committee and the study types of misplacement were recorded, namely medial cortical perforation (MCP), lateral cortical perforation (LCP), antewas approved by the regional ethics committee. rior cortical perforation of vertebral body (ACP), endplate Radiological work-up perforation (EPP), and perforation of neural foramen (FP). Plain radiographs were obtained pre- and postoperatively with The medial and lateral cortical perforations were graded as low-dose radiation; this included posteroanterior and lateral follows: grade 1 when the pedicle screw was partially medialviews. CT examinations were performed with low-radiation- ized or lateralized, and grade 2 when the pedicle screw was dose spine CT at 6–8 weeks after surgery. All examinations totally medialized or lateralized (Figure 1). All examinations were performed on a 16-slice CT scanner (SOMATOM Sensa- were read on 2 different occasions by a senior neuroradiolotion 16; Siemens AG, Forchheim, Germany) with the follow- gist experienced in radiology of spinal deformity. In cases of ing scan parameters: slice collimation 16 × 0.75 mm, rotation different grading on the 2 occasions of evaluation, a consensus time 0.75 s, pitch 1.5, tube voltage 80 kV, and quality refer- was reached by joint evaluation with another neuroradiologist. ence for the effective tube current-time product 25 mAs (Abul- The misplacement rate for every individual patient was calcuKasim et al. 2009). The effective radiation dose of low-dose lated as the number of misplaced screws times 100 divided by CT including 15 vertebral bodies has been reported to be 0.37 the total number of screws inserted.


52

Acta Orthopaedica 2011; 82 (1): 50–55

Operative technique Table 1. Rate of different types of screw misplacement assessed in the study All operations were performed through a standard posterior exposure. The entry points 2005–2009 Period 1 Period 2 Period 3 Period 4 for screws were determined after identificaLCP1 75 20 12 18 25 tion of the bony landmarks. At each assumed LCP2 82 38 13 14 17 entry point, a 3-cm long and 1-mm wide titaLCP, total 157 (7) 58 25 32 42 MCP1 67 24 13 10 20 nium bone marker was inserted. By intermitMCP2 44 17 7 9 11 tent exposure with a C-arm fluoroscope in MCP, total 111 (5) 41 20 19 31 the anteroposterior view, the position and the ACP 11 4 4 1 2 EPP 11 6 0 3 2 degree of tilt of the bone marker were estiFP 11 5 1 2 3 mated. The screw canal was prepared with Combination of different types a 14 4 2 4 4 a hand-driven drill. After drilling, a probe Total rate of misplacement 315 118 52 61 84 Misplacement rate per patient b 14 20 13 12 11 or “feeler” was used to palpate the bottom Total number of screws 2201 586 383 493 739 and borders of the screw canal. Thereafter, Total number of patients 116 34 21 25 36 self-tapping transpedicular screws were Screws per patient 19 17.2 18.2 19.7 20.5 sequentially introduced. Curve correction a Combination of different types: includes combinations of LCP and 1 of the other types, was performed by simple rod derotation as namely ACP, EPP, or FP. Figures in parentheses represent percentage. b Misplacement rate per patient (%) equals the number of misplaced screws x 100 well as direct vertebral rotation (DVR) when divided by the total number of screws inserted. having only the concave rod in place (Lee et al. 2004). The posterior elements were decorticated, and finally the contoured stabilizing rod was inserted. All operations were performed under surgery was 56° (SD 10) and 17° (SD 8), which means an spinal cord monitoring by motor-evoked potential (MEP). The angle reduction by 38° (SD 9) (69%) (p < 0.001). implants used were made of titanium alloy (EXPEDIUM Spine System). Screws were of uniplanar type with a diameter of 4–6 Rate of screw misplacement mm, depending on the pedicular width. We evaluated 2,201 pedicle screws (116 patients with an average of 19 screws per patient). The mean rate of screw Statistics misplacement for the whole study population operated on Estimates are presented as mean (SD). Sampling uncertainty of between the autumn of 2005 and 2009 was 14 (median 13). estimates is presented as the 95% confidence interval (CI). Pos- The rate of misplacement was zero in 16 patients (14%). The sible correlation between the operation date (i.e. the number of rate of lateral and medial cortical perforation was 7% and 5% the patient) and the misplacement rate was also tested. Student (Table 1). 1 patient reported postoperative pain and paresthet-tests were performed to find out whether there were break- sia in a specific dermatome distribution (T8–T10 on the right points in the learning curve that defined the maximal learning side). Low-dose CT showed no medial cortical perforation by comparing the rate of misplacement: (a) in patients oper- or foraminal perforation that could explain the neurological ated on during the first period with that in patients operated deficit. 11 of 18 screw tips that ended in the vicinity of the on during the remaining study periods (2007–2009), and (b) aorta (1–4 mm) were inserted in 2005–2006 (p < 0.001). The in patients operated on during the first 2 periods (2005–2007) number of laterally placed screws that passed through the pedwith that in patients operated on during the last 2 periods icle rib unit (using the in-out-in screw technique) increased (2008–2009). Chi-square test was performed to test the asso- successively from 3 in 2005–2006 to 9 in 2009 (p < 0.001). No ciation of the occurrence of screw misplacement with differ- single pedicular fracture related to the correction maneuver ent categorical variables (sex, side, level of scoliotic apex, and was observed. scoliotic convexity/convexity). Differences with p-values of < 2 patients were reoperated for removal of misplaced pedicle 0.05 were considered statistically significant. We used SPSS screws: 1 with ACP with the screw tip in close proximity to software version 17 for all statistical analyses. the aorta to prevent direct vascular injury or development of pseudoaneurysm, and 1 with MCP and 6–7 mm of spinal canal encroachment.

Results Characteristics of curve The distribution of different types of scoliotic curves according to the Lenke classification was as follows (n/%): type 1 (60/52), type 2 (7/6), type 3 (19/16), type 4 (5/4), type 5 (15/13), and type 6 (10/9). The Cobb angle before and after

Rate of screw misplacement during different study periods We observed a statistically significant inverse correlation between the occurrence of misplacement and the patient number i.e. the date of operation (correlation coefficient of –0.35; p < 0.001) (Figure 2A). The skillfulness of screw inser-


Acta Orthopaedica 2011; 82 (1): 50–55

53

cortical perforation in the first period and the 3 remaining study periods was 2.8% (p = 0.05) (Table 2). Corresponding values for differences in the misplacement rate in patients who were operated in the first 2 periods and the second 2 periods were 2.2% (p = 0.08) for lateral cortical perforation and 2.4% (p = 0.07) for medial cortical perforation. Association between screw misplacement and different study variables No statistically significant correlation was found between the patient’s age and the rate of misplacement (correlation coefficient –0.10; p = 0.3). No statistically significant association was found between the occurrence of screw misplacement and sex, concavity of the scoliotic curvature, or the side of the apex of Figure 2. A. Scatter plot showing the results of the linear regression: patient number (date of operathe major curve. However, the protion) against the misplacement rate (R2 = 0.12). B–D. Line graphs showing the misplacement rate for different study periods: B. Misplacement rate for all types of misplacement. C. Rate of lateral corportion of patients with no screw tical perforation (all LCP, LCP grade 1, and LCP grade 2). D. Rate of medial cortical perforation (all misplacement was higher in males MCP, MCP grade 1, and MCP grade 2). Periods 1–4 indicate patients operated during 2005–2006, (6/22 males as opposed to only 10/94 2007, 2008, and 2009, respectively. females) (Table 3). The rate of screw misplacement in thoracic pedicles was 14% and it was 10% in lumbar pedicles (p = 0.001). The Table 2. Rate of misplacement per patient: Student’s t-test to find out if there was a breakpoint at the end of the first period, i.e. at the rate of misplacement was 20% in the upper thoracic levels end of 2006 compared with the period 2007–2009 (T1–T6) and 13% in the lower thoracic levels (T7–­T12) (p < 0.001) (Table 3). Type of Period Period Mean misplacement 1 2–4 difference Total LCP MCP

20 9.6 7.1

12 6.2 4.3

8 3.4 2.8

95% CI

p-value

3.9–12 < 0.001 0.7–6.1 0.01 –0.0001–5.6 0.05

tion improved significantly, with reduction of the rate of screw misplacement from 20% in 2005–2006 to 11% in 2009. The same applied to the rate of lateral and medial cortical perforation (Table 1 and Figure 2). The rate of medial cortical perforation increased slightly, however, from 3.9% in 2008 to 4.2% in 2009, which mainly depended on increase in grade-1 medial cortical perforation. We found a difference between the rate of misplacement per patient operated on during the first period of the study (20%) and the rate of misplacement per patient operated on during the remaining 3 periods (12%) (p < 0.001). This also applied to lateral cortical perforation with a mean difference of 3.4% (p = 0.01). The mean difference between the rate of medial

Association between screw misplacement and type of curve Although patients with a Lenke type-1 curve showed a higher rate of screw misplacement (16%), no statistically significant correlation was found between the Lenke type and the rate of screw misplacement (p = 0.6). The association between the Lenke type on one hand and the LCP and MCP on the other was not statistically significant (p = 0.9 and p = 0.08). However, the rate of MCP was twice as common in Lenke 1 than in Lenke 3 and 5 (7% vs. 3%) (p = 0.08). The number of fixed levels and screws inserted was highest in patients with Lenke 3, 4, and 6 curves, depending on whether they were double or triple curves (Table 4).

Discussion Although segmental pedicle screw fixation is widely used nowadays to provide anchors for rods in the posterior fixation following correction of the scoliotic deformity in AIS and


54

Acta Orthopaedica 2011; 82 (1): 50–55

Table 3. Chi-square test of association between the occurrence of screw misplacement and different categorical variables Different variables Patients with optimally placed screws misplaced screws Optimally placed screws Misplaced screws Optimally placed screws Misplaced screws Optimally placed screws Misplaced screws Optimally placed screws Misplaced screws Optimally placed screws Misplaced screws

p-value Male 6 16 Male 384 (88%) 54 (12%) Convex 787 (85%) 134 (15%) Right 727 (86%) 120 (14%) Thoracic 1,458 (84%) 269 (14%) T1–6 537 (80%) 137 (20%)

Female 10 84 0.08 Female 1,502 (85%) 261 (15%) 0.2 Concave 1,099 (86%) 181 (14%) 0.8 Left 1,159 (86%) 195 (14%) 0.9 Lumbar 428 (90%) 46 (10%) 0.001 T7–12 921 (87%) 132 (13%) < 0.001

the rate of neurovascular complication is often less than 1%, the occurrence of such complications should be avoided by all means available. Increased familiarity with the surgical technique and also experience are the most effective means of avoiding screw misplacement. The learning curve for thoracic pedicle screw insertion has recently been studied regarding improvement of different parameters such as duration of surgery, amount of blood loss, major curve correction, hospital stay, and complication rate (Lonner et al. 2009) but we could not find any studies on the learning curve regarding accurate screw placement. We found an average misplacement rate of 14%, with a clear learning curve. The rate of misplacement dropped from 20% at the end of the first period of the study to 13% at the end of the second period. We believe that the continued reduction of the misplacement rate to 11% at the end of the study period can be attributed to the experience gathered, and increased familiarity with the surgical technique.

The senior surgeon received a yearly report on the misplacement rate from the radiologist evaluating the screw placement. The radiological and medical records of the patients were scrutinized at a joint evaluation in 2007 (the second period of the study), which resulted in measures that we believe also contributed to the continued improvement in misplacement rate during the remaining periods of the study: (1) The choice of the screws adjusted to the pedicle diameters measured at the preoperative low-dose spine CT contributed to a general decrease in the misplacement rate. (2) Shorter screw length used during the last 2 years of the study period resulted in a reduction in anterior cortical perforation from 8 (in 2005– 2007) to 3 (in 2008–2009), and subsequently a decrease in the number of screws (with ACP or LCP) whose tip was close to the aorta from 16 (in 2006–2007) to 2 (in 2008–2009). (3) The number of screws inserted using the “in-out-in technique” through the pedicle rib unit increased from 6 during the first 2 periods to 17 during the last 2 periods. We believe that the latter at least partly contributed to reduction of the rate of lateral cortical perforation, as the surgeons intentionally used this technique in thoracic segments with narrow pedicles. The decreasing rate of medial cortical perforation was just above the limit for statistical significance, but this type of misplacement decreased continuously throughout the study period. The slight increase in the last year of the study period was attributed to a slight increase in the rate of partial grade of misplacement (MCP1) whereas the rate of total medial cortical perforation (MCP2) continued to decrease. 1 patient reported a postoperative neurological deficit related to the right-sided T8–T10-dermatome. As no screw misplacement was found at these levels, the neurological deficit was considered to be due to a local extraforaminal nerve injury, which may have been related to some operative injury during preparation before the pedicle insertion. 2 patients (2 screws) were reoperated for removal of misplaced pedicle screws. In our opinion, removal of misplaced pedicles screws should be restricted to: (1) misplaced pedicle screws in patients with neurological deficit corresponding to the level of misplacement, (2) misplaced pedicle screws with screw tip abutting the

Table 4. The number of screws inserted, the number of the level fixed, and the rate of screw misplacement in relation to different types of scoliotic curves (according to the Lenke classification) Lenke Number of type levels fixed 1 2 3 4 5 6

12 (10–15) 11 (7–14) 13 (11–15) 13 (11–15) 11 (7–13) 13 (10–15)

Number of pedicle screws inserted

Rate of all types of misplacement

Rate of LCP

Rate of MCP

18 (14–26) 17 (13–23) 22 (19–25) 21 (17–24) 20 (14–24) 21 (16–27)

16 (0–47) 12 (0–33) 13 (0–41) 13 (0–29) 13 (0–32) 12 (0–25)

8 (0–41) 6 (0–16) 7 (0–20) 6 (0–13) 6 (0–21) 6 (0–14)

7 (0–35) 4 (0–27) 3 (0–23) 5 (0–8) 3 (0–13) 5 (0–15)


Acta Orthopaedica 2011; 82 (1): 50–55

aorta, and (3) medially misplaced pedicle screws with canal encroachment exceeding the pedicle diameter. Studies dealing with screw misplacement following posterior spinal fixation with segmental pedicle screw fixation have shown varying degrees of screw misplacement depending on (a) whether only thoracic, lumbar spine, or both were included, (b) the radiological modality used for evaluation of screw insertion, (c) the grading system used during the evaluation, and (d) the peroperative guide used during screw insertion. Misplacement rates as low as 2.7% were reported for screw insertion of the lumbar spine (Schwarzenbach et al. 1997) and as low as 1.5% when screw insertion was evaluated with plain radiography (Suk et al. 2001). Previous studies have, however, shown that CT may reveal a 10 times higher misplacement rate than does plain radiography (Farber et al. 1995). A study using a strict grading system that classified even 1 mm of cortical violation as misplacement reported misplacement rates of up to 50% (Upendra et al. 2008). Screw misplacement was found to be substantially reduced using computerassisted orthopedic surgery (CAOS) in pedicle screw insertion in a porcine cadaver model (Richards et al. 2007) and using a CT-guided O-arm (Nottmeier et al. 2010). Our reported rate of screw misplacement of 14% for the whole study period (11% for the last period (2009)) is in accordance with most of the reports in the literature for patients with scoliosis (Halm et al. 1996, Liljenqvist et al. 1997, Di Silvestre et al. 2007). However, the total rate of misplacement often overshadows the successive improvement over time—as our study revealed, namely a clear improvement in screw placement during the last year of the study (11%). Our findings validate our previous conclusions about the reliability of low-dose spine CT and our proposed grading system in the evaluation of screw misplacement (Abul-Kasim et al. 2009b, 2010). As in other surgical fields, training, coaching, team work, feedback to surgeons, and experience are among the most important tools to reduce the rate of misplacement and other complications—as well as reduction of the duration of surgery, blood loss, and hospital stay. Surgical navigation systems such as O-arm providing 2-D and 3-D intraoperative imaging may further reduce the rate of screw misplacement in spinal deformity surgery. However, this technique means increased exposure of young individuals with AIS to radiation (and thus needs to be optimized with regard to radiation dose), increased cost, and extra training of the surgeons and operating staff.

Both authors contributed to the conception and design of the study, critical analysis of the data, interpretation of the findings, and preparation and critical revision of the manuscript.

We are grateful to Jan Åke Nilsson (B.S.), Department of Orthopedic Surgery, University of Lund, University Hospital, Malmö for his advice on statistics and for his assistance with the statistical analysis.

55

No competing interests declared.

Abul-Kasim K, Overgaard A, Maly P, Ohlin A, Gunnarsson M, Sundgren P C. Low-dose helical computed tomography (CT) in the perioperative workup of adolescent idiopathic scoliosis. Eur Radiol 2009a; 19 (3): 610-8. Abul-Kasim K, Strombeck A, Ohlin A, Maly P, Sundgren P C. Reliability of low-radiation dose CT in the assessment of screw placement after posterior scoliosis surgery, evaluated with a new grading system. Spine 2009b; 34 (9): 941-8. Abul-Kasim K, Ohlin A, Strombeck A, Maly P, Sundgren P C. Radiological and clinical outcome of screw placement in adolescent idiopathic scoliosis: evaluation with low-dose computed tomography. Eur Spine J 2010; 19 (1): 96-104. Belmont P J, Jr., Klemme W R, Dhawan A, Polly D W, Jr. In vivo accuracy of thoracic pedicle screws. Spine 2001; 26 (21): 2340-6. Coe J D, Shaffrey C I, Alert V, Berven S H. Complications in spinal fusion for adult scoliosis. A report of the scoliosis research society morbidity and mortality committee. In: SRS Annual Meeting, 2006. Di Silvestre M, Parisini P, Lolli F, Bakaloudis G. Complications of thoracic pedicle screws in scoliosis treatment. Spine 2007; 32 (15): 1655-61. Farber G L, Place H M, Mazur R A, Jones D E, Damiano T R. Accuracy of pedicle screw placement in lumbar fusions by plain radiographs and computed tomography. Spine 1995; 20 (13): 1494-9. Halm H, Liljeqvist U, Link T, Jerosch J, Winkelmann W. Computerized tomography monitoring of the position of pedicle screws in scoliosis surgery. Z Orthop Ihre Grenzgeb 1996; 134 (6): 492-7. Kim Y J, Lenke L G. Thoracic pedicle screw placement: free-hand technique. Neurol India 2005; 53 (4): 512-9. Kuklo T R, Potter B K, Polly D W, Jr., Lenke L G. Monaxial versus multiaxial thoracic pedicle screws in the correction of adolescent idiopathic scoliosis. Spine 2005; 30 (18): 2113-20. Lee S M, Suk S I, Chung E R. Direct vertebral rotation: a new technique of three-dimensional deformity correction with segmental pedicle screw fixation in adolescent idiopathic scoliosis. Spine 2004; 29 (3): 343-9. Liljenqvist U R, Halm H F, Link T M. Pedicle screw instrumentation of the thoracic spine in idiopathic scoliosis. Spine 1997; 22 (19): 2239-45. Lonner B S, Kondrachov D, Siddiqi F, Hayes V, Scharf C. Thoracoscopic spinal fusion compared with posterior spinal fusion for the treatment of thoracic adolescent idiopathic scoliosis. J Bone Joint Surg (Am) 2006; 88 (5): 1022-34. Lonner B S, Auerbach J D, Estreicher M B, Kean K E. Thoracic pedicle screw instrumentation: the learning curve and evolution in technique in the treatment of adolescent idiopathic scoliosis. Spine 2009; 34 (20): 2158-64. Nottmeier E W, Pirris S M, Balseiro S, Fenton D. Three-dimensional imageguided placement of S2 alar screws to adjunct or salvage lumbosacral fixation. Spine J 2010 Apr 29. [Epub ahead of print] Richards P J, Kurta I C, Jasani V, Jones C H, Rahmatalla A, Mackenzie G, Dove J. Assessment of CAOS as a training model in spinal surgery: a randomised study. Eur Spine J 2007; 16 (2): 239-44. Schwarzenbach O, Berlemann U, Jost B, Visarius H, Arm E, Langlotz F, Nolte L P, Ozdoba C. Accuracy of computer-assisted pedicle screw placement. An in vivo computed tomography analysis. Spine 1997; 22 (4): 452-8. Suk S I, Lee C K, Kim W J, Chung Y J, Park Y B. Segmental pedicle screw fixation in the treatment of thoracic idiopathic scoliosis. Spine 1995; 20 (12): 1399-405. Suk S I, Kim W J, Lee S M, Kim J H, Chung E R. Thoracic pedicle screw fixation in spinal deformities: are they really safe? Spine 2001; 26 (18): 2049-57. Upendra B N, Meena D, Chowdhury B, Ahmad A, Jayaswal A. Outcomebased classification for assessment of thoracic pedicular screw placement. Spine 2008; 33 (4): 384-90.


56

Acta Orthopaedica 2011; 82 (1): 56–63

Would loss to follow-up bias the outcome evaluation of patients operated for degenerative disorders of the lumbar spine? A study of responding and non-responding cohort participants from a clinical spine surgery registry Tore K Solberg1,2, Andreas Sørlie1,2, Kristin Sjaavik1,2, Øystein P Nygaard2,3, and Tor Ingebrigtsen1,4 1Department of Neurosurgery, University Hospital of Northern Norway, Tromsø; 2Norwegian Registry for Spine Surgery (NORspine), Northern Norway Regional Health Authority, Tromsø; 3National Centre for Spinal Disorders and Department of Neurosurgery, University Hospital of St. Olav, Trondheim; 4Institute of Clinical Medicine, University of Tromsø, Norway Correspondence: tore.solberg@unn.no Submitted 10-01-13. Accepted 10-09-21

Background and purpose Loss to follow-up may bias the outcome assessments of clinical registries. In this study, we wanted to determine whether outcomes were different in responding and non-responding patients who were included in a clinical spine surgery registry, at two years of follow-up. In addition, we wanted to identify risk factors for failure to respond. Methods 633 patients who were operated for degenerative disorders of the lumbar spine were followed for 2 years using a local clinical spine registry. Those who did not attend the clinic and those who did not answer a postal questionnaire—for whom 2 years of outcome data were missing—and who would be lost to follow-up according to the standard procedures of the registry protocols, were defined as non-respondents. They were traced and interviewed by telephone. Outcome measures were: improvement in health-related quality of life (EQ-5D), leg pain, and back pain; and also general state of health, employment status, and perceived benefits of the operation. Results We found no statistically significant differences in outcome between respondents (78% of the patients) and nonrespondents (22%). Receipt of postal questionnaires (not being summoned for a follow-up visit) was the strongest risk factor for failure to respond. Forgetfulness appeared to be an important cause. Older patients and those who had complications were more likely to respond. Interpretation A loss to follow-up of 22% would not bias conclusions about overall treatment effects and, importantly, there were no indications of worse outcomes in non-respondents. 

Clinical registries are increasingly being used to monitoring treatment effectiveness and for evaluation of risk factors associated with different outcomes. Loss to follow-up may seriously bias the outcome assessments of clinical registries, and will reduce the statistical power due to smaller sample size (Hunt and White 1998, Hollis and Campbell 1999, Parker and Dewey 2000, Shih 2002, Gluud 2006). Information about outcomes of patients who do not respond at followup is valuable both for clinicians and researchers. In limited clinical trials, one can make vigorous attempts to trace and retain cohort members. Such efforts would be too expensive and resource-demanding in large population-based registries (Roder et al. 2005, Fritzell et al. 2006). Thus, researchers who use registry data will have to deal with higher numbers of non-respondents being lost to follow-up (Hunt and White 1998). If the outcomes of non-respondents and respondents are different, wrong conclusions could be drawn about the beneficial and harmful effects of interventions (Gluud 2006). Several studies have indicated that individuals who drop out of clinical trials have worse outcomes than those who do not (Sims 1973, Murray et al. 1997, Norquist et al. 2000, Ludemann et al. 2003, Kim et al. 2004). Different imputation methods have been developed to compensate for missing outcomes (Rubin and Schenker 1991, Little and Yau 1996, Shih and Quan 1997, Wood et al. 2004), but these methods are also susceptible to bias, since they rely on assumptions made about the dropouts (Hollis and Campbell 1999, Shih 2002). Studies of the “true” outcomes in non-respondents may help us to make the right assumptions about outcomes

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2010.548024


Acta Orthopaedica 2011; 82 (1): 56–63

57

Patients operated n=667 Operated but not recorded in the registry (by mishap) n=10 (1.5%) Died, n=11 (1.6%), cause: – cerebrovascular disease, n=4 – cancer, n=3 – pneumonia/sepsis, n=3 – myocardial infarction, n=1

Excluded, n=13 (1.9%), due to: – trauma, n=7 – psychosis, n=2 – drug abuse, n=2 – language barriers, n=2

Included in the analysis n=633 (94.9%)

Respondents n=491 (77.6%)

Non-respondents n=142 (22.4%) (11/142 consistent non-respondents) Lost, not traceable n=4/142 (2.8%) (1/4 consistent non-respondents)

Figure. Study population.

Interviewed n=138/142 (97.2%)

of patients who are lost to follow-up. In addition, to prevent loss to follow-up, we need information about risk factors for failure to respond. Here we present a prospective study of patients who were operated for degenerative disorders of the lumbar spine. We assessed the outcomes of non-respondents, who would be lost to follow-up according to the standard procedures of registry protocols, and compared their outcomes with those of patients who responded, in order to evaluate whether the missing outcomes would bias conclusions about treatment effectiveness. We also wanted to identify risk factors for failure to respond.

Patients and methods Study population This study comprised all consecutive patients (n = 633) registered with 1 operation for degenerative disorders of the lumbar spine at the Department of Neurosurgery, University Hospital of Northern Norway (UNN), from Jan 1, 2000 through Dec 31, 2003 (Figure). Data collection and registration was part of the daily routines of the department, involving the entire staff, and the study population represented the total population operated and included in the registry at the unit (Solberg et al. 2005a, b). The mean age of the patients (63% men) was 45 (16–83) years (Table 1). All patients were operated at 1 or 2 levels between L2 and S1. 557 (88%) were operated for the first time, and 76 (12%) had been operated previously. Of these 76 patients, 47 (62%) were reoperated at the same level, 25

(33%) at different level(s), and 4 (5%) were reoperated at both the same and different level(s). Follow-up time from the date of operation (baseline) was 2 years. The registry database was linked to the National Population Registry of Norway through the national 11-digit personal identification number. In this way, we obtained continuously updated information about changes of home address and dates of death in the study population. Causes of death were available from the medical records of the hospitals in our region. We excluded participants who died within 2 years of followup. The causes of death were not related to the initial surgery. However, 1 patient (aged 67) died 26 days after the operation, of an acute myocardial infraction. We excluded 13 patients whose outcome evaluations would be biased by other severe, conflicting problems—as described in Figure. Informed consent was obtained from all participants. The registry protocol was approved by the Data Inspectorate of Norway. Registry protocols/follow-up In the year 2000, a comprehensive clinical spine surgery registry for quality control and research was established at UNN. Based on experiences from the Swedish Spinal Register (SweSpine) (Fritzell et al. 2006) and previous validation studies from the local clinical registry at UNN (Solberg et al. 2005a, Solberg et al. 2005b), the local registry of UNN was expanded to a national registry in 2007: the Norwegian Registry for Spine Surgery (NORspine). We have evaluated data obtained from the 2 protocols of the local registry at UNN. Protocol A was used in 2000 and 2001 and was changed to protocol B, which was used in 2002 and 2003. The only difference between the two protocols was how data were collected at 2 years of follow-up. Patients operated before 2002 (protocol A) were summoned for follow-up visits at the outpatient clinic at 24 months, whereas patients operated later (protocol B) received postal questionnaires. We could therefore investigate how these differences in obtaining follow-up data influenced response rates. All patients were summoned for follow-up visits at 3 and 12 months at an outpatient clinic. The questionnaires and a stamped, addressed return envelope were distributed by ordinary postal mail, to be completed at home by the patients. An independent observer, a research nurse responsible for all follow-up visits, collected and checked all the returned questionnaires and interviewed the patients about employment status and complications. Travel expenses were covered by the public National Insurance Organization. At 2 years, patients who did not attend the clinic (protocol A) got one reminder by telephone within a few days, from the research nurse. They were asked to make a new appointment for a follow-up visit or to respond by postal mail. Patients who did not return the questionnaire at 2 years (protocol B) got 1 reminder with a new copy of the postal questionnaire and a stamped, addressed return envelope.


58

Acta Orthopaedica 2011; 82 (1): 56–63

Table 1. Characteristics of the study population All I II n= 633 Respondents Non-respondents n =491 n= 142

III a Consistent non-respondents n = 12

p-value of difference I vs. II I vs. III

Age, median (95% CI) 42 (41–44) 43 (41–44) 40 (37–44) 34 (29–51) 0.04 0.04 Females (%) 233 (36.8) 187 (38.1) 46 (32.4) 4(33.3) 0.2 0.7 BMI, median (95% CI) kg/m2 b 26 (25–26) 26 (25–26) 26 (25–26) 25 (22–27) 0.7 0.5 Smokers (%) 296 (47) 224 (46) 71 (50) 8 (67) 0.4 0.2 Living alone, n (%) 171 (27) 126 (26) 45 (32) 8 (67) 0.2 0.002 Had child less than 8 years, n (%) 178 (28) 133 (27) 45 (32) 3 (25) 0.3 0.9 Weeks on sick leave, median (95% CI) 8 (5–11) 8 (6–12) 4 (1–13) 1 (0–24) 0.1 0.2 Days of hospital stay, median (95%CI) 4 (4–4) 4 (4–4) 3 (3–4) 3 (2–4) 0.02 0.1 Previous low back operation, n (%) 76.0 (12) 64 (13) 12.0 (9) 1 (8) 0.1 0.6 Had complication to the surgery, n (%) 31 (5) 29 (6) 2 (1) 0 0.03 0.4 EQ-5D score, median (95% CI) 0.16 (0.12–0.19) 0.16 (0.09–0.19) 0.16 (0.09–0.33) 0.23 (-0.10–0.69) 0.4 0.9 Health state, median (95% CI) 40 (35–40) 40 (35–40) 43 (38–50) 36 (22–55) 0.08 0.9 Leg pain, median (95% CI) 67 (64–69) 68 (65–70) 65 (60–69) 59 (26–84) 0.5 0.6 Back pain, median (95% CI) 54 (50–60) 55 (50–61) 52 (45–57) 51 (45–83) 0.1 0.6 Were anxious and/or depressed, n (%) c 304 (48) 231 (47) 73 (51) 8 (67) 0.4 0.2 Educational level, n (%) 1.0 d 0.8 d Primary school 189 (30) 148 (30) 41 (29) 5 (42) 218 (34) 168 (34) 50 (35) 3 (25) Vocational school 74 (12) 60 (12) 14 (10) 1 (8) Gymnasium / high school 81 (13) 66 (13) 15 (11) 1 (8) University or college < 4 years 71 (11) 49 (10) 22 (16) 2 (17) University or college > 4 years Employment status, n (%) 0.7 e 0.8 e On sick leave 345 (55) 276 (56) 69 (49) 6 (50) On partial sick leave 33 (5) 24 (5) 9 (6) Working full time 79 (13) 54 (11) 25 (18) 2 (17) Homemaker 10 (2) 9 (2) 1 (1) Student 28 (4) 12 (4) 7 (5) 1 (8) Unemployed 9 (1) 7 (1) 2 (1) 1 (8) Retired pensioner 62 (10) 53 (11) 9 (6) f On rehabilitation 27 (4) 19 (4) 8 (6) Disability pensioner 40 (6) 28 (6) 12 (9) 2 (17) ASA grade I–V, n (%) g 0.8 h 0.4 h Grade I 135 (40) 116 (36) 48 (40) 2 (22) Grade II 189 (57) 192 (59) 69 (57) 7 (78) Grade III 10 (3) 17 (5) 4 (3) a Group III was a subgroup of group II. b Body mass index. c Mild to severe problems. d University or college education? (yes/no). e On full or partial sick leave, on rehabilitation, or disability pensioner? (yes/no). f Patients having received worker’s compensation for more than 12 months with the prospect of returning back to work, or permanent disability status. g No patients had ASA grade > III. h ASA grade I vs. grade II and III.

Respondents/non-respondents Patients for whom 2-years of follow-up data were missing, despite these measures, would be lost to follow-up under standard protocol conditions. They were defined as non-respondents (group II, n = 142; protocol A, n = 37; protocol B, n = 105) and they were invited to participate in the study by telephone interview. Patients who did not respond at 3, 12, or 24 months were classified as consistent non-respondents (group III, n = 12: protocol A, n = 8; protocol B, n = 4). Thus, group III was a subgroup of group II. The rest of the patients were defined as respondents (group I, n = 491) (Figure). We used 3 sources for tracing the non-respondents: the National Population Registry of Norway, publicly available online telephone directories (Harvey et al. 2003), and the electronic medical records of the hospital. 138 of the 142 nonrespondents were interviewed by telephone in a standardized

fashion (Hunt and White 1998) by the same interviewer (AS). These patients were instructed to report their condition at 2 years after surgery. The patients were also asked to give their main reason for not responding. When data collection was complete, the study group had a consensus meeting where patients’ answers were categorized into 5 main reasons for not responding: “forgot to complete or return the questionnaire”, “questionnaire fatigue”, “sickness”, “could not remember having received questionnaires”, and “family- or work-related problems”. Baseline data At admission, the patients completed the baseline questionnaire. During their hospital stay, the surgeon recorded data concerning diagnosis, treatment, employment status, and


Acta Orthopaedica 2011; 82 (1): 56–63

duration of symptoms according to a standard registration form. Finally, all questionnaires and forms were collected and checked for completeness by a dedicated research nurse. Questionnaires The questionnaires completed by the patients at baseline and follow-up were identical, and were used for outcome assessments, including interviews. The baseline questionnaire contained additional questions about demographics and lifestyle issues. The primary outcome measure was the EuroQol-5D (EQ-5D) questionnaire. Secondary outcome measures were perceived benefit of the operation, employment status, and visual analog scales (VAS) for leg pain, back pain, and state of health. EQ-5D EQ-5D is a generic and preference-weighted measure of health-related quality of life (HRQL). It evaluates 5 dimensions: mobility, self-care, activities of daily life, pain, and anxiety and/or depression. For each dimension, the patient describes 3 possible levels of problems (none, mild to moderate, or severe). Hence, this descriptive system contains 243 (35) combinations or index values for health states (the EuroQol Group 1990). We used the value set based on the main survey from the EuroQol group (Dolan et al. 1996, Dolan 1997), which has been validated for this patient population (Solberg et al. 2005b). Total range of score is from –0.594 to 1, where 1 corresponds to perfect health and 0 to death. Negative values are considered to be worse than death (the EuroQol Group 1990). Health state EuroQol VAS forms the second part of the EQ-5D questionnaire. The patients rate their general state of health by drawing a line from a box marked “your health state today” to the appropriate point on the 20-cm VAS scale, which ranges from 0 to 100 (worst to best imaginable health) (the EuroQol Group 1990). Benefit of the operation At follow-up, the patients were asked: “How much benefit have you had from the operation?” The response alternatives were “very much”, “quite a lot”, “some”, “none at all” or “uncertain” (Solberg et al. 2005a, b). Leg pain and back pain Pain intensity was graded by the patient in 2 separate 100-mm VAS for leg and back pain (where 0 = no pain). The American Society of Anesthesiologists (ASA) grading system ASA grade was registered for each patient by a doctor or a specialized nurse before surgery. ASA grade (I–V) classifies patients according to their vulnerability, i.e. physical condi-

59

tion (from no disease to life-threatening systemic disease) (Dripps 1963). Before 2002, data on ASA grade were not registered systematically (62% missing data), and they were therefore omitted from the analysis. Of the data from 2002 and 2003, only 9% were missing. These values (except 1) could be obtained from the medical records of the patients. Statistics We tested whether within-group change scores were statistically significant (change from baseline to follow-up), using paired t-test or Wilcoxon’s matched-pairs signed rank test depending on the distribution of the data. Baseline characteristics and differences in outcome between subgroups (I–III) were assessed with independent-samples t-test, Mann-Whitney U-test, or Chi-square test. Central tendency is presented as mean when normally distributed, and as median when skewed. Confidence intervals for medians were calculated according to McKean and Schrader (1984). We assessed risk factors for not responding at 2 years of follow-up in multivariate analysis, using respondents (value = 0) vs. non-respondents (value = 1) as dependent variable. Being summoned for a follow-up visit (protocol A) vs. receiving a postal questionnaire (protocol B) was used as exposition variable. We adjusted for covariates obtained from baseline data (Table 1) using a backward logistic regression model, only if the covariates were judged to be clinically relevant and if baseline values differed significantly (level 0.1) between respondents and non-respondents. To get a better model-data fit, we had dichotomized two covariates: living alone and complications (yes/no). SPSS for Windows version 14.0 was used for all analyses.

Results Non-respondents were younger, were hospitalized for fewer days, and had more complications than the respondents. Consistent non-respondents were more likely to live alone (Table 1). We found no difference in ASA grade between the groups. However, this result is uncertain since we lacked data from 2000 and 2001, when the response rate was highest. Disc herniation treated by microdiscectomy was the commonest operation (Table 2). Response rates The overall response rate declined during the follow-up period, to 77.6% at 24 months. When the protocol was changed from A to B in 2002, the response rate decreased considerably. Patients who were invited for a follow-up visit at the outpatient clinic at 2 years (protocol A) had a higher response rate than patients who only received questionnaires by mail (protocol B) (88% vs. 69%, p < 0.001). 4 patients could not be traced (Figure); among them, 1 was a consistent non-respondent. After obtaining the missing outcomes of the non-respondents by telephone interview, the


60

Acta Orthopaedica 2011; 82 (1): 56–63

Table 2. Indications for and types of surgery among respondents non-respondents Group All I Respondents n = 633 n = 491

II III a Non- Consistent respondents non-respondents n = 130 n = 12

Indications for surgery, n (%) Lumbar disc herniation 519 (82) 399 (81) 120 (84) 12 (100) Central spinal stenosis 94 (15) 78 (16) 16 (11) Lateral spinal stenosis 39 (6) 31 (6) 8 (6) Segmental instability 31 (5) 24 (5) 7 (5) b Sum 683 532 151 12 Types of surgery, n (%) Microdiscectomy 476 (75) 362 (74) 114 (80) 11 (92) Laminectomy 111 (17) 90 (18) 21 (15) Instrumented fusion 30 (5) 24 (5) 6 (4) Chemonucleolysis 16 (3) 15 (3) 1 (1) 1 (8) Sum 633 491 142 12 a

Group III was a subgroup of group II. b Patients could have more than one indication for surgery.

Table 3. Sequential outcomes of the study population during 2 years of follow-up Follow-up n (response rate) EQ-5D score b Health state b Leg pain b Back pain b Benefited from the operation, n (%) c Received worker’s compensation, n (%) d

3 months

12 months

2 years

598 (95%) 0.45 (0.41–0.48) < 0.001 29 (27–31) < 0.001 43 (41–46) < 0.001 29 (27–31) < 0.001

574 (91%) 0.46 (0.43–0.50) < 0.001 31 (29–34) < 0.001 41 (38–44) < 0.001 28 (25–31) < 0.001

629 (99%) a 0.46 (0.43–0.49) < 0.001 30 (28–32) < 0.001 41 (38–44) < 0.001 27 (24–30) < 0.001

539 (90)

527 (92)

571 (91)

335 (57)

166 (31)

181 (29)

a Includes non-respondents interviewed by telephone. b Absolute values (improvements from baseline) are shown as mean change, (95% CI) and p-value c Patients who stated that they had “some”, “much”, or “very much” benefit from the operation . d Patients who were on full or partial sick leave, on rehabilitation, or disability pensioners.

outcome data were 99% complete (Table 3). None of the nonrespondents refused to be interviewed. To trace and interview non-respondents was time consuming. The mean time from the operation until all the data concerning 24 months of follow-up had been collected was 2 years for the respondents and 3 years for the non-respondents. We identified 5 main reasons for not responding: forgot to complete or return the questionnaire (n = 87, 63%), questionnaire fatigue (n = 23, 17%), sickness (n = 15, 11%), could not remember having received questionnaires that had been sent (n = 7, 5%), and family- or work-related problems (n = 5, 4%). Information from 1 patient was missing. Outcome assessment Both primary and secondary outcome measures improved after the operation. These effects persisted throughout the observation period (Table 3). There were no statistically significant differences in outcome between respondents and non-respondents or between

respondents and consistent non-respondents, measured by employment status and perceived benefits of the operation at 2 years of follow-up, and improvements in HRQL, health state, leg pain, and back pain (Table 4). For the non-respondents, there were no statistically significant differences in outcomes between those who did not attend the outpatient clinic (protocol A) and those who did not respond to a postal questionnaire (protocol B) (data not shown). Complications 31 patients (5%) had 34 complications (Table 5). Complications were more frequent among the respondents than among the non-respondents (7% vs. 1%, p = 0.03). Risk factor analysis 2 independent risk factors for failure to respond were found by multivariate analysis (Table 6). Patients (operated in 2002 and 2003) who only received postal questionnaires (protocol


Acta Orthopaedica 2011; 82 (1): 56–63

61

Table 4. Subgroup analyses of respondents and non-respondents at 2 year Outcome a EQ-5D score c Health state d Leg pain d Back pain c Benefited from the operation, n (%) e Received workers compensation, n (%) f

I . Respondents n = 491 (77.6%) 0.46 (0.36–0.60) 31 (28–34) 40 (37–43) 22 (18–28)

< 0.001 < 0.001 < 0.001 < 0.001

II . Non-respondents n = 138 (21.8%) 0.41 (0.30–0.64) 27 (22–32) 44 (38–50) 26 (17–32)

< 0.001 < 0.001 < 0.001 < 0.001

III. Consistent non-respondents b n = 11 0.64 (0.19–0.76) 28 (13–43) 43 (25–61) 40 (17–66)

0.003 0.002 0.001 0.008

p-value of difference I vs. II I vs. III 0.8 0.1 0.3 1.0

0.6 0.7 0.8 0.3

447 (91)

124 (91)

11 (100)

0.8

0.3

141 (29)4

0 (29)

4 (36)

1.0

0.6

a Improvements from baseline (absolute values) are shown. b Group III is a subgroup of group II. c Median change, (95% CI) and p-value d Mean change, (95% CI) and p-value e Patients who stated that they had “some”, “much”, or “very much” benefit from the operation. f Patients who were on full or partial sick leave, on rehabilitation, or disability pensioners.

