The Finnish Orthopaedic Association began to register arthroplasties in 1980. At present, the registry is run by the National Agency for Medicine. Registration of joint replacements was initially voluntary, but since the beginning of 1989, it has been mandatory for all Finnish orthopaedic surgeons who perform a primary or revision total hip arthroplasty to file standard forms with the National Arthroplasty Registry. The arthroplasty registry is linked and matched with related national data registries, such as the Population Register Centre, which detects whether the patient has died or moved to another area or emigrated, and to the Finnish Hospital Discharge Register. The Finnish Hospital Discharge Register was founded in 1976, is administered by the National Research and Development Centre for Welfare and Health, and records about 95% of all discharges from all hospitals in Finland1. In addition to demographic data, the dates of admission and discharge, the primary diagnosis and up to three subsidiary diagnoses, the surgical procedures performed, and the type of hospital are recorded. Data from the different databases are easily pooled with use of each patient's unique Social Security number, which is assigned to all Finnish citizens and immigrants. This means that the treatment of any individual patient can be tracked for his or her entire lifetime (called “treatment paths” in Finland) even if they move to another hospital district or are treated in a hospital other than the one where the primary operation was performed.
In the present study, data from the Finnish Arthroplasty Registry were used to analyze the results of primary total hip arthroplasty performed between 1980 and 2004 in patients eighty years of age and older and to analyze some of the factors that affected outcomes.
The goal of this study was to determine the effect of sex, diagnosis, year of the primary operation, design of the prosthesis, fixation method (cemented, hybrid, or cementless), type of operating hospital (university hospital, central hospital, regional hospital, or other hospital), and side of the involved hip on the outcome of primary total hip arthroplasty in patients eighty years of age and older.
Materials and Methods
The database maintained by the Finnish Arthroplasty Registry was used as the source for records. The database is subjected annually to an internal validation procedure in which the figures of the Finnish Arthroplasty Registry are compared with the nationwide Hospital Discharge Registry. Only records of primary total hip replacements in patients eighty years of age and older were included in the present study.
The Finnish Arthroplasty Registry contained information on 6989 primary total hip replacements performed between 1980 and 2004 in 6540 patients who were eighty years of age or older at the time of total hip arthroplasty. Only the hip that was operated on first was included in the analysis so that the study group consisted of 6540 hips. However, the data were separately analyzed for unilateral and bilateral operations. Population data from the same time period were collected from the Finnish Population Register.
The mean age of the patients at the time of the primary total hip arthroplasty was 82.7 years (range, eighty to ninety-seven years). The mean longevity of 3065 patients in this age group who died following total hip arthroplasty was 5.1 years (range, zero to twenty-one years). Demographic data on the patients are presented in Table I.
One hundred different designs of femoral components and 101 different designs of acetabular components were used. These were combined in 393 different ways. In one-third of these combinations, both the femoral and the acetabular component (121 combinations) were from the same manufacturer, but in the rest, the femoral and the acetabular components (272 combinations) were from different manufacturers. The most commonly used designs of the femoral component were the Exeter Universal (Stryker Orthopaedics, Mahwah, New Jersey), which was implanted in 29% of the hips, and the Lubinus SPII (Waldemar Link, Hamburg, Germany), which was used in 27% of the hips. The most commonly used designs of the acetabular component were the Lubinus STD (Waldemar Link) in 22% of the hips, the Exeter All-Poly (Stryker Orthopaedics) in 12% of the hips, and the Exeter Contemporary (Stryker Orthopaedics), which was implanted in 11% of the hips. The numbers for many of the designs were so small that they did not provide enough power for effective statistical analysis.
In 5047 (82%) of the 6175 hips included in the analysis, both components were fixed with cement. Seven hundred and twenty-nine hips (12%) had hybrid fixation, that is, either the femoral or acetabular component was fixed with cement, and the remaining 399 hips (6%) were inserted with cementless fixation. Of the hips with hybrid fixation, 91% had the femoral component fixed with cement and 9% had the acetabular component fixed with cement.
We defined (the first) revision as any operation on the hip that involved removal, exchange, or reimplantation of one, or both, of the prosthetic components.
