Sevoflurane Ameliorates Gas Exchange and Attenuates Lung Damage in Experimental Lipopolysaccharide-induced Lung Injury

After approval was obtained from the local animal care and use committee (Zurich, Switzerland), pathogen-free, male Wistar rats weighing 350–500 g (Charles River, Sulzfeld, Germany) were used. The rats were housed in standard cages at 22°± 1°C under a 12/12-h light–dark regimen. Food and water were supplied ad libitum .

Rats were anesthetized with intraperitoneal sodium thiopental (100 mg/kg; Pentothal, Ospedalia AG, Hünenberg, Switzerland). For continuous propofol infusion and fluid administration (10 ml · kg⁻¹· h⁻¹sodium chloride), the tail vein was cannulated with a sterile 22-gauge catheter (BD Insyte; Becton Dickinson S.A., Madrid, Spain). A sterile polyethylene catheter for blood sampling and blood pressure monitoring was placed into the left carotid artery (pressure transducer; Spacelabs, Hertford, United Kingdom). The rats were tracheotomized, and a sterile metal cannula was inserted into the trachea, followed by mechanical ventilation in pressure-controlled mode (Servo Ventilator 300; Maquet, Solna, Sweden). Peak inspiratory pressure was 14 cm H²O, with a positive end-expiratory pressure of 3 cm H²O. The fractional inspired oxygen concentration (Fio²) was 1.0, inspiratory:expiratory ratio was 1:2, and respiratory frequency was 30 min⁻¹. Arterial blood samples were analyzed at 0, 2, 4, and 6 h for arterial oxygen tension (Pao²) and arterial carbon dioxide tension (Paco²). Body temperature was maintained at 37°C by a warming lamp.

To evaluate the oxygenation capability of the lung over time, the ratio of oxygen tension to inspired oxygen fraction (Pao²/Fio²) was calculated at defined time points for each group (0, 2, 4, and 6 h), as well as alveoloarterial oxygen tension difference with values obtained from the Federal Office of Meteorology and Climatology MeteoSwiss (Zurich, Switzerland).

Rats were randomly assigned to four different groups: (1) propofol–lipopolysaccharide (n = 6), (2) propofol–phosphate-buffered saline (PBS) (n = 4), (3) sevoflurane–lipopolysaccharide (n = 6), and (4) sevoflurane–PBS (n = 4). Rats in the lipopolysaccharide groups were intratracheally instilled with 150 μg Escherichia coli –lipopolysaccharide (serotype 055:B5; Sigma Aldrich, Buchs, Switzerland) in 300 μl PBS.16Both control groups (propofol–PBS and sevoflurane–PBS) received 300 μl intratracheally instilled PBS. After the application of either lipopolysaccharide or PBS, rats were ventilated as described and propofol was infused intravenously at a dose of 10–20 mg · kg⁻¹· h⁻¹to maintain anesthesia. Propofol, 97% (Sigma Aldrich, Buchs, Switzerland), was dissolved in a 14% Cremophor EL (Biochemika Fluka, Buchs, Switzerland) solution to a final concentration of 10 mg/ml.17Two hours after the onset of lung injury, the anesthetic was changed according to the protocol to either propofol or sevoflurane for a subsequent 4 h (6-h injury model with 4 h of postconditioning). Sevoflurane was administered using the AnaConDa system. The expiratory concentration of sevoflurane was measured with a multigas analyzer (VEO Multigas Monitor; PHASEIN Medical Technologies, Danderyd, Sweden). In all experiments, the concentration of sevoflurane was 1–2 vol% (0.5–1 minimum alveolar concentration, respectively).

At the end of the experiment, animals were killed. The right heart was flushed with 10 ml PBS, after which a bronchoalveolar lavage was performed (3 × 10 ml PBS, pooled). The collected fluid was centrifuged at 4°C (1,500g  for 10 min), and aliquots of the supernatant were frozen at −20°C. The cell pellet was resuspended in 1 ml PBS. After the cells were dyed with trypan blue, they were counted with a Neubauer chamber.

