Thrombotic obstruction of Bjork-Shiley valves--diagnostic and surgical considerations.

We have evaluated 12 patients with thrombotic obstruction of the Bjork-Shiley valve since 1975, 11 in the mitral and 1 in the aortic position. During this period 442 patients with 303 mitral (181 plano-convex and 122 convexo-concave) and 205 aortic (112 plano-convex and 93 convexo-concave) Bjork-Shiley valves were available for follow-up. The incidence of thrombosis for the plano-convex model was 1.06% per patient year for the mitral position and 0.19% per patient year for the aortic position. The new convexo-concave model has brought down the incidence to 0.78% per patient year for the mitral (P less than 0.01) and 0% per patient year for the aortic valve. The onset of symptoms was acute (less than 15 days) in 41.7% and subacute (greater than 15 days) in 58.3% patients. All patients presented with pulmonary edema. Evidence of inadequate anticoagulation was present in only 3 (25%) patients. Reduction of prosthetic sounds and appearance of a new murmur was highly suggestive of valve thrombosis. Echocardiography and cinefluoroscopy was very useful for the instant recognition of this condition and had obviated the need for cardiac catheterisation in the last 6 patients. Emergency surgery was obligatory in all. Thrombectomy alone was successful in 9 patients. Three patients required replacement of the prosthesis. There was one operative death (mortality 8.3%). The long-term follow-up of these patients (3-82 months, mean 34 months) is excellent. We conclude that thrombotic obstruction of the Bjork-Shiley valve is often not related to inadequate anticoagulation, and more than half of the patients do not present with abrupt onset of symptoms. The convexo-concave model has significantly reduced this problem. Emergency surgery with thrombectomy is the procedure of choice for clotted prostheses.