Clinical temporary ventricular assist. Pathologic findings and their implications in a multi-institutional study of 41 patients.

Forty-one patients, distributed among four centers, had left (33 patients), right (five), or bilateral (three) temporary ventricular assistance with textured (24) or smooth (17) surfaced diaphragm pumps, during an evaluation supported by the National Institutes of Health. Cardiac failure had occurred in 39 postoperative patients (after aorta-coronary bypass [23], valve replacement [four], both [nine], or other [three]), with total cardiopulmonary bypass time mean 306 minutes (range 69 to 600). Two patients had cardiomyopathy. Death of 35 nonsurvivors was due to myocardial necrosis (14), hemorrhage (nine), cerebrovascular accidents (three), infection (three), and other (six). Mean duration of support in all patients was 62 hours. In 16 patients (40%) whose condition improved, cardiac assist duration was mean 127 hours (range 48 to 264), compared with mean 19 hours (range 1 to 120) in 25 who did not. Of 17 patients in whom duration of support exceeded 72 hours, 15 (88%) improved, 11 were weaned, and six survived long term. Tissue examination (in 33 patients) by biopsy at pump implantation or autopsy revealed coagulation or contraction band myocyte necrosis, with or without hemorrhage, in 26 patients; of these, 10 improved and six were long-term survivors. Pump-related complications (two) included pulmonary embolism, most likely related to a cannulation site thrombus, and an aortic cannulation site infection in one patient each. This study suggests that mechanical cardiac assist may be accomplished with a low complication rate; should not necessarily be denied to patients with existing necrosis, because myocardial necrosis does not preclude improvement or survival; and frequently leads to functional myocardial recovery if patients survive early noncardiac complications, often the result of long duration of cardiopulmonary bypass.