Ventilator-induced lung injury: historical perspectives and clinical implications

Abstract

Mechanical ventilation can produce lung physiological and morphological alterations termed ventilator-induced lung injury (VILI). Early experimental studies demonstrated that the main determinant of VILI is lung end-inspiratory volume. The clinical relevance of these experimental findings received resounding confirmation with the results of the acute respiratory distress syndrome (ARDS) Network study, which showed a 22% reduction in mortality in patients with the acute respiratory distress syndrome through a simple reduction in tidal volume. In contrast, the clinical relevance of low lung volume injury remains debated and the application of high positive end-expiratory pressure levels can contribute to lung overdistension and thus be deleterious. The significance of inflammatory alterations observed during VILI is debated and has not translated into clinical application. This review examines seminal experimental studies that led to our current understanding of VILI and contributed to the current recommendations in the respiratory support of ARDS patients.

Figures

Figure 1

Macroscopic aspect of rat lungs after mechanical ventilation at 45 cm H2O peak airway pressure. Left: normal lungs; middle: after 5 min of high airway pressure mechanical ventilation. Note the focal zones of atelectasis (in particular at the left lung apex); right: after 20 min, the lungs were markedly enlarged and congestive; edema fluid fills the tracheal cannula. Used with permission. From Dreyfuss et al. [15].

Figure 2

Effect of gradual exposure of normal rats to 45 cmH2O peak airway pressure ventilation on lung water content and pulmonary permeability. Pulmonary edema was assessed by measuring extravascular lung water content (Qwl/BW) and permeability alterations by measuring bloodless dry-lung weight (DLW/BW) and the distribution space in the lungs of 125I-labeled albumin (Alb. Space). Permeability pulmonary edema developed after only 5 minutes of mechanical ventilation. After 20 min of mechanical ventilation, there was a dramatic increase in lung water content and pulmonary permeability (p < 0.01 vs. other groups). Used with permission. From Dreyfuss et al. [8].

Figure 3

Changes in the ultrastructural appearance of the blood-air barrier after 5 min (A) and 20 min (B) mechanical ventilation of a closed-chest rat at 45 cm H2O peak airway pressure. (A) The thin part of an endothelial cell (En) is detached from the basement membrane (arrowhead) forming a bleb. (B) Very severe changes in the alveolar-capillary barrier resulting in diffuse alveolar damage. The epithelial layer is totally destroyed (upper right quadrant) leading to denudation of the basement membrane (arrows). Hyaline membranes (HM), composed of cell debris and fibrin (f), occupy the alveolar space. IE, interstitial edema. Used with permission. From Dreyfuss et al. [15] (panel A) and [8] (panel B).

Figure 4

Comparison of the effects of high peak (45 cm H2O) positive inspiratory pressure plus high tidal-volume ventilation (HiP-HiV) with the effects of negative inspiratory airway pressure plus high tidal-volume ventilation (iron lung ventilation = LoP-HiV) and of high peak (45 cm H2O) positive inspiratory pressure plus low tidal-volume ventilation (thoracoabdominal strapping = HiP-LoV). Pulmonary edema was assessed by the determination of extravascular lung water content (Qwl/BW) and permeability alterations by the determination of bloodless dry-lung weight (DLW/BW) and of the distribution space in the lungs of 125I-labeled albumin (Alb. Sp.). Dotted lines represent the upper 95 percent confidence limit for control values. Permeability edema occurred in both groups receiving high tidal-volume ventilation. Animals ventilated with a high peak-pressure and a normal tidal volume had no edema. Used with permission. From Dreyfuss et al. [29].

Figure 5

Relationship between plateau pressure (Pplat) and 111In-transferrin lung-to-heart ratio slope (an index of lung microvascular permeability; left axis, open circles) and alveolar 99 mTc-albumin permeability-surface area product (an index of alveolar epithelium permeability; right axis, full circles) in mechanically ventilated rats. Both indexes dramatically increased for plateau pressures comprised between 20 and 25 cm H2O. Used with permission. From de Prost et al. [32].

Figure 6

Effect of increasing PEEP from 0 to 15 cm H2O during ventilation with two levels of tidal volume (VT, 7 ml/kg of body weight = Lo VT; 14 ml/kg BW of body weight = Med VT). When PEEP was increased, pulmonary edema, as evaluated by extravascular lung water (Qwl), occurred. The level of PEEP required to produce edema varied with VT; 15 cm H2O PEEP during ventilation with a low VT versus 10 cm H2O PEEP during ventilation with a moderately increased VT. *p < 0.05; **p < 0.01 vs. ZEEP and the same VT. Used with permission. From Dreyfuss and Saumon [33].

Figure 7

Interaction between previous lung alterations and mechanical ventilation on pulmonary edema. Effect of previous toxic lung injury. Extravascular lung water (Qwl) after mechanical ventilation in normal rats (open circles) and in rats with mild lung injury produced by α-naphthylthiourea (ANTU) (closed circles). Tidal volume (VT) varied from 7 to 45 ml/kg body weight. The solid line represents the Qwl value expected for the aggravating effect of ANTU on edema caused by ventilation, assuming additivity. ANTU did not potentiate the effect of ventilation with VT up to 33 ml/kg body weight. In contrast, VT 45 ml/kg body weight produced an increase in edema that greatly exceeded additivity, indicating synergy between the two insults. Used with permission. From Dreyfuss et al. [38].

Figure 8

Effect of different ventilator strategies on cytokine concentrations in lung lavage of isolated unperfused rat lungs. Four ventilator settings were used: controls (C = normal tidal volume), moderate tidal volume + high PEEP (MVHP), moderate tidal volume + zero PEEP (MVZP), high tidal volume + zero PEEP (HVZP) resulting in the same end-inspiratory distension as MVHP. Major increases in cytokine concentrations were observed with HVZP. Used with permission. From Tremblay et al. [57].

Figure 9

TNF-α, IL-1β, and MIP-2 concentrations in bronchoalveolar lavage fluid of isolated, nonperfused rat lungs maintained for 2 h in a statically inflated state at 7 cm H2O airway pressure (VT0), ventilated with 7 mL/kg tidal volume and 3 cm H2O positive end-expiratory pressure (VT7), or ventilated with 42 mL/kg VT and zero end-expiratory pressure (VT42). IL-1β and MIP-2 concentrations were slightly higher (*p < 0.05) in the VT42 group. There was no difference in TNF-α concentration. Used with permission. From Ricard et al. [59].

Figure 10

(A) Comparison of lung cytokine levels in nonventilated rats (Cont) and in rats subjected to an injurious mechanical ventilation strategy (30 mL/kg tidal volume and zero end-expiratory pressure) alone (HV) or to hemorrhagic shock-reperfusion alone (HSR). Compared with controls, lung cytokine concentrations were higher after HSR but not after HV. (B) Comparison of lung cytokine levels in rats subjected to HSR alone, HSR combined with conventional ventilation (HSR-CV) and HSR followed by HV (HSR-HV). Injurious ventilation (HV) after HSR significantly increased mediator release above the levels observed after HSR alone or combined with conventional ventilation. Results are expressed in pg/g. *p < 0.05 as compared with controls (A) or HSR (B). The same observations were made in bronchoalveolar lavage fluid and in plasma. Adapted with permission. From Bouadma et al. [60].

Figure 11

Examples of scintigraphy images integrating the 15 min following tracer instillation (t0-t15; left panels) and the last 15 min of the experiment (t195-t210; right panels). Regions of interest (ROIs) were drawn around initial focus of edema (ROIE), the apex of the same lung (ROIA), the contralateral lung (ROICL), and over the thorax (ROIT). At baseline (left panels), all animals were ventilated with a tidal volume of 8 mL/kg and a PEEP of 2 cmH2O and exhibited focalized localization of the tracer in the left lung. When the same ventilator settings were kept during the experiment (a), the tracer remained remarkably confined to the initial zone; there was no contralateral and slight homolateral dissemination. High-volume ventilation (Pplat = 30 cmH2O) with no PEEP (b) induced strong homo- and contralateral dispersion of the tracer and systemic leakage, as attested by the evident decrease in overall activity. High-volume ventilation with 6 cmH2O PEEP (c) induced systemic, but not contralateral, dissemination of the tracer. Used with permission. From de Prost et al. [35].