Abstract P300: C-reactive Protein and Racial Differences in Type 2 Diabetes Incidence: Reasons for Geographic and Racial Differences in Stroke (REGARDS)

Abstract

Introduction: Black Americans have a higher incidence of diabetes and have elevated inflammatory biomarkers compared to white Americans. Elevated inflammation is a risk factor for diabetes but the impact of inflammation on the racial disparity in diabetes is unknown. Hypothesis: Elevated C-reactive protein (CRP) attenuates the observed black-white difference in incident diabetes. Methods: REGARDS enrolled 30,239 black and white adults aged ≥45 years from the contiguous US in 2003-07. This analysis included REGARDS participants without baseline diabetes who were assessed for diabetes 9 years later. RRs for incident diabetes by race were calculated using modified Poisson regression adjusting for risk factors known to contribute to the racial difference in diabetes incidence. The attenuation by CRP of the black-white RR of incident diabetes was calculated as the percent difference in the race RR in models with and without CRP adjustment; 95% CI for the difference was estimated using bootstrapping. Results: Of 11,073 participants without baseline diabetes (33% black, 67% white), black participants had higher CRP than white participants, and 12.5% developed incident diabetes. The black-white RR for incident diabetes in the base model was 1.74 (95% CI: 1.52, 1.99) for women and 1.44 (1.25, 1.66) for men. Baseline CRP mediated 21% (14, 29%) of this association in women and 20% (12, 34%) in men. These percent attenuations were similar in models adjusting for other diabetes risk factors but were diminished in a fully adjusted model; 5% (-4, 25%) in women and 7% (-43, 50%) in men (Figure). Conclusion: Adjustment for CRP in base models accounted for 20% and 21% of the excess risk of incident diabetes observed in black men and women, respectively, in this study. This substantial mediation persisted after adjusting for other risk factors but was diminished in the fully adjusted model. This suggests a role of inflammation in the diabetogenic effects of risk factors contributing to the observed racial difference in diabetes incidence.