SARS-CoV-2 Among Infants <90 Days of Age Admitted for Serious Bacterial Infection Evaluation Free

Electronic medical records (EMRs) of infants <90 days of age admitted to any of 3 New York University Langone Health System inpatient facilities in Brooklyn, Long Island, and Manhattan, or NYC Health + Hospitals (NYCHHC) Bellevue Hospital between March 15, 2020, and December 15, 2020, were screened for eligibility. Records from the NYCHHC Bellevue Hospital were only available for review through October 1, 2020. The start date reflects when routine in-house SARS-CoV-2 nasopharyngeal reverse transcription polymerase chain reaction (RT-PCR) testing became available at 3 of the 4 study facilities. Because of the limited availability of SARS-CoV-2 testing supplies early in the NYC COVID-19 epidemic, SARS-CoV-2 RT-PCR testing was limited to hospitalized patients at the time of study initiation; therefore, infants undergoing SBI evaluation in the emergency department who were not hospitalized were not included.

Inclusion criteria included age <90 days at the time of hospitalization and admission for SBI evaluation. Admission for SBI evaluation was determined by investigator review of the emergency department and admitting history as to whether evaluation and treatment of bacterial infection was the main reason for admission. Febrile and nonfebrile infants, including newborns undergoing SBI evaluation because of exposure to maternal infection, were included in the overall study analysis, and separate subgroup analysis was performed on febrile infants. Infants without SARS-CoV-2 testing during admission were excluded. Abstracted data from the EMR were entered into a Health Insurance Portability and Accountability Act of 1996–compliant database (Research Electronic Data Capture), including demographics (age, sex, ethnicity, and race), medical comorbidities, mode of birth delivery, history of exposure to ill contacts and confirmed COVID-19 cases, clinical symptoms at presentation (documented as negative if absent or not noted on admission notes), initial laboratory (white blood cell [WBC] count, absolute neutrophil count [ANC], absolute lymphocytic count [ALC], hemoglobin, platelets, C-reactive protein [CRP]) and chest radiograph findings (the presence of infiltrates or consolidations per radiology report, further categorized as focal or bilateral), microbiologic culture and/or BioFire FilmArray Respiratory Panel (BioFire Diagnostics, Salt Lake City, UT) results, respiratory and nonrespiratory interventions required, ICU admission status, clinical outcome, and hospital length of stay (LOS).

Categorical variables were analyzed via Pearson’s χ² test. Continuous variables with normalized distributions (skewness: ≤1) are reported as a mean with the SD and were analyzed by using Student t tests with Welch’s correction, and those with nonnormalized data distributions (skewness: >1) are reported as the median with the interquartile range (IQR) and were analyzed by using Mann–Whitney U tests. Prism 8.4 (GraphPad, San Diego, CA) was used for statistical analysis, and P values < 0.05 were considered statistically significant. Classification of high and low community circulation periods were determined by examination of daily COVID-19 testing positivity data provided by the NYC Department of Health and Mental Hygiene, with high community circulation defined by 7-day rolling average testing positivity rates ≥5%.²⁵ 

This study was approved by the New York University Grossman School of Medicine Institutional Review Board and NYCHHC System to Track and Approve Research with a waiver of consent and authorization for data collection.

Medical records of 1119 infants <90 days of age admitted to one of the study facilities during the study period were reviewed. Of these, 148 infants (13%) met the eligibility criteria. The majority were male (n = 86; 58%), and the most frequently identified race and ethnicity were white (n = 71; 48%) and non-Hispanic (n = 74; 50%; Table 1). No differences in sex, racial, or ethnic distribution were observed between febrile and nonfebrile infants (Table 1). The mean age at admission was 25 (SD: 22) days (Fig 1) and was lower among nonfebrile infants (8 days [SD: 14]), compared with febrile infants (34 days [SD: 20]; P < .001). Most infants (n = 135; 91%) had no pre-existing medical conditions. A total of 22 of the 148 infants (15%) admitted for SBI evaluation were positive for SARS-CoV-2 by using a nasopharyngeal RT-PCR.

By using the ≥5% 7-day rolling average COVID-19 testing positivity threshold, March 15 to May 28, 2020, and November 28 to December 15, 2020, were defined as periods of high community circulation; the remainder of the study period was considered to have low community circulation. During high community circulation periods, 19 of 62 infants (31%) tested positive for SARS-CoV-2, whereas only 3 of 86 infants (3%) tested positive while community circulation was low (P < .001).

