A user’s guide
Associate Professor Barry Chatterton, Mr
Dr Tony Roberts
Associate Professor Allan Need
The Department of Nuclear Medicine & Bone Densitometry has been involved in bone densitometry as part of a multi-disciplinary team involved with osteoporosis research since 1982. It continues to provide state of the art bone densitometry and general nuclear medicine services, which are available on direct referral to all medical practitioners.
This booklet was prepared in response to frequent requests from referring practitioners for information to allow the best use of results. It is meant to be a brief, general guide. If you have any further questions or comments to help us improve this guide or our service, please contact the department.
The department funds a large proportion of its clinical and educational activity from fees paid by private patients. Referrals are welcomed.
Associate Professor Barry Chatterton, Senior Director, Dept Nuclear Medicine & Bone Densitometry
The use of bone densitometry as a clinical tool in the diagnosis and management of osteoporosis is now established, but the choice of technique, interpretation of results, decision to institute therapy and time to repeat the examination are all subject to debate. The following may help in making an informed decision.
Increasing age & menopause are two major factors likely to be associated with increasing risk of osteoporosis and therapeutic intervention at this time is also most likely to be effective. Some authorities recommend a study in all women about the time of the menopause. In addition, 25% of hip fractures occur in men, and should be considered for testing where appropriate.
Most other risk factors are either uncommon or not discriminating and no combination of these allows good prediction of bone density (which therefore should be measured).
The risk factors recognised by the Medicare rebate schedule are (in both sexes) :
Additionally, The Australian Federal Government announced on 18-Dec-2006 that changes to eligibility criteria for alendronate and bone mineral density tests for patients over the age of 70 would apply from 1-Apr-2007. The following extracts are from the Minister for Healths speech:"From 1 April 2007, alendronate (in the form of Fosamax® Once Weekly and Alendro® Once Weekly) will be available on the PBS for patients with osteoporosis aged 70 years and over who are at high risk of fracture as measured by a bone mineral density test. Alendronate is already available on the PBS for patients with osteoporosis who have suffered a bone fracture following minimal trauma.
To coincide with this extension, bone mineral density tests for all patients aged 70 years and over will be covered by Medicare from 1 April 2007."
At present, universal screening is not recommended, and individualised assessment should be performed.
Patients receiving Strontium ranelate as a therapy for osteoporosis may exhibit a spuriously increasing "BMD".
The presently widely used techniques use ionising radiation, and have high levels of precision.
DXA has a very low radiation dose (a few days of background).
The existence of a Medicare rebate only for DXA and QCT of the axial skeleton has influenced the choice of technique.
The NIH Consensus Development Conference of 2000 stated "BMD measurements have been shown to correlate strongly with load-bearing capacity of the hip and spine and with the risk of fracture". Furthermore, measurement of BMD for each scan site is the best predictor of fracture at that site. A more detailed discussion of some of these techniques appears later in this guide.
The World Health Organisation (WHO) has recently reported on assessing fracture risk and developed the FRAX tool to determine an individual's 10 year probability of fracture at the hip and other sites based on BMD and clinical risk factors. Given a patient's clinical history, height & weight along with femoral neck T or Z score, a 10 year fracture risk is determined.
or Statistics! Statistics! The ZZZZZZZzscore!
Unfortunately, the interpretation of bone densitometry depends on a statistical understanding, which is sometimes a little different from the accepted way of interpreting other laboratory tests eg. a haemoglobin where the result is usually expressed as a range which encompasses 95% of normals (ie. ± 2 Standard Deviations, SD) with each range neatly defined according to age and sex-matched controls (see between arrows below).
This diagram represents the usual (normal) distribution of biological variables - bone density included. 66% of normal values will fall between -1 and +1 SD (and therefore 34% outside, 17% above and 17% below). Commonly, bone density reports plot the age matched mean ± 1 SD . This serves to make a larger proportion of the population seem abnormal (osteoporotic) than is usual for most diagnostic tests. The reason for this is somewhat arbitrary, but it is generally regarded as appropriate to treat more than the lowest 2.5% of densities.
For bone density at any age, the shape of the distribution is normal, but the absolute values fall after the peak bone density is reached in early adulthood.
