Respiratory aspects of Covid-19 in children: what pediatricians need to know

This is an article prepared collectively by members of the Scientific Department of Pulmonology of Brazilian Society of Pediatrics (BSP) on the respiratory aspects of COVID-19 in childhood, considering the clinical and diagnostic peculiarities of this age group and discussing the imaging methods that can assist in this process. Pulmonary involvement in the disease is notorious and occurs in varying degrees of severity. Although less frequently than in adults, children and adolescents can also develop severe conditions. Imaging exams are part of the investigation of the patient with COVID-19, since they can assist in the initial diagnosis, in the assessment of the evolution and prognosis of the disease. The indications of chest X-rays and computed tomography (CT) and their most relevant characteristics are placed. In most studies, it is emphasized that the radiological findings in children are similar to those found in adults, but with less frequency, intensity and extension. Recently, however, authors who studied 34 children with COVID-19 in China reported that irregular, high-density opacities were common, while the ground-glass pattern, typical in adults, was rarely seen on CT’s scans. Chest X-rays is less sensitive to identify alterations, with CT being the best imaging test to visualize SARS-CoV-2 lesions, but it must be ordered with precise indications, as alone is not sufficient for diagnosis. Monitoring of cases through oximetry was also discussed. It is concluded after a vast literature review that the severity of the cases must be observed by clinical signs, imaging exams and oximetry, among others, always together.


INTRODUCTION
In December 2019, China informed the World Health Organization (WHO) that an outbreak of pneumonia of unknown origin occurred in the city of Wuhan, Hubei province. On January 7, 2020, a new type of Coronavirus was identified as the etiologic agent of these pneumonias 1,2 . On 02/11/2020, this disease was called COVID-19 (Coronavirus disease-2019) and on the same day the WHO named the SARS-CoV-2 virus. On 01/20/2020, the first case involving a child was reported, in the city of Shenzen, also in China 3 . Since then, the disease has spread throughout the world, and was classified as a pandemic by the WHO on 11/03/2020 2 .
SARS-CoV-2 is the seventh virus identified in the Coronavirus family, with a single-stranded RNA, being common to different species, including humans. It probably originated from bats due to the similarity with the virus that affects this species 4 .

Transmission
The most important form of virus transmission is through respiratory secretions (droplets and aerosols containing viruses inside), released through speech, breathing, coughing and sneezing. These secretions infect other people up to 2 meters away, and are the main route of spread of the disease 4,5 . Sick individuals are the main contaminants, but people who are asymptomatic or who are still within the incubation period are also potential contaminants 4,5 . Within this context, although children present milder symptoms than adults, contaminants and sources of disease spread are also possible, due to their nature of contamination and because they adhere to the respiratory etiquette 1,2 .
The most severe cases have a higher viral load and are significant sources of infection, hence the cause of the high number of healthcare professionals getting sick, making nosocomial transmission an important point of concern in the epidemiology of the disease 4 .
Another form of contamination is direct contact with secretions taken to the mouth, nose or eyes by contaminated hands that have had contact with surfaces containing the virus 4 . SARS-CoV-2 can survive for different periods on different surfaces, with contamination of some of them, such as door handles, sinks, protective equipment, bathroom towels, among others, being common. The virus survives longer on surfaces at lower temperatures. There is also the possibility of contamination by air vents, such as air conditioning devices 4 .
Vertical transmission has already been reported in the literature, but in small numbers, as case reports, especially during the last weeks of pregnancy, which promotes virulence in the newborn and possible neurological manifestations 6,7 . Yagnin 3 reported a series of 10 cases of mothers with positive RT-PCR for COVID-19, but in none of the newborns did the test become positive soon after birth. To date, there is no evidence of the presence of SARS-CoV-2 in breast milk 7 .
SARS-CoV-2 has already been found in the urine and feces of infected children, which is yet another epidemiological concern. It was isolated in feces weeks after diagnosis, suggesting contamination by this route for a prolonged period 1 .

