| BackgroundA large number of patients with bacterial infections appear every year around the world,and bacterial infections have quietly become one of the leading causes of death for many patients in the world.However,due to the characteristics of blood system diseases and the need for long-term and high-dose application of chemotherapy drugs,glucocorticoids and various immunosuppressive drugs in the course of disease treatment,it is difficult for patients in the department of hematology to avoid damage to immune function.Many patients often have different degrees of agranulocytosis,and the infection rate is very high.Bacterial infection has also become one of the common complications of hematology patients.Carbapenem antibiotics are one of the most commonly used antibacterial drugs in hematology department and have saved the lives of a large number of patients.However,with the continuous emergence of carbapenem antibiotic-resistant strains in recent years,the management and treatment of infected patients has also become more difficult,and the mortality rate of patients with blood diseases associated with bacterial infections continued to rise.Understanding the risk factors of poor prognosis in patients with carbapenem-resistant organism infection in the Department of Hematology,and quantitatively assessing the risk of death of patients is beneficial to the management and treatment of infected patients,and is of great significance to the patients themselves and the majority of clinicians.Object(1)To explore the influencing factors of death in patients who are with carbapenem-resistant organism infection in the Department of Hematology;(2)To establish a mortality risk prediction model for the patients with carbapenem-resistant organism infection in hematology department,and to evaluate the predictive value of the model for the prognosis of patients.MethodsThe clinical data of 168 patients with carbapenem-resistant organism infection were selected and inputted,who were hospitalized in the Department of Hematology,First Affiliated Hospital of Zhengzhou University from November 2018 to February 2021.The patients,who were in hospital from November 2018 to October 2020,were used as the modeling group(129 cases in total),and the patients who were in hospital from November 2020 to February 2021,were used as the validation group(39 cases in total).The patients in the modeling group were divided into survival subgroup and death subgroup according to their status at the time of discharge,with 61 and 68 cases.respectively.Between the survival subgroup and the death subgroup,patients were compared for general conditions,antibiotic use,blood diseases,some of the patients’biochemistry,whether they had had invasive procedures,had received hematopoietic stem cell,had been intensive-care unit,and had other underlying diseases.The death risk factors of the infected patients were obtained by single factor analysis after comparing the death sub-group and the survival sub-group.In order to include more risk factors to the next analysis,the index P<0.20 obtained from single factor analysis was further included in the multiple Logistic regression equation,and establish the factors affecting the risk of death in patients with carbapenem-resistant organism infection in hematology department,and the correlation coefficients from the results of multivariate binary Logistic regression equation were used,to establish and evaluate a predictive model of post-infection mortality risk in patients with carbapenem-resistant bacteria in hematology department.The receiver operating characteristic(ROC)curves of the model group and the validation group were drawn to predict the mortality risk of carbapenem-resistant organism infection.ResultsIn the model group,patients in the death subgroup and the survival subgroup had statistical significance in the primary disease type of blood system,the days of hospitalization before infection of drug resistant bacteria,the days of antibiotic use before infection,the kinds of antibiotics used before infection,the most kinds of antibiotics used together before infection,the use of Carbapenem antibiotics before infection,the use of other special grade antibiotics before infection,the neutrophil count and the duration of agranulocytosis,etc.,p<0.05.Whether suffering from cardiovascular and cerebrovascular diseases,received chemotherapy,received hematopoietic stem cell transplantation(HSCT),used advanced antibiotics before infection,serum albumin before infection,etc.in the univariate analysis results were 0.05<P<0.20.According to the principle of introducing independent variables into the logistic regression equation,the factors with P<0.20 in the univariate analysis were substituted into the multivariate regression analysis equation as independent variables,and the results showed that Logistic(p)=0.833 ×a+1.803 ×b+0.130 ×c-0.107 ×d+0.928×e-0.320×f-2.560(In which,whether suffering from cardiovascular and cerebrovascular diseases is recorded as a,whether HSCT has been performed is recorded as b,the hospitalization time before infection is recorded as c,the antibiotic use time before infection is recorded as d,and whether the other types of special use grades have been used before infection.Antibiotics were recorded as e,and neutrophil counts before infection were recorded as f).The likelihood ratio chi-square test showed:=37.388,p<0.00001;the H-L test showed:chi-square=7.787,P=0.455.The area under the ROC curve(AUC)was 0.804,the best cut-off value was 0.4531,the sensitivity was 86.8%,the specificity was 65.6%,and the Youden index was 0.524.The area under the ROC curve(AUC)was 0.810,the best cut-off value was 0.4994.The sensitivity,specificity and Youden index were 73.9%,81.2%and 0.551 respectively.ConclusionFor the death of patients infected with carbapenem-resistant organism in the blood department,there are the independent risk factors were history of HSCT,long pre-infection hospital stay,inadequate pre-infection antibiotic use,use of other special grade antibiotics,history of cerebrovascular disease,and low pre-infection neutrophil count.Using these risk factors as predictors,we can construct a model to predict the risk of death.The model has a good effect in predicting the prognosis of patients with carbapenem-resistant organism infection in hematology department. |
PDF Full Text Request