Prediction of body water compartments in preterm infants by bioelectrical impedance spectroscopy.

Author: Collins CT, Reid J, Makrides M, Lingwood BE, McPhee AJ, Morris SA, Gibson RA, Ward LC.
Affiliation:
FOODplus Research Centre, Women's and Children's Health Research Institute, Women's and Children's Hospital and Flinders Medical Centre, Adelaide, South Australia, Australia. carmel.collins@health.sa.gov.au
Conference/Journal: Eur J Clin Nutr.
Date published: 2013 Jan
Other: Volume ID: 67 Suppl 1 , Pages: S47-53 , Special Notes: doi: 10.1038/ejcn.2012.164 , Word Count: 225



BACKGROUND/OBJECTIVES:
To evaluate nutritional interventions in preterm infants, a simple, accurate assessment of the type of growth, that is, change in body composition through the relative contributions of lean body tissue and fat mass to weight gain, is needed. Bioelectrical impedance may provide such a method. The aim of this study was to develop resistivity coefficients appropriate for use in bioelectrical impedance spectroscopy (BIS) analysis of body water volumes in preterm infants.
SUBJECTS/METHODS:
A total of 99 preterm infants were enrolled (mean gestational age 32 completed weeks). Total body water (TBW) and extracellular water (ECW) were determined using the reference methods of deuterium and bromide dilution. BIS measurements taken at the same time allowed calculation of resistivity coefficients. Predictions of TBW and ECW obtained using these coefficients were then validated against volumes determined using the reference methods in a separate cohort of infants.
RESULTS:
Data were available for 91 preterm infants. BIS-predicted TBW and ECW correlated well with the measured volumes (Pearson's r(p)=0.825 and 0.75, respectively). There was a small bias (TBW 10 ml and ECW 40 ml) but large limits of agreement (TBW ± 650 ml and ECW ± 360 ml).
CONCLUSIONS:
BIS appears to have limited clinical utility; however, the relatively small bias means that it may be useful for measurements within a population or for comparisons between groups in which population means rather than individual values are compared.
PMID: 23299871

BACK