The association of hormone therapy with interval breast cancer
Mammography is widely accepted as an effective modality in breast cancer screening, particularly in older, postmenopausal women. Yet, some breast cancers are not detected at routine screening but are subsequently diagnosed in the interval after a negative mammogram (interval cancers). Hormone therapy (HT) may reduce the sensitivity of mammography by its effects on breast density or on tumor growth. Purpose: To investigate the association of the use of HT on the likelihood of interval vs. screen-detected breast cancer and to assess the effects of formulation and cumulative dose of HT. Methods: Study subjects were drawn from a cohort of women enrolled in the Breast Cancer Screening Program (BCSP) at the Group Health Cooperative of Puget Sound (GHC), a health maintenance organization in western Washington State. The BCSP offers routine screening at typically a two-year interval. Post-menopausal women under 80 years of age were selected if they were diagnosed with invasive breast cancer in 1988 to 1995 and had at least one prior screen during 1988 and 1993. Women with interval breast cancers diagnosed within two years of a negative screening mammogram (n=l 10), were compared to women with screendetected cancers (n=353). HT use was determined from the GHC computerized pharmacy database. The association of HT use with interval cancers vs. screen-detected cancers was estimated by odds ratios after adjustment for the confounding effects of age, body mass index and screening history. Results: Among current users of HT, the likelihood of interval cancer in women who used both estrogen and progestogen was higher than in women who had never used hormones (adjusted OR=2.97, 95% 0=1.61-5.47).’ This risk was further elevated for women who had used both estrogen and progestogen for more than 45 months in the 5 years prior to the index screen (adjusted OR=3.47, 95% 0=1.58, 7.59), for women with a body mass index > 25 kg/m2 (adjusted OR=3.46, 95% CI=1.26, 8.89), and for the first screening mammogram (adjusted C)R=4.02, 95% 0=1.52,10.65). Neither current use of estrogen alone, duration of HT use, or time since last use among former users, appeared to be associated with the risk of interval cancer. Conclusion: Screening mammography is more likely to miss abnormalities in current users of HT, particularly in women who currently use both estrogen and progestogen.
- Epidemiology