Finally, lungs were shock-frozen in liquid nitrogen and stored at −80°C for isolation of RNA.

To assess the differences in lung permeability between the study groups, total protein and albumin were measured in bronchoalveolar lavage fluid (BALF). Total protein was determined using a Bradford assay (Bio-Rad, Hercules, CA). Albumin levels were assessed using an enzyme-linked immunosorbent assay (ELISA; Bethyl Laboratories Inc., Montgomery, TX) according to the manufacturer’s protocol. The detection range for albumin was 7.8–10,000 ng/ml.

Sandwich ELISAs were performed according to the manufacturer’s protocol assessing the chemokines cytokine-induced neutrophil chemoattractant-1 (CINC-1; R&D Systems Europe Ltd., Abingdon, United Kingdom) and monocyte chemoattractant protein-1 (MCP-1; BD Biosciences, San Diego, CA). The detection range was 7.8–1,000 pg/ml for CINC-1 protein and 62.5–16,000 pg/ml for MCP-1.

Total RNA was isolated from lung tissue using the RNeasy Mini Kit (Qiagen, Basel, Switzerland) according to the manufacturer’s protocol. Tissue was lysed in the provided buffer and subsequently loaded on RNeasy mini spin columns. RNA was eluted with RNAse-free water. Total amounts and purity of RNA were determined by absorbance at 260 nm and the 260/280-nm absorbance ratio, respectively.

Reverse transcription was performed with 0.8 μg total RNA at 20°C for 5 min, 42°C for 30 min, and 95°C for 5 min. Random hexanucleotide primers and murine leukemia virus reverse transcriptase were used for complementary DNA synthesis.

Real-time quantitative TaqMan polymerase chain reaction (PCR) was performed on a GeneAmp 5700 system (P.E. Applied Biosystems, Waltham, MA). Specific primers (Microsynth, Balgach, Switzerland) and labeled TaqMan probes (Roche Applied Science, Basel, Switzerland) were designed for MCP-1, CINC-1, and 18S. The TaqMan universal PCR Master Mix (Applied Biosystems, Branchburg, NJ) was used for the assays in a final reaction volume of 15 μl. All primers and probes used in the experiments are presented in table 1. Each experimental PCR run was performed in duplicate with simultaneous assays for controls with no template.

For quantitation of gene expression, the comparative C^tmethod was used as described by Livak et al.  18The C^tvalues of samples (propofol– lipopolysaccharide and sevoflurane–lipopolysaccharide) and controls (propofol–PBS and sevoflurane–PBS) were normalized to the housekeeping gene (18S) and calculated as 2^−ΔΔCt, where ΔΔC^t=ΔC^t,samples−ΔC^t,controls.

For histologic examination, lungs (previously not flushed in the respiratory compartment) were fixed with 3% paraformaldehyde in PBS and then imbedded in TissueTek (Sakura Finetec Inc., Torrance, CA). A series of microsections (7 μm) of every study group was stained with hematoxylin and eosin. Lung injury was quantified by three blinded researchers, using a lung injury score described previously.19,20The lung pathology was assessed by various degrees of edema (0 = no, 1 = mild, 2 = moderate, 3 = severe) and reactive cell infiltration (0 = no, 1 = mild, 2 = moderate, 3 = severe). Adding these two individual scores resulted in the final score ranging from 0 to 6.

Surfactant was pelleted by high-speed centrifugation (30,000g  for 45 min at 4°C). The crude pellet was resuspended in 110 μl saline, 0.9%, and total phospholipid content was measured using the method of Stewart.21Fifty microliters of sample was added to glass tubes containing 2 ml spectroscopic-grade chloroform. Two milliliters ammonium ferrocyanate, 3.04% (wt/vol), and ferric chloride hexahydrate, 2.7% (wt/vol), in distilled H²O was added, and the mixture was vortexed for 1 min. Standards (0–100 mg/ml) were prepared with phosphatidylcholine in chloroform. The lower chloroform phase was withdrawn, and absorption was measured at 488 nm with a quartz cuvette.