The end point for Kaplan-Meier survivorship analysis was revision for any reason. Data on patients who did not undergo revision were censored on December 31, 2004, or at death, if death occurred before December 31, 2004.
The factors included in the statistical analyses are presented in Table I. Variable descriptors were checked in order to find any extreme values or errors in data input. Categorical variables were dummy-coded. For survival analyses, the original data file from the National Implant Registry was organized so that each row represented one hip. The steps in the analysis included testing for significant differences in survival with use of Kaplan-Meier survival analysis2 (with revision for any reason as the end point) and log-rank tests; testing for significant predictors, checking the adequacy of the proportional hazards (the probability of an end event) assumption3 by graphical examination of the partial residuals and, more formally, by testing the significance of time dependency (a trend in the partial residuals with time and the significance of the time-dependent covariate); calculating univariate statistics for each variable; entering significant variables into a multivariate Cox model; and Cox regression model diagnostics in order to determine whether the model adequately describes the data. In addition to the analysis of the proportional hazards assumption, as detailed above, model diagnostics included checking for influential observations4. In order to detect any exceptionally influential observations or outliers, dfbeta values, which estimate the changes in the regression coefficients on deleting each observation in turn, were calculated.
The initial significance level (p value) was set at 0.05. Because multiple predictors were analyzed, the p value was adjusted so that the more conservative value was used. Therefore, the exact p value for each statistical analysis is provided. The results are given as the mean (and 95% confidence interval) if not otherwise indicated. Binomial confidence intervals were calculated for the survival figures with use of Clinstat5.
The proportional hazards assumption was tested and was met for sex and fixation. In other words, the hazards (the risks of revision) associated with these variables were not dependent on time. The proportional hazards assumption was not met for the year of the primary operation and the design of the prosthesis. The hazards associated with these variables varied over time.
The nonproportional hazards of the year of the primary operation were accommodated for by use of a stratified Cox model, where each stratum (category of the year of the primary operation) is allowed to have a different baseline hazard function, while proportional hazards are assumed for the remaining covariates across the strata.
No extreme values or outlier observations that would have dominated the model individually and given a nonrepresentative model were found. With so many factors under investigation, multiple-comparisons techniques were employed and p values were appropriately adjusted. This was done with use of log-rank (Mantel-Cox) tests for overall comparisons of the predictors and for comparisons of each pair of factors composing the predictor.
Octogenarians as Total Hip Replacement Patients
Between 1980 and 2004, the number of people eighty years of age and over in Finland increased 2.35-fold (by 135%), from 86,480 to 203,318, and the number of primary total hip replacements performed increased from thirty-five to 583. The ratio (%) of the population eighty years of age and older undergoing total hip arthroplasty increased by sevenfold from 0.0405% to 0.2867% (see Appendix).
Reasons for Revisions
Of the total of 6989 primary total hip replacements, 195 (2.8%) were revised. For the 183 hips with complete data on the revision, the most common reasons for revision were aseptic loosening of the femoral or acetabular component, or both, which occurred in eighty-four hips (46%); recurrent dislocation, in thirty-six (20%); periprosthetic fracture, in twenty-four (13%); and infection, in twenty-three hips (13%). Malposition or fracture of the prosthesis and other causes were less frequent reasons for revision. In 150 (77%) of the 195 revisions, the femoral and/or acetabular components were exchanged. In twenty-three revisions (12%), the implants were removed but were not replaced with new implants. Eighteen (78%) of those implants were removed to treat infection. The remaining twenty-two revisions (11%) consisted of an isolated change of the liner and/or the modular femoral head or an open reduction for a recurrent dislocation.
Survivorship of the total hip replacement was significantly associated with female sex (p < 0.0005), hybrid fixation compared with cementless fixation (p < 0.05), and an earlier time period during which the primary total hip arthroplasty was performed (1980 to 1989 compared with 1995 to 1999; p < 0.05) (see Appendix).
The log-rank (Mantel-Cox) tests indicated that the diagnosis, design of the prosthesis, type of hospital, and side of the involved hip did not affect the survival of the primary hip replacements. These nonsignificant variables were excluded from the multivariate Cox analysis (see Appendix).
