|| Checking for direct PDF access through Ovid
Introduction: Standard lipid panel measures (e.g., low HDL cholesterol and high triglyceride levels) are associated with developing diabetes. Lipoprotein and metabolomics profiles measured by nuclear magnetic resonance (NMR) carry additional information and may add to knowledge of future diabetes risk.Hypothesis: We assessed the hypothesis that two multimarkers of insulin resistance derived from NMR LipoProfile® analysis, LP-IR (lipoprotein insulin resistance index) and IRDRF (insulin resistance diabetes risk factor index), are associated with incident diabetes independent of traditional risk factors in non-diabetic, apparently healthy subjects.Methods: LP-IR and IRDRF were measured at LipoScience/LabCorp for 3,564 non-diabetic participants (mean age 32±4 years, 46% [n=1,646] male) in the Coronary Artery Risk Development in Young Adults (CARDIA) study. LP-IR was calculated from 6 lipoprotein subclass and size parameters. IRDRF was calculated by combining LP-IR with levels of branched-chain amino acids, the inflammatory marker GlycA, and 2 additional lipoprotein subclasses. Diabetes status was ascertained by fasting glucose levels, hemoglobin A1c, and diabetes medications.We used Cox proportional hazards to model the relationship between LP-IR and IRDRF with incident diabetes. Participants were followed for a mean time of 21±5 years. We considered 4 models that examined LP-IR and IRDRF both as continuous and categorical variables and adjusted for various traditional predictors of diabetes.Results: Model 3, which adjusted for components of metabolic syndrome, had the best model fit in all cases (Table 1). LP-IR and IRDRF had a significant linear relationship with incident diabetes. Participants who had LP-IR and IRDRF scores in Q4 were significantly more likely to develop diabetes than those with scores in Q1.Conclusions: In conclusion, LP-IR and IRDRF can be derived from a clinically-available NMR LipoProfile test, and both were associated with incident diabetes independent of traditional predictors.