Endothelial cells can be protected against cytokine-induced toxicity by volatile anesthetics. The authors tested whether inhalation of sevoflurane at subanesthetic concentrations provides protection against postocclusive endothelial dysfunction induced by ischemia-reperfusion injury of the forearm in humans.
Five healthy male volunteers were enrolled in this study with crossover design. Each subject was randomly exposed to 15 min of forearm ischemia in the presence or absence of sevoflurane. Sevoflurane was inhaled at 0.5-1 vol% end-tidal concentrations from 15 min before ischemia until 5 min after the onset of reperfusion. Hyperemic reaction, an indicator of ischemic injury and endothelial function, was determined at 15 and 30 min of reperfusion using venous occlusion plethysmography. Also, markers of leukocyte activation (CD11b, CD42b) were measured by flow cytometry during reperfusion.
Fifteen minutes of forearm ischemia followed by reperfusion diminished postocclusive endothelium-dependent hyperemic reaction at 15 and 30 min of reperfusion. Peri-ischemic inhalation of sevoflurane, targeted at 0.5-1 vol% end-tidal concentrations, markedly improved postocclusive hyperemic reaction. In addition, inhalation of sevoflurane attenuated activation of leukocytes, as measured by CD11b expression, after ischemia-reperfusion injury. No changes in CD42b expression were observed after ischemia-reperfusion of the forearm.
These data suggest that human endothelium, a key component of all vital organs, is receptive to protection by sevoflurane in vivo. Peri-ischemic administration of sevoflurane mimics a combination of pharmacologic preconditioning and postconditioning and protects at even low sedative concentrations (< 1 vol%). Inhibition of leukocyte adhesion is likely to be involved in the protection.
VOLATILE anesthetics protect against ischemia–reperfusion injury and reduce myocardial infarct size.1,2This protection is thought to be due to pharmacologic preconditioning and postconditioning.3–5So far, several clinical studies in patients undergoing coronary artery bypass graft surgery have shown cardioprotection by volatile anesthetics.6,7A double-blinded placebo-controlled trial in patients undergoing coronary artery bypass graft surgery demonstrated not only a decrease of postoperative myocardial dysfunction but also attenuation of renal damage implying “multiorgan” protection by these agents.6Interestingly, decreased expression of platelet endothelial cell adhesion molecule 1, an adhesion molecule responsible for endothelial migration of leukocytes, was found to be reduced in atrial biopsies of sevoflurane-preconditioned patients and was associated with improved event-free long-term survival.8
Vascular endothelium is critically involved in many steps of tissue damage originating from ischemia–reperfusion.9,10Intact vascular endothelium is essential in maintaining arterial tone and coagulation status. Because human endothelium can be protected against ischemia–reperfusion by ischemic10,11and pharmacologic12preconditioning and postconditioning and is ubiquitous in the vasculature of all tissues, it has been speculated that endothelial protection may play a central role in the observed “multiorgan” protection by volatile anesthetics.8So far, it has been shown experimentally that human endothelial cells exposed to volatile anesthetics developed a pronounced resistance against cytokine-induced toxicity, consistent with a preconditioning-like effect.13However, it is unclear whether endothelial protection by volatile anesthetics also occurs against ischemia–reperfusion injury in vivo in humans. To address this important question, we used an established model of endothelial dysfunction after sustained forearm ischemia.10,14In these studies, it has been clearly demonstrated that ischemia–reperfusion injury, as reflected by reduced postocclusive forearm blood flow, is exclusively due to endothelial dysfunction.
Here, we demonstrate that sevoflurane inhalation at subanesthetic concentrations preserves postocclusive blood flow, representing the first in vivo human evidence of endothelial protection against ischemia–reperfusion injury by volatile anesthetics. Future studies should confirm our results in diseased elderly patients under clinical conditions.
Materials and Methods
Subjects
Five healthy male volunteers aged 30 ± 7 (23–43) yr with a body mass index of 25 ± 3 (22–27) kg/m2gave informed signed consent and participated in this study with crossover design. All participants were nonsmokers and refrained from caffeine and black chocolate for 24 h. The subjects fasted overnight and took a single dose of oral ranitidine (300 mg) the evening before participation. Each subject underwent the same experimental protocol with and without sevoflurane at least 7 days apart. The research was conducted in accordance with the Declaration of Helsinki (2000). The study was approved by the local ethics committee of the University Hospital Zurich, Zurich, Switzerland.
