Alanine to glycine ratio is a novel predictive biomarker for type 2 diabetes mellitus

Abstract Aim We aimed to evaluate the metabolite ratios that could predict the clinical incidence or remission of type 2 diabetes mellitus (T2D). Methods The Cox proportional hazards regression model was used to assess 1813 individuals without T2D to test the predictive value of metabolite ratios for T2D incidence and 451 newly diagnosed T2D for remission. The receiver operating characteristic curve analysis was performed to determine the best cut‐off values for the metabolite ratios. Survival analyses were performed to compare the four subgroups defined by baseline metabolite ratios and clinical status of obesity. Results The alanine/glycine was the most significant marker for T2D incidence (hazard ratio per SD: 1.24; p < .001). On the other hand, metabolite hydroxy sphingomyelin C22:2 was most specific for T2D remission (hazard ratio per SD: 1.32; p = .029). Survival analysis of T2D incidence among the subgroups defined by the combination of alanine/glycine and obesity showed the group with a high alanine/glycine and obesity had the highest risk of T2D incidence ( p < .001). The alanine/glycine as a T2D risk marker was also validated in the independent external data. Conclusions The combination of obesity and the alanine/glycine ratio can be used to evaluate the diabetes risk.