S3 heart sound amplitudes measured by accelerometer reflect auscultated S3 heart sound volume grades

Background The third heart sound (S3) is regarded as one of the earliest signs of heart failure (HF) and is typically assessed via auscultation clinically. The embedded accelerometer sensor in implantable devices can pick up the mechanical vibration caused by S3. We compared the device-measured S3 amplitude (devS3) to auscultated S3 volume grades (audS3). Methods In the MultiSENSE study, ambulatory HF patients implanted with a CRT-D were enrolled and followed up for 1 year. DevS3 amplitude was trended daily. Clinicians were asked to auscultate for S3 intensity on a graded scale (None/Low/Moderate/Loud) at baseline and each regular follow-up. Clinicians were blinded to the devS3 data at the time of assessment. We compared devS3 on the day of assessment in patients with different level of audS3 intensity. ANOVA test was used to compare the mean S3 amplitudes among groups and unpaired t-tests for between-group comparison. Conclusion This demonstrates that device-measured S3 amplitude reflects auscultated S3 heart sound volume grades. Since device-measured S3 is an objective measure on a continuous scale as opposed to auscultation which is subjectively categorized, it could avoid inter-observer auscultation variability and enables monitoring of this clinical variables in ambulatory HF patients. Result A total of 5211 auscultation reports were obtained from the 900 patients in the study. AudS3 intensity was None in 4709, Low in 391, Moderate in 108, and Loud in 3 reports (moderate and loud were combined for this analysis). Figure represents the devS3 amplitude (mean ± standard error) on the day of the follow-up as a function of the audS3 grades reported, the mean devS3 values for None, Low, and Moderate/Loud audS3 was 0.96 ± 0.01 mG, 1.02 ± 0.02 mG, and 1.27± 0.06 mG respectively. Both ANOVA test and unpaired tests between groups were significant and remained significant after adjusting for repeated measures.