A Unitary Model for Involutional Osteoporosis: Estrogen Deficiency Causes Both Type I and Type II Osteoporosis in Postmenopausal Women and Contributes to Bone Loss in Aging Men


  • B. Lawrence Riggs,

    Corresponding author
    1. Endocrine Research Unit, Division of Endocrinology, Nutrition, Metabolism, and Internal Medicine, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, U.S.A.
    • Address reprint requests to: B. Lawrence Riggs, M.D., Mayo Clinic and Mayo Foundation, 200 First Street SW, North 6 Plummer Building, Rochester, MN 55905 U.S.A.
    Search for more papers by this author
  • Sundeep Khosla,

    1. Endocrine Research Unit, Division of Endocrinology, Nutrition, Metabolism, and Internal Medicine, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, U.S.A.
    Search for more papers by this author
  • L. Joseph Melton III

    1. Department of Health Sciences Research, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, U.S.A.
    Search for more papers by this author


We propose here a new unitary model for the pathophysiology of involutional osteoporosis that identifies estrogen (E) deficiency as the cause of both the early, accelerated and the late, slow phases of bone loss in postmenopausal women and as a contributing cause of the continuous phase of bone loss in aging men. The accelerated phase in women is most apparent during the first decade after menopause, involves disproportionate loss of cancellous bone, and is mediated mainly by loss of the direct restraining effects of E on bone cell function. The ensuing slow phase continues throughout life in women, involves proportionate losses of cancellous and cortical bone, and is associated with progressive secondary hyperparathyroidism. This phase is mediated mainly by loss of E action on extraskeletal calcium homeostasis which results in net calcium wasting and increases in the level of dietary calcium intake required to maintain bone balance. Because elderly men have low circulating levels of both bioavailable E and bioavailable testosterone (T) and because recent data suggest that E is at least as important as T in determining bone mass in aging men, E deficiency may also contribute substantially to the continuous bone loss of aging men. In both genders, E deficiency increases bone resorption and may also impair a compensatory increase in bone formation. For the most part, this unitary model is well supported by observational and experimental data and provides plausible explanations to traditional objections to a unitary hypothesis.


IN 1941, ALBRIGHT ET AL.1 called attention to estrogen (E) deficiency as the cause of postmenopausal osteoporosis, and this mechanism has been strongly supported by subsequent studies showing that E administration prevented bone loss induced by oophorectomy in perimenopausal women2,3 and that bone cells were targets of E action.4–6 However, factors in addition to E deficiency have been implicated as causes of bone loss in aging women, including secondary hyperparathyroidism,7,8 impaired osteoblast function, possibly due to changes in local cytokines9 or systemic growth factors,10,11 and, at least in some elderly persons, nutritional deficiency of vitamin D.12 Moreover, aging men also lose bone despite clinically evident hypogonadism occurring only infrequently. These considerations have militated against Albright's unitary hypothesis that a relative or absolute deficiency in sex steroids causes all, or almost all, involutional bone loss.

Any comprehensive model of pathogenesis should explain the differences and similarities of bone loss in aging women and men. Women undergo two phases of involutional bone loss whereas men undergo one.13 After achieving peak bone mass in young adulthood, bone mineral density (BMD) remains relatively constant in both genders until middle life, although Slemenda et al.14 found small decreases in BMD at the hip in premenopausal women. At menopause, women undergo an accelerated, transient phase of bone loss that is most apparent over the subsequent decade and accounts for cancellous bone losses of 20–30% and cortical bone losses of 5–10%.13 Because the age of menopause varies and because bone loss precedes menopause by several years in some women,14,15 this phase is less evident in cross-sectional studies but is readily apparent in longitudinal studies following oophorectomy.2,3 The accelerated phase of bone loss in women merges asymptotically with an underlying late phase of slow bone loss which then continues indefinitely. The slow, continuous phase of bone loss in aging men resembles the late slow phase in postmenopausal women. Over life, this slow phase accounts for losses of about 20–30% of cancellous bone and about 20–30% of cortical bone in both genders.13 These patterns are shown schematically in Fig. 1.

