No abstract

No abstract

No abstract

Growth standards are key tools in assigning fetal smallness. Growth charts are central to this. The availability of growth charts with varying conceptual methodology and design makes their comparison imperative to ensure wise clinical decision making. This was a prospective, descriptive and correlational study performed at two fetal medicine centers, on 1019 unselected third trimester South Indian women with singleton pregnancies. The estimated fetal weight (EFW) was calculated from one dataset per woman using the Hadlock III formula. The EFW centiles were obtained from eight prenatal growth charts: Hadlock, FMF, Spanish, INTERGROWTH, WHO, NICHD, Mikolajczyk and GROW (fully customized), and categorized to B 5th, 5.1 to B 10th, 10.1-89.9th and C 90 centiles. Comparison was done with similar categories of neonatal birthweight centiles obtained from Fenton, INTERGROWTH and GROW customized neonatal standards. At EFW cut-off of B 10th centile, the sensitivity range of the fetal growth charts were between 9.5 and 60% and the false positive rates (FPR) between 1.9 and 18.38%. Similar figures for EFW B 5th centile, were 9.5-64.2% and 1.0-12.8%, respectively. The INTERGROWTH chart had the highest positive predictive value of 54.6-63.6%. The FMF chart had the highest sensitivity and the highest FPR. The sensitivity, at a cut-off of B 5th centile, of Mikolajczyk (9.5-12.6%), and the GROW (14.4-18.9%) prenatal charts were closest to the incidence of uteroplacental insufficiency (7.9%) in our study. Wide variations noted in the performance of prenatal and neonatal growth charts in detecting fetal-neonatal smallness indicates the need for critical selection of growth charts and possibly additional supportive information in clinical decision making. Keywords Fetal growth charts Á Fetal growth standards Á Small for gestational age Á Fetal growth restriction Á Third trimester fetal ultrasound Á Hadlock Á Intergrowth Á GROW Á Neonatal growth standards Electronic supplementary material The online version of this article (

No abstract