Abstract
There are racial/ethnic differences in the incidence of hormone receptor positive and negative breast cancer. To understand why these differences exist, we investigated associations between hormone-related factors and breast cancer risk by race/ethnicity in the Multiethnic Cohort (MEC) Study.
Among 81,511 MEC participants (Native Hawaiian, Japanese American, Latina, African American, and White women), 3,806 estrogen receptor positive (ER+) and 828 ER- incident invasive breast cancers were diagnosed during a median of 21 years of follow-up. We used Cox proportional hazards regression models to calculate associations between race/ethnicity and breast cancer risk, and associations between hormone-related factors and breast cancer risk by race/ethnicity.
Relative to White women, ER+ breast cancer risk was higher in Native Hawaiians and lower in Latinas and African Americans; ER- disease risk was higher in African Americans. We observed interaction with race/ethnicity in associations between oral contraceptive use (OC; pint 0.03) and body mass index (BMI; pint 0.05) with ER+ disease risk; ever vs never OC use increased risk only in Latinas and positive associations for obese versus lean BMI were strongest in Japanese Americans. For ER- disease risk, associations for OC use, particularly duration of use, were strongest for African Americans (pint 0.04).
Our study shows that associations of OC use and obesity with ER+ and ER- breast cancer risk differ by race/ethnicity, but established risk factors do not fully explain racial/ethnic differences in risk. Further studies are needed to identify factors to explain observed racial/ethnic differences in breast cancer incidence.
Keywords: Breast cancer, estrogen receptor, hormone-related risk factors, race/ethnicity, risk
Introduction
Breast cancer is the most common cancer in women in the US, and incidence rates vary between racial/ethnic groups; age-adjusted 2012–16 annual incidence rates (per 100,000) are highest in non-Hispanic White women (130.8) and lowest in Asian/Pacific Islander women (93.2) [1]. Breast cancer subtypes also differ by race/ethnicity, with 81% of tumors in non-Hispanic White women being estrogen receptor (ER) positive (ER+), 79% in Asian/Pacific Islander women, 75% in Latina women, and 67% in African American women [1]. The Asian/Pacific Islander group includes multiple different subgroups, such as Japanese Americans and Native Hawaiians, and marked variation in the incidence of overall breast cancer and the proportion of tumor subtypes between these groups has been reported [2, 3].
Established hormone-related risk factors for postmenopausal breast cancer have mostly been studied among White women; a later age at menarche, parity, and an earlier age at first birth are established protective factors, particularly for hormone receptor (HoR) positive (HoR+; i.e. ER+ and/or estrogen and progesterone receptor positive (ER+/PR+) tumors), whereas a high BMI and postmenopausal hormone use (PMH; especially combined estrogen and progesterone (E+P) formulations) may increase risk [4, 5]. Results for HoR- disease are less consistent, but generally suggest weaker or no association for hormone-related factors [4, 6–8], with the exception of a positive association for parity in African American women [9, 10]. While it has been shown that factors such as age at menarche or parity may have a weaker association with HoR+ disease risk in African American than in White women [11], there is scarce literature on risk factors for HoR+ and HoR- breast cancer in Latina, Japanese American, and Native Hawaiian women.
To address the limited knowledge of breast cancer risk factors among non-White women, we analyzed differences in breast cancer risk accounting for HoR status by race/ethnicity and investigated the impact of hormone-related factors on risk in the major racial/ethnic groups in the Multiethnic Cohort (MEC) Study. While three previous studies in the MEC have evaluated hormone-related breast cancer risk factors [12–14], the current study is the first to evaluate risk factors for HoR+ and HoR- breast cancer by racial/ethnic group. Our objective here is to inform on breast cancer risk factors in a racially/ethnically diverse population which can help improve avenues for prevention among these underrepresented populations.
Materials & methods
Study population
The MEC Study has been described previously [15, 16]. Briefly, between 1993–1996 over 215,000 men and women from Hawaii and California (primarily Los Angeles County) aged 45–75 years completed a mailed baseline questionnaire including questions on lifestyle and reproductive factors. Incident, invasive breast cancers were identified by linkage to the Surveillance, Epidemiology, and End Results (SEER) cancer registries for Hawaii and California using International Classification of Diseases for Oncology 3rd revision site codes C500-C509 (excluding histology codes: 9050–9055, 9140, and 9590–9992), and information on tumor histology, stage and grade was obtained. Vital status was determined by linkage to state death files and the National Death Index. Case and death ascertainment were complete through December 31, 2014.
From 110,712 women in the five main self-reported racial/ethnic groups in the MEC Study (non-Hispanic White [referred to as White], Black or African American [referred to as African American], Native Hawaiian, Japanese American, and Latina) we excluded 4,973 participants diagnosed with breast cancer before cohort entry as reported at baseline or identified through linkage to the tumor registries, 15,647 participants who reported being premenopausal at baseline or had missing information on menopausal status, 7,939 participants missing information on relevant factors (any of these factors: use of postmenopausal hormones containing estrogen [PMH], use of oral contraceptives [OCs], age at menarche, number of children, age at first birth, or body mass index [BMI]), and 91 participants with <1 year of follow-up. We additionally excluded 551 incident breast cancer cases missing information on ER status. Comparing cases missing ER status to those with data on ER status, we observed similarities in age at diagnosis and tumor stage while the proportion of missing ER status data was somewhat higher in African American and Latina cases (16% and 18%, respectively vs ≤8% in the other groups). This left 81,511 participants among whom 3,806 ER+ and 828 ER- incident invasive breast cancers were diagnosed during a median of 21 years of follow-up (interquartile range [IQR] 16–21 years).
Exposures and covariates
All exposures and covariates were assessed at baseline. Hormone-related exposures of interest were parity (no, yes), number of children (0, 1, 2, 3, ≥4), age at menarche (<13, 13–14, ≥15 years), age at natural menopause (<45, 45–49, 50–54, ≥55 years; in women reporting a natural menopause, or including a category for surgical/unknown cause of menopause when used as a covariate), OC use (never, ever [defined as ≥1 year] use), duration of OC use (never, ≥1 year to <5 years, ≥5 years), PMH use (never, former, current use at baseline), PMH duration (never, former, current <5, current ≥5 years), type of PMH use (never, former, current E-alone, current E+P), BMI (<25, 25–29, ≥30 kg/m2). The MEC Study baseline questionnaire did not include questions on breastfeeding. Smoking was not included as a risk factor in the current study because a detailed analysis in the MEC was recently published [17]. Additional variables that were evaluated as confounders include: first-degree family history of breast and/or ovarian cancer (no, yes), diabetes prevalence at baseline (has a doctor ever told you that you have diabetes (high blood sugar)?; no, yes), average alcohol consumption (non-drinker, <1 drink/day, ≥1 drink/day), smoking status (never, former, current), and education (12th grade or less, vocational training or some college education, college graduate or higher).
Statistical analyses
Cox proportional hazards regression models, with the time (years) between age at study baseline and exit as the time scale, were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for ER+ and ER- breast cancer risk. Study exit was defined as 1) the date of diagnosis of incident invasive breast cancer, 2) date of death, or 3) end of follow-up (December 31, 2014), whichever occurred first. Associations for ER+ vs ER- disease, as well as ER+/PR+ vs estrogen receptor and progesterone negative (ER-/PR-) disease, were compared using competing risks Cox regression models [18, 19]. In analysis specific to ER+ breast cancer risk, ER- cases were censored at the date of diagnosis, and vice versa.
Multivariable models included a priori selected covariates: continuous age at baseline, ever OC use, number of children, and age group at first birth. Models in the overall population were additionally adjusted for race/ethnicity (strata). All other factors listed under the exposures and covariates heading were evaluated as potential confounders using forward selection and did not impact results (<10% HR change) [20]. Our primary analyses considered ER+ and ER- status separately, and we carried out sensitivity analyses considering ER+PR+ and ER-PR- status. Missing data in adjustment variables were assigned the value corresponding to the category with the most participants or, if ≥5% of data were missing, were included in a separate category. P-values for trends were calculated using continuous variables (where available; continuous BMI was natural log transformed). Interactions between race/ethnicity and the exposure variables were evaluated by comparing multivariable models with and without multiplicative interaction terms using likelihood ratio tests. The proportional hazards assumption was assessed using Schoenfeld residuals and no violation was observed. Descriptive analysis showing the population distribution of exposure variables were age-standardized for baseline age (per year) using indirect standardization. A two tailed P<0.05 was considered statistically significant. Analyses were performed using SAS 9.4 (SAS Institute Inc., Cary, NC, USA). Our study was approved by the Institutional Review Boards at the University of Hawaii and the University of Southern California, who considered return of the completed baseline questionnaire as active consent.
