Few controlled studies have evaluated the long-term continuation of pharmacotherapy for relapse prevention in patients with obsessive-compulsive disorder (OCD). This study assessed efficacy and safety of fluoxetine versus placebo in preventing relapse of OCD during a 52-week period in responders to short-term administration of fluoxetine. Patients who met DSM-IV criteria for OCD and had a Yale-Brown Obsessive Compulsive Scale score > or = 19 were treated with single-blind fluoxetine 20, 40, or 60 mg/day (based on physician assessment of response and tolerability). After 20 weeks, responders were randomly assigned to receive continued treatment with fluoxetine or placebo and were monitored for relapse for up to 52 weeks. Of 130 patients who entered the study, 71 (55%) were randomly assigned to receive fluoxetine (N = 36) or placebo (N = 35). Patients who received fluoxetine had numerically lower relapse rates compared with those who received placebo, although the difference was not significant (Kaplan-Meier 1-year relapse rates: fluoxetine, 20.6%; placebo, 31.9%; one-tailed p value = 0.137). In additional analyses evaluating patients on the basis of fluoxetine dose at randomization, patients who continued treatment with fluoxetine 60 mg/day (N = 52) had significantly lower rates of relapse than those who were switched to placebo (Kaplan-Meier 1-year relapse rates: fluoxetine, 17.5%; placebo, 38.0%; one-tailed p value = 0.041). Those who responded to the acute treatment phase with 40 (N = 18) or 20 (N = 1) mg/day had low overall rates of relapse, and the difference between continued fluoxetine and placebo treatment for these patients was not significant. For responders to the 60 mg/day dosage, those patients who continued treatment with fluoxetine were provided greater protection against relapse than those patients switched to placebo.