Unfractionated heparin (UFH) remains the most widely used anticoagulant for neonatal and pediatric mechanical circulatory devices, specifically ECMO (5). UFH is a sulfated polysaccharide that produces its anticoagulant effect by inactivating free thrombin and activated factor X (FXa) by binding to antithrombin (AT), which induce a confirmational change in AT structure and enhance its serine protease inhibiting activity by 500–1000 folds. Heparin binds to AT by a high-affinity pentasaccharide, which is present on about a third of heparin molecules. Thrombin inhibition is dependent upon heparin binding to both the coagulation enzyme and AT; whereas inhibition of FXa is dependent only upon binding to AT (6). Thrombin bound to fibrin or subendothelium is not inhibited by the AT-UFH complex. UFH also has the ability to bind to plasma proteins, platelet factor 4, high molecular weight von Willebrand factor (vWF) and endothelial cells independently of AT which results in a variable response to heparin and even heparin resistance (6). In addition, these proteins are released during increased inflammation and activation of the coagulation system, leading to inefficient anticoagulant activity. The possibility of heparin induced thrombocytopenia (HIT) as well as wide interpatient variability are major shortcomings of UFH. Despite this, the use of UFH is well-known to treating clinicians with a short half-life and easy reversibility. Therapeutic efficacy of UFH can be measured by the activated clotting time (ACT), activated partial thromboplastin time (aPTT), anti-factor Xa activity (anti-FXa) and VET.

AT is a serine protease inhibitor necessary for heparin to exert its anticoagulant effect. Some centers include routine monitoring of AT and replacement with either fresh frozen plasma or AT concentrates. The rationale for the replacement of AT is to improve the effectiveness of heparin and reduce thrombotic complications. However, target AT levels remain unclear and guidelines for replacement with fresh frozen plasma (FFP) or AT concentrates vary significantly across centers. AT use remains controversial, as there is the potential risk of bleeding and added cost. A Cochrane review of 20 trials using AT replacement in critically ill patients showed no significant effect on mortality in all subgroup populations, including children, but had a 1.5-fold increased risk of bleeding (7). Until larger clinical trials can establish the effectiveness of AT replacement in pediatric ECMO/VAD, AT concentrates should be used with caution.

The use of direct thrombin inhibitors (DTIs) is becoming increasingly more common in the ECMO/VAD population, even in non-HIT scenarios. DTIs inhibit thrombin irrespective of whether the thrombin is free, or fibrin bound. Bivalirudin is the most commonly utilized direct thrombin inhibitor in ECMO and VADs. Bivalirudin is eliminated primarily by a non-organ mechanism (proteolysis) with only a minor (20%) component of renal clearance. Bivalirudin has the shortest half-life (about 25 min) and also has an intermediate affinity to thrombin in comparison to lepirudin (highest affinity) and argatroban (lowest affinity) (8).

Though studies are limited, bivalirudin thus far has been used effectively in both pediatric ECMO and VAD patients. In a single-center retrospective cohort, Seelhammer et al. evaluated a total of 422 ECMO patients (89 pediatric) and did not find an increase in adverse outcomes in the pediatric population. In the adult subgroup, bivalirudin was associated with improved overall survival (9). There is no consensus on bolus or continuous infusion dosing of bivalirudin. Loading dose has been reported as 0.4–0.5 mg/kg and continuous infusion rates from 0.05 to 0.5 mg/kg/h (10–12) in ECMO patients, with higher dosing reported in VADs (13). Pediatric studies are limited with mostly retrospective single-center data or small case series in the ECMO literature. Some pediatric programs have reported using bivalirudin as a first line agent with good outcomes; however, high-quality studies are lacking (11, 13–15). In the Berlin Heart EXCOR and Advanced Cardiac Therapies Improving Outcome Network (ACTION) post-approval surveillance study report, 92% of patients in the post surveillance group (post-market approval of EXCOR VAD) were transitioned to bivalirudin and also had decreased bleeding events, improved survival, lower stroke, and fewer thrombotic events (16).