Results
As shown in Table 1, the median age at cohort entry among postmenopausal female participants was 61 years (interquartile range (IQR) 55, 68). Over half of Native Hawaiians and Latinas had 4 or more children, compared to <36% in the other racial/ethnic groups. Japanese Americans had the fewest children and were older at their first birth (41% were ≥26 years vs ≤25% in the other groups). Whites, along with Japanese Americans, were the most likely to report current use of PMH (at baseline; both 45% vs ≤28% in the other groups). African Americans were most likely to be obese (37% with a BMI ≥30kg/m2) and Japanese Americans were least likely (6%). Among breast cancer cases (Table 2), the median age at diagnosis was 72 years (IQR: 65, 78) and was lowest for Native Hawaiians (68 years, IQR: 62, 74) years. In the overall population, 82% of breast cancers were ER+ (65% were ER+/PR+). The proportion of ER- disease was highest in African Americans (27%; 20% in Latinas, and 15–16% in the other groups), with a similar pattern for ER-/PR- tumors.
Table 1.
Age-standardized population characteristics, overall and by race/ethnicity, among postmenopausal women in the Multiethnic Cohort Study.
| Overall | White | African American | Native Hawaiian | Japanese American | Latina | |
|---|---|---|---|---|---|---|
| n=81,511 | n=20,003 | n=16,060 | n=5,468 | n=22,333 | n=17,647 | |
|
| ||||||
| Age at cohort entrya (years; median, IQR) | 61 (55, 68) | 61 (54, 68) | 62 (55, 69) | 58 (52, 65) | 64 (57, 69) | 60 (55, 65) |
| Duration of follow-upa (years; median, IQR) | 21 (16, 21) | 21 (16, 21) | 20 (13, 22) | 20 (14, 21) | 21 (18, 21) | 21 (18, 22) |
| Parity | ||||||
| Nulliparous | 12% | 15% | 13% | 7% | 14% | 8% |
| Parous | 88% | 85% | 87% | 93% | 86% | 92% |
| 1 child | 11% | 11% | 15% | 6% | 11% | 7% |
| 2 children | 22% | 25% | 19% | 14% | 31% | 15% |
| 3 children | 21% | 22% | 17% | 18% | 25% | 19% |
| ≥4 children | 34% | 26% | 36% | 56% | 19% | 52% |
| Age at first birth b | ||||||
| <21 years | 34% | 31% | 54% | 49% | 12% | 44% |
| 21–25 years | 41% | 44% | 31% | 40% | 47% | 38% |
| ≥26 years | 25% | 25% | 15% | 12% | 41% | 18% |
| Age at menarche | ||||||
| <13 years | 49% | 50% | 49% | 55% | 48% | 48% |
| 13–14 years | 39% | 40% | 38% | 34% | 39% | 39% |
| ≥15 years | 13% | 11% | 13% | 12% | 13% | 14% |
| Age at natural menopause c | ||||||
| <45 years | 16% | 15% | 20% | 19% | 11% | 21% |
| 45–49 years | 31% | 33% | 32% | 30% | 28% | 34% |
| 50–54 years | 41% | 41% | 36% | 37% | 48% | 35% |
| ≥55 years | 11% | 11% | 11% | 13% | 12% | 8% |
| Missing | 1% | 0% | 1% | 1% | 1% | 1% |
| OC use at baseline (n, %) | ||||||
| Never | 70% | 62% | 67% | 74% | 73% | 75% |
| Ever (≥1 year) | 30% | 38% | 33% | 26% | 27% | 25% |
| <5 years | 16% | 19% | 16% | 14% | 15% | 15% |
| ≥5 years | 14% | 19% | 16% | 12% | 11% | 10% |
| Missing duration of OC use | 1% | 0% | 1% | 0% | 0% | 1% |
| PMH use at baseline | ||||||
| Never | 45% | 35% | 55% | 53% | 40% | 53% |
| Former | 20% | 21% | 24% | 20% | 15% | 22% |
| Current | 35% | 45% | 21% | 28% | 45% | 25% |
| Current: <5 years | 15% | 17% | 10% | 14% | 21% | 12% |
| Current: ≥5 years | 19% | 28% | 11% | 14% | 23% | 12% |
| Missing duration of PMH-E use | 1% | 0% | 1% | 0% | 0% | 1% |
| Current: E-alone | 20% | 25% | 16% | 17% | 22% | 16% |
| Current: E+P | 14% | 19% | 5% | 11% | 22% | 9% |
| Missing current PMH-P use | 1% | 0% | 0% | 0% | 1% | 1% |
| Body Mass Index at baseline | ||||||
| <25 kg/m2 | 46% | 52% | 25% | 34% | 69% | 32% |
| 25–29 kg/m2 | 32% | 30% | 37% | 34% | 24% | 40% |
| ≥30 kg/m2 | 22% | 18% | 37% | 32% | 6% | 28% |
All percentages represent the full population unless otherwise stated
not age-standardized for baseline age
among 71,481 parous women
among 46,895 women who reported a natural menopause.
Abbreviations: IQR: interquartile range, OC: Oral contraceptive, PMH: postmenopausal hormone.
Table 2.
Age-standardized characteristics of postmenopausal breast cancer cases, overall and by race/ethnicity.
| Overall | White | African American | Native Hawaiian | Japanese American | Latina | |
|---|---|---|---|---|---|---|
| n=4,634 | n=1,252 | n=802 | n=450 | n=1,447 | n=683 | |
|
| ||||||
| Age at diagnosisa (years; median, IQR) | 72 (65, 78) | 71 (65, 77) | 73 (67, 79) | 68 (62, 74) | 73 (66, 78) | 71 (66, 77) |
| Duration of follow-upa (years; median, IQR) | 10 (6, 15) | 9 (5, 15) | 11 (6, 16) | 10 (5, 16) | 9 (5, 15) | 12 (6, 16) |
| Tumor hormone receptor status at diagnosis | ||||||
| ER+ | 82% | 85% | 73% | 85% | 84% | 80% |
| ER+/PR+ | 65% | 68% | 51% | 73% | 69% | 60% |
| ER+/PR- | 13% | 13% | 12% | 10% | 12% | 14% |
| ER- | 18% | 15% | 27% | 15% | 16% | 20% |
| ER-/PR+ | 2% | 1% | 3% | 2% | 2% | 2% |
| ER-/PR- | 16% | 14% | 23% | 13% | 14% | 18% |
| Missing PR status | 5% | 4% | 11% | 2% | 3% | 6% |
| Tumor grade at diagnosis | ||||||
| Grade I | 26% | 27% | 20% | 26% | 31% | 21% |
| Grade II | 41% | 42% | 33% | 43% | 44% | 40% |
| Grade III & IV | 27% | 24% | 38% | 26% | 21% | 32% |
| Missing | 6% | 6% | 9% | 6% | 4% | 7% |
| Disease stage at diagnosis | ||||||
| Localized | 72% | 72% | 67% | 71% | 78% | 68% |
| Regional & distant | 27% | 28% | 32% | 28% | 22% | 30% |
| Missing | 1% | 0% | 1% | 1% | 0% | 2% |
All percentages represent the total group size
Abbreviations: ER: Estrogen receptor, PR: progesterone receptor, IQR: interquartile range
not age-standardized for baseline age.
In competing risks models by ER status in the overall population (Supplementary table 1), we observed differences in associations with risk of ER+ and ER- breast cancer for OC use (phet ≤0.01), type of PMH use (phet <0.01) and BMI (phet <0.01), and a non-significant difference for parity (phet 0.07; phet 0.19 for number of children). Current (baseline) PMH E+P use, obesity, and nulliparity increased risk for ER+ cancers; ever OC use increased ER- disease risk. Results from competing risks models considering both ER/PR status were similar.
