[Statistical relations between blood and cerebrospinal fluid serology in patients with syphilis]

In any case of latent syphilis without conclusive history, an examination of the cerebrospinal fluid (CSF) is generally recommended. The present paper deals with the question whether particular findings obtained by serologic tests on the blood can supply sufficient proof to rule out any specific involvement of the central nervous system in an individual patient. Serologic investigation on 148 blood-CSF sample pairs revealed striking correlations between several blood and CSF parameters on a high level of statistical significance. Based on blood serology data alone, a multi-dimensional, non-parametric statistical test (the k-nearest-neighbor method) was able to rule out specific involvement of the CSF in more than one third of the patients. We consider our approach useful in certain cases where CSF examination should rather be avoided, especially in patients with impaired general condition due to other diseases. Yet, in any case of neurological irregularity, CSF examination should generally be carried out regardless of the results of blood serology.