This chapter is motivated by the clinical question of how mechanical ventilation might contribute to the development of MODS. We will extract and integrate information from several different fields of investigation, and wish to highlight some of the most recent and pertinent findings that have contributed to our understanding of the mechanisms by which mechanical ventilation contributes to MODS. MODS is often irreversible, with mortality ranging from 60 to 98% [16]. To date, there is neither an effective treatment for MODS nor an effective means for preventing its onset. By understanding the mechanisms by which mechanical ventilation contributes to MODS, this new conceptualization of VILI could lead to a paradigm shift for therapies to prevent ventilated ARDS patients from developing MODS.

Physiological Effects of Mechanical Ventilation

It has been known for decades that mechanical ventilation can have important effects on systemic and regional hemodynamics, as well as global oxygen delivery and consumption, as highlighted in many excellent reviews [17-19]. The application of positive end-expiratory pressure (PEEP) is one of the key components of the ventilatory management ofARDS.PEEPdecreases cardiac output bydecreasing

Mechanical Ventilation

Fig. 1. Postulated mechanisms whereby mechanical ventilation contributes to MODS. From [2] with permission

Distal Organs

■ tissue injury secondary to inflammatory mediators/cells ' impaired oxygen delivwy • bacteremia venous return and by an increase in right ventricular afterload [20-22]; as a result, global oxygen delivery falls. This effect can usually be reversed by volume loading and inotropic agents. PEEP may also induce regionalhemodynamic alterations due to the combined effects of decreases in cardiac output andblood flow redistribution [17]. In addition, intraabdominal pressure may increase in response to diaphragmatic swings, so that the splanchnic organs may be compressed [23].

In experimental conditions, a number of authors have reported that PEEP decreases splanchnic blood flow [24,25], although splanchnic oxygen consumption can usually be maintained by a compensatory increase in oxygen extraction [26]. A similar response to PEEP has been observed in patients during abdominal surgery [27]. Trager and colleagues [28] progressively increased the PEEP level up to 15 cmH2O in patients with septic shock. Cardiac output decreased in all patients, and hepatic vein oxygen saturation and hepatic vein oxygen saturation decreased more at PEEP 15 than at 10 cmH2O. Kiefer and colleagues [29] reported that in patients with ALI, moderate levels of PEEP did not significantly alter splanchnic blood flow as well as indices of tissue hypoxia such as gastric mucosal PO2 and the blood lactate to pyruvate ratio, provided fluid resuscitation was able to maintain cardiac output. Fournell and colleagues [30] found in anesthetized dogs that gastric mucosal oxygen saturation decreased at PEEP 15 cmH2O despite the maintenance of cardiac output by fluid infusion. The effect of PEEP on gastric mucosal oxygenation may be dissociated from the effect on cardiac output and splanchnic blood flow. Seeman-Lodding and colleagues [31] measured arterio-venous concentration gradients of tissue-type plasminogen activator (t-PA) and the respective blood flow across the pulmonary, coronary, hepatic and preportal vascular beds in pigs after the application of PEEP up to 10 cmH2O. They found thatwith PEEP, the magnitude of the preportal net release of t-PA was markedly enhanced, with a concomitant decrease in liver blood flow, suggesting that clinically used levels of PEEP induce increases in net release of t-PA within preportal organs.

Reduction of tidal volume (Vt) is another key component of the ventilatory management of ALI/ARDS. Vt reduction may exert systemic and regional circulatory effects. The increase in cardiac output resulting in oxygen delivery during Vt reduction has been observed in several studies ofpatientswithARDS [32,33].There are two mechanisms that could be responsible for the increased cardiac output with reduced Vt: First, the decrease in airway pressure, by inducing a reduction in pleural pressure, will lead to an increase in venous return. In addition, a decrease in transpulmonary pressure could decrease overdistension of the lung, leading to a decrease in the resistance of alveolar microvessels, hence imposing a relatively reduced transient resistance to right ventiricular output. Another mechanism leading to an increase in cardiac output during Vt reduction is the resulting hypercapnia. A number of studies have demonstrated that hypercapnia induces an increase in sympathetic activity, which may enhance cardiac output [34, 35]. In addition to the central hemodynamic effects, Vt reduction may also induce regional hemodynamic alterations such as effects on the gut circulation. The effect of Vt reduction on gut regional blood flow is controversial. Cardenas and colleagues [36] reported a parallel increase in cardiac output and mesenteric blood flow after Vt reduction in animals. In contrast, despite an increase in cardiac output, Sitbon and colleagues [37] observed no increase in gastric mucosal blood flow after Vt reduction resulting in hypercapnia in ARDS patients. They speculated that the heterogeneity in the individual response of gastric mucosal blood low during Vt reduction resulting in hypercapnia could be due to opposing direct (i.e., local vasodilatory effect) and indirect (i.e., global sympathetic stimulation) effects of hypercapnia on gut vessels.

