Characterization of an autosomal dominant bleeding disorder caused by a thrombomodulin mutation.

We describe a family with an autosomal dominant disorder characterized by severe trauma- and surgery-related bleeding. The proband, who experienced life-threatening bleeding during a routine operation, had normal clotting times, but markedly reduced prothrombin consumption. Plasma levels of all coagulation factors and of the main coagulation inhibitors were normal. Thrombin generation at low triggers was severely impaired and mixing experiments suggested the presence of a coagulation inhibitor. Using whole exome sequencing, the underlying genetic defect was identified as the THBD c.1611C>A mutation (p.Cys537Stop), which predicts a truncated form of thrombomodulin that is shed from the vascular endothelium. The patient had decreased expression of endothelium-bound thrombomodulin, but extremely elevated levels of soluble thrombomodulin in plasma, impairing the propagation phase of coagulation via rapid activation of protein C and consequent inactivation of factors Va and VIIIa. The same thrombomodulin mutation has been recently described in an unrelated British family with strikingly similar features.

Thrombin-dependent Incorporation of von Willebrand Factor into a Fibrin Network.

Attachment of platelets from the circulation onto a growing thrombus is a process involving multiple platelet receptors, endothelial matrix components, and coagulation factors. It has been indicated previously that during a transglutaminase reaction activated factor XIII (FXIIIa) covalently cross-links von Willebrand factor (VWF) to polymerizing fibrin. Bound VWF further recruits and activates platelets via interactions with the platelet receptor complex glycoprotein Ib (GPIb). In the present study we found proof for binding of VWF to a fibrin monomer layer during the process of fibrinogen-to-fibrin conversion in the presence of thrombin, arvin, or a snake venom from Crotalus atrox. Using a domain deletion mutant we demonstrated the involvement of the C domains of VWF in this binding. Substantial binding of VWF to fibrin monomers persisted in the presence of the FXIIIa inhibitor K9-DON, illustrating that cross-linking via factor XIII is not essential for this phenomenon and suggesting the identification of a second mechanism through which VWF multimers incorporate into a fibrin network. Under high shear conditions, platelets were shown to adhere to fibrin only if VWF had been incorporated. In conclusion, our experiments show that the C domains of VWF and the E domain of fibrin monomers are involved in the incorporation of VWF during the polymerization of fibrin and that this incorporation fosters binding and activation of platelets. Fibrin thus is not an inert end product but partakes in further thrombus growth. Our findings help to elucidate the mechanism of thrombus growth and platelet adhesion under conditions of arterial shear rate.

Whole-blood thrombin generation monitored with a calibrated automated thrombogram-based assay.

BACKGROUND: The calibrated automated thrombogram (CAT) assay in plasma is a versatile tool to investigate patients with hypo- or hypercoagulable phenotypes. The objective was to make this method applicable for whole blood measurements. METHODS: Thin-layer technology and the use of a rhodamine 110-based thrombin substrate appear to be essential for a reliable thrombin generation (TG) assay in whole blood. Using this knowledge we developed a whole blood CAT-based assay. RESULTS: We demonstrated that the whole blood CAT-based assay is a sensitive and rapid screening test to assess function of the hemostatic system under more nearly physiological conditions than the TG assay in plasma. Under conditions of low tissue factor concentration (0.5 pmol/L) and 50% diluted blood, the intraassay CV of the thrombogram parameters, endogenous thrombin potential and thrombin peak height, were 6.7% and 6.5%, respectively. The respective interassay CVs were 12% and 11%. The mean interindividual variation (SD) of 40 healthy volunteers was 633 (146) nmol · min/L for the endogenous thrombin potential and 128 (23) nmol/L for the thrombin peak. Surprisingly, erythrocytes contributed more than platelets to the procoagulant blood cell membranes necessary for optimal TG. Statistically significant (P < 0.001) and potentially clinically significant correlations were observed between circulating factor-VIII concentrations in blood of hemophilia A patients and endogenous thrombin potential (r = 0.62) and thrombin peak height (r = 0.58). CONCLUSIONS: We have developed a reliable method to measure TG in whole blood. The assay can be performed with a drop of blood and may provide a useful measurement of TG under more physiological conditions than plasma.

Absence of platelet-dependent fibrin formation in a patient with Scott syndrome.

