Characteristic response patterns are described for two antidepressant drugs and placebo in post-traumatic stress disorder. Early onset and steady improvement occurred on a global rating scale for both drugs and placebo in those who ultimately met responder criteria at the end of treatment. In certain cases, the magnitude of global response was greater for drug than for placebo. At weeks 2 or 4, the Clinical Global Impressions score for fluoxetine but not for amitriptyline, served as a good predictor of eventual response. In a review of numerous completed placebo-controlled trials, antidepressants were superior to placebo in seven out of eight comparisons using the Clinical Global Impressions, although specific effects on post-traumatic stress disorder scales were more variable. Drug response rates are similar for combat and civilian trauma samples, but placebo response rates may be higher in the latter. Effect sizes suggested a moderate-to-good effect for drug therapy.