The Effect of Pump Speed Settings on Suspected Pump Thrombosis in Patients Supported With Continuous-Flow Left Ventricular Assist Devices

Background: Pump speed is the only modifiable setting after LVAD implantation. In current practice, pump speeds are selected to optimize LV unloading while minimizing suction events. Echocardiographic and hemodynamic studies have taken an increasing role to ensure appropriate settings. Low pump speed has been theorized to increase the risk of thrombus formation in the pump. In this exploratory analysis, we sought to determine the effect of pump speed on the risk of pump thrombosis (PT) in patients supported with continuous-flow LVADs. Methods: We retrospectively identified 418 patients who underwent implantation of a HeartMate II® LVAD at our institution from 6/2005–9/2014. For each patient, pump speed settings were obtained post-operatively during index hospitalization for LVAD implantation and throughout the follow-up period. Patients were included for analysis if they survived index hospitalization after LVAD and were maintained on a pump speed that remained consistent for the duration of support. Patients with below median pump speed (<9400 RPM) were compared with those with above median pump speed (≥9400 RPM). Baseline characteristics were compared between groups with Fisher's exact test and Student's t-test used to compare categorical and continuous variables, respectively. The primary outcome was freedom from suspected PT with Kaplan-Meier analysis performed to determine outcomes over time. Results: After exclusions, 276 patients were included for analysis with 124 patients with pump speed <9400 RPM compared with 152 patients with pump speed ≥9400 RPM. Patients with low pump speed were older and more likely to be implanted as destination therapy and to have lower BMI, ischemic cardiomyopathy and tobacco use. After median follow-up of 391 days, there was a non-significant trend toward increased suspected PT in patients with low (<9400 RPM) pump speed (P = .11, Fig. 1). When comparing patients with a pump speed <9000 RPM to those with standard (9000–9600 RPM) and high (>9600 RPM) pump speeds, there was a persistent trend toward increased suspected PT at lower speeds that did not reach statistical significance. Conclusion: There was a non-significant trend toward increased suspected PT in patients with low pump speed throughout the duration of LVAD support. Further studies are needed in larger cohorts to define the role of low pump speed in the development of PT and to determine the need to consider the possibility of adverse outcomes such as PT when determining pump speed settings.