重症肺炎患儿预后的危险因素及其列线图预测模型构建研究

Risk Factors and Nomogram Prediction Model of the Prognosis of Children with Severe Pneumonia

背景小儿重症肺炎的发生与病毒感染、环境、免疫功能等多种因素有关,临床常采用机械通气、抗感染等治疗措施缓解患儿的临床症状,但其病死率仍居高不下,因此分析重症肺炎患儿预后的影响因素并进行个体化干预具有重要的临床意义。目的探讨重症肺炎患儿死亡的危险因素,并构建列线图预测模型,以期为不良预后患儿的管理提供一定指导。方法回顾性分析2017年2月至2021年2月扬州大学附属医院收治的210例重症肺炎患儿的临床资料。根据预后将患儿分为死亡组与存活组,比较两组临床资料。采用多因素Logistic回归分析探讨重症肺炎患儿预后的影响因素;使用R软件(R 3.6.3软件包)和rms程序包构建预测重症肺炎患儿预后的列线图模型;采用Bootstrap法(原始数据重复抽样100次)对列线图模型进行内部验证,采用一致性指数(CI)、校正曲线、受试者工作特征(ROC)曲线及Hosmer-Lemeshow拟合优度检验评估列线图模型对重症肺炎患儿预后的预测价值。结果重症肺炎患儿均至少检出1种病原菌,共检出病原菌232株,其中革兰阴性菌占比最高,为60.34%。所有患儿中,死亡31例,存活179例。单因素分析结果显示,两组患儿年龄、并发症情况、病程、机械通气情况、感染病原菌株数比较,差异有统计学意义(P<0.05)。多因素Logistic回归分析结果显示,年龄≥3岁〔OR=0.852,95%CI(0.789,0.919)〕、有并发症〔OR=2.968,95%CI(1.543,5.709)〕、病程>7 d〔OR=4.421,95%CI(1.628,12.004)〕、机械通气〔OR=2.648,95%CI(1.353,5.186)〕、感染病原菌株数≥2株〔OR=2.168,95%CI(1.123,4.188)〕是重症肺炎患儿死亡的独立影响因素(P<0.05)。基于影响重症肺炎患儿预后的独立危险因素建立列线图模型,内部验证结果显示:CI为0.826;列线图模型预测重症肺炎患儿预后的ROC曲线下面积为0.822〔95%CI(0.751,0.994)〕。列线图模型预测重症肺炎患儿预后的校正曲线趋近于理想曲线;Hosmer-Lemeshow拟合优度检验结果显示,列线图模型预测重症肺炎患儿死亡风险的一致性良好(χ^(2)=12.351,P=0.117)。结论年龄<3岁、有并发症、病程>7 d、机械通气、感染病原菌株数≥2株是重症肺炎患儿死亡的独立危险因素,基于这些危险因素构建的列线图模型对预测重症肺炎患儿预后具有良好的准确性、区分度及一致性。

Background The occurrence of severe pneumonia in children is related to multiple factors such as viral infection,environment and immune function.The clinical treatment measures of mechanical ventilation,anti-infection can alleviate the clinical symptoms of children,but the mortality rate is still high.Therefore,it is of important clinical significance to analyze the influencing factors for the prognosis of children with severe pneumonia and carry out individualized intervention.Objective To explore the risk factors for the prognosis of children with severe pneumonia,and to construct a Nomogram prediction model,in order to provide certain guidance for the management of children with poor prognosis.Methods The clinical data of 210 children with severe pneumonia admitted to the Affiliated Hospital of Yangzhou University from February 2017 to February 2021 were retrospectively analyzed,which were divided into the death group and the survival group according to the prognosis of children.Clinical data were compared between the two groups.Multivariate Logistic regression analysis was used to explore the influencing factors for the prognosis of children with severe pneumonia;R software(R 3.6.3 software package)and rms program package were used to construct a Nomogram model for predicting the prognosis of children with severe pneumonia;Bootstrap method(repeat sampling 100 times)was used to internally verify the Nomogram model,the consistency index(CI),calibration curve,receiver operating characteristic curve(ROC curve)and Hosmer-Lemeshow goodness of fit test were used to evaluate the predictive value of Nomogram model for prognosis of children with severe pneumonia.Results At least one pathogen was detected in children with severe pneumonia,and a total of 232 pathogens were detected,of which gram-negative bacteria accounted for the highest proportion,which was 60.34%.Among all children,31 died and 179 survived.The results of univariate analysis showed that,there was significant difference of age,complications,course of disease,mechanical ventilation and infection with pathogenic strains between the two groups(P<0.05).The results of multivariate Logistic regression analysis showed that,age≥3 years old[OR=0.852,95%CI(0.789,0.919)],complipcated with complications[OR=2.968,95%CI(1.543,5.709)],course of disease>7 days[OR=4.421,95%CI(1.628,12.004)],mechanical ventilation[OR=2.648,95%CI(1.353,5.186)],infection with pathogenic strains≥2 strains[OR=2.168,95%CI(1.123,4.188)]were independent influencing factors for the prognosis of children with severe pneumonia(P<0.05).A Nomogram prediction model was established based on independent risk factors affecting the prognosis of children with severe pneumonia,and the internal verification results showed that,CI was 0.826;the area under the ROC curve of the Nomogram model for predicting the prognosis of children with severe pneumonia was 0.822[95%CI(0.751,0.994)].The calibration curve of the Nomogram model in predicting the prognosis of children with severe pneumonia was closed to the ideal curve;the Hosmer-Lemeshow goodness of fit test results showed that,the Nomogram model in predicting the prognosis in children with severe pneumonia had good agreement(χ^(2)=12.351,P=0.117).Conclusion Age<3 years old,complipcated with complications,course of disease>7 days,mechanical ventilation,infection with pathogenic strains≥2 strains are independent risk factores for the prognosis of children with severe pneumonia,and the Nomogram model constructed based on these risk factors has good accuracy,discrimination and consistency.