Serologic surveillance of maternal Zika infection in a prospective cohort in Leon, Nicaragua during the peak of the Zika epidemic

Abstract

Background: Zika virus caused thousands of congenital anomalies during a recent epidemic. Because Zika emerged in areas endemic for dengue and these related flaviviruses elicit cross-reactive antibodies, it is challenging to serologically monitor pregnant women for Zika infection.

Methods: A prospective cohort of 253 pregnant women was established in León, Nicaragua. Women were followed during prenatal care through delivery. Serologic specimens were obtained at each visit, and birth outcome was recorded. Established flavivirus serologic methods were adapted to determine Zika seroprevalence, and a stepwise testing algorithm estimated timing of Zika infection in relation to pregnancy.

Results: Zika seroprevalence was approximately 59% among women tested. Neutralization testing was highly concordant with Zika NS1 BOB results. Per study algorithm, 21% (40/187) of women were classified as experiencing Incident ZIKV infection during pregnancy. Importantly, the Incident ZIKV group included mostly women pregnant during the 2016 Zika epidemic peak and the only 3 subjects in the cohort with RT-PCR-confirmed infections. Approximately 17% of births had complications; 1.5% (3/194) manifesting clinical criteria of congenital Zika syndrome, one was RT-PCR-confirmed as a case of congenital Zika syndrome. Adverse birth outcome did not correlate with timing of Zika infection.

Conclusions: By leveraging prenatal care systems, we developed a simple algorithm for identifying women who were likely infected by Zika during pregnancy.

Fig 1. Serologic algorithm for determining maternal ZIKV serostatus.

a) A series of serologic tests with increasing specificity were used to determine ZIKV prevalence at the time of birth. b) For samples with prior ZIKV infection, further testing of birth and prenatal specimens was pursued to determine the timing of ZIKV infection in relation to pregnancy.

Fig 2. Proportion of women in each serologic category.

The percentage of each serostatus is shown for the total cohort as well as subsets of the cohort stratified by LMP, indicating which pregnancies occurred during dates inclusive of peak ZIKV transmission (“Peak”) or after that period (“Post-peak”). Naïve specimens were excluded for this analysis. 183 of the 187 subjects had data for LMP, n = 108 before 30 Sept 2016 and n = 75 after 30 Sept 2016. LMP, last menstrual period.

Fig 3. Timing of pregnancy relative to Zika epidemic.

The timing of pregnancies in our cohort is shown as a function of last menstrual period (LMP) reported by subjects in relation to the ZIKV epidemiologic curve. The date September 30, 2016 (arrow) divides our cohort into two categories, those LMP on or before this date (green box) were pregnant during known ZIKV transmission in Nicaragua; those with later LMP (red box) were pregnant after the vast majority of reported ZIKV transmission. Epidemiologic data are publically available: (https://www.paho.org/hq/dmdocuments/2017/2017-phe-zika-situation-report-nic.pdf).