Skip to main content

Foundational Trials

PreVent

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Bag-mask ventilation during tracheal intubation of critically ill adults
  • Acronym: PreVent
  • Year: 2019
  • Journal published in: The New England Journal of Medicine
  • Citation: Casey JD, Janz DR, Russell DW, Vonderhaar DJ, Joffe AM, Dischert KM, et al. Bag-mask ventilation during tracheal intubation of critically ill adults. N Engl J Med. 2019;380:811-821.

Context & Rationale

  • Background
    • Peri-intubation hypoxaemia is common in ICU airway management and is strongly associated with peri-procedural cardiovascular collapse (including cardiac arrest) and subsequent morbidity.
    • Classic rapid-sequence intubation (RSI) doctrine discouraged mask ventilation after induction, aiming to reduce gastric insufflation and aspiration risk, but this approach assumes a short apnoeic interval and adequate oxygen reserve.
    • Critically ill adults often have reduced functional residual capacity, high shunt fraction, poor preoxygenation efficiency, and a longer/less predictable interval from induction to tracheal intubation (multiple attempts, physiologically difficult airway), all of which increase desaturation risk.
    • Clinicians therefore faced an unresolved trade-off: tolerate apnoea without ventilation (potentially severe hypoxaemia) versus provide “gentle” bag-mask ventilation (potential aspiration risk), in a setting with limited randomised evidence.
  • Research Question/Hypothesis
    • In critically ill adults undergoing tracheal intubation, does protocolised bag-mask ventilation between induction and laryngoscopy (vs no ventilation during that interval) increase the lowest oxygen saturation and reduce severe hypoxaemia, without increasing aspiration or pulmonary complications?
    • Hypothesis: bag-mask ventilation with controlled technique would improve oxygenation during the peri-induction apnoeic window and not increase clinically important aspiration events.
    1
  • Why This Matters
    • Bag-mask ventilation is ubiquitous, low-cost, and immediately deployable; if beneficial and safe, it offers a pragmatic method to reduce severe hypoxaemia during one of the highest-risk procedures in ICU care.
    • A clear answer would shift peri-intubation practice away from tradition-driven “no ventilation” RSI dogma and towards physiology-driven airway management.

Design & Methods

  • Research Question: Among critically ill adults undergoing tracheal intubation, does bag-mask ventilation during the interval between induction and laryngoscopy improve peri-intubation oxygenation compared with no ventilation during that interval?
  • Study Type: Pragmatic, multicentre, parallel-group randomised trial; open-label; ICU setting (7 ICUs in the United States); ClinicalTrials.gov NCT03026322.
  • Population:
    • Adults in an ICU undergoing tracheal intubation with planned administration of induction medication (typical ICU RSI practice).
    • Inclusion required clinical equipoise regarding whether bag-mask ventilation between induction and laryngoscopy should be performed or withheld.
    • Key exclusions included situations in which the treating clinician judged bag-mask ventilation was required to prevent hypoxaemia, or judged that bag-mask ventilation should not be provided because aspiration risk was unacceptably high (e.g., active emesis/haematemesis or analogous high-risk scenarios), and situations in which there was insufficient time to enrol prior to induction.
  • Intervention:
    • Bag-mask ventilation starting after induction and continuing until laryngoscopy.
    • Technique: use of an oropharyngeal airway; two-handed mask seal with upper airway manoeuvres (jaw-thrust and/or head-tilt–chin-lift); oxygen flow ≥15 L/min; PEEP valve set at 5–10 cm H2O; ventilation rate 10 breaths/min; smallest tidal volume to produce visible chest rise.
  • Comparison:
    • No bag-mask ventilation between induction and laryngoscopy (“apnoea” strategy during that interval).
    • Therapeutic bag-mask ventilation was permitted for rescue (e.g., following a failed laryngoscopy attempt or if oxygen saturation fell below 90%).
  • Blinding: Unblinded allocation (procedural intervention not feasibly blinded); key physiological outcomes were objective (pulse oximetry), but co-interventions (preoxygenation method, timing) could be influenced by clinician knowledge of group assignment.
  • Statistics: A total of 350 patients were planned to detect a 5 percentage-point difference in lowest oxygen saturation (SD 14), with 90% power and two-sided alpha 0.05; an interim assessment found SD 15, increasing the target sample size to 400 (final enrolment 401). Primary analysis was intention-to-treat (primary outcome analysed using rank-based testing with complementary regression-based effect estimates). 1
  • Follow-Up Period: Primary and most secondary physiological outcomes measured from induction through 2 minutes post-intubation; radiographic outcomes assessed within 48 hours where imaging was obtained; clinical outcomes followed to hospital discharge and ventilator-free days to day 28.