To gain insight into the contribution of platelet-dependent thrombin formation in haemostasis and thrombosis, we investigated under flow conditions the haemostatic functions of platelets from a patient with Scott syndrome. Scott platelets are characterised by a diminished platelet-dependent thrombin generation. Thrombin generation was determined by calibrated automated thrombography and flow-based experiments were performed to reveal collagen-mediated platelet activation and fibrin deposition. Our studies indicate that adherent Scott platelets do not differ from control platelets in the formation of stable platelet aggregates under static and flow conditions. While for adherent control platelets a shape change, e.g. balloon formation, and externalisation of phosphatidylserine (PS) is associated with an increase in intracellular calcium concentration, this is not the case for Scott platelets. The calcium-induced morphological changes in control platelets are accompanied with a diminished recruitment of free flowing platelets. Scott platelets, not showing a calcium-induced shape change, also lost the ability to recruit free flowing platelets. These findings rebut the hypothesis that the mild bleeding tendency of Scott syndrome patients is due to a preserved adhesive activity of patient's platelets. Perfusion of tissue factor (TF)-activated control blood over immobilised collagen results in the formation of fibrin fibers that radiate from platelet aggregates. Although platelet aggregates were also observed after perfusion with TF-activated Scott blood, fibrin deposition was not observed. In conclusion, our findings indicate that platelet adhesion and spreading on a collagen matrix in the absence of fibrin formation is sufficient to sustain haemostasis under non-traumatic conditions.

Factor Xa and thrombin evoke additive calcium and proinflammatory responses in endothelial cells subjected to coagulation.

Endothelial cells react to factor Xa and thrombin by proinflammatory responses. It is unclear how these cells respond under physiological conditions, where the serine proteases factor VIIa, factor Xa and thrombin are all simultaneously generated, as in tissue factor-driven blood coagulation. We studied the Ca(2+) signaling and downstream release of interleukins (ILs), induced by these proteases in monolayers of human umbilical vein endothelial cells. In single cells, factor Xa, but not factor VIIa, complexed with tissue factor, evoked a greatly delayed, oscillatory Ca(2+) response, which relied on its catalytic activity and resembled that of SLIGRL, a peptide specifically activating the protease-activated receptor 2 (PAR2). Thrombin even at low concentrations evoked a rapid, mostly non-oscillating Ca(2+) response through activation of PAR1, which reinforced the factor Xa response. The additive Ca(2+) signals persisted, when factor X and prothrombin were activated in situ, or in the presence of plasma that was triggered to coagulate with tissue factor. Further, thrombin reinforced the factor Xa-induced production of IL-8, but not of IL-6. Both interleukins were produced in the presence of coagulating plasma. In conclusion, under coagulant conditions, factor Xa and thrombin appear to contribute in different and additive ways to the Ca(2+)-mobilizing and proinflammatory reactions of endothelial cells. These data provide first evidence that these serine proteases trigger distinct signaling modules in endothelium that is activated by plasma coagulation.

Factor Xa-driven thrombin generation in plasma: dependency on the aminophospholipid density of membranes and inhibition by phospholipid-binding proteins.

Phosphatidylserine (PS) externalization of activated platelets plays a pivotal role in haemostasis and thrombosis. In the present study we have explored the relationship between the PS density of membranes and the rate of thrombin generation in plasma. Factor (F)Xa-initiated thrombin generation was measured in platelet-free plasma (PFP) containing either phospholipid vesicles of varying PS-content or non-stimulated platelets (reconstituted PRP). The duration of the initiation phase of FXa-driven thrombin generation decreased dramatically with increasing PS density. Concomitantly, the maximal rate of thrombin generation during the propagation phase (maxR) increased non-linearly, with the steepest incline between 5 and 10 mol% PS. Titration of FVa into plasma containing 2 mol% PS increased maxR proportionally and diminished the lag phase. In contrast, platelet-dependent thrombin generation was not influenced by addition of FVa. With increasing platelet concentration, the duration of the initiation phase drastically decreased, and maxR increased proportionally. At a physiologically relevant platelet concentration, maxR corresponded with the maxR found with 2 microM of 10 mol% PS. Annexin A5 (AnxA5) and lactadherin appeared to be powerful inhibitors of in-situ thrombin generation under all conditions examined, with AnxA5 being three- to four-fold more potent than lactadherin. In conclusion, maximal thrombin generation in plasma requires membranes with a density of 10-20 mol% PS. Our data further indicate that thrombin formed in situ induces externalization of PS to approx 10 mol% in a substantial platelet subpopulation.

Near-Patient Thrombin Generation in Patients Undergoing Elective Cardiac Surgery.

