Quinidine pharmacokinetic behaviour was evaluated in 139 adult hospitalised men receiving oral quinidine therapy. A total of 391 serum quinidine concentrations were measured by enzyme immunoassay for routine clinical purposes. The NONMEM programme was used to examine the relationship between quinidine pharmacokinetics and several potential covariates. A 1-compartment open model with first-order absorption and elimination was assumed. The mean apparent volume of distribution (Vd) was about 230L. When measured,&agr;1-acid glycoprotein (AAG) levels were not included in the analysis. Oral quinidine clearance (CL) decreased with age, severe congestive heart failure and renal disease, and increased in patients with a history of alcohol abuse. The interpatient variability in CL and the intrapatient residual variability expressed as coefficients of variation (CV) were 28 and 31%, respectively. When AAG values were incorporated into the analysis, the only important covariates of CL were the AAG measurements and the presence of renal dysfunction as indicated by a calculated creatinine clearance of less than 50 ml/min (3 L/h). The interpatient variability in CL and the residual intrapatient CVs decreased to approximately 24 and 26%, respectively. Improvement of the CL model by inclusion of measured AAG strongly suggests that quinidine elimination is dependent on the free concentration of drug in plasma and supports the use of free serum quinidine concentrations when evaluating and monitoring quinidine therapy.