Michael E. Halkos1, Vinod H. Thourani1, Omar M. Lattouf1, Patrick Kilgo2, Faraz Kerendi1, Robert A. Guyton1, John D. Puskas1; 1Emory University School of Medicine, Atlanta, GA; 2Emory University Rollins School of Public Health, Atlanta, GA
Objective: Diabetes mellitus (DM) is associated with an increased risk of adverse outcomes following coronary artery bypass grafting (CABG). Hemoglobin A1c (HbA1c) is a blood measure of long-term glucose control. It is unknown whether adequacy of diabetic control, as measured by preoperative HbA1c, is a predictor of adverse outcomes after CABG.Methods: Of 3,555 consecutive patients who had primary, elective CABG at a single academic center from 4/1/2002-6/30/2006, 3,089 patients (86.9%) had preoperative HbA1c levels obtained and entered prospectively into a computerized database. All patients were treated with a uniform perioperative intravenous insulin protocol (target blood glucose level below 120mg/dl). A multivariable logistic regression model (adjusted for 16 known risk factors) was used to determine whether HbA1c, as a continuous variable, was associated with in-hospital mortality, stroke, myocardial infarction (MI) and deep sternal wound infection (DSWI) after CABG. Receiver operator characteristic (ROC) curve analysis identified the HbA1c cutoff value that maximally discriminated outcome dichotomies.
Results: In-hospital mortality for all patients was 1.0% (31/3,089). HbA1c level (odds ratio (OR) 1.47 per unit increase, p=.0027) and preoperative diagnosis of DM (OR 3.61, p=.0126) predicted mortality after CABG. ROC curve analysis revealed that a HbA1c above 7.7% was associated with increased mortality, with a 47% increase in mortality for each unit increase of HbA1c above 7.7%. Higher HbA1c, but not preoperative diagnosis of DM, was associated with postoperative MI (p=.0134) and DSWI (p=.0045) after CABG with ROC value cutoffs of 9.8% and 7.7%, respectively. Neither HbA1c nor preoperative diagnosis of DM was predictive of perioperative renal failure or stroke.
Conclusion: Poor preoperative glycemic control, as measured by an elevated HbA1c, is more strongly associated with adverse events following CABG than is a preoperative diagnosis of DM. These differences were observed despite an aggressive perioperative insulin infusion protocol. Preoperative HbA1c testing may allow for better risk stratification in patients with a known history of DM, and in diabetic patients with no prior diagnosis of the disease.
Back to Annual Meeting
Back to Program Outline
