RESUMEN
BACKGROUND: Total joint replacements (TJRs) should be considered as one of few definite endpoints in osteoarthritis research. We analyzed factors associated with late-life prevalence and risk factors for incidence of TJRs due to osteoarthritis in a population based cohort. METHODS: After exclusion of inflammatory arthritis and fractures as causes of TJR, 5170 participants in the AGES-Reykjavik Study (mean age (SD) 76.4(6), 58% females) were included for osteoarthritis studies. Three thousand one hundred thirty-three of them had a follow-up visit 5 years later. RESULTS: The prevalence of having at least one joint replacement operation due to OA was 13.6% and the yearly incidence was 1.4%/year during the five-year follow-up. Factors positively associated with late life prevalence of TJR included BMI, hand OA severity, female gender, finger length ratio and spine BMD. Risk factors for TJRs in the incidence group were symptoms at initial visit, prior TJR in the contralateral joint and BMI. Much stronger associations were seen for TKR than for THR with discriminatory analysis showing an AUC 0.71 for late life prevalence and 0.84 for the incidence. CONCLUSIONS: This study illustrates the importance of the different information expressed by late life prevalence vs. incidence on the factors associated with severe osteoarthritis of the knee and hip. The observation that prior TJR is a risk factor for subsequent TJR in the contralateral joint has not been described previously. The high power predictions for TKR suggest that a predictive model may be feasible, particularly if it can be extended by the addition of further predictive variables, perhaps through genetic, biomarker or imaging data.