An altered plasma lipidome-phenome network characterizes heart failure with preserved ejection fraction.

Abstract

Heart failure with preserved ejection fraction (HFpEF) is a multifactorial, multisystemic syndrome that involves alterations in lipid metabolism. This study aimed to test whether distinct plasma lipid profiles or lipid entities or both are associated with clinical and functional echocardiographic parameters in HFpEF. We examined the human plasma lipidome in HFpEF patients (n=18) with left ventricular ejection fraction ≥50% and N-terminal pro-brain natriuretic peptide (NT-proBNP) >125pg/mL and control subjects (n=12) using mass spectrometry-based shotgun lipidomics. The cohort included 8 women and 22 men with average age of 67.8±8.6 SD. The control and disease groups were not significantly different with respect to age, body mass index, systolic and diastolic blood pressure, and waist-to-hip ratio. The disease group experienced more fatigue (P<0.001), had more often coronary artery disease (P=0.04), and received more medications (beta-blockers, P<0.001). The disease group had significantly different levels of HFpEF-relevant parameters, including NT-proBNP (P<0.001), left ventricular mass index (P=0.005), left atrial volume index (P=0.001), and left ventricular filling index (P<0.001), and lower left ventricular end-diastolic diameter (P=0.014), with no difference in left ventricular ejection fraction. Significant differences in lipid profiles between HFpEF patients and controls could not be detected, including no significant differences in abundance of circulating lipids binned by carbon chain length or by double bonds, nor at the level of individual lipid species. However, there was a striking correlation between selected lipids with smoking status that was independent of disease status, as well as between specific lipids and hyperlipidaemia [with corresponding significance of either false discovery rate (FDR) <0.1 or FDR<0.01]. In an exploratory network analysis of correlations, we observed significantly stronger correlations within the HFpEF group between individual lipids from the cholesterol ester and phosphatidylcholine (PC) classes and clinical/echocardiographic parameters such as left atrial volume index, left ventricular end-diastolic diameters, and heart rate (FDR<0.1). In contrast, the control group showed significantly stronger negative correlations (FDR<0.1) between individual species from the PC and sphingomyelin classes and left ventricular mass index or systolic blood pressure. We did not find significant direct associations between plasma lipidomic parameters and HFpEF and therefore could not conclude that any specific lipids are biomarkers of HFpEF. The validation in larger cohort is needed to confidently conclude the absence of first-order associations.