We enrolled 597 infants and young children with acute COVID-19 with verifiable documentation of vaccination status, of whom 371 (62.1%) were 8 months to <2 years of age and 226 (37.9%) were 2 to <5 years of age (Fig. 1 and Table 1). Table 1 shows the demographic and clinical characteristics of these children by age group. The largest percentage of children were enrolled in the South and the lowest number in the Northeast, compared with the West and Midwest census regions. The proportion of patients with at least one underlying condition differed by age group: among children <2 years of age, 44.7% had at least one underlying medical condition, compared with 69.9% of those 2 to <5 years old (P < 0.001). Among these children, the most common underlying conditions were respiratory or neurologic/neuromuscular. Over 80% of children (n = 499/597, 83.6%) were tested for at least 1 non-SARS-CoV-2 respiratory virus, all of whom were tested for influenza, and 490 (82.1%) were tested for respiratory syncytial virus (RSV) (Table 1). Among the 499 children with any other viral testing, 188 (37.7%) tested positive for any virus. Co-detection of influenza virus was infrequent (1.2%). RSV was co-detected in 11.4% of children <5 years old tested during the study period.

Clinical outcomes by age group are shown in Table 2. Median length of hospital stay was 2 days (IQR: 1–4 days) among children 8 months to <2 years and 3 days (IQR: 2–5 days) in children 2 to <5 years. Over one-third (n = 228, 38.2%) of children received respiratory support above low-flow oxygen, with high-flow nasal cannula (21.8%) being the most common, followed by IMV (8.5%) and bilevel positive airway pressure or continuous positive airway pressure (7.9%). There were 174 (29.1%) patients admitted to the intensive care unit (ICU); 75 (12.6%) had a life-threatening illness. Three patients with no underlying medical conditions required ECMO and one of these patients died. Except for longer length of stay, outcomes were similar among children 8 months to <2 years and 2 to <5 years of age.

Figure 2 shows the demographic and clinical characteristics of children with acute COVID-19 categorized by the highest levels of respiratory support (none or low-flow oxygen only to IMV) by percentage of patients. Of 51 children requiring IMV, approximately half (25, 49.0%) were non-Hispanic White and 33 (64.7%) had underlying medical conditions (Figure 2, Supplemental Digital Content 2, http://links.lww.com/INF/F364), most commonly cardiorespiratory or neurologic/ neuromuscular disease. Of those children with no underlying medical conditions, 63.7% required no respiratory support or low-flow oxygen supplementation only; however, this declined to 48.4% and 53.3% of patients with underlying respiratory and neurologic conditions, respectively (Fig. 2). Nearly 60% of children requiring IMV who had any other viral testing had only SARS-CoV-2 detected, with no detection of a second virus (Table, Supplemental Digital Content 2, http://links.lww.com/INF/F364). Among 86 patients who required noninvasive or IMV and received additional testing, 40 (46.5%) had no viral co-detections, 46 (53.5%) had at least 1 viral co-detection and 6 (7.0%) had multiple viral co-detections (Table, Supplemental Digital Content 2, http://links.lww.com/INF/F364). Among the 3 children requiring ECMO, 2 had coinfection with human metapneumovirus, including the 1 patient who died (Table, Supplemental Digital Content 3, http://links.lww.com/INF/F365).

Most children (n = 528, 88.4%) hospitalized with acute COVID-19 were not vaccinated against COVID-19 despite being eligible for vaccination. Only 27 (4.5%) patients had received their primary series [3 doses of Pfizer-BioNTech (n = 8) or 2 doses of Moderna (n = 19)]. White non-Hispanic patients comprised most patients who completed the primary series (n = 20, 74.1%). Completion of a primary mRNA vaccination series was low across all regions but was highest in the Northeast (11/90, 12.2%) and lowest in the South (3/205, 1.5%). Most children with critical illness were unvaccinated or had not completed their primary series, including 164/174 (94.3%) children with COVID-19 who were admitted to the ICU, 45/51 (88.2%) who received IMV, and 69/75 (92.0%) with life-threatening COVID-19. This included all 3 patients who required ECMO (1 died) and all 29 patients who required vasopressors (Table 3). All 6 patients who completed their primary series and were hospitalized with a life-threatening illness had at least 1 underlying respiratory or neurologic condition, including 4 patients who had both (Table, Supplemental Digital Content 4, http://links.lww.com/INF/F366).

