In a double-blind crossover study of 39 children with established migraine, there was no difference between treatment with propanolol and placebo as regards frequency, severity or average duration of migraine attacks. There was some evidence that propanolol increased the average length of headaches.