Diabetic patients have a higher risk of undergoing any type of surgical intervention compared to the non-diabetic population. Additionally, after the procedure, they also present higher perioperative morbidity and mortality^[1^]. Currently, there is a progressive increase in the prevalence of diabetes in patients referred for cardiovascular surgery. Besides higher perioperative morbidity and mortality, these patients present lower angina-free survival and a significant decrease in long-term survival^[2^,3^]. In this scenario, evidence shows that appropriate glycemic management can improve these outcomes.

Insulin deficiency and insulin resistance are known to be aggravated by surgery and anesthesia and may lead to lipolysis and ketogenesis, which may result in metabolic acidosis and subsequent electrolyte changes. Protein catabolism increases due to increased degradation and decreased synthesis. Insulin administration may reverse most of these disorders, so, in type 2 diabetic patients, long-acting sulfonylureas, like chlorpropamide, should be discontinued and replaced with agents of short duration. Metformin should always be discontinued. Type 2 diabetic patients with marked hyperglycemia and on oral treatment should receive insulin prior to surgery. Insulin requirements may range from 0.25-0.40 U per gram of glucose in diabetic patients with normal weight undergoing surgery to 0.4-0.8 U per gram of glucose in diabetic patients with obesity, liver disease, on treatment with steroids, or sepsis, and 0.8-1.2 U per gram of glucose in diabetic patients undergoing surgery with use of extracorporeal circulation (ECC). Therefore, the appropriate dose should be personalized, knowing that the protocol preferred by most authors is based on the variable rate of insulin infusion.

Next, we will present recommendations based on international evidence and adapted to our clinical practice for the perioperative management of hyperglycemia in adult patients with and without diabetes undergoing cardiovascular surgery at the Instituto do Coração da Faculdade de Medicina da Universidade de São Paulo.

The limitation in glycemic management is associated with increased morbidity and mortality.

McAlister et al.^[4^] found, in a retrospective analysis of 291 patients undergoing coronary artery bypass grafting (CABG), that the mean blood glucose level on the first postoperative day adequately predicted the development of an adverse outcome. Harmful effects with elevated intraoperative blood glucose levels were also reported by Gandhi et al.^[5^] in a retrospective study of 409 patients undergoing cardiac surgery. In this study, pre- and intraoperative hyperglycemias were independent risk factors for perioperative complications, including death.

A retrospective study of 6280 patients undergoing cardiac procedures found that high blood glucose levels were a predictor of mortality in patients with and without diabetes^[6^]. Fish et al.^[7^] showed that glycemic levels above 250 mg/dL increased 10 times the number of complications in patients undergoing CABG.

Finally, Anderson et al.^[8^] showed that, in 1375 patients undergoing CABG, preoperative hyperglycemia increased by twice the mortality rate in one year compared to patients with normal preoperative blood glucose.

Together, these studies strongly suggest that hyperglycemia during the perioperative period of cardiovascular surgery is a predictor of morbidity and mortality, regardless of whether or not the patient is diabetic. Next, we will present recommendations supported by evidence of how the reduction of blood glucose levels decreases the incidence of negative outcomes in this group of patients.

Individualization is the key point in elderly population, per the Brazilian Society of Diabetes (SBD), the American Diabetes Association (ADA), and the European Association for the Study of Diabetes (EASD), the fasting glycemic value is up to 150 mg/dL, the postprandial glycemic value is < 180 mg/dL, and the less stringent hemoglobin A1c (HbA1c) goal is < 8%. According to the critical analysis of four major studies - United Kingdom Prospective Diabetes Study (UKPDS), Action in Diabetes and Vascular Disease: Preterax and Diamicron MR Controlled Evaluation (ADVANCE), Action to Control Cardiovascular Risk in Diabetes (ACCORD), and Veterans Affair Diabetes Trial (VADT) - the attempt to rigorously control the glycemia in elderly patients, especially in those with known atherosclerotic diseases, in addition to not preventing cardiovascular events, may increase mortality (ACCORD), possibly but not necessarily due to hypoglycemia. Therefore, the SBD recommends that an overall assessment should be made of each case in order to flexibilize or consolidate the therapeutic targets, following the same principles described^[9^].

