The relationship between plasma concentrations of cibenzoline and its antiarrhythmic effect was evaluated in patients receiving the drug orally as part of an ascending multiple dose efficacy and tolerance study. Twenty-five patients participated in a 3-day placebo period, 3 days of 32.5 mg cibenzoline every 6 hr, 3 days of 65 mg cibenzoline every 6 hr, 3 days of 81.25 mg cibenzoline every 6 hr, and 3 final placebo days. Arrhythmia frequency was monitored by 24-hr Holter monitoring and blood samples were drawn during and after dosing. Percent reduction in baseline premature ventricular complex (PVC) frequency for the 25 subjects demonstrated considerable interpatient variability in antiarrhythmic response. Cibenzoline plasma concentrations over 300 ng/ml were associated with some decrease in PVC frequency in virtually all cases. The relationship between plasma concentration and PVC frequency was studied more rigorously in eight of the 25 patients and that for ventricular couplet (VC) frequency was studied in six. For these analyses, PVC and VC frequency data were averaged over 6-hr intervals and plotted against trough cibenzoline concentrations. The data from each patient were fitted with a concentration-effect function (Hill equation) by means of least squares regression. With the exception of two extreme values, the concentration corresponding to 90% reduction in PVC frequency (C90) ranged from 215 to 405 ng/ml. In five of the six patients with arrhythmia in whom VC data were also evaluated, the individual C90 for VCs were considerably less than those for PVCs. The agreement between the observed concentration-response relationships and those predicted by curve-fitting the data suggests that the antiarrhythmic effect of cibenzoline is proportional to its plasma concentration, and that the Hill equation provides an accurate mathematic description of the concentration-response relationship.