Comparing BMD results between two similar DXA systems using the generalized least significant change.

One of the long-standing frustrations of clinical densitometry practice is not being able to compare bone mineral density (BMD) measures taken on different densitometers and know if the difference represents a true change. Recently, a method for comparing measures on different systems was published. This method, called the generalized least significant change (GLSC) requires after a procedure to quantify the precision of both systems as well as the in vivo cross-calibration relationship when there is a difference in the technology of the systems. We followed this procedure when a Hologic QDR-4500A was replaced with a Hologic Discovery/W even though these systems would be similar if not identical for hip and spine measures. Thirty participants were scanned twice on each system at the hip and spine. We found that the precisions of each system were similar and the differences in the average BMD values from the spine phantom and in vivo measures for the total spine, total hip, and neck regions were less than 1%. However, the correlation coefficients ranged from 0.96 to 0.98. The magnitude of change needed for significance was typically twice as large for intersystem scan (6-10%) comparisons than intrasystem (3-6%). In summary, we have presented an example of how the GLSC is calculated and used in a clinical practice. The results show that there is a substantial loss in sensitivity to change when comparing scans taken on different systems even in this case of similar technology. A revision of the International Society for Clinical Densitometry's policies for comparing scans from systems of the same technology may be appropriate.