The EMRA Critical Care Committee's Roadblock series is the resuscitationist's guide to overcoming obstacles in the normal algorithm of critically ill patients.
Refractory Hypoxia on the Ventilator
Special thanks to John Gaillard, MD, Associate Professor of Anesthesiology and Emergency Medicine at Wake Forest University School of Medicine, and Andrew Petrilli, MD, Assistant Professor of Anesthesiology at Wake Forest University School of Medicine and Assistant Program Director of the Wake Forest Critical Care Medicine Fellowship, who are both emergency medicine-trained intensivists practicing critical care in the Cardiothoracic ICU as well as a mixed Medical/Surgical ICU. Dr. Gaillard additionally works in the ED at Wake Forest Baptist Medical Center. Thanks also to EMRA Critical Care Committee Chair Dustin Slagle, MD, who reviewed this paper.
Some people use mnemonics (like DOPES) to guide their initial assessment of the hypoxic patient on the vent. Do you use a specific structure like this, or what is your initial approach in evaluating such patients?
Dr. Gaillard: I first evaluate O2 sat: is it a good signal, is the patient actually hypoxemic? Then I evaluate the patient, the endotracheal tube (ETT), and the ventilator. In terms of the patient, I listen for breath sounds, watch their breathing for signs of asymmetry (pneumothorax or severe atelectasis) or hyperexpansion (bronchospasm or breath stacking). I suction the ETT to ensure patency. I then look at the ventilator, focusing on things like ensuring everything is connected, my vent numbers - specifically I look at the mode and settings, the FiO2, the patient's peak and plateau pressures, and tidal volumes including comparison of exhaled vs. inhaled tidal volume (looking for air leak).
Petrilli: I like to approach it by ruling out everything but the lungs first. You can use mnemonics like DOPES; I tend to start at the ventilator and work my way towards the patient, asking myself, "Is this a mechanical problem (tube dislodgement or obstruction, circuit kink, equipment problem, etc.), sedation issue, obstructive problem (bronchospasm, breath stacking, mucus plug), etc.?" Once I've ruled out everything else, I focus on the lungs themselves and what I can do about sick lungs getting sicker.
What diagnostic tests do you get initially to guide management?
Gaillard: I like to throw the ultrasound probe on the chest to look for things like pneumothorax, consolidations, right heart strain, and globally assess cardiac function. I will also get a chest X-ray to evaluate for lung pathology and check ETT position. If one lung is completely whited out I consider mucus plug and that patient may need chest PT and a bronch. If the lungs look perfectly clear, I start to think about pulmonary embolism or other causes of shunt physiology. If the lungs show worsening pneumonia or ARDS, the problem may simply be that your patient's lungs are getting sicker.
Petrilli: I agree with the chest X-ray and ultrasound as well as a good physical exam. With regards to getting an ABG specifically, I think a baseline ABG can be helpful to establish that it is correlating well with the pulse ox, especially for patients in which pulse ox may be less reliable like those with darker skin tones. I do not believe in monitoring serial ABGs, however. It is painful and more time consuming in patients without an arterial line, and I think that monitoring VBGs and the pulse ox works just as well.
What information are you getting from the ventilator when assessing the hypoxic patient?
Gaillard: Of course I'm looking at the mode and settings first and any ventilator alarms. Then I’ll look at peak and plateau pressures. If tidal volume (TV) is set, is the patient reaching that? What is the exhaled TV compared to inhaled TV (is there an air leak)?
Petrilli: I'm looking at the pressures needed to overcome airway resistance and oxygenate and ventilate lung units with respect to the patient’s lung compliance. If peak pressures are >40 cm H2O, that indicates that the ventilator may be utilizing an excessive amount of pressure to overcome pathologic airway resistance or poor lung compliance. The inspiratory hold maneuver described below helps delineate problems with resistance versus compliance.
What ventilator maneuvers can you try to either troubleshoot or correct refractory hypoxemia?
Gaillard: You should know how to do an inspiratory and expiratory hold. The inspiratory hold shows your plateau (alveolar) pressure (goal is <30 cm H2O to reduce risk of lung injury). High plateau pressures can indicate too much PEEP or tidal volume, air trapping, patient-ventilator dyssynchrony, or poor compliance. The expiratory hold will allow you to calculate intrinsic (“auto”) PEEP. Auto PEEP indicates breath stacking which can occur with obstructive lung disease such as asthma or COPD, or other causes of air trapping like mucus plugging or inappropriate vent settings (too high TV or too short exhalation time).
Petrilli: The inspiratory hold can help you determine if there is a lung parenchyma problem or an airway resistance problem in the case of high pressure alarms. If the plateau pressure and the peak pressure are both high, this most often indicates a lung compliance issue (the sick lungs themselves are the problem). High intra-abdominal pressure can also cause an elevated plateau pressure via transmission of pressure via the diaphragm, so that is one confounder of which to be aware. If the plateau pressure is low or normal with high peak pressure, this indicates a resistance problem (e.g., mucus plug, patient biting tube or other tube issue, bad obstructive disease).
In the past there was a push towards recruitment maneuvers such as “40 for 40” (increasing PEEP to 40 mmHg for 40 seconds). More recent data have demonstrated that this actually leads to worse outcomes, particularly in moderate to severe ARDS, so I try to stay away from those. The current American Thoracic Society (ATS) guidelines on management of ARDS also recommend against recruitment maneuvers.
