Hormonal replacement theraphy: Fact or fiction? (Editorial).

For over half a century hormonal replacement therapy (HRT) has been accepted as the standard of care for postmenopausal symptoms and many other illnesses associated with and following the menopause. HRT was introduced in the pharmacopeia before rigorous evidence of the efficacy and safety of medications were required by the FDA and other regulatory agencies.

The underlying principle for the use of HRT appears to be quite simple: at the menopause, the ovaries stop producing estrogen (and progesterone); postmenopausal symptoms readily respond to estrogen replacement, so estrogen is recommended for the relief of postmenopausal symptoms. Similarly, the lack of estrogen may lead to atrophy of the genital organs that may lead to dyspareunia and urinary incontinence; estrogen replacement often relieves these conditions. The menopause is associated with hyperlipidemia, estrogen replacement reduces the total cholesterol and low-density lipids and increases the high-density lipids, so it was felt that estrogen would probably be useful for the prevention and treatment of hyperlipidemias and cardiovascular diseases. The administration of estrogen stops the bone loss, and increases the bone mass, so estrogen was felt to be useful for the prevention and treatment of osteoporosis.

The next step was to assume that estrogen is also useful for a number of conditions that are more prevalent after than before the menopause. Cognitive deficits and Alzheimer's disease are more prevalent after the menopause, so it was felt that HRT also might be useful for the prevention and possibly management of these conditions. Cancer of the colon also is more prevalent after than before the menopause, so estrogen might be useful for the prevention of cancer of the colon. Osteoarthritis is more prevalent after than before the menopause, so estrogen also might be useful for this condition. As time went by, more and more potential uses for estrogen replacement were identified. Estrogen rapidly became the panacea for a number of diseases, and many clinicians advocated that unless there is a contraindication, estrogen should be routinely given, and preferably indefinitely, to all postmenopausal women.

Unfortunately, many of these beneficial effects of estrogen were accepted without the hard evidence that is now required from other medications. The results of the HERS study (Hulley et al, JAMA 1998; 280:605-613) cast a serious shadow on the use of estrogen/progesterone for the secondary prevention of cardiovascular diseases. Indeed, that study had to be prematurely terminated because more patients in the estrogen/progesterone group sustained cardiovascular events during the first year of the study than those who were taking a placebo. Estrogen is now no longer recommended for the secondary prevention of cardiovascular disease. Its use for the primary prevention of cardiovascular disease is still accepted, but is being questioned.

Although estrogen prevents the bone loss associated with the menopause and actually increases the bone mineral density, its role in the treatment of osteoporosis is questionable. There is no large scale, prospective, placebo-controlled study to demonstrate that estrogen reduces the fracture risk. Observational studies suggest that HRT is associated with a 50% reduction in vertebral fractures and a 25% to 30% reduction in the risk of hip fractures (Barett-Connor, BMJ 1998; 317:457-461). Observational and retrospective studies, however, are open to a number of biases.

The PEPI study (JAMA 1995; 273:199-208), on the other hand, a prospective study that included only 875 subjects randomly allocated to either placebo or estrogen with or without progesterone only showed that estrogen increases the bone mineral density, but there was no evidence of a reduction in fracture risk. Admittedly a population of 875 subjects is too small to draw any useful conclusion on the effects of fractures. But the HERS study, which included 2,763 subjects, monitored for a mean of 4.1 years, also did not show any difference between the active and placebo group on fractures of the long bones. Although this study was not powered to examine the effect of estrogen on fractures, it is somewhat disheartening that there was no difference in fracture rates between the two groups, given the mean age of this population: about 68 years and the duration of the study. There is indeed a paucity of studies to document the beneficial effects of estrogen on the reduction of fractures in patients with osteoporosis (Manson, Martin, N Engl J Med 2001; 345:34-40). Most of the data on fracture risk reduction is derived from observational or cross-sectional studies.

The beneficial effects of estrogen on cognitive functions are also questionable. In the study conducted by Matthews et al (J Am Geriatr Soc 1999; 47:518-523), women who were on estrogen replacement exhibited a slower decline in mental functions than those who were not on this therapy. The women who were on estrogen, however, had a higher level of formal education than the women who were not on estrogen. An analysis of the data showed that whereas current oral estrogen use does not protect against age-related declines in cognitive functions, formal education protects against such a decline. Women who have reached a high level of formal education are more likely to take estrogen than those who have not reached such a level. The apparent beneficial effects of estrogen on cognitive functions therefore may be due to some other factor, such as the level of formal education and not to the estrogen administration itself.

This is probably a major issue with estrogen replacement: it is not known whether many of the observed beneficial effects of estrogen are actually due to the estrogen per se, or whether they are due to some other common denominator, such as level of formal education, lifestyle, nutritional intake, level of exercise taken, cigarette smoking, general concerns about health, and access to medical care. We still do not have irrefutable evidence about the beneficial effects of estrogen on a number of conditions such as osteoporosis, coronary artery disease, and Alzheimer's disease.

The definitive answer will not be available until the Women's Health Initiative Study conducted under the auspices of the National Institute of Health is completed. The results are expected in 2005. The results of the Women's International Study of Long Duration Estrogen after Menopause, conducted in 14 countries, are not expected until 2012. Hopefully, these studies will determine whether estrogen is useful in a variety of clinical conditions. Unfortunately, these results will not be available for quite a few years. So what are we to do in the meantime?

Although the practice of Medicine is still an art, it is now based on firm scientific foundations. It is reassuring that in many instances firm data are available to help clinicians. Clinicians, however, need to ask themselves what exactly are they trying to achieve while treating the particular patient sitting across from their desk. They then need to refer to the evidence-based data concerning the best available medication to treat that condition. If they are trying to relieve postmenopausal symptoms estrogen is probably the best medication available. If, on the other hand, they are trying to treat hypercholesterolemia or osteoporosis, then maybe they should consider other medications which have been shown to be effective in the management of these particular conditions. There is, as yet, no panacea that can effectively treat all conditions at the same time.

Ronald C. Hamdy, MD, FRCP, FACP 

Editor