BoneKEy-Osteovision | Meeting Reports
Meeting report from the IOF world congress on osteoporosis
DOI:10.1138/2002051
Pathophysiology of osteoporosis Gideon A. Rodan, Department of Bone Biology & Osteoporosis, Merck Research LaboratoriesOsteoporosis is a multifactorial disease characterized by a reduction in bone mass and a deterioration in microarchitecture, leading to an increase in fracture risk. The susceptibility to fracture is determined by the amount of bone, the bone material properties, bone micro- and macroarchitecture, and exogenous trauma. Information pertaining to the pathophysiology of each of these factors was presented at the meeting. Each of them, with the possible exception of falls, is probably influenced by genetics (), covered in a separate report.
Previous cross-sectional and prospective studies of the natural history of osteoporotic fractures have shown that fracture incidence in women correlates with bone mineral density (BMD). Comparison of BMD in Rotterdam men and women with hip fractures indicates that both genders have a similar fracture risk at the same absolute BMD (). In 7,070 early postmenopausal women (mean age 55) from Sweden, previous fracture combined with forearm BMD was a strong predictor of new fractures, increasing the risk by a factor of 2.1 (). The mechanism for the contribution of previous fractures to fracture risk has not been elucidated.
The contributions to fracture reduction of changes in BMD and bone turnover produced by antiresorptive therapy were further analyzed and discussed, but mechanistic insights are still lacking. A metaanalysis (54,615 patient years) of the risk of non-vertebral fractures (), showed that for 10% decrease in resorption markers there was a 6% decrease in fracture risk (FR) and for 10% decrease in formation markers there was a 13% decrease in FR.
The role of bone structure in FR was a major theme () at the meeting. An analysis of hip fractures in men and women showed that both fractured at similar "buckling ratio thresholds", resulting however from different geometries, larger diameter and thinner cortices in women and smaller femoral neck width in men (). Differences were attributed to sex steroid effects (see below). Mechanical testing of the strength of femoral neck in pathological specimens of osteoporotic bones showed that strength is determined primarily by cortical bone (). An ex vivo biomechanical analysis of 120 specimens () showed that cervical BMD gave the highest correlation for both side impact (r = 0.73) and for vertical loading (r = 0.72). Lower correlations were found with DXA of the spine (r=0.64) and radius (r=0.61). Interestingly, when areal BMD and geometric properties of the femoral neck were evaluated, nulliparous women had a weaker hip bone structure than parous women (), further implicating a role for sex hormones. In 152 bone biopsies in men with low BMD (-2.5< t <-5), hypogonadism, low BMI, smoking and chronic disease correlated with microstructural deterioration ().
Do mechanical stimuli influence BMD? A study on young adults (52 men, 69 women, mean age 23.3) and older adults (81 men, 81 women, mean age 63.2) reported a significant correlation between total bone area and maximum strength of the hand in young adults and between hand strength and trabecular BMD in older women (). A separate study on 218 postmenopausal women, age 41 to 83, 138 of whom were osteoporotic (t <-2.5) showed a positive correlation between hand grip strength and femoral neck BMD (). However, in the same patients there was a negative correlation between handgrip strength and age, as well as menopause duration, suggesting that changes in muscle strength and BMD may not be causally related. In another study on 1,044 women 75 years old, the strongest determinant of BMD was gain in body weight between ages 50 and 75 (0.4-0.65 t score difference/10 kg), with very minor contributions of physical activity, isometric muscle strength or lean mass (). Two year results of a 3 year high impact exercise study in 137 postmenopausal osteoporotic women (-4<t<-2.5, spine or total hip) reports a 0.7% increase in L1-L4 BMD vs. 2.3% loss in control women. Significant differences were also observed in BMD by QCT and in isometric flexor force (). As previously reported, it seems to be difficult to increase bone mass in osteoporotic women, if at all, by mechanical stimuli.
Almost all non-vertebral fractures occur during falls (). Pathophysiological factors that contribute to falls include (): muscle strength in lower extremities, balance, gait disorders, vision, polypharmacy (drugs) and cognitive impairment. A positive correlation was found between muscle strength and 1,25(OH)2D3 levels in the elderly. In the European prospective osteoporosis study (EPOS) on 2,510 men and 3,034 women aged 50 to 80 in 22 centers, 4,136 of whom had BMD measurements, the risk of falls and personal or parental fracture history were a strong predictor of fractures ().
Peak bone mass is probably determined by multiple genes but among environmental factors, dietary calcium and vitamin D, as well as immobilization, were again reported to play a role. In 100 Greek healthy men, age 19-22, radial BMC and BMD were lower in those ingesting less than 400 mg Ca per day and those immobilized for more than one month (). In 133 male and 279 female first grade students in Kyoto, Japan, BMD was higher in girls with a dietary history of milk and yogurt and boys with greater intake of milk (). In another study in adults, BMD was also positively correlated with calcium intake (). In Sweden, a survey of 26 million observation years found higher fracture incidence and seasonal variations in fractures at northern latitudes, that may be related to differences in vitamin D status (). Seasonal variations were also observed in forearm BMD evaluated in 10,364 women age 49-69, in Goteborg and Varberg, Sweden, which may also be related to seasonal variations in vitamin D. The previous report on reduced hip fracture incidence in institutionalized women receiving calcium and vitamin D was confirmed in 583 ambulatory institutionalized French women, mean age 85.2, and was associated with decreased circulating PTH. Femoral neck BMD, measured in 114 of these patients, increased relative to controls by 2.65% ().
Is there any evidence for changes in bone material properties in osteoporosis? Few measurements have been conducted and there is no compelling evidence so far. In osteogenesis imperfecta, where bone fragility is increased, there is a significant reduction in trabecular bone density and an increase in bone turnover (), as observed in other osteoporoses.
The rate of bone loss has a large impact on BMD levels in the elderly and elevated bone turnover, usually associated with bone loss, is a pathophysiological risk factor on its own. The pathophysiology of bone loss thus deserves special attention. Epidemiologically, the major cause of bone loss is estrogen deficiency (). A study on 40 postmenopausal women using newly developed ultrasensitive assays concluded that osteoporotic women are much more sensitive to estrogen deficiency compared to controls and have higher circulating osteoprotogerin (OPG) levels (). Higher OPG levels were also observed in 177 Austrian osteoporotic women, where a strong correlation with the bone turnover marker SCTX (C-terminal crosslinked telopeptide of type I collagen) was observed (). These data are explained as an unsuccessful compensatory mechanism. In men, estradiol is also the hormone that suppresses bone resorption, while testosterone acts mainly as a stimulator of bone formation ().
