Flow transport and not ejection fraction determines blood stasis in patients with impaired left ventricular systolic function.

Impaired left ventricular (LV) systolic function is a risk factor for intraventricular thrombosis and cardioembolism. However, below a given threshold, LV ejection fraction (EF) poorly predicts these events, suggesting the existence of additional sources of variability. We introduce queue models of LV blood transit connecting flow component analysis and residence time (RT) mapping. These models yield closed-form expressions for the average RT of blood in the LV as a function of (1) EF, (2) direct flow (DF), and (3) residual volume (RV). Models' performance was tested against RT obtained from vector flow mapping in 332 subjects, including controls and patients with acute myocardial infarction (AMI), hypertrophic (HCM), and dilated cardiomyopathy (DCM). Queue models show RT is increasingly sensitive to DF as EF decreases, contradicting the traditional view of large DF as a teleological advantage. Instead, RT is minimized when blood transits in a first-in-first-out (FIFO) manner, while DF short-circuits the FIFO pattern, prolonging RT for other flow components. FIFO models showed a good performance in assessing RT in the studied subjects. Our results show that large DFs increase blood stasis when EF is low. These models also explain why EF is a poor marker of the risk of intraventricular thrombosis.