Abstract

Childbearing at an older age has been associated with a lower risk of endometrial cancer, but whether the association is independent of the number of births or other factors remains unclear. Individual-level data from 4 cohort and 13 case-control studies in the Epidemiology of Endometrial Cancer Consortium were pooled. A total of 8,671 cases of endometrial cancer and 16,562 controls were included in the analysis. After adjustment for known risk factors, endometrial cancer risk declined with increasing age at last birth (Ptrend < 0.0001). The pooled odds ratio per 5-year increase in age at last birth was 0.87 (95% confidence interval: 0.85, 0.90). Women who last gave birth at 40 years of age or older had a 44% decreased risk compared with women who had their last birth under the age of 25 years (95% confidence interval: 47, 66). The protective association was similar across the different age-at-diagnosis groups and for the 2 major tumor histologic subtypes (type I and type II). No effect modification was observed by body mass index, parity, or exogenous hormone use. In this large pooled analysis, late age at last birth was independently associated with a reduced risk of endometrial cancer, and the reduced risk persisted for many years.

Exposure of the endometrium to estrogen “unopposed” by progesterone is a common cause of endometrial cancer (1). Pregnancy results in large changes in endogenous estrogen and progesterone levels, and for this reason it is plausible that pregnancy influences the incidence of endometrial cancer. Most epidemiologic studies have observed that nulliparity is associated with an elevated risk of endometrial cancer, whereas an increasing number of births is associated with a reduction in risk (1). Later age at last birth (≥35 years of age) is associated with a lower risk of endometrial cancer (2–10), but whether the association is independent of the number of births or other known risk factors (e.g., obesity, exogenous hormone use) requires clarification. In the present study, we combined individual-level data from 17 epidemiologic studies and performed a pooled analysis of 8,671 endometrial cancer cases and 16,562 controls to examine how age at last birth affects the risk of endometrial cancer and how this is influenced by other risk factors for the disease. The large number of cases and controls also allowed us to evaluate associations by tumor subtype.

MATERIALS AND METHODS

Participating studies

Seventeen studies (4 cohort and 13 case-control) in the Epidemiology of Endometrial Cancer Consortium (E2C2) that collected information on age at last birth or age at last pregnancy were included in the pooled analysis (3, 11–27). The E2C2 is an international consortium established to pool resources and data from many endometrial cancer studies in an effort to identify genetic and environmental risk factors for endometrial cancer (28). Cohort studies were analyzed as nested case-control studies, with up to 4 controls randomly selected (among women with an intact uterus and without endometrial cancer before the diagnosis of the index case) for each case based on exact year of birth, date of cohort entry (±6 months), and other criteria as appropriate for each individual study (e.g., race/ethnicity, study area, etc.). The characteristics of the individual studies and the matching variables for both the cohort and case-control studies are shown in Table 1. For most studies, the large majority of participants were non-Hispanic whites. Two studies, the Hawaii Endometrial Cancer Study and the Shanghai Endometrial Cancer Study, included mainly or exclusively nonwhite populations. The number of endometrial cancer cases in each study ranged from 232 to 1,396. Informed consent was obtained from all study participants as part of the original studies and in accordance with each study's institutional review board.

Table 1.

Characteristics of the 17 Studies Included in This Pooled Analysis

Study, Year (Reference No.) Location Recruitment Period Matching Factors White Race, % No. of Cases No. of Controls Mean Age, yearsa Measure Used
 
Age at Last Birth or Full Term Pregnancyb Age at Last Pregnancyc 
Cohort studies 
 Breast Cancer Detection Demonstration Project, 2005 (17) 29 US clinics 1979–1980 Birth year, cohort entry, race, and clinic 92 664 2,418 65.4 Yes  
 California Teachers Study, 2010 (13) California 1995–1996 Birth year, cohort entry, race, and area 91 757 3,010 65.7  Yes 
 Iowa Women's Health Study, 1996 (21) Iowa 1986 Birth year, cohort entry, race, and area 98 553 2,212 71.6 Yes  
 Nurses Health Study, 2010 (3) 11 US states 1976 Birth year, cohort entry, race, and area 95 650 1,641 60.8 Yes  
Case-control studies 
 Australian National Endometrial Cancer Study, 2011 (25) Australia 2005–2007 Age (in 5-year increments) and state 88 1,396 740 61.2 Yes  
 Bay Area Women's Health Study, 2003 (16) California 1996–1999 Age (in 5-year increments) and ethnicity 89 500 470 61.6 Yes  
 Connecticut Endometrial Cancer Study, 2011 (18) Connecticut 2004–2009 Age (in 5-year increments) 93 668 665 61.0 Yes  
 Estrogen, Diet, Genetics, and Endometrial Cancer Study, 2008 (22) New Jersey 2001–2005 Age (in 5-year increments) 90 469 467 62.6 Yes  
 Fred Hutchinson Cancer Research Center Study, 2009 (27) Washington 1994–2005 Age (in 5-year increments) 95 881 865 59.6 Yes  
 Hawaii Endometrial Cancer Study, 1997 (15) Hawaii 1988–1993 Age (±2.5 years) and ethnicity 24 471 511 57.2 Yes  
 Polish Endometrial Cancer Study, 2007 (12) Poland 2000–2003 Age (±5 years) and site 100 551 1,925 57.2 Yes  
 Patient Epidemiologic Data System Study, 2009 (20) New York 1982–1998 Age (±5 years) 97 541 541 62.7  Yes 
 Shanghai Endometrial Cancer Study, 2007 (26) China 1997–2003 Age (±5 years) 1,199 1,212 54.5 Yes  
 US Endometrial Cancer Study, 1992 (11) 5 US clinics 1987–1990 Age (±5 years), race, and telephone exchange 92 433 320 58.7 Yes  
 Case-control study conducted by USC, 1997 (23) Los Angeles 1987–1993 Age (±5 years) 100 833 791 63.1 Yes  
 Women's Insight and Shared Experience Study, 2006 (24) Philadelphia 1999–2002 Age (±5 years) and race 79 616 1,583 61.8 Yes  
 Western New York Diet Study, 2000 (19) New York 1986–1991 Age (±5 years) and county of residence 100 232 639 57.9 Yes  
Study, Year (Reference No.) Location Recruitment Period Matching Factors White Race, % No. of Cases No. of Controls Mean Age, yearsa Measure Used
 