Estrogen (E2), the aromatization product of testosterone (T), and aromatase gene polymorphism are associated with bone loss in elderly men (). The ratio between T and E2, an indirect measure for aromatase activity, was higher in normal than in osteoporotic men, further implicating E2 in bone metabolism in elderly men (). No new mechanistic insights into the mode of action of E2 on the skeleton were presented.
In addition to sex steroids, other prominent steroids responsible for a large proportion of iatrogenic osteoporosis are the glucocorticoids. Given continuously or intermittently, orally or by inhalation, corticosteroids (CS) increase significantly the risk for osteoporosis (). CS treatment seems to increase the risk of fracture, estimated in a metaanalysis of 5,704 men and 12,253 women aged 21 to 103, independent of its effect on BMD (), through a mechanism that has not been elucidated. Interestingly, bone loss measured in 25 patients with pituitary Cushing's disease and nine patients with adrenal Cushing syndrome is more severe in the latter, presumably due to lower androgen levels (). GCs also contribute to the low bone density and fracture risk in rheumatoid arthritis (), but RA is a separate risk factor for osteoporosis. Interesting, a single infusion of the TNF antibody inflixinab reduced bone resorption (). TNF has also been implicated in the bone loss caused by under nutrition in rodents, where transgenic mice expressing the soluble tumor necrosis factor receptor 1, which blocks TNF action, were resistant to hyponutrition, that caused bone loss and strength reduction ().
Systemic lupus erythematosis (SLE) is associated with low BMD in 30% of pre-menopausal and 36% of post-menopausal patients (). The low BMD in SLE patients was confirmed in a separate study on 16 females, aged 6-17, and there was no correlation to disease activity or glucocorticoid use in these patients (). Low radius BMD was also observed in patients with ankylosing spondylitis (). Other conditions of compromised health, such as alcoholic cirrhosis (), Huntington's disease () and eating disorders () were also reported to correlate with low BMD. The detailed pathophysiological reasons remain to be elucidated.
In summary, data presented at the WCO provided further support for BMD and bone turnover playing an important role in the risk for osteoporotic fractures. Data were presented on the role of bone structure in femoral neck fractures and the putative role of sex steroids in determining bone architecture in males and females. Additional data supported the role of estrogen in the suppression of bone resorption in men. Glucocorticoids and secondary hyperparathyroidism were further confirmed as risk factors for osteoporotic fractures, the latter being reduced by calcium and vitamin D. Glucocorticoid treatment was identified as a risk for fractures, independent of other variables. The biochemical mechanisms for most pathophysiological factors, such as the positive effects of estrogen and the negative effects of rheumatic disease, remain to be elucidated.
The structural basis of bone fragility Ego Seeman, Austin and Repatriation Medical Centre, University of Melbourne, Melbourne, AustraliaFractures are uncommon in young adult women and men because the usual compressive, tensile or bending loads per unit area imposed on the bone are well below its strength (). Bone fragility emerges during aging because bone remodeling, the process that replaces old with new bone, and bone modeling, the process that constructs the size and shape of bone during growth, fail to maintain the material and structural properties of bone that have been selected for during phylogenesis and developed during ontogenesis.
At some time during adulthood that is still poorly defined, the balance between the volumes of bone removed and replaced in each basic multicellular unit (BMU) changes from being positive during growth, to become zero then negative because bone formation decreases (). In men, the volume of bone resorbed in each BMU probably remains unchanged, except perhaps in a subset of men with very low estrogen levels in whom prolongation of the life span of osteoclasts increases the volume of bone removed in each BMU (). In women, bone formation also declines before menopause and more so after menopause because estrogen deficiency reduces the life span of the osteoblast. In addition, the volume of bone resorbed increases in women so that the combined effect of a larger volume of bone removed plus a smaller volume of bone replaced in each BMU increases the negative balance in each BMU. This is the structural basis of bone loss and the structural damage. If remodeling rate increased without any imbalance in each BMU there would be no structural damage, only a fall in the mineral content of the bone tissue.
Men lose bone more slowly than women during young adulthood and midlife because remodeling rate continues slowly. In women, menopause is accompanied by an increase in bone remodeling rate so that the sum of all the negative bone balances in each BMU produce an acceleration in bone loss and accelerated damage as trabeculae are lost and become disconnected. Loss of complete elements, as occurs in women, produces a greater loss of strength than trabecular thinning in men (which is the effect of reduced bone formation without increased resorptive volume) (). Loss of horizontal trabeculae predisposes to bending and failure in buckling of the remaining unsupported thinner vertical trabeculae ().
The higher remodeling rate that accompanies estrogen deficiency after menopause produces more severe structural damage in women at cortical sites, as well as in trabecular bone, as the many more BMUs remodeling bone on its endosteal surfaces, each with its negative BMU balance, thin the cortices, and make them more porous (). The loss of bone from the cortical component accelerates as remodeling increases the surface to volume ratio ‘trabecularizing’ cortical bone. That is, the intensity of remodeling remains unchanged (or may increase as secondary hyperparathyroidism occurs). A lower mass of bone is eroded by the same intensity of remodeling, pores in cortical bone coalesce, and the cortices continue to become thinner. As the cortices thin and become more porous the cross-sectional area of cortical bone in the axial direction decreases, so at the same load the load per unit area now is greater. Likewise, as the trabecular spongiosa erodes in volume the cross-sectional area also declines. The same loads then produce greater deformation in bone in bending or compression, predisposing to microdamage, cracking and, if microdamage is not repaired, fracture may follow in cortical and trabecular bone.
The high remodeling rate itself alters the material properties of bone as well as its structure. The bone mineral content of the cortical and trabecular bone remaining decreases, reducing its stiffness, and producing greater deformation and fracture. In regions less exposed to remodeling such as the interstitial bone, the mineral density may be highest, providing the least resistance to microcrack propagation. Thus, even though loads probably diminish with age because physical activity is less and muscle mass declines, bone strength, a function of its material and structural properties, declines more greatly than the decline in usual loading. The ratio of load to strength, a measure of fracture risk, approaches or exceeds unity so that the margin of safety from structural failure diminishes ().
