The aim of this study was 1) to assess the incidence of electrocardiographic changes after treatment with halofantrine and 2) to study the relationship between these changes and plasma levels of halofantrine and its main metabolite, N-desbutyl-halofantrine. Thirty-four male patients with uncomplicated falciparum malaria were enrolled in this study. Halofantrine was administered on two separate days at a total oral dosage of 24 mg/kg/day in three doses over a 12-hr period. The interval between the two treatments was seven days. Twelve-lead electrocardiography (ECG) was performed to measure the QT interval (QTc), ambulatory ECG monitoring was done to detect ventricular arrhythmia, signal-averaged ECG was performed to detect late ventricular potentials, and blood tests were performed to determine plasma concentrations of halofantrine and N-desbutyl-halofantrine. Maximum QTc was observed at 12 hr after both the first (P < 0.0002) and second treatments (P < 0.03). Signal-averaged ECG revealed late potentials in four cases (72 hr after the first treatment in one case and at 24 hr after the second treatment in three cases). Ventricular arrhythmia was not observed. Significantly higher plasma concentrations of halofantrine were observed 2 hr after the second treatment. At this time, both the time effect and time interaction were significant (P < 0.008 and P < 0.02, respectively). The QTc interval was significantly correlated with the plasma halofantrine level (r = 0.41, P < 0.01) but not with the plasma N-desbutyl-halofantrine level (r = 0.30, not significant). In three cases, late ventricular potentials were associated with a maximum concentration of halofantrine. Our findings indicate that electrocardiographic changes are dose-dependent and that a second treatment at the same dosage may be hazardous.