We estimate that in England, 88 CYP died of COVID-19 during the first 26 months of the COVID-19 pandemic, with an estimated mortality rate of 3.59 (2.88–4.42) per 1,000,000 person years. At the beginning of the pandemic, most CYP who were found to be SARS-CoV-2 positive around their death, were thought likely to have died of the virus itself. This proportion reduced during 2020, but there was a rise in this proportion in the first months of 2021, at the time of the Omicron outbreak. The largest number of deaths from COVID-19 in CYP occurred in the youngest (under 1 year), and oldest (over 10 years), in contrast to previous work suggesting lower population risk in infants. This translated to a highest predicted risk of death from SARS-CoV-2 in the youngest and oldest CYP. The risk of dying of COVID-19 was also highly patterned by ethnicity and deprivation, with the risk substantially higher for those CYP from Asian or Black backgrounds, or in the more deprived areas. We identified a number of underlying conditions that may be associated with a higher mortality risk due to COVID-19. Children with neurodisability, but also Sickle cell disease, Trisomy 21, Oncological disease, Congenital Cardiac Disease, and Cystic Fibrosis were all likely to have a much higher chance of SARS-CoV-2 being considered a cause of death compared to the overall population. Although, as in our previous work, all children with epilepsy or asthma, who likely died of COVID-19 also had an LLC. Indeed, overall, over 80% of CYP who died of COVID-19 had a LLC and 90% had an underlying chronic condition. We identified no deaths in vaccinated CYP, perhaps for two reasons; a delay in vaccine availability for children, because of prioritisation of adults (although vaccine was recommended for neurodisabled older children soon after it was licensed in the UK (Jan 2021)), as well as low uptake of vaccine in this proposed high risk group early in the pandemic5. There were also no deaths in infants born to vaccinated mothers during the first 2 years of the pandemic, although again, effective take-up was only seen towards the end of the study periods6.
This work is based on statutory data reported to NCMD, and previous work has shown good validation and coverage7. However, there were some missing demographic data (e.g. ethnicity), and some linkage with routine data may have been insufficient; and this should be considered when interpreting our findings. We could only identify a HES record in 83% of CYP. For the main analysis we excluded those where a record was not obtained. If these records were not present (possibly due to the child not previously having accessed healthcare), then the proportion of those CYP dying of COVID-19, without an underlying LLC or chronic conditions, appears to double.
The coding of cause of death (i.e. COVID-19 or not) has potential limitations, but we used a similar process to previous work1, and we were able to access a broad range of data to assess the contribution of COVID-19 to the death in each case. We also utilised routine HES records (a routine administrative database) to identify underlying conditions, so mis-diagnosis for some cases is possible. In addition, the estimates of underlying conditions used here, especially for uncommon conditions, may not be accurate, and therefore, the relative mortality rates between different conditions needs to be made with caution. The main limitation of this work, despite national coverage over the first 26 months of the COVID-19 pandemic, is the very small number of COVID-19 fatalities in CYP, which inevitably limits interpretation of sub-group analyses, and precision of our estimates. We also did not have infection status data across the whole population, so mortality rates presented here are population rates, rather than case-specific fatality rates, and differences between groups may represent differential exposure.
Interpretation
It is important to distinguish between the wider impact of the COVID-19 pandemic, and the direct infectious burden of COVID-19 in CYP. In England, child mortality dropped to its lowest level ever recorded during the first year of the pandemic, with reductions in deaths from other infectious agents, substance abuse and from underlying, chronic disease, reducing substantially during the main period of the pandemic, and the social lockdowns around it3,8,9. The 88 deaths identified in this analysis occurred despite this overall reduction in mortality, and despite the social-distancing and other public-health initiatives implemented while the vaccination roll-out was realised. Furthermore, while others have reported lower levels of severe disease in CYP during the omicron variant era2, in this work we saw a rise in the number of deaths during this time, most likely due to a higher number of infections. However, infection rates, patterned by underlying vulnerabilities are complex and interpretation consequently limited.
Data from the United States estimated a crude death rate of 1.0 per 100,000 population for individuals aged 0 to 19 years during a similar surveillance period; around three times higher than that seen in England (0.36 per 100,000). Indeed, in the US, COVID-19 mortality between August 2021 and July 2022, was the 8th most common cause of death, causing 2% of deaths in this age group10 (compared to 1.4% in this work). Some of the difference may be explained by the wider age range reported in that work (0–19 years)10, and the limited data available on the US deaths compared to the NCMD dataset.
Consistent with other work10, mortality in the first year of life, and again in later childhood, appears to be higher than in some older age groups. The higher mortality rate seen here in infants may have been previously unrecognised due to the substantial mortality already seen in this age range11, although may reflect higher infection rate. However, maternal vaccination is likely to reduce stillbirth, preterm births, neonatal ICU admission, in addition to reductions in COVID-19 infection during the first 6 months of life12,13.
The association of COVID-19 deaths with underlying conditions in CYP (especially severe neurodisability) is not unexpected, and consistent with other work14. Around 10% of COVID deaths had no evidence of any long term disease, including common childhood aliments, such as asthma, and is similar to the proportion of COVID-19 deaths across all ages, including adults, where no pre-existing condition was identified (England and Wales data, 11.8%)15. However, for CYP with underlying conditions, it can be difficult to assess the contribution of the virus infection to death, compared to the contribution of the underlying condition itself. For some groups (e.g. malignancies and neonatal deaths), overall mortality has not measurably changed over the 3 years of the pandemic8 as, even if, the individual additional risk of COVID in these groups may be substantial, it remains difficult to identify where the base-line mortality is already very high. In contrast, the social changes around the pandemic reduced the circulation of other infections, and without COVID-19 many of these CYP may have succumbed to a different pathogen. Currently, with all mitigations removed, other viruses are circulating, and it is essential we continue to assess the ongoing contribution of this virus to severe disease, and deaths, in CYP. Finally, of course, all the mortality rates, and rate ratios, reported here were after lockdown implementations, and personal shielding of these vulnerable groups. While social distancing, and other interventions may have reduced some of the health inequalities in England3, the stark associations with ethnicity and deprivation are consistent with ONS data in adults15, and appear disproportionate to the already recognised inequalities in child mortality16; possibly offsetting any possible progress seen over the last decade.