When evaluating associations between race/ethnicity and breast cancer risk by ER status (Table 3) we found that, relative to White women, age-adjusted ER+ breast cancer risk was higher in Native Hawaiians (HR 1.42 [CI 1.26, 1.60]) and lower in African Americans (HR 0.72 [CI 0.66, 0.80]) and Latinas (HR 0.54 [CI 0.49, 0.60]). Age-adjusted risk of ER- disease was significantly higher only in African Americans (HR 1.41 [CI 1.15, 1.72]). After adjustment for number of children, and ever OC use, differences in risk relative to White women remained and the association with ER- disease was somewhat stronger in Native Hawaiians (HR 1.33 [CI 1.00, 1.77]). Results by ER/PR status were similar, but the suggestive positive association with ER- disease among Native Hawaiians was attenuated for ER-/PR- disease (HR 1.26 [CI 0.93, 1.72]).
Table 3.
Associations between race/ethnicity and risk of hormone receptor positive and negative breast cancer in postmenopausal women, from competing risks models.
| White (ref.) | African American | Native Hawaiian | Japanese American | Latina | Pheta | |
|---|---|---|---|---|---|---|
|
|
||||||
| ER+ breast cancer | ||||||
| Number of cases/population | 1,063/20,003 | 598/16,060 | 387/5,468 | 1,216/22,333 | 542/17,647 | |
| HR (CI) Age-adjusted | 1.00 | 0.72 (0.66, 0.80) | 1.42 (1.26, 1.60) | 0.97 (0.89, 1.05) | 0.54 (0.49, 0.60) | <0.0001 |
| HR (CI) Multivariable adjustedb | 1.00 | 0.74 (0.67, 0.82) | 1.52 (1.35, 1.71) | 0.97 (0.89, 1.05) | 0.58 (0.52, 0.65) | <0.0001 |
| ER+/PR+ breast cancer c | ||||||
| Number of cases/population | 853/20,003 | 412/16,060 | 336/5,468 | 994/22,333 | 407/17,647 | |
| HR (CI) Age-adjusted | 1.00 | 0.62 (0.55, 0.70) | 1.53 (1.35, 1.74) | 0.99 (0.90, 1.08) | 0.51 (0.45, 0.57) | <0.0001 |
| HR (CI) Multivariable adjustedb | 1.00 | 0.63 (0.56, 0.71) | 1.64 (1.44, 1.86) | 0.99 (0.91, 1.09) | 0.54 (0.48, 0.61) | <0.0001 |
| ER- breast cancer | ||||||
| Number of cases/population | 189/20,003 | 204/16,060 | 63/5,468 | 231/22,333 | 141/17,647 | |
| HR (CI) Age-adjusted | 1.00 | 1.41 (1.15, 1.72) | 1.26 (0.94, 1.67) | 1.07 (0.88, 1.30) | 0.81 (0.65, 1.00) | |
| HR (CI) Multivariable adjustedb | 1.00 | 1.44 (1.18, 1.76) | 1.33 (1.00, 1.77) | 1.08 (0.89, 1.32) | 0.86 (0.69, 1.07) | |
| ER-/PR- breast cancer c | ||||||
| Number of cases/population | 172/20,003 | 177/16,060 | 54/5,468 | 204/22,333 | 128/17,647 | |
| HR (CI) Age-adjusted | 1.00 | 1.34 (1.09, 1.65) | 1.19 (0.88, 1.62) | 1.03 (0.84, 1.27) | 0.80 (0.64, 1.01) | |
| HR (CI) Multivariable adjustedb | 1.00 | 1.37 (1.11, 1.69) | 1.26 (0.93, 1.72) | 1.05 (0.85, 1.28) | 0.85 (0.68, 1.08) | |
Abbreviations: CI: 95% confidence interval, ER: estrogen receptor, HR: hazard ratio, PR: progesterone receptor
Pheterogeneity comparing model assuming the same exposure between race/ethnicity and hormone receptor positive and negative breast cancer subtypes to model allowing different associations using likelihood ratio tests
adjusted for baseline age, number of children, and ever OC use
ER+/PR+ is defined as ER+ and PR+ disease and ER-/PR- is defined as ER- and PR- disease, other subtypes (i.e., ER+/PR- and ER-/PR+) and those missing PR status were censored at their date of diagnosis.
In analyses of hormone-related factors in relation to ER+ breast cancer risk, we observed interaction with race/ethnicity in the association between ever OC use and ER+ breast cancer risk (pint 0.03; duration of use pint 0.10; Table 4). Compared to never use, OC ever use was positively associated with risk in Latinas (HR 1.38 [CI 1.14, 1.67]) irrespective of duration of use, while we saw no association in the other groups. There was a suggestive interaction between BMI and race/ethnicity (pint 0.05); an obese (≥30kg/m2) versus lean (<25kg/m2) BMI was associated with increased ER+ disease risk in all groups (HR ranging from 1.17 to 1.63), while an overweight (25–29kg/m2) vs lean BMI was associated with risk only in Japanese American (HR 1.41 [CI 1.24, 1.59]) and White women (HR 1.17 [CI 1.02, 1.34]). When stratifying these analyses for PMH use, positive associations between a high BMI and ER+ disease were limited to women who reported no PMH use at baseline (Supplementary Table 2). We observed no significant interaction between PMH use and race/ethnicity (pint ≥0.11) and current (baseline) PMH use increased risk irrespective of duration in all groups except African American and Native Hawaiian women (current vs never use; HRs ranging 1.35–1.45), with stronger associations for PMH E+P than E-alone. For African American women, a suggestive positive association was seen only for E+P use (current vs never use; HR 1.37 [CI 0.98, 1.91]). Although associations between other hormone-related factors and ER+ breast cancer risk did not significantly differ between racial/ethnic groups (Pint race/ethnicity ≥0.23), associations for age at menarche and age at first birth were more pronounced in Japanese Americans (e.g. age at first birth, 21–25 vs <21 years: HR 1.38 [CI 1.09, 1.75]; ≥26 vs <21 years: HR 1.56 [CI 1.22, 1.99]). Associations between hormone-related factors and ER+/PR+ breast cancer risk (Supplementary tables 3 & 4) were similar to those for ER+ status, with the exception of an attenuated association for PMH E+P in African Americans (HR 1.01 [CI 0.64, 1.59]).
Table 4.
Associations (HR & CI) between hormone-related factors and ER+ breast cancer riska in postmenopausal women, by race/ethnicity.
