Parameters describing disposition and absolute oral bioavailability of quinidine were determined in ten normal male volunteers using a specific assay. Various models were compared for their ability to describe the experimental data. An intravenous quinidine gluconate and an oral quinidine sulfate solution were administered (3.74 mg/kg quinidine base). In three subjects the intravenous and oral studies were repeated. One-, two-, and three-compartment models with zero and first-order input were fitted to the plasma concentrations. The selection of the best model was made by the Akaike information criterion and by eye. After intravenous administration, plasma concentration-time curves could be adequately described by a two-compartment model. Mean disposition constants (+/- SD) were obtained from individualized fits (V1: 0.398 +/- 0.336 LITER/KG, Vdarea: 2.53 +/- 0.72 liter/kg, alpha: 0.316 +/- 0.294 min -1, beta: 0.00204 +/- 0.00262 min -1, k2: 0.0305 +/- 0.0101 min -1). A clearance of 4.9 +/- 1.5 ml/min/kg was observed. After oral administration, three-compartment models were needed to describe the observed data in some cases. Absorption was in most cases best described by a zero-order rather than by a first-order process. The time to peak concentration varied from 23 to 121 min, the lag time was always less than 3 min, and the mean elimination rate constant was 0.00171 min -1. The mean oral bioavailability of quinidine was 0.70 +/- 0.17.