Keywords
ECMO - History - Anticoagulation
Dawn of Extracorporeal Membrane Oxygenation and Its Evolution
Dawn of Extracorporeal Membrane Oxygenation and Its Evolution
Extracorporeal membrane oxygenation (ECMO) is the only clinical circumstance in which all the cardiac output is circulated outside the body for days or weeks or sometimes months. All the many factors which cause clotting are activated and are controlled using a single anticoagulant. It is surprising that this works at all, and yet, it is successful for all types of patients for long durations of time. This review will evaluate all of those factors and consider clotting and bleeding during ECMO.
Prior to 1960, cardiopulmonary bypass (CPB) was developed and first used in 1954 for cardiac surgery.[1] Blood was anticoagulated with high doses of heparin and reversed with protamine after the heart was repaired. This led to the development of all of cardiac catheterization and cardiac surgery. However, the CPB system used in those days was very damaging to blood and was lethal if used for more than 2 hours. This was caused by the direct interface between blood and air or oxygen.
Prolonged extracorporeal circulation (ECC) is possible because membrane oxygenators (membrane lungs) were developed in the 1960s, which prevented the interface between blood and air in the original CPB devices. In 1965, Bartlett et al demonstrated that ECC could be maintained for 4 days in dogs using a membrane lung which we designed in our laboratory.[2]
[3]
[4] To prevent clotting in the pump, tubing, and oxygenator, they used continuous infusion of heparin as in CPB. Initially, the bleeding was uncontrollable. They decreased the heparin dose in stages, monitored by the whole blood activated clotting time (ACT).[5]
Bartlett et al found that maintaining the ACT about twice the baseline prevented clotting and bleeding in these prolonged animal studies. For the activator, they used rabbit brain thromboplastin, which is used in the clinical laboratory to measure the prothrombin time (PT) in plasma. They used this approach to develop and study all the physiology and hematology variables in prolonged ECC in laboratory animals.[6]
[7] Others did the same.[8]
[9]
The first successful clinical cases of prolonged ECC were in 1971 to 1975. This was used for potentially fatal respiratory and cardiac failure.[10]
[11]
[12] In these studies continuous infusion of heparin was titrated to two times normal ACT. This technology became known as ECMO. Over 100 cases were reported by 1980.[13] Through the 1980s, ECMO became the standard treatment for infants and children with cardiac and respiratory failure unresponsive to other treatments. The survival rate was 70 to 90% in children who were otherwise moribund and expected to die. This was proven in prospective randomized trials.[14]
[15]
[16] There are three types of ECMO: venovenous (VV) ECMO, venoarterial (VA) ECMO, and central ECMO. VV ECMO is indicated to support the lung (i.e., for respiratory failure). VA ECMO and central ECMO are used to support both the lung and the heart. ECMO is now used for severe cardiac and respiratory failure in all age groups. There are now over 170,000 cases in the Extracorporeal Life Support Organization (ELSO) Registry.
The ELSO was formed in 1989. It is a consortium of medical centers devoted to the study and improvement of ECMO practice. ELSO maintains a registry of cases, provides education and certification, publishes guidelines, and publishes the definitive text of all things ECMO. This text (called the Red Book) is updated every 5 years. The sixth edition was published in 2022.[17] There are thorough discussions of thrombosis and anticoagulation in the Red Book.[18]
Bartlett et al switched from rabbit brain thromboplastin-driven (extrinsic) PT to the contact activator kaolin to initiate (intrinsic) coagulation because it was easier to manage at the bedside. Contact activation-started, activated partial thromboplastin time (aPTT) in plasma was also used to titrate heparin. This could be done in the central laboratory and did indeed measure the effect of heparin on coagulation, but as this is measured in cell-free plasma, it does not measure the effects of platelets, white cells, or red cells. Continuous infusion of heparin titrated to ACT or aPTT 1.5 to two times normal remains the standard of care in ECMO today. Direct thrombin inhibitors (DTIs) such as bivalirudin or argatroban are used as anticoagulants in patients suspected of having heparin-induced thrombocytopenia (HIT). These drugs are also used as the primary anticoagulant in many centers. The dose of DTI is titrated in the same way as heparin, to an ACT or aPTT of 1.5 to two times normal. When ECMO is used at high flow rates, it is possible to discontinue anticoagulation altogether and some centers have reported managing ECMO with only minimal subcutaneous doses of heparin.[19]
[20] Results are good, in fact better than with full anticoagulation.