Some patients initiated but did not complete the COVID-19 primary vaccination series before hospitalization (n = 42, 7.0%). Of these, 33 were overdue for a vaccine dose at the time of hospitalization while 9 were hospitalized between doses but were not overdue for their next dose (Figure 3). Overdue doses were common among those who initiated the Pfizer-BioNTech primary series [n = 27 (81.8%)] compared with those who initiated the Moderna primary series [n = 6 (18.2%)]. Likewise, among the 9 nonoverdue patients who were hospitalized in between primary series doses, 8 initiated the Pfizer-BioNTech primary series.

Across 28 centers in 21 states from September 20, 2022 to May 31, 2023, we identified 597 children between 8 months and <5 years of age hospitalized for COVID-19. Nearly half were previously healthy with no underlying conditions. Over a quarter of these hospitalized children required ICU admission, and life-threatening COVID-19 occurred in over 1 in 10. All 3 of the children who required ECMO for acute COVID-19 were previously healthy, and none of these children nor the 29 children requiring vasopressors were vaccinated, and 1 died. Vaccine coverage in this cohort was very low despite being eligible, including among those developing severe and life-threatening disease. Furthermore, over 60% of those hospitalized for acute COVID-19 who had initiated vaccination had not completed their primary series prior to hospitalization. Given the low overall vaccination coverage among these patients, it is imperative to understand the underlying reasons for such limited vaccine uptake in this age group.

Vaccine efficacy has primarily been ascertained through immunogenicity studies that were not designed to assess protection against severe clinical outcomes, which occurred rarely among these trial cohorts. Our platform is designed specifically to enroll hospitalized patients; however, fewer than 5% of children hospitalized for acute COVID-19 and approximately 7% of SARS-CoV-2 negative patients enrolled through this same network had completed the primary COVID-19 vaccination series before admission, precluding real-world evaluation of vaccine effectiveness against severe COVID-19 illness. This low coverage stands in contrast to influenza vaccine uptake in this age group. In previous reports of influenza vaccine effectiveness among children <5 years of age hospitalized for acute respiratory infection within the New Vaccine Surveillance Network and the Pediatric Intensive Care Influenza Network, between 30% and 40% of children <9 years of age hospitalized for influenza and 50%–70% of hospitalized influenza negative controls had received a seasonal influenza vaccine.^16–18 Notably, over the first 4 months following the rollout of the Pfizer-BioNTech and Moderna monovalent primary series in this age group, US real-world vaccine effectiveness against outpatient symptomatic SARS-CoV-2 infection among children 3–5 years old was demonstrated to be 31% and 36% effective, respectively.⁸ In Singapore, vaccine effectiveness against outpatient symptomatic SARS-CoV-2 infection among infection-naïve children 1–4 years of age was 63%.⁹ Vaccine effectiveness should be continuously monitored, given the recent incorporation of doses targeting the Omicron subvariant XBB.1.5, into all primary mRNA vaccination series.

