Interpretation
Reinforcing our findings, four larger cohorts6,7,9,13reported lower estimates of pregnancy losses and/or birth defects (5-7%) while we detected 8.4% of the combined outcomes with an overall agreement among our rate of pregnancy losses (4.7%) and these studies (3-10%). The latter of these cohorts also reported that a laboratory confirmed postnatal diagnosis of CZS, what was not feasible in this study, contributes to 47% of adverse outcomes and 61% of severe adverse outcomes described. A metanalysis of studies encompassing 2941 pregnancies27 estimated a prevalence of 2.3% among all pregnancies and 2.7% in live births, while a recent review showed that among fetuses exposed to ZIKV during pregnancy, fetal losses occurred in 4-7%, microcephaly in 4-6% and CZS in 5-14%2, which are similar to this study findings (4.7%, 3.3% and 2.9%, respectively).
An inverse association between the gestational age of ZIKV infection and the severity of fetal outcomes has been reported with the worst outcomes resulted from first-trimester infection6,7,9,28-30. However, no study has yet systematically and prospectively approached the exact gestational age of the ZIKV-infection or explored this as a continuous variable. Most studies estimated maternal infection by trimester3,5-9,29-32. Also, only one of the previous studies13 has ruled out congenital cytomegalovirus infection based on the detection of CMV-DNA within three weeks of age, which is an important etiology of CNS abnormalities that might confound ZIKV data. We demonstrated that severe adverse outcomes were a consequence of maternal ZIKV infection at < 11 weeks of gestation. We found no infant with typical CZS whose maternal infection occurred after 14 weeks gestation. Due to the relatively low number of affected infants in this cohort, we could not estimate the influence of potential covariates. Nonetheless, the fact that worst outcomes were 14-fold more likely < 11 weeks of gestation is a very relevant finding regarding clinical management.
The single study that systematically tested infants born to ZIKV-infected mothers have estimated a 20-30% maternal-fetal transmission rate13. However, versus other congenital infections, confirmation of infant ZIKV infection remains a challenge because no nucleic acid or serological tests have been validated for this purpose. Only a portion of infants with the typical CZS phenotype are positive33. Consequently, figures based on detection of nucleic acid or antibodies might underestimate the true vertical transmission rate and all ZIKV-exposed babies should be considered at risk for potential late consequences. We evaluated a subgroup of infants without microcephaly or CNS abnormalities to identify early signs of the disease; 8% of them had potential neurologic abnormalities. Indeed, CNS structural findings related to ZIKV infection can be silent in the neonatal period for infants not submitted to neuroimaging examinations34. However, less is known about isolated early warning signs of neurological dysfunctions in young infants without microcephaly and CNS malformations. Prior work showed that 18%35 to 50%36 infants had neurological findings. However, some of these infants had obvious CNS abnormalities limiting comparisons with our study. We continue to follow the infants to confirm the persistence or appearance of neurodevelopmental findings as suggested by others12,37. Certainly, upcoming controlled cohort studies are needed for a better understanding of the long-term adverse outcomes of infants born to ZIKV-infected pregnant women.