Our data indicate that between 1990 and 2002, the number and prevalence of primary hip and knee replacement procedures increased substantially in the United States. The trend was considerably more pronounced for primary total knee arthroplasty, the prevalence of which almost tripled during the study period. The explosive growth in both the number and the prevalence of primary total knee arthroplasty for both genders and all age categories can be explained, at least in part, by the increased recognition, by the candidates for surgery and the members of the orthopaedic community, of the effectiveness of the procedure in treating degenerative disease of the knee. Furthermore, the importance of this treatment modality was reflected in the National Institutes of Health (NIH) Consensus Development Conference on Total Knee Replacement in 200311. In addition, there is a strong relationship between obesity and the risk of osteoarthritis, especially in the knee12,13. Recent epidemiological research has indicated that the prevalence of obesity in the United States has increased over the past decade14,15, which may help to explain the increase in total knee arthroplasty rates. Although we could not directly correlate the prevalence of obesity and the prevalence of joint replacement on the basis of the data available in the NHDS, the major increase in the population of obese patients in the United States appears to contribute to the increasing trends in primary joint replacement observed in this study.
The number and prevalence of revision total hip and knee arthroplasties also increased substantially in the United States between 1990 and 2002. The number of primary and revision hip procedures increased by 62% and 79%, respectively, and the number of both primary and revision knee procedures increased by approximately 200%. Thus, the increasing number of revision knee procedures appears to be fueled by the growth in primary procedures.
The discrepancy between overall hip and knee revision burdens in the United States may be partly explained by differences in the sizes of the patient populations treated with each type of joint replacement. In 1990, the patient population in which a hip prosthesis was implanted (2.17 million) was 2.4 times greater than that in which a knee prosthesis was implanted (0.91 million), so there was a greater pool of hip implants from which revisions may result. This partly explains the greater hip revision burden (16.7%) compared with the knee revision burden (8.4%) in 1990. In 2002, the installed base for hip prostheses was only 1.3 times that of the base for knee prostheses. However, the hip revision burden (18.1%) remained more than twice that of the knee revision burden (8.4%), which suggests that this difference may only be partially explained by differences in the patient populations treated with the different implants.
With the exception of total hip arthroplasties performed in Sweden, which will be discussed below, the revision burden in the United States compared favorably with that in several countries with established total joint registries (Table II). Overall, the total hip arthroplasty revision burden of 17.5% in the United States from 1990 through 2002 fell within the range of revision burdens of 15.0% to 18.3% observed in Norway, Finland, and Australia16-18. In Canada, the revision burden for total hip arthroplasty was somewhat lower (13.1% for 2002 to 200319). Similarly, the United States revision burden for total knee arthroplasty of 8.2% was comparable with the revision burdens of 7.2% to 8.0% for Norway, Sweden, and Finland16,18,20. The Canadian knee revision burden was slightly lower (6.1% for 2002 to 2003)19, whereas the Australian knee revision burden was slightly greater (10.8% for 1999 through 2002)17.
The overall revision burden for total hip arthroplasty in the United States was substantially greater than the revision burden of 7.7% to 11% reported for Sweden10,21 (Table II). The differences between the Swedish and United States hip revision burdens may be due to the differences in prosthetic survivorship in the elderly population with hip implants. The revision hip burden (from 1992 through 2000) of 6.4% for the elderly population (more than sixty-five years old) in Sweden was substantially lower than the burden of 11% for the general Swedish population. Because the majority of patients requiring total hip arthroplasty are elderly, the hip revision burden for the overall population will be significantly affected by the revision burden for elderly patients. Conversely, the revision hip burden from 1990 through 2002 for the elderly population in the United States (16.9%) was similar to that for the general population (17.5%).
In the United States, the revision burden for hip arthroplasty was approximately twice that for knee arthroplasty from 1990 through 2002, a finding that is consistent with the differences in hip and knee revision burdens reported in Finland, Norway, Canada, and Australia. Conversely, the Swedish hip-to-knee revision burden ratio was essentially unity. The steady trend in revision burden over the years is especially disconcerting when the installed patient base continues to increase, as was observed in the United States between 1990 and 2002. This is particularly true in the case of total knee replacements.
