ObjectiveTo explore the causal relationship between different types of drug poisoning and different types of stroke using Mendelian randomization (MR).MethodsDrug poisoning types included anticonvulsant and anti-Parkinsonism drug poisoning, as well as poisoning from analgesics, antipyretics, and antirheumatic drugs. The types of strokes encompassed ischemic stroke, large artery atherosclerotic ischemic stroke, small-vessel ischemic stroke, cardiogenic ischemic stroke, lacunar stroke, transient ischemic attack, intracerebral hemorrhage, and subarachnoid hemorrhage. The above data is sourced from the genome-wide association study (GWAS) public database. Bidirectional MR analysis was performed using six methods, including inverse variance weighted (IVW). Multiple sensitivity analyses were employed to test the results for pleiotropy and heterogeneity. The MRlap function was used to correct for causality in the presence of sample overlap, and Meta-analysis was utilized to deal with bias due to different outcome sources. Finally, reverse MR analyses were performed for drug intoxication types that were found to be causally associated with stroke in the forward MR analyses.ResultsAnticonvulsant and anti-Parkinsonism drug poisoning were protective factors for intracerebral hemorrhage (source from European Molecular Biology Laboratory) [OR(95%CI)=0.969(0.945-0.995), P < 0.05], while Meta-analysis showed anticonvulsant and anti-Parkinsonism drug poisoning had no protective effect on intracerebral hemorrhage [OR(95%CI)=0.938(0.845-1.042), P > 0.05]. Analgesics, antipyretics, and antirheumatic drugs poisoning were risk factors for small-vessel ischemic stroke and lacunar stroke[OR(95%CI)=1.074(1.011-1.142), 1.094(1.031-1.162); P < 0.05]. All results were free of heterogeneity and horizontal pleiotropy. In addition, there was no reverse causality between drug poisoning and stroke.ConclusionInappropriate use of different types of drugs can affect the risk of stroke.