The pharmacokinetics of phenytoin is complicated by the nonlinearity of the dose-concentration relationship which is a consequence of capacity-limited metabolism. Individualized therapy with phenytoin is therefore optimally required. As no data are available on the population pharmacokinetics of phenytoin in the Western Cape, this study was undertaken to address this issue. This study was conducted prospectively primarily to: (1) investigate the influence of various patient variables on the population pharmacokinetic parameters of phenytoin, (2) assess whether the parallel Michaelis-Menten and first-order elimination model provides a better fit to the data than the Michaelis-Menten model, (3) determine population pharmacokinetic parameter estimates of phenytoin representative of the patient population, and (4) validate and compare the clinical applicability of the parameter estimates and the models. The study population comprised 332 black and coloured, adult, male and female epileptic patients residing in the Western Cape, South Africa. All patients were on phenytoin monotherapy for the management of their epilepsy and no drugs known to interfere with phenytoin pharmacokinetics were taken concurrently. Clinical pharmacokinetic dosing services were initiated at 9 clinics from which patients were selected for this study. The service entailed a patient interview, a chart review, drug analysis and provision of either a written or verbal consultation report. The data were analyzed using NONMEM (nonlinear mixed-effects modelling), a computer programme designed for population pharmacokinetic analysis that allows pooling of data from many individuals. The Michaelis-Menten and the parallel Michaelis-Menten and first-order elimination models were fitted to 853 steady-state dose: serum concentration pairs.