If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Division of Pediatric Infectious Disease, Department of Pediatrics, University of Alabama at Birmingham, AL, USADepartment of Paediatrics, King Abdulaziz University, Jeddah, Saudi Arabia
Human coronaviruses (HCoVs) are important respiratory pathogens in humans and animals. Most HCoVs are emerging pathogens, with five known human pathogens identified in the last two decades.
Aim
To examine the clinical course of HCoV infection in children to improve understanding of severity and outcomes.
Methods
A retrospective review was undertaken of all encounters of children with known HCoV infection at a tertiary paediatric hospital from January 2015 to January 2018. Electronic medical records were reviewed for demographic data, HCoV type, viral co-pathogens, time to testing, need for hospitalization, requirement for higher-level care (HLC) including intensive care unit management and requirement for oxygen support, radiographic findings suggestive of lower respiratory tract (LRT) disease, and length of stay (LOS).
Findings
In total, 450 encounters for 430 different patients were identified, with the majority (85%) being inpatient. OC43 was the most common HCoV. Younger patients (age <5 years) had higher probability of hospitalization [adjusted odds ratio (aOR) 2.2, 95% confidence interval (CI) 1.2–4.1], requirement for HLC (aOR 1.8, 95% CI 1.0–3.1) and presence of LRT findings on chest radiographs (aOR 1.7, 95% CI 1.01–2.9). Clinical outcomes did not differ between HCoV types, except LOS which was longer for 229E. Fifty-two (11%) encounters were detected after 3 days of hospitalization (median 25.5 days), suggesting possible nosocomial infection.
Conclusion
HCoVs are important respiratory pathogens in the paediatric population, especially among patients aged <5 years who are at increased risk for severe disease. The role of HCoVs as hospital-acquired pathogens may be underappreciated.
]. Most are considered as emerging human pathogens, as their origin has been linked to cross-species transmission associated with intermediate animal hosts. HCoVs are divided into four groups – alpha, beta, gamma and delta – with human pathogens belonging to the first two groups. Alpha HCoVs include 229E and NL63. Beta coronaviruses include severe acute respiratory syndrome coronavirus (SARS-CoV), SARS-CoV-2, Middle East respiratory syndrome coronavirus, OC43 and HKU1 [
]. This study was designed and performed prior to the emergence of SARS-CoV-2, the cause of the coronavirus disease 2019 (COVID-19) pandemic – a global health threat of unprecedented severity [
Disease severity attributed to different HCoV types can vary. Endemic HCoVs (OC43, NL63, 229E and HKU1) primarily cause upper respiratory tract infections, particularly among children [
Community surveillance of respiratory viruses among families in the Utah Better Identification of Germs-Longitudinal Viral Epidemiology (BIG-LoVE) study.
Epidemiology and clinical presentations of the four human coronaviruses 229E, HKU1, NL63, and OC43 detected over 3 years using a novel multiplex real-time PCR method.
The COVID-19 pandemic highlights the importance of understanding the disease pathogenesis related to HCoV infection. The purpose of this study was to examine the clinical course of HCoV infection in children to improve understanding of disease severity and outcomes.
Methods
A single-centre retrospective chart review of a cohort involving all children with a positive test for HCoV who received care at Children's of Alabama, a large tertiary paediatric hospital in Birmingham, Alabama, USA, was undertaken. This work did not involve human participants and was approved by the Institutional Review Board of the University of Alabama at Birmingham.
Patients
All patients with a positive viral respiratory panel (VRP) on polymerase chain reaction (PCR) for HCoVs were included. PCR was performed using GenMark e-Sensor [
]. Only the first encounter was analysed if the same patient presented on more than one occasion for the same HCoV, since prolonged shedding or re-infection could not be ruled out. Patients with repeated positive PCR tests (more than one encounter for the same HCoV during the study period) and patients with mixed infections (more than one HCoV simultaneously) are described separately.
