Serum metabolomics for the diagnosis and classification of myasthenia gravis

Abstract

Myasthenia gravis (MG) is a chronic autoimmune neuromuscular disease with few reliable diagnostic measures. Therefore, it is great important to explore novel tools for the diagnosis of MG. In this study, a serum metabolomic approach based on LC–MS in combination with multivariate statistical analyses was used to identify and classify patients with various grades of MG. Serum samples from 42 MG patients and 16 healthy volunteers were analyzed by liquid chromatography Fourier transform mass spectrometry (LC-FTMS). MG patients were clearly distinguished from healthy subjects based on their global serum metabolic profiles by using orthogonal partial least squares (OPLS) analysis. Moreover, different changes in metabolic profiles were observed between early- and late-stages MG patients. Nine biomarkers, including gamma-aminobutyric acid and sphingosine 1-phosphate were identified. In addition, 92.8% sensitivity, 83.3% specificity and 90% accuracy were obtained from the OPLS discriminant analysis (OPLS-DA) class prediction model in detecting MG. The results presented here illustrate that serum metabolomics exhibits great potential in the detecting and grading of MG, and it is potentially applicable as a new diagnostic approach for MG.