The cell line, kindly provided by Roscoe Warner, Ph.D. (Research Assistant Professor, Department of Pathology, University of Michigan, Ann Arbor, Michigan), was cultured in Dulbecco modified Eagle medium (DMEM; Invitrogen AG, Basel, Switzerland), supplemented with 10% fetal bovine serum (FBS), 1% penicillin–streptomycin, and 1% HEPES buffer in an incubator with 5% CO². They were grown in uncoated 35 × 10-mm plates (Corning Inc., Corning, NY) to more than 95% confluence. DMEM–10% FBS was replaced by DMEM–1% FBS 24 h before lipopolysaccharide stimulation. Rat pulmonary artery endothelial cells (RPAECs) were stimulated with lipopolysaccharide from Escherichia coli , serotype 055:B5 (Sigma, Buchs, Switzerland), in a concentration of 20 μg/l in DMEM–1% FBS for 6 h (control group only stimulation with PBS in DMEM–1% FBS instead of lipopolysaccharide).

The L2 cell line (CCL 149; American Type Culture Collection, Rockville, MD) was derived through cloning of adult female rat lung of alveolar epithelial cell (AEC) type II origin.22The cells were cultured and stimulated in the same way as RPAECs.

For the incubation time of 6 h, the following carbon dioxide concentrations were chosen: 5% (control), 7.5%, and 10%. After the incubation supernatants were collected, ELISAs were performed and expression of CINC-1 and MCP-1 was analyzed.

Control and stimulated RPAECs and AECs were exposed to propofol diluted in 14% Cremophor EL for 6 h. After the incubation, supernatants were collected, ELISAs were performed, and expression of CINC-1 and MCP-1 was analyzed.

For all experiments, cell viability was 95% as determined by measurement of lactate dehydrogenase (CytoTox 96, Non-Radioactive Cytotoxicity Assay; Promega, Madison, WI).

Values were expressed as mean ± SD.

The ratio of oxygen tension to inspired oxygen fraction and alveoloarterial oxygen tension difference (Po²difference) data were tested by analysis of variances for repeated measurements (two-way analysis of variance). The interaction testing between group and time from the repeated measures has been performed. ELISA data were tested by analysis of variance for repeated measurements (one-way analysis of variance) with a Tukey–Kramer multiple post hoc  test. Real-time PCR data were tested using a t  test with two-tailed hypothesis testing. GraphPad Prism4 and GraphPad Instat3 (GraphPad Software, La Jolla, CA) were used for statistical analyses. P  values of 0.05 or less were considered statistically significant.

Intratracheal lipopolysaccharide resulted in a significant decrease of Pao²/Fio²for both anesthetics (propofol and sevoflurane) compared with the PBS controls after 6 h of injury (fig. 1). Animals in the sevoflurane–lipopolysaccharide group had a significantly higher Pao²/Fio²(243 ± 94 mmHg [32.4 kPa]) compared with the propofol–lipopolysaccharide group (88 ± 19 mmHg [11.7 kPa]) after 6 h of lipopolysaccharide injury. There were no significant differences between the two PBS groups (sevoflurane–PBS, 415 ± 28 mmHg [55.3 kPa]; propofol–PBS, 433 ± 32 mmHg [57.7 kPa]; fig. 1). The influence of factor sevoflurane and time was P = 0.0169 and P = 0.0202, respectively. No significant interaction could be found between sevoflurane and time (P = 0.3284).

Accordingly, intratracheal lipopolysaccharide resulted in an increase of the alveoloarterial oxygen tension difference (Po²difference) for both anesthetics. The propofol–lipopolysaccharide group had a significantly higher Po²difference compared with the sevoflurane–lipopolysaccharide group after 6 h, whereas no differences were found in the PBS groups (data not shown).

