Background Streptococcus pneumoniae,as a respiratory tract common pathogen,can cause invasive disease and sepsis.This study aimed to investigate the association of factors with clinical outcomes in sepsis with strepto...Background Streptococcus pneumoniae,as a respiratory tract common pathogen,can cause invasive disease and sepsis.This study aimed to investigate the association of factors with clinical outcomes in sepsis with streptococcus pneumoniae infection based on MIMIC-IV database.Methods The sepsis patients with streptococcus pneumoniae infection were included.Different variables between the survivor group and the non-survivor group were analyzed.Multivariable logistic regression was applied to identify the factors which were associated with clinical outcomes.Results A total of 80 sepsis patients with streptococcus pneumoniae infection were included.The in-hospital mortality was 23.75%(n=19).Significant differences were found in heart rate,white blood cell,RDW,MCV and hematocrit between the survivor group and the non-survivor group.The area under the ROC curve of hematocrit was 0.758 with a sensitivity of 73.7%and a specificity of 72.1%.The cut-off value of hematocrit was 30.8%.Conclusions Hematocrit level was identified to be negatively associated with in-hospital mortality in sepsis with streptococcus pneumoniae infection.展开更多
目的:研究血红蛋白-红细胞分布宽度比值(hemoglobin/red blood cell distribution width ratio,HRR)与冠脉搭桥手术(coronary artery bypass grafting,CABG)后急性肾损伤(acute kidney injury,AKI)发生风险的相关性。方法:选择重症监护...目的:研究血红蛋白-红细胞分布宽度比值(hemoglobin/red blood cell distribution width ratio,HRR)与冠脉搭桥手术(coronary artery bypass grafting,CABG)后急性肾损伤(acute kidney injury,AKI)发生风险的相关性。方法:选择重症监护数据集(Medical Information Mart for Intensive Care Database,MIMIC)-Ⅳ中所有CABG术后患者作为研究对象,根据是否发生AKI分为2组。比较分析2组一般资料,并将有统计学差异的变量纳入logistic单因素回归分析,以单因素分析中P<0.05的变量纳入多因素logistic回归分析。Logistic回归模型评估HRR对CABG术后AKI风险的预测价值。亚组分析采用层次回归模型。结果:共5 623例患者纳入研究,AKI组4 342例,非AKI组1 281例。AKI组患者入院时HRR水平明显低于非AKI组(P<0.001)。多因素回归模型显示HRR是预测CABG术后AKI风险的独立危险因子(OR=0.92,95%CI=0.88~0.96,P<0.001)。Logistic回归模型发现,在模型Ⅲ中(调整潜在混杂因素),HRR水平降低仍然是CABG术后发生AKI的独立影响因素。亚组分析发现HRR与AKI发生的相关性在大多数协变量中相似。结论:低HRR水平是CABG术后AKI发生风险的独立危险因素。展开更多
背景重症监护病房内多重耐药菌(multidrug-resistant organisms,MDRO)感染发生率高,对接受重症监护的急性胰腺炎患者的抗生素治疗带来严峻挑战。目的对解放军总医院(以下简称“我院”)重症胰腺炎治疗中心重症监护病房和美国重症监护医...背景重症监护病房内多重耐药菌(multidrug-resistant organisms,MDRO)感染发生率高,对接受重症监护的急性胰腺炎患者的抗生素治疗带来严峻挑战。目的对解放军总医院(以下简称“我院”)重症胰腺炎治疗中心重症监护病房和美国重症监护医疗信息库Ⅳ版(Medical Information Mart for Intensive Care-Ⅳ,MIMIC-Ⅳ)中急性胰腺炎(acute pancreatitis,AP)病例的病原菌和耐药情况进行分析,为抗感染的经验性治疗提供依据,探讨中美抗生素耐药形势差异及带来的启示。方法回顾性收集我院2018-2019年以及MIMIC-Ⅳ中2014-2019年的AP患者送检标本培养结果和药敏结果,分析AP患者感染的主要病原菌和耐药情况。结果我院AP患者数量为314例,培养鉴定出微生物菌株570株;MIMIC-Ⅳ中AP患者数量为604例,培养鉴定出微生物菌株368株。我院不同标本来源中(除尿液外)的微生物均以肺炎克雷伯菌(17.24%~26.92%)和鲍曼不动杆菌(13.56%~46.15%)为主;而MIMIC-Ⅳ中腹水、尿液和静脉血中检出最多的微生物分别为铜绿假单胞菌(17.39%)、大肠埃希菌(22.22%)和肺炎克雷伯菌(16.90%)等,呼吸道标本和中心静脉导管中以金黄色葡萄球菌(分别为28.37%和100%)最多,胆汁中凝固酶阴性葡萄球菌(21.43%)和金黄色葡萄球菌(14.29%)较多。我院肺炎克雷伯菌和鲍曼不动杆菌中MDRO的占比在90%以上,分别为94.55%和99.02%,铜绿假单胞菌和大肠埃希菌中MDRO的占比在80%以上,分别为82.14%和83.87%;MIMIC-Ⅳ中肺炎克雷伯菌、铜绿假单胞菌和大肠埃希菌中MDRO的占比分别为25.93%,27.59%和26.83%。结论国内外重症数据库中急性胰腺炎患者标本分离到的微生物主要为革兰阴性菌,国外首位为金黄色葡萄球菌,国内首位为肺炎克雷伯菌,主要阴性菌耐药情况较国外明显严重。展开更多
