Left atrial function and the risk of new-onset atrial fibrillation in cardiac amyloidosis

Abstract Background Cardiac amyloidosis (CA) is an infiltrative disease characterized by the accumulation of misfolded proteins into the extracellular matrix of the myocardium. Ventricular but also atrial myopathy are described in CA, with higher risk of atrial fibrillation (AF) and cardioembolic events even during sinus rhythm. Purpose We aimed to investigate which parameters of left atrial (LA) structure and function could predict new-onset AF (NOAF) in patients with CA, aiding in improved follow-up. Methods We prospectively included patients diagnosed with CA, both light chain (AL) and variant transthyretin (ATTRv) with no history of AF, from 2 tertiary European centers. Cardiac involvement was assessed according to current European recommendations. Comprehensive echocardiographic studies were performed at baseline with measurement of both classical structure and function parameters parameters derived from speckle tracking echocardiography measured with a vendor-dependent analysis software and the primary outcome was NOAF. Results Of the 179 patients diagnosed with CA, 86 were excluded due to history of AF, unacceptable image quality and non-AL or non-ATTRv diagnoses. Finally, 93 patients with CA were included (mean age 54.3±9.8, 58% males), and 44 patients developed NOAF during a median follow up of 11 (32.5) months. Patients with NOAF had heavier hearts (LVMi 155.5±39.6 vs. 136.6±44.3 g/m2, p= 0.03), worse global LV function (LVEF 46.7±10.9 vs. 53.8±11.2 %, p= 0.003), more LV longitudinal impairment (GLS -10.0±4.7 vs. -13.4±3.7 %, p= 0.0002), larger atria (LAVi 44.6±14.3 vs. 35.4±13.6 mL/m2, p= 0.002) and worse LA function (LA emptying fraction 28.5±15.5 vs. 44.9±14.6 %, p< 0.001; LA peak longitudinal strain (PALS) 13.6±11.3 vs. 20.5±10.4 %, p= 0.003). In the multivariate analysis adjusting for confounding factors, including LA diameter, LAVi and other LA function parameters, lower PALS was independently associated with higher risk for NOAF (HR: 0.874, 95% CI: 0.784-0.974, p= 0.014). Using the receiver operating characteristics curve analysis, we identified the optimal cut-off of PALS for predicting NOAF as 14.3% (Sn 71%, Sp 73%, AUC 0.75). Furthermore, PALS ≤14.3% was an independent predictor for NOAF (HR: 5.259, 95% CI: 2.653–10.423, P< 0.001). The Kaplan–Meier curves showed that patients with PALS ≤14.3% had a higher risk for NOAF than those with PALS >14.3% (P<0.001 by log-rank test, Figure 1). Conclusions PALS is an independent predictor for NOAF in patients with CA.Figure 1.Kaplan-Meier curves for NOAF