Estrogens and Postmenopausal Osteoporosis: Was Albright Right After All?


  • John P. Bilezikian

    Corresponding author
    1. Departments of Medicine and Pharmacology, College of Physicians and Surgeons, Columbia University New York, New York, U.S.A.
    • Address reprint requests to: Dr. John P. Bilezikian, Department of Medicine, College of Physicians and Surgeons, 630 West 168th Street, New York, NY 10032 U.S.A.
    Search for more papers by this author

THE PRIMARY ROLE OF estrogen deficiency in postmenopausal osteoporosis has been a credo in this field since it was first proposed by Fuller Albright in 1941.1 Over the past 60 years, abundant experimental evidence has continued to provide unequivocal support for the validity of this concept.2 In 1983, Riggs and Melton proposed a refinement to the simple, unitary concept that estrogen deficiency is responsible for osteoporosis by suggesting that postmenopausal osteoporosis consists of two phases, which they termed Type I and Type II.3 In Type I osteoporosis, accelerated bone loss occurring during the first 4–8 menopausal years was due rather exclusively to the acute estrogen deficiency state. We are beginning to understand more completely the mechanisms by which estrogen deficiency leads to rapid bone loss at this time. Estrogen appears to control the local production of bone-resorbing cytokines and other factors. In the wake of estrogen deficiency, these local bone-resorbing factors emerge as a major pathophysiological mechanism for osteoclast activation.4 Women vary in the extent to which they experience accelerated bone loss due to estrogen withdrawal. The Riggs–Melton hypothesis required other determinants, besides estrogen deficiency, to account for the variability in rates of bone loss among early postmenopausal women. Genetic, racial, and nutritional factors, for example, are likely to be important determinants with respect to these early skeletal consequences of estrogen deficiency. Another determinant that dictates whether such processes will eventuate into osteoporosis is the bone mass accrued earlier during the formative period of skeletal growth. The greater the peak bone mass achieved in childhood and adolescence, the greater bone loss has to be in the menopausal years to lead to osteoporosis.

The second phase of the Riggs–Melton hypothesis (Type II) describes a slower process that is less well accounted for by estrogen deficiency than by mechanisms associated with aging per se. Because both genders are believed to share in this aspect of bone loss, estrogen was relegated to a relatively unimportant role. In both women and men, other factors were considered to be important, such as an increase in parathyroid hormone (PTH) concentration with advancing years. The age-related rise in PTH and other homeostatic “breakdowns” were implicated in the slow but inexorable loss of bone mass associated with aging in both sexes.

The distinction between Type I (estrogen dependent) and Type II (estrogen independent) osteoporosis was conceptually easy, but the facile separation in time (early-Type I vs. late-Type II) was somewhat artificial. In women, the processes associated with aging are occurring at the same time as the processes associated with acute estrogen deficiency. In women, however, accelerated bone loss predominates in the early menopausal years, the slower process becoming apparent only thereafter. In men, only the slow, age-related phenomenon is appreciated from the onset of bone loss. The lack of an accelerated phase of bone loss in men is explained by the fact that men do not experience the equivalent of a menopause (andropause) in their middle years. Most cartoons describing the time line of adult bone loss show two components in women (fast followed by slow) and a single phase in men (slow only).

In the current article, Riggs, Khosla, and Melton take a bold conceptual step by proposing that estrogen, in fact, is primarily responsible for both types of bone loss.5 Having thus “refined” the Albright hypothesis by distinguishing between early and late menopausal bone loss 15 years ago, they now attempt to explain both events under the unitary conceptual umbrella of estrogen deficiency. They argue further that uniphasic, age-related bone loss in men can also be accounted for, at least in part, by age-related changes in estrogen levels.

