Hormone replacement therapy and accuracy of mammographic screening

Lancet. 2000 Jan 22;355(9200):270-4. doi: 10.1016/S0140-6736(99)07319-5.


Background: Hormone replacement therapy (HRT) is commonly used and may affect the accuracy of mammographic screening.

Methods: We examined the sensitivity, specificity, and small-cancer detection rate according to HRT use in 103,770 women in Victoria, Australia, who attended first-round screening in 1994 and who did not have a personal history of breast cancer or a breast lump or a bloodstained or watery nipple discharge at the time of screening. BreastScreen Victoria provides mammography to women aged 40 years and older every 2 years. Unconditional logistic modelling was used to adjust for age, family history, and symptom status.

Findings: The sensitivity of screening mammography for a 2-year screening interval was lower in HRT users (64.8% [95% CI 58-72]) than non-users (77.3% [74-81]). In the target group (50-69 years), the sensitivity was 64.3% (57-72) in HRT users and 79.8% (76-84) in non-users. Among women who were diagnosed with cancer during the 2-year screening interval, HRT users were more likely to have a false negative result than non-users (odds ratio 1.60 [1.04-2.21]) after adjusting for potential confounding factors. Specificity was 0.6% lower in HRT users compared with non-users. Among women who did not have cancer diagnosed in the interval, HRT users were more likely to have a false positive result (adjusted odds ratio 1.12 [1.05-1.19]).

Interpretation: We show that HRT use reduces the sensitivity of mammographic screening. In countries where HRT use is widespread, the reduction in sensitivity with HRT use may undermine the capacity of population-based mammographic-screening programmes to realise their potential mortality benefit.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Aged
  • Breast Neoplasms / diagnostic imaging*
  • Breast Neoplasms / epidemiology
  • Female
  • Hormone Replacement Therapy*
  • Humans
  • Logistic Models
  • Mammography / statistics & numerical data*
  • Mass Screening / statistics & numerical data*
  • Middle Aged
  • Sensitivity and Specificity
  • Victoria / epidemiology