Background: The purpose of this study was to quantify the procedural rate and revision burden of total hip and knee arthroplasty in the United States and to determine if the age or gender-based procedural rates and overall revision burden are changing over time.

Methods: The National Hospital Discharge Survey (NHDS) for 1990 through 2002 was used in conjunction with United States Census data to quantify the rates of primary and revision arthroplasty as a function of age and gender within the United States with use of methodology published by the American Academy of Orthopaedic Surgeons. Poisson regression analysis was used to evaluate the procedural rate and to determine year-to-year trends in primary and revision arthroplasty rates as a function of both age and gender.

Results: Both the number and the rate of total hip and knee arthroplasties (particularly knee arthroplasties) increased steadily between 1990 and 2002. Over the thirteen years, the rate of primary total hip arthroplasties per 100,000 persons increased by approximately 50%, whereas the corresponding rate of primary total knee arthroplasties almost tripled. The rate of revision total hip arthroplasties increased by 3.7 procedures per 100,000 persons per decade, and that of revision total knee arthroplasties, by 5.4 procedures per 100,000 persons per decade. However, the mean revision burden of 17.5% for total hip arthroplasty was more than twice that for total knee arthroplasty (8.2%), and this did not change substantially over time.

Conclusions: The number and prevalence of primary hip and knee replacements increased substantially in the United States between 1990 and 2002, but the trend was considerably more pronounced for primary total knee arthroplasty.

Clinical Relevance: The reported prevalence trends have important ramifications with regard to the number of joint replacements expected to be performed by orthopaedic surgeons in the future. Because the revision burden has been relatively constant over time, we can expect that a greater number of primary replacements will result in a greater number of revisions unless some limiting mechanism can be successfully implemented to reduce the future revision burden.