Left atrial area as a simple predictor of paroxysmal AF in patient with acute ischemic stroke

Abstract

Abstract Funding Acknowledgements Type of funding sources: None. Introduction The most important cardiogenic reason of acute Ischemic stroke (AIS) is atrial fibrillation (AF). Left atrial (LA) enlargement (atriopathy) is one of the risk factors of atrial fibrillation. Basic bio-metric measurements (height , weight and BSA) in patients with AIS can be challenging due to immobilization, therefor estimation of LA size using recommended parameter (left atrial volume index) is often not possible in every day practice. Aim To assess the discriminant value of different LA size parameters for detection of known paroxysmal AF in patients with acute ischemic stroke. To estimate best cut-off value of LA size parameters which predict presence of paroxysmal AF in patients with AIS. Materials and methods A database of 663 consecutive patients admitted to a neurology department with the diagnosis of acute ischemic stroke (AIS) who underwent transthoracic echocardiography (TTE) were analysed. After exclusion of 117 patient with persistent or permanent AF (present during TTE) the examined group comprised of 546 patients. Average age was 70,1years (20-97yrs), 53% were women. From the examined group 56 (10,3%) had known (from past medical history or diagnosed during the hospitalization in ecg, telemetry or holter monitoring) paroxysmal AF. 35,6% of examined patients required bedside echocardiography due to immobilization. The predictive values for the presence of paroxysmal AF were assessed for: LA antero-posterior diameter (LAd) in parasternal long axis , LA area in four chamber diameter (LAA), as well as LAd and LAA indexed to aorta diameter(Ao) (innere-to-inner method on the level of sinus of Valsalvy). Receiver Operating Characteristic (ROC) curves, area under the curve (AUC), specificity and sensitivity were analysed, Younden index was used for the estimation of cut-off values. Results ROC – curves of all examined parameters are presented in fig. 1. The AUC of LAd, LAA, LAd/Ao and LAA/Ao were: 0,60; 0,73 0,72; 0,81 respectively. Differences in AUC were statistically significant (p<0,05). The best cut of value for LAA/Ao was 7,2cm2/cm , with sensitivity of 0,90 and specificity of 0,59. 112 of 490 (23%) patient without known paroxysmal AF exceeded the estimated cut-off point. Conclusions 1. In 1/3 of patients with AIS the assessment of recommended parameter of LA size (LAVI) can be difficult in clinical practice. 2. A simple, widely available parameter of left atrial size (LAA/Ao) has the best predictive value for the detection of patients with known paroxysmal AF. 3. Nearly ¼ of patients without known AF presenting with AIS have atriopathy which can be connected to not recognized arrhythmia. 4. LAA/Ao can be used as a simple, widely available parameter for the selection of patient for more profound prospective screening for AF in patients with AIS.