Geczik et al. Breast Cancer Research (2022) 24:9 https://doi.org/10.1186/s13058-022-01500-8 RESEARCH ARTICLE Open Access Measured body size and serum estrogen metabolism in postmenopausal women: the Ghana Breast Health Study Ashley M. Geczik1, Roni T. Falk1, Xia Xu2, Daniel Ansong3, Joel Yarney4, Beatrice Wiafe‑Addai5, Lawrence Edusei4, Florence Dedey4, Verna Vanderpuye4, Nicholas Titiloye6, Ernest Adjei6, Francis Aitpillah6, Ernest Osei‑Bonsu6, Joseph Oppong6, Richard Biritwum7, Kofi Nyarko7, Seth Wiafe8, Baffour Awuah6, Joe‑Nat Clegg‑Lamptey4, Thomas U. Ahearn1, Jonine Figueroa9, Montserrat Garcia‑Closas1, Louise A. Brinton1† and Britton Trabert1*† Abstract Background: Several anthropometric measures have been associated with hormone‑related cancers, and it has been shown that estrogen metabolism in postmenopausal women plays an important role in these relationships. However, little is known about circulating estrogen levels in African women, and the relevance to breast cancer or breast cancer risk factors. To shed further light on the relationship of anthropometric factors and estrogen levels in African women, we examined whether measured body mass index (BMI), waist‑to‑hip ratio (WHR), height, and self‑ reported body size were associated with serum estrogens/estrogen metabolites in a cross‑sectional analysis among postmenopausal population‑based controls of the Ghana Breast Health Study. Methods: Fifteen estrogens/estrogen metabolites were quantified using liquid chromatography‑tandem mass spectrometry in serum samples collected from postmenopausal female controls enrolled in the Ghana Breast Health Study, a population‑based case–control study conducted in Accra and Kumasi. Geometric means (GMs) of estrogens/ estrogen metabolites were estimated using linear regression, adjusting for potential confounders. Results: Measured BMI (≥ 30 vs. 18.5–24.9 kg/m2) was positively associated with parent estrogens (multivariable adjusted GM for unconjugated estrone: 78.90 (66.57–93.53) vs. 50.89 (43.47–59.59), p‑value < 0.0001; and unconju‑ gated estradiol: 27.83 (21.47–36.07) vs. 13.26 (10.37–16.95), p‑value < 0.0001). Independent of unconjugated estradiol, measured BMI was associated with lower levels of 2‑pathway metabolites and higher levels of 16‑ketoestradriol. Similar patterns of association were found with WHR; however, the associations were not entirely independent of BMI. Height was not associated with postmenopausal estrogens/estrogen metabolite levels in African women. Conclusions: We observed strong associations between measured BMI and parent estrogens and estrogen metabo‑ lite patterns that largely mirrored relations that have previously been associated with higher breast cancer risk in postmenopausal White women. The consistency of the BMI‑estrogen metabolism associations in our study with those *Correspondence: britton.trabert@nih.gov †Co‑senior authors: Louise A. Brinton and Britton Trabert. 1 Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health (NIH), DHSS, 9609 Medical Center Dr., Bethesda, MD 20892, USA Full list of author information is available at the end of the article © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/b y/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/z ero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Geczik et al. Breast Cancer Research (2022) 24:9 Page 2 of 12 previously noted among White women suggests that estrogens likely explain part of the BMI‑postmenopausal breast cancer risk in both groups. These findings merit evaluation in Black women, including prospective studies. Keywords: Measured body mass index, Height, Waist‑to‑hip ratio, Estrogen metabolism, Postmenopausal Black women Introduction compared with lean women, and estrone was 60% higher Anthropometric measures, such as body mass index [12]. In a study in the Women’s Health Initiative Obser- (BMI) and height, are associated with increased post- vational Study, BMI was positively associated with cir- menopausal breast cancer risk in studies conducted culating parent estrogens and reduced methylation of among predominantly non-Hispanic White women catechol estrogen metabolites [13]. These findings are [1–3]. Recent meta-analyses have estimated 15% higher consistent with the patterns associated with higher breast risk of postmenopausal breast cancer [2] in overweight cancer risk [11], however, the study populations evalu- or obese compared with lean women. Central adiposity ated to date have predominantly included non-Hispanic measured by waist circumference (WC) or waist-to-hip White women. ratio (WHR) has also shown strong positive associations Little is known about estrogen levels in African with breast cancer risk [3, 4] although these are attenu- women, and the relevance to breast cancer or breast can- ated after accounting for BMI. One of the hypothesized cer risk factors. It has been reported previously that US mechanisms for these relations is that overweight and Black women have higher circulating estradiol independ- obese women have elevated levels of circulating estro- ent of adiposity and experience less reduction in levels gens [5, 6], as adipocytes produce estrogens from andro- with weight loss than White women [14]. Further, it has gens via aromatase activity [7]. Height may indicate early been hypothesized that these hormonal differences may life nutritional status and high levels of endogenous pro- contribute to differential patterns of breast cancer inci- liferative hormone such as estrogens. Adult height is also dence and mortality by race [15, 16]. Thus, investigations a risk factor for breast cancer, most notably hormone to understand sex steroid hormone associations with receptor positive tumors in both pre- and postmenopau- anthropometric characteristics in African women will sal women [8]. contribute to understanding breast cancer risk factors in In Westernized countries and based on predominantly this population. We had the opportunity to examine this White study populations, it has been established that using the population-based postmenopausal controls of estrogens are key hormones in breast carcinogenesis. the Ghana Breast Health Study conducted in two large Parent estrogens (estradiol and estrone) stimulate cell metropolitan areas where obesity has been demonstrated proliferation via estrogen receptor-mediated