More than 1.3 million (12%) cases of COVID-19 in children were reported in the United States as of the end of November 2020.1 Several case series have described the clinical features of COVID-19 in pediatric patients 2-8 and suggest a disease milder in children compared to adults.4,5,7-11
However, children can present with a wide spectrum of diseases ranging from asymptomatic infection to severe respiratory disease and various inflammatory complications.12-14
Severe or life-threatening illness occurs in approximately 10% to 20% of adults with COVID-19.15,16 Increasing age, male sex, and certain comorbid conditions are risk factors for severe COVID-19 in adults.17 -twenty
Medically complex children may be more likely to require intensive care, but most early pediatric studies have not evaluated specific risk factors in greater detail.4,5,7,8,11,21 Much remains to be learned about the factors epidemiological, demographic and clinical risk of severe disease from COVID-19 in children.
Colorado has reported more than 36,000 pediatric cases of SARSCoV-2 as of early December 2020, representing 14% of all cases in the state.22 Children’s Hospital Colorado (CHCO) has established one of the Larger cohorts of pediatric patients from a single institution with COVID-19 disease in the United States.
In this study, the authors evaluated risk factors for severe disease among children with SARSCoV-2 infection. The findings may help pediatric healthcare providers and public health stakeholders adapt clinical management and improve pandemic planning and resource allocation.
| Materials and methods |
> Studio design and environment
In this retrospective cohort study, the authors captured clinical and epidemiological data of all pediatric patients with SARS-CoV-2 infection at CHCO between March 15 and July 8, 2020. CHCO is the largest pediatric referral center. large for children in a 7-state region and includes a 434-bed acute care hospital in Aurora, Colorado, a 111-bed acute care hospital in Colorado Springs, Colorado, and 13 additional network locations offering patient services outpatient clinics, specialties and emergency care. All sites use a common electronic health record (EHR; Epic systems, Verona, WI).
> SARS-CoV-2 test
Molecular testing for SARS-CoV-2 was performed on nasopharyngeal swabs, nasopharyngeal lavages/aspirates, tracheal aspirates, and bronchoalveolar lavage specimens using 1 of 4 qualitative real-time reverse transcription PCR assays for molecular detection of SARS-CoV-2. , all of which have been granted Emergency Use Authorization (EUA) by the Food and Drug Administration (FDA): CDC 2019-nCoV Real-Time RT-PCR Diagnostic Panel (CDC, Atlanta, GA), Directo Simplexa COVID-19 Assay (DiaSorin Molecular LLC, Cypress, CA), Abbott RealTime SARS-CoV-2 Assay (Abbott Molecular Inc., Des Plaines, IL), and Xpert Xpress SARS-CoV-2 Assay ( Cepheid, Sunnyvale, CA).
These 4 assays had 100% positive and negative agreement in preclinical studies, did not demonstrate cross-reactivity with other human coronaviruses or with a diverse panel of other common respiratory pathogens.
> Study population
All patients <21 years of age with positive molecular tests for SARS-CoV-2 performed at CHCO were included. Patients ≥21 years of age were included only if they were followed by CHCO for a chronic medical condition. Molecular testing for SARS-CoV-2 became available on March 16, 2020, with the first positive patient identified on March 19.
For the first 4 weeks, testing was limited to patients admitted with fever and respiratory symptoms who tested negative for other respiratory viruses on a multiplex panel and symptomatic immunocompromised outpatients. On April 10, 2020, the CHCO began testing all patients for SARS-CoV-2 upon admission.
On April 18, a testing site was opened for outpatients and before procedures. They excluded patients tested outside the state of Colorado, parents/caregivers of pediatric patients, pregnant women, and healthcare workers.
> Data collection and management
The authors received a daily report from the EHR for each patient with a positive SARS-CoV-2 PCR result in the CHCO system. Demographic and clinical data were extracted from the EHR in real time and entered into standardized data collection forms developed in REDCap, hosted by the University of Colorado, Denver.
EHR data included all urgent/emergency, primary care, or specialist care services, inpatient notes, and diagnostic test results, as available.
Ethnicity was recorded as documented in the EHR. Body mass index (BMI) was calculated from height and weight documented in the EHR (measured in the past month) for children <12 months of age, within the past 6 months for those aged 1-5 years. , and within the last year for children > 5 years).
