Preconditioning by Sevoflurane Decreases Biochemical Markers for Myocardial and Renal Dysfunction in Coronary Artery Bypass Graft Surgery: A Double-blinded, Placebo-controlled, Multicenter Study

The local ethics committees of the three hospital centers approved this study, and written informed consent was obtained from all patients. Seventy-two patients scheduled for elective CABG surgery were enrolled at three hospital centers in Switzerland, the University Hospital Zurich, the City Hospital Triemli Zurich, and the University Hospital Lausanne.

Inclusion criteria were being scheduled for elective CABG surgery on CPB circuit with cardiac arrest and age of 40–80 yr. Exclusion criteria were as follows: concomitant aortic or valvular surgery, elevated cardiac enzyme concentrations within 24 h before surgery, unstable angina, angina within 24 h before surgery, left ventricular bundle branch block or marked resting ST–T-segment abnormalities precluding electrocardiography interpretation, cardiac pacemaker dependency, hemodynamic instability with the need for medical or mechanical inotropic support, anesthesia or surgery within 24 h before CABG surgery, and administration of adenosine triphosphate–sensitive potassium channel agonists or antagonists such as diazoxide, nicorandil, sulfonylurea, or theophylline.

The immediate study period included the preoperative period with a duration of at least 12 h before CABG surgery through 72 h after surgery (fig. 1). On the preoperative day, monitoring with Holter electrocardiography was started until 72 h after CABG surgery. On the same day, patients were randomly allocated (by opening of an envelope) to preconditioning with placebo or sevoflurane. All patients received midazolam or flunitrazepam for premedication. Anesthesia was induced in all patients with propofol or etomidate, opioids including fentanyl or remifentanil, and the muscle relaxants pancuronium or vecuronium. All necessary monitoring lines were then inserted, and anesthesia was maintained with propofol infusion, continuous infusion or repeated doses of opioids, and pancuronium or vecuronium administration as required. Median sternotomy and pericardiotomy were performed, and the right atrium and the ascending aorta were cannulated. After administration of heparin (300 U/kg), standard CPB with a disposable hollow fiber oxygenator and aprotinin (2 × 10⁶U) administered to the priming volume was started. With completely established CPB (2.4 l·min⁻¹·m⁻²body surface area) and with the heart totally decompressed, the vaporizer was set to 4 vol% for exactly 10 min. To apply the preconditioning stimulus, a sevoflurane vaporizer Sevotec 5 (Abbott, Baar, Switzerland) was integrated into the CPB machine between the fresh gas flow inlet and the membrane oxygenator. Preconditioning with placebo was achieved by administration of an oxygen–air gas mixture without sevoflurane. Patients assigned to preconditioning with sevoflurane received sevoflurane 4 vol% corresponding to 2 minimum alveolar concentration. Sevoflurane concentrations were also measured by gas chromatography (Perkin–Elmer, Norwalk, CT) in selected patients (n = 10); these concentrations reached 1.4 ± 0.4 mm in the venous reservoir of the CPB circuit at the end of the preconditioning procedure. Blinding of the treatment to surgeons, anesthesiologists, and perfusionists was guaranteed by covering the anesthetic liquid label of the vaporizer, which was either completely filled with sevoflurane or totally empty depending on the randomization. Phenylephrine was administered to maintain aortic blood pressure above 50 mmHg. Immediately after this preconditioning period, the aorta was cross clamped and the cold (4°C) cardioplegic solution was administered to achieve cardiac arrest. No volatile anesthetics were administered during the study except for the preischemic period in patients randomized to the sevoflurane group. Depending on the anatomic circumstances and the surgeons’ preferences, cardioplegic solution with or without blood was administered anterogradely, retrogradely, or both every 10 min intermittently. Distal anastomoses were performed during a single period of aortic cross clamping. After completion of the distal anastomoses, cold cardioplegic solution was administered through the vein grafts. The proximal anastomoses were conducted during side clamping. In all patients, at least one internal mammary artery graft was used. Core temperature was actively cooled to 32°C or was allowed to drop spontaneously with α-stat regulation of blood pH. Atrial samples were taken before preconditioning at the time of cannulation and at the end of the preconditioning stimulus immediately before induction of cardioplegic arrest (fig. 1). After termination of CPB, heparin was antagonized by protamine. Hemoglobin concentrations were kept above 7 g/dl during CPB and above 9 g/dl postoperatively. After surgery, all patients were admitted to the intensive care unit. Patients from both treatment groups received the same routine postoperative care as determined by the caring physicians.

