Do the risks of estrogen plus progestin outweigh the benefits in healthy post-menopausal women?

ABSTRACT

BACKGROUND: One of the proposed benefits of postmenopausal hormone replacement therapy (HRT) is the prevention of coronary heart disease. This proposal is based on evidence from nonrandomized observational studies and intermediate outcomes such as improved lipid profiles. The possibility of harm from HRT has also been reported, particularly regarding breast cancer and thromboembolic disease. The Heart and Estrogen/progestin Replacement Study recently challenged the benefits of HRT, showing no overall protective effect on coronary heart disease (and an increased risk of harm in the first year of treatment) for women with prior coronary heart disease.

POPULATION STUDIED: The Women’s Health Initiative is a set of clinical trials with more than 160,000 women enrolled in studies of low-fat diet, calcium and vitamin D supplementation, and post-menopausal hormone use. This particular report focused on the trial of estrogen plus progestin in women with an intact uterus. A total of 16,608 post-menopausal women were randomized to receive either 1 daily tablet of conjugated equine estrogen 0.625 mg and medroxyprogesterone acetate 2.5 mg (Prempro) or placebo. Women were excluded if they had a history of breast cancer, other cancer within 10 years, hysterectomy, anemia, thrombocytopenia, alcoholism, or dementia. Ages ranged from 50 to 79 years (mean 63 years). Approximately 36% of the women were being treated for hypertension, 4.4% had diabetes, and 7.7% reported a history of cardiovascular disease.

STUDY DESIGN AND VALIDITY: This was a well-designed double-blind, randomized controlled trial with concealed allocation. Baseline characteristics were similar between groups. Follow-up was conducted 6 weeks after randomization, every 6 months with questionnaires, and annually with in-clinic visits. Intention-to-treat analysis was appropriate and would tend to find smaller differences between groups given the high dropout rates (42% in the HRT group and 38% in the placebo group). The trial was originally designed to last more than 8 years, but the independent safety monitoring board recommended stopping the trial when the difference in breast cancer rates exceeded a predetermined threshold and the global index was supportive of harm. When the trial was stopped in the spring of 2002, the average follow-up period was 5.2 years.

OUTCOMES MEASURED: The primary outcome measure was the rate of coronary heart disease, defined as acute myocardial infarction requiring overnight hospitalization, silent myocardial infarction determined from serial electrocardiograms, or coronary heart disease death. The secondary outcome measure was hip fracture rate. The primary adverse outcome measure was invasive breast cancer rate. Reported outcomes also included other cancers, total fractures, stroke, pulmonary embolism, deep vein thrombosis, and total mortality. A global index of outcomes was also calculated as a summary measure of risks and benefits. No measures of vasomotor symptoms or quality of life were reported.

RESULTS: Women in the HRT group had a higher annual incidence of coronary heart disease (0.37% vs 0.30%, NNH = 1429), invasive breast cancer (0.38% vs 0.30%, NNH = 1250), stroke (0.29% vs 0.21%, NNH = 1250), and venous thromboembolic disease (0.34% vs 0.16%, NNH = 556). Bone fractures were less prevalent in the HRT group (total annual fracture rate, 1.47% vs 1.91%, NNT = 228), as was colorectal cancer (0.10% vs 0.16%, NNT = 1667). All of these differences except thromboembolic disease lost statistical significance when adjusting for multiple comparisons, but subgroup analyses showed these differences in adverse events regardless of baseline risks of coronary heart disease and breast cancer. Individuals who adhered to the study medication showed greater differences in adverse events, and individuals who had already used HRT before the study had higher rates of breast cancer. Overall mortality was not different in the 2 groups. The excess risk of events in the global index was 19 per 10,000 person-years. In other words, an average of 1 additional adverse event would be expected over 5 years of treatment for every 100 women meeting these criteria.