Figure FIG. 1.

Schematic representation of changes in bone mass over life in cancellous (broken line) and cortical (solid line) bone in women (left panel) and men (right panel) from age 50 onward. Note that men have only one phase of continuous bone loss but women have two—an early accelerated phase and a late slow phase. Note also that the accelerated phase, but not the slow phase, involves disproportionate loss of cancellous bone.


Early effects of menopause on bone loss

The accelerated phase of bone loss begins at menopause, can be prevented by E replacement, and clearly results from cessation of ovarian function. E acts through high affinity E receptors (ERs) in osteoblasts4,5 and osteoclasts6 to restrain bone turnover, and when this restraint is lost at menopause, overall bone turnover increases. In a population-based study of 653 French women, Garnero et al.15 found that menopause induced a net increase in bone resorption: bone formation markers were increased by about 45% and bone resorption markers were increased by about 90%. In addition, E deficiency increases the sensitivity of bone to parathyroid hormone (PTH)16 and, perhaps, to other resorption-inducing agents, and this further enhances the resorption defect. Compensatory increases in urinary calcium excretion17,18 and decreases in intestinal calcium absorption17,19 prevent the resultant skeletal outflow of calcium into the extracellular fluids from producing hypercalcemia. Consequently, serum intact PTH levels remain within normal limits in early postmenopausal women, although there is a trend toward slightly decreased levels.20–22 In at least some studies, institution of E replacement results in increased serum PTH levels18,23,24 (Fig. 2), suggesting that there was indeed a small compensatory decrease in PTH secretion at steady state.

Figure FIG. 2.

Changes in serum ionic calcium and intact PTH and urine free deoxypyridinoline cross-links, an index of bone resorption, in 18 early postmenopausal women studied at baseline (BSL) and after 6 months of treatment with 0.1 mg/day transdermal estradiol patches (ERT). Data were obtained by analysis of samples stored at −70°C from the study reported by McKane et al.18 Note that homeostatic increases in serum PTH levels maintained a constant level of serum, ionic calcium despite reduction of bone resorption by estrogen therapy.

Secondary hyperparathyroidism

Serum intact PTH levels and indices of bone turnover increase with equal pace in aging women, and these increases correlate directly with each other, even after adjusting for the effect of age (Table 1). That suppression of PTH secretion by intravenous calcium infusion abolished the differences in bone resorption markers between young and elderly women25 (Fig. 3) strongly suggests that the increase in bone resorption in aging women is PTH dependent. Levels of serum PTH and bone resorption markers also increase in aging men, and the pattern of the increase closely resembles that in aging women.7,8,24,26 Thus, a considerable body of data implicates secondary hyperparathyroidism as the cause of the slow, age-related phase of bone loss in both genders.

Table Table 1. Changes in Serum PTH and in Biochemical Markers for Bone Turnover in 304 Women Residents of Rochester, MN from the Third into the Tenth Decade of Life
original image
Figure FIG. 3.

Changes in excretion of urine N-telopeptide cross-links of type I collagen (NTx), a marker for bone resorption, as assessed by measurements in 4-h specimens over 24 h in 10 young adult premenopausal women (A) and 10 elderly postmenopausal women (B) before (open circles) and after (closed circles) 24-h calcium infusion.25 This completely suppressed serum PTH in the young adult women but it only partially suppressed serum PTH in the elderly women (data not shown). Nonetheless, the age-dependent increase in bone resorption was almost completely eliminated by the calcium infusion, suggesting that secondary hyperparathyroidism was involved in causation.

