P989Determinants of beat-to-beat left ventricular function during atrial fibrillation: a combined clinical-computational study

Abstract

Abstract Funding Acknowledgements Netherlands Organisation for Scientific Research (NWO-ZonMw, VIDI grant 016.176.340) Background The rapid irregular atrial electrical activity during atrial fibrillation (AF) is associated with an irregular and variable left ventricular (LV) systolic pump function. The mechanisms determining LV function during AF remain incompletely understood. Purpose To assess the reliability of global longitudinal strain (GLS) as a measure of LV function during AF, and to elucidate how beat-to-beat changes in LV preload and afterload affect LV function during AF. Methods Beat-to-beat speckle-tracking echocardiography was performed in patients with persistent AF. A hundred consecutive beats in each patient were imaged during AF and we evaluated the relation between GLS in the image plane (4-chamber view) and (pre-)preceding cycle length (CL) in these patients. We used the CircAdapt cardiovascular system model to simulate cardiac mechanics and hemodynamics during AF for each individual patient 1) by imposing the exact irregular sequence of CL as measured in the patient and 2) by making the atrial myocardium non-contractile. These simulations enabled beat-to-beat quantification of preload (end-diastolic volume, EDV), afterload (systolic aortic pressure) and GLS during AF. Results Clinical data and simulations both showed a negative non-linear relation between preceding CL and GLS (Panel A). Non-linearity was more pronounced at low preceding CLs (<750ms), while GLS at preceding CLs >750ms showed less dependence on CL. Simulating ventricular failure by a reduction in ventricular contractility led to a lower overall GLS and a loss of non-linear response at low CL (Panel A, right, diamonds). Increased GLS at high preceding CL (purple box) was explained by a higher preceding EDV (higher preload) (p < 0.002, Panel B), reflecting the Frank-Starling mechanism of contractile myocardium. At a given preceding CL, variability in GLS was explained by the afterload of the preceding beat, with a lower preceding afterload (systolic aortic pressure) leading to higher GLS (Panel C, yellow box, p < 0.002), but not by changes in preceding preload (Panel C, p non-significant). Preload of the pre-preceding beat also correlated with changes in GLS (Panel C, p < 0.005). Conclusions During AF, GLS depends non-linearly on the preceding CL, with GLS measures performed at longer preceding CLs (>750ms in our cohort) showing a low beat-to-beat variability. Beat-to-beat hemodynamic changes in preceding afterload could explain differences in LV function at same preceding CL. Our combined clinical-computational study highlights the variability in GLS measurement during AF and provides new insight into the potential hemodynamic mechanisms determining LV function in AF patients. Abstract Figure.