Abstract
Endometrial cancer risk has been directly associated with glycemic load. However, few studies have investigated this link, and the etiological role of specific dietary carbohydrate components remains unclear. Our aim was to investigate associations of carbohydrate intake, glycemic index, and glycemic load with endometrial cancer risk in the US Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial. Recruitment took place in 1993–2001. Over a median of 9.0 years of follow-up through 2009, 386 women developed endometrial cancer among 36,115 considered in the analysis. Dietary intakes were assessed using a 124-item diet history questionnaire. Cox proportional hazards models were applied to calculate hazard ratios and 95% confidence intervals. Significant inverse associations were detected between endometrial cancer risk and total available carbohydrate intake (hazard ratio (HR) = 0.66, 95% confidence interval (CI): 0.49, 0.90), total sugars intake (HR = 0.71, 95% CI: 0.52, 0.96), and glycemic load (HR = 0.63, 95% CI: 0.46, 0.84) when women in the highest quartile of intake were compared with those in the lowest. These inverse associations were strongest among overweight and obese women. No associations with endometrial cancer risk were observed for glycemic index or dietary fiber. Our findings contrast with previous evidence and suggest that high carbohydrate intakes and glycemic loads are protective against endometrial cancer development. Further clarification of these associations is warranted.
Keywords: carbohydrate, dietary fiber, endometrial cancer, glycemic index, glycemic load
Endometrial cancer is a relatively common malignancy affecting women in developed countries (1), where type 1 endometrial tumors have been increasing in incidence (2–4). As such, factors associated with Western lifestyles have been closely linked with endometrial cancer development. It is the female malignancy most highly related to excess body weight (5), yet despite this clear direct association with body size, very few dietary risk factors have been implicated in endometrial cancer etiology (6). In particular, no conclusions were made with respect to carbohydrate and dietary fiber intakes and endometrial cancer risk in the World Cancer Research Fund/American Institute of Cancer Research 2007 global food, nutrition, physical activity, and cancer report (6).
A recent systematic review showed that diets high in glycemic load are directly associated with endometrial cancer development (7). Glycemic load is a carbohydrate-related measure designed to reflect the overall glucose demand of a food over the 2-hour period after consumption; it incorporates both the glycemic index value and the total available carbohydrate content of the usual portion size of the food (8). Glycemic load can therefore be interpreted as a measure of carbohydrate quality and quantity. Glycemic index values rank the glucose demand of the standardized portion size (usually 50 g) of a food in comparison with white bread or glucose over the 2-hour postprandial period. Glycemic index can therefore be viewed as a measure of carbohydrate quality only. Findings from the systematic review also indicated that a high-glycemic-load diet increases the risk of endometrial cancer with higher body mass index (BMI; weight (kg)/height (m)2), suggesting that BMI may be an effect modifier of the association between glycemic load and endometrial cancer (7). However, only 4 prospective cohort studies were included in the review (9–12). A subsequent study originating from the National Institutes of Health-AARP Diet and Health Study suggested that women with the highest glycemic loads were at a nonsignificantly heightened risk of endometrial cancer (13). In contrast, glycemic index has not been associated with endometrial cancer risk (7, 13), while reports on other carbohydrate components, including sugars, starch, and fiber, have produced conflicting findings (9, 14–16). Because of the relatively small number of studies conducted to date, we sought to investigate the associations of glycemic index, glycemic load, and other dietary carbohydrate components with endometrial cancer risk within a large US prospective study, the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial.
MATERIALS AND METHODS
Study design
The PLCO Cancer Screening Trial is a large population-based trial in which participants have been randomized to undergo screening procedures for the 4 cancer types of interest or to follow their usual medical care practices (17). Recruitment of participants occurred between 1993 and 2001 at 10 study centers throughout the United States: Georgetown University/Lombardi Cancer Center (Washington, DC), Henry Ford Health System (Detroit, Michigan), Marshfield Clinic Research Foundation (Marshfield, Wisconsin), Pacific Health Research and Education Institute (Honolulu, Hawaii), University of Alabama at Birmingham (Birmingham, Alabama), University of Colorado (Denver, Colorado), University of Minnesota (Minneapolis, Minnesota), University of Pittsburgh (Pittsburgh, Pennsylvania), University of Utah (Salt Lake City, Utah), and Washington University School of Medicine (St. Louis, Missouri) (18). Approval for the study was granted by the relevant institutional review boards at all study centers.
Female PLCO study participants
Women were eligible for enrollment if they were aged ≥55 years and <75 years; had no prior diagnosis of primary or metastatic colon, rectal; lung, or ovarian cancer; had not undergone surgical removal of both ovaries (prior to October 2006), the entire colon, or 1 lung; and had no previous or current use of tamoxifen or raloxifene during the 6 months prior to randomization (prior to April 1999). At recruitment, participants completed a baseline questionnaire that included questions about sociodemographic background, smoking history, personal and family medical histories, anthropometric factors, reproductive factors, medication and hormone use, and recent history of screening examinations for prostate, lung, colorectal, and ovarian cancers.
At baseline, of the 78,216 women recruited, 2,094 had more than 8 missing frequency responses for the dietary questionnaires, 81 had a personal history of endometrial cancer at study recruitment, and 5 had invalid follow-up times and were excluded from analysis. Further exclusion criteria applied to the remaining 76,036 women in this data set (in order) were: noncompletion of the validated Diet History Questionnaire (DHQ) (n = 15,714), having undergone a hysterectomy prior to study entry (n = 21,974), having the questionnaire completed by a proxy respondent (n = 255), having missing information on height or weight (n = 457), and having an extreme (greater than twice the interquartile range) log energy intake or BMI (n = 428). A further 201 and 892 women who developed endometrial cancer or were censored from follow-up after study recruitment but prior to completion of the DHQ, respectively, were also excluded. This left 36,115 women for inclusion in this analytical data set.
Participants were followed from the date of DHQ completion to study year 13, December 31, 2009, the date of death, or the date of endometrial cancer diagnosis, whichever came first. The median time elapsed between recruitment/completion of the baseline questionnaire and completion of the DHQ was 3.0 years (interquartile range, 2.9–4.0). Tumors classified as malignant neoplasms of the corpus uteri were considered cases of endometrial cancer (codes C.54–C.55 in the International Classification of Diseases for Oncology, Third Edition). The majority of cases included in the study (92.2%) were neoplasms of the endometrium (code C54.1), with the remainder being classified as malignant neoplasms of the corpus uteri or uterus.