As endosteal bone loss proceeds, periosteal apposition continues, but markedly less so than during growth of course. Whether periosteal apposition is independent of endosteal bone loss or is an adaptative response to increased relative loading (on a bone of diminishing mass and compromised architecture) is uncertain. Age-related periosteal apposition occurs in both sexes but it is greater in men than in women so that vertebral cross sectional area (CSA) increases more in men than women (). Consequently, the load imposed per unit area of bone decreases more in men than women. Periosteal apposition also offsets endosteal bone loss more in men than in women so that net bone loss from the whole bone is less in men than women. Thus, the load on the bone decreases more, and the strength of the bone decreases less, in men than women. This sex difference may be a model for the structural basis of differences in bone fragility between races and between individuals of the same sex who come to sustain fractures compared to those who do not. African Americans have fewer fractures than Caucasians perhaps because peak trabecular thickness is greater, remodeling rate is lower by virtue of the lower surface to volume ratio of the trabecular bone mass, so trabecular bone loss and architectural disruption may be less during aging. Cortices of long bones are thicker so that the endocortical thinning compromises CSA less than in Caucasians; these features have not been directly studied at this time.
Sex steroids play a most important role the regulation of bone remodeling and bone structure. Estrogen deficiency results in increased bone resorption, while estrogen and testosterone deficiency may contribute to a reduction in bone formation, in part, by modifying the life span of the cells regulating remodeling (). Several studies at the World Congress in Osteoporosis held in Lisbon addressed the structural basis of bone fragility and the role of sex steroid deficiency in producing structural failure. Szulc et al., in a study of 934 men aged 19 to 85 years, report that the extent of periosteal expansion in long bones is, in part, explained by circulating testosterone concentrations while endosteal bone loss is associated with circulating estrogen (). It is not known whether the smaller bone size in men with fractures is due to a reduction in periosteal apposition during aging or whether the smaller bone size is growth-related.
Gennari et al. suggest aromatisation of androgens to estrogens is important in determining rates of bone loss in men (). In their prospective study of rates of bone loss in 200 men, those subjects with a free estrogen index below the median had greater rates of loss than those above the median; no such relationship was found for the free androgen index. It is not clear whether the division into these two groups produced differences in other covariates that may explain the greater rates of bone loss. Nevertheless, Goemaere et al. also report higher rates of loss associated with estrogen not testosterone levels in a 4 year prospective study of 273 men (). The authors also report that individuals homozygous for the shortest allele length of the TTTA aromatase gene had more rapid bone loss. Legrand et al. report that in men, greater numbers of risk factors for osteoporosis, including hypogonadism, chronic illness, tobacco use, and low body mass index were associated with reduced trabecular connectivity (). Whether the hypogonadism reflects low testosterone or estrogen was not examined. Uebelhart et al. report that raloxifene reduces remodeling markers in men with reduced circulating estrogen levels below about 100 pg/ml, but not in men with higher levels (). This observation supports the notion that men with lower estrogen are predisposed to bone loss, as reported by Khosla et al. (). Taken together, these studies support the suggestion by Riggs et al. that estrogen deficiency plays a central role in bone loss in both men and in women ().
Duan et al. report that women with hip fractures have increased femoral neck diameter while men with hip fractures have reduced femoral neck diameter (). Due to the differences in bone size, the section modulus, a measure of bending strength, is reduced in men with hip fractures but not in women with hip fractures. The buckling index, ratio of cortical thickness to the femoral neck radius, was increased in both sexes with hip fractures. These differences in bone size may be growth or age-related and may be the result of sex hormone deficiency, which in females may produce increased bone size (as estrogen inhibits periosteal apposition during growth) while testosterone deficiency may limit periosteal expansion during growth or aging.
The growth hormone and insulin like growth factor 1 (IGF-1) axis is likely to be important in establishing structural features of bone during growth as well as aging. However, patients with acromegaly do not have high bone density when the larger bone size is take into account. Eckstein et al. report that transgenic mice overexpressing GH had larger bone size but the ratio of bone size to cortical area was normal as was bone mass to body weight (). Females had increased trabecular bone volume due to increased trabecular thickness but males had reduced trabecular thickness while tissue matrix density was reduced in both sexes. Explanations for these interesting findings are not apparent.
A more structural approach to the analysis of drug action is also beginning to emerge. Antiresorptive agents reduce remodeling, reduce the life span of osteoclasts and perhaps increase the lifespan of osteoblasts, effects that will reduce the progressive loss of trabecular architecture, lessen trabecular thinning and cortical porosity, and increase bone mineral content of bone tissue by allowing more complete secondary mineralization (). Dufresne et al. used micro-computed tomography to examine microstructural changes in response to risedronate. In minipigs and in iliac crest bone biopsy from human subjects treated with risedronate, trabecular microarchitecture was better preserved than in control minipigs and placebo treated subjects respectively ().
Although antiresorptive drugs reduce progression of the architectural decay, these agents do not restore the structure of bone beyond what is produced by reducing the size of the reversible remodeling space. Legrand et al. report that, in a primate model ovariectomized 10 months before, treatment with a PTHrP analog resulted in an increase in trabecular bone volume above sham and ovariectomized controls (). Cortical porosity was similar to ovariectomized animals with biochemical and histological changes in bone formation, reflecting the anabolic effect of this drug unaccompanied by change in resorption indices.
Progress in our understanding of the pathogenesis of bone fragility is occurring. As the specific material and structural properties of the bone that determine its strength are defined in unambiguous terms, differences in bone strength between individuals of the same or opposite sex can be specified and the mechanisms responsible differences in bone size, cortical thickness and porosity, trabecular numbers, thickness, connectivity, tissue mineral density, microfracture burden, can be defined. In time, the vagaries contained in terms like bone ‘quality’ and ‘idiopathic’ osteoporosis will become obsolete.
