key: cord-1003387-kqenysho
authors: Satoskar, S.; Hindman, D.; Husain, A.; Prichett, L.; Badaki, O.; Kane, A.
title: Correlation of Suspected COVID-19 Symptoms with COVID-19 Positivity in Children
date: 2022-05-05
journal: nan
DOI: 10.1101/2022.05.03.22274641
sha: b67991aca610db798ba6ed270c6da69b92c6e9bb
doc_id: 1003387
cord_uid: kqenysho

Background: Early in the pandemic, COVID-19 was found to infect adults at higher rates than children, leaving limited data on disease presentation in children. Further understanding of the epidemiology of COVID-19 symptoms among children is needed. Our aim was to explore how symptoms vary between children testing positive for COVID-19 infection versus children testing negative. Methods: Data analysis of symptom prevalence among pediatric patients presenting to emergency departments (ED) in the Johns Hopkins Health System (JHHS) with concern for COVID-19 who subsequently received COVID-19 testing. Inclusion criteria included patients 0-17 years-of-age, ED evaluation between March 15th, 2020 - May 11th, 2020, and those who were ordered for COVID-19 testing. Chart review was performed to document symptoms using ED provider notes. Comparisons were made using chi-squared t-tests and Students t-tests. Results: Fever (62.6%) and cough (47.9%) were the most prevalent symptoms among children with suspected COVID-19 infection. Compared to children with a negative COVID-19 test, children who tested positive had higher prevalence of myalgia (21.7% vs 6.0%) and loss of taste/smell (15.2% vs 0.9%). Over half of the children who tested positive for COVID-19 had public insurance (52.2%) and 58.7% of the positive tests occurred among children with Hispanic ethnicity. Conclusions: Myalgia and loss of taste/smell were found to be significantly more prevalent among COVID-19 positive children compared to children testing negative. Additionally, children with public insurance and those with Hispanic ethnicity were more likely to test positive, emphasizing the importance of social factors in the screening and decision-making process.

Although many children have asymptomatic infections [6] , studies show that children with 80 symptomatic SARS-CoV-2 infections develop symptoms that range in severity across multiple organ 81 systems. These symptoms have been shown to primarily affect the respiratory system, and include 82 cough, sore throat, and shortness of breath [7] . Gastrointestinal (GI) symptoms are also common and 83 include diarrhea and vomiting. Recent evidence shows that the Multisystem Inflammatory Syndrome 84 (MIS-C), a serious condition resulting in systemic inflammation that adversely affects the heart [8], 85 lungs, kidneys and GI organs, is also associated with COVID-19 in children [9] [10] [11] . 86 However, an analysis to further explore the association between individual symptoms with 87 COVID-19 test positivity has not been performed in children. Assessing whether certain symptoms 88 are more specific for a positive COVID-19 test has utility to the clinical provider, providing a gauge 89 that helps to discern whether any particular symptom is likely to place a patient in one of two groups: 90 COVID-19 illness or not COVID-19 illness. This is crucial as it can help aid in risk stratification and 91 efficient utilization of testing resources in children, which is critical to pandemic control [12] [13] 92 especially when resources are scarce, as was evident during the early months of the pandemic.

Given the varying presentations of pediatric COVID-19 illness, we conducted an analysis of 94 symptom prevalence among pediatric patients presenting to emergency departments with concern for 95 COVID-19 who subsequently received COVID-19 testing. We collected data from three pediatric 96 emergency department (ED) facilities in a single health system during the initial stages of the 97 pandemic. Our primary specific aim was to explore how symptoms vary between children testing 98 positive for COVID-19 infection versus children testing negative, which was used to associate 99 individual symptoms to COVID-19 test positivity. 100 2 Methods 101 2.1 Design and Sample: 102 We performed a retrospective secondary data analysis of children from 0 through 17 years of 103 age who presented with possible COVID-19 disease to three Pediatric ED facilities in a single health 104 system -two suburban community-based hospitals (PED 2, PED 3) and an urban designated level 1 105 pediatric trauma center (PED 1). Relevant encounters took place between March 15 th , 2020 and May 106 11 th , 2020. We included any patient who was ordered for COVID-19 testing during the time period. 107 COVID-19 testing was ordered based on if the patient met the testing algorithm criteria set at the . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted May 5, 2022. ; https://doi.org/10.1101/2022.05.03.22274641 doi: medRxiv preprint 108 time, and this algorithm changed over the study period as the pandemic evolved. SARS-CoV-2 109 testing was performed using a PCR assay. Molecular assays used for diagnosis at the time included 110 Xpert Xpress SARS-CoV-2/Flu/RSV (Cepheid), NeuMoDx SARS-CoV-2 (Qiagen), Cobas SARS-111 CoV-2 (Roche), ePlex Respiratory Pathogen Panel 2 (Roche), Aptima SARS-CoV-2 (Hologic), 112 Accula SARS-CoV-2 assays (ThermoFisher Scientific), and RealStar® SARS-CoV-2 RT-PCR 113 (Altona Diagnostics) [14] [15] [16] [17] . Choice of test method was based on what was optimal for fastest result 114 turnaround time given the staffing during the shift and test volumes. ED visits that were transfers 115 from one of the other two facilities were excluded. This study was approved by the Johns Hopkins 116 University Institutional Review Board (IRB00246826). 117 2.2 Data Source: 118 We utilized data collected via a Precision Medicine Analytics Data Commons Platform 119 (PMAP) based on electronic medical records (EMR) [18] . Relevant variables included the date of the 120 visit, the associated pediatric emergency department, vital signs, and patient demographics, including 121 age, sex, race, ethnicity, and insurance. We also included receipt of SARS-CoV-2 testing, results of 122 SARS-CoV-2 testing, isolation orders, acuity measured with the emergency severity index (ESI), 123 presence of influenza-like illness (ILI) symptoms, and the patient's area deprivation index (ADI) 124 based on the census block.

