Maintenance of Oxygenation During Rapid Sequence Intubation in the Emergency Department

Rapid sequence intubation (RSI) is the most common method of airway control in the emergency department (ED).1, 2 Administration of an anesthetic agent and a neuromuscular blocking agent (NMBA) optimizes conditions for tracheal intubation and is thought to minimize the risk of aspiration.3-10 Evidence suggests that RSI improves first-pass success and reduces complications in the critically ill.11-14 However, RSI in the ED is not without its drawbacks.15, 16 The use of an NMBA results in the cessation of spontaneous ventilation and consequently an interruption of alveolar oxygen delivery. A crucial step in maximizing the safety of RSI is preoxygenation, which creates a large intrapulmonary oxygen reservoir that can be utilized during the period of apnea until mechanical ventilation can be initiated.17-25 Even with preoxygenation, a significant number of patients undergoing RSI in the ED experience desaturation.26, 27 Most of the time, this is transient and is easily reversed with manual positive pressure ventilation. But occasionally desaturation becomes critical and may result in life-threatening complications such as dysrhythmias, cardiovascular collapse, hypoxic brain injury, or cardiac arrest.28-30 Recently, apneic oxygenation has been promoted as an adjunct to RSI to help reduce the risk of desaturation.23 While the principle and technique of apneic oxygenation have been known for over half a century, the past few years have seen a great deal of research, interest, and discussion regarding its role during RSI in the critically ill.23, 31-50 During apnea, the ongoing uptake of oxygen from the alveoli into the bloodstream creates a negative pressure gradient between the upper airway and the lungs. By providing a continual supply of oxygen to the upper airway, a reservoir of oxygen is maintained, and this oxygen can move into the alveoli in the absence of any airflow.31, 32 Studies performed on elective surgical patients in the operating room have demonstrated that apneic oxygenation can significantly prolong the safe apnea time.51-55 In this issue of Academic Emergency Medicine, Caputo and colleagues56 present the results of a randomized controlled trial of apneic oxygenation during RSI in the ED. In this study, 206 patients were randomized to undergo standard RSI or standard RSI with the addition of apneic oxygenation. The intervention group had oxygen supplied at ≥15 L/min via a standard nasal cannula (intranasally) and also via an end-tidal CO2 nasal cannula (infranasally) and thus received a minimum of 30 L/min during intubation. The authors sought to determine if the use of apneic oxygenation would result in a decrease in the prevalence of hypoxemia during the peri-intubation period. They found no difference in the incidence of moderate hypoxemia (SpO2 < 90%) or severe hypoxemia (SpO2 < 80%). There also was no difference in the mean lowest saturation (control group 93% vs. apneic oxygenation group 92%). At first glance, these results may seem surprising, but close analysis of the data reveals the reason why no benefit was seen: most of the patients in this trial were intubated very rapidly. The mean apnea time was 58 seconds in the control group and 64 seconds in the apneic oxygenation group. Ninety percent of the patients were intubated in less than 100 seconds; all were intubated by 3.3 minutes. We know from physiologic modeling experiments and human studies that with complete preoxygenation several minutes of safe apnea time can be expected in adults.57-59 Since most of the patients in this study were intubated within a couple of minutes, the preoxygenation alone likely would have provided an adequate oxygen reservoir to prevent desaturation. Apneic oxygenation would only be beneficial in those patients who have exhausted their oxygen reserves, for example, patients who had prolonged laryngoscopies. This study illustrates the problem of studying apneic oxygenation in the critically ill—the necessity of rapid intubation and restoration of alveolar oxygen delivery precludes evaluation of the outcome measure of interest—an extended safe apnea time. Interestingly, the results of this study are similar to those found by another randomized controlled trial that was performed in the intensive care unit (ICU).34 In this study of 150 patients, Semler and colleagues34 evaluated the impact of apneic oxygenation at 15 L/min via nasal cannula on patients undergoing RSI in the ICU. They found that the use of apneic oxygenation had no effect on the mean lowest saturation during the peri-intubation period (90% in the control group vs. 92% in the apneic oxygenation group). It is notable however, that many of the patients in this trial had hypoxemic respiratory failure requiring noninvasive ventilation (NIV), suggesting that they had significantly reduced functional residual capacities (FRC). Previous work has demonstrated that patients with reduced FRC to body weight ratios have limited tolerance to apneic oxygenation.60 Additionally, one-third of the patients in the apneic oxygenation group were preoxygenated with a nonrebreather (NRB) mask at 15 L/min, a technique that has been shown to result in poor preoxygenation.61-63 Inadequate preoxygenation likely would have reduced the benefit of apneic oxygenation.60 These two randomized control