key: cord-0770169-rws51hkl
authors: Zavala, Maria; Ireland, Georgina; Amin-Chowdhury, Zahin; Ramsay, Mary E; Ladhani, Shamez N
title: Acute and persistent symptoms in children with PCR-confirmed SARS-CoV-2 infection compared to test-negative children in England: active, prospective, national surveillance
date: 2021-11-29
journal: Clin Infect Dis
DOI: 10.1093/cid/ciab991
sha: ba10a3a34bf59ace7847b6b688321a42a81fbc00
doc_id: 770169
cord_uid: rws51hkl

BACKGROUND: Most children recover quickly after COVID-19, but some may have on-going symptoms. Follow-up studies have been limited by small sample sizes and lack of appropriate controls METHODS: We used national testing data to identify children aged 2-16 years with a SARS-CoV-2 PCR test during 01-07 January 2021 and randomly selected1,500 PCR-positive cases and 1,500 matched PCR-negative controls. Parents were asked to complete a questionnaire about the acute illness and pre-specified neurological, dermatological, sensory, respiratory, cardiovascular, gastrointestinal, mental health (including emotional and behavioural well-being) and other symptoms experienced at least five times at one month after the PCR test. RESULTS: Overall, 35.0% (859/2456) completed the questionnaire, including 38.0% (472/1242) cases and 32% (387/1214) controls. of whom 68% (320/472) and 40% (154/387) were symptomatic, respectively. The most prevalent acute symptoms were cough (249 /859, 29.0%), fever (236/859, 27.5%), headache (236/859, 27.4%) and fatigue (231/859, 26.9%). One month later, 21/320 (6.7%) of symptomatic cases and 6/154 (4.2%) of symptomatic controls (p=0.24) experienced on-going symptoms. Of the 65 on-going symptoms solicited, three clusters were significantly (p<0.05) more common, albeit at low prevalence, among symptomatic cases (3-7%) than symptomatic controls (0-3: neurological, sensory and emotional and behavioural wellbeing. Mental health symptoms were reported by all groups but more frequently among symptomatic cases than symptomatic controls or asymptomatic children. CONCLUSIONS: Children with symptomatic COVID-19 had a slightly higher prevalence of on-going symptoms than symptomatic controls, and not as high as previously reported. Healthcare resources should be prioritised to support the mental health of children.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) virus was first identified in China in December 2019 and spread rapidly across the continents, causing millions of cases and deaths worldwide [1] . In the United Kingdom, the first imported cases were identified in January 2020, with the first confirmed cases in children on 29 February 2021 [2] . typically presents as fever, cough or loss of smell or taste, and is usually self-limiting, although, in some, may progress to severe pneumonia, multiorgan failure and death [3] [4] [5] .

Age is the most important risk factor for severe and fatal COVID-19, although other factors including ethnicity, co-morbidities and obesity are also associated with increased risk [6] .

While most adults recover without sequelae, some survivors will have on-going symptoms, commonly known as long COVID or post-COVID syndrome, which may last for weeks to months after acute infection [7, 8] . So far, more than 200 on-going symptoms have been reported in adults and children [9] . The most prevalent symptoms across multiple studies include fatigue, shortness of breath and loss of smell or taste [7, 10] . In an attempt to standardise the case condition, the UK National Institute for Health and Care Excellence (NICE) has proposed two categories based on symptoms duration: ongoing symptomatic COVID-19 in patients with symptoms persisting for 4-12 weeks after an infection consistent with COVID-19 and post-COVID-19 syndrome, when symptoms develop during or after an infection consistent with COVID-19, continue for >12 weeks and are not explained by an alternative diagnosis [11] .

Unlike adults, children have a lower risk of severe disease, hospitalisation or death due to COVID-19 [2, 12] . Most children exposed to SARS-CoV-2 remain asymptomatic or develop mild, self-limiting upper respiratory tract symptoms which resolve within a few days [13] .

