As a student, I never understood why some physicians were so obsessed with The Airway. Put a tube through the cords. How can such a simple subject merit tomes in a library and hours of lectures?
I have since garnered new appreciation. Although difficult airways are infrequent and airway management have low morbidity and mortality, there are few things in medicine more sphincter tightening. It only took one difficult airway before I became as obsessive, if not more, than the physicians I previously could not understand.
In a nutshell, airway management is like playing a video game. The goal is simple and there are many avenues to achieve it. Novices learn to succeed by listening to tips or following basic rules of a checklist. They have fun when things go well and quit the game forever if things go poorly.
An airway expert thinks about airway with every waking moment and then dreams about it when asleep. They deconstruct this "simple subject" down to the most minute details and study them in isolation to find the optimal approach. Every piece of advice they received as a novice is questioned and checklists are supplanted by critical thinking.
The theory is simple. Denitrogenate the lungs in order to prolong time to desaturation during apnea. In healthy subjects, the procedure is also simple.
Drugs - Sedatives
The purpose of sedatives is to provide amnesia, analgesia, and lack of awareness. Having volunteered to be a human guinea pig of awake intubations, I can certainly attest to the unpleasantries. But being dead is worse. And sedatives increase the risk of hemodynamic instability and thereby increase the risk of death. Thus, the provider and the patient together have to earn the right to use sedatives via aggressive resuscitation.
The biggest error committed with sedatives is not the medication themselves but the dose. As an intern, I proudly committed to memory a table drugs and their doses. I quickly learned that that table likely led to more harm than good for critically ill patients. Their pharmacodynamics are so different from healthy persons such that traditional dosing regimens are likely to lead to adverse hemodynamic consequences. Using bispectral index (BIS) to judge depth of sedation, pigs in hemorrhagic shock should receive 10-20% of the propofol dose compared to their healthy counterparts.
Unfortunately, this is too difficult to study in humans and no guidelines exist for appropriate dose of sedatives for the patient in shock. My practice has been to give quarter to half dose sedatives to patients in undifferentiated shock. I closely follow the clinical exam and am ready to additionally dose if there is purposeful movement or response. Anecdotally, I have not had any patients require repeat dosing; none have remembered the day, much less the act, of intubation.
The following chart was extrapolated from pig studies to illustrate the precentage reduction in dose of sedatives necessary to achieve BIS 40-60 in hemorrhagic shock.
- Egan TD, Kuramkote S, Gong G, Zhang J, Mcjames SW, Bailey PL. Fentanyl pharmacokinetics in hemorrhagic shock: a porcine model. Anesthesiology. 1999;91(1):156-66.
- Johnson KB, Egan TD, Kern SE, Mcjames SW, Cluff ML, Pace NL. Influence of hemorrhagic shock followed by crystalloid resuscitation on propofol: a pharmacokinetic and pharmacodynamic analysis. Anesthesiology. 2004;101(3):647-59.
- Johnson KB, Egan TD, Kern SE, et al. The influence of hemorrhagic shock on propofol: a pharmacokinetic and pharmacodynamic analysis. Anesthesiology. 2003;99(2):409-20.
- Johnson KB, Egan TD, Layman J, Kern SE, White JL, Mcjames SW. The influence of hemorrhagic shock on etomidate: a pharmacokinetic and pharmacodynamic analysis. Anesth Analg. 2003;96(5):1360-8.
- Johnson KB, Kern SE, Hamber EA, Mcjames SW, Kohnstamm KM, Egan TD. Influence of hemorrhagic shock on remifentanil: a pharmacokinetic and pharmacodynamic analysis. Anesthesiology. 2001;94(2):322-32.
To add to the evils of sedatives, the following ex vivo study details the effect of sedatives on human atrial contractility. Hashed boxes indicate total and free drug at clinical concentrations in humans.
Again, it is important to emphasize that pharmacodynamics are drastically different in shock patients and that every drug has the potential for negative inotrope. There have been several case reports of asystole after ketamine - traditionally considered a symphathomimetic. I have personally witnessed such a case along with another case of bradycardia/hypotension following ketamine induction.
