Osteopaenia – a marker of low bone mass and fracture risk

Introduction Absolute fracture risk is greatest for individuals with osteoporosis, more than half of these fractures arise from those with osteopaenia and normal bone mineral density, a probable consequence of greater numbers at risk in these categories. However, areal bone mineral density measurements used commonly in clinical practice do not detect differences in bone tissue properties, geometry and microarchitecture, contributing to bone strength. This critical review discusses osteopaenia as a marker of low bone mass and fracture risk. Conclusion Newer technologies such as high-resolution peripheral computed tomography have the advantage of assessing trabecular and cortical components of bone separately. Quantifying these parameters and considering clinical risk factors that affect fracture risk independent of bone quantity and quality, may better discriminate between high and low-risk individuals, improving the decision-making for targeting appropriate interventions to reduce the public health burden of fractures. BMD is a continuous variable, which approximates a normal distribution, and it is commonly categorised into normal BMD, osteopaenia and osteoporosis on the basis of nominal thresholds recommended by an expert panel of the World Health Organization3. Osteopaenia is the low bone mass category defined by BMD T-scores between −1.0 and −2.5. Using these cut-points, 16% of young normal women are defined as having osteopaenia and 5% have osteoporosis, but these individuals may make little, if any, contribution to the population burden fragility fracture. The osteopaenia threshold was based on data that derived a theoretical fracture threshold4, whereas the osteoporosis threshold was based on the prevalence of fracture among postmenopausal Caucasian women. Although these thresholds were devised for epidemiological purposes appropriate for Caucasian women, they have been widely adopted for clinical use in broader populations. thresholds for men aged 50 years and older indicates that just over half have osteopaenia corresponding to BMD T-score from −1.0 to −2.5 (and 6% have osteoporosis, T-score < −2.5)5. As there is discordance in BMD between skeletal sites, such estimates depend on the site scanned, as well as the reference range used to determine T-scores6,7. However, while osteoporosis confers the greatest risk for fracture, fracture risk is not negligible in persons with more moderate deficits in BMD8–10. Age-standardised 5-year absolute fracture risk derived from total hip BMD at baseline for post-menopausal women in Australia are 30.8% (95%CI 22.0–39.6) for women with osteoporosis, 17.5% (95%CI 13.2–21.7) for women with osteopaenia and 7.2% (95%CI 3.7– 10.7) for women with normal BMD8. The aim of this critical review was to explore osteopaenia as a marker of low bone mass and fracture risk.

BMD is a continuous variable, which approximates a normal distribution, and it is commonly categorised into normal BMD, osteopaenia and osteoporosis on the basis of nominal thresholds recommended by an expert panel of the World Health Organization 3 .Osteopaenia is the low bone mass category defined by BMD T-scores between −1.0 and −2.5.Using these cut-points, 16% of young normal women are defined as having osteopaenia and 5% have osteoporosis, but these individuals may make little, if any, contribution to the population burden fragility fracture.The osteopaenia threshold was based on data that derived a theoretical fracture threshold 4 , whereas the osteoporosis threshold was based on the prevalence of fracture among postmenopausal Caucasian women.Although these thresholds were devised for epidemiological purposes appropriate for Caucasian women, they have been widely adopted for clinical use in broader populations.thresholds for men aged 50 years and older indicates that just over half have osteopaenia corresponding to B-MD T-score from −1.0 to −2.5 (and 6% have osteoporosis, T-score < −2.5) 5 .As there is discordance in BM-D between skeletal sites, such estimates depend on the site scanned, as well as the reference range used to determine T-scores 6,7 .However, while osteoporosis confers the greatest risk for fracture, fracture risk is not negligible in persons with more moderate deficits in BMD [8][9][10] .Age-standardised 5-year absolute fracture risk derived from total hip BMD at baseline for post-menopausal women in Australia are 30.8%(95%CI 22.0-39.6)for women with osteoporosis, 17.5% (95%CI 13.2-21.7)for women with osteopaenia and 7.2% (95%CI 3.7-10.7)for women with normal BMD 8 .The aim of this critical review was to explore osteopaenia as a marker of low bone mass and fracture risk.