Genetics of osteoporosis John Eisman, Garvan Institute of Medical Research, St Vincent's Hospital, Sydney, AustraliaThe importance of genetics for osteoporosis research, both as a tool and a target, was evidenced by three Plenary talks and a further four oral and several poster presentations that focussed on Genetics at this IOF World Congress. Roland Baron addressed the potential for genetic tools to help identify new targets for anabolic agents in bone and for different antiresorptives. He distinguished the concept of genomics (the use of sequencing and identification of genes and gene families) from functional genomics (the use of microarray profiling and proteomics to identify expressed genes under specific conditions and in specific tissues). He stressed the advantage to be gained from using human and murine genetic models of existing mutations, as well as random mutagenesis in combination with positional cloning, to identify novel genes and gene regions. Transgenic approaches also have a tremendous potential using knock-ins, knock-outs and conditional alterations to identify specific targets. Baron also focused on the large number of genes now known to be involved in both endocrine, paracrine and autocrine regulation of bone biology, as well as many metabolic and transcriptional factors which have been identified through genetic approaches. One example was the identification of Cathepsin K, first as a major gene expressed in an osteoclastoma, and then as the gene in which a mutation is responsible for the human disorder, pycnodysostosis. Another major example was the identification of the lipoprotein receptor-like protein 5 (LRP5) gene from two diverse genetic disorders: osteoporosis-pseudoglioma with a loss-of-function mutation associated with low bone mass, and the familial high-bone-mass phenotype with a gain-of-function mutation associated with high bone density and complete lack of fractures in affected family members (). In these loss- and gain-of-function examples, further study using genetic models identified the strong involvement of the Wnt and frizzled pathways.
Gerard Karsenty addressed the incomplete understanding of molecular mechanisms underlying the complex control of the three major “unknown” regulatory pathways, i.e. control of longitudinal growth, bone mineralization and bone remodeling. He stressed the power of combination studies of human mutations with mouse genetics and in vitro cell culture models to identify genes and gene pathways, including “new pathways we do not suspect”. He did note, however, that approaches were unlikely to identify environmental factors or replace biochemical studies. He used the example of how the knock-out of Cbfa1, thought to be involved in the regulation of the immune cell development, revealed a marked bone phenotype in mouse with no mineralized bone despite a normal cartilaginous skeleton; the heterozygote mutant mouse had a syndrome comparable to the human craniocleidodysostosis. He noted that of the many new regulators and pathways recently shown to be involved in osteoblast and osteoclast regulation, almost all had been identified through genetic tools. Karsenty went on to discuss the concept that bone mass, body weight and gonadal function have common “endocrine” regulators. The obese Ob/Ob mouse has high body weight and high bone mass despite low sex hormone levels. Other mice, including the fat-free mouse and lipodystrophic patients, without excess fat and with low body weight also have high bone mass associated with high leptin levels. He described how recent studies using intracerebroventricular leptin induced marked bone loss in the Ob/Ob and normal, wild type mice (). This has been complemented by other recent work demonstrating the critical role of the hypothalamic NPY Y2 receptor pathway in mediating this effect (). In this work targeted ablation of the Y2 receptor led to a doubling in bone volume. He concluded by mentioning the potential of the LRP5 and OPG pathways and stressing how the combination of human and animal mutation and in vitro biochemical studies was providing major new insights and targets in bone biology and therapeutics. An animal study () with some parallels to the Cbfa-1 story, looked at the knock-out of a gene, SCA-1, associated with peripheral mononuclear cells. Initially these animals had quite normal development but then developed much lower bone density and worsened bone fragility with advancing age. It remains to be shown whether this particular gene is associated with similar changes in humans.
Stuart Ralston reviewed data indicating that different bone “phenotypes”, including bone mineral density, bone geometry and bone turnover were considered to have 50 to 85% genetic components. Even fracture outcomes had 25 to 35% heritability. To demonstrate the importance of heritability, he pointed out that a family history of hip fracture conferred as high a risk of subsequent fracture as a personal history of a low trauma fracture. He described the approaches for gene mapping, including linkage in extended families and sib pairs, animal studies and association studies in human models using candidate gene approaches. He summarized the large number of loci that have now been identified in both human and mouse, consistent with the involvement of multiple genetic factors in bone health. He reviewed a number of human bone diseases that had allowed the identification of specific gene pathways in bone development including the genes for cathepsin K, TGF and carbonic anhydrase type II. Ralston, as had Baron and Karsenty, focused on the new area developed in relation to LRP5 from the combined approaches of the human high bone mass and the osteoporosis-pseudoglioma inherited conditions. He went on to discuss the large range of candidate genes that have been shown to be associated with osteoporosis, as measured by bone density, and by fracture independent of bone density. He reviewed in more detail the work on the collagen 1(I) gene that had come predominantly from his own group. Initially identified by its association with low bone density, this polymorphism in an Sp1 binding site in intron 1 of the collagen 1A1 gene conferred greater transcriptional activity (). This led to a distortion of the normal ratio of 2:1 1(I) and 2(I) gene products resulting in reduced collagen strength. In vitro osteoblasts with this polymorphism generated fewer and less mineralised bone nodules. Interestingly, this polymorphism was associated with increased risk of fracture, independent of bone density, despite its identification through association with lower bone density. He commented that, although a high proportion of bone density variance appears to be genetically determined, in the groups he has studied 70 to 80% of the variance remained unexplained even after accounting for major known genetic and environmental factors. He expressed the view that, with the rapid advance in the number of genes that had been identified in relation to bone in the last few years, it was likely that more and more genes would be identified in which polymorphic variance could be associated with clinical bone outcomes. He supported the view of Baron and Karsenty that better understanding of genetic factors in osteoporosis pathogenesis would lead to targets for new drugs, potentially better fracture risk assessment and also better understanding of response to treatment.
The oral presentations featured a report chromosomal mapping in the Framingham population (). Bone density and genotyping data were available in more than 1,500 family members from 330 nuclear pedigrees with 2-29 members. A large number of loci with suggestive LOD scores were identified, including that on 4q34, which appeared to be related to femoral sites as well as lumbar spine. Other loci, particularly on 8q24, 9q22 and 12q23, had variable associations with spine and proximal femur regions. All of these regions require much further in-depth analysis. A repeat polymorphism in the aromatase (CYP19) gene on 15q21.1 was studied in 273 community dwelling men aged 71-86 years. A variable association between the shortest repeat (7) allele length and more rapid bone loss at both the hip and forearm was found, partly related to bioavailable oestradiol. In the discussion Richard Prince (Perth) commented on a similar finding in over 1,000 postmenopausal women. In another repeat polymorphism in the androgen receptor gene in 300 men aged 55-85 years, greater repeat length was associated with lower bone density, more bone loss and lesser phalangeal thickness, as assessed by quantitative ultrasound (). Of interest, the polymorphism effect appeared to be greatest in the lowest quartile of serum testosterone, suggesting an important interaction between the hormone level and its receptor for this relatively common genetic variant. An interaction between the vitamin D receptor and oestrogen receptor gene polymorphisms was correlated with fracture risk in more than 600 women aged 55-80 years from the Rotterdam study. The presence of one haplotype of the vitamin D receptor in combination with another relatively uncommon haplotype of the oestrogen receptor was associated with a ten-fold increase in fracture risk, four-fold greater than that associated with either polymorphism alone. Although this analysis was based on relatively small numbers, the effect was very strong. An association between height and these polymorphisms was also reported from the Rotterdam study. Adult height was significantly lower (-3.8cm) in those homozygotic for one VDR haplotype and for one oestrogen receptor haplotype (-1.6cm).