Table 5. Types of complications in 31 (5%) of 633 patients a Complications All n = 633 Dural tear Deep wound infection Superficial wound infection Urinary bladder infection Reoperation within the same hospital stay Intraoperative nerve root injury Postoperative muscle hernia Gall bladder infection Deep leg vein thrombosis Gastric ulcer hemorrhage Minor myocardial infarction Sum, n (%)

9 5 10 2

I. Respon- II. Nondents respondents n = 491 n = 142 9 5 9 2

1

tient clinic (protocol A) (odds ratio (OR) = 3, 95% CI: 2–5). A 1-year increase in age increased the probability of responding by 2% (OR = 0.98). Having had a complication and living alone were not independent risk factors in the multivariate analysis (Table 6).

Discussion

We found similar outcomes between respondents and nonrespondents at 2 years of follow-up in patients who were oper1 1 ated for degenerative disorders of the lumbar spine, assessed 1 1 1 1 as changes in HRQL (EQ-5D) score, pain, and state of health, 1 1 or employment status and perceived benefit. Importantly, 1 1 the non-respondents did not have poorer outcomes than the 1 1 34 (6) 32 (7) 2 (1) respondents. However, better outcome in consistent nonrespondents might have reached statistical significance if the a 3 of the patients had 2 complications. No complications occurred in sample size had been larger. The patients reported forgetfulconsistent non-respondents (group III). ness as the main reason for not responding. The patients most likely to respond were those who were summoned for followup visits and older patients. B) at 2 years of follow-up were less likely to respond than It has been suggested that as a rule of thumb, a loss to folthose who were summoned for a follow-up visit at the outpa- low-up of greater than 20% probably leads to assessment bias, whereas a rate of less than 5% would not (Sackett et al. 2000, Table 6. Risk factors for failure to respond at 2 years of follow-up in 633 patients Schulz and Grimes 2002). Our results indicate that a 22% loss Factors OR a, b 95% CI p-value OR c 95% CI p-value to follow-up does not alter the conclusions about the overall Only received postal questionnaires at 2 years d 3.2 (2.1–4.9) < 0.001 3.2 (2.1–4.9) < 0.001 effects of treatment within the Age 0.99 (0.97–1.0) 0.05 0.98 (0.97–1.0) 0.05 whole, large cohort. In staHad any complication 0.23 (0.54–1.0) 0.05 0.26 (0.06–1.1) 0.07 tistical terms, we could treat Days of hospital stay 0.89 (0.82–0.96) 0.01 0.99 (0.90–1.1) 0.9 1.0 (0.99–1.0) 0.1 Health state the non-respondents as if they Living alone 1.3 (0.89–2.0) 0.2 were missing at random (Shih a OR: odds ratio. b Univariate analysis. c Multivariate analysis. d Protocol A, operated in 2000 and 2001. 2002). However, by simply 2

2


62

ignoring the non-respondents, somewhat older patients and those who had complications would be over-represented. Where there were lower response rates, this could confound the overall assessments towards poorer treatment effects if older patients and those who had complications tended to report poorer outcomes. To prevent selection bias, for example when comparing subgroups of patients with different response rates, the treatment effects should be adjusted for clinically relevant risk factors associated with responding (Etter and Perneger 1997, Wood et al. 2004). The safest way to avoid bias is to reduce loss to follow-up. Our study shows that patients who only received postal questionnaires were 3 times less likely to respond than those who were summoned for follow-up visits. Similar results have been published previously (Sitzia and Wood 1998). It would be too demanding on resources to arrange long-term follow-up visits for the participants in large clinical registries (Roder et al. 2005, Fritzell et al. 2006). The patients would therefore have to be contacted at home. Several ways of increasing response rates to postal questionnaires have been recommended (Etter and Perneger 1997, Edwards et al. 2002, 2007, Etter et al. 2002, Schulz and Grimes 2002). We found that forgetfulness was the most important reason for failure to respond. This problem can be prevented by sending early reminders to study participants, for example by using modern telecommunication. SMS and e-mail are now widely available, especially to younger patients who are less likely to respond.We assessed a homogenous patient population living in a typical Northern European society where most public health services are free, national population registries are updated, and the level of social security is high. Thus, people from lower socioeconomic classes and patients with disability can afford to respond, and can be given help to respond. This might explain why we did not find worse outcomes in the non-respondents. Our findings may not be valid for populations living under other ethnic and socioeconomic conditions. One weakness of this study is that only non-respondents were interviewed by telephone, with a time delay of 12 months. The delayed interviews may have introduced recall bias. However, previous reports on sequential long-term outcomes in similar patient populations have shown that the outcomes are relatively stable (Findlay et al. 1998, Amundsen et al. 2000, Atlas et al. 2000). Thus, we would expect recall bias to be small. Some studies have indicated that interview subjects tend to overestimate favorable outcomes (Burroughs et al. 2001, Ludemann et al. 2003), but the opposite has also been suggested (Wildner 1995). In our study, the non-respondents did not report better outcomes, even though they were somewhat younger and had fewer complications than patients who responded. It was beyond the scope of this study to evaluate assessment bias due to deaths in study participants. Cohort members who die during follow-up must be accounted for and handled separately in the analyses, as previously described (Lachin 1999, Shih 2002).

Acta Orthopaedica 2011; 82 (1): 56â&#x20AC;&#x201C;63

TKS and AS: idea, protocol, data collection, data analysis, and writing. KS: protocol, data collection, and writing. Ă&#x2DC;PN: protocol, data analysis, and writing. TI: idea, protocol, data analysis, and writing.

Amundsen T, Weber H, Nordal H J, Magnaes B, Abdelnoor M, Lilleas F. Lumbar spinal stenosis: conservative or surgical management?: A prospective 10-year study. Spine 2000; 25: 1424-35. Atlas S J, Keller R B, Robson D, Deyo R A, Singer D E. Surgical and nonsurgical management of lumbar spinal stenosis: four-year outcomes from the maine lumbar spine study. Spine 2000; 25: 556-62. Burroughs T E, Waterman B M, Cira J C, Desikan R, Claiborne D W. Patient satisfaction measurement strategies: a comparison of phone and mail methods. Jt Comm J Qual Improv 2001; 27: 349-61. Dolan P. Modeling valuations for EuroQol health states. Med Care 1997; 35: 1095-108. Dolan P, Gudex C, Kind P, Williams A. The time trade-off method: results from a general population study. Health Econ 1996; 5: 141-54. Dripps R. New classification of physical status. Anesthesiology 1963; 24: 111. Edwards P, Roberts I, Clarke M, DiGuiseppi C, Pratap S, Wentz R, Kwan I. Increasing response rates to postal questionnaires: systematic review. BMJ 2002; 324: 1183. Edwards P, Roberts I, Clarke M, DiGuiseppi C, Pratap S, Wentz R, Kwan I, Cooper R. Methods to increase response rates to postal questionnaires. Cochrane Database Syst Rev 2007;MR000008. Etter J F, Perneger T V. Analysis of non-response bias in a mailed health survey. J Clin Epidemiol 1997; 50: 1123-8. Etter J F, Cucherat M, Perneger T V. Questionnaire color and response rates to mailed surveys. A randomized trial and a meta-analysis. Eval Health Prof 2002; 25: 185-99. Findlay G F, Hall B I, Musa B S, Oliveira M D, Fear S C. A 10-year followup of the outcome of lumbar microdiscectomy. Spine 1998; 23: 1168-71. Fritzell P, Stromqvist B, Hagg O. A practical approach to spine registers in Europe: the Swedish experience. Eur Spine J (Suppl 1) 2006; 15: S57-S63. Gluud L L. Bias in clinical intervention research. Am J Epidemiol 2006; 163: 493-501. Harvey B J, Wilkins A L, Hawker G A, Badley E M, Coyte P C, Glazier R H, Williams J I, Wright J G. Using publicly available directories to trace survey nonresponders and calculate adjusted response rates. Am J Epidemiol 2003; 158: 1007-11. Hollis S, Campbell F. What is meant by intention to treat analysis? Survey of published randomised controlled trials. BMJ 1999; 319: 670-4. Hunt J R, White E. Retaining and tracking cohort study members. Epidemiol Rev 1998; 20: 57-70. Kim J, Lonner J H, Nelson C L, Lotke P A. Response bias: effect on outcomes evaluation by mail surveys after total knee arthroplasty. J Bone Joint Surg (Am) 2004; 86: 15-21. Lachin J M. Worst-rank score analysis with informatively missing observations in clinical trials. Control Clin Trials 1999; 20: 408-22. Little R, Yau L. Intent-to-treat analysis for longitudinal studies with dropouts. Biometrics 1996; 52: 1324-33. Ludemann R, Watson D I, Jamieson G G. Influence of follow-up methodology and completeness on apparent clinical outcome of fundoplication. Am J Surg 2003; 186: 143-7. McKean J W, Schrader R M. A comparison of methods for studentizing the sample mean. Commun Statist 1984; 13: 751-73. Murray D W, Britton A R, Bulstrode C J. Loss to follow-up matters. J Bone Joint Surg (Br) 1997; 79: 254-7. Norquist B M, Goldberg B A, Matsen F A, III. Challenges in evaluating patients lost to follow-up in clinical studies of rotator cuff tears. J Bone Joint Surg (Am) 2000; 82: 838-42.


Acta Orthopaedica 2011; 82 (1): 56–63

Parker C, Dewey M. Assessing research outcomes by postal questionnaire with telephone follow-up. TOTAL Study Group. Trial of Occupational Therapy and Leisure. Int J Epidemiol 2000; 29: 1065-9. Roder C, Chavanne A, Mannion A F, Grob D, Aebi M. SSE Spine Tango— content, workflow, set-up. www.eurospine.org-Spine Tango. Eur Spine J 2005; 14: 920-4. Rubin D B, Schenker N. Multiple imputation in health-care databases: an overview and some applications. Stat Med 1991; 10: 585-98. Sackett D L, Richardson W S, Rosenberg W, Haynes R B. Evidence-based medicine: how to practice and teach EBM. Churchill Livingstone, Edinburgh 2000. Schulz K F, Grimes D A. Sample size slippages in randomised trials: exclusions and the lost and wayward. Lancet 2002; 359: 781-5. Shih W. Problems in dealing with missing data and informative censoring in clinical trials. Curr Control Trials Cardiovasc Med 2002; 3: 4. Shih W J, Quan H. Testing for treatment differences with dropouts present in clinical trials--a composite approach. Stat Med 1997; 16: 1225-39.

63

Sims A C. Importance of a high tracing-rate in long-term medical follow-up studies. Lancet 1973; 2: 433-5. Sitzia J, Wood N. Response rate in patient satisfaction research: an analysis of 210 published studies. Int J Qual Health Care 1998; 10: 311-7. Solberg T K, Nygaard O P, Sjaavik K, Hofoss D, Ingebrigtsen T. The risk of “getting worse” after lumbar microdiscectomy. Eur Spine J 2005a; 14: 49-54. Solberg T K, Olsen J A, Ingebrigtsen T, Hofoss D, Nygaard O P. Healthrelated quality of life assessment by the EuroQol-5D can provide costutility data in the field of low-back surgery. Eur Spine J 2005b; 14: 1000-7. The EuroQol Group. EuroQol--a new facility for the measurement of healthrelated quality of life. The EuroQol Group. Health Policy 1990; 16: 199208. Wildner M. Lost to follow-up. J Bone Joint Surg (Br) 1995; 77: 657. Wood A M, White I R, Thompson S G. Are missing outcome data adequately handled? A review of published randomized controlled trials in major medical journals. Clin Trials 2004; 1: 368-76.


64

Acta Orthopaedica 2011; 82 (1): 64–68

Transposition of the apophysis of the greater trochanter for reconstruction of the femoral head after septic hip arthritis in children 4 children followed for more than 15 years Pål Benum Department of Orthpaedics, St. Olav’s University Hospital and Department of Neuroscience, Norwegian University of Science and Technology, Trondheim, Norway Correspondence: pal.benum@ntnu.no Submitted 10-06-09. Accepted 10-10-02

Background and purpose Total necrosis of the femoral head after infection in children during their first months of life gives a dislocated hip with severe leg shortening. A new femoral head can be achieved with subtrochanteric osteotomy and transposition of the apophysis of the greater trochanter into the acetabulum. Previous reports have dealt with short-term results (up to 12 years). Here I present some results of this procedure 15–24 years after operation. Patients and methods 4 children aged 1–6 years with complete necrosis of the femoral head were operated on with transposition of the greater trochanter. Secondary shelf plasty was performed later in 1 child, distal femoral epiphysiodesis in another, and femoral bone lengthening in 1 child. The mean follow-up period was 19 (15–24) years. Results A new femoral head developed in all hips. 2 of them had a spherical head with a good acetabular cover, and without any osteoarthritis except for slight reduction of cartilage height. These hips were painless, with a mobility that allowed good walking function after 16 and 24 years, respectively. In the other 2 patients, in which there was a severe acetabular dysplasia at the primary operation, the new femoral head was somewhat flattened; painful osteoarthritis led to hip replacement 15 and 21 years after trochanter arthroplasty. Even these patients had a relatively good walking function until the last couple of years before hip replacement. Maximum leg length discrepancy was 7 cm. Interpretation Trochanter arthroplasty with subtrochanteric osteotomy in total femoral head necrosis after septic arthritis in children may give satisfactory long-term results provided adequate acetabular cover is obtained. Although the method cannot provide a normal hip, it can contribute to less length discrepancy, less pain, improved gait, and more favorable conditions for later hip replacement. 

Acute septic arthritis of the hip in children during their first months of life occasionally leads to total necrosis of the femoral head. This results in dislocation of the joint and severe limb shortening. A 6-year-old girl suffering from this condition was admitted to our department in 1985. Bearing in mind that apophyseal cartilage has the potential to develop into a kind of joint cartilage when transplanted to a joint (Benum 1974), I transposed the apophysis of the greater trochanter into the acetabulum in an attempt to reconstruct the femoral head. Encouraged by the short-term result after this operation, and by medium-term results reported after similar operations (Hunka et al. 1982, Dal Monte et al. 1984), I later used this method in 3 other children. I now present the method and the long-term results after this procedure.

Patients and methods 2 girls and 2 boys were operated on (Table 1). They had all developed necrosis of the femoral head due to hip infection following sepsis during their first months of life. In 2 of the children, the sepsis was due to infection with Staphylococcus aureus during intravenous treatment in the first week after birth. One of the children had been treated with antibiotics because of a suspected joint infection at the age of two months. The reason for infection was unknown in 1 child, who was born in India. One of the children (no. 1), who suffered from congenital hydrocephalus, had previously been operated with a valgus osteotomy of the actual right hip since a very small remnant of a cartilaginous head without any continuity to a very short femoral neck had been found at arthrotomy. The opposite (left) hip of this patient had also been treated with a valgization osteotomy (Figure 1). It is uncertain whether the

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2010.548030


Acta Orthopaedica 2011; 82 (1): 64–68

A

65

B

Figure 1. Patient no. 1. A. Preoperatively, showing no femoral head on the right side and the femur positioned considerably more proximally than normal. A valgization osteotomy had been performed previously. A valgization osteotomy had also been performed in the left hip. B. Postoperatively, following transposition of the greater trochanter into the acetabulum on the right side. The subtrochanteric osteotomy was fixated with a Steinman pin. C. 24 years postoperatively, showing a spherical femoral head well covered by a congruent acetabular roof. The cartilage thickness seems adequate. Note the steep femoral neck and the low medial femoral head offset. Severe osteoarthritic changes can be seen in the left hip.

A

C

B

Figure 2. Patient no 2. A. Preoperatively, showing a luxated hip without any signs of a femoral neck and femoral head in the left hip. B. After trochanter arthroplasty and subtrochanteric osteotomy fixated with Steinman pins. Note severe acetabular dysplasia. C. 21 years after trochanter arthroplasty. The femoral head is flattended and poorly covered by the acetabulum. Osteoarthritic changes are present.

varus deformity of the left hip was due to infection of the joint or congenital dysplasia. Age at arthroplasty ranged from 1 to 6 years. At the time of operation, all the children had a proximally dislocated femur and shortening of the limb (Figures 1–4). An anterior approach was used. Total necrosis of the femoral head was found at the operation in all of the children. In child no. 1, in whom a small remnant of a cartilaginous femoral head had been found at the previous operation, there was still no continuity to the remnants of the femoral neck. After removal of fibrous tissue from the acetabulum, the insertion of the gluteal tendons and the upper part of the insertion of the external rotator tendons were released. Then a subtrochanteric osteotomy was performed. The apophysis was slightly trimmed to fit into the acetabulum. Following transposition of the apophysis into the acetabulum, the osteotomy was fixated with Steinmann pins, screws, or a plate—aiming at a neck-shaft angle of approximately 130 degrees. The gluteal tendons were reinserted into the lateral subtrochantric area of the femur as far distally as possible. In child no. 1, an adductor tenotomy had to be peformed to enable

C

abduction of the hip. A spica cast was used for 3 months. In child no. 4, a shelf procedure was performed 6 months after the primary operation due to severe acetabular dysplasia. The patients have now been followed for 24, 21, 16, and 15 years, respectively.

Results Radiographic follow-up showed development of a new femoral head in all hips. Following appearance of a center of ossification in the transposed greater trochanter, a growth zone was visible until the end of the growth period, as shown in Figure 3. In 2 of the hips a nearly spherical head developed, covered by a well-developed acetabulum (Figures 1 and 3). At the final follow-up 24 and 16 years postoperatively, when the patients were 30 and 19 years old, radiographs showed a slight reduction in cartilage height. These hips were painless, with a mobility that allowed good walking function. The other 2 hips, where a good cover for the new femoral head had not been


66

Acta Orthopaedica 2011; 82 (1): 64–68

A

B

C

Figure 3. Patient no 3. A. Preoperatively, showing a totally luxated hip. B. 18 months after arthroplasty, the osteotomy had been fixated with a plate. The arrow points to a tiny ossification center in the new femoral head. C. 6 years after arthroplasty. Note the growth plate below the transposed greater trochanter. D. 6 years postoperatively. The new head is nearly spherical and well covered by a congruent acetabular roof.

achieved, a dysplastic joint developed (Figures 2 and 4). These hips were also painless, and the operation gave an improved gait until secondary osteoarthritis developed and gave indication for hip replacement, 21 and 15 years after primary operation when the patients were 22 and 21 years old, respectively. During growth, a coxa valga with a reduced medial femoral head offset developed in all hips and some degree of limping was seen in all patients. Leg length discrepancy was treated with distal femoral epiphysiodesis in 1 child (no. 2), and lengthening osteotomy of the femur in another (no. 3) (Table 1). At final follow-up, the leg length discrepancies were less than 3 cm, except in 1 patient. The average total range of motion was 120 (105–135) degrees (Table 2).

A

D

Discussion As early as in 1874, severe loss of bone from the femoral head following septic arthritis of the hip was described by Thomas Smith at St. Bartholomew’s Hospital in London (Smith 1874). He showed by postmortem studies that the head and neck had

B

Figure 4. Patient no 4. A. Preoperatively, showing a luxated hip without any signs of a femoral head. Note the severe dysplasia of the acetabulum. B. 15 years after operation, showing that the new femoral head, which is nearly spherical, is poorly covered by a neoacetabulum above a severely dysplastic acetabulum. Osteoarthritis has developed. C. After insertion of a total hip prosthesis.

C

Table 1. Age at primary disease, type of infection, age at trochanter arthroplasty, and secondary operations Patient Age at primary disease 1, 2, 3, 4,

F F M M

3 weeks 10 days 6 months 2 months

Reason for Bacteriology sepsis Intravenous infusion Intravenous infusion Unknown Unknown

S. aureus S. aureus Unknown Unknown

Affected joints Both hips Both hips, left ankle Right hip Right hip

Age at arthro- plasty (years) 6 1 3 6

Secondary operation None Epiphysiodesis Lengthening osteotomy Shelf operation


Acta Orthopaedica 2011; 82 (1): 64–68

67

Table 2. Findings at the final follow-up examination Patient Obs. Radiographic findings Clinical findings at final follow-up period Evaluation Leg length Range of motion b c (years) of the joint difference Pain shape OA a (cm) Ext. Flex. Abd. Add. Int. rot. Ext. rot. Total 1, F 24 Spherical no 0 (+2.0) e None –5 75 10 10 10 5 105 2, F 21 Dysplastic yes –1.3 (–2.4) Moderate 0 80 15 20 10 10 135 3, M 16 Spherical no –2.0 None 0 90 20 10 10 5 135 4, M 15 Dysplastic yes –5.0 (–7.0) Moderate 0 90 15 0 0 0 105 a No OA means no osteoarthritic changes except slight reduction in cartilage height. b Values in parentheses indicate leg length difference including the difference due to subluxation. c Moderate pain indicates pain during activity. d Slight limp indicates limp that does not significantly influence the walking capacity, in contrast to e In patient no. 1, the hip on the opposite side was subluxated.

been totally destroyed in some cases. In clinical practice, the differential diagnosis between necrosis of the femoral head due to septic arthritis and severe dysplasia of the hip may sometimes be difficult. In my series, the operative findings strongly indicated that joint infection was the reason for femoral head necrosis even in the 2 children without information about bacterial growth. An attempt to stabilize the hip in a child with total necrosis of the femoral head was reported by L’Episcopo (1936). He split the proximal end of the remaining femur and bent the medial portion into the acetabulum. The Colonna procedure, as used in non-unions of the femoral neck (Colonna 1935), was also occasionally used in necrosis of the femoral head. However, according to the reports by Hallel and Salvati (1977) and Freeland et al. (1980), the hips tended to end up in a dislocated or subluxated position when this method was used in femoral head necrosis following septic arthritis. Stiffness and ankylosis were also mentioned as problems with this procedure. Primary subtrochanteric osteotomy combined with transposition of the trochanteric apophysis was first described by Weissman (1967). In his case, the procedure resulted in an ankylosed hip. Furthermore, Stetson et al. (1968) reported one case operated with the same procedure. There was practically no motion in this hip 11 years after operation. Hunka et al. (1982) reported satisfactory results in 3 of 5 children 8–12 years after trochanteric arthroplasty with primary osteotomy. Monte et al. (1984) described good results in 6 of 16 children at an average of 4 years after secondary osteotomy. Finally, Dobbs et al. (2003) presented results after 5 such operations. 1 hip that was observed for 16 years had nearly autofused, whereas one hip observed for 18 years had a flattened femoral head and moderate osteoarthritic changes. In my study of 4 hips, satisfactory results with development of a spherical femoral head well contained within a nearly concentric and well-located acetabulum were found in 2 hips, 24 and 16 years after surgery. In both of these cases, the acetabulum gave a nearly normal cover of the femoral head after the operation. In the 2 cases where substantial osteoar-

Limp d Slight Marked Slight Marked

marked limp.

thritic changes developed, a severely dysplastic acetabulum gave insufficient support to the reconstructed femoral head. This demonstrates that satisfactory long-term results depend on adequate acetabular cover after operation. It should be noticed, however, that the method may give reasonably good function of the hip during childhood and adolescence, and to some degree prevent development of severe shortening of the limb, even if ideal cover of the femoral head is not obtained. Some tendency to limp was seen in all patients, as in other studies of patients operated with this procedure. This is probably partly caused by the reduced lever arm of the gluteal muscles due to the development of coxa valga. The growth plate of the greater trochanter only contributes to increased length of the new femoral neck and femoral shaft after arthroplasty, and not to increased width of the proximal femur at the level of the insertion of the gluteal muscles. Although the growth plate contributes to increased length of the femur, this contribution is too small to fully compensate for the loss of the original proximal femoral growth plate. Thus, secondary operative procedures to avoid unacceptable leg-length discrepancies might still be indicated after this type of arthroplasty, as done in 2 of my patients. However, the leg length discrepancies will usually be less in children in whom trochanter arthroplasty is performed than in untreated patients, since the method prevents severe proximal migration of the femur. From a biological point of view, it is interesting that an apophysis has the potential to develop into an epiphyseal-like structure when traction is replaced by pressure and friction. The osseous nucleus grows and is remodeled into a spherical joint-end which adapts to the opposite joint surface. Under normal conditions, all the cartilage of an apophysis is completely ossified at the end of the growth period. After transposition to the joint, this process seems to be modified leaving the most superficial layers of the new articulating surface unossified, like what was seen when apophyseal cartilage was transplanted to joint defects (Benum 1974). There is no agreement about the most suitable age for doing a trochanter arthroplasty. Probably the apophysis has the


68

Acta Orthopaedica 2011; 82 (1): 64–68

greatest potential for remodeling before the ossification has become too advanced. On the other hand, it seems likely that the remodeling process is greatly influenced by the loading and motion that takes place during gait. Hence, the best results will probably be obtained if the child has already passed the first learning period of walking. In my child patients, all operations were performed before the age of 7 years. As in other studies, however, the number of hips was too small in this study to be able to draw any firm conclusions about the ideal age for this type of operation. In summary, previous studies have shown some satisfactory results up to 12 years after subtrochanteric osteotomy and transposition of the apophysis of the greater trochanter into the acetabulum in children with total femoral head necrosis due to septic arthritis. My findings indicate that satisfactory results may even last for a considerably longer period. A nearly perfect acetabular cover of the transposed apophysis is a prerequisite for a long-lasting favorable result. However, the procedure might be beneficial even if a perfect cover is not achieved. In such a case, the operation might reduce the tendency to shortening of the limb and give a reasonable function until osteoarthritis develops. Furthermore, it appears likely that a total hip replacement might give a better result in such a hip compared to an untreated hip with total dislocation and severe shortening

Benum P. Tansposition of apophyseal cartilage to osteochondral defects of joints. An experimental study in dogs. Acta Orthop Scand (Suppl 156) 1974. Colonna P C. A new type of reconstruction operation for old ununited fracture of the neck of the femur. J Bone Joint Surg 1935; 17: 110-22. Dal Monte A, Capelli O, Donzelli O, Libri L, Soncini G. Trochanteroplasty in the treatment of infantile septic arthritis of the hip. Ital J Orthop Traumat 1984; 10/2: 145-52. Dobbs M B, Sheridan J J, Gordon J E, Corley C L, Szymanski D A, Schoenecker P L. Septic arthritis of the hip in infancy: Long term follow up. J Pediatr Orthop 2003; 23: 162-8. Freeland A E, Sullivan D J, Westin G W. Greater trochanteric hip arthroplasty in children with loss of the femoral head. J Bone Joint Surg (Am) 1980; 62: 1351-61. Hallel T, Salvati E A. Septic arthritis of hip in infancy. End result study. Clin Orthop 1977; (132): 115-28. Hunka L, Said S E, MacKenzie D A, RogalA E J, Cruess R L. Classification and surgical management of the severe sequela of septic hips in children. Clin Orthop 1982; (171): 30-6. L’Episcopo J B. Stabilization of pathological dislocation of the hip in children. J Bone Joint Surg 1936; 18: 737-42. Smith T. On the acute arthritis of infants. St. St. Bartholomew’s Hospital Reports 1874; 10: 189. Stetson J W, DePonte R J, Southwick W O. Acute septic arthritis of the hip in children. Clin Orthop 1968; (56): 105-16. Weissman S L. Transplantation of the trochanteric epiphysis into the acetabulum after septic arthritis of the hip. Report of a case. J Bone Joint Surg (Am) 1967; 49: 1647-51.

The author is most grateful to Kjell Harbo of the Department of Orthopaedics, Stavanger University Hospital for admission and partial follow-up of 2 of the children in this study.


Acta Orthopaedica 2011; 82 (1): 69–75

69

Range of motion and strength after surgery for brachial plexus birth palsy 107 patients followed for 12-year Mikko O Kirjavainen1, Yrjänä Nietosvaara2, Sanna M Rautakorpi2, Ville M Remes1, Tiina H Pöyhiä3, Ilkka J Helenius4, and Jari I Peltonen2 1Department

of Orthopedics and Traumatology and Hospital for Children and Adolescents, Helsinki University Central Hospital; 2Helsinki Medical Imaging Center, University of Helsinki, Helsinki; 3Department of Orthopedics and Traumatology, Turku University Central Hospital, Turku, 4Department of Orthopedics and Traumatology, Helsinki University Central Hospital, Finland Correspondence: mikko.kirjavainen@hus.fi Submitted 09-10-20. Accepted 10-08-30

Background There is little information about the range of motion (ROM) and strength of the affected upper limbs of patients with permanent brachial plexus birth palsy. Patients and methods 107 patients who had brachial plexus surgery in Finland between 1971 and 1998 were investigated in this population-based, cross-sectional, 12-year follow-up study. During the follow-up, 59 patients underwent secondary procedures. ROM and isometric strength of the shoulders, elbows, wrists, and thumbs were measured. Ratios for ROM and strength between the affected and unaffected sides were calculated. Results 61 patients (57%) had no active shoulder external rotation (median 0° (-75–90)). Median active abduction was 90° (1–170). Shoulder external rotation strength of the affected side was diminished (median ratio 28% (0–83)). Active elbow extension deficiency was recorded in 82 patients (median 25° (5–80)). Elbow flexion strength of the affected side was uniformly impaired (median ratio 43% (0–79)). Median active extension of the wrist was 55° (-70–90). The median ratio of grip strength for the affected side vs. the unaffected side was 68% (0–121). Patients with total injury had poorer ROM and strength than those with C5–6 injury. Incongruity of the radiohumeral joint and avulsion were associated with poor strength values. Interpretation ROM and strength of affected upper limbs of patients with surgically treated brachial plexus birth palsy were reduced. Patients with avulsion injuries and/or consequent joint deformities fared worst. 

Most brachial plexus birth palsy (BPBP) patients (66–92%) recover spontaneously (Michelow et al. 1994, Noetzel et al. 2001, Hoeksma et al. 2004, Pöyhiä et al. 2010). Indications for brachial plexus surgery vary (Kay 1998, Rust 2000, O’Brian et al. 2006). However, severe total injury or upper-middle

plexus injury with no signs of spontaneous recovery within 3–6 months is widely accepted as an indication for early operative treatment (Gilbert et al. 1988, Clarke and Curtis 1995, Strömbeck et al. 2000, Smith et al. 2004). The severity of neural involvement in BPBP varies from transient neurapraxia to avulsion-type root injuries. Upper plexus (C5-6) injury affects shoulder and elbow function. Furthermore, wrist function is affected to varying degrees in more extensive injuries that involve the upper and middle plexus (C5-7). In total injuries (C5-T1), finger function is also compromised (Bager 1997, Sheburn et al. 1997). Muscle weakness and joint contractures of the affected upper limb are common in patients with permanent BPBP (Zancolli 1981, Waters et al. 1998, Hoeksma et al. 2003, Kirjavainen et al. 2007, Strömbeck et al. 2007). Muscle imbalance in BPBP patients can lead to soft tissue contractures and eventually to joint deformities (Pollock and Reed 1989, Waters et al. 1998, Nath et al. 2007). There is a negative correlation between degree of osseous deformity of the glenohumeral joint and shoulder range of motion (ROM) (Hoeksma et al. 2003, Kozin 2004). In this population-based, cross–sectional, long-term followup study, we assessed ROM and isometric maximal muscle strength of the upper limbs of surgically treated BPBP patients.

Patients and methods We performed a population-based, cross-sectional followup study on data from 107 patients treated for BPBP; these data were obtained from the hospital discharge register of the National Research and Development Center for Welfare and Health (STAKES) for 1971 through 1998. 1,706 patients had received treatment for BPBP, of which 124 had had brachial

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2010.539499


70

Acta Orthopaedica 2011; 82 (1): 69–75

Table 1. Types of secondary procedures performed during the follow-up period Extent of injury

Median age Median Soft-tissue in years at number of shoulder Humeral operation (range) procedures (range) procedure osteotomy

Soft-tissue forearm procedure

Bone forearm procedure

Hand procedure

C5–6 (n = 51) C5–7 (n = 31) Total (n = 25)

7 (1–13) 5 (2–11) 6 (5–11)

1 (1–3) 2 (1–4) 2 (1–3)

15 16 6

6 3 3

4 5 11

0 2 3

0 5 5

Sum

37

12

20

5

10

plexus surgery over the 27 year period. Of these patients, 3 were excluded from the study: 2 due to hemiparesis that resulted from cerebral palsy and 1 because of mental retardation. The remaining 121 patients were invited to participate in this study and 107 (64 females) agreed. All physical examinations were performed in the Hospital for Children and Adolescents, Helsinki University Central Hospital, between September 2002 and October 2003. This study was part of a larger investigation carried out at Helsinki University Central Hospital. Patients The median ages of the patients at the time of brachial plexus surgery were 2.9 (0.4–11.0) months, 2.2 (0.9–6.8) months, and 2.4 (0.8–4.9) months for the C5-6 injury group, the C5-7 injury group, and the total injury group (see below). Direct neuroraphy was the most common procedure (n = 65). Lower plexus reconstruction was performed on one fifth of the patients who had total palsy. For the remaining patients with total palsy, either upper or middle plexus reconstruction was performed. Physical examinations were carried out at a median of 12 (5–32) years after the primary surgery. The right upper limb was affected in 56 of the patients and the left upper limb in 51. 3 patients had temporary contralateral BPBP. The extent of the injury was estimated by physical examination, electromyography (EMG), cervical myelography, and intraoperative findings. Of the 107 patients, 51 had a C5-6 palsy, 31 had a C5-7 palsy, and 25 had a total brachial plexus palsy. None of the patients had isolated lower nerve root injury. Root avulsion was diagnosed in 34 patients by intraoperative findings or by myelography. The numbers of patients who had avulsion involvement were: 17 for 1 root, 11 for 2 roots, 3 for 3 roots, 1 for 4 roots, and 1 for 5 roots. No other illness causing contracture of the joints or muscle atrophy of the upper limb was noted in these patients. During the follow-up period, secondary reconstructive procedures were performed on the shoulders or forearms, or hands of 59 patients. 29 patients underwent more than one secondary operation (Table 1). Physical examination Passive and active ROM measurements of the affected and the unaffected upper limbs were taken by two unblinded phys-

The Metitur Good Strength device for measurement of elbow flexion strength.

iotherapists. They used a standard goniometer, which was marked in 1-degree divisions. External rotation in adduction, abduction, flexion, and extension of the shoulder, elbow flexion, elbow extension, forearm pronation and supination, extension, flexion, and ulnar and radial deviation of the wrist were measured. Radial abduction of the thumb was also recorded. Maximum isometric muscle strength was measured using a Jamar dynamometer (Asimow Engineering Co., Los Angeles, CA) and also a Good Strength Metitur adjustable dynamometer chair (Metitur Oy, Jyväskylä, Finland) (Figure). The validity and reliability of these two devices have been reported (Mathiowetz et al. 1984, Hovi et al. 1993, Era et al.1994, Bellace et al. 2000, Curb et al. 2006). Shoulder (external and internal rotation strength in adduction), elbow (flexion and extension strength), and grip strength measurements were obtained from patients in a sitting position on a custom-made dynamometer chair. Grip strength was measured using the Jamar dynamometer and the Good Strength dynamometer chair for comparison. 3 measurements each for the affected and the unaffected sides were taken and the best attempt was taken as the result. Ratios of the strength of the unaffected hand to that of the affected hand were calculated and values of


Acta Orthopaedica 2011; 82 (1): 69–75

less than 89% were judged to be subnormal in further analysis (Petersen et al. 1989). The strength ratio was also evaluated in relation to the patient’s age. Radiographs of the shoulder (anteroposterior AP and axillary) and elbow (AP and lateral) were taken in 95 patients. 2 patients were pregnant and 10 patients declined. Congruity of the glenohumeral joint was classified according to an arbitrary scale: (1) congruent, (2) posteriorly subluxated, (3) posteriorly dislocated, (4) anteriorly subluxated, and (5) anteriorly dislocated. Similarly, congruency of the radiohumeral joint was estimated by evaluating the radiographs as follows: (1) congruent, (2) subluxated, and (3) dislocated. Some of these radiographic data have been reported in an earlier study (Kirjavainen et al. 2007). Statistics Values are expressed as median (range). We used MannWhitney U-test to examine the differences for continuous variables and chi-squared test for categorical variables. Correlations were analyzed using the Spearman rank correlation test. 2-tailed p-values of less than 0.05 were considered statistically significant. Strength and ROM measurements were correlated with the type of injury (avulsion vs. no avulsion), secondary procedures, and the congruity of glenohumeral and radiohumeral joint by univariate analysis. Odds ratios (ORs) and their 95% confidence intervals (CIs) were used to analyze the presence of good clinical outcome for patient injury groups using multivariate analysis logistic regression models (NCSS statistical software version 6.0; NCSS, Kaysville, UT). Age at the time of the operation, sex, type of injury (avulsion vs. no avulsion), and congruity of the glenohumeral and radiohumeral joints were considered independent covariates in these analyses. The type of surgery (neurolysis, neuroraphy, grafting, or neurotization) was closely associated with the type (avulsion vs. no avulsion) or with the extent of the neural injury. Thus, both of these covariates could not be included in the same logistic regression model. Ethics The Ethics Research Board of Helsinki University Central Hospital approved this study (no. 79/E7/2001). Permission to contact the subjects was received from different hospital districts throughout Finland. A signed informed consent form was obtained from the patients, or from their parents if the patients were minors.

Results The shoulder Range of motion. 61 of the 107 patients had limited active external rotation of the affected shoulder (≤ 0°). The extent of the injury did not have an effect on the incidence or absence of external rotation (C5-6: 31/51; C5-7: 12/31; and C5-T1:

71

Table 2. Median active range of motion (ROM) of the affected upper limb and the ratio (%) between affected and unaffected limb according to the extent of injury

C5-6 ROM Ratio

C5-7 ROM Ratio

Total ROM Ratio

Shoulder External rotation in adduction –8 0 7 11 –15 Abduction 125 75 68 36 50 Extension 31 43 10 17 11 Flexion 126 75 82 50 51 Elbow and forearm Extension deficiency 20 94 25 90 35 Flexion 135 94 128 90 128 Pronation 80 79 67 58 54 Supination 50 71 30 42 0 Wrist Extension 65 86 44 60 35 Flexion 65 100 60 92 44 Radial deviation 24 100 0 0 –2 Ulnar deviation 40 100 41 78 20

0 28 20 44 44 96 54 0 47 75 0 44

18/25). Only those patients with C5-6 injury had a median active abduction ROM value of greater than 90° (median 125°). Moreover, 29 patients had no active extension of the shoulder. The shoulder flexion was the best preserved function. The ratios of passive ROM values between the affected and the unaffected sides were well preserved, apart from the external rotation of the shoulder (41% for C5-6, 56% for C5-7, and 43% for C5-T1). The ROM and ratios of active motion for the extent of injury are given in Table 2. Strength. The median ratio of 44% (0–147) for internal rotation strength of the shoulder indicated that this was better preserved than that of external rotation, at 28% (0–105). However, only one patient of the C5-7 group and 6 patients (12%) of the C5-6 group reached or surpassed the predefined cut-off ratio of > 89% for normal internal rotation strength measurements. Similarly, normal values for external rotation strength were obtained for 1 of the 25 patients with total injury and 2 of the 51 patients with C5-6 injury. None of the patients in the C5-7 injury group had normal external rotation strength (Table 3). The elbow and forearm Range of motion. Lack of full extension of the elbow was the most common restriction. 89 patients had an active extension deficiency (median 25° (5–80)). All except 1 patient with total injury had an active extension deficiency. The incidences of active extension deficiency for the C5-6 and C5-7 injury groups were 78% (40/51) and 81% (25/31). Similarly, a median of 15° (5–55) for passive extension deficiency ROM was found in 82 patients. No differences in passive motion of elbow flexion between the groups were detected, whereas for active motion only a 7% difference between the affected and unaffected sides was found. 12 of the 25 patients with total


72

Acta Orthopaedica 2011; 82 (1): 69–75

Table 3. Isometric strength (N) measurements using the Metitur Good Strength device. Grip strength was also measured by the Jamar dynamometer (kg). Max. value of 3 attempts

Shoulder external rotation C5-6 C5-7 total Shoulder internal rotation C5-6 C5-7 total Elbow flexion C5-6 C5-7 total Elbow extension C5-6 C5-7 total Grip C5-6 C5-7 total Grip (Jamar) C5-6 C5-7 total

Median total values Affected Unaffected side side 18 25 16 9 43 53 36 42 59 60 66 46 55 69 44 34 123 151 159 38 8 11 7 1

70 66 72 74 103 99 118 118 151 133 157 158 114 101 129 116 228 207 241 226 21 20 23 20

Ratio % Range 28 38 23 12 44 60 37 35 43 51 42 36 49 80 40 27 68 88 64 21 35 61 30 6

0–100 0–81 0–105 18–147 9–96 0–62 5–90 0–88 0–62 5–185 4–80 0–81 41–121 32–91 0–86 19–111 3–80 0–33

palsy and only 5 of the 51 with C5-6 injury had lack of active supination (< 0°) (p < 0.001). The incidence for the lack of supination for the C5-7 group was 7/31. The ratio between the affected and unaffected sides for forearm pronation indicated the function was well preserved. The ROM and ratios of active motion for the extent of injury are given in Table 2. Strength. The median ratios of strength between the affected and unaffected sides were 43% (0–90) for elbow flexion and 49% (0–185) for elbow extension (Table 3). None of the patients in the C5-7 injury group or the total injury group exceeded the normal predefined ratio (> 89%) for elbow flexion or for elbow extension strength. 1 patient in the C5-6 group had a normal ratio for elbow flexion and 17 patients had normal ratios for elbow extension strength. The wrist and hand Range of motion. Extension and radial deviation of the wrist were impaired most often. 16 patients had no active extension (< 0°) and 46 had no active radial deviation (< 0°). The incidences for the C5-6 group were 3/51 for no active extension and 6/51 for no active radial deviation. The corresponding values for the C5-7 injury group were 5/31 and 21/31. For patients with total palsy, 8/25 had no active extension and 19/25 had no active radial deviation. The incidences for patients with no active radial abduction of the thumb (< 0°) were: 7/25 for the total palsy group, 2/31 for the C5-7 group,

and 1/51 for the C5-6 group. The ratios of passive ROM between the affected and unaffected sides were well preserved except for those of radial deviation of the wrist (C5-6: 100%; C5-7: 40%, and C5-T1: 67%). The ROM and ratios of active motion for each injury group are given in Table 2. Strength. Higher grip strength measurements were obtained from the patients by the Good Strength device than by the Jamar dynamometer. The ratio of strength between the affected and the unaffected sides was 68% (0–121) with the Good Strength device and 35% (0–11) with the Jamar dynamometer (p < 0.001). 9 of the 107 patients could not hold and compress the Jamar dynamometer; their results were therefore recorded as zero. However, 5 of the same 9 individuals could produce a positive result using the Good Strength device (Table 3). Age was not related to the strength ratios obtained for any patient group; nor was it associated with any other strength measurement. Radiographic findings Radiographic findings, which were used for further statistical analyses, have been reported in our earlier publication (Kirjavainen et al. 2007), which was based on a slightly larger patient population (n = 112). The congruity of the glenohumeral joint was: 80 congruent, 9 posteriorly subluxated, 3 posteriorly dislocated, and 3 anteriorly subluxated. Anterior dislocation was not noted. Moreover, the radiohumeral joint was congruent in 77 patients, subluxated in 14, and dislocated in 4. Predictors of upper extremity function Univariate analyses showed that there was a negative correlation between avulsion-type injury and grip strength for the Jamar (p = 0.007) and for the Good Strength (p < 0.001) device measurements, but not with any other strength measurements or for ROM. Similarly, the incongruency of the glenohumeral joint was associated with a lack of extension of the shoulder joint (p = 0.02), but not with other ROM or strength measurements. There was an inverse correlation between secondary procedures and elbow flexion strength (p = 0.02) but this did not apply to other strength measurements or to ROM. The following factors were adjusted in the analyses: age at surgery, sex, type of injury (avulsion vs. no avulsion) and congruency of the glenohumeral joint. After these adjustments were made, the avulsion type of injury was predictive of poor shoulder internal rotation strength (OR = 0.4, CI: 0.14–0.89; p = 0.03), impaired grip strength according to Jamar measurement (OR = 0.3, CI: 0.11–0.76; p = 0.01) and poor elbow flexion strength (OR = 0.3, CI: 0.12–0.79; p = 0.02) (Tables 4 and 5). Similarly, an incongruent radiohumeral joint was associated with a poor outcome for grip strength as measured by the Jamar device (OR = 0.2, CI: 0.035–0.79; p = 0.02) and the Good Strength device (OR = 0.2, CI: 0.073–0.70; p = 0.01), and elbow extension strength (OR = 0.3, CI: 0.093–0.95; p = 0.04) (Table 5).


