Protocol and Statistical Analysis Plan for the Pragmatic Trial Examining Oxygenation Prior to Intubation of Preoxygenation With Noninvasive Ventilation vs Oxygen Mask in Critically Ill Adults

Abstract

BACKGROUND:

Hypoxemia is a common and life-threatening complication during emergency tracheal intubation of critically ill adults. The administration of supplemental oxygen before the procedure (ie, preoxygenation) decreases the risk of hypoxemia during intubation.

RESEARCH QUESTION:

Does preoxygenation with noninvasive ventilation prevent hypoxemia during tracheal intubation of critically ill adults compared with preoxygenation with an oxygen mask?

STUDY DESIGN AND METHODS:

The Pragmatic Trial Examining Oxygenation Prior to Intubation (PREOXI) is a prospective, multicenter, nonmasked randomized comparative effectiveness trial being conducted in seven EDs and 17 ICUs across the United States. The trial compares preoxygenation with noninvasive ventilation vs oxygen mask among 1,300 critically ill adults undergoing emergency tracheal intubation. Eligible patients are randomized in a 1:1 ratio to receive either noninvasive ventilation or an oxygen mask before induction. The primary outcome is the incidence of hypoxemia, defined as a peripheral oxygen saturation of < 85% between induction and 2 min after intubation. The secondary outcome is the lowest oxygen saturation between induction and 2 min after intubation. Enrollment began on March 10, 2022, and is expected to conclude in 2023.

INTERPRETATION:

The PREOXI investigation will provide important data on the comparative effectiveness of preoxygenation with noninvasive ventilation vs oxygen mask for the prevention of hypoxemia during emergency tracheal intubation. Specifying the protocol and statistical analysis plan before the conclusion of enrollment increases the rigor, reproducibility, and interpretability of the trial.

TRIAL REGISTRY:

ClinicalTrials.gov; No.: NCT05267652; URL: www.clinicaltrials.gov

Life-threatening hypoxemia occurs in 10% to 20% of emergency tracheal intubations in the ED and ICU.^1,2 Hypoxemia during intubation is associated with an increased risk of cardiac arrest and death.^3,4 Identifying interventions to prevent hypoxemia during emergency tracheal intubation is a high priority for clinicians and researchers.^5,6 Because patients typically are apneic between induction of anesthesia and intubation, but continue to consume oxygen, the oxygen content in the lungs at the time of induction is a primary determinant of whether the patient will experience hypoxemia.

Preoxygenation, the administration of supplemental oxygen before induction of anesthesia, increases lung oxygen content at induction and decreases the risk of hypoxemia.^7,8 In clinical practice, preoxygenation for emergency tracheal intubation of critically ill adults is most commonly administered using either an oxygen mask or noninvasive ventilation.¹

Preoxygenation with an oxygen mask typically is performed using either a non-rebreather mask or a bag-mask device. A non-rebreather mask is a loose-fitting mask with an oxygen reservoir connected to an oxygen source. A bag-mask device is a mask capable of forming a tight seal over the mouth when held in place by a clinician and can be used to provide supplemental oxygenation alone, or both supplemental oxygen and ventilation.⁹ Both types of oxygen mask can deliver up to 100% oxygen, are simple to set up, and have low potential for gastric insufflation. However, oxygen masks may deliver oxygen less effectively in critically ill patients when high minute ventilation and poor mask seal allow the entrainment of ambient air, with resulting alveolar oxygen concentrations as low as 50%.^7,10

Preoxygenation also is administered routinely via noninvasive ventilation, in which a tight-fitting mask is connected to a machine capable of providing both 100% oxygen and positive-pressure ventilation. Compared with an oxygen mask, noninvasive ventilation may reduce air entrainment by delivering a higher inspiratory flow rate of oxygen and by minimizing leaks. Additionally, noninvasive ventilation increases the mean airway pressure and recruits atelectatic lung, potentially reducing shunting. Compared with an oxygen mask, noninvasive ventilation may take longer to initiate and potentially may increase the risk of gastric insufflation and aspiration.

