Article, Respiratory Medicine

Face mask leak with nasal cannula during noninvasive positive pressure ventilation: A randomized crossover trial

a b s t r a c t

Background: Nasal cannula can achieve Apneic oxygenation during emergency intubation. However, pre-proce- dure nasal Cannula placement may be difficult in patients undergoing non-invasive positive pressure ventilation (NPPV) prior to intubation. Our objective was to compare mask leak during NPPV with versus without simulta- neous application of nasal cannula. We hypothesized mask leak would be no worse with concomitant use of nasal cannula (non-inferiority design).

Methods: We performed a randomized crossover non-inferiority study of healthy volunteers. We randomized subjects undergoing 60 s trials of NPPV (10 cm H2O continuous positive airway pressure) to either NPPV alone (NPPV-a) or NPPV with nasal cannula at 15 L/min (NPPV-nc). After a brief rest period, all subjects underwent the alternative intervention. The primary outcome was time averaged mask leak over 60 s (L/min). We defined a non-inferiority margin of 5 L/min.

Results: We enrolled 64 subjects. Mean time-averaged mask leak was 2.2 L/min for NPPV-a versus 4.0 L/min for NPPV-nc for a difference of 1.7 L/min (one-sided 95% CI -? to 3.2 L/min). NPPV-a resulted in higher mean min- ute volume received (13.5 versus 12.2 L) and higher mean respiratory rates (14.8 versus 13.5 breaths per minute).

Conclusion: The addition of nasal cannula during NPPV does not significantly increase mask leak. The simulta- neous application of nasal cannula with NPPV may be a useful strategy to streamline airway management among patients undergoing NPPV prior to intubation.

Introduction

Background

Preoxygenation techniques optimize intubating conditions during Emergency airway management by prolonging time to desaturation [1]. Historically, the standard technique entailed administering a high fraction of inspired oxygen (FiO2) via non-rebreather mask (NRB) for 3 min or eight vital capacity breaths to wash out nitrogen in the lungs and achieve preoxygenation [2]. However, these strategies may not be practical or sufficient in critically ill patients and may lead to peri-intu- bation hypoxia.

Another option to minimize peri-procedural desaturation is apneic oxygenation via standard nasal cannula placement during intubation [1]. This strategy can prolong safe apnea time, increase time oxygen

* Corresponding author at: MCHE-EMR, 3551 Roger Brooke Dr., Fort Sam Houston, TX 78234, United States.

E-mail address: [email protected] (D.J. Brown).

1 Department of Emergency Medicine, Emory University School of Medicine, Atlanta, GA.

saturations remain above 95%, and limit desaturations during intuba- tion [3,4]. Apneic oxygenation via standard nasal cannula is associated with increased endotracheal intubation First pass success without hyp- oxemia in the Emergency Department (ED) setting [5].

Patients exhibiting shunt physiology may not achieve adequate oxy- gen saturations using these techniques alone. They may require pre-in- tubation non-invasive positive pressure ventilation (NPPV) to augment mean airway pressure, improving oxygenation and ventilation of shunted alveoli. Critically ill patients preoxygenated with NPPV have higher mean oxygen saturations prior to intubation and higher peri-in- tubation nadirs [6]. Pre-intubation NPPV has also been shown to benefit patients with primary lung disorders not involving shunt, such as asth- ma and COPD [7].

Patients requiring NPPV to improve oxygen saturations prior to intu- bation may benefit from continuation of passive oxygenation via nasal cannula during the intubation Apneic period [8]. Rapid transition from preoxygenation to intubation is imperative in these critically ill patients. Hence, simultaneous application of nasal cannula during NPPV prior to intubation would be preferable to application of nasal cannula after discontinuing NPPV prior to initiating intubation. One potential concern with the simultaneous application of NPPV and nasal cannula is that the

https://doi.org/10.1016/j.ajem.2017.10.055 0735-6757/

nasal cannula tubing may compromise the NPPV mask seal and there- fore compromise delivery of positive pressure during preoxygenation.

