A study of the potential use of digitally recorded heart soundsto screen for heart failure with reduced ejection fraction andoptimize response to cardiac resynchronization therapy

Heart failure with reduced ejection fraction (HFrEF) is a leading cause of cardiovascular mortality and morbidity. Diagnosis of heart failure requires clinical assessment of a patient and the use of echocardiography or other clinical tools that are not immediately accessible to patients at the bedside. This thesis aimed to study the potential use of externally recorded heart sounds recorded with an electronic stethoscope (estethoscope) to screen for heart failure with reduced ejection fraction and optimise response to cardiac resynchronisation therapy (CRT) as treatment for HFrEF. Based on recording of heart sounds with an e-stethoscope from 37 subjects without diagnosis of HFrEF and 59 subjects with HFrEF, differences in acoustic features of heart sounds in the frequency domain were found between presence and absence of HFrEF. 53 subjects with HFrEF were indicated for implantation of CRT. Heart sounds were recorded along with recording of ECG, non-invasive haemodynamics and echocardiographic measurements before CRT implant and at 4 to 6 months and 10 to 12 months after implantation of CRT. Response to CRT was assessed using a combination of subjective and objective clinical parameters including NYHA class, 6- minute walk test and Minnesota Living with Heart Failure score. A frequency-based acoustic parameter was found to increase from baseline in responders to CRT but not in non-responders to CRT or in patients with undetermined response to CRT. In 11 patients who underwent echocardiography-guided optimisation of paced atrioventricular delay (PAVD) of CRT devices, heart sounds were recorded simultaneously with an estethoscope. It was found that the optimal PAVD obtained from a frequency-based acoustic feature of heart sounds differed from those obtained from conventional Doppler echocardiography methods but the differences in mitral inflow duration and aortic valve velocity time integral were unlikely to be clinically significant. Therefore, the acoustic feature may still have some potential to predict the optimal PAVD. Digitally recorded heart sounds with an e-stethoscope have potential clinical applications in screening for HFrEF, monitoring response to treatment with CRT and optimisation of CRT.