Secondary Logo

Journal Logo

Review Article

Management of Refractory Hypoxemia During Venovenous Extracorporeal Membrane Oxygenation for ARDS

Montisci, Andrea; Maj, Giulia; Zangrillo, Alberto; Winterton, Dario; Pappalardo, Federico

Author Information
doi: 10.1097/MAT.0000000000000207
  • Free


According to the recent criteria of the Berlin definition, acute respiratory distress syndrome (ARDS) is defined as an acute, diffuse lung injury syndrome that appears within 1 week of a known clinical insult or new or worsening respiratory symptoms.1

Despite the progress in medical therapies and ventilation management, ARDS still carries a high mortality risk increasing with stages of ARDS from mild (20%) to moderate (41%) to severe (52%).1

Extracorporeal membrane oxygenation (ECMO) is a modified cardiopulmonary bypass that provides gas exchange (venovenous [VV] ECMO) and ensures systemic perfusion (venoarterial [VA] ECMO) to sustain the cardiopulmonary function of the patient.2

The recent pandemic of H1N1 influenza, and its considerable risks of evolution to ARDS in previously healthy young people, has kindled the interest and the research on ECMO, a supportive therapy of the failing heart and lungs already used in past decades with various indications.

Current evidence and future applications of VV ECMO have been recently summarized by Agerstrand et al.3 Two studies conducted in 2009, involving the use of current ECMO technology, showed a survival benefit in ARDS patients treated with ECMO. The Conventional ventilation or ECMO for Severe Adult Respiratory failure (CESAR) trial,4 a multicenter randomized trial conducted in United Kingdom, in which ARDS patients were randomized to receive conventional mechanical ventilation versus referral to an ECMO center, showed a 16% reduction in the ECMO-referred group in the primary end point considered (death or severe disability at 6 months).

Davies et al.5,6 in the Australia and New Zealand ECMO experience, reported an in-hospital mortality rate of 25% in the group of ARDS patients treated with ECMO during the H1N1 influenza pandemic. A common observation after venovenous bypass positioning is that hypoxic and cyanotic patients, with SaO2 values close to 80%, reacquire normal oxygenation and CO2 tension.

Extracorporeal Life Support Organization (ELSO) guidelines suggest that during VV ECMO for ARDS the achievement of a SaO2 > 80% and a SvO2 > 70% indicates adequate support,7 and that the compensatory increase of cardiac output (CO) and the achievement of an hematocrit more than 40% ensure sufficient O2 delivery (DO2).7,8

A possible objection could be that O2 tension might be too low and worsen the patient’s prognosis.9 Moreover, this concept assumes a normal cardiac function, which can compensate for hypoxemia, and that the patient is ventilated in a protective manner.

This review focuses on that subset of patients who remain hypoxemic during VV ECMO, on its pathophysiological issues and on the possible strategies for treatment, addressing clinical scenarios in which the hypoxemia is not related to treatable causes (i.e., pneumothorax or equipment failure) but just to the severity of underlying lung disease.

The management of mechanical ventilation during VV ECMO is beyond the scope of this review, and a comprehensive review on this topic is referenced.10

What O2 Peripheral Saturation Target During VV ECMO?

As stated above, during VV ECMO for ARDS, patients are often hypoxic despite the extracorporeal support, with partial pressure of oxygen (PaO2) values of 45–55 mm Hg.

Neither the safe PaO2 lower limit in critically ill patients nor the threshold at which the brain is injured, is known.

In many clinical situations, as in utero development or in case of exposure to hypobaric hypoxia at high altitude, humans are exposed to sustained hypoxia, during which many nonessential cellular processes may be temporarily suspended. This phenomenon, called oxygen conformance, is reversible and not associated with long-term negative effects.11

A great number of studies have evaluated the lower limit of PaO2 at which healthy, acclimatized subjects can be exposed without evidence of neurocognitive impairment. In a recent study, arterial blood gases in climbers breathing ambient air on Mount Everest were evaluated, reporting values among the lowest ever documented in humans: at 8400 m, mean PaO2, mean PaCO2, and mean SaO2 were, respectively, 24.6 mm Hg, 13.3 mm Hg, and 54%.12

However, such reports are not applicable to critically ill patients, who might be not able to activate the necessary compensatory mechanisms, like increase in minute ventilation, CO, and hemoglobin levels.

The key question is to understand if hypoxemia can be accepted without fear of a worse prognosis that is to say if these oxygen values can be considered adequate to meet the patient’s needs.

To date, the pivotal question of tolerability of subnormal PaO2 in an ARDS patient, raised by the consensus conference in 1994, is still unanswered. Moreover, there is no data definitively demonstrating that higher PaO2 values are linked to better outcome.

Indeed, in the setting of ARDS patients, therapies associated with better outcome are not directly linked to improved oxygenation, as demonstrated by a review of 101 studies on this topic,13 in which the authors stated that PaO2/inspired fraction of oxygen (FiO2) index was not a reliable predictor of outcome.

A therapeutic strategy often used in ARDS patients is permissive hypoxemia that consists of the tolerance of values of SaO2 close to 80%, on the basis that normal oxygen saturation is not needed or impossible to achieve, recognizing that other variables of oxygen delivery, like CO or hemoglobin concentration, may be more effectively modified.

Moreover, employing an oxygenation target of 55–80 mm Hg, the potential oxidative stress and the harms of an aggressive management of mechanical ventilation are avoided.

Three main objections to this treatment strategy may be highlighted: 1) randomized controlled trials (RCTs) have never demonstrated a survival benefit in critically ill patients when the CO or hemoglobin concentration was modified to augment DO2 to supernormal values, compensating for the hypoxemia; 2) conflicting results in animal studies do not demonstrate that a constant oxygen delivery predicts a constant oxygen uptake14; 3) there are no definitive endpoints available to confirm the adequacy of tissue oxygen supply, above all this during VV ECMO, as SvO2 is not a reliable marker of global oxygen balance. Moreover, if permissive hypoxemia may be agreeable in a short-term prospective, the long-term prospective must prioritize care that results in optimal long-term result, i.e., in the light of positive neurologic outcomes.15

The burden of cognitive impairment in survivors from ARDS is high. Mikkelsen et al.16 in the ARDS cognitive outcomes study (ACSOS) assessed neuropsychological functions in a subset of survivors enrolled in Fluid and Catheter Treatment Trial (FACCT). They found a cognitive impairment in 41 of 75 (55%) survivors at 12 months. Lower PaO2 and enrollment in the conservative-fluid management strategy of the FACCT study were independently associated with cognitive impairment at 12 months. The difference in terms of PaO2 was 15 mm Hg (71 mm Hg in impaired patients and 86 mm Hg in not impaired patients, P = 0.02).