Arterial carbon dioxide tension levels were higher in both lipopolysaccharide groups compared with the PBS groups. Paco²was significantly higher in the propofol–lipopolysaccharide group (56.6 ± 8.1 mmHg [7.5 kPa]) compared with the sevoflurane–lipopolysaccharide group (42.2 ± 7.1 mmHg [5.6 kPa]) after 6 h of lipopolysaccharide injury (table 2).

Mean arterial pressure decreased in all four study groups during the course of the experiment. There were no significant differences in mean arterial pressure between the four groups at any time (table 2).

The recovery of BALF was comparable in all study groups. Seventy percent of administered fluid was retrieved.

Total cell count in BALF was determined as a measure of effector cell recruitment. Cells in BALF of PBS animals were identified as alveolar macrophages, whereas 99.5% of the cells in lipopolysaccharide animals were neutrophils. Cell count increased significantly in both lipopolysaccharide groups compared with both control groups. The sevoflurane–lipopolysaccharide group showed a significantly lower total cell number compared with the propofol–lipopolysaccharide group (sevoflurane–lipopoly-saccharide, 14.94 ± 5.72 cells/10⁶/ml; propofol–lipopolysaccharide, 27.18 ± 7.75 cells/10⁶/ml; fig. 2). There were no significant differences between the PBS groups.

Albumin concentration in BALF, reflecting alveolocapillary permeability, was significantly lower in the sevoflurane-LPS group compared with the propofol–lipopolysaccharide group (sevoflurane–lipopolysaccharide, 4.9 ± 3.8 μg/ml; propofol–lipopolysaccharide, 10.4 ± 3.5 μg/ml; fig. 3). The alveolar protein content as a measure of accumulation of proteins upon inflammation was significantly higher in the lipopolysaccharide groups compared with the PBS groups. In addition, a significantly lower protein concentration was found in the sevoflurane–lipopolysaccharide group compared with the propofol–lipopolysaccharide group (sevoflurane–lipopolysaccharide, 1.39 ± 0.51 mg/ml; propofol–lipopolysaccharide, 2.26 ± 0.32 mg/ml; table 2).

The protein concentration of the chemokines CINC-1 and MCP-1 in BALF was assessed by ELISA. CINC-1 and MCP-1 level increased significantly in both lipopolysaccharide groups compared with both PBS groups. The sevoflurane–lipopolysaccharide group showed significantly lower levels of CINC-1 and MCP-1 compared with the propofol–lipopolysaccharide group (figs. 4A and B). In the sevoflurane–lipopolysaccharide group, CINC-1 and MCP-1 expression decreased 29% and 53%, respectively, compared with the propofol–lipopolysaccharide group.

The expression of messenger RNA (mRNA) of CINC-1 and MCP-1 was analyzed in total lung tissue by real-time PCR. Values were normalized to 18S and expressed relatively to controls (PBS groups). The mRNA expression in both lipopolysaccharide groups was significantly increased compared with both PBS groups. Again, the sevoflurane–lipopolysaccharide group showed significantly lower mRNA levels compared with the propofol-lipopolysaccharide group (figs. 5A and B): in the sevoflurane–lipopolysaccharide group, CINC-1 mRNA and MCP-1 mRNA expression decreased by 42% and 53%, respectively, compared with the propofol–lipopolysaccharide group.

As expected, intratracheal lipopolysaccharide resulted in a pulmonary edema with inflammatory cell recruitment (fig. 6). Quantification of the injury showed a significant increase of the lung injury score in both lipopolysaccharide groups compared with the PBS groups. However, there was no significant difference between the sevoflurane–lipopolysaccharide group and the propofol–lipopolysaccharide group (table 2).

