Establishment And Clinical Validation Of The Prediction Model Of Acute Renal Injury After Cardiac Surgery

Part One. Validation of5prediction models for AKI and its outcome after cardiac surgery in Chinese patientsObjective:To assess the clinical usefulness and value of the5models for the prediction of Acute kidney injury (AKI),severe AKI which renal replacement treatment was needed(RRT-AKI) and death after cardiac surgery in Chinese patients.Methods:1067patients who underwent cardiac surgery in the department of cardiac surgery in the Zhongshan hospital, Fudan university between May,2010and January,2011were involved in this research. Evaluating the predicting value for AKI (AKICS),RRT-AKI (Cleveland score,SRI and Mehta score) and death (EUROscore) after cardiac surgery by Hosmer-Lemeshow goodness-of-fit test for the calibration and area under receiver operation characteristic curve(AUROC) for the discrimination.Results:The incidence of AKI was20.33%(217/1067),and63.13%of their renal function recovered completely. The incidence of RRT-AKI was3.56%(38/1067) and the mortality of AKI and RRT-AKI was9.77%(21/217)and44.73%(17/38) respectively. The total mortality was3.28%(35/1067).The discrimination and calibration (AUROC=0.687; x2=14.66, p<0.05) for the prediction of AKI of AKICS were low. For the prediction of RRT-AKI,the discrimination and calibration (AUROC=0.736; x2=4.02, p>0.05) of Clevel and score were high enough, but the predicated value was lower than the real value(1.7%vs3.86%p<0.05,). The discrimination of Mehta score(AUROC=0.687) and the calibration of SRI (x2=6.59, p<0.05)were low. The discrimination and calibration (AUROC=0.685; x2=8.42, p>0.05) for the prediction of death of EURO score was low. Conclusions:According to the2012KDIGO AKI definition, none of the5models above is good at predicting AKI after cardiac surgey. Cleveland score has been validated to have a proper impact on predicting RRT-AKI after cardiac surgery but the predicting value is still in doubt. EURO score has been validated to have an inaccurate predicting value for death after cardiac surgery. Part Two.A clinical risk model to predict acute kidney injury after cardiac surgery.Background:AKI is a major complication following cardiac surgery and is associated with increased postoperative morbidity and mortality. So it is essential to build up an efficient risk model which was based upon Chinese patients with the AKI definition and stage well recognized. The aim of this study was to identify the risk factors for AKI after cardiac surgery and build up the prediction risk model.Methods:1394patients who underwent cardiac surgery(CABG, OPCAB,Valve surgery and CABG with/without valve surgery)in the department of cardiac surgery in the Zhongshan hospital, Fudan university between May,2010and January,2011were involved in this research. Logistic regression was used to analyze the incidence and risk factors of AKICS among1067patients, and the other300patients were set for validation, by means of Hosmer-Lemeshow goodness-of-fit test for the calibration and receiver operation characteristic(ROC) curves with area under ROC curve(AUC) for the discrimination.Results:The incidence of AKI after cardiac surgery was20.08%(280/1394), among which the RRT rate was3.30%(46/1394) and the mortality was10.35%(29/280).The total mortality was2.72%(38/1394).According to the logistic regression:Male (OR=1.645),Valve with CABG surgery (OR=3.511),cerebrovascular disease (OR=5.534), preoperative radiography (OR=1.609),preoperative NYHA>2(OR=1.509), preoperative serum creatine>115μmol/l (OR=2.325),intra-operative plasma infusion>400ml(OR=2.466), low cardiac output syndrome(OR=3.602) and post-operative central vein pressure>14mmH2O (OR=2.156) were defined as independent risk factor for AKI after cardiac surgery.The AUROC for AKICS of the validation set was0.81,which meant a good discrimination, and Hosmer-Lemeshow goodness-of-fit test showed a proper calibration.Conclusions:The risk model was based upon chinese patients, with AKI being defined with the2012KDIGO AKI guideline and proper prediction value for AKICS was validated.