The overall survivorship of the prosthesis in the hips that had a primary operation, with revision for any reason as the end point, was 99% (95% confidence interval, 98% to 99%) at two years (4778 hips), 97% (95% confidence interval, 96% to 97%) at five years (2617 hips), and 94% (95% confidence interval, 93% to 95%) at ten years (532 hips) (Fig. 1). (Expanded ordinates from 90% to 100% for the Kaplan-Meier curves are shown in the Appendix.)
Prosthetic survivorship at five years, with revision for any reason as the end point, was 97% (95% confidence interval, 97% to 98%) for women and 95% (95% confidence interval, 94% to 97%) for men (p < 0.0005) (Fig. 2).
Prosthetic survivorship was slightly better for patients who had the primary operation in the 1980s than it was for those who had the operation in more recent years (Fig. 3). The difference in survivorship reached significance (p < 0.05) only for those who underwent surgery between 1980 and 1989 compared with those who had the operation between 1995 and 1999; the survival rate at five years was 98% (95% confidence interval, 97% to 99%) and 96% (95% confidence interval, 95% to 97%), respectively.
Patients with hybrid fixation had slightly better prosthetic survival than those with cementless or cemented fixation (Fig. 4). The difference reached significance (p < 0.05) only when patients with hybrid fixation were compared with those with cementless fixation; survival at five years was 98% (95% confidence interval, 97% to 99%) and 95% (95% confidence interval, 92% to 98%), respectively. A subanalysis indicated that, for the patients with cementless acetabular fixation, survival was better for those with a cemented femoral component compared with those with a cementless femoral component (p < 0.05). In the patients with a cemented acetabular component, there was no significant difference in the survival of the prosthesis between those with a cemented femoral component and those with a cementless femoral component.
No significant differences in survivorship were found between matched (from the same manufacturer) and mismatched femoral and acetabular components. The diagnosis, design of prosthesis, type of hospital, and side of the involved hip did not significantly affect the survival.
Multivariate Cox Regression Analysis
Significant variables were included in the multivariate Cox analysis. Sex was again found to be highly significant (p < 0.0005), with women having better prosthetic survivorship. The risk of revision was 1.8 (95% confidence interval, 1.3 to 2.5) times higher for men than for women.
Hybrid fixation was also associated with better prosthetic survivorship compared with cementless fixation (p < 0.05). The risk of revision for patients with hybrid fixation was 0.4 (95% confidence interval, 0.2 to 1) times that for patients with cementless fixation.
As the year of the primary operation was used as a stratification variable, it was not possible to examine the effect of this covariate. Estimated hazard ratios in multivariate Cox analyses are presented in a table in the Appendix.
In Finland, the increase in the number of patients eighty years of age and older has been projected to be 41% from 2002 to 2015 and 140% from 2002 to 20306. This increase will be accompanied by an increased demand for total hip replacement7-13. However, in this age group, it is equally important to note that the likelihood of death is a “competing risk” reducing the likelihood of the demand.
In our registry as a whole, the most common reason for patients of all ages to undergo revision total hip replacement was aseptic loosening, which was the cause of 82% of the revisions in 1991 and 59% of the revisions in 200114. In contrast, the prevalence of revision for a diagnosis of aseptic loosening in patients who were eighty years of age and older was only 46% of all revisions, and it may be explained by the lower physical demands15 and shorter longevity of these patients. In contrast, dislocation, periprosthetic fracture, and infection were more common reasons for revision in the octogenarians compared with all other patients.
Our end point, revision for any reason, may not represent the actual number of patients in this age group who might be candidates for a revision as some may have refused surgery for a variety of reasons. The “risk of revision” in this context should be interpreted cautiously, as it may represent a willingness of the patient to undergo revision, once a diagnosis has been made, which may be different from the actual prevalence of prosthetic failure.
For patients in this age group, the surgeon may not propose revision as aggressively as he or she might to a younger patient, particularly if the loosening is relatively symptom-free and mainly diagnosed on the basis of radiographic findings. Even if the surgeon proposes revision, the patient may refuse and prefer nonoperative treatments.
