Secondary Logo

Journal Logo

Cardiovascular Anesthesia: Research Report

The Pulmonary Immune Effects of Mechanical Ventilation in Patients Undergoing Thoracic Surgery

Schilling, Thomas MD, DEAA*; Kozian, Alf MD*; Huth, Christof MD; Bühling, Frank MD; Kretzschmar, Moritz cand med*; Welte, Tobias MD§; Hachenberg, Thomas MD, PhD*

Author Information
doi: 10.1213/01.ane.0000172112.02902.77
  • Free
  • Chinese Language Editions

Mechanical ventilation with a large tidal volume (Vt) may generate an inflammatory response in the alveoli. This ventilator-induced lung injury is characterized by cytokine production, leukocyte recruitment and neutrophil-dependent tissue destruction (1–3).

Different mechanisms may lead to cellular activation and mediator release. These include mechanical forces applied to lung tissue during mechanical ventilation that may result in alveolar cell stretch and overdistension, shear forces secondary to repeated tidal collapse, and reopening of alveolar units and increased vascular shear stresses (4–7).

One-lung ventilation (OLV) is an established procedure during thoracic surgery. Commonly, Vt used during two-lung ventilation (TLV) in the range of 10 mL/kg and zero end-expiratory pressure are recommended during OLV to maintain arterial oxygenation and carbon dioxide elimination (8).

However, the capacity of the aerated lung during OLV is greatly reduced compared with TLV. Therefore, increased mechanical strains of the dependent lung produced by increased mean airway pressures, followed by compression of alveolar vessels and increased pulmonary vascular resistance (PVR) may induce proinflammatory effects in the ventilated alveoli. The use of smaller tidal volumes with lower inspiratory pressures (volume- and pressure-limited ventilation) is favorable in patients undergoing thoracotomy (9). Furthermore, experimental data from an isolated rabbit lung model suggest that protective OLV including Vt and positive end-expiratory pressure (PEEP) set to avoid lung collapse and over-distension in ventilated lungs is able to minimize ventilator-induced lung injury (10).

Previous experimental and clinical studies on TLV (11,12) showed a progressive alteration of pulmonary immune function during anesthesia and surgery. Accordingly, in surgical patients undergoing OLV with high and low Vt, a time-dependent increase of proinflammatory variables was found; however, concentrations of inflammatory mediators in tracheal aspirate were not different between the two ventilator settings and neither time course nor concentrations of pulmonary or systemic mediators differed between patient groups (13).

In this prospective, randomized, clinical study, we therefore examined whether a standard ventilation setting (VT = 10 mL/kg) may result in a time-dependent alteration of pulmonary immune function in patients undergoing open thoracic surgery and OLV. Furthermore, we tested the hypothesis that reduction of Vt (5 mL/kg) would decrease ventilation-associated lung inflammation during and after OLV.


Thirty-two patients (ASA physical status II-III) scheduled for elective open thoracic surgery were included in the study (Table 1). Approval for the study was obtained from the IRB of the Otto-von-Guericke-University Magdeburg. Written informed consent for enrollment in the trial was obtained from each patient.

Table 1
Table 1:
Patient Characteristics and Surgical Criteria

Preoperative evaluation included a complete history, physical examination, measurements of actual weight and height, electrocardiogram, chest roentgenogram, pulmonary function tests (vital capacity [VC] and forced expiratory volume in 1 s, FEV1, expressed as percentage of VC), echocardiogram, and blood gas tension analysis.

Exclusion criteria were decompensated cardiac (>New York Heart Association II) or pulmonary diseases (VC or FEV1<50% of the predicted values), pulmonary hypertension (mean pulmonary artery pressure [MPAP] >30 mm Hg) and pre-existing coagulation disorders. Patients treated with immune modulators (cytostatic drugs, corticosteroids and nonsteroidal antiinflammatory drugs, vaccination, blood products) within 3 mo before surgery were not enrolled into the study. Patients with current nicotine abuse and with symptoms of an acute inflammatory process (clinically defined or abnormal data for C-reactive protein, leukocyte count, or body temperature) were also excluded.

