Many experimental studies (1) have suggested that a patient with acute lung injury or adult respiratory distress syndrome (ARDS) is likely at risk of ventilator-induced lung injury. This injury may relate to supportive care following initial insult to the lung and is thought to arise from alveolar overdistension and conceivably the repetitive opening and closing of lung units associated with tidal volume cycling (1). Strategies to avoid this injury have been proposed, including low tidal volume ventilation, permissive hypercapnea, and “open lung” ventilation. The last ventilator strategy couples a positive end-expiratory pressure (PEEP) level sufficient to prevent alveolar collapse during deflation with tidal volume limitation to avoid overdistension during lung inflation.

Central to bench and clinical studies of ventilator-induced lung injury is the assessment of lung mechanics. At the bedside in particular, the volume–pressure (V–P) relationship of the respiratory system is used as a surrogate measure of the degree of alveolar inflation and recruitment. Remarkably, the correlation of lung mechanics to respiratory system mechanics has not been much questioned, and the factors that might cause patient-to-patient variability have not been much explored. Thus, the report by Gattinoni and colleagues (2) in this issue of the Journal is of considerable importance to the critical care community.

These investigators studied respiratory system mechanics, partitioned between lung and chest wall components, in two groups of patients with ARDS. They prospectively classified patients as having ARDS resulting from either pulmonary disease (ARDS_p, primarily pneumonia) or extrapulmonary disease (ARDS_exp, primarily trauma and intra-abdominal sepsis). The differences between the two groups were notable. At PEEP levels of 0 cm H₂O, both groups had similar respiratory system elastance and similar end-expiratory lung volume. However, patients with ARDS_p had much stiffer lungs than patients with ARDS_exp. When patients with ARDS_p had their PEEP levels increased to 15 cm H₂O, the respiratory system became stiffer yet, with minimal recruitment of lung volume. Patients with ARDS_exp exhibited a much greater chest-wall stiffness than patients with ARDS_p, and this correlated with an elevated intra-abdominal pressure. Increasing PEEP levels in ARDS_exp patients from 0 to 15 cm H₂O decreased respiratory system elastance and resulted in a significant lung recruitment of approximately 0.3 L. The difference in lung and chest wall mechanics between the two patient groups suggested fundamental differences in lung pathophysiology to the authors. They proposed that patients with ARDS_p had substantial lung consolidation with minimal recruitment with PEEP. Patients with ARDS_exp, on the other hand, were suspected to have alveolar flooding and collapse that was more amenable to recruitment with PEEP.

Are these results valid and the interpretations appropriate? The measurements of lung and chest wall mechanics were straightforward. Unfortunately, gas exchange data were not reported, which would have determined whether PEEP effects on mechanics were linked to reduction in intrapulmonary shunt. The classification of patients into two groups was based upon three independent and blinded clinicians' decisions, using history, clinical information, and microbiologic data. The ARDS_p and ARDS_exp groups were clearly different with regard to apparent etiology of lung injury, and both were studied at a similar time in their clinical course to avoid confounding time-related changes in lung function. The only significant demographic difference between the groups was age (ARDS_exp patients were older), which would not likely explain the differences in mechanics. Thus, the differences in mechanics seem quite real and not well explained by anything other than the cause of lung failure. The speculation that these differences arise from different lung histopathologies, however, is just that—speculation. Future studies would advance our knowledge in this area by correlating differences in lung mechanics between such groups to either lung histology, a difficult enterprise in a clinical investigation, or to lung imaging.

How do these data relate to other investigations in similar patients? Ranieri and colleagues (3) compared patients with ARDS consequent to major abdominal surgery (“surgical” ARDS) to patients with “medical” ARDS. In surgical ARDS, the static inspiratory V–P relationship of the respiratory system and lung showed a downward concavity, indicating that elastance increased with tidal volume, suggesting alveolar overdistension. Patients with ARDS consequent to medical conditions had an upward concavity on the static inspiratory V–P curves of the respiratory system and lung, indicating a decrease in elastance and alveolar recruitment during inflation. PEEP effects were not examined. Surgical ARDS patients who underwent operative abdominal decompression for bleeding exhibited an upward and leftward shift of the lung, chest wall, and respiratory system V–P curves.

There is at least superficial similarity between the groups being compared in these two studies: Both appear to roughly compare surgical ARDS to other conditions. Thus it is notable that Gattinoni's ARDS_p patients did not exhibit lung recruitment with PEEP (2), while Ranieri's medical patients exhibited an upward concavity of the static V–P inflation curve, suggesting lung recruitment during inflation (3). It is difficult to say a great deal more about these differences because the maneuvers are different and a number of patients in the Ranieri study with medical ARDS (drug overdose, urosepsis) would have been classified as having ARDS_exp in the Gattinoni study.

Margoni and colleagues (4) studied 13 patients with acute respiratory failure, partitioning respiratory system mechanics and describing PEEP effects on mechanics and gas exchange. These investigators showed that in a subset of patients with ARDS the lower inflection point (LIP) of the static V–P curve of the respiratory system was due to the shape of the V–P curve of the chest wall and not the lung. When this was true, PEEP was not as effective in improving Pa_O2 , suggesting that in these patients the LIP is not a measure of alveolar recruitment. Again, a comparison with the Gattinoni study is made difficult by the inability to match patient groups with much confidence.

How do these investigations inform our understanding of recent trials of lung-protective ventilator strategies? A single but prominent study by Amato and colleagues (5) has demonstrated that open-lung ventilation is associated with improved outcome, including decreased mortality, in patients with ARDS. A large fraction of patients in this study were, by clinical description, similar to the ARDS_p patients described by Gattinoni. Yet the lung mechanics of the patients reported by Amato appear to parallel the medical patients reported by Ranieri and not the ARDS_p patients. Other reports have not confirmed a benefit of low tidal volume ventilation in patients with ARDS (6-8), but detailed patient characteristics of either the etiology of ARDS or the respiratory system mechanics observed have not been reported. Thus, post hoc comparisons of patient groups in these outcome studies to the detailed studies of mechanics is difficult, if not impossible.

Finally, how may we best use this information for current clinical practice or future clinical studies? The lessons from these recent investigations of respiratory system, lung, and chest-wall mechanics in patients with ARDS are these:

1. Lung mechanics in patients with ARDS cannot be confidently inferred from measurement of respiratory system mechanics at the mouth, and partitioning of separate lung and chest wall effects is advisable.

2. Upper and lower inflection points on the V–P curve must be interpreted in light of contributions of deranged chest wall (including abdominal) mechanics, and hence may not simply reflect properties of the injured lung.

3. Overdistension of the lung cannot be predicted confidently by an end-inspiratory pressure, particularly when chest wall or abdominal elastance is markedly increased.

4. Lung recruitability may be influenced by the nature of the lung insult itself, with lung consolidation perhaps somewhat refractory to PEEP maneuvers.

These lessons have intuitive appeal and make one wonder why such obvious implications of basic respiratory system mechanics had not been demonstrated earlier. The partitioning of lung mechanics in these studies required esophageal manometry to estimate pleural pressure, but this measurement has become a reasonably standard clinical tool. It is certainly premature to recommend routine measurements at the bedside to partition lung mechanics in individual patients, although clinicians who are actively employing ventilator strategies to limit alveolar distension should be aware that simple measurement of plateau pressure at end inspiration is likely to overestimate the true transpulmonary pressure when abdominal distension and other causes of chest wall stiffness are present. Most importantly, as future studies are designed to further elucidate the clinical significance of ventilator-induced lung injury and its prevention, these principles should help researchers design protocols to define patient groups more precisely and select appropriate methods for measuring mechanics at the bedside.

References

Section:

ChooseTop of pageReferences <<CITING ARTICLES