Search

Epidemiologic evidence regarding the relationship between smoking and breast cancer risk is inconsistent. Some studies suggest that the relationship depends on interaction with other factors, such as alcohol use, body mass index (BMI), and menopausal hormone therapy (MHT). We investigated the relationship between smoking and breast cancer risk and interactions with breast cancer risk factors in the NIH-AARP Diet and Health Study, a large prospective cohort. Postmenopausal women ages 50-71 years (N = 192,076) in six US states and two metropolitan areas were followed from 1995-1996 through 2006. Risk factor information was self-reported at baseline. Smoking status was based on whether participants smoked ≥ 100 cigarettes in their lifetime and whether they currently smoked (current -15%, former -40%, never -45%). Alcohol use was estimated from a dietary questionnaire, and categorized based on drinking ≤ 5 or > 5 g/day. BMI was calculated from reported height and weight and categorized as ≥ 30 or < 30 kg/m2. MHT use was categorized as current, former, or never use of any estrogen or progestin preparation. Cancer diagnosis, estrogen receptor (ER), and progesterone receptor (PR) data were reported by state registries. After a mean 9.6 years of follow-up, 7,698 women were diagnosed with primary invasive breast cancer. Multivariable-adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated using Cox proportional hazards regression. Multiplicative interactions between smoking and covariates were evaluated using the likelihood ratio test (LRT). Overall, smokers were at an increased risk of breast cancer compared to women who never smoked (current HR 1.19, 95% CI 1.10, 1.28; former HR 1.08, 95% CI 1.02, 1.13). Excess risk diminished as time since quitting increased and was close to null for women who quit smoking ≥ 10 years prior to study enrollment compared to never smokers (HR 1.04, 95% CI 0.98, 1.11). Relative risks differed significantly based on alcohol use (P-LRT < 0.01), but not BMI or MHT use (P-LRT > 0.05). The HR associated with current smoking was 1.15 (95% CI 1.05, 1.25) among women who drank H 5 g/day, but was higher among women who drank > 5 g/day (HR 1.41, 95% CI 1.22, 1.61). The relationship for those who drank > 5 g/day persisted after adjustment for the amount of alcohol (5-10, 10-20, 20-35, >35 g/day) consumed (HR 1.36, 95% CI 1.18, 1.56). Among women who drank > 5 g/day, current smoking was associated with increased risks of hormone receptor-positive tumors (ER+/PR+ HR 1.29, 95% CI 1.01, 1.64; ER+/PR- HR 2.11, 95% CI 1.27, 3.50), but not ER-/PR- tumors (HR 1.07, 95% CI 0.64, 1.79). In summary, we found that smoking was associated with elevated breast cancer risk which was strongest among women who drank > 5 g of alcohol per day. Among these women, smoking-associated increases in breast cancer risk were limited to hormone receptor-positive tumors, consistent with the known relationship between alcohol use and ER+ breast cancer risk. Findings were similar after additional adjustment for the amount of alcohol consumed, suggesting that the increased risks were not due to residual confounding by alcohol dose, although further analyses are needed to fully understand the interaction between these two factors. [ABSTRACT FROM AUTHOR]