Blood samples were obtained preoperatively, at arrival in the intensive care unit, and 24, 48, and 72 h after surgery; they were stored at −20°C until analysis. The following parameters were determined using the Roche/Hitachi 917 or the Roche Elecsys 2010 (Roche Diagnostics, Mannheim, Germany): NT-proBNP (electrochemiluminescence sandwich immunoassay)—sensitivity, >5 ng/l; intra- and interassay coefficients of variance, <3%; normal value (97.5th percentile, age of older than 50 yr) for men, <334 ng/l; normal value for women, <227 ng/l; total creatine kinase (enzymatic reaction with NADPH formation)—sensitivity, 2 U/l; intra- and interassay coefficients of variance, <2%; normal value for men, ≤170 U/l; normal value for women, ≤145 U/l; CK-MB activity (immunologic ultraviolet assay)—sensitivity, 5 U/l; intra- and interassay coefficients of variance, <2%; normal value, ≤24 U/l; cTnT (electrochemiluminescence sandwich immunoassay)—sensitivity, >0.01 μg/l; intra- and interassay coefficients of variance, <5%; normal value, <0.01 μg/l; creatinine (Jaffé reaction)—sensitivity, 8.8 μm; intra- and interassay coefficients of variance, <2.5%; normal range for men, 62–106 μm; normal range for women, 44–80 μm; and high-sensitivity C-reactive protein (immunoturbidimetric assay)—sensitivity, 0.3 mg/dl; intra- and interassay coefficients of variance, <6%; normal value, <1 mg/dl. CysC assays were purchased from Dako A/S, Glostrup, Denmark (particle-enhanced turbidimetric assay: sensitivity, 0.2 mg/l; intra- and interassay coefficients of variance, <3.5%; normal range, 0.74–1.50 mg/l).

Three-channel digital Holter electrocardiography monitoring was begun at least 12 h before surgery and continued for 72 h postoperatively (8500 series; GE Marquette Medical Systems, WI). Seven bipolar leads were simultaneously recorded using silver–silver chloride electrodes. Lead resistance was tested every 6 h, and faulty leads were replaced. The effect of patient positional variation was measured in supine and upright positions before the study. Holter electrocardiography data were analyzed for ST-segment changes (Marquette Holter analysis system software version 8500; GE Marquette Medical Systems, WI) indicative of ischemia after exclusion of abnormal QRS complexes, such as ventricular ectopic beats or beats with conduction abnormalities. ST-segment changes were trended in three leads for the duration of the recording. Baseline ST-segment levels were defined as the average ST-segment during stable periods (at least 10 min) preceding each ischemic period. Two independent assessors blinded to group assignment, clinical course, and the various plasma parameters reviewed all ischemic periods. Disagreement was resolved by consensus. Electrocardiographic ischemic periods were defined as reversible ST-segment changes lasting at least 1 min and involving either a shift from baseline of ≥0.1 mV of ST depression or a shift from baseline of ≥0.2 mV of ST–T elevation at the J point. ST-segment depression was measured 60 ms after the J point, unless that point fell within the T wave, in which case it was shortened to the J point plus 40 ms. Holter analysis was corrected for positional variation by taking the maximum shift noted by positional changes. The following parameters were measured as indicators of the severity of each ischemic episode: total episodes per patient with ischemia, maximal ST depression, duration of longest ischemic episode, total area under the curve (defined as the integral of ST depression in mV × duration), and maximal area under the curve. Perioperative time was divided into three separate episodes (T1 = preoperative period; T2 = first 24 h postoperatively; T3 = 24–72 h postoperatively). Intraoperative Holter electrocardiography recordings were not used for comparison of the two anesthetic groups due to the many confounding variables such as movement, artifacts due to chest opening, electrocautery interference, ventricular pacing, and electrolyte artifacts. Holter electrocardiography recordings were further analyzed for the occurrence of dysrhythmias using the above-mentioned software. Ventricular and supraventricular ectopic beats were identified and counted as isolated, bigeminal cycles, couplets, or runs. The number of runs and the percent of recorded time with atrial fibrillation were calculated separately for each period.