Do you typically disconnect the ventilator in your troubleshooting process?
Gaillard: Some people do this to differentiate between a ventilator issue and an ETT or patient problem. I personally think you can get the same information from the ventilator graphics and numbers themselves. If the graphics are concerning for significant auto PEEP, I will consider disconnecting the patient from the vent.
Does any ventilator mode itself show a benefit over others in refractory hypoxemia?
Petrilli: There is no good evidence for this. A lot of it is institution-dependent or personal preference, but no mode has been shown to improve outcomes over others. Some individuals are big proponents of airway pressure release ventilation (APRV) and I think this can be a good mode if you know how to use it and titrate it (this is not a "set it and forget it" type of mode), but there isn't good evidence for true benefit over other modes at this point. Personally I feel that if you are going to use it, it should be started early on in the treatment course of severe disease and not relegated to a "rescue" mode.
Once you've ruled out other causes besides sick lungs, what is your stepwise approach to refractory hypoxia in ARDS?
Petrilli: My practice is generally consistent with the most recent ATS ARDS guidelines. A basic foundation to managing these patients is to ensure lung-protective ventilation with low tidal volumes (6-8 cc/kg or even 4-6 cc/kg ideal body weight) and low airway pressures (maintaining plateau pressure <30 cm H2O and driving pressure < 15 cm H2O). Systemic corticosteroids are also controversial but currently recommended for all patients with ARDS regardless of disease severity, with the DEXA-ARDS trial supporting early use.
The simplest way to start addressing hypoxia itself is to titrate PEEP and FiO2. The updated 2023 ATS ARDS guidelines now recommend a high PEEP strategy in patients with oxygenation and recruitment difficulty (using the high PEEP table from the ARDSNet trial). If you have esophageal manometry available, you can use this to estimate pleural pressure and thus calculate transpulmonary pressure and more accurately titrate PEEP. If you don't have this technology, you can sequentially increase the PEEP while checking plateau pressure via the inspiratory hold maneuver - if you’re increasing PEEP and your plateau pressure is still <30 cm H2O and driving pressure < 15 cm H2O, this can indicate that you are recruiting more lung units to address hypoxemia. This just needs to be done with caution, as excess PEEP can cause hypotension due to increased intrathoracic pressure and right ventricular dysfunction due to increased pulmonary vascular resistance.
The next things I consider if the patient remains hypoxemic are patient positioning and synchrony. The PROSEVA trial showed that prone positioning provides a mortality benefit in patients with moderate to severe ARDS (PaO2/FiO2 [P/F] ratio <150) by improving V/Q matching. Prone positioning is especially beneficial in patients with dependent atelectasis and posterior disease when supine. I prone patients for 16 hours at a time when hypoxemia does not improve with titration of PEEP and FiO2 as above (ie, PaO2/FiO2 remains below 150).
I will then trial temporary paralysis in patients with severe ARDS or those with dyssynchrony on the vent despite adjustments to sedation. The ACURASYS trial compared 48-hour paralysis with cisatracurium to heavy sedation and demonstrated a comparative mortality benefit with paralysis in patients with severe ARDS. The more recent Rose Trial, however, compared neuromuscular blockade to light sedation in severe ARDS patients and showed no benefit of paralysis. Because of the conflicting evidence this is a bit controversial, but I tend to trial paralysis in severe ARDS patients that remain dyssynchronous and have poor compliance despite heavy sedation. If their lung compliance improves with the paralytic trial, I will start them on a cisatracurium infusion for 48 hours or until their pulmonary mechanics and hypoxia improve.
Another option to consider is use of inhaled pulmonary vasodilators (eg epoprostenol [Flolan], nitric oxide, milrinone). Current evidence does not support any mortality benefit or reduction in number of days on the ventilator with these agents, but they do improve oxygenation by helping to reduce ventilation-perfusion mismatch and are low-risk medications. Given this, many clinicians will still try them in patients with refractory hypoxemia. I think the patient population that may most benefit from these agents are patients with pulmonary hypertension at baseline or right heart failure.
Gaillard: I typically walk through a similar stepwise approach of considering a change in ventilator mode, increasing FiO2 and PEEP while being mindful of peak and plateau pressures, proning, starting an inhaled pulmonary vasodilator (understanding that this improves oxygenation without improving mortality), trialing heavy sedation, and paralyzing if needed.
At what point are you considering V-V ECMO and in which patients?
Petrilli: Once I've tried all the strategies above, if the patient is still severely hypoxic, I consider V-V ECMO based on evidence (though mixed and difficult to interpret) from the EOLIA and CESAR trials. The Extracorporeal Life Support Organization (ELSO) guidelines recommend consideration of V-V ECMO in patients with potentially reversible respiratory failure refractory to conventional treatments who have: refractory hypoxemia (P/F <80) or severe hypercapnic respiratory failure (pH <7.25 with PaCO2 ≥60 mmHg). ELSO and ATS additionally recommend that patients should be early in the ARDS course (on ventilator for <7 days), which I generally agree with, though if the patient is young with reversible disease I will sometimes extend this window.
Gaillard: If all of the other strategies fail, I consider V-V ECMO in patients with reversible disease and without significant comorbidities who have been on the ventilator for <7 days. The RESP score can be used to predict survival after ECMO for patients with severe respiratory failure and can also be helpful in determining whom to cannulate.