Sustained Weight Loss and Risk of Breast Cancer in Women 50 Years and Older

A Pooled Analysis of Prospective Data

Lauren R. Teras; Alpa V. Patel; Molin Wang; Shiaw-Shyuan Yaun; Kristin Anderson; Roderick Brathwaite; Bette J. Caan; Yu Chen; Avonne E. Connor; A. Heather Eliassen; Susan M. Gapstur; Mia M. Gaudet; Jeanine M. Genkinger; Graham G. Giles; I-Min Lee; Roger L. Milne; Kim Robien; Norie Sawada; Howard D. Sesso; Meir J. Stampfer; Rulla M. Tamimi; Cynthia A. Thomson; Shoichiro Tsugane; Kala Visvanathan; Walter C. Willett; Anne Zeleniuch-Jacquotte; Stephanie A. Smith-Warner


J Natl Cancer Inst. 2020;112(9):929-937. 

In This Article


In this large prospective study of women aged 50 years and older, sustained weight loss of at least 2 kg was associated with a lower risk of breast cancer compared with stable weight. The lowest risk, 32% lower than stable weight, was for women who sustained at least 9 kg of weight loss and were not taking HT. Women who regained some of the weight after losing 9 kg were still at a lower risk of breast cancer than women with stable weight. These results were particularly striking for overweight and obese women. Most other patterns of weight change conferred the same breast cancer risk as maintaining stable weight, including women who gained weight in the first 5 years but subsequently lost weight.

Most previous studies of weight change have focused on weight change from early adulthood (eg, age ~18 years) to middle or later adulthood (age 40 years and older). Adult weight gain from young adulthood was consistently associated with postmenopausal breast cancer risk in these studies.[3] In the present study, we were interested in the impact of weight loss after age 50 years. We did not observe a strong association with weight gain for this age group, consistent with the findings from the Women's Health Initiative (WHI) Observational Study (not included in this pooled study population). In both studies the weight gain interval was no more than 5 years; therefore, participants did not gain as much weight as they might have over a longer time period. Previously observed associations with weight gain were only noteworthy at no less than 22.7 kg of gain.

Fewer studies have examined weight loss and risk of breast cancer. Studies of weight loss from early to middle or later adulthood have mostly been null. Weight loss during this time period is relatively rare, and the null results may reflect a lack of power. Furthermore, most studies did not examine whether the weight loss was sustained. To our knowledge, four studies[11–14] prospectively examined associations of weight change in middle or later adulthood and breast cancer risk, and only two examined sustained weight loss.[11,12] The Nurses' Health Study[11] reported that sustained weight loss of no less than 10 kg after menopause was associated with a 57% lower breast cancer risk (95% CI = 14% to 79%), although this result was based on only nine exposed cases. The Cancer Prevention Study-II suggested a similar, but not statistically significant, inverse association.[12] The WHI Observational Study (not included in the pooled study because it did not have three weight measures) observed an association with weight loss (≥5% loss vs stable weight: HR = 0.88, 95% CI = 0.78 to 0.98), but the (included) WHI Clinical Trial study[14] did not. However, neither WHI study examined sustained weight loss. In our study-specific results, sustained weight loss was associated with a suggestive inverse association with breast cancer risk in the WHI clinical trial population (Supplementary Table 1, available online). Our results underscore the importance of accounting for the subsequent weight gain that often immediately follows weight loss, because only weight loss that was sustained was inversely associated with breast cancer.

Mechanistic studies have consistently shown that blood levels of postmenopausal endogenous estrogens are strongly associated with higher BMI and breast cancer risk,[29] and there is evidence to suggest that circulating sex hormone concentrations can be reduced by weight loss. In both weight loss intervention arms of the SHAPE-2 trial, estradiol, free estradiol, and testosterone concentrations were statistically significantly reduced (and sex hormone binding globulin was increased) compared with the control group, which experienced no weight loss.[30] Other studies have also shown that weight loss reduces sex steroid hormone concentrations,[31,32] as well as C-reactive protein, interleukin 6, tumor necrosis factor alpha, insulin-like growth factor 1, and insulin-like growth factor binding protein.[33]

Our finding of a stronger association for non-HT users was not unexpected. This difference by HT use is likely due to the increase in circulating sex hormones caused by exogenous hormone use,[34] which may overwhelm more moderate changes in hormones because of adipose tissue alone.[35] In women not taking HT, the endogenous hormone shift caused by weight loss may be more influential. The importance of this interaction, however, has lessened in recent years because of the drastic decline in the use of postmenopausal hormones, estimated to be down to less than 5% of US women in 2010.[36] We were also not surprised to find that the greatest benefit of weight loss was among women with BMI greater than 25 kg/m2, because very few women with BMI less than 25 kg/m2 lost substantial amounts of weight.

Limitations of the present study include the varying number of years in each weight-change interval from each of the contributing cohort and the inability to use more than three weight assessments. Although we evaluated the influence of most known breast cancer risk factors, we acknowledge that unknown confounders or survival factors (associated with both weight change and breast cancer risk) may have influenced our results through uncontrolled confounding or selection bias. We also cannot be sure that the weight loss in our study was intentional; however, we do not think reverse causality had a major impact on our results, because weight loss does not typically accompany a breast cancer diagnosis (unlike for other cancers).[37] It is also possible that associations of weight loss and breast cancer may be different today than they were when the women in this study lost weight (mid- to late 1980s and 1990s), given that the prevalence of overweight and obesity has increased for all age groups worldwide,[1,38] and women today are more likely to have carried excess adiposity from an earlier age than the women in our study population. We also note that self-reported, rather than measured, weight and height were used in eight of 10 cohorts, and body composition information at the same time points was not available. Although the correlation between self-reported and measured weight has previously been shown to be high (r = 0.97), individual participant reports may be less accurate.[39,40] To address this concern, we stratified our results by method of weight and height ascertainment (Supplementary Table 8, available online) and did not detect differences; likewise, in a study of more than 10 million participants from 239 prospective studies, the Global BMI Mortality Collaboration did not detect differences in results by method of weight and height ascertainment.[41] In addition, we have no reason to believe that potential misclassification by self-reported weight would be differential by case status, although it is more likely to occur among the obese group.[39] We also note that we did not have objectively measured physical activity data in this study, but validation studies have shown that self-reported physical activity can accurately rank adult physical activity levels.[42–44] Finally, the study population was predominately white and college educated and largely included US-based cohorts. Future research is needed to confirm these findings in other populations such as women from different countries, nonwhite populations, contemporary cohorts, and women who have a more disadvantaged socioeconomic status. Despite this set of limitations, this study addressed major limitations of previous studies, specifically to prospectively examine sustained weight loss during middle-to-later adulthood with adequate sample size in a general population study.

In conclusion, we found that losing weight—and keeping it off—was associated with lower breast cancer risk for women aged 50 years and older. This message is particularly important for the two-thirds of the US population who are overweight or obese and therefore at higher risk of breast cancer. Perhaps equally as important, these results suggest that gaining weight, and then losing it, confers the same breast cancer risk as keeping a stable body weight. In other words, it is not too late to lower your risk of breast cancer if you have gained weight after 50 years of age. Prevention of the most common cancer worldwide may be a particularly motivating factor for the near epidemic numbers of overweight women.