Eye (Lond). 2025 Oct 10. doi: 10.1038/s41433-025-04035-2. Online ahead of print.
ABSTRACT
OBJECTIVE: The clinical evidence linking medications to retinal detachment (RD) remains limited. This study utilises real-world data from the U.S. FDA Adverse Event Reporting System (FAERS) to identify drugs associated with RD and characterise their classifications.
METHODS: A disproportionality analysis was performed on over 17 million FAERS reports. Drugs with statistically significant disproportionate RD reporting were identified and categorised by therapeutic class, signal strength, latency period, subgroup factors (e.g., off-label use, age, gender, subtypes, regions), and sensitivity analysis.
RESULTS: Thirty drugs showed significant associations with RD, including ophthalmic agents, anticancer therapies, corticosteroids, and erectile dysfunction medications. Signal strength varied: pilocarpine, encorafenib, and ocriplasmin exhibited the highest signal strength, while prednisolone and bevacizumab showed lower strength. Latency periods differed significantly: erectile dysfunction drugs had the longest median latency (365 days), whereas anticancer drugs had the shortest (14 days). Subgroup analyses revealed elevated RD signal strength with off-label use and distinct susceptibility patterns: younger adults (<65) and males had higher signal strength for specific drug classes (e.g., ophthalmic agents), while older adults (>65) were more susceptible to RD with corticosteroids. Subtype analysis highlighted drug-specific associations with exudative, tractional, and rhegmatogenous RD. Sensitivity analysis restricted to healthcare professional reports further confirmed the robustness of the findings.
CONCLUSIONS: This study provides a stratified list of medications with disproportionate reporting signals for RD. Further high-quality epidemiological and mechanistic studies are warranted to validate the potential causality associations between these medications and RD. Assessment of drug-related retinal detachment using real-world large-scale data.
PMID:41073807 | DOI:10.1038/s41433-025-04035-2
