MCS Minute Series: Differential Gas Exchange on Peripheral Femoral VA ECMO: Dual Circulations, Differential Carbon Dioxide, and More

By Trina Augustin, MD, BSN

Peer reviewed by: Lewis McLean MD and Eric Leiendecker, MD

You just came on service and a new ECPR patient is rolling into your unit from the catheterization lab after having PCI in the setting of an acute myocardial infarction and OHCA. You glance at the monitor and see a perfusing rhythm, pulse pressure >20, MAP of 74, end tidal of 25, and a pulse oximetry of 96%. The nurses are settling the patient in and your respiratory therapist tells you he put the patient on “standard ECMO rest settings” with pressure control ventilation (Inspiratory pressure of 10 over PEEP of 10, RR 10, and FIO2 30%).

Done thinking? Or is there more to this story?

We are so conditioned to think “rest settings on VV ECMO” that sometimes it is easy to forget the incredibly nuanced gas exchange physiology we face in our patients on peripheral femoral VA ECMO.

Disclaimer 1:

This post does not insinuate that the management of gas exchange on VV ECMO does not hold an incredible amount of complexity. Complexities abound from the merits of spontaneous breathing on VV ECMO, with or without mechanical ventilation, to determining the best ventilator mode, optimal PEEP/driving pressure, mechanical power, respiratory rate, and more. But this debate will be a topic of a future post so stay tuned.

Disclaimer 2: Mechanical ventilation on VA ECMO is largely devoid of direct evidence, with most of our practice guided by expert opinion, consensus statements, retrospective studies, and reviews.

Now for a quick dive into the nuances of managing gas exchange on peripheral femoral VA ECMO.

Key Concepts:

• Gas exchange during peripheral femoral VA ECMO involves complex interactions. PMID: 34869455

Source: Diagram by Lewis McLean

• Patients on peripheral femoral VA ECMO with native cardiac ejection (or antegrade CO, such as with an Impella for LV venting) will develop a dual circulation—think two circulations “in parallel”. In a dual circulation system, some parts of the body are perfused by the native cardiac output, while others are supplied by blood flow from the ECMO circuit. This results in competitive flow between the blood ejected antegrade into the aorta from the heart and the blood ejected retrograde up the aorta from the ECMO circuit. The location where these two blood flows meet, known as the mixing cloud, primarily depends on the native cardiac function and the level of ECMO support (blood flow rate).

• • It is imperative to remember the location of the mixing cloud is dynamic. For example, increasing inotropic support, and thus native CO, can shift the mixing cloud more distally in the aorta. On the other hand, increasing ECMO flows (increasing LV afterload while simultaneously decreasing transpulmonary flow) will decrease native CO and shift the mixing cloud more proximal in the aorta.

• • A patient with no native ejection, which can occur right after ECPR cannulation, will not experience dual circulation. Instead, retrograde flow from the ECMO circuit will extend to the aortic root, providing perfusion to the entire body. However, this condition is typically short-lived, with dual circulation developing from antegrade cardiac output as native cardiac function recovers or an LV vent (such as an Impella) is placed.

• As VA ECMO features dual circulations, the oxygen and carbon dioxide content of the blood ejected by the heart antegrade into the aorta is determined primarily by gas exchange in the native lungs. In contrast, the ECMO membrane lung gas exchange determines the oxygen and carbon dioxide content of the blood ejected retrograde up the aorta. This distinction is crucial for managing gas exchange. Dual circulations can result in not only differential hypoxia but also differential carbon dioxide levels. PMID: 37891688

• How is VV ECMO different? In VV ECMO, the ECMO circuit dictates gas exchange for the entire body, with minimal contribution from the diseased lungs. Since VV ECMO flow is “in series” with native cardiac output, oxygenated blood from the ECMO circuit mixes with deoxygenated blood in the right atrium/vena cava, flows through the heart and lungs, and is ejected as a homogenous arterial flow providing oxygen delivery (DO2) for the entire body.

