Left Ventricular Assist Device Thrombosis Is Associated With an Increase in the Systolic-to-Diastolic Velocity Ratio Measured at the Inflow and Outflow Cannulae.

OBJECTIVE: To determine whether the ratio of peak systolic-to-nadir diastolic velocity (S/D ratio) measured using Doppler at the left ventricular assist device (LVAD) inflow and outflow cannulae is associated with pump thrombosis and to determine whether there is an absolute decrease in the diastolic cannula velocities in LVAD thrombosis. DESIGN: Retrospective chart review. SETTING: University hospital. PARTICIPANTS: Patients who underwent LVAD exchange. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Transesophageal echocardiograms were reviewed from all patients with the HeartMate II device (Thoratec Corporation, Pleasanton, CA) over a 6-year period and who underwent LVAD exchange for pump thrombosis. The following 3 time points were evaluated: (1) initial LVAD placement (prethrombosis), (2) thrombosis, and (3) exchanged LVAD placement (postthrombosis). Systolic and diastolic flow velocities were examined using pulse-wave spectral Doppler at the inflow and outflow cannulae, and the S/D ratio for each was determined. Statistical analysis was performed with SAS, version 9.4 (SAS Institute, Cary, NC), using 2-tailed tests and alpha = 0.05. Thirteen patients were included in the study. Significant differences were observed in S/D ratios among the 3 phases at both the inflow (p = 0.0234) and outflow (p = 0.0047) cannulae. Pairwise tests of the inflow cannulae showed that the mean S/D ratio at the time of thrombosis (mean ± standard deviation [SD], 4.29 ± 1.74) was significantly greater than the prethrombosis ratio (2.49 ± 0.65; p = 0.0069). Among outflow measurements, the mean S/D ratio at thrombosis (3.94 ± 1.34) was significantly higher than both the prethrombosis (2.63 ± 0.56; p = 0.0025) and postthrombosis (2.74 ± 0.83) (p = 0.0093) ratios. Decreases in diastolic velocities were not statistically significant at the inflow cannula. At the outflow cannula, there was a significant difference in diastolic velocity among the phases (p = 0.0233). Specifically, the postthrombosis diastolic measurements (41.50 ± 9.94) were significantly higher than both the prethrombosis (26.85 ± 10.13; p = 0.0140) and thrombosis (26.7 ± 15.35; p = 0.0151) values. CONCLUSIONS: An increased S/D ratio measured with Doppler at the LVAD inflow and outflow cannulas may be associated with pump thrombosis. Decreased diastolic cannula velocities were not observed in LVAD thrombosis.