Twenty-seven patients with combat-related posttraumatic stress disorder (PTSD) entered an open, prospective, 10-week trial of fluoxetine, beginning with 20 mg/day and increasing to 80 mg/ day until response was optimal or side effects prohibited dose increase. Nineteen patients completed 3 or more weeks and were included in the data analysis. Total Clinician-Administered PTSD Scale scores decreased from a mean of 64.5 at baseline to 42.7 at endpoint (F=7.17p< 0.001), and improvement was significant in each of the three PTSD subscales (reexperiencing, avoidance/ numbing, and hyperarousal). Depression and anxiety ratings showed similar improvements, and suicidally ratings did not increase. Global improvement scores decreased from 4.0 at baseline to 2.67 at endpoint (F= 12.08,p< 0.001); however, improvement in social and occupational functioning was minimal. Appreciable improvement tended to occur after 6 weeks, suggesting that higher fluoxetine doses and/or duration than that used for depression may be indicated in this population. Panic attack frequency decreased by at least 50% in six of eight patients who kept panic diaries. The high dropout rate reflects problems with side effects, anxiety symptoms, external events, and substance abuse. Our data suggest that fluoxetine is effective in reducing reexperiencing, avoidance, and hyperarousal symptoms of PTSD, and this improvement is independent of comorbid panic disorder. In addition, fluoxetine appears to be effective in reducing panic attacks in PTSD patients. The efficacy of fluoxetine for some PTSD patients is interesting in light of emerging neuropharmaco-logic data suggesting serotonergic dysregulation in some PTSD patients. Noradrenergic hypotheses are also discussed. The findings should be confirmed by double-blind, placebo-controlled studies.