Risk Factors for Severe COVID-19 in Children

As described previously,^2,24  the Coronavirus Disease 2019–Associated Hospitalization Surveillance Network (COVID-NET) conducts population-based surveillance of laboratory-confirmed COVID-19–associated hospitalizations in 250 acute-care hospitals located in a defined catchment area that includes ∼10% of the US population. COVID-NET conducts surveillance in 99 counties across 14 states (California, Colorado, Connecticut, Georgia, Iowa, Maryland, Michigan, Minnesota, New Mexico, New York, Ohio, Oregon, Tennessee, and Utah) located in each of the 10 Health and Human Services Department regions.

Hospitalizations that meet the case definition of a COVID-19–associated hospitalization are included in surveillance if the patient resided in the catchment area and had a positive molecular or rapid antigen SARS-CoV-2 test during hospitalization or up to 14 days before admission. COVID-NET cases can include patients with acute COVID-19, MIS-C, or asymptomatic SARS-CoV-2 infection detected through screening. Trained surveillance officers identify cases using notifiable disease and laboratory databases as well as hospital admission and infection control practitioner logs. Medical chart abstraction is completed for all COVID-19–associated hospitalizations among children aged <18 years by using a standard case report form; hospitalizations are then reported to COVID-NET. This activity was reviewed by the Centers for Disease Control and Prevention (CDC) and conducted consistent with applicable federal law and CDC policy.¹  Participating sites obtained approval from their respective state and local institutional review boards, as required.

During March 2020 to May 2021, 3106 children with COVID-19 –associated hospitalization were identified by COVID-NET. To identify risk factors for severe COVID-19 among hospitalized children, we analyzed data from 2293 (73.8%) hospitalizations. Hospitalizations were excluded if chart abstraction was incomplete (n = 81; 2.6%), the patient was pregnant at the time of admission (n = 127; 4.1%), or outcome data were unknown (n = 15; 0.5%). Additionally, we excluded 718 children who had any of the following primary reasons for admission, which were not likely related to COVID-19: psychiatric admissions (n = 270), obstetrics or gynecology (n = 162), trauma (n = 136), inpatient procedures (n = 118), or other with no symptoms consistent with COVID-19 at the time of admission (n = 32). Excluded children differed on demographic and medical characteristics, compared with included children (Supplemental Table 4).

The dependent variable was severe COVID-19, defined as requiring ICU admission or invasive mechanical ventilation or in-hospital death.

Demographic variables included age group, sex, housing type, and race and/or Hispanic ethnicity group (Hispanic, non-Hispanic Black, non-Hispanic White, non-Hispanic Asian or Pacific Islander, non-Hispanic other or unknown). Children with unknown ethnicity (n = 79; 3.5%) were presumed to be non-Hispanic.

COVID-NET collects information on 14 categories of underlying medical conditions (Supplemental Table 5); other medical conditions reported in free text were categorized after review by a pediatrician. The underlying conditions considered for each age group were determined on the basis of clinical relevance and sample size. Among children aged <2 years, 7 underlying conditions were considered: airway abnormality, cardiovascular disease, chronic lung disease, feeding tube dependence, neurologic disorders, prematurity (gestational age: <37 weeks), or other conditions (immunocompromised condition, gastrointestinal or liver disease, chronic metabolic disease, blood disorders, renal disease, or other condition). Among children aged 2 to 17 years, 13 underlying conditions were considered: airway abnormality, asthma, blood disorders, cardiovascular disease, developmental delay, diabetes mellitus (type I or 2), feeding tube dependence, immunocompromised conditions, obesity (BMI kg/m² ≥95th percentile for age and sex based on CDC growth charts; International Classification of Diseases, 10th Revision, codes for obesity; or obesity selected on the case report form), nonasthma chronic lung disease, nondiabetes chronic metabolic disease, nondevelopmental delay neurologic disorders, or other conditions (gastrointestinal or liver disease; renal disease; or rheumatologic, autoimmune, or inflammatory disease).