Key Results

This trial was not stopped early. Enrolment proceeded to the revised target sample size (401 patients) after a pre-specified variance re-estimation.

Outcome Bag-mask ventilation No ventilation Effect p value / 95% CI Notes
Lowest oxygen saturation (primary) 96% (IQR 87–99) (n=193) 93% (IQR 81–99) (n=197) Adjusted MD +4.7 percentage points 95% CI 2.5 to 6.8; P=0.01 Primary comparison by rank-based testing; regression-based adjusted estimate reported.
Severe hypoxaemia (SpO2 <80%) 21/193 (10.9%) 45/197 (22.8%) RR 0.48 95% CI 0.30 to 0.77 Clinically salient secondary outcome; absolute reduction 11.9 percentage points.
Hypoxaemia (SpO2 <90%) 57/193 (29.5%) 79/197 (40.1%) RR 0.74 95% CI 0.56 to 0.97 Directionally consistent with primary outcome.
Profound hypoxaemia (SpO2 <70%) 8/193 (4.1%) 20/197 (10.2%) RR 0.41 95% CI 0.18 to 0.90 Low event rates; confidence interval remains wide but excludes unity.
Aspiration (operator-reported) 5/199 (2.5%) 8/202 (4.0%) RR 0.63 95% CI 0.21 to 1.91; P=0.41 Trial not powered to exclude modest harm; ascertainment limited to recognised events.
New radiographic opacity <48 h post-intubation 30/183 (16.4%) 28/189 (14.8%) RR 1.11 95% CI 0.70 to 1.77; P=0.73 Non-specific surrogate (atelectasis, aspiration, evolving pneumonia all plausible).
In-hospital mortality 52/196 (26.5%) 53/201 (26.4%) RR 1.00 95% CI 0.71 to 1.41; P=0.99 Not designed/powered for mortality benefit.
Ventilator-free days to day 28 19 (IQR 0–25) 18 (IQR 0–25) MD +0.6 days 95% CI −1.7 to 2.9; P=0.61 Consistent with a peri-procedural physiological benefit without downstream signal.
    • Bag-mask ventilation increased median lowest SpO2 from 93% to 96% and reduced severe hypoxaemia from 22.8% to 10.9%.
    • Aspiration was uncommon (2.5% vs 4.0%) with wide confidence intervals, meaning clinically important harm could not be excluded despite the absence of a harm signal.
    • Clinical endpoints (ventilator-free days, in-hospital mortality) were not different, consistent with the trial’s peri-procedural mechanistic focus.