BACKGROUND: Measuring thrombin generation (TG) in plasma increasingly gained attention as a diagnostic tool in the field of thrombosis and hemostasis. To include the contribution of all blood cells, recently, the whole blood TG method was developed. METHODS: We changed the calculation method of the standard calibrated automated thrombography (CAT) to a method only taking into account the data until the peak of TG, thereby considerably reducing the time from blood draw to result. By redesigning the method, the blood volume per test was reduced to 15 µL. RESULTS: For all TG parameters, the interassay variation proved to be below 15%. The interindividual variation of all parameters was comparable to the CAT method. Thirty-three patients undergoing cardiothoracic surgery were included to investigate whether our assay correlates with postoperative blood loss. On dividing patients into severe and mild bleeders, significant differences between both groups were found for the peak endogenous thrombin potential (peakETP) and peak values determined by our near-patient device. Importantly, patients with a peakETP below the median experienced significantly more blood loss compared to those with a peakETP above the median. A similar division based on the peak as well as the body mass index of the patient yielded similar significant differences. A combination of the peakETP, the body mass index, and the lag time even resulted in a better predictor of blood loss compared to each parameter separately. CONCLUSIONS: Our adapted whole blood TG assay can be used near patients and is indicative for the amount of blood loss post cardiothoracic surgery.

Synergistic effect of thrombin on collagen-induced platelet procoagulant activity is mediated through protease-activated receptor-1.

OBJECTIVE: In the blood coagulation process, the rate of thrombin formation is critically dependent on phosphatidylserine (PtdSer) at the surface of activated platelets. Thrombin synergistically enhances the collagen-induced platelet procoagulant response. The objective of this study is to elucidate the mechanism of this synergistic action with a focus on the intracellular Ca2+ concentration ([Ca2+]i) and the various platelet receptors for thrombin. METHODS AND RESULTS: We demonstrate that procoagulant activity is related to a sustained increased [Ca2+]i, which in turn depends on extracellular Ca2+ influx. Increased PtdSer exposure coincides with increased [Ca2+]i and was observed in a subpopulation (approximately 14%) of the platelets after stimulation with thrombin plus collagen. 2D2-Fab fragments against the thrombin binding site on GPIbalpha made clear that this receptor did not signal for platelet procoagulant activity. Inhibition of protease-activated receptor 1 (PAR-1) and PAR-4 by selective intracellular inhibitors and selective desensitization of these receptors revealed that PAR-1, but not PAR-4, activation is a prerequisite for both sustained elevations in [Ca2+]i and procoagulant activity induced by collagen plus thrombin. CONCLUSIONS: The interaction of thrombin with PAR-1 mediates a synergistic effect on collagen-induced procoagulant activity by inducing a sustained elevation in [Ca2+]i in a subpopulation of platelets.

Thrombin Generation in Zebrafish Blood.

To better understand hypercoagulability as an underlying cause for thrombosis, the leading cause of death in the Western world, new assays to study ex vivo coagulation are essential. The zebrafish is generally accepted as a good model for human hemostasis and thrombosis, as the hemostatic system proved to be similar to that in man. Their small size however, has been a hurdle for more widespread use in hemostasis related research. In this study we developed a method that enables the measurement of thrombin generation in a single drop of non-anticoagulated zebrafish blood. Pre-treatment of the fish with inhibitors of FXa and thrombin, resulted in a dose dependent diminishing of thrombin generation, demonstrating the validity of the assay. In order to establish the relationship between whole blood thrombin generation and fibrin formation, we visualized the resulting fibrin network by scanning electron microscopy. Taken together, in this study we developed a fast and reliable method to measure thrombin generation in whole blood collected from a single zebrafish. Given the similarities between coagulation pathways of zebrafish and mammals, zebrafish may be an ideal animal model to determine the effect of novel therapeutics on thrombin generation. Additionally, because of the ease with which gene functions can be silenced, zebrafish may serve as a model organism for mechanistical research in thrombosis and hemostasis.

Simultaneous measurement of thrombin generation and fibrin formation in whole blood under flow conditions.

