Are P2Y12 reaction unit (PRU) and % inhibition index equivalent for the expression of P2Y12 inhibition by the VerifyNow assay? Role of haematocrit and haemoglobin levels

The results of the whole blood VerifyNow P2Y12 assay can be expressed as platelet reaction units (PRU) or % inhibition index (%inh), but an optimal cut-off for the assessment of high on-treatment platelet reactivity (HPR) predictive of clinical events has been validated only for PRU. The aim of the study was to study the influence of haematological variables, such as platelet and leukocyte counts or haematocrit / haemoglobin, within the limits indicated by the manufacturer for assay validity, on the results of the test. We performed a comparison of PRU and %inh in a series 186 samples obtained from a clinical trial on patients under dual antiplatelet therapy. The results show that PRU significantly decreases with increasing haematocrit / haemoglobin, whereas %inh does not, due to a parallel change in PRU and iso-TRAP baseline value. PRU and % inhibition index are not equivalent for the definition of HPR, because of their different sensitivities to haematocrit / haemoglobin.

Internalization of microparticles by endothelial cells promotes platelet/endothelial cell interaction under flow

BACKGROUND

Microparticles (MPs) released by activated or apoptotic cells increase in number in the blood of subjects with vascular or metabolic diseases and may contribute to thrombotic complications.

OBJECTIVES

In this study, we investigated whether MPs promoted platelet recruitment to endothelial cells in flow conditions, and by which mechanism.

METHODS

Human umbilical vein endothelial cells (HUVECs) grown in microslide perfusion chambers were exposed to MPs prepared in vitro from HUVECs, monocytes or platelets.

RESULTS

Videomicroscopy of DIOC-labelled blood perfused at arterial rate on human umbilical vein ECs demonstrated that, irrespective of their cell origin, MPs promoted the formation of platelet strings at the surface of HUVECs. This platelet/endothelial cell interaction was dependent on von Willebrand factor (VWF) expression at the HUVEC surface and involved Glycoprotein Ib and P-selectin. Interestingly, HUVECs internalized MPs within a few hours through a process involving anionic phospholipids, lactadherin and αvβ3 integrin. This uptake generated the production of reactive oxygen species via the xanthine/xanthine oxidase system (inhibited by allopurinol and the ROCK inhibitor Y-27632) and the NADPH oxidase (inhibited by SOD). Reactive oxygen species appeared essential for VWF expression at the endothelial cell surface and subsequent platelet/endothelial cell interaction under flow. The pathophysiological relevance of this process is underlined by the fact that circulating MPs from Type I diabetic patients induced platelet/endothelial cell interaction under flow, with an intensity correlated with the severity of the vasculopathy.

French clinical practice guidelines on the management of patients on vitamin K antagonists in at-risk situations (overdose, risk of bleeding, and active bleeding)

The present report from several French medical societies in the field and the French National Authority for Health provides an expert consensus for the management of patients on vitamin K antagonists in at-risk situations (overdose, risk of bleeding, and active bleeding). Asymptomatic VKA overdose is defined as an International Normalized Ratio (INR) value above the upper limit of the therapeutic target. In this case, the guidelines describe the rapid reduction of the INR down to the therapeutic range, either by omitting a dose or using vitamin K. Regarding the haemorrhagic complications, the guidelines address the management of these patients according to the severity of bleeding, and especially focus on the use of prothrombin complex concentrate. Finally, the consensus addresses the management of patients in cases of elective or emergency surgery or other invasive procedures, and discusses whether treatment should be continued or not, and whether VKA substitution by heparin--"bridging anticoagulation"--is needed.

Microparticles from apoptotic monocytes induce transient platelet recruitment and tissue factor expression by cultured human vascular endothelial cells via a redox-sensitive mechanism

Circulating microparticles derived from different types of blood cells have been reported to impair endothelial function and to induce pro-inflammatory and prothrombotic endothelial phenotypes. Although the number of monocyte-derived microparticles (M-MPs) is elevated in the blood of patients with various inflammatory conditions, their interaction with endothelial cells has been poorly investigated so far. In this study, we produced microparticles in vitro from apoptotic human monocytes and examined the effects of their interaction with cultured human umbilical vascular endothelial cells (HUVECs). We found that low concentrations of M-MPs induced the production of reactive oxygen species (ROS), mainly anion superoxide, by the endothelial cells. At sub-toxic concentrations, M-MPs induced a rapid expression of von Willebrand factor at the cell surface, which mediated the transient attachment of non-activated platelets to the endothelium in flow conditions. In parallel, M-MPs up-regulated the expression of functional tissue factor by the endothelial cells. ROS controlled these two major changes and the process involved the phosphorylation of p38 mitogen activated protein kinase. We conclude that M-MPs may contribute to thrombotic events by producing redox signalling in endothelial cells.