Two small randomized trials have compared these two approaches to preoxygenation. The first trial found that, among 53 patients in the ICU undergoing tracheal intubation in two hospitals, noninvasive ventilation increased the lowest oxygen saturation compared with the effects of an oxygen mask (mean lowest oxygen saturation, 93% vs 81%, respectively; P < .001) with no difference between groups in the incidence of aspiration (6% vs 8%).¹¹ Among 201 patients in the ICUs of six hospitals, the second trial found no difference in the severity of illness in the 7 days after intubation and an incidence of hypoxemia during intubation of 18.4% in the noninvasive ventilation group vs 27.7% in the oxygen mask group (P = .10).¹² Thus, whether noninvasive ventilation for preoxygenation in critically ill adults undergoing emergency tracheal intubation decreases the incidence of hypoxemia compared with an oxygen mask remains unknown. We designed the Pragmatic Trial Examining Oxygenation Prior to Intubation (PREOXI) to test the hypothesis that, among critically ill adults undergoing emergency tracheal intubation in the ED and ICU, preoxygenation with noninvasive ventilation decreases the incidence of hypoxemia compared with preoxygenation with an oxygen mask.

Study Design and Methods

This article was written in accordance with Standard Protocol Items: Recommendations for Interventional Trials guidelines (Table 1, e-Appendix 1).¹³

TABLE 1].

Schedule of Enrollment, Interventions, and Assessments in PREOXI

Time Point	Eligibility Screening, Decision to Perform Tracheal Intubation	Randomization and Allocation, Before Tracheal Intubation	During the Procedure				Final Outcome Assessment, Discharge or 28 d After Enrollment
			Induction	Tracheal Intubation	0–2 min After Tracheal Intubation	0–24 h After Tracheal Intubation
Enrollment
Eligibility screen	X	…	…	…	…	…	…
Enrollment	…	X	…	…	…	…	…
Allocation	…	X	…	…	…	…	…
Interventions
Preoxygenation with NIPPV	…	X	X	…	…	…	…
Preoxygenation with face mask oxygen	…	X	X	…	…	…	…
Screening for contraindications	X	X	X	…	…	…	…
Assessments
Baseline variables	X	X	…	…	…	…	…
Periprocedural variables	…	X	X	X	X	…	…
Adverse events	…	X	X	X	X	X	X
Clinical outcomes	…	…	…	…	X	X	X

Table entries without data mean that data elements were not collected at that time. NIPPV = noninvasive positive-pressure ventilation; PREOXI = Pragmatic Trial Examining Oxygenation Prior to Intubation.

Study Design

PREOXI is a pragmatic, multicenter, nonmasked, parallel-group randomized trial comparing preoxygenation using noninvasive ventilation with preoxygenation using an oxygen mask among critically ill adults undergoing emergency tracheal intubation in the ED and ICU. The primary outcome is the incidence of hypoxemia, defined as a peripheral oxygen saturation of < 85% between induction of anesthesia and 2 min after intubation. The trial is conducted by the Pragmatic Critical Care Research Group (www.pragmaticcriticalcare.org). The trial was registered before initiation of enrollment (ClinicalTrials.gov Identifier: NCT05267652).

Study Population

Patients located in a participating ED or ICU who are undergoing tracheal intubation using a laryngoscope and sedation are eligible. Complete lists of inclusion and exclusion criteria are provided in Table 2.

TABLE 2].

Inclusion and Exclusion Criteria

Inclusion criteria

Patient is located in a participating unit.

Planned procedure is tracheal intubation using a laryngoscope and sedation.

Planned operator is a clinician expected to perform tracheal intubation routinely in the participating unit.

Exclusion criteria

Patient is receiving positive-pressure ventilation by a mechanical ventilator, bag-mask device, or laryngeal mask airway.

Patient is known to be younger than 18 years.

Patient is known to be pregnant.

Patient is known to be in prison.

Immediate need for tracheal intubation precludes safe performance of study procedures.

