ABSTRACT
BACKGROUND: Body composition could help identify malnutrition in pediatric patients, but there is uncertainty over which techniques are most suitable and prevailing opinion that measurements are difficult to obtain in practice. This study examined the acceptability, practicality, reliability, and validity of different anthropometric and body composition measurements in patients with complex diagnoses in a tertiary pediatric hospital. METHODS: A total of 152 children aged 5-18 years had weight, height, body mass index (BMI), mid-upper arm circumference (MUAC), 4-site skinfold thicknesses (SFT), bioelectrical impedance analysis (BIA), and dual-energy x-ray absorptiometry (DXA) assessed on admission and discharge. Acceptability was assessed in a continuous scale, practicality with number/percentage of successful measurements, reliability with intraclass correlation coefficients and coefficients of repeatability, and validity between "simpler" techniques and DXA with Bland-Altman analysis of agreement and Cohen kappa. RESULTS: Techniques were overall acceptable. Measurements were successful in >50%, with patient refusal uncommon. Coefficients of repeatability were good (0.3 cm MUAC and height, 0.2 kg weight, and 1.0 mm SFTs). All techniques significantly overestimated DXA fat mass, but BMI and triceps SFT better identified abnormal fat mass (κ = 0.46 and 0.49). BIA fat-free mass was not significantly different from DXA, with substantial agreement between techniques (κ = 0.65). CONCLUSION: Body composition by a range of techniques is acceptable, practical, and reliable in a diverse group of children with complex diagnoses. BIA seems a good alternative to DXA for assessing fat-free mass, triceps SFT, and BMI for fat mass but should be used with care as it could overestimate total fat mass in individuals.
Subject(s)
Absorptiometry, Photon , Body Composition , Body Mass Index , Electric Impedance , Skinfold Thickness , Humans , Child , Female , Male , Absorptiometry, Photon/methods , Adolescent , Reproducibility of Results , Child, Preschool , Anthropometry/methods , Body Weight , Malnutrition/diagnosis , ArmABSTRACT
BACKGROUND: Better tools are needed to diagnose and identify children at risk of clinical malnutrition. OBJECTIVES: We aimed to compare body composition (BC) and malnutrition screening tools (MSTs) for detecting malnutrition on admission; and examine their ability to predict adverse clinical outcomes [increased length of stay (LOS) and complications] in complex pediatric patients. METHODS: This was a prospective study in children 5-18 y old admitted to a tertiary pediatric hospital (n = 152). MSTs [Pediatric Yorkhill Malnutrition Score (PYMS), Screening Tool for the Assessment of Malnutrition in Pediatrics (STAMP), and Screening Tool for Risk of Impaired Nutritional Status and Growth (STRONGkids)] were completed on admission. Weight, height, and BC [fat mass (FM) and lean mass (LM) by DXA] were measured (n = 118). Anthropometry/BC and MSTs were compared with each other and with clinical outcomes. RESULTS: Subjects were significantly shorter with low LM compared to reference data. Depending on the diagnostic criteria used, 3%-17% were classified as malnourished. Agreement between BC/anthropometric parameters and MSTs was poor. STAMP and STRONGkids identified children with low weight, LM, and height. PYMS, and to a lesser degree STRONGkids, identified children with increased LOS, as did LM compared with weight or height. Patients with complications had lower mean ± SD LM SD scores (-1.38 ± 1.03 compared with -0.74 ± 1.40, P < 0.05). In multivariable models, PYMS high risk and low LM were independent predictors of increased LOS (OR: 3.76; 95% CI: 1.36, 10.35 and OR: 3.69; 95% CI: 1.24, 10.98, respectively). BMI did not predict increased LOS or complications. CONCLUSIONS: LM appears better than weight and height for predicting adverse clinical outcomes in this population. BMI was a poor diagnostic parameter. MSTs performed differently in associations to BC/anthropometry and clinical outcomes. PYMS and LM provided complementary information regarding LOS. Studies on specific patient populations may further clarify the use of these tools and measurements.
