A 10-year cohort study was performed, involving all of the 118 patients treated with bucillamine in our hospital between 1988 and 1990. Evaluation was made on the basis of erythrocyte sedimentation rate, grip strength, joint score, duration of morning stiffness, and Lansbury index consisting of the above four parameters. Eleven patients were male and 107 were female, with a mean age of 53 years (range: 20-79 years) and a mean duration of illness of 8.2 years (range: 2-31 years). Lansbury index remained significantly suppressed throughout the 10-year period of treatment. Continuous treatment was possible for 10 years in 18 patients (15%: 2 men and 16 women). Stage of disease did not advance in 14 patients. Six patients met the criteria for remission. Of all patients, 50% dropped out of treatment at 2.4 years after the start of treatment and 75% at 5 years. The 100 patients who dropped out could be roughly divided into three groups. One third of them dropped out because of lack of or attenuation of response. Another third dropped out because of referral to other medical facilities or discontinuation of visits to our hospital, and the remaining third dropped out because of adverse reactions to treatment. There was no particular trend in terms of sex, age, duration of sickness, drugs used before bucillamine, or level of activity of rheumatoid arthritis. There were no significant difference in the stage and class of the disease, and other backgrounds between 10-year treatment group and dropout group.