RESUMEN
Objective Background Predicting risk factors for valve replacements is important both for informed consent of patients and objective review of surgical outcomes. Development of reliable prediction rules requires large data sets with ap-propriate risk factors that are available before surgery. Methods Data were from Belling Anzhen Institute of heart, pulmonary and vascular diseases in the period of January 1993 to December 2004. 4482 heart valve replacement patients were analyzed.There. were 848 aortic valve replacements, 2202 mitral valve replacements and 1387 double valve replacements. Logistic regres-sion was used to examine the relationship between risk factors and in-hospital mortality. Results In the multivariable analysis,5 variables in the aortic model (older age, body area, NYHA class IV, creatin, CPB time) , 8 variables in the mitral model ( NYHA class Ⅳ, congestive heart failure, cardiac/thoracic ratio, FS, etiology, LVESD, CPB time, use of IABP) and 7 var-iables in the double valve model (older age, NYHA class Ⅳ, previous myocarditis, diabetes, CPB time, weight index, previ-ous percutaneous mitral balloon valvotomy ) remained independent predictors of the outcome. The mathematical models were highly significant predictors of the in-hospital mortality, and the results were in general agreement with those of others. The area uoder the receiver operating characteristic curve for the aortic model was 0. 921 [ 95% confidence interval ( CI ), 0. 874 to 0. 967 ], for the mitral model was 0. 859 ( 95% CI, 0. 813 to 0. 905 ) aod for dnuhle model was 0. 868 ( 95% CI, 0. 827 to 0.908). The goodness-of-fit statistic for the aortic model was χ~2 = 1.463, P=0.993, for the mitral model was χ~2 = 8.720,P = 0. 366 and for the double valve model was χ~2 = 8 . 134, P = 0. 420. Conclusion We print results and methods for use in day-to-day practice to calculate patient-specific in-hospital mortality after aortic and mitral valve surgery, by the logistic e-quation for each model or a simple scoring system with a look-up table for mortality rate.