Mechanical ventilation is a supportive treatment allowing survival in patients with cerebral, neurological, muscular or lung diseases. Mechanical ventilation is most often defined as a mechanical inflation of gas in the lungs with a positive inflation pressure, a variable tidal volume (VT) and respiratory rate. In animals exact definitions for low, moderate and high VT do not exist. However, VT <10 mL kg−1 was used for low, and >20 mL kg−1 for high VT ; in comparison, spontaneously breathing healthy rats have a VT of 8 mL kg−1 . Cut-offs for low- or high-inflation pressures were never specified for animal studies, but pressures above 30 cmH2O may induce lung injuries with endothelial and epithelial alterations .
Mechanical ventilation in human and animal studies with high VT (i.e. ‘aggressive’ mechanical ventilation) producing lung damage related to alveolar overdistension or volutrauma is defined as ventilator-induced lung injury (VILI) . The application of high VT in mechanically ventilated rats over a period of several minutes has been shown to produce epithelial and endothelial disruption triggering increased vascular filtration pressures and pulmonary oedema [5,6]. A further mechanism responsible for VILI is the stretch injury related to repetitive collapse and reopening of atelectatic regions with each breath (atelectrauma). Stretch injury was observed most often in acute lung injury models [7,8]. Volutrauma and atelectrauma may induce pulmonary cell activation with neutrophil sequestration in the alveolar space and local inflammation .
Not only high VT with any positive pressure and stretch phenomena but also the duration of mechanical ventilation may injure lung tissue. The impact of mechanical ventilation on the lungs in a time-dependent manner has rarely been investigated in animals. This limited evidence may be due to the difficulty to examine separately the effects of mechanical ventilation and duration. The relationship may be analogous to the administration of a toxin: both single doses and repeated doses must be taken into account . Among the rare studies that have addressed this question in healthy animals the following results were observed: Healthy baboons, ventilated for several days with VT of 12-15 mL kg−1 showed signs of a mild inflammatory response in histological sections of the lungs [10,11]. In piglets ventilated with VT of 15 mL kg−1 for 3 days oxygenation decreased more than 20% . After prolonged mechanical ventilation oxygenation seems to be more impaired in smaller animals. In rats ventilated for 48 h with low to moderate VT (10 mL kg−1), there was a 40% decrease of oxygenation compared with spontaneous breathing animals . The application of a moderate VT (19 mL kg−1) in rats induced major lung injuries after mechanical ventilation of 7 h , the application of a low VT (7 mL kg−1) did not injure lungs after ventilation for 5 h .
The aim of this study was to evaluate the effects of the duration of mechanical ventilation with different VT on lung mechanics, gas exchange, pulmonary capillary leakage and broncho-alveolar lavage (BAL) fluid cell distribution in rats without pre-existing lung injury.
Materials and methods
The study protocol was approved by the Institutional Ethics Committee for animal research and experiments were conducted according to regulations of the local veterinary office.
Sprague-Dawley rats, 300 ± 3.2 g (mean ± SEM); were separated into five experimental groups (n = 12-15 rats per group) mechanically ventilated with different VT: 9 mL kg−1 (VT9); 18 mL kg−1 (VT18); 27 mL kg−1 (VT27); 36 mL kg−1 (VT36) and 45 mL kg−1 (VT45). Half of the animals from each group were ventilated for 1 h and half for 7 h. Spontaneously breathing rats were included to determine confounding effects of mechanical ventilation and tracheotomy: blood and BAL fluid samples were collected from spontaneously breathing rats immediately following anaesthesia induction (n = 3) or after 7 h of anaesthesia (n = 4), as well as from rats breathing spontaneously through a tracheotomy following 1 h (n = 5) or 7 h of respiration (n = 5).