Those with a BMI in the 85th percentile to < 95th percentile for age were considered overweight, those ≥ the 95th percentile were considered obese, and those ≥ 120% of the 95th percentile for age were considered severely obese.23,24 Active comorbid conditions were recorded as documented in the EHR and categorized according to body system.
Exposure history, social history, and other risk factors for SARS-CoV-2 acquisition were recorded if documented in the EHR. The results of diagnostic tests were recorded for all hospitalized patients during the period of admission or resolution of symptoms, for outpatients from symptom onset to resolution, and for asymptomatic patients if performed any laboratory or imaging test on the same day as the SARS-CoV-2 PCR test.
Chest radiographs were considered abnormal if the impression included opacity, consolidation, or peribronchial airspace thickening.25,26 Patients were classified as requiring intensive care if they (1) were admitted to the pediatric intensive care unit (ICU). for symptomatic COVID-19, or (2) were admitted to the neonatal ICU for symptomatic COVID-19 and required a higher level of respiratory support than low-flow nasal cannula. This study was approved by the Colorado Multiple Institutional Review Board with a waiver of informed consent. Statistic analysis
The authors described demographic characteristics using summary statistics. Descriptive data include all eligible patients with SARS-CoV-2 testing and are presented as means and standard deviations; medians and interquartile ranges; or accounts and proportions.
Among all patients with symptomatic SARS-CoV-2 infection, they analyzed the odds of severe outcomes, defined as:
(1) hospitalization;
(2) need for intensive care; or
(3) need for breathing support.
Asymptomatic hospitalized patients who were incidentally identified with SARS-CoV-2 were excluded from these analyses. Simple logistic regression was performed for admission, respiratory support, and intensive care.
Multivariable regression pooled variables to identify demographic predictors, comorbidity predictors, and symptom predictors of outcomes. Each variable within a group was evaluated using the following criteria: P < 0.10 for entry and P < 0.05 for the remainder.
Those variables with >20% missing data were excluded from the multivariable analyses; these included BMI, laboratory values, baseline oxygen use, and exposure history; all other variables had zero missing values, so no imputation was required.
All analyzes were 2-tailed and P values <0.05 were considered significant. Multivariable analysis was not performed for the intensive care outcome due to the small number in this category. All analyzes were performed in SAS 9.4 (SAS Institute, Inc., Cary, NC).
| Results |
Study population
A total of 454 children and young people with a mean age of 11 years (upper limit 23 years) were identified, 42.1% being women, with SARS-CoV-2 infection during the study period. The vast majority (N = 427, 94%) were identified during expanded testing in which asymptomatic and symptomatic children could be tested.
The number of cases per week during the study period increased steadily and peaked in the second week of July. Patients identified as Latino or Hispanic contributed 54.6% of cases, followed by non-Hispanic whites (22.9%) and non-Hispanic blacks (5.1%). The proportion of patients with SARS-CoV-2 infection with Hispanic ethnicity was significantly higher than the patient population seen at CHCO in 2019 (29%, P < 0.0001).
> Clinical presentation
Among all SARS-CoV-2 PCR-positive patients, 315 (69%) were symptomatic and 80 (18%) were asymptomatic. The most common symptom at the time of positive SARS-CoV-2 PCR was fever (>100.4 F) (27%), followed by cough (23%), and congestion or rhinorrhea (18%).
The most frequently documented risk factor for exposure to COVID-19 was a family member testing positive for SARS-CoV-2 (N = 114, 25%) followed by social gatherings of more than 10 people (N = 85 , 19%).
Symptomatic cases more frequently reported these risk factors with 98 (31%) reporting a family member with COVID-19 versus 16 (12%) of asymptomatic cases and 79 (25%) of symptomatic cases reporting a meeting social compared to 6 (4%) of asymptomatic cases.
More than one-third of symptomatic cases (N = 115, 37%) had no documented exposure, while more than two-thirds (N = 98, 71%) of asymptomatic cases had no documented risk factor.
Almost half (45%) of children with SARS-CoV-2 had at least 1 comorbid condition. The most common types of comorbidities identified were pulmonary (16.7%), gastrointestinal (10.8%), and neurological disease (10.6%). Of the 211 children with BMI data available, almost half (45%) were overweight. Of these, 30% were classified as obese and 11% were considered severely obese.
A total of 85 children (19%) were admitted, of whom 66 (78%) were symptomatic. The remaining 19 patients (22%) were admitted for other reasons and were never symptomatic for COVID-19.
