Outcomes of SARS-CoV-2–Positive Youths Tested in Emergency Departments

Key Points Question What proportion of SARS-CoV-2–positive youths tested in emergency departments (ED) experience severe outcomes (ie, intensive interventions, severe organ impairment, or death) within 14 days? Findings Among 3221 SARS-CoV-2–positive youths enrolled in a global prospective cohort study with outcome data, 3.3% had severe outcomes within 14 days. Across a subgroup of 2510 SARS-CoV-2–positive youths discharged home after testing, 0.5% had severe outcomes during the 2-week follow-up period. Meaning The findings of this study suggest that risk factors such as age, underlying chronic illness, and symptom duration may be useful for clinicians to consider when evaluating pediatric patients with SARS-CoV-2 infection.


Introduction
2][3][4] These early estimates likely underreported the true number of children infected with SARS-CoV-2 because of testing capacity and the generally mild, or even asymptomatic, nature of the disease in children. 5,6However, the pandemic has evolved, and in the United States, youths now represent 25% of all new COVID-19 cases. 7Similarly, pediatric hospitalizations due to COVID-19, which increased 8-fold between May and November of 2020, 8 have seen a further 5-fold increase between June and August 2021. 91][12][13] The risk of severe outcomes among youths with SARS-CoV-2 infection is poorly understood with estimates varying considerably between study designs, settings, and regions. 1,14,15Studies generally include large administrative databases (ie, community based), 16 hospitalized populations, 17 and youths admitted to the intensive care unit (ICU). 139][20][21] However, data from large prospective cohort studies which include youths with early or mild stages of disease seeking emergency department (ED) care are lacking.
We sought to quantify the frequency of and risk factors for severe outcomes in SARS-CoV-2positive children enrolled in a prospective ED-based cohort study.

Setting, Design and Participants
The Pediatric Emergency Research Network (PERN)-COVID-19 prospective cohort study enrolled participants between March 7, 2020, and June 15, 2021, who were tested for SARS-CoV-2. 22Children and adolescents aged younger than 18 years who had a SARS-CoV-2 test performed because of suspected acute infection based on symptoms or exposure were eligible.Sites attempted to recruit the first 5 eligible youths each day, using a consecutive approach starting with the first test performed on chronological time.As this led to variations in the percentage of positive youths recruited across sites, the approach was revised to target the recruitment of the first 2 test-positive and 4 test-negative youths each day.For this study, we focused on youths with positive SARS-CoV-2 nucleic acid tests [eg, polymerase chain reaction (PCR)] enrolled in 38 of the PERN-COVID-19 participating EDs across 8 countries including Argentina, Canada, Costa Rica, Italy, Paraguay, Singapore, Spain, and the United States.We included those who were test-negative as a comparator group.All participants enrolled in 3 study sites in 2 countries were SARS-CoV-2 negative and were included in relevant analyses.

Data Collection
At the time of enrollment, the participant's guardian provided data regarding demographic characteristics, epidemiological risk factors, and clinical symptoms.Two weeks after the index ED visit, guardians were contacted via telephone, text, or email to determine if there were subsequent health care visits, treatments, or interventions received.This information was supplemented by a medical record review performed a minimum of 30 days after the index ED visit to collect data related to medical care and interventions provided, disposition, and clinical outcomes.Participants were deemed lost-to-follow-up at 14 days if 5 attempts to contact the participant's legal guardian were unsuccessful and a medical record review was not performed.Race and ethnicity data were only available for a subset of the study population and were not reported.

Symptoms of SARS-CoV-2 Infection
Symptom duration was defined as time from reported onset of symptoms to time of ED evaluation and was grouped according to investigator consensus as 0 to 3 days, 4 to 7 days, and 8 or more days.
Febrile or respiratory illness was defined as having any of the following symptoms: fever, cough, rhinorrhea, congestion, wheezing, difficulty breathing, sputum production, sore throat, or apnea.

SARS-CoV-2 Testing
Testing protocols varied by site and over time.Participants were classified as SARS-CoV-2 positive if they had a positive nucleic acid test (eg, PCR) during the qualifying ED encounter or in the subsequent 2 weeks in instances of repeat testing, on any of the following samples: nasal swab, nasopharyngeal swab, oropharyngeal/throat swab, and saliva.Dates of infection were grouped according to enrollment and severity trends (ie, waves).