Relatively little is known with regards to changes in regional circulation and global oxygen delivery due to mechanical ventilation in relation to protective ventilatory strategies (small Vt and higher PEEP) versus conventional ventilatory strategies (large Vt and low PEEP) in patients with ARDS.

Mechanical Strain-induced Release of Inflammatory Mediators in vitro

Because of the complexity of lung structure, the variety of cell types, and the variety of mechanical forces to which these cells are exposed, there are a number of mechanisms by which mechanical stimulation may alter cellular responses. There are several excellent recent reviews of this topic [38-42]. Dos Santos and Slutsky [38] have detailed putative mechanoreceptive mechanisms leading to proinflammatory signaling by lung cells in VILI. In this review, the authors emphasized both stretch-activated ion channels and plasma membrane disruption. Grembowicz and colleagues demonstrated in a series of studies that calcium influx accompanying plasma membrane disruption is associated with translocation of nuclear factor kappa B (NF-kB) and upregulation of stress response genes [3]. These, in turn, could represent up-stream events leading to pro-inflammatory signaling. In a recent review, Vlahakis and Hubmayr [43] emphasized that plasma membrane stress failure is a central event in the pathophysiology of injury from alveolar overdistension.

Vlahakis and colleagues [7] found that in cultured A549 cells, mRNA for inter-leukin (IL)-8 increased 4-fold after four hours of cyclic strain sufficient to change the cell surface area by 30%. Continued strain for up to 48 hours resulted in a nearly 50% increase in IL-8 secretion compared with non-strained controls. This finding was confirmed by Quinn and colleagues [6], who also found that the increase in IL-8 secretion was associated with activation ofthe JNK family of mitogen activated protein kinases (MAPKs). Mourgeon et al. [4] found that in cultured fetal rat lung cell, macrophage inflammatory protein-2 (MIP-2), a rodent analog of human IL-8, was liberated into supernatant after four hours of cyclic strain.

Pugin and colleagues [5] cultured human alveolar macrophages on flexible silastic membranes and exposed the cells to cyclic stretch for up to 32 hours. They found that cyclic strain increased secretion of IL-8 and MMP-9 (gelatinase b - a type IV collagenase), but not tumor necrosis factor (TNF)-a or IL-6. When the macrophages were pretreated with lipopolysaccharide (LPS), TNF-a and IL-6 secretion increased to a greater extent in strained cells compared with static cultures. Mechanical strain also activated NF-kB in macrophages after 30 minutes. In another study, these same authors found that a bronchial epithelial cell line, an endothelial cell line, and primary lung fibroblasts exposed to the same cyclic strain, did not secrete IL-8, but macrophages and A549 cells secreted IL-8 in response to mechanical distention. The relative amount ofIL-8 secreted from macrophages was much greater than that from A549 cells [44].

Pulmonary endothelial cells form a continuous monolayer on the luminal surface of the lung vasculature. During mechanical ventilation,pulmonary endothelial cells are exposed to shear stress, in particular repetitive opening and collapse of atelectatic regions of the lung, as well as changes in transluminal pressure during alveolar inflation. Changes in shear stress could be sensed directly by cell membrane components such as membrane proteins, ion channels, or caveolae or by alterations of the cellular cytoskeleton; subsequent cellular signaling cascades through phosphorylation events can lead to diverse effects such as the release of cytokines and other mediators, activation of transcription factors, altered gene expression and protein expression, and cell division or death [39]. Gan and colleagues [45, 46] found that various combinations of shear and intraluminal pressures in human umbilical veins caused gene and protein expression of vascular endothelial growth factor as well as c-jun and c-fos.

