| |
Osteoporosis
|
| |
| |
The Lancet June 2006; 367:2010-2018
Philip Sambrooka and Cyrus Cooperb
a. Institute of Bone and Joint Research, University of Sydney, Sydney 2065, NSW, Australia
b. MRC Epidemiology Resource Centre, University of Southampton, Southampton, UK
"....the estimated number of hip fractures worldwide will rise from 1-7 million in 1990 to 6-3 million in 2050....the number of hip fractures worldwide could be as high as 8-2 million by 2050.....combined annual costs of all osteoporotic fractures have been estimated to be $20 billion in the USA and about $30 billion in the European Union..... the benchmark for diagnosis of osteoporosis has been the assessment of bone mineral density.... several risk factors independent of bone mineral density have been identified; these are, history of fracture, glucocorticoid use, family history of fracture, cigarette smoking, excessive alcohol consumption, and low bodyweight....."
Summary
Osteoporosis is a serious public health issue. The past 10 years have seen great advances in our understanding of its epidemiology, pathophysiology, and treatment, and further advances are rapidly being made. Clinical assessment will probably evolve from decisions mainly being made on the basis of bone densitometry, to use of algorithms of absolute fracture risk. Biochemical markers of bone turnover are also likely to become more widely used. Bisphosphonates will probably remain the mainstay of therapy, but improved understanding of the optimum amount of remodelling suppression and duration of therapy will be important. At the same time, other diagnostic and therapeutic approaches, including biological agents, are likely to become more widespread.
Epidemiology
Osteoporosis is a skeletal disease characterised by low bone mass and microarchitectural deterioration with a resulting increase in bone fragility and hence susceptibility to fracture.1 It is an important public health issue because of the potentially devastating results2 and high cumulative rate of fractures; in white populations, about 50% of women and 20% of men older than 50 years will have a fragility fracture in their remaining lifetime.3 Indeed, in white women, the one in six lifetime risk of hip fracture is greater than the one in nine risk of developing breast cancer.4
Fractures of the hip, vertebral body, and distal forearm have long been regarded as the typical osteoporotic fractures. However, the effect of osteoporosis on the skeleton is systemic and prospective studies have shown that there is a heightened risk of almost all types of fracture in individuals with low bone density and, irrespective of fracture site, adults who sustain a fracture are at substantially greater risk of sustaining another fracture of a different type.5 Worldwide, elderly people represent the fastest growing age-group, and the yearly number of fractures is likely to rise substantially with continued ageing of the population. Thus even if age-adjusted incidence rates for hip fracture remain stable, the estimated number of hip fractures worldwide will rise from 1-7 million in 1990 to 6-3 million in 2050. Additionally, fracture rates seem to be rising in many parts of the world. On the assumption that age-adjusted rates will rise by only 1% per year, the number of hip fractures worldwide could be as high as 8-2 million by 2050. Osteoporotic fractures also impose a major economic burden on health-care systems worldwide. In 1997, a conservative estimate of the worldwide direct and indirect annual costs of hip fracture was US$131-5 billion.6 More recently, the combined annual costs of all osteoporotic fractures have been estimated to be $20 billion in the USA and about $30 billion in the European Union.1
Hip fractures are the most devastating result of osteoporosis; they require the patient to be admitted to hospital and cause serious disability and excess mortality.2 Most hip fractures take place after a fall; 80% occur in women and 90% in people older than 50 years. The incidence of hip fracture increases exponentially with age (figure 1). There is substantial variation in hip fracture rates between populations, and hip fracture has been used as an international index of the frequency of osteoporosis. Age-adjusted rates are highest in Scandinavian and North American populations, with almost seven-fold lower rates in southern European countries. Hip fracture risk is also low in Asian and Latin American populations and rates seem to be lower in rural areas than in urban areas in any country. The striking rise in hip fracture rates with age in both men and women in most regions of the world results from both an age-related decrease in bone mass at the proximal femur and the age-related increase in falls.
