key: cord-266150-wox7pnkr authors: Torres, Juan Pablo; Piñera, Cecilia; De La Maza, Verónica; Lagomarcino, Anne J; Simian, Daniela; Torres, Bárbara; Urquidi, Cinthya; Valenzuela, María Teresa; O’Ryan, Miguel title: SARS-CoV-2 antibody prevalence in blood in a large school community subject to a Covid-19 outbreak: a cross-sectional study date: 2020-07-10 journal: Clin Infect Dis DOI: 10.1093/cid/ciaa955 sha: doc_id: 266150 cord_uid: wox7pnkr BACKGROUND: A SARS-CoV-2 outbreak affecting 52 people from a large school community in Santiago, Chile was identified (March 12), nine days after the first country case. We assessed the magnitude of the outbreak and the role students and staff played using a self-administered antibody detection test and survey. METHODS: The school was closed on March 13, and the entire community was placed under quarantine. We implemented a home-delivery, self-administered, IgG/IgM antibody test and survey to a classroom stratified sample of students and all staff from May 4-19. We aimed to determine overall seroprevalence rates by age group, reported symptoms, contact exposure and to explore dynamics of transmission. RESULTS: Antibody positivity rates were 9.9% (95%CI: 8.2-11.8) for 1,009 students and 16.6% (95%CI: 12.1-21.9) for 235 staff. Among students, positivity was associated with younger age (P=0.01), lower grade level (P=0.05), prior RT-PCR positivity (P=0.03), and history of contact with a confirmed case (P<0.001). Among staff, positivity was higher in teachers (P=0.01) and in those previously RT-PCR positive (P<0.001). Excluding RT-PCR positive individuals, antibody positivity was associated with fever in adults and children (P=0.02; P=0.002), abdominal pain in children (P=0.001), and chest pain in adults (P=0.02). Within antibody positive individuals, 40% of students and 18% of staff reported no symptoms (P=0.01). CONCLUSIONS: Teachers were more affected during the outbreak and younger children were at higher infection risk, likely because index case(s) were teachers and/or parents from preschool. Self-administered antibody testing, supervised remotely, proved to be a suitable and rapid tool. Our study provides useful information for school re-openings. As of May 30, 2020, reported Covid-19 cases had reached 5,817,385 globally, including 362,705 deaths [1] , with the majority of infections occurring in adults 18-64 years of age (74%) [2] . Children under 18 years of age have made up a relatively small proportion of all cases (3% in the US) [2] , and children under 5 years of age account for less than 1% of reported cases [3] . However, early studies from China suggested that as the epidemic progressed in a large community, children became infected at higher rates [4] . Further information from Europe and the Americas also seem to indicate that children could play a minor role in the infection process, at least in terms of detected cases and transmission [5, 6] . nce the true percentage of the childhood population infected with SARS-CoV-2 is not known, it is unclear the role that children play in transmission of the virus. However, they may interact with a greater number of contacts, as compared to adults, under normal non-social distancing conditions [7] . Additionally, they may have higher viral loads compared to adults [8] , suggesting that they may still be transmitters of disease, similar to other respiratory viruses [9] . The possibility of children transmitting infection, especially to adult populations, led many countries worldwide to implement school closures as an important component of SARS-CoV-2 transmission mitigation policies [10] ; the impact of such measures is still being explored [11] [12] [13] [14] . The reference method for screening and diagnosing Covid-19 infection is RT-PCR, nevertheless detection of antibodies against SARS-CoV-2 (IgG, IgM and IgA) play a complementary role, with particular importance in providing epidemiological information [15] . Seroprevalence has been extensively explored in Covid-19 patients confirmed by RT-PCR, as recently reviewed [16] . A few studies have assessed seroprevalence, primarily in asymptomatic individuals using different methods, with rates among healthcare workers who had direct contact with Covid-19 patients ranging from 1.6-44% [17, 18] and in the general population 2.8-5% [19] [20] [21] . Covid-19 was first detected in Chile on March 3, 2020; over the following two weeks, cases increased slowly, largely via travelers arriving from Asia and Europe [22] . The first communities to report an increase in detected cases were in eastern Santiago, and the first Covid-19 outbreak to be reported in Chile was in a private school in this area. The school outbreak began on March 12, nine days after the index Covid-19 case was detected in the country. The school year began on March 4, and during the first week of school parent-teacher meetings were held nearly every evening (Figure 1 ). On March 13, when there were 58 confirmed cases nationwide, two cases were reported within the school community, one teacher and another staff member. That same day, the Chilean Ministry of Health declared quarantine for the entire school community. In the next few days, confirmed symptomatic cases (RT-PCR) continued to appear, primarily amongst staff and parents, and to a