In order to study the effect of cisapride on gastric stasis and to evaluate the possible risk of cholestasis, 20 premature neonates born in the hospital during one year were orally treated with cisapride 0.15 mg/kg q.i.d., over a mean period of 38 days. The gestational age ranged from 26 to 34 weeks and the mean age at the start of the cisapride treatment was 18 days. All patients were ventilated, 13 had a respiratory distress syndrome (hyaline membrane disease), and 9 had gastro-oesophageal reflux (GOR). All patients were given a semielementary formula by means of a continuous nasogastric infusion. The gastric residue was studied during three days: 24 hour baseline and 48 hours under cisapride treatment. The mean residue decreased (p less than 0.0001) from 50.6% during the last 6 baseline hours to 12.1% during the last 6-hours of the cisapride period. The mean feeding volume increased from 24.2 ml to 34.2 ml (p less than 0.001). A group of four patients had reversible cholestasis against the background of an outbreak of Candida, three before and one during cisapride treatment. Therefore, it could not be demonstrated that cisapride plays a role in the development of cholestasis. Because of the risks of GOR and the drawbacks of delayed enteral feeding, it is concluded that the use of cisapride is justified in premature neonates with gastric stasis.