Oregon Health Sciences University, Portland VA Medical Center, Portland, Oregon
Address all correspondence and requests for reprints to: Eric S. Orwoll, Bone and Mineral Unit, Oregon Health Sciences University, 3181 S.W. Sam Jackson Park, Portland, Oregon 97201.
"We Walk by Faith Not by Sight" (2 Corinthians 5:7)
The importance of estrogen in bone and mineral metabolism is incontrovertible. It has many important actions on skeletal function at molecular, cellular and tissue levels, and has been more intensively studied than perhaps any other compound used in skeletal therapeutics. Even the low levels of endogenous estrogens present in postmenopausal women have recently been noted to be associated with meaningful skeletal effects (1). The exciting discoveries of a second form of the estrogen receptor and a complex array of molecular events at the level of estrogen-mediated transcription indicate that the mechanism of estrogens effects is more intricate than ever imagined. As a direct result of new molecular insights, selective estrogen receptor modulators (SERMs) have been developed and promise to further expand estrogens impact on bone health. Since estrogen also has a fundamental role in the management of postmenopausal symptoms and may be an important modulator of cardiovascular and neurological health, it has long been rightfully considered a mainstay in the management of disorders that are associated with menopauseincluding osteoporosis. The perception that estrogens benefits are comprehensive (and at a reasonable cost!) has lead to a strong advocacy of its role as the primary defense against osteoporotic fractures.
Unfortunately, the vigorous espousal of the undoubted skeletal value of estrogen has probably led to overconfidence. In reality, as we discuss below, the ability of estrogen to reduce the burden of osteoporotic fractures, albeit real, seems quite modest. Moreover, the evidence that buttresses the practical usefulness of estrogen is delicate. Rather than take too much solace in the merits of estrogen, it is vitally important that we better understand and accept the limitations of estrogen therapy and begin to look hard for approaches that can extend the benefits it provides.
Estrogen and bone mass
In clinical terms, the prevention of early postmenopausal bone loss is one of the most salient manifestations of estrogens pleotropic effects. A reduction in trabecular and endocortical bone mass, in the face of a dramatic increase in the rate of remodeling, has been repeatedly documented to follow the menopausal decline in gonadal function. Appropriate replacement doses of estrogen ameliorate, if not completely prevent, these changes in skeletal homeostasis. For 510 yr after the menopause, virtually all measures of bone loss and remodeling activity can be maintained at premenopausal levels with exogenous estrogen use, whereas a withdrawal of estrogen therapy results in a recurrence of menopausal remodeling dynamics and bone loss. These findings have been found to be essentially universal in healthy early postmenopausal women and have led to the belief that estrogen replacement continues to be effective in the prevention of bone loss throughout the postmenopausal period. Although this contention is actually untested, it is widely assumed to be true. For instance, the American Association of Clinical Endocrinologists Clinical Guidelines on Osteoporosis recommend that an evaluation of skeletal health in postmenopausal women on estrogen is not needed. Similarly, the Canadian Medical Association Practice Guidelines on Osteoporosis suggest that when a woman and her physician agree that estrogen replacement therapy will be used in the postmenopausal period, bone mass measurements are unnecessary.
