Right and left atrial echocardiographic parameters and outcome in patients with atrial fibrillation

Abstract Background Atrial remodeling in atrial fibrillation (AF) can be evaluated by measuring atrial dimensions, volumes, or function, in the left atrium (LA) but right atrial (RA) changes are often neglected. Purpose The aim of the analysis was to evaluate the power of LA and RA parameters in predicting a composite endpoint (CEP) of all-cause death, any thromboembolism, any acute coronary syndrome, and hospitalization for new or worsening heart failure. Methods From March 2016 to December 2021, patients with ECG-documented AF within 1 year were prospectively enrolled. We retrieved echo data by reviewing clips and collected clinical and laboratory data. Patients were followed up after 1 month and every 6 months thereafter unless clinically relevant events occurred. We considered left and right atrial antero-posterior diameter (iLAAPD, iRAAPD), left and right atrial volume (LAVI, RAVI), left and right atrial sphericity index (LASI, RASI), right and left atrial emptying fraction (RAEF, LAEF). Volumes and diameters were indexed by body surface area. Echo parameters were collected according to the 2015 recommendations of the American Society of Echocardiography. Results We enrolled 489 patients, 300 males (61.3%), median age 75 (IQR 66-80) years, median CHA2DS2VASc 3 (2-5), median HAS-BLED 1 (1-2). Four-hundred fifty-one (92.2%) patients were anticoagulated. Eighty-seven patients (17.8%) had a paroxysmal AF, 141 (28.8%) a persistent AF, 198 (40.5%) a permanent AF, and 63 (12.9%) a first-detected AF. After a median follow-up of 1114 (392-1384) days, 129 (26.3%) achieved the CEP. LAEF < 28% and iRAAPD ≥ 44 mm/m2 showed better sensitivity (respectively 72% and 73%) while iLAAPD and LAVI showed better negative predictive values (both 81%). At multivariable Cox regression analysis (adjusted for CHA2DS2VASc, AF type, and pattern of anticoagulation), building one model for every echocardiographic parameter, LA dimension, volume and function were associated with a higher risk of worse outcome (LAAPD > 22 mm/m2 HR 1.78; 95%CI 1.24-2.57; LAVI ≥ 42 ml/m2 HR 1.50 95%CI 1.03-2.19; LAEF < 28% HR 1.74 95%CI 1.21-2.48) while RASI, LASI, and RA parameters (except for iRAAPD ≥ 44 mm/m2: HR 1.65; 95%CI 1.10-2.47) were not. Comparing those models with a null one (i.e. variables used for adjustments) using the likelihood ratio test, the above-mentioned models demonstrated a significant improvement in predicting the risk of CEP (iLAAPD p=0.0018; LAVI p=0.0347; LAEF p=0.0028; iRAAPD p=0.0148). Visual analysis of multivariable restricted cubic splines regression graphs confirmed iLAAPD and LAVI as the only atrial parameters that refined risk prediction, with their lower limits of CI exceeding the 1 HR line at reference points (Figure). Conclusions In patients with AF, LAVI and iLAAPD allow better discrimination and risk prediction for the composite outcome, while sphericity indexes and most of the RA parameters do not provide adequate predictionFigure