Comparative prognostic value of nine cardiorenal biomarkers for mortality among adults with prediabetes and diabetes

BACKGROUND: Cardiorenal biomarkers are increasingly recognized as valuable tools for risk stratification in individuals with dysglycemia. However, their comparative prognostic value for long-term all-cause and cardio-cerebrovascular disease (CCD) mortality remains unclear. METHODS: We analyzed data from 4087 adults with prediabetes or diabetes from NHANES 1999-2004, with mortality follow-up through 2019. Nine biomarkers were assessed: NT-proBNP, hs-cTnT, hs-cTnI, CRP, UACR, BUN/Cr ratio, uric acid, cystatin C, and β2-microglobulin. Survey-weighted Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for all-cause and CCD mortality. Time-dependent receiver operating characteristic (ROC) analysis evaluated discriminative performance. Weighted quantile sum (WQS) regression was applied to assess the collective impact of biomarkers. RESULTS: During a median follow-up of 16.5 years, 1855 all-cause deaths and 630 CCD deaths were recorded. In fully adjusted models, NT-proBNP (HR = 2.19; 95% CI, 1.73-2.78), hs-cTnT (HR = 2.30; 1.63-3.24), hs-cTnI (HR = 1.80; 1.44-2.24), cystatin C (HR = 1.76; 1.36-2.28), β2-microglobulin (HR = 1.72; 1.36-2.16), and UACR (HR = 1.50; 1.23-1.81) were significantly associated with all-cause mortality, with similar findings for CCD mortality. NT-proBNP and hs-cTnT demonstrated the strongest discrimination for 10-year all-cause (AUCs: 0.80 and 0.81) and CCD mortality (AUCs: both 0.85). Adding NT-proBNP or hs-cTnT to the base model significantly improved predictive accuracy. WQS regression confirmed a significant positive association with all-cause (HR = 1.56; 95% CI, 1.32-1.84) and CCD mortality (HR = 1.39; 95% CI, 1.14-1.70), with NT-proBNP, hs-cTnT, and cystatin C contributing most strongly. CONCLUSIONS: Among nine evaluated cardiorenal biomarkers, NT-proBNP, hs-cTnT, and cystatin C demonstrated the strongest and most consistent associations with all-cause and CCD mortality among adults with prediabetes and diabetes, as well as the highest incremental predictive value. These biomarkers may serve as key components in future risk stratification models for individuals with prediabetes or diabetes.