急性主动脉夹层诊断和预警模型构建
发布时间：2019-06-01 11:20
【摘要】：研究目的:急性主动脉夹层(acute aortic dissection,AAD)是一种发病突然、进展迅速、致死率高的心血管疾病;早期诊断并进行合理的治疗是降低急性主动脉夹层患者死亡率的关键。自1761年Morgagni第一次报道以来,急性主动脉夹层的诊疗水平不断进步,但其临床误诊率仍然高达39%。目前急性主动脉夹层的确诊主要是根据影像学检查结果的单一模式,缺乏有针对性的实验室检验指标来辅助诊断。肝肾功能、血常规、凝血功能等临床最常用检验项目具有检查迅速、成本低等优点,且已在基层医疗中心广泛普及;更重要的是,上述检验项目中的血脂、白细胞计数、中性粒细胞计数、血小板计数、D-二聚体、纤维蛋白原等等众多指标已被众多研究证实在急性主动脉夹层患者循环血中发生明显改变。所以,本研究以急性主动脉夹层患者上述临床常用检验项目和部分病史信息为研究对象,寻找基于临床最常用实验室检验方法的,可用于急性主动脉夹层早期诊断和病情评估的实用工具。研究方法:本研究分为模型构建和模型评价两个环节。在模型构建环节,按照纳入和排除标准,纳入2014年10月至2016年8月于第二军医大学附属长海医院治疗的患者共392例,其中急性主动脉夹层患者230例(包括:急性期生存患者,n=175;急性期死亡患者,n=55;Stanford A型患者,n=113;Stanford B型患者,n=113;因发病后迅速死亡未能确定Stanford分型者,n=4)和对照组患者162例(包括:急性心肌梗死[acute myocardial infarction,AMI],n=55;急性肺动脉栓塞[acute pulmonary embolism,APE],n=39;腹主动脉瘤[abdominal aortic aneurysm,AAA],n=68)。首先我们将急性主动脉夹层患者的血液检验指标分别与上述3组对照组患者的血液检验指标进行单因素变量分析,同时也根据急性主动脉夹层患者的急性期死亡情况和Stanford分型进行亚组间的单变量分析。通过上述分析过程,确定各组及各亚组间具有明显差异的血液学检验指标。然后将上述差异指标纳入多变量分析,确定每个指标在多变量分析中的回归系数、标准误、Wald卡方值、P值、其对应的OR值及其95%置信区间。最后将多变量分析的结果带入R软件,制作各对照组和亚组的列线图,得到急性主动脉夹层与急性心肌梗死鉴别诊断模型(diagnostic model between AAD and AMI,DMDI)、急性主动脉夹层与急性肺动脉栓塞鉴别诊断模型(diagnostic model between AAD and APE,DMDE)、急性主动脉夹层与腹主动脉瘤鉴别诊断模型(diagnostic model between AAD and AAA,DMDA)、Stanford A、B型急性主动脉夹层鉴别诊断模型(diagnostic model between Stanford A AAD and Stanford B AAD,DMAB)、急性主动脉夹层患者急性期死亡预警模型(modle of death risk judgment for AAD sufferer,MDRJ)。在模型评价环节,纳入同期在第二军医大学附属长海医院治疗的因患者159例,其中急性主动脉夹层患者60例(包括:急性期生存患者,n=34;急性期死亡患者,n=26;stanforda型患者,n=32;stanfordb型患者,n=28)和对照组患者99例(包括:急性心肌梗死[acutemyocardialinfarction,ami],n=36;急性肺动脉栓塞[acutepulmonaryembolism,ape],n=20;腹主动脉瘤[abdominalaorticaneurysm,aaa],n=43)。我们将上述不同分组患者临床检验指标带入上述模型,利用相应模型计算各组患者预测分数,利用诊断试验的评价方法,计算上述五个模型的准确度、灵敏度和特异度,综合评价上述五个模型。结果:(1)通过对急性主动脉夹层患者和急性心肌梗死患者的临床检验指标和病史资料进行单因素和多因素分析,最终筛选患者年龄(age)、高密度脂蛋白(hdl)、血小板计数(plt)、d-二聚体(d-dimer)、纤维蛋白原(fib)、中性粒细胞计数与淋巴细胞计数比值(neutrophiltolymphocyteratio,nlr)、甘油三酯与高密度脂蛋白比值(triglyceride/hdl,tg/hdl)纳入急性主动脉夹层与急性心肌梗死鉴别诊断模型;经验证该模型准确率=80.49%,灵敏度=74.29%,特异度=85.11%。(2)通过对急性主动脉夹层患者和急性肺动脉栓塞患者的临床检验指标和病史资料进行单因素和多因素分析,最终筛选患者年龄(age)、吸烟史(smoke)、白细胞计数(wbc)、单核细胞计数(mono)、中性粒细胞计数(gran)和凝血酶原时间(pt)纳入急性主动脉夹层与急性肺动脉栓塞鉴别诊断模型;经验证该模型准确率=66.18%,灵敏度=61.90%,特异度=68.09%。(3)通过对急性主动脉夹层患者和腹主动脉瘤患者的临床检验指标和病史资料进行单因素和多因素分析,最终筛选患者年龄(age)、胆固醇(cholesterol)、甘油三酯(triglyceride)、低密度脂蛋白(ldl)和中性粒细胞计数(gran)纳入急性主动脉夹层与腹主动脉瘤鉴别诊断模型;经验证该模型准确率=78.31%,灵敏度=72.22%,特异度=82.98%。(4)通过对急性主动脉夹层stanforda型患者和stanfordb患者的临床检验指标和病史资料进行单因素和多因素分析,最终筛选患者年龄(age)、中性粒细胞计数(gran)、d-二聚体(d-dimer)、纤维蛋白原(fib)和活化部分凝血活酶时间(aptt)纳入stanforda、b型急性主动脉夹层鉴别诊断模型;经验证该模型准确率=72.34%,灵敏度=76.19%,特异度=69.23%。(5)通过对急性主动脉夹层急性期生存患者和死亡患者的临床检验指标和病史资料进行单因素和多因素分析,最终筛选白细胞计数(wbc)、中性粒细胞比值(neut_ratio)、血小板计数(plt)、d-二聚体(d-dimer)、患者年龄(age)纳入急性主动脉夹层患者急性期死亡预警模型;经验证该模型准确率=76.59%,灵敏度=78.13%,特异度=73.33%。结论:通过本研究,我们构建了急性主动脉夹层与急性心肌梗死鉴别诊断模型、急性主动脉夹层与急性肺动脉栓塞鉴别诊断模型、急性主动脉夹层与腹主动脉瘤鉴别诊断模型、Stanford A、B型急性主动脉夹层鉴别诊断模型、急性主动脉夹层患者急性期死亡预警模型;上述模型可以协助临床医师对急性主动脉夹层患者进行快速诊断和病情评估以及急性期死亡风险评估;对进一步完善目前急性主动脉夹层的诊断模式和丰富急性主动脉夹层患者的病情评估方法具有重要意义。
[Abstract]:The purpose of this study is that the acute aortic dissection (AAD) is a kind of cardiovascular disease with rapid onset, rapid progress and high fatality rate, and the early diagnosis and reasonable treatment is the key to reduce the mortality of the patients with acute aortic dissection. Since the first report of Morgagni in 1761, the level of diagnosis and treatment of acute aortic dissection has advanced, but its clinical misdiagnosis rate is still as high as 39%. At present, the diagnosis of the acute aortic dissection is mainly based on the single pattern of the results of the imaging examination and the lack of targeted laboratory