Data collection
Data collected from review of the subjects' electronic medical records (EMRs) included demographics (date of birth, gender and race/ethnicity), admission status (inpatient or outpatient), chest x-ray (CXR) findings (when obtained for clinical indication), need for higher-level care (HLC) [defined as admission to intensive care (paediatric, neonatal or cardiac) or special care units], HCoV type, viral co-pathogens (influenza virus, parainfluenza virus, respiratory syncytial virus, adenovirus, human metapneumovirus and rhinovirus/enterovirus), time to testing (from admission date to VRP result date), requirement for oxygen support and length of stay (LOS). Underlying medical comorbidities were also recorded, including prematurity (<37 weeks of gestation), immunosuppression (cancer, primary immunodeficiency or use of chemotherapy, immunomodulators such as steroids or biologic response modifiers), underlying respiratory disease (chronic lung disease, cystic fibrosis or asthma), cardiac disease, requirement for oxygen support at baseline, presence of tracheostomy tube, neuromuscular disease (cerebral palsy, seizure disorder or muscular atrophy) and other complex medical problems (chronic diagnoses not included above).
Definitions
Age was calculated in days from date of birth to encounter date, and dichotomized for those aged <5 years or ≥5 years. Admissions for observation alone were counted as ‘outpatient’ unless LOS exceeded 24 h. CXR findings were recorded for LRT disease (defined as peribronchial thickening, consolidations, patchy infiltrates or worsening of chronic parenchymal disease). Requirement for oxygen support was recorded as positive if the patient developed a new oxygen requirement or required an increase in oxygen support relative to baseline. Encounters with the identification of a new respiratory pathogen after 3 days of hospitalization with onset of new symptoms were labelled as possibly nosocomial. Encounters were labelled as recurrent if the same patient had a different HCoV detected at separate evaluations.
Statistical analyses
Descriptive statistics used proportions and percentages for categorical variables, and medians and interquartile ranges (IQR) for continuous variables. Bivariate analysis was conducted using Chi-squared test or Fisher's exact test, as appropriate. Significance was determined as P<0.05. LOS was rounded to the nearest integer and analysed using non-parametric tests for bivariate and Poisson distribution in the multi-variate model. Multi-variate analysis was undertaken using generalized linear mixed models with participants as random effect. Adjusted odds ratios (aOR) with 95% confidence intervals (CI), and P values are presented. Analyses were performed using SAS Version 9.3 (SAS Institute, Cary, NC, USA). Patients with mixed HCoVs (more than one HCoV simultaneously, N=7) were excluded from bivariate and multi-variate analyses due to the small number of observations, and are therefore described separately.
Results
Population
The dataset included 430 different patients with 455 encounters (407 patients had one encounter, 21 patients had two encounters and two patients had three encounters): 389 (85%) inpatient encounters and 66 (15%) outpatient encounters (Figure 1). Table I provides the demographic characteristics and spectrum of clinical disease. Viral infections were more common during the winter and spring (Figure 2). Viral co-pathogens were present in 234 children (51%), with rhinovirus/enterovirus being the most common (Table I).
Figure 1Flow chart showing the original and final datasets and participants. HCoVs, human coronaviruses.
Figure 2Distribution of encounters with positive total human coronaviruses over the study period. Yearly positive encounters were as follows: 101 encounters in 2015, 134 encounters in 2016, 199 encounters in 2017, and 21 encounters in January 2018.
The distribution of HCoVs and the presence of viral co-pathogens did not differ by age, gender or race/ethnicity. The detection of viral co-pathogens was associated with age <5 years (57% vs 39%; P=0.0006), immunocompetence (54% vs 38%; P=0.013) and absence of neuromuscular disorders (54% vs 42%; P=0.038). HCoVs and the presence of any viral co-pathogens did not differ by prematurity, cardiac disease, other complex medial problems, requirement for oxygen support at baseline, and presence of a tracheostomy tube.
Clinical outcome
Inpatient encounters were more common in patients aged <5 years, patients with a history of prematurity, patients with a tracheostomy tube at baseline, and immunocompetent patients (Table II). The other outcomes did not differ between inpatient and outpatient encounters.