For every encounter where SARS-CoV-2 testing was obtained, we performed chart review 126 based on a pre-determined protocol to document the presence or absence of symptoms using provider 127 notes in the ED. Pertinent symptoms included fever, headache, sore throat, cough, shortness of 128 breath, myalgias, diarrhea, and loss of taste or smell. We used REDCap for data entry. All charts that 129 were flagged for questions were reviewed by a single senior investigator to resolve any ambiguity. 130 Missing data were not imputed, however there were very little missing data. Out of the total 516 131 patients, 18 patients had missing data for some of the variables and these patients were excluded from 132 the analysis involving those variables. 133 2.3 Data Analysis: 134 We analyzed data using Microsoft Excel for Mac 2021 and StataIC 16.1 (StataCorp. 2019. 135 Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC). The prevalence of each 136 pertinent symptom was found by dividing the number of children presenting to the ED with that 137 symptom over the total ED visits in our cohort during that time period. The prevalence rate of each 138 symptom was then compared between children who tested positive for SARS-CoV-2 to those who 139 tested negative for SARS-CoV-2. Subjective parameters including myalgias, headaches, and loss of 140 taste or smell were analyzed by a subgroup analysis which included only children above 5 years-of-141 age. We made comparisons using chi-squared t-tests for categorical variables and Student's t tests for 142 continuous variables.

We identified 516 patient encounters where providers were concerned for possible COVID-19 145 and subsequently tested for SARS-CoV-2. There were 46 (8.9%) children who tested positive for 146 SARS-CoV-2 and 470 (91.1%) who tested negative. Table 1 presents summary descriptive 147 characteristics of the study population.

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 5, 2022. ; https://doi.org/10.1101/2022.05.03.22274641 doi: medRxiv preprint

The majority of COVID-19 cases occurred in the 12-17 years age group (47.8%). The mean 149 age among those who tested positive was 8.76 years compared to 6.64 years among those who were 150 negative (p = 0.01). Although the majority of our study population consisted of non-Hispanics 151 (78.5%), 58.7% of positive tests occurred among children with Hispanic ethnicity. The majority of 152 children who tested negative for SARS-CoV-2 were either non-Hispanic Black (35.1%) or non-153 Hispanic White (30.4%). Over half of the children who tested positive for SARS-CoV-2 had public 154 health insurance (52.2%).

155

Fever (62.6%) and cough (47.9%) were the most prevalent symptoms among children 156 suspected of SARS-CoV-2. When compared to children with a negative SARS-CoV-2 test, children 157 who tested positive had a significantly higher prevalence of myalgia (21.7% vs 6.0%, p-value = 158 0.001) and loss of taste or smell (15.2% vs 0.9%, p-value < 0.001), as summarized in Table 2 . There 159 were no significant differences in the prevalence of sore throat, cough, shortness of breath, and 160 diarrhea symptoms between children testing positive versus negative for SARS-CoV-2.

In this retrospective secondary data analysis utilizing EMR data and chart review, we found 163 that among children who present with suspected COVID-19 infection, fever and cough are the most 164 prevalent symptoms. However, we find that symptoms of myalgia and loss of taste or smell are 165 associated with COVID-19 positivity in our pediatric patient population.

Although many studies have documented a variety of symptoms, including these [6-7], fever 167 and cough have been reported as two predominant symptoms associated with SARS-CoV-2 infection 168 in children [7-8]. Fever and cough were also found to be present in the majority of children who 169 underwent testing for COVID-19 infection in our study. However, interestingly, it was myalgia and 170 loss of taste or smell that were significantly associated with a positive COVID-19 test result. Myalgia 171 symptoms have been implicated in many inflammatory conditions [9]. It may be possible that in 172 children suspected of COVID-19 infection, myalgia may be a direct manifestation of the systemic 173 inflammatory state implicated in COVID-19 pathogenesis, and therefore a more specific indicator for 174 positive COVID-19 infection compared to other viral respiratory symptoms alone.

Our study took place during the early months of the pandemic, when testing resources were 176 scarce and limited, and as a result resource allocation and utilization had to be done prudently. 177 Currently, COVID-19 infection rates are increasing in the pediatric population [19-20], and several 178 health systems are starting to face similar testing resource shortages as was evident early in the 179 pandemic [21]. Our study identifies symptoms that are most specific for SARS-CoV-2 positivity in 180 our pediatric study population, which is essential to aid in strengthening COVID-19 test specificity to 181 aid in efficient resource utilization.