Severe Outcomes
A priori, a severe outcome was defined by the occurrence of any of the following complications:

Sample Size
The PERN-COVID-19 prospective cohort study originally planned to recruit up to 12 500 participants to obtain a subsample of test-positive cases that included at least 50 with severe outcomes, which would provide 93.9% power to detect when a predictive model discriminates severe from nonsevere outcomes in the larger population of test-positive youths, assuming a true C statistic of 0.70. 22These calculations used a variance inflation factor of 2.0 to account for model complexity as measured by degrees of freedom. 23As outcomes were classified in a delayed manner, we ended up having an excess number of severe outcomes prior to achieving our target sample size, which led to a termination of recruitment.

Statistical Analysis
Pearson χ 2 or Fisher exact tests were used to compare the prevalence of participant characteristics across age groups.All youths with ED outcome (day 0) data were included in the primary analysis.
Risk differences were calculated by subtracting the cumulative incidence between categories of the following predefined variables: sex, age, and chronic conditions.Median length of hospitalization and ICU stay were compared across dichotomous strata using the Mann-Whitney test and across the 5 age groups using the Kruskal-Wallis test.Outcomes among children discharged home from the ED on day 0 and who completed 14-day follow-up were summarized using similar methods.
A multivariable logistic regression model that included the country and time period of enrollment, and accounted for clustering by site, was used to identify factors associated with severe outcomes.The following variables were specified a priori for inclusion into the model: sex, age group, chronic underlying condition as reported by the guardian (excluding asthma), history of asthma, prior pneumonia, symptoms of febrile or respiratory illness, duration of symptoms, and date of infection.
From this model, we derived additional models by removing nonsignificant variables (P > .05) in sequence and then computing the Akaike Information Criteria (AIC) weights for each model.The model with the lowest AIC was selected as the best fit model. 24To evaluate associations with race and ethnicity, a similar multivariable logistic regression analysis was performed for participants recruited in EDs in the United States alone.We compared the proportion of severe outcomes by time (each month) of study enrollment using a trend test that adjusted for cluster effects.
Baseline characteristics of enrolled youths who tested SARS-CoV-2 positive and negative were compared using the χ 2 test and R × C contingency tables as appropriate.Severe outcomes among SARS-CoV-2-negative children were compared with SARS-CoV-2-positive children overall and based on index ED visit hospitalization status stratified by country using risk difference calculation methods.
No missing data approaches were used as only 0.8% of participants lacked data required to complete primary outcome analyses.All analyses were 2-sided and conducted using STATA version 16 (StataCorp) with statistical significance defined as P < .05.Statistical analysis was performed from September to October 2021.
b Includes 8 youths transferred to other tertiary facilities.
c As per analysis plan, considers only youths with full 14-day follow-up.

JAMA Network Open | Pediatrics
Other severe illness, intervention, death Abbreviation: ICU, intensive care unit.
a No outcome information is known for the 1 youth without emergency department discharge/hospitalization information; and for 38 of 3221 youths, the , outcome data were missing after day 0 (26 youths lost to follow-up who were discharged at day 0) or at the time of initial hospitalization discharge (12 lost to follow-up youths discharged from hospital within 14 days).

JAMA Network Open | Pediatrics
Outcomes of SARS-CoV-2-Positive Youths Tested in Emergency Departments

Discussion
In this global cohort study, after 14 days of prospective follow-up for 3221 youths who tested positive for SARS-CoV-2 infection, 23% had been hospitalized, 3% experienced severe outcomes, and 4 children died.We identified the following risk factors for severe outcomes: aged greater than 5 years, having a preexisting chronic illness, previous episode of pneumonia, and presenting to the hospital 4 to 7 days after symptom onset.6][27] In contrast with some other studies, we did not find that very young infants were at a higher risk for severe outcomes. 18,21,27In some studies where very young infants were identified as being at higher risk, the outcome of interest was hospitalization or ICU admission, 21,27 whereas we required specific intensive care interventions or complications.As indications for ICU admission differ substantially by time and place, it is a strength of our study that we did not consider ICU admission in isolation to be an indicator of severe disease.
Although asthma has been suggested as a risk factor for severe illness in youths with COVID-19, 18 our study, as well as a registry-based study in the United States, did not confirm this association. 27To the best of our knowledge, no other studies have identified symptom duration prior to hospital presentation as a risk factor for severe pediatric COVID-19.As effective therapeutics for youths with detected SARS-CoV-2 are infrequently administered, it is unlikely that this association indicates a beneficial effect of earlier presentation to care.Rather, this may reflect the natural history of infection in youths, with symptom progression appearing between 4 and 7 days being more likely to lead to both hospital presentation and severe outcomes.Previous studies have shown that compared with White non-Hispanic persons in the United States, Black race and Hispanic ethnicity are associated with increased test-positivity and hospitalizations, [28][29][30][31] including among children. 27[34][35] Our findings from the prospective follow-up of the subgroup of SARS-CoV-2-positive youths who were discharged home from their index ED visits reflect the natural history of mild-to-moderate pediatric SARS-CoV-2 infection.Among these 2510 youths, approximately 13% returned to either a clinic or hospital for care within 14 days due to worsening, persistent or new symptoms, 2% were hospitalized, 0.5% had severe outcomes, and none died.

