Stratification of Morbidity and Mortality Outcome by Preoperative Risk Factors in Coronary Artery Bypass Patients

TLDR

The study aimed to relate perioperative morbidity and mortality risk to preoperative severity of illness in coronary artery bypass graft patients through a retrospective analysis of 5,051 cases using univariate and logistic regression. Researchers performed logistic regression on 5,051 patients to identify risk factors, then prospectively validated the resulting equations and an additive 1–6 point score in a 4,069‑patient cohort from the Cleveland Clinic Foundation. The resulting models accurately predicted mortality and morbidity, with higher scores indicating greater morbidity, and outperformed existing methods, making the clinical scoring system useful for preoperative risk estimation.

Abstract

To relate morbidity and mortality risk to preoperative severity of illness in patients undergoing coronary artery bypass grafting.Retrospective analysis of 5051 patients using univariate and logistic regression to identify risk factors associated with perioperative morbidity and mortality. Prospective application of models to a subsequent 2-year validation cohort (n = 4069).Cleveland Clinic Foundation.All adult patients undergoing coronary artery bypass graft surgery between July 1, 1986, and June 30, 1988 (reference group), and July 1, 1988, and June 30, 1990 (validation group).Mortality and morbidity (myocardial infarction and use of intra-aortic balloon pump, mechanical ventilation for 3 or more days, neurological deficit, oliguric or anuric renal failure, or serious infection).Emergency procedure, preoperative serum creatinine levels of greater than 168 mumol/L, severe left ventricular dysfunction, preoperative hematocrit of 0.34, increasing age, chronic pulmonary disease, prior vascular surgery, reoperation, and mitral valve insufficiency were found to be predictive of mortality. In addition to these factors, diabetes mellitus, body weight of 65 kg or less [corrected], aortic stenosis, and cerebrovascular disease were predictive of morbidity. Logistic regression equations were developed, and a simple additive score for clinical use was designed by allocating each of these risk-factor values of 1 to 6 points. Both methods predict mortality. Increased morbidity was demonstrated with increases in score.The logistic or clinical models developed are superior to the currently available methods for comparing mortality outcome and provide previously unavailable information on morbidity based on preoperative status. The clinical scoring system is useful for preoperative estimates of morbidity and mortality risks.