RESUMO
BACKGROUND: Coffee consumption has been associated with a lower risk of diabetes, but little is known about the mechanisms responsible for this association, especially related to the time when coffee is consumed. OBJECTIVE: We examined the long-term effect of coffee, globally and according to the accompanying meal, and of tea, chicory, and caffeine on type 2 diabetes risk. DESIGN: This was a prospective cohort study including 69,532 French women, aged 41-72 y from the E3N/EPIC (Etude Epidémiologique auprès de Femmes de la Mutuelle Générale de l'Education Nationale/European Prospective Investigation into Cancer and Nutrition) cohort study, without diabetes at baseline. Food and drink intakes per meal were assessed by using a validated diet-history questionnaire in 1993-1995. RESULTS: During a mean follow-up of 11 y, 1415 new cases of diabetes were identified. In multivariable Cox regression models, the hazard ratio in the highest category of coffee consumption [> or =3 cups (375 mL)/d] was 0.73 (95% CI: 0.61, 0.87; P for trend < 0.001), in comparison with no coffee consumption. This inverse association was restricted to coffee consumed at lunchtime (hazard ratio: 0.66; 95% CI: 0.57, 0.76) when comparing >1.1 cup (125 mL)/meal with no intake. At lunchtime, this inverse association was observed for both regular and decaffeinated coffee and for filtered and black coffee, with no effect of sweetening. Total caffeine intake was also associated with a statistically significantly lower risk of diabetes. Neither tea nor chicory consumption was associated with diabetes risk. CONCLUSIONS: Our data support an inverse association between coffee consumption and diabetes and suggest that the time of drinking coffee plays a distinct role in glucose metabolism.