Establishment of risk prediction model for polymyxin-and carbapenem-resistant Klebsiella pneumoniae infection
10.11816/cn.ni.2025-240989
- VernacularTitle:耐多黏菌素和碳青霉烯类肺炎克雷伯菌感染风险预测模型构建
- Author:
Xiaqin HE
1
;
Meng LIU
1
;
Yi ZHANG
1
;
Xiaoqian WANG
1
;
Weiqi WANG
1
;
Xiaoqin WANG
1
Author Information
1. 西安交通大学第一附属医院检验科,陕西西安 710061
- Publication Type:Journal Article
- Keywords:
Polymyxin-resistant;
Carbapenem-resistant Klebsiella pneumoniae;
Risk prediction model;
Risk fac-tor;
Antibiotic
- From:
Chinese Journal of Nosocomiology
2025;35(5):776-781
- CountryChina
- Language:Chinese
-
Abstract:
OBJECTIVE To explore the risk factors for polymyxin-and carbapenem-resistant Klebsiella pneumoni-ae(PR-CRKP)infection and establish the prediction model.METHODS The clinical data were retrospectively col-lected from the patients with CRKP infection who were treated in the First Affiliated Hospital of Xi'an Jiaotong University from Jan.2023 to Mar.2024.The enrolled patients were divided into the CRKP group and the PR-CRKP group according to the result of drug susceptibility testing for polymyxin B.Totally 203 patients who were treated from Jan.2023 to Dec.2023 were assigned as the modeling group,and 91 patients who were treated from Jan.2024 to Mar.2024 were assigned as the validation group.Multivariate logistic regression analysis was per-formed for the risk factors for PR-CRKP infection,nomogram was built up for prediction of PR-CRKP infection by R software,and the predictive efficacy of the model was evaluated by means of receiver operating characteristic(ROC)curves.RESULTS The result of univariate analysis showed that the proportions of patients who received fi-berobronchoscopy,endotracheal intubation/tracheotomy,were complicated with other carbapenem-resistant or-ganisms(CROs)infections,bloodstream infections,were treated in intensive care unit(ICU)and had the history of exposure to polymyxins and carbapenems 3 months before the admission were higher in the PR-CRKP group than in the CRKP group(P＜0.05);the length of hospital stay and duration of use of polymyxins and carbapene-ms were longer in the PR-CRKP group than in the CRKP group(P＜0.05).Multivariate logistic regression analy-sis indicated that complication with other CROs infections,history of exposure to polymyxins and carbapenems 3 months before the admission,duration of use of polymyxins and ICU stay were the risk factors for the PR-CRKP infection(P＜0.05).The area under the curve(AUC)of the predictive model was 0.898 in the modeling group,with the sensitivity 80.33％,the specificity 84.51％.ROC curve analysis showed that the AUC was 0.901 in the validation group,with the sensitivity 75.00％,the specificity 92.73％.CONCLUSION The prediction model that is established based on the result of multivariate analysis has high value in prediction of PR-CRKP infection.