trials contrast with some recent observational studies that have demonstrated a benefit with the use of apneic oxygenation during emergency intubation.35-37 Wimalasena and colleagues35 incorporated apneic oxygenation into their aeromedical RSI protocol and found in a before-and-after study that the prevalence of desaturation to <93% decreased from 23% to 17%. In the author's ED, the introduction of apneic oxygenation was found to be associated with an increase in first-pass success without hypoxemia (82% vs. 69%).36 In a subset of patients with intracranial hemorrhages, who were at great risk of harm from hypoxemia, apneic oxygenation was associated with a reduced prevalence of desaturation to <90% (7% vs. 29%), to <80% (4% vs. 18%), and to <70% (3% vs. 9%).37 It is important to note, however, that in these studies nasal oxygen administration was initiated during the preoxygenation phase, and thus part of the improvement seen may in fact be due to improved preoxygenation from a higher fraction of inspired oxygen (FiO2) that the patients in the apneic oxygenation groups received.59 Recent studies have demonstrated that the addition of nasal cannula to NRB mask preoxygenation results in an improved fraction of expired oxygen (FeO2), an indicator of the completeness of denitrogenation.61, 62 Several meta-analyses of apneic oxygenation studies have been published in the past year and all have demonstrated a clinical benefit with the use of apneic oxygenation during emergent RSI.42-46 Holyoak and colleagues45 performed a meta-analysis of 17 studies involving 2,422 patients in the ED, ICU, or prehospital setting and found that there was a significant reduction in both the prevalence of mild desaturation to <90% (relative risk [RR] = 0.65) and critical desaturation to <80% (RR = 0.61). When broken down by indication for intubation, the effect became nonsignificant for patients intubated for respiratory failure, but remained significant for patients intubated for other indications. Additionally, seven studies in this meta-analysis showed a statistically significant improvement in first-pass success (RR = 1.06). Again, on subgroup analysis this effect became nonsignificant in patients intubated for respiratory failure, but remained significant for those intubated for other indications. Pavlov and colleagues43 performed a meta-analysis of eight studies involving 1,953 patients in the ED, ICU, and prehospital setting and found that apneic oxygenation reduced the RR of clinically significant hypoxemia by 30%. Oliveira J E Silva and colleagues46 performed a review of 14 studies on apneic oxygenation in the ED or ICU involving 2,023 patients. A meta-analysis of eight of these found that apneic oxygenation was associated with decreased hypoxemia (SpO2 < 93%), but not decreased severe hypoxemia (SpO2 < 80%) or life-threatening hypoxemia (SpO2 < 70%). Several points regarding these meta-analyses are worth noting. One, is that these reviews used many of the same studies in their analyses, so it is not surprising that they all found similar benefits with apneic oxygenation. Second, the quality of studies included in the analyses was relatively low; thus there is a low to moderate level of certainly in the point estimates presented. Third, the meta-analyses included both low-flow (≤15 L/min) and high-flow (≥50 L/min) apneic oxygenation studies. While the literature on the efficacy of high-flow apneic oxygenation is quite robust, that for low-flow apneic oxygenation is less so. Inclusion of the high-flow studies might have inflated the overall benefit of apneic oxygenation, particularly in regard to low-flow apneic oxygenation. Physiologic modeling data has demonstrated that the duration of safe apnea time is highly dependent on the ambient fraction of oxygen that the airway is exposed to, suggesting that high-flow and low-flow apneic oxygenation may be very different in their efficacy.64 Finally, as pointed out by Wong and colleagues47 in their narrative review of apneic oxygenation, adverse events were not reported, and the studies were not adequately powered to detect them. What are we to conclude from all of these interesting yet conflicting studies? Should we alter our current practice and routinely use apneic oxygenation for RSI in the ED? Certainly, apneic oxygenation is a simple and low-cost intervention to institute in most EDs. The brief period of oxygen administration is unlikely to result in any harm to most patients; however, care should be taken in patients who can be adversely affected by the concomitant hypercarbia that develops with apnea, such as those with increased intracranial pressure, metabolic acidosis, or pulmonary hypertension.41, 65-67 Also, it is important to note that rare cases of gastric rupture, pneumothorax, and pneumomediastinum have been reported with nasopharyngeal oxygen administration, even at very low oxygen flow rates.68-72 Apneic oxygenation can potentially be of benefit to patients undergoing emergent RSI, particularly those with difficult airways who have prolonged intubation attempts. Since many difficult airways are not able to be identified prospectively, it would not be unreasonable to provide apneic oxygenation to all patients undergoing RSI in the ED.73 Certainly it should be used on any patient that is anticipated to have a potentially long laryngoscopy or is likely to have rapid desaturation