Prevention of postmenopausal osteoporosis and associated fractures: Clinical evaluation of the choice between estrogen and bisphosphonates

In the aftermath of the publications of the Women's Health Initiative (WHI) studies, the majority of women ceased using well-suited hormonal supplementation due to fear of the potential increase in heart attacks or breast cancer. This fear was fostered by the uncritical and poorly informed decisions of the regulatory authorities in the USA and Europe that instructed doctors to avoid estrogen-based hormonal treatments. The WHI results demonstrated unequivocally that postmenopausal hormone replacement therapy (HRT) most powerfully reduces the risks of hip and vertebral fractures. Instead, these authorities recommended declassification of HRT as a first line of prevention of osteoporosis.

Later publications emanating from the same WHI studies have documented that estrogen-only therapy in the early postmenopausal years protects against coronary heart disease and reduces the incidence of breast cancer. Herein, the attributes of estrogen-based hormonal supplementation are contrasted with those of bisphosphonates in the prevention of postmenopausal osteoporosis. HRT still offers net benefits that outweigh harm, provided that the appropriate regimen is individualized in terms of dose, route of administration, combination of hormones and adequate monitoring.

Osteoporosis, characterized by low bone mass and increased susceptibility to fracture, is the consequence of inadequate peak bone mass, excessive bone resorption, reduced bone formation or a combination thereof. The disease is also influenced by genetic, hormonal and environmental factors. Metabolic changes may influence the more sensitive trabecular bone compared with cortical bone.

At the menopause, a phase of accelerated bone loss of about 4% per annum has been recognized, which affects mainly trabecular (cancellous) bones and lasts for about 4 years [1-3]. Afterwards, the annual rate of bone loss slows back to 1%[4]. By the age of 80 years, a non-user of HRT will have lost 30–50% of her bone mass. A reduction of bone density by one standard deviation, in the hip for example, increases the risk of hip fractures by 2.6-fold [5]. Femoral neck fractures carry a 15% fatality within 3 months and, in addition to the immediate morbidity associated with fractures at any site, long-term disability has a great impact on quality of life of the sufferers.

The lifetime risk of osteoporotic fractures for a 50-year-old white man is 2% for vertebral fractures, 2% for distal radius fractures and 3% for femoral neck fractures; the corresponding figures for a woman of the same age and ethnicity are 11%, 13% and 14%, respectively [6]. Therefore, more than one-third of women will sustain one or more osteoporotic fractures postmenopause.

Regulation of bone remodeling

As a result of terrestrial life and its impact on skeleton, a bone reshaping (remodeling) mechanism has evolved to repair the microscopic damage in areas that sustain these stresses and strains. Osteoclasts, derived from monocyte precursors and bone marrow stem cells that circulate within the hematopoietic cell pool, are cytokine-driven into proliferation and differentiation. A tightly coupled cellular function involving the osteoclasts and osteoblasts ensures that, concomitant with osteoclastic bone resorption, osteoblastic bone formation is in train. Activated osteoblasts release cytokines, such as the key regulator NFκB ligand (also called receptor activator of nuclear factor-κB ligand; RANK-L) and MCSF (monocyte colony-stimulating factor), which activate their corresponding receptors on osteoclasts – RANK and c-Fms (colony-stimulating factor receptor-1), respectively. Both cytokines are essential for proliferation of precursors, differentiation and activation of osteoclasts. Osteoprotegerin, a soluble decoy receptor released by the osteoblasts, may bind RANKL, thus preventing its stimulation of osteoclasts [7]. Further, as a result of osteoclastic activation, interleukin (IL)-1β induces osteoblastic production of osteoprotegerin [8]. This cross-talk between the main two cell types is conducted through transforming growth factor-β (TGFβ) released from matrix degradation [9]. At this level of cellular cross-talk many other regulators such as parathormone, estrogen and insulin-like growth factor-1 fine-tune this remodeling function [10-12].

Other cytokines produced by T cells may participate in the upregulation of osteoclastic activity during estrogen deficiency [13]. Tumor necrosis factor (TNF) directly stimulates osteoclastic activity [14] and inhibits osteoblastogenesis [15]. TNF directly stimulates the formation of osteoclasts precursors, acts synergistically with the levels of NFκB and activator protein-1 (AP-1) signaling to increases RANL activity and synergistically with RANKL to increase the expression of RANK on osteoclasts [16]. IL-1 mediates TNF action on osteoclast precursors and osteoclast differentiation and activation. Similar to TNF, IL-1 prolongs osteoclast lifespan through a potent anti-apoptotic activity [17].