| White | African American | Native Hawaiian | Japanese American | Latina | Pintb | |
|---|---|---|---|---|---|---|
|
|
||||||
| Number of cases/population | 1,063/20,003 | 598/16,060 | 387/5,468 | 1,216/22,333 | 542/17,647 | |
| Parity | ||||||
| Nulliparous (ref.) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | |
| Parous | 0.84 (0.72, 0.99) | 0.90 (0.72, 1.14) | 0.77 (0.54, 1.09) | 0.83 (0.71, 0.97) | 0.76 (0.57, 1.02) | 0.88c |
| 1 child | 0.94 (0.75, 1.18) | 1.00 (0.74, 1.33) | 0.86 (0.51, 1.43) | 0.96 (0.77, 1.19) | 0.77 (0.50, 1.18) | |
| 2 children | 0.86 (0.71, 1.04) | 0.82 (0.62, 1.10) | 0.79 (0.52, 1.22) | 0.81 (0.68, 0.97) | 0.94 (0.67, 1.33) | |
| 3 children | 0.93 (0.77, 1.13) | 1.04 (0.78, 1.38) | 0.79 (0.52, 1.18) | 0.85 (0.71, 1.02) | 0.88 (0.63, 1.23) | |
| ≥4 children | 0.70 (0.57, 0.85) | 0.84 (0.65, 1.09) | 0.74 (0.51, 1.07) | 0.73 (0.60, 0.89) | 0.67 (0.49, 0.91) | 0.86d |
| ptrend number of children (incl. 0) | <0.0001 | 0.04 | 0.22 | 0.0005 | <0.0001 | |
| ptrend number of children (excl. 0)e | 0.002 | 0.05 | 0.69 | 0.01 | 0.0005 | |
| Age at first birth e | ||||||
| <21 years (ref.) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | |
| 21–25 years | 1.14 (0.96, 1.34) | 1.12 (0.92, 1.36) | 0.95 (0.75, 1.19) | 1.38 (1.09, 1.75) | 1.22 (1.00, 1.50) | |
| ≥26 years | 1.34 (1.11, 1.62) | 1.23 (0.96, 1.58) | 0.87 (0.60, 1.27) | 1.56 (1.22, 1.99) | 1.44 (1.11, 1.85) | 0.23 |
| Age at menarche | ||||||
| <13 years (ref.) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | |
| 13–14 years | 0.94 (0.82, 1.07) | 0.99 (0.83, 1.17) | 0.79 (0.63, 0.99) | 0.85 (0.75, 0.96) | 1.02 (0.85, 1.23) | |
| ≥15 years | 0.91 (0.74, 1.11) | 0.69 (0.52, 0.92) | 0.86 (0.62, 1.20) | 0.69 (0.57, 0.83) | 1.00 (0.77, 1.29) | 0.29 |
| Age at natural menopause f | ||||||
| <45 years (ref.) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | |
| 45–49 years | 1.22 (0.94, 1.58) | 1.19 (0.82, 1.72) | 1.05 (0.73, 1.53) | 1.03 (0.79, 1.36) | 0.96 (0.69, 1.32) | |
| 50–54 years | 1.22 (0.95, 1.57) | 1.24 (0.86, 1.77) | 1.21 (0.85, 1.74) | 1.26 (0.98, 1.62) | 1.13 (0.83, 1.53) | |
| ≥55 years | 1.68 (1.23, 2.28) | 1.49 (0.96, 2.31) | 1.04 (0.62, 1.73) | 1.39 (1.03, 1.87) | 1.49 (0.99, 2.24) | 0.80 |
| OC use | ||||||
| Never (ref.) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | |
| Ever | 0.98 (0.85, 1.12) | 1.14 (0.94, 1.38) | 0.83 (0.65, 1.05) | 0.97 (0.83, 1.12) | 1.38 (1.14, 1.67) | 0.03g |
| Ever: <5 years | 1.03 (0.87, 1.22) | 1.09 (0.85, 1.39) | 0.81 (0.60, 1.08) | 0.91 (0.75, 1.09) | 1.32 (1.05, 1.67) | |
| Ever: ≥5 years | 0.94 (0.79, 1.11) | 1.21 (0.95, 1.54) | 0.86 (0.64, 1.17) | 1.06 (0.87, 1.28) | 1.46 (1.12, 1.89) | 0.10h |
| PMH use | ||||||
| Never (ref.) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | |
| Former | 1.10 (0.92, 1.32) | 1.16 (0.96, 1.41) | 1.08 (0.82, 1.40) | 1.17 (0.98, 1.39) | 1.06 (0.85, 1.32) | |
| Current | 1.43 (1.24, 1.65) | 1.17 (0.95, 1.43) | 1.02 (0.81, 1.29) | 1.45 (1.27, 1.64) | 1.35 (1.11, 1.65) | 0.11i |
| Current: <5 years | 1.57 (1.30, 1.88) | 1.23 (0.93, 1.63) | 1.08 (0.82, 1.40) | 1.17 (0.98, 1.39) | 1.06 (0.85, 1.32) | |
| Current: ≥5 years | 1.36 (1.16, 1.59) | 1.15 (0.89, 1.48) | 0.93 (0.69, 1.24) | 1.28 (1.09, 1.51) | 1.33 (1.03, 1.72) | 0.20j |
| Current: E-alone | 1.07 (0.90, 1.27) | 1.10 (0.87, 1.38) | 0.90 (0.68, 1.20) | 1.22 (1.04, 1.42) | 1.20 (0.95, 1.52) | |
| Current: E+P | 1.97 (1.68, 2.32) | 1.37 (0.98, 1.91) | 1.22 (0.91, 1.65) | 1.69 (1.46, 1.95) | 1.65 (1.27, 2.13) | 0.17k |
| Body Mass Index | ||||||
| <25 kg/m2 (ref.) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | |
| 25–29.9 kg/m2 | 1.17 (1.02, 1.34) | 1.17 (0.95, 1.45) | 1.15 (0.88, 1.49) | 1.41 (1.24, 1.59) | 1.12 (0.91, 1.38) | |
| ≥30 kg/m2 | 1.17 (1.00, 1.38) | 1.26 (1.01, 1.56) | 1.47 (1.14, 1.89) | 1.63 (1.32, 2.02) | 1.43 (1.15, 1.78) | 0.05 |
| ptrend log BMI | 0.09 | 0.02 | 0.001 | <0.0001 | 0.001 | |
Cox proportional hazards regression models adjusted for age at baseline, number of children, and ever OC use.
Abbreviations: BMI: body mass index, CI: 95% confidence interval, ER: estrogen receptor, HR: hazard ratio, OC: oral contraceptive, PMH: postmenopausal hormone.
ER- breast cancer cases censored at date of diagnosis
pint comparing model without to model with interaction term for exposure and race/ethnicity using log-likelihood ratio tests
pint parity (no, yes)
pint number of children (0, 1, 2, 3, ≥4)
among parous women (16,879 White, 13,872 African American, 5,078 Native Hawaiian, 19,360 JA and 16,292 Latina)
among postmenopausal women who reported a natural menopause (11,574 White, 7,395 African American, 3,002 Native Hawaiian, 14,127 Japanese American and 10,474 Latina)
pint never vs ever OC use
pint duration of OC use (never, <5, ≥5 years)
pint never, former, current PMH use
pint duration of PMH use (never, former, current: <5, current: ≥5 years)
pint type of PMH use (never, former, current E-alone, current E+P).
In analyses of ER- breast cancer risk, we observed an interaction between race/ethnicity and duration of OC use (pint 0.04), but not for ever OC use (pint 0.25; Table 5). In African Americans, ever OC use was positively associated with ER- breast cancer risk (HR 1.77 [CI 1.29, 2.43]) irrespective of duration of use, while associations were significant only for ≥5 years of use in White and Japanese American women (HRs 1.63–1.67). Although there was no interaction with race/ethnicity for other variables (Pint ≥0.09), current PMH use at baseline was associated with higher ER- disease risk only in Latinas (HR 1.66 [CI 1.14, 2.40]; no association seen for the E+P formulation or ≥5-year duration of PMH use). Aside from an attenuated association for OC use in Japanese Americans (≥5 years vs never use; HR 1.45 [CI 0.93, 2.26]), results were unchanged in analyses of ER-/PR- breast cancer (Supplementary table 5).
Table 5.