Assays based on the formation of thrombin and fibrin are frequently used, and results are considered exchangeable in research/clinical settings. However, thrombin generation and fibrin formation do not always go hand in hand and flow profoundly influences thrombus formation. We describe the technical/clinical evaluation of an assay to simultaneously measure thrombin generation and fibrin formation under conditions of flow. Introduction of a fluorometer into a 'cone and base principle'-based rheometer allowed the measurement of thrombin generation (using a thrombin-sensitive substrate) and fibrin formation (changes in viscosity), while applying a linear shear flow. Increasing shear rates inversely related with thrombin generation and fibrin formation. Increasing fibrinogen concentrations in defibrinated plasma resulted in increased thrombin generation and fibrin formation. In pre-operative samples of 70 patients undergoing cardiothoracic surgery, fibrin formation and thrombin generation parameters correlated with fibrinogen content, rotational thromboelastometry (ROTEM) and whole blood Calibrated Automated Thrombinography (CAT) parameters, respectively. Upon dividing patients into two groups based on the median clot strength, a significant difference in perioperative/total blood loss was established. In conclusion, we clinically evaluated a method capable of simultaneously measuring thrombin generation and fibrin formation in plasma/whole blood under continuous flow, rendering our method one step closer to physiology. Importantly, our test proved to be indicative for the amount of blood loss during/after cardiothoracic surgery.

Complementary roles of glycoprotein VI and alpha2beta1 integrin in collagen-induced thrombus formation in flowing whole blood ex vivo.

Platelets interact vigorously with subendothelial collagens that are exposed by injury or pathological damage of a vessel wall. The collagen-bound platelets trap other platelets to form aggregates, and they expose phosphatidylserine (PS) required for coagulation. Both processes are implicated in the formation of vaso-occlusive thrombi. We previously demonstrated that the immunoglobulin receptor glycoprotein VI (GPVI), but not integrin alpha2beta1, is essential in priming platelet-collagen interaction and subsequent aggregation. Here, we report that these receptors have yet a complementary function in ex vivo thrombus formation during perfusion of whole blood over collagen. With mice deficient in GPVI or blocking antibodies, we found that GPVI was indispensable for collagen-dependent Ca2+ mobilization, exposure of PS, and aggregation of platelets. Deficiency of integrin beta1 reduces the GPVI-evoked responses but still allows the formation of loose platelet aggregates. By using mice deficient in G(alpha)q or specific thromboxane A2 and ADP antagonists, we show that these autocrine agents mediated aggregation but not collagen-induced Ca2+ mobilization or PS exposure. Collectively, these data indicate that integrin alpha2beta1 facilitates the central function of GPVI in the platelet activation processes that lead to thrombus formation, whereas the autocrine thromboxane A2 and ADP serve mainly to trigger aggregate formation.

Factor XI-dependent reciprocal thrombin generation consolidates blood coagulation when tissue factor is not available.

OBJECTIVE: Feedback activation of factor XI by thrombin is a likely alternative for tissue factor-dependent propagation of thrombus formation. However, the hypothesis that thrombin can initiate and propagate its formation in a factor XI-dependent and platelet-dependent manner has not been tested in a plasma milieu. METHODS AND RESULTS: We investigated thrombin generation in recalcified platelet-rich plasma activated with varying amounts of thrombin or factor VIIa. Thrombin initiates and propagates dose-dependently thrombin generation only when platelets and plasma factor XI are present. Incubation of thrombin-activated platelets with a tissue factor neutralizing antibody had no effect on thrombin formation, indicating that platelet-associated tissue factor, if present at all, is not involved. In the absence of factor VIII, thrombin could not initiate its own formation, whereas factor VIIa-induced thrombin generation was reduced. Collagen strongly stimulated both thrombin-initiated and factor VIIa-initiated thrombin generation. CONCLUSIONS: These findings support the notion that platelet-localized feedback activation of factor XI by thrombin plays an important role in maintaining normal hemostasis as well as in sustaining thrombus formation when the TF pathway is inhibited by tissue factor pathway inhibitor.

Covalently-bound heparin makes collagen thromboresistant.

OBJECTIVE: Blood compatibility of artificial surfaces depends on their immunogenic and thrombogenic properties. Collagen's weak antigenicity makes it an attractive candidate for stent coatings or fabrication of vascular grafts. However, the thrombogenic nature of collagen limits its application. We examined whether heparinization can make collagen more thromboresistant. METHODS AND RESULTS: Collagen was heparinized by crosslinking collagen with extensively periodate oxidized heparin and/or by covalently bonding of mildly periodate oxidized heparin. Both ways of heparinization have no effect on platelet adhesion and could not abolish induction of platelet procoagulant activity. However, thrombin generation was completely prevented under static and flow conditions. The functionality of immobilized heparin was confirmed by specific uptake of antithrombin, 13.5+/-4.7 pmol/cm2 and 1.95+/-0.21 pmol/cm2 for mildly and heavily periodated heparin, respectively. CONCLUSIONS: These results indicate that immobilization of heparin on collagen, even as a crosslinker, is a very effective way to prevent surface thrombus formation. These data encourage the application of heparinized collagen as stent-graft material in animal and eventually human studies.

