Results
Overall, 2934 children participated in the study of whom 2906 were tested successfully with at least one of the two applied antibody tests and 2832 (96.5%) had also entered necessary study data in the online tool. Demographic data of the children participating in the study are given in table 1 and locations of test-centers across counties are depicted in figure 1.
Both tests consistently identified the same children with antibody responses with some exceptions (figure 2). A positive result in at least one of the two tests defined a positive case. Overall, 5.7% of successfully tested children were positive for SARS-CoV-2 antibodies: 138 were positive in both tests, one only in the Elecsys test and 23 by the in-house ELISA (figure 2).
Strong regional differences were observed in the prevalence of SARS-CoV-2 antibodies in children (figure 1). Overall, children in the heavily affected county of Tirschenreuth (with 1638 positive PCR tests/100.000 inhabitants) had positive antibody response 3-4 times more often than in the two other test regions. When only those children randomly selected (approach 1) and only one child (youngest) per family were included in the analysis, 7.2% of tested children where positive in Tirschenreuth, 3.1% in Regensburg and 1.8% in Oberbayern/ Alpine region. In those who volunteered for testing, e.g. due to symptoms that may have been related to COVID-19 or suspected contact to a COVID-19 patient (approach 2), 15.9% were found positive in Tirschenreuth, 2.3% in Regensburg and 7.8% in Oberbayern/ Alpine region, again taking only one child per family into account.
The older the children, the more positive SARS-CoV-2 tests were found, with 4.9% positive in the 0-6 year-olds (n=1299), 5.7% in the 7-10 year-olds (n=849) and 7.3% positive in the 11-17 year-olds (n=684). Children with chronic diseases tended to be slightly less often positive (4.3% of 344) than those without chronic diseases. Within the study population, 263 children had already received a SARS-CoV-2 PCR test previously and 21 had a positive test result. Of these, 15 individuals showed elevated antibody responses (71.4%) while in 6 subjects no response in any of the two tests could be found. 238 children lived in a household with a positively tested family member and of these 32.4% developed antibodies against SARS-CoV-2. Thus, living with a SARS-CoV-2 positive family member is the single most prominent association with a SARS-CoV-2 infection in children in our study population. We assessed symptoms potentially related with SARS-CoV-2 infections in our study population but found very few specific features that would allow to discriminate COVID-19 from common viral infections in children (supplementary table 1).
Next, we invited children with highly positive antibody responses to participate in a clinical follow-up examination (figure 3). For this study, in the absence of any better criteria for selection at the time point of designing the study, children with S/Co ≥ 100 in the Elecsys Anti-SARS-CoV-2 were considered to have a strong overall antibody response, even though the test is licensed for a qualitative interpretation of results only. Values of that kind were observed in 3.4% of all children tested and in 69.8% of the positive cases. As testing was performed in an anonymous way (in which only the parents/ participants could link test results to the proband), children were invited through the study app to participate in a clinical follow-up with their pediatrician. Thus, positive and highly positive children could not be approached directly by the study team or their pediatrician and only those children who responded (52 out of 97, 53.6%) could be screened for signs and symptoms of PMIS similar to Kawasaki disease in a three-stage approach. Of these, 18 (34.6% of the responders and 18.6% of all those with highly elevated measurements in the Elecsys test) showed at least one clinical symptom related to PMIS; while 8 (15.4%/8.2%) showed two or more clinical signs. One child showed additional changes in the laboratory parameters (elevated lactate dehydrogenase (LDH)) while two presented with cardiac findings and one with excessive chronic fatigue.
As highlighted in figure 2, N-protein specific antibody titers (Elecsys) did not correlate with our in-house S-protein ELISA, which has been demonstrated earlier to nicely correlate with neutralization. In apost hoc analysis, we determined the levels of neutralizing antibody titers within this subgroup of 52 children, selected based on their high N-specific antibody signal with follow up data available. While the neutralizing titers correlated with the S-protein ELISA (Spearman’s p=0.0002, R=0.495) no correlation could be found with the N-protein based Elecsys test (figure 4a, b). Analysing the distribution of neutralizing antibody titers (figure 4c), 48% (n=15) of children with neutralization capacity below mean turned out to be symptomatic, while only 15% (n=3) of children with high neutralizing activity showed symptoms (supplementary table 2). This effect was even more pronounced when IC50 values were split into more defined groups (figure 4d). This effect was also seen (to a lesser degree) for the S-protein based ELISA, confirming the observation that S-protein specific antibody titers correlate with neutralizing capacity of the tested sera. In sum, this subgroup analysis revealed significantly lower neutralization titers in symptomatic patients as compared to children (p=0.034), who did not develop signs and symptoms suggestive for PMIS.