pathways. to be a breast cancer risk factor [17]. When parent estrogens are hydroxylated at one of three carbon positions of the steroid ring, metabolites are Materials and methods formed along three different pathways (i.e., 2-, 4-, and Study population 16-hydroxylation pathways). The carcinogenicity of indi- For the current cross-sectional analysis, we utilized data vidual estrogen metabolites can vary. Catechol estrogen from postmenopausal female controls enrolled in the metabolites can stimulate cell proliferation via estrogen Ghana Breast Health Study, a multi-disciplinary pop- receptor-dependent pathways and induce DNA damage ulation-based case–control study. The methodology of directly by forming quinone DNA adducts or indirectly the original study is described in more detail elsewhere via redox cycling [9]. Methylation of the catechol estro- [17, 18]. In brief, population controls were selected on gen metabolites can prevent mutagenic quinone forma- the basis of frequency matching to breast cancer cases tion [10]. Prospective epidemiologic studies of breast (enrolled at Korle Bu Teaching Hospital in Accra and cancer risk have suggested that metabolism favoring par- Komfo Anoyke Teaching Hospital and Peace and Love ent estrogens into the 2- and 4-pathway over the 16-path- Hospital in Kumasi) on age, with similar restrictions way is associated with lower postmenopausal breast regarding case catchment areas, and at least 1  year of cancer risk [11]. residence in these areas. The 2011 Ghanaian census Epidemiologic evidence supports that among women was used to select enumeration areas (areas comprised not using exogenous hormones, circulating levels of of ~ 750 residents) of the districts from which cases were estrogens are higher in overweight and obese US women expected to derive. Trained census workers enumerated [1, 5, 6]. In a pooled analysis of eight studies, estra- all households with respect to the sex and age of the resi- diol was 83% higher in postmenopausal obese women dents. When households were enumerated, a brochure G eczik et al. Breast Cancer Research (2022) 24:9 Page 3 of 12 was left explaining the study and encouraging participa- analyses as slight (1 or 2), average (3 or 4), slightly heavy tion should an individual be selected for inclusion. After (5 or 6), or heavy (7, 8, or 9). Weight, standing height, selected areas had been enumerated, individuals were waist circumference, and hip circumference were meas- randomly selected to approximate the age distribution ured by trained staff at an in-person interview. Measure- of female breast cancer cases expected during the study. ments were made at least twice in the same setting and Study personnel visited subjects’ homes to determine eli- averaged. BMI (kg/m2) was calculated as weight (in kilo- gibility, inform them of study selection and invite them grams) divided by height (in meters squared). WHR was for a hospital visit. calculated based on measured waist circumference (cm) Controls were approached for in-person interviews divided by measured hip circumference (cm). by trained personnel who recorded information on standardized questionnaires. Interviews were generally Laboratory assays conducted in the hospitals, although a few were admin- Details of the hormone assay have been published previ- istered at the subjects’ homes. The interview response ously [19–21]. Briefly, stable isotope dilution liquid chro- rate was 91.9%. Of 2106 potentially eligible controls, we matography-tandem mass spectrometry (LC–MS/MS) excluded the following women who at interview indi- was used to quantify 15 estrogens and estrogen metabo- cated that they were premenopausal (n = 1237), did not lites including: estrone, estradiol, 2-pathway metabolites know their menopause status (n = 9), reported current (2-hydroxyestrone, 2-methoxyestrone, 2-hydroxyestra- hormone use (n = 10), or could not provide informa- diol, 2-methoxyestradiol, and 2-hydroxyestrone-3-me- tion on current hormone use (n = 16). An additional 199 thyl ether); 4-pathway metabolites (4-hydroxyestrone, women were excluded because they did not have enough 4-methoxyestrone, and 4-methoxyestradiol); and serum volume available for the estrogen/estrogen metab- 16α-pathway metabolites (16α-hydroxyestrone, estriol, olite assays. Of the 635 samples from which estrogens 16-ketoestradiol, 16-epiestriol, and 17-epiestriol). This were measured, an additional 3 were excluded because method detects 15 estrogens and estrogen metabolites of missing age and 47 because of reports of menstrual in serum which circulate, at least in part, as sulfated bleeding on the blood draw questionnaire, suggesting and/or glucuronidated conjugates to facilitate storage, that they were either perimenopausal or premenopausal. transport, and excretion. Five of the estrogens (estrone, The final analytic population consisted of 585 postmeno- estradiol, estriol, 2-methoxyestrone and 2-methoxyestra- pausal controls that had information on at least one vari- diol) were also measured in unconjugated forms in cir- able related to body size. culation. For those metabolites with both combined and unconjugated measurements, the concentration of the Exposure assessment conjugated form was calculated as the difference between The study questionnaire focused on established breast the combined estrogen measurement and the uncon- cancer risk factors including demographic factors, men- jugated estrogen measurement; for estradiol that calcu- strual and reproductive characteristics, family history lation was (conjugated estradiol = combined estradiol of breast cancer, medical history, occupational history, – unconjugated estradiol). The limit of detection for each and anthropometric and physical activity variables. estrogen and estrogen metabolite measured using this Anthropometric measures included a 9-scale pictogram LC–MS/MS assay was 10  fg on column (approximately for participants to self-identify body shape (Fig. 1), with 0.33–0.37  pmol/L) [19, 22]. There were no samples in 1 corresponding to the slimmest body shape and 9 the the current study with undetectable levels for any of the heaviest body shape. Based on the self-reported body hormones measured. Laboratory coefficients of variation size silhouette scale (1–9), women were categorized for (CV) of blinded quality control duplicates