Acta Orthopaedica 2011; 82 (1): 69–75

73

Our patients were relatively young at the time of plexus surgery, which—along with the small size of the child—possibly facilitated end-to-endtype repair by neuroraphy. Direct neuroraphy was used more often in patients with C5-6 injury than in those with total injury. In contrast, grafting and Shoulder external rotation Shoulder internal rotation (ratio>0.89) (ratio >0.89) neurotization were used more often in patients Characteristic n OR 95% CI p-value OR 95% CI p-value with total injury. Root avulsion had been sustained in almost one-third of the patients in our study. Age (months) <3 67 1 1 The strength of our study is the long-term > 3 40 1.8 0.66–4.8 0.3 2.2 0.89–5.40 0.09 follow-up. On the other hand, the 27-year-long Sex patient inclusion period, which lasted from 1971 F 65 1 1 M 42 1.1 0.41–2.9 0.9 1.8 0.79–4.3 0.2 to 1998, can be considered a weakness—as indiType of injury cations, timing, and surgical techniques for prino avulsion 69 1 1 mary plexus reconstruction and secondary proceavulsion 38 0.4 0.13–1.2 0.09 0.4 0.14–0.89 0.03 Congruency of dures have changed over such a long time. The glenohumeral joint long-term results of patients with or without secyes 80 1 1 ondary operations cannot be compared with each no 15 1.3 0.36–4.7 0.7 2.2 0.69–6.7 0.2 other, due to the cross-sectional study design. Furthermore, caution should be exercised when the results of this nationwide investigation are compared with results from modern tertiary institutions specialDiscussion ized in the treatment of BPBP. Patients with permanent BPBP have various degrees of functional impairment of their affected upper extremities, due to a The shoulder limited ROM and a reduction in muscle power (Hardy 1981, A strong association between shoulder contracture and osseBoome and Kaye 1988, Hoeksma et al. 2003). Secondary joint ous deformity was reported by Hoeksma et al. (2003). Half of procedures have been reported to improve function in patients their patients had a shoulder contracture of > 10°. We found with permanent palsy (Pearl et al. 2006, Waters et al. 2006, a correlation between the incongruency of the glenohumeral 2008). Comparing the results obtained from different institu- joint and the poor extension of the shoulder. A lack active tions is difficult, as all the commonly used scales for ROM and external rotation was the most common shoulder impairment strength of the affected upper limb in BPBP are descriptive: and was found in more than half of our patients, which supMallet classification, Toronto test score, and active movement ports the findings by Hoeksma et al. (2003). Pagnotta et al. scale (Bae et al. 2003). (2004) reported a series of 203 patients who had undergone Table 4. Odds ratios (ORs) and their 95% confidence intervals (CIs) for shoulder rotation strength (Good Strength; Metitur). Age at plexus surgery, sex, type of injury, and congruency of glenohumeral joint were considered as independent covariates in logistic regression models. A strength ratio of > 0.89 between the affected and the unaffected side was considered normal

Table 5. Odds ratios (ORs) and their 95% confidence intervals (CIs) for isometric elbow and hand strength measured by the Good Strength and Jamar dynamometers. Age at plexus surgery, sex, type of injury, and congruency of radiohumeral joint were considered as independent covariates in the logistic regression models. A strength ratio of > 0.89 between the affected and the unaffected side was considered normal ´ Characteristic

n

Grip strength (Jamar) (> 0.89) OR 95% CI p-value

Grip strength (GS) (> 0.89) OR 95% CI p-value

Elbow flexion (GS) (> 0.89) OR 95% CI p-value

Elbow extension (GS) (> 0.89) OR 95% CI p-value

Age (months) <3 67 1 1 1 1 > 3 40 1.2 0.48–3.0 0.7 0.5 0.56–4.7 0.3 0.6 0.25–1.6 0.3 0.8 0.34–2.02 Sex F 65 1 1 1 1 M 42 1 0.70–2.4 0.9 1 0.93–8.0 0.07 0.6 0.26–1.5 0.3 1.9 0.81–4.5 Type of injury no avulsion 69 1 1 1 1 avulsion 38 0.3 0.11–0.76 0.01 0.4 0.16–1.04 0.06 0.3 0.12–0.79 0.01 0.5 0.23–1.3 Congruency of radiohumeral joint yes 89 1 1 1 1 no 18 0.2 0.035–0.79 0.02 0.2 0.073–0.70 0.01 0.4 0.14–1.7 0.3 0.3 0.093–0.95

0.7 0.1 0.2

0.04


74

latissimus dorsi transfer for restoring shoulder abduction and external rotation with 10 to 15-years of follow-up. These authors obtained satisfactory results with active external rotation. They also noted that this improvement was sustained after the procedure for all patient groups—except the C5-6 group, in which they found that a slight deterioration had occurred between 10 and 15 years of follow-up. On the other hand, they found progressive deterioration of active abduction after 6–10 years of surgery. They also reported a mean abduction of 105° for the C5-6 group and 75° for the C5-7 group at 15 years of follow–up, as compared to only 68° for the total injury group at 10 years of follow-up. The median values for the corresponding injury groups in our more heterogeneous population were 125°, 68°, and 50° at 12 years of follow-up. The results of a 30-year-long follow-up study by Kirkos et al. (2005) also suggest that short-term satisfactory results with soft tissue procedures to restore active abduction and external rotation of the shoulder were not maintained. According to our results, patients with secondary reconstructive procedures generally had poorer ROM values than those patients who had been treated with primary plexus surgery. This finding can probably be partly explained by those patients with plexus surgery alone having sustained a less severe BPBP than patients who subsequently underwent secondary operations. To our knowledge, there have been no previous reports in the literature on shoulder strength measurements in operatively treated patients with BPBP. In the present study, both external and internal rotation strength were markedly diminished. Patients with an avulsion-type injury had poorer strength in internal rotation than those patients without root avulsions, which seems logical since patients with root avulsions have potentially fewer motor neurons. The elbow and forearm A Swedish study reported that 90% of all patients with permanent BPBP had an elbow extension deficit (Strömbeck et al. 2007). The authors also found that elbow extension deteriorated substantially during the 5-year-long follow-up period, whereas shoulder and forearm rotation remained unchanged. Restriction of rotation of the active forearm was reported by Sibinski et al. (2007) who found that most of their patients had subnormal (< 80° ROM) active pronation and active supination. We found extension deficit in 89% of all our patients and a lack of active supination in 56% of them, which is in accordance with earlier studies. Both elbow flexion and extension strength were markedly reduced in our patients (usually with > 50% reduction). Furthermore, we found that none of the patients with C5-7 or total injury reached the normal ratio for these strength measurements. The wrist and hand Changes in hand and wrist movements have been reported in patients with either C5-7 or total injury (Strömbeck et al. 2007). We found limited wrist and hand ROM mainly in

Acta Orthopaedica 2011; 82 (1): 69–75

the C5-7 group and the total injury group. However, we also found that 5 patients (10%) with C5-6 palsy had limited hand and wrist ROM. This finding indicates that such patients may have had a more extensive plexus lesion than that in an upper root injury, which reflects the difficulty in reliably determining the extent of the root injury. Recently, Strömbeck et al. (2007) found that grip strength as measured by the Jamar dynamometer was subnormal (ratio 80%) for all 12 patients with total injury. Their finding is in line with that in our series, where we found that none of the patients with total injury attained or exceeded the normal predefined ratio (> 89%) as measured by either the Jamar dynamometer or the Good Strength device. However, these authors also reported that 22% of patients with C5-6 injury (9/41) had subnormal grip strength. Their finding is similar to the 34% (Jamar) and 14% (Good Strength) that we found. The differences between our study data and theirs might be explained by the more lenient criteria they used for defining normal grip strength, which perhaps underestimated the numbers of individuals with subnormal grip strength. In our study, grip strength measurement was higher using the Good Strength device than that measured by the Jamar dynamometer. This is probably due to the fixed platform of the Good Strength device, which gives support to the arm. In the present study, ROM and strength of the shoulder and elbow reflect the outcome of BPBP after plexus reconstruction and also secondary surgery, while in every patient’s upper nerve root (responsible for shoulder and elbow) neuroma (5–6) was cut/resected and after that reconstructed. Later on perhaps secundary surgery may have performed. However, the results for hand function mainly reflected the natural history of lower root recovery in total injuries, as most patients with total injury had only upper or middle plexus reconstruction. Median strength of the affected upper limb was less than 50% of that of the unaffected upper limb in every measurement except for grip strength, as measured by the Good Strength device. The poorest results were obtained for external rotation strength of the shoulder. Active external rotation ROM of the shoulder was the most adversely affected parameter, with 0° as the median value. As expected, passive movements were generally better maintained than active movements. Avulsiontype injury and incongruent radiohumeral joint were associated with poor results for ROM and strength measurements.

MK: study concept and design, physical examinations, interpretation of the data, and drafting of the manuscript. YN, JP, and VR: study concept and design, review of the manuscript. SR: assessment of range of motion and strength measurements. IH: statistical analysis. TP: evaluation of radiographs.

This study was supported by the Foundation for Pediatric Research, a Finnishstate study grant, the Finnish Cultural Foundation, and the Päivikki and Sakari Sohlberg Foundation.


Acta Orthopaedica 2011; 82 (1): 69–75

No competing interests declared.

Bae D S, Waters P M, Zurakowski D. Reliability of three classification systems measuring active motion in brachial plexus birth palsy. J Bone Joint Surg (Am) 2003; 85: 1733-8. Bager B. Perinatally acquired brachial plexus palsy – a persisting challenge. Acta Peaditr 1997; 86: 1214-9. Bellace J V, Healy D, Besser M P, Byron T, Hohman L. Validity of the Dexter evaluation system’s Jamar dynamometer attachment for assessment of the hand grip strength in a normal population. J Hand Ther 2000; 13: 46-51. Boome R S, Kaye J C. Obstetric traction injuries of the brachial plexus. Natural history, indication for surgical repair and results. J Bone Joint Surg (Br) 1988; 70: 571–6. Clarke H M, Curtis C G. An approach to obstetrical brachial plexus injuries. Hand Clin 1995; 11: 563–80. Curb J D, Ceria-Ulep C D, Rodriguez B L, Grove J, Guralnik J, Willcox B J, Donlon T A, Masaki K H, Chen R. Performance-based measures of physical function for high-function populations. J Am Geriatr Soc 2006; 54: 737-42. Era P, Rantanen T, Avlund K, Gause-Nilsson I, Heikkinen E, Schroll M, Steen B, Suominen H. Maximal isometric strength and antropometry among 75-year-old men and women in three Nordic localities. Scand J Med Sci Sports 1994; 4: 26-31. Gilbert A, Razaboni R, Amar-Khodja S. Indications and results of brachial plexus surgery in obstetrical palsy. Orthop Clin North Am 1988; 19: 91-105. Hardy A E. Birth injuries of the brachial plexus. Incidence and prognosis. J Bone Joint Surg (Br) 1981; 63: 98–101. Hoeksma A F, ter Steeg A M, Dijkstra P, Nelissen R G, Beelen A, de Jong B A. Shoulder contracture and osseous deformity in obstetrical brachial plexus injury. J Bone Joint Surg (Am) 2003; 85: 316-22. Hoeksma A F, ter Steeg A M, Nelissen R G, van Ouwerkerk W J, Lankhorst G J, de Jong B A. Neurological recovery in obstetrical brachial plexus injuries: an historical cohort study. Dev Med Child Neurol 2004; 46: 76–83. Hovi L, Era P, Rautonen J, Siimes M A. Impaired muscle strength in female adolescents and young adults surviving leukemia in childhood. Cancer 1993; 72: 276-81. Kay S P. Obstetrical brachial palsy. Br J Plast Surg 1998; 51: 43-50. Kirjavainen M, Remes V, Peltonen J, Kinnunen P, Pöyhiä T, Telaranta T, Alanen M, Helenius I, Nietosvaara Y. Long-term results of obstetrical brachial plexus surgery – a population- based clinical and radiological study. J Bone Joint Surg (Am) 2007; 89: 18-26. Kirkos J M, Kyrkos M J, Kapetanos G A, Haritidis J H. Brachial plexus palsy secondary to birth injuries. J Bone Joint Surg (Br) 2005; 87: 231-5. Kozin S H. Correlation between external rotation of the glenohumeral joint and deformity after brachial plexus birth palsy. J Pediatr Orthop. 2004; 24: 189-93. Mathiowetz V, Weber K, Volland G, Kashman N. Reliability and validity of grip and pinch strength evaluations. J Hand Surg (Am) 1984; 9: 222-6.

75

Michelow B J, Clarke H M, Curtis C G, Zuker R M, Seifu Y, Andrews D F. The natural history of obstetrical brachial plexus palsy. Plast Reconstr Surg 1994; 93: 675–80. Nath R K, Lyons A B, Melcher S E, Paizi M. Surgical correction of the medial rotation contracture in obstetric brachial plexus palsy. J Bone Joint Surg (Br) 2007; 89: 1638-44. Noetzel M J, Park T S, Robinson S, Kaufman B. Prospective study of recovery following neonatal brachial plexus injury. J Child Neurol 2001; 16: 488–92. O’Brien D F, Park T S, Noetzel M J, Weatherly T. Management of birth brachial plexus palsy. Childs Nerv Syst 2006; 22: 103-12. Pagnotta A, Haerle M, Gilbert A. Long-term results on abduction and external rotation of the shoulder after latissimus dorsi transfer for sequelae of obstetric palsy. Clin Orthop 2004; (426): 199-205. Pearl M L, Edgerton B W, Kazimiroff P A, Burchette R J, Wong K. Arthroscopic release and latissimus dorsi transfer for shoulder internal rotation contractures and glenohumeral deformity secondary to brachial plexus birth palsy. Bone Joint Surg (Am) 2006; 88: 564-74. Petersen P, Petrick M, Connor H, Conklin D. Grip strength and hand dominance: challenging the 10% rule. Am J Occup Ther 1989; 43: 444-7. Pollock A N, Reed M H. Shoulder deformities from obstetrical brachial plexus paralysis. Skeletal Radiol 1989; 18: 295-7. Pöyhiä T H, Lamminen A E, Peltonen J I, Kirjavainen M O, Willamo P J, Nietosvaara Y A. Brachial plexus birth injury: US screening for glenohumeral joint instability. Radiology. 2010; 254: 253-60. Rust R S. Congenital brachial plexus palsy: Where have we been and where are we now? Semin Pediatr Neurol 2000; 7: 58–63. Sherburn E W, Kaplan S S, Kaufman B A, Noetzel M J, Park T S. Outcome of surgically treated birth-related brachial plexus injuries in twenty cases. Pediatr Neurosurg 1997; 27: 19–27. Sibinski M, Sherlock D A, Hems T E, Sharma H. Forearm rotational profile in obstetric brachial plexus injury. J Shoulder Elbow Surg. 2007; 16: 784-7. Smith N C, Rowan P, Benson L J, Ezaki M, Carter P R. Neonatal brachial plexus palsy. Outcome of absent biceps function at three months of age. J Bone Joint Surg (Am) 2004; 86: 2163-70. Strömbeck C, Krumlinde-Sundholm L, Forssberg H. Functional outcome at 5 years in children with obstetrical brachial plexus palsy with and without microsurgical reconstruction. Dev Med Child Neurol 2000; 42: 148–57. Strömbeck C, Krumlinde-Sundholm L, Remahl S, Sejersen T. Long-term follow-up of children with obstetric brachial plexus palsy I: functional aspects. Dev Med Child Neurol 2007; 49: 198-203. Waters P M, Smith G R, Jaramillo D. Glenohumeral deformity secondary to brachial plexus birth palsy. J Bone Joint Surg (Am) 1998; 80: 668–77. Waters P M, Bae D S. The effect of derotational humeral osteotomy on global shoulder function in brachial plexus birth palsy. J Bone Joint Surg (Am) 2006; 88: 1035-42. Waters P M, Bae D S. The early effects of tendon transfers and open capsulorrhaphy on glenohumeral deformity in brachial plexus birth palsy. J Bone Joint Surg (Am) 2008; 90: 2171-9. Zancolli E A. Classification and management of the shoulder in birth palsy. Orthop Clin North (Am) 1981;12: 433-57.


76

Acta Orthopaedica 2011; 82 (1): 76â&#x20AC;&#x201C;81

Wrist function recovers more rapidly after volar locked plating than after external fixation but the outcomes are similar after 1 year A randomized study of 63 patients with a dorsally displaced fracture of the distal radius Maria K T Wilcke, Hassan Abbaszadegan, and Per Y Adolphson Division of Orthopaedics, Karolinska Institutet, Department of Clinical Sciences, Danderyd Hospital, Stockholm, Sweden Correspondence: maria.wilcke@ds.se Submitted 10-06-23. Accepted 10-10-11

Background and purpose Promising results have been reported after volar locked plating of unstable dorsally displaced distal radius fractures. We investigated whether volar locked plating results in better patient-perceived, objective functional and radiographic outcomes compared to the less invasive external fixation. Patients and methods 63 patients under 70 years of age, with an unstable extra-articular or non-comminuted intra-articular dorsally displaced distal radius fracture, were randomized to volar locked plating (n = 33) or bridging external fixation. Patientperceived outcome was assessed with the Disability of the Arm, Shoulder, and Hand (DASH) questionnaire and the Patient-Rated Wrist Evaluation (PRWE) questionnaire. Results At 3 and 6 months, the volar plate group had better DASH and PRWE scores but at 12 months the scores were similar. Objective function, measured as grip strength and range of movement, was superior in the volar plate group but the differences diminished and were small at 12 months. Axial length and volar tilt were retained slightly better in the volar plate group. Interpretation Volar plate fixation is more advantageous than external fixation, in the early rehabilitation period. î Ž

The risk of a poor outcome after a fracture of the distal radius increases with malunion (Grewal and MacDermid 2007) and in highly unstable fractures, operative fixation is required to maintain a satisfactory anatomical position. Closed reduction and bridging external fixation rely on ligamentotaxis to reduce and keep the fracture in alignment. It has been used for unstable distal radius fractures for several decades. External fixation requires 5â&#x20AC;&#x201C;6 weeks of immobilization, and some fracture redisplacement often occurs after the fixation device has been removed (Dicpinigaitis et al. 2004, Handoll et al. 2007). In later years, there has been a strong trend towards

open reduction and internal fixation with volar locked plating in the management of unstable, dorsally displaced, fractures of the distal radius. Volar locked plating facilitates an anatomical reduction of the fracture, it stabilizes the fracture during the entire healing process, and it allows early wrist mobilization. Good results in terms of patient-rated outcome scores, objective function, and radiographic outcome have been reported both in younger and older patients (Orbay and Fernandez 2002, 2004, Kamano et al. 2005, Chung et al. 2006, Oshiege et al. 2007, Jupiter et al. 2009). Several studies have compared dorsal plating (Grewal et al. 2005, Kateros et al. 2010), fragment-specific systems (Abramo et al. 2009), or a mixture of dorsal and volar plating techniques (Kreder et al. 2005, Leung et al. 2008) with external fixation, but there is no substantial evidence to support the use of internal fixation instead of external fixation (Margaliot et al. 2005). Few studies have compared volar locked plating with external fixation, and there is still insufficient evidence regarding which gives the best outcome (Wright et al. 2005, Egol et al. 2008, Rizzo et al. 2008, Wei et al. 2009). We have carried out a randomized comparison of open reduction with volar locked plating and closed reduction with bridging external fixation for unstable dorsally displaced extra-articular and non-comminuted intra-articular fractures of the distal radius. Our hypothesis was that volar locked plating would result in better patient-perceived, objective functional, and radiographic outcome after 12 months than external fixation.

Patients and methods Eligibility criteria The inclusion criteria were: an acute unilateral dorsally dis-

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2011.552781


Acta Orthopaedica 2011; 82 (1): 76–81

placed fracture of the distal radius (extra-articular AO class A fractures and C1 fractures with only one intra-articular fracture line (Fernandez and Wolfe 2005)), an axial shortening of ≥ 4 mm, or a dorsal angulation of ≥ 20°) in patients aged 20–70 years with no previous fracture of either wrist. Patients were excluded if they had a concurrent fracture of the upper extremities, were medicated with warfarin, were unable to co-operate with follow-up (dementia, substance abuse, psychiatric illness, language problems), had an open fracture or a fracture that was not amenable to both methods, i.e. a distal fragment that was too small (less than 10 mm of intact volar cortex) or too comminuted (AO class C2 and C3 fractures) to allow plate fixation. Sample size The Disabilities of the Arm, Shoulder, and Hand (DASH) score (Hudak et al. 1996) at 12 months was the primary outcome variable used to determine statistical power. The level of significance was set at p < 0.05. Power analysis showed that a 90% power to discover a difference of 10 points (SD 12) in DASH score would require 62 patients. The standard deviation of 12 points was based on data from a group of similar patients in a previous study (Wilcke et al. 2007). Ethics Ethical permission was given by the ethics board of the local university (D.nr. 2005/4:8). Written informed consent was obtained from all patients. The trial was registered at www. ClinicalTrials.gov (NCT00989222). Recruitment and randomization During the recruitment period (January 1, 2006 to May 9, 2008), 82 patients met the inclusion criteria (Figure 1). Primary reduction was performed and a temporary plaster cast was applied at the emergency ward. Patients were informed of the study and asked to participate during the visit at the emergency ward or by telephone on the following day. 63 patients were included and randomized by a sealed envelope procedure to either open reduction and internal fixation with a volar locking plate or closed reduction and bridging external fixation. The patients were informed of the randomized method immediately. Randomization was conducted in blocks of 20 with age stratification (over or under 50 years of age). Interventions The patients were planned for day surgery based on the availability of theaters according to normal routines, and the operations were performed after a mean of 4 (1–9) days. The attending trauma surgeon performed the operation; when needed, he/she was assisted by a more experienced colleague. All patients were operated on an out-clinic basis in plexus anesthesia and began finger exercises immediately postoperatively. In plate-fixated patients, a volar flexor carpi radialis approach was used and a volar locked plate with 4 optional distal locked

77

Assessed for eligibility n=121 (20–70 years, dorsally displaced distal radius fracture, dorsal angulation of >20° or axial shortening of >4 mm, attended the emergency ward during the inclusion period)

Excluded due to given criteria or not meeting inclusion criteria n=39: – Previous wrist fracture: 17 – Bilateral fracture: 1 – Concurrent other fracture of upper extremity: 1 – Fracture not amenable to both methods: 12 – Psychiatric illness: 2 – Heavy alcohol abuse: 1 – Open fracture: 4 – Warfarin treatment: 1

Missed to inclusion: 8

Rejected participation: 11 63 patients included and randomized

Volar plate n=33

External fixation n=30

3 months 31 completed DASH, PRWE and physical examination

3 months 27 completed DASH, PRWE and physical examination

6 months 31 completed DASH, PRWE and physical examination

6 months 28 completed DASH, PRWE and physical examination

12 months All completed DASH and PRWE, 32 physical examination; All radiological examination

12 months All completed DASH, PRWE, physical examination and radiological examination

Figure 1. Flow chart of the patients.

screws was applied (Königsee; Swemac, Sweden). Cancellous bone graft was not used. The patients were given a dorsal below-elbow plaster cast for 10–12 days, after which active wrist mobilization began supervised by a physiotherapist. In patients with external fixation, a Hoffmann device (Stryker, NJ) was applied using 2 pins in the second metacarpal and 2 pins in the radius proximal to the fracture. A closed reduction was performed aided by fluoroscopy, and the fixation device was locked. Supplementary K-wires were not used routinely but at the surgeon’s discretion. In 1 case, a radial pin was used. The external fixation was removed after 5 weeks, in the outpatient clinic without anesthesia, and wrist mobilization began supervised by a physiotherapist. Outcome assessment Clinical follow-up was conducted at 10 days, 5 weeks, and 3, 6, and 12 months postoperatively. The validated Swedish versions of the DASH score (Hudak et al. 1996, Atroshi et al. 2000) and the Patient-Rated Wrist Evaluation (PRWE) outcome questionnaire (MacDermid et al. 1998, Wilcke et al. 2009) were used to assess the patient-rated functional result.


78

Acta Orthopaedica 2011; 82 (1): 76–81

Table 1. Clinical data

Volar plating

Mean age (range) in years Female, n (%) AO fracture type: A/C, n (%) Dominant hand injury, n (%) Retired, n (%)

Figure 2. Radiological measurements: axial shortening (a), dorsal angulation (b), and radial angulation (c).

Both questionnaires yield a score from 0 to 100, where higher scores represent more disability. The questionnaires were completed at baseline, and at 3, 6, and 12 months. Scores are presented with the baseline scores subtracted. Objective function was assessed with grip strength and range of movement by a physiotherapist at 3, 6, and 12 months. Grip strength was measured with a dynamometer (Grippit; AB Detector, Sweden). The range of movement was measured using a standard goniometer. Both wrists were assessed and the uninjured wrist was used as a control. Grip strength and range of movement are expressed as percentages of those of the uninjured wrist. Grip strength was adjusted by 10% for the non-dominant side (Bechtol 1954). Group assignment could not be blinded to the assessors due to different scars. Standard anteroposterior and lateral radiographs were taken directly after surgery, after 5 weeks, and after 12 months. The fractures were classified by one of the authors (MW) according to the AO classification system. Axial shortening, dorsal angulation, and radial angulation were also measured (Figure 2). Radiographs of the uninjured wrist were used as a control. Axial shortening is expressed in mm, and dorsal and radial angulation in degrees. Values are presented as the difference between the injured wrist and the uninjured wrist. Statistics The groups were compared by the Chi-square test for sex, fracture classification, hand dominance, and retirement; Wilcoxon rank sum test for DASH and PRWE scores; and unpaired Student t-test for objective physical and radiographic measurements. The statistical analyses were performed with the statistical package JMP (SAS Institute Inc., NC).

Results 33 patients were randomized to volar plating and 30 patients to external fixation. The groups were similar in age, sex, and fracture type, and most fractures were extra-articular (Table 1). External fixation was performed by less experienced surgeons to a greater extent than volar plating.

55 (20–69) 25 (76) 26 (79)/7 (21) 15 (45) 9 (28)

External fixation 56 (21–69) 23 (77) 22 (73)/8 (27) 14 (47) 9 (30)

Table 2. Patient-perceived results measured by the DASH and PRWE scores

Volar plating

External fixation

p-valuea

DASH 3 months 9 (6–12) 27 (20–33) < 0.001 6 months 6 (3–9) 14 (9–19) 0.008 12 months 7 (4–11) 11 (6–16) 0.1 PRWE 3 months 14 (8–20) 31 (23–39) < 0.001 6 months 9 (5–14) 17 (11–22) 0.02 12 months 11 (6–16) 15 (9–21) 0.3     Values are presented as points (95% CI) corrected for baseline values. Higher scores indicate more disability. a Wilcoxon rank-sum test.

Patient-perceived outcome The patients in the volar plate group reported statistically significantly better DASH and PRWE scores than the patients in the external fixation group at 3 and 6 months. However, after 12 months, the differences had diminished and were no longer significant (Table 2). Objective function Grip strength and range of movement were superior in the volar plate group but the differences diminished with time (Table 3). Radiographic results At final follow-up, the patients in the volar plate group had less axial shortening and dorsal angulation of the distal radius than the patients in the external fixation group (Table 4). Complications In the external fixation group, 1 patient was reoperated with a supplementary volar plate within a week due to an unacceptable fracture position postoperatively. At the 12-month evaluation, a corrective osteotomy was planned in 1 patient, due to a painful malunion. 4 patients suffered from pin tract infections and in 1 of these, pin loosening occurred with malunion as a consequence. 1 patient suffered from a mild complex regional pain syndrome and 1 patient reported a disturbing skin adhesion after a Hoffmann pin. In 4 patients, a light sensory deficit was noted, corresponding to a superficial radial nerve branch.


Acta Orthopaedica 2011; 82 (1): 76–81

79

Table 3. Objective physical measurements (95% CI) expressed as a percentage of the uninjured side

Volar plating

External fixation p-value a

3 months Grip strength 72 (64–80) 46 (37–55) < 0.001 Extension 84 (78–90) 59 (49–69) < 0.001 Flexion 81 (77–85) 71 (65–77) 0.009 Ulnar deviation 89 (81–97) 74 (63–86) 0.04 Radial deviation 89 (80–98) 75 (55–95) 0.2 Supination 95 (91–98) 76 (68–85) < 0.001 Pronation 98 (96–100) 89 (84–94) < 0.001 6 months Grip strength 89 (83–95) 72 (65–78) < 0.001 Extension 92 (86–97) 77 (68–84) 0.001 Flexion 88 (84–91) 83 (78–88) 0.1 Ulnar deviation 99 (92–106) 91 (79–103) 0.2 Radial deviation 103 (92–115) 89 (70–107) 0.2 Supination 98 (95–100) 88 (84–93) < 0.001 Pronation 100 (100–100) 95 (92–99) 0.005 12 months Grip strength 94 (86–102) 85 (79–91) 0.08 Extension 94 (90–98) 85 (77–93) 0.04 Flexion 89 (86–92) 83 (77–89) 0.08 Ulnar deviation 96 (87–105) 83 (72–93) 0.05 Radial deviation 97 (88–106) 89 (77–100) 0.3 Supination 99 (97–100) 89 (81–98) 0.02 Pronation 99 (98–100) 92 (86–99) 0.04 a Student’s

t-test

Table 4. Mean (95% CI) radiological measurements presented as the difference from the uninjured wrist

Volar plating

External fixation

p-value a

At injury Axial shortening (mm) 4 (3-5) 4 (3-5) 0.9 Dorsal angulation (degrees) 41 (39–44) 43 (40–46) 0.4 Radial angulation (degrees) 11 (8–12) 9 (7–11) 0.5 Post reduction Axial shortening (mm) -2 (-2 to -1) 0 (0–1) < 0.001 Dorsal angulation (degrees) 6 (5–8) 9 (6–12) 0.1 Radial angulation (degrees) 2 (1–3) 1 (0–2) 0.4 5 weeks postoperatively Axial shortening (mm) 0 (-1–0) 1 (1–2) < 0.001 Dorsal angulation (degrees) 7 (5–9) 11 (7–14) 0.07 Radial angulation (degrees) 2 (1–4) 2 (1–4) 0.9 12 months postoperatively Axial shortening (mm) 0 (0–1) 2 (1–3) < 0.001 Dorsal angulation (degrees) 7 (5–9) 11 (7–15) 0.05 Radial angulation (degrees) 2 (1–4) 3 (1–5) 0.5 a

Student’s t-test.

In the volar plate group, 1 patient was operated on with a median nerve decompression and plate extraction after 6 months due to carpal tunnel syndrome. A transient affection of the median nerve was reported by 3 other patients postoperatively. 1 patient suffered from a rupture of the flexor pollicis longus tendon. This patient started treatment with high-dose corticosteroids during the follow-up period due to a newly dis-

covered cancer. At the 12-month follow-up, 2 patients were planned for plate removal due to irritation from the distal radial edge of the plate and 1 of these patients also had complaints of extensor tenosynovitis superficial to the proximal cortical screws.

Discussion We found that volar locked plating was advantageous in the early rehabilitation period, compared to bridging external fixation, but at 1 year the outcomes were similar. Egol et al. (2008) compared volar locked plating and bridging external fixation with supplementary Kirschner wires in 77 patients, and found an improved range of movement early after volar plating but at 1 year the range of movement was similar—as were grip strength and DASH score at any time. These results are in accordance with ours, although we found better grip strength and DASH scores in the plating group at 3 and 6 months. One possible explanation for this difference could be that our plate-fixated patients had a more active early mobilization regime. Another reason for the disparities in DASH scores could be that the variances (standard deviations) were larger in the study by Egol et al. (2008), which might have affected their statistical power negatively. Our external-fixated patients had a somewhat poorer radiographic result than the patients in the study by Egol et al. (2008). This can be explained by the fact that, with 1 exception, our patients did not receive supplementary Kirschner wires since this was not the routine at our department. The use of augmenting pins enhances fragment stability (Dunning et al. 1999) and one could question whether we have compared an insufficient external fixation method with volar plating, and perhaps this would explain why our early findings favored volar plating to a greater extent. However, we do not believe that the small differences in the radiographic results between the groups could account for the differences in grip strength and DASH score in the early rehabilitation period. Rather, we would expect that a worse radiographic outcome would be more likely to affect the result after 1 year. It is interesting however, that even a potentially suboptimal external fixation method gives a similar outcome to that of volar plating after 1 year. By omitting the pins, the risk of radial nerve and skin complications is reduced. Rozental et al. (2009) randomized 45 patients similar to ours to open reduction and volar plating versus indirect reduction and percutaneous pinning or external fixation. They found that the radiographic outcome was similar between the groups but DASH score, grip strength, and range of movement were superior in the plate-fixated group up to 12 weeks after the injury. The outcomes were similar after 1 year, however. In our material, the plate-fixated patients were also better at 1 year regarding range of movement but the absolute differences were small and probably without any clinical relevance.


80

Marcheix et al. (2010) randomized 103 patients over the age of 50 years with unstable extra- and intra-articular fractures (including comminuted C2 and C3 fractures) to volar locked plating or “mixed pinning”. At 3 and 6 months, the plated patients had better objective functional results and reported better DASH scores, in accordance with our findings. The 1-year results were not reported. Wei et al. (2009) compared external fixation (22 patients) with locked radial (12 patients) or volar plating (12 patients) and found that volar-plated patients had better DASH scores in the first 3 months. At 6 and 12 months, however, the DASH scores were similar between the groups. Our interpretation is that the more rapid recovery of patientperceived and objective wrist function in patients treated with volar locked plating, compared to external fixation, found by us and by the other authors mentioned above, is due to the earlier active wrist mobilization postoperatively that is allowed by the volar locked plating technique. A limitation of our study is that it may have been underpowered. It can be argued that we might have found a statistically significant difference in DASH and PRWE scores at 12 months if the patient groups had been larger. However, a retrospective power analysis showed that 70–110 patients in each group would be required to make the differences significant. Considering the small absolute differences at 12 months and the findings of other investigators, as discussed above, we believe that our results are valid. We used a volar plate with a simple construction, which limited the type of fractures that could be included to extra-articular and non-comminuted AO type C1 fractures. Considering the later risk of osteoarthritis, open reduction and volar plating may also be advantageous in the long term for more complex intra-articular fractures compared to external fixation, since the method allows specific fragment reduction. However, the clinical relevance of radiographic degenerative changes in the wrist is being debated (Downing and Karantana 2008) and there is still little evidence in the literature that volar plating is better than external fixation for intra-articular fractures (Wright et al. 2005, Rizzo et al. 2008). As the patient-rated function appears to be similar between volar plating and external fixation at 1 year, it is important to consider the complications of the 2 methods. Our study was performed when the volar locked plating method was new to our department; therefore, the good results and relatively few reported complications in the plate group—despite the simple and rather bulky design of the plate—in combination with the learning curve for our surgeons, imply that volar plating is a forgiving method. However, high complication rates have been reported by other authors (Rozental and Blazar 2006, Casaletto et al. 2009, Knight et al. 2010) after volar plating, with fracture collapse and rupture of the extensor pollicis longus and flexor pollicis longus tendons. We did not observe any fracture collapse, perhaps because we did not include the most osteoporotic elderly patients or comminuted intra-articular

Acta Orthopaedica 2011; 82 (1): 76–81

fractures. 1 patient in our material suffered from a flexor pollicis tendon rupture. If the follow-up period had been longer, we might have seen further cases since this complication can present several years after the fracture (Casaletto et al. 2009). In the external fixation group, the serious complications were 2 cases of fracture collapse with malunion. Concerning the complications, we do not believe that our study has demonstrated that either treatment has a clear advantage. Randomization was not performed in the operating theater, with the possibility of surgeon bias. To a large extent, external fixation was performed by less experienced surgeons than was volar plating. This may have contributed to the slightly poorer radiographic result in this group and the 2 cases of fracture collapse. The large number of volar plates that have been implanted in recent years may result in unexpected late complications in the future. What remains to be evaluated is the long-term risk of rupture of the flexor pollicis longus tendon after volar plating. It is a serious complication that requires tendon transfer with a long rehabilitation period. Furthermore, one can speculate that remaining volar plates might potentially lead to more complicated fracture patterns or other problems after a subsequent wrist trauma, especially since patients will age and their osteoporosis will increase. The advantage of external fixation is that complications can be expected to occur within the first year after the injury. Thus, a relevant question is whether the earlier recovery of wrist function after volar plating might be gained at the expense of serious late complications. The clinical implication of our results is that for unstable extra-articular and simple intra-articular fractures of the distal radius in a patient under the age of 70 years, volar plating should be considered when rapid recovery of wrist function is important. However, overlooking the slower return of wrist function, external fixation is still an effective, inexpensive, and less invasive method. MW designed and conducted the study, analyzed and interpreted the data, and wrote the manuscript. HA and PA participated in designing the study and revised the manuscript. PA also participated in acquisition of data.

No competing interests declared.

Abramo A, Kopylov P, Geijer M, Tägil M. Open reduction and internal fixation compared to closed reduction and external fixation in distal radial fractures: a randomized study of 50 patients. Acta Orthop 2009; 80 (4): 478-85. Atroshi I, Gummesson C, Andersson B, Dahlgren E, Johansson A. The disabilities of the arm, shoulder and hand (DASH) outcome questionnaire. Reliability and validity of the Swedish version evaluated in 176 patients. Acta Orthop Scand 2000; 71 (6): 613-8. Bechtol C O. Grip test: The use of a dynamometer with adjustable handle spacings. J Bone Joint Surg (Am) 1954; 36: 820-4. Casaletto J A, Machin D, Leung R, Brown D J. Flexor pollicis longus tendon ruptures after palmar plate fixation of fractures of the distal radius. J Hand Surg (Eur) 2009; 34 (4): 471-4.


Acta Orthopaedica 2011; 82 (1): 76–81

Chung K C, Watt A J, Kotsis S V, Margaliot Z, Haase S C, Kim H M. Treatment of unstable distal radius fractures with the volar locking plating system. J Bone Joint Surg (Am) 2006; 88 (12): 2687-94. Dicpinigaitis P, Wolinsky P, Hiebert R, Egol K, Koval K, Tejwani N. Can external fixation maintain reduction after distal radius fractures? J Trauma 2004; 57 (4): 845-50. Downing N D, Karantana A. A revolution in the management of fractures of the distal radius? J Bone Joint Surg (Br) 2008; 90 (10): 1271-5. Dunning C E,Lindsay C S, Bicknell R T, Patterson S D, Johnson J A, King G J W. Supplemental pinning improves the stability of external fixation in distal radius fractures during simulated finger and forearm motion. J Hand Surg (Am) 1999; 24: 992-1000. Egol K, Walsh M, Tejwani N, McLaurin T, Wynn C, Paksima N. Bridging external fixation and supplementary Kirschner-wire fixation versus volar locked plating for unstable fractures of the distal radius: a randomised, prospective trial. J Bone Joint Surg (Br) 2008; 90 (9): 1214-21. Fernandez D L, Wolfe S W. Distal radius fractures. In: Green’s operative hand surgery. (Ed. Green D P, Hotchkiss R, Pederson W, Wolfe S). Elsevier Churchill Livingstone. Philadelphia; 2005; 1: 645-710. Grewal R, MacDermid J C. The risk of adverse outcome in extra-articular distal radius fractures is increased with malalignment in patients of all ages but mitigated in older patients. J Hand Surg (Am) 2007; 32: 962-70. Grewal R, Perey B, Wilmink M, Stothers K. A randomized prospective study on the treatment of intra-articular distal radius fractures: open reduction and internal fixation with dorsal plating versus mini open reduction, percutaneous fixation, and external fixation. J Hand Surg (Am) 2005; 30 (4): 764-72. Handoll H H, Huntley J S, Madhok R. External fixation versus conservative treatment for distal radial fractures in adults. Cochrane Database of Systematic Reviews 2007, Issue 3. Hudak P L, Amadio P C, Bombardier C. Development of an upper extremity outcome measure: the DASH (Disabilities of the arm, shoulder and hand). The Upper Extremity Collaborative Group (UECG). Am J Int Med 1996; 29 (6): 602-8. Jupiter J B, Marent-Huber M, LCP Study Group. Operative management of distal radial fractures with 2.4-millimeter locking plates. A multicenter prospective case series. J Bone Joint Surg (Am) 2009; 91 (1): 55-65. Kamano M, Koshimune M, Toyama M, Kazuki K. Palmar plating for Colles’ fractures - a preliminary report. J Hand Surg (Am) 2005; 30: 750-5. Kateros K, Macheras G, Galanakos S P, Sofianos I, Papakostas I, Papadakis SA. External fixation versus ”pi” plate for distal radius fractures. J Trauma 2010; 68 (1): 166-72. Knight D, Hajducka C, Will E, McQueen M. Locked volar plating for unstable distal radial fractures: clinical and radiological outcomes. Injury 2010; 41 (2): 184-9. Kreder H J, Hanel D P, Agel J, McKee M, Schemitsch E H, Trumble T E, Stephen D. Indirect reduction and percutaneous fixation versus open reduction and internal fixation for displaced intra-articular fractures of the distal radius: a randomised, controlled trial. J Bone Joint Surg (Br) 2005; 87 (6): 829-36.

