By Vibhu Sharma, MD, MS

Attending Physician, Division of Pulmonary and Critical Care Medicine, John H. Stroger Hospital of Cook County; Assistant Professor of Medicine, Rush University Medical Center, Chicago

Dr. Sharma reports no financial relationships relevant to this field of study.

SYNOPSIS: In this multicenter, randomized, unblinded trial, bag-mask ventilation from time of induction to laryngoscopy was associated with higher oxygen saturations and lower incidence of severe hypoxemia compared to those not “bagged.”

SOURCE: Casey JD, Janz DR, Russell DW, et al. Bag-mask ventilation during tracheal intubation of critically ill adults. N Engl J Med 2019;380:811-821.

Casey et al designed a trial to determine whether bag-mask ventilation (BMV) starting at induction prevents a drop in oxygen saturation (SpO2). Further, the authors wanted to know whether BMV increased the risk for aspiration. The primary outcome was the lowest SpO2 documented between induction and two minutes after endotracheal intubation. The secondary outcome was the incidence of severe hypoxemia (SpO2 < 80%). Other prespecified outcomes included operator-visualized aspiration (oropharyngeal or gastric) and the presence of a new opacity on chest X-ray. Exploratory SpO2 outcomes also were compared (lowest SpO2 < 90%; lowest SpO2 < 70%; and decrease in SpO2). Seven academic medical center ICUs in the United States participated; 266 of 667 patients enrolled were excluded (those with emesis, hematemesis, hemoptysis, and those with a full stomach. Patients who needed urgent intubation also were excluded, as were those with severe acidemia or severe hypoxemia in whom ventilation was deemed indicated during induction.

Thus, 401 patients undergoing tracheal intubation were randomized to receive either ventilation with a bag-mask device or no ventilation between induction and laryngoscopy. Overall, 60% of enrolled patients experienced hypoxemic respiratory failure, and half had sepsis or septic shock. All patients received preoxygenation. The technique was not mandated and left to the operator, with an equivalent proportion in each group receiving positive pressure oxygenation (defined as either bi-level positive airway pressure [BiPAP], high-flow nasal cannula [HFNC] oxygen, or BMV). BiPAP and HFNC settings were not standardized.

One group (BMV group) was randomized to receive BMV with 5-10 cm H2O positive end-expiratory pressure (PEEP) and 15 L/minute oxygen at 10 breaths per minute starting from the time of induction until laryngoscopy was performed. The other group (no BMV group) received no ventilation. Oral airways and jaw thrust chin lift maneuvers were allowed as needed. Apneic oxygenation (Apox) was not mandated post-induction in the no BMV group, but was allowed per intubator preference (78% chose to provide Apox with a 100% non-rebreather mask or nasal cannula). BMV was not allowed per the protocol in patients randomized to the no BMV group unless first laryngoscopy failed (n = 44) or SpO2 < 90% necessitated BMV (n = 5). This latter group was analyzed with their initial assigned group (i.e., no BMV).

All patients underwent rapid sequence intubation with etomidate/ketamine as the induction agent and neuromuscular blockade. Almost all patients in each group were intubated by trainees (critical care fellows or anesthesia residents composed approximately 96% in each group). The intubators in each group had intubated a median of 50 patients (range, 32-100), and the first-pass success rate was about 80% in each group.

The authors found that BMV starting at induction compared to no BMV was associated with a median lowest oxygen saturation of 96% compared to 93%, a statistically significant difference (P = 0.01). Those patients with the lowest SpO2 at the time of induction benefited the most from BMV. If BMV was used, the lowest SpO2 for the sickest patients (with SpO2 < 97% at induction), after all preoxygenation interventions had ceased and induction medications were administered, was higher (89%) compared to the lowest SpO2 for those who did not receive BMV (82%). Severe hypoxia (SpO2 < 80%) occurred less frequently in the BMV group (relative risk [RR], 0.48; 95% confidence interval [CI], 0.30-0.77). Similarly, the lowest SpO2 also was higher in those “bagged” with SpO2 ≥ 97% at induction. Overall, the median decrease in SpO2 from induction to two minutes after intubation was lower in the BMV group compared to the no BMV group (median decrease of 1% vs. 5%; mean difference, 4.5%; 95% CI, 2.2-6.8). The incidence of lowest SpO2 < 90% and < 70% also was lower in the BMV group. Visualized aspiration during intubation and new opacities on chest X-ray within the 24 hours following intubation were not higher among those receiving BMV.


