Evaluation of Eulerian and Lagrangian Models for Hemolysis Estimation

TLDR

Hemolysis from flow‑induced mechanical damage remains a challenge in devices such as VADs, artificial lungs, and heart valves, and although power‑law models are popular, many different implementations exist. This study aimed to evaluate various power‑law hemolysis models by computing hemolysis in a custom shearing device and a clinical VAD and comparing the predictions to experimental measurements. The authors applied both Eulerian scalar‑transport and multiple Lagrangian models to the two devices, then compared the calculated hemolysis to laboratory data. The models exhibited large absolute errors (up to 91 % for Eulerian and 57 % for Lagrangian) and could not predict magnitude accurately, but the Eulerian approach showed strong correlation (>0.99), indicating it can reliably rank devices or guide design optimization.

Abstract

Hemolysis caused by flow-induced mechanical damage to red blood cells is still a problem in medical devices such as ventricular assist devices (VADs), artificial lungs, and mechanical heart valves. A number of different models have been proposed by different research groups for calculating the hemolysis, and of these, the power law-based models (HI(%)=Ct(α)τ(β)) have proved the most popular because of their ease of use and applicability to a wide range of devices. However, within this power law category of models there are a number of different implementations. The aim of this work was to evaluate different power law-based models by calculating hemolysis in a specifically designed shearing device and a clinical VAD, and comparing the estimated results with experimental measurements of the hemolysis in these two devices. Both the Eulerian scalar transport and all the Lagrangian models had fairly large percentage of errors compared with the experiments (minimum Eulerian 91% and minimum Lagrangian 57%) showing they could not accurately predict the magnitude of the hemolysis. However, the Eulerian approach had large correlation coefficients (>0.99) showing that this method can predict relative hemolysis, which would be useful in comparative analysis, for example, for ranking different devices or for design optimization studies.