Left atrial cross-sectional area is a novel measure of atrial shape associated with cardioembolic strokes

Timothy C Tan,Maria Carmo Pereira Nunes,Mark Handschumacher,Octavio Pontes-Neto,Yong-Hyun Park,Cashel O'Brien,Victoria Piro,Gyeong-Moon Kim,Johanna Helenius,Xin Zeng,Jose Luiz Padilha da Silva,Karen Furie,Hakan Ay,Judy Hung
DOI: https://doi.org/10.1136/heartjnl-2019-315964
IF: 5.7
Heart
Abstract:Objective: Cardioembolic (CE) stroke carries significant morbidity and mortality. Left atrial (LA) size has been associated with CE risk. We hypothesised that differential LA remodelling impacts on pathophysiological mechanism of major CE strokes. Methods: A cohort of consecutive patients hospitalised with ischaemic stroke, classified into CE versus non-CE strokes using the Causative Classification System for Ischaemic Stroke were enrolled. LA shape and remodelling was characterised by assessing differences in maximal LA cross-sectional area (LA-CSA) in a cohort of 40 prospectively recruited patients with ischaemic stroke using three-dimensional (3D) echocardiography. Flow velocity profiles were measured in spherical versus ellipsoidal in vitro models to determine if LA shape influences flow dynamics. Two-dimensional (2D) LA-CSA was subsequently derived from standard echocardiographic views and compared with 3D LA-CSA. Results: A total of 1023 patients with ischaemic stroke were included, 230 (22.5%) of them were classified as major CE. The mean age was 68±16 years, and 464 (45%) were women. The 2D calculated LA-CSA correlated strongly with the LA-CSA measured by 3D in both end-systole and end-diastole. In vitro flow models showed shape-related differences in mid-level flow velocity profiles. Increased LA-CSA was associated with major CE stroke (adjusted relative risk 1.10, 95% CI 1.04 to 1.16; p<0.001), independent of age, gender, atrial fibrillation, left ventricular ejection fraction and CHA2DS2-VASc score. Specifically, the inclusion of LA-CSA in a model with traditional risk factors for CE stroke resulted in significant improvement in model performance with the net reclassification improvement of 0.346 (95% CI 0.189 to 0.501; p=0.00001) and the integrated discrimination improvement of 0.013 (95% CI 0.003 to 0.024; p=0.0119). Conclusions: LA-CSA is a marker of adverse LA shape associated with CE stroke, reflecting importance of differential LA remodelling, not simply LA size, in the mechanism of CE risk.
What problem does this paper attempt to address?