Dietary assessment
Respondents who completed the validated 124-food item DHQ were considered in the current dietary analyses of carbohydrate intake, glycemic index, and glycemic load. Comparisons between the DHQ and four 24-hour dietary recalls revealed the deattenuated correlation coefficients for energy-adjusted carbohydrate and fiber intakes in women to be 0.69 and 0.77, respectively (19). Dietary nutrient intakes were calculated by multiplying the daily frequency of each consumed food item by the nutrient value of the sex-specific portion size based on the nutrient database from the US Department of Agriculture's 1994–1996 Continuing Survey of Food Intakes by Individuals (19). In the US Department of Agriculture's food composition tables, total carbohydrate values include both available carbohydrate, including sugars and starches, and dietary fiber. Therefore, dietary fiber intake was subtracted from total carbohydrate intake to obtain a value for total available carbohydrate, as previously described (20). Glycemic index and glycemic load values were assigned to each food using published international tables of values (21), finding the best match possible as previously described (22). Sex- and serving-size-specific glycemic load values were then calculated for 225 nutritionally similar food groups using the weighted mean method (19, 23), and daily glycemic load was calculated according to the frequency of intake. Daily glycemic index was then determined by dividing glycemic load by total available carbohydrate intake and multiplying the result by 100 (22). Total sugars intake was classified as total available carbohydrate intake minus starch intake.
Statistical analysis
Cox proportional hazards models were applied to calculate hazard ratios and 95% confidence intervals for endometrial cancer risk, using person-years of follow-up as the underlying time variable. Glycemic load, total available carbohydrate, and intakes of fiber, starch, and sugars were adjusted for energy intake using the nutrient density method (24). Cox proportional hazards plots were visually inspected to ensure that assumptions were met. Dietary variables were categorized according to energy-adjusted quartiles of intake (except for glycemic index, which was not highly correlated with energy; r = 0.02), based on the distribution in the whole analytical cohort. In addition, dietary variables were explored as continuous variables by dividing energy-adjusted intake (except glycemic index) by the difference between the 25th and 75th percentiles, after checking for a normal data distribution.
In minimally adjusted models, we included energy intake (log kcal/day) and age at completion of the DHQ (years) as covariates. For potential confounders, the mode was imputed for missing or unknown values for education (n = 58; <0.2%), history of diabetes (n = 181; <0.5%), age at menarche (n = 57; <0.2%), age at menopause (n = 324; <0.9%), use of hormone replacement therapy (n = 212; <0.6%), and use of oral contraceptives (n = 20; <0.1%). In multivariable-adjusted models, additional confounders included ethnicity (non-Hispanic white, non-Hispanic black, Asian, or other), age at menarche (<12, 12–13, or ≥14 years), age at menopause (<50 years vs. ≥50 years), oral contraceptive use (ever/never), and BMI (<18.5, 18.5–<25, 25–<30, or ≥30). Other potential confounders that were tested but did not significantly influence results (P > 0.25 in models) and therefore were not included in the models were history of diabetes, study center, education, height, parity, hormone replacement therapy, personal history of endometriosis, personal history of uterine fibroid tumors, alcohol intake, physical activity level, and smoking status. In further models, the nutrient density method was also used to evaluate possible confounding effects of other macronutrients (24), and these energy-adjusted intakes were used to assess correlations between nutrients using Pearson's partial correlation coefficient. P values for trend across quartiles were estimated by assigning the median intake value for the quartile to each person and including this as a continuous variable in the hazards model. We also conducted analyses stratified a priori to explore effect modification for tertiles (data not shown because of limited statistical power) and continuous increments of dietary intake and endometrial cancer risk by BMI category. Tests for interaction were conducted by including an additional multiplicative variable for BMI and dietary intake in the regression model in tertile analysis and by means of the Wald test for continuous analysis. Stratified analysis was also performed for endometrial adenocarcinomas only and according to time to diagnosis for cases diagnosed during the first 2, 2–4, and >4 years of follow-up. All statistical analysis was conducted using Intercooled Stata, version 11 (StataCorp LP, College Station, Texas).
RESULTS
Over a median follow-up period of 9.0 years, 386 women developed endometrial cancer among 36,115 women included in this analysis (corresponding to a total of 305,360 person-years of follow-up). Characteristics of participants are displayed according to quartile of energy-adjusted glycemic load in Table 1. Persons with the highest dietary glycemic loads were slightly older and less likely to be Caucasian, to be obese, to have a college or postgraduate education, to be a current or former smoker, and to have used oral contraceptives, but they reported undertaking vigorous physical activity more frequently than women with the lowest glycemic loads, although activity levels were unknown for more than half of participants. History of diabetes, use of hormone replacement therapy, parity, and ages at menarche and menopause were largely similar across categories of glycemic load.
Table 1.
