6.3 Serial DXA Interpretation & Follow-up Intervals
Key Takeaways
Compare absolute BMD at the same site on the same scanner or on instruments with valid cross-calibration for that site.
Use matched evaluable vertebrae with the LSC for that combination. If both hips were scanned on both studies, use mean bilateral total hip BMD for monitoring.
A change at least as large as the applicable LSC reaches statistical significance. When the difference is below LSC, report no statistically significant change.
ISCD individualizes follow-up timing by purpose, age, baseline BMD, treatment, and factors associated with rapid BMD change.
Establish a valid comparison
Follow-up dual-energy X-ray absorptiometry (DXA) can help monitor therapy, identify unexpected loss, or inform treatment decisions. The International Society for Clinical Densitometry (ISCD) advises ordering a follow-up bone mineral density (BMD) measurement with a defined objective and when the result is likely to affect management. Begin interpretation by identifying the comparison study and checking whether the two measurements are technically comparable.
- Device. Prefer the same scanner, with known and acceptable in vivo precision. When devices differ, verify cross-calibration at the site and use the applicable inter-system least significant change (LSC). If there is no valid comparison, document that limitation and establish a new baseline.
- Image and analysis. Compare scan mode, positioning, anatomy, and region of interest (ROI). Review the original images and analysis. A change in hip rotation, a different side, or a new artifact can alter the result independently of bone change. Software can help reproduce the prior analysis, but the interpreter must verify the resulting ROIs.
- Same measurement. When both hips have been scanned at both time points, ISCD recommends mean bilateral total hip BMD for monitoring. The lowest valid hip score serves diagnostic classification. For spine monitoring, compare the same evaluable levels. When a valid comparison uses fewer than four levels, use an LSC for that combination. Precision generally worsens as vertebrae are removed. One vertebra does not suffice for monitoring.
- Absolute BMD. Subtract prior BMD from current BMD in g/cm². Scores rounded to one decimal place and calculated from reference databases are unsuitable for calculating serial change. To express the change as a percentage, divide the difference by prior BMD and multiply by 100. Use the applicable absolute LSC for the significance test when it is provided in g/cm².
Apply the correct threshold and report the result
Compare the magnitude of the absolute difference to the facility LSC for the specific measurement. A comparable change at least as large as the LSC is statistically significant at the confidence level used to calculate it. Report a significant increase or decrease in g/cm² and percent. When the change falls below the LSC, report no statistically significant change, without quantifying the subthreshold difference as a gain or loss in the report.
| Teaching calculation | Prior BMD | Current BMD | Absolute LSC | Interpretation |
|---|---|---|---|---|
| Spine after 18 months of teriparatide | 0.750 | 0.810 | 0.032 | Gain of 0.060 g/cm² (8.0%), statistically significant |
| Total hip during alendronate treatment | 0.820 | 0.808 | 0.025 | Difference of −0.012 g/cm² (−1.46%) is below LSC. Report no statistically significant change. |
Each BMD and LSC value in the table is in g/cm². The numerical subthreshold difference is shown to teach the calculation. On the spine comparison, check images for artifacts and interpret the gain with the treatment history. On the hip comparison, an apparent decrease below LSC cannot establish treatment failure. When an unexpected change reaches the threshold and scan quality has been confirmed, consider reassessing the patient and plan of care.
Individualize the follow-up interval
- Starting or changing treatment. In its U.S. clinician's guide, the Bone Health and Osteoporosis Foundation recommends BMD testing one to two years after initiating or changing osteoporosis therapy, then at intervals guided by the clinical situation.
- Rapid change may affect care. ISCD recognizes that shorter intervals can be useful with glucocorticoids, aromatase inhibitors, androgen deprivation, osteoanabolic therapy, malabsorption, severe inflammatory disease, prolonged immobilization, bariatric surgery, or surgical menopause. It does not assign a single interval to all of these conditions. In its specific U.S. glucocorticoid guidance, the American College of Rheumatology recommends reassessment every one to two years in defined ongoing-exposure groups and discusses yearly BMD until stable in patients at very high fracture risk.
- Stable patients and planned interruptions. Consider age, baseline BMD, therapy and new risk factors when setting the interval. Before and during a planned bisphosphonate interruption, follow-up BMD can inform reassessment. Avoid assuming that every patient should return annually or after a fixed number of years.
- New fracture or major risk factor. Reassess and arrange follow-up testing when indicated. Do not delay secondary fracture prevention while awaiting DXA.
ISCD Adult Official Positions · BHOF Clinician’s Guide · ACR Glucocorticoid-Induced Osteoporosis Guideline
A patient receiving denosumab has comparable lumbar spine BMD measurements of 0.780 g/cm² at baseline and 0.805 g/cm² at follow-up. The facility’s LSC is 0.030 g/cm². What should the report state?
The gain is significant because the BMD increased.
The patient’s response cannot be evaluated until the T-score improves by one standard deviation.
There is no statistically significant change at the lumbar spine.
The result proves treatment failure and warrants immediate discontinuation.
A 52-year-old postmenopausal woman with active rheumatoid arthritis is beginning prednisone 20 mg daily for at least six months. What do the ISCD Adult Positions say about the timing of her follow-up BMD testing?
Use a fixed three- to five-year interval after a normal scan.
Use a fixed two-year interval for all patients on systemic glucocorticoids.
Arrange yearly scans for life regardless of changes in treatment or risk.
Individualize the interval to clinical objectives and consider shorter follow-up because glucocorticoids can cause rapid BMD change.
Why should serial DXA change be calculated from BMD values rather than by subtracting T-scores or Z-scores?
Scores use reference databases and rounding, which can change the displayed score independently of measured BMD.
T-scores are only valid on peripheral instruments.
T-scores use a logarithm that cancels BMD units only when the scans are simultaneous.
Z-scores cannot be calculated for anyone aged 50 or older.
Sections you finish are checked off in the contents.