If the increase in bone resorption in aging women was caused only by the loss of E action on bone cells, the resulting outflow of calcium from the skeleton would decrease PTH secretion. As noted earlier, this appears to occur during the accelerated early phase of bone loss. The late slow phase of bone loss, however, is associated with a progressive increase in serum PTH levels and thus must be due to another mechanism: a compensatory secondary hyperparathyroidism for net calcium losses from the body due to several age-related abnormalities in extraskeletal calcium homeostasis. Renal function declines with aging but not to levels that would increase parathyroid function.27 Renal calcium conservation25 and intestinal calcium absorption, however, also decrease in elderly women.28–31 The decreases in calcium absorption with aging have been attributed variously to impaired metabolism of vitamin D to 1,25-dihydroxyvitamin D (1,25(OH)2D),30,32 primary impairment in calcium absorption,7,33 or decreased concentrations of intestinal vitamin D receptors.34 Finally, some elderly women, especially those who are housebound and reside in countries that do not fortify milk products with vitamin D, have vitamin D deficiency that may contribute to the secondary hyperparathyroidism.12

Because of these age-related abnormalities in extraskeletal calcium homeostasis, the elderly must consume more dietary calcium to prevent negative calcium balance. By combined calcium balance and kinetic methods, Heaney et al.17 found that premenopausal women required 1000 mg/day, whereas postmenopausal women required 1500 mg/day to maintain balance, which is far greater than the average calcium intake of only 700 mg/day among American postmenopausal women.35 Also, a number of recent randomized clinical trials have shown that calcium supplements are effective in reducing bone loss in elderly women.36–39 Additionally, McKane et al.40 performed intensive metabolic studies in a group of young adult premenopausal women on their usual calcium intake, a group of elderly women on their usual calcium intake, and a group of elderly women who had received a high calcium supplement for 3 years (Table 2). The unsupplemented elderly women had the expected increases in parathyroid function and in bone resorption markers, whereas the calcium-supplemented elderly women had values that were indistinguishable from those in the young premenopausal women. These data support the hypothesis that a failure of elderly women to increase their calcium intake to a level high enough to offset the age-related increases in net calcium losses contributes substantially to their development of secondary hyperparathyroidism and increased bone resorption.

Table Table 2. Comparative Effects of Age and Calcium Intake in Women as Previously Described by McKane et al.(40)
original image

Effect of continued E deficiency in late postmenopausal women

All agree that E deficiency causes the accelerated transient phase of bone loss in early postmenopausal women. It is less clear, however, what role E deficiency plays in the pathogenesis of the secondary hyperparathyroidism and increased bone turnover associated with slow bone loss in late postmenopausal women. In theory, there could be no interaction (the abnormalities are independent of E deficiency), a partial interaction (E deficiency contributes to the abnormalities), or a complete interaction (E deficiency causes or is permissive for the abnormalities).

Somewhat surprisingly, recent studies indicate that there is a complete interaction. In contrast to prevailing belief, Prestwood et al.41 found that short-term E treatment of elderly women (>80 years old) decreased values for biochemical markers of bone turnover significantly. McKane et al.42 studied three groups of women: a young adult premenopausal group (Group A), a group of untreated elderly women (Group B), and a group of elderly women who were receiving long-term E therapy (Group C). Because Groups B and C were of comparable age, but differed in estrogen sufficiency, and Groups A and C were comparable with respect to estrogen sufficiency, but differed in age, the effect of age and E deficiency could be disassociated. As shown in Table 3, after E deficiency was corrected, there were no effects on serum intact PTH and bone resorption levels that could be attributed to aging per se. Also, in a population-based observational study, Khosla et al.24 found that serum intact PTH and bone turnover increased progressively in untreated postmenopausal women, whereas there were no increases in postmenopausal women receiving E therapy. In addition, Heshmati et al.43 demonstrated that reduction of the already low levels of serum E to near undetectable levels in a group of late (mean, age 69 years) postmenopausal women by administration of letrozole, an aromatase inhibitor that blocks conversion of weak androgens to E in adipose and other peripheral tissues, significantly increased bone resorption markers by about 15%. Thus, even the low levels of serum sex steroids in late postmenopausal women modulate bone turnover. These recent studies documenting a central role of E deficiency in the etiology of bone loss in late postmenopausal women are consistent with the earlier study by Richelson et al.44 who compared values for BMD at the lumbar spine, proximal femoral, and radius in three groups of women: a normal group of premenopausal women, a group of oophorectomized women who were of comparable age but were 20 years postmenopausal, and a group of postmenopausal women who were of comparable years postmenopausal but were 20 years older. They found that the values for BMD in the young, E-deficient women were similar to those of the older postmenopausal women but were significantly less that those of the young normal premenopausal women.

