社区获得性肺炎患儿感染多重耐药菌的危险因素评估模型

摘要

Objective To explore the risk factors of multi-drug resistant bacteria infection in children with community-acquired pneumonia,and to establish the relevant evaluation model.Methods The clinical data of children with community acquired pneumonia(CAP)hospitalized in the department of respiration of Shanghai Children′s Medical Center from Jan. 1st 2014 to Dec.31st 2015 were collected.Children infected with multi-drug resistant strains were allocated to a case group,children not infected with multi-drug resistant bacteria were included in a control group,and the clinical data of the two groups were analyzed by binary Logistic regression analysis,and the evaluation model was established.The receiver operating characteristic(ROC)curve was used to evaluate the diagnostic accuracy.Results Altogether 1 303 cases were included,and binary Logistic regression analysis showed that age less than 3 years old,expectoration,shortness of breath,the usage of antibiotics 30 days before the diagnosis of CAP,history of preterm birth,blood oxygen saturation≤0.92 when admitted,interleukin-6≥15 ng/L and chest X-ray showing two lungs′involvement were independent risk factors of CAP children infected by multi-drug resistance bacteria(all P<0.05).The area under the ROC curve(AUC)of CAP children infected by multi-drug resistant bacteria was 0.825(SE=0.029,95%CI 0.767-0.882,P<0.001)under the established scoring system.The best diagnostic boundary value for the evaluation model scoring was 5 with a sensitivity of 70.3% and specificity of 83.7%.Conclusion This study suggests that the risk factors evaluation model has good accuracy in diagnosis of CAP children infected by multi-drug resistant bacteria.It can help pediatricians identifying children infected by multi-drug resistant bacteria to prescribe broad-spectrum antibiotics rationally.%目的 探究社区获得性肺炎(CAP)患儿感染多重耐药菌的危险因素,试建立相关评估模型.方法 收集2014年1月1日至2015年12月31日上海儿童医学中心呼吸科住院的确诊为CAP患儿的临床资料,将感染多重耐药细菌的患儿纳入病例组,未感染多重耐药细菌的病例纳入对照组,对两组临床数据进行二分类Logistic回归分析,构建CAP患儿感染多重耐药菌的危险因素评估模型,利用受试者工作特征曲线(ROC曲线)评价其诊断准确性.结果 共纳入1303例病例,二元Logistic回归分析发现年龄<3岁、咳痰、气促、确诊肺炎前30 d有抗生素使用史、早产、入院时脉搏血氧饱和度≤0.92、白细胞介素6≥15 ng/L以及胸部X线提示两肺受累是儿童CAP多重耐药的危险因素(均P<0.05),构建的联合评分系统判断CAP患儿感染多重耐药菌的ROC曲线下面积为0.825 (SE=0.029,95%CI 0.767~0.882,P<0.001),联合评分的最佳诊断界值为5分,其灵敏度为70.3%,特异度为83.7%.结论 本研究提出的CAP患儿多重耐药的危险因素评估模型效能较高,可以帮助儿科医师早期识别需要应用广谱抗生素的多重耐药菌感染患儿.