143 Body composition measurement methods in preterm infants 7 Results Quality assessment Bias fat free mass -179 g (old predictive equation including weight z-scores and clinical parameters) Limits of agreement old equation + 249 g, i.e. poor agreement (mean fat free mass study population 2500 g) Bias fat free mass -12 g (new predictive equation, including weight z-socres and taking into account new feeding regimen) Limits of agreement new equation + 210 g, i.e. good agreement (mean fat free mass population 2500 g) Level of evidence 2 Strengths & limitations Assessment of agreement with reference method + Assessment of intra- and interobserver variability – Repeated measurements – Coefficients of variance assessed + Sensitivity analysis + Bootstrapping analysis + Cross validation group – External validation – Large study population + Exclusively preterm infants + 82% of the variance in fat free mass (g) was explained by predictive equation including clinical factors and change in weight z-score 24% of variance in fat mass percentage was explained by clinical factors and change in weight z-score Level of evidence 4 Strengths & limitations Assessment of agreement with reference method – Assessment of intra- and interobserver variability – Repeated measurements + Coefficients of variance assessed – Sensitivity analysis – Bootstrapping analysis – Cross validation group – External validation – Large study population + Exclusively preterm infants + Results include preterm as well as term infants The bias + limits of agreement of the predictive equation including GA, gender and weight was + 6.1 g for fat free mass adjusted by length (g/cm), i.e. poor agreement (study population mean fat free mass adjusted for length 55 g/cm) The bias + limits of agreement of the predictive equation including GA, gender and weight was + 4.9 g for fat mass adjusted by length, i.e. poor agreement (study population mean fat mass adjusted by length 7.4 g/cm) Level of evidence 4 Strengths & limitations Assessment of agreement with reference method + Assessment of intra- and interobserver variability – Repeated measurements – Coefficients of variance assessed – Sensitivity analysis – Bootstrapping analysis – Cross validation group – External validation – Large study population + Exclusively preterm infants – Weight explained 97% of the variance of fat free mass (g) Body mass index explained 27% of the variance in fat mass percentage and had a prediction error of 3.5% Level of evidence 4 Strengths & limitations Assessment of agreement with reference method – Assessment of intra- and interobserver variability – Repeated measurements – Coefficients of variance assessed – Sensitivity analysis – Bootstrapping analysis – Cross validation group – External validation – Large study population + Exclusively preterm infants +
RkJQdWJsaXNoZXIy MTk4NDMw