Risk Factors Of Mortality And Establishment Of A Prediction Score System For The Hospitalized Patients With Acute Kidney Injury

Objective1. To study the risk factors of90days and1year mortality in the acute kidney injury for hospitalized patients.2. To establish a prediction score system for the hospitalized patients with acute kidney injury to predict the prognosis of AKI.Methods1. We constructed a study of1169hospitalized AKI patients at the Second Xiangya Hospital of Central South University between January1996to April2013. The diagnosis and classification of AKI were reconfirmed and categorized according to KDIGO criteria. In this study, the731AKI patients between January1996to December2007were assigned as the test group, and the438AKI patients between January2008to April2013as the validation group. We collected the patient’s clinical basic data, followed-up the90days and1year for every patient. Calculatting the90days and1year mortality. The test group was divided into Surviving group and death group. Multivariate logistic regression analysis was applied to identify the risk factors for AKI. Based on the odds ratios, we derived a new prediction score system, from which the forewarning model for AKI mortality within90days and1year were established. The ROC curve and chi-squared test were used to assess the accuracy and efficacy of the scoring system.Results1. The90days mortality of the test group and validation group were13.8%and11.6%. The1year mortality of the test group and validation group were14.8%and12.6%.2. Univariate analysis suggest the age, type of AKI, AKI etiology, mechanical ventilation, hypotension, Shock, heart failure, respiratory failure, gastrointestinal failure, central nervous system failure, peak blood urea nitrogen, peak serum potassium and ATN-ISI score in the prognosis of90days and1year have statistically significant difference. Multivariate Logistic Regression Analysis suggest age, the type of AKI, respiratory failure, central nervous system failure, hypotension, and ATNISI score>0.4are the independent risk factors for hospitalized patients with acute kidney injury in90days; age, the type of AKI, respiratory failure, central nervous system failure and hypotension are the independent risk factors for hospitalized patients with AKI in1year.3. Using the chi-square trend test to test prognostic scoring systems of90days and1year, showing statistically significant changes in mortality(P<0.001). The area under the ROC curve of predictting90days and1-year mortality were0.833and0.817.4. In90days and1year early-warning model of prognosis, the mortality between test group and the validation group were not statistically different(x2=1.7958,P=0.1802; x2=0.1006, P=0.7511). The scoring system can accurately predict the mortality of AKI patients developed during hospitalization.Conclusions1. The mortality of the hospitalized patients with acute kidney injury was11.6%-14.8%.2. Six variables were identified as the independent risk factors for AKI, including age, the type of AKI, respiratory failure, Central nervous system failure, Hypotension, and ATNISI score>0.4.3. We establish two prediction score system for90days and1year mortality hospitalized patients with acute kidney injury, respectively. The higher the score, the higher the mortality rate, it maybe help the clinicians to identify the high risk factors of patients early.