BackgroundPrevious studies have found that inflammatory proteins are involved in the pathogenesis of atrial fibrillation (AF). We used mendelian randomization to explore the potential pathogenic inflammatory proteins of AF. MethodsThis study adopts a Mendelian randomization design to primarily assess causal associations using the Wald ratio and the inverse variance weighting method. It leverages protein quantitative trait locus (pQTL) data encompassing 91 types of inflammatory proteins from 14,824 participants of European ancestry. The primary analysis phase utilizes AF GWAS data from 55,106 participants, with an additional 237,690 participants included in the validation stage. Sensitivity analyses, including reverse causality analysis, Bayesian colocalization analysis, and phenotype scanning, were conducted. Finally, the study explores potential targeted drugs. ResultsThe findings highlight a causal link between 7 inflammatory proteins and AF, with 2 showing positive correlations and 5 exhibiting negative correlations. Among these, fibroblast growth factor 5 (FGF5) emerges as particularly robust in sensitivity analysis. Colocalization analysis indicates a shared genetic variation between FGF5 and AF, supporting its potential as a targeted therapy for AF. Importantly, this causal relationship remains unaffected by reverse causality. Furthermore, significant pleiotropic effects were observed in phenotype scanning. Finally, the causal association between FGF5 and AF was successfully replicated during the validation phase. ConclusionFGF5 may become an intervention target for AF targeted therapy.