A Calf Model for Left Ventricular Centrifugal Mechanical Assist.

Abstract

The increased use of centrifugal mechanical assist (CMA) for treatment of refractory postcardiotomy cardiogenic shock highlights the need for experimental testing to improve clinical results. This report describes the preoperative conditioning, anesthetic and surgical technique, and postoperative management of a reliable calf model refined in this laboratory for in vivo sub-chronic (96 h) testing of CMA. Holstein bull calves (2 to 3 months old; mean body weight, 78 kg; n = 35) were instrumented for left ventricular CMA; 4 of these calves were sham-operated controls. Anesthetic recovery and postoperative restraint were accomplished in a specially designed crate to which each calf was preconditioned extensively. Younger calves were more readily conditioned and more tolerant of postoperative restraint than older calves. One calf died of ventricular fibrillation in-traoperatively. One calf that had been heparinized developed uncontrollable hemothorax and died 12 h postoperatively. One calf prematurely dislodged his aortic cannula 15 h postoperatively and exsanguinated. Six calves developed pelvic limb paresis or paralysis because of lumbar spinal cord thromboembolism by 36 h postoperatively, and 3 of these calves were sacrificed by 42 h postoperatively. Fifteen calves required sedation in the first 12 h after the operation. Tachycardia associated with bottle feeding occurred in 15 calves. Second-degree atrioventricular block was noted frequently during deep relaxation. Postmortem examination demonstrated the absence of surgical wound and distant infection, security of cannulae in all but the calf that prematurely dislodged the aortic cannula, absence of thrombus formation at cannu-lation sites, and presence of thromboembolism in 51% of the calves. The incidence of thromboembolic lesions was not influenced by the need for chemical restraint, by the occurrence of feeding-associated tachycardia, or by the presence of atrioventricular block. There were no thromboembolic lesions in any of the sham-operated controls.