Translocation of PKC isoforms in response to preconditioning was assessed by immunofluorescence in all samples. 27,28Briefly, right atrial tissue samples were placed in Hanks solution, immediately frozen in liquid nitrogen, and stored at −70°C. Cryosections (5 μm) were prepared with a cryostat (Cryo-star HM 560 M; Microtom, Kalamazoo, MI) and collected on slides precoated with gelatin. All sections were fixed for 10 min in 100% acetone at −20°C, rinsed with phosphate-buffered saline, and incubated in 10% normal goat serum for 30 min to block nonspecific binding. Sections were incubated for 1 h at room temperature with primary antibodies. Rabbit polyclonal antibodies to PKC δ and PKC ε (Santa Cruz Biotechnology, Santa Cruz, CA) diluted in phosphate-buffered saline (1:400–1:1,200) containing 2% normal goat serum were added. Antibodies to PKC were combined with the following: mouse monoclonal α-prohibitin-1 antibody (Research Diagnostics, Flanders, NJ) (1:25) as a mitochondrial marker; guinea pig polyclonal m-dystrophin antibodies (gift from Richard A. Zuellig, Ph.D., Institute of Pharmacology, University Zurich, Switzerland) (1:2,000) as a sarcolemmal marker; mouse monoclonal α-N -cadherin antibody (Sigma, St. Louis, MO) as a marker for the intercalated disks (1:500); and rabbit polyclonal α-myomesin antibodies as a marker for cardiomyocytes (gift from Hans M. Eppenberger, Ph.D., Professor of Cell Biology, Department of Cell Biology, Swiss Federal Institute of Technology Zurich, Zurich, Switzerland) (1:200). All sections were then washed with phosphate-buffered saline twice for 5 min each and incubated for 1 h with a mixture of secondary antibodies conjugated to Alexa Fluor 555 goat α-rabbit, Alexa Fluor 488 goat α-mouse, or Alexa Fluor 488 goat α-guinea pig antibody (Molecular Probes, Eugene, OR) (1:500) and 4′,6-diamidino-2-phenylindole-2 hydrochloride (DAPI) (10 ng/ml) (Sigma, St. Louis, MO) in phosphate-buffered saline at room temperature. After washing with phosphate-buffered saline, sections were protected with coverslips using DAKO mounting medium (DAKO, Carpinteria, CA). Sections were analyzed by epifluorescence microscopy using an upright microscope (Axioplan 2; Zeiss, Jena, Germany) with appropriate filter blocks for the detection of fluorescein isothiocyanate, tetramethyl rhodamine isothiocyanate, CY5, and ultraviolet fluorescence. In addition, confocal images were obtained with an LSM Pascal confocal microscope (Zeiss, Jena, Germany) using the appropriate laser lines and filter blocks (488, 540, and 640 nm). Since aged human tissue has a significant amount of lipofuscin-induced autofluorescence and the CY5 filter shows only the lipofuscin autofluorescence, the relative signal intensity in the three channels (fluorescein isothiocyanate, tetramethyl rhodamine isothiocyanate, and CY5) was used to separate on the composite image PKC labeling from unspecific lipofuscin signals. 29Randomly chosen fields of sections from all samples were examined for translocation of PKC δ and PKC ε to sarcolemma (colocalization with dystrophin), mitochondria (colocalization with prohibitin-1), intercalated disks (colocalization with N -cadherin), or nuclei (colocalization with DAPI) without prior knowledge of the treatment. Confocal imaging was used to quantify PKC ε translocation to nuclei of cardiomyocytes. Five randomly chosen fields at a ×400 magnification from sections of samples from each treatment group (placebo and sevoflurane) collected before and after preconditioning were analyzed for colocalization of PKC δ with nuclei (DAPI staining). Double staining with myomesin was used to assure that only PKC ε translocation to cardiomyocytes was counted. The number of PKC ε–positive nuclei was expressed as the percentage of total nuclei of myocytes per field.

Medical charts were reviewed, and the caregivers were interviewed daily for the occurrence of postoperative cardiovascular and renal adverse events. Adverse events (as opposed to myocardial and renal injury markers determined at the end of the study) were diagnosed by the independently managing clinicians. The diagnosis of a new postoperative myocardial infarction required a new Q wave, persistent ST–T-segment changes as defined by Minnesota Codes, 30and/or association with elevated CK-MB isoenzyme activity (concentration, >100 U/l). The diagnosis of a cerebrovascular insult required the presence of clinical symptoms and/or a positive computer-assisted tomography scan. The diagnosis of significant postoperative renal dysfunction required newly established postoperative hemodialysis or hemofiltration.

