Introduction
Breast cancer continues to be an important health problem and is one of the most common causes of cancer deaths in women worldwide, with an estimated 358,967 new cases and 90,665 breast cancer-related deaths in the European Union annually [
1].
Although genetic profiling, age of menarche and menopause, parity, age of the first child, previous occurrence of cancer, and breast density are all well-known risk factors for breast cancer, lifestyle is considered an increasingly important, modifiable contributing factor to breast cancer etiology [
2].
Obesity, defined as a body mass index (BMI) of ≥ 30 kg/m
2, affects over 600 million adults worldwide, and the World Health Organization estimates that 40% of adult women are overweight (BMI of ≥ 25 kg/m
2), with the prevalence tripling between 1975 and 2016 [
3]. The prevalence of obesity varies widely by country, with low rates in countries such as Vietnam (2.1%) or Japan (4.4%) compared with 37.3% in the United States and the highest rates in Oceania (Nauru 61%, Cook Islands 55%) [
1]. Except for some regions in sub-Saharan Africa and Asia, more people are now obese than underweight [
1].
Several studies [
4‐
6] have shown a significantly strong association between increased BMI and higher breast cancer incidence and specific mortality in postmenopausal women. However, in premenopausal women, high BMI is associated with a reduced risk of breast cancer [
7].
The precise mechanisms whereby obesity plays a protective role against breast cancer in premenopausal women, but represents a risk factor after menopause, remain elusive [
2].
Furthermore, two meta-analyses described that, in premenopausal women, obesity is associated with high-risk estrogen receptor (ER)-negative and triple-negative breast cancer but, in postmenopausal women, obesity seems to be a risk for hormone receptor-positive breast cancer [
8,
9]. However, another meta-analysis that studied the association between obesity, hormone receptor, and menopausal status, reported an increased hazard ratio for overall survival in heavier versus lighter women independently of hormone receptor or menopausal status [
10]. Consequently, more studies are currently needed to elucidate the role of obesity in different breast cancer subtypes.
Mammographic breast density (MBD) is based on the proportion of stromal, epithelial, and adipose tissue in the breast. MBD is also an independent risk factor for the development of breast cancer, with a higher risk in women with high density. A systematic review and meta-analysis of 42 studies found that the relative risk of incidental breast cancer is 2.92 for women with heterogeneously dense breasts (type C) and 4.64 for women with extremely dense breasts (type D), compared to women with almost entirely fatty breasts (types A and B) [
11].
MBD is influenced by factors such as age and BMI (MBD decreases with increasing age and BMI), and increases with hormone replacement therapy [
12] Therefore, there is a possible paradox in the relationship between breast cancer risk and fat tissue depending on its localization (high risk for body fatness but not for breast adipose tissue) [
13].
Fat tissue has been described as a microenvironment promoting carcinogenesis through different mechanisms, in particular, chronic inflammation [
14], but it also has a potentially protective role, especially as a source of vitamin D [
13,
15]. Furthermore, recent studies have shown that increased levels of leptin and decreased adiponectin secretion are directly associated with breast cancer development [
2].
Dyslipidemia is strongly associated with obesity and has been independently linked with breast cancer risk and survival [
16], but data are conflicting. The ACALM study demonstrated that women aged above 40 years with high cholesterol were 45% less likely to develop breast cancer than women with normal cholesterol levels [
17]. Moreover, some studies observed that low HDL-cholesterol was associated with higher estrogen levels and absolute mammographic density (both independent risk factors for breast cancer) [
18]; and intratumor cholesteryl ester accumulation was associated with more aggressive tumors, including grade 3, HER2-positive, and triple-negative breast cancers [
19]. However, based on the results of recent studies, 27-OH-cholesterol is potentially a better biomarker than total cholesterol [
2].
Vitamin D is known for its anti-cancer properties, including induction of apoptosis and inhibition of angiogenesis and metastasis [
20]. Low vitamin D levels were shown to be associated with increased overall and disease-specific breast cancer mortality [
20,
21]. Furthermore, vitamin D deficiency increased the risk of recurrence of luminal breast cancer, but this relationship was not found in patients with HER2-positive or triple-negative cancer subtypes [
22].
Hyperinsulinemia is an independent risk factor for poor breast cancer prognosis and is associated with low adiponectin levels and shorter breast cancer survival [
23]. Moreover, elevated HOMA-IR scores and low adiponectin levels are both associated with obesity and increased breast cancer mortality. However, in premenopausal women, high circulating insulin levels may protect against breast cancer, the same as obesity [
24].
The objectives of the current study were 1) to analyze the association between BMI and MBD with breast cancer molecular subtypes and 2) to study the possible differences between cholesterol, vitamin D, and insulin levels in recently diagnosed early breast cancer.
Methods
The study included women with a recent diagnosis of early breast cancer during a 1-year period at three Spanish breast cancer units (MD Anderson Cancer Center Madrid, Segovia Hospital, and San Pedro Hospital of Logroño).
Oncologists at the breast cancer units completed a questionnaire at diagnosis of all included women about lifestyle (e.g., diet, exercise, smoking habit). Clinical characteristics (hypertension, diabetes, menopausal status, breast density, weight, height, and abdominal size) and tumor characteristics (TNM, estrogen receptors [ER], progesterone receptors [PR], Ki67, and HER2) were recorded. Blood tests for total cholesterol, LDL-cholesterol, HDL-cholesterol, triglycerides, insulin, and vitamin-D (25-OH vitamin D) were conducted at diagnosis.