Characteristics of Women in the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial According to Energy-adjusted Quartile of Glycemic Load, United States, 1993–2001
| Characteristic | Energy-adjusted Quartile of Glycemic Loada |
|||||||
|---|---|---|---|---|---|---|---|---|
| Quartile 1 (n = 9,029) |
Quartile 2 (n = 9,029) |
Quartile 3 (n = 9,029) |
Quartile 4 (n = 9,028) |
|||||
| No. | %b | No. | % | No. | % | No. | % | |
| Median age at DHQ completion, years | 64 | 65 | 65 | 66 | ||||
| Race/ethnicity | ||||||||
| Non-Hispanic white | 8,545 | 94.6 | 8,537 | 94.6 | 8,354 | 92.5 | 7,746 | 85.8 |
| Non-Hispanic black | 183 | 2.0 | 182 | 2.0 | 245 | 2.7 | 477 | 5.3 |
| Asian | 145 | 1.6 | 195 | 2.2 | 296 | 3.3 | 672 | 7.4 |
| Other | 156 | 1.7 | 115 | 1.3 | 134 | 1.5 | 133 | 1.5 |
| Body mass indexc | ||||||||
| <18.5 | 85 | 0.9 | 97 | 1.1 | 92 | 1.0 | 126 | 1.4 |
| 18.5–<25 | 3,726 | 41.3 | 3,690 | 40.9 | 3,730 | 41.3 | 4,265 | 47.2 |
| 25–<30 | 3,046 | 33.7 | 3,149 | 34.9 | 3,194 | 35.4 | 2,965 | 32.8 |
| ≥30 | 2,172 | 24.1 | 2,093 | 23.2 | 2,013 | 22.3 | 1,672 | 18.5 |
| Education | ||||||||
| Less than high school | 337 | 3.7 | 335 | 3.7 | 374 | 4.1 | 478 | 5.3 |
| High school/12 years | 3,249 | 36.0 | 3,121 | 34.6 | 3,155 | 34.9 | 3,104 | 34.4 |
| Some college/post-high school | 2,185 | 24.2 | 2,376 | 26.3 | 2,442 | 27.1 | 2,597 | 28.8 |
| College/postgraduate | 3,258 | 36.1 | 3,197 | 35.4 | 3,058 | 33.9 | 2,849 | 31.6 |
| Smoking status | ||||||||
| Never smoker | 3,944 | 43.7 | 5,027 | 55.7 | 5,408 | 59.9 | 5,729 | 63.5 |
| Former smoker | 3,870 | 42.9 | 3,247 | 36.0 | 2,995 | 33.2 | 2,663 | 29.5 |
| Current smoker | 1,215 | 13.5 | 755 | 8.4 | 626 | 6.9 | 636 | 7.0 |
| History of diabetes | ||||||||
| No | 8,581 | 95.0 | 8,574 | 95.0 | 8,621 | 95.5 | 8,623 | 95.5 |
| Yes | 448 | 5.0 | 455 | 5.0 | 408 | 4.5 | 405 | 4.5 |
| Age at menarche, years | ||||||||
| <12 | 1,768 | 19.6 | 1,642 | 18.2 | 1,709 | 18.9 | 1,642 | 18.2 |
| 12–13 | 4,926 | 54.6 | 5,085 | 56.3 | 4,972 | 55.1 | 4,878 | 54.0 |
| ≥14 | 2,335 | 25.9 | 2,302 | 25.5 | 2,348 | 26.0 | 2,508 | 27.8 |
| Age at menopause, years | ||||||||
| <50 | 3,039 | 33.7 | 2,939 | 32.6 | 3,016 | 33.4 | 3,106 | 34.4 |
| ≥50 | 5,990 | 66.3 | 6,090 | 67.5 | 6,013 | 66.6 | 5,922 | 65.6 |
| Median no. of livebirths | 3 | 3 | 3 | 3 | ||||
| Use of oral contraceptives | ||||||||
| Never use | 3,469 | 38.4 | 3,981 | 44.1 | 4,369 | 48.4 | 4,616 | 51.1 |
| Ever use | 5,560 | 61.6 | 5,048 | 55.9 | 4,660 | 51.6 | 4,412 | 48.9 |
| Use of hormone replacement therapy | ||||||||
| Never use | 3,303 | 36.6 | 3,515 | 38.9 | 3,651 | 40.4 | 3,823 | 42.4 |
| Former use | 4,182 | 46.3 | 4,013 | 44.5 | 3,793 | 42.0 | 3,641 | 40.3 |
| Current use | 1,544 | 17.1 | 1,501 | 16.6 | 1,585 | 17.6 | 1,564 | 17.3 |
| Vigorous physical activity, hours/week | ||||||||
| <1 | 1,487 | 16.5 | 1,282 | 14.2 | 1,200 | 13.3 | 1,193 | 13.2 |
| 1–4 | 1,932 | 21.4 | 1,985 | 22.0 | 1,933 | 21.4 | 1,766 | 19.6 |
| >4 | 844 | 9.4 | 873 | 9.7 | 906 | 10.0 | 924 | 10.2 |
| Unknown | 4,766 | 52.8 | 4,889 | 54.2 | 4,990 | 55.3 | 5,145 | 57.0 |
Abbreviation: DHQ, Diet History Questionnaire.
a Cutpoints for energy-adjusted quartiles of glycemic load were as follows: quartile 1, <57.8 units/1,000 kcal/day; quartile 2, 57.8–<65.2 units/1,000 kcal/day; quartile 3, 65.2–<72.6 units/1,000 kcal/day; quartile 4, ≥72.6 units/1,000 kcal/day.
b Percentages may not total 100 because of rounding.
c Weight (kg)/height (m)2.
As shown in Table 2, women with the highest energy-adjusted glycemic loads also had higher intakes of available carbohydrate, starch, total sugars, and fiber and a higher glycemic index but lower intakes of total energy, fat, saturated fat, protein, and alcohol compared with counterparts with lower glycemic loads.
Table 2.
Median Daily Nutrient Intakes of Women in the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial According to Energy-adjusted Quartile of Glycemic Load, United States, 1993–2001
| Median Daily Nutrient Intake | Energy-adjusted Quartile of Glycemic Loada |
|||
|---|---|---|---|---|
| Quartile 1 (n = 9,029) | Quartile 2 (n = 9,029) | Quartile 3 (n = 9,029) | Quartile 4 (n = 9,028) | |
| Energy, kcal | 1,501 | 1,474 | 1,399 | 1,300 |
| Total available carbohydrate, g/1,000 kcal | 99.0 | 116.4 | 128.5 | 145.2 |
| Glycemic index, units | 52.2 | 53.0 | 53.5 | 54.5 |
| Glycemic load, units/1,000 kcal | 52.3 | 61.7 | 68.6 | 78.2 |
| Starches, g/1,000 kcal | 39.9 | 46.3 | 49.5 | 52.3 |
| Total sugars, g/1,000 kcal | 56.8 | 70.1 | 78.9 | 92.4 |
| Sucrose, g/1,000 kcal | 18.4 | 23.1 | 26.1 | 29.8 |
| Fructose, g/1,000 kcal | 8.8 | 11.6 | 13.8 | 17.9 |
| Fiber, g/1,000 kcal | 7.6 | 9.9 | 12.0 | 15.5 |
| Total fat, g/1,000 kcal | 42.1 | 37.1 | 33.1 | 27.2 |
| Saturated fat, g/1,000 kcal | 12.9 | 11.4 | 10.1 | 8.1 |
| Protein, g/1,000 kcal | 40.8 | 39.9 | 38.5 | 35.1 |
| Alcohol, g | 4.2 | 1.5 | 0.8 | 0.4 |
a Cutpoints for energy-adjusted quartiles of glycemic load were as follows: quartile 1, <57.8 units/1,000 kcal/day; quartile 2, 57.8–<65.2 units/1,000 kcal/day; quartile 3, 65.2–<72.6 units/1,000 kcal/day; quartile 4, ≥72.6 units/1,000 kcal/day.
