Matrix metalloproteinase (MMP)-1 and MMP-3 levels were measured in serum samples from rheumatoid arthritis (RA) patients undergoing a double-blinded placebo-controlled trial with the chimaeric anti-tumour necrosis factor (TNF)-alpha antibody cA2. Both MMP-1 (P < 0.015), but to a larger extent MMP-3 (P < 0.001) levels were elevated in all RA patients prior to the commencement of the trial compared with normal control sera. Following cA2 therapy, MMP-1 and MMP-3 levels were assessed in the placebo, and 1 and 10 mg/kg cA2-treated groups at 7, 14, 21 and 28 days. In both the 1 and the 10 mg/kg cA2-treated groups, a significant decrease in serum MMP-3 levels at all time points was observed, reducing maximally to 41% of pre-infusion values at day 7. MMP-1 levels were also reduced, but less dramatically than MMP-3, to 85% of pre-infusion values after 14 days in the 10 mg/kg cA2 treated group. In a separate non-placebo-controlled study, we also evaluated the tissue inhibitor of metalloproteinase (TIMP)-1 levels in plasma following cA2 infusion. Pre-infusion TIMP-1 levels were above the normal control range, but were significantly reduced (P < 0.035) 14 days after infusion to 72% of pre-infusion values. This study confirms previous reports that MMP-3 levels are elevated and correlate with measures of inflammation in RA, and furthermore demonstrate that serum MMP-3 and MMP-1 levels are downmodulated following anti-TNF-alpha antibody therapy. Whilst serum MMP-3 levels correlated with C-reactive protein (CRP) both prior to and following anti-TNF-alpha antibody therapy, it remains to be demonstrated that serum MMP-3 and/or MMP-1 levels reflect the cartilage and bone resorptive processes which are evident in this disease.