Biomarkers in paediatric bacterial meningitis: a systematic review and meta-analysis of diagnostic test accuracy.

Abstract

Biomarkers for paediatric bacterial meningitis are essential for accurate diagnosis. To perform a systematic review of diagnostic accuracy on cerebrospinal fluid (CSF) and blood biomarkers for paediatric bacterial meningitis. Databases Medline, Excerpta Medica Database, Scopus and Web of Science were used. Eligible studies were those on novel diagnostic CSF and blood biomarkers from which data on biomarker concentration or diagnostic accuracy could be abstracted. Paediatric patients (0-18years) suspected of a central nervous system (CNS) infection. The QUADAS-2 tool was used to assess risk of bias. Difference in biomarker concentrations were assessed by calculating standardised and weighted mean differences. A random-effects meta-analysis model was used. Hierarchical summary receiver-operating characteristic curves were constructed. We identified 3,435 studies, of which 112 articles on 113 individual biomarkers (CSF n=90, blood n=23) were included. In CSF, CRP, IL-6, TNF-α and IL-8 showed the largest mean differences between bacterial meningitis and viral meningitis and IL-6, TNF- α and IL-8 between bacterial meningitis and no CNS infection/inflammation. CSF CRP and ferritin showed excellent discrimination for bacterial versus viral meningitis (summary area-under-the-curve [sAUC] 0.94, 95% CI 0.92-0.97, sAUC 0.94, 95% CI 0.90-1.0). CSF IL-6 and procalcitonin showed excellent discrimination for bacterial versus non-bacterial meningitis and versus no CNS infection/inflammation (sAUC IL-6: 0.98; 95% CI 0.96-1.00, sAUC procalcitonin: 0.96; 95% CI 0.94-0.99). Procalcitonin in blood showed good discrimination (AUC 0.89; 95% CI 0.68-1.00). We identified several CSF biomarkers with high diagnostic accuracy for the diagnosis of bacterial meningitis, including IL-6, procalcitonin, CRP and ferritin. None of the blood biomarkers exhibited excellent discrimination for paediatric bacterial meningitis. Validation of these biomarkers in prospective well-designed studies of diagnostic accuracy performed in children with suspected meningitis is needed. CRD42022361785.