Emergency airway management is considered an essential skill in critical care settings (prehospital, emergency departments, intensive care units [ICUs]), but carries an inherent risk of critical hypoxemia and other major adverse events and complications reported in up to 40% to 45% of cases [1–3]. Hypoxemia before intubation, or significant desaturation during intubation, pose a risk peri-intubation cardiac arrest that occurs in up to one in 25 critically ill patients [3]. Recent studies indicate that peri-intubation hypoxemia and absence of preoxygenation carry significantly increased risk of intubation related cardiac arrest [4,5]. All patients require appropriate preoxygenation, positioning and hemodynamic optimization to allow for appropriate rapid sequence intubation (RSI), but patients with obesity, late term pregnancy, pediatric patients, and patients with acute hypoxemic respiratory failure are at particular risk of rapid desaturation and often require special attention or modifications to preoxygenation and/or intubation strategy.

Traditional RSI requires rapid successive administration of a sedative/hypnotic agent and a neuromuscular blocking agent without subsequent positive pressure ventilation before laryngoscopy. This requires adequate preoxygenation to provide an apnea time long enough for drug onset, laryngoscopy, and tube placement; a difficult task to achieve in patients with profound hypoxemia [6]. Preoxygenation should span the entire peri-intubation period to maximize oxygen reserve and account for continued oxygen consumption during the hypopneic and apneic periods [7]. Induction medications may take up to 90 seconds to optimize intubating conditions, and preoxygenation can be lost after five breaths following oxygen source removal [8]. Thus, optimizing preoxygenation includes continuing oxygen administration up until the period of apnea and often even during apnea to reduce the risk of cardiovascular collapse due to hypoxemia with intubation [8,9]. The differences in patient presentation and underlying physiology complicate planning a preoxygenation strategy. This review seeks to describe the methods for preoxygenation, appropriate endpoints to preoxygenation and the potential use of apneic oxygenation in patients undergoing rapid sequence intubation.

A MEDLINE search strategy using PubMed (US National Library of Medicine) was used with the following search terms from March 2008 to March of 2023: “rapid sequence intubation” OR “hypoxia” OR “intubation” OR “bag mask ventilation” OR “pregnancy critical care” OR “obesity critical care” OR “obesity intubation” OR “preoxygenation.” The search was conducted over a 4-month period from February to May of 2023. Articles related to the emergency department, ICU, and operating room, and physiologic studies, where appropriate, were included. Non-English articles, those related to pediatric patients (<18 years of age), and animal models were excluded. Articles related to prehospital airway management were excluded due to differences in availability of equipment and setting. Other exclusion criteria included: case reports/series, supraglottic airways for preoxygenation, studies primarily pertaining to analysis of video laryngoscopy, fiberoptic intubation or direct laryngoscopy, and those related to COVID-19 viruses.

This search resulted in an extensive list of resources, of 30,993 results. Following a search of the best matches and articles meeting the above criteria, an initial 10,000 articles were reviewed for evaluation of inclusion criteria with 298 further isolated out of this subgroup based on appropriate criteria, such as discussion of preoxygenation specifically in the emergency department. In the case of meta-analyses frequently referencing the same primary sources, with five or more common articles referenced between sources, attempts were made to select the article with the most recent publication or with inclusion of more randomized controlled trials or other sources. There were limited studies found on pregnancy related randomized controlled trials (RCTs) and meta-analysis and reviews were considered adequate sources as this is a special patient population less likely to be subject to RCTs. The final articles selected are included in Table 1 [1,5,6,10–27].

Preoxygenation prior to RSI is the standard of care, but strategies have evolved since RSI was first described [11]. RSI requires preoxygenation with set endpoints based on available equipment and oxygenation values to increase the time to desaturation after induction. The time to desaturation, or “safe apnea time” is traditionally considered the time from administration of RSI medications until the saturation reaches 90%, and the fundamental goal of preoxygenation is to prolong this time interval. Methods for preoxygenation, ranging from a standard nasal cannula to positive pressure ventilation, are summarized in Table 2 [6,22,25,28–33].

As part of the preoxygenation strategy, the plan should include thresholds for adequate preoxygenation. In the emergency department, peripheral oxygen saturation (SpO₂) is frequently used with goal saturations of 100% for preoxygenation, and <90% often considered the stopping point for reoxygenation before severe desaturation (<80%) occurs [14]. Other traditionally common end points include absolute time or duration of tidal breathing (3 minutes) or vital capacity breaths of 100% oxygen (typically eight), although it has limitations in the critically ill [29]. End-tidal oxygen (EtO₂) in combination with other parameters, such as SpO₂, the partial pressure of oxygen in arterial blood (PaO₂), allows for granular evaluation of the underlying physiology to optimize preoxygenation methods [29]. EtO₂ has been more closely studied in anesthesia literature due to availability in the operating room with additional note of fraction of expired oxygen <85% shown to be inadequate for intubation attempts [13]. In the presence of shunt physiology or uncertain reliability of oxygen monitoring devices, PaO₂ provides a direct way to assess oxygenation of the blood and acts a gauge for the amount of gas exchange in the alveolar capillary beds [6,29]. The advantages and limitations of these methods are discussed in Table 3 [6,13,29].