Age at Last Birth or Full Term Pregnancyb Age at Last Pregnancyc 
Cohort studies 
 Breast Cancer Detection Demonstration Project, 2005 (17) 29 US clinics 1979–1980 Birth year, cohort entry, race, and clinic 92 664 2,418 65.4 Yes  
 California Teachers Study, 2010 (13) California 1995–1996 Birth year, cohort entry, race, and area 91 757 3,010 65.7  Yes 
 Iowa Women's Health Study, 1996 (21) Iowa 1986 Birth year, cohort entry, race, and area 98 553 2,212 71.6 Yes  
 Nurses Health Study, 2010 (3) 11 US states 1976 Birth year, cohort entry, race, and area 95 650 1,641 60.8 Yes  
Case-control studies 
 Australian National Endometrial Cancer Study, 2011 (25) Australia 2005–2007 Age (in 5-year increments) and state 88 1,396 740 61.2 Yes  
 Bay Area Women's Health Study, 2003 (16) California 1996–1999 Age (in 5-year increments) and ethnicity 89 500 470 61.6 Yes  
 Connecticut Endometrial Cancer Study, 2011 (18) Connecticut 2004–2009 Age (in 5-year increments) 93 668 665 61.0 Yes  
 Estrogen, Diet, Genetics, and Endometrial Cancer Study, 2008 (22) New Jersey 2001–2005 Age (in 5-year increments) 90 469 467 62.6 Yes  
 Fred Hutchinson Cancer Research Center Study, 2009 (27) Washington 1994–2005 Age (in 5-year increments) 95 881 865 59.6 Yes  
 Hawaii Endometrial Cancer Study, 1997 (15) Hawaii 1988–1993 Age (±2.5 years) and ethnicity 24 471 511 57.2 Yes  
 Polish Endometrial Cancer Study, 2007 (12) Poland 2000–2003 Age (±5 years) and site 100 551 1,925 57.2 Yes  
 Patient Epidemiologic Data System Study, 2009 (20) New York 1982–1998 Age (±5 years) 97 541 541 62.7  Yes 
 Shanghai Endometrial Cancer Study, 2007 (26) China 1997–2003 Age (±5 years) 1,199 1,212 54.5 Yes  
 US Endometrial Cancer Study, 1992 (11) 5 US clinics 1987–1990 Age (±5 years), race, and telephone exchange 92 433 320 58.7 Yes  
 Case-control study conducted by USC, 1997 (23) Los Angeles 1987–1993 Age (±5 years) 100 833 791 63.1 Yes  
 Women's Insight and Shared Experience Study, 2006 (24) Philadelphia 1999–2002 Age (±5 years) and race 79 616 1,583 61.8 Yes  
 Western New York Diet Study, 2000 (19) New York 1986–1991 Age (±5 years) and county of residence 100 232 639 57.9 Yes  

a Age at diagnosis for cases and age at interview or reference date for controls.

b Longer than 6 months, live or still birth.

c Any length of pregnancy.

Data collection

Data from individual studies were received at the E2C2 data coordinating center at Memorial Sloan-Kettering Cancer Center with personal identifiers removed. Each data set was checked for inconsistencies and completeness. Queries were sent to the investigators to resolve inconsistencies and any other data issues. Data on age at last birth (or full-term pregnancy, i.e., pregnancy lasting more than 6 months) were provided by 15 of the 17 studies; 2 studies (the California Teachers Study and the Patient Epidemiologic Data System Study) only provided data on age at last pregnancy. These variables were considered the same in the main analysis, but results are also given excluding these 2 studies. Each study also provided information regarding tumor characteristics, demographic variables, and known/potential risk factors and covariates (Table 2). These variables were defined and uniformly recoded in accordance with the E2C2 data dictionary (available upon request).

Table 2.

Characteristicsa of Endometrial Cancer Cases and Controls in the Pooled Analysis

 No. of Cases (n= 8,671)
 
No. of Controls (n= 16,562)
 
 Mean (SD) % Mean (SD) % 
Age, yearsb 61.9 (9.1)  62.1 (10.1)  
Body mass indexc 28.8 (7.4)  25.6 (5.0)  
Race 
 Non-Hispanic white  80.1  84.9 
 Black  2.2  3.4 
 Asian  15.0  9.7 
 Hawaiian/Pacific Islander  0.5  0.5 
 Otherd  2.3  1.5 
Body mass indexc 
  <25  36.5  54.0 
 25–<30  28.7  29.8 
  ≥30  34.8  16.3 
Age at menarche, years 
  <12  20.0  15.8 
 12  23.4  21.6 
 13  26.6  27.5 
 14  14.7  16.1 
  ≥15  14.8  18.5 
 Missing/unknown  0.4  0.5 
Parity 
 1  20.7  18.1 
 2  35.6  34.4 
 3  24.3  24.6 
  ≥4  19.5  22.9 
Menopausal statuse 
 Premenopausal  12.0  14.9 
 Perimenopausal  0.2  0.2 
 Postmenopausal  66.5  63.6 
 Missing/unknown  21.4  21.4 
Menopausal hormone use 
 Never  60.6  63.0 
 Ever  36.4  33.9 
 Missing/unknown  2.9  3.1 
Menopausal estrogen-only usee 
 Never  66.5  70.1 
 Ever  17.1  12.4 
 Missing/unknown  16.4  17.4 
Oral contraceptive use 
 Never  60.8  53.0 
 Ever  38.7  46.6 
 Missing/unknown  0.5  0.4 
Diabetese 
 No  76.1  82.4 
 Yes  14.3  7.1 
 Missing/unknown  9.6  10.5 
Smokingf 
 Never  61.5  55.5 
 Past  25.7  27.1 
 Current  8.4  13.0 
 Missing/unknown  4.5  4.4 
 No. of Cases (n= 8,671)
 
No. of Controls (n= 16,562)
 
 Mean (SD) % Mean (SD) % 
Age, yearsb 61.9 (9.1)  62.1 (10.1)  
Body mass indexc 28.8 (7.4)  25.6 (5.0)  
Race 
 Non-Hispanic white  80.1  84.9 
 Black  2.2  3.4 
 Asian  15.0  9.7 
 Hawaiian/Pacific Islander  0.5  0.5 
 Otherd  2.3  1.5 
Body mass indexc 
  <25  36.5  54.0 
 25–<30  28.7  29.8 
  ≥30  34.8  16.3 
Age at menarche, years 
  <12  20.0  15.8 
 12  23.4  21.6 
 13  26.6  27.5 
 14  14.7  16.1 
  ≥15  14.8  18.5 
 Missing/unknown  0.4  0.5 
Parity 
 1  20.7  18.1 
 2  35.6  34.4 
 3  24.3  24.6 
  ≥4  19.5  22.9 
Menopausal statuse 
 Premenopausal  12.0  14.9 
 Perimenopausal  0.2  0.2 
 Postmenopausal  66.5  63.6 
 Missing/unknown  21.4  21.4 
Menopausal hormone use 
 Never  60.6  63.0 
 Ever  36.4  33.9 
 Missing/unknown  2.9  3.1 
Menopausal estrogen-only usee 
 Never  66.5  70.1 
 Ever  17.1  12.4 
 Missing/unknown  16.4  17.4 
Oral contraceptive use 
 Never  60.8  53.0 
 Ever  38.7  46.6 
 Missing/unknown  0.5  0.4 
Diabetese 
 No  76.1  82.4 
 Yes  14.3  7.1 
 Missing/unknown  9.6  10.5 
Smokingf 
 Never  61.5  55.5 
 Past  25.7  27.1 
 Current  8.4  13.0 
 Missing/unknown  4.5  4.4 

aRisk factor distributions were standardized to the number of cases in each study.

bAge at diagnosis for cases and age at interview/reference date for controls.

cWeight (kg)/height (m)2.

dIncluding mixed, Hispanic whites, other, and unknown.

e Available for 15 studies.

f Available for 16 studies.

