Characteristics of Hospitalized Children With SARS-CoV-2 in the New York City Metropolitan Area

This was a multicenter, retrospective cohort study including patients aged 0–21 years who were admitted to any of the pediatric acute or critical care units during hospitalization at a large health care system and its academic affiliate, between March 1 and May 10, 2020. All patients were managed until discharge or end of the study period (June 10, 2020). The patients were admitted to acute or critical care units on the basis of individual unit admission guidelines at clinical teams’ discretion. The diagnosis of SARS-CoV-2 was made by real-time reverse transcriptase polymerase chain reaction of nasopharyngeal swabs of patients with any evolving symptoms of COVID-19 as described by the Centers for Disease Control and Prevention,¹⁶  such as fever, respiratory symptoms, fatigue, headache, gastrointestinal symptoms, and new loss of taste or smell and on risk assessment based on history. The testing was mostly uniform across all institutions due to being a part of the same health care system or university affiliation. Initially, testing samples were sent to New York City (NYC) Department of Health and Mental Hygiene for processing. By April 1, all participating hospitals were performing “in-house” testing after Food and Drug Administration granted Emergency Use Authorization of reverse transcriptase polymerase chain reaction test for SARS-CoV-2 (Roche Cobas SARS-CoV-2 Test and Cepheid Xpert Xpress SARS-CoV-2 Test). The local institutional review boards approved the study with a waiver of consent and relevant data user agreements. Seven of these patients were included in other case series publications.^7,8,17,18 

Information collected included demographics, presenting symptoms, laboratory and imaging results, treatments, and clinical outcomes. Patients were subdivided and compared by highest level of care during admission (acute care versus critical care unit) and by age groups (<1, 1–5, 6–14, and 15–21 years). Further subanalyses of the data were performed comparing patients who received any respiratory support to those who required no respiratory support during hospitalization. BMI was calculated by using a BMI calculator from Centers for Disease Control and Prevention for age and sex, and subjects were categorized as obese if they were ≥95th percentile or BMI ≥30 and as overweight if they were 85th–95th percentile or BMI 25 to <30 for age ≥2 years.¹⁹  Because of the nature of the study, data collection for obtaining history was not standardized within the cohort, and all testing and management plans were undertaken at the clinical team’s discretion. Categorical variables were compared by using Pearson χ² test or Fisher exact test. Medians of continuous variables were compared by using the Mann–Whitney test, and means were compared by using Student t test. GraphPad Prism8 software was used for statistical analysis. A P value of <.05 was considered significant.

We identified 82 hospitalized children who tested positive for SARS-CoV-2. The majority were male (63%), and the median age was 5 years (interquartile range [IQR]: 2.5 months to 15.2 years) (Table 1). Among all patients, 23 (28%) were admitted to critical care units during hospitalization. One patient required venovenous extracorporeal membrane oxygenation (ECMO). At the end of the study period, all patients remained alive and were discharged from the hospital. The median duration from symptom onset to hospital admission was 1 day (IQR: 1–4 days) and length of stay was 3 days (IQR: 2–5 days). The majority of the patients were admitted because of respiratory distress, sepsis evaluation, or severity of gastrointestinal symptoms.

About half of all patients (49%) were Hispanic and Latino, 22% were non-Hispanic white, 9% were non-Hispanic African American, 5% were non-Hispanic Asian, and 15% were multiple, other, or unknown race. This was consistent with overall patient population at these 4 institutions. African American, Hispanic, and Asian patients together were represented more frequently among those requiring critical care versus acute care but this difference was not statistically significant (74% vs 59%, P = .22). Thirty-three patients (40%) were aged <1 year, 9 (11%) were aged 1 to 5 years, 19 (23%) were aged 6 to 14 years, and 21 (26%) were aged 15 to 21 years. Distribution of presenting symptoms and need for admission to critical care or acute care units among patients of all age groups is shown in Figs 1 and 2. Fourteen patients (67%) in the 15 to 21 years age group and 9 patients (47%) in the 6 to 14 years age group had a BMI in the overweight or obese range.