The epidemiology of vertebral fractures has been clarified with the development of universally accepted definitions and large population-based radiographic studies.7 Geographical variation in the prevalence and incidence of vertebral fractures seems to be substantially less than that for hip fracture. Only about a third of all radiographically identified vertebral deformities come to specialist attention and less than 10% result in hospital admission. In cases that are identified medically, there is substantial disability from pain and increased thoracic kyphosis. Only about a quarter of vertebral fractures result from falls,8 and most result from routine activities such as bending or lifting light objects. The prevalence of vertebral fractures is almost as high in men as in women, which is thought to be due to occupation-associated trauma in men.
Wrist fractures have a different pattern of occurrence from hip and vertebral fractures (figure 1). There is an increase in incidence in white women between the ages of 45 and 60 years, followed by a plateau or more attenuated rise thereafter. Most wrist fractures happen in women, 50% of whom are older than 65 years. The incidence in men is low and does not increase much with age.
Patients with any type of fragility fracture are at increased risk of other types of fracture.9,10 A vertebral fracture leads to a ten-fold increase in risk of subsequent vertebral fractures.11 Hip and forearm fractures predict a similar rise in risk of future fracture at the same site. There is, however, some attenuation in the risk of future fractures at distant sites from the index fracture; patients with a history of vertebral fracture have a 2-3-fold increased risk of future hip fracture and 1-4-fold increase in risk of distal forearm fracture.
The adverse outcomes of osteoporotic fracture fall into three broad categories: mortality, morbidity, and cost. The effect of fractures on survival is dependent on fracture type. Hip fractures are the most serious, with 10-20% excess mortality in the first year after fracture.12 The risk of death is greatest in the first 6 months after the fracture and decreases over time. However, few of these deaths are directly attributable to hip fracture; most result from chronic illnesses that lead to both fracture and early death. Acute events such as infections and postoperative complications are also important. The excess mortality associated with vertebral fractures extends well beyond the first year. Again, reduced survival is difficult to attribute to direct effects of the fracture and increased mortality seems to be a result of comorbidity.
Pathophysiology
Osteoporotic fractures result from a combination of reduced bone strength and increased rate of falls. Although bone mineral density remains the best available non-invasive assessment of bone strength in routine clinical practice, many other skeletal characteristics also contribute to bone strength. These include bone macroarchitecture (shape and geometry), bone microarchitecture (both trabecular and cortical), matrix and mineral composition, as well as the degree of mineralisation, microdamage accumulation, and the rate of bone turnover, which can affect the structural and material properties of bone.13,14 The recognition of these other measures (often referred to as bone quality) is becoming more important, and their incorporation into algorithms of fracture detection remains the subject of continuing translational research.15 The bone mass of an individual in later life is a result of the peak bone mass accrued during intrauterine life,16 childhood, and puberty, as well as the subsequent rate of bone loss. Although genetic factors strongly contribute to peak bone mass,17 environmental factors in intrauterine life, childhood, and adolescence modulate the genetically determined pattern of skeletal growth.18 Moreover, although various candidate genes (most notably those encoding the vitamin-D receptor, collagen I_1, LDL receptor-related protein 5 [LRP5], and oestrogen receptor) have been linked to bone mineral density, attempts to relate polymorphisms in such genes to fracture risk have generally been unsuccessful, apart from the Sp1-binding-site polymorphism in the collagen I_1 gene.19-21
Bone loss takes place as a result of oestrogen deficiency in postmenopausal women, as well as through oestrogen-independent age-related mechanisms (such as secondary hyperparathyroidism and reduced mechanical loading). At the cellular level, bone loss occurs because of an imbalance between the activity of osteoclasts and osteoblasts. In adult life, the skeleton is continually remodelled in an orderly sequence of bone resorption followed by bone formation-referred to as coupling. If the processes of resorption and formation are not matched, there is a remodelling imbalance; this imbalance can be magnified by a rise in the rate of initiation of new bone remodelling cycles (activation frequency).
