Neurosyphilis: Concordance between cerebrospinal fluid analysis and subsequent antibiotic strategy for patients undergoing evaluation of a diagnosis of neurosyphilis

•Given the long-term sequelae of untreated neurosyphilis and poor sensitivity of tests to detect treponemes in the cerebrospinal fluid (CSF), questions regarding the utility of a lumbar puncture and CSF analysis to either confirm or exclude neurosyphilis are raised.•The authors reviewed the concordance between CSF analysis and the subsequent antibiotic strategy for patients undergoing evaluation of a diagnosis of neurosyphilis.•While 10.7% of patients met existing definitions of neurosyphilis based on CSF parameters, 44.3% were treated with a neurosyphilis-directed antibiotic regimen.•Conversely, 35 patients (26.7% of the cohort) with CSF findings supportive of a diagnosis of neurosyphilis did not receive neurosyphilis therapy.•These observations suggest that clinicians are making therapeutic decisions in spite of the results of CSF analysis. The confirmation or analysis and exclusion of a diagnosis of neurosyphilis has long presented a challenge for infectious diseases clinicians. The authors reviewed the concordance between cerebrospinal fluid (CSF) analysis and the subsequent antibiotic strategy for patients undergoing evaluation of a diagnosis of neurosyphilis. All patients with positive serum syphilis serology referred for CSF analysis between January 2009 and May 2016 were included. Indications for CSF analysis were determined by review of the hospital electronic medical records. CSF parameters were determined from the hospital pathology database. Cases were defined as either ‘confirmed’, ‘supportive’ of, or ‘not supportive’ of a diagnosis of neurosyphilis based on existing definitions. Subsequent therapy was defined as for neurosyphilis, late latent primary syphilis or no therapy based on existing guidelines. Of 131 patients reviewed, 95.4% were male and HIV co-infected (74%). A confirmed diagnosis of neurosyphilis was met by fourteen patients (10.7%). All but two of these were treated with a neurosyphilis-directed regimen. Of the 58 patients treated with neurosyphilis antibiotics, 17.2% had no CSF findings suggestive of the diagnosis. Seventy-three patients were not treated for neurosyphilis; however 35 of these met the CSF criteria for a diagnosis supportive of neurosyphilis. The results of routine CSF analysis in patients with a possible diagnosis of neurosyphilis are inconsistently applied in the clinical setting, calling into question the value of routine CSF. Empirical neurosyphilis treatment should be considered up front in patients with high