血常规指标在COVID-19和甲流病毒感染鉴别中的价值及预测模型的构建

目的探讨血小板压积、淋巴细胞百分比和单核细胞百分比在COVID-19和甲型流感病毒感染鉴别诊断中的价值及预测模型构建。方法收集2022年12月至2023年6月感染新型冠状病毒(SARS-CoV-2)和甲型流感病毒的患者101例,对两组患者血常规指标进行统计学方法。应用Logistic回归分析构建诊断SARS-CoV-2感染和甲型流感病毒感染的10个预测模型;随机选择各100例确诊患者的血常规指标代入10个预测模型进行验证;绘制ROC曲线评估各指标及10个预测模型诊断效能。结果单因素分析显示,SARS-CoV-2与甲型流感病毒感染患者的嗜碱性粒细胞计数、中性粒细胞百分比、淋巴细胞百分比、单核细胞百分比、红细胞分布宽度-CV、血小板计数、血小板压积差异有统计学意义(P<0.05)。应用ROC曲线显示,血小板压积、淋巴细胞百分比和单核细胞百分比联合模型9:Y=10.023×PCT+0.051×LYMPH_p-0.140×MONO_p-1.953鉴别诊断COVID-19的效能最好,测试集和验证集的AUC分别为0.737和0.756。据测试集最优模型ROC曲线确定的最佳cut-off值,在验证集中进行测试,仍可达到较好特异度。结论血小板压积、淋巴细胞百分比和单核细胞百分比模型在SARS-CoV-2感染和甲型流感病毒感染鉴别诊断中具有一定临床指导意义。

Objective To explore the value of platelet ocrit,lymphocyte percentage and monocyte percentage in the differential diagnosis of COVID-19 and influenza A virus infection and the derivation of the prediction model.Methods Collecting data of 101 patients successively infected with SARS-CoV-2 and A influenza virus in the Affiliated Hospital of Shaoxing University from December 2022 to June 2023,statistically analysed the blood routine indexes and their ratio of the two groups;multivariate logistic regression analysis was applied to derive 10 prediction models for diagnose COVID-19 and influenza A virus infection.The blood routine indexes of 100 confirmed patients were randomly selected into 10 prediction models for validation.Receiver operating characteristic(ROC)curves were drawn to evaluate the efficacy of each single index and 10 indicators to predict the differential diagnosis of COVID-19.Results Univariate statistical analysis showed that basophil count,neutrophil percentage,lymphocyte percentage,percentage of monocytesred,blood cell distribution width-CV,platelet count and platelet pressure were statistical differences present in patients successively infected with SARS-CoV-2and A influenza virus(P<0.05).Applying the ROC to evaluate the results of the univariate and 10 logistic regression models and the validation set found that the combined model of platelet ocrit,lymphocyte percentage and monocyte percentage standard 9:Y=10.023×PCT+0.051×LYMPH_p-0.140×MONO_p-1.953 had the best efficacy for differential diagnosis COVID-19,The AUC for the test and validation sets were 0.737 and 0.756.According to the best cutoff value determined by the optimal model of ROC curve in the test set,the test could still achieve good specificity in the verification set.Conclusion The platelet ocrit,lymphocyte percentage and monocyte percentage model have clinical significance in the differential diagnosis of COVID-19 and A virus infection.