Among all infants admitted for SBI evaluation, the mean age of infants with a positive SARS-CoV-2 RT-PCR was higher than that of infants who tested negative (33 [SD: 17] days, to 23 [SD: 23] days; P = .03), although no age difference was observed between SARS-CoV-2–positive and SARS-CoV-2–negative infants in the febrile subgroup (33 days [SD: 14] to 35 days [SD: 22]; P = .63). Comorbidities were more common among infants negative for SARS-CoV-2 compared with those who were positive among all infants (n = 13 [10%] to n = 0 [0%]; P < .001) and the febrile subgroup (n = 7 [9%] to n = 0 [0%]; P = .007). Infants with SARS-CoV-2 were more likely to have any ill contact documented at admission, compared with those who tested negative, both among all infants (n = 9 [41%] to n = 9 [7%]; P = .006) and the febrile subgroup (n = 9 [47%] to n = 6 [9%]; P = .006). No differences were observed in admission documentation of confirmed COVID-19 contacts between all infants (P = .75) or febrile infants (P = .48) regardless of SARS-CoV-2 status (P = .75) or community SARS-CoV-2 circulation level (P = .94; Table 1).

Nearly all infants with SARS-CoV-2 presented with a fever (n = 20; 91%), whereas less than two-thirds of infants without SARS-CoV-2 had a fever before admission (n = 74 [59%]; P < .001; Table 2). An isolated fever was the most common presentation of SARS-CoV-2 (n = 13; 59%), and an isolated fever was more frequent among infants with SARS-CoV-2 compared with those who tested negative (n = 17; 13%; P < .001), including in the febrile subgroup (P = .002; Table 2). Hypothermia (n = 0 [0%] to n = 8 [6%]; P = .004), irritability and/or fussiness (n = 2 [9%] to n = 37 [29%]; P = .01), diarrhea (n = 0 [0%] to n = 7 [6%]; P = .008), and lethargy (n = 0 [0%] to n = 9 [7%]; P = .002) were less frequently documented among all SBI evaluation infants who tested positive for SARS-CoV-2 than those with negative test results. Irritability and other symptoms were less frequently documented among febrile infants with SARS-CoV-2 compared with those without SARS-CoV-2 (n = 2 [10%] to n = 35 [47%]; P < .001 and n = 0 [0%] to n = 7 [9%]; P = .007, respectively), but no differences in the frequency of other assessed symptoms was observed in the febrile subgroup (Table 2).

Infants with SARS-CoV-2 had significantly lower mean WBC counts (7.3 [IQR: 5.5–11.3] x 10³ cells per µL to 13.0 [IQR: 8.8–17.7] x 10³ cells per µL; P < .001), lower median ANCs (1.7 [IQR: 1.4–3.1] x 10³ cells per µL to 5.8 [IQR: 2.5–9.9] x 10³ cells per µL; P < .001), lower median ALCs (3.3 [IQR: 2.7–4.7] x 10³ cells per µL to 4.6 [IQR: 3.0–6.6] x 10³ cells per µL; P = .04), and lower mean hemoglobin levels (12.8 [SD: 2.1] g/dL to 14 [SD: 3.7] g/dL; P = .04), compared with those without SARS-CoV-2 (Table 3). Lower WBC, ANC, and ALC counts were also observed in the febrile subgroup with SARS-CoV-2, although no differences in hemoglobin were observed (Table 3). Additionally, infants with SARS-CoV-2 infection had lower median CRP values than those without, both among all infants (0.8 [IQR: 0.4–2.5] mg/L to 5.7 [IQR: 1.5–49] mg/L; P = .005) and the febrile subgroup (0.8 [IQR: 0.3–2.4] mg/L to 13.6 [IQR: 1.5–79.2]; P < .001).

