The apparent paradox of maternal seropositivity as a risk factor for congenital cytomegalovirus infection: a population-based prediction model.

Because maternal seropositivity for CMV is associated with substantial protection against congenital CMV infection, prevention measures have focused mainly on seronegative pregnant women for decades. However, population-wide insight in the contribution of nonprimary infection (reactivation and/or re-infection with a different strain) on the most common sequela, hearing loss, is missing. A population-based prediction model was developed to estimate the proportion of congenital CMV-related hearing loss resulting from nonprimary maternal infection. Incorporated was a meta-analysis of the risk of hearing loss, calculating pooled proportions of children with hearing loss after nonprimary and primary infection. Subsequently, the model was applied for worldwide present population seroprevalences (range 30-95%). It was estimated that, for all population seroprevalences, nonprimary maternal infections are responsible for the majority of congenital CMV infections. This proportion increased with seroprevalence, ranging from 57% (95%CI 24-85%) to 96% (95% CI 88-99%) for seroprevalences of 30% to 95%. Our meta-analysis (six reports) showed that the risk of hearing loss after nonprimary infection was 11% (28/253 children, 95% CI 7-15%) versus 13% (50/385 children, 95% CI 10-16%) after primary infection. Incorporating this risk into our model, we estimated that nonprimary infections also accounted for the majority of CMV-related hearing loss. This proportion ranged from 53% (95% CI 13-86%) to 95% (95% CI 62-99%) for seroprevalences of 30% to 95%. Our data underline the worldwide contribution of nonprimary infections in causing CMV-related hearing loss. These results imply that prevention research such as vaccine and hygiene studies should not only be directed at seronegative but also seropositive pregnant women.