I have concluded that any sedative can be used in critically ill patients as long as one is comfortable with the dosing and redosing interval. With fentanyl, I use 25-50mcg boluses q3min, titrating to effect; doubling the dose every other dose. With versed, I use 1-2mg in a similar strategy to fentanyl or in combination with fentanyl. With propofol, I use 25mcg q2min, titrating to effect; doubling the dose every other dose. The major detriment with this strategy is that its not truly rapid sequence; I am waiting to reach the appropriate sedation depth before pushing the paralytic. This delay can sometimes be 6min - agonizingly long.
Ketamine and etomidate are attractive options because they can be quickly titrated q2min, and - I find - rarely need titration. For the unstable patient, I start with 0.15mg/kg etomidate or 0.5mg/kg ketamine. I think there is a lot of sense in mixing sedatives as you get the combined sedative effects with different dose-dependent side effect profiles.
Drugs - Paralytic
Whereas the sedative drugs provide comfort to the patient, paralytic drugs benefit the provider. They improve the provider's view, first pass success, and time to intubation. The caveat is that, even more-so than the induction drugs, paralytics is a point of no return. Ventilation and oxygenation is entirely in your hands.
Out of the nondepolarizing paralytics, I think that the pharmacodynamics of rocuronium make it the most favorable. At doses of 1.2mg/kg, it has a 1min onset. Vecuronium, by comparison is 3-5min. The debate is therefore mostly rocuronium vs succinylcholine.
Succinylcholine produces paralysis in 45s-1min with full recovery in 10-15min in patients with normal plasma pseudocholinesterases. Succinylcholine fans remain advocates primarily citing that it will wear off in 5-15min if one encounters a difficult airway. However in my world, letting the patient wake up and resume spontaneous breathing is rarely an option because my intubations are rarely elective.
Potassium will increase by about 0.5-1mEq/L on average but it can also increase much more substantially in patients with proliferated extrajunctional receptors (i.e. 48h after burns, trauma, or denervation; boys with muscular dystrophy).
Succinylcholine also activates muscarinic receptors of the heart, leading to bradycardia, junctional rhythms, and even sinus arrest. In children, this an occur after one dose. In adults, it is more common with a second dose 5 minutes after the first. This effect can be very unpredictable as succinylcholine can activate muscarinic receptors at other sites, paradoxically increasing SBP and HR.
There are a litany of other disease specific concerns: malignant hyperthermia, increased intraocular pressure, increased intracranial pressure, increased gastric pressure, rhabdo, worsening fractures, myasthenia gravis, Eaton-Lambert, plasma pseudocholinesterase disorders.
Even with the best patient selection, I remain more partial to rocuronium for critically ill patients. Depolarizing paralytics cause fasciculations and increase myocyte oxygen consumption. This can theoretically decrease safe apnea time especially in hypoxemic respiratory failure. This is very difficult to study but seems to pan out in healthy patients. 60 healthy patients were randomized to rocuronium vs succinylcholine during intubation for elective surgery. After intubation, the ETT was left open to air until SpO2 declined to 92%. The succinylcholine group on average reached 92% 46seconds before the rocuronium group.
Another study randomized 60 patients to induction with 1) lidocaine, fentanyl, propofol, roc 2) lidocaine, fentanyl, propofol, sux 3) propofol, sux. The patient was intubated and time was recorded until desaturation to 95%. Propofol and sux caused significantly more muscle fasciculations than the other groups and desaturated to 95% 136s faster than lidocaine, fentanyl, propfol, roc.
It is unclear how this would translate to critically ill patients with poor oxygen reserve and hypoxemia prior to intubation. I only know that I would want all of my stars aligned for such a patient where failure to establish an airway is obviously not an option.
The only concern for rocuronium is the package insert: 5.11 Rocuronium bromide may be associated with increased pulmonary vascular resistance, so caution is appropriate in patients with pulmonary hypertension or valvular heart disease. I could not find any literature to substantiate these claims.