Study objective

The objective of this study is to compare measured mask leak flow (L/min) while on NPPV in healthy volunteers with versus without the simultaneous use of a nasal cannula. We hypothesize that mask leak will not be significantly greater with the use of a nasal cannula versus not using a nasal cannula (non-inferiority design).

Methods

Study design and setting

We performed a randomized crossover non-inferiority study using healthy volunteers. The study setting was an urban academic tertiary care hospital. Our institutional review board reviewed and approved the study. We registered the study on ClinicalTrials.gov (NCT02743936).

Subjects

We recruited a convenience sample of staff members affiliated with our ED. Subjects included medics, nurses, residents, and attending phy- sicians. Inclusion criteria consisted of healthy adult volunteers aged 18 years or older. Exclusion criteria included craniofacial abnormalities precluding the application of either nasal cannula or NPPV, inability to tolerate NPPV during the acclimation phase of the protocol, or active cardiac or pulmonary disease (including any respiratory infection).

We obtained written informed consent from all subjects. No subjects received compensation for participating. We documented subject flow in accordance with the Consolidated Standards for Reporting Trials (CONSORT) Statement (Fig. 1) [9].

Study protocol

We randomized subjects to one of two initial study arms, NPPV alone (NPPV-a) or NPPV with nasal cannula (NPPV-nc). In both study arms, subjects underwent NPPV using a Respironics V60 non-invasive ventilator with a Respironics AF531 EE headgear and mask (Phillips Healthcare, Andover, MA). This is a facemask NPPV system that covers the patient’s nose and mouth. Investigators determined subject mask sizes using standard sizing charts accompanying each mask. For the NPPV-nc arm only, subjects underwent NPPV with the simultaneous use of an AirLife standard nasal cannula (CareFusion, San Diego, CA). After screening, consent, and enrollment, we allocated subjects to their initial study arm using a randomization sequence with permuted blocks. Both subjects and investigators were aware of the subject alloca- tion (open label design).

We administered Continuous positive airway pressure at 10 cm H2O for NPPV in both study arms. We added nasal cannula with 15 L/min oxygen flow in the NPPV-nc arm. For the NPPV-a arm, subjects underwent NPPV only without placement of a nasal cannula.

Prior to the start of each intervention investigators fit the NPPV facemask. In the NPPV-nc arm we placed a nasal cannula in the standard fashion before placing the NPPV facemask. We instructed subjects to relax and breathe naturally. At the start of both interventions, subjects underwent 2 min of acclimation to facilitate spontaneous restful venti- lation and ensure appropriate mask fit. Subjects tolerating the acclima- tion period continued spontaneous restful ventilation for the 60 s data collection period. We made no adjustments to mask fit during the 1 min of data collection. All subjects underwent a two-minute rest (washout period) before proceeding to the alternate intervention (Fig. 1).

Measurements

Investigators obtained video recordings of the Respironics V60 ven- tilator data output including measurements of time-averaged mask leak flow (L/min) and tidal volume (L) for each breath. The recordings allowed accurate data collection and transcription of these parameters for every breath taken by each study subject during the trial period. The V60 ventilator measures flow rate and pressure at the machine out- put and patient mask, comparing both ends of the patient circuit to cal- culate measurements and perform automatic leak compensation. Regarding mask leak flow, the V60 ventilator compares end-exhalation actual flow of each respiratory cycle and the original baseline flow to es- timate unintentional leak at each end-exhalation (reported in L/min), assuming that any discrepancy is due to mask leak. Regarding tidal vol- ume, the ventilator compares delivered expiratory and inspiratory tidal volumes and assumes discrepancies represent mask leak [10]. We situ- ated the ventilator monitors such that the subjects could not see and use the monitor output as Real-time feedback to adjust breathing or positioning.

After undergoing both interventions, investigators asked subjects to rate their discomfort associated with the NPPV with and without nasal cannula using a verbal numerical rating scale (VNRS). This scale ranged from 0 (“no discomfort”) to 10 (“maximal discomfort”). Investigators recorded these responses onto hard-copy data collection forms.