Open thoracic surgical procedures were done for established or suspected malignancies, respectively (carcinomas, metastases). Patients with infectious processes were excluded. Lung resections were performed through a standard posterolateral or an anterolateral muscle-sparing thoracotomy.

Two hours preoperatively, each patient received 0.15 mg/kg diazepam orally for premedication. Before anesthesia, a thoracic epidural catheter was inserted at the level of the thoracic segments T4-5 to T6-7. Epidural catheterization was performed by loss of resistance technique via the median approach. The catheter was advanced cranially 2-3 cm into the epidural space and a test dose (3 mL bupivacaine 0.5% with epinephrine, 5 μg/mL) was given to exclude intrathecal or intravascular position of the catheter.

General anesthesia was induced IV with propofol (1-1.5 mg/kg), cisatracurium (0.1 mg/kg), and continuous infusion of remifentanil (0.3 μg·kg−1·min−1). Anesthesia was maintained by continuous infusions of propofol (4.0 ± 1.2 mg·kg−1·h−1), remifentanil (0.2 ± 0.08 μg·kg−1·min−1), and cisatracurium (2 μg·kg−1·min−1). For prophylactic antibiosis, a single infusion of cefotiam (2 g) was administered.

After tracheal intubation with a left- or right-sided standard double-lumen tube (DLT) (Broncho-Cath® 39F or 41F; Mallinckrodt Medical Ltd., Dublin, Ireland), the patients’ lungs were mechanically ventilated with intermittent positive pressure ventilation with Fio2 of approximately 0.45 in air, PEEP 3 mm Hg. Mechanical ventilation was provided with an anesthesia ventilator connected to a circle system (Cicero®; Dräger, Lübeck, Germany). Gas flow and airway pressure were measured at the proximal end of the endotracheal tube with a standard monitor for ventilatory measurements (Capnomac-Ultima®; Datex-Ohmeda, Helsinki, Finland)

All patients received an arterial catheter for continuous arterial blood pressure measurements and arterial blood sampling (20-gauge Angiocath; Becton Dickinson, Franklin Lakes, NJ) and a pulmonary arterial catheter (Swan-Ganz thermodilution catheter; B Braun, Melsungen, Germany) for cardiac output measurements (Hewlett Packard monitor M1106C; Palo Alto, CA) and sampling of mixed venous blood. After sampling, blood gas tension analysis was performed immediately with standard blood gas electrodes (ABL; Radiometer, Copenhagen, Denmark).

After surgery, all patients were admitted to the intensive care unit (ICU) and monitored for at least 24 h. During postoperative stabilization, fluids and blood transfusions were given in order to maintain central venous pressure (CVP) ≥3 cm H2O, urine output ≥1 mL·kg−1·h−1, and hemoglobin concentration ≥6.0 mmol/L−1. A chest radiograph examination was done in the ICU before tracheal extubation. After tracheal extubation, supplemental oxygen was given in order to keep peripheral oxyhemoglobin saturation >95%.

After OLV, pain therapy was started intraoperatively by bolus injection of ropivacaine 0.375% (8–10 mL) with sufentanil (1 μg/mL) and was continued by infusion of 4–6 mL/h ropivacaine 0.2% with sufentanil (1 μg/mL) via the thoracic epidural catheter for 2–4 days.

After induction of anesthesia and tracheal intubation with a standard DLT, patients were randomly assigned to two groups (random numbers generated by the Excel® computer program, Microsoft, Redmond, WA): group A (n = 16) received TLV with a Vt = 10 mL/kg; group B (n = 16) received TLV with reduced Vt = 5 mL/kg. Respiratory frequencies were adjusted to achieve a normal arterial Paco2.

During OLV, the nondependent lung collapsed completely, and ventilatory settings were as follows: Vt remained unchanged at 10 mL/kg and 5 mL/kg, respectively. Respiratory frequencies were adjusted to achieve a Paco2 of 35–45 mm Hg. Fio2 was set to 0.8–1.0 to achieve a Pao 2 >80 mm Hg. PEEP was set to zero during OLV while avoiding peak inspiratory pressures >30 mm Hg and gas trapping at end-expiration. In three patients of group A the threshold of 30 mm Hg was reached. In these cases the Vt was reduced by 1 mL/kg. End-expiratory PEEP records (0–2 mm Hg) revealed no differences between both groups.

