Interpretation
There is a trade-off between maximizing sensitivity (few false negatives) and specificity (few false positives) in the chart selection. For SGA screening, using data from a previous large cohort study conducted in France, a false negative conferred an adjusted 2.1-increased risk for stillbirth (68). In absolute terms (according to a prevalence of stillbirth among detected SGA of 1%), this means one additional fetal death for each 87 non-detected SGA. The WHO charts exhibited higher sensitivity for SGA-associated adverse outcomes and a low ponderal index. However, false positives are also an issue to consider. A false positive for SGA means unnecessary follow-up and planned delivery, which should be at term in adherence to the international guidelines. A large cohort study in the UK showed that two otherwise normal small babies are picked up for every SGA fetus with complications identified (69). There is evidence from nationwide studies that compared with true negatives, iatrogenic preterm deliveries were 4.6 times higher than false positives. Thus, the ideal chart for fetal growth assessment should combine a good capacity to rule in and rule out SGA-associated complications. Under the assumption that the same weight is given to false negatives and positives, the diagnostic odds ratio [DOR] (+LHR/-LHR) estimates the performance. Especially for the definition of FGR, the IG-21st charts exhibited a better overall performance in predicting low Apgar scores. Furthermore, the diagnostic performance for a low ponderal index (a surrogate of the thirty phenotypes) was better when SGA and FGR were defined using the IG-21st charts.