BACKGROUND: We assessed the tolerability of and response to citalopram in depressed patients who had discontinued fluoxetine treatment due to adverse events. METHOD: Fifty-five outpatients with DSM-IV major depressive disorder and a confirmed history of intolerance to fluoxetine (mean final dose = 24.6 mg/day) were switched to citalopram (20 mg/day) after a 2- to 4-week single-blind placebo washout period. During a 6-week, open-label treatment protocol, citalopram could be titrated up to 40 mg/day. Safety and tolerability, including reemergence of symptoms that previously had been associated with fluoxetine, were assessed by recording all spontaneously reported or observed adverse events. Efficacy was evaluated using the Hamilton Rating Scale for Depression (HAM-D), the Clinical Global Impressions (CGI) scale, and several other measures. Response was defined as a CGI-Improvement score at endpoint of 1 or 2 (i.e., very much or much improved). RESULTS: Ninety-five percent of patients (N = 52) completed the citalopram trial. The only adverse events reported by more than 5 patients (>or= 10%) were pharyngitis (15%) and constipation (11%), and none of the 3 early terminations were attributed to adverse events. The rate of recurrence of the fluoxetine-associated adverse events was low, with headache (3 [27%] of 11 cases), nausea (2 [22%] of 9 cases), and decreased libido (5 [18%] of 28 cases) being the most common. Significant improvement from baseline HAM-D (p <.001) was observed by the first week of citalopram therapy and continued until study end. The intent-to-treat CGI response rate was 65% (36 of 55 patients) at study endpoint; 69% (36 of 52 patients) of the completers responded. CONCLUSION: These data suggest that fluoxetine-intolerant patients can be treated effectively with citalopram.