A number of other studies also looked at the vitamin D receptor. A large study from Japan () analyzed 1,434 of a group of 4,500 women. They reported slightly lower spine and hip bone density with the tt allele (equivalent to BB) of the vitamin D receptor, but only in premenopausal women. A smaller Korean study () in 276 postmenopausal woman observed modest relationships of the Bsm, Apa, Taq and a polyA length allele polymorphisms with bone density. A small Argentinean study () in 110 postmenopausal women found no overall association between VDR genotypes from bone density. However in the osteoporotic subset, the VDR bb genotype was associated with higher bone mass. Another small study examined the Fok1 VDR polymorphism in 114 healthy postmenopausal Czech women () and found that proximal femur bone density was lower with the FF genotype. The same group reported associations of DHEA S levels () with vitamin D and oestrogen receptor polymorphisms and a weak association of apolipoprotein E genotype () with bone density. The latter three reports from the same group indicate the performance of multiple studies in the same small population sample, without explicit adjustment for the number of different polymorphisms and genes studied, raising the major potential risk of confounding by Type 2 errors. This is also a significant issue in whole genome scans.
A number of smaller studies investigated other gene polymorphisms. A UK study of the growth hormone gene () in 196 men and 124 women aged 63-73 years related polymorphisms to weight at birth and at age one year. One GH allele was associated with lower weight at age year and also more rapid bone loss at the spine in adulthood. They found no such relationship with polymorphisms in the IGF-1 gene. A Danish study () in 133 woman reported an association of polymorphisms of osteoprotegerin and possibly bone sialoprotein genes with lower forearm bone density. A relatively large Swedish study of 425 women found no association of a PPAR polymorphism with bone density at any site ().
Some studies examined polymorphisms in relation to response to treatment. A Japanese study () of polymorphisms in the vitamin D receptor, oestrogen receptor and ApoE genes reported that the Xba I estrogen receptor genotype was related to response to estrogen replacement therapy and the ApoE genotype to both baseline bone density and response to vitamin K treatment. In a very small study of only 21 individuals, TGF β1 polymorphism () was associated with change in bone density after completion of bisphosphonate treatment. A large Austrian study () with data on 1,254 elderly (aged 70+ years) noted the potential to examine relationships with bone metabolism and neuromuscular outcomes.
Overall, the plenary, oral and poster presentations on genetics focused on the potential value of human and animal studies to identify new pathways and targets for treatments. The clinical studies extended this concept to examine how polymorphisms might help identify high risk individuals in terms of having low bone density or losing bone more rapidly or having an adverse effect in response to treatment. The plenary, oral and poster presentations all stressed the potential use of using genetic tools to develop new treatments and to help identify individuals likely to have better (or worse) outcomes to such treatments. This was also recently addressed in an entire issue of Endocrinology on “Bringing Genomics Research to Endocrinology” ().
Epidemiology of osteoporosis Kassim Javaid and Cyrus Cooper, MRC Environmental Epidemiology Unit, University of Southhampton, UK Burden of diseaseL. Joseph Melton summarized the worldwide perspective on osteoporosis. Projections of burden of disease, despite the uncertainty, provide a guide for service provision at local, national and international levels. It is estimated that currently there are one million hip fractures globally each year, and that by 2050 this will increase to 4.5 million per annum. While in Europe the rates are stable, other countries in the world are experiencing rises in hip fracture of up to 3% per annum. A one percent rise in the incidence of hip fracture would result in eight million hip fractures per annum globally by 2050. Hence, the burden of osteoporosis will switch from western populations to those of Africa and the East as the number of elderly increases in both absolute and relative terms.
Peak bone massThe risk of osteoporosis in later years is related to both peak bones mass attained by early adulthood and the rate of bone loss on later years. Skeletal growth tracks from early life and is a determinant of peak bone mass. While maternal factors such as smoking, fat stores and physical activity may relate to fetal bone accrual (), the mechanism is unknown and may involve programming hormones such as leptin. In a study of 115 neonates, cord leptin was related to infant body composition and also, cord leptin appeared to mediate the relationship between a mothers fat store and her baby's bone mass ().
The sex specific growth of the skeleton during puberty may in part explain the marked gender difference in osteoporotic risk in later years. Further evidence for sex specific growth during puberty comes from the Minos Study (), which demonstrated that in men, testosterone was related to the external diameter of the femoral neck, radius and ulna. In addition low concentrations of estradiol were related to increased rates of bone loss. The relationship between high sex steroid markers and reduced bone turnover markers was apparent in adolescents but not during later adulthood.
Male osteoporosisThe definition of male osteoporosis using BMD measurements depends critically on whether sex specific reference ranges are used. Using data from the Rotterdam study, men with incident hip fracture had a higher absolute BMD than women (). However, the relationship found between hip fracture and BMD was similar in men and women. The authors concluded that the use of male-specific T scores to guide fracture risk from BMD measurement in men was the better model.
GeneticsFrom twin studies the heritability, the variation of phenotype due to genetic variation, of skeletal status ranges from 60-85% for BMD, 70-85% for hip geometry, 50-75% for bone turnover and 25-35% for fragility fracture, according to Stuart Ralston. A positive family history of osteoporotic fracture increases an individuals risk of fracture by 3.7 fold. This has lead to studies for candidate genes using linkage analysis. A goal of genetic research is the identification of a cluster of genes whose polymorphisms are responsible for determining bone mass across populations. One candidate is the Col1A1/2 gene polymorphisms, which are found in Caucasian but not African and Asian populations and are associated with an increase in fracture risk independent of BMD ().
Another aspect of genetic research is to identify novel pathways for targeting treatments. Studies of the genetics of osteoporosis-pseudoglioma syndrome and high bone mass syndrone families have shown involvement of the same gene, the ldl-receptor-related protein 5. In osteoporosis-pseudoglioma syndrome (), the receptor is inactive () while the receptor is constitutively active in siblings with high bone mass (). While this polymorphism may not account for variation in bone mass in the population as a whole, it has open new avenues for pharmacological research.
