Abstract
PURPOSE: Real-world evidence on fluoroquinolone-associated retinal detachment is contradictory. Therefore, we aim to examine the association between newly prescribed fluoroquinolones and the occurrence of retinal detachment with recent data from a large central European country. METHODS: A cohort study with an active comparator new user design was conducted. Dispensings of fluoroquinolone episodes were compared to a group of reference antibiotic episodes (amoxicillin, amoxicillin clavulanic acid, azithromycin, cefuroxime, cephalexin, clindamycin, sulfamethoxazole-trimethoprim, and doxycycline). Data from one of the largest statutory health insurances in Germany, the AOK, were used to follow up individuals with new antibiotic dispensing during the years 2014-2018 for the occurrence of retinal detachment. Piece-wise exponential additive mixed models adjusted for person-time, age, gender, comorbidities, year, and quarter at index were applied to estimate adjusted hazard ratios (aHR) with corresponding 95% confidence intervals (95% CI). RESULTS: In total, 15,232,585 antibiotic episodes were included in the cohort of which 0.05% episodes had an incident retinal detachment. The covariate-adjusted hazard ratio for fluoroquinolone episodes was 1.01 [0.95;1.08]. Likewise, in the propensity score-matched cohort the covariate-adjusted hazard ratio was 0.99 [0.92;1.07]. Moreover, there was little evidence for differences in age and gender subgroups, by follow-up time, selection of active comparator agent, dosage category, or censoring approach. CONCLUSIONS: This large German cohort study found no meaningful real-world evidence for the association between fluoroquinolones and retinal detachment compared to a group of active comparator antibiotics.