Discussion
The Pathophysiology of Hypoxemia
To understand oxygenation, it is essential to understand the causes of hypoxemia. These causes are inadequate alveolar oxygenation (low environmental oxygen pressure or alveolar hypoventilation), diffusion abnormalities, dead space (high ventilation, low perfusion [V/Q] mismatch), low V/Q mismatch, shunt, and low venous blood saturation. In the Emergency Department (ED) patient placed on ≥ 0.4 fraction-inspired oxygen (fiO2), all of these problems have inconsequential effects on oxygenation except shunt and low venous blood saturation. See Figure 1 for an explanation of these two phenomena.
Figure 1.
Ventilation/perfusion units. In the normal lung, oxygen enters the alveoli and raises the saturation from the venous level of 70% to 100% by the time it reaches the arterial side. In shunt, no oxygen can get in to the alveoli, so the venous saturation is never increased. In low SvO2 situations, the alveoli are not able to raise the low venous saturation to the normal arterial level. When these two problems are both present, the arterial desaturation becomes even worse.
An anatomical shunt is a direct connection between the arterial and venous blood flow, for example, a septal defect in the heart. When we speak about shunt as the cause of hypoxemia, we are rarely referring to anatomical shunts. Physiologic shunt is the major cause of poor oxygenation in ill ED patients already on supplemental oxygen. A physiologic shunt is caused by areas of alveoli that are blocked from conducting oxygen, but still have intact blood vessels surrounding them. This perfusion without any ventilation leads to a direct mixing of deoxygenated venous blood into the arterial blood. Causes of shunt include pneumonia, atelectasis, pulmonary edema, mucus plugging, and adult respiratory distress syndrome. No matter how high the fiO2, these areas will never have an improved oxygenation because inhaled gas never reaches the blood. The only way to improve oxygenation in these areas of the lungs is to fix the shunt.
Low venous oxygen saturation is also an important cause of hypoxemia in the ED. Venous blood is never fully desaturated when it reaches the lungs. In normal patients, the hemoglobin reaching the lungs has a saturation of ~ 65–70%, therefore, only a small amount of exposure to oxygen can rapidly bring the saturation to 100%. In shock states, the venous blood will arrive at the lungs with lower saturations due to greater tissue extraction. This venous blood will require more exposure to oxygen to reach a saturation of 100%; in injured lungs this may not occur. This problem becomes much more deleterious when combined with physiologic shunt. In this combination, the already abnormally low saturation venous blood mixes directly into the arterial supply.
This should impel the practitioner to always consider the circulatory system when evaluating the patient's respiratory status. If the patient about to be intubated is in shock, attempts to improve and prevent the reduction of cardiac output become methods to improve oxygenation. Tailoring sedative medications to the patient's cardiac status and blood volume is critical.[1,2] If time allows, these patients will also benefit from aggressive preintubation normalization of preload, afterload, and inotropy.[3,4]
Standard ED Preoxygenation
The standard recommended technique for ED preoxygenation is tidal volume breathing of oxygen from a high fiO2 source for at least 3 min or eight vital capacity breaths.[5] When possible, a maximal exhalation preceding the tidal volume breathing improves preoxygenation.[6,7] The non-rebreather mask (NRB), though the routine oxygen source, provides only 65–80% of fiO2.[8] In a healthy non-obese adult patient, these standard techniques have been shown to provide a buffer as long as 8 min before the saturation drops below the critical 90% threshold.[9] In the ill patient with injured lungs, abnormal body habitus, or upregulated metabolism, this time is significantly shortened.[9] In some cases it is impossible to obtain a saturation > 90% before the intubation attempt, regardless of the duration of standard preoxygenation.
A patient with a saturation < 95% on a nasal cannula set to 6 L/min of oxygen is exhibiting at least some degree of shunting, as this setting will provide ~ 0.4 fiO2.[8] If the saturation is < 95% on a NRB, the patient is exhibiting signs of moderate to severe shunting. These latter patients are at risk for a precipitous and dramatic decline in oxygen saturation during the intubation procedure.
We have seen many situations in which a patient preintubation is saturating < 90% even with a NRB; the providers become frustrated, abandon further attempts at preoxygenation, and proceed to the immediate intubation of the patient to improve the saturation. However, if the patient is saturating < 90% before rapid sequence intubation (RSI), they may have an immediate and profound desaturation almost immediately after the RSI drugs are administered. Figure 2 shows the oxygen-hemoglobin dissociation (saturation) curve. The patient in this circumstance is already on the steep portion of this curve and will shortly be at critically low pressures of oxygen.
Figure 2.
Oxyhemoglobin dissociation curve. The shape of the curve demonstrates that at 90% saturation, the patient is at risk of critically low oxygen levels (< 40 mm Hg PaO2) if even a brief period of time elapses without reoxygenation. Patients will take a much longer time to desaturate from 100% to 90% than to go from 90% to 70%.
