COVID-19 in Brazilian Pediatric Patients: A Retrospective Cross-Sectional Study with a Predictive Model for Hospitalization.

BACKGROUND

This study was conducted to ascertain the most frequent symptoms of COVID-19 infection at first consultation in a pediatric cohort and to devise a predictive model for hospitalization.

METHODS

This is a retrospective cross-sectional study of 1028 Brazilian patients aged <18 years with SARS-CoV-2 infection in a single reference hospital in the first year of the pandemic. Clinical, demographic, laboratory, and disease spectrum data were analyzed via multivariate logistic regression modeling to develop a predictive model of factors linked to hospitalization.

RESULTS

The majority of our cohort were schoolchildren and adolescents, with a homogeneous distribution concerning sex. At first consultation, most patients presented with fever (64.1%) and respiratory symptoms (63.3%). We had 204 admitted patients, including 11 with Pediatric Multisystem Inflammatory Syndrome. Increased D-dimer levels were associated with comorbidities (p = 0.018). A high viral load was observed in patients within the first two days of symptoms (p < 0.0001). Our predictive model included respiratory distress, number and type of specific comorbidities, tachycardia, seizures, and vomiting as factors for hospitalization.

CONCLUSIONS

Most patients presented with mild conditions with outpatient treatment. However, understanding predictors for hospitalization can contribute to medical decisions at the first patient visit.