Optimizing Extracorporeal Membrane Oxygenation Gas Exchange: Key Insights for Clinical Management

Abstract

This review examines the role of extracorporeal membrane oxygenation (ECMO) in the management of severe acute respiratory distress syndrome (ARDS), emphasizing its contribution to lung-protective ventilation through optimizing oxygenation and ensuring optimal decarboxylation. Key determinants of oxygen delivery during ECMO include circuit blood flow, cannula size and positioning, and hemoglobin concentration. Strategies for troubleshooting oxygenation issues are reviewed, including recirculation, increased oxygen consumption, and oxygenator dysfunction. In contrast, carbon dioxide removal (decarboxylation), which ECMO circuits efficiently achieve, is primarily influenced by sweep gas flow and the patient’s systemic PaCO₂. Effective management of these factors is crucial to ensure optimal ECMO support, enable ultra-protective lung ventilation, and improve outcomes in critically ill patients with severe ARDS.

Introduction

Acute respiratory distress syndrome (ARDS) is a life-threatening condition that is characterized by hypoxemic respiratory failure [1]. While mechanical ventilation serves as the cornerstone of supportive care, severe cases may necessitate high pressures and tidal volumes, which increase the risk of ventilator-induced lung injury (VILI) [2,3]. This emphasizes the need for alternative strategies that ensure gas exchange while minimizing pulmonary damage [4,5]. Veno–venous extracorporeal membrane oxygenation (V-V ECMO) serves a critical function in managing severe ARDS by facilitating gas exchange, specifically oxygenation and carbon dioxide removal, via an extracorporeal circuit. This support enables the use of ultra-protective ventilation strategies, thereby minimizing VILI [2,5,6]. However, in certain patients, V-V ECMO may not achieve sufficient oxygenation (e.g., SpO₂ <90% or inadequate oxygen delivery) or effective decarboxylation (PaCO₂ ≥45 mm Hg), necessitating continued reliance on aggressive mechanical ventilation [2,7,8]. This undermines one of the primary benefits of ECMO: permitting lung rest. Consequently, optimizing ECMO efficiency in oxygen and carbon dioxide exchange is essential to reduce VILI and improve clinical outcomes in patients with severe ARDS.

This review discusses the physiological principles of ECMO oxygenation and carbon dioxide removal, strategies to enhance its effectiveness, and troubleshooting approaches to address oxygenation challenges during ECMO support.

Mechanism of V-V ECMO

V-V ECMO facilitates gas exchange by withdrawing deoxygenated blood from the venous circulation, circulating it through an extracorporeal circuit, and reinfusing oxygenated, decarboxylated blood into the venous system [8]. The circuit comprises three key components: (1) a centrifugal pump, (2) a hollow-fiber membrane oxygenator, and (3) an oxygen–air blender at 1 to 14 L/min for gas exchange. Blood flows outside the fibers while sweep gas circulates, enabling oxygen diffusion into the blood and CO₂ removal. Typically, drainage occurs from the right atrium (RA) or inferior vena cava (IVC) via a multiperforated cannula. In the femoral jugular configuration, blood is drained from the RA and reinfused into the superior vena cava (SVC; via the right internal jugular vein). Drainage is obtained from the IVC in the femoral–femoral configuration (Fem–Fem), while reinfusion occurs into the RA.

Alternatively, a dual-lumen cannula offers a single-site approach for V-V ECMO [9]. The Avalon Elite Bi-Caval Dual-Lumen Catheter is inserted via the right internal jugular vein, and enables simultaneous drainage from both the SVC and IVC, and reinfusion into the RA, thereby minimizing recirculation. The ProtekDuo cannula, also inserted through the right internal jugular vein, features a distal tip positioned in the pulmonary artery, providing both oxygenation and right ventricular support [9,10]. However, since both the Avalon and Protek-Duo cannulas require placement via the internal jugular vein, which has a smaller diameter than the femoral vein, the maximum achievable drainage flow is generally lower at 3 to 5 L/min, compared to dual cannula configurations. This may limit oxygenation in patients with high cardiac output (CO) or those with minimal native lung function, thereby affecting the suitability of dual-lumen cannulas during ECMO initiation.

