Background:
The usefulness of biomarkers from different biologic pathways for predicting cardiovascular disease (CVD) events among African Americans is not well understood.
Methods:
We evaluated prospectively 3,102 Jackson Heart Study participants (mean age 54 years; 64% women) with data on a panel of 9 biomarkers representing inflammation (high sensitivity C - reactive protein), adiposity (adiponectin, leptin), neurohormonal activation (B-type natriuretic peptide [BNP], aldosterone, and cortisol); insulin resistance (HOMA-IR); and endothelial function (endothelin and homocysteine). We used Cox proportional hazard regression to relate the biomarker panel to the incidence of CVD (stroke, coronary heart disease, angina, heart failure and intermittent claudication) adjusting for standard CVD risk factors.
Results:
On follow-up (median 8.2 years), 224 participants (141 women) experienced a first CVD event, and 238 (140 women) died. Circulating concentrations of aldosterone, BNP and HOMA-IR were associated with CVD (multivariable-adjusted hazard ratios [HR] and 95% confidence interval [CI] per standard deviation (SD) increase in log-biomarker) were, respectively 1.15, (95% CI 1.01-1.30, p=0.016), 1.97, (95% CI 1.22-2.41, p<0.0001), and 1.30, (95% CI 1.10-1.52, p=0.0064). Blood cortisol and homocysteine were associated with death (HR per SD increment log-biomarker, respectively, 1.17, (95% CI 1.01-1.35, p=0.042), and 1.24, (95% CI 1.10-1.40, pvalue=0.0005). Biomarkers improved risk reclassification by 0.135; 0.120 of which was gained in classification of participants that experienced CVD events and 0.015 from participants that did not. Also, biomarkers marginally increased the model c-statistic beyond traditional risk factors.
Conclusions:
In our community-based sample of African Americans, circulating aldosterone, BNP and HOMA-IR predicted CVD risk, whereas serum cortisol and homocysteine predicted death. However, the incremental yield of biomarkers over traditional risk factors for risk prediction was minimal.