Oral sumatriptan is an effective acute treatment for migraine in adults, but its efficacy in children is still undetermined. Twenty-three children, aged 8.3 to 16.4 years, took both sumatriptan and placebo in a randomized, double-blind, placebo-controlled, crossover trial. The primary endpoint was a greater than or equal to 50% decrease in pain intensity on a 100-mm visual analogue scale at 2 hours. Other endpoints of efficacy were pain intensity difference (PID), showing pain relief at each time point; summed pain intensity differences (SPIDs), estimating overall pain relief; and preference. Two hours after sumatriptan, 7 of 23 reached the primary endpoint, and after placebo, 5 of 23 (difference 9%, 95% CI for difference, -21 to 38%; p = ns). Within 2 hours, the headache disappeared completely in 5 of 23 children after sumatriptan and in 2 of 23 children after placebo (p = ns). Median PIDs were slightly better for sumatriptan between 0.5 and 4 hours (p = ns). Median SPIDs increased almost identically up to 2 hours. Thereafter, median SPIDs for placebo remained practically constant, whereas for sumatriptan, the improvement continued. At 4 hours, the median SPID for sumatriptan was 2.4 times as high as for placebo. However, the maximum differences between median SPIDs at 4 hours (38.5, 95% CI, -75.8 to 57.5; Wilcoxon signed rank test, p = 0.4) or at any other point were not statistically significant. Of the 23 children, 13 preferred sumatriptan and 2 placebo (sign test, p = 0.004). The failure of this and previous controlled studies suggests that the response of children to sumatriptan may be different from adults.