食管癌切除术后围术期死亡危险因素的分析及预测模型的建立

Identification of risk factors for perioperative mortality with esophagectomy and establishment of a prediction model

目的探讨食管癌围术期死亡原因及危险因素,建立Logistics回归模型.方法收集2002-01/2015-12于新疆医科大学第一附属医院进行食管癌切除术患者临床资料,食管癌850例,根据围术期及住院期间结局分为死亡和非死亡各1组,将64个可能对食管癌手术死亡有相关影响的常见因素进行回顾性分析,通过计算机用Logistics回归模型研究食管癌围术期与死亡原因相关的危险因素,进行单因素、多因素分析,建立回归方程,总结实际意义.结果850例食管癌中手术死亡39例,手术死亡率4.59%.单因素Logistics回归分析,64个因素中,25个因素与食管癌切除术后围术期死亡有关,分别为民族、年龄、术前合并症、术前血红蛋白值、手术切除方式、吻合方式、吻合水平、手术年代、术后白蛋白值、肠内营养时间、ICU治疗时间、心律失常、肺炎、肺部感染,重症肺炎,单侧胸腔积液、双侧胸腔积液、感染性休克、MODS、吻合口瘘、乳糜胸、呼吸衰竭、循环衰竭、呼吸循环衰竭、缺血缺氧综合征等,统计学差异有意义(P<0.05).多因素Logistics回归分析提示手术年代,术后白蛋白值,肠内营养时间,心律失常,MODS,乳糜胸,呼吸衰竭等因素(P<0.05),有统计学意义,为本研究食管癌围术期死亡原因独立危险因素并进入回归方程,获得预测模型为:P=1/1+e^(2.951X_(25)+1.922X_(27)+0.445X_(30)-2.066X_(32)+4.313X_(50)+2.333X_(55)-2.689X_(58)-8.999),(0≤P≤1).结论食管癌围术期死亡原因分析中,手术年代、术后白蛋白值、心律失常、MODS、乳糜胸、呼吸衰竭等为本研究食管癌围术期死亡原因独立危险因素并进入回归方程,并获得预测模型;欲降低食管癌围术期死亡,需改善医疗环境,提高医疗技术,并在围术期严密观察和评估血清白蛋白值,科学并谨慎开始肠内营养时间,积极预防术后发生心律失常、MODS、乳糜胸以及呼吸衰竭等并发症,并尽早处理.

AIM To identify the risk factors for perioperative mortality in patients with esophageal cancer,and establish a logistics regression model.METHODS From January 2002 to December 2015,the clinical data for 850 patients who were diagnosed with esophageal cancer and underwent esophagectomy at our hospital were divided into a non-death group and a death group according to the presence of perioperative mortality or not.Sixty-four factors that may have an effect on surgical mortality in esophageal cancer were analyzed retrospectively.The risk factors for surgical death were then identified by univariate and multivariate analyses.A logistics regression model was finally established based on the identified risk factors.RESULTS Of 850 cases of esophageal cancer included,perioperative death occurred in 39（4.59%）cases.Univariate logistics regression analysis showed that 25 of 64 factors were significantly related with perioperative death,including ethnicity,age,preoperative comorbidities,preoperative hemoglobin,surgical resection,anastomosis,anastomosis level,operative time,postoperative albumin,enteral nutrition time,ICU treatment time,arrhythmia,pneumonia,lung infection,severe pneumonia,unilateral pleural effusion,bilateral pleural effusion,septic shock,MODS,anastomotic fistula,chylothorax,respiratory failure,circulatory failure,respiratory and circulatory failure,ischemia and hypoxia syndrome（P 〈 0.05）.Multivariate regression analysis showed that operative time,arrhythmia,postoperative albumin,enteral nutrition time,MODS,chylothorax,and respiratory failure were independent risk factors for perioperative mortality（P 〈 0.05）.Based on these identified risk factors,we obtained a prediction model：P =1/1 ＋ e∧（2.951χ25＋1.922χ27＋ 0.445χ30-2.066χ32＋4.313χ50＋2.333χ55-2.689χ58-8.999）（0 〈 p≤1）.CONCLUSION Operative time,arrhythmia,postoperative albumin,enteral nutrition time,MODS,chylothorax,and respiratory failure are independent risk factors for perioperative mortality in esophageal cancer.To reduce perioperative mortality in esophageal cancer,it needs to improve the medical environment and enhance surgical techniques,enhance preoperative,intraoperative and postoperative evaluation of postoperative albumin,start enteral nutrition carefully and scientifically,prevent the occurrence of postoperative arrhythmia,MODS,chylothorax,respiratory failure and other complications,and give treatment as soon as possible.