Fifty percent of the patients had a known sick contact with confirmed or suspected COVID-19. Forty-five percent of all patients had at least 1 comorbidity at initial presentation, with obesity, asthma, and malignancy as the most common pre-existing conditions (Table 1). Fever (80%) was the most common presenting symptom, followed by cough (39%) and shortness of breath or dyspnea (29%) (Fig 1). Overall, older children had similar symptomatology to those described in adult patients with COVID-19.²⁰  There were 9 patients (11%) with acute appendicitis (7 were aged 6–14 years and 2 were aged 15–21 years; 3 patients had perforated appendix) and 2 children (2%) with intussusception.

Leukopenia and lymphopenia were observed in 22% and 36% of the patients, respectively. C-reactive protein (CRP) was elevated in 67% (n = 54) of patients, and procalcitonin was elevated in 34% (n = 35) of patients. Other inflammatory markers, such as interleukin-6 (IL-6), D-dimer, and lactate dehydrogenase, were elevated in two-thirds and ferritin was increased in one-third of the patients (Table 2).

Forty-eight (59%) patients had chest radiographs performed, of which 42 (88%) were abnormal. Chest radiographic findings included bilateral ground glass or hazy opacities in 20 (42%) patients, focal consolidation in 6 (13%), small airway inflammation in 15 (31%), pleural effusion in 3 (6%), and pneumothorax in 1 (1%) patient. Five children (6%) had computerized tomography of the chest, and 4 of them had bilateral ground glass or patchy airway opacities in the lungs.

Twelve patients (15%) in this cohort had coinfections. Sterile site culture results were negative, with the exception of 2 positive blood culture results (Klebsiella pneumoniae in a patient requiring mechanical ventilation and ECMO support; Klebsiella oxytoca and Streptococcus constellatus in a patient with a perforated appendix) and 2 positive urine cultures (Pseudomonas aeruginosa in patient with history of posterior urethral valve and Hutch diverticulum; and Candida dubliniensis in a patient with diabetic ketoacidosis). Three patients had viral coinfections (rhino/enterovirus, respiratory syncytial virus, and influenza B).

Hydroxychloroquine (alone or in combination with other therapies) was administered to 26% of patients, predominately early in the reporting period. Other therapies included tocilizumab (10%), remdesivir (6%), lopinavir plus ritonavir (2%), and convalescent plasma (1%). Patients with respiratory symptoms or with comorbidities were more likely to receive medications for the treatment of COVID-19. In total, 76% patients requiring respiratory support received therapies for COVID-19, compared with 6% of those who did not require any respiratory support (P < .001). Similarly, higher proportions of patients with any comorbidity received medications for COVID-19 than those who had no comorbidities (53% vs 11%, P < .001).

Twenty-nine patients (35%) required some form of respiratory support, of which 7 patients (9%) needed mechanical ventilation. The noninvasive ventilation modalities (maximum respiratory support required) included 12 patients (15%) on nasal cannula, 9 (11%) on high-flow nasal cannula, and 1 (1%) on bilevel positive airway pressure support. We performed a subanalysis and comparison between children who required any form of respiratory support and those who required no respiratory support during hospitalization. Patients who required respiratory support were older (median age: 17.1 years IQR: 8.1–18.8 years] vs 9 months [IQR: 5 months–10.2 years], P < .001) and more frequently had pre-existing comorbidities (69% vs 32%, P = .001). Children with obesity had increased need for respiratory support, but this was not statistically significant (41% vs 30%, P = .21]. Children with a history of asthma were more likely to require respiratory support during hospitalization (28% vs 8%, P = .02).