Coagulation can also be measured by viscoelastography, such as thromboelastography (TEG) or rotational thromboelastometry (ROTEM), which provides much more information than clotting times, regarding clotting and lysis in a single whole blood sample. The amount of heparin in plasma can also be measured by the anti-factor Xa (anti-Xa) assay in plasma, but this does not measure the heparin effect in whole blood, so it is less valuable in managing ECMO patients at the bedside. It should be noted that ACT, aPTT, anti-Xa, and TEG/ROTEM, all measure different components of clotting. They should not and do not correlate with each other during ECMO. The best measurement of clotting and anticoagulation in ECMO is ACT, TEG, or ROTEM because they measure coagulation in whole blood.
Of course, there are many factors which contribute to the clotting of blood in ECMO systems. Platelet activation and release of platelet molecules predominate, but surface conditions, surface absorption, rheology at different flow rates, step-off and irregularities in the circuit, and patient conditions such as sepsis are just a few examples. All of these variables are discussed in detail in this article.
There are many surface modifications and coatings that are used on the plastic devices used for ECMO. Despite this, systemic anticoagulation is still required.[21] This laboratory has also been investigating generating nitric oxide (NO) at the plastic surfaces to prevent platelet adhesion and activation. NO serves this role in the normal endothelium.[22] This is one approach to eliminating systemic anticoagulation during ECMO. With all the research in this area, we expect that it will be possible to manage ECMO without systemic anticoagulation in the near future.
Evolution of Anticoagulation in Adults
Evolution of Anticoagulation in Adults
Anticoagulation is an integral part to maintain the circuit patency, to reduce the risk of thrombosis in the circuit and the patient during ECMO support. Heparin as bolus dose followed by continuous infusion has been the main stay of anticoagulation in patients supported with ECMO for many years.[23] This is based on the experience from patients undergoing CPB. Although heparin is inexpensive, has antidote (protamine sulfate) in case of significant bleeding, and has additional actions such as anticomplement and anti-inflammatory effects in addition to its anticoagulant effect (which may be beneficial to patients supported with ECMO), it has some disadvantages.
Heparin is a sulfated polysaccharide, and its major anticoagulant action is by inactivating thrombin and activated factor X (factor Xa) through an antithrombin (AT)-dependent mechanism.[24] Therefore, AT is essential to produce its anticoagulant effect. Heparin exhibits a marked variability in anticoagulant response due to its highly negatively charged nature leading to its binding for positively charged plasma proteins, proteins released from platelets, and endothelial cell proteins and surfaces which is a prominent feature in patients supported with ECMO.[24] More importantly, HIT is a significant complication with heparin use and can be a more frequent complication in patients supported with ECMO compared with others treated with heparin.[25]
Despite wider use of heparin in ECMO, target anticoagulant intensity, and the best methods to monitor anticoagulation in patients supported with ECMO are still under debate and there is a significant variation in the clinical practice. Generally, a bolus of heparin (25–100 units/kg) with maximum 5,000 units is given at the time of ECMO cannulation followed by continuous infusion to maintain a specific target depending on the method of the monitoring according to the ECMO center[23]
[26] in the absence of a significant bleeding especially intracerebral bleeding ([Table 1]).
Table 1
Methods of monitoring heparin in patients supported with ECMO
Assay
|
Advantages
|
Disadvantages
|
Comments
|
Activated clotting time (ACT)
|
Rapid, point of care test, especially with high concentration of heparin such as during cardiopulmonary bypass when aPTT is unrecordable
|
Not specific for heparin
ACT can be affected by
temperature, coagulation factors (fibrinogen, factor XII, XI, X, IX, V, or II), platelet count and their function
|
ACT of 180–220 s is generally used during ECMO support but this can be varied depending on the center, device including activator used
|
Activated partial thromboplastin time (aPTT)
|
Can be point of care test or laboratory-based assay
Easily available
|
Not specific for heparin
aPTT can be prolonged due to coagulation factor deficiency (fibrinogen, factor II, V, VIII, IX, X, XI, or XII)
The presence of inhibitors including lupus anticoagulant
Affected by acute phase reaction (high factor VIII, fibrinogen, or C-reactive protein)
|
aPTT ratio 1.5–2.5 with the midpoint of normal aPTT range is generally regarded as the therapeutic target for thrombosis
There is a variation in the accepted aPTT ratio or aPTT range in patients supported with ECMO (i.e., aPTT 40–60 s in VV-ECMO or 60–80 s in VA-ECMO)[23]
[26]
|
Heparin anti-Xa level
|
Specific for heparin and not affected by other factors such as thrombocytopenia or coagulation factor deficiency
|
May not be easily available and the results may be affected by haemolysis and hyperbilirubinemia
|
Target heparin anti-Xa is 0.3–0.7 units/mL in patients with thrombosis
Target heparin anti-Xa level in patients supported with ECMO may be varied (i.e., anti-Xa 0.2–0.3 units/mL in VV-ECMO and 0.3–0.5 units/mL in VA-ECMO)[23]
[26]
|
Abbreviations: ECMO, extracorporeal membrane oxygenation; VA, venoarterial, VV, venoarterial.
Traditionally, DTIs such as argatroban and bivalirudin have been used in patients who developed HIT.[25] However, recently many ECMO centers have taken interest in the use of DTIs in patients supported with ECMO due to lack of dependency on AT, no risk of HIT, and more predictable anticoagulant effect due to no significant nonspecific binding to positively charged proteins and molecules.[27]
[28]
Evolution of Anticoagulation in Pediatrics including Newborns
Evolution of Anticoagulation in Pediatrics including Newborns
The research and trials on novel antithrombotic agents in pediatrics have increased in recent years and their adaptations have accelerated. The introduction of these novel agents provides new and exciting therapeutic options; however, they also present new challenges to our established clinical practices. In the pediatric ECMO setting, the decision of prescribing an alternative anticoagulant in place of the traditional heparin is complicated by an array of considerations including the native hemostatic state of the patient, evidence on efficacy, and appropriate therapeutic monitoring. A small total blood volume of neonates gives a challenge of anticoagulation due to the almost equal volume of patient and extracorporeal volume in the ECMO circuit, which is approximately 300 mL.
Developmental Hemostasis and Its Challenges in Extracorporeal Membrane Oxygenation
At birth, there are differences in the hemostatic system as compared with adults which can lead to management challenges.[29]
[30] Specifically, at birth many of the procoagulant proteins are almost half of adult normative levels, for example, factor IX activity. Some other procoagulant proteins are elevated in neonates, such as von Willebrand factor. These levels will not begin to reach adult norms until roughly 6 months of life. Levels of fibrinogen, factor V, factor VIII, and factor XIII are some of the few that are normal at birth. Additionally, neonates have developmentally different levels of thrombin inhibitors. Antithrombin, protein C, protein S, and tissue plasminogen activator levels are all lower at birth. Furthermore, infants are more dependent on α2-macroglobulin for thrombin inhibition than adults. Lastly, blood vessels in neonates, specifically intracerebral, are inherently fragile and susceptible to vascular injury. Taken collectively, these natural differences result in a delicate balance between hemostasis and thrombosis that poses a particular challenge. The PT and aPTT need to be interpreted with age-appropriate reference ranges, and anti-Xa may be more suitable for drug monitoring. Antithrombotic agents, such as heparin, are dependent on antithrombin and therefore require higher dosing. The use of exogenous antithrombin to achieve therapeutic drug levels in neonates remains controversial.[30]
Anticoagulants Used in Pediatrics and Neonates
