Effect on perfusion chamber thrombus size in patients with atrial fibrillation during anticoagulant treatment with oral direct thrombin inhibitors, AZD0837 or ximelagatran, or with vitamin K antagonists

Thrombosis Research 129 (2012) e83–e91

Contents lists available at SciVerse ScienceDirect

Thrombosis Research journal homepage: www.elsevier.com/locate/thromres

Regular Article

Effect on perfusion chamber thrombus size in patients with atrial fibrillation during anticoagulant treatment with oral direct thrombin inhibitors, AZD0837 or ximelagatran, or with vitamin K antagonists Michael Wolzt a,⁎, Ulf G. Eriksson c, Ghazaleh Gouya a, Nicolai Leuchten a, Stylianos Kapiotis b, Margareta Elg c, Kajs-Marie Schützer c, Sofia Zetterstrand c, Malin Holmberg c, Karin Wåhlander c a b c

Department of Clinical Pharmacology, Medical University of Vienna, Austria Clinical Institute for Laboratory Medicine, Medical University of Vienna, Austria AstraZeneca R&D Mölndal, Sweden

a r t i c l e

i n f o

Article history: Received 2 June 2011 Received in revised form 12 August 2011 Accepted 17 August 2011 Available online 16 September 2011 Keywords: AZD0837 ximelagatran direct thrombin inhibitor perfusion chamber thrombus size atrial fibrillation

a b s t r a c t Introduction: AZD0837 and ximelagatran are oral direct thrombin inhibitors that are rapidly absorbed and bioconverted to their active forms, AR-H067637 and melagatran, respectively. This study investigated the antithrombotic effect of AZD0837, compared to ximelagatran and the vitamin K antagonist (VKA) phenprocoumon (Marcoumar ®), in a disease model of thrombosis in patients with non-valvular atrial fibrillation (NVAF). Methods: Open, parallel-group studies were performed in NVAF patients treated with VKA, which was stopped aiming for an international normalized ratio (INR) of ≤ 2 before randomization. Study I: 38 patients randomized to AZD0837 (150, 250 or 350 mg) or ximelagatran 36 mg twice daily for 10–14 days. Study II: 27 patients randomized to AZD0837 250 mg twice daily or VKA titrated to an INR of 2–3 for 10–14 days. A control group of 20 healthy elderly subjects without NVAF or anticoagulant treatment was also studied. Size of thrombus formed on pig aorta strips was measured after a 5-minute perfusion at low shear rate with blood from the patient/control subject. Results: Thrombus formation was inhibited by AZD0837 and ximelagatran. Relative to untreated patients, a 50% reduction of thrombus size was estimated at plasma concentrations of 0.6 and 0.2 μmol/L for ARH067637 and melagatran, respectively. For patients receiving VKA treatment, the thrombus size was about 15% lower compared with healthy elderly controls. Conclusions: Effects of AZD0837 and ximelagatran on thrombus formation were similar or greater than for VKA therapy and correlated with plasma concentrations of their active forms. © 2011 Elsevier Ltd. All rights reserved.

Oral direct thrombin inhibitors (DTIs) have been developed and recently introduced for the treatment and prevention of systemic thromboembolism [1]. As a principal advantage over therapy with

Abbreviations: NVAF, Non-valvular atrial fibrillation; VKA, Vitamin K antagonist; INR, International normalized ratio; DTI, Direct thrombin inhibitor; TTA, Total thrombus area; IR, Immediate-release; PK, Pharmacokinetic(s); PD, Pharmacodynamic(s); APTT, Activated partial thromboplastin time; F1 + 2, Prothrombin fragment 1 + 2; TAT, Thrombin–antithrombin complex; ACT, Activated coagulation time; TCT, Thrombin clotting time; PAP, Plasmin–antiplasmin complex; FPA, Fibrinopeptide A; β-TG, β-thromboglobulin; CRP, C-reactive protein; ICAM-, Intracellular adhesion molecule 1; VCAM-1, Vascular cell adhesion molecule 1; Cmax, Maximum plasma concentration; tmax, Time to maximum plasma concentration; AUCτ, Area under the plasma concentration–time curve during the dosing interval at steady state; t½, Half-life; CI, Confidence interval; ULN, Upper limit of normal; CV, Coefficient of variation; SD, Standard deviation. ⁎ Corresponding author at: Universitätsklinik für Klinische Pharmakologie, Allgemeines Krankenhaus Wien, Währinger Gürtel 18–20, A-1090 Vienna, Austria. Tel.: +43 1 40400 2981; fax: +43 1 40400 2998. E-mail address: [email protected] (M. Wolzt). 0049-3848/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.thromres.2011.08.018

vitamin K antagonists (VKAs), these drugs offer a predictable anticoagulant effect with an improved safety profile, and they do not require frequent coagulation monitoring and dose adjustments. Thrombin has a central role in the coagulation cascade and is responsible for clot formation via its potent activation of platelets and formation of the fibrin network [2,3]. The novel oral anticoagulant AZD0837 is a prodrug, which, after oral administration, is rapidly absorbed and bioconverted to its active form AR-H067637 [4], a selective and reversible DTI [5]. Animal studies have demonstrated that the active form inhibits thrombus formation in rat venous and arterial thrombosis models with no or minor increases in bleeding [6]. In clinical Phase II studies, treatment with AZD0837 for the prevention of stroke and systemic embolic events in patients with atrial fibrillation has shown promising safety and antithrombotic effect [7,8]. Investigation of pharmacological profiles of novel antithrombotic agents in experimental disease models and effects on biomarkers with potential clinical relevance may guide dose selection. The perfusion chamber method, originated by Badimon et al. [9], has been used

