In this study, we investigated the association between BMI and the incidence of breast cancer in approximately 800,000 women aged < 45 years using data from a Japanese nationwide database. A high BMI was associated with a significantly lower incidence of breast cancer compared to normal BMI. Whereas the incidence of breast cancer with hormonal drug administration also decreased in participants with a high BMI, the incidence of breast cancer with trastuzumab administration increased in the obese group. Our findings demonstrated that the East Asian population exhibited a similar association between BMI and the incidence of breast cancer in premenopausal women to that has been observed in the Western population [1–7].
Although several studies, including meta-analyses, have reported that obesity could have been a risk factor for premenopausal breast cancer in Asian women [3, 6, 7], our study showed that obesity was associated with a reduced risk for premenopausal breast cancer (except for HER-2 positive breast cancer) in the Japanese population, similar to that in the Western population. We consider that the analyses of previous studies were suboptimal due to their study design. The meta-analyses included many Asian case-control studies with a limited number of participants, especially few obese populations [6, 7]. Additionally, in a previous study showing that obesity was a risk factor for premenopausal breast cancer in large Japanese cohorts, 51 years of age was defined as the menopausal threshold during the follow-up period because 50% of women had been reported to become postmenopausal at this age [3]. As approximately 40% of Japanese women reach menopause between 45 and 51 years of age [21], the results of that study may have been influenced by postmenopausal breast cancer, where obesity is a risk factor [1–6]. Since the menopausal threshold in our study was defined as 45 years, we consider that our results would indicate the risk of breast cancer in premenopausal women more accurately than the previous study.
The age-specific incidence of breast cancer in East Asian countries has been reported to be different from that in Western countries. Women in Japan, Taiwan, and Korea demonstrated a peak incidence of breast cancer in their 40s, whereas Western women showed a peak incidence in their 70s [31–34]. The reason for the difference has been unresolved. We consider that this can be explained by the difference in BMI distribution, as women in East Asian countries have considerably lower BMI than those in other countries [35]. Since a high BMI generally reduces the risk of premenopausal breast cancer (similar to our study findings) and increases the risk of postmenopausal breast cancer [1–6], many women with a low BMI in these countries would be at a high risk for developing breast cancer in their 40s (i.e., the end of premenopausal status) and at a relatively low risk after menopause. Thus, the age to be focused on during breast cancer screening may depend on the BMI distribution of the population.
Our study showed an association between BMI and the incidence of breast cancer, categorized by subtypes, as BMI and body composition are considered to differ by subtypes [36]. A recent study of 758,592 premenopausal women in the US found that HR-positive breast cancer was associated with BMI in women aged between 18 and 54 years despite no significant association for HER2-positive breast cancer [2]. Our study’s results correspond with that of this previous study. Together, these would indicate that low BMI is a risk factor, especially for HR-positive breast cancer, regardless of ethnicity. Indeed, premenopausal women in East Asia, who were likely to have a low BMI, showed a higher incidence of HR-positive breast cancer and equivalent incidence of HER2-positive breast cancer compared to those in the US [32]. This association implies BMI-related differences in sex-hormone profiles, which may affect the pathogenesis of breast cancer [2]. For example, serum levels of several sex hormones such as estradiol, progesterone, testosterone, and sex hormone-binding globulin may differ by BMI [37–43]. Furthermore, rich subcutaneous fat in obese women where approximately 5% of total estradiol is synthesized can lead to irregular menstrual cycles through negative feedback in the hypothalamic-pituitary axis [44–46]. However, these have not been proven to have a direct effect on the development of premenopausal breast cancer [2, 47]. We believe that elucidating the paradoxical association between premenopausal adiposity and breast cancer development might result in identifying modifiable factors in premenopausal women.
Several limitations regarding this study should be acknowledged. First, we were unable to investigate several factors that are known to be risks for breast cancer; for example, marriage status [48], number of deliveries [11, 49], and breastfeeding [1, 50]. However, because these factors are not confounding factors but can be intermediate factors between BMI and breast cancer development, they should not be adjusted for to investigate the effect of BMI on breast cancer development [51]. We were also unable to assess the family history of breast cancer [52–54] and BRCA mutation [1, 11, 53]. Since these factors have not been reported to be associated with BMI, they may not be confounding factors necessarily to be adjusted for. Additionally, we were unable to obtain information on drug use such as hormone replacement therapy and oral contraceptives [54, 55]. There were very few participants in the current database who received these hormonal drugs before the first health checkup, presumably because almost all participants were enrolled in the company insurance immediately preceding the checkup. Furthermore, a case-control study and a cohort study in Japanese women showed no increase in the incidence of breast cancer in those who received hormone replacement therapy [49, 56]. Second, information on menstruation such as the timing of menarche and menopause was also unavailable. Instead of the timing of menopause, we used the age 45 years as a threshold based on previous studies [21–23, 57]. Even though the current study population included some postmenopausal women, for whom obesity is a risk factor for breast cancer, obesity in women < 45 years was significantly associated with decreased breast cancer incidence. Third, the follow-up period was shorter than that of previous prospective cohort studies [2, 3, 12]. The current database collected data from Japanese companies, and some participants could have exited the database soon due to job change or retirement. Whereas the short follow-up period could cause the risk of observing reverse causality, skewing the results of the analysis regarding the association between cancer and obesity [58]. Therefore, we excluded participants with a short follow-up period to reduce this risk, similar to previous large-scale retrospective studies [59, 60]. Finally, there were some missing data in some background factors; i.e., smoking history, alcohol consumption, and physical activity. Hence, we performed several multivariable analyses with and without adjustment for these factors and showed that the results did not differ by the adjusted factors.
In conclusion, the RCS analysis of this study, which used a Japanese nationwide database, and included a general population of 785,703 individuals, demonstrated that a high BMI was associated with a low incidence of breast cancer in women aged < 45 years. The association was also observed in breast cancer with hormonal drug administration, whereas the occurrence of breast cancer with trastuzumab administration did not significantly differ by BMI. Our findings suggest that BMI is inversely associated with the incidence of breast cancer in East Asian premenopausal women, similar to that observed in Western women. Additionally, the results would contribute to explaining the incidence of breast cancer peaks in women in their 40s in East Asia.