81

Leung F, Tu Y K, Chew W Y, Chow S P. Comparison of external and percutaneous pin fixation with plate fixation for intra-articular distal radial fractures. A randomized study. J Bone Joint Surg (Am) 2008; 90 (1): 16-22. MacDermid J C, Turgeon T, Richards R S, Beadle M, Roth J H. Patient rating of wrist pain and disability: a reliable and valid measurement tool. J Orthop Trauma 1998; 12 (8): 577-86. Marcheix P S, Dotzis A, Benkö P E, Siegler J, Arnaud JP, Charissoux JL. Extension fractures of the distal radius in patients older than 50: a prospective and randomized study comparing fixation using mixed pins or a palmar fixed-angle plate. J Hand Surg (Eur) 2010; Mar 17. [Epub ahead of print] Margaliot Z, Haase S C, Kotsis S V, Kim H M, Chung K C. A meta-analysis of outcomes of external fixation versus plate osteosynthesis for unstable distal radius fractures. J Hand Surg (Am) 2005; 30 (6): 1185-99. Orbay J L, Fernandez D L. Volar fixation for dorsally displaced fractures of the distal radius; a preliminary report. J Hand Surg (Am) 2002; 27: 205-15. Orbay J L, Fernandez D L. Volar fixed-angle plate fixation for unstable distal fractures in the elderly patient. J Hand Surg (Am) 2004; 29: 96-102. Oshiege T, Sakai A, Zenke Y, Moritani S, Nakamura T. A comparative study of clinical and radiological outcomes of dorsally angulated, unstable distal radius fractures in elderly patients: intrafocal pinning versus volar locking plating. J Hand Surg (Am) 2007; 32: 1385-92. Rizzo M, Katt B A, Carothers J T. Comparison of locked volar plating versus pinning and external fixation in the treatment of unstable intraarticular distal radius fractures. Hand (N Y) 2008; 3 (2): 111-7. Rozental T D, Blazar P E. Functional outcome and complications after volar plating for dorsally displaced, unstable fractures of the distal radius. J Hand Surg (Am) 2006; 31 (3): 359-65. Rozental T D, Blazar P E, Franko O I, Chacko A T, Earp B E, Day C S. Functional outcomes for unstable distal radial fractures treated with open reduction and internal fixation or closed reduction and percutaneous fixation. A prospective randomized trial. J Bone Joint Surg (Am) 2009; 91 (8): 1837-46. Wei D H, Raizman N M, Bottino C J, Jobin C M, Strauch R J, Rosenwasser M P. Unstable distal radius fractures treated with external fixation, a radial column plate, or a volar plate. J Bone Joint Surg (Am) 2009; 91 (7): 1568-77. Wilcke M K, Abbaszadegan H, Adolphson P Y. Patient-perceived Outcome After Displaced Distal Radius Fractures. A Comparison Between Radiological Parameters, Objective Physical Variables, and the DASH Score. J Hand Ther 2007; 20 (4): 290-8. Wilcke M T, Abbaszadegan H, Adolphson P Y. Evaluation of a Swedish version of the patient-rated wrist evaluation outcome questionnaire: good responsiveness, validity, and reliability in 99 patients recovering from a fracture of the distal radius. Scand J Plast Reconstr Surg Hand Surg 2009; 43 (2): 94-101. Wright T W, Horodyski M B, Smith D W. Functional outcome of unstable distal radius fractures: ORIF with a volar fixed-angle tine plate versus external fixation. J Hand Surg (Am) 2005; 30: 289-99.


82

Acta Orthopaedica 2011; 82 (1): 82–89

Health-related quality of life (EQ-5D) before and after orthopedic surgery Karl-Åke Jansson1 and Fredrik Granath2 1Orthopedics Section, Department of Molecular Medicine and Surgery, Karolinska Institutet, at Karolinska University Hospital; 2Clinical Epidemiology Unit, Department of Medicine, Karolinska Institutet, at Karolinska University Hospital, Stockholm, Sweden Correspondence: karl-ake.jansson@karolinska.se Submitted 10-03-04. Accepted 10-09-06

Background and purpose Population data on mortality and life expectancy are generally available for most countries. However, no longitudinal data based on the health-related quality of life outcome from the EQ-5D instrument have been reported for orthopedic patients. We assessed the effect of orthopedic surgery as measured by EQ-5D. Methods We analyzed EQ-5D data from 2,444 patients who were operated at the Department of Orthopedic Surgery at Karolinska University Hospital, 2001–2005. We also made a comparison between results from this cohort and those from a Swedish EQ-5D population survey. Results The mean EQ-5D index score improved from 0.54 to 0.72. Hip and knee arthroplasty, operations related to previous surgery, trauma-related procedures, and rheumatoid arthritis surgeries had preoperative EQ-5D index scores of 0.48 to 0.52. All of these groups showed substantial improvement in scores (0.63 to 0.80). Patients with tumors or diseases of the elbow/hand showed higher preoperative scores (0.66 to 0.77), which were similar postoperatively. In most patients, the EQ-5D index score improved but did not reach the level reported for an age- and sex-matched population sample (mean difference = 0.11). Interpretation Our results can be used as part of the preoperative patient information to increase the level of patient awareness and cooperation, and to facilitate rehabilitation. In future it will be possible—but not easy—to use the EQ-5D instrument as a complementary consideration in clinical priority assessment. 

Musculoskeletal conditions are the leading cause of severe long-term pain and disability in the world, affecting hundreds of millions of people (Woofle and Pfleger 2003). They are also the main cause of disability in older age groups, and rank among the top 10 causes of disability-adjusted life-years (DALY) in Europe (WHO 2006). This has been recognized by the World Health Organization, endorsing the Bone and Joint Decade (2000–2010) (Woolfe 2000). Osteoarthritis is

the fifth greatest cause of years lived with disability (YLD) in high-income countries (The Word Bank 2006). During the year 2007, 114,000 patients underwent a primary hip or knee joint replacement operation in the UK (England and Wales National Joint Registry 2009). Prevalence data from Sweden for the same year show that 1 in 15 elderly women had a knee arthroplasty (Swedish Knee Arthroplasty Register 2009). One of the major goals of the Bone and Joint Decade has been to reduce the burden and cost of musculoskeletal disorders for individuals, healthcare providers, and society in general. At the end of the decade, it is now appropriate to reflect on the outcome of orthopedic surgery. Improvement in health-related quality of life (HRQOL) is one of the most important goals of orthopedic surgery (Ethgen et al. 2006, Jansson et al. 2009). There are several HRQOL instruments available. Among these, the generic instruments can be used for diverse patient groups independently of the underlying disease or disability. Generic instruments include, for example, the EQ-5D (EuroQol), the SF-6D (derived from RAND-36/SF-36), the HUI (Health Utilities Index Mark II/Mark III), and the AQoL (Assessment of Quality of Life) (Kopec and Willison 2003). The SF-36 instrument is most commonly used. Most studies have concentrated on specific orthopedic interventions, and most of them show improved HRQOL after surgery (Towheed and Hochberg 1996). HRQOL has been used to evaluate the effect of surgical procedures (Hoffmann et al. 2006, Akahane et al. 2007). Treatment outcome across various elective orthopedic surgical procedures has been compared (Hansson et al. 2008, Anderson et al. 2009, Osnes-Ringen et al. 2009). Generic tools have also been used for the estimation of orthopedic effectiveness of healthcare (Räsänen et al. 2006). The generic health-related quality of life instrument—EQ-5D—allows both a description of health status along 5 dimensions and the evaluation of health or the estimation of a health summary score: the EQ-5D score on a scale where 0 is death and 1 is full health (Dolan 1997, Brooks et al. 2003). The instru-

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2010.548026


Acta Orthopaedica 2011; 82 (1): 82–89

ment has been included in population surveys in more than 10 countries (Kind et al. 1998, Burström et al. 2001, Scende and Williams 2004). HRQOL and health status measures have often been used as outcomes in clinical trials and studies assessing a variety of orthopedic interventions (Tidermark et al. 2003, Cockerill et al. 2004, Jansson et al. 2005, Löfvendahl et al. 2005, Rivero-Arias et al. 2005, Odenbring et al. 2008, Giannoudis et al. 2009). The EQ-5D is short and easy to use, and shows good responsiveness (Tidermark et al. 2003), i.e. it is capable of capturing clinically important changes. Moreover, it also allows combination of different dimensions of health to form an overall index, the EQ-5D index score, as required for healthcare evaluations and for construction of quality-adjusted life-years (QALYs), a measure frequently used in cost-effectiveness analyses (Gold et al. 1996, Meunning and Gold 2001, Drummond et al. 2005). Population data on mortality and life expectancy are generally available for most countries. However, no longitudinal data based on the inclusion of the HRQOL outcome by the EQ-5D have been reported in a clinical setting of orthopedic patients. We therefore introduced the EQ-5D instrument at our department in order to measure all patients selected for elective orthopedic operations. The aim of this study was to preoperatively evaluate the HRQOL in our cohort regardless of other co-morbidity factors and also to make a comparison between this cohort and a Swedish EQ-5D population survey. In addition, we wanted to assess the postoperative outcome by the EQ-5D instrument in order to have output data to explore the potential of EQ-5D for medical priority and health economy calculations. We report data from 2,444 patients.

Patients and methods Study population Between January 2001 and May 2005, 4,715 elective orthopedic operations were performed at the Department of Orthopedic Surgery, Karolinska University Hospital. We included 4,011 patients during this period, all of whom completed the EQ-5D questionnaire. Acute operations were not included. The enrollment of patients was done at the ward, and informed consent was given by all patients. At baseline, i.e. on the day before surgery, the first EQ-5D questionnaire was completed by the patient at the ward. The 12-month EQ-5D questionnaire was sent once to the patients by mail 11 months after surgery once, with no reminders. To be included in the 12-month follow-up, patients had to have answered the EQ-5D questionnaire within 3 months. 2,444 patients completed the 12-month EQ-5D questionnaire within 15 months postoperatively. We performed a drop-out analysis of the 1,567 patients who did not answer the 12-month follow-up questionnaire. Age, sex, type of surgery, and preoperative EQ-5D data were scrutinized.

83

We divided the cohort into 15 groups according to anatomical region and type of surgery. We also compared EQ-5D results for patients older than 20 years of age with those from a Swedish population survey involving 3,069 individuals (Burström et al. 2001, 2003). The study design was approved by the ethics committee of Karolinska Institutet (no. 03-631). Outcomes: the EQ-5D measure Health-related quality of life data were obtained from the EQ-5D, a self-administered patient questionnaire (EuroQol Group 1990, Brooks 1996, Dolan et al. 1996). The EQ-5D respondents classify their own health status into 5 dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression with 3 levels of severity (no problems, moderate problems, or severe problems). Dolan et al. (1996) used the time trade-off (TTO) method to rate these different states of health in a large UK population (UK EQ-5D index tariff). As there is no Swedish TTO tariff for EQ-5D health states, and since the only Swedish population survey to assess the EQ-5D used the UK tariff, we used the preference scores generated from the UK population when calculating the EQ-5D index scores for our study population. The patients completed the Swedish-translated questionnaire (EQ-5D 2009). By design, this descriptive system is able to identify 243 unique health states. An index score can be assigned to each of these health states to indicate its value or desirability from the point of view of the general public. Scores in the UK EQ-5D value set range from –0.594 for the worst possible health state to 1.0 for a perfect state of health, with 0 on the scale representing the state of being dead. Negative scores suggest that the corresponding health states are considered worse than being dead. Normally, the EQ-5D questionnaire needs 1 to 3 min for self-completion. Statistics The EQ-5D index scores are reported as mean (SD). Age and sex standardized EQ-5D index scores at baseline (preoperatively and at 12 month follow-up (postoperatively) was calculated as the difference between observed scores and the age- (10 year age-groups) and sex-specific mean scores in the population survey. These preoperative and (12-month) postoperative EQ-5D index standardized scores are reported as mean (SD). The changes in EQ-5D index score from baseline (preoperatively) and 12 months (postoperatively) were calculated and a paired t-test was used to test whether the change from baseline was equal to 0. We also analyzed the fraction of patients (by number and percentage) whose EQ-5D index score changed from baseline to 12 months (improved or deterioriated by > 0.1). Responders and non-responders at the 12-month follow-up were compared regarding age, sex, type of surgery, and preoperative EQ-5D by chi-square tests for qualitative variables and t-tests for quantitative variables. Even though the distribution of the change from baseline was not normal,


84

Acta Orthopaedica 2011; 82 (1): 82–89

Table 1. Details of the 2,444 patients in the study at baseline (preoperatively), including surgical procedures and anatomical regions Age Females N % mean SD %

EQ-5D index score at baseline mean SD

Standardized score a at baseline mean SD

All patients 2,444 100 56 18 56 0.54 0.35 –0.29 Sex Women 1,359 55.6 59 18 – 0.50 0.37 –0.36 Men 1,085 44.4 50 18 – 0.59 0.33 –0.26 Op. procedure Hip arthroplasty 370 15.1 69 11 55 0.49 0.34 –0.31 Knee arthroplasty 365 14.9 67 12 61 0.51 0.33 –0.29 Complications 326 13.3 53 19 52 0.52 0.37 –0.31 Trauma 287 11.7 46 19 48 0.52 0.36 –0.34 Knee 210 8.6 43 17 45 0.65 0.30 –0.21 Benign tumor 173 7.1 43 17 58 0.77 0.28 –0.09 Rheumatoid arthritis 159 6.5 59 13 84 0.48 0.36 –0.34 Malignant tumor 119 4.9 58 19 53 0.71 0.31 –0.11 Spine 119 4.9 58 16 54 0.30 0.35 –0.53 Hip 95 3.9 54 18 60 0.41 0.36 –0.43 Shoulder 74 3.0 51 16 43 0.62 0.36 –0.23 51 2.1 50 15 63 0.56 0.35 –0.28 Foot Elbow/hand 37 1.5 55 14 59 0.67 0.29 –0.16 Diabetes/infections 26 1.1 62 17 42 0.40 0.37 –0.41 Unknown 33 1.4 45 17 30 0.58 0.35 –0.28

0.35 0.36 0.33 0.34 0.34 0.36 0.35 0.30 0.27 0.36 0.31 0.35 0.35 0.35 0.36 0.29 0.38 0.35

a

Standardized EQ-5D score: difference between the preoperative EQ-5D index score (baseline) and that of the reference population survey (age- and sex-specific mean EQ-5D index scores).

the central limit theorem implies valid inference using the t-test when the sample size is more than about 30, and all but 1 subgroup had larger sample sizes. Since the fraction of responders was different for the different types of surgery, the comparisons with respect to age, sex, and preoperative EQ-5D score were also adjusted for this difference by analysis of variance and logistic regression.

Results The final analysis included 2,444 patients, 57% of whom were women, and the mean age at surgery was 56 (SD 18) years (Table 1). One third of the patients had osteoarthritis and were operated on for hip or knee replacement. 13% of the patients had operations due to complications after previous surgery. 1 in 10 had trauma related to surgery, and 1 in 10 was operated due to knee disorders. 1.4% had an unknown, unidentified operation procedure code. The mean preoperative EQ-5D index score at baseline was 0.54, which is 0.29 units lower than would be expected in a population-based sample of the same age and sex distribution. On average, women had lower scores (0.50) than men (0.59) before surgery (p < 0.001), which remained unchanged after adjustment for age and type of surgery. Age did not affect the preoperative score substantially, with the exception of patients younger than 30 years, who had a higher mean score (by 0.12 units) than the average patients. This age effect could to some

extent be explained by type of surgery. The different surgical procedures showed a wide spectrum of average EQ-5D index scores at baseline (0.30–0.77). When comparing the different surgical procedures for the overall mean EQ-5D index score at baseline, procedures related to benign or malignant tumors and elbow/hand diseases scored statistically significantly higher than average, which is important clinically, while patients with hip and spine procedures scored significantly lower than average. Preoperatively, at baseline, the standardized EQ-5D index score (mean difference between the age- and sex-matched population) was –0.29. All 15 groups of patients had a lower EQ-5D index score than in the matched population (–0.09 to –0.53). At the 12-month follow-up, the mean EQ-5D index score had increased statistically significantly by 0.18 units from baseline to 0.72 (Table 2). The mean EQ-5D in women increased almost to the level of that in men: 0.71 in comparison to 0.73. Patients younger than 30 years had a 12-month mean EQ-5D index score of 0.79 and patients older than 80 years had a 12-month mean score of 0.66. Patients who underwent hip or knee arthroplasty, had complications after surgery, underwent other knee surgery, had trauma-related procedures, had rheumatoid arthritis or who underwent spine, hip, or infection-related surgery showed statistically significant improvements in mean EQ-5D index score (0.09 to 0.31). Patients with benign or malignant tumors or elbow/hand diseases showed no statistically significant changes in EQ-5D index score.


Acta Orthopaedica 2011; 82 (1): 82–89

85

Table 2. Details of the 2,444 patients in the study at 12 months postoperatively

N

%

EQ-5D index score 12-month mean SD

Standardized score a 12-month mean SD

Change from baseline mean p-value b

Change from baseline > 0.1 % n

Change from baseline < –0.1 % n

All patients 2,444 100 0.72 0.30 –0.11 0.30 0.18 <0.0001 49 1193 14 Sex Women 1,359 56 0.71 0.30 –11 0.30 0.21 <0.0001 50 680 12 Men 1,085 44 0.73 0.30 –12 0.31 0.14 <0.0001 47 510 16 Op. procedure Hip arthroplasty 370 15 0.80 0.25 0.00 0.25 0.31 <0.0001 69 254 6 Knee arthroplasty 365 15 0.73 0.27 –0.07 0.28 0.22 <0.0001 54 196 9 Complications 326 13 0.63 0.34 –0.20 0.34 0.11 <0.0001 40 132 18 Trauma 287 12 0.73 0.29 –0.12 0.28 0.21 <0.0001 56 162 12 Knee 210 8.6 0.73 0.29 –0.13 0.29 0.09 <0.0001 35 74 13 Benign tumor 173 7.1 0.80 0.28 –0.06 0.28 0.03 0.09 32 56 18 Rheumatoid arthritis 159 6.5 0.64 0.31 –0.18 0.31 0.16 <0.0001 48 76 11 Malignant tumor 119 4.9 0.71 0.28 –0.11 0.28 –0.00 0.97 24 28 31 Spine 119 4.9 0.61 0.35 –0.21 0.35 0.31 <0.0001 56 67 14 Hip 95 3.9 0.68 0.34 –0.15 0.33 0.27 <0.0001 58 55 13 Shoulder 74 3.0 0.73 0.32 –0.12 0.33 0.11 0.005 46 34 19 Foot 51 2.1 0.69 0.28 –0.15 0.29 0.13 0.02 41 21 18 37 1.5 0.70 0.27 –0.13 0.27 0.03 0.57 24 9 24 Elbow/hand Diabetes/infections 26 1.1 0.66 0.30 –0.15 0.28 0.27 0.002 62 16 8 Unknown 33 1.4 0.69 0.32 –0.17 0.32 0.11 0.06 39 13 24

334 163 174 22 34 60 35 27 31 17 37 17 12 14 9 9 2 8

a

Standardized EQ-5D score: difference between the postoperative EQ-5D index score and that of the reference population survey (age- and sex-specific mean EQ-5D index score). b p-value for testing if the change from baseline is equal to 0.

The standardized EQ-5D index score (mean difference between the age- and gender-matched population) at 12 months of follow-up was –0.11. Hip arthroplasty patients had a mean standardized EQ-5D index score preoperatively of –0.31 but their EQ-5D score improved and reached the level of that of the age- and sex-matched population (standardized EQ-5D index score of 0.00 at 12-month follow-up). Knee arthroplasty, trauma-related operations, other hip and knee surgery, rheumatoid arthritis surgery, surgery after complications, and spine surgery showed major improvements in EQ-5D index score 12 months after operation. However, they did not reach that of the matched population. The mean difference in score from that of the matched population postoperatively varied from –0.07 to –0.21. One year after surgery, half of the patients experienced an improvement of > 0.1 in their EQ-5D index score and a small group (14 %) reported deterioration in their scores of > –0.1. 69% of the hip arthroplasty patients improved by at least 0.1 and only 6% deteriorated in their EQ-5D index score, in contrast to malignant tumor surgery where only 24% improved more than 0.1 and 30% deterioriated by > –0.1. We found that the distribution of the EQ-5D index score was bimodal, and very few individuals scored around the average (Figure 1A and B). Preoperatively, the EQ-5D index score had a bimodal distribution around 0.1 and 0.7. At 12 months, the distribution was still bimodal but most patients now had scores within the range 0.7–1.0. The pre-and postoperative EQ-5D index scores showed 4 major groups of patients (Figure 2). The

first group of patients (26%) had experienced great improvement, while a second group of patients with high preoperative scores (58%) had improved slightly. A third group with low EQ-5D scores preoperatively (12%) were unchanged, and a fourth small group (4%) perceived a decline in their HRQOL. The mean response rate of those who completed the EQ-5D questionnaire at baseline was 85% (Table 3). The response rate varied considerably (59–100%), with the lowest response rates for patients with diabetes/infection (59%) and malignant tumors (68%). In the dropout analyses (Appendix) we found that the responders were more likely to be women, to be older, or to have a low preoperative EQ-5D index score. The response rate also depended on the type of surgery. The responders were on average 5 years older than non-responders (p < 0.001). However, after adjustment for type of surgery this difference was reduced to 3 years, but it was still highly significant (p < 0.001). A comparison between gender and response rate showed that women had a higher response rate (unadjusted comparison, p = 0.0009). Adjustment of the association between gender and response rate for type of surgery reduced the association between gender and response rate (adjusted, p = 0.02). Similarly, after adjusting the difference in mean EQ-5D index score at baseline between responders and non-responders for type of surgery, the difference became less pronounced. On average, the responders had a lower score than non-responders by 0.05 units (p < 0.001). However, after adjustment for type of surgery, this difference was reduced to 0.03 units (p = 0.02).


86

Acta Orthopaedica 2011; 82 (1): 82â&#x20AC;&#x201C;89

Figure 1. A. Bar chart showing preoperative health-related quality of life (EQ-5D) in the orthopedic cohort. Baseline EQ-5D index scores; n = 2,444. B. Bar chart showing postoperative health-related quality of life (EQ-5D) in the orthopedic cohort. 12-month follow-up EQ-5D index scores; n = 2,444.

Table 3. Total number of elective operations in the study at baseline and responder rate at baseline

Elective operations at baseline N

Responders at baseline N

All 4,715 4,011 Op. procedure Hip arthroplasty 754 533 Knee arthroplasty 612 583 Complications 542 524 Trauma 576 514 Knee 383 383 Benign tumor 398 343 Reumatoid arthritis 260 220 Malignant tumor 317 217 Spine 239 186 Hip 194 144 Shoulder 176 143 Foot 85 76 Elbow/hand 59 57 Diabetes/infections 64 38 Unknown 56 50 Figure 2. Graph showing health-related quality of life (EQ-5D) in an orthopedic cohort. Preoperative and 12-month postoperative EQ-5D index scores. The first group of patients (26%) had experienced a great improvement (diamonds); a second group of patients (58.3%) with high preoperative scores were slightly improved (triangles). A third group (11.4%) were unchanged, with low EQ-5D index scores (squares), and a fourth, small group (4.3%) had a decline in their scores (circles).

Discussion We found that most patients who were operated on for orthopedic conditions experienced an improved health-related

% 85 71 95 97 89 100 86 85 68 78 74 81 89 97 59 89

quality of life and that their mean EQ-5D index score increased from 0.54 to 0.72 one year after surgery. As expected, we noted large differences between surgical groups. In contrast to patients with tumor diseases, who scored high with a mean EQ-5D of 0.71, patients scheduled for hip or knee arthroplasty scored considerably lower (0.49 and 0.51, respectively). The indication for surgery is however, totally different in these cases, which must be kept in mind when interpreting these data. Notably, the group of patients treated with hip arthroplasty improved considerably


Acta Orthopaedica 2011; 82 (1): 82–89

and reached the scores of the age- and sex-matched reference population. Interestingly, patients with tumors improved in HRQOL to some extent in spite of their malignant conditions. In a review evaluating changes after hip replacement, the results from all studies were consistent in showing beneficial and often dramatic improvements in HRQOL after elective procedures (Towheed and Hochberg 1996). Another review analyzing Short Form-36 and the Western Ontario and McMaster University osteoarthritis index after hip and knee arthroplasties showed similar results, and both procedures were found to be quite effective in terms of improvement in health-related quality-of-life dimensions (Ethgen et al. 2004). Surgery for lumbar spinal stenosis can give improvement in self-reported quality of life similar to that in hip and knee arthroplasty (Rampersaud 2008). A recently published study demonstrated that spinal surgery can return patients’ HRQL to that of age-matched population norms and yield outcomes similar to those in hip and knee replacement patients (Mokhtar et al. 2010). As other authors have shown (Hansson et al. 2008, Anderson et al. 2009), our study confirms that patients who have undergone spine procedures improve in HRQOL as excellently as the arthroplasty patients do. In a study evaluating patients with inflammatory arthritis using both EQ-5D and SF-6D health assessment questionnaires, the authors recommended the inclusion of at least one preference-based measure in future clinical studies (Harrison et al. 2010). We noticed in our study that inflammatory arthritis (rheumatoid arthritis (RA)) patients had a positive effect on HRCOL but the improvement was less than for patients treated with joint replacement. The reason for this could be that surgery had an effect on pain in the actual joint treated but less improvement in other dimensions of health (OsnesRingen 2009). The minimal important difference (MID) is important for interpreting the impact of score changes, and is also an important measure for power calculations in studies (Walters and Brazier 2003). MID for EQ-5D index score has been reported by Walters and Brazier (2005). For those subjects who reported some changes, a mean EQ-5D index score of 0.07 was found. In our orthopedic cohort, half of the patients had elevated EQ-5D index scores (by more than 0.1) after the operation. 14% had reduced EQ-5D index scores—by more than 0.1—one year after the operation, and one third had less changes (less than 0.1) in their EQ-5D index scores. This first attempt to collect a whole sample of orthopedic conditions makes it possible to perform cost-utility analysis. A QALY is defined as 1 year in full health. Estimation of QALYs requires data on survival and the corresponding health state score, the health status reflecting the HRQOL of the individual, on a scale from 0 (dead) to 1 (full health) (Gold et al. 1996, Meunning and Gold 2001, Drummond et al. 2005). If utilities are multiplied by the amount of time spent in that particular health state, then they become QALYs. QALYs allow for varying times spent in different states by calculat-

87

ing an overall score for each patient. For the studies in which the follow-up is 1 year, the mean change in utility scores over the 1-year period can be directly interpreted as the MID for a QALY. QALYs may have the potential to influence public policy and decisions about resource allocation. If baseline characteristics are controlled for the EQ-5D data, our findings could be used for comparison between hospitals. Comparison between provider units in different hospitals or between consultant specialities within a single institution can provide important information that might be applied for benchmarking or performance management. In addition to clinical priority assessment, criteria in elective orthopedic surgery EQ-5D could be used (NKO 2009). Patients with low scores have low autonomy and should be given high priority (Government 2003, NKO 2009). We found that one third of all patients had a low preoperative HRQOL according to EQ-5D index score and two-thirds of them improved considerably. In future, “soft” HRQOL data (e.g. EQ-5D) might be included in the preoperative evaluation as well as more old-fashioned “hard” data such as radiology. However, to use the instrument in order to make priorities between groups of orthopedic surgical procedures seems to be more controversial, as the patients’ individual EQ-5D index scores differed substantially. The present study has several limitations. It is a prospective follow-up study of patients who underwent surgery, not a prospective randomized controlled trial comparing surgery to nonoperative treatment. However, most of the surgery performed involved accepted interventions (NKO 2009). At baseline, we lost 15% of all patients scheduled for elective surgery. If not all patients are reached at baseline, the patients with the most severe symptoms could be left out and the results would be biased towards patients with less symptoms. However, the numbers of patients included and the response rates were high, apart from for the group of patients with diabetes/infection. For patients who were operated on for diabetes/infection, our results may therefore have been underestimated. The department only mailed 1 follow-up questionnaire to the patients and no reminders, which led to a loss of more than 40% of those initially included in the study. In the dropout analysis, no major difference was found in the preoperative EQ-5D index scores between the responders and the nonresponders. However, the responders tended to be women and to be older, causing our results to be a conservative interpretation. In this analysis, no information on patient co-morbidities or on other types of interfering conditions was collected. Thus, the study can be considered to represent a cross-section of orthopedic patients who undergo surgery at a university hospital. We selected 1 year as a time outcome measurement because it was an easy endpoint. In some groups of patients (e.g. elderly), it might have been better to have had a shorter time frame because many other factors may have impaired the results.


88

The choice of algorithm used to convert self-classification scores can affect the index-based score, as shown in a study that compared UK and US scoring algorithms in patients undergoing percutaneous coronary intervention (PCI) (Shrive et al. 2007). However, while country-specific societal preferences may reduce the scope in comparing HRQoL estimates across studies from different countries, they are more helpful for local decision-making, especially when allocating resources within national healthcare programs. The EQ-5D instrument has potential limitations. It may lack responsiveness to small but clinically important changes in health (Dawson et al. 2001). In the subgroups of patients who were operated on for elbow/hand, shoulder, and foot problems, we noted only minor health changes. The lack of minimal important differences (MID) for this group must be considered. It is also important to add condition-specific instruments in evaluating outcome after orthopedic surgery. The bimodal distribution of EQ-5D scores that we found preoperatively and at the 12-month follow-up has also been reported by others (Conner-Spady et al. 2001, Xie et al. 2007). The EQ-5D algorithm tends to cluster scores in the upper extremity close to 1.0, and around 0.45. We strongly believe that it is the structure of the instrument that causes this phenomenon rather than the fact that it appears to highlight 2 subgroups of patients. This has also been noted in other studies (Rivero-Arias et al. 2005, Jansson et al. 2009). We consider that our cohort represents patients in general who have undergone orthopedic surgery. This is the largest orthopedic cohort to be studied regarding HRQL so far, with 426 diagnoses and 446 orthopedic procedures. It could be questioned why we divided the cohort into 15 groups according to anatomical region and type of surgery, but it would have been difficult to present the results in any other way due to the large number of procedures. The drawback of this is that we lost the possibility of presenting details of specific diagnoses and procedures. We notice that our large cohort had a low HRQOL according to EQ-5D index score. A major strength in our report is that we matched our cohort with the Swedish EQ-5D reference population survey (Burström et al. 2001, 2003). We compared all patients older than 20 years of age and in spite of the finding that most patients felt an improved quality of life, the average preoperative EQ-5D index score of 0.54 is among the lowest reported in the literature so far. In the population survey (Burström et al. 2001) it was found that patients with low back pain scored 0.55, patients with stroke 0.43 and those with depression 0.38.

KÅJ designed the study, compiled and analyzed the data, and wrote the manuscript. FG participated in the analysis, and in writing and editing of the manuscript.

We are grateful for the information we received from Professor Gunnar Németh, former head of the Department of Orthopedic Surgery at Karolinska

Acta Orthopaedica 2011; 82 (1): 82–89

University Hospital, and to Eila Sterner and Kerstin Holmlund who helped with patient recruitment. Karl-Åke Jansson was funded in part by grants from Karolinska Institutet and the Department of Orthopedics at Karolinska University Hospital, Solna.

No competing interests declared.

Akahane T, Shimizu T, Isobe K, Yoshimura Y, Fujioka F, Kato H. Evaluation of postoperative general quality of life for patients with osteosarcoma around the knee joint. J Pediatr Orthop B 2007; 16 (4): 269-72. Anderson P A, Puschak T J, Sasso R C. Comparison of short-term SF-36 results between total joint arthroplasty and cervical spine decompression and fusion or arthroplasty. Spine (Phila Pa 1976) 2009; 15: 34 (2): 176-83. Brooks, EuroQol Group. EuroQol. The current state of play. Health Policy 1996: 37: 53-72. Brooks R, Rabin R, de Charro F. The measurement and valuation of health status using EQ-5D: A European perspective. Evidence from the EuroQol BIOMED Research Programme. Dordrecht: Kluwer Academic Publishers 2003. Burström K, Johannesson M, Diderichsen F. Swedish population healthrelated quality of life results using the EQ-5D. Quality of Life Reserch 2001: 621-35. Burström K. Population Health and Inequalities in Health. Measurement of health - related quality of life and changes in QALYs over time in Sweden. Thesis Stockholm: ReproPrint AB 2003. Cockerill W, Lunt M, Silman A J, et al. Health-related quality of life and radiographic vertebral fracture. Osteoporos Int 2004; 15: 113-9. Conner-Spady B, Cumming C, Nabholtz JM, Jacobs P, Stewart D. Responsiveness of the EuroQol in breast cancer patients undergoing high dose chemotherapy. Qual Life Res 2001; 10 (6): 479-86. Dawson J, Fitzpatrick R, Frost S, Gundle R, McLardy-Smith P, Murray D. Evidence for the validity of a patient-based instrument for assessment of outcome after revision hip replacement. J Bone Joint Surg (Br) 2001; 83 (8): 1125-9. Dolan P. Modeling valuations for EuroQol health status. Med Care 1997: 1095-108. Dolan P, Gudex C, Kind P, Williams A. The time-trade-off method: Results from a general population study. Health Econ 1996: 141-54. Drummond M, O´bien B, Stoddart G, Torrance G. Methods for the economic evaluation of health care programmes. 3rd edn. New York: Oxford University Press 2005. England & Wales - National Joint Registry 2009. Available from http://wwwnew.njrcentre.org.uk/njrcentre/Default.aspx EQ-5D 2009: Available from http://www.euroqol.org/eq-5d/eq-5d-versions/ eq-5d-languages/available-versions.html Ethgen O, Bruyère O, Richy F, Dardennes C, Reginster JY. Health-related quality of life in total hip and total knee arthroplasty. A qualitative and systematic review of the literature. J Bone Joint Surg (Am) 2004; 86: 963-74. EuroQol Group. EuroQol - A new facility for the measurement of healthrelated quality of life. Health Policy 1990: 199-208. Giannoudis P V, Harwood P J, Kontakis G, Allami M, Macdonald D, Kay S P, Kind P. Long-term quality of life in trauma patients following the full spectrum of tibial injury (fasciotomy, closed fracture, grade IIIB/IIIC open fracture and amputation). Injury 2009; 40 (2): 213-9. Gold M, Siegel J, Russel L, Weinstein M (Eds.) Cost-effectiveness in health and medicine. New York: Oxford University Press 1996. Government Bill. National Goals for Public Health. Stockholm: Elanders Gotab, Government Bill 2003/03:35. (In Swedish) 2002. Hansson T, Hansson E, Malchau H. Utility of spine surgery: a comparison of common elective orthopedic surgical procedures. Spine 2008; 33 (25): 2819-30.


Acta Orthopaedica 2011; 82 (1): 82–89

Harrison M J, Lunt M, Verstappen S M, Watson K D, Bansback N J, Symmons D P. Exploring the validity of estimating EQ-5D and SF-6D utility values from the health assessment questionnaire in patients with inflammatory arthritis. Health Qual Life Outcomes. 2010; 11; 8:21. Hoffmann C, Gosheger G, Gebert C, Jürgens H, Winkelmann W. Functional results and quality of life after treatment of pelvic sarcomas involving the acetabulum. J Bone Joint Surg (Am) 2006; 88 (3): 575-82. Jansson K A, Nemeth G, Granath F, Blomqvist P. Health-related quality of life in patients before and after surgery for a herniated lumbar disc. J Bone Joint Surg (Br) 2005; 7: 959-64. Jansson K A, Németh G, Granath F, Jönsson B, Blomqvist P. Health-related quality of life (EQ-5D) before and one year after surgery for lumbar spinal stenosis. J Bone Joint Surg (Br) 2009; 91 (2): 210-6. Kind P, Dolan P, Gudex C, Williams A. Variations in population health status: results from a United Kingdom national questionnaire survey. BMJ 1998; 316: 736-741. Kopec J A, Willison K D. A comparative review of four preference-weighted measures of health-related quality of life. J Clin Epidemiol 2003; 56: 31725. Löfvendahl S, Eckerlund I, Hansagi H, Resch S, Hanning M. Waiting for orthopaedic surgery: factors associated with waiting times and patients’ opinion. Int J Qual Health Care 2005; 17 (2): 133–40. Meunning P, Gold M. Using the years-of health-life measure to calculate QUALYs. Am J Prev Med 2001; 20: 35-9. Mokhtar S A, McCombe P F, Williamson O D, Morgan M K, White G J, Sears W R. Health-related quality of life: a comparison of outcomes after lumbar fusion for degenerative spondylolisthesis with large joint replacement surgery and population norms. Spine J 2010; 10 (4): 306-12. NKO. Swedish National Competence Centre for Musculoskeletal Disorders (NKO) 2009. Medical indications for orthopaedic procedures. Available from www.nko.se Odenbring S, Wagner P, Atroshi I. Long-term outcomes of arthroscopic acromioplasty for chronic shoulder impingement syndrome: a prospective cohort study with a minimum of 12 years’ follow-up. Arthroscopy 2008; 24 (10): 1092-8. Osnes-Ringen H, Kvien T K, Henriksen J E, Mowinckel P, Dagfinrud H. Orthopaedic surgery in 255 patients with inflammatory arthropathies: longitudinal effects on pain, physical function and health-related quality of life. Ann Rheum Dis 2009; 68 (10): 1596-601. Rampersaud Y R, Ravi B, Lewis S J, Stas V, Barron R, Davey R, Mahomed N. Assessment of health-related quality of life after surgical treatment of focal symptomatic spinal stenosis compared with osteoarthritis of the hip or knee. Spine J 2008; 8 (2): 296-304.

89

Rivero-Arias O, Campbell H, et al. Surgical stabilisation of the spine compared with a programme of intensive rehabilitation for the management of patients with chronic low back pain: cost utility analysis based on a randomised controlled trial. BMJ 2005; 330: 1239. Räsänen P, Roine E, Sintonen H, Semberg-Konttinen V, Ryynänen O P, Roine R. Use of quality-adjusted life years for the estimation of effectiveness of health care: A systematic literature review. Int J Technol Assess Health Care 2006; 22 (2): 235-41. Scende A, Williams A (Eds). Measuring self-reported population health: An international perspective based on the EQ-5D. Rotterdam: Euroqol Group 2004. Shrive F M, Ghali W A, Johnson J A, Donaldson C P, Manns B J. Use of the U.S. and U.K. scoring algorithm for the EuroQol-5D in an economic. Evaluation of cardiac care. Med Care 2007; 45: 269-73. Swedish Knee Arthroplasty Register 2009. Available from http://www.knee. nko.se/english/online/thePages/index.php The Word Bank. Global Burden of Disease and Risk Factors ISBN-10:08213-6262-3. The Word Bank, Washington DC: Oxford University Press and the World Bank 2006. Tidermark J, Bergström G, Svensson O, Törnkvist H, Ponzer S. Responsiveness of the EuroQol (EQ 5-D) and the SF-36 in elderly patients with displaced femoral neck fractures. Qual Life Res 2003; 12 (8): 1069-79. Towheed T E, Hochberg M C. Health-related quality of life after total hip replacement. Semin Arthritis Rheum 1996; 26 (1): 483-91. Xie F, Li S C, Luo N, Lo N N, Yeo S J, Yang K Y, Fong K Y, Thumboo J. Comparison of the EuroQol and short form 6D in Singapore multiethnic Asian knee osteoarthritis patients scheduled for total knee replacement. Arthritis Rheum 2007; 15: 57 (6): 1043-9. Walters S J, Brazier J E. What is the relationship between the minimally important difference and health state utility values? The case of the SF-6D. Health Qual Life Outcomes 2003; 1: 4. Walters S J, Brazier J E (2005). Comparison of the minimally important difference for two health state utility measures: EQ-5D and SF-6D. Qual Life Res 2005;14:1523-32. WHO. Gaining Health. The European Strategy for Prevention and Control of Noncommunicable Diseases 2006. Report of the 56th Session of the WHO Europe Regional Committee for Europe, Copenhagen, 11-14 September. Available from http://www.euro.who.int/Document/RC56/edoc08. pdf.2006. Woolfe A D. The bone and joint decade 2000 –2010. Ann Rheum Dis 2000; 59: 81–2. Woofle A D, Pfleger B. Burden of major musculoskeletal conditions. Bull World Health Organ 2003; 81: 646–56.


