We conducted a randomize study of fentanyl compared to isoflurane anesthesia in patients with pulmonary hypertension undergoing mitral valve surgery. Patients were premedicated and randomly assigned to one of two groups: 21 patients had anesthesia induced with thiopental and maintained with isoflurane; 23 patients had anesthesia induced with a fentanyl bolus and maintained with a fentanyl infusion. Adjustments of fentanyl infusion and isoflurane concentration, as well as fentanyl boluses and vasoactive/positive inotropic medication, were administered to maintain preoperative arterial blood pressure. Both groups exhibited similar demographics, similar duration of cardiopulmonary bypass (CPB) surgery, anesthesia, and time from entrance into the surgical intensive care unit (SICU) to endotracheal extubation. However, the time from entrance into the SICU to awake was significantly (P < 0.05) shorter in patients given isoflurane anesthesia. Hemodynamic variables were recorded at baseline and 12 surgical events and compared between and within groups. Significant changes from baseline were demonstrated in both groups upon institution and discontinuation of CPB. Patients receiving isoflurane anesthesia exhibited cardiovascular depression as compared to their baseline. There were no deaths in either patient group. Adequate hemodynamic profiles were achieved in both groups with comparable use of inotropic and vasoactive medication, with the exception of norepinephrine that was administered intraoperatively to significantly (P < 0.05) more patients in the isoflurane-based anesthesia group. Neither technique was associated with acute improvement of right heart performance or pulmonary hypertension, in large part because of morphologic changes of the pulmonary arterial bed, occurring with long-standing mitral valve disease. We conclude that isoflurane-based anesthesia is adequate for this type of surgery, although there is a higher anesthetic algorithm failure rate than with fentanyl-based anesthetic technique.