Table IIBivariate analysis showing number (percentage) of encounters with different clinical outcomes [hospitalization, requirement for higher-level care (HLC), lower respiratory tract (LRT) disease on chest x-ray (CXR) and requirement for oxygen support] across different risk factors, not adjusted for multiple occurrences per patient and excluding patients with mixed human coronaviruses
Encounters where HLC was required for admitted patients (N=384)
Risk factor
Inpatient (N=384)
Outpatient (N=64)
P-value
<5 years of age
278 (89%)
33 (11%)
0.0008
Immunocompromised
52 (76%)
16 (24%)
0.018
Tracheostomy
46 (96%)
2 (4%)
0.034
History of prematurity
99 (92%)
9 (8%)
0.0424
Encounters where HLC was required for admitted patients (N=384)
Risk factor
HLC (N=225)
No HLC (N=159)
P-value
<5 years of age
180 (65%)
98 (35%)
<0.0001
Respiratory disease
51 (42%)
70 (58%)
<0.0001
Immunocompromised
20 (38%)
32 (62%)
0.0015
Complex medical problems
39 (46%)
45 (54%)
0.0104
Encounters with requirement for oxygen support for admitted patients (N=384)
Risk factor
Received oxygen support (N=239)
Did not receive oxygen support (N=145)
P-value
Immunocompromised
19 (37%)
33 (63%)
<0.0001
Requirement for oxygen support at baseline
64 (91%)
6 (9%)
<0.0001
Tracheostomy
42 (91%)
4 (9%)
<0.0001
Neuromuscular disease
73 (84%)
14 (16%)
<0.0001
Cardiac disease
50 (81%)
12 (19%)
0.0011
History of prematurity
72 (73%)
27 (27%)
0.0125
Encounters with LRT disease on CXR (N=334)
Risk factor
LRT disease (N=193)
No LRT disease (N=141)
P-value
<5 years of age
142 (62%)
87 (38%)
0.0210
Complex medical problems
51 (69%)
23 (31%)
0.0279
Cardiac disease
29 (46%)
34 (54%)
0.0360
Viral co-pathogens present
116 (63%)
69 (37%)
0.0426
P-values obtained from Chi-squared test comparing proportions of different outcomes across selected risk factor groups. Risk factors with significant P-values are presented here.
Among inpatient encounters, HLC was associated with patients aged <5 years, immunocompetent patients, and patients who did not have underlying respiratory disease or other complex medical problems (Table II). HLC was not associated with other outcomes of interest.
Among inpatient encounters, the requirement for oxygen support was associated with the presence of a tracheostomy tube, oxygen support at home (above baseline home requirement), prematurity, cardiac disease, neuromuscular disorders and immunocompromised status (Table II). The need for oxygen support was not significantly associated with the other tested outcomes.
A CXR was obtained in 334 encounters (315 inpatient and 19 outpatient encounters). Findings suggestive of LRT disease were more frequently associated with patients aged <5 years, patients with other complex medical problems, patients without cardiac disease, and encounters positive for viral co-pathogens (Table II), while immunosuppression and other outcomes were not associated with LRT disease findings on CXR.
The median LOS was longer for patients who had neuromuscular disease, patients aged ≥5 years, patients with cardiac disease and patients requiring oxygen support at baseline, and varied by HCoV type (Table III). LOS did not differ by the other tested outcomes.