Patient is apneic, is hypopneic, or has another condition requiring positive-pressure ventilation between enrollment and induction.

Operator has determined that preoxygenation with noninvasive ventilation or preoxygenation with an oxygen mask is required or contraindicated for optimal care of the patient.

Randomization and Treatment Allocation

Patients are randomized in a 1:1 ratio to undergo preoxygenation with noninvasive ventilaton vs oxygen mask in permuted blocks of variable size, stratified by study site. Study-group assignments are generated using a computerized randomization sequence, placed in sequentially numbered opaque envelopes and distributed to enrolling sites. Before opening the envelope, the operator determines that the patient meets eligibility criteria. Study group assignment remains concealed to study personnel and clinicians until after the decision has been made to enroll the patient and the envelope is opened. Patients are enrolled as soon as the operator opens the trial envelope to reveal study group assignment. After randomization, patients, clinicians, and study personnel are not masked to group assignment because of the nature of the intervention.

Study Interventions

Training

Before beginning enrollment at each site, study investigators provide training on study procedures including instructional videos with consensus best practice recommendations for preoxygenation. Descriptions of the training videos and website links can be found in e-Appendix 1.

Noninvasive Ventilation Group

For patients assigned to the noninvasive ventilation group, operators are instructed to administer noninvasive ventilation using either a conventional mechanical ventilator or a dedicated noninvasive ventilator (ie, a bilevel positive airway pressure machine). The operator determines the ventilator type, ventilator mode, inspiratory and expiratory pressures, and mandatory respiratory rate. Operators receive the following best practice recommendations for the administration of preoxygenation using noninvasive ventilation: (1) preoxygenate ≥ 3 min (if feasible); (2) continue noninvasive ventilation until initiation of laryngoscopy, (3) FIO₂ of 100%, (4) expiratory pressure of ≥ 5 cm H₂O; (5) use an inspiratory pressure of ≥ 10 cm H₂O, and (6) respiratory rate of ≥ 10 breaths per minute.

Oxygen Mask Group

For patients assigned to the oxygen mask group, clinicians are instructed to administer supplemental oxygen via a non-rebreather mask or bag-mask device without bag-mask ventilation from the initiation of preoxygenation until the induction of anesthesia. The operator determines whether to use a non-rebreather mask or a bag-mask device (without ventilation). Operators receive the following best practice recommendations for the administration of preoxygenation using an oxygen mask: (1) preoxygenate ≥ 3 min (if feasible), (2) use the maximum oxygen flow rate possible (≥ 15 L/min), and (3) continue oxygenation from induction to laryngoscopy.

Cointerventions

In both trial groups, trial protocol allows the administration of supplemental oxygen by standard nasal cannula or high-flow nasal cannula during preoxygenation, between induction of anesthesia and initiation of laryngoscopy, and between initiation of laryngoscopy and intubation of the trachea. Bag-mask ventilation between induction and laryngoscopy also may be provided at the discretion of the operator in both groups. Figure 1 shows the interventions, cointerventions, and phases of the intubation process during which these interventions are applied.

Figure 1 –

Diagram showing oxygenation and ventilation interventions in each trial group. The methods of oxygen delivery and ventilation allowed, recommended, and mandated during each phase of the tracheal intubation procedure are displayed for the noninvasive ventilation group and the oxygen mask group. For patients assigned to the noninvasive ventilation group, clinicians are instructed to administer noninvasive ventilation via either a dedicated noninvasive ventilator or an invasive ventilator connected to a tight-fitting mask from the initiation of preoxygenation until induction of anesthesia. Trial protocol recommends that noninvasive ventilation be continued between induction of anesthesia and the initiation of laryngoscopy. For patients assigned to the oxygen mask group, clinicians are instructed to administer supplemental oxygen without ventilation via either a nonrebreather mask or a bag-mask device from the initiation of preoxygenation until induction of anesthesia. Trial protocol recommends that the administration of supplemental oxygen by mask be continued between induction of anesthesia and the initiation of laryngoscopy. In either trial group, trial protocol allows the administration of supplemental oxygen by high-flow nasal cannula or standard nasal cannula in any phase of the intubation procedure, including under a mask during preoxygenation. In either trial group, trial protocol allows ventilation using a bag-mask device between induction and the initiation of laryngoscopy in place of noninvasive ventilation or an oxygen mask. NIV = noninvasive ventilation.