Hypoxia Induces a Prothrombotic State Independently of the Physical Activity.

Hypoxia (oxygen deprivation) is known to be associated with deep vein thrombosis and venous thromboembolism. We attempted to get a better comprehension of its mechanism by going to high altitude, thereby including the potential contributing role of physical activity. Two groups of 15 healthy individuals were exposed to hypoxia by going to an altitude of 3900 meters, either by climbing actively (active group) or transported passively by cable car (passive group). Both groups were tested for plasma fibrinogen, von Willebrand factor and factor VIII levels, fibrinolysis, thrombin generating capacity, heart rate, oxygen saturation levels and blood pressure. As a control for the passive group, 7 healthy volunteers stayed immobile in bed for 7 days at normoxic conditions. The heart rate increased and oxygen saturation levels decreased with increasing altitude. Fibrinolysis and fibrinogen levels were not affected. Factor VIII and von Willebrand factor levels levels increased significantly in the active group, but not in the passive group. Plasma thrombin generation remained unchanged in both the active and passive group with increasing altitude and during 7 days of immobility in healthy subjects. However, by applying whole blood thrombin generation, we found an increased peak height and endogenous thrombin potential, and a decreased lagtime and time-to-peak with increasing levels of hypoxia in both groups. In conclusion, by applying whole blood thrombin generation we demonstrated that hypoxia causes a prothrombotic state. As thrombin generation in plasma did not increase, our results suggest that the cellular part of the blood is involved in the prothrombotic phenotype induced by hypoxia.

Platelet activation and blood coagulation.

Platelet activation and blood coagulation are complementary, mutually dependent processes in haemostasis and thrombosis. Platelets interact with several coagulation factors, while the coagulation product thrombin is a potent platelet-activating agonist. Activated platelets come in a procoagulant state after a prolonged elevation in cytosolic [Ca2+]i. Such platelets, e.g. when adhering to collagen via glycoprotein VI, expose phosphatidylserine (PS) at their outer surface and produce (PS-exposing) membrane blebs and microvesicles. Inhibition of aminophospholipid translocase and activation of phospholipid scramblase mediate the exposure of PS, whereas calpain-mediated protein cleavage leads to membrane blebbing and vesiculation. Surface-exposed PS strongly propagates the coagulation process by facilitating the assembly and activation of tenase and prothrombinase complexes. Factor IXa and platelet-bound factor Va support these activities. In addition, platelets can support the initiation phase of coagulation by providing binding sites for prothrombin and factor XI. They thereby take over the initiating role of tissue factor and factor VIIa in coagulation activation.

Fibrinogen adsorption, platelet adhesion and thrombin generation at heparinized surfaces exposed to flowing blood.

Thrombus formation at an artificial surface in contact with blood is a complex process that encompasses accretion of platelets from flowing blood and fibrin deposition. Platelet adhesion and fibrin formation are intimately intertwined reactions that are triggered by different sets of surface adsorbed plasma proteins. To dissect the contribution of protein adsorption and platelet adhesion to thrombin formation, a coherent study was performed with non-coated (NC) and heparin-coated (HC) surfaces. Thrombin production in whole blood, platelet adhesion and protein adsorption were studied using an amidolytic thrombin assay, a dynamic platelet adhesion assay and ellipsometry, respectively. Thrombin generation in flowing whole blood exposed to HC surfaces was greatly diminished when compared with NC surfaces. However, separate platelet adhesion and protein adsorption studies with anticoagulated whole blood revealed that platelets do not adhere because fibrinogen is not available in the protein layer that was deposited during the perfusion. These findings indicate that the in vitro thrombogenicity of a material cannot be predicted from platelet adhesion and protein adsorption data when these measurements are performed with anti-coagulated blood or platelet rich plasma. Preincubation of NC and HC surfaces with fibrinogen or 2000-fold diluted plasma resulted in similar amounts of surface-bound fibrinogen and mediated massive platelet adhesion from flowing whole blood. These results indicate that a) platelet adhesion correlates with the availability of surface-bound fibrinogen and b) NC and HC surfaces are indistinguishable with respect to protein (fibrinogen) adsorption and platelet adhesion. It is apparent that the heparinized surface used in our studies exerts its anti-thrombogenic properties by neutralizing locally formed thrombin and not by reducing fibrinogen-dependent platelet adhesion.

