Predictors of intensive care unit admission in pediatric respiratory infections: the role of inflammatory indices and pathogenic profiles

Aim This study investigated clinical, laboratory, and respiratory pathogen-related factors associated with pediatric intensive care unit (PICU) admission among hospitalized children, emphasizing the role of inflammatory indices and the systemic immune-inflammation index (SII). Methods This retrospective study included children aged 1 month to 18 years hospitalized in the pediatric ward or PICU of Giresun University Women and Children Training and Research Hospital between January 2023 and March 2025. Eligible patients had at least one positive viral respiratory polymerase chain reaction (PCR) result. The primary outcome was PICU admission. We used multivariable binary logistic regression to evaluate associations between respiratory pathogens, clinical and laboratory parameters, and PICU admission. Results Among 835 patients, 291 (34.9%) required PICU admission. In the pathogen-based model, rhinovirus/enterovirus detection was independently associated with increased odds of PICU admission (adjusted odds ratio [aOR], 1.604), whereas adenovirus, respiratory syncytial virus (RSV), and influenza/parainfluenza were associated with lower odds. In the clinical/laboratory model (n = 655), higher SII values (aOR, 1.051 per 100,000-unit increase; 95% CI: 1.029–1.074), heart rate (aOR, 1.025 per beat/min increase; 95% CI: 1.015–1.035), and plateletcrit (aOR, 1.036 per 0.1% increase; 95% CI: 1.020–1.052) were independently associated with PICU admission, whereas higher albumin (aOR, 0.913; 95% CI: 0.876–0.951) and female sex (aOR, 0.667; 95% CI: 0.454–0.979) were inversely associated. SII discrimination was modest (area under the curve [AUC], 0.618). Conclusion Respiratory pathogen profile, SII, heart rate, plateletcrit, albumin, and sex were independently associated with PICU admission. These routinely available parameters may support early risk stratification when interpreted alongside clinical assessment.