Abstract
Flavonoids have anti-inflammatory and antioxidative effects and thus may protect against diabetes. Therefore, we hypothesized that consumption of flavonoids and specific food and beverage sources of flavonoids would be associated with reduced risk of incident diabetes. At baseline (1986), diet (by food frequency questionnaire) and health information were collected from 35,816 postmenopausal women free of diabetes. Self-reported incident diabetes was ascertained 5 times during the study (1987, 1989, 1992, 1997, and 2004). Cox proportional hazards regression was used to calculate hazard ratios for incident diabetes according to categories of total flavonoids and anthocyanidins, flavones, flavanones, flavonols, flavan-3-ol monomers, isoflavones, and proanthocyanidins. Hazard ratios according to intake categories of flavonoid-rich foods and beverages were also calculated (apples, pears, berries, broccoli, bran, citrus, tea, and red wine). Flavonoid consumption was not associated with diabetes risk after multivariable adjustment. Although other flavonoid-rich foods and beverages were not associated, red wine was inversely associated with diabetes. Women who reported drinking red wine >or=1 time/wk had a 16% reduced risk of diabetes than those drinking wine <1 time/wk [HR (95% CI): 0.84 (0.71, 0.99)], with parallel findings for white wine, beer, and liquor. In conclusion, these data do not support a diabetes-protective effect of flavonoids. The suggestive evidence of a protective effect of regular red wine consumption is shared with an inverse association between alcohol drinks in general and diabetes risk and may reflect the effects of nonflavonoid constituents that are common to all alcohol drinks.