e84

M. Wolzt et al. / Thrombosis Research 129 (2012) e83–e91

to assess the antithrombotic effect of new anticoagulants in healthy subjects [10–12]. A comparison of drugs can be performed in this human disease model, using a perfusion chamber that gives a low shear rate equivalent to flow conditions in the venous system and assumed to simulate the thrombus formation in the cardiac atria of the heart in patients with atrial fibrillation. The primary aim of the present study was to investigate the effect of oral AZD0837 treatment on thrombus formation in patients with non-valvular atrial fibrillation (NVAF). Low-shear-rate perfusion chambers with denuded pig aorta as the thrombogenic surface were used. The first study compared the effect of AZD0837 with ximelagatran, which is an oral DTI that has been studied previously. Ximelagatran is also a prodrug, which is bioconverted to the active form, melagatran. The patients received AZD0837 at different doses and ximelagatran at a therapeutic dose regimen used in clinical Phase III studies for prevention of stroke in patients with NVAF [13,14]. The second study compared AZD0837 with dose-adjusted VKA (target international normalized ratio [INR] 2–3). The size of the thrombus formed on the thrombogenic surface of the perfusion chamber was evaluated by measurement of D-dimer concentration of the plasmindegraded thrombus. This approach is more accurate than the traditional morphometric analysis of ex vivo thrombus formation in studies employing collagen-coated parallel-plate perfusion chambers [15]. In study I, total thrombus area (TTA) was also evaluated by morphometric analysis of selected slices of the paraffin-embedded thrombus, which was the method used in previous perfusion chamber studies with denuded porcine aorta strips [10–12]. In addition, the effects on coagulation-time assays and biomarkers of thrombogenesis were investigated. For comparison, in a third study, ex vivo thrombus formation and biomarkers were assessed in healthy elderly subjects without NVAF or anticoagulant treatment. Methods Study participants Patients with atrial fibrillation and anticoagulant treatment Caucasian patients with persistent or permanent NVAF, and with at least one additional risk factor for stroke, were enrolled into two studies – 46 patients to study I (study code D1250C00011) and 32 patients to study II (study code D1250C00025). Medical history was obtained and a complete health examination that included a physical examination, monitoring of vital signs, electrocardiogram and laboratory screen was done at the pre-entry and follow-up visits. Study I: Thirty-eight patients (33 men and 5 women) with a mean age of 64 years were randomized to receive drug treatment; 35 completed the study. Study II: Twenty-seven patients (22 men and 5 women) with a mean age of 65 years were randomized to receive drug treatment; 25 completed the study (16 of these patients had previously been included in, and completed, study I). Elderly subjects without atrial fibrillation or anticoagulant treatment Study III: Twenty Caucasian subjects (12 men and 8 women) with a mean age of 75 years were included (study code D1250M00002). A medical history and information about use of medication were registered and a physical examination including a laboratory screen was performed. The study was performed to obtain control data from elderly subjects without NVAF or anticoagulant and antiplatelet treatment within 10 days before the experimental day. All three studies were approved by the Ethics Committee of the Medical University of Vienna and the Allgemeines Krankenhaus, Vienna, Austria. Studies were conducted in compliance with the Declaration of Helsinki, including current revisions, and Good Clinical Practice guidelines. Written informed consent was obtained from all participants. For study I, enrolment began in January 2005

and the last patient completed the study in March 2006. For study II, enrolment began in June 2006 and the last patient completed the study in December 2006. For the elderly control subjects (study III), the first subject was enrolled in July 2006 and the last subject completed the study in August 2006.

Study Design Patients with atrial fibrillation and anticoagulant treatment Studies I and II followed an open, randomized, parallel-group design and enrolled patients with NVAF who were treated with VKAs. Treatment with antiplatelet agents (including acetyl salicylic acid) was not allowed within 10 days before randomization, and treatment with fibrinolytic agents was not allowed within 30 days before randomization. After enrolment, VKA treatment was discontinued and when the international normalized ratio (INR) was ≤ 2 the patients were randomised to study drug (in study II it was also required that after the partial wash-out INR should increase at least 0.3 for patients receiving VKA during the study period). The minimum change in INR after partial VKA washout was selected to allow the presence of an additional therapeutic drug effect following randomization and study drug treatment. A complete VKA washout was not possible for ethical reasons due to the risk of thromboembolic events. Study I: Patients were randomized to either AZD0837 at a dose of 150 mg (n = 12), 250 mg (n = 2) or 350 mg (n = 11), or ximelagatran 36 mg (n = 13), twice daily for 10 to 14 days. All doses were given as immediate-release (IR) tablets. After an amendment, the 250 mg treatment group was added. However, due to a premature discontinuation of the study when ximelagatran was withdrawn from the market because of safety concerns, only two patients were included in this group. The measurements of thrombus formation in the perfusion chamber were conducted at steady state (after 10 to 14 days of treatment) on three occasions: i) at trough (before the dose given in the morning on the last treatment day), ii) at peak plasma concentrations (3 hours after dosing on the last treatment day) and iii) in the ximelagatran group 12 hours after the last dose and for the AZD0837 treatment groups 24 hours after the last dose. Study II: Patients received either AZD0837 (IR tablets) 250 mg twice daily (n = 14) or VKA (phenprocoumon, Marcoumar ®, Roche, Austria; n = 13) titrated to an INR of 2–3 for 10–14 days. The measurements of thrombus formation in the perfusion chamber were conducted on four occasions: i) before cessation of previous VKA treatment at an INR of 2–3, ii) after cessation of VKA and partial washout, achieving an INR of ≤ 2 and following randomization on the last treatment day, iii) before dose and iv) at 3 hours after dosing. For both studies, a parallel-group design was chosen as the variability in the primary variable was expected to be related to the perfusion chamber method itself, rather than to interpatient variability in the response to the drug, and to minimize the length of the study period for each patient. Due to practical reasons, and as the assessment of the primary variable was obtained by objective laboratory methods, these were open-label studies. A treatment period of a minimum of 10 days ensured that steady-state pharmacokinetics (PK) of the drugs under study were achieved without a residual effect of preceding VKA therapy on the thrombin inhibitors.

Elderly subjects without anticoagulant treatment Study III: In the control group with elderly subjects without anticoagulant treatment, venous blood sampling and a perfusion chamber experiment were conducted on a study day up to 21 days after the health screening.

M. Wolzt et al. / Thrombosis Research 129 (2012) e83–e91

Perfusion chamber experiments Perfusion chambers were made of a Plexiglas block through which a cylindrical hole of 0.2 cm in diameter was machined [9]. Three serially placed chambers were used in study I and two chambers were used in studies II and III. Each chamber contained a thrombogenic surface (pig aorta tunica media). The rheological conditions in the chambers simulated venous blood flow (the estimated shear rate calculated for an equivalent circular tube is 212 s-1; this condition is not fully met when the thrombogenic surface is introduced into the chamber). Before blood perfusion, the system was perfused with 0.9% sodium chloride to ensure no leaks were present and to remove air bubbles. Venous blood was drawn from a vein in the arm through an 18 G cannula with a pump (Masterflex® L/S™, Cole-Parmer Instrument Company, Vernon Hills, IL, USA). Five millilitres of blood were discarded before each perfusion. The aorta pieces were perfused at 10 mL/min for 5 minutes, followed by a 30-second perfusion with 0.9% sodium chloride. Size of thrombus formed on the thrombogenic surface of two chambers was evaluated by measurement of the concentration of D-dimers (Asserachrome D-Dimer, Diagnostica Stago) of the plasmindegraded thrombus, using the mean value of the concentrations in the two perfusion chambers at each time point. The thrombus on the pig aorta was degraded with plasmin solution (0.5 mL of 0.1 M phosphate buffered saline pH 7.4 and 0.05 mL plasmin, 10 U/mL, Chromogenix, Milano, Italy). The tubes were placed in a water bath (37 °C) for 60 minutes, mixed every 15 minutes and the reaction was stopped by addition of 50 μL aprotinin (6150 U/mL, from bovine lung, Fluka Chemie GmbH, Buchs, Switzerland). In study I, from the third perfusion chamber, TTA was also evaluated using microscopic morphometry [16,17]. The pig aorta tunica media sample was prepared and embedded in paraffin. These samples were then sectioned and stained for the presence of total thrombus [17]. Evaluations were made of the TTA on up to 6 sections taken perpendicular to the long axis of the pig aorta tunica media surface from the perfusion chamber and a mean area value (μm 2/mm) was calculated.

