It has recently been shown that acute changes of venous blood ammonia (NH3) may predict short-term adverse effects of valproic acid (VPA). In the present study, the time course of NH3 concentration after a single oral dose of VPA (800 mg) was monitored in 68 epileptic patients. Patients were classified into four groups: previously untreated patients (group A, n = 21), patients under treatment with either phenobarbital (group B, n = 14) or phenytoin (group C, n = 13) or both (group D, n = 20). In each patient, venous blood for the NH3 assay was taken before the VPA dose (predose level) and at 1, 2, 3, and 4 h after the dose (postdose levels). While in patients receiving only VPA the postdose NH3 concentrations did not differ from the predose level, in each of groups B, C, and D the postdose concentrations appeared to be significantly higher than the predose concentration. The greatest increase was observed in group D. In the light of the data reported in the literature, those patients whose NH3 concentration after the VPA dose exceeds 100 micrograms/dl should be considered at higher risk for short-term, VPA-induced adverse effects during long-term therapy. Thus, our data suggest that caution should be exercised in adding VPA to anticonvulsant treatments including phenobarbital or phenytoin or both.