| アブストラクト | BACKGROUND AND AIMS: Drug-induced dystonia is a serious, potentially disabling adverse event (AE) associated with certain medications. Despite its clinical relevance, the existing literature is largely limited to case reports or analyses of individual drugs, and a systematic evaluation of medications implicated in dystonia remains lacking. This study undertook one of the first comprehensive screenings and risk signal rankings of drugs linked to dystonia using data from the FDA adverse event reporting system (FAERS), aiming to inform clinical drug safety. METHODS: Reports of dystonia-related AEs from Q1 2004 to Q3 2024 were retrieved from FAERS using standardized MedDRA queries. Disproportionality analyses were conducted using the reporting odds ratio (ROR), proportional reporting ratio (PRR), Bayesian Confidence Propagation Neural Network (BCPNN), and Multi-item Gamma Poisson Shrinker (MGPS) methods to identify potential drug-dystonia signals. The Kaplan-Meier method was applied to assess the time to dystonia onset after drug exposure. Sensitivity analyses using the Omega shrinkage estimator were performed to explore potential drug-drug interactions. RESULTS: A total of 27,618 patients with 28,938 dystonia reports were included. Metoclopramide (7178 reports) and aripiprazole (1595 reports) were most frequently reported. Among the top 50 drugs, metoclopramide showed the strongest disproportionality signal, followed by prochlorperazine and haloperidol. Notably, dystonia was not listed in the package inserts of eight identified drugs, including certain antiepileptics, antidepressants, and antiparkinsonian agents. Most events occurred within 0-30 days of drug initiation. The combination of aripiprazole and risperidone was frequently reported and showed notable interaction signals. CONCLUSIONS: Metoclopramide and several antipsychotics were strongly associated with reported drug-induced dystonia. These findings highlight the need for cautious prescribing and close monitoring, particularly with antipsychotic combinations, and may support improved risk awareness and labeling. |