Evaluation of surfactant protein B (SP-B) RNA expression in lung tissue revealed a decrease in the expression of SP-B in both lipopolysaccharide groups compared with controls. However, decrease of SP-B in the sevoflurane–lipopolysaccharide animals was less accentuated compared with propofol–lipopolysaccharide animals (fig. 7).

Furthermore, analysis of the phospholipid content in bronchoalveolar lavage revealed an increase in the expression of phospholipids in both lipopolysaccharide groups compared with controls. The propofol–lipopolysaccharide animals showed significantly higher phospholipid levels compared with the sevoflurane–lipopolysaccharide animals (fig. 8).

Because a significantly higher Pco²was observed in the lipopolysaccharide–propofol group after 6 h of injury, we analyzed the possible proinflammatory effect of hypercapnia on RPAECs and AECs with or without lipopolysaccharide stimulation. Carbon dioxide values of 7.5% or 10% did not seem to have an impact on the inflammatory reaction in RPAECs or AECs compared with 5% CO²(figs. 9A and B). Similarly, we analyzed the possible proinflammatory effects of propofol in 14% Cremophor. Because AECs are not in direct contact with the anesthetic, we used smaller concentrations of propofol for the in vitro  approach. No proinflammatory effects were shown in nonstimulated RPAECs or AECs. Stimulation with lipopolysaccharide in the presence of propofol resulted in the same increase of CINC-1 and MCP-1 levels as observed in the lipopolysaccharide group (figs. 9C and D).

The current study demonstrates that anesthetic postconditioning with sevoflurane improves oxygenation and attenuates lung damage as indicated by less recruitment of effector cells into the respiratory compartment, decreases expression of the proinflammatory mediators CINC-1 and MCP-1, and reduces lung hyperpermeability in an in vivo  model of lipopolysaccharide-induced lung injury.

These results corroborate our previous in vitro  studies, where we showed a significant reduction of proinflammatory mediators by preconditioning7and by postconditioning10of AECs with sevoflurane in in vitro  models of lipopolysaccharide-induced injury. This is the first in vivo  study comparing the postconditioning effects of sevoflurane and propofol in a model of ALI.

First, we focused on the effect of both anesthetics on oxygenation capability of the lung. The significant improvement of Pao²/Fio²by postconditioning with sevoflurane after 6 h is most likely due to a less impaired gas exchange compared with propofol sedation. This was also reflected in the calculations of alveoloarterial oxygen tension difference. As discussed below, the reason for this seems to be an attenuation of lung damage after lipopolysaccharide challenge. To our knowledge, the amelioration of Pao²by postconditioning with a volatile anesthetic in an in vivo  model of ALI has not yet been described in the literature.

A possible explanation for the deteriorated Pao²/Fio²ratio could be an inhibition of the hypoxic pulmonary vasoconstriction (HPV) by both anesthetics. Clinical investigations are not conclusive regarding the possible effect of anesthetics on HPV. In animals models, volatile anesthetics seem to inhibit HPV, and increase intrapulmonary shunt fraction or reduce arterial oxygen tension in a dose–response manner,15,23,24whereas propofol does not affect HPV.25In the clinical scenario, however, in patients undergoing one-lung ventilation, sevoflurane and propofol have been shown to have similar effects on shunt fraction and arterial oxygen tension.26,27In our model, the impact of the volatile anesthetic–induced inhibition of HPV cannot be excluded.

Second, the expression of CINC-1 and MCP-1 was studied. These chemoattractants have been shown to possess potent chemotactic activity for neutrophils (CINC-1 and MCP-1) and monocytes (MCP-1) and therefore play a significant role in the acute inflammatory response in ALI.28–31The decrease of CINC-1 and MCP-1 proteins in bronchoalveolar lavage and of the mRNA in lung tissue by postconditioning with sevoflurane on the molecular level suggests a functional attenuation of inflammation by reduction of effector cell recruitment. In fact, we were able to prove this reduction of effector cells in the BALF (total cell count).