Tumor histology

Only incident cases of endometrial cancer (primary site codes C54 and C55.9) were included in this analysis. Twelve studies (the Australian National Endometrial Cancer Study, Bay Area Women's Health Study, Breast Cancer Detection Demonstration Project, Connecticut Endometrial Cancer Study, California Teachers Study, Estrogen, Diet, Genetics, and Endometrial Cancer Study, Fred Hutchinson Cancer Research Center Study, Hawaii Endometrial Cancer Study, Iowa Women's Health Study, Patient Epidemiologic Data System Study, US Endometrial Cancer Study, and a case-control study conducted by USC) provided the International Classification of Diseases for Oncology, Third Edition, histology codes for each case. Five studies (the Nurses' Health Study, Women's Insight and Shared Experience Study, Shanghai Endometrial Cancer Study, Polish Endometrial Cancer Study, and Western New York Diet Study) provided a summary histologic type for each case (i.e., endometrioid, serous, clear cell, adenocarcinoma not otherwise specified, etc). Unopposed estrogen use is suspected to affect the risk of type I tumors but not type II tumors (29). On the basis of widely accepted guidelines (30–32), we classified endometrial cancer cases as type I or type II. Type I tumors included endometrioid (codes 8380, 8381, 8382, and 8383), adenocarcinoma tubular (codes 8210 and 8211), papillary adenocarcinoma (codes 8260, 8262, and 8263), adenocarcinoma with squamous metaplasia (code 8570), mucinous adenocarcinoma (codes 8480 and 8481), and adenocarcinoma not otherwise specified (code 8140) tumors. Type II tumors included serous (code 8441), papillary serous (codes 8460 and 8461), squamous cell (codes 8050, 8070, 8071, and 8072), adenosquamous (code 8560), small-cell carcinoma (code 8041), and mixed-cell adenocarcinoma (code 8323) tumors. We excluded 175 women with a sarcoma diagnosis from the present analysis.

Exclusion criteria

Women who were nulliparous (2,168 cases and 2,561 controls) and women for whom we were missing information on parity (83 cases and 125 controls) or age at last birth, full-term pregnancy, or last pregnancy (248 cases and 550 controls) were excluded from the analyses. Women with missing body mass index (BMI, measured as weight in kilograms divided by height in meters squared) values (185 cases and 280 controls) were also excluded from the analysis because BMI is an important potential confounder. After these exclusions, 8,671 endometrial cancer cases and 16,562 controls remained for analysis.

Statistical methods

Age at last birth was categorized as <25, 25–29, 30–34, 35–39, or ≥40 years. Covariates considered in the present analysis included age (at diagnosis for cases and at interview or reference date for controls), race/ethnicity, BMI, age at menarche, parity, age at first birth, menopausal status, menopausal hormone use (any kind), menopausal estrogen use, oral contraceptive (OC) use, smoking status, and diabetes mellitus. Data on age, race/ethnicity, BMI, age at menarche, parity, menopausal hormone, and OC use were available from all 17 studies. Menopausal status was not available from the Fred Hutchinson Cancer Research Center Study or the Australian National Endometrial Cancer Study, and data on menopausal estrogen use were not available from the Patient Epidemiologic Data System Study or the Iowa Women's Health Study. Smoking status was not available in the Bay Area Women's Health Study, and diabetes information was not available in either the Bay Area Women's Health Study or the Patient Epidemiologic Data System Study. We created standardized categories for the covariates: age (<50, 50–54, 55–59, 60–64, 65–69, or ≥70 years), study type (cohort or case-control), race/ethnicity (non-Hispanic white, African American/black, Asian, Hawaiian/Pacific Islander, or other), BMI (<25, 25–29, or ≥30), age at menarche (<12, 12, 13, 14, or ≥15 years), age at first birth (<25, 25–29, or ≥30 years), parity (1, 2, 3, or ≥4 children), OC use (never or ever), menopausal hormone use (never or ever), menopausal estrogen use (never or ever), diabetes (no or yes), and smoking status (never, past, or current). Categories for missing data were created for age at menarche, OC use, menopausal hormone use, diabetes, and smoking status. The association between endometrial cancer and age at last birth in each study was assessed by calculating odds ratios and 95% confidence intervals using logistic regression stratified by age × race/ethnicity, and when data were available, they were adjusted for BMI, age at menarche, parity, OC use, menopausal hormone use, diabetes, and smoking status. We estimated the combined odds ratio and 95% confidence interval for age at last birth from these individual study results using a random-effects model and tested for the statistical significance of between-study heterogeneity using the Q test statistics (33, 34). No asymmetry was observed in the funnel plot of the effect estimates from individual studies (data not shown).

We also analyzed the complete individual data using a pooled analysis stratified by study × age × race/ethnicity and adjusted for BMI, age at menarche, parity, OC use, menopausal hormone use, diabetes, and smoking status. Tests for trend were performed by entering the ordinal values representing categories of age at last birth as a continuous variable in the models. We investigated effect modification of the association between age at last birth and endometrial cancer risk by BMI, parity, OC use, and menopausal estrogen use. Interaction analysis with menopausal estrogen use was limited to postmenopausal women in the 13 studies with information on both menopausal status and menopausal estrogen use. Women who had ever used estrogen alone were compared with women who had never used any type of menopausal hormone. Tests of interaction were assessed using log-likelihood test statistics comparing models with and without the interaction term. We investigated how time since last birth affected the association of age at last birth with endometrial cancer by estimating the effect of age at last birth in increasing age at diagnosis categories. We also examined the effect of age at last birth for the 2 major tumor histologic classifications (type I and type II) using a polytomous logistic regression model and tested the null hypothesis that there was no difference in the pooled odds ratios between type I and type II tumors. All statistical tests (P values quoted) were 2-sided. Statistical analyses were performed using Stata, version 11.0 (StataCorp LP, College Station, Texas) and SAS, version 9.2 (SAS Institute, Inc., Cary, North Carolina).

RESULTS

Characteristics of the 8,671 cases and 16,562 controls included in the pooled analysis are shown in Table 2. The majority of women were non-Hispanic white (80.1% of cases and 84.9% of controls) and postmenopausal (66.5% of cases and 63.6% of controls). On average, cases had higher BMIs than did controls. Cases were more likely to have had an earlier age at menarche, had fewer births, have used menopausal hormones, and be diabetic. Cases were less likely to have used OCs or to have ever smoked.

A forest plot of the study-specific risk estimates for endometrial cancer per 5-year increase in age at last birth is shown in Figure 1. All but one study (US Endometrial Cancer Study) reported an inverse association. The combined odds ratio per 5-year increase (OR5) in age at last birth and risk of endometrial cancer was 0.88 (95% CI: 0.85, 0.91), with no significant heterogeneity observed among studies (P = 0.12). When we excluded the 2 studies (California Teachers Study and Patient Epidemiologic Data System Study) with data on age at last pregnancy of any length rather than age at last birth from the pooled analysis, the combined OR5 was unchanged (OR5 = 0.88, 95% CI: 0.84, 0.92).

Figure 1.

The risk of endometrial cancer associated with a 5-year increase in age at last birth by study. The numbers of cases and controls shown in each study were after all exclusions had been applied. Odds ratios were stratified by age and race/ethnicity and, when possible, were adjusted for parity, age at menarche, body mass index, oral contraceptive use, menopausal hormone use, smoking, and diabetes. The combined odds ratio per 5-year increase of 0.88 (95% confidence interval: 0.85, 0.91) was calculated using a random-effects model with a P for heterogeneity of 0.12. ANECS, Australian National Endometrial Cancer Study; BAWHS, Bay Area Women's Health Study; BCDDP, Breast Cancer Detection Demonstration Project; CECS, Connecticut Endometrial Cancer Study; CI, confidence interval; CTS, California Teachers Study; EDGE, Estrogen, Diet, Genetics, and Endometrial Cancer Study; FHCRC, Fred Hutchinson Cancer Research Center Study; HAW, Hawaii Endometrial Cancer Study; IWHS, Iowa Women's Health Study; NHS, Nurses' Health Study; PECS, Polish Endometrial Cancer Study; PEDS, Patient Epidemiologic Data System Study; SECS, Shanghai Endometrial Cancer Study; US, US Endometrial Cancer Study; USC, case-control study conducted by USC; WISE, Women's Insight and Shared Experience Study; and WNYDS, Western New York Diet Study.

