Postnatally acquired cytomegalovirus infections in preterm infants

Abstract

A postnatal cytomegalovirus (CMV) infection is common in very low birth weight infants with an estimated prevalence of 6–59%. Breast milk from CMV seropositive mothers is the main source of postnatal CMV infection. Ninety-six percent of these mothers shed CMV in their breast milk after delivery due to local reactivation in the breast. The great majority of postnatally infected infants does not present with symptoms and signs of CMV infection, which include sepsis-like illness with thrombocytopenia, pneumonia or hepatitis. However, because of concerns regarding short- and long-term consequences of postnatal CMV infection and in the absence of conclusive studies on neurodevelopmental outcome of infected infants, various preventive measures including freezing or pasteurization of breast milk, are recommended in several countries to prevent CMV transmission from mother to infant. This thesis, performed at the NICU of the University Medical Center Utrecht, The Netherlands, showed that 85% of the admitted preterm infants with postnatal CMV infection did not develop any clinical symptoms of postnatal CMV infection. Still, one-third of the infected infants developed lenticulostriate vasculopathy in the postnatal period, assessed using cranial ultrasonography. None of the infants with postnatal CMV infection in our study has been treated with antiviral medication or intravenous anti-CMV antibodies and the CMV disease was self-limiting in all symptomatic cases. An increased risk of postnatal CMV infection was associated with non-native Dutch maternal origin, decreased gestational age, breastfeeding and low serum anti-CMV IgG infant-mother ratio. Absolute serum anti-CMV IgG titers of mothers and their infants were not associated with risk of CMV transmission. CMV UL55 and UL144 genotype distribution is similar in congenital and postnatal CMV infection, suggesting that not viral genotype but stage of brain maturation is responsible for the development of severe disease. Using cerebral diffusion tensor imaging, a significantly decreased fractional anisotropy was seen in the occipital region of postnatally infected preterm infants compared to non-infected infants, suggesting microstructural changes in that region. However, at 16 months corrected age no differences in neurodevelopmental outcome between infected and non-infected infants were found. Furthermore, postnatal CMV infection was not associated with sensorineural hearing loss in preterm infants during the first and second year of life. Finally, neurodevelopmental outcome was assessed in a large cohort of infected and non-infected preterm infants and did not show significant differences. While this thesis has shown that postnatal CMV infection in preterm infants is usually asymptomatic and does not lead to hearing problems or neurodevelopmental delay at 2 years of age, preventive interventions like freezing, pasteurization or withholding of breast milk do not seem to be necessary. The benefits of feeding preterm infants with breast milk likely outweigh the risk of symptomatic postnatal CMV infection or impaired neurodevelopmental outcome.