The PREOXI Trial: Pre-Oxygenation with NIV vs Facemask

Background: Tracheal intubation is frequently performed in critically ill patients.  Hypoxemia is one of the major adverse events that can occur during intubation and it increases the risk of morbidity and mortality. Preoxygenation before induction of anesthesia increases the amount of oxygen in the lungs and decreases the risk of hypoxemia by prolonging the safe apneic time. Most patients receive preoxygenation in the form of non-rebreather  (NRB) masks which do not provide positive pressure or ventilatory support. The absence of positive pressure provision is important as recruitment of alveoli can result in better oxygenation and, in theory, longer safe apneic times. Noninvasive ventilation (NIV) provides positive pressure and ventilatory support and in theory should result in safer intubations. However, NIV does take time to set up and many worry that provision of positive pressure can increase the risk aspiration of gastric contents(Although not shown in the PreVent trial or this trial).

Paper: Gibbs KW et al. Noninvasive Ventilation for Preoxygenation During Emergency Intubation. NEJM 2024. PMID: 38869091

Clinical Question: In critically ill adult patients does the use of NIV for preoxygenation reduce hypoxemia during intubation compared to an oxygen mask?

What They Did:

  • Pragmatic Trial Examining Oxygenation Prior to Intubation (PREOXI)
  • Pragmatic, multicenter, unblinded, randomized, parallel-group, controlled trial
  • 7 EDs and 17 ICUs across the US
  • Critically ill adults (Age ≥18 years) undergoing tracheal intubation
  • Randomized to preoxygenation (At least 3min) before induction of anesthesia) with:
    • NIV: Use of a tight-fitting mask connected to either a conventional mechanical ventilator or a dedicated noninvasive ventilator before the induction of anesthesia
      • Operators selected type of ventilator and the mode
      • FiO2 100%
      • ePAP 5cmH20
      • iPAP at least 10cmH20
      • RR 10BPM (backup rate)
    • O2 Mask: Supplemental O2 using either a NRB or BVM device without manual ventilation before induction of anesthesia
      • Choice of NRB or BVM was up to operator
      • Administer highest flow rate of oxygen available (≥15LPM)

Trial Interventions

  • Operators could administer ventilation with BVM to patients in either group
  • O2 mask group: 30% got BVM with ventilations; 11.3% got BVM without ventilations during apneic period
  • Unclear if PEEP valves were used
    • Operators could provide supplemental oxygen through standard nasal cannula or HFNC to patients in either trial group during preoxygenation during the interval between induction of anesthesia and initiation of laryngoscopy and during the interval between initiation of laryngoscopy and tracheal intubation (apneic oxygenation)
  • 3% got some form of NC in NIV group
  • 16% got some form of NC in O2 mask group


  • Primary: Hypoxemia during intubation (Defined as O2 sat <85% during interval between induction of anesthesia and 2 min after tracheal intubation)
  • Secondary:
    • Lowest oxygen saturation during the interval between induction of anesthesia and 2min after tracheal intubation
  • Exploratory:
    • Hemodynamic events that could result from PPV or severe hypoxemia (Including SBP <65mmHg (Hypotension) or increased use of vasopressors, and cardiac arrest) during interval between induction of anesthesia and 2min after tracheal intubation
  • Safety:
    • Gastric aspiration during intubation (As reported by operator or a new infiltrate on CXR in 24hours after induction or O2 sat and FiO2 at 24hrs after induction)


  • Critically ill adults (Age ≥18 years)
  • Undergoing tracheal intubation that involved sedation and a laryngoscope


  • Pregnant
  • Prisoner
  • Already receiving positive-pressure ventilation
  • Apnea or hypopnea
  • Immediate need for tracheal intubation that precluded randomization
  • Clinician performing the procedure determined that preoxygenation with NIV or an oxygen mask was either necessary or contraindicated


  • 4567 patients assessed for eligibility
    • 4462 met inclusion criteria
      • Most common reasons for exclusion
        • 924 underwent intubation too urgently to complete trial procedures
        • 840 already receiving PPV
        • 389 had vomiting, hematemesis, hemoptysis or epistaxis precluding NIV
        • 313 were eligible but not enrolled
        • 248 had severe agitation precluding noninvasive ventilation
      • 1301 patients enrolled
        • 1% had hypoxemic respiratory failure
        • Tracheal intubation performed in ICU in 73.2%
        • Tracheal intubation performed in ED in 26.8%
        • 9% of intubations performed by a resident or a fellow
          • Operators had performed a median of 50 previous tracheal intubations
        • 5% of NIV group received NIV
        • 8% in the O2 mask group received mask for preoxygenation
        • Duration of preoxygenation was at least 3min in 96.6% of patients n NIV group and 95.0% in the O2 mask group
        • O2 saturation of ≤95%
          • NIV: 8.3%
          • O2 Mask: 17.4%
      • Hypoxemia (<85%)
        • NIV: 9.1%
        • O2 Mask: 18.5%
        • NNT = 10
        • Diff -9.4%; 95% CI -13.2 to -5.6; P <0.001
        • Subgroup Analysis