To identify risk factors for severe COVID-19, we specified bivariate and multivariable log-linked Poisson generalized estimating equations in SAS (version 9.4; SAS Institute, Inc, Cary, NC) using robust variance estimators to account for clustering of hospitalizations within 10 Health and Human Services Department regions. We present unadjusted and adjusted risk ratios (aRRs), 95% confidence intervals (CIs), and P values using a type I error rate of 5%. Multivariable models identifying demographic characteristics associated with severe COVID-19 included all children aged <18 years and were adjusted for the presence of ≥1 underlying medical conditions. Multivariable models identifying underlying medical conditions associated with severe COVID-19 were adjusted for demographic characteristics and specified separately for children aged <2 years and 2 to 17 years because some underlying medical conditions (eg, prematurity and obesity) are only clinically relevant for specific pediatric subgroups. In supplementary analyses, we stratified models by additional age groups (ie, <6 months, 6–23 months, 2–4 years, 5–11 years, and 12–17 years) and by race and ethnicity group among children aged 2 to 17 years. We also conducted sensitivity analyses, excluding patients with a discharge diagnosis of MIS-C.⁸ 

Age-adjusted cumulative population-based rates of severe COVID-19 during March 2020 to May 2021 were calculated by using the number of catchment area residents aged <18 years hospitalized with severe COVID-19 as the numerator and 2019 bridged-race postcensal population estimates from the National Center for Health Statistics as the denominator.²⁵  Rates and rate ratios (RRs) by age group, sex, and race and ethnicity groups are presented. Hospitalizations with complete data on age group, sex, and race and ethnicity were included; no additional exclusion criteria were applied to the numerator data. Rates with relative SEs ≥30% were suppressed.

Of 2293 pediatric hospitalizations, 745 (32.5%) were infants and children aged <2 years; 1548 (67.5%) were children aged 2 to 17 years (Table 1). One-half (53.4%) were male, and the median age was 7 years (interquartile range [IQR]: 1–14). Most were Hispanic (33.7%) or non-Hispanic Black (32.1%), followed by non-Hispanic White (22.9%) or non-Hispanic Asian/Pacific Islander (5.2%). More than one-half (55.0%) had ≥1 underlying medical conditions, although the prevalence varied by age group (28.7% among infants and children <2 years old; 67.7% among children 2‒17 years old). The most common underlying conditions were obesity, chronic lung disease, neurologic disorders, cardiovascular disease, and blood disorders. Within the latter categories, asthma, developmental delay, congenital heart disease, and sickle cell disease were the most common conditions respectively.

Among hospitalized children aged <18 years, 30.1% had severe COVID-19 (Table 1). In multivariable analyses, the risk of severe COVID-19 among hospitalized children was higher among children with ≥1 underlying medical condition (aRR: 1.5; 95% CI: 1.2–1.9; P = .001; Fig 1). Severe disease was significantly less likely in infants aged <6 months (aRR: 0.7; 95% CI: 0.5–0.9; P = .004). Race and ethnicity group were not statistically significantly associated with severe COVID-19. In sensitivity analyses excluding children with MIS-C, children living in a congregate, other, or unknown residence type also had a higher risk of severe disease (aRR: 1.5; 95% CI: 1.0‒2.2; P = .3) and children aged 5 to 11 years had a lower risk of severe COVID-19, relative to children aged 12 to 17 years (aRR: 0.9; 95% CI: 0.8–1.0; P = .3; Supplemental Table 6).

Among hospitalized children aged <2 years, 22.0% had severe COVID-19 (Table 1). In multivariable analyses, the risk of severe COVID-19 was higher among children with chronic lung disease (aRR: 2.2; 95% CI: 1.1‒4.3; P = .03), neurologic disorders (aRR: 2.0; 95% CI: 1.5‒2.6; P < .0001), cardiovascular disease (aRR: 1.7; 95% CI: 1.2‒2.3; P = .004), prematurity (aRR: 1.6; 95% CI: 1.3‒2.1; P ≤ .0001) or airway abnormality (aRR: 1.6; 95% CI: 1.1‒2.2; P = .02; Fig 2A). Other conditions were not significantly associated with an increased risk of severe COVID-19. Results from sensitivity analyses excluding children with MIS-C were similar (Supplemental Table 7).

Among hospitalized children aged 2 to 17 years, 34.0% had severe COVID-19 (Table 1). In multivariable analyses, the risk of severe COVID-19 was higher among children with feeding tube dependence (aRR: 2.0; 95% CI: 1.5‒2.5; P < .0001), diabetes mellitus (aRR: 1.9; 95% CI: 1.6‒2.3; P < .0001) and obesity (aRR: 1.2; 95% CI: 1.0‒1.4; P = .0003; Fig 2B). Other conditions were not significantly associated with increased risk of severe COVID-19. In sensitivity analyses excluding children with MIS-C, developmental delay was also a risk factor for severe COVID-19 (aRR: 1.3; 95% CI: 1.1‒1.6; P = .004; Supplemental Table 7).