Survivorship was worse for the period from 1995 to 1999 compared with the period from 1980 to 1989. The extensive use of the cementless Biomet cups with HexLoc liners (Biomet, Warsaw, Indiana) in Finland may explain this shortened overall survival of hip implants during 1995 to 1999, as most of them were revised during that period16 (the nine-year survival of Biomet cups was only 65% and that of T-Tap cups [Biomet] was only 58%).
In North America, there is a large reluctance to use mismatched components. In the present series, almost 400 different stem and cup combinations were used, often by matching components from different manufacturers. Despite this, the present analysis did not show a significant difference in survival with mismatched components. This may in part be due to the lack of statistical power as so many different mismatched components were used and as the information on the type of component used was often missing. One reason for the use of ad hoc mismatched components in Finland might be the relatively small size of the inventories of prostheses in many orthopaedic units and a lack of adequate preoperative planning so that, at the time of the surgery, appropriately sized and matched components were not available to the surgeon.
We found that hybrid fixation in elderly patients was significantly better than cementless fixation. However, hybrid fixation was not significantly better than cemented fixation.
Our results agree with those of other investigators17-23 who have reported that women had better overall results than men. This difference has been attributed to less biomechanical loading in women, who in general are smaller than men17. In addition, more men than women had a moderate activity level (more than four hours per week)24, especially with regard to rates of heavy work, such as gardening25.
Because of a lack of national guidelines and because of aggressive marketing, Finland has traditionally been an experimental ground for implant companies and for new innovations and prosthetic designs. At present, this shows up in the extensive use of “hybrid” fixation and the use of multiple stem-cup combinations. Most of these combinations are so rare that the important task of performing postmarketing registry surveillance is not possible because the analyses currently lack statistical power. With further follow-up, some trends may emerge.
Figures and tables showing survivorship curves (with the Y axis from 90% to 100%), the number of octogenarians in Finland from 1980 to 2004 compared with the number and ratio of senior citizens undergoing total hip replacement, selected factors associated with survival, the log-rank (Mantel-Cox) tests for overall comparisons of the predictors, and estimated hazard ratios in multivariate Cox analyses are available with the electronic versions of this article, on our web site at jbjs.org (go to the article citation and click on “Supplementary Material”) and on our quarterly CD/DVD (call our subscription department, at 781-449-9780, to order the CD or DVD).
NOTE: The authors thank Juha Nevalainen, MD, Associate Professor and former Head of the Finnish National Arthroplasty Registry, for his help with the revision of this manuscript.
Disclosure: In support of their research for or preparation of this work, one or more of the authors received, in any one year, outside funding or grants in excess of $10,000 from the Sigrid Jusélius Foundation, EVO grants, and Finska Läkaresällskapet. Neither they nor a member of their immediate families received payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity. No commercial entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, division, center, clinical practice, or other charitable or nonprofit organization with which the authors, or a member of their immediate families, are affiliated or associated.
Investigation performed at COXA Hospital for Joint Replacement, Tampere; ORTON Orthopaedic Hospital, Invalid Foundation, Helsinki; National Agency for Medicines, Helsinki; and Department of Medicine and Department of Orthopaedics and Traumatology, Helsinki University Central Hospital, Helsinki, Finland
1. Gissler M, Haukka J. Finnish health and social welfare registers in epidemiological research. Norsk Epidemiologi. 2004;14:113-20.
2. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457-81.
3. Bland M. An introduction to medical statistics. 3rd ed. New York: Oxford University Press; 2000. Survival data using Cox regression; p 324-5.
4. Fox J. Cox proportional-hazards regression for survival data. Appendix to “An R and S-PLUS companion to applied regression.” http://cran.r-project.org/doc/contrib/Fox-Companion/appendix-cox-regression.pdf. Accessed 2006 Mar 24.
5. Bland M. An introduction to medical statistics. 3rd ed. New York: Oxford University Press; 2000. Statistics and computing; p 2-3.
6. Martelin T, Sainio P, Koskinen S. Ikääntyvän väestön toimintakyvyn kehitys. [The development of functional ability of aging population]. In: Kautto M, ed. Ikääntyminen voimavarana—Tulevaisuusselonteon liiteraportti 5. Valtioneuvoston kanslian julkaisusarja. [Aging as a resource-Report appendix 5 of the review of the future. Publication series of the Office of the Council of State.] 2004;33:117-31. Finnish.
