To assess the relative contributions of encainide and its putatively active metabolites, O-demethyl encainide (ODE) and 3 methoxy-O-demethyl encainide (3MODE), to the drug's pharmacologic effects, we compared intravenous infusions and sustained oral therapy in two phenotypically distinct groups of patients, extensive and poor metabolizers of encainide. Unlike poor metabolizers, extensive metabolizers had appreciable quantities of both metabolites detectable in plasma and had fourfold shorter elimination half-lives for encainide. By quantitating electrocardiogram intervals, arrhythmia frequency, and plasma concentrations, we found that, in poor metabolizers, arrhythmia suppression and ventricular complex (QRS) prolongation were correlated positively with encainide concentrations (r greater than or equal to 0.570, P less than 0.014). In these two subjects, antiarrhythmic concentrations of encainide (greater than 265 ng/ml) were at least fivefold higher than those sustained in the six extensive metabolizers during steady state oral therapy. In extensive metabolizers, encainide concentrations were uncorrelated with effects. Arrhythmia suppression and QRS prolongation in extensive metabolizers correlated best with ODE (r greater than or equal to 0.816, P less than 0.001); QTc change correlated positively with both 3MODE and ODE. Arrhythmia suppression paralleled QRS prolongation; the relationship between them appeared similar in both phenotypic groups. In most patients, extensive metabolizers, encainide effects during oral therapy are mediated by metabolites, probably ODE.