Introduction
Taller height and higher postmenopausal body mass index (BMI) are well-established risk factors for breast cancer [
1-
5]. While weight gain in adulthood increases the risk of postmenopausal breast cancer, higher weight in adolescence and early adulthood has been associated with decreased risk of pre- and postmenopausal breast cancer [
3,
6-
8]. The latter may be due in part to the fact that the differentiation of the mammary glands is promoted under higher estrogen levels in obese adolescents and young adults [
4,
9-
12].
Carriers of germline pathogenic variants in the
BRCA1 or
BRCA2 breast cancer predisposition genes have high breast cancer risks [
13]. However, the extent to which these risks are modified by anthropometric factors is unknown. Most previous studies have been retrospective, some of them with limited sample sizes, and their results have generally been inconclusive [
14-
21]. Three previous studies examined the association of height with breast cancer risk in pathogenic variant carriers [
17,
18,
20]. While two studies reported a positive association between height and breast cancer risk [
17,
20], one did not observe an association [
18]. These studies also analyzed body weight and breast cancer risk by menopausal status. Manders et al. found an association of higher current weight with postmenopausal breast cancer risk in a cohort of 299 carriers of pathogenic variants in
BRCA1 or
BRCA2 [
17]. They also found weak evidence of a positive association between weight gain and postmenopausal breast cancer [
17]. The retrospective study by Qian et al. reported an inverse association for higher BMI at age 18 years and risk of premenopausal breast cancer in a cohort of 14,676 carriers of pathogenic variants in
BRCA1 and 7912 in
BRCA2 [
20]. Using a Mendelian randomization approach, they additionally showed a similar inverse association between genetically determined BMI and premenopausal breast cancer risk, consistent with that seen in the general population [
3,
5-
8]. The only prospective study on BMI and breast cancer risk in carriers of a pathogenic variant in
BRCA1 or
BRCA2 found weak evidence for an association of higher BMI at age 18 years with lower postmenopausal breast cancer risk, but no association of current BMI or weight change and breast cancer risk [
18].
The identification of non-genetic risk factors for breast cancer in high-risk populations is important for developing more accurate risk prediction models and designing risk-adapted prevention strategies. In the present work, we evaluated the associations of height, BMI, and change in weight with pre- and postmenopausal breast cancer risk for carriers of a pathogenic variant in
BRCA1 or
BRCA2, using data from the International
BRCA1 and
BRCA2 Cohort Consortium [
22-
25].
Methods
Study sample
We used pooled data from three large prospective cohort studies of
BRCA1 and
BRCA2 pathogenic variant carriers: the International
BRCA1/2 Carrier Cohort Study (IBCCS, consisting of 19 national multi- and single-center prospective cohort studies) [
22], the Kathleen Cuningham Foundation Consortium for research into Familial Breast cancer (kConFab) [
24,
26], and the Breast Cancer Family Registry (BCFR) [
23,
25] (Additional file
1: Table S1). The study sample comprised women who were 18 to 80 years of age at recruitment and tested positive for a pathogenic germline variant in
BRCA1 or
BRCA2. Women with pathogenic variants in both genes were excluded.
Data collection
Study participants completed a baseline questionnaire and one or more follow-up questionnaires. The questionnaires asked about risk factors for breast cancer, including height, young-adult weight (age 18 or 20 years), weight at questionnaire completion, reproductive and medical history, surgical interventions, and menstrual history (age at menarche, age at last menstruation, whether the woman had had any periods in the past year, the number of years/months since the last menstruation, and reason(s) for periods stopping) [
27].
BMI was calculated as weight (kg) divided by height (m) squared. Weight change was calculated as the difference (in kg) between baseline weight and young-adult weight.
For women who indicated no periods in the past year, age at menopause was determined by adding one year to “age at last menstruation.” Women below the age of 60 years were considered premenopausal if they indicated that they had had a period in the past year, or if the “reason for periods stopping” was medication, oral contraceptive use, pregnancy, or breastfeeding. Women below the age of 60 years reporting risk-reducing salpingo-oophorectomy (RRSO) as the reason for menopause were considered premenopausal until RRSO. For women who reported a hysterectomy without bilateral oophorectomy, menopausal status was considered unknown. For women who were still menstruating when they started hormonal therapy and those who took hormonal contraceptives at older ages, age at menopause was classified as unknown.
