Tianjin Medical Journal ›› 2025, Vol. 53 ›› Issue (3): 242-246.doi: 10.11958/20242177

• Clinical Research • Previous Articles     Next Articles

Construction and effect evaluation of multi-drug resistant bacterial infection prediction model for patients with esophageal cancer after operation

BU Jing1(), WANG Pengyu2, YANG Xingxiao1,()   

  1. 1 Department of Infection Management, the Fourth Hospital of Hebei Medical University, Shijiazhuang 050011, China
    2 Department of Clinical Laboratory, the Fourth Hospital of Hebei Medical University, Shijiazhuang 050011, China
  • Received:2024-12-11 Revised:2025-01-20 Published:2025-03-15 Online:2025-03-31
  • Contact: E-mail:48401334@hebmu.edu.cn

Abstract:

Objective To analyze the risk factors of postoperative multi-drug resistant bacteria (MDRO) infection in patients with esophageal cancer, construct the nomogram model and evaluate fitting effect of the model, so as to help doctors make accurate clinical decisions. Methods A total of 116 patients with esophageal cancer who received surgical treatment were selected and divided into the infected group (25 cases) and the uninfected group (91 cases) according to whether they were infected with MDRO. American anesthesia association of physicians rating (ASA) score and tumor locations (upper, middle and lower segments) on admission were compared between the two groups. Surgical method (endoscopic, open), clinical stage, history of chemoradiotherapy, preoperative nutritional status, white blood cell count at admission, retention time of drainage tube, retention time of central venous catheter, length of ICU stay and total length of hospital stay were also compared between the two groups. Principal component analysis (PCA) was used to screen and reduce the dimension of the data. Multivariate Logistic regression analysis was performed to analyze the risk factors of postoperative MDRO infection. A nomogram model of MDRO infection risk was constructed, and the predictive fitting effect of the model was evaluated by receiver operating characteristic (ROC) curve and calibration curve. Results Compared with the uninfected group, higher ASA score (≥3 points), laparotomy and clinical stage Ⅲ, higher proportion of patients with poor nutritional status before surgery, longer drainage tube retention time, central venous catheter retention time and long ICU stay time were found in the infected group (P < 0.05). Multivariate Logistic regression analysis suggested that open surgery, long stay in ICU and poor preoperative nutritional status were risk factors for postoperative MDRO infection in patients with esophageal cancer (P<0.05). Based on this, the area under ROC curve of the nomogram model was 0.828 (0.759-0.897). The results of Hosmer-Lemeshow test showed χ2=0.426, P=1.000, and the model had good goodness-fit, high calibration degree and clinical application degree. Conclusion The nomogram risk prediction model based on the mode of operation,length of ICU stay and preoperative nutritional status has good prediction ability.

Key words: esophageal neoplasm, laparotomy, nomograms, risk factors, length of stay, multidrug-resistant bacteria

CLC Number: