Sensitivity and specificity of cerebrospinal fluid CXCL13 for diagnosing Lyme neuroborreliosis - a study on 1410 patients and review of the literature

Abstract

IntroductionThe B-cell chemoattractant CXCL13 has been suggested as a cerebrospinal fluid (CSF) biomarker for Lyme neuroborreliosis (LNB). Our aim was to substantiate the value of CXCL13 in a large unselected cohort and determine a practical cut-off value to diagnose LNB. MethodsWe retrospectively studied clinical and CSF data of consecutive patients who underwent CSF CXCL13 testing over a period of three years (February 2015 to January 2018) at our academic teaching hospital. Patients were classified into 12 groups according to their final diagnosis. To diagnose LNB (definite or probable/possible), definitions of the respective guideline of the German Neurological Society were applied. ResultsOf 1410 patients, 29 were diagnosed with definite LNB and 9 with probable/possible LNB. Median CXCL13 levels were highly elevated in both LNB groups (554 pg/mL and 649 pg/mL, respectively) and the group with bacterial/fungal CNS infections (410 pg/mL; n = 6), while all other groups had markedly lower median CXCL13 levels (p < .001). For definite LNB, the best CXCL13 test cut-off was 55.5 pg/mL with a sensitivity of 96.6% (95% confidence interval, CI, 80.4%–99.8%) and a specificity of 94.9% (95% CI 93.5%–95.9%). All patients with LNB showed clinical improvement after antibiotic treatment. ConclusionIn this large monocentric cohort, CSF CXCL13 was found to be a highly sensitive and useful marker for LNB. In conditions with low index of suspicion for LNB, CXCL13 testing may be unwarranted. A review of the literature on the sensitivity and specificity of CSF CXCL13 in the differential of LNB is provided.