Table Table 3. Comparative Effects of Age and Estrogen Status in Women as Described by McKane et al.(40)
original image

Collectively, these data indicate that E deficiency is responsible for all, or almost all, of the bone loss in aging women, at least up to the age of 75 years. If so, how is it possible to reconcile the age-related secondary hyperparathyroidism that drives the slow phase of bone loss with the presence of continued E deficiency, which in the accelerated transient phase of bone loss in early postmenopausal women suppresses PTH secretion? The answer to this question requires a rethinking of our traditional ideas about how E modulates skeletal and extraskeletal calcium metabolism.

Effect of E deficiency on extraskeletal calcium homeostasis

That E affects bone turnover directly through its action on bone cells is well established. That it also affects bone turnover indirectly through an action on extraskeletal calcium homeostasis is a newer and less well accepted concept.45 By combined balance and kinetic studies in transmenopausal women, Heaney et al.17 found that about half of the negative calcium balance induced by menopause was due to decreased calcium absorption and about half to increased renal calcium excretion, although they did not determine whether these abnormalities were the cause or the result of the negative calcium balance.

Several studies have documented effects of E deficiency on extraskeletal calcium homeostasis. Gallagher et al.30 found that E treatment increased both serum total 1,25(OH)2D levels and calcium absorption in postmenopausal osteoporotic women. Some investigators have attributed the increase in serum total 1,25(OH)2D levels after orally administered estrogen to an increase in serum levels of vitamin D binding protein induced by a “first pass” stimulation of hepatic synthesis,46,47 whereas others have found that free serum 1,25(OH)2D levels increased,48,49 suggesting that estrogen may alter the metabolism of vitamin D. Moreover, in perimenopausal women before and 6 months after oophorectomy, Gennari et al.19 found that the increase in calcium absorption in response to treatment with 1,25(OH)2D was blunted in the presence of E deficiency, suggesting a direct enhancing effect of E on intestinal calcium absorption. Using estimates based on regression analysis, Nordin et al.50 found that early postmenopausal women had a “renal calcium leak” that they attributed to E deficiency. McKane et al.18 assessed renal calcium transport by direct measurements at baseline and during administration of a saturating dose of PTH in early postmenopausal women before and after 6 months of E treatment. They demonstrated a PTH-independent decrease in tubular calcium absorption in the E-deficient as compared with the E-replete women, but no effect on tubular reabsorption of other cations—observations that are consistent with a direct effect of E on renal calcium conservation. Finally, Cosman et al.51 found that E treatment of postmenopausal women decreased PTH secretory reserve as assessed by measurement of serum PTH during hypocalcemia induced by infusion of EDTA, suggesting strongly that E is acting directly on the parathyroid gland to decrease PTH secretion. Also, Naveh-Many et al.52 reported that parathyroid glands contain ER,52 but others have failed to confirm this.53 Thus, E appears to have skeletal effects that decrease bone turnover directly and extraskeletal effects that decrease it indirectly.

E deficiency and decreased bone formation

Although age-related bone loss is caused mainly by increased bone resorption, impaired bone formation also contributes. At menopause, bone resorption increases more than bone formation does.15,54 From analysis of calcium kinetic data, Ivey and Baylink55 pointed out that inadequate compensatory increases in bone formation to offset the increase in bone resorption was an important cause of bone loss in early postmenopausal women. Using histomorphometry of bone biopsy samples from late postmenopausal women, Lips and Meunier56 demonstrated decreased wall thickness of trabecular packets, which is incontrovertible evidence of decreased bone formation at the cellular level.

