Periprosthetic Femoral Fracture within Two Years After Total Hip Replacement

Transcription

1 e167(1) CPYRIGHT Ó 2014 BY THE JURNAL F BNE AND JINT SURGERY, INCRPRATED Periprosthetic Femoral Fracture within Two Years After Total Hip Replacement Analysis of 437,629 perations in the Noric Arthroplasty Register Association Database Truike M. Thien, MD, PhD, Georgios Chatziagorou, MD, Göran Garellick, MD, PhD, ve Furnes, MD, PhD, Leif I. Havelin, MD, PhD, Keijo Mäkelä, MD, PhD, Søren vergaar, MD, PhD, Alma Peersen, MD, PhD, Antti Eskelinen, MD, PhD, Pekka Pulkkinen, MD, PhD, an Johan Kärrholm, MD, PhD Investigation performe at the Institute of Clinical Sciences, The Sahlgrenska Acaemy, University of Göteborg, Göteborg, Sween Backgroun: We use the Noric Arthroplasty Register Association atabase to evaluate whether age, sex, preoperative iagnosis, fixation, an implant esign influence the risk of revision arthroplasty ue to periprosthetic fracture within two years from operation of a primary total hip replacement. Methos: Inclue in the stuy were 325,730 cemente femoral stems an 111,899 uncemente femoral stems inserte from 1995 to Seven frequently use stems (two cemente stems [Exeter an Lubinus SP II] an five uncemente stems [Bi-Metric, Corail, CLS Spotorno, ABG I, an ABG II]) were specifically stuie. Results: The incience of revision at two years was low: 0.47% for uncemente stems an 0.07% for cemente stems. Uncemente stems were much more likely to have this complication (relative risk, 8.72 [95% confience interval, 7.37 to 10.32]; p < ). Age ha no consistent influence on the risk for revision of cemente stems, but revision in the uncemente group increase with increasing age. A cemente stem was associate with a higher risk in male patients compare with female patients (hazar ratio, 1.95 [95% confience interval, 1.51 to 2.53]; p < ), whereas an uncemente stem was associate with a reuce risk in male patients compare with female patients (hazar ratio, 0.74 [95% confience interval, 0.62 to 0.89]; p = 0.001). The risk for revision ue to early periprosthetic fracture increase uring the 2003 to 2009 perio compare with the 1995 to 2002 perio both before an after ajustment for emographic factors an fixation (relative risk, 1.44 [95% confience interval, 1.18 to 1.69]; p < ). The hazar ratio for the Exeter stem was about five times higher than that for the Lubinus SP II stem (hazar ratio, 5.03 [95% confience interval, 3.29 to 7.70]; p < ). f the five uncemente stems, the ABG II stem showe an increase hazar ratio of 1.63 (95% confience interval, 1.16 to 2.28) (p = 0.005), whereas the Corail stem showe a ecrease hazar ratio of 0.47 (95% confience interval, 0.34 to 0.65) (p < ) compare with the reference Bi-Metric esign. Conclusions: The shape an surface finish of the femoral stem an its fixation coul be relate to the increase risk of some prosthetic esigns. Even if the incience of early periprosthetic fracture in general is low an other reasons for revision must be consiere, specific attention shoul be given to the choice of fixation an stem esign in risk groups. Level of Evience: Prognostic Level III. See Instructions for Authors for a complete escription of levels of evience. Peer Review: This article was reviewe by the Eitor-in-Chief an one Deputy Eitor, an it unerwent bline review by two or more outsie experts. It was also reviewe byan expert in methoologyan statistics. The Deputy Eitor reviewe each revision of the article, an it unerwent a final reviewbythe Eitor-in-Chief prior to publication. Final corrections an clarifications occurre uring one or more exchanges between the author(s) an copyeitors. Disclosure: None of the authors receive payments or services, either irectly or inirectly (i.e., via his or her institution), from a thir party in support of any aspect of this work. None of the authors, or their institution(s), have ha any financial relationship, in the thirty-six months prior to submission of this work, with any entity in the biomeical arena that coul be perceive to influence or have the potential to influence what is written in this work. Also, no author has ha any other relationships, or has engage in any other activities, that coul be perceive to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitte by authors are always provie with the online version of the article. J Bone Joint Surg Am. 2014;96:e167(1-7)

2 e167(2) Previous reports from the Sweish Hip Arthroplasty Register note that the early survival of uncemente total hip replacement is inferior to cemente fixation 1,partlyueto early periprosthetic femoral fracture. Previous hip fracture is a risk factor for subsequent periprosthetic fracture 2-5 as well as age 6. It is not clear if this early complication varies between ifferent uncemente stem esigns. The incience is expecte to rise in the future because of an increasing number of uncemente total hip arthroplasties an a longer life expectancy. Intraoperative femoral fractures appear to be more common because of increasing use of cementless fixation 2. Perhaps the risk of fracture can be reuce with careful preoperative planning an surgical technique, but no ata to ate support this assumption. Because of the comparatively rare occurrence of periprosthetic femoral hip fractures, large patient cohorts are neee to evaluate probable causes. Hip arthroplasty registries are an important tool for continuous monitoring