The sample size was calculated based on previously reported data for BNP concentrations. 12–16With an expected difference of 30% between group means, 40% SD of the means, α= 0.05, and β= 0.8, a sample size of 24 patients per group was necessary. A logarithmic transformation was applied to all data to ensure a normal distribution before statistical analysis. Two-factor repeated-measures analysis of variance was used to evaluate differences over time between groups for all parameters determined in plasma samples. The Greenhouse–Geisser correction was used to address the deviation from sphericity. Multiple paired t  tests were used to compare the parameters at each time point with the respective preoperative baseline measurements within groups, and unpaired t  tests were used to compare these parameters at each time point between groups. All other data were analyzed using unpaired t  tests for parametric data or Mann–Whitney tests for nonparametric data. Categorical data were analyzed using the two-tailed Fisher exact test or chi-square test, as appropriate. If not otherwise indicated, values are mean ± SD. P  values were multiplied by the number of comparisons that were made (Bonferroni correction), and corrected P < 0.05 was considered statistically significant. Analyses were performed using SigmaStat and SPSS (SPSS, Chicago, IL).

Patient characteristics are listed in table 1. The placebo and sevoflurane groups were similar with respect to all clinical data. In addition, there was no difference between groups regarding the operative data (table 2), except for phenylephrine administration during the preconditioning process. Sevoflurane-treated patients required significantly more phenylephrine during the preconditioning process than did placebo-treated patients to maintain blood pressure above 50 mmHg. However, there was no difference in mean arterial blood pressure between the groups during preconditioning, and sevoflurane could be administered safely. Two patients in the placebo group and one patient in the sevoflurane group had a postoperative myocardial infarction as determined by the caring physicians. One placebo-treated patient required transient postoperative inotropic support with an intraaortic balloon pump due to hemodynamic instability. No cerebrovascular insult occurred in any of the patients, and none of the patients required new postoperative hemofiltration or dialysis.

Preoperative NT-proBNP concentrations were similar in placebo- and sevoflurane-treated patients (fig. 2A). In both groups, a marked increase was observed from 24 to 72 h after surgery (time effect, P < 0.001). However, postoperative plasma concentrations were significantly lower in sevoflurane-treated patients than in placebo-treated patients (group effect, P < 0.001; group–time interaction, P = 0.003). In addition, postoperative peak plasma NT-proBNP concentrations were markedly lower in sevoflurane-treated patients than in placebo-treated patients (2,180 ± 1,118 ng/l vs.  4,841 ± 2,937 ng/l, respectively;P < 0.001).

Preoperative plasma concentrations of cTnT, total creatine kinase, and CK-MB activity were within normal limits in all patients and were not different between groups (figs. 2B–D). A significant postoperative increase in concentrations of cTnT, total creatine kinase, and CK-MB activity was observed in the placebo and sevoflurane groups (time effect for all parameters, P < 0.001). Although CK-MB activity peaked immediately after surgery, cTnT demonstrated a more protracted course, with the highest plasma concentration 48 h after surgery. There was no significant difference over time between groups (cTnT: group effect, P = 0.56; group–time interaction, P = 0.98; total creatine kinase: group effect, P = 0.89; group–time interaction, P = 0.92; CK-MB activity: group effect, P = 0.79; group–time interaction, P = 0.88), nor was there a difference with respect to peak concentrations between groups (postoperative peak cTnT concentration: 0.52 ± 0.37 μg/l in the sevoflurane group vs.  0.57 ± 0.53 μg/l in the placebo group;P = 0.65) (figs. 2B–D). Eighteen patients (25%) had postoperative peak cTnT concentrations of >0.65 μg/l, indicative of major perioperative myocardial damage 31: 11 (31%) in the placebo group (1.25 ± 0.69 μg/l) and seven (19%) in the sevoflurane group (0.91 ± 0.18 μg/l) (P = 0.28). There was no significant increase in plasma NT-proBNP concentrations in patients with postoperative peak cTnT concentrations of >0.65 μg/l, as compared with patients with postoperative peak cTnT concentrations of <0.65 μg/l (4,504 ± 3,025 ng/l vs.  3,337 ± 2,686 ng/l, respectively;P = 0.12).