To our knowledge, this is the first population-based study, based on NHDS data, of the rates of revision of both total hip and knee replacements. The procedural rates of primary and revision hip arthroplasty between 1995 and 1996 (Figs. 2-A and 2-B) were consistent with those previously calculated for the same time period in a Medicare population-based study5. In the present study, year-matched procedural rates for men and women were significantly different for all operations except for revision total knee arthroplasty. The procedure rates for women were generally greater than those for men. The small differences between the rates determined in our study and the rates reported by Mahomed et al.5 are due to differences in the data source, exclusion criteria, and age-group stratification.
The hip and knee revision burdens identified in this study have important implications for health costs in the United States. For example, if the 2002 hip revision burden of 18.1% was reduced by 1% (a decrease of approximately 2844 hip revision procedures), the potential cost savings could range from $42.5 million to $112.6 million, on the basis of recent procedural cost estimates of $15,000 to $40,000 for revision total hip arthroplasty7,8,22,23. Similarly, the potential cost savings for a 1% reduction in the knee revision burden (which was 8.4% in 2002) could range from $53.5 million to $98.4 million, on the basis of a decrease of approximately 4497 knee revision procedures and procedural cost estimates of $11,900 to $21,90022,23. Hence, a 1% decrease in the revision burden for both hip and knee arthroplasty would translate into an overall savings of between $96.0 million and $211.0 million for the United States health care system.
Although NHDS data for the age-groups of less than forty-five years old and of eighty-five years old or more were included in the present analyses, conclusions about age effects should be limited to the remaining age-groups (forty-five to eighty-four years old). The small sample sizes (less than thirty patients) resulted in large data variability within these two age-groups. The use of the National Inpatient Sample, which has a substantially (twenty-five times) larger number of sampled records than does the NHDS, may alleviate the limitations regarding the ability to analyze age effect with use of the NHDS data. However, the AAOS used NHDS data to make projections of total knee arthroplasty and total hip arthroplasty discharges, within the same age-groups and for the same procedures that we studied, so we used NHDS data for consistency and comparative purposes. In addition, since the number of years (ten to twenty) spanned by each age classification is large compared with the total thirteen-year time period evaluated in this study, the study may have had some limitations associated with demographic changes within the age-groups themselves. Perhaps, if we had evaluated smaller age-groups (e.g., spanning five years) among the elderly, we would have been better able to capture those differences.
Another limitation of the study pertains to the system for coding of the type of arthroplasty. Although the number of total knee replacements far exceeds the number of unicondylar knee replacements, the current ICD-9 coding system does not allow detailed classification of knee arthroplasty. As surgical techniques improve and the demand for less invasive surgery and shorter recovery time increases, the number of unicondylar knee procedures may increase substantially. Such an increase would require clearer differentiation between total and unicondylar knee procedures in order to better assess the clinical and economic impacts of the two procedures. Also, primary partial hip replacements (hemiarthroplasties) were excluded from the present study to limit the comparisons to primary total and revision arthroplasties.
As noted in previous studies5,24, analysis of total joint replacements with the use of data derived from administrative claims may be limited by random coding inaccuracies as well as by the more serious problem of systematic bias (e.g., underreporting), which are inherent in the design and implementation of claims databases. In the present study, we focused on joint replacement procedures, not the underlying diagnoses of the patients receiving them. Furthermore, coding inaccuracies may be considered as random uncertainty, which was accounted for in our statistical analysis of the entire thirteen-year NHDS data set.
We validated our procedure counts by using the four-year NHDS data (for 1996 through 1999) that had been previously reported by the AAOS9, and we confirmed that we were accurately estimating the national total-joint-replacement rates on the basis of the available data. The accuracy of NHDS-based national projections of joint arthroplasty procedure counts has been verified by the Department of Health and Human Services with use of the independent National Inpatient Sample (NIS) database25. The discrepancies between national estimates of hip and knee arthroplasty rates based on 1995 NHDS and NIS data were 9% and 6%, respectively26. Consequently, the systematic bias in population-based estimates of total hip and knee arthroplasty rates based on NHDS data was judged to fall within an acceptable margin of uncertainty (≤10%) for the type of analyses performed in this study.