Pathogenesis
After transmission, the virus is deposited along the respiratory tract 5 . It uses the Angiotensin-2 Converting Enzyme (ECA-2) to penetrate the cell where it can replicate 8 .
In the initial phase, the virus promotes local symptoms of the upper airways and general symptoms such as adynamia, myalgia and fever. This phase is contaminating and can end at this stage, leading to the end of the disease, which occurs in around 80% of those infected 4 . Seven days after this period, the pulmonary phase with infiltration and proliferation of the virus in the lungs can occur, causing pneumonia with vasodilation, increased endothelial permeability, leukocyte recruitment and lung injury with hypoxia. There may be cardiopulmonary stress associated 4 . In the third stage (inflammatory phase) there is systemic inflammation, called cytokine storm, with the protagonist Interleukin-6 (IL-6) activated by leukocytes, acting on a large number of cells, triggering this response. There is an increase in serum levels of ferritin, interleukins and C-reactive protein. This inflammatory process can affect other organs. Cardiac injury with myocarditis, cardiac muscle contraction deficit and liver injury with elevated transaminases is more frequently described 4 . Another impairment is disseminated vascular coagulation, with an increased risk of pulmonary thromboembolism or in other regions 4 .
Children have had milder symptoms and signs compared to adults, especially the elderly and adults with comorbidities, such as arterial hypertension, heart disease and diabetes 1,2 . The causes for these milder signs and symptoms are still unknown. Some hypotheses would be: children tend not to have dysregulation of the immune system as it occurs with adults, since they maintain normal lymphocyte counts (reduced by 3.5% against 70% of adults), as well as C-reactive protein dosage, normal D-dimer and liver function 4,8 . Another hypothesis is that children have lower expression of ACE-2, which makes it difficult for the virus to enter the cell cytoplasm 8 . Finally, they present significant production of antibodies against other viruses, which would somehow act against SARS-CoV-2 4 .
Clinical status: how to assess the respiratory disease?
Based on current data, children have less clinical severity in about 80% of reported cases. In those with comorbidities, however, the disease may have a severe course, progressing to severe acute respiratory syndrome (SARS) and multiple organ dysfunction. The main symptoms are similar to those from common viral diseases, frequently found among children attending schools or daycare centers 9 .
The clinical spectrum of COVID-19 in children varies from asymptomatic to severe acute respiratory distress (Table 1) 10 . Dong et al. 11 published a pediatric series involving 2,143 pediatric patients registered in the China Center for Disease Control and Prevention database: laboratory confirmed cases corresponded to 34%, and 66% were characterized as suspect. The median age (IQR) was 7 (2-13) years and 1,213 cases (57%) were boys. Among laboratory confirmed cases, the proportion of asymptomatic, mild, moderate, severe and critically ill infections was 12.9%, 43.1%, 41%, 2.5% and 0.4%, respectively. Children with a confirmed diagnosis had a milder clinical course than suspected cases (which may include other prevalent viruses, such as respiratory syncytial virus, influenza and para-influenza etc. in this age group), indicating that the severity of the disease by SARS-CoV-2 may be milder than other acute respiratory infections in this age group. The average time from the beginning of the disease until the diagnosis was 2 days (range: 0 to 42 days). The proportion of "serious and critical" cases was 10.6%, 7.3%, 4.2%, 4.1% and 3.0% for the age group of "1, 1 to 5, 6 to 10 , 11 to 15 and > 15 years, respectively, indicating that young children, especially babies, were more vulnerable to severe SARS-CoV-2 infection; one child (14 years old) died. This study did not describe the frequency of individual symptoms in its population 11 .
According to another series of pediatric cases, with 171 patients, with a median age of 6.7 years (range from 1 day to 15 years), admitted to a hospital in Wuhan, China, all patients tested positive for COVID-19; there were 27 (15.8%) asymptomatic patients; 33 (19.3%) with upper airway symptoms and 111 (64.9%) with pneumonia. Seventy-one patients had fever (41.5%), lasting from 1 to 16 days (median, 3 days). Three patients were admitted to the intensive care unit: all with comorbidities (hydronephrosis, leukemia-during chemotherapy and intussusception). The intussusception patient was 10 months old and died 12 . Several skin rashes have been recently seen in some pediatric cases with variable clinical presentations 10 . Table 1 shows standards of COVID-19 clinical presentations, according to severity.
In a systematic review involving 38 studies (1,124 cases); the authors described the main clinical, laboratory and radiological characteristics of children infected with SARS-CoV2. Of all cases, 1,117 had their severity classified: 14.2% were asymptomatic, 36.3% were mild, 46% were moderate, 2.1% were severe and 1.2% were critical. The most prevalent symptom was fever (47.5%), followed by cough (41.5%), nasal symptoms (11.2%), diarrhea (8.1%) and nausea/vomiting (7.1%). One hundred and forty-five (36.9%) children were diagnosed with pneumonia, and 43 (10.9%) had infections of the upper airways; the authors concluded that the clinical manifestations of children with COVID-19 differ widely from the cases recorded in adults. Fever and respiratory symptoms should not be considered a registered trademark of COVID-19 in children. The distribution of the clinical manifestations of children with COVID-19 in the selected studies is shown in table 2 1 .
Patients with more severe clinical manifestations develop hypoxemia and poor perfusion, usually by the end of the first week. Complications commonly described are severe acute respiratory syndrome (SARS), myocarditis, septic shock, disseminated intravascular coagulation, acute kidney injury and liver dysfunction. Increased concentrations of procalcitonin, CRP and IL-10 and decreased IgA level and percentage of CD4 + CD25 + T lymphocytes have been associated with pneumonia in children with COVID-19 13 .
The main clinical manifestations of children with COVID-19 described in selected studies are contained in Table 2.

