Septic complications in pediatric community-acquired pneumonia represent a significant clinical challenge, as severe forms of lower respiratory tract infections in school-aged children and adolescents often remain underestimated. Despite improvements in diagnostic criteria, the early identification of organ dysfunction remains problematic. In this context, the need for highly informative biomarkers that can enhance prognostic accuracy and early risk stratification continues to grow. One such potential indicator is SLPI, which reflects the activity of neutrophil-driven inflammation and proteolytic stress.
A total of 349 children aged 5–17 years hospitalized with community-acquired pneumonia were included in the study. Among them, 135 patients formed the prospective cohort, where sequential follow-up and biospecimen collection were performed according to a standardized protocol, while 214 children constituted the retrospective cohort, for whom medical records from the past five years were analyzed. The control group consisted of 40 practically healthy children of comparable age without acute or chronic respiratory diseases. All patients underwent comprehensive clinical and laboratory evaluation, which included an assessment of general condition, respiratory function, and systemic physiological response. The laboratory assessment included measurement of standard infectious-inflammatory markers–C-reactive protein, erythrocyte sedimentation rate, leukocyte count, interleukin-1, interleukin-6, and lactate. Additionally, the concentration of SLPI was determined as a marker of neutrophil activation and proteolytic imbalance. Disease severity was evaluated using the updated pediatric Phoenix Sepsis Score, with values ≥2 indicating organ dysfunction and fulfilling criteria for sepsis.
Children with severe community-acquired pneumonia demonstrated significantly higher levels of classical systemic inflammatory markers. IL-6 concentrations reached 18.92±1.53 pg/mL, while IL-1 levels were 6.38±0.42 pg/mL (both p<0.001), indicating pronounced cytokine activation. Lactate levels in the severe group were 2.84±0.21 mmol/L (p=0.008), reflecting altered tissue perfusion and metabolic stress. Particular attention was drawn to the SLPI profile: its concentration was 51.18% higher in children with severe pneumonia compared to those with a non-severe course (p=0,0003). Analysis according to the Phoenix criteria further demonstrated that children with sepsis (≥2 points) had SLPI concentrations 45% higher than those without signs of organ dysfunction (p=0.0067), supporting the hypothesis of SLPI involvement in the progression from localized inflammation to systemic response.
A number of statistically significant associations between SLPI and infectious-inflammatory markers were established. Weak positive correlations were observed with CRP (rxy=+0.30; p<0.001), ESR (rxy=+0.22; p=0.011), leukocyte count (rxy=+0.18; p=0.041), and IL-1 (rxy=+0.24; p=0.004), confirming the integrative nature of SLPI as a marker linked to neutrophil-driven inflammation. In contrast, negative correlations were found with PaO₂/FiO₂ ratio (rxy=–0.27; p=0.002) and Glasgow Coma Scale score (rxy=–0.32; p<0.001).
According to ROC curve analysis, isolated SLPI demonstrated moderate but statistically significant diagnostic accuracy for identifying children at risk of sepsis (AUC=0.626; p=0.021). The optimal SLPI cut-off determined by Youden’s index (≈970 ng/mL) yielded a sensitivity of 72.5 % and a specificity of 43.2 %, confirming its ability to detect early changes in patients with potential progression toward organ dysfunction. The baseline inflammatory marker model (CRP, ESR, leukocyte count, IL-1, IL-6, lactate) demonstrated an AUC of 0.684. The addition of SLPI increased AUC to 0.707 (ΔAUC=+0.025; p<0.001), indicating a substantial improvement in discriminatory power. A simplified model incorporating only CRP and SLPI showed comparable accuracy (AUC=0.704), highlighting the significant value of SLPI in predicting sepsis development.
The evaluation of net reclassification improvement (NRI) and integrated discrimination improvement (IDI) demonstrated favorable reclassification after incorporating SLPI into the predictive model, with IDI=+0.021 (p=0.04). This indicates a higher proportion of correctly reclassified patients and improved overall model accuracy. Based on the obtained results, a prognostic model for sepsis risk in children with community-acquired pneumonia was developed. Its application enables more precise early risk stratification, identification of high-risk groups, and optimization of clinical decision-making regarding monitoring and treatment. A promising direction for further development is the integration of SLPI as a predictive variable into machine-learning algorithms to enhance predictive accuracy in large-scale clinical datasets.