Occurrence of breast cancer was derived from follow-up questionnaires and, for some studies, through linkage to cancer registries. Information on vital status was obtained from municipal or death registries, medical records, or family members. For details, see Kuchenbaecker et al. [
13]. All study participants provided written informed consent, and each study was approved by the relevant ethics committee at each participating institution.
Statistical analysis
Associations with breast cancer risk were evaluated using Cox proportional hazards regression with age as the timescale, estimating hazard ratios (HR) and corresponding 95% confidence intervals (CI). Exposure variates analyzed were height, young-adult BMI, baseline BMI, and weight change between early adulthood and baseline. Each anthropometric variable was analyzed in a separate model. All analyses were stratified by year of birth (< 1950, 1950–1959, 1960–1969, ≥ 1970) and study and were adjusted for age at menarche, number of full-term pregnancies, oral hormonal contraceptive use (ever/never), and hormone replacement therapy (ever/never), as reported in the baseline questionnaire. Robust variance estimation was used to account for familial clustering of study participants. Associations were examined separately for retrospective and prospective observation times, i.e., before and after baseline questionnaire, respectively, separately for BRCA1 and BRCA2 pathogenic variant carriers, and separately for pre- and postmenopausal women, resulting in eight sets of analyses.
Retrospective analysis
Associations of breast cancer risk with height and young-adult BMI were analyzed. For premenopausal women, the observation time started at birth and ended at diagnosis of the first primary breast cancer (invasive or in situ) as the event of interest, or was censored at diagnosis of any other type of cancer, risk-reducing mastectomy (RRM), completion of baseline questionnaire, or menopause, whichever came first. For postmenopausal women, observation started at menopause (if menopause occurred before baseline) and was censored at diagnosis of the first primary breast cancer (event), or at diagnosis of any other type of cancer, RRM, or baseline, whichever came first. Due to the non-random sampling of prevalent breast cancer cases, retrospective analyses were performed using the weighted retrospective cohort approach described by Antoniou et al. [
28]. Individuals were weighted such that the observed breast cancer incidence rates in the pre- and postmenopausal cohorts were consistent with established age-specific risk estimates for
BRCA1 and
BRCA2 variant carriers [
29].
Prospective analysis
Associations of breast cancer risk with height, young-adult BMI, baseline BMI, and weight change between early adulthood and baseline were analyzed. For premenopausal women, observation time started at baseline questionnaire completion (if women were premenopausal at that time point) and ended at diagnosis of the first primary breast cancer as the event of interest. Observation time was censored at diagnosis of any other type of cancer, RRM, last follow-up, or menopause, whichever came first. For postmenopausal women, observation started at baseline or menopause, whichever came last, and ended at diagnosis of the first primary breast cancer, or at diagnosis of any other type of cancer, RRM, or last follow-up, whichever came first.
If a woman’s prospective observation period had included both pre- and postmenopausal periods, then follow-up periods were assigned to the pre- and postmenopausal analyses, as appropriate.
All reported
P values are two-sided.
P values < 0.05 were considered statistically significant. Statistical analyses were conducted using R 4.1.1 for Windows (R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL:
https://www.R-project.org) and IBM SPSS Statistics for Windows Version 26 (IBM Corp, Armonk, NY)
.
Discussion
In this large cohort of women carrying a pathogenic variant in BRCA1 or BRCA2, we found associations of height, BMI, and weight gain with breast cancer risk in pre- and postmenopausal women, which were generally consistent in direction and magnitude with those described in the general population. This is the first study with a large prospective component to analyze the association anthropometric measures with breast cancer risk by menopausal status for BRCA1 and BRCA2 variant carriers separately.
Height as a risk factor for pre- and postmenopausal breast cancer has been described in prospective cohort studies, and in meta-analyses for the general population [
1,
2]. We found positive associations for
BRCA2 variant carriers between height and risks of pre- and postmenopausal breast cancer in the retrospective analyses, and with premenopausal breast cancer risk in the prospective analysis; no associations with height were observed for
BRCA1 variant carriers, although the 95% CI include a 20% increased risk per 10 cm for all analyses except the retrospective analysis of premenopausal women. Height as a breast cancer risk factor for
BRCA2, but not for
BRCA1 variant carriers, was also found in a retrospective study by Qian et al. [
20]. The difference in the association between
BRCA1 and
BRCA2 carriers may be explained by the difference in proportions of estrogen receptor (ER)-positive breast cancer (22% for
BRCA1, 77% for
BRCA2) [
27]. Zhang et al. showed, both in prospective studies and using Mendelian randomization, that height was a risk factor for hormone receptor positive breast cancer, and a weak or no association for receptor negative disease [
2].
