Background: Breastfeeding practices strongly influence the epidemiology of human cytomegalovirus infection. By contrast with term neonates, few data are available on transmission of infection from mothers to preterm infants during breastfeeding.
Methods: 151 mothers and their 176 preterm infants (gestational age at birth <32 weeks or birthweight <1500 g) were prospectively screened for cytomegalovirus infection by serology, virus culture, and PCR. The roles of cell-free and cell-associated cytomegalovirus excretion during lactation were analysed longitudinally in relation to transmission, by maximum-likelihood estimates.
Findings: Of the 69 seronegative breastfeeding control mothers, none had detectable cytomegalovirus DNA in breastmilk and none of their 80 infants shed the virus in urine. The proportion of cytomegalovirus reactivation in seropositive breastfeeding mothers was 96% (73 of 76). The early appearance of viral DNA in milk whey (median 3.5 days post partum in transmitters; 8 days in non-transmitters; p=0.025) and infectious virus in milk whey (10 days and 16 days, respectively; p=0.005) were risk factors for transmission. The cumulative rate of transmission was 37% (27 of 73 mothers; 33 infants). The infection of the neonates had a mean incubation time of 42 days (95% CI 28-69). About 50% of the infected infants had no symptoms, but four had sepsis-like symptoms.
Interpretation: The proportion of cytomegalovirus reactivation during lactation almost equals maternal seroprevalence. Breastfeeding as a source of postnatal cytomegalovirus infection in preterm infants has been underestimated and may be associated with a symptomatic infection.