beyond 6 ml/kg: can we personalize lung protective ventilation? a CCN Re-post

Note: This post initially appeared on in its original form. It is shared here as part of an ongoing conversation about how to make mechanical ventilation work for the patient, rather than vice versa. See our other vent content on CCN here!


A tale of two patients, the same age, sex, and height, who are intubated due to respiratory failure, subsequently diagnosed with ARDS. Both are placed on a tidal volume of 6 ml/kg ideal body weight (420 ml in this case), with PEEP 12. Initial ABG: 7.24/55/90 on 0.5 FiO2

Patient A has a plateau pressure of 30.

Patient B has a plateau pressure of 20.

Should/could they be managed differently? How would you decide?

This conversation focuses on the recent work “Effect of Lowering Tidal Volume on Mortality in ARDS Varies with Respiratory System Elastance” by Goligher, Costa, and colleagues.

Study Summary:

  • This work is predicated on the driving pressure hypothesis, popularized by the landmark paper by Amato and colleagues. In that analysis, driving pressure was the ventilation parameter that best stratified mortality risk, at a cutoff of 15 cm H2O.
    • As a reminder, driving pressure is defined as Plateau minus PEEP.
  • The authors sought to determine if the mortality benefit of low tidal volume ventilation (LTVV) differs based on respiratory system elastance.
    • As a reminder, elastance is calculated as (Plateau pressure – PEEP)/Tidal volume; it could also be written as Driving Pressure/Tidal Volume. It is the inverse of compliance. 
    • In this study, elastance was normalized to the predicted body weight (PBW). Healthy patients typically have a normalized elastance < 1 cm H2O/(ml/kg).
  • This work is a secondary analysis of five ARDS trials that tested high vs low tidal volume strategies. In total, data on over 1,000 patients was analyzed.
  • Using Bayesian multivariable logistic regression, the posterior probability that 60-day mortality benefit from LTVV varies with elastance was 93%. 
    • This model was adjusted for P:F ratio, severity of illness scores, and the mortality rate of the control group of each original study.
    • The posterior probability of absolute risk reduction in mortality at least 5% rose with increasing elastance, from 29% in the “low” elastance (<2 cmH2O/(ml/kg)) group, compared to 82% in the high (>3 cmH2O/(ml/kg)) group.
    • Notably, the mortality benefit of LTVV did not vary with P:F ratio!
  • Important limitations of this study include the age of the trials (over 20 years old at this point) and the post-hoc design. 

Putting it into context

What does this mean for 6 ml/kg tidal volumes in ARDS?

  • There is consistent (observational) data about the prognostic value of driving pressure. What has been less clear, though, is if it is a reasonable therapeutic target.
  • The design of this study suggests a potential causal relationship of LTVV on mortality, dependent on elastance. The easiest way to conceptualize this at the bedside is observing the driving pressure at the current set tidal volume, and making adjustments accordingly.
  • Blanket application of a set tidal volume for all ARDS patients, regardless of underlying physiology, probably never made sense. We are overdistending some lungs, and probably being too restrictive in others.
  • For some patients, 6 ml/kg may still not be “lung protective” if the delivered tidal volume exceeds the amount of lung available for ventilation. This analysis suggested that a driving pressure > 15 cm H2O (a high elastance situation) should make one consider lowering the tidal volume below 6 ml/kg PBW. The excess tidal volume delivered would likely cause tidal hyperinflation of the aerated lung, more than any potential recruitment of the atelectatic/flooded alveoli.
  • Conversely, a patient with low elastance and driving pressure < 15 may be able to tolerate careful increases in tidal volume. Potential benefits could include decreased dyssynchrony and decreased need for sedation and neuromuscular blockade. The authors suggest (and I agree) that as long as the driving pressure is in a safe range, this strategy is supportable, and physiologically rational.

It’s worth emphasizing again that the P:F ratio did not predict response to LTVV. Future intervention trials may be better served by stratifying patients based on driving pressure rather than P:F ratio.

Back to the cases!

Assuming predicted body weight of 70 kg for both patients:

Patient A has a normalized elastance of 3 cm H2O/(ml/kg)

  • The patient with high elastance (low compliance) has an elevated driving pressure (18). 
  • Lowering the tidal volume (to as low as 4 ml/kg PBW, in a stepwise fashion) should be considered, with the goal of decreasing driving pressure to less than 15 cm H2O.
    • The physiologic goal here is to scale the tidal volume delivered to the amount of lung available for ventilation. 
    • Of note, elastance could also be decreased by PEEP titration based on best driving pressure. There is no “best practice” as to whether one should titrate PEEP or tidal volume first, but both may be necessary. Each approach carries its own risks and benefits.
  • As the fraction of dead space is expected to increase when tidal volume is decreased, pay careful attention to PaCO2 levels. Permissive hypercapnia is standard of care in ARDS, but in certain patients it can predispose to complications such as acute cor pulmonale.
    • Extracorporeal CO2 removal (ECCO2R) is a physiologically plausible intervention to facilitate these “ultra lung protective” tidal volumes, but further study is needed. Trials are in progress.    

Patient B has a normalized elastance of 1.3 cm H2O/(ml/kg)

  • This patient with lower elastance (higher compliance) has an acceptable driving pressure (8).
  • It would be reasonable to carefully liberalize the tidal volume, which may decrease sedation requirements and dyssynchronies. 
    • This strategy may promote comfort, mobility, and perhaps decrease need for other interventions (neuromuscular blockade use, use of vasopressors to mitigate sedation-related hypotension). 
  • If the tidal volume is liberalized, the available data suggest that the driving pressure should still be kept below 15 cm H2O.


  • A secondary analysis of ARDS clinical trials by Goligher, Costa and colleagues demonstrates that low tidal volumes likely have the most mortality benefit when elastance is highest (i.e., compliance is lowest). This supports a driving pressure strategy.
  • If a driving pressure strategy is chosen, a target of < 15 cm H2O should be the used
  • Potential benefits of a driving pressure and elastance-driven strategy include: 
    • Better scaling of tidal volume to available lung in patients with high elastance 
    • Ability to provide more liberal tidal volumes to patients with low elastance, thereby (likely) decreasing need for deep sedation, neuromuscular blockade, etc.
  • This rationale likely explains why post-operative patients with normal pulmonary parenchyma (and thus normal elastance) can safely tolerate higher tidal volumes, as the driving pressure is expected to be low in these cases. 
  • The available data suggest that driving pressure may be a better target (compared to P:F ratio, or plateau pressure alone, for example) to titrate ventilation parameters such as tidal volume and PEEP. However, randomized clinical trial data is lacking.
  • Addendum: I need to emphasize that best evidence still supports tidal volumes of 4-8 ml/kg IBW. This framework, in my estimation, is still helpful in tidal volume selection within that range. Moving outside this range is definitely evidence-limited.




  1. How does one incorporate the “hidden“ pleural pressure swings into the assessment of driving pressure? Are we calculating elastance with spont effort aborted? Similarly, do you think that a non intubated patient with a hypothetical pleural pressure of greater than -15 is at risk of p-sili compared to one of, say -8?

    Liked by 1 person

    1. Thank you for the excellent comment. In these trials (and in my practice), driving pressure is assessed with a passive patient.

      The PSILI question is so important, but the data are still unclear. In your example, I do think the more neg pleural pressure patient is at higher risk. We need to study this phenomenon better!


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s