利用临床特征与生化指标建立细菌性脑膜炎转归不良的风险预测模型  

Establishing a risk prediction model for adverse outcomes of bacterial meningitis using clinical features and biochemical indicators

作　　者：李博[1] 迟洁[1] 高蒙 Li Bo;Chi Jie;Gao Meng(Department of Laboratory Medicine,Shaanxi Provincial People's Hospital,Xi'an 710068,China;Department of Laboratory Medicine,General Hospital of Ordnance Industry,Xi'an 710068,China)

机构地区：[1]陕西省人民医院检验科,西安710068 [2]兵器工业总医院检验科,西安710068

出　　处：《国际医药卫生导报》2025年第4期663-668,共6页International Medicine and Health Guidance News

基　　金：陕西省自然科学基础研究计划(2022JQ-953)。

摘　　要：目的分析细菌性脑膜炎患儿的临床特征与生化指标,构建转归不良的风险预测模型。方法对陕西省人民医院2020年9月至2024年5月收治的274例细菌性脑膜炎患儿的临床资料展开回顾性分析,其中男156例、女118例,年龄范围0~18岁。根据患儿院内治疗结局将其分为转归不良组和转归良好组,分析其转归不良的危险因素。利用临床特征与生化指标建立细菌性脑膜炎转归不良的风险预测列线图模型,并进一步评价模型的区分度、预测性能及实用性。采用Mann-Whitney U、t、χ^(2)检验进行统计比较,模型的区分度、预测性能、实用性分别采用受试者操作特征曲线(ROC)及曲线下面积(AUC)、校准曲线及Hosmer-Lemeshow检验、决策分析曲线进行评价。结果274例细菌性脑膜炎患儿转归不良65例(23.72%)。反复惊厥发作、脑脊液蛋白>1 g/L、脑脊液葡萄糖/血糖降低、血/脑脊液细菌培养阳性是细菌性脑膜炎转归不良的危险因素(均P<0.05)。基于上述4个危险因素建立细菌性脑膜炎转归不良的风险预测列线图模型,ROC显示模型预测的AUC为0.855(95%置信区间0.808~0.895);校准曲线显示模型预测的曲线与理想曲线走向相同且贴近,Hosmer-Lemeshow检验显示,模型预测转归不良概率与实际转归不良概率接近(P=0.394);决策分析曲线显示,当高风险阈值在0.06~0.77时应用模型预测具有较好价值。结论基于反复惊厥发作、脑脊液蛋白、脑脊液葡萄糖/血糖、血/脑脊液细菌培养4个危险因素构建细菌性脑膜炎转归不良的风险预测列线图模型具有良好的应用价值。Objective To analyze the clinical characteristics and biochemical indicators of children with bacterial meningitis,and to construct a risk prediction model for poor outcomes.Methods A retrospective analysis was conducted on the clinical data of 274 children with bacterial meningitis treated at Shaanxi Provincial People's Hospital from September 2020 to May 2024.There were 156 boys and 118 girls.They were 0-18 years old.The children were divided into a poor outcome group and a good outcome group based on the in-hospital treatment outcomes.The clinical characteristics and biochemical indicators of the two groups were compared,and the risk factors for poor outcomes were analyzed.A nomogram model for predicting the risk of adverse outcomes in the children with bacterial meningitis was established using the clinical features and biochemical indicators.The model's discrimination,predictive performance,and practicality were evaluated.Mann-Whitney U test,t test,bandχ^(2)test were used for the statistical comparisons.The model's discriminability was evaluated by the receiver operating characteristic curve and the area under the curve.The model's performance was assessed by the calibration curve and Hosmer-Lemeshow test.The model's practicality was evaluated by the decision curve analysis.Results Among the 274 children,65 cases had poor outcomes,with a poor outcome rate of 23.72%.Repeated convulsive seizures,cerebrospinal fluid protein>1 g/L,decreased cerebrospinal fluid glucose/blood glucose,and positive blood/cerebrospinal fluid bacterial culture were risk factors for the childrens poor outcomes(all P<0.05).A nomogram model was established to predict the risk of adverse outcomes in the children with bacterial meningitis based on the four risk factors mentioned above.The receiver operating characteristic curve showed that the area under the curve predicted by the model was 0.855(95%CI 0.808-0.895).The calibration curve showed that the predicted curve of the model had the same and close direction as the ideal curve;the Ho

关 键 词：细菌性脑膜炎 儿童 临床特征 生化指标 病情转归 列线图模型 

分 类 号：R72[医药卫生—儿科]