Already in 1999, the study of Hopkins et al.,17 evaluating neuropsychological sequelae in survivors of ARDS, enrolled 168 patients, 55 of which completed the follow up at 1 year. At the time of hospital discharge, all patients had a cognitive impairment. At 1 year, a significant improvement was observed, but 17 of 55 (30%) experienced deficit in cognitive function, and 43 of 55 (78%) had an impairment of at least one cognitive function, such as memory, attention, or concentration. There were significant correlations between the amount of time spent below normal values (<90%) and the impaired performance on intelligence, attention, speed of processing, visuospatial skills, and executive function test.

The authors, on the basis of correlation of hypoxemia and neurocognitive sequelae, hypothesized that the cerebral hypoxia might be a possible explanation of this phenomenon, not previously described in ARDS.

Cognitive impairment was correlated with neuroradiologic findings in another study conducted by the same author. Hopkins et al.18 found that in 15 ARDS patients, in comparison with matched control subjects, 53% had atrophy or lesions by radiological report. Even if a correlation between ventricular volumes was not established with duration or severity of hypoxemia, the authors hypothesized that the observed brain atrophy might be attributed to hypoxia, on the basis of similarity of the cerebral lesions to those reported in other pathologic states where hypoxia is a landmark (cardiac arrest, asthma, and monoxide poisoning).

In the light of these unsolved questions, Mikkelsen et al.19 have proposed a resetting of arterial oxygenation target in ARDS, balancing potential advantages and disadvantages of each strategy waiting for a more sound knowledge of long-term effects of oxygenation on short and long outcomes of ARDS patients.

Finally, it must to be underlined that the majority of medical literature on the topic of hypoxia is based on the concept of oxygen content and oxygen delivery, whose formulas are often shortened dropping the contribution of dissolved oxygen. This approach, however, is misleading, as tissues derive the means to support their metabolic aerobic activity from dissolved oxygen in the blood, and therefore, the contribution of dissolved O2 is never negligible and becomes particularly important when anemia is present.20

Pathophysiology of Hypoxemia During ARDS

A recent study, performed by Messaï et al.,21 identified the main causes of hypoxemia during VV ECMO in low oxygen partial pressure in blood leaving the oxygenator, high recirculation, elevated patient CO, and low mixed venous saturation. In their study, evaluating a new formula linking all these parameters for determining arterial oxygenation in ARDS patients treated with VV ECMO, the authors found a good correlation between predicted SaO2 and measured SaO2.

Excluding equipment failure, factors that contribute to hypoxemia during VV ECMO are 1) mixture between ECMO-oxygenated blood and patient’s own venous blood, 2) recirculation, 3) intrapulmonary shunt, and 4) flow exceeding oxygenator performances (not further treated because of the rarity of this scenario with the current technology).9

Mixture Between Blood Oxygenated by ECMO and Patient’s Own Venous Blood

During venovenous bypass, oxygenated blood from the infusion cannula reaches the venous circulation and mixes with the patient’s systemic venous return in the right atrium.

This situation implies that, assuming that there is no gas exchange across the lung, the oxygen saturation of the blood in right ventricle is equal to oxygen saturation of the arterial side of circulation; and, furthermore, that the use of mixed venous saturation as an index of systemic tissue oxygenation is misleading because the O2 saturation in pulmonary artery results from the mixture between oxygenated (from the reinfusion cannula) and deoxygenated blood (patient’s systemic venous return).22–24

Interactions between native CO and ECMO flow are well depicted in a recent study,25 elegantly demonstrating that, during VV ECMO for ARDS in patients whose native lung function was completely abolished, the factors determining the patient’s arterial oxygenation were VV ECMO blood flow and the inspired fraction of O2 of sweep gas (ECMO FiO2). Specifically, a ratio between VV ECMO flow and patient’s CO ≥ 60% was always associated with SaO2 > 90% and arterial PaO2 > 60 mm Hg. Moreover, a strong linear correlation was found between ECMO FiO2 and SaO2 and PaO2.

For these reasons, ELSO guidelines3 suggest setting a blood flow of 60–80 ml/kg for adults, which usually guarantees an oxygen delivery of 3 ml/kg/min. At these values of blood flow, the ratio between pump flow and deoxygenated arterial blood is usually 3:1.8

However, this ratio may be quite different if we look at the situation of a septic patient, in which CO is usually elevated, and the ratio with ECMO flow can be maintained only at expense of flow augmentation or CO manipulation.


Recirculation in VV ECMO is defined as the fraction of oxygenated blood that is infused into the right atrium and is then aspirated back into the venous line of the ECMO circuit.

A recent review on this topic identified in pump speed; blood flow rates; intrathoracic, intracardiac, and intra-abdominal pressure; cannula type, size, and position; and direction of extracorporeal blood flow the main factors to which recirculation is linked.26

Although there are many methods for the calculation of the recirculation fraction, like the use of indicator dilution technique or thermodilution or ultrasound-based technique, these are expensive or unavailable in many clinical settings; nevertheless, it is very important to have simple methods to exclude large fraction of recirculation.

A published formula27 is based on the blood oxygen saturation (SO2) of the drainage and the reinfusion cannulas and the pulmonary artery O2 saturation: R = (SO2preoxy − SvO2)/(SO2postoxy − SvO2), where SO2preoxy is the saturation of blood in the venous (entering the oxygenator) cannula, SO2postoxy is the saturation of blood in the arterial cannula (leaving the oxygenator), and SvO2 is the mixed venous saturation.

