浙江省终末期肾病患者流行病学调查及死亡风险预测模型构建

发布时间：2018-07-04 11:55

本文选题：透析登记 + 终末期肾脏病　；参考：《浙江大学》2015年博士论文

【摘要】：第一部分浙江省新增终末期肾病患者流行病学调查分析目的：透析登记系统是终末期肾病(end stage renal disesase, ESRD)患者规范化管理的重要组成部分。浙江省透析质量控制中心(Zhejiang Dialysis Quality and Management Center, ZDQM)建立于2007年,数据库主要登记内容包括透析患者的基本信息、病情发展与转归、药物使用情况及化验室指标等。本研究对2008～2013年ZDQM中新增ESRD患者的资料进行分析,了解其流行病学现状,以期为临床医师、科研人员及医疗卫生政策的制定与修改提供参考数据。方法：收集2008年1月至2013年12月,ZDQM数据库中新增ESRD患者的基本信息,包括性别、年龄、透析方式、原发病构成、转归情况、实验室数据等。对其年发病率、原发病构成及变迁、性别年龄结构及死亡原因等进行统计分析。结果：2008～2013年浙江省新增ESRD透析患者共26310例(其中维持性血液透析(maintenance hemodialysis, MHD)患者19143例,维持性腹膜透析(maintenance peritoneal dialysis, MPD)患者5199例),年发病率从46.3/百万人上升至119.9/百万人；平均年龄从53.0岁上升至58.7岁,65岁以上的患者比例从25.1%上升到37.8%；男性患者略多于女性(1.3：1)；原发病诊断前三位依次是慢性肾小球肾炎(chronicglomerulonephritis, CGN)(占全体51.3%),糖尿病肾病(diabetic nephropathy, DN)(占全体17.3%)和高血压肾病(hypertensive nephrosclerosis, HTN)(占全体6.4%),DN呈现缓慢上升趋势,登记的死亡原因中,心血管事件排首位。结论：2008～2013年浙江省ESRD年发病率持续上升,透析方式以血透治疗为主,腹透治疗有所增加,老年患者比例逐年上升,原发病中以慢性肾小球肾炎为主,糖尿病肾病呈现缓慢上升的趋势,心血管事件为ESRD透析患者的主要死因。第二部分倾向性评分法在血液透析和腹膜透析患者生存比较中的应用目的：终末期肾病(end stage renal disease, ESI ID)患者采用何种透析方式生存预后更好,目前尚存争议,本研究利用倾向性评分法(propensity score method, PSM)将浙江省新发ESRD透析患者的基本资料进行匹配,比较血液透析(hemodialysis, HD)和腹膜透析(peritoneal dialysis, PD)患者的生存情况。方法：研究对象为2008年7月1日至2013年6月30日,浙江省透析质量控制中心(Zhejiang Dialysis Quality and Management Center, ZDQM)新接收,年龄≥18岁、透析时间90天的ESRD患者,收集患者人口统计学资料、诊断、血管通路等临床基线信息,随访至2014年6月30日,随访期间发生死亡作为观察的最终结局。将PD患者的性别、年龄、原发病及合并症作为参照,多因素Logistic回归计算出倾向性评分,应用PSM,随机地在HD组中匹配具有近似评分的患者,再通过Kaplan-Meier法、 log-rank检验及Cox比例风险回归模型对患者的生存情况进行比较分析。结果：共纳入透析患者19846例,HD患者15610例(78.7%)、PD患者4236例(21.3%),中位随访时间为29个月(3-72个月)。应用PSM匹配出4233对患者,两组患者基线资料中,年龄、性别、原发病、合并症(除慢性肝病),均无显著性差异。Kaplan-Meier生存曲线显示在PSM之前,PD患者生存率高于HD患者(log-rank检验,P0.001),PSM后,两组患者总体生存率差异无统计学意义(log-rank检验,P=0.979),在开始透析后1年内PD患者生存率优于HD患者(P0.05)。亚组分析结果显示：1)在性别、年龄、原发病、合并症各亚组中,两种透析方式生存率无显著性差异(log-rank检验,P0.05)；2)在年龄65岁的患者亚组中,两组患者生存率无明显差异,其中合并糖尿病(diabetes mellitus, DM)的患者,HD组生存率高于PD组(log-rank检验,P0.05),非DM患者,PD组生存率高于HD组(log-rank检验,P0.05)；3)年龄≥65岁的患者亚组中,无论合并DM与否,两组患者生存率无显著性差异(log-rank检验,P0.05)；4)多因素Cox回归分析结果显示,年龄、原发病、心血管疾病、恶性肿瘤及慢性阻塞性肺部疾病(chronic obstructive pulmonary disease, COPD)为透析患者生存预后的独立危险因素,其中原发病为慢性肾小球肾炎(chronic glomerulonephritis, CGN)和高血压肾病(hypertensive nephrosclerosis, HTN)的患者生存率明显高于糖尿病肾病(diabetic nephropathy, DN)患者(P0.05)；5)使用动静脉内瘘(arteriovenous fistula,AVF)作为血管通路的HD患者生存率高于PD患者及使用中心静脉导管的HD患者(log-rank检验,P0.001)。结论：在新发ESRD患者中,HD和PD患者总体生存率无显著性差异,透析龄≤1年的PD患者生存率高于HD患者。