due to reduced oxygen reserves. Use of a standard nasal cannula at 15 L/min is a reasonable approach for brief periods of apneic oxygenation. End-tidal CO2 nasal cannulae should not be used, as these devices do not deliver oxygen through the nasal prongs, but instead through a series of small holes in the tubing just below the nose. Moreover, they are only rated to deliver oxygen at ≤5 L/min. In patients who are deemed very high risk for desaturation, apneic oxygenation using a heated and humidified high-flow nasal cannula (HFNC) system, such as Optiflow™ or Vapotherm™, is recommended, particularly if increasing PaCO2 is also a concern.74-79 In the operating room, the Optiflow HFNC system at an oxygen flow of 70 L/min (THRIVE) has been demonstrated to produce prolonged safe apnea times, with the added benefit of minimizing increases in PaCO2 by also providing a modicum of apneic ventilation.74 Some data on the use of THRIVE in the critically ill requiring emergent intubation suggests that it can reduce desaturation events.76, 77 It should be emphasized that when apneic oxygenation is used as an adjunct during RSI, a patent airway must be maintained at all times to allow movement of oxygen from the upper airway into the lungs. Additionally, it should be recognized that the benefit of apneic oxygenation will only be realized if it is preceded by good preoxygenation.60 While apneic oxygenation recently has come into vogue for emergent intubation, perhaps the issue that warrants greater consideration is the provision of proper preoxygenation in the critically ill.22-25, 42, 120-122 Increased attention to this crucial step of RSI is likely to reduce the likelihood of desaturation in the critically ill and thus have the greatest impact on patient safety.80, 82, 83 Much of the literature concerning preoxygenation is based on elective intubations in the operating room using an anesthetic circuit and mask with a good seal. This is not practical for emergent intubations and thus alternative methods must be used. In some countries, flow-dependent Mapleson breathing systems are available in the ED and have been shown to be very effective for preoxygenation.85-89 In the United States, Mapleson circuits are not typically available, and the only manual resuscitators available for preoxygenation are flow-independent self-inflating bags, also known as a bag-valve-masks (BVMs). Unfortunately, BVMs from different manufacturers vary considerably in their design and thus deliver a widely variable FiO2 in spontaneously breathing patients, ranging from 40% to 95%.90-98 Specifically, BVMs without a one-way valve on the expiratory port allow entrainment of large amounts of room air during spontaneous ventilation and thus deliver a greatly reduced FiO2.90 Also, BVMs that use corrugated tubing for the oxygen reservoir, as opposed to an inflatable plastic bag, have been found deliver a lower FiO2.95 Due to the lack of consistency in the ability of BVMs to deliver a high FiO2, many emergency physicians instead use a NRB mask with a bag reservoir for preoxygenation. A common practice in the ED has been to preoxygenate using a NRB mask at 15 L/min. This flow is unlikely to meet the peak inspiratory flow of many patients in the ED and recent studies have in fact demonstrated the inadequacy of this method of preoxygenation.61, 62, 99 Groombridge and colleagues62 have demonstrated that in healthy adults a NRB mask at 15 L/min only results in a mean FeO2 of 52%, significantly less than the 80% that can be achieved with an anesthetic circuit or a BVM with a functional expiratory valve. This is because many NRB masks are rigid and nonconforming to the face and thus permit the entrainment of room air, which significantly reduces the FiO2. Interestingly, a recent study by Driver and colleagues63 demonstrated that the leak from a NRB mask can be compensated for by increasing the oxygen flow to the “flush rate” (flowmeter fully open). In this study, using a NRB mask with an oxygen flow of 50 L/min resulted in a mean FeO2 of 86% in healthy adults, which was significantly higher than the mean FeO2 achieved with a NRB mask at 15 L/min (54%) and which was comparable to the mean FeO2 that was achieved with a BVM at 15 L/min (77%). The real question is, what is the best approach to preoxygenation in the ED? In patients with inadequate spontaneous ventilation, the decision is straightforward—they require manually assisted ventilations with a BVM. Fortunately, the reduced FiO2 seen with some BVMs during spontaneous breathing is not an issue when manual positive pressure ventilation is provided by squeezing the bag. The decision on the appropriate preoxygenation method in patients with adequate spontaneous ventilation is more complicated. In cooperative patients, the use of a BVM is the best option for preoxygenation, if it has a functional expiratory valve, and if a tight face mask seal can be achieved. Great care must be taken to maintain a good mask seal, as even a small leak can result in poor preoxygenation due to entrainment of room air.100-104 Unfortunately, small leaks can be difficult to appreciate, especially without monitoring end-tidal O2 or end-tidal CO2.104 In patients who are uncooperative, or in whom an adequate mask seal cannot be maintained, a NRB mask at flush rate is an acceptable alternative, if the actual flush rate is known and is ≥50 L/min. It