[Identification of metabolic biomarkers associated with the onset of type 2 diabetes based on a nested case-control study].

Objective: To explore metabolic biomarkers associated with the onset of type 2 diabetes. Methods: Cluster random sampling method was used to select 10 867 local residents aged ≥ 20 years in Liangxi district of Wuxi City, Jiangsu Province in 2007. The baseline survey and physical examination were conducted to collect participants' information, including demographic characteristics, behavior and lifestyles, disease history, family history of diabetes, height, weight, waist circumference and blood pressure, etc. Blood samples were collected and biochemical indexes (high density lipoprotein cholesterol, total cholesterol, triglyceride, fasting blood glucose, etc.) were tested. By June 30, 2020, 220 newly diagnosed patients with type 2 diabetes during the follow-up were selected as cases, and 220 healthy individuals were matched as controls with age (±5 years) and the same sex. High performance liquid chromatography mass spectrometer was used to detect and identify metabolites in serum samples of two groups at baseline. Lasso regression and multivariate conditional logistic regression were used to explore the metabolites associated with the onset of type 2 diabetes. Results: The age of participants at baseline was (53±7) years, and 41.82% were male. 25 out of 1 579 metabolites were selected to be potentially associated with the onset of type 2 diabetes in the lasso regression model. The multivariable conditional logistic regression analysis showed that only 7-Methylxanthine had an independent effect on type 2 diabetes (P=0.019). The area under the receiver operating characteristic curve (AUC) (95%CI) of the prediction model of type 2 diabetes based on traditional risk factors was 0.80 (0.76-0.85). After the 7-methylxanthine in the model, the AUC (95%CI) increased to 0.92 (0.89-0.95) (P<0.001). From the second year, 7-methylxanthine could improve the prediction performance (P=0.007). Conclusion: The level of 7-methylxanthine is related to the onset of type 2 diabetes, and can be used as a biomarker to predict its incidence risk.目的： 探索与2型糖尿病发病相关的代谢标志物。 方法： 于2007年采用整群随机抽样方法在江苏省无锡市梁溪区抽取10 867名社区内20岁及以上本地居民进行基线调查，收集对象基本资料（人口社会学信息、行为和生活方式、疾病史和糖尿病家族史等）进行体格检查（测量身高、体重、腰围和血压等），采集血样并进行血液生化指标检测（高密度脂蛋白胆固醇、总胆固醇、甘油三酯和空腹血糖等）。随访至2020年6月30日，以随访期新发的220例2型糖尿病患者为病例，以年龄（±5岁）和性别为配对因素，在队列中选择220名健康个体作为对照，采用高效液相色谱质谱仪对两组对象基线调查采集的血清样品进行代谢物检测和鉴定。采用lasso回归模型和多因素条件logistic回归模型探索与2型糖尿病发病相关的代谢物。 结果： 基线调查时对象年龄为（53±7）岁，男性占41.82%。对象血清中检测出1 579种代谢物，经lasso回归模型初筛得到25个新发2型糖尿病相关代谢物，将其纳入多因素条件logistic回归模型，最终显示仅7-甲基黄嘌呤显示与2型糖尿病发病相关的独立效应（P=0.019）。基于传统危险因素的2型糖尿病预测模型的受试者工作特征曲线下面积（AUC）（95%CI）为0.80（0.76~0.85），模型中增加7-甲基黄嘌呤后，AUC（95%CI）上升至0.92（0.89~0.95）（P<0.001），且从第2年开始，增加代谢物7-甲基黄嘌呤即可提高模型预测效能（P=0.007）。 结论： 7-甲基黄嘌呤水平与2型糖尿病发病相关，可作为2型糖尿病发病风险预测的生物标志物。.