Inconsistent Harmonic Peaks in Patients with Continuous Left Ventricular Assist Devices at Baseline

Abstract

Introduction Despite improvements in continuous left ventricular assist devices (LVADs), 8-11% of patients annually experience thrombosis. Quantitative acoustic analysis of LVADs has recently been investigated as a novel method of early thrombosis detection. Previous studies have shown the existence of peaks in sound amplitude at expected harmonics based on VAD speed, and changes in amplitude with thrombosis. However prior to utilizing this method to detect thrombosis in patients, it is important to establish a baseline of acoustics in VAD patients with no known thrombosis. Hypothesis In patients with HeartMateII (HMII) and HeartMateIII (HMIII) LVADs, with no clinical indication of thrombosis, there will be consistent characteristic peaks at the first 4 harmonics. Methods 19 HMII patients and 3 HMIII patients were recorded at the heart apex for 6-8 seconds using 3M™ Littmann® Electronic Stethoscope Model 3200. Patients were recorded at least once over 2 years, with 4 HMII and 2 HMIII patients recorded 2-3 times. Software used includes 3M™ Littmann® StethAssist™ to upload sound files, Audacity 2.1.2 to generate acoustic spectrum data via Fourier transform, and Microsoft Excel to visualize sound amplitude peaks in the frequency spectrum. Expected harmonic frequencies (Hλ) based on LVAD speed were calculated with the equation: Hλ = (n*RPM)/60. Patient demographics, VAD parameters, and laboratory studies were collected from the electronic medical record. Acoustics and patient data were analyzed, focusing on the presence or absence of the first four harmonics. Results In all HMII recordings, peaks were present at the 1st and 3rd harmonics. In 8 out of 19 HMII patients, all peaks 1-4 were present (Group 1). In 11 out of 19 HMII patients, either the 2nd +/- 4th peaks were missing (Group 2). Between the two groups there was no significant difference in vital signs, VAD parameters or labs. In Group 2, 3 out of 4 patients are missing the same harmonic peaks on subsequent recordings. Only one patient in Group 2 was inconsistent in peaks missing on subsequent recordings. In all HMIII recordings peaks at the 1st, 2nd, and 4rd harmonics were present. Four out of five recordings showed an extra peak between the 2nd and 3rd harmonics. The 3rd harmonic was present in only one recording and had a smaller sound amplitude than other peaks. Conclusions Although other studies have shown data with consistent peaks at all expected 1-4th harmonics, our study did not find this to be true. In HMII devices, it appears 1st and 3rd harmonics are most consistent across patients, while the 1st, 2nd, and 4th harmonics are most consistent in HMIII devices. Individual patients remain relatively consistent in harmonics missing, and therefore baseline acoustic spectrums may be patient specific similar to other VAD parameters. Consequentially, it may be necessary to determine a patient's own baseline acoustics and use the deviation as an indicator of possible pathology or significant physiologic change.