Understanding the controversy about hormonal replacement therapy: insights from estrogen effects on experimental and clinical atherosclerosis

Whereas hormonal replacement/menopause therapy (HRT) in post-menopausal women increases the coronary artery risk, epidemiological studies (protection in pre-menopaused women) suggest and experimental studies (prevention of the development of fatty streaks in animals) demonstrate a major atheroprotective action of estradiol (E2). The understanding of the deleterious and beneficial effects of estrogens is thus required. The immuno-inflammatory system plays a key role in the development of fatty streak deposit as well as in the atherosclerotic plaque rupture. Whereas E2 favors an anti-inflammatory effect in vitro (cultured cells), it rather elicits pro-inflammation in vivo, at the level of several subpopulations of the immuno-inflammatory system, which could contribute to plaque destabilization. Endothelium is another important target for E2, since it stimulates endothelial NO and prostacyclin production, thus promoting beneficial effects of vasorelaxation and platelet aggregation inhibition. Prostacyclin, but not NO, appears to be involved in the atheroprotective effect of E2. Estradiol accelerates also endothelial regrowth, thus favoring vascular healing. Finally, most of these effects of E2 are mediated by estrogen receptor alpha, and are independent of estrogen receptor beta. In summary, a better understanding of the mechanisms of estrogen action is required not only on the normal and atheromatous arteries, but also on innate and adaptive immune responses. This should help cardiovascular disease prevention optimization after menopause. These mouse models should help to screen existing and future Selective Estrogen Receptor Modulators (SERMs).

Antiphosphatidylethanolamine antibodies are associated with an increased odds ratio for thrombosis. A multicenter study with the participation of the European Forum on antiphospholipid antibodies

A multicenter study was set up to evaluate the prevalence, clinical and biological significance of antiphosphatidylethanolamine antibodies (aPE) in thrombotic patients with or without the main known clinical and biological risk factors for thrombosis. APE and antibodies, defined as the laboratory criteria of antiphospholipid syndrome (APS) -lupus anticoagulant, anticardiolipin and anti-beta(2)-GPI antibodies were measured in 270 patients with thrombosis (234 venous and 37 arterial) and 236 matched controls. APE were found in 15% of thrombotic patients compared to 3% of controls (p < 0.001) with no predominant isotype, no association with the main known clinical or biological risk factors for thrombosis neither with a type of thrombosis, arterial or venous. In a multivariate logistic regression analysis of antibodies, aPE showed the highest association with thrombosis (odds ratio [OR]: 4.2, p < 0.001). Moreover, using a multivariate analysis in a case-control subgroup study on 158 patients, IgGaPE were found to be significantly associated with venous thrombosis (OR:6;p = 0.005). Interestingly, 25 of the 40 aPE-positive patients (63%) were negative for the APS laboratory criteria. Most of them (21/25) had venous thrombosis, recurrent in ten of them. Four patients also suffered from early or late miscarriages. Our results underline the strength of the association between the presence of aPE and thrombosis and suggest their measurement in thrombotic patients, especially when lupus anticoagulant, anticardiolipin or anti-beta(2)-GPI antibodies are absent.

Central laboratory and point of care assessment of perioperative hemostasis

PURPOSE

To review laboratory investigations required for the diagnosis of a constitutional or acquired defect of hemostasis before surgery, or during the perioperative period in context of rapid evolution and possible therapeutic adjustment.

METHODS

A review of the literature.

PRINCIPAL FINDINGS

Systematic preoperative screening is poorly efficient. It should be restricted to patients selected on clinical history and physical examination. Intra- and postoperative investigation is oriented by the clinical circumstances. In most instances, laboratory investigation is organized as hierarchical steps, the first one including a platelet count, a prothrombin time and activated partial thromboplastin time. According to the results of these assays and to the clinical circumstances, further tests specific of primary hemostasis, coagulation or fibrinolysis are performed. Point of care (POC) monitoring has been developed more recently for the investigation of a perioperative bleeding. Several tests examine platelet functions, others measure whole blood activated partial thromboplastin time or prothrombin time and some explore global hemostasis, allowing the detection of excessive fibrinolysis. Point of care testing provides a rapid and valuable answer but, if one accepts the monitoring of unfractionated heparin during extracorporeal circulation, most assays have not been properly validated.