Microbiologic samples obtained included blood cultures (n = 146; 98%), urine cultures (n = 102, 69%), BioFire FilmArray respiratory panel testing (n = 97; 66%), and cerebrospinal fluid (CSF) cultures (n = 92; 62%). Twenty-two blood cultures were positive (15%), 9 of them (41%) were deemed by clinicians to represent true infections, whereas the remainder (n = 13; 59%) were deemed to represent nonclinically significant contaminants. No difference was observed in overall blood culture positivity (n = 3 [14%] to n = 19 [15%]; P = .83) or in the proportion of true infections compared with contaminants (n = 1 [33%] to n = 8 [42%]; P = .65) between infants with and without SARS-CoV-2 infection, including the febrile subgroup (n = 3 [15%] to n = 18 [24%]; P = .16; and n = 3 [15%] to n = 7 [9%]; P = .54). Infants without SARS-CoV-2 were more likely than those with SARS-CoV-2 to have positive urine culture results (n = 25 [20%] to n = 1 [5%]; P =.002) and positive CSF culture results (n = 5 [4%] to n = 0 [0%]; P = .02). Among febrile infants, positive urine culture results were more likely among SARS-CoV-2–negative compared with positive infants (n = 22 [33%] to n = 1 [6%] P = .03). No positive CSF cultures occurred among febrile infants. Detection of other viruses on the respiratory viral panel (RVP) was low (n = 9; 6%) and no differences in RVP positivity rates were observed between those with and without SARS-CoV-2 in all infants (n = 2 [9%] to n = 7 [6%]; P = .85), febrile infants (n = 1 [6%] to n = 6 [10%]; P = .69), or between high or low community SARS-CoV-2 circulation periods (n = 2 [3%] to n = 7 [8%]; P = .16). Chest radiography was performed in 56 infants (38%); no differences in the presence or distribution of infiltrates were present between all infants with and without SARS-CoV-2 or in the febrile subgroup (Table 3).

A minority of infants (n = 32; 22%) required respiratory support during hospitalization (Table 4). The need for any respiratory support did not differ for infants with and without SARS-CoV-2 (n = 2 [9%] to n = 20 [24%]; P = .16). Infants with SARS-CoV-2 were significantly less likely to require an ICU admission (n = 2 [9%] to n = 47 [37%]; P < .001) but had a similar hospital LOS to infants without SARS-CoV-2 (2 [IQR: 2–3] days to 3 [IQR: 2–7] days; P = .08). No differences in respiratory interventions, ICU admissions, or LOS were observed in the febrile subgroup (Table 4). All infants with SARS-CoV-2 survived to discharge.

In this study, 15% of the 148 infants admitted for an SBI evaluation during the 9-month study period had positive SARS-CoV-2 nasopharyngeal RT-PCR tests. We found SARS-CoV-2 positivity rates among infants hospitalized for an SBI evaluation significantly differed between periods of low (3%) and high (31%) community SARS-CoV-2 circulation. Our finding of 31% SARS-CoV-2 positivity during high community circulation periods is consistent with a 30% SARS-CoV-2 prevalence reported in a case series of febrile infants <60 days of age evaluated in a different facility in NYC during the peak of the epidemic.²⁶ 

In adults and older children, COVID-19 can present with a wide variety of symptoms, including systemic, dermatologic, respiratory, and gastrointestinal symptoms.^{1,7,10–12,16,18,20,27}  However in this study, isolated fever was the most common presentation of SARS-CoV-2, with a higher percentage of infants in this study presenting with an isolated fever compared with the 40% reported in another study of infants.²⁸  A lower proportion of infants with SARS-CoV-2 in this study had dermatologic, respiratory, or gastrointestinal symptoms noted at admission, compared with other pediatric and young infant studies.^{5,9–14,20,26,27,29–32} 

The mean age of all infants with SARS-CoV-2 infection in our study was higher than that of SARS-CoV-2–negative infants but in line with mean or median ages (16 to 39 days) reported in other studies of infants undergoing SBI evaluation.^{9,12,26,28,30}  Although this could reflect longer exposure to the extrauterine environment and more opportunities for viral transmission, inclusion of asymptomatic newborns exposed to maternal infections may have decreased the average age of SARS-CoV-2–negative infants, which may explain why this age difference was not observed in the febrile subgroup. The lower age among infants without SARS-CoV-2 infection may confound the differences in hemoglobin observed between all infants with and without SARS-CoV-2, because hemoglobin may be elevated shortly after birth, and no hemoglobin differences were observed between SARS-CoV-2–positive and SARS-CoV-2–negative infants in the older, febrile subgroup. Lower WBC, ANC, and ALC values were observed among SARS-CoV-2 infants in all analysis groups, consistent with what has been reported in some pediatric COVID-19 studies.^13,33–37  No differences in chest imaging results were observed among infants with and without SARS- CoV-2, consistent with findings from older children described in the literature.²⁹  We recommend clinicians avoid using laboratory or imaging results to guide decision-making on whether to test hospitalized infants for SARS-CoV-2.