Third, alveolar albumin and protein influx as markers of increased influx of inflammatory proteins and alveolocapillary leakage, respectively, were evaluated. Lung hyperpermeability causing pulmonary edema is thought to be a main mechanism inducing ARDS.5,32Again, postconditioning with sevoflurane significantly decreased albumin and protein influx. Recently, it was shown that reduction of lung hyperpermeability protects against lipopolysaccharide-induced lung injury.33Therefore, the therapeutic effects of sevoflurane on ALI could be mediated by reduction of lung hyperpermeability.

Fourth, SP-B RNA expression in lung tissue was significantly less decreased upon lipopolysaccharide injury in the sevoflurane group compared with the propofol group, indicating a milder degree of injury. SP-B plays a critical role for maintenance of stability of surfactant. As shown in previous experimental approaches, expression of SP-B is decreased upon injury, probably as a consequence of destruction of the alveolocapillary unit with alveolar epithelial type II cells.34,35 

Fifth, lipopolysaccharide–propofol animals showed a significantly higher expression of phospholipids in BALF. We hypothesize that increases in phospholipids in the alveolar space could be due to decreases in surfactant clearance by type II cells and the cells resident in the alveolar space. Summarized, both results regarding SP-B and phospholipids underline a less deteriorated surfactant function by postconditioning with sevoflurane compared with propofol after lipopolysaccharide challenge.36 

Up to now, several in vivo  studies have explored the effects of sevoflurane on lung tissue but with inconsistent results. Takala et al.  37compared sevoflurane anesthesia with thiopentone anesthesia in a model of ventilated healthy pigs. It was demonstrated that AEC type II cell integrity and ultrastructure remained unchanged after long-term (6-h) high-concentration exposure to sevoflurane (1.5 minimum alveolar concentration). Furthermore, a lower gene expression of tumor necrosis factor-α and interleukin-1β was detectable in the intact porcine lung tissue after sevoflurane anesthesia.38On the other hand, an increase of pulmonary inflammatory mediators and pulmonary NO³and NO²production after sevoflurane anesthesia was revealed by another study using the same model.39However, this study was not based on an ALI model. In addition, the sevoflurane concentration of 4 vol% was rather high compared with our model.

To exclude a proinflammatory effect of propofol dissolved in Cremophor on pulmonary cells, we performed in vitro  experiments. RPAECs were coexposed to propofol in concentrations previously reported.40,41No increased cytotoxicity or enhanced inflammatory response could be observed. In addition, it should be mentioned that several studies have pointed out a protective effect of propofol as well.42–44 

Another component, which theoretically could enhance inflammatory injury, is the increased content of carbon dioxide after 6 h of injury. We discussed this increase as a consequence of injury. In vitro  experiments underlined our hypothesis by showing that increased concentrations of carbon dioxide did not interfere with the inflammatory reaction. This is in accord with the literature, where only carbon dioxide values of 15% or 20% induced an additional injury.45 

Few reports exist focusing on the postconditioning capabilities of sevoflurane in acute lung injury. In a recent publication, Hofstetter et al.  46examined the antiinflammatory effects of sevoflurane in an in vivo  model of lipopolysaccharide-induced endotoxemia in rats. In this study, administration of sevoflurane 15 min after stimulation with lipopolysaccharide resulted in a decrease of tumor necrosis factor-α and interleukin-1β plasma levels. In contrast to our study, lipopolysaccharide was given intravenously with an early administration of sevoflurane after the injury. In the current study, we were able to show antiinflammatory effects of sevoflurane even when administered 2 h after a lipopolysaccharide stimulation, i.e. , with late initiation of postconditioning. This may be of clinical relevance for patients who have already experienced a trigger event that may result in ALI, or even ARDS in that sevoflurane may beneficially interfere with the further development of the lung injury.

