We prospectively compared the hemodynamic performance of Hancock and Carpentier-Edwards bioprosthetic aortic valves in a randomized study of 100 patients. A total of 47 patients received the Hancock valve, and 53 received the Carpentier-Edwards valve. Mean pressure gradients were measured using micromanometer catheters and cardiac outputs by thermodilution. Multiple measurements were made in each patient with atrial pacing, volume infusion, and inotropic drugs for a total of 319 observations. The pressure gradients and Gorlin valve areas showed significant scatter caused by both flow-dependent and patient-dependent variability. Mean transvalvular pressure gradients were therefore compared after adjustment for flow rate and for random interpatient differences using analysis of variance and covariance. Pressure gradients were lower and Gorlin valve areas larger for the Hancock valve than for the Carpentier-Edwards valve, but the differences were significant only for the smaller valve sizes. Compared with the Carpentier-Edwards valve, the mean pressure gradients were significantly lower for the Hancock 19-mm modified orifice (MO) valves (16.9 versus 31.7 mm Hg, p = 0.04), for the 21-mm valves (15.2 versus 22.4 mm Hg, p = 0.003), and for the 23-mm MO valve (9.2 versus 13.8 mm Hg, p = 0.04). The Gorlin areas were also significantly larger for the Hancock 19-mm MO valve (0.85 versus 0.77 cm2, p = 0.004) and the 21-mm MO valve (1.11 versus 0.89 cm2, p = 0.0009) but not for the 23-mm MO valve (1.59 versus 1.14, p = 0.08). Mean gradients and valve areas were not different for any of the larger valve sizes.(ABSTRACT TRUNCATED AT 250 WORDS)