In outpatient pediatric settings, choices for anticoagulation range from the historically used low molecular heparin (LMWH) and warfarin to the novel pharmaceuticals like direct-acting oral anticoagulants.[31] However, for hospitalized pediatric patients, heparin and heparin-analog such as LMWH remain the ideal anticoagulants of choice.[32] In the setting of ECMO and mechanical circulatory devices, heparin and its analogs have been proven to be effective in both preventing and treating thrombosis, resulting in the established safety portfolio and general familiarity with heparin in this population.[33]
[34] However, nonheparin-based anticoagulants such as DTIs are steadily gaining popularity, especially when the more conventional anticoagulation strategies with heparin fail.[35] DTIs are utilized in adult ECMO mostly due to concerns for HIT,[36] but the incidence of HIT remains low in children. The clinical experiences and published data for DTI use in pediatric ECMO, however, remain limited to scenarios where heparin has failed resulting in thromboembolic events. It has been shown that bivalirudin is associated with reduced incidence of major bleeding, thrombosis, and in-hospital mortality in children and can be preferred over heparin in pediatric ECMO.[37] As more data on DTI in the setting of pediatric ventricular assist devices (VADs) becomes available, DTIs may prove to be effective alternatives to heparin, due to the ease of maintaining steady-state levels compared with heparin.[38]
Challenges with Monitoring Anticoagulation for Extracorporeal Membrane Oxygenation in Pediatrics and Newborns
The goal of monitoring anticoagulation in patients on ECMO is to prevent bleeding and thrombosis. The lack of an “ideal coagulation assay” to predict these events is a major drawback. The most common tests used to monitor anticoagulation include aPTT, ACT, and anti-Xa as heparin remains the most common anticoagulant in use. The aPTT has been shown to correlate poorly with anticoagulation and complications of bleeding and thrombosis and is affected by elevated activity of acute phase reactants (factor VIII, fibrinogen, and C-reactive protein) and the presence of lupus anticoagulant, making the assay unreliable. Its universal availability, however, makes it an attractive target. The ACT, being a point-of-care assay continues to be utilized, despite several studies challenging its utility especially in pediatrics.[39] The anti-Xa has been shown to be more specific to the effect of heparin anticoagulation[40] and correlates well with heparin levels but correlation with bleeding and thrombotic complications remains controversial. Using anti-Xa assay utilizing endogenous antithrombin, not exogenously added antithrombin, is recommended since it reflects actual anti-Xa activity in vivo due to heparin antithrombin complex. Platelet function is known to be affected in patients on ECMO, but aggregation studies require large volumes of blood and specialized laboratory facilities and therefore not often feasible. The utility of functional platelet studies in the setting of ECMO is still unclear, although the passage of platelets through the ECMO circuit could potentially result in activation and subsequent consumption or platelet dysfunction resulting in bleeding. Its utility in thrombotic situations may include defining the optimal level of platelet inhibition in patients on antiplatelet agents to prevent thrombosis. Viscoelastic tests (ROTEM or TEG) have been used more often and shown to be beneficial as they are able to assess the overall hemostatic capacity and may compliment the routine assays but have not been shown to be independently predictive of bleeding or thrombosis.[41] The availability and utilization of newer anticoagulants like DTI, specifically bivalirudin, adds to the controversies. These agents have been shown to be more effective, but data in pediatrics are limited. The aPTT has been used to monitor DTI and has several limitations including the plateau effect at higher concentrations of bivalirudin, placing the child at risk for bleeding.[42] The dilute thrombin time and anti-factor IIa assay are deemed to be more specific but are still being investigated for their utility in monitoring bivalirudin.[43]
[44] Monitoring anticoagulation in pediatric ECMO, therefore, remains a challenging and evolving subject, with a dire need for data to correlate with the complications of bleeding and thrombosis.
Final Words
Thrombosis and hemorrhage continue to be one of the most common complications and causes of mortality among pediatric ECMO. With new research, accumulated experience, and advances in laboratory technologies, it remains to be seen if novel antithrombotic agents like bivalirudin or traditional heparin will be the defacto agent for ECMO anticoagulation.
