Clinical Presentation and Outcome of Acute Respiratory Illnesses in South African Children During the COVID-19 Pandemic

Background: Data from low- and middle-income countries (LMICs) show higher morbidity and mortality in children with acute respiratory illness (ARI) from severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2). However, whether SARS-CoV-2 infection is distinct from other causes of ARI in this regard is unclear. We describe clinical characteristics and outcomes of South African children with SARS-CoV-2 and non-SARS-CoV-2 ARIs. Methods: We performed a cross-sectional study including 0–13 years old children admitted to Tygerberg Hospital between May and December 2020 with an ARI. Routine clinical data were collected by the attending clinicians. All children underwent SARS-CoV-2 polymerase chain reaction testing. For severity of disease, the need for respiratory support and duration of support was considered. Multivariable logistic regression models were built to determine the factors associated with SARS-CoV-2 infection and severity. Results: Data for 176 children were available, 38 (22%) children were SARS-CoV-2 polymerase chain reaction positive and 138 (78%) were negative. SARS-CoV-2 positive children were more likely to be female (OR: 2.68, 95% CI: 1.18–6.07), had lower weight-for-age Z score (OR: 0.76, 95% CI: 0.63–0.93), presented more frequently with fever (OR: 3.56, 95% CI: 1.54–8.24) and less often with cough (OR: 0.27, 95% CI: 0.11–0.66). SARS-CoV-2 infection was associated with significantly longer duration of oxygen treatment (median 8 vs. 3 days; OR: 1.1, 95% CI: 1.01–1.20). Overall, 66% of children had viral coinfection, with no significant difference between the groups. In total, 18% of SARS-CoV-2 positive children were readmitted within 3 months for a respiratory reason, compared with 15% SARS-CoV-2 negative children (P = 0.64). Conclusions: Our data show that ARIs from SARS-CoV-2 cannot be easily differentiated, but were associated with a higher morbidity compared with ARIs from other causes. Overall outcomes were good. The long-term implications of severe SARS-CoV-2 pneumonia in young children in low- and middle-income countries require further study.


Study Design
We performed a cross-sectional study including 0-to 13-year-old children, who were admitted to Tygerberg Hospital between May 5, 2020 (week 19) and December 5, 2020 (week 48) with an ARI.This period falls within the beginning of autumn until the beginning of summer.South Africa reported its first case of SARS-CoV-2 on March 5, 2020, with the first wave officially starting at the beginning of May 2020 in the Western Cape (week 19). 14

Setting
Tygerberg Hospital is a large public tertiary hospital in the Western Cape province of South Africa providing specialist and subspecialist pediatric services.General pediatric services in South Africa provide care for children up to the age of 13 years, which is the rationale for the age cutoff for our study population.The majority of patients accessing the hospital come from poor socio-economic circumstances.Due to limited critical care beds, respiratory support in the form of nasal continuous positive airway pressure using Fisher and Paykel variable flow system, or high-flow nasal cannula oxygen therapy using the Airvo Optiflow system are mostly provided in the pediatric emergency unit and general pediatric wards.

Study Participants and Definitions
Children admitted with symptoms of an ARI and who had a SARS-CoV-2 PCR test were included.ARIs were classified into one of the 3 diagnostic categories according to the main presenting feature, namely upper respiratory tract infection (URTI), lower respiratory tract infection (LRTI) or asthma/wheeze.URTI was defined as having at least 1 respiratory symptom, without any signs of tachypnea or respiratory distress.LRTI was defined as any child with respiratory symptom(s) and evidence of tachypnea or respiratory distress, including acute viral bronchiolitis.Asthma (>5 years) or wheeze (≤5 years) were classified if the predominant clinical presentation was wheeze/reversible bronchospasm.We excluded babies admitted to the neonatal service, children whose SARS-CoV-2 PCR was positive but who did not have any respiratory symptoms, and children in whom a SARS-CoV-2 PCR result was not available.Children with multisystem inflammatory syndrome in children were only included if their presentation was as an ARI.