distributed 1 2 3 4 5 6 7 8 9 Fig. 1 Body size silhouettes as shown in study questionnaire. Question asked respondents to indicate which silhouette best represented their current body shape Geczik et al. Breast Cancer Research (2022) 24:9 Page 4 of 12 within and across batches were < 5% for all hormones combinations of BMI comparisons (25–29.9 vs. 18.5– measured. Intraclass correlation coefficients (ICCs) 24.9 kg/m2; ≥ 30 vs. 18.5–24.9 kg/m2; ≥ 30 vs. 25–29.9 kg/ ranged from 0.97 to 0.998 with a median value of 0.99. m2) and six current body size comparisons (heavy vs. slight, slightly heavy vs. slight, average vs. slight, heavy vs. Statistical analysis average, slightly heavy vs. average, and heavy vs. slightly After log-transformation of data to improve normality, heavy). geometric means (GM) (pmol/L) of individual serum Finally, because underlying diseases may influence estrogens/estrogen metabolite concentration by expo- the associations, we performed sensitivity analyses after sure categories were estimated using linear regression excluding women diagnosed with a history of diabetes adjusting for potential confounders: age at blood draw, (n = 49) and excluding women with low BMIs (< 18.5 kg/ blood draw year, smoking status (never, former, current, m2) (n = 20). unknown/missing), time since menopause (≤ 2, 3–5, All statistical tests were two-sided with 5% type I error. 6–10, > 10  years, missing), and oral contraceptive use Q-values reflecting the false discovery rates (FDR) were (ever, never). We performed a test for trend by includ- calculated to address multiple comparisons (25 tests per ing the exposure in the model as an ordinal variable. The exposure) separately for the original model, the model percent change (%Δ) in GMs between the highest and the with additional adjustment for unconjugated estradiol, lowest categories was estimated by taking the ratio of the and the model with additional adjustment for measured GM difference between the two categories over the GM BMI (where applicable). Analyses were conducted with of the reference category, multiplied by 100. We statisti- SAS version 9 (SAS Institute). cally tested for the difference using a Wald test. Several secondary analyses were performed. First, for Results BMI (and other measures of body size), we additionally Among 585 postmenopausal African women, the aver- adjusted for unconjugated estradiol to examine whether age age at blood draw was 56.8 years (standard deviation the associations with other estrogen metabolites were 8.1 years) (Table 1). Most women reported never smok- driven by their correlations with unconjugated estradiol, ing (95.0%), not having a history of diabetes (87.9%), the estrogen most strongly correlated with measured never using oral contraceptives (84.8%), and having BMI (Spearman r = 0.43). Next, we investigated whether given birth to at least 3 children (81.6%). The distribu- BMI was associated with altered patterns of estrogen tion of measured BMI categories among study partici- metabolism, using pathway groups. We compared the pants was as follows: 3.4% underweight (< 18.5  kg/m2), mean proportions of parent estrogens out of summed 33.0% healthy weight (18.5- < 25.0  kg/m2), 28.9% over- estrogens/estrogen metabolites across BMI categories weight (25.0- < 30.0 kg/m2), and 26.3% obese (≥ 30.0 kg/ with adjustment for the summed concentration of estro- m2); 8.4% of women had missing data on either measured gens/estrogen metabolites. Further, because 2-, 4-, and height and/or weight. 16-pathway metabolites (“child metabolites”) are metab- olized from a limited pool of shared precursors (parent Measured BMI estrogens), an increase in the level of one downstream Obese BMI (≥ 30  kg/m2 vs. 18.5- < 25.0  kg/m2) was pathway indicates a reduction in levels of other compet- associated with higher levels of parent estrogens ing pathways. To address this, we modeled proportions of (unconjugated estrone: 78.90 (66.57–93.53) vs. 50.89 each child metabolite pathway group (2-pathway metab- (43.47–59.59), p-value < 0.0001; unconjugated estra- olites: 2-catechols [2-hydroxyestrone, 2-hydroxyestra- diol: 27.83 (21.47–36.07) vs. 13.26 (10.37–16.95), diol] and methylated 2-catechols [2-methoxyestrone and p-value < 0.0001) (Fig.  2, Table  2). Positive associations 2-methoxyestradiol]; 4-pathway metabolites: 4-catechols between high BMI and 2-hydroxyestone, 4-hydrox- [4-hydroxyestrone] and methylated 4-catechols [4-meth- yestrone, and most of the 16-alpha pathway estrogen oxyestrone, 4-methoxyestradiol]; 16-pathway metabo- metabolites (5 out of 7) were also observed, while asso- lites: [16α-hydroxyestrone, estriol, 16-ketoestradiol, ciations with the 2- and 4- pathway methylated catechols 16-epiestriol, 17-epiestriol]) out of summed child metab- were null (Table  2). After adjustment for unconjugated olites, with adjustment for the summed concentration estradiol, the positive associations between high BMI and of child metabolites. This approach estimates the asso- many of the estrogen metabolites noted did not persist ciation with replacement of one pathway group for other but instead we observed inverse associations between pathway groups while holding summed child metabolites higher BMI and many of the 2-pathway estrogen metabo- constant. We tested for any difference across BMI catego- lites (including 2-hydroxyestrone, 2-hydroxyestradiol, ries using global F test; if there was a significant differ- and 2-methoxyestrone), a suggestion of an inverse asso- ence (p < 0.05), we performed pairwise t-tests for three ciation between higher BMI and 4-methoxyestrone, G eczik et al. Breast Cancer Research (2022) 24:9 Page 5 of 12 Table 1 Distribution of demographic and measured anthropometric factors in Ghana Breast Health Study Postmenopausal Controls (n = 585) Mean Standard Deviation Age at blood draw 56.8 8.1 Age at menopause 48.3 5.1 N Percent Time since menopause ≤ 2 104 17.8 3–5 105 18.0 6–10 130 22.2 > 10 153 26.2 Missing 93 15.9 Year of blood draw 2013 229 39.2 2014 193 33.0 2015 163 27.9 Smoking status Current 0 0.0 Former 4 0.7 Never 556 95.0 Unknown 5 0.9 Missing 20 3.4 Diabetes Ever 49 8.4 Never 514 87.9 Unknown 22 3.8 Age at menarche < 14 112 19.2 15 171 29.2 16 107 18.3 ≥ 17 109 18.6 Unknown 86 14.7 Parity/Number of births Nulliparous 16 2.7 1–2 91 15.6 3–4 200 34.2 5 + 277 47.4 Unknown 