关键词: 颅脑损伤, 肺炎克雷伯菌, 预测模型, 列线图, 耐药性

Abstract:

Objective To analyze the antimicrobial resistance profile of carbapenem-resistant Klebsiella pneumoniae (CRKP) isolated from neurosurgical patients with severe traumatic brain injury (sTBI), to identify independent risk factors for CRKP infection in this population, and to construct a nomogram for individualized risk prediction. Methods We retrospectively enrolled 524 sTBI patients admitted to the Department of Neurosurgery, Ganzhou People's Hospital, between September 2019 and September 2024. Patients were grouped by CRKP infection status into a CRKP infection group (n=95) and a non-CRKP infection group (n=429). Antimicrobial susceptibility of CRKP isolates was analyzed. Variables showing statistically significant differences between the two groups were entered into multivariable logistic regression to identify independent risk factors. A nomogram based on these independent predictors was developed to estimate the risk of CRKP infection. Model discrimination was assessed by receiver operating characteristic (ROC) curve and area under the curve (AUC); calibration was evaluated with a calibration curve. Results Antimicrobial susceptibility testing showed high resistance rates of CRKP to commonly used antibiotics, while resistance rates to tigecycline and ceftazidime-avibactam were 0%. Between the two groups, the proportions of patients with age ≥60 years, hospital stay ≥30 days, operative time ≥60 minutes, number of operations ≥2, postoperative prophylactic antibiotics after first surgery, antibiotic exposure ≥30 days, combination antibiotic therapy, use of ≥3 antibiotic agents, admission GCS ≤5, coma, tracheotomy, mechanical ventilation, indwelling urinary catheter, central venous catheter, indwelling gastric tube, repeat catheterization, fever ≥10 days, blood transfusion, hypoproteinemia, and use of immunosuppressants or steroids differed significantly (all P<0.05). Multivariable logistic regression identified the following independent risk factors for CRKP infection in neurosurgical sTBI patients: age ≥60 years, antibiotic exposure ≥30 days, use of ≥3 antibiotic classes, indwelling urinary catheter, indwelling gastric tube, fever ≥10 days, blood transfusion, and hypoproteinemia (all P<0.05). A nomogram constructed from these predictors yielded an AUC of 0.965(95%CI: 0.950-0.980, P<0.05), with sensitivity 0.895 and specificity 0.944. The calibration curve showed good agreement between predicted and observed risks. Conclusion Age ≥60 years, prolonged antibiotic exposure (≥30 days), use of multiple antibiotic classes (≥3), indwelling urinary catheter, indwelling gastric tube, prolonged fever (≥10 days), blood transfusion, and hypoproteinemia are independent risk factors for CRKP infection in neurosurgical patients with sTBI. The nomogram based on these factors demonstrates excellent discriminative ability and good calibration.

Key words: craniocerebral trauma, Klebsiella pneumoniae, predictive model, nomograms, drug resistance