A number of laboratory variables, including Hb., ESR and several phase proteins, fluctuated in concord with the clinical signs of synovitis activity in patients with rheumatoid arthritis during a controlled study of 3 disease-modifying anti-rheumatic drugs (DMARD). The correlation between laboratory variables and clinical synovitis was significant in a large patient population but the correlation coefficients were not of such magnitude that any of the laboratory variables reflected clinical synovitis activity in a reliable manner in the individual patients. In patients treated with azathioprine, the response of the Hb, (and consequently of the ESR), was reduced compared to patients given other DMARD. This phenomenon was caused by the bone marrow suppressing effect of azathioprine. However, the effect of azathioprine on the clinical synovitis activity did not differ from that of the 2 other drugs. Similar results were found by reviewing the literature about controlled trials of DMARD. In the present trial the clinical evaluation was performed under optimal conditions. In daily clinical practice the evaluations of the joints may be less than optimal since they may be performed by different rheumatologists with varying experience. Consequently, it may be difficult to do without the unreliable laboratory variables mentioned in the routine assessments of disease activity, unless the quality of routine evaluations of synovitis activity is improved considerably.