Data Collection
A standard data collection form was used to capture routine clinical and laboratory information at the time of presentation to pediatric emergency unit or admission to a ward.All cases were reviewed 3-6 months after the initial admission to assess midterm outcome and record any readmissions within 3 months of initial presentation.Missing information was subsequently retrospectively captured using patient clinical records and the National Health Laboratory system.Unfortunately, in cases where the data collection form was not fully completed at the time of admission, certain data were missing despite retrospective review of the medical records, and consequently not all variables were captured for all cases.De-identified data were captured onto a secure RedCap database.A waiver of individual consent was obtained from Stellenbosch University Human Research Ethics committee (HREC N20/04/013_COVID).Data of some children were previously published. 6,7e-identified digital chest radiographs (CXR) were reported by 2 external expert reviewers, who were blinded to any clinical information apart from age.Radiological features were described in accordance with World Health Organization classification of pneumonia as either alveolar infiltrate, other infiltrate (which included interstitial infiltrates) or no pneumonia. 15Weight-for-age Z-scores (WAZ) were calculated using a World Health Organization-based app developed by the Canadian Pediatric Endocrine group. 16hroughout the study period, all children admitted with ARI symptoms underwent real-time reverse-transcription polymerase chain reaction testing for SARS-CoV-2 via a single nasopharyngeal swab as part of routine care.Testing for other respiratory viruses was not performed routinely, but at the discretion of the clinical team or retrospectively as part of the study if samples were still available.Tests were performed in a registered virology laboratory using Allplex 2019-nCoV assay (Seegene Inc) and manufacturer's cycle threshold cutoff values were used.Multiplex PCR testing for respiratory viral panel was performed using either the Anyplex II RV16 detection assay or the Allplex RV Essential Assay (both from Seegene Inc).Epidemiologic data from the National Institute of Communicable disease shows that the Ancestral type variant (May 3, 2020 to August 16, 2020) and Beta variant (November 8, 2020 to February 7, 2021) were present during the study period. 17

Data Analysis
Data analysis was done using IBM SPSS Statistics v27.The χ 2 test was used to compare categorical variables, and independent samples t test or Mann-Whitney U test were used to compare continuous variables.Multivariable analysis was performed using logistic regression models to study the association of demographic/clinical variables with SARS-CoV-2 infection and severity.Variables included in multivariable analysis model assessing SARS-CoV-2 positivity were age, sex, WAZ, any comorbidity, COVID contact, HIV exposure or disease, TB exposure or disease and reported symptoms and signs as listed in Table 2. Variables included in the model assessing severity and outcome were age, gender, need for oxygen or further respiratory support, need for pediatric intensive care unit (PICU) admission, length of stay and number of days on oxygen.Level of significance was set at 0.05.Laboratory markers were excluded from the multivariable models because of a high percentage of missing data.

RESULTS
A total of 176 children were included in our study, of which 38 (21.6%) were SARS-CoV-2 PCR positive and 138 (78.4%) were SARS-CoV-2 PCR negative.

Clinical Presentation
Comparison of the clinical features, special investigations, treatment and outcomes in SARS-CoV-2 positive and negative children is shown in Table 2.In univariable analysis, measured or reported fever was more common in children with SARS-CoV-2 infection compared with those without (68% vs. 40%; P = 0.01).Children with SARS-CoV-2 were less likely to present with cough (63% vs. 85%; P = 0.01), tight chest (48% vs. 70%; P = 0.02) or wheeze (33% vs. 51%; P = 0.03).There were no other symptoms or signs, which were significantly different between the groups.

Severity of Disease and Treatment
A large percentage of children required oxygen supplementation (132/176, 74%) or further respiratory support in the form of high-flow, continuous positive airway pressure or invasive positive pressure ventilatin (IPPV) (51/136, 28%), in keeping with their need for admission at a referral hospital.SARS-CoV-2 positive children were not more likely to need oxygen (76% vs. 75%, P = 0.83) or respiratory support (38% vs. 26%, P = 0.11), but were more likely to require intensive care admission (18% vs. 7%, P = 0.03), required oxygen supplementation for longer (median 6 vs. 2 days, P = 0.01) and had significantly longer hospital stay (median 7 vs. 3 days, P = 0.01).Multivariable logistic regression analysis was performed, which included age, gender, need for oxygen or further respiratory support, PICU admission, length of stay and number of days on oxygen.This showed number of days on oxygen (OR: 1.1, 95% CI: 1.01-1.20)as the only factor independently associated with SARS-CoV-2 infection.In terms of treatment received, steroid therapy, antibiotics and bronchodilator therapy were similar between the groups.

Clinical Outcomes
There were no deaths, apart from a child with acute myeloid leukemia who died of complications related to his underlying condition 3 months after the initial presentation.A total of 27 children were readmitted within 3 months for a respiratory reason, but there was no significant difference between SARS-CoV-2 positive and negative groups (18% vs. 15%, P = 0.64).Of the readmitted children who were initially SARS-CoV-2 negative, 17 of 20 had a repeat SARS-CoV-2 PCR upon readmission, of which none were positive.Of the 7 children readmitted who were initially SARS-CoV-2 positive, 2 tested SARS-CoV-2 PCR positive again between 1 and 3 months later.