Study Protocols
Figure 1summarizes the two protocols of this crossover study. Ischemia–reperfusion of the forearm was used as a model of endothelial dysfunction.10,14An intravenous line was placed in a large cubital vein of the nondominant arm, and 500 ml lactated Ringer’s solution was infused to correct for the fluid deficit. Odansetron (0.5 mg) was administered intravenously to decrease the risk of nausea and vomiting independent of whether sevoflurane was inhaled. Studies were performed in a temperature-controlled and quiet room (24°C) during the early morning hours. The nondominant forearm was rendered ischemic for 15 min by inflating a 12-cm-wide blood pressure cuff placed around the upper arm to 180 mmHg to induce endothelial dysfunction (test ischemia). In the protocol with sevoflurane inhalation, sevoflurane in 50% oxygen was inhaled by the spontaneously breathing volunteers using a tight facemask connected to the common gas outlet of an anesthesia machine (Siemens Servo 900D ventilator; Siemens Life Support Systems, Sona, Sweden) to achieve an end-tidal concentration between 0.5 and 1.0 vol%. Previous studies reported loss of consciousness with sevoflurane inhalation at approximately 1 vol% end-tidal concentration.15At the end of each protocol, 0.4 mg sublingual nitroglycerin was administered to assess endothelium-independent dilation 3 min after nitroglycerin intake. Blood samples were taken from a cubital vein of the nondominant arm before test ischemia (baseline) and after 2, 5, and 10 min of reperfusion. Monitoring consisted of intermittent noninvasive blood pressure measurements, five-lead electrocardiogram, end-tidal carbon dioxide and sevoflurane concentrations (Hellige VICOM-SM SMU 612; PPG, Freiburg, Germany), and Bispectral Index (A2000 monitor® with three adhesive electrodes to the forehead, single channel: Fp1–Fpz, version 3.3; Aspect Medical Systems, Inc., Natick, MA) for depth of sedation.
Assessment of Resistance Vessel Endothelial Function Using Hyperemic Blood Flow Response in the Forearm
Study subjects were positioned comfortably on a bed in a supine position with a slight elevation of the head and with both arms extended (45° abducted) and elevated above the level of the right atrium to ensure complete and rapid venous return to the heart. Mercury-in-silastic strain-gauge venous occlusion plethysmography (Vasoquant 4000; ELCAT GmbH, Wolfratshausen, Germany) was used to measure forearm blood flow. Strain-gauges were positioned tightly around the largest circumferences of the forearms. Venous congestion was achieved by inflating a cuff around the upper arm to 40 mmHg. The recording period of blood flow consisted of four cycles of venous occlusion followed by deflation (each 5 s). Reactive hyperemia was induced by 4 min of blood flow arrest to the forearm by inflating the blood pressure cuff placed around the upper arm 20 mmHg above the systolic blood pressure. Hyperemic reaction was divided into an early hyperemic reaction (EHR) reflecting peak flow at the onset of reperfusion (first measurement after deflation of cuff), and late hyperemic reaction (LHR) reflecting maintenance of hyperemia (mean of the three subsequent measurements). Previous studies showed that hyperemic blood flow regains basal conditions approximately 40–60 s after a 4-min occlusion period.16Reactive hyperemia was determined at baseline before test ischemia on the dominant forearm (to avoid inadvertent ischemic preconditioning), 15 min after test ischemia on the nondominant ischemic forearm, and 30 min after test ischemia on both forearms (fig. 1). During recording, the hands were excluded from the circulation by inflation of wrist cuffs to 220 mmHg to minimize the contribution of the hand skin flow.