However, as already highlighted,28 the use of SvO2 during VV ECMO is misleading because the pulmonary artery saturation measures the mixing of deoxygenated blood of the native circulation and the oxygenated blood from the venous cannula of ECMO, and therefore, it is impossible to refer to SvO2 in conventional terms.22–24

One study29 suggests that it is possible to exclude at least significant recirculation volume if, by performing a conventional blood gas analysis from the drainage and the reinfusion cannulas, the pO2 of the drained blood is in the normal range (median 42 mm Hg) and below 10% of pO2 of the infusion cannula; moreover, a blood oxygen saturation in the drainage cannula below 75% may exclude meaningful recirculation.7

Intrapulmonary Shunt and Its Modulation

The main mechanism of hypoxemia during ARDS is intrapulmonary shunt30: this is linked to the pathological changes of the alveoli and the interstitium, with alveolar filling of protein rich fluid, red blood cells, and neutrophil infiltration31 (Figure 1).

Figure 1
Figure 1:
Chest X-ray showing a whiteout appearance of the lungs, with a femoro-jugular dual cannulas configuration for venovenous extracorporeal membrane oxygenation (VV ECMO).

According with Sangalli et al.,32 for the purposes of this review, we have defined intrapulmonary shunt as the extreme of ventilation/perfusion mismatching, in which an alveolus is perfused but not ventilated. Blood passing through such alveolar units reaches the left side of the heart without gas exchange. Intrapulmonary shunt or true shunt can be calculated using values from a pulmonary artery catheter, while breathing oxygen at 100%: Qs/Qt = 100 × (CcO2 – CaO2)/(CcO2 – CvO2), where CcO2, CaO2, and CvO2 are, respectively, the content of oxygen in pulmonary capillary, arterial, and venous blood.

In healthy subjects there is a certain amount of shunt, caused by the bronchial circulation and to the coronary venous blood drained directly into the left ventricle. This can be calculated from the same formula but at FiO2 lower than 100%, and encompasses three components: ventilation/perfusion mismatching, diffusion limitation, and true (intrapulmonary) shunt.

In ARDS, when the pulmonary function is completely abolished, we can assume that the shunt fraction is close to 100%.

The degree of intrapulmonary shunting is influenced by several factors, such as vascular pressures, vasoactive substances, and the degree of lung inflation.33 Several studies, however, have highlighted the importance of CO per se in determining the degree of shunt.

Lynch et al.34 have reported the results of pharmacologic and mechanical modulation of CO on intrapulmonary shunt: there is a significant linear correlation between CO and the fraction of intrapulmonary shunt.

In a series of ARDS patients, Dantzker et al.35 noted a strong correlation between CO reduction and intrapulmonary shunt reduction, when the CO was reduced by lowering of the venous return with incremental positive end-expiratory pressure (PEEP) or incremental tidal volumes. They also observed an improvement in arterial oxygenation; however, they highlighted that this strategy, while improving arterial oxygenation, may lead to reduced overall tissue oxygenation.

Even though data on arterial oxygenation are controversial, with nonhomogeneous findings in animal and human studies, modulation of CO can be a strategy to optimize peripheral oxygenation in VV ECMO, with two major objectives: ameliorate the balance between pump flow and patient’s native flow and reduce the degree of hypoxemia directly attributable to intrapulmonary shunt.

Strategies to Improve Peripheral Oxygenation

Available treatment strategies can be classified as follows:

  1. Strategies to increase the blood’s oxygen content:
    • i.increase of ECMO flow;
    • ii.increase of blood oxygen-carrying capacity.
  2. Strategies to reduce recirculation.
  3. Strategies to reduce oxygen consumption:
    • i.sedation and neuromuscular blockade;
    • ii.therapeutic hypothermia.
  4. Manipulation of CO and intrapulmonary shunt:
    • i.β-blockers infusion;
    • ii.prone positioning.
  5. Switch to venoarterial or a hybrid configuration.

Strategies Addressed to an Increase of Blood’s Oxygen Content

Increase of ECMO flow.

During VV ECMO for ARDS, it is a common practice8 to start with a flow of 60–80 ml/kg, which usually corresponds to an oxygen delivery of 3 ml/kg/min at CO values of 5 L/min and hemoglobin concentration of 15 g/dl.

The first action often undertaken in case of hypoxemia and pump flow below the maximum level allowed by the circuit is to increase the flow rate (assuming that the ECMO FiO2 is 1).

Indeed, the major determinants of oxygenation during extracorporeal support are ECMO flow and FiO2 of sweep gas.8,25

The possibility of developing a certain flow is determined by the characteristics of the vascular access, drainage tubing resistance, and pump properties.

Besides this, often the increase of ECMO flow does not improve the oxygenation because of the augmented recirculation fraction.

Higher flows also carry a greater risk of hemolysis. With current ECMO technology, this risk is related to the contemporary presence of these conditions: elevated RPM and occlusion of the venous line (clinically evident as venous line chattering), i.e., during coughing, kinking of the cannulas or hypovolemia, thereby determining an extreme negative pressure across the pump head. In these conditions, a vacuum is created in the pump head (because the pump continues to run and ejects the blood whereas the venous line is occluded), and it leads to cavitation and hemolysis.36

Increase of blood oxygen-carrying capacity.

Extracorporeal membrane oxygenation support is one of the most blood-expensive settings, partly owing to the high frequency of bleeding-related complications (in respiratory ECMO, the 2012 ELSO Registry reported bleeding from cannulas sites and surgical sites as a significant problem associated with worse prognosis).37

Besides bleeding, transfusion needs are because of the attempt to increase peripheral oxygen delivery in hypoxic patients with the increase of the blood’s oxygen-carrying capacity through administration of packed red blood cells.

Extracorporeal Life Support Organization guidelines, in fact, state that during ECLS it is mandatory to maintain an hemoglobin level of 12–14 g/dl and a normal hematocrit, owing to the fact that DO2 is determined by blood flow through the artificial lung, and if an anemic condition is present, a higher blood flow will be necessary to obtain the same level of oxygen delivery.

To optimize the risk–benefit ratio related to such a transfusional strategy, it is experts’ opinion that the need of higher blood flow, the associated risks of hemolysis, the reduction of blood volume, and the related administration of fluids are a much greater hazard than the transfusion-related risks.38

To date, there are no clinical prospective studies on the impact of restrictive transfusion protocols on ECMO patients for cardiac or respiratory failure, and there are also little data on transfusion strategies in adult patients receiving ECMO.