Ultra-Protective Lung Ventilation

Extracorporeal oxygenation via V-V ECMO facilitates complete oxygenation and carbon dioxide removal, while supplementing invasive mechanical ventilation. This enables substantial reductions in ventilator settings, thereby attenuating mechanical stress on the lungs, and supporting lung-protective ventilation strategies. Decreasing ventilatory support lowers mechanical power delivery, mitigating VILI and promoting pulmonary recovery [11-13]. As V-V ECMO fully supports gas exchange, the intensity of mechanical ventilation can be markedly reduced, helping to prevent further lung damage. In this context, international guidelines and expert consensus advocate ultra-lung-protective ventilation strategies that are characterized by substantial reductions in tidal volume (<4 mL/kg predicted body weight), respiratory rate of 10 to 20 breaths/min, plateau pressure (<25 cmH₂O), and driving pressure (<15 cmH₂O).

However, the optimal level of ventilatory support reduction, including the potential use of apneic ventilation and the best positive end expiratory pressure strategy, remains subject to ongoing debate. Ultra-protective ventilation has been associated with improved survival outcomes in randomized controlled trials and meta-analyses [2,7,14].

Optimizing Oxygenation in V-V ECMO

The conventional ventilatory support versus extracorporeal membrane oxygenation for severe adult respiratory failure (CESAR) and ECMO to Rescue Lung Injury in Severe ARDS (EOLIA) trials, the two landmark randomized controlled trials in the field, both advocated targeting a PaO₂ between 65 and 90 mm Hg or an arterial oxygen saturation (SaO₂) >90% by adjusting the ECMO flow and FmO₂ [2,7]. Oxygenation in V-V ECMO is primarily influenced by three crucial factors: the efficiency of the membrane oxygenator (which depends on its surface area and oxygen diffusion capacity), hemoglobin concentration, and ECMO blood flow (Q_ECMO) (Table 1).

Table 1.

Key factors influencing oxygenation on venovenous ECMO

Factor	Description
Oxygenator function	The membrane oxygenator's surface area and material properties influence gas exchange. Over time, performance may decline due to clot formation, which makes monitoring circuit pressure essential.
Hemoglobin concentration	Affects the oxygen-carrying capacity of blood. Increased hemoglobin raises CaO₂, thereby enhancing DO₂ to tissues.
QECMO	The volume of blood flowing through the ECMO circuit is the most significant modifiable factor affecting oxygenation. To maintain sufficient systemic oxygenation, QECMO should surpass 60% of cardiac output, usually necessitating flows of at least 3 L/min.
Cannula size	Larger cannulas significantly improve flow according to the Hagen–Poiseuille equation.
Cannula position	Positioning the tip of the drainage cannula in the right atrium enables more stable and effective drainage, supporting consistent QECMO.
Venous return	Venous return is influenced by the patient’s volume status and intrathoracic pressure, which can impede flow into the drainage cannula. Hypovolemia or increased intrathoracic pressure—such as in pneumothorax or cardiac tamponade—can greatly diminish venous return.

ECMO: extracorporeal membrane oxygenation; CaO₂: arterial oxygen content; DO₂: oxygen delivery; Q_ECMO: ECMO blood flow.

1. Size of the membrane oxygenator

Each oxygenator has a maximum flow capacity that determines oxygen delivery to the blood. For example, with a hemoglobin level of 12 g/dL and a venous oxygen saturation of 75%, a flow rate of 6 L/min ensures that post-oxygenator saturation remains above 95%. However, exceeding the threshold Q_ECMO rate can lead to inadequate oxygenation and cause desaturation. The amount of oxygen added at a flow rate of 6 L/min is calculated as:

(0.95−0.75)×12 g/dL×1.36×1.06 L/min=196 mL/min.