of outcome after total hip replacement 7-9. The Noric Arthroplasty Register Association 10,11 (NARA) has compile a common atabase base on the National Registers in Denmark, Finlan, Norway, an Sween, which, because of its size, may enable stuies of this complication on the level of a specific implant esign. The aim of the present stuy was to evaluate whether age, sex, preoperative iagnosis, fixation, an specific implant esign influence the risk of revision ue to periprosthetic femoral fracture within two years from a primary total hip replacement. Materials an Methos From January 1, 1995, through 2009, 449,930 femoral stem implants (surface replacements exclue) were reporte to the arthroplasty registers in Denmark, Finlan, Norway, an Sween. These four Noric registers have nearly complete coverage. The completeness of ata varies between 86% an 99% in the four countries. The NARA inclues both primary total hip replacement an revision, as well as the ata on the reason for revision. Patients are registere on the basis of their personal ientity number, a national ientification number covering the total resient population of a country. Numbers are issue by the tax agency as part of the population register. All revisions are continuously reporte to the registers. The reason for revision is entere into a form by the operating surgeon an is valiate by stuies of case recors. Revision is efine as exchange or removal of the entire prosthesis or at least one of its parts. In this stuy, only revision of the femoral component ue to periprosthetic fracture within two years from the primary total hip replacement was use as an en point. Reoperation, efine as any other hip-relate surgery following primary total hip replacement, leaving the primary total hip replacement implant intact, was not inclue as an en point as the registration of reoperation is not uniform in all four countries. In 12,301 patients (2.7%), ata were incomplete because of at least one missing parameter of sex, age, or type of implant, so these cases were exclue, leaving 437,629 total hip replacements, of which 325,730 (74.4%) were cemente (Table I). The patients were ivie into five age groups (younger than fifty years, fifty to fifty-nine years, sixty to sixty-nine years, seventy to seventy-nine years, an eighty years or more) an into six groups of iagnoses. To enable analysis of iniviual stem esigns as one group, the ifferent coe numbers use in the four countries for a particular stem were recoe for each of the stems selecte, which were the esigns most frequently use. To be inclue, they shoul also have been use in at least three of the four countries inclue in the NARA collaboration. Seven implant esigns were ientifie for further analysis: the two most frequently use cemente stems, the polishe wege Exeter (Stryker, Portage, Michigan; an the anatomic Lubinus SP II (Walemar Link, Hamburg, Germany; an five uncemente stems, the polishe wege CLS Spotorno stem (Zimmer, Warsaw, Iniana; an four more anatomically shape stems, Bi- Metric (Biomet, Warsaw, Iniana; Corail (DePuy, Warsaw, Iniana; an ABG I an ABG II (Stryker) (Table II). All stems available were inclue ignoring the choice of fixation on the cup sie. All variations of each stem esign were inclue provie that they ha been use in at least fifty arthroplasties each. Statistics We use escriptive statistics for the presentation of emographic factors such as age, sex, preoperative iagnosis, fixation, an stem esign. The crue survival at six months an two years following primary total hip replacement was compute for the seven specifie implants. We use Cox regression analyses to calculate the crue an ajuste relative risk of revision ue to periprosthetic fracture within two years of primary total hip replacement operation with 95% confience interval (95% CI). We ajuste for the type of fixation (cemente or uncemente), age (less than fifty years, fifty to fifty-nine years, sixty to sixtynine years, seventy to seventy-nine years, or eighty years or more), iagnosis (primary osteoarthritis, fracture, peiatric hip isease, inflammatory isease, iiopathic femoral hea necrosis, an other iagnoses), an involve sie (right or left). In a secon moel, we ae year of operation separate into two groups (1995 to 2002 an 2003 to 2009) to evaluate if there was a tren over time. All patients were followe until ate of revision, eath, or December 31, Sex was exclue from these analyses because of non-proportionality an interaction between stem fixation an sex (p < ). There was also an interaction between stem fixation an age group (p < ) an a three-way interaction among stem fixation, sex, an age group (p < ), motivating a stratifie analysis. Therefore, subgroup analyses on each sex separately were performe comparing cemente an uncemente stems. In further analyses, TABLE I Choice of Stem Fixation in the Four Countries an Demographic Patient Data Cemente Uncemente Total no. of patients* 325, ,899 Share of cemente stems per country Denmark 63% Finlan 52% Norway 79% Sween 90% Sex Male 36% 53% Female 64% 47% Patient age # (yr) 71.4 ± ± 11.3 *This category inclue all patients with cemente stems, uncemente stems, hybri stems, an reverse hybri stems. The values are given as the percentage of patients. Significant ifferences among countries were note at p < Significant values between groups were note at p < #The values are given as the mean an the stanar eviation.