Mechanically ventilated rats were anaesthetized with isoflurane (1.5-2.0 vol.%) and tracheotomized with a 14 G polyethylene cannula (Braun, Melsungen, Germany). Respiratory rate was 90 min−1 for Group VT9, 45 min−1 for VT18, 19 min−1 for VT27, 14 min−1 for VT36 and 10 min−1 for VT45, and was adjusted (in preliminary investigations) to maintain PaCO2 in the limits of 40 ± 8 mmHg with a constant volume-cycled rodent ventilator (model 683, Harvard Apparatus Corp, South Natick, MA; inspiratory time: expiratory time = 1 : 1, with a sinusoidal flow pattern). Positive end-expiratory pressure (PEEP) was 2.5 cmH2O. Inspiratory fresh gas flow was administered with calibrated rotameters at a flow of 0.6L min−1 containing oxygen and air to an inspired fraction of oxygen (FiO2) of 0.4. Inspiratory gases were measured with a Datex multiple gas analyser (Ultima™; Datex/Instrumentarium, Helsinki, Finland) and mixed in a 0.5 L capacity airbag with an open overflow valve before connection to the inspiratory port of the Harvard ventilator. The spontaneously breathing rats were placed in a box of Plexiglas and anaesthetized with isoflurane between 1.0 and 1.2 vol.% using an inspiratory flow of 1L min−1 (FiO2 of 0.4 in air).
Airway pressure (Paw) was continuously measured using a calibrated pressure transducer (Validyne DP 45; Northridge, CA); lung volume was determined with a pneumotachometer (Harvard Apparatus, model 3.0) connected to a pneumotachograph (Gould, Godart V, type 17212; Bilthoven, The Netherlands).
The femoral artery was cannulated for blood sampling and continuous arterial blood pressure monitoring. In mechanically ventilated animals (3 mL kg−1 h−1) 0.9% NaCl, 0.1 mg kg−1 h−1 morphine and 1 mg kg−1 pancuronium were administered. Rectal temperature was maintained at 37 ± 0.5°C with a heating pad. Variables were stored at a sampling rate of 50 Hz via an analogue/digital interface converter (Biopac; Santa Barbara, CA) on a microcomputer.
At the end of the experiment all rats were killed by an overdose of thiopental independent of the duration of ventilation.
Intact chest pressure-volume (P-V) curves were determined by inflating the lungs with air (0.3 mL s−1) to a peak Paw of 30 cmH2O from functional residual capacity (ambient Paw without PEEP) using a combined infusion/withdrawal pump. This method is an adaptation of the original description of P-V manoeuvre in rats with an inflation flow rate of 0.66 mL s−1 . Following a 1-2 min period of VT ventilation for haemodynamic recovery, a deflation P-V curve was obtained by withdrawing insufflated air from a Paw of 30 cmH2O at a constant rate of 0.3 mL s−1 back to ambient Paw. P-V curves were performed after 1, 3, 5 and 7 h in groups for 7 h and after 1 h in ventilated groups for 1 h. P-V curves were described with a sigmoid equation  including the determination of the inspiratory lower and upper inflection points on the inspiratory limb and collapse pressure (maximum curvature point) on the deflation limb. Inspiratory and expiratory total thoraco-abdominal compliance was determined arbitrarily at an inspiratory or expiratory Paw of 10 cmH2O (V10). Collapse pressure was defined as the inflection point on the expiratory curve, which divided the part with low compliance at high Paw from the part with high compliance at lower Paw.
Arterial blood was sampled for blood gas analysis, pH and haematocrit (Radiometer, model 505; Acid Base Laboratory, Copenhagen, Denmark) as well as for white blood cell count (Sysmex; Kobe, Japan) after 1, 3, 5 and 7 h in groups for 7 h and after 1 h in groups ventilated for 1 h.
A whole lung BAL was performed instilling 4 mL physiological saline into the trachea, followed by ventilation for 15 s, and then withdrawing two aliquots, one of 1 mL (tracheo-bronchial lavage fluid) and one of approximately 2 mL (BAL fluid). The BAL fluid was halved. One half was immediately centrifuged for 5 min (3000 cycles min−1) at 4°C and the clear supernatant stored at −70°C for protein analysis. The other half was centrifuged for 5 min at 550 cycles min−1 and the smear was stained with Papanicolau stain. Differential cell count was performed on 400 cells. Percentages of number of neutrophils to number of total cells per BAL were calculated . Total protein concentration (plasma and BAL fluid proteins) at the end of the experiment (or just before death, if applicable) was measured in the supernatant by spectrophotometry (Bio-Rad Protein Assay; Bio-Rad Laboratories, Hercules, CA). The alveolar-arterial protein concentration ratio between alveolar and arterial plasma protein concentration per group was calculated to estimate pulmonary permeability [19,20].
