key: cord-0785591-7nyd0imv
authors: Levorson, R. E.; Christian, E.; Hunter, B.; Sayal, J.; Sun, J.; Bruce, S. A.; Garofalo, S.; Southerland, M.; Ho, S.; Levy, S.; Defillipi, C.; Peake, L.; Place, F. C.; Hourigan, S. K.
title: SARS-CoV-2 Seroepidemiology in Children and Adolescents
date: 2021-01-31
journal: nan
DOI: 10.1101/2021.01.28.21250466
sha: 2285dc68e54e315f5a54e34d58f7b8bf1cf72310
doc_id: 785591
cord_uid: 7nyd0imv

ObjectivesPediatric SARS-CoV-2 data remain limited and seropositivity rates in children were reported as <1% early in the pandemic. Seroepidemiologic evaluation of SARS-CoV-2 in children in a major metropolitan region of the United States was performed. MethodsChildren and adolescents [≤]19 years were enrolled in a cross-sectional, observational study of SARS-CoV-2 seroprevalence from July-October 2020 in Northern Virginia, United States. Demographic, health, and COVID-19 exposure information was collected, and blood was analyzed for SARS-CoV-2 spike protein total antibody. Risk factors associated with SARS-CoV-2 seropositivity were analyzed. Orthogonal antibody testing was performed, and samples were evaluated for responses to different antigens. ResultsIn 1038 children, the anti-SARS-CoV-2 total antibody positivity rate was 8.5%. After multivariate logistic regression, significant risk factors included Hispanic ethnicity, public or absent insurance, a history of COVID-19 symptoms, exposure to person with COVID-19, a household member positive for SARS-CoV-2 and multi-family or apartment dwelling without a private entrance. 66% of seropositive children had no symptoms of COVID-19. Orthogonal antibody testing with a receptor binding domain specific antigen revealed a high concordance of 80.5%. Children also demonstrated a robust immune response to the nucleocapsid antigen. ConclusionsA much higher burden of SARS-CoV-2 infection, as determined by seropositivity, was found in children than previously reported; this was also higher compared to adults in the same region at a similar time. Contrary to prior reports, we determined children shoulder a significant burden of COVID-19 infection. The role of childrens disease transmission must be considered in COVID-19 mitigation strategies including vaccination. Article Summary8.5% of children had SARS-CoV-2 antibodies in Fall 2020, double the adult rate. The role of pediatric infection is important to consider in mitigation strategies. Whats Known on This SubjectSARS-CoV-2 pediatric seroepidemiologic data is limited. Reported viral rates underestimate the burden of infection in children due to mild or asymptomatic disease. Limited cohorts of children suggest low seropositivity rates compared to adults. What This Study AddsUS children in the largest SARS-CoV-2 seroepidemiology study to date had double the rate of antibodies compared to adults. Most children were asymptomatic. Risk factors include age, ethnicity and living conditions. Most children made antibodies to different antigens of SARS-CoV-2.

criteria included receipt of immunoglobulin therapy within the past 11 months, including intravenous 1 7 9

immunoglobulin, whole blood, fresh frozen plasma, experimental therapy with convalescent plasma, or 1 8 0 other blood product transfusions, except for packed red blood cells and platelets. 1 8 1 Enrolled participants completed a questionnaire collecting demographic, health and potential COVID-19 1 8 2 exposure information. Blood was collected by venipuncture, either for the sole purpose of the study or in 1 8 3 conjunction with another clinical blood draw. Blood was tested using the US Food and Drug 1 8 4 Administration emergency use authorized Ortho Clinical Diagnostics VITROS Immunodiagnostic 1 8 5 Products Anti-SARS-CoV-2 Total test (Ortho Clinical) performed on the Vitros 3600 system (Ortho 1 8 6 Clinical Diagnostics, Raritan, NJ) to detect total antibody (IgG, IgA and IgM) responses against the 1 8 7 SARS-CoV-2 spike protein. This assay has a reported clinical specificity of 100% (95% CI: 99.1-100%) 1 8 8 [9] . Samples were documented as reactive (≥1.00 S/Co) or non-reactive (<1.00 S/Co) for anti-SARS-1 8 9