CONCLUSION

The investigation of hemostasis at the central laboratory and POC testing have distinct objectives. The utility of the former for the diagnosis and the adjustment of therapeutics have been well demonstrated. In contrast, the experience with POC testing is relatively recent, and its utility for patient management remains to be demonstrated in proper clinical trials.

Integrin-dependent interaction of lipid rafts with the actin cytoskeleton in activated human platelets

Dynamic connections between actin filaments and the plasma membrane are crucial for the regulation of blood platelet functions. Protein complexes associated with αIIbβ3 integrin-based cytoskeleton structures are known to play a role in these processes. However, mechanisms involving lateral organizations of the plasma membrane remain to be investigated. Here, we demonstrate that a large fraction of platelet lipid rafts specifically associates with the actin cytoskeleton upon activation. This association was inhibited by antagonists of fibrinogen-αIIbβ3 binding and did not occur in type I Glanzman's thrombasthenic platelets. The raft-cytoskeleton interaction is a reversible process correlating with the intensity and stability of platelet aggregation. Although only a minor fraction of αIIbβ3 was recovered in rafts upon activation, this integrin specifically upregulated the level of PtdIns(4,5)P2 in membrane microdomains and induced the recruitment of several actin-modulating proteins known to directly or indirectly interact with this lipid. Controlled disruption of rafts did not affect αIIbβ3-mediated platelet aggregation in response to high concentrations of thrombin but significantly inhibited fibrin clot retraction. We propose that rafts participate in the organization of membrane-cytoskeleton interactions where αIIbβ3-mediated tension forces apply during the late phase of platelet activation.

Shedding of active tissue factor by aortic smooth muscle cells (SMCs) undergoing apoptosis

As apoptosis of neo-intimal SMCs is a feature of advanced atherosclerotic plaques, the procoagulant properties of SMCs of synthetic phenotype undergoing apoptosis were investigated. SMCs isolated from rat aorta obtained 10 days after balloon injury, previously found to up-regulate Tissue Factor (TF) and Tissue Factor Inhibitor (TFPI) and to release large amounts of TFPI (Ghrib et al. Thromb Haemost 2002;87:1043-50), were sensitive to the apoptosis induced by Fas-ligand. During this process, surface TF activity rose by a factor 10 over 6 hours, in parallel with a proportional increase in prothrombinase, while TF protein expressed at the membrane significantly decreased. The microparticles (MPs) produced during SMC death bore intact and functional TF, but the release of TFPI did not change, so that the balance shifted to a procoagulant state during apoptosis. Shed MPs enhanced thrombus formation in flowing whole blood over collagen coated-glass slides. Apoptotic SMCs in atherosclerotic plaques represent a reservoir of highly thrombogenic material, released into the blood stream in case of spontaneous or mechanical plaque disruption.

Standardization of factor VII/activated factor VII measurement in plasma of patients treated with recombinant factor activated VII

To improve the standardization of the factor VII clotting activity (FVII:C) assay in patients treated with recombinant activated factor VII (rFVIIa), we conducted a multicentre study on plasma samples from four patients with haemophilia A, before and at various times after injection of a single dose of rFVIIa. FVII:C and prothrombin time were measured with the methods and reagents routinely used in each laboratory. Strong inter-laboratory variability of FVII:C values was found. The main source of variability was the type of thromboplastin. FVII:C values measured using rabbit thromboplastin were very close to activated factor VII clotting activity values (FVIIa:C) measured with a commercial assay (Staclot VIIa-TF). FVII:C values obtained with human placental thromboplastin were about three times lower than those obtained with rabbit and recombinant thromboplastins, and with the FVIIa:C assay. There was a good relationship between FVIIa:C and activated factor VII antigen values measured using a commercial immunoassay (Imubind FVIIa ELISA). In conclusion, rFVIIa at pharmacological concentrations can be easily monitored on the basis of FVII:C, using rabbit and probably also recombinant thromboplastin; equivalent results are obtained with a specific activated factor VII bioassay.

Utilisation and safety of low molecular weight heparins: prospective observational study in medical inpatients

AIMS

Low molecular weight heparins (LMWHs) are widely used as curative or preventive treatments of thromboembolic diseases. The aim of our study was to: investigate the pattern of prescription of LMWHs in different departments of French teaching hospitals; andestimate the incidence of adverse drug reactions (ADRs) induced by LMWHs and associated risk factors for the occurrence of bleeding events.

