Dual-energy X-ray absorptiometry (DXA) to measure bone mineral density (BMD) for diagnosis of osteoporosis - experimental data from artificial vertebrae confirms significant dependence on bone size.

Introduction: The WHO definition of osteoporosis, and published BMD (Bone Mineral Density) references ranges, do not consider differences in bone size. Because it is a two-dimensional technique, and cannot measure bone depth, aBMD (areal BMD) measured using DXA (Dual-energy X-Ray Absorptiometry) is affected by bone size variability. Mathematical models have been devised to correct aBMD for bone size, but these are confounded by variations in soft tissue surrounding bone. Confirmation of the actual quantitative effect on clinical results for patients requires precise changes in bone size and mineral density, but studies of humans and animals are limited by the inability to precisely control these in natural bones. Purpose: The objectives of this experiment were to obtain precise, repeatable, quantitative data from sets of artificial vertebrae to confirm the dependence of aBMD on bone size in clinical practice, and to test the effect of applying corrections based on assumptions that the vertebrae were simple geometric shapes to produce corrected BMAD (Bone Mineral Apparent Density). Methods and materials: Four sets of artificial bones, each set containing four cylinders of different diameters but identical in height, were constructed by casting a mixture of epoxy resin and calcium carbonate powder into a mould. The cylinders were considered to be artificial vertebrae L1 to L4 so that all four in a set may be tested in a single scan. The X-Ray attenuation of the material used was varied between the sets, to represent differences in BMD. Each set of vertebrae was inserted into a soft-tissue analogue and DXA scanned, in the anteroposterior position, with the GE Lunar Prodigy and the Hologic Discovery. Results: The results verify the theoretical direct proportionality between aBMD and diameter, confirming that aBMD is significantly affected by bone size. Applying a BMAD correction, by assuming the vertebrae to be cylinders, reduced the effect of change in bone diameter by approximately two orders of magnitude to an insignificant level. Conclusion: This experiment has confirmed that BMD measured using DXA, accepted in clinical practice as the