Common Questions
How precise are repeat measures with DXA?
The DXA technique has a high level of precision: On repeat measures using the same machine, values will vary by only 1 to 2 percent, making it one of the most precise tests in medicine. Although small, this variation impacts:
- Interpretation of statistically important differences
- Interval of repeat DXA scanning
For example, assuming a 1 percent variation in measurement (precision error rate), two DXA results would have to differ by more than 2.77 percent before you could be confident that a real change had occurred (see table below). Therefore, if the bone density has decreased by less than 2.77 percent, you cannot know whether this is due to true bone density loss or simply is due to the error rate of the machine.
This also impacts the repeat scanning interval. If your patient loses bone density at the relatively rapid rate of 3 percent per year, such as may occur in postmenopause or during steroid therapy, at least a year should elapse before the next scan (1). For the slower standard rate of loss, 1 percent per year, an even longer interval may be appropriate, e.g., 3 to 5 years.
| Precision Error of Technique | Statistically Meaningful Difference* | Assumed Rate of Bone Loss (per year) | Appropriate Follow-up Interval (in years) |
|---|---|---|---|
| 1% | +/- 2.77% | 1% | 3 years |
| 1% | +/- 2.77% | 3% | 1 year |
| 2% | +/- 5.54% | 1% | 5 to 6 years |
| 2% | +/- 5.54% | 3% | 2 years |
* This number reflects the amount two scans would have to differ for you to be sure a real change has occurred, with 95% confidence. Adapted from: Levis S, Altman R. Bone densitometry: Clinical considerations. Arthritis and Rheumatism. 1998; 41:577-587.
- National Osteoporosis Foundation. Osteoporosis: Cost-effectiveness analysis and review of the evidence for prevention, diagnosis and treatment: Executive Summary. Osteoporosis International 1999; 1 Suppl: 3-78.