For the last bronchoalveolar lavage (BAL) and postoperative ventilatory support the endotracheal tubes were changed to standard single-lumen tubes (inner diameter, 8.5 or 9 mm, Mallinckrodt Medical Ltd.). All patients were tracheally extubated 1-3 h after bronchoscopy.

Cardiopulmonary data were evaluated at three stages: (A) during TLV before thoracic surgery, (B) during OLV, and (C) postoperatively. Measurement of cardiac output was performed by a Swan-Ganz thermodilution catheter. Furthermore, the following cardiorespiratory variables were recorded: heart rate, mean arterial blood pressure (MAP), MPAP, CVP, and pulmonary artery occlusion pressure (PAOP), as well as arterial and mixed venous blood gas tensions. Data were documented continuously and systemic vascular resistance, PVR, oxygen delivery index (DO2i), oxygen consumption index (VO2i) and shunt fraction were calculated according to standard formulas.

BAL was based on a standardized procedure. Before surgery, the correct position of the DLT was confirmed by fiberoptic bronchoscopy (models BF-P40 and BF3-C40; Olympus, Melville, NY), and first BAL was performed through the DLT 30 min after tracheal intubation. BAL of the dependent, ventilated lung was repeated after OLV immediately at the end of the surgical procedures and 2 h postoperatively.

The tip of the bronchoscope was brought into wedge position in a segmental bronchus of the left-sided lower lobe or the right middle lobe. A different, randomly chosen, segment was lavaged each time. For BAL, 80 mL of 0.9% saline solution was sequentially instilled and suctioned in 20-mL portions; 50%–60% of which was recovered. The volume of return of lavage fluid did not differ between patient groups.

Lavage fluids were filtered through sterile gauze filters, collected on ice, and immediately centrifuged at 200g for 10 min. Supernatant aliquots were kept frozen at −80°C for subsequent analysis. The cell pellets were resuspended in ice-cold phosphate buffer with 0.01% sodium azide and 2% bovine serum for staining and counting.

Concentrations of interleukin (IL)-8, IL-10, and soluble intercellular adhesion molecule (sICAM)-1 in the BAL fluids were determined by commercially available quantitative sandwich enzyme immunoassays (Quantikine®; R&D Systems Ltd., Abingdon, UK). Tumor necrosis factor (TNF) α and polymorphonuclear (PMN) cell elastase immunoassays were provided by Immunotech, France and Milenia Biotech, Germany, respectively. Protein concentrations were measured by an assay for the colorimetric detection and quantitation of total protein (Micro BCA ™ Protein Assay Reagent Kit; Pierce, Rockford, IL).

All samples of one patient were analyzed in the same assay run. The samples were measured in duplicates and the assays were performed according to the manufacturer’s instructions. The optical density of the samples was determined by a microplate reader (Safire®; Tecan Ltd., Salzburg, Austria) and analyzed using the Safire microplate reader software by interpolation from standard curves. The sensitivities of the test kits were as follows: IL-8: 3.5 pg/mL, IL-10: 0.5 pg/mL, sICAM-1: 0.35 ng/mL, PMN elastase: 3 ng/mL, TNFα: 5 pg/mL, protein: 0.5 μg/mL.

Albumin concentrations were estimated by a Nephelometer (BN 2000; Dade/Behring, Liederbach, Germany). Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS Inc., Chicago, IL). Data are presented as mean ± sd in the case of normal distribution (cardiopulmonary and ventilation variables) and are expressed as median and interquartile range in the case of non-normal distribution. Based on previous studies (14,15) the hypothesis was tested that after OLV cell numbers, IL-8, or PMN elastase would increase by more than 25% with a power of 0.8 and a two-tailed P < 0.05. This required a sample size of 12 to 14 subjects in each group.

The analysis of normally distributed data was performed by a repeated-measures one-way analysis of variance with post hoc Bonferroni correction. Non-normally distributed data were logarithmically transformed to achieve homogenous variances of data sets; however, cytokine concentrations in BAL fluids still differed from normal distribution even after transformation. These data were analyzed using Kruskal-Wallis H-Test with adjustment of α-level for repeated measurements.