The primary outcome measure was time-averaged mask leak flow (L/min). Secondary outcomes included received minute volume over the minute-long NPPV period (L), respiratory rate (breaths per minute), the percentage of breaths with any mask leak, and subject discomfort as reported by VNRS (0-10).

Data analysis

We based our sample size estimate upon the primary outcome of time-averaged mask leak for which we assumed normally-distributed data. We planned a non-inferiority analysis. Specifically, the null hy- pothesis was that mask leak would be at least 5 L/min (non-inferiority margin) higher in the NPPV-nc arm compared to the NPPV-a arm. The alternative hypothesis was that we would observe no such difference in mask leak. Given our non-inferiority design, we planned one-sided inferential statistical testing and so assumed ? = 0.025 as is the conven- tion for non-inferiority testing instead of ? = 0.05 as is the convention in superiority testing [11,12]. We anticipated standard deviation in mask leak measurements of 9 L/min based on preliminary data. Given these assumptions, enrollment of 63 subjects would achieve 80% power using one-sided to reject the null hypothesis (that mask leak is greater with NPPV-nc compared to NPPV-a) when the alternative is true. Rejecting the null hypothesis would lead to the conclusion that NPPV-nc is non-inferior (in terms of greater mask leak) compared to NPPV-a.

Investigators double entered all video recording and hard-copy form data into a secure Excel database (version 14; Microsoft, Redmond, WA). We exported all data into SPSS for statistical analysis (version 22; IBM, Armonk, NY). We calculated descriptive statistics to report pa- tient characteristics. We compared the primary outcome of mean mask leak by calculating the one-sided 95% confidence interval (CI) of the dif- ference in mean mask leak between NPPV-a versus NPPV-nc (equiva- lently a 97.5% confidence interval given our assumption that ? = 0.025). If the upper bound of the one-sided 95% CI was less than the in- feriority margin (5 L/min) then we rejected inferiority. We chose 5 L/min as the non-inferiority margin based upon consensus between the investigators that this value represents the minimal clinically signif- icant difference. We repeated these analyses of the primary outcome stratified by intervention sequence. We performed similar one-sided analyses of secondary outcomes including the percentage of breaths with any mask leak, mean minute volume received, mean respiratory

Fig. 1. Consolidated Standards of Reporting Trials (CONSORT) diagram of enrollment, allocation, follow-up, and analysis of subjects undergoing non-invasive positive pressure ventilation (NPPV).

Table 1

Subject demographics and characteristics (n = 64).

All (n = 64)

First Arm

NPPV-a (n = 32)

NPPV-nc (n = 32)

n

64

32

32

Mean age, years

32.6

32.4

32.8

Male sex (%)

48 (75%)

23 (48%)

25 (52%)

Mean height, cm

Mask size (%)

175.5

175.8

175.0

Small

18 (28%)

10 (31%)

8 (25%)

Medium

38 (59%)

20 (63%)

18 (56%)

Large

8 (13%)

2 (6%)

6 (19%)

Abbreviations: cm-centimeters; NPPV-a-non-invasive positive pressure ventilation alone; NPPV-nc-non-invasive positive pressure ventilation with nasal cannula.

rate, and median discomfort. We calculated one-sided 95% CI of ordinal variables using a Hodges-Lehmann estimator.

Results

Characteristics of study subjects

All 64 healthy volunteers screened were eligible for study inclusion and agreed to participate. No recruited subjects declined to participate or met any exclusion criteria. All male subjects were cleanly shaven except for two subjects. All enrolled subjects underwent randomization to initial study arm: 32 to NPPV-a and 32 to NPPV-nc. All participants successfully crossed over to the alternate arm and completed the study. There were no losses to follow-up (Fig. 1). Patient characteristics were comparable between subjects initially allocated to either arm (Table 1).