Data Collection

An observer not directly involved with the intubation procedure collects data for key periprocedural outcomes, including oxygen saturation at induction and the lowest value for oxygen saturation between induction and 2 min after successful intubation. Observers may be trained clinical personnel (eg, physicians or nurses) or research personnel. Immediately after the intubation procedure, the operator completes a paper data collection form to record the device used for preoxygenation, the duration of preoxygenation, the devices used for oxygenation and ventilation between induction and laryngoscopy, and complications of intubation (eg, aspiration).¹⁴ Study personnel at each site review the medical record to collect data on baseline characteristics, management before and after laryngoscopy, and clinical outcomes. A complete list of baseline, periprocedural, and in-hospital variables are provided in e-Table 1.

Primary Outcome

The primary outcome is the incidence of hypoxemia, defined as a peripheral oxygen saturation of < 85% during the interval between induction and 2 min after intubation.

Secondary Outcome

The sole secondary outcome is the lowest oxygen saturation during the interval between induction and 2 min after tracheal intubation.

Additional Outcomes

Table 3 reports the safety outcomes, exploratory procedural outcomes, and exploratory clinical outcomes.

TABLE 3].

Study Outcomes

Outcome Type	Description
Primary	Incidence of hypoxemia, defined as a peripheral oxygen saturation of < 85% during the interval between induction and 2 min after tracheal intubation.
Secondary	Lowest oxygen saturation during the interval between induction and 2 min after tracheal intubation.
Safety	Incidence of operator-reported aspiration.
	${\text{F}}_{{\text{IO}}_{\text{2}}}$ at 24 h after induction.
	Oxygen saturation at 24 h after induction.
	Incidence of pneumothorax, defined as radiology report of new pneumothorax on chest imaging in the 24 h after induction. If no chest imaging is available before induction, any pneumothorax in the 24 h after induction will be assumed to be new.
	New infiltrate, defined as radiology report of new infiltrate on chest imaging in the 24 h after intubation. If no chest imaging is available before induction, any infiltrate in the 24 h after induction will be assumed to be new.
Exploratory procedural	Incidence of severe hypoxemia, defined as the lowest oxygen saturation of < 80% between induction and 2 min after tracheal intubation.
	Incidence of very severe hypoxemia, defined as the lowest oxygen saturation of < 70% between induction and 2 min after tracheal intubation.
	Oxygen saturation at induction.
	Systolic BP at induction.
	Duration from induction to successful intubation (duration of the intubation procedure).
	Cormack-Lehane grade of glottic view on first attempt.
	Successful intubation on the first attempt
	No. of laryngoscopy attempts.
	No. of attempts at passing a bougie.
	No. of attempts at passing an endotracheal tube.
	Incidence of cardiovascular collapse, defined as a composite of ≥ 1 of the following between induction and 2 min after intubation: systolic BP < 65 mm Hg, new or increased vasopressor administration, cardiac arrest not resulting in death within 1 h of induction, cardiac arrest resulting in death within 1 h of induction.
Exploratory clinical	28-d in-hospital mortality.
	Ventilator-free days to 28 d.
	ICU-free days to 28 d.

Sample Size Estimation

The current trial is designed to detect a 6% absolute difference between groups in the incidence of hypoxemia, a difference that is similar to or smaller than the difference considered to be clinically meaningful in the design of prior trials of oxygenation strategies during tracheal intubation.^2,15 Assuming an incidence of hypoxemia of 17% in the oxygen mask group based on data from two recently completed trials by this network in similar ED and ICU settings, detecting a 6% absolute decrease in the incidence of hypoxemia with 85% power at a two-sided α level of .05 would require enrollment of 1,264 patients (632 per group).^16,17 Anticipating missing data for up to 3% of patients, we plan to enroll a maximum of 1,300 total patients (650 per group). This sample size calculation was performed in PS version 3.1.2 software (Vanderbilt University Medical Center).

