Our study corroborate the mild disease in children during the second wave of the SARS-COV-2 pandemic in Spain2,23,24. In keeping with other paediatric series, the most commonly reported symptoms were fever followed by mild respiratory syndrome, headache and gastrointestinal disease6.
In a Spanish Paediatric Association review of COVID-19 manifestations in childhood, 18 studies were selected, mainly from Europe (52%) and mostly in patients admitted to hospital. Some differences were found compared to our group of patients. Fever followed by mild respiratory and gastrointestinal symptoms were the ones most commonly reported25. General as well as neurological symptoms were also reported, mainly fatigue, and headache, similar to our study. Dyspnoea was more common amongst the European group, and abdominal pain was the most frequent gastrointestinal manifestation25. Few children in our study required hospitalisation or specific treatment. During the second wave of the pandemic a nasopharyngeal sample RT-PCR was performed per protocol to every child with symptoms consistent with COVID-19. This might explain why in our series a mild course of the disease was more common compared to the first wave studies24. As previously reported in the literature, the index case at home was a child in less than 50% of cases, suggesting that adults are the main source of infection7.
In our study a large proportion of children were living with adults with symptomatic COVID-19 and adults with confirmed disease by RT-PCR.
The proportion of non seroconverters is high in our study with either the 3 serological techniques and seems to be higher than the figures reported in adulthood11.
Given the fact that the infection in childhood is frequently asymptomatic or paucisymptomatic it is possible that the proportion of individuals in whom no antibodies are detected is higher amongst the paediatric population as compared to the adult patients.
A progressive decay of antibodies titers is observed both in children and adults in the following months after the infection. The levels of antibodies vary alongside, anti-spike ones may be found earlier than anti nucleoproteins and the latter can also disappear faster26. However, it is possible that despite the absence of IgG antibodies against S protein and nucleoprotein, other neutralising antibodies might be produced. These ones may last longer and be responsible for a long-lasting protection against the disease. In adulthood this protection has been reported up to 15 months18.
Information about the length of antibodies titers against different SARS-CoV-2 antigens is scarce in childhood20 and little is known about their progress in time. A decrease of IgG antibodies titers is observed six months after the infection both in children and adults20. However neutralising antibodies might stay high in the following months specially in children aged 3 or less, even with higher levels than the ones encountered in older children and in adults20. In a recent Italian study a persistence of neutralising antibodies is observed after 7-8 months of infection20. While IgG antinucleoprotein and anti-spike decrease over time and even disappear in up to 54% of patients20. In adults there is evidence of a correlation between antibodies against RBD of spike protein with neutralising antibodies. Nevertheless, the kinetics of these two types of antibodies might be different.27
Both the presence of anti-spike and antinucleoprotein antibodies are associated with a lower risk of reinfection in the following 6 months28. The infection gives rise to an immune response in most cases however the protection against a reinfection is unclear. Immune response can be either humoral or celular and more studies are needed26.
It is possible that the variability in the kinetics of the different antibodies accounts for the slightly observed difference in sensitivity among the types of serologic tests used in our study. In any case, the proportion of non-seroconverter children is high and seemingly higher than the one reported in adults, although its significance in terms of protection remains to be elucidated.
In our series there is a statistically significant association between seronegativity and the age group 2-10 years, or being contact with a non-confirmed COVID-19 household. Although the significance of these findings is uncertain, it could be related to the paucisymptomatology in this age group or due to a less exposure to the virus outside the familiar group. More studies are needed to determine factors involving the lack of seroconversion.
In terms of serology, we found a high concordance in the 3 tests used. Detection of antibodies using ELISA Human IgG/IgA/IgM anti SARS-CoV-2 yielded a higher sensitivity than with Abbot and Siemens techniques although the number of patients was small and therefore comparisons are not possible. Concordance between the 3 techniques adds value to our results on high grade of seronegativity in childhood as compared to adults11.
This study has somes limitations. The number of patients included is relatively small although it is related to the number of patients consecutively identified during the study period. It is a prospective analysis that may have missed some retrospective data obtained in a clinical and epidemiological questionnaire. Therefore, memory bias of parents and children may be present. Data were collected via questionnaires in an outpatient facility and some of them were not obtained in the precise moment the patients attended the emergency room. In some cases, we do not have accurate information about some symptoms such as loss of smell. The external validity of our study is limited because it only includes patients attended in the Emergency Department and may not be representative of the general population. Although SARSCoV2 diagnosis was based on a single positive PCR, due to the high prevalence at the time of the study (second wave of the pandemic in Spain), its positive predictive value is very high, in addition all patients had symptoms or an epidemiological situation that had a high clinical suspicion of infection that supports the positive PCR result
However, our study has several strengths. On the one hand its prospective design, in the second wave of the pandemic in Spain with a selection according to pre-established criteria of consecutive patients with a RT- PCR confirmed diagnosis. And on the other hand, the serologic testing were performed at the same time lapse after the infection, and using different serologic testing, giving strength to our results.
In summary, our results offer additional data of the mild clinical picture in children, who usually are not the index case in the family. In addition, our series suggest that the proportion of children in whom seroconversion is not detected may be higher than in adults. More studies are needed to determine the factors involved in the humoral response and its significance in SARS-CoV-2 infected children.