In some cases, mediator concentrations were below the detection limits of the assays. These data were entered into the statistical analysis with a value of 0.01. A P value of <0.05 was considered significant for all statistical procedures.


Patient characteristics and details of thoracic surgery are shown in Table 1. Surgical admission criteria included lobectomy (13 of 32), pneumonectomy (3 of 32), and atypical pulmonary resection (16 of 32). There were no significant differences of demographic data between the patient groups. Duration of OLV in relation to time of surgery was also not different between groups (median duration of OLV, 70 min; range, 20-119 min).

The administration of buffy-coat free erythrocyte concentrates (EC) during the study period is documented in Table 1. Nine of 32 patients in both groups received 1 or 2 EC. One patient of group A received 4 EC during and after thoracic surgery. There were no differences in the use of blood products between groups. Furthermore, there were no correlations between immune variables and administration of erythrocytes.

Mean values for hemodynamic data, ventilation, and gas exchange variables are presented in Tables 2 and 3. In all patients, an increase of MPAP, PAOP, CVP, and shunt (Qs/Qt) was observed; however, differences between the groups were not significant. MAP and cardiac index remained unchanged.

Table 2
Table 2:
Ventilation and Gas Exchange Data at the Beginning of the Study (TLV, Preoperatively), During OLV (After 30 min) and at TLV Postoperatively
Table 3
Table 3:
Hemodynamic Data at the Beginning of the Study (TLV, Preoperatively), During OLV (After 30 min) and at TLV Postoperatively

Comparison of patients ventilated with either small or large Vt revealed significant differences in peak and plateau airway pressures (PAW peak, PAW plateau) but not in Pao2. During mechanical ventilation with Vt 5 mL/kg, a more rapid ventilation rate was required to achieve the desired Paco2 range (A: f = 8-10 Vt/min; B: f = 14-16 Vt/min), whereas minute volume ventilation and inspiratory/cycle time ratio were not different between groups. However, Paco2 was significantly higher in patients receiving OLV with low Vt, representing higher dead space ventilation.

In both large and small Vt groups, the number of intra-alveolar cells, protein and albumin concentrations increased during mechanical ventilation (Fig. 1). In the small Vt group, the increase of cells as well as of protein and albumin concentrations in the BAL fluid was significant during and after mechanical ventilation.

Figure 1.
Figure 1.:
Time course of intra-alveolar cell numbers, polymorphonuclear neutrophils (PMN) elastase, protein, and albumin concentrations in patients undergoing thoracic surgery and one-lung ventilation (OLV). Data are presented as ranges, medians, and interquartile ranges; the symbols * and # indicate significant differences within a group, § indicates differences between groups, P < 0.05). Vt, tidal volume.

In all patients, an increase of the alveolar concentrations of PMN elastase (Fig. 1), IL-8, and TNF-α was observed during and after OLV (Fig. 2). However, in contrast to patients ventilated with Vt 10 mL/kg, mechanical ventilation with reduced Vt resulted in a significant decrease of alveolar TNF-α concentrations after OLV.

Figure 2.
Figure 2.:
Time course of the alveolar inflammatory markers interleukin (IL)-8, tumor necrosis factor (TNF)-α, soluble intercellular adhesion molecule (sICAM)-1 and IL-10 before and after one-lung ventilation (OLV) and 2 h postoperatively. Mediator concentrations in the bronchoalveolar lavage (BAL) fluid are presented as ranges, medians and interquartile ranges, P25-P75; the symbols * and # indicate significant differences within a group, § indicates differences between groups, P < 0.05. Vt, tidal volume

In contrast to PMN elastase, IL-8, and TNF-α, sICAM-1 alveolar concentrations were significantly decreased in the small Vt group during ventilation (Fig. 2).

Intra-alveolar IL-10 concentrations decreased significantly during ventilation with Vt 10 mL/kg but remained unchanged in the small Vt group (Fig. 2).