Oestrogen has a central role in normal physiological remodelling, and oestrogen deficiency after the menopause results in a remodelling imbalance with a substantial increase in bone turnover. This imbalance leads to a progressive loss of trabecular bone, partly because of increased osteoclastogenesis. Enhanced formation of functional osteoclasts seems to be the result of increased elaboration of osteoclastogenic proinflammatory cytokines such as interleukin-1 and tumour necrosis factor, which are negatively regulated by oestrogen22-24 (figure 2). A direct effect of oestrogen in accelerating osteoclast apoptosis has also been attributed to increased production of transforming growth factor _.25
Understanding of the cellular basis of remodelling has advanced rapidly lately. The receptor activator of NFƒÈB (RANK), its ligand (RANKL), and the decoy receptor osteoprotegerin are now known to be key regulators of osteoclastic bone resorption in vitro and in vivo.26 Osteoblasts express RANKL constitutively on their cell surface; RANKL interacts with its cognate receptor, RANK, which is expressed on osteoclast precursors and promotes osteoclast differentiation. Interaction of RANKL with RANK on mature osteoclasts results in their activation and extended survival (figure 2). Osteoprotegerin, present in the bone microenvironment, is mainly secreted by osteoblasts and stromal cells; in vivo, osteoprotegerin blocks the interaction of RANKL with RANK and thus acts as a physiological regulator of bone turnover. Oestrogen might also exert part of its antiresorptive effects on bone by stimulating osteoprotegerin expression in osteoblasts.27
Important novel genes and pathways for osteoblast differentiation and function have been discovered. LRP5 is a modulator of osteoblast function and hence bone formation. It is a coreceptor for a series of osteoblast-stimulating proteins operating through the Wnt signalling pathway. LRP5 is expressed on the osteoblast membrane between two other receptors, Frizzled and Kremen. Frizzled and LRP5 bind to Wnt, thereby activating bone formation. Wnt inhibitors such as the Dickkopf protein (Dkk) bind to Kremen and LRP5, causing LRP5 to be internalised and therefore be unable to bind Wnt, leading to inhibition of bone formation. Polymorphisms in the LRP5 gene have been described in rare disorders including gain-of-function mutations resulting in high bone mass phenotypes28,29 and loss-of-function mutations resulting in osteoporosis-pseudoglioma syndrome.30 Polymorphisms of LRP5 have been associated with differences in bone density and fractures.31,32
Several inhibitors interact with the Wnt signalling pathway. One of these, sclerostin, the product of the SOST gene, inhibits Wnt signalling.33 Inactivating mutations in the SOST gene lead to hyperactivation of Wnt signalling, which mediates the bone overgrowth seen in sclerosteosis or Van Buchem disease34,35 and polymorphisms in the SOST gene have been associated with bone mineral density in elderly people.36
Assessment of fracture risk
Since 1994, the benchmark for diagnosis of osteoporosis has been the assessment of bone mineral density. The ability to predict fracture risk from this measure is at least as good as if not better than the ability to predict heart disease risk from blood cholesterol concentrations and to predict stroke risk from blood pressure values.37 However, low bone mineral density alone does not mean an individual will have a fracture, and although the widely accepted diagnostic threshold of a T score less than -2-5 seems useful on the basis of results from many clinical trials, the clinical use of the associated term osteopenia (T score between -1 and -2-5) is less clear and quite broad; a T score of -1-1 and -2-4 are both osteopenic yet they confer vastly different fracture risks.
