The role of left atrium function parameters in predicting low-voltage areas in the left atrium in patients with atrial fibrillation

Abstract Background Low-voltage areas (LVA) assessed in the left atrium (LA) electroanatomical map are potent predictors of the efficacy of atrial fibrillation (AF) ablation. Clinical factors predicting the presence of LVA have moderate accuracy. The aim of this analysis was to assess the role of LA functional echocardiographic parameters (i.e., LA strains) in predicting the presence of LVA in patients undergoing AF ablation. Methods Consecutive, unselected patients undergoing AF ablation were included in this retrospective analysis. All patients underwent transthoracic (including LA strains) and transesophageal echocardiography, pulmonary vein isolation, and LA electroanatomical mapping using a multipolar catheter. LVA was defined as an area greater than 2 cm2 with a voltage below 0.5 mV (or 0.3 mV when measured during AF). A logistic regression LVA predictive model based on clinical parameters and echocardiographic morphology parameters was built. Receiver Operating Characteristic (ROC) analysis and area under the curve (AUC) calculations of LA functional parameters and logistic regression models were performed to identify the best predictors. Results A total of 793 patients were available for analysis; 47 (6%) were excluded due to poor echocardiographic image quality or significant tachycardia during atrial fibrillation (130/min), and in 61 patients information about LVA was not available. Finally, 685 patients were included in the analysis, of whom 157 (22.9%) had LVA (LVA (+) group). To test parameters specific to sinus rhythm (conduit and contraction strains), the patient population was divided according to the rhythm during echocardiography testing into SR patients (n = 369) and AF patients (n = 316). In the SR group, the best parameter for predicting the presence of LVA was LA contraction strain (LASct, AUC 0.862, 95% Confidence Interval (CI): 0.816–0.908); in AF patients, it was LA reservoir strain (LASr, AUC 0.711, 95% CI: 0.653–0.769). In SR patients, adding LASct to the clinical logistic regression predictive model markedly improved its predictive value (AUC from 0.86, 95% CI: 0.79–0.92 to 0.90, 95% CI: 0.85–0.95). In AF patients, adding LASr to the clinical model only moderately improved its performance (AUC from 0.84, 95% CI: 0.79–0.88 to 0.85, 95% CI: 0.80–0.89). Conclusions In patients undergoing AF ablation, LASct had the best predictive value among LA strains and added significant independent information to the predictive model for the presence of LVA (SR patients). In patients tested during AF, LASr had the best predictive value for LVA, but when added to the clinical model, it only moderately improved its performance.