In fact, this degree of faith in the effectiveness of estrogen is probably misplaced. Of most importance, it must be remembered that osteoporotic fracture is primarily a disorder of older women. The average age of women who sustain a hip fracture is at least 75; 25 yr after the usual time of menopause. In spite of the impressive actions of estrogen on bone mass in the early postmenopausal years, the effects in older women (those most likely to suffer fractures) are much less reassuring. In the longest follow-up of women placed on hormone replacement in the early menopause (approximately 10 yr), a decline in bone mass seems to commence in the later stages of observation (2). This is a phenomenon predicted by Heaney, Nordin and others (3, 4), who postulated that other aspects of age-related bone loss (rather than sex steroid deficiency) could dominate the older years. In the Framingham experience, a positive effect of estrogen on bone density was not evident in women older than 75 yr (although there were few who had taken estrogen continuously since menopause) (5). Most impressive, in older women (mean age 72) participating in the Study of Osteoporotic Fractures, the rate of bone loss was noted to accelerate with increasing age, and those on estrogen replacement continued to lose bone density, as did those without estrogen (6). Estrogen use reduced the rate of bone loss but by only 35% at the proximal femur and 33% at the calcaneus. Even though estrogen clearly had a positive effect, the continuation of bone loss, despite the use of estrogen, is of obvious clinical concern, particularly in light of the strong relationship between bone density and fracture risk in that same population. The finding that older postmenopausal women lose bone despite estrogen replacement strongly suggests that the causes of bone loss at these ages are more complex than in the early postmenopausal period and that estrogen alone is incapable of ensuring skeletal integrity.
The usefulness of estrogen supplementation, initiated later in the postmenopausal period, has been examined in a very small number of short-term trials. In some, women were given estrogen and compared with placebo-treated controls. On average, bone loss was prevented, and even a slight increase in bone mass was observed. However, these reports did not provide the proportion of women who continued to lose bone while taking estrogen, how they differed in other ways from those who did not, or how long-lasting the positive effect was. Essentially, the confidence with which we can advise estrogen for skeletal protection in individual older women is limited.
Most women take estrogen for few of the postmenopausal years; and because the cost, inconvenience, and risk of some complications increases with duration of use, it has been hoped that a limited period of estrogen use can provide ongoing benefit. In fact, some have suggested that 510 yr of estrogen therapy after menopause can yield a lasting betterment of bone mass. As might be expected, the usefulness of previous (but discontinued) estrogen use in the maintenance of bone mass is actually quite unclear. Not only are the available trials uncontrolled and usually retrospective, but they are also inconsistent. Nevertheless, most suggest that there is a rapid diminution of bone mass after estrogen is withdrawn, so that 5 yr after discontinuation, measures of bone density are indistinguishable from those in women who have never taken estrogen. By the time women reach the age at which fractures become most common (>75 yr), the benefits of estrogen taken years before may be minuscule.
Estrogen and fractures
The most important goal of estrogen administration for osteoporosis is the prevention of fractures. Unlike trials of the effects of estrogen replacement on bone mass in the early menopause (often placebo controlled), the ability of estrogen to reduce fracture risk has been noted only in uncontrolled, frequently retrospective, observational or case-controlled studies. The apparent fracture protection afforded by estrogen therapy in these studies has sometimes been impressive (relative risk of fracture < 0.5), but the lack of randomization or placebo controls limits the usefulness of these results. Despite attempts to statistically adjust for other potentially interacting variables, the relative effectiveness of estrogen in reducing fracture risk cannot be confidently estimated without rigorously designed, randomized, controlled trials (like the Womens Health Initiative). In addition, women who have chosen to take estrogen replacement have, on average, healthier lifestyles than those who do not (7, 8), and it is very possible that an important part of the protective effect associated with taking estrogen is mediated via other factors (exercise, diet, and others).
Nevertheless, the evidence that supports the efficacy of estrogen in fracture risk reduction deserves attention. A relationship between estrogen use and lower fracture risk has been a consistent finding, and it is highly probable that estrogen exerts a beneficial effect on fracture risk in older women. Recent meta- and decision analyses estimated that the relative risk of hip fracture in estrogen-treated women is 0.460.75 (9, 10). These reductions in fracture risk are important, but certainly incomplete. Because half of menopausal women are estimated to experience a nontraumatic fracture, these effect sizes would still leave an enormous public health problem. To compound the limitation of the estrogen effect, the largest trials suggest that the greatest fracture risk reductions occur in the youngest women (those generally at least risk for fracture). Once again, it is important to emphasize that all the data concerning fracture risk are uncontrolled, and estrogen-taking women have been shown, in general, to be more involved with preventive health practices. The pooled estimates of estrogens effect (meta- and decision analyses) do not formally take this issue into account. Therefore, even the limited degree of protection noted above must be considered optimistic when projected into the general postmenopausal population.
