The most important findings of the present study are as follows: Vt and PEEP act synergistically to stabilize alveoli; increasing PEEP is more effective at stabilizing alveoli than reducing Vt; stabilizing alveoli and preventing microatelectasis with low Vt/high PEEP reduces VILI; and the mechanism of VILI in this three hour animal model appears to be mechanical rather than inflammatory. Ventilating the surfactant-injured lung with high Vt/low PEEP results in a continuum of abnormal alveolar mechanics ranging from slightly unstable alveoli to complete recruitment/derecruitment (Additional file
2). Conversely, ventilation with low Vt/high PEEP stabilizes alveoli and provides an important means of defense against VILI in the setting of abnormal surfactant function. The issues are more complex clinically because the impact of improper mechanical ventilation may vary with the degree of initial lung injury and the heterogeneity of ventilation.
Although low Vt ventilation is not new a concept in protective mechanical ventilation [
18], the observations that high PEEP and low Vt work synergistically to stabilize alveoli and that increasing PEEP is more effective than reducing Vt at stabilizing alveoli are unique. If alveolar instability causes lung injury, as both our previous study [
27] and present one suggest, it appears that increasing PEEP would provide a greater degree of 'protection' than that provided by reduction in Vt. Examining the trends in I-EΔ when Vt was changed with a similar PEEP reveals that there was a 47.6% decrease in I-EΔ (alveoli were stabilized) between Vt 15 (cc/kg)/PEEP 5 (cmH
2O) and Vt 6/PEEP 5; a 31.2% decrease between Vt 15/PEEP 10 and Vt 6/PEEP 10; and a 58.7% decrease between Vt 15/PEEP 20 and Vt 6/PEEP 20 (Figure
2b and Table
1). However, I-EΔ decreased to a much greater degree, especially at lower Vt, when PEEP was changed with similar Vt; we saw a 1067% decrease in I-EΔ between Vt 6/PEEP 5 and Vt 6/PEEP 20; a 660% decrease between Vt 12/PEEP 5 and Vt 12/PEEP 20; and a 64.1% decrease between Vt 15/PEEP 5 and Vt 15/PEEP 20. These data demonstrate that PEEP can have a much greater impact on alveolar stabilization than reduced Vt, and they suggest that increasing PEEP may be more beneficial in the prevention of VILI than lowering Vt. In addition, we noted that even when using the ventilator strategy that resulted in the best stabilization of alveoli (low Vt/high PEEP), these alveoli were less stable than normal ones. It is known that unstable alveoli cause VILI [
27], but the degree of instability necessary to cause injury is not known. It is possible that the slight increase in instability above normal stability (Figure
2) could be sufficient to cause alveolar damage. If this is true, then other modes of protective ventilation such as high-frequency oscillatory ventilation may cause less VILI than low Vt/high PEEP.
Examination of alveolar mechanics also provides new insight as to the time course of development of VILI. When animals were initially placed on high Vt/low PEEP ventilation, alveoli were unstable compared with those in the low Vt/high PEEP group, but the number of patent alveoli was similar between groups for the first hour (Figure
3a). Mean alveolar stability improved over time in the high Vt/low PEEP group because unstable alveoli progressively derecruited (Figure
3a), suggesting that unstable alveoli will eventually collapse (Figure
3b). Progressive alveolar derecruitment is a concern with low Vt ventilation [
19,
28‐
30]; however, progressive derecruitment was also observed with high Vt ventilation in this study. Thus, it appears that with sufficient injury to the alveolus progressive derecruitment can occur even if PEEP is elevated.
Protection with low tidal volume and elevated positive end-expiratory pressure
Reduced lung injury with low Vt ventilation has been the subject of much investigation, and this strategy has become the standard-of-care for ARDS patients [
1,
18]. A study by Frank and coworkers [
31] demonstrated reduced atelectasis and alveolar epithelial injury when Vt was reduced from 12 to 6 cc/kg. In a clinical trial involving 44 ARDS patients [
10], reduction in mean tidal volumes (11.1 versus 7.6 cc/kg) produced a marked reduction in BAL fluid levels of TNF-α, IL-1, IL-6 and IL-8, suggesting that lower Vts may reduce biotrauma-induced VILI.
Although low Vt ventilation has become the standard-of-care for ARDS patients, it may exacerbate lung injury if insufficient PEEP is applied to prevent end-expiratory alveolar collapse [
32]. One of the aims of the present study was to show the relative value of lowering Vt versus raising PEEP in reducing alveolar stability. We demonstrated that increasing PEEP from 5 to 10 cmH
2O with a Vt of 6 cc/kg provided much greater alveolar stability (53.7% decrease in I-EΔ) than reducing Vt from 15 to 6 cc/kg at either 5 cmH
2O (46.7% decrease) or 10 cmH
2O (31.2% decrease) PEEP. If these results can be extrapolated to clinical treatment of acute lung injury/ARDS, then there is certainly a clear benefit from low Vt ventilation, but there is a potentially greater benefit from even modest increases in PEEP.
Richard and coworkers [
20] demonstrated that alveolar derecruitment is more a function of reduced plateau pressures than of low Vt. In addition, they showed that increased levels of PEEP could prevent derecruitment. These findings are consistent with the results of the present study. In addition, low Vt/high PEEP ventilation – similar to that used in our low Vt/high PEEP group – has yielded improvements in oxygenation [
33‐
35] and reduces both intra-alveolar protein levels [
33] and lung injury [
34,
35], supporting the hypothesis that decreased Vt and increased PEEP work synergistly to reduce alveolar instability and reduce VILI.