Results from Cox proportional hazards regression analysis of carbohydrate intake and endometrial cancer risk are displayed in Table 3. After adjustment for confounders, total available carbohydrate was strongly inversely associated with endometrial cancer risk when the highest quartile of intake was compared with the lowest (hazard ratio (HR) = 0.66, 95% confidence interval (CI): 0.49, 0.90; P for trend = 0.01) and in continuous analysis. Similar risk reductions were observed for glycemic load and endometrial cancer (quartile 4 vs. quartile 1: HR = 0.63, 95% CI: 0.46, 0.84; P for trend = 0.002). Similar or potentially stronger inverse associations were detected when the nutrient density method of energy adjustment was not applied to the analysis (see Supplementary Data, available at http://aje.oxfordjournals.org/). No clear pattern of association between glycemic index and endometrial cancer risk was observed in either minimally adjusted models or fully adjusted models. The similar results seen for endometrial cancer risk and energy-adjusted glycemic load and energy-adjusted total available carbohydrate reflect their very high correlation (r = 0.94). Neither dietary glycemic load nor available carbohydrate was highly correlated with glycemic index after adjustment for energy intake (r = 0.34 and r = 0.004, respectively). After energy adjustment, dietary glycemic load was also highly correlated with intakes of fat (r = −0.67) and saturated fat (r = −0.56) and was moderately correlated with intakes of dietary fiber (r = 0.31) and protein (r = −0.35). Similar patterns were seen for correlations between energy-adjusted available carbohydrate and other nutrients. In multivariable-adjusted models, further adjustments for energy-adjusted total fat, saturated fat, fiber, or protein intake did not alter the associations seen for glycemic load or total available carbohydrate intake and endometrial cancer risk (data not shown).
Table 3.
Risk of Endometrial Cancer in Relation to Quartile of Dietary Carbohydrate Intake Among Women in the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial, United States, 1993–2001
| Carbohydrate and Quartile of Intake | No. of Controls | No. of Cases | Age- and Energy-adjusted |
Multivariable-adjusteda |
||
|---|---|---|---|---|---|---|
| HR | 95% CI | HR | 95% CI | |||
| Total available carbohydrate, g/1,000 kcal/day | ||||||
| Q1 (<109.1) | 8,921 | 108 | 1.00 | 1.00 | ||
| Q2 (109.1–<122.6) | 8,919 | 110 | 1.00 | 0.77, 1.31 | 1.00 | 0.77, 1.31 |
| Q3 (122.6–<136.1) | 8,930 | 99 | 0.90 | 0.68, 1.18 | 0.91 | 0.69, 1.19 |
| Q4 (≥136.1) | 8,959 | 69 | 0.62 | 0.46, 0.85 | 0.66 | 0.49, 0.90 |
| P for trendb | 0.002 | 0.01 | ||||
| Continuous measurec | 35,729 | 386 | 0.80 | 0.70, 0.91 | 0.82 | 0.72, 0.93 |
| Glycemic index, units/day | ||||||
| Q1 (<51.1) | 8,940 | 89 | 1.00 | 1.00 | 1.00 | |
| Q2 (51.1–<53.3) | 8,911 | 118 | 1.31 | 0.99, 1.72 | 1.29 | 0.98, 1.70 |
| Q3 (53.3–<55.4) | 8,939 | 90 | 0.99 | 0.74, 1.33 | 0.97 | 0.72, 1.30 |
| Q4 (≥55.4) | 8,939 | 89 | 0.98 | 0.73, 1.31 | 0.94 | 0.70, 1.26 |
| P for trend | 0.49 | 0.34 | ||||
| Continuous measure | 35,729 | 386 | 0.95 | 0.86, 1.05 | 0.94 | 0.85, 1.04 |
| Glycemic load, units/1,000 kcal/day | ||||||
| Q1 (<57.8) | 8,911 | 118 | 1.00 | 1.00 | ||
| Q2 (57.8–<65.2) | 8,929 | 100 | 0.83 | 0.64, 1.09 | 0.82 | 0.63, 1.07 |
| Q3 (65.2–<72.6) | 8,934 | 95 | 0.79 | 0.60, 1.03 | 0.78 | 0.59, 1.02 |
| Q4 (≥72.6) | 8,955 | 73 | 0.60 | 0.44, 0.81 | 0.63 | 0.46, 0.84 |
| P for trend | 0.001 | 0.002 | ||||
| Continuous measure | 35,729 | 386 | 0.80 | 0.71, 0.91 | 0.81 | 0.72, 0.92 |
| Starches, g/1,000 kcal/day | ||||||
| Q1 (<39.7) | 8,928 | 100 | 1.00 | 1.00 | ||
| Q2 (39.7–<46.6) | 8,934 | 95 | 0.94 | 0.71, 1.24 | 0.92 | 0.69, 1.22 |
| Q3 (46.6–<53.7) | 8,932 | 97 | 0.95 | 0.72, 1.25 | 0.92 | 0.69, 1.21 |
| Q4 (≥53.7) | 8,935 | 94 | 0.91 | 0.68, 1.20 | 0.90 | 0.67, 1.19 |
| P for trend | 0.53 | 0.46 | ||||
| Continuous measure | 35,729 | 386 | 0.99 | 0.87, 1.11 | 0.99 | 0.87, 1.12 |
| Total sugars, g/1,000 kcal/day | ||||||
| Q1 (<61.2) | 8,924 | 105 | 1.00 | 1.00 | ||
| Q2 (61.2–<73.9) | 8,918 | 110 | 1.04 | 0.80, 1.36 | 1.07 | 0.82, 1.40 |
| Q3 (73.9–<87.8) | 8,929 | 101 | 0.95 | 0.72, 1.25 | 0.98 | 0.75, 1.29 |
| Q4 (≥87.8) | 8,958 | 70 | 0.67 | 0.49, 0.90 | 0.71 | 0.52, 0.96 |
| P for trend | 0.007 | 0.02 | ||||
| Continuous measure | 35,729 | 386 | 0.79 | 0.69, 0.91 | 0.81 | 0.71, 0.93 |
| Fiber, g/1,000 kcal/day | ||||||
| Q1 (<8.8) | 8,938 | 91 | 1.00 | 1.00 | ||
| Q2 (8.8–<10.9) | 8,926 | 103 | 1.12 | 0.85, 1.49 | 1.12 | 0.84, 1.49 |
| Q3 (10.9–<13.3) | 8,936 | 93 | 1.02 | 0.76, 1.36 | 1.03 | 0.77, 1.36 |
| Q4 (≥13.3) | 8,929 | 99 | 1.08 | 0.81, 1.44 | 1.13 | 0.85, 1.51 |
| P for trend | 0.76 | 0.53 | ||||
| Continuous measure | 35,729 | 386 | 0.99 | 0.87, 1.14 | 1.02 | 0.89, 1.17 |
Abbreviations: CI, confidence interval; HR, hazard ratio; Q, quartile.
a Adjusted for age at completion of the Diet History Questionnaire (years), body mass index (weight (kg)/height (m)2; <18.5, 18.5–<25, 25–<30, or ≥30), age at menarche (<12, 12–13, or ≥14 years), age at menopause (<50 years vs. ≥50 years), race/ethnicity (non-Hispanic white, non-Hispanic black, Asian, or other), oral contraceptive use (ever/never), and energy intake (calculated by the nutrient density method, except for glycemic index, and including log kcal/day in the model).
b P values for trend were estimated by assigning the median intake value for the quartile to each person and including this as a continuous variable in the model.
c Continuous intakes were calculated by dividing the energy-adjusted intake (except glycemic index) by the difference between the 75th and 25th percentiles.