How to manage the children with respiratory disease
All children with severe acute respiratory syndrome (SARS) should undergo RT-PCR for SARS-CoV-2. However, in many services, children with mild illness without a history of contact are not tested for SARS-CoV-2. Due to the high percentage of acute respiratory tract infections due to other etiologies, such as acute viral bronchiolitis or severe acute asthma, many patients may meet the case definition of Severe Acute Respiratory Syndrome (SARS), being admitted to a

Asymptomatic infection
No clinical signs and symptoms of the disease, normal chest x-ray or chest CT scan, associated with a positive test for SARS-CoV2.

Mild infection
Symptoms of upper airway involvement, such as fever, cough, sore throat, rhinorrhea and sneezes, besides myalgia and fatigue. Normal tests of the respiratory system. Some cases may course without fever and other gastrointestinal symptoms, such as vomits, nausea, abdominal pain and diarrhea.

Moderate infection
Clinical signs of pneumonia. Persistent fever, dry cough in the beginning, and productive later; there may be lung crackling and wheezing upon respiratory auscultation, but in this phase, without respiratory distress. Some patients may not show clinical signs or symptoms, but the chest CT scan may show typical pulmonary lesions.

Severe infection
The initial respiratory symptoms may be associated with gastrointestinal symptoms, such as diarrhea. Clinical deterioration usually happens within one week, with the patient developing dyspnea and hypoxemia (SpO2 < 94%).

Critical infection
Patients may soon deteriorate to an acute respiratory distress syndrome or respiratory failure, and may develop shock, encephalopathy, myocardial damage or heart failure, coagulopathy, acute kidney damage and multiple organ dysfunction.
Adapted from: Carlotti et al. 38 hospital. In view of the current probability of children being exposed to the new virus, although with a lower likelihood of infection than adults, it is appropriate to examine them in separate places from other children suspected of having COVID-19. Thus, those with suspected SARS-CoV-2 infection awaiting laboratory results should be admitted separately from adults and other children. Parents should stay with isolated children and receive appropriate PPE (s). A quick response time for diagnostic test results will also reduce the risk of exposure. Official guidelines currently recommend hospital admission for confirmed cases, particularly those patients classified as having "Severe Pneumonia" and "in critical condition" for greater care ( Figure 1). The following criteria can be particularly considered for admission (any of the following criteria): 1. Respiratory discomfort (tachypnea, correlate with age group) 2. SpO2 < 92% in room air 3. Shock/poor peripheral perfusion 4. Poor oral intake, especially in babies and young children 5. Lethargy, especially in babies and young children 6. Seizures/encephalopathy To date, we do not have specific guidelines for children with underlying diseases, such as chronic respiratory diseases, immunosuppression, uncorrected heart disease, chronic kidney disease, etc. This group needs more intensive monitoring and early therapy 14 .

Differential Diagnosis
The differential diagnosis of COVID-19 in children is complex, since many pediatric diseases have similar signs and symptoms.
Cough and fever of mild to moderate intensity were the symptoms most frequently reported in a meta-analysis involving 551 children with positive tests for SARS-CoV-2. 15 Thus, in general, a differential diagnosis of COVID-19 is considered for all respiratory conditions associated with acute infections, from those with isolated upper airway involvement (differential with mild cases of COVID-19), to pneumonia (differential with moderate cases of COVID-19), to Severe Acute Respiratory Syndrome SARS (differential with severe cases of COVID-19) and severe sepsis from other etiologies, such as bacterial sepsis, staphylococcal or streptococcal toxic shock syndrome, as well as severe infections that develop with myocarditis, such as those caused by other enteroviruses (differential with critical cases of COVID -19). 11,16 Various respiratory viruses, common bacteria and atypical bacteria, present with clinical syndromes similar to those caused by SARS-COV-2, according to Table 1. 17,18 Other infectious diseases can also present with similar symptoms especially in the presence of fever and systemic symptoms such as those listed on Chart 2. 19 More recently, we know that children of different nationalities presented a multisystemic inflammatory syndrome with clinical manifestations and changes in complementary exams similar to those found in children and adolescents with Kawasaki syndrome, incomplete Kawasaki and/or toxic shock syndrome, opening a new front in differential diagnosis. These children had high and persistent fever (38-40°C), rash of different presentations, non-purulent conjunctivitis, hands and feet edema, abdominal pain, vomiting and diarrhea. The vast majority evolved to shock (with arterial hypotension and tachycardia), mainly cardiogenic, with elevation of myocardial enzymes, requiring vasoactive drugs for hemodynamic stabilization. Many had inflammation of the serosa, with pleural, pericardial effusion and ascites. Almost all of them needed ventilatory support, even though respiratory manifestations are not relevant. Some cases are so severe that they become a differential diagnosis of familial hemophagocytic lymphohistiocytosis syndrome or macrophage activation syndrome, seen in children with rheumatological diseases. 20,21 Today, considering the moment we are living through, the investigation of epidemiological data, laboratory and image exams and, especially, the search for the virus identification, are essential for the confirmation of the disease.  Low fever, paroxystic cough in spells followed by vomit and perioral cyanosis, excess saliva.
There may be cyanosis and apnea.

