Preoxygenation With Flush Rate Oxygen: Comparing the Nonrebreather Mask With the Bag-Valve Mask

Study objective Nonrebreather masks and bag-valve masks are used for preoxygenation before emergency intubation. Flush rate oxygen delivered with a nonrebreather mask is noninferior to bag-valve mask oxygen at 15 L/min. We seek to compare the nonrebreather mask with flush rate oxygen to a bag-valve mask with flush rate oxygen (with and without inspiratory assistance) and determine whether the efficacy of bag-valve mask with flush rate oxygen is compromised by a simulated mask leak. Methods We conducted 2 prospective studies in healthy, adult volunteers. All devices in both studies used flush rate oxygen, achieved by rotating the flowmeter dial counterclockwise until it could not be rotated farther, which delivered oxygen at 40 to 60 L/min. Study 1 compared preoxygenation with nonrebreather mask to bag-valve mask (modified with a one-way exhalation port) with and without a simulated mask leak. Study 2 compared nonrebreather mask to bag-valve mask with inspiratory assistance. The primary outcome was FeO 2 . For each comparison, we prespecified a noninferiority margin of FeO 2 for the nonrebreather mask (compared with the bag-valve mask, bag-valve mask with mask leak, and bag-valve mask with assistance) of 10%. Results Thirty subjects were enrolled in study 1 and 27 subjects were enrolled in study 2. For study 1, mean FeO 2 values for nonrebreather mask, bag-valve mask, and bag-valve mask with leak were 81% (95% confidence interval [CI] 78% to 83%), 76% (95% CI 71% to 81%), and 30% (95% CI 26% to 35%), respectively. FeO 2 for the nonrebreather mask was noninferior to the bag-valve mask at flush rate (difference 5%; 95% CI –1% to 10%). FeO 2 was higher for the nonrebreather mask compared with the bag-valve mask with a simulated mask leak (difference 51%; 95% CI 46% to 55%). For study 2, mean FeO 2 values for nonrebreather mask and bag-valve mask with assistance were 83% (95% CI 80% to 86%) and 77% (95% CI 73% to 80%), respectively. FeO 2 for the nonrebreather mask was noninferior to the bag-valve mask with assistance at flush rate (difference 6%; 95% CI 3% to 10%). Conclusion With flush rate oxygen, the nonrebreather mask is noninferior to the bag-valve mask, with and without inspiratory assistance. Bag-valve mask performed poorly with a mask leak, even with flush rate oxygen. Flush rate oxygen with a nonrebreather mask is a reasonable default preoxygenation method in spontaneously breathing patients with no underlying respiratory pathology.