肝癌患者外科切除术后感染风险预测模型的建立

Construction of predictive model of infection risk of patients with liver cancer after surgical resection

摘要:

目的 分析肝癌患者外科切除术后感染的危险因素，建立肝切除术后感染的预测模型。 方法 回顾性分析2017年2月—2019年10月苏州大学附属第一医院普外科收治的施行肝癌外科切除术的患者。调查患者的一般资料、实验室资料、手术资料、术后感染情况等。通过单因素χ²检验和多因素logistic回归确定患者术后发生感染的独立危险因素。基于独立危险因素，建立logistic风险预测模型。利用ROC曲线下面积评价模型的预测效果。另选取2020年1—6月肝癌外科切除术患者100例建立验证组，对建模组数据进行验证。 结果 共纳入患者310例，外科手术后发生医院感染45例，医院感染率为14.52%。其中手术部位感染15例（占33.33%），肝周围感染12例（占26.67%），其他远处部位感染18例（占40.00%）。血清清蛋白（ALB） < 35 g/L、手术时间>240 min、输血量>1 000 mL、引流时间>7 d、ASA评分>Ⅱ级是肝癌患者外科术后发生感染的独立危险因素。根据危险因素构建风险预测模型。该预测模型的ROC曲线下面积为0.904，灵敏度为0.889，特异度为0.766，Youden指数为0.655。 结论 该模型对肝癌患者外科切除术后发生感染的预测效果良好，可为临床医务人员预防高危人群发生感染提供一定理论依据。

Abstract:

Objective To analyze the risk factors of post-operative infection in patients with liver cancer after hepatectomy, and construct a predictive model for post-hepatectomy infection. Methods Patients who underwent hepatectomy in the First Affiliated Hospital of Soochow University from February 2017 to October 2019 were retrospectively analyzed. General data, laboratory data, operation data and post-operative infection of patients were investigated. Univariate χ² test and multivariate logistic regression were used to determine the independent risk factors for post-operative infection. Based on independent risk factors, logistic risk predictive model was constructed. The area under the receiver operating characteristic (ROC) curve was used to evaluate the predictive effect of the model. In addition, 100 cases of liver cancer patients undergoing surgical resection from January to June in 2020 were selected to establish the validation group, data of modeling group were verified. Results A total of 310 patients were inclu-ded in analysis, 45 cases of healthcare-associated infection (HAI) developed after operation, incidence of HAI was 14.52%, there were 15 cases (33.33%) of surgical site infection, 12 cases (26.67%) of peri-hepatic infection and 18 cases (40.00%) of other distant site infection. Serum albumin (ALB) < 35 g/L, operation time>240 minutes, blood transfusion volume >1 000 mL, drainage time>7 days, ASA score>grade II were independent risk factors for post-operative infection in patients with liver cancer. According to risk factors, risk predictive model was constructed. The area under the ROC curve of predictive model was 0.904, sensitivity, specificity and Youden index were 0.889, 0.766 and 0.655 respectively. Conclusion The model has a good predictive effect on post-operative infection in liver cancer patients after hepatectomy, which can provide a theoretical basis for clinical medical staff to prevent infection in high-risk groups.