A 12-week single-blind study used valproic acid in the treatment of refractory, simple and complex absence (petit mal) seizures in 17 children. The clinical seizure control was compared with the number and duration of electrical discharges on six-hour EEG recordings before and during therapy. Sixteen patients showed clinical improvement. Fourteen had greater than 75% clinical improvement and ten patients became free of absence seizures. Side effects were mild, but two patients experienced transient, asymptomatic thrombocytopenia. Good correlation between clinical response and decreasing seizure activity on the EEG occurred in the seizure-free patients.