Because the increased bone resorption and the impaired compensatory bone formation occur concurrently at the menopause, both are almost certainly caused by E deficiency. In contrast, the decreased bone formation associated with the late postmenopausal phase of bone loss has generally been attributed to age-related factors, particularly to decreased paracrine production of growth factors13,57 or to decreased circulating levels of growth hormone58 and insulin-like growth factor I.11,58 However, the observation that E treatment increases production of IGF-I59 and transforming growth factor-β by osteoblastic cells in vitro,60 raises the interesting possibility that impaired osteoblast function in the late postmenopausal women is also caused by E deficiency. Also, an effect of E on stimulating skin collagen synthesis by fibroblasts has been repeatedly demonstrated61,62 and, as with fibroblasts, collagen production by osteoblasts represents >90% of cell biosynthetic capacity. However, E treatment has been reported both to stimulate63 and to inhibit64 bone formation in experimental animals and to stimulate65 and inhibit66 proliferation of human osteoblastic cells in vitro.

Thus, although not as yet conclusively demonstrated, a considerable body of evidence suggests that E deficiency impairs compensatory increases in bone formation in both the early and late phases of involutional bone loss in women.


Men do not have an equivalent of the rapid phase of bone loss that women experience following menopause. After accounting for the absence of this phase, the pattern of the continuous phase of bone loss and the increases in serum PTH and bone resorption markers in aging men are virtually superimposable on those occurring in women,8,22,26 which is consistent with a common causal mechanism for the slow bone loss in both genders. If E deficiency underlies the slow phase of bone loss in postmenopausal women, how can the causal mechanism of the bone loss in men be the same as that in women? Conventional wisdom holds that bone mass is maintained mainly by E in women and mainly by testosterone (T) in men, and this long-standing belief is supported by the rapid bone loss after oophorectomy in women2,3 and after orchiectomy in men.67 Only a few aging men develop overt hypogonadism, however, and serum total T decreases only slightly with aging in men.68 We suggest here the radical hypothesis that E deficiency makes a substantial contribution to the continuous phase of bone loss in men. Because the testes produce both sex steroids and because T can be converted in peripheral tissues to E, a deficiency of either or both could mediate postorchiectomy bone loss in men. Also, human osteoblasts contain both ER4,5 and androgen receptors,69 although, thus far, osteoclastic cells have been found to contain only ER.6,70

Data from several recent “experiments of nature” are consistent with the concept that E plays a major role in maintaining bone mass in men. Smith et al.71 found that a 28-year-old man who was homozygous for null mutations of the ER gene was eunachoid, had unfused epiphyses, and was severely osteopenic despite normal levels of serum T and elevated levels of serum E. Carani et al.72 and Morishima et al.73 each studied a young adult male with null homozygous mutations of the gene for the P-450 enzyme, aromatase, which is required for conversion of androgens to E. Both men had unfused epiphyses and osteopenia, and E treatment increased bone mass in both. Because all three cases were young adult men, the finding that their osteoporosis was associated with T sufficiency and E deficiency is related to their failure to achieve peak bone mass. However, it is likely that the same hormonal relationships will apply to bone in aging men.

In a relevant study in aged male rats, Vanderschueren et al.74 found no differences in the effects of orchiectomy or treatment with aromatase inhibitor on decreasing bone density, suggesting that the aromatization of androgens to estrogens was playing a major role in skeletal maintenance. However, certain geometrical features of the proximal femur were decreased to a greater degree in the rats who underwent both orchiectomy and aromatase inhibition, suggesting that both E and T have a synergistic role in establishing some features of peak bone mass in the rat.

Four recent population-based, observational studies26,75–77 involving an aggregate total of 1410 men from young adulthood to old age have demonstrated by multivariate analysis that free serum E rather than free serum T was the main predictor of bone mass at all measured sites, except at some cortical bone sites in the appendicular skeleton. Moreover, whereas serum total T and E decreased only modestly in these studies, there were large decreases in serum free T and E,26,75–77 which is consistent with the findings of other groups.68,78,79 Moreover, when serum bioavailable T and E (estradiol plus estrone) were measured, levels over life were reduced by 66% and 48%, respectively.26 These decreases were partially due to increases in the levels of serum sex hormone binding globulin, which decreased the availability of T and E to peripheral tissues. Even these findings are conservative, however, because some of the effects of circulating T on bone cells may still be mediated by E after local aromatization of T to E in target tissue. Finally, Bernecker et al.80 found that mean levels of serum E but not T were significantly reduced in 56 men with established idiopathic osteoporosis.

