住院AKI患者病死率危险因素分析及预警模型建立和应用

发布时间：2018-08-29 17:10

【摘要】：目的 1.研究住院AKI患者90天及1年病死率相关的危险因素。 2.建立住院AKI患者90天及1年预后预警模型评价住院AKI患者的临床预后。方法选取1996年1月到2013年4月中南大学湘雅二医院各科室的成人住院AKI患者共1169例,采用2012年改善全球肾脏病预后组织(KDIGO)颁布的诊断标准,1996年1月至2007年12月的731例构成试验组,2008年1月至2013年4月的438例构成验证组。记录入选患者的临床基本资料,分别进行90天及1年的随访,或以患者死亡为终点事件,统计90天及1年病死率。将试验组分为死亡组和存活组,比较两组的参数差异,进一步应用多因素Logistic回归分析分别确定AKI患者90天及1年死亡的独立危险因素,根据各危险因素对应的OR值按照四舍五入的方法赋予其相应积分,计算各病例的总评分,计算各评分所对应的病死率。对各评分所对应的病死率行卡方趋势性检验及作ROC曲线,秩和检验评价试验组与验证组死亡率的拟合度。分别建立90天及1年预后的预警模型,在验证组中进行初步应用,利用CMHχ2检验(cochran mantel haeszel statistics)法验证预警模型的预测性。结果 1.试验组与验证组的90天病死率分别为13.8%、11.6%；1年病死率分别为14.8%、12.6%。 2.单因素分析显示90天及1年预后中年龄、AKI类型、AKI病因、机械通气、低血压、休克、心衰、呼衰、胃肠道衰竭、中枢神经系统衰竭、BUN峰值、K+峰值、ATN-ISS评分方面具有显著统计学差异。多因素Logistic回归分析,确定其90天死亡相关的危险因素,发现年龄、AKI类型、呼吸衰竭、中枢神经系统衰竭、低血压、ATNISI评分0.4为90天死亡相关的独立危险因素。确定其1年死亡相关危险因素,发现年龄、AKI类型、呼吸衰竭、中枢神经系统衰竭、低血压为1年死亡相关的独立危险因素。 3.使用卡方趋势性检验对90天及1年预后评分系统进行趋势检验,显示病死率的变化有统计学意义(P0.001),预测90天及1年病死率的ROC曲线下面积分别为0.833(95%CI:0.788～0.879), P0.001;0.817(95%CI:0.771～0.864), P0.001 4.住院AKI患者90天及1年预后预警模型中试验组与验证组的病死率均无统计学差异(χ2=1.7958,P=0.1802;χ2=0.1006, P=0.7511),显示这两个预警模型对AKI的病死率均具有良好的预测能力。结论 1.住院AKI患者的病死率为11.6%-14.8%。 2.年龄、AKI类型、呼吸衰竭、中枢神经系统衰竭、低血压、ATNISI评分0.4为住院AKI患者死亡相关的独立危险因素。 3.本研究建立了两个住院AKI患者死亡的预警模型,分数越高,病死率越高,预测价值均良好,有助于临床医师早期识别高危患者。
[Abstract]:Objective 1. To study the risk factors associated with 90 days and 1 year mortality in AKI patients. 2. A 90-day and 1-year prognostic early warning model was established to evaluate the clinical prognosis of hospitalized AKI patients. Methods from January 1996 to April 2013, 1169 adult AKI patients in Xiangya second Hospital of Central South University were selected. According to the diagnostic criteria issued by (KDIGO) for improving the prognosis of Nephropathy in 2012, 731 cases from January 1996 to December 2007 constituted a trial group, and 438 cases from January 2008 to April 2013 constituted a validation group. The basic clinical data were recorded, followed up for 90 days and 1 year, or the terminal events were the death of the patients, and the mortality of 90 days and 1 year were counted. The experimental group was divided into death group and survival group. The parameters of the two groups were compared, and the independent risk factors for 90 days and 1 year death of AKI patients were determined by multivariate Logistic regression analysis. According to the corresponding OR value of each risk factor according to the rounding method the corresponding integral was given to calculate the total score of each case and calculate the case fatality rate corresponding to each score. The mortality corresponding to each score was tested by chi-square trend test, ROC curve and rank sum test to evaluate the fitness of mortality between the test group and the validation group. The early warning models of 90 days and 1 year prognosis were established and applied in the validation group. CMH 蠂 2 test was used to verify the predictive value of the early warning model by (cochran mantel haeszel statistics) method. Result 1. The 90-day mortality of the test group and the validation group were 13.8and 11.60.The 1-year mortality was 14.8and 12.6, respectively. Univariate analysis showed significant difference in age type AKI etiology, mechanical ventilation, hypotension, shock, heart failure, respiratory failure, gastrointestinal failure, central nervous system failure, bun peak K peak and ATN-ISS score in 90 days and 1 year prognosis. Multivariate Logistic regression analysis showed that age type respiratory failure central nervous system failure and hypotension score 0. 4 were independent risk factors for 90 days death. The risk factors associated with one year death were determined. Age AKI type, respiratory failure, central nervous system failure and hypotension were found to be independent risk factors for 1 year death. The trend test of 90 days and 1 year prognostic scoring system by chi-square trend test showed that the change of mortality was statistically significant (P0. 001). The area under the ROC curve for predicting the mortality of 90 days and 1 year was 0.833 (95%CI:0.788~0.879), P0.001 + 0.817 (95%CI:0.771~0.864) and P0.001 4, respectively. There was no significant difference in mortality between the experimental group and the validation group in 90 days and 1 year prognostic early warning model of AKI patients in hospital (蠂 ~ (2 +) 1.7958 (P ~ (0.1802); 蠂 ~ (2 +) 0.1006, P ~ (0.7511), which indicated that the two early warning models had a good predictive ability to the mortality of AKI. Conclusion 1. The fatality rate of hospitalized patients with AKI was 11.6- 14.8. Age AKI type, respiratory failure, central nervous system failure, hypotension and ATNISI score 0.4 were independent risk factors for death in hospitalized AKI patients. In this study, two early warning models of death in AKI patients were established. The higher the score, the higher the mortality rate, and the better the predictive value was, which was helpful for clinicians to identify high-risk patients early.
【学位授予单位】：中南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：R692.5

【参考文献】