Objective: Examine prevalence and diagnostic utility of cerebrospinal fluid (CSF) treponemal antibodies in early syphilis.
Design: Comparison study.
Setting: Sexually transmitted diseases clinic.
Patients: Forty patients with untreated early syphilis who underwent lumbar puncture. Fifteen were human immunodeficiency virus seropositive.
Measurements: Cerebrospinal fluid cell count, protein, VDRL test, and antibodies to Treponema pallidum by microhemagglutination test for T pallidum (MHA-TP) and fluorescent treponemal antibody absorption test (FTA-ABS); albumin ratio; and IgG index.
Results: Cerebrospinal fluid cell count was not available for one sample, and this patient was excluded from analysis. Of 39 patients, eight (21%) had reactive CSF-VDRL (definite neurosyphilis). Eleven (28%) had mildly elevated cell count or protein concentration, but nonreactive CSF-VDRL (possible neurosyphilis). Twenty had normal cell count and protein concentration, and non-reactive CSF-VDRL (normal). Cerebrospinal fluid MHA-TP and CSF FTA-ABS were reactive in all eight with neurosyphilis. Cerebrospinal fluid MHA-TP was reactive in seven (70%) of 10 with possible neurosyphilis and in six (32%) of 19 with normal CSF. Cerebrospinal fluid FTA-ABS was reactive in four (36%) of 11 with possible neurosyphilis and in five (28%) of 18 with normal CSF. A reactive CSF treponemal test was associated with higher mean CSF cell count and reactive CSF-VDRL.
Conclusion: When criteria to define neurosyphilis depend on cell count or CSF-VDRL reactivity, the sensitivity of CSF treponemal antibodies is high. Nonreactive CSF treponemal tests may help to exclude a diagnosis of neurosyphilis in patients with early syphilis.