There have, however, been increasing reports of long COVID in children, with one online survey indicating that 90% of children did not regain normal health up to 6-8 months after A c c e p t e d M a n u s c r i p t confirmed or suspected COVID-19 [14, 15] . Many of the earlier publications, however, have been limited by a lack of laboratory-confirmed diagnosis among cases or comparison with a control group that did not have COVID-19, although more recent studies/surveys have tried to address some of these issues [13, [16] [17] [18] . This investigation aimed to determine the course of illness and on-going symptoms in children aged 2-16 years with laboratoryconfirmed SARS-CoV-2 infection compared to test-negative children in England during January 2021, when the Alpha variant was prevalent. Station, Texas), cases and controls were stratified by age in years and partial postcode in eight and 78 stratas respectively, making a total of 939 stratas. A random sample from each strata was taken to produce a total sample size of 1,500 cases with the same number of controls matched. On 22 February, an invitation letter was sent to cases and controls with an online link, a QR code to access the online survey developed using SnapSurvey v.11 (see Supplementary Questionnaire), and a paper version of the questionnaire with a stamped, A c c e p t e d M a n u s c r i p t self-addressed envelope. The questionnaire was designed to be completed by the child's parent/guardian of the child and was developed following an extensive literature review of common symptoms associated with long COVID in the scientific literature and online, including those reported in adults and assessed by the study investigators to be applicable to children. The questionnaire was designed to be completed at least one month after the initial RT-PCR test. The final questionnaire requested about demographics, COVID-19 symptoms at the time of testing, household composition, confirmed cases in the household and pre-specified symptoms at the time of the RT-PCR test and at least one month later.

Symptomatic cases and controls were defined as children with a positive or negative PCR test, respectively, during 01-07 January 2021 who reported any symptom at the time of the test. Asymptomatic cases and controls were defined as children who had a positive or negative PCR test, respectively, and whose parents reported no symptoms for their child at the time of the test. Elicited on-going symptoms in cases and controls were divided into neurological, dermatological, sensory, respiratory and cardiovascular, gastrointestinal, mental health (including emotional and behavioural wellbeing) and other, allowing parents to report any other symptoms not included in the questionnaire. Symptoms were defined as ongoing if experienced at least five times a month at least one month after the PCR test.

Two reminders were sent to the families did not complete the questionnaire two and four weeks after the initial invitation letter. Index of multiple deprivation (IMD) was used as a proxy measure for socio-economic status and was obtained using postcodes from the A c c e p t e d M a n u s c r i p t significant in the univariate analysis comparing symptomatic cases vs controls and asymptomatic cases vs controls, after adjusting for age, sex, ethnicity, socio-economic status and comorbidities. Results are reported as odd ratios (OR) and adjusted odd ratios (aOR) with 95% confidence intervals (CI).

Of the 1,500 cases and 1,500 controls identified initially, 1,257 cases and 1,217 controls had Table 1) . Symptomatic cases had mild illness and only one symptomatic case required hospitalisation. Comparison of demographic data showed no significant differences between respondents and non-respondents except older age among respondents (median age, 10 vs 6 years).

At the time of the RT-PCR test, the most prevalent symptoms among children who displayed any symptom included cough (29.0%, 249 /859), fever (27.5%, 236/859), headache (27.5%, 236/859) and fatigue (26.9%, 231/859) ( Table 2) . Thirteen of the eighteen elicited symptoms were significantly more prevalent among cases than controls ( Table 2 ). The most common reason for testing was because the child was identified as a contact of a household case A c c e p t e d M a n u s c r i p t (38.7%, 327/859), the child had COVID-19 symptoms (27.6%, 233/859) or due to another reason not specified in the survey (17.06%, 144/859), such as school testing or parental choice. More than half the symptomatic cases were tested because they were contacts of a household case (51.9%, 166/320) compared to symptomatic controls (14.3%, 22/154; p<0.001). The same was also true for asymptomatic cases (63.6%, 96/152) compared to asymptomatic controls (19.63%, 43/233; p<0.001).

One month after the RT-PCR test, a higher proportion of parents of symptomatic cases (6.73%, 21/320) reported that their child was still unwell than symptomatic controls (4.20%, 6/154; p==0.24).24). The 64 elicited on-going symptoms were categorised as follows:

neurological, dermatological, sensory, respiratory, mental health, gastrointestinal and other.

Of these, nine were significantly more prevalent among symptomatic cases compared to symptomatic controls (Table 3 ) and clustered into three categories ( Figure 2 In the univariate analysis, none of the symptoms were significantly more prevalent among asymptomatic cases compared to asymptomatic controls. Although not statistically significant, reports of mental health symptoms were abundant among both asymptomatic cases and asymptomatic controls.

At the time of their RT-PCR test, symptomatic children with laboratory-confirmed COVID-19 had a higher prevalence of elicited symptoms than test-negative symptomatic children, especially cough (36.2%), fatigue (35.0%) and fever (33.1%). One month later, 6.7% of symptomatic cases and 4.2% of symptomatic controls had on-going symptoms. Of the 64 elicited symptoms, nine were more common among symptomatic cases than symptomatic controls, but not among asymptomatic cases or asymptomatic controls.

