Detection of drug-induced acute respiratory distress syndrome risk signals in FAERS: a real-world pharmacovigilance study.

Background: Drug-induced acute respiratory distress syndrome (ARDS) represents an often overlooked yet potentially severe adverse reaction in clinical practice. Existing evidence predominantly stems from isolated case reports, and systematic investigations based on large-scale real-world data remain limited. This study aimed to comprehensively evaluate the risk signals of drug-related ARDS and identify potential high-risk drug classes using the US FDA Adverse Event Reporting System (FAERS). Methods: We retrieved data from FAERS and extracted ARDS-related reports based on the MedDRA Preferred Terms (PT), followed by deduplication and data cleaning. Four disproportionality analysis methods were used to detect drug signals. LASSO regression and multivariable logistic regression were applied to identify potential independent risk factors. The time interval from drug exposure to ARDS onset was also assessed. Results: A total of 15,986 drug-related ARDS reports were included, with the highest proportion occurring in middle-aged and older adults, and slightly more cases in males than in females. The five most frequently reported drugs were mycophenolic acid, methotrexate, rituximab, tacrolimus, and amiodarone. Signal detection indicated strong associations for prednisone, mycophenolic acid, amiodarone, and cytarabine. Multivariable analyses further identified 22 drugs significantly associated with ARDS risk. The median time to ARDS onset was 30 days, and approximately 75% of cases occurred within 150 days after treatment initiation. Conclusion: Using real-world data, this study identified multiple drug classes with significantly elevated ARDS risk, including immunosuppressants, antineoplastic agents, glucocorticoids, cardiovascular drugs, and nonsteroidal anti-inflammatory drugs (NSAIDs). These findings provide important evidence for clinical monitoring of high-risk medications.