Zika virus (ZIKV) is transmitted primarily through the bite of an infected Aedes species mosquito; however, it can also be transmitted between sexual partners or from mother-to-child in utero or during delivery [1,2]. Numerous studies have documented the serious consequences of ZIKV infection during pregnancy, including fetal or infant brain and eye abnormalities, such as intracranial calcifications and microcephaly [3-7]. Key knowledge gaps still exist, such as understanding the full spectrum of adverse outcomes in pregnant women, fetuses, and infants associated with ZIKV infection in pregnancy; the relative contributions of sexual transmission and mosquito-borne transmission to the occurrence of infections in pregnancy; and the risk of adverse fetal and infant outcomes by gestational week of maternal infection or presence of ZIKV symptoms. Limited evidence suggests that ZIKV infection during pregnancy may be associated with long-term sequelae, including poor neurodevelopmental outcomes for those without congenital abnormalities [8-13]; however, the lack of an adequate comparison group in existing analyses limits the interpretation of whether these adverse outcomes can be attributed to ZIKV infection. There is a need to describe the risk of adverse developmental outcomes in infants and children following ZIKV exposure in utero.

The Colombian Instituto Nacional de Salud (INS) in collaboration with the United States Centers for Disease Control and Prevention (CDC) established the Zika en Embarazadas y Niños (ZEN) prospective cohort study in multiple Colombian cities. The primary objectives were to: (1) identify risk factors for ZIKV infection in pregnant women, partners, and infants; (2) assess the risk for adverse maternal, fetal, infant, and childhood outcomes associated with ZIKV infection and other congenital infections; and (3) assess modifiers of risk.

Pregnant women aged 16 years or older were recruited in their first trimester of pregnancy from participating public and private prenatal care clinics in Colombia. The 13 participating clinics were located in cities in 3 departments: Atlántico, Santander, and Valle del Cauca (Figure 1), representing 50,000 live births in 2017 [14]. Cities/clinics were selected based on national surveillance data indicating ZIKV disease in the prior year, a large volume of pregnant women receiving prenatal care, and support from local health authorities.

Pregnant women were recruited by study staff in clinic waiting rooms, by referral from prenatal care providers, or by self-referral after seeing recruitment brochures or posters. Among the pregnant women screened for eligibility (n=2,737), 879 (32.1%) were excluded because they did not meet eligibility criteria (Table 1 and Figure 2).

Some eligible women (n=339, 12.4%) declined participation after screening. One participant was enrolled twice, and the second pregnancy was excluded from the analysis. A total of 1,519 pregnant women were enrolled across sites from February 2017 to January 2018. Upon enrollment, pregnant women received ZIKV prevention kits with educational materials, condoms, diethyltoluamide (DEET)-containing mosquito repellent, an appointment card, and a thermometer.

For enrolled pregnant women ≥ 18 years, staff asked if they had a male partner who lived with them, and if staff could contact him for participation. Of the 1,399 enrolled pregnant women aged ≥ 18 years, 156 (11.1%) did not have a male partner or one who lived with them, 460 (32.9%) did not agree to have their partners contacted, 4 did not respond to the question, and 779 (55.7%) agreed for staff to contact their partner. Of those who agreed, 123 partners could not be reached. Among the 656 men screened, 33 (5.0%) were not eligible (Table 1 and Figure 3), and 336 (51.2%) declined participation. A total of 287 male partners were enrolled across sites from February 2017 to February 2018.

Among the 1,519 enrolled pregnant women, 1,337 (88.0%) participated throughout their pregnancies, with 1,221 women delivering at least 1 liveborn infant (n=1,239 infants; Figure 2). A total of 1,108 infants were enrolled in the first phase of follow-up, which occurred from birth through approximately 6 months of age (Figure 4). Around the 6-month follow-up, parents (or guardians) were assessed for interest and eligibility for themselves and their infant to participate in the second phase of follow-up through 18 months of age. Parents were given a study kit at enrollment and each follow-up, which included developmentally appropriate toys and supplies and health education materials. Of the 1,011 infants who completed the 6-month follow-up, all met eligibility criteria for the 18-month follow-up, but the parents of 161 (15.9%) infants declined further participation (Figure 4). A total of 850 infants (including 11 multiple birth sets) and their corresponding 839 parents/legal guardians enrolled in the 18-month follow-up. A total of 499 infant-parent pairs completed follow-up to 18 months.

Data are presented for the pregnancy and male partner cohort. Among the 1,519 enrolled pregnant women, the median gestational age at enrollment was 10 weeks (Table 2).