Mechanical trauma versus 'biotrauma'
It has been suggested that injurious mechanical ventilation, such as high Vt and/or low PEEP levels, produces lung injury through biotrauma. Stretch imposed on alveolar epithelial cells has demonstrated dramatic increases in IL-8 release as well as IL-8 gene transcription
in vitro [
13]. A clinical study involving 44 ARDS patients [
12] identified a significant reduction in IL-6, IL-8, and TNF in those patients ventilated with a low Vt in combination with elevated PEEP. In the present study histologic injury was significantly worse in the high Vt/low PEEP group, but the levels of inflammatory mediators were not significantly increased by this strategy in either serum or BAL fluid. Furthermore, neutrophil elastase actually declined over time, regardless of ventilation strategy. These data suggest that mechanical trauma (shear stress from unstable alveoli) rather than biotrauma is the initial mechanism of VILI. If this study had been conducted for a longer time, then we hypothesize that inflammatory mediators would have increased in the high Vt/low PEEP group. In our previous study [
27] we did identify increases in IL-6 and IL-8 when we extended the study by an additional hour, although proteases were not increased, similar to the present study. There was a significant increase in the number of polymorphonuclear leukocytes in lung tissue in the high Vt/low PEEP group compared with the low Vt/high PEEP group, even though there was no difference in the measured inflammatory mediators. It is known that cytokines are not free floating in the plasma but can be bound to cells. This suggests that there was an increase in the tissue-specific cytokines in lung in the high Vt/low PEEP group that resulted in increased polymorphonuclear leukocyte sequestration. We previously showed in a similar animal model that there can be an increase in tissue bound cytokines (TNF and IL-6) [
27].
Critique of methods
Detailed critiques of this
in vivo microscopic technique have previously been reported [
6,
22‐
24,
27,
36,
37]. This
in vivo microscopic technique allows measurements of alveoli in only two dimensions, and thus we measured alveolar cross-sectional area at inspiration and expiration and these data were used to calculate changes in alveolar size with ventilation (I-EΔ). Although this technique only measures alveolar mechanics in two dimensions, the mechanics of alveoli in the normal and surfactant-deactivated lung are profoundly different. Therefore, our hypothesis that alveolar instability is injurious to the lung appears valid, despite our inability to measure precise changes in alveolar volume. Additionally, our
in vivo microscopic technique does not provide us with a global measure of alveolar mechanics, but rather we are restricted to the subpleural alveoli in our microscopic field. We have recently demonstrated that subpleural alveoli do not over-distend even at very high airway pressure (60 cmH
2O; see the data repository by DiRocco and coworkers [
37]), and so we did not expect to observe alveolar over-distension in the PEEP 20/Vt 15 group.
Although not ideal, this technique provides a bridges between purely physiologic approaches to assessment of alveolar mechanics (such as pressure-volume curve analysis) and purely anatomic approaches (such as computed tomography scanning). The short duration of the study might not have been sufficient time to allow a change in inflammatory mediators to take place. Ventilation with low Vt resulted in a significant increase in PCO
2, which could not be normalized by increasing respiratory rate. It has been shown that high PCO
2 can protect against VILI [
38], and so it is possible that the reduction in tissue injury in the low Vt/high PEEP group could have been due to high PCO
2 rather than stabilization of alveoli. Finally, we did not use a recruitment maneuver before setting PEEP and Vt, and it is possible that the results of the experiment would have been altered if a recruitment maneuver had been performed.
Although we used a small number of animals in each group (n = 3/group), the facts that the data were very tight (low standard error) and that we achieved statistical significance in our primary end-point (alveolar stability) suggest that the study had sufficient power to address the the issue considered in the present study. No bias was introduced by a single animal; otherwise, the data would have had a very high standard error.
In this study we considered whether a combination of Vt and PEEP that resulted in alveolar instability cause lung injury. To address this issue, we directly observed subpleural alveoli for stability and, at necropsy, removed the lung tissue that had been observed using the in vivo microscope for histologic analysis. This methodology allowed used to correlate alveolar instability with lung injury and test our hypothesis. However, there were several confounding factors that do not allow us to extrapolate these results readily to the ARDS patient. First, our alveolar sample size was very small and included subpleural alveoli only, so we do not know whether the area of lung sampled is representative of the entire lung. Second, our samples were from nondependent lung areas and our results might have been different in the dependent lung. Finally, we must open the chest to attach the in vivo microscope, and so we do not know whether our findings would have been different if we had been able to obtain the same information with a closed chest.
Tween causes serious lung injury, regardless of the type of mechanical ventilation that the Tween-injured lung is subjected to. Static compliance fell significantly in both groups following Tween instillation and did not significantly recover with time. This suggests that the static compliance was at a nadir following Tween and could not be further reduced by VILI. However, we did observe a significant improvement in partial arterial oxygen tension in the low Vt/high PEEP group, suggesting protection of the lung from VILI.
Cytokines were not significantly increased in this study, which is not consistent with many other experiments demonstrating that high Vt/low PEEP ventilation strategies increase plasma and BAL fluid cytokine levels. This could be for two reasons. Tween is a unique injury model, and other studies demonstrating cytokine increase have used other lung injury models. Also, this experiment was very short, and if we had extended the diuration of the study we might have identified significantly increased cytokine levels. Finally, not all studies have demonstrated that cytokines are released with injurious ventilation [
39], and our findings support this hypothesis.