Early reports of long-COVID, initially among hospitalised adults, and then among community A c c e p t e d M a n u s c r i p t Whilst acknowledging potential biases of the population participating in the ZOE app study (white, middleclass, female participants) and parental reporting of PCR-testing and results, a strength of this analysis was the inclusion of symptomatic but SARS-CoV-2 RT-PCR negative controls, matched 1:1 for age, gender, and week of testing. The same methodology reported significantly higher rates of persistent symptoms in adults compared to children:

13.3% lasting ≥4 weeks, 4.5% lasting ≥8 weeks and 2.3% lasting ≥12 weeks [8].

In our cohort, the reported symptoms at the time of infection and at follow-up are consistent with the reported literature for symptomatic COVID-19 in children [30, 31] . Reassuringly, the vast majority of children recovered after acute infection and, while 6.7% of symptomatic cases had on-going symptoms a month later, so did 4.2% of symptomatic controls. The difference in prevalence of on-going symptoms between symptomatic cases and symptomatic controls (2.5%) is similar to the ZOE app study for children (3.5%) but substantially different the ONS survey which reported "any symptoms ≥12 weeks after assumed date of infection" of 9.8% in 2-11 year-olds and 13.0% in 12-16 year-olds who tested positive for SARS-CoV-2 compared to 2.0% and 1.7% among test-negative controls, respectively, between April 2020 and March 2021 [27].

In our cohort, most elicited symptoms were as common among symptomatic cases as among symptomatic controls. We did, however, identify three symptom clusters among symptomatic cases, although individually these symptoms were relatively uncommon among cases. Sensory symptoms affecting taste, smell and vision had the strongest associations and were reported almost exclusively among symptomatic cases except for one symptomatic control. Similarly, too, confusion (3.8%) was only reported among symptomatic cases. These symptoms appear to be specific to COVID-19 and are well-reported in adults and children [15, 29] . Reassuringly, on-going symptoms in asymptomatic cases and control were rare. 

One of the main strengths of this study is the random sampling of PCR-confirmed SARS-CoV-2 cases from a national dataset alongside contemporaneous, matched controls, during a period when widespread testing was available nationally. We only included cases tested in the community so that the findings were not biased towards hospitalised cases with more severe illness, which is rare in children. Another strength is the minimisation of recall bias by activating the survey within two months of the children´s PCR-test. Additionally, to improve questionnaire uptake and reduce bias against families with limited internet access, a paper version of the questionnaire was offered.

There are, however, some limitations. Despite three attempts, questionnaire completion rate was 35% and there may be some selection bias between respondents and non-respondents.

Parents of children with on-going symptoms, for example, may be more likely to complete the questionnaire than those whose children recovered. Comparison of demographic data, however, found no significant differences between respondents and non-respondents except younger age among the latter (median age, 6 years), who are more likely to recover uneventfully than older children. Some children may also have been wrongly misclassified because of false-positive or false-negative PCR-tests, but this is likely to be a minority, given A c c e p t e d M a n u s c r i p t that England was experiencing high case-rates because of the alpha variant. Moreover, it is possible that some controls may have been exposed to SARS-CoV-2 but remained asymptomatic or opted not to be tested. Ideally, cases and controls would have had a SARS-CoV-2 antibody test to confirm their infection status a month after the RT-PCR test.

Given our sample size, too, comparisons are hampered by the sparsity of reported symptoms. For instance, our sample size had 80% power to detect a difference in proportion with symptoms of 6% vs. 1% (relative risk of around 6) and for more common symptoms, M a n u s c r i p t Age (median, IQR) (859/859) 10 (6, 13) 10 (6,13) 9 (6.5,12) 8 (5,12) 12 (9,14)

Age M a n u s c r i p t M a n u s c r i p t Table 3 . A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t 

Coronavirus (COVID-19) Dashboard

COVID-19 in children: analysis of the first pandemic peak in England

Clinical manifestations of COVID-19 in the general population: systematic review

Features of 20 133 UK patients in hospital with covid-19 using the ISARIC WHO Clinical Characterisation Protocol: prospective observational cohort study

Multicentre Italian study of SARS-CoV-2 infection in children and adolescents, preliminary data as at 10

Factors associated with COVID-19-related death

A c c e p t e d M a n u s c r i p t