Although the fact that the risk of future fracture rises steeply with declining bone mineral density is well established, the view is growing that assessment of fracture risk should encompass all aspects of risk and that intervention should not be guided solely by bone mineral density. Therefore, WHO, analysed all international cohort studies in which information on clinical risk factors and bone mineral density are available and incident fractures have been ascertained.38 On the basis of this information, several risk factors independent of bone mineral density have been identified; these are, history of fracture, glucocorticoid use, family history of fracture, cigarette smoking, excessive alcohol consumption, and low bodyweight. The combined use of these risk factors together with age and bone mineral density in multivariate models allows the 10-year probability of hip and other fractures to be predicted. Thus a woman at the age of 60 years with an average bone mineral density (about T -1-4) has an average 10-year probability of hip fracture of around 2-4%. If she has a history of fragility fracture, this risk is increased two-fold to 4-8%. Intervention thresholds for various agents to combat osteoporosis can be derived with use of health economic modelling. Although population screening strategies have not been validated, a case-finding approach based on the fracture assessment probability with clinical risk factors and, where appropriate, additional testing of bone mineral density can be useful. The widespread use of such algorithms of risk prediction will enhance the ability to target treatments to clinical situations in which they are most cost effective.
Several serum and urine biochemical markers of bone turnover have been developed. These provide non-invasive and fairly inexpensive methods for assessing rates of bone formation and resorption in vivo. As with any new technology, however, the precise positioning of these biochemical markers in the clinical approach to osteoporosis management has not been established. The most widely available markers include serum bone-specific alkaline phosphatase and the amino-terminal propeptide of type 1 procollagen, which are markers of bone formation, and urine or serum telopeptides of collagen crosslinks, which are markers of bone resorption. Disadvantages of biochemical markers include their indication of whole-body bone turnover (mainly cortical bone) and day-to-day variability, although this variation is no greater than that for circulating markers of other chronic diseases. Despite these drawbacks, there are several ways in which biochemical markers might help in osteoporosis. They may be used to: (1) study the pathogenesis of osteoporosis, (2) predict the risk of future fracture (independently of bone loss), and (3) predict and monitor the response to therapy. Prospective studies have shown an association between osteoporotic fracture and indices of bone turnover independent of bone mineral density in women during the menopause and in elderly women.39 In elderly women with values for resorption markers exceeding the reference range for premenopausal women, fracture risk is increased by about two-fold after adjustment for bone mineral density. These findings suggest that a combined approach using bone mineral density, clinical risk factors, and markers of bone turnover could improve fracture prediction.40
Management
Since most fractures happen as a result of falls, attention to reducing the risk of falls seems important. Although no studies are available that show that strategies to reduce the rate of falls will reduce fractures, the use of hip protectors to reduce the impact of falls has proven effective in high-risk individuals,41 although compliance remains an issue.42 At a mechanistic level, drugs can be considered in terms of whether they act mainly on bone resorption (antiresorptive agents) or on bone formation (anabolic agents). With this classification, antiresorptive treatments include calcium, vitamin D, hormone therapy, bisphosphonates, selective oestrogen-receptor modulators, and calcitonin.
Antiresorptives
Calcium and vitamin D play distinct parts in bone physiology, yet many trials have examined different combinations of these two treatments and the relative effect of each is not clear. Calcium supplementation alone provides small beneficial effects on bone mineral density throughout postmenopausal life and might slightly reduce fracture rates.43 Low vitamin D status has been associated with reduced bone mineral density, high bone turnover, and increased risk of falls and of hip fracture in elderly people, but differences in baseline vitamin D concentrations make comparisons between trials difficult.
The first appropriately powered fracture trial examined a high risk group (3270 elderly women in residential care) treated with a combination of calcium and vitamin D supplementation (800 IU of cholecalciferol) for 18 months and showed a 43% reduction in hip fractures and 32% in total non-vertebral fractures. 44 A subsequent trial in 2686 people living in the community examined oral cholecalciferol 100 000 IU given every 4 months for 5 years and showed a decrease in the risk of first hip, wrist, forearm, or vertebral fracture by 33% compared with placebo.45 In the RECORD trial,46 5292 ambulatory patients who had sustained a recent low-trauma fracture were randomised to either calcium alone, vitamin D (800 IU) alone, combination therapy, or placebo. There were no significant differences in fracture rates between the four groups, but interpretation of the study is limited by the low compliance rate and absence of information about baseline vitamin D status. A meta-analysis concluded that vitamin D reduced the risk of hip fracture by 26% and non-vertebral fracture by 23% in a dose-dependent manner in individuals with vitamin D deficiency.47 There is controversy about whether the amount of vitamin D needed for bone health is 50 or 80 nmol/L and in an individual with a baseline serum 25 hydroxy vitamin D concentration of less than 50 nmol/L, doses higher than 1000 IU vitamin D per day could be needed to reach 80 nmol/L.
