This prospective, open-label, non-randomized trial at the University Departments of Infectious Diseases in Ljubljana, Slovenia, and Göteborg, Sweden, was conducted to compare the kinetics of the cerebrospinal fluid (CSF) mononuclear cell count after 10-14 d of ceftriaxone or doxycycline for treatment of Lyme neuroborreliosis. 29 patients were treated with intravenous ceftriaxone 2 g daily in Ljubljana and 36 patients with oral doxycycline 400 mg daily in Göteborg. The study protocol included lumbar puncture before and 6-8 weeks after treatment initiation. There was a marked decrease (1.2 log10 x 10(6)/l) of the median CSF mononuclear cell count following treatment. With the assumption of a linear regression of the logarithmic mononuclear cell counts between the 2 lumbar punctures, no significant difference between the 2 antibiotic treatments could be found. All patients were clinically much improved after treatment. At 6 months follow-up 23 (79%) of the ceftriaxone- and 26 (72%) of the doxycycline-treated patients were completely recovered. Intravenous ceftriaxone or oral doxycycline was found to be effective, safe, and convenient for treatment of Lyme neuroborreliosis.