Breast cancer is the most common cancer in women in the United States, but its incidence varies substantially by race/ethnicity.1 Age-adjusted incidence of breast cancer is highest among non-Hispanic white (NHW) women followed by African American and Hispanic women.1-3 Among women aged <40 years, breast cancer incidence is highest in African Americans; rates among young Hispanic women are lower than those of both African Americans and NHWs.4 Between 2001 and 2004, breast cancer incidence fell approximately 14% in NHWs, but the decrease was much smaller in Hispanics (−3%), and there was no change in the rate among African Americans.5, 6 The reasons for these disparities in incidence probably are multifactorial and are not well understood, and more etiologic studies of breast cancer in minorities are needed to clarify the complex relations between race, socioeconomic factors, biology, and breast cancer.
Well established risk factors for breast cancer have been identified in studies conducted primarily among NHW women. Family history,7, 8 reproductive and menstrual factors,8-10 and breast health history8, 11 all are associated with the risk of developing breast cancer. Suspected risk factors for breast cancer include mammographic breast density,12 plasma estrogen levels,13, 14 weight gain during adulthood,15 and body mass index (BMI).15-17 A recent review of risk factors for breast cancer identified BMI and weight gain between ages 20 years and 50 years as the most important variables in predicting breast cancer behind Gail model risk factors,8 breast density, and free estradiol concentrations.18 Because obesity is an epidemic in the United States, particularly among African American and Hispanic women,19 it is essential to determine its association with breast cancer in these under-studied populations.
Breast Cancer in Hispanics
Relatively few epidemiologic studies of breast cancer have been conducted in Hispanic women. Although the age-adjusted risk of breast cancer is approximately 30% lower in Hispanics compared with NHWs (hazard ratio [HR], 0.70; 95% confidence interval [CI], 0.57-0.85), the risk estimate is attenuated to nearly 1.0 after adjustment for socioeconomic factors and screening history.20 Family history21 and older age at first live birth22 reportedly increase the risk of breast cancer in Hispanic women, and early age at menarche may decrease the risk,21 a result that is contradictory to studies of breast cancer in NHW women. Higher parity, which is a protective factor in NHW women, is not associated with breast cancer in Hispanic women.22 These contradictory results in studies of known risk factors suggest that the etiology of breast cancer may differ between Hispanics and NHWs.
Hispanic women are diagnosed with breast cancer at a more advanced stage23-25 and present with higher grade, larger tumors than NHWs after adjusting for age at diagnosis.26 Breast tumors in Hispanic women are more commonly estrogen receptor (ER)-negative compared with those in NHW women.20, 23 These poor prognostic factors lead to significantly lower 5-year survival rates in Hispanics compared with NHWs,23, 27 and recent statistics indicate that Hispanic women are 22% more likely to die of their breast cancer than NHW women.1
Breast Cancer in African Americans
African American women are significantly less likely to develop breast cancer than NHW women.20 Even after adjustment for Gail model risk factors, socioeconomic factors, smoking status, and mammography screening, African American women remain 25% less likely to develop the disease than NHW women.20 However, African American women are diagnosed with this disease at a younger age than NHW women,28 and the rate of breast cancer in premenopausal African Americans significantly exceeds that of NHWs.29, 30 The reasons for this crossover in incidence rates in premenopausal African American women remain largely unknown. Studies of risk factors identified in NHW women have reported a similar risk profile among African American women.31-33
African American women present with larger, more advanced, higher grade breast tumors and are diagnosed with triple-negative breast cancer more frequently than NHW women.34, 35 African American women with late-stage, triple-negative disease have the poorest survival rates of all racial/ethnic groups.36, 37
Although deaths from breast cancer have declined steadily over the last 20 years overall, African Americans are significantly less likely to survive than NHW women regardless of their disease stage at diagnosis.38 This difference in survival is most pronounced for women who are diagnosed at ages <40 years, in which African Americans are more than twice as likely to die as NHWs with the same stage of breast cancer.38 Recent calculations estimate that the age-adjusted mortality rate is nearly 70% greater in African American women than in NHW women.39 At each successive year of diagnosis, African Americans are more likely to die from their disease.40 This indicates that African American women are at higher risk of death when diagnosed both at young ages38 and at later ages40 in all stages of disease at diagnosis.38, 40
Because most studies are conducted primarily in NHWs, the etiology of breast cancer in minorities is not well understood. The purposes of this review were 1) to identify and critique the epidemiologic literature on the association of obesity, weight, and weight gain and breast cancer risk in Hispanic and African American women, with a brief summary of some of the literature in NHW women; 2) to identify gaps in the literature, particularly gaps in the study of Hispanic and African American women; and 3) to review the biologic mechanisms in which obesity may be associated with breast cancer.
