The present study represents an open-label clinical trial comparing treatment with a benzodiazepine (lorazepam) to divalproex in 11 inpatients with uncomplicated alcohol withdrawal syndrome. The trial used the Clinical Institute Withdrawal Assessment for Alcohol-Revised (CIWA-Ar) scale. There were no significant differences in demographics or substance use parameters between the divalproex group (n = 6) or the lorazepam group (n = 5). A significant Group x CIWA-Ar score interaction [F(8,72) = 2.57, p < or = .01] was confirmed and further substantiated by a quadratic trend component for the interaction [F(1,9) = 24.9, p < or = .001]. This preliminary study supports further investigation of divalproex in the treatment of alcohol withdrawal.