1 0.2 Oral contraceptive use Ever 89 15.2 Never 496 84.8 Unknown 0 0.0 Age at menopause < 45 95 16.2 45–54 349 59.7 ≥ 55 48 8.2 Unknown 93 15.9 BMI (kg/m2) Underweight (< 18.5) 20 3.4 Healthy weight (18.5–24.9) 193 33.0 Geczik et al. Breast Cancer Research (2022) 24:9 Page 6 of 12 Table 1 (continued) N Percent Overweight (25.0–29.9) 169 28.9 Obese (30 +) 154 26.3 Unknown/Missing 49 8.4 Current body size from pictogram Slight (Fig. 1, 1 or 2) 64 10.9 Average (3 or 4) 189 32.3 Slightly heavy (5 or 6) 175 29.9 Heavy (7, 8, or 9) 87 14.9 Unknown 70 12.0 Waist-to-hip ratio < 0.86 180 30.77 0.86–0.93 176 30.09 > 0.93 182 31.11 Missing 47 8.03 Height (cm) < 155 168 28.7 155–159.9 170 29.1 160 + 233 39.8 Missing 14 2.4 Fig. 2 The proportion of parent estrogen concentrations increased and child pathway metabolites overall decreased (left panel) across BMI categories. When evaluating the proportion of the 2‑, 4‑, and 16‑pathway estrogen metabolites out of the combined concentration of metabolites, 2‑pathway metabolism decreased, and 16‑pathway metabolism increased across increasing BMI categories (right panel) and a positive association between higher BMI and After adjusting for unconjugated estradiol, the posi- 16-ketoestradiol (Additional file 1: Table S1). tive association between estrone and self-reported body size remained, while the associations with the estrogen Self‑reported body size metabolites attenuated substantially (Additional file  1: Consistent with the associations between circulating Table S2). estrogens and measured BMI, women who self-reported the highest body size categories (7,8, or 9 = heaviest) had Waist‑to‑hip ratio the highest estrogen levels (parent estrogens, 2-hydrox- Estrone levels and many of the 16-pathway metabo- yestrone, 4-hydroxyestrone, and five of seven 16-path- lites increased across increasing tertiles of WHR way metabolites) compared with women who reported (WHR > 0.93 (T3) vs. < 0.86 (T1): estrone 332.58 the lowest body size categories (1 or 2 = slight) (Table 3). (247.85–446.28) vs. 265.63 (196.64–358.83), G eczik et al. Breast Cancer Research (2022) 24:9 Page 7 of 12 Table 2 Geometric means (pmol/L) and 95% CIs of serum estrogens/estrogen metabolites by current body mass index in postmenopausal control women not using menopausal hormone therapy in the Ghana Breast Health Study Geometric mean (95% CI) p‑trend %Δ p‑diff Underweight Healthy weight Overweight Obese (30 + kg/m2) (< 18.5 kg/m2) (18.5–24.9 kg/m2) (25.0–29.9 kg/m2) N 20 193 169 154 Median BMI (kg/m2) 18 22 27 34 Estrone 144.97 (88.25–238.13) 224.99 (170.21– 318.61 (239.30–424.21) 448.63 (336.80– < 0.0001 99.4 < 0.0001 297.39) 597.58) Unconjugated 41.70 (32.20–54.00) 50.89 (43.47–59.59) 64.64 (54.90–76.12) 78.90 (66.57–93.53) < 0.0001 55.0 < 0.0001 Conjugated 93.85 (50.12–175.77) 165.03 (117.93– 251.15 (178.34–353.68) 366.55 (260.70– < 0.0001 122.1 < 0.0001 230.94) 515.40) Estradiol 16.77 (10.73–26.23) 19.68 (14.73–26.31) 29.63 (21.93–40.02) 41.81 (31.07–56.27) < 0.0001 112.4 < 0.0001 Unconjugated 11.50 (8.05–16.42) 13.26 (10.37–16.95) 19.80 (15.25–25.71) 27.83 (21.47–36.07) < 0.0001 109.9 < 0.0001 Conjugated 2.24 (0.83–6.03) 3.32 (1.65–6.67) 5.02 (2.50–10.08) 7.32 (3.67–14.59) < 0.0001 120.4 < 0.0001 2‑Hydroxyestrone 45.44 (37.43–55.16) 51.36 (45.24–58.31) 53.48 (46.66–61.30) 58.42 (51.15–66.73) 0.01 13.7 0.03 2‑Hydroxyestradiol 12.35 (8.23–18.53) 9.38 (7.04–12.51) 8.39 (6.22–11.31) 8.50 (6.37–11.36) 0.06 − 9.4 0.24 2‑Methoxyestrone 19.75 (14.41–27.06) 21.17 (17.44–25.70) 21.43 (17.55–26.17) 23.20 (18.91–28.46) 0.15 9.6 0.20 Unconjugated 8.54 (5.90–12.36) 8.56 (6.84–10.71) 9.07 (7.22–11.40) 8.42 (6.68–10.62) 0.93 − 1.6 0.84 Conjugated 8.67 (5.29–14.21) 9.57 (7.26–12.60) 9.94 (7.58–13.02) 12.75 (9.61–16.92) 0.02 33.3 0.02 2‑Methoxyestradiol 10.55 (7.23–15.38) 12.92 (10.00–16.68) 12.82 (10.01–16.42) 13.17 (10.18–17.03) 0.44 1.9 0.78 Unconjugated 2.88 (2.02–4.11) 2.86 (2.27–3.60) 2.89 (2.27–3.68) 2.94 (2.32–3.73) 0.65 3.0 0.62 Conjugated 6.39 (3.61–11.29) 9.09 (6.33–13.04) 8.36 (5.88–11.89) 8.89 (6.22–12.70) 0.76 − 2.2 0.83 2‑Hydroxyestrone‑ 3.52 (2.56–4.83) 3.68 (2.97–4.58) 3.69 (2.96–4.60) 3.79 (3.06–4.71) 0.61 3.0 0.69 3‑methyl ether 4‑Hydroxyestrone 6.29 (4.35–9.09) 6.62 (4.76–9.21) 7.13 (5.16–9.84) 8.36 (6.06–11.52) 0.004 26.3 0.007 4‑Methoxyestrone 3.95 (2.96–5.27) 3.58 (2.87–4.46) 3.62 (2.89–4.54) 3.60 (2.89–4.47) 0.85 0.5 0.94 4‑Methoxyestradiol 1.48 (1.04–2.11) 1.54 (1.24–1.92) 1.45 (1.18–1.78) 1.58 (1.28–1.97) 0.73 2.6 0.71 16α‑Hydroxyestrone 23.04 (14.78–35.90) 29.08 (20.92–40.42) 37.98 (27.19–53.04) 50.14 (36.11–69.62) < 0.0001 72.4 < 0.0001 Estriol 82.19 (58.33–115.80) 82.48 (65.95–103.16) 95.37 (75.16–121.02) 105.42 (83.45–133.17) 0.001 27.8 0.002 Unconjugated 9.28 (7.62–11.29) 9.43 (8.38–10.62) 10.05 (8.97–11.27) 9.21 (8.21–10.33) 0.78 − 2.4 0.56 Conjugated 70.77 (47.79–104.79) 70.06 (54.67–89.77) 82.52 (63.50–107.24) 94.71 (73.17–122.60) 0.0006 35.2 0.0007 16‑Ketoestradiol 11.97 (8.87–16.16) 17.24 (13.76–21.59) 19.72 (15.58–24.97) 27.89 (22.13–35.14) < 0.0001 61.8 < 0.0001 16‑Epiestriol 17.41 (11.22–27.02) 16.63 (13.07–21.16) 18.96 (14.91–24.09) 20.47 (16.06–26.08) 0.011 23.0 0.008 17‑Epiestriol 28.50 (19.39–41.90) 25.07 (18.83–33.38) 26.02 (19.77–34.26) 27.13 (20.51–35.88) 0.44 8.2 0.27 Geometric means adjusted for age at blood draw (continuous), blood draw year (2013, 2014, 2015), smoking status (never, former, current, missing), diabetes (yes, no, missing), time since menopause (≤ 2, 3–5, 6–10, > 10, missing), ever used oral contraceptives (yes, no, missing) p-trend was estimated using the Wald test for ordinal BMI category %Δ indicates the percentage change in estrogen/estrogen metabolite levels, comparing women with current BMI ≥ 30 vs. 18.5–24.9 kg/m2, and was estimated by taking the ratio of the geometric mean difference in estrogen/estrogen metabolite levels between women with current BMI ≥ 30 vs. 18.5–24.9 kg/m2 to the geometric mean of women with current BMI 18.5–24.9 kg/m2, multiplied by 100 p-diff was estimated using the Wald test and indicates a p-value comparing estrogen/estrogen metabolite levels of women with current BMI ≥ 30 vs. 18.5–24.9 kg/m2 Bold p-values represent FDR q-value ≤ 0.05 p-trend = 0.03) (Table  4). In contrast, unconjugated (Additional file  1: Table  S3). The associations with 2-methoxyestrone and unconjugated 