CoV-2. 1 9 0

Orthogonal antibody testing was performed on reactive samples as the estimated prevalence prior to the 1 9 1 study was 1%, based on very limited data at the time. 4 Orthogonal serologic testing was recommended by 1 9 2 the CDC for populations with a low pre-test probability, including low or unknown prevalence of disease 1 9 3

[10]. In such conditions, an initial antibody assay is performed, and then on all positive tests, a secondary 1 9 4 assay with a different or more focused target is performed. In this study, the Ortho Clinical assay 1 9 5 (targeting the total protein -S1 and S2 subunits) was used as the primary assay, followed by the Siemens 1 9 6 SARS-CoV-2 IgG assay solely targeting the S1 unit receptor binding domain (RBD) (Siemens 1 9 7

Healthineers, Erlangen, Germany). Additionally, to assess the antibody response in children to different 1 9 8 SARS-CoV-2 antigens, all initially reactive samples were tested using the Abbot nucleocapsid IgG 1 9 9 antibody assay (Abbott Laboratories, Chicago, IL). Furthermore, a random 10% sample of the initial 2 0 0 non-reactive samples on the Ortho Clinical assay were analyzed on the Siemens (RBD) and Abbot 2 0 1 (nucleocapsid) IgG assays. 2 0 2 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted January 31, 2021. ; https://doi.org/10.1101/2021.01.28.21250466 doi: medRxiv preprint Sample size justification was based on early estimates suggesting a SARS-CoV-2 seroprevalence in 2 0 3 children of approximately 1% [4] . Thus, in order to estimate the seropositivity in demographic subgroups 2 0 4 while considering the clustering effect due to testing at different enrollment sites, 80% Wilson score 2 0 5 intervals were used to determine a total sample size of at least 1,000 children.

Statistical analysis methods included both comprehensive descriptive statistics and statistical modeling. 2 0 7

Positive frequency counts were used to estimate both the overall and subgroup prevalence rates to 2 0 8 delineate the effects of different factors. The corresponding Wilson score intervals provided the 2 0 9 uncertainty quantification of the prevalence estimates. Graphical diagnostics and chi-squared tests of 2 1 0 independence were used to determine whether selected covariates were correlated. Univariate and 2 1 1 multivariate logistic regression analyses of antibody presence were used to explore potential effects from 2 1 2 covariates. A final logistic regression model was chosen using a stepwise model selection. Odds ratios 2 1 3 with corresponding confidence intervals were calculated for the coefficients in the model. The 2 1 4 distribution of antibody titer levels across different factor levels was examined. Kruskal-Wallis tests were 2 1 5 used to determine if these distributions differed across potential risk factors. Simple inter-rater reliability 2 1 6 rates were calculated to assess orthogonal testing results. R version 4.0.3 was used to perform statistical 2 1 7

analyses. 2 1 8

This study included 1038 children. Demographic and clinical data are shown in Table 1 . All age groups 2 2 0 between 0-19 years were well represented, and racial and ethnic diversity was reflective of the Northern 2 2 1 Virginia populace. 2 2 2 Table 1 shows the prevalence rates for anti-SARS-CoV-2 total antibody (Ortho Clinical assay) across 2 2 3 demographic and clinical covariates of interest. The overall positivity rate for anti-SARS-CoV-2 total 2 2 4 antibody was 8.5% (88/1038). SARS-CoV-2 antibodies were found in 8.2% of White children, 5.2% of 2 2 5

Black or African American children, 5.7% of Asian children, and 16.2% of children identified as other 2 2 6 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted January 31, 2021. ; https://doi.org/10.1101/2021.01.28.21250466 doi: medRxiv preprint racial origin. When compared by age groups a bimodal distribution was noted with a seropositive rate of 2 2 7

13.7% in young children 0-5 years, 7.5% in school-age children 6-10 years, 5.1% in early adolescents 2 2 8 11-15 years and 10.8% in older adolescents 16-19 years (Supplemental Figure 1 ).

