Preoperative fasting blood glucose levels and increased mortality after CABG surgery

Cardiology Review® Online, Decmber 2005, Volume 22, Issue 12

In their study, Anderson and coworkers (page 10) report that patients with impaired fasting glucose levels have an increased risk of death following coronary artery bypass graft (CABG) surgery that is similar to patients with suspected and confirmed diabetes mellitus. Their results are in keeping with a growing body of literature that suggests that hyperglycemia during coronary revascularization procedures results in increased short- and long-term mortality. While I agree with this concept, there are several major limitations in both the design and methodology of this study that limit its impact on clinical practice.

The study was retrospective, and the patient group with abnormal fasting glucose levels was small, resulting in widened confidence intervals. We are not told when fasting glucose levels were drawn (for example, on the day of or several days before surgery). It is unknown whether protocols were in place to maintain tight glycemic control during the perioperative period and whether or not patients with impaired fasting glucose and those with suspected diabetes received any antihyperglycemic therapy. This could have easily affected the outcome.

I am surprised at the cumulative 1-year mortality in the three non-normal groups, which ranged from 5.7% to 7.4%. The authors fail to tell us what percentage of these patients received a left internal mammary artery to left anterior descending graft (which has been shown to enhance survival in patients with diabetes), and the incidence and type of perioperative complications in each group. More importantly, we are not told what the causes of deaths were among the groups and whether they were related to cardiovascular etiologies (myocardial infarction, stroke, congestive heart failure, arrhythmias) or noncardiovascular entities. For example, it is well known that the incidence of sternal wound infections is increased in diabetic patients with elevated perioperative glucose levels. Did a significant number of the patients with impaired fasting glucose have wound infections that contributed to their mortality? Finally, the authors noted that the 30-day and 1-year mortality rates were similarly increased among the three non-normal groups. This suggests that some underlying disease process was not effectively treated. The authors conclude their discussion by stating that the mechanism for their findings remains unclear, and they question whether normalizing glucose levels perioperatively would have altered the results. In this commentary, data will be provided that may help to answer these questions.

Diabetic patients, and those with abnormal fasting glucose levels, develop insulin resistance during CABG surgery that contributes to increased levels of free fatty acids and decreased myocardial uptake of glucose.1 The resulting hyperglycemia contributes to oxidative stress and augmentation of the inflammatory response, which leads to endothelial dysfunction. Diabetic patients with coronary artery disease have increased levels of proinflammatory cytokines, such as interleukin-6 and tumor necrosis factor, which stimulate the synthesis of acute-phase proteins, such as C-reactive protein.2 Hyperglycemia impairs endothelial function by depleting NAD (P)H, which ultimately decreases endothelial nitric oxide (NO) synthase and reduces NO release, resulting in vasoconstriction.3 Bioassays from internal mammary arteries and from saphenous vein grafts of diabetic patients show decreased NO activity and increased production of superoxide and NAD (P)H oxidase.4 Hyperglycemia also results in increased levels of plasminogen activator inhibitor-1 and adhesion molecules, which impairs platelet function, resulting in increased adhesiveness and hyperaggregability.5 These changes in endothelial function, the inflammatory response, and platelet function contribute to arterial thrombosis and ultimately the patency of bypass grafts, as well as native coronary vessels.

Insulin acts to improve myocardial metabolism, reduce the inflammatory response, preserve endothelial integrity, and maintain platelet function during periods of myocardial ischemia. It decreases levels of free fatty acids and increases the myocardial uptake of glucose; decreases the levels of reactive oxygen species, adhesion molecules, and C-reactive protein; increases the bioavailability of NO; and decreases serum levels of plasminogen activator inhibitor-1.1,5,6

In summary, the mechanism for the favorable actions of insulin during CABG surgery in diabetic patients may result from its ability to limit injury to vein grafts and native coronary vessels owing to its preservation of endothelial function. This results in enhanced recovery of ischemic myocardium and better vein graft patency, which contributes to superior long-term survival.

Is there any evidence to show that normalizing glucose levels with insulin results in superior short- and long-term outcomes in diabetic CABG surgery patients? We designed a study to answer this question in 141 diabetic CABG surgery patients prospectively randomized to tight glycemic control (125—200 mg/dL) with continuous intravenous insulin therapy, or moderate control (< 250 mg/dL) with intermittent subcutaneous insulin therapy.7 Following CABG surgery, patients with tight glycemic control had lower serum glucose levels (138 ± 4 SE versus 260 ± 6 mg/dL; P < .001), decreased serum lactate (1.75 ± 0.14 versus 2.33 ± 0.31 mmol/L; P = .04), and less serum free fatty acids (0.33 ± 0.03 versus 0.57 ± 0.05 mEq/L; P < .001). As a result, patients with tight glycemic control required less pacing and inotropic support, spent less time on the ventilator (6.9 ± 0.3 versus 10.7 ± 0.6 hours; P = .002), and had a lower incidence of atrial fibrillation (16.6% versus 42%; P = .0017) and infectious complications (0% versus 13%; P = .01), all of which contributed to a shorter postoperative length of stay (6.5 ± 0.1 versus 9.2 ± 0.3 days; P = .003). Patients with tight glycemic control also showed a significant survival advantage 3 years after surgery (98% versus 90%; P = .04), had decreased episodes of ischemia (5% versus 19%; P = .01), and maintained a lower angina class.

What are the clinical implications from Anderson’s study and our own data? Tight glycemic control in diabetic CABG surgery patients improves both short- and long-term outcomes. All patients exhibiting abnormal fasting glucose levels or elevated perioperative serum glucose values should receive continuous intravenous insulin infusions to maintain serum glucose at less than 200 mg/dL for 12 to 18 hours post-CABG. It is no longer a question of whether treatment for elevated perioperative glucose values is necessary, but whether superior results can be obtained if glucose levels are lowered to 100 mg/dL or lower as advocated by van den Berghe and coworkers in a study involving critically ill patients receiving mechanical ventilation.8 This will be the subject of our future studies, as we attempt to determine the optimal method for glycemic control in the postoperative period.