2-methoxyestra- estrone and 16-pathway metabolites were no longer diol levels decreased across increasing tertiles of WHR apparent in models adjusted for BMI, but the inverse [9.79 (7.81–12.26) vs. 8.36 (6.67–10.49) vs. 8.25 (6.64– associations between unconjugated 2-methoxyestrone, 10.24), 0.03; 3.20 (2.57–3.98) vs. 2.84 (2.28–3.53) vs. unconjugated 2-methoxyestradiol, and estriol levels 2.76 (2.22–3.43), 0.02, respectively]. These associations and WHR remained after adjustment for BMI (Addi- persisted in models adjusted for unconjugated estradiol tional file 1: Table S3). Geczik et al. Breast Cancer Research (2022) 24:9 Page 8 of 12 Table 3 Geometric means (pmol/L) and 95% CIs of serum estrogens/estrogen metabolites by current body size assessed using pictogram in postmenopausal control women not using menopausal hormone therapy in the Ghana Breast Health Study Geometric mean (95% CI) p‑trend %Δ p‑diff Slight Average Slightly heavy Heavy N 64 189 175 87 Median BMI (kg/m2) 22 25 29 31 Estrone 277.88 (168.63– 472.03 (299.00– 562.73 (351.76– 748.98 (449.60– < 0.0001 58.7 < 0.0001 457.91) 745.19) 900.24) 1,247.71) Unconjugated 52.63 (40.17–68.96) 66.56 (52.15–84.95) 74.68 (57.90–96.31) 94.03 (70.48–125.45) < 0.0001 41.3 < 0.0001 Conjugated 189.16 (105.62– 397.12 (241.72– 474.59 (284.04– 637.85 (365.42– < 0.0001 60.6 < 0.0001 338.78) 652.42) 792.96) 1,113.38) Estradiol 28.47 (18.21–44.53) 41.16 (27.80–60.95) 49.45 (32.77–74.60) 66.41 (42.42–103.95) < 0.0001 61.3 < 0.0001 Unconjugated 19.55 (13.51–28.28) 30.52 (22.13–42.09) 33.98 (24.18–47.74) 47.33 (31.93–70.18) < 0.0001 55.1 < 0.0001 Conjugated 3.86 (1.59–9.38) 6.37 (2.83–14.32) 10.37 (4.54–23.68) 10.18 (4.16–24.94) 0.0001 59.8 0.0011 2‑Hydroxyestrone 52.77 (45.52–61.18) 62.25 (55.20–70.19) 66.33 (57.88–76.03) 70.17 (58.93–83.57) 0.0008 12.7 0.0005 2‑Hydroxyestradiol 12.07 (8.09–18.01) 10.33 (7.24–14.74) 10.47 (7.25–15.12) 10.17 (6.93–14.92) 0.30 − 1.6 0.18 2‑Methoxyestrone 17.04 (13.44–21.59) 18.17 (14.85–22.23) 18.99 (15.31–23.57) 21.42 (16.80–27.31) 0.02 17.9 0.02 Unconjugated 6.21 (4.89–7.89) 6.58 (5.44–7.96) 6.38 (5.18–7.87) 6.79 (5.30–8.70) 0.62 3.1 0.44 Conjugated 8.59 (5.85–12.62) 10.03 (7.58–13.27) 11.61 (8.57–15.73) 14.57 (10.30–20.61) 0.001 45.3 0.003 2‑Methoxyestradiol 9.77 (7.65–12.49) 10.17 (8.26–12.51) 10.40 (8.36–12.94) 9.52 (7.50–12.08) 0.83 − 6.4 0.79 Unconjugated 2.63 (2.09–3.32) 2.86 (2.38–3.45) 2.88 (2.36–3.52) 2.84 (2.29–3.52) 0.49 − 0.7 0.40 Conjugated 6.59 (4.56–9.54) 6.96 (5.23–9.27) 6.78 (4.98–9.22) 6.42 (4.59–8.97) 0.72 − 7.8 0.87 2‑Hydroxyestrone‑ 3.00 (2.32–3.89) 3.12 (2.45–3.96) 3.29 (2.56–4.24) 3.37 (2.56–4.44) 0.19 8.2 0.27 3‑methyl ether 4‑Hydroxyestrone 6.60 (4.51–9.65) 7.60 (5.46–10.58) 8.46 (5.96–12.00) 9.47 (6.58–13.63) 0.0007 24.6 0.002 4‑Methoxyestrone 3.07 (2.52–3.74) 3.00 (2.58–3.48) 3.19 (2.72–3.76) 3.08 (2.55–3.71) 0.63 2.6 0.98 4‑Methoxyestradiol 1.51 (1.16–1.95) 1.58 (1.29–1.92) 1.61 (1.29–2.00) 1.63 (1.30–2.05) 0.42 3.3 0.44 16α‑Hydroxyestrone 21.42 (14.99–30.61) 30.66 (22.61–41.57) 29.25 (21.00–40.73) 37.29 (25.86–53.75) 0.002 21.6 0.0001 Estriol 85.14 (69.73–103.95) 92.92 (79.22–109.00) 95.18 (79.99–113.26) 113.48 (90.92–141.64) 0.014 22.1 0.01 Unconjugated 9.53 (8.17–11.13) 9.74 (8.64–10.98) 10.44 (9.09–11.99) 9.86 (8.50–11.44) 0.28 1.3 0.61 Conjugated 73.92 (59.01–92.59) 80.71 (67.41–96.64) 81.25 (66.65–99.06) 102.40 (80.16–130.81) 0.017 26.9 0.01 16‑Ketoestradiol 11.97 (9.51–15.07) 15.74 (13.33–18.57) 15.69 (12.99–18.95) 19.92 (15.74–25.21) 0.0002 26.6 < 0.0001 16‑Epiestriol 16.64 (12.80–21.63) 19.37 (15.52–24.17) 19.88 (15.86–24.92) 22.85 (17.56–29.74) 0.008 18.0 0.005 17‑Epiestriol 19.86 (13.54–29.15) 20.60 (14.42–29.42) 20.95 (14.54–30.19) 20.68 (14.15–30.22) 0.70 0.4 0.71 Geometric means adjusted for age at blood draw (continuous), blood draw year (2013, 2014, 2015), smoking status (never, former, current, missing), diabetes (yes, no, missing), time since menopause (≤ 2, 3–5, 6–10, > 10, missing), ever used oral contraceptives (yes, no, missing) p-trend was estimated using the Wald test for ordinal body size category %Δ indicates the percentage change in estrogen/estrogen metabolite levels, comparing women with heavy body size category to average sized women, and was estimated by taking the ratio of the geometric mean difference in estrogen/estrogen metabolite levels between women reporting heavy body size categories minus average body size category to the geometric mean of women with average body size, multiplied by 100 (we could do this in the table to slight) p-diff was estimated using the Wald test and indicates a p-value for comparing estrogen/estrogen metabolite levels of women with current heavy body size vs. average body size Bold p-values represent FDR ≤ 0.05 Height comparisons using FDR, most associations with a nom- There were no clear patterns of increasing or decreas- inal p-value less than or equal to 0.01 had q-values less ing estrogen metabolism across categories of increasing than or equal to 0.05 (as indicated with bold font in the measured height (Additional file 1: Table S4). manuscript tables). Sensitivity analyses Discussion Results did not change substantively after exclud- In this novel cross-sectional study of circulating estro- ing women with diabetes at blood draw or those who gen metabolism in postmenopausal African women, had an underweight BMI. When considering multiple measured BMI was positively associated with higher G eczik et al. Breast Cancer Research (2022) 24:9 Page 9 of 12 Table 4 Geometric means (pmol/L) and 95% CIs of serum estrogens/estrogen metabolites by waist‑to‑hip ratio (WHR) tertile in postmenopausal control women not using menopausal hormone therapy in the Ghana Breast Health Study Geometric mean (95% CI) (model 1) p‑trend %Δ < 0.86 0.86–0.93 > 0.93 N 180 176 182 Median BMI (kg/m2) 24.0 26.0 29.0 Estrone 265.63 (196.64–358.83) 280.96 (208.05–379.43) 332.58 (247.85–446.28) 0.03 25.2 Unconjugated 59.62 (50.09–70.96) 57.50 (48.32–68.41) 65.72 (55.47–77.88) 0.18 10.2 Conjugated 195.47 (136.03–280.86) 215.62 (149.69–310.61) 261.63 (183.72–372.58) 0.02 33.8 Estradiol 25.85 (19.75–33.82) 25.68 (19.62–33.60) 30.30 (23.19–39.59) 0.16 17.2 Unconjugated 17.65 (13.75–22.66) 17.14 (13.38–21.96) 20.13 (15.63–25.92) 0.23 14 Conjugated 4.11 (2.15–7.84) 4.49 (2.34–8.61) 5.05 (2.64–9.67) 0.28 23 2‑Hydroxyestrone 55.59 (48.69–63.48) 