Abstract
BACKGROUND: Stroke is a major cause of serious disability in the United States. Previous studies found multiple associations of serum metabolites with acute ischemic stroke (AIS) compared to controls, but few of them evaluated metabolome in a longitudinal fashion. Therefore, we compared the metabolome of the acute and chronic stages of ischemic stroke. METHODS: We evaluated 1295 serum metabolites from the cohort of 60 stroke patients at acute and chronic stages by performing global metabolomics using ultra-high-performance liquid chromatography/mass spectrometry (LC-MS) and gas chromatography/mass spectrometry (GC-MS). We used Orthogonal Partial Least Square-Discrimination Analysis (OPLS-DA) to inspect group disparity and a mixed regression model to compare metabolites in the acute and chronic stages with Two-Stage Benjamini & Hochberg (TSBH) and Bonferroni correction for multiple testing. RESULTS: The OPLS-DA revealed significant separation of acute and chronic stage metabolites. Mixed regression identified 228 metabolites with TSBH, and 29 metabolites with Bonferroni correction different in acute and chronic stages. At the acute stage, there was a consistent increase of the metabolites of mono/diacylglycerols, sphingolipids, medium/long-chain fatty acids, and amino acids glycine, valine, and tyrosine. At the same time, there was a consistent decrease of the metabolites of acyl-choline related fatty acids, phospholipids, and amino acids alanine, aspartate, and tyramine. Additionally, we identified eight novel metabolites significantly altered at the acute stage of stroke. CONCLUSION: Our pilot study demonstrated significant alterations in metabolomic patterns between the acute and chronic stages of stroke, validating some case-control findings. Future investigation in a larger independent cohort is warranted to identify early biomarkers of acute ischemic stroke.