3.1 Introduction

Bone mineral density is the major criterion used for the diagnosis and monitoring of osteoporosis. BMD differs between different sites around the body and there is only a moderate correlation between BMD at different sites.⁶² BMD of a specific site is the best predictor of fracture at that particular site.⁶³ Techniques available for measuring BMD are shown in Table 2.

Table 2: Techniques for measuring BMD

Other techniques are available that measure properties related to bone density. Quantitative Ultrasound (QUS) can be used to measure properties of the calcaneus related to bone quality and structure, though it cannot be used to diagnose osteoporosis or to target treatment. Biochemical markers such as resorption markers can be used to assess bone turnover.

This section focuses on the techniques of DXA, QCT, and QUS, and the use of biochemical markers for the diagnosis and monitoring of osteoporosis.

3.2 Plain radiographs

Assessment of bone density from plain radiographs is not appropriate as it is open to marked observer variation and apparently normal density does not reliably exclude osteoporosis.⁶⁴ Although severe osteopaenia on plain films correlates reasonably with low BMD measured by DXA, there is a wide overlap.⁶⁵ Treatment should not be instituted on the basis of plain film findings as this could lead to many patients being treated unnecessarily. Similarly, many patients with osteoporosis would be missed. The use of digital radiography, which allows the image to be manipulated electronically, has introduced a degree of uncertainty that makes it even more difficult to reliably assess bone density. Evidence level 1+

Conventional radiographs should not be used for the diagnosis or exclusion of osteoporosis.

When plain films are interpreted as ''severe osteopaenia'' it is appropriate to suggest referral for DXA.

Grading of vertebral fractures and the number of fractures will influence management. Standardisation of reporting of vertebral fractures identified on plain radiographs would be helpful. There is an established method for such reporting.⁶⁶

The presence of vertebral fractures should be included in reports on conventional radiographs along with a recommendation for further action. (See Section 6.2)

3.3 Peripheral techniques

DXA scanning is the current standard technique for the diagnosis of osteoporosis due to its ability to measure BMD at a variety of sites. Peripheral imaging techniques such as pQCT, pDEXA, SXA, RA, phalangeal ultrasound, and peripheral radiographic fractal analysis are often used as screening methods for subsequent DXA, for diagnosis of osteoporosis, or the monitoring of treatment. Their principal advantages compared to DXA are their relatively modest cost and the portability of the equipment. Few studies have been done to compare these techniques against the current standard of DXA.

It has been suggested that patients with a forearm T-score of less than -2.5 on DXA should begin treatment, while those with a score greater than -1 should simply be reassured that fracture risk is low.⁶⁷ This, however, only applies to the diagnosis of fracture risk in postmenopausal women. A significant proportion of women with T-scores between -1 and -2.5 would still have to be referred for subsequent axial DXA. It should also be noted that there is only a moderate correlation between forearm or heel and axial BMD and therefore forearm or heel BMD is not appropriate for making treatment decisions. In addition the forearm is not a suitable site for monitoring response to treatment.⁶⁸

There is no role for forearm or heel scanning in the diagnosis of osteoporosis in targeting therapy to reduce fracture risk. In remote or rural areas provision of a mobile DXA service is a viable alternative.

3.3.1 Quantitative ultrasound

Quantitative ultrasound equipment is available that measures a range of parameters using several different methods. Most systems measure speed of sound (SOS) and broadband ultrasound attenuation (BUA) of the calcaneus. Different systems produce different values both in absolute terms and in relation to age-matched subjects. It is not possible to extrapolate the findings from one instrument or technique and apply it to another, and this limits the amount of generally applicable evidence.

This complexity is reflected in the literature, where there is a considerable variation in the design of studies and the presentation of conclusions.

Some studies are based on populations of elderly patients living in sheltered or supervised accommodation. Conclusions from these studies may not be generalisable to populations living independently, particularly when looking at the development of new fractures where protection from falls may have a bearing on fracture incidence.

There is evidence³² that QUS of the heel can predict fracture risk of hip and spine independently of BMD measurements. There is also some evidence that QUS in addition to BMD evaluation by DXA may give a better estimate of fracture risk than DXA scanning alone.³² The precision of QUS is generally poor and changes in QUS of the heel may not reflect changes in BMD at the spine or hip.

The ultimate conclusion must be that though QUS may have a role in improving estimates of fracture risk, this is at best a proxy for the assessment of bone density. QUS of the heel cannot, therefore, be recommended for the investigation or monitoring of patients suspected of being at risk for osteoporosis or to justify initiation of treatment.

