Proteins associated with rehospitalization, mortality and diuretic resistance in acutely decompensated heart failure

Most proteomic analyses in patients with heart failure (HF) focus on the stable ambulatory setting; pathways that drive acute decompensated heart failure (ADHF) are largely uncharacterized. We aimed to examine the associations of cardiometabolic proteins with HF rehospitalization, mortality and diuretic response in patients with ADHF.

The Mechanisms of Diuretic Resistance Cohort comprised patients hospitalized for ADHF at Yale New Haven Hospital from 2015 to 2020. We quantified 369 serum proteins and evaluated their associations with rehospitalization, mortality or diuretic response, using individual and combined biomarkers. Models adjusted for sociodemographic/lifestyle factors, comorbidities and clinical factors. For individual biomarker models, significant proteins were those with a false-discovery rate q-value (FDRq) <1%, while combined biomarker models used a penalized cross-validated log-likelihood method to reduce false positives.

Of 401 patients enrolled, median age was 65 years; 36% were female. Six (C1QTNF1, CTSD, FCN2, SERPINA12, TFRC and TNFRSF10C) were associated with increased risk of HF rehospitalization [hazard ratio (HR) ranging from 1.2 to 1.35 per standard deviation (SD) protein concentration] while one (SDC1) was associated with decreased risk (HR = 0.77 per SD protein concentration). Three (CDH1, FABP6 and TNC) were associated with mortality (HR ranging from 1.35 to 1.64 per SD protein concentration). Three (MMP7, PGLYRP1 and REN) were associated with reduced diuretic response (% Estimate ranging from -12.2% to -18.6% per SD protein concentration). Among all proteins, those associated with lower diuretic response correlated with greater mortality risk [Spearman r = -0.63 (95% confidence interval: -0.56, -0.69)].

Several proteins with known associations with cardiometabolic disease were associated with poor outcomes or reduced diuretic response. They have known roles in inflammatory, metabolic and cardiac or kidney remodelling pathways and may point to novel pathways in ADHF that warrant investigation as potential prognostic tools and targets for intervention.