COVID-19 might out-strip the number of mechanical ventilators available to us. This has led to interest in using a single ventilator to support multiple patients. This post will review the theory and evidence regarding this (with the admission that I don’t have experience with this).
Bedrock principle: Patient-Ventilator Independence
Normally, we adjust the ventilator so that the ventilator is adapted well to suit an individual patient’s needs. The adage is “fit the ventilator to the patient, don’t fit the patient to the ventilator” – in other words, adjust the ventilator to keep the patient comfortable, rather than over-sedating the patient to tolerate an uncomfortable ventilator mode.
We cannot do this when splitting the ventilator. In fact, any interaction where the patient drives the ventilator is problematic (because this allows one patient to affect another patient’s ventilation). For example, we wouldn’t want one patient’s tachypnea to cause other patients attached to the same ventilator to be hyperventilated.
Thus, a bedrock principle of multiple-patient ventilation is that each patient should have no effect on the ventilation of other patients attached to the ventilator. This is achievable, as described below. The alternative is chaos.
Pressure-cycled ventilation is superior to volume-cycled ventilation
The debate regarding pressure-cycled ventilation versus volume-cycled ventilation is perpetual in critical care. Most units and physicians have some preference, but either strategy works well for most patients. In short, the advantages of each are as follows:
- Volume-cycled ventilation: Advantage is delivery of a guaranteed tidal volume (disadvantage is lack of control over peak pressure).
- Pressure-cycled ventilation: Advantage is guaranteed limitation of the peak pressure (disadvantage is lack of control over tidal volume).
Once we start splitting a ventilator between multiple patients, this debate largely evaporates. Using a volume-cycled mode has numerous, major disadvantages:
- Using a volume-cycled mode with multiple patients provides no control over the tidal volume of any patient, and also provides no control on the maximal airway pressure. This is literally the worst of both worlds.
- A volume-cycle mode will introduce the possibility of deleterious interactions between patients. For example, let’s suppose Patient A’s endotracheal tube gets kinked. This will cause Patient B to receive dangerously large tidal volumes!
- Patients sharing the ventilator must have similar size, similar FiO2 and similar PEEP requirements.
Using a pressure-cycled mode solves these problems:
- With a pressure-cycled mode, we retain control over the maximal airway pressure and the driving pressure. We cannot deliver a guaranteed tidal volume to any patient, but that is no different from having any patient on pressure-cycled ventilation. Ability to control and limit the driving pressure may allow this strategy to be reasonably lung-protective.1
- Deleterious interactions between patients are avoided using a pressure-cycled mode. For example, if Patient A’s endotracheal tube gets kinked in a pressure-cycled mode, then Patient A will receive a reduced tidal volume. However, this will have no impact on Patient B.
- Patients sharing the ventilator don’t have to have a similar size. Larger patients will tend to have a greater absolute compliance, so they will receive larger breaths.
Continuous mandatory ventilation (CMV) is required
Normally, patients are able to trigger the ventilator to deliver a breath. This isn’t possible if a single ventilator is being used to support multiple patients (because, as mentioned earlier, one patient’s tachypnea could cause all patients attached to the ventilator to be hyperventilated).
Therefore, the mode of ventilation which must be used is continuous mandatory ventilation (CMV). What this means is that the ventilator fires at a set rate. The patient has no control over the respiratory rate (i.e. the patient cannot trigger a breath). This is an antiquated mode of ventilation, because it’s generally uncomfortable. However, it’s the only way to achieve patient-ventilator independence.
Modern ventilators may lack a continuous mandatory ventilation mode. However, the same effects might be achieved as follows:
- Increase the ventilator trigger threshold as high as possible, so that it’s impossible for patients to trigger a breath (a.k.a. “lock out” the ventilator).
- If #1 is unsuccessful, respirolytic sedation (using drugs that suppress respiratory drive such as opioids and propofol) could be used to reduce patients’ respiratory drive and prevent them from triggering the ventilator. Paralysis would be used only as a last resort.
Ventilation efficacy will be sub-optimal
Carbon dioxide clearance will not be optimized by a multi-ventilator strategy for a few reasons:
- Tidal volumes will be difficult to track and optimize.
- Y-site connections and tubing increase dead space
It will probably be necessary to accept permissive hypercapnia. For patients with substantial acidosis, IV bicarbonate may be needed to support pH (more on this here). As is usually the case in ARDS, the primary focus is providing lung-protective ventilation, rather than optimal blood gas parameters.
Management of profound hypoxemia: Pressure Control Inverse Ratio Ventilation (PC-IRV)
Patients with COVID-19 appear to be relatively responsive to PEEP. Of course, PEEP is merely one way to increase the mean airway pressure – which is the most important variable affecting lung recruitment. For profound hypoxemia, inverse ratio ventilation may be used to increase the mean airway pressure even further.
Inverse ratio ventilation involves increasing the inspiratory time, so that the patient is spending most of the time in the inspiratory phase (inspiratory time > expiratory time). Inverse ratio ventilation is generally not used because it’s uncomfortable, but in this context patients will be deeply sedated anyway. The overall concept here is similar to APRV – trying to maintain an “open lung” with ongoing application of gentle amounts of pressure (rather than abrupt high-pressure recruitment maneuvers).
basic setup to split a single ventilator
So, what this leaves with is the following:
- Multiple patients attached to a single ventilator. The patients don’t need to be the same size, but ideally they should have roughly similar severity of lung injury (e.g. similar PEEP and FiO2 requirements)(more on achieving this matching below).