Antiplatelet agents are infrequently used during pediatric ECMO. The non-biologic surface of the ECMO circuit induces ongoing platelet activation and adhesion, which leads to platelet dysfunction (17, 18). Initiation of ECMO is associated with a decline in platelet count in neonates and pediatric patients. In a large multicenter study, the unadjusted association of average daily platelet count with mortality had a roughly linear association with the log odds of mortality up to a platelet count of approximately 115 × 10⁹/L. On multivariable analyses, average daily platelet count was not independently associated with mortality however platelet transfusion volume was associated with mortality. This suggests that the platelet count was not the reason for increased mortality but factors associated with platelet transfusion (19). There is a paucity of data about adjunct antiplatelet therapy in pediatric ECMO patients with hypercoagulable states. Recently, two hypercoagulable COVID-19 adults were supported with venovenous (VV) ECMO and thromboelastography-guided antiplatelet therapy was initiated with good outcomes (20). In a case series of three pediatric patients with the multisystem inflammatory syndrome in children (MIS-C) supported with venoarterial (VA) ECMO, aspirin was used pre-ECMO cannulation due to coronary dilation. Interestingly, the only patient not treated with aspirin pre-cannulation had early circuit thrombosis requiring circuit change despite similar elevation in inflammatory markers in all patients (21). However, the Extracorporeal Life Support Organization (ELSO) registry data suggests similar frequency of thrombotic events in adult COVID-19 ECMO patients compared to the overall ELSO registry (22). Thus, there is no evidence to support routine use of antiplatelet medications in ECMO, however they may be beneficial in select cases.

Antiplatelet agents have been routinely used in the VAD population. The Edmonton anticoagulation and platelet inhibition protocol was the first widely utilized detailed guideline for pediatric VADs and included treatment with aspirin and clopidogrel (23). Antiplatelet therapy was guided by TEG®-platelet mapping. A more aggressive triple weight-based antiplatelet strategy was employed by Rosenthal et al. and this group reported 84% lower rate of stroke then the Edmonton guideline (24). With the more widespread use of DTIs for pediatric VAD patients, aspirin and dipyridamole are the most commonly used anti-platelet agents (3, 25, 26). A recent report using data from 10 centers affiliated with ACTION, showed lower stroke and major late bleeding in patients anticoagulated with DTIs. In this study, 88% of patients were also treated with antiplatelet agents, but dual antiplatelet therapy was used in less than a third of patients (27).

The ACT is a whole blood, point-of-care test that measures the time it takes for clot formation using an intrinsic pathway activator. Advantageous to pediatrics, this point-of-care test is easy to perform with fast results and requires only a small amount of blood. Initially, this test was used during cardiopulmonary bypass where heparin dosing is much higher than what is used during ECMO. During ECMO, the dosing of heparin is lower and there has been poor correlation to the ACT (29). In conjunction, it is important to remember that the duration of ECMO is much longer (days to weeks) than that of cardiopulmonary bypass (hours) leading to a sustained hyperinflammatory state which may further interfere with anticoagulant monitoring. The ACT is also impacted by several other factors including hemodilution, hypothermia, coagulation factor deficiencies and platelet dysfunction, all of which may occur during ECMO therapy (30). Previous adult and pediatric ECMO studies have reported that the ACT had no correlation to UFH dose and was relatively insensitive to changes in UFH dose (31, 32). Despite these limitations, the ACT remains in broad use in up to 68 to 97% of institutions (5, 33). The most commonly reported and recommended goal ACT for ECMO is 180–220 s (28, 34, 35), but it is unclear if this reflects adequate anticoagulation for pediatric patients (36–38).

The aPTT is a plasma-based assay that evaluates the intrinsic and common factor pathways (factors XII, XI, IX, X, V, II and fibrinogen). This test is used to monitor anticoagulation effects of both UFH and direct thrombin inhibitors. aPTT testing is inexpensive and has the ability to detect underlying factor deficiencies (inherited or acquired), vitamin K deficiency or the presence of disseminated intravascular coagulation. However, the sensitivity of aPTT to heparin effect is both analyzer and method dependent and thus each laboratory must have their own determination of normal range. These reference ranges may change frequently due to variations in aPTT reagents between lots (39). Baseline aPTT vary significantly with age due to developmental hemostasis, and thus neonates and young children have different reference ranges than adult patients. In addition, there is an age-dependent effect of UFH on the prolongation of the aPTT (40). Generally, programs target an aPTT goal of 1.5–2.5 times baseline, which is based upon the normal range of the anti-factor Xa assay at most centers. Interestingly, there is a wide variation of reported aPTT values (in pediatric ECMO and VAD) that fall in the therapeutic window of anti-Xa (28, 40, 41). Considering developmental hemostasis in this age group and wide variation of aPTT ranges, using the aPTT for monitoring alone may be insufficient.

