From the beginning of 1987 to the end of 1989, 72 rheumatoid arthritis patients (RA) whose disease could not be controlled by a single disease modifying antirheumatic drug (DMARD) were selected for the trial treatment. They continued the DMARD treatment used initially at its regular dose, and then started another DMARD regimen at 1/3 to 1/2 of the regular dose as an additive DMARD treatment, which we have designated as Additive Two DMARD Therapy (ATDT). The patients were followed until the end of 1992. In the 3 months of ATDT, the effectiveness of ATDT was obtained in 42 (58.3%) patients who showed more than a 30% decrease in the initial Lansbury's activity index (AI). The rate of side effects at 3 months were 5.6%. Tiopronin, bucillamine or salazopirine added to gold sodium thiomalate or tiopronin were suggested as the recommended DMARD combinations for ATDT. The suppressive effects on AI, ESR, CRP and rheumatoid factor continued for as long as 18 to 24 months. The mean period of ATDT was 21.7 months and that at which ATDT proved useful was 31.9 months. A discontinuation of the first DMARD treatment without any following disease aggravation was obtained in 10 of 15 patients whose disease activity had been sufficiently suppressed for longer than a year. In conclusion, ATDT was suggested to be a useful way of treating RA patients whose disease activity could not be controlled by a single DMARD treatment, as well as a way of evaluating the next DMARD while the ongoing DMARD was observed to gradually lose its initial drug effect.