The efficacy of moclobemide (378 mg +/- 76 mg/day) and fluoxetine (36 mg +/- 8 mg/day) in preventing relapse was studied during 12 weeks of continuation treatment after a 6-week initial trial. Fifty-nine patients with Hamilton Depression Rating Scale (HDRS) scores of 16 or less were enrolled; 29 continued to receive moclobemide and 30 fluoxetine. Efficacy was measured using a 17-item HDRS, the Montgomery-Asberg Depression Rating Scale and the Clinical Global Impression. Improvement in quality of life was measured using a Medical Outcome Study Short-form General Health Survey and the 15D Measure of Quality of Life. Twenty-three per cent of the patients in the fluoxetine group dropped out of the study and 10% in the moclobemide group. Two patients (7%) in the moclobemide group and one (3%) in the fluoxetine group suffered a relapse. Health status and quality of life improved in both drug groups during a 12-week continuation period. The reports of adverse events fell to one third during the continuation phase. The results indicate that benefits may be gained from extending acute treatment.