METHODS

This prospective study was performed in two teaching hospitals in Toulouse (south-western France) in March 1999 in different medical wards. All patients receiving a prescription for a LMWH were included in the survey. All data were prospectively recorded in each ward.

RESULTS

A total of 334 patients were included. Sex ratio (male/female) was 1.25 and mean age was 72.5 +/- 16.3 years (extremes:18-101). 450 prescriptions for LMWHs were collected (1.34 prescription per patient) and involved mainly enoxaparin (61%), which was more frequently used than tinzaparin in patients over 75 years old (71.7 vs 28.3%; p < 0.0001). Ninety-nine patients received a LMWH for curative treatment (corresponding to 127 prescriptions of which 99 were for enoxaparin and 28 were for tinzaparin [p < 0.0001]). Indications included therapy for deep venous thrombosis, pulmonary embolism, acute coronary syndrome, unstable angina pectoris, non-Q-wave myocardial infarction. Serious renal insufficiency was significantly more frequent in patients from the geriatrics department (p < 0.00001). Enoxaparin was prescribed more frequently in patients with serious or moderate renal insufficiency than tinzaparin (72 vs 61%, p < 0.05). The incidence of LMWHs-induced ADRs was 10.5% occurring in 22 cases during preventive treatment of deep venous thrombosis and in 13 cases during curative therapy. ADRs were classified as 'serious' in 11 cases (31.4%). Reported ADRs were bleeding events (n = 15), thrombocytosis (n = 13), thrombopenia (n = 4) and hepatic cytolysis (n = 1). The mean delay for the occurrence of bleeding effects was 8.0 +/- 9.1 days (range 1-40). Multivariate analysis of the influence of several criteria on the occurrence of haemorrhagic effects showed that the decrease of creatinine clearance (10 ml/min) was associated with an increased haemorrhagic risk (relative risk [RR] = 1.34, 95% CI 1.12-1.65; p < 0.05). Moreover, the risk of adverse bleeding effects increased for patients with a creatinine clearance <20 ml/min (RR = 2.8; 95% CI 1.00-7.8).

CONCLUSION

Our data firstly show a different pattern of LMWHs prescription in different clinical wards. Secondly, the risk of bleeding ADRs in patients treated by LMWHs increases significantly with renal function impairment for the two LMWH preparations studied. More pharmacoepidemiological studies are necessary in patients with several risk factors, particularly in elderly people who often have renal impairment, in order to determine the optimal pattern use of each LMWH.

Reassessment of von Willebrand factor (VWF), VWF propeptide, factor VIII:C and plasminogen activator inhibitors 1 and 2 during normal pregnancy

The use of von Willebrand factor [VWF antigen (WF:Ag)] measurement as a marker of endothelial cell activation for monitoring hypertensive pregnancies is limited by the poor definition of reference values. We reassessed these reference values using different assays, and those of the propeptide (VWF:Ag II) and factor VIII coagulant activity (factor VIII:C), in a large population of normal pregnancies, at 3-week intervals of gestational age. Plasminogen activator inhibitors (PAI-1 and PAI-2) were measured in parallel. Blood was collected at single time points between 12 weeks and delivery in 306 women undergoing normal singleton pregnancy. For clinical purposes, the VWF:Ag reference values were assay independent and the influence of ABO blood group on VWF:Ag or factor VIII:c was found to be limited during the third trimester. VWF:Ag II was not influenced by the ABO blood group. The ratio, VWF:Ag/factor VIII:C was close to 1.0 throughout pregnancy. In contrast, VWF:Ag II increased more slowly than VWF:Ag and the ratio of VWF:Ag II to VWF:Ag in plasma decreased from 1.00 to 0.5 at term. PAI-1 and PAI-2 increased with gestational age, but PAI-2 decreased during the last 2 weeks, indicating physiological placental regression at the very end of pregnancy.