In this study, we focused on the difference between the intravenous anesthetic propofol and the volatile anesthetic sevoflurane. However, it remains questionable whether the observed difference would also be found with other intravenous anesthetics. Interestingly, in cardiac ischemia–reperfusion injury, protection by volatile anesthetics, morphine, and propofol is relatively well investigated.47It is generally agreed that these agents reduce the myocardial damage caused by ischemia and reperfusion. Other anesthetics, which are often used in clinical practice, such as fentanyl, ketamine, barbiturates, and benzodiazepines, have been much less studied, and their potential as cardioprotectors is currently unknown. Therefore, general conclusions should not be drawn.

Today, sedation of patients with ALI/ARDS in the ICU is commonly performed using propofol. In the last years, the antiinflammatory effects of this intravenous anesthetic have been extensively studied in several in vivo  studies. It has been shown that propofol has antiinflammatory effects that attenuate cytokine response after endotoxin shock in rats.48,49Several studies suggest that pretreatment and posttreatment with propofol provides protective effects in endotoxin-induced ALI19,50and lipopolysaccharide-induced shock.51However, the antiinflammatory effects of propofol are thought to be at least in part due to containing EDTA, which is a component of the commercially used propofol formulation.52In our study, we used a propofol formulation in 14% Cremophor without EDTA as clinically used propofol would induce hypervolemia in rats because of the low concentration of propofol. This could explain why fewer antiinflammatory effects in the propofol groups were found. However, a recent clinical trial comparing the antiinflammatory property of sevoflurane and propofol in patients undergoing thoracic surgery with one-lung ventilation has also shown less inflammatory response in the sevoflurane group, even in the presence of EDTA.53 

Since the AnaConDa was approved for the use in ICUs, it is now possible to take advantage of the properties of volatile anesthetics, such as fast induction, fast awakening, and easy titration, for sedation of postoperative and critically ill patients. Few studies have assessed the use of volatile anesthetics, especially sevoflurane, via  AnaConDa in ICU patients so far.6,54Recently, a significantly shorter recovery time and a significantly shorter hospitalization time with sevoflurane sedation compared with propofol was demonstrated in patients after cardiothoracic surgery.54Up to now, there have been no clinical studies regarding the effects of sevoflurane sedation in patients with ALI or ARDS. The results of this in vivo  study indicate that sevoflurane sedation of patients with ALI may be beneficial.

Our study has several limitations. First, as already discussed, we used a special formulation of propofol in 14% Cremophor without EDTA, which is not commonly used in the ICU. This could be a reason for the reduced immune response in the propofol group. In addition, findings of this study could be specific to this animal model. However, the lipopolysaccharide injection model has recently been evaluated to promise the most direct clinical relevance considering gram-negative sepsis in which ALI is most common.55,56Second, our observations are based on a model of a beginning ALI and therefore may not be applicable in already established ARDS. Moreover, we studied the effect of sevoflurane only during a very short period (6 h) compared with the clinical situation. In addition, we administered an Fio²of 1.0 in our model, which is not commonly used in ICUs except for severe cases of ARDS. To our knowledge, nothing is known about any interaction of hyperoxia and sevoflurane that may influence the antiinflammatory effects of sevoflurane. According to the literature, hyperoxia-induced toxic effects on cells appear only after exposure times of more than 12 h.57However, we cannot exclude that hyperoxia influences the antiinflammatory effects of sevoflurane in our model.

Despite these limitations, this study might be of clinical relevance. We could show that in developing ARDS, gas exchange deteriorates significantly less by just using sevoflurane as a sedative compared with propofol. This property of sevoflurane seems to be mediated by inhibition of lung inflammation as indicated by lower levels of cytokines and less recruitment of effector cells into the lung tissue. Sedating ICU patients with sevoflurane using the AnaConDa system might therefore be a promising new therapeutic approach for ALI and ARDS. Moreover, the application of sevoflurane can be easily combined with protective ventilation strategies, generating further interesting treatment options.

In conclusion, the current study indicates that anesthetic postconditioning with sevoflurane offers beneficial properties compared with propofol in a model of ALI in vivo .