Data presentation and statistical analysis
All results are expressed as mean ± SEM. Data for the P-V curves was constructed by averaging volumes at each centimetre of Paw increments. For comparisons between the groups a one-way analysis of variance (ANOVA) with Dunnett's post hoc analysis was performed. Repeated measures ANOVA was used to analyse continuous variables over time. Statistical significance was accepted at the P < 0.05 level.
All animals survived the 7 h observation period, except animals in Groups VT36 and VT45 (mortality rates of 62.5% and 100%, respectively; Fig. 1) which died from circulatory failure as manifested by a progressive decrease in mean arterial pressure, bradycardia and metabolic acidosis associated with a high-permeability pulmonary oedema. After 7 h of mechanical ventilation, peak Paw, mean Paw and mean arterial pressure of surviving animals was not significantly different from the 1 h time point (Table 1).
After 1 h of mechanical ventilation inspiratory thoraco-abdominal compliance was similar for all ventilated groups (Fig. 2). The lower inflection point on the inspiratory limb was slightly but significantly higher in the VT27 and VT36 groups compared with the reference VT9 group (Fig. 3), but the upper inflection point on the inspiratory limb was comparable in all groups (between 12.0 ± 0.3 and 12.8 ± 0.2 cmH2O). Similarly, both inspiratory and expiratory volumes at an inspiratory Paw of 10 cmH2O (V10) were not significantly different between groups. Finally, P-V curve parameters obtained from animals sacrificed after the 1 h time point were not different from the ones scheduled to be investigated during the whole 7 h follow-up period (data not shown).
After 7 h of mechanical ventilation, the shape of the P-V curve was altered by prolonged mechanical ventilation with high VT (Fig. 2). The lower and upper inflection points on the inspiratory limb were significantly raised at 7 h in the three surviving animals of Group VT36, and already after 3 h of ventilation in animals of Group VT45 (Fig. 3). Inspiratory, and to a lesser extent expiratory volume at Paw of 10 cmH2O were profoundly decreased in Group VT45. Finally, there was a VT-dependent rightward shift of the collapse pressure on the deflation limb of the P-V curve in animals ventilated with the two highest VT.
Partial pressure of arterial oxygen (PaO2) remained unchanged over time in animals of Groups VT9, VT18 and VT27 (Fig. 4). In contrast, PaO2 decreased with time in Groups VT36 and VT45, being significantly reduced 3 h after mechanical ventilation in Group VT45 when compared with the other groups at this time point and after 7 h in surviving animals of Group VT36.
In animals breathing spontaneously, PaO2 remained stable (animals without a tracheotomy (Group C): 180 ± 2 mmHg (FiO2 of 0.4) at 1 h and no change over time; animals with a tracheotomy (Group T): 156 ± 9 mmHg at 1 h and no change over time).
Neutrophils in BAL fluid
White blood cell count and the percentage of neutrophils in blood were not different between the groups at 1 h of mechanical ventilation. In BAL fluid, the percentage of neutrophils after 1 h of ventilation was not significantly different between groups (Fig. 5a). Spontaneously breathing rats without tracheotomy (Group C) had a significantly lower percentage of neutrophils after 1 h. After 7 h, neutrophil accumulation in BAL fluid was increased similarly in all ventilated groups (>60% of total cells) compared with the percentage of neutrophils in BAL after 1 h (<40% of total cells) and compared with spontaneously breathing rats during anaesthesia of 7 h (15 ± 5% of total cells). There was a large variability of the absolute number of total white blood cells and neutrophils in BAL between the rats. Total cells ranged between 103 000 and 777 000 cells mL−1 after 1 h of mechanical ventilation or spontaneous respiration, and between 115 000 and 560 000 cells mL−1 after 7 h of mechanical ventilation or spontaneous respiration.
