OSCAR Study Group: High-Frequency Oscillation for Acute RespiratoryDistress Syndrome

Young et al for the OSCAR Study Group. N Engl J Med 2013; 368:806-813.

Clinical Question

  • Does High Frequency Oscillatory Ventilation (HFOV), compared with conventional ventilation, improve 30 day mortality in patients with ARDS?


  • Randomised controlled trial
  • Permuted block randomisation
  • Stratified by baseline characteristics


  • 27 general ICUs in UK with varying experience of HFOV
  • Dec 2007 – July 2012


  • Mechanical ventilation for <7 days; and ARDS defined by all of the following:
    • PaO2/FiO2 <200mmHg (26.7 kPa)
    • PEEP >5
    • Bilateral pulmonary infiltrates, without evidence of left atrial hypertension

The Definition of ARDS has since been revised and this entry criteria match the draft Berlin definition of ‘moderate’ and ‘severe’ ARDS. The lastest Berlin Definition uses PEEP > 10cm water as part of its criteria for severe ARDS 

  • 795 patients randomised (28.7% of the 2769 patients screened)
  • Exclusions: Ventilated for > 7 days; < 16 years of age; weighed < 35 kg; particpating in other research trials; lung disease characterised by airway narrowing or gas trapping’ recent lung surgery


  • HFOV using Novalung R100 ventilator (Metran) until the start of weaning.
  • Initial settings were:
    • ventilation frequency of 10 Hz
    • mean airway pressure of 5 cm water above the plateau airway pressure at enrollment
    • bias flow rate of 20 litres per minute
    • cycle volume of 100 ml (the volume of gas used to move the oscillating diaphragm; the tidal volume delivered to the alveoli is a fraction of this volume)
    • FiO2 1.0
    • Fixed I:E ratio 1:1
  • Algorithm for weaning. If patients reached a mean airway pressure of 24 cm water, at an FiO2 0.4 or less with a paO2 level of 60mmHg (8 kPa) or more for at least 12 hours, they were switched to pressure-controlled ventilation for weaning from mechanical ventilation
  • Patients could be restarted on HFOV up to 2 days after the start of weaning


  • Ventilated according to local practice.  
  • Encouraged to use ARDSnet style ventilation (different in OSCILLATE where this was mandated) i.e pressure-controlled ventilation at 6-8 ml/kg of ideal body weight with combinations of PEEP and FiO2 values used in the ARDSnet study


  • Primary Outcome: Mortality at 30 days
    • 41.7% in HFOV vs. 41.1% in control. Absolute difference 0.6% (95% C.I. -6.1 to 7.5)
  • Oxygenation improved with HFOV when compared with conventional ventilation
    • PaO2:FiO2 ratio (mmHg) on Day 2: 212 vs. 163

Authors’ Conclusions

  • No benefit or harm was found from the use of HFOV in adult patients with ARDS. HFOV should not be used for routine care


  • Multi-centre
  • Intention-to-treat
  • Independent telephone randomisation system
  • Appropriate power calculation
  • APACHE II mean score of 21.8 i.e sick patients


  • 2/3rds of patients screened were not recruited. It is unclear why in a 1/3 of the cases (‘had other reasons’)
  • Slow recruitment. Average 0.82 patients per ICU month
  • 20 of the 29 units had not previously had any experience with using HFOV. However > 2000 people were trained before recruitment started and reflective of UK practice at the time
  • ARDSnet ventilation was encouraged rather than mandated
  • The Novalung R100 ventilator was used in the HFOV group. This has not been used before in clinical trials. Other studies (including OSCILLATE) have used the SensorMedics 3100B ventilator, a device that has an electromechanically given diaphragm, which normally oscillates with an I:E fixed 1:1 ratio

The Bottom Line

  • HFOV improved oxygenation but did not improve 30 day mortality when compared with conventional ventilation in patients with ARDS (approximates to ‘moderate to severe’ using new definition)
  • HFOV should not be used for routine care in ARDS.
  • The OSCILLATE trial which was published concurrently reported a lower mortality in  conventional ventilation group (35% vs. 41%). One of the reasons may be that ARDSnet ventilation was mandated in OSCILLATE rather than suggested in OSCAR
  • Before selling your oscillator on e-bay, some may still consider it an option for rescue therapy in ARDS whilst making an ECMO referral

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.