Heart valve replacement surgery mortality risk model and the use of extracorporeal membrane oxygenation as a replacement to cardiopulmonary bypass

Abstract

Objective To form a scoring model to predict the risk of in-hospital mortality of patients undergoing heart valve replacement. Subsequently, we attempt to assess whether the mortality of patients undergoing heart valve procedure can be decreased by performing earlier ECMO support named replaced extracorporeal circulation of ECMO (REC-ECMO) once patients condition predicted severe by this model preoperatively. Methods We retrospectively reviewed the medical records of 4482 critically ill patients in the database of Anzhen Hospital in Beijing, who underwent cardiac valve replacement from January 1994 to December 2004. Several demographic, clinical, and laboratory variables were retrospectively gathered as predicators of death. A scoring model predicting the risk of postcardiotomy mortality were established, in which variables that were statistically significant (P < 0.05) in the univariate analysis were included in multivariate analysis by applying a multiple logistic regression. The predicted mortality of patients who were going to receive heart valve replacement was calculated using this model. Those whose predicted mortality were more than 10% received valve procedure with normal extracorporeal circulation or REC-ECMO respectively. The difference of the observed mortality were tested by applying Pearson Chi-Square (α = 0.05). Results Six out of seven selected variables that were statistically significant in the univariate analysis were included in multivariate analysis. The risk of mortality according to scoring model was: Ln(P/1-P)=-4.3742+0.5192 renal function + 0.1467 age+ 0.3947 EF+0.9390 CAD +0.5888 PAP+ 0.3287 LVEDD. The predicted mortality of This two groups had no significant difference[(16.35± 4.3) % vs( 16.21 ± 5.19) %, P < 0.05]. While the observed mortality differed significantly(Pearson Chi-Square, P =0.047, P <0.05) and was 30.76%(16/52)and 9.09%(2/22) respectively. Conclusion A risk of postoperative valve replacement mortality can be predicted before the procedure is performed by using our scoring model. Patients with predicted mortality more than 10% by the scoring model gain lower observed mortality if REC-ECMO rather than normal extracorporeal circulation is used. Key words: Extracorporeal membrane oxygenation; Heart valve diseases; Heart valve prosthesis implantation; Proportional hazards models