50.51 (44.35–57.51) 54.15 (47.66–61.52) 0.66 − 2.6 2‑Hydroxyestradiol 8.99 (6.59–12.27) 8.83 (6.53–11.96) 9.19 (6.83–12.37) 0.78 2.3 2‑Methoxyestrone 22.16 (18.06–27.20) 21.67 (17.70–26.54) 21.32 (17.62–25.80) 0.58 − 3.8 Unconjugated 9.79 (7.81–12.26) 8.36 (6.67–10.49) 8.25 (6.64–10.24) 0.03 − 15.7 Conjugated 9.40 (7.00–12.63) 10.65 (7.98–14.20) 10.96 (8.44–14.23) 0.20 16.6 2‑Methoxyestradiol 12.73 (9.96–16.28) 12.52 (9.72–16.12) 12.84 (10.02–16.46) 0.89 0.9 Unconjugated 3.20 (2.57–3.98) 2.84 (2.28–3.53) 2.76 (2.22–3.43) 0.02 − 13.7 Conjugated 7.85 (5.57–11.06) 8.42 (5.92–11.99) 8.98 (6.37–12.65) 0.18 14.3 2‑Hydroxyestrone‑3‑methyl ether 3.66 (2.94–4.57) 3.64 (2.94–4.52) 3.76 (3.05–4.64) 0.72 2.6 4‑Hydroxyestrone 7.04 (5.14–9.66) 7.27 (5.25–10.06) 7.35 (5.37–10.05) 0.61 4.3 4‑Methoxyestrone 3.93 (3.15–4.90) 3.41 (2.75–4.22) 3.58 (2.90–4.43) 0.14 − 8.8 4‑Methoxyestradiol 1.46 (1.18–1.80) 1.51 (1.22–1.87) 1.57 (1.28–1.93) 0.26 7.6 16α‑Hydroxyestrone 31.89 (22.93–44.34) 32.62 (23.35–45.57) 42.96 (31.11–59.32) 0.003 34.7 Estriol 82.26 (66.81–101.28) 91.05 (74.10–111.88) 102.32 (83.39–125.56) 0.01 24.4 Unconjugated 9.21 (8.28–10.25) 10.13 (9.01–11.39) 9.40 (8.47–10.43) 0.62 2 Conjugated 71.43 (56.65–90.07) 77.59 (61.76–97.49) 90.38 (72.03–113.40) 0.01 26.5 16‑Ketoestradiol 17.82 (14.10–22.51) 18.56 (14.71–23.41) 22.85 (18.09–28.86) 0.002 28.2 16‑Epiestriol 17.45 (13.81–22.04) 17.79 (14.15–22.36) 19.87 (15.80–24.98) 0.08 13.9 17‑Epiestriol 25.37 (19.23–33.49) 27.06 (20.42–35.87) 26.27 (19.89–34.70) 0.63 3.5 Geometric means adjusted for age at blood draw (continuous), blood draw year (2013, 2014, 2015), smoking status (never, former, current, missing), diabetes (yes, no, missing), time since menopause (≤ 2, 3–5, 6–10, > 10, missing), ever used oral contraceptives (yes, no, missing) p-trend was estimated using the Wald test for ordinal WHR category %Δ indicates the percentage change in estrogen/estrogen metabolite levels, comparing women with highest WHR tertile to lowest WHR tertile, and was estimated by taking the ratio of the geometric mean difference in estrogen/estrogen metabolite levels between women with highest WHR tertile minus lowest WHR tertile to the geometric mean of women with lowest WHR tertile, multiplied by 100 Bold p-values represent FDR ≤ 0.05 levels of most of the measured estrogens. After adjust- distribution. Height was not associated with differences ment for unconjugated estradiol, which showed in estrogen metabolism. the strongest association with BMI, positive asso- Our findings of positive associations between current ciations between BMI and estrone as well as BMI and BMI and parent estrogen levels are consistent with stud- 16-ketoestradiol persisted, whereas BMI was inversely ies conducted predominantly among White women [12]. associated with most of the 2-pathway metabolites. These findings are also in line with biological evidence Like measured BMI, self-reported body size was posi- supporting the major source of estrogens in postmeno- tively associated with higher levels of most estrogens/ pausal women derives from aromatization of andro- estrogen metabolites. Consistent with this, we observed gens (androstenedione and testosterone) to estrogens BMI attenuated the associations between WHR and (estriol and estradiol) in adipose tissue. To date, a lim- estrogens; together this suggests metabolite concentra- ited number of studies have examined current BMI in tions were driven by overall adiposity rather than fat relation to estrogen metabolism. Earlier studies using Geczik et al. Breast Cancer Research (2022) 24:9 Page 10 of 12 ELISA-based assays measured only two estrogen metab- lack of signal for the metabolites with self-reported body olites thought to be the most and the least carcinogenic: size could be due to the imprecise nature of the picto- 16α-hydroxyestrone and 2-hydroxyestrone, respectively gram or from collapsing across categories. [23–26]. Results from these earlier studies supported Current adult height in African women was not associ- an inverse association between adiposity and the ratio ated with differences in estrogen metabolism. This find- of urinary 2-hydroxyestrone to 16α-hydroxyestrone in ing is consistent findings in predominantly non-Hispanic both pre- and postmenopausal women [24, 25, 27]. In a White postmenopausal women from the Women’s Health study nested in the Prostate, Lung, Colorectal, and Ovar- Initiative Observational Study [13]. ian Cancer Screening Trial, self-reported BMI was posi- Limitations of the current study include the use of tively correlated with all 15 serum estrogens/estrogen measured circulating estrogens/estrogen metabolites at metabolites among postmenopausal women [28]; how- a single point in time. However, a previous study using ever, that study did not assess whether the associations our same assay has shown moderate to high 1-year ICCs with estrogen metabolites remained after accounting in postmenopausal women [31], suggesting that meas- for correlations with unconjugated estradiol. In a study ured serum estrogens/estrogen metabolites may also nested in the Women’s Health Initiative Observational adequately represent postmenopausal levels over at Study, positive associations between increasing BMI and least 1 year. While we used established BMI cutpoints to estrogen metabolites did not remain after adjustment for facilitate comparison with previous research conducted unconjugated estradiol; however, consistent with the cur- among predominantly White postmenopausal women, rent study, inverse associations between BMI and meth- measures of obesity are not well established for African ylated 2-catechols became apparent after adjusting for populations, and as such may not be as informative of unconjugated estradiol in postmenopausal women [13]. disease risk. This latter study also demonstrated that obese women Our study has notable strengths. Measurement error appear in general to be less likely to metabolize parent for the anthropometric measures in the current study estrogens into child metabolites, but more likely to favor was reduced by using measured height, weight, and waist metabolism of parent estrogens into 16-pathway estrogen and hip circumferences, as compared to other studies metabolites over 2- or 4-pathway metabolites. Our find- that used self-reported height/weight, etc. Other