90

Acta Orthopaedica 2011; 82 (1): 90–95

Development of simulated arthroscopic skills A randomized trial of virtual-reality training of 21 orthopedic surgeons Christine Andersen1, Trine N Winding2, and Martin S Vesterby2 1Department

of Orthopedics and 2Orthopedic Skills Laboratory, Silkeborg Regional Hospital, Silkeborg, Denmark Correspondence: chrisand79@gmail.com Submitted 09-11-26. Accepted 10-08-30

Background and purpose Previous studies have shown that there is a correlation between arthroscopic experience and performance on a virtual-reality (VR) unit. We analyzed the development inexperienced surgeons went through during VR training of shoulder arthroscopy. Methods 14 inexperienced surgeons from Silkeborg Regional Hospital were randomized into an intervention group and a control group. 7 experienced surgeons constituted another control group. All were tested twice on insightMIST—an advanced arthroscopic VR trainer—within a period of 6–15 days. The intervention group also received a 5-hour training program on the VR unit. Results The average time for the arthroscopy in the intervention group was reduced from 720 (SD 239) seconds to 223 (SD 114) seconds (p = 0.03 compared to the inexperienced control group). Distance travelled by the camera was reduced from 367 (SD 151) cm to 84 (SD 44) cm in the intervention group (p = 0.02 compared to the inexperienced control group). Depth of collisions was also significantly reduced, whereas distance travelled by the probe and number of collisions were improved in the intervention group, although not statistically significantly. Interpretation VR training is a possible way for young and inexperienced surgeons to achieve basic navigation skills necessary to perform arthroscopic surgery. Further studies regarding the transferability of the skills acquired on the VR unit to the operating theater are desirable. 

Previous studies have shown that there is a correlation between arthroscopic experience and performance on a virtual reality (VR) unit, which is an advanced arthroscopic training simulator that allows users to learn and improve in minimally invasive surgical techniques. Both single-point estimates and longitudinal studies have indicated that increased real-time arthroscopic experience correlates with performance on a VR unit (Bliss et al. 2005, McCarthy et al. 2006, Ceponis et al. 2007, Gomoll et al. 2008, Howells et al. 2008b). There have

been no studies, however, showing the development that inexperienced orthopedic surgeons go through when receiving training on a VR unit, and comparing the skills they acquire to those of experienced surgeons who regularly deal with arthroscopic surgery—or to a complementary control group of inexperienced surgeons not receiving any training. The number of arthroscopic surgical procedures is increasing. In the Danish healthcare system, where there is a high demand for productivity, it can be hard to find the time to educate the new generation of orthopedic specialists. The educational environment is not optimal for the inexperienced surgeon who is aware of the time pressure. Furthermore, patient safety must be considered. Thus, it is critical to consider alternative methods of raising the standard of education and for this purpose VR training has proven beneficial in other surgical fields, such as gynecology and laparoscopy (Larsen et al. 2006, Gurusamy et al. 2009, Kossi and Luostarinen 2009, Larsen et al. 2009). Other studies have shown that experienced surgeons have markedly better performance than inexperienced doctors on a VR unit, indicating that the skills acquired in the operating theater are transferable to the VR unit and visa versa (Pedowitz et al. 2002, Srivastava et al. 2004, Gomoll et al. 2007). We assessed the effect of training on a VR unit for a group of doctors with no arthroscopic experience. Our hypothesis was that after undergoing a training program on a VR unit, inexperienced doctors would improve compared to doctors of similar experience who did not undergo the program, and perform at the same level or better than experienced surgeons—as measured by performance on the VR unit.

Methods 21 doctors working at Silkeborg Regional Hospital in May and June 2008 were included in the study. The reason for choosing 21 was mainly based on practical circumstances such as numbers of newly started interns in the department. Before enrollment, all doctors filled in a questionnaire regarding their

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2011.552776


Acta Orthopaedica 2011; 82 (1): 90–95

91

Figure 1. InsightMIST.

arthroscopic experience and were divided into the following groups. Group 1—the experienced control group—consisted of 7 experienced arthroscopic surgeons who did at least 1 independent arthroscopic procedure a week. Group 2—the intervention group—consisted of 7 interns from the orthopedics department with no arthroscopic experience. Group 3— the inexperienced control group—consisted of 7 interns from the same hospital with no arthroscopic experience. The interns were randomized into either the control or the intervention group by drawing lots. The questionnaire also included questions regarding the amount of time spend on computer games by the subjects on a daily basis and showed no difference between the 3 groups. Data collection All the subjects went through a standard test in shoulder arthroscopy on insightMIST—an advanced arthroscopic virtual-reality trainer manufactured by GMV (Tres Cantos, Madrid, Spain). The set-up is illustrated in Figure 1. The unit was chosen after testing 2 different VR simulators, and the GMV VR trainer provided the more realistic VR environment both regarding the forced feedback and the anatomical visualization. Furthermore, GMV showed great inter-

A

B

est in continued development and improvement of the system. The unit consisted of a high-performance computer, a 19-inch touch monitor, 2 haptic devices (SensAble PHANTOM; 1 simulated camera and 1 simulated probe), 1 integration platform with dynamic elements, and 1 physical (left) shoulder model. The purpose of the test was to complete a shoulder arthroscopy by identifying a number of spheres in the joint placed on anatomic landmarks, centering the spheres one by one with a camera and palpating them with a probe, as illustrated in Figure 2. The sequence by which the spheres were placed changed from test to test, making it impossible to know the position of the next sphere. The test was executed as follows. The subject was alone in the training room during the test, only accompanied by the instructor. Before the test, the subject was handed standard written instructions explaining the purpose of and how to complete the test. Before starting the test, the subject had the opportunity to discuss and clarify any doubts with the instructor. The instructor was the same person throughout the experiment and was present through the entire duration of the test, but was not allowed to give any advice during the test. All the subjects underwent the test twice with a minimum of 6 days between tests. In the intervening period, the control groups were not allowed to use the VR unit at all. For practical reasons, the upper time limit to complete the test was set at exactly 15 min. The intervention group started their training day by completing the standard test in the same way as the subjects in the control groups. After completing the standard test, they did not repeat that particular test again until they were retested approximately 1 week later, as illustrated in Figure 3. Instead, they went through a 5-hour training program on the VR unit, starting out by receiving advice in basic navigation skills by a surgeon of more experience than themselves (although not an specialist in orthpedics). After that, they went systematically through the tests available on the VR unit. An instructor was present throughout the day and was available for guidance. After completing the 5-hour program, the subjects were not allowed to use the VR unit before repeating the standard test on the day of follow-up. All the subjects were evaluated on 5 different parameters: time to complete the exercise, number of collisions with surrounding tissue, maximum depth of collision with surrounding tissue, and paths travelled with both camera and probe. All

C

Figure 2. Locating the sphere (A). Palpating the sphere with the probe (B). Touching the sphere for 2 seconds (C).


92

Acta Orthopaedica 2011; 82 (1): 90–95

The results for the intervention group were all characterized by a vast improvement from the first to the second test, measured using all 5 parameters. The experienced control group had RANDOMIZATION only minor variations between the 2 tests. The inexperienced control group was characterized 7 subjects 7 subjects 7 subjects by what appeared to be random results, with Standard test Standard test Standard test Day 0 Day 0 Day 0 large inter- and intrapersonal variation. The intervention group went from an average of 720 seconds in the first test to 223 seconds 7 subjects in the second test, an average improvement of Education 497 seconds. The average improvement for the Day 0 inexperienced control group was 65 seconds, whereas the experienced control group neither increased nor decreased their average time 7 subjects 7 subjects 7 subjects consumption, since the difference was only 1 Standard test Standard test Standard test second. The intervention group improved more Day 6–15 Day 6–15 Day 6–15 than the inexperienced control group (p = 0.03). After going through the training program, the Figure 3. Flow chart of the experiment. intervention group completed the exercise faster that the experienced control group, but this difTable 1. Comparing group 2 (inexperienced control) and group 3 ference was not statistically significant (p = 0.4). (intervention) in test 1 The same tendency was seen when analyzing distance travelled by the camera. The intervention group had an average Difference between P-value comparing improvement of 283 cm for the camera whereas both of the the groups the groups control groups increased this distance from the first to the Group 3 – Group 2 second test. The intervention group showed the worst result Time 47 s 0.7 in the first test, but the best in the second test. The improveDistance camera 162 mm 0.9 ment made by the intervention group compared to the inexpeDistance probe 1099 mm 0.4 rienced control group was statistically significant (p = 0.02). Depth of collisions –0.02 kN 0.9 Number of collisions 52 0.4 The distance travelled by the probe showed the same tendency as for distance travelled by camera, although the improvement made by the intervention group was not statistically significant compared to the inexperienced control group. the data were computed automatically and no manual registra- The intervention group provided a better result in the second tion was involved. test than the experienced control group on distance travelled by probe and camera, although the numbers could not be Statistics shown to be statistically significantly different. Analysis of the potential damage caused to the joint by conThe data were normally distributed and were analyzed with a t-test using Stata statistical software version 9.0. A p-value of sidering the number of times either camera or probe collided with surrounding tissue, and the force by which the instru< 0.05 was considered to be statistically significant. ments collided, also showed a vast improvement in the intervention group—which went from causing the highest number of collisions in the first test to the lowest in the second test. Results (Figure 4) The average improvement was 129 less collisions for the interThe 21 subjects all completed the standard test twice at a mean vention groups, whereas the number of collisions increased interval of 10 days. We intended to keep this interval at around in both of the control groups. The improvement was not sta1 week since this was estimated to be an appropriate and tistically significant (p = 0.07) when comparing the intervenrealistic length of time for future interns between the train- tion group with the inexperienced control group. Regarding ing session on the VR unit and doing supervised but real-time the depth of collisions, the intervention group improved more arthroscopic surgery in the theater. The intervention group and than the inexperienced control group (p = 0.02) (Table 2; see the inexperienced control group were judged in a similar way supplementary data). by their performance in test 1, based on the 5 different parameters (Table 1). 7 subjects Experienced control group

14 subjects Inexperienced group


Acta Orthopaedica 2011; 82 (1): 90â&#x20AC;&#x201C;95

Group 1

93

Group 2

Group 3

Figure 4. Individual development from the first to the second test concerning the 5 different parameters, for all 3 groups. Group 1: experienced control group; group 2: inexperienced control group; group 3: intervention group


94

Discussion After completing a 5-hour training program on a VR unit, the 7 doctors who had never performed independent arthroscopic procedures showed marked improvement in their skills, placing them on the same level or making them even better than experienced surgeons (also measured on performance on the VR unit). Our study was based on a small number of subjects, thus providing some statistical uncertainty since minor variations in the subjects’ performance had a relatively large effect on the outcome. This was seen, for example, with one of the subjects in the experienced control group who provided a good result measured on all the parameters in the first test, whereas the performance in the second test was markedly worse—and worse than all the other subjects in that group. The second test probably did not reflect the subject’s skills, but might have been the result of a bad day, and given the small size of the groups the result provided by that particular person affected the average results rather much. We have found no reviews on virtual-reality training in arthroscopy and related topics. In the field of laparoscopic surgical training, we found 1 review published by the Cochrane Collaboration in January 2009. Gurusamy et al. (2009) included 23 studies (24 references) that were all randomized clinical trials. The conclusion was that “Virtual reality training can supplement standard laparoscopic surgical training of apprenticeship and is at least as effective as video trainer training in supplementing standard laparoscopic training. Further research of better methodological quality and with more patient-relevant outcomes is needed.” The number of studies focusing on individual effects of VR training and the ability to transfer skills gained on a VR unit to the operating room is rising. Hariri et al. (2004) found that using a surgical simulator was as least as effective as textbook images for learning anatomy and, more importantly, the students: “rated the simulator higher as an effective learning tool than the textbook group rated the textbook”. McCarthy et al. (2006) showed: “significant improvements in: task completion time, shorter arthroscopic path lengths, shorter probe path lengths, and fewer arthroscopic tip contacts”. 2 control laboratory studies showed a correlation between surgical experience and an objective assessment of arthroscopic skills evaluated with a VR simulator (Gomoll et al. 2007, Howells et al. 2008a, b). Howells et al. (2008b) showed in a randomized clinical trial that the orthopedic surgical trainees who had undergone a period of laboratory-based arthroscopic simulator training demonstrated improved technical performance in the operating theater compared to an untrained group. A follow-up study showed that the skills gained in a virtual environment could still be found (in 4 of 5 parameters) when the subjects were retested after 3 years (Gomoll et al. 2008). In spite of the small number of subjects, our study shows beneficial effects of training on a VR unit for inexperienced

Acta Orthopaedica 2011; 82 (1): 90–95

doctors. The improvement compared to doctors who were at the same level regarding arthroscopic skills (measured by performance on the VR unit) is striking. The fact that the intervention group outperformed the experienced group in the second test indicates that the simulator does show deviations from reality. On the other hand, the mere fact that experienced surgeons generally performed markedly better than inexperienced doctors on a VR unit indicates that the skills are somewhat transferable; some of the skills necessary to perform arthroscopic surgery such as hand-eye coordination, triangulation, and the ability to work in 3 dimensions while watching a 2-dimensional screen are skills that can be trained using a VR unit. This can lead to improved accuracy, a reduced number of errors, a reduced number of unnecessary movements, and reduced time consumption—skills that are necessary in order to perform arthroscopy in an operating theater as well as on a simulator. It would be of interest, however, to compare the performance of doctors with no VR training to doctors with VR training in an operating theater. It would also be desirable to perform similar studies on larger populations to increase the statistical certainty, and also to expand the follow-up period to investigate whether the improvements made by the intervention group would be as strong after several months or even years. Supplementary data Table 2 is available at our website (www.actaorthop.org), identification number 3843.

All authors contributed to conception and design of the study. Education on the VR unit was conducted by MV and CA. Data collection and writing of the manuscript was mainly done by CA assisted by MV. Analysis and interpretation of data was mainly done by TW.

The authors thank Dr Lars Hoej, CEPOME, for giving editorial advice and Niels Trolle of the Department of Biostatistics, University of Aarhus, for help with data analysis.

No competing interests declared.

Bliss J P, Hanner-Bailey H S, Scerbo M W. Determining the efficacy of an immersive trainer for arthroscopy skills. Stud Health Technol Inform 2005; 111: 54-6. Ceponis P J, Chan D, Boorman R S, Hutchison C, Mohtadi N G. A randomized pilot validation of educational measures in teaching shoulder arthroscopy to surgical residents. Can J Surg 2007; 50 (5): 387-93. Gomoll A H, O’Toole R V, Czarnecki J, Warner J J. Surgical experience correlates with performance on a virtual reality simulator for shoulder arthroscopy. Am J Sports Med 2007; 35 (6): 883-8. Gomoll A H, Pappas G, Forsythe B, Warner J J. Individual skill progression on a virtual reality simulator for shoulder arthroscopy: A 3-year follow-up study. Am J Sports Med 2008; 36 (6): 1139-42.


Acta Orthopaedica 2011; 82 (1): 90â&#x20AC;&#x201C;95

Gurusamy K S, Aggarwal R, Palanivelu L, Davidson B R. Virtual reality training for surgical trainees in laparoscopic surgery. Cochrane Database Syst Rev 2009; (1) (1): CD006575. Hariri S, Rawn C, Srivastava S, Youngblood P, Ladd A. Evaluation of a surgical simulator for learning clinical anatomy. Med Educ 2004; 38 (8): 896902. Howells N R, Brinsden M D, Gill R S, Carr A J, Rees J L. Motion analysis: A validated method for showing skill levels in arthroscopy. Arthroscopy 2008a; 24 (3): 335-42. Howells N R, Gill H S, Carr A J, Price A J, Rees J L. Transferring simulated arthroscopic skills to the operating theatre: A randomised blinded study. J Bone Joint Surg (Br) 2008b; 90 (4): 494-9. Kossi J, Luostarinen M. Virtual reality laparoscopic simulator as an aid in surgical resident education: Two yearsâ&#x20AC;&#x2122; experience. Scand J Surg 2009; 98 (1): 48-54.

95

Larsen C R, Sorensen J L, Ottesen B S. Simulation training of laparoscopic skills in gynaecology. Ugeskr Laeger 2006; 168 (33): 2664-8. Larsen C R, Soerensen J L, Grantcharov T P, Dalsgaard T, Schouenborg L, Ottosen C, Schroeder T V, Ottesen B S. Effect of virtual reality training on laparoscopic surgery: Randomised controlled trial. BMJ 2009; 338: b1802. McCarthy A D, Moody L, Waterworth A R, Bickerstaff D R. Passive haptics in a knee arthroscopy simulator: Is it valid for core skills training? Clin Orthop 2006; (442): 13-20. Pedowitz R A, Esch J, Snyder S. Evaluation of a virtual reality simulator for arthroscopy skills development. Arthroscopy 2002; 18 (6): E29. Srivastava S, Youngblood P L, Rawn C, Hariri S, Heinrichs W L, Ladd A L. Initial evaluation of a shoulder arthroscopy simulator: Establishing construct validity. J Shoulder Elbow Surg 2004; 13 (2): 196-205.


96

Acta Orthopaedica 2011; 82 (1): 96–101

Surgery for skeletal metastases in lung cancer Complications and survival in 98 patients Rüdiger J Weiss and Rikard Wedin Department of Molecular Medicine and Surgery, Section of Orthopaedics and Sports Medicine, Karolinska University Hospital, Karolinska Institutet, Stockholm, Sweden Correspondence: rudiger.weiss@karolinska.se Submitted 10-04-30 Accepted 10-09-17

Background and purpose Most lung cancer patients with skeletal metastases have a short survival and it is difficult to identify those patients who will benefit from palliative surgery. We report complication and survival rates in a consecutive series of lung cancer patients who were operated for symptomatic skeletal metastases. Methods This study was based on data recorded in the Karolinska Skeletal Metastasis Register. The study period was 1987–2006. We identified 98 lung cancer patients (52 females). The median age at surgery was 62 (34–88) years. 78 lesions were located in the femur or spine. Results The median survival time after surgery was 3 (0–127) months. The cumulative 12-month survival after surgery was 13% (95% CI: 6–20). There was a difference between the survival after spinal surgery (2 months) and after extremity surgery (4 months) (p = 0.03). Complete pathological fracture in non-spinal metastases (50 patients) was an independent negative predictor of survival (hazard ratio (HR) = 1.8, 95% CI: 1–3). 16 of 31 patients with spinal metastases experienced a considerable improvement in their neurological function after surgery. The overall complication rate was 20%, including a reoperation rate of 15%. Interpretation Bone metastases and their subsequent surgical treatment in lung cancer patients are associated with high morbidity and mortality. Our findings will help to set appropriate expectations for these patients, their families, and surgeons. 

Lung cancer has become one of the most common cancers worldwide and is the predominant cause of death among cancer patients. The American Cancer Society estimated that almost 160,000 patients would die from lung and bronchus cancer in the USA in 2009 (Jemal et al. 2009). Some authors have stated that lung cancer is one of the most important challenges in oncology at the present time (Boyle and Dresler 2005). Historically, about one third of all lung cancer patients are found to have bone metastases during the course of the dis-

ease. Symptoms and events of skeletal origin such as pain, pathological fractures, spinal cord compression with paraparesis, and hypercalcemia are common complications. The decline in quality of life and eventual death of these patients can be explained to some extent by skeletal complications and their treatment (Coleman 1997). Lung cancer patients with skeletal events have a short expected survival; however, some case reports have involved patients who survived several years after pathological fractures (Agarwala and Hanna 2005, Hirano et al. 2005). A major problem in selecting patients for surgery is to avoid operating on those who are likely to die very soon after surgery. Although several features help to identify patients with long survival (Bauer and Wedin 1995, Tomita et al. 2001), it is still difficult to identify those who will die early. We analyzed a consecutive series of lung cancer patients who were operated on for skeletal metastases at our department, to determine the complications and reoperation rates after surgery and to identify risk factors for early death.

Patients and methods This study was based on data recorded in the Karolinska Skeletal Metastasis Register (Wedin and Bauer 2005). The register is a quality-control database that prospectively collects individual-based information for cancer patients admitted to Karolinska University Hospital in Stockholm. All data are collected based on the national registration number (a 10-digit number), which is unique for each Swedish resident. The criterion for inclusion in the database is surgical treatment for complete or impending fractures due to skeletal metastasis. The register gathers data on patient identity, age, sex, primary tumor, location of metastases, type of pathological fracture, and surgical treatment such as method of fixation, type of implant, and postoperative complications. It includes staging information for patients diagnosed with lung cancer

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2011.552779


Acta Orthopaedica 2011; 82 (1): 96–101

97

Table 1. 98 lung cancer patients with skeletal metastases

Total

Males Females

Number of patients 98 46 Median age at diagnosis (years) 62 60 range 34–88 43–88 Smoking status at diagnosis Former smoker 47 21 Current smoker 41 20 Never smoker 10 5 Histopathology Non-small cell cancer Adenocarcinoma 58 25 Squamous cell carcinoma 14 8 Low differentiated non-small cell cancer 8 4 Large cell carcinoma 7 3 Small cell cancer 6 3 Other cancers 5 3 Skeletal metastasis Femur 46 18 Vertebra 31 24 Humerus 16 2 8 2 Other Type of surgical procedure 34 10 Internal fixation 32 11 Endoprosthesis Spinal decompression 17 12 + instrumentation 15 12 Spinal decompression 3 1 Other Preoperative treatment (primary tumor) 52 24 Chemotherapy 27 10 Radiotherapy 12 5 Lobectomy Postoperative treatment (bone metastasis) 59 28 Radiotherapy

52 63 34–84 26 21 5 33 6 4 4 3 2 28 7 14 6 24 21 5 3 2 28 17 7 31

(Mountain 1997). Neurological function in patients with spinal metastases is assessed by the Frankel classification of motor and sensory compromise (Frankel et al. 1969, Davis et al. 1993). The neurological function is assessed preoperatively and within 2 weeks postoperatively. For the present study, we included all lung cancer patients who had surgery due to pathological fractures. None of the patients were excluded for any reason. In patients who had surgery for more than 1 metastasis, all sites were included in the analysis. However, only the first surgical procedure was accounted for in the survival analysis. The study period was 1987–2006. We identified 98 individual lung cancer patients (52 females) treated surgically for skeletal metastatic lesions. The median age at surgery was 62 (34–88) years. Most patients in this cohort (n = 88) were current or former smokers at diagnosis of the disease. Adenocarcinoma was found in most cases (n = 58) followed by squamous cell carcinoma (n = 14) (Table 1). The lung cancer patients were grouped in the following stages: 11 IA + IB, 1 IIA + IIB, 5 IIIA, 73 IIIB + IV, and such data were missing for 8 patients.

Statistics Median values and ranges were used as descriptive statistics. Kaplan-Meier analysis was used to construct the cumulative survival with 95% confidence intervals (CIs). The time between diagnosis of lung cancer and death and time between surgical procedure and death were included in the survival analysis. The log-rank test was used to compare survival after extremity surgery with spine surgery. The Cox multiple-regression model was used to study risk factors for death related to the patient and to the surgical procedure. The results were expressed as hazard ratios (HRs) with corresponding 95% CI. If a HR was > 1, the patients at risk were dying at a faster rate than the patients in the reference group. The assumption of proportional hazards was investigated by hazard function plots and log-log plots for all covariates. No signs of insufficient proportionality were detected in the hazard functions and the log-log plots ran parallel for all covariates. The factors studied in the univariate Cox model were as follows: age, sex, smoking status, type of lung cancer, staging, location of skeletal metastasis, type of pathological fracture, preoperative hemoglobin levels, time period of surgery (1987–1996 vs. 1997–2006) and perioperative chemoand radiotherapy (Table 1). Any variable whose univariate test had a p-value of < 0.25 was considered as a candidate for the multivariate model along with all variables of known biological importance. Wilcoxon’s signed ranks test was used to compare preoperative and postoperative neurological function in patients with spine metastases. The level of significance was set at p ≤ 0.05. All statistical analyses were performed using the PASW statistics package version 18 (SPSS Inc., Chicago, IL).

Results Most metastatic bone lesions were located in the femur and vertebra. Of the 31 vertebral lesions, 22 were located in the thoracic spine and 9 in the lumbar spine. 53 of 70 non-spinal lesions had caused complete pathological fractures. Regarding treatment of the primary tumor, more than half of the patients (n = 52) had chemotherapy and 27 had radiotherapy. Postoperative radiotherapy for the operated lesions was registered in 59 patients (Table 1). None of the patients were lost to followup, and none were still alive at the time of this analysis. The neurological function in patients with spine metastases improved after surgery (p = 0.001). 16 of the 31 patients gained at least 1 Frankel grade, 14 maintained their neurological function, and 1 patient deteriorated (Table 2). The main indication for surgery in these patients was spinal cord compression (n = 29) or painful instability (n = 2). Survival analysis The median survival time after surgery of skeletal metastatic lesions was 3 (0–127) months. Males had a median survival


98

Acta Orthopaedica 2011; 82 (1): 96–101

Table 2. Pre- and postoperative neurological function in 31 patients with spine metastasis graded according to Frankel

Preop.

A. Complete paraplegia B. No motor function C. Motor function useless D. Slight motor deficit E. No motor deficit

Follow-up

2 6 14 6 3

A

All 10 (0–168) Males 8 (1–60) Females 10 (0–168) Histopathology Non-small cell cancer 10 (0–168) Adenocarcinoma 11 (1–168) Squamous cell carcinoma 9 (0–43) Low differentiated non-small cell cancer 8 (1–12) Large cell carcinoma 11 (6–60) Small cell cancer 8 (1–23) Other cancers 5 (1–24) Spinal surgery All 7 (2–29) Males 7 (2–29) Females 8 (3–20) Extremity surgery All 10 (0–168) Males 10 (1–60) Females 11 (0–168) Pathological fracture Complete 8 (0–80) Impending 12 (2–168)

1 2 7 16 5

Survival (%) 100

80

60

40

B 3 (0–127) 2 (0–59) 4 (0–127) 4 (0–127) 4 (0–127) 2 (0–18) 2 (0–8) 7 (0–59) 2 (0–3) 1 (1–15) 2 (0–15) 2 (0–15) 1 (1–7) 4 (0–127) 3 (0–59) 4 (0–127) 2 (0–61) 4 (0–127)

A: time between diagnosis of lung cancer and death; B: time between surgical procedure and death.

20

0 0 n = 98

Table 3. Median survival of lung cancer patients with skeletal metastases. Time in months (range)

2 57

4 39

6 25

8 18

10 15

12 14

14 10

16 9

18 6

Months after surgery/number at risk Cumulative survival of lung cancer patients after surgery for skeletal metastases, with 95% confidence intervals.

Table 4. Risk of death after surgery for skeletal metastases in lung cancer patients (Cox regression analysis) Factor Complete pathological fracture a Vertebral skeletal metastasis b

time of 2 (0–59) and females of 4 (0–127) months (Table 3). The cumulative 6-, 12-, and 18-month survival after surgery was 24% (CI: 15–33), 13% (CI: 6–20), and 6% (1–11) respectively (Figure). There was a difference between the median survival time after spinal surgery (2 (0–15) months) and after extremity surgery (4 (0–127) months) (p = 0.03) (Table 3). The univariate Cox regression analysis revealed an increased risk of death after surgery for patients with complete pathological fracture, metastatic lesions in the vertebra, preoperative hemoglobin levels less than 10 g/dL, small cell cancer, and the absence of postoperative radiotherapy. The multivariate Cox analysis with the above 5 variables and age and sex showed that complete pathological fracture was a significant predictor of survival (Table 4). All the other factors studied—such as age, sex, smoking status, staging, time period of surgery, and chemotherapy—were not associated with an increased risk of death. Reoperations and complications The overall complication rate including reoperations was

Hemoglobin level < 10 g/dL c Small cell cancer d Postoperative radiotherapy (no) e

HR

95% CI

p-value

1.7 f 1.8 g 1.7 f 1.7 g 1.9 f 1.8 g 2.6 f 1.9 g 1.7 f 1.6 g

1.1–2.6 1.0–3.0 1.0–2.7 0.9–3.4 1.1–3.4 0.9–3.7 1.2–6.0 0.8–4.8 1.1–2.6 0.9–2.9

0.02 0.05 0.03 0.1 0.03 0.09 0.02 0.2 0.03 0.1

Factor versus a impending pathological fracture, b femur, c ≥ 10 g/ dL,d adenocarcinoma, e yes; HR = hazard ratio, f unadjusted, g adjusted (HRs are adjusted for type of pathological fracture, location of skeletal metastasis, preoperative hemoglobin levels, type of lung cancer, postoperative radiotherapy, age, and sex).

20/98. 15 failed reconstructions in 14 patients led to a new operation. The reasons for these reoperations were poor initial fixation, local tumor progression, periprosthetic fracture, deep infection, paraplegia, material failure, nonunion, and technical error. The median time to failure was 1 (0–13) months (Table 5). Complications that were treated non-surgically were seen in 5 of the patients, including wound infection, gastric bleeding, poor initial fixation, and technical error (Table 6).


Acta Orthopaedica 2011; 82 (1): 96â&#x20AC;&#x201C;101

99

Table 5. Reoperations after surgery for skeletal metastases in lung cancer patients Sex, age, Localization Primary operation histology F, 57, b F, 63, c F, 62, a F, 62, a F, 74, a F, 66, a M, 45, a F, 73, b F, 62, a F, 53, a M, 56, a M, 55, d F, 48, a M, 72, b

Humerus, proximal Intramedullary nail Humerus, distal Plate Acetabulum Acetabular reinforcement ring + total joint arthroplasty Femur, proximal Total joint arthroplasty Femur, proximal Hemiarthroplasty Femur, trochanteric Hemiarthroplasty Femur, trochanteric Sliding hip screw Femur, trochanteric Sliding hip screw Femur, subtrochanteric Hemiarthroplasty Femur, distal Plate Vertebra, thoracic Decompression Vertebra, thoracic Decompression + instrumentation Vertebra, thoracic Decompression + instrumentation Vertebra, thoracic Decompression

a adenocarcinoma. b

Time (months) to and reason for failure 5, local tumor progression 3, material failure 1, deep infection 9, periprosthetic fracture 3, periprosthetic fracture 1, poor initial fixation 1, poor initial fixation 1, poor initial fixation 4, poor bone stock 0, technical error 13, non-union + local tumor progression 0, paraplegia (hematoma)

Treatment Hemiarthroplasty Plate Debridement Modular tumor prosthesis Modular tumor prosthesis Total joint arthroplasty Total joint arthroplasty Exchange of screw Total joint arthroplasty Hemiarthroplasty Plate Removal of hematoma

8, local tumor progression

Decompression

0, poor initial fixation 1, deep infection

Reinstrumentation Debridement

squamous cell carcinoma. c large cell carcinoma. d other.

Table 6. Complications after surgery for skeletal metastasis in lung cancer patients Sex, age, Localization Primary operation histology

Time (months) to and reason for failure

Treatment

F, 75, a M, 48, c F, 48, b M, 62, d M, 55, d

0, fatal gastric bleeding 0, wound infection 0, gastric bleeding 0, technical error 1, wound infection

Non-surgical Non-surgical Non-surgical Non-surgical Non-surgical

Femur, distal Vertebra, thoracic Vertebra, thoracic Vertebra, thoracic Vertebra, thoracic

a adenocarcinoma. b

Intramedullary nail Decompression + instrumentation Decompression + instrumentation Decompression + instrumentation Decompression

large cell carcinoma. c small cell lung cancer. d other.

Discussion Our main finding was the short survival of lung cancer patients after surgery for skeletal metastases. Complete pathological fracture was identified as an independent risk factor for death. Moreover, a high complication rate poses a considerable burden to the patients. The overall median survival time after surgery was only 3 months, and just 13% of the patients were still alive 1 year postoperatively. 3 of 4 patients were already classified as having an advanced stage of lung cancer at the diagnosis of the primary tumor. The 1-year survival rate of patients operated for skeletal metastases from various primary tumor sites is limited, and ranges between 30 and 54% (Bono et al. 1991, Bauer and Wedin 1995, Durr and Refior 1998, Bohm and Huber 2002, Hansen et al. 2004, Lin et al. 2007). Differences in patient selection, type of primary tumor, and indications for surgery may explain the spread in survival time. We have previously reported on the surgical outcome of 107 patients with

skeletal breast cancer metastases. The postoperative median survival in these patients was 8 months (Wedin et al. 2001). Sugiura et al. (2008) reported a median survival time of 7 months in 118 patients with bone metastasis from lung cancer; however, only one third of these patients underwent surgery. Nathan et al. (2005) found that in patients with various types of cancer metastases, lung cancer patients fared the worst with a median survival time of 4 months. The short survival time has also been reported by other authors (Bauer and Wedin 1995, Durr and Refior 1998, Katagiri et al. 2005). We found a shorter survival after operations for skeletal metastases of the spine (2 months) as compared to the extremities (4 months). Bauer and Wedin (1995) described a median 1-year survival of 25% after spinal surgery and 31% after extremity surgery, based on all types of cancer metastases. However, this difference was not statistically significant. The decision between a surgical and non-surgical treatment of a patient with bone metastasis is influenced by the anatomical location, the tumor type, the extent of the tumor, the gen-


100

eral medical condition of the patient, and expected survival time. Surgery may be chosen if the patient is healthy enough and life expectancy is sufficiently long to benefit from pain relief, improved mobility, and care. If there is no expectation that the survival may be long enough for the patient to recover and truly benefit from surgical procedures, other palliative therapies may be chosen instead. However, this may restrict the patients involved to bed or chair. New therapies have been introduced to improve the survival of patients with lung cancer (Giaccone et al. 2006, Sandler et al. 2006). The median survival of patients with advanced lung cancer has increased from 6 months to 12 months with the introduction of new treatment regimens such as chemotherapy in combination with a monoclonal antibody, which acts as an angiogenesis inhibitor (Rosen et al. 2003, Sandler et al. 2006). However, the efficacy of promising new agents in lung cancer patients concerning treatment of metastases in bone, which is one of the most common metastatic sites, is unknown (Langer and Hirsh 2010). Some authors have stated that a life expectancy of at least 2 months is usually required for meaningful surgery of limb metastases (Harrington et al. 1976), and of 3 to 6 months for spinal lesions (Cybulski et al. 1987, Atanasiu et al. 1993, Tomita et al. 1994). We think that stabilization of long-bone fractures is almost always justified unless the patient has reached a terminal stage. However, we must raise the question of whether spinal surgery would be justified in our cohort with a median survival time of barely 2 months. Decision making regarding the management of pathological fractures is complex, balancing tumor biology, biomechanics, and functional outcome goals. We believe that surgery for vertebral metastases may be the best alternative in patients who are expected to live for at least another 2–3 months if surgery leads to a substantial difference, i.e. if the patient can avoid being bedridden or can retain the ability to ambulate independently. More than half of our patients with spine disease improved considerably, i.e. at least 1 Frankel grade. Still, a 2-month survival will be necessary in most cases to gain real benefit from the procedure, considering postoperative pain, surgical complications, and sometimes slow neurological recovery. Today, we are more cautious with patient selection for surgery because of the short survival rate in combination with a high complication rate in this cohort. We identified low preoperative hemoglobin levels as a negative predictor of survival. Hemoglobin has been used as a prognostic factor in metastatic prostate cancer (Matzkin et al. 1993) and in heterogeneous groups of cancer patients (Hansen et al. 2004, Nathan et al. 2005). In the multivariate analysis, we could show that complete pathological fractures had an unfavorable prognosis for survival, as described by other authors (Sugiura et al. 2008). Almost all of the patients (14/16) with humeral metastases had complete fractures. This is probably because impending

Acta Orthopaedica 2011; 82 (1): 96–101

fractures in non-weight bearing bones frequently escape diagnosis due to the relative absence of load-related pain. We found a complication rate of 20%. The total failure rate leading to reoperation of previously operated metastatic sites was 15%. We have previously reported a reoperation rate of 12% in skeletal breast cancer metastases (Wedin et al. 2001) and a failure rate of 11% in patients treated for metastatic lesions of the long bones (Wedin et al. 1999). Nathan et al. (2005) described a similar reoperation rate (11%) after surgery in patients with different types of cancer. Failures are often due to local tumor progression, stress fracture, implant failure due to nonunion, or poor fixation of osteosynthetic devices in insufficient bone stock. They appear equally often in the extremities and in the spine (Yazawa et al. 1990, Bono et al. 1991), as in our series. The high reoperation rate in our study may also be partly explained by the fact that during the 20-year time frame, many orthopedic surgeons—not all sub-specialized in orthopedic oncology—performed the operations. Our study was limited by confounding factors. Changes in diagnostics, in selection of patients, in perioperative treatment, in indications for surgery, and in methods of fixation during the 20-year study period may have influenced the results. To address the question of whether or not to operate on lung cancer patients with bone metastases, one would want to know the survival prognosis for all patients—not only the ones who (historically) have undergone surgery—as they represent a selected group of patients. This could not be investigated, however, as our register covers only surgically-treated cancer patients. Despite its shortcomings, the present work represents the largest follow-up study of lung cancer patients after surgery for skeletal metastases. Most studies have analyzed heterogeneous populations of patients with bone metastasis. Most analyses on indications for surgery have been based on the treatment of breast and prostate cancer, which is not necessarily applicable to lung cancer patients. In conclusion, our findings highlight the fact that surgical treatment of bone metastases in lung cancer patients is associated with high morbidity and mortality. Thus, the selection process for the (probably) few lung cancer patients who will benefit from surgical treatment is important, especially those with spinal cord compression.

Both authors contributed to the planning of the study, interpretation of the results, and editing of the manuscript. RJW performed the data analysis and wrote the manuscript. RW collected the data.

The study was supported by grants from Karolinska Institutet and Stockholm County Council.

No competing interests declared.


Acta Orthopaedica 2011; 82 (1): 96â&#x20AC;&#x201C;101

Agarwala A K, Hanna N H. Long-term survival in a patient with stage IV nonsmall-cell lung carcinoma after bone metastasectomy. Clin Lung Cancer 2005; 6: 367-8. Atanasiu J P, Badatcheff F, Pidhorz L. Metastatic lesions of the cervical spine. A retrospective analysis of 20 cases. Spine (Phila Pa 1976) 1993; 18: 1279-84. Bauer H C, Wedin R. Survival after surgery for spinal and extremity metastases. Prognostication in 241 patients. Acta Orthop Scand 1995; 66: 143-6. Bohm P, Huber J. The surgical treatment of bony metastases of the spine and limbs. J Bone Joint Surg (Br) 2002; 84: 521-9. Bono B, Cazzaniga P, Pini V, Zurrida S M, Spagnolo R, Torelli L, Corona C, Bono A. Palliative surgery of metastatic bone disease: a review of 83 cases. Eur J Cancer 1991; 27: 556-8. Boyle P, Dresler C. Preventing the lung cancer epidemic. Ann Oncol 2005; 16: 1565-6. Coleman R E. Skeletal complications of malignancy. Cancer 1997; 80: 158894. Cybulski G R, Von Roenn K A, Dâ&#x20AC;&#x2122;Angelo C M, DeWald R L. Luque rod stabilization for metastatic disease of the spine. Surg Neurol 1987; 28: 277-83. Davis L A, Warren S A, Reid D C, Oberle K, Saboe L A, Grace M G. Incomplete neural deficits in thoracolumbar and lumbar spine fractures. Reliability of Frankel and Sunnybrook scales. Spine (Phila Pa 1976) 1993; 18: 257-63. Durr H R, Refior H J. Prognosis of skeletal metastases. Orthopade 1998, 27: 294-300. Frankel H L, Hancock D O, Hyslop G, Melzak J, Michaelis L S, Ungar G H, Vernon J D, Walsh J J. The value of postural reduction in the initial management of closed injuries of the spine with paraplegia and tetraplegia. I. Paraplegia 1969; 7: 179-92. Giaccone G, Gallegos Ruiz M, Le Chevalier T, Thatcher N, Smit E, Rodriguez J A, Janne P, Oulid-Aissa D, Soria J C. Erlotinib for frontline treatment of advanced non-small cell lung cancer: a phase II study. Clin Cancer Res 2006; 12: 6049-55. Hansen B H, Keller J, Laitinen M, Berg P, Skjeldal S, Trovik C, Nilsson J, Walloe A, Kalen A, Wedin R. The Scandinavian Sarcoma Group Skeletal Metastasis Register. Survival after surgery for bone metastases in the pelvis and extremities. Acta Orthop Scand (Suppl 311) 2004, 75: 11-5. Harrington K D, Sim F H, Enis J E, Johnston J O, Diok H M, Gristina A G. Methylmethacrylate as an adjunct in internal fixation of pathological fractures. Experience with three hundred and seventy-five cases. J Bone Joint Surg (Am) 1976; 58: 1047-55. Hirano Y, Oda M, Tsunezuka Y, Ishikawa N, Watanabe G. Long-term survival cases of lung cancer presented as solitary bone metastasis. Ann Thorac Cardiovasc Surg 2005, 11: 401-4. Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun M J. Cancer statistics, 2009. CA Cancer J Clin 2009; 59: 225-49.

101

Katagiri H, Takahashi M, Wakai K, Sugiura H, Kataoka T, Nakanishi K. Prognostic factors and a scoring system for patients with skeletal metastasis. J Bone Joint Surg (Br) 2005; 87: 698-703. Langer C, Hirsh V. Skeletal morbidity in lung cancer patients with bone metastases: demonstrating the need for early diagnosis and treatment with bisphosphonates. Lung Cancer 2010; 67: 4-11. Lin P P, Mirza A N, Lewis V O, Cannon C P, Tu S M, Tannir N M, Yasko A W. Patient survival after surgery for osseous metastases from renal cell carcinoma. J Bone Joint Surg (Am) 2007; 89: 1794-801. Matzkin H, Perito P E, Soloway M S. Prognostic factors in metastatic prostate cancer. Cancer 1993; 72: 3788-92. Mountain C F. Revisions in the International System for Staging Lung Cancer. Chest 1997; 111: 1710-7. Nathan S S, Healey J H, Mellano D, Hoang B, Lewis I, Morris C D, Athanasian E A, Boland P J. Survival in patients operated on for pathologic fracture: implications for end-of-life orthopedic care. J Clin Oncol 2005; 23: 6072-82. Rosen L S, Gordon D, Tchekmedyian S, Yanagihara R, Hirsh V, Krzakowski M, Pawlicki M, de Souza P, Zheng M, Urbanowitz G, Reitsma D, Seaman J J. Zoledronic acid versus placebo in the treatment of skeletal metastases in patients with lung cancer and other solid tumors: a phase III, doubleblind, randomized trial--the Zoledronic Acid Lung Cancer and Other Solid Tumors Study Group. J Clin Oncol 2003; 21: 3150-7. Sandler A, Gray R, Perry M C, Brahmer J, Schiller J H, Dowlati A, Lilenbaum R, Johnson D H. Paclitaxel-carboplatin alone or with bevacizumab for nonsmall-cell lung cancer. N Engl J Med 2006; 355: 2542-50. Sugiura H, Yamada K, Sugiura T, Hida T, Mitsudomi T. Predictors of survival in patients with bone metastasis of lung cancer. Clin Orthop 2008; (466): 729-36. Tomita K, Kawahara N, Baba H, Tsuchiya H, Nagata S, Toribatake Y. Total en bloc spondylectomy for solitary spinal metastases. Int Orthop 1994; 18: 291-8. Tomita K, Kawahara N, Kobayashi T, Yoshida A, Murakami H, Akamaru T. Surgical strategy for spinal metastases. Spine (Phila Pa 1976) 2001; 26: 298-306. Wedin R, Bauer H C. Surgical treatment of skeletal metastatic lesions of the proximal femur: endoprosthesis or reconstruction nail? J Bone Joint Surg (Br) 2005; 87: 1653-7. Wedin R, Bauer H C, Wersall P. Failures after operation for skeletal metastatic lesions of long bones. Clin Orthop 1999: (358): 128-39. Wedin R, Bauer H C, Rutqvist L E. Surgical treatment for skeletal breast cancer metastases: a population-based study of 641 patients. Cancer 2001; 92: 257-62. Yazawa Y, Frassica F J, Chao E Y, Pritchard D J, Sim F H, Shives T C. Metastatic bone disease. A study of the surgical treatment of 166 pathologic humeral and femoral fractures. Clin Orthop 1990; (251): 213-9.