In sensitivity analyses among additional age groups, the categories of underlying medical conditions associated with increased risk of severe disease varied by age groups (Supplemental Table 8). Among children aged 6 to 23 months, chronic lung disease (aRR: 2.4; 95% CI: 1.3–4.3), neurologic disorders (aRR: 2.3; 95% CI: 1.8–2.9), and cardiovascular disease (aRR: 1.4; 95% CI: 1.2–1.5) were associated with increased risk of severe COVID-19. Among children aged 2 to 4 years, feeding tube dependence (aRR: 2.1; 95% CI: 1.1–3.9) and chronic metabolic disease (aRR: 1.3; 95% CI: 1.1–1.5) were associated with increased risk of severe COVID-19. Among children aged 5 to 11 years, obesity (aRR: 1.4; 95% CI: 1.2–1.6) was associated with increased risk of severe COVID-19. Among children aged 12 to 17 years, feeding tube dependence (aRR: 3.0; 95% CI: 2.6–3.5), chronic metabolic disease (aRR: 1.7; 95% CI: 1.5–1.9), and obesity (aRR: 1.3; 95% CI: 1.0–1.6) were associated with increased risk of severe COVID-19. Other conditions were not significantly associated with increased risk of severe COVID-19.

In sensitivity analyses among children aged 2 to 17 years hospitalized with COVID-19, the categories of underlying medical conditions associated with increased risk of severe disease varied by race and ethnicity group (Supplemental Table 9). Among Hispanic children aged 2 to 17 years, chronic metabolic disease (aRR: 1.6; 95% CI: 1.1‒2.5; P = .03) and obesity (aRR: 1.4; 95% CI: 1.0‒1.8; P = .03) were associated with increased risk of severe COVID-19. Among non-Hispanic Black children aged 2 to 17 years, obesity (aRR: 1.2; 95% CI: 1.0‒1.4; P = .01) and cardiovascular disease (aRR: 1.1; 95% CI: 1.1‒1.1; P < .0001) were associated with increased risk of severe COVID-19. Among non-Hispanic White children aged 2 to 17 years, chronic metabolic disease (aRR: 1.6; 95% CI: 1.1‒2.3; P = .01), obesity (aRR: 1.5; 95% CI: 1.0‒2.2; P = .04), and feeding tube dependence (aRR: 2.1; 95% CI: 1.3‒3.3; P = .002) were associated with increased risk of severe COVID-19. Other underlying medical conditions were not significantly associated with increased risk of severe COVID-19 among race and ethnicity groups.

Among children <18 years of age hospitalized with COVID-19, 12 (0.5%) died during hospitalization (Table 2). The median age was 7 years (IQR: 0–14), and most were male (n = 7; 58%), Hispanic (n = 6; 50%), or non-Hispanic Black (n = 4; 33%). The majority (n = 10; 83%) had ≥1 underlying medical conditions; neurologic disorders (n = 7; 58%) were the most common underlying condition.

During March 2020 to May 2021, the overall cumulative population-based rate of hospitalization was 43.2 per 100 000 children aged <18 years, and the rate of severe COVID-19 was 12.0 per 100 000 children aged <18 years (Table 3). Hospitalization rates were highest among infants aged <12 months (177.5 per 100 000; RR: 3.2 versus children aged 2–17 years), Hispanic children (71.2 per 100 000; RR: 3.3 versus non-Hispanic White children), and non-Hispanic Black children (63.0 per 100 000; RR: 2.9 versus non-Hispanic White children). Rates of severe disease were highest among infants aged <12 months (36.8 per 100 000; RR: 2.4 versus children 2–17 years), Hispanic children (17.6 per 100 000; RR: 3.3 versus non-Hispanic White children), and non-Hispanic Black children (21.1 per 100 000; RR: 3.9 versus non-Hispanic White children).

Almost one-third of hospitalized children with SARS-CoV-2 infection required ICU admission or invasive mechanical ventilation, and children with specific underlying medical conditions were at greater risk of severe COVID-19. A strength of this investigation was the large, geographically diverse sample of children hospitalized with SARS-CoV-2 infection, which enabled the identification of demographic and medical characteristics associated with severe COVID-19 among pediatric subgroups, including by age and race and ethnicity. These results provide needed descriptive information about COVID-19 severity in children before widespread availability of pediatric COVID-19 vaccination and can serve as a baseline to assess changes in trends as COVID-19 vaccines are approved for use in younger age groups, to compare with severe adverse vaccine reactions, and as new SARS-CoV-2 variants emerge. Additionally, information about pediatric risk factors for severe COVID-19 can inform clinical decision-making by identifying children who may benefit from closer monitoring after hospitalization. These results may also guide other prevention measures, including health education and risk communication campaigns and recommendations for vaccination,^26–28  once COVID-19 vaccines are approved for use among younger children.