Fracture and BMD
The authors have referenced some of their own studies in this review.These referenced studies have been conducted in accordance with the Declaration of Helsinki (1964) and the protocols of these studies have been approved by the relevant ethics committees related to the institution in which they were performed.All human subjects, in these referenced studies, gave informed consent to participate in these studies.Populationbased studies reveal that the burden of fracture arises, not from the relatively small, high-risk group with osteoporosis, but from the larger group with inter-mediate risk.Different studies have used different inclusion and exclusion criteria, defined low-trauma fractures in different ways and ascertained different combina-

Defining osteopaenia and osteoporosis
For nearly three decades, areal bone mineral density (BMD) has been measured using either dual-photon absorptiometry or, more recently, dual-energy X-ray absorptiometry (DX-A) as a marker of osteoporosis Based on proximal femur BMD from a population-based cohort study of 616 postmenopausal Australian women followed for 5.6 years, 26.9% of radiologically confirmed fractures arose from women with osteoporosis and 73.1% from women without osteoporosis (56.5% from women with osteopaenia and 16.6% from women with normal BMD measured at the total hip) 8 .
Similarly, in a cohort study of 149,524 postmenopausal white women enrolled in National Osteoporosis Risk Assessment (NORA) from primary care practices in the United States and followed for 12 months after baseline assessment using BMD measurements obtained from a variety of peripheral devices (heel, finger or forearm), 18% of self-reported fractures arose from women with osteoporosis and 82% from women without osteoporosis (52% from women with osteopaenia and 30% from women with normal BMD) 10 .
The NORA data excluded those with a history of known osteopo-rosis or anti-fracture therapy and this may explain the difference in the proportion of fractures arising from each BMD category between the Australian and US studies.Furthermore, the prevalence of osteoporosis and osteopaenia varies depending on the device, the site of BMD measurement and the reference population used to define T-scores.Same-site differences exist, not only between machines from the same manufacturer, but also across devices made by different manufacturers 11 ; peripheral devices generate T-scores that may differ from those derived using hip or spine DXA 12 and the choice of reference population impacts on the BMD cut-points for identifying osteopaenia and osteoporosis 6 .The Geelong Osteoporosis Study (GOS) data used proximal femur BMD because of the well-recognised influence of age-related artefacts on spine BMD measurements 13 .However, despite these methodological limitations, greater numbers of fractures are consistently shown to arise in women with normal or osteopaenic BMD and this may simply be a consequence of larger numbers of women at risk of fracture in these groups (Figure 1).
However, the risk for fracture is multi-factorial.Many clinical risk factors for fracture operate through reduced BMD; however, others act independent of BMD 14 .Increasing age contributes independently to the risk of fracture; for the same BMD, the risk of fracture varies by a factor of 8-10 between women aged <45 years and 80 years and older 15 .Even though the majority of individuals who sustain a fragility fracture do not have a prevalent fracture (this proportion is 75% among women with osteopaenia), a prior fracture independently doubles the risk of subsequent fracture; women with osteopaenia and a prevalent fracture are at comparable risk to those with osteoporosis on BMD criteria 8 .Low body mass index is recognised as a risk factor for fracture that is essentially independent of age and sex, but dependent on BMD 16 .Falls independently increase the risk of fracture 8,17,18 .

Fracture risk assessment
Various models for predicting fracture have been developed that involve BMD in conjunction with clinical risk factors with an aim of improving risk stratification, particularly within the large group with moderate bone deficits categorised as osteopaenia.The World Health Organization collaborating centre developed the FRAX algorithm as a tool based on clinical risk factors, with and without BMD, using primary data from multi-national prospective cohort studies 19,20 .The FRAX estimates 10-year probability of hip fractures and major osteoporotic fractures (including fractures of the hip, spine, humerus and wrist).
In Australia, data from two population-based studies, the GOS 21   as the GOS Fracture Risk (FRISK) Score 17,22 and the Garvan algorithm 23 , respectively.The GOS FRISK estimated 10-year probability of lowtrauma fractures at the hip, spine, forearm and humerus, whereas the Garvan predicted 5-and 10-year probability of fragility fractures of the hip, spine, wrist, humerus, hand, scapula, clavicle, pelvis, lower limb and sternum.The FRAX includes multiple clinical risk factors, whereas both the GOS FRISK and Garvan utilise fewer risk factors (Table 1).The FRAX did not include falls because these data were not consistently collected across the multiple population-based studies from which the data were derived 14 .The purpose of developing such models is to provide clinically useful tools to better identify, with high sensitivity and specificity, individuals in the population who are at greatest risk for fracture.