Management of Bleeding and Thrombosis during Extracorporeal Membrane Oxygenation
Management of Bleeding and Thrombosis during Extracorporeal Membrane Oxygenation
ECMO is associated with high rates of bleeding and thrombosis, inherent with the risk of thrombosis is the potential risk of circuit loss and hence the need for anticoagulation. Worldwide, heparin is the most used anticoagulant although there is an increasing use of the DTIs such as argatroban and bivalirudin. Major hemorrhage has been reported in up to 50% of patients requiring ECMO, and it is associated with increased mortality.[45] Intracranial hemorrhage is of particular concern as well as other “medical” bleeds such as pulmonary and gastrointestinal.[45]
[46] Thrombosis can occur in the patient or within the circuit despite the use of anticoagulation. Deep vein thrombosis is the most common thrombotic event due to the use of central venous catheters.[47] However, pulmonary emboli and ischemic stroke are also recognized in 5 to 10% of patients.[45]
[46]
[48] Although rates remain high, international registries have shown that they have decreased over the last decade.[45] An increased awareness and surveillance for these complications and monitoring of hemostatic defects are likely to be key contributors.
Thrombotic events typically prompt clinicians to increase target ranges of anticoagulation or introduce an antiplatelet agent, although this should be guided by the severity and site of thrombosis versus the risk of bleeding. Alternatives to anticoagulation can be considered in situations with an unacceptable risk of bleeding, with the use of mechanical thrombectomy and catheter-directed thrombolysis described.[49]
[50]
Hemostatic defects contributing to bleeding due to ECMO include acquired von Willebrand syndrome, factor XIII deficiency, hypofibrinogenemia, thrombocytopenia, and platelet dysfunction.[51]
[52]
[53] International guidelines recognize the need to monitor platelet counts, fibrinogen levels, PT, and aPTT; some centers have adopted monitoring of these additional components to provide more bespoke, personalized treatment for bleeding with promising results in terms of survival, bleeding, and total blood product use.[54] Treatment of bleeding is multifaceted incorporating the cessation/reversal of anticoagulation, identification, and treatment of the source of bleeding, blood product replacement to correct hemostatic defects, and the use of antifibrinolytic agents. Evaluation of the bleeding source is crucial including surgical intervention, endoscopic procedures, and interventional radiological procedures. Adjuncts such as topical antifibrinolytic agents have successfully been employed particularly in endobronchial bleeding.[55]
Although there is a concern that stopping anticoagulation increases the risk of circuit occlusion, recent data showed similar outcomes if anticoagulation is ceased for short durations.[56] This is due to the increasingly recognized effect of hemorrhage on survival outcomes. Protamine can be used to reverse anticoagulation with heparin, but the risk of circuit loss is too high with ECMO.
Systemic antifibrinolytic agents such as tranexamic acid and aminocaproic acid are widely used during surgery especially cardiac surgery and trauma as they reduce mortality due to bleeding.[57]
[58] However, there is concern that circuit occlusion and thrombosis rates are increased in ECMO studies, largely limited to single centers.[59] Therefore, while these drugs should be considered during major hemorrhage, extrapolation to continuous use needs to be with caution and studied further. Further trials to establish best practice and assess other treatment options for bleeding such as desmopressin (DDAVP) to promote platelet function and recombinant activated factor VII during major hemorrhage are required.[55]
Venoarterial Extracorporeal Membrane Oxygenation—A Cardiology Perspective
Venoarterial Extracorporeal Membrane Oxygenation—A Cardiology Perspective
VA ECMO delivers a nonphysiological parallel circulation that provides effective cardiopulmonary support, even in the setting of circulatory arrest. With improved access and maturation of the technology, it was perhaps predictable that VA ECMO would be enthusiastically adopted, particularly by critical care cardiology and cardiac surgery. The utilization of VA ECMO has increased exponentially over the last two decades. Germany provides a stark illustration—the annual number of VA ECMO procedures increased from 80 in 2007 to 2614 procedures in 2015, an increase of more than 30-fold.[60] In the United States, VA ECMO is increasingly deployed to bridge patients with advanced heart failure to heart transplantation since the revision of the heart allocation scheme in 2018.[61]
