Abstract
BACKGROUND: Studying a spontaneous migraine attack is challenging, particularly the earliest components. Nitroglycerin is a potent, reliable and reproducible migraine trigger of the entirety of the migraine attack, making its use experimentally attractive. METHODS: Fifty-three subjects with migraine with a history of spontaneous premonitory symptoms were exposed to a 0.5 mcg/kg/min nitroglycerin infusion. Eighty-three percent (n = 44) developed typical premonitory and headache symptomatology. Fifty-seven percent (n = 25) were invited back to further study visits, during which they were re-exposed to nitroglycerin or placebo infusion in a double-blind randomised design. The phenotype of premonitory symptoms and headache was captured and compared to spontaneous attacks and between triggered attacks using agreement analysis. RESULTS: More premonitory symptoms were triggered with nitroglycerin than placebo (mean symptom difference = 4, t(20) = 7.06, p < 0.001). The agreement in triggering for the most commonly reported premonitory symptoms (concentration difficulty and tiredness) was >66%. The retriggering agreement for all but one premonitory symptom was >60%. The agreement in timing to onset of premonitory symptoms was reliable across two triggered attacks. The agreement with spontaneous attacks and between attacks for headache and its associated symptoms, including laterality, was less reliable. CONCLUSIONS: Nitroglycerin can reliably and reproducibly provoke premonitory symptomatology associated with migraine. This forms an ideal model to study the earliest manifestations of migraine attacks.