• This well-known phenomenon occurs when LV cardiac output (or antegrade flow from Impella) is present but native gas exchange is impaired, causing hypoxemic blood to enter the aorta proximal to the mixing point. Consequently, poorly oxygenated blood is delivered proximally, while well-oxygenated blood from the ECMO circuit perfuses the aorta distal to the mixing point. The end organ effects of the dual circulation here concern two organs with high oxygen extraction ratios – the brain and the heart. If the mixing point is in the proximal descending aorta, the coronaries, brain, and upper extremities may receive severely impaired oxygen delivery, while the lower body receives excellent oxygen delivery, resulting in differential hypoxia. The placement of a right upper extremity (RUE) arterial line is to alert the clinician to hypoxemic blood perfusing the coronaries and cerebral circulation.

We spend a lot of time talking about differential hypoxia but what about differential carbon dioxide levels?

• Dual circulations with different carbon dioxide levels: The PaCO2 level proximal to the mixing cloud is primarily determined by native lung ventilation, while PaCO2 levels distal to the mixing cloud is determined by ECMO Sweep gas flow rate. PMID: 37891688. Warning this may be oversimplified.

Disclaimer 3: Native gas exchange can be influenced by ECMO gas exchange through renal compensation. Additionally, limited data suggests that increasing sweep gas flow for non-intubated patients on peripheral VA ECMO can reduce dyspnea and respiratory drive. This pathophysiology is not completely elucidated. One hypothesis for the decreased respiratory drive was increased sweep gas decreasing PaCO2 (altering pH) and triggering chemoreceptors, while another suggests decreased PACO2 (alveolar partial pressure in carbon dioxide) leading to decreased stimulation of carbon dioxide- sensitive pulmonary C-fibers thus decreasing respiratory drive. While the mechanism is somewhat unclear yet when considering dual circulation physiology on peripheral VA ECMO (where we typically think of sweep gas changes primarily effected the body distal to the mixing cloud and not affecting chemoreceptors/respiratory drive primarily), it likely involves an “ECMO-related reduction of carbon dioxide stores” thereby allowing ECMO gas exchange to impact native respiratory drive. PMID: 38436930

Since we might not be accustomed to considering differential carbon dioxide levels, let's explore an example to understand why this is important…

Imagine you're in the ICU, managing a critically ill patient who was placed on VA ECMO during a peri-arrest situation. His native LV function is recovering with PP>30 and he appears to have intact native gas exchange. His mixing cloud is distal to the innominate artery. He is on a Bumetanide (BUMEX) drip to optimize his volume status before working towards ECMO liberation.

Vital parameters

• Minute ventilation is 5.1 Liters
• ECMO blood flow 4L, FdO2 60%, and ECMO sweep gas rate 1.5L
• RUE ABG 7.53/47/105/39

His metabolic alkalosis is attributed to ongoing diuresis with loop diuretics. But is this the true cause?

What if his post oxygenator gas yesterday was 7.29/66/130/28 on a sweep gas flow of 1.5L? Hence, his kidneys were being perfused with acidotic blood (respiratory acidosis in the setting of inadequate sweep gas flow rate) from the ECMO circuit causing renal compensation with increased hydrogen ion excretion and HCO3 reclamation with a resultant metabolic alkalosis. This is where thinking about the end organ effects of the dual circulation becomes very important.

Remember, while you may have a dual circulation—two circulations in parallel—they are still very interdependent.

Steps to consider when managing Differential Gas Exchange on peripheral femoral VA ECMO

• Ensure appropriate monitoring of the dual circulation (“both circulations”) PMID: 37891688
  • RUE arterial line for periodic ABGs (+/-RUE pulse oximetry)
    • Assess oxygen delivery (DO2) to the innominate artery (1^st branch of the aorta)
    • Classically used as a surrogate to approximate carotid/cerebral DO2, thought to best approximate what the brain “is seeing”
    • Native ejection + CXR “whiteout” (assumed impaired native gas exchange)
      • Well oxygenated RUE ABG: Either native gas exchange better than expected or very proximal mixing cloud with innominate artery perfused by ECMO flow and hence brain as well. Coronary perfusion is unknown. Can compare post-oxygenator and RUE ABG to further elicit the source.
      • Poorly oxygenated RUE ABG: assume mixing cloud distal to innominate with poor DO2 to coronaries and brain (DO2 to the brain can be variable if mixing cloud is in aortic arch)

• Post oxygenator gas: Assess the contribution of ECMO gas exchange.
  • We usually use this to assess membrane lung function, but it’s important to remember this is what the ECMO is delivering in terms of oxygenation and decarboxylation.
  • We always focus on what the coronaries are seeing but what the kidneys (and other organs) are seeing is also important!

