Early application of airway pressure release ventilation may reduce the duration of mechanical ventilation in acute respiratory distress syndrome

Zhou. Intensive Care Medicine 2017; 43:1648-1659. doi:10.1007/s00134-017-4912-z

Clinical Question

  • In mechanically ventilated patients with ARDS, does airway pressure release ventilation (APRV) compared to conventional low tidal volume ventilation (LTV) reduce duration of mechanical ventilation?


  • Acute respiratory distress syndrome (ARDS) is an inflammatory process featuring increased lung vascular permeability resulting in hypoxaemia and reduced lung compliance
  • Conventional ventilation is based upon the ARMA trial, which has become known as ‘ARDSnet’ or low tidal volume ventilation (LTV)
  • Small trials and animal studies have suggested that longer periods within the ventilation cycle at high pressure with intermittent release to low pressure, known as airway pressure release ventilation (APRV), may be beneficial by improving alveolar recruitment, gas exchange and reduced lung injury


  • Randomised, controlled trial
  • Ethics approval in accordance with Helsinki Declaration with informed written consent
  • Trial registered as “BILEVEL-APRV” a priori with no change in primary outcome: NCT02639364
  • Randomisation sequence generated by computer in block design
  • Concealment of allocation by opaque envelopes
  • Sample population drawn from consecutive patients
  • Allocation not blinded (“Open-label” design)
  • Power calculation based upon previously published data
    • LTV 14.5 ventilator-free days
    • Expected absolute difference with APRV +5 days
    • Powered at 80%
    • Significance set at 0.05
    • 138 patients required, which includes allowance for drop-outs
  • Statistical analysis using appropriate tests


  • Single centre study in Sichuan province, China
  • May 2015 – October 2016


  • Inclusion:
    • Diagnosis of ARDS according to the Berlin Criteria
    • PaO2 : FiO2 < 250 mmHg (< 33.3 kPa)
    • Endotracheal intubation and mechanical ventilation for < 48 hours
  • Exclusion:
    • Anticipated mechanical ventilation < 48 hours
    • Neuromuscular disorders that prolong mechanical ventilation need
    • Intracranial hypertension (suspected or proven)
    • Severe COPD
    • Treatment with ECMO at enrollment
    • Refractory shock
    • Known barotrauma
    • Age <18 or > 85 years
    • Pregnancy
    • Less than 6-month predicted life or lack of commitment to life support
  • 251 screened; 138 patients randomised; 138 in intention-to-treat analysis; 118 in the per-protocol analysis
  • Baseline characteristics between groups, including reason for ARDS, were not significantly different except for the presence of coexisting disease (APRV vs LTV)
    • Age: 51.5 vs 52.0 years
    • APACHE 2: 22.0 vs 20.2
    • Diagnosis of cancer: 9.9% vs 17.9% (P = 0.22)
    • Any coexisting disease: 32.4% vs 50.7% (P = 0.04)
    • Pneumonia as cause of ARDS: 25.4% vs 38.8% (P = 0.10)


  • Airway Pressure Release Ventilation (APRV)
    • Initiation: patients were transitioned from previous Volume Assist-Controlled Ventilation (VCV) to APRV
      • Phigh was set to previous Pplat measured, and not exceeding 30 cmH2O
      • Plow was set a 5 cmH2O
      • Tlow (the duration at the lower pressure) was set at 1.0–1.5x expiratory time constant (TE)
        • TE is the time required for 63% volume to be expelled
        • TE = Resistance of the respiratory system (Rrs) x Static compliance (Cstat)
          • Rrs and Cstat are measured during an inspiratory pause
        • ~3x TE is the time required to expel full volume down to residual volume
        • Tlow was subsequently adjusted so that expiratory flow rate was 50% peak expiratory flow rate
      • Release frequency was set at 10–14 /min
      • Thigh was dependent upon Tlow and release frequency
      • Initial spontaneous respiratory level was targeted as 30% of total minute ventilation
    • Weaning: if cardiopulmonary function was stable and sedation was adequate, the ventilatory support was reduced twice daily
      • Phigh was reduced by 2 cmH2O
      • Release frequency was reduced by 2 /min
    • Termination: patients underwent a spontaneous breathing trial when:
      • Phigh 20 cmH2O or less
      • FiO2 40% or less