We reviewed the literature on the association between obesity, weight, and weight gain and breast cancer. Both PubMed and Medline were used to identify all publications on these topics from January 1970 through July 2009. Medline was searched using a combination of the Medical Subject Heading (MeSH) terms “obesity,” “body mass index,” “weight,” “weight gain,” “Hispanic,” and “African American.” References listed in articles of interest also were cross-checked for appropriateness for inclusion in this review article. Studies in NHWs were included only if risk estimates and 95% CIs were given so that the results could be displayed graphically in the accompanying figures. Only publications in English were considered.
Studies of the effect of body size on the risk of breast cancer vary in the method of assessing anthropometrics. Common anthropometric measurements used in these epidemiologic studies include BMI, weight, and weight gain throughout adulthood. Studies of body size and breast cancer have involved primarily NHW women, and those are summarized below. The limited studies in Hispanics and African Americans also are discussed below.
Obesity and BMI in NHW Women
BMI, which is measured as weight in kilograms divided by height in meters squared, is categorized as underweight (BMI <18.5 kg/m2), normal weight (BMI 18.5-24.9 kg/m2), overweight (BMI 25-29.9 kg/m2), and obese (BMI ≥30.0 kg/m2).41, 42 Obesity is an epidemic in the United States, because approximately 33% of all adults aged >20 years are considered obese,19 and it has been well studied as a risk factor for several cancers, including breast cancer.43
Most16, 44-49 but not all50, 51 case-control studies have reported an inverse association between obesity and premenopausal breast cancer. An early meta-analysis of case-control studies indicated that there was a 12% decrease in premenopausal risk for every 8 kg/m2 increase in BMI; this is equivalent to the difference between thin (ie, a BMI of 22 kg/m2) and obese (ie, a BMI of 30 kg/m2).52
Cohort studies in premenopausal women have been less consistent, but most report an inverse association between obesity and breast cancer risk.17, 53-57 A meta-analysis of cohort studies concluded that the risk of premenopausal breast cancer is reduced significantly by approximately 30% for each 8 kg/m2 increase in BMI.52 The Pooling Project in Diet and Cancer analyzed data from 7 prospective cohort studies and reported that premenopausal women with BMI >31 were 46% less likely to develop breast cancer than those with BMI <21 (HR, 0.54; 95% CI, 0.34-0.85).53 Figure 1 summarizes studies of the association between obesity and premenopausal breast cancer that were conducted mostly in NHWs.
Many epidemiologic studies of obesity and postmenopausal breast cancer have been conducted, and most46, 49, 51, 58-62 but not all44, 47, 50, 63 case-control studies have concluded that obesity is associated with an increased risk. Although many cohort studies have reported a positive association between BMI and postmenopausal breast cancer risk,53, 64-70 several have not.17, 57, 71-75 Two pooled analyses reported a positive association between BMI and postmenopausal breast cancer.53, 64 The risk of obesity-related postmenopausal breast cancer appears to be increased only in those who have not received hormone replacement therapy (HRT).15, 17, 68 In non-HRT users, obesity increased the risk of postmenopausal breast cancer by 2-fold compared with normal weight women,68, 76 whereas there was a null effect in ever-users.15, 17, 68, 71, 76 Figure 2 summarizes the results from these studies, which were conducted mostly in postmenopausal NHW women.
Body Weight in NHW Women
Nearly all studies of body weight and breast cancer risk were conducted predominantly in NHW study populations, so the effects of weight and weight gain remain largely unknown in Hispanic and African American women. Among premenopausal women, there is no association between weight at diagnosis and breast cancer.16, 50, 60, 72, 77-79 Figure 3 summarizes the results from these studies.
Studies of the effect of body weight on breast cancer risk among postmenopausal women have reported mostly null results.16, 72, 75, 77, 78 There is some evidence that the effect of weight on postmenopausal breast cancer may be modified by HRT use, because a significant association was reported in non-HRT users only.76, 79 Studies of the association between weight and postmenopausal breast cancer risk are summarized in Figure 4.
