简易Nomogram图模型在评估成年人肺炎支原体肺炎严重程度中的作用

The role of a simple Nomogram model in evaluating the severity of mycoplasma pneumonia pneumonia in adults

目的探讨简易Nomogram图模型在评估成年人肺炎支原体肺炎(MPP)严重程度中的作用。方法回顾性分析温州市中心医院2015年3月至2022年10月收治的MPP患者162例的临床资料,根据是否确诊为重症MPP分为重症组(n=67)与普通组(n=95)。提取患者的临床资料,筛选14个临床变量,包括年龄、性别、发病季节、是否发热、热峰、发热持续时间、咳嗽持续时间、白细胞计数、中性粒细胞比例、淋巴细胞比例、血红蛋白、血小板计数、C反应蛋白(CRP)、降钙素原(PCT),对单因素分析中有统计学意义的变量进行多因素logistic回归分析,应用R语言软件包(3.6.2版本)构建Nomogram图模型,并运用校准曲线及受试者工作特征曲线(ROC曲线)对模型进行验证。结果单因素分析结果显示,重症组的热峰更高(Z=5.03,P<0.001),发热(χ^(2)=27.55,P<0.001)、咳嗽时间(χ^(2)=28.72,P<0.001)更长,白细胞总数(t=2.93,P=0.004)、中性粒细胞比例(t=9.08,P<0.001)、CRP(t=35.05,P<0.001)及PCT(t=15.09,P<0.001)高于普通组,淋巴细胞比例(t=11.16,P<0.001)、血红蛋白(t=1.22,P<0.001)及血小板(t=2.82,P<0.001)则低于普通组,差异均有统计学意义。多因素logistic回归分析结果显示,热峰、咳嗽时间、CRP与MPP严重程度呈正相关,淋巴细胞比例、血红蛋白浓度及血小板计数与MPP严重程度呈负相关(均P<0.05)。Nomogram图模型的建立及验证结果显示,模型的准确性较好,敏感度为88.73%,特异度为77.61%,C指数为0.904。结论热峰、咳嗽持续时间、淋巴细胞比例、血红蛋白浓度、血小板计数和CRP等指标与早期MPP的严重程度密切相关,简易Nomogram图模型可成为早期评估MPP严重程度的工具之一。

Objective To investigate the role of a simple Nomogram model in evaluating the severity of mycoplasma pneumoniae pneumonia(MPP)in adults.Methods The clinical data of 162 patients with MPP who received treatment in Wenzhou Central Hospital from March 2015 to October 2022 were retrospectively analyzed.These patients were divided into a severe group(n=67)and a common group(n=95)according to whether they were diagnosed with severe MPP.The clinical data of patients were recorded.Fourteen clinical variables were screened,including age,sex,onset season,fever,heat peak,fever duration,cough duration,white blood cell count,percentage of neutrophils,percentage of lymphocytes,hemoglobin,platelet count,C-reactive protein,and procalcitonin.Multivariate logistic regression analysis of statistically significant variables in univariate analysis was performed.The Nomogram model was constructed with the R language software package(version 3.6.2).The model was verified with a calibration curve and receiver operating characteristic curve.Results Univariate analysis results showed that in the severe group,the fever peak(Z=5.03,P<0.001)was higher,fever duration(χ^(2)=27.55,P<0.001),and cough duration(χ^(2)=28.72,P<0.001)were longer,white cell count(t=2.93,P=0.004),percentage of neutrophils(t=9.08,P<0.001),C-reactive protein(t=35.05,P<0.001),and procalcitonin level(t=15.09,P<0.001)were greater compared with the common group.The percentage of lymphocytes(t=1.16,P<0.001),hemoglobin level(t=1.22,P<0.001),and platelet count(t=2.82,P<0.001)in the severe group were significantly lower than those in the common group.Multivariate logistic regression analysis results showed that heat peak,cough duration,and C-reactive protein were positively correlated with the severity of MPP(all P<0.05).The percentage of lymphocytes,hemoglobin concentration,and platelet count were negatively correlated with the severity of MPP(all P<0.05).The establishment and validation results of the Nomogram model showed that the accuracy of the model was good,with a sensitivity of 88.73%,a specificity of 77.61%,and a C-index of 0.904.Conclusion Heat peak,cough duration,percentage of lymphocytes,platelet count,and C-reactive protein are closely related to the severity of early MPP.A simple Nomogram model can be one of the tools for early assessment of the severity of MPP.