Effect of Metabolite Levels on Type 2 Diabetes Mellitus and Glycemic Traits: A Mendelian Randomization Study

Objective: To assess the causal associations of plasma levels of metabolites with type 2 diabetes mellitus (T2DM) and glycemic traits. Methods: Two-sample mendelian randomization (MR) was conducted to assess the causal associations. Genetic variants strongly associated with metabolites at genomewide significance level (P < 5 x 10(-8)) were selected from public genome-wide association studies, and single-nucleotide polymorphisms of outcomes were obtained from the Diabetes Genetics Replication and Meta-analysis consortium for T2DM and from the Meta-Analyses of Glucose and Insulin-related Traits Consortium for fasting glucose, insulin, and glycated hemoglobin (HbA1c). The Wald ratio and inverse-variance weighted methods were used for analyses, and MR-Egger was used for sensitivity analysis. Results: The beta estimates per 1-SD increase of arachidonic acid (AA) level was 0.16 (95% CI, 0.078-0.242; P < 0.001). Genetic predisposition to higher plasma AA levels were associated with higher fasting glucose levels (beta 0.10 [95% CI, 0.064-0.134], P < 0.001), higher HbA1c levels (beta 0.04 [95% CI, 0.027-0.061]), and lower fasting insulin levels (beta -0.025 [95% CI, -0.047 to -0.002], P = 0.033). Besides, 2-hydroxybutyric acid (2-HBA) might have a positive causal effect on glycemic traits. Conclusions: Our findings suggest that AA and 2-HBA may have causal associations on T2DM and glycemic traits. This is beneficial for clarifying the pathogenesis of T2DM, which would be valuable for early identification and prevention for T2DM.