158 Chapter 7 inaccurate predictor of regional body composition.25 In our view, currently there is inconclusive evidence on the predictive value of body area circumferences and more research is needed to assess the potential of mid-arm circumference as a predictor of whole body or regional fat mass. Accuracy of skinfold measurements Several studies assess SFT in preterm infants. 24,51-56 For example, Daly-Wolfe and colleagues investigated thepredictive valueof SFTand found that, togetherwith the mid-arm circumference, SFT explained 49% of the variance in fat mass percentage. 24 However, only one study included in the review assessed the validity of SFT in preterm infants 26. Unfortunately, only 10 late preterm infants were included in this study and analysis included term infants as well, so no robust conclusions can be drawn from their findings.26 Recently, we found poor agreement between SFT and body composition in a study with exclusively preterm infants.56 Moreover, SFT is influenced by the fluid status and there is a high inter-observer variability.13 In addition, it could be deemed controversial to use SFT calipers in extremely preterm infants in light of their vulnerable skin. Therefore, there is insufficient evidence to support SFT as an clinically useful measure of fat mass in preterm infants at this time. Accuracy of bioelectrical impedance Both BIA, as well as bioelectrical impedance spectroscopy, have a poor predictive value for TBWmeasured with isotope dilution analyses 28,35,57 and body composition measured with DXA 27. BIA did not seem to provide an additive predictive effect for fat-free mass or TBW, compared to body weight alone.27,28 Though, both Kushner and colleagues as well as Tang and colleagues found the impedance index (cm2/ Ω) to significantly improve the prediction of TBW, the majority of variance in TBW measured with isotope dilution was still explained by weight. 33,57 Kushner and colleagues concluded that the impedance index (cm2/ Ω) explained 99% of the variance in TBW. However, in their subgroup of preterm infants there was a significant bias: a higher variance was found for higher values of the impedance index. This bias was only eliminated by the addition of weight to the prediction equation 33 – implying that bioelectrical impedance on its own is not an adequate predictor of body composition in preterm infants. Accuracy of ultrasound Ahmadetal. previouslydemonstratedthatultrasoundmeasurementscorrelatewith fat mass in preterm infants.58 Depending on the site of ultrasound measurement the intra-observer variability was reported to be up to 14.7% in preterm infants59. To our knowledge the inter-observer variability has not been investigated in preterm infants, but has been reported to show high inter-observer agreement (0.89-0.95)
RkJQdWJsaXNoZXIy MTk4NDMw