Extracorporeal Circulation Without Bleeding

See allHide authors and affiliations

Science Translational Medicine  05 Feb 2014:
Vol. 6, Issue 222, pp. 222fs7
DOI: 10.1126/scitranslmed.3008497


A newly created antibody to the active site of Factor XIIa protects against clotting of extracorporeal bypass circuitry without bleeding risk in vivo, obviating the need for heparin (Larsson et al., this issue).

A cardiovascular surgeon has three requirements for performance of surgery with extracorporeal blood oxygenation [that is, circulation and oxygenation outside the body, such as in cardiopulmonary bypass (CPB) or extracorporeal membrane oxygenation (ECMO)]: (i) Instant anticoagulation of blood with minimal platelet and leukocyte activation; (ii) real-time anticoagulation monitoring for intraoperative adjustments; and (iii) the ability to stop anticoagulation instantly when the surgery is completed. No currently available anticoagulant meets all of these requirements, and any new agent must be assessed against these parameters. Now, Larsson et al. describe an antibody to activated factor XII (XIIa) that allows blood flow in an artificial circuit without clotting or bleeding (1).


During surgery, extracorporeal perfusion changes the composition of blood by dilution and physical stress of blood interacting with an artificial material. No man-made material has the physiological anticoagulant nature of the intravascular compartment, and contact of the blood with the artificial surfaces of the extracorporeal circulatory system activates several plasma proteolytic systems that increase the risk of thrombosis by initiating the intrinsic coagulation, complement, kinin pathways (Fig. 1). For the past 50 years, unfractionated heparin (UFH) has been widely used as an anticoagulant to prevent thrombosis during CPB and ECMO (UFH targets are shown in Fig. 1). An inexpensive drug, UFH provides instant anticoagulation, a function that can be monitored in the operating room (OR) by measuring activated blood-clotting time (ACT), an surface-activated blood coagulation test. UFH anticoagulation is rapidly neutralized with protamine sulfate when patients are disconnected from the blood oxygenation apparatus. For these reasons, the surgical community feels comfortable using UFH.

Fig. 1. Thwarting thrombosis in extracorporeal circulation.

Shown are heparin and 3F7 targets in three biological pathways in plasma and on endothelial cell membranes in subjects receiving extracorporeal circulation. Blue lines represent the complement pathway, which is important for leukocyte activation and cytokine release during extracorporeal circulation; red lines represent the contact-activation pathway, with XII autoactivation on the artificial surface of the bypass circuitry, which leads to prekallikrein (PK) activation with XII-PK reciprocal activation and amplification. XIIa formation also promotes intrinsic blood coagulation activation via factor XI with downstream thrombin formation. Thrombin proteolyzes fibrinogen to make a fibrin clot. Excess thrombin formation leads to clotting of the circuitry; excess anticoagulation leads to excessive bleeding. Green lines characterize the kallikrein/kinin system, whereby plasma kallikrein cleaves high–molecular weight (HMW) kininogen to liberate bradykinin (BK). BK binds its receptors, B2R and B1R, which liberate tissue plasminogen activator (tPA) to initiate fibrinolysis, and nitric oxide (NO) and prostacyclin to produce vasodilation and platelet inhibition. BK fragment 1-5 [K-(1-5)] is a thrombin inhibitor. XIIf (βXIIa), Hageman factor fragment; B1R, bradykinin B1 receptor; B2R, BK B2 receptor; PAR, protease-activated receptor.


However, UFH is far from the ideal anticoagulation agent. UFH is a glycosaminoglycan isolated from mucosal tissues of pigs or cows, and recently, an impure preparation caused death in some recipients (2). Further, ACT monitoring of heparin activity is not precise, which contributes to bleeding in some patients receiving a CPB or ECMO. In addition, UFH use is associated with the development of antiheparin antibodies, which activate platelets and induce thrombocytopenia, vessel occlusion, and thrombosis. And recently, some patients suffered postoperative thrombocytopenia caused by antibodies to protamine sulfate (3).

