Predicting Device-Induced Platelet Damage in Assisted Circulation Using a New Biophysical Model and Experimental...

Platelet activation and receptor shedding caused by non-physiological shear stress remain a critical challenge in the design and clinical performance of circulatory assist devices. Computational fluid dynamics (CFD)-based approaches have been widely used by coupling flow-derived scalar shear stress with empirical damage models to predict platelet activation. In this study, we evaluate the effectiveness of our recently proposed biophysical damage model for predicting platelet activation and receptor shedding caused by circulatory assist devices.

Four rotary blood pumps have been chosen for this study: One axial pump (HeartMate II) and three centrifugal pumps (HeartMate 3, BrioVAD, and CentriMag). CFD simulations were performed under three clinically relevant circulatory support conditions: HeartMate II, HeartMate 3, BrioVAD at a pressure head of 75 mmHg, along with CentriMag at pressure heads of 75, 150, and 350 mmHg. The resulting platelet damage indices were compared with experimentally measured platelet activation (PAC-1 expression) and receptor shedding (GPIbα and GPVI) data.

HeartMate 3, BrioVAD, and CentriMag operating at 75 mmHg caused the lowest levels of platelet activation and receptor shedding, while CentriMag, operating at 350 mmHg, produced the highest, 10 times more activation, and six times more receptor shedding compared to at 75 mmHg. Strong correlations were observed between CFD predicted and in vitro data for all biomarkers ( $R^2>0.76$ ), with the highest agreement observed for GPVI shedding ( $R^2>0.95$ ).

The biophysical model preserved the relative trends observed previously with the power-law formulation, while demonstrating improved agreement with experimental data.