To study prospectively the effects of cisapride on ventricular repolarization, depolarization, and arrhythmia markers in neonates, we determined before and three days after starting cisapride (1 mg/kg/day): corrected QT interval (QTc) and QT dispersion (QTd) on standard ECGs, and duration of filtered QRS (fQRS) and of low amplitude (<40 microV) terminal signals (LAS40, ms) and root mean square of the last 40 ms (RMS40, microV) using high-gain signal-averaged ECG (SAECG). Twenty-four term and 11 preterm infants (gestational age 23-35 weeks) were studied at a median chronological age of 32 days. QTc and QTd were not different between term and preterm infants. Cisapride lengthened QTc (mean +/- SD; ms: 396.6 +/- 24.8 before vs. 417.0 +/- 35.2 after, p < 0.001). Three term and two preterm infants (5/35 = 14%; 95% CI: 5-30%) had a QTc >450 ms after cisapride. QTd after cisapride increased significantly in all infants with prolonged QTc. Filtered QRS, LAS40, and RMS40 before and after cisapride were within our normal values. We conclude that cisapride prolongs ventricular repolarization in neonates and infants without altering depolarization. Although no clinical arrhythmias were observed the dose of 0.8 mg/kg/day should not be exceeded.