The effect of diabetic medications on creatine kinase-myocardial band levels in patients undergoing coronary artery bypass surgery☆

Abstract

Ischemic preconditioning has been shown to attenuate the rise in creatine kinase-myocardial band levels that occur with coronary artery bypass surgery (CABG). Recently, concerns have been raised that some sulfonylureas particularly glibenclamide may block ischemic preconditioning. The purpose of this study was to determine the effect of various diabetic medicines on creatine kinase-myocardial band levels after CABG. In this retrospective study of 799 patients undergoing CABG, patients continued their routine diabetic medicines up to the day of surgery. Intra-operatively and postoperatively, tight glycemic control was maintained with an insulin infusion. Anesthesia was maintained with isoflurane supplemented by fentanyl. Creatine kinase-myocardial band levels were determined the day after surgery at 05:00 h and the mean levels compared between diabetics and non-diabetics and further compared by type of diabetic medicine. After univariable comparisons, linear regression was used to determine the statistically significant predictors of creatine kinase-myocardial band levels. After correction for other factors, none of the diabetic medicines was a statistically significant predictor of creatine kinase-myocardial band levels. We found that the use of glibenclamide or other diabetic medications had no effect on creatine kinase-myocardial band levels the morning after patients underwent CABG.

1. Introduction

Ischemic preconditioning, where brief ischemia promotes the release of protective substances that limit myocyte death, may be cardioprotective in patients undergoing coronary artery bypass surgery (CABG) [1,2]. These protective substances, such as adenosine and bradykinin, activate signal transduction cascades, such as phosphatidylinositol-3-kinase-Akt and extracellular signal regulated kinase 1 and 2, which lead to opening of the mitochondrial K_ATP channel, release of reactive oxygen species, further signaling kinases, which mediate the transcriptions of further mediators and effectors, such as inducible nitric oxide synthase, manganese superoxide dismutase, heat shock proteins, and cyclooxygenase 2. All these factors, and probably other currently undescribed factors, then converge on the mitochondria permeability transition pore (PTP) to limit its opening and preserve the myocyte [3]. Studies showed that blockage of a key component of this pathway K_ATP channels prevented ischemic preconditioning in vitro and that the sulfonylurea glibenclamide (or glyburide) blocked K_ATP channels and prevented ischemic preconditioning during coronary angioplasty [4].

While hyperglycemia and diabetes mellitus (DM) are risk factors for coronary artery disease and death, the choice of which medicines to use to control glucose levels and the level of glycemic control to achieve are controversial and their use has not consistently been shown to reduce mortality. Furthermore, DM itself may be protective or may interfere with protection. Study in diabetic rats showed that compared to non-diabetic rats or rats with mild hyperglycemia, rats with severe hyperglycemia had resistance to ischemia [5]. Conversely, Kersten et al. found that hyperglycemia and insulin deficiency abolished ischemic preconditioning [6].

Release of cardiac enzymes, such as CK-MB, is considered indicative of myocyte death and even small increases after cardiac surgery have been associated with increased long-term risk of death [7]. Studies have shown that ischemic preconditioning before CABG can occur by at least two mechanisms: preoperative angina and some anesthetic agents. In a small study, Vahlhaus et al. found that patients who suffered preoperative angina had lower CK-MB values after surgery (18.7±1.3 vs. 13.8±1.5 U/l, mean±S.E.M., P<0.05), suggestive of an ischemic preconditioning benefit in CABG patients [8]. Anesthesthetic agents including isoflurane have also been shown to promote ischemic preconditioning and that this benefit may be lost in patients taking some sulfonylureas [9].

We hypothesized that sulfonylureas interfered with ischemic preconditioning and resulted in higher CK-MB levels than in non-diabetic patients and diabetic patients being treated with other medications. The purpose of this study was to determine the effect of diabetic medication on CK-MB release after CABG. A secondary purpose was to determine if hard outcomes of electrocardiographic myocardial infarction or 30-day mortality differed between groups.

2. Methods

This study was approved by the Institutional Review Board. As it was a retrospective analysis of routinely collected data, patient consent was waived. All patients who underwent CABG without concomitant valve, aortic, or other operation between 1 October 2006 and 30 September 2008, were included. Patients were excluded if they had a myocardial infarction within seven days before surgery. The departmental database consisting of demographic, preoperative, intra-operative, and postoperative factors was combined with the laboratory database for CK-MB levels (normal level: 0=5 ng/ml), drawn the morning after surgery, hemoglobin A1c levels, drawn in surgery at anesthesia induction, and mean glucose levels from arrival in OR to 05:00 h the next morning. Mean glucose was calculated by the trapezoidal method.