3 e167(3) TABLE II Numbers Revise within Six Months an Two Years for the Stem Designs in the Stuy Numbers Revise* Within Six Months Within Two Years Numbers Available All cemente stems 117 (0.04%) 238 (0.07%) 325,730 (74.4%) All uncemente stems 474 (0.42%) 530 (0.47%) 111,899 (25.6%) Cemente stem esigns Exeter 52 (0.06%) 120 (0.14%) 85,336 (26.2%) Lubinus SP II 18 (0.02%) 32 (0.03%) 94,917 (29.1%) Uncemente stem esigns Bi-Metric 129 (0.54%) 135 (0.56%) 23,943 (21.4%) CLS Spotorno 22 (0.29%) 25 (0.33%) 7692 (6.9%) Corail 51 (0.28%) 51 (0.28%) 17,932 (16.0%) ABG I 16 (0.38%) 19 (0.45%) 4186 (3.7%) ABG II 39 (0.78%) 46 (0.92%) 5024 (4.5%) *The values are given as the numbers of patients, with the row percentage in parentheses. The values are given as the number of patients available, with the percentage of the total number of either cemente stems or uncemente stems in parentheses. only the two selecte cemente stems (Lubinus SP II an Exeter) or the five selecte uncemente stems (Bi-Metric, CLS Spotorno, Corail, ABG I, an ABG II) were inclue in aition to the covariates presente above. All calculations use the compile atabase incluing all four countries. We use IBM SPSS Statistics, Version 20 (IBM, Armonk, New York). The proportional hazar assumption was controlle by plotting survival curves an TABLE III Relative Risk of Revision Due to Periprosthetic Fracture Variable No. of Patients Relative Risk* P Value Type of stem (unajuste) < Uncemente 111, (5.80 to 7.88) Cemente 325,730 1 (1) Type of stem (ajuste) < Uncemente 111, (7.37 to 10.32) Cemente 325,730 1 Age Younger than fifty years 22, (0.30 to 0.58) < Fifty to fifty-nine years 61, (0.44 to 0.69) < Sixty to sixty-nine years 131, (0.70 to 1.02) 0.07 Seventy to seventy-nine years 157,269 1 Eighty years or more 64, (0.91 to 1.49) 0.22 Diagnosis Primary osteoarthritis 353,446 1 Fracture 34, (1.95 to 2.98) < Peiatric hip isease 15, (0.76 to 1.57) 0.65 Inflammatory isease 15, (0.84 to 1.84) 0.27 Iiopathic femoral hea necrosis (1.35 to 2.95) ther (1.46 to 2.79) < Involve sie 0.30 Right 239, (0.81 to 1.07) Left 198,020 1 *The values are given as the relative risk, with the 95% CI in parentheses. This is the reference category for all cemente an uncemente stems, with sex exclue (see text).

