Online app predicts COVID-19 severity in children

A Bayesian model helps predict the risk of severe COVID-19 in children, which is associated with inflammation, cytopaenia, age, comorbidities, and organ dysfunction, suggests a study.

“The more severe the syndrome, the more the risk factor increases the risk of critical illness,” the researchers said.

A multicentre, prospective study was conducted on children infected with SARS-CoV-2 in 52 Spanish hospitals to identify risk factors causing critical disease and to generate a predictive model that can anticipate the probability of need for critical care. A multivariate Bayesian model was used to estimate the probability of such need.

Three hundred fifty children were enrolled from 12 March 2020 to 1 July 2020, of whom 24.2 percent required critical care. The following clinical syndromes of decreasing severity were identified: multi-inflammatory syndrome (MIS-C; 17.3 percent), bronchopulmonary (51.4 percent), gastrointestinal (11.6 percent), and mild syndrome (19.6 percent). [Pediatr Infect Dis J 2021;40:e287-e293]

Predictors seen for severe COVID-19 included high C-reactive protein (CRP) and creatinine concentration, lymphopaenia, low platelets, anaemia, tachycardia, age, neutrophilia, leukocytosis, and low oxygen saturation. These factors elevated the risk of critical disease depending on the syndrome: “the more severe the syndrome, the more risk the factors conferred,” according to the researchers.

Based on these findings, an online risk prediction tool was developed, which can be accessed through this link: https://rserver.h12o.es/pediatria/EPICOAPP/ (username: user; password: 0000).

The identified risk factors for critical disease were comparable to those of other studies. Of note, some predictors for critical disease, such as age and comorbidities, were patient-dependent, while others indicated immune dysregulation and severe inflammation in critical patients. [J Pediatr 2021;230:23-31.e10; BMJ 2020;370:m3249]

Some biomarkers, such as D-dimer, interleukin-6, or proBNP, could not be used in the predictive model due to inconsistent measurements in patients with mild COVID-19. However, these biomarkers were found to be significantly increased in children requiring critical care. Moreover, high inflammatory markers as CRP or blood cell disorders, such as leukocytosis, neutrophilia, anaemia, lymphopaenia, and thrombopaenia, had been observed in severe cases. [J Infect 2021;82:e16-e18]

“In our analysis, we found that the best clinical cutoff point for platelets was 220,000/mm³ instead of 150,000/mm³, which is classically used for thrombopaenia,” the researchers said. “The cytopenia found in severe cases suggests either damage to bone marrow or peripheral cells or migration of activated cells to tissues.”

With the novel predictive model, physicians can now anticipate the probability of critical care in children with COVID-19, as early recognition of this need is relevant for initiating necessary treatments.

“Through a rapid, inexpensive, and comprehensive web app, the attending physician can introduce the patient’s data at admission and the risk of severe disease can be obtained. The evaluation of the algorithm showed significant accuracy and sensitivity,” the researchers said.

“To our knowledge, this is the first model with an online app to help and recognize the need for critical care in children with COVID-19,” they added.

The model was created for hospitalized children with severe COVID-19 and must not be applied to outpatients, and the risk predicted considered those of patients receiving care only. Finally, this predictive model must be updated regularly since the dynamics of the disease and management strategies could change over time, the researchers noted. [BMJ 2020;371:1-2]