Anti-FXa is another plasma-based assay that determines the anticoagulant activity of UFH by measuring the ability of heparin-bound AT to inhibit FXa. This assay is a photometric test; thus, it is important to note that hyperbilirubinemia, elevated triglyceride and heme levels interference may cause falsely low results. In addition, AT levels will influence the assay and results may also be inaccurate if the laboratory adds exogenous AT to the specimen (28). This is particularly important in neonates and young children, as these differences are likely to be more prominent (42). The main advantage of anti-FXa is that it directly measures the heparin effect rather than heparin concentration. Therapeutic levels are considered to be 0.3–0.7 IU/ml. Specifically, in mechanical circulatory devices, anti-FXa has been shown to correlate better with heparin dosing than the ACT (31), and in one study, an increase in heparin by 10 units resulted in a reproducible increase in anti-FXa level by 0.07 IU/ml (28). Thus, anti-FXa was better correlated with heparin effect than ACT which is poorly correlated with heparin dosing (32). Unfortunately, discordance between aPTT values and anti-FXa is common in pediatric ECMO and makes interpretation of these labs challenging (31).

As previously discussed, some centers routinely monitor AT levels. AT has been reported to increase by about 1% for each ECMO day and is not related to daily volume of FFP transfusions. (28). Bembea et al. found that the median baseline AT activity for which patients received treatment was 43%, range 38–64%. AT levels were inversely correlated with the ACT, even after adjustment for heparin dose, but had a strong positive correlation to anti-FXa level (28).

The dilute thrombin time is a plasma-based assay that is used to determine the anticoagulant activity of DTIs, like bivalirudin. Here patient plasma is diluted with human fibrinogen or normal human plasma and then clotting is initiated by the addition of thrombin; the clotting time is then measured and compared to a laboratory generated standard reference curve (43). A recent study demonstrated that the aPTT and DTT have poor correlation in samples from pediatric VAD patients (44). Because the DTT is not affected by heparin, lupus anticoagulants or vitamin-K dependent factors it may be superior for monitoring anticoagulant activity; however, studies are limited and it remains unclear if the DTT is preferred over the aPTT for monitoring (43).

Most commonly, the samples obtained for VET are collected in sodium citrate tubes which necessitates timely processing. Standard values have been reported in healthy children and in children with congenital heart disease (51, 52). However, it has been reported that the use of samples collected in this way may result in significant artifacts on TEG® analysis and could lead to heparin overdosing in these patients (53). In addition, discrepant results have been reported using TEG® 5000, TEG® 6S and ROTEM® on the same whole blood sample in orthotopic liver transplant and trauma patients; although all three tests could identify abnormalities in clot formation (50, 54). If TEG® 6S provides an accurate assessment of coagulopathy in the ECMO and VAD population, its use at the bedside would allow for point of care results and decreased need for skilled laboratory personnel.

Another benefit of VET testing is the value in patients anticoagulated with DTIs. Dilute thrombin time is not readily available in many institutions and standard laboratory monitoring may not accurately reflect anticoagulant effect of DTIs. TEG® ecarin clotting time has shown excellent correlation with chromogenic bivalirudin levels, r² = 0.746 (55). In a pediatric study including 106 blood samples of children supported with ECMO on bivalirudin, prolongation of clotting time on ROTEM intrinsic coagulation pathway (INTEM) and INTEM with heparinase (HEPTEM) showed moderate to strong correlation with aPTT and hepzyme aPTT suggesting that ROTEM may provide a good alternative to those standard assays. In this report, the target range of INTEM CT was 260–300 s for children without evidence of heparin-like activity anticoagulated with bivalirudin during ECMO (56).