Bone microarchitecture and structure
The observation that age and previous fracture independently increase the risk for fracture 8,17 is consistent with the notion that increasing age 24 and fragility fracture 25 are markers for greater material or structural deterioration in bone, not quantified by BMD.Bone morphology and microarchitecture contribute to the breaking strength of bone 26 .To be strong, bones need to be stiff enough to withstand deformation under loading, yet adequately elastic to absorb energy during compression and tension 27 .Recently developed technologies for assessing bone structure include high-resolution peripheral quantitative computed tomography (pQCT) and magnetic resonance imaging that have the advantage of simultaneously assessing trabecular and cortical components of bone separately, in addition to geometric characteristics of the peripheral skeleton 28,29 .
In a matched case-control study of postmenopausal French women, 101 cases with fragility fracture over 13 years of follow-up were matched with fracture-free controls 30 .Vertebral and non-vertebral fractures were associ-ated with low volumetric BMD and structural deterioration of trabecular and cortical bone as assessed by high-resolution pQCT at the distal radius and tibia, independent of areal BMD.Cases had decreased trabecular volume, cortical thickness, trabecular number and trabecular thickness.Similarly, in another study using high-resolution pQCT, osteoporotic women had lower density, cortical thickness and increased trabecular separation than osteopaenic women.Among osteopaenic women, those with fracture had lower trabecular density and more heterogeneous trabecular distribution 25 .These women were defined as having osteoporosis or osteopaenia based on measurements of BMD at the lumbar spine or proximal femur.The lower T-score was used to categorise subjects.A proportion of those with BMD in the osteopaenic range at the lumbar spine, alone, are likely to have had osteoporosis with the BMD measurement being spuriously increased by artefact.The apparent greater micro-architectural deterioration among those women with osteopaenia and fracture may therefore have been related to miscategorisation.
The pQCT assessment of the ultradistal radius in the United States shows that the structural basis for the observed decrease in trabecular volume differs between men and women.With ageing, women undergo loss of trabeculae with an increase in trabecular separation, whereas men start with thicker trabeculae and experience less agerelated microstructural damage.Because decreases in trabecular number substantially affect bone strength, this finding may explain, at least in part, the protection men have against age-related increases in distal forearm fractures.More recent findings suggest that development of intracortical porosity may play an important role in compromising bone strength 24,31 and that this could explain the high proportion of non- vertebral fractures that occur with ageing at predominantly cortical sites 24 .
In an Australian study of 185 female twin pairs aged 40-61 years, postmenopausal women were found to have higher levels of remodelling markers that were associated with larger intracortical surface area rather than with the progressively diminishing trabecular surface area 31 .Identification of intracortical, endocortical and trabecular bone surface area are beyond the resolution of contemporary DXA analysis and are, therefore, not accounted for using BMD from DXA.

Conclusion
Fragility fractures pose a considerable health burden to the community.Effective strategies to reduce the burden of fractures depend on the development of preventive measures to target lifestyle or pharmacological interventions, based on identification of individuals at risk.The burden of fractures arises, not from the relative few with severely low BMD identified as osteoporosis, but from those with mild to moderate bone deficits.Individuals with osteopaenia are commonly not treated because there is a lack of data relating to anti-fracture therapies in this group and, based on post hoc analyses from osteoporosis clinical trials, the numbers needed to treat are too large to be economically feasible if the whole group is to be considered.Yet, over half of the fractures in the population arise from this group.Those at highest risk for fracture within this group need to be identified and evidence-based treatment strategies developed to reduce the public health burden of fractures.
Improved risk stratification may be achieved by quantifying factors that contribute to bone strength, such as bone morphology and microarchitecture, which are properties beyond the resolution of conventional densitometry by DXA.It needs to be demonstrated that such predic-tors of risk are amenable to reduction with osteoporosis therapies and that anti-fracture treatment reduces fracture risk before recommendations are deemed appropriate.Furthermore, non-bone risk factors, that are amenable to modification, also need to be considered.

Figure 1 :
Figure 1:The distribution of bone mineral density (BMD) at the total hip for women is shown, together with the cut-off points for osteoporosis and osteopaenia (shaded in grey).Absolute risk for fracture (%) is represented by the unshaded columns, indicating that absolute risk for the group with osteopaenia is intermediate between those with osteoporosis and normal BMD.The proportion of fractures arising from those with osteoporosis, osteopaenia and normal BMD is represented by columns shaded in black, indicating that most fractures arise from the group with osteopaenia.Data relate to postmenopausal Australian women8 .