test indicators to assist in the diagnosis. The clinical most common test items such as the liver and kidney function, the blood routine, the blood coagulation function and the like have the advantages of rapid examination, low cost and the like, and have been widely popularized in the basic medical center; and more importantly, the blood fat, the white blood cell count, the neutrophil count and the platelet count in the above test items are more important, A number of indicators such as D-dimer, fibrinogen and so on have been demonstrated to have a significant change in circulating blood in patients with acute aortic dissection. Therefore, this study is based on the clinical common test items and some medical history information of the patients with acute aortic dissection as the study object, and the utility of the early diagnosis and evaluation of the acute aortic dissection can be found based on the most common laboratory test methods. The research method: this study is divided into two parts: model construction and model evaluation. In the model building,392 patients with acute aortic dissection (including patients with acute stage survival, n = 175, and n = 55) were included in the second military medical university in August 2014 to August 2016, in accordance with the inclusion and exclusion criteria. A type of Stanford A patient, n = 113; a Stanford B-type patient, n = 113; and 162 of the patients in the control group (including: acute myocardial infarction[acute myoctal infraction, AMI], n = 55; acute pulmonary embolism[acute pulmonary, APE], n = 39; abdominal aortic aneurysm[abdominitic anaplastic, AAA],n=68). First, we compare the blood test index of the patients with acute aortic dissection with the blood test index of the 3-group control group, and the single-factor analysis between the sub-groups according to the acute death of the acute aortic dissection and the Stanford type. Hematology test indicators with significant differences between the groups and subgroups were determined by the above analysis process. And then the difference index is included in the multivariate analysis to determine the regression coefficient, the standard error, the Wald card square value, the P value, the corresponding OR value and the 95% confidence interval of each index in the multivariate analysis. and finally, the results of the multi-variable analysis are carried into the R software, and a column chart of each control group and a subgroup is manufactured to obtain an acute aortic dissection and an acute myocardial infarction differential diagnosis model (DMDI), An acute aortic dissection and an acute pulmonary embolism differential diagnosis model (DMDE), an acute aortic dissection and an abdominal aortic aneurysm differential diagnosis model (DMDA), Stanford A, Model of acute aortic dissection (AD and Stanford B AAD, DMAB) and the model of death early-warning for acute aortic dissection (MDRJ). In the model evaluation,159 patients with acute aortic dissection (including patients with acute stage survival, n = 34, patients with acute phase death, n = 26; stanford type), n = 32; stanford db, were included in the second military medical university attached to the long-sea hospital in the same period. N = 28) and in the control group,99 (including: acute myocardial infarction[acutumardihalation, ami], n = 36; acute pulmonary embolism[acutepulmonarymal, ape], n = 20; abdominal aortic aneurysm[abdominalaoricaneureysm, aaa], n = 43). In that model, the clinical examination index of the different group of patients is carry into the model, and