RSV Influenza A and B Parainfluenza 1 and 3 Coronavírus (SARS-CoV, SARS-CoV2, MERS-CoV) Common and atypical bacteria
Fever, cough, tachypnea, chest pain, desaturation. There may be systemic involvement with prostration and signs of toxemia.

Hantavirus
High fever, myalgia, headache, the first week courses with acute respiratory failure.

Chart 2.
Other diseases in the differential diagnosis of COVID-19. Faced with a child with nonspecific symptoms, the critical and systematic approach is prudent, categorizing the cases initially as suspect (with clinical and/or epidemiological and/or radiological criteria). During follow-up and investigation, the cases will be classified as confirmed, of high or low probability according to the suggestion presented in Chart 3. 22

Risk factors for severe COVID-19 in children
Descriptive, observational studies reported the presence of certain pre-existing conditions in children who developed severe conditions and fatal outcomes in SARS-CoV2 infection, pointing to a tendency for certain underlying diseases to act as risk factors. Dong et al. 11 reported that younger children, particularly younger than one year, were more vulnerable to severe conditions. Lu et al. 12 , during an observation period, reported three children who needed ventilatory support and all had pre-conditions. (hydronephrosis, leukemia under chemotherapy and intussusception).
Likewise, She et al. 23 , described the only two cases of critically ill patients who had a history of underlying disease (congenital heart disease with malnutrition and bilateral hydronephrosis with lithiasis).
Another study described two cases of severe SARS-CoV2 pneumonia that occurred in one student in remission of acute lymphoblastic leukemia, and another in an obese adolescent. The same study draws attention to the careful assessment of risk factors, since children with COVID-19 without severity sometimes had the same status of comorbidities (severe group and the non-severe group (P = 1.00). 24 Like the aforementioned studies, underlying conditions such as congenital heart disease, bronchial pulmonary hypoplasia, abnormalities of the respiratory tract, abnormal hemoglobin level, severe malnutrition, primary immune deficiency or the prolonged use of immunosuppressants, seem to be criteria for more severe disease in children. 25 From the point of view of complementary assessment, the impairment of more than three lung segments was associated with a higher risk of severity (odds ratio = 25.0, p = 0.006). Elevations in IL-6levels, high total bilirubin and D-dimer can also help identify patients with potential severity early on 24 .
The characteristics of the risk groups for severe COVID-19 in children are not yet clearly defined. So far, the literature suggests certain underlying diseases and some laboratory and imaging findings, but a longer observation period and a larger group of children may, in the near future, accurately define this group of special interest.

Radiological findings
Imaging exams are part of the investigation of patients with COVID-19, since they can assist in the initial diagnosis, in addition to contributing to the assessment of the disease's evolution and prognosis. 2,26 The exam considered the gold standard for diagnosis, the identification of viral RNA by Polymerase Chain Reaction by Reverse Transcriptase (RT-PCR), can present false-negative results in around 30% of cases, depending on the quality of the sample collection and laboratory logistics. 26 The literature is poor in relation to the assessment of radiological findings in children. Between 1 and 2% of disease notifications occur in patients younger than 18 years of age. Most of these patients do not need to undergo imaging tests, since only 1 to 4% develop more severe conditions. 2 Chest radiography is easy to perform, inexpensive and carries little radiation, but it is less accurate than chest tomography (CT) to show changes resulting from COVID-19, both in children and in adults. 2,27 The most frequently found findings on chest x-ray are: 2 -Peripheral ground glass pattern in lower regions -Irregular bilateral consolidations CT provides a better visualization of unobservable changes in the x-ray. Its use must be judicious, especially in children, due to the higher dose of radiation, in addition to the higher cost and not being so accessible in smaller centers. The main changes found are: 1,2,4,27,28 -Peripheral and irregular ground glass pattern -Peripheral consolidations -Inverted halo signal -Mosaic perfusion -Irregular and discreet opacity These are preferably located in the subpleural region, in the lower lobes, and bilaterally, affecting more than one lobe. (Figures 3,4,5