Figure 1.

The risk of endometrial cancer associated with a 5-year increase in age at last birth by study. The numbers of cases and controls shown in each study were after all exclusions had been applied. Odds ratios were stratified by age and race/ethnicity and, when possible, were adjusted for parity, age at menarche, body mass index, oral contraceptive use, menopausal hormone use, smoking, and diabetes. The combined odds ratio per 5-year increase of 0.88 (95% confidence interval: 0.85, 0.91) was calculated using a random-effects model with a P for heterogeneity of 0.12. ANECS, Australian National Endometrial Cancer Study; BAWHS, Bay Area Women's Health Study; BCDDP, Breast Cancer Detection Demonstration Project; CECS, Connecticut Endometrial Cancer Study; CI, confidence interval; CTS, California Teachers Study; EDGE, Estrogen, Diet, Genetics, and Endometrial Cancer Study; FHCRC, Fred Hutchinson Cancer Research Center Study; HAW, Hawaii Endometrial Cancer Study; IWHS, Iowa Women's Health Study; NHS, Nurses' Health Study; PECS, Polish Endometrial Cancer Study; PEDS, Patient Epidemiologic Data System Study; SECS, Shanghai Endometrial Cancer Study; US, US Endometrial Cancer Study; USC, case-control study conducted by USC; WISE, Women's Insight and Shared Experience Study; and WNYDS, Western New York Diet Study.

Table 3 shows the results of the pooled analysis of complete individual data. The OR5 in all women was 0.87 (95% CI: 0.85, 0.90), essentially the same as the OR5 estimated using meta-analysis. Table 3 also shows the results of examining age at last birth categorically. Women who had their last birth at 40 years of age or older were at a 44% decreased risk of endometrial cancer (OR = 0.56, 95% CI: 0.47, 0.66) than were women who had their last child before 25 years of age. We further examined this association among women with 2 or more births both with and without adjustment for age at first birth (categorical and linear term) and also stratified by age at first birth categories; results were similar (Web Tables 1 and 2, available at http://aje.oxfordjournals.org/).

Table 3.

Association Between Age at Last Birth and Endometrial Cancer by Age at Diagnosis in the Pooled Analysis

Age at Last Birth, years No. of Cases No. of Controls Odds Ratio 95% Confidence Interval 
All Agesa 
<25 1,401 2,189 1.00 Referent 
25–29 3,208 5,211 0.95 0.87, 1.05 
30–34 2,601 5,295 0.83 0.76, 0.92 
35–39 1,166 2,983 0.68 0.61, 0.76 
≥40 295 884 0.56 0.47, 0.66 
Ptrend    <0.0001 
 Per 5-year increase   0.87 0.85, 0.90 
Age <50 Yearsb 
<25 131 276 1.00 Referent 
25–29 360 698 0.84 0.62, 1.15 
30–34 184 461 0.75 0.53, 1.05 
35–39 51 199 0.46 0.29, 0.72 
≥40 10 42 0.40 0.18, 0.92 
Ptrend   0.0003  
 Per 5-year increase   0.81 0.72, 0.91 
Age 50–59 Yearsb 
<25 496 744 1.00 Referent 
25–29 1,050 1,643 0.95 0.81, 1.11 
30–34 803 1,581 0.80 0.68, 0.94 
35–39 301 743 0.64 0.52, 0.79 
≥40 65 190 0.55 0.39, 0.77 
Ptrend   <0.0001  
 Per 5-year increase   0.87 0.82, 0.91 
Age 60–69 Yearsb 
<25 574 798 1.00 Referent 
25–29 1,310 1,903 0.99 0.85, 1.14 
30–34 1,058 1,900 0.89 0.76, 1.03 
35–39 492 1,118 0.75 0.63, 0.89 
≥40 102 315 0.48 0.36, 0.64 
Ptrend   <0.0001  
 Per 5-year increase   0.88 0.84, 0.92 
Age ≥70 Yearsb 
<25 200 371 1.00 Referent 
25–29 488 967 0.94 0.74, 1.18 
30–34 556 1,353 0.83 0.66, 1.04 
35–39 322 923 0.68 0.53, 0.87 
≥40 118 337 0.67 0.49, 0.92 
Ptrend   0.0001  
 Per 5-year increase   0.89 0.83, 0.94 
Age at Last Birth, years No. of Cases No. of Controls Odds Ratio 95% Confidence Interval 
All Agesa 
<25 1,401 2,189 1.00 Referent 
25–29 3,208 5,211 0.95 0.87, 1.05 
30–34 2,601 5,295 0.83 0.76, 0.92 
35–39 1,166 2,983 0.68 0.61, 0.76 
≥40 295 884 0.56 0.47, 0.66 
Ptrend    <0.0001 
 Per 5-year increase   0.87 0.85, 0.90 
Age <50 Yearsb 
<25 131 276 1.00 Referent 
25–29 360 698 0.84 0.62, 1.15 
30–34 184 461 0.75 0.53, 1.05 
35–39 51 199 0.46 0.29, 0.72 
≥40 10 42 0.40 0.18, 0.92 
Ptrend   0.0003  
 Per 5-year increase   0.81 0.72, 0.91 
Age 50–59 Yearsb 
<25 496 744 1.00 Referent 
25–29 1,050 1,643 0.95 0.81, 1.11 
30–34 803 1,581 0.80 0.68, 0.94 
35–39 301 743 0.64 0.52, 0.79 
≥40 65 190 0.55 0.39, 0.77 
Ptrend   <0.0001  
 Per 5-year increase   0.87 0.82, 0.91 
Age 60–69 Yearsb 
<25 574 798 1.00 Referent 
25–29 1,310 1,903 0.99 0.85, 1.14 
30–34 1,058 1,900 0.89 0.76, 1.03 
35–39 492 1,118 0.75 0.63, 0.89 
≥40 102 315 0.48 0.36, 0.64 
Ptrend   <0.0001  
 Per 5-year increase   0.88 0.84, 0.92 
Age ≥70 Yearsb 
<25 200 371 1.00 Referent 
25–29 488 967 0.94 0.74, 1.18 
30–34 556 1,353 0.83 0.66, 1.04 
35–39 322 923 0.68 0.53, 0.87 
≥40 118 337 0.67 0.49, 0.92 
Ptrend   0.0001  
 Per 5-year increase   0.89 0.83, 0.94 

aAge, study, and race/ethnicity stratified and adjusted for age at menarche, parity, body mass index, oral contraceptive use, menopausal hormone use, smoking, and diabetes.

bStudy and race/ethnicity stratified and adjusted for age at menarche, parity, body mass index, oral contraceptive use, menopausal hormone use, smoking, and diabetes.

