Introduction: An adverse cardiovascular disease (CVD) risk factor profile is typically present well before the clinical diagnosis of diabetes. Diabetes and hyperglycemia are risk factors for myocardial damage and subclinical atherosclerosis, but it is unclear if subclinical cardiac damage is present before the clinical diagnosis of diabetes. Hypothesis: Subclinical myocardial damage_as indicated by elevated cardiac troponin measured by a novel highly sensitive assay (hs-cTnT) will be associated with risk of diabetes. Methods: We conducted a prospective cohort analysis of hs-cTnT and incidence of self-reported diabetes among ARIC study participants with no known history of diabetes or CVD at baseline (1996-98). Adjusted Cox proportional hazard models included hs-cTnT as a categorical variable or as a spline. Results: During a median of 12 years of follow-up of 7,898 participants, there were 1,809 self-reported cases of newly diagnosed diabetes. Compared to those with undetectable hs-cTnT (<0.003ug/L), participants with hs-cTnT 0.009-0.0013 or ≥0.014ug/L had a significantly increased risk for diabetes after adjustment for diabetes risk factors: HRs of 1.18, 95% CI 1.0-1.40 and 1.32, 95% CI 1.04 - 1.69, respectively. After further adjustment for glucose, the relationship was attenuated (p-values for linear trend 0.01 before and 0.19 after adjustment) but was significant when modeled non-linearly (p-value=0.003 for spline with knot at the mean hs-cTnT of 0.005 ug/L) (Figure). Conclusions: Elevated levels of hs-cTnT are associated with an increased risk of developing diabetes even in the absence of a history of CVD and after adjusting for diabetes risk factors but the association is non-linear after adjustment for fasting glucose. This suggests that diabetes and heart disease may develop in parallel with myocyte necrosis preceding as well as following diabetes.