Pulmonary and Systemic Release of Inflammatory Mediators in VILI ex vivo and in vivo

Under experimental conditions, high VT and/or low PEEP induces release of pro-inflammatory cytokines into the airspaces andbloodstream, increased neutro-phil infiltration into the lung, and activation of alveolar macrophages. There is considerable evidence supporting a role for the release of inflammatory mediators in VILI ex vivo and in vivo.

Tremblay and colleagues [9] found that ventilation of isolated, non-perfused rat lungs with Vt of 40 ml/kg and zero PEEP for 2 hrs resulted in large increases in lavage concentrations of TNF-a, IL-1p, IL-6, and MIP-2. The increase in these cytokines was greater if rats were pretreated with LPS. Northern blot analysis of whole lung homogenates revealed increased expression of c-fos mRNA, with both high and moderate Vt and/or zero PEEP ventilation.

Chiumello and colleagues [10] demonstrated that mechanical ventilation for 4 hrs with a Vt of 16 ml/kg, zero PEEP versus Vt of 9 ml/kg, PEEP of 5 cmH2O produced an increased release of pro-inflammatory cytokines (TNF-a and MIP-2) in the lung as well as in the systemic circulation in an in vivo acid aspiration model in rats. Haitsma and colleagues [47] found that mechanical ventilation for 20 min with peak inspiratory pressure/PEEP ratio 45/0 versus 45/10 increased release of TNF-a into the systemic circulation in rats pretreated with LPS intratracheally, while increased bronchoalveolar lavage (BAL) TNF-a in rats pretreated with intraperitoneal LPS, suggested loss of alveolar and systemic compartmentalization of TNF-arespectively.

Evidence for the importance of these inflammatory mediators in the development of VILI comes from experimental studies of the effects of an anti-TNF antibody and an IL-1 receptor antagonist on lung injury following saline-lavage. Imai and colleagues [48] employed anti-TNF-a antibodies and observed improvements in oxygenation and respiratory compliance, decreased lavage neutrophil counts, as well as reduced histological evidence of lung injury. Narimanbekov and colleagues [49] used an IL-ip receptor antagonist and found reduced lung lavage concentrations of a number of markers of lung injury (i.e., albumin, elastase, and neutrophils) in a saline-lavaged rabbit model.

Recently Held and colleagues [50] demonstrated in isolated perfused mouse lungs that both LPS and overventilation caused translocation of NF-kB to the nucleus, leading to release of MIP-2. However, there were major differences in response to the two different stimuli when they used LPS-resistant C3H/HeJ mice, which have abnormalities in Toll like receptor (TLR)-4. In LPS-resistant C3H/HeJ mice overventilation, but not LPS, caused translocation of NF-kB and release of MIP-2, suggesting that initial signaling steps between LPS and ventilation differ and that the NF-kB translocation elicited by overventilation is independent of TLR4.

Passage of Mediators from Lung to Bloodstream

One important question is where the increased systemic pro-inflammatory mediators originate from. Loss of compartmentalization of local pulmonary inflammatory mediators due to mechanical ventilation has been recently proposed. The alveolar barrier restricts transport of macro-molecules of a size similar to that of cytokines (15-20 kDa). After secretion in the lung, cytokines such as TNF-a remain in the alveolar space and leak into the circulation only if there is injury of the alveolar barrier. Several investigators have shown that increased permeability of the alveolar-capillary interface, as a result of lung injury leads to the release of mediators into circulation that would normally have remained compartmentalized within the alveolar space [51, 52]. Tutor and colleagues [51] employed an isolated perfused rat lung model in which they injected TNF-a into the lung and measured its appearance in the perfusate. They found that the perfusate TNF-a concentrations were increased only when alveolar-capillary permeability was increased and not in normal lung, suggesting that loss of compartmentalization of alveolar TNF-a could occur, but only in the context of damage to the alveolar-capillary membrane. Von Bethman and colleagues [8] reported that in an isolated perfused murine lung model, ventilation with higher transpulmonary pressure (25 cmH2O) versus normal pressure (10 cmH2O) led to a significant increase in concentration of both TNF-a and IL-6 in the perfusate. As compartmentalization of the local pulmonary response is lost, systemic release of inflammatory mediators may promote the massive inflammatory response that underlies MODS.

