The purpose of this study was to investigate the relationships between depressive subtypes and response to fluoxetine treatment in a large cohort of outpatients. We studied 294 outpatients with major depressive disorder who were then treated with fluoxetine 20 mg/day for 8 weeks. Treatment outcome was evaluated with the Hamilton Depression Rating Scale (HDRS)-17, the Clinical Global Impressions-Severity, and with the HDRS-8; the latter is proposed to be a relatively more specific measure of depression severity than the HDRS-17. We assessed the relationships between degree of treatment response and several depressive subtypes (melancholic, atypical, hostile, and anxious depression, double depression, and depression with comorbid personality disorders), after adjusting for baseline depression severity. We found that nonanxious depressives (patients without any comorbid anxiety disorder) improved slightly but significantly more during treatment than anxious depressives on all outcome measures. Melancholic depression was associated with slightly less improvement on the HDRS-17 only, whereas the other subtypes of depression were not associated with differences in treatment outcome.