should be recognized that the flow achieved from wall oxygen at flush rate can vary considerably from institution to institution. Recent reports from Australia found that flush rate preoxygenation with a NRB mask only resulted in an oxygen flow of 19 L/min and only marginally improved FeO2.105, 106 Ideally, wall oxygen flow meters that are calibrated to 15 L/min should be replaced with ones that are calibrated up to 70 L/min so the exact flow can be controlled during preoxygenation. Use of a soft, pliable NRB mask that conforms well to the face and has two functional expiratory valves is recommended for preoxygenation to minimize entrainment of room air and maximize the FiO2. Whenever preoxygenation is performed it should be done with patients in the head-up position, as this increases the FRC and thus the volume of oxygen stores that can be achieved with preoxygenation.107, 108 Multiple studies on patients in the operating room have demonstrated that the head-up position results in improved preoxygenation and increases the safe apnea time.109-112 Interestingly, a recent randomized controlled trial in critically ill patients in the ICU found that preoxygenation in the head-up position resulted in no improvement in oxygenation compared to the supine position.113 The majority of the patients in this trial, however, had respiratory failure, many with shunt physiology, and this many have limited the benefits of upright preoxygenation. Based upon all data currently available, it would seem prudent to preoxygenate critically ill patients in the ED in the head-up position when feasible. While 3 minutes of tidal volume breathing is recommended for preoxygenation, it should be appreciated that this many not be sufficient in some patients.19-21, 24, 100, 104, 114, 115 While the above preoxygenation techniques will be satisfactory in most patients requiring intubation in the ED, patients with a significant right-to-left intrapulmonary shunt from air space disease will not be able to be adequately preoxygenated without positive end expiratory pressure (PEEP).24, 25, 48 These patients require alveolar recruitment and thus should be preoxygenated with a BVM using a PEEP valve or with NIV. Studies in the ICU have demonstrated a clear benefit of NIV preoxygenation in patients with hypoxemic respiratory failure, with better preoxygenation achieved and a reduced prevalence of desaturation.79, 116 A recent study also has demonstrated that the combination of both NIV and HFNC was effective in reducing the severity of desaturation in profoundly hypoxemic patients.117 If a patient cannot tolerate NIV preoxygenation, the use of HFNC at ≥50 L/min is an acceptable alternative, as data suggest that excellent preoxygenation can be achieved with this technique in patients with mild to moderate hypoxemia.75, 76, 78 In a hypoxemic patient who cannot tolerate preoxygenation efforts, consideration can be given to using a delayed sequence intubation (DSI) technique.118 This involves the administration of a dissociative dose of ketamine to allow the patient to tolerate preoxygenation. It should be borne in mind that some patients can be susceptible to apnea even from small doses of ketamine.119 Thus, once a DSI is initiated and ketamine is administered, one must be prepared to immediately airway It is important to that many patients with difficult airways have a greatly reduced tolerance for apnea, and thus consideration should be given to a RSI with the provision of manual ventilation 120-122 are available for BVMs and their use is recommended to manual ventilation which can increase the risk of gastric and RSI and an intubation is the most prudent of in the It is worth that a system can also be used during intubations to prevent While to the of good preoxygenation in the ED is the is that the of intrapulmonary oxygen stores can be without an indicator of alveolar oxygen In the operating room, this is easily as anesthetic have in that end-tidal In the ED, such is not However, the recent introduction of a the of FeO2 in the ED. this unit a as well as a of both FiO2 and This regarding the of ventilation, oxygen and FeO2 has some and must be with it can be an in the of the 25, We have used it in our ED for the past year and have found it very in to detect of inadequate preoxygenation. It has to preoxygenation by to their such as the face mask seal, increasing the oxygen flow or the duration of preoxygenation. Use of such as this may help to the safety of intubation in the ED by improved preoxygenation and thus potentially reducing desaturation events and the associated with them. It is important to note that of FeO2 a system and thus can only be done when with a technique that a FeO2 can be when using a BVM with a good face mask seal or NIV if the unit has a expiratory FeO2 is not with preoxygenation methods such as NRB masks or HFNC tracheal intubation is one of the most yet an emergency can during RSI in the ED is common when can result in life-threatening The risk of desaturation can be reduced with attention to preoxygenation, using the most appropriate method for the given clinical oxygen are available and may be in maximizing preoxygenation emergent intubation. Apneic oxygenation is a simple intervention that may help reduce the risk of desaturation, but it will not poor preoxygenation. on the of oxygenation during RSI in the ED is to the safety of this