Patients took their usual diabetic medicines through the day prior to surgery. None was given the morning of surgery. All patients underwent a standardized anesthetic consisting of induction with diazepam, fentanyl, thiopental, and pancuronium. Vecuronium was used if the patient had renal failure. Anesthesia was maintained with isoflurane, which was started on intubation and maintained at ∼1 MAC throughout the case, and additional small doses of fentanyl. Starting at anesthesia induction, glucose levels were measured hourly and an intravenous insulin infusion was titrated according to our standard protocol to achieve and maintain glucose levels between 80 and 110 mg/dl.

Standard cardiopulmonary bypass with normothermic perfusion was used. Myocardial protection was achieved by antegrade blood cardioplegia solution. When clinically indicated, retrograde blood cardioplegia was also used. Myocardial cooling was accomplished by topical application of ice saline solution in the pericardium.

Patients who received each diabetic medicine were compared to those who did not. Student t-test was used to compare normally distributed continuous data. Data are presented as mean±S.D. Fisher's exact test and χ²-test were used to compare categorical data and proportions. To control for other factors that may affect CK-MB levels, linear regression was performed on each group. CK-MB level was the dependent variable with all the variables in Table 1 entered in forward stepwise models. P<0.05 and 95% confidence intervals that excluded one denoted statistical significance. Based on a power analysis with CK-MB=29±45 ng/ml, 799 patients is sufficient to detect a two-tailed difference in CK-MB=9 ng/ml with α=0.05, and power=0.8. SPSS 16 (SPSS, Inc, Chicago, IL) was used.

3. Results

Of the 799 patients, 570 (71%) were male with a mean age=66±11 years (Table 1). Two hundred and thirty-five (66%) of the 358 diabetic patients were receiving medications for their DM (Table 2 ). While diabetic patients had higher glycosylated hemoglobin A1c (HbA1c), 7.7%±1.9 vs. 5.9%±0.5, P<0.001, than non-diabetic patients, they achieved similar intra- and postoperative glucose control, 119±27 vs. 116±26, P=0.13. By univariable analysis, the preoperative use of glargine insulin, rosiglitazone, and glibeclamide was associated with lower postoperative levels of CK-MB. However, after correction for other factors, none of the diabetic medicines remained a statistically significant predictor of CK-MB (Table 3 ). We also found that the occurrence of myocardial infarction by electrocardiography and peri-operative death was low and did not differ by diabetic medications (Table 3).

4. Discussion

We found that, after multivariable correction, none of the diabetic medicines was associated with CK-MB levels that differed from levels in non-diabetic patients. There are few previous studies evaluating the effects of diabetic medications on ischemic preconditioning for patients undergoing CABG. Forlani et al. randomized 40 diabetic patients to continue taking glibenclamide for two days before and on the morning of surgery or to have glibenclamide stopped at least two days before surgery and be started on subcutaneous insulin then [1]. Anesthesia was randomized to either fentanyl-based or fentanyl-based supplemented by isoflurane for 15 min immediately prior to cardiopulmonary bypass. In the two groups who did not receive isoflurane, the group who received insulin and not glibenclamide had lower CK-MB levels (85±28 vs. 47±7, P<0.05) at 1 h and (65±42 vs. 39±20, P<0.05) at 24 h after surgery, but not at 48 or 96 h. In the two groups who received isoflurane, there was only a difference at 1 h (50±23 vs. 41±5, P<0.05) but not at 24, 48 or 96 h after surgery. In the group that received isoflurane, the switch to insulin produced lower troponin-I levels only at 24 h. That study showed that the use of glibenclamide instead of insulin produced more cardiac enzyme release, which could be mostly attenuated by administering isoflurane.

Our study differs in that we measured CK-MB levels at only one time – the morning after surgery. Measuring CK-MB more frequently and at different times may have detected differences. However, measuring CK-MB only on the morning after surgery has been shown to be a significant predictor of mortality [7]. Our studies further differ in that we used isoflurane continuously starting at induction compared to their 15 min prior to cardiopulmonary bypass. The effects of isoflurane may be dose dependent and 15 min may have provided a dose that produced only partial benefit. Also, our use of isoflurane post cardiopulmonary bypass may have provided postischemic protection [9]. All these differences may have contributed to our not finding a difference in CK-MB levels based on use or non-use of glibenclamide or any other diabetic medication.

