Defining and evaluating a novel outcome measure representing end-stage knee osteoarthritis: data from the Osteoarthritis Initiative.

We described a definition of end-stage knee osteoarthritis (esKOA) and evaluated its association with health outcomes and osteoarthritis risk factors. We included Osteoarthritis Initiative participants with or at risk for knee osteoarthritis who had complete baseline data. We defined esKOA by adapting a validated appropriateness algorithm for total knee replacement based on data from baseline and the first four follow-up visits. We performed person-based analyses, including both knees from all participants. Participants met the definition of esKOA at the visit at which ≥1 knee reached the esKOA criteria. We assessed differences in individual characteristics between groups at baseline and over time and tested if incident esKOA (outcome) was associated with osteoarthritis risk factors (e.g., age, maximum adult weight, and quadriceps strength). The cohort consisted of 3916 participants with mean age of 61 (SD = 9) years and mean body mass index of 28.4 (4.7) kg/m(2); 59 % were female and 9.7 % developed incident esKOA. Those with incident esKOA had poorer health outcomes at baseline and greater declines in health outcomes, with the exception of SF-12 mental health score. Five out of nine tested risk factors were associated with incident esKOA in unadjusted analyses, with older age (≥65 years; odds ratio = 1.44, 95 % confidence interval = 1.19 to 1.83) and quadriceps weakness (odds ratio = 0.78, 95 % confidence interval = 0.71 to 0.86) remaining significant in adjusted models. Older age and quadriceps weakness predicted esKOA. esKOA is also characterized by poor health-related outcomes. This definition of esKOA could be a new clinically relevant outcome measure for osteoarthritis research.