Mechanical ventilation is critical for survival from acute respiratory distress syndrome (ARDS). However, mechanical ventilation can also cause ventilator-induced lung injury (VILI), which may delay or prevent recovery (1, 2). One of the causes of VILI is excessive lung stretch during inspiration (3, 4).

Some investigators and clinicians recommend monitoring end-inspiratory plateau pressure (Pplat) as a surrogate for maximal lung stretch. Smaller tidal volumes can reduce Pplat and stretch-induced VILI, and this approach was associated with improved clinical outcomes in patients with ARDS (5). Pplat is easy to measure and is directly related to the risk of death (6). Patients with more severe disease tend to have higher Pplat, and they may benefit from more aggressive tidal volume reduction to control stretch-induced VILI. However, Pplat reflects chest wall as well as lung mechanics. In patients with increased chest wall elastance or with large pleural effusions, Pplat may be elevated even though lung stretch is not high. A better way of monitoring lung stretch is required.

In this issue of the Journal (pp. 346–355), Chiumello and colleagues present a study examining lung stress and strain in four groups of subjects, including those with ARDS (7). To a mechanical engineer, stress is the force per unit of cross-sectional area of a structure subjected to such a force (8). Chiumello and colleagues calculated stress as the difference in transpulmonary pressure (P_L, the formal units of which are force/area) between functional residual capacity (FRC, with zero positive end-expiratory pressure [PEEP]) and end-inspiration. A simple expression of strain is the ratio of the increase in length of a structure divided by its resting length (8). For the three-dimensional lung, the investigators calculated strain as the change in lung volume between FRC and end-inspiration normalized to FRC. Estimates of lung stress and strain varied as predicted among subject groups. Moreover, at any Pplat, estimates of lung stress (ΔP_L) and strain varied substantially within each patient group. In patients with ARDS with Pplat = 30 cm H₂O, ΔP_L varied from approximately 10 to 30 cm H₂O because elastances of both lung and chest wall varied widely.

It was beyond the scope of this study to define the specific thresholds for “harmful” stress and strain, but the authors suggested a ΔP_L of approximately 27 cm H₂O and a strain of 2 as possible critical limits. The key question arising from this work is whether use of stress or strain as defined would be better than use of Pplat for guiding ventilator management.

We share the authors' belief that the use of predictors of stretch-induced VILI that are independent of chest wall mechanics would be ideal. However, we think there are some limitations to the proposed approaches for measuring stress and strain. First, a better indication of end-inspiratory stress would be the absolute value of P_L at end-inspiration. Unfortunately, artifacts in esophageal pressure, especially in supine, critically ill patients, make it very difficult to measure absolute P_L accurately (9–12). By using the difference in P_L between FRC and end-inspiration, the artifacts could arithmetically cancel out in the calculation of ΔP_L. Assuming that the artifact in esophageal pressure is constant at the two lung volumes and that stress at FRC is low, ΔP_L could be a fair estimate of end-inspiratory stress. Another concern regarding ΔP_L is that esophageal pressure is measured in only one location, but changes in pleural pressure during inspiration may vary depending on location in the chest. Thus, a single value of ΔP_L may not adequately reflect stress in some lung regions.

As discussed by Chiumello and colleagues, there are important implications in the estimates of strain due to lung recruitment during inspiration. Without taking such factors into account, strain may be overestimated. As aptly expressed by the authors, some patients would be condemned to very low tidal volumes or PEEPs, similar to what would happen if clinicians simply monitored Pplat.

The proposed expressions of end-inspiratory stress and strain in this study are the same at any given lung volume, regardless of how the lungs arrive at that lung volume. They do not account for the “trajectory” to end-inspiration, which also affects pathogenesis of VILI. At a given level of end-inspiratory lung stretch, VILI is probably greater with larger tidal increments in volume and pressure. This has been demonstrated in several experimental models, perhaps most dramatically in the classic study of VILI by Webb and Tierney (13): injury to the lungs of rats ventilated with a peak inspiratory pressure of 45 cm H₂O and no PEEP was much greater than in lungs of rats ventilated with the same peak inspiratory pressure but with 10 cm H₂O PEEP. Strain as defined by Chiumello and colleagues does not account for VILI from mechanical ventilation at low end-expiratory volumes and pressures.

Clinicians need reliable targets for safe limits of stress or strain. We anticipate that safe limits will vary with age, time on mechanical ventilation, presence of inflammatory conditions such as pneumonia and sepsis, and trajectory to end-inspiration. More information is needed to define safe limits for stress and strain.

The work by Chiumello and colleagues is technically and intellectually stimulating. The authors have elevated our thinking and challenged us to utilize better, more rational approaches to ventilator management. But we need refined methods for estimating stress and strain and more information regarding safe limits under various physiologic and pathologic conditions before we can apply these ideas to the clinical management of our patients.

References

Section:

ChooseTop of pageReferences <<CITING ARTICLES