The number of standard deviations away from the mean is represented in most reports as a Z score. A Z score of -1 means that the subject is 1 SD below the mean, and +2, 2 SD above the mean. The Z score may be expressed relative to normal people of the same age, or compared with young normals. It is the latter comparison (frequently designated the T score) which relates more closely to fracture risk. Medicare rebate eligibility is related to both T & Z score (see Rebates on page 14)
Although there is no true "fracture threshold" (a value of bone density above which fractures do not occur - sufficient trauma will break any bone), the risk increases considerably at bone densities which are less than 2 SD below the young normal mean. However, Medicare has decided that osteoporosis occurs at -2.5 SD or lower for the purpose of generating a rebate for repeat studies.
The graph adjacent is a schematic of a typical normal bone density curve from virtually any bone in women (menopause at 50). The horizontal lines are 1 SD (ie 1 "T" unit) apart, and the curve shows the mean value for age with heavier bars which are ± 1 SD, which is the range usually quoted in bone density reports, but for reasons mentioned above, excludes 29% of "normals". The thick horizontal line is 2.5 SD below the young normal mean, and therefore represents, by some definitions, the "Fracture Threshold", or the upper limit of the osteoporotic range. Note also the increased loss rate immediately after the menopause.
It can be seen that many elderly women may have a bone density that is 2.5 or more standard deviations below the young normal mean. This makes it "osteoporotic" by the above definition, despite being average for age. This corresponds with the increased fracture risk in the elderly. For comparative purposes, fracture risk approximately doubles for every SD reduction in bone density, so the fracture risk at -3 SD (ie T score = -3) is about 8 times greater (2x2x2) than the young normal, eg. O on the graph.
The elderly patient O marked above would therefore have a Z score of 0, but a T score (or young normal Z score) of -3. If osteoporosis is diagnosed on the basis of the young normal distribution, then about 45% of 50 year olds will have at least one site (hip, spine or forearm) affected, and virtually 100% of the very elderly will be osteoporotic. Bone density is normally higher in the heavier and taller patient. Many instruments will correct not only for age but also these factors, so occasionally one may see a heavy post menopausal patient with a greater bone density than a young normal, but nevertheless a corrected "Z" score of less than age matched!
Z and to a lesser extent T scores may be adjusted for weight or ethnicity. Densitometer manufacturers provide this as an option, and many DXA services utilise one or both adjustments.
Weight adjustment of the Z score is only provided on GE-Lunar systems, with weights above the population mean (about 70kg in women and 80 in men in US, UK and Australian reference populations) being raised when above the mean, and lowered when below the mean, but only in a range from 35-100kg. Outside of the weight adjustment range, the adjustment is limited to the maximum change. The magnitude of change per kg is small - typically 0.003 to 0.004 g/cm2 per kg (about 0.025 to 0.03 SD) , depending on scan site and gender. In practice, this means that a female patient with a mass of 100kg will have their expected BMD for age adjusted up by 0.12g/cm2 - or about 1 standard deviation. This may mean, for example, that the patient can have a T score of 0, but a Z score of -1 at age 50.
The NHANES study reported that non-Caucasian ethnic groups may have different normal ranges resulting in a higher expected BMD in the total femur (Looker A. et al, Osteoporos Int 1998; 8:468). Caucasian women were observed (on a Hologic DXA) to have an average peak BMD of 0.94g/cm2 with an SD of 0.12 in the total femur whilst African/American women had a peak BMD of 1.03g/cm2 with and SD of 0.16. This could lead to over treatment of non-Caucasian women.
An ethnic adjustment of the Z score is provided as an option by all DXA manufacturers, but Hologic and Norland also adjust the T score for ethnicity, whilst GE-Lunar do not.
The NOS(UK) and ISCD recommend that Z score adjustment be applied for ethnicity, and weight correction turned off, but the scientific basis for utilisation of weight correction is poorly documented, as is the ethnicity adjustment in the various Caucasian derived races.
Many practices have historically used these adjustments in different combinations. Suddenly changing the adjustment may cause difficulty for referrers using the values longitudinally to manage patients. It may be more appropriate to follow these recommendations in new DXA services but retaining existing adjustments in well established services but highlighting the adjustment.
