OBJECTIVE

To determine whether baseline or serial plasma concentrations of stromelysin (matrix metalloproteinase 3 [MMP-3]) protein might distinguish subjects with progressive radiographic knee osteoarthritis (OA) from those with stable disease.

METHODS

Subjects were 120 women with unilateral knee OA who participated in a 30-month randomized, placebo-controlled trial of structure modification with doxycycline. Anteroposterior views of both knees in a semiflexed position were obtained at baseline, 16 months, and 30 months. Subjects were selected to obtain comparisons of plasma MMP-3 levels between 60 progressors (21 taking doxycycline, 39 taking placebo) and 60 nonprogressors (30 taking doxycycline, 30 taking placebo) with respect to medial joint space narrowing (JSN) in the index knee. Each group consisted of 30 subjects who exhibited significant increases in knee pain. Blood samples were obtained semiannually for MMP-3 assay.

RESULTS

Subjects in the placebo group whose MMP-3 concentration was in the upper tertile of the baseline distribution showed a 4-fold increase in the odds of progression of JSN as compared with the lower tertile (odds ratio 4.12, P = 0.037). Baseline MMP-3 levels were unrelated to knee pain. The within-subject mean of serial MMP-3 concentrations was associated with concurrent JSN in the placebo group over the 0-16-month interval (b = 0.18 mm/SD increase in the mean MMP-3, P < 0.01) and over the 16-30-month interval (b = 0.15, P < 0.05). Similar evidence of concurrent validity was found in the placebo group for the maximum of intercurrent MMP-3 values.

CONCLUSION

The baseline MMP-3 level was a significant predictor of JSN in this pilot study. Moreover, serial plasma MMP-3 levels reflected concurrent JSN in the placebo group over the 30-month period of observation.