Why do some patients' saturation crash during laryngoscopy, whereas other patients are fine? What can we do to prevent this?
Traditional model of desaturation during intubation
The above figure illustrates a common physiological model for desaturation during intubation. According to this model:
- The lungs function as a tank containing oxygen.
- Preoxygenation with 100% FiO2 fills the lung-tank with oxygen.
- After paralysis, the oxygen tank is gradually used up over several minutes. When the oxygen runs out, the patient desaturates.
- Apneic oxygenation can be used to keep the tank full during laryngoscopy, extending the safe apneic period almost indefinitely.
This model was created and validated in the elective anesthesia world. It may work well in that context. The issue is how useful this model is among patients who are being intubated because of respiratory failure – patients with deranged physiology.
Most experienced critical care practitioners realize that this model doesn’t quite work. For example, according to the above graph, severe desaturation should never occur before 3 minutes. However, it’s common for critically ill patients to severely desaturate within 1-2 minutes of anesthesia induction.
Why the model fails: Physiologic shunting
Why do some patients desaturate so quickly? The oxygen consumption of an average adult is ~250 ml/min. Total lung capacity of an adult is usually >4,000 ml. Therefore, if the patient is fully pre-oxygenated, then there is plenty of oxygen in the lungs to prevent desaturation for a long time (>>10 minutes).
The answer is physiologic shunting. If about a quarter of the lung tissue collapses, this causes a shunt whereby deoxygenated blood effectively bypasses the lungs and goes directly into the systemic circulation. No matter how much oxygen might be in the remainder of the lungs, a 25% shunt will cause severe desaturation.
Many critically ill patients may start out with a physiologic shunt to begin with (e.g. parts of the lung are atelectatic or socked full with pneumonia). Upon induction of anesthesia, diseased lungs (which are often filled with secretions and deficient of surfactant) may collapse rapidly, increasing the shunt. Obesity may also contribute by causing immediate diaphragmatic elevation which collapses the lung bases.
Combatting atelectasis: Vent-As-Bag strategy
The antidote to atelectasis is positive pressure. For patients with severe hypoxemia, the most powerful oxygenation strategy is to recruit the lungs with BiPAP and then to continue positive pressure ventilation right until the moment of laryngoscopy. The ideal way to do this involves using a machine capable of delivering either BiPAP or mechanically delivered breaths (e.g. a mechanical ventilator or a sophisticated BiPAP machine). The patient remains connected to the machine throughout the induction period. The machine is set to deliver machine-initiated breaths immediately after the patient’s own respiratory rate drops, in a seamless transition. Using a machine to “bag” the patient during induction allows for the delivery of perfectly controlled breaths with a pressure cap that won’t insufflate the esophagus (e.g. 20 cm or less). This has been described in detail in a previous blog here.
Vent-as-bag is a great technique. From a physiologic standpoint, this is nearly an ideal approach to oxygenation during intubation (it could theoretically be improved by bleeding in some nitric oxide to improve ventilation-perfusion matching). The problems are logistic. For patients who aren’t already on BiPAP, this requires obtaining a BiPAP mask and attaching it to the ventilator. Using a BiPAP mask for every patient can consume a lot of masks. However, compared to the risks involved with intubation of critically ill patients (e.g. ~2% mortality), this could be a reasonable price to pay for improved safety.
Rapid sequence intubation (RSI) versus bagging before laryngoscopy
One perpetual debate is whether patients should be bag-mask ventilated after induction medications are given and before laryngoscopy. There are roughly two schools of thought on this:
- Rapid sequence intubation (RSI) – In its purest form, this involves pushing medication, waiting for paralysis without any ventilation, and then proceeding with laryngoscopy. Avoidance of any mask ventilation should theoretically reduce the risk of gastric insufflation and aspiration.
- Bag-Mask Ventilation – After pushing medication, the patient is gently mask ventilated while the drugs take effect. This may help prevent atelectasis and shunt physiology, thereby improving oxygenation.
This debate has raged on for decades in the absence of any high-quality evidence. A new multi-center RCT may finally offer some answers on the topic.
Casey et al: Bag-mask ventilation during tracheal intubation of critically ill adults.
This is a prospective, multi-center trial comparing pure rapid-sequence intubation (without any ventilation) to a strategy that involves bag-mask ventilation during induction.1 Bag-mask ventilation was performed using the following parameters:
- A PEEP valve was connected and set to 5-10 cm.
- A two-handed mask seal was used by the intubating clinician (with a head-tilt and chin-lift maneuver).
- Ventilation was performed at 10 breaths per minute.