Main results

Mean time-averaged mask leak was 2.2 L/min for NPPV-a versus

4.0 L/min for NPPV-nc for a difference of 1.7 L/min (one-sided 95% CI

-? to 3.2 L/min, Table 2, Fig. 2). The mean difference in mask leak be- tween the two interventions among subjects allocated to first undergo NPPV-a was 2.0 L/min (one-sided 95% CI -? to 4.5 L/min). The differ- ence in mean mask leak between the two interventions among subjects allocated to first undergo NPPV-nc was 1.5 L/min (one-sided 95% CI -? to 4.5 L/min). The highest recorded time-averaged mask leak was

41.9 L/min for a NPPV-a experiment and 53.6 L/min for a NPPV-nc experiment.

Regarding secondary outcomes, we observed a trend towards a lower proportion of breaths with any detectable mask leak with NPPV-a as compared to NPPV-nc (Table 2, Fig. 3). NPPV-a appeared to result in slightly higher mean minute volume received (13.5 versus 12.2 L, Table 2 and Fig. 4) and higher mean respiratory rates (14.8 versus 13.5 breaths per minute, Table 2). Finally, we observed a trend towards lower median subject 10-point VNRS discomfort scores: 1 for NPPV-a versus 3 for NPPV-nc (Fig. 5).

Discussion

Overview of results

We sought to determine whether the addition of nasal cannula to NPPV would result in greater mask leak. Our results suggest that the mask leak during NPPV with the addition of nasal cannula is not signif- icantly greater than the mask leak with a NPPV mask alone. There was a trend to increased time-averaged mask leak with the addition of nasal cannula, but the increase of 1.7 L/min is likely not clinically significant. As with many modern noninvasive ventilators, this particular NPPV ma- chine can deliver flow rates up to 130 L/min when delivering 100% FiO2 and compensate for leaks up to 60 L/min, which is larger than any leak recorded in this study [13]. These results suggest that the simultaneous use of nasal cannula during NPPV is not likely to compromise ventilation.

Apneic oxygenation has been shown to increase safe apnea time in obese patients in the Operating Room environment [3], reduce the inci- dence of peri-intubation desaturation in patients with intracranial hem- orrhage [4], and increase endotracheal intubation first pass success without hypoxia in the ED [5]. In contrast to this data, the FELLOW study did not find any difference in lowest Arterial oxygen saturations among patients intubated in the ICU setting randomized to apneic oxy- genation versus Standard care [14]. However, this study did not fully replicate emergency airway management performed in the ED because approximately one-third of the patients received BVM ventilation dur- ing the apneic period, in contrast to common ED practice of not ventilat- ing during the apneic period to avoid inducing aspiration. It is likely that apneic oxygenation would be more beneficial in patients who did not receive BVM ventilations during the apneic period.

The use of nasal cannula for apneic oxygenation in addition to NPPV is a relatively unproven modality. One recent study evaluated the use of high flow nasal cannula in addition to NPPV for preoxygenation and apneic oxygenation of hypoxic ICU patients. This study found a lower rate of severe hypoxia and higher oxygen saturation nadirs during the intubation procedure [15]. Grant et al. recently published a case series of eight ED patients at risk for hypox- ia that were managed safely using NPPV and NC with ventilator de- livered breaths during the apneic period. All patients improved their oxygen saturations and had no peri-intubation desaturations [16]. Additionally, a study of healthy volunteers found comparable mean end tidal oxygen levels (EtO2, a marker of denitrogenation) utilizing a BVM with a PEEP (positive end-expiratory pressure) valve plus nasal cannula compared to BVM with PEEP valve alone (75.5% versus 78.9%) [17].

A similar study of healthy volunteers examined EtO2 levels during simultaneous use of nasal cannula and BVM versus BVM alone. In scenarios without simulated mask leak, EtO2 with simul- taneous nasal cannula and BVM was comparable to BVM alone (75% versus 79%). In scenarios with a simulated mask leak created by inserting a nasogastric tube under the BVM, EtO2 was 66% for BVM plus nasal cannula and 41% for BVM alone. The study had similar results for experiments using a non-rebreather mask in- stead of BVM [18].

Table 2

Outcomes for each study arm (n = 64).