Especially high prevalence rates were noted in children with Hispanic ethnicity (26.8%, 55/207), those 2 3 0 with public insurance (Medicaid, 21.6% 27/125), those without insurance (55.0%, 11/20), children living 2 3 1 in multi-family or apartment dwellings without a private entrance (28.7%, 33/115), and children recruited 2 3 2 from safety net primary care clinic (primary care location 2, 25%, 16/64) ( Figure 1 ).

Of children exposed to an individual with a known history of COVID-19, 33.0% (35/106) had antibodies.

Living in the same household with a person who tested positive for SARS-CoV-2 increased the 2 3 5 seroprevalence rate to 52.5% (31/59). Having a personal history of symptoms consistent with COVID-19 2 3 6 was associated with a seroprevalence of 11.8% (30/255). However, 65.9% (58/88) of children with 2 3 7 positive antibody testing had no personal history of symptoms, and 54.6% (48/88) had no known 2 3 8 exposure.

After multiple regression, significant factors associated with seropositivity were found to include 2 4 0 ethnicity, age group, insurance status, symptoms consistent with COVID-19, having a household member 2 4 1 test positive for COVID-19 and dwelling type ( Of those who tested positive for anti-SARS-CoV-2, titer levels ranged from 1.05 to 1000 S/Co. Across 2 4 8 all age groups, titer levels did not significantly differ (Kruskal-Wallis, p=0.088) (Supplemental Figure 2 ).

There was no significant difference in titer levels between those with no symptoms, those with symptoms 2 5 0 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted January 31, 2021. ; https://doi.org/10.1101/2021.01.28.21250466 doi: medRxiv preprint 1 1

An independently determined Northern Virginia adult seroprevalence was 4.4% in a similar study that 2 7 4 concluded 2 months prior to this pediatric study (personal communication, author Christopher Difillipi).

As of mid-October, the Centers for Disease Control reported a Virginia adult seroprevalence of 4.1% 2 7 6

[12]. The seropositivity rate in children in this current study was more than double the reported 2 7 7 concurrent adult seroprevalence.

The higher than expected pediatric seropositivity rate in our study is in part likely attributable to 2 7 9 continued viral transmission over time. In addition, early in the pandemic, it was perceived that children 2 8 0 were at reduced risk for acquiring COVID-19, either from innate protective physiologic characteristics or 2 8 1 reduced exposure [1, 3] . Schools closed rapidly at the start of the pandemic with reduced communal 2 8 2 exposure. However, the unrecognized burden of mild or asymptomatic disease, resulted in testing bias 2 8 3

and under-representation of children in early testing schemes [1, [13] [14] [15] . The high SARS-CoV-2 2 8 4 seropositivity in children we identified was an unexpected finding and it has important implications for 2 8 5 this silent burden of disease and risk of transmission to and from children.

In our population, after multi-regression analysis, identified significant risk factors for seropositivity 2 8 7

included: age-group, Hispanic ethnicity, insurance status, and residence within a multi-family dwelling or 2 8 8 apartment without a private entrance. The highest risk belonged to children living in the same home as an 2 8 9

individual who had tested positive for SARS-CoV-2, with an odds ratio of 22.

Children under 5 years were statistically more likely to be seropositive for SARS-CoV-2 compared with 2 9 1 other age groups in univariate analysis. This finding was also surprising and has implications for the need 2 9 2 for different considerations in mitigation and protection strategies in the younger age groups. Neither the 2 9 3