As shown in Table 3, the inverse association between carbohydrate intake and endometrial cancer risk appeared to be attributable to total sugars intake (quartile 4 vs. quartile 1: HR = 0.71, 95% CI: 0.52, 0.96) rather than starch intake. In further analysis, when the associations between endometrial cancer risk and carbohydrate as individual monosaccharides or disaccharides were explored, no associations were detected for fructose intake (quartile 4 vs. quartile 1: HR = 0.89, 95% CI: 0.66, 1.21); however, sucrose intake was protective (quartile 4 vs. quartile 1: HR = 0.64, 95% CI: 0.47, 0.88). There was no evidence of a protective association between dietary fiber intake and endometrial cancer risk. Similar results were evident when analysis was restricted to endometrial adenocarcinomas only (Supplementary Data).
Table 4 shows results from stratified analysis of continuous measures of dietary variables (categorical data are not shown because of limited statistical power) and endometrial cancer risk by BMI category. The previously observed inverse associations between total available carbohydrate, glycemic load, and total sugars intake and endometrial cancer risk appeared to be strongest for overweight persons (and obese persons, for total sugars only), although tests for interaction were not statistically significant. The nonsignificant associations of glycemic index and starch and dietary fiber intakes with endometrial cancer risk were not markedly altered by varying BMI categories. In further analysis, there was no evidence of the protective associations for total available carbohydrate, total sugars, and glycemic load attenuating over time (data not shown).
Table 4.
Multivariable-adjusted Association between Energy-adjusted Dietary Factors and Risk of Endometrial Cancer Among Women in the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial, by Body Mass Index Category, United States, 1993–2001
| Dietary Factor and Body Mass Indexa Category | No. of Controls | No. of Cases | Hazard Ratiob,c | 95% Confidence Interval | P for Interactiond |
|---|---|---|---|---|---|
| Total available carbohydrate | |||||
| 18.5–<25 | 15,296 | 115 | 0.89 | 0.71, 1.13 | 0.55 |
| 25–<30 | 12,236 | 118 | 0.74 | 0.59, 0.94 | |
| ≥30 | 7,802 | 148 | 0.82 | 0.66, 1.01 | |
| Glycemic index | |||||
| 18.5–<25 | 15,296 | 115 | 0.90 | 0.75, 1.08 | 0.37 |
| 25–<30 | 12,236 | 118 | 0.83 | 0.69, 1.00 | |
| ≥30 | 7,802 | 148 | 1.05 | 0.89, 1.23 | |
| Glycemic load | |||||
| 18.5–<25 | 15,296 | 115 | 0.87 | 0.70, 1.10 | 0.31 |
| 25–<30 | 12,236 | 118 | 0.71 | 0.56, 0.89 | |
| ≥30 | 7,802 | 148 | 0.85 | 0.69, 1.05 | |
| Starches | |||||
| 18.5–<25 | 15,296 | 115 | 0.95 | 0.76, 1.19 | 0.22 |
| 25–<30 | 12,236 | 118 | 0.90 | 0.71, 1.13 | |
| ≥30 | 7,802 | 148 | 1.07 | 0.88, 1.37 | |
| Total sugars | |||||
| 18.5–<25 | 15,296 | 115 | 0.91 | 0.71, 1.17 | 0.29 |
| 25–<30 | 12,236 | 118 | 0.77 | 0.60, 0.99 | |
| ≥30 | 7,802 | 148 | 0.78 | 0.63, 0.98 | |
| Fiber | |||||
| 18.5–<25 | 15,296 | 115 | 1.05 | 0.83, 1.33 | 0.15 |
| 25–<30 | 12,236 | 118 | 1.07 | 0.83, 1.38 | |
| ≥30 | 7,802 | 148 | 0.96 | 0.75, 1.23 |
a Weight (kg)/height (m)2.
b Hazard ratio for continuous increments of intake, calculated by dividing the energy-adjusted intake (except glycemic index) by the difference between the 75th and 25th percentiles.
c Adjusted for age at completion of the Diet History Questionnaire (years), body mass index (continuous), age at menarche (<12, 12–13, or ≥14 years), age at menopause (<50 years vs. ≥50 years), race/ethnicity (non-Hispanic white, non-Hispanic black, Asian, or other), oral contraceptive use (ever/never), and energy intake (calculated by the nutrient density method, except for glycemic index, and including log kcal/day in the model).
d Estimated by means of the Wald test.
DISCUSSION
Analysis from the PLCO Cancer Screening Trial demonstrated that total available carbohydrate and glycemic load were significantly inversely associated with endometrial cancer risk. Inverse associations were stronger for total sugars intake than for starch intake, and this relationship was particularly evident in overweight and obese women. Glycemic index and dietary fiber intake were not related to endometrial cancer risk in this study population.
We identified no significant association between dietary fiber and endometrial cancer risk in either categorical or continuous analysis. This contrasts with results from a meta-analysis of case-control studies, in which Bandera et al. (25) found significant 20%–30% reduced risks of endometrial cancer for women with the highest dietary fiber intakes. However, our nonsignificant findings corroborate those from the only prospective case-cohort study identified in that systematic review (25) and 2 further prospective studies originating from the European Prospective Investigation into Cancer and Nutrition (EPIC) (9) and Nurses' Health Study (26) cohorts. This suggests that fiber intake is not related to endometrial cancer risk and perhaps that dietary measurement error may have influenced the results seen in previous case-control studies.
No evidence of an association between glycemic index and endometrial cancer risk was identified either, which is consistent with null findings from previous work (7, 13). However, the surprising observation that a high glycemic load was inversely associated with endometrial cancer risk in the PLCO Trial challenges previous conclusions from systematic reviews and meta-analyses (7, 27–29). Additionally, total available carbohydrate intake was strongly associated with a reduced risk of endometrial cancer. Conversely, the majority of investigations to date have shown no significant links between carbohydrate intake and endometrial cancer risk (10, 11, 30, 31). In the current study, glycemic load was very highly correlated with total available carbohydrate intake. Nonetheless, similar correlations were observed between glycemic load and total carbohydrate intake in EPIC (9), which found positive associations between these dietary factors and endometrial cancer risk, in contrast to the current study findings.