lesser extent within the student population. At the time of the study, PCR testing was available with medical referral when patients were symptomatic or had contact with a known case. As of April 6, 52 M a n u s c r i p t 5 members of the school community had been confirmed positive for SARS-CoV-2 and there had been one associated death; cases were distributed as follows: 7 (13%) students, 18 (35%) staff, and 27 (52%) parents. The index case was a staff member who worked with the entire preschool and elementary school staff, and was present at all of the parent-teacher meetings for Prekinder through 4 th grade. The school outbreak provided a unique opportunity to determine the role of children as potential asymptomatic transmitters of SARS-CoV-2. While children are believed to play a minor role in detected symptomatic cases, if a large number of students were asymptomatically infected, this could be determined after the fact by testing for anti-SARS-CoV-2 antibodies. Thus, we designed a contactfree field study using finger-prick chromatographic-based IgM/IgG antibodies tests, in order to evaluate a large number of students and school staff, 8-10 weeks after the start of the outbreak. Our primary aim was to determine the overall SARS-CoV-2 antibody prevalence in blood during the outbreak and characterize antibody prevalence among students by grade level, as well as in school staff. Secondary aims were to retrospectively identify symptoms, confirmed Covid-19 cases, and contact history within the school community in order to correlate with antibody positivity and explore possible transmission dynamics within the school community. The school is co-educational, located in the community of Vitacura (population 85,000; 2017 Census), in the Metropolitan Region (population 7.1 million). The school has 2,616 students in 14 grade levels. There are 318 staff members, of which 195 are teachers. Classrooms in the preschool have two head teachers and 25-27 students; in the rest of the school classes are 36-38 students, and there is one teacher in 3 rd -12 th grades. The academic researchers presented the study proposal to the school authorities in mid-April, followed by submission to the Ethical Committee for Research in Human Subjects of the Faculty of Medicine, Universidad de Chile, which was approved on April 30. All school staff and a randomized selection of students, evenly distributed by classroom, were invited to participate in the study. Participants were sent an email describing the study and asking them to provide their address if they were interested in participating. In the case of students, parents and/or legal guardians were initially contacted, and if they declined, a random replacement was selected. After indicating an intention to participate, a package containing a flyer with simple step-by-step instructions (supplement) and the antibody testing supplies (test card, lancet, a cryovial with diluent, 2 capillary tubes and an alcohol pad) were distributed to each household using a georeferenced-home delivery service (Uber Flash). The flyer included a link/QR code that directed participants to a web M a n u s c r i p t 6 page where study data were collected using REDCap electronic data capture tools hosted at the Medical School of the Universidad de Chile. [23] Using the REDCap platform, participants could read and digitally sign the informed consent, and in the case of children 8 years and older, assent. Once these forms were signed, a copy was emailed to participants for their records and they were directed to a secure survey that i) asked basic demographic questions, ii) requested information on any previous RT-PCR test for SARS-CoV-2 and potential contact with any Covid-19 positive cases, and iii) asked about symptoms experienced since the outbreak (date and duration in days of each symptom). Next, a detailed video (supplement), prepared by the research team, demonstrated how to perform the antibody test. Participants were asked to attach a photo of the test after 15 minutes had elapsed and self-report the appearance of the three lines, G (IgG), M (IgM), and C (test control). After completion, participants were told they would receive results as interpreted by the research team, within the next few days. All tests were to be performed within one hour of opening the sealed aluminum foil bag. The Novel Coronavirus (2019-nCoV) IgG/IgM Test Kit (Colloidal gold) from Genrui Biotech Inc, China, was used according to the manufacturer's specifications. For details on training staff in interpreting photos of the test cards see the supplement. In a first screening phase, the study nurse and/or technician viewed the photo provided by the participant along with the participant's self-report as to the visibility of the three bands, and determined whether the tests were IgG+, IgM+, IgG & IgM+, Negative, Invalid, or Indeterminate. In this screening phase, all self-reported positive responses, in addition to any visual presence according to the study nurse and/or technician, no matter how faint, were to be considered. All samples considered positive, including those where parents/participants and study staff did not agree, and indeterminate samples (doubt of line presence), along with a random selection of negative samples, were then individually reviewed