We found that higher young-adult BMI is associated with lower premenopausal breast cancer risk in both
BRCA1 and
BRCA2 variant carriers in the retrospective analysis, with consistent, although not statistically significant, risk estimates in the prospective analysis. This is in line with the associations seen in the general population, Mendelian randomization studies, and earlier retrospective studies by Manders et al. [
17] and Qian et al. [
20] for
BRCA1 and
BRCA2 variant carriers, the data of which partially overlapped with the retrospective data of the present study. The reasons for the inverse association of high young-adult BMI with breast cancer risk are unclear, but might be mediated through early breast tissue differentiation [
4,
9]. In particular, childhood adiposity has been linked to a lower risk of benign breast disease and lower mammographic density which are risk factors for breast cancer [
10,
11,
30]. Understanding the biological mechanism underlying the inverse association of higher young-adult BMI and lower premenopausal breast cancer risk could potentially identify modifiable pathways and provide new insights for prevention in the future.
Higher young-adult BMI is also associated with a lower risk of postmenopausal breast cancer in the general population [
3,
5]. We found no association of higher young-adult BMI with postmenopausal breast cancer risk, which is also consistent with the findings by Qian et al. [
20]. However, the retrospective cohort for postmenopausal women was much smaller than the premenopausal cohort and the HR estimates did not differ significantly.
In the prospective analysis of
BRCA1 pathogenic variant carriers, we found that higher baseline BMI and adult weight gain were associated with higher risk of postmenopausal breast cancer, but not with risk of premenopausal breast cancer. This is in line with the finding by Manders et al. for
BRCA1/2 variant carriers and with the findings in the general population [
17,
31,
32]. However, for
BRCA2 variant carriers, we found no evidence of association, although a HR of similar magnitude compared to
BRCA1 variant carriers cannot be excluded, because of the wide confidence intervals.
A recent study has suggested that weight may be more predictive of postmenopausal breast cancer risk than BMI [
33]. We analyzed the variables baseline weight and young-adult weight separately and found no large differences in association with breast cancer compared to baseline BMI and young-adult BMI (Additional file
4: Table S4).
Two biological explanations for our findings of high baseline BMI and of weight change and higher breast cancer risk in the postmenopausal cohort of carriers of
BRCA1 can be considered. A similar pattern (with an association for
BRCA1- but not
BRCA2-associated breast cancer) has been observed for oral contraceptive use [
34]. Both findings were unexpected, since hormonal exposure is known to especially promote the ER-positive breast cancer subtype. However, in the general population, a weak association between elevated BMI and hormone receptor negative postmenopausal breast cancer and an association of oral contraception use with early onset triple negative breast cancer was reported [
35-
38]. One explanation for hormone-induced triple negative breast cancer might be the activation of the paracrine pathway via the receptor activator of nuclear factor kB (RANK) [
39-
46]. The pathway seems to be more relevant for carriers of variants in
BRCA1 compared to
BRCA2 or
PALB2 [
47].
Early surgical menopause may lead to use of hormone therapy, usually a combination of estrogen and progestin. If BRCA1-deficient breast cells are particularly sensitive to hormonal treatment, this could lead to a significantly higher risk of breast cancer compared with BRCA2 variant carriers associated with very early menopause and hormonal contraception use. In our study, however, the comparison of analysis with and without adjustment for use of hormonal contraception and postmenopausal hormone therapy did not support this hypothesis, since an association of high baseline BMI and weight gain with postmenopausal breast cancer risk was consistently seen in carriers of a pathogenic variant in BRCA1, but not in BRCA2.
Another mechanism of action promoting
BRCA1-associated postmenopausal breast cancer may be the lack of suppression of aromatase.
BRCA1 is known to be an inhibitor of aromatase (CYP19) transcription and also interferes with transcriptional activation mediated by the estrogen receptor alpha (ER
α) [
47]. High BMI and weight gain might therefore lead to higher estradiol levels due to increased aromatase activity and consecutively to postmenopausal
BRCA1-associated breast cancer. Interestingly, after menopause the percentage of ER-positive
BRCA1-associated breast cancer increases [
48].