Continuous nasal oxygen throughout the intubation period after induction, termed “apneic oxygenation,” can augment preoxygenation and potentially prolong the safe apnea time. The literature on apneic oxygenation is mixed and plagued with confounders. Gleason et al. [10], showed decreased desaturation when adding apneic oxygenation, except for patients with primary respiratory failure. In contrast, Semler et al. [21] demonstrated no difference in lowest arterial oxygen saturation in patients intubated in the ICU setting, though it should be noted that most patients were intubated for primary respiratory failure. An additional study by Caputo et al. [20] showed no benefit when applying a standard nasal cannula (NC) at flow rates >15 L/min after preoxygenation with bag valve mask (BVM), nonrebreathing masks, or noninvasive positive pressure ventilation. McQuade et al. [28] found an increase in EtO₂ with NC at rates >15 L/min, deleterious effect with <5 L/min, and no benefit when an adequate seal was achieved with BVM alone, as the addition of NC disrupted the mask seal, potentially explaining the study results by Caputo et al. [20]. Yet, in cases of a mask leak, there was benefit from adding a nasal cannula [28]. Russotto et al. [14] found that apneic oxygenation with low flow or high flow nasal cannula decreased desaturation.

A 2016 study by Sakles et al. [27] demonstrated increased first pass success without hypoxia in patients with apneic oxygenation, though it should be noted that these patients were primarily intubated for airway protection and traumatic injuries. In a 2017 meta-analysis, obesity, elective surgery, and those intubated emergently for nonpulmonary causes had improved oxygen saturations with apneic oxygenation further emphasizing the potential benefit from apneic oxygenation [24]. Cabrini et al. [12] and White et al. [24] demonstrated that there is likely no benefit from high flow NC (HFNC) administration during the apneic period which contrasts with operating room data showing prolonged safe apnea time and potential benefits in select patient populations. An additional study by Sakles et al. [34] from 2016 demonstrated decreased desaturation when apneic oxygenation was used for intubation.

While apneic oxygenation can be quite helpful in some patients, such as those intubated for primary neurologic injury or airway protection, and should be used, it is less likely to prevent desaturation in some patients with acute hypoxemic respiratory failure. In these patients, the underlying physiology presents significant challenges to preoxygenation, especially in patients with severe hypoxemic respiratory failure. Underlying pathophysiology must be considered when selecting the best preoxygenation strategy for a given patient. These include optimizing denitrogenation in all patients, maximizing functional residual capacity in at risk patients, reducing shunt physiology in high-risk patients, and recognizing failed preoxygenation in refractory patients.

Denitrogenation is the rate limiting step for preoxygenation in healthy patients, and at times in critically ill patients [11]. Thus, traditional preoxygenation with tidal breathing of 100% oxygen for 3 minutes or eight vital capacity breaths with a tight fitting mask is, in actuality, denitrogenation [29,35]. Methods of denitrogenation all involve overcoming nitrogen from ambient air with high flows of 100% oxygen to “washout” alveolar nitrogen, leaving only oxygen, water vapor, and CO₂. The nonrebreathing masks commonly used in the ED do not achieve a tight mask seal, thus room air is entrained around the mask with each inspiration [29,33]. This reduces the effective fraction of inspired oxygen (FiO₂) despite delivering 100% oxygen from the nonrebreather reservoir [33]. Groombridge et al. [33] and Mosier et al. [29] showed improved preoxygenation (measured by the fractional of exhaled oxygen) by adding a nasal cannula to attenuate the poor mask seal, but it does not completely compensate for the leak. Alternative methods for complete denitrogenation include using high flow nasal oxygen at flow rates greater than the inspiratory flow rate, or noninvasive positive pressure with a tight fitting mask [33]. EtO₂ is a relatively easy measure to obtain, either with a single breath, or continuous monitoring, and can be helpful to objectively evaluate when adequate denitrogenation has been achieved (EtO₂ >85%–90%) [7,8]. Fully denitrogenating the alveoli will result in a large alveolar oxygen tension, as the only other gasses available in the alveoli after denitrogenation would be water vapor and carbon dioxide as dictated by the alveolar gas equation.

Following denitrogenation, renitrogenation can rapidly occur when the oxygen source is removed before apnea [7,8]. West et al. [7] demonstrated renitrogenation may occur as quickly as 160 seconds, and Mosier et al. [8] found renitrogenation occurs after about five breaths once the oxygen source is removed in healthy patients. Gentle mask ventilation after induction can prevent renitrogenation, and has been shown to reduce the risk of severe hypoxemia [1].

