COVID-19 in 17 Italian Pediatric Emergency Departments

Discussion

In this study, we have presented our experience with the diagnosis and treatment of 170 children with COVID-19, from the perspective of 17 EDs in Italy. With the data presented, we confirm what we reported in a preliminary report⁷ and add new knowledge on the possible late presentation of MIS-C. Even if rare, this condition might be serious and requires prompt recognition, which may not be immediately feasible because of the early presentation of patients (before all symptoms develop) or the presence of different features of the MIS-C (eg, toxic shock syndrome, secondary hemophagocytic lymphohistiocytosis, or macrophage activation syndrome).¹⁰

Our results reveal differences compared with 3 recent Chinese cohorts reported in the literature.^2–4 The higher number of patients aged <1 year is the primary difference finding observed when compared with the wider population of the Chinese CDC data,² which has a prevalence of COVID-19 in patients aged ≥10 years.

The exposure of the Italian cohort differs from that of the Chinese cohorts because it was highly (>50%) due to exposure outside family clusters (nonfamiliar). This difference may be associated with the delayed lockdown in Italy, imposed on March 8, 5 weeks after the pandemic outbreak. In China, the lockdown occurred 3 weeks after the outbreak was declared a public health emergency of international concern.

Another major difference from previously described COVID-19 cohorts^2–4 is the significant difference in distributions of patients within the disease severity classification. Specifically, our data revealed that children were more frequently categorized as asymptomatic or mild, whereas both Lu et al³ and the CDC² described a prevalence of moderate cases. These data should not lower the attention and preparedness to COVID-19 because children may represent an important source of viral transmission and amplification.¹¹

One of the main features defining a moderate, instead of mild, disease status is the presence of pneumonia, which is diagnosed by either chest radiograph or CT scan,² regardless of the clinical status of the patient (the authors state that this stage may be asymptomatic). In Italian centers, acting independently from one another because no official protocols were available at the time, chest radiographs were performed in a limited number of cases. On the basis of clinical findings, CT scans were performed only in 3 patients because the use of CT scan was considered unnecessary for COVID-19 diagnosis, without any aggravation of the final outcome.

Radiologic imaging was obtained at the discretion of the treating physician at every ED. For this reason, the number of radiologic images obtained may reflect the true behavior of Italian pediatricians in obtaining imaging, which, in such circumstances, was clinically oriented, in the absence of official specific guidelines.

Diversely, all of the patients included in the Qiu et al⁴ series underwent a chest CT. In Lu et al³ and the CDC report,² there is no information about the rate of chest radiograph or CT used to define the severity of disease. The minimal use of chest CT in our experience could have determined the reduced number of cases classified as moderate, also defined as patients presenting no signs and symptoms at clinical evaluation and radiologic but subclinical lung lesions by means of a chest CT.

To accurately compare reports by using homogeneous classifications, we suggest that a clinically driven classification, rather than a radiologic one, could be more helpful for the appropriate clinical diagnosis of children with COVID-19.

A classification of disease severity will help clinicians predict patient needs, plan appropriate ED flow management, and obtain optimal allocation of resources (eg, necessity of a blood test, imaging, or admission).

Conversely, including a radiologic diagnosis of pneumonia as a disease-classification criterion may lead physicians to routinely perform chest CT in children with COVID-19, increasing radiation risk, resource use, and the contagion risk of other healthy workers,¹² without additional benefit to the child.

A key question remains unsolved: that is, trying to manage children (as well adults) with COVID-19 in an evidence-based way, despite the limitation of evidence-based interventions at our disposal. Within this context, radiologic imaging (especially chest CT) does not represent, in our opinion, best practice because of the high number of asymptomatic and mild cases and, in addition, no single finding can reliably differentiate pneumonia of any etiology from other causes of childhood respiratory illness,¹³ and pediatricians are aware of the radiation-induced risk for children.^14–18

Chest CT improves diagnostic capabilities, but its use comes with a well-demonstrated increased risk of solid cancer or leukemia.¹⁴ From a patient perspective, the benefits of a medically necessary CT scan exceed the small increase in radiation-induced cancer risk. Once more, because of the high number of asymptomatic and mild cases of COVID-19 (which in our cohort resulted even higher), chest CTs could be unnecessary for the assessment of the disease severity.

Point-of-care lung ultrasound, which allows for accurate and high-quality investigations without using ionizing radiation, could represent a reasonable alternative for assessing interstitial syndrome and, also, detecting consolidations in children.¹⁹ In our experience, 11 of 13 (84%) patients investigated by using a lung ultrasound has positive results for sonographic findings of interstitial syndrome, with 5 of them showing an associated consolidation. Because, in adults who are COVID-19–positive, lung ultrasound abnormalities have been described before clinical manifestations and virus detection, ultrasound has been proposed as a tool for early diagnosis.^19,20 For children presenting to EDs, an ultrasound approach by expert sonographers could represent an adjunct tool for achieving a rapid severity assessment of COVID-19 lung involvement and tracing the disease evolution. However, we must acknowledge that not all pediatricians may have the proper skills to perform this examination properly, although they could be rapidly trained in lung ultrasound during an emergency situation, such as this.

Considering the different mortality rates across nations and variation of our cohort from the other pediatric cohorts, in further studies, researchers should investigate the possibility of a variable expression of COVID-19 in populations. Some of the differences we highlighted could be explained by the disparity between health systems, which can lead to a limited number of patients accessing hospitals resources.

Our study has some limitations. First, the nature of the study is retrospective. Secondly, the real number of children with a SARS-CoV-2 infection is still unknown in Italy and other countries currently affected by this epidemic. Although we provided a statistical comparison between an Italian cohort and the most representative Chinese studies to date, the scope of this work is not epidemiological because there are currently no estimations of the real disease prevalence among populations, especially in the asymptomatic patient group. For this reason, the real prevalence of COVID-19, its spectrum of presentation, and the real mortality rate remain unknown. Moreover, the reported death rates across countries are heterogeneous, suggesting a lack of uniform case definitions,²¹ possibly due to different patient enrollment settings (hospitalized children [Lu et al³ and Qiu et al⁴] and the Chinese CDC registry [Dong et al²]). Last, we may have missed some of the initial clinical features related to COVID-19. Patients presented to EDs at different stages of illness onset, and the clinical spectrum of the disease may have varied in those patients. Our data collection included only clinical information gathered during ED evaluation and not reports of previously experienced disease-related symptoms, possibly resulting in bias during each clinical observation.

This may lead to an overestimation of more severe cases (moderate to critical), although, with our results, we demonstrated that these numbers are lower than what has been previously reported.

Our population may be not representative of the real COVID-19 spectrum of disease in Italian children, and the comparison with Chinese cohorts may, therefore, be biased. For further analysis, we will collect additional data and perform more effective statistical estimations.