Sodium valproate enteric-coated tablets were administered as monotherapy to 118 patients (median age, 19 years) with primary generalized epilepsies. More than half (56%) of these patients were transferred from prior drug therapy, most of them because of inadequate seizure control, and some because of adverse effects. Seventy-one percent of the patients experienced tonic-clonic seizures, either alone or in combination with other types of seizures, principally absences. Mean duration of follow-up was 18 months (median, 17 months; range, 1-68 months). At a mean daily dosage of less than 20 mg/kg, 83% of the patients became seizure-free. Therapy was equally effective against tonic-clonic seizures, absences, and myoclonic seizures. Tonic-clonic seizures were suppressed in 85% of cases (89% when patient had only one seizure type), absences in 82% (95% when patient had only one seizure type), and myoclonic seizures in 82%. Paroxysmal activity was present in 88% of the electroencephalogram (EEG) records before valproate monotherapy, and in 32.4% at the study's end. These results were achieved with generally mild and mostly transient side effects; side effects were reported by 16% of patients during the first month, and 2% at the last follow-up. No hematologic or hepatic toxicity was observed. The lag time between attaining steady-state serum concentrations and achieving maximal clinical improvement suggests that sodium valproate monotherapy should be given an adequate trial to ensure that patients derive the greatest possible benefit before adding or switching to another drug.