Prog Cardiovasc Dis. 2026 Sep 30:S0033-0620(26)00109-X. doi: 10.1016/j.pcad.2026.09.010. Online ahead of print.
ABSTRACT
BACKGROUND: Atrial fibrillation (AF) reflects systemic cardio-metabolic dysfunction. While myocardial stress and injury biomarkers are associated with incident AF, the role of ketone bodies (KB) remains incompletely defined.
OBJECTIVES: To evaluate a multi-biomarker risk score integrating KB with NT-proBNP and hs-cTnT for incident AF prediction and its incremental value beyond established clinical models.
METHODS: We studied 6429 MESA participants free of AF and clinical CVD at baseline. A multi-biomarker risk score using total KB >252 μmol/L, NT-proBNP ≥125 pg/mL, and hs-cTnT ≥14 ng/L was categorized as Very Low Risk Score = 0, Low Risk Score = 1-3, Intermediate Risk Score = 4-5, and High Risk Score > 5. Associations with incident AF were evaluated using Cox proportional hazards and Fine-Gray competing risk models.
RESULTS: Over 16.4 years, 1269 participants developed AF before death. Higher risk score categories were associated with a strong, graded increase in AF risk independent of demographic, socioeconomic, and clinical covariates. High Risk Score > 5 had nearly threefold greater AF risk versus Very Low Risk Score = 0 (HR 2.72; 95% CI, 2.11-3.49; P < 0.001). At 5 years, AUC was 0.768 for CHARGE-AF, 0.784 after adding NT-proBNP, and 0.787 after adding hs-cTnT. Adding KB after NT-proBNP and hs-cTnT did not materially change the AUC; 10-year findings were similar. Decision curves showed no clear additional net benefit.
CONCLUSIONS: A multi-biomarker risk score integrating KB with NT-proBNP and hs-cTnT was strongly associated with incident AF and provided modest incremental predictive value beyond CHARGE-AF. External validation is warranted before clinical application.
PMID:42815840 | DOI:10.1016/j.pcad.2026.09.010