Vitamin DWhile hypovitaminosis D is an important component of age related bone loss in the elderly (), evidence is accruing to its importance in the attainment of peak bone mass. In a three year prospective study of adolescent girls, those in the lowest third of base line 25-hydroxy-vitamin D3 had lower gains in spinal BMD than those in the highest third (). These differences persisted even after adjustment for height gain, physical activity and dietary intake of calcium.
Immigrant populations from the Indian subcontinent to Western Europe are recognized to have a higher prevalence of hypovitaminosis D due to reduced sunlight exposure and dietary deficiencies. In a cohort of Pakistani immigrants living in north Norway, 20% of women and 13% of men had biochemical evidence of hypovitaminosis D with secondary hyperparathyroidism (). This high prevalence was found across age groups, including those women of child bearing age. In women but not men, low vitamin D status was related to reduced forearm BMD as measured by single X-ray absorptiometry.
Variation in vitamin D status leading to increased fracture may not be limited to immigrant groups living in Scandinavian countries, but also account for differences in hip fracture in the indigenous population. In a study of the entire Swedish population over 50 years old for 10 years, generating 26 million observation years, a 10 degree increase in latitude was associated with a 46% (male), 27% (female) increase in hip fracture (). Similar differences were found with seasonal variation.
FallingFragility fractures occur when the mechanical load applied to a bone exceeds its strength, thus encompassing both bone strength and trauma load. The prediction of forearm fracture by either falls’ risk or bone strength was analyzed from the European Prospective Osteoporosis Study (EPOS) (). While in men, BMD was a stronger predictor, in women, fall related characteristics were better predictors of future limb fracture than BMD. Once again this highlights the importance of falls assessment as well as bone densitometry in the management of osteoporosis.
Other high risk groupsClinical risk factors for osteoporosis are often used to identify those with low BMD. However the association between clinical risk factors and osteoporosis may extend beyond differences in mineralization and include structural changes, which predict vertebral fracture independently of BMD (). In men with osteoporosis, those referred for assessment with greater than three clinical risk factors were more likely to have micro-architectural changes such as reduced interconnectivity index and increased free end score (). Of the risk factors, smoking, hypogonadism and low body mass index were specifically related to these deleterious structural changes.
Which cluster of risk factors to use and their ability to identify those with osteoporosis is dependent on the sample group studied. A Scandinavian study identified age >82.5 years, low body weight, difficulty in rising from a chair, lack of physical exercise, previous wrist fracture and non-use of vitamin D supplements as predictors of osteoporosis in 4,347 free living women aged over 75 years (). However, even in the presence of all 6 risk factors the positive predictive value, using hip BMD <2.5 as the outcome, was 66%.
Previous fracture is a recognized risk factor for future hip fracture. Whether the type of fracture influences risk in men versus women was analyzed in a metaanalysis of previously published studies of Caucasian women and men aged over 50 years. While vertebral fracture was predictive of any future fracture in both sexes, in men - but not women - a previous Colles’ fracture was associated with a higher risk of future hip fracture (). While the prevalence of Colles’ fracture in men is low, these data suggest that men who sustain one are at high risk for future hip fracture and warrant further investigation.
HIV-infected patients have a higher incidence of osteoporosis and the etiology is thought to be multifactorial, including treatment with protease inhibitors, disease activity, lipodystrophy and wasting (). The effect of protease inhibitor medication is thought to be the principal factor. A case control study of 119 male HIV-infected patients was unable to find a relationship between bone mass and different types of treatment but showed the principal cause for bone loss in these patients was weight loss ().
Hyperthyroidism is a cause of reduced bone mass (), however whether treatment with thyroxine is deleterious is controversial (). From a large cohort of peri-menopausal women, treatment with thyroxine per se was not associated with increased rates of either bone loss or fracture (). However, this may reflect euthyroid status of those on thyroxine treatment in this sample. In different groups of thyroxine treated patients with poorer control, thyroxine users have increased bone loss ().
The effect of parity on risk of osteoporosis in later years is controversial, with studies suggesting benefit (), harm () or no benefit. The mechanism underlying this association is not known but may reflect beneficial endocrine or turnover effects. An alternative mechanism may involve the unique loading of the hip during pregnancy, a time of active bone remodeling. Using hip structural analysis software to analysis the geometry of the proximal femur, nulliparous women had lower cross sectional area, sub-periosteal width and section modulus compared with parous women (). These findings were independent of lactation history and BMD. Each additional birth improved hip structure. Changes in hip geometry during pregnancy have not yet been studied and are indicated to determine whether hip geometry influences or is influenced by parity.
Meeting the expertsAdachi discussed the relationship between corticosteroid use and increased fracture, highlighting the postmenopausal woman on long term corticosteroid treatment being at exceptionally high risk of fragility fracture. Also emphasized was the lack of a lower dose threshold with doubling of vertebral fracture risk even with 2.5 mg daily ().
Compston brought to attention organ transplant patients as being at particularly high risk. The pathophysiology in the short term is due to primarily corticosteroid exposure. Other factors include cyclosporin use and pre-transplant bone disease. In the longer term, premature ovarian failure or hypogonadism and renal dysfunction add to the risk of osteoporosis. The diminishing use of long term corticosteroid therapy and the transplantation of patients earlier on in their disease were cited as important causes for the reduction in fracture incidence in patients after a liver transplant in the UK. The role of pre-transplant assessment of skeletal status and prophylactic amino-bisphosphonate therapy following heart/lung and renal transplant was discussed.
Bone densitometry Paul D. Miller, University of Colorado Health Sciences Center and Colorado Center for Bone Research Lakewood, Colorado USAThere were a number of interesting presentations in the area of bone densitometry. An in-situ analysis of femoral load failure of cadaveric specimens was correlated with the predictive capacity of various bone mass measurement (BMM) technologies to predict the variability of femoral failure loads (). In-situ measurements were made by dual energy x-ray absorptiometry (DXA) of the hip, spine and radius; peripheral quantitative computerized tomography (pQCT) of the radius, tibia and femur; and calcaneal ultrasound (US). Side impact as well as vertical loading were tested. Hip (cervical region) bone mineral content (BMC) or bone mineral density (BMD) had the highest correlation with either side or vertical femoral load failure, while pQCT showed lower predictive value, and heel US had no independent predictive value for estimating femoral strength. The authors make the suggestion that "clinical diagnosis of femoral fracture risk should thus rely on site-specific DXA of the femur." In this reviewer's opinion, while these in-situ data are scientifically important, the authors should not extrapolate in-situ data to clinical conclusions. In addition, diagnosis of osteoporosis and risk prediction are distinct intents of BMM and are not to be confused as the same purpose, as they were by these authors. Diagnosis of osteoporosis by BMM is an application of the World Health Organization (WHO) criteria, not examined in this study (). Risk prediction for hip fracture in the clinical realm is a function of not only BMD but additional risk factors (). Other in vivo human studies that incorporate low BMD into the entire clinical picture of hip fracture risk have shown that factors other than BMD contribute independently to risk; and that low BMD by any device in the elderly population, including heel DXA, heel US and wrist DXA predicts the risk of hip fracture to a similar degree (as expressed as relative risk/standard deviation of BMD) as measuring BMD at the hip alone. Conclusions extrapolated from cadaver in-situ samples should be cautiously applied to living humans.
