Driving Pressure: Critical Respiratory Variable
The authors of a recent study published in the New England Journal of Medicine were able to show that driving pressures have a greater effect on mortality than any other individual respiratory variable. The driving pressure was calculated by subtracting positive end-expiratory pressure (PEEP) from plateau pressure in patients who were not spontaneously breathing. In practical terms, the driving pressure can be thought of as the force imposed on the healthy areas of the lung during a tidal breath. The more pressure above PEEP that is required to achieve tidal volume, the more cyclic strain the lung will experience.
Amato and colleagues conducted their analysis by combining patient-level data from nine randomized controlled trials enrolling patients with acute respiratory distress syndrome (ARDS). They used mediation analysis to isolate and quantify the effect of driving pressure on mortality. They adjusted for mortality risk (using the Acute Physiologic Score and Chronic Health Evaluation [APACHE] II or the Simplified Acute Physiologic Scale [SAPS], depending on which score the individual trial used) and severity of lung injury (using blood gas values and measures of tidal elastance). They concluded that driving pressure matters more than low tidal volume ventilation (adjusted for body weight) or PEEP. Strain is the issue, and changes to tidal volume or PEEP are unlikely to be effective if they don't result in a driving pressure reduction.
The physiologic concept that these investigators are promoting is not new, but it is appealing. It fits with the "baby lung" concept and gives physicians a target respiratory variable that can easily be measured at the bedside, assuming that the patient isn't tachypneic or breathing spontaneously. It provides a nice, tidy explanation for why PEEP trials have been inconsistent. Higher PEEP only helps if it recruits lung and increases compliance. If it does, the driving pressure required to achieve a given tidal volume will go down and there will be less cyclic strain on the lung. If it doesn't, the opposite can occur. In part, these investigators have found yet another bedside method for measuring recruitable lung and properly setting PEEP.
Caveat for Bedside Clinicians
As always, though, there are caveats. First, they don't define a target driving pressure above which strain is more likely to occur, because they have entered driving pressure into their models as a continuous variable. This is good in that it provides a more accurate representation of the relationship between driving pressure and mortality (for each 7-cm H2O increase in driving pressure, clinicians can expect the odds of death to increase by 1.3-1.5). Unfortunately, clinicians (and all humans, really) have trouble translating such relationships into reality. We prefer to be given a target below which there is no excess mortality (like plateau pressure < 30 cm H2O or tidal volume of 6 cc/kg). Without such a threshold value, whether artificial or not, how does one set the driving pressure at the bedside?
The issue of spontaneous breathing is also troublesome. They excluded such patients from their analysis because plateau pressure is difficult to measure in a spontaneously breathing patient and because the generation of negative pleural pressure makes compliance and elastance of the respiratory system impossible to measure. The resulting increase in variability would likely weaken the association between driving pressure and mortality. Clinicians measuring driving pressure at the bedside face a dilemma, however: Do they sedate their patients to obtain measurements? If so, do they keep them sedated and risk delayed extubation or do they allow the patients to breathe on their own and potentially lose the predictive capacity that comes with driving pressure measurement?
In conclusion, this study is important. I'll be measuring driving pressure at the bedside while anxiously awaiting the randomized controlled trial that is sure to come. The issues with spontaneous breaths and compliance/elastance certainly aren't specific to this measurement. They have been a source of frustration and uncertainty for some time. Clinicians should factor in the driving pressure when they are setting PEEP and tidal volume.