OBJECTIVE To discern the frequency of torsades de pointes and QT prolongation in patients receiving intravenous erythromycin lactobionate; to examine the degree of QT prolongation and QT dispersion due to intravenous erythromycin in a typical clinical setting; and to identify any concurrent factors that might predispose patients to excessive QT prolongation or torsades de pointes while receiving intravenous erythromycin. METHODS Retrospective cohort trial. METHODS A university teaching hospital. METHODS All inpatients who received intravenous erythromycin lactobionate during a 1-year period. RESULTS The records of 278 consecutive patients were analyzed, of whom 49 had 12-lead electrocardiograms while receiving and not receiving erythromycin. The dosages of erythromycin ranged from 18-83 (42 +/- 18) mg/kg/day. Of the 49 patients, the baseline QTc was 432 +/- 39 msec, compared with 483 +/- 62 msec during erythromycin therapy (p < 0.01). In 30 of 49 patients with heart disease, the increase in QTc due to erythromycin was 15 +/- 11%, compared with 8.6 +/- 10% in the 19 patients without heart disease (p < 0.05). The degree of QTc dispersion was 34 +/- 16 msec at baseline, compared with 80 +/- 35 msec with erythromycin (p < 0.01). Overall, 19 (39%) of 49 patients had a moderate to severe delay in ventricular repolarization (QTc > or = 500 msec). Of the 278 patients prescribed intravenous erythromycin over the year, it caused torsades de pointes in just one (< or = 0.4%). CONCLUSIONS Erythromycin lactobionate-induced torsades de pointes is rare, although QT prolongation is common. Some patients may be at risk for suffering torsades de pointes due to this agent, particularly if heart disease or other factors that may further delay ventricular repolarization are present.