- The tidal volume was limited to the lowest volume which caused visible chest rise.
- Oropharyngeal airway was used as needed.
Patients were excluded if the treating clinician felt that there was a high risk of desaturation or aspiration. The study involved 401 patients being intubated in intensive care units, of whom 80% had respiratory failure:
Details of the intubation procedure are shown below. First-pass success was obtained in 82% of patients, with the median time from induction to intubation being very short (~2.5 minutes). 78% of patients in the no-ventilation group received apneic oxygenation.
The efficacy results were striking. The nadir oxygen saturation was lower in the no-ventilation group, with 10% of patients falling to a saturation below 70% (figure below). Differences were statistically significant using a variety of different metrics (in the p<0.01 range).
The safety outcomes are perhaps even more striking (table below). Most importantly, there was no signal for increased aspiration in patients receiving bag-mask ventilation (in fact, there were fewer witnessed aspiration events in the bag-mask ventilation group).
- More patients in the bag-mask ventilation group received bag-mask ventilation as a preoxygenation strategy (whereas patients in the no-ventilation group were more likely to be preoxygenated using high-flow nasal cannula or BiPAP). Therefore, superior results in the bag-mask ventilation group could theoretically reflect superiority of bag-mask ventilation as a preoxygenation technique. However, high-flow nasal cannula and BiPAP are outstanding strategies for preoxygenation, so it’s unlikely that using these techniques would cause inferior preoxygenation compared to bag-mask ventilation. Indeed, there was a trend towards worse oxygen saturation at the time of induction among the patients in the bag-mask ventilation group. Furthermore, adjusting for the mode of preoxygenation in a post-hoc analysis didn't affect the results.
- All patients in the bag-mask ventilation group received apneic oxygenation, whereas only 77% of patients in the no-ventilation group received apneic oxygenation. Therefore, better outcomes in the bag-mask ventilation group might theoretically reflect higher compliance with apneic oxygenation (100% vs. 77%). However, studies of apneic oxygenation have shown its efficacy to be limited in most patients.2 Thus, a 23% difference in apneic oxygenation is unlikely to have significant impact on the outcome of the study.
- Bag-mask ventilation is an operator-dependent skill. Mask ventilation within this study was performed in a specific fashion with several safety features (e.g., slow rate, low tidal volumes, two-handed technique). This study shouldn’t be interpreted to mean that mask ventilation is harmless, or that it can be performed in a careless fashion.
- Results may not apply to all situations. In particular, a no-ventilation strategy may still be safer among patients deemed to be at high risk of aspiration (patients who would have been excluded from this study).
Conclusions on this study
No single study is perfect, nor does any study establish 100% certainty on a topic (for example, additional studies are needed to confirm these findings in other contexts). That being said, this is an excellent multi-center trial. It unequivocally represents the best available evidence on this topic. In particular, its performance in ICUs (rather than the operating theater) is an important strength, improving generalizability to the critical care arena.
I think that this study gives the beneficial information that ventilation should be utilized in the interval between induction and tracheal intubation in cases in which the ‘RSI’ protocol is utilized. What do you think Stu?
— James DuCanto, M.D. (@jducanto) February 18, 2019
This study should change practice. Currently, the question probably isn’t whether we should provide some positive pressure to patients during induction. The real question is how to best achieve this, which will be explored next.
Unpacking the physiology of bag-mask ventilation during induction
Bag-mask ventilation achieves to functions: ventilation and oxygenation. For most intubations, only one of these matters: oxygenation.
Ventilation during induction is nearly always irrelevant
If a patient is totally apneic, the pCO2 will rise by ~1.8 mm per minute.3 This means that the actual increase in pCO2 during intubation is tiny. For example, in the Casey study above, the median apnea duration was ~2 minutes. Without any ventilation, this would increase pCO2 by only ~4 mm.
With modern airway techniques, it would be rare for the duration of apnea to be >5 minutes. This means that at most the pCO2 would rise by ~9 mm.
For the vast majority of patients, a pCO2 rise of 3-9 mm isn’t clinically relevant. Perhaps this could make a small difference for a profoundly acidemic patient, or a patient with severe pulmonary hypertension. But for the great majority of patients, a minor increase in pCO2 just doesn't matter.
Therefore, when providing bag-mask ventilation during induction, the volume of gas moving in and out of the lungs (the minute ventilation) is irrelevant. Giving big breaths to increase the patient’s minute ventilation isn't helpful.
