Toward a clinically based classification of disease severity for paediatric COVID-19 - Authors' reply.

We thank Danilo Buonsenso and colleagues for their excellent suggestion in response to our observational cohort study. We think that defining the moderate clinical type of coronavirus disease 2019 (COVID-19) severity should be based on what the presence of pneumonia means in the progression of COVID-19 and what classifying measures are more practical for clinicians. COVID-19 was initially understood as pneumonia caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Now, after the reporting of a great number of mild and asymptomatic cases of COVID-19 (the so-called iceberg phenomenon that is common for most infectious diseases) and the institution of RT-PCR as a diagnostic standard, radiological examination seems less important as a diagnostic tool than it used to be. Pneumonia is often considered a leading life-threatening risk for children. Radiological evidence is crucial in assessing organ damage, as indicated by the presence of pneumonia when COVID-19 progresses from mild or asymptomatic to severe. This potential for rapid disease progression justifies the definition of the moderate clinical type, which is mainly defined by the presence of mild pneumonia.

Therefore, we used the same definitions of disease severity for our follow-up study (unpublished) as used in our previous cohort study. After analysing the prognostic value of the chest CT scans from 127 paediatric patients with COVID-19 in our follow-up study, we found that length of hospitalisation and body temperature were significantly higher for children with pneumonia than for those without pneumonia. Bilateral pneumonia was often associated with a higher white blood cell count, suggesting mixed infections. Right-lobe pneumonia persisted after 1 month of follow-up, but left-lobe pneumonia did not. Although only a small proportion of paediatric patients progress to severe illness or become critically ill, and COVID-19 pneumonia seems milder than H1N1 pneumonia, the outcome is often serious once patients with COVID-19 enter a severe condition. Among the three patients in our follow-up cohort who had severe disease, one patient died and one became critically ill.

Because of the paucity of long-term follow-up data from the COVID-19 pandemic, more time is required to fully understand the value of radiological examinations. Additionally, from the onset of SARS-CoV-2 infection to the diagnosis of COVID-19, the sensitivity of chest CT scans appears to be higher than that of RT-PCR. Digital radiology might be a suitable choice to reduce the amount of radiation exposure to children. We agree that radiological examinations might not be necessary when looking after a large number of patients, where radiological examinations are not available, or when patients can be diagnosed with mild COVID-19 by an experienced clinician. Nevertheless, using CT scans to diagnose pneumonia helps to define moderate disease severity, provides valuable information about outcome, and increases the accuracy and sensitivity of screening.

The proposition by Buonsenso and colleagues to define moderate disease by a combination of symptoms might produce ambiguous results compared with the direct and objective evidence gained from radiological scans. Patients with COVID-19 often have lung lesions detectable by CT scanning before they exhibit symptoms. Furthermore, a more clinically based diagnosis of pneumonia requires experienced clinicians, and the wheezing, abnormal breath sounds, and snoring in Buonsenso's definition might result from mixed infections or be present intermittently.