腹壁切口疝修补术后补片感染列线图预测模型构建

目的分析腹壁切口疝修补术后补片感染的危险因素，建立切口疝患者补片感染的预测模型，为临床预测切口疝修补术后发生补片感染提供一种可视化评价工具。 方法回顾性分析2016年1月至2018年12月在首都医科大学附属北京朝阳医院就诊的475例切口疝患者的临床资料，收集患者的一般资料、手术资料、术后恢复情况，随访补片感染情况。使用Lasso回归筛选预测因子，在此基础上通过多因素Logistic回归进一步分析并建立列线图预测模型，采用受试者工作特征曲线下面积评估模型的预测效力。 结果475例接受切口疝修补手术的患者中有11例出现补片感染，发生率为2.3%。Lasso回归结合多因素Logistic回归分析结果显示，体质量指数（OR=1.206，95% CI 1.034~1.407）、糖尿病史（OR=6.484，95% CI 1.233~34.108）、术后外科手术部位感染（OR=37.095，95% CI 4.253~323.532）是切口疝患者发生补片感染的影响因素（P<0.05），利用上述变量建立列线图预测模型，列线图预测模型预测补片感染发生AUC为0.880（95% CI 0.785~0.975）。 结论本研究成功建立一种具有良好预测效力的列线图预测模型，有助于提高对补片感染高危切口疝患者的早期鉴别能力，为改善切口疝患者预后提供帮助。

A predictive nomogram for the risk of mesh infection after incisional hernia repair

ObjectiveTo explore the risk factors of mesh infection after incisional hernia repair, and to develop a predictive nomogram as visualized tool for clinical prediction of mesh infection. Methods475 patients with incisional hernias were retrospectively selected from Beijing Chaoyang Hospital of Capital Medical University from January of 2016 to December of 2018. Patients' basic information, surgery information and recovery information after surgery were collected, and mesh infection information was followed up. Predictors of mesh infection were screened by Lasso regression analysis, and further analyzed by multivariate logistic regression analysis. Then the final determined ones were applied to develop a predictive nomogram. The area under the ROC curve (AUC) was used to evaluate the predictive utility of the nomogram. ResultsAmong the 475 cases, 11 developed a mesh infection; the incidence was 2.3%. The findings of Lasso regression with multivariate logistic analysis demonstrated that body mass index (BMI) OR=1.206, 95% CI (1.034, 1.407)]、diabetes OR=6.484, 95% CI (1.233, 34.108)] and surgical site infection (SSI) OR=37.095, 95% CI (4.253, 323.532)] were associated with mesh infection. The predictive nomogram was established using the above-mentioned variables. The AUC of the nomogram was 0.880 95% CI (0.785, 0.975)]. ConclusionWe successfully established a predictive nomogram with high accuracy, which may be used to improve the early identification of mesh infection and the clinical outcome of patients with incisional hernias.