Abstract
Acute ischemic stroke (AIS) subtypes exhibit distinct pathophysiological mechanisms. While current classification methods predominantly depend on neuroimaging, there remains a critical need for sensitive biomarkers to complement imaging and enhance early subtype identification in clinical settings. This study employed metabolomics to identify such biomarkers. A total of 320 AIS patients within 48 h of symptom onset were enrolled, including 227 with large artery atherosclerosis (LAA) and 93 with small vessel occlusion (SVO). Participants were divided into a discovery cohort (n = 177) and a validation cohort (n = 143) based on enrollment order. Pseudotargeted serum metabolomic profiling was performed using liquid chromatography-tandem mass spectrometry (LC-MS/MS). Distinct metabolomic signatures were observed between LAA and SVO subtypes. Three differentiating metabolites-glycoursodeoxycholic acid, docosapentaenoic acid (22n-6), and FAHFA 38:4-were consistently identified and validated across both cohorts. Combined analysis of these metabolites significantly enhanced the discriminatory power for AIS subtype differentiation. This study presents these three circulating metabolites that have the potential to serve as novel biomarkers for the early differentiation between LAA and SVO subtypes of AIS.