Diagnosis–Treatment Discordance in Suspected Lyme Neuroborreliosis: A Retrospective Cohort Study from a Romanian Tertiary Centre (2008–2025)

Abstract

Background: European and Romanian recommendations for Lyme neuroborreliosis (LNB) emphasise cerebrospinal fluid (CSF) pleocytosis and a positive intrathecal antibody index (IAI) for diagnostic confirmation. We assessed whether antibiotic treatment in suspected LNB aligns with diagnostic recommendations. Methods: We conducted a retrospective, single-centre cohort study of patients with suspected LNB between 2008 and 2025, all with positive first-tier ELISA (enzyme-linked immunosorbent assay) Borrelia serology. Patients were classified as confirmed, probable, possible, or ruled-out LNB based on clinical presentation, serology, and CSF findings. We assessed factors associated with antibiotic treatment and temporal changes in diagnostic work-up and treatment patterns. Results: Among 58 patients, LNB was confirmed in 4 (6.9%), probable in 11 (19.0%), possible in 31 (53.4%), and ruled out in 12 (20.7%). Antibiotics were administered to 47 patients (81.0%), of whom 68.1% had possible or ruled out LNB. Antibiotic treatment was more frequent in early-stage than in late-stage disease (38/42, 90.5% vs. 9/16, 56.2%; p = 0.006); no significant association was detected with confirmed (p = 1.000) or ruled-out (p = 0.681) status. Between 2008 and 2019 and 2023–2025, IAI testing increased significantly (35.1% vs. 71.4%, p = 0.013), yet overall antibiotic use remained unchanged (81.1% vs. 81.0%, p = 1.000). Conclusions: Antibiotic treatment was poorly aligned with diagnostic classification, with most treated patients lacking confirmation. Improved diagnostic evaluation over time was not accompanied by a reduction in antibiotic use, highlighting a critical need for stronger diagnostic and antibiotic stewardship.