Weight Gain in NHW Women
It has been reported consistently that weight gain from puberty through adulthood increases the risk of postmenopausal breast cancer regardless of study design,17, 68, 71, 74, 75, 80-86 and only 1 cohort study demonstrated a protective effect of weight gain.87 This effect is obvious even in elderly women (aged >70 years).74 Similar to overweight/obesity and body weight, the effect of weight gain throughout adulthood is modified by HRT use, with a strong association reported in nonusers and a null association reported in ever users.68, 71 Nearly all studies that we reviewed for the current review used self-reported heights and weights in adolescence to assess adolescent obesity and weight gain throughout adulthood, and the results from studies of weight gain have been consistent despite differences in study design, definition of weight gain, and analytic procedures. It is possible that this consistency is because weight gain is an increase in body fat and represents a more accurate measurement of adiposity than body weight; body weight contains both fat and lean mass and can be misleading.82 The results from studies of the association between weight gain and postmenopausal breast cancer risk are summarized in Figure 5.
Obesity and Breast Cancer in Hispanic Women
Few studies of the association between obesity and breast cancer have been conducted in Hispanics, although the Hispanic population is the fastest-growing subgroup in the United States,88 and obesity rates are high and rising in this ethnicity.19 The earliest studies reported that obese Hispanic women were nearly 2.5 times more likely to develop breast cancer than lean women22 (Table 1). A population-based case-control study of Hispanics in the United States indicated that adult weight gain was associated with an increased risk of breast cancer, regardless of menopausal status.89 Hispanic women in the highest quartile of weight change between weight at age 18 years and usual adult weight were nearly 2.5 times more likely to develop breast cancer than those in the lowest quartile, but weight and BMI at age 18 years had no effect on risk.89
Table 1. Body Mass Index and the Risk of Breast Cancer in Hispanic and African American Women
Results (Comparing Highest Level With Lowest Level of BMI)
Abbreviations: BMI, body mass index; CI, confidence interval; HRT, hormone-replacement therapy; OR, odds ratio; RR, relative risk.
Premenopausal: OR, 0.96 (95% CI, 0.63-1.46; Ptrend=.86); postmenopausal with no history of HRT use: OR, 0.80 (95% CI, 0.44-1.45; Ptrend=.61); postmenopausal with current or past use of HRT: OR, 0.74 (95% CI, 0.47-1.15; Ptrend=.17)
A more recent case-control study conducted in Hispanics living in the Southwestern United States concluded that the effect of body size on breast cancer risk varied by menopausal status.90 Obesity at age 15 years decreased the risk of premenopausal breast cancer but had no association in postmenopausal women.90 Adult weight gain between the ages 15 and 50 years was associated inversely with postmenopausal breast cancer risk in Hispanic women, a finding that is contradictory to what has been reported in NHW women.90 These results suggest that the effect of obesity on breast cancer risk may be different between Hispanic and NHW women, and more studies are needed to confirm these findings.
Only 1 study has assessed the association between body size and tumor characteristics in Hispanic women.91 This small study (N = 150) indicated that Hispanic women in the highest quartile of BMI were significantly less likely to be diagnosed with tumors >1cm (odds ratio [OR], 0.16; 95% CI, 0.03-0.84).
Obesity and Breast Cancer in African American Women
The prevalence of obesity in African American women is higher than that of either Hispanic or NHW women.19 Although obesity is a risk factor for postmenopausal breast cancer and has a suspected inverse association with premenopausal breast cancer in NHW women, relatively few studies of obesity have been conducted in African American women (Table 1). Although most case-control studies reported no association between obesity and premenopausal breast cancer risk, the findings in postmenopausal women were inconsistent.21, 44, 92-95 An early case-control study of African American women aged <40 years was the only 1 to report a positive association between current BMI and premenopausal breast cancer risk21; those results conflict with results from premenopausal studies in NHW women. However, Mayberry observed an inverse association between BMI at age 18 years and premenopausal breast cancer, a result that is consistent with findings in NHW women.21 The most recent case-control study of African American women indicated that obesity at age 18 years reduced the risk of premenopausal breast cancer by 33% (OR, 0.67, 95% CI, 0.47-0.96), but there was no association between current BMI and premenopausal breast cancer.96 Among case-control studies in postmenopausal African American women, significant inverse,44, 93 positive,94, 95 and null associations between obesity and breast cancer have been reported.92, 96
In the Black Women's Health Study, the only cohort study of obesity and breast cancer in African American women, 1062 incident cases of invasive breast cancer were identified among 59,000 enrolled women.97 In premenopausal women, current BMI, weight gain since age 18 years, waist circumference, and waist-to-hip ratio were not associated with breast cancer risk, but high BMI at age 18 years significantly reduced the risk by 32%.97 The results were similar among postmenopausal women. Although BMI at age 18 years had an inverse association with postmenopausal breast cancer risk, there was a null effect of all other anthropometric measures.97 These results, together with the findings reported by Mayberry,21 suggest that early life BMI may be important in the development of African American breast cancer.