7. 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.
8. 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.
9. Frankel S, Eachus J, Pearson N, Greenwood R, Chan P, Peters TJ, Donovan J, Smith GD, Dieppe P. Population requirement for primary hip-replacement surgery: a cross-sectional study. Lancet. 1999;353:1304-9.
10. Ostendorf M, Johnell O, Malchau H, Dhert WJ, Schrijvers AJ, Verbout AJ. The epidemiology of total hip replacement in The Netherlands and Sweden: present status and future needs. Acta Orthop Scand. 2002;73:282-6.
11. Wells VM, Hearn TC, McCaul KA, Anderton SM, Wigg AE, Graves SE. Changing incidence of primary total hip arthroplasty and total knee arthroplasty for primary osteoarthritis. J Arthroplasty. 2002;17:267-73.
12. Bourne RB, Maloney WJ, Wright JG. An AOA critical issue. The outcome of the outcomes movement. J Bone Joint Surg Am. 2004;86:633-40.
13. Pedersen AB, Johnsen SP, Overgaard S, Søballe K, Sørensen HT, Lucht U. Total hip arthroplasty in Denmark: incidence of primary operations and revisions during 1996-2002 and estimated future demands. Acta Orthop. 2005;76:182-9.
14. Nevalainen J, Keinonen A, Mäkelä A. The 2000-2001 implant yearbook on orthopaedic endoprostheses. Finnish Arthroplasty Register. Helsinki, Finland: National Agency for Medicines; 2003.
15. Schmalzried TP, Szuszczewicz ES, Northfield MR, Akizuki KH, Frankel RE, Belcher G, Amstutz HC. Quantitative assessment of walking activity after total hip or knee replacement. J Bone Joint Surg Am. 1998;80:54-9.
16. Puolakka TJ, Pajamäki KJ, Pulkkinen PO, Nevalainen JK. Poor survival of cementless Biomet total hip: a report on 1,047 hips from the Finnish Arthroplasty Register. Acta Orthop Scand. 1999;70:425-9.
17. Schurman DJ, Bloch DA, Segal MR, Tanner CM. Conventional cemented total hip arthroplasty. Assessment of clinical factors associated with revision for mechanical failure. Clin Orthop Relat Res. 1989;240:173-80.
18. Havelin LI, Espehaug B, Vollset SE, Engesæter LB. Early failures among 14,009 cemented and 1,326 uncemented prostheses for primary coxarthrosis. The Norwegian Arthroplasty Register, 1987-1992. Acta Orthop Scand. 1994;65:1-6.
19. Johnsson R, Franzén H, Nilsson LT. Combined survivorship and multivariate analyses of revisions in 799 hip prostheses. A 10- to 20-year review of mechanical loosening. J Bone Joint Surg Br. 1994;76:439-43.
20. Lucht U. The Danish Hip Arthroplasty Register. Acta Orthop Scand. 2000;71:433-9.
21. Malchau H, Herberts P, Söderman P, Odén A. Prognosis of total hip replacement: update and validation of results from the Swedish National Hip Arthroplasty Register 1979-1998. Presented as a scientific exhibit at the Annual Meeting of the American Academy of Orthopaedic Surgeons; 2000 Mar 15-18; Orlando, FL.
22. Rand JA, Trousdale RT, Ilstrup DM, Harmsen WS. Factors affecting the durability of primary total knee prostheses. J Bone Joint Surg Am. 2003;85:259-65.
23. Nevalainen J, Hirvonen A, Pulkkinen P. The 1997 implant yearbook on orthopaedic endoprostheses. Helsinki, Finland: National Agency for Medicines; 1998.
24. Høidrup S, Sørensen TI, Strøger U, Lauritzen JB, Schroll M, Grønbaek M. Leisure-time physical activity levels and changes in relation to risk of hip fracture in men and women. Am J Epidemiol. 2001;154:60-8.
25. Sims J, Hill K, Davidson S, Gunn J, Huang N. A snapshot of the prevalence of physical activity amongst older, community dwelling people in Victoria, Australia: patterns across the ‘young-old’ and ‘old-old’. BMC Geriatr. 2007;7:4.