Reproducibility of DXA estimations of body fat in HIV lipodystrophy: implications for clinical research.

Dual-energy X-ray absorptiometry (DXA) estimates of body fat are increasingly used for the evaluation of human immunodeficiency virus lipodystrophy (HIV LD); however, limited data are available on their reproducibility. This information is essential for using this tool as an end point in treatment trials or as a diagnostic tool. This study evaluates the reproducibility of DXA body fat estimation in HIV-positive subjects with and without lipodystrophy. Thirty subjects representing a spectrum of severity of fat redistribution underwent same-day repeat whole-body DXA scans (Hologic QDR 4500A scanner). Root mean square coefficients of variation (RMS-CV) were used to estimate minimum detectable differences (MDDs) for body fat content in different regions. Absolute MDD was calculated by multiplying the MDD by the mean fat-mass value for each anatomical area. The RMS-CV ranged from 4.0% for arm fat to 1.6% for total fat. Relative and absolute MDD values ranged from 11.0% or 160 g for arm fat to 4.3% or 628 g for total fat. DXA measurements of regional body fat mass in subjects with HIV show similar reproducibility to other populations. Minimal detectable differences were smaller than differences observed in published studies for all measurements. DXA is a sensitive tool for detecting changes in peripheral fat among patients with HIV lipodystrophy.