Establishment and validation of a predictive clinical model for postoperative surgical site infection in patients with colorectal surgery
10.3760/cma.j.cn441530-20230619-00217
- VernacularTitle:结直肠术后手术部位感染临床预测模型的建立和验证
- Author:
Yiyu YANG
1
;
Xufei ZHANG
;
Jianwei ZHU
;
Peige WANG
;
Wenjing LIU
;
Xiuwen WU
;
Jian'an REN
Author Information
1. 东南大学医学院附属第二临床医院（东部战区总医院）全军普通外科研究所，南京 210002
- Keywords:
Surgical site infection;
Colorectal surgery;
Clinical prediction models;
Nomogram
- From:
Chinese Journal of Gastrointestinal Surgery
2023;26(9):837-846
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To investigate the risk factors of surgical site infection (SSI) after colorectal surgery, and to establish and validate a risk prediction model nomogram.Methods:An observational study was conducted to retrospectively collect data of 6527 patients aged ≥16 years who underwent colorectal surgery in 56 domestic hospitals from March 1, 2021 to February 28, 2022 from the national Surgical Site Infection Surveillance network. The incidence of SSI after surgery was 2.3% (149/6527). According to the ratio of 7:3, 6527 patients were randomly divided into the modeling cohort (4568 cases) and the validation cohort (1959 cases), and there was no statistically significant difference between the two datasets ( P>0.05). Univariate analysis was performed using t test /Mann-Whitney U test /χ 2 test. Multivariate analysis was performed using binary logistic regression to establish a preliminary model and select variables using Lasso analysis to establish an optimized model nomogram. The discrimination and calibration of the model were evaluated by ROC curve, calibration curve, and Hosmer-Lemeshow test. AUC value>0.7 is considered a good discrimination of the model. The Bootstrap method (repeated self-sampling 1000 times) was used to verify the constructed model internally and externally to evaluate the accuracy of the constructed model. Results:Multivariate analysis showed that history of chronic liver disease (OR=3.626, 95%CI: 1.297-10.137, P<0.001) and kidney disease (OR=1.567,95%CI:1.042-2.357, P=0.038), surgical antibiotic prophylaxis (OR=1.564, 95%CI:1.038-2.357, P=0.035), and emergency surgery (OR=1.432,95%CI: 1.089-1.885, P=0.021), open surgery (OR=1.418, 95%CI:1.045-1.924, P=0.042), preoperative stoma (OR=3.310, 95%CI:1.542-7.105, P<0.001), postoperative stoma (OR=2.323,95%CI: 1.537-8.134, P<0.001), surgical incision type above grade II (OR=1.619,95%CI:1.097-2.375, P=0.014), and each unit increase in total bilirubin (OR=1.003,95%CI:-0.994-1.012, P=0.238), alanine aminotransferase (OR=1.006, 95%CI:1.001-1.011, P=0.032), blood urea nitrogen (OR=1.003,95%CI:0.995-1.011, P=0.310), blood glucose (OR=1.024, 95%CI:1.005-1.043, P=0.027), C-reactive protein (OR=1.007, 95%CI:1.003-1.011, P<0.001), length of incision (OR=1.042, 95%CI:1.002-1.087, P=0.031), surgical duration (OR=1.003,95%CI:1.001-1.005, P=0.017), and surgical blood loss (OR=1.001,95%CI: 1.000-1.002, P=0.045) were risk factors for SSI after colorectal surgery. Each unit increase in albumin level (OR=0.969,95%CI:0.941-0.998, P=0.036) was an independent protective factor for SSI after colorectal surgery. The area under the curve of the optimized model obtained by internal and external validation were 0.768 (95%CI: 0.723-0.813) and 0.753 (95%CI: 0.680-0.832), respectively. The predicted value of the calibration curve was basically consistent with the actual value. Conclusions:The risk prediction model for SSI after colorectal surgery constructed in this study has good discrimination and calibration. The nomogram created in this model can provide an evaluation basis for the observed rate and expected event rate of SSI after clinical colorectal surgery.
