Left atrial appendage morphology and mitral annular plane systolic excursion predict stroke in patients with atrial fibrillation and high thromboembolic risk

Background Previous research has revealed a relationship between the morphology and function of the left atrial appendage (LAA) and the rate of thromboembolic events in patients with atrial fibrillation (AF). There is limited data on the impact of the left ventricular (LV) function on LAA thrombus formation. Emptying of the LAA depends on the active contraction of this structure as well as on the passive motion due to external compression from the LV base during ventricular diastole. Purpose Given these considerations the goal of this study was to investigate by transesophageal and transthoracic echocardiography the morphology and function of the LAA in relationship with the LV function and to assess the impact on incidence of stroke in a group of patients with AF with high thromboembolic risk who underwent TEE-guided direct current cardioversion. Methods We included 148 patients (age 65 ±11 yrs, 79men, median CHA2DS2-VASc score 4 (IQR 2-6), left ventricular ejection fraction 48±11%), who underwent transesophageal echocardiography before cardioversion. During a median follow-up period of 23 months (range 1-107) 34 patients developed stroke or TIA. Results The distribution of the four LAA morphologies (Figure 1) was: Cactus [47 (32%)], Cauliflower [30 (20%)], Windsock [38(26%)], and Chicken Wing [33 (22%)]. During the follow-up period stroke/TIA occurred significantly more frequently in patients with cactus or cauliflower LAA morphology vs. patients with windsock or chicken wing morphology (25 (35%) vs 9 (11%); p value=0.001). Patients who developed stroke or TIA had significantly lower MAPSE vs patients with event free survival (8±3 vs 12±3mm; p value=0.001). Kaplan-Meier survival curves for different categories of MAPSE and LAA morphology are shown in Figure 2. The significant relationship between MAPSE and stroke-free survival was retained in Cox multivariable analysis that included LAA morphology, CHA2DS2-VASc score and AF types. This association was present whether MAPSE was entered into the multivariable model as a continuous or dichotomized variable. Conclusions In patients with high thromboembolic risk, MAPSE predicts the occurrence of stroke independent of LAA morphology and CHA2DS2-VASc score. This phenomenon could be related to the impact LV diastole on LAA emptying.