Associations (HR & CI) between hormone-related factors and ER- breast cancer riska in postmenopausal women, by race/ethnicity
| White | African American | Native Hawaiian | Japanese American | Latina | Pintb | |
|---|---|---|---|---|---|---|
|
|
||||||
| Number of cases/population | 189/20,003 | 204/16,060 | 63/5,468 | 231/22,333 | 141/17,647 | |
| Parity | ||||||
| Nulliparous (ref.) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | |
| Parous | 1.10 (0.73, 1.67) | 0.89 (0.59, 1.35) | 0.99 (0.36, 2.73) | 1.08 (0.73, 1.61) | 0.72 (0.41, 1.25) | 0.77c |
| 1 child | 0.91 (0.50, 1.66) | 0.88 (0.52, 1.49) | 1.06 (0.27, 4.26) | 1.34 (0.80, 2.23) | 0.79 (0.35, 1.77) | |
| 2 children | 1.22 (0.76, 1.95) | 0.88 (0.54, 1.46) | 1.28 (0.41, 3.99) | 1.03 (0.66, 1.60) | 1.01 (0.53, 1.92) | |
| 3 children | 1.01 (0.61, 1.66) | 1.18 (0.72, 1.91) | 0.72 (0.22, 2.36) | 1.08 (0.68, 1.70) | 0.59 (0.30, 1.16) | |
| ≥4 children | 1.15 (0.71, 1.84) | 0.76 (0.48, 1.20) | 1.00 (0.35, 2.83) | 1.02 (0.63, 1.66) | 0.68 (0.38, 1.20) | 0.59d |
| ptrend number of children (incl. 0) | 0.99 | 0.33 | 0.88 | 0.56 | 0.02 | |
| ptrend number of children (excl. 0)e | 0.71 | 0.38 | 0.88 | 0.34 | 0.05 | |
| Age at first birth e | ||||||
| <21 years (ref.) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | |
| 21–25 years | 1.53 (1.04, 2.24) | 1.05 (0.76, 1.47) | 1.02 (0.58, 1.79) | 0.99 (0.61, 1.60) | 1.76 (1.18, 2.63) | |
| ≥26 years | 1.22 (0.76, 1.95) | 0.85 (0.53, 1.37) | 0.99 (0.40, 2.41) | 1.29 (0.79, 2.11) | 1.38 (0.80, 2.36) | 0.16 |
| Age at menarche | ||||||
| <13 years (ref.) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | |
| 13–14 years | 0.79 (0.57, 1.08) | 0.82 (0.60, 1.10) | 1.33 (0.79, 2.26) | 1.03 (0.78, 1.36) | 1.00 (0.70, 1.43) | |
| ≥15 years | 1.38 (0.91, 2.09) | 0.94 (0.61, 1.44) | 1.12 (0.49, 2.55) | 0.80 (0.52, 1.23) | 0.97 (0.58, 1.61) | 0.62 |
| Age at natural menopause f | ||||||
| <45 years (ref.) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | |
| 45–49 years | 1.10 (0.59, 2.03) | 0.74 (0.42, 1.29) | 1.06 (0.39, 2.88) | 1.30 (0.69, 2.43) | 0.74 (0.36, 1.55) | |
| 50–54 years | 1.08 (0.58, 2.00) | 0.80 (0.47, 1.38) | 0.91 (0.34, 2.45) | 1.25 (0.68, 2.28) | 1.85 (0.99, 3.46) | |
| ≥55 years | 1.27 (0.56, 2.85) | 1.05 (0.52, 2.11) | 1.48 (0.46, 4.74) | 1.14 (0.55, 2.38) | 0.93 (0.33, 2.63) | 0.35 |
| OC use | ||||||
| Never (ref.) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | |
| Ever | 1.34 (0.97, 1.86) | 1.77 (1.29, 2.43) | 1.28 (0.72, 2.26) | 1.31 (0.95, 1.81) | 1.21 (0.82, 1.77) | 0.25g |
| Ever: <5 years | 0.99 (0.64, 1.52) | 1.63 (1.11, 2.40) | 1.39 (0.71, 2.70) | 1.08 (0.71, 1.63) | 1.58 (1.03, 2.40) | |
| Ever: ≥5 years | 1.67 (1.16, 2.40) | 1.88 (1.29, 2.73) | 1.20 (0.57, 2.52) | 1.63 (1.10, 2.42) | 0.77 (0.41, 1.46) | 0.04h |
| PMH use | ||||||
| Never (ref.) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | |
| Former | 0.76 (0.49, 1.19) | 0.97 (0.69, 1.36) | 0.98 (0.51, 1.86) | 1.34 (0.93, 1.94) | 0.87 (0.54, 1.41) | |
| Current | 1.14 (0.83, 1.58) | 0.89 (0.63, 1.26) | 0.75 (0.41, 1.37) | 1.15 (0.86, 1.54) | 1.66 (1.14, 2.40) | 0.12i |
| Current: <5 years | 1.18 (0.77, 1.80) | 1.01 (0.65, 1.58) | 0.53 (0.23, 1.24) | 1.19 (0.82, 1.72) | 2.20 (1.42, 3.39) | |
| Current: ≥5 years | 1.12 (0.78, 1.61) | 0.78 (0.48, 1.26) | 0.92 (0.44, 1.93) | 1.14 (0.81, 1.60) | 1.26 (0.76, 2.09) | 0.09j |
| Current: E-alone | 1.07 (0.74, 1.55) | 0.97 (0.66, 1.42) | 0.81 (0.41, 1.63) | 1.13 (0.79, 1.60) | 1.81 (1.20, 2.73) | |
| Current: E+P | 1.25 (0.85, 1.85) | 0.57 (0.28, 1.17) | 0.67 (0.28, 1.62) | 1.21 (0.85, 1.71) | 1.26 (0.71, 2.24) | 0.20k |
| Body Mass Index | ||||||
| <25 kg/m2 (ref.) | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | |
| 25–29.9 kg/m2 | 0.91 (0.66, 1.27) | 0.89 (0.63, 1.26) | 1.31 (0.71, 2.42) | 1.30 (0.97, 1.74) | 0.91 (0.63, 1.33) | |
| ≥30 kg/m2 | 0.88 (0.59, 1.32) | 0.96 (0.68, 1.37) | 1.10 (0.58, 2.10) | 1.02 (0.57, 1.80) | 0.74 (0.47, 1.15) | 0.61 |
| ptrend log BMI | 0.06 | 0.93 | 0.44 | 0.20 | 0.44 | |
Cox proportional hazards regression models adjusted for age at baseline, number of children, and ever OC use.
Abbreviations: BMI: body mass index, CI: 95% confidence interval, ER: estrogen receptor, HR: hazard ratio, OC: oral contraceptive, PMH: postmenopausal hormone.
ER+ breast cancer cases censored at date of diagnosis
pint comparing model without to model with interaction term for exposure and race/ethnicity using log-likelihood ratio tests
pint parity (no, yes)
pint number of children (0, 1, 2, 3, ≥4)
among parous women (16,879 White, 13,872 African American, 5,078 Native Hawaiian, 19,360 JA and 16,292 Latina)
among postmenopausal women who reported a natural menopause (11,574 White, 7,395 African American, 3,002 Native Hawaiian, 14,127 Japanese American and 10,474 Latina)
pint never vs ever OC use
pint duration of OC use (never, <5, ≥5 years)
pint never, former, current PMH use
pint duration of PMH use (never, former, current: <5, current: ≥5 years)
pint type of PMH use (never, former, current E-alone, current E+P).
Discussion
The current study of racial/ethnic differences in postmenopausal breast cancer risk by tumor HoR status builds on previous studies in the full MEC Study population [12–14]. While our results in the overall population are comparable to these previous analyses, an additional ~1500 incident invasive breast cancers were diagnosed during the extended follow-up period, enabling us to investigate associations stratified for both HoR status and racial/ethnic group. Compared with White women, Native Hawaiian women had a higher risk of developing HoR+ disease, whereas Latina and African American women had a lower risk. Risk of developing HoR- disease was higher in African American than in White women. In analyses comparing hormone-related risk factor associations by race/ethnicity, we observed interactions for the associations with OC use for both HoR+ and HoR- disease, and a suggestive interaction for the association with obesity for HoR+ disease. Compared with never users, Latinas who ever used OCs had a higher risk of developing HoR+ disease while these associations were not apparent in other racial/ethnic groups. Obesity increased the risk for HoR+ disease across all racial/ethnic groups with more pronounced associations observed for Native Hawaiians, Japanese Americans and Latinas. In relation to HoR- disease risk, compared with never use, ever OC use increased risk among African Americans irrespective of duration of use, while increased risk was seen only for ≥5 years of OC use in Japanese Americans.
Racial/ethnic differences in breast cancer risk
The higher risk of HoR- breast cancer in African American relative to White women observed in our study is well established [1, 21, 22]. Although Japanese Americans have also been suggested to have lower risk of HoR+ disease than White women [2], we did not observe risk differences for HoR+ or HoR- breast cancer between Japanese American and White women. Japanese American and Native Hawaiian women are often included together in the combined Asian/Pacific Islander group; our study adds to the literature showing substantial variation in breast cancer risk and subtypes within this group [2, 3, 23, 24]. Importantly, Native Haw7aiians in our study had a higher risk of developing HoR+ breast cancer than Whites, which is in line with a previous study using data from the Hawaii Tumor Registry [3]. High, comparable, breast cancer incidence rates in Native Hawaiian and Non-Hispanic White women are also reported in SEER, however this study did not account for breast cancer subtype [25].