and 6)
In a study involving 20 SARS-CoV-2 positive RT-PCR children with and a mean age of 2 years and 1 month, Xia et al. 28 reported that the x-ray did not detect lung injuries and such injuries were more visible on CT. Of the latter, 20% were normal; among the altered CTs, 60% had ground-glass opacity, 50% consolidations with an inverted halo sign, irregular opacity in 20%, and small nodules in 3%. There was no pleural effusion or lymphadenomegaly.
Steinberger et al 29 evaluated 30 children with SARS-CoV-2, positive RT-PCR, and a mean age of 10 years, from 6 different centers. Nine patients were asymptomatic, all with normal CT. Of the total, 23/30 (77%) had normal CT and only 7 (23%) had changes: 86% with ground-glass pattern, 14% with ground-glass + consolidations, 29% with mosaic perfusion and 29% with signal inverted halo. The changes were more frequent in patients over 14 years of age, with 71% in more than one lobe, 71% bilaterally and 86% in the periphery. There was no pleural effusion or lymphadenomegaly.
In a study with 171 children with confirmed SARS-COV2, Lu et al. 12 found 32.7% with ground-glass pattern on CT; 18.7% with local opacity and 12.3% with bilateral opacity; 15.8% had no radiological changes. Twelve patients (9.8%) had radiological changes, but were asymptomatic.
Wang et al. 30 reported abnormalities in 66% of CT scans of children with COVID-19, with the ground-glass pattern in more than one lobe, the most frequent alteration in 35% of cases. Carlotti et al. 10 showed that the most severe cases had bilateral condensations in posterior basal regions.
Zhang et al. 31 studied 34 children with COVID-19 in 4 hospitals in China and reported that irregular, high-density opacities were common while the ground-glass pattern was rarely seen on CT.
The image alterations in COVID 19, described in papers involving pediatric patients, would be more evident from the fourth day of the disease, both on X-Ray and CT. The evolution of the images is described in Figure 7. 32 In summary, the radiological findings in children are similar to those found in adults, but at a lower frequency, intensity and extension. The X-ray is less sensitive to identify changes, and the CT scan is the best imaging test to show SARS--CoV-2 lesions, but it must be ordered with precise indications, as alone is not sufficient for diagnosis.
Other viral infections may show similar images and there may be co-infections. There is an international concern    regarding the excessive use of CT in children, as it may cause future damage by ionizing radiation (low radiation devices are lacking). In addition, it may contaminate healthcare professionals and other patients when using the equipment.

When to order a CT scan?
The discussion about the role of radiological examinations in patients with COVID 19 in the pediatric population has great use and repercussions in clinical practice. Some questions intrigue physicians who work with children with COVID 19, namely: i) When should imaging tests be performed in the initial diagnosis? Ii) How should the disease progression be assessed? and iii) Is there a radiological pattern with prognostic value. 2,33,34 In adults, chest tomography has been performed early in the initial approach, and as a form of diagnostic screening. This is because laboratory diagnostic tests require more time for its result and, in addition, some RT-PCR results -the gold standard for the disease -may be false negative, due to the time of disease or the collection technique. It is important to stress that the Radiology Society of North America (SNRC) and the Brazilian College of Radiology do not corroborate this practice. 2 At the beginning of the pandemic, with the general fear of the medical community about how the disease would develop in the pediatric population, we believed that chest computed tomography (CT) should be performed early for the initial diagnosis in children. However, with the publication of a series of papers showing the most favorable evolution in children, a common sense was reached that chest CT can be performed only in selected cases. Furthermore, the initial changes may be similar to those of other respiratory viruses, therefore uncharacteristic. In pediatric practice, a CT scan is indicated for patients who progress to more severe conditions and hospitalization. 3,29 Thus, as in the adult population, chest radiography has less specificity than CT, which should be indicated for those patients with unsatisfactory clinical evolution or in groups at risk. However, it is important to note that, whenever possible, it should be performed under low radiation protocols. 44 In general, CT in adults shows multifocal peripheral opacities, ground-glass opacities and predominant distribution in the lower and posterior lobes, bearing a bilateral presentation. In the researched literature, tomographic findings in the pediatric age group are more discreet, with fewer lung lobes involved and even with unilateral presentations. The main tomographic findings found in children infected with SARS-CoV-2 range from the presence of multiple irregular bilateral "ground-glass" opacities, sparse and irregular "ground-glass" opacities, and/or infiltrating the middle third or periphery of the lung or subpleural space. 41 There are descriptions of images of subpleural infiltrates, condensations in uni or bilateral halos. Upon tomographic examination, the progression of the disease determines image increase and bilaterality, with high-density condensations and peribronchial thickening, affecting both lungs. Pleural effusion is not described. In the remission phase of the disease, that is, from the 14th day on, there is still some halo images and parenchymal bands that can persist for months. 44 It is important to note that 30% of patients may have negative CT scans, and a normal CT does not rule out the diagnosis. In addition, CT is not indicated in asymptomatic patients. 3 In adults, CT should be performed in mild, moderate and severe symptomatic patients with risk factors for disease progression such as the elderly, diabetics, hypertension, in the search for complications and to rule out alternative diagnoses. Likewise, in patients who worsen their respiratory condition. It is important to note that in general it is not indicated in mildly symptomatic patients without risk factors. In the pediatric population, this criterion of greater attention for the risk group can be followed and, evidently, in patients with clinical deterioration. 34 Other situations in which CT scans have been ordered in Brazil are patients residing in cities without access to laboratory tests, or awaiting examination, or with nasal swab examination with negative RT-PCR, but with strong clinical suspicion, if moderately and severely symptomatic. 3 One of the classifications that has been used to quantify the extent of the disease is based on the percentage of the affected lung: i) mild involvement: < 25%; ii) moderate involvement: 25 to 50%; iii) marked involvement : > 50%. 2 The main tomographic findings are ground-glass opacity, reticular opacities, consolidations and inverted halo sign. The findings can be classified into 4 categories 2 : a. Findings compatible with a viral infection (

