Treatment with an oral direct thrombin inhibitor decreases platelet activity but increases markers of inflammation in patients with myocardial infarction


Christina Christersson MD, PhD, Department of Cardiology, Uppsala University, S-751 85 Uppsala, Sweden.
(fax: +46-18-50-66-38; e-mail:


Abstract.  Christersson C, Oldgren J, Wallentin L, Siegbahn A (Uppsala University, Uppsala, Sweden). Treatment with an oral direct thrombin inhibitor decreases platelet activity but increases markers of inflammation in patients with myocardial infarction. J Intern Med 2011; 270: 215–223.

Background.  Thrombin has a role not only in the coagulation process but also in inflammatory responses. Oral direct thrombin inhibitors (DTIs) are currently being evaluated in patients with thromboembolic diseases.

Objective.  To investigate whether an oral DTI affects markers for platelet and inflammatory activity after myocardial infarction (MI).

Methods.  A total of 518 patients with MI were randomly assigned to ximelagatran treatment (four different dose groups) in combination with aspirin, or aspirin alone for 6 months. The levels of soluble (s) P-selectin, soluble tissue factor, C-reactive protein (CRP), interleukin (IL)-10 and IL-18 were analysed in serial blood samples.

Results.  sP-selectin concentration increased after 1 week and persisted at an elevated level for 6 months in all study groups (P < 0.001). In the two highest ximelagatran dose groups, there was a reduced increase in sP-selectin compared to treatment with lower doses of ximelagatran and aspirin alone (P = 0.01 and P = 0.002, respectively). IL-18 levels did not change in the aspirin alone treatment group. By contrast, there was an elevation in IL-18 level in the lower and higher ximelagatran dose groups after 6 months (P = 0.006 and P < 0.001, respectively). Ximelagatran increased IL-10 levels (P = 0.002) and reduced the decrease in CRP levels after 6 months compared to treatment with aspirin alone (P = 0.002).

Conclusion.  A persistent elevation of platelet activity is found in patients with a recent MI after the cessation of acute antithrombotic treatment, and the addition of an oral DTI at higher doses decreases the activity. By contrast, long-term treatment with a DTI increases the levels of several markers of inflammation. Further studies with prolonged exposure of oral DTIs are needed for evaluation of the effect on inflammatory processes and to determine whether these agents influence clinical outcomes.


The progression of atherosclerosis because of the accumulation of monocytes/macrophages and lipids within the vessel wall generates an inflammatory state [1]. Upon disruption of a vulnerable plaque, there will be exposure of tissue factor (TF) release of inflammatory cytokines leading to the activation of platelets, formation of thrombin and progression of atherothrombosis [2]. An increase in the levels of cell-activating markers such as soluble TF (sTF) and soluble P-selectin (sP-selectin) has been found in patients with myocardial infarction (MI) [3, 4]. sTF originates from truncated TF from extravascular cells and monocytes, from TF-bearing microparticles derived from platelets and monocytes, and from alternatively spliced human TF [5]. The blood-borne TF maintains the procoagulant and inflammatory state in atherosclerosis [6]. P-selectin is expressed both in endothelial cells and in platelets. The source of sP-selectin is unclear. The results from several studies suggest that the marker predominantly reflects platelet activity [7–9]. However, sP-selectin has also been related to endothelial cell activity [8]. Levels of inflammatory markers such as C-reactive protein (CRP) and interleukin (IL)-10 are elevated in acute MI and predict the risk of future ischaemic events [10, 11]. IL-18 is an independent inflammatory cytokine, and concentrations in patients with previous MI are important for plaque vulnerability, although the long-term changes in the levels of this marker after the acute event have not been described [12, 13].

New oral antithrombotic drugs such as direct thrombin inhibitors (DTIs) are under development as additional treatment after MI and as alternative treatment to warfarin for venous thromboembolism and atrial fibrillation [14–16]. Thrombin has a central role in coagulation through cleavage of fibrinogen to fibrin, but it also participates in inflammatory processes both as a pro- and an anti-inflammatory agent [17–19]. Because of the pluripotent effects of thrombin, concerns have been raised regarding long-term treatment with DTIs, and an increased frequency of MI has been reported in some trials of DTIs [14, 20]. The first available DTI, ximelagatran, was evaluated for secondary prevention after MI in the Efficacy and Safety of the oral Thrombin inhibitor ximelagatran in combination with aspirin, in patiEnts with rEcent Myocardial damage (ESTEEM) trial [15]. Addition of ximelagatran to aspirin reduced the risk of new ischaemic events and decreased thrombin generation and fibrin turnover [21]. The objective of the present study was to investigate whether an oral DTI affects markers for cell activation and inflammation when given as long-term treatment to patients with recent MI.