Bag-mask ventilation during induction provides goodness by delivering pressure
The mechanism of benefit from bag-mask ventilation probably comes from increasing airway pressure, to prevent de-recruitment. The most important parameter here is the mean airway pressure – the average pressure in the airways over time. For example:
- A high peak airway pressure with zero PEEP is ineffective at keeping the lungs open.
- A lower peak airway pressure combined with PEEP is far better at preventing atelectasis.
The best way to safely recruit lungs during bag-mask ventilation is to slowly insufflate low tidal volumes (thereby achieving a low peak pressure) while using lots of PEEP. Keeping the peak pressure low avoids esophageal insufflation. Meanwhile, applying PEEP is the most efficient strategy to keep the lungs open (it directly boosts the mean airway pressure). Use of a PEEP valve during bag-mask ventilation is absolutely essential – without a PEEP valve, bag-mask ventilation is terrible at lung recruitment.
Would adding a nasal cannula under the mask help?
The graphs above show an idealized situation wherein the mask seal is perfect and maintained 100% of the time. In reality, there are likely to be intermittent mask leaks (especially if a person is physically holding the mask on, as opposed to a mask which is being held on with straps). Mask leaks will lead to de-recruitment:
One solution to this problem is to add a nasal cannula running at 15 liters/minute underneath the mask. Addition of oxygen within the mask should help maintain pressure and avoid de-recruitment due to mask leaks. The cannula may be kept on during laryngoscopy to provide apneic oxygenation. This strategy was designed by Scott Weingart and described further here:
Using a mask plus nasal cannula with no bagging at all could be a reasonable strategy for a patient at high risk of regurgitation. This would provide a low level of continuous positive airway pressure (CPAP) without any risk of esophageal insufflation (the PEEP valve prevents high levels of pressure from accumulating in the pharynx, making it impossible to insufflate the esophagus).
Algorithm for providing positive pressure during induction
Below is one approach to preventing de-recruitment. As discussed above, the goal here isn’t to provide ventilation (i.e. reduction of CO2). The aim is purely to provide sufficient pressure to prevent lung collapse leading to shunt physiology and hypoxemia. Focusing on this more narrow physiologic target allows for optimization of oxygenation, while avoiding aspiration.
The mechanism of insufflation-induced regurgitation is shown below (top panel). Insufflated gas distends the stomach. If enough fluid is present in the stomach, pressurized gas may propel this fluid backwards up the esophagus.
Placing the patient in head-up positioning may theoretically prevent this, as shown above. This positioning may also off-load the diaphragm and thereby reduce lung de-recruitment. It’s possible that head-up positioning could thus enhance the efficacy and safety of bag-mask ventilation during induction.
- Bag-mask ventilation during induction of anesthesia avoided desaturation, without increasing the risk of aspiration.
- The physiology of rapid desaturation during intubation involves atelectasis creating shunt physiology. Ventilation likely prevents this by using positive pressure to keep the lung inflated.
- The benefit of ventilation during intubation isn’t to promote CO2 clearance. Given how rapidly intubation is performed, CO2 is unlikely to increase significantly (even if the patient is completely apneic).
- Safety and efficacy of ventilation during induction might be enhanced by the use of a PEEP valve and relatively slow/small breath delivery. The goal is to achieve a high mean airway pressure (by using PEEP), while avoiding high peak airway pressures (to avoid esophageal insufflation).
- Mask ventilation in this study was performed meticulously using a two-hand technique, PEEP valve, slow rate (10/min), and low volumes (the smallest volume required to generate visible chest rise). This high-quality technique must be replicated in clinical practice in order to obtain the benefits described in the trial.
- EMCrit compendium of preox, reox, and deox information.
- Vent-as-bag technique to provide precise ventilation during induction (VAPOX)
- Using a pressure-limited ventilator (PulmCrit)
- Using a volume-limited ventilator (EMCrit)
- Scott Weingart and I differ slightly in our preferred techniques for vent-as-bag (probably due to our practice habits rather than any fundamental theoretical disagreement). I would use whatever technique seems more comfortable to you based on the type of machines you're used to and your typical practice patterns.
- EMCrit: head to sternal line concept – If you’re using head-up positioning, this is an essential anatomic concept to understand.
- Epic video by Reuben Strayer on high-quality two-handed ventilation is below (more on this at his blog here):
- PulmCrit Theoretical Post – The COVID Severity Index (CSI 1.0) - April 2, 2020
- PulmCrit wee – Why the SCCM/AARC/ASA/APSF/AACN/CHEST joint statement on split ventilators is wrong. - March 29, 2020
- PulmCrit- Is Lopinavir/Ritonavir down and out? - March 19, 2020