This systematic review fully assesses epidemiological and clinical characteristics of SARS-CoV-2 infections reported in pediatric age.
Children are less likely to develop severe symptoms of COVID-19 than adults, with 95% of all cases ranging from asymptomatic to mild-moderate clinical patterns, as described by different case series[10, 11, 24, 27–56, 67, 68]. Moreover, less than 2% of patients were admitted to PICU or required MV[10, 25, 26]. Overall, three deaths were reported (mortality rate 0.07%): all patients developed complications[10, 16], including a preterm newborn who died from sepsis[59]. In adulthood, over two-third of those who died from COVID-19 had a comorbidity[73], while a very limited number of children with underlying disease could be identified, and none of them showed worse clinical course of the infection in comparison to previously healthy patients[10, 24, 26, 55, 56].
However, children might not be tested for SARS-CoV-2 as frequently as adults. Moreover, the current gold standard for the diagnosis of SARS-CoV-2 infection is Real Time-Polymerase Chain Reaction (RT-PCR) on respiratory tract specimen. The diagnostic accuracy of RT-PCR highly depends on the virus-specific diagnostic window, and the analytical sensitivity of this assay is potentially plagued by false SARS-CoV-2 negativity attributable to the low viral loads, especially in asymptomatic or mild symptomatic patients that might transmit the disease as well[74].
Clinical features
Infected children usually show typical symptoms of acute respiratory infections including fever (49.2%) and cough (44.1%)[10, 11, 16, 25–29, 31–33, 35–37, 41, 45, 47, 49, 53–56]. In particular, some authors reported that up to one-third of symptomatic children may have high fever[10], but generally below 39°C[32]. Differently from adults, children are more likely to present with extrarespiratory symptoms[75]: diarrhea (8.9%) and vomiting (5.5%) are the most frequently reported ones. It has been showed that, when present, gastrointestinal symptoms usually anticipate the typical respiratory pattern[76]. Previous studies on SARS-CoV cases demonstrated the viral detection in gut biopsy specimens and stool of recovered patients, indicating a possible gastrointestinal tract tropism that may partially provide explanations for extrarespiratory symptoms and persistent viral shedding through fecal-oral route[77]. There is growing evidence that this mechanism of excretion may be typical also for SARS-CoV-2[2, 31]. As described in a case series of ten infected children, SARS-CoV-2 remained detectable in rectal swabs after nasopharyngeal swabs turned negative[31]. However, the extrapulmonary detection of viral RNA does not mean infectious virus is present, but two independent laboratories from China declared that they have successfully isolated live 2019-nCoV from the stool of patients (unpublished)[76]. Moreover, given the fact that pathogenesis of human coronavirus mainly depends on the interactions between its transmembrane spike glycoprotein (S-protein) and specific cell receptors of angiotensin converting enzyme II (ACE2)[78], recent analysis revealed that ACE2 was expressed also in upper esophagus and absorptive enterocytes from ileum and colon[55].
According to evaluated studies, clinical presentation in newborns and infants below 3 months could be slightly different than in older children, with a higher proportion of them (17%) presenting with a severe pattern[59, 60]. Even if the vertical transmission for SARS-CoV-2 has not been demonstrated[19, 79], in 80% of neonatal cases the mother was infected[57–60, 63, 66]. Moreover, nosocomial infection may also occur, and strict measures to reduce this risk should be always observed[20]. However, some authors hypothesized that 2019-nCoV infection and morbidity in newborns may be related to possible hypoxemia in the infected mother that increases the risk of perinatal adverse events such as birth asphyxia and premature birth[59, 80]. Evidence remains still too limited.