Pharmacodynamic biomarkers Pharmacodynamic (PD) biomarkers were assessed in plasma from venous blood samples collected via an indwelling catheter at the same time points as when perfusion chamber experiments were performed. Activated partial thromboplastin time (APTT, STA-APTT Automate 5/Diagnostica Stago, Asnieres, France), prothrombin time (expressed as INR, STA Neoplastin Plus, Diagnostica Stago), D-dimers (TintElize, Trinity Biotech/Biopool), and prothrombin fragment 1 + 2 (F1 + 2, Enzygnost® F1 + 2 micro, Dade Behring, Marburg, Germany) were measured using standardized laboratory methods at the Clinical Institute for Laboratory Medicine, Allgemeines Krankenhaus, Vienna, Austria. Thrombin generation (Calibrated Automated Thrombogram, CAT, Synapse, Maastricht, The Netherlands) was measured at AstraZeneca R&D Mölndal, Sweden. Peak thrombin activity after 5 pmol/L tissue factor activation was assessed in study subjects in relation to pooled human platelet-poor plasma from untreated healthy volunteers [18]. In addition, thrombin–antithrombin complex (TAT, Enzygnost ® TAT, Dade Behring) was measured in study II. Additional methods assessed in study I were low-range activated coagulation time (ACT, bedside test measured in whole blood collected by repeated venous puncture, Medtronic HemoTec, Minneapolis, USA), thrombin clotting time (TCT, STA Thrombin Diagnostica Stago), plasmin–antiplasmin complexes (PAP Complex ELISA kit Technoclone, Vienna, Austria), fibrinopeptide A (FPA; ZYMUTEST FPA, Hyphen BioMed, Andrésy, France), soluble P-selectin (human P-Selectin/CD62P, R&D Systems, Minneapolis, USA), β-thromboglobulin (β-TG, Asserachrome β-Thromboglobulin, Diagnostica Stago), high-sensitive Creactive protein (CRP, Abbott Diagnostics, Wiesbaden, Germany),

e85

intracellular adhesion molecule 1 (ICAM-1/CD54, R&D Systems) and vascular cell adhesion molecule 1 (VCAM-1/CD106, R&D Systems). Pharmacokinetic analysis Venous blood samples for PK analysis were collected via an indwelling catheter at the same time points as when perfusion chamber experiments were performed. Plasma concentrations of AZD0837, the intermediate metabolite AR-H069927 and AR-H067637 were analysed by Analytico Medinet B.V., Breda, The Netherlands. Melagatran plasma concentrations were analysed by Quintiles AB, Uppsala, Sweden. Both methods used high-performance liquid chromatography mass spectrometry. The lower limit of quantification was 10 nmol/L for all analytes. In study I, frequent plasma samples were collected at pre-dose and 1, 2, 3, 4, 5, 6, 8, 10, 12 and 24 hours after the morning dose on Days 10–14, and PK parameters for AR-H067637 and melagatran were estimated by non-compartmental analysis using WinNonlin Professional version 4.1 (Pharsight Corporation, Mountain View, USA). The PK parameters estimated were the maximum plasma concentration (Cmax), the time to Cmax (tmax), area under the plasma concentration–time curve during the dosing interval at steady state (AUCτ) and terminal half-life (t½). AUCτ was calculated by the log-linear trapezoidal method and t½ as ln2/λ, where λ is derived from regression analysis of the terminal part of the log concentration–time curve. Statistical Analysis Patients in studies I and II were randomized sequentially to treatment. Study III was non-randomized. All studies were exploratory and thus no formal determinations of sample size were made and no formal statistical hypothesis testing was done in any of the studies. The thrombus size (D-dimer concentration; studies I and II) and TTA (study I) were log-transformed and analysed using a mixedeffect model with fixed effects for treatment and time and the interaction between treatment and time, and patient as a random effect. Results were transformed back to the original scale and are presented graphically as geometric means with 95% confidence intervals (CIs). Potential relationships between plasma concentration of AR-H067637 or melagatran and PD variables were investigated using an exploratory approach. For APTT and ACT the square root of plasma concentration was used in the model. The model for TCT included plasma concentration, and peak thrombin activity was log-transformed before being entered into the model, which included plasma concentration. All models included treatment and interaction between the plasma concentration variable and treatment, and for all models the repeated statement in the SAS v8.2 proc mixed procedure was used to model the covariance structure within patients. The covariance type was compound symmetric. Results are displayed graphically. To explore the relationship between thrombus size (D-dimer concentration) and plasma concentration, the AZD0837 data from studies I and II were pooled. Thrombus size was log-transformed before being entered into the model, which included plasma concentration. A corresponding model was fitted for ximelagatran. The covariance structure within patients was handled as described above. Results are displayed graphically and parameter estimates are tabulated. Results Patient characteristics Table 1 summarizes baseline characteristics of study participants. Many of the NVAF patients had hypertension (60 of 65 subjects) and dyslipidaemia (27 of 65 subjects), thirteen patients had a history of transient cerebral ischaemia or stroke and fifteen patients had diabetes mellitus. Cardiovascular disease was also reported for several patients.

e86

M. Wolzt et al. / Thrombosis Research 129 (2012) e83–e91

Table 1 Characteristics of patients with NVAF (studies I and II) and elderly control subjects (study III). NVAF

NVAF

Control

Study I

Study II

Study III

Age, years (range) Male/female, n Body mass index, kg/m² (range) Cardiovascular history, n (%) CHF Diabetes mellitus Dyslipidaemia Hypertension Myocardial infarction Stroke TIA