The Women's Health Initiative reported its findings in the calcium plus vitamin D group consisting of 36 282 postmenopausal women.48 Bone mineral density increased significantly at the hip (1-06% during 9 years) but not at other skeletal sites in patients given 1000 mg calcium plus 400 IU vitamin D compared with those given placebo. In the intention-to-treat analysis, the reduction in hip fracture with calcium plus vitamin D was not significant (hazard ratio 0-88, 95% CI 0-72-1-10). However, the power of the study was reduced from a predicted 85% to 48% by a lower than projected hip-fracture rate and the reduction in hip-fracture risk was significant in the adherent population (0-71, 0-52-0-97). The risk of renal calculi was also increased with calcium plus vitamin D (1-17, 1-02-1-34); however, the baseline calcium intake of the study population was greater (mean intake 1150 mg per day) than in most osteoporotic individuals.
Taken together, these studies suggest that calcium alone or in combination with vitamin D has only a modest effect on fracture risk and that vitamin D is most likely to be effective in deficient individuals. Although most clinical trials of drugs for osteoporosis have used calcium and vitamin D as placebo therapy, whether either agent alone or in combination has an adjunctive role-ie, whether they are needed for the full antifracture effect of any particular osteoporosis treatment-or further enhance the activity of that therapy, remains unclear. The Women's Health Initiative showed a possible synergistic effect of calcium plus vitamin D with oestrogen therapy, although the effect was not significant. Active formulations of vitamin D such as calcitriol and alfacalcidol have also been studied, but whether they confer additional benefit to standard vitamin D is controversial.
The role of long-term postmenopausal hormone therapy in the prevention and management of osteoporosis is controversial after publication of the Women's Health Initiative study of combined oestrogen and progestagen therapy49 and its sister study of oestrogen alone.50 The study population consisted of women aged 50-79 years, many of whom had cardiovascular risk factors. Women were not selected to have low bone mineral density, unlike in most osteoporosis trials. Nevertheless, both trials showed substantial reductions in subsequent osteoporotic fractures. There were many endpoints, and controversy exists about the appropriate adjusted CIs. Overall, between ages 50 and 79 years, the increase in stroke was eight per 10 000 person-years, although the absolute risk was low in women in their 50s and rose with age. A raised risk of cardiovascular events was shown with combined hormone therapy, especially in those starting treatment when older than 70 years.
In the combined hormone therapy study, there was an increase in breast cancer by 5 years of eight per 10 000 person years (ie <0-1%). This increase was matched by a similar reduction in other major cancers, and there were no changes in overall cancer or mortality rates.49 The oestrogen alone study was stopped after 6-8 years because of stroke events but was by then showing a reduction (p=0-06) in breast cancer of seven per 10 000 person years.50 These data suggest a different risk profile for opposed therapy compared with an unopposed oestrogen and for older versus younger women.51 From the Women's Health Initiative results, one can conclude that, in women with osteoporosis and cardiovascular risk factors, hormone therapy should be avoided in favour of alternative antiresorptive agents, and that hormone therapy remains an option only for short-term early use around the menopause in symptomatic women with high risk of fracture.