Hormone-related factors and hormone receptor positive breast cancer risk
When investigating risk factors for HoR+ disease, we observed significant interaction with race/ethnicity for OC use, which was positively associated with HoR+ breast cancer risk only in Latinas. While we are not aware of any studies that have evaluated OC use and HoR+ breast cancer risk in Latinas, previous studies in White and African American women suggest a null association [6, 26] or a small increase in HoR+ disease risk with OC use among recent OC users [6, 27, 28]. Time since last use was not available in the current study. Although the association with PMH use did not differ significantly by race/ethnicity, we observed positive associations between current PMH use at baseline, but not former use, and risk of ER+ breast cancer only among White, Japanese American and Latina women. Consistent with earlier reports (mostly in White women) [4], we observed stronger associations for E+P formulations. Use of exogenous hormones, both OCs and PMH, may influence exposure to estrogen and progesterone, and has been linked to increased breast density and breast tumor development and growth [29–31]. An obese BMI was associated with increased HoR+ breast cancer risk in all racial/ethnic groups, although the association in White women did not reach significance. When accounting for PMH use, we also noted that positive associations between an obese BMI and HoR+ breast cancer were limited to women who reported never use of PMH, consistent with previous reports [32]. We additionally observed positive associations for an overweight BMI in Japanese American and White women. Adipose tissue is the main source of estrogen in postmenopausal women [33] and our findings are in line with previous studies reporting positive associations between BMI and risk of HoR+ breast cancer risk among postmenopausal White, African American, and Asian women [34–36]. Results for Latinas have been less consistent [34, 36]. Additionally, associations for overweight were stronger in Japanese Americans (41% increased risk) than in White women (17%) in our study. While a higher proportion of Japanese Americans had a lean BMI out of all the racial/ethnic groups in our study, they have greater visceral adiposity and liver fat than Whites after adjusting for total fat mass [37], which could contribute to the strong association observed for these women in our study and others [34, 38]. Although we observed no interactions with race/ethnicity for other reproductive factors, there were some notable differences by racial/ethnic group. Parity was inversely associated with HoR+ disease risk in White, Japanese American and Latina women. Later ages at first birth and natural menopause were positively associated with risk in White, Japanese American and Latina women, whereas a later age at menarche was inversely associated with HoR+ disease risk in African American and Japanese American women. These findings are generally in line with previous studies on reproductive factors in White [6, 8], African American [9, 11, 39], and Latina women [40, 41]. Reports including multiple racial/ethnic groups indicate that, compared with associations in White women, associations for HoR+ disease may be somewhat weaker in African American women [11, 39] and (in contrast to our findings for Japanese Americans) in Asian American women [39].
Hormone-related factors and hormone receptor negative breast cancer risk
Associations between hormone-related factors and HoR- breast cancer risk did not differ by race/ethnicity with the exception of OC duration; OC use for five or more years (vs never use) was associated with a ≥63% increased HoR- disease risk in White, Japanese American and African American women while there was no association for Latinas and Native Hawaiians. In line with our findings for postmenopausal African Americans, two previous studies (including both pre- and postmenopausal African American women) reported increased (menopausal status adjusted) ER- breast cancer risk with both ever and increasing duration of OC use [26, 28]. Studies in mostly White women similarly suggest a positive association between OC use and HoR- disease risk [6, 27, 42, 43]. In the present study, current (baseline) vs never use of PMH E-alone influenced ER- disease risk only in Latinas (+81%). A recent meta-analysis of studies in mostly White women reported positive associations with current/recent PMH use, with stronger associations for E+P than E-alone formulation [4]. In agreement with most [6, 8, 9, 11, 41], but not all [10, 39] previous studies in White, African American, Asian American, and Latina women, most reproductive factors such as parity and age at menopause were not associated with HoR- breast cancer risk in our study. To our knowledge ours is the first study to report on a range of hormone-related risk factors for HoR- disease in Japanese American and Native Hawaiian women.
Strengths and limitations
The strengths of our study include the prospective, racially/ethnically diverse, population-based design of the MEC Study, which is largely representative of its source populations [15]. This enabled us to conduct a comprehensive comparative analysis of risk factors for HoR+ and HoR- breast cancer across racial/ethnic groups. To our knowledge this study is the first to evaluate risk factors for HoR+ and HoR- breast cancer in Native Hawaiian women, a population with very high breast cancer incidence rates [3, 44]. Our study had a number of limitations. First, the MEC Study baseline questionnaire did not include questions on breastfeeding and we were unable to investigate or adjust for this factor that has been linked to decreased risk of both HoR+ and HoR- breast cancer in parous women [8, 41, 45]. A second limitation is the relatively small sample size for Native Hawaiian women; the lack of associations for hormone-related factors, with the exception of BMI and HoR+ disease, may be due to the lower power to detect associations in this group. Third, the modest number of HoR- cases limited our ability to investigate associations for HoR- cancers by race/ethnicity. Additionally, data on HER2 status were available for less than half (46%) of the breast cancer cases in our study; while we know that triple negative tumors (i.e., ER-PR-HER2-) are more common in African Americans [1, 46] our study was insufficiently powered to investigate racial/ethnic differences by combined ER, PR, and HER2 expression. Finally, due to multiple comparisons some of our findings may be due to chance and should be confirmed in additional studies.
Conclusion
Breast cancer risk differs by race/ethnicity and HoR status; compared to White women, risk of HoR+ disease is higher among Native Hawaiian and lower in Japanese American and African American women, while risk of HoR- disease is higher for African American women. Although we observed differences in race/ethnicity-specific associations for several hormone-related factors and breast cancer risk, particularly OC use and BMI, these factors do not appear to explain the observed racial/ethnic differences in HoR+ and HoR- breast cancer risk. Our study highlights differences in hormone-related risk factor associations across racial/ethnic groups that can help improve breast cancer prevention strategies for these understudied populations. Further work is needed to identify novel risk factors that may explain racial/ethnic differences in risk for hormone receptor positive and negative breast cancer.
Supplementary Material
Novelty and impact statement:
Racial/ethnic differences in hormone receptor positive and negative breast cancer risk are well-documented; the factors contributing to these differences less so. The authors compared hormone-related risk factors for breast cancer among a multiethnic population. They identified a number of notable differences in the impact of established risk factors - particularly oral contraceptive use and BMI - between racial/ethnic groups. These results reinforce the need to consider race/ethnicity in breast cancer studies and prevention recommendations.
Acknowledgements and disclosures
Funding: The Multiethnic Cohort is supported by National Cancer Institute grant U01CA164973 (L. Le Marchand, C. Haiman, L. Wilkens). The content is solely the responsibility of the authors and do not represent the official view of NIH.
Abbreviations:
- BMI
body mass index
- CI
95% confidence interval
- ER
estrogen receptor
- ER+
estrogen receptor positive
- ER-
estrogen receptor negative
- ER/PR
estrogen receptor and progesterone receptor
- ER+/PR+
estrogen receptor and progesterone receptor positive
- ER-/PR-
estrogen receptor and progesterone receptor negative
- HoR
hormone receptor (i.e. ER and/or ER/PR)
- HoR+
hormone receptor positive
- HoR-
hormone receptor negative
- HR
hazard ratio
- IQR
interquartile range
- MEC
Multiethnic Cohort
- OC
oral contraceptive
- PMH
postmenopausal hormones
- PMH E-alone
postmenopausal hormones containing only estrogen
- PMH E+P
postmenopausal hormones containing both estrogen and progesterone
- SEER
Surveillance, Epidemiology, and End Results
Footnotes
Data Availability Statement: For information on applications to gain access to data from the Multiethnic Cohort please see: https://www.uhcancercenter.org/for-researchers/mec-data-sharing. Further details and other data that support the findings of this study are available from the corresponding author upon request.
Ethics Statement: Our study was approved by the Institutional Review Boards at the University of Hawaii and the University of Southern California. Informed consent to participate in the Multiethnic Cohort Study was obtained via return of the completed baseline questionnaire.
Conflict of Interest: The authors declare that they have no conflicts of interest.
