A randomized, crossover study was conducted in healthy male volunteers to assess the diurnal variation in the steady-state pharmacokinetics of valproate after multiple 250-mg oral and intravenous infusion doses after an intravenous 750-mg loading dose. Multiple blood samples were collected throughout each 168-hour study period, and plasma valproate concentrations were quantitated using a gas chromatographic technique. Within-regimen comparisons indicated statistically significant differences for mean steady-state peak (Cmax) and trough (Cmin) plasma concentrations and area under the plasma concentration-versus-time curve (AUC) between the second and third doses on day 4 after oral dosing, indicating a diurnal variation in the rate of valproate absorption from the delayed-release tablet preparation. Between-regimen steady-state comparisons of pharmacokinetic parameters revealed some significant differences in mean time to Cmax (Tmax), Cmax, Cmin, AUC, and total body clearance for respective day-4 dosing intervals, but not for the entire day 4, except for Cmin and Tmax. Mean elimination rate constant and half-life did not significantly differ between the regimens. The regimens were bio-equivalent at steady state, as assessed by 90% confidence intervals (two one-sided test procedures) for Cmax, Cmin, and AUC, with similarity in degrees of fluctuation ((Cmax-Cmin)/Caverage). Despite the presence of diurnal variation in valproate absorption after oral dose administration, the steady-state plasma concentration-versus-time profile was well maintained by both regimens within the accepted therapeutic range of 50 to 100 micrograms/mL.