study ings corroborate these results and provide novel informa- strengths include the use of the high-performance LC– tion about patterns of estrogen metabolism with BMI in MS/MS assay that provided a comprehensive evaluation African women. of individual estrogens/estrogen metabolites with high In our study, most of the WHR-metabolite associa- reliability, sensitivity, and specificity. Further, use of a tions were not independent of BMI. In studies where large sample size limited to postmenopausal women not body fat distribution was measured by dual-energy using hormones at blood collection and careful adjust- X-ray absorptiometry scan [29] or measured WHR ment for potential confounders assessed at blood collec- [30], central obesity was not associated with circulating tion increased the validity of the results. unconjugated estradiol independent of BMI among pre- dominantly White postmenopausal women, suggesting Conclusions that body fat distribution does not impact circulating In this comprehensive analysis of measured anthropo- estradiol beyond that of overall adiposity. However, the metrics and serum estrogens/estrogen metabolites in inverse associations between WHR and unconjugated African women, we observed strong, positive associa- 2-methoxyestrone and unconjugated 2-methoxyestradiol tions between measured BMI and parent estrogens in and positive association between WHR and estriol per- postmenopausal women. After adjustment for uncon- sisted in models additionally adjusted for BMI. This sug- jugated estradiol, measured BMI was also associated gests that, in African women, the association between with lower levels of 2-pathway metabolites and higher these metabolites and WHR may represent a pattern of levels of 16-ketoestradiol. In  studies of predominantly estrogen exposure that is potentially relevant for disease White women, it has been suggested that endogenous risk and that warrants further exploration. As such, the estrogen metabolism at least partially mediates the independent association between these metabolites and association between BMI and increased risk of post- WHR may represent a pattern of estrogen exposure that menopausal estrogen receptor positive (ER +) breast is potentially relevant for disease risk in African women cancer, given the observation that increasing BMI is and warrants further exploration. associated with higher levels of parent estrogens and Measured BMI and self-reported body size were asso- reduced concentrations of 2-pathway metabolites. The ciated with similar increases in parent estrogens, but only consistency of the BMI-estrogen metabolism associa- BMI measurements were related to the metabolites. The tion in this study of African women suggests that these Geczik et al. Breast Cancer Research (2022) 24:9 Page 11 of 12 same mechanisms may be relevant for postmenopausal Declarations ER + breast cancer risk in African women. Our data also suggest that WHR in African women may explain Ethics approval and consent to participateAll questionnaires were administered after obtaining written informed con‑ differences in circulating estrogen metabolite levels sent on forms approved by institutional review boards in the U.S. and Ghana. independent of BMI. These findings merit further eval- uation in prospective studies. Consent for publicationAll authors have read the manuscript, accept responsibility for the manu‑ scripts contents and agree that the manuscript is ready for submission to your Supplementary Information journal. The online version contains supplementary material available at https:// doi. Competing interests org/ 10.1 186/ s13058‑ 022‑ 01500‑8. The authors declare that they have no competing interests. Additional file 1: Author details Supplemental Tables summarizing geometric means 1 Division of Cancer Epidemiology and Genetics, National Cancer Institute, (pmol/L) and 95% CIs of serum estrogens/estrogen metabolites by current National Institutes of Health (NIH), DHSS, 9609 Medical Center Dr., Bethesda, body mass index with additional adjustment for unconjugated estra‑ MD 20892, USA. 2 Protein Characterization Laboratory, Leidos‑Frederick, diol (Table S1); by current body size assessed using pictogram with addi‑ Inc., Frederick National Laboratory for Cancer Research, Frederick, MD, USA. tional adjustment for unconjugated estradiol (Table S2); by waist‑to‑hip 3 Kwame Nkrumah University of Science and Technology, Kumasi, Ghana. ratio (WHR) tertile with additional adjustment for measured BMI (Table S3); 4 Korle Bu Teaching Hospital, Accra, Ghana. 5 Peace and Love Hospital, Kumasi, and by height tertile (Table S4) in postmenopausal control women not Ghana. 6 Komfo Anokye Teaching Hospital, Kumasi, Ghana. 7 University using menopausal hormone therapy in the Ghana Breast Health Study. of Ghana, Accra, Ghana. 8 Loma Linda University, School of Public Health, Loma Linda, CA, USA. 9 The University of Edinburgh, Cancer Research UK Edinburgh Acknowledgements Center, Edinburgh, Scotland. The success of this investigation would not have been possible without exceptional teamwork and the diligence of the field staff who oversaw Received: 28 July 2021 Accepted: 10 January 2022 the recruitment, interviews and collection of data from study subjects. Special thanks are due to the following individuals: Korle Bu Teaching Hospital,Accra—Dr. Adu‑Aryee, Obed Ekpedzor, Angela Kenu, Victoria Okyne, Naomi Oyoe Ohene Oti, Evelyn Tay; Komfo Anoyke Teaching Hospital, Kumasi— Marion Alcpaloo, Bernard Arhin, Emmanuel Asiamah, Isaac Boakye, References Samuel Ka‑chungu and; Peace and Love Hospital, Kumasi—Samuel Amanama, 1. Bruning PF, et al. Body measurements, estrogen availability and the risk of Emma Abaidoo, Prince Agyapong, Thomas Agyei‑Ansong, Debora Boateng, human breast cancer: a case‑control study. Int J Cancer. 1992;51(1):14–9. Margaret Frempong, Bridget Nortey Mensah, Richard Opoku, and Kofi Owusu 2. Cheraghi Z, et al. Effect of body mass index on breast cancer during Gyimah. The study was further enhanced by surgical expertise provided by Dr. premenopausal and postmenopausal periods: a meta‑analysis. PLoS ONE. Lisa Newman of the University of Michigan and by pathological expertise pro‑ 2012;7(12):e51446. vided by Drs. Stephen Hewitt and Petra Lenz of the National Cancer Institute, 3. Connolly BS, et al. A meta‑analysis of published literature on waist‑to‑hip and Dr. Maire A. Duggan from the Cumming School of Medicine, University ratio and risk of breast cancer. Nutr Cancer. 