儿童社区获得性肺炎进展为重症肺炎的风险预测

Prediction of the risk of community acquired pneumonia progressing to severe pneumonia in children

目的 利用常见实验室检验结果建立区别重症肺炎与普通肺炎的模型,为早期识别重症肺炎提供依据、指导临床决策。方法 收集住院肺炎患儿实验室检查结果,按照最终诊断分为重症肺炎组和普通肺炎组。统计分析检查结果,建立并验证列线图预测模型。结果 共纳入720例患儿:重症肺炎174例,普通肺炎546例,随机分为建模组(70%)和验证组(30%)。发现年龄≤12.5月、中性粒细胞百分比≥58.4%、C-反应蛋白≥4.9 mg/L,白蛋白≤37.7g/L、乳酸脱氢酶>504.5 U/L、谷丙转氨酶≥21.5 U/L等是重症肺炎的独立危险因素,根据Logistic回归结果建立列线图预测模型,验证显示模型预测性能良好。结论 建立的列线图模型可以很好地预测重症肺炎,使用该模型可以确定重症肺炎的风险。

Objective To distinguish severe pneumonia from common pneumonia via a model established with laboratory findings,and to provide evidence for early identification of severe pneumonia so as to guide clinical decision-making.Method The laboratory findings of hospitalized children with pneumonia were collected and divided into a severe pneumonia group and a common pneumonia group according to the final diagnosis.A statistical analysis was made from these data,and a nomogram prediction model was established and verified finally.Results A total of 720 patients,including 174 cases with severe pneumonia and 546 with common pneumonia,were randomly selected for the modeling group(70%)and the validation group(30%).Age≤12.5 months,percentage of neutrophil≥58.4%,C-reactive protein≥4.9 mg/L,albumin≤37.7 g/L,lactate dehydrogenase>504.5 U/L,glutamic-pyruvic transaminase≥21.5 U/L were independent risk factors of severe pneumonia.The prediction model of nomogram was established based on the result of a Logistic regression.The validation of the model with the modeling and the validation group showed that the prediction performance of the model was good.Conclusion The nomogram based on these factors could effectively identify severe pneumonia.This model can help determine the risk of severe pneumonia.