We examined total sugars and starches in an attempt to further clarify the inverse association between total available carbohydrate intake and endometrial cancer risk in the PLCO Trial. The protective association was stronger for total sugars intake than for starch intake, and this was attributed to sucrose intake. This contrasts with EPIC, which found a significant 36% increased risk of endometrial cancer per 50-g/day increment in total sugars intake (9). Few other studies have examined carbohydrate constituents in relation to endometrial cancer risk, with some reporting largely nonsignificant associations (14, 15), while another study observed direct associations with sucrose intake (32). Thus, it is unclear why glycemic load and total carbohydrate and sugars intakes were inversely related to endometrial cancer risk in our study. There is some evidence that a high-fat diet can increase the risk of developing endometrial cancer (33, 34), but this is certainly not conclusive, as other studies have observed no associations (14, 35) or indeed have found inverse associations for animal and saturated fat intakes and endometrial carcinogenesis (30). In our explorations, intake of dietary fat, including saturated fat, could not explain the inverse associations seen for glycemic load and available carbohydrate in the present analysis.
We had postulated that a high-glycemic-load diet may increase endometrial cancer risk via hyperinsulinemia and the insulin-like growth factor (IGF) system (36). Hyperinsulinemia has been shown to lower concentrations of IGF-binding protein, thereby increasing levels of insulin-like growth factor 1 (IGF-1), which inhibits apoptosis and sex hormone-binding globulin synthesis and stimulates cell proliferation and sex steroid synthesis, promoting tumor development (37). However, perhaps the results from the current study should not have been entirely unexpected. Free IGF-1 concentrations have been inversely related to endometrial adenocarcinoma development in a US prospective study, despite illustrating a strong positive association between hyperinsulinemia and risk, especially among overweight and obese participants (38). Moreover, the inverse association between IGF-1 and endometrial cancer risk was also strongest among overweight and obese women, rather than normal-weight women (38). The authors proposed that the seemingly paradoxical association may have been due to a lack of connection between circulating IGF-1 concentrations and those found in uterine cells (38). Since dietary carbohydrate factors are thought to affect IGF concentrations (39), it is plausible that these factors may explain the inverse associations with endometrial cancer risk in the current study, but they do not explain the lack of consistency with positive associations seen in other studies. However, these observations require confirmation before further speculation on the potential underlying mechanisms involved.
There is overwhelming evidence of a dose-response relationship between BMI and the risk of endometrial cancer (6, 40–42). Alternatively, overweight and obese persons are known to underreport total energy consumption in dietary assessments (43), so we cannot rule out the possibility that differential underreporting of energy intake in these women may have exaggerated the observed associations. It is also plausible that methodological issues regarding glycemic load calculations may explain the discrepancy between our findings and published data; however, this would not account for the inverse associations between total carbohydrate and sugars intakes and endometrial cancer risk. Moreover, the same method was applied in the study by George et al. (13); therefore, it may be that there are different underlying food sources at any given value of dietary glycemic index or glycemic load in the data from the PLCO Trial compared with previous studies. If this were the case, different dietary patterns with the same overall dietary glycemic index and glycemic load values might explain the divergent effects of carbohydrate quality observed between studies. It is also possible that residual confounding from unmeasured factors may have influenced the results shown or that some findings were due to chance. For example, we were unable to fully test adjustment for physical activity level, which has a protective role in endometrial carcinogenesis (44) but was unknown for 55% of women in our data set. Note that a large proportion of women in the original trial had undergone a hysterectomy and were subsequently excluded from the analyses. Although it was necessary to exclude these women since they were not at risk of endometrial cancer, they may have had different dietary habits, which in turn may have influenced the associations seen. However, the prevalence of women with hysterectomies in the PLCO Trial is broadly in line with the prevalences seen nationally in the United States for women aged 55–74 years (45, 46); therefore, this should have had a minimal impact on the representativeness of women in the PLCO Trial.
Strengths of the current study include its large sample size and prospective design, making it unlikely that any associations seen were attributable to reverse causation and that subclinical cancer may have contributed to changes in the dietary habits of cases. The dietary questionnaire has been shown to assess total carbohydrate and fiber intakes of women residing in the United States with fair validity and reproducibility (19), and a rigorous method was used to assign glycemic index and glycemic load values to foods, primarily using data on American foods where possible (22).
There are some limitations of the current analysis. Dietary questionnaire data were self-reported, which may have incorporated some measurement error from participants’ overemphasizing healthy eating patterns and underestimating intakes of unhealthier foods. A large number of women (n = 15,714) were excluded because they had not completed the validated DHQ, and it is possible that this may have introduced some respondent bias. In addition, dietary intake was assessed only once at baseline; therefore, we were unable to account for subsequent changes in diet. There was also limited statistical power to examine interactions between dietary factors and endometrial cancer risk by BMI stratum.
In conclusion, high glycemic load, available carbohydrate, and total sugars intake were significantly inversely related to endometrial cancer risk in the PLCO Trial. Further studies are needed to determine whether the results seen in the PLCO Trial are atypical or whether the relationship of glycemic load and total carbohydrate and sugars intakes with endometrial cancer risk is not straightforward and varies in different population groups.
Supplementary Material
ACKNOWLEDGMENTS
Author affiliations: Cancer Epidemiology and Health Services Research Group, Centre for Public Health, Queen's University Belfast, Belfast, Northern Ireland (Helen G. Coleman, Liam J. Murray, Marie M. Cantwell); Division of Cancer Epidemiology and Genetics, National Cancer Institute, Rockville, Maryland (Cari M. Kitahara, Amanda Black, Rachael Stolzenberg-Solomon); and Division of Cancer Prevention, National Cancer Institute, Rockville, Maryland (Kevin W. Dodd).
Conflict of interest: none declared.
