CARD18: In vitro and in silico Characterization of Aggregation and Embolization of Thrombi on Ventricular Assist Device Cannula

Abstract

Background: The unacceptably high stroke rate of HeartMate III VAD without signs of pump thrombosis is hypothesized to be the result of the thrombi originating on the inflow cannula, ingesting and ejecting emboli from the VAD. Therefore, inflow cannula thrombosis has been an emerging focus. The inflow cannula of contemporary VADs, which incorporate both polished and rough regions serve as useful benchmarks to study the effects of roughness and shear on thrombogenesis. An in-vitro study was conducted to emulate the micro-hemodynamic condition on a sintered inflow cannula, and to observe the deposition and detachment patterns. With a computational fluid dynamic tool, this study aimed to provide insight into the optimization of inflow cannula and potentially reducing adverse neurological events due to upstream thrombus. Methods: A parallel plate channel was designed for this study which incorporates periodic teeth (approximately 75-micron Ra) that mimic the sintered surface treatment of a left ventricular inflow cannula. Anti-coagulated whole human blood (3.2% sodium citrate), collected from 17 healthy donors, was perfused through the microchannel by a syringe pump at calibrated flow rates to achieve shear rates of 1000, 2500, and 7500s-1 (Reynolds numbers 3.06 - 22.88). Deposition of fluorescently labeled platelets was visualized using inverted epifluorescence microscopy. Mean fluorescence intensity, deposition probability, and embolization pattern were collected. Numerical simulations of a previously validated thrombosis model were performed using OpenFOAM. A hexahedral dominant mesh was used with the transition to prism elements near the refinement region at the boundary. Results: The average fluorescence intensity map (Fig.1) depicts the time-course deposition within the serrated micro-channel for three perfusion rates studied. Sustained growth of platelets adhesion was seen in all shear conditions. Increasing the shear rate results in less adhesion and more bulk embolization, which was captured by real-time video. Variance in deposition location was also found associated with an increasing shear rate, confirmed by CFD simulation. At the lowest shear rate, platelets aggregated onto the tips of the teeth and were trapped within the valleys between teeth. At 7500 s-1, less deposition was found on the tips, and emboli were observed washing off from the tips. Conclusion: This in vitro and in silico thrombus formation study is the first to focus on emulating clinical-related surface roughness and supraphysiological shear of a VAD inflow cannula. The simulation reproduced the experimental observation of the shear-dependent spatial distribution of platelet deposition. The findings in the variance of adherent thrombus and mean intensity of platelet deposition may suggest that a supraphysiological flow induces embolization around the cannula’s rough surface, which raises concerns about hemocompatibility of inflow cannula without signs of adherent thrombus.Figure 1. (A) Experiment set-up to emulate the micro-hemodynamic condition on a sintered inflow cannula. (B) Mean fluorescent intensity of platelet deposition for three shear rates studied. N=5. (C) The results of thrombosis CFD simulation at the corresponding shear rates confirm the in vitro observation.