by a three-physician panel to determine their final positivity/negativity (details in supplement). Data was analyzed in STATA version 15 [24] . Statistical differences for categorical variables were calculated using Pearson's Chi-squared test or Fisher's Exact test; continuous variables were tested using the Mann-Whitney U test or Kruskal-Wallis test. Sample size calculations and details on model adjustment are in the supplement. M a n u s c r i p t 7 Parents of 1,224 children were contacted by email; 1,105 manifested interest in participating and were sent a study package. A total of 1,029 child/parents signed consent and participated in the study, of whom 20 (1.9%) were excluded from analysis due to invalid or indeterminate results. This resulted in 1,009 students for inclusion in the analysis, representing 38% of the entire student body (Supplementary Figure 1a) . All staff (n=318) were contacted by email, and a total of 272 responded and were sent packages; 240 completed the online consent process, of which 235 were included in the final analysis, representing 74% of the entire school staff (Supplementary Figure 1b) . Overall antibody positivity rates were 9.9% (95%CI: 8.6-11.5) for students and 16.6% (95%CI: 12.1-21.9) for school staff ( Table 1 ). High school students had a lower positivity rate compared to younger levels. There were no differences in positivity between women and men in either students or staff, or between teachers and support staff. Among students, antibody positive children were younger, had a higher PCR positivity rate (in those who underwent PCR testing during the outbreak), and were more likely to self-report contact with one or more confirmed cases, as compared to seronegative children ( Table 2 ). In staff, the only variable that differed by antibody positivity was a previous history of PCR positive results; most individuals reported contact with one or more Covid-19 cases and there was no difference by antibody positivity (74 vs. 54%). Overall 490/1,009 (49%) students and 171/235 (73%) staff declared the presence of one or more symptoms since the start of the outbreak (Figure 2) . In children, abdominal pain and fever were significantly more common among antibody positive (21% and 17%) compared to antibody negative children (9% and 6%, respectively). After removing children that were PCR positive, these differences remain significant (22% and 14% compared to 9% and 5%). Among school staff (Figure 2) , antibody positive individuals reported significantly higher frequency of weakness (54% vs. 30%), abdominal pain (21% vs. 7%), fever (36% vs. 7%), myalgia (46% vs. 24%), dyspnea (26% vs. 6%), chest pain (33% vs. 8%), and hyposmia (31% vs. 5%). After removing PCR positive individuals, these differences remain significant only for fever and chest pain (14% vs. 6% and 25% vs. 7%, respectively). When comparing symptoms between staff and students, a significantly higher frequency of headache, weakness, fever, myalgias, dyspnea, chest pain, sore throat, and hyposmia was observed in staff. Within antibody positive individuals, 40% (95%CI: 30-50%) of students and 18% (95%CI: 8-34%) of staff reported no symptoms (P=0.01). M a n u s c r i p t 8 Regarding the timing of onset of reported symptoms, in the staff the median date was significantly earlier in antibody positive versus negative participants (P=0.02). There was greater dispersion in the timing of symptom onset in seronegative participants (March-May), while in antibody positive subjects symptoms predominantly occurred during March (closer to the school outbreak). In students, the median date of symptom onset was also significantly earlier in antibody positive children (P=0.04). The median duration from self-reported symptom onset to sampling tended toward fewer days for an IgM positive result (44 days for students and 51 for staff), compared to an IgM/IgG positive result (51 and 60 days) or an IgG positive result (58 and 56 days) (Students P=0.06, Staff P=0.17). For seropositive individuals, symptoms occurred earlier in adults compared to students (median: March 15 vs. March 21; P=0.03). Antibody positive students were distributed across all grade levels at the school (Figure 3 ), but were significantly less common in high school students. There was a significantly greater number of teachers that were antibody positive and/or with a history of positive RT-PCR results in the preschool compared to the other levels. The median percent of antibody positive students per classroom was 8.3% (IQR 1.6-14.3%). In 7 classrooms, over 25% of students were positive for antibodies, of which four had a primary teacher who was antibody positive and/or RT-PCR positive. Antibody positive children had an average of 1.8 contacts with a confirmed RT-PCR Covid-19 case, while antibody negative children had 1.4 contacts (P=0.01). We found that the greater the number of contacts, the greater the probability that the child was antibody positive (OR: 1.4; P=0.05; bivariate logistic model; response variable: positive/negative test result; explanatory variable: number of Covid-19 positive contacts). The most common Covid-19 contact in antibody positive (versus negative) children was their teacher (21% vs. 12%; P= 0.022), a household relative (11% vs. 2%; P<0.001), a classmate (9 % vs. 4%; P= 