Considerations about imaging tests and oximetry:
is there already a severity score for children?
Children usually have mild cases of COVID-19 and chest radiography is not able to identify lesions or details. Chest computed tomography (CT) is more sensitive for identifying lesions, but care should be taken not to unnecessarily expose children to radiation. Thus, the question remains whether the x-ray or the measurement of hemoglobin saturation by oximetry can be useful in the evaluation of children with COVID-19, considering that the findings in children are different from those in adults. 3 There are few studies in children with COVID-19, and studies in adults are referred with the objective of verifying whether it is possible to establish an association between clinical picture, image and oximetry variables.
Steinberger et al. 29 evaluated 30 patients aged 10 months to 18 years (median = 10 years), immunocompetent and without other morbidities. Of these, 9/30 (30%) were asymptomatic. Fourteen patients out of 23 (61%) had normal CT and had at least one symptom. All 7 patients with abnormal CT (ground-glass opacities and/or consolidation were symptomatic. The most frequent clinical findings were fever > 38 C (53%) and cough (27%). No patient required oxygen, intubation or ICU admission. The number of affected lobes and the degree of impairment of each lobe was assessed. The severity of the impairment was classified according to the findings from Chang et al 48 , namely: without lobe impairment, score zero; minimum (1 to 25%) score 1; mild 26 to 50% score 2; moderate 51 to 75%, score 3; and severe 76 to 100% score 4. Severity was assessed by the sum of the scores of the 5 evaluated lobes; 5 lobes, each lobe from zero to 4 ; score variation (0 to 20); 2 patients had involvement of 1 lobe; 3 patients with 2 lobes and 2 patients with 3 and 4 pulmonary lobes involved.
The findings of this study (TABLE 3) show us that there is no relationship between tomographic findings and arterial desaturation, since there is no report that patients with lung injuries required O2, that is, they did not desaturated.
Mussolino et al. 36 consecutively evaluated 10 children (median age 11 years) admitted with COVID-19. Ultrasonography was performed in all of them and 1 patient underwent CT with similar findings to those of Steinberger et al. 29 All patients were symptomatic upon admission: fever (80%); cough (50%) and diarrhea (20%) and had pulmonary involvement (areas of the lung with subpleural consolidation and pleural irregularities. There were no reports of dyspnea, desaturation and need for oxygen. That is, the initial clinical condition (fever and cough), the findings of lung lesions (consolidation) in the initial phase in children do not correlate with desaturation and dyspnea. admission was 98%, but 6 (17%) of the children with moderate disease needed oxygen. All progressed well, the average length of stay was 14 days, although of these n = 10 (28%) were asymptomatic, 7 (19%) had upper airway manifestations and one patient had dyspnea. A striking feature of COVID-19 in these patients is the involvement of vital organs such as the lungs and heart (alteration of myocardial enzymes), even in patients with mild and moderate disease (31%). Although the study did not evaluate criteria for CT severity and relationship with oximetry, it is possible to observe that patients with mild to moderate clinical disease present radiological changes, elevation of cardiac enzymes, with normal oxygen saturation, but who progressively needed O2. This allows us to state that, in children with mild to moderate clinical disease, it is not possible to establish a severity score, or even an association between image and transcutaneous measurement of hemoglobin saturation, and despite myocardial changes, all presented with good progress.
Severity markers of lung injury that use partial oxygen saturation (SpO2), are suitable substitutes for those who use blood pressure (PaO2) in children with respiratory failure with SpO2 between 80% and 97%. Both should be used in clinical practice to characterize risks, increase participation in clinical trials and check for disease prevalence 39 .
Pulse oximetry and pulmonary ultrasound can be useful tools to track or rule out low oxygenation or pulmonary changes consistent with severe acute respiratory syndrome (SARS) in places with few resources, where arterial blood gases and chest x-ray are not available 40 .