4 e167(4) TABLE IV Relative Risk of Revision Due to Periprosthetic Fracture Ajuste for Diagnosis (Primary or Seconary steoarthritis) an Involve Sie in Age Groups an Separate by Female an Male Patients Age Group an Fixation No. of Patients Relative Risk* P Value Female Younger than fifty years 0.02 Uncemente (1.28 to 14.06) Cemente Fifty to fifty-nine years < Uncemente 16, (3.40 to 13.73) Cemente 15,895 1 Sixty to sixty-nine years Uncemente 21, (8.30 to 19.36) < Cemente 54,367 1 < Seventy to seventy-nine years < Uncemente 10, (11.71 to 24.22) Cemente 91,000 1 Eighty years or more < Uncemente (9.52 to 24.33) Cemente 43,146 1 Male Younger than fifty years 0.09 Uncemente (0.70 to 5.99) Cemente Fifty to fifty-nine years Uncemente 15, (1.56 to 5.35) Cemente 12,669 1 Sixty to sixty-nine years Uncemente 19, (2.83 to 6.06) < Cemente 36,719 1 < Seventy to seventy-nine years < Uncemente (4.75 to 10.73) Cemente 47,623 1 Eighty years or more < Uncemente (2.45 to 14.15) Cemente 17,204 1 *The values are given as the relative risk, with the 95% CI in parentheses. This is the reference category for all cemente an uncemente stems. by computing an plotting the Schoenfel resiuals for each covariate using Rstatistics. Source of Funing No external funing source was use in this stuy. Results The incience of periprosthetic femoral fracture at two years was low: 0.47% for uncemente stems an 0.07% for cemente stems (Table II). There were 768 revisions (238 cemente an530uncemente)uetoperiprostheticfracturewithintwo years (Table II), constituting 9.5% of all revisions. Nearly all of the fractures with uncemente stems occurre within the first six months, an fractures with cemente stems occurre mainly after six months. At six months, the crue revision rate of uncemente stems ue to periprosthetic femoral fracture was about ten times higher than that observe for cemente stems. At two years, the risk ifference ha ecrease an was about six times higher (Table II). Uncemente stems were more commonly use in male patients than in female patients an the mean patient age of the group with the uncemente stem at the time of arthroplasty was about ten years younger than in the group who unerwent arthroplasty with a cementestem(tablei),butallstemswereuseinallagegroups. The survival (an stanar eviation) at two years regarless of the reason for revision was 98.3% ± 0.05% for cemente stems an 97.1% ± 0.04% for uncemente stems. The corresponing survival (an stanar eviation) base on early periprosthetic

5 e167(5) femoral fracture was 99.9% ± 0.01% for cemente femoral stems an 99.5% ± 0.04% for uncemente femoral stems. verall, the unajuste relative rate for revision ue to periprosthetic fracture was higher for uncemente stems (relative risk, 6.76 [95% CI, 5.80 to 7.88]) (Table III). After ajustment for age, iagnosis, an involve sie, the relative risk increase further to 8.72 (95% CI, 7.37 to 10.32) (Table III). The risk of periprosthetic femoral fracture was increase for both preoperative femoral neck fracture an iiopathic femoral hea necrosis (Table III). The uncemente stems ha a higher relative risk for revision compare with cemente stems increasing with age in both male patients an female patients (Table IV). In male patients younger than fifty years of age, we foun no ifference (relative risk, 2.05 [95% CI, 0.70 to 5.99]), but the number of revise stems ue to fractures was very low (four cemente stems [0.1%] an twenty uncemente stems [0.2%]). From 1995 to 2002, 15.3% of all stems were uncemente, increasing to 34.7% from 2003 to The unajuste relative risk for revision within two years more than ouble in the perio 2003 to 2009 compare with the perio 1995 to 2002 (relative risk, 2.23 [95% CI, 1.95 to 2.67]; p < ). After ajustment for age, iagnosis, an sex, the relative risk ha still increase, but ha roppe to 1.44 (95% CI, 1.18 to 1.69) (p < ; etaile analysis not shown). A separate analysis of the entire cemente an uncemente groups (see Appenix) showe that the influence of sex an age iffere between the two types of stem fixation. In the cemente group, male patients were at a higher risk of revision compare with female patients (unajuste hazar ratio, 1.84 [95% CI, 1.44 to 2.35]) (see Appenix), an previous hip fracture was a risk factor for subsequent periprosthetic femoral fracture (see Appenix). In the uncemente group, male patients were at a lower risk of revision ue to periprosthetic fracture (unajuste hazar ratio, 0.69 [95% CI, 0.58 to 0.82]) (see Appenix). The Lubinus SP II an Exeter stems constitute 55.3% of all of the cemente cases. There were thirty-two periprosthetic fractures in the Lubinus SP II group an 120 periprosthetic fractures in the Exeter group. In the regression analyses, the Exeter stem ha about five times an increase risk to be associate with revision ue to this complication (see Appenix). The uncemente stem esigns specifically stuie constitute 52.5% of all of the uncemente stems. Compare with the reference an most frequently use Bi-Metric stem, the Corail stem was associate with a ecrease risk both before an after ajustment for the covariates stuie. The ABG II esign showe an increase risk, whereas the CLS Spotorno an ABG I esigns i not significantly iffer from the Bi-Metric esign (see Appenix). Discussion Periprosthetic femoral fracture is more common in uncementestemsaninpolishecementestemsanismost frequent uring the early postoperative months, increasing with age, especially in oler women. We observe a variation in the risk ratio among the five specific uncemente stem esigns stuie. The ABG II esign showe an increase risk an the Corail esign showe a ecrease risk compare with the Bi-Metric esign. verall, the number of revisions ue to early periprosthetic femoral fracture in the Scaninavian countries base on ata from all national registers is low (Table II). The avantages of this stuy were that it represents a wie spectrum of orthopaeic surgeons with variable clinical experience an covers the whole Noric region. Nonetheless, only a few esigns of stems coul be ientifie in sufficient numbers for a reliable evaluation of probable esign-relate features. However, confouning factors associate with register stuies require consieration. Uncomplicate periprosthetic fractures (e.g., those classifie as Vancouver type A 12 ) rarely result in revision an are not inclue in our stuy. If Vancouver type-b an C fractures with a high incience of complications an reoperations 1 are treate without exchange of the prosthesis, this surgical proceure is not recore as a revision an consequently isnotinclueinthepresentanalysis.thisunerreportingcoul also contribute to a istorte view of the fracture incience. Some implant-relate parameters may be biase by factors not known by us, such as use of ifferent types of incision, which is not consistently recore in the NARA atabase. Minimally invasive incisions, which theoretically coul prouce a higher number of unrecognize fractures, have not been extensively use in the Noric countries. In Sween, for example, this approach was use in <1% of the cases. Not unexpectely, previous hip fracture was a risk factor for subsequent periprosthetic femoral fracture. The reason why iiopathic necrosis of the femoral hea also is associate with an increase risk is not quite clear, but it coul also be because of osteoporosis or poor bone quality ue to other comorbiities 13. Systemic abnormalities 14, treatment with corticosterois, an alcoholism or substance abuse are known to be associate with reuce bone-mineral ensity 10 an increase risk of fracture. Most of the early fractures aroun uncemente stems were revise within six months. Some of them might have appeare uring surgery as minor fissures, which progresse to obvious clinical fractures uring the rehabilitation perio. Nonetheless, our analysis inicates that the incience of early periprosthetic femoral fracture resulting in revision has increase uring the later perio analyze even after ajustment for choice of stem fixation, age, an iagnosis. The reason for this increase is not known. Factors such as faster rehabilitation, shorter training perio before surgeons may operate inepenently, change of implant, an patient selection coul be possible causes. Uncemente stems may be more prone to fracture shoul the patient sustain trauma to the hip, as long as the stem has not establishe biological fixation. ne coul speculate that similar mechanisms are responsible for the increase risk of fracture with polishe stems not bone to the cement mantle. Some previous stuies using ual x-ray absorptiometry (DXA) have shown that, uring the postoperative year, the loss of bone mineral ensity is most pronounce, which in some of the regions might be followe by a small recovery For the Exeter stem, the unajuste survival showe a weak tenency to become less steep after about four to five months, possibly reflecting changes in bone metabolism. In the uncemente cases, the steep course of the survival curve up to two months postoperatively coul reflect elaye