There is insufficient evidence to support the use of any of these techniques for population screening, or for pre-screening for DXA.

3.4 Quantitative computed tomography

Quantitative Computed Tomography has been widely used to measure BMD, particularly in the spine. It can be performed on conventional CT scanners by purchasing special software. An advantage of QCT is that it can measure cortical and trabecular bone separately. Disadvantages are the relatively high radiation dose involved and the high cost of scans.

Given the limited availability of CT scanners in Scotland and the competing demands for their use, it would not be appropriate to use QCT for the routine diagnosis or monitoring of osteoporosis in NHSScotland.

3.5 Dual energy x-ray absorptiometry

DXA can measure BMD at the spine, hip, forearm, heel, and in the total body. The diagnosis of osteoporosis varies greatly depending on the measurement site, and on the number of sites measured. It is not possible to exclude a diagnosis of osteoporosis on the basis of a DXA measurement at only one site.³²

Bone mineral density measurement at the hip provides the best prediction of hip fracture risk⁶⁹ but does not exclude the possibility of osteoporosis at the spine. As scans are not analysed while the patient is still lying on the scanner, it is normal to acquire both hip and spine scans at the one visit. The spine is the preferred site for monitoring the response to treatment. Careful interpretation of the spine image, and comparison of the T-scores of individual vertebrae is required as the measured BMD may be affected by factors such as vertebral fracture or degenerative changes.

A well conducted systematic review on the diagnosis and monitoring of osteoporosis in postmenopausal women has demonstrated the effectiveness and usefulness of DXA.³² Evidence level 1++

BMD should normally be measured by DXA scanning performed on two sites, preferably anteroposterior spine and hip.

3.5.1 monitoring

The precision of DXA varies with BMD and the measurement site. In this context, precision is a measure of how reproducible BMD is if it is measured several times during the same patient visit, with the patient re-positioned on the scanner between measurements. This determines how large the change in BMD must be between successive patient visits before it can be confidently interpreted as a genuine change.

In the spine precision can be affected by artefacts associated with degenerative changes. When outliers were removed, long term precisions of 1.1, 2.2 and 1.3% were achieved for the lumbar spine, femoral neck and total hip BMD respectively.⁷⁰ To detect changes at the 95% confidence level they must be at least 2.8 times the precision error. Changes in BMD of the spine, femoral neck and total femur cannot therefore be detected with confidence unless they are around 3, 6 and 4% respectively.

Careful examination of DXA spine images is required to ensure that apparent changes in BMD are not due to artefacts. Misleading changes have been reported in the monitoring of osteoporosis therapy where it has been shown that women who lose BMD during the first year of treatment can gain BMD if the same treatment is continued for a second year. Effective treatments for osteoporosis should not normally be changed because of loss of BMD during the first year of use.⁷¹ There is insufficient evidence to determine whether repeating BMD measurements two years after starting treatment is useful.^4,32 Evidence level 1++

Repeat measurements should only be performed if they influence treatment.

3.5.2 risk

The radiation dose from DXA scans varies depending on the scanner type and the site measured. The combined effective dose from AP spine, lateral spine, and hip scans is typically less than 30 mSv, which corresponds to only a few days natural background radiation or a single transatlantic flight.

Patients should be reassured that the radiation dose from DXA is extremely small.

There is no significant difference in accuracy or precision between older generation pencil beam and newer fan beam DXA scanners. Fan beam systems offer shorter scanning times leading to a high patient throughput. Some models of fan beam DXA scanners also offer high resolution lateral spine imaging.

Spinal degenerative disease is prevalent among the elderly and may result in an artefactual increase in spine BMD measured in the AP view. Lateral spine DXA selectively measures the trabecular rich vertebral bodies and is less affected by spinal degenerative disease. Lateral spine DXA identifies more osteoporotic patients and is more sensitive to age related bone loss than AP spine DXA.⁷² Lateral spine DXA is, however, less precise and a greater treatment effect for lateral spine rather than AP spine DXA may be offset by greater precision errors.⁷³ The lateral spine view is not available on many DXA systems.

If DXA investigations are repeated, AP spine and total hip measurements should be used to follow response to treatment.

Although the greatest treatment effect is often observed in the spine, it is often also helpful to monitor changes in hip BMD because spinal images may be affected by artefacts.

A model DXA report is included in Annex 2.

3.6 Fracture prediction

Annual hip fracture risk can be estimated from age, sex and femoral neck BMD^74,75,76,77 An example of the annual hip fracture risk for females as a function of age and Z-score is shown in Figure 2. To extend the annual fracture risk to a 10 year fracture risk requires the assumption that the patient's Z-score will remain constant.