The functional residual capacity (FRC) provides the lung volume for oxygen available during apnea. Denitrogenating an optimized FRC will provide the largest alveolar oxygen reservoir [8,10]. However, FRC is often reduced, either externally by compression from the chest wall, abdominal contents, pneumothorax/hemothorax or pleural effusion; or internally, by loss of functional airspace from pneumonia, edema, etc. [6]. Optimizing the FRC in at-risk patients involves upright positioning, draining relevant thoracic contents (effusions, hemothorax/pneumothorax), and recruiting lung units.

In patients where FRC is reduced because of body habitus, the dependent portions of the lung are compressed by the chest wall or abdomen, such as obesity. Upright positioning alone can improve FRC in these patients, with De Jong et al. [16] recommending reverse Trendelenburg positioning in combination with noninvasive ventilation for preoxygenation. Decreased FRC in pregnancy presents difficulties from the fixed compression of the gravid uterus on the chest cavity as well as on the inferior vena cava, specific positioning to overcome a decreased FRC includes ramped positioning with a left tilt [18]. Positioning, combined with preoxygenation with positive pressure ventilation can be quite advantageous in these patients [6,16]. For patients where external compression is due to body cavity fluids or air, such as ascites, hemothoraces, pleural effusions, or pneumothraces, drainage in parallel with preoxygenation can improve FRC and thus reduce ventilation/perfusion (V/Q) mismatch.

Given that positioning has an outsized role in improving FRC, a logical workflow would be to preoxygenate all patient in an upright position when able [6,11,29]. These patients can be repositioned after induction for the desired laryngoscopy position and to avoid increased intubation attempts with ramped positioning as stated by Cabrini et al. [12] and De Jong et al. [16]. There is potential for atelectasis by attempting nitrogen washout with 100% FiO₂, a potential threat to the FRC [5,29]. Delay et al. [15] demonstrated that positive end-expiratory pressure during preoxygenation resulted in higher EtO₂, a surrogate maker for denitrogenation, compared to spontaneous breathing and is likely a way to overcome the atelectasis of 100% FiO₂ administration and maintain the FRC.

The most modifiable way to improve V/Q mismatch is via increasing the FRC, though in some patients with severe disease may still have intrapulmonary shunt significant enough to limit the effectiveness of preoxygenation [29]. In these patients, EtO₂ may be appropriately high, yet that reservoir of oxygen does not sufficiently interface with the pulmonary blood flow to resaturate hemoglobin [29]. This is the final hurdle for some critically ill patients. Shunt fraction is on one extreme end of V/Q mismatch. However, unlike V/Q mismatch, shunt is refractory to both improving FRC and maximizing denitrogenation [5,29]. Shunt occurs when deoxygenated blood is not reoxygenated due to decreased alveolar-capillary gas exchange in the lung and will lead to further hypoxemia when the deoxygenated blood returns to the heart and the peripheral circulation [5].

When shunt is the rate limiting step for an adequate safe apnea time in a patient where RSI is the planned strategy, high flow nasal oxygen (HFNO) or noninvasive positive pressure ventilation (NIPPV) should be first-line for preoxygenation to maximize denitrogenation and reduce V/Q mismatch to the greatest possible degree through improved alveolar recruitment [5,6,29]. Fong et al. [25] and Mosier et al. [5] demonstrated that patients with acute hypoxemic respiratory failure receiving NIPPV had a decreased incidence of desaturation during intubation compared to conventional methods and HFNC, likely secondary to increased alveolar recruitment and gas exchange and thus potentially decreased shunt. In cases of shunt, PaO₂ may be helpful compared to EtO₂ and SpO₂ to determine the degree of shunt and if safe apnea time is possible or if alternative intubated strategies should be considered [6,29]. In the setting of a fully denitrogenated patient, if the PaO₂ is low, the shunt fraction is large and desaturation will occur quickly.

Patients with refractory shunt propose a unique challenge for preoxygenation management. Even with appropriate preoxygenation, they often do not tolerate periods of apnea with RSI [6]. In these patients, awake intubation may be preferred to maintain spontaneous breathing while still utilizing HFNO or nasal NIPPV during the procedure [5,6].

Preoxygenation remains a necessary step in all intubations and should include apneic oxygenation. Selecting the optimal preoxygenation method requires considering the underlying physiology and optimizing the FRC, reducing shunt and achieving maximal denitrogenation. Positioning is equally important, and critical in certain populations such as obesity and pregnancy. Utilizing the right preoxygenation strategy tailored to the patient’s anatomy, underlying physiology, and intubation plan will improve intubation safety in critically ill patients. In the most severe cases of refractory hypoxemia, preoxygenation may not be possible despite optimizing preoxygenation.