Abstract MP21: Circulating Biomarkers of Dairy Fat and Incident Type 2 Diabetes in Two US Prospective Cohorts

Background: Prior observational studies suggest that self-reported consumption of dairy foods is associated with lower risk of type 2 diabetes (DM). Few studies have used circulating biomarkers that provide objective measures of dairy fat consumption. Aim: To test the hypothesis that plasma fatty acid biomarkers of dairy fat, 15:0, 17:0, 16:1 n-7t and 14:0, are associated with lower incidence of DM. Methods: We used gas-liquid chromatography to measure plasma 15:0, 17:0, 16:1 n-7t and 14:0 biomarkers in 3,347 adults aged 30-75 years and free of prevalent DM at baseline in two separate U.S. prospective cohorts, Nurses’ Health Study (NHS) and Health Professionals’ Follow Up Study (HPFS). Incident DM was identified through 2008 and confirmed by validated supplementary questionnaire using symptoms, diagnostic tests, and medical therapy. Cox proportional hazards regression was used and cohort findings pooled by fixed-effects meta-analysis. Results: During mean ± SD follow-up of 14.0 ± 4.9 years, 254 new cases of DM were diagnosed. Correlations with self-reported dairy fat consumption were modest for 17:0 (r=0.19), 16:1 n-7t (r=0.21) and 15:0 (r=0.27), and weaker for 14:0 (r=0.12). In pooled multivariate analyses, comparing highest to lowest quartiles, lower risk of DM was seen for 17:0 [HR=0.57 (95% CI: 0.39 - 0.82), P-trend=0.001] and 16:1 n-7t [HR=0.60 (0.42 - 0.87), P-trend=0.007], while 14:0 was positively associated [HR=1.98 (1.35 - 2.91), P-trend <0.0001] and 15:0 was not associated [HR=1.11 (0.75 - 1.63), P-trend=0.80] ( Table ). Conclusion: In separate prospective cohorts, two biomarkers of dairy fat (17:0 and 16:1 n-7t) were associated with lower incidence of DM. 14:0, which is also obtained from beef, and is a marker of de novo lipogenesis, was associated with increased risk. Further research is needed on plausible biological and mechanistic pathways.