Background: Lyme neuroborreliosis (LNB) is characterized by cerebrospinal fluid (CSF) inflammation with several cytokines/chemokines and B-lymphocytes. Clinically, LNB in children may be difficult to discriminate from non-Lyme aseptic meningitis (NLAM). We aimed to identify CSF cytokine/chemokine patterns in children with LNB, NLAM and controls and elucidate the diagnostic value of these cytokines/chemokines alone or in combination to discriminate between LNB and NLAM.
Methods: Children with symptoms suggestive of LNB were included prospectively and categorized as LNB, NLAM or controls (no pleocytosis). Cytokines/chemokines in CSF were measured by multiplex bead assays and levels were compared between the three groups by nonparametric statistical tests. Previous results from the same children on the established biomarker, CXCL13, were included in the statistical analyses. The diagnostic properties of cytokines/chemokines to discriminate between LNB and NLAM were determined by receiver operating characteristic curve analyses with estimates of area under curve (AUC). To explore diagnostic properties of combinations of cytokines/chemokines, prediction models based on logistic regression were used.
Results: We included 195 children with LNB (n = 77), NLAM (n = 12) and controls (n = 106). Children with LNB had higher CSF levels of CCL19, CCL22 and CXCL13 compared to NLAM and controls, whereas INFγ was higher in NLAM than in LNB and controls. CXCL13 was the superior single cytokine/chemokine to discriminate LNB from NLAM (AUC 0.978). The combination CXCL13/CCL19 (AUC 0.992) may possibly improve the specificity for LNB, especially for children with moderate CXCL13 levels.
Conclusions: The intrathecal immune reaction in LNB is characterized by B cell associated chemokines. Whether the combination CXCL13/CCL19 further improves discrimination between LNB and NLAM beyond the diagnostic improvements by CXCL13 alone needs to be tested in new studies.
Keywords: Chemokine; Cytokine; Diagnosis; Meningitis; Neuroborreliosis; Neuroinflammation.
Copyright © 2020 Elsevier Ltd. All rights reserved.