Is There a Place for Preconditioning During Cardiac Operations in Humans?

Transcription

1 Is There a Place for Preconditioning During Cardiac Operations in Humans? Bruno Pouzet, MD, Jean-Baptiste Lecharny, MD, Monique Dehoux, MD, Sandrine Paquin, MD, Masafumi Kitakaze, MD, PhD, Jean Mantz, MD, PhD, and Philippe Menasché, MD, PhD Departments of Cardiovascular Surgery, Anesthesiology, and Biochemistry, Groupe Hôspitalier Bichat-Claude Bernard, Paris, France, and National Cardiovascular Center, Fujishirodai Suita, Osaka, Japan Background. Activation of the kinase cascade (protein kinase C (PKC), tyrosine kinase (TK), and mitogenactivated protein kinase (MAPK) is a key feature of the transduction pathway, elicited by preconditioning signals and mediating their cardioprotective effects. We assessed whether such an activation occurred during cardiac operations and could thus represent a target for cardioprotective strategies. Methods. A total of 20 patients undergoing coronary artery bypass grafting surgery were studied. During the first 10 minutes of cardiopulmonary bypass (CPB), 10 were treated with sevoflurane (2.5 minimum alveolar concentration), an inhalational anesthetic that mimics preconditioning through a similar activation of the kinase cascade. Ten case-matched patients undergoing 10 minutes of sevoflurane-free CPB served as controls. Right atrial biopsies were taken before and 10 minutes after CPB and were then processed for the measurement of PKC, TK, and p38 MAPK activities by enzyme assay techniques. Troponin I was also monitored over the first 2 postoperative days. Results. Compared with pre-cpb values, PKC and p38 MAPK activities (in nanomoles per milligram of protein per minute and arbitrary units, respectively) increased significantly and to the same extent in both groups: PKC, from to in controls (p 0.037) and from to in sevoflurane (p 0.016); p38 MAPK, from to in controls (p 0.005) and from to in sevoflurane (p 0.045). Conversely, sevoflurane triggered a significant increase in TK activity (from to picomoles per milligram of protein per minute p ) which did not occur in controls (from to picomoles per milligram of protein per minute, p 0.09). Likewise, the peak postoperative value of troponin I was not different between controls and sevoflurane-treated patients ( vs , p 0.21). Conclusions. Cardiopulmonary bypass triggers an activation of the kinase cascade that is mechanistically linked to opening of potassium channels. The direct opening of these channels by the anesthetic sevoflurane does not increase kinase activation further, nor does it improve markers of cell necrosis, thus suggesting that pharmacologically targeting potassium channels may overlap the preconditioning-like effects of CPB alone. (Ann Thorac Surg 2002;73:843 8) 2002 by The Society of Thoracic Surgeons Accepted for publication Nov 1, Address reprint requests to Dr Menasché, Service de Chirurgie, Cardiovasculaire B, Groupe Hôspitalier Bichat-Claude Bernard, 46, rue Henri Huchard, Paris Cedex 18, France; ccv-bloc.sec3@ bch.ap-hop-paris.fr. The efficacy of ischemic preconditioning in reducing myocardial cell necrosis and improving recovery of function has stimulated intense research for pharmacologic agents that could duplicate the cardioprotective effects of preconditioning and that would be safe to use in the clinical setting of cardiac surgery. The list of available compounds is, however, limited. Adenosine has yielded mixed results [1]; in fact, a trend toward improved patient outcomes has been reported only when the drug was present before, during, and after cardioplegia, ie, in a setting quite different from the preconditioning scenario [2]. In contrast, inhalational anesthetic agents are raising increasing interest because of their purported ability to open adenosine triphosphate (ATP) dependent potassium channels, which are thought to play a major role in mediating the salutary effects of preconditioning. It becomes attractive, then, to use agents that, in addition to keeping the patient asleep, could elicit cardioprotection. A previous study from our group [3] has shown that isoflurane caused upregulation of ecto-5 -nucleotidase, a surrogate marker for activation of protein kinase C (PKC), which is a key step in the preconditioningtriggered signaling pathway, thus supporting the idea that these commonly used anesthetic agents could act as preconditioning mimetics. However, in our study, evidence for such a role was largely indirect. Consequently, the present trial was designed to assess more directly whether a parent drug of isoflurane, sevoflurane, could also turn on the preconditioning pathway. Because there is compelling experimental evidence that sevoflurane also acts through opening of ATP-dependent potassium channels [4 7], the state of which cannot be assessed directly in the in situ human heart, the primary 2002 by The Society of Thoracic Surgeons /02/$22.00 Published by Elsevier Science Inc PII S (01)

