The Real-world Study In Prognosis And Treatment Of Critically Ill Patients With Sepsis-associated Acute Kidney Injury Based On MIMIC Database

Chapter one:Background of the studySepsis-associated acute kidney injury(SA-AKI)has a high incidence in critically ill patients and is closely related to mortality and the length of staying in hospital.Therefore,early prognosis prediction and treatment of SA-AKI is very important but only a few clinical studies focused on it.The MIMIC database,as a large critical illness database,integrates the collection of real-world data,solves the phenomenon of small sample size and poor data integrity in the local database,and provides great convenience for research of SA-AKI disease severity and prognosis.Chapter two:A prediction model for assessing prognosis in critically ill patients with SA-AKI.Objective:To investigate the possible predictors of prognosis in critically ill patients with SA-AKI using real-world data and develop a prediction model.Methods:We conducted a retrospective cohort analysis based on a training cohort of 2066 patients enrolled from the Multiparameter Intelligent Monitoring in Intensive Care Database ?(MIMIC ?)and a validation cohort of 102 patients treated at local intensive care unit.Least absolute shrinkage and selection operator(LASSO)regression and multivariate Cox regression analysis were used to identify predictors for survival.Areas under the ROC curves(AUC),the concordance index(C-index)and calibration curves were used to evaluate the efficiency of the prediction model in both cohorts.Results:The overall mortality of SA-AKI was approximately 18%.Age,admission type,liver disease,metastatic cancer,lactate,BUN/SCr,admission creatinine,positive culture and AKI stage were independently associated with survival and combined in the Cox regression model.The C-index in the training and validation cohorts was 0.73 and 0.72.The AUC in the training cohort was 0.77,0.72,and 0.70 for the 7-day,14day and 28-day probability of in-hospital survival,respectively,while in the external validation cohort,it was 0.83,0.73 and 0.67.SAPSII and SOFA scores showed poorer performance.Calibration curves demonstrated a good consistency.Conclusion:The prognosis model has predictive value for in-hospital mortality of SAAKI in critically ill patients and outperforms generic scores.Chapter three:Association between dexmedetomidine administration and outcomes in critically ill patients with SA-AKI.Objective:To investigate the association between dexmedetomidine administration and outcomes in critically ill patients with SA-AKI using real-world data.Methods:Critically ill patients with SA-AKI were identified from the Medical Information Mart for Intensive Care(MIMIC)-? database.Propensity score matched(PSM)analysis was used to match patients receiving dexmedetomidine to those without treatment.Linear regression,logistic regression model,and Cox proportional hazards model were used to assess the associations between dexmedetomidine and inhospital mortality,recovery of renal function,vasopressor requirements and length of stay,respectively.Results:A total of 2192 SA-AKI patients were included in the data analysis.After PSM,710 pairs of patients were matched between the patients who received dexmedetomidine(DEX group)and those without treatment.Dexmedetomidine was associated with the higher recovery of renal function(61.9%vs 55.5%,HR 1.35;P=0.0l),and reduced in-hospital mortality(28.3%vs 41.3%,HR 0.56;P<0.001)as well as prolonged ICU stay(15.8 d vs 12.6 d,HR 2.34,P<0.001)and hospital stay(23.7d vs 19.7d,HR 4.47;P<0.001).No significant difference was found in vasopressor requirements in over-all SA-AKI patients.Conclusion:Dexmedetomidine administration was associated with improvement inhospital survival and recovery of renal function in critically ill patients with SA-AKI.