Internal Validity

    • Randomisation and Allocation: Randomised, stratified by ICU; allocation concealment was operationally feasible up to enrolment, but blinding post-allocation was not feasible.
    • Drop out or exclusions: 401 patients randomised; lowest SpO2 missing for 11/401 (2.7%) (193 analysed in bag-mask group; 197 analysed in no-ventilation group for the primary outcome).
    • Performance/Detection Bias: Open-label design creates risk of co-intervention differences; detection bias was mitigated for the primary outcome by its objective nature (pulse oximetry), but aspiration ascertainment relied on clinical recognition and thus was vulnerable to under-detection.
    • Protocol Adherence: Excellent separation for the intended exposure between induction and laryngoscopy: bag-mask ventilation 198/199 (99.5%) vs 5/202 (2.5%).
    • Separation of the Variable of Interest: Contamination/rescue ventilation was clinically common later in the peri-intubation period: bag-mask ventilation between induction and intubation occurred in 198/199 (99.5%) vs 44/202 (21.8%).
    • Baseline Characteristics: Oxygen saturation at induction was similar (median 99% [IQR 97–100] vs 99% [IQR 96–100]), suggesting comparable immediate pre-induction oxygenation status, but preoxygenation approach differed after allocation.
    • Adjunctive therapy use: Positive-pressure preoxygenation was more frequent in the bag-mask group (99/199 [49.7%] vs 47/202 [23.4%]); preoxygenation with bag-mask specifically occurred in 79/199 (39.7%) vs 22/202 (10.9%), consistent with potential allocation-influenced co-intervention.
    • Timing: The intervention arm had longer time from induction to laryngoscopy (median 98 s [IQR 65–135] vs 72 s [IQR 52–120]) and longer induction-to-intubation time (144 s [IQR 107–164] vs 119 s [IQR 90–191]), reflecting real-world workflow changes induced by the assigned strategy.
    • Dose: Ventilation “dose” was defined pragmatically (10 breaths/min; minimal tidal volume for chest rise; PEEP 5–10 cm H2O; oxygen flow ≥15 L/min), which standardised a gentle approach but may not match all clinician techniques.
    • Outcome Assessment: Lowest SpO2 is objective but subject to oximeter lag and artefact; aspiration and radiographic opacity are clinically meaningful but imperfect and variably sensitive.
    • Statistical Rigor: Sample size re-estimation was prespecified and addressed variance mis-specification; primary analysis remained intention-to-treat with sensitivity analyses reported as consistent with the main findings.

Conclusion on Internal Validity: Overall, internal validity is moderate-to-strong: randomisation and near-complete follow-up support causal inference for oxygenation outcomes, but open-label design, allocation-influenced co-interventions (notably preoxygenation method), and limited power/ascertainment for aspiration constrain certainty about harms.

External Validity

    • Population Representativeness: Adult ICU cohort across seven ICUs in the United States; applicability is strongest to medical ICU intubations performed with RSI-style induction where clinicians are uncertain about whether to ventilate during the apnoeic window.
    • Important exclusions: Patients in whom clinicians believed ventilation was mandatory to avoid immediate hypoxaemia, and patients in whom ventilation was considered unsafe because of aspiration risk, were excluded; this limits generalisability precisely in the extremes of physiological and aspiration risk.
    • Setting and operator effects: The protocolised two-handed mask seal, PEEP valve use, and airway adjunct (oropharyngeal airway) may not be routinely available or consistently performed in all ICUs, emergency departments, or resource-limited settings.
    • Applicability: Where skilled mask ventilation and PEEP are available, the intervention is highly scalable; where staffing/skills are limited, the balance between benefit and potential harm may differ.

Conclusion on External Validity: Generalisability is moderate: results are highly applicable to typical ICU RSI intubations without prohibitive aspiration risk, but less certain for crash intubations, profoundly unstable patients, or those at very high aspiration risk.

Strengths & Limitations

  • Strengths:
    • Pragmatic multicentre randomised design in the ICU environment (high clinical relevance).
    • Clear, protocolised description of the bag-mask technique (oxygen flow, PEEP, rate, airway adjuncts), enabling reproducibility.
    • Objective primary physiological endpoint with low missingness.
    • Clinically meaningful reduction in severe hypoxaemia (SpO2 <80%).
  • Limitations:
    • Open-label design with plausible allocation-influenced co-interventions (notably preoxygenation modality and timing), potentially inflating or attenuating the measured effect.
    • Primary endpoint is a surrogate (lowest SpO2) rather than patient-centred outcomes; the trial was not powered for mortality or longer-term clinical benefit.
    • Aspiration and pulmonary harms were uncommon with wide confidence intervals; ascertainment relied on recognised events and non-specific radiographic surrogates.
    • Exclusion of patients at very high aspiration risk or those deemed unable to tolerate apnoea limits inference about safety in those populations.