2 844 POUZET ET AL Ann Thorac Surg SEVOFLURANE AND CARDIAC PRECONDITIONING 2002;73:843 8 Table 1. Baseline Values Measured Before Sevoflurane Preconditioning Patients Troponin I (ng/ml) PKC (nmol mg 1 min 1 ) TK (pmol mg 1 min 1 ) p38 MAPK (a.u.) Controls Sevoflurane-preconditioned a b c d a p 0.3. b p c p d p a.u. arbitrary units; p38 MAPK p38 mitogen-activated protein kinase; PKC protein kinase C; TK tyrosine kinase. end point of this study was the myocardial tissue levels of PKC, tyrosine kinase (TK), and p38 mitogen-activated protein kinase (MAPK), the cascade activation of which is temporally linked to the potassium channels along the signaling pathway of preconditioning. Patients and Methods Protocol A total of 20 patients undergoing elective coronary artery bypass grafting (CABG) were prospectively studied after approval by our institutional review committee. All patients were anesthetized with an intravenous combination of fentanyl, midazolam, and pancuronium, with care being taken to avoid the use of volatile anesthetic agents until the onset of CPB. Once bypass had been instituted, patients were randomly allocated to receive sevoflurane for 10 minutes or to serve as controls. Sevoflurane was added to the gas mixture of the oxygenator (2.5 minimum alveolar concentration). At the completion of the 10- minute treatment (or after a time-matched period of drug-free CPB in control patients), the aorta was crossclamped and myocardial protection was achieved by continuous retrograde low-dilution blood cardioplegia. Perioperative Myocardial Necrosis Assessment To detect myocardial infarction, blood levels of troponin I were serially measured with the Dimension RXL-HM analyzer using a two-site fluorometric enzyme immunoassay (Dade Behring, La Défense, France) after the induction of anesthesia, 20 hours and 2 days after the surgery. Values are expressed as nanograms per milliliter. Measurement of PKC, TK, and MAPK Activities To detect kinase activation, right atrial biopsy samples were harvested before bypass and then either at the end of the 10-minute preconditioning regimen in sevoflurane-treated patients or after an equivalent pump time in control patients, immediately before aortic crossclamping. Tissue specimens were frozen into liquid nitrogen and stored at 80 C. The activity of PKC was measured, as described previously [8], by enzyme assay with the RPN 77A kit (Amersham), which provides a simple and reliable method of estimating PKC without extensive purification of the samples. Activity of PKC is expressed as nanomoles per milligram of protein per minute. Protein concentrations were measured by the method of Lowry and colleagues [9], with bovine serum albumin used as a standard. Tyrosine kinase activity was assessed by the enzymatic assay technique [10, 11] (using the AUSA tyrosine kinase assay kit, Transbio Corporation, Baltimore, MD) and is expressed as picomoles per milligram of protein per minute. To immunoprecipitate p38 MAPK, 1 g of cardiac tissue was homogenized and an in vitro kinase assay was carried out as described previously [12]. At the end of the kinase reaction, the immunoprecipitates were separated by electrophoresis on a sodium dodecyl sulfate polyacrylamide gel, which was fixed, dried, and subjected to autoimmunography. The activity of p38 MAPK was then measured by an image analyzer and is expressed in arbitrary units. Statistical Analysis The data are represented as the mean 1 SEM, with a value of p less than 0.05 considered significant. All analyses were performed with appropriate software (StatView; SAS Institute, Inc, Cary, NC). Longitudinal studies comparing values over time within each group were achieved using paired t tests. One-way analysis of variance were used to test for differences between groups. Results The base line demographic characteristics of the patients did not differ between the two groups. Left ventricular ejection fraction was 46% 8% in controls and 45% 6% in sevoflurane-treated patients. Aortic cross-clamping times were also comparable between the control and treated groups (47 3 and 53 5 minutes, respectively). Likewise, as shown in Table 1, there was no significant difference in base line variables between the two groups. Compared with pre-cpb values, PKC and p38 MAPK activities both increased significantly and to the same extent in the two groups: PKC, from to in controls (p 0.037) and from to in sevoflurane-treated patients (p 0.016); p38 MAPK, from to in controls (p 0.005) and from to in sevoflurane-treated patients (p 0.045) (Figs 1 and 2). Conversely, sevoflurane triggered a significant increase in TK activity (from to , p ), which did not occur in controls (from to , p 0.09) (Fig 3). The peak postoperative values of troponin I were not