Similar to previous investigations,^29,30  we found that the presence of ≥1 underlying medical conditions was associated with increased risk of severe COVID-19 and identified the specific underlying conditions associated with severe COVID-19 within pediatric subgroups. Results underscore the importance of obesity and diabetes, which have previously been documented as risk factors for severe COVID-19 among both adults and children^{4,12,24,30–33}  and identify additional risk factors, including neurologic and cardiovascular disease, feeding tube dependence, airway abnormality, and prematurity among specific pediatric subgroups. These results highlight the potential importance of neurologic disorders (including developmental delay), which were reported in more than one-half of the 12 in-hospital pediatric deaths and were associated with increased risk of severe COVID-19 across several pediatric population subgroups, including children 2 to 17 years of age without a discharge diagnosis of MIS-C and infants and children <2 years. Neurologic disorders have been shown to increase risk of severe illness in other respiratory diseases, potentially through decreased muscle tone and strength, impaired mobility, or structural conditions that diminish pulmonary function.^34–36  Consistent with findings from influenza-associated hospitalizations,³⁷  we found that some underlying medical conditions, including immunocompromised conditions and blood disorders, were not associated with increased risk of severe COVID-19, which may be explained by lower thresholds for hospital admission among children with conditions such as sickle cell disease.

When we examine population-based rates, which have not been calculated in most other studies, infants aged <12 months had the highest rates of hospitalization and severe COVID-19, compared with those of all other pediatric age groups, an important finding when assessing infant risk from COVID-19 disease.³  However, similar to other studies,^4,12,29  we found that, once hospitalized with SARS-CoV-2 infection, infants were not at significantly increased risk of severe COVID-19 relative to older children. Young infants may have a lower threshold for admission compared with older children and, therefore, some hospitalized infants may not be as seriously ill from COVID-19 illness as hospitalized older children, potentially obscuring the true risk to infants compared with older children. More research is needed to evaluate the risk of severe COVID-19 among infants relative to older children.

Consistent with other studies,^2,3  Hispanic and non-Hispanic Black children had higher population-based rates of hospitalization and severe disease relative to non-Hispanic White children. However, as reported elsewhere,^6,12,23,29  once hospitalized with SARS-CoV-2 infection, Hispanic and Black children aged 2 to 17 years were not at increased risk of severe COVID-19 relative to White children, after controlling for the presence of ≥1 underlying medical conditions. These results may suggest that Hispanic and non-Hispanic Black children may be at greater risk of SARS-CoV-2 infection, COVID-19 illness, and associated hospitalization but are not necessarily at greater risk of severe disease outcomes after accounting for variation in the prevalence of underlying medical conditions. Hispanic and non-Hispanic Black children may be at greater risk of SARS-CoV-2 infection relative to non-Hispanic White children by, for example, increased risk of infection with SARS-CoV-2 among household members who may be disproportionately represented among essential occupations, structural barriers to accessing health care, or other mechanisms.^38–40 

Limitations of this investigation include geographic and temporal variability in testing availability, capacity, and performance across contributing sites. Additionally, this investigation may have had limited statistical power to detect differences in severe COVID-19, particularly by less prevalent underlying medical conditions or among pediatric subgroups. In addition, ICU admission and invasive mechanical ventilation may not be proxies for disease severity for all children, particularly if the threshold for admission to the ICU for monitoring varied by pediatric subgroup. Also, children could have been misclassified if underlying conditions were not noted on their electronic medical records or case report form. Finally, these results are from a network of acute-care hospitals in 14 states and may not be generalizable to all hospitalized children with SARS-CoV-2 infection in the United States.

Children experience severe COVID-19, and, in hospitalized children, the presence of specific underlying medical conditions may be associated with greater risk of severe COVID-19 outcomes. Results provide baseline information about disease severity among children before widespread pediatric COVID-19 vaccination, and can be used to inform clinical decision-making, monitor population-based trends over time, and improve risk communication.