2002;44(2):127–38. of Calgary, Canada. Study management assistance was received from Ricardo 4. Harvie M, Hooper L, Howell AH. Central obesity and breast cancer risk: a Diaz, Shelley Niwa, Usha Singh, Ann Truelove and Michelle Brotzman at Westat, systematic review. Obes Rev. 2003;4(3):157–73. Inc. Appreciation is also expressed to the many women who agreed to partici‑ 5. Hankinson SE, et al. Alcohol, height, and adiposity in relation to estrogen pate in the study and to provide information and biospecimens in hopes of and prolactin levels in postmenopausal women. J Natl Cancer Inst. preventing and improving outcomes of breast cancer in Ghana. 1995;87(17):1297–302. 6. Lukanova A, et al. Body mass index, circulating levels of sex‑steroid hor‑ Authors’ contributions mones, IGF‑I and IGF‑binding protein‑3: a cross‑sectional study in healthy AMG: statistical analysis, data interpretation, writing, and critical review of the women. Eur J Endocrinol. 2004;150(2):161–71. manuscript. BT: study conception, study design, statistical analysis, interpre‑ 7. Lorincz AM, Sukumar S. Molecular links between obesity and breast tation of results, writing, and critical review of the manuscript. LAB: study cancer. Endocr Relat Cancer. 2006;13(2):279–92. conception, study design, data acquisition, data interpretation, writing, and 8. Zhang B, et al. Height and breast cancer risk: evidence from pro‑ critical review of the manuscript. XX: laboratory analysis and critical review of spective studies and mendelian randomization. J Natl Cancer Inst. the manuscript. JF: study design, data acquisition, and critical review of the 2015;107(11):djv219. manuscript. TUA, MGC: data/study management and critical review of the 9. Yager JD, Davidson NE. Estrogen carcinogenesis in breast cancer. N Engl J manuscript. RTK, DA, JY, BWA, LE, FD, VV, NT, EA, FA, EOB, JO, RB, KN, SW, BA, Med. 2006;354(3):270–82. JNCL: data acquisition and critical review of the manuscript. All authors read 10. Zahid M, et al. Inhibition of catechol‑O‑methyltransferase increases estro‑ and approved the final manuscript. gen‑DNA adduct formation. Free Radic Biol Med. 2007;43(11):1534–40. 11. Sampson JN, et al. Association of estrogen metabolism with breast Funding cancer risk in different cohorts of postmenopausal women. Cancer Res. This research was supported in part by funds from the intramural research 2017;77(4):918–25. program of the National Cancer Institute, National Institutes of Health. 12. Key TJ, et al. Body mass index, serum sex hormones, and breast cancer risk in postmenopausal women. J Natl Cancer Inst. 2003;95(16):1218–26. Availability of data and material 13. Oh H, et al. Anthropometric measures and serum estrogen metabolism The datasets generated or analyzed for the current study are not publicly avail‑ in postmenopausal women: the Women’s Health Initiative Observational able due to data privacy of patients but are available from the corresponding Study. Breast Cancer Res. 2017;19(1):28. author upon reasonable request. 14. Stolzenberg‑Solomon RZ, et al. Sex hormone changes during weight loss and maintenance in overweight and obese postmenopausal African‑American and non‑African‑American women. Breast Cancer Res. 2012;14(5):R141. 15. Joslyn SA. Hormone receptors in breast cancer: racial differences in distri‑ bution and survival. Breast Cancer Res Treat. 2002;73(1):45–59. Geczik et al. Breast Cancer Research (2022) 24:9 Page 12 of 12 16. Pinheiro SP, et al. Racial differences in premenopausal endogenous hormones. Cancer Epidemiol Biomark Prev. 2005;14(9):2147–53. 17. Brinton LA, et al. Design considerations for identifying breast can‑ cer risk factors in a population‑based study in Africa. Int J Cancer. 2017;140(12):2667–77. 18. Brinton L, et al. Factors contributing to delays in diagnosis of breast can‑ cers in Ghana, West Africa. Breast Cancer Res Treat. 2017;162(1):105–14. 19. Xu X, et al. Quantitative measurement of endogenous estrogens and estrogen metabolites in human serum by liquid chromatography‑tan‑ dem mass spectrometry. Anal Chem. 2007;79(20):7813–21. 20. Brinton LA, et al. Serum estrogens and estrogen metabolites and endo‑ metrial cancer risk among postmenopausal women. Cancer Epidemiol Biomark Prev. 2016;25(7):1081–9. 21. Trabert B, et al. Circulating estrogens and postmenopausal ovarian cancer risk in the women’s health initiative observational study. Cancer Epide‑ miol Biomark Prev. 2016;25(4):648–56. 22. Loud JT, et al. Circulating estrogens and estrogens within the breast among postmenopausal BRCA1/2 mutation carriers. Breast Cancer Res Treat. 2014;143(3):517–29. 23. Matthews CE, et al. Physical activity, body size, and estrogen metabolism in women. Cancer Causes Control. 2004;15(5):473–81. 24. Coker AL, et al. Re: ethnic differences in estrogen metabolism in healthy women. J Natl Cancer Inst. 1997;89(1):89–90. 25. Schneider J, et al. Effects of obesity on estradiol metabolism: decreased formation of nonuterotropic metabolites. J Clin Endocrinol Metab. 1983;56(5):973–8. 26. Modugno F, et al. Obesity, hormone therapy, estrogen metabolism and risk of postmenopausal breast cancer. Int J Cancer. 2006;118(5):1292–301. 27. Fishman J, Boyar RM, Hellman L. Influence of body weight on estradiol metabolism in young women. J Clin Endocrinol Metab. 1975;41(5):989–91. 28. Schairer C, et al. Quantifying the role of circulating unconjugated estra‑ diol in mediating the body mass index‑breast cancer association. Cancer Epidemiol Biomark Prev. 2016;25(1):105–13. 29. Mahabir S, et al. Usefulness of body mass index as a sufficient adiposity measurement for sex hormone concentration associations in postmeno‑ pausal women. Cancer Epidemiol Biomark Prev. 2006;15(12):2502–7. 30. Liedtke S, et al. Postmenopausal sex hormones in relation to body fat distribution. Obesity (Silver Spring). 2012;20(5):1088–95. 31. Falk RT, et al. Relationship of serum estrogens and estrogen metabolites to postmenopausal breast cancer risk: a nested case‑control study. Breast Cancer Res. 2013;15(2):R34. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations. Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced rese archers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions