Objective: Development and validation of a nomogram for predicting the 4-year incidence of type-2 diabetes mellitus (T2DM) in a Chinese population was attempted. Methods: This prospective cohort study was conducted in Shijingshan District Pingguoyuan Community (Beijing, China) from December 2011 to April 2012 among adults aged≥40 years not suffering from T2DM. Finally, 8 058 adults free of T2DM were included with a median duration of follow-up of 4 years. Participants were divided into a modeling group and verification group using simple random sampling at a ratio of 7∶3. Univariate and multivariate Cox proportional risk models were applied to identify the independent risk predictors in the modeling group. A nomogram was constructed to predict the 4-year incidence of T2DM based on the results of multivariate analysis. The Concordance Index and calibration plots were used to evaluate the differentiation and calibration of the nomogram in both groups. Results: A total of 5 641 individuals were in the modeling group and 2 417 people were in the validation group, of which 265 and 106 had T2DM, respectively, at 4-year follow-up. In the modeling group, age (HR=1.349, 95%CI 1.011-1.800), body mass index (HR=1.347, 95%CI 1.038-1.746), hyperlipidemia (HR=1.504, 95%CI 1.133-1.996), fasting blood glucose (HR=4.189, 95%CI 3.010-5.830), 2-h blood glucose level according to the oral glucose tolerance test (HR=3.005, 95%CI 2.129-4.241), level of glycosylated hemoglobin (HR=3.162, 95%CI 2.283-4.380), and level of γ-glutamyl transferase (HR=1.920, 95%CI 1.385-2.661) were independent risk factors for T2DM. Validation of the nomogram revealed the Concordance Index of the modeling group and validation group to be 0.906 (95%CI 0.888-0.925) and 0.844 (95%CI 0.796-0.892), respectively. Calibration plots showed good calibration in both groups. Conclusion: These data suggest that our nomogram could be a simple and reliable tool for predicting the 4-year risk of developing T2DM in a high-risk Chinese population.
目的： 构建预测成年人4年后2型糖尿病（T2DM）发病风险的列线图并进行验证。 方法： 前瞻性队列研究。于2011年12月至2012年4月在北京市石景山区苹果园社区年龄≥40岁且无T2DM的成年人中进行基线调查，4年后进行随访调查。最后共纳入8 058名受试者，使用随机数字表法按照7∶3的比例分成建模组和验证组。采用单因素和多因素Cox比例风险模型确定建模组发生T2DM的独立危险因素，并根据多因素分析的结果构建列线图，用C指数和校准图评估列线图在建模组和验证组中预测4年发生T2DM风险的区分度和校准度。 结果： 基线调查无T2DM的5 641名建模组受试者和2 417名验证组受试者4年后随访时分别有265例和106例发生T2DM。建模组Cox回归多因素分析显示年龄（HR=1.349，95%CI 1.011~1.800）、体重指数（HR=1.347，95%CI 1.038~1.746）、高脂血症（HR=1.504，95%CI 1.133~1.996）、空腹血糖（HR=4.189，95%CI 3.010~5.830）、糖负荷后2 h血糖（HR=3.005，95%CI 2.129~4.241）、糖化血红蛋白（HR=3.162，95%CI 2.283~4.380）、γ-谷氨酰转移酶（HR=1.920，95%CI 1.385~2.661）是T2DM发生的独立危险因素。根据独立危险因素构建的列线图经验证显示，建模组和验证组C指数分别为0.906（95%CI 0.888~0.925）和0.844（95%CI 0.796~0.892）。校准图显示，T2DM的估计概率与实际观测概率之间具有良好的一致性。 结论： 构建的列线图是预测成年人4年患T2DM风险的简单可靠工具，便于早期识别高危人群。.