PICC相关性上肢静脉血栓风险评估模型的构建与初步评价

发布时间：2018-06-17 05:07

本文选题：PICC相关性血栓 + 危险因素　；参考：《天津医科大学》2017年硕士论文

【摘要】：目的探索和分析PICC相关性上肢静脉血栓形成的相关危险因素,构建PICC相关性上肢静脉血栓风险评估模型并对其进行评价,为临床PICC相关性血栓的评估、预防提供参考。同时,对二维超声及血流显像技术下难以确切诊断但临床考虑静脉血栓的患者实施超声造影,以探讨实施超声造影在PICC相关性血栓诊断中的应用价值。方法收集2015年4月-2017年1月于天津市某三级甲等综合医院留置PICC导管的所有符合纳入排除标准的患者为研究对象,并依据研究对象纳入的先后时间顺序分为训练样本和验证样本。从置管当天开始对研究对象实施为期一个月的B超随访。随访时间顺应PICC导管维护要求,每7天一次于导管换药前进行,共计随访4次。收集资料包括:患者一般资料、置管穿刺过程相关信息、身体疾病和病史相关信息、部分临床化验结果信息以及置管侧肢体临床表现状况。采用常规多普勒超声技术对患者上肢静脉的血栓形成状况进行诊断。对于常规多普勒超声难以判断者,实施超声造影检查并保存造影图像,以明确血栓形成状况。研究依据有无静脉血栓形成,将训练样本和验证样本的研究对象分为无血栓组和有血栓组,使用SPSS18.0软件对数据进行分析和处理。采用?2检验、t检验以及Mann-Whitney U检验对训练样本各因素变量进行分析,找出存在统计学差异的变量。然后,利用二分类Logistic回归分析筛选出独立危险因素并建立PICC相关性血栓风险评估模型,依据优势比对模型各因素赋值,并且对血栓形成风险度进行了分级。利用灵敏度、特异度以及ROC曲线分别在训练样本和验证样本中对模型进行验证和评价。最后,对PICC导管超声造影的图像资料进行分析,初步总结和分析超声造影在PICC相关性血栓中的应用价值。结果1.研究共纳入研究对象375例,训练样本275例,验证样本100例,共计发生血栓72例(训练样本53例,验证样本19例),静脉血栓总发生率为19.20%。2.训练样本单因素分析表明,与PICC相关性血栓形成有关的因素为肥胖与否、疾病诊断、血栓史、化疗史、主要输注内容物、导管尖端位置、CVC或PICC导管置入史、糖尿病、活动能力(KPS评分)以及D-二聚体(≥0.5mg/L)。3.采用Logistic回归对训练样本中有统计学差异的因素的进一步分析得出,PICC相关性血栓形成的独立危险因素包括:是否肥胖(OR=3.449,95%CI:1.634-7.493,P=0.001);糖尿病(OR=3.016,95%CI:1.342-6.775,P=0.008);血栓史(OR=2.887,95%CI:1.256-6.632,P=0.012);化疗史(OR=2.268,95%CI:1.070-4.808,P=0.033);D-二聚体(OR=2.283,95%CI:1.132-4.604,P=0.021)。4.依据Logistic回归分析的OR值(四舍五入法)对模型各独立因素进行风险积分赋值,分别为肥胖(BMI≥25kg/m2)3分;糖尿病3分;血栓史3分;化疗史2分,D-二聚体(≥0.5mg/L)2分。同时,建立的评估模型为:Y=肥胖×3+糖尿病×3+血栓史×3+化疗史×2+D-二聚体(≥0.5mg/L)×2-3.19。5.通过计算训练样本中有血栓组和无血栓组的95%置信区间,得出血栓形成的危险度分级:3分为低危级;3-5分为中危级;5分为高危级。6.训练样本和验证样本的ROC曲线分析:在训练样本中采用ROC曲线分析PICC相关性血栓风险评估模型的曲线下面积为0.753(95%CI:0.676-0.830),灵敏度为0.632,特异度为:0.797,最佳截断值为4.5分。在验证样本中,PICC相关性血栓风险评估模型的ROC曲线下面积为0.758(95%CI:0.684-0.883),灵敏度为0.784,特异度为:0.691,最佳截断值同样为4.5分。说明模型的预测效果呈中等水平。7.PICC超声造影的诊断效果及应用价值:超声造影检查可对存在彩色外溢、低流速下无明显血流信号、临床症状严重等PICC相关性血栓可疑者做出准确诊断。结论1.本研究得出肥胖、糖尿病、血栓史、化疗史以及D-二聚体(≥0.5mg/L)五项因素是影响PICC相关性血栓形成的独立危险因素。2.依据上述因素构建所的PICC相关性血栓风险评估模型具有较好的评估效果,并且危险分级明确。可在一定程度上对患者的静脉血栓发生风险进行评估。3.超声造影和常规超声在PICC相关性血栓的诊断上各有优势,对于临床怀疑PICC相关性血栓但常规超声不能确诊血栓形成状况者,可进一步实施超声造影检查,其过程安全,结果可靠,值得推广应用。
[Abstract]:Objective to explore and analyze the related risk factors of PICC related upper limb venous thrombosis, to construct and evaluate the risk assessment model of PICC related upper limb venous thrombosis, and to provide reference for the evaluation of clinical PICC related thrombus and prevention. The patients with thrombus were performed by contrast-enhanced ultrasound to explore the application value of contrast-enhanced ultrasound in the diagnosis of PICC related thrombus. Methods all the patients who had retained the PICC catheter in a class three class a general hospital in Tianjin, April 2015, -2017 years, were collected for the study of the image, and the time and the time of the study were included. The sequence was divided into training samples and validation samples. The subjects were followed up for one month from the day of catheterization. The follow-up time was followed by the PICC catheter maintenance requirements, every 7 days before the catheterization, a total of 4 times. The data included the general information of the patients, the related information of the catheterization process, the physical disease and the history of the disease. Relevant information, some clinical test results information and the clinical manifestation of the lateral limb of the tube. Conventional Doppler ultrasound technique was used to diagnose the condition of the thrombosis of the upper limb veins. For those who were difficult to judge the conventional Doppler ultrasound, the ultrasound examination and preservation of the image were carried out to clarify the status of thrombus formation. There were no venous thrombosis. The subjects of training samples and samples were divided into non thrombus group and thrombus group. The data were analyzed and processed using SPSS18.0 software. 