AZD0837

AZD0837

AZD0837

Ximelagatran

AZD0837

VKA

Elderly subjects

150 mg (n = 12)

250 mg (n = 2)

350 mg (n = 11)

36 mg (n = 13)

250 mg (n = 14)

(n = 13)

(n = 20)

67 (62–76) 8/4 32 (25–40)

55 (47–63) 2/0 26 (24–28)

64 (49–78) 10/1 27 (20–35)

62 (48–79) 13/0 31 (23–40)

65 (50–84) 11/3 30 (22–40)

66 (50–78) 11/2 29 (23–42)

75 (70–83) 12/8 28 (21–36)

1 (8) 3 (25) 8 (67) 11 (92) 1 (8) 2 (17) 0 (0)

1 (50) 0 (0) 1 (50) 2 (100) 0 (0) 0 (0) 0 (0)

0 (0) 1 (9) 2 (18) 11 (100) 2 (18) 3 (27) 1 (9)

0 (0) 3 (23) 5 (38) 11 (85) 5 (38) 1 (8) 1 (8)

1 (7) 4 (29) 6 (43) 13 (93) 4 (29) 2 (14) 2 (14)

3 (23) 4 (31) 7 (54) 12 (92) 2 (15) 1 (8) 0 (0)

0 (0) 1 (5) 8 (40) 10 (50) 0 (0) 1 (5) 1 (5)

CHF, congestive heart failure; NVAF, non-valvular atrial fibrillation, TIA, transient ischemic attack; VKA, vitamin K antagonist.

Co-medication was frequently used but there was no use of disallowed medication that was considered to affect the validity of the results. The most common medications were angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, β-adrenoceptor antagonists, calcium channel blockers and statins (co-medication data not shown). To a lesser extent, subjects in the elderly control group had cardiovascular risk factors, disease history and related treatments.

For the group receiving ximelagatran, comparable trough measurements 12 hours after dose were carried out on two occasions. The coefficient of variation (CV) for the within-subject ratio in thrombus size between the

Safety Study I: Three patients discontinued the study prematurely: two patients on AZD0837 350 mg and one on ximelagatran. One patient on AZD0837 was discontinued due to predefined stop criterion prolongation in APTT N3x upper limit of normal (ULN) after 8 days of treatment. Another patient on AZD0837 was discontinued due to hospitalization with heart failure, which was not considered to be related to the study drug. One patient receiving ximelagatran was discontinued due to high serum bilirubin concentrations, which were already present before start of treatment. Study 2: Two patients discontinued the study prematurely: one patient on AZD0837 discontinued due to increased APTT N3 x ULN and another patient randomized to the VKA group did not fulfil the protocol inclusion criteria. Adverse events were few and without apparent differences between the treatment groups. A reversible mean increase of 10% in serum creatinine was seen during exposure to AZD0837. Effect on thrombus formation Study I: AZD0837 compared with ximelagatran The time course of thrombus size, measured as plasmin-degraded thrombus D-dimers, and TTA is presented in Fig. 1. At steady state, AZD0837 at a dose of 150 mg and 350 mg reduced thrombus size from mean (standard deviation [SD]) 48.1 (13.7) and 44.6 (10.5) μg/ml to 27.7 (13.8) and 21.1 (14.9) μg/ml at 3 hours after dosing, respectively (all p b 0.01). For ximelagatran, a thrombus size reduction from 52.1 (30.4) to 22.2 (24.3) μg/ml was noted (p b 0.001). This reduction by 50-70% returned to pre-dose values 24 hours after the last dose of AZD0837, and 12 hours after the last dose of ximelagatran. Comparable changes were observed for TTA, which was reduced by 30–60% at 3 hours compared with pre-dose. The correlation between thrombus size D-dimer and TTA estimated by linear regression was R = 0.207 (p = 0.034); evaluated by Kendall's Tau correlation coefficient it was R = 0.253 (p = 0.0001). Thrombus size evaluated by plasmin-degraded thrombus appeared to have lower variability than TTA obtained by morphometric analysis.

Fig. 1. Thrombus size (Panel A) and TTA (Panel B) in patients randomized to AZD0837 (open circles 150 mg bid; open triangles 350 mg bid) or ximelagatran (closed squares). Thrombus size was measured at steady state before dosing (pre-dose) and post-dose at 3 h and 24 h after the last morning dose. Ximelagatran was also given in the evening so the final measurement was at 12 hours after the last dose. Data are presented as geometric means with 95% CIs. For thrombus size shown in Panel A, the dashed horizontal line and grey shading represent geometric mean thrombus size and corresponding 95% CIs in elderly controls.

M. Wolzt et al. / Thrombosis Research 129 (2012) e83–e91 Table 2 Plasma concentrations of AR-H067637 and melagatran at steady state in patients with NVAF treated with AZD0837 or ximelagatran, respectively. Time after last dose (h)

Pre-dosea 3h 12 or 24 hb

AZD0837

AZD0837

AZD0837

Ximelagatran

150 mg

250 mg

350 mg

36 mg

(study I)

(study II)

(study I)

(study I)

0.41 (0.15) 0.70 (0.22) 0.18 (0.08)

0.54 (0.26) 0.75 (0.27) –c

0.67 (0.19) 1.20 (0.32) 0.37 (0.13)

0.16 (0.15) 0.46 (0.17) 0.17 (0.13)

a Pre-dose samples were obtained at 12 hours after the last dose. bIn both studies, blood samples were collected during steady -state at the same time as the perfusion chamber experiments were done. In study I, the last sample was obtained at 24 hours after the last dose of AZD0837 and at 12 hours after the last dose of ximelagatran. cNo sample collected. Data are presented as means (SD).

