The Edinburgh Cardiac Surgery Score survival prediction in the long-stay ICU cardiac surgical patient

Abstract

Predictors of outcome in long-stay patients following cardiac surgery have hitherto been ill defined. The aims of this study were to test the Parsonnet pre-operative scoring system and to define a scoring system for inhospital mortality applicable post-operatively to strengthen the clinical decision-making process. Following case note review of 262 consecutive patients who stayed 7 days or more in intensive care, a total of 110 pre-, intra- and post-operative factors were documented. In this long-stay group the Parsonnet score was confirmed to be predictive of 30 day mortality. Univariate analysis identified significant association between mortality in the Intensive Care Unit (ICU) and the following: inotrope days, (defined as number of inotropes times number of days) ventilation, units of platelets (P = < 0.00001), chest reopening, fresh frozen plasma units (P < 0.002), total parenteral nutrition, noradrenaline, Parsonnet score (P = 0.005), dopamine, bypass time, vasodilators, intra-aortic balloon counterpulsation, enteral nutrition and other major surgery (P < 0.05). Stepwise logistic regression on these significant factors was used to produce the Edinburgh Cardiac Surgery Score (ECS) applicable from Day 10 onwards in the intensive care unit: ECS Score = (Inotrope days) +2 (Ventilation) + (Platelets) + (Parsonnet) -3. The ECS score may be a useful predictor of ICU mortality probability for cardiac surgical patients requiring 10 days or more intensive care and is presently undergoing prospective evaluation in our centre.