Children are unlikely to have been the primary source of household SARS-CoV-2 infections

INTRODUCTION

On the 31^st of December 2019, the World Health Organisation (WHO) was informed of a series of unknown pneumonia cases emerging in Wuhan City, Hubei Province, China. On the 7^th of January, 2020 the causative agent of these infections was identified as a novel coronavirus subsequently named SARS-CoV-2. This virus eventually spread to more than 180 different countries worldwide and in March 2020 the WHO declared SARS-CoV-2 a pandemic. At the time of writing (22^nd March, 2020) SARS-CoV-2 has infected >274,000 people resulting in >11,000 deaths.¹ Large data analyses have shown that the elderly are particularly susceptible to severe forms of COVID-19 (the clinical manifestation of SARS-CoV-2 infection).² However, disease severity and transmission amongst the paediatric population is less well defined. In an analysis of the 44,672 confirmed COVID-19 cases in China only 2·1% were found to be in children (≤19 years of age),² suggesting a reduced rate of infection in this age group. However, this observation remains controversial as others have suggested that the rate of SARS-CoV-2 infection in children under ten (7·4%) is similar to the population average (7·9%).³ The current consensus is that clinical symptoms of COVID-19 are, on average, milder in children (although severe infections can still occur).^4–8 Such observations are consistent with the lower incidence of severe disease amongst children following infection with other coronaviruses including severe acute respiratory distress syndrome (SARS) or Middle East respiratory syndrome (MERS).^9,10 However, this lower incidence of clinical symptoms (and therefore lower incidence of detected disease) raises concerns that children could be an important source of SARS-CoV-2 in household transmission clusters.⁸

To address this question we performed an observational study using currently published literature of the clinical features of SARS-CoV-2 in children and descriptions of household transmission clusters of SARS-CoV-2.

METHODS

Definitions of household transmission clusters

A household transmission cluster was defined as a group of ≥ 2 confirmed cases of SARS-CoV-2 or H5N1 influenza virus infection in co-habiting individuals in whom the onset of cases occurred within 2 weeks of each other. Co-habiting individuals were defined as parents, children, partners, siblings, grandparents, grandchildren, uncle/aunts, nieces/nephews, family-in-law and live-in employees. The index case of each cluster was defined as the individual in the household cluster who first developed symptoms. Adults were defined as individuals equal to or over the age of 18 whilst children were defined as individuals less than 18 years of age. The decision tree used to classify household transmission clusters is shown in Figure 1.

• Download figure
• Open in new tab

Role of the funding source

The sponsors of this study had no role in the study design, collection, analysis, and interpretation of data, report writing, or the decision to submit for publication.

RESULTS

We identified 166 articles that described SARS-CoV-2 infection in children, rejected 126 articles due to a lack of sufficient and or appropriate data and derived a total of 40 articles in addition to case information obtained from local news sources.

A broad range of disease severity was observed (Table 1). Whilst approximately 19% of paediatric cases were asymptomatic, severe cases (12%) were still observed. In children, the predominant symptoms associated with disease onset were fever (77%) and cough (59%), whilst only a minority of children were reported to display fatigue (2%). The majority of children showed abnormal chest computer tomography (CT) although almost 30% of cases had normal chest CTs. The mean incubation period across these studies was 7.74 d ± 3.22.

Mean complete blood counts were largely unremarkable, with the mean values falling within the normal reference ranges for children aged 6-12 years⁴⁶. (Table 1). In contrast, blood chemistry analysis showed that the mean value of c-reactive protein (18.46 mg/L) was elevated compared to the 10mg/L suggested as the upper limit of the normal reference range for children⁴⁷. Similarly, the mean level of lactate dehydrogenase level detected in the blood of children infected with SARS-CoV-2 (315.75U/L) was elevated relative to the recommended reference range for children aged 1 to 10 years (<305U/L)⁴⁸. Together, these data suggest that SARS-CoV-2 induces a broad range of clinical symptoms in children and can manifest as either an asymptomatic, mild or severe infection.

