The Society of Thoracic Surgeons, The Society of Cardiovascular Anesthesiologists, and The American Society of ExtraCorporeal Technology

Clinical Practice Guidelines—Anticoagulation During Cardiopulmonary Bypass

Linda Shore-Lesserson, MD; Robert A. Baker, PhD, CCP; Victor A. Ferraris, MD, PhD; Philip E. Greilich, MD; David Fitzgerald, MPH, CCP; Philip Roman, MD, MPH; John W. Hammon, MD


Anesth Analg. 2018;126(2):413-424. 

In This Article

Heparin Dosing for Initiation and Maintenance of CPB

Class I Recommendation

  • A functional whole blood test of anticoagulation, in the form of a clotting time, should be measured and should demonstrate adequate anticoagulation before initiating and at regular intervals during CPB. (Level of Evidence C)

Class IIa Recommendations

  • Bolus administration of unfractionated heparin based on weight is reasonable for achieving adequate anticoagulation, but individual response to heparin is heterogeneous and requires a therapeutic functional test of clot inhibition before initiation of CPB, independent of the bolus dose used. (Level of Evidence C)

  • It is reasonable to use activated clotting time (ACT) tests that produce "maximally activated" clotting times as these tests mitigate ACT variability, are less susceptible to hypothermia, and correlate more closely with factor Xa activity compared with tests that use a single activator. (Level of Evidence B)

  • It is reasonable to maintain activated clotting time above 480 seconds during CPB. However, this minimum threshold value is an approximation and may vary based on the bias of the instrument being used. For instruments using maximal activation of whole blood or microcuvette technology, values above 400 seconds are frequently considered therapeutic. (Level of Evidence C)

Class IIb Recommendations

  • Use of a heparin dose-response formula may identify reduced sensitivity to heparin, but has not been shown to be more useful than weight-based heparin dosing in determining the heparin dose required to achieve an adequate ACT for initiation of CPB. (Level of Evidence B)

  • Use of heparin concentration monitoring in addition to ACT might be considered for the maintenance of CPB, as this strategy has been associated with a significant reduction in thrombin generation, fibrinolysis, and neutrophil activation. However, its effects on postoperative bleeding and blood transfusion are inconsistent. (Level of Evidence B)

  • During CPB, routine administration of unfractionated heparin at fixed intervals, with ACT monitoring, might be considered and offers a safe alternative to heparin concentration monitoring. (Level of Evidence C)

Activated clotting time is considered the gold standard in monitoring anticoagulation for CPB. The establishment of a safe or optimal range for ACT dates back to data published in the 1970s when Bull and colleagues[3] showed no development of clot in the oxygenator or circuit when ACT was maintained above 300 seconds. However, Young and coauthors[4] challenged this threshold when they demonstrated fibrin formation in the circuits of rhesus monkeys maintained on CPB with a minimum ACT value of 300 seconds, and they recommended that this threshold value be increased to 400 seconds by showing it was safe in 5 pediatric patients on CPB. To maintain a margin of safety above 400 seconds, the minimum acceptable ACT value of approximately 480 seconds became a "standard of care" that was used in numerous future studies and in clinical practice, but was based on limited evidence. Despite this widely accepted level of anticoagulation, there is no clear consensus on the accurate calculation of this initial dose of unfractionated heparin. Options for calculating the initial heparin bolus include a fixed, weight-based dose, (eg, 300 IU/kg), or use of point-of-care tests that measure the whole blood sensitivity to heparin using an associated dose response.

In addition to the heterogeneity of heparin formulations themselves, individual responsiveness to heparin is variable. The pharmacodynamics of unfractionated heparin are highly dependent on the level and function of plasma antithrombin III. In patients with preoperative hypercoagulability or reduced antithrombin III responsiveness, increased levels of circulating heparin are necessary to achieve a therapeutic ACT value before CPB.[5] Na and coauthors[5] reported significant variations to heparin responsiveness in an observational study of patients with known, stabilized infectious endocarditis. Garvin and associates[6] also reported observed variations in heparin response in patients having CPB. In a retrospective institutional database review of 3,880 patients, these investigators found wide variation in the heparin bolus dose required to obtain a target ACT. The initial unfractionated heparin bolus dose did not correlate well with the first post-heparin ACT (R 2 = 0.03).