Table IIIMedian [interquartile range (IQR)] length of stay (days) for inpatient encounters by different risk factors, excluding possible nosocomial infections (N=332)
Multi-variate analyses were performed to help delineate the relevant predictors of outcome (Table IV). When adjusted for immunosuppression, presence of tracheostomy tube and prematurity, patients aged <5 years were more likely to be admitted to hospital compared with patients aged ≥5 years (aOR 2.2, 95% CI 1.2–4.1). Furthermore, a model predicting HLC (adjusted for chronic respiratory disease, cardiac disease and immunosuppression) indicated that patients aged <5 years were more likely to require HLC (aOR 1.8, 95% CI 1.0–3.1). Surprisingly, HCoV-infected children with immunosuppression (aOR 0.3, 95% CI 0.1–0.5), chronic respiratory disease (aOR 0.3, 95% CI 0.2–0.6) and other complex medical problems (aOR 0.5, 95% CI 0.3–0.9) were less likely to require HLC. When adjusted for prematurity, cardiac disease, neuromuscular disease, requirement for oxygen support at baseline, tracheostomy and immunosuppression, patients with neuromuscular disorders (aOR 2.5, 95% CI 1.3–5.7), cardiac disorders (aOR 3.6, 95% CI 1.66–7.9) or requirement for oxygen support at baseline (aOR 3.7, 95% CI 1.3–11) were more likely to require oxygen support while hospitalized. Immunosuppressed children were less likely to require oxygen support during admission (aOR 0.4, 95% CI 0.2–0.7). Finally, a multi-variate model for LOS showed that encounters with 229E compared with OC43 (aOR 1.7, 95% CI 1.1–2.6; P=0.0174), presence of neuromuscular disorders (aOR 1.4, 95% CI 1.0–2.0; P=0.0433), presence of cardiac disease (aOR 1.8, 95% CI 1.2–2.6; P=0.0113) and requirement for oxygen support at baseline (aOR 1.5, 95% CI 1.0–2.2; P=0.0361) were associated with longer LOS, while the presence of a tracheostomy tube at baseline was associated with shorter LOS (aOR 0.6, 95% CI 0.43–0.96; P=0.0356).
Table IVMulti-variate analyses for different clinical outcomes among patients with different risk factors
Risk factor
Adjusted OR
95% CI
P-value
Odds for having inpatient encounter (N=448)
<5 vs ≥5 years of age
2.2
1.2–4.1
0.0130
Tracheostomy vs no tracheostomy
3.8
0.8–18.0
0.0916
Immunosuppressed vs non-immunosuppressed
0.7
0.3–1.4
0.2817
History of prematurity vs no history of prematurity
1.4
0.6–3.2
0.4056
Odds for requiring HLC care while hospitalized (N=384)
Respiratory disease vs no respiratory disease
0.3
0.2–0.6
0.0005
Immunosuppressed vs non-immunosuppressed
0.3
0.1–0.5
0.0011
Complex medical problems vs no complex medical problems
0.5
0.3–0.9
0.0184
<5 vs ≥5 years of age
1.8
1.0–3.1
0.0390
Odds for requiring oxygen support while hospitalized (N=384)
Cardiac disease vs no cardiac disease
3.6
1.6–7.9
0.0035
Immunosuppressed vs non-immunosuppressed
0.4
0.2–0.7
0.0080
Neuromuscular disease vs no neuromuscular disease
2.5
1.3–5.7
0.0102
Requirement for oxygen support at baseline vs no requirement for oxygen support at baseline
3.7
1.3–11.0
0.0209
Tracheostomy vs no tracheostomy
1.9
0.5–7.2
0.3013
History of prematurity vs no history of prematurity
1.2
0.6–2.1
0.6169
Odds for having LRT disease among patients who received CXR (N=334)
Fifty-two (11%) encounters were identified as possibly nosocomial (HCoV detected after 3 days of hospitalization), with a median LOS of 26.5 (IQR 7.50–99) hospital days when the first VRP was obtained. Nosocomial infections were associated with longer median LOS (87.5 days, IQR 27.5–223 days) compared with community-acquired infections (median 4 days, IQR 2–9 days; P<0.0001). Nosocomial infections were more likely in patients with a tracheostomy tube at baseline (28% vs 12%; P=0.0019), immunosuppressed patients (27% vs 11%; P=0.0024), patients without viral co-pathogens (19% vs 9%; P=0.0046), patients who required HLC (17% vs 9%; P=0.0226) and patients aged <5 years (16% vs 8%; P=0.0340). HCoV types did not differ significantly between nosocomial and community-acquired infections.