However, in the setting of ARDS, two studies39,40 have identified red blood cell transfusion as clinical predictor of mortality.

Strategies to Reduce Recirculation

There are various systems to diminish the recirculation fraction.

The most used technique is to maximize the distance between the two cannulas, thereby reducing the possibility that the blood leaving the reinfusion cannula reaches the drainage cannula.

Addressing cannulas configuration, one study41 has compared the femoroatrial (FA) configuration (femoral drainage and atrial reinfusion; Figure 2) with the atriofemoral (AF) configuration (atrial drainage and femoral reinfusion). Femoroatrial configuration provides higher maximal ECMO flow and higher pulmonary arterial mixed venous saturation and requires comparatively less flow to maintain an equivalent mixed venous oxygen saturation than the AF configuration.

Figure 2
Figure 2:
Venovenous extracorporeal membrane oxygenation (VV ECMO) with femoroatrial configuration for VV ECMO in an extubated patient.

Ichiba et al.42 described a three cannulas VV ECMO circuit composed of two drainage cannulas (one of these in the internal jugular vein and the other one in the left femoral vein) and one long return cannula in the right femoral vein, allowing an improved venous drainage.

Another study43 shows that the so-called χ-configuration—in which a multihole drainage cannula is positioned in the right atrium with the terminal segment just below the superior vena cava, and the infusion cannula customized in a manner to form an angle of 60° in its terminal segment, so as to be positioned in proximity to the tricuspid valve—allows near complete drainage of desaturated blood and a preferential inflow through the tricuspid valve, thereby reducing the recirculation and improving the patient’s oxygenation.

A new double lumen cannula (DLC) has recently been introduced, designed to be inserted through the internal jugular vein44 (Figures 3 and 4). The cannula has a drainage lumen whose tip ends, if properly inserted, in inferior vena cava (IVC). The drainage lumen drains blood from the superior and IVC through proximal and distal openings, respectively. The infusion lumen opens in the right atrium 10 cm from the cannula tip toward the tricuspid valve (Figure 5). In the animal experiments, this technique allows a very low recirculation fraction. Camboni et al.45 performed a comparison of flow characteristics of DLC with standard two vessels cannulation. The two vessel configuration allows higher flows, less negative pressures for drainage at a fixed flow because of the possibility to use larger cannulas.

Figure 3
Figure 3:
Detail of a double lumen cannula inserted percutaneously through the right internal jugular vein.
Figure 4
Figure 4:
Venovenous extracorporeal membrane oxygenation (VV ECMO) with double lumen cannula in a patient suffering from acute respiratory distress syndrome (ARDS) because ofNeisseria meningitidis purpura fulminans.
Figure 5
Figure 5:
Drawing showing the correct position of a double lumen cannula (DLC) with the infusion port located toward the tricuspid valve and the draining holes located in the superior and inferior vena cava.

Double lumen cannula can offer an alternative to two site cannulation in the case of prolonged support, allowing a more safe mobilization of the patient, or in the evenience of unfavorable groin anatomy.

Strategies to Reduce the Oxygen Consumption

Sedation and neuromuscular blockade.

Neuromuscular blocking drugs (NMBDs) administration can have multiple beneficial effects on ARDS patients, but, while some of these have been demonstrated, others remain matter of speculation.

The ARDS and Curarisation Systématique (ACURASYS) study reported that in patients with severe ARDS early administration of a NMBD improved the adjusted 90-day survival and increased the time off of the ventilator without increasing muscle weakness.46

The effects of NMBDs on mortality in ARDS are the object of a recent review47 in which the authors concluded that a short-term cysatracurium besylate infusion is associated with a reduced 28-day, intensive care unit (ICU) and in-hospital mortality, without increasing the risk of ICU-acquired muscular weakness.

During VV ECMO, deep sedation and NMBDs are almost always used during the early phase of support, to aid cannulation and cardiorespiratory stabilization.

In the context of hypoxemia volume-controlled ventilation, deep sedation and neuromuscular blockade are established. On the basis of the aforementioned studies, it is conceivable that the suppression of breathing work during a phase of refractory hypoxemia can reduce the systemic oxygen needs, but there is no data about the direct contribution of such maneuvers to hypoxemia during VV ECMO.

Therapeutic hypothermia.

The term therapeutic hypothermia (TH) refers to deliberate lowering of body temperature to 36–25°C and can be classified as mild (34–35.9°C), moderate (32–33.9°C), moderate-deep (30–31.9°C), and deep (<30°C) hypothermia.48

The physiologic response to TH and its protective effects are very complex and are not just related to the achieved temperature, but also to the speed of induction, duration, speed of rewarming, and prevention of side effects.

For the purpose of this review, we will only discuss the effects and application of mild TH.

To date, there are no studies on TH during VV ECMO in ARDS.

The main method to achieve TH during ECMO is the manipulation of the heat exchanger.

Many mechanisms can affect the oxygen transport and its peripheral utilization during TH: metabolic rate (MR), O2 solubility, acid–base status, oxygen–hemoglobin dissociation, CO or regional blood flow, and hemoglobin concentration.49

In many biological systems, there is a reduction of 50% of MR for 10°C reduction.

Hypothermia can also affect the solubility of O2 in blood and water: at a given partial pressure, hypothermia augments the amount of dissolved O2. This phenomenon is very interesting in the ARDS setting when dealing with very low values of PaO2. CO is globally reduced, but the contemporary drop of oxygen consumption (VO2) causes the VO2/DO2 ratio to remain the same.

A common side effect of TH, shivering, determines a large increase in MR but can be effectively controlled with many drugs.

The use of TH in ARDS is object of anecdotal experiences: Villar and Slutsky50 assigned 19 ARDS and septic patients to receive conventional treatment or conventional treatment plus hypothermia as a last resort. They found a significant improvement in oxygenation, in the presence of unchanged VO2 and increased O2 extraction ratio, and increased survival in the hypothermia-treated patients.

In conclusion, on the basis of available literature, the cooling of a hypoxemic patient during VV ECMO can be considered as a possible strategy to reduce the oxygen consumption and to improve, in doing so, the DO2/VO2 ratio.

Manipulation of CO and intrapulmonary shunt

β-Blocker infusion.