Case reports and series have identified concomitant urinary tract infections and bacteremia among young infants with COVID-19.^9,13,26–28  We found lower rates of positive urine and CSF cultures in infants with SARS-CoV-2 compared to those without SARS-CoV-2, but no difference in blood culture positivity rates. Given the potentially severe consequences of untreated bacterial infections, we recommend clinicians continue to assess young febrile infants for bacterial infections, regardless of SARS-CoV-2 status. We found a relatively low incidence (6%) of viral co-infection among infants with and without SARS-CoV-2. This likely reflects community-wide decreases in other respiratory viruses reported in New York during the study period^38,39  because of enhanced infection control practices during the COVID-19 pandemic. Additional studies are warranted to determine the prevalence of viral coinfections when respiratory virus circulation resembles more typical, seasonal epidemiological patterns.

Overall illness among infants with SARS-CoV-2 infection in this study was mild to moderate, consistent with the severity of disease in this age group previously reported.^26,28  A minority of infants with SARS-CoV-2 required respiratory support or ICU admission, consistent with the reported literature,^{8,9,12,13,19,28–30,32}  although this may be confounded by our observed higher prevalence of comorbidities among infants without SARS-CoV-2 compared with infants with SARS-CoV-2. The median hospital LOS for SARS-CoV-2–positive infants in this study was shorter than previously reported,^10,30  and all infants with SARS-CoV-2 survived to discharge, consistent with good outcomes reported across other studies.^{10–12,17,26} 

Our study has several limitations. The number of infants with SARS-CoV-2 in our sample is small, and most participants were not identified in the EMR as belonging to a racial or ethnic minority group. Therefore, our results may not accurately reflect the prevalence and/or severity of infection among infants of racial or ethnic backgrounds that have been associated with more severe COVID-19. Because of limited SARS-CoV-2 testing capacity at the beginning of the NYC epidemic, testing was only available for hospitalized patients at the start of this study; therefore, the study population was limited to hospitalized infants and is not representative of all SARS-CoV-2 infections among infants <90 days of age undergoing SBI evaluation in an emergency care setting. We anticipate that the exclusion of infants evaluated for an SBI in the emergency department but subsequently discharged may bias our results in favor of reflecting illness among younger infants and capture more moderate to severe cases, potentially underrepresenting infants 60-89 days of age and those with mild infections. However, despite a potential bias toward enrolling sicker infants, we found a low incidence of ICU admission and need for respiratory support among hospitalized infants with SARS-CoV-2.

We used 7-day rolling average testing positivity rates, with a cutoff threshold of 5% as a surrogate marker for levels of community SARS-CoV-2 circulation, although use of this measure has limitations. Changes in testing capacity throughout the NYC COVID-19 epidemic, from initially being limited to hospitalized patients to widespread outpatient testing, makes this an imprecise surrogate marker of community circulation because of differences in the population receiving testing over time. However, these testing limitations also affect other markers of community SARS-CoV-2 circulation, such as the total number of confirmed cases or case rates in the NYC population. We opted to use testing positivity because of the availability of data from the NYC Department of Health and Mental Hygiene during the entire study period without any data gaps. The cutoff threshold of 5% testing positivity was chosen on the basis of agreement between the timing of this testing positivity value with significant inclines or declines on the NYC COVID-19 epidemic curve.

Enhancing our knowledge of how SARS-CoV-2 infection affects young infants is important for informing clinical practice, planning public health measures (such as COVID-19 vaccine distribution), and furthering our understanding of COVID-19 in this age group. The prevalence of SARS-CoV-2 infection among hospitalized young infants in 2020 varied with levels of community transmission. Although the transmissibility of SARS-CoV-2 from infants at this age is unknown, identification of infants with SARS-CoV-2 infection has important implications for hospital and local community infection control. Additional research is needed on young infants evaluated in a wider range of settings, including the emergency department and outpatient clinics, to truly understand the full impact of SARS-CoV-2 infection in this age group. During the 9-month study period, we witnessed rapidly evolving changes in SARS-CoV-2 epidemiology and variations in testing capacity from limited inpatient testing only to broad-scale outpatient availability. As the epidemiology of COVID-19 continues to evolve with the emergence of variant viruses and vaccination implementation, continued monitoring of infection in this age group is warranted to determine if, when, and in what settings a selective testing strategy, as opposed to universal testing, may be appropriate in the future.

ALC

absolute lymphocytic count

ANC

absolute neutrophil count

COVID-19

coronavirus disease 2019

CRP

C-reactive protein

CSF

cerebrospinal fluid

EMR

electronic medical record

IQR

interquartile range

LOS

length of stay

NYC

New York City

NYCHHC

New York City Health + Hospitals

RT-PCR

reverse transcription polymerase chain reaction

SARS-CoV-2

severe acute respiratory syndrome coronavirus 2

SBI

serious bacterial infection

WBC

white blood cell count