Alveolar-arterial protein concentration ratio
The alveolar-arterial protein ratio in all mechanically ventilated rats after 1 and 7 h of ventilation was significantly higher compared with spontaneously breathing animals without tracheotomy (Group C; Fig. 5b). There was a significant increase of this ratio measured before death in Groups VT36 and VT45, and in surviving animals of Group VT36 the alveolar-arterial protein ratio tended to be increased after 7 h of mechanical ventilation, though not reaching statistical significance. Furthermore, after 7 h of spontaneous respiration, animals with a tracheotomy (Group T) had a higher alveolar-arterial protein ratio than spontaneously breathing animals without tracheotomy (Group C).
The main results of the present study performed in healthy rats demonstrate the importance of the duration of mechanical ventilation on high volume VILI (VT > 27 mL kg−1), a pathological condition that was not detectable after 1 h of ‘aggressive’ mechanical ventilation, but which resulted in an increased mortality rate after 7 h of ventilation. Furthermore, the degree of VILI and survival after prolonged mechanical ventilation was VT-dependent, with a cut-off VT at values above 27 mL kg−1. To our knowledge this is the first investigation that systematically tested different VT over a longer period of time in ventilated, healthy rats.
Prolonged duration of mechanical ventilation, independent of the applied VT, induced some degree of VILI. There was a significant increase of the percentage of neutrophils in BAL fluid in all ventilated groups, and this increase was significantly different from the groups with spontaneous breathing at 7 h of mechanical ventilation. Furthermore, prolonged duration of mechanical ventilation increased this neutrophil sequestration in the alveolar space, which may be interpreted as an aggravation of local inflammation. This increased percentage of neutrophils after 7 h even in groups with assumed low pulmonary overdistention (VT9 and VT18) may favour the hypothesis that stretch injuries related to repetitive collapse and reopening of atelectatic regions may be the principal reason for these observations.
The alveolar-arterial protein concentration ratio increased significantly over time in the groups with high VT. This is strongly suggestive of an increased pulmonary permeability developing with the duration of mechanical ventilation despite most likely normal pulmonary venous pressure. High peak end-inspiratory volume seems to be the major determinant of VILI with oedema and major permeability alterations at least in rats . Scanning electron microscopy showed at higher inflations a higher number of endothelial and epithelial breaks per millimetre cell lining and more interstitial oedema . The alveolar-arterial protein concentration ratio used in this model has some limitations. This ratio is an estimate of pulmonary permeability [19,20]; however, protein measurement in BAL may also include proteins from the local inflammatory reaction and surfactant proteins. Thus, the reported alveolar-arterial protein concentration ratio may be an overestimation of the pulmonary permeability.
The increase of the alveolar-arterial protein ratio over time in groups with high VT was associated with a significant decrease in PaO2, a decreased lung compliance with increased lower and upper inflection points and expiratory collapse pressures. Similar increased lower and upper inflection points on the inspiratory limb of the P-V curve have been described in a model with intratracheal viscous liquid administration simulating pulmonary oedema with distal airway occlusions . Thus, there seems to be an association between right and down shifting of the inspiratory P-V curve and lung oedema. The deflation limbs of the pressure volume curves of the Groups VT36 and VT45 were right and down shifted (Fig. 2) with significantly increased collapse pressures. Similar curves have been observed in patients with adult respiratory distress syndrome of pulmonary origin  suggesting reduced alveolar airspace.
At 1 h, none of the investigated parameters distinguished deleterious from non-deleterious mechanical ventilation. In contrast to these results, Imanaka and colleagues observed after 40 min of mechanical ventilation with high VT and high-peak pressure (45 cmH2O) a decrease in PaO2, a huge neutrophil infiltration into alveolar spaces, an upregulation of CD54 and CD11b on alveolar macrophages and transforming growth factor-beta 1 mRNA in lung tissues . Our results differ also from the results of Dreyfuss and colleagues; they observed ventilator-induced lung oedema within 20 min after the application of a VT of 40 mL kg−1 (peak Paw 45 cmH2O) . This may be explained by the fact that our peak inspiratory pressures were slightly lower (maximal average peak pressure of 37 cmH2O) and that we always used a PEEP of 2.5 cmH2O. This low PEEP may have produced a protective effect on lung tissue by reducing shear forces on epithelial cells related to the containing alveolar recruitment and derecruitment . Indeed, Gajic showed that a PEEP of 3 cmH2O reduces significantly lung injuries in lungs of rats ventilated with VT of 40 mL kg−1 .