• Is there native ejection? Obtain gases as discussed above to help delineate location of mixing cloud
  • In some cases echocardiography can help delineate the position of the mixing cloud….

Video: Mixing cloud in the descending aorta on TEE. Video credit Lewis McLean

• Keep in mind that the location of the mixing cloud is dynamic and requires continuous reassessment with any changes in left ventricular cardiac output (such as native LV recovery, use of inotropes, vasodilators), ECMO flow adjustments, or the placement of an LV vent.

• LV unloading and its role in the mixing cloud….
  • Proximal aortic/subclavian return cannulation (sport mode), Surgical LV vent, atrial septostomy, pulmonary artery vent, and LAVA ECMO will push the mixing point back towards the aortic root, resulting in a more proximal mixing point.
    • Surgical LV venting, atrial septostomy, PA vent, and LAVA ECMO do this by returning more of the transpulmonary flow back to the access side of the circuit
    • Sport mode cannulation achieves this by pressurizing the proximal arterial tree directly with the return flow, usually eliminating a clinically relevant mixing point
  • ECPELLA, and to a lesser degree IABP, will move the mixing point away from the aortic root, resulting in a more distal mixing point
    • This will be relatively fixed with IABP
    • This may be adjusted by manipulation of Impella flow
      • Low flow/P level – more proximal mixing point
      • High flow/P level – more distal mixing point

• Imperative to check both RUE ABG and post oxygenator gas periodically

• While they are two circulations in parallel, remember they are still interdependent through the body’s compensatory mechanisms

• The upper body typically has a higher oxygen extraction than the lower body, so if your venous multistage drain is not bicaval (also draining the SVC) this could further contributing to variations in regional gas exchange PMID: 33232

• VA ECMO significantly decreases transpulmonary blood flow by rerouting a significant portion of venous return in the ECMO drainage cannula; hence, minute ventilation on mechanical ventilation (MV) typically needs to be decreased to match ventilation to perfusion

• The goal is to match alveolar ventilation to transpulmonary blood flow/perfusion, so remember to account for physiological (anatomic + alveolar) dead space and adjust accordingly (approximately 2ml/kg of ideal body weight) PMID: 29494107

Clinical Pearl: if your patient’s CXR is “whited out” from pulmonary edema, contusions post CPR, or ARDS from PNA—they may have increased physiological dead space not participating in gas exchange. These patients may not require as much of a decrease in their minute ventilation post cannulation as you would expect.

• In VV ECMO the focus is typically “lung rest” settings (i.e. ultra-lung protective ventilation) and we expect minimal contribution from native gas exchange; whereas, in VA ECMO the focus should be on lung protective ventilation while ensuring adequate gas exchange to the portion of the body perfused by the native circulation

• Inability to maintain adequate gas exchange with lung protective ventilation may require additional therapies-not addressed in this post. Stay tuned for future discussions…

• Ideally target normoxia and normocapnia from both the native gas exchange and ECMO circuit (dual circulations)

• Recent expert consensus on gas exchange post ECPR recommend the following:
  • PaO2 100-150
    • Hyperoxia associated with increased ROS, organ dysfunction, and increased morbidity and mortality PMID: 36871240, PMID: 38061577
  • PaCO2 40-45
    • Avoid rapid drops in PaCO2, potentially associated with increased neurological complications PMID: 32574466
  • pH >7.2
  • While this consensus recommended mechanical ventilation with AC/VC ( TV 6ml/kg of ideal body weight, RR 12, PEEP 8), a recent international multicenter survey indicated that pressure control ventilation was the most commonly utilized mode of ventilation in VA ECMO

• There is ongoing equipoise regarding gas exchange targets and optimal ventilator management inpatients on VA ECMO. Most ELSO data is focused on VV ECMO with recommendations for plateau pressure ≤25, DP ≤15, low FiO2, low RR (4-15), individualized PEEP ( ≥10 cm H2O) and low TV (typical <4 ml/kg) PMID: 33965970. VA ECMO patients are a different population with different physiology so it is difficult to extrapolate this data; however, it is still important to avoid lung injurious settings, especially if concurrent ARDS

Trina's approach…

First, my strategy for mechanical ventilation (MV) in these patients is influenced by my preference for awake ECMO. I aim to avoid MV entirely in some patients, and in others, I strive for early MV liberation to reap the benefits of reduced sedation, lower risk of delirium and ventilator-associated pneumonia, and the ability to mobilize patients and engage in rehabilitation. My incredible team routinely walks femorally cannulated VA ECMO patients.

