Current views on cluster headache pathogenesis indicate a primary central nervous system dysfunction, in particular a hypothalamic involvement. To confirm the hypothalamic involvement in cluster headache we evaluated the hypothalamic-pituitary axis responsiveness with the thyrotrophin releasing hormone (TRH) test. A dose of 200 micrograms of TRH was administered i.v. to nine healthy controls, 32 patients with cluster headache during cluster period and 16 in remission period. Delta maximum thyrotrophin (TSH) was significantly lower in patients with cluster headache during cluster period (p less than 0.05 versus healthy controls and cluster headache patients in remission). No difference was observed between healthy controls and cluster headache patients in remission. A monoaminergic dysfunction at the hypothalamic level is hypothesized.