102

Acta Orthopaedica 2011; 82 (1): 102–111

Time course of skeletal muscle regeneration after severe trauma Muscle function against the background of MRI and histological findings Tobias Winkler1, Philipp von Roth1, Georg Matziolis1, Maria R Schumann1, Sebastian Hahn1, Patrick Strube1, Gisela Stoltenburg-Didinger2, Carsten Perka1, Georg N Duda1, and Stephan V Tohtz1 1Center for Musculoskeletal Surgery and Julius Wolff Institute Berlin, Brandenburg Center for Regenerative Therapies; and 2Institute of Neuropathology, Charité – Universitaetsmedizin Berlin, Berlin, Germany Correspondence: tobias.winkler@charite.de Submitted 09-09-13. Accepted 10-07-27.

Background and purpose Animal models of skeletal muscle injury should be thoroughly described and should mimic the clinical situation. We established a model of a critical size crush injury of the soleus muscle in rats. The aim was to describe the time course of skeletal muscle regeneration using mechanical, histological, and magnetic resonance (MR) tomographic methods. Methods Left soleus muscles of 36 Sprague-Dawley rats were crushed in situ in a standardized manner. We scanned the lower legs of 6 animals by 7-tesla MR one week, 4 weeks, and 8 weeks after trauma. Regeneration was evaluated at these times by in vivo measurement of muscle contraction forces after fast-twitch and tetanic stimulation (groups 1W, 4W, 8W; 6 per group). Histological and immunohistological analysis was performed and the amount of fibrosis within the injured muscles was determined histomorphologically. Results MR signals of the traumatized soleus muscles showed a clear time course concerning microstructure and T1 and T2 signal intensity. Newly developed neural endplates and myotendinous junctions could be seen in the injured zones of the soleus. Tetanic force increased continuously, starting at 23% (SD 4) of the control side (p < 0.001) 1 week after trauma and recovering to 55% (SD 23) after 8 weeks. Fibrotic tissue occupied 40% (SD 4) of the traumatized muscles after the first week, decreased to approximately 25% after 4 weeks, and remained at this value until 8 weeks. Interpretation At both the functional level and the morphological level, skeletal muscle regeneration follows a distinct time course. Our trauma model allows investigation of muscle regeneration after a standardized injury to muscle fibers. 

Injuries of skeletal muscle tissue cause deficiencies in local muscle function that can also affect bone healing (Duda et

al. 2003, Schaser et al. 2003, Utvag et al. 2003, Harry et al. 2008) or the long-term success of a prosthesis (Kleemann et al. 2003, Perka et al. 2005). Despite the fact that there have been several experimental approaches, no methods for the treatment of the main causes of these deficiencies—i.e. the loss of contractile muscle substance and the formation of fibrosis—have been taken into routine clinical use. One reason is that the results of putative therapies are often only described mono-dimensionally, in most cases histologically, and rarely functionally or with in vivo diagnostic methods. Animal models of muscle contusion injuries should closely mimic the clinical situation. Among them, open crush injuries allow standardized evaluation of regeneration in a selected muscle (Schultz et al. 1985, Rushton et al. 1997). For the execution of the trauma, either forceps (McGeachie and Grounds 1987, Kurek et al. 1997, Fink et al. 2003) or custom-made devices have been used (Jarvinen and Sorvari 1975, Rushton et al. 1997, Bunn et al. 2004). Previous muscle-crush models have had the disadvantage of either affecting only part of the muscle or of impairing the organ innervation and blood supply. Two types can be found in the literature: the segmental crush (Schmalbruch 1976) and the complete crush, where only 4–6% of the muscle fibers remain intact (Fink et al. 2003). In the latter, myoneuronal junctions are damaged, which triggers not only regeneration of muscle substance but also initial innervation deficits. These deficits always lead to impaired healing (Saunders and Sissons 1953, Koishi et al. 1995, Dedkov et al. 2001, Pereira et al. 2006). Histological analysis of the regenerative process after crush injury has been done by a number of authors, describing the initial phase of inflammation followed by satellite cell activation, myotube regeneration, and fibrosis of the muscle (Jarvinen and Sorvari 1975, Schmalbruch 1976, Sorokin et al.

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2010.539498


Acta Orthopaedica 2011; 82 (1): 102–111

2000, Bunn et al. 2004). It has been shown that the development of fibrotic tissue is one of the key factors for regenerative deficits of muscle function after trauma (Anderson et al. 1999, Nozaki et al. 2008). The development of intramuscular fibrosis over time after crush trauma has not, however, been quantitatively described. Magnetic resonance imaging (MRI) can be used in the clinical evaluation of skeletal muscle injuries. Not only are T1and T2- weighted spin echo sequences highly sensitive for detection of edema and bleeding in anatomical relation to the structures affected (Elsayes et al. 2006), but MR scans also allow the determination of intra-individual time courses after trauma. The evaluation of muscle regeneration by contraction force measurements provides information about the functional outcome of therapeutic approaches (Sato et al. 2003, Negishi et al. 2005, Winkler et al. 2009). The description of a muscle trauma model should therefore not solely include morphological but also mechanical data. One aim of our study was to establish a standardized model of selective trauma of the muscle substance of a single muscle, leaving the blood and nerve supply undamaged. This was achieved by developing a global crush injury of the rat soleus muscle with respect to the topographic arrangement of the innervating endplate zone. We also wanted to describe the regeneration of muscle contraction forces against the background of morphological changes within the muscle tissue following a critical size of injury in a skeletal muscle.

Material and methods Animals 36 male Sprague Dawley rats weighing 450–550g were used for the study. The rats were housed at a constant temperature of 25°C with free access to pellet food and water. All animal experiments were carried out according to the policies and principles established by the Animal Welfare Act, the NIH Guide for Care and Use of Laboratory Animals, and the German national animal welfare guidelines. The study was approved by the local institute of health. Experimental procedure All animals received an open crush trauma of the left soleus muscles. 1, 4, and 8 weeks after injury, muscle contraction forces were measured in vivo and the animals were killed (groups 1W, 4W, and 8W; n = 6 per group). The soleus muscles of both sides were harvested for histological evaluation. Soleus muscles of 4 animals were harvested on days 1, 2, and 4 after trauma for histological evaluation of the first days after injury. 6 animals were examined by sequential MRI measurements at 1, 4, and 8 weeks (group MRI, n = 6).

103

Figure 1. Innervation of soleus muscle and vascular supply. Arrowheads: lateral border of soleus muscle (SO); arrow: neurovascular bundle coming from the gastrocnemius muscle (GM) and entering at the medial border of the SO, the tip of the arrow indicating the soleus nerve; PE: peroneal muscles. The asterisk shows the Achilles tendon.

Muscle trauma The rats were anesthetized by subcutaneous injections of 0.15 mL ketamine (100 mg/mL) and 0.15 mL xylazine (2%) diluted with 0.2 mL 0.9% saline. The left leg was shaved and disinfected with povidone-iodine. Through a 2-cm posterolateral longitudinal skin incision from the lateral gastrocnemius head to the Achilles tendon, the soleus muscle was mobilized and the soleus artery, vein and nerve, which arise from the midpart of the gastrocnemius, were located. The medial border of the soleus muscle was dissected cranially and caudally to the neurovascular structures. A curved artery forceps, the jaws of which were protected by polyethylene tubes to avoid lesions of the muscle fascia, was introduced in direct proximity to the Achilles tendon and closed for 20 seconds. This procedure was repeated proximally 3 times, always in direct continuity to the respective distal crush. The insertion of the neurovascular bundle and a trapezoid region with a 3-mm long base at the insertion of the bundle and a 2-mm short base at the lateral margin of the soleus muscle was spared and 3 crushes were exerted proximal to this area in the manner described above (Figure 1). Standardized pressure on the muscle was ensured by closing the forceps to its third stage at each crush. The closing pressure at this stage corresponds to 112 (SD 5.1) N (data obtained from preliminary tests with the material testing device Zwick 1455 (Zwick GmbH, Ulm, Germany). After multiple irrigations, the superficial muscle and skin were closed. Measurement of muscle force The animals were anesthetized as described before and received bilateral surgery. The sciatic nerve and the soleus muscle were exposed, sparing all neurovascular structures.


104

The Achilles tendon was cut and the lower extremity was connected to a muscle force measuring device (Experimentria, Budapest, Hungary). The distal part of the soleus muscle was connected to a force transducer via a suture (4-0, silk). The sciatic nerve was subsequently stimulated with 9 mA/75 Hz bipolar pulses 5 times, for 0.1 second each (8 periods) with 5-second intervals between the pulses (Broniatowski et al. 1992, Silverman and Brull 1993). After this short twitch stimulation, the maximal muscle strength was measured using a protocol of 9 mA/75 Hz pulses 5 times, for 3 seconds each at 5-second intervals, reaching tetany in all cases. After completion of the muscle strength measurements, the animals were killed with an overdose of anesthetic and the soleus muscles were fixed in buffered paraformaldehyde (4%) for histological examination. MRI 6 animals were examined by MRI one, 4, and 8 weeks after muscle injury. Measurements were taken with a 7-tesla MRI spectrometer for small animals (Pharmascan 70/16; Bruker, Ettlingen, Germany). Images were processed with ParaVision software (Bruker BioSpin). Rats were anesthetized by isofluran/O2 inhalation and were introduced into the spectrometer in a special animal handling system equipped with a heating pad and continuous monitoring (ECG and respiratory frequency). A rat brain coil with an inner diameter of 38 mm was used for transmission and reception. Each measurement involved one lower limb of the animal from the tibial plateau to the ankle joint. T1 (TR/TE: 1000/10.6 ms; scan time 12 min) and T2-turbo-rare sequences (TR/TE: 6351.6/75.6 ms; scan time 12 min) were performed with a resolution of 176 × 176 µm and a slice thickness of 500 µm. Analysis of grayscale values was performed with ImageJ version 1.38i (National Institutes of Health, Bethesda, MD). The signal intensities of the voxels of the injured soleus muscles were quantified in the T1- and T2-weighted scans and compared to those of an uninjured reference muscle (anterior tibial muscle). The differences between the mean grayscale values of the soleus muscles were calculated for each time point (Figure 4). Histology Hematoxylin and eosin (HE) staining and Picro-sirius red staining were used for descriptive analysis of the trauma. The stain Picro-sirius red selectively highlights collagenous connective tissues with a well-defined contrast. This stain was therefore used for quantitative analysis of fibrosis in the traumatized muscles, as previously described (Sweat et al. 1964, Wake et al. 2005). Furthermore, muscles were stained for nestin to localize the positioning of myotendinous junctions (MTJs) before and after trauma (Aarimaa et al. 2004) and with a-bungarotoxin to visualize the distribution of neuromuscular junctions within the soleus muscle. The endothelial marker factor VIII was used for the detection of vessels.

Acta Orthopaedica 2011; 82 (1): 102–111

Soleus muscles were either paraformaldehyde-fixed, dehydrated, and embedded in paraffin or fresh-frozen in liquid nitrogen-cooled 2-methylbutan and embedded in Tissue Tek. The specimens were sectioned longitudinally (4 and 10 µm). For evaluation of connective tissue, sections were incubated for 60 min in Sirius red solution (5g Sirius red dissolved in 500 mL saturated picric acid). Differentiation was achieved by 2 washes with diluted acetic acid followed by a short dehydration in a series of graded alcohols. The following antisera and antibodies were used for immunohistochemistry: mouse monoclonal antibody to rat nestin (1:800)  (BD Biosciences Pharmingen, Franklin Lakes, NJ) and mouse monoclonal antibody to factor VIII (1:400) (Daco Cytomation, Glostrup, Denmark). For light microscopy, the primary antibodies were visualized using appropriate avidin–biotin–peroxidase kits (Vector Laboratories, Burlingame, CA) and counterstained with hematoxylin. For fluorescence microscopy, Alexa Fluor 546-conjugated goat anti-mouse antibody (1:300) (Invitrogen Corporation, Camarillo, CA) was used as secondary antibody. DAPI was used as counterstain for all fluorescence stains. For endplate staining, the frozen sections were incubated for 2 h at room temperature with Alexa Fluor 488-conjugated a-bungarotoxin (2 µg/mL) (Invitrogen). A blinded investigator evaluated the amount of collagenous connective tissue in the muscles. In order to measure the total area of endo- and perimysial fibrosis digitally, microscopic images of whole longitudinal sections of the muscle were collected from the distal to the proximal tendon. Images were correlated and connected to obtain the whole longitudinal sections with the help of the Axio Vision program (release 4.4; Carl Zeiss, Göttingen, Germany). These compound pictures were then edited with an image analysis system (KS 400 3.0; Carl Zeiss). Tendinous structures were removed by the investigator and the absolute area of red connective tissue was measured and normalized to the total muscle area. Statistics The arithmetic mean and SD were determined for each measure. Analysis of statistical significance was performed using the non-parametric Wilcoxon test for dependent samples when comparing measures intra-individually, and the non-parametric Mann-Whitney U test was used for independent values. The level of significance was set at 0.05.

Results Mechanical testing The maximally reached twitch contraction force remained almost constant with differences of less than 5% after 5 consecutive stimulations. Control muscles without trauma had a mean fast-twitch contraction force of 0.85 (SD 0.22) N. Stimulation of the sciatic nerve over a second yielded tetanic contractions in all the muscles tested, resulting in a maximum contraction force of 1.49 (SD 0.32) N in the control muscles,


Acta Orthopaedica 2011; 82 (1): 102â&#x20AC;&#x201C;111

105

Contraction force normalized to intact control (=1) 1.0

Fast twitch Tetany

0.8

0.6

0.4

0.2

0 1

4

8

Weeks after trauma Figure 2. Time course of muscle function. Contraction forces were normalized within individuals to the forces of uninjured control muscles.

which amounted to 175% of the value after fast-twitch stimulation (p < 0.001). During the evaluated regeneration period, a continuous increase in contraction force after fast-twitch stimulation and after tetanic stimulation could be observed, describing the functional regeneration of the traumatized soleus muscle over time (Figure 2). One week after crush trauma, the mean contraction forces after tetanic stimulation amounted to 0.34 (SD 0.05) N. After fast-twitch stimulation, almost the same maximum force (0.32 (SD 0.04) N) was reached, indicating that the injured muscles had no possibility to generate additional power through activation of reserve fibers at that time. Normalization against the right soleus muscle, which served as an internal control, showed a lag of regeneration of tetanic compared to twitch contraction force of 27% of the force generated by the healthy muscles 1 week after trauma (fast twitch 49% (SD 7), tetany 23% (SD 4) of the control side, p < 0.001). This lag of the tetanic contraction forces decreased over time, with the tetanic being 21% and 13% less than the normalized twitch contraction forces at weeks 4 and 8 after crush injury (week 4: fast twitch 61% (SD 13) of the control side, tetany 40% (SD 12), p = 0.028; week 8: fast twitch 68% (SD 19) of the control side, tetany 55% (SD 13), p = 0.028). This could be described by the correlation between the twitch/tetanus ratio and time after trauma (weeks), which showed a coefficient of correlation of 0.79 (p < 0.001). MRI findings Sequential MRI measurements showed a continuous time course of soleus muscle morphology. 1 week after trauma, T1 and T2 scans demonstrated a clear contrast between the injured muscles and the surrounding superficial and deep flexors, the latter presenting with low signal intensity (Figure 3). The signals of the soleus muscles showed pronounced inho-

Figure 3. Time course of T1 and T2 signals after crushing of soleus muscle. Axial scans of the left lower leg of rat 2 in the MRI group, 1, 4, and 8 weeks after trauma. The asterisk shows the tibia and the arrowhead shows the fibula. Arrows: soleus muscle.

mogeneities in both scan modes. The alterations observed affected the whole muscle tissue. In the T2-weighted images, interstitial edema could be detected within the soleus fascia and at the site of the approach. A posttraumatic seroma was located between the soleus and the gastrocnemius muscles. 4 weeks after injury, these seroma and the edematous changes within the fascia had been resorbed to a great extent. The elevated signal intensity of the soleus muscles had decreased, but there was still a clear contrast between the injured muscles and their surroundings. Also, the inhomogeneities within the traumatized muscle tissue had decreased. 8 weeks after injury, the soleus muscles were still slightly elevated in their signal intensity in the T2-weighted scans, whereas in the T1-weighted scans the signal had almost adapted to the surrounding muscle tissue. The signal elevation was homogenous. The mean signal elevations of the injured soleus muscles in the T2-weighted scans were quantified at each time point, as shown in Figure 4. A clear time course of an exponential decrease in the grayscale values over time could be discerned, with the differences between traumatized and healthy muscle tissue being only minimal for T1 signals 8 weeks after injury. Descriptive histology Sirius red-stained sections from the first day after injury showed red swollen, prenecrotic myofibers with intact boundaries on the one hand, and ruptured myofibers on the other, indicating the prevalence of the two main types of muscle injury: in-situ necrosis and shearing (Figure 5a). The fact that the basal laminae were injured in the latter case can be deduced from the myofiber ruptures initially observed (Figure 5b) and from the amount of isolated regenerating myofibers on the days studied thereafter (Figure 5d-f). What can also be observed in sections of the first and following days is the nature of the traumaâ&#x20AC;&#x201D;a global crush injury of the soleus


106

Acta Orthopaedica 2011; 82 (1): 102–111

Relative amount of collagenous connective tissue over time referenced to the whole area of the muscle sections in percent (SD)

Grayscale value 1200

T1 T2

Weeks after trauma 1 Area of collagenous connective tissue 40.4% (3.5)

1000

2 23.5% (7.1)

3 25.4% (7.3)

800 600 400 200 0 1

4

8

Weeks after trauma Figure 4. Time course of the differences in the grayscale values of soleus muscles and uninjured anterior tibial muscles 1, 4, and 8 weeks after soleus trauma in T1- and T2-weighted sequences. P-values: 1 week vs. 4 weeks, T1 + T2: p < 0.001; 1 week vs. 8 weeks, T1 + T2: p < 0.001; 4 weeks vs. 8 weeks, T1 and T2: p = 0.22 (ns) and p = 0.02, respectively.

muscle with a histologically obvious spared zone at the site of the neurovascular insertion (Figure 5a and d). The spared region was also affected by the trauma, as seen by the initial infiltration of inflammatory cells and interstitial hematoma, but this area showed a pronounced conservation of myofibers compared to the proximal and distal regions. This observation is shown in Figure 5 panels a, d and e, and can be interpreted according to the localization of the survival zone and regeneration zone after laceration trauma described in the work of Kääriäinen et al. (1998). The infiltration of inflammatory cells could be observed immediately after injury, showing a maximum on days 2 and 4 and remaining up to the end of the first week. At this time, most of the initial hematoma had already been removed and been replaced by loose connective tissue infiltrated mainly by monocytes and macrophages (Figure 6b). Centronucleated regenerating muscle fibers constituted the majority of the fibers at one week after trauma. The distribution of newly formed blood vessels was uniform, indicating that there was a homogenous regeneration process in the crush zones, as can be seen in a representative section stained for factor VIII (Figure 7a). Vessels were detected at day 4 after trauma since the quick onset of the angiogenic process in skeletal muscle after trauma peaked within the first 5 days (Jarvinen 1976, Jarvinen et al. 1983). The distribution of newly developed MTJs was also homogeneous and not confined to certain parts of the injured soleus. Figure 7b shows the myotendinous transition zone of a healthy soleus muscle, where immunoreactivity to nestin could be seen at the end zones of the soleus myofibers but there was no reactivity to a-bungarotoxin. In the crush-injured soleus muscles, nestin immunoreactivity was dispersed throughout the muscle and was not only found at the tips of regenerating

myofiber stumps but also at the lateral aspects of the fibers, indicating connections with the interstitially developed collagenous tissue and participation of the fibers in the contraction process. Co-reactivity to a-bungarotoxin was observed in only a few cases (Figure 7c). Figure 7d-f shows the a-bungarotoxinstained neural endplates 4 weeks after injury, which could be detected on myofibers disseminated in the crushed areas of the muscle. We also observed this 8 weeks after injury. Before injury, neural endplates could only be seen in a linear distribution in the region of the entry point of the neurovascular structures. 4 weeks after injury, inflammation had vanished and fibrotic tissue of a dense type had developed, although fibroblast nuclei still had a loose chromatin structure in many places, being a sign of activity. Apart from places of local scar formation, fibrotic tissue was evident especially in the interstitium between the myofibers (Figure 6c). 8 weeks after injury, maturation of connective tissue had already taken place, with spindle-shaped nuclei being the dominant fibroblast phenotype. Mature collagen fibers had replaced the loose reticular network to a large extent, as seen with polarization microscopy in slides stained with Sirius red. Few regenerating muscle fibers were still present. Mature muscle fibers were surrounded by an interstitial mesh of fibrotic tissue, representing endo- and perimysial collagen deposition (Figure 6d-f). Histomorphometry Histomorphometric evaluation showed that 1 week after severe crush injury, 40% of the muscles were occupied by collagenous connective tissue, representing the loose, still active tissue described above. After 4 weeks, this area was reduced to 24% (p = 0.05) and it was still at this value at week 8 after trauma (Table).

Discussion Experimental research on skeletal muscle trauma requires reliable animal models. These should afford the possibility of studying new therapeutic methods, especially with regard to their influence on functional and microstructural regeneration. We used a modified standardized blunt trauma of the rat soleus muscle to establish a model of a selective injury. The regeneration process of traumatized skeletal muscle tissue could therefore be analyzed and described over time from a


Acta Orthopaedica 2011; 82 (1): 102–111

Figure 5. Longitudinal cryo-sections of crush-injured soleus muscles stained with Picrosirius red. a. Overview showing the angular uninjured region at the middle of the section (from day 1 after injury). The insertion of the neurovascular bundle is indicated by an arrow. b. Detail of panel a showing ruptured myofibers as evidence of shearing-type injury. c. Detail of panel a showing swollen, prenecrotic myofibers as evidence of in situ necrosis happening in parallel to the latter type of injury. d. Overview of the uninjured zone and the distal crush zone of a soleus muscle from day 2 after crush injury. e. Detail of panel d showing the border zone of injured and uninjured muscle with an interstitial inflammatory reaction. f. Myofibers in the crush zone surrounded by inflammatory cells in loose connective tissue.

107

mechanical, histological, and MR-morphological point of view. In contrast to previously described whole soleus crush injuries (Bassaglia and Gautron 1995, Fink et al. 2003), the area of the neuromuscular junctions in the soleus muscle, a well-defined region at the midbelly of the muscle, was spared in order to create a model of a trauma selectively affecting the muscle fibers and the interstitial tissue without harming the main innervation of the organ. We therefore avoided an initial complete denervation, which would have made it impossible to differentiate between the consequences of denervation and of myofiber trauma. The advantage of the trauma model presented here is that with the greatest degree of possible myofiber injury without denervation injury, the effects of new therapies for skeletal muscle trauma can be analyzed with the lowest bias in the intrinsic regeneration capacities of the muscle itself with the maximum range for a possible therapeutic effect. Mechanical evaluation showed a tetanic contraction force—describing the capacity of force development of a muscle—of 1.5 N in the uninjured soleus muscles, which is in accordance with the literature (Anderson et al. 1999). In contrast to other settings (Meffert et al. 2008), the force measurement protocol used in this study allowed us to analyze the function of a single muscle without any interference from other musculature. 1 week after severe crush injury, contraction forces in both stimulation modes had almost the same values (tetany 0.3 N, fast twitch 0.3 N). Normalization of the forces to the uninjured internal control muscle (= 1.0) demonstrated that the injury had its greatest functional impact on tetanic contractions with a reduction to 0.23, while twitch contraction remained at almost half of the value of the healthy muscles after the first week. Two conclusions can be made from these findings: (1) facilitation of muscle force development was not possible for the damaged fibers, and (2) additional reserve fibers could not be recruited during repetitive stimulation of the muscle. Facilitation is an elementary phenomenon when muscle fibers are prompted with successive electrical stimuli. A second stimulus results in a markedly increased peak force when given at a short interval (optimally,


108

Acta Orthopaedica 2011; 82 (1): 102–111

Figure 6. Histological staining of soleus muscles 4 days and 1, 4, and 8 weeks after crush injury. a. Cryosection with HE staining of cross-sectioned myofibers surrounded by inflammatory cells on day 4 after injury. b. Paraffin section with HE staining of regenerating myofibers showing loose deposition of connective tissue with a high degree of cellularity and regenerating muscle fibers one week after trauma. c. Paraffin section with HE staining showing dense connective tissue with local scar formation and interstitial fibrosis 4 weeks after injury. d. Cryo-section with Sirius red staining: an overview of the injured soleus muscle 8 weeks after trauma showing typical interstitially pronounced fibrosis in the distal and proximal crush zone. e. Detail of panel d showing cross-sectioned myofibers surrounded by collagenous fibrotic tissue. f. Detail of panel d showing (longitudinally) sections of myofibers and interstitial fibrosis.

1.3–1.4 times the duration of the previous twitch) (Parmiggiani and Stein 1981). A prolongation and increase in local Ca2+ influx (Endo 1977) and increased muscle stiffness (Rubinstein et al. 1998) after the first twitch are thought to be responsible for the mechanism, both being properties of the muscle that were altered by the injury. In our model, both types of skeletal muscle injury—in situ necrosis and shearing—co-exist. The latter leads to the separation of myofibers, which results in split regenerating fibers (as also described by Schmalbruch (1976) after whole muscle crush and by Äärimaa et al. (2004) after transection injury). These myofibers have no contact with the neuromuscular junctions at the mid-belly of the soleus muscle and need to re-establish their contact with the nervous system. This could

be demonstrated in our model by the development of new neuromuscular endplates with a homogenous distribution within the crush area. Due to the character of the trauma (lack of a linear fibrotic border), we cannot deduce whether the nerve outgrowths pierced through fibrous tissue or whether they went along only mildly altered interstitial paths. The extent of the trauma and the amount of fibrosis indicate, however, that at least part had to establish contact via scar tissue. Posttraumatic soleus microstructure is dominated by the developing fibrosis. After transection or segmental crush injury, fibrosis is confined to a small area. The regenerating myofibers then develop new myotendinous junctions within this dense connective tissue (Hurme and Kalimo 1992), allowing the muscle almost complete functional regenera-


Acta Orthopaedica 2011; 82 (1): 102â&#x20AC;&#x201C;111

109

from the nerve supply (Garrett et al. 1984). The establishment of new myotendinous junctions of the regenerating myofibers with the surrounding fibrotic interstitium together with the development of new neuromuscular endplates shows that the fibers again participate in the contraction process after regeneration. The mechanical evaluation described the functional regeneration of the muscle, with an increase in fast-twitch strength to twothirds of that in the uninjured control muscle at 8 weeks after trauma, being only 13% higher than the normalized tetanic strength. This shows the recovery of the chemical and mechanical properties of the soleus muscle over time. However, contraction forces at 8 weeks in our model lacked half (tetany) and one third (fast twitch) of the forces of the uninjured muscle. Histological evaluation at this time showed that there was pronounced interstitial fibrosis, which differs from the local fibrosis described above. It can be deduced that not only contractile properties but also passive mechanical properties such as muscle stiffness are altered by this tight collagenous network, also affecting force development via reduced compliance of the muscle. Together with the replacement of contractile muscle substance with fibrotic tissue, this reduction in compliance can still be responsible for the residual lag of contraction forces 8 weeks after trauma. Although the character of the collagenous tissue changes, the amount appears to stay at a constant level of 25% of the muscle tissue. Regenerative therapies for supFigure 7. a. Factor VIII staining of a longitudinally sectioned soleus muscle 4 weeks after crush ported muscle healing should take injury, showing the distribution of vessels in the injury zones. b. Cryo-section with an immunohisthis into account, since not addressing tochemical stain for nestin in a healthy soleus muscle (counterstain: DAPI), myotendinous transifibrosis means not being able to fully tion zone. c. Staining for nestin in a longitudinally sectioned soleus muscle (counterstain: DAPI). Nestin immunoreactivity was dispersed throughout the muscle and could be found not only at reconstitute muscle forces after injury. the tips of regenerating myofiber stumps but also at the lateral aspects of the fibers. d. Overview The increase in contraction forces of a-bungarotoxin staining of a longitudinally sectioned soleus muscle (counterstain: DAPI). The during the observation period was white arrows indicate multiple, newly developed neural endplates in the proximal and distal crush zones of the muscle. e. Detail of panel d depicting a linear distribution of endplates in the uninjured accompanied by a decrease in T1 and region. f. Detail of panel d depicting endplates within the crushed area. T2 hyperintensity in the MR measurements. The time course of this develtion (Kaariainen et al. 1998, Menetrey et al. 1999), provided opment described the decrease in the acuity of the trauma. the distal part of the affected muscle has not been severed Morphologically, MR scans showed a disorganized aspect of


110

the muscle tissue one week after trauma with an inhomogeneity of voxel signals, which could only be observed to a small degree 4 weeks after trauma. This development corresponded well with the histological sections showing short regenerating muscle fibers separated by loose connective tissue at the first time investigated, and showing successive reorganization 4 and 8 weeks after trauma. Fibrotic tissue exhibits low signal intensity on all pulse sequences (Elsayes et al. 2006). 8 weeks after injury, the structure of the soleus muscles appeared to be homogenous and could hardly be differentiated from the surrounding muscles. This demonstrates that the interstitial fibrosis, which is one of the main problems for the contractile function, cannot be detected by standard MRI sequences, even when high field scanners are used. We believe that our model will be of value for studying the influence of therapeutic approaches at several different levels of skeletal muscle regeneration.

TW: design of the study, animal surgery, and preparation of the manuscript. PvR and CP: MRI investigations. GM: muscle force measurements. MRS, SH, and GS: histological analysis. PS: animal surgery. GND: statistical analysis. ST: design of the study and preparation of manuscript.

The authors thank Susanne Mueller, Camilla Bergmann, and Gabriela Korus for providing excellent technical assistance. The study was supported by a kick-off grant from the Berlin-Brandenburg Center for Regenerative Therapies.

No competing interests declared.

Aarimaa V, Kaariainen M, Vaittinen S, Tanner J, Jarvinen T, Best T, et al. Restoration of myofiber continuity after transection injury in the rat soleus. Neuromuscul Disord 2004; 14 (7): 421-8. Anderson J, Almeida-Silveira M I, Perot C. Reflex and muscular adaptations in rat soleus muscle after hindlimb suspension. J Exp Biol 1999; 202 (Pt 19): 2701-7. Bassaglia Y, Gautron J. Fast and slow rat muscles degenerate and regenerate differently after whole crush injury. J Muscle Res Cell Motil 1995; 16 (4): 420-9. Broniatowski M, Azar K, Davies C R, Jacobs G B, Tucker H M. Selective feedback control of spastic musculature in a canine model. Asaio J 1992; 3 (3): M248-52. Bunn J R, Canning J, Burke G, Mushipe M, Marsh D R, Li G. Production of consistent crush lesions in murine quadriceps muscle--a biomechanical, histomorphological and immunohistochemical study. J Orthop Res 2004; 22 (6): 1336-44. Dedkov E I, Kostrominova T Y, Borisov A B, Carlson B M. Reparative myogenesis in long-term denervated skeletal muscles of adult rats results in a reduction of the satellite cell population. Anat Rec 2001; 263 (2): 139-54. Duda G N, Bartmeyer B, Sporrer S, Taylor W R, Raschke M, Haas N P. Does partial weight bearing unload a healing bone in external ring fixation? Langenbecks Arch Surg 2003; 388 (5): 298-304. Elsayes K M, LammLe M, Shariff A, Totty W G, Habib I F, Rubin D A. Value of magnetic resonance imaging in muscle trauma. Curr Probl Diagn Radiol 2006; 35 (5): 206-12.

Acta Orthopaedica 2011; 82 (1): 102â&#x20AC;&#x201C;111

Endo M. Calcium release from the sarcoplasmic reticulum. Physiol Rev 1977; 57 (1): 71-108. Fink E, Fortin D, Serrurier B, Ventura-Clapier R, Bigard A X. Recovery of contractile and metabolic phenotypes in regenerating slow muscle after notexin-induced or crush injury. J Muscle Res Cell Motil 2003; 24 (7): 421-9. Garrett W E, Jr., Seaber A V, Boswick J, Urbaniak J R, Goldner J L. Recovery of skeletal muscle after laceration and repair. J Hand Surg (Am) 1984; 9 (5): 683-92. Harry L E, Sandison A, Paleolog E M, Hansen U, Pearse M F, Nanchahal J. Comparison of the healing of open tibial fractures covered with either muscle or fasciocutaneous tissue in a murine model. J Orthop Res 2008; 26 (9): 1238-44. Hurme T, Kalimo H. Adhesion in skeletal muscle during regeneration. Muscle Nerve 1992; 15 (4): 482-9. Jarvinen M. Healing of a crush injury in rat striated muscle. 3. A microangiogra-phical study of the effect of early mobilization and immobilization on capillary ingrowth. Acta Pathol Microbiol Scand 1976; 84A: 85-94. Jarvinen M, Sorvari T. Healing of a crush injury in rat striated muscle. 1. Description and testing of a new method of inducing a standard injury to the calf muscles. Acta Pathol Microbiol Scand (A). 1975; 83 (2): 259-65. Jarvinen M, Aho A J, Lehto M, Toivonen H. Age dependent repair of muscle rupture. A histological and microangiographical study in rats. Acta Orthop Scand 1983; 54: 64-74. Kaariainen M, Kaariainen J, Jarvinen T L, Sievanen H, Kalimo H, Jarvinen M. Correlation between biomechanical and structural changes during the regeneration of skeletal muscle after laceration injury. J Orthop Res 1998; 16 (2): 197-206. Kleemann R U, Heller M O, Stoeckle U, Taylor W R, Duda G N. THA loading arising from increased femoral anteversion and offset may lead to critical cement stresses. J Orthop Res 2003; 21 (5): 767-74. Koishi K, Zhang M, McLennan I S, Harris A J. MyoD protein accumulates in satellite cells and is neurally regulated in regenerating myotubes and skeletal muscle fibers. Dev Dyn 1995; 202 (3): 244-54. Kurek J B, Bower J J, Romanella M, Koentgen F, Murphy M, Austin L. The role of leukemia inhibitory factor in skeletal muscle regeneration. Muscle Nerve 1997; 20 (7): 815-22. McGeachie J K, Grounds M D. Initiation and duration of muscle precursor replication after mild and severe injury to skeletal muscle of mice. An autoradiographic study. Cell Tissue Res 1987; 248 (1): 125-30. Meffert R H, Frey S P, Jansen H, Ochman S, Raschke M J, Langer M. Muscle strength quantification in small animals: a new transcutaneous technique in rabbits. J Orthop Res 2008; 26 (11): 1526-32. Menetrey J, Kasemkijwattana C, Fu F H, Moreland M S, Huard J. Suturing versus immobilization of a muscle laceration. A morphological and functional study in a mouse model. Am J Sports Med 1999; 27 (2): 222-9. Negishi S, Li Y, Usas A, Fu F H, Huard J. The effect of relaxin treatment on skeletal muscle injuries. Am J Sports Med 2005; 33 (12): 1816-24. Nozaki M, Li Y, Zhu J, Ambrosio F, Uehara K, Fu F H, et al. Improved muscle healing after contusion injury by the inhibitory effect of suramin on myostatin, a negative regulator of muscle growth. Am J Sports Med 2008; 36 (12): 2354-62. Parmiggiani F, Stein R B. Nonlinear summation of contractions in cat muscles. II. Later facilitation and stiffness changes. J Gen Physiol 1981; 78 (3): 295-311. Pereira B P, Tan J A, Zheng L, Tan B L, Lahiri A, Lim A Y, et al. The cut intramuscular nerve affects the recovery in the lacerated skeletal muscle. J Orthop Res 2006; 24 (1): 102-11. Perka C, Heller M, Wilke K, Taylor W R, Haas N P, Zippel H, et al. Surgical approach influences periprosthetic femoral bone density. Clin Orthop 2005; (432): 153-9. Rubinstein I, Abassi Z, Coleman R, Milman F, Winaver J, Better O S. Involvement of nitric oxide system in experimental muscle crush injury. J Clin Invest 1998; 101 (6): 1325-33.


Acta Orthopaedica 2011; 82 (1): 102â&#x20AC;&#x201C;111

Rushton J L, Davies I, Horan M A, Mahon M, Williams R. Production of consistent crush lesions of murine skeletal muscle in vivo using an electromechanical device. J Anat 1997; 190 ( Pt 3): 417-22. Sato K, Li Y, Foster W, Fukushima K, Badlani N, Adachi N, et al. Improvement of muscle healing through enhancement of muscle regeneration and prevention of fibrosis. Muscle Nerve 2003; 28 (3): 365-72. Saunders J H, Sissons H A. The effect of denervation on the regeneration of skeletal muscle after injury. J Bone Joint Surg (Br) 1953; 35 (1): 113-24. Schaser K D, Zhang L, Haas N P, Mittlmeier T, Duda G, Bail H J. Temporal profile of microvascular disturbances in rat tibial periosteum following closed soft tissue trauma. Langenbecks Arch Surg. 2003; 388 (5): 323-30. Schmalbruch H. The morphology of regeneration of skeletal muscles in the rat. Tissue Cell 1976; 8 (4): 673-92. Schultz E, Jaryszak D L, Valliere C R. Response of satellite cells to focal skeletal muscle injury. Muscle Nerve 1985; 8 (3): 217-22. Silverman D G, Brull S J. The effect of a tetanic stimulus on the response to subsequent tetanic stimulation. Anesth Analg 1993; 76 (6): 1284-7.

111

Sorokin L M, Maley M A, Moch H, von der Mark H, von der Mark K, Cadalbert L, et al. Laminin alpha4 and integrin alpha6 are upregulated in regenerating dy/dy skeletal muscle: comparative expression of laminin and integrin isoforms in muscles regenerating after crush injury. Exp Cell Res 2000; 256 (2): 500-14. Sweat F, Puchtler H, Rosenthal S I. Sirius Red F3ba as a stain for connective tissue. Arch Pathol 1964; 78: 69-72. Utvag S E, Grundnes O, Rindal D B, Reikeras O. Influence of extensive muscle injury on fracture healing in rat tibia. J Orthop Trauma 2003; 17 (6): 430-5. Wake R, Kim-Mitsuyama S, Izumi Y, Yoshida K, Izumiya Y, Yukimura T, et al. Beneficial effect of candesartan on rat diastolic heart failure. J Pharmacol Sci 2005; 98 (4): 372-9. Winkler T, von Roth P, Matziolis G, Mehta M, Perka C, Duda G N. Doseresponse relationship of mesenchymal stem cell transplantation and functional regeneration after severe skeletal muscle injury in rats. Tissue Eng Part A 2009; 15 (3): 487-92.

.


102

Acta Orthopaedica 2011; 82 (1): 102–111

Time course of skeletal muscle regeneration after severe trauma Muscle function against the background of MRI and histological findings Tobias Winkler1, Philipp von Roth1, Georg Matziolis1, Maria R Schumann1, Sebastian Hahn1, Patrick Strube1, Gisela Stoltenburg-Didinger2, Carsten Perka1, Georg N Duda1, and Stephan V Tohtz1 1Center for Musculoskeletal Surgery and Julius Wolff Institute Berlin, Brandenburg Center for Regenerative Therapies; and 2Institute of Neuropathology, Charité – Universitaetsmedizin Berlin, Berlin, Germany Correspondence: tobias.winkler@charite.de Submitted 09-09-13. Accepted 10-07-27.

Background and purpose Animal models of skeletal muscle injury should be thoroughly described and should mimic the clinical situation. We established a model of a critical size crush injury of the soleus muscle in rats. The aim was to describe the time course of skeletal muscle regeneration using mechanical, histological, and magnetic resonance (MR) tomographic methods. Methods Left soleus muscles of 36 Sprague-Dawley rats were crushed in situ in a standardized manner. We scanned the lower legs of 6 animals by 7-tesla MR one week, 4 weeks, and 8 weeks after trauma. Regeneration was evaluated at these times by in vivo measurement of muscle contraction forces after fast-twitch and tetanic stimulation (groups 1W, 4W, 8W; 6 per group). Histological and immunohistological analysis was performed and the amount of fibrosis within the injured muscles was determined histomorphologically. Results MR signals of the traumatized soleus muscles showed a clear time course concerning microstructure and T1 and T2 signal intensity. Newly developed neural endplates and myotendinous junctions could be seen in the injured zones of the soleus. Tetanic force increased continuously, starting at 23% (SD 4) of the control side (p < 0.001) 1 week after trauma and recovering to 55% (SD 23) after 8 weeks. Fibrotic tissue occupied 40% (SD 4) of the traumatized muscles after the first week, decreased to approximately 25% after 4 weeks, and remained at this value until 8 weeks. Interpretation At both the functional level and the morphological level, skeletal muscle regeneration follows a distinct time course. Our trauma model allows investigation of muscle regeneration after a standardized injury to muscle fibers. 

