The pharmacokinetic profile of D-3-acetoxy-cis-2,3-dihydro-5-(2-dimethylamino-ethyl)-2-(p-methoxy-phenyl)-1, 5-benzothiazepin-4(5H)-one hydrochloride (diltiazem . HCl) following i.v. and p.o. administration has been studied in six healthy subjects using a new sensitive GLC method. The volunteers received an i.v. infusion of 20 mg in 20 min, a peroral solution of 120 mg and two 60-mg tablets (Dilzem) in a randomized sequence. The plasma level time courses of the unchanged compound following infusion and peroral solution were simultaneously evaluated by nonlinear regression analysis. The best model was chosen by means of statistical criteria. In all subjects the experimental data could be adequately described by an open three-compartment model with zero order input following infusion and first order absorption following p.o. solution. Using this procedure the following common disposition parameters were obtained for both routes of administration: t1/2 alpha = 0.1 h, t1/2 beta = 2.1 h, t1/2 gamma = 9.8 h (harmonic means derived from individualized fits), Vc = 0.9 +/- 0.4 l/kg, Vss = 5.2 +/- 2.4 l/kg, Cltot = 11.5 +/- 1.8 ml/min/kg. Compared to the beta-phase the terminal gamma phase represents a smaller contribution to the total AUC. The blood/plasma distribution ratio was found to be 1.00 +/- 0.08 (N = 4), the mean hepatic extraction ratio was 0.54. Peak levels appeared 0.6 +/- 0.3 h after the p.o. solution. The mean absolute bioavailability of diltazem based on the individual AUC infinity 0 of the p.o. solution and the infusion was 0.44 +/- 0.10. Following tablet administration, delayed maximum levels were found after 2.8 +/- 0.9 h. Comparing the AUC infinity 0 of tablets and p.o. solution, there was no significant difference between both dosage forms.