References
- 1.DeSantis CE, Ma J, Gaudet MM, Newman LA, Miller KD, Goding Sauer A, Jemal A, Siegel RL. Breast cancer statistics, 2019. CA Cancer J Clin 2019;69:438–51. [DOI] [PubMed] [Google Scholar]
- 2.Parise C, Caggiano V. Disparities in the risk of the ER/PR/HER2 breast cancer subtypes among Asian Americans in California. Cancer epidemiology 2014;38:556–62. [DOI] [PubMed] [Google Scholar]
- 3.Loo LWM, Williams M, Hernandez BY. The high and heterogeneous burden of breast cancer in Hawaii: A unique multiethnic U.S. Population. Cancer epidemiology 2019;58:71–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Collaborative Group on Hormonal Factors in Breast C. Type and timing of menopausal hormone therapy and breast cancer risk: individual participant meta-analysis of the worldwide epidemiological evidence. Lancet 2019;394:1159–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Fortner RT, Katzke V, Kühn T, Kaaks R. Obesity and Breast Cancer. In: Pischon T, Nimptsch K. Obesity and Cancered. Cham: Springer International Publishing, 2016:43–65. [DOI] [PubMed] [Google Scholar]
- 6.Ritte R, Tikk K, Lukanova A, Tjonneland A, Olsen A, Overvad K, Dossus L, Fournier A, Clavel-Chapelon F, Grote V, Boeing H, Aleksandrova K, Trichopoulou A, Lagiou P, Trichopoulos D, Palli D, Berrino F, Mattiello A, Tumino R, Sacerdote C, Quiros JR, Buckland G, Molina-Montes E, Chirlaque MD, Ardanaz E, Amiano P, Bueno-de-Mesquita HB, van Gils CH, Peeters PH, Wareham N, Khaw KT, Key TJ, Travis RC, Weiderpass E, Dumeaux V, Lund E, Sund M, Andersson A, Romieu I, Rinaldi S, Vineis P, Merritt MA, Riboli E, Kaaks R. Reproductive factors and risk of hormone receptor positive and negative breast cancer: a cohort study. BMC Cancer 2013;13:584. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Yang XR, Chang-Claude J, Goode EL, Couch FJ, Nevanlinna H, Milne RL, Gaudet M, Schmidt MK, Broeks A, Cox A, Fasching PA, Hein R, Spurdle AB, Blows F, Driver K, Flesch-Janys D, Heinz J, Sinn P, Vrieling A, Heikkinen T, Aittomaki K, Heikkila P, Blomqvist C, Lissowska J, Peplonska B, Chanock S, Figueroa J, Brinton L, Hall P, Czene K, Humphreys K, Darabi H, Liu J, Van ‘t Veer LJ, van Leeuwen FE, Andrulis IL, Glendon G, Knight JA, Mulligan AM, O’Malley FP, Weerasooriya N, John EM, Beckmann MW, Hartmann A, Weihbrecht SB, Wachter DL, Jud SM, Loehberg CR, Baglietto L, English DR, Giles GG, McLean CA, Severi G, Lambrechts D, Vandorpe T, Weltens C, Paridaens R, Smeets A, Neven P, Wildiers H, Wang X, Olson JE, Cafourek V, Fredericksen Z, Kosel M, Vachon C, Cramp HE, Connley D, Cross SS, Balasubramanian SP, Reed MW, Dork T, Bremer M, Meyer A, Karstens JH, Ay A, Park-Simon TW, Hillemanns P, Arias Perez JI, Menendez Rodriguez P, Zamora P, Benitez J, Ko YD, Fischer HP, Hamann U, Pesch B, Bruning T, Justenhoven C, Brauch H, Eccles DM, Tapper WJ, Gerty SM, Sawyer EJ, Tomlinson IP, Jones A, Kerin M, Miller N, McInerney N, Anton-Culver H, Ziogas A, Shen CY, Hsiung CN, Wu PE, Yang SL, Yu JC, Chen ST, Hsu GC, Haiman CA, Henderson BE, Le Marchand L, Kolonel LN, Lindblom A, Margolin S, Jakubowska A, Lubinski J, Huzarski T, Byrski T, Gorski B, Gronwald J, Hooning MJ, Hollestelle A, van den Ouweland AM, Jager A, Kriege M, Tilanus-Linthorst MM, Collee M, Wang-Gohrke S, Pylkas K, Jukkola-Vuorinen A, Mononen K, Grip M, Hirvikoski P, Winqvist R, Mannermaa A, Kosma VM, Kauppinen J, Kataja V, Auvinen P, Soini Y, Sironen R, Bojesen SE, Orsted DD, Kaur-Knudsen D, Flyger H, Nordestgaard BG, Holland H, Chenevix-Trench G, Manoukian S, Barile M, Radice P, Hankinson SE, Hunter DJ, Tamimi R, Sangrajrang S, Brennan P, McKay J, Odefrey F, Gaborieau V, Devilee P, Huijts PE, Tollenaar RA, Seynaeve C, Dite GS, Apicella C, Hopper JL, Hammet F, Tsimiklis H, Smith LD, Southey MC, Humphreys MK, Easton D, Pharoah P, Sherman ME, Garcia-Closas M. Associations of breast cancer risk factors with tumor subtypes: a pooled analysis from the Breast Cancer Association Consortium studies. J Natl Cancer Inst 2011;103:250–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Fortner RT, Sisti J, Chai B, Collins LC, Rosner B, Hankinson SE, Tamimi RM, Eliassen AH. Parity, breastfeeding, and breast cancer risk by hormone receptor status and molecular phenotype: results from the Nurses’ Health Studies. Breast cancer research : BCR 2019;21:40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Ambrosone CB, Zirpoli G, Ruszczyk M, Shankar J, Hong CC, McIlwain D, Roberts M, Yao S, McCann SE, Ciupak G, Hwang H, Khoury T, Jandorf L, Bovbjerg DH, Pawlish K, Bandera EV. Parity and breastfeeding among African-American women: differential effects on breast cancer risk by estrogen receptor status in the Women’s Circle of Health Study. Cancer causes & control : CCC 2014;25:259–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Palmer JR, Viscidi E, Troester MA, Hong CC, Schedin P, Bethea TN, Bandera EV, Borges V, McKinnon C, Haiman CA, Lunetta K, Kolonel LN, Rosenberg L, Olshan AF, Ambrosone CB. Parity, lactation, and breast cancer subtypes in African American women: results from the AMBER Consortium. J Natl Cancer Inst 2014;106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Cui Y, Deming-Halverson SL, Shrubsole MJ, Beeghly-Fadiel A, Fair AM, Sanderson M, Shu XO, Kelley MC, Zheng W. Associations of hormone-related factors with breast cancer risk according to hormone receptor status among white and African American women. Clin Breast Cancer 2014;14:417–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Pike MC, Kolonel LN, Henderson BE, Wilkens LR, Hankin JH, Feigelson HS, Wan PC, Stram DO, Nomura AM. Breast cancer in a multiethnic cohort in Hawaii and Los Angeles: risk factor-adjusted incidence in Japanese equals and in Hawaiians exceeds that in whites. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 2002;11:795–800. [PubMed] [Google Scholar]
- 13.Setiawan VW, Monroe KR, Wilkens LR, Kolonel LN, Pike MC, Henderson BE. Breast cancer risk factors defined by estrogen and progesterone receptor status: the multiethnic cohort study. American journal of epidemiology 2009;169:1251–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.White KK, Park SY, Kolonel LN, Henderson BE, Wilkens LR. Body size and breast cancer risk: the Multiethnic Cohort. International journal of cancer 2012;131:E705–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kolonel LN, Altshuler D, Henderson BE. The multiethnic cohort study: exploring genes, lifestyle and cancer risk. Nature reviews. Cancer 2004;4:519–27. [DOI] [PubMed] [Google Scholar]
- 16.Kolonel LN, Henderson BE, Hankin JH, Nomura AM, Wilkens LR, Pike MC, Stram DO, Monroe KR, Earle ME, Nagamine FS. A multiethnic cohort in Hawaii and Los Angeles: baseline characteristics. American journal of epidemiology 2000;151:346–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Gram IT, Park SY, Maskarinec G, Wilkens LR, Haiman CA, Le Marchand L. Smoking and breast cancer risk by race/ethnicity and oestrogen and progesterone receptor status: the Multiethnic Cohort (MEC) study. International journal of epidemiology 2019;48:501–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Wang M, Spiegelman D, Kuchiba A, Lochhead P, Kim S, Chan AT, Poole EM, Tamimi R, Tworoger SS, Giovannucci E, Rosner B, Ogino S. Statistical methods for studying disease subtype heterogeneity. Stat Med 2016;35:782–800. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Lunn M, McNeil D. Applying Cox regression to competing risks. Biometrics 1995;51:524–32. [PubMed] [Google Scholar]
