老年胃肠癌术后全身炎症反应综合征危险因素分析及预测模型建立  

Risk factors analysis and a prediction model of SIRS in elderly patientswith gastrointestinal cancerrelated surgery

作　　者：林洪岩 刘馨 李涛[1] 陈东旭 李茜[1] LIN Hongyan;LIU Xin;LI Tao;CHEN Dongxu;LI Qian(Department of Anesthesiology,West China Hospital,Sichuan University,Chengdu 610041,China)

机构地区：[1]四川大学华西医院麻醉科,四川成都610041

出　　处：《西部医学》2023年第8期1213-1218,共6页Medical Journal of West China

基　　金：国家重点研发计划项目(2018YFC2001802-2)。

摘　　要：目的基于老年胃肠癌术后围术期指标构建全身炎症反应综合征(SIRS)预测模型。方法回顾性收集2015年1月—2020年12月我院择期胃肠癌手术患者围术期诊疗数据,基于Logistics回归和LASSO回归筛选特征变量,建立列线图预测模型,并采用受试者工作特征曲线下面积(ROC)和校准曲线评估预测模型的区分度和一致性。结果总计纳入1169例患者,其中99例(8.5%)患者发生SIRS。SIRS组患者术后30 d死亡率和术后感染率增加、ICU入住率、住院时间和医疗总费用增加。Logistics分析结果显示,患者BMI、手术时间、术后白细胞计数(WBC)、中性粒细胞计数与淋巴细胞计数比值(NLR)是术后SIRS的危险因素(P<0.05),术后血红蛋白计数越高,SIRS发生风险越低(P<0.05)。由以上变量建立的预测模型的曲线下面积(AUC)为0.760(95%CI:0.671~0.772,P<0.05),敏感性为0.818,特异性为0.570。结论老年胃肠癌患者术后发生SIRS严重影响患者预后;以BMI、手术时间、术后WBC、NLR和HGB建立的预测模型对术后SIRS的发生具有较高的预测效能和良好的临床适用性。Objective This study aimed to develop a model to predict the risk of SIRS entirely based on readily available electronic health records.Methods The perioperative diagnosis and treatment data of elective major gastrointestinal cancer surgery in West China Hospital of Sichuan University from January 2015 to December 2020 were retrospectively collected.Based on Logistics and LASSO regression,feature predictors were identified and construed a prediction model.The discriminative ability of the prediction model was assessed in two ways,based on areas under the receiver operating characteristic curves(AUCs)and visual agreement between the calibration curve for the prediction model and the ideal calibration curve.Results Among of 1169 patients were enrolled,99(8.5%)cases of SIRS were identified.A higher incidence of postoperative 30-day mortality,postoperative infection rates,and ICU occupancy rates,longer length of stay and total medical costs were found in SIRS group.The analysis showed that body mass index,operation time,postoperative white blood cell count,neutrophil count to lymphocyte count ratio were risk factors for postoperative SIRS(P<0.05),while the higher postoperative hemoglobin count,the lower the risk of SIRS(P<0.05).The area under the curve(AUC)was 0.760(95%CI:0.671-0.772,P<0.05),the sensitivity was 0.818,and the specificity was 0.570,which showed a satisfactory discrimination.Conclusion Occurring SIRS after operation in elderly patients with gastrointestinal cancer,can affect the prognosis of patients.The prediction model developed in current study can predict the risk of SIRS with satisfactory discrimination.

关 键 词：老年 胃肠癌 全身炎症反应综合征 危险因素 预测模型 

分 类 号：R735[医药卫生—肿瘤]