老年非胆源性AP继发IPN的危险因素及评分体系

Study on Risk Factors and Scoring System of IPN Secondary to Senile Non-Biliary AP

目的探讨老年急性非胆源性胰腺炎患者继发感染性胰腺坏死的高危因素并构建临床风险分层的评分体系。方法采用回顾性病例对照研究方法,收集2018年9月至2020年5月锦州医科大学附属第一医院收治的确诊并经规范治疗及随诊的129例老年急性非胆源性胰腺炎患者的临床资料;男62例,女67例;年龄为(70.62±4.63)岁,年龄范围为65~91岁,按是否继发感染性胰腺坏死将患者分为感染组(n=44)和未感染组(n=85),比较两组患者病程、体重指数(body mass index,BMI)、是否合并基础疾病、是否有器官受累及血液指标,正态分布的计量资料以均数±标准差(χ^(2)±s)表示,组间比较采用独立样本t检验;偏态分布的计量资料以M表示,组间比较采用Mann-Whitney U检验。计数资料以例(百分率)表示,组间比较采用χ^(2)检验;单因素分析采用χ^(2)检验。多因素采用Logistic回归模型,P<0.05为差异有统计学意义,对具有统计学意义的变量进行筛选独立危险因素并建立风险分层评分,同时通过绘制(receiver operation characteristics,ROC)受试者特征曲线检测证明其价值。结果(1)影响老年急性非胆源性胰腺炎感染性胰腺坏死的危险因素分析。单因素分析结果显示:患者合并腹水、BMI、病程(h)、白蛋白(albumin,ALB)、血清钙离子(Ca^(2+))、血小板计数(platelet count,PLT)、降钙素原(procalcitonin,PCT)、C-反应蛋白(c-reactive protein,CRP)、凝血酶原时间(prothrombin time,PT)、血乳酸(blood lactic acid,BLA)是影响老年急性非胆源性胰腺炎感染性胰腺坏死的相关因素(P<0.05)。多因素分析结果显示:合并腹水,ALB<34 g/L、PLT<214×10^(9)/L、CPR>132.4 mg/L、PT>14.27 s是影响老年急性非胆源性胰腺炎感染性胰腺坏死的独立危险因素(P<0.05);(2)风险评分预测模型的建立及验证。根据多因素Logistic回归结果建立风险预测P=1/1+exp(19.35^(2+)2.792X_(1)-0.452X 2-0.028X 3+0.018X 4+0.480X 5),采用Hosmer-Lemeshow检验回归方程拟合优度较高(P=0.518),采用ROC曲线评价回归方程的区分度,曲线下面积(area under curve,AUC)为0.911(95%可信区间为0.848~0.974)。风险预测评分及不同评分患者发生感染性胰腺坏死的概率。患者合并腹水、ALB<34 g/L、PLT<214×10^(9)/L、CPR>132.4 mg/L、PT>14.27 s发生感染性胰腺坏死风险评分分别为2、1、1、1、1分。总分为0、1、2、3、4、5、6分,评分≥3分发生胰腺坏死的概率为25.1%,评分<3分发生感染性胰腺坏死的概率为1.6%。结论合并腹水、ALB、PLT、CRP、PT为老年急性非胆源性胰腺炎合并感染性胰腺坏死的独立危险因素,建立感染性胰腺坏死风险预测评分模型,能有效识别感染性胰腺坏死的高危患者,可为临床诊断和治疗及预后提供参考。

Objective To investigate the high risk factors of secondary infectious pancreatic necrosis in elderly patients with acute non-biliary pancreatitis and to establish a clinical risk stratified scoring system.Methods A retrospective case-control study was conducted to collect the clinical data of 129 elderly patients with acute non-biliary pancreatitis who received standard treatment and follow-up in the First Affiliated Hospital of Jinzhou Medical University from September 2018 to May 2020.62 males and 67 females,aged(70.62±4.63)years,ranging from 65 to 91 years old,were included in this study.The patients were divided into infected group(n=44)and uninfected group(n=85)according to whether secondary infectious pancreatic necrosis occurred.Course of disease,body mass index(BMI),complications with underlying diseases,organ involvement and blood indexes were compared between the two groups.Measurement data of normal distribution were expressed as mean±standard deviation(χ^(2)±s),and independent sample test was used for the comparison between the groups.The measurement data of skewed distribution was expressed as M,and the comparison between groups was performed by test.Enumeration data were expressed by means of cases(percentage),and χ^(2) test was used for the comparison between the groups.Univariate analysis was performed by using χ^(2) test.Logistic regression model was used for multiple factors,with statistically significant difference(P<0.05).Independent risk factors were screened for variables with statistical significance and risk stratified score was established.At the same time,the value was proved by drawing the receiver characteristic curve(ROC).Results(1)Analysis of risk factors affecting infectious pancreatic necrosis in elderly patients with acute non-biliary pancreatitis.Univariate analysis results showed that the complications including ascites,BMI,course of disease(h),albumin(ALB),serum calcium(Ca^(2+)),platelet count(PLT),procalcitonin(PCT),C-reactive protein(CRP),prothrombin time(PT),blood lactic acid(BLA)were associated with infectious pancreatic necrosis in elderly patients with acute non-biliary pancreatitis(P<0.05).Multivariate analysis showed that ascites,ALB<34 g/L,PLT<214×10^(9)/L,CPR>132.4 mg/L and PT>14.27 s were independent risk factors for infective pancreatic necrosis in elderly patients with acute non-biliary pancreatitis(P<0.05).(2)Establishment and verification of prediction model for risk score.The risk prediction was established based on multivariate Logistic regression results,P=1/1+exp(19.35^(2+)2.792X_(1)-0.452X 2-0.028X 3+0.018X 4+0.480X 5).Hosmer-Lemeshow test showed that the regression equation had a higher goodness-of-fit(P=0.518).The ROC curve was used to evaluate the differentiation of regression equations,and the area under curve(AUC)was 0.911(95%confidence interval was 0.848~0.974).Risk prediction score and the probability of infectious pancreatic necrosis in patients with different scores were as below.The risk scores of infective pancreatic necrosis in patients with ascites,ALB<34 g/L,PLT<214×10^(9)/L,CPR>132.4 mg/L and PT>14.27 s were 2,1,1,1,1,respectively.The total score was 0,1,2,3,4,5,and 6.The probability of pancreatic necrosis with score≥3 was 25.1%,and that of infectious pancreatic necrosis with score<3 was 1.6%.Conclusion Complications including ascites,ALB,PLT,CRP and PT are independent risk factors for elderly acute non-biliary pancreatitis complicated with infectious pancreatic necrosis.The establishment of prediction model for risk score for infectious pancreatic necrosis can effectively identify high-risk patients with infectious pancreatic necrosis,which can provide reference for clinical diagnosis,treatment and prognosis.