Avoiding over- and underweight is important for both pathogenic variant carriers and non-carriers. Even if the relative risks for anthropometric measures and breast cancer in
BRCA1 and
BRCA2 variant carriers are similar to those found in the general population, the absolute excess risks will be higher for
BRCA1 and
BRCA2 variant carriers because of the higher background breast cancer risk. Additionally, as shown by Hopper et al., the postmenopausal elevation of breast cancer risk by far outweighs the premenopausal protection associated with high BMI [
49].
Variant carriers may be highly motivated to attempt to modify their breast cancer risk through a healthy lifestyle [
50,
51]. Healthy behavior change might include better nutrition and weight loss, although a persistent change in lifestyle is hard to achieve, especially after menopause [
51]. Our retrospective results of low BMI and higher premenopausal breast cancer risk imply that underweight and unhealthy weight loss should be discouraged in premenopausal women. But further research on a larger dataset is necessary. The first intervention study on nutrition and physical activity for
BRCA1/2-pathogenic variant carriers (the LIBRE trial in Germany) is currently recruiting [
52]. It aims to show that physical fitness and maintaining a healthy body weight are key to good quality of life and—in the long run—prevent breast cancer in both affected and unaffected carriers.
Anthropometric measures are included in cancer risk prediction models such as BOADICEA [
53]. In BOADICEA, the relative risks for height and BMI among variant carriers are assumed to be the same as those for the general population. The present analyses suggest that, for height, this is a reasonable assumption for
BRCA2, but some adjustment may be needed for
BRCA1. The association effect sizes for BMI are generally consistent with the general population estimates, at least for young-adult BMI [
53].
The study has some limitations. Despite the large size of the cohort, the number of incident cases was still limited and the confidence limits were correspondingly wide, especially in subgroup analyses, such as postmenopausal high young-adult BMI. Thus, the results are still heavily dependent on the retrospective data, which are potentially subject to more bias. Much larger prospective cohorts with a broader age range and a longer follow-up are therefore needed.
A general challenge is the definition of menopausal status. Ideally, this would be defined by taking a detailed menstrual history and, in many cases, a blood test—but this is not feasible in large epidemiological studies, which rely on self-reporting. This may have led to some misclassification of postmenopausal women as premenopausal women, diluting the difference between the two groups.
A known limitation of retrospective studies is the potential for survival bias introduced by including prevalent cancer cases. There is evidence in the general population that higher BMI is associated with advanced tumor stage and a worse prognosis [
54]. If the same is true for variant carriers, the inverse association with high BMI might be overestimated [
55]. Some studies have addressed this by defining a pseudoincident cohort, which includes breast cancer cases only if they occurred within the last 5 years to avoid survival bias [
34]. We did not utilize this approach, which would have substantially reduced the sample size. However, the retrospective and prospective findings were broadly consistent, though some associations (e.g., the inverse association with young-adult BMI in premenopausal women) were not statistically significant in the prospective analysis. Again, larger prospective analyses are needed.
This is the first prospective study to analyze the association of breast cancer risk with anthropometric measures by menopausal status, separately for BRCA1 and BRCA2 variant carriers. Height was a risk factor for premenopausal breast cancer in BRCA2 variant carriers, but not BRCA1 carriers. High young-adult BMI was associated with decreased risk of premenopausal breast cancer, but not for postmenopausal breast cancer. Higher BMI and weight gain in adult life were risk factors for postmenopausal breast cancer in BRCA1 variant carriers.
In conclusion, our findings indicate that the associations of height, BMI and weight gain with breast cancer risk in BRCA1 and BRCA2 variant carriers are broadly similar to those reported in the general population when taking into account menopausal status. This research is not only important to inform carriers about age-specific cancer risks, it might also open up new risk reduction strategies by elucidating yet unknown signaling pathways.
