In this large, population-based study of black women and white women diagnosed with invasive breast cancer between ages 35 and 64 years, black women had greater risk of mortality than white women overall. High BMI 5 years before diagnosis was associated with increased risk of all-cause and breast cancer–specific mortality among white women, but not among black women.
Because black women generally have a higher prevalence of obesity than white women at the time of breast cancer diagnosis, several authors have postulated that obesity could explain some of the observed racial differences in mortality.21–24
However, our data do not support this hypothesis. Black women were more likely than white women to have lower education levels, ER-negative tumors, nonlocalized tumors, and comorbidities. All of these factors were associated with increased risk of mortality. After considering age, education, study site, tumor stage, ER status, and comorbidities, obesity did not further explain any differences in mortality risk between black women and white women. The study by Chlebowski et al25
showed a continuing racial difference in mortality after adjusting for BMI. However, the number of deaths among black women was small (n = 21), and the median follow-up was only 3.1 years. Further, this study did not address the extent to which inclusion of BMI in the model affected the risk estimates, nor were they able to look separately for black and white women at how BMI affected risk.25
Our results showing increased risk of mortality for obese white women compared with normal-weight white women are consistent with those of most previous studies,4–13
although results stratified by tumor characteristics are not totally consistent with previous studies. Our data and others4,5,10
show that obese women were more likely to be diagnosed with later-stage tumors. We, and one other study,4
observed similar obesity-survival risk patterns among women with different tumor stages. In contrast, another study reported a stronger association among women with early-stage tumors.6
However, the number of women with later-stage tumors was relatively small in the latter study (n = 192).6
Previous studies have demonstrated that obesity is more strongly associated with developing ER-positive tumors than ER-negative tumors.25
For survival, our study and others have suggested that no differences exist in the obesity-survival association by ER status.4,6,8
However, a recent study with a small number of deaths (n = 87) reported that obesity was not associated with overall survival among women with triple-negative breast cancer.26
These results suggest that mechanisms involved in the obesity-incidence association may differ from those involved in the obesity-survival association.
It has been hypothesized that adiposity affects breast cancer prognosis by elevating circulating levels of endogenous estrogen that result from the conversion of adrenal androgens to estrogen in peripheral adipose tissue,27
lowering circulating levels of sex hormone–binding globulin,28
increasing insulin resistance,29
and increasing insulin-like growth factor-1 levels.30
Furthermore, adipose tissue produces leptin and a group of other growth factors (eg, interleukin-6, tumor necrosis factor-α, vascular endothelial growth factor) that may stimulate angiogenesis, lead to more rapid growth of malignant cells, and promote metastasis.31–33
Another possible explanation for the adverse effect of obesity on mortality is attributed to the more advanced tumor stage associated with obesity.33–35
Among women with nonlocalized tumors, but not among women with localized tumors, we observed increased risk of mortality with obesity, especially breast cancer–specific mortality, for both black women and white women. Thus the obesity-mortality association may reflect the more advanced stage at cancer diagnosis.
It is not clear why obesity-associated mortality risk patterns differ between black women and white women with breast cancer. The Cancer Prevention Study II with a cohort of participants free of cancer at baseline examined the effect of BMI on the risk of fatal breast cancer.22
Of interest, the authors found that higher BMI at baseline was positively associated with fatal breast cancer among white women, but not among black women.22
Furthermore, increasing evidence suggests a stronger association between BMI and coronary heart disease, stroke, and cardiovascular disease mortality among white than among black individuals.36
In our study, normal-weight black women had 53% (95% CI, 24% to 88%) greater risk of all-cause mortality than normal-weight white women, but obese black women did not have greater risk of all-cause mortality than obese white women (RR, 1.02; 95% CI, 0.78 to 1.32); these risk patterns were consistent across all strata examined. Thus other factors such as different cultural, social environmental, psychological, and behavioral factors; health care quality and access; and biologic characteristics that are associated with both obesity and survival may be a partial explanation for black women's higher mortality.
A major strength of our study is the large, population-based sampling of black women and white women. To our knowledge, our study is the first to examine the association of mortality with BMI between black women and white women with invasive breast cancer living in the same geographical regions. The detailed information on a large number of potential risk factors for breast cancer incidence and mortality enabled us to assess many potential confounders, build appropriate multivariate models, and evaluate possible effect modifiers. Finally, our high rate and long duration of follow-up (median, 8.6 years; 25th to 75th percentiles, 7.1 to 10.1 years) enabled us to have large numbers of events.
A limitation is that our results are not generalizable to all patients with breast cancer. In the United States, 57% of women diagnosed with invasive breast cancer are between the ages of 35 and 64 years.3
Our results apply only to women in this age range. Further, self-reported measures of body size may be inaccurate, with heavier women more likely to underreport their weights.14,37
Such measurement error would be expected to be nondifferential with respect to mortality, resulting in attenuation of the true underlying associations. Importantly, the validity of self-reported data on weight among black women is similar to that of white women.38
We were unable to assess BMI at diagnosis or weight change after diagnosis, which were of prognostic value in some studies.39,40
Prior studies show no evidence to suggest that women's weight in the few years before cancer diagnosis differs much from their weight at diagnosis.41,42
Further, prediagnosis body weight may be a stronger predictor of mortality than weight gain after diagnosis.9,12
Although we lacked information on the breast cancer therapies, we have presumed that controlling for age, stage of disease, and hormone receptor status has provided some control for treatment. Nevertheless, previous studies have suggested that black women may have less optimal treatment than white women. They are more likely to delay the initiation of treatment,43
less likely to receive radiation therapy or appropriate dose-intensities of adjuvant chemotherapy,44,45
more likely to terminate treatment prematurely,46
and less likely to adhere to recommended treatment regimens44,45
than white women. However, we found no data regarding whether obese black women's treatment differs from that of normal weight black women. Further, the lack of data on human epidermal growth factor receptor 2 expression prevented informative analyses of the obesity-survival association for women with triple-negative subtype.
Because we did not adjust for multiple testing in our assessment of main effects and stratified analyses, some results (eg, that for the inverse BMI association with decreasing breast cancer mortality among black women) may be false positives and due to few deaths in some strata. Additionally, the results for thin women (BMI < 20 kg/m2) may not be as informative as those for heavier women given the small number of thin women.
In summary, we find that obesity 5 years before breast cancer diagnosis is an independent predictor of survival among white women ages 35 to 64 years, but not among black women in this age group. These findings suggest that differences in the distribution of obesity among black women and white women diagnosed with breast cancer are unlikely to account for the poorer survival of black women.