Patient population and study design

The ESTEEM trial was a phase II study of the safety and efficacy of the oral DTI ximelagatran in patients with recent MI. Of 1883 patients, 518 were included in the biomarker substudy. High-risk patients with at least one risk factor for the recurrence of ischaemic events were randomly assigned to one of four doses of ximelagatran twice daily: 24 mg (n = 84), 36 mg (n = 79), 48 mg (n = 88) or 60 mg (n = 88), or placebo (n = 179) for 6 months. All patients received 160 mg aspirin per day, and oral ADP inhibitors were not permitted in the study protocol [15].

Blood sampling and laboratory methods

Venous blood was drawn, before intake of the morning dose of DTI or placebo, from the decubital vein with a 21-/22-gauge needle, at randomization, before the start of study treatment, after 1 week and after 6 months of treatment. After discarding the first portion of the sample, blood was collected into Vacutainer tubes (4.5 mL) containing citrate (3.8%) or ethylenediaminetetraacetic acid. The blood was centrifuged within 30 min at 2000 g for 20 min and stored at −70 °C until analysis.

At randomization, there were 491 and 487 blood samples available for analysis of sP-selectin and IL-18, respectively. In total, 39% of patients discontinued the study treatment prematurely. The main reason for discontinuation in the ESTEEM trial was the occurrence of cardiovascular events (prespecified in the study protocol), bleeding or raised concentrations of alanine transaminase [15]. Therefore, at 6 months, 309 and 303 samples were available for analysis of sP-selectin and IL-18, respectively. To further evaluate the long-term effect of DTIs on markers for cell activation and inflammation, we analysed sTF, high-sensitivity (hs) CRP and IL-10 in patients treated with the highest ximelagatran dose (60 mg) who had at least two available serial blood samples (n = 77). An equal number of randomly selected samples from patients in the placebo group were analysed (n = 77).

Laboratory methods

Levels of sP-selectin, IL-10 (R&D Systems, Minneapolis, MN, USA), IL-18 (MBL; Naka-ku Nagoya, Japan), sTF (American Diagnostics, Greenwich, CA, USA) and hsCRP (Architec, Abbott, IL, USA) were measured by immunoassay. All analyses were performed at the Laboratory for Coagulation Research, Department of Medical Sciences, Uppsala University, Sweden. The intra-assay precision was 5%, 6%, 10%, 4% and 8% for sP-selectin, IL-18, sTF, hsCRP and IL-10, respectively.

Statistical analysis

The proportion or mean values are given for demographic variables. The levels of markers for platelet activation and inflammation are given as medians (25th–75th percentiles). Analysis of markers for platelet and inflammatory activity was prespecified in the substudy protocol. The parameters were not normally distributed, and therefore, nonparametric tests were used (the Wilcoxon signed ranks test for within group comparisons and the Mann–Whitney U-test or the Kruskal–Wallis test for between group comparisons). Within each group, percentage changes from randomization were given as median values. Spearman’s rank correlation coefficients were used for the calculation of correlation between plasma concentration of markers for cell activation and inflammation. As the primary objectives included several markers and multiple tests, a value of P ≤ 0.01 was considered significant.


Baseline characteristics

Patients were included in the study 6 days after an MI. The mean age of study population was 71 years and 31% were woman. An ST-elevation MI was the index event in 57% of patients. A total of 339 patients were randomly assigned to one of four doses of ximelagatran administered twice a day, and 179 were assigned to placebo. Baseline characteristics were similar across treatment groups (Table 1). Between 76% and 89% of patients in the different study groups had been treated with heparin or low molecular-weight heparin until 26.2 and 25.9 h (mean values) before randomization in the placebo and the combined ximelagatran groups, respectively. Treatment at randomization with beta-blockers, agents acting on the renin–angiotensin system and statins was also similar across the treatment groups (Table 2). Patient characteristics and medical treatment in the randomly selected subset of the placebo group used in the analyses of sTF, hsCRP and IL-10 did not differ from those of the whole placebo group (Tables 1 and 2).