年龄≥65岁的患者中,无论合并DM与否,HD治疗与PD治疗生存率无明显差异；年龄65岁的患者中,合并DM者,HD治疗生存率高于PD治疗,未合并DM者,PD生存率高于HD。年龄、原发病、心血管疾病、恶性肿瘤及COPD是影响透析患者生存率的独立危险因素,原发病中DN患者的死亡风险最高,新发ESRD患者使用AVF的HD患者生存率明显高于PD患者和使用中心静脉导管的HD患者,PD治疗可能会增加老年患者的死亡风险。第三部分维持性血透患者死亡风险预测及评分模型构建目的：血液透析(hemodialysis, HD)患者死亡风险明显高于普通人群,为提高临床医师对其死亡风险的预判能力,本研究通过临床易于获取的基本资料和实验室数据,对新发维持性血透(maintenance hemodialysis, MHD)患者透析后2年内的死亡风险进行预测,同时构建相应的风险评分标准。方法：纳入浙江省透析质量控制中心(Zhejiang Dialysis Quality and Management Center, ZDQM)2008年1月1日到2012年6月30日,年龄≥18岁、透析时间90天的新增MHD患者,收集患者开始透析时的人口统计学资料、诊断、血管通路、实验室数据等信息,所有患者随访至2014年6月30日,随访期间发生死亡作为观察的最终结局。将患者随机分为构建模型的训练组(占总体60%)和验证模型效度的验证组(占总体40%),利用训练组的临床基本资料和实验室数据,采用多因素logistic回归统计方法筛选出与患者透析开始后2年内全因死亡有关的独立影响因素构建模型,验证组通过受试者工作特征(Receiver operator characteristic,ROC)曲线下面积、敏感度及特异度,来反映预测模型的鉴别效度及符合程度；确立风险预测模型后,根据模型中各危险因素的权重系数及其变量类型,共同建立评分标准。结果：共纳入4295例患者,2年内死亡691人(占总数16.1%),多因素logistic回归分析结果显示年龄、原发病、血管通路、恶性肿瘤、血清白蛋白及血总钙是患者死亡的独立影响因素,其中血清白蛋白为保护性因素；确立模型公式=0.614*年龄(=1,2,3,4,或5)+0.864*血管通路(=0,或1)+(0.784*(若糖尿病肾病(diabetic nephropathy, DN)=1),或0.217(若高血压肾病(hypertensive nephrosclerosis, HTN)=1),或0.796(若原发病诊断为其他及未知=1))+0.709*恶性肿瘤(=0,或1)-0.554*血清白蛋白(=0,或1)+0.270*血总钙(=0,1,或2)-4.943；训练组ROC曲线下面积为0.767,95%CI(0.744～0.790),H-L检验,χ2=3.144,P=0.925,验证组验证,ROC曲线下面积为0.732,95%CI(0.700～0.765),敏感度和特异度分别为71.5%和64.2%；评分标准：年龄30岁：3分,每增加15岁增加3分,75岁为15分；原发病：慢性肾小球肾炎(chronic glomerulonephritis, CGN):0分,DN：4分,HTN：1分,其他：4分；血管通路：动静脉内瘘(arteriovenous fistula, AVF):0分,中心静脉导管：4分；合并症：恶性肿瘤：3分；实验室数据：血白蛋白：35g/1：0分,35g/1：-3分；血总钙：2.1mmol/1:1分,2.1-2.6mmol/1:0分,2.6mmol/l:3分。按最终评分分为：低危组(9分),中危组(10～13分),高危组(14～17分)和极高危组(≥18分)。结论：本研究回顾性分析ZDQM数据库中2008～2012年新发MHD患者信息,利用临床易于获得的患者基本资料和实验室数据,构建出MHD患者透析后2年内的死亡风险预测模型,并设立了相应的评分标准,经验证具有良好的鉴别度和准确性,在今后临床科研相关人员预判患者生存预后和制定临床决策的过程中具有一定的参考价值。
[Abstract]:Part one epidemiological analysis of newly diagnosed end-stage renal disease in Zhejiang Province
Objective: the dialysis registration system is an important part of the standardized management of patients with end stage renal disesase (ESRD). The Zhejiang dialysis quality control center (Zhejiang Dialysis Quality and Management Center, ZDQM) is established in 2007. The main contents of the database include the basic information of the dialysis patients and the development of the disease. In this study, the data of the new ESRD patients in the 2008~2013 year ZDQM were analyzed to understand the epidemiological status in order to provide reference data for the clinicians, scientific researchers and medical health policies.