Recurrent presentations and mixed infections
Out of the 430 patients, 23 were identified to have a total of 48 encounters with a different HCoV infection (two patients had three encounters each and 21 patients had two encounters each). Among those recurrent encounters, there was a median interval of 357 days (IQR 87–456 days) between HCoV detection. Fifty-four percent of recurrent infections occurred in patients aged ≥5 years, 85% occurred in inpatients, 28% occurred in immunocompromised patients, 44% occurred in patients with viral co-pathogens, 33% occurred in patients with chronic respiratory disease, and 23% were classified as possible nosocomial infections. Oxygen support was provided to patients in 67% of the encounters, and HLC was needed in 38% of the encounters. Seven patients were found to have two different HCoVs simultaneously. Of these mixed HCoV infections, five patients were hospitalized, three patients needed oxygen support, and no patients required HLC. OC43 was the most common HCoV detected in this subgroup (detected in six patients).
Discussion
This study presents one of the largest cohorts of paediatric patients with endemic HCoVs. Key findings provide insight into future needs for the management of HCoV infections in young children. As predicted, the findings show that OC43 was the most common HCoV detected during the study period, similar to other studies undertaken in the USA [
Epidemiology and clinical characteristics of human coronaviruses OC43, 229E, NL63, and HKU1: a study of hospitalized children with acute respiratory tract infection in Guangzhou, China.
Only HCoV type was associated with LOS, as inpatient encounters with 229E were associated with longer LOS. Other outcomes did not differ between HCoV types. Similar to the present findings, some studies [7,26] found no significant differences between HCoV types and disease severity. In contrast, other reports noted few differences in clinical presentation between HCoV types [
Epidemiology and clinical characteristics of human coronaviruses OC43, 229E, NL63, and HKU1: a study of hospitalized children with acute respiratory tract infection in Guangzhou, China.
This study found that younger patients (age <5 years) were more likely to be admitted to hospital, require HLC, and have CXR findings suggestive of LRT disease. Studies of paediatric patients with HCoVs have yielded variable results. Some studies have shown that younger patients had a higher likelihood of developing LRT disease and a higher need for care escalation or oxygen support [
Genetic variability of human coronavirus OC43-, 229E-, and NL63-like strains and their association with lower respiratory tract infections of hospitalized infants and immunocompromised patients.
Importantly, the present study showed that immunosuppressed patients were less likely to require HLC or need oxygen support when hospitalized. Patients in this study with underlying respiratory diseases were also less likely to need HLC. Other studies in paediatric patients provided inconsistent results about the role of immunosuppression or the presence of respiratory diseases as possible risk factors for disease enhancement. For example, in a review of similar patients with HCoV infections, immunosuppression was associated with severe LRT disease. Further, underlying respiratory disease was associated with a higher risk of developing LRT disease and severe LRT disease [
]. Cardiovascular disease and congenital diseases, but not immunosuppression, have been associated with the need for respiratory support and intensive care unit admission [
]. In contrast, the present study found that patients with complex medical problems were less likely to require HLC. The findings of better outcomes in some potentially high-risk groups may reflect PCR utilization at the study institution, where, especially in the early stages of PCR deployment, it was not unusual to test high-risk patients with milder disease while testing of their healthier counterparts was reserved for when escalation of care was needed. With COVID-19 being driven by an over-response of the immune system, to a large extent, it is possible to hypothesize that disease caused by the endemic HCoVs may have a similar pathogenesis.
Seasonal changes in detected HCoV infections were similar throughout the study period (Figure 2) for all different types. The sharp increase in annual HCoV-positive tests was likely due to increased utilization of PCR-based testing at the study institution. Different seasonal distribution of HCoV types has been reported in some studies [
], while the present study showed that OC43 was the most common during the study period.