Table 1. Baseline characteristics
n = 179
Placebo random subgroup
n = 77
Ximelagatran combined groups
n = 339
Ximelagatran 24 mg
n = 84
Ximelagatran 36 mg
n = 79
Ximelagatran 48 mg
n = 88
Ximelagatran 60 mg
n = 88
  1. LVEF, left ventricular ejection fraction; MI, myocardial infarction.

Mean age (year)72727272717170
Man/woman (%)72/2871/2967/3367/3366/3465/3570/30
Diagnosis at randomization
 ST-elevation MI (%)54555849584760
 Non-ST-elevation MI (%)46454251425340
Additional risk factors
 Previous MI (%)25232327132724
 Angina pectoris (%)42364044314539
 Diabetes mellitus (%)12161510111822
 Peripheral occlusive arterial disease (%)68979109
 Hypertension (%)44404343404245
 Symptomatic heart failure or LVEF <0.4 (%)1214131491715
 Hyperlipidaemia (%)49485051465052
Table 2. Treatment at randomisation
n = 179
Placebo random subgroup
n = 77
Ximelagatran combined groups
n = 339
Ximelagatran 24 mg
n = 84
Ximelagatran 36 mg
n = 79
Ximelagatran 48 mg
n = 88
Ximelagatran 60 mg
n = 88
  1. ACE, angiotensin-converting enzyme; AT, angiotensin.

Beta-blocking agents (%)94989390898995
ACE inhibitors or AT II blocking agents (%)54525455385957
Statins (%)57526560586573


At randomization, the concentration of sP-selectin was 32 (25–41) ng mL−1 in the placebo group and 30 (23–40) ng mL−1 in the combined ximelagatran groups (P = 0.2). In all study groups, there was a persistent and significant elevation of sP-selectin levels (P < 0.001) during the study, which was most pronounced in patients receiving placebo (Fig. 1). Patients in the combined ximelagatran group had lower sP-selectin levels after 1 week than those who received placebo [34 (26–43) ng mL−1 vs. 36 (30–46) ng mL−1 (P = 0.01)]. Patients who received the two highest ximelagatran doses showed a significantly lower increase of 7% in sP-selectin levels after 1 week compared to 15% (P = 0.002) and 13% (P = 0.01) in those who received placebo and the combined two lower doses of ximelagatran, respectively. There was no significant difference in sP-selectin levels after 6 months compared to 1 week either in the placebo group (P = 0.3) or in the ximelagatran groups (P = 0.3–0.4).

Figure 1.

sP-selectin concentrations during the study period in the placebo, low-dose ximelagatran (24 + 36 mg) and high-dose ximelagatran (48 + 60 mg) groups. Box plots demonstrate the 25th and 75th percentiles and the median is shown with a black line. Whiskers denote the 10th and 90th percentiles. *** P < 0.001 and ** P ≤ 0.01, concentrations at 1 week versus at randomization within each group. ##P ≤ 0.01, ximelagatran 48 + 60 mg versus placebo groups.


The levels of IL-18 were 240 (176–342) and 246 (183–320) pg mL−1 in the placebo and the combined ximelagatran groups, respectively, at randomization (P = 0.8), with no difference between the ximelagatran groups. There was no significant change in the concentration of IL-18 in the placebo group after 1 week or after 6 months (Fig. 2). In the ximelagatran groups, no changes in the levels were found after 1 week. After 6 months of ximelagatran treatment, the IL-18 levels were increased by 6% in the two lower-dose groups and 10% in the two higher-dose groups, compared to the levels at 1 week (P = 0.006 and P < 0.001, respectively). There was a significant increase in IL-18 levels after 6 months in the ximelagatran-treated patients, compared to those who received placebo (P < 0.001).

Figure 2.

Concentrations of IL-18 during the study period in the placebo, low-dose ximelagatran (24 + 36 mg) and high-dose ximelagatran (48 + 60 mg) groups. Box plots demonstrate the 25th and 75th percentiles and the median is shown with a black line. Whiskers denote the 10th and 90th percentiles. †††P < 0.001, concentrations at 6 months versus 1 week within each group. ##P ≤ 0.01, ximelagatran versus placebo groups.


