Prenatal diagnosis of congenital cytomegalovirus infection: prospective study of 237 pregnancies at risk.

OBJECTIVE

To develop recommendations for prenatal diagnosis of congenital cytomegalovirus (CMV) infection and evaluate possible prognostic markers.

METHODS

We studied 237 pregnant women who had suspected or confirmed primary CMV infections by amniocenteses with or without funipuncture. Diagnosis of CMV was based on culture and polymerase chain reaction (PCR) done on amniotic fluid (AF) samples; fetal blood tests for CMV immunoglobulin M antibodies, PCR, and nonspecific biologic markers; and repeated ultrasound examinations. In cases of pregnancy termination, viral and pathologic examinations of fetuses were done. At birth, CMV infections were sought in newborns. Pediatric follow-up was scheduled for at least 2 years.

RESULTS

Of 210 fetuses and newborns correctly evaluated, 55 had CMV infections. Ten of 38 fetuses infected before 20 weeks' pregnancy had severe congenital disease. The global sensitivity of prenatal diagnosis was 80%. Best sensitivity and 100% specificity were achieved by PCR done on AF sampled after 21 weeks' gestation, respecting a mean interval of 7 weeks between diagnosis of maternal infection and prenatal diagnosis. Fetal thrombocytopenia was associated with severe fetal disease. Ultrasound follow-up missed two fetuses who presented with neurologic impairment due to CMV after birth.

CONCLUSION

A reliable prenatal diagnosis of congenital CMV infection based on PCR on amniocentesis samples can be made after 21 weeks' pregnancy, after a 7-week interval between diagnosis of maternal infection and antenatal procedure. Ultrasound and nonspecific biologic parameters are not sufficient to identify all fetuses at risk of severe sequelae.