The route and timing of the initial administration of unfractionated heparin has a direct impact on the ability to obtain a therapeutic ACT. A small randomized trial done by Grima and colleagues[7] found that intermittent doses of unfractionated heparin administered before CPB (100 IU/kg for 3 doses) maintained adequate levels of anticoagulation during CPB better than a single bolus dose of 300 IU/kg. Intermittent pre-CPB heparin treatment resulted in lower mean decreases in factor VIII, fibrinogen, antithrombin III, and platelet count than if a large bolus dose were administered. In a prospective nonrandomized trial performed by Neema and colleagues,[8] 6 of the 100 patients who received 300 IU/kg of unfractionated heparin before CPB had a resultant post-heparin ACT less than 350 seconds. Other pathologic disturbances such as thrombocytosis may limit the effectiveness of weight-based heparin bolus administration.

Owing to the heterogeneity of the pharmacodynamic response to unfractionated heparin, the utilization of ex vivo heparin dose-response technologies was studied as a more accurate prediction of initial heparin dosing. Although ex vivo heparin dose-response technologies may identify patients who have a reduced sensitivity to conventional doses of heparin, these tests have limited ability to calculate correctly an optimal initial unfractionated heparin bolus dose. The observational study by Garvin and colleagues[6] demonstrated poor correlation of the calculated in vitro heparin dose response curve compared with the actual patient heparin dose response, resulting in a failure to reach therapeutic ACT values in nearly 17% of patients. During CPB, an overestimation of heparin concentration may occur when using the ACT assay alone. Falsely elevated ACT values may be observed under conditions of hypothermia, reduced hemoglobin concentration, hypofibrinogenemia, and pharmacologic agents that are not associated with a concomitant increase in heparin concentration.[9] In a controlled, nonrandomized study of 42 patients, Machin and colleagues[10] demonstrated prolongation of ACT values during hypothermic CPB when compared with normothermic CPB. Leyvi and colleagues[11] reported similar ACT prolongation under conditions of both hypothermia and hemodilution using a number of ACT technologies while plasma antifactor Xa heparin level activity remained constant. Maintaining ACT values during CPB without heparin concentration monitoring may result in lower doses of heparin. These known sensitivity limitations in ACT monitoring may result in subclinical plasma coagulation occurring during CPB.

Whole blood heparin concentration assays are statistically more closely correlated with plasma anti-Xa levels than the ACT.[12] Clinically, heparin concentration tests are performed alongside a functional test of clotting, such as an ACT, because a therapeutic functional confirmation of anticoagulation provides important safety data. In a randomized controlled trial of 200 patients, Koster and colleagues[13] found that adhering to a heparin concentration maintenance protocol led to a significant reduction in thrombin generation, fibrinolysis, and neutrophil activation, when compared with ACT monitoring alone (480 seconds). Despotis and associates[14] randomized patients to ACT-based (using 5,000 units unfractionated heparin doses to maintain ACT values >480 seconds) versus heparin concentration-based management (with minimum ACT >480 seconds), and reported a higher heparin total dose in patients in the heparin concentration group (612 ± 147 U/kg versus 462 ± 114 U/kg, p < 0.0001). Patients in the heparin concentration group also had lower protamine to heparin ratios and required significantly fewer blood product transfusions (platelets, plasma, and cryoprecipitate) than the ACT-based control group. Another randomized trial of 31 patients scheduled for reoperation resulted in significant reductions in perioperative blood loss and blood product usage when maintaining higher patient-specific heparin dosing during CPB.[15] Another study found reduced platelet activation and evidence of reduced thrombin generation with heparin concentration monitoring compared with routine ACT monitoring.[16] Together, these studies suggest that whole blood heparin concentration monitoring results in larger doses of unfractionated heparin during CPB and improved hemostatic suppression compared with ACT monitoring alone. However, these results did not translate into improved clinical outcomes and have not been wholly reproducible in the literature. A retrospective analysis involving 686 patients favored ACT-based monitoring compared with heparin concentration monitoring because of less postoperative bleeding and transfusion requirements associated with ACT-based monitoring.[17]

Traditionally, the gold standard for measuring the anticoagulant effects of heparin is inhibition of factor Xa (anti-Xa) activity. Factor Xa is a major target for unfractionated heparin and can be readily measured in plasma using laboratory assays. The various studies that seek to validate a new measure of heparin activity, or a clotting time assay, use anti-Xa activity as the gold standard comparison. However, plasma assays for anti-Xa activity are not ideally suited for point-of-care testing. Anti-Xa measurement serves as a validating test for novel point-of-care assays that reflect anti-Xa activity. Hansen and associates[18] studied a whole blood modified ACT test and found it to be highly correlated to laboratory anti-Xa measurement. Helstern and associates[19] reported another one-step clotting assay that correlates well with anti-Xa tests and is not influenced by hemodilution, but clinical studies are lacking.

