OBJECTIVE
The hand is the major site of joint involvement in osteoarthritis (OA), but the clustering of this condition at different joint sites within the hand as part of an entity known as generalized OA remains contentious. We examine this issue in a population sample of 967 peri and postmenopausal women.
METHODS
The pattern of radiographic involvement of hand joints was examined using the Kellgren and Lawrence system in a sample of women aged 45-64 years, selected from the general population in Chingford, East London. Log linear modelling techniques were used to examine clustering of OA in different joint sites and to examine whether this clustering remained significant after age adjustment.
RESULTS
There was clear evidence of clustering in joint involvement: Thus 20 women had 4 or more affected joints compared with only 2 expected (chi 2 = 72.0, 4df, p < 0.001). The risk of multiple distal interphalangeal (DIP) joint involvement, after age adjustment, in a woman with a single affected DIP joint was significantly increased (OR 10.0, 95% CI 7.3-13.7) and was substantially greater than the corresponding risk for proximal interphalangeal joint (OR 3.1, 95% CI 1.4-6.8) involvement. Symmetry was the strongest determinant of pattern with an OR 38.8 (95% CI 14.5-103.5) for corresponding DIP joint involvement in the contralateral hand.
CONCLUSION
These data provide clear evidence for a polyarticular subset of hand OA in women. There are 3 major determinants of the pattern of polyarticular involvement; symmetry, clustering by row and clustering by ray, in descending order of importance.