In our study population, racial and socioeconomic factors had a strong association with 183 SARS-CoV-2 positivity. Children who had public insurance or no insurance were significantly more 184 likely to test positive for SARS-CoV-2 compared to children with private insurance. Additionally, we 185 found ethnicity to be significantly associated with SARS-CoV-2 positivity where Hispanic children 186 had a higher likelihood of testing positive compared to non-Hispanic children. Although this 187 phenomenon has been documented in adults, our study provides further insight among the pediatric 188 population. This is especially crucial since the urban hospitals under the Johns Hopkins Health 189 System (JHHS) predominantly serve Black and Hispanic populations, mainly from within and 190 surrounding the Baltimore area. Interestingly in our study, most of our study population consisted of . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 5, 2022. ; https://doi.org/10.1101/2022.05.03.22274641 doi: medRxiv preprint

As evidence continues to grow, our study provides further insight into COVID-19 symptoms 216 among the pediatric population. However, the presence or absence of any single symptom, other than 217 possibly the sudden loss of taste or smell, is not sufficient to guide clinical decision-making in testing 218 for SARS-CoV-2 among pediatric patients presenting to emergency facilities. When considering 219 testing for COVID-19, clinicians should be attuned to other factors that may predispose patients to 220 increased risk, such as social determinants of health and patients' social habits. Additionally, given 221 that COVID-19 infectivity is growing in the pediatric population with the addition of newer variants 222 as compared to the start of the pandemic, it is essential that clinicians maintain a low threshold for 223 testing children with a variety of symptoms, allowing for improved disease detection and aiding the 224 prevention of subsequent spread. Table 1 . Patient Characteristics of those tested for COVID-19. Note. NH = non-Hispanic, ADI = area 234 deprivation index. The a p-value is a comparison of patient characteristics between children who 235 tested positive versus children who tested negative for . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted May 5, 2022. ; https://doi.org/10.1101/2022.05.03.22274641 doi: medRxiv preprint

Although the increased 197 COVID-19 positivity rates that were observed in Hispanics and in patients with public or no 198 insurance in our study could be in part due to the sociodemographic structure of the surrounding 199 areas, several studies have shown that minority racial groups and those belonging to lower 200 socioeconomic statures were at increased risk for serious and fatal COVID-19

We did not test children at random, 205 rather we tested children based on the presence of certain symptoms, which were pre-determined by 206 our testing algorithms at that time. Initially, these symptoms included only fever and respiratory 207 symptoms. Diarrhea and loss of taste/smell were added late in the study period, which likely explains 208 the observed lower prevalence of those symptoms in our study population. Second, we assessed 209 patient encounters that occurred during the early months of the pandemic. Data on COVID-19 210 symptoms were just beginning to emerge at that time, which may have played a role in symptom 211 documentation and triage. Third, our study was limited to hospitals that are situated in urban settings 212 in a single state

Fare Better than Adults

How COVID-19 Affects Children Compared to Adults

Clinical characteristics of COVID-19 in 273 children compared with adults in Shandong Province

Acute Heart 279 Failure in Multisystem Inflammatory Syndrome in Children in the Context of Global

CoV-2 Pandemic

285 COVID-19 associated Multisystem Inflammatory

Western New York approach

COVID-19 and multisystem 289 inflammatory syndrome in children and adolescents

COVID-19 recognition and 292 digital risk stratification

Development and validation of a risk 294 stratification model for screening suspected cases of COVID-119 in China

Clinical performance 297 of the GenMark Dx ePlex respiratory pathogen panels for upper and lower respiratory tract 298 infections

Comparison of the 303 analytical sensitivity of seven commonly used commercial SARS-CoV-2 automated 304 molecular assays

Comparing the 306 analytical performance of three SARS-CoV-2 molecular diagnostic assays

Precision Medicine 309 Analytics Platform (PMAP)

Epidemiology of COVID-19 in Children Aged 5-11 years. Epidemiology Task Force 315 CDC COVID-19 Public Health Response

Hospital are experiencing supply shortages for COVID tests

Differential COVID-19 case positivity in New York City 323 neighborhoods: Socioeconomic factors and mobility

326 Racial And Ethnic Inequities In Children's Neighborhoods: Evidence From The New Child

Johns Hopkins Medicine. The Johns Hopkins Hospital Fiscal Year

Association of 333 Race and Ethnicity with COVID-19 Test Positivity and Hospitalization Is Mediated by 334 Socioeconomic Factors

COVID in Babies and Kids: Symptoms and Prevention

348 Burden of Illness in Households With Severe Acute Respiratory Syndrome Coronavirus 2-349 Infected Children

Systematic 351 review of reviews of symptoms and signs of COVID-19 in children and adolescents. Archives 352 of Disease in Childhood

Molecular diagnostic assays 356 for COVID-19: an overview