Response: Re: Declines in Invasive Breast Cancer and Use of Postmenopausal Hormone Therapy in a Screening Mammography Population

The magnitude of decline in invasive breast cancer has been remarkably consistent across studies ( 1–4 ). Two of the main proposed causes for the decline are a decrease in screening mammography and a decline in use of postmenopausal hormone therapy (HT). We reported a decline in invasive cancer of 5% annually or 15% over a period of 3 years and a decline of estrogen receptor (ER)–positive invasive cancer of 13% annually or 26% over a period of 2 years in a population of women undergoing routine mammography in which the time between screening examinations was similar among current users of HT, former users of HT, and nonusers. The annual decline in ER-positive breast cancer remained comparable and statistically significant ( Ptrend = .01) to what we reported even when we omit the first two data points in 2001 and last data point in 2003, despite Anderson’s suggestion that these data points drove our statistically significant findings. In summary, we have shown that a decline in screening mammography rates is unlikely to account for the recent decline in US breast cancer incidence. Anderson suggests that, because the breast cancer rate we reported was virtually unchanged between 1997 and 2004 and the decline in breast cancer starting in 2000 was not proportional to the decline in HT use, it is hard to conclude that there is any relation between HT use and decline in invasive breast cancer. However, given that mammography use does not account for the decline in breast cancer rate that we observed, our results—particularly the corresponding changes in rates of HT use and breast cancer over time—do suggest that the precipitous decline in HT use likely played a part in the change in breast cancer rates. For example, we observed a statistically significant ( Ptrend <.001) 2% annual increase in HT use from 1997 to 2000 and a concomitant statistically significant 9% annual increase in invasive cancer from 1998 to 2000 ( Ptrend = .03). Likewise, as the rate of HT use started to decline in 2000, we observed a statistically significant 5% annual decline in invasive cancer, as noted above. Robbins and Clarke state that a percentage change in breast cancer incidence is dependent on the absolute change in prevalence of HT use and the relative increase in breast cancer risk among HT users compared with nonusers. We observed a 31.6% absolute decrease in HT prevalence from 2000 to 2004, but a lower 15% overall decline in breast cancer rate because invasive breast cancer risk is increased only 24% (95% confidence interval [CI] = 2% to 50%) in HT users relative to nonusers ( 5 ). We observed a greater decline in the rate of ER-positive breast cancer (26% over a period of 2 years) because ER-positive breast cancer risk is 72% (95% CI = 55% to 90%) higher in HT users than in nonusers ( 6 ). Potential mechanisms that could result in a decline in breast cancer incidence are shown in Table 1 . Of these potential mechanisms, we have shown that screening mammography use is not likely to account for a decline in breast cancer rates but that a decline in HT use likely plays a role. Measuring ER-positive breast cancer rates in continuous, former, and never users of postmenopausal HT may inform our understanding of the magnitude of the contribution of declines in postmenopausal HT use on breast cancer incidence. Potential mechanisms to explain recent decline in breast cancer incidence * HT = postmenopausal hormone therapy; DCIS = ductal carcinoma in situ. Mechanisms not mutually exclusive.