REFERENCES
- 1.Amant F, Moerman P, Neven P, et al. Endometrial cancer. Lancet. 2005;366(9484):491–505. doi: 10.1016/S0140-6736(05)67063-8. [DOI] [PubMed] [Google Scholar]
- 2.Evans T, Sany O, Pearmain P, et al. Differential trends in the rising incidence of endometrial cancer by type: data from a UK population-based registry from 1994 to 2006. Br J Cancer. 2011;104(9):1505–1510. doi: 10.1038/bjc.2011.68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Boll D, Karim-Kos HE, Verhoeven RH, et al. Increased incidence and improved survival in endometrioid endometrial cancer diagnosed since 1989 in the Netherlands: a population based study. Eur J Obstet Gynecol Reprod Biol. 2013;166(2):209–214. doi: 10.1016/j.ejogrb.2012.10.028. [DOI] [PubMed] [Google Scholar]
- 4.Duong LM, Wilson RJ, Ajani UA, et al. Trends in endometrial cancer incidence rates in the United States, 1999–2006. J Womens Health (Larchmt) 2011;20(8):1157–1163. doi: 10.1089/jwh.2010.2529. [DOI] [PubMed] [Google Scholar]
- 5.Renehan AG, Tyson M, Egger M, et al. Body-mass index and incidence of cancer: a systematic review and meta-analysis of prospective observational studies. Lancet. 2008;371(9612):569–578. doi: 10.1016/S0140-6736(08)60269-X. [DOI] [PubMed] [Google Scholar]
- 6.World Cancer Research Fund/American Institute for Cancer Research. Food, Nutrition, Physical Activity, and the Prevention of Cancer: A Global Perspective. Washington, DC: American Institute for Cancer Research; 2007. [Google Scholar]
- 7.Mulholland HG, Murray LJ, Cardwell CR, et al. Dietary glycaemic index, glycaemic load and endometrial and ovarian cancer risk: a systematic review and meta-analysis. Br J Cancer. 2008;99(3):434–441. doi: 10.1038/sj.bjc.6604496. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Salmeron J, Ascherio A, Rimm EB, et al. Dietary fiber, glycemic load, and risk of NIDDM in men. Diabetes Care. 1997;20(4):545–550. doi: 10.2337/diacare.20.4.545. [DOI] [PubMed] [Google Scholar]
- 9.Cust AE, Slimani N, Kaaks R, et al. Dietary carbohydrates, glycemic index, glycemic load, and endometrial cancer risk within the European Prospective Investigation into Cancer and Nutrition cohort. Am J Epidemiol. 2007;166(8):912–923. doi: 10.1093/aje/kwm161. [DOI] [PubMed] [Google Scholar]
- 10.Silvera SA, Rohan TE, Jain M, et al. Glycaemic index, glycaemic load and risk of endometrial cancer: a prospective cohort study. Public Health Nutr. 2005;8(7):912–919. doi: 10.1079/phn2005741. [DOI] [PubMed] [Google Scholar]
- 11.Larsson SC, Friberg E, Wolk A. Carbohydrate intake, glycemic index and glycemic load in relation to risk of endometrial cancer: a prospective study of Swedish women. Int J Cancer. 2007;120(5):1103–1107. doi: 10.1002/ijc.22422. [DOI] [PubMed] [Google Scholar]
- 12.Folsom AR, Demissie Z, Harnack L, et al. Glycemic index, glycemic load, and incidence of endometrial cancer: the Iowa Women's Health Study. Nutr Cancer. 2003;46(2):119–124. doi: 10.1207/S15327914NC4602_03. [DOI] [PubMed] [Google Scholar]
- 13.George SM, Mayne ST, Leitzmann MF, et al. Dietary glycemic index, glycemic load, and risk of cancer: a prospective cohort study. Am J Epidemiol. 2009;169(4):462–472. doi: 10.1093/aje/kwn347. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Salazar-Martinez E, Lazcano-Ponce E, Sanchez-Zamorano LM, et al. Dietary factors and endometrial cancer risk: results of a case-control study in Mexico. Int J Gynecol Cancer. 2005;15(5):938–945. doi: 10.1111/j.1525-1438.2005.00253.x. [DOI] [PubMed] [Google Scholar]
- 15.Lucenteforte E, Talamini R, Montella M, et al. Macronutrients, fatty acids and cholesterol intake and endometrial cancer. Ann Oncol. 2008;19(1):168–172. doi: 10.1093/annonc/mdm446. [DOI] [PubMed] [Google Scholar]
- 16.Goodman MT, Wilkens LR, Hankin JH, et al. Association of soy and fiber consumption with the risk of endometrial cancer. Am J Epidemiol. 1997;146(4):294–306. doi: 10.1093/oxfordjournals.aje.a009270. [DOI] [PubMed] [Google Scholar]
- 17.Prorok PC, Andriole GL, Bresalier RS, et al. Design of the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial. Control Clin Trials. 2000;21(6 suppl):273S–309S. doi: 10.1016/s0197-2456(00)00098-2. [DOI] [PubMed] [Google Scholar]
- 18.Gohagan JK, Prorok PC, Hayes RB, et al. The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial of the National Cancer Institute: history, organization, and status. Control Clin Trials. 2000;21(6 suppl):251S–272S. doi: 10.1016/s0197-2456(00)00097-0. [DOI] [PubMed] [Google Scholar]
- 19.Subar AF, Thompson FE, Kipnis V, et al. Comparative validation of the Block, Willett, and National Cancer Institute food frequency questionnaires: the Eating at America's Table Study. Am J Epidemiol. 2001;154(12):1089–1099. doi: 10.1093/aje/154.12.1089. [DOI] [PubMed] [Google Scholar]
- 20.Flood A, Peters U, Jenkins DJ, et al. Carbohydrate, glycemic index, and glycemic load and colorectal adenomas in the Prostate, Lung, Colorectal, and Ovarian Screening Study. Am J Clin Nutr. 2006;84(5):1184–1192. doi: 10.1093/ajcn/84.5.1184. [DOI] [PubMed] [Google Scholar]
- 21.Foster-Powell K, Holt SH, Brand-Miller JC. International table of glycemic index and glycemic load values: 2002. Am J Clin Nutr. 2002;76(1):5–56. doi: 10.1093/ajcn/76.1.5. [DOI] [PubMed] [Google Scholar]
- 22.Flood A, Subar AF, Hull SG, et al. Methodology for adding glycemic load values to the National Cancer Institute Diet History Questionnaire database. J Am Diet Assoc. 2006;106(3):393–402. doi: 10.1016/j.jada.2005.12.008. [DOI] [PubMed] [Google Scholar]
- 23.Subar AF, Midthune D, Kulldorff M, et al. Evaluation of alternative approaches to assign nutrient values to food groups in food frequency questionnaires. Am J Epidemiol. 2000;152(3):279–286. doi: 10.1093/aje/152.3.279. [DOI] [PubMed] [Google Scholar]