6 e167(6) revisions of intraoperative fractures an an increase risk of fracture before ingrowth of the stem. Somewhat surprisingly, the influence of sex was reverse between use of cemente an uncemente fixation of the stem. Cementing the stem seems to have a protective effect against early periprosthetic fracture an the overall risk becomes very low. Physical activity an certain comorbiities associate with increase risk of trauma ue to fall might become equally an more ecisive factors than the bone quality. ur observation, contrary to previous finings 6, that age ha no certain influence on the risk of early periprosthetic fracture aroun a cemente stem might support this theory. Use of uncemente stems ha an eightfol to ninefol increase risk for revision ue to early periprosthetic fracture, which has been reporte previously 2. This risk increase was particularly high in female patients an, contrary to the finings with use of cement, the risk ecrease in the younger age groups. Early periprosthetic fracture is the thir most common complication following operation with hemiarthroplasty after femoral neck fracture 18 an the incience is higher with use of an uncemente moern stem esign, suggesting that uncemente fixation is not the first choice in oler female patients an patients with previous femoral neck fracture. The two most frequently use cemente femoral stems with high survival in the NARA atabase have completely ifferent shapes an surface finishes. The anatomic Lubinus stem is esigne to become fixe in the cement mantle, whereas the tapere Exeter stem is esigne to subsie insie the cement mantle to achieve an even loa bearing. Thus, the reason for the increase fracture risk with this stem might be similar to the one presente for uncemente stems. This material property is likely use in the Exeter esign as well as in the majority of polishe stems. Polishe stems generally are at higher risk for a periprosthetic fracture 4,19,20. However, polishe stems have an excellent overall track recor, particularly relate to the risk of loosening an osteolysis. In the NARA atabase, the stem survival (an stanar eviation) at fourteen years incluing all reasons for revision is 94.5% ± 1.0% for the Exeter stem an 95.4% ± 0.6% for the Lubinus stem, which is only marginally higher than the Exeter stem. The increase risk for the ABG II stem an the corresponing ecrease risk for the Corail stem are ifficult to interpret. It seems that a wege shape is superior to a more anatomic esign even if, from a theoretical point of view, the situation shoul be reverse. A wege-shape stem coul be expecte to more frequently act as a stress riser with its comparatively sharp corners compare with a roune esign. It might be that other factors such as the time between the insertion an rigi osseous fixation is shorter for the Corail stem. The length of the stem coul also be an issue, but oes not agree with the observation that the relative frequency of periprosthetic fracture aroun the ABG I stemwasonlyhalfofthatobservewiththeabgiistem. Revision ue to early periprosthetic femoral fracture is increasing in our stuy. Even if our analyses o not allow for establishment of a istinct age limit, they inicate that specific attention shoul be given to the bone quality in relation to choice of fixation in patients oler than sixty years of age an especially in female patients. Appenix Tables showing the unajuste an ajuste hazar ratios of risk of revision ue to periprosthetic fracture in male an female patients in the cemente an uncemente groups an the unajuste an ajuste hazar ratios of risk of revision ue to periprosthetic fracture for the two selecte cemente stems an the five selecte uncemente stems are available with the online version of this article as a ata supplement at jbjs.org. n Truike M. Thien, MD, PhD Georgios Chatziagorou, MD Göran Garellick, MD, PhD Johan Kärrholm, MD, PhD Institute of Clinical Sciences, The Sahlgrenska Acaemy, University of Göteborg, Box 426, Göteborg, Sween. aress for T.M. Thien: truike.thien@capio.se ve Furnes, MD, PhD Leif I. Havelin, MD, PhD The Norwegian Arthroplasty Register, Department of Clinical Meicine, University of Bergen, Jonas Lies vei 65, 5021 Bergen, Norway. aress for. Furnes: ove-furnes@helse-bergen.no. aress for L.I. Havelin: leif.havelin@helse-bergen.no Keijo Mäkelä, MD, PhD Department