Among the 66 children admitted with symptoms, 55% required respiratory support and 17% required critical care. A total of 40 admitted patients had chest radiographs performed on admission, of which 25 (63%) had abnormal findings. Five of 39 (13%) inpatients evaluated with a panel of respiratory pathogens had coinfection with an additional respiratory tract virus.
> Treatment and results
Of patients admitted with symptomatic COVID-19, the median hospital stay was 3 days (IQR, 1-6). Among the 36 children who required respiratory support, 26 (72%) received low-flow oxygen through the nasal cannula, 2 (6%) high-flow oxygen through the nasal cannula, 5 (14%) pressure ventilation non-invasive positive, 2 (6%) intubation/ventilation and 1 (3%) extracorporeal membrane oxygenation (ECMO).
Eight patients (12%) received antiviral therapy for COVID and 10 (15%) received immunomodulatory therapy directed against COVID-19. Five patients required readmission for COVID-19-related illness, 1 of whom was readmitted twice. Three patients (4.5%) had complications of venous thromboembolism. One medically complex patient with multiple pre-existing comorbidities died from complications of ECMO.27 All other patients were discharged from the hospital.
> Risk factors for severe COVID-19
Risk factors associated with symptomatic hospital admission were evaluated. There were no sex or race/ethnicity differences associated with need for admission. Compared with children aged 11 to 15 years, infants aged 0 to 3 months [odds ratio (OR), 7.86; P < 0.001] and young adults > 20 years (OR, 5.09; P = 0.03) were more likely to require admission. Comorbid conditions emerged as significant predictors of admission.
The presence of any comorbid condition increased the odds of admission (OR, 2.73; P = 0.0003), and the odds increased with each additional comorbidity (OR, 1.36; P < 0.0001).
Several categories of comorbidities increased the risk of admission, including pulmonary, gastrointestinal, endocrine, neurological, and psychiatric diseases; immunocompromising conditions; and history of premature birth.
Specific diagnoses including asthma, obstructive sleep apnea, baseline oxygen requirement, and diabetes or prediabetes were significantly associated with admission.
Obesity resulted in more than double the odds of admission and severe obesity increased the odds almost 5 times. Symptoms of respiratory tract infections (including fever, cough, shortness of breath) were predictive of admission, as were diarrhea, abdominal pain, and fatigue.
Many of the same risk factors were associated with the need for respiratory support and intensive care. Demographic risk factors associated with the need for respiratory support included Hispanic ethnicity, age 0 to 3 months, or >20 years.
Comorbid conditions such as obesity and asthma were associated with the need for respiratory support, as were several symptoms of respiratory infection at the time of SARS-CoV-2 testing.
The need for intensive care was associated with obstructive sleep apnea (OR, 4.7; P = 0.04) as well as elevated C-reactive protein (CRP) at admission. Median CRP among patients requiring intensive care was 17.7 mg/dl (IQR, 5.3-22.9) compared to those not requiring intensive care (1.95 mg/dl; IQR, 0 ,7–5,5).
For every 1-unit increase in CRP on admission, the odds of requiring critical care increased by 1.2 (P = 0.02). Other laboratory values on admission were not associated with the need for intensive care, including white blood cell count, absolute neutrophil count, platelet count, procalcitonin, ferritin, D-dimer, and lactate.
In multivariable analysis, age 0 to 3 months or > 20 years continued to be a significant demographic predictor of need for admission and respiratory assistance. Comorbid conditions that remained significantly associated with admission included immunocompromising conditions, gastrointestinal disease, history of preterm birth, asthma, and diabetes/prediabetes, while only gastrointestinal disease and asthma remained significantly associated with respiratory support.
Fever, shortness of breath, and vomiting at presentation remained significantly associated with admission and respiratory support.
| Discussion |
In this large pediatric cohort identified in a network of tertiary referral hospitals, the authors identified important factors associated with severe COVID-19 in children. Importantly, they focused on symptomatic children who were admitted and required respiratory support or who received critical care to better inform healthcare providers about the population at risk for severe COVID-19.
In this cohort, 1 in 5 symptomatic children with SARS-CoV-2 infection required hospital admission. Clear risk factors for admission or need for respiratory support emerged, including extremes of age, obesity, and other underlying comorbidities. Elevated CRP was predictive of the need for intensive care.
