Descriptive statistics for the study population and the distribution of risk factors for cases and controls are shown in . The mean age was 61.6 years for cases and 64.3 years for controls (p = 0.02). Overall, consistent with other studies, obesity was strongly associated with increased risk for endometrial cancer, while current smoking, parity, and oral contraceptive use appeared to decrease endometrial cancer risk. Age-adjusted mean values for the HEI-2005 score as well as individual food components in cases and controls are shown in . The age-adjusted mean values for the overall HEI-2005 score, total fruit, whole fruit, total vegetables, dark green and orange vegetables, whole grains, dairy, meat and beans, and oils were lower for cases when compared to controls; although the mean difference was borderline significant only for the whole fruit component and for the dairy component. The age-adjusted mean values for total grain, saturated fat, sodium, and SoFAAS were marginally higher for cases when compared to controls. However, none of these mean differences were statistically significant.
Select characteristics of cases and controls participating in the EDGE study, New Jersey
Age-adjusted means for the HEI score and its components based on daily food consumption in cases and controls
Endometrial cancer risk estimates associated with the HEI-2005 score and its individual components are shown in . As mentioned earlier, by convention, a HEI score of more than 80 indicates a “good” diet, scores between 51 and 80 indicate a diet that “needs improvement,” and scores of less than 51 indicate a “poor” diet [9
]. Using these cutpoints, only 11 cases and 7 controls were in the “poor” diet category, with 85% of the cases and 83% of the controls being in the “needs improvement” category. Because this was not an optimal distribution, we decided to categorize the HEI score variable in quartiles (). The OR for the highest vs. the lowest quartile of the HEI-2005 score was below one, but the 95% confidence interval included the null value (OR = 0.83, 95% CI: 0.52–1.34). Nevertheless, we also computed the OR comparing a HEI-2005 score of >80 (“good diet”) vs. <51 (“poor diet”), with similar results (OR: 0.64; 95% CI: 0.20–2.05) (data not shown).
Association between the Healthy Eating Index score (HEI-2005) and its components and endometrial cancer risk
Overall, there was little evidence for a statistically significant association between any of the individual food components and endometrial cancer risk (). Although the point estimates for the highest vs. the lowest quartile of total fruit, whole fruit, total vegetable, total grain, whole grain, dairy, meat and beans, oils, saturated fat, and sodium were below one, the confidence intervals included the null value. For the meat and beans component, the OR was 0.70 (95% CI: 0.45–1.11; p for trend: 0.07). We also analyzed separately the foods included in the meat and beans component. We did not find any significant associations with any of the foods, although the effect estimates for the highest quartiles of fish, beans, and legumes when compared to the lowest quartiles were below one or close to one (data not shown). For red meat, the OR for the highest quartile compared to the lowest was above one, but there was little evidence of an association (OR = 1.11, 95% CI: 0.70–1.75).
To evaluate whether an individual's body mass index (BMI) modified the relationship between HEI score and endometrial cancer, stratified analyses were conducted. Although risk estimates for the highest HEI score category were of different magnitude in normal, overweight, and obese women, the confidence intervals overlapped, and the p for interaction was not statistically significant (p = 0.11; data not shown). Overall, our analyses provided little support for effect modification by BMI.