1118-P: Prediction of Clamp-Derived Insulin Sensitivity in African Americans from Clinical Risk Factors

Surrogate indices of insulin sensitivity based on fasting glucose and insulin values fail to adequately predict insulin sensitivity in African-Americans (AA) , a group with high risk for type 2 diabetes. The objective of this study was to evaluate the individual and combined utility of clinically accessible risk factors (demographic, anthropometric, and metabolic traits) to predict clamp-derived insulin sensitivity in 53 AA adults without diabetes. Insulin sensitivity was measured with the euglycemic clamp (SI-clamp) at an insulin dose of 120 µU/m2/min/kg lean body mass to assess primarily skeletal muscle glucose disposal. Concentrations of lipid/lipoproteins, adipokines, glucose, insulin, C-peptide, proinsulin, and inflammatory markers were obtained from fasting sera. Individual associations between SI-clamp and predictors were evaluated with linear regression, adjusting for age and sex. Multiple regression analyses with backward stepwise selection of all candidate predictors were used to derive prediction models. Body mass, triglycerides (TAG) , glucose, insulin, C-peptide, proinsulin, and TNF-α were inversely associated with SI-clamp, whereas female sex, HDL-cholesterol, adiponectin, and IFN-γ were positively associated (P< 0.05) . For the prediction models, HDL-cholesterol, TAG, C-peptide, adiponectin, leptin, and IL-6 explained the highest proportion of variance in SI-clamp (AIC = 15.8 and adjusted R2 = 54.27%) . A second, more clinically applicable model was also derived using HDL-cholesterol, TAG, C-peptide, and adiponectin (AIC = 18.136 and adjusted R2 = 50.63%) . These results suggest that in AA, fasting levels of lipid/lipoproteins, C-peptide, and adiponectin are stronger predictors of insulin sensitivity than fasting glucose and insulin. Development of a surrogate index for insulin sensitivity with these predictors may aid in accurately identifying AA at increased risk for type 2 diabetes in clinical settings and epidemiological studies.