Oral steroids can interrupt bouts of cluster headache (CH) attacks, but recurrence is frequent and may lead to steroid-dependency. Suboccipital steroid injection may be an effective 'single shot' alternative, but no placebo-controlled trial is available. The aim of our study was to assess in a double-blind placebo-controlled trial the preventative effect on CH attacks of an ipsilateral steroid injection in the region of the greater occipital nerve. Sixteen episodic (ECH) and seven chronic (CCH) CH outpatients were included. ECH patients were in a new bout since no more than 1 week. After a one-week run-in period, patients were allocated by randomization to the placebo or verum arms and received on the side of attacks a suboccipital injection of a mixture of long- and rapid-acting betamethasone (n=13; Verum-group) or physiological saline (n=10; Plac-group). Acute treatment was allowed at any time, additional preventative therapy if attacks persisted after 1 week. Three investigators performed the injections, while four others, blinded to group allocation, followed the patients. Follow-up visits were after 1 and 4 weeks, whereafter patients were followed routinely. Eleven Verum-group patients (3 CCH) (85%) became attack-free in the first week after the injection compared to none in the Plac-group (P=0.0001). Among them eight remained attack-free for 4 weeks (P=0.0026). Remission lasted between 4 and 26 months in five patients. A single suboccipital steroid injection completely suppresses attacks in more than 80% of CH patients. This effect is maintained for at least 4 weeks in the majority of them.