My PharmD friend managed to find this article (thx RB!) which studied rocuronium in isolation by administering it to cardiac surgery patients already intubated with succinylcholine and after succinylcholine had worn off. Rocuronium decreased the mean pulmonary artery pressures on average but in 3/20 cardiac surgery patients caused a 20% rise in the diastolic PAP. Overall though, rocuronium proved to be very hemodynamically neutral and the anesthesiology literature supports the use of rocuronium in pulmonary hypertension.
However, if rocuronium is to be used, it is important to dose the medication appropriately. Novices will use label's recommended dosing of 0.6mg/kg. But at this dose, the onset and potency of action is reduced. 60 ASA I and II patients were randomized to five different dosing regimens of rocuronium. The outcome variable was intubating conditions as defined by vocal cord position, vocal cord movement, ease of laryngoscopy, airway reaction, and movement of limbs. The results heavily favored higher dosing with no negative effects on hemodynamics. Moreover, patients in shock have reduced perfusion of myocytes and may need the higher doses to achieve the same effect.
Drugs - Sequence
The default sequence to intubation seems to be rapid-sequence intubation. But survey the medical field or literature and it's hard to find a consensus on its definition. In 1951, Morton and Wylie first described administering sedatives immediately followed by paralytics with the patient sitting upright to minimize the risk of aspiration. In 1961, Sellick introduced cricoid pressure in addition to concomitant administration of sedatives and paralytics. However, these physicians also utilized positive pressure ventilation prior to endotracheal intubation. In 1963, Wylie recommended avoiding positive pressure ventilation until endotracheal intubation to minimize the risk of gastric insufflation. Today, most physicians will vaguely agree that it involves 1) administering sedatives and paralytics in close temporal proximity and 2) not utilizing positive pressure ventilation until endotracheal intubation.
RSI was designed to minimize the risk of aspiration, which is cited as the most common complication to airway management in anesthesia. The event rate is estimated to be 1/600-800 emergency surgeries. Fatality is estimated to be 1/340 000 in the UK. I have two contentions with RSI. Firstly, it has never been shown to reduce rates of aspiration. Secondly, the rate of aspiration and its associated mortality rates are minuscule in the face the overall mortality rate of patients requiring intubation in the emergency department and the ICU. For many cases outside of the anesthesia world, RSI is not the appropriate sequence.
My choice of sequence and approach depends on anticipated anatomic difficulties and anticipated physiologic difficulties,
Anatomically difficult airway. Need video options and strongly consider awake fiberoptic intubation. Prepare for surgical airway.
Physiologically difficult airway. The risks of intubating a critically ill patient extends far beyond aspiration. Generally, intubation can kill the patient through hypoxia, hypotension, acidosis, aspiration, and intracranial hypertension.
The mortality risk for a relatively healthy patient requiring intubation for airway protection during surgery can be hypothetically ranked as aspiration > hypoxia > hypotension > acidosis > ICH. A sequence that emphasizes reduced aspiration - like RSI - is ideal.
However, a patient requiring intubation for pulmonary edema with a starting SpO2 of 70% on HHFNC is much different: hypoxia > hypotension > aspiration > acidosis > ICH. RSI on this patient heralds worsening hypoxia and impending hypoxic cardiac arrest. Noninvasive positive pressure ventilation may be necessary in order to oxygenate, and sedation without paralytics may be necessary so that the patient would tolerate NIPPV. This is a sequence known as delayed sequence intubation (DSI) with violates both tenants of RSI and potentially increases the risk of aspiration. Nevertheless, it is appropriate when mortality from hypoxia greatly outweighs mortality from aspiration.
Consider a patient with hematemesis and hemorragic shock - HR 140s and BP 70/30. This patient has yet another mortality profile: hypotension > hypoxia > aspiration > acidosis > ICH. Positive pressure ventilation decreases venous return and sedatives reduce sympathetic drive. Both exacerbate hypotension in this already hypotensive patient and herald PEA arrest. Furthermore, active hematemesis and aspiration makes visualizing airway structures incredibly difficult and will likely contribute to prolonged intubation attempt and worsening hypoxia. On the other hand, blood is a very neutral aspirate. The rate of pneumonitis and ARDS from aspirating blood or water is low. My choice in this situation is resuscitation sequence intubation: focus first on volume resuscitation, and an NGT to decompress the stomach. Again the risk of aspiration is higher but at the benefit of reducing the risk of hypotension and hypoxia.