The expression of tissue factor and tissue factor pathway inhibitor in aortic smooth muscle cells is up-regulated in synthetic compared to contractile phenotype

Tissue factor (TF) and its specific inhibitor TF pathway inhibitor (TFPI) are produced by vascular smooth muscle cells (SMCs) in vitro and are increased in vivo in atherosclerotic compared to normal vessels. Besides local regulation of the hemostatic balance, this may be related to non-hemostatic TF/protease dependent functions such as SMC proliferation, adhesion and migration. The aim of the study was to compare the expression of both proteins between the contractile (normal adult) and synthetic (neo-intimal) SMC phenotypes. Primary cultures of SMCs isolated from rat thoracic aorta before and 10 days after balloon injury displayed stable characteristics of the contractile and synthetic phenotype, respectively. Synthetic SMCs expressed more TF mRNA than contractile SMCs, but released excess TF in the conditioned medium, so that the cell-associated TF activity measured by a factor Xa generating assay remained similar in the two subtypes. Accordingly, cell surface thrombogenicity measured under blood flow conditions was also similar. The production and release of functional TFPI was enhanced by a factor 3 to 6 (p < 0.01) in synthetic SMCs. A difference in the quantitative expression of TF and TFPI is a new distinctive feature of SMC phenotypes. Matrix-associated TFPI derived from synthetic SMCs may serve as an anchorage for their migration and regulate protease-activated processes during neo-intima formation.

Quantification of lupus anticoagulants in clinical samples using anti-beta2GP1 and anti-prothrombin monoclonal antibodies

Quantification of lupus anticoagulant (LA) in clinical samples is hampered by the lack of a suitable standard of activity. We evaluated the use of mAbs displaying LA activity for this purpose. As most patient samples contain both beta2Glycoprotein I (beta2GP1) and prothrombin dependent LA, a combination of two mAbs, one of each specificity, was added to normal plasma in a concentration from 0 to 60 microg/ml. Eight assay systems using different reagents and instruments were used. The calibration curves were linear for all but one, with marked differences between the responsiveness to each mAb. A panel of plasmas from 69 patients with persistent LA diagnosed using the SSC-ISTH criteria was tested. An antiphospholipid syndrome (APS) was present in 40, whereas 29 were asymptomatic. LA activities of individual plasmas varied between assays (p < 10(-4)), but homogeneous subgroups were identified. In a majority of samples, LA activity displayed a prothrombin-dependent profile, with a variable contribution of beta2GP1-dependent activity. The latter was associated to beta2GP1 antibodies detected by solid-phase immunoassay. By using 3 dilute Russell viper venom time assays, higher LA titers were found in APS, compared to asymptomatic patients (p <0.05).

The effect of platelet PlA polymorphism on experimental thrombus formation in man depends on blood flow and thrombogenic substrate

A number of studies have reported conflicting data on the association of the PlA1/PlA2 polymorphism of the GPIIIa gene and coronary syndromes. We have investigated the effect of this polymorphism on experimental platelet thrombus formation in man. Forty healthy male volunteers were genotyped for the PlA1/PlA2 polymorphism. Thrombus formation was induced ex vivo by exposing a tissue factor (TF) or a collagen-coated coverslip in a parallel plate perfusion chamber to native blood for 2 and 4 min. The shear rates at these surfaces were 650 and 2,600 s(-1). Platelet and fibrin deposition was quantified by immunoenzymatic methods. The frequencies of PlA1/PlA1 and PlA1/PlA2 genotypes were 52.5% and 47.5%, respectively. Ex vivo deposition of fibrin on TF was not affected by the PlA1/PlA2 polymorphism. However, the ex vivo platelet deposition at 650 s(-1) was higher in blood from PlA1/PlA1 individuals than in PlA1/PlA2 individuals (P= 0.008 at 4 min). On collagen, neither fibrin nor platelet deposition was significantly affected by the PlA1/PlA2 polymorphism. Platelet thrombus formation is significantly influenced by genetic variations in the GPIIIa platelet receptor. This effect depends on the blood flow properties and the nature of the thrombogenic stimulus.

The French haemophilia cohort: rationale and organization of a long-term national pharmacosurveillance system

Medicinal products of biological origin still carry a specific iatrogenic risk, mainly because of their starting material, mode of preparation and variability. Careful postmarketing surveillance systems are therefore necessary. To assess the long-term safety of haemophilia treatment with plasma-derived and recombinant clotting factor products, a cohort study was set up in France in 1994. Participants were patients with haemophilia A and B, with or without previous clotting factor therapy. Clinical events, treatments, biological data and adverse events were recorded on standard forms. Blood samples were separated into serum, plasma and peripheral blood mononuclear cells, frozen, and banked in a central laboratory. The same data and samples were collected at yearly follow-up visits. As of December 1999, 1234 haemophiliacs were enrolled in 39 haemophilia centres. At enrollment, 50.2% of patients were under 15 years of age, and the cumulative number of days of exposure to the product was below 50 in 35.1% of cases. The median duration of follow-up was 26.9 months, with a total of 2729 patient-years (135,947 days of exposure and 211 million units of factor VIII or IX). To date, only 17 patients were lost to follow-up. The initial results show good compliance with this health-watch policy among patients and clinicians specializing in haemophilia. The regular follow-up data and centralized sample bank will serve to investigate rapidly any suspected outbreaks as soon as reliable biological tests become available in the future.