If oxygen consumption (VO₂) is approximately 250 mL/min with no native lung function, hypoxemia may occur. In addition, the patient’s VO₂ may fluctuate during the ECMO course in response to physiological changes, such as sepsis, agitation, or increased metabolic demand. In such cases, VO₂ may exceed the oxygen delivery capacity of ECMO, making the ability to increase ECMO flow a critical consideration. Thrombosis of the oxygenator membrane or simply clots can reduce its functional surface, impairing gas exchange. Monitoring ΔP (the difference between pre- and post-oxygenator pressures) is crucial, with values exceeding 60 mm Hg suggesting membrane thrombosis. Similarly, a gradual decline in Q_ECMO at a constant rate of revolutions per minute (RPM), in the absence of drainage issues, also indicates this complication. Continuous assessment of these parameters is essential for early detection of oxygenator failure and optimization of ECMO support.

2. Hemoglobin pptimization

Based on the calculation of arterial oxygen content (CaO₂), which is equal to SaO₂×Hb×1.34, optimizing hemoglobin levels enhances oxygen transport capacity, thereby improving systemic oxygenation during ECMO support. This effect primarily results from an increase in CaO₂, and consequently, oxygen delivery (DO₂), rather than from direct changes in SaO₂ or PaO₂15. A low DO₂ may be suspected when SaO₂ is <90% and lactate levels remain elevated. In such cases, increasing the hemoglobin concentration may help restore adequate DO₂.

3. ECMO blood flow management

Adequate Q_ECMO is critical, as insufficient flow is a leading cause of oxygenation failure [16-18]. While the oxygenator’s transfer capacity and fractional membrane oxygen (FmO₂) remain fixed, Q_ECMO is a key modifiable factor that is influenced by cannula size, positioning, and patient volume status. To ensure effective extracorporeal oxygenation, Q_ECMO should ideally reach at least 60% of the patient’s CO19. Since only the blood passing through the ECMO circuit undergoes oxygenation, maintaining an appropriate Q_ECMO-to-CO ratio is essential for systemic oxygen delivery, particularly in cases of impaired native lung function. In most patients, circuit flows ranging 3 to 6 L/min are required to sustain SaO₂ levels above 88%−90% while enabling lung-protective ventilation [19,20].

Determinants of ECMO Blood Flow

1. Cannula size

The size of the cannula significantly impacts Q_ECMO, as described by the Hagen–Poiseuille equation, which states that flow is proportional to the fourth power of the cannula radius (Figure 1). Larger diameters and shorter lengths improve flow while minimizing resistance. Manufacturers indicate the maximum flow rate for each cannula size in French units, although the actual blood flow is influenced by viscosity (Figure 2). Ultrasound-guided vessel assessment is critical to proper cannula selection, ensuring that the cannula diameter is 10% to 20% smaller than the vessel diameter to prevent obstruction21. The conversion from vessel diameter (mm) to French size follows the equation: Fr=-diameter (mm)×3. Proper drainage cannula selection is critical in V-V ECMO18. Unlike venoarterial–ECMO, where lower flows of approximately 2 to 3 L/min may suffice, inadequate cannula size in V-V ECMO can significantly impair oxygenation and drainage efficiency. Multihole large drainage cannula >24 Fr is recommended to ensure optimal oxygenation on V-V ECMO.

Fig. 1.

Hagen–Poiseuille equation demonstrating how extracorporeal membrane oxygenation flow (Q) is affected by the radius (r) of the cannula, the length of the cannula (L), and the viscosity (η) of the blood. This equation shows that flow rate increases with higher pressure difference and larger radius, and decreases with higher viscosity and longer tube length. The dependence on radius to the fourth power makes small changes in radius especially significant for flow. ΔP: pressure difference (between the two ends of the tube); Q: volumetric flow rate of the fluid; η: dynamic viscosity of the fluid; L: length of the tube; r: radius of the tube.

Fig. 2.

Graph showing the relationship between pressure drop and flow rate for different cannula sizes. Each curve represents a different cannula size (19, 21, 23, 25, and 29 Fr), with flow rates on the x-axis and pressure drop (in mm Hg) on the y-axis. This figure illustrates how the diameter of the drainage cannula influences the pressure drop. To achieve a flow of 6 L/min, a 21 Fr drainage cannula results in a pressure drop of –130 mm Hg, which can expose the patient to severe hemolysis and cavitation. In contrast, a 29 Fr cannula ensures a significantly lower pressure drop of –40 mm Hg.