Injuries of skeletal muscle tissue cause deficiencies in local muscle function that can also affect bone healing (Duda et

al. 2003, Schaser et al. 2003, Utvag et al. 2003, Harry et al. 2008) or the long-term success of a prosthesis (Kleemann et al. 2003, Perka et al. 2005). Despite the fact that there have been several experimental approaches, no methods for the treatment of the main causes of these deficiencies—i.e. the loss of contractile muscle substance and the formation of fibrosis—have been taken into routine clinical use. One reason is that the results of putative therapies are often only described mono-dimensionally, in most cases histologically, and rarely functionally or with in vivo diagnostic methods. Animal models of muscle contusion injuries should closely mimic the clinical situation. Among them, open crush injuries allow standardized evaluation of regeneration in a selected muscle (Schultz et al. 1985, Rushton et al. 1997). For the execution of the trauma, either forceps (McGeachie and Grounds 1987, Kurek et al. 1997, Fink et al. 2003) or custom-made devices have been used (Jarvinen and Sorvari 1975, Rushton et al. 1997, Bunn et al. 2004). Previous muscle-crush models have had the disadvantage of either affecting only part of the muscle or of impairing the organ innervation and blood supply. Two types can be found in the literature: the segmental crush (Schmalbruch 1976) and the complete crush, where only 4–6% of the muscle fibers remain intact (Fink et al. 2003). In the latter, myoneuronal junctions are damaged, which triggers not only regeneration of muscle substance but also initial innervation deficits. These deficits always lead to impaired healing (Saunders and Sissons 1953, Koishi et al. 1995, Dedkov et al. 2001, Pereira et al. 2006). Histological analysis of the regenerative process after crush injury has been done by a number of authors, describing the initial phase of inflammation followed by satellite cell activation, myotube regeneration, and fibrosis of the muscle (Jarvinen and Sorvari 1975, Schmalbruch 1976, Sorokin et al.

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2010.539498


Acta Orthopaedica 2011; 82 (1): 102–111

2000, Bunn et al. 2004). It has been shown that the development of fibrotic tissue is one of the key factors for regenerative deficits of muscle function after trauma (Anderson et al. 1999, Nozaki et al. 2008). The development of intramuscular fibrosis over time after crush trauma has not, however, been quantitatively described. Magnetic resonance imaging (MRI) can be used in the clinical evaluation of skeletal muscle injuries. Not only are T1and T2- weighted spin echo sequences highly sensitive for detection of edema and bleeding in anatomical relation to the structures affected (Elsayes et al. 2006), but MR scans also allow the determination of intra-individual time courses after trauma. The evaluation of muscle regeneration by contraction force measurements provides information about the functional outcome of therapeutic approaches (Sato et al. 2003, Negishi et al. 2005, Winkler et al. 2009). The description of a muscle trauma model should therefore not solely include morphological but also mechanical data. One aim of our study was to establish a standardized model of selective trauma of the muscle substance of a single muscle, leaving the blood and nerve supply undamaged. This was achieved by developing a global crush injury of the rat soleus muscle with respect to the topographic arrangement of the innervating endplate zone. We also wanted to describe the regeneration of muscle contraction forces against the background of morphological changes within the muscle tissue following a critical size of injury in a skeletal muscle.

Material and methods Animals 36 male Sprague Dawley rats weighing 450–550g were used for the study. The rats were housed at a constant temperature of 25°C with free access to pellet food and water. All animal experiments were carried out according to the policies and principles established by the Animal Welfare Act, the NIH Guide for Care and Use of Laboratory Animals, and the German national animal welfare guidelines. The study was approved by the local institute of health. Experimental procedure All animals received an open crush trauma of the left soleus muscles. 1, 4, and 8 weeks after injury, muscle contraction forces were measured in vivo and the animals were killed (groups 1W, 4W, and 8W; n = 6 per group). The soleus muscles of both sides were harvested for histological evaluation. Soleus muscles of 4 animals were harvested on days 1, 2, and 4 after trauma for histological evaluation of the first days after injury. 6 animals were examined by sequential MRI measurements at 1, 4, and 8 weeks (group MRI, n = 6).

103

Figure 1. Innervation of soleus muscle and vascular supply. Arrowheads: lateral border of soleus muscle (SO); arrow: neurovascular bundle coming from the gastrocnemius muscle (GM) and entering at the medial border of the SO, the tip of the arrow indicating the soleus nerve; PE: peroneal muscles. The asterisk shows the Achilles tendon.

Muscle trauma The rats were anesthetized by subcutaneous injections of 0.15 mL ketamine (100 mg/mL) and 0.15 mL xylazine (2%) diluted with 0.2 mL 0.9% saline. The left leg was shaved and disinfected with povidone-iodine. Through a 2-cm posterolateral longitudinal skin incision from the lateral gastrocnemius head to the Achilles tendon, the soleus muscle was mobilized and the soleus artery, vein and nerve, which arise from the midpart of the gastrocnemius, were located. The medial border of the soleus muscle was dissected cranially and caudally to the neurovascular structures. A curved artery forceps, the jaws of which were protected by polyethylene tubes to avoid lesions of the muscle fascia, was introduced in direct proximity to the Achilles tendon and closed for 20 seconds. This procedure was repeated proximally 3 times, always in direct continuity to the respective distal crush. The insertion of the neurovascular bundle and a trapezoid region with a 3-mm long base at the insertion of the bundle and a 2-mm short base at the lateral margin of the soleus muscle was spared and 3 crushes were exerted proximal to this area in the manner described above (Figure 1). Standardized pressure on the muscle was ensured by closing the forceps to its third stage at each crush. The closing pressure at this stage corresponds to 112 (SD 5.1) N (data obtained from preliminary tests with the material testing device Zwick 1455 (Zwick GmbH, Ulm, Germany). After multiple irrigations, the superficial muscle and skin were closed. Measurement of muscle force The animals were anesthetized as described before and received bilateral surgery. The sciatic nerve and the soleus muscle were exposed, sparing all neurovascular structures.


104

The Achilles tendon was cut and the lower extremity was connected to a muscle force measuring device (Experimentria, Budapest, Hungary). The distal part of the soleus muscle was connected to a force transducer via a suture (4-0, silk). The sciatic nerve was subsequently stimulated with 9 mA/75 Hz bipolar pulses 5 times, for 0.1 second each (8 periods) with 5-second intervals between the pulses (Broniatowski et al. 1992, Silverman and Brull 1993). After this short twitch stimulation, the maximal muscle strength was measured using a protocol of 9 mA/75 Hz pulses 5 times, for 3 seconds each at 5-second intervals, reaching tetany in all cases. After completion of the muscle strength measurements, the animals were killed with an overdose of anesthetic and the soleus muscles were fixed in buffered paraformaldehyde (4%) for histological examination. MRI 6 animals were examined by MRI one, 4, and 8 weeks after muscle injury. Measurements were taken with a 7-tesla MRI spectrometer for small animals (Pharmascan 70/16; Bruker, Ettlingen, Germany). Images were processed with ParaVision software (Bruker BioSpin). Rats were anesthetized by isofluran/O2 inhalation and were introduced into the spectrometer in a special animal handling system equipped with a heating pad and continuous monitoring (ECG and respiratory frequency). A rat brain coil with an inner diameter of 38 mm was used for transmission and reception. Each measurement involved one lower limb of the animal from the tibial plateau to the ankle joint. T1 (TR/TE: 1000/10.6 ms; scan time 12 min) and T2-turbo-rare sequences (TR/TE: 6351.6/75.6 ms; scan time 12 min) were performed with a resolution of 176 × 176 µm and a slice thickness of 500 µm. Analysis of grayscale values was performed with ImageJ version 1.38i (National Institutes of Health, Bethesda, MD). The signal intensities of the voxels of the injured soleus muscles were quantified in the T1- and T2-weighted scans and compared to those of an uninjured reference muscle (anterior tibial muscle). The differences between the mean grayscale values of the soleus muscles were calculated for each time point (Figure 4). Histology Hematoxylin and eosin (HE) staining and Picro-sirius red staining were used for descriptive analysis of the trauma. The stain Picro-sirius red selectively highlights collagenous connective tissues with a well-defined contrast. This stain was therefore used for quantitative analysis of fibrosis in the traumatized muscles, as previously described (Sweat et al. 1964, Wake et al. 2005). Furthermore, muscles were stained for nestin to localize the positioning of myotendinous junctions (MTJs) before and after trauma (Aarimaa et al. 2004) and with a-bungarotoxin to visualize the distribution of neuromuscular junctions within the soleus muscle. The endothelial marker factor VIII was used for the detection of vessels.

Acta Orthopaedica 2011; 82 (1): 102–111

Soleus muscles were either paraformaldehyde-fixed, dehydrated, and embedded in paraffin or fresh-frozen in liquid nitrogen-cooled 2-methylbutan and embedded in Tissue Tek. The specimens were sectioned longitudinally (4 and 10 µm). For evaluation of connective tissue, sections were incubated for 60 min in Sirius red solution (5g Sirius red dissolved in 500 mL saturated picric acid). Differentiation was achieved by 2 washes with diluted acetic acid followed by a short dehydration in a series of graded alcohols. The following antisera and antibodies were used for immunohistochemistry: mouse monoclonal antibody to rat nestin (1:800)  (BD Biosciences Pharmingen, Franklin Lakes, NJ) and mouse monoclonal antibody to factor VIII (1:400) (Daco Cytomation, Glostrup, Denmark). For light microscopy, the primary antibodies were visualized using appropriate avidin–biotin–peroxidase kits (Vector Laboratories, Burlingame, CA) and counterstained with hematoxylin. For fluorescence microscopy, Alexa Fluor 546-conjugated goat anti-mouse antibody (1:300) (Invitrogen Corporation, Camarillo, CA) was used as secondary antibody. DAPI was used as counterstain for all fluorescence stains. For endplate staining, the frozen sections were incubated for 2 h at room temperature with Alexa Fluor 488-conjugated a-bungarotoxin (2 µg/mL) (Invitrogen). A blinded investigator evaluated the amount of collagenous connective tissue in the muscles. In order to measure the total area of endo- and perimysial fibrosis digitally, microscopic images of whole longitudinal sections of the muscle were collected from the distal to the proximal tendon. Images were correlated and connected to obtain the whole longitudinal sections with the help of the Axio Vision program (release 4.4; Carl Zeiss, Göttingen, Germany). These compound pictures were then edited with an image analysis system (KS 400 3.0; Carl Zeiss). Tendinous structures were removed by the investigator and the absolute area of red connective tissue was measured and normalized to the total muscle area. Statistics The arithmetic mean and SD were determined for each measure. Analysis of statistical significance was performed using the non-parametric Wilcoxon test for dependent samples when comparing measures intra-individually, and the non-parametric Mann-Whitney U test was used for independent values. The level of significance was set at 0.05.

Results Mechanical testing The maximally reached twitch contraction force remained almost constant with differences of less than 5% after 5 consecutive stimulations. Control muscles without trauma had a mean fast-twitch contraction force of 0.85 (SD 0.22) N. Stimulation of the sciatic nerve over a second yielded tetanic contractions in all the muscles tested, resulting in a maximum contraction force of 1.49 (SD 0.32) N in the control muscles,


Acta Orthopaedica 2011; 82 (1): 102â&#x20AC;&#x201C;111

105

Contraction force normalized to intact control (=1) 1.0

Fast twitch Tetany

0.8

0.6

0.4

0.2

0 1

4

8

Weeks after trauma Figure 2. Time course of muscle function. Contraction forces were normalized within individuals to the forces of uninjured control muscles.

which amounted to 175% of the value after fast-twitch stimulation (p < 0.001). During the evaluated regeneration period, a continuous increase in contraction force after fast-twitch stimulation and after tetanic stimulation could be observed, describing the functional regeneration of the traumatized soleus muscle over time (Figure 2). One week after crush trauma, the mean contraction forces after tetanic stimulation amounted to 0.34 (SD 0.05) N. After fast-twitch stimulation, almost the same maximum force (0.32 (SD 0.04) N) was reached, indicating that the injured muscles had no possibility to generate additional power through activation of reserve fibers at that time. Normalization against the right soleus muscle, which served as an internal control, showed a lag of regeneration of tetanic compared to twitch contraction force of 27% of the force generated by the healthy muscles 1 week after trauma (fast twitch 49% (SD 7), tetany 23% (SD 4) of the control side, p < 0.001). This lag of the tetanic contraction forces decreased over time, with the tetanic being 21% and 13% less than the normalized twitch contraction forces at weeks 4 and 8 after crush injury (week 4: fast twitch 61% (SD 13) of the control side, tetany 40% (SD 12), p = 0.028; week 8: fast twitch 68% (SD 19) of the control side, tetany 55% (SD 13), p = 0.028). This could be described by the correlation between the twitch/tetanus ratio and time after trauma (weeks), which showed a coefficient of correlation of 0.79 (p < 0.001). MRI findings Sequential MRI measurements showed a continuous time course of soleus muscle morphology. 1 week after trauma, T1 and T2 scans demonstrated a clear contrast between the injured muscles and the surrounding superficial and deep flexors, the latter presenting with low signal intensity (Figure 3). The signals of the soleus muscles showed pronounced inho-

Figure 3. Time course of T1 and T2 signals after crushing of soleus muscle. Axial scans of the left lower leg of rat 2 in the MRI group, 1, 4, and 8 weeks after trauma. The asterisk shows the tibia and the arrowhead shows the fibula. Arrows: soleus muscle.

mogeneities in both scan modes. The alterations observed affected the whole muscle tissue. In the T2-weighted images, interstitial edema could be detected within the soleus fascia and at the site of the approach. A posttraumatic seroma was located between the soleus and the gastrocnemius muscles. 4 weeks after injury, these seroma and the edematous changes within the fascia had been resorbed to a great extent. The elevated signal intensity of the soleus muscles had decreased, but there was still a clear contrast between the injured muscles and their surroundings. Also, the inhomogeneities within the traumatized muscle tissue had decreased. 8 weeks after injury, the soleus muscles were still slightly elevated in their signal intensity in the T2-weighted scans, whereas in the T1-weighted scans the signal had almost adapted to the surrounding muscle tissue. The signal elevation was homogenous. The mean signal elevations of the injured soleus muscles in the T2-weighted scans were quantified at each time point, as shown in Figure 4. A clear time course of an exponential decrease in the grayscale values over time could be discerned, with the differences between traumatized and healthy muscle tissue being only minimal for T1 signals 8 weeks after injury. Descriptive histology Sirius red-stained sections from the first day after injury showed red swollen, prenecrotic myofibers with intact boundaries on the one hand, and ruptured myofibers on the other, indicating the prevalence of the two main types of muscle injury: in-situ necrosis and shearing (Figure 5a). The fact that the basal laminae were injured in the latter case can be deduced from the myofiber ruptures initially observed (Figure 5b) and from the amount of isolated regenerating myofibers on the days studied thereafter (Figure 5d-f). What can also be observed in sections of the first and following days is the nature of the traumaâ&#x20AC;&#x201D;a global crush injury of the soleus


106

Acta Orthopaedica 2011; 82 (1): 102–111

Relative amount of collagenous connective tissue over time referenced to the whole area of the muscle sections in percent (SD)

Grayscale value 1200

T1 T2

Weeks after trauma 1 Area of collagenous connective tissue 40.4% (3.5)

1000

2 23.5% (7.1)

3 25.4% (7.3)

800 600 400 200 0 1

4

8

Weeks after trauma Figure 4. Time course of the differences in the grayscale values of soleus muscles and uninjured anterior tibial muscles 1, 4, and 8 weeks after soleus trauma in T1- and T2-weighted sequences. P-values: 1 week vs. 4 weeks, T1 + T2: p < 0.001; 1 week vs. 8 weeks, T1 + T2: p < 0.001; 4 weeks vs. 8 weeks, T1 and T2: p = 0.22 (ns) and p = 0.02, respectively.

muscle with a histologically obvious spared zone at the site of the neurovascular insertion (Figure 5a and d). The spared region was also affected by the trauma, as seen by the initial infiltration of inflammatory cells and interstitial hematoma, but this area showed a pronounced conservation of myofibers compared to the proximal and distal regions. This observation is shown in Figure 5 panels a, d and e, and can be interpreted according to the localization of the survival zone and regeneration zone after laceration trauma described in the work of Kääriäinen et al. (1998). The infiltration of inflammatory cells could be observed immediately after injury, showing a maximum on days 2 and 4 and remaining up to the end of the first week. At this time, most of the initial hematoma had already been removed and been replaced by loose connective tissue infiltrated mainly by monocytes and macrophages (Figure 6b). Centronucleated regenerating muscle fibers constituted the majority of the fibers at one week after trauma. The distribution of newly formed blood vessels was uniform, indicating that there was a homogenous regeneration process in the crush zones, as can be seen in a representative section stained for factor VIII (Figure 7a). Vessels were detected at day 4 after trauma since the quick onset of the angiogenic process in skeletal muscle after trauma peaked within the first 5 days (Jarvinen 1976, Jarvinen et al. 1983). The distribution of newly developed MTJs was also homogeneous and not confined to certain parts of the injured soleus. Figure 7b shows the myotendinous transition zone of a healthy soleus muscle, where immunoreactivity to nestin could be seen at the end zones of the soleus myofibers but there was no reactivity to a-bungarotoxin. In the crush-injured soleus muscles, nestin immunoreactivity was dispersed throughout the muscle and was not only found at the tips of regenerating

myofiber stumps but also at the lateral aspects of the fibers, indicating connections with the interstitially developed collagenous tissue and participation of the fibers in the contraction process. Co-reactivity to a-bungarotoxin was observed in only a few cases (Figure 7c). Figure 7d-f shows the a-bungarotoxinstained neural endplates 4 weeks after injury, which could be detected on myofibers disseminated in the crushed areas of the muscle. We also observed this 8 weeks after injury. Before injury, neural endplates could only be seen in a linear distribution in the region of the entry point of the neurovascular structures. 4 weeks after injury, inflammation had vanished and fibrotic tissue of a dense type had developed, although fibroblast nuclei still had a loose chromatin structure in many places, being a sign of activity. Apart from places of local scar formation, fibrotic tissue was evident especially in the interstitium between the myofibers (Figure 6c). 8 weeks after injury, maturation of connective tissue had already taken place, with spindle-shaped nuclei being the dominant fibroblast phenotype. Mature collagen fibers had replaced the loose reticular network to a large extent, as seen with polarization microscopy in slides stained with Sirius red. Few regenerating muscle fibers were still present. Mature muscle fibers were surrounded by an interstitial mesh of fibrotic tissue, representing endo- and perimysial collagen deposition (Figure 6d-f). Histomorphometry Histomorphometric evaluation showed that 1 week after severe crush injury, 40% of the muscles were occupied by collagenous connective tissue, representing the loose, still active tissue described above. After 4 weeks, this area was reduced to 24% (p = 0.05) and it was still at this value at week 8 after trauma (Table).

Discussion Experimental research on skeletal muscle trauma requires reliable animal models. These should afford the possibility of studying new therapeutic methods, especially with regard to their influence on functional and microstructural regeneration. We used a modified standardized blunt trauma of the rat soleus muscle to establish a model of a selective injury. The regeneration process of traumatized skeletal muscle tissue could therefore be analyzed and described over time from a


Acta Orthopaedica 2011; 82 (1): 102–111

Figure 5. Longitudinal cryo-sections of crush-injured soleus muscles stained with Picrosirius red. a. Overview showing the angular uninjured region at the middle of the section (from day 1 after injury). The insertion of the neurovascular bundle is indicated by an arrow. b. Detail of panel a showing ruptured myofibers as evidence of shearing-type injury. c. Detail of panel a showing swollen, prenecrotic myofibers as evidence of in situ necrosis happening in parallel to the latter type of injury. d. Overview of the uninjured zone and the distal crush zone of a soleus muscle from day 2 after crush injury. e. Detail of panel d showing the border zone of injured and uninjured muscle with an interstitial inflammatory reaction. f. Myofibers in the crush zone surrounded by inflammatory cells in loose connective tissue.

107

mechanical, histological, and MR-morphological point of view. In contrast to previously described whole soleus crush injuries (Bassaglia and Gautron 1995, Fink et al. 2003), the area of the neuromuscular junctions in the soleus muscle, a well-defined region at the midbelly of the muscle, was spared in order to create a model of a trauma selectively affecting the muscle fibers and the interstitial tissue without harming the main innervation of the organ. We therefore avoided an initial complete denervation, which would have made it impossible to differentiate between the consequences of denervation and of myofiber trauma. The advantage of the trauma model presented here is that with the greatest degree of possible myofiber injury without denervation injury, the effects of new therapies for skeletal muscle trauma can be analyzed with the lowest bias in the intrinsic regeneration capacities of the muscle itself with the maximum range for a possible therapeutic effect. Mechanical evaluation showed a tetanic contraction force—describing the capacity of force development of a muscle—of 1.5 N in the uninjured soleus muscles, which is in accordance with the literature (Anderson et al. 1999). In contrast to other settings (Meffert et al. 2008), the force measurement protocol used in this study allowed us to analyze the function of a single muscle without any interference from other musculature. 1 week after severe crush injury, contraction forces in both stimulation modes had almost the same values (tetany 0.3 N, fast twitch 0.3 N). Normalization of the forces to the uninjured internal control muscle (= 1.0) demonstrated that the injury had its greatest functional impact on tetanic contractions with a reduction to 0.23, while twitch contraction remained at almost half of the value of the healthy muscles after the first week. Two conclusions can be made from these findings: (1) facilitation of muscle force development was not possible for the damaged fibers, and (2) additional reserve fibers could not be recruited during repetitive stimulation of the muscle. Facilitation is an elementary phenomenon when muscle fibers are prompted with successive electrical stimuli. A second stimulus results in a markedly increased peak force when given at a short interval (optimally,


108

Acta Orthopaedica 2011; 82 (1): 102–111

Figure 6. Histological staining of soleus muscles 4 days and 1, 4, and 8 weeks after crush injury. a. Cryosection with HE staining of cross-sectioned myofibers surrounded by inflammatory cells on day 4 after injury. b. Paraffin section with HE staining of regenerating myofibers showing loose deposition of connective tissue with a high degree of cellularity and regenerating muscle fibers one week after trauma. c. Paraffin section with HE staining showing dense connective tissue with local scar formation and interstitial fibrosis 4 weeks after injury. d. Cryo-section with Sirius red staining: an overview of the injured soleus muscle 8 weeks after trauma showing typical interstitially pronounced fibrosis in the distal and proximal crush zone. e. Detail of panel d showing cross-sectioned myofibers surrounded by collagenous fibrotic tissue. f. Detail of panel d showing (longitudinally) sections of myofibers and interstitial fibrosis.

1.3–1.4 times the duration of the previous twitch) (Parmiggiani and Stein 1981). A prolongation and increase in local Ca2+ influx (Endo 1977) and increased muscle stiffness (Rubinstein et al. 1998) after the first twitch are thought to be responsible for the mechanism, both being properties of the muscle that were altered by the injury. In our model, both types of skeletal muscle injury—in situ necrosis and shearing—co-exist. The latter leads to the separation of myofibers, which results in split regenerating fibers (as also described by Schmalbruch (1976) after whole muscle crush and by Äärimaa et al. (2004) after transection injury). These myofibers have no contact with the neuromuscular junctions at the mid-belly of the soleus muscle and need to re-establish their contact with the nervous system. This could

be demonstrated in our model by the development of new neuromuscular endplates with a homogenous distribution within the crush area. Due to the character of the trauma (lack of a linear fibrotic border), we cannot deduce whether the nerve outgrowths pierced through fibrous tissue or whether they went along only mildly altered interstitial paths. The extent of the trauma and the amount of fibrosis indicate, however, that at least part had to establish contact via scar tissue. Posttraumatic soleus microstructure is dominated by the developing fibrosis. After transection or segmental crush injury, fibrosis is confined to a small area. The regenerating myofibers then develop new myotendinous junctions within this dense connective tissue (Hurme and Kalimo 1992), allowing the muscle almost complete functional regenera-


Acta Orthopaedica 2011; 82 (1): 102â&#x20AC;&#x201C;111

109

from the nerve supply (Garrett et al. 1984). The establishment of new myotendinous junctions of the regenerating myofibers with the surrounding fibrotic interstitium together with the development of new neuromuscular endplates shows that the fibers again participate in the contraction process after regeneration. The mechanical evaluation described the functional regeneration of the muscle, with an increase in fast-twitch strength to twothirds of that in the uninjured control muscle at 8 weeks after trauma, being only 13% higher than the normalized tetanic strength. This shows the recovery of the chemical and mechanical properties of the soleus muscle over time. However, contraction forces at 8 weeks in our model lacked half (tetany) and one third (fast twitch) of the forces of the uninjured muscle. Histological evaluation at this time showed that there was pronounced interstitial fibrosis, which differs from the local fibrosis described above. It can be deduced that not only contractile properties but also passive mechanical properties such as muscle stiffness are altered by this tight collagenous network, also affecting force development via reduced compliance of the muscle. Together with the replacement of contractile muscle substance with fibrotic tissue, this reduction in compliance can still be responsible for the residual lag of contraction forces 8 weeks after trauma. Although the character of the collagenous tissue changes, the amount appears to stay at a constant level of 25% of the muscle tissue. Regenerative therapies for supFigure 7. a. Factor VIII staining of a longitudinally sectioned soleus muscle 4 weeks after crush ported muscle healing should take injury, showing the distribution of vessels in the injury zones. b. Cryo-section with an immunohisthis into account, since not addressing tochemical stain for nestin in a healthy soleus muscle (counterstain: DAPI), myotendinous transifibrosis means not being able to fully tion zone. c. Staining for nestin in a longitudinally sectioned soleus muscle (counterstain: DAPI). Nestin immunoreactivity was dispersed throughout the muscle and could be found not only at reconstitute muscle forces after injury. the tips of regenerating myofiber stumps but also at the lateral aspects of the fibers. d. Overview The increase in contraction forces of a-bungarotoxin staining of a longitudinally sectioned soleus muscle (counterstain: DAPI). The during the observation period was white arrows indicate multiple, newly developed neural endplates in the proximal and distal crush zones of the muscle. e. Detail of panel d depicting a linear distribution of endplates in the uninjured accompanied by a decrease in T1 and region. f. Detail of panel d depicting endplates within the crushed area. T2 hyperintensity in the MR measurements. The time course of this develtion (Kaariainen et al. 1998, Menetrey et al. 1999), provided opment described the decrease in the acuity of the trauma. the distal part of the affected muscle has not been severed Morphologically, MR scans showed a disorganized aspect of


110

the muscle tissue one week after trauma with an inhomogeneity of voxel signals, which could only be observed to a small degree 4 weeks after trauma. This development corresponded well with the histological sections showing short regenerating muscle fibers separated by loose connective tissue at the first time investigated, and showing successive reorganization 4 and 8 weeks after trauma. Fibrotic tissue exhibits low signal intensity on all pulse sequences (Elsayes et al. 2006). 8 weeks after injury, the structure of the soleus muscles appeared to be homogenous and could hardly be differentiated from the surrounding muscles. This demonstrates that the interstitial fibrosis, which is one of the main problems for the contractile function, cannot be detected by standard MRI sequences, even when high field scanners are used. We believe that our model will be of value for studying the influence of therapeutic approaches at several different levels of skeletal muscle regeneration.

TW: design of the study, animal surgery, and preparation of the manuscript. PvR and CP: MRI investigations. GM: muscle force measurements. MRS, SH, and GS: histological analysis. PS: animal surgery. GND: statistical analysis. ST: design of the study and preparation of manuscript.

The authors thank Susanne Mueller, Camilla Bergmann, and Gabriela Korus for providing excellent technical assistance. The study was supported by a kick-off grant from the Berlin-Brandenburg Center for Regenerative Therapies.

No competing interests declared.

Aarimaa V, Kaariainen M, Vaittinen S, Tanner J, Jarvinen T, Best T, et al. Restoration of myofiber continuity after transection injury in the rat soleus. Neuromuscul Disord 2004; 14 (7): 421-8. Anderson J, Almeida-Silveira M I, Perot C. Reflex and muscular adaptations in rat soleus muscle after hindlimb suspension. J Exp Biol 1999; 202 (Pt 19): 2701-7. Bassaglia Y, Gautron J. Fast and slow rat muscles degenerate and regenerate differently after whole crush injury. J Muscle Res Cell Motil 1995; 16 (4): 420-9. Broniatowski M, Azar K, Davies C R, Jacobs G B, Tucker H M. Selective feedback control of spastic musculature in a canine model. Asaio J 1992; 3 (3): M248-52. Bunn J R, Canning J, Burke G, Mushipe M, Marsh D R, Li G. Production of consistent crush lesions in murine quadriceps muscle--a biomechanical, histomorphological and immunohistochemical study. J Orthop Res 2004; 22 (6): 1336-44. Dedkov E I, Kostrominova T Y, Borisov A B, Carlson B M. Reparative myogenesis in long-term denervated skeletal muscles of adult rats results in a reduction of the satellite cell population. Anat Rec 2001; 263 (2): 139-54. Duda G N, Bartmeyer B, Sporrer S, Taylor W R, Raschke M, Haas N P. Does partial weight bearing unload a healing bone in external ring fixation? Langenbecks Arch Surg 2003; 388 (5): 298-304. Elsayes K M, LammLe M, Shariff A, Totty W G, Habib I F, Rubin D A. Value of magnetic resonance imaging in muscle trauma. Curr Probl Diagn Radiol 2006; 35 (5): 206-12.

Acta Orthopaedica 2011; 82 (1): 102â&#x20AC;&#x201C;111

Endo M. Calcium release from the sarcoplasmic reticulum. Physiol Rev 1977; 57 (1): 71-108. Fink E, Fortin D, Serrurier B, Ventura-Clapier R, Bigard A X. Recovery of contractile and metabolic phenotypes in regenerating slow muscle after notexin-induced or crush injury. J Muscle Res Cell Motil 2003; 24 (7): 421-9. Garrett W E, Jr., Seaber A V, Boswick J, Urbaniak J R, Goldner J L. Recovery of skeletal muscle after laceration and repair. J Hand Surg (Am) 1984; 9 (5): 683-92. Harry L E, Sandison A, Paleolog E M, Hansen U, Pearse M F, Nanchahal J. Comparison of the healing of open tibial fractures covered with either muscle or fasciocutaneous tissue in a murine model. J Orthop Res 2008; 26 (9): 1238-44. Hurme T, Kalimo H. Adhesion in skeletal muscle during regeneration. Muscle Nerve 1992; 15 (4): 482-9. Jarvinen M. Healing of a crush injury in rat striated muscle. 3. A microangiogra-phical study of the effect of early mobilization and immobilization on capillary ingrowth. Acta Pathol Microbiol Scand 1976; 84A: 85-94. Jarvinen M, Sorvari T. Healing of a crush injury in rat striated muscle. 1. Description and testing of a new method of inducing a standard injury to the calf muscles. Acta Pathol Microbiol Scand (A). 1975; 83 (2): 259-65. Jarvinen M, Aho A J, Lehto M, Toivonen H. Age dependent repair of muscle rupture. A histological and microangiographical study in rats. Acta Orthop Scand 1983; 54: 64-74. Kaariainen M, Kaariainen J, Jarvinen T L, Sievanen H, Kalimo H, Jarvinen M. Correlation between biomechanical and structural changes during the regeneration of skeletal muscle after laceration injury. J Orthop Res 1998; 16 (2): 197-206. Kleemann R U, Heller M O, Stoeckle U, Taylor W R, Duda G N. THA loading arising from increased femoral anteversion and offset may lead to critical cement stresses. J Orthop Res 2003; 21 (5): 767-74. Koishi K, Zhang M, McLennan I S, Harris A J. MyoD protein accumulates in satellite cells and is neurally regulated in regenerating myotubes and skeletal muscle fibers. Dev Dyn 1995; 202 (3): 244-54. Kurek J B, Bower J J, Romanella M, Koentgen F, Murphy M, Austin L. The role of leukemia inhibitory factor in skeletal muscle regeneration. Muscle Nerve 1997; 20 (7): 815-22. McGeachie J K, Grounds M D. Initiation and duration of muscle precursor replication after mild and severe injury to skeletal muscle of mice. An autoradiographic study. Cell Tissue Res 1987; 248 (1): 125-30. Meffert R H, Frey S P, Jansen H, Ochman S, Raschke M J, Langer M. Muscle strength quantification in small animals: a new transcutaneous technique in rabbits. J Orthop Res 2008; 26 (11): 1526-32. Menetrey J, Kasemkijwattana C, Fu F H, Moreland M S, Huard J. Suturing versus immobilization of a muscle laceration. A morphological and functional study in a mouse model. Am J Sports Med 1999; 27 (2): 222-9. Negishi S, Li Y, Usas A, Fu F H, Huard J. The effect of relaxin treatment on skeletal muscle injuries. Am J Sports Med 2005; 33 (12): 1816-24. Nozaki M, Li Y, Zhu J, Ambrosio F, Uehara K, Fu F H, et al. Improved muscle healing after contusion injury by the inhibitory effect of suramin on myostatin, a negative regulator of muscle growth. Am J Sports Med 2008; 36 (12): 2354-62. Parmiggiani F, Stein R B. Nonlinear summation of contractions in cat muscles. II. Later facilitation and stiffness changes. J Gen Physiol 1981; 78 (3): 295-311. Pereira B P, Tan J A, Zheng L, Tan B L, Lahiri A, Lim A Y, et al. The cut intramuscular nerve affects the recovery in the lacerated skeletal muscle. J Orthop Res 2006; 24 (1): 102-11. Perka C, Heller M, Wilke K, Taylor W R, Haas N P, Zippel H, et al. Surgical approach influences periprosthetic femoral bone density. Clin Orthop 2005; (432): 153-9. Rubinstein I, Abassi Z, Coleman R, Milman F, Winaver J, Better O S. Involvement of nitric oxide system in experimental muscle crush injury. J Clin Invest 1998; 101 (6): 1325-33.


Acta Orthopaedica 2011; 82 (1): 102â&#x20AC;&#x201C;111

Rushton J L, Davies I, Horan M A, Mahon M, Williams R. Production of consistent crush lesions of murine skeletal muscle in vivo using an electromechanical device. J Anat 1997; 190 ( Pt 3): 417-22. Sato K, Li Y, Foster W, Fukushima K, Badlani N, Adachi N, et al. Improvement of muscle healing through enhancement of muscle regeneration and prevention of fibrosis. Muscle Nerve 2003; 28 (3): 365-72. Saunders J H, Sissons H A. The effect of denervation on the regeneration of skeletal muscle after injury. J Bone Joint Surg (Br) 1953; 35 (1): 113-24. Schaser K D, Zhang L, Haas N P, Mittlmeier T, Duda G, Bail H J. Temporal profile of microvascular disturbances in rat tibial periosteum following closed soft tissue trauma. Langenbecks Arch Surg. 2003; 388 (5): 323-30. Schmalbruch H. The morphology of regeneration of skeletal muscles in the rat. Tissue Cell 1976; 8 (4): 673-92. Schultz E, Jaryszak D L, Valliere C R. Response of satellite cells to focal skeletal muscle injury. Muscle Nerve 1985; 8 (3): 217-22. Silverman D G, Brull S J. The effect of a tetanic stimulus on the response to subsequent tetanic stimulation. Anesth Analg 1993; 76 (6): 1284-7.

111

Sorokin L M, Maley M A, Moch H, von der Mark H, von der Mark K, Cadalbert L, et al. Laminin alpha4 and integrin alpha6 are upregulated in regenerating dy/dy skeletal muscle: comparative expression of laminin and integrin isoforms in muscles regenerating after crush injury. Exp Cell Res 2000; 256 (2): 500-14. Sweat F, Puchtler H, Rosenthal S I. Sirius Red F3ba as a stain for connective tissue. Arch Pathol 1964; 78: 69-72. Utvag S E, Grundnes O, Rindal D B, Reikeras O. Influence of extensive muscle injury on fracture healing in rat tibia. J Orthop Trauma 2003; 17 (6): 430-5. Wake R, Kim-Mitsuyama S, Izumi Y, Yoshida K, Izumiya Y, Yukimura T, et al. Beneficial effect of candesartan on rat diastolic heart failure. J Pharmacol Sci 2005; 98 (4): 372-9. Winkler T, von Roth P, Matziolis G, Mehta M, Perka C, Duda G N. Doseresponse relationship of mesenchymal stem cell transplantation and functional regeneration after severe skeletal muscle injury in rats. Tissue Eng Part A 2009; 15 (3): 487-92.

.


112

Acta Orthopaedica 2011; 82 (1): 112–113

Diaphyseal femoral fatigue fracture associated with bisphosphonate therapy – 3 more cases Kaiya Osugi1, Shigeki Miwa2, Shinobu Marukawa3, Kouhei Marukawa3, Yoshiharu Kawaguchi4, and Shinichi Nakato1 1Department

of Orthopaedic Surgery, Asahi General Hospital; 2Miwa Orthopedic Clinic,3Marukawa Hospital; 4Department of Orthopedic Surgery, University of Toyama, Japan Submitted 10-03-28. Accepted 10-08-17

Case 1 In April 2006, a 77-year-old female fractured her left femur after falling. The cortex had a spike-shaped edge at the fracture site and also showed thickening (Figure 1). Intramedullary nailing was performed. The fracture gap remained open 1 year after surgery (Figure 2), but healing was confirmed by radiography 2 years after the injury. 2.5 years after the first injury, the patient fell while walking and contracted a diaphyseal transverse fracture of the right femur. The cortex had a spike-shaped edge at the fracture site. The patient had been given ethidronate (4.8 g/12 weeks) from January 1999 to October 2001 because of postmenopausal osteoporosis. Thereafter, she was given alendronate (5 mg/day) for 7 years until the right femoral fracture had healed. Urine analysis conducted in February 2004 showed a free deoxypyridinoline/creatinine ratio (f-DPD/Cre) of 3.1, which is at the lower end of the normal range (2.8–7.6). Case 2 In May 2009, a 78-year-old female felt that she was going to fall while walking, and contracted a diaphyseal transverse fracture. The cortex had a spike shaped-edge at the fracture site (Figure 3). The lateral cortex of the contralateral femur was also thickened (Figure 4). This patient had been administered risedronate (2.5 mg/day) for 2 years, from November 2002 until November 2004, because of postmenopausal osteoporosis. Thereafter, she was given alendronate (5 mg/day) until October 2008. In the next half-year period from October 2008, she was administered raloxifene (60 mg/day). Her tartrate-resistant acid phosphatase 5b (TRACP-5b) 2 days after the injury was 111 mU/dL, which is below the standard values for postmenopausal females (250–260). Case 3 In March 2004, a 70-year-old female fell and contracted a diaphyseal transverse fracture of her right femur with spikeshaped cortical thickening laterally. The lateral cortex of the left femur was also thickened. Union was confirmed by radiographs 6 months after intramedullary nailing. In September 2004, she felt that she was falling and experienced pain in her left femur. Radiographs showed an incomplete fracture, and

she was advised to rest and support her leg. She was able to walk again, putting weight on the leg in December. In March 2005, she felt that she was about to fall and fractured the left femur. Again, the cortex presented a spike-shaped edge at the fracture. The patient had been administered risedronate (2.5 mg/day) from June 2002 through January 2005.

Discussion Suppression of bone resorption is the principal pharmacological activity of bisphosphonates. Bisphosphonates are therefore widely used in the treatment of osteoporosis. They reduce the risk of fractures (Kwek et al. 2008). However, the possibility of severely suppressed bone turnover (SSBT) under long-term administration has been cited (Black et al. 2006). In other words, repair of the microdamage to bones in daily life is inhibited by bone resorption suppression agents, and the cumulative effect of this microdamage is the mechanism behind vulnerability of the bone (Mashiba et al. 2000). Some characteristics of meta/diaphyseal stress fractures include (1) that they result from light trauma, (2) that they are simple and transverse, (3) that one side of the bone cortex shows a spike-shaped edge, (4) that there is thickening of the lateral cortex (Odvina et al. 2010), (5) that there is delayed union, and (6) that there is presentation of symptoms before the fracture (Ciarella et al. 2003). Points (1)–(4) were characteristics of all our cases. Furthermore, the fracture gap that persisted for a year in case 1 might be an example of (5). In addition, a reduction in bone resorption markers was confirmed in cases 1 and 2. Since the first case of a fatigue bone fracture during treatment with alendronate, reported in 2005 by Odvina et al., there have been a number of such cases described in the literature (Black et al. 2006). Furthermore, there have been some rarer reports with risedronate (Ensrud et al. 2004, Goh et al. 2007). To our knowledge, however, there have been no previous reports regarding raloxifene. Here, in case 2, we report a fatigue bone fracture after 6 years of bisphosphonate treatment, followed by a switch to raloxifene. In case 3, a complete fracture occurred 2 months after cessation of bisphosphonate

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2010.539500


Acta Orthopaedica 2011; 82 (1): 112â&#x20AC;&#x201C;113

Figure 1. Case 1. The right femur showed lateral cortical thickening with a jagged pattern (arrow).

Figure 2. Case 1. Reduced fracture gap of the left femur (arrowheads) 1 year postoperatively.

113

Figure 3. Case 2. Right transverse diaphyseal fracture with a medial spike (arrow).

administration. This is the only report of fatigue bone fractures after transition from bisphosphonate to raloxifene treatment or after suspension of risedronate administration. Several cases of fatigue fractures of the femur in patients on long-term bisphosphonate treatment were reported by Aspenberg (2009). Almost half of the patients had bilateral fractures and a quarter of them took corticosteroids. On the other hand, Schilcher and Aspenberg (2009) reported that the incidence density of stress fractures associated with bisphosphonate use was low and acceptable, considering that bisphosphonate treatment is likely to reduce the incidence density of any fracture. Black et al. (2010) concluded that the risk of fracture of the subtrochanteric or diaphyseal femur associated with bisphosphonate use was very low.

Aspenberg P. Bisphosphonates and fatigue fractures. IBMS BoneKEy 2009; 6: 465-9. Black D M, Schwartz A V, Ensrud K E, Cauley J A, Levis S, Quandt S A, Satterfield S, Wallace R B, Bauer D C, Palermo L, Wehren L E, Lombardi A, Santora A C, Cummings S R, FLEX Research Group. Effects of continuing or stopping alendronate after 5 years of treatment: the Fracture Intervention Trial Long-term Extension (FLEX): a randomized trial. JAMA. 2006; 296: 2927-38.

