The role of smoking in breast cancer development: an analysis of a Mayo Clinic cohort.

The purpose of this study was to assess the predictive value of smoking history on breast cancer diagnosis in a referral clinic population. We conducted a case-control study using clinical data collected on 8,097 female patients (1,225 breast cancer cases and 6,872 controls) seen in the Mayo Clinic Breast Clinic between August 1, 1993 and November 31, 2003. Breast cancer patients and noncancer patients significantly differed with respect to age at time of the index visit (p < 0.001), number of pregnancies (p = 0.006), number of live births (p = 0.002), vital status at last known follow-up (p < 0.001), current menstruation (p < 0.001), age at menopause (p < 0.001), history of hysterectomy (p < 0.001), use of oral contraception (p = 0.05), duration of oral contraception use (p = 0.001), use of other exogenous hormones (p < 0.001), duration of exogenous hormone use (p = 0.05), breast pain at time of index visit (p = 0.002), smoking status (p < 0.001), and use of five or more alcoholic beverages per week (p = 0.002). After adjustment for these baseline characteristics, having a personal history of smoking was found to be predictive of breast cancer diagnosis (odds ratios [OR] = 1.25, p = 0.004). Other positive predictors for breast cancer diagnosis were: age (OR = 1.02, p < 0.001), history of hysterectomy (OR = 0.66, p < 0.001), prior use of oral contraception for more than 11 years (OR = 2.10, p < 0.001), and prior use of other exogenous hormones/estrogen (OR = 1.81, p < 0.001). In this referral practice having a personal history of smoking is predictive of breast cancer diagnosis. Further studies are needed to further explore this relationship.