The associations between age at last birth and risk of endometrial cancer were similar across the 4 age-at-diagnosis/reference groups (<50, 50–59, 60–69, or ≥70 years) (Table 3). A significant trend was observed in all age groups (Ptrend ≤ 0.0003). The OR5 was 0.81 (95% CI: 0.72, 0.91) for women younger than 50 years of age, 0.87 (95% CI: 0.82, 0.91) for women aged 50–59 years, 0.88 (95% CI: 0.84, 0.92) for women aged 60–69 years, and 0.89 (95% CI: 0.83, 0.94) for women aged 70 years or older. The test for interaction between age group (at diagnosis/reference) and age at last birth was not statistically significant (P = 0.67).

We examined the race/ethnicity-specific association between age at last birth and endometrial cancer in the 3 major racial/ethnic groups: whites (6,946 cases and 14,059 controls), blacks (186 cases and 566 controls), and Asians (1,297 cases and 1,598 controls) (Web Table 3). An inverse association between increasing age at last birth and endometrial cancer risk was observed in whites (OR5 = 0.87, 95% CI: 0.84, 0.89; Ptrend < 0.0001) and to a lesser extent in Asians (OR5 = 0.92, 95% CI: 0.84, 1.02; Ptrend = 0.09). No association was observed in the small subset of black women (OR5 = 0.98, 95% CI: 0.83, 1.16; Ptrend = 0.70). The P value for test for interaction between race/ethnicity and age at last birth (per 5-year increase) was 0.025.

We also examined the association between age at last birth and endometrial cancer by study type (cohort vs. case-control). The associations in the cohort studies (OR5 = 0.90, 95% CI: 0.85, 0.95; Ptrend < 0.0001) and in the case-control studies (OR5 = 0.86, 95% CI: 0.83, 0.90; Ptrend < 0.0001) were similar.

Table 4 shows the association between age at last birth and endometrial cancer by parity. A significant inverse association with increasing age at last birth was observed among women with at least 2 births (Ptrend < 0.0001). Among uniparous women, we observed a modest inverse association with increasing age at last birth (OR5 = 0.92, 95% CI: 0.87, 0.98; Ptrend = 0.058). The test of interaction was not statistically significant across the 4 parity groups (Pinteraction = 0.09) or when we limited the analysis to multiparous women (Pinteraction = 0.30).

Table 4.

Association Between Age at Last Birth and Endometrial Cancer by Parity in the Pooled Analysis

Age at Last Birth, years No. of Cases No. of Controls Odds Ratioa 95% Confidence Interval 
Women With 1 Birth 
<25 540 861 1.00  
25–29 676 1,022 0.83 0.69, 0.98 
30–34 370 602 0.84 0.69, 1.02 
35–39 163 300 0.82 0.63, 1.05 
≥40 42 82 0.77 0.48, 1.22 
Ptrend    0.058 
 Per 5-year increase   0.92 0.87, 0.98 
Women With 2 Births 
<25 584 925 1.00  
25–29 1,287 2,241 1.00 0.87, 1.15 
30–34 873 1,777 0.88 0.76, 1.03 
35–39 288 691 0.71 0.58, 0.86 
≥40 51 168 0.48 0.33, 0.69 
Ptrend   <0.0001  
 Per 5-year increase   0.87 0.84, 0.93 
Women With 3 Births 
<25 216 307 1.00  
25–29 808 1,264 0.93 0.75, 1.16 
30–34 738 1,512 0.77 0.61, 0.96 
35–39 278 738 0.61 0.48, 0.79 
≥40 63 166 0.57 0.39, 0.84 
Ptrend   <0.0001  
 Per 5-year increase   0.85 0.79, 0.90 
Women With ≥4 Births 
<25 61 96 1.00  
25–29 437 684 1.14 0.76, 1.71 
30–34 620 1,404 0.88 0.59, 1.30 
35–39 437 1,254 0.74 0.50, 1.12 
≥40 139 468 0.59 0.38, 0.92 
Ptrend   <0.0001  
 Per 5-year increase   0.83 0.78, 0.89 
Age at Last Birth, years No. of Cases No. of Controls Odds Ratioa 95% Confidence Interval 
Women With 1 Birth 
<25 540 861 1.00  
25–29 676 1,022 0.83 0.69, 0.98 
30–34 370 602 0.84 0.69, 1.02 
35–39 163 300 0.82 0.63, 1.05 
≥40 42 82 0.77 0.48, 1.22 
Ptrend    0.058 
 Per 5-year increase   0.92 0.87, 0.98 
Women With 2 Births 
<25 584 925 1.00  
25–29 1,287 2,241 1.00 0.87, 1.15 
30–34 873 1,777 0.88 0.76, 1.03 
35–39 288 691 0.71 0.58, 0.86 
≥40 51 168 0.48 0.33, 0.69 
Ptrend   <0.0001  
 Per 5-year increase   0.87 0.84, 0.93 
Women With 3 Births 
<25 216 307 1.00  
25–29 808 1,264 0.93 0.75, 1.16 
30–34 738 1,512 0.77 0.61, 0.96 
35–39 278 738 0.61 0.48, 0.79 
≥40 63 166 0.57 0.39, 0.84 
Ptrend   <0.0001  
 Per 5-year increase   0.85 0.79, 0.90 
Women With ≥4 Births 
<25 61 96 1.00  
25–29 437 684 1.14 0.76, 1.71 
30–34 620 1,404 0.88 0.59, 1.30 
35–39 437 1,254 0.74 0.50, 1.12 
≥40 139 468 0.59 0.38, 0.92 
Ptrend   <0.0001  
 Per 5-year increase   0.83 0.78, 0.89 

a Age, study, and race/ethnicity stratified and adjusted for age at menarche, body mass index, oral contraceptive use, menopausal hormone use, smoking, and diabetes.

We also examined effect modification by BMI (Web Table 4) and exogenous hormone use (OC and menopausal estrogen) (Web Table 5). We observed no significant interaction between age at last birth and any of these hormonal exposures (Pinteraction ≥ 0.12).

Table 5.

Association Between Age at Last Birth and Endometrial Cancer Histology in the Pooled Analysis

Age at Last Birth, years No. of Controls Type I Tumorsa
 
Type II Tumorsb
 
No. of Cases ORc 95% CI No. of Cases ORc 95% CI 
<25 2,189 1,235 1.00  106 1.00  
25–29 5,211 2,846 0.96 0.88, 1.06 232 1.00 0.78, 1.28 
30–34 5,295 2,302 0.84 0.76, 0.92 190 0.81 0.62, 1.05 
35–39 2,983 1,003 0.67 0.60, 0.75 117 0.84 0.63, 1.12 
≥40 884 257 0.55 0.46, 0.65 29 0.61 0.39, 0.94 
Ptrend   <0.0001   0.01  
 Per 5-year increase   0.87 0.84, 0.89  0.90 0.84, 0.97 
Age at Last Birth, years No. of Controls Type I Tumorsa
 
Type II Tumorsb
 
No. of Cases ORc 95% CI No. of Cases ORc 95% CI 
<25 2,189 1,235 1.00  106 1.00  
25–29 5,211 2,846 0.96 0.88, 1.06 232 1.00 0.78, 1.28 
30–34 5,295 2,302 0.84 0.76, 0.92 190 0.81 0.62, 1.05 
35–39 2,983 1,003 0.67 0.60, 0.75 117 0.84 0.63, 1.12 
≥40 884 257 0.55 0.46, 0.65 29 0.61 0.39, 0.94 
Ptrend   <0.0001   0.01  
 Per 5-year increase   0.87 0.84, 0.89  0.90 0.84, 0.97 

Abbreviations: CI, confidence interval; OR, odds ratio.