It is possible to compare Z or T scores with percentiles. A Z or T score of +2 means that 97.5% of the reference population have less density
Z (or T) score
Some authorities recommend more intensive monitoring and preventative treatment of the patients with bone densities in the lowest quartile (25%) which correspond to a T score of < -0.7, particularly in peri-menopausal women. A more conservative approach is to use a T score of < -1.
Where a child’s bone density needs to be assessed, most densitometer manufacturers provide a normal range for children in the lumbar spine and the total body. However, the paediatric normal range shows relatively high variability due to differential growth rates and puberty. Bone Mineral Content should increase until maturity and may be a better prospective parameter than BMD.
It is recommended that children be assessed and monitored using total body bone densitometry and bone mineral content until physical maturity. At physical maturity, an additional lumbar spine and femur scan may be acquired and used to continue monitoring.
Different densitometers express the results using different units of the amount of bone present. The BMD (from which the Z and T score are usually derived) is used almost exclusively (in hip and spine densitometry) for both diagnosis and follow-up. In some forearm machines and the total body, the bone mineral content (BMC) may be used for follow up. Bone mineral density (BMD) is more commonly used to compare a patient with others of the same age by taking into account the patients size and the size of the measured region.
The bone density estimate gives a rough guide to the risk of future fractures, with a similar predictive value to that of blood pressure or cholesterol levels on stroke or myocardial infarction as an example. As well as the level of bone density estimated on an initial study, it maybe prudent to check for rapid bone loss using biochemical tests or even follow-up bone density measurements. Treatment with hormones in particular will need a full understanding of the patient’s wishes and her gynaecological status. As in any other field of medicine, the patient rather than a laboratory test is the subject of treatment.
A synopsis of management is included in the second half of this guide.
The 1996 Consensus Conference on Osteoporosis recommended the following :
T score (Young Normal BMD)
> + 1
No further measurements
0 to 1
Repeat measure 5 to10 years
-1 to 0
Repeat measure 2 to 5 years
-2.5 to –1
An journal article by M Whyte (J Clin Dens, Misinterpretation of Osteodensitometry with High Bone Density - BMD Z2.5, 2005, 8(1): 1-6) illustrated that very high Z scores may be due to underlying causes - other than fracture. The article describes a case study, and includes the following list of possible caused for very high bone mineral density.
|Dysplasias and Dysostoses||Metabolic Disorders||Other disorders|
|Craniodisphyseal dysplasia||Carbonic anhydrase II deficiency||Diffuse ideopathic skeletal hyperostosis (DISH)|
|Craniometaphyseal dysplasia||Fluorosis||Hypertrophic osteoarthropathy|
|Heavy metal poisoning||Ionising radiation|
|Frontometaphyseal dysplasia||Hepatisis C-associated osteosclerosis||Leukemia|
|High bome mass (LRP5) phenotype/disease||Hyper-vitaminosis A or D||Lymphoma|
|Infantile cortical hyperostosis (Caffey disease)||Hypoparathyroidism||Mastocytosis|
|Juvenile Paget disease
|Milk-alkali syndrome||Multiple myeloma|
|Metaphyseal dysplasia (Pyle disease)||X-linked hypophosphatemia||Osteomyelitis|
|Mixed-sclerosing, bone dystrophy||Osteonecrosis|
|Osteopetrosis (several types)||Paget’s bone disease|
|Progressive diaphyseal dysplasia (Engelmann disease)||Skeletal metastases (especially breast and prostate)|
With DXA technology, the usual projection for spine measurement is A-P and limited to the lumbar spine to eliminate confounding structures. The scan image will contain posterior elements (laminae, spinous processes), as well as the vertebral body in the measurement. Osteophytes and more anterior calcification (eg aorta) will also be included.
In the proximal femur ("hip"), the computer produces semi-automated regions of interest. The femoral neck and trochanteric regions are self-explanatory, Ward’s triangle is shown on the scan as a rectangular region! It represents the region in the femoral neck with the fewest trabeculae, and therefore the least density. For this reason, it is a sensitive indicator of risk, but exact placement (& re-positioning) of regions of interest is difficult and therefore it is the least reproducible of these measurements. More recently, the sum of these areas – the total hip – has been shown to have the best precision for followup purposes.
