INTERPRETATION OF
BONE DESITOMETRY MEASUREMENTS

THOMAS HIGGINS, MD

Dual photon desitometry measurements are very useful in the assessment of fracture risk, age related bone loss and the response to osteoporosis therapy. Currently dual photon x-ray (DPX) is the technique of choice for bone densitometry. Dual photon isotope techniques using gadolinium and Quantitative CT (QCT) are older techniques which have generally been replaced by DPX. Generally the hip & spine are measured. Measurement of peripheral sites like the wrist are useful in certain circumstances. Heal measurements are frequently used for screening but my experience has been that heal measurements are unreliable for screening and may both over and under estimate fracture risk.

Normal bone density increases with age to a peak bone mass in the late twenties or early thirties and then there is a gradual age related bone loss in both sexes. Regional differences in relative bone density in the same individual are routinely seen in both normal and osteoporotic patients.

A difference of between 3 to 5% in any two bone density measurements is required to be certain of a real change although smaller changes can be indicative of a trend. Since changes in bone density are generally slow and relatively large changes are necessary to be absolutely sure the change is real, repeat bone density measurements are usually not performed any more frequently than once a year. Use of the same densitometer or at least the same equipment manufacturer is desirable.

The actual bone density scan contains an image of the bone being measured which is useful in checking for positioning or artifactual errors. There is also a set of numbers and calculations used in compiling the final report. Bone density is reported as gm/cm2 and the results are compared to a large population database for each sex. The population comparisons are reported as both standard deviations above or below a predicted peak bone mass and as standard deviations above or below an average person of the same age and sex.

The standard deviation from predicted peak bone mass is reported as the "T score" when using Hologic densitometers and the "young adult Z score" when using Lunar densitometers. In general, fracture risk is felt to increase by 1.5 to 2.5 times for each standard deviation below the predicted peak bone mass. A "Z score" when using Hologic or "age matched Z score" when using Lunar equipment represents the age matched standard deviations and gives an indication of age related bone loss. It is important to note here that a bone density which is "normal for age" does not exclude a significant fracture risk in older individuals due to normal age related bone loss. Another comparison often made in reports is to the "fracture level". The fracture level is an arbitrary bone density above which fragility fractures are rarely seen.

The World Health Organization has defined normal bone density as +/- 1 standard deviation from predicted peak bone mass, osteopenia to be 1 to 2.5 standard deviations below peak, osteoporosis to be 2.5 standard deviations or more below peak and severe osteoporosis to be a density 2.5 standard deviations or more below peak together with evidence of a fragility fracture. It is important to note that osteoporosis is defined by the "T score" in Hologic reports and the "young adult Z score" in Lunar reports. It should also be stressed that the WHO criteria are not meant to be thresholds for treatment. For instance a younger person with a particular bone density may have a more significant problem than an older person with the same bone density because of a greater potential for future age related bone loss in the younger person. Decisions regarding treatment must be made on clinical grounds and not on density measurements alone.

Density measurement of the lumbar spine between L1 and L4 is generally the most reliable site for density measurements, especially in younger persons. There are, however, many sources of artifact in lumbar area such as the presence of osteoarthritis, compression fractures, prior back surgery, severe scoliosis or aortic calcifications. In general, artifacts are most frequently seen in patients over sixty years of age. All artifacts in lumbar density measurements increase the bone density except for a lamenectomy which will decrease it. Careful attention should be directed toward any artifact prior to final interpretation of a bone density report. Occasionally only one or two of the vertebral bodies will be able to be interpreted. If artifacts are not likely to be an issue, ordering the lumbar spine alone may be adequate.

Hip density measurements are less subject to artifactual errors than are lumbar measurements but are also less accurate because they are more subject to technical and positioning problems. When serial measurements of the hip are performed, they should be done of the same hip wherever possible. Hip density measurements provide density values for the femoral neck, the trochanteric region, the intertrochanteric region and Ward’s Triangle. There is also a total hip density measurement in most reports which is a weighted average of the trochanteric, intertrochanteric and femoral neck areas. In general the femoral neck area is considered to be most accurate for prediction of fracture risk but the total hip density is considered to be the most reproducible for serial testing.

Occasionally, bone density measurements of other sites including the wrist, lateral spine, os calcus and whole body are seen but these are relatively unusual. It is often said that the more sites that are assessed, the better the overall assessment of the bone density status and fracture risk. This statement must be balanced against the time and cost involved in testing. Currently a DPX using pencil beam technology takes 30 minutes for each scan. Newer fan beam DPX technology will reduce the time factor but will probably not significantly reduce the cost.

Single measurements of bone density should be considered to assess fracture risk in individuals with risk factors for fragility fractures. Serial measurements should be considered to assess age related bone changes and when following response to therapy. As noted previously, serial measurements at intervals of less than one year are generally not useful. In some patients, especially those with severe osteoporosis or in whom compliance is in question, it may be necessary to have more timely reassurance that therapy is effective. Urinary testing for collagen cross links is now available and can be considered in such situations. Levels of cross links fall rapidly after initiation of effective therapy. Testing can be done on a spot urine but the specimen should be a "second morning void" because of a large diurnal variation in excretion rates.

SUMMARY

1) Fracture risk and the labeling of bone density (normal, osteopenia or osteoporosis) is based on a comparison of the density obtained to the predicted peak bone mass. This measurement is called the "T score" or "young adult Z score" depending on the equipment utilized.

2) Fracture risk increases by 1.5 to 2.5 times for each standard deviation below peak bone mass.

3) Comparison of bone density to age related normals gives an indication of age related bone changes. Values "normal for age" do not mean there is not a significant risk of fracture in older patients. "Z score" or "age matched Z score".

4) Slightly different degrees of bone density at different sites is not unusual as osteoporosis is not uniform throughout the body and different disease states affect different sites differently.

5) Lumbar desitometries are more accurate but more subject to artifact. Hip desitometries are less accurate but less subject to artifact.

6) In hip densitometry, the comparison of femoral neck density to predicted peak bone mass is most closely related to fracture risk and the total hip density is most useful in serial measurements.

7) Bone density measurements should generally not be preformed any more frequently than yearly.

8) Screening bone density measurements should be considered in persons with risk factors for bone loss such as perimenopausal state, secondary amenorrhea, advanced age, eating or gastrointestinal disorders, COPD, prolactinoma, hyperparathyroidism, hyperthyroidism or thyroid hormone use, hypercortisolism or prednisone use, immobilization or disuse, kidney stones or hypercalciuria, excessive tobacco or alcohol use, etc. Screening in women taking estrogens may be useful as not all patients respond to estrogens and there is a high non compliance rate with ERT.

9) Serial bone density measurements should be considered in patients with risk factors for bone loss not on therapy and in patients being treated for osteoporosis.

10) Treatment decisions must be based on clinical judgment and not on bone density values. For instance, levels of bone density which are not low enough to qualify for a diagnosis of osteoporosis may still be quite significant, especially in younger individuals.