August 1999 Bulletin


Bone density tests

John D. Kaufman, MD, explains bone density test to Helene Kaufman, as Carol Hauenstein, technologist at computer, gets ready to begin test.

By John D. Kaufman, MD

The history of bone density measurement goes back to the 1940s. Attempts to measure bone density at that time utilized plain radiographs. Since demineralization of bone is not apparent on a plain X-ray until about 40 percent of the bone has been lost, different methods of bone density determination were developed by measuring size or shape of different anatomical structures.

Grading systems were developed based on the appearance of trabecular patterns. The Singh Index was based on the trabecular pattern in the proximal femur. Radiographs were graded 1 through 6 based on the disappearance of the normal trabecular pattern usually seen in the femoral neck. Studies showed a correlation between a Singh Index of less than 3 and fractures of the hip, wrist and spine.

Radiographic absorptiometry was developed during the late 1980s as an easy way to determine BMD with plain X-rays. An X-ray of the hand is taken incorporating an aluminum reference wedge. This film is then analyzed and the density of the bone compared to the reference wedge. The correlation between the RA values and the actual density is excellent.

Single photon absorptiometry

In the early 1960s, a new method of determining bone density using a radioactive isotope was developed-single photon absorptiometry (SPA). A single energy photon beam was passed through bone and soft tissue to a detector. The amount of mineral in the path could then be quantified. The amount of soft tissue the beam had to penetrate needed to be small so the distal radius was usually utilized. SPA measurements are accurate and the test usually takes about 10 minutes. The radioactive source gradually decays, however, and must be replaced after a time.

Dual photon absorptiometry

The principle of dual photon absorptiometry (DPA) is the use of a photon beam that has two distinct energy peaks. One energy peak will be more absorbed by soft tissue and the other by bone. The soft tissue component then can be mathematically subtracted and the BMD thus determined. For the first time, BMD of the spine and proximal femur could be determined. Although accurate for predicting fracture risk, precision is poor due to decay of the isotope, and the machine has limited usefulness in monitoring BMD changes over time.

Dual-energy X-ray absorptiometry

Dual-energy X-ray absorptiometry (DXA) works in a similar fashion to DPA, but uses an X-ray source instead of an isotope. This is superior because the radiation source does not decay and the energy stays constant over time. Scan times are much shorter than with DPA and radiation dose is very low. The skin dose for an AP spine scan is in the range of 3 mrem. DXA scans are extremely precise. Precision in the range of 1 percent to 2 percent has been reported. DXA can be used as an accurate and precise method to monitor changes in bone density in patients undergoing treatments. The first generation DXAs used a pencil beam type scanner. The X-ray source moves with a single detector. Second generation machines use a fan beam scanner that incorporates a group of detectors instead of a single one. These machines are considerably faster and produce a higher resolution image. DXA has become the "gold standard" for BMD measurement today.

Quantitative computed tomography

Measurement of BMD by quantitative computed tomography (QCT) uses most standard CT scanners with software packages that allow them to determine bone density in the hip or spine. This technique is unique in that it provides for true three-dimensional imaging and reports BMD as true volume density measurements. The advantage of QCT is its ability to isolate the area of interest from surrounding tissues. It can, therefore, localize an area in a vertebral body of only trabecular bone leaving out the elements most affected by degenerative change and sclerosis. The QCT radiation dose is about 10 times that of DXA and QCT tests may be more expensive than DXA.

Peripheral bone density testing

Lower cost portable devices that determine bone mineral density at peripheral sites, such as the radius, the phalanges or the calcaneus are being utilized more and more for osteoporosis screening. The advantage of these devices is the ability to bring bone density assessment to a large portion of the population who otherwise would not be able to have the test. These machines cost considerably less than those that measure the hip and spine.

The Norland pDXA is a true dual energy X-ray device that is dedicated to measure BMD in the distal radius. It is portable and weighs about 59 pounds.

Hologic makes a quantitative ultrasound (QUS) unit, called the Sahara. This device measures the speed of sound and the attenuation of the sound signal through the calcaneus. It gives an "estimated" BMD. In two recent prospective studies of postmenopausal women over age 65, QUS predicted hip fractures as well as BMD measurements of the hip by DXA.