  • All levels of oxygen therapy benefitted from NIV: This goes against the thought that it was all FiO2 alone (i.e. Breaths may have helped with alveolar recruitment)
  • Effect of NIV was higher in patients with higher body mass index (≥30)
  • Cardiac Arrest
    • NIV: 0.2% (1/645)
    • O2 Mask: 1.1% (7/656)
    • Diff -0.9%; 95% CI -1.8 to -0.01
  • Aspiration
    • NIV: 0.9%
    • O2 Mask: 1.4%
    • Diff -0.4%; 95% CI -1.6 to 0.7
  • All Outcomes

  • Lowest Oxygen Saturation


  • Asks a clinically important question
  • Multicenter randomized clinical trial increases generalizability of findings
  • Large sample size provides sufficient statistical power to detect clinically meaningful differences between groups
  • Overseen by an independent data and safety monitoring board
  • Collection of outcome data by an independent observer minimizes observer bias
  • Trial protocol and statistical analysis plan were published before the conclusion of enrollment


  • Excluded patients already receiving PPV which does not inform decisions regarding preoxygenation in this population
  • Excluded patients with vomiting, hematemesis, hemoptysis or epistaxis which does not allow for assessment of safety or effectiveness of NIV in this population of patients
  • Did not evaluate the use of HFNC for preoxygenation/apneic oxygenation
  • Patients, clinicians, and trial personnel were aware of the trial group assignments which could cause bias
  • Intubations were performed by residents/fellows which may not have as much experience as more seasoned clinicians
  • Unclear if PEEP valves were used in the O2 mask arm


  • PreVent Trial [Link is HERE]
    • We already knew that PPV prior to intubation in critically ill patients improves lowest O2 sat after induction prior to tracheal intubation
    • In this RCT patients were randomized to BVM vs no BVM during induction to initiation of laryngoscopy
    • There was a slight difference in median lowest O2 saturation (primary outcome) 96% BVM vs 93% no BVM
    • Additionally severe hypoxemia (<80%) (a secondary outcome) was 10.9% in BVM group compared to 22.8% in the no BVM group
    • There was no difference in aspiration: 2.5% vs 4.0%
  • Time from Induction to Intubation/1st Pass Success Rate
    • Intubations took 115 seconds in the NIV group vs 113 seconds in the O2 mask group. Would these have been quicker intubations in more experienced hands
    • First pass successful intubation was 88.9% in NIV group and 87.0% in the O2 mask group
  • Assumptions/Unknowns
    • There were several things that would affect preoxygenation specifically in the O2 mask group that were not recorded in this trial. The assumption is these things were done (That is a big assumption), but we just don’t know for sure…
    • Sitting Up: Were patients in the O2 mask group sitting up like the patients in the NIV group. This is important for several reasons including removing weight off the chest/diaphragm, recruiting posterior alveoli, and using gravity to prevent n/v
    • PEEP Valve on BVM: It is unclear if PEEP valves were used in the patients that got a BVM.  A PEEP of 10 or 15cmH20 on the PEEP valve will certainly recruit alveoli better than a BVM alone at flush rate oxygen
    • Flush Rate O2: In the text the authors write O2 ≥15LPM but was it at 15LPM or flush rate.  There is a big difference between 15LPM and 30 to 45LPM
  • Pragmatically using NIV is not going to be necessary for every intubation IMHO (Here is what I do in my practice)…
    • I preoxygenate all intubations with NRB @flush rate oxygen + NC @15LPM. IF I get O2 sats >95% then I proceed with induction and leave NC on for apneic oxygenation prior to laryngoscopy
    • If I am unable to get O2 sats >95% or I am concerned the patient has some shunt physiology, then these patients need PEEP (I am now adding obese patients to this list). I do this with BVM @flush rate/PEEP valve @15cmH20 as apneic CPAP + NC @15LPM.  Once I start to push induction meds I start to gently bag the patient to help with alveolar recruitment prior to laryngoscopy (This was shown in the PreVent trial (Discussed above)
    • PREOXI shows us that PPV/PEEP can also be safely achieved with NIV instead of BVM @flush rate/PEEP valve @15cmH20, especially when there is time available to get equipment ready

Author Conclusion: “Among critically ill adults undergoing tracheal intubation, preoxygenation with noninvasive ventilation resulted in a lower incidence of hypoxemia during intubation than preoxygenation with an oxygen mask.” 

Clinical Take Home Point: This is potentially a practice changing paper. We should be using some form of PPV/PEEP for preoxygenation in critically ill patients. I don’t however think it has to be in the form of NIV as a BVM @flush rate/PEEP valve @15cmH20 should/could achieve the same thing (Not compared in PREOXI trial).


  1. Gibbs KW et al. Noninvasive Ventilation for Preoxygenation During Emergency Intubation. NEJM 2024. PMID: 38869091

For More Thoughts on This Topic Checkout:

Post Peer Reviewed By: Anand Swaminathan, MD (Twitter/X: @EMSwami)

Cite this article as: Salim Rezaie, "The PREOXI Trial: Pre-Oxygenation with NIV vs Facemask", REBEL EM blog, June 28, 2024. Available at:

Like this article?

Share on Facebook
Share on Twitter
Share on Linkdin
Share via Email

Want to support rebelem?