ARDS is characterized by a loss of integrity of the alveolar capillary barrier due to severe diffuse alveolar damage, leading to bidirectional protein flux. Therefore, not only pro-inflammatory cytokines butalsolocallysecretedproteins,particularly the surfactant-associated protein (SP) may pass into the systemic circulation. SP-A, SP-B and SP-D, which have anti-inflammatory properties, have been detected in serum of ARDS patients, and have been associated with outcome of patients with ARDS [53, 54]. The balance between the pro- and anti-inflammatory cytokines passing from the lung to the bloodstream may be more important in determining the subsequent effect than the absolute values of any single mediators.

Bacterial Translocation in Mechanical Ventilation

Another mechanism whereby mechanical ventilation may contribute to the development of a systemic inflammatory response is by promoting bacterial translocation from the air spaces into the circulation. Two recent studies evaluated the influence of mechanical ventilatory strategy on the translocation of bacteria from the lung into the bloodstream in dogs and rats[55, 56]. After intratracheal instillation of bacteria, these animals were ventilated with a high transpulmonary pressure (-30 cmH2O) and minimal (0 to 3 cmH2O) or 10 cmH2O PEEP. Bacteremia seldom occurred in control animals ventilated with low airway pressure, whereas it was found in nearly all animals ventilated with high Vt and a low PEEP. In contrast, ventilation with the same transpulmonary pressure but with 10 cmH2O PEEP resulted in rates of bacteremia as low as in controls. In a saline-lavaged rabbit lung injury model, mechanical ventilation with a Vt of 12 ml/kg and without PEEP resulted in translocation of intratracheally instilled endotoxin into the systemic circulation, but ventilation with a Vt of 5 ml/kg and a PEEP of 10 cm H2O did not. The appearance of endotoxin in the blood stream was associated with an increase in plasma TNF-a [57]. Since the gut could be a "motor" of MODS, bacterial and endotoxin translocation caused by mechanical ventilation may play a role in development of MODS [15].

Pulmonary and Systemic Inflammatory Mediators in VILI in Clinical Studies

Elevated levels of pro-inflammatory mediators have been measured in airspace lavage fluid and in the plasma of patients with ARDS. Ranieri and colleagues [11] measured BAL and plasma levels of several pro-inflammatory cytokines in 44 patients with ARDS. At study entry, patients were randomized to receive mechanical ventilation with a conventional strategy (mean Vt of 11.1 ml/kg, mean plateau airway pressure 31 cmH2O, and mean PEEP of 6.5 cmH2O), or a protective ventilatory strategy with Vt (7.6 ml/kg) and higher PEEP (14.8 cmH2O) with a mean plateau airway pressure of 24.6 cmH2O. PEEP in the latter group was set above the lower inflection point of the respiratory system pressure volume curve. Baseline measurements of cytokines were made at the time of admission (study entry), and were then measured serially for three days. By 36 hours, BAL fluid from patients in the protective ventilation group had significantly fewer polymorphonuclear cells and lower concentrations ofTNF-a, IL-1 p, IL-6, and IL-8. Plasma levels of IL-6 were also significantly lower in the patients receiving protective ventilation. The NIH ARDS Network study found lower levels of plasma IL-6 at three days in patients ventilated with low Vt compared with conventional Vt [12].

