Time to solve the puzzle of heart failure with preserved ejection fraction by biomarkers, echocardiography and cardiac magnetic resonance

Abstract

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – National budget only. Main funding source(s): HEART PRESERVED onbehalf HEART PRESERVED Background Heart failure with preserved ejection fraction (HFpEF) is now recognized as a major and growing public health problem worldwide. Clinical trials investigating different treatment strategies had disappointing results. Several biomarkers of inflammation, endothelial dysfunction, and myocardial fibrosis appeared to be promising in understanding HFpEF pathophysiology. Methods. We enrolled prospectively 94 patients with HFpEF in sinus rhythm (according to 2019 scoring system) (67 ± 9 yrs, 33 men). We evaluated them by 2D and speckle tracking echocardiography (STE). 80 patients had also a cardiac magnetic resonance (CMR) 1.5T evaluation. We measured LV ejection fraction (LVEF), mean E’ (E’m), E/E’ ratio, sPAP, left atrial volume indexed (LAVi), and global longitudinal stain by STE (GLS). By CMR we evaluated LVEFcmr, LV mass, T1 mapping with mean extracellular volume (ECVm), and pre-gadolinium mean times quantification (preGDT1m) as markers of myocardial fibrosis. All patients had NTproBNP and biomarkers for systemic inflammation (IL6, cystatin C, pentraxin-3, GDF15), endothelial dysfunction: soluble E -selectin, VCAM, von Willebrand factor (vWf), and myocardial fibrosis: Galectin-3. Results. LVEF was 60.5 ± 6 % and LVEFcmr 61 ± 6.6%. All parameters from the scoring system were as we expected: E’m = 7.6 ± 1.8 cm/s, E/E’ ratio = 11 ± 3.4, sPAP = 34 ± 8 mmHg, LAVi = 47 ± 11 ml/m2, GLS=-18.3 ± 2.9, and NTproBNP of 282 ± 294 pg/ml. NTproBNP significantly correlated with sPAP, LAVi, preGDT1m, ECVm, galectin-3, GDF15, and pentraxin-3 (all r > 0.4, p < 0.05). The best predictor for NTproBNP level was GDF15 (r = 0.4, r2 = 0.25, p = 0.001). LAVi significantly corelated with E/E’ ratio, sPAP, NTproBNP, galectin-3 (r > 0.4, p < 0.05). GLS correlated with LVEFcmr, LV mass, ECVm, preGDT1m, LAVi, E/E’ ratio, NTproBNP, GDF15, vWf, Eselectin, VCAM (all r = 0.4, p < 0.05). The best predictor model for GLS was LV mass, NTproBNP, E-selectine, and vWf (r = 0.67, r2 = 0.45, P < 0.001). sPAP was best predicted by a model composed by IL6, VCAM, LAVi (r = 0.5, r2 = 0.25, p < 0.001). E’m significantly correlated with vWf, GHD15, VCAM, LV mass, and preGDT1 (all r > 0.4, p < 0.05), but the best predictor model included only LV mass and GDF15 (r = 0.57, r2 = 0.32, P < 0.001). Galectin-3 significantly correlated with LAVi, preGDT1m, and NTproBNP, but the only predictor for galectin-3 level was preGDT1 (r = 0.4, r2 = 0.2, p = 0.007). Conclusions In HFpEF NTproBNP is significantly correlated with markers of inflammatory status, endothelial dysfunction, and fibrosis, but the level is mainly determined by inflammation (GDF15). Diastolic dysfunction parameters are mainly correlated with inflammatory and endothelial dysfunction biomarkers . Only LAVi was correlated with myocardial fibrosis. Sub-clinical systolic dysfunction is mainly determined by proinflamatory status and endothelial dysfunction, but not by fibrosis.