Holter electrocardiography–detected ST-segment changes were separately analyzed for the preoperative period, the first 24 h immediately after surgery, and from 24 to 72 h after surgery. There was no difference with respect to the incidence (table 3) as well as the severity of ischemic events (table 4) between both groups for any of the analyzed periods. In addition, no difference was detected regarding perioperative ventricular or supraventricular arrhythmias. The incidence of dysrhythmias at the time of aortic cross clamp release was not different between the groups (table 5).

Translocation of the PKC isoforms δ and ε in response to sevoflurane treatment was assessed by colocalization with sarcolemma (dystrophin), mitochondria (prohibitin-1), intercalated disks (N -cadherin), and nuclear staining with DAPI. There was clear translocation of PKC δ to sarcolemma in sevoflurane-treated atrial samples as compared with samples without the preconditioning stimulus (figs. 3A–C). In addition, PKC ε translocated to mitochondria (figs. 3D–F), intercalated disks (figs. 3G–I), and nuclei after sevoflurane treatment. To quantify the PKC ε translocation in response to sevoflurane exposure, the percentage of nuclei with clear PKC ε colocalization was determined in cardiomyocytes of atrial samples before and after sevoflurane preconditioning as well as in time-matched placebo-treated samples using confocal microscopy (fig. 4A–D). The results of these experiments demonstrated that the percentage of PKC ε–positive nuclei was markedly increased after the application of sevoflurane (before sevoflurane application, 14 ± 11%; after sevoflurane application, 63 ± 14%;P < 0.001).

There was no difference in baseline plasma CysC concentrations between groups. CysC concentrations significantly increased immediately postoperatively and peaked at 48 h after surgery for both groups (time effect, P < 0.001). CysC concentrations were markedly higher for placebo-treated patients than for sevoflurane-treated patients (group effect, P < 0.009; group–time interaction, P < 0.001), including postoperative peak plasma CysC concentrations (1.58 ± 0.67 mg/l vs.  1.21 ± 0.28 mg/l, respectively;P = 0.0039) (fig. 5A). Using the prospectively defined cut-off concentration of 1.5 mg/l for CysC, 10 patients in the placebo group but only two patients in the sevoflurane group had new postoperative renal dysfunction (P < 0.015). Plasma creatinine concentrations were slightly, but not significantly, increased in placebo-treated patients (time effect, P < 0.001; group effect, P = 0.28; group–time interaction, P = 0.98) (fig. 5B). However, postoperative peak plasma creatinine concentrations were significantly higher in the 10 placebo-treated patients with newly detected postoperative renal dysfunction (148 ± 33 μm) than in patients without postoperative renal dysfunction (96 ± 21 μm) (P < 0.0001). As a result of dilution, plasma creatinine concentrations decreased in both groups during the first hours after surgery.

To our knowledge, this study is the first double-blinded, placebo-controlled, randomized clinical trial evaluating pharmacological preconditioning elicited by volatile anesthetics. The principal new findings of the study are as follows. First, sevoflurane preconditioning significantly decreased postoperative release of NT-proBNP, a highly sensitive biochemical marker of the myocardial contractile state, in patients undergoing CABG surgery after cardioplegic arrest. Second, the concept of PKC translocation as a pivotal signaling step in the initiation of preconditioning elicited by volatile anesthetics could be directly visualized and confirmed for the first time in human myocardium. Finally, although not a primary outcome measure, the postoperative renal glomerular filtration rate was markedly higher among sevoflurane-preconditioned patients, indicating a preconditioning-like preservation of renal function by systemically administered sevoflurane.

