To the Editor.
—Colditz et al1 have dismissed the issue raised by Dr Dupont2 that increased surveillance for breast cancer among women receiving estrogen therapy may explain the increased diagnosis of breast cancer seen in this cohort. They argue that since there were no differences in tumor size between estrogen users and nonusers, surveillance bias is an unlikely explanation for their findings. However, we believe that several additional issues require clarification before detection bias can be ruled out as an explanation for their findings.First, Dupont is correct in noting that estrogen users had a 31% higher rate of mammography than nonusers (64% of users and 49% of nonusers had mammography in the previous year; 64 is 31% higher than 49). This increased rate of mammography (31%) is nearly identical to the increased risk of diagnosis of breast cancer seen in nurses using estrogen therapy (36%).Second, although
Bush TL, Helzlsouer K. Estrogen Replacement Therapy and Risk of Breast Cancer. JAMA. 1991;266(10):1357. doi:10.1001/jama.1991.03470100049015