The development of extracorporeal life support modalities has added a new dimension to the care of critically ill patients who fail conventional treatment options. Extracorporeal circuits, like those used in hemodialysis, cardiopulmonary bypass, ventricular assist devices, extracorporeal membrane oxygenation (ECMO), and therapeutic apheresis, are common in modern medicine.
Extracorporeal membrane oxygenation presents specialized temporary life support for patients with severe cardiac or pulmonary failure, bridging time for organ recovery, transplant, or permanent assistance. The beginning of ECMO support dates from 1971, when the first prolonged extracorporeal oxygenation and perfusion were used in the case of a patient with severe acute respiratory distress syndrome (ARDS)
[1]. Over the last decade, the indications for ECMO support have expanded beyond severe respiratory failure and refractory cardiogenic shock
[2] to include an assortment of clinical presentations, including a bridge to heart or lung transplantation
[3], extracorporeal cardiopulmonary reanimation (ECPR)
[4], resuscitation of patients with severe traumas
[5], and rewarming due to accidental deep hypothermia
[6]. Recently, the use of ECMO support for out-of-hospital cardiac arrest has been popularized in some countries, with reported improvements in outcome
[7][8][9][10].
2. ECMO Configurations and Circuits
There are two main ECMO configurations: venoarterial (va-ECMO), used for a refractory cardiogenic shock, and venovenous (vv-ECMO), used for a severe respiratory failure, both of which can be subject to several modifications
[20]. Venoarterial ECMO combines adequate oxygen delivery and carbon dioxide removal with circulatory support
[21]. Vascular access is obtained by the placement of a drainage cannula in a large central vein supplying the ECMO system with patient blood. A second cannula returns the oxygenated blood either to the venous (vv-ECMO) or the arterial system (va-ECMO).
The traditional ECMO circuit utilizes technology as a cardiopulmonary bypass, i.e., it is a closed circuit with a membrane-type gas-exchange system
[21]. The main distinctions between those two extracorporeal life modalities are in the duration of support, the existence of a venous reservoir, the air–blood interface, and the cardiotomy reservoir. Cardiopulmonary bypass is usually only employed for the duration of surgery, while ECMO support may be needed for weeks or even months
[21].
The ECMO circuit consists of three main components: the pump, the gas, and the heat-exchange device connected with the polyvinyl chloride tubing (usually UFH coated). The earlier roller blood pumps have now been exchanged with more advanced and magnetically actuated centrifugal pumps that control the required blood flow
[22][23]. Gas-exchange devices, fulfilling the patient’s metabolic needs for oxygen and the removal of carbon dioxide, evolved from direct air–blood contact systems to membrane-type gas-exchange devices (oxygenators). Since the early 2000s, polymethylpentene hollow fiber membranes are increasingly used, where the gas is ventilated through hollow fiber bundles and the blood circulates around the fibers (
Figure 1)
[21][24].
Figure 1. Schematic presentation of a diffusion membrane showing blood flow between the gas and water-filled network of hollow fibers. With permission of Maquet.
The amount of oxygen in the gas mixture depends on the metabolic need of the patient and can be increased up to 100% oxygen. The gas–blood flow ratio is typically adjusted to maintain normocapnia. Increasing the sweep gas flow will lead to an increased clearance of carbon dioxide while not altering oxygenation. Decreasing the sweep flow will result in increased partial pressure of carbon dioxide in the blood
[21].
Keeping the circulating blood pressure in the oxygenator higher than the pressure of the circulating gas is of utmost importance to prevent the passage of air bubbles across the membrane, which can result in air embolism. Therefore, the oxygenator device should always be located below the level of the patient’s heart. Furthermore, insertion or removal of central catheters or interventions with the opening of blood vessels could result in aspiration of air by the ECMO system
[21][25].
Finally, the third component of the ECMO system is a heat exchanger preventing circuit-related heat dispersion but also giving the option of a targeted temperature management, for example, in the treatment of sepsis, metabolic crisis, rewarming of accidental hypothermia, or as therapeutic hypothermia after cardiac arrest. The heat exchanger may be integrated into the gas-exchange device or be a separate component. It is usually based on nonpermeable hollow fiber bundles with circulating nonsterile water (
Figure 1)
[21].
3. Inflammation, Coagulation, and ECMO
The normal hemostasis in critically ill patients receiving ECMO support is distorted. Surgical trauma (ECMO implantation or cardiac surgery) and exposure of blood to the large surfaces of the ECMO circuits initiate and propagate immediate inflammatory response and activation of the coagulation cascade. Furthermore, the complexity of critical illness and its inflammatory response may additionally imbalance patient hemostasis
[12]. The systemic inflammatory response in patients receiving cardiopulmonary bypass is well established and discussed extensively in the literature
[26][27][28][29][30], but the information on this complex and multifaceted inflammatory response to ECMO support is still limited.
