To determine the individual relative risk to neurologically affected patients of infection with Borrelia burgdorferi, within the framework of a multicenter case control study encompassing the four neurological departments of the Universities of Basel, Bern, Zurich and Lausanne, 378 patients and 1134 healthy blood donors serving as controls underwent analysis for antibodies against Borrelia burgdorferi by ELISA. The seroprevalence were estimated for a cut-off point of 2 standard deviations from the mean, these values corresponding to cut-off points of 1:32 for IgM and 1:256 for IgG by immunofluorescence testing. For IgM, 4.8% of the neurological patients were positive versus 4.1% of controls; the corresponding values for IgG were 10.1% versus 10.7% respectively. Hence, neurologically affected patients were not at higher risk for infection with Borrelia burgdorferi than were controls. We found no elevated relative risk in any diagnostic subgroup. The results of a positive Lyme serology must therefore be interpreted with care and in relation to clinical and CSF findings. On the basis of our results, screening for Lyme borrelioses serves no purpose.