P value for the difference in odds ratio per 5-year increase between type I and type II tumors = 0.32.

aType I tumors include endometrioid, adenocarcinoma not otherwise specified, papillary adenocarcinoma, adenocarcinoma tubular, mucinous adenocarcinoma, and adenocarcinoma with squamous metaplasia.

bType II tumors include serous, papillary serous, squamous, adenosquamous, mixed cell parts serous.

cThe odds ratio was calculated using polytomous logistic regression adjusted for age, race/ethnicity, study, age at menarche, parity, body mass index, oral contraceptive use, menopausal hormone use, smoking, and diabetes.

Finally, Table 5 shows the pooled odds ratios relating age at last birth to histology-specific risks. Increasing age at last birth was associated with a reduced risk of both types of endometrial cancer: 7,643 type I tumor cases (OR5 = 0.87, 95% CI: 0.84, 0.89; Ptrend < 0.0001) and 674 type II tumor cases (OR5 = 0.90, 95% CI: 0.84, 0.97; Ptrend = 0.01). The P value for the difference in the OR5 between type I and type II tumors was 0.32. Results for type I tumors were similar when we restricted the analysis to the “purest” type I tumors, that is, grade 1 endometrioid tumors (1,168 cases; OR5 = 0.87, 95% CI: 0.81, 0.93; Ptrend = 0.0001).

DISCUSSION

The present pooled analysis of individual data from 17 epidemiologic studies is to our knowledge the largest investigation to date of the risk of endometrial cancer in relation to age at last birth and how tumor histology and other risk factors influence the relation. The analysis provides strong evidence that late age at last birth is independently associated with a lower risk of endometrial cancer. Risk decreases by approximately 13% per 5-year delay in last birth. Similar associations were seen for both type I and type II tumors and across all categories of age, parity, BMI, OC use, and menopausal estrogen use.

Several studies have examined the association between age at last birth or pregnancy and endometrial cancer risk (2–8, 10, 12, 21, 35–37), with the majority of studies reporting an inverse association (2–10); 5 of these published studies (the Fred Hutchinson Cancer Research Center Study, Iowa Women's Health Study, Nurses' Health Study, Polish Endometrial Cancer Study, and Shanghai Endometrial Cancer Study) (3, 12, 21, 36, 37) were included in the present pooled analysis. Seven studies have examined the association between time since last birth or pregnancy and endometrial cancer risk (2, 9, 12, 35, 36, 38, 39), with the majority of studies reporting an increasing risk with increasing time since last birth or pregnancy (2, 9, 35, 38, 39). In our pooled analysis, we did not include time since last birth in the statistical models because once the age at diagnosis/reference date is fixed, age at last birth and time since last birth are perfectly correlated. We have approached the issue of distinguishing between age at last birth and time since last birth by investigating the relation of age at last birth to risk of endometrial cancer within strata of age at diagnosis.

Although we found no modification of effect by most factors examined, we did not find an association between age at last birth and endometrial cancer risk in black women. It is possible that it was due to the small number of black women (186 cases and 566 controls) in our pooled analysis. The distribution of age at last birth among blacks was not strikingly different from the distributions among whites or Asians. In our study, type II tumors were more common in blacks (19%) than in whites (8%) or Asians (4%). However, we found that the association of age at last birth with type II tumors was only slightly weaker than that with type I tumors. When the analysis in blacks was restricted to type I tumors (141 cases), the odds ratio for endometrial cancer per 5-year increase in age at last birth was 0.94 (95% CI: 0.78, 1.14), slightly lower than the odds ratio for all tumors combined among black women (odds ratio = 0.98, 95% CI: 0.83, 1.16). The apparent lack of association between age at last birth and risk of endometrial cancer in blacks warrants additional study among larger groups of black women.

Several potential biologic mechanisms have been suggested for why late age at last birth might protect against endometrial cancer, including: 1) women capable of becoming pregnant at an older age may possess a “healthy” endometrium or experience fewer anovulatory cycles (3, 6); 2) prolonged exposure to progesterone during pregnancy may be particularly beneficial at older ages, the critical period for endometrial cancer development (3); and 3) shedding of premalignant and malignant cells, which are more likely to exist with increasing age, from the mucosal lining of the uterine cavity during childbirth (40). Whatever the underlying mechanism(s) is/are, our findings that the association was seen even among women aged 70 years or older at diagnosis suggest that the reduced risk associated with having a late age at last birth is long-lasting. Also, the fact that the association was similar in type I and type II tumors suggests that the impact of late age at last birth is probably estrogen-independent.

The strengths of the present study include the large sample size, which provides greater statistical power than any individual study alone, especially with regard to type II tumors and subgroup analyses; the minimal, if any, publication bias because inclusion of an individual study in our analysis was not dependent on whether results had been previously published; and the comparability across studies, in that we used individual-level data to standardize the definitions and modeling approaches for the exposure and potential confounders, which is not possible in meta-analyses based on published estimates. Nonetheless, heterogeneity in exposure assessment in how each study asked about certain exposures is a limitation of pooled analyses. In the present project, this limitation is of more concern for certain potential confounders (e.g., menopausal hormone use) than for age at last birth. Furthermore, the fact that there was no central pathologic review of tumor histology might lead to misclassification of tumor types.

In summary, the present large pooled analysis provides strong evidence that late age at last birth is associated with a reduced risk of endometrial cancer independent of other known risk factors and that this reduction in risk is long-lasting. Future investigations should focus on biologic mechanisms underlying this association to further clarify endometrial carcinogenesis.

ACKNOWLEDGMENTS

Author affiliations: Department of Preventive Medicine, Keck School of Medicine, University of Southern California, Los Angeles, California (Veronica Wendy Setiawan, Malcolm C. Pike, Wendy Cozen, Thomas Mack); Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, New York (Malcolm C. Pike, Xiaolin Liang, Sara H. Olson, Radhai Rastogi); Department of Epidemiology, Harvard School of Public Health, Boston, Massachusetts (Stalo Karageorgi, Immaculata De Vivo); Vanderbilt University Medical Center, Nashville, Tennessee (Sandra L. Deming, Hui Cai, Xiao-Ou Shu); School of Public Health, University of Minnesota, Minneapolis, Minnesota (Kristin Anderson, Kim Robien); Beckman Research Institute of City of Hope, Duarte, California (Leslie Bernstein, James V. Lacey, Jr.); National Cancer Institute, Bethesda, Maryland (Louise A. Brinton, Catherine Schairer, Nicolas Wentzensen, Hannah P. Yang); Mayo Clinic, Rochester, Minnesota (James R. Cerhan); Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, Washington (Chu Chen, Jennifer Doherty); School of Public Health and Health Professions, University at Buffalo, SUNY, Buffalo, New York (Jo L. Freudenheim); University of Hawaii Cancer Center, Honolulu, Hawaii (Marc T. Goodman, Galina Lurie, Rayna K. Matsuno, Pamela J. Thompson); M. Sklodowska-Curie Memorial Cancer Center and Institute of Oncology, Warsaw, Poland (Jolanta Lissowska); Yale School of Public Health, New Haven, Connecticut (Lingeng Lu, Harvey Risch, Herbert Yu); Channing Laboratory, Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts (Susan E. Hankinson, Immaculata De Vivo); Department of Cancer Prevention and Control, Roswell Park Cancer Institute, Buffalo, New York (Susan McCann, Kirsten B. Moysich); Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania (Timothy R. Rebbeck, Brian L. Strom); Division of Genetics and Population Health, Queensland Institute of Medical Research, Brisbane, Australia (Amanda B. Spurdle, Penelope M. Webb); Cancer Registry of Norway, Oslo, Norway (Giske Ursin); Department of Nutrition, University of Oslo, Norway (Giske Ursin); School of Public Health, University of Washington, Seattle, Washington (Noel S. Weiss); Shanghai Cancer Institute, Shanghai, China (Yong-Bing Xiang); and Cancer Prevention Institute of California, Fremont, California (Pamela L. Horn-Ross).