At randomization, the levels of sTF were 191 (162–254) and 205 (147–272) pg mL−1 in the placebo and 60-mg ximelagatran groups, respectively (P = 0.7). The concentration of sTF did not change within the placebo group during the study period (Table 3). Within the 60-mg ximelagatran group, the level of sTF was significantly reduced after 1 week (P = 0.008). After 6 months of ximelagatran treatment, the sTF level was increased by 11% compared to the level after 1 week (P < 0.001). There was a significant elevation of sTF concentration from 1 week to 6 months in the ximelagatran group compared to the placebo group (P = 0.002).

Table 3. The concentrations of sTF, CRP and IL-10 in the randomly selected placebo subgroup (n = 77) and ximelagatran 60 mg groups (n = 77)
 Placebo random subgroupXimelagatran 60 mg
  1. Results presented as medians (25th–75th percentiles). ***P < 0.001 and **P ≤ 0.01, concentrations at 1 week versus at randomization within the different groups. †††P < 0.001 and ††P ≤ 0.01, concentrations at 6 months versus 1 week within the different groups.

  2. CRP, C-reactive protein; sTF, soluble tissue factor.

Soluble TF pg mL−1
 Randomization191 (162–254)205 (147–272)
 1 week186 (144–230)188 (150–269)**
 6 months189 (150–232)199 (177–265)†††
CRP μg mL−1
 Randomization10.9 (4.5–29.8)10.1 (4.4–23.5)
 1 week3.4 (1.8–9.4)***4.2 (2.1–9.2)***
 6 months1.4 (1.0–3.2)†††2.8 (1.4–4.9)††
IL-10 pg mL−1
 Randomization1.7 (1.0–3.2)1.3 (0.8–2.2)
 1 week1.0 (0.5–2.5)***1.2 (0.6–1.9)
 6 months1.0 (0.5–2.1)1.4 (0.6–2.4)††


The levels of CRP were 10.9 (4.5–29.8) μg mL−1 in the placebo group and 10.1 (4.4–23.5) μg mL−1 in the 60-mg ximelagatran group at randomization (P = 0.8). CRP was markedly decreased by 65% after 1 week with a further decrease of 53% between 1 week and 6 months in the placebo group (P < 0.001). Similar results were demonstrated in the 60-mg ximelagatran group with a 59% reduction after 1 week of study treatment (P < 0.001). In the 60-mg ximelagatran group, CRP levels decreased by 24% after 6 months compared to the levels at 1 week (P < 0.001). There were no significant differences in CRP levels or in the change in levels from the time of randomization when comparing the 60-mg ximelagatran and the placebo groups after 1 week. However, after 6 months, the CRP concentrations were higher in the ximelagatran group compared to the placebo group (P = 0.002) (Table 3).


At randomization, the concentrations of IL-10 were 1.7 (1.0–3.2) and 1.3 (0.8–2.2) pg mL−1 in the placebo and 60-mg ximelagatran groups, respectively (P = 0.07). In the placebo group, the level of IL-10 was reduced by 18% after 1 week compared to at randomization (P < 0.001) and remained at a similar level at 6 months (P = 0.5). Within the 60-mg ximelagatran group, there was no change in the levels of IL-10 after 1 week of study treatment (P = 0.5). However, after 6 months, the concentration of IL-10 had increased by 22% compared to the levels at 1 week (P = 0.002) (Table 3). The change in IL-10 level in the ximelagatran group after 6 months of treatment was significantly higher compared to the change in the placebo group (P = 0.005).

Associations between markers for platelet activation, coagulation and inflammation

After 6 months of study treatment, there was a weak but significant correlation between sP-selectin and sTF (r = 0.22, P = 0.01). There was no correlation between sP-selectin and the other markers of inflammation. IL-18 was associated with IL-10 levels (r = 0.26, P = 0.006) after 6 months. There was also a relation between the concentration of sTF and the inflammatory marker IL-10 after 6 months (r = 0.23, P = 0.01). Within the 60-mg ximelagatran group, the correlation between sTF and IL-10 was even stronger (r = 0.53, P < 0.001). The analysis of prothrombin fragment 1 + 2 (F1 + 2) and d-dimer as markers for thrombin generation and fibrin turnover in the ESTEEM trial has previously been reported; a persistent reduction of these markers in the ximelagatran groups was found after 6 months [21]. There was no significant correlation between F1 + 2 or d-dimer and the concentrations of sTF in the ximelagatran group after 6 months.