Collectively, these data support the hypothesis that E deficiency plays a major role in involutional bone loss in men as well as in women. However, T clearly accounts for the sexual dimorphism of the skeleton that develops following puberty and, probably also, stimulates periosteal growth of cortical bone.81 More studies must be made to define the additional effects of T on the male skeleton and to determine the relative contributions of deficiencies of E and T in causation of the slow phase of bone loss in aging men.


Limited duration of the accelerated phase of bone loss in women

In postmenopausal women, the rapid phase of bone loss as assessed by densitometry subsides after about a decade, whereas the slow phase in women and the continuous phase in men persist indefinitely. Because of the high rate of cancellous bone loss in the early accelerated phase, this bone compartment is rapidly depleted. When the amount of cancellous bone falls below some critical level, biomechanical forces act to limit the rate of further bone loss.82

Alternatively, the apparent two-component model of postmenopausal bone loss could be explained by geometric differences in surface-to-volume ratios of cancellous and cortical bone leading to different bone loss kinetics. Thus, cancellous bone loss may follow first-order kinetics (the rate of loss is proportional to the initial bone mass and decreases as the bone mass decreases), whereas cortical bone loss may follow simple linear kinetics (the rate of bone loss is constant and is independent of the initial bone mass). Computer simulations indicate that such a process would closely resemble the bone loss curves shown in Fig. 1. However, from a biologic standpoint, the consequences for the pathophysiology of bone loss would not change.

Two types of parathyroid function in postmenopausal women

According to our hypothesis, both phases of involutional bone loss in women begin at menopause. However, these processes have opposing effects on PTH secretion. During the early accelerated phase, the loss of the direct restraining effect of E on bone resorption leads to increased outflow of skeletal calcium into the extracellular fluids and compensatory decreases in PTH secretion, whereas loss of the extraskeletal effects of E result in increased whole-body losses of calcium and compensatory increases in PTH secretion. In the early years after menopause, the skeletal effects of E deficiency appear to be somewhat more important than the extraskeletal effects because E treatment during this phase leads to an increase in serum PTH levels18,23 (Fig. 2), an effect suggesting some degree of suppression of PTH secretion. As the accelerated phase of bone loss gradually subsides, the effect of E deficiency on extraskeletal calcium homeostasis becomes dominant, leading to progressive increases in PTH secretion. During this latter phase, E treatment decreases serum PTH levels.24,42 In observational studies, Koh et al.20 studied 12,238 postmenopausal women who were selected for having low hip BMD values, and Prince et al.22 studied 655 normal postmenopausal women; both found that serum intact PTH values were unchanged or decreased slightly during the initial 15–20 years after menopause and then increased progressively. Thus, the pattern of changes in PTH levels may be a more sensitive indicator of the transition between the phases of bone loss than is the pattern of changes in BMD, which suggests an earlier transitional point.

Progression of secondary hyperparathyroidism

Ledger et al.83 found that PTH secretory dynamics in elderly women were indistinguishable from those occurring in women with early chronic renal failure and documented parathyroid gland hyperplasia. This is consistent with an autopsy study that found a trend to parathyroid hyperplasia in elderly women and men.84 It is characteristic for even mildly hyperplastic parathyroid tissue, such as we postulate to be present in elderly women and men, to become increasingly less sensitive to negative feedback to changes in serum ionic calcium levels. This would lead to the observed progressive increases in PTH secretion and, thus, to progressive increases in bone turnover.