There was a reduced survival over time in groups with high VT and mean Paw >27 cmH2O (Groups VT36 and VT45). In these groups, prolonged mechanical ventilation (>3 h) was lethal in 62.5% and 100% of rats, respectively. This reduced survival was associated with significant alterations in gas exchange, lung mechanics and progressive circulatory failure. High lethality was also observed in a model with healthy baby pigs ventilated with a high peak Paw (40 cmH2O), a PEEP of 3 to 5 cmH2O and a FiO2 of 0.4 . This high Paw mechanical ventilation was terminated within 22 ± 11 h due to severe hypoxaemia. This later time point of death compared with our data seems to be related to species as suggested by Dreyfuss and colleagues  with an earlier hypoxia and mortality in smaller animals.
In isolated, perfused rabbit lungs a low respiratory rate of 3 min−1 reduced VILI . The reduction of the respiratory rate and the corresponding prolongation of inspiratory time in the high VT groups in our entire animal model seem not to counterbalance the deleterious effects of alveolar overdistension. Similar results have been observed in a sheep model. The variation of respiratory rate combined with a constant high peek inspiratory pressure of 50 cmH2O did not change the degree of VILI .
Spontaneous breathing during anaesthesia even with the risk of atelectasis offered the best protection against alveolar neutrophil sequestration and capillary leakage. This observation is in accordance with a recent clinical investigation showing that spontaneous breathing with ventilatory support increased respiratory system compliance, and PaO2 compared with controlled mechanical ventilation in patients at risk for adult respiratory distress syndrome .
In summary, we have developed an experimental model of slowly evolving, progressive, lethal VILI in anaesthetized normal rats demonstrating the preponderant effect of the duration (>3 h) of ‘aggressive’ ventilation and the cut-off value of the level of VT applied (>27 mL kg−1). The relevant end points for estimation of this injury may thus only be detectable after more than 1 h of mechanical ventilation. Our model investigating systematically the effect of both the duration and VT of mechanical ventilation offers investigators of VILI a basis for optimal model adaptation for their needs. In particular, we propose that the time factor should also be considered in this kind of research allowing more clinically relevant models and more information on slow acting processes such as mechanical stress (but not trauma), inflammation or infection and reparative mechanisms in lung tissue.
We would like to thank Jennifer Hantson, Division d'Investigations Anesthésiologiques, Hôpitaux Universitaires de Genève, for excellent technical assistance and Jérome Pugin, MD, Division des Soins Intensifs de Médecine, Hôpitaux Universitaires de Genève, for his helpful comments.
This work was supported by a research grant from the Department of Anesthesiology, Pharmacology and Surgical Intensive Care, University Hospitals of Geneva, Switzerland.
1. Dreyfuss D, Soler P, Saumon G. Mechanical ventilation-induced pulmonary edema. Interaction with previous lung alterations. Am J Respir Crit Care Med
2. Golder FJ, Fuller DD, Davenport PW, Johnson RD, Reier PJ, Bolser DC. Respiratory motor recovery after unilateral spinal cord injury: eliminating crossed phrenic activity decreases tidal volume
and increases contralateral respiratory motor output. J Neurosci
3. Tsuno K, Prato P, Kolobow T. Acute lung injury from mechanical ventilation at moderately high airway pressures. J Appl Physiol
4. Dreyfuss D, Saumon G. Ventilator-induced lung injury. Lessens from experimental studies. Am J Respir Crit Care Med
5. Parker JC, Hernandez LA, Peevy KJ. Mechanisms of ventilator-induced lung injury. Crit Care Med
6. Gajic O, Lee J, Doerr CH, Berrios JC, Myers JL, Hubmayr RD. Ventilator-induced cell wounding and repair in the intact lung. Am J Respir Crit Care Med
