Dietary Energy Density, Glycemic Load, Glycemic Index, and Risk for Endometrial Cancer in the CPS-II Nutrition Cohort Free

There is growing interest in the role of dietary factors related to energy balance in endometrial cancer etiology. The association between energy density (ED), an estimate of energy per unit of food (kcal/g; ref. 1), and endometrial cancer risk has not been studied. Glycemic index (GI) reflects the difference in glycemic response between consumption of a food compared with glucose. Glycemic load (GL) reflects carbohydrate quality and quantity. GI and GL could influence cancer risk through effects on insulin and glucose metabolism, adiposity, and related conditions (2). The World Cancer Research Fund/American Institute for Cancer Research concluded that GL (but not GI) is a probable cause of endometrial cancer (3). The lack of convincing evidence may be due to inconsistencies in results among prospective studies and other methodologic concerns. The ability to update hysterectomy status and censor during follow-up is important for minimizing bias, especially if hysterectomy rates differ by GL or related factors (e.g., adiposity). Residual confounding by smoking, which is inversely associated with endometrial cancer risk, is another potential concern (4, 5). Finally, type II diabetes (T2D) is a risk factor for endometrial cancer. Cohort participants with prevalent T2D may have already made changes to lower carbohydrate (GL) intakes. At least one study (4) observed that GL was lower in women with T2D and positively associated with risk only among women without baseline T2D (4). Careful consideration of these potential sources of bias is necessary when examining associations of ED, GL, and GI with risk.

The American Cancer Society established the Cancer Prevention Study II Nutrition Cohort (CPS-II NC) in 1992 (6). In 1999, 73,641 CPS-II NC women completed a 152-item food frequency questionnaire (FFQ; 17–19). We excluded women who completed no subsequent surveys (n = 1,741), had a history of any cancer prior to 1999 (except nonmelanoma skin cancer; n = 13,633), were perimenopausal, premenopausal, of unknown menopausal status (n = 189), had an unverified self-reported cancer (n = 38), reported implausible dietary data [very low (<500 kcal/day) or high (>3,500 kcal/day) energy intakes or missing FFQ sections (n = 1,255)], or reported having a hysterectomy (n = 22,452) or diabetes (n = 3,336), leaving 30,997 women in this analysis.

Endometrial cancer cases (ICD 10:C54.1 and C55; ICD 9:179 and 182) were identified through self-report on follow-up questionnaires and verified via medical records or linkage with state registries or the National Death Index. Person-years of follow-up were computed from completion of the 1999 FFQ to the date of last return of a subsequent biennial questionnaire, reported hysterectomy, diagnosis of endometrial cancer, death, or June 30, 2013.

Statistical analyses were performed using SAS version 9.4. Multivariable-adjusted HRs and 95% confidence intervals (CIs) were estimated for quartiles of dietary ED, GL, and GI using Cox proportional hazards regression models with the lowest category (Q₁) as the referent. These models are stratified by individual age and adjusted for hormone replacement therapy (HRT) use (never, current, former, or missing use of estrogen or combined HRT), detailed smoking history, age at menarche, parity, age at menopause, oral contraceptive use, and physical activity. Final models were evaluated with and without body mass index (BMI; kg/m²) as a continuous variable. Other potential covariates included in Table 1 were considered but did not confound associations. Linear trends were assessed by assigning the median value from each quartile of the exposures and modeling this as a continuous variable. The Cox proportional hazards assumption was tested for each dietary variable, and no violations were observed. Interactions by BMI, HRT, and physical activity with dietary variables were assessed by including each factor and its cross-product term with quartile of diet variables in separate multivariable models using the likelihood ratio test. Results from two-sided χ² tests were considered statistically significant at P < 0.05.

During a median follow-up of 13.6 years, 425 endometrial cancer cases were identified. Median dietary ED was 1.5 kcal/g [interquartile range (IQR) = 1.3–1.7 kcal/g]. Median (IQR) GL and GI were 113.7 (100.5–126.8) and 52.5 (50.4–54.5), respectively. Participants who consumed higher ED diets were more likely to have higher BMI, larger waist circumference, and greater adult weight gain, and diets higher in total energy and fat (Table 1). They were also more likely to be current smokers and more sedentary. In contrast, GL and GI were generally inversely associated with these lifestyle characteristics.

We found no associations of ED, GL, or GI with endometrial cancer risk (Table 2). Furthermore, there was no evidence of statistical interactions with BMI, physical activity, or HRT use (not shown).

These results do not support an association between ED and postmenopausal endometrial cancer risk. Our null results for GL and GI are consistent with those of the two largest endometrial cancer studies, the NIH-AARP (n = 1,041 cases; ref. 5) and the European Prospective Investigation into Cancer and Nutrition cohorts (n = 710 cases; ref. 7).

No potential conflicts of interest were disclosed.

Conception and design: T.J. Hartman, M.M. Gaudet, S.M. Gapstur

Development of methodology: T.J. Hartman, S.M. Gapstur

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M.L. McCullough, S.M. Gapstur

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): T.J. Hartman, M.L. McCullough, J.M. Hodge, M.M. Gaudet, S.M. Gapstur

Writing, review, and/or revision of the manuscript: T.J. Hartman, M.L. McCullough, J.M. Hodge, M.M. Gaudet, Y. Wang, S.M. Gapstur

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.M. Hodge

The authors would like to acknowledge the contribution to this study from central cancer registries supported through the Centers for Disease Control and Prevention's National Program of Cancer Registries and cancer registries supported by the NCI's Surveillance Epidemiology and End Results Program. The American Cancer Society funds the creation, maintenance, and updating of the Cancer Prevention Study II cohort.

The authors express sincere appreciation to all Cancer Prevention Study II participants, and to each member of the study and biospecimen management group.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.