3 Ann Thorac Surg POUZET ET AL 2002;73:843 8 SEVOFLURANE AND CARDIAC PRECONDITIONING 845 Fig 1. Protein kinase C (PKC) activities in controls and sevofluranepreconditioned patients. (Baseline PKC activity in right atrial biopsy samples taken before bypass; 10 min PKC activity at end of 10-minute preconditioning regimen in sevoflurane-treated patients or after an equivalent pump time in control patients; *p vs baseline; #p vs baseline; p 0.2 vs controls.) different between controls and sevoflurane-treated patients (at 20 hours: ng/ml versus ng/ml, p 0.62; at 48 hours: ng/ml versus ng/ml, p 0.21) (Fig 4). Clinical outcomes did not differ between the two groups. One patient in each group required inotropic agents. Comment The basic objective of this study was to assess whether the inhalational anesthetic sevoflurane could mimic classic ischemic preconditioning by opening ATP-dependent Fig 3. Effects of sevoflurane preconditioning on tyrosine kinase (TK) activities. (Baseline protein kinase C activity in right atrial biopsy samples taken before bypass; 10 min protein kinase C activity at end of 10-minute preconditioning regimen in sevofluranetreated patients or after an equivalent pump time in control patients; *p 0.09 vs baseline; #p vs baseline; p 0.2 vs controls.) sarcolemmal/mitochondrial potassium channels during on-pump coronary artery bypass operations with a subsequent reduction of postoperative cell necrosis. Our results only partially support this hypothesis, in that neither the intraoperative biochemical markers of the preconditioning pathway nor the enzymatic markers of postoperative myocardial cell necrosis were found to differ between the two groups. Justification of End Points A major issue associated with studies of preconditioning during clinical cardiac operations is the selection of end points providing reasonable evidence that the pathway has been turned on. Because the states of opening and closing of potassium channels cannot be assessed directly in the in situ human heart, one has to rely on Fig 2. Comparisons of p38 mitogen-activated protein kinase (MAPK) activities before and after sevoflurane preconditioning. (a.u. arbitrary units; Baseline p38 MAPK activity before bypass; 10 min p38 MAPK activity at end of 10-minute preconditioning regimen in sevoflurane-treated patients or at a time-matched study in control patients; *p vs baseline; #p vs baseline; p 0.22 vs controls.) Fig 4. Troponin I release in controls and sevoflurane-preconditioned patients over first 48 hours postoperatively. Baseline indicates troponin I blood levels at the induction of anesthesia. (*p 0.62 vs controls, #p 0.21 vs controls.) (20 hours 20 hours postoperatively; 48 hours 48 hours postoperatively.)

4 846 POUZET ET AL Ann Thorac Surg SEVOFLURANE AND CARDIAC PRECONDITIONING 2002;73:843 8 surrogate markers. In this study, we measured myocardial tissue levels of PKC, TK, and p38 MAPK because a large number of experimental studies have brought compelling evidence that activation of this cascade is mechanistically linked to opening of ATP-sensitive potassium channels. Most recently, the codependent relationship between these two events has been well demonstrated in a sheep model of regional ischemia [13], in which the infarct-limiting effect of preconditioning was abolished by opposite blockade (ie, a potassium channel opener an inhibitor of PKC or an agonist of PKC an inhibitor of potassium channels). That such a relationship also exists in human is suggested by two other studies showing, in human ventricular myocytes, that PKC activates the K ATP current at reduced intracellular concentrations of ATP [14], and that delayed preconditioning involves both the p38 MAPK pathway and mitochondrial potassium channels [15]. We acknowledge that a more direct evidence that kinase activation is a surrogate marker for potassium channel opening would have required pharmacologic manipulations with potassium channel blockers, but this was not attempted for obvious ethical reasons. In fact, there is ongoing controversy regarding the critical juncture at which potassium channels come into play temporally along this pathway. According to the classic scheme, signaling is initiated by activation of various membrane