Hormone Replacement Therapy: Relationship to Survival
ABSTRACT & COMMENTARY
This is the third major report of the Nurses’ Health Study, and this report is a critical analysis of mortality related to hormone replacement therapy (HRT), with a focus on cardiovascular disease and cancer deaths. Grodstein and colleagues were particularly interested in the relationship of current vs. past use of hormones and the underlying risk profile for heart disease and breast cancer. The prior reports from this longitudinal study appeared in 1995 (N Engl J Med 332:1589) and 1996 (N Engl J Med 335:1406) and emphasized the relationship of breast cancer to hormone use, as well as the effect of combined estrogen-progestin administration vs. estrogen alone. This study began in 1976 when almost 122,000 female registered nurses, age 30-55, were enrolled. A questionnaire was used every two years to track updated information on risk factors, HRT status, and any new diagnosis of significant illness or death. The primary end point was death from any cause; coronary artery disease, stroke, or cancer risks were emphasized. The observational study was terminated in 1994; almost 100% of the 3600 deaths were investigated. Overall, less than 15% were due to coronary artery disease, while more than 50% were from cancer (425 breast cancer deaths). Each case subject who died was matched with 10 controls chosen at random who met the same criteria. Biennial questionnaires were filled out by the women between the years 1978 and 1992. In the case of a new diagnosis of cancer or other illness, the health status of the previous questionnaire was used to avoid bias that might lead to hormone discontinuation because of illness. Risk of coronary artery disease was assessed using standard risk factors; risk of breast cancer was considered for those with a mother or sister with breast cancer.
Results: There was a strong reduction in all-cause mortality among active hormone users, of whom 16% reported current use on the last questionnaire, and 28% were past users. Current HRT was associated with a 42% relative risk reduction in total deaths, and after adjusting for CAD risk factors, particularly cigarette smoking, there remained a 37% lower risk of death. However, prior hormone use did not increase survival. CAD mortality was markedly reduced by 53% in current users, and cancer mortality was also lowered in current hormone users by 29%. However, these survival benefits were attenuated among long-term HRT users, with the relative reduction in risk for 10 or more years of current HRT use declining to 20% (vs 37% for all current users); this reduction of benefit appeared to be mostly due to a 43% increase in death from breast cancer in the long-term users. Women who had stopped using HRT within three years had a 22% decrease in risk of death; there was no benefit in former users who had not taken HRT for more than five years. As previously reported, combination therapy with progestin did not decrease the benefit of HRT; in fact, among current users, there was a decrease of 54% in risk of death for combination therapy vs. 31% for estrogen alone.
Careful analysis of CAD risk and the relationship to benefit from estrogen indicated that presence of major CAD risk factors substantially increases the benefit of protection from HRT. Thus, women with any CAD risk factor (smokers, high cholesterol, hypertension, diabetes, family history, high BMI) had a 49% decrease in death from all causes compared to only an 11% decrease in mortality among the women at low risk with no CAD risk factors (who represented only 1/7 of the entire cohort). Even in those with a family history of breast cancer (mother, sister) present in 11% of the entire population, there was a 35% decrease in all-cause mortality among current users. The few younger women (less than 50) who were current users had no survival benefit. Conversely, for women older than 60, current use resulted in a 42% decrease in mortality, with use for more than 10 years imparting less benefit than use for less than 10 years. Grodstein et al conclude that current hormone users had a substantially lower all-cause mortality rate than never users, particularly due to death from CAD; this benefit was no longer present five years after stopping HRT. HRT women with CAD risk factors had the greatest increase in survival. Other major HRT studies have emphasized the dilemma of prevention bias or a "healthy user bias," which is well known and reflects the fact that women who choose to use HRT are generally healthier than non-users. The Nurses’ Health Study would appear to reduce such bias, in that the presumed range of socioeconomic status among participants is narrow. Grodstein et al conclude that "confounding in this study was more rigorously controlled" than in other reports. This study demonstrated a 35% reduction in the incidence of colorectal cancer, as well as a robust decrease in mortality in those women who developed this cancer who were current users. Similarly, the risk of dying from breast cancer among current users was reduced by 24% vs. non-users, even though the relative risk of developing breast cancer was increased. Grodstein et al conclude that, "on average, the survival benefits appeared to outweigh the risks, but the risks and benefits vary depending on existing risk factors and duration of hormone use must be carefully considered for each woman." (Grodstein F, et al. N Engl J Med 1997;336:1769-1775.)
COMMENT BY JONATHAN ABRAMS, MD
These data speak for themselves and clearly support a major protective role for HRT with respect to fatal and nonfatal CAD in women who are currently using estrogen replacement. Similar to the men in the Western Scotland trial of cholesterol lowering, higher CAD risk women had greater benefit, and those free of CAD risk factors appeared to have little prolongation of survival. These data should assist the physician and the patient in making decisions regarding HRT, as this risk analysis provides newer insights as to the risk-benefit ratio of HRT. In younger women and those individuals who have a family history of breast cancer, there was a greater likelihood of developing breast cancer with HRT, particularly with long-term use. Former hormone users did not have an increased likelihood of breast cancer. Of particular interest is the fact that survival with breast cancer was improved in hormone users, which may reflect earlier detection or a therapeutic benefit relating to cessation of estrogen once breast cancer develops. HRT is not generally recommended solely for protection against CAD risk, and in fact, recent surveys indicate that even in women who use HRT, understanding of the risk of CAD and concern about developing CAD is a minor issue. Women and their physicians need to be far better educated regarding the actual status of HRT, a subject in which fear of breast cancer often dominates informed decisionmaking.
The Nurses’ Health Study overwhelmingly demonstrates that there is a major improvement of survival in current users, but the study also provides information as to how to better stratify benefit and risk. Furthermore, the intriguing suggestion that long-term continued use (> 10 years) is associated with a reduction of survival benefit (mostly due to the development of breast cancer) provides considerable ammunition for the concept that HRT is not necessarily best for life-long use. In an accompanying editorial, the issue of breast cancer and women at low- and high-risk for this condition is discussed without a definite conclusion, although a question is raised regarding the possibility of restricting the duration of HRT. New estrogen preparations will be available in the near future that are tissue specific, and it is hoped that risk of breast (and uterine) cancer will be substantially diminished with these preparations. Until that time, the current data strongly favor HRT in general, with the admonition that careful consideration of this decision must be made by any woman and her physician contemplating such therapy. In women with a bad CAD risk factor profile, and, without question, those who have established CAD, the impetus to use HRT should be favorably influenced by the robust epidemiologic data demonstrating a major decrease in CAD mortality.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content