LAUSR

BACKGROUND Diagnosis of neurosyphilis is currently based on the cerebrospinal fluid (CSF) assessments and CSF-Venereal Disease Research Laboratory (CSF-VDRL) is the traditional "gold standard". In the real world, CSF assessments or CSF-VDRL are not always available. This study aimed to identify non-invasive predictors of neurosyphilis based on real-world clinical parameters and diagnostic criteria in populations with different HIV status. METHODS In this retrospective cohort study, syphilis patients with different HIV statuses hospitalized for neurosyphilis screening were retrospectively recruited at an infectious disease hospital. Neurosyphilis was defined by real-world diagnostic criteria. Logistic regression and receiver operating characteristic curve analysis were used to investigate and evaluate predictors of neurosyphilis. RESULTS In total, 528 patients were enrolled, including 143 syphilis patients without HIV infection and 385 HIV/syphilis-coinfected patients. One hundred twelve and 304 neurosyphilis patients were identified in the HIV-negative and HIV-positive groups, respectively. A high serum toluidine red unheated serum test (TRUST) titre was a robust predictor of neurosyphilis in all participants. An age ≥ 50 years old (aOR = 5.062, 95% CI, 1.449-17.680) in the HIV-negative group and CD4+ T cell count < 330/μL (< 300 as reference, aOR = 0.552, 95% CI, 0.315-0.966) in the HIV-positive group were predictors of asymptomatic neurosyphilis. CONCLUSIONS In real-world situations, for asymptomatic syphilis patients, relatively old age and a high serum TRUST titre in HIV-negative populations, and CD4+ T cells < 330/μL and/or serum TRUST titre > 1:64 in HIV-positive populations might predict neurosyphilis.