缺血性脑血管病急性期危险分层对复发预测的对比分析

发布时间：2018-04-28 16:38

本文选题：缺血性脑卒中 + 短暂性脑缺血发作　；参考：《复旦大学》2013年硕士论文

【摘要】：背景：脑血管病具有高发病率、高复发率、高致残率和高死亡率的特点,早期进行复发风险评估、并干预危险因素可以预防脑血管病的复发,对该病的整体防治具有十分重要的意义。现在常用的对脑卒中复发风险分层的工具有ESRS和ABCD2评分,其中ABCD2评分主要用于评估TIA患者是否有早期脑卒中复发风险、是否需要紧急处理或住院观察,ESRS评分主要用于预测缺血性脑卒中发病后1年的脑卒中复发风险；这两个评分重在利用临床症状和体征等进行评分,虽然简易,但没有考虑到对脑卒中预后有重要影响的实验室及血管成像危险因素,势必影响对脑卒中复发的预判效果。本研究在ABCD2和ESRS评分的基础上,综合考虑临床症状、体征、实验室和影像学危险因素,建立新的脑卒中风险评估模型,提高对脑卒中复发的预测能力,以指导缺血性脑血管病的二级预防工作。目的：(一)建立以缺血性脑血管病患者的临床特征、影像学、实验室危险因素为基础的脑卒中风险评分模型——SRS (Stroke Risk Score);(二)对所建立的SRS评分模型与ESRS和ABCD2评分对缺血性脑血管病患者的复发预测进行有效性评估。方法：(一)SRS评分模型的建立：(1)SRS评分原始模型的潜在构成包括：年龄、血压、血糖、低密度脂蛋白胆固醇、同型半胱氨酸、C反应蛋白、心房颤动、颈动脉成像、既往缺血性脑卒中或短暂性脑缺血发作史、既往心肌梗死史、外周动脉疾病等危险因素；(2)用单变量Cox比例风险回归分析确定缺血性脑血管病后2年内脑卒中复发的危险因素,首先在集合变量模型A中引入单变量Cox比例风险回归分析中显著性为P≤0.10单变量危险因素；所有其他没有被纳入A模型的、先前定义的脑卒中复发独立预测因子,逐个考察其对脑卒中复发的预测价值,以P0.05为有显著性差异,逐个纳入能整体提高每个模型的C-统计值的独立预测因子。通过逐步分析最终确定了SRS评分模型的独立预测因子构成及其最佳加权方案。(二)预测效果评价,分组为：推导组300例患者为2009年9月1日至2010年8月31日缺血性脑血管病住院患者,随访2年；验证组315例患者为2010年9月1日至2011年8月31日缺血性脑血管病住院患者,随访1年；集合组为2009年9月1日至2011年8月31日住院的推导组和验证组615例患者；TIA组为2009年9月1日至2011年8月31日推导组和验证组中的65例TIA住院患者。分析：(1)计算各评分量表的C-统计值,以评估各评分量表的综合预测效果。(2)应用ROC曲线评价各评分将集合组患者评估为低危组和危险组的敏感度、特异度,比较各评分模型对脑卒中复发高风险患者的辨别能力。(3)应用Kaplan-Meier曲线法估计各模型评估为低危组和危险组患者的生存曲线,比较低危组和危险组患者的无脑卒中复发生存率高低。(4)考察观察性风险与预测性风险的变化趋势。结果：(一)SRS评分模型的建立：SRS评分模型构成为：年龄、糖尿病史、低密度脂蛋白胆固醇、同型半胱氨酸、纤维蛋白原、心房颤动、颈动脉成像、既往缺血性脑卒中或短暂性脑缺血发作史、外周动脉疾病、冠状动脉疾病、充血性心力衰竭、吸烟；通过逐步分析,剔除了对推导组患者无预测价值的危险因素变量,引入了对推导组患者有预测价值的危险因素变量；并逐个分析了能够使评分模型获得最大C-统计值的每个危险因素变量的最佳加权方案；最终建立了SRS评分模型,见下表所示。 (二)预测效果评价：在对推导组患者30天、90天、180天、1年、2年脑卒中复发预测中,ESRS评分预测价值良好(C-统计值0.6063-0.6936),ABCD2评分预测价值较低(G-统计值0.3970-0.5702),SRS评分(C-统计值0.5863-0.6940)与ESRS评分相似,均较ABCD2评分明显提高。对验证组患者30天、90天、180天、1年脑卒中复发预测中,ABCD2评分预测价值适度(C-统计值0.5329-0.5790),SRS评分与ESRS评分相似(C-统计值分别为0.4726-0.5335,0.4656～0.5019)。在对集合组30天、90天、180天、1年脑卒中复发预测中,SRS与ESRS评分相似(C-统计值分别为0.5438-0.5925,0.5487～0.5884),两者均较ABCD2评分提高(C-统计值0.4957-0.5728)。在对TIA组30天、90天、180天、1年脑卒中复发预测中,SRS评分与ABCD2评分相似(C-统计值分别为0.3047～0.5000,0.4703～0.5391),均较ESRS评分高(C-统计值0.1719-0.4209)。在对集合组患者脑卒中复发风险评估中,在30天、90天、180天、1年内脑卒中复发高风险患者评估为危险组的准确性,SRS评分的敏感度和特异度分别为77.8～82.7%、33.2～35.2%,ESRS评分分别为66.7～73.0%、40.1～41.7%,ABCD2评分分别为89.2～100%、8.8～9.3%。除外验证组、TIA组,对于推导组、集合组患者脑卒中复发风险评估中,SRS与ESRS评分评估为危险组的患者均较低危组有较高的脑卒中复发率(P0.05),但前者较后者区别更明显；ABCD2评分评估为低危组和危险组的患者脑卒中复发率无统计学差异(P0.05)。结论：在以急性缺血性卒中为主体的急性缺血性脑血管病患者队列中,SRS、ESRS评分对脑卒中复发风险的预测准确性相似,SRS较ESRS评分整体预测准确性稍高,而ABCD2评分的预测准确性较低。SRS、ESRS评分对于急性缺血性脑血管病患者中脑卒中复发高风险患者的区别能力相似,SRS评分整体区别能力较ESRS评分稍高,ABCD2评分的区别能力较差。相关结果结论需进一步前瞻性研究证实。
[Abstract]:Background : cerebrovascular disease has the characteristics of high incidence rate , high recurrence rate , high disability rate and high mortality rate , early relapse risk assessment and intervention of risk factors can prevent the recurrence of cerebral vascular disease .
This study , based on the ABCD2 and ESRS scores , comprehensively considers clinical symptoms , signs , laboratory and imaging risk factors , establishes a new risk assessment model of stroke , and improves the prediction ability of the recurrence of stroke to guide the secondary prevention of ischemic cerebrovascular disease .