The study (CONFIDENCE -The Coronavirus Infection in Pediatric Emergency Departments study) evaluated a cohort of 100 Italian children under 18 with Covid-19, confirmed by nasopharyngeal swab PCR. The median age was 3.3 years and the origin of the infection outside the home occurred in 55% of the cases. The most frequent signs were: sick aspect (12%), fever ≥ 37.6 C (54%), cough 44% and difficulty or refusal to eat in 23%. There were 4% of the children with Sp02 < 95%, assessed by pulse oximetry. All of them presented an image of pulmonary involvement (x-ray or pulmonary ultrasound); 9 needed oxygen during hospitalization due to desaturation or respiratory distress. Eleven patients with pulmonary impairment (chest x-ray or ultrasound) had normal oxygenation (pulse oximetry), probably due to lesions in the initial stage. No patient got worse. Nine patients were admitted to the ICU (4 neonates, 3 infants < 3m, others due to clinical condition or comorbidities). Of these, one required mechanical ventilation (patient with encephalopathy and epilepsy). This study demonstrates in children that the indication for hospital admission was dependent on the clinical condition, other morbidities and not on image evaluation. During admission, 4 children had SpO2 < 95%, but nine needed oxygen support. The indication for the use of oxygen was not based on the imaging results, as 11 patients had changes in radiography or ultrasound and did not require oxygen. It was not possible to Qiu et al 32 evaluated children with COVID-19 (n = 36), with mean age 8 ± 3.5. Severity was classified as mild or asymptomatic (n = 17) and moderate (n = 19) according to the criteria by Chen et al 38 . Ground-glass opacity findings were seen in 53% and 100% of patients with mild and moderate disease, respectively. The arterial saturation measure upon establish an association between clinical severity, saturation levels and radiological changes in the early stages of COVID-19 in children 27 .
Pulse oximeters have accuracy of ± 1-2% when SpO2 > 75-80% and there is no accuracy reported for values below 70%, because ethically there is no way to test it. This problem can be mitigated when monitoring patients with COVID-19 by setting saturation values above 75-80%, although it must be recognized that accuracy information is generally not reported, regardless of the degree of hypoxemia for pulse oximeters that are cheap and easy to purchase. Although there is knowledge about the accuracy and possible failures that may interfere with the values obtained, the ease of use, low cost of this equipment, associated with the burden of COVID-19 and the risks of silent hypoxemia, make it a reasonable solution for monitoring individuals at risk 41 .
Lipnick et al 42 compared 6 low-cost oximeters in 22 healthy individuals to assess measures of arterial saturation with a range of 70 to 100%. Of the 6 devices tested, only 2 achieved accuracy criteria established by the International Organization for Standardization accuracy criterion (Accuracy root mean square-Arms < 3%). Four of the oximeters showed Arms values > 3.0% and, in three of them, when the saturation was 80 to 90%. This value was > 5% on four devices when the saturation values are 70-80%. This may partly justify the difficulty of obtaining an association between variables (for example, CT versus pulse oximetry), the lack of accuracy of the device used.
The decision to institute mechanical ventilation is made according to the physician's clinical judgment, in addition to SpO2, dyspnea, respiratory rate, chest x-ray and other factors. Many patients with COVID-19 are intubated due to hypoxemia; still, with little dyspnea or distress 43 .
In order to verify whether the chest x-ray score correlates with hospitalization, intubation, length of stay and death, adults (n = 338) between 21 and 50 years of age were evaluated with COVID-19. The chest -ray score was obtained by dividing each hemithorax into 3 parts and the presence of opacities (score = 1 if present in one area). In the initial assessment, radiography with a score ≥ 2 was an independent predictor of hospital admission (n = 145), and a score ≥ 3 (n = 28) for intubation. The authors concluded that for patients with COIVID-19, between 21 and 50 years old, the chest x-ray score is an independent and predictive factor of severity for hospital admission and intubation. These findings probably occur because despite the low sensitivity of chest x-rays, it is altered in the most severe cases. 44 Yang et al. 45 , evaluated 102 patients with COVID-19, aged 15-79 years, 84 with mild disease.