Asymptomatic infections experienced by children infected with SARS-CoV-2 could have important implications for transmission. Specifically, a number of asymptomatic or mild paediatric cases could result in children driving the intra-household transmission of SARS-CoV-2.⁸ To investigate this hypothesis, we examined the available published data on SARS-CoV-2 household transmission clusters. A total of 31 household transmission clusters were identified with sufficient information available to include in this analysis (Table 2). These cases were drawn from China, Singapore, the USA, Vietnam, and South Korea. Of these transmission clusters, only three (9·7%) were identified as having a paediatric index case (Table 2). Importantly, these data do not suggest that children are not becoming infected with SARS-CoV-2, as isolated paediatric SARS-CoV-2 infections would not be captured in reports of household transmission (and are indeed shown in Table 1). Rather, these data suggest that to date, children have not been the primary source of household SARS-CoV-2 infections.

The limited number of defined SARS-CoV-2 household clusters with children as the index case could have been influenced by the fact that disease in children can be asymptomatic. Accordingly, it is possible that within a household cluster children were not correctly identified as the index case of the infection (i.e. the first to develop symptoms) and were instead mistakenly identified as a contact case. To exclude this possibility, we reanalysed the data looking at the number of families where an adult was identified as the index but one or more children were identified as asymptomatically infected (Figure 1). However, even if we assume that all asymptomatic children in these families were in fact the index case, only 6/28 (21%) children were identified as the index case in the household cluster (Table 2). These data suggest that even if children are being mistakenly overlooked as the index case, they are still likely to only have accounted for a limited percentage of household cluster transmissions of SARS-CoV-2.

It is also possible that these data were influenced by the fact that early in the SARS-CoV-2 outbreak, infections were associated with travel to outbreak areas (i.e. initially to Wuhan itself and later to the entirety of the Hubei province). Travel is much more likely to be undertaken by an adult in the family, rather than a child under the age of 16 years, potentially cofounding the results. To control for this factor, we reanalysed the data only including household transmission clusters where the index case had no history of travel or the whole family was located in (or had a history of travel) to an outbreak area. This resulted in a total 23 cases, two of which (9%) were identified as having a child as the index case in the cluster. Finally, we wished to determine if the limited number of paediatric index cases was specific to SARS-CoV-2 or if it was a typical phenomenon associated with zoonotic infections. To do so, we examined 56 family clusters of highly pathogenic H5N1 avian influenza transmission events (Table 3). Here, 54% (30/56) of transmission clusters identified children as the index case. These data suggest that whilst children can facilitate the intra-household transmission of some viral infections, the data available to date suggest that children have not played a substantive role in the intra-household transmission of SARS-CoV-2.

DISCUSSION

Understanding the pathogenesis and transmission of SARS-CoV-2 in children is of paramount importance to reducing the severity of the pandemic and implementing appropriate public health controls.

Here, we analyse previously published literature from China, Singapore, South Korea, Japan, and Iran to show that children infected with SARS-CoV-2 displayed a broad range of clinical characteristics. Whilst some children developed severe disease, others presented asymptomatically with normal chest CT findings. These data are consistent with current consensus is that clinical symptoms of COVID-19 are, on average, milder in children (although severe infections can still occur).^4–8 Similarly findings have been made in children infected with either SARS or MERS.^9,10 At present, it remains unclear why children may develop less severe forms of COVID-19. It has been suggested that the receptor for SARS-CoV-2, ACE2, may be expressed at a lower level in children than in adults.¹⁶ Alternatively, these observations may reflect the fact that the paediatric immune system is more adept than that of adults at dealing with infections for which there is no pre-existing immunity.^58,59 This is reflected in the comparatively mild infections experienced by children upon infection with measles virus and varicella zoster virus.^60,61 Understanding age-dependent differences in the pathogenesis of SARS-CoV-2 remains a key priority for future research.