Children with any pre-existing comorbidity were more likely to require critical care than acute care during hospitalization (70% vs 37%, P = .008). Obesity is the most common risk factor associated with the need for critical care (63% vs 28%, P = .02). There was no difference in the proportions of overweight children (≥2 years) admitted to acute or critical care (25% vs 13%, P = .41). Children admitted to critical care were more likely to have sick contacts than those admitted to acute care units (78% vs 45%, P = .008). In addition, they were more likely to have elevated CRP, IL-6, and alanine aminotransferase (Table 2). Children admitted to critical care had higher rates of renal dysfunction at presentation than those admitted to acute care with elevated creatinine levels for age²¹  (43% vs 10%, P = .002) and low estimated glomerular filtration rate^22,23  (56% vs 19%, P = .01). Initial vital signs were not significantly different between 2 groups. The median length of stay was higher among patients needing critical care (8 days [IQR: 4–24] vs 2 days [IQR: 2–4], P < .001).

In this study, we describe the largest pediatric case series to date in the United States of hospitalized children with SARS-CoV-2. In contrast to previously published pediatric SARS-CoV-2 inpatient cohorts, in which the final clinical outcome of many patients (up to one-third) had not yet been determined at the time of publication, we capture clinical outcomes through hospital discharge for all patients.^12–14  We compared these patients on the basis of their level of care and age groups. Additionally, we described a more-granular comparison between children who required respiratory support to those with no respiratory support than previously described in the literature.

Categorization of the disease severity among children with SARS-CoV-2 has been variable among limited published data,^3,11–15  with the main focus being patients requiring mechanical ventilation (Supplemental Table 3). Noninvasive respiratory support poses a significant morbidity burden for both patients and health care systems. The guidance regarding performing early intubation or trying noninvasive ventilation modalities has rapidly evolved for patients with COVID-19 and worsening respiratory status. Therefore, in our study, we have looked at the data in a more comprehensive manner and compared children on the basis of their level of care to understand the burden of acute and critical care admissions in pediatric patients as well as elucidated the features of those children who required respiratory support (invasive or noninvasive). We recognize the need for standardizing descriptions of disease acuity and clinical course in the pediatric literature so we can arrive at a more meaningful understanding of the impact of SARS-CoV-2 infection in children.

Children with comorbidities were noted to be at increased risk for critical care admission and need for respiratory support in this study. Obesity is an important risk factor for SARS-CoV-2 admission in patients <60 years of age²⁴  and is associated with an increased need for mechanical ventilation in at least 1 pediatric study.¹⁴  We speculate that children with obesity may be at increased risk for more severe illness because of SARS-CoV-2, as demonstrated for other viral infections, such as influenza A,²⁵  and among children requiring critical care during hospitalization.^26–28  This may be related to an altered immune response,²⁵  reduced baseline pulmonary function, and association of progression to an inflammatory syndrome as described with other comorbid conditions in patients with obesity.^29,30  In previously published case series from the United States,^11,12  children with asthma were not at increased risk for critical care admission, but morbidities such as need for respiratory support were not described among them. In our larger cohort, asthma as a comorbidity was significantly higher among children who required some form of respiratory support when compared with those who required no respiratory support, although there was no statistically significant increase in the need for critical care admission. We suggest that pediatricians and other care providers consider all comorbidities, particularly obesity and asthma, as potential risk factors for severe illness while caring for children during this pandemic.

In the described cohort, patients aged 15 to 21 years were the most severely affected, with three-quarters of them requiring some form of respiratory support. Patients aged <1 and 1 to 5 years had relatively milder symptoms, and few of them required respiratory support. These findings are in contrast to a large retrospective pediatric study from China, in which the largest proportion of severe and critical cases were in children aged <1 and 1 to 5 years.³  One-third of those cases were not laboratory confirmed for SARS-CoV-2, and there was no standardized testing for other viral illnesses in the Chinese study. The reasons for differences observed in disease severity among various age groups, despite higher nasopharyngeal viral load observed among infants compared with older children,³¹  is yet to be determined. Some possible explanations could be from evolving immunologic response to the viral infections in early life and age-dependent differential expression of cell surface angiotensin-converting enzyme 2 receptors.^32–34 

Children <1 year of age were the most-represented age group in this study, which is similar to or slightly higher than other smaller studies from NYC.^15,31  In the beginning of pandemic spread in the United States, when data on children with COVID-19 from China suggested higher severity of infection among infants and younger children, it is possible that more infants with confirmed infection were admitted to ensure safety of these children. We did observe a decreasing trend of admitting infants from the emergency department over time during course of pandemic as our understanding of this infection in children increased.