Benefits of glycemic management (< 180 mg/dL) in diabetic patients during cardiac surgery:

It reduces mortality;

The impact of glycemic management during heart surgery was reported by Furnary et al.^[10^] analyzing 3554 patients undergoing CABG between 1987 and 2001. Patients were divided into three groups: from 1987 to 1991, they received insulin subcutaneously once every 4 hours to maintain blood glucose level at 200 mg/dL; from 1991 to 1998, continuous intravenous insulin infusion was used to maintain blood glucose levels between 150 and 200 mg/dL; and from 1998 to 2001, the Portland protocol was instituted, administering continuous insulin drip to maintain blood glucose levels between 100 and 150 mg/dL. Continuous infusions of insulin resulted in lower mean glucose levels than that obtained with intermittent subcutaneous insulin therapy. From 1992 and after the establishment of continuous insulin protocols, perioperative mortality in diabetic patients undergoing CABG was reduced by 50% (4.5% vs. 1.9%, P=0.0001), which was similar to that of non-diabetic patients undergoing CABG. There was also a significant decrease in the incidence of surgical wound infections (P=0.001). Furnary et al.^[11^] expanded its original series to include an additional population of 1980 patients treated with the Portland protocol between 2001 and 2005. They introduced a new method for assessing glycemic management, called "Glucose 3", which consisted of the mean glucose values obtained in surgery and on the first and second postoperative days. An increase in "Glucose 3" was an independent predictor of perioperative mortality (P<0.001). On the other hand, this value was also related to an increase in the incidence of deep infections of the sternum, time of hospitalization, blood transfusions, atrial fibrillation, and low cardiac output syndrome.

The importance of the use of insulin in diabetic patients undergoing CABG was presented by Lazar et al.^[12^] and consists of a modified glucose, insulin, and potassium solution. In this study, 141 diabetic patients undergoing CABG were randomized to receive glucose, insulin, and potassium to maintain a blood glucose level between 120 and 180 mg/dL or a mobile insulin scale to maintain glucose level < 250 mg/dL. The combination of glucose, insulin, and potassium was initiated in the induction of anesthesia and continued for 12 hours in the intensive care unit (ICU). This protocol obtained better glycemic management immediately before cardiopulmonary bypass (169 mg/dL vs. 209 mg/dL, P<0.0001) and after 12 hours in the ICU (134 mg/dL vs. 266 mg/dL, P<0.0001). Patients treated with strict glycemic control had higher cardiac indices (P<0.0001) and less need for vasoactive drugs (P<0.05) and cardiac stimulation (P<0.05). Strict glycemic control resulted in a lower incidence of infections (0 vs. 13%, P=0.01) and atrial fibrillation (15% vs. 60%, P=0.007). All this contributed to the reduction of hospitalization time (6.5 days vs. 9.2 days, P=0.0003). After five years, Kaplan-Meier curves showed survival advantage (P=0.04) for patients who received better glycemic control. They had a lower incidence of recurrent angina (P=0.01), wound infection (P=0.03), and were able to maintain a lower class of angina (P=0.03).

Evidence on the importance of rigorous glycemic management in patients undergoing CABG has also been demonstrated in a study by Van den Berghe et al.^[13^], involving 1548 patients on mechanical ventilation admitted to a surgical ICU. In this prospective, randomized study, 62% of the patients underwent cardiac surgery and only 13% had a history of diabetes. Patients were randomized into a conventional therapeutic group, in which insulin was administered only if the blood glucose level was > 215 mg/dL to maintain a target of 180-200 mg/dL, and into another group that received continuous infusion of insulin to maintain glucose levels between 80 and 110 mg/dL. Intensive insulin therapy resulted in a significant reduction in mortality (10% vs. 20%, P=0.005). Cardiac surgery mortality was only reduced in those patients who required three days of ICU care. Hospital mortality for all cardiac surgery patients, regardless of ICU stay, was reduced from 5.1% to 2.1% (P<0.05). Intensive glycemic management had no effect on the morbidity and mortality of patients who stayed more than three days in the ICU. In another study to identify patients who could benefit more from strict glycemic management, D'Alessandro et al.^[14^] sought to correlate the effect of strict glycemic management with the results expected by EuroSCORE in diabetic patients undergoing CABG. Using the propensity score, 300 diabetic patients undergoing CABG who received strict glycemic management (150 to 200 mg/dL in the operating room, 140 mg/dL in the ICU) and 300 diabetic patients undergoing CABG when the insulin protocols did not yet exist were compared. The rigid glycemic management group had a significantly lower mortality than that expected by EuroSCORE (1.3% vs. 4.3%, P=0.01). Mortality was low, mainly in the highest risk group (EuroSCORE 4; 2.5% vs. 8.0%, P=0.03). In contrast, there was no difference between observed and expected mortalities in the group without strict glycemic management in patients with EuroSCORE < 4. Two other additional studies showed the importance of glycemic management in reducing sternal infection. Zerr et al.^[15^] studied the effects of glycemic management on the incidence of sternal infection in 1585 diabetic patients undergoing CABG. Sternal infection increased by 1.3% in patients with mean glucose values of 100 to 150 mg/dL to 6.7% in patients with levels of 250 to 300 mg/dL. In a retrospective study of diabetic patients undergoing CABG, Hruska et al.^[16^] observed that continuous infusion of insulin maintaining blood glucose levels between 120 and 160 mg/dL significantly reduced the incidence of sternal infection compared with intermittent subcutaneous injections.