Routine redosing of unfractionated heparin at fixed intervals during CPB, despite a therapeutic ACT, is commonly used when heparin concentration assays are not available to simulate the practice of "higher heparin dosing." This practice prescribes additional fixed doses of unfractionated heparin at specific timepoints, even though ACT may be above target. In a prospective trial of 100 patients presenting for cardiac surgery, one third of the initial heparin bolus was administered at the 90-minute point of CPB, with repeat doses every 60 minutes thereafter.[8] This strategy maintained adequate anticoagulation during the entire period of hypothermic CPB; bleeding variables were not reported.

Despite the reported benefits of higher heparin dosing, other studies seemingly contradict these results. In a small prospective trial of 21 patients, Gravlee and colleagues[20] concluded that subclinical plasma coagulation occurs during CPB despite heparin concentrations greater than 4.1 IU/mL. Furthermore, postoperative mediastinal chest tube drainage correlates with increased heparin concentration, especially if heparin rebound is not carefully monitored. A subsequent, prospective study of 63 patients by Gravlee and colleagues[21] showed that subjects who received an unfractionated heparin bolus of 400 IU/kg and had heparin concentration maintained greater than 4 IU/mL did not differ in mediastinal drainage or transfusion products from a control group of patients receiving a bolus dose of 200 IU/kg plus additional heparin for ACT values less than 400 seconds. A prospective trial in 31 patients undergoing cardiac surgery revealed that all patients had a residual circulating heparin level after protamine administration (mean 0.18 IU/mL), detected by a chromogenic anti-Xa assay. This residual heparin concentration did not correlate with ACT or whole blood heparin concentration nor did it correlate with postoperative mediastinal tube drainage volume.[22] Although the studies supporting higher unfractionated heparin doses are greater in size and number, the impact of using higher doses of heparin on postoperative bleeding appears to be unclear, especially if residual effects of heparin are not detected or treated.

Documented therapeutic anticoagulation treatment of patients having CPB is necessary and is routinely performed using an ACT. However, ACT devices vary considerably in their measurement platforms, activators, sample volumes, and sensitivities to external elements such as hemodilution, hypothermia, and concomitant drug therapies.[23,24] It appears that arterial versus venous blood sampling and a wait period as long as 15 minutes do not significantly affect the ACT result.[25,26] Currently there are many instruments and platforms available that purport to measure ACT values. To rationally utilize an ACT device for patients undergoing CPB, it is important to understand how the testing platform works, the therapeutic target that corresponds to a historical ACT of 480 seconds, and how well the results correlate with anti-Xa activity. In an early study of heparin monitoring and ACT threshold values, it was noted that the two most commonly used ACT devices correlated with each other, yet there was significant bias with one of the instruments.[27] Another observational study showed that many ACT tests correlated poorly with heparin level as assessed by anti-Xa plasma activity.[28] Patteril and associates[29] demonstrated that after switching their cohort population to a newer ACT device, the new instrument yielded a lower mean ACT value compared with temporal controls (557 versus 618 seconds, p < 0.05), and a higher dose of unfractionated heparin was needed to achieve a minimum ACT of 480 seconds.[29] A certain level of validation has been performed for other ACT instruments as well.[30]

Tests that use a maximal degree of activation of the blood sample by using multiple or more potent activators produce shorter clotting times relative to the standard ACT with a single activator.[11] The tests that utilize a maximally activated sample also report less variability in clotting times and are less susceptible to prolongation by hypothermia and artifacts.[10] The maximal activation removes the variability induced by hemodilution of clotting factors. Maximal activation is also accomplished in the microcuvette ACT technology owing to the small sample volume and minimization of sample dilution. A plasma supplemented ACT accomplishes a similar result. This test has been shown to mitigate the ACT variability to more closely mirror anti-Xa levels; however, it is cumbersome and difficult to perform at the point of care.[19]

The viscoelastic tests have been modified for point-of-care measurement of the ACT and in a small ex vivo analysis in CPB patients, the two tests performed similarly to standard ACT tests with respect to heparinization and hemodilution.[30,31] Another observational study involving 50 CPB patients demonstrated that a viscoelastic measurement of ACT activity mirrored the activity of both standard ACT tests and anti-Xa levels.[32] It remains uncertain what the threshold minimum safe values are for the viscoelastic clotting times in CPB and how they correspond with the historical 480 second target. Further clinical and outcome studies are warranted before switching patient management to a viscoelastic ACT test.