The time courses of intra-alveolar cell numbers, protein and albumin concentrations as well as IL-8, PMN elastase, IL-10, with the exception of TNF-α, did not differ significantly between the groups after OLV and 2 h postoperatively. Intra-alveolar sICAM-1 concentrations were significantly smaller after ventilation with Vt = 5 mL/kg.

Nevertheless, 30 min of TLV with Vt 5 mL/kg in contrast to 10 mL/kg before thoracic surgery resulted in decreased alveolar concentrations of protein, albumin, sICAM-1 and IL-10. However, 2 h postoperatively there were no significant differences between groups.


The data of our study suggest that OLV initiates a proinflammatory response in the alveolar compartment of the dependent lung. Intra-alveolar concentrations of IL-8, TNF-α, PMN elastase, protein, and albumin, as well as cell numbers in the BAL fluids increased during and after OLV, and antiinflammatory IL-10 decreased. Ventilation with Vt 5 mL/kg in comparison with Vt 10 mL/kg significantly decreased alveolar TNF-α and sICAM-1 and increased IL-10 after OLV. However, 2 hours postoperatively there were no differences between groups; only sICAM-1 showed a consistent trend to decrease after mechanical ventilation with Vt 5 mL/kg.

Cytokines are involved in different pathologic states of the lung (16). TNF-α is a polypeptide cytokine that is principally produced by macrophages/monocytes and commonly associated with critical inflammatory conditions(17). Experimental data suggest that alveolar macrophages produce less IL-1 but more TNF-α than plasma monocytes (18). This may explain the unchanged TNF-α plasma concentrations that were found during conventional mechanical ventilation in anesthetized patients (19). In the present study TNF-α in BAL fluid was assessed, reflecting TNF-α secretion of alveolar macrophages.

IL-8 is one of the most important cytokines responsible for the recruitment of inflammatory cells to the alveoli. It is increased in the BAL fluid of patients with acute respiratory distress syndrome, sepsis, and multiorgan failure (20), and IL-8 concentrations have been correlated to the degree of pulmonary dysfunction (21). The cytokine is a powerful chemotactic factor for PMNs and stimulates adherence of PMN to pulmonary epithelial and endothelial cells, producing a discrete granulocytosis. Besides their beneficial antimicrobial effects, neutrophils liberate potentially toxic products, such as PMN elastase, leukotrienes, and free radicals, into the alveolar space. These mediators can produce irreversible damage to the host tissue (22). In the present study we found an increased alveolar IL-8 concentration that corresponded to the increased amounts of cells in the BAL fluid. This was not influenced by the ventilation regimen. In addition, alveolar epithelial cells express molecules, such as ICAM-1, that interact with leukocytes and thus promote their transepithelial migration and activation (23). Interestingly, the expression of sICAM-1 was significantly less after ventilation with Vt 5 mL/kg. This may suggest that ventilation with reduced Vt can partially prevent epithelial cell activation. Antiinflammatory cytokines such as IL-10 are expressed in response to cell activation and function directly to diminish the severity of lung injury by modulating the expression of proinflammatory cytokines (24). We found a significant suppression of the IL-10 secretion after ventilation with large Vt but not after ventilation with small Vt, advocating an improved antiinflammatory immunoregulation in the small Vt group.

Several studies suggest that hyperinflation of the lungs may lead to the development of ventilator-induced lung injury and may promote proinflammatory conditions (3,5,7). Additional forces result from increased pulmonary transmural capillary pressure (25). In animals subjected to transmural pulmonary capillary pressures exceeding 24 mm Hg, stress failure of the pulmonary blood-gas barrier was observed, including disruptions of the capillary endothelial and alveolar epithelial layers (26). Experimental data suggest that capillary pressure is about halfway between MPAP and left atrial pressure (27). This means, in our patients, that at midlung level the pulmonary capillary pressure would be approximately 16 mm Hg. However, in the lateral decubitus position the capillaries of the dependent lung are localized 15–20 cm lower. This leads to an additional hydrostatic gradient and to altered distribution of pulmonary perfusion resulting in a capillary pressure of more than 24 mm Hg (28). An increased pulmonary capillary pressure may contribute to the failure of the blood-gas-barrier followed by increased albumin concentrations in the BAL fluids. In this context, it is remarkable that MPAP and PAOP increased to the same degree during OLV in both groups. This may explain why alveolar albumin concentrations did not differ between patients ventilated with large or small Vt.