Objective To compare the performance of five machine learning models and SAPSⅡ score in predicting the 30-day mortality amongst patients with sepsis.Methods The sepsis patient-related data were extracted from the MIM...Objective To compare the performance of five machine learning models and SAPSⅡ score in predicting the 30-day mortality amongst patients with sepsis.Methods The sepsis patient-related data were extracted from the MIMIC-Ⅳ database.Clinical features were generated and selected by mutual information and grid search.Logistic regression,Random forest,LightGBM,XGBoost,and other machine learning models were constructed to predict the mortality probability.Five measurements including accuracy,precision,recall,F1 score,and area under curve(AUC) were acquired for model evaluation.An external validation was implemented to avoid conclusion bias.Results LightGBM outperformed other methods,achieving the highest AUC(0.900),accuracy(0.808),and precision(0.559).All machine learning models performed better than SAPSⅡ score(AUC=0.748).LightGBM achieved 0.883 in AUC in the external data validation.Conclusions The machine learning models are more effective in predicting the 30-day mortality of patients with sepsis than the traditional SAPS Ⅱ score.展开更多
Background This study aimed to explore the associations of factors with in-hospital mortality in sepsis with MRSA infection based on the MIMIC-Ⅳdatabase.Methods All sepsis patients with MRSA infection were included.M...Background This study aimed to explore the associations of factors with in-hospital mortality in sepsis with MRSA infection based on the MIMIC-Ⅳdatabase.Methods All sepsis patients with MRSA infection were included.Multivariate regression analysis was applied to identify predictive factors.Models were applied to explore the relationship between the factors and in-hospital mortality after adjusting potential cofounders.Results A total of 157 patients were included including a survivor group(n=111)and a non-survivor group(n=46)were included.RDW was identified as an independent factor with in-hospital mortality.A smooth fitting curve for a positive linear relationship between RDW and in-hospital mortality was constructed.The area under the ROC curve(AUC)was 0.652 with a sensitivity of 0.565 and a specificity of 0.676.The optimal cut-off value of RDW was 16.55.Compared with SOFA and APACHEⅡscores,RDW had a better predictive performance.Conclusion RDW was associated with clinical outcomes in sepsis with MRSA infection.展开更多
文摘Background Streptococcus pneumoniae,as a respiratory tract common pathogen,can cause invasive disease and sepsis.This study aimed to investigate the association of factors with clinical outcomes in sepsis with streptococcus pneumoniae infection based on MIMIC-IV database.Methods The sepsis patients with streptococcus pneumoniae infection were included.Different variables between the survivor group and the non-survivor group were analyzed.Multivariable logistic regression was applied to identify the factors which were associated with clinical outcomes.Results A total of 80 sepsis patients with streptococcus pneumoniae infection were included.The in-hospital mortality was 23.75%(n=19).Significant differences were found in heart rate,white blood cell,RDW,MCV and hematocrit between the survivor group and the non-survivor group.The area under the ROC curve of hematocrit was 0.758 with a sensitivity of 73.7%and a specificity of 72.1%.The cut-off value of hematocrit was 30.8%.Conclusions Hematocrit level was identified to be negatively associated with in-hospital mortality in sepsis with streptococcus pneumoniae infection.
