The activity of urinary N-acetyl-beta-glucosaminidase (NAG) was measured in random urines using the ratio (NAG index) of NAG to grams of urine creatinine in 102 epileptic children taking antiepileptic drugs (AEDs). A high urinary NAG index (2 SD or more above the mean for the age-matched control/normal subjects) was detected in 40 (39%, 40/102) epileptic children with AEDs. None of the 40 epileptic children with abnormal urinary NAG excretion had significant proteinuria. Among the 83 epileptic children under monotherapy, 29 cases (35%) had elevated urinary NAG excretion. Valproic acid presented the highest incidence of abnormal urinary NAG index (78%, 7/9 cases) within the monotherapy group, and the incidence was statistically significantly higher than that (26%, 14/55) in the carbamazepine group (p < 0.005). In the monotherapy group, no significant difference in serum levels of AEDs was found between children with normal urinary NAG excretion and those with abnormal. Nineteen epileptic children were treated with more than one AEDs. Eleven of them (58%, 11/19 cases) had a high urinary NAG index. The incidence of high urinary NAG index in the polytherapy group and that in monotherapy group was not significantly different (p > 0.05). This study suggests that AEDs may be potentially nephrotoxic and that urinary NAG may play a role in screening renal tubular injury in epileptic children under therapy of AEDs. Further investigation should be conducted to determine whether the effect of AEDs on renal tubular cells is or is not reversible.