Analyses were based on 61,335 women who had normal or higher body weight and were cancer free at baseline, survived at least 3 years, and had adequate data.
Overall, about 13.3% of women lost at least 5% of their body weight between baseline and year 3; 7.9% did so intentionally, losing an average of 19.6 pounds, and 5.5% did so unintentionally, losing an average of 16.9 pounds.
“We used the 5% decrease because this level has been shown to change some biochemical markers potentially associated with cancer. And it has been shown in a different study population, a randomized trial, to reduce the frequency of diabetes,” Dr. Chlebowski noted.
“There was nothing that the study did to induce the weight loss. And based on a partial look at the data, it doesn’t look like many of the women went to some kind of program to lose weight. So it was probably self-directed weight loss,” he noted. “Interestingly, the body mass index of this group was 29.9, so they were on the verge of going into obesity. We wondered in retrospect whether that was a motivating factor for them.”
In multivariate analysis, relative to peers having a stable weight over time, the women losing at least 5% of their weight had a lower risk of breast cancer (hazard ratio, 0.88; 95% confidence interval, 0.78-0.98; P = .02).
Findings were unchanged after further adjustment for mammography frequency. In addition, risk reduction was statistically indistinguishable whether the weight loss was intentional or not (P = .2 for interaction), and whether women were normal weight, overweight, or obese at baseline (P = .4 for interaction).
The 19.6% of women who had a weight gain of at least 5% did not have a significantly elevated risk of breast cancer overall, but they did have a significantly elevated risk of triple-negative breast cancer (HR, 1.54; 95% CI, 1.16-2.05). “We really don’t have a good explanation for this,” Dr. Chlebowski said.
Tackling the global obesity epidemic by conventional means to reduce cancer risk is an uphill battle, he concluded. “We are going to be looking for possible mediating factors. If you can find the mediating factors, pharmacologic intervention would have a much greater chance of success.”
Dr. Chlebowski disclosed that he had no relevant conflicts of interest. Research was supported by the National Heart, Lung and Blood Institute; National Institutes of Health; Department of Health and Human Services; and American Institute for Cancer Research.
SOURCE: Chlebowski et al., SABCS 2017 Abstract GS5-07