The ability of 3 different ultrasound measurements, heel (by Lunar and Hologic) and phalanges (by IGEA), to predict hip fracture risk was compared in a prospective study of 7,494 women followed for a mean of 2.5 years (). During this time there were 62 hip fractures. The predictive value of the 2 heel devices was similar and both predicted fracture risk, both as RR/SD (~2.0) and as the area under the curve (AUC) in receiver operating characteristic (ROC) analysis, while the phalangeal US parameters were not predictive of hip fracture risk. For the heel devices the RR/SD as well as the ROC curves seem to parallel those same values for hip fracture prediction observed in the study of osteoporotic fracture (SOF) as well as the EPIDOS study, which have been the only studies simultaneously comparing the risk prediction of hip DXA and heel US measurements () The heel US RR/SD and ROC curves also are similar to the predictive value for hip fracture seen by heel and wrist DXA in the NORA study (National Osteoporosis Risk Assessment) ().
Another presentation provided the preliminary results of a comparison of hip BMD to fall history and personal/parental fracture history in men and women 50-80 years of age. While upper-limb fracture risk appeared to be predominately related to fall and/or personal/parental fracture history rather than hip BMD, lower limb fracture risk was determined by both low hip BMD and personal history of fracture. Investigations into the etiologies of falls may impact the high incidence of fracture rates, especially in Scandinavian countries.
Finally, to predict WHO cut-offs for the diagnosis of osteoporosis, the predictive value of several osteoporosis-specific questionnaires was compared to both central DXA and several peripheral BMM devices, with the T scores being calculated from the CaMos (Canadian multicentre osteoporosis study) reference population database in over 800 Canadian women (). In regression models the peripheral devices explained between 63-75% of the central BMD variance while the risk profiles explained only 2-5%. Yet the ROC curves for the diagnosis of osteoporosis by WHO criteria were no different between BMM devices and the risk indices. Combining peripheral BMM results with risk indices increased sensitivity but decreased specificity. The authors conclude that, in the absence of central DXA, peripheral BMM devices offer a greater sensitivity and specificity for the diagnosis of osteoporosis than do risk profiles. This reviewer and others have previously shown that the T score discordance between BMM devices is wide and to a large degree can be mitigated by calculation of the T score from a consistent young-normal reference population database (). Concordance between different manufacturer BMM devices is also shown in this abstract, where the T score was calculated from the consistent CaMos reference population database. These data also provide more scientific support for the need to create a standardized reference population database between all BMM devices, in order to reduce WHO diagnostic discrepancies that now currently exist between various BMM devices. Furthermore, while the NORA study of > 200,000 postmenopausal women also showed the value of peripheral devices to predict the risk for all fractures, the percent of women diagnosed as being osteoporotic in NORA by WHO criteria (7%) was less than one-half of the percentage diagnosed by hip BMD (16%) in the postmenopausal population (). Much of this underdiagnosis by peripheral technologies is related to the inconsistent young-normal reference population databases used to calculate the T score used by the four separate peripheral devices in NORA. The findings in this paper (), which shows similar positive-predictive value of various peripheral devices as compared to central DXA to diagnose osteoporosis by the WHO criteria, may be related to the use of a consistent (CaMos) database.
Clinical decision making in osteoporosis David J. Hosking, Nottingham City Hospital, Nottingham, UKThe belated recognition that osteoporosis is a problem in men has stimulated the expenditure of considerable effort in order to identify mechanisms and processes leading to hip and spine fracture.
One approach is to model assumptions about the relationship between BMD and hip fracture risk in men and compare the results with prospective data, as has been done for the Rotterdam Study (). In this population the ratio of men to women of a given age with a hip fracture is 1:1.7, although the overall ratio is 1:3 because of the preponderance of elderly women. Men fracture their hips at a BMD approximately 0.7g/cm2 higher than women but the relationship between BMD and fracture risk is similar for men and women. In women the relative risk of hip fracture increases by 2.6 for every SD decreased in BMD. Using a gender-specific SD the same relationship appears to hold for men.
The next issue to consider is the threshold T score used for screening and intervention. Using a gender- specific T score, a value of -2.5 identifies the same proportion of men as women with a hip fracture. In terms of the absolute risk, men and women fracture their hips at the same absolute BMD. However the proportion of men who fracture below this level will be too small for it to be used as a screening threshold.
Several groups have also examined the issue of spine fractures in men and the relationship to vertebral dimensions. In one of these studies () it was found that young men with vertebral fractures did not appear to have smaller vertebrae (where the force/cross sectional area will be increased), but their fragility seems to relate to a lower volumetric BMD. In this respect they differ from older men with spine fractures who have both smaller vertebrae and a lower volumetric BMD. This underlines a recurrent theme of the heterogeneity of the pathogenesis of osteoporotic fractures.
However another study also examined the width of L3 and found that although this was 7-19% lower in the first degree relatives (male and female) of young men (mean age 48 years) with idiopathic osteoporosis, only the differences in the fathers reached statistical significance. It was suggested that the normal periosteal apposition which causes an age related increase in vertebral width did not occur in the family members of these patients — the implication being that this might be under genetic control ().
