The aims of this randomized study were (1) to determine if isoflurane is effective in controlling blood pressure during thoracic aortic cross-clamping, and (2) to compare its effects on hemodynamics and oxygen transport to those of sodium nitroprusside. Sodium nitroprusside (SNP group, n = 10) or isoflurane (ISO group, n = 10) was started 2 minutes before cross-clamping and was adjusted to maintain systolic arterial pressure as near as possible to preinduction values. The duration of thoracic aortic cross-clamping was 26 +/- 4 minutes in the SNP group and 30 +/- 4 minutes in the ISO group. Administration of isoflurance and sodium nitroprusside was stopped 2 minutes before unclamping. The same anesthetic technique using fentanyl, 6 micrograms/kg, flunitrazepam, 0.02 mg/kg, pancuronium, 0.1 mg/kg, and 50% N2O was used for all patients. At the time of clamping, either isoflurance (maximal expired concentration, 2.5% +/- 0.3%) or sodium nitroprusside (cumulative dose, 11.1 +/- 1.0 mg) was effective in maintaining the systolic blood pressure below 160 mm Hg, whereas the pulmonary capillary wedge pressure did not change. However, only SNP was able to bring the arterial pressure above the cross-clamp back to postinduction levels. During clamping, stroke index values were similar in both groups, but cardiac index increased only in patients receiving SNP. In both groups, at clamping and unclamping, PvO2 was higher than postinduction values, indicating that throughout the study the oxygen needs of the perfused area were adequately met. There was no evidence of acute left ventricular decompensation because pulmonary capillary wedge pressures did not abruptly increase, nor did pulmonary edema occur.(ABSTRACT TRUNCATED AT 250 WORDS)