Determination of CD11b and CD42b Expression in Leukocytes Using Flow Cytometry
CD11b expression is an indicator of neutrophil activation, whereas CD42b expression on leukocytes indicates formation of leukocyte–thrombocyte complexes. CD11b and CD42b were determined as previously described.10Collected blood samples were immediately processed for flow cytometry. Heparinized blood (50 μl) was incubated with antibody solution for 10 min at room temerature. FACS lysing solution (450 μl; Becton Dickinson, Basel, Switzerland) was added and incubated for 20 min. Samples were fixed by adding 500 μl 0.2% formaldehyde in phosphate-buffered saline for 30 min. The FACSCalibur (Becton Dickinson) flow cytometer was used to measure R-phycoerythrin (PE) fluorescence at 580 nm and fluorescein isothiocyanate (FITC) fluorescence at 515 nm. Leukocytes were distinguished from each other by typical physical characteristics, resulting in well-delineated cellular subpopulations that are easily identified on forward and side-scatter plots. Monoclonal antibodies for polymorphonuclear granulocytes (CD15, PE-labeled, clone 80H5, Immunotech, Marseille, France) and monocytes (CD14, FITC-labeled, clone 61D3; eBioscience, Wembley, United Kingdom) further served for identification of cellular subgroups. CD11b and CD42b expression were measured by fluorescence intensity of PE-conjugated and FITC-labeled monoclonal antibodies directed against CD11b (PE-labeled, clone 2LPM19C; DAKO, Glostrup, Denmark) and CD42b (FITC-labeled, clone AN51; DAKO). Results were compared with isotype-matched antibodies staining as controls (PE-labeled immunoglobulin G; eBioscience, and FITC-labeled immunoglobulin G; Becton Dickinson). A minimum of 20,000 events was counted on each sample.
Statistical Analysis
Data are given as mean ± SD. Forearm blood flow was measured in milliliters per 100 g tissue per minute and expressed as percent change compared with baseline flow measurements. For each participating subject, activation of leukocytes on the ischemic limb was expressed as the ratio of CD11b mean fluorescence intensities (MFIs) between time-matched control and sevoflurane blood samples as follows:
where i = 1, 2, …, 5 and j = baseline, 2 min, 5 min, 10 min.
Subsequently, data were normalized to baseline. Paired t test and repeated-measures analysis of variance were used for comparison. P < 0.05 was considered significant. Analyses were performed using SigmaStat Version 2 (SPSS Science, Chicago, IL).
Results
All study subjects tolerated the procedure without complications, except for one experiencing transient nausea without vomiting. Pain caused by the 15 min of forearm ischemia was rated as 2–3 on a visual analog scale by all participants in the protocol without sevoflurane inhalation as opposed to 0 on the visual analog scale in the protocol with sevoflurane inhalation. The 15 min of test ischemia and inhalation of sedative concentrations of sevoflurane had only marginal effects on blood pressure and end-tidal carbon dioxide concentrations (table 1). All participants responded to verbal commands and could be easily aroused on tactile stimulation. None of them experienced an excitement phase during sedation.
Sustained Ischemia Followed by Reperfusion Deteriorates Endothelium-dependent Hyperemic Reaction in the Forearm
A test ischemia of 15 min on the forearm was used to elicit endothelial dysfunction. Hyperemic blood flow response was determined using venous occlusion plethysmography after 15 and 30 min of reperfusion on the ischemic (nondominant) side and after 30 min of reperfusion on the control arm. EHR (peak flow at reopening of the forearm vessels) was unaffected by test ischemia, whereas LHR (maintenance of hyperemic reaction) was markedly diminished (percent change from baseline at 15 min: −38 ± 18%, P = 0.01; percent change from baseline at 30 min: −32 ± 13%, P = 0.006; figs. 2A and B). Administration of sublingual nitroglycerin did not reverse the impairment in forearm blood flow (P = 0.55), indicating postocclusive endothelium-dependent but smooth muscle–independent dysfunction in vasomotion. Hyperemic blood flow response was unaffected by test ischemia on the control (nonischemic, dominant) arm (data not shown).
Sevoflurane Inhalation Potentiates Endothelium-dependent Hyperemic Reaction
Each subject underwent the same study protocol with and without sevoflurane inhalation. Sevoflurane inhalation was initiated 15 min before the onset of test ischemia and was maintained throughout the first 5 min of reperfusion (fig. 1). Thus, the total time of sevoflurane administration was 35 min. There was no consistent improvement of EHR after sevoflurane inhalation (fig. 2A). However, sevoflurane administration consistently and markedly improved LHR after test ischemia (figs. 2B and 3). Likewise, sevoflurane inhalation enhanced LHR but not EHR on the nonischemic arm 25 min after cessation of sevoflurane inhalation (fig. 4). Notably, end-tidal concentrations, as measured after deeply exhaling, were 0% in all subjects 10 and 25 min after cessation of sevoflurane inhalation. Sublingual administration of nitroglycerin did not affect forearm perfusion (P = 0.95), pointing to the importance of endothelium-dependent vasomotion in the model used. Collectively, sevoflurane efficaciously protects the endothelium against ischemia–reperfusion injury.
