Abstract
Osteoarthritis (OA) is the most common joint disease in the world, characterized by pain and loss of joint function, which has led to a serious reduction in the quality of patients' lives. In this work, ultrahigh performance liquid chromatography coupled with quadrupole time-of-flight tandem mass spectrometry (UPLC-QToF/MS) in conjunction with multivariate pattern recognition methods and an univariate statistical analysis scheme were applied to explore the serum metabolic signatures within OA group (n = 31), HC (healthy controls) group (n = 57) and non-OA group (n = 19) for early diagnosis and differential diagnosis of OA. Based on logistic regression analysis and receiver operating characteristic (ROC) curve analysis, seven metabolites, including phosphatidylcholine (18:0/22:6), p-cresol sulfate and so on, were identified as critical metabolites for the diagnosis of OA and HC and yielded an area under the curve (AUC) of 0.978. The other panel of unknown m/z 239.091, phosphatidylcholine (18:0/18:0) and phenylalanine were found to distinguish OA from non-OA and achieved an AUC of 0.888. These potential biomarkers are mainly involved in lipid metabolism, glucose metabolism and amino acid metabolism. It is expected to reveal new insight into OA pathogenesis from changed metabolic pathways.