A score to estimate 30-day mortality after intensive care admission after cardiac surgery.

OBJECTIVE: Several risk-scoring systems have been developed to predict surgical mortality and complications in cardiac surgical patients, but none of the current systems include factors related to the intraoperative period. The purpose of this study was to develop a score that incorporates both preoperative and intraoperative factors so that it could be used for patients admitted to a cardiac surgical intensive care unit (ICU) immediately after surgery.

METHOD: Preoperative and intraoperative data from 30,350 patients in four hospitals were used to build a multiple logistic regression model estimating 30-day mortality after cardiac surgery. Sixty percent of the patients were used as a derivation group and forty percent as a validation group.

RESULTS: Mortality occurred in 2.6% of patients (n = 790). Preoperative factors identified in the model were age, female sex, emergency status, pulmonary hypertension, peripheral vascular disease, renal dysfunction, diabetes, peptic ulcer disease, history of alcohol abuse, and refusal of blood products. Intraoperative risk factors included the need for an intra-aortic balloon pump, ventricular assist device or extracorporeal membrane oxygenation leaving the operating room, presence of any intraoperative complication reported by the surgeon, the use of inotropes, high-dose vasopressors, red blood cell transfusion, and cardiopulmonary bypass time. When used after surgery at ICU admission, the model had C-statistics of 0.86 in both derivation and validation sets to estimate the 30-day mortality.

CONCLUSIONS: Preoperative and intraoperative variables can be used on admission to a cardiac surgical ICU to estimate 30-day mortality. The score could be used for risk stratification after cardiac surgery and evaluation of performance of cardiac surgical ICUs.