Data and Safety Monitoring Board and Interim Analysis

The data and safety monitoring board and interim analysis are described in e-Appendix 1.

Statistical Analysis Principles

Analyses will be conducted following reproducible research principles using R software (R Foundation for Statistical Computing). Categorical variables will be presented as number and percentage and will be compared between groups using a χ² test. Continuous variables will be presented as mean ± SD or median (interquartile range) and will be compared between groups using a Wilcoxon rank-sum test. We will present absolute between-group differences with associated 95% CIs. A two-sided P value of < .05 will define a statistically significant between-group difference in the primary outcome. With a single primary outcome, no adjustment for multiplicity will be made. For secondary, safety, and exploratory analyses, emphasis will be placed on the magnitude of differences between groups with 95% CIs, rather than statistical significance.

Main Analysis of the Primary Outcome

The main analysis will be an unadjusted intention-to-treat comparison of the primary outcome of hypoxemia between patients randomized to the noninvasive ventilation group vs patients randomized to the oxygen mask group using a χ² test. The absolute difference in proportions, associated 95% CI, and a P value for the comparison will be presented. The primary analysis will be conducted among patients for whom the primary outcome is available without imputation of missing data. Planned tables of baseline characteristics, procedural characteristics, and outcomes are presented in e-Tables 2 and 3.

Additional Analyses of the Primary Outcome

Multivariable Analysis:

To account for relevant baseline covariates, we will fit a generalized linear mixed effects model using a logit link function with the primary outcome as the dependent variable, study site as a random effect, and fixed effects of study group and the following prespecified baseline covariates: age, sex, race and ethnicity, BMI, location at enrollment (ED or ICU), highest FIO₂ 1 h before initiation of preoxygenation, Acute Physiology and Chronic Health Evaluation II score,¹⁸ indication for intubation (hypoxemic respiratory failure: yes vs no), and operator experience (number of previous intubations the operator has performed). All continuous variables will be modelled assuming a nonlinear relationship to the outcome using restricted cubic splines with between 3 and 5 knots.

Effect Modification:

We will examine whether prespecified baseline variables modify the effect of study group assignment on the primary outcome using a formal test of statistical interaction in a logistic regression model with the primary outcome as the dependent variable and independent variables of study group, the proposed effect modifier, and the interaction between the two. For categorical variables, we will present the OR and 95% CIs within each prespecified subgroup. Continuous variables will not be dichotomized for analysis of effect modification, but may be dichotomized for data presentation. In accordance with the Instrument for Assessing the Credibility of Effect Modification Analyses recommendations, we have prespecified the following baseline variables as potential effect modifiers and hypothesized the direction of effect modification for each.

Patient Location (ED vs ICU):

We hypothesize that patient location will not modify the effect of study group assignment on the primary outcome.

BMI (kg/m²):

We hypothesize that BMI will modify the effect of study group assignment on the primary outcome, with a greater decrease in the incidence of hypoxemia in the noninvasive ventilation group compared with the oxygen mask group among patients with higher BMIs, as compared with patients with lower BMIs. This hypothesis is supported by evidence from multiple prior studies that patients with obesity are more likely to have early airway closure and atelectasis-dependent shunting that is likely to improve with positive-pressure ventilation.^19,20

FIO₂ 1 h Before Intubation:

We hypothesize that FIO₂ 1 h before intubation will modify the effect of study group assignment on the primary outcome, with a greater decrease in the incidence of hypoxemia in the noninvasive ventilation group compared with the oxygen mask group among patients with higher FIO₂ 1 h before intubation compared with patients with lower FIO₂. This hypothesis is supported by evidence from multiple prior studies that patients requiring higher FIO₂ have more atelectasis-dependent shunting that is likely to improve with positive-pressure ventilation.^5,11,21

Acute Physiology and Chronic Health Evaluation II Score:

We hypothesize that Acute Physiology and Chronic Health Evaluation II score will not modify the effect of study group assignment on the primary outcome.