The CT score was adapted from Chang et al. 35 and obtained by dividing 20 regions; each region received 1 or 2 points if the involvement was < 50% or ≥ 50%, respectively; that is, the score could vary from (0 to 40). The patients were divided into 2 groups (mild and severe); severe = RR ≥ 30; O 2 saturation ≤ 93%; PaO 2 / FiO 2 ≤ 300 mmHg and or need for mechanical ventilation, shock, failure of another organ.
The best CT score value capable of discriminating between mild and severe patients was 19.5, with 83.3% sensitivity and 94% specificity. In patients with mild disease, O 2 saturation was 97% (96-98) and 92% (88 to 93) for critically ill patients. It was possible to state that patients with more severe disease had a higher CT score and lower levels of oximetry. However, the SpO 2 variable has an assessment bias since it is an inclusion criterion for severe disease.
In order to check whether CT findings may be related to the evolution of patients with COVID-19, 380 patients with a mean age of 53.62 ± 16.66 years were evaluated in a cross-sectional study (66.1% male). The most frequent CT findings (low radiation dose and 4mm slices) were peripheral changes (86.6%) and peribroncovascular interstitium (34.6%) ground-glass (54.1%). The CT score was assessed according to Pan et al 46 .
In another study, the 5 pulmonary lobes were evaluated and scored from 1 to 5 according to their involvement; zero = without involvement; 1 = < 5%; 2 = from 5 to 25%; 3 = 26 to 49%; 4 = 50 to 75% and 5 = > 75% of involvement. There was a correlation between the mean score of CT severity and mortality (13.68 versus 8.72) p < 0.0001). SpO 2 levels were higher for patients who survived: 93.82 ± 5.88 versus 87.13 ± 6.72, for those who died (p = 0.002). Interestingly, when evaluating blood gas values, there was no difference for PO 2 and PCO 2 . This puts in check the hemoglobin saturation values found and the difficulty of establishing the relationship between oximetry and CT scores. There is no description on how to measure the saturation levels or which device was used, challenging the accuracy of the device. 47 To check whether the quantitative CT image can determine clinical severity, Li et al. 48 evaluated patients with COVID -19 who were divided into 3 groups according to Chinese guidelines: mild (few symptoms and normal CT); moderate (respiratory symptoms with pneumonia) and severe-critical (RR ≥ 30; O 2 saturation ≤ 93%; PaO 2 /FiO 2 ≤ 300 mmHg) -(respiratory failure requiring mechanical ventilation, shock or failure of other organs requiring ICU). Tomographic findings were evaluated according to the score by Chung et al. 49 and compared with the clinical classification. There were 78 patients included, 40 females. They were classified as mild: 24 (30.8%); moderate: 46 (59%) and severe-critical 8 (10.2%). The median CT score was higher in the severe-critical group (10), with variation (8-18) when compared to the moderate group (5) with variation (5 to 11) with p < 0.001. However, 32/46 (70%) of moderately-ill patients, with SpO2 ≥ 94%, had CT showing involvement of more than 2 pulmonary lobes, and 37/46 (80.4%) had the involvement of 2 lungs. Severelycritically ill patients (n = 8), who by criterion had SpO2 ≤ 93%, had the same image involvement. It was not possible to establish a relationship between the number of pulmonary lobes involved and the SpO 2 level. There is a need to analyze not only the number but also the degree of involvement of each lobe.

FINAL REMARKS
a. There are few COVID-19 studies in children, and with a small number of patients evaluated with CT; thus, it limits our capacity to establish score parameters that may be associated with oximetry and CT or chest x-ray and clinical severity.
b. There is no standardized CT severity score in children with COVID-19. The studies use adult scores, which in turn also have limitations, as they were applied to patients with SARS after discharge and not upon hospital admission.
c. Chest x-rays have limitations concerning their use in children due to a low sensitivity and the fact that children have milder disease. If it is necessary to use CT, it should be performed with low radiation dosages. In adults, it is a marker for hospitalization and intubation.
d. Oximeters should be used, due to their low cost and ease of use, but it is important to know their limitations, difficulty of use in young children, lack of accuracy and the need for proper use.
e. Severity assessment should be based on the following information: clinical signs, oximetry, image and not on isolated variables, even because it is a dynamic situation.