Presence of a rapid phase of bone loss in women but not in men

During the accelerated bone loss in women during the first 4–8 years after menopause, the calcium fluxes resulting from skeletal and extraskeletal effects of E deficiency are balanced, or nearly so, because serum PTH levels remain relatively constant. During this accelerated phase, the loss of cancellous bone is 3- to 5-fold greater than the loss of cortical bone. The rapid cancellous bone loss is mediated by a large increase in the number and activity of osteoclasts, resulting in perforative resorption of trabecular plates with loss of structural elements.85 Only later as the rapid phase of cancellous bone loss wanes do the extraskeletal effects of E deficiency predominate, which results in progressive secondary hyperparathyroidism and slow bone loss. Aging men have only the slow phase of bone loss associated with secondary hyperparathyroidism and do not have the early osteoclast-perforative phase that results in rapid depletion of cancellous bone. We suggest that induction of the accelerated phase in early postmenopausal women is triggered by the relatively rapid fall of serum E over a short period of only a few years to values that are about 15–20% of the mean of premenopausal women. Indeed, orchiectomy in men is associated with similar decreases in serum E and T and is associated with an accelerated phase of bone loss that is similar to that occurring in women at menopause.67 In contrast, the continuous phase of bone loss in men is associated with a gradual but protracted decrease of serum bioavailable E and T of only about 15% per decade and, in postmenopausal women after the rapid phase of bone loss subsides, the already low levels of E decrease little more.24 This more gradual decline in serum bioavailable E and T results in slow bone loss that is mediated mainly by the extraskeletal effects of sex steroid deficiency.

The presence of the accelerated transient phase of predominantly cancellous bone loss in early postmenopausal women explains the pattern of distal forearm (Colles') fracture that increases during the first 10–15 years after menopause and then plateaus. Conversely, the absence of this accelerated phase of loss in aging men may explain the low incidence of this forearm fracturing in aging men.13 At first glance, this may seem at variance with the relatively high prevalence of vertebral fractures among middle-aged men, but this may be attributed to occupational trauma rather than to osteoporosis.86


In the new unitary model, as shown schematically in Figs. 4A and 4B, E deficiency plays a central role in the pathophysiology of both phases of involutional bone loss in women and makes a major contributory role in the continuous phase of bone loss in men. At menopause in women, the acute loss of the restraining effects of E on bone cell activity leads to an accelerated phase of loss of predominantly cancellous bone that decreases after about 4–8 years and disappears after about 15–20 years when severe depletion of cancellous bone stimulates counter-regulatory forces that limit further loss. The slow phase of bone loss, which also begins at menopause, then becomes dominant. It involves loss of both cancellous and cortical bone and continues throughout the remainder of life. It is caused by the loss of E effects on extraskeletal calcium homeostasis leading to decreased intestinal calcium absorption, increased renal calcium wasting, and, perhaps also, effects on vitamin D metabolism and loss of a direct effect on the parathyroid gland that decreases PTH secretion. These extraskeletal alterations lead to a net loss of calcium from the body that increases the level of dietary calcium intake required to prevent secondary hyperparathyroidism and increased bone turnover. These manifestations can be reversed by either E replacement (which restores extraskeletal calcium fluxes to premenopausal levels) or by large increases in dietary calcium (which offsets the net calcium losses induced by postmenopausal abnormalities in extraskeletal calcium fluxes). An age-related impairment in bone formation also contributes to the slow phase of bone loss and may be caused, at least in part, by the loss of E-stimulated synthesis of bone matrix proteins by osteoblasts, although direct evidence for this at present is lacking. Because recent data suggest that bone mass in men is regulated by E as well as by T and because elderly men have low levels of serum bioavailable E, E deficiency may also make a major contribution to bone loss in aging men. The gradual induction of E deficiency in aging men leads to bone loss by mechanisms that are similar to those operative in the slow phase of bone loss in postmenopausal women.

Figure FIG. 4.

Schematic representative of unitary model for bone loss in postmenopausal women (A) and in aging men (B). See text for details.