Hypoxemic Respiratory Failure as the Indication for Intubation (Yes vs No):

We hypothesize that hypoxemic respiratory failure as the indication for intubation will modify the effect of study group assignment on the primary outcome, with a greater decrease in the incidence of hypoxemia in the noninvasive ventilation group compared with the oxygen mask group among patients with hypoxemic respiratory failure compared with patients without. This hypothesis is supported by evidence from two prior randomized trials suggesting a potential benefit for noninvasive ventilation among patients with acute hypoxemic respiratory failure.^11,12

Analysis of the Secondary Outcome

We will perform an unadjusted intention-to-treat comparison of patients randomized to the noninvasive ventilation group vs the oxygen mask group regarding the secondary outcome of lowest oxygen saturation between induction and 2 min after intubation using the Wilcoxon rank-sum test.

Analyses of Additional Outcomes

The analysis plan for additional outcomes is described in e-Appendix 1.

Handling of Missing Data

The plan for handling of missing data is described in e-Appendix 1.

Ethics and Dissemination

Critically ill patients undergoing tracheal intubation are at significant risk for morbidity and mortality from the underlying illness. Most patients undergoing tracheal intubation in routine clinical care receive preoxygenation with either noninvasive ventilation or an oxygen mask. Any benefits or risks of these two approaches are experienced by patients undergoing tracheal intubation in clinical care, outside the context of research. As a requirement for enrollment in PREOXI, the patient’s treating clinician must determine that either preoxygenation with noninvasive ventilation or preoxygenation with an oxygen mask would be a safe and reasonable approach for the patient. Otherwise, the patient is excluded. Therefore, making the decision between the two approaches randomly in the context of a trial, rather than by a clinician who thinks either approach is safe, is expected to pose no more than minimal additional risk.

Obtaining informed consent from potential study participants or their legally authorized representatives would be impracticable. Most critically ill patients undergoing emergency tracheal intubation lack decisional capacity because of the critical illness and surrogate decision-makers frequently are unavailable. Further, emergency tracheal intubation is a time-sensitive procedure with only minutes between the decision to intubate and the completion of the procedure. Meaningful informed consent could not be executed in this brief window.

Because the study involves minimal incremental risk, the study would not adversely affect the welfare or privacy rights of the participant, and because obtaining informed consent would be impracticable, a waiver of informed consent was requested from and approved by the single institutional review board at Vanderbilt University Medical Center (Identifier: 211271). Conduct of this trial with waiver of informed consent is consistent with previous randomized trials comparing alternative approaches to tracheal intubation in widespread clinical use.^{2,15–17,22,23} This approach was approved secondarily by the US Department of Defense Health Agency Human Research Protection Office. Institutional review boards at participating sites reviewed the protocol, addressed any local contextual factors, and ceded responsibility for ethics approval and study oversight to the single institutional review board.

Discussion

Herein, we report the rationale, design, and analysis plan for PREOXI, a 1,300-patient randomized trial of preoxygenation with noninvasive ventilation vs preoxygenation with oxygen mask among critically ill adults undergoing emergency tracheal intubation in the ED or ICU. Several important trial design elements bear additional consideration.

We selected hypoxemia (as a binary variable), rather than lowest oxygen saturation (a continuous variable), as the primary outcome for the trial for several reasons. First, experiencing hypoxemia in the range associated with an increased risk of adverse clinical outcomes (eg, cardiac arrest) may be more clinically relevant than changes in oxygen saturation within the normal range. For example, a change in oxygen saturation of 5 percentage points from 87% to 82% may be associated more closely with adverse outcomes than a change in oxygen saturation of 10 percentage points from 100% to 90%. Second, values for oxygen saturation are right-censored because oxygen saturation reaches 100% with a PaO₂ of approximately 100 mm Hg, but patients may have a higher PaO₂ after preoxygenation. The approach of analyzing hypoxemia as a binary variable, rather than lowest oxygen saturation, has been used by many prior trials and has been endorsed by airway experts.^24–29