In addition to a low BMD it is recognized that a number of clinical risk factors predispose to future fracture and a number of presentations addressed this issue. The limitations of single risk factors, with the exception of a prevalent fracture, are well recognized. Several groups have examined the use of multiple risk factors for predicting low BMD or increased fracture risk. In a large prospective trial of bisphosphonate therapy, a combination of six risk factors was examined for the ability to predict hip osteoporosis (). In a population with a prevalence of hip osteoporosis of 19.1% all those risk factors which proved to be significantly associated with osteoporosis and had a prevalence of >5% were entered into a multivariate model. Six independent variables (age >82.3 years, low body weight, difficulty in rising from a chair, self reported wrist fracture, lack of current exercise and non-use of Vitamin D) were examined singly or in combination for their predictive capacity. This varied from a positive predictive value of 6% for the lack of current exercise (population prevalence 3.6%, RR 0.32) to 77% for the combination of all six risk factors (population prevalence 0.3%, RR 4.0). Identifying a low risk population with a 10% probability of hip osteoporosis would only exclude one third of the sample from screening, but raising the threshold to a 20% probability would exclude approximately two thirds of the sample which would be a more pragmatic solution. Multiple risk factors perform better than single factors at higher absolute risk, as might be expected ().
A prevalent fracture is one of the most powerful predictors of future fracture risk, and since forearm fractures often occur at an earlier age their utility for predicting later events was examined by several groups. In a meta-analysis () it was found that men have a greater risk of a hip fracture after a wrist fracture compared to women. (Male pooled RR 3.26, Female 1.53). In men, a spine or wrist fracture appeared equally predictive of future hip fracture, while in women, only spine fractures had this effect. In the Rotterdam Study the RR of an incident vertebral fracture appeared greater in women than men. (Female 4.1, Male 2.8) although the confidence intervals around the estimate were similar ().
Although osteoporosis is more common among Colles fracture patients (44%) than controls (27%), the depressing finding is that 37% of the Colles fracture patients had had a previous fracture which had failed to elicit a therapeutic response ().
Predicting the response to treatment also continues to stimulate interest in new markers, and not surprisingly attention focuses on OPG as a marker for effects of bisphosphonate therapy (). After an initial fall which parallels that of CTX there is a return towards baseline, which was attributed to the effect of bisphosphonates to increase secretion/cell, offsetting the reduction in osteoblast number as turnover decreases. The gain in OPG correlated well with the increase in trochanteric BMD after a year of bisphosphonate therapy, while the combination of the decrease in CTX and the gain in OPG accounted for 71% of the variance in BMD gain. Of less certainty is the pathogenetic mechanisms which are being reflected by such measures of the ‘cytokine soup’.
Treatment of osteoporosis Juliet E. Compston, University of Cambridge, UKAlthough there is already a range of options available for the management of osteoporosis in postmenopausal osteoporosis, new therapies are still emerging. There is particular interest in intermittent bisphosphonate regimens, which may have better tolerability than continuous ones, and in the development of drugs with anabolic skeletal effects, which produce larger increases in bone mineral density than antiresorptive agents and thus might be expected to have greater anti-fracture efficacy (although this currently remains unproven).
Ibandronate is currently being evaluated for use in the prevention of osteoporotic fractures in postmenopausal women and two oral presentations () were devoted to this topic. One of these () reported the effects of intermittent intravenous ibandronate, 1 or 2 mg given at 3 monthly intervals, in postmenopausal women with low spine bone mineral density. The increase in spine bone mineral density achieved with the 2 mg regimen was significantly greater than that seen in women treated with 1 mg, (5% vs 3% at one year, p<0.0001). Furthermore, in this study a significant reduction in vertebral fracture risk was documented (62% and 50% respectively for the two regimens). Thus ibandronate can be added to the list of bisphosphonates already shown to have anti-fracture efficacy () although further data are required to establish whether this extends to non-vertebral and hip fractures. Provided a sufficient dose is given, intermittent regimens appear to be as effective as continuous daily administration.
A critical issue in the management of osteoporosis is the rate at which treatment effects wear off after cessation of therapy. In the case of the bisphosphonates, there may be differences between compounds, but data are relatively sparse (,,,). A study of postmenopausal women who had received alendronate () suggests that the longer alendronate therapy is continued, the more lasting its effects after withdrawal of therapy. Furthermore, the rate of bone loss after withdrawal does not appear to be accelerated, but is similar to normal postmenopausal loss; hence, a finite treatment period results in a lasting treatment benefit.
Another potential application for bisphosphonates is in the management of osteogenesis imperfecta. To date, most studies have focused on children with this disease () but in an oral presentation () it was reported that cyclic administration of intravenous neridronate produces significant increases in bone mineral density at the spine and hip in affected adults. In this study the regimen used was 2mg/kg intravenously every 3 months; whether the increases in bone mineral density are associated with fracture reduction remains unknown although there was a trend towards fewer fractures in treated adults in this study.
Anabolic agents are currently a topical issue in osteoporosis treatment () and both parathyroid hormone peptide [PTH(1-34)] and strontium were topics for oral presentations. In men, beneficial effects on bone mineral density of daily injections of recombinant human PTH(1-34) (teriparatide) have been shown () and in analysis of data from another study it was shown that the treatment response in spine bone mineral density is independent of age, previous fracture, tobacco smoking, alcohol intake or sex steroid levels ().
Strontium is a bone seeking element that stimulates bone formation and inhibits resorption (). In a Phase II study, increases in bone mineral density in the spine and hip were recently reported in postmenopausal women with osteoporosis treated with oral strontium ranelate, and the incidence of new vertebral deformities (a secondary end-point) was also reduced in women treated with the highest dose (2g/day) (). The results of a large Phase III randomised controlled trial of the effects of this strontium compound given in a daily dose of 2g, were reported in the Late Breaking News session (). This multicentre study was conducted in 1,649 postmenopausal women with established spinal osteoporosis. After three years, a significant treatment benefit was seen for spine bone mineral density (11.4% vs -1.3% in treated and placebo groups respectively, p<0.001) and a 41% reduction in vertebral fractures was also observed in women treated with strontium (relative risk 0.59, 95% confidence intervals 0.48: 0.73; p<0.001). Interestingly, the changes in biochemical markers of bone turnover with treatment indicated uncoupling of bone turnover, with an increase in bone specific alkaline phosphatase (a marker of formation) and decrease in serum CTX (a marker of resorption). Strontium is well tolerated and no specific adverse events emerged from this large study.
NOTE: This meeting report is made possible with help from IOF staff. Free access to all meeting abstracts and late breaking news from the WCO Meeting are made available through the Osteoporosis International and IOF web sites. Osteoporosis International is a joint initiative of the International Osteoporosis Foundation and the National Osteoporosis Foundation of the USA, and is published by Springer.
This work is licensed under a
Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 United States License.