An emerging technique to treat refractory hypoxemia in ARDS patients during VV ECMO is esmolol administration.

Esmolol is a cardioselective β1-blocker agent, characterized by ultrashort half-life that allows for quickly reversible modulation of its pharmacologic effects.

When esmolol is administered as a bolus followed by continuous infusion, the onset of activity occurs within 2 minutes, with 90% of β-blockade at 5 minutes. Full recovery from β-blockade takes 18–30 minutes after stopping the infusion.51

A recent case series of three ARDS patients treated with esmolol for refractory hypoxemia during VV ECMO revealed the feasibility of this strategy to augment the PaO2 reducing CO and thereby ameliorating the match CO/Pump flow (PF), without a significant reduction of oxygen delivery.9

More specifically, three patients with ARDS and refractory hypoxemia despite protective mechanical ventilation and high flows of VV ECMO, with an hemodynamic profile characterized by high CO (>7 L/min), were treated with a continuous infusion of esmolol at a dosage of 50–80 mcg/kg/min. Hemodynamic evaluation during the first 12 hours demonstrated a significant reduction of CO and heart rate (HR) and a significant improvement of peripheral oxygenation, as demonstrated by the increase of PaO2 from the radial artery samples.

Furthermore, calculated DO2 did not significantly vary during the treatment. The absence of metabolic acidosis and the decreasing trend of blood lactates during the treatment further confirm the absence of peripheral hypoperfusion because of reduced CO.

A possible objection could be that reduction of CO might jeopardize peripheral oxygen delivery. However, this technique does not entail a low CO state, which should be strictly monitored by means of markers of tissue hypoxia, such as lactates or metabolic acidosis.

On the pathophysiological side, the most prominent effect of esmolol infusion is the reduction of HR and CO in patients characterized by tachycardia and high CO states, like ARDS septic patients.

By reducing CO, esmolol produces two main effects, namely improves the ratio between ECMO flow and patient’s native flow and, in doing so, prevents the need of dangerous ECMO flow increases; moreover, as stated above, the reduction of CO is a well-known mechanism of intrapulmonary shunt reduction that is a paramount mechanism of hypoxemia in patients with ARDS.

However, β-blocker administration may have other advantages.

Myocardial dysfunction is often present in the shocked septic patient, and various studies reported high incidence of reduced left ventricular ejection fraction, ranging from 24% to 50% of affected patients,52 and there are also consistent findings regarding impaired left ventricular diastolic function and right ventricular dysfunction.53

During septic shock, alterations of β-adrenergic signaling altered oxygen use in cardiomyocites, because of macrocirculatory and microcirculatory changes, disoxya, because of altered mitochondrial function and direct effects on cardiac performance, because of circulating myocardial depressant factor like cytokines and bacterial endotoxins may all play a role in the etiology of cardiac dysfunction.

Recently, Macchia et al.54 have conducted a pharmacoepidemiologic study, evaluating whether septic patients taking chronically β-blocker therapy had a different mortality rate than those who did not receive chronic treatment.

They found lower mortality rate at 28 days in patients previously taking β-blockers admitted to ICU for sepsis and developing organ dysfunction.

Interestingly, Morelli et al.55 have recently reported that, in septic patients with a HR of 95/min or higher, requiring high-dose norepinephrine to maintain a mean arterial pressure of 65 mm Hg or higher, treated with esmolol versus standard care to achieve a target value of HR between 80 and 94 beats per minute, there was a strong benefit in mortality.

However, although increasing evidence suggests the possible beneficial role of β-blockers in sepsis, the topic is still controversial.

Prone positioning.

Prone positioning is a well-known system to improve oxygenation in patients under mechanical ventilation suffering from ARDS.

The physiologic effects of prone positioning are attributable to changes in the stiffness of the whole chest wall; redistribution of regional alveolar inflation from ventral dependent regions to dorsal nondependent regions; relief from the heart pressure on the lungs; changes of hypoxic vasoconstriction; and reduction of ventilator induced lung injury (VILI).56,57

Very recently, a multicenter RCT assigned 466 patients with severe ARDS to prone-positioning session of at least 16 hours daily, or supine positioning. The prone group had a significant lower 28-day and 90-day mortality, with an incidence of complications that did not differ between the two groups, except for cardiac arrest, more frequent in the supine-positioned patients.58

However, the pronation of ECMO patients poses major dilemmas regard as catastrophic complications as cannulas dislodgement or pump failure.

Very recently, Guervilly et al.59 reported the largest series of patients turned to prone positioning during VV ECMO therapy. Fifteen ARDS patients were turned into prone position, without major complications, if they had severe hypoxemia (PaO2/FiO2 ratio below 70) despite maximal oxygenation, injurious ventilation parameters with plateau pressure exceeding 32 cm H2O, or failure of attempt to wean ECMO after at least 10 days on ECMO support. The main findings of the study are significant improvement in PaO2/FiO2 ratio at 6 hours (P = 0.03) and 12 hours (P = 0.007) after reversal. The improvement in oxygenation persisted 1 hour (P = 0.017) and 6 hours (P = 0.013) after being turned back to the supine position.

Before this experience, the combined treatment of VV ECMO and prone positioning has been reported only in small case series and in a retrospective study involving few patients, without control group.60,61

A retrospective cohort of 9 ARDS patients treated with VV ECMO and positioning therapy was published.62 The patients received a median of 20 hours of positioning therapy during ECMO course. An improvement in oxygenation and lung compliance was observed after 72 hours from initiation of positioning therapy, with no differences in outcome.

These limited experiences does not allow inference about the effects on the outcome of patients treated with VV ECMO and positioning therapy, but there is convincing evidence about the feasibility and safety of such a strategy in case of severely hypoxemic patients, provided by trained teams in patient prone positioning (Figure 6).

Figure 6
Figure 6:
Prone positioning during venovenous extracorporeal membrane oxygenation (VV ECMO).

Transition to VA ECMO

There is a paucity of data addressing this specific issue.

In ARDS, we can consider switching to VA ECMO as a consequence of two main conditions:

  1. refractory hypoxemia, when all the strategies listed above have failed, but the cardiac function is normal;
  2. when a cardiac dysfunction supervenes in terms of septic myocardial dysfunction or right ventricular dysfunction (acute cor pulmonale) during respiratory failure.

