Eggs, dietary cholesterol, choline, betaine, and diabetes risk in the Women's Health Initiative: a prospective analysis.

BACKGROUND

Epidemiological studies have been inconsistent regarding the relations between diabetes risk and the consumption of eggs and nutrients in eggs, such as choline, betaine, and cholesterol. There have been few studies among elderly women.

OBJECTIVES

The objective of this study was to examine associations between consumption of eggs, cholesterol, choline, and betaine and the risk of diabetes among elderly US women.

METHODS

Multivariable Cox regression was used with data from the prospective Women's Health Initiative. Population attributable risks were calculated. Consumption of eggs alone (not mixed in foods) and nutrients were assessed with an FFQ. Diabetes incidence was defined as the first incidence of self-reported diabetes treated with oral diabetes medication or insulin injections.

RESULTS

There were 46,263 women at follow-up baseline. During 13.3 y and 592,984 person-years of follow-up, there were 5480 incident diabetes cases. Higher egg, cholesterol, and choline consumption were each significantly associated with increases in diabetes risk. The associations for eggs and choline were not significant after adjustment for cholesterol consumption. The association for eggs was attenuated after adjustment for non-egg cholesterol consumption, with 1 significant HR in the top consumption quintile (≥3 eggs/wk) of 1.15 (95% CI: 1.05, 1.27; P for linear trend = 0.0001). The population attributable risks for obesity, overweight, consumption of ≥3 eggs/wk, inadequate exercise, and poor diet were 25.0 (95% CI: 22.3, 27.6), 12.8 (95% CI: 11.1, 14.5), 4.2 (95% CI: 2.3, 6.1), 3.5 (95% CI: 1.2, 5.8), and 3.1 (95% CI: 0.5, 5.7), respectively.

CONCLUSIONS

As egg consumption increased to ≥3 eggs/wk, there was a steady increase in diabetes risk that may have been due to the cholesterol in the eggs. The population attributable risk for ≥3 eggs/wk was far lower than that for being obese or overweight.