碳青霉烯耐药肺炎克雷伯菌血流感染预测模型的开发与验证
该研究通过检索PubMed、Cochrane Library和Embase数据库(建库至2022年7月)获取碳青霉烯耐药肺炎克雷伯菌血流感染(CRKP-BSI)的危险因素,合并相对危险度(RR)后基于因素权重建立风险评分模型。模型纳入14项危险因素,总分0~173.5分。在上海第一人民医院2016年1月至2022年1月住院的230例CRKP感染患者中进行外部验证,41例发生CRKP-BSI,模型曲线下面积(AUC)为0.783(95%CI 0.689~0.876),最佳截断值81.25分,敏感度75.6%、特异度81.0%。高风险组CRKP-BSI发生率显著高于低中风险组[64.2%(43/67)比4.9%(8/163);RR=13.175(95%CI 5.920~29.319),P<0.001]。作者认为该模型可能有助于临床医生早期识别高风险患者。
要点
- 验证队列230例CRKP感染患者中41例发生CRKP-BSI,模型AUC为0.783(95%CI 0.689~0.876)。
- 最佳截断值81.25分时敏感度75.6%、特异度81.0%,据此将患者分为低中风险和高风险两组。
- 高风险组CRKP-BSI发生率为64.2%(43/67),低中风险组为4.9%(8/163),RR=13.175(95%CI 5.920~29.319),P<0.001。
- 模型纳入14项常规临床参数:心血管疾病、严重中性粒细胞减少或免疫抑制、ICU住院史、既往住院、碳青霉烯暴露、氨基糖苷类暴露、抗真菌药暴露、气管插管或气管切开、机械通气、血液透析、中心静脉导管、留置导尿管、CRKP定植、非感染部位肺炎克雷伯菌阳性。
- 总分范围0~173.5分,作者认为该模型可能辅助临床早期识别CRKP-BSI高风险患者。
适用人群与场景
- 验证队列为上海第一人民医院2016年1月至2022年1月住院的CRKP感染患者,共230例。
诊断标准与关键阈值
- 风险评分模型总分0~173.5分,最佳截断值81.25分;评分≥81.25分为高风险组,<81.25分为低中风险组。
研究设计与关键结果
- 研究类型为预测模型开发与外部验证;开发基于PubMed、Cochrane Library和Embase数据库建库至2022年7月检索的文献,合并14项危险因素的RR并基于权重建立评分模型。
- 验证队列230例CRKP感染患者中41例发生CRKP-BSI,模型AUC为0.783(95%CI 0.689~0.876)。
- 最佳截断值81.25分,敏感度75.6%、特异度81.0%。
- 按截断值分组:低中风险组163例,高风险组67例;高风险组CRKP-BSI发生率64.2%(43/67),低中风险组4.9%(8/163),RR=13.175(95%CI 5.920~29.319),P<0.001。
证据类型、等级与局限
- 证据类型:预测模型开发与外部验证研究。
- 局限:原文未报告局限性。
对床旁实践的改变
- 该模型基于14项常规可获取的临床参数,作者认为可能辅助临床医生早期识别CRKP-BSI高风险患者。
提炼说明
- 本条依据论文摘要整理;摘要未涉及「处置流程与药物剂量」「红旗征与不能漏的情况」「指南或共识出处」,故未列出这些栏目;全文请点「阅读原文」。
英文原文(摘要)· 点击展开
OBJECTIVE: To develop and validate a predictive model for the risk of bloodstream infection (BSI) caused by carbapenem-resistant Klebsiella pneumoniae (CRKP).
METHODS: A literature search was conducted in PubMed, Cochrane Library, and Embase databases from inception to July 2022 to identify studies reporting statistically significant risk factors for CRKP-BSI. Relative risks (RR) were extracted and pooled. Based on factor weights, a risk-scoring model was established. For external validation, hospitalized CRKP-infected patients from January 2016 to January 2022 at Shanghai First People's Hospital were included. Clinical data were used to calculate individual risk scores. The predictive accuracy was assessed using receiver operator characteristic curve (ROC curve). Patients were stratified into low-to-intermediate-risk and high-risk groups based on the optimal cut-off, and CRKP BSI incidence was compared between groups.
RESULTS: The literatures related to the risk factors of CRKP-BSI published from database inception to July 2022 was retrieved and screened from PubMed, Cochrane Library, and Embase. Fourteen risk factors were included in the scoring model: cardiovascular disease, severe neutropenia or immunosuppression, intensive care unit (ICU) stay history, prior hospitalization, carbapenem exposure, aminoglycoside exposure, antifungal exposure, endotracheal intubation or tracheostomy, mechanical ventilation, hemodialysis, central venous catheter, indwelling urinary catheter, CRKP colonization, and Klebsiella pneumoniae positivity at non infection sites. The total score ranged from 0 to 173.5 points. In the validation cohort of 230 CRKP-infected patients, 41 developed CRKP BSI. The model yielded an area under the curve (AUC) of 0.783 (95%CI was 0.689-0.876). The optimal cut off was 81.25 points, with sensitivity of 75.6% and specificity of 81.0%. Based on this cut off, 163 patients were categorized as low-to-intermediate risk and 67 patients as high risk. The incidence of CRKP BSI in the high-risk group was significantly higher than in the low-to-intermediate-risk group [64.2% (43/67) vs. 4.9% (8/163); RR = 13.175 (95%CI was 5.920-29.319), P < 0.001].
CONCLUSIONS: The model, based on 14 routinely available clinical parameters, demonstrated good performance in predicting CRKP BSI risk and may assist clinicians in early identification of high risk patients.
参考来源
- Jin S, Zhou F, Wei D, et al. [Development and validation of a prediction model for bloodstream infection caused by carbapenem-resistant Klebsiella pneumoniae]. Zhonghua Wei Zhong Bing Ji Jiu Yi Xue. 2025. PMID: 41146503. doi:10.3760/cma.j.cn121430-20241014-00840
供执业医师抢救与诊疗参考,不替代临床判断;剂量请结合体重、肝肾功能和本院药品规格复核,以指南原文为准。