DXA is the current gold standard for the diagnosis of osteoporosis. The best sites at which to measure BMD for prediction of future fracture risk are the lumbar spine and the proximal femur.^2,4 Both sites should be measured with consideration to dual hip scanning. DXA is reliable, with a reported precision of approximately 1%, although in routine clinical practice this is closer to 2%.² At this level of precision, the least significant change at the lumbar spine would be 5.6% between measurements, with 95% confidence that the change is real.

Each SD reduction in FN BMD increases the age-adjusted risk of hip fracture by a factor of approximately 2.5 (range 2.0–3.5), whereas the risk attributable to any minimal trauma fracture is almost the same (range 1.7–2.4). Similarly, each SD reduction in lumbar spine BMD increases the risk of spinal fracture by a factor of approximately 2.3 (range 1.9–2.8). FN and total hip BMD appear the best overall predictors of fracture risk. Total hip is the better site for monitoring BMD because it has good precision (less affected by positioning) and is relatively unaffected by osteoarthritis, which can spuriously elevate spinal BMD values, as can vertebral fractures and arterial calcification.^2,5

The initial assessment of BMD by DXA has the following aims:

• to determine the patient’s BMD: Fracture risk is multifactorial and may be significantly elevated in individuals outside the osteoporotic range. However, the use of the osteoporotic T-score threshold is the criterion by which healthcare funders define osteoporosis, as well as being consistent with studies in which antifracture effects of anti-osteoporotic drugs have been demonstrated.
• to determine the precise extent of BMD reduction: This is important for refining assessment of individual fracture risk and the extent of recommended therapeutic measures. Absolute fracture risk algorithms (e.g., FRAX^® [available at https://fraxplus.org] or the Garvan Fracture Risk Calculator [available at www.garvan.org.au/bone-fracture-risk]) may be useful in more accurately determining individual fracture risk and assisting the patient in making a treatment decision (refer to Section 1.3).

Repeat DXA scans at intervals of two years or longer can be considered to assist risk assessment or when a change or interruption in treatment is being considered (also refer to Section 4).^6–8 The treatment-related change in BMD correlates with the proportion of fracture risk reduction.^7–9 In addition, repeat BMD measurement can identify people with ongoing bone loss, which is an independent predictor of fracture risk.⁶ Repeat DXA scans may improve adherence to therapy in some people.¹⁰ However, a minimum of two years may be needed to reliably measure a change in BMD due to limitations in DXA precision. If BMD is stable and/or the individual is at low risk of fracture (normal or mild osteopenia; T-score >–1.5), less frequent monitoring, up to an interval of 5–15 years, can be considered. Shorter intervals between repeat DXA scans at intervals of one year may be appropriate in high-risk individuals (e.g., patients on corticosteroid therapy or ADT for prostate cancer). In all cases, the expected rate of change in BMD and fracture risk should guide repeat measurement.⁶

Quantitative computed tomography (QCT) BMD measurement can provide equivalent hip BMD to DXA scans and may be interpreted using the WHO T-score criteria.¹ Spinal QCT also provides information on fracture risk, but it is important to note the WHO T-score osteoporotic criteria cannot be applied in this situation.¹¹

Fracture risk using QCT of the spine is mostly interpreted using American College of Radiology criteria.¹² There are no data demonstrating a reduction in fracture risk by specific anti-osteoporotic treatment chosen based on QCT measurements. However, given the equivalency of hip QCT to hip DXA, there is no reason to doubt the utility of hip QCT in guiding therapy.¹³ The disadvantage of QCT remains the significantly higher radiation exposure compared with DXA,¹³ particularly at the hip. DXA of the spine and hip remains the recommended measurement for the diagnosis of osteoporosis and baseline BMD assessment. In some patients with moderate-to-severe osteoarthritic changes, spine QCT may have a particular advantage because it is less affected by osteoarthritic changes than DXA.

Quantitative ultrasound

Quantitative ultrasound of the heel and other sites can provide information on fracture risk.² However, quantitative ultrasound has not been demonstrated to provide information on absolute fracture risk and reduction of fracture risk by anti-osteoporotic treatment. DXA measurements at the spine and proximal femur are preferred for making therapeutic decisions and should be used, if possible. Quantitative ultrasound is not recommended as a routine diagnostic test for osteoporosis.

Biochemical markers of bone turnover

Increased biochemical markers of bone turnover in the blood and/or urine (e.g., serum C-terminal telopeptide or serum alkaline phosphatase) have been shown in trials to be independent risk factors for fractures in women and men.¹⁴ Bone turnover markers are useful markers of medication adherence and response to treatment, and may help guide choice of treatment. Short-term treatment-related changes in bone turnover markers account for a large proportion of the treatment effect of vertebral fracture.¹⁵ However, variability in analysis and lack of standardisation may reduce the utility of these assessments on an individual basis in routine clinical practice.

• For patients with ready access to DXA, a BMD measurement before commencing therapy is recommended. A normal or near-normal BMD despite existing fractures should prompt a more extensive work-up to exclude other causes of fracture.
• A normal BMD despite typical vertebral fractures also poses a problem regarding the usefulness of anti-osteoporotic treatments that have not been tested in such a population. Such discrepant findings should be resolved on an individual basis and may require specialist (e.g., endocrinologist, rheumatologist) referral.
• A history of high-trauma falls resulting in vertebral fracture can leave evidence of vertebral deformities that may not indicate underlying osteoporosis. In such situations, consultation with a specialist may be warranted.
• Conventional radiographs should not be used for the diagnosis or exclusion of osteoporosis.
• Evaluation of osteoporosis is based on the lower T-score of either the lumbar spine, FN or total hip.¹
• The BMD at the forearm may be measured by DXA, but caution is advised because there are limited data on its use in guiding therapy.
• Repeat BMD measurements may be performed to assess the efficacy of treatment and residual fracture risk or to assist in improving patient medication adherence.
• If possible, it is recommended repeat BMD tests are performed using the same instrument or at least the same make of instrument (manufacturer and model type) to improve the comparability of results in interpreting BMD change.⁶
• Relevant blood and urine studies should be obtained prior to initiating therapy if the medical history and/or clinical examination is suggestive of secondary osteoporosis, or the DXA Z-score is ≤–2.0 (i.e., two or more SDs different from age- and sex-matched controls).¹⁶
• If radiographs reveal one or more vertebral fractures typical of osteoporosis, BMD measurement may not be essential before starting medical therapy, if clinically appropriate. There are a limited number of scenarios in which meaningful evaluation of BMD is not possible (e.g., bilateral hip replacements and osteoporotic fractures in the lumbar spine region of BMD measurement [L2–4]). In such cases, it should be assumed that BMD measurement would have been low and that therapy is likely to be beneficial. Forearm BMD may be useful; however, its precise value has not been as well characterised as spine and hip BMD.