Accuracy of body composition measurements by dual energy x-ray absorptiometry in underweight patients with chronic intestinal disease and in lean subjects.

BACKGROUND: To assess the accuracy of Dual-energy X-ray absorptiometry (DXA) in underweight patients with chronic gastrointestinal disease, we investigated the ability of DXA to detect variations in body composition induced by infusion of parenteral nutrition (PN). Furthermore, the influence of a low body weight per se on the accuracy of DXA was studied by placing packets of lard on lean healthy subjects. METHODS: The hydration study included 11 patients with short bowel syndrome on long-term home parenteral nutrition (9 women and 2 men), and (mean +/- SD) 49.5 +/- 17.1 yr., 19.3 +/- 3.1 kg/m2. The lard study, where packets of lard were placed either over the thighs or the trunk region, was performed in 8 healthy lean male volunteers, 26.4 +/- 7.4 yr., and 21.0 + 0.9 kg/m2. Body composition, including measures of the total mass (TM), soft tissue mass (STM), lean tissue mass (LTM), fat mass (FM), and total body mineral content (TBBMC), was assessed by DXA. The fat fraction of the lard packets (3.49 kg), measured in triplicate by chemical fat extraction, was 52.2%. RESULTS: Hydration study; The increase in scale weight (BW) of approximately 0.90 kg due to infusion of PN correlated significantly to the increase in TM (R-square = 0.72, SEE 0.36 kg, p < 0.01), and the increase in STM (R-square = 0.69, SEE 0.38 kg, p < 0.01), however not with the increase LTM (R-square = 0.30, SEE 1.06 kg, p = 0.08). Mean changes in TM (0.88 kg), STM (0.88 kg), and LTM (0.81 kg) were not significantly different from changes in BW (p > 0.05). Lard study; Regardless of position, measurements of FM and LTM of the added lard were not significantly different from expected values. However, the composition of the lard packets into FM and LTM was more accurately detected when the packets were placed over the thighs than over the trunk region. The accuracy of DXA in individual subjects, expressed as the SD of the difference between expected and measured values, was 1.03 kg and 1.06 kg for the detection of changes in LTM and FM, respectively, and 0.18 kg for the detection of changes in STM and TM. CONCLUSIONS: On a group level, DXA provided sufficient accuracy to detect small changes in body composition in underweight patients with chronic gastrointestinal disease. However, the accuracy errors were higher than reported in normal weight subjects. The accuracy was not influenced by a low body weight per se.