0.03), followed by a home caregiver (4% vs. 0.1%; P<0.001). Sources with the greatest likelihood of possible contagion in students were: a home caregiver (OR: 27.9), a household relative (OR: 5.4), a classmate (OR: 3.2), and teacher (OR: 2.2). SARS-CoV-2 antibody prevalence was 10% in students and 17% in staff from a large school in Santiago, Chile, 8 to 10 weeks after a Covid-19 outbreak affecting mostly staff and parents, and to a lesser extent students. These prevalence rates are higher than reports from community settings, where rates reported to date are about 5% [19] [20] [21] . However, due the fact that not everyone who develops Covid-19 develops an immune response [26] , true infection rates may have been even higher. Consistent with the detection of PCR confirmed cases during the outbreak, more adults than students were antibody positive. Among students, antibody positivity was higher in younger children, pre-high M a n u s c r i p t 9 school, and among teaching staff as compared to non-teaching staff. Increased positivity in younger children was likely due to the fact that the index case was a member of the preschool community. Factors associated with antibody positivity in children were SARS-CoV-2 PCR positivity during the outbreak and a history of contact with an infected case; in staff, only PCR positivity was significantly associated with antibody positivity. Overall, PCR testing and contact history was significantly higher in staff compared to students, which in addition to the higher antibody positivity observed in this study, support the more significant role of adults within the outbreak, in proportion to the overall population. Nineteen participants had prior positive PCR results, of which 14 (73.7%) were also antibody positive. This outbreak has particular epidemiological features that may differ from other current or future school outbreaks. The index case(s) was(were) incoming travelers, teachers, parents and possibly students. Initial cases in non-traveling members of the community were mostly related to parentteacher meetings, especially among adults related to the preschool. It is possible that a future outbreak, with a different index case(s), may result in a different age distribution of infection within the school. The duration of the virus's circulation period within the school lasted a maximum of 10 days from the beginning of the school year to school closure (if an infected individual circulated on opening day). The virus most likely continued to circulate in the households of infected individuals, and a thorough investigation of intra-familiar outbreaks was not performed. Additional external exposures both during and following the outbreak period, while possible, seem less significant, as a majority of families in the area were under relatively strict stay-at-home orders imposed on March 20. Our findings, if replicated in other schools, have several potential implications for the prevention of future Covid-19 outbreaks in schools. In this outbreak, school closure occurred within two days of detection of the first two cases and within a few days of the start of the school year. Nevertheless, newly detected cases occurred mostly after school closure, with a declining trend in school staff and an increase in parents and students. We hypothesize that adults, mainly through adult-to-adult contact seem to have been most affected during this outbreak. Students on the other hand, were most likely to be infected by adults, either their teacher or parent, who most likely were infected during parentteacher meetings. Thus, reopening schools, under scenarios in which community transmission levels may be as low as during the pre-epidemic phase, should focus on avoiding new cases among teachers. Assuring household detection of cases in adults will also be important. General recommendations for reopening schools are focused on maximizing person-to-person distancing as much as possible given the reality of a school setting: reducing both the number of students and hours within classrooms, limiting group activities, rapid identification and isolation of symptomatic cases, robust parent-school reporting of cases and action plans in case of outbreaks, among others. Our study, does not contradict these recommendations, but may lead to prioritize actions, by especially focusing on reducing adult-to-adult transmission. Certainly, school related outbreak studies from other settings and situations, especially after the first wave, are required to further increase our knowledge of transmission dynamics within M a n u s c r i p t 10 schools, in order to pinpoint recommendations which may end up being most effective. It is likely that current infection awareness will allow early identification of future school outbreaks and implementation of school measures within a short period, which should rapidly curtail a potential outbreak. Antibody detection tests have limitations. Although manufacturers claim sensitivities and specificity of 72-100% and 98.7-100% respectively [25] , how these values relate to the true infection status of individuals in a community setting is not fully clear. According to the manufacturer, any visible band at the IgG and/or IgM level indicates a positive test. Nevertheless, visibility of the band is subject