Abstract Cardiac output was measured by the dye-dilution method in 181 patients up to nine days after cardiac surgery. Three hundred thirty-two valid determinations were performed in duplicate or triplicate. Cardiac index could not be predicted from any of fourteen patient-related variables or from an optimal linear regression function of them. The ability of any variable to predict survival could not be tested because of the small number of patients who died. However, survival for more than seven days after operation was most closely related to high values of cardiac power and power index, but not to cardiac output or cardiac index. Patients with valve replacement had greater central venous pressure, 14% lower stroke work index, and 16% lower power index than patients without valve replacement. Cardiac and power indices increased with postoperative time. Central venous pressure and peripheral resistance index were unrelated to postoperative time. These findings suggest that cardiac competence was increasing, while peripheral resistance was relatively constant.