7. Nahum A, Hoyt J, Schmitz L, Moody J, Shapiro R, Marini JJ. Effect of mechanical ventilation strategy on dissemination of intratracheally instilled Escherichia coli in dogs. Crit Care Med
8. Muscedere JG, Mullen JB, Gan K, Slutsky AS. Tidal ventilation at low airway pressures can augment lung injury. Am J Respir Crit Care Med
9. Belperio JA, Keane MP, Burdick MD et al
. Critical role for CXCR2 and CXCR2 ligands during the pathogenesis of ventilator-induced lung injury. J Clin Invest
10. Anzueto A, Peters JI, Tobin MJ et al
. Effects of prolonged controlled mechanical ventilation on diaphragmatic function in healthy adult baboons. Crit Care Med
11. Simonson SG, Huang YC, Fracica PJ et al
. Changes in the lung after prolonged positive pressure ventilation in normal baboons. J Crit Care
12. Goldstein I, Bughalo MT, Marquette CH, Lenaour G, Lu Q, Rouby JJ. Mechanical ventilation-induced air-space enlargement during experimental pneumonia in piglets. Am J Respir Crit Care Med
13. Le Bourdelles G, Viires N, Boczkowski J, Seta N, Pavlovic D, Aubier M. Effects of mechanical ventilation on diaphragmatic contractile properties in rats. Am J Respir Crit Care Med
14. Behnia R, Molteni A, Waters CM et al
. Early markers of ventilator-induced lung injury in rats. Ann Clin Lab Sci
15. Cilley RE, Wang JY, Coran AG. Lung injury produced by moderate lung overinflation in rats. J Pediatr Surg
16. Lai YL, Diamond L. Comparison of five methods of analyzing respiratory pressure-volume curves. Respir Physiol
17. Venegas JG, Harris RS, Simon BA. A comprehensive equation for the pulmonary pressure-volume curve. J Appl Physiol
18. Capron F. Lavage broncho-alvéolaire. Arch Anat Cytol Path
19. Walder B, Brundler MA, Totsch M, Elia N, Morel DR. Influence of the type and rate of subarachnoid fluid infusion on lethal neurogenic pulmonary edema in rats. J Neurosurg Anesthesiol
20. Schneuwly OD, Licker M, Pastor CM et al
. Beneficial effects of leukocyte-depleted blood and low-potassium dextran solutions on microvascular permeability in preserved porcine lung. Am J Respir Crit Care Med
21. Dreyfuss D, Saumon G. Role of tidal volume
, FRC, and end-inspiratory volume in the development of pulmonary edema following mechanical ventilation. Am Rev Respir Dis
22. Fu Z, Costello ML, Tsukimoto K et al
. High lung volume increases stress failure in pulmonary capillaries. J Appl Physiol
23. Martin-Lefevre L, Ricard JD, Roupie E, Dreyfuss D, Saumon G. Significance of the changes in the respiratory system pressure-volume curve during acute lung injury in rats. Am J Respir Crit Care Med
24. Albaiceta GM, Taboada F, Parra D, Blanco A, Escudero D, Otero J. Differences in the deflation limb of the pressure- volume curves in acute respiratory distress syndrome
from pulmonary and extrapulmonary origin. Intens Care Med
25. Imanaka H, Shimaoka M, Matsuura N, Nishimura M, Ohta N, Kiyono H. Ventilator-induced lung injury is associated with neutrophil infiltration, macrophage activation, and TGF-beta 1 mRNA upregulation in rat
lungs. Anesth Analg
26. Dreyfuss D, Soler P, Basset G, Saumon G. High inflation pressure pulmonary edema. Respective effects of high airway pressure
, high tidal volume
, and positive end-expiratory pressure. Am Rev Respir Dis
27. Schiller HJ, McCann II UG, Carney DE, Gatto LA, Steinberg JM, Nieman GF. Altered alveolar mechanics in the acutely injured lung. Crit Care Med
28. Tsuno K, Miura K, Takeya M, Kolobow T, Morioka T. Histopathologic pulmonary changes from mechanical ventilation at high peak airway pressures. Am Rev Respir Dis
29. Hotchkiss Jr JR, Blanch L, Murias G et al
. Effects of decreased respiratory frequency on ventilator-induced lung injury. Am J Respir Crit Care Med
30. Rich PB, Reickert CA, Sawada S et al
. Effect of rate and inspiratory flow on ventilator-induced lung injury. J Trauma
31. Putensen C, Zech S, Wrigge H et al
. Long-term effects of spontaneous breathing during ventilatory support in patients with acute lung injury. Am J Respir Crit Care Med