Spanish nor Switzerland cohorts examined children <5 years, but adolescents (ages 10-19) in Switzerland 2 9 4

were more likely to test positive for antibodies than younger children (ages 5-9) [3, 4] . Young children are 2 9 5 cared for more closely by adults with possible higher risk of exposure due to prolonged physical 2 9 6 proximity and are more likely to interact with each other with less protection. Additionally, even though 2 9 7 there was an indoor mask mandate outside the home, this was only for children ≥ 10 years. Moreover, 2 9 8 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted January 31, 2021. ; https://doi.org/10.1101/2021.01.28.21250466 doi: medRxiv preprint even as public schools were closed for in-person learning, some childcare facilities for younger children 2 9 9 remained open. However, consistent with other reports, our analyses did not indicate a significant 3 0 0 difference in seropositivity in children cared for outside of their home, probably because greater efforts 3 0 1 are taken to prevent transmission in childcare settings [16] . For seropositive individuals, mean antibody titers did not appear to differ between those with or without 3 0 7 symptoms, recent or remote. It is interesting that for some seropositive children, symptoms were reported 3 0 8 <2 weeks from testing as the reported kinetics of the antibody response to SARS-CoV-2 demonstrate that 3 0 9

at least a week is necessary to mount IgM responses after onset of symptomatic disease [15] . Since 3 1 0 COVID-19 symptoms overlap greatly with other childhood viral syndromes, it is possible that these 3 1 1

reported symptoms were not truly caused by COVID-19 infection. It is also possible that some children 3 1 2 with remote symptomatology had seroreversion and so were not captured in the positive cohort [17, 18] .

Exposure to infectious persons with COVID-19 is a well-recognized risk factor for acquiring COVID-19 3 1 4 and our analysis confirms this [19] [20] [21] [22] . Consistent with other reports, the household attack rate was even 3 1 5

higher; over 50% of children reported as exposed to a test-positive household member were seropositive 3 1 6 [23] . Conversely, almost 60% of seropositive children had no known contact with test-positive SARS-3 1 7

CoV-2. This emphasizes the insidious effects of asymptomatic spread with viral shedding in symptomatic 3 1 8 and asymptomatic individuals [24] .

Ethnically and socioeconomically marginalized groups unable to shelter in place have shouldered a 3 2 0 disproportionate burden of COVID-19 disease globally [23, 26, 27] . Hispanic ethnicity, the dominant 3 2 1 minority in Northern Virginia, conferred four-fold odds of seropositivity. Publicly provided health 3 2 2 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted January 31, 2021. ; https://doi.org/10.1101/2021.01.28.21250466 doi: medRxiv preprint insurance and lack of insurance is an accepted marker for lower socioeconomic status, and families with 3 2 3 limited financial means are more likely to live in multi-family dwellings with a common entrance.

Significantly higher seropositivity rates in each of these groups demonstrated the disproportionate impact 3 2 5

of COVID-19 on the underprivileged.

Orthogonal testing with an antibody assay specific to the S1/RBD antigen of the spike protein 3 2 7 demonstrated high concordance. Those samples which failed to show concordance in orthogonal testing 3 2 8

to the S1/RBD exhibited very low titers to the full spike protein (Ortho Clinical assay). Future 3 2 9

consideration should be given to redefining the lower limit of detection for full spike protein antibody 3 3 0 assays as it is possible that these lower titer levels may represent cross-reactivity to conserved epitopes in 3 3 1 benign seasonal human coronaviruses [8] . In addition, some samples that may have been negative on 3 3 2 orthogonal testing may have been due to the Ortho Clinical assay also detecting IgM; however those 3 3 3 negative on orthogonal testing were not more likely to have recent symptoms consistent with COVID-19 3 3 4 within 2 weeks of antibody testing.

Children with a response to the full spike protein assay also frequently manifested antibodies to the 3 3 6 nucleocapsid antigen. Interestingly, it was those in older age groups that more likely failed to produce a 3 3 7 nucleocapsid antibody response, contrary to previously published data suggesting that young children 3 3 8 may fail to elicit a nucleocapsid antibody response [7] . representativeness of the regional population, as it was focused on children having blood drawn for 3 4 2