This work was supported by a grant from the National Cancer Institute (NCI) of the National Institutes of Health (NIH) (grant R03 CA135632 to V. W. S.). V. W. S is supported in part by a National Cancer Institute K07 Career Development Award (CA116543). P. M. W. and A. B. S. are supported by fellowships from the National Health and Medical Research Council of Australia (NHMRC). The individual studies were funded by the following grants and agencies: Australian National Endometrial Cancer Study: NHMRC grant 339435 and the Cancer Councils of Queensland and Tasmania; Bay Area Women's Health Study: NIH grant R01 CA74877; controls were collected under NIH grant R01 63446, U.S. Army Medical Research Program DAMD grant 17-96-607, and California Breast Cancer Research Program (CBCRP) grant 4JB-1106; Breast Cancer Detection Demonstration Project: Intramural Research Programs of the Department of Health and Human Services, NCI, NIH; Connecticut Endometrial Cancer Study: NIH grant R01 CA098346; California Teachers Study: NIH grant R01CA77398 and the CBCRP fund; the collection of cancer incidence data was supported by the California Department of Public Health as part of the statewide cancer reporting program mandated by California Health and Safety Code Section 103885, the NCI's Surveillance, Epidemiology, and End Results Program awarded to the Cancer Prevention Institute of California, the Public Health Institute, University of Southern California, and the Centers for Disease Control and Prevention's National Program of Cancer Registries; Estrogen, Diet, Genetics, and Endometrial Cancer Study: NIH grant R01 CA83918; Fred Hutchinson Cancer Research Center: NIH grants R35 CA39779, R01 CA75977, N01 HD23166, K05 CA92002, R01 CA105212, and R01 CA87538; Hawaii Endometrial Cancer Study: NIH grants P01 CA33619, R01 CA58598, N01 CN67001, and N01 PC35137; Iowa Women's Health Study: NIH grant R01 CA39742; Nurses' Health Study: NIH grants P01 CA87262 and R01 CA082838; Polish Endometrial Cancer Study: Intramural Research Funds of the Department of Health and Human Services, NCI, NIH; Shanghai Endometrial Cancer Study: NIH grant R01 CA092585; US Endometrial Cancer Study: Intramural Research Funds of the Department of Health and Human Services, NCI, NIH; case-control study conducted by USC: NIH grants R01 CA48774 and P30 CA14089; Women's Insight and Shared Experience Study: NIH grant P01 CA77596; and Western New York Diet Study: NIH grant CA11535.

The authors thank Jianning Luo for assistance in data cleaning and management and Dr. Robert A. Soslow at Memorial Sloan-Kettering Cancer Center for sharing his expertise in endometrial cancer pathology and assistance in the classification of type I and type II tumors. The authors also thank Dr. Leah Sansbury and Dr. Leah Mechanic at the National Cancer Institute for their support on the activities of the Epidemiology of Endometrial Cancer Consortium.

Conflict of interest: none declared.

REFERENCES

1
Cook
LS
Weiss
NS
Doherty
JA
, et al.  . 
Schottenfeld
D
Fraumeni
JF
Jr
Endometrial cancer
Cancer Epidemiology and Prevention
 , 
2006
New York, NY
Oxford University Press
(pg. 
1027
-
1043
)
2
Albrektsen
G
Heuch
I
Tretli
S
, et al.  . 
Is the risk of cancer of the corpus uteri reduced by a recent pregnancy? A prospective study of 765,756 Norwegian women
Int J Cancer
 , 
1995
, vol. 
61
 
4
(pg. 
485
-
490
)
3
Karageorgi
S
Hankinson
SE
Kraft
P
, et al.  . 
Reproductive factors and postmenopausal hormone use in relation to endometrial cancer risk in the Nurses' Health Study cohort 1976–2004
Int J Cancer
 , 
2010
, vol. 
126
 
1
(pg. 
208
-
216
)
4
Kvåle
G
Heuch
I
Ursin
G
Reproductive factors and risk of cancer of the uterine corpus: a prospective study
Cancer Res
 , 
1988
, vol. 
48
 
21
(pg. 
6217
-
6221
)
5
Lambe
M
Wuu
J
Weiderpass
E
, et al.  . 
Childbearing at older age and endometrial cancer risk (Sweden)
Cancer Causes Control
 , 
1999
, vol. 
10
 
1
(pg. 
43
-
49
)
6
Lesko
SM
Rosenberg
L
Kaufman
DW
, et al.  . 
Endometrial cancer and age at last delivery: evidence for an association
Am J Epidemiol
 , 
1991
, vol. 
133
 
6
(pg. 
554
-
559
)
7
Parazzini
F
La Vecchia
C
Negri
E
, et al.  . 
Reproductive factors and risk of endometrial cancer
Am J Obstet Gynecol
 , 
1991
, vol. 
164
 
2
(pg. 
522
-
527
)
8
Salazar-Martinez
E
Lazcano-Ponce
EC
Gonzalez Lira-Lira
G
, et al.  . 
Reproductive factors of ovarian and endometrial cancer risk in a high fertility population in Mexico
Cancer Res
 , 
1999
, vol. 
59
 
15
(pg. 
3658
-
3662
)
9
Pfeiffer
RM
Mitani
A
Landgren
O
, et al.  . 
Timing of births and endometrial cancer risk in Swedish women
Cancer Causes Control
 , 
2009
, vol. 
20
 
8
(pg. 
1441
-
1449
)
10
Bevier
M
Sundquist
J
Hemminki
K
Does the time interval between first and last birth influence the risk of endometrial and ovarian cancer?
Eur J Cancer
 , 
2011
, vol. 
47
 
4
(pg. 
586
-
591
)
11
Brinton
LA
Berman
ML
Mortel
R
, et al.  . 
Reproductive, menstrual, and medical risk factors for endometrial cancer: results from a case-control study
Am J Obstet Gynecol
 , 
1992
, vol. 
167
 
5
(pg. 
1317
-
1325
)
12
Brinton
LA
Sakoda
LC
Lissowska
J
, et al.  . 
Reproductive risk factors for endometrial cancer among Polish women
Br J Cancer
 , 
2007
, vol. 
96
 
9
(pg. 
1450
-
1456
)
13
Canchola
AJ
Chang
ET
Bernstein
L
, et al.  . 
Body size and the risk of endometrial cancer by hormone therapy use in postmenopausal women in the California Teachers Study cohort
Cancer Causes Control
 , 
2010
, vol. 
21
 
