ABSTRACT
Accelerated fatigue testing of clinical heart valves has been performed at cyclic rates of 33 to 35 cycles per second at 37 degrees C using water for non-biological valves and glutaraldehyde solutions for tissue valves. Flows were in the physiological range, and the pressure difference across each valve during closure was 100 +/- 25 mm Hg. The results showed that major fatigue occurred for the Starr-Edwards 2320 at 150 million cycles, the Hufnagel trileaflet at 124 million cycles, the Björk-Shiley Delrin disc at 140, the Björk-Shiley Pyrolite disc at 973, the Beall 103 at 60, the Hancock porcine at 62, the Carpentier-Edwards porcine at 34, and the Ionescu-Shiley porcine pericardial prosthesis at 65 million cycles. The Lillehei-Kaster was removed after 762 million cycles without discernible wear. Three facts emerged from the testing data: (1) the component worn in vitro wears in vivo; (2) the sites of in vitro fatigue on the component are identical to clinical specimens; and (3) those valves that have high durability in vitro have given similar performance in patients. The in vitro and clinical data for tissue valves do not correlate. The possible reasons for the discrepancy are discussed, and a note of caution is made regarding realistic expectations of clinical durability of tissue valves.
