Left atrial contraction strain as a predictor of device-detected atrial fibrillation in patients with heart failure with reduced ejection fraction

Background Atrial fibrillation (AF) is the most common atrial arrhythmia in patients with heart failure (HF) and is closely linked to structural and functional remodelling of the left atrium (LA). Left atrial strain (LAS) parameters are sensitive markers of early LA dysfunction, and their impairment has been shown to predict AF in several clinical settings. Identifying patients at increased AF risk may improve prevention and management, particularly in high-risk groups such as those with HF. Purpose In this study, we aimed to evaluate whether LAS measures can predict new-onset AF (NOAF) in patients with heart failure with reduced ejection fraction (HFrEF) in sinus rhythm undergoing implantable cardioverter-defibrillator (ICD) or cardiac resynchronization therapy (CRT) implantation. Methods In this prospective single-centre observational study, consecutive patients with symptomatic HFrEF (LVEF ≤40%, regardless of aetiology), sinus rhythm, and no prior AF were enrolled at referral for an ICD/CRT implantation. All patients underwent standardized transthoracic echocardiography, including LAS assessment by two-dimensional speckle-tracking, as well as laboratory testing and ECG at baseline and at 12 months after implantation. The primary outcome was NOAF within 12 months, detected by ICD/CRT interrogation and confirmed by intracardiac electrogram analysis (atrial rate >200/min, episode duration >30 s). Results A total of 88 patients were analysed, and NOAF occurred in 15 (17%). Mean age was 65 ± 12 years, median BMI 26.6 (IQR 24–29), and 83% were men (n=73). Ischaemic aetiology was present in 64.8% (n=57), and 60.2% received CRT (n=53). Patients who developed NOAF had lower LA reservoir (LASr) and contraction strain (LASct), lower LA emptying fraction (LAEF), and higher LA stiffness index (LASI) (all p<0.001). Other between-group differences are shown in Table 1. In univariable logistic regression, lower LASr (OR 0.802), lower LASct (OR 0.768), and higher LASI (OR 2.218) were significantly associated with NOAF (all p<0.001). Other univariable predictors included LAEF, E/e′avg, E-wave velocity, and E/A ratio, whereas clinical variables (age, sex, ischaemic aetiology), conduction abnormalities, and LV parameters were not significant. LASct showed the best discrimination (AUC 0.829, 95% CI 0.721–0.938), at a Youden-derived cut-off of 8%, sensitivity was 79.5% and specificity 73%. In multivariable analysis, several two-parameter models reached an AUC of 0.831, but improvement over LASct alone was minimal and not clinically meaningful (Table 2). Conclusions LASct was the strongest predictor of NOAF and showed the highest standalone diagnostic accuracy. Adding other echocardiographic parameters provided only marginal incremental value beyond LASct alone. Incorporating LAS assessment may improve atrial arrhythmia risk stratification in patients with HFrEF undergoing device therapy.