These findings support demographic trends identified in the literature from the United States and elsewhere.3–5 Similar to other studies, they found a bimodal age distribution among those requiring admission with infants and young adults having the highest probability of being admitted.5,11,28
They detected a male predominance among children and young people who tested positive, but sex was not a predictive factor for severe COVID-19. Almost 70% of the children in the cohort with positive PCR for SARS-CoV-2 were symptomatic, and 21% required hospital admission and 3.5% required critical care, similar to other pediatric reports.4,5
Encouragingly, the length of stay was relatively short and the need for intubation/mechanical ventilation or ECMO was infrequent, consistent with published literature on children with COVID-19.4,29,30
They identified an ethnic disparity with overrepresentation of children who identify as Hispanic compared to the population served by CHCO and the general Colorado population.11,31 Racial and ethnic disparities have emerged as a key predictor of COVID-19 in adults and children throughout the United States. 9,32–37 In Colorado, while Hispanic and Latino populations make up 21.7% of Colorado residents, they have accounted for 36.9% of COVID-19 cases.22
This study demonstrates a greater disparity among Hispanic children and youth in Colorado. Furthermore, Hispanic ethnicity was associated with increased odds of respiratory support in univariate analysis but not in multivariate analysis, indicating that some other factor among Hispanic children may have been confounding this association.
Several factors may be contributing to these ethnic and racial disparities, which merit further exploration, including a higher prevalence of medical comorbidities, high-density settings, stigma, and barriers to accessing health care.38-41
This study investigated risk factors for severe COVID-19 defined as need for admission, respiratory support, or intensive care. They identified particular comorbid conditions, including asthma, diabetes and obesity, to predict more severe pediatric COVID-19 illnesses.
There is growing evidence that children with comorbidities (particularly medically complex patients) may be at increased risk of hospitalization and intensive care.3,5,42
One-third of the children in this study, who were categorized as having gastrointestinal comorbidity, required a gastrostomy or jejunostomy tube, supporting the association of severe COVID-19 with medical complexity.
Among children with endocrine disease, one-third had diabetes mellitus or prediabetes, which are related to COVID-19 outcomes in adult studies.43,44 Obesity has emerged as an independent risk factor for severe COVID-19 in adults. , with increasing evidence of the same trend in children.11,17-20,45,46
Several factors may be contributing to this association, including a higher prevalence of other medical comorbidities, obesity-related complications, race/ethnicity, and socioeconomic or behavioral factors.46,47 Thirty percent of patients in this cohort with BMI data were obese, which is alarming, as the reported obesity rate in children ages 10-17 in Colorado is only 10.7%.48
Obesity and severe obesity in this cohort were significantly associated with admission and respiratory support; however, due to missing data, they were unable to include this variable in the multivariable analyses.
Importantly, elevated CRP emerged as a clear risk factor for children requiring intensive care, consistent with 2 other US studies in children.3,49,50
A history of sleep apnea also emerged as a risk factor for intensive care in this cohort. However, we did not identify any other demographic or clinical factors associated with the need for intensive care, possibly due to inadequate power to detect these associations given the small number of intensive care admissions in this cohort.
This is one of the largest pediatric cohorts reported with more than 400 children and youth. They conducted a comprehensive analysis of factors associated with more severe COVID-19 in children, including demographics, comorbid conditions, and symptoms at presentation. The authors focused on symptomatic children to ensure that children admitted for other reasons did not skew the results. Additionally, they performed a multivariate analysis to address potential confounders.
Data were limited to those available in the EHR; Although sufficient data were available for most participants, limited information was available in some outpatient cases. Changing recommendations about who to test may also have affected the findings; The early focus on testing only symptomatic admitted children may have biased the cohort toward including more sick children, but only 6.9% of the cohort was diagnosed while restricted testing was being performed, so we do not anticipate this has created significant bias.
In conclusion, they found that age, comorbid conditions, and elevated CRP levels are risk factors for severe COVID-19 in children. Findings from this study can inform pediatric providers and public health officials to tailor clinical management, pandemic planning, and resource allocation.
Counseling for families with children with comorbid medical conditions should include a discussion of increased risk of serious illness. Healthcare providers may consider CRP screening at admission to inform the need for intensive care. Additional research should evaluate approaches to mitigate these risk factors and explore associations of ethnicity and COVID-19 in children.
