This study explored serum concentrations of fluoxetine and norfluoxetine in subjects treated under different dosing regimens. Using a placebo-controlled, randomized, double-blind design, we compared 20 mg fluoxetine daily, 60 mg fluoxetine weekly, and a placebo. All subjects had responded positively to a 7-week open-label trial of 20 mg fluoxetine daily. Consistent with numerous attempts to correlate fluoxetine and norfluoxetine blood levels to clinical response, we found no significant relationships between serum concentrations and Hamilton Rating Scale for Depression (HAM-D) scores. Serum concentrations were significantly different among the three treatment groups. Pairing the fact that the serum concentrations of subjects given 60 mg fluoxetine weekly were significantly lower than those given 20 mg daily with the finding that the 60 mg weekly group had lower HAM-D scores suggests that, for stabilized subjects, this nontraditional dosing strategy is at least as effective in treating depression as the standard 20 mg fluoxetine daily treatment.