of rthopaeics an Traumatology, Turku University Hospital, P.. Box 52, Turku, Finlan. aress: keijo.makela@tyks.fi Søren vergaar, MD, PhD Institute of Clinical Research, University of Southern Denmark, Sr. Boulevar 29, 5000 ense, Denmark. aress: soeren.overgaar@ouh.regionsyanmark.k Alma Peersen, MD, PhD Department of Clinical Epiemiology, Aarhus University Hospital, lof Palmes Alle 43-45, 8200 Aarhus, Denmark. aress: abp@ce.au.k Antti Eskelinen, MD, PhD The Coxa Hospital for Joint Replacement, Box 652, Tampere, Finlan. aress: antii.eskelinen@coxa.fi Pekka Pulkkinen, MD, PhD Department of Public Health, University of Helsinki, Box 41, Helsinki, Finlan. aress: pekka.pulkkinen@tyks.fi

7 e167(7) References 1. Garellick G, Kärrholm J, Rogmark C, Herberts P. Annual report from the Sweish Hip Arthroplasty Register Accesse 2014 Apr Davison D, Pike J, Garbuz D, Duncan CP, Masri BA. Intraoperative periprosthetic fractures uring total hip arthroplasty. Evaluation an management. J Bone Joint Surg Am Sep;90(9): Franklin J, Malchau H. Risk factors for periprosthetic femoral fracture. Injury Jun;38(6): Epub 2007 Apr Linahl H, Malchau H, Herberts P, Garellick G. Periprosthetic femoral fractures classification an emographics of 1049 periprosthetic femoral fractures from the Sweish National Hip Arthroplasty Register. J Arthroplasty ct;20(7): Sarvilinna R, Huhtala HS, Sovelius RT, Halonen PJ, Nevalainen JK, Pajamäki KJ. Factors preisposing to periprosthetic fracture after hip arthroplasty: a case (n = 31)- control stuy. Acta rthop Scan Feb;75(1): Cook RE, Jenkins PJ, Walmsley PJ, Patton JT, Robinson CM. Risk factors for periprosthetic fractures of the hip: a survivorship analysis. Clin rthop Relat Res Jul;466(7): Epub 2008 May Herberts P, Malchau H. How outcome stuies have change total hip arthroplasty practices in Sween. Clin rthop Relat Res Nov;(344): Herberts P, Malchau H. Long-term registration has improve the quality of hip replacement: a review of the Sweish THR Register comparing 160,000 cases. Acta rthop Scan Apr;71(2): Malchau H, Herberts P, Eisler T, Garellick G, Söerman P. The Sweish Total Hip Replacement Register. J Bone Joint Surg Am. 2002;84(Suppl 2): Havelin LI, Fensta AM, Salomonsson R, Mehnert F, Furnes, vergaar S, Peersen AB, Herberts P, Kärrholm J, Garellick G. The Noric Arthroplasty Register Association: a unique collaboration between 3 national hip arthroplasty registries with 280,201 THRs. Acta rthop Aug;80(4): Havelin LI, Robertsson, Fensta AM, vergaar S, Garellick G, Furnes. A Scaninavian experience of register collaboration: the Noric Arthroplasty Register Association (NARA). J Bone Joint Surg Am Dec 21;93(Suppl 3): Bray H, Garbuz DS, Masri BA, Duncan CP. Classification of the hip. rthop Clin North Am Apr;30(2): Singh JA, Jensen MR, Harmsen SW, Lewallen DG. Are gener, comorbiity, an obesity risk factors for postoperative periprosthetic fractures after primary total hip arthroplasty? J Arthroplasty Jan;28(1): e1: 2. Epub 2012 Apr Chang CC, Greenspan A, Gershwin ME. steonecrosis: current perspectives on pathogenesis an treatment. Semin Arthritis Rheum Aug;23(1): Rahmy AI, Gosens T, Blake GM, Tonino A, Fogelman I. Periprosthetic bone remoelling of two types of uncemente femoral implant with proximal hyroxyapatite coating: a 3-year follow-up stuy aressing the influence of prosthesis esign an preoperative bone ensity on periprosthetic bone loss. steoporos Int Apr;15(4): Epub 2003 Dec Thien TM, Thanner J, Kärrholm J. Ranomize comparison between 3 surface treatments of a single anteverte stem esign: 84 hips followe for 5 years. J Arthroplasty Apr;25(3): e1. Epub 2009 Feb Thien TM, Thanner J, Kärrholm J. Fixation an bone remoeling aroun a lowmoulus stem seven-year follow-up of a ranomize stuy with use of raiostereometry an ual-energy x-ray absorptiometer. J Arthroplasty Jan; 27(1): e1. Epub 2011 May Leonarsson, Kärrholm J, Åkesson K, Garellick G, Rogmark C. Higher risk of reoperation for bipolar an uncemente hemiarthroplasty. Acta rthop ct;83(5): Epub 2012 Sep Nieuwenhuijse MJ, Valstar ER, Kaptein BL, Nelissen RG. The Exeter femoral stem continues to migrate uring its first ecae after implantation: years of follow-up with raiostereometric analysis (RSA). Acta rthop Apr;83(2): Epub 2012 Mar Sarvilinna R, Huhtala H, Pajamäki J. Young age an wege stem esign are risk factors for periprosthetic fracture after arthroplasty ue to hip fracture. A casecontrol stuy. Acta rthop Feb;76(1):56-60.