目的 分析影响1型糖尿病患者白蛋白尿的危险因素，筛查1型糖尿病肾病高风险人群，以期指导早期防治。 方法 采用横断面回顾性研究的方法，收集中山大学附属第三医院2008年至2015年394例1型糖尿病患者的临床资料，根据尿白蛋白量分为无白蛋白尿组（299例）、微量白蛋白尿组（73例）和大量白蛋白尿组（22例）。使用分类树模型分析筛选尿白蛋白量的相关危险因素，通过结点增益分析筛选微量和大量白蛋白尿的高风险人群。 结果 建立的分类树模型共包括3层，10个结点，从27个预测变量中筛选出4个重要解释变量包括：糖尿病视网膜病变、纤维蛋白原（Fib）、腰臀比、红细胞分布宽度（RDW）。糖尿病视网膜病变是糖尿病白蛋白尿的重要危险因素。糖尿病视网膜病变合并腰臀比＞0.82者出现大量白蛋白尿的概率为43.8%；糖尿病视网膜病变合并RDW＞0.14者，有88.9%的概率出现大量白蛋白尿。 结论 糖尿病视网膜病变是1型糖尿病白蛋白尿的重要的危险因素。

Objective To analyze the related factors of micro-albuminuria and macro-albuminuria in type 1 diabetes mellitus (DM) in the classification tree model, and to screen the high risk population of diabetic kidney disease. Methods 394 patients with type 1 diabetes were enrolled in the hospital from 2008 to 2015. According to glomerular filtration rates and urine albumin quantification, the patients were divided into type 1 diabetes group (299 cases), micro-albuminuria group (73 cases) and macro-albuminuria group (22 cases). The classification tree model was utilized to analyze related factors of the different stages of proteinuria, and the high-risk population was screened by node gain analysis. Results Four important explanatory variables were screened out by the classification tree model from the 27 candidate variables related to primary renal damage, including retinopathy, fibrinogen, waist-hip ratio (WHR), red blood cell distribution width (RDW). Retinopathy was an major factor of DKD. The probability of macro-albuminuria in retinopathy and WHR＞0.82 group was 43.8%, and if at the same time RDW＞0.14, the probability of macro-albuminuria was 88.9%. Conclusions The classification tree model can analyze factors of the separate stages of proteinuria in type 1 diabetic patients effectively. Retinopathy is the major influential factors of type 1 diabetic patients with proteinuria.