Variables

NPPV-a (n = 64)

NPPV-nc (n = 64)

Effect Size Difference (one-sided 95% CI)

Mean time averaged mask leak, L/min (95% CI)

2.2 (0.8-4.1)

4.0 (2.0-6.5)

1.7 (-?-3.2)

Mean breaths with any mask leak, % (95% CI)

28.1 (18.9-37.8)

33.3 (22.6-45.1)

5.2 (-?-18.0)

Mean minute volume, L (95% CI)

13.5 (12.3-14.7)

12.2 (10.9-13.6)

-1.3 (-? to -0.5)

Mean respiratory rate, breaths/min (95% CI)

14.8 (13.3-16.3)

13.5 (12.3-14.8)

-1.2 (-? to -0.5)

Median discomfort, 0-10 VNRS (IQR)

1 (1-2)

3 (2-5)

2 (-?-2)

Abbreviations: CI-confidence interval; IQR-interquartile range; L-liters; NPPV-a: non-invasive positive pressure ventilation alone; NPPV-nc: non-invasive positive pressure ventilation with nasal cannula; VNRS-verbal numerical rating scale.

Fig. 2. Time-averaged mask leak over 1 min. The vertical axis represents time-averaged mask leak (liters/minute) as measured by a ventilator during 1 min of non-invasive positive pressure ventilation (10 cm H2O of continuous positive airway pressure) without nasal cannula. The horizontal axis represents time-averaged mask leak during 1 min of non-invasive positive pressure ventilation with nasal cannula. Each black diamond represents the measurements for one of 64 study subjects.

These studies suggest that the addition of nasal cannula to BVM has a minimal impact on mask seal insofar as it does not compromise oxygenation. At the same time, these data demonstrate that, in the

presence of significant mask leak (as simulated by a nasogastric tube under the mask), the addition of nasal cannula improves oxy- genation. However, oxygen measurements in this situation are

Fig. 3. Proportion of breaths with any mask leak. The vertical axis represents the proportion of breaths (%) noted to have any mask leak as measured by a ventilator during 1 min of non- invasive positive pressure ventilation (10 cm H2O of continuous positive airway pressure). The horizontal axis represents individual study subjects (n = 64). The subject numbers reflect ascending proportions of breaths with mask leak during experiments without nasal cannula and do not reflect chronological order of study participation. Each pair of bars represents data for an individual subject. The gray bars represent proportions of breaths with mask leak during experiments with NPPV-a. The black bars represent proportions of breaths with mask leak during experiments with simultaneous NPPV-nc.

Fig. 4. Minute volumes. The vertical axis represents the minute volumes (liters) as measured by a ventilator during 1 min of non-invasive positive pressure ventilation (10 cm H2O of continuous positive airway pressure). The horizontal axis represents individual study subjects (n = 64). The subject numbers reflect ascending minute volumes during experiments without nasal cannula and do not reflect chronological order of study participation. Each pair of bars represents data for an individual subject. The gray bars represent minute volumes during experiments with NPPV-a. The black bars represent minute volumes during experiments with simultaneous NPPV-nc.

lower than those observed without a substantial mask leak (75% ver- sus 66%) [18]. Consequently, it seems there are levels of mask leak which will result in desaturation for which the addition of nasal

Fig. 5. Boxplot of subject discomfort as reported on an ordinal verbal numerical rating scale from 0 to 10. The midlines represent the medians. The bottom and top box edges represent the 25th and 75th percentiles, respectively. The bottom and top “whiskers” represent the minimum and maximum values, respectively. Subjects reportedly significantly less discomfort with non-invasive positive pressure ventilation alone compared to non-invasive positive pressure ventilation with simultaneous application of nasal cannula.

cannula will not fully compensate. In these studies, the mask leak caused by the nasal cannula alone was not sufficient to compromise oxygenation. Our study provides important additional information by specifically quantifying the mask leak associated with use of nasal cannula during NPPV. Our results offer additional data supporting that the mask leak caused by nasal cannula is minimal and clinically insignificant.