In all patients, virtually the same plateau airway pressures were recorded during OLV; whereas peak pressures were significantly different between groups. Airway pressures measured at the Y-piece of the tube are determined by different factors, including inspiratory flow rate, airway resistance, and compliance of lung and chest wall gas flow. In fact, the point of greatest resistance and least compliance in the entire ventilator-patient circuit is the tracheal tube. As a consequence, airway pressure measurements coinciding with high inspiratory gas flows provide little information concerning alveolar pressures. For a given inspiratory flow pattern, comparable plateau pressures may be observed during different Vt, as resistance and compliance of the DLT are constant.

Studies in patients undergoing major abdominal and thoracic surgery have shown that short-term TLV or OLV did not induce a systemic inflammatory response (19). Furthermore, Wrigge et al. (13) analyzed plasma and pulmonary cytokine concentrations during ventilation with a protective setting compared with larger Vt. No significant differences of pulmonary and systemic mediators between the two ventilatory settings were found. However, pulmonary mediators were determined in tracheal aspirates rather than in BAL fluids obtained under standardized conditions. In the present study alveolar concentrations of TNF-α and sICAM-1 were significantly smaller in patients ventilated with a Vt 5 mL/kg. Patients undergoing extended pulmonary resections or pneumonectomies are at risk for endothelial injury in the non-operated lung associated with high-protein pulmonary edema (29). Possibly, OLV with increased intrapulmonary pressures contributes to the development of postpneumonectomy pulmonary edema.

It should be noted that outcome variables were not specified and recorded in the present study. Additional limitations may include the small sample size and lack of true blinding. Because of the design of the study, the patients were ventilated with a Vt 5 mL/kg or 10 mL/kg immediately after tracheal intubation. Therefore, a control BAL of all patients during TLV with Vt 10 mL/kg was not obtained to limit the number of BAL procedures that may have induced pulmonary inflammatory effects. Specifically, it can not be excluded that the increase in cells before lung surgery is partially influenced by the first BAL procedure.

In summary, OLV may promote production and release of proinflammatory substances in the alveoli of the dependent lung. Reductions of Vt and subsequently decreased peak airway pressures have significant effects on alveolar TNF, sICAM-1 and IL-10 concentrations after OLV and in the postoperative course. Whether lung protective ventilation approaches, such as pressure-limited ventilation with sufficient PEEP and a decelerating flow pattern, further reduce lung damage during OLV remains to be studied.


