On-pump or off-pump? Impact of risk scores in coronary artery bypass surgery

Remain controversies about the use of cardiopulmonary bypass (CPB) in coronary artery bypass grafting (CABG). The aim of this study was to evaluate the impact of the 2000 Bernstein Parsonnet (2000BP) and EuroSCORE (EU) for choice CPB in CABG.

1.551 consecutive patients underwent CABG. CPB was used in 1.121 (72.3%) patients. The performance of 2000BP and EU was assessed by calibration, discrimination and correlation tests. For both risk scores, increasing the value of the score and presence of CPB were directly related to a higher risk of death (P<0.05). Therefore with these two variables was constructed a logistic regression model for each risk score, in order to determine in which value of score the presence of CPB increases significantly the risk of death.

The calibration, like the area under the ROC curve for the group with CPB [2000BP=0.80; EU=0.78] and without CPB [2000BP=0.81; EU=0.85] were appropriate. The Spearman correlation for groups with and without CPB was 0.66 (P<0.001) and 0.62 (P<0.001), respectively. Using the 2000BP, for a value>17.75 the presence of CPB increased the chance of death to 7.4 [CI 95% (4.4-12.3), P<0.0001]. With the EU, for a value >4.5 the presence of CPB increased the chance of death to 5.4 [CI 95% (3.3-9), P<0.0001].

In decision making, the 2000BP>17.75 or the EU>4.5 guide to identify patients who underwent CABG with CPB increases significantly the chance of death.