The role of fasting glucose, insulin levels, and C-peptide in coronary heart disease (CHD) in non-diabetic individuals remains uncertain. We examined the association between fasting glucose, insulin and C-peptide with the long-term incidence of CHD in Japanese-American men. In 1980-1982, from a random sample of the Honolulu Heart Program men (n = 1378), aged 61-81 years, data on several CHD and metabolic risk factors were obtained to examine the relation of fasting glucose, insulin and C-peptide to 19-year CHD incidence. Age-adjusted incidence of CHD increased with increasing quintiles of glucose, insulin and C-peptide. Age-adjusted CHD rates in the glucose quintiles were 11.9, 11.6, 14.4, 18.1 and 24.1 per 1000 person-years (trend p < 0.001). In individual Cox models (lowest quintiles of glucose, insulin and C-peptide as reference) the relative risks (95% confidence interval) of CHD incidence for the glucose quintiles adjusting for age, smoking, hypertension, cholesterol, physical activity, and body mass index, were 0.9 (0.6-1.4), 1.2 (0.8-1.8), 1.4 (0.9-2.2), and 1.7 (1.1-2.6), respectively (trend p = 0.004). Insulin and C-peptide were not significantly associated with CHD on multivariate analysis. Fasting glucose remained the only significant predictor of increased CHD risk (p = 0.003) in a model combining all 3 metabolic variables. In this cohort, only fasting glucose independently predicts long-term incidence of CHD. Age-adjusted insulin and C-peptide levels were associated with CHD incidence, but after adjustment for other risk factors, do not independently predict CHD.
Keywords: coronary disease; diabetes mellitus; follow-up studies; glucose; insulin.