Bisphosphonates represent the biggest advance in the treatment of osteoporosis in the past decade, with results of clinical trials showing reductions in the risk of vertebral fractures (40-50%) and non-vertebral fractures (20-40%), including hip fractures.52 Several bisphosphonates have reduced vertebral fractures. Daily alendronate and risedronate have reduced the risk of single and multiple spine fractures, asymptomatic (morphometric) and symptomatic spine fractures in women with bone mineral density T scores of less than -2-5 and one or more prevalent spine fractures.53-56 Daily ibandronate can reduce the risk of vertebral fractures in women with low bone mineral density and one or more baseline fractures.57 Etidronate has also been shown to prevent vertebral fractures, but problems in design and analyses make the results of these trials unclear.58,59
Non-vertebral fracture reduction with bisphosphonates has shown more variable results. For example, overall non-vertebral fracture rates have been substantially reduced in some studies with alendronate and risedronate,53,55,60 but not in others,54,56,61 apart from in post-hoc subgroups, such as individuals with baseline T scores of less than -2-5.61 Non-vertebral fracture rates were not reduced with ibandronate in the overall study population, but were reduced in a post-hoc analysis in people with a baseline T score less than -3-0. Ibandronate has been specifically approved only for the prevention of vertebral fractures. In the sole trial powered for hip fractures as the primary endpoint,60 an overall reduction by 30% was seen in hip fracture risk in women treated with risedronate. However, in women aged older than 80 years, selected mainly on the basis of fall related risk factors but not low bone mineral density, there was no significant reduction.
Because the dosing regimen (which required the patients to fast and remain upright for at least 30 min) and upper gastrointestinal side-effects were often limiting factors in daily bisphosphonate therapy, weekly formulations were developed. Although the bone mineral density response and the suppression of bone turnover with once weekly alendronate or risedronate is no different from that with the daily formulations,62,63 there are no fracture studies with these weekly formulations. However, in the ibandronate phase III trial, there was a significant reduction in vertebral fracture rates with an intermittent, non-daily regimen.57 Bisphosphonates and other antiresorptives are often compared in trials using bone mineral density or biomarker endpoints,64-66 but conclusions about the relative efficacy of different agents is not possible without trials using fracture endpoints. Such trials are unlikely in view of the enormous sample size that would be needed to show a difference between agents.67
Despite their impressive antifracture efficacy, several issues are now arising with respect to bisphosphonates. These drugs remain in the skeleton for decades and their duration of physiological effect is unclear, but bone turnover markers can remain suppressed for at least 5 years after their discontinuation.68 Bone biopsy studies show that bone-forming surfaces are suppressed by 40-80% with usual doses of the bisphosphonates69 and an unresolved question remains as to whether such potent inhibition of bone turnover can be harmful over time. The biological purpose of bone remodelling is to remove microdamage and replace it with new bone. As bone mineralisation increases, bone can become brittle. Moreover, if bone resorption is strongly inhibited, microdamage might not be repaired and damage could accumulate. Benefits of bisphosphonates on fracture endpoints are proven by randomised controlled trials only for the first 4-5 years of treatment,61,70 and the optimum duration of therapy remains unclear. Observational, uncontrolled follow-up to 10 years in women from the original phase III alendronate trials has shown no suggestion of negative effects, with the fracture rate being similar during years 6-10 to years 1-3.71 In these circumstances, the skeletal affinity of the particular bisphosphonate used,72 the patient's age, the pre-existing fracture risk, and the bone mineral density achieved with treatment should be considered when deciding on the duration of treatment.
There has also been growing recognition that the effects of bisphosphonates on markers of bone turnover account for a large part of their antifracture action. With the bisphosphonate risedronate, vertebral fracture reduction seems to plateau when markers are suppressed to about 40-50% of postmenopausal levels.73 Whether greater suppression with other bisphosphonates will result in further risk reduction or, indeed, what the optimum amount of suppression of bone turnover is with bisphosphonate therapy remains unclear.