Although this model explains much of the pathophysiological mechanisms for involutional bone loss in men and women, it does not take into account the contributions of differences among individuals in attainment of peak bone mass. However, studies comparing the effect of orchiectomy and treatment with the aromatase inhibitor in growing male rats showed comparable effects indicating that much of the effect of T on skeletal growth may have been mediated by E.87 The model also does not take into account the effects of sundry environmental and behavioral factors that increase bone loss in some, but not in other, members of the population, such as nutritional vitamin D deficiency. However, Garnero et al.15 found that the effect of decreasing levels of serum 25-hydroxyvitamin D in women accounted for only 5–8% of the variance in bone turnover in elderly French women, a population that has been demonstrated to be relatively deficient in vitamin D.12


In 1983,88 and in subsequent publications,13,89 Riggs and Melton proposed that involutional osteoporosis in women could be divided into two distinct syndromes that differed with respect to changes in regional BMD, pattern of fractures, hormonal changes, and etiology. The processes leading to the type I and II osteoporosis syndromes correspond well to the two phases of involutional bone loss in women described here, with certain modifications.

Type I osteoporosis presents during the first 15–20 years after menopause and is characterized by an excessive and disproportionate loss of cancellous bone over cortical bone, leading to acute vertebral compression fractures and distal forearm (Colles') fractures. Although there are some exceptions, most studies in patients with type I osteoporosis have found values for serum E and related hormones that are not different from those of age-matched controls.90 Although it has been widely believed that we stated that type I osteoporosis was due to only E deficiency, this is incorrect. According to the original hypothesis,88 type I osteoporosis results from E deficiency plus some additional factor, operative only in the presence of E deficiency, that produces an exaggeration of the rate and duration of the rapid postmenopausal phase of bone loss. As has been previously reported,91–93 cytokines, such as interleukin-1, interleukin-6, tumor necrosis factor, and prostaglandin E2, may act as paracrine mediators of E action in bone. In the presence of E deficiency, perhaps a genetically determined increased responsiveness of these E-stimulated paracrine effectors may predispose some postmenopausal women, but not others, to excessive cancellous bone loss and, thus, to type I osteoporosis. Genetic polymorphisms resulting in differences in the number or the function of ER could also augment the effect of E on bone cells, although this has not as yet been convincingly demonstrated. Another potential predisposing factor may be the presence of impaired renal tubular calcium transport leading to chronic renal calcium losses in women with type I osteoporosis.94 These abnormalities are not mutually exclusive and both could result from local increases of the same cytokine(s) in bone and kidney.

The process causing type II osteoporosis in women and involutional osteoporosis in men corresponds to the slow phase of age-related bone loss and involves comparable losses of both cancellous and cortical bone. This process appears to involve virtually the entire population of aging women and men and, as more and more bone is lost, those in the lower part of the age-specific normal distribution for BMD will be at the greatest risk for fracture. The major fractures associated with this process are those of the proximal femur and wedge fractures of the vertebrae, although various other fractures occur also at sites with mixtures of cancellous and cortical bone.13 The original 1986 hypothesis proposed that the slow phase of bone loss resulted from a combination of secondary hyperparathyroidism and decreased bone formation, and these remain as the main proximate causes. However, this part of the hypothesis must now be modified to recognize that both abnormalities may be manifestations of the underlying E deficiency. Thus, we believe that the type I/II model, with the modifications indicated, is still valid and should be merged with the unitary model described here.


The poet T.S. Eliot wrote “We shall not cease from exploration, and the end of all our exploring will be to arrive where we started and know the place for the first time.” Thus, if the unitary model is correct, we will have come full circle to the original 1941 view of Albright that E deficiency is virtually the sole cause of postmenopausal osteoporosis. However, the recent information about the different mechanisms by which E deficiency affects bone and calcium metabolism leads to a modern view that is much more comprehensive and complex than Albright had supposed. Thus, E deficiency accounts for the early accelerated and the slow late phases of bone loss in women and for much of the continuous phase of bone loss in men but acts at the tissue level through different mechanisms to produce its various manifestations. Although there are supporting data for all aspects of this unitary model, more experimental studies are needed, particularly to determine how much of bone loss in aging men is due to E deficiency and how much is due to T deficiency and, in both genders, to what extent E deficiency contributes to the age-related decrease in osteoblast function with aging.


This work was supported by National Institutes of Health Grants AG04875, MO-00585, and AR27065.