We selected an oxygen saturation of < 85% as the threshold for the primary outcome based on several physiologic and procedural factors. First, an oxygen saturation of 85% corresponds with the inflection point on the oxyhemoglobin dissociation curve, at which further decrements in arterial oxygen concentrations result in rapid and critical desaturation.⁹ Second, an oxygen saturation of < 85% has been associated with an increased risk of cardiac arrest during tracheal intubation³⁰ and may be associated with increased mortality.⁵

Another trial design element that warrants consideration is the decision to allow the use of either a non-rebreather mask or a bag-mask device to provide preoxygenation in the oxygen mask group. Both devices can be used to provide supplemental oxygen without positive-pressure ventilation, but differ in some respects. A non-rebreather mask uses a loose-fitting interface and requires no provider involvement, whereas a bag-mask device requires a provider to hold the mask in place and can provide a tight mask seal. A bag-mask device also can be used to provide positive-pressure ventilation, which in the oxygen mask group of the current trial was allowed only after induction of anesthesia. Although these differences theoretically could affect preoxygenation, available evidence suggests that the two devices perform similarly.³¹ Any theoretical difference in preoxygenation between two devices providing oxygen is likely to be much smaller than the difference between these devices, which provide supplemental oxygen alone, and devices that provide both supplemental oxygen and positive-pressure ventilation (ie, the noninvasive ventilation group in PREOXI). Allowing both types of oxygen masks routinely used in clinical care also increases the generalizability of trial results.

The use of positive-pressure ventilation after induction of anesthesia (the end of preoxygenation) is allowed in both trial groups. Evidence from a prior randomized trial conducted among patients in the ICU suggests that bag-mask ventilation between induction and laryngoscopy reduces the risk of hypoxemia.² However, the safety and efficacy of bag-mask ventilation has not been evaluated among patients in the ED. Some experts have advocated avoiding bag-mask ventilation among patients in the ED who are less likely to have fasted than patients in the ICU and therefore may be at higher risk of aspiration.³² Because PREOXI enrolls patients in both the ED and ICU and focuses on preoxygenation and not management between induction and laryngoscopy, trial protocol allows bag-mask ventilation after induction of anesthesia, but does not mandate it.

The final design consideration is the administration of supplemental oxygen via standard nasal cannula and high-flow nasal cannula. Because oxygen facemasks and facemasks for noninvasive ventilation must be removed during laryngoscopy, some experts advocate for the administration of supplemental oxygen via standard nasal cannula oxygen or high-flow nasal cannula during laryngoscopy. The randomized trials available to date have not demonstrated consistently a benefit from administration of supplemental oxygen via standard nasal cannula or high-flow nasal cannula during tracheal intubation.^15,33–38 Therefore, the trial protocol for PREOXI allows, but does not require, the administration of supplemental oxygen via standard nasal cannula and high-flow nasal cannula in both trial groups.

Interpretation

PREOXI will provide important evidence regarding the effectiveness of common preoxygenation strategies for the prevention of hypoxemia during emergency tracheal intubation. To aid in the transparency and interpretation of trial results, this article detailing the protocol and statistical analysis plan for PREOXI has been finalized before the conclusion of patient enrollment.

Supplementary Material

Additional information: The e-Appendix and e-Tables are available online under “Supplementary Data.”

Take-home Points.

Study Question:

Does preoxygenation with noninvasive ventilation prevent hypoxemia during tracheal intubation of critically ill adults compared with preoxygenation with an oxygen mask?

Results:

This article describes the protocol and statistical analysis plan for the Pragmatic Trial Examining Oxygenation Prior to Intubation investigation of preoxygenation with noninvasive ventilation vs with oxygen mask.

Interpretation:

Prespecifying the full statistical analysis plan before completion of enrollment increases rigor, reproducibility, and transparency of the trial results.

ABBREVIATIONS:

PREOXI

Pragmatic Trial Examining Oxygenation Prior to Intubation