The COVID-NET Surveillance Team Authors and Affiliations are as follows: Pam Daily Kirley, MPH (California Emerging Infections Program); Nisha Alden, MPH (Colorado Department of Public Health and Environment); Kimberly Yousey-Hindes, MPH, CPH (Connecticut Emerging Infections Program, Yale School of Public Health); Emily Fawcett, MPH (Georgia Emerging Infections Program; Georgia Department of Public Health; Atlanta Veterans Affairs Medical Center); Patricia A. Ryan, MS (Maryland Department of Health); Justin Henderson, MPH (Michigan Department of Health and Human Services); Ruth Lynfield, MD (Minnesota Department of Health); Sarah A. Khanlian, MPH (University of New Mexico, New Mexico Emerging Infections Program); Grant Barney, MPH (New York State Department of Health); Christina B. Felsen, MPH (University of Rochester School of Medicine); Jess Shiltz, MPH (Ohio Department of Health); Nasreen Abdullah, MD, MPH (Public Health Division, Oregon Health Authority); William Schaffner, MD (Vanderbilt University Medical Center); Mary Hill, MPH (Salt Lake County Health Department)

We acknowledge with gratitude the clinicians and surveillance officers at the COVID-NET sites included in this study. We also wish to acknowledge the following people: Onika Anglin (COVID-NET Surveillance Team, Centers for Disease Control and Prevention; General Dynamics Information Technology); Jennifer L. Milucky (COVID-NET Surveillance Team, Centers for Disease Control and Prevention); Roxanne Archer, Brooke Heidenga, Jeremy Roland, Maria Rosales, Monica Napoles (California Emerging Infections Program); Rachel Herlihy, Sarah McLafferty, Millen Tsegaye (Colorado Department of Public Health and Environment); Amber Maslar, Paula Clogher, Adam Misiorski, Christina Parisi, Maria Correa, Tessa Carter, Carol Lyons, Daewi Kim, Gaggan Brar, Linda Niccolai (Connecticut Emerging Infections Program, Yale School of Public Health); Allison Roebling, Katelyn Ward, Jana Manning, Asmith Joseph, Chandler Surell, Gracie Chambers, Grayson Kallas, Lauren Russell, Daniel Pizarro, Jeremiah Williams, Rayna Ceaser, Stephanie Lehman, Taylor Eisenstein, Suzanne Segler, Kyle Openo (Georgia Emerging Infections Program, Georgia Department of Public Health; Veterans Affairs Medical Center; Foundation for Atlanta Veterans Education and Research); Kenzie Teno (Iowa Department of Public Health); Elisabeth Vaeth, Cindy Zerrlaut, David Blythe, Alicia Brooks (Maryland Department of Health); Rachel Park, Michelle Wilson (Maryland Emerging Infections Program – The Johns Hopkins Bloomberg School of Public Health); Jim Collins, Shannon Johnson, Sue Kim, Alexander Kohrman, Sam Hawkins, Val Tellez Nunez, Lauren Leegwater, Sierra Peguies-Khan, Chloe Brown (Michigan Department of Health and Human Services); Kathy Como-Sabetti, Austin Bell, Kalyla Bilski, Emma Contestabile, Claire Henrichsen, Katherine Schleiss, Samantha Siebman, Emily Holodick, Lisa Nguyen, Kristen Ehresmann, Richard Danila (Minnesota Department of Health); Kathy M. Angeles, Emily B. Hancock, Yadira Salazar-Sanchez, Meaghan Novi, Nancy Eisenberg, Melissa Christian, Dominic Rudin (New Mexico Emerging Infections Program); Nancy Spina, Suzanne McGuire, Adam Rowe, Kerianne Engesser (New York State Department of Health); Sophrena Bushey, Kevin Popham, Virginia Cafferky, Christine Long, RaeAnne Kurtz, Maria Gaitan (University of Rochester School of Medicine and Dentistry); Ama Owusu-Dommey, Breanna McArdle (Public Health Division; Oregon Health Authority); Kylie Seeley (Oregon Health & Science University School of Medicine); Katie Dyer, Karen Leib, Tiffanie Markus, Terri McMinn, Danielle Ndi, John Ujwok, Anise Elie, Kathy Billings, Manideepthi Pemmaraju (Vanderbilt University Medical Center); Amanda Carter, Andrea George, Andrea Price, Andrew Haraghey, Ashley Swain, Caitlin Shaw, Ian Buchta, Ilene Risk, Jake Ortega, Laine McCullough, Melanie Crossland, Ryan Chatelain, Tyler Riedesel (Salt Lake County Health Department).

aRR

adjusted risk ratio

CDC

Centers for Disease Control and Prevention

CI

confidence interval

COVID-19

coronavirus disease 2019

COVID-NET

Coronavirus Disease 2019–Associated Hospitalization Surveillance Network

ICD-10

International Classification of Diseases, 10th Revision

IQR

interquartile range

MIS-C

multisystem inflammatory syndrome in children

SARS-CoV-2

severe acute respiratory syndrome coronavirus 2

RR

rate ratio