Performance Of SinoSCORE In Predicting In-hospital Mortality And Postoperative Complications After Mitral Valve Replacement In Ningxia

Objective:To study the SinoSCORE evaluative value on the mortality risk ofmitral valve replacement patients and postoperative complications.Methods:1. A total of505cases of mitral valve replacement（MVR） fromDecember2003to December2013were chosen for retrospective analyze;2.UsingSinoSCORE scoring model to predict postoperative hospital mortality in patientswith mitral valve replacement and major complications (low cardiac outputsyndrome, renal failure, pulmonary infection and multiple organ failure) toevaluate its predictive ability. using the receiver-operating Characteristic (ROC)test the discrimination of the modle. Determine the optimum diagnostic cut-offvalue.Results: The mean age of the505enrolled patients was (49.25±10.15)years． The area under ROC was0.78(95%CI:0.678—0.882)， indicating gooddiscrimination power of the model in predicting in-hospital mortality．The overallin-hospital mortality was4.75%(24/505). Hosmer-Lemeshow goodness of fit test P=0.744(X2=4.304),，suggesting a high calibration ability of the model．The areasunder ROC of low cardiac output syndrome, renal failure and multiple organfailure were0.712,0.701and0.802，respectively，suggesting that SinoSCORE hada satisfactory performance in predicting post-operative low cardiac outputsyndrome, renal failure, pulmonary infection and multiple organ failure． And the optimal cut-off points for the above four complications predicted by SinoSCOREwere6.0,4.5and6.5.Conclusions: SinoSCORE has a better performance in predicting thein-hospital mortality risk in Chinese patients undergoing valve surgery，and it canbetter predict low cardiac output syndrome, renal failure and multiple organfailure.