However, this approach may not be feasible for many patients, especially the subset placed on VA ECMO during ECPR. Mechanical ventilation can provide significant cardiac benefits, such as optimizing left ventricular preload and afterload, reducing the work of breathing and thus oxygen demand, and enhancing gas exchange with increased myocardial oxygen delivery.

• First, assess the “lung rest” PCV 10/10/10/variable FIO2 that will have invariably been set automatically after ECPR cannulation :). Not in frequently this may actually be working and giving you appropriate minute ventilation + gas exchange as it often decreases their minute ventilation to target the sudden decrease in pulmonary blood flow.

• Set ventilator parameters with individual specific cardiac and lung physiology in mind. Do they have concomitant ARDS, pulmonary edema, isolated RV failure, etc.

• Typically start with a volume control or pressure control mode. My goals are to minimize mechanical power with driving pressure <15, TV≤6ml/kg, RR (as low as possible for minute ventilation needs), and plateau pressure ≤25-30

• I individualize PEEP to optimize their functional residual capacity (FRC) and maintain “open lung ventilation”

• Typically titrate ventilator FIO2 to PaO2 on RUE ABG. If suspicious of mixing point between innominate & coronaries may maintain higher FIO2 to ensure adequate coronary DO2

• Strive to avoid ventilator dyssynchrony as well as patient self-induced lung injury (P-SILI) especially as I like to minimize sedation and work towards awake ECMO. Tend to find PCV or PS more well tolerated. As always close monitoring needed to ensure appropriate TV, DP, and patient effort

• Hypoxic patient despite optimized PEEP: Increase mean airway pressure while aggressively optimizing lung function and determining if additional interventions needed
  • Increased inspiratory time (I:E 1:1) on PCV (inverse ratio ventilation)
  • APRV
  • Consider V-AV ECMO configuration

Take Home points:

• Patients with native ejection on peripheral femoral VA ECMO will develop dual circulations—”two circulations in parallel”
• Dual circulations can cause differential gas exchange, resulting in differential hypoxia and differential carbon dioxide levels
• It is crucial to ensure adequate gas exchange in both circulations while remembering that are still interdependent
• Peripheral VA ECMO decreases transpulmonary blood flow, so its import to decrease minute ventilation to match lung ventilation to perfusion
• The goal of mechanical ventilation on peripheral VA ECMO is to provide “lung-protective ventilation” while ensuring adequate gas exchange to the body areas it perfuses

Additional reading:

• ELSO general guidelines for all ECLS
• Asija, R., et. Al. How I manage differential gas exchange in peripheral venoarterial extracorporeal membrane oxygenation. Critical care (London, England), 27(1), 408.

Additional New Information

More on EMCrit

• EMCrit Wee – Jason Bartos on the Minneapolis ECPR Experience
• EMCrit 265 – ECPR 2.0 (ECMO CPR)
• EMCrit 222 – Demetris Yannopoulos on ECPR-the Minneapolis Way [ECMO]

• Author
• Recent Posts

Katrina Augustin

Cardiovascular Intensivist and emergency Medicine Consultant at Mayo Clinic, Rochester, MN Campus

Katrina (Trina) Augustin is a Cardiovascular Intensivist and Emergency Medicine consultant at Mayo Clinic in Rochester Minnesota where she also works as part of the ECMO team. She completed an Emergency Medicine Residency and chief residency at the University of Pittsburgh Medical Center. Trina's interests include cardiovascular critical care with a focus on mechanical circulatory support devices including ECMO. She is passionate about using FOAMed as an educational opportunity to improve the care of patients presenting with cardiovascular emergencies.

Latest posts by Katrina Augustin (see all)

• CV-EMCrit – MCS Minute Series: Differential Gas Exchange on Peripheral Femoral VA ECMO with Trina - September 16, 2024
• CV-EMCrit – Cardiovascular Intensivist Series:Post Cardiopulmonary Bypass Care (CABG) with Trina Augustin - May 25, 2024
• CV-EMCrit Wee – MCS Minute: ECMO and the DO2/VO2 ratio with Trina Augustin - January 31, 2024