Previous investigators have raised concerns about the implantation of joint replacements in younger patients27. Our results showed that the rates of primary hip and knee procedures were highest in people between the ages of sixty-five and eighty-four years. Similar trends were observed for revision hip and knee replacements, with the greatest prevalences in people between the ages of sixty-five and eighty-four years. Although we found little evidence to suggest that the increases in the prevalence of joint replacements between 1990 and 2002 were due to a disproportionate increase in the prevalence of patients younger than sixty-five years old, an evaluation of the raw procedure counts suggested otherwise. In the population of patients treated with primary total hip arthroplasty between the ages of forty-five and eighty-four years, the relative proportion who were between forty-five and sixty-four years old increased from 28% to 40% between 1990 and 2002. In the population of patients treated with primary total knee arthroplasty between the ages of forty-five and eighty-four years, the relative proportion who were between forty-five and sixty-four years old increased from 26% to 36% between 1990 and 2002. However, a similar relative increase was not observed in the relative proportion of patients between the ages of forty-five and sixty-four years undergoing revision total hip arthroplasty (36% in 1990 and 32% in 2002) or revision total knee arthroplasty (49% in 1990 and 25% in 2002). The discrepancy between the relative proportion and the prevalence (procedure rate) of primary total hip and knee arthroplasties among the younger patients is masked by the 42% increase in the forty-five to sixty-four-year-old patient population from 1990 to 2002 compared with the 1% and 26% increases in the sixty-five to seventy-four-year-old and seventy-five to eighty-four-year-old patient populations. Hence, the increasing number of procedures performed on younger patients is not reflected by the population-normalized prevalence (procedure rate). However, the use of procedural prevalence is relevant to account for year-to-year variation in patient populations. Future studies will be done to examine age and gender-stratified procedure counts in more detail.
The prevalence trends reported in the current study have important ramifications with regard to the number of hip and knee replacements expected to be performed in the future. The projected increase in the number of hip and knee replacements in a previous AAOS model9 was fueled solely by the anticipated growth of the United States population and was based on a constant prevalence of arthroplasty for a four-year window of data reported in the NHDS. However, our study has demonstrated that the mean procedural rates from 1996 through 1999 underestimate the future total hip and knee arthroplasty rates, as evidenced by the rapidly increasing rates of both primary and revision hip and knee arthroplasties between 1990 and 2002. Consequently, the projections of future primary joint replacement rates reported by the AAOS should be interpreted as a conservative lower boundary if the prevalence trends observed during the thirteen-year time-period that we studied continue forward in time indefinitely. Because the revision burden has been relatively constant between 1990 and 2002, we can expect that a greater number of primary replacements will, in turn, result in a greater number of revisions unless some limiting mechanism can be implemented to reduce the future revision burden.
Tables representing the overall procedure counts and rates and the annual changes in rate ratios based on age and gender as well as figures depicting procedural rates of revision arthroplasties stratified by gender and by age-group 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-ROM (call our subscription department, at 781-449-9780, to order the CD-ROM).
Investigation performed at Exponent Inc., Philadelphia, Pennsylvania
The authors did not receive grants or outside funding in support of their research or preparation of this manuscript. None of the authors 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, educational institution, or other charitable or nonprofit organization with which the authors are affiliated or associated.
1. , Barrett J, Katz JN, Liang MH. Total hip arthroplasty: use and select complications in the US Medicare population. Am J Public Health. 1996;86: 70-2.
2. , Losina E, Barrett J, Phillips CB, Mahomed NN, Lew RA, Guadagnoli E, Harris WH, Poss R, Baron JA. Association between hospital and surgeon procedure volume and outcomes of total hip replacement in the United States medicare population. J Bone Joint Surg Am. 2001;83: 1622-9.
3. , Losina E, Baron JA, Fossel AH, Guadagnoli E, Lingard EA, Miner A, Phillips CB, Katz JN. Contribution of hospital characteristics to the volume-out-come relationship: dislocation and infection following total hip replacement surgery. Arthritis Rheum. 2002;46: 2436-44.
4. , Phillips CB, Baron JA, Fossel AH, Mahomed NN, Barrett J, Lingard EA, Harris WH, Poss R, Lew RA, Guadagnoli E, Wright EA, Losina E. Association of hospital and surgeon volume of total hip replacement with functional status and satisfaction three years following surgery. Arthritis Rheum. 2003;48: 560-8.
5. , Barrett JA, Katz JN, Phillips CB, Losina E, Lew RA, Guadagnoli E, Harris WH, Poss R, Baron JA. Rates and outcomes of primary and revision total hip replacement in the United States medicare population. J Bone Joint Surg Am. 2003;85: 27-32.