Figure 2: Annual hip fracture risk for women as a function of age and femoral neck BMD

Following a DXA scan of the hip, the annual hip fracture risk (or 10 year fracture risk) should be included in the DXA report.

Existing vertebral fractures increase the risk of a subsequent vertebral fracture by a factor of four, and double the risk of a subsequent hip fracture.⁷⁸ Identification of existing vertebral fractures is therefore an important factor in the assessment of future fracture risk. Lateral spine DXA can detect vertebral deformity. Visual assessment of lateral spine images from high resolution fan beam DXA agrees as well with lateral radiographs as radiographs interpreted by different radiologists.^79,80 Fan beam DXA results in a radiation dose of only 1% of a lateral radiograph. In addition to its role in identifying existing fractures, lateral DXA can identify artefacts such as aortic calcification and degenerative changes that affect BMD measured in the AP view.

Where lateral spine scans acquired with fan-beam DXA are available, visual assessment should be used to assess prevalent vertebral fractures.

Evidence of existing vertebral deformity should be used to modify the hip fracture risk estimated from age, sex, and BMD.

3.7 Cost of DXA scanning

Provision of, and access to, DXA scanners is variable across Scotland. As a consequence, not all patients suspected of having osteoporosis are scanned. Without scans, patients may be treated inappropriately or not at all. Annex 3 provides an example annual costing of a DXA service. Once a service has been established, the cost per scan is relatively low.

Since some of the costs are fixed it is clear that increased use of a scanner reduces the unit cost. We do not, however, have sufficient evidence to draw conclusions about the impact of DXA on treatment costs, or the effect on healthcare expenditure that may arise from a reduction in fracture incidence.

A complete DXA investigation, including patient preparation, image acquisitions, analysis, printouts and report generation, is typically performed in 20 to 25 minutes, making each DXA system capable of a throughput of approximately 4,000 patients per year. Assuming an annual referral rate of 1% of the population,⁸¹ a minimum of 13 DXA scanners would be required for the Scottish population. Taking into account geographical factors and the distribution of population density, however, a larger number would be required if all communities were to be given equal access.

DXA should be available in all Health Board areas

When new DXA scanners are purchased these should be high resolution systems with fast acquisition times. Consideration should be given to purchasing a system that is capable of lateral spine imaging. Where this is available a typical study should consist of quantitative measurements of AP lumbar spine and hip, and visual assessment of the lateral spine.

3.8 Biochemical markers

Biochemical markers of bone turnover have the potential to have a major clinical impact on the investigation and management of osteoporosis in Scotland. Individual marker assays are simple and inexpensive to perform and modern laboratory technology has the capacity to cope with the maximum likely workload.

By definition, the diagnosis of osteoporosis is directly linked to the measurement of BMD. Several studies have sought to use biochemical markers to select patients at risk of rapid bone loss for subsequent BMD measurement, but have failed to demonstrate a consistent relationship between marker results and bone loss. The sensitivity and specificity of the bone marker assays were too low to be useful.³²

Biochemical markers of bone turnover should have no role in the diagnosis of osteoporosis or in the selection of patients for BMD measurement.

There is evidence from recent studies that resorption markers measured in urine or more recently in serum, can predict increased fracture risk (OR~2) independently of BMD. However, there is no conclusive evidence that has demonstrated the value of one or more specific markers, either alone or in combination with other factors, in the prediction of fracture risk in the individual patient.⁸² At the present time biochemical markers of bone turnover have not been proven to have clinical value in the prediction of fracture risk in individual subjects.

Changes in biochemical marker concentrations alter with therapy and these changes may be used to predict subsequent changes in BMD. It has been suggested that regular monitoring of biochemical markers can increase patient compliance with therapy and/or assist with the alteration of therapy to achieve optimal effect on improving BMD. Although several original studies support this view,^83,84 they have used different markers and different study protocols resulting in variable outcomes.³² A recent meta-analysis shows that the greater the increase in BMD at the spine and hip, or decrease in bone markers at one year, the greater the reduction in the risk of non-vertebral fracture.⁸⁵

There is currently no agreement on the marker(s) of choice for this application or on the preferred strategy for optimal use. Currently there is insufficient evidence to support a recommendation for the routine use of a specific panel of biochemical markers of bone turnover in monitoring and adjusting the treatment being given to patients with osteoporosis. However, this position will change in the foreseeable future and there is every likelihood that evidence will emerge to establish a definite role for biochemical markers in this application.