Correlation Study Between Inflammatory Markers and Pathogenic Characteristics in Patients with Severe Pneumonia in Intensive Care Units

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Objective To determine pathogen-specific inflammatory biomarker profiles, develop predictive models for Gram-negative infections and evaluate the added value of serial biomarker monitoring in critically ill patients with severe pneumonia in intensive care units (ICUs). Methods This retrospective cohort study included 126 adults admitted to a tertiary ICU between 1 June 2021 and 31 May 2024. Procalcitonin (PCT), C-reactive protein (CRP) and interleukin-6 (IL-6) were measured at admission and on ICU days 3, 5 and 7. Biomarker-pathogen associations were assessed using Spearman correlation, receiver operating characteristic (ROC) analysis, multivariate regression and stratified analysis of polymicrobial infection and ICU secondary superinfection. Results Predominant pathogens were Acinetobacter baumannii (15.9%), Candida spp. (14.3%), Klebsiella pneumoniae (9.5%), Escherichia coli (7.1%), Pseudomonas aeruginosa (6.3%) and Staphylococcus aureus (3.2%). PCT and IL-6 correlated with Gram-negative infection, especially K. pneumoniae (PCT: r = 0.56, p < 0.001; IL-6: r = 0.51, p < 0.001). PCT ≥ 2.8 ng/mL predicted Gram-negative pneumonia with an AUC of 0.83 (95%CI: 0.78-0.88), 79% sensitivity and 74% specificity. Combining PCT>2.8 ng/mL with IL-6 > 95 pg/mL increased specificity to 87% in monomicrobial Gram-negative infection and 83% in polymicrobial cases. Sustained dual elevation over 72 h predicted 28-day ICU mortality (HR = 2.14, 95%CI: 1.32-3.47) and dynamic trends predicted secondary superinfection better than admission biomarkers (AUC: 0.81 vs 0.65). PCT and IL-6 did not distinguish multidrug-resistant from antimicrobial-susceptible K. pneumoniae. Conclusion PCT and IL-6 help identify Gram-negative pneumonia and serial monitoring adds prognostic value, but biomarker thresholds should be interpreted with clinical findings and microbiological susceptibility results.