Many alternatives to UFH have been considered for anticoagulation in extracorporeal circulation. Most agents have focused on anticoagulation by inhibition of thrombin or activated factor X (Xa) (Fig. 1). Recombinant mutated alpha-1-antitrypsin Pittsburgh is an extremely potent antithrombin agent that has been used in a simulated extracorporeal circulation model; however, this agent is an irreversible anticoagulant and thus dangerous for use in humans (4). Danaparoid sodium (DS), a mixture of hog-intestine glycosaminoglycans, has been used successfully in CPB in patients with active heparin-induced thrombocytopenia and thrombosis syndrome, but it is difficult to monitor DS levels in the OR, and the lack of an antidote requires natural clearance to remove the agent after CPB (5). The thrombin inhibitor bivalirudin has been used widely and successfully in patients undergoing CPB (6) and can be monitored during surgery with both direct thrombin assays and new ACT protocols. The drug’s short half-life and rapid renal clearance recommend its use when an UFH alternative is needed, such as with acute heparin-induced thrombosis–thrombocytopenia syndrome (HITTS) and emergency CPB surgery. Unfamiliarity is the limitation to bivalirudin’s acceptance in the cardiovascular surgical community.


Clearly, new drugs with fewer off-target effects are needed for extracorporeal circulation to prevent thrombosis risk without excess bleeding. Activation of the serine protease factor XII (XII) on the artificial surfaces of extracorporeal circuits is a major problem that initiates several proteolytic cascades in blood vessels. Contact of XII with the artificial surface of the bypass circuitry leads to autoactivation and formation of XIIa, which increases the risk of thrombosis by activating the intrinsic coagulation, complement, and kallikrein/kinin systems (Fig. 1). Activation of the intrinsic coagulation system leads to thrombin and fibrin formation with platelet activation, while activation of the complement and kallikrein/kinin systems contributes to inflammation, blood pressure alteration, and fibrinolysis.

Previously, XII was shown to influence thrombosis risk but not bleeding (hemostasis) in mice with the XII-encoding gene knocked out (F12−/− mice) (7); F12−/− mice are protected from ferric chloride–induced carotid artery thrombosis, contact activation–induced pulmonary emboli, and stroke. Later, a potent XIIa inhibitor, rHA-infestin-4, was shown to prevent fibrin formation and arterial thrombosis but has an off-target effect on activated factor X (Xa), which inhibits Xa cleavage of prothrombin (8). Another XIIa inhibitor, aprotinin, prevents blood loss associated with extracorporeal circulation. However, at concentrations required for efficacy in CPB, aprotinin promotes off-target inhibition of activated protein C, which leads to increased thrombosis in patients (9).

In this issue of Science Translational Medicine, Larsson et al. (1) present a study with a new antibody (3F7) directed to the active site of XIIa that allows blood flow without thrombosis in a rabbit model of ECMO. 3F7 was discovered by screening a phage library that encoded fragment antigen binding regions (Fab fragments) of human antibodies with plasma-derived human βXIIa (Hageman factor fragment), the catalytic domain of XIIa. The selected phage clones were sequenced and their protein products analyzed for binding to XIIa and βXIIa. Only Fab fragments that bound exclusively to βXIIa were reformatted as intact human IgG4 antibodies and expressed in human 293T cells.

The resulting antibodies were characterized by their ability to bind βXIIa with a high degree of specificity, and antibody 3F7 was selected for further study. In vitro, 3F7 inhibited βXIIa amidolytic activity, activated partial thromboplastin time (aPTT) and 3F7 ellagic acid–, platelet-, or long chain polyphosphate–, but not tissue factor–, induced thrombin-generation times. 3F7 also blocked contact activation–induced cleavage of high–molecular weight kininogen and XIIa complex formation with its plasma protease inhibitor, C1 inhibitor. When rabbit blood was flowed over a collagen matrix, 3F7 reduced thrombus formation at both venous and arterial shear rates with reduced fibrin, platelet, and leukocyte accumulation. The ability of adenosine diphosphate (ADP) and collagen to activate platelets from the 3F7-treated animals was reported as being unaltered, but more studies are needed on platelet qualitative function after circuitry with 3F7.

When administered directly to mice at concentrations ≥5 mg/kg, 3F7 blocked ferric chloride–induced carotid artery thrombosis with prolongation of plasma aPTT and without influence on the prothrombin time (PT) or tail-bleeding time. A single bolus infusion of 3F7 into rabbits did not increase ear, skin, or cuticle bleeding times in rabbits, and, similar to UFH (300 IU/kg), reduced thrombus weight and increased time to arterial-venous shunt occlusion.