DISCUSSION
Our data in hospitalized South African children show that SARS-CoV-2-associated ARI was associated with younger age, female sex and lower WAZ score, and had higher morbidity compared with ARI in SARS-CoV-2 negative children.Although we found that SARS-CoV-2 positive children presented more frequently with fever, but less frequently with cough and wheeze, they cannot be easily distinguished based on the clinical presentation.This has implications for hospital testing and isolation strategies during future waves.
Nearly one third of all children admitted with ARI had comorbid conditions, which was similar between the groups.In total, 10 children had current TB disease or history of TB in this study, including 2 children with SARS-CoV-2 infection.Adult data show that a history of TB is a risk factor for more severe disease, but our numbers were too small to analyze this further. 18,19Although currently data in children is lacking, the potential impact of the COVID-19 pandemic on disruption of TB national control programs, as well as the potential for more severe TB or COVID-19, means that this is an important area for further research.No other studies, as far as we are aware, have reported lower WAZ score in children with SARS-CoV-2.The reasons for this are not clear.The lower WAZ score may reflect poorer socio-economic circumstances, and SARS-CoV-2 may spread more in overcrowded areas where there is potentially less ability to isolate and adhere to control measures, but this would probably apply to other respiratory pathogens too.
We found no difference in laboratory parameters, in contrast to Jimenez-Garcia et al and other studies of COVID-19 in children. 3,6,20The reason for this may be the young median age of our cohort or the high prevalence of coinfection.We found alveolar infiltrates in 48% of SARS-CoV-2 positive children, compared with 31% who were negative, and no interstitial infiltrates in the positive group.Ground glass infiltrates and consolidation are the commonest radiographical findings reported in children with SARS-CoV-2.Toba et al 3 reported CXR findings of interstitial infiltrates in 18% of SARS-CoV-2 children and consolidation in 15%.The reason for lack of finding interstitial infiltrates in our study is not clear.
Severity of the disease and outcomes were worse in the SARS-CoV-2 positive children.They tended to require more  Respiratory Illness in Children During COVID respiratory support and had increased PICU admission.Most notably, children with SARS-CoV-2 were significantly longer on oxygen compared with SARS-CoV-2 negative children, as well as having an increased length of hospital stay.This is consistent with our previous findings, which showed that SARS-CoV-2 infection in younger children was associated with hospital admission and need for respiratory support, but this also highlights the difference between SARS-CoV-2 positive and negative children. 7lthough SARS-CoV-2 infections have been generally reported as mild in children, our data show that there is substantial morbidity in children.Jimenez-Garcia et al 21 also found longer oxygen duration among SARS-CoV-2 positive children, but no increased length of stay.This high morbidity may be partly explained by the lower median age of the SARS-CoV-2 positive group.In contrast to most studies of pediatric COVID-19, nearly half of our study population were infants, and infants were more likely to be SARS-CoV-2 positive. 3,8,20,22There is increasing evidence that young age is a risk factor for severe disease from SARS-CoV-2.Bhuiyan et al showed that half of "young COVID-19" (ie, <years) cases occurred in infants, with 7% of these cases requiring intensive care unit admission. 22raff et al reported that infants <3 months of age with COVID-19 were more likely to require admission. 23A large study in the 6 sub-Saharan African countries of 469 children, which included children reported in this study, showed a higher rate of morbidity and mortality in children compared with HICs, with young age and comorbidities being the significant predictors of severe disease. 7he role of coinfection requires further consideration.A reduction in the detection of respiratory viruses and seasonal variation due to SARS-CoV-2 control measures was reported worldwide during the initial months of the pandemic. 9,10,24,25Our study took place during the first and beginning of the second COVID-19 wave when heavy lockdown restrictions were in place, which included travel restrictions, school closure, no gatherings, social distancing and wearing of masks in public places.Despite these measures, the detection of other respiratory viruses in 66% of patients is much higher than that reported in other studies. 3,12RSV, HRV and adenovirus were most commonly detected, but notably there was almost no influenza.This observation emphasizes that the effect of infection control/lockdown measures on the prevalence of other viral pathogens is highly setting-specific.The catchment area of our hospital mainly consists of urban slums and informal settlements, where the effect of infection prevention measures on viral transmission is likely to be minimal.
Although no children died from ARI in our study, the risk of excess mortality from SARS-CoV-2 in LMIC settings with younger populations and less healthcare resources necessitates comprehensive public health interventions.Furthermore, the potential effects of early SARS-CoV-2 infection on the long-term lung health of children requires further study.7][28][29] A significant proportion of children in our study were readmitted within 3 months for a respiratory reason.At this stage, the potential long-term impact of mild or severe SARS-CoV-2 infection is unknown, but its apparent predilection to cause severe infection in young infants suggests that follow up of these children is an area for further research to understand the long-term impact.
Although we were able to compare COVID-19 and non-COVID-19 respiratory disease and viral coinfections in a cohort of South African children, our study has several limitations.First, it is a convenience sample from a single center of admitted children, reflecting the more severely ill children.As we tried to capture as many patients as possible during the study period, we did not attempt to match SARS-CoV-2 cases with controls.Therefore, there is a discrepancy in the group sizes, which possibly influenced the analysis.Second, due to the young median age of our group, symptoms such as headache, altered taste or smell and some abdominal symptoms may have been under-reported.Lastly, we describe only the initial and beginning of the second wave of the pandemic in South Africa, and it is now known that subsequent variants of the SARS-CoV-2 virus may affect children differently.