In patients with ARDS, concentration ofTNF-a, IL-1p and IL-6 were higher in the arterialized blood (obtained via a wedged Swan-Ganz catheter), as compared with mixed venous blood, suggesting that the lungs in these patients were contributing cytokines to the systemic circulation [58]. Recently, Stuber and colleagues [59] studied patients with ALI and foundthat switching to conventional mechanical ventilation (Vt of 12 ml/kg, PEEP of 5 cmH2O) from a lung protective strategy (Vt of 5 ml/kg, PEEP of 15 cmH2O) was associated with a marked increase in plasma cytokine levels within 1 h, while plasma cytokine levels returned to baseline when lung protective settings was reestablished. In contrast, in 39 patients with normal lungs, without ARDS, Wrigge and colleagues [60] found that ventilation with a high Vt (15 ml/kg) and zero PEEP did not affect plasma levels of either IL-6, TNF-a, IL-1 receptor antagonist, or IL-10. These data suggest that if the lung is normal, mechanical ventilation, even with relatively large Vt, is unlikely to increase alveolar capillary permeability and augment the pulmonary inflammatory response, leading to increased production of inflammatory mediators.

Multiple Organ Dysfunction and VILI in Clinical Studies

There are now prospective data in the literature examining the hypothesis that a protective ventilatory strategy can have an impact on the development of MODS. As previously mentioned, Ranieri and colleagues reported that the use of a lung-protective strategy in patients with ARDS attenuated an increase in pulmonary and systemic cytokine levels including TNF-a and IL-6 [11]. In a subsequent post hoc analysis (Fig. 2), they found a higher incidence of renal failure in ARDS patients ventilated with conventional ventilation compared with a lung protective strategy. Furthermore, they found a significant correlation between overall MODS score with changes in plasma concentration of a number of inflammatory mediators (IL6, TNF-a, IL-1 p and IL-8), which are involved in MODS [13]. Similarly, the NIH ARDS Network reported the results of a randomized, clinical trial comparing a Vt of 12 ml/kg with 6 ml/kg (predicted body weight). They found lower levels of plasma IL-6 in the 6 ml/kg tidal volume group, associated with a greater number of organ failure-free days and a 22% reduction in mortality rate [12]. These data suggest that mechanical ventilation, which is invariably used in the care of patients with ARDS, can increase alveolar capillary permeability and augment the pulmonary inflammatory response, leading to increased production of inflammatory mediators. If these mediators (e.g., cytokines) enter the circulation, they could contribute to the development of MODS. It is important to emphasize that MODS is a complex syndrome, often precipitated and intensified by a series of events rather than a single event. A likely scenario is that there is an ongoing inflammatory response as a result of the persistence of the factors that either initiated or exacerbated the response, and/or failure of intrinsic regulatory mechanisms. Conversely, patients with normal lungs who receive prolonged ventilation (e.g., neuromuscular disease with respiratory failure) would not develop VILI nor MODS.


In this review, we have focused mainly on the effects of mechanical ventilation on hemodynamics and systemic inflammation, but how might these lead to end organ dysfunction? Recently, it has been shown that inflammatory cytokines/chemokines can modulate apoptosis in various cell types [61] and increased apoptosis has been detected in animal model of CLP [62, 63] as well as in patients dying of sepsis and

Mechanical Organ

Fig. 2. a. Number of patients with organ system failure in the 2 groups of patients on entry and within 72 hours of mechanical ventilation. Patients may have failure of more than 1 organ system. b. Changes in total number of failing organs and in plasma concentration of interleukin 6. Asterisk indicates number of failing organ systems at 72 hours minus number on entry. Dagger indicates change in interleukin 6 derived from value at 36 hours minus entry value. From [13] with permission.

Fig. 2. a. Number of patients with organ system failure in the 2 groups of patients on entry and within 72 hours of mechanical ventilation. Patients may have failure of more than 1 organ system. b. Changes in total number of failing organs and in plasma concentration of interleukin 6. Asterisk indicates number of failing organ systems at 72 hours minus number on entry. Dagger indicates change in interleukin 6 derived from value at 36 hours minus entry value. From [13] with permission.

MODS [64]. Imai and colleagues have demonstrated increases in end organ epithelial cell apoptosis after injurious mechanical ventilation in an in vivo rabbit ARDS model [65]. Apoptosis might thus be an important down-stream effect in the development of systemic inflammatory response syndrome (SIRS)/MODS caused by VILI in ALI/ARDS, although there are insufficient data to confirm this hypothesis at present.

Was this article helpful?

0 0

Post a comment