One of the problems with peripheral testing is that only one site is tested and low bone density in the hip or spine may be missed. This results because of discordance of bone density between different skeletal sites. Although these peripheral machines are considered accurate, there have been doubts raised about their precision. The reproducibility of peripheral machines may not be good enough to monitor patients undergoing treatment for osteoporosis.

Discordance in BMD among various skeletal sites is more common in the years just following menopause. BMD may be normal at one site and low at another site. In these early postmenopausal years, bone density in the spine decreases first because the turnover in this highly trabecular bone is higher than other skeletal sites. Bone density at various skeletal sites begins to coincide at about age 70.

In early post menopausal women, therefore, up to the age of about 65, the most accurate site to measure is probably the spine. In older women over the age of 65, the concordance of skeletal sites is much closer and it may not make much difference which site is measured. Caution must be used in interpreting spine scans in elderly patients because of degenerative changes falsely elevating the BMD values. Measurements are, however, mostly site specific and the most accurate predictor of fracture risk at any site is a bone density measurement at that site.

The peripheral devices are, at present, good screening devices because of their portability, availability and lower cost, but patients may still need central testing, even in light of a normal peripheral test. The following can be used as a guide to which patients with a normal peripheral test should be tested centrally:

Bone density report interpretation

The main purpose of obtaining a bone density test is determining fracture risk. The bone mineral density correlates very well with risk of fracture. It is more powerful in predicting fractures than cholesterol is in predicting myocardial infarction or blood pressure in predicting stroke.

  1. Postmenopausal patients not on hormone replacement therapy (HRT), concerned about osteoporosis and concerned about prevention, who would consider HRT, bisphosphonates or SERMs, if a low bone mass is discovered
  2. Maternal history of hip fracture, smoking, tallness (more than 5'7") or thinness (less than 125 pounds)
  3. Patients on medications associated with bone loss
  4. Patients with secondary conditions associated with low bone mass (hyperthyroidism, posttransplantation, malabsorption, hyperparathyroidism, alcoholism, etc.
  5. Patients found to have high urinary collagen crosslinks (NTx, etc.)
  6. History of previous fragility fracture

Adapted from Miller PD et al, J of Clin

Densitometry, vol. 1, no3, 214, 1998


The T-score is the number of standard deviations (SD) above or below the young adult mean. The young adult mean is the expected normal value for the patient compared to others of the same sex and ethnicity in a reference population the manufacturer builds into the DXA software. It is approximately what the patient should have been at their peak bone density at about age 20.

As a general rule, for every SD below normal the fracture risk doubles. Thus, a patient with a BMD of 1 SD below normal (a T-score of - 1) has 2 times the risk of fracture as a person with a normal BMD. If the T-score is -2 the risk of fracture is 4 times normal. A T-score of -3 is 8 times the normal fracture risk. Patients with a high risk can then be treated and future fractures prevented.

Different manufactures of DXA equipment report the hip BMD differently. In LunarTM machines, the average value for the femoral neck (designated as "Neck" in the report) is usually the most accurate for the hip. In HologicTM machines, the "Total" value should be used. In the spine, the most accurate T-score is the average of the first four lumbar vertebrae (L1-L4) in all manufactures of DXA machines.

There are other factors that determine fracture risk such as a person's eyesight, balance, leg strength, and conditions that might cause them to fall more easily. Age itself is an independent risk factor for fracture independent of bone density. Anyone with osteoporosis that has had a previous fragility fracture is considered to have severe osteoporosis and has a very high risk for future fractures.


The Z-score is the number of standard deviations the patient's bone density is above or below the values expected for the patient's age. By comparing the patient to the expected BMD for his or her own age, the Z score can help classify the type of osteoporosis. Primary osteoporosis is age-related where no secondary causes are found. Secondary osteoporosis occurs when underlying agents or conditions induce bone loss. Some common causes of secondary osteoporosis are thyroid or parathyroid abnormalities, malabsorption, alcoholism, smoking and the use of certain medications especially corticosteriods.

A Z-score lower then -1.5 should make you suspicious of secondary osteoporosis. If secondary causes are suspected, the patient should usually undergo further work up including laboratory testing to find out if there is an underlying reason for the osteoporosis. This is important because treating the underlying condition may be necessary to correct the low bone density.

John D. Kaufman, MD, is chairman of the Academy's Osteoporosis Interest Group and is in private practice in Santa Clarita, Calif.

Home Previous Page