2. Cannula tip positioning

The optimal placement of the drainage cannula tip has traditionally been near the junction of the IVC and the RA [22]. However, a preferred alternative is direct positioning within the RA, rather than the intrahepatic IVC [22,23]. This adjustment enhances venous drainage, ensuring stable Q_ECMO. Cannula placement within the intrahepatic IVC may lead to flow restrictions, particularly during increases in intra-abdominal or intrathoracic pressure (e.g., coughing) [20]. This phenomenon, often observed as ‘chattering,’ can necessitate unnecessary fluid boluses, which if the underlying issue is suboptimal cannula positioning, may be detrimental in ARDS cases. Given that conservative fluid strategies are generally recommended in ARDS care, indiscriminate volume administration should be avoided.

3. Venous return and drainage insufficiency

Compromised Q_ECMO occurs when venous return is inadequate relative to drainage pressures at the cannula ports. Signs of drainage insufficiency include tubing movement (i.e., ‘chattering’), fluctuating blood flow, and a decrease in pre-pump pressure (i.e., P1). If pre-pump pressure falls below −100 mm Hg or exhibits significant oscillations, cavitation and hemolysis may ensue. In addition, the failure of flow to increase despite higher pump speeds suggests drainage insufficiency. Venous return reductions may result from hypovolemia, vasodilation, or elevated intrathoracic pressure, all of which impair ECMO oxygenation capacity [24]. Hypovolemia can be diagnosed via subcostal or transhepatic ultrasound, revealing a reduced IVC diameter with the cannula appearing ‘molded’ within the vessel [21]. Also, tension pneumothorax or cardiac tamponade, both of which increase intrathoracic pressure, must be ruled out, using chest radiography and echocardiography [24].

At the bedside, indiscriminate fluid bolus administration should be avoided in response to drainage insufficiency. Instead, temporarily reducing the pump RPM may help assess whether the issue is resolved. Persistent chattering or unexplained flow reduction necessitates investigating the underlying cause. If hypovolemia is confirmed, a fluid bolus may be appropriate. However, if repeated volume resuscitation is required to maintain ECMO flow and clinical targets remain unmet, the addition of a secondary drainage cannula might be considered [25].

Hypoxemia Despite Adequate ECMO Flow

Hypoxemia may persist despite sufficient Q_ECMO due to several factors that are related either to ECMO circuit factors, or patient factors (Figures 3, 4).

Fig. 3.

Mechanisms and locations of oxygenation failure in venovenous extracorporeal membrane oxygenation. Each boxed item (e.g., membrane failure, mechanical problems) represents a separate category of oxygenation failure with its underlying causes. PmO₂: pression of oxygen on post membrane ECMO; FmO₂: fraction of membrane oxygen; ΔP: pressure difference; RA: right atrium; LA: left atrium; LV: left ventricle; SvO₂: venous saturation on oxygen.

Fig. 4.

Stepwise approach to managing refractory hypoxemia in venovenous extracorporeal membrane oxygenation (ECMO). Q_ECMO: ECMO blood flow; FmO₂: fraction of membrane oxygen; SGF: sweep gas flow; PCO2: arterial pressure in carbon dioxide; SaO2: arterial oxygen saturation; SvO₂: venous saturation on oxygen; PmO₂: pression of oxygen on post membrane ECMO; VO2: oxygen consumption; ΔP: pressure difference.

1. Circuit factors

1) Recirculation

The most common cause is recirculation, which is present across all ECMO settings, though not always associated with severe hypoxemia, where oxygenated blood from the return cannula is inadvertently drawn into the drainage cannula, instead of deoxygenated venous blood. This occurrence is widespread at high Q_ECMO [16], but other factors may exacerbate it, such as cannula malposition, a small drainage cannula, elevated CO, and low central venous pressure. A key clinical indicator is the drainage cannula; when the blood color closely resembles that of the return cannula, this suggests recirculation, which can be quantified as 100% when the pre-pump oxygen saturation (SpreO₂) equals the post-pump oxygen saturation (SpostO₂). Although no universally accepted threshold exists, an SpreO₂ below 75% generally indicates minimal recirculation [20]. In clinical practice, the most practical way to assess relevant recirculation is by trending SpreO₂ alongside peripheral SaO₂. An increasing SpreO₂, coupled with a decreasing SaO₂, particularly when SpreO₂ exceeds or closely approaches SaO₂, serves as a strong indicator of significant recirculation [26]. To minimize recirculation, it is advisable to adjust the positioning of the cannula to ensure sufficient separation. This can be confirmed through chest X-ray or echocardiography, with a recommended minimum distance of at least 10 cm between the tips of the two cannulas. Certain cannulation configurations, such as Fem–Fem ECMO, are more susceptible to recirculation. Similarly, excessive proximity of the cannula tips can exacerbate this issue in femoral–jugular ECMO. The use of a dual-lumen cannula, such as the Avalon, could reduce the risk of recirculation when the reinfusion flow is optimally directed toward the tricuspid valve.