Acknowledgements
The Hereditary Breast and Ovarian Cancer Research Group Netherlands (HEBON) consists of the following Collaborating Centers: Netherlands Cancer Institute (coordinating center), Amsterdam, NL: M.K. Schmidt, F.B.L. Hogervorst, F.E. van Leeuwen, M.A. Adank, D.J. Stommel-Jenner, R. de Groot, L. Hordijk; Erasmus Medical Center, Rotterdam, NL: J.M. Collée, I. Geurts-Giele, M.J. Hooning, I.A. Boere; Leiden University Medical Center, NL: C.J. van Asperen, P. Devilee, R.B. van der Luijt, T.C.T.E.F. van Cronenburg; Radboud University Medical Center Nijmegen, NL: M.R. Wevers, A.R. Mensenkamp; University Medical Center Utrecht, NL: M.G.E.M. Ausems, M.J. Koudijs; Amsterdam UMC, Univ of Amsterdam, NL: K.Y. van Spaendonck; Amsterdam UMC, Vrije Universiteit Amsterdam, NL: K. van Engelen, J.J.P. Gille; Maastricht University Medical Center, NL: E.B. Gómez García, M.J. Blok, M. de Boer; University of Groningen, NL: L.P.V. Berger, A.H. van der Hout, M.J.E. Mourits, G.H. de Bock; The Netherlands Comprehensive Cancer Organisation (IKNL): S. Siesling, J. Verloop; The nationwide network and registry of histo- and cytopathology in The Netherlands (PALGA): Q.J.M. Voorham. HEBON thanks the study participants and the registration teams of IKNL and PALGA for part of the data collection. We acknowledge the GENEPSO Centers: the Coordinating Center: Institut Paoli-Calmettes, Marseille: Catherine Noguès, Emmanuel Breysse, Pauline Pontois, Lilian Laborde, and the Collaborating Centers: Institut Curie, Paris: Dominique Stoppa-Lyonnet, Marion Gauthier-Villars; Bruno Buecher, Institut Gustave Roussy, Villejuif: Olivier Caron; Hôpital René Huguenin/Institut Curie, Saint Cloud: Emmanuelle Fourme-Mouret; Centre Paul Strauss, Strasbourg: Jean-Pierre Fricker; Centre Leon Berard, Lyon: Christine Lasset, Valerie Bonadona; Hôtel Dieu -Centre Hospitalier, Chambery: Sandra Fert-Ferrer; Centre François Baclesse, Caen: Pascaline Berthet; CHRU Dupuytren, Limoges: Laurence Venat-Bouvet; CHU La Milétrie, Poitiers: Brigitte Gilbert-Dussardier; Hôpital d’Enfants CHU Dijon –Centre Georges François Leclerc, Dijon: Laurence Faivre; Centre Alexis Vautrin, Vandoeuvre-les-Nancy: Elisabeth Luporsi; Institut Claudius Regaud, Toulouse: Laurence Gladieff; Réseau Oncogénétique Poitou Charente, Niort: Paul Gesta; Institut Paoli-Calmettes, Marseille: Catherine Nogès, Hagay Sobol, François Eisinger; Institut Bergonié, Bordeaux: Michel Longy, Centre Eugène Marquis, Rennes: Catherine Dugast†; GH Pitié Salpetrière, Paris: Chrystelle Colas, Florent Soubrier; CHU Arnaud de Villeneuve, Montpellier: Isabelle Coupier, Pascal Pujol, Carole Corsini; Centres Paul Papin, and Catherine de Sienne, Angers, Nantes: Alain Lortholary; Centre Oscar Lambret, Lille: Philippe Vennin†, Claude Adenis; Institut Jean Godinot, Reims: Tan Dat Nguyen; Institut Jean-Godinot and ICC Courlancy, Reims: Clotilde Penet; Centre René Gauducheau, Nantes: Capucine Delnatte; Centre Henri Becquerel, Rouen: Julie Tinat, Isabelle Tennevet; Hôpital Civil, Strasbourg: Jean-Marc Limacher; Christine Maugard; Polyclinique Courlancy, Reims: Liliane Demange†; Clinique Sainte Catherine, Avignon: Hélène Dreyfus; Hôpital Saint-Louis, Paris: Odile Cohen-Haguenauer; Couple-Enfant-CHU de Grenoble: Dominique Leroux; Hôpital de la Timone, Marseille: Hélène Zattara-Cannoni; CHU Fort de France, Fort de France: Odile Bera. In grateful memory of Hakan Olsson, MD PhD, Professor, Oncology, Clinical Sciences in Lund, University Hospital, Lund, Sweden. We thank kConFab participants, kConFab investigators and all the kConFab coordinators, research nurses, interviewers, and data management staff, and the heads, and staff of the participating Family Cancer Clinics. We thank GC-HBOC participants, GC-HBOC investigators and staff of the participating GC-HBOC Centers of the University Hospital of Cologne and University Hospital of Dresden, Germany. Interdisciplinary Health Research International Team Breast Cancer Susceptibility (INHERIT BRCAs): We would like to thank Dr Martine Dumont and Martine Tranchant (Cancer Genomics Laboratory, CRCHUQ) for sample management.