An example of APRV.
By Je.rrt (Own work) [CC BY-SA 3.0 (], via Wikimedia Commons


  • Low Tidal Volume Ventilation (LTV)
    • Initiation: patients were continued on the Volume Assist-Control Ventilation mode (VCV)
      • Tidal volume (VT) target was 6 ml/kg predicted body weight
      • Acceptable range for VT was 4–8 ml/kg
      • PEEP level was adjusted according to FiO2 using low-PEEP table from ARMA trial
        • If ARDS was severe (PaO2:FiO2 < 100 mmHg) then PEEP could be titrated according to optimal oxygenation, optimal compliance or clinician’s discretion
      • Ventilation frequency was set to maintain target pH according to ARMA trial
    • Weaning: PEEP and frequency were titrated down as oxygenation and ventilation improved as outlined above
      • Daily sedation breaks occurred if Richmond Agitation Sedation Scale score was less than -2
    • Termination: patients underwent a spontaneous breathing trial when:
      • FiO2 40% or less
      • PEEP 8 cmH2O or less
      • PaO2 60 mmHg (8 kPa) or more
      • Acceptable spontaneous breathing effort
      • Systolic BP 90 mmHg or more with not more than
        • Dobutamine 5 µg/kg/min
        • Noradrenaline 2 µg/kg/min
      • No neuro-muscular blocking drugs

Management common to both groups

  • All patients were initially ventilated with volume assist-control ventilation (VCV) according to the LTV strategy using a Puritan Bennett 840 Ventilator from Covidien
  • Additional measures that could be used in severe hypoxia were:
    • Recruitment manoeuvres
      • LTV: incremental PEEP increase up to 40 cmH2O
      • APRV: Phigh and Plow increased incrementally until Phigh 40 cmH2O
    • Prone positioning
    • Neuro-muscular blockade
    • Inhaled nitric oxide
  • spontaneous breathing trials were:
    • 30 minutes
    • PEEP 5 cmH2O
    • Pressure support ventilation 5–7 cmH2O
    • FiO2 40%


  • Primary outcome: the number of ventilator-free days up to day 28 was significantly greater in the APRV group compared to the LTV group
    • Median ventilator-free days:
      • APRV group: 19 [IQR 8–22]
      • LTV group: 2 [IQR 0–15]
      • P = < 0.001
  • Secondary outcome:
    • APRV resulted in significantly better markers of oxygenation and ventilation, such as better PaO2:FiO2 ratio, better lung compliance, lower Pplat and Ppeak
    • APRV resulted in significantly lower heart rate and higher blood pressure
    • APRV resulted in less sedation depth (i.e. more awake patients) with lower doses of opiates and benzodiazepines (propofol doses were similar)
    • Successful extubation rates were higher in the APRV group, and fewer tracheostomies occurred
    • ICU length of stay was shorter in the APRV group
    • Mortality and hospital length of stay were similar between the groups
    • Rescue measures for severe hypoxia occurred more frequently in the LTV group

Authors’ Conclusions

  • Early application of APRV was associated with better oxygenation, less sedation, fewer days of mechanical ventilation and shorter ICU stays


  • Interesting theory lacking substantial evidence so far
  • Reasonable hypothesis and appropriate methodology to test it
  • Randomised, controlled trial minimises bias from confounding factors and provides best evidence from testing two interventions
  • Appropriate randomisation process
  • Consecutive sampling reduces selection bias – clinicians could not choose who to include in the trial as all patients were considered
  • Adequate concealment using opaque envelopes, which minimises selection bias, but web-based system would have been better
  • Clear protocolised care reduces possible bias from open-label design – clinicians knew which group the patient was in but treatment was directed by objective protocols


  • Single centre conduct limits certainty about validity of results and prevents meaningful generalisation of those results – generally repeat studies with multi-centre designs find a small effect from the intervention
  • Important baseline difference exist (by chance) – the APRV group had less coexisting chronic disease, which may have added to the size of the observed effect
    • Multiple linear regression analysis was conducted to statistically correct for this and the authors report the conclusion remained the same
  • Open-label design introduces possible allocation bias as treatment unrelated to the hypothesis can be different between groups due to clinician awareness of allocation – this tends to lead to exaggerated effect size
  • The intervention and control group protocols differed in numerous ways and not just the specific ventilator settings
    • This included high PEEP recruitment manoeuvres (such as those found to be harmful in the ART Trial), weaning strategies, thresholds for spontaneous breathing trials
    • It could be said that this trial compares a ventilation protocol that includes APRV vs a ventilation protocol that includes LTV
    • Many centres no longer use LTV in the way this protocol specified, which may reduce the generalisability of the results
    • Whilst this doesn’t reduce the validity of the conclusion, it does imply that other centres should not expect to see improvement in patient outcomes from simply changing the ventilator mode
  • Respiratory therapists implemented all ventilator associated therapies, which may not be standard practice in all Intensive Care Units
    • This reduces the generalisability of the results
    • Such large differences may not be observed if the same intervention is performed by Critical Care doctors and nurses rather than dedicated respiratory therapists

The Bottom Line

  • The reported benefits of APRV are clinically significant and the conduct of the trial is reasonable enough to conclude the results are probably valid
  • However, given the single centre design and the extensive use of respiratory therapists, I would like to see a multi-centre, multi-national trial before concluding APRV is far better than LTV
  • For now, I shall consider APRV in selected patients with severe ARDS

External Links


Summary author: Duncan Chambler
Summary date: 17 January 2018
Peer-review editor: Segun Olusanya

Leave a Reply