to individual interpretation, as evaluated in our pilot testing. Bands can range from clearly visible to faintly visible. To deal with this issue we defined criteria for positivity, with four levels of visibility, based on a medical panel's review of photos of the test card. Overall there was agreement between parents and the medical panel in 931/1,009 student samples (92%) and 213/235 (91%) staff samples. Discrepancies largely occurred in samples deemed negative by parents, but positive by the panel. Antibody positivity levels would have been different based exclusively on parent/staff ascertainment: 65/1,009 (6.4%) and 31/235 (13%), respectively. A second test to confirm results in samples with discordant interpretations between the panel and participants would have been desirable (e.g. serum antibody test by ELISA); however this was not possible due to a citywide quarantine, which is still in place to date. Importantly, the self-applied tests were successful in terms of self-application with only 3.7% determined to be invalid and 0.5% indeterminate (1.9% were invalid/indeterminate after the option to perform a second test). While it would have been ideal to test parents as well, due to funding limitations this was not possible. Symptoms were retrospectively reported with the known risk of recall bias. Importantly, we asked participants to report symptoms prior to viewing the instructional sampling video, thus it is unlikely that test results influenced self-report of symptoms. The fact that both antibody negative students and staff reported symptoms at later dates leads us to conclude that their symptoms were due to other causes. Self-application of the antibody test, delivered using a home delivery system, in combination with the use of electronic data capture via the participant's cell phone in a secure web platform proved to be an effective, contact-free epidemiological tool in the current pandemic setting, where limiting person-toperson contact is a relevant public health measure. Such methodology could be considered in other settings worldwide, including remote areas. M a n u s c r i p t 11 In this school-based Covid-19 outbreak in Chile, affecting nearly 50 people among school and household members, antibody positivity rates based on a self-administered test were 10% and 17% among students and staff, respectively. Adults, mainly through adult-to-adult contact seem to have been most affected during this outbreak. Self-administered, remotely supervised antibody testing allowed us to determine the magnitude of infection and characterize outbreak features. M a n u s c r i p t 12 M a n u s c r i p t M a n u s c r i p t M a n u s c r i p t 17 World Health Organization. Coronavirus disease (COVID-19) Situation Report-132 COVID-19): Cases in the The Novel Coronavirus Pneumonia Emergency Response Epidemiology T. The Epidemiological Characteristics of an Outbreak of 2019 Novel Coronavirus Diseases (COVID-19) -China Epidemiology and transmission of COVID-19 in 391 cases and 1286 of their close contacts in Shenzhen, China: a retrospective cohort study Coronavirus Disease 2019 in Children -United States Spread of SARS-CoV-2 in the Icelandic Population Changes in contact patterns shape the dynamics of the COVID-19 outbreak in China An analysis of SARS-CoV-2 viral load by patient age Journal of clinical virology : the official publication of the Pan American Society for United Nations Educational SaCO. COVID-19 educational disruption and response School closure and management practices during coronavirus outbreaks including COVID-19: a rapid systematic review Impact of school closures for COVID-19 on the US healthcare workforce and net mortality: a modelling study School Opening Delay Effect on Transmission Dynamics of Coronavirus Disease 2019 in Korea: Based on Mathematical Modeling and Simulation Study Impact assessment of non-pharmaceutical interventions against coronavirus disease 2019 and influenza in Hong Kong: an observational study Connecting clusters of COVID-19: an epidemiological and serological investigation Antibody Tests in Detecting SARS-CoV-2 Infection: A Meta-Analysis SARS-CoV-2-specific antibody detection in healthcare workers in Germany with direct contact to COVID-19 patients Asymptomatic Seroconversion of Immunoglobulins to SARS-CoV-2 in a Pediatric Dialysis Unit COVID-19 Antibody Seroprevalence Seroprevalence of SARS-CoV-2-Specific Antibodies Among Adults Informe de situación COVID-19 N16 Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support StataCorp. Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC Diagnostic performance of COVID-19 serology assays Clinical and immunological assessment of asymptomatic SARS-CoV-2 infections We thank the School Rector David Halm, c.s.c. and Vicerector Macarena López for their enthusiastic support for the study and the entire school community for participation in such difficult times. We thank Dr. Sergio George for his support in the writing of the manuscript, Uber for the support with free rides for package delivery, Cowork Latam for granting us office space, and Nicole Huerta for her revision of photos of test cards. The authors have no conflicts of interest to declare. The ethics committee of the Medical School of the Universidad de Chile approved of all aspects of this study.