Methods: the basic information of new ESRD patients in ZDQM database from January 2008 to December 2013 was collected, including sex, age, dialysis mode, primary pathogenesis, prognosis, laboratory data and so on. The annual incidence, the Constitution and changes of the primary disease, the gender age structure and the cause of death were analyzed.
Results: in 2008~2013 years, there were 26310 new ESRD dialysis patients in Zhejiang province (including 19143 patients with maintenance hemodialysis, MHD) and 5199 cases of maintenance peritoneal dialysis (maintenance peritoneal dialysis, MPD). The annual incidence rate rose from 46.3/ million to 119.9/ million; the average age increased from 53 years. The proportion of patients aged 58.7 and above rose from 25.1% to 37.8%; male patients were slightly more than women (1.3:1); the first three were chronic glomerulonephritis (chronicglomerulonephritis, CGN) (51.3%), diabetic nephropathy (diabetic nephropathy, DN) (all 17.3%) and hypertensive nephropathy (hypertensive nephro). Sclerosis (HTN) (accounting for 6.4% of all), DN showed a slow upward trend, and cardiovascular events ranked first among registered deaths.
Conclusion: in the past 2008~2013 years, the incidence of ESRD in Zhejiang province continued to rise. Hemodialysis was the main method of dialysis, the treatment of peritoneal dialysis was increased, the proportion of elderly patients increased year by year. The primary disease was chronic glomerulonephritis, and diabetic nephropathy showed a slow rising trend. The main cause of ESRD dialysis patients was the heart blood tube event.
The second part is the application of propensity score in hemodialysis and peritoneal dialysis patients.
Objective: the survival of patients with end-stage renal disease (end stage renal disease, ESI ID) has a better survival prognosis and is still in dispute. This study uses the tendency score (propensity score method, PSM) to match the basic data of the new ESRD dialysis patients in Zhejiang Province, and compare the hemodialysis (hemodialysis, hemodialysis) and peritoneal dialysis. An analysis of the survival of patients (peritoneal dialysis, PD).
Methods: from July 1, 2008 to June 30, 2013, the Zhejiang dialysis quality control center (Zhejiang Dialysis Quality and Management Center, ZDQM) was newly received, ESRD patients aged over 18 years and dialysis time for 90 days were collected, and the patient's demographic data, diagnosis, vascular access and other clinical baseline information were collected and followed up to 3 in June 2014. 0 days, during the follow-up period, death was the final outcome of the observation. The gender, age, primary disease and complication of PD patients were taken as reference, the tendency score was calculated by multiple factor Logistic regression, and PSM was used to match the patients with approximate score in the HD group, and then the Kaplan-Meier, log-rank test and Cox proportional risk regression model were used. The survival of the patients was compared and analyzed.
Results: 19846 cases of dialysis patients were included, 15610 cases (78.7%) of HD patients, 4236 cases (21.3%) of PD patients and 29 months (3-72 months). PSM was used to match 4233 pairs of patients. In the baseline data of the two groups, age, sex, primary disease, and complication (except chronic liver disease), there was no significant difference in the.Kaplan-Meier survival curve in PSM Before, the survival rate of PD patients was higher than that of HD (log-rank test, P0.001). After PSM, there was no significant difference in the total survival rate between the two groups (log-rank test, P=0.979). The survival rate of PD patients was superior to HD within 1 years after beginning dialysis (P0.05). The subgroup analysis showed 1) two kinds of dialysis in the sex, age, primary disease, and the subgroups of the complication. There was no significant difference in the survival rate (log-rank test, P0.05); 2) there was no significant difference in survival rate between the two groups of patients aged 65 years old, and the survival rate of group HD with diabetes mellitus, DM was higher than that of the PD group (log-rank test, P0.05), non DM patients, and the PD group survival rate was higher than that of the HD group (3); In the subgroup of patients aged more than 65 years old, no matter the combination of DM or not, there was no significant difference in survival rate between the two groups (log-rank test, P0.05), and 4) multiple factor Cox regression analysis showed that age, primary disease, cardiovascular disease, malignant tumor and chronic obstructive pulmonary disease (chronic obstructive pulmonary disease, COPD) were the survival precondition of dialysis patients. The postoperative independent risk factors, of which patients with chronic glomerulonephritis, CGN, and hypertensive nephrosclerosis, HTN, were significantly higher than patients with diabetic nephropathy (diabetic nephropathy, DN) (P0.05); 5) use arteriovenous fistula (arteriovenous fistula) as a blood vessel. The survival rate of HD patients was higher than that of PD patients and HD patients using central venous catheter (log-rank test, P0.001).
