Incident heart failure in chronic kidney disease: proteomics informs biology and risk stratification

Abstract

Background and aims: Incident heart failure (HF) among individuals with chronic kidney disease (CKD) incurs hospitalizations that burden patients and health care systems. There are few preventative therapies, and the Pooled Cohort equations to Prevent Heart Failure (PCP-HF) perform poorly in the setting of CKD. New drug targets and better risk stratification are urgently needed.

Methods: In this analysis of incident HF, SomaScan V4.0 (4638 proteins) was analysed in 2906 participants of the Chronic Renal Insufficiency Cohort (CRIC) with validation in the Atherosclerosis Risk in Communities (ARIC) study. The primary outcome was 14-year incident HF (390 events); secondary outcomes included 4-year HF (183 events), HF with reduced ejection fraction (137 events), and HF with preserved ejection fraction (165 events). Mendelian randomization and Gene Ontology were applied to examine causality and pathways. The performance of novel multi-protein risk models was compared to the PCP-HF risk score.

Results: Over 200 proteins were associated with incident HF after adjustment for estimated glomerular filtration rate at P < 1 × 10-5. After adjustment for covariates including N-terminal pro-B-type natriuretic peptide, 17 proteins remained associated at P < 1 × 10-5. Mendelian randomization associations were found for six proteins, of which four are druggable targets: FCG2B, IGFBP3, CAH6, and ASGR1. For the primary outcome, the C-statistic (95% confidence interval [CI]) for the 48-protein model in CRIC was 0.790 (0.735, 0.844) vs. 0.703 (0.644, 0.762) for the PCP-HF model (P = .001). C-statistic (95% CI) for the protein model in ARIC was 0.747 (0.707, 0.787).

Conclusions: Large-scale proteomics reveal novel circulating protein biomarkers and potential mediators of HF in CKD. Proteomic risk models improve upon the PCP-HF risk score in this population.

Keywords: Chronic kidney disease; Heart failure; Mendelian randomization; Risk model.

Structured Graphical Abstract

Large-scale proteomics analysis of heart failure in CKD cohorts. CKD, chronic kidney disease; PCP-HF, Pooled Cohort equations to Prevent Heart Failure; CRIC, Chronic Renal Insufficiency Cohort; ARIC, Atherosclerosis Risk in Communities; HERMES, Heart Failure Molecular Epidemiology for Therapeutic Targets Consortium; MR, Mendelian randomization; eGFR, estimated glomerular filtration rate; HF, heart failure; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; pQTL, protein quantitative trait loci.

Figure 1

Individual proteins associated with heart failure after adjustment for eGFR in CRIC volcano plots of the associations of 4638 proteins with HF and HF subtypes, adjusted for eGFR. Extreme HR values are HR > 2 or <0.5

Figure 2

Mendelian randomization. Mendelian randomization (MR) in the HERMES genome-wide association study suggested causal associations for 6 proteins significant at P-value < .05 (corresponding to FDR < 0.2). pQTLs were identified in deCODE and the instrumental variable for each protein is shown above, along with the OR (95% CI) for MR. Three proteins with (*) replicated in ARIC at P < .05. Four proteins with (†) are potentially druggable targets (Therapeutic Target Database, at http://db.idrblab.net/)

Figure 3

Clinical, protein, and hybrid risk models for incident heart failure in CRIC discrimination for risk models are shown in the 20% testing set (479 participants) and calibration in the 80% training set. In the testing set, there were 31 events for 4-year HF, 63 for 14-year HF, 23 for HFrEF, and 22 for HFpEF. Calibration is performed for quintiles of predicted risk, except for those marked with asterisk, which were analysed in quartiles. C-statistics are compared by t-test using a one-sided P-value. C-statistics for 14-year and 4-year HF PCP and protein models in gender- and race-specific subgroups are listed in Supplementary data online, Table S11

Figure 4

Time-dependent area under the curve. Time-dependent AUCs are shown for each risk model in the 20% testing set in CRIC (n = 479 participants)