Acute coronary syndrome and acute decompensated heart failure are the main causes of cardiogenic shock.[62] With the exception of revascularization, effective therapeutic interventions in cardiogenic shock are limited, and morbidity and mortality remain considerable. Inadequate oxygen delivery due to circulatory failure is the defining abnormality in cardiogenic shock, and this deficiency in oxygen delivery can be effectively remedied by VA ECMO support. The ELSO recently reported on the outcomes of VA ECMO support in cardiogenic shock.[63] The ELSO Registry data were notable for two reasons. First, the proportion of patients with cardiogenic shock supported with VA ECMO who were discharged alive was unchanged annually between 2014 and 2018 (38.0, 36.4, 42.7, 43.5, and 38.5%; p = 0.66). Second, major bleeding complications were common, including central nervous system hemorrhage (2.1%), pulmonary hemorrhage (4.1%), and gastrointestinal bleeding (6.1%). Other studies have similarly shown a high incidence of major bleeding in cardiogenic shock and VA ECMO support. In the 2017 Culprit Lesion Only PCI versus Multivessel PCI in Cardiogenic Shock (CULPRIT-SHOCK) substudy, 21.5% of patients suffered at least one bleeding event within 30 days, including fatal bleeding in 5.4%. VA ECMO support was a risk factor for bleeding in cardiogenic shock, independent of heparin and antiplatelet therapy. Bleeding was associated with a significantly increased risk of early mortality.[64]
A multipronged approach is required to reduce morbidity and mortality in patients with cardiogenic shock and VA ECMO support. First, greater adoption of protocolized team-based care to optimize the delivery of mechanical circulatory support in patients with cardiogenic shock at specialized centers may improve outcomes.[65] Second, there is urgency in understanding the hemostatic dysfunction and management of antithrombotic therapy to minimize bleeding complications associated with VA ECMO. Therapeutic monitoring of heparin remains controversial due to the limitations of the available whole blood and plasma-based tests and highly variable practices between centers. TEG-guided anticoagulant management have shown promise in VV ECMO.[66] However, it is not clear if these results could be extrapolated to VA ECMO in cardiogenic shock, with reported hemostatic and coagulation differences between VA and VV ECMO.[67] Finally, clinical trials are needed to assess the efficacy and role of VA ECMO as part of a mechanical circulatory support strategy and TEG-guided anticoagulant management in VA ECMO in cardiogenic shock.
Hemocompatibility of the Extracorporeal Membrane Oxygenation Circuit
Hemocompatibility of the Extracorporeal Membrane Oxygenation Circuit
ECMO is often needed for many days to weeks to support refractory lung or heart failure. Despite systemic anticoagulation, prolonged blood exposure to the artificial circuit results in the deposition of plasma proteins, activation of platelets, and binding of circulatory cells on its surface, leading to circuit-associated thromboinflammation.[68]
[69] These circuit-associated thromboinflammatory responses may exacerbate the patient's multiorgan dysfunction and increase the risk of thrombotic events (e.g., oxygenator failure, circuit clots, and embolic events)[70]
[71] Ideally, an artificial circuit surface should replicate the endothelium's antithrombotic and anti-inflammatory properties to prevent circuit-associated thromboinflammation and preclude the necessity of systemic anticoagulation on ECMO. Over the years, ECMO circuit components and materials have been modified to enhance their hemocompatibility to reduce circuit-associated thromboinflammation.[68]
[69] The surface coating method modifications focusing on hemocompatibility can be classified into three categories: biomimetic (or bioactive), biopassive, and endothelization ([Table 2]).[69]
Table 2
Types of hemocompatibility ECMO circuit coatings[69]
Coating type
|
Technology
|
Name
|
Manufacturer
|
Biomimetic
|
Heparin coating
|
Carmeda
|
Medtronic
|
Nitric oxide releasing
|
Under development
|
Commercially unavailable
|
Complement inhibitor
|
Under development
|
Commercially unavailable
|
Argatroban
|
Under development
|
Commercially unavailable
|
Biopassive
|
Poly-2-methoxyethyl acrylate (PMEA)
|
X-coating
|
Terumo
|
Tribloc copolymer (polycaprolactone–polymethylsiloxane–polycaprolactone)
|
Smart -X
|
Sorin
|
Phosphocholine
|
P.h.i.s.o
|
Sorin
|
Albumin
|
Safeline
|
Maquet
|
Amphophilic polymer
|
Softline
|
Maquet
|
|
Zwitterionic polymer
|
Under development
|
Commercially unavailable
|
Tethered liquid perfluorocarbon
|
Under development
|
Commercially unavailable
|
Combination
(biomimetic and biopassive)
|
Covalently bonded heparin, sulfate and sulfonate groups (negatively charged), and polyethylene oxide (hydrophilic)
|
Trillium
|
Medtronic
|
Covalently bonded heparin with albumin
|
Bioline
|
Maquet
|
Endothelization
|
In-vitro and in-vivo endothelization
|
Under development
|
Commercially unavailable
|
Abbreviation: ECMO, extracorporeal membrane oxygenation.
