中文摘要：

目的：探索性构建稳定期冠心病患者1年内发生急性心血管事件的预警模型,应用 ROC曲线评价效能。方法选择中日友好医院稳定期冠心病患者,且完成了1年随访者共397例,结合随访1年后发生急性心血管事件情况,对患者的临床资料进行多因素分析。结果建立预警方程：Logit（p）=-2.319-0.779（ApoA1分层）-0.978运动＋0.908脑卒中病史＋0.877（hs CRP分层）＋0.797胸闷。应用 ROC曲线评价效能,评价指标结果：曲线下面积（AUC）为0.744,敏感度为81.4%,特异度为62.2%, Youden指数为0.436,阳性似然比为2.153,阴性似然比为0.299,阳性预报值为0.218,阴性预报值为0.964。结论建立的预警方程主要应用于预测1年内稳定期冠心病患者发生心血管事件的危险。

英文摘要：

Objective To establish a recurrent event early alarming model for stable coronary heart disease （CHD）.Methods Three hundreds and ninety seven patients with stable CHD followed up for one year were enrolled.All patients were treated with acute cardiovascular event（ACE）.The clinical data were collected and analyzed.Results Early alarming model：Logit（p）=-2.319 0.779 ApoA1-0.978 physical exercises ＋0.908 stroke history＋0.877 hs CRP ＋0.797 suffocating.Then the effectiveness of the model was verified by ROC curve,and the results showed that AUC=0.744,sensitivity was 81.40%,specificity was 62.20%,Youden index=0.436,positive likelihood=2.153,negative likelihood=0.29,positive predictive=0.218,negative predictive=0.964.Conclusion A recurrent event early alarming model for stable CHD was established to predict the risk of stable CHD patient with ACE within one year.