慢性肾衰竭并院内下呼吸道细菌感染的风险预测列线图模型构建研究

Construction of nomograph model for risk prediction of chronic renal failure complicated with bacterial infection of lower respiratory tract in hospital

目的探讨慢性肾衰竭并院内下呼吸道细菌感染的风险因素,据此构建列线图预测模型。方法回顾性分析2019年2月至2022年1月商丘市第三人民医院收治的721例慢性肾衰竭患者临床资料,将其作为模型组,根据患者是否合并院内下呼吸道细菌感染将模型组分为并发组与未并发组,且并发组患者均进行痰培养、细菌鉴定试验。比较两组患者一般资料,采用多因素Logistic回归分析法分析模型组慢性肾衰竭并院内下呼吸道细菌感染的影响因素,并基于多因素分析结果使用R3.4.3软件包绘制列线图模型;另收集2016年1月至2019年1月该院收治的784例慢性肾衰竭患者临床资料,将其作为验证组,绘制校准曲线采用Bootstrap法检验预测列线图的校准度。结果本研究模型组721例慢性肾衰竭患者中共43例并发院内下呼吸道细菌感染,发生率为5.96%,共分离出病原菌58株,其中革兰阴性菌33株(56.90%)、革兰阳性菌22株(37.93%)、真菌3株(5.17%);模型组年龄≥60岁、抽烟史、住院时间≥15 d、不合理使用抗生素、Hb<60 g/L、Alb<35 g/L是慢性肾衰竭并院内下呼吸道细菌感染的危险因素(P<0.05);将上述危险因素作为预测指标,构建模型组慢性肾衰竭并院内下呼吸道细菌感染的风险预测列线图模型,经Bootstrap法进行验证,模型组的一致性指数(C-index)为0.829,验证组的C-index为0.825,且校正曲线的校准度较好。结论年龄≥60岁、抽烟史、住院时间≥15 d、不合理使用抗生素、Hb<60 g/L、Alb<35 g/L均是慢性肾衰竭并院内下呼吸道细菌感染的危险因素,综合上述因素构建的列线图预测模型对个体化预测慢性肾衰竭并院内下呼吸道细菌感染的发生具有良好的应用价值。

【Objective】To explore the risk factors of chronic renal failure with bacterial infection of lower respiratory tract in hospital,and build a nomograph prediction model.【Methods】The clinical data of 721 patients with chronic renal failure admitted to our hospital from February 2019 to January 2022 were retrospectively analyzed.They were treated as model group.The model group were divided into concurrent group and non-concurrent group according to whether they were complicated with bacterial infection of lower respiratory tract in hospital,and the sputum culture and bacterial identification tests were performed on the patients in the concurrent group.The general data of the two groups of patients were compared,the influencing factors of the model group with chronic renal failure complicated with bacterial infection of lower respiratory tract in hospital were analyzed by multifactor logistic regression analysis method,and R3.4.3 software package was used to draw the nomogram model based on the results of multifactor analysis.In addition,the clinical data of 784 patients with chronic renal failure admitted to our hospital from January 2016 to January 2019 were collected as the validation group,and the calibration curve was drawn to test the calibration of the prediction nomogram using the Bootstrap method.【Results】There were 43 cases of bacterial infection of lower respiratory tract in hospital in 721 patients with chronic renal failure in the model group,the incidence was 5.96%.A total of 58 strains of pathogenic bacteria were isolated from 43 patients with chronic renal failure and bacterial infection of lower respiratory tract in hospital,including 33 strains of Gram-negative bacteria(56.90%),22 strains of Gram-positive bacteria(37.93%)and 3 strains of Fungi(5.17%).Age≥60 years old,smoking history,hospital stay≥15 days,unreasonable use of antibiotics,Hb<60 g/L,Alb<35 g/L in the model group were the risk factors of chronic renal failure with bacterial infection of lower respiratory tract in hospital(P<0.05).The above risk factors were used as predictors to construct a risk prediction nomogram model for chronic renal failure and bacterial infection of lower respiratory tract in hospital in the model group.The Bootstrap method was used for verification.The consistency index(C-index)of the model group was 0.829.The C-index of the verification group was 0.825,and the calibration curve was well calibrated.【Conclusion】Age≥60 years old,smoking history,hospital stay≥15 days,unreasonable use of antibiotics,Hb<60 g/L,Alb<35 g/L are all risk factors of chronic renal failure and bacterial infection of lower respiratory tract in hospital.The nomograph prediction model constructed by combining the above factors have a good application value for individualized prediction of the occurrence of chronic renal failure and bacterial infection of lower respiratory tract in hospital.