Generalized osteoarthritis in women: pattern of joint involvement and approaches to definition for epidemiological studies.

OBJECTIVE: To ascertain whether clustering between joint sites in osteoarthritis (OA) is more common than would be expected simply from the rising prevalence of the disorder with age, and to explore a definition of generalized OA (GOA) by determining the pattern of joint group involvement, in a population sample of peri and postmenopausal women. METHODS: Radiographs of the hands, knees, and hips were obtained in a population sample of 702 women aged 45 to 64 years. Distal interphalangeal, proximal interphalangeal, carpometacarpal, knee, and hip OA were assessed using the Kellgren-Lawrence grading system. Logistic regression was used to test for overall clustering of OA between joint sites, and log linear models were used to study the patterns of association between different sites. RESULTS: Multiple involvement of the 5 joint groups studied occurred significantly more frequently than could be expected by chance alone (chi 2 = 52.3, df = 5, p < 0.001), and this clustering remained significant after age adjustment (chi 2 = 26.1, df = 5, p < 0.001). Thresholds could be defined for the number of involved joint groups that distinguished a polyarticular subset of OA. These thresholds varied with age and the radiographic cutoff at which OA was assigned. Thus, for grade 2+ disease, GOA could be defined by involvement of 2 or more joint groups at age 45-47 years, but required involvement of all 5 joint groups at age 60-64 years. Symmetry within joint groups was the most pronounced feature in the pattern of joint involvement in the sample as a whole, with associations between different joint groups being substantially weaker than those for symmetrical bilateral involvement of a particular joint. CONCLUSION: There is a clear tendency towards polyarticular OA among women aged 45-64 years. However, there is no single threshold number of joint sites that can be used to define GOA. The pattern of joint involvement in OA is primarily symmetrical, and this pattern strongly suggests a systemic etiology in this subset of postmenopausal women.