A model to predict survival at time of postcardiotomy intraaortic balloon pump insertion

To facilitate timely application of new forms of cardiac support to patients at highest risk after cardiotomy despite conventional support with the intraaortic balloon pump, an accurate prediction of survival must be available at the time of weaning from cardiopulmonary bypass. We, therefore, acquired 240 demographic, disease, and perioperative characteristics of 322 patients (mortality rate, 48.4%) who required IABP support to separate from bypass. Four variables available before or within 10 minutes of the first attempt at weaning from bypass significantly predicted mortality by stepwise logistic regression: complete heart block as demonstrated by need for temporary pacing at weaning (p < 0.001), advanced age (p < 0.002), preoperative blood urea nitrogen concentration (p = 0.036), and female sex (p = 0.048). An equation generated by the logistic model predicted a 72.2% survival rate in the 25% of patients at least risk (actual survival rate, 71.6%); in the 25% at greatest risk, death was predicted in 73.0%, and the actual mortality rate was 74.1%. The equation was then prospectively applied to 330 intraaortic balloon pump-supported patients managed at another institution. The overall mortality rate there was 41.2%; in the 25% at least risk, predicted survival rate was 70.5% (actual survival rate, 77.1%), and in the 25% at greatest risk, predicted mortality rate was 75.7% (actual mortality rate, 62.7%). Thus, retrospectively at one institution and prospectively at another, the equation generated by this model based only on data available at the time of weaning from bypass was able to define one subgroup of patients 2.6 to 2.7 times as likely to die as another subgroup from within similar cohorts.(ABSTRACT TRUNCATED AT 250 WORDS)