Аtrial cardiomyopathy as a determinant of stroke risk in heart failure with preserved ejection fraction and non-permanent atrial fibrillation

Background Heart failure with preserved ejection fraction (HFpEF) is a systemic inflammatory–fibrotic syndrome characterized by myocardial stiffness and progressive atrial remodeling. In this setting, atrial fibrillation (AF) often reflects advanced atrial cardiomyopathy (ACM) rather than isolated electrical instability. Structural remodeling, mechanical dysfunction and pro-thrombotic activation of the left atrium (LA) may represent a key mechanism linking HFpEF to cerebrovascular events beyond conventional clinical risk scores. Aim To determine whether biomarkers of inflammation and fibrosis together with echocardiographic markers of ACM are independently associated with transient ischemic attack (TIA) or stroke in HFpEF patients with paroxysmal/persistent AF. Methods This multicenter retrospective cohort study included 345 HFpEF patients with non-valvular paroxysmal/persistent AF from four tertiary cardiovascular centers, followed for 3 years. Inflammatory and fibrotic biomarkers - tissue factor (TF), high-sensitivity C-reactive protein (hsCRP), interleukin-6 (IL-6), transforming growth factor beta 1 (TGF-β1) were measured using ELISA. Transthoracic echocardiography assessed ACM using peak atrial longitudinal strain (PALS), indexed minimal and maximal LA volumes (LAVImin, LAVImax),left ventricular global longitudinal strain (LV GLS) and average E/e′ ratio. LA stiffness index was calculated as E/e′ divided by PALS. The primary endpoint was TIA or ischemic stroke. Multivariable logistic regression was adjusted for age, sex, CHA2DS2-VASc score, anticoagulation therapy, eGFR and NT-proBNP. Results TIA/stroke occurred in 94 patients (27.2%). Patients with events had higher median NT-proBNP levels (860 [520–1450] vs 520 [310–880] pg/ml, p<0.001) and elevated TF (1350±65 vs 850±33 pg/ml, p<0.001), hsCRP (8.6±1.4 vs 4.2±1.6 mg/L, p=0.002) and TGF-β1 (1240±75 vs 900±43 pg/ml, p=0.02). Echocardiography revealed reduced PALS (20.4±4.2% vs 28.6±5.1%, p<0.001), increased LAVImin (28 ± 6.0 vs 18 ± 5.0 ml/m², p<0.001), LAVImax did not differ between groups (39.4±6.6 vs 37.1±6.2 ml/m², p=0.28) and higher E/e′ (16.2±3.4 vs 11.8±2.9, p<0.001). PALS <25% was associated with approximately two-fold higher odds of TIA/stroke. In multivariable analysis, reduced PALS (OR 3.48, p<0.001), elevated TF (OR 3.14, p=0.02), hsCRP (OR 1.83, p<0.002) remained independently associated with TIA/stroke after adjustment for clinical covariates and NT-proBNP. Conclusions In HFpEF patients with non-permanent AF, ACM—characterized by mechanical dysfunction and inflammatory–prothrombotic activation—emerges as a central pathophysiological substrate linking myocardial stiffness to cerebrovascular events. These findings suggest that stroke risk in this population reflects progression of atrial disease rather than arrhythmia alone and that LA mechanical assessment may refine stroke risk stratification beyond conventional clinical models.