Screening for coarctation of the aorta (CoA) during fetal ultrasound (US) in a low-risk population is based on the great vessels asymmetry (GVA) with small aorta. Over the last ten years, our screening rate was 49%. Some of the undetected CoA, that we had in care, seemed to us, to the view of the postnatal US, identifiable during pregnancy. We sought to assess the possible performance of fetal cardiac screening of CoA in an unselected population by the evaluation of aortic valve size measurement.

We retrospectively analysed data from 199 neonates born after 32weeks of gestation between January 2003 and December 2012: 77 were diagnosed for CoA among them 38 were prenatally considered at risk, 122 were referred in utero due to GVA to evaluate during pregnancy the risk of CoA after birth, but did not developed CoA. We compared them to a control group of 166 patients without significant congenital heart disease. US aortic annulus valve (AVA) sizes were measured (as we did prenatally) in parasternal long axis view between hinge points during systole and normalized to body size with Pettersen's Z-score equation.

Mean AVA Z-scores were −0,05 in the controls, −1 in the GVA and −2,4 in the CoA group. AVA Z-scores were significantly smaller in the CoA group than in the control (P<0.05) and in the GVA group (P<0.05). The best cut-off point for Coa discrimination with ROC analysis was a Z-score of −1,84. Among the 39/77 CoA not identified prenatally, 19 had a Z-score<−2 and should probably have been identified. Five had a Z-score>−0,05 and could not be diagnosed.

This study confirms that aortic annulus of newborns with CoA are significantly smaller than those with normal or even asymmetric in late pregnancy hearts. Although, on that basis alone, it is not possible to identify before birth all newborns who will develop CoA, as at least 6.5% presented CoA despite a normal AVA Z-score.