2 test, t test and Mann-Whitney U test were used to analyze the variables of the training samples to find out the variables with statistical differences. Two Logistic regression analysis was used to screen out independent risk factors and establish a PICC related thrombus risk assessment model. According to the advantages of the model, the risk degree of thrombus formation was graded. The sensitivity, specificity and ROC curve were used to verify and evaluate the model in the training sample and the validation sample respectively. Finally, the image data of PICC catheter ultrasonography were analyzed, and the application value of ultrasound contrast in PICC related thrombosis was preliminarily summarized and analyzed. Results 1. the study included 375 subjects, 275 training samples and 100 samples, 72 cases of thrombosis (training samples 53 cases, 19 samples), and the total incidence of venous thrombosis. The single factor analysis for 19.20%.2. training samples showed that the factors related to PICC related thrombosis were obesity or not, disease diagnosis, thrombus history, chemotherapy history, main infusion content, catheter tip position, CVC or PICC catheter implantation history, diabetes, activity ability (KPS score) and D- two polymer (0.5mg/L).3. using Logistic regression to training Further analysis of factors with statistical differences found that independent risk factors for PICC related thrombosis include obesity (OR=3.449,95%CI:1.634-7.493, P=0.001); diabetes (OR=3.016,95%CI:1.342-6.775, P=0.008); thrombotic history (OR=2.887,95%CI:1.256-6.632, P=0.012); chemotherapy history (OR=2.268,95%CI:1.070-4.808, P=0.033). D- two polymer (OR=2.283,95%CI:1.132-4.604, P=0.021).4. was based on the OR value of Logistic regression analysis (four RMS and five entry method) to evaluate the risk integral of the model independent factors, including obesity (BMI > 25kg/m2) 3 points, diabetes 3 points, thrombus history 3; chemotherapy history 2, D- two polymer (> 0.5mg/L) 2. Meanwhile, the evaluation model was: Y= obesity * * 3+ diabetes mellitus x 3+ thrombosis history * 3+ chemotherapy history * 2+D- two polymer (> 0.5mg/L) x 2-3.19.5. through calculating the training samples with thrombus group and non thrombus group, the risk degree of thrombosis is classified: 3 is low risk grade; 3-5 is medium danger grade; 5 is the ROC curve analysis of high risk class.6. training sample and verification sample: in training sample The area under the curve of ROC curve analysis of PICC related thrombus risk assessment model was 0.753 (95%CI:0.676-0.830), the sensitivity was 0.632, the specificity was 0.797, and the best truncated value was 4.5. In the verified samples, the area of the ROC curve of the PICC related thrombus risk assessment model was 0.758 (95%CI:0.684-0.883) and the sensitivity was 0.784. The difference is: 0.691, the best truncation value is also 4.5 points. It shows that the prediction effect of the model is the diagnostic effect and the application value of the medium level.7.PICC sonography: the ultrasonic contrast examination can make the accurate diagnosis of the suspected PICC phase thrombus in the presence of color spillover, low flow rate and serious clinical symptoms. Conclusion 1. studies It is concluded that obesity, diabetes, thrombus history, chemotherapy history, and D- two polymer (> 0.5mg/L) are independent risk factors affecting the formation of PICC related thrombosis..2. based on the above factors has a better evaluation effect on the PICC related thrombus risk assessment model, and the risk classification is clear. To a certain extent, the patient can be used. The risk of venous thrombosis is evaluated by.3. ultrasound and conventional ultrasound in the diagnosis of PICC related thrombus. For the clinical suspected PICC related thrombus but the conventional ultrasound can not diagnose the thrombus formation, the ultrasound examination can be further carried out, the process is safe and the result is reliable. It is worth popularizing.
【学位授予单位】：天津医科大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R472

【参考文献】