the prediction score of each group is calculated by using the corresponding model, and the accuracy, the sensitivity and the specificity of the five models are calculated by using the evaluation method of the diagnosis test, and the five models are comprehensively evaluated. Results: (1) The age (age), high density lipoprotein (hdl) and platelet count (plt) of patients with acute aortic dissection and patients with acute myocardial infarction were analyzed by single factor and multi-factor analysis. The ratio of d-dimer (d-dimer), fibrinogen (fib), neutrophil count and lymphocyte count (nlr), triglyceride to high-density lipoprotein (tg/ hdl) was included in the differential diagnosis model of acute aortic dissection and acute myocardial infarction. The accuracy of the model was 80.49%, the sensitivity was 74.29%, and the specificity was 85.11%. (2) The patient age (age), smoking history (smoke), white blood cell count (wbc), and monocytic count (mono) were selected by single-factor and multi-factor analysis of the clinical and medical history data of patients with acute aortic dissection and acute pulmonary embolism. Neutrophil count (gran) and prothrombin time (pt) were included in the diagnosis model of acute aortic dissection and acute pulmonary embolism. The accuracy of the model was 66.18%, the sensitivity was 61.90%, and the specificity was 68.09%. and (3) finally screening the age (age), the cholesterol (cholesterol) and the triglyceride of the patients by performing single-factor and multi-factor analysis on the clinical examination indexes and the medical history data of the patients with the acute aortic dissection and the abdominal aortic aneurysm, Low density lipoprotein (ldl) and neutrophil count (gran) were included in the diagnosis model of acute aortic dissection and abdominal aortic aneurysm. The accuracy of the model was 78.31%, the sensitivity was 72.22%, and the specificity was 82.98%. (4) The patient age (age), the neutrophil count (gran), and the d-dimer (d-dimer) were selected by single-factor and multi-factor analysis of the clinical and medical history data of the patients with an acute aortic dissection and the stanford-type patient. Fibrinogen (fib) and activated partial thromboplastin time (aptt) were included in the differential diagnosis model of the type b acute aortic dissection. The accuracy of the model was 72.34%, the sensitivity was 76.19%, and the specificity was 69.23%. (5) By single-factor and multi-factor analysis of the clinical examination index and the medical history data of the patients with acute aortic dissection and the patients with death, the white blood cell count (wbc), the neut _ ratio and the platelet count (plt) were selected. D-dimer (d-dimer) and age of the patient were included in the early-stage death-warning model of acute aortic dissection; the accuracy of the model was 76.59%, the sensitivity was 78.13%, and the specificity was 73.33%. Conclusion: Through this study we constructed the diagnosis model of acute aortic dissection and acute myocardial infarction, the diagnosis model of acute aortic dissection and acute pulmonary embolism, the diagnosis model of acute aortic dissection and abdominal aortic aneurysm, Stanford A, The model can be used to assist the clinician in the rapid diagnosis and assessment of the acute aortic dissection and the assessment of the risk of death in the acute stage. It is of great significance to further improve the diagnosis model of acute aortic dissection and the method of evaluating the condition of patients with acute aortic dissection.