Thrombogenicity of polysaccharide-coated surfaces.

Heparinization of artificial surfaces has been proven to reduce the intrinsic thrombogenicity of such surfaces. The mechanism by which immobilized heparin reduces thrombogenicity is not completely understood. In the present study heparin-, alginic acid- and chondroitin-6-sulphate-coated surfaces were examined for protein adsorption, platelet adhesion and thrombin generation. The protein-binding capacity from solutions of purified proteins was significantly higher for heparin-coated surfaces when compared with alginic acid- and chondroitin sulphate-coated surfaces. Yet, when the surfaces were exposed to flowing plasma, only the heparinized surface adsorbed significant amounts of antithrombin. None of the surfaces adsorbed fibrinogen under these conditions, and as a result no platelets adhered from flowing whole blood. Our results indicate that protein adsorption and platelet adhesion from anticoagulated blood cannot be used to assess the thrombogenicity of (coated) artificial surfaces. Indeed, the thrombin generation potentials of the different surfaces varied remarkable: while non-coated surface readily produced thrombin, alginic acid- and chondroitin sulphate-coated surfaces showed a marked reduction and virtually no thrombin was generated in flowing whole blood passing by heparinized surfaces.

Coils and tubes releasing heparin. Studies on a new vascular graft prototype.

Coiled metallic guidewires find widespread use, for instance in interventional cardiology. It is known that release of heparin from the surface of guidewires is advantageous to prevent formation of thrombotic emboli. New coiled tubular structures, having larger inner and outer diameter as compared to guidewires, are presented. In theory these tubes can be used as interposition vascular grafts. Ten coiled tubes with an internal diameter of 690 microm were made. Five different adherent polymeric coatings with increasing hydrophilicity were used. Five tubes contained heparin in the coating and the other five were unheparinised controls. The five tubes containing heparin were studied with respect to heparin release in vitro (amount released, kinetics), and immobilised heparin that is exposed at the surface. All tubes were studied with a direct cell contact assay using 3T3 mouse fibroblast cells, a dynamic thrombin generation test, and endothelial cell growth onto the coils. It was found that the heparinised tubes lead to very little thrombin formation. It is argued that this is due to heparin that is immobilised and exposed at the inner surface of such tubes. Furthermore the coils showed to be cytocompatible and endothelial cells adhere and proliferate well onto the coils. This concept is believed to hold promise for further development of small vascular grafts.

Thrombin generation during cardiopulmonary bypass: the possible role of retransfusion of blood aspirated from the surgical field.

BACKGROUND: In spite of using heparin-coated extracorporeal circuits, cardiopulmonary bypass (CPB) is still associated with an extensive thrombin generation, which is only partially suppressed by the use of high dosages of heparin. Recent studies have focused on the origins of this thrombotic stimulus and the possible role of retransfused suctioned blood from the thoracic cavities on the activation of the extrinsic coagulation pathway. The present study was designed to find during CPB an association between retransfusion of suctioned blood from the pericardium and pleural space, containing activated factor VIIa and systemic thrombin generation. METHODS: Blood samples taken from 12 consenting patients who had elective cardiac surgery were assayed for plasma factor VIIa, prothrombin fragment 1+2 (F1+2), and thrombin-antithrombin (TAT) concentrations. Blood aspirated from the pericardium and pleural space was collected separately, assayed for F1+2, TAT, and factor VIIa and retransfused to the patient after the aorta occlusion. RESULTS: After systemic heparinization and during CPB thrombin generation was minimal, as indicated by the lower than base line plasma levels of F1+2, and TAT after correction for hemodilution. In contrast, blood aspirated from the thoracic cavities had significantly higher levels of factor VIIa, F1+2, and TAT compared to the simultaneous samples from the blood circulation (P < 0.05). Furthermore, after retransfusion of the suctioned blood (range, 200-1600 mL) circulating levels of F1+2, and TAT rose significantly from 1.6 to 2.9 nmol/L (P = 0.002) and from 5.1 to 37.5 µg/L (P = 0.01), respectively. The increase in both F1+2, and TAT levels correlated significantly with the amount of retransfused suctioned blood (r = 0.68, P = 0.021 and r = 0.90, P = 0.001, respectively). However, the circulating factor VIIa levels did not correlate with TAT and F1+2 levels. CONCLUSIONS: These data suggest that blood aspirated from the thoracic cavities during CPB is highly thrombogenic. Retransfusion of this blood may, therefore, promote further systemic thrombin generation during CPB.