Regional anthropometric features to identify newborns at higher risk of poor outcome

BACKGROUND/OBJECTIVE The main problem related to the identification of small for gestational age (SGA) fetuses using the current ultrasonographic (US) approaches is related to the fact that a discrete number of cases are misdiagnosed therefore SGA often show a poor outcome at birth. We applied different models of US assessment of fetal growth to evaluate their efficacy in identifying those newborns. METHODS A group of 5956 singleton pregnant women that delivered from 0 to 6 days from US, by computing fetal routinary biometry and estimated fetal weight (EFW) was evaluated. Measurements were compared to Lubchenco, Yudkin and Festini nomograms of neonatal weight and abdominal circumference (AC) values were compared to the Italian Reference Curve (IRC). Sensitivity, specificity, predictive values and likelihood ratios of fetal biometry as diagnostic tests for the detection of poor neonatal outcome in women with SGA were assessed by using Receiver Operating Curve (ROC) test. Therefore, the probability of SGA after having the test positive (higher than the cut-off) was estimated and compared with the pretest probability, defined as the prevalence of SGA in the whole group of cases. RESULTS According to the values of nomograms referred by Lubchenco and Yudkin, at the best cut-off chosen by the ROC curve analysis, EFW achieved a sensitivity of 46.5% and 66.4% and a specificity of 98.1% and 94.3% respectively, as diagnostic test for SGA. Using the Festini nomogram, EFW achieved a sensitivity of 62.8% and a specificity of 95.3%. Using the IRC charts for AC the sensitivity was 49.5% and the specificity 96.0%. At birth 893 cases were SGA, giving an overall prevalence of the disease in this population of 14.9%. By using the Lubchenco nomogram, the probability of a SGA (PPV) was as high as 81.3%, whilst with values below the cut-off (NPV) it was 8.8%. By using Yudkin growth chart PPV was 67.3% and NPV 5.9%. PPV and NPV by using Festini nomogram and IRC were 70.4% and 68.7%, 93.6% and 91.5% respectively. CONCLUSIONS The Festini nomogram better estimates the risk of having or not a SGA at birth. This nomogram has been computed and validated on a regional (Tuscany) population and suggests that referring to a regional anthropometric evaluation better estimates the probability of identifying newborns at higher risk of poor outcome.