Part of our findings of no effect of any sulfonylurea on CK-MB release may be related to our fortuitous choice of anesthetic agents. Studies in an open-chest rat model of myocardial ischemia, where the left coronary artery is occluded, found that morphine produced a similar reduction in infarct size as did three episodes of 5-min occlusions separated by 5-min reperfusions, comparing both to placebo. Interestingly, both glibenclamide and naloxone, a non-selective opioid receptor antagonist, completely blocked the effects of morphine [10]. Further research using β-funaltrexamine, a specific μ receptor antagonist; D-Ala, 2N-Me-Phe, 4glycerol5-enkephalin, a specific μ receptor agonist; nor-binaltorphimine, a κ receptor antagonist; 7-benzylidenenaltrexone, a selective δ1-opioid receptor antagonist; or naltriben, a selective δ2-opioid receptor; showed that the beneficial effects were mediated via the δ1-opioid receptor in conjunction with the K_ATP channel [9]. However, while fentanyl is sometimes felt to have minimal effects on the δ1-opioid receptor compared to morphine, hydromorphone, and meperidine, another study has suggested that fentanyl and its derivatives act at the δ1-opioid receptor and induce myocardial protection [11].

While fentanyl is used as an adjunct in our anesthetics, isoflurane is used as the predominant agent at concentrations to produce alveolar levels of ∼1 MAC (age adjusted). At these concentrations, isoflurane provides myocardial protection against ischemia. Several studies show benefits from isoflurane. In a small study, Tomai et al. found that isoflurane found no overall difference in troponin-I and CK-MB levels [4]. However, in the group with depressed left ventricular function (<50%), isoflurane produced significantly lower levels of the two biomarkers of ischemia: (troponin-I 1.1±0.7 vs. 2.3±1.3 ng/ml, P=0.03, and CK-MB 39±10 vs. 57±22 U/l, P=0.04). Similarly, Belhomme et al. found that isoflurane produced smaller release of troponin-I and CK-MB than did the control group [2]. Raphael et al. found that isoflurane induced protection was dependent on phosphatidylinositol-3-kinase/Akt signaling [12]. Phosphatidylinositol-3-kinase/Akt is currently felt to be one of the main intracellular factors responsible for antiapoptic signaling and survival of cells. However, ischemic preconditioning via phosphatidylinositol-3-kinase/Akt was not associated with the mitochondrial K_ATP channels and that, in guinea pig hearts at least, ischemic preconditioning can occur via either system [13]. Di Lisa et al. have recently proposed that the final common pathway for ischemic preconditioning is inhibition of the PTP, and that K_ATP was only one of several factors that modulate PTP state [14].

Those studies, taken in combination, suggest that using an agent, such as isoflurane, that produces ischemic preconditioning via phosphatidylinositol-3-kinase/Akt may override any harmful effects of inhibition of the K_ATP channels by glibenclamide.

There are several limitations to this study. The lack of effect of diabetic medicines on CK-MB levels may be an effect of the anesthetic regimen or the myocardial and cardioplegic techniques for protecting the heart. Specifically, our use of isoflurane as the main anesthetic agent may have sufficiently protected the heart and obviated any harmful effects of sulfonylureas. Studies need to be conducted using other anesthetic drugs as they may provide a different level of protection than does isoflurane. Similarly, our use of tight glycemic control may have provided benefits and in centers where looser glycemic control is employed, the effects of isoflurane will need to be further evaluated [6]. Additionally, we only measured CK-MB once; serial measurements of CK-MB or more sensitive markers of myocardial ischemia such as troponin-I may find differences where single measurement of CK-MB does not [15]. Finally, our study was underpowered (and not designed) to find differences in peri-operative myocardial infarction by electrocardiography or in 30-day mortality. However, the study was adequately powered to find small (9 ng/ml) differences in CK-MB levels.

One strength of our study is that we evaluated effects of diabetic medicines in typical clinical practice, where patients are frequently admitted the day of surgery or have surgery the day after cardiac catheterization. In the current economic climate, admitting patients for at least two days before surgery [1] to switch glibenclamide to subcutaneous insulin is generally not feasible.

In conclusion, we found that the use of glibenclamide or other diabetic medications had no effect on CK-MB levels the morning after patients underwent coronary artery bypass.

References