Switching to a VA ECMO circuit allows for supernormal PaO2 on the arterial side as the native lung is completely bypassed and also provides hemodynamic support in case of failing left and right ventricle. However, this is not straightforward in acute lung failure. As part of cannulation technique, if the lung is failing, arterial cannulation should be as close as possible to the heart to avoid “Harlequin syndrome,” in which the patients appear with a blue head and red legs. This is because of the competition between the anterograde blood flow related to native CO and ECMO flow delivered by a femoral cannula and may compromise an adequate perfusion of the upper body.

Femoral cannulation should be avoided unless using a long cannula to reach the aortic arch, entailing a high risk of hemolysis. Appropriate strategies include axillary artery or central cannulation through the chest. Axillary artery cannulation has gained widespread application for aortic surgery, but it has many limitations for ECMO: upper limb ischemia, edema, bleeding, and infection (Figure 7). Javidfar et al.63 reported good results in terms of adequate oxygenation, sufficient ventricular unloading, and low complication rate with subclavian artery cannulation. Moreover, as the heart is normal, it will continue to beat even if fully drained from the circuit, and therefore, deoxygenated blood will be ejected into the coronary arteries and brain. Also, as the right upper arm is perfused by the ECMO, SpO2 monitoring will prove unreliable.

Figure 7
Figure 7:
Axillary artery cannulation complicated by right-hand ischemia treated by distal reperfusion.

However, the need to switch to a VA configuration may be suggested by an ongoing hemodynamic instability because of right ventricular dysfunction. Recently, Vieillard-Baron et al.64 reported the results of a series of studies addressing right ventricular failure in ARDS patients. In the era of protective ventilation, Boissier et al.65 reported an incidence of 22% of acute cor pulmonale (ACP) in patients who met the criteria of the Berlin definition for moderate to severe ARDS. Hemodynamic consequences of ACP included tachycardia, hypotension, shock, and the need for hemodynamic support.65 In such a scenario, only VA ECMO may provide adequate support. The general risks of VA versus VV ECMO must be taken into account, with the VA configuration usually being associated with more frequent and severe complications. Cheng et al.66 recently published a meta-analysis including 1866 patients treated with VA ECMO, reporting a cumulative rate of neurologic complications associated to hemorrhagic or ischemic stroke of 13.3%, a cumulative rate of major or significant bleeding up to 40.8%, and a pooled estimate rate of vascular complication of 16.9%.

Finally, as reported by the experience of Biscotti et al.,67 the consideration of an hybrid configuration of the circuit may overcome both the limitations in hemodynamic support linked to VV ECMO and in oxygenation of the upper body linked to VA ECMO. For the purpose of this review, the discussion will be limited to venous-venoarterial (V-VA) ECMO. Although the concept of reinfusing blood from both the arterial and venous side is conceptually appealing, it is, however, technically demanding as pressures, and therefore, flows through the circuit cannot be effectively controlled; it requires continuous attendance of a perfusionist and does not provide any consistent benefit, as arterial cannulation is required (Figure 8).

Figure 8
Figure 8:
Hybrid configuration.

In clinical practice, transition to VA ECMO in patients with ARDS should, therefore, encompass a multifaceted evaluation addressing not only arterial oxygen tension, but also a thorough evaluation of the risks involved with arterial cannulation.


Management of refractory hypoxemia during VV ECMO for ARDS poses a great number of questions that still remain unanswered by current knowledge.

If some of the strategies proposed here are supported by sound experimental and clinical evidence, like the determinants of patient’s oxygenation during VV ECMO or the transfusion management, others, like therapeutic hypothermia and prone positioning, still need validation in larger studies.

Among promising approaches, β-blockade seems to be of utmost interest, on the basis of pleiotropic beneficial effects that β-blockers appear to have in septic shock, but these findings need to be evaluated in adequately powered studies.

However, a critical topic is still waiting for response: paradoxically, even if hypoxemia is the trigger for starting extracorporeal life support in ARDS patients, little is known about the limits of tolerability of sustained hypoxia in critically ill patients, especially in a long-term perspective.

Therefore, despite the improved results guaranteed by the use of extracorporeal support, many issues in this field require a thorough focus, which might translate into further improvements.

Author Contributions

A.M. and G.M. contributed toward conception of the study, literature search, and writing up the first draft of the paper; A.Z. and D.W. contributed toward critical revision of the paper; and F.P. contributed toward conception of the study, critical revision, and final approval of the paper.


