Breast cancer was associated with postmenopausal estrogen and progestin hormone therapy
ACP J Club. 1996 Jan-Feb;124:20. doi:10.7326/ACPJC-1996-124-1-020
Colditz GA, Hankinson SE, Hunter DJ, et al. The use of estrogens and progestins and the risk of breast cancer in postmenopausal women. N Engl J Med. 1995 Jun 15;332:1589-93. [PubMed ID: 7753136]
To examine the risk for breast cancer in postmenopausal women from hormone replacement therapy (HRT) that included progestins.
16-year cohort analysis of women in the Nurses' Health Study.
Community-based study in the United States.
A cohort of 121 700 female registered nurses, aged 30 to 55 years, completed mailed questionnaires in 1976 and every 2 years thereafter until 1992. In 1976, women who reported breast cancer or other cancer (except nonmelanoma skin cancer) and all premenopausal women were excluded from the analysis. As of 1990, 69 586 women were classified as postmenopausal. Follow-up was 95% complete for the diagnosis of nonfatal breast cancer and 98% complete for fatal breast cancer.
Assessment of Risk Factors
Current and past use of postmenopausal hormone therapy and type of hormone (conjugated estrogen, estrogen plus progestin, progestins alone, or estrogen plus testosterone), year of birth, and age at menopause.
Main Outcome Measure
Relative risk (RR) for breast cancer associated with postmenopausal hormone therapy.
Breast cancer was identified in 1935 postmenopausal women. The multivariate adjusted RR for breast cancer (adjusted for age at menopause and type of menopause, parity, age at first delivery, age at menarche, family history of breast cancer, and history of benign breast cancer) was 1.32 (95% CI 1.14 to 1.54) among women using conjugated estrogens alone, 1.41 (CI 1.15 to 1.74) among women using estrogen plus progestin, and 2.24 (CI 1.26 to 3.98) among women using progestins alone. Among women currently taking hormones, the RR for breast cancer was highest in the oldest age group (RR 1.69 for women 65 to 69 y) and lowest in the youngest age group (RR 1.0 for women < 50 y). The risk for breast cancer increased in women currently taking hormone therapy for ≥ 5 years duration (RR 1.46, CI 1.22 to 1.74 for hormone use for the last 5 to 10 y), whereas women with a history of former but not current hormone therapy use had no more risk than women who had never taken hormone therapy (RR 1.00, CI 0.80 to 1.26 for 5 to 10 y of former hormone use). The RR for death from breast cancer was 1.14 (CI 0.85 to 1.51) for women currently taking hormones and 0.80 (CI 0.60 to 1.07) for women who had taken hormones in the past.
The relative risk for breast cancer was increased in postmenopausal women currently taking hormone replacement therapy, particularly estrogen alone or estrogen plus progestin. The risk was increased in women > 55 years of age and in those with > 5 years of hormone use.
Sources of funding: National Institutes of Health and American Cancer Society.
For article reprint: Dr. G. Colditz, Harvard Medical School, Channing Laboratory, 180 Longwood Avenue, Boston, MA 02115, USA. FAX 617-731-1541.
Although clinicians may despair at the seemingly contradictory conclusions of the studies by Colditz and Stanford and their colleagues, these studies provide important new information. In keeping with previous research, both studies support the conclusion that short-term HRT (< 5 years) has no important effect on the risk for breast cancer. Similarly, neither study found evidence that progestins either protect against or dramatically increase the effects of estrogen on the breast. The former claim was made on the basis of an earlier study that has since been discounted because of methodologic problems; the latter concern, raised in a large study done in Sweden, has diminished with longer follow-up from the study (1). Still in dispute are the effects of long-term HRT, which are now commonly prescribed to protect against osteoporosis and coronary heart disease. An increase in breast cancer was shown only among current users in the Nurses' Health Study but not in any group of HRT users in the study by Stanford and colleagues. Several considerations lend greater weight to the findings of the Nurses' Health Study: Hormone exposure data, collected prospectively from medically knowledgeable persons, appear more reliable; analyses excluded in situ disease (present in 16% of women with cancer in the study by Stanford and colleagues); and fewer eligible persons were lost to follow-up. Could the associations be caused by confounding (underlying differences in cancer risk between users and nonusers) or bias (i.e., closer surveillance of hormone users)? Confounding is unlikely because many characteristics that are more common among women taking HRT (lower body mass index, early or surgical menopause, favorable family history) place these women at lower risk for cancer. Also, evidence of detection bias was not found in the Nurses' Health Study and probably cannot explain observed increases in breast cancer mortality.
Nonetheless, the clinical implications of these findings are less dire than they might first appear. In the Nurses' Health Study, risk was confined to current users, even among women who had taken HRT for extended periods, and mortality was not significantly increased among current users as a group. Also, the increase in death from breast cancer among a subgroup of current, long-term users (≥ 5 y; RR 1.45, CI 1.01 to 2.09) was offset by a trend toward decreased risk among former users (RR 0.80, CI 0.60 to 1.07). These findings are consistent with the hypothesis that current use of HRT promotes the growth of existing cancers rather than initiating new cancers. Other reports have documented that most women who develop cancer while taking HRT have a favorable prognosis (2).
These studies reinforce several recommendations. First, clinicians should feel comfortable recommending short-term HRT to relieve perimenopausal symptoms (e.g., hot flashes). Second, combination therapy (estrogen plus progestin) should be limited to women with an intact uterus, to protect against endometrial cancer. Third, a possible increase in the risk for breast cancer must be considered along with the expected benefits of long-term HRT and the preference of each patient. The absolute increases in breast cancer (if real) from HRT seem modest: an increase of 1 to 2 cases of breast cancer per 1000 patient-years of treatment, or a 1% absolute increase in the lifetime risk for dying from breast cancer (i.e., from 3% to 4%). Because CHD and fractures account for more morbidity and mortality than does breast cancer in older women, the benefits of hormone therapy are still likely to exceed the risks for most women, especially those at increased risk for CHD or osteoporosis (3).
It may be many years before the Women's Health Initiative provides more reliable data about the risks and benefits of long-term HRT. Until then, clinicians should engage their perimenopausal patients in an open discussion of symptoms attributable to menopause, the potential benefits and risks of HRT (including effects on the quality of life), and individual concerns and health priorities. Clinicians can offer cautious reassurance to those with indications for long-term HRT, recommend alternative preventive measures for women reluctant to take HRT, and make sure all women taking postmenopausal estrogen receive regular mammograms.
David Atkins, MD, MPH
U.S. Preventive Services Task ForceWashington, DC, USA