An interesting finding from this study was the presence of 52 (11%) possible nosocomial infections. Nosocomial HCOV infections have been reported in previous studies. One report utilizing indirect immunofluorescence and cell culture found that HCoVs were the most common reason for viral nosocomial infections in a neonatal intensive care unit [
]. Although the definition of possible nosocomial infections used in the present study may be strict, more than 50% of those encounters were for patients who had been admitted for ≥3 weeks before testing positive for HCoV. Respiratory viral nosocomial infections (including HCoVs) in neonates have been associated with the escalation of care and longer LOS, especially in premature neonates [
This study also serves to highlight the prevalence of recurrent and mixed HCoV infections in the paediatric population. Twenty-three patients had more than one encounter due to different HCoV types during the study period. The majority were inpatient encounters in patients aged ≥5 years. Similarly, seven patients with more than one HCoV infection simultaneously were identified. The detection of multiple HCoV types in the same sample has been reported with varying frequencies [
Genetic variability of human coronavirus OC43-, 229E-, and NL63-like strains and their association with lower respiratory tract infections of hospitalized infants and immunocompromised patients.
]. However, due to the small sample size (N=7), it is unclear whether having more than one HCoV type simultaneously is associated with different clinical presentations or outcomes.
Twelve patients had repeated detection of the same HCoV type over 29 encounters. These encounters raise the possibility of re-infection with the same HCoV type or prolonged shedding of the virus, although the retrospective nature of this study precludes definitive analysis in this regard. It is postulated that some of these encounters may be less likely due to re-infection, as the intervals between the two encounters were relatively short (within 2 months) for the majority. Prolonged shedding of HCoV for up to 60 days has been described in immunocompromised patients [
]. On the other hand, re-infection with the same HCoV type in children has also been demonstrated. HCoV endemic infections in coastal Kenya were tracked, and NL63, 229E and OC43 re-infections were observed with varying intervals between re-infections. Re-infections were less common with 229E and OC43. Genome sequencing data suggest limited genetic switching, which may be an indication for lack of immune selection [
]. The immune response to endemic HCoV infections may not be long-lasting in some patients, and re-infections may be possible.
Limitations of this study included its retrospective nature and reliance on information as recorded in the EMRs. Moreover, the study was limited by the under-representation of paediatric patients who did not have major risk factors for severe disease. Although significance can be demonstrated, the differences between some groups were small; for example, increased median LOS in patients with certain risk factors (Table III).
In conclusion, HCoVs are important pathogens in the paediatric population that have a significant burden of disease. Hospitalization and care escalation are more common in younger patients (age <5 years). This study of the pre-COVID-19 era demonstrates that the clinical impact of the four endemic HCoV types is similar. The impact of HCoVs as nosocomial infections may be under-recognized. Future prospective studies are warranted to delineate the roles of HCoV infections in the paediatric population, especially in high-risk groups. Importantly, as antiviral drugs are developed to manage other coronavirus infections, namely COVID-19, therapy in this population warrants consideration.
Conflict of interest statement
None declared.
Funding sources
None.
Ethical approval
This work did not involve human participants and was approved by the Institutional Review Board of the University of Alabama at Birmingham.
References
Cherry J.
Demmler-Harrison G.
Kaplan S.
Steinbach W.
Hotez P.
Feigin and Cherry's textbook of pediatric infectious diseases.
Community surveillance of respiratory viruses among families in the Utah Better Identification of Germs-Longitudinal Viral Epidemiology (BIG-LoVE) study.
Epidemiology and clinical presentations of the four human coronaviruses 229E, HKU1, NL63, and OC43 detected over 3 years using a novel multiplex real-time PCR method.
Epidemiology and clinical characteristics of human coronaviruses OC43, 229E, NL63, and HKU1: a study of hospitalized children with acute respiratory tract infection in Guangzhou, China.
Genetic variability of human coronavirus OC43-, 229E-, and NL63-like strains and their association with lower respiratory tract infections of hospitalized infants and immunocompromised patients.
00014-2&id=gr1.jpg){kind=link}
00014-2&id=gr2.jpg){kind=link}