Intraoperative glycemic management with continuous infusion of intravenous insulin is not necessary in non-diabetic patients undergoing cardiac surgery, provided that blood glucose levels remain < 180 mg/dL.

Is rigorous glycemic control necessary for all patients undergoing cardiovascular surgery?

Butterworth et al.^[17^] showed the effects of strict blood glucose control in 381 non-diabetic patients undergoing isolated CABG. In this prospective, randomized study, one group received continuous infusion of insulin when intraoperative blood glucose levels were > 100 mg/dL. The other group did not receive insulin coverage. The primary outcome was the incidence of new neurological events, neuro-ophthalmologic and neurobehavioral deficits, and neurological-related deaths. Intraoperative blood glucose levels were significantly lower in patients receiving insulin infusion, although there was no difference in the incidence of neurological complications. Likewise, there were no differences between the groups regarding operative mortality, need for vasoactive drugs, and time of hospitalization; however, patients who did not receive intraoperative insulin had blood glucose levels at 200 mg/dL. In this study, intraoperative glycemic management did not improve short- and long-term clinical outcomes in the non-diabetic patients' group.

Gandhi et al.^[18^] studied the effects of intensive intraoperative insulin therapy in 400 patients undergoing elective CABG. Patients were randomized prospectively into a group with continuous insulin, to maintain blood glucose levels between 80 and 100 mg/dL, and into a conventional group, to maintain blood glucose levels up to 200 mg/dL using intermittent boluses of intravenous insulin. The incidence of diabetes was 20% in both groups. There was no difference between the groups in the composite outcome: death, sternal infection, prolonged ventilation, cardiac arrhythmias, stroke, and renal failure up to 30 days after surgery. There was also no difference in ICU time and hospital admission between the groups. And there was a tendency for more deaths (P=0.06) and stroke (P=0.02) in the intensive care group with insulin. The limitations of this study were the inclusion of patients with and without diabetes and that both groups received intensive insulin therapy in the immediate postoperative period.

The FREEDOM study was designed to evaluate the benefits of current angioplasty techniques (which uses drug-eluting stents) and of CABG in combination with aggressive drug therapy in diabetic patients^[26^]. The results of overall mortality, acute myocardial infarction, and stroke were significantly lower in the surgery group. This study confirmed that CABG is the preferred revascularization strategy in diabetic patients.

Important issues regarding the management of hyperglycemia during cardiovascular surgery should still be elucidated. In this scenario, future research will determine: (1) the ideal level of glycemic management and, if appropriate, the most important perioperative period to maintain this control; (2) whether the blood glucose level reached is as important as the amount of insulin delivered; (3) the importance of preoperative levels of HbA1c and whether surgery should be postponed in patients with high HbA1c levels; (4) if the handling of elderly patients should be protocolized or individualized, could them tolerate higher levels of HbA1c without risk morbidities or their increased risk of hypoglycemia is so higher than of the general population that a more permissive HbA1c would significantly reduce their morbidity and mortality? Answers to these questions will increase our understanding of hyperglycemia during cardiovascular surgery and will help us determine more appropriate methods to achieve glycemic control and improve clinical outcomes in this group of high-risk patients.