Hypophyseal–pituitary hormones

Acting through their cognate receptors and downstream from RANK-L, thyroid-stimulating hormone (TSH) and follicle-stimulating hormone (FSH) have reciprocal effects on bone resorption through the interaction with MAP kinases, NFκB and Akt kinases [18],[19]. TSH and FSH differentially regulate TNFα, an inducer of osteoclasts [20-22]. Given its pivotal importance in the survival of the terrestrial species, the control of the bone-modeling unit is governed by many other complex gene-activating mechanisms; the corollary being that only hyposufficiency-related genetic mutations or manifest disease states can result in bone disease, be that bone loss or increased bone formation conditions.

Estrogen and bone metabolism

Estrogen plays an essential role in skeletal growth and bone homeostasis in both men and women. Estrogen is established to have direct effects on bone cells, but the mechanisms involved appear to be complex and multifaceted. The duration of estrogen deficiency negatively impacts bone mass. This does not necessarily refer to the menopause status, since ‘relative’ estrogen deficiency, for example in perimenopausal smokers, may result in net bone loss and during lactation to help mobilizing calcium into milk. Therefore definition of estrogen's role in the process of bone turnover is required. Within 4–6 weeks of the loss of estrogen, increased osteoclastic activity supervenes. Increased osteoclastogenesis has been documented in terms of increased number of bone-remodeling units, increased depth of resorption surfaces as well as increased duration of resorption due to inhibition of osteoclast apoptosis. While a concomitant osteoblastogenic activity is stimulated, it falls short of matching osteoclastic activity. This may be explained in part by increased apoptosis in osteoblasts as a result of estrogen deficiency [23].

In general, estrogen action is mediated via its cognate nuclear receptors, members of the steroid receptor superfamily of nuclear receptors. Both the classical estrogen receptor ERα and ERβ are differentially expressed during osteoblast differentiation [24],[25], implying their functional roles in bone metabolism. Data to support this notion come from studies of ER-knockout mice and the single case of ERα deficiency in a man [26]. The male patient has an ERα null mutation, unfused epiphyses and severe osteoporosis, due to increased bone degradation.

Estrogen regulation of osteoblast life span

In addition to its actions through its nuclear receptors, some estrogen actions are due to binding its plasma membrane receptors and the resulting effects are accomplished within seconds or minutes. Estrogen induces apoptosis, but not complete elimination, of osteoclastic cells and inhibits apoptosis of osteoblasts; actions that are mediated through cell membrane-bound estrogen receptors and their ability to increase ERK1 and ERK2 phosphorylation and repress JNK activity [23],[27]. These cytoplasmic phosphorylated kinases are transported to the nucleus where they modulate transcription factors required for the anti-apoptotic effect of estrogen.

Numerous studies refer to the role of ovariectomy in stimulating T cells and consequent increased production of TNF, which in turn increases RANKL production by T cells [28-30]. This may in part account for accelerated bone loss following the menopause. Estrogen suppresses this clonal expansion of TNF-producing T cells, a suppressor of IL-1 and IL-6 production by osteoclasts [31],[32], and, as such, estrogen is an inhibitor of the key steps in osteoclastic cell function. Estrogen induces apoptosis in osteoclasts mainly through induction of TGFβ[33]. This activity is shared to a lesser extent by the selective estrogen-receptor modulators: tamoxifen and raloxifene, manifested by lower accrual of bone mass in randomized comparative trials [34]. Estrogen, and not progesterone, induces the synthesis of osteoprotegerin in human osteoblast-like cells, suggestive of estrogen's role in the stimulation of bone formation [35],[36].

Estrogen exerts positive effects on calcium balance as it improves its absorption, facilitates vitamin D activation and reduces renal tubular loss of calcium. Further, estrogen exerts other effects on calcium balance by induction of vitamin D receptor [37], a process that is suppressed by estrogen deficiency and re-activated by estrogen administration. Further, estrogen replacement increases the synthesis of calcitonin [38].