1. Ranieri VM, Rubenfeld GD, Thompson BT, et al.ARDS Definition Task Force. Acute respiratory distress syndrome: the Berlin Definition. JAMA. 2012;307:2526–2533
2. MacLaren G, Combes A, Bartlett RH. Contemporary extracorporeal membrane oxygenation for adult respiratory failure: life support in the new era. Intensive Care Med. 2012;38:210–220
3. Agerstrand CL, Bacchetta MD, Brodie D. ECMO for adult respiratory failure: current use and evolving applications. ASAIO J. 2014;60:255–262
4. Peek GJ, Mugford M, Tiruvoipati R, et al.CESAR trial collaboration. Efficacy and economic assessment of conventional ventilatory support versus extracorporeal membrane oxygenation for severe adult respiratory failure (CESAR): a multicentre randomised controlled trial. Lancet. 2009;374:1351–1363
5. Davies A, Jones D, Bailey M, et al. Extracorporeal membrane oxygenation for 2009 influenza A (H1N1) acute respiratory distress syndrome. JAMA. 2009;302:1888–1895
6. Davies A, Jones D, Gattas D. Extracorporeal membrane oxygenation for ARDS due to 2009 influenza A (H1N1) – Reply. JAMA. 2010;303:942
7. Extracorporeal Life Support Organization. ELSO Guidelines for Cardiopulmonary Extracorporeal Life Support. 2009Version 1.1 Ann Arbor, MI April
8. Zwischenberger JB, Bartlett RH. Management of blood flow and gas exchange during ECLS ECMO Extracorporeal Cardiopulmonary Support in Critical Care. 20124th ed Ann Arbor, MI ELSO:149–156
9. Guarracino F, Zangrillo A, Ruggeri L, et al. β-Blockers to optimize peripheral oxygenation during extracorporeal membrane oxygenation: a case series. J Cardiothorac Vasc Anesth. 2012;26:58–63
10. Schmidt M, Pellegrino V, Combes A, Scheinkestel C, Cooper DJ, Hodgson C.. Mechanical ventilation during extracorporeal membrane oxygenation. Crit Care. 2014;18:203
11. Schumacker PT, Chandel N, Agusti AG. Oxygen conformance of cellular respiration in hepatocytes. Am J Physiol. 1993;265(4 Pt 1):L395–L402
12. Grocott MP, Martin DS, Levett DZ, McMorrow R, Windsor J, Montgomery HECaudwell Xtreme Everest Research Group. . Arterial blood gases and oxygen content in climbers on Mount Everest. N Engl J Med. 2009;360:140–149
13. Krafft P, Fridrich P, Pernerstorfer T, et al. The acute respiratory distress syndrome: definitions, severity and clinical outcome. An analysis of 101 clinical investigations. Intensive Care Med. 1996;22:519–529
14. Abdelsalam M, Cheifetz IM. Goal-directed therapy for severely hypoxic patients with acute respiratory distress syndrome: permissive hypoxemia. Respir Care. 2010;55:1483–1490
15. Mikkelsen ME, Anderson B, Christie JD, Hopkins RO, Lanken PN. Can we optimize long-term outcomes in acute respiratory distress syndrome by targeting normoxemia? Ann Am Thorac Soc. 2014;11:613–618
16. Mikkelsen ME, Christie JD, Lanken PN, et al. The adult respiratory distress syndrome cognitive outcomes study: long-term neuropsychological function in survivors of acute lung injury. Am J Respir Crit Care Med. 2012;185:1307–1315
17. Hopkins RO, Weaver LK, Pope D, Orme JF, Bigler ED, Larson-LOHR V. Neuropsychological sequelae and impaired health status in survivors of severe acute respiratory distress syndrome. Am J Respir Crit Care Med. 1999;160:50–56
18. Hopkins RO, Gale SD, Weaver LK. Brain atrophy and cognitive impairment in survivors of acute respiratory distress syndrome. Brain Inj. 2006;20:263–271
19. Mikkelsen ME, Anderson B, Christie JD, Hopkins RO, Lanken PN. Can we optimize long-term outcomes in acute respiratory distress syndrome by targeting normoxemia? Ann Am Thorac Soc. 2014;11:613–618
20. Spiess BD. Perfluorocarbon emulsions as a promising technology: a review of tissue and vascular gas dynamics. J Appl Physiol (1985). 2009;106:1444–1452
21. Messaï E, Bouguerra A, Harmelin G, Di Lascio G, Cianchi G, Bonacchi M. A new formula for determining arterial oxygen saturation during venovenous extracorporeal oxygenation. Intensive Care Med. 2013;39:327–334
22. Sidebotham D, Allen SJ, McGeorge A, Ibbott N, Willcox T. Venovenous extracorporeal membrane oxygenation in adults: practical aspects of circuits, cannulae, and procedures. J Cardiothorac Vasc Anesth. 2012;26:893–909
23. Sidebotham D, McGeorge A, McGuinness S, Edwards M, Willcox T, Beca J. Extracorporeal membrane oxygenation for treating severe cardiac and respiratory disease in adults: part 1 – overview of extracorporeal membrane oxygenation. J Cardiothorac Vasc Anesth. 2009;23:886–892
24. Sidebotham D, McGeorge A, McGuinness S, Edwards M, Willcox T, Beca J. Extracorporeal membrane oxygenation for treating severe cardiac and respiratory failure in adults: part 2-technical considerations. J Cardiothorac Vasc Anesth. 2010;24:164–172
25. Schmidt M, Tachon G, Devilliers C, et al. Blood oxygenation and decarboxylation determinants during venovenous ECMO for respiratory failure in adults. Intensive Care Med. 2013;39:838–846
26. Abrams D, Bacchetta M, Brodie D. Recirculation in venovenous extracorporeal membrane oxygenation. ASAIO Journal. 2014 [Epub ahead of print]
27. Fortenberry JD, Pettignano R, Dykes F. Principles and practice of venovenous ECMO. ECMO Extracorporeal Cardiopulmonary Support in Critical Care. 20053rd ed. Ann Arbor, MI Extracorporeal Life Support Organization:94
28. Walker JL, Gelfond J, Zarzabal LA, Darling E. Calculating mixed venous saturation during veno-venous extracorporeal membrane oxygenation. Perfusion. 2009;24:333–339
29. Locker GJ, Losert H, Schellongowski P, et al. Bedside exclusion of clinically significant recirculation volume during venovenous ECMO using conventional blood gas analyses. J Clin Anesth. 2003;15:441–445
30. Rodríguez-Roisin R, Roca J. Mechanisms of hypoxemia. Intensive Care Med. 2005;31:1017–1019
31. Bachofen M, Weibel ER. Alterations of the gas exchange apparatus in adult respiratory insufficiency associated with septicemia. Am Rev Respir Dis. 1977;116:589–615
32. Sangalli F, Patroniti N, Pesenti ASangalli. Basic aspects of physiology during ECMO support ECMO – Extracorporeal Life Support in Adults. 2014 Springer:19–36