The patient who overdosed on TCA or aspirin: acidosis > hypotension > aspiration > hypoxia > ICH. Generally, patients can tolerate a remarkable degree of acidosis. However in toxicology, acidosis increases free drug of TCA causing hypotension, seizures, and cardiac arrest; acidosis also facilitate salicylate penetration into the blood brain barrier causing irreversible brain damage. Although RSI is "rapid", there is still apnea time so acidosis is unavoidable. One may consider rapid sequence induction followed by hyperventilation with positive pressure ventilation to buffer the impending acidosis. Or one may consider rapid sequence airway where an LMA is inserted immediately after rapid sequence induction to provide uninterrupted ventilation before the LMA is exchanged for an ETT. Even more exciting is fiberoptic intubation either awake or during positive pressure ventilation through a mask and three way swivel adaptor.
The patient with intracranial hemorrhage: ICH > hypotension > aspiration > hypoxia > acidosis. Here, RSI is appropriate after pretreatment with beaucoup fentanyl (3-5mcg/kg) ± lidocaine and push dose pressors.
The most appropriate sequence depends on weighing the patient-specific risk factors and selecting that which addresses the patient's greatest risk. Sometimes this comes at the cost of increasing other risks like aspiration. But the risk of aspiration is, in my opinion, low and its attributable mortality lower still.
Can oxygenation be maintained without ventilation? Eight healthy patients were intubated and connected to an anesthesia bag with 100% FiO2. The anesthesiologist did not squeeze the bag (i.e. no ventilation) and monitored the SpO2 and ABGs. No one desaturated after 18min and one patient lasted 53min. The experiment was aborted for respiratory acidosis though there was no clinical change.
But of course, those patients were already intubated. What about preoxygenation into the nasopharynx? Twelve healthy patients were involved in a cross over trial. They underwent preoxygenation and then induction/paralysis. A 36Fr nasal airway was inserted into one nare followed by a 8Fr catheter inserted 2cm into the nasal airway (sounds like a nasal cannula inside an NPA) and either 3L O2 or 3L Air was administered into the catheter. The anesthesiologist was blinded. This was maintained until the patient desaturated or until 10min elapsed. Afterwards, the patient was bagged up with 100% FiO2 and was crossed-over into the other arm. No one receiving 3Lpm O2 desaturated whereas most desaturated before the 10min mark in the air group.
But what about critically ill patients? Many of whom are getting intubated because of hypoxemia despite aggressive support. The FELLOWS trial randomized 150 ICU patients intubated for hypoxemic respiratory failure to receiving 15LPM via HFNC during laryngoscopy vs no nasal cannula during laryngoscopy. They found no difference and concluded that apneic oxygenation via NODESAT (nasal oxygenation during efforts securing a tube) is not helpful.
However, is 15Lpm really "high flow"? It's certainly the top number on the wall oxygen but the wall oxygen is rated to deliver 45-70Lpm of flow. You can hear the difference when you dial the oxygen past the 15Lpm mark. Did the FELLOWS trial simply not give enough? Especially with the use of HFNC, you can easily administer 60LPM at 100% FiO2.
The OPTINIV trial randomized 50 ICU patients undergoing intubation for hypoxemic respiratory failure to HFNC at 60LPM vs nothing during laryngoscopy AFTER preoxygenation with noninvasive ventilation (interestingly with a ventilator delivering breaths through a face mask). There was less incidence of hypoxemia with the HFNC group though no differences in clinical outcomes (ICU LOS, ventilator days, hospital LOS).
One might make the argument that if there is no difference in clinical outcomes, does all of this matter? To this I answer: in a bad airway, you will do anything and pay anything to have even 5 more seconds. I probably wouldn't use HFNC on every patient I intubate but in light of this data, I would certainly use NC fully cranked up during apneic oxygenation.
Curving the stylet and the endotacheal tube to the shape of the laryngoscope and the tongue seems to make sense intuitively. But this actually reduces the operator’s view of the cords. A more ergonomic option is to keep the stylet straight until the cuff with a 30deg bend at the cuff.
Bougie and Aintree