Thromboelastography-based anticoagulation management during extracorporeal membrane oxygenation: a safety and feasibility pilot study

During extracorporeal membrane oxygenation (ECMO), systemic anticoagulation is routinely used to avoid thrombin generation due to the contact of blood with non-biological surfaces of the extracorporeal circuit. Despite this, some degree of coagulopathy can be detected almost immediately after the start of the extracorporeal circulation [1, 2]. This may results in both bleeding and thrombotic events, which are the most frequent adverse effects during ECMO jeopardizing the final outcome [3, 4]: whether hemostasis-related complications could be attributed to: (1) the underlying patient conditions (such as sepsis or inflammation); (2) the effect of anticoagulant drugs; or (3) the suboptimal anticoagulation monitoring, is still matter of debate.

So far, we do not have many answers regarding the first two issues. Interactions between coagulation, inflammation and immune pathways are increasingly recognized but still represent an open field of research [5, 6] and unfractionated heparin remains the anticoagulant drug of choice in most ECMO centers [7, 8]. Anticoagulation monitoring methods are also far from being ideal [9]. Activated partial thromboplastin time (aPTT) is universally recognized as the standard assay for monitoring heparin therapy [10, 11] and it is considered a good choice in the adult population treated with ECMO [7]. However, aPTT is performed on plasma samples and may not provide a comprehensive picture of the hemostatic profile in the absence of the cellular components of the blood. In contrast, thromboelastography (TEG) is a whole blood viscoelastic point-of-care test, which provides information regarding the entire coagulation cascade (including platelet function, platelet–fibrin interactions with red blood cells and fibrinolysis). At present, TEG has been recommended only to manage coagulation abnormalities during ECMO [12]. Nevertheless, the TEG reaction time (the time needed to change the physical nature of blood from liquid to gel) could act as a surrogate for thrombin generation and drive heparin infusion [13‐15]. In a retrospective study on 32 consecutive patients treated with ECMO for severe respiratory failure, we frequently observed a marked heparin effect on the TEG tracing despite an aPTT ratio and activated clotting time (ACT) within the therapeutic anticoagulation range [16]. Aware of these findings, since excessive anticoagulation remains a reasonable risk factor for bleeding during ECMO [17, 18], we raised the question as to whether using a TEG-driven strategy to titrate heparin during ECMO could represent a safe and feasible alternative to the “conventional” approach based on aPTT monitoring.

We designed the present pilot study to compare two protocols for managing heparin therapy during veno-venous ECMO: one based on thromboelastography and the other based on aPTT ratio, in a population of patient with severe respiratory failure.

Between September 2, 2014, and November 30, 2016, we screened 64 patients; of these, 42 were enrolled (21 in the TEG group and 21 in the aPTT group). (In Fig. 2 reasons for exclusion of the remaining patients are depicted.) Baseline clinical characteristics and coagulation parameters of the patients are summarized in Table 1. There appeared to be a degree of imbalance only in the proportion of women enrolled in the two groups. There was no difference in baseline coagulation parameters; only one patient for each group presented with DIC at study entry. C-reactive protein was not different in the two groups at study entry: 16.2 (8.2–24.1) mg/dL in the TEG group and 14.5 (6.7–25.2) mg/dL in the aPTT group, p = 0.63.

Table 2 shows blood and coagulation parameters during the study period. Heparin infusion rate, aPTT ratio, R-K and anti-Xa were lower in the TEG group compared to the aPTT group.

As for the feasibility, the number of protocol violations was small in both groups, indicating that the algorithm was easy to follow. Interestingly, the number of analyses in the target range was lower in the TEG group, resulting in a higher number of heparin dosing adjustments. This may be explained by the sensitivity of TEG (even in the early phases of clot formation) to low heparin dosing [11] and to fibrinogen and platelet levels [22].