CONCLUSIONS
In this study, we found that among children admitted with ARI, younger age, female gender and low WAZ were associated with SARS-CoV-2 infection.SARS-CoV-2 infection was associated with significantly longer oxygen requirement.This study provides further evidence of severe COVID-19 infection among infants, with high morbidity and health resource implications despite low mortality rates.The potential impact on the long-term lung health of these babies requires further investigation.

Current Abstracts
Edited by: Robert J. Leggiadro Erythema migrans (EM), the most common manifestation of Lyme disease, can appear as a classic bullseye or as a homogenous erythematous lesion.Prompt recognition of an EM lesion has important clinical implications because patients who receive early antibiotic therapy have an excellent prognosis and are unlikely to develop cardiac, musculoskeletal, and neurologic manifestations of Lyme disease.However, the prompt recognition of an EM lesion can prove challenging.
One challenge is that skin type and pigment can impact the appearance of EM, making it harder to recognize.In particular, darker skin tones are underrepresented in clinician educational materials, further limiting clinicians' ability to recognize disease in these individuals.The goal of this multicenter prospective study was to identify the association between Black race and the diagnosis of Lyme disease in a cohort of children undergoing evaluation for suspected Lyme disease.In addition, differences between Lyme disease symptoms at presentation by race were examined.
The study was performed at 8 emergency departments participating in the Pedi Lyme Net clinical research network between June 2015 and August 2022.Primary outcome was a case of Lyme disease, characterized by either an EM lesion measuring at least 5 cm in diameter and/or a positive 2-tiered Lyme disease serology in the appropriate clinical scenario.Using United States census crosswalks, each 5-digit residential ZIP code was mapped to one or more counties contained within.Using an online Lyme disease dashboard, incidence was assigned as the population-weighted mean of county incidences contained within the ZIP code.
Overall, 957 children (23.9% of enrolled) had Lyme disease.Of these, 88 (9.2%) had a diagnostic EM lesion alone, 781 (81.6%) had a positive 2-tier serology alone, and 88 (9.2%) had both.Of the 176 with a diagnostic EM lesion, 106 had a single and 70 had multiple EM lesions measuring 5 cm or greater in diameter.Most children had neurologic symptoms or arthritis.Children with Lyme disease were clustered around participating centers and in counties of high Lyme disease incidence.
Compared with all other races, Black children were less likely to have Lyme disease diagnosed in the emergency department after adjustment for age and local Lyme disease incidence (adjusted odds ratio (aOR), 0.63; 95% confidence interval (CI),.48-.81).For children with Lyme disease, Black children were less likely to be diagnosed with cutaneous manifestations of Lyme disease (aOR, 0.34; 95% CI,.14-.79) but more likely to be diagnosed with a swollen joint on examination (aOR, 3.68; 95% CI, 2.13-6.36).
Comment: Black children were less likely to be diagnosed with Lyme disease when compared to children of other races in this study.Interestingly, Black children with Lyme disease were less likely to be diagnosed with cutaneous manifestations and more likely to be diagnosed with arthritis after adjustment for other factors related to Lyme disease risk.Improving education and guidelines around Lyme disease diagnosis across different skin types is warranted to reduce disparities in Lyme disease diagnosis.

TABLE 1 .
Demographic Characteristics of Patients

TABLE 4 .
Results of Respiratory Viral Panel Tests