Migraine and cluster headache are the most common disabling primary headache syndromes and are typically episodic. A reliable method of triggering such headache attacks facilitates the study and treatment of these disorders. There is sufficient clinical and laboratory evidence to suggest that somatostatin withdrawal may be a useful way of triggering headache. We studied 15 subjects, eight migraineurs, four cluster headache sufferers and three healthy controls. Each subject had a standard somatostatin infusion, 250 microg/h for 3.5h. Subjects were followed for 24h post-infusion. Growth hormone was suppressed in each subject demonstrating a biologically active infusion of somatostatin. None of the non-headache sufferers had pain. Seven of eight migraine sufferers had no immediate headache and no delayed headache. One migraineur experienced short lasting headache with no migrainous features. Three of four patients with cluster headache had no significant pain with the infusion, while one had pain after 1h. The results suggest that somatostatin infusion is not a reliable way to produce headache in experimental settings in either migraine or cluster headache. The data do not exclude a role for somatostatinergic mechanisms in primary headache.