9
(pg. 
1407
-
1416
)
14
Doherty
JA
Cushing-Haugen
KL
Saltzman
BS
, et al.  . 
Long-term use of postmenopausal estrogen and progestin hormone therapies and the risk of endometrial cancer
Am J Obstet Gynecol
 , 
2007
, vol. 
197
 
2
pg. 
139
  
e1–139.e7
15
Goodman
MT
Wilkens
LR
Hankin
JH
, et al.  . 
Association of soy and fiber consumption with the risk of endometrial cancer
Am J Epidemiol
 , 
1997
, vol. 
146
 
4
(pg. 
294
-
306
)
16
Horn-Ross
PL
John
EM
Canchola
AJ
, et al.  . 
Phytoestrogen intake and endometrial cancer risk
J Natl Cancer Inst
 , 
2003
, vol. 
95
 
15
(pg. 
1158
-
1164
)
17
Lacey
JV
Jr
Brinton
LA
Lubin
JH
, et al.  . 
Endometrial carcinoma risks among menopausal estrogen plus progestin and unopposed estrogen users in a cohort of postmenopausal women
Cancer Epidemiol Biomarkers Prev
 , 
2005
, vol. 
14
 
7
(pg. 
1724
-
1731
)
18
Lu
L
Risch
H
Irwin
ML
, et al.  . 
Long-term overweight and weight gain in early adulthood in association with risk of endometrial cancer
Int J Cancer
 , 
2011
, vol. 
129
 
5
(pg. 
1237
-
1243
)
19
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
, vol. 
11
 
10
(pg. 
965
-
974
)
20
McCann
SE
Yeh
M
Rodabaugh
K
, et al.  . 
Higher regular coffee and tea consumption is associated with reduced endometrial cancer risk
Int J Cancer
 , 
2009
, vol. 
124
 
7
(pg. 
1650
-
1653
)
21
McPherson
CP
Sellers
TA
Potter
JD
, et al.  . 
Reproductive factors and risk of endometrial cancer. The Iowa Women's Health Study
Am J Epidemiol
 , 
1996
, vol. 
143
 
12
(pg. 
1195
-
1202
)
22
Olson
SH
Orlow
I
Bayuga
S
, et al.  . 
Variants in hormone biosynthesis genes and risk of endometrial cancer
Cancer Causes Control
 , 
2008
, vol. 
19
 
9
(pg. 
955
-
963
)
23
Pike
MC
Peters
RK
Cozen
W
, et al.  . 
Estrogen-progestin replacement therapy and endometrial cancer
J Natl Cancer Inst
 , 
1997
, vol. 
89
 
15
(pg. 
1110
-
1116
)
24
Rebbeck
TR
Troxel
AB
Wang
Y
, et al.  . 
Estrogen sulfation genes, hormone replacement therapy, and endometrial cancer risk
J Natl Cancer Inst
 , 
2006
, vol. 
98
 
18
(pg. 
1311
-
1320
)
25
Rowlands
IJ
Nagle
CM
Spurdle
AB
, et al.  . 
Gynecological conditions and the risk of endometrial cancer. Australian National Endometrial Cancer Study Group; Australian Ovarian Cancer Study Group
Gynecol Oncol
 , 
2011
, vol. 
123
 
3
(pg. 
537
-
541
)
26
Xu
WH
Dai
Q
Xiang
YB
, et al.  . 
Interaction of soy food and tea consumption with CYP19A1 genetic polymorphisms in the development of endometrial cancer
Am J Epidemiol
 , 
2007
, vol. 
166
 
12
(pg. 
1420
-
1430
)
27
Bodelon
C
Doherty
JA
Chen
C
, et al.  . 
Use of nonsteroidal antiinflammatory drugs and risk of endometrial cancer
Am J Epidemiol
 , 
2009
, vol. 
170
 
12
(pg. 
1512
-
1517
)
28
Olson
SH
Chen
C
De Vivo
I
, et al.  . 
Maximizing resources to study an uncommon cancer: E2C2—Epidemiology of Endometrial Cancer Consortium
Cancer Causes Control
 , 
2009
, vol. 
20
 
4
(pg. 
491
-
496
)
29
Sherman
ME
Theories of endometrial carcinogenesis: a multidisciplinary approach
Mod Pathol
 , 
2000
, vol. 
13
 
3
(pg. 
295
-
308
)
30
Felix
AS
Weissfeld
JL
Stone
RA
, et al.  . 
Factors associated with Type I and Type II endometrial cancer
Cancer Causes Control
 , 
2010
, vol. 
21
 
11
(pg. 
1851
-
1856
)
31
Prat
J
Prognostic parameters of endometrial carcinoma
Hum Pathol
 , 
2004
, vol. 
35
 
6
(pg. 
649
-
662
)
32
Ronnett
BM
Zaino
R
Ellenson
LH
, et al.  . 
Endometrial cancer. In: Kurman RJ
Blaustein's Pathology of the Female Genital Tract
 , 
2002
New York, NY
Springer
(pg. 
501
-
559
)
33
Higgins
JP
Thompson
SG
Deeks
JJ
, et al.  . 
Measuring inconsistency in meta-analyses
BMJ
 , 
2003
, vol. 
327
 
7414
(pg. 
557
-
560
)
34
Stukel
TA
Demidenko
E
Dykes
J
, et al.  . 
Two-stage methods for the analysis of pooled data
Stat Med
 , 
2001
, vol. 
20
 
14
(pg. 
2115
-
2130
)
35
Dossus
L
Allen
N
Kaaks
R
, et al.  . 
Reproductive risk factors and endometrial cancer: the European Prospective Investigation into Cancer and Nutrition
Int J Cancer
 , 
2010
, vol. 
127
 
2
(pg. 
442
-
451
)
36
Xu
WH
Xiang
YB
Ruan
ZX
, et al.  . 
Menstrual and reproductive factors and endometrial cancer risk: results from a population-based case-control study in urban Shanghai
Int J Cancer
 , 
2004
, vol. 
108
 
4
(pg. 
613
-
619
)
37
Pocobelli
G
Doherty
JA
Voigt
LF
, et al.  . 
Pregnancy history and risk of endometrial cancer
Epidemiology
 , 
2011
, vol. 
22
 
5
(pg. 
638
-
645
)
38
Parazzini
F
Negri
E
La Vecchia
C
, et al.  . 
Role of reproductive factors on the risk of endometrial cancer
Int J Cancer
 , 
1998
, vol. 
76
 
6
(pg. 
784
-
786
)
39
Hinkula
M
Pukkala
E
Kyyrönen
P
, et al.  . 
Grand multiparity and incidence of endometrial cancer: a population-based study in Finland
Int J Cancer
 , 
2002
, vol. 
98
 
6
(pg. 
912
-
915
)
40
Kvåle
G
Heuch
I
Nilssen
S
Reproductive factors and cancers of the breast and genital organs—are the different cancer sites similarly affected?
Cancer Detect Prev
 , 
1991
, vol. 
15
 
5
(pg. 
369
-
377
)

Author notes

Abbreviations: BMI, body mass index; CI, confidence interval; E2C2, Epidemiology of Endometrial Cancer Consortium; OC, oral contraceptive; OR5, odds ratio per 5-year increase.