In a rabbit model of extracorporeal circulation using an infant membrane oxygenator and roller pump (ECMO), blood flow through the circuit was maintained for up to 6 hours after intravenous administration of either 3F7 (7 mg/kg; n = 4) or UFH (n = 5; 50 IU/kg, a standard loading dose). Similar to UFH, 3F7 significantly prolonged ACT and aPTT (relative to untreated animals) without causing significant differences in plasma concentrations of prothrombin1+2 peptide, soluble fibrin, C3a, or hemoglobin, or in plasma pH or oxygen saturation. As observed with UFH, 3F7 treatment prevented the increase in ECMO circuit pressure that was observed in untreated animals. 3F7 infusion into rabbits did not prolong ear or kidney bleeding time on the ECMO circuit. Last, like heparin, 3F7-treated gas-exchanging capillaries in the oxygenators displayed a reduction in the physical presence of fibrin relative to untreated capillaries.

When UFH is used as an anticoagulant, it inhibits blood clotting at multiple points in the coagulation cascades, both in the vasculature and in blood flowing through artificial circulatory devices (Fig. 1). UFH functions as an anticoagulant by potentiating the plasma serine protease inhibitor (serpin) antithrombin to inhibit thrombin, factor Xa, and every other serine protease in the blood coagulation system, including XIIa and plasma kallikrein. Thrombin inhibition also helps to protect from platelet activation in the circuit as well. UFH also potentiates another serpin C1 inhibitor, which accounts for 90% XIIa and 45% plasma kallikrein inhibition, to reduce contact activation, bradykinin formation, and complement activation through both the classic and alternative pathways.

3F7 is much more parochial in its choice of targets (Fig. 1): It inhibits only XIIa and βXIIa. Even with this selective XIIa inhibition, Larsson et al. (1) show that 3F7 blocked XI activation and intrinsic blood coagulation in the plasma of 3F7-treated rabbits but not classic or alternative complement activation or kallikrein/kinin system activation. Thus, one can cogently conclude that 3F7 provides ECMO circuit thrombosis protection without anticoagulation.


How good is 3F7 compared to UFH? The data provided are on a limited number of animals, and when one critically inspects the data, one observes several results that caution enthusiasm. At five hours into circuit use, there was a significant increase in the PT-INR (prothrombin time–International Normalized Ratio) and d-dimer (16-fold) in plasma samples from the 3F7-treated relative to UFH-treated animals, suggesting that there may be more thrombin activation and fibrinolysis in the treated animals. Also at five hours into circuit use, there was a significant decrease in fibrinogen (P = 0.014) in the plasma of 3F7-treated animals relative to UFH-treated ones. The elevated d-dimer and lowered fibrinogen might result from secondary fibrinolysis and fibrinogenolysis not inhibited by 3F7. No studies on fibrinolytic parameters were presented in the investigation. At present, the numbers are too small to be certain if the differences are significant.

We do know that 3F7 has the potential to decrease bleeding in CPB and ECMO to the degree observed with UFH treatment. It is not yet clear whether alterations in platelet function are also prevented, but the changes in platelet count with 3F7 parallel those seen with UFH. Most importantly, 3F7 prevents thrombosis, although the data are not yet crystal clear as to whether it is as good as UFH. Still, as a potential agent for use in CPB and ECMO, 3F7 appears to be quite attractive. First, it can provide instant anticoagulation. Second, it can be adequately monitored in the OR with the ACT assay. Last but not least, 3F7 does not need to be neutralized after surgery because it does not induce bleeding, regardless of its effects on the ACT and aPTT; thus, it will not matter how long it takes to eliminate 3F7 from a patient’s circulation.

3F7 also may be effective in management of other types of contact activation–induced disease, such as catheter and vessel thrombosis, sepsis, adult respiratory distress syndrome, and hereditary angioedema (especially type III, which results from activation of XII). The current work represents a new concept for physicians to appreciate—the prevention of thrombosis without anticoagulation and increased bleeding risk. However, once understood, it should be readily adopted.


View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article