OBJECTIVE: The authors determined the effects of antipanic treatment with fluoxetine on human growth hormone (GH) response to the alpha 2 agonist clonidine. METHOD: Seventeen patients with panic disorder and 15 healthy volunteers were challenged with clonidine. Thirteen of the patients and 12 of the volunteers were given a second challenge with clonidine 12 weeks later. The patients received open fluoxetine and the healthy subjects received no treatment between challenges. Subjects with high baseline human GH levels (greater than 2 ng/ml) at the first and second challenges were excluded from further analysis. RESULTS: The patients with panic disorder (N = 13 for the first challenge and N = 9 for the second) had significantly lower human GH responses to clonidine than the healthy subjects (N = 14 during the first challenge and N = 9 for the second) during both challenges, despite clinical improvement in eight of the nine patients at the time of the second challenge. CONCLUSIONS: Blunted secretion of human GH in response to clonidine in patients with panic disorder persists despite clinical recovery.