Casey et al concluded that “bagging” patients during induction is safe and allows for higher SpO2 during a critical period of airway management. There are several caveats to the findings.

Patients who needed urgent intubation were excluded. In essence, the authors assessed the effect of study interventions in those patients in whom intubation could proceed with adequate procedures to preoxygenate. Indeed, median SpO2 in each group at the start of induction was 99%, a relative luxury in the setting of airway management in the ICU. Only a small number of patients in each group (27 in the BMV group and 17 in the no ventilation group) had SpO2 < 92% at start of induction. The exclusion criteria necessarily introduced selection bias, and the number of patients excluded was substantial. While this was appropriate to maintain safety, the study population skewed toward patients with relatively preserved ventilation/perfusion (V/Q) matching (those who could be “bagged up”).

Second, the preoxygenation technique was not standardized. HFNC, BiPAP, and BMV as techniques for preoxygenation were grouped as “preoxygenation with positive pressure.” In a post-hoc analysis, while the preoxygenation device did not seem to modify between-group differences in terms of lowest SpO2, preoxygenation device settings may have been selected intuitively to be more aggressive in patients perceived to be sicker. For example, noninvasive ventilation with higher pressure settings may have been selected for those patients perceived to have worse V/Q mismatch or a higher HFNC oxygen flow rate may have been selected for patients perceived to be sicker. Intubators were not blinded to any patient characteristics, including clinical diagnosis or SpO2 at start of preoxygenation, given the pragmatic nature of the study.

Third, the majority of patients were intubated by critical care fellows and anesthesia residents with approximately 50 intubations under their belt on average (range, 30-100). A survey of pulmonary/critical care fellowship programs in the United States revealed that 67% of fellows graduated with ≤ 50 laryngoscopies performed.1 Although Casey et al did not find that operator experience modified the effect of BMV on the lowest SpO2, the results of other studies have shown that the level of experience matters.2,3 Given the absence of harm and evidence of benefit, BMV during induction ought to be the standard, especially with trainees as intubators.

Fourth, while the proportion of patients with one or more difficult airway characteristics was no different between groups, those with obstructive sleep apnea were overrepresented in the no BMV group (26 vs. 15). This may have skewed toward lower SpO2 with a longer time to intubation and (potentially) a failed first laryngoscopy, especially with neuromuscular blockade. The mean difference between groups with respect to the lowest SpO2 was driven by two of seven sites (and mostly by a single center that recruited the largest number of patients). The authors offered no details with respect to characteristics of patients or preoxygenation techniques analyzed by institution.

Finally, none of the institutions performed preinduction gastric antral ultrasound. This technique has been well described and may prevent clinically relevant aspiration events around the time of intubation.4 If the stomach is empty, the effect of BMV on gastric aspiration events is irrelevant. BMV with oxygen delivered at 15 L/minute and 10 breaths per minute with 5-10 cm H2O PEEP during induction is associated with higher SpO2, prevents severe hypoxia (number needed to treat = 9 to prevent one episode of severe hypoxemia), and is not associated with a higher risk of aspiration. In carefully selected patients, BMV prevents critical desaturation. Desaturation by 5% from time of induction to laryngoscopy may be expected if no BMV is provided.


  1. Joffe AM, Liew EC, Olivar H, et al. A national survey of airway management training in United States internal medicine-based critical care fellowship programs. Respir Care 2012;57:1084-1088.
  2. Simpson GD, Ross MJ, McKeown DW, Ray DC. Tracheal intubation in the critically ill: A multi-center national study of practice and complications. Br J Anaesth 2012;108:792-799.
  3. Smischney NJ, Seisa MO, Heise KJ, et al. Predictors of arterial desaturation during intubation: A nested case-control study of airway management-part I. J Thorac Dis 2017;9:3996-4005.
  4. Koenig SJ, Lakticova V, Mayo PH. Utility of ultrasonography for detection of gastric fluid during urgent endotracheal intubation. Intensive Care Med 2011;37:627-631.