Several humoral and cellular systems are involved in complex interactions between inflammation and coagulation during ECMO support. Acute inflammation initiates clotting, compromises the fibrinolytic system, and reduces the activity of natural anticoagulant mechanisms. Moreover, endotoxin, IL-1β, and tumor necrosis factor-α (TNF-α) downregulate thrombomodulin and neutrophil elastase cleaves thrombomodulin from the endothelial cell surfaces
[31][32]. P-selectin and E-selectin are synthesized or expressed on endothelial and platelet surfaces. Tissue factor, from the cell surface of leucocytes and monocytes, is induced by endotoxin, CD40 ligand, or TNF-α. It further binds factor VIIa and converts factor X to its activated form (Xa), which together with factor Va generates thrombin from prothrombin
[33]. Additionally, inflammation reduces protein C levels, probably due to a combination of consumption and associated liver dysfunction with a consequent nonactivation of factor Va leading to the stabilization of prothrombin activation complexes
[34]. Increased C-reactive protein levels facilitate monocyte–endothelial cell interactions, promote plasminogen activator inhibitor-1 and tissue factor formation, and induce complement activation
[35][36][37]. The activation of platelets, key elements of hemostasis and inflammation, occurs as a result of complement activation and thrombin generation with a consequent release of a variety of mediators (proinflammatory cytokines, chemokines, adhesion factors, proteases, hemostatic factors, etc.). This all together plays a role in the development of a systemic inflammatory response
[38][39][40][41].
Antithrombin is inactivated and/or consumed, while the levels of vascular heparin-like molecules may be reduced due to neutrophil activation products and inflammatory cytokines
[42][43][44]. Finally, detailed information on the role of the fibrinolytic system in patients receiving ECMO support is still lacking, but recent studies reported on the association between increased fibrinolysis and bleeding complications
[45][46].
Furthermore, different configurations of ECMO may also alter hemostasis. In the prospective HECTIC trial, the rate of thrombosis in va-ECMO was around 40%, with rates twice as high in vv-ECMO
[47]. Moreover, in an ex vivo model, ECMO flow rates below 1.5 L/min were shown to decrease platelet aggregation, weaken clot firmness, and surprisingly increase hemolysis (despite the lower pump speed)
[48]. Given the sparsity of evidence for low ECMO flows in humans, researchers establish more anticoagulation at lower flow rates (e.g., ACT 150–170 s at 2–3 L/min). Moreover, researchers use the same anticoagulation protocol per se in va- and vv-ECMO configurations but strive to tailor anticoagulation to each patient using viscoelastic monitoring on a routine basis.
Therefore, homeostasis and the balance between the procoagulant and anticoagulatory factors are crucial to avoid hemorrhagic or thromboembolic complications, and for the patency of the extracorporeal circuit and its components (see Figure 2).
Figure 2. Presentation of prothrombotic and prohemorrhagic factors with an influence on homeostasis. Achieving a balance between the risk of bleeding and thrombosis is both critical and complex in patients receiving ECMO support. Aside from the initiation and propagation of the inflammatory response (proinflammatory state) and the activation of the coagulation cascade (prothrombotic state), ECMO may also lead to platelet dysfunction, fibrinolysis, malfunction of von Willebrand factor, and consumption of coagulation factors leading to a prohemorrhagic state.
Interestingly, within the first 10 min after ECMO support initiation and contact of blood with the artificial surfaces, factor XII cleaves into factor XIIa and XIIf. Factor XIIa has an important role in the activation of kallikrein and bradykinin, both strong drivers of inflammation and coagulation
[49][50][51]. The role of bradykinin in inflammation may be even more interesting in va-ECMO with the lungs, the major site of bradykinin inactivation, being bypassed. As a response to the role of factor XII, its neutralization may lead to a reduction in inflammation, which has been recently shown in ex vivo and animal ECMO models with the use of a plasma protease factor XII function-neutralizing antibodies
[52][53]. Further studies focused on potential uses in humans are warranted.
Lastly, it should be mentioned that, like any exposure to mechanical support devices, patients on ECMO support can develop increased human leukocyte antigen (HLA) sensitization, which is of relevance in bridge-to-transplant therapeutic considerations
[54].
Given the above, systemic anticoagulation and coagulation monitoring are of immense importance for adverse events prevention in patients receiving ECMO support. The association of inflammation and thrombosis, known as thromboinflammation, is well reported in the literature, especially in COVID-19 patients
[55][56]. Hyperinflammation may lead to a limitation of the UFH effect by decreasing antithrombin levels or increasing heparin binding to acute phase proteins
[42][43][44]. A recent report on a possible association between bleeding and unintended excessive anticoagulation in ECMO patients without hyperinflammation remains to be confirmed in larger cohorts
[13]. Finally, despite the extensive development of anticoagulants, ECMO pumps, oxygenators, and tubing systems, the systemic inflammatory response syndrome and distorted hemostasis remain a clinical concern. It is still unclear if the extent of inflammation may also benefit the patient, beyond its deleterious effects. To warrant a more detailed understanding of the underlying pathophysiological processes, the reporting on inflammatory response during ECMO support should be improved in forthcoming studies.