Background and aim of the study: The current standard of in vitro and in vivo preclinical heart valve testing has recently been questioned because of its failure to reveal the thrombogenic potential of the Medtronic Parallel prosthetic valve. The aim of this study was to develop a swine model for the in vivo preclinical evaluation of mechanical heart valves, and to assess the ability of this model to identify mechanical heart valve design features that result in valve-related thrombosis.
Methods: Twenty-two swine underwent mitral valve replacement (MVR) using three different bileaflet mechanical valve designs (St Jude Medical, CarboMedics, Medtronic Parallel). Each animal was placed in an anticoagulation protocol (group I, INR 3.0-3.5; group II, INR 2.0-2.5; group III, no anticoagulation) and followed for up to 20 weeks. Terminal studies were performed on all animals surviving for more than 30 days.
Results: Twenty-one animals survived the immediate postoperative period. Four of six group I animals died from hemorrhagic (large wound hematoma; hemopericardium) complications early in the study. In the two long-term (61 and 89 days) survivors, INRs of 3.0 to 3.5 were never achieved (61-day survivor, mean INR 2.0 +/- 1.03; range: 0.8-5.4; 89-day survivor, mean INR 1.92 +/- 1.34; range: 1.0-7.9). Pathological analysis of explants from group I survivors revealed minimally obstructive fibrous sheathing on the inflow orifice without restriction of bileaflet motion (61 and 89 days), and two large perivalvular defects (61 days). Six of seven group II animals died from early hemorrhagic complications (hemopericardium) (mean INR 2.32 +/- 1.84; range: 0.8-8.2). Vegetations resulting in obstruction of both sides of the valve orifice and restriction of bileaflet motion were observed in a group II survivor (mean INR 2.33 +/- 1.58; range: 0.9-7.0). Group III animals (n = 8) survived for a mean of 106 +/- 60 days (range: 1-177 days). In group III, fibrous sheathing was present on all explanted valves and organized thrombi in six valves; orifice obstruction (seven valves) and restriction of bileaflet motion (three valves) were also observed.
Conclusion: The use of MVR in swine as a preclinical model to evaluate the safety and performance of mechanical heart valves is limited by: (i) difficulty in maintaining safe levels of anticoagulation with warfarin, resulting in a high incidence of hemorrhagic complications; (ii) marked fibrous sheath formation and associated thrombosis; and (iii) an increased incidence of perivalvular defects, believed to result from normal somatic growth occurring in young swine.