Management of anticoagulation is predominantly assessed by using the R-time (TEG®) or CT (ROTEM®) and comparing the values with and without the presence of heparin. Reference ranges are based upon individual lab standardization. In an adult randomized controlled trial, anticoagulation using aPTT based protocol compared to TEG® based protocol targeting TEG®-Kaolin R-time 16–24 min (normal value: 4–8 min) found that TEG based titration was safe, feasible, and resulted in lower heparin dose compared to aPTT based protocol (57). A pediatric retrospective study found that anti-FXa activity > 0.25 IU/ml (sensitivity 81%, specificity 67%, PPV 81%, and NPV 58%) and a TEG® R-time > 17.85 min (sensitivity 84%, specificity 68%, PPV 82%, and NPV 59%) may minimize the risk of thrombosis in pediatric and neonatal ECMO patients but could not find an optimal target to minimize bleeding (58). In general, programs target somewhere between 2 and 4 times the baseline TEG® R-time. In the event of incomplete correction of the R/CT time with the addition of heparinase, it may suggest coagulation factors deficiency and consideration may be given to replacement with FFP based on presence or absence of clinical bleeding. In addition, some patients develop acquired von Willebrand deficiency or other clotting factor deficiencies due to consumption in the circuit and abnormalities can prompt additional laboratory testing to identify deficiencies. Clot strength as assessed by the MA/MCF is influenced by fibrinogen, absolute platelet count and function. Increased lysis time may be indicative of hyperfibrinolysis (circuit disseminated intravascular coagulopathy or acquired factor XIII deficiency). Importantly, viscoelastic tests may help diagnose covert coagulopathy that may not be identified on standard laboratory assays and could contribute to bleeding risk (59).

Although limited, studies evaluating the use of VET in adult ECMO and VAD have increased over the last decade. Pediatric ECMO/VAD studies have reported VET-based anticoagulation algorithms and associations with clinical outcomes including bleeding and thrombosis. Figures 3, 4 describe the clinical features and TEG®/ROTEM® parameters that have been associated with increased bleeding and thrombotic complications. In a recent report using data from 265 children in the Pediatric ECMO Outcomes Registry (PEDECOR) VET was utilized in 34 of centers contributing data, although bleeding and thrombotic complications were not different between centers utilizing VET compared to centers that did not (60). We review these pediatric studies below and in Table 3.

While the use of VET in pediatric ECMO/VAD has increased, limitations exist. Importantly, technical considerations, including skills to properly maintain the device and expert proficiency in handling blood samples, are needed for reliable results. After sample procurement, activation of the sample is performed. In TEG® and ROTEM, contact activation (intrinsic coagulation pathway) using ellagic acid or kaolin can provide information like activated partial thromboplastin time. While tissue factor activation (extrinsic coagulation pathway) provides information like prothrombin time. Lyophilized heparinase neutralization of heparin, blocking of platelet contribution to clot formation with platelet inhibitors, or inhibiting fibrinolysis with antifibrinolytics can be used. Thus, activators for both the intrinsic and extrinsic systems are available. When celite and kaolin are used as activators for TEG® in the presence of aprotinin, there is a prolongation of R time, a decrease in α angle, and a decrease in MA, suggesting that these are not ideal activators for a TEG®-guided transfusion algorithm (88). TEG® has a poor correlation with frequently used laboratory testing and has not been shown to predict bleeding or thrombosis during ECMO. Currently, the cut-off values based on TEG® and ROTEM-based parameters are not well established for pediatric patients. In addition, results should be interpreted by an experienced clinician who understands the intricacies of the test and variables affecting the hemostatic system. Today, the amount of blood required to run the VET is small. Still, it may be problematic in the neonatal population, especially those awaiting cardiac transplantation, where it is essential to limit blood product exposure. Another limitation is that von Willebrand deficiency cannot reliably be detected using VETs. In addition to all of these factors, the child's clinical status is the most critical variable, and decisions based on VET/laboratory tests may not reflect the dynamic changes in the critically ill child.