We showed that of the 31 recorded SARS-CoV-2 household transmission clusters there were only three incidences of children being identified as the index case in the family. This observation is supported by previous evidence from China, where a study of 66 family transmission clusters showed that children were never the first in the family to be diagnosed with COVID-19.⁶² Similarly, a separate study of 419 household SARS-CoV-2 transmission clusters in China did not detect a single cluster where the index case was under the age of 15, and only three where the index case was under the age of 18.⁶³ Early observations from Wuhan also suggested that children were most likely to be diagnosed with SARS-CoV-2 following an exposure history to a household infection, indicating that even in epidemic areas children were more likely to be contact, rather than index cases.^64,65 The reasons for these observations remain unclear. These data may reflect an overall lower incidence of infection in children.² Alternatively, these data may reflect altered viral shedding (either in titre or duration) in children or an alternate, less efficient route of virus transmission from infected children to adults.^24,66 It is also possible that children simply have fewer interactions outside of the home than adults and are therefore less likely to be index cases. This may be magnified in countries which have implemented school closures in an attempt to control the outbreak. However, it is worth noting that Singapore did not mandate the closure of schools and yet still did not detect any household clusters of transmission of SARS-CoV-2 with children as the index case.^13,67 Similarly, should children have fewer interactions outside the home it would suggest that children would be unlikely to be the index case in any household transmission event. In contrast, our analysis of household transmission clusters of highly pathogenic H5N1 avian influenza show that children were frequently the index case. Whether this difference simply reflects the different route of transmission of these infections (poultry-to-human vs person-to-person spread) remains to be determined.

There is a growing body of evidence suggesting that asymptomatic individuals play an important role in the transmission of SARS-CoV-2 in the community^68-70; However, such studies do not negate the results of the present study, as both children AND adults can present with an asymptomatic infection. This is perhaps best demonstrated by the outbreak of SARS-CoV-2 on board the Diamond Princess cruise ship⁶⁶. Here, 634 individuals were infected. Of these infected individuals, only six were aged 0–19 years, 152 were aged 20–59 years and 476 were 60 years and older. Despite the near absence of children on the ship, 328 of the 634 infected individuals were asymptomatic. These data clearly indicate that asymptomatic infections are not limited to the paediatric population. Therefore, perhaps one of the most effective public health intervention strategies would be to identify and isolate asymptomatically infected adults, as it is these individuals who have numerous contacts outside the home and may be contributing to viral spread.

There are important caveats that need to be acknowledged when interpreting these data for public health policy. Our analysis of household transmission clusters was also based on the assumption that the incubation period for SARS-CoV-2 was approximately equivalent in children and adults. Whilst we did not observe any significant difference in the incubation time of infected children and their infected adult relatives (data not shown), others have suggested there might be a longer incubation period for SARS-CoV-2 in children⁷¹. It is also possible that the analysis performed herein was impaired by the policy within particular countries to only test contacts of a known COVID-19 patient when and if they become symptomatic. This could result in the number of asymptomatically infected individuals being underestimated. Importantly, China rapidly implemented a policy to test both symptomatic and non-symptomatic contacts, suggesting that this limitation would only apply to transmission clusters detected outside of China. Nevertheless, despite these important limitations, these data suggest that children are unlikely to be driving the household transmission of SARS-CoV-2.

As the global SARS-CoV-2 situation continues to evolve it will be important to validate these data with a larger number of case studies across a wider geographic distribution. It is also possible that as community spread of the virus becomes more common, the transmission dynamics and clinical characteristics of disease will change. However, in the interim our analyses show that, to date, children have not been the primary source of household transmission clusters.

FUNDING

KRS was supported by the Australian Research Council [DE180100512]. ACB receives funding from the National Health and Medical Research Council with an Investigator Award (1175509).

CONTRIBUTIONS

YZ and KRS contributed to data acquisition; study design; data analysis; manuscript writing and revising. CJB contributed to data analysis; manuscript writing and revising. KDH contributed to data acquisition; data analysis; manuscript writing and revising. JES contributed to data acquisition; data analysis; manuscript revising. MT, LES, ECN, JL contributed to data acquisition; data analysis. KYC contributed to the study design; data acquisition; data analysis. JLP contributed to data interpretation and critical review of the manuscript. CG contributed to the contributed to critical review of the manuscript. ACB contributed to critical review of the manuscript and reviewed for clinical content.

DECLARATION OF INTERESTS

No conflict of interests to declare.