Comparison of dual-energy X-ray absorptiometry machines for measuring fat distribution changes of HIV-associated lipodystrophy.

BACKGROUND Dual-energy X-ray absorptiometry (DXA) is regarded as a useful objective method for diagnosis of HIV-associated lipodystrophy (LD). OBJECTIVE To determine whether the use of different DXA machines may affect the diagnosis of LD. METHODS Body composition in 24 HIV-infected patients, 12 of whom had clinically diagnosed LD, was measured on two DXA machines (Lunar and Hologic). RESULTS Hologic gave significantly higher values than Lunar DXA for total body fat percent (22.9 vs 20.5%, P<0.001), arm fat percent (27.9 vs 14.5%, P<0.001) and leg fat percent (23.4 vs 17.5%, P<0.001), and significantly lower values than Lunar for trunk fat percent (22.3 vs 24.7%, P<0.001) and trunk-to-limb fat percent ratio (0.89 vs 1.62, P<0.001). When measured by Lunar, patients with LD had significantly lower leg fat percent (13.5 vs 21.6%, P=0.04) and higher trunk-to-limb fat percent ratio (1.85 vs 1.38; P=0.012) than those without LD. When measured by Hologic, patients with LD had values of leg fat percent (20.1 vs 26.8%, P=0.11) and trunk-to-limb fat percent ratio (0.95 vs 0.85, P=0.43) that did not differ significantly from those without LD. There was a significant difference in the contribution to a composite LD score between Lunar and Hologic measurements. CONCLUSION DXA machines from different manufacturers give major differences in measurements of body fat content and distribution, and this may affect the ability to distinguish patients with LD from those without LD. Further standardization of DXA technology is needed before widespread application in the clinical diagnosis of LD.