Total body measurements divide the skeleton into arbitrary regions, which have some diagnostic value, and may also report other parameters of body composition such as lean body mass, and total body fat. The total body BMD is the value most commonly used.
Typical reproducibility of bone density estimations is 1% (one of the most precise measurements in medicine)! This means that (statistics again!) 67% of measurements repeated on the same day will fall within ± 1%. To be 99% sure that a change has really occurred, a change of three times this magnitude should occur for single serial measurements. Real annual changes of this degree occur at the time of the menopause or in patients on corticosteroids. Therefore in the individual, there is little point in repeating the study in less than 12 months and often 2 years. This limitation of course may not apply in research studies, where averaging from many patients allows meaningful results from shorter follow up periods.
Referrers should be aware that patients treated with Strontium ranelate (PROTOS or PROTELOS) may have significant increases in BMD, with up to 50% of the change in BMD being due to Strontium incorporation into the bone due to the higher atomic number of Strontium affecting DXA measurements. The magnitude of change in BMD may, therefore, not be accurately determined with this treatment.
There is incomplete agreement between measurements at different sites, partly due to machine errors in measurements, and selection of the normal ranges, but also due to real biological variation. (One may have a "thick" waist and thin wrists, and similar variation in bone density is also explicable.) Measurements at each site may predict fracture to a similar extent. It is probably prudent to use only a predetermined site or sites for measurements.
There are 3 major brands of densitometer currently operated in Australia (Hologic, LUNAR & Norland). Unfortunately, the calibration of these 3 instruments differ, resulting in different BMD and BMC values, and therefore preventing direct numerical comparison of results. The normal ranges on these instruments has also been defined from different populations.
Studies have been performed which allow for cross calibration of the different instruments, but this is of limited value, as there will be some discrepancy (due to statistics again!). We can provide some of the cross calibration equations on request, but as new models are introduced, the number of conversion factors increases.
When performing follow up studies on a patient, it is strongly recommended that the same brand of densitometer (and for preference, the same machine) be used to ensure the most valid comparisons.
The highly computed modern densitometers produce an excess of data for clinical use, and the raw report is often difficult to quickly assess. Values are often given for BMC, BMD, Z scores & percentiles (relating both to young normals and age matched controls), for each of the lumbar vertebrae and permutations and combinations of these, and in the hip, neck, Ward’s triangle and trochanteric region. Using the mean of several vertebrae (often L2-L4), the femoral neck or total hip probably obtains the best information. In the analysis of a study the operator will exclude vertebrae that appear crushed or have an excessive amount of sclerosis relating to degenerative change. As Z scores will be different for the same density as the patient ages, the BMD for DXA (or forearm BMC) should be used in follow-up.
The RAH has developed an automated reporting system which provides uniform interpretation, and emphasises the most important data, but may not take into account unusual clinical situations. A sample report (as PDF) is provided below.
Analysis of the density of the lumbar vertebral body imaged in the lateral position was seen as a method to reduce errors due to arthritic changes and calcification in the aorta. Unfortunately, in patients who would best benefit from this study it has been found that overlying pelvis and ribs and the greater path of soft tissues that the X-ray beam must penetrate, limit both the precision and number of vertebrae which may be studied. Routine use is not recommended, but it may have some value in the elderly. Lateral projections are obtained in morphometry, which is an image processing rather than a densitometric technique.
Specialised ultrasound machines produce an index of the bony properties of the heel, tibia, phalanges and other sites. The degree to which the sound is attenuated gives an estimate of the quality of the bone (ie trabecular structure) and the speed of sound through the heel may indicate the amount of bone in the path of the sound beam. The combination of these results has been shown to stratify patients into their relative risk of future fracture at the hip almost as well as DXA at the femoral neck.
The result is not strictly bone density, and the precision is such that follow-up of treatment is not yet possible. Because of the imprecision in the method, and also because it measures a different bone, the correlation between DXA measurements at other sites is only moderate. Medicare does not reimburse this study. Future developments may make it more applicable.
Due to the lack of ionising radiation, quantitative ultrasound tests are often performed on a self referral basis.