Sevoflurane Inhalation Attenuates Activation of Leukocytes after Ischemia–Reperfusion Injury
To elucidate possible mechanisms underlying the protection by sevoflurane, we determined the expression of CD11b and CD42b from forearm blood samples of the ischemic side. Activation of leukocytes on the ischemic limb was expressed as the ratio of CD11b mean fluorescence intensities between time-matched control and sevoflurane blood samples. Sevoflurane administration reduced the ratio by more than 50% at 2, 5, and 10 min of reperfusion in granulocytes (P = 0.002) and monocytes (P = 0.005) indicating a marked reduction in CD11b expression by sevoflurane during reperfusion (fig. 5). No changes in the percentages of CD11b expressing granulocytes or monocytes were observed during reperfusion. In contrast to the surface marker CD11b, no alterations in CD42b expression or in the percentage of CD42b-positive granulocytes or monocytes (leukocyte–platelet complexes) were detected (data not shown).
Sevoflurane Provides Endothelial Protection at Sedative (Subanesthetic) Concentrations (< 1 vol% End-tidal)
Sevoflurane was administered to the study subjects to obtain an end-tidal concentration between 0.5 and 1 vol% (table 1). All participants responded to verbal commands and exhibited arousal in response to tactile stimulation compatible with a mean Bispectral Index value of 74 (table 1), but experienced amnesia for the entire duration of sevoflurane inhalation. Importantly, sevoflurane at even low subanesthetic concentrations provided endothelial protection.
Discussion
This study shows for the first time in humans in vivo that peri-ischemic sevoflurane inhalation attenuates ischemia–reperfusion injury of the endothelium, as assessed by hyperemic blood flow response and CD11b activation in the forearm. The second important finding of this study is that even subanesthetic concentrations of sevoflurane, which induced sedation in the participants (mean Bispectral Index ranging from 68 to 81), were efficacious to elicit protection. Finally, enhanced hyperemic reaction was observed on both sides, including the nonischemic forearm, providing evidence that sevoflurane acts systemically on the endothelium. Taken together, our data suggest that endothelial protection by sevoflurane may be possible to harness in unanesthetized but sedated subjects. Nevertheless, our results should be confirmed in diseased elderly patients under clinical conditions.
The current study used strain-gauge venous occlusion plethysmography to measure hyperemic blood flow response after prolonged test ischemia of the forearm. Venous occlusion plethysmography is a reliable noninvasive tool for investigating vascular function,17and its reproducibility is excellent.18The rate and swelling of the forearm during occlusion of venous return is used to assess the rate of arterial inflow, which is dependent on normal function of the endothelium. Reactive hyperemia in the peripheral arteries is mediated by the endothelial release of nitric oxide, and changes in vascular resistance and blood flow are largely attributable to the bioavailability of nitric oxide.19Other vasodilators such as prostaglandins and adenosine may also contribute to reactive hyperemia.20As reported previously,10no effect of nitroglycerin on hyperemic reaction after prolonged test ischemia of the forearm was observed in our study, consistent with an endothelium-dependent but smooth muscle–independent vasomotion effect. Interestingly, EHR representing blood flow at the onset of vessel opening did not show consistent improvement after sevoflurane treatment as opposed to LHR. However, at very high flow rates typically occurring at the onset of hyperemia, arterial inflow may be impeded before the end of venous occlusion and may thus falsify the relation between the rate of increase in forearm volume and arterial flow. Also, EHR measurements may be more prone to technical artifacts such as brachial artery collapse because of the blood pressure cuff. Finally, we cannot completely rule out that with a higher number of participants, EHR might have become also significantly regulated after sevoflurane treatment. Irrespectively, our results clearly demonstrate marked improvement in LHR after peri-ischemic sevoflurane inhalation.