文摘目的:研究血红蛋白-红细胞分布宽度比值(hemoglobin/red blood cell distribution width ratio,HRR)与冠脉搭桥手术(coronary artery bypass grafting,CABG)后急性肾损伤(acute kidney injury,AKI)发生风险的相关性。方法:选择重症监护数据集(Medical Information Mart for Intensive Care Database,MIMIC)-Ⅳ中所有CABG术后患者作为研究对象,根据是否发生AKI分为2组。比较分析2组一般资料,并将有统计学差异的变量纳入logistic单因素回归分析,以单因素分析中P<0.05的变量纳入多因素logistic回归分析。Logistic回归模型评估HRR对CABG术后AKI风险的预测价值。亚组分析采用层次回归模型。结果:共5 623例患者纳入研究,AKI组4 342例,非AKI组1 281例。AKI组患者入院时HRR水平明显低于非AKI组(P<0.001)。多因素回归模型显示HRR是预测CABG术后AKI风险的独立危险因子(OR=0.92,95%CI=0.88~0.96,P<0.001)。Logistic回归模型发现,在模型Ⅲ中(调整潜在混杂因素),HRR水平降低仍然是CABG术后发生AKI的独立影响因素。亚组分析发现HRR与AKI发生的相关性在大多数协变量中相似。结论:低HRR水平是CABG术后AKI发生风险的独立危险因素。
文摘背景重症监护病房内多重耐药菌(multidrug-resistant organisms,MDRO)感染发生率高,对接受重症监护的急性胰腺炎患者的抗生素治疗带来严峻挑战。目的对解放军总医院(以下简称“我院”)重症胰腺炎治疗中心重症监护病房和美国重症监护医疗信息库Ⅳ版(Medical Information Mart for Intensive Care-Ⅳ,MIMIC-Ⅳ)中急性胰腺炎(acute pancreatitis,AP)病例的病原菌和耐药情况进行分析,为抗感染的经验性治疗提供依据,探讨中美抗生素耐药形势差异及带来的启示。方法回顾性收集我院2018-2019年以及MIMIC-Ⅳ中2014-2019年的AP患者送检标本培养结果和药敏结果,分析AP患者感染的主要病原菌和耐药情况。结果我院AP患者数量为314例,培养鉴定出微生物菌株570株;MIMIC-Ⅳ中AP患者数量为604例,培养鉴定出微生物菌株368株。我院不同标本来源中(除尿液外)的微生物均以肺炎克雷伯菌(17.24%~26.92%)和鲍曼不动杆菌(13.56%~46.15%)为主;而MIMIC-Ⅳ中腹水、尿液和静脉血中检出最多的微生物分别为铜绿假单胞菌(17.39%)、大肠埃希菌(22.22%)和肺炎克雷伯菌(16.90%)等,呼吸道标本和中心静脉导管中以金黄色葡萄球菌(分别为28.37%和100%)最多,胆汁中凝固酶阴性葡萄球菌(21.43%)和金黄色葡萄球菌(14.29%)较多。我院肺炎克雷伯菌和鲍曼不动杆菌中MDRO的占比在90%以上,分别为94.55%和99.02%,铜绿假单胞菌和大肠埃希菌中MDRO的占比在80%以上,分别为82.14%和83.87%;MIMIC-Ⅳ中肺炎克雷伯菌、铜绿假单胞菌和大肠埃希菌中MDRO的占比分别为25.93%,27.59%和26.83%。结论国内外重症数据库中急性胰腺炎患者标本分离到的微生物主要为革兰阴性菌,国外首位为金黄色葡萄球菌,国内首位为肺炎克雷伯菌,主要阴性菌耐药情况较国外明显严重。
文摘Objective To compare the performance of five machine learning models and SAPSⅡ score in predicting the 30-day mortality amongst patients with sepsis.Methods The sepsis patient-related data were extracted from the MIMIC-Ⅳ database.Clinical features were generated and selected by mutual information and grid search.Logistic regression,Random forest,LightGBM,XGBoost,and other machine learning models were constructed to predict the mortality probability.Five measurements including accuracy,precision,recall,F1 score,and area under curve(AUC) were acquired for model evaluation.An external validation was implemented to avoid conclusion bias.Results LightGBM outperformed other methods,achieving the highest AUC(0.900),accuracy(0.808),and precision(0.559).All machine learning models performed better than SAPSⅡ score(AUC=0.748).LightGBM achieved 0.883 in AUC in the external data validation.Conclusions The machine learning models are more effective in predicting the 30-day mortality of patients with sepsis than the traditional SAPS Ⅱ score.
文摘Background This study aimed to explore the associations of factors with in-hospital mortality in sepsis with MRSA infection based on the MIMIC-Ⅳdatabase.Methods All sepsis patients with MRSA infection were included.Multivariate regression analysis was applied to identify predictive factors.Models were applied to explore the relationship between the factors and in-hospital mortality after adjusting potential cofounders.Results A total of 157 patients were included including a survivor group(n=111)and a non-survivor group(n=46)were included.RDW was identified as an independent factor with in-hospital mortality.A smooth fitting curve for a positive linear relationship between RDW and in-hospital mortality was constructed.The area under the ROC curve(AUC)was 0.652 with a sensitivity of 0.565 and a specificity of 0.676.The optimal cut-off value of RDW was 16.55.Compared with SOFA and APACHEⅡscores,RDW had a better predictive performance.Conclusion RDW was associated with clinical outcomes in sepsis with MRSA infection.