1. Broccard AF, Hotchkiss JR, Suzuki S, et al. Effects of mean airway pressure and tidal excursion on lung injury induced by mechanical ventilation in an isolated perfused rabbit lung model. Crit Care Med 1999;27:1533–41.
2. Slutsky AS. Lung injury caused by mechanical ventilation. Chest 1999;116:9–15.
3. Dreyfuss D, Ricard JD, Saumon G. On the physiologic and clinical relevance of lung-borne cytokines during ventilator-induced lung injury. Am J Respir Crit Care Med 2003;167:1467–71.
4. Tremblay L, Valenza F, Ribeiro SP, et al. Injurious ventilatory strategies increase cytokines and c-fos m-RNA expression in an isolated rat lung model. J Clin Invest 1997;99:944–52.
5. Dreyfuss D, Saumon G. Ventilator-induced lung injury: lessons from experimental studies. Am J Respir Crit Care Med 1998;157:294–323.
6. Tremblay LN, Miatto D, Hamid Q, et al. Injurious ventilation induces widespread pulmonary epithelial expression of tumor necrosis factor-alpha and interleukin-6 messenger RNA. Crit Care Med 2002;30:1693–700.
7. Pugin J. Molecular mechanisms of lung cell activation induced by cyclic stretch. Crit Care Med 2003;31:S200–6.
8. Brodsky JB, Fitzmaurice B. Modern anesthetic techniques for thoracic operations. World J Surg 2001;25:162–6.
9. Tugrul M, Camci E, Karadeniz H, et al. Comparison of volume controlled with pressure controlled ventilation during one-lung anaesthesia. Br J Anaesth 1997;79:306–10.
10. de Abreu MG, Heintz M, Heller A, et al. One-lung ventilation with high tidal volumes and zero positive end-expiratory pressure is injurious in the isolated rabbit lung model. Anesth Analg 2003;96:220–8.
11. Kotani N, Lin CY, Wang JS, et al. Loss of alveolar macrophages during anesthesia and operation in humans. Anesth Analg 1995;81:1255–62.
12. Kotani N, Hashimoto H, Sessler DI, et al. Intraoperative modulation of alveolar macrophage function during isoflurane and propofol anesthesia. Anesthesiology 1998;89:1125–32.
13. Wrigge H, Uhlig U, Zinserling J, et al. The effects of different ventilatory settings on pulmonary and systemic inflammatory responses during major surgery. Anesth Analg 2004;98:775–81.
14. Kotani N, Hashimoto H, Sessler DI, et al. Neutrophil number and interleukin-8 and elastase concentrations in bronchoalveolar lavage fluid correlate with decreased arterial oxygenation after cardiopulmonary bypass. Anesth Analg 2000;90:1046–51.
15. Frass OM, Buhling F, Tager M, et al. Antioxidant and antiprotease status in peripheral blood and BAL fluid after cardiopulmonary bypass. Chest 2001;120:1599–608.
16. Strieter RM, Belperio JA, Keane MP. Host innate defenses in the lung: the role of cytokines. Curr Opin Infect Dis 2003;16:193–8.
17. Maus U, Rosseau S, Knies U, et al. Expression of proinflammatory cytokines by flow-sorted alveolar macrophages in severe pneumonia. Eur Respir J 1998;11:534–41.
18. Rich EA, Panuska JR, Wallis RS, et al. Dyscoordinate expression of tumor necrosis factor-alpha by human blood monocytes and alveolar macrophages. Am Rev Respir Dis 1989;139:1010–6.
19. Wrigge H, Zinserling J, Stuber F, et al. Effects of mechanical ventilation on release of cytokines into systemic circulation in patients with normal pulmonary function. Anesthesiology 2000;93:1413–7.
20. Schutte H, Lohmeyer J, Rosseau S, et al. Bronchoalveolar and systemic cytokine profiles in patients with ARDS, severe pneumonia and cardiogenic pulmonary oedema. Eur Respir J 1996;9:1858–67.
21. Boutten A, Dehoux MS, Seta N, et al. Compartmentalized IL-8 and elastase release within the human lung in unilateral pneumonia. Am J Respir Crit Care Med 1996;153:336–42.
22. Kinoshita M, Ono S, Mochizuki H. Neutrophils mediate acute lung injury in rabbits: role of neutrophil elastase. Eur Surg Res 2000;32:337–46.
23. Lee JH, Del Sorbo, L, Uhlig S, et al. Intercellular adhesion molecule-1 mediates cellular cross-talk between parenchymal and immune cells after lipopolysaccharide neutralization. J Immunol 2004;172:608–16.
24. Farivar AS, Krishnadasan B, Naidu BV, et al. Endogenous interleukin-4 and interleukin-10 regulate experimental lung ischemia reperfusion injury. Ann Thorac Surg 2003;76:253–9.
25. Fu Z, Costello ML, Tsukimoto K, et al. High lung volume increases stress failure in pulmonary capillaries. J Appl Physiol 1992;73:123–33.
26. Tsukimoto K, Mathieu-Costello O, Prediletto R, et al. Ultrastructural appearances of pulmonary capillaries at high transmural pressures. J Appl Physiol 1991;71:573–82.
27. Bhattacharya J, Staub NC. Direct measurement of microvascular pressures in the isolated perfused dog lung. Science 1980;210:327–8.
28. West JB, Tsukimoto K, Mathieu-Costello O, Prediletto R. Stress failure in pulmonary capillaries. J Appl Physiol 1991;70:1731–42.
29. Licker M, Spiliopoulos A, Frey JG, et al. Risk factors for early mortality and major complications following pneumonectomy for non-small cell carcinoma of the lung. Chest 2002;121:1890–7.
© 2005 International Anesthesia Research Society