Long-term adherence to treatment is needed for optimum fracture reduction in patients with osteoporosis;74 however, even with weekly bisphosphonate regimens, compliance remains disappointing.75 Less frequent, more potent bisphosphonate regimens such as once monthly76 or once yearly intravenous administration77 could improve convenience for patients and potentially help with compliance. There have been reports of the rare but serious disorder of osteonecrosis of the jaw associated with bisphosphonate use, mainly in patients treated with high doses and often several bisphosphonates for cancer.78,79 The actual nature of the pathophysiology of osteonecrosis of the jaw and the role of bisphosphonates needs further investigation.
Selective oestrogen-receptor modulators represent a chemically diverse set of compounds that do not have the steroid structure of oestrogen, but have a tertiary structure that allows binding to the oestrogen receptor to exert selective agonist or antagonist effects on different oestrogen target tissues. Those that have been or are being studied for effects on osteoporosis, breast cancer, and cardiovascular disease include raloxifene, arzoxifene, and lasofoxifene. The most studied is raloxifene, and its effects on markers of bone turnover have generally been less (eg, 30-40% reduction) than with bisphosphonate therapy.64,65 Similarly, the response in terms of bone mineral density is smaller than with bisphosphonates, with increases averaging between 2 and 3% at different skeletal sites over 3 years.80
In the phase III study of 7705 postmenopausal women, raloxifene 60 mg daily significantly reduced the rates of vertebral fractures by almost 50% in individuals without previous fractures and 34% in women with previous (prevalent) vertebral fractures.81,82 This effect on vertebral fractures is similar to that with bisphosphonates.83 However, there was no significant reduction in the rate of non-vertebral fractures or hip fractures, apart from in the subgroup with severe vertebral fractures (>40% loss of vertebral height) in post-hoc analyses.84 One explanation for this apparent absence of non-vertebral effect is that raloxifene's more slight antiresorptive effects can return high bone turnover to normal and prevent microarchitectural deterioration in trabecular bone, but reduction fracture risk at sites of cortical bone such as the hip, needs more potent antiresorptive effects.85 In an extension of the phase III study for an additional 4 years in 4011 women, non-vertebral fracture was a secondary endpoint, but again, there was no significant difference in overall non-vertebral fracture rates.86 In the absence of unequivocal proof of a non-vertebral fracture effect, raloxifene should probably mainly be used in postmenopausal women with milder osteoporosis or in those with predominantly spinal osteoporosis. Potential side-effects include an increased risk of venous thrombosis similar to that with hormone therapy, and exacerbation of hot flushes.87 Although only approved for the treatment of osteoporosis, in an individual, raloxifene's potential to prevent the development of breast cancer might improve its risk benefit ratio. A large phase III trial with calcitonin produced equivocal results,88 and was affected by a large number of patients lost to follow-up and the trial design of only partial blinding, so the role of calcitonin remains unclear.89
Anabolic agents
Unlike the previously mentioned therapies that act mainly to reduce bone resorption, the first clearly anabolic therapy that stimulates bone formation is parathyroid hormone.90 Clinical trials have been done with intact parathyroid hormone (hPTH 1-84) and with the 34 aminoacid peptide (hPTH 1-34) now called teriparatide. Results from a phase III trial with teriparatide in 1637 postmenopausal women showed a 65% reduction in the risk of new vertebral fractures and a 53% reduction for non-vertebral fracture with the 20 _g dose.91 However, the benefit in terms of bone mineral density seemed to wane after discontinuation unless followed by an antiresorptive agent. What does this mean in terms of its effects on fracture? Because of safety concerns, therapeutic courses are limited to 24 months and cost also remains an issue. Alternating 3 monthly cycles of teriparatide might dissociate the stimulation of bone formation from activation of bone resorption and improve the anabolic effect of parathyroid hormone in patients receiving bisphosphonates,92 but this possibility needs confirmation. A clinical trial with hPTH 1-84 showed a 68% reduction in the rate of new vertebral fractures in patients without a prevalent vertebral fracture.93
There is also evidence that the response to parathyroid hormone might be modified by previous treatment with antiresorptives. For example, previous treatment with alendronate resulted in an attenuated bone mineral density response with teriparatide, whereas in those treated with previous raloxifene, the bone mineral density response was no different than in treatment-naive patients.94 Similarly, treatment with hPTH 1-84 alone increased bone mineral density to a greater extent than the combination of hPTH 1-84 and alendronate.95 However, none of these studies have addressed determinants of bone strength other than bone mineral density and the effectiveness of combined parathyroid hormone and an antiresorptive agent in reducing fractures remains unclear.