- The ventilator is set to pressure-cycled ventilation with a high PEEP (noting that patients with COVID-19 seem to be highly PEEP-responsive) and a low driving pressure (to achieve lung protection). For example, a setting of 30 cm / 18 cm might be reasonable for many patients.
- The ventilator trigger is locked out, to prevent patients from triggering breaths.
- Patients will likely require deep sedation to render them passive on the ventilator (e.g. propofol plus opioids). Paralysis isn’t necessarily required, but it may be necessary in some cases, depending on how sensitive patients are to sedation.
- Ventilation efficacy of each patient can be tracked using an end-tidal CO2 monitor placed in-line with their own endotracheal tube (in a shortage of etCO2 sensors, it might be possible to use a single sensor and rotate it between patients to spot-check the pCO2 of each patient sequentially).
- Permissive hypercapnia will need to be anticipated and managed, as discussed above.
- Viral filters should be used to prevent cross-contamination of pathogens between different patients.
In theory, a single ventilator could be used to support multiple patients (e.g. 2-4 patients, possibly even 6 or 8?). At some point the ventilator may not be powerful enough to support the summed tidal volumes of all the patients.
bigger picture: Five ventilators to provide personalized settings to 20 patients
A major drawback of the above setup is that patients must be matched based on relative severity of lung injury (PEEP and FiO2 requirements). This issue could be overcome as follows.
Imagine that we set up five ventilators:
- Ventilator 1: Mild injury settings (FiO2 50%, PEEP 10 cm, Peak pressure 20 cm)
- Ventilator 2: Moderate injury settings (FiO2 60%, PEEP 14 cm, Peak pressure 26 cm)
- Ventilator 3: High injury setting (FiO2 80%, PEEP 18 cm, Peak pressure 30 cm)
- Ventilator 4: Refractory hypoxemia settings (FiO2 100%, PEEP 22 cm, peak pressure 35 cm).
- Ventilator 5: Salvage settings (FiO2 100%, PEEP 22 cm, peak pressure 35 cm, inverse ratio ventilation with inspiratory time >> expiratory time).
Each of the ventilators could be connected to 1-4 patients. If patients deteriorated, they could be moved to a higher-number ventilator (e.g. from Ventilator #2 to #3). Alternatively, as patients improved, they could be shifted to a lower-number ventilator. This system could allow a handful of ventilators to provide reasonably personalized settings to a large number of patients.
This general concept has been demonstrated using animals and lung models.2,3 However, depending on exactly how the model is constructed and what gauge is used to determine “success,” different results can be obtained. For example, Branson et al. demonstrated that multi-lung ventilation cannot be used to deliver precise tidal volumes.4 That’s wholly predictable based on physics, so it doesn’t actually reveal anything. So, if we are using delivery of a fixed tidal volume as a gauge for success, then multi-patient ventilation will fail. However, if we are using delivery of a fixed driving pressure as a gauge for success, then multi-patient ventilation may succeed.
One published report does describe the use of split ventilation in two volunteers (using a facemask interface, rather than intubation). Pressure cycled ventilation was successfully applied with good results.5
- It is almost certainly possible to ventilate several patients with a single ventilator. This probably can be achieved with reasonably lung-protective settings (i.e. low driving pressure). However, the cost of this strategy is loss of control over precise tidal volumes and suboptimal ventilation (with high pCO2).
- A fundamental goal of multi-patient ventilation is to prevent any patient from affecting the other patients. This may be achievable using pressure-cycled ventilation without the ability of any patient to trigger the ventilator.
- Patients would need to be deeply sedated and passive on the ventilator (or paralyzed if necessary).
- Each patient’s ventilatory efficiency could be monitored using end tidal CO2. This would be required as a surrogate for tidal volume or minute ventilation (which will not be measurable).
- By using a small number of ventilators with a range of different settings, a large group of patients could be supported with fairly personalized settings.
- COVID IBCC chapter here (with additional links to other COVID resources at the bottom)
- Columbia Presbyterian protocol for splitting ventilators here.
- More advanced and granular exploration about how to hook everything up here.
- 1.Amato M, Meade M, Slutsky A, et al. Driving pressure and survival in the acute respiratory distress syndrome. N Engl J Med. 2015;372(8):747-755. doi:10.1056/NEJMsa1410639
- 2.Neyman G, Irvin C. A single ventilator for multiple simulated patients to meet disaster surge. Acad Emerg Med. 2006;13(11):1246-1249. doi:10.1197/j.aem.2006.05.009
- 3.Paladino L, Silverberg M, Charchaflieh J, et al. Increasing ventilator surge capacity in disasters: ventilation of four adult-human-sized sheep on a single ventilator with a modified circuit. Resuscitation. 2008;77(1):121-126. doi:10.1016/j.resuscitation.2007.10.016
- 4.Branson R, Blakeman T, Robinson B, Johannigman J. Use of a single ventilator to support 4 patients: laboratory evaluation of a limited concept. Respir Care. 2012;57(3):399-403. doi:10.4187/respcare.01236
- 5.Smith R, Brown J. Simultaneous ventilation of two healthy subjects with a single ventilator. Resuscitation. 2009;80(9):1087. doi:10.1016/j.resuscitation.2009.05.018
- 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