Exercise was recorded as a ‘Yes/No’ response regarding whether the subject completed more than 150 min per week of moderate exercise (OMS recommendations). Diet was studied by collecting information about fruit and vegetable consumption, weekly alcohol consumption, olive oil used, and processed foods intake. Breast density information was obtained from the breast radiology report. Tumors were classified, according to the 13th St Gallen International Breast Cancer Panel, into luminal-A like (ER/PR positive, Ki-67 < 20%), luminal-B like (ER/PR positive, Ki-67 ≥ 20%), HER2-positive (ER and PR positive/negative, HER2-positive) and triple-negative (ER-, PR-, and HER2-negative).
The study received the approval of the hospital MD Anderson Cancer Center, all data of patients were coded and do not suppose any risk to the integrity of the patients.
All statistical analyses were performed with R software, version 4.1.1. Quantitative variables were described as median [IQR] and qualitative variables as absolute (n) and relative (%) frequencies. Chi-square or Fisher test were used to evaluate significant differences between qualitative data. A non-parametric Kruskal–Wallis rank sum test was used to evaluate differences in quantitative variables (total cholesterol, LDL-cholesterol, HDL-cholesterol, triglycerides, insulin, and vitamin-D) within MBD or molecular subtype subpopulations; postmenopausal and premenopausal differences in each MBD or molecular subtype subpopulation were evaluated by non-parametric Wilcoxon rank sum test. Correlations between BMI and blood test variables were determined by Spearman rank correlation coefficient (ρ). Differences with a p-value ≤ 0.05 were considered to be statistically significant.
Discussion
In the current study of women with early breast cancer, data showed that almost 50% of patients were overweight (32%) or obese (16%), which aligns with data for the general female population in Spain (30.6% overweight, 15.5% obesity) [
25], but is lower than in other countries (USA, 37.3% obesity) [
1]. Our findings are in line with previous reports on the association between increased BMI and higher breast cancer incidence in postmenopausal, but not in premenopausal, women [
4‐
7]: 68% of obese patients were postmenopausal and 61.9% of patients with normal weight were premenopausal.
Only 7 patients (5.1%) in our study population were T3–T4, so we were unable to evaluate any association between tumor size and BMI. Engin et al. described more aggressive breast cancers with large tumor size, high-histological grade, and estrogen receptor-negative in patients with low adiponectin levels [
26].
In the current study, 49.7% of patients reported doing exercise and 21% reported smoking; these data are similar to general population data for women in Spain with 54.8% of individuals reporting being sedentary and 20% being smokers [
25].
We did not find any association between BMI and molecular subtype but, according to menopausal status, we observed higher BMI in postmenopausal luminal A and HER2-positive patients, as published in other studies that showed obesity as a risk factor for hormone receptor-positive breast cancer in postmenopausal women [
2]. However, in contrast to previous reports [
8,
9], we did not find a higher incidence of triple-negative breast cancer in obese/overweight premenopausal patients.
High MBD is an independent risk factor for breast cancer [
11], and 76.2% of our patients had MBD type C or D (55.1% and 21.1%, respectively). Our study findings align with previous reports that MBD and BMI are independent risks factors of breast cancer [
27] because, in our population, we observed decreased MBD with increasing BMI; this difference was more notable in premenopausal women, with high BMI in premenopausal patients with low MBD, and low BMI in premenopausal patients with high MBD. Furthermore, unlike other authors that described a higher risk of ER-negative breast cancer in premenopausal women with high MBD and high BMI [
28], we found no association between MBD and molecular subtypes.
Half of our patients had hypercholesterolemia at diagnosis (52.1%), similar to data for the general Spanish population (50.5%) [
29] and, unlike other studies [
18,
19], we did not find any association between total cholesterol and MBD or more aggressive subtypes (triple-negative or HER2-positive); on the contrary, we found higher levels of cholesterol in postmenopausal patients with luminal A subtype. These findings suggest that total cholesterol may not be the most appropriate biomarker and future studies may need to assess 27-OH-cholesterol.
We were unable to study the potentially protective effect of vitamin D because we only collected data about vitamin D levels at diagnosis; 39.8% of patients had low vitamin D and there was no association with BMI. Ismail et al. described vitamin D deficiency in 30% of Egyptian females with breast cancer and an association with the HER2-positive subtype and worse prognosis [
30].
In line with previous literature [
24], we found that hyperinsulinemia was associated with obesity, with no differences according to menopausal status.
Conflict of interest
Dr. Isabel Calvo received speaker honoraria from Gilead, Roche, and MSD and financial support for educational activities from Roche and Daiichi-Sankyo. Dr. M González-Rodríguez received speaker honoraria from Gilead and financial support for educational activities from Roche and Gilead. Dr. F. Neria declares no conflict of interest. Dr. I. Gallegos received consultancy honoraria from Roche, Bristol, and MSD; advisory board honoraria from Roche, Pfizer, Novartis, Daiichi-Astra, Janssen, Sanofi, Eisai, and Lilly and support for educational activities from Roche, Pfizer, and Astellas. Dr. L. García-Sánchez declares no conflict of interest. Dr. R. Sánchez-Gómez declares no conflict of interest. Dr. S. Pérez received speaker honoraria from GE Healthcare, BD, Motiva and financial support for educational activities from GE Healthcare, BD, and Motiva. Dr MF Arenas declares no conflict of interest. Dr L.G. Estévez received consultant/Advisor/speaker honoraria from AstraZeneca, Daiichi Sankyo, Gilead Sciences, MSD, Lilly, and Roche; and from research (paid to Institution) by Roche.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.