receptors by their agonists (particularly adenosine, bradykinin, opioids, and -agonists). Receptor activation then causes phosphorylation and translocation of PKC and other downstream kinases [16 18], especially TK and p38 MAPK, the phosphorylation of which correlates with the protection afforded by preconditioning [19]. This would ultimately lead to opening of ATP-dependent sarcolemmal/mitochondrial potassium channels [20], the role of which in mediating a preconditioning-type of cardioprotection in the human heart has been established by studies using right atrial trabeculae harvested during cardiac operations [21, 22]. This paradigm, however, has been recently challenged by some studies suggesting that mitochondrial potassium channels could actually act as triggers, rather than end effectors, of the signaling pathway [23, 24]. According to this hypothesis, receptor activation opens mitochondrial potassium channels, causing the mitochondria to produce oxygen-derived free radicals that would then set the heart in a preconditioned state by oxidative activation of the kinase cascade [25 27]. The end effector of this pathway is not yet precisely identified; however, heat shock protein (HSP) 27 is one possible candidate. This small molecular weight compound has been shown to play an important role in maintaining the integrity of the actin cytoskeleton [28] and it is mechanistically linked to the kinase cascade, as it can be phosphorylated by the p38 MAPK [29]. However, regardless of whether potassium channels are upstream or downstream from the kinase cascade, all data converge toward a tight coupling between these two events. Data Analysis In the present study, patients exposed to sevoflurane demonstrated a significant upregulation of PKC, TK, and p38 MAPK shortly after the onset of bypass, thus suggesting that the biochemical pathway involved in preconditioning had been turned on. Our choice of sevoflurane was based on some attractive characteristics of this inhalational anesthetic, which include short half-life, easy handling, and marked cardioprotective properties manifest as a reduction in infarct size [30], an improved recovery of function after unprotected global ischemia [5, 6, 31] and cardioplegic arrest [4, 32], and a better preservation of postischemic coronary flow and nitric oxide release [6]. The observation that these protective effects are consistently abrogated by potassium channel blockers [4 7] strongly suggests the involvement of these channels, similar to what occurs in classic ischemic preconditioning. Unexpectedly, however, upregulation of the kinase cascade suggestive of potassium channel activation was also found to occur in control patients after a timematched period of bypass, and its magnitude was not different from that seen after exposure to sevoflurane except in the case of TK. One explanation for the greater activation of TK in the sevoflurane group could be a more specific effect of the drug on TK receptors, a finding that has actually been reported in skeletal muscle [33]. Likewise, enzymatic markers of cell necrosis, as assessed by troponin I levels, were similar in the two groups. This observation suggests a preconditioning-like effect of CPB itself, which could be explained by the endogenous release of membrane receptor activators of the signal transduction pathway such as adenosine or catecholamines. This hypothesis is consistent with the previous findings of Burns and coworkers [34], who demonstrated that the infarct-limiting effect of ischemic preconditioning could actually be duplicated by a brief period of CPB before the coronary occlusion. Taken together, our data suggest that CPB triggers an activation of the kinase cascade, which makes likely an opening of potassium channels. Administration of a direct potassium channel opener such as sevoflurane does not induce a greater upregulation of the kinase cascade, nor does it translate into a reduced postoperative release of enzymatic markers of myocardial cell necrosis. Thus, in the clinical setting of on-pump coronary artery operations, a sole pretreatment by potassium channel openers might be redundant with CPB, thereby generating uncertainty as to whether these drugs can elicit additional cardioprotective effects through the kinase-potassium channel pathway. This redundancy could have been missed by studies using human right atrial trabeculae to demonstrate a preconditioning effect because of the pre-cpb timing of tissue biopsies. Study Limitations This somewhat negative conclusion should, however, be tempered by the methodologic limitations inherent in the design of this pilot study. First, the small sample size may

