Is changes in body weight an error source on Hologic DXA systems?

Can you provide citations (papers or abstracts or “personal communications”) on the effect of changes in fat mass on the accuracy of BMD by DXA, using Hologic 4500A scanners?

1. You must be very careful to exclude scans with pannus overlay on hip scans. If you did not systematically check for this in an obese population, this can generate unpredictable hip values.
2. There will be some changes in accuracy for DXA due to large changes in weight. Evans (1) did a nice study on QDR-4500As by overlaying 1kg of fat over the scan regions. These lard packs were about 1 “ thick. The spine values decreased by 1.6% with the lard packs and 0.7% decrease at the lumbar spine. Thus, when patients get fatter, there is seemingly an error that makes the bone density appear to decrease.
3. If we put this in perspective, Evans added an inch to the apparent diameter of the patient. If we think of this in belt sizes, this would be an increase in belt size of 2*pi*(r2-r1) = pi ~ 3”. This is a pretty good weight change, a 3” lengthening of the belt.
4. Tothill (2,3) did work primarily looking at whole body scans and found a similar effect of increasing bmd with decreasing weight.

At the 12-month visit, there is increased bone loss at the hip, but not the spine or whole body, in the weight loss intervention group.  The SC is concerned that the bone loss might be an artifact of the weight loss.  

1.       For Hologic, the error seems to be a decrease in weight causes an error by increasing BMD.

2.       With significant weight loss, the skeletal loading is greatly diminished and you would expect there to be a decrease in bone density just associated with lower skeletal loading.

3.       Total body bmc is very very highly correlated with total body lean mass, even moreso with lean skeletal muscle mass. I would look at the correlation of LSMM and TBBMC at base line and at last visit, as well as the correlation of change in LSMM and change in TBBMC. If you have patients that lost fat but not lean, you could use this as a discriminator between error and truth. i.e. no fat changes but lean changes, BMC changes might be just artifactural, no lean changes but fat changes, BMC changes may be all error.

I hope this helps!!!

John

Reference:

1.         Ellen M. Evans, Mina C. Mojtahedi, Renee B. Kessinger and Mark M. Misic Simulated Change in Body Fatness Affects Hologic QDR 4500A Whole Body and Central DXA Bone Measures Journal of Clinical Densitometry Volume 9, Issue 3, July-September 2006, Pages 315-322

2.         Peter Tothill  Dual-Energy X-Ray Absorptiometry Measurements of Total-Body Bone Mineral During Weight Change Journal of Clinical Densitometry, vol. 8, no. 1, 31–38, 2005

3.         Tothill P, Avenell A. Errors In Dual-Energy X-Ray Absorptiometry Of The Lumbar Spine Owing To Fat Distribution And Soft Tissue Thickness During Weight Change. Br J Radiol. 1994 Jan;67(793):71-5.