two ximelagatran trough assessments was 23% for thrombus size and 89% for TTA. Plasma concentrations of the active forms of AZD0837 and ximelagatran, measured at the same time points as the perfusion chamber experiments, are shown in Table 2. The thrombus size decreased with increasing plasma concentrations of the active forms of the DTIs (Fig. 3). Parameters of the regression analysis of thrombus size versus plasma concentrations are shown in Table 3. Frequent blood samples were collected during the dosing intervals on the experimental days that allowed description of the full PK profile for the active forms of AZD0837 and ximelagatran. Plasma concentrations of AZD0837 and the intermediary metabolite were also measured (data not shown), and these PK profiles supported a rapid absorption of AZD0837 and metabolism via the intermediate metabolite to the active form, AR-H067637. The tmax for AR-H067637 varied from 2 to 6 hours with a median of about 3 hours. The mean (SD) Cmax of AR-H067637 was 0.72 (0.22) and 1.25 (0.29) μmol/L for AZD0837 150 mg and 350 mg treatment groups, respectively. The mean t½ of AR-H067637 ranged from 13 to 17 hours after the different doses of AZD0837. The AUCτ for AR-H067637 was 6.04 and 10.3 μmol·h/L after the 150 mg and 350 mg AZD0837 doses, respectively, with low interindividual variabilities (CV) of 33% and 19%. These AUCτ values correspond to average steady-state concentrations for AR-H067637 of 0.50 and 0.86 μmol/L after the 150 mg and 350 mg AZD0837 doses, respectively. The mean (SD) Cmax of melagatran was 0.48 (0.19) μmol/L and the mean t½ was 5.3 hours, with a median tmax of 3 hours. The AUCτ of melagatran was 3.42 μmol·h/L, corresponding to a steady-state concentration of 0.29 μmol/L, with a CV of 58%. Study II: AZD0837 compared with VKA The time course of thrombus size is presented in Fig. 2. The duration of the partial VKA washout was between 2 and 13 days after the last VKA intake. After partial washout, mean (SD) thrombus size was 46.0 (15.8) and 49.7 (18.0) μg/ml in patients randomized to AZD0837 and VKA, respectively (p=n.s. between groups). During steady-state treatment with AZD0837, pre-dose thrombus size (12 hours after previous dose) was 38.9 (13.6) and decreased to 31.0 (14.9) μg/ml 3 hours after AZD0837 administration (pb 0.05 vs. pre-dose). Thrombus size remained unchanged

e87

in patients on VKA treatment in steady state at pre-dose (49.6 [12.1]) and 3 hours after dosing (52.1 [23.4], p=n.s. vs. pre-dose). Mean steady state plasma concentrations of AR-H067637 are presented in Table 2. The effect on thrombus size was associated with the plasma concentration of AR-H067637 (Fig. 3). The intercept and slope parameters from the regression analysis for study II were similar to those estimated in study I (Table 3); therefore, an analysis was carried out where AZD0837 data were combined for both studies. Based on the regression analysis for combined AZD0837 data (Table 3), the estimate of thrombus size for untreated patients was 63.5 μg/mL. Compared with this baseline estimate, the thrombus size for patients treated with VKA (44 μg/mL; 95% CI 33–57 at 3 hours after dose) was decreased by about 30%. From the concentration–effect relationship estimated for AR-H067637 from studies I and II combined, a similar decrease in thrombus size is predicted at a plasma concentration of about 0.3 μmol/L. Study III: Elderly subjects without anticoagulant treatment The least squares geometric mean value of thrombus size was 51.5 μg/mL (95% CI 36–73), which is indicated in Figs. 1 and 2. Pharmacodynamic markers Study I: AZD0837 compared with ximelagatran After cessation of VKA, mean (SD) prothrombin time expressed as INR was reduced to 1.2 (0.1) when measured at pre-dose after 10– 14 days of treatment with AZD0837 or ximelagatran. The coagulation assays APTT, ACT and TCT showed prolonged coagulation times, and peak thrombin activity was decreased. These effects correlated to the plasma concentrations of the active forms of AZD0837 and ximelagatran (Fig. 4). For peak thrombin activity, the plasma concentration–effect relationships for AR-H067637 and melagatran were similar, while for APTT, ACT and TCT there was a steeper linear plasma concentration-dependent relationship for melagatran compared with AR-H067637. The largest relative difference was observed for TCT, which was the most sensitive coagulation assay. Venous D-dimer, F1 + 2, PAP, FPA, β-TG (platelet corrected), ICAM, VCAM-1 and high-sensitive CRP concentrations were below the upper reference limit for all treatment groups and similar across treatments and sampling points (data not shown). Concentrations of soluble P-selectin (platelet corrected) were lower in subjects randomized to AZD0837 compared with ximelagatran at steady state, but were unchanged after dosing of the DTIs (data not shown). Study II: AZD0837 compared with VKA In patients randomized to VKA, INR was reduced from 2.5 (0.6) to 1.7 (0.5) after temporary cessation of VKA and increased to 2.6 (0.5) at steady state. In patients receiving AZD0837, INR was reduced from 2.6 (0.4) to 1.6 (0.3) after VKA washout. At steady state with AZD0837, INR was 1.6 (0.3) before dosing and increased to 1.9 (0.5) 3 hours after dosing. PD markers in venous plasma are presented in Table 4. The peak thrombin activity increased after partial washout of VKA, consistent with a reduced anticoagulation. Peak activity decreased

Table 3 Regression parameters estimated from regression analysis of the correlation between thrombus size (μg/mL) and plasma concentrations of AR-H067637 or melagatran (μmol/L). Treatment (study)

AZD0837 (study I) AZD0837 (study II) AZD0837 (combined) Ximelagatran (study I)

Intercept

95% CI

Slope

Estimate

Lower

Upper

4.16 4.12 4.15 4.31

3.86 3.76 3.90 3.97

4.47 4.48 4.40 4.65

− 1.22 − 1.00 − 1.16 − 3.56

Baselinea

95% CI Lower

Upper

− 1.60 − 1.45 − 1.47 − 4.27

− 0.85 − 0.54 − 0.85 − 2.85

b IC50

(μmol/L) 64.2 61.6 63.5 74.4

0.57 0.70 0.60 0.19

a Baseline is the thrombus size estimated without treatment (zero concentration) obtained from the intercept of the regression analysis of the log-transformed response (natural logarithms). bIC50 is the concentration of AR-H067637 or melagatran that reduces thrombus size by 50% compared with no treatment.

e88

M. Wolzt et al. / Thrombosis Research 129 (2012) e83–e91

concentrations of the active form (data not shown), which is consistent with the results in study I. Study III: Elderly subjects without anticoagulant treatment Values for the PD markers in venous plasma are shown in Table 4. Discussion

Fig. 2. Thrombus size in patients randomized to AZD0837 (open circles) or VKA (closed squares). Thrombus size was measured during VKA treatment, after partial VKA washout and at steady state before dosing (pre-dose) and 3 hours after dosing (post-dose). Data are presented as geometric means with 95% CIs. The dashed horizontal line and grey shading represent geometric mean thrombus size and corresponding 95% CIs in elderly controls.