While the available data suggests simultaneous nasal cannula with NPPV does not materially decrease oxygen saturations or con- tribute to mask leak, our results suggest the addition of nasal cannula decreases minute ventilation. During the NPPV-nc experiments, me- dian respiratory rate and median minute ventilation was significant- ly lower compared to the experiments using the NPPV mask only. The significance of this finding is unclear. Given the minimal mask leak in our study, we believe it is unlikely that the decreased minute volumes reflect increased difficulties with ventilation. The decreased minute volumes may instead reflect a change in respiratory effort with additional mean airway pressure augmentation provided by the nasal cannula.

Future research should examine the impact of simultaneous nasal cannula and NPPV on patients experiencing respiratory distress rather than healthy subjects. Much of the existing data (to include our study) is from healthy volunteers and suggests that the addition of nasal can- nula does not compromise NPPV mask seal. The next step will be to demonstrate that the addition of nasal cannula does not compromise mask leak in actual ED patients. Measuring ventilation will also be im- portant given our findings to suggest lower minute volumes in patients using nasal cannula with NPPV. It will also be important to examine the subset of this patient population undergoing intubation to determine if the addition of nasal cannula optimizes patient survival and peri-intu- bation oxygen saturations given the intent of achieving apneic oxygen- ation [3,4].

Limitations

Our study population was limited to cooperative young healthy vol- unteers with a larger proportion of males than females. In addition, all but two males were cleanly shaven and all subjects had grossly normal dentition. Data collection occurred during a 60 s period, shorter than the duration a patient may utilize this treatment in a clinical setting. This limits generalizability to patients who may have more challenging air- way characteristics such as facial hair, poor dentition, craniofacial ab- normalities, significant cardiopulmonary abnormalities, or significant anxiety related to their critical illness.

Blinding was not feasible with our study design. Objective end- points minimized investigator bias. Subjects had the potential to bias our results with changes in respiratory effort and mask leak. We mini- mized this risk by instructing all subjects to relax and breathe naturally and situated the ventilator monitors outside each subject’s line of sight to remove real-time feedback.

For this study we utilized the Respironics V60 ventilator for both our intervention and outcome measurements. This ventilator pro- vides breath-to-breath compensation for mask leaks up to 60 L/min [13]. It is unclear the extent to which the mask leaks and volumes we observed in our study will extrapolate to other ventilator or mask models [19-23], including disposable NPPV models commonly used in the pre-hospital setting that cannot compensate for mask leak.

Additionally, it is unclear if the addition of nasal cannula flow will significantly interfere with the ventilators ability to calculate mask leak and minute ventilation. Based upon the Respironics V60 Ventilator user Manual, the ventilator calculates leak as the difference in the ma- chine end exhalation flow from the baseline flow. The ventilator further re-calculates baseline at the end of each exhalation and as such should control for the effects of nasal cannula flow in its mask leak computa- tions [10]. Nevertheless, we cannot definitively prove this is the case as the exact formula and software are proprietary (personal communi- cation, Respironics representative).

Conclusions

This study suggests that the mask leak resulting from the simulta- neous use of nasal cannula at 15 L/min and NPPV via facemask at CPAP 10 cm H2O is not significantly greater as defined by a non-inferior- ity margin of 5 L/min than the mask leak with a NPPV facemask alone.

Disclaimer

The view(s) expressed herein are those of the author(s) and do not reflect the official policy or position of Brooke Army Medical Center, the

U.S. Army Medical Department, the U.S Army Office of the Surgeon Gen- eral, the Department of the Army or the Department of Defense, or the

U.S. Government.

Funding

None.

Conflicts of interest

DB reports no conflicts of interest; SC reports no conflicts of interest; MDA reports no conflicts of interest.

References

  1. Weingart SD, Levitan RM. Preoxygenation and prevention of desaturation during emergency airway management. Ann Emerg Med 2012;59(3):165-75 [e1].
  2. Carmichael FJ, et al. Preoxygenation: a study of denitrogenation. Anesth Analg 1989;

    68(3):406-9.