Conclusion: there is no significant difference in the overall survival rate of patients with HD and PD in the new ESRD patients. The survival rate of PD patients with dialysis age less than 1 years is higher than that of the HD patients. There is no significant difference in the survival rate between HD and PD in patients with age of 65 years old. The survival rate of HD is higher than that of PD, and the survival rate of HD is higher than that of PD, and the survival rate of HD is higher than that of PD treatment, and the survival rate is higher than that of PD treatment. And DM, the survival rate of PD is higher than that of HD. age. Primary disease, cardiovascular disease, malignant tumor and COPD are independent risk factors affecting the survival rate of dialysis patients. The mortality risk of DN patients in primary disease is the highest. The survival rate of HD patients using AVF in the new ESRD patients is significantly higher than those of PD patients and HD patients using central venous catheterization, PD treatment may increase. The risk of death in elderly patients.
The third part is the prediction of death risk and the construction of scoring model for maintenance hemodialysis patients.
Objective: the risk of death in hemodialysis (HD) patients was significantly higher than that of the general population, in order to improve the ability of clinicians to prejudge the risk of death. The risk of death within 2 years after dialysis in patients with new maintenance hemodialysis (MHD) was increased by clinical data and laboratory data. Make a forecast and build a corresponding risk scoring standard.
Methods: Zhejiang Dialysis Quality and Management Center, ZDQM) from January 1, 2008 to June 30, 2012, more than 18 years old and 90 days of dialysis time were added to the new MHD patients. The demographic data, diagnosis, vascular access, laboratory data, and so on were collected and all the patients were collected. The patients were followed up to June 30, 2014, and death was the final outcome of observation during follow-up. The patients were randomly divided into a training group (60%) and a validation group (40%) to verify the validity of the model. Using the clinical data of the training group and the laboratory data, the multiple factor logistic regression method was used to select the patients and the patients. The model was constructed for all the independent factors related to death within 2 years after the beginning of dialysis. The validation group reflected the identification validity and compliance of the prediction model through the area, sensitivity and specificity under the Receiver operator characteristic (ROC) curve, and the risk factors in the model were established, and the risk factors were based on the risk factors. The weight coefficient and the type of variables are used to establish the scoring standard.
Results: a total of 4295 patients were included, and 691 people died within 2 years (16.1%). Multiple factor Logistic regression analysis showed that age, primary disease, vascular access, malignant tumor, serum albumin and blood total calcium were independent factors of death, and serum albumin was a protective factor; the model formula =0.614* age (=1,2,3,4, Or 5) +0.864* vascular access (=0, or 1) + (0.784* (if diabetic nephropathy (diabetic nephropathy, DN) =1), or 0.217 (if essential hypertensive nephropathy (hypertensive nephrosclerosis, HTN) =1), or 0.796 (if the primary disease is diagnosed as other and unknown =1)) The area of the ROC curve in the training group was 0.767,95%CI (0.744 ~ 0.790), H-L test, X 2=3.144, P=0.925, and verification group proved that the area under the ROC curve was 0.732,95%CI (0.700 ~ 0.765), the sensitivity and specificity were 71.5% and 64.2%, respectively, the age 30 years old: 3, each increase 15 years old, 3 points, 75 years as 15 points; the primary disease: Chronic Renal Small Chronic glomerulonephritis (CGN): 0 points, DN:4 points, HTN:1 points, and other: 4 points; vascular access: arteriovenous fistula (arteriovenous fistula, AVF): 0 points, central venous catheter: 4; complication: malignant tumor: 3 points; laboratory data: serum albumin: 35g/1:0, 35g/1:-3 points; blood total calcium: 2.1mmol/1:1 score 2.1-2.6mmol/1:0 scores and 2.6mmol/l:3 scores were divided into low risk group (9 points), middle risk group (10~13 points), high risk group (14~17 points) and extremely high risk group (18 points).
Conclusion: This study reviewed the information of new MHD patients in the ZDQM database for 2008~2012 years, and constructed the prediction model of death risk within 2 years after dialysis by using the basic data and laboratory data that were easy to obtain, and set up a corresponding grading standard, which proved to have good identification and accuracy. It is of some reference value for future clinical research personnel to predict survival prognosis and make clinical decisions.
【学位授予单位】：浙江大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：R692.5

【参考文献】