Laboratory findings
Twenty-nine studies accounting for a total of 474 patients reported information on blood investigations in pediatric cases with COVID-19. Overall, no significant abnormalities were observed and this was consistent with the results of a previously published review including a total sample size of 66 children with confirmed SARS-CoV-2 infection[71]. In particular, full blood cell count was normal in most patients. Two case series reported high rates of lymphopenia (10/25, 40%[26] and 11/36, 30.1%[54] respectively). However, this finding seems to be in contrast with adult data, as low lymphocyte count has been noted in up to 80% of infected critically-ill subjects[74, 81]. The limited number of severe COVID-19 infection may in part explain the lack of significant lymphopenia in children.
Our results suggested that inflammatory indexes may be abnormal in more than one-third of children with SARS-CoV-2 infection (38.7%), while Henry et al. described only 10-13% of cases with high CRP and/or PCT[71]. This controversial finding could be explained by the high heterogeneity in defining a cut-off of abnormal values across all the included studies. However, in adults a PCT value of ≥0.5 ng/ml was observed to be associated with a near 5-fold increase in risk of severe clinical course of COVID-19 infection[82].
Other significant reported laboratory investigations were represented by high CPK values (13.1%) and liver enzymes (12.9%). These enzymes could be often altered during viral infections[83]. In particular, high CPK levels or aspartate aminotransferase activity correlated with more severe clinical patterns in adult patients with SARS-CoV-2 infection[84]. Abnormal transaminases levels may also express a sign of direct liver damage. Recently published data demonstrated enrichment of ACE2 expression in cholangiocytes (59.7% of cells) suggesting that SARS-CoV-2 might lead to direct damage of intrahepatic bile ducts[85].
Radiology findings
A total of 493 children had radiological exams, with half of them showing abnormalities. The sensitivity of chest X-ray is supposed to be inferior to that of CT-scan: in adults, nonspecific patchy peripheral and peribronchovascular opacities may be shown in all lung zones, according to the severity of the infection[86]. As children usually develop milder patterns of the disease, chest X-ray may fail to identify typical lesions, and it is mainly adopted in newborns and younger infants[59, 63, 66]. CT-scan is frequently performed in children with suspected or confirmed SARS-CoV-2 infection (up to 85% of all cases reported). The most frequently recognizable lesions are represented by GGO, with unilateral or bilateral distribution [27–29].However, it should be noted that CT-scan was performed also in asymptomatic patients[10, 27–30, 38, 39, 49, 57], and that more than one third of all patients who underwent this exam resulted completely normal. The use of CT-scan as a diagnostic screening tool in confirmed or suspected COVID-19 patients is supported by recent evidence showing that its sensitivity could be greater than that of real time-PCR in detecting the virus (98% vs. 71%, respectively)[87]. However, routine use of CT-scan has several obvious implications, in particular in a pediatric setting where concern about unnecessary exposure to radiation source should be raised. Therefore, other possible diagnostic imaging tools may be used. Lung ultrasound has been successfully adopted in adult subjects with SARS-CoV-2 infection[88], but, to date, there is no data in children.
Treatment
Currently, there is no ongoing drug trial specifically aimed at children. Symptomatic treatment alone was used in most cases, in particular in newborns[59].
The only therapeutic recommendation in pediatric age is to use nebulized IFN and oral antiviral agents (i.e., lopinavir/ritonavir), with CCS for complications (acute respiratory distress syndrome, encephalitis, hemophagocytic syndrome or septic shock) and IVIg for severe cases[22]. However, none of these therapies have shown a clear benefit in the treatment of SARS-CoV-2.
The main limitation of our review is represented by the difficulty to retrieve the full text of some Chinese studies, and we had to rely on English-language summaries, or publications that referenced papers published in Chinese. Moreover, our findings are essentially based on limited case-series and case reports, therefore laboratory parameters of interest were not consistently reported and reference ranges were not always clearly defined. Similarly, radiological patterns were difficult to compare, as the description of the included cases was not standardized.
However, despite the scarce and extremely heterogeneous evidence on pediatric patients with COVID-19, the use of systematic databases allowed us to review at a glance and interpret the majority of published studies up to 7 April 2020.