Author Affiliations: Obstetrics and Gynecology Epidemiology Center, Department of Obstetrics and Gynecology (Drs Xue and Michels), and the Channing Laboratory, Department of Medicine (Drs Willett, Rosner, Hankinson, and Michels), Brigham and Woman's Hospital, and Harvard Medical School, Boston, Massachusetts; and Departments of Epidemiology (Drs Xue, Willett, Rosner, Hankinson, and Michels), Nutrition (Dr Willett), and Biostatistics (Dr Rosner), Harvard School of Public Health, Boston.
Johnson and colleagues raise insightful questions on our assessment of passive smoking related to the amount of detailed measurement included, potential inclusion of women exposed to 1 measurement in the nonexposed group of another measurement, and potential confusion about the derived passive smoking index.
Because the assessment of passive smoking relies heavily on subjective recall, which is likely to differ between cases and noncases, results from case-control studies are likely biased.1 Our results of no association between passive smoking and the incidence of breast cancer is consistent with 7 of 8 cohort studies published up to December 2006.2 Three more recent cohort studies produced conflicting results3- 5: one identified an increased incidence only associated with the longest duration of exposure in childhood and adulthood,3 one identified an marginally increased incidence associated with the highest cumulative exposure in terms of both duration and intensity after age 20 years,4 and the other identified no association.5 Another recent large prospective cohort study in combination with a meta-analysis of all published studies up to January 2008 concluded there was no association as well.1 Therefore, evidences from the majority of cohort studies suggest a null association between passive smoking and breast cancer risk.
Xue F, Willett WC, Rosner BA, Hankinson SE, Michels KB. Analysis of Passive Smoking and Breast Cancer of Limited Value—Reply. Arch Intern Med. 2011;171(16):1508-1510. doi:10.1001/archinternmed.2011.384