Assessment of Neurodevelopment in Infants With and Without Exposure to Asymptomatic or Mild Maternal SARS-CoV-2 Infection During Pregnancy

Key Points Question Is asymptomatic or mild maternal SARS-CoV-2 infection compared with no infection during pregnancy associated with observable infant neurodevelopmental differences at ages 5 to 11 months? Findings In this cohort study involving a geographically diverse cohort of 407 infants born to 403 mothers, no association was found between mild or asymptomatic maternal SARS-CoV-2 infection during pregnancy and infant cognition, language, or motor development as assessed by a novel telehealth-adapted version of the Developmental Assessment of Young Children, second edition. Meaning Given the continued high prevalence of SARS-CoV-2 infection globally, these data offer information regarding infant neurodevelopment that may be helpful for pregnant individuals with asymptomatic or mild SARS-CoV-2 infections.

participated in SARS-CoV-2 surveillance, 689 were invited to participate in the ESPI COMBO sub-study because they were adherent to study surveillance activities for >40% of their enrollment weeks and their infants were <6 months of age during the ESPI COMBO enrollment period. The ESPI COMBO sub-study mirrored COMBO data collection from delivery to 6 months postpartum. ESPI COMBO infants who contributed to analyses presented here were born between January 2021 and September 2021.

eMethods 2. SARS-CoV-2 Classification and Potential for Misclassification in the COMBO Study
The COMBO Initiative utilizes NYP's Clinical Data Warehouse to automatically extract data, including maternal SARS-CoV-2 test results by PCR and serology, on all delivering mothers and their infants (once born) every two weeks. Following automated EHR extraction of mothers infected with SARS-CoV-2 infections during pregnancy, a detailed chart review was conducted for each exposed mother to determine symptom severity and timing. This chart review was primarily conducted by pediatricians and obstetricians providing patient care during the pandemic, or research assistants trained by these clinicians. For women identified as exposed by serology only, the chart was reviewed for outside records scanned into our EHR system, in person and telehealth visits, and all other notes between onset of the pandemic (early March 2020) and birth. The occurrence of a SARS-CoV-2 infection during pregnancy was determined by PCR test results with or without serology for 21 (32.8%), by serology for 42 (65.6%), and by self-report and EHR documentation of non-specific outside positive COVID testing for 1 (1.6%) of the exposed participants from the COMBO cohort included in this analysis. Given that false positive SARS-CoV-2 results by both PCR and serology are extremely rare and most of our patients had either multiple positive tests or experienced symptoms, the potential for misclassifying unexposed women into the exposed group is very low and was not considered in our exploration of effects of misclassification. Based on the detailed chart review, the SARS-CoV-2 infection was classified as either asymptomatic or symptomatic.
Chart review was also conducted for dyads selected for approach for enrollment into the unexposed group and if any evidence of COVID-19 was found during pregnancy, the mother was not selected or was reclassified to the exposed group. However, two sources of misclassification exist: asymptomatic disease and nonseroconversion. The asymptomatic disease in our population can be estimated as follows: 19 (30%) of 64 exposed women in our cohort were asymptomatic, a rate consistent with reports from larger cohorts of pregnant women 4 . The most conservative estimate of misclassification of exposed dyads into the unexposed group prior to 7/20/2020 (when universal testing by serology was initiated) is therefore 30% of 15% of women expected to have COVID-19 disease at some point in pregnancy 5 , which results in 4.5%. After 7/20/2020, misclassification would occur most likely due to asymptomatic disease in nonseroconverters. Therefore, the most conservative rate of misclassification would be 15% of the 4.5% of asymptomatic women or 0.67% of all women initially screened into the unexposed group.

eMethods 3. Classification of SARS-CoV-2 Status in the ESPI COMBO Substudy
Classification was based on a combination of maternal self-report of COVID positivity prior to ESPI enrollment, PCR and symptoms during the ESPI phase, and serological testing during the ESPI phase. The occurrence of a SARS-CoV-2 infection during pregnancy was determined by PCR test results for 36 (72%) and by serology for 14 (28%) of the exposed participants from the ESPI COMBO cohort included in this analysis.
Serological testing in ESPI was conducted from sera collected up to three times during pregnancy (at enrollment, at end of second trimester and at end of pregnancy). All participants had serological testing performed at the end of pregnancy. Samples were tested at the CDC using Luminex xMAP-SARS-CoV-2 Multi Antigen Assay, which is a qualitative assay. The assay tests against 3 antigens: S1, RBD, and nucleocapsid (N) proteins. The N protein is unique to natural infection, whereas S1 and RBD antibodies form in response to both natural infection and vaccination. Therefore, looking at all three antibodies helps differentiate between positive testing in response to vaccination versus natural infection.
However, the ESPI sample contained individuals positive for S1 and/or RBD (negative for N), who had never been vaccinated, which was attributed to quicker waning of antibodies against N than of antibodies against S1 and RBD. Therefore, the following classification taking into account self-report, PCR results, and antibody positivity was implemented:

. Power Analysis
Our power analyses were based on our primary outcome, which aimed to investigate differences in DAYC-2 subdomain scores between infants with and without in utero exposure to maternal SARS-CoV-2 infections. The DAYC-2 is a normative standardized assessment with a population mean score of 100 and standard deviation (SD) of 15. Using these normative values, our power analysis aimed to ensure 90% power to detect a 0.5 SD difference between groups, with alpha set to <0.05. Given the enrollment ratios of approximately 1:1 (COMBO) and 1:3 (ESPI COMBO), achieving this requires a sample size of a minimum of 168 or 225 infants, respectively, for all subdomains. The current manuscript includes a total of 403 infants and therefore our analyses were well-powered to detect differences in DAYC-2 subdomain scores of at least 0.5 SD between SARS-CoV-2 exposed versus unexposed infants. 16 7 Some infants included in this analysis did not complete all of the administered subdomains of the neurobehavioral assessment due to 1) technical issues with Zoom and/or internet connection, 2) infant fussiness, or 3) erroneous administration of the assessment that resulted in our inability to determine a ceiling, or test-terminating score, for the infant.  This analysis only includes mothers who were confirmed SARS-CoV-2 negative during pregnancy and confirmed SARS-CoV-2 positive during pregnancy. Mothers with a pre-pregnancy/indeterminant SARS-CoV-2 infection were excluded (n=37). 2 Adjusted model includes race, ethnicity, maternal age at birth, insurance status, parity, delivery method, gestational age at birth, baby's sex, site, and language of assessment.