"Diet and Risk of Clinical Diabetes in Women" American Journal of Clinical Nutrition 55: 1018-1023 1992
To determine the relations of diet with risk of clinical noninsulin-dependent diabetes, we analyzed data from a prospective cohort of 84,360 US women. During 6 y of follow-up we identified 702 definite incident cases. Because body mass index (BMI) is a powerful risk factor for diabetes, we examined the relations of fat (including type), fiber, sucrose, and other components of diet to risk of diabetes, among women with BMIs (in kg/m2) < 29 kg/m2. After controlling for body mass index, previous weight change, and alcohol intake, we observed no associations between intakes of energy, protein, sucrose, carbohydrate, or fiber and risk of diabetes. Compared with women in the lowest quintile of energy-adjusted intake, the relative risks (and tests for trend) for those in the highest quintile were 0.61 (P trend = 0.03) for vegetable fat, 0.62 (P trend = 0.008) for potassium, 0.70 (P trend = 0.005) for calcium, and 0.68 (P trend = 0.02) for magnesium. These inverse associations were attenuated among obese women (BMIs >= 29).
Additional quote from text:
"We also examined specific foods in relation to risk of diabetes among
women with a body mass index < 29. We added each of the 61 food items to a
model that included body mass index, alcohol intake, energy intake, and
prior weight change; we observed that many items were significantly related
to risk. Skim milk, cabbage, carrots, homemade pie, cake, dark bread,
peanut butter, and wine were each inversely related to risk. Whole milk,
yogurt, processed meats, eggs, cola, noncarbonated beverages, home-fried
food, white bread, potato chips, rice and pasta, and French fries were each
positively related to risk.
To examine associations for fruit and vegetables, we summed the daily intake of reported intakes for fruits, vegetables, and fruit and vegetables combined. Among women with a body mass index < 29, intake of vegetables was inversely related to risk of diabetes, but intake of fruit was not related to risk."
Additional comment by poster:
The effect of glycemic index (GI) on diabetes was not separately examined in this study. The reason why vegetables, but not fruit are associated with reduced risk, may be because vegetables, but not fruit are often consumed with the major meals. Although most vegetables have high GIs themselves, it has been found that addition of vegetables such as carrots to a large meal lowers the GI of the entire meal. (European Journal of Clinical Nutrition 48: 386-396 1994)