Mechanisms Underlying Endothelial Protection by Sevoflurane
Volatile anesthetics protect against ischemia–reperfusion. One suggested mechanism is the attenuation of leukocyte activation,21a primary mediator of local and remote tissue damage after ischemia–reperfusion. Mobert et al. 22simulated ischemia–reperfusion by incubating endothelial cells with H2O2for activation. Immediate expression of adhesion molecules occurred, which was abrogated by isoflurane exposure.22De Rossi et al. 23reported that isoflurane at 1 minimum alveolar concentration (MAC) inhibited the activation of L-selectin, CD11a, and CD11b. Interestingly, although 0.5 MAC of isoflurane was unable to affect expression of CD11a/b, shedding of L-selectin, responsible for neutrophil tethering and rolling, was attenuated. Another in vitro study showed decreased expression of intercellular adhesion molecule 1 and E-selectin after tumor necrosis factor α stimulation in desflurane-treated human umbilical vein endothelial cells.24Although we observed a reduction in CD11b expression in our study during reperfusion in the sevoflurane protocol, this reduction is unlikely to be the sole mechanism of protection. No significant changes in platelet–granulocyte or platelet–monocyte complexes were observed after forearm ischemia in our model. Hence, other mechanisms related to adenosine triphosphate–dependent potassium channels,3G protein–coupled receptors,25,26activation of protein kinase C,27and reperfusion injury salvage kinases including protein kinase B and extracellular signal–regulated kinase might be involved in the observed protection. These signaling components are closely linked to the ability of volatile anesthetics to elicit pharmacologic preconditioning and postconditioning,28,29and some of them have been already been demonstrated to be important in the endothelial protection by preconditioning of the forearm in healthy volunteers.12,14
Questioning the “1-MAC Dogma” of Organ Protection?
So far, most experimental and clinical studies showed organ protection with volatile anesthetics at concentrations of 1 MAC or higher. Using a cellular model of simulated ischemia, Zaugg et al. 3reported a concentration-dependent protection by isoflurane and sevoflurane in isolated rat ventricular myocytes. Isoflurane and sevoflurane at 0.5 MAC already elicited marked protection against ischemia, whereas 2 MAC did not provide better protection than 1.5 MAC. Obal et al. 30demonstrated the highest degree of protection by sevoflurane postconditioning with 1 MAC compared with 2 MAC in an in vivo rat model of regional ischemia of the heart. Interestingly, Morisaki et al. 31found an up-regulation of the deleterious P-selectin in rat mesenteric vessels in situ after increasing the sevoflurane concentration from 1 to 2 MAC, raising the possibility that leukocyte–endothelium interaction may be most favorably affected at lower rather than higher concentrations of volatile anesthetics. In our study, sevoflurane was titrated to a subanesthetic target end-tidal concentration between 0.5 and 1 vol%, and Bispectral Index was used as an additional indicator of depth of sedation. The observations in our study now provide evidence that the endothelium in man may be receptive to protection at concentrations well below 1 MAC.
Clinical Implications
Endothelial function is deteriorated in a wide range of conditions such as mental and physical stress, pain, diabetes mellitus, and arterial hypertension,32–34which are associated with increased perioperative cardiovascular morbidity and mortality. Reduced endothelial dilator function during ischemia–reperfusion may exacerbate vasospasm and thus is a critical early event of organ injury. The results of this study now imply that sevoflurane administration may protect endothelial function against ischemia–reperfusion and thereby increase postocclusive organ blood flow and tissue oxygenation. Because the endothelium is a key component in all body tissues, it can be speculated that sevoflurane inhalation may exert whole body protection. The fact that this protection can be achieved under conscious sedation further opens the possibility to use inhalational sedation during diagnostic and interventional procedures in cardiology or endovascular procedures of high-risk patients.
Limitations of the Study
We have carefully excluded confounding factors that may affect vasoreactivity, and a crossover study design was chosen to overcome interindividual variability. Nonetheless, the small sample size of the study may potentially infiltrate type I error. Moreover, the forearm ischemia model does not induce necrosis of endothelial cells, but rather a state of “endothelial stunning,” which resolves spontaneously within 60 min.35,36Whether the observed protection is also effective in more substantial endothelial injury, particularly in patients with preexisting endothelial dysfunction, must be established. Also, it is unclear to what extent these results in the brachial vasculature can be extrapolated to the coronary vasculature. Finally, additional experiments are required to unravel the underlying mechanisms.
In conclusion, our data suggest that in man, the endothelium, a key component of all vital organs, is receptive to protection by sevoflurane at even low sedative concentrations (< 1 vol% end-tidal).
The authors thank Ursula Steckholzer, B.Sc. (Technician, Department of Trauma Surgery, University Hospital Zurich, Zurich, Switzerland), for her expert technical assistance as well as the postanesthesia care unit nurses, colleagues, and volunteers who participated in this study.