There were few children and adolescents with confirmed infection in Norway in the first phase of the pandemic, and very few were hospitalised. Diagnostic groups where there is an indicated increased risk of COVID-19 in adults, for example patients with diabetes and cardiac/pulmonary diseases (17), were not over-represented among individuals under the age of 20 with confirmed infection. A generally low number of children and adolescents with underlying conditions tested positive for the virus.
Other studies have also found that children and adolescents constitute a small proportion of all cases of confirmed SARS-CoV-2. In a European register, the European Surveillance System (TESSy), with over 576 000 registered infection cases, 1.3 % are under the age of 10 years (11), which is comparable with 1.4 % in our data. In the Madrid region and in the USA, 0.8 % and 1.7 %, respectively, of those with confirmed infection were under 18 years old (1, 9), and in China 2.2 % were less than 20 years old (10). We found a somewhat larger share of 6.1 % under 20 years old, which may be due to greater testing capacity and activity in Norway. In Iceland, 3 % of all positive tests were in children aged under 10 years. This may also be due to higher testing activity and a lower threshold for testing (3). However, none of the in all 848 under 10-year olds in the Icelandic study of a random sample tested positive, compared to 0.8 % of adults. In the area in Italy that was first affected, 2.6 % tested positive for SARS-CoV-2 in a population screening, but none of them were children under 10 years old. The percentage of positive tests increased particularly for over-50s (4).
We found no notable gender difference for those with confirmed infection, and equal distribution between the sexes also applies to Norway's population as a whole (18). Icelandic data show a higher risk for men, both through population screening and in targeted testing according to symptoms and risk (3), and studies from China and the US also find a non-significant excess of boys (1, 19). There is a somewhat larger percentage of men than women among hospitalised adults in Norway (18), and other countries also have an excess of adult men among hospital patients (17, 20). The figures for children and adolescents hospitalised in Norway are too low for any conclusions to be drawn.
The proportion of under-20s among patients hospitalised with SARS-CoV-2 was 1.3 % in our dataset. In other places there is a similarly small share of approximately 1 % (1.3 % in New York, 0.9 % under the age of 15 in China) (17, 20).
Reports from the US indicate that about 6 % of children and adolescents with confirmed SARS-CoV-2 were hospitalised (1). The corresponding figures from Italy are 11 % (21) and from Spain a full 60 % (9). A smaller percentage of children and adolescents are hospitalised in Norway, which may be due to greater testing activity and to more youngsters being tested because of mild symptoms or contact tracing. To prevent the spread of infection, some countries have practised hospitalisation instead of isolation at home of confirmed infected persons, and this will also result in a far larger percentage of SARS-CoV-2 patients in hospital (10). The percentage of infected persons who were hospitalised in Norway is far lower among children and adolescents than in the over-20s population (18).
Our study has several strengths, not least that the compilation of data from national registers makes it possible to analyse the whole population. We have looked at large groups with underlying conditions identified through diagnostic codes from both the primary and the specialist health service, and have thus included data from groups of diseases that pose a potentially higher risk of an infectious disease taking a serious course. With such a low number of infected cases, it is not possible to study groups of less common underlying diseases. Our estimates even for combined groups of less common underlying diseases are equivocal. Mild cases of the disease, where health services have not been contacted or no final diagnosis has been made, will also be absent from data sources. Registered data on underlying conditions for the youngest age groups may be incomplete.
A weakness of the study is that testing capacity was limited during the first weeks of the pandemic, and groups such as health workers, potential hospital patients and close contacts of infected persons received priority. This may have led to somewhat fewer children and adolescents with SARS-CoV-2 being tested, compared with adults. We therefore cannot draw conclusions on the incidence of infection, and wait for serological studies to provide greater clarification. Data from Iceland, where large-scale testing was conducted both as screening of a random sample and on the basis of symptoms and risk, indicate a real lower incidence of infection among children under 10 years old (3). A study based on transmission in Wuhan and Shanghai estimates that the risk of being infected with SARS-CoV-2 is three times as high for adults as for children and adolescents (7).
There has probably been a lower threshold for testing children and adolescents with chronic diseases and higher susceptibility to infection. Healthy children with mild symptoms are probably tested to a lesser extent. Testing practice may therefore create a false association (confounding by indication). The association between increased infection incidence and neuromuscular disease must therefore be interpreted very cautiously, also because the number infected in groups with rare diagnoses such as cerebral palsy and muscular diseases is very low. We found that the other diagnosis groups were not over-represented among infected under-20s. However, the figures are small, and any differences may therefore be difficult to demonstrate (type 2 error). We do not have a sufficiently large dataset to study the association between risk of hospitalisation and underlying disease. Studies in the US have found that eight of ten hospitalised children and adolescents in ICUs have an underlying condition (22), and that especially children with chronic complex disorders were over-represented.
Our data show that the incidence of confirmed SARS-CoV-2 is low among children and adolescents, and that very few of those with confirmed infection require hospitalisation. Groups with underlying conditions such as cardiac and pulmonary disease, diabetes or impaired immunity were not over-represented among infected under-20s in Norway in the initial period of the outbreak of the coronavirus.