What are the differences between ultrasound tests for osteoporosis and that available from DXA service providers?
Vertebral crush fractures are a common, usually diagnostic, feature of osteoporosis. These have been previously diagnosed using standard X-ray techniques. Vertebral compression fractures are independent predictors of future fracture and may qualify patients for specific therapy under the PBS .
Some of the more recent generation of bone densitometers use multiple detectors to produce a high-resolution image from which semi-automated measurements of vertebral dimensions are obtained. The ratio of the vertical heights of the anterior and posterior edges of a vertebral body allows the diagnosis of a wedge fracture to be made, for example. This technique allows vertebral fracture to be diagnosed with much less radiation exposure than a standard radiograph. Morphometry may represent an independent method of following the progress of osteoporosis.
The output gives vertebral heights and A/P ratios relative to expected standard deviations. A single 4 SD or two or more 3 SD differences are generally regarded as a significant fracture. The percent height loss is also displayed.
NOTE: DXA morphometry measurements do not meet the HIC requirements for PBS applications.
The aim in preventing or treating osteoporosis is to stop further bone loss, so the best form of monitoring is sequential bone density. However, because of the relationship between the precision of bone densitometry (1%) and the average annual rate of loss (1.2%) it is necessary to wait a year or more after treatment begins for densitometry to confirm it is working. Biochemical markers for bone resorption & formation show effects within weeks of starting treatment. Biochemical tests can also be used to select the most appropriate form of treatment. Intestinal malabsorption of calcium is common in osteoporosis and responds to calcitriol therapy (calcium absorption can be measured by a radiocalcium absorption test at the Institute of Medical & Veterinary Science). Osteoporotic women with a high fasting urinary calcium excretion appear to best respond to hormone therapy. Women with low bone turnover may respond best to anabolic steroid therapy.
Tests currently being used in the management of osteoporosis include :
Women more than 1 SD below (T < -1) the young normal mean for bone density at the menopause can rapidly become osteoporotic since women lose over 10% of their bone mass within 5 years of the menopause. We recommend these women receive immediate prophylaxis to prevent further bone loss.
Women with bone density within 1 SD (T > -1) of the young normal mean at the menopause should have their bone density remeasured after two years in case they are losing rapidly. Perimenopausal women in the top quartile of bone density (T > 0.7) do not need to have another measurement for at least five years.
Women with osteoporotic fractures should definitely be treated and women with a bone density more than 2.5 SD below the young normal range should probably be treated, but it is worth remembering that the lifetime risk of fracture depends on the number of remaining years of life. Thus there is much more value (and less inconvenience) in treating someone aged 65 with a low bone density than in treating someone aged 85. It is also worth remembering that 75% of hip and spine fractures occur in women with a bone density more than three standard deviations below the young normal mean.
Osteoporosis in men is less likely to be treated, although many of the risk factors are common to both sexes, and 25% of fractures occur in men.
General measures should include:
Specific medications to prevent further bone loss include:
A calcium supplement of 1 gram daily (preferably taken at night) will often arrest bone loss but may be ineffective if intestinal calcium absorption is impaired. Patients who excrete less than 3mmol per day of urinary calcium are likely to have calcium malabsorption.
Oestrogens will stop bone loss, help prevent heart disease, and may be indicated for women with postmenopausal symptoms. There may be a small increase in risk of breast cancer. They must be given with Progestagens in women who still have a uterus to protect against endometrial cancer.
Available oestrogens include:
Oestrogen skin patches or implants are available when oral oestrogens are not tolerated.
The progestagen, norethisterone (Primolut-N, 2.5mg daily), can prevent bone loss in its own right and investigations are under way to assess the relative benefits and side effects.
Calcitriol (Rocaltrol, 0.25m g b.d. or 0.25m g daily with
1g Calcium) is available on the Pharmaceutical Benefits List (authority required)
for patients with vertebral compression fractures or other fractures occurring
with minimal trauma.
It has been shown prospectively to reduce the incidence of fracture in women.
Etidronate (Didrocal, 400 mg daily for two weeks every three months), Alendronate (Fosamax and Fosamax Plus, 10mg daily or 70mg per week) and Risedronate (Actonel and Actonel Combi, 5mg daily or 35mg per week) are effective and are on the PBS (authority required for similar indications) for both men and women.