Bisphosphonates' effects on bone

The bisphosphonates are a family of compounds that are synthetic analogs of inorganic pyrophosphate with a high affinity for hydroxyapatite, particularly at the resorption site. Unlike natural pyrophosphates (P–O–P), simple bisphosphonates such as etidronate and clodronate contain the P–C–P configuration, which when incorporated into the ATP molecule results in ATP analogs that are resistant to metabolic degradation by endogenous phosphatases [39],[40]. As such, a number of ATP-dependent intracellular enzymes fail to function. One such target is increased mitochondrial permeability that sets the caspase-3 apoptosis pathway with consequent apoptosis of osteoclasts [41].

Nitrogen-containing bisphosphonates are more potent than simple bisphosphonates and inhibit different enzymes at different sites along the mevalonic acid pathway which is primarily concerned with sterol synthesis and the synthesis of isoprenoid lipids. In particular, the inhibition of farnesyl pyrophosphate synthase (FDD synthase) prevents the formation of isoprenoid lipids. Isoprenoid lipids are required for many synthetic functions within the cell including the post-translational modification of proteins (prenylation) necessary for their correct function by conferring a characteristic carboxy-terminal motif [42],[43]. Inhibition of protein prenylation of small GTPases disrupts many osteoclastic cellular processes such as loss of the ruffled border of osteoclasts [44], the loss of actin rings and cytoskeletal organization [45]. Nitrogen-containing bisphosphonates affect the resorptive function of ostreoclasts and apoptosis may occur as a result.

Prevention of postmenopausal osteoporosis and associated fractures

Estrogen and estrogen-based hormonal supplementation

Treatment of postmenopausal women with 17β-estradiol (E₂), alone or combined with progestogen, has been shown to prevent postmenopausal bone loss effectively [46-50], but convincing fracture data were not available. The reports of the WHI continuous combined conjugated equine estrogen/medroxyprogesterone acetate (CEE/MPA) and the CEE-alone studies have alerted clinicians to the long-term effects of progestogens, and these two studies have demonstrated unambiguously the fracture-preventing role of estrogens [51],[52]. Furthermore, long-term estrogen replacement therapy at high doses not only blunts bone resorption but also stimulates bone formation, leading to a net anabolic effect [53].

Bisphosphonates

Published placebo-controlled trials demonstrate the efficacy of bisphosphonates in increasing bone mineral density (BMD), suppressing bone resorption markers and reducing vertebral and hip fractures [54-57]. A lasting effect on bone turnover has been frequently reported after the discontinuation of bisphosphonates [58]. A plausible explanation is that bisphosphonates become sequestered in the hydroxyapatite and remain inactive until released by bone resorption during normal remodeling processes, when they exert their inhibitory effect on osteoclasts and therefore suppress bone remodeling [58],[59]. However, this interval-dosing effect is lower that that observed during active therapy.

Bisphosphonates are effective inhibitors of bone resorption, but they also inhibit bone mineralization by blocking calcium phosphate crystal growth and dissolution in vitro[60]. In vivo, bisphosphonates are capable of inhibiting bone mineralization at roughly equivalent doses to those required to inhibit osteoclasts, and subtle differences in selectivity of action of these compounds determine their net effect on bone turnover [61]. Further, concerns are raised due to pronounced inhibition of osteoclastic activity that can be detrimental to bone formation due to the coupling of these two cellular events [62-64] leading to changes in bone quality and strength [65]. This has been demonstrated further in trials of combination therapy of alendronate with teriparatide 1–34 and parathyroid hormone 1–84. The bisphosphonate blunted the anabolic effects of parathyroid hormones [66],[67] while the lesser suppressor of osteoclastic activity, raloxifene, did not impair the anabolic effects.

The choice between estrogen-based treatment and bisphosphonates

Estrogen restores bone homeostasis

The negative impact of estrogen deficiency on bone homeostasis is witnessed during the perimenopausal years, which becomes more pronounced when menstruation ceases and during the following 4 years. A rational strategy to prevent bone loss, therefore, is to restore the premenopausal endocrine milieu and to allow rebuilding of the bone collagen framework, as optimized in that particular woman in her own premenopausal environment. Estrogen replacement in postmenopausal women restores the balance between bone resorption and bone formation to premenopausal state.

In a systematic review of randomized clinical trials, estrogen replacement commenced in the perimenopausal period for an average of 6.2 years reduced incident fractures by 52% (95% confidence interval 18–64%) cost-effectively [68]. Given that the average age of the menopause is around 51 years, it follows that women at high risk of developing osteoporosis under the age of 60 are prime treatment targets with estrogen in order to maximize the impact on prevention of osteoporosis.

The adverse publicity that followed the WHI report in 2002 swayed the general prescribing trends to move rapidly from estrogen-based replacement in favor of other agents. If we were to replace estrogen-based treatment with another agent, equivalent homeostatic effect had to be demonstrated in the new agent. Where compromise had to be accepted, the new agent should be shown to have a greater safety profile.

Bisphosphonates suppress bone formation

Bisphosphonates increase BMD in hip and spine, but in parallel group comparisons the increase in HRT users was greater [54]. Lindsay and colleagues showed in postmenopausal women that alendronate significantly decreased both bone formation and bone resorption while those who received HRT continued to show stable levels of bone resorption [69]. This report demonstrated the restoration to normal premenopausal values of bone resorption and bone formation in the HRT-treated group. Other publications corroborated these findings using therapeutic doses of risedronate and alendronate compared with HRT [70],[71].

Significance of osteoclast suppression

Gene ablation studies of factors essential for osteoclasts function were found to result in osteoporotic bones, disorganized bone structure and fragility fractures [72-76]. Such experiments of impaired osteoclast function have also shown malformations of the jaw, periodontal diseases and poor teeth development [77],[78]. Animal studies showed an increase in micro-crack fractures of bone tissues, particularly of cortical bone, as a result of long-term bisphosphonate treatment and this increase was also dose-dependent [75],[79-82]. This was not observed with long-term treatment with estrogen or raloxifene in monkeys [83], further suggesting that moderate suppression of osteoclasts to premenopausal values is more protective of bone strength.

Concerns were made over delayed or absent fracture healing in patients receiving alendronate who sustained non-vertebral fractures [65]. Histomorphometry of biopsies of fracture sites showed reduced or absent osteoclastic surface and extensively suppressed bone formation. Bisphosphonate suppression of osteoclast function and its impact on bone quality and risk of subsequent fractures have been reviewed by Karsdal and associates [84].

Bisphosphonates and osteonecrosis of the jaw

High doses of bisphosphonates suppressed angiogenesis [85]. This becomes relevant with the reported increased incidence of osteonecrosis of the jaw in association with long-term use of bisphosphonates [86-92]. As many of these patients were receiving cancer chemotherapy, it suggests a possible synergism of the two modalities of drugs on the cellular components of bone remodeling, in addition to possible suppressive effect on angiogenesis.

Estrogen, bisphosphonates and bone histomorphometry

The ability of histomorphometry to predict vertebral bone strength was comparable to that of bone densitometry. Bone structure as assessed by connectivity density did not improve the correlation between static histomorphometric measures and vertebral bone strength. Neither static histomorphometry nor biomechanical testing of iliac crest bone biopsies proved to be a good predictor of vertebral bone strength [93]. However, in addition to its suppression of osteoclastic activity, estrogen is unique in its ability to regenerate bone collagen after its disintegration, an effect that has not been shown with bisphosphonates. Estrogen maintains homeostasis of the bone-remodeling unit. This observation was made possible by studies on iliac crest biopsies submitted to histomorphometric assessment in postmenopausal women receiving long-term estradiol implants [94].

There is insufficient data on deterioration in bone qualities and micro-cracks in patients on long-term bisphosphonates, probably due to inadequate numbers in the published studies. In a cross-sectional study, increased bone microdamage was observed following an average of 5 years of alendronate treatment of postmenopausal women with low bone density compared with untreated controls [95].

Adverse effects

Bisphosphonates

Treatment with bisphosphonates is recognized to cause upper gastrointestinal side-effects: esophagitis, esophageal ulcers, esophageal stricture and esophageal erosions have all been reported. In addition, other adverse effects included abdominal pain and distension, dyspepsia, regurgitation, melena, diarrhea or constipation, flatulence and musculoskeletal pain [96],[97]. Indeed, when many of these adverse events were treated with proton pump inhibitors a rise in the incidence of hip fractures was noted [98].

Estrogen

Estrogen-based hormonal supplementation may cause mastalgia, headache and dyspepsia, but it offers relief from an array of estrogen deficiency symptoms. Venous thromboembolic events with oral estrogen use have been reported (hazard ratio 2.11; 95% confidence interval 1.26–3.55) [51], but there are no such reports in parenterally administered estrogen. Increased ischemic heart disease as reported by the WHI study groups was not substantiated in younger postmenopausal women (aged 50–59 years) [99], with evidence of protection in the estrogen-only arm of the study [100]. Concerns about increased risk of breast cancer as a result of estrogen-only therapy are allayed by the results of the WHI studies [101] and the increased incidence of breast cancer with continuous combined estrogen/progestin therapy [51] questions the safety aspect of continuous progestin administration. Sequentially combined progestin to continuous estrogen therapy confers lesser risk of breast cancer than that observed with continuous combined regimens. In a recent report on data from the French E3N cohort study, micronized progesterone or dydrogesterone has not been shown to heighten the risk of breast cancer [102]. This treatment approach has to take into account the woman's acceptance of re-establishing menstruation.

Evolution of hormonal supplementation in the treatment of postmenopausal women

Most symptomatic women describe estrogen therapy as the magic pill that cures many ills. Such biofeedback has facilitated long-term adherence to an estrogen-based treatment. But two important statistical observations were made that challenged the acceptability of estrogen on its own. The estrogen-only regimen for women with intact uterus was found to increase the number of episodes of irregular bleeding, endometrial hyperplasia and the highly differentiated early stage endometrial cancer [103],[104]. Irregular bleeding in postmenopausal women using hormonal treatment is not acceptable to users, requires endometrial investigation and as such is considered costly. The early writings on the use of progestogens documented the induction of regular withdrawal bleeding with short courses of progestogens [105]. In further attempts to eliminate the possibility of endometrial hyperplasia, the duration of progestogen usage was extended to 12 days [106].

The development of endometrial hyperplasia has been unduly emphasized as a precursor of endometrial cancer. A recent study on the use of unopposed estrogen in women with intact uterus showed that 9% of women develop endometrial hyperplasia during 3 years of treatment, 90% of which were of the simple variety [107]. Endometrial cancer affected only a small minority of users, 9.6 per 1000 women per year after 10 years of use [108]. Despite the lack of a mechanism for causality, estrogen has become an accepted carcinogen for the endometrium. Concomitantly, scientists working on cancer cell lines (endometrium and breast) discovered that estrogen accelerated the growth of these cancer cell lines. An unwarranted enthusiasm for such reports assumed causality of estrogen in breast carcinogenesis and became simply an accepted fact.

It has largely been recognized that the dose and duration of sequentially added progestogens to continuous estrogen need to be adjusted to minimize the adverse effects of these steroids while maintaining an acceptable pattern of withdrawal bleeding. Endpoint measurements of the metabolic consequences of progestogen use were not sensitive enough to alert prescribers or manufacturers for that matter, who stated in a leap of faith that progestogens were metabolically benign. It followed that continuous combined estrogen and progestogen was introduced as an amenorrhea regimen, which received greater acceptance by postmenopausal women.

Numerous studies published over the last 15 years have demonstrated, in a variety of biological systems, the adverse effects of progestogens as they oppose the actions of estrogen either directly increasing estrogen catabolism or antagonizing the renewal of estrogen receptors: for example, blockage of reverse cholesterol transport [109],[110], impairment of coronary artery reactivity [111], increased invasiveness of cervical cancer [112-114] and impairment of vaginal immunity toward viruses [115-117]. While it is difficult to extrapolate these findings to clinical diseases, these studies suggest the biological plausibility for the adverse potential of progestogens, particularly when continuously administered. Indeed, many trial reports of combined estrogen and progestogens emphasized the notion that the progestogen under investigation was not shown to interfere with estrogen action.

Conclusions

In preventing postmenopausal bone loss, the most effective method is to promote a premenopausal estrogen milieu. This is particularly relevant to the first decade after the menopause. The target population is, therefore, the young postmenopausal woman with high risk of developing osteoporosis. Where estrogen was thought to be contraindicated or feared for its adverse effects on breast cancer, heart attacks and stroke, it became reasonable to seek an alternative. The literature that followed the initial WHI reports has methodically shredded the statistical analyses deployed in early reporting of the WHI studies. All adjusted relative risks were shown to be insignificant changes [118], confirmed the cardioprotective effects of HRT in early postmenopausal women [99], and indeed demonstrated the neutral effects of estrogen-only therapy on breast carcinogenesis [101]. Further, the WHI reports confirmed the long awaited fracture reduction data as a result of HRT. Given the limited, yet significant, reporting on micro-cracks and osteonecrosis of the jaw, possibly the result of profound suppression of osteoclastic function, it is timely to rethink our knee-jerk reaction in denouncing HRT. In the management of disease, holistic approaches mark the quality of the attending physician.