Biomimetic (or bioactive) surfaces include heparin-coated, NO-coated, covalent C1-esterase inhibitor-coated, and argatroban-coated circuits.[72]
[73]
[74]
[75]
[76] Heparin-coated circuits are the most well-developed and widely available commercially. The first commercially available heparin-coated system was developed by the company Carmeda in 1983.[77] In a systemic review, heparin-coated systems have shown reduced contact activation of coagulation, complement system activation, alteration of granulocytes, inflammation, activation of platelets, disturbance of homeostasis, loss of blood, and cerebral damage compared with uncoated systems.[77] However, the impact of heparin-coated circuits on clinical outcomes remains undetermined. NO, an endothelial-derived relaxing factor, inhibits collagen-induced platelet activation. The most recent NO coating consists of a lipophilic-NO donor complex within polyvinyl chloride polymers for a more controlled NO release in the blood.[78] The major limiting factor for NO-coated circuits is a short shelf-life of fewer than 4 weeks. Currently, NO-coated systems are commercially unavailable. Gerling et al have reported the development of C1-esterase inhibitor-based circuit coating, inhibiting complement and factor XII activation to minimize circuit-associated thromboinflammation.[79] Yu et al have reported decreased clot burden and fibrinogen conservation by using argatroban-coated circuit in an in-vitro model.[80]
Biopassive surfaces include albumin, phosphorylcholine, poly-2-methoxyethyl acrylate (PMEA), polyethylene oxide, and poly MPC-co-BMA-co-TSMA.[69] Albumin coating, available since 1980, creates a hydrophilic base layer on the artificial circuit to reduce thromboinflammation by preventing the binding of platelets, leukocytes, and coagulation proteins on the circuit. Many manufacturers use albumin as part of a multilayer, bioactive coating alternating with a heparin layer ([Table 2]). Phosphorylcholine is a hydrophilic polar head of phospholipids containing negatively charged phosphate bound to positively charged choline. It exerts antithrombotic activity via reducing platelet activation and protein adsorption. Similarly, PMEA with a hydrophobic polyethylene backbone can decrease thromboinflammation by preventing platelet activation and coagulation protein adsorption. It is unclear if above-mentioned biopassive coatings are comparable to or better than commercially available heparin-coated circuits in improving clinical outcomes by reducing circuit-associated thromboinflammation.
Endothelization of the surface by in-vitro pre- or in-vivo self-endothelization has been attempted in synthetic vascular grafts, stents, and tissue-engineered vessels.[81]
[82] However, there are many drawbacks, as the completion of endothelization of the different stents can take months to years. In synthetic vascular grafts, there is insufficient endothelial cell migration and proliferation to the middle section of the lumen. In-vitro endothelization techniques using surface modification molecules such as fibronectin are being developed to improve endothelial cells adhesion, expansion, and migration. The tenuous endothelization techniques are further limited by a long culture time, risk of contamination and infection, and cost-effectiveness.
Overcoming the circuit-associated thromboinflammation by using surface modifications could expand the use of ECMO support without the risk of bleeding or thrombotic complications and minimize the need for systemic anticoagulation. Hopefully, in near future, this will make possible long-term ECMO support without disturbing the patient's blood hemostasis or inflammatory system.
