Diaphragm-Protective Mechanical Ventilation to Improve Outcomes in ICU Patients?

The respiratory muscle pump drives alveolar ventilation. The diaphragm, a thin (2–3 mm in healthy adults) dome-shaped muscle, is an essential component of the respiratory muscle pump. In critically ill patients, ventilatory demands may exceed the capacity of the respiratory muscle pump, and in this case mechanical ventilation is used to unload the respiratory muscles to prevent the development of muscle fatigue and injury. However, in the past decade, several studies have provided evidence that respiratory muscle weakness may develop in ICU patients on ventilator support (1–4). In fact, diaphragm dysfunction is highly prevalent. For instance, Dres and colleagues (5) demonstrated that at the time of the first spontaneous breathing trial in patients who had undergone mechanical ventilation, 63% of the patients had developed diaphragm dysfunction, which has recently been termed “critical illness–associated diaphragm weakness” (6).

In this issue of the Journal, Goligher and colleagues (pp. 204–213) present their findings from a large prospective cohort study in invasively ventilated ICU patients (7). This study was designed to examine whether changes in the thickness of the diaphragm, as assessed by ultrasound, were associated with adverse outcomes, including prolonged ventilator dependence, reintubation, and death. At least two thickness measurements were obtained in 191 ventilated patients. A rather complex statistical plan was used, but the main message of the study can be summarized as follows: diaphragm muscle thickness decreased by ≥10% in 41% of patients by median day 4, and this was associated with a lower probability of ventilator liberation, prolonged ICU admission, and respiratory complications (including reintubation and tracheostomy) compared with patients with a <10% change in diaphragm thickness. Remarkably, in 24% of the patients, diaphragm thickness increased by ≥10% in the first week of ventilation, and this was also associated with prolonged ventilation. The authors also demonstrated that the change in diaphragm thickness varied with diaphragm effort as assessed by muscle-thickening fraction or diaphragm electrical activity: low effort was associated with reduced thickness and high effort was associated with increased thickness.

This study is very important for a number of reasons that should be emphasized. First, this is the first prospective study with a large number of patients to report a relationship between changes in diaphragm muscle thickness and clinical outcomes. Second, this study demonstrates a relationship between changes in diaphragm thickness and diaphragm effort. The development of diaphragm atrophy under conditions of disuse or low effort is rather intuitive and was previously demonstrated by analyses of diaphragm muscle biopsies from ventilated, brain-dead patients (1, 2). The pathophysiology and histological substrate for increased diaphragm thickness during mechanical ventilation are less intuitive, but the fact that increased diaphragm thickness resulted in a lower cumulative incidence of ventilation liberation suggests that the increased thickness is not just a reflection of diaphragm hypertrophy (i.e., increased contractile tissue in the muscle fibers). An alternative hypothesis is that increased diaphragm thickness results from diaphragm injury. The authors propose that the high breathing effort induced diaphragm muscle fiber injury, resulting in an inflammatory response and fiber swelling. This is an intriguing proposition, and is supported by a previous study in humans that showed excessive diaphragm fiber injury after acute inspiratory muscle loading (8), and another study that demonstrated elevated numbers of inflammatory cells in diaphragm fibers of critically ill patients (3). In the latter study, increased expression of inflammatory cells was also present in the diaphragm of nonseptic patients, suggesting that this does not reflect systemic inflammation but rather a local inflammatory response, possibly elicited by diaphragm injury. Further evaluation of these findings, however, does not show a positive correlation between the number of neutrophilic granulocytes and macrophages in the diaphragm and the cross-sectional area of fibers in those specimens (n = 14; unpublished observations). The absence of such a correlation does not seem to support the proposition that increased local inflammation causes fiber swelling. Furthermore, it remains to be established that fiber swelling occurs, as previous analyses of diaphragm fiber cross-sectional areas in critically ill patients did not provide evidence in support of swelling (3, 9). Whether diaphragm injury in critically ill patients causes interstitial fluid accumulation (i.e., swelling of the space in between muscle fibers) and thus provides an alternative explanation for increased diaphragm thickness remains to be determined and warrants further investigations.

Taken together, these studies demonstrate that diaphragm structure and function are very sensitive to high and low loading conditions, both of which have been observed in ICU patients (2, 10) and may occur at different time points during the ICU stay (Figure 1). Goligher and colleagues (7) now convincingly demonstrate that this has an important impact on patient outcomes. Based on these data, it seems reasonable to conclude that prevention of disuse and high breathing effort for prolonged periods of time is the cornerstone of diaphragm-protective mechanical ventilation. It may be possible to target physiological loading of the diaphragm by adjusting the level of inspiratory ventilator support and/or dose of sedatives. However, this requires monitoring of the respiratory muscle effort. Several techniques are feasible in critically ill patients, including monitoring of esophageal pressure, diaphragm electromyography, and ultrasound (11, 12). The relevant question is, which level of activity can be considered diaphragm protective? A pragmatic approach would be to aim for values reported for healthy subjects (e.g., when using the esophageal balloon, a muscle pressure of 5–10 cm H₂O [13] or diaphragm electrical activity of 8% of maximum value [14]). The data from Goligher and colleagues (7) indicate that an inspiratory thickening fraction of ∼40% prevents changes in diaphragm thickness.

Figure 1.

In patients with a critical illness, the respiratory muscle pump may exert a level of effort beyond its physiological limits for hours or even days. Before ICU admission, patients with impending respiratory failure may exhibit very high respiratory muscle effort. After endotracheal intubation and institution of controlled mechanical ventilation, the respiratory muscles are at risk for disuse atrophy. In subsequent days, both high and low breathing efforts may occur, particularly if the activity of the diaphragm is not monitored. This may result in functional and structural modifications of the diaphragm muscle, especially in a hostile environment due to hypercapnia, hypoxemia, and inflammation.

Obviously, more studies are required to identify the optimal range of diaphragm effort in critically ill patients under different clinical conditions, and, more importantly, to confirm that a treatment algorithm that uses diaphragm monitoring can improve outcomes. But let us not forget that even with regard to arterial blood pressure, we are still debating optimal targets for specific patient populations (15), despite the fact that intra-arterial catheters have been used for decades. Goligher and colleagues provide important evidence that both high and low levels of breathing effort are associated with adverse outcomes, and therefore, in addition to lung-protective ventilation, we now may incorporate diaphragm-protective ventilation into our clinical protocols. Undoubtedly, future studies will provide better guidance and even personalized target values. For now, let not better be the enemy of good.