Epidemiology and Mortality Risk of Severe Viral Pneumonia During the Pre-Pandemic, COVID-19 Pandemic and Post-Pandemic Era: A Retrospective Study of Hospitalized Children in ShenZhen, China Between 2017 and 2023.

Purpose: This study aims to investigate the spectrum of viruses leading to severe viral pneumonia (SVP) and the associated risk factors for mortality among pediatric patients in the pediatric intensive care unit (PICU).

Methods: Taking the outbreak and end of the COVID-19 pandemic as a aboundary, The pre-pandemic period of COVID-19 spans from 01/2017 to 12/2019, the pandemic period from 01/2020 to 12/2021, and the post-pandemic period from 01/2022 to 12/2023. Patients were subsequently stratified into survivor and non-survivor groups based on clinical outcomes.

Results: A total of 1007 patients (median age 1.42 years, range 0.58-4.00; male: female ratio 1.7:1) diagnosed with SVP. Cases were stratified into pre-pandemic (n = 419, 41.6%), pandemic (n = 272, 27.0%), and post-pandemic (n = 316, 31.4%) periods. Viral predominance varied across phases: Pre-pandemic: Influenza A (IVA, 37.0% [155/419]), respiratory syncytial virus (RSV, 29.8%), adenovirus (19.8%), and influenza B (15.5%). Pandemic phase: Human rhinovirus (HRV, 40.1% [109/272]), RSV (33.1%), parainfluenza viruses (11.4%), and bocavirus (HBoV, 10.7%). Post-pandemic: HRV (24.4% [77/316]), RSV (22.8%), HBoV (14.2%), and IVA (13.6%). Comparative analysis revealed significant intergroup differences in the proportion of patients aged < 3 years, primary immunodeficiency disorders (PIDs), and sepsis between pure viral infection deaths and coinfection-associated fatalities among SVP cases. Logistic regression identified eight independent mortality predictors: acute leukemia, other malignant tumors, PIDs, moderate-to-severe underweight, rhabdomyolysis, acute respiratory distress syndrome (ARDS), infection-related encephalopathy, and multiorgan dysfunction syndrome (MODS). The prediction model demonstrated robust discriminative capacity for SVP mortality: sensitivity 73.8%, specificity 90.2%, and AUC 0.888 (95%CI 0.838-0.938) via ROC curve analysis.

Conclusions: The COVID-19 pandemic has altered the landscape of respiratory viruses causing SVP in children. The presence of underlying health conditions, particularly acute leukemia, other malignancies, and immunodeficiency, significantly increases the risk of death in children with viral pneumonia. The risk prediction model offers a reliable tool for clinical practice to predict mortality in these patients.