another clinical purpose and self-referral for the study. However, these factors are accounted for in this 3 4 3 analysis. In addition, Northern Virginia is a major metropolitan area, and so results may not be 3 4 4 generalizable to areas that have more rural demographics. As noted, cross-reaction with other common 3 4 5 endemic human beta-coronaviruses may result in false positive antibody results, particularly at currently 3 4 6 . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted January 31, 2021. ; https://doi.org/10.1101/2021.01.28.21250466 doi: medRxiv preprint participants of the study and their families; such studies can only happen because people find 3 8 0 that helping in research, especially in the midst of a pandemic, is worthwhile, and therefore, . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted January 31, 2021. ; https://doi.org/10.1101/2021.01.28.21250466 doi: medRxiv preprint . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted January 31, 2021. Supplemental Figure 1 : Pairwise plot of age vs. seropositivity rate (%). The red points represent the observed positivity rate in the cohort at each age, and the line is the predicted positivity rate produced by a quadratic regression model. The estimated seropositivity rate is higher for younger and older ages of children.

Supplemental figure 2 : A multiple boxplot that shows the SARS-CoV-2 total spike protein titer levels (S/Co) for those who tested positive separated by age group and summary statistics.

Supplemental Figure 3 : A multiple boxplot that showing SARS-CoV-2 total spike protein titer levels (S/Co) for those who tested positive separated by presence of symptoms and summary statistics.

. CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted January 31, 2021. ; https://doi.org/10.1101/2021.01.28.21250466 doi: medRxiv preprint . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted January 31, 2021. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted January 31, 2021. ; https://doi.org/10.1101/2021.01.28.21250466 doi: medRxiv preprint 1 All other Primary Care clinics 9 0%* 0% 29.9% . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted January 31, 2021. . CC-BY-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted January 31, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted January 31, 2021. 

CDC COVID-19 Response Team. Coronavirus Disease 2019 in Children -United States

Seroprevalence of SARS-CoV-2 immunoglobulin antibodies in Wuhan, China: part of the city-wide massive testing campaign

Prevalence of SARS-CoV-2 in Spain (ENE-COVID): a nationwide, population-based seroepidemiological study

Seroprevalence of anti-SARS-CoV-2 IgG antibodies in Geneva, Switzerland (SEROCoV-POP): a population-based study

Prevalence of SARS-CoV-2 Infection in Children and Their Parents in Southwest Germany

SARS-CoV-2 antibody prevalence in blood in a large school community subject to a Covid-19 outbreak: a cross-sectional study

Distinct antibody responses to SARS-CoV-2 in children and adults across the COVID-19 clinical spectrum

Preexisting and de novo humoral immunity to SARS-CoV-2 in humans

Symptomatic and Asymptomatic Viral Shedding in Pediatric Patients Infected With Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2): Under the Surface

Clinical Characteristics and Viral RNA Detection in Children With Coronavirus Disease 2019 in the Republic of Korea

Interpreting Diagnostic Tests for SARS-CoV-2

Limited Secondary Transmission of SARS-CoV-2 in Child Care Programs -Rhode Island

Coronavirus Infections in Children Including COVID-19: An Overview of the Epidemiology, Clinical Features, Diagnosis, Treatment and Prevention Options in Children

Decline in SARS-CoV-2 Antibodies After Mild Infection Among Frontline Health Care Personnel in a Multistate Hospital Network -12 States

Factors Associated with Positive SARS-CoV-2 Test Results in Outpatient Health Facilities and Emergency Departments Among Children and Adolescents Aged <18 Years -Mississippi

Household Transmission of SARS-CoV-2: A Systematic Review and Meta-analysis

Transmission of SARS-COV-2 Infections in Households -Tennessee and Wisconsin

Humoral Immune Response to SARS-CoV-2 in Iceland

Assessment of 135 794 Pediatric Patients Tested for Severe Acute Respiratory Syndrome Coronavirus 2 Across the United States

SARS-CoV-2, SARS-CoV, and MERS-CoV viral load dynamics, duration of viral shedding, and infectiousness: a systematic review and meta-analysis. The Lancet Microbe

COVID-19 Trends Among School-Aged Children -United States

Clinical Characteristics of 58 Children With a Pediatric Inflammatory Multisystem Syndrome Temporally Associated With SARS-CoV-2

Severe Coronavirus Disease-2019 in Children and Young Adults in the Washington, DC, Metropolitan Region