5 Ann Thorac Surg POUZET ET AL 2002;73:843 8 SEVOFLURANE AND CARDIAC PRECONDITIONING 847 have precluded the demonstration of between-group differences in enzyme release. This issue will be addressed by an ongoing trial that has been powered to detect such differences if they occur. However, it is noteworthy that in our previous small study of isoflurane [3], a trend towards lower postoperative troponin I levels could be demonstrated, and it is thus possible that inhalational anesthetic agents (such as isoflurane, sevoflurane, enflurane, desflurane, and halothane) display different degrees of cardioprotection [30, 35 37]. Second, the right atrial site of tissue sampling and the measurement of global enzymatic activities versus specific kinases isoforms pose additional methodologic limitations inherent in the use of human specimens. However, a recent study of the MAPK kinase/extracellularregulated kinase pathway has shown that atrial and ventricular myocardium responded similarly to CPB [38]. We finally acknowledge that the timing of drug administration may have not been optimal. In fact, to maximize the chances of demonstrating an effect of sevoflurane, the anesthetic was given as an immediate pre cross-clamp treatment, thereby skipping the washout period intrinsic to true preconditioning protocols. It is conceivable that a better therapeutic exploitation of the cardioprotective properties of the drug would be obtained by extending its time of administration, as suggested by the cardioprotective effects of sustained mitochondrial potassium channel opening during index ischemia [39] and reperfusion [40]. In support of this assumption, exposure to sevoflurane during the sole period of reperfusion has also been shown to elicit cardioprotection [30, 32, 37]. In conclusion, taking activation of the kinase cascade as a marker for potassium channel activation, our data failed to demonstrate that the inhalational anesthetic sevoflurane turned on this pathway characteristic of preconditioning to a greater extent than CPB alone. However, the rationale for the use of this type of anesthetic agents and the preclinical data are strong enough to warrant further studies to assess how the cardioprotective effects of these agents can best be therapeutically exploited. References 1. Belhomme D, Peynet J, Florens E, Tibourtine O, Kitakaze M, Menasché P. Is adenosine preconditioning truly cardioprotective in coronary artery bypass surgery. Ann Thorac Surg 2000;70: Mentzer RM, Birjiniuk V, Khuri S, et al. Adenosine myocardial protection. Preliminary results of a phase II clinical trial. Ann Surg 1999;229: Belhomme D, Peynet J, Louzy M, Launay J-M, Kitakaze M, Menasché P. Evidence for preconditioning by isoflurane in coronary artery bypass graft surgery. Circulation 1999; 100(Suppl II):II Coetzee JF, le Roux PJ, Genade S, Lochner A. Reduction of postischemic contractile dysfunction of the isolated rat heart by sevoflurane: comparison with halothane. Anesth Analg 2000;90: Mathur S, Farhangkhgoee P, Karmazyn M. Cardioprotective effects of propofol and sevoflurane in ischemic and reperfused rat hearts: role of K ATP channels and interaction with the sodium-hydrogen exchange inhibitor HOE 642 (cariporide). Anesthesiology 1999;91: Novalija E, Fujita S, Kampine JP, Stowe DF. Sevoflurane mimics ischemic preconditioning effects on coronary flow and nitric oxide release in isolated hearts. Anesthesiology 1999;91: Toller WG, Kersten JR, Pagel PS, Hettrick DA, Warltier DC. Sevoflurane reduces myocardial infarct size and decreases the time threshold for ischemic preconditioning in dogs. Anesthesiology 1999;91: Kitakaze M, Hori M, Morioka T, et al. 1 -Adrenoreceptor activation increases ecto-5 -nucleotidase activity, and adenosine release in rat cardiomyocytes by activating protein kinase C. Circulation 1995;91: Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurement with folin phenol reagent. J Biol Chem 1951; 193: Kemp BE, Pearson RB. Design and use of peptide substrates for protein kinases. Methods Enzymol 1991;200: Tremblay L, Gingras D, Boivin D, Beliveau R. Tyrosine protein kinase activity in renal brush-border membranes. Biochim Biophys Acta 1992;1108: Ishizuka T, Terada N, Gerwins P, et al. Mast cell tumor necrosis factor productions is regulated by MEK kinases. Proc Natl Acad Sci USA 1997;94: Gaudette GR, Krukenkamp IB, Saltman AE, Horimoto H, Levitsky S. Preconditioning with PKC and the ATP-sensitive potassium channels: a codependent relationship. Ann Thorac Surg 2000;70: Hu K, Duan D, Li GR, Nattel S. Protein kinase C activates ATP-sensitive K current in human and rabbit ventricular myocytes. Circ Res 