Molecular bases of antithrombin deficiency in French families: identification of seven novel mutations in the antithrombin gene

We have investigated the molecular bases of familial antithrombin deficiency in eight French families. Eight mutations in the antithrombin coding exons were identified, seven of which were novel mutations. In all cases, individuals were heterozygous for the mutation. We found two small frameshift deletions in exon 3a, leading to type I deficiency. Five missense mutations in exons 3b or 5 also caused type I deficiency and their potential consequences on the antithrombin three-dimensional structure were analysed. The last mutation in exon 4 was associated with a type II 'reactive site' deficiency: a dysfunctional antithrombin that is affected in its interaction with thrombin was present in circulation.

Inhibition of tissue factor synthesis by disruption of ERK kinases and PKC signaling pathways in human vascular SMCs

Tissue Factor (TF), the receptor for plasma VII/VIIa and the initiator of blood coagulation, is inducible in vascular smooth muscle cells (SMCs) by growth factors and bacterial lysopolysaccharides (LPS) and is expressed in vivo after vascular injury. As TF expression is a determinant of the thrombogenicity of vascular lesions, we investigated the signal pathways involved in this process. Human vascular SMCs were obtained from normal arteries and made quiescent by serum deprivation. Baseline TF antigen and activity were up-regulated by various agonists: fetal calf serum (FCS), LPS, and platelet derived growth factor (PDGF) being the most effective but with different kinetics. TF expression induced by LPS was transient with a maximum 6 h after stimulation and returned to baseline levels after 24 h whereas TF expression induced by serum or PDGF was sustained for at least 24 h. Rapid and transient activation of Extracellular signal-Regulated Kinase (ERK) was observed after stimulation by PDGF and FCS, but not by LPS. The role of ERK, Ras and protein kinase C activities were investigated using specific inhibitors, PD 98059, manumycin A and calphostin C respectively. For TF induction by LPS, PKC activity was required and the ERK/Ras pathway was not involved. In contrast, the effect of PDGF was strictly ERK and Ras dependent, but partially prevented by PKC inhibitors. TF induction by FCS was ERK dependent but partially Ras and PKC dependent. In conclusion, TF expression appears to be a non-specific response of SMCs to numerous stimuli through multiple signal pathways which differ according to the inducing agent.

Further studies on the mechanism for the antithrombotic effects of naroparcil, an orally active thiozyloside compound

The antithrombotic beta-D-xyloside, naroparcil, has previously been shown to induce a dose-related increase of circulating glycosaminoglycans (GAGs) together with an antithrombin activity (anti-IIa) via heparin cofactor II (HCII) in the rabbit. In order to go further in the mechanisms, the relationship between the antithrombotic activity, the HCII-mediated anti-IIa activity and the plasma GAG content was investigated. We showed that the in vitro specific activity on the inhibition of thrombin by HCII of the plasma GAG extract from naroparcil-treated rabbits was increased by a factor of 60 when compared to controls. In addition, the fractionation of the plasma GAG extract by affinity chromatography on immobilized HCII led to a more potent material whereas the low-affinity fraction was shown to be inactive in thrombin inhibition by HCII. The qualitative analysis of GAGs showed the presence of the deltaDi-4S DS disaccharide, undetectable in control, which accounted for 22% in the unfractionated GAG extract and for 60% in the high affinity fraction. In vitro experiments using immuno-depleted plasma in antithrombin III (ATIII), HCII or both, indicated that the anti-IIa activity of the plasma GAG extract from naroparcil-treated rabbits was mainly due to HCII potentialisation. The unfractionated GAG extract and the high affinity fraction were shown to be antithrombotic in a Wessler-based model in the rat, giving ED80 values of 610 UA/kg and 56 UA/kg respectively whereas the low-affinity fraction was devoid of any antithrombotic activity. These results show that the antithrombotic activity of naroparcil is dependent on modification in the plasma GAG profile which inactivates thrombin via the HCII.