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


Descriptive Statistics

The final analytic cohort included 180 885 women. Women aged 50 years and older who were part of the DCPP, but not the analytic cohort, were more likely to be current smokers (14.9% vs 9.9%, respectively), in the low physical activity group (48.2% vs 42.9%), and not have a college education (46.1% vs 63.6%). Otherwise the study population was similar to the overall DCPP population (Supplementary Table 3, available online). Among the women in the analytic cohort, 6930 breast cancers were identified during follow-up (Table 1). Median follow-up after the 10-year weight-change period was 8.3 [interquartile range (IQR) = 6.9–12.8] years. At the start of interval 1, the median age and BMI of participants was 59 (IQR = 55–64) years and 25.1 (22.7–28.3) kg/m2, respectively. At the end of interval 2, median BMI was 25.6 (IQR = 22.9–29.1) kg/m2. Of women in the analytic cohort, 20.3% maintained stable weight (±2 kg) over the two study intervals (n = 36 744). Another 21.8% of the cohort lost weight in interval 1 (n = 39 371), but only half of these women sustained it over the second interval (n = 19 694). More than one-third of the cohort gained weight in interval 1 (n = 67 394), and most sustained the weight gain through interval 2 (n = 42 912).

Women with sustained weight loss had a lower risk of breast cancer than women whose weight remained stable (Table 2). The larger the amount of sustained weight loss, the lower the risk of breast cancer (>2–4.5 kg: HR = 0.87, 95% CI = 0.77 to 0.99; >4.5–<9 kg: HR = 0.84, 95% CI = 0.73 to 0.96; ≥9 kg: HR = 0.74, 95% CI = 0.58 to 0.94). Women who lost at least 9 kg in interval 1 and gained back some (but not all) of the weight also had a lower risk of breast cancer compared with stable weight (HR = 0.77, 95% CI = 0.62 to 0.97). Results for women who gained up to 13.4 kg in the first interval but subsequently lost all the gained weight suggested a slightly lower risk of breast cancer also (HR = ~0.9), but the confidence interval included the null value. Hazard ratios for women with sustained weight gain relative to stable weight were generally greater than 1, but these estimates were largely not statistically significant. All other weight-change patterns had a similar breast cancer risk to stable weight.

Subgroup and Sensitivity Analyses

The breast cancer risk reduction for sustained weight loss was linear (Figure 2) and specific to women not using HT at the start of breast cancer follow-up (>2–4.5 kg lost: HR = 0.82, 95% CI = 0.70 to 0.96; >4.5–<9 kg lost: HR = 0.75, 95% CI = 0.63 to 0.90; ≥9 kg lost: HR = 0.68, 95% CI = 0.50 to 0.93). No associations were observed among HT users. Sustained weight gain was not associated with breast cancer risk for either group (Supplementary Table 4, available online).

The associations we observed for sustained weight loss were also observed for ER+ tumors; similar but not statistically significant associations were found for ER- tumors, possibly because of limited sample size (Supplementary Table 5, available online). Overall, our results showed stronger associations for weight loss among overweight and obese women compared with women with BMI less than 25 kg/m2 (Supplementary Table 6, available online). For sustained weight loss of at least 9 kg, women with a baseline BMI greater than 25 kg/m2 had a 25% (95% CI = 3% to 42%) lower breast cancer risk compared with women with stable weight. As would be expected, very few women with BMI less than 25 kg/m2 (n = 5 cases) had a sustained weight loss of at least 9 kg, and we did not report hazard ratios for this group because of statistical instability. Among women who lost more than 4.5 kg to less than 9 kg, only overweight women were at a statistically significant lower breast cancer risk compared with stable weight (HR = 0.78, 95% CI = 0.61 to 0.98), but an inverse association was also suggested among obese women (HR = 0.88, 95% CI = 0.67 to 1.16) and less so among normal weight women (HR = 0.93, 95% CI = 0.72 to 1.22). No discernable pattern was observed when results were stratified by baseline physical activity (Supplementary Table 7, available online).

In sensitivity analyses, we found no evidence that any individual study was heavily influencing the results. No statistical evidence of heterogeneity across cohorts was observed (P heterogeneity > 0.2 for all categories), but sample sizes were extremely small for individual cohorts, and weight change categories had to be collapsed (Supplementary Table 1, available online). Our findings were also extremely consistent when we dropped each study one at a time, and when we restricted to US studies only (results not shown). Sensitivity analyses using percent, instead of absolute, weight change also showed a similar pattern of results (not shown).