目的：探讨急性缺血性脑卒中（AIS）患者1年发生预后不良的危险因素，并建立其列线图风险预测模 型。方法：通过西安脑卒中数据库平台，连续收集本市4所三级甲等医院2015年1~12月入院的AIS患者， 并在入院确诊后随访1年。采用单因素和多因素Logistic回归分析，探讨AIS患者1年预后不良的相关危 险因素，采用R软件及rms程序包构建AIS患者1年预后不良的列线图预测模型。结果：多因素Logistic回 归分析显示，年龄（OR＝1.069，95%CI 1.052～1.087，P=0.000）、合并肺炎（OR＝3.121，95%CI 1.595～ 6.107，P=0.001）、白细胞计数（OR＝1.137，95%CI 1.062～1.217，P=0.000）、心房颤动（OR＝1.816，95%CI 1.059～3.115，P=0.030）、入院NIHSS评分（OR＝1.196，95%CI 1.153～1.241，P=0.000）是西安地区AIS患者 随访1年发生预后不良的独立危险因素。根据上述危险因素，成功建立预测AIS患者1年发生预后不良的 列线图预测模型。该模型ROC曲线下面积（AUC）为0.846，具有良好的区分度；Hosmer-Lemeshow拟合优 度检验结果显示差异无统计学意义（χ2=12.22，df=8，P=0.142）。结论：成功建立用于预测AIS患者随访1年 预后不良的列线图风险预测模型，该模型具有良好区分度与校准度。

To investigate the risk factors of poor prognosis 1 year after acute ischemic stroke (AIS) and to establish a nomogram risk prediction model. Methods: Clinical data of AIS patients who were consecutively admitted to 4 tertiary-grade A class hospitals from January to December 2015 were collected through the Xi’an Stroke Data Bank. The patients were followed up for 1 year after diagnosis. Univariate and multivariate Logistic analysis were applied to analyze the risk factors of poor 1-year prognosis. R software and the rms package were used to establish a nomogram risk prediction model for poor 1-year prognosis in AIS patients. Results: Multivariate Logistic regression analysis showed that age (OR=1.069, 95% CI 1.052～ 1.087, P=0.000), complication by pneumonia (OR=3.121, 95% CI 1.595～6.107, P=0.001), leukocyte count (OR=1.137, 95%CI 1.062～1.217, P=0.000), atrial fibrillation (OR=1.816, 95%CI 1.059～3.115, P=0.030), and NIHSS score at admission (OR=1.196, 95%CI 1.153～1.241, P=0.000) were independent risk factors for poor 1-year prognosis of AIS patients in the Xi’an area. Based on the above independent risk factors, a nomogram prediction model was established to predict poor 1-year prognosis in AIS patients. ROC curve analysis showed that the area under the curve (AUC) was 0.846, indicating good discrimination. The Hosmer-Lemeshow test showed no significant difference (χ2=12.22, df=8, P=0.142). Conclusion: A nomogram risk prediction model for poor 1-year prognosis in AIS patients was successfully established. This model has good differentiation and calibration.