In Reply: In response to Drs MacLennan and Baber, we agree that detection bias is always a relevant issue to consider in observational studies. Consistent with other studies published since 1966, we reported an increased risk of ovarian cancer in women using HT (relative risk, 1.38; 95% confidence interval, 1.26-1.51).1,2 If users of HT had more pelvic examinations, it would appear reasonable to assume an earlier detection of the ovarian cancer in these women. However, the excess detection of ovarian cancer is only in an introduction phase (comparable with start of a screening program). In a steady-state situation with a stable proportion of HT users, such an earlier detection of ovarian cancer in women using hormones should not produce a higher incidence rate, because after some time this early detection is eventually compensated for by a lower detection rate. The mean follow-up of 8 years in our study should include both an initial potentially higher detection rate followed by a lower detection rate after this introduction.
Mørch LS, Lidegaard Ø. Hormone Therapy Use and Risk of Ovarian Cancer—Reply. JAMA. 2009;302(20):2203-2204. doi:10.1001/jama.2009.1725