Air Displacement Plethysmography is an Accurate and Feasible Noninvasive Measure of Fat-Free Mass in Children With Intestinal Failure.

BACKGROUND: The nutritional status of children with intestinal failure (IF) can be difficult to determine using body weight and currently available anthropometric techniques. Air displacement plethysmography (ADP) is a noninvasive measure of whole-body composition that measures body mass and volume, with a calculation of percent body fat (%BF) and fat-free mass (FFM) that may be useful during the provision of specialized nutrition. OBJECTIVES: To evaluate the validity and feasibility of measuring body composition in children with IF using ADP compared with deuterium dilution (DD), as well as secondarily with other measures of body composition, namely bioelectrical impedance analysis (BIA), dual-energy X-ray absorptiometry (DXA), and four-site skinfold anthropometry. METHODS: We conducted a prospective cohort study of 18 children recruited through the Center for Advanced Intestinal Rehabilitation at Boston Children's Hospital. Patients 2-17 years of age with IF dependent on parenteral nutrition (PN) for more than 90 days were included. Spearman rank correlation and Bland-Altman limits of agreement (LOA) analysis were used to compare ADP to 4 alternative measures of body composition. RESULTS: Eighteen children with IF, median age 7.1 [interquartile range (IQR) 5.4-9.3] years, 9 female (50%), and median residual bowel length 31 (IQR 22-85) cm were enrolled. Median PN energy intake was 46 (IQR 39-49) kcal/kg/day. Incomplete bladder emptying lead to invalid measures of DD in 4 subjects. Spearman correlation coefficients for %BF were low to moderate between ADP and DD ( r = 0.29), DXA ( r = 0.62), BIA ( r = 0.50), and skinfold ( r = 0.40). Correlations for FFM were high between ADP and these other measures (range 0.95-0.98). Comparing ADP with DD and skinfold measures, Bland-Altman analysis showed small mean bias (-1.9 and +1.5 kg) and acceptable 95% LOA ranges (10.7 and 22.9 kg), respectively, with larger bias (-10.7 and -7.7 kg) and LOA ranges (38.7 and 45.2 kg) compared to DXA and BIA. %BF by ADP and skinfold thickness were moderately correlated ( r = 0.43) with low bias (-0.2%) but very wide LOA (25.7%). CONCLUSIONS: Body composition via ADP is feasible and valid in children with IF as a measure of FFM but appears less suitable for the measurement of %BF. The technique holds promise as a noninvasive measure of body composition to assess the efficacy of nutritional, medical, and surgical interventions.