A few studies have evaluated the attribution of obesity to racial differences in stage at diagnosis.98-100 African Americans are between 1.5 times and 2 times more likely to be diagnosed with stage II or greater breast cancer than NHWs,98, 100 and the risk of morbid obesity for African Americans is nearly 10 times that of NHWs (OR, 9.7; 95% CI, 6.5-14.5).99 Adjusting for obesity reduces the risk of advanced stage breast cancer in African Americans compared with NHWs by approximately 33%.99, 100 Although obesity may explain some of the racial differences in stage, most of the differences remain unexplained and probably can be attributed to socioeconomic factors, access to care, and genetic differences.98
Biologic Mechanisms of Obesity and Breast Cancer
Obesity and body size may be involved in the estrogen and insulin pathways, and it is hypothesized that both pathways are associated with the development of breast cancer.90 Women with greater amounts of body fat have higher levels of estrogen.90, 101 Insulin levels, which are positively associated with obesity, also are related to estrogen concentrations.90, 102
Endogenous Sex Hormones
Estrogen is mitogenic and influences many processes within the body, including cell growth.103 Prospective studies of primarily NHW women have indicated that women with the highest estrogen levels are twice as likely to develop breast cancer as those with the lowest levels.14 Because levels of estrogen vary throughout the menstrual cycle, there are limited data on its association with premenopausal breast cancer,104 and most hypotheses involving obesity and endogenous sex hormones are limited to postmenopausal women.
Total serum estradiol levels are associated inversely with BMI in premenopausal non-Hispanic women but associated positively in postmenopausal women.105 This may help explain the differing effects of obesity on breast cancer risk in premenopausal and postmenopausal women.106 Sex hormone-binding globulin (SHBG) is the primary protein carrier of estradiol, and the proportion of free estradiol in circulation is associated inversely with SHBG.15 Obesity reduces the levels of SHBG and, thus, increases concentrations of free estradiol.64 In a pooled analysis of 8 cohort studies, BMI had a significant, positive association with breast cancer risk, but the effect was reduced substantially after adjustment for serum estrogen concentration.64 After adjusting for free estradiol levels, the risk associated with BMI was attenuated toward the null.64 This is consistent with the hypothesis that the positive association between BMI and postmenopausal breast cancer risk is primarily because of the obesity-related increase in estrogens, and particularly free estradiol.64
However, there have been few studies of the biologic mechanisms in minority women. Both premenopausal and postmenopausal African American women have higher total and free estradiol concentrations compared with NHW women.107, 108 Although there is an inverse association between SHBG and obesity in premenopausal African American women, no association has been observed in postmenopausal African Americans.109 Results from the limited number of studies in Hispanic women have been inconsistent, with 1 reporting similar levels of endogenous estrogen concentrations between postmenopausal Hispanic and NHW women108 and another reporting significantly higher estrogen levels in postmenopausal Hispanics compared with NHWs.110 Additional studies in minorities are needed to understand the relation between obesity, endogenous estrogens, and breast cancer in minority women.