The most-common presenting symptoms in this study were fever and/or dry cough. Several children had atypical nonrespiratory presenting symptoms, such as acute abdominal pain, new-onset seizure, and new-onset diabetes. Although some of these associations have been described in adult and pediatric literature,^14,35,36  it is unclear whether SARS-CoV-2 was the causative agent or if it was coincidentally diagnosed. Children admitted to critical care were more likely to have sick contacts than those admitted to acute care units. It is possible that families of patients admitted to critical care units may have reflected more on sick contacts or other aspects related to history (recall bias) or the clinical teams obtaining detailed history. Inflammatory markers were elevated in many of our patients, but only a few were associated with the need for critical care (CRP and IL-6). These findings suggest a similar, but likely less-severe, inflammatory process to that described in adult studies.^37,38  Renal dysfunction was observed more in patients requiring critical care, which has been previously described in pediatric patients with greater illness severity.³⁹  None of the patients required renal replacement therapy. The reasons for higher rate of renal dysfunction could be multifactorial, including hyperinflammation, cytokine storm, and complement dysregulation associated with COVID-19.^40,41  In addition, kidneys have increased expression of angiotensin-converting enzyme 2 receptor, potentially leading to viral entry and injury to the cells.⁴²  Among radiographic findings, a majority of the patients requiring respiratory support had bilateral ground glass or patchy opacities in lungs, whereas in a few cases these findings did not correlate with the level of clinical support required. Potential targeted therapies for COVID-19 have limited efficacy data among children. However, these treatments were given on the basis of individual patient clinical characteristics and in an effort to treat the most severely affected children.

Thirteen percent of patients in our described cohort were diagnosed with acute appendicitis or intussusception on presentation. This is a higher rate than previously published case series from NYC area.¹⁴  All of these patients had gastrointestinal symptoms (abdominal pain, vomiting, or diarrhea), three-quarters of them had fever, and one-quarter had respiratory symptoms. It is possible that children taken to operating theater were more likely to get tested as per local infection prevention measures and were coincidentally found positive for SARS-CoV-2 during high prevalence of infection in the region. Viral infections are known to cause lymphadenopathy and hyperplasia of other lymphatic tissues, such as Peyer’s patches in the intestines,^43–45  which may serve as a lead point resulting in intussusception,^46,47  and inflammation of lymphoid tissue around the appendix, resulting in acute appendicitis. This association of acute appendicitis and intussusception with SARS-CoV-2 infection needs further exploration, including analysis of pathologic changes in the surgically removed appendix.

Our study has several important limitations, including a small sample size, limited geographical range, and its observational nature. Because of rapidly evolving testing and management guidelines for suspected cases with SARS-CoV-2 at all study locations, there may have been some initial inconsistency in diagnosis and management over this study period. Also, some of the associations and risk factors observed in our relatively small study cohort need to be explored in detail on a larger data set. Nonetheless, our focus on hospitalized pediatric patients with SARS-CoV-2 in 4 NYC Metropolitan area hospitals with widespread community transmission provides important data for pediatric clinicians.

This described cohort is the largest case series to date of hospitalized children with confirmed SARS-CoV-2 managed from admission to discharge. We observed that children with comorbidities, particularly obesity and asthma, are at increased risk for critical care admission and/or the need for respiratory support. Children with renal dysfunction at presentation may be at greater risk for the need for critical care. Understanding risk factors associated with severity of illness among children requires further investigation on a larger data set.

We acknowledge the extraordinary work of all staff members who took care of these patients in the middle of this pandemic. We thank the patients and families of the children included in this study.