The B-type natriuretic peptide (BNP) together with the A-type natriuretic peptide regulate blood pressure by modulating water and salt homeostasis. 32Although BNP was first isolated from porcine brain in 1988, 33studies over the last decade clearly have shown that the heart is the main source of this hormone. 34The physiologically active form of BNP (C-terminal BNP) derives from the precursor proBNP, is composed of a 17-amino acid residue ring structure, and is continuously released from ventricular myocytes into the circulation in response to intracavitary pressure and diastolic wall stress (Laplace law). 35Conversely, the biologically inactive N-terminal cleavage product, NT-proBNP, is also circulating at blood concentrations similar to C-terminal BNP in healthy volunteers, 36but it exhibits greater proportional and absolute increments for a defined degree of cardiac dysfunction than does C-terminal BNP, suggesting its superior role as an indicator of cardiac impairment and cardiovascular prognosis. 13,37,38BNP mediates its biologic effects including natriuresis, diuresis, vasodilation, and antagonism to renin, aldosterone, and catecholamine action by the guanylate cyclase–linked receptor natriuretic receptor A, which forms cGMP as a second messenger and is found in a variety of tissues including the endothelium. 39Results from numerous studies over the last decade strongly confirm the close correlation of blood BNP concentrations increased in proportion to the severity of cardiac dysfunction. Accordingly, increasing BNP concentrations correspond to New York Heart Association (New York, NY) functional classes, closely reflect simultaneously measured hemodynamic indices including pulmonary capillary wedge pressure, and correlate inversely with cardiac output and ejection fraction as determined by angiography, echocardiography, and radionuclide ventriculography. 19–23After myocardial infarction, C-terminal BNP and particularly NT-proBNP show an inverse relation with ventricular function including left ventricular ejection fraction and ventricular systolic and diastolic volumes, which are indicators of the postinfarct ventricular remodeling process. 13Moreover, increased BNP concentrations are independent predictors for future cardiovascular adverse events, including heart failure and death, in a variety of clinical settings without the use of other invasive or expensive diagnostic tests. 12,14,17,18Notably, elevated blood NT-proBNP concentrations were even successfully used as targets for the titration of effective angiotensin-converting enzyme inhibitor therapy for patients with congestive heart failure. 38Recently, BNP also received attention as a marker of functional recovery of the myocardium following CPB. Morimoto et al.  16found in a heterogeneous population undergoing cardiac surgery that blood BNP concentrations became markedly and acutely elevated starting 12 h after CPB, returning to baseline not until 3 weeks postoperatively. In this study, peak postoperative BNP concentrations were inversely correlated with postoperative left ventricular ejection fraction and cardiac output. In another study of patients undergoing CABG surgery with CPB, peak and cumulative BNP release, as determined from coronary sinus blood during myocardial reperfusion, closely correlated with myocardial lactate production. 40Taken together, BNP has a pivotal role in the counterregulatory response to cardiac failure, thereby acting as a highly sensitive index of the extent and severity of cardiac dysfunction.

On the basis of this pathophysiologic background, we chose NT-proBNP, beside other cardiac injury markers, to assess the myocardial preconditioning effects of sevoflurane in patients undergoing elective CABG surgery with total CPB. The results of our study demonstrated that a single brief exposure of sevoflurane immediately prior to cardioplegic arrest dramatically reduced the postoperative NT-proBNP concentration, a biochemical marker of myocardial dysfunction. This is the first time the beneficial cardiac effects of preconditioning could be directly linked to and determined by the neurohormonal indicator BNP. Surprisingly, we did not find a significant difference between groups with respect to markers of myocardial necrosis such as CK-MB activity or cTnT concentration. However, since increased concentrations of myocardial enzymes after CABG surgery derive from cannulation of the right atrium, cardioplegia itself, inadequate delivery of cardioplegia in the presence of stenosis or hypertrophy, vigorous manipulations of the heart, prolonged surgery, and differences in surgical techniques, they may not properly reflect the protection afforded by preconditioning. Moreover, properly administered cardioplegia may prevent major structural damage and yet leave postischemic myocardium amenable to functional improvement by preconditioning. The observation that cTnT and NT-proBNP release did not parallel each other in fact supports the recently observed concept that NT-proBNP reflects the contractile state of the myocardium per se  rather than structural damage or ischemia. 14,17No difference was observed between groups with respect to perioperative ST-segment changes. However, electrocardiography is less sensitive for the diffuse forms of postcardiac surgery–related myocardial ischemia and may underestimate myocardial injury. 41We did not observe an antiarrhythmic effect of preconditioning elicited by volatile anesthetics. Preischemic administration of volatile anesthetics can decrease the incidence of postischemic ventricular arrhythmias, particularly in small animals. However, in larger animals and humans, this effect is more controversial. 9,42Together, each parameter (cTnT, Holter electrocardiography, and BNP) reflects a unique independent axis in the evolution of cardiac outcome and adds complementary incremental information about the cardioprotective mechanisms underlying anesthetic preconditioning in human myocardium.

