The Role of Inflammatory Biomarkers and Microbiological Profile in Clinical Decision-Making for Antibiotic Therapy in Pediatric Patients: A Retrospective Study

Background: Distinguishing bacterial from non-bacterial infections remains challenging in pediatric practice. This study evaluated the relationship between inflammatory biomarkers, microbiological evidence, and antibiotic prescribing, with particular emphasis on agreement between treatment decisions and a study-specific bacterial evidence classification. Methods: This retrospective observational study included 2688 pediatric patients evaluated between January 2023 and March 2026. A nonvalidated, study-specific Bacterial Evidence Score (BES; 0–6 points) combined fever, C-reactive protein, white blood cell count, absolute neutrophil count, procalcitonin, and positive bacterial culture. The primary outcome was disagreement between antibiotic prescribing and the study-defined BES threshold (≥2), rather than independently confirmed inappropriate prescribing. Multivariable logistic regression was used to evaluate factors associated with antibiotic prescribing and BES-based decision disagreement. Results: Antibiotics were prescribed to 1974 patients (73.4%), of whom 1391 (70.5%) received agents classified as broad-spectrum. Microbiological testing was performed in 913 patients (34.0%), yielding 362 positive cultures. Antibiotic prescribing increased overall across BES levels, from 36.9% at score 0 to 98.1% at score 6 (adjusted odds ratio per point, 1.28; 95% confidence interval, 1.17–1.41). Under the study-specific BES ≥ 2 classification, 825 decisions (30.7%) disagreed with the scoring rule; this proportion does not represent the prevalence of clinically inappropriate prescribing. Among culture-positive patients, 167 (46.1%) received pathogen-directed antibiotic therapy supported by microbiological findings, and 91 (25.1%) underwent microbiology-guided treatment modification. Clinical resolution or improvement was documented in 2224 patients (82.7%). Conclusions: Antibiotic prescribing was associated with cumulative inflammatory and microbiological evidence. However, the BES-based classification was exploratory and sensitive to the selected threshold, while microbiological findings were not necessarily available at initial prescribing. These findings support further evaluation of integrated diagnostic and antimicrobial stewardship approaches, with prospective validation required before clinical application.