Time domain features of heart sounds for determining mechanical valve thrombosis

Abstract

Thrombosis of implanted heart valve is a rare but lethal complication for patients with mechanical heart valve. Echocardiogram of mechanical heart valves is necessary to diagnose valve thrombosis definitely. Because of the difficulty in making early diagnosis of thrombosis, and the cost of diagnosis equipment and operators, improving noninvasive, cheap and simple methods to evaluate the functionality of mechanical heart valves are quite significant especially for first step medical center. Because of this, time domain features obtained from auscultation of heart sounds are proposed to evaluate mechanical heart valve thrombosis as a simple method in this chapter. For this aim, heart sounds of one patient with mechanical heart valve thrombosis and five patients with normally functioning mechanical heart valve were recorded. Time domain features of recorded heart sounds, the skewness and kurtosis, were calculated and statistically evaluated using paired and unpaired t-test. As a result, it is clearly seen that the skewness of first heart sound is the most discriminative features (p < 0.01) and it may be used fairly well in differentiating normally functioning mechanical heart valve from malfunctioning mechanical heart valve. © 2011 Springer Science+Business Media B.V.