Abstract
Background: Utilization of estrogen hormone replacement therapy (HRT) has dropped dramatically since the publication of the Women’s Health Initiative findings questioning its use for prevention of chronic conditions. Recent publications have suggested a link between this drop in HRT use and a parallel drop in breast cancer.
Methods: Maccabi Healthcare Services provide services to nearly 1.8 million members. We tracked estrogen use, mammography screening utilization, and breast cancer diagnoses in a cohort of female members from 2000 to 2006. Inclusion criteria were: age 45 and over on 1/1/2000, active HMO member from 1/1/2000 until death or 12/31/2006, no history of breast cancer prior to 1/1/2000. We used a Cox proportional hazards model to estimate the contributions of mammography screening, age at the start of the study period, and use of oral estrogen therapy on breast cancer risk.
Results: The study, cohort consisted of 168,732 women, of whom 35,606 (21.1%) purchased oral HRT, 116,373 (69%) had at least one screening mammogram, and 3,939 (2.3%) were diagnosed with breast cancer between 1/1/2000 and 12/31/2006. The percent of women undergoing screening mammography was 16.7% in 2000 and peaked at 34.3% in 2005. The proportion purchasing estrogens dropped from 14.4% in 2000 to 5.0% in 2006. Annual cancer incidence was 3.2/1000 in 2000, and peaked at 3.5/1000 in 2006, in parallel with the peak in mammography screening. A Cox proportional hazards model including an interaction term for mammography* oral estrogen use indicated a significant positive association of mammography screening (HR 6.96, 95% C 1.305 – 1.042), oral HRT (HR 2.0, 95% CI 1.32 – 2.90) and age (HR 1.035, 95% CI 1.035 – 1.042) on cancer incidence.
Discussion: We detected an increase in cancer incidence concurrent with and following a rapid decline in HRT use. Our findings demonstrate the competing impacts of increased mammography screening and decreased HRT use on breast cancer incidence and underscore the value of person-level analysis in assessing such trends. Future work will focus on the impact of estrogen-only and estrogen-progestin, and characterize the effects of HRT estrogen receptor-positive and -negative cancers.