1996;78: Carroll R, Yellon DM. Delayed cardioprotection in a human cardiomyocyte-derived cell line: the role of adenosine, p38 MAP kinase and mitochondrial KATP. Basic Res Cardiol 2000;95: Baines CP, Wang L, Cohen MV, Downey JM. Protein tyrosine kinase is downstream of protein kinase C for ischemic preconditioning s anti-infarct effect in the rabbit heart. J Mol Cell Cardiol 1998;30: Maulik N, Yoshida T, Zu Y-L, Sato M, Banerjee A, Das DK. Ischemic preconditioning triggers tyrosine kinase signaling: a potential role for MAPKAP kinase 2. Am J Physiol 1998;44: H Miura T, Liu Y, Kita H, Ogawa T, Shimamoto K. Roles of mitochondrial ATP-sensitive K channels and PKC in antiinfarct tolerance afforded by adenosine A 1 receptor activation. J Am Coll Cardiol 2000;35: Weinbrenner C, Liu GS, Cohen MV, Downey JM. Phosphorylation of tyrosine 182 of p38 mitogen-activated protein kinase correlates with the protection of preconditioning in the rabbit heart. J Mol Cell Cardiol 1997;29: Gross GJ, Fryer RM. Sarcolemmal versus mitochondrial ATP-sensitive K channels and myocardial preconditioning. Circ Res 1999;84: Ghosh S, Standen NB, Galinanes M. Evidence for mitochondrial K ATP channels as effectors of human myocardial preconditioning. Cardiovasc Res 2000;45: Pomerantz BJ, Robinson TN, Morrell TD, Heimbach JK, Banerjee A, Harken AH. Selective mitochondrial adenosine triphosphate-sensitive potassium channel activation is sufficient to precondition human myocardium. J Thorac Cardiovasc Surg 2000;120: Pain T, Yang X-M, Critz SD, et al. Opening of mitochondrial K ATP channels triggers the preconditioned state by generating free radicals. Circ Res 2000;87: Gross GJ, Fryer RM. Mitochondrial K ATP channels. Triggers or distal effectors of ischemic or pharmacological preconditioning. Circ Res 2000;87: Baines CP, Goto M, Downey JM. Oxygen radicals released during ischemic preconditioning contribute to cardioprotection in the rabbit myocardium. J Mol Cell Cardiol 1997;29:

6 848 POUZET ET AL Ann Thorac Surg SEVOFLURANE AND CARDIAC PRECONDITIONING 2002;73: O Rourke B. Myocardial K ATP channels in preconditioning. Circ Res 2000;87: Forbes RA, Steenbergen C, Murphy E. Diazoxide-induced cardioprotection requires signaling through a redoxsensitive mechanism. Circ Res 2001;88: Baines CP, Liu GS, Birincioglu M, Critz SD, Cohen MV, Downey JM. Ischemic preconditioning depends on interaction between mitochondrial K ATP channels and actin cytoskeleton. Am J Physiol 1999;276:H Sugden PH, Clerk A. Stress-responsive mitogen-activated protein kinases (c-jun N-terminal kinases and p38 mitogenactivated protein kinases) in the myocardium. Circ Res 1998; 83: Preckel B, Schlack W, Comfère T, Obal D, Barthel H, Thamer V. Effects of enflurane, isoflurane, sevoflurane and desflurane on reperfusion injury after regional myocardial ischaemia in the rabbit heart in vivo. Br J Anaesth 1998;81: Li F, Hayes JK, Wong KC, Szakacs J. Administration of sevoflurane and isoflurane prior to prolonged global ischemia improves heart function of isolated rat heart. Acta Anaesthesiol Sin 2000;38: Preckel B, Thamer V, Schlack W. Beneficial effects of sevoflurane and desflurane against myocardial reperfusion injury after cardioplegic arrest. Can J Anaesth 1999;46: Kudoh A, Matsuki A. Sevoflurane stimulates inisitol 1,4,5- triphosphate in skeletal muscle. Anesth Analg 2000;91: Burns PG, Krukenkamp I, Caldarone CA, Gaudette GR, Bukhari EA, Levitsky S. Does cardiopulmonary bypass alone elicit myoprotective preconditioning. Circulation 1995; 92(Suppl II):II Conradie S, Coetzee A, Coetzee JF. Anesthetic modulation of myocardial ischemia and reperfusion injury in pigs: comparison between halothane and sevoflurane. Can J Anaesth 1999;46: Roscoe AK, Christensen JD, Lynch C III. Isoflurane, but not halothane, induces protection of human myocardium via adenosine A 1 receptors and adenosine triphosphatesensitive potassium channels. Anesthesiology 2000;92: Schlack W, Preckel B, Stunneck D, Thamer V. Effects of halothane, enflurane, isoflurane, sevoflurane and desflurane on myocardial reperfusion injury in the isolated rat heart. Br J Anaesth 1998;81: Araujo EG, Bianchi C, Sato K, Faro R, Li XA, Sellke FW. Inactivation of the MEK/ERK pathway in the myocardium during cardiopulmonary bypass. J Thorac Cardiovasc Surg 2001;121: Wang S, Cone J, Liu Y. Dual roles of mitochondrial K ATP channels in diazoxide-mediated protection in isolated rabbit hearts. Am J Physiol 2001;280:H Fryer RM, Hsu AK, Gross GJ. Mitochondrial K ATP channel opening is important during index ischemia and following myocardial reperfusion in ischemic preconditioned rat hearts. J Mol Cell Cardiol 2001;33:831 4.