平均动脉压联合心率对脓毒症患者预后评估作用的回顾性研究

Retrospective study on the prognostic assessment by mean arterial pressure combined with heart rate in patients with sepsis

目的回顾性分析脓毒症患者的临床资料,探索一个评估脓毒症预后的指标,并验证其有效性及准确性。方法筛选危重症监护数据库(MIMIC)-IV数据库中,在重症监护室(ICU)住院且诊断为脓毒症的患者6548例。导出平均动脉压(MAP)和心率(HR)等相关临床资料。据确诊脓毒症28 d内患者死亡情况分为存活组4306例和死亡组2242例。使用受试者操作特征(ROC)曲线评估指标的预测作用,根据最大AUC数据的MAPmin×√√HR_(min)截断值,分为A组(≤439)2982例和B组(>439)3566例。Cox回归分析各项指标对28 d死亡率的影响,采用Kaplan-Meier法评估两组间预后差异,最后建立模型并验证指标的有效性。结果ROC曲线分析比较,最大的AUC指标MAPmin×√HR_(min)值为0.677。经多因素Cox回归分析发现年龄、合并症(严重肝功能不全、恶性肿瘤)、急性生理学评分Ⅲ、血红蛋白、血尿素氮、乳酸、外周血氧饱和度、MAP、机械通气以及血管活性药物是脓毒症患者的独立预测因素(P<0.05)。构建增加MAPmin×√HR_(min)的模型后发现,其AUC值最大,敏感度、特异度、准确度以及C-index值分别为89.36%、54.64%、69.46%、0.771。比较模型间净重新分类指数以及综合判别改善指标后发现,加入MAPmin×√HR_(min)的模型预测预后能力更强,差异有统计学意义(P<0.01)。结论MAPmin×√HR_(min)是评估脓毒症患者28 d死亡率的有效指标,能有效提高脓毒症患者预后模型的准确性。

Objective To retrospectively analyze the clinical data of patients with sepsis,explore an index for assessing the prognosis of sepsis,and validate its effectiveness and accuracy.Methods A total of 6548 patients diagnosed with sepsis and admitted to the intensive care unit(ICU)were screened from the Medical Information Mart for Intensive Care(MIMIC)-IV database.Relevant clinical data,including mean arterial pressure(MAP)and heart rate(HR),were extracted.The patients were divided into the survival group(4306 cases)and the death group(2242 cases)based on the 28-day mortality of confirmed sepsis patients.The receiver operating characteristic(ROC)curve was used to assess the predictive value of the indicators.According to the MAPmin×√HR_(min)cut-off value determined by the maximum AUC,the patients were divided into Group A(≤439,2982 cases)and Group B(>439,3556 cases).Cox regression analysis was used to assess the impact of various indicators on the 28-day mortality,and Kaplan-Meier method was used to evaluate the prognostic differences between the two groups.Finally,a model was established,and the effectiveness of the indicators was validated.Results ROC curve analysis showed that the maximum AUC indicator MAPmin×√HR_(min)was 0.677.Multivariate Cox regression analysis revealed that age,comorbidities(severe hepatic insufficiency,malignancy),and acute physiology scoreⅢ(APSⅢ),hemoglobin(Hb),blood urea nitrogen(BUN),lactate,peripheral capillary oxygen saturation(SpO2),MAP,mechanical ventilation,and vasoactive drugs were independent prognostic factors for patients with sepsis(P<0.05).After including MAPmin×√HR_(min)into the model,the AUC value of the model was optimal,and the sensitivity,specificity,accuracy and C-statistic values were 89.36%,54.64%,69.46%and 0.771.By comparing the net reclassification index(NRI)and integrated discrimination improvement(IDI)index between the models,it was found that the model with the addition of MAPmin×√HR_(min)could effectively improve the prognostic effecacy for patients with sepsis,the difference was statistically significant(P<0.01).Conclusion MAPmin×√HR_(min)is an effective indicator for assessing 28-day mortality rate of septic patients and can significantly improve the accuracy of the prognostic model for septic patients.