again after treatment with VKA or AZD0837, indicating that both treatments exerted systemic anticoagulation. At steady state, peak thrombin activity during VKA treatment was stable, while for AZD0837 the peak thrombin activity mirrored the changes in the plasma concentration of AR-H067637. Peak thrombin activity was higher pre-dose and decreased at 3 hours post-dose of AZD0837 to a level comparable with that in the VKA group. As observed in study I, venous D-dimer plasma concentrations were below the upper reference limit for all treatment groups and similar for both treatment groups at all sampling occasions. F1 + 2 and TAT plasma concentrations were also generally low and below the upper limit of normal. For subjects receiving AZD0837, the means and the variability of TAT and F1 + 2 levels were higher than during VKA treatment. Median and range of D-dimers, TAT and F1 + 2 levels are given in Table 4 because of the skewed distributions of these variables especially for the AZD0837 group. For patients treated with AZD0837, the prolongation of APTT and decrease in peak thrombin activity correlated with the plasma

Fig. 3. Effects on thrombus size versus plasma concentration of AR-H067637 and melagatran. Data shown for ximelagatran 36 mg bid (crosses) and AZD0837 150 mg and 350 mg bid (triangles) treatments given in study I and AZD0837 250 mg bid treatment given in study II (inverse triangles). The lines show the estimated concentration–effect relationships from the regression analyses of AZD0837 (solid line) and ximelagatran (dashed line) data.

The antithrombotic effect of the oral DTI AZD0837 was investigated in patients with NVAF, using the perfusion chamber method first described by Badimon et al. [9]. To our knowledge, this is the first perfusion chamber study performed in patients with NVAF. At the doses studied, AZD0837 reduced thrombus size to an extent comparable to ximelagatran and greater than VKA, which were used as active controls. The results complement dose–response data obtained in clinical studies and guide dose selection for pivotal clinical studies to document efficacy and safety of AZD0837. Two Phase II dose-finding studies in patients with NVAF treated with an IR formulation of AZD0837 for 3 months [7] or an extended-release formulation for 3–9 months [8], suggest that a daily dose of 300 mg provides a similar suppression of thrombogenesis as dose-adjusted VKA. Thrombus size was assessed by analysis of the D-dimer concentration of the plasmin-degraded fibrin in the thrombus. It has been reported that this quantification of fibrin deposition on a thrombogenic surface is more accurate than TTA analysis in a different perfusion chamber setting [15]. In study I it was shown that the within-subject coefficient of variation of the D-dimer method was lower, suggesting improved reproducibility, compared with the TTA method used in previous studies [10–12] that measures TTA by morphometric analysis from selected paraffin-embedded slices. TTA determined by morphometric histological analysis depends on the analysis of the two-dimensional irregular thrombus area of selected slices and is influenced by the embedding procedure of the tissue. Thus, the signal-to-noise ratio favours analysis of thrombus size by measurement of D-dimer concentrations of the plasmin-degraded fibrin content in the whole thrombus formed on the thrombogenic surface in the perfusion chamber. Furthermore, both fibrin and platelets are included in the measurement of TTA. It has been demonstrated that thrombin inhibition with ximelagatran results in greater inhibitory effect on fibrin formation than on platelet adhesion and aggregation [10], and fibrin content measured from the degraded thrombus in the present studies is therefore expected to be a more sensitive method for evaluating the effect. However, both thrombus D-dimer and TTA measurements can only serve as pharmacodynamic biomarkers and do not allow extrapolation to clinical efficacy. AZD0837 and ximelagatran dose-dependently decreased thrombus size and exerted greater inhibitory effects on thrombus size 3 hours after dosing at steady state compared to pre-dose (trough plasma concentration). This effect was also demonstrated by the relationship between thrombus size and the plasma concentrations of the active forms of AZD0837 (AR-H067637) and ximelagatran (melagatran). Regression lines for the association between thrombus Ddimer and TTA were similar for the direct thrombin inhibitors under study (data not shown). The results are consistent with previously reported associations between melagatran plasma concentrations and the effects on TTA as well as total fibrin area for perfusion chamber thrombus formation studied in young healthy subjects [10]. In the present comparison of ximelagatran and AZD0837, essentially complete inhibition of thrombus formation was achieved for both compounds. Melagatran was effective at lower concentrations than ARH067637, for which the concentration–effect relationships were similar in studies I and II. Based on the IC50 estimated from the regression analyses, melagatran was approximately three times more potent. This is in accordance with previously reported results from an in vitro study using the same method with whole blood from young healthy volunteers with increasing concentrations of AR-H067637 or melagatran added [19].

M. Wolzt et al. / Thrombosis Research 129 (2012) e83–e91

e89

Fig. 4. Effects on coagulation time assays, APTT, ACT and TCT, and peak thrombin activity versus plasma concentrations of AR-H067637 (open circles) and melagatran (crosses).

compared with untreated healthy controls. Relative to the estimate for untreated NVAF patients, the thrombus size for patients treated with VKA was decreased by about 30%. A direct comparison between the effects of VKA and the thrombin inhibitors in the same group of NVAF patients cannot be made with the present study design, and the differences across the studied NVAF patient cohorts should be interpreted with caution. After partial VKA washout (INR ≤2), period systemic anticoagulation was still detectable, and the changes in thrombus size in patients receiving VKA was minor. APTT was prolonged and peak thrombin generation was lower compared with the untreated healthy controls. This suggests that residual activity of indirectly acting VKA persists at INR ≤2. Consistent with the delayed pharmacodynamic onset of action for VKA, there were no apparent changes between pre-dose and 3 hours after dosing at steady state for the effects on thrombus size and other coagulation markers. Effects on coagulation time assays and biomarkers of thrombogenesis were investigated to complement the comparisons of antithrombotic ef-

Treatment with VKA, dose adjusted to an INR of 2–3, was less effective in reducing perfusion chamber thrombus formation than the DTIs. Untreated NVAF patients were not studied for ethical reasons because of the risk of stroke and thromboembolic events without anticoagulant therapy. Instead, data for comparison were obtained from healthy elderly subjects without NVAF or anticoagulant treatment. Thrombus size was about 15% lower in patients receiving VKA relative to the healthy elderly subjects without anticoagulant treatment, although confidence intervals were overlapping. A greater difference might have been expected, but the comparison is confounded by differences between these two groups of subjects, for example in terms of age and disease history. The impact of the prothrombotic state of untreated NVAF patients in this chamber model is not known. However, the intercepts of the concentration–effect relationships provided estimates of the thrombus size for untreated NVAF patients that were higher than thrombus size for elderly subjects without NVAF or anticoagulant treatment, suggesting a more prothrombotic state and greater thrombus formation for an untreated patient with NVAF

Table 4 Coagulation variables (mean and SD) in venous blood in patients with atrial fibrillation on anticoagulant treatment randomized to AZD0837 or VKA (Study II), and in elderly subjects without anticoagulant treatment (Study III). APTT, sec Time point