Strengths and Limitations
This study had some strengths and limitations.A strength of our study was our ability to compare outcomes among SARS-CoV-2 test-positive children to similar, albeit not identical, test-negative children recruited prospectively, prior to the occurrence of outcomes.Although overall the risk of severe outcomes did not differ between those who were test-positive relative to those who were test-negative, the risk of severe outcomes was higher among hospitalized SARS-CoV-2 positive children when compared with those who were test-negative.This finding does not stand in isolation; in a large retrospective study that included 242 158 children with COVID-19 and 2 084 180 with influenza infection, hospitalization (5-to 13-fold), hypoxemia, and pneumonia were more frequent in those with COVID-19. 16 recruitment of study participants took place in EDs, our findings overestimate the risk of severe outcomes among SARS-CoV-2-positive youths and should not be interpreted to reflect the risk faced by community-based cases; rather, they are meant to provide an estimate of this risk among an ED-screened pediatric population.Sites were asked to enroll participants consecutively based on time of testing, however, the proportion of potentially eligible participants consenting to participate likely varied by site and period, which challenges the representativeness of our study population to reflect all youths tested for SARS-CoV-2 in pediatric EDs.Similarly, although participating EDs were given the same study protocol for recruitment, various factors including regional case definitions, screening criteria, and testing capacity, were not controlled by the study and could have differed by site and period.Thus, 5% of our SARS-CoV-2-positive participants were asymptomatic-most of whom were tested as they were positive contacts of known cases or as part of routine screening procedures.To account for this concern, our multivariable logistic regression model adjusted for country and period.Because participating EDs were located in academic pediatric institutions, we cannot generalize our results to all community EDs nor can we generalize to countries beyond those included in our analysis.Finally, as testing for variants of concern was not universal, we were unable to include circulating variants in our model.

Conclusions
The findings from this large global prospective cohort study support a growing body of literature on the risk of severe outcomes and factors associated with these events in SARS-CoV-2-infected youths.
Among youths who sought care in the ED, these events occurred in approximately 3% of SARS-CoV-2-positive youths, with the risk varying by age, history of underlying conditions, symptom duration and country.Our findings suggest a low risk of severe outcomes among youths discharged to home.
However, among hospitalized SARS-CoV-2-positive youths, the risk of severe outcomes exceeds that of test-negative youths.

Figure 1 .
Figure 1.Flow of SARS-CoV-2-Positive Study Participants Including Outcomes and Follow-up 3222 SARS-CoV-2 positive children, March 7, 2020, to June 15, 2021 Outcomes of SARS-CoV-2-Positive Youths Tested in Emergency Departments Enrolling sites had local institutional review board approval or established a reliance agreement with the Cincinnati Children's Hospital Medical Center institutional review board.The legal guardians of all participants provided informed consent (written or verbal based on site) to participate in this study.This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

Table 2 .
Cumulative Outcomes of 3221 SARS-CoV-2-Positive Youths Within 14 Days of Index Emergency Department Visit a,b youths from the United States, similar risk factors were identified; and race and ethnicity were not associated with severe outcomes (eTable 4 in Supplement 1).

Table 3 .
Association of Demographic Factors and Medical History With Severe Outcomes Among the 3141 SARS-CoV-2-Positive Youths With Complete Covariate Data a,b b This table reflects the final parsimonious model.c Severe outcomes did not occur outside of these countries.
17ong 2510 SARS-CoV-2-positive youths discharged home, only 0.5% had severe outcomes during the follow-up period.Although the overall proportion of SARS-Similarly, while 20% of youths hospitalized in a US-based retrospective cohort study had severe or very severe disease,17in a prospective Canadian study of 150 hospitalized SARS-CoV-2-positive youths, 50% had severe or critical illness.26Thus,an understanding of the study population is crucial to interpreting risk estimates.Our study population provides a risk estimate for youths brought for ED care.Our lower estimate of severe disease likely reflects our stringent definition which required the occurrence of complications or specific invasive interventions. 25