So far, preconditioning by volatile anesthetics has been evaluated in four unblinded studies with patients undergoing on-pump CABG surgery. 8–11An additional small study compared sevoflurane with propofol anesthesia in patients who underwent CABG surgery and found improved myocardial function and decreased release of cardiac troponin I with the sevoflurane-based anesthetic. 43These studies, although partially contradictory in their results, provide evidence that volatile anesthetics may protect human myocardium and support our observation of a cardiac benefit from sevoflurane preconditioning. Penta de Peppo et al.  11applied enflurane, 1.3 vol%, via  the respirator over 5 min immediately before CPB. Consistent with our results, preconditioning afforded increased postoperative left ventricular contractility, but no decrease in CK-MB or troponin T release was found. In another study, 8administration of isoflurane, 2.7 vol%, for 5 min on the fully established CPB followed by a 10-min washout period before aortic cross clamping only demonstrated a tendency to lower CK-MB and troponin I release (not statistically significant). Moreover, Haroun-Bizri et al.  9demonstrated increased postoperative cardiac index after administering isoflurane, 0.5–2 vol%, shortly before CPB via  the respirator. This study further showed decreased intraoperative ST-segment changes, but no reduction was found in reperfusion dysrhythmias after the release of the aortic cross clamp in preconditioned patients. No ST-segment changes or arrhythmias were assessed postoperatively in this study. Tomai et al.  10applied isoflurane, 1.5 vol%, for 15 min through the respirator followed by a washout period of 10 min before starting CPB. In contrast to the aforementioned studies, no difference in postoperative cardiac index and left ventricular ejection fraction was noticed between the control group and the preconditioned group. However, a decrease in postoperative CK-MB and troponin I release could be detected in a subgroup of patients with poor preoperative left ventricular ejection fraction (<50%).

Although a series of experimental studies implicated PKC activation as a pivotal step in anesthetic preconditioning, 5to date it is not clear whether preconditioning-specific translocation of PKC to subcellular targets also occurs in response to volatile anesthetics, particularly in human myocardium. Volatile anesthetics prime the activation of sarcolemmal and mitochondrial K^ATPchannels, 5,6the putative end-effectors of preconditioning via  stimulation of adenosine receptors and subsequent activation of PKC as well as via  increased formation of NO and free oxygen radicals. Activated PKC acts as a pivotal amplifier of the preconditioning stimulus and stabilizes, by phosphorylation, the open states of the mitochondrial and sarcolemmal K^ATPchannels. The opening of K^ATPchannels ultimately elicits cytoprotection by decreasing cytosolic and mitochondrial Ca²⁺overload. PKC with its isoform-specific translocation is one of the most prominent kinases associated with the preconditioned state. 44The results of this study clearly showed that PKC δ was translocated to the sarcolemma and PKC ε was translocated to mitochondria, intercalated disks, and nuclei in response to sevoflurane preconditioning. Translocation of PKC from the cytosolic to the particulate compartments is commonly used as an index of PKC activation. We used immunohistochemical analysis to determine subcellular translocation of PKC isoforms. 27,28Since this technique primarily provides qualitative data, we also determined the number of nuclei with clear PKC ε colocalization to quantify the preconditioning effect. The results of these studies strongly support the concept that volatile anesthetics induce PKC translocation in vivo . These findings are also in clear accordance with a previous observation that ecto-5′-nucleotidase, a marker of PKC activity, is increased in response to isoflurane administration in human myocardium. 8Interestingly, because PKC δ and PKC ε are Ca²⁺-insensitive isoforms, it appears that volatile anesthetics mainly affect Ca²⁺-independent PKC signaling pathways. We recently demonstrated that cardioprotection afforded by volatile anesthetics was sensitive to modulations of the NO signaling pathway. 5Further, NO is known to induce activation and translocation of PKC ε in cardiomyocytes. 45Thus, it may well be possible that volatile anesthetics elicit the subcellular redistribution of PKC in human myocardium at least, in part, by the NO signaling pathway.

