Incidence and Factors Associated with Newly Recorded Database-Defined Central Nervous System-Related Diagnoses During and Shortly After Metronidazole Therapy: A Large-Scale Japanese Hospital-Based Administrative Claims Database Study.
Takada Keisuke K, Imai Shungo S, Enoki Yuki Y, Igarashi Yuki Y et al.
Metronidazole (MNZ) is widely used to treat anaerobic and protozoal infections. However, evidence regarding the factors associated with newly recorded, database-defined central nervous system (CNS)-related diagnoses during or shortly after MNZ therapy remains limited. This study evaluated the incidence of newly recorded, database-defined CNS-related diagnoses during and shortly after MNZ therapy and identified the associated factors using a Japanese hospital-based administrative claims database. This retrospective cohort study used the JMDC hospital-based administrative claims database from April 2014 to August 2022. Adults prescribed oral or injectable MNZ were included. The primary outcome was a newly recorded, database-defined CNS-related diagnosis during the risk period after MNZ initiation. Prespecified factors were evaluated using conventional and Firth-penalized Cox proportional hazards regression. Among 58,195 patients, 59 (0.1%) had a newly recorded, database-defined CNS-related diagnosis, including 52 coded as R26.8 (unsteadiness on the feet). Age ≥ 60 years (adjusted hazard ratio [HR] 4.18; 95% confidence interval [CI] 1.70-10.30), female sex (adjusted HR 1.96; 95% CI 1.16-3.31), and renal failure (adjusted HR 1.93; 95% CI 1.08-3.47) were significantly associated with the outcome. In contrast, liver disease and initial daily MNZ dose > 1500 mg/day showed no significant associations. This study estimated the incidence of newly recorded, database-defined CNS-related diagnoses during and shortly after MNZ therapy and identified age ≥ 60 years, female sex, and renal failure as associated factors in the primary Cox regression analysis. Age ≥ 60 years showed the most consistent association across the sensitivity analyses, whereas the associations with female sex and renal failure were less precise in some analyses. However, these findings do not establish causality or a dose threshold.