Data from BreastScreen Victoria, which provides free biennial breast screenings for women over 40 years of age, was used to evaluate the association of hormone replacement therapy (HRT) use with the diagnosis and survival of breast cancer in postmenopausal women. Between January 1993 and December 2000, 4402 eligible women over age 55 were diagnosed with invasive breast cancer and ductal carcinoma in situ.
The rate of use of HRT at the time of breast cancer diagnosis ranged from 20% to 24% in the years from 1993 to 1999 and increased to 29% in 2000. Compared with women with breast cancer who were not taking hormones, those who were taking HRT were younger (mean 62.3 years vs 66.7 years; P > .01) and were more regular attendees at screening appointments with an average interval between screenings of 25.3 months compared with 36.1 months (P = .02) for nonHRT users. Women taking hormones were more likely to be diagnosed with lobular tumors that were well differentiated and more likely to have an interval cancer (P < .01). Lumpectomy was more common in HRT users.
Of 312 deaths in the study population during the study period, 157 were attributed to invasive breast cancer. The average length of survival among these women was 46.0 months. The average follow up for all women in the study was 6.1 years. The 5-year disease-specific survival rate for women taking HRT at the time of diagnosis of breast cancer was 98.2% compared with 97.1% for those not taking HRT (P = .07). For death from any cause, the 5-year survival rate was 97.4% for HRT users and 94.5% for nonusers (P < .01).
Analyses of variables contributing to death from breast cancer were made using a proportional hazards model. High-risk factors included grade III tumor, positive lymph nodes, and tumor size greater than 20 mm. Considering these variables, analysis of risk of death associated with surgery type found that women who underwent lumpectomy had a longer survival compared with women who had mastectomy or no surgery (P < .01).
Analyses according to HRT use considering all variables generated a hazard ratio for death from breast cancer of 0.67 (95% confidence interval [CI], 0.37–1.22) for HRT users compared with nonusers when breast cancer was detected in the first round of screening. When detection was made in subsequent screenings, the hazard ratio for death from breast cancer was 0.61 (95% CI, 0.31–1.18).
Hazard ratios for death from all causes for users versus nonusers were 0.88 (95% CI, 0.58–1.32) in the first screening and 0.49 (95% CI, 0.28–0.86) in subsequent screenings.