    Ramachandran SK, et al. Apneic oxygenation during prolonged laryngoscopy in obese patients: a randomized, controlled trial of nasal Oxygen administration. J Clin Anesth 2010;22(3):164-8.

  3. Sakles JC, et al. Apneic oxygenation is associated with a reduction in the incidence of hypoxemia during the RSI of patients with intracranial hemorrhage in the emergen- cy department. Intern Emerg Med 2016;11(7):983-92.
  4. Sakles JC, et al. First pass success without hypoxemia is increased with the use of ap- neic oxygenation during rapid sequence intubation in the Emergency Department. Acad Emerg Med 2016;23(6):703-10.
  5. Baillard C, et al. noninvasive ventilation improves preoxygenation before intubation of hypoxic patients. Am J Respir Crit Care Med 2006;174(2):171-7.
  6. Ram FS, et al. Non-invasive positive pressure ventilation for treatment of respiratory failure due to exacerbations of chronic obstructive pulmonary disease. Cochrane Da- tabase Syst Rev 2004(3):Cd004104.
  7. Weingart SD. Preoxygenation, reoxygenation, and delayed sequence intubation in the emergency department. J Emerg Med 2011;40(6):661-7.
  8. Schulz KF, et al. CONSORT 2010 statement: updated guidelines for reporting parallel group randomized trials. Ann Intern Med 2010;152(11):726-32.
  9. Philips Healthcare. Respironics V60 ventilator user manual. Andover, MA: Philips Healthcare; 2009.
  10. Kaji AH, Lewis RJ. Are we looking for superiority, equivalence, or noninferiority? Asking the right question and answering it correctly. Ann Emerg Med 2010;55(5): 408-11.
  11. Kaji AH, Lewis RJ. Noninferiority trials: is a new treatment almost as effective as an- other? JAMA 2015;313(23):2371-2.
  12. Oto J, et al. A comparison of leak compensation during pediatric noninvasive venti- lation: a lung model study. Respir Care 2014;59(2):241-51.
  13. Semler MW, et al. Randomized trial of apneic oxygenation during endotracheal intu- bation of the critically ill. Am J Respir Crit Care Med 2016;193(3):273-80.
  14. Jaber S, et al. apnoeic oxygenation via high-flow nasal cannula oxygen combined with non-invasive ventilation preoxygenation for intubation in hypoxaemic patients in the intensive care unit: the single-centre, blinded, randomised controlled OPTINIV trial. Intensive Care Med 2016;42(12):1877-87.
  15. Grant S, et al. Ventilator-assisted preoxygenation: protocol for combining non-inva- sive ventilation and apnoeic oxygenation using a portable ventilator. Emerg Med Australas 2016;28(1):67-72.
  16. Groombridge C, et al. Assessment of common preoxygenation strategies outside of the operating room environment. Acad Emerg Med 2016;23(3):342-6.
  17. Hayes-Bradley C, et al. Efficacy of nasal cannula oxygen as a preoxygenation adjunct in emergency airway management. Ann Emerg Med 2015;68(2):174-80.
  18. Amack AJ, Barber GA, Ng PC, Smith TB, April MD. Comparison of ventilation with one-handed mask seal with an intraoral mask versus conventional cuffed face mask in a cadaver model: a randomized crossover trial. Ann Emerg Med 2017; 69(1):12-7.
  19. Lee HY, et al. The performances of standard and ResMed masks during bag-valve- mask ventilation. Prehosp Emerg Care 2013;17(2):235-40.
  20. Stewart RD, et al. Influence of mask design on Bag-mask ventilation. Ann Emerg Med

    1985;14(5):403-6.

    Zobrist B, Casmaer M, April MD. single rescuer ventilation using a bag-valve mask with internal handle: a randomized crossover trial. Am J Emerg Med 2016;34: 1991-6.

  21. Reed P, Zobrist B, Casmaer M, Schauer SG, Kester N, April MD. Single rescuer venti- lation using a bag valve mask with removable external handle: a randomized cross- over trial. Am J Emerg Med 2017;15:1-6.