2) Membrane oxygenator dysfunction

A progressive reduction in the oxygenator’s functional surface area due to clot formation can impair oxygen transfer. Oxygenator dysfunction is often accompanied by additional clinical and laboratory indicators, such as visible clots on the arterial side of the membrane, elevated hemolysis markers (e.g., free hemoglobin), or a progressively rising delta pressure. However, these signs are not specific. A post-oxygenator PaO₂ <300 mm Hg, despite an FmO₂ of 100%, strongly suggests membrane dysfunction [27]. Less common mechanisms of oxygenator failure include lipid embolism, drug-related membrane damage (e.g., prolonged high-dose propofol), and heparin-induced thrombocytopenia.

2. Patient factors

1) Increased oxygen demand

As seen in sepsis, elevated metabolic demand can lead to increased CO, reducing the proportion of blood oxygenated by ECMO. Since effective Q_ECMO should exceed 60% of CO to ensure optimum systemic oxygenation [19], a rise in CO may necessitate an increased Q_ECMO. Echocardiography can help assess CO, and if persistently high, interventions such as deep sedation or mild hypothermia may help reduce metabolic demand and improve the Q_ECMO-to-CO ratio [21]. Specifically, these strategies reduce oxygen consumption and CO, thereby increasing the proportion of blood oxygenated by the ECMO circuit and improving SaO₂. Some studies suggest beta-blockers as a potential option for refractory hypoxemia, even when Q_ECMO is sufficient [28]. By decreasing CO, beta-blockers may enhance the Q_ECMO/CO ratio, potentially leading to improved SaO₂ and CaO₂. However, since DO₂ is a function of both CO and CaO₂, lowering CO with beta-blockers may ultimately compromise tissue oxygenation, which is a key objective of ECMO therapy [29]. As a first-line approach, management should prioritize optimizing the Q_ECMO-to-CO balance through interventions such as mild hypothermia, adequate analgesia, and deeper sedation. Beta-blockers should be considered only in select cases where high CO is unrelated to metabolic needs. If used, close monitoring of DO₂, oxygen extraction, and signs of anaerobic metabolism, such as elevated lactate levels, is essential to ensure adequate tissue perfusion.

2) Low hemoglobin levels

Anemia can contribute to inadequate oxygen delivery despite sufficient ECMO flow. The multicenter international prospective cohort Prospective Robservational study on Transfusion practice in ECMO patients (PROTECMO) study reported that red blood cell transfusion was consistently associated with lower mortality only when administered at a hemoglobin concentration of less than 7 g/dL, with no significant mortality benefit observed for transfusions at higher hemoglobin levels [30]. When adequate DO₂ is not achieved, particularly due to low SaO₂, hemoglobin levels should be evaluated, and if necessary, transfusion above the standard 7 g/dL threshold should be considered to optimize oxygen- carrying capacity [30].

By promoting improved dorsal lung recruitment and ventilation/perfusion matching, prone positioning is a highly effective and recommended intervention to improve gas exchange without ECMO [1]. However, its use during ECMO remains more controversial, as the Prone Positioning During Extracorporeal Membrane Oxygenation in Patients With Severe ARDS (PRONECMO) randomized controlled trial did not demonstrate significant physiological or outcome benefits [31]. Nevertheless, since when performed in experienced centers, the procedure is safe, it may be considered as a last-resort option in cases of refractory hypoxemia despite adequate ECMO and ventilator support.

Determinants of Decarboxylation on V-V ECMO

With the implementation of ultra-protective lung ventilation during ECMO, tidal volumes are typically reduced to less than 4 mL/kg of predicted body weight. Under these conditions, CO₂ accumulation becomes an inevitable consequence [32]. While mild CO₂ retention is tolerated, excessive hypercapnia can lead to pulmonary vasoconstriction, and in severe cases, right ventricular failure [33-35]. Therefore, mitigating hypercapnia when possible is clinically advantageous [35]. In the context of extracorporeal circulation, CO₂ removal is more efficiently achieved than oxygenation, and targeting a PaCO2 in the range 35 to 45 mm Hg is usually recommended [2,7,8]. CO₂ transfer through the membrane lung is influenced by Q_ECMO, with maximal CO₂ clearance exceeding 300 mL/min at flows above 6 L/min using the Quadrox oxygenator (MAQUET Cardiopulmonary AG, Hirrlingen, Germany) (Table 2). However, due to the high diffusibility of CO₂, approximately 20 times greater than that of oxygen, substantial decarboxylation can occur, even at low Q_ECMO [36]. Studies have demonstrated that even when Q_ECMO is reduced below 2.5 L/min, PaCO₂ remains stable. This principle has facilitated the development of low-flow extracorporeal CO₂ removal devices, which are capable of eliminating more than 70 mL/min of CO₂ at Q_ECMO as low as 450 mL/min.

Table 2.

Key factors influencing CO₂ removal on venovenous ECMO

Factor	Description
QECMO	The volume of blood passing through the ECMO circuit directly affects CO₂ removal. Increased QECMO enhances decarboxylation. However, a flow at 2 L/min is frequently enough to remove all CO₂ production.
Sweep gas flow	The rate of sweep gas flow across the oxygenator is a major determinant of CO₂ removal. Higher flow rates facilitate greater CO₂ clearance.
Membrane size	The rate of CO₂ removal depends on the membrane properties especially the size, the material and structure of the membrane (e.g., polymethylpentene).
Systemic PCO₂	The higher the pre-ECMO PaCO₂, the faster its clearance through the membrane oxygenator. Controlled CO₂ reduction is essential to minimize the risk of intra-cranial bleeding.

CO₂: carbon dioxide; ECMO: extracorporeal membrane oxygenation; Q_ECMO: ECMO blood flow; PCO₂: arterial pressure in carbon dioxide; PaCO₂: partial pressure of carbon dioxide.

A critical determinant of CO₂ clearance is the sweep gas flow across the oxygenator, which is the primary driver of extracorporeal decarboxylation. Given ECMO’s powerful CO₂ removal capacity, careful titration is required to prevent an excessively rapid decrease in PaCO₂. Clinical studies have shown that a PaCO₂ reduction exceeding 27 mm Hg is significantly associated with an increased risk of intracranial hemorrhage [37]. This risk is particularly relevant in patients with severe hypercapnia before ECMO initiation. Since CO₂ removal follows a diffusion-driven process, the higher the pre-ECMO PaCO₂, the faster its clearance through the membrane oxygenator. As controlled CO₂ reduction is essential to minimize the risk of neurological complications in these cases, it is recommended to induce a gentle decrease in PaCO₂ through a very gradual increase in sweep gas flow (steps of 0.5 to 1 L/min) over the next 24 hours, with close monitoring of arterial blood gases or end-tidal CO₂.

Conclusion

Optimizing oxygenation and CO₂ removal via the ECMO membrane is fundamental in managing severe ARDS. Enhanced ECMO-mediated oxygenation further reduces mechanical ventilation, facilitating ultra-lung-protective strategies. Ensuring adequate Q_ECMO is paramount, and requires appropriate cannula selection and positioning, as these directly impact circuit performance. Persistent hypoxemia despite sufficient flow necessitates evaluation for recirculation, increased VO₂, anemia, or oxygenator dysfunction. On the other hand, careful titration of the sweep gas flow is required to prevent an excessively rapid decrease in PaCO₂. Addressing these factors is essential for maximizing DO2 and CO2 removal and improving the efficacy of V-V ECMO support in critically ill patients.