Background: Consumption of whole fruits, but not fruit juice, has been associated with lower risk of type 2 diabetes. However, individual fruits have different compositions of carbohydrate, phytochemicals and other nutrients, and may thus have different effects on diabetes risk. We examined whether consumption of specific fruits was prospectively associated with risk of type 2 diabetes in US men and women. In addition, we evaluated whether the associations of fruits with diabetes were determined by the glycemic load (GL) of fruits consumption. Methods: After excluding participants with chronic diseases at baseline, we evaluated 66,720 women from the Nurses' Health Study (1984-2008); 85,961 women from the Nurses’ Health Study II (1991-2009); and 26,149 men from the Health Professionals Follow-up Study (1986-2008). Validated semi-quantitative food frequency questionnaires were administered to assess habitual consumption of fruits and other foods every two to four years. Incidences of type 2 diabetes were identified using biennial self-reported questionnaires and confirmed using supplementary questionnaires. The associations were prospectively assessed in each cohort, using Cox proportional hazard regression. Sociodemographics, lifestyle, caloric intakes, intakes of other fruits, and other dietary factors were adjusted for to control for confounding. Cohort-specific estimates were pooled by a random-effects meta-analysis. Results: During 3,447,866 person-years of follow-up, 11,521 participants were newly diagnosed with type 2 diabetes. The pooled multivariable-adjusted hazard ratios (HRs) for type 2 diabetes (for every 3 servings/week increase of fruit intake) were 0.74 [95% confidence interval (CI): 0.66, 0.83] for blueberries; 0.86 (0.80, 0.93) for grapes or raisins; 0.86 (0.75, 0.99) for prunes; 0.91 (0.81, 1.03) for bananas; 0.93 (0.88, 0.97) for apples or pears; 0.95 (0.91, 1.00) for grapefruits; 0.96 (0.90, 1.02) for peaches, plums or apricots; 1.01 (0.97, 1.05) for oranges; 1.06 (0.96, 1.19) for strawberries; and 1.11 (1.03, 1.20) for cantaloupe. No significant heterogeneity was found among the three cohorts, except associations with banana consumption. In contrast, each drink per day of fruit juice was associated with a HR (95% CI) of 1.07 (1.00, 1.14). The HRs for type 2 diabetes were 0.81 (0.69, 0.96) per 1 serving/day of high GL fruits, 0.96 (0.86, 1.07) for moderate GL fruits, and 1.05 (0.92, 1.19) for low GL fruits. Conclusion: Our data suggest that intakes of certain whole fruits, including grapes or raisins, prunes, apples or pears, grapefruits, and blueberries, are associated with lower risk of type 2 diabetes, whereas high consumption of fruit juice may lead to increased risk. Future research is needed to confirm our findings and elucidate mechanism underlying the associations for individual fruits.
Read full abstract