Interpretation & Why It Matters

  • Practice implication
    • For ICU intubations in which aspiration risk is acceptable, a protocolised “gentle” bag-mask ventilation strategy between induction and laryngoscopy can materially reduce severe hypoxaemia (10.9% vs 22.8%) while improving lowest oxygen saturation (median 96% vs 93%).
    • The intervention reframes the peri-intubation apnoeic window as an active therapeutic period rather than an obligatory “no ventilation” interval, challenging tradition-based RSI practice in critical illness.
    • Interpretation emphasises physiology over dogma: the price of apnoea in critical illness is often severe desaturation; controlled mask ventilation with PEEP aims to mitigate that cost while minimising gastric insufflation risk.
    2

Controversies & Subsequent Evidence

    • Aspiration risk remains the central residual uncertainty: aspiration events were uncommon (2.5% vs 4.0%) with wide confidence intervals (RR 0.63; 95% CI 0.21 to 1.91), leaving limited ability to exclude clinically important harm; reliance on clinically recognised aspiration and non-specific radiographic opacities constrains inference about occult aspiration. 3
    • Open-label co-interventions complicate mechanistic attribution: clinicians provided positive-pressure preoxygenation more often in the bag-mask group (49.7% vs 23.4%), plausibly amplifying oxygen reserve before induction and contributing to outcome differences beyond “between induction and laryngoscopy” ventilation alone.
    • Endpoint hierarchy and clinical meaning: lowest SpO2 is a surrogate, but the consistent reduction across clinically consequential thresholds (SpO2 <80%, <70%) supports relevance to peri-intubation safety, even if downstream patient-centred benefits were not demonstrated.
    • Subsequent RCT evidence in peri-intubation oxygenation emphasises benefit of positive-pressure strategies: a later large trial of noninvasive ventilation for preoxygenation before ICU intubation demonstrated improved oxygenation outcomes, reinforcing the broader concept that active ventilatory support around induction can be beneficial when carefully applied. 4
    • Synthesis evidence: a contemporary network meta-analysis comparing peri-intubation oxygenation strategies supports the view that strategy selection (NIV, HFNC, facemask, combinations) meaningfully affects oxygenation endpoints, with the practical implication that “no assistance” during critical windows is rarely optimal for hypoxaemia-prone patients. 5
    • Guidelines: post-trial critical care guidance on RSI increasingly incorporates oxygenation-first principles and acknowledges ventilation strategies around induction when aspiration risk is acceptable and technique is controlled. 6

Summary

    • In 401 ICU patients, bag-mask ventilation between induction and laryngoscopy increased lowest oxygen saturation (median 96% vs 93%) compared with no ventilation during that interval.
    • Severe hypoxaemia (SpO2 <80%) was reduced from 22.8% to 10.9% (RR 0.48; 95% CI 0.30 to 0.77).
    • Aspiration was uncommon (2.5% vs 4.0%) and not increased, but confidence intervals were wide and ascertainment imperfect.
    • There was no difference in in-hospital mortality (26.5% vs 26.4%) or ventilator-free days to day 28 (median 19 vs 18).
    • Open-label co-interventions (notably increased positive-pressure preoxygenation in the intervention group) and rescue ventilation in the control group limit purity of mechanistic attribution but do not negate the pragmatic clinical signal.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • Where volume/issue/page details are not displayed above, the DOI link resolves to the definitive journal landing page containing full bibliographic metadata.

Overall Takeaway

PreVent is a landmark ICU airway trial because it directly tested (and largely refuted) the assumption that “no ventilation after induction” should be default RSI practice in critical illness. By demonstrating a substantial reduction in severe hypoxaemia with a standardised, gentle bag-mask technique and no clear signal of increased aspiration, it shifted peri-intubation thinking from tradition-based dogma to physiology-based risk management—while appropriately leaving residual uncertainty about rare harms in the highest aspiration-risk populations.

Overall Summary

  • Gentle bag-mask ventilation between induction and laryngoscopy improved peri-intubation oxygenation and reduced severe hypoxaemia in critically ill adults, without a detected increase in aspiration—but uncommon harms could not be definitively excluded.

Bibliography