Centrifugal ventricular assist device for support of the failing heart after cardiac surgery

Objective: To determine the morbidity and mortality associated with use of centrifugal ventricular assist devices for postcardiotomy cardiogenic shock and to determine factors that might influence outcome and thus, aid in patient selection. Design: A retrospective study. Setting: Surgical intensive care unit in a university hospital. Patients: During a 6-yr period, a total of 7,385 adult patients underwent cardiac operations requiring cardiopulmonary bypass. Myocardial protection consisted of single-dose cold crystalloid cardioplegia and continuous topical hypothermia by saline lavage. A total of 72 (1%) patients developed postcardiotomy cardiogenic shock. Of 72 patients, 28 met the institutional criteria and were placed on centrifugal ventricular assist devices. Interventions: Twenty-eight adult patients with postcardiotomy cardiogenic shock were supported with centrifugal ventricular assist devices. Measurements and Main Results: A total of 15 patients received left ventricular assist devices, five received right ventricular assist devices, and eight received both right and left ventricular assist devices. Mean age of ventricular assistance patients was 50.8 ± 12.9 yrs (range 22 to 72), and mean duration of ventricular assistance was 2.8 ± 2.5 days (range 4 hrs to 10 days; median 2 days). Twenty-five complications occurred in 16 patients and included bleeding (13), tamponade (2), systemic embolism (6), seizures (2), and sepsis (2). Nine patients required reexploration for bleeding or tamponade. Nine (32%) of 28 patients were discharged from the hospital. Ventricular assistance for cardiac failure after transplantation was associated with improved survival (p < .10), while age >50 yrs and postoperative tamponade each showed trends toward association with mortality (p = .10). Survival was not predicted by gender, weight, time on cardiopulmonary bypass, aortic cross-clamp time, urgency of operation, or preoperative congestive heart failure. At 27 ± 20 months follow-up, all survivors were alive and New York Heart Association functional class I or II. Conclusions: These results document a low incidence of ventricular assist device use in a surgical practice that employs a relatively simple method of myocardial protection. When post-cardiotomy ventricular assistance was necessary, a centrifugal pump was used and successful outcome and satisfactory long-term results were possible in nearly one third of patients. Ventricular assistance for cardiac failure after transplantation was associated with improved survival. Older age is a relative contraindication to mechanical ventricular assistance.