Abstract 11: Associations of Gluten Intake With Type 2 Diabetes Risk and Weight Gain in Three Large Prospective Cohort Studies of US Men and Women

Background: Gluten-free diets have grown in popularity, but evidence is lacking regarding gluten intake and long-term health. Methods: In Nurses’ Health Study (NHS, n=69,276), NHSII (n=88,610), and the Health Professionals Follow-Up Study (HPFS, n=41,908), we estimated gluten intake using a validated food-frequency questionnaire collected every 2-4 years. Incident T2D was defined as physician diagnosed diabetes and confirmed with supplementary information. Results: Gluten intake (mean ± standard deviation) was 5.83±2.23, 6.77±2.50, and 7.06±2.76 grams/day in NHS, NHSII, and HPFS, respectively, and strongly correlated with intakes of carbohydrate sources, especially refined grains, starch, and cereal fiber (Spearman correlation coefficients > 0.6). During 4.24 million years of follow-up, 15,947 T2D cases were confirmed. An inverse association between gluten intake and T2D risk was observed in all three cohorts after multivariate adjustment (table), and hazard ratio (HR, 95% confidence intervals [95%CI]) comparing extreme quintiles was 0.80(0.76, 0.84; P<0.001). The associations were slightly attenuated after further adjusting for cereal fiber (HR[95%CI]= 0.87[0.81, 0.93]), but not other carbohydrate components. Among participants without major chronic diseases and aged <65 years, changes in gluten intake were not significantly associated with weight gain in multivariate adjusted model: 4-year weight change (95%CI, lb) was 0.08(-0.06, 0.22; P=0.25) in NHS, -0.05(-0.18, 0.08; P=0.43) in NHSII, and 0.36(-0.24, 0.96; P=0.24) HPFS for each 5 grams increase in gluten intake. Conclusions: Our findings suggest that gluten intake may not exert significant adverse effects on the incidence of T2D or excess weight gain. Limiting gluten from diet is thus unlikely to facilitate T2D prevention and may lead to reduced consumption of cereal fiber or whole grains that help reduce diabetes risk.