Objective : ( 1 ) To establish a model _ SRS ( Stroke Risk Score ) based on clinical features , imaging and laboratory risk factors in patients with ischemic cerebrovascular disease .

Methods : ( 1 ) The establishment of SRS scoring model : ( 1 ) The potential constituents of the original SRS scoring model include : age , blood pressure , blood sugar , low density lipoprotein cholesterol , homocysteine , C - reactive protein , atrial fibrillation , carotid imaging , history of previous ischemic stroke or transient ischemic attack , history of previous myocardial infarction , peripheral arterial disease and other risk factors ;
( 2 ) Cox proportional hazards regression analysis was used to determine the risk factors of stroke recurrence within 2 years after ischemic cerebrovascular disease .
All the other independent predictors of stroke recurrence , which were not included in model A and previously defined , were investigated on a one - by - one basis to predict the predictive value of the recurrence of stroke . The independent predictors of the C - statistic value of each model were included one by one .
In the validation group , 315 patients were hospitalized for ischemic cerebrovascular disease from September 1 , 2010 to August 31 , 2011 , followed up for 1 year ;
The collection group was 615 patients in the derivation group and validation group , hospitalized from September 1 , 2009 to August 31 , 2011 ;
( 2 ) Using the Kaplan - Meier curve method to estimate the survival curves of patients with low risk group and risk group , compare the survival curves of each model to low risk group and risk group , compare the survival curves of the low risk group and the risk group , compare the low risk group and the risk group patient ' s survival curve , compare the low risk group and the risk group patient ' s survival curve , and ( 4 ) investigate the trend of the observational risk and the predictive risk .

Results : ( 1 ) The establishment of SRS scoring model : SRS scoring model consisted of age , diabetes history , low density lipoprotein cholesterol , homocysteine , fibrinogen , atrial fibrillation , carotid imaging , previous ischemic stroke or transient ischemic attack history , peripheral artery disease , coronary artery disease , congestive heart failure , smoking ;
By stepwise analysis , the risk factor variables of no predictive value in the derived group were excluded , and the risk factor variables with predictive value for the derived patients were introduced .
and analyzing the optimal weighting scheme of each risk factor variable which can make the scoring model obtain the maximum C - statistic value one by one ;
The SRS scoring model is finally established , as shown in the following table .

( C - Statistical value : 0.5329 - 0.6936 ) , the score of SRS was similar to that of ESRS ( 0.5438 - 0.5925 , 0.5487 - 0.519 ) .
The patients with low risk group and risk group had no statistical difference in the ABCD2 score ( P0.05 ) .

Conclusion : In the cohort of patients with acute ischemic cerebrovascular disease , the accuracy of SRS and ESRS is similar to that of the risk of recurrence of stroke , and the accuracy of the overall prediction of the SRS is slightly higher than that of the ESRS . The difference between the overall difference of the SRS and ESRS is slightly higher than that of the ESRS . The difference of the ABCD2 scores is poor . The conclusion of the relevant results is confirmed by further prospective studies .

【学位授予单位】：复旦大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：R743.3

【参考文献】