Abstract
AIMS/HYPOTHESIS: Elevated fasting NEFAs are thought to promote type 2 diabetes. Three prospective studies support this concept, showing increased diabetes risk associated with fasting NEFA. However, these prospective associations may be confounded by strong cross-sectional correlations between fasting NEFA and metabolic predictors of diabetes. To examine this assumption, we used cohort data from the Insulin Resistance Atherosclerosis Study (IRAS). METHODS: Within the IRAS cohort (n = 902, 145 incident cases), we examined nine metabolic variables for their confounding effect on the fasting NEFA-diabetes association: 2 h glucose; fasting plasma glucose; body mass index; waist circumference; waist-to-hip ratio; weight; insulin sensitivity (S (I)); fasting insulin; and acute insulin response. We compared odds ratios for fasting NEFA (log( e ) transformed and adjusted for age, sex, ethnicity and clinic) before and after inclusion of each metabolic variable into a logistic regression model. RESULTS: Three variables (2 h glucose, BMI and S (I)) cross-sectionally correlated with fasting NEFA (r > or = 0.1, p < 0.05). Unadjusted for metabolic predictors, fasting NEFA levels were positively associated with diabetes risk: OR 1.37 (95% CI 0.87-2.15) per unit on a log scale. All metabolic variables except AIR showed confounding. Inclusion of 2 h glucose reversed the positive association (OR 0.50 [95% CI 0.30-0.82]), whereas other predictors reduced the association to the null. The final model included the variables correlated with baseline fasting NEFA (2 h glucose, BMI and S (I)) and the demographic variables resulting in OR 0.47 (95% CI 0.27-0.81). CONCLUSIONS/INTERPRETATION: Our results indicate that 2 h glucose strongly confounds the prospective association between fasting NEFA and diabetes; carefully adjusted fasting NEFA levels are inversely associated with diabetes risk.