The aim of this study was to evaluate the efficacy and tolerance of intermittent oral administration of diazepam during hyperthermia for reducing the recurrence of febrile seizure: 185 children, between 8 months and 3 years of age, with a first febrile seizure and normal neurologic development, were randomly assigned in a double-blind fashion to receive orally administered diazepam (0.5 mg/kg, then 0.20 mg/kg, every 12 hours) or placebo, whenever the rectal temperature was more than 38 degrees C. The main criterion of efficacy was the seizure recurrence rate 1 year after the first seizure. The duration of the study was 3 years; eight different centers in France participated. There were 462 febrile episodes and 1000 days with prophylactic treatment. The recurrence rates did not differ between the diazepam group (16%) and the placebo (19.5%) group. The children with recurrent seizures were significantly younger at the time of the first seizure (17 +/- 6.9 months) than children without a recurrent seizure (21 +/- 8.5 months). In children with recurrent seizures, prophylactic treatment was correctly administered to only 1 of 15 children in the diazepam group and to 7 of 18 children in the placebo group. The following were the reasons for this poor cooperation: convulsion being the first manifestation of the fever (seven cases in each group), parents neglecting to give treatment (nine cases), and refusal to take treatment by two children. Side effects were similar in the two groups except for hyperactivity, which was more frequent in the diazepam (138 days) than in the placebo (34 days) group. Intermittent oral administration of diazepam at the onset of fever offered no advantage over placebo in preventing recurrence of seizure. This finding probably reflects a lack of efficacy of the intermittent method rather than of diazepam itself.