A Theory Nixed for Why More Heft Means Less Premenopausal Breast Cancer

BOSTON, Nov. 27 -- Cross ovulation and menstrual cycle patterns off the list for an explanation of why overweight women have less of a risk for premenopausal breast cancer.

So suggests a Harvard and Brigham and Women's Hospital group on the basis of an analysis of more than 100,000 women in the Nurses' Health Study II.

The puzzling protective effect is more likely to be hormonal, concluded Karin B. Michels, Sc.D., Ph.D., and colleagues, in the Nov. 27 issue of the Archives of Internal Medicine.

Among the 113,130 premenopausal women included in the analysis of the Nurses' Health Study II, those with a current body mass index of 30 or higher had 19% lower risk of breast cancer compared to women with a normal 20.0 to 22.4 BMI.

This was true even after adjusting for menstrual variables, infertility due to an ovulation disorder, probable polycystic ovary syndrome (PCOS), and other factors that indirectly measure anovulation (hazard ratio 0.81, 95% confidence interval 0.68 to 0.96).

Higher current BMI was significantly associated with estrogen receptor-positive breast cancer incidence (P=0.02) but not with estrogen receptor-negative (P=0.52) or progesterone receptor-positive (P=0.12) or -negative (P=0.87) breast cancer incidence. The same pattern was seen for the association with BMI at age 18 (P for heterogeneity 0.68 for ER and 0.78 for PR).

"Because BMI was more clearly related to ER-positive than ER-negative breast cancer, a role of sex steroid hormones is likely," Dr. Michels and colleagues wrote.

Numerous previous studies have shown lower breast cancer risk with higher BMI in premenopausal women, the investigators wrote, "but the biological mechanisms underlying this perplexing link have remained largely unresolved."

Since high BMI can be associated with irregular or long menstrual cycles or with PCOS, some researchers had suggested anovulation as a possible mechanism.

So the investigators analyzed data from the Nurses Health Study II, which began in 1989 with questionnaires completed by women ages 25 to 42 at baseline. Every two years, the participants were followed for information on demographic variables, lifestyle factors, and medical events including breast cancer. The analysis including only those who had not reported cancer at baseline and were still premenopausal.

From 1989 to 2003, the women reported 1,398 total cases of incident invasive breast cancer, with a significant inverse trend between current BMI and cancer incidence. The findings were:

• In multivariate analysis, hazard ratios ranged from 1.13 (95% CI 0.94 to 1.37) for BMI less than 20 to 0.81 (95% CI 0.68 to 0.96) for BMI 30 or higher (P=0.002 for trend).
• In multivariate analysis controlling for menstrual pattern (categorized by regularity and length), the hazard ratios ranged from 1.13 (0.93-1.37) for BMI less than 20 to 0.78 (95% CI 0.66 to 0.93) for BMI 30 or higher (P=0.001 for trend).
• In multivariate analysis controlling for ovulatory infertility, the hazard ratios ranged from 1.13 (95% CI 0.94 to 1.37) for BMI less than 20 to 0.81 (95% CI 0.68 to 0.96) for BMI 30 or higher (P=0.002 for trend).
• In multivariate analysis controlling for polycystic ovary syndrome, the hazard ratios ranged from 1.13 (95% CI 0.94 to 1.37) for BMI less than 20 to 0.81 (95% CI 0.68 to 0.97) for BMI 30 or higher (P=0.003 for trend).

The striking similarity of these analyses indicates that indirect measures of anovulation do not explain the link between BMI and breast cancer incidence.

The strongest predictor of breast cancer incidence was self-reported BMI at age 18. Being overweight (BMI 27.5 or higher) at this age conferred 43% less risk than normal weight after adjusting for measures of anovulation (HR 0.57, 95% CI 0.41 to 0.81).

Adjusting for BMI at age 18 attenuated the link between current BMI and breast cancer incidence, making it a better predictor of risk than anovulation. The hazard ratios in this multivariate analysis ranged from 1.13 (95% CI 0.93 to 1.37) for BMI less than 20 to 0.92 (95% CI 0.76 to 1.12) for BMI 30 or higher (P=0.16 for trend).

The researchers cautioned that the findings cannot rule out a role for anovulation because they did not measure anovulation directly.

"However, because adjustment for menstrual cycle patterns, infertility due to ovulatory disorder, probable PCOS, and use of oral contraceptives did not even slightly attenuate the association with BMI," they wrote, "anovulation does not seem to be a primary explanation for the reduced risk in heavier women."

They also noted that detection bias was possible since obese women are less likely to seek breast cancer screening than normal-weight women. "It is possible that obese women with preclinical breast cancer delay their diagnosis, moving the detection of their cancer from the premenopausal to the postmenopausal phase."

The analysis was supported by the Massachusetts Department of Public Health. The study was funded by the National Institutes of Health.