Accuracy of prenatal ultrasound screening to identify fetuses infected by cytomegalovirus who will develop severe long-term sequelae

International audienceObjectives - To compare the ability of detailed routine ultrasound examination, performed without knowledge of maternal serology and fetal status, with that of targeted prenatal imaging performed in prenatal diagnostic units in cases of known fetal infection to identify cytomegalovirus (CMV)-infected fetuses that will develop long-term sequelae. Methods - All prenatal imaging reports were collected for 255 children with congenital CMV in a registered cohort between 2013 and 2017 (NCT01923636). All women had undergone detailed routine fetal ultrasound examination at 20-24 and 30-34 weeks as part of routine antenatal care. All cases of known fetal CMV infection had also undergone targeted prenatal ultrasound examination. Postnatal structured follow-up for up to 48 months of age involved clinical, audiological and neurological assessment, including Brunet-Lezine scoring. Long-term sequelae (> 12 months) were considered to be mild in cases with isolated unilateral hearing loss and/or vestibular disorders, and severe in cases with bilateral hearing loss and/or neurological sequelae. All imaging reports were analyzed retrospectively with the knowledge of congenital CMV infection, searching for reference to findings that were, or could have been, related to fetal infection. Findings were analyzed in relation to whether the cases were diagnosed with CMV in utero or only postnatally. Results - There were 237 children with complete follow-up data (> 12 months), for a median of 24 (range, 12-48) months. Of these, 30% (71/237) were diagnosed with CMV prenatally and 70% (166/237) were diagnosed within 3 weeks after birth. 72.5% (29/40) of children with long-term sequelae, including 74% (14/19) with severe long-term sequelae, were not identified in the prenatal period. Among those diagnosed prenatally, the sensitivity of prenatal imaging for predicting long-term sequelae and severe long-term sequelae was 91% and 100%, respectively, while, in the group diagnosed only postnatally, non-specific infection-related ultrasound findings had been reported without raising suspicion in 48% of cases with long-term sequelae and 64% of those with severe long-term sequelae. Conclusions - Routine detailed ultrasound examination in pregnancy is not an appropriate screening tool for congenital CMV infection that leads to long-term sequelae, in contrast with the high performance of targeted prenatal imaging in known cases of fetal infection. The non-specific nature of ultrasound features of CMV and their evolution, and a lack of awareness of caregivers about congenital CMV, are likely explanations. Awareness of the sonologist regarding congenital CMV and knowledge of the maternal serological status in the first trimester seem key to the performance of prenatal ultrasound. Copyright © 2020 ISUOG. Published by John Wiley & Sons Ltd