It is well documented that CABG surgery with the use of CPB causes ischemia, inflammation, and a significant embolic load to the whole body, leading to global organ dysfunction. We thought that systemically administered sevoflurane might exert beneficial effects on other vital organs over the effects on myocardial tissue. Since renal failure is an important complication in patients undergoing CABG surgery with CPB with an incidence of up to 16% and an associated mortality rate of 13%, 46we decided to determine the protective effect of sevoflurane preconditioning on postoperative kidney function. Mangano et al.  47recently showed that after cardiac surgery, a reduction of creatinine clearance by only 10% was significantly and independently associated with intensive care unit stay, prolonged hospitalization, and resource utilization. In addition, a mild elevation in the preoperative serum creatinine concentration (130–150 μm) significantly increases the need for postoperative mechanical renal support and in-hospital mortality. 48Although there was no clinically overt new renal dysfunction in both of the groups in our study, sevoflurane preconditioning markedly improved the postoperative glomerular filtration rate, as determined by serial CysC measurements. CysC is a basic low-molecular-weight cysteine proteinase inhibitor (13,359 Da), which is the product of a “housekeeping” gene expressed in all nucleated cells at a constant rate. 25It is freely filtered by the glomerulus but not secreted by the tubulus. Thus, its serum concentration is entirely determined by the glomerular filtration rate. Although plasma creatinine concentrations and even creatinine clearance may significantly overestimate the glomerular filtration rate among patients with impaired renal function, plasma CysC concentrations are independent of age, sex, muscle mass, or dietary factors and more readily and accurately reflect even mild renal dysfunction. 24The observation that sevoflurane preconditioning may prevent the CPB-associated subtle decrements in the glomerular filtration rate provides evidence that pharmacological preconditioning may protect the kidney and even other vital organs subjected to CPB-associated damage. 49Whether the observed renoprotective effect is the direct result of a preconditioning effect by sevoflurane operative on renal tissue or the renal vasculature or whether this functional improvement merely reflects the preserved cardiac function, as determined by NT-proBNP, cannot be decided from the present study. However, preconditioning in our study prevented transient renal dysfunction. At present, it is unclear to what extent this treatment may prevent manifestation of clinical disease.

This study has some important limitations. The study was designed with sufficient power to detect differences in postoperative NT-proBNP concentrations. It was not designed to detect differences in in-hospital clinical outcome. In addition, several variables associated with the protective effects of preconditioning were evaluated as secondary outcome measures to support the results of our primary outcome variable NT-proBNP. For these variables, no individual sample size was calculated a priori . However, this study had a power of 80% to detect a reduction of 20% in plasma cTnT 43or CysC concentrations 47,48and the incidence of arrhythmias but a reduction of 70% in the incidence of ischemic ST-segment changes. 50Although this study is the largest evaluating anesthetic preconditioning in humans to date, we cannot exclude that for the variables where no difference was found a larger sample size might have resulted in significant differences. The present study did not simultaneously measure hemodynamic parameters to evaluate the cardiac contractile state. However, the diagnostic and prognostic value of BNP as a correlate of myocardial function was previously carefully evaluated and well established by angiography, echocardiography, concomitant measurements of hemodynamic parameters, and radionuclide ventriculography in a variety of clinical settings including the specific situation of cardiac recovery following CABG surgery. 15,16Specifically, since the BNP concentration is known to remain elevated 1 month after the end of surgical stress, 16it is highly unlikely that increased postoperative blood NT-proBNP concentrations merely reflect transient volume overload due to perioperative fluid management. Since more phenylephrine was administered during the process of preconditioning in the sevoflurane group to maintain blood pressure above 50 mmHg, we cannot entirely exclude that the stimulation of α-adrenergic receptors may have contributed to the observed cardiac and renal effects. However, in a separate multivariate analysis, the amount of administered phenylephrine did not prove to be a significant predictor of postoperative peak NT-proBNP or CysC concentrations, making a significant contribution of phenylephrine to the observed effects unlikely. Since all patients received aprotinin in their priming volume, it is unlikely that this treatment accounted for the observed differences between the groups. Because this study only evaluated immediate perioperative effects of sevoflurane preconditioning, we can only speculate on potential beneficial long-term effects of this treatment. Future studies should address this important issue.

In summary, pharmacological preconditioning with sevoflurane was found to reduce postoperative myocardial dysfunction in patients undergoing CABG surgery as assessed by the biochemical marker NT-proBNP. This study also directly visualized for the first time the translocation of PKC δ and ε to subcellular targets in human myocardium in response to sevoflurane. Finally, we also observed a decrease in CPB-induced renal dysfunction in the preconditioned patients.

The authors thank the cardiac surgeons, intensive care physicians and nurses, and local research associates from the three Swiss medical centers who facilitated the completion of these studies. They also thank Sandra Bühlmann, M.D. (Resident in Anesthesiology, City Hospital Triemli, Zurich, Switzerland), and Christoph Hofer, M.D. (Attending Physician in Anesthesiology, City Hospital Triemli), for the collection of clinical data, and Eliana Lucchinetti, Ph.D. (Laboratory for Biomechanics, Swiss Federal Institute of Technology, Zurich, Switzerland), for helpful discussions and for proofreading the manuscript.