A recent evaluation of incident fracture rates in the Study of Osteoporotic Fractures population illustrated the limitations of estrogen therapy. In women who reported taking estrogen continuously since the menopause (an average of 26 yr), almost 20% experienced a nontraumatic, nonvertebral fracture during a 9-yr follow-up period (11). This was fully two thirds the number observed in women who had never taken estrogen. The extent to which estrogen was associated with lower rates of vertebral and hip fracture was similar. Clearly, many women who take estrogen will ultimately suffer osteoporotic fractures.
Summary
In sum, the skeletal benefits provided by hormone replacement therapy are important, and estrogen should be justifiably considered one of the fundamentals of menopausal management. However, its efficacy in the prevention of osteoporotic fractures should not be overstated. Low bone mass and fractures remain serious threats in older postmenopausal women, even in the presence of hormone replacement. With the recognition of that reality comes a variety of challenges to clinicians and investigators.
Clinicians should use estrogen to its best advantage, though, at the same time, remain vigilant of its limitations. Older postmenopausal women who are, or who have been, taking estrogen should not be a priori considered adequately protected against fracture. The same careful clinical evaluation recommended for the protection of those at increased fracture risk, including bone mass measures, should be available to estrogen-taking women in the later postmenopausal period. Recognizing this need, the National Osteoporosis Foundation has recently recommended bone mineral density testing in older women, regardless of estrogen status, and HCFA reimbursement has become available for this indication. In the presence of low bone density, estrogen-taking women should be afforded appropriate levels of diagnostic and therapeutic attention, with the intent to further reduce fracture risk. Once estrogen therapy has been elected, patients and their clinicians should be aware that bone loss can still be expected in the later postmenopausal years. Periodic reevaluations of bone density and other risks for fracture (e.g. falls) may be appropriate. In the face of continued good health, those reevaluations can be infrequent, but women who have medical conditions associated with adverse skeletal effects should be followed more closely despite their estrogen therapy.
If we are to build on the value of estrogen to improve our approach to osteoporosis, additional data are needed. When fracture risk has been considered in complex models that adjust for account other medical and lifestyle variables, a greater protective effect of estrogen has emerged (12), suggesting that there are factors whose actions attenuate those of estrogen. Prominent candidates include genetics, tobacco and alcohol use, medications, body composition, and propensity to fall. There are undoubtedly others yet unidentified. These should be further defined, and the basic mechanisms for interactions between them and estrogen should be examined. It would be clinically advantageous to use these factors to accurately identify not only the women who would benefit from estrogen replacement but also those who would not.
Because estrogen has important beneficial effects, an essential research objective should be the development of therapeutic strategies that combine the advantages of estrogen with other modalities (pharmacological or otherwise), to maximally protect the many estrogen-taking women who may be considered at continued risk for fracture. To whit, the initial reports of combination therapy with bisphosphonates are encouraging (13, 14). Other important skeletal agents can be anticipated to interact with estrogen in as-of-yet unforeseen ways. On another front, it can be anticipated that selective estrogen receptor modulators will be subject to estrogen-like limitations when used for osteoporosis prevention/therapy, and their effects will be influenced by a similar, but distinct, array of covariates.
Our vision of the role of hormone replacement in the reduction of osteoporotic fracture risk should be considerably sharpened. At one time, estrogen was the lone best hope for the avoidance of fracture, but no longer should we be satisfied with that imperfect solution. The development of much more efficient and effective methods for osteoporosis risk assessment, prevention, and treatment now provides a chance to surpass previous expectations. Because many women who have elected to take estrogen will suffer fractures nonetheless, efforts should be directed at identifying those with continued risk, and at gaining a greater understanding of how to manage osteoporosis while using estrogen in the context of complementary approaches.
Received February 3, 1999.
Accepted February 26, 1999.
References