This is a big article because it proposes a new way of thinking about estrogen deficiency and bone loss, harkening also back to the original concept of Albright. The authors propose that two separate pathophysiological dynamics are associated with estrogen deficiency: one responsible for early accelerated bone loss in women and another dynamic process responsible for slower, later bone loss in women and men. In the first phase, acute estrogen deficiency causes uncompensated skeletal resorption. The ensuing efflux of calcium from bone tends to suppress PTH levels and is associated with an increase in urinary calcium excretion and a reduction in gastrointestinal calcium absorption. These renal and gastrointestinal consequences are explained by homeostatic mechanisms to protect the organism from hypercalcemia. Although not explicitly proposed, the relative decrease in PTH occurring at this time could be part of this homeostatic adjustment. The reduction in PTH, for which there is admittedly only circumstantial evidence, could play a role by facilitating renal calcium excretion and limiting 1,25-dihydroxyvitamin D [1,25(OH)2D] production. In the short term, the net effect is substantial negative calcium balance and rapid bone loss.

In the second phase, Riggs et al. propose that newly appreciated extraskeletal actions of estrogen deficiency predominate over the skeletal actions that are so evident in the early postmenopausal years. According to their hypothesis, aging unmasks the extraskeletal consequences of estrogen deficiency. The extraskeletal effects of estrogen deficiency lead directly to an increase in renal calcium excretion and a reduction in gastrointestinal calcium absorption. It should be noted that these same target organs are implicated in the earlier phase of estrogen deficiency but the authors choose to explain a similar pathophysiology at that earlier time by obscure mechanisms not related to estrogen deficiency. The gastrointestinal consequences at this later time are believed to be due to a direct effect of estrogen deficiency on calcium absorption as well as to an indirect effect of estrogen deficiency to reduce the formation of 1,25(OH)2D.

The hypothesis to account for this second phase of bone loss proceeds to invoke a secondary, compensatory rise in PTH due to these estrogen-related renal and gastrointestinal effects as well as perhaps to a direct consequence of estrogen deficiency to stimulate parathyroid glandular activity. It is proposed further that this formulation could also explain age-related bone loss in men who experience reductions in estrogen levels as they age. Recent observations are supportive, correlating bone loss in aging men more significantly with declining estrogen levels than with declining androgen levels.6–8

As is true for most provocative hypotheses, it is a challenge to reconcile all elements of the concept with what is really known. With incomplete knowledge, one can only point out areas in which observations may be a variance or for which more data are needed. For example, the hypothesis does not explain why or how the skeletal consequences of estrogen deficiency would predominate in the early postmenopausal years while the putative extraskeletal effects would predominate later. The authors propose that the effects of estrogen deficiency in the skeleton wane over time but that the extraskeletal effects do not. The “biological clock” by which this time sequence of events unfolds is obscure.

A second point is that our knowledge of the extraskeletal actions of estrogens is woefully incomplete. Very little is known about the effects of estrogen upon the renal handling of calcium. Equally sparse are data pointing to direct actions of estrogen on gastrointestinal absorption of calcium or on vitamin D metabolism. The validity of the hypothesis hinges, in part, upon documenting these extraskeletal actions of estrogens more completely.

A third area for which more data are needed pertain to the role of PTH. The conundrum is that in the early phase of estrogen deficiency, associated with accelerated bone loss, PTH appears to be relatively suppressed, whereas in the later phase of estrogen deficiency, associated with a slower rate of bone loss, PTH appears to be relatively increased. One is tempted to say that you cannot have it both ways but the authors propose that you can. The only way you can have it both ways is to “buy into” the time sequence of initial skeletal and subsequent extraskeletal effects of estrogen. If one could show that in fact this chronology occurs, as well as show that estrogen really does harbor potent extraskeletal actions, the sequence of changes in PTH dynamics could be explained.

Another area for which the hypothesis has to account is a time-related slowing of the profound estrogen effect on cancellous bone while accomodating an effect of increasing PTH levels on cortical bone. The authors address this dilemma simply by stating that the skeletal actions of estrogens wane over time. But how? It is possible that the increasing PTH levels are responsible for mitigating the resorptive effects of estrogen deficiency on cancellous bone. Such a formulation gains support by observations in primary hyperparathyroidism in which the protective effect of PTH on cancellous bone in postmenopausal women has been well documented.9 In fact this and other observations have led many investigators to consider PTH as a potential therapy for postmenopausal osteoporosis.10 It is ironic, and a sobering comment on the state of our knowledge, that Riggs et al. view PTH rather exclusively as a culprit while others regard PTH as a potential cure for postmenopausal osteoporosis!