【学位授予单位】：第二军医大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R543.1
本文编号：2490232
[Abstract]:The purpose of this study is that the acute aortic dissection (AAD) is a kind of cardiovascular disease with rapid onset, rapid progress and high fatality rate, and the early diagnosis and reasonable treatment is the key to reduce the mortality of the patients with acute aortic dissection. Since the first report of Morgagni in 1761, the level of diagnosis and treatment of acute aortic dissection has advanced, but its clinical misdiagnosis rate is still as high as 39%. At present, the diagnosis of the acute aortic dissection is mainly based on the single pattern of the results of the imaging examination and the lack of targeted laboratory test indicators to assist in the diagnosis. The clinical most common test items such as the liver and kidney function, the blood routine, the blood coagulation function and the like have the advantages of rapid examination, low cost and the like, and have been widely popularized in the basic medical center; and more importantly, the blood fat, the white blood cell count, the neutrophil count and the platelet count in the above test items are more important, A number of indicators such as D-dimer, fibrinogen and so on have been demonstrated to have a significant change in circulating blood in patients with acute aortic dissection. Therefore, this study is based on the clinical common test items and some medical history information of the patients with acute aortic dissection as the study object, and the utility of the early diagnosis and evaluation of the acute aortic dissection can be found based on the most common laboratory test methods. The research method: this study is divided into two parts: model construction and model evaluation. In the model building,392 patients with acute aortic dissection (including patients with acute stage survival, n = 175, and n = 55) were included in the second military medical university in August 2014 to August 2016, in accordance with the inclusion and exclusion criteria. A type of Stanford A patient, n = 113; a Stanford B-type patient, n = 113; and 162 of the patients in the control group (including: acute myocardial infarction[acute myoctal infraction, AMI], n = 55; acute pulmonary embolism[acute pulmonary, APE], n = 39; abdominal aortic aneurysm[abdominitic anaplastic, AAA],n=68). First, we compare the blood test index of the patients with acute aortic dissection with the blood test index of the 3-group control group, and the single-factor analysis between the sub-groups according to the acute death of the acute aortic dissection and the Stanford type. Hematology test indicators with significant differences between the groups and subgroups were determined by the above analysis process. And then the difference index is included in the multivariate analysis to determine the regression coefficient, the standard error, the Wald card square value, the P value, the corresponding OR value and the 95% confidence interval of each index in the multivariate analysis. and finally, the results of the multi-variable analysis are carried into the R software, and a column chart of each control group and a subgroup is manufactured to obtain an acute aortic dissection and an acute myocardial infarction differential diagnosis model (DMDI), An acute aortic dissection and an acute pulmonary embolism differential diagnosis model (DMDE), an acute aortic dissection and an abdominal aortic aneurysm differential diagnosis model (DMDA), Stanford A, Model of acute aortic dissection (AD and Stanford B AAD, DMAB) and the model of death early-warning for acute aortic dissection (MDRJ). In the model evaluation,159 patients with acute aortic dissection (including patients with acute stage survival, n = 34, patients with acute phase death, n = 26; stanford type), n = 32; stanford db, were included in the second military medical university attached to the long-sea hospital in the same period. N = 28) and in the control group,99 (including: acute myocardial infarction[acutumardihalation, ami], n = 36; acute pulmonary embolism[acutepulmonarymal, ape], n = 20; abdominal aortic aneurysm[abdominalaoricaneureysm, aaa], n = 43). In that model, the clinical examination index of the different group of patients is carry into the model, and the prediction score of each group is calculated by using the corresponding model, and the accuracy, the sensitivity and the specificity of the five models are calculated by using the evaluation method of the diagnosis test, and the five models are comprehensively evaluated. Results: (1) The age (age), high density lipoprotein (hdl) and platelet count (plt) of patients with acute aortic dissection and patients with acute myocardial infarction were analyzed by single factor and multi-factor analysis. The ratio of d-dimer (d-dimer), fibrinogen (fib), neutrophil count and lymphocyte count (nlr), triglyceride to high-density lipoprotein (tg/ hdl) was included in the differential diagnosis model of acute aortic dissection and acute myocardial infarction. The accuracy of the model was 80.49%, the sensitivity was 74.29%, and the specificity was 85.11%. (2) The patient age (age), smoking history (smoke), white blood cell count (wbc), and monocytic count (mono) were selected by single-factor and multi-factor analysis of the clinical and medical history data of patients with acute aortic dissection and acute pulmonary embolism. Neutrophil count (gran) and prothrombin time (pt) were included in the diagnosis model of acute aortic dissection and acute pulmonary embolism. The accuracy of the model was 66.18%, the sensitivity was 61.90%, and the specificity was 68.09%. and (3) finally screening the age (age), the cholesterol (cholesterol) and the triglyceride of the patients by performing single-factor and multi-factor analysis on the clinical examination indexes and the medical history data of the patients with the acute aortic dissection and the abdominal aortic aneurysm, Low density lipoprotein (ldl) and neutrophil count (gran) were included in the diagnosis model of acute aortic dissection and abdominal aortic aneurysm. The accuracy of the model was 78.31%, the sensitivity was 72.22%, and the specificity was 82.98%. (4) The patient age (age), the neutrophil count (gran), and the d-dimer (d-dimer) were selected by single-factor and multi-factor analysis of the clinical and medical history data of the patients with an acute aortic dissection and the stanford-type patient. Fibrinogen (fib) and activated partial thromboplastin time (aptt) were included in the differential diagnosis model of the type b acute aortic dissection. The accuracy of the model was 72.34%, the sensitivity was 76.19%, and the specificity was 69.23%. (5) By single-factor and multi-factor analysis of the clinical examination index and the medical history data of the patients with acute aortic dissection and the patients with death, the white blood cell count (wbc), the neut _ ratio and the platelet count (plt) were selected. D-dimer (d-dimer) and age of the patient were included in the early-stage death-warning model of acute aortic dissection; the accuracy of the model was 76.59%, the sensitivity was 78.13%, and the specificity was 73.33%. Conclusion: Through this study we constructed the diagnosis model of acute aortic dissection and acute myocardial infarction, the diagnosis model of acute aortic dissection and acute pulmonary embolism, the diagnosis model of acute aortic dissection and abdominal aortic aneurysm, Stanford A, The model can be used to assist the clinician in the rapid diagnosis and assessment of the acute aortic dissection and the assessment of the risk of death in the acute stage. It is of great significance to further improve the diagnosis model of acute aortic dissection and the method of evaluating the condition of patients with acute aortic dissection.
【学位授予单位】：第二军医大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R543.1
【参考文献】
相关期刊论文 前3条
1 刘心甜;贺行巍;谭蓉;刘婉君;王贝;刘玉建;王涛;刘成伟;苏f^;曾和松;;High-density Lipoprotein Cholesterol and In-hospital Mortality in Patients with Acute Aortic Dissection[J];Journal of Huazhong University of Science and Technology(Medical Sciences);2016年03期
2 管珩,吴庆华;努力提高我国主动脉瘤的诊治水平[J];中华普通外科杂志;2005年01期
3 段志泉,罗英伟,张强,王春喜,董雨亭;65例腹主动脉瘤的诊断与治疗[J];中华普通外科杂志;2000年11期
，本文编号：2490232
本文链接：https://www.wllwen.com/yixuelunwen/xxg/2490232.html
最近更新
教材专著
