Discussion
In our study, performed in regions of Germany with a relatively high incidence of COVID-19 in adults in the first phase of the pandemic, approximately 6% of children tested positive for SARS-CoV-2 antibodies in two tests directed against the N and S proteins of the virus. Of those children with high N-protein specific antibody levels, 18 showed signs and symptoms suggestive for mild to moderate forms of PMIS weeks to months after the suspected infection with SARS-CoV-2. Within this subgroup, sera of children with symptoms showed significantly lower neutralizing capacity as compared to sera from children without PMIS or PMIS-like symptoms.
This study was not primarily aimed to investigate representative prevalence rates of SARS-CoV-2 infections in children in Germany in the first wave,15 but to answer the question if screening of children for SARS-CoV-2 antibodies may help to detect those at increased risk for PMIS. Despite the closing of schools, kindergartens and nurseries very early on in the pandemic in Bavaria, a surprisingly high number of children showed antibodies in our study. One possible explanation for that could be that many parents who participated in the study suspected a coronavirus infection in their children due to symptoms or outbreaks in the community. Indeed, children were explicitly not tested in the beginning of the pandemic when test capacities were limited. Thus, the study may have addressed an unmet need of parents to get their children tested, which was further supported by the observation that participation in the study was overwhelming.
We used values beyond S/Co of 100 in the Elecsys test as a proxy for strong antibody responses, as this value is approximately the 97th percentile of all previous available test values (provided by Roche, personal communication). We are aware that the assay is a qualitative test but nevertheless, in the absence of a better measure to determine strong antibody responses at the early point in the pandemic, we applied it in the context of a study in this first screening approach. Surprisingly, almost 70% of seropositive children with a mild to asymptomatic course of the initial SARS-CoV-2 infection, had such values compared to only 21% of seropositive adults with mild disease in one of our previous studies.16 These data may suggest that children mount stronger antibody responses to SARS-CoV-2 than adults on a regular basis. It can also be speculated that this may contribute to milder acute course of the initial infection as observed in children. In contrast, high levels of S-specific antibodies seem to worsen the outcome in severe courses of COVID-19 in adults.17 This hypothesis should be further l investigated in longitudinal studies.
Many children show values of S/Co beyond 100 in the Elecsys test, and a similar observation was made for the S protein based in-house test, where also high values were observed in more than half of the positively tested individuals. As we included only children with strong N-specific antibody responses in our follow-up, we may have missed some children that developed PMIS. Due to the anonymous design of our study, we cannot go back to study participants and extend the follow-up to those with lower antibody values. However, this would be needed to investigate possible cutoff points in antibody responses, which could predict PMIS with greater specificity and/or sensitivity. Therefore, it seems a good strategy at this point to evaluate all children who show antibody responses to SARS-CoV-2 for PMIS symptoms.
Our study indicates that very few symptoms are specific for COVID-19 in children. On the other hand, only 23% of the 162 children with detectable SARS-CoV-2 antibodies were free of symptoms in the weeks before the antibody test. Interestingly, even children as young as 6 years of age were able to indicate loss of smell and taste - the only specific symptom for COVID-19 we could identify in children. Thus, screening for SARS-CoV-2 infections in children by symptoms does not seem to be useful.
A large number of children acquired antibodies against SARS-CoV-2 when family members had developed COVID-19. Therefore, we suggest that children confronted with COVID-19 in the household should systematically be screened for SARS-CoV-2 antibody responses, e.g. 4 weeks after the diagnosis in the index case, thereby not missing out on potential childhood SARS-CoV-2 infections despite of mild or absent symptoms in children.
In our study population, at least 2 children showed PMIS with cardiac affection (A-V block grade I; thoracic pain and left axis deviation), even though the children were not recognized to be sick enough to be referred to the hospital. Of note, our study also identified 2 siblings of study participants (tested positive for SARS-CoV-2 antibodies outside the study) with cardiac PMIS symptoms. Furthermore, it became clear in the course of the study that initially mildly affected children suffer from long-term COVID-19 effects. As reported in adults ,18 we identified a case of chronic fatigue syndrome timely associated with SARS-CoV-2 infection. Thus it seems necessary to follow up children after a COVID-19 infection for neuropsychological effects of COVID-19.
We found a significant correlation between PMIS / PMIS-like symptoms and low neutralization titers within a subgroup of children with high titers of N-specific antibodies. Lower neutralization titers were also found in severe PMIS compared to active COVID-19.19 Our observation that lower neutralization capacities are associated even with mild PMIS symptoms suggests that this may at least contribute if not cause the development of PMIS. Further studies need to confirm the predictive value of high N-specific antibodies and clarify whether or not high N-specific antibody titers and low neutralization capacity are independent predictors for the development of PMIS and PMIS-like symptoms.
Based on our results we propose to screen children from households with COVID-19 cases on a regular basis for SARS-CoV-2 antibodies as well as children from areas with high prevalence of COVID-19, if any symptoms suggestive for COVID-19 occur. Those that are found to have antibodies, no matter if S protein or N protein directed, should be offered a clinical follow-up for PMIS. Based on the experience with Kawasaki disease,20 early intervention seems to have the potential to avoid long-term or chronic illness also in PMIS.