The Future of Anticoagulation in Extracorporeal Membrane Oxygenation
The Future of Anticoagulation in Extracorporeal Membrane Oxygenation
During ECMO a delicate anticoagulation therapy is needed to prevent thrombotic occlusions within the oxygenator and extracorporeal circuit while balancing the bleeding risks in the patients. Despite being associated with excess bleeding, heparin has remained the most commonly used anticoagulation agent in ECMO. A particularly challenging aspect in heparin therapy is the supplementation with AT that is required for heparin anticoagulant activity. Furthermore, monitoring heparin dosing by the aPTT, thrombin time or ACT is error prone as the endogenous factors that drive clotting in these coagulation assays are consumed to unknown levels. Balancing anticoagulation therapy during ECMO has remained so challenging that due to coagulopathy and/or life-threatening bleeding, a considerable portion of patients does not receive any anticoagulation treatment at all. In a recent French cohort of coronavirus disease 2019 (COVID-19) patients supported by ECMO, both thromboembolic events and bleeding incidence were high and associated with mortality, supporting the urgent need to optimize anticoagulation strategies during ECMO.[83] Starting from the seminal observation that thrombus formation, but not hemostasis, is defective in factors XII- and XI-deficient mice,[84] an array of experimental and preclinical data have shown that some coagulation mechanisms principally differ in thrombosis and hemostasis. The plasma contact system proteins factors XII and XI, high molecular weight kininogen, and plasma kallikrein initiate the “intrinsic” coagulation pathway. Inherited deficiency in contact system proteins is not associated with increased bleeding in human and animal models. However, and challenging the classical concept of a coagulation balance, these contact system factors critically contribute to thrombosis, despite the fact that they have a minor if any role in hemostatic coagulation mechanisms.[85] The novel data indicate that pharmacological agents that target the contact system including its endogenous activator the inorganic polymer polyphosphate provide the opportunity for safe anticoagulation that in sharp contrast to current antithrombotic therapies is not associated with an increase in bleeding.[86] Indeed, in an experimental murine stroke model, the peptide-based factor XII inhibitor PCK (H-D-Pro-Phe-Arg-chloromethylketone) potently interfered with fibrin formation and vascular occlusion upon ischemia/reperfusion injury and, however, did not increase bleeding from injury sites in inhibitor-treated mice.[87] PCK has limited specificity and is rapidly cleared from circulation stressing development for improved factor XII inhibitors. The factor XII neutralizing antibody 3F7 specifically blocks the activated coagulation factor with high affinity in the nanomolar range. In a preclinical ECMO model, in rabbits, 3F7 infusions provided long-term thromboprotection with similar efficiency as compared with heparin.[88] In contrast to heparin treatment, bleeding times and blood loss from injury sites in 3F7-infused rabbits were not increased and similarly to saline-treated controls. The anticoagulant activity of 3F7 was so potent that the factor XII inhibitor largely blocked pathologic coagulation and thrombosis in the oxygenator despite the fact that the gas exchanging capillaries in the artificial lung were not coated with heparin and therefore highly procoagulant. In contrast to currently used anticoagulant drugs that interfere with thrombosis in a dose-dependent manner, thrombo-protection conferred by new contact system-neutralizing agents requires a substantial reduction of factor activities to <25%. Therefore, monitoring of new anticoagulants that target contact system proteins requires novel diagnostic assays to ensure patient safety.[89] In addition to safe anticoagulant targeting factor XII provides anti-inflammatory activities by interfering with the formation of the proinflammatory mediator bradykinin. In a recent phase 3 clinical trial, prophylactic therapy with the 3F7 antibody derivative CSL312 (also called Garadacimab) potently interfered with bradykinin-mediated tissue swellings.[90] Together, pharmacologic targeting of factor XII provides the exciting opportunity for safe anticoagulation in ECMO and possible other indication with additional antithrombo-inflammatory beneficial effects.
Conclusion
In the past 50 years, the technology of ECMO has advanced dramatically. To further mitigate risks of bleeding and thrombosis during ECMO and improve outcomes, the pursuit of better anticoagulants, better monitoring methods, and better ECMO circuits that include tubing, pump, and oxygenator will continue. However, it should be emphasized that no matter the advances in technologies, there are no substitutes for a multidisciplinary team approach that includes experts from various subspecialties in the care of these critically ill patients.