Abstract 4139938: Lipidomic signature of acute ischemic and hemorrhagic stroke

Introduction: Stroke is the second leading cause of death worldwide and the leading cause of disability. Without a plasma biomarker care is often delayed. The brain is rich in lipids, which readily cross the blood-brain barrier, and therefore represent a target for biomarkers of stroke. The study aimed to determine the lipidomic changes in plasma in stroke patients through an untargeted and targeted analysis. Methods: A cohort of 482 patients from the INTERSTROKE study was included in the analysis. This included 241 stroke patients (120 ISCH, 121 HEM) and 241 age and sex-matched controls. A LC/MS/MS platform was used to perform a detailed lipidomic and oxylipidomic analysis of plasma. Biomarker analysis was performed using the Random Forest Classification algorithm (RFC). Results: Lipidomic analysis identified 141 lipid species and 32 oxylipins significantly altered in ISCH stroke compared to control and 167 lipid species and 34 oxylipins in HEM stroke. Between ISCH and HEM stroke there were in 87 significant lipids. There was a 141 % increase in phosphatidylserine (PS) 40:6 (p< 0.0001) in patients with HEM stroke compared to ISCH, whereas prostaglandin E2 was found to be 94% higher (p<0.0001) in ISCH stroke. RFC model identified lysophosphatidylcholine (LPC) 14:0 and 15-oxoETE as the most important features contributing to classification between control and stroke with a mean AUC score of 0.81 and 0.69 respectively. PS 40:6 and 14-HDoHE are the most important features contributing to separating ISCH and HEM stroke, with mean AUC scores of 0.84 and 0.67, respectively. PC 34:1 and PC-O 36:3 were found to be the biggest predictors of stroke outcomes at 30 days, measured by The Modified Rankin Score (mRS) score. In addition, four lipids: Ceramide 24:1, C14 Tetradecanoyl Carnitine, trihexosylceramide 16:0, and PS 40:6 were found to be the independent predictors of mortality in patients with stroke. Conclusion: We have demonstrated significant alterations in the human plasma lipidome during acute stroke, revealing distinct differences between HEM and ISCH subtypes. These lipid profiles not only differentiate between stroke subtypes but also predict clinical outcomes. Given the current lack of plasma biomarkers for stroke, our study underscores the potential of lipid molecules as valuable biomarkers for stroke diagnosis and prognosis.