- 20.Modeling Greenland S. and variable selection in epidemiologic analysis. Am J Public Health 1989;79:340–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Clarke CA, Keegan TH, Yang J, Press DJ, Kurian AW, Patel AH, Lacey JV, Jr. Age-specific incidence of breast cancer subtypes: understanding the black-white crossover. J Natl Cancer Inst 2012;104:1094–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hall IJ, Moorman PG, Millikan RC, Newman B. Comparative analysis of breast cancer risk factors among African-American women and White women. American journal of epidemiology 2005;161:40–51. [DOI] [PubMed] [Google Scholar]
- 23.Chuang E, Paul C, Flam A, McCarville K, Forst M, Shin S, Vahdat L, Swistel A, Simmons R, Osborne M. Breast cancer subtypes in Asian-Americans differ according to Asian ethnic group. J Immigr Minor Health 2012;14:754–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Gomez SL, Glaser SL, Horn-Ross PL, Cheng I, Quach T, Clarke CA, Reynolds P, Shariff-Marco S, Yang J, Lee MM, Satariano WA, Hsing AW. Cancer research in Asian American, Native Hawaiian, and Pacific Islander populations: accelerating cancer knowledge by acknowledging and leveraging heterogeneity. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 2014;23:2202–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Liu L, Noone AM, Gomez SL, Scoppa S, Gibson JT, Lichtensztajn D, Fish K, Wilkens LR, Goodman MT, Morris C, Kwong S, Deapen D, Miller BA. Cancer incidence trends among native Hawaiians and other Pacific Islanders in the United States, 1990–2008. J Natl Cancer Inst 2013;105:1086–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Rosenberg L, Boggs DA, Wise LA, Adams-Campbell LL, Palmer JR. Oral contraceptive use and estrogen/progesterone receptor-negative breast cancer among African American women. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 2010;19:2073–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Collaborative Group on Hormonal Factors in Breast C. Breast cancer and hormonal contraceptives: collaborative reanalysis of individual data on 53 297 women with breast cancer and 100 239 women without breast cancer from 54 epidemiological studies. Lancet 1996;347:1713–27. [DOI] [PubMed] [Google Scholar]
- 28.Bethea TN, Rosenberg L, Hong CC, Troester MA, Lunetta KL, Bandera EV, Schedin P, Kolonel LN, Olshan AF, Ambrosone CB, Palmer JR. A case-control analysis of oral contraceptive use and breast cancer subtypes in the African American Breast Cancer Epidemiology and Risk Consortium. Breast cancer research : BCR 2015;17:22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Byrne C, Ursin G, Martin CF, Peck JD, Cole EB, Zeng D, Kim E, Yaffe MD, Boyd NF, Heiss G, McTiernan A, Chlebowski RT, Lane DS, Manson JE, Wactawski-Wende J, Pisano ED. Mammographic Density Change With Estrogen and Progestin Therapy and Breast Cancer Risk. J Natl Cancer Inst 2017;109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Bonfiglio R, Di Pietro ML. The impact of oral contraceptive use on breast cancer risk: State of the art and future perspectives in the era of 4P medicine. Semin Cancer Biol 2021. [DOI] [PubMed] [Google Scholar]
- 31.Kuhl H, Schneider HP. Progesterone--promoter or inhibitor of breast cancer. Climacteric : the journal of the International Menopause Society 2013;16 Suppl 1:54–68. [DOI] [PubMed] [Google Scholar]
- 32.Munsell MF, Sprague BL, Berry DA, Chisholm G, Trentham-Dietz A. Body mass index and breast cancer risk according to postmenopausal estrogen-progestin use and hormone receptor status. Epidemiol Rev 2014;36:114–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Perel E, Killinger DW. The interconversion and aromatization of androgens by human adipose tissue. J Steroid Biochem 1979;10:623–7. [DOI] [PubMed] [Google Scholar]
- 34.Bandera EV, Maskarinec G, Romieu I, John EM. Racial and ethnic disparities in the impact of obesity on breast cancer risk and survival: a global perspective. Adv Nutr 2015;6:803–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Pischon T, Nimptsch K. Obesity and Risk of Cancer: An Introductory Overview. Recent Results Cancer Res 2016;208:1–15. [DOI] [PubMed] [Google Scholar]
- 36.John EM, Sangaramoorthy M, Hines LM, Stern MC, Baumgartner KB, Giuliano AR, Wolff RK, Slattery ML. Body size throughout adult life influences postmenopausal breast cancer risk among hispanic women: the breast cancer health disparities study. Cancer Epidemiol Biomarkers Prev 2015;24:128–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Lim U, Monroe KR, Buchthal S, Fan B, Cheng I, Kristal BS, Lampe JW, Hullar MA, Franke AA, Stram DO, Wilkens LR, Shepherd J, Ernst T, Le Marchand L. Propensity for Intra-abdominal and Hepatic Adiposity Varies Among Ethnic Groups. Gastroenterology 2019;156:966–75 e10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Le Marchand L, Wilkens LR, Castelfranco AM, Monroe KR, Kristal BS, Cheng I, Maskarinec G, Hullar MA, Lampe JW, Shepherd JA, Franke A, Ernst T, Lim U. Circulating Biomarker Score for Visceral Fat and Risks of Incident Colorectal and Postmenopausal Breast Cancer: The Multiethnic Cohort Adiposity Phenotype Study. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 2020;29:966–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.John EM, Phipps AI, Hines LM, Koo J, Ingles SA, Baumgartner KB, Slattery ML, Wu AH. Menstrual and reproductive characteristics and breast cancer risk by hormone receptor status and ethnicity: The Breast Cancer Etiology in Minorities study. International journal of cancer 2020;147:1808–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Sweeney C, Baumgartner KB, Byers T, Giuliano AR, Herrick JS, Murtaugh MA, Slattery ML. Reproductive history in relation to breast cancer risk among Hispanic and non-Hispanic white women. Cancer causes & control : CCC 2008;19:391–401. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Sangaramoorthy M, Hines LM, Torres-Mejia G, Phipps AI, Baumgartner KB, Wu AH, Koo J, Ingles SA, Slattery ML, John EM. A Pooled Analysis of Breastfeeding and Breast Cancer Risk by Hormone Receptor Status in Parous Hispanic Women. Epidemiology (Cambridge, Mass.) 2019;30:449–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Sweeney C, Giuliano AR, Baumgartner KB, Byers T, Herrick JS, Edwards SL, Slattery ML. Oral, injected and implanted contraceptives and breast cancer risk among U.S. Hispanic and non-Hispanic white women. International journal of cancer 2007;121:2517–23. [DOI] [PubMed] [Google Scholar]
- 43.Huang WY, Newman B, Millikan RC, Schell MJ, Hulka BS, Moorman PG. Hormone-related factors and risk of breast cancer in relation to estrogen receptor and progesterone receptor status. American journal of epidemiology 2000;151:703–14. [DOI] [PubMed] [Google Scholar]
- 44.Hawaii Tumor Registry. Hawaii Cancer At A Glance 2012–2016 2020. [Google Scholar]
- 45.Islami F, Liu Y, Jemal A, Zhou J, Weiderpass E, Colditz G, Boffetta P, Weiss M. Breastfeeding and breast cancer risk by receptor status--a systematic review and meta-analysis. Annals of oncology : official journal of the European Society for Medical Oncology 2015;26:2398–407. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Chen L, Li CI, Tang MT, Porter P, Hill DA, Wiggins CL, Cook LS. Reproductive Factors and Risk of Luminal, HER2-Overexpressing, and Triple-Negative Breast Cancer Among Multiethnic Women. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 2016;25:1297–304. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.