急性白血病患者化疗后肺部感染特点与影响因素及风险预测模型

Characteristics and influencing factors of lung infection in patients with acute leukemia after chemotherapy and risk prediction model

目的探究急性白血病患者化疗后肺部感染特点及易感因素,并建立风险预测模型。方法选取2016年1月-2020年5月开封市中心医院血液内科收治的86例急性髓系白血病患者,均给予化疗治疗,对肺部感染的病原菌分布情况及临床特征进行统计,并分析急性白血病患者化疗后肺部感染发生的影响因素;建立急性白血病患者化疗后肺部感染的风险预测模型,并采用受试者工作特征曲线(ROC曲线)进行验证。结果35例肺部感染患者的痰标本中共检出41株病原菌,其中革兰阴性菌、革兰阳性菌、真菌分别占73.17%、24.39%、2.44%(P<0.05);肺部感染患者首发症状主要为发热、咳嗽、咳痰等,主要症状为咳嗽、咳痰等,体征为肺湿罗音、呼吸音减弱等,CT征象包括磨玻璃影、实变影、索条影等;年龄≥60岁、有侵入性操作、未使用抗生素、白细胞计数<1×10^(9)/L、住院时间≥20 d是导致急性白血病患者化疗后肺部感染发生的影响因素(P<0.05);急性白血病患者化疗后肺部感染预测模型为P=1/[1+e^((0.182+0.653×年龄+0.592×侵入性操作-0.312×使用抗生素-0.285×白细胞数目+0.564×住院时间))],经ROC曲线显示该预测模型预测肺部感染发生的AUC为0.921,95%CI为0.857~0.984。结论急性白血病患者化疗后肺部感染的主要病原菌为革兰阴性菌,且基于影响因素建立的预测模型对肺部感染风险具有预测价值。

OBJECTIVE To explore the characteristics and susceptibility factors of lung infection in patients with acute leukemia after chemotherapy,and to establish the risk prediction model.METHODS A total of 86 patients with acute myeloid leukemia admitted to Hematology Department of Kaifeng Central Hospital were enrolled between Jan 2016 and May 2020.All underwent chemotherapy.The distribution and clinical characteristics of lung infection pathogens were statistically analyzed.The risk factors for lung infection were analyzed.The risk prediction model for lung infection was established,which was verified by receiver operating characteristic curves(ROC curves).RESULTS There were 41 strains of pathogens in the sputum samples from 35 patients with lung infection.The detection rates of Gram-negative bacteria,Gram-positive bacteria and fungi were 73.17%,24.39%and 2.44%,respectively(P<0.05).The first-onset symptoms in patients with lung infection included fever,cough and expectoration,and main symptoms included cough and expectoration.The signs included wet lung rales and reduced breath sounds.CT signs included ground-glass opacity,consolidation shadow and cord-like shadow.Age≥60 years,invasive operation,no usage of antibiotics,white blood cell count<1×10^(9)/L and hospitalization time≥20 d were risk factors for lung infection in patients with acute leukemia after chemotherapy(P<0.05).The prediction model of risks for lung infection was as the following:P=1/[1+e^((0.182+0.653×age+0.592×invasive operation-0.312×antibiotics use-0.285×leukocyte count+0.564×hospitalization time))].ROC curve showed that AUC and 95%CI of the prediction model for predicting lung infection were 0.921 and 0.857-0.984,respectively.CONCLUSION The main pathogen of lung infection in patients with acute leukemia after chemotherapy is Gram-negative bacteria.The prediction model of lung infection based on the risk factors has predictive value for lung infection.