6. , Barrett J, Mahomed NN, Baron JA, Katz JN. Early failures of total hip replacement: effect of surgeon volume. Arthritis Rheum. 2004;50: 1338-43.
7. . Economics of revision total hip arthroplasty. Clin Orthop Relat Res. 1995;319: 209-14.
8. , Sawhney J, Hsu J, Cofield RH. Cost analysis of revision total hip arthroplasty. A 5-year followup study. Clin Orthop Relat Res. 1999;369: 175-8.
9. Frankowski JJ, Watkins-Castillo S. Primary total knee and hip arthroplasty projections for the U.S. population to the year 2030. 2002. http://www.aaos.org
10. , Herberts P, Eisler T, Garellick G, Soderman P. The Swedish Total Hip Replacement Register. J Bone Joint Surg Am. 2002;84 Suppl 2:2-20. Erratum in: J Bone Joint Surg Am. 2004;86: 363.
11. National Institutes of Health. NIH Consensus Development Conference on Total Knee Replacement. 2003. consensus.nih.gov/cons/117/117cdc_intro.htm.
12. , Hart D, Spector TD, MacGregor AJ. The association of body mass index and osteoarthritis of the knee joint: an examination of genetic and environmental influences. Arthritis Rheum. 2003;48: 1024-9.
13. , March L, Sambrook PN. Knee osteoarthritis: the influence of environmental factors. Clin Exp Rheumatol. 2002;20: 93-100.
14. , Ford ES, Bowman BA, Dietz WH, Vinicor F, Bales VS, Marks JS. Prevalence of obesity, diabetes, and obesity-related health risk factors, 2001. JAMA. 2003;289: 76-9.
15. , Ogden CL, Johnson CL, Carroll MD, Curtin LR, Flegal KM. Prevalence of overweight and obesity among US children, adolescents, and adults, 1999-2002. JAMA. 2004;291: 2847-50.
16. . The Norwegian Joint Registry. Bull Hosp Jt Dis. 1999;58: 139-47.
17. Australian Orthopaedic Association. National Joint Replacement Registry annual report 2003. 2003. http://www.dmac.adelaide.edu.au
18. National Agency for Medicines. The 2000-2001 implant yearbook on orthopaedic endoprostheses: Finnish Arthroplasty Register. 2003. http://www.nam.fi
19. , Canadian Joint Replacement Registry (CJRR). 2004 report: total hip and total knee replacements in Canada. 2004.
20. , Knutson K, Lewold S, Lidgren L. The Swedish Knee Arthroplasty Register 1975-1997: an update with special emphasis on 41,223 knees operated on in 1988-1997. Acta Orthop Scand. 2001;72: 503-13.
21. Malchau H, Herberts P, Garellick G, Soderman P, Eisler T. Prognosis of total hip replacement. Update of results and risk-ratio analysis for revision and re-revision from the Swedish National Hip Arthroplasty Register 1979-2000. 2002. http://www.jru.orthop.gu.se.
22. , Healy WL, Richards JA. Comparison of the hospital cost of primary and revision total hip arthroplasty after cost containment. Orthopedics. 1999;22: 185-9.
23. , Drakeford MK, Tsao AK, Gittelsohn A, Krackow KA, Hungerford DS. Revision and primary hip and knee arthroplasty. A cost analysis. Clin Orthop Relat Res. 1995;311: 136-41.
24. , Barrett J, Baron JA, Katz JN. Accuracy of Medicare claims data for rheumatologic diagnoses in total hip replacement recipients. J Clin Epidemiol. 2003;56: 515-9.
25. Agency for Healthcare Research and Quality. Comparisons by hospital characteristics. In: comparative analysis of HCUP and NHDS inpatient discharge data. Technical supplement 13, NIH Release 5. Rockville, MD: Agency for Health Care Policy and Research. http://www.ahrq.gov
26. Agency for Healthcare Research and Quality. NIS and NHDS comparisons by primary procedures ranked by NIS data, 1995. In: comparative analysis of HCUP and NHDS inpatient discharge data. Technical supplement 13, NIH Release 5. Rockville, MD: Agency for Health Care Policy and Research. http://www.ahrq.gov
Copyright 2005 by The Journal of Bone and Joint Surgery, Incorporated
27. . Implant wear in total joint replacement: clinical and biological issues, material and design considerations. Rosemont, IL: American Academy of Orthopaedic Surgeons; 2001.