Hemodynamic changes during four anesthetic techniques were studied in 80 preterm neonates. Atropine, 0.02 mg/kg, and pancuronium, 0.1 mg/kg, were given intravenously to all patients, who were ventilated with oxygen and air. Each group of 20 patients then received 0.75% isoflurane, 0.5% halothane, 20 micrograms/kg fentanyl, or 2 mg/kg ketamine. Heart rate (HR), systolic blood pressure (SAP), and mean blood pressure (MAP) were recorded at 1-min intervals until surgical stimulation. HR remained at or above control level in all groups. Statistically significant decreases (P less than 0.01) in SAP and MAP occurred following administration of each anesthetic. SAP decreased 30% during isoflurane administration, 25% during halothane, 21% following fentanyl, and 16% following ketamine. Clinically important decreases (25% or greater) in SAP were observed in some patients in each group, but the incidence was significantly less in patients receiving ketamine (P less than 0.02). The covariables of conceptual age, postnatal age, weight, urine specific gravity, hematocrit, and presence of patent ductus arteriosus did not have statistically significant effects on SAP and MAP changes. The authors conclude that SAP and MAP decrease significantly during each of the anesthetic techniques studied and that clinically important decreases in SAP occur less frequently during the technique using ketamine.