This investigation highlights the need to improve COVID-19 vaccination coverage and awareness of severe COVID-19 illness in young children. Despite an observed decrease in the proportion of hospitalized children <5 years of age requiring ICU admission during the period of Omicron predominance, similar proportions are still requiring IMV.¹⁹ In this analysis, those children with underlying respiratory or neurologic conditions developed more severe COVID-19, requiring higher levels of respiratory support. Although children with these conditions are known to be at higher risk for COVID-19-related complications,²⁰ only a few were vaccinated. Among the small percentage of children who were fully vaccinated, 70% were non-Hispanic White, and the highest proportion of children completing their primary series reside in the Northeast census region. Disparities in vaccination coverage by race and ethnicity and geographic region comport with previous findings in this age group, likewise finding disproportionately higher vaccine coverage among non-Hispanic White children and those residing in the Northeast.²¹

Low vaccine coverage in this age group may indicate logistical barriers to provision of COVID-19 vaccine to very young children, which include (1) challenges in vaccine distribution (where doses may be available only in pediatric hospitals or large pediatric practices), (2) reliance on pediatricians and hospitals over pharmacies to administer doses to very young or medically complex children and (3) the relative complexity of the Pfizer-BioNTech primary series (three doses, with different inter-dose intervals), which may reduce parental compliance with the primary series schedule. Reflecting this, we noted that a higher proportion of patients initiated but did not complete the Pfizer-BioNTech primary series compared with those who initiated the Moderna primary series. While nearly 70% of those who completed the primary series initiated the Moderna series, over 80% of children who did not complete their primary series initiated the Pfizer-BioNTech primary series. The longer timeframe required to complete the 3-dose Pfizer-BioNTech primary series may also allow for more opportunities for infection during inter-dose intervals. In addition, caregiver perceptions of COVID-19 severity and vaccine safety may predict COVID-19 vaccination intention for children <5 years. In the early months of eligibility in this age group, there was substantial evidence of parental hesitancy and low vaccine confidence.²² Despite this, pre- and early post-marketing monitoring for severe adverse events indicated that COVID-19 vaccination was overwhelmingly safe and effective among children 6 months to <5 years of age, with adverse events rarely reported.^11,12,23,24 Of note, 73% of US children have completed their routine pediatric vaccines, all consisting of multidose series, and 17% have initiated but not completed one or more of their routine series. While these figures highlight a need for promotion and programmatic optimization to improve routine pediatric vaccine uptake, COVID-19 vaccination coverage in young children remains far below coverage of routine pediatric vaccination.²⁵ Expanded efforts to address gaps in vaccine access and additional caregiver outreach may help promote adherence to COVID-19 vaccination schedules.

Strengths of this investigation include representative enrollment of hospitalized COVID-19 patients by site, wide geographic breadth of included sites covering 28 pediatric hospitals in 21 states, rigorous vaccine verification procedures, and timing of patient enrollment that aligned with vaccine eligibility in this age group. This investigation, however, was subject to several limitations. First, even with simultaneous enrollment of a separate population of hospitalized SARS-CoV-2-negative patients, vaccine coverage in this population was too low to evaluate vaccine effectiveness. Second, we did not universally test all patients for other respiratory viruses; rather, clinical data on viral co-detections were obtained solely through medical records abstraction. Influenza affected relatively few patients, but it is possible that RSV, human metapneumovirus or other respiratory viral co-detections influenced disease severity. Third, we could not reliably differentiate between patients who received low-flow oxygen supplementation compared with those who received no oxygen support, as low-flow oxygen supplementation is frequently transient and challenging to ascertain through retrospective medical record abstraction. Finally, patients were enrolled largely from tertiary referral pediatric centers, so while participating sites represented geographic breadth, the findings from this investigation may not be generalizable to all patients 8 months to 5 years hospitalized for severe COVID-19 illness in the community.

Despite eligibility, most children hospitalized for COVID-19, including most children with underlying medical conditions, had not initiated their COVID-19 mRNA primary vaccination series before their hospitalization. Of the children who had initiated vaccination, particularly those initiating the Pfizer-BioNTech primary series, most had not completed their primary series by the time they were hospitalized. Our investigation shows that some young children, including those who are otherwise previously healthy, are at risk of developing severe COVID-19. Strategies to reduce barriers to vaccine access among young children, as well as caregiver education, are vital components of efforts to expand coverage and promote completion of all recommended COVID-19 mRNA vaccine doses.