Neonatal and long‐term outcomes of infants with congenital cytomegalovirus infection and negative amniocentesis: systematic review and meta‐analysis

Objective Cytomegalovirus (CMV) DNA is detectable in the amniotic fluid collected by amniocentesis in cases in which the fetus has been infected. However, cases of congenital neonatal CMV infection with a negative amniocentesis result have also been reported in the literature. The aim of the present study was to compare pregnancies with a negative amniocentesis result to those with a positive amniocentesis result in terms of incidence of fetal insult and long‐term sequelae. Methods Observational studies that included pregnant women with CMV infection who underwent amniocentesis and that reported their results together with neonatal and/or long‐term outcomes of the offspring were included. The risk of bias in included studies was assessed using the Newcastle–Ottawa Scale. The rate of severe symptoms at birth, defined as neurological symptoms or multiorgan involvement at birth, and the rate of severe sensorineural hearing loss (SNHL) and/or neurodevelopmental impairment at follow‐up were the main outcomes of the study. The secondary outcome was the rate of pregnancy termination due to the presence of CMV‐associated central nervous system (CNS) findings or multiorgan involvement on ultrasound/magnetic resonance imaging (MRI). Results Seven studies were included in the systematic review and meta‐analysis. The pooled false‐negative rate of amniocentesis was 8.0% (95% CI, 5.0–13.0%). The pooled rate of severe symptoms at birth was 0.0% (95% CI, 0.0–1.0%; I2 = 0%) in fetuses with a negative amniocentesis result and 22.0% (95% CI, 11.0–38.0%; I2 = 75%) in those with a positive amniocentesis result. The pooled odds ratio (OR) was 0.03 (95% CI, 0.01–0.10; I2 = 0%). The pooled rate of severe SNHL and/or neurodevelopmental impairment at follow‐up in fetuses with a negative amniocentesis result was 0.0% (95% CI, 0.0–1.0%; I2 = 0%) and, in those with a positive amniocentesis result, it was 14.0% (95% CI, 7.0–26.0%; I2 = 64%). The pooled OR was 0.04 (95% CI, 0.01–0.14; I2 = 0%). The pooled rate of pregnancy termination due to the presence of CMV‐associated CNS findings or multiorgan involvement on ultrasound/MRI was 0.0% (95% CI, 0.0–2.0%; I2 = 0%) in fetuses with a negative amniocentesis result and 20.0% (95% CI, 10.0–36.0%; I2 = 82%) in those with a positive amniocentesis result. The pooled OR was 0.03 (95% CI, 0.01–0.08; I2 = 0%). A subgroup analysis including only pregnancies with primary CMV infection and a sensitivity analysis including only prospective studies were carried out, showing very similar results to those of the main analysis. Conclusion A negative amniocentesis result in pregnant women with CMV infection ensures lack of fetal insult and long‐term sequelae to the child, even if transmission has occurred.