AZD0837

During VKA (INR 2–3) After partial VKA washout (INR ≤2) 43.5 (7.9) At steady state pre-dose 64.8 (14.5) At steady state 3 h post-dose 76.4 (20.8)

Peak thrombin activity, %

TAT, μg/L

D-dimer, μg/L

F1 + 2, pmol/L

VKA

AZD0837 VKA

AZD0837

VKA

AZD0837

VKA

AZD0837

VKA

45.6 (7.1) 58.2 (12.7) 58.8 (13.9)

36 58 54 36

2.6 2.6 2.5 2.1

2.3 2.1 2.0 1.9

20 27 31 26

52 (7–144) 46 (7–208) 31 (10–227) 31 (14–245)

68 (24–115) 93 (33–196) 122 (54–490) 103 (72–1709)

68 80 59 46

(11) (16) (23) (22)

33 55 35 37

(10) (19) (22) (18)

(1.6–6.3) (1.6–8.6) (1.6–67) (1.6–60)

(1.6–4.7) (1.6–4.6) (1.6–4.9) (1.6–11)

(7–476) (7–182) (7–121) (7–146)

(39–132) (44–135) (14–94) (23–149)

Elderly subjects without anticoagulant treatment APTT, sec

Peak thrombin activity, %

TAT, μg/L

D-dimer, μg/L

F1 + 2, pmol/L

33.9 (3.8)

116 (21)

2.3 (1.6–40)

145 (54–1000)

253 (130–621)

APTT, activated partial thromboplastin time; TAT, thrombin antithrombin complex; F1 + 2, prothrombin fragment 1 + 2; INR, international normalized ratio; VKA, vitamin K antagonist. Peak thrombin activity is given as relative values (%) to a pool of plasma from untreated healthy volunteers. APTT and peak thrombin activity data are presented as means (SD). TAT, D-dimer and F1 + 2 are presented as median (range) as the distribution is skewed with some high values. Reference range or ULN was 1 to 4.1 μg/L for TAT, below 130 μg/L for D-dimer and 69 to 229 nmol/L for F1 + 2.

e90

M. Wolzt et al. / Thrombosis Research 129 (2012) e83–e91

fects on ex vivo thrombus formation for AZD0837, ximelagatran and VKA treatment. For the coagulation time assays APTT and ACT, and for peak thrombin activity, comparable plasma concentration–effect relationships were observed for the active forms of AZD0837 and ximelagatran. In contrast, the regression line for the TCT assay, where thrombin itself is added to the test assay, was steeper for melagatran compared to AR-H067637. This difference may be explained by the higher protein binding for AR-H067637 (fraction bound 71–88% [4]) compared with melagatran (less than 15% [20]) and thus a lower bioactive fraction in plasma, which gives a lower effect on TCT when thrombin is added directly. The difference in protein binding may be less important for the APTT, ACT and thrombin generation assays, where thrombin is generated from prothrombin and compound bound to plasma protein may have time to equilibrate with formed thrombin [21]. Thrombin generation in plasma, assessed as peak thrombin activity, was reduced as expected and as previously shown for ximelagatran [22]. In the presence of a thrombin inhibitor, the thrombin generation assay may underestimate the effect due to interaction with multiple substrates in the assay [18]. Compared to the measurement of the area under the thrombin generation curve (endogenous thrombin potential), the peak thrombin activity is influenced less by the measurement error that occurs because of the interaction with α2-macroglobulin. Dose-adjusted VKA (INR 2–3) decreased thrombin generation by about 70%, while AZD0837 and ximelagatran, depending on dose, depressed thrombin peak activity by 40–85%. Levels of venous D-dimers and other biomarkers of thrombogenesis were low and similar across patients at all measured time points at steady state and remained unchanged after dosing of the drugs under study, indicating continued antithrombotic status during the studies. In a previous clinical study, AZD0837 treatment was shown to suppress D-dimer levels in untreated patients with NVAF (VKA naïve) and the onset of effect was observed during the first 8 weeks of treatment [8]. F1+ 2, the activation peptide released after activation of prothrombin by factor Xa, was slightly lower in the VKA group compared with the AZD0837 group. This could be explained by the observation that VKA treatment results in lower concentrations of carboxylated prothrombin available for activation to thrombin. AZD0837 was well tolerated. The reversible 10% mean increase in serum creatinine with AZD0837 seen during the present study is consistent with findings in larger Phase II studies [7,8]. A study to evaluate renal function in healthy elderly subjects found that glomerular filtration rate measured by iohexol clearance was unaffected by AZD0837 [23], and that the likely mechanism for the serum creatinine increase is an inhibitory effect on transporter-mediated tubular secretion of creatinine. The modest elevation in serum creatinine is therefore unlikely to have any safety implications. An IR formulation of AZD0837 was used in the present studies. Pharmacokinetic properties of an extended-release formulation of AZD0837, as used in a dose-finding study in patients with NVAF [8], reflect a slower drug absorption. Corresponding changes in pharmacodynamics, such as antithrombotic effects on ex vivo thrombus formation, would be expected accordingly. Effects on coagulation time assays and thrombin generation were shown in the dose finding study [8] that are consistent with the results in the present studies. In summary, the present studies demonstrated an antithrombotic effect of the oral DTI AZD0837 in patients with NVAF that was similar to the effects of ximelagatran and more effective than VKA. In addition, the effects on several potential biomarkers of thrombogenesis and coagulation assays supported that treatment with AZD0837 resulted in comparable anticoagulation as ximelagatran and VKA. Evaluation of the antithrombotic effects using this ex vivo human disease model in the intended patient population with NVAF is useful to guide the dose selection for clinical trials of this novel oral anticoagulant.

Conflict of interest Michael Wolzt has received honoraria for consultancy work and has been a lecturer in symposia sponsored by AstraZeneca. Ghazaleh Gouya, Nicolai Leuchten and Stylianos Kapiotis have no conflicts of interest. Ulf G. Eriksson, Margareta Elg, Kajs-Marie Schützer, Sofia Zetterstrand, Malin Holmberg, and Karin Wåhlander are employees of the sponsor with stock ownership.

Acknowledgements The authors gratefully acknowledge the contributions of Carola Fuchs, RN to this study. The trial was sponsored by AstraZeneca, who were involved in the study design, data interpretation and, in conjunction with the authors, the decision to publish. Employees of the sponsor collected, and managed the data, and performed the data analysis. All authors had access to the clinical study data, and took part in their interpretation. Dr M Wolzt wrote the first draft of the manuscript and all authors reviewed and contributed to subsequent drafts of the manuscript, and approved the final version. Editorial assistance was provided by Lee Kempster, MediTech Media, UK, and was funded by AstraZeneca.