As expected from previous research [16], our TEG-based protocol resulted in the prescription of a much lower amount of heparin, even inferior to usual prophylactic doses. This was confirmed by significantly lower values in the TEG group of anti-Xa activity, which is considered the gold standard assay to measure heparin effect; importantly, in our study this test was performed without adding antithrombin to the plasma sample. It must be noted, however, that anti-Xa values were at the lower end of the recommended level for proper anticoagulation [12, 23] even in the aPTT group. Nevertheless, the number of thrombotic episodes was small and not different between the study groups.

Since this was the first time a TEG-based protocol was used to adjust heparin infusion, we planned a safety rule of not increasing heparin dose in the TEG group if the corresponding aPTT was 2.5 or higher. Yet, this never happened.

One of the major concerns of using a low level of anticoagulation during ECMO is the development of consumptive coagulopathy [1, 24]; this was not the case since D-dimer levels were comparable in the two groups. Importantly, we tried to standardize the timing of circuit exchange by means of daily calculation of a “circuit change score” based on coagulation alterations, heparin requirements, visual evaluation of clotting in the system and functional ECMO parameters. This score together with the evidence of clinically relevant coagulation alterations (i.e., bleeding), prompted the decisions to exchange the circuit. This may have prevented the development of a severe form of coagulopathy. However, score values and the number of exchanged ECMO system were similar in the two groups.

Since excessive anticoagulation increases the risk of bleeding during ECMO [17] and advances in technology (i.e., circuit coating with biocompatible surfaces) are supposed to reduce the incidence of circuit thrombosis, several reports of minimal anticoagulation during ECMO have been published. In a retrospective study by Agerstrand et al., incidence of bleeding was reduced and circuit or oxygenator thrombosis was rare, using a low target aPTT [25]. Yeo et al. described a lower incidence of bleeding and blood products transfusions in patients anticoagulated according to a lower than conventional ACT target [26]. Recently, Krueger et al. reported that a prophylactic regimen with only subcutaneous enoxaparin was feasible and safe in 61 patients on veno-venous ECMO [27]. Anyhow, due to the high variability in patients’ response to heparin [28], we believe in the added value of monitoring anticoagulation also when low levels of heparin are used. In our study, we confirmed that R-K is sensitive to low doses of heparin, but it loses accuracy at recommended anticoagulant ranges (i.e., anti-Xa 0.3–0.7 IU/mL) [23, 29]. Interestingly, the frequency of flat-line TEG tracings was lower in this study compared to our previous experiences [16]. This may be explained by a strict and prospective adherence to the protocol for heparin adjustment, which was not the case in our previous retrospective study.

Since the anticoagulant effect of heparin is mediated mainly by antithrombin [10], we supplemented antithrombin to achieve an activity level above 70%. We observed a tendency for lower antithrombin levels in the aPTT group, probably explained by increased consumption mediated by higher administered heparin dose.

Our study revealed that the use of TEG resulted in a higher costs compared to the aPTT group. Performing a TEG in fact is more expensive than requesting an aPTT from the laboratory, at least in the two facilities where the study was conducted. Notwithstanding, this was balanced in part by the lower expenses for the anticoagulant.

Our study has several limitations. First, it is a pilot study on a relatively small number of patients. Second, the target ranges for anticoagulation applied in both groups were arbitrarily chosen. Indeed, anticoagulation targets are different between ECMO centers [8]. We used an aPTT range between 1.5 and 2 times control, which is considered adequate in other settings [28] [30, 31] and that was common practice in the two participating institutions; however, it has never been validated during ECMO. Likewise the target R-K time of 16–24 min (equivalent to 2–3 times the upper limit of normal values) had not been validated before. Both of them may require to be better defined. Third, the use of FFP and antithrombin, the target hemoglobin, fibrinogen and platelets is not standard across centers. Hence, the results of our study might have also been influenced by the contributions of these variables.

We found that using a TEG-driven protocol seems to be feasible to manage heparin anticoagulation during VV-ECMO and does not look to be associated with an increased rate of complications. Furthermore, even if our study was not powered to detect major differences in outcomes, the TEG protocol allowed the administration of lower heparin doses without an increase in thrombotic complications and with a tendency for less bleeding compared to the aPTT protocol. A larger trial is warranted to confirm these preliminary findings.