Sixty outpatients with obsessive-compulsive disorder (OCD, 22 men, 38 women) were randomized to receive 6 months of antiexposure therapy with fluvoxamine (group F), exposure therapy with fluvoxamine (group Fe), or exposure therapy with placebo (group Pe). Patients in group F did not comply with antiexposure therapy, so it was in fact a neutral condition. Patients began with depressed mood (mean Hamilton depression score = 19). Fifty patients were reevaluated at week 8, 44 at week 24 (posttest), 37 at week 48, and 33 at 18 months, 1 year posttreatment (group F, n = 10; group Fe, n = 12; group Pe, n = 11). The three groups improved on rituals and depression. There was a drug effect on rituals at week 8 and on depression at week 24; both these effects disappeared at week 48. The 33 18-month completers had been comparable at baseline to those not followed up, apart from having more severe behavioral avoidance. At 18-month followup, patients as a whole remained improved with no between-group differences; over 80% of the Fe and Pe patients versus 40% of the F patients were not receiving antidepressant treatment (Fe vs. F: p < 0.04; Pe vs. F: p = 0.053; Fe vs. Pe: NS). In OCD fluvoxamine and exposure therapy were synergistic in the short term, and exposure reduced subsequent need for antidepressants in the followup year after they had been stopped.