Metabolomic Association and Risk Prediction With Heart Failure in Older Adults
Guning Liu,Ngoc Quynh H Nguyen,Kari E Wong,Sunil K Agarwal,Eric Boerwinkle,Patricia P Chang,Brian L Claggett,Laura R Loehr,Jianzhong Ma,Kunihiro Matsushita,Carlos J Rodriguez,Joseph S Rossi,Stuart D Russell,R Brandon Stacey,Amil M Shah,Bing Yu,Ngoc Quynh H. Nguyen,Kari E. Wong,Sunil K. Agarwal,Patricia P. Chang,Brian L. Claggett,Laura R. Loehr,Carlos J. Rodriguez,Joseph S. Rossi,Stuart D. Russell,R. Brandon Stacey,Amil M. Shah
DOI: https://doi.org/10.1161/circheartfailure.123.010896
2024-03-03
Circulation Heart Failure
Abstract:Circulation: Heart Failure, Ahead of Print. BACKGROUND:Older adults have markedly increased risks of heart failure (HF), specifically HF with preserved ejection fraction (HFpEF). Identifying novel biomarkers can help in understanding HF pathogenesis and improve at-risk population identification. This study aimed to identify metabolites associated with incident HF, HFpEF, and HF with reduced ejection fraction and examine risk prediction in older adults.METHODS:Untargeted metabolomic profiling was performed in Black and White adults from the ARIC study (Atherosclerosis Risk in Communities) visit 5 (n=3719; mean age, 75 years). We applied Cox regressions to identify metabolites associated with incident HF and its subtypes. The metabolite risk score (MRS) was constructed and examined for associations with HF, echocardiographic measures, and HF risk prediction. Independent samples from visit 3 (n=1929; mean age, 58 years) were used for replication.RESULTS:Sixty metabolites (hazard ratios range, 0.79–1.49; false discovery rate, <0.05) were associated with incident HF after adjusting for clinical risk factors, eGFR, and NT-proBNP (N-terminal pro-B-type natriuretic peptide). Mannonate, a hydroxy acid, was replicated (hazard ratio, 1.36 [95% CI, 1.19–1.56]) with full adjustments. MRS was associated with an 80% increased risk of HF per SD increment, and the highest MRS quartile had 8.7× the risk of developing HFpEF than the lowest quartile. High MRS was also associated with unfavorable values of cardiac structure and function. Adding MRS over clinical risk factors and NT-proBNP improved 5-year HF risk prediction C statistics from 0.817 to 0.850 (∆C, 0.033 [95% CI, 0.017–0.047]). The association between MRS and incident HF was replicated after accounting for clinical risk factors (P<0.05).CONCLUSIONS:Novel metabolites associated with HF risk were identified, elucidating disease pathways, specifically HFpEF. An MRS was associated with HF risk and improved 5-year risk prediction in older adults, which may assist at at-risk population identification.
cardiac & cardiovascular systems