The largest proportions of enrolled pregnant women were 18-24 years old (44.3%), had at least secondary education (48.1%), had public health insurance (68.5%), were living with a partner (73.7%), and were from Atlántico (41.1%). Of 1,519 enrolled pregnant women, 176 (11.6%) withdrew and 6 (0.4%) were lost to follow-up at time of delivery (Figure 2). Of the remaining 1,337 women who remained in the study, 118 (8.8%) had a pregnancy loss and 1,221 (83.8%) had a live birth (includes two multiple gestation pregnancies, which resulted in discordant pregnancy outcomes). All 1,519 enrolled pregnant women had at least 1 serum sample obtained for ZIKV testing; 93% had at least 1 urine sample, and 2 had amniotic fluid samples (taken for other clinical indications); placental tissue samples from 4 losses were submitted for ZIKV testing. On average, pregnant women had 5 serum samples (range, 1 to 11) and 5 urine samples (range, 0 to 15) collected. The total number of samples collected per pregnant woman for ZIKV testing ranged from 1 to 18, with a mean of 9. Sixteen women (1%) had at least 1 positive sample for ZIKV by rRT-PCR.

Among the 287 male partners, the largest proportions of men were 25-34 years of age (49.1%), had at least secondary education (48.1%), had private health insurance (47.1%), and were from Atlántico (41.5%) (Table 2). All male partners had a serum sample taken at enrollment, and 91% had at least 1 urine sample taken for ZIKV testing during their partners’ pregnancy. One participant had semen samples collected. On average, male partners had 3 urine samples (range, 0 to 6) collected for ZIKV testing. The total number of samples collected per male partner ranged from 1-7, with a mean of 4. Two men (0.7%) had at least 1 positive sample for ZIKV by rRT-PCR. Compared to pregnant women with no male partner enrolled, women with a male partner enrolled had higher educational achievement, were more likely to have private insurance, and were more likely to be married (Supplementary Material 5).

The ZEN cohort study aims to provide a prospective evaluation of adverse maternal, fetal, infant, and early childhood outcomes associated with ZIKV and other infections during pregnancy. The study results may guide recommendations for preventing congenital infections including ZIKV, improve counseling of patients about the risks of congenital infections, and help agencies prepare to monitor neurodevelopment and provide services to affected children and their families.

This cohort has several strengths. First, the study design allowed for intensive and frequent follow-up to examine incident infections during pregnancy and infancy. Second, this is one of the few ZIKV cohort studies that included male partners, which will provide information on the protective and risk behaviors for male-to-female sexual transmission of ZIKV in pregnancy. Third, the study collected information on multiple prenatal exposures and risk factors for poor pregnancy and child outcomes that can be examined along with physical and developmental outcomes at several time points between delivery and 18 months of age. In addition, data from the ZEN cohort can help fill gaps in knowledge about the developmental trajectory of children in Colombia more generally, and can also be used as a comparison population for other Colombian cohorts of children exposed in utero to ZIKV.

The ZIKV outbreak was unprecedented in its devastating impact on pregnant women; however, the outbreak had waned by early 2017 when enrollment in the ZEN cohort began [22]. The ZIKV infection rate in this cohort was substantially lower (1%) than the 10% incidence at the peak of the epidemic; thus, the study is under-powered to study the association between ZIKV and adverse pregnancy outcomes. A prospective cohort study enrolling high-risk pregnancies found that 8% of participants at any gestational age were ZIKV-positive by rRT-PCR, and that Zika infection was associated with an increased risk of microcephaly, but not of some adverse pregnancy outcomes (e.g., preterm delivery, small for gestational age) [23,24]. Other studies enrolling pregnant women during the first trimester to monitor incident infections have yet to publish their results [25-27]. Data from the ZEN cohort will be pooled with those from other prospective cohort studies for an individual participant data meta-analysis of ZIKV infection during pregnancy to characterize pregnancy outcomes and longer-term sequelae in children [28]. A third of pregnant women did not consent for their partners to be contacted and only half of the eligible partners agreed to participate in the study, which aligned with the prenatal care appointments when possible. Women with a male partner enrolled had higher educational achievement, were more likely to have private insurance, and were more likely to be married. Furthermore, the scope of the research study was broadened to include additional congenital infections (e.g., cytomegalovirus, lymphocytic choriomeningitis virus), which will allow us to answer key questions about the prevalence and impact of these infections on pregnancy and child neurodevelopment.

There are several lessons learned that could be applied in future infectious disease outbreaks in which pregnant women are especially vulnerable. The study leveraged the existing strong national surveillance and laboratory testing infrastructure of Colombia; however, challenges existed in assembling a cohort in the midst of an epidemic. Delays in enrollment were in part due to the need to build local capacity, consolidate research support, and obtain the required reviews in the United States and Colombia [29]. Study investigators worked collaboratively with local health authorities to identify sites and qualified study staff, develop culturally appropriate materials, and conduct training to expand ZIKV testing capacity and to conduct developmental assessments. The ZEN study established a research network in Colombia that could be leveraged for studies of pregnant women and infants. Finally, the capacity-building of healthcare workers and the emphasis on recommended developmental evaluation for all children, including those affected by ZIKV and other congenital infections, strengthened the health systems in Colombia to improve the health of women, children, and families.

Opportunities for collaboration should be directed to the coinvestigator, Diana Valencia (ile9@cdc.gov).