Artificial circulatory support with textured interior surfaces. A counterintuitive approach to minimizing thromboembolism

Although numerous left ventricular assist devices (LVADs) have been used clinically, frequent thromboembolic complications have been reported despite the smooth interior LVAD surfaces and systemic anticoagulant medication. In contrast, the Thermo Cardiosystems HeartMate 1000 IP LVAD has textured interior surfaces that are promptly covered by a densely adherent neointima. We hypothesize that elimination of a direct interface between prosthetic material and blood elements reduces the risk of peripheral embolization and minimizes the necessity for systemic anticoagulant medication. This report defines the thromboembolic risk of this type of LVAD and characterizes the nature and effectiveness of the various anticoagulation regimens that were tested during the initial clinical trial with this device. All values are reported as mean +/- SD. Fifty-four males and three females with an average age of 47 +/- 11 years were supported with the HeartMate 1000 IP LVAD for an average of 62 +/- 76 days at 11 clinical centers in the United States. Patients were prospectively evaluated for thromboembolic complications. Five different anticoagulation regimens were used during the first 4 postoperative weeks: no anticoagulants, low-molecular-weight dextran, heparin, dipyridamole plus aspirin, or miscellaneous agents. After the first 4 weeks, the patients were treated with aspirin plus dipyridamole or miscellaneous agents. Prothrombin time (PT), partial thromboplastin time (PTT), and fibrinogen values for the patients were measured at 0.1, 1, 2, 4, 8, 12, 16, 20, 24, 32, and 46 weeks during support. Two patients (3.5%) suffered thromboembolic cerebrovascular complications, an incidence of 0.2 episodes per patient-year of observation. One episode was due to fungal vegetation developing on the device and the other was due to embolization from a previously placed native mechanical aortic valve prosthesis. In the absence of infection, there were no device-related thromboembolic complications. Mean prothrombin time for all groups was 13.3 +/- 0.5 seconds with no significant intergroup differences. Mean partial thromboplastin time during the first 4 weeks for the heparin-treated group was 53.3 +/- 6.6 seconds, which was significantly longer than for all other groups, but fell to control values after heparin was discontinued at 4 weeks. Mean fibrinogen level for all groups was 370 +/- 48 mg/dL, with no intergroup differences. The HeartMate 1000 IP LVAD provides adequate circulatory su