New agents
Strontium ranelate is a fairly new antiosteoporotic agent that has been approved in the European Union for the treatment of postmenopausal osteoporosis. Although postulated to increase bone formation while reducing bone resorption, its mechanism of action remains unclear. Results from a phase III clinical trial96 in 1649 postmenopausal women showed that strontium ranelate reduced the risk of new vertebral fractures by 49% after 1 year and by 41% over 3 years. Clinical vertebral fracture risk was also reduced by 52% in the first year and in the longer term. Results from a second large-scale phase III clinical trial showed that strontium ranelate slightly reduced the risk of non-vertebral fractures in postmenopausal women (overall effect 16%), and the risk of hip fracture in a subgroup of women aged 74 years or older with low bone mineral density (T <-3-0).97 In both these trials, strontium ranelate was well tolerated apart from a low rate of gastrointestinal side-effects and an unexplained increased risk of venous thrombosis. In view of its effectiveness against vertebral and non-spine fractures, strontium ranelate could be an important alternative to bisphosphonates in the treatment of postmenopausal osteoporosis.
The future
Appreciation of the RANKL mechanism linking osteoblast and osteoclast function has opened the door to new biological therapies in the treatment of osteoporosis, for example the use of antibodies to RANKL. AMG 162 is a fully human monoclonal antibody that specifically binds with high affinity to RANKL and prevents it from binding to its receptor, so inhibiting osteoclast differentiation, activation, and survival. This effect is quite long lasting and, in a phase-II dose-ranging trial, twice yearly injections significantly increased bone mineral density at the total hip to a similar or greater extent than that seen with alendronate.98 A phase-III fracture trial is now in progress.
Other biological targets that offer promise include cathepsin K, Dickkopf, and sclerostin. Cathepsin K is a tissue-specific cysteine protease that plays a part in the degradation of protein components of the bone matrix in bone resorption. Phase II studies of cathepsin K inhibitors are now underway. Although most existing treatments target osteoclast function, the functional role of LRP5 in osteoblast biology makes it a potential target for anabolic therapy via Dickkopf inhibition. Similarly, inhibitors of the SOST gene product, sclerostin, might also be a therapeutic target.
Search strategy and selection criteria
The information in this Seminar is based on MEDLINE and PubMed searches with the search terms "osteoporosis" or "fracture" in combination with the keywords "calcium", "vitamin D", "bisphosphonates", "selective estrogen receptor modulator", "parathyroid hormone", "strontium", "RANKL", "Receptors-LDL" or terms such as "randomized trials" or "meta-analyses". We mainly selected papers from the past 5 years, but also included frequently referenced and highly regarded older papers. Some review articles or book chapters were included because they provide comprehensive overviews that are beyond the scope of this Seminar.
Conflict of interest statement
Philip Sambrook receives honoraria and speakers' fees including reimbursement of travel and accommodation expenses from Merck, Sanofi-Aventis, Servier, Roche, Eli Lilly, and Bone. Research funding is provided by the Australian National Health and Medical Research Council and Arthritis Australia, as well as companies for individual clinical trials. Cyrus Cooper has received honoraria and speakers fees, including reimbursement of travel and accommodation expenses, from Procter & Gamble Pharmaceuticals, Servier, GSK-Roche, Eli Lilly, and Novartis. Research funding is provided by the Medical Research Council, Arthritis Research Campaign, and National Osteoporosis Society.
|
|
| |
| |
|
|
|
