Antibiotic use in children with suspected CNS infection in the molecular diagnostic era: a cohort study of 1198 children

Abstract Objectives Central nervous system (CNS) infections in children require rapid diagnostic evaluation and prompt empiric antimicrobial therapy. However, clinical presentation is often non-specific, leading to frequent use of broad-spectrum antibiotics while awaiting cerebrospinal fluid (CSF) results. We evaluated antibiotic treatment modifications during the diagnostic work-up and quantified changes in antibiotic spectrum coverage (ASC) before and after lumbar puncture (LP). Patients and methods We conducted a retrospective single-centre study including children <18 years who underwent LP for suspected CNS infection between January 2016 and December 2024. CSF was analysed using the BioFire® FilmArray® Meningitis/Encephalitis (ME) Panel. Antibiotic treatment modifications after LP-related diagnostic information became available were assessed descriptively. Paired ASC scores before and after LP were compared using the Wilcoxon signed-rank test among treated children. Results Among 1198 children undergoing LP, 730 (60.9%) received antibiotic therapy. Antibiotic use was highest in neonates (227/246; 92.3%) and infants aged 29 days to <6 months (118/140; 84.3%) and lowest in adolescents (106/329; 32.2%). Treatment was modified after LP-related diagnostic information became available in 412/1198 children (34.4%). Among treated children, 313 (42.9%) underwent de-escalation, 99 (13.6%) escalation and 318 (43.6%) had no change. Overall, mean ASC decreased from 8.5 before LP to 6.6 afterwards (ΔASC −1.9; P < 0.001). Conclusions LP-related diagnostic information was associated with frequent antibiotic treatment modification and reduced antibiotic spectrum coverage. These findings support integrated CSF diagnostics, including multiplex polymerase chain reaction, as a component of antimicrobial stewardship in children with suspected CNS infection.