Novacor left ventricular assist system long-term performance: comparison of clinical experience with demonstrated in vitro reliability.

Since the first implant of the Novacor wearable left ventricular assist system (LVAS) in 1993, median implant duration worldwide has increased from 93 days (max 2.2 years) to 202 days (max 4.1 years) in May 2001. In vitro reliability/durability testing of the Novacor LVAS has previously demonstrated a mean time to failure of 4.2 (3.04-5.59) years. These tests revealed a single failure mode--main bearing wear--with measurable symptoms gradually appearing before degradation of pump function. An ongoing clinical study of 37 recipients implanted for more than 1 year has shown that a simple noninvasive method of pump surveillance, derived from the in vitro experience, is well tolerated in the clinical setting. The overall clinical experience is consistent with in vitro reliability/durability tests that demonstrated 99.9%, 98.5%, and 87.4% freedom from wear at 1, 2, and 3 years (80% confidence). So far, the clinical study has shown 99.4%, 91.5%, and 91.5% freedom from wear at the same 1, 2, and 3 year intervals (95% confidence). Furthermore, the clinical findings have corroborated the in vitro experience that the wear mechanism is generally measurable and gradual, enabling elective clinical LVAS replacement or transplantation. Clinical valve performance was also monitored, using an exercise protocol and collecting comparative data on peak flows across the valves. It was determined that developing valve dysfunction could be diagnosed early and, in the failures that did occur (n = 2), these were related to the patient's disease state. In conclusion, although clinical conditions exposed the LVAS to a wide range of different environmental and hydraulic stresses, the surveillance program described appears practical and reliable, and its findings broadly parallel those of the earlier in vitro study. Additional data needed to complete formal validation continue to be collected.