Osteonecrosis of the jaw has been reported as a result of bisphosphonate therapy. Whilst the incidence is very low, it is a potential serious side effect of such therapy. Several review articles and comments are available:
Please refer to these articles for further infornation.
Selective oestrogen receptor modulators are now available (PBS authority required and only for post menopausal osteoporosis). Raloxifene (Evista, 60mg daily) has been shown to reduce the risk of fractures and may also be shown to reduce the risk of breast cancer.
Anabolic steroids are useful in patients with low bone turnover as they appear to increase bone formation. They are limited by virilising side effects. After six months of therapy (50mg of nandrolone decanoate - Deca-Durabolin – given i.m. every 4 weeks) a significant risk of voice change and hirsutism exists.
Calcitonin injections are effective for pain relief following acute vertebral fracture, but are expensive and include side effects such as flushing and nausea. Calcitonin is not currently approved for the treatment of osteoporosis in Australia.
This compound has recently been approved by the PBS for treatment of post-menopausal
osteoporosis in Australia. Strontium ranelate has been reported to reduce
the risk of new vertebral fractures by about 30 percent over four years. The
strontium is reported to incorporate into new bone, not old bone, probably
changing bone architecture.
Treatment requires consumption of a 2g sachet of Strontium ranelate powder in water daily.
Patients treated with Strontium ranelate will exhibit significant increases in BMD. Changes of up to 4% per annum in the lumbar spine and 2% per annum at the femoral neck have been reported with DXA measurement. It is recommended that BMD only be used to determine compliance with therapy as changes in BMD cannot be accurately determined.
Bisphosphonates, Calcitriol, SERMS, HRT and Anabolic steroids have a role in the prevention and treatment of steroid induced osteoporosis.
Once therapy is commenced, it is best to ensure bone loss is controlled. A reduction of markers of bone reabsorption (urinary cross-links) is a useful early marker. A repeat bone mineral density measurement after twelve months to two years will usually suffice unless bone loss is expected to be rapid (eg. during corticosteroid therapy).
Patients with side effects from treatment, poor response to treatment or without access to densitometry services may benefit from assessment by specialists in skeletal medicine.
Interventions to reduce the risk of falling include:
A number of research studies have shown that use of padding around the hips significantly reduces the risk of femoral neck fractures in those at risk of falling. The Health Promotion Unit of the Royal Adelaide Hospital can supply a Hip Safety kit, which includes:
Further information and assistance may be obtained from a number of sources listed at the end of this document
The schedule of benefits as of 1 November 2008 are reprinted below.
Details may be found in the Medicare Benefits Schedule book or online. The following notes should only be considered after referral to this section.
12306 DXA/12309 QCT
12315 DXA/12318 QCT - monitoring (where result will assist management)
[one service only in a period 24 consecutive months]
12321 DXA monitoring
12323 DXA measurement for a person 70 years or older
Note - screening at the time of the menopause remains specifically excluded.
Sample DXA report
For Bone Densitometry, including AP Spine, femoral neck, forearm, total body DXA, lateral morphometry, ultrasound and rural Mobile DXA service, please contact the staff in the Dept Nuclear Medicine & Bone Densitometry at the Royal Adelaide Hospital.
Phone: (08) 8222 5408
E-mail: Email a query
For biochemical diagnosis & management,
Dr Tony Roberts, Dept of Endocrinology, Royal Adelaide Hospital
Phone (08) 8222 5520
Relevant RAH clinics :
Osteoporosis (08) 8222 5586
Endocrine (08) 8222 5586
Women’s Health Centre (08) 8222 5587
For information on fall prevention services and fall management:
Health Promotion Unit, Royal Adelaide Hospital
Phone (08) 8222 5193
Falls and Injury Assessment Clinic, Repatriation General Hospital
Phone : (08) 8275 1103
Flinders Medical Centre Falls and Injury Prevention Service
Phone : (08) 8204 5511
Homefront (for Dept of Veterans Affairs)
Phone General Enquiries : (08) 8290 0478
Veterans only : 1800 801 945
The full NIH consensus statement on Osteoporosis is available at: