Risk Factors for Infection After Endoscopic Submucosal Dissection for Early Colorectal Cancer and Construction of a Risk Prediction Model
Objective To screen the risk factors for infection after endoscopic submucosal dissection for early colorectal cancer and to construct and validate a risk prediction model for postoperative infection,thus providing a certain basis for clinical early prevention and treatment.Methods Clinical data of 96 patients who underwent endoscopic submucosal dissection for early colorectal cancer were retrospectively analyzed,and patients were classified into infection group and non-infection group according to the presence or absence of postoperative infection.The risk factors of postoperative infection after endoscopic submucosal dis-section for colorectal cancer were identified via univariate and multivariate Logistic regression analysis.Then a risk prediction model was constructed and the predictive efficacy was validated using receiver operating characteristic(ROC)curve.Results Postoperative infection occurred in 11 of 96 patients with colorectal cancer after endoscopic submucosal dissection,with an inci-dence rate of 11.46%.The types of infections were predominantly urinary tract infections,followed by abdominal infections and lung infection.Univariate and multivariate Logistic regression analysis denoted that age(X1),comorbid diabetes mellitus(X2),preoperative hypoproteinemia(X3),operation time(X4)and postoperative perforation(X5)were all risk factors for postoperative infection after endoscopic submucosal dissection for early colorectal cancer(P＜0.05).Establishment of the regression equation was Logit(P)=-11.368+0.367X1+0.563X2+0.656X3+0.515X4+0.750X5.The validation of the prediction model equa-tions showed a good model fit with a C-index of 0.853(95%CI:0.706 to 0.911).ROC curve indicated that the area under the curve,standard error(SE),Youden index,sensitivity and specificity of the model was 0.882(95%CI:0.789～0.975),0.048,0.677,83.3%and 80.2%,suggesting that the prediction model had good predictive efficacy.Conclusion Age,comorbid diabe-tes mellitus,preoperative hypoproteinemia,operation time and postoperative perforation are all risk factors for postoperative infec-tion after endoscopic submucosal dissection for early colorectal cancer.The established risk prediction model shows a good fit,pre-dictive efficacy and clinical value,which can provide guidance for the formulation of clinical prevention measures.
Colorectal cancerEndoscopic submucosal dissectionPostoperative infectionRisk factorsRisk prediction model
NETL
NSTL
万方数据