This abandonment of preoxygenation and rush to premature intubation may be predicated on the fallacy that saturation declines in a linear fashion over time. The shape of the curve in Figure 2 demonstrates that the time to go from 100% to 90% is dramatically longer than the time it takes to go from 90% to injuriously low levels of oxygen pressure resulting in dysrhythmia, seizure, and cardiac arrest.
In this circumstance of low saturation before RSI, many airway experts recommend preoxygenation with a bag/valve/mask device (BVM). When the BVM is manufactured with an appropriate exhalation port and a tight mask seal is obtained, it can deliver > 0.9 fiO2 both when the patient spontaneously breathes and with assisted ventilations.[10] However, this increase from a fiO2 of ~ 0.7 (NRB) to ~ 0.9 (BVM) will do nothing to ameliorate shunt and little to correct low V/Q mismatched alveoli. In addition, it requires a practitioner to maintain an ideal mask seal during the stressful moments of preparing for RSI. If the mask seal is inadequate, room air will be entrained.
Preoxygenation in High-risk Patients
Non-invasive ventilation (NIV) has become a mainstay in the management of respiratory emergencies in most EDs. NIV is also the optimal technique for preoxygenation of high-risk patients. With a properly fitted, full-face NIV mask, fiO2 of ~ 1.0 is assured, and because these masks strap around the patient's head, no practitioner is needed to maintain the mask seal. With a setting of continuous positive airway pressure (CPAP) at 0 cm H2O, this NIV set-up will simply provide a source of nearly 100% oxygen. With increased CPAP settings, shunt can actually be treated and the patient's oxygenation significantly improved.[11-15]
Starting with a CPAP setting of 5 and titrating up to a maximum of 15 cm H2O, 100% saturation can be achieved in patients in whom NRB or BVM preoxygenation did not result in adequate saturations. This strategy requires the NIV machine or, preferably, a standard ventilator standing by in the ED. Unless the ED is consistently staffed with an in-department respiratory therapist, it is also necessary for the clinicians to know how to immediately set up and apply NIV themselves.
In EDs where neither a ventilator nor a NIV machine is available, the patient can be preoxygenated by spontaneously breathing through a BVM with a positive end-expiratory pressure (PEEP) valve attached. This is suboptimal, as a provider must hold the mask tightly over the patient's face and even a slight break in the mask seal eliminates the PEEP. PEEP valves will be discussed in more detail below.
Oxygenation During the Apneic Period
In standard RSI, the oxygen mask is left on the patient's face until the time of intubation. However, nothing is done to maintain a patent connection between the mouth and the glottis. As the sedative and paralytic drugs take effect, the tongue and the posterior pharyngeal tissues can occlude the passageway of oxygen to the glottis. Although this seems irrelevant as the patient is no longer breathing, it ignores the benefits of apneic oxygenation.
Apneic Oxygenation
In an experiment by Frumin et al., patients were preoxygenated, intubated, paralyzed, and placed on an anesthesia machine that provided 1.0 fiO2 and no ventilations.[16] These patients were maintained in this apneic state for between 18 and 55 min. None of these patients desaturated below 98%, despite being paralyzed and receiving no breaths. Although their CO2 levels rose, their oxygenation was maintained due to apneic oxygenation. Oxygen was absorbed from the patients' alveoli by pulmonary blood flow; this established a gradient for the continued pull of oxygen from the endotracheal tube and anesthesia circuit. In another study, Teller et al. showed that pharyngeal insufflation with oxygen significantly extended the time to desaturation during apnea.[17] Numerous studies on apneic oxygenation during brain death testing confirm that even without any respiratory effort, oxygen saturation can be maintained.[18-20]
If a continuous path of oxygen is maintained from the pharynx to the glottis during the apneic period of RSI, the patient will continue to oxygenate. This has led us to perform a jaw thrust in all high-risk patients during their apneic period. In some cases, we also place nasopharyngeal airways to augment the passage of oxygen. These techniques, combined with high-flow O2 from a NRB mask, NIV mask, or the facemask of a BVM, will allow continued apneic oxygenation.
Another problem during the apneic period is absorption atelectasis due to alveoli filled with near 100% oxygen. The nitrogen in normally ventilated alveoli serves to maintain their patency. When we preoxygenate with high fiO2, our goal is to completely wash out this nitrogen. This can lead to alveolar collapse as the oxygen is taken up by pulmonary blood; further shunt is the result.[21] The use of NIV ventilation with CPAP can maintain these alveoli in an open state during the apneic period. When NIV is combined with a jaw thrust and patent oro/nasopharyngeal passage of air, the potential benefits of apneic oxygenation can be fully realized.