A second hypothesis involves the biosynthesis of estrogen in women. Whereas premenopausal women mostly synthesize estrogen in the ovary, postmenopausal estrogen biosynthesis occurs in peripheral sites, including adipose tissue.104 Aromatase is a family of enzymes present in adipose tissue and within breast tumor tissue.111 Among obese postmenopausal women, estrogen biosynthesis occurs in adipose tissue as a result of aromatization of androgens to estrogen.15, 90 Thus, it is hypothesized that postmenopausal breast cancer risk is elevated in obese women as a result of increasing exposure to estrogens.15
Insulin and the Insulin-Like Growth Factor Axis
Insulin-like growth factor I and II (IGF-I and IGF-II) are mitogenic in both normal and cancerous cells and stimulate cell proliferation and inhibit apoptosis.112-114 Both IGF-I and IGF-II bind to IGF receptors on the surface of the cell, and elevated levels of circulating IGFs are associated with an increasing risk of cancer.112 IGF binding proteins (IGFBPs) are responsible for binding the IGFs to the cell, and IGFBP-3 binds >95% of the IGF in the blood.115 It has been demonstrated that increasing levels of IGF-I and decreasing levels of IGFBP-3 are associated with cancer risk.115
Obesity may affect breast cancer risk through these insulin-related pathways. Receptors for IGFs may be synthesized by fat tissue and are expressed in breast tumor cells.116 Early animal studies demonstrated that a reduction in body weight significantly decreased IGF-I levels and breast cancer incidence in mice.117 In postmenopausal NHW women, body weight and weight gain are associated with elevated IGF-I levels and reduced IGFBP-3 levels,80 and obesity significantly increases IGF-I concentrations.118 The available data are scarce on the association between insulin, IGFs, and breast cancer in minority women. A recent study indicated that African American women have higher IGF-I concentrations and lower IGFBP-3 concentrations compared with NHW women119; however, to our knowledge, there have been no studies on the effects of insulin on breast cancer risk in African American women. In the only study we identified among Hispanic women, there was no association between IGF-I, IGFBP-3, and breast cancer risk.120
Insulin and IGFs may be involved in the development of postmenopausal breast cancer through the estrogen pathway. Serum insulin levels are associated inversely with SHBG concentrations in NHW women,121 resulting in an increase in free estradiol. Furthermore, insulin may induce aromatization in the adipose tissue, which would raise estrogen levels and ultimately increase the risk of breast cancer.122 Nearly all studies of insulin, IGF-I, IGFBP-3, and breast cancer have been conducted in NHWs.120 With such limited data on mechanisms in minority women, it is not possible to conclude that the hypothesized biologic mechanisms are the same in NHWs and minorities.
In conclusion, African Americans have a 51% greater prevalence of obesity than NHWs.123 Nearly 40% of African American women are obese, followed by Hispanics (29%), and NHWs (22%).123 Thus, as obesity rates among Hispanics and African Americans continue to rise, there is an urgent need to identify the role that both obesity and adult weight gain play in the development of breast cancer in these minorities. Studies consisting predominantly of NHW women have demonstrated that obesity is associated with an increased risk of postmenopausal breast cancer15, 46, 49, 51, 53, 58-62, 64-70 and a reduced risk of premenopausal breast cancer.15-17, 44-49, 53-57 There also is evidence of effect modification by HRT, because a strong positive association between obesity and postmenopausal breast cancer has been observed in non-HRT users, but not in ever-HRT users or current-HRT users.15, 17, 68, 71, 76 Weight gain during adulthood is associated consistently with an increased risk of postmenopausal breast cancer, regardless of study design.17, 68, 71, 74, 75, 80-86 In the recently released report by the World Cancer Research Fund/American Institute for Cancer Research, body fatness as measured by BMI was identified as a probable risk-reducing factor for premenopausal breast cancer and a convincing cause of postmenopausal breast cancer.124 Furthermore, weight gain during adulthood was identified as a probable cause of postmenopausal breast cancer.124 That systematic review did not stratify by race/ethnicity, but most studies were comprised primarily of NHW women.
In this review, 45 studies of body size and breast cancer risk, predominantly composed of NHW women, were identified. The results were consistent among most studies but provided little insight into the etiology of breast cancer in Hispanic and African American women. By comparison, only 3 studies of body size and breast cancer were conducted in Hispanic women,22, 89, 90 and only 8 studies were conducted in African American women.21, 44, 92-97
With such scant data, it is difficult to draw conclusions about the association between body size and breast cancer in Hispanics and African Americans. The results from the small number of studies are inconsistent but contradict findings from studies of NHW women. Little research has been conducted into possible biologic mechanisms between body size and breast cancer in minorities, and it is possible that the effects of obesity on breast cancer risk differ between the racial/ethnic groups. Without additional studies of both body size and biologic mechanisms in Hispanics and African Americans, our understanding of breast cancer development in these minority women will remain limited. Although most research is conducted in NHW women, guidelines regarding risk factors are generalized for women of all racial/ethnic groups.124 It is essential for breast cancer researchers to recognize that risk factors may vary between racial/ethnic groups. Studies that comprise only minority women are crucial for understanding the existing breast cancer disparities and hopefully will narrow the gap in breast cancer survival that is so wide today.
This work was supported by a training grant in breast cancer disparities from Susan G. Komen for the Cure (Grant KG090010), the Avon Foundation, and a supplement to a grant to The University of Texas MD Anderson Cancer Center as a Specialized Program of Research Excellence (SPORE) in Breast Cancer (P50 CA-116199-02S1).