Novel risk score model for non-proliferative diabetic retinopathy based on untargeted metabolomics of venous blood

Background and Purpose Nonproliferative diabetic retinopathy (NPDR) occurs in the early stages of Diabetic retinopathy (DR), and the study of its metabolic markers will help to prevent DR. Hence, we aimed to establish a risk score based on multiple metabolites through untargeted metabolomic analysis of venous blood from NPDR patients and diabetic non-DR patients. Experimental Approach Untargeted metabolomics of venous blood samples from patients with NPDR, diabetes melitus without DR were performed using high-performance liquid chromatography-mass spectrometry. Results Detailed metabolomic evaluation showed distinct clusters of metabolites in plasma samples from patients with NPDR and diabetic non-DR patients. NPDR patients had significantly higher levels of phenylacetylglycine, L-aspartic acid, tiglylglycine, and 3-sulfinato-L-alaninate, and lower level of indolelactic acid, threonic acid, L-arginine (Arg), and 4-dodecylbenzenesulfonic acid compared to control. The expression profiles of these eight NPDR risk-related characteristic metabolites were analyzed using Cox regression to establish a risk score model. Subsequently, univariate and multivariate Cox regression analyses were used to determine that this risk score model was a predictor of independent prognosis for NPDR. Conclusions Untargeted metabolome analysis of blood metabolites revealed unreported metabolic alterations in NPDR patients compared with those in diabetic non-DR patients or MH. In the venous blood, we identified depleted metabolites thA and Arg, indicating that they might play a role in NPDR development.