- 24.Willett W, Stampfer MJ. Total energy intake: implications for epidemiologic analyses. Am J Epidemiol. 1986;124(1):17–27. doi: 10.1093/oxfordjournals.aje.a114366. [DOI] [PubMed] [Google Scholar]
- 25.Bandera EV, Kushi LH, Moore DF, et al. Association between dietary fiber and endometrial cancer: a dose-response meta-analysis. Am J Clin Nutr. 2007;86(6):1730–1737. doi: 10.1093/ajcn/86.5.1730. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Cui X, Rosner B, Willett WC, et al. Dietary fat, fiber, and carbohydrate intake in relation to risk of endometrial cancer. Cancer Epidemiol Biomarkers Prev. 2011;20(5):978–989. doi: 10.1158/1055-9965.EPI-10-1089. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Galeone C, Augustin LS, Filomeno M, et al. Dietary glycemic index, glycemic load, and the risk of endometrial cancer: a case-control study and meta-analysis. Eur J Cancer Prev. 2013;22(1):38–45. doi: 10.1097/CEJ.0b013e328354d378. [DOI] [PubMed] [Google Scholar]
- 28.Gnagnarella P, Gandini S, La Vecchia C, et al. Glycemic index, glycemic load, and cancer risk: a meta-analysis. Am J Clin Nutr. 2008;87(6):1793–1801. doi: 10.1093/ajcn/87.6.1793. [DOI] [PubMed] [Google Scholar]
- 29.Nagle CM, Olsen CM, Ibiebele TI, et al. Glycemic index, glycemic load and endometrial cancer risk: results from the Australian National Endometrial Cancer Study and an updated systematic review and meta-analysis. Eur J Nutr. 2013;52(2):705–715. doi: 10.1007/s00394-012-0376-7. [DOI] [PubMed] [Google Scholar]
- 30.Jain MG, Rohan TE, Howe GR, et al. A cohort study of nutritional factors and endometrial cancer. Eur J Epidemiol. 2000;16(10):899–905. doi: 10.1023/a:1011012621990. [DOI] [PubMed] [Google Scholar]
- 31.Xu WH, Dai Q, Xiang YB, et al. Nutritional factors in relation to endometrial cancer: a report from a population-based case-control study in Shanghai, China. Int J Cancer. 2007;120(8):1776–1781. doi: 10.1002/ijc.22456. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Friberg E, Wallin A, Wolk A. Sucrose, high-sugar foods, and risk of endometrial cancer—a population-based cohort study. Cancer Epidemiol Biomarkers Prev. 2011;20(9):1831–1837. doi: 10.1158/1055-9965.EPI-11-0402. [DOI] [PubMed] [Google Scholar]
- 33.Potischman N, Swanson CA, Brinton LA, et al. Dietary associations in a case-control study of endometrial cancer. Cancer Causes Control. 1993;4(3):239–250. doi: 10.1007/BF00051319. [DOI] [PubMed] [Google Scholar]
- 34.Littman AJ, Beresford SA, White E. The association of dietary fat and plant foods with endometrial cancer (United States) Cancer Causes Control. 2001;12(8):691–702. doi: 10.1023/a:1011292003586. [DOI] [PubMed] [Google Scholar]
- 35.McCann SE, Freudenheim JL, Marshall JR, et al. Diet in the epidemiology of endometrial cancer in western New York (United States) Cancer Causes Control. 2000;11(10):965–974. doi: 10.1023/a:1026551309873. [DOI] [PubMed] [Google Scholar]
- 36.Du H, Van der ADL, Feskens EJ. Dietary glycaemic index: a review of the physiological mechanisms and observed health impacts. Acta Cardiol. 2006;61(4):383–397. doi: 10.2143/AC.61.4.2017298. [DOI] [PubMed] [Google Scholar]
- 37.Kaaks R, Lukanova A, Kurzer MS. Obesity, endogenous hormones, and endometrial cancer risk: a synthetic review. Cancer Epidemiol Biomarkers Prev. 2002;11(12):1531–1543. [PubMed] [Google Scholar]
- 38.Gunter MJ, Hoover DR, Yu H, et al. A prospective evaluation of insulin and insulin-like growth factor-I as risk factors for endometrial cancer. Cancer Epidemiol Biomarkers Prev. 2008;17(4):921–929. doi: 10.1158/1055-9965.EPI-07-2686. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Biddinger SB, Ludwig DS. The insulin-like growth factor axis: a potential link between glycemic index and cancer. Am J Clin Nutr. 2005;82(2):277–278. doi: 10.1093/ajcn.82.2.277. [DOI] [PubMed] [Google Scholar]
- 40.Friedenreich C, Cust A, Lahmann PH, et al. Anthropometric factors and risk of endometrial cancer: the European Prospective Investigation into Cancer and Nutrition. Cancer Causes Control. 2007;18(4):399–413. doi: 10.1007/s10552-006-0113-8. [DOI] [PubMed] [Google Scholar]
- 41.Furberg AS, Thune I. Metabolic abnormalities (hypertension, hyperglycemia and overweight), lifestyle (high energy intake and physical inactivity) and endometrial cancer risk in a Norwegian cohort. Int J Cancer. 2003;104(6):669–676. doi: 10.1002/ijc.10974. [DOI] [PubMed] [Google Scholar]
- 42.Schouten LJ, Goldbohm RA, van den Brandt PA. Anthropometry, physical activity, and endometrial cancer risk: results from the Netherlands Cohort Study. J Natl Cancer Inst. 2004;96(21):1635–1638. doi: 10.1093/jnci/djh291. [DOI] [PubMed] [Google Scholar]
- 43.Black AE, Prentice AM, Goldberg GR, et al. Measurements of total energy expenditure provide insights into the validity of dietary measurements of energy intake. J Am Diet Assoc. 1993;93(5):572–579. doi: 10.1016/0002-8223(93)91820-g. [DOI] [PubMed] [Google Scholar]
- 44.Voskuil DW, Monninkhof EM, Elias SG, et al. Physical activity and endometrial cancer risk, a systematic review of current evidence. Cancer Epidemiol Biomarkers Prev. 2007;16(4):639–648. doi: 10.1158/1055-9965.EPI-06-0742. [DOI] [PubMed] [Google Scholar]
- 45.Merrill RM. Hysterectomy surveillance in the United States, 1997 through 2005. Med Sci Monit. 2008;14(1):CR24–CR31. [PubMed] [Google Scholar]
- 46.Whiteman MK, Hillis SD, Jamieson DJ, et al. Inpatient hysterectomy surveillance in the United States, 2000–2004. Am J Obstet Gynecol. 2008;198(1):34.e1–34.e7. doi: 10.1016/j.ajog.2007.05.039. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.