预测腹部手术切口感染风险的列线图模型的建立

Nomogram model of predicting postoperative surgical wound infection in patients with abdominal surgery

目的通过对腹部手术切口感染相关危险因素进行分析，建立个体化预测腹部手术切口感染风险的列线图模型。方法纳入2014年6月-2017年6月符合人选标准的腹部手术患者2931例，收集临床资料，采用单因素分析筛选手术切口感染相关因素，将有统计学意义的变量纳入多因素Logistic回归模型，分析切口感染危险因素；应用R软件建立预测腹部手术切口感染风险的列线图模型，采用Bootstrap法进行模型验证，并应用ROC曲线探索列线图模型对腹部手术切口感染的预测效率。结果本组腹部手术患者发生切口感染84例，感染率为2．87％；共检出84株病原菌，以革兰阴性菌为主，占59．52％。Logistic回归分析显示，年龄（OR=4．196）、营养不良（OR=2．559）、糖尿病（OR=2．949）、急诊手术（OR=3．326）及术后未进行康复教育（OR=1．198）是腹部手术切口感染的独立危险因素（P〈0．05）。列线图模型预测腹部手术切口感染风险的一致性指数为0．726；ROC曲线显示列线图模型预测腹部手术切口感染的曲线下面积为0．735（95％CI：0．672～0．827）。结论本研究建立的预测腹部手术切口感染风险的列线图模型，具有良好的区分度与准确度，临床应用价值高，对甄别高风险人群，制订干预对策具有指导意义。

Objective To explore the risk factors for abdominal surgical wound infection, and to establish a nomogram model for the prediction of the risk factors. Methods A total of 2 931 patients underwent abdominal surgery were enrolled in this study from June 2014 to June 2017. The clinical data were collected. The univariate and multivariate logistic regression were used to analyse the risk factors for abdominal surgical wound infection. A predictive nomogram model was established using the R software. Bootstrap method was used to validate the nomogram model and ROC curve was used to explore the predictive efficacy of the model in predicting the abdominal surgical wound infection. Results Of 2 931 enrolled patients, 84 of them had surgical wound infection and the infection rate was 2.87%. A total of 84 pathogens were detected, which approximately 59.52% of them were gram-negative bacteria. The muhivariable logistic regression showed that age, malnutrition, diabetes mellitus, emergency operation, and unsuccessful rehabilitation education were independent risk factors for abdominal surgical wound infection （OR=4.196, 2.559, 2.949, 3.326, 1.198; P 〈 0.05）. The nomogram model showed a C-index of 0.726 with good discrimination and accuracy. The ROC curve showed that the area under the curve （AUC） for the nomogram model was 0.735 （95%CI： 0.672-0.827）. Conclusions The established nomogram model has good discrimination and accuracy for predicting abdominal surgical wound infection. It is helpful for distinguishing high-risk populations and improving intervention strategies in clinical practice.