In time, the “waning” skeletal effects of estrogen and the emerging greater presence of PTH raise issues regarding the kind of bone loss that occurs in this later, Type II, phase of bone loss. One would not expect to see disproportionate reductions in cortical bone—as one sees in primary hyperparathyroidism—because the increases in PTH levels are small and virtually always still within the normal range. One would not expect to see continued disproportionate reductions in cancellous bone in this later period because the predominant estrogen effect on cancellous bone is ameliorated. Thus, the rather similar slow kinetics of bone loss at both cancellous and cortical sites in the later period of postmenopausal bone loss can be accomodated by the hypothesis of Riggs et al., although they do not choose to account for it in the way I have suggested.

In the aging male, the notion that estrogen deficiency may also be pivotal in the mechanisms of bone loss is of great interest. Observations in the few men described with a genetic defect in the estrogen receptor,11 a state of estrogen resistance, or in aromatase activity12,13 in which estrogen levels are absent demonstrate a critical role for estrogens in the establishment of peak bone mass in men. The estrogen-resistant or estrogen-deficient young men are osteoporotic. In the men with aromatase deficiency, treatment with estrogen markedly improves bone mass.13,14 Recent tantalizing data also suggest that estrogen might contribute to the events associated with bone loss in men.6–8 It is premature, however, to speculate on how important estrogen is in the processes associated with bone loss in the aging male.

I do not think Riggs, Khosla, and Melton would want the readers of their excellent article to miss the point that there are many factors, besides estrogen deficiency, that contribute to bone loss in the later postmenopausal years. The possibility that estrogen deficiency could be important in the slower, age-related mechanisms by which bone is lost does not exclude a number of other important mechanisms that are independent of estrogen. We certainly recognize, as do the authors, that 1,25(OH)2D levels fall with age-related reductions in renal function, that the gastrointestinal tract becomes less responsive to vitamin D with aging, and that many aging individuals are deficient in dietary and solar sources of vitamin D. These changes in vitamin D are not features of estrogen deficiency per se. We also recognize the importance of adequate calcium intake in the overall skeletal health of postmenopausal women. Deficiencies in dietary calcium, combined with vitamin D deficiency, can contribute importantly to the increases in PTH associated with aging. These forces can give rise to age-associated abnormalities in calcium balance independent of estrogen. Other nutritional points, life style elements, and genetic determinants are clearly some of those “other factors” that distinguish further the estrogen-deficient woman who develops osteoporosis from the estrogen deficient woman who does not develop osteoporosis. Women who take estrogens in their postmenopausal years are not invariably protected from osteoporosis. Conversely, women who choose not to take postmenopausal estrogens are not necessarily doomed to this disease. Clearly, the development of postmenopausal osteoporosis is determined by factors that go well beyond an exclusive role for estrogens.

Finally, it should not be overlooked that other pathological events attributable to intervening disease processes, drugs, and life style can contribute to bone loss in the postmenopausal years. Postmenopausal women are subject to bone loss if primary hyperparathyroidism, thyrotoxicosis, or Cushing's syndrome develop; if they receive glucocorticoids, other immunosuppressive agents, or anticonvulsants; and if they smoke, drink alcohol excessively, or live a sedentary life.

Riggs and his colleagues are to be congratulated for a worthy attempt to formulate a coherent, unitary hypothesis to account for postmenopausal bone loss, as first suggested by Albright. They add an important concept to the dialogue, namely that estrogen deficiency may well help to explain not only the early phase of postmenopausal bone loss but also the later, slower phase associated with aging. With this view, they also are able to accomodate the possibility that men too lose bone with age, in part because of declining estrogen levels. Although Albright first proposed that estrogen deficiency is responsible for postmenopausal osteoporosis, Riggs, Khosla, and Melton are correct in concluding that their newer formulation of Albright's unitary concept is considerably more complex than he proposed. Future studies are likely to reveal even more of this complexity.