Background and aim of the study: Currently, the sheep model is preferred for preclinical in-vivo evaluation of prosthetic heart valves implanted in the mitral position. In sheep, the anatomy and tissue characteristics in, and around, the native mitral valve's posterior commissure (12:00-3:00 quadrant) makes valve implantation technically challenging. As the majority of non-infectious paravalvular leaks occurred in this quadrant, the surgical technique was modified to offer greater exposure of the annulus in this region and permit more accurate placement of sutures.
Methods: A total of 223 valve implantations (138 bioprostheses, 85 mechanical valves) performed between 1991 and 1998 using the sheep model was retrospectively reviewed. No evidence of endocarditis was found, indicating that etiology of the leaks was due to surgical technique. The incidence of paravalvular leak in the 12:00-3:00 quadrant was compared with that in all other quadrants, both before and after the surgical technique change.
Results: The incidence of paravalvular leak was 33% (28/85) for mechanical valves and 13.8% (19/138) for bioprosthetic valves. Leaks in the 12:00-3:00 quadrant accounted for 82% (14/17) of those occurring before the surgery change, and only 43% (13/30) afterwards. Before the change, the incidence of paravalvular leak was 20.3% (14/69) in the 12:00-3:00 position, and 4.3% (3/69) in the non-12:00-3:00 quadrants. After the change, incidence in the 12:00-3:00 position fell to 8.4% (13/154; p < 0.05), but that in other quadrants was not significantly affected. Analysis by valve type showed that leak incidence in the 12:00-3:00 quadrant fell from 20.3% (12/59) to 3.8% (3/79) in the bioprosthesis group (p < 0.05), and from 20.0% (2/10) to 13.3% (10/75) in the mechanical valve group (p = 0.56). In all other quadrants there was no statistically significant change in either groups.
Conclusion: Technical refinement of surgery has led to a reduced incidence of paravalvular leak in the 12:00-3:00 quadrant, thereby providing a standard by which to compare prostheses implanted in the mitral position. These data should also assist in improving prosthetic heart valve design by enabling device-related complications to be distinguished from model-related complications.