Figure 4. Case 2. The left femur showed cortical thickening with a jagged pattern (arrow).

Black D M, Kelly M P, Genant H K, et al. Bisphosphonates and fractures of the subtrochanteric or diaphyseal femur. N Engl J Med. 2010; 362: 176171. Ciarella T E, Fyhrie D P, Parfitt A M. Effects of vertebral bone fragility and bone formation rate on the mineralization levels of cancellous bone from white females. Bone 2003; 32: 311-5. Ensrud K E, Barrett-Connor E L, Schwartz A, et al. Randomized trial of effect of alendronate continuation versus discontinuation in women with low BMD: results from the Fracture Intervention Trial long-term extension. J Bone Miner Res 2004; 8: 1259-69. Goh S K, Yang K Y, Koh J S, et al. Subtrochanteric insufficiency fractures in patients on alendronate therapy. J Bone Joint Surg (Br) 2007; 89: 349-53. Kwek E B K, Goh S K, Koh J S B, et al. An emerging pattern of subtrochanteric stress fractures: A long-term complication of alendronate therapy. Injury 2008; 39: 224-31. Mashiba T, Hirano T, Turner C H, et al. Suppressed bone turnover by bisphosphonates increases microdamage accumulation and reduces biomechanical properties in dog rib. J Bone Miner Res 2000; 15: 613-20. Odvina C V, Levy S, Rao S, Zerwekh J E, Sudhaker Rao D. Unusual mid-shaft fractures during long term bisphosphonate therapy. Clin Endocrinol 2010; 72:161-8. Schilcher J, Aspenberg P. Incidence of steress fractures of the femoral shaft in women treated with bisphosphonate. Acta Orthop 2009; 80 (4): 413-5.


114

Acta Orthopaedica 2011; 82 (1): 114–115

Technical note

Pseudoarthrosis repair after failed metatarsophalangeal 1 arthrodesis Turnaround inlay graft in 26 patients Imre M Takács and Bart A Swierstra Department of Orthopaedic Surgery, Sint Maartenskliniek, Nijmegen, The Netherlands Correspondence: b.swierstra@maartenskliniek.nl Submitted 10-09-05. Accepted 10-11-04

We present an inlay bone grafting technique for the treatment of failed first metatarsophalangeal joint (MTP-1) arthrodesis. To our knowledge, this technique has not been described before for this specific condition.

Technique The patient is placed in supine position with a tourniquet around the upper or lower leg. An incision is made over the dorsum of the first ray. The pseudoarthrosis is exposed. Existing hardware is removed when necessary. A corticocancellous 0.5–0.75 × 5 cm bone block, including healthy metatarsal bone and the pseudoarthrosis site, is osteotomized and mobilized. The bone block is turned around 180 degrees and put back into its bone bed, bridging the pseudoarthrosis site with a continuous piece of healthy bone. Fixation is done with a plate (AO third tube) and screws (Figure 1). The patient is mobilized with heel weight bearing for 8 weeks.

Patients Between 2001 and 2009, 26 patients (26 feet) with pseudoarthrosis of a MTP-1 arthrodesis were operated with this technique. Their mean age at operation was 63 (52–76) years. The previous fixation method was most often with screws, or with a K-wire or a plate. The average time from the last arthrodesis attempt until revision arthrodesis was 12 (4–31) months. The average time to follow-up was 4.5 (0.5–9) years. At follow-up, all patients were sent a Foot Function Index (FFI) form and a questionnaire containing a VAS score for pain, a question about satisfaction, and a question about orthopedic footwear.

Figure 1. Schematic representation of turnaround inlay graft for MTP 1 pseudarthrosis.

25 of 26 patients returned the questionnaire. The average postoperative FFI score was 23 (0–47). The average postoperative VAS score for pain was 2.5 (0–9). 21 patients were satisfied with the postoperative result. 10 patients used prescription shoes. 2 patients presented with plate breakage because of non-union at 10 months and at 5 years after surgery, but they did not have pain and did not need treatment. In the other 24 patients, there was no sign of non-union at discharge from routine follow-up (Figures 2 and 3). In 2 patients, hardware was removed because of irritation. 1 patient had a relapse of Sudeck dystrophy. 1 patient had a correction osteotomy because of malposition of the hallux 6 years after surgery.

Discussion Pseudoarthrosis after attempted MTP-1 arthrodesis occurs in 1–14% of cases (Coughlin and Shurnas 2003, Coughlin et al. 2005, Gaucher and Coughlin 2006, Bennet and Sabatta 2009).

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2011.552782


Acta Orthopaedica 2011; 82 (1): 114–115

115

B Figure 3. A and B. 6 months after turn around inlay graft.

A Figure 2. MTP-1 pseudarthrosis after removal of hemiprosthesis.

Hope et al. (2010) described acceptable outcome scores with only removal of material and debridement in 7 of 11 patients. In our series, the 2 patients with hardware failure did not want further treatment, which confirms that some patients tolerate a non-union well. Hamilton et al. (2009) described other surgical treatment options. A Keller resection arthroplasty has several mechanical disadvantages and is not recommended for active patients. Resection of the pseudoarthrosis and refixation will lead to shortening of the first ray. This might be overcome with a bone block distraction arthrodesis. Fusion after this procedure varied from 79–92% in small series of patients with different indications (Hecht et al. 1997, Brodsky et al. 2000, Meyerson et al. 2000). Our fusion rate in 23/25 patients is similar or better, but our technique is less demanding. Our technique has some advantages. Because no additional incision for graft harvesting is needed, there is no donor site morbidity. As the surgical approach is only from the dorsal side, disruption of local vascularization is limited, which is beneficial for fusion. The alignment of the hallux remains unchanged. But if the position is unsatisfactory, small corrections are possible without complete disruption of the pseudarthrosis. A disadvantage of the technique is the sometimes difficult radiographic judgment of the fusion of the graft. The patients with the plate breakage were already discharged from routine postoperative follow-up and were satisfied for 10 months and 5 years, respectively, but apparently the graft was not incorporated. Despite satisfaction in 21/25 patients, the mean FFI of 23 (0–49) is not as good as our results after primary MTP-1 fusion with an FFI of 8 (0–59) (Van Doeselaar et al. 2010).

IT performed the follow-up and prepared the manuscript. BS performed all operations and finalized the manuscript.

Bennet G L, Sabetta J. First metatarsophalangeal arthrodesis: evaluation of plate and screw fixation. Foot Ankle Int 2009; 30: 752-6. Brodsky J W, Ptaszek A J, Morris S G. Salvage first MTP arthrodesis utilizing ICBG, clinical evaluation and outcome. Foot Ankle Int 2000; 21: 290-6. Coughlin M J, Shurnas P S. Hallux rigidus grading and long-term results of operative treatment. J Bone Joint Surg (Am) 2003; 85: 72-88. Coughlin M J, Grebing B R, Jones C P. Arthrodesis of the first metatarsophalangeal joint for idiopathic hallux valgus: intermediate results. Foot Ankle Int 2005; 26: 783-92. Gaucher N R, Coughlin M J. Hallux metatarsophalangeal joint arthrodesis using dome-shaped reamers and dorsal plate fixation: a prospective study. Foot Ankle Int 2006; 27: 869-76. Hamilton G A, Ford L A, Patel S. First metatarsophalangeal joint arthrodesis and revision arthrodesis. Clin Podiatr Med Surg 2009; 26: 459-73. Hecht P J, Gibbons M J, Wapner K L, Cooke C, Hoisington S A. Arthrodesis of the first metatarsophalangeal joint to salvage failed silicone implant arthroplasty. Foot Ankle Int 1997; 18: 383-90. Hope M, Savva N, Whitehouse S, Elliot R, Saxby T S. Is it necessary to refuse a non-union of a hallux metatarsophalangeal joint arthrodesis? Foot Ankle Int 2010; 31: 662-9. Myerson M S, Schon L C, McGuigan F X, Oznur A. Result of arthrodesis of the hallux metatarsophalangeal joint using bone graft for restoration of length. Foot Ankle Int 2000; 21: 297-306. Van Doeselaar D J, Heesterbeek P J C, Louwerens J W K, Swierstra B A. Foot function after fusion of the first metatarsophalangeal joint. Foot Ankle Int 2010; 8: 670-5.


116

Acta Orthopaedica 2011; 82 (1): 116â&#x20AC;&#x201C;118

Management of severe tibial bony defects with double metal blocks in knee arthroplastyâ&#x20AC;&#x201D;a technical note involving 9 cases Seung-Wook Baek and Choong H Choi Department of Orthopaedic Surgery, Hanyang University College of Medicine, South Korea Correspondence : chhchoi@hanyang.ac.kr Submitted 10-06-17. Accepted 10-09-28

Severe bony defects during either primary or revision knee arthroplasty are common. If the defect is > 10 mm in its greatest depth, metal augmentation or a bone graft should be considered. We report the operative technique of metal augmentation using double metal blocks for severe uncontained proximal tibial defects. Surgical technique and postoperative protocol Through an anterior midline skin incision and medial parapatellar approach, the knee joint is exposed (Figures 1 and 2). Soft tissue is released to obtain balancing of varus or valgus deformities. The intramedullary alignment instrumentation is used to prepare the femoral side. The extramedullary tibial alignment guides are applied to the tibia. Once proper axial alignment is verified, the proximal end of the tibia can be cut at usually 10 mm depth from the unaffected tibial condyle. The tibial surface is then prepared to accept the tibial base tray. The sclerotic base of the defect is cut to expose a flat, cancellous bony surface and the concave, irregular defect is converted to a flat one by minimal bone removal with a saw. The tibial bone defect is then

Figure 1. Preoperative radiographs with marked medial tibial bone loss.

assessed, and an appropriate-sized metal block is selected. A cutting guide for the block is assembled and a matching bone resection carried out. Care must be taken not to over-resect the bone, since the tibial blocks should be inserted in a tight manner. The positions of the central keels on the baseplate are prepared on the tibial surface. The trial tibial component with the block and intramedullary stem is assembled and inserted. A trial reduction of the prosthesis is done, and alignment and stability with patellar tracking is assessed. After lavage, using pulsed normal saline and drying out of the prepared surfaces, blood and fat are kept out of the cement-prosthesis interface. The real components are assembled and cemented. We use one mix of PMMA cement with gentamicin to cement the tibial and femoral components separately. Metal blocks of 10 mm + 10 mm or 10 mm + 5 mm are used on the medial aspect of the tibial component to compensate for bone defects. Both the tray and the block have a waffled surface that allows interdigitation of cement. The first block is attached to the tibial tray with screws. After that, the next block is cemented to the first one (Figure

Figure 2. Intraoperative photograph after arthrotomy.

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2010.548031


Acta Orthopaedica 2011; 82 (1): 116–118

Figure 3. A. The first metal block is attached to the tibial tray with screws.

117

B. After that, the next block is cemented to the first one.

Patients Between 2004 and 2007, we carried out metal augmentation of tibial defects in 9 patients during either primary or revision knee arthroplasty, using the operative technique described. All the procedures were carried out by a single surgeon. Primary diagnosis were osteoarthritis in 4 cases, rheumatoid arthritis in 1, aseptic loosening in 2, and septic loosening in 1 case. Patients’ mean age at operation was 65 (51–80) years. The mean follow-up period was 5 (2.5–6) years. During the follow-up period, all the patients had a pain-free knee. The mean range of motion was 127° (120–135), and there was no radiographic evidence of prosthesis loosening or subsidence (Figures 5 and 6). Figure 4. If the tibial blocks protrude over the cortical rim, downsized metal blocks may be available (another case).

3). The intramedullary stems on the tibial components are commonly used to protect the peripheral bone from stress. Finally, the real prosthesis with cemented tibial stem and with 10 mm + 10mm or 10 mm + 5 mm double metal augmentation are cemented into place. Downsized metal blocks can be used when the tibial blocks protrude over the cortical rim because of the natural taper of the proximal tibia (Figure 4). Once the cement is cured and excess is removed, a polyethylene implant is inserted and the knee is reduced. Final checks of soft tissue tension, alignment, patellar tracking, and range of knee movements is made before wound closure. A continuous passive motion is started within 24 h postoperatively, progressing slowly in flexion. Passive knee extension is encouraged by placing the patient’s foot on a pillow while in bed. Weight bearing with the aid of crutches or a cane starts on the fifth or sixth postoperative day.

Discussion The indication for augmentation of bone defects in TKA is inability to achieve stability of the trial implants at the time of trial reduction. This generally occurs when 40% or more of the bone-implant interface is unsupported by host bone (Cuckler 2004). Treatment options for large tibial bone defects include polymethylmethacrylate cement, morselized or structural allograft, metal augments, and custom or hinge/tumor prostheses (Radnay and Scuderi 2006). Excessive cement augmented by screws and mesh techniques has been abandoned because the construct is not mechanically sound, and may result in cement fragmentation and early failure of the replacement (Brooks et al. 1984, Brand et al. 1989, Ritter et al. 1993). The technique for the application of an autograft in primary knee arthroplasty has been well described by Windsor et al. (1986). It is physiologically sound, cheap, and reproducible— and has the advantage of bone stock preservation (Toms et al. 2009). However, Laskin (1989) reported a 33% failure rate at 5-year follow-up of autografts in primary arthroplasty.


118

Acta Orthopaedica 2011; 82 (1): 116â&#x20AC;&#x201C;118

Figure 5. Immediate postoperative radiographs of knee replacement with double metal blocks.

Larger defects may require allograft augmentation (Cuckler 2004, Engh and Ammeen 2007). However, allograft bone has the disadvantages that is difficult to achieve a proper fit with the host bone (Ries 1996) and that although they remain structurally intact, they are frequently not revascularized; new bone is laid down only in the periphery of a dead allograft (Parks and Engh 1997). Custom implants can be used for large bone defects, and should theoretically provide the best fit and force transmission of any of the methods used to address bone deficiency (Brooks et al. 1984). However, they are expensive and often require considerable time to manufacture. Modular implants with metal augmentation facilitate the treatment of bony defects (Patel et al. 2004, Fehring et al. 1996, Nelson et al. 2003, Pagnano et al. 1995). The advantage of modular metal augments is that they offer flexibility during the operation but still provide the surgeon with the ability to deal with defects of bone ranging from 5 to 10 mm in depth in various locations on the tibial plateau (Rand 1998). However, custom prostheses may be required for larger defects. We have modified the operative method by attaching 2 blocks with the use of cement and screws to allow up to 20 mm of segmental bone defects to be restored. They can be applied quickly, allow intraoperative custom fabrication, and help restore an anatomic joint line. The technique is simple and has no learning curve. Brand M G, Daley R J, Ewald F C, Scott R D. Tibial tray augmentation with modular metal wedges for tibial bone stock deficiency. Clin Orthop 1989; (248): 71-9.

Figure 6. Postoperative radiographs at 5-year follow-up.

Brooks P J, Walker P S, Scott R D. Tibial component fixation in deficient tibial bone stock. Clin Orthop 1984; (184): 302-8. Cuckler J M. Bone loss in total knee arthroplasty: graft augment and options. J Arthroplasty (Suppl 1) 2004; 19 (4): 56-8. Engh G A, Ammeen D J. Use of structural allograft in revision total knee arthroplasty in knees with severe tibial bone loss. J Bone Joint Surg (Am) 2007; 89 (12): 2640-7. Fehring T K, Peindl R D, Humble R S, Harrow M E, Frick S L. Modular tibial augmentations in total knee arthroplasty. Clin Orthop 1996; (327): 207-17. Laskin R S. Total knee arthroplasty in the presence of large bony defects of the tibia and marked knee instability. Clin Orthop 1989; (248): 66-70. Nelson C L, Gioe T J, Cheng E Y, Thompson R C, Jr. Implant selection in revision total knee arthroplasty. J Bone Joint Surg (Am) (Suppl 1) 2003; 85: S43-51. Pagnano M W, Trousdale R T, Rand J A. Tibial wedge augmentation for bone deficiency in total knee arthroplasty. A followup study. Clin Orthop 1995; (321): 151-5. Parks N L, Engh G A. The Ranawat Award. Histology of nine structural bone grafts used in total knee arthroplasty. Clin Orthop 1997; (345): 17-23. Patel J V, Masonis J L, Guerin J, Bourne R B, Rorabeck C H. The fate of augments to treat type-2 bone defects in revision knee arthroplasty. J Bone Joint Surg (Br) 2004; 86 (2): 195-9. Radnay C S, Scuderi G R. Management of bone loss: augments, cones, offset stems. Clin Orthop 2006; (446): 83-92. Rand J A. Modular augments in revision total knee arthroplasty. Orthop Clin North Am 1998; 29 (2): 347-53. Ries M D. Impacted cancellous autograft for contained bone defects in total knee arthroplasty. Am J Knee Surg 1996; 9 (2): 51-4. Ritter M A, Keating E M, Faris P M. Screw and cement fixation of large defects in total knee arthroplasty. A sequel. J Arthroplasty 1993; 8 (1): 63-5. Toms A D, Barker R L, McClelland D, Chua L, Spencer-Jones R, Kuiper J H. Repair of defects and containment in revision total knee replacement: a comparative biomechanical analysis. J Bone Joint Surg (Br) 2009; 91 (2): 271-7. Windsor R E, Insall J N, Sculco T P. Bone grafting of tibial defects in primary and revision total knee arthroplasty. Clin Orthop 1986; (205): 132-7.


Acta Orthopaedica 2011; 82 (1): 119–121

119

Correspondence

Failed internal fixation Sir—We read this article (Zustin and Winter 2009) with great interest and we appreciate the great efforts the authors have put in to publish this case report and extensive review of literature. However, we would like to discuss some primary issues regarding the case that has been reported. The authors made an excellent effort to salvage the prosthesis by internal fixation. However, this fixation method is not ideal even for a simple fractured neck of femur let alone periprosthetic fracture following hip resurfacing. It is well established that the more vertical the fracture neck of femur, higher is the incidence of non-union and avascular necrosis (Liporace et al. 2008) It would be of some interest to review the pre-injury radiograph with the hip resurfacing in situ to rule out superior notching as this fracture pattern is usually associated with superior notching and trauma in a young patient with good quality of bone. In a morphological study (Zustin et al. 2010c), osteonecrosis was the most frequent cause of fracture-related failures. The authors also suggested that intraoperative mechanical injury of the femoral neck such as notching and/or malpositioning of the femoral component might lead to changes in the loading pattern or in the resistance to fracture of the femoral neck and may result in both acute and chronic biomechanical femoral neck fractures. We note the authors comments that the avascular necrosis was possibly after the second procedure from the histology findings. However, it is difficult to ascertain whether the avascular necrosis was the cause of the fracture or it was secondary to the trauma that possibly disrupted the blood supply from the minimally displaced fracture. Praveen Mereddy 9 St Brides Court, Ingleby Barwick, Stockton-on-Tees, TS17 5HF Cleveland, UK e-mail: mpkr3@yahoo.com

Sir—I appreciate Dr. Mereddy’s comments regarding our article (Zustin and Winter 2009). In response, I would first like to discuss the issue of periprosthetic fractures following total hip resurfacing arthroplasty (THRA) somewhat more generally, and then comment on those issues with reference to the presented case. Periprosthetic fractures following THRA can be either atraumatic, or more rarely, caused by sufficient trauma. Atraumatic fractures occur typically during the first few months after implantation and the bone hidden inside the femoral component can be either viable or display theoretically diverse pathological changes (e.g. osteonecrosis, renal osteodystrophy, tumor). Therefore, we suggested that if the femoral remnant was viable and did not display any pathological changes that might have caused its weakening, then biomechanical factors played a substantial role in the pathogenesis of those fractures. Therefore, these fractures were defined as mechanical fractures (Zustin et al. 2010b). Interestingly, the biopsy showed either no reaction to fracture (acute mechanical fracture) or findings characteristic of fracture callus or pseudoarthrosis (chronic mechanical fracture). Both fracture types are most probably associated with mechanical insufficiency of the proximal femur caused by an inaccurate surgical technique or accumulation of compressive damage (Long et al. 2009). Currently, there is no doubt about the causative role of osteonecrosis (ON) in the pathogenesis of periprosthetic fractures following THRA (Little et al. 2005, Campbell et al. 2006, Steffen et al. 2010, Zustin et al. 2010b, c). Given the fact that the femoral head contains viable bone tissue prior to implantation, the absence of a vital reaction at the bone- (cement) -implant interface is regarded sufficient for the diagnosis of postoperative osteonecrosis and I am not aware of any data contrary to this concept, published by Bogoch et al. (1982). Nonetheless, Little et al. (2005) analyzed intraoperative biopsies from resurfaced femoral heads and found evidence of postoperative ON in 12 of 13 failures of THRA caused by fracture. Furthermore, the incidence of postnecrotic fractures was related to the extent of the ON lesion (Zustin et al. 2010a). Because healing cannot proceed within dead bone, it was not possible to further differentiate between acute or chronic postnecrotic fractures. It is true that fibrous membrane or pseudoarthrosis-like morphological changes can occur particular in cases with a collapsed ON lesion, and if we radiographically observed the pedestal sign (Madhu et al. 2010), migration and radiolucent lines in such cases, we placed them preferentially in the group

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2011.555373


120

Postoperative radiograph of the presented patient after implantation of the THRA.

of failures caused by loosening of the femoral component (Zustin et al. 2010a). Pathological fractures caused by other potential bone disease (renal osteodystrophy, tumor) have not yet been reported. Although ON was suggested to be causative in most periprosthetic fractures in the Hamburg THRA retrieval study (2003â&#x20AC;&#x201C;2009), it is noteworthy that we observed substantial changes in the failure pattern when we compared those fractures with the retrieved hips analyzed in our laboratory at the end of the study. More specifically, in retrieved femoral specimens from the latter period of time, periprosthetic fractures were less frequent and the specimens displayed smaller ON lesions. This may possibly be explained by the better surgical technique. Traumatic fractures seem to be somewhat different from the former type because, theoretically, the femoral components are both well-seated and well-fixed during implantation, and, despite the presence of viable good quality bone remnant under the femoral component, the fractures are caused by sufficient trauma. Furthermore, notching and/or a thick cement mantle or other findings linked with inaccurate surgical technique do not seem to play any substantial role in the pathogenesis of these fractures. This complication can occur even at late follow-up and the majority of the fractures are located sub-capitally, transcervically or even in the inter- or subtrochanteric portion of the femur. When compared with

Acta Orthopaedica 2011; 82 (1): 119â&#x20AC;&#x201C;121

the published cadaver studies (Angli et al. 2007, Richards et al. 2008, Vail et al. 2008), traumatic fractures showed some similar features (vertical fracture line in the subcapital location, possible location of the fracture line in the trochanteric area or even distally to it, ultimate load failure in experiment versus trauma). In the presented case, we did not recognize any radiographic signs of post-implantation notching (Figure). Also, despite the sufficient time (11 weeks) between the surgical therapy of the periprosthetic fracture and the revision surgery, we did not find any cellular reaction at the bone-metal interface on the surface of the screws. In contrast to this finding, a vital reaction was apparent at the bone-cement interface of the femoral component. Moreover, although the bone tissue proximal to the fracture line was areactive and multifocally fragmented due to the collapse of necrotic bone trabeculae adjacent to the fracture line, mineralized callus formation was apparent distally to the fracture line consistent with a vital reaction to the multifocal fractured and collapsed bone. Therefore I strongly believe that in our case, the femoral remnant was viable after the implantation of the THRA but not following the periprosthetic fracture caused by the motorcycle accident and its surgical therapy. Because traumatic fractures are quite rare, I am looking forward to seeing the results of morphological retrieval analyses by other colleagues and encourage the sharing of knowledge regarding the pathogenetic mechanisms of THRA complications. Jozef Zustin, MD Orthopaedic Pathology Unit, University Medical Center Hamburg-Eppendorf, Martinistr. 52, DE-20246 Hamburg j.zustin@uke.de

Anglin C, Masri B A, Tonetti J, Hodgson A J, Greidanus N V. Hip resurfacing femoral neck fracture influenced by valgus placement. Clin Orthop 2007; (465):71-9. Bogoch E R, Fornasier V L, Capello W N. The femoral head remnant in resurfacing arthroplasty. Clin Orthop 1982; (167): 92-105. Campbell P, Beaule P E, Ebramzadeh E, LeDuff M, De Smet K, Lu Z, et al. The John Charnley Award: a study of implant failure in metal-on-metal surface arthroplasties. Clin Orthop 2006; (453): 35-46. Liporace F, Gaines R, Collinge C, Haidukewych G J. Results of internal fixation of Pauwels type-3 vertical femoral neck fractures.J Bone Joint Surg (Am) 2008; 90 (8): 1654-9. Little C P, Ruiz A L, Harding I J, McLardy-Smith P, Gundle R, Murray D W, et al. Osteonecrosis in retrieved femoral heads after failed resurfacing arthroplasty of the hip. J Bone Joint Surg (Br) 2005; 87 (3): 320-3. Long J P, Santner T J, Bartel D L. Hip resurfacing increases bone strains associated with short-term femoral neck fracture. J Orthop Res 2009; 27 (10): 1319-25. Madhu T S, Akula M R, Raman R N, Sharma H K, Johnson V G. The Birmingham hip resurfacing prosthesis. An independent single surgeonâ&#x20AC;&#x2122;s experience at 7-year follow-up. J Arthroplasty 2010; Jan 5.


Acta Orthopaedica 2011; 82 (1): 119â&#x20AC;&#x201C;121

Richards C J, Giannitsios D, Huk O L, Zukor D J, Steffen T, Antoniou J. Risk of periprosthetic femoral neck fracture after hip resurfacing arthroplasty: valgus compared with anatomic alignment. A biomechanical and clinical analysis. J Bone Joint Surg (Am) (Suppl 3) 2008; 90: 96-101. Steffen R T, Athanasou N A, Gill H S, Murray D W. Avascular necrosis associated with fracture of the femoral neck after hip resurfacing: histological assessment of femoral bone from retrieval specimens. J Bone Joint Surg (Br) 2010; 92 (6): 787-93. Vail T P, Glisson R R, Dominguez D E, Kitaoka K, Ottaviano D. Position of hip resurfacing component affects strain and resistance to fracture in the femoral neck. J Bone Joint Surg Am. 2008 Sep;90(9):1951-60.

121

Zustin J, Winter E. Failed internal fixation due to osteonecrosis following traumatic periprosthetic fracture after hip resurfacing arthroplasty. Acta Orthop 2009; 80 (6): 666-9. Zustin J, Hahn M, Morlock M M, Ruther W, Amling M, Sauter G. Femoral component loosening after hip resurfacing arthroplasty. Skeletal Radiol 2010a; 39 (8): 747-56. Zustin J, Sauter G, Morlock M M, Ruther W, Amling M. Association of osteonecrosis and failure of hip resurfacing arthroplasty. Clin Orthop 2010b; (468) (3): 756-61 Zustin J, Krause M, Breer S, Hahn M, von Domarus C, RĂźther W, Sauter G, Morlock M M, Amling M. Morphologic analysis of periprosthetic fractures after hip resurfacing arthroplasty. J Bone Joint Surg (Am) 2010c; 92 (2): 404-10.


122

Acta Orthopaedica 2011; 82 (1): 122–123

Correspondence

Overtreatment of cruciate ligament injuries Sir–In an editorial in the October issue of Acta Orthopaedica 2010, Per Aspenberg (Aspenberg 2010) comments on the article recently published in New England Journal of Medicine (Frobell et al. 2010). We congratulate the authors on this well performed randomized controlled trial on the treatment of ACL injuries were they found that rehabilitation + ACL reconstruction was not superior to rehabilitation + optional delayed ACL reconstruction when using the mean of 4 of the 5 subscales of the KOOS score at two years as the primary outcome variable. Aspenberg concludes in his editorial that most patients who are operated on early after ACL injury undergo the procedure in vain. However, in our opinion the difference in meniscal surgery between the groups in Frobell’s study may indicate the opposite. Preservation of the menisci is a key factor in preventing later osteoarthritis in ACL-deficient knees as previously shown in another study, also from Lund University, Sweden (Neumann et al. 2008). As shown in table D in supplementary material (Frobell et al. 2010) the total number of treated menisci at baseline was 34 in the early ACL reconstruction group versus 21 in the rehabilitation + optional delayed ACL reconstruction group. During follow up these numbers were 6 versus 29 respectively (<0.001). The authors say that meniscal tears were managed more aggressively in the subjects assigned to early ACL reconstruction and were more likely to be left untreated in the subjects assigned to rehabilitation plus optional delayed ACL reconstruction, and believe that this difference probably explains the greater frequency of meniscal surgery during follow-up in the latter group. In our view another possible explanation could be that meniscal tears that were small and non-symptomatic in the early phase became larger and symptomatic with time in the nonreconstructed knees. Thus, an injury that could be treated with either repair, or a small resection or left untreated at the time of an early ACL-reconstruction could end up as a large bucket handle tear leading to a subtotal menisectomy in an unstable knee. Secondly, the risk of developing new injuries to the menisci and articular cartilage may be higher in a non-reconstructed knee. This is supported by registry data (Granan et al. 2009). Finally, the difference in the frequency of meniscal injuries between the groups may increase further with longer follow-up time. In summary, Frobell’s study could just as well support early reconstruction to protect the menisci. We look forward to the long term follow-up in this study. If the frequency of meniscal tears continue to accumulate at the same rate in the non-reconstructed knees this study may end

up showing that most early ACL reconstructions are not performed in vain. Sverre Løken, Asbjørn Årøen, Lars Engebretsen Orthopaedic Department, Oslo University Hospital, Oslo, Norway

Sir–The main goal of cruciate ligament reconstruction in the sports medicine community has been to enable the patients to go back to a high functional level (often sports activities) in a short time. The study by Frobell et al. shows that, for this end, surgery is need-less for most patients if appropriate physiotherapy is provided. In that respect, I maintain that most patients may be operated on in vain, and that beginning with a structured training program is the treatment of choice for a vast majority. The letter from Løken et al. now brings up the subject of long-term results, i.e. the risk of osteoarthritis. This is a commendable and much awaited change of focus in the cruciate ligament replacement field. It remains to see whether reconstruction can prevent – or possibly cause – osteoarthritis. Acta Orthopaedica has asked Drs Lohmander and Frobell to answer the letter regarding this point. Per Aspenberg Department of Orthopaedics, IKE, Faculty of Health Sciences, Linköping University, Sweden per.aspenberg@liu.se

Sir–We thank Dr. Løken and colleagues for their interest in our study on anterior cruciate ligament (ACL) injuries. At 2 years of follow-up our randomized controlled study showed no advantage of early ACL reconstructive surgery with structured rehabilitation over that of structured rehabilitation only with optional reconstruction ‘as needed’, as monitored by

Open Access - This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the source is credited. DOI 10.3109/17453674.2011.555374


Acta Orthopaedica 2011; 82 (1): 122–123

KOOS (Frobell et al. 2010). Moreover, we found no difference in any of the secondary outcomes, or in the number of meniscal surgeries during the 2-year follow-up. We share the concern of Løken and colleagues for the long-term outcome after injury to the anterior cruciate ligament (ACL), in particular the development of osteoarthritis, and continue to monitor this patient group long term by patient-reported outcomes, activity level, plain radiographs, magnetic resonance imaging, and biomarkers. Results for these outcomes will be reported. Løken and colleagues raise three points. Firstly, they suggest that the literature shows that a functionally intact meniscus may be important to prevent future osteoarthritis in the ACL-injured knee. We agree on the likely importance of a functionally intact meniscus to prevent the development of knee osteoarthritis, but note that not all reports on the ACLinjured knee are consistent with this hypothesis. Secondly, they suggest that the risk of re-injury may be increased in the non-reconstructed knee. This is certainly possible, but all studies published so far with the exception of one (Frobell et al. 2010), are confounded by indication. We agree that the registry study of Granan (Granan et al. 2009) shows that the frequency of observed cartilage and meniscal damage increases with time after injury. This is to be expected, has been reported (Roos et al. 1995, Lohmander et al. 2007), and may be interpreted as early-stage osteoarthritis development. In the Granan study, no results were provided to report later cartilage and meniscus status by follow-up of those already reconstructed, so we are unable to draw any conclusions on the benefit of reconstruction on these outcomes from their study. Further, no data are available for those not included in the registry, and meniscus injury caused by the initial trauma is not reported. Recent observational studies find no difference in later osteoarthritis between those reconstructed or not (Lohmander et al. 2007, Meuffels et al. 2009). Thirdly, Løken and colleagues suggest that a difference in meniscal surgery rate between our study groups may develop with time. This is certainly possible. We found no difference between the two study groups in the number of meniscal surgeries at two years. Whether differences will show at later times remain to be reported. Until then we can only speculate. A high proportion of those with an ACL tear become ‘young patients with old knees’. The rate of osteoarthritis development after these injuries remains a major clinical challenge: with regard to preventing these injuries, preventing osteoarthritis

123

developing after the injury, and treating severe osteoarthritis in the young and active patient. We encourage further basic research on disease mechanisms, randomized trials to identify the best treatments, and large and long-term registry studies with minimal loss to follow-up to monitor long-term outcome and complications. In these studies, patient reported outcomes are central, other outcomes such as radiographic changes or numbers of surgeries are at best surrogates. L Stefan Lohmander Department of Orthopedics, Clinical Sciences Lund, Lund University, Sweden, and Research Unit for Musculoskeletal Function and Physiotherapy and Department Orthopaedics and Traumatology, University of Southern Denmark Richard B Frobell Department of Orthopedics, Clinical Sciences Lund, Lund University, Sweden

Aspenberg P. Overtreatment of cruciate ligament injuries. Acta Orthop 2010; 81 (5): 524-5. Frobell R B, Roos E M, Roos H P, Ranstam J, Lohmander L S. A randomized trial of treatment for acute anterior cruciate ligament tears. N Engl J Med 2010; 363 (4): 331-42. Frobell R B, Roos E M, Roos H P, Ranstam J, Lohmander L S. Supplementary material at: http://www.nejm.org/doi/suppl/10.1056/NEJMoa0907797/ suppl_file/nejmoa0907797_appendix.pdf Granan L P, Bahr R, Lie S A, Engebretsen L. Timing of anterior cruciate ligament reconstructive surgery and risk of cartilage lesions and meniscal tears: a cohort study based on the Norwegian National Knee Ligament Registry. Am J Sports Med 2009; 37 (5): 955-61. Lohmander L S, Englund M, Dahl, L L, Roos E M. The long-term consequence of anterior cruciate ligament and meniscus injuries: Osteoarthritis. Am J Sports Med 2007; 35: 1756-69. Meuffels D E, Favejee M M, Vissers M M, Heijboer M P, Reijman M, Verhaar J A N. Ten year follow-up study comparing conservative versus operative treatment of anterior cruciate ligament ruptures. A matched-pair analysis of high level athletes. Br J Sports Med 2009; 43: 347-51. Neuman P, Englund M, Kostogiannis I, Fridén T, Roos H, Dahlberg L E. Prevalence of tibiofemoral osteoarthritis 15 years after nonoperative treatment of anterior cruciate ligament injury: a prospective cohort study. Am J Sports Med 2008; 36 (9): 1717-25. Roos H, Adalberth T, Dahlberg L, Lohmander L S. Osteoarthrosis of the knee after injury to the anterior cruciate ligament or meniscus - the influence of time and age. Osteoarthritis Cartilage 1995; 3: 261-7.


124

Acta Orthopaedica 2011; 82

Information to authors Submission of manuscripts Manuscripts should be submitted online, www.actaorthopscand.org. Authors are encouraged to provide the names, addresses, and e-mail of potential reviewers. Note the following points: a. The manuscript should be prepared according to Uniform Requirements for Manuscripts submitted to Biomedical Journals (www.icmje.org/). American spelling is preferred. b. Authors submitting a paper do so on the understanding that it has not been published and is not being considered for publication elsewhere. The authors should provide a statement about previous submissions and reports that might be regarded as duplicate publication of the same or a similar paper. Copies of that study should be submitted with the paper. c. Authors should provide a description of their individual contributions to the study. d. Reports of interventional studies (defined by human participants being prospectively assigned to one or more health-related interventions to evaluate the effects on health outcomes) should be registered in a public trials registry, and the registration id should be presented at the bottom of the abstract page. The manuscript should comply with the CONSORT statement (www.consort-statement.org). The study protocol approved by the ethics committee and a completed CONSORT checklist and flowchart should be appended to the manuscript. e. Systematic reviews and meta-analyses of interventional studies should comply with the PRISMA statement (www. prisma-statement.org), and the PRISMA checklist and flowchart should be appended to the manuscript. For systematic reviews and meta-analyses of observational studies the MOOSE checklist and flowchart (www.equator-network. org) should be used. f. Manuscripts presenting results on diagnostic accuracy should comply with the STARD-statement (www.stardstatement.org), and the STARD checklist and flowchart should be appended to the manuscript. g. Reports of observational studies should comply with the STROBE-statement (www.strobe-statement.org) and the STROBE checklist should be appended to the manuscript. Manuscript layout Title. The title should include information on the results of the investigation, including the number of patients, average follow-up, and animal or cadaver experiments. The first name, middle initial(s), and last name of each author should be given with department affiliations as well as the name, complete address and e-mail of the corresponding author.

Abstract. The abstract should not exceed 200 words. It should consist of 4 parts: Background, Patients/material/animals and methods, Results and Interpretation. Introduction. The introduction should focus on the state of knowledge at the beginning of the study. The aims and main hypothesis of the study should be stated clearly. Only in exceptional cases should it exceed one typewritten page. Case reports. These should be short, include the Case history and Discussion without an Introduction. Patients/material/animals and methods. The selection of subjects and the inclusion or exclusion criteria should be described. Subjects who declined to participate, withdrawals and those with incomplete follow-up should be accounted for. Describe in detail how the measurements were made and the techniques used. Statistics. All statistical methods should be mentioned. Unusual methods should be referenced. The tests should be two-sided, unless the use of one-sided tests is justified. No data should be removed, imputed, weighted, adjusted or trimmed unless they are specifically described and justified and its consequences are given. Hypothesis tests (p-values). These tests should be used, in combination with a defined effect size and when statistical power has been considered. Give p-values with real numbers, if these are greater than 0.001 (1 digit except zeros), otherwise use ‘p < 0.001’. Do not use ‘ns’, ‘p > 0.05’ or asterisks. In most instances, the 95% confidence intervals should be given, especially in exploratory analyses and when estimating the effects or differences. Ethics. Approval by the local ethics committee should be stated with date of issue and registration number. Results. Do not give the same data in more than one way. When summarizing the data, always include measures of variability and the number of subjects. Give the median and range—e.g., 60 (35–70) years, the mean and standard deviation—59 (SD 15) years, and the frequencies for nominal data. The results of matched data should be given in relevant form (e.g., the distribution of pairwise differences). Percentages should not be used if the sample size is less than 100. Discussion. This section should contain a critical discussion of the results—e.g., the quality of the data (selection and information bias) and adequacy of the statistical analysis (confounding bias). It should also assess the relevant literature for or against the findings and if possible, the conclusions as regards clinical application or further research. Discuss, do not recapitulate, your results. Tables. Each table should be self-explanatory, with an adequate title and a logical presentation of the data. Abbreviate words in the columns and explain in footnotes. Each column heading for numerical data should include the unit of mea-

Copyright © Informa Healthcare Ltd 2011. ISSN 1745–3674. Printed in Sweden – all rights reserved.


Acta Orthopaedica 2011; 82

surement. Use SI units. Avoid unnecessary decimals! It is seldom advisable to use more than 2 digits for biologic measurements. Digital illustrations should be used. A scanned picture or digital photo should usually be about 85 mm (i.e. 1000 pixels) wide with a resolution of 300 DPI (dots per inch), which gives a square picture an uncompressed file size of 4.5 MiB in color and 1 MiB in gray-scale. A black and white drawing or graph may be scanned in 800 DPI bitmap, i.e. 1 bit TIFF. The preferred format is a TIFF-file. The resolution of WEB illustrations (gif) is usually too poor. Color photos should be in CMYK colors. The cost of color will be paid by Acta. Highlights, arrows and letters may be added to digital photos but keep a clean version of the photo for the layout process. Acta prefers to add these as a separate layer. For digital graphs, use a graphics program that can export EPS-files. Avoid frames around diagrams, diagrams with perspective drawing, and bar graphs or histograms (use tables). Symbols should be consistent throughout a series of figures. Each axis should be horizontally labeled, with a description of the variable it represents. Use sans serif letters (Arial or Helvetica). Make diagrams in black-and-white with gray or color areas, but avoid complex patterns. Contributions of authors. Describe in short what each author did. Acknowledgments. Technical help and financial or other sponsorship may be acknowledged. Conflicts of interest, if any, should be clearly stated in the manuscript on submission. When the manuscript is accepted for publication, all authors are asked also to sign a conflict of interest statement. References. We prefer the references to be cited by name and year (chronologically) of publication in the text instead of giving their numbers, i.e. one author, two authors or author et al. and year. Thus, the list of references should be given in alphabetical order.

125

Submission Create an Adobe Portable Document File (PDF) of your full manuscript. You will need the Adobe Acrobat Writer Program or similar PDF creator program installed on your computer. We recommend that you write your manuscript as a Microsoft Word file and embed all images, tables, figures etc. within the one file. Now convert the Microsoft Word Document to a PDF file by choosing the appropriate icon in the task bar of the Microsoft Word Application. Alternatively, right hand click on the icon of your manuscript file and choose, convert to Adobe PDF. If you do not have the Abobe PDF writer installed on your computer, you can use the ‘try it for free’ feature found on the Adobe website at this address: http://createpdf.adobe.com/ Important: Please line number the Microsoft Word document before converting to PDF by following the procedure below. Insert line numbers: 1. Switch to print layout view. 2. Do one of the following: • To add line numbers to the entire document, click Select All on the Edit menu. • To add line numbers to part of a document, select the text you want to number. • To add line numbers to existing sections, click in a section or select multiple sections. 3. On the File menu, click Page Setup, and then click the Layout tab. 4. If you’re adding line numbers to part of a document, click Selected text in the Apply to box. 5. Click Line Numbers. 6. Select the Add line numbering check box, and then select the options you want.

Acta Orthopaedica 1/11  

Magazine about special subjects in orthopaedical medicine.