33. West JB. Blood flow to the lung and gas exchange. Anesthesiology. 1974;41:124–138
34. Lynch JP, Mhyre JG, Dantzker DR. Influence of cardiac output on intrapulmonary shunt. J Appl Physiol Respir Environ Exerc Physiol. 1979;46:315–321
35. Dantzker DR, Lynch JP, Weg JG. Depression of cardiac output is a mechanism of shunt reduction in the therapy of acute respiratory failure. Chest. 1980;77:636–642
36. Toomasian JM, Bartlett RH. Hemolysis and ECMO pumps in the 21st Century. Perfusion. 2011;26:5–6
37. Paden ML, Conrad SA, Rycus PT, Thiagarajan RR. ELSO registry: extracorporeal life support organization registry report 2012. ASAIO J. 2013;59:202–10
38. Bartlett RHAnnich GM, Lynch WR, MacLaren G, Wilson JM, Bartlett RH. Physiology of extracorporeal life support ECMO Extracorporeal Cardiopulmonary Support in Critical Care. 20124th ed Ann Arbor, MI ELSO:11–32
39. Netzer G, Shah CV, Iwashyna TJ, et al. Association of RBC transfusion with mortality in patients with acute lung injury. Chest. 2007;132:1116–1123
40. Gong MN, Thompson BT, Williams P, Pothier L, Boyce PD, Christiani DC. Clinical predictors of and mortality in acute respiratory distress syndrome: potential role of red cell transfusion. Crit Care Med. 2005;33:1191–1198
41. Rich PB, Awad SS, Crotti S, Hirschl RB, Bartlett RH, Schreiner RJ. A prospective comparison of atrio-femoral and femoro-atrial flow in adult venovenous extracorporeal life support. J Thorac Cardiovasc Surg. 1998;116:628–632
42. Ichiba S, Peek GJ, Sosnowski AW, Brennan KJ, Firmin RK. Modifying a venovenous extracorporeal membrane oxygenation circuit to reduce recirculation. Ann Thorac Surg. 2000;69:298–299
43. Bonacchi M, Harmelin G, Peris A, Sani G. A novel strategy to improve systemic oxygenation in venovenous extracorporeal membrane oxygenation: the “χ-configuration”. J Thorac Cardiovasc Surg. 2011;142:1197–1204
44. Bermudez CA, Rocha RV, Sappington PL, Toyoda Y, Murray HN, Boujoukos AJ. Initial experience with single cannulation for venovenous extracorporeal oxygenation in adults. Ann Thorac Surg. 2010;90:991–995
45. Camboni D, Philipp A, Lubnow M, et al. Extracorporeal membrane oxygenation by single-vessel access in adults: advantages and limitations. ASAIO J. 2012;58:616–621
46. Papazian L, Forel JM, Gacouin A, et al.ACURASYS Study Investigators. Neuromuscular blockers in early acute respiratory distress syndrome. N Engl J Med. 2010;363:1107–1116
47. Alhazzani W, Alshahrani M, Jaeschke R, et al. Neuromuscular blocking agents in acute respiratory distress syndrome: a systematic review and meta-analysis of randomized controlled trials. Crit Care. 2013;17:R43
48. Polderman KH. Mechanisms of action, physiological effects, and complications of hypothermia. Crit Care Med. 2009;37:S186–202
49. Villar J, Espinosa STisherman AS, Stertz F. Adult respiratory distress syndrome and sepsis Therapeutic Hypothermia. 2005 New York Springer:169–178
50. Villar J, Slutsky AS. Effects of induced hypothermia in patients with septic adult respiratory distress syndrome. Resuscitation. 1993;26:183–192
51. Wiest DB, Haney JS. Clinical pharmacokinetics and therapeutic efficacy of esmolol. Clin Pharmacokinet. 2012;51:347–356
52. Rudiger A, Singer M. The heart in sepsis: from basic mechanisms to clinical management. Curr Vasc Pharmacol. 2013;11:187–195
53. Hunter JD, Doddi M. Sepsis and the heart. Br J Anaesth. 2010;104:3–11
54. Macchia A, Romero M, Comignani PD, et al. Previous prescription of β-blockers is associated with reduced mortality among patients hospitalized in intensive care units for sepsis. Crit Care Med. 2012;40:2768–2772
55. Morelli A, Ertmer C, Westphal M, et al. Effect of heart rate control with esmolol on hemodynamic and clinical outcomes in patients with septic shock: a randomized clinical trial. JAMA. 2013;310:1683–1691
56. Gattinoni L, Carlesso E, Taccone P, Polli F, Guérin C, Mancebo J. Prone positioning improves survival in severe ARDS: a pathophysiologic review and individual patient meta-analysis. Minerva Anestesiol. 2010;76:448–454
57. Dickinson S, Park PK, Napolitano LM. Prone-positioning therapy in ARDS. Crit Care Clin. 2011;27:511–523
58. Guérin C, Reignier J, Richard JC, et al.PROSEVA Study Group. Prone positioning in severe acute respiratory distress syndrome. N Engl J Med. 2013;368:2159–2168
59. Guervilly C, Hraiech S, Gariboldi V, et al. Prone positioning during veno-venous extracorporeal membrane oxygenation for severe acute respiratory distress syndrome in adults. Minerva Anestesiol. 2014;80:307–313
60. Otterspoor LC, Smit FH, van Laar TJ, Kesecioglu J, van Dijk D. Prolonged use of extracorporeal membrane oxygenation combined with prone positioning in patients with acute respiratory distress syndrome and invasive Aspergillosis. Perfusion. 2012;27:335–337
61. Litmathe J, Sucker C, Easo J, Wigger L, Dapunt O. Prone and ECMO—a contradiction per se? Perfusion. 2012;27:78–82
62. Kredel M, Bischof L, Wurmb TE, Roewer N, Muellenbach RM. Combination of positioning therapy and venovenous extracorporeal membrane oxygenation in ARDS patients. Perfusion. 2014;29:171–177
63. Javidfar J, Brodie D, Costa J, et al. Subclavian artery cannulation for venoarterial extracorporeal membrane oxygenation. ASAIO J. 2012;58:494–498
64. Vieillard-Baron A, Price LC, Matthay MA. Acute cor pulmonale in ARDS. Intensive Care Med. 2013;39:1836–1838
65. Boissier F, Katsahian S, Razazi K, et al. Prevalence and prognosis of cor pulmonale during protective ventilation for acute respiratory distress syndrome. Intensive Care Med. 2013;39:1725–1733
66. Cheng R, Hachamovitch R, Kittleson M, et al. Complications of extracorporeal membrane oxygenation for treatment of cardiogenic shock and cardiac arrest: a meta-analysis of 1,866 adult patients. Ann Thorac Surg. 2014;97:610–616
67. Biscotti M, Lee A, Basner RC, Agerstrand C, Abrams D, Brodie D, et al. Hybrid configuration via percutaneous access for extracorporeal membrane oxygenation: a single center experience. ASAIO J. 2014;60:635–642

ARDS; ECMO; Hypoxemia

Copyright © 2015 by the American Society for Artificial Internal Organs