References [1] Eriksson BI, Quinlan DJ, Weitz JI. Comparative pharmacodynamics and pharmacokinetics of oral direct thrombin and factor Xa inhibitors in development. Clin Pharmacokinet 2009;48:1–22. [2] Dutt T, Toh CH. The Yin-Yang of thrombin and activated protein C. Br J Haematol 2008;140:505–15. [3] Weitz JI. Factor Xa or thrombin: is thrombin a better target? J Thromb Haemost 2007;5:65–7. [4] Johansson S, Cullberg M, Eriksson UG, Elg M, Dunér K, Jensen E, et al. Single-dose pharmacokinetics, pharmacodynamics and safety of AZD0837, a novel oral direct thrombin inhibitor, in young healthy male subjects. Int J Clin Pharmacol Therap 2011;49:258–67. [5] Deinum J, Mattsson C, Inghardt T, Elg M. Biochemical and pharmacological effects of the direct thrombin inhibitor AR-H067637. Thromb Haemost 2009;101: 1051–9. [6] Pehrsson S, Johansson K, Kjaer M, Elg M. Evaluation of AR-H067637, the active metabolite of the new direct thrombin inhibitor AZD0837, in models of venous and arterial thrombosis and bleeding in anaesthetised rats. Thromb Haemost 2010;104:1242–9. [7] Olsson SB, Rasmussen LH, Tveit A, Jensen P, Wessman P, Paniflov S, et al. Safety and tolerability of an immediate-release formulation of the oral direct thrombin inhibitor AZD0837 in the prevention of stroke and systemic embolism in patients with atrial fibrillation. Thromb Haemost 2010;103:604–12. [8] Lip GYH, Rasmussen LH, Olsson SB, Jensen EC, Persson AL, Eriksson U, et al. Oral direct thrombin inhibitor AZD0837 for the prevention of stroke and systemic embolism in patients with non-valvular atrial fibrillation: a randomised doseguiding, safety and tolerability study of four doses of AZD0837 versus vitamin K antagonists. Eur Heart J 2009;30:2897–907. [9] Badimon L, Turitto V, Rosemark JA, Badimon JJ, Fuster V. Characterization of a tubular flow chamber for studying platelet interaction with biologic and prosthetic materials: deposition of indium 111-labeled platelets on collagen, subendothelium, and expanded polytetrafluoroethylene. J Lab Clin Med 1987;110:706–18. [10] Sarich TC, Osende JI, Eriksson UG, Fager GB, Eriksson-Lepkowska M, Ohlsson L, et al. Acute antithrombotic effects of ximelagatran, an oral direct thrombin inhibitor, and r-hirudin in a human ex vivo model of arterial thrombosis. J Thromb Haemost 2003;1:999–1004. [11] Wahlander K, Eriksson-Lepkowska M, Nystrom P, Eriksson UG, Sarich TC, Badimon JJ, et al. Antithrombotic effects of ximelagatran plus acetylsalicylic acid (ASA) and clopidogrel plus ASA in a human ex vivo arterial thrombosis model. Thromb Haemost 2006;95:447–53. [12] Zafar MU, Vorchheimer DA, Gaztanaga J, Velez M, Yadegar D, Moreno PR, et al. Antithrombotic effects of factor Xa inhibition with DU-176b: Phase-I study of an oral, direct factor Xa inhibitor using an ex-vivo flow chamber. Thromb Haemost 2007;98:883–8. [13] Olsson SB. Executive Steering Committee on behalf of the SPORTIF III Investigators. Stroke preventions with the oral direct thrombin inhibitor ximelagatran compared with warfarin in patients with non-valvular atrial fibrillation (SPORTIF III): randomised clinical trial. Lancet 2003;362:1691–8. [14] SPORTIF Executive Steering Committee for the SPORTIF V Investigators. Ximelagatran vs warfarin for stroke prevention in patients with nonvalvular atrial fibrillation: a randomized trial. JAMA 2005;293:690–8.

M. Wolzt et al. / Thrombosis Research 129 (2012) e83–e91 [15] Orvim U, Barstad RM, Stormorken H, Brosstad F, Sakariassen KS. Immunologic quantification of fibrin deposition in thrombi formed in flowing native human blood. Br J Haematol 1996;95:389–98. [16] Hanson SR, Sakariassen KS. Blood flow and antithrombotic drug effects. Am Heart J 1998;135:132–45. [17] Dangas G, Badimon JJ, Coller BS, Fallon JT, Sharma SK, Hayes RM, et al. Administration of abciximab during percutaneous coronary intervention reduces both ex vivo platelet thrombus formation and fibrin deposition. Implications for a potential anticoagulant effect of abciximab. Arterioscler Thromb Vasc Biol Arterioscler Thromb Vasc Biol 1998;18:1342–9. [18] Hemker HC, Giesen P, Al Dieri R, Regnault V, de Smedt E, Wagenvoord R, et al. Calibrated automated thrombin generation measurement in clotting plasma. Pathophysiol Haemost Thromb 2003;33:4–15. [19] Elg M, Zachrisson H. Investigation of the antithrombotic effect of AR-H067637, the active metabolite of the oral anticoagulant AZD0837, using human whole blood in a flow chamber model in vitro. Haematologica 2008;93(S1):463 Abstract 1186.

e91

[20] Eriksson UG, Bredberg U, Hoffmann K-J, Thuresson A, Gabrielsson M, Ericsson H, et al. Absorption, distribution, metabolism, and excretion of ximelagatran, an oral direct thrombin inhibitor, in rats, dogs, and humans. Drug Metab Disp 2003;31:294–305. [21] Wagenvoord R, Deinum J, Elg M, Hemker HC. The paradoxical stimulation by a reversible thrombin inhibitor of thrombin generation in plasma measured with thrombinoscopy is caused by α2-macroglobulin-thrombin. J Thromb Haemost 2010;8:1281–9. [22] Wolzt M, Boström SL, Svensson M, Wåhlander K, Grind M, Sarich TC. Effects of the oral direct thrombin inhibitor ximelagatran on P-selectin expression and thrombin generation in atrial fibrillation. Pathophysiol Haemost Thromb 2003;33: 68–74. [23] Schützer KM, Svensson MK, Zetterstrand S, Eriksson UG, Wåhlander K. Reversible elevations of serum creatinine levels but no effect on glomerular filtration during treatment with the direct thrombin inhibitor AZD0837. Eur J Clin Pharmacol 2010;66:903–10.