P-selectin, a marker for cell activation, is expressed both by platelets and endothelial cells. The soluble form of P-selectin is related to other markers for platelet activity, and studies indicate that the majority of sP-selectin is shed from platelets [7–9]. sP-selectin levels are elevated in unstable angina [3]. The first evaluation of sP-selectin in the present study reflected the platelet activity in patients with recent MI treated with the standard acute antithrombotic treatment. However, after the cessation of these standard drugs, there was an increase in the initial levels of sP-selectin in all study groups, and the elevation of these levels persisted for up to 6 months after MI. In the ximelagatran groups, the two higher doses showed an effect on sP-selectin levels with an attenuated increase compared to the placebo group, and this effect was stable for up to 6 months. In experimental models, DTIs have been found to reduce platelet P-selectin expression upon thrombin stimulation [22–24]. Compared to warfarin, DTIs also decreased platelet P-selectin expression in patients after elective percutaneous coronary intervention and in patients with atrial fibrillation [25, 26]. The effect on platelet activity seems to be related to the dose of the oral DTI, in contrast to the dose-independent effect on thrombin generation and fibrin turnover after MI [21].

Increased IL-18 plasma levels found in patients with coronary artery disease predict death and are correlated with the severity of coronary atherosclerosis [12, 27]. In the present study, we found that the concentrations of IL-18 after recent MI remained unchanged for up to 6 months in the group treated with aspirin combined with standard secondary prevention. Thrombin is involved in inflammatory processes by forming complexes with thrombomodulin and activated protein C (APC). The complexes reduce leucocyte adhesion, monocyte infiltration of endothelium and the release of cytokines through binding to the endothelial cell protein C receptor and to the protease-activated receptor 1 [18]. Thrombin inhibitors reduced APC in endothelial cells in vitro, and hirudin increased leucocyte–endothelial cell interactions in a sepsis microcirculation model [28, 29]. Both endothelial cells and macrophages produce IL-18 upon activation, and therefore, the increased IL-18 concentration found in the ximelagatran-treated groups may be explained by a mechanism that involves a decrease in the amount of available thrombin and reduced APC levels [30, 31]. Whether DTIs can directly interact with endothelial cells or leucocytes and thereby induce production of IL-18 is unknown at present, but this could also contribute to the ximelagatran-induced increase in IL-18 concentration.

Similarly, an increase in IL-10 levels after 6 months of ximelagatran treatment was observed in the present study. IL-10 is an anti-inflammatory cytokine that is important in the regulation of atherosclerotic lesions [32, 33]. IL-10 reduced IL-18-induced matrix metalloproteinase-9 levels in mononuclear cells, and the IL-18/IL-10 ratio has been related to poor outcomes in acute MI [34, 35]. The increase in IL-18 concentration may induce production of IL-10 to keep the balance between pro- and anti-inflammatory processes. Conflicting data have been reported with regard to high concentrations of IL-10 and the risk of new MI and death in patients with acute coronary syndrome [11, 36]. In the present study, there was a reduction in IL-10 levels in the placebo group after 1 week, which might indicate that IL-10 reflects the acute inflammatory response during MI.

In accordance with previous studies, we found a reduction of CRP for up to 6 months after MI [37, 38]. However, ximelagatran treatment reduced the decrease in CRP level compared with treatment with aspirin alone. The correlations between the inflammatory markers indicate an increased inflammatory turnover in the ximelagatran-treated patients. The effects of the DTI on the inflammatory response seem to progress more slowly than the previously described effects on thrombin generation and fibrin turnover [21]. It is important to evaluate the long-term inflammatory response initiated by DTIs in patients and to determine whether the changes in levels of markers are related to outcomes. In mouse models, IL-18 has been found to enhance atherosclerosis [39]. A small but significant increase in frequency of MI has been found in some clinical trials with oral DTIs, both with ximelagatran and the new compound dabigatran [14, 20, 40]. Recently, when all available trials comparing warfarin and ‘anticoagulant equivalents’ in atrial fibrillation were analysed, warfarin appeared to be associated with a lower frequency of MI [41].

The mechanisms underlying these results and whether the DTI-induced increase in inflammatory activity is related to the risk of MI need to be further investigated. In addition, whether the increase in levels of inflammatory markers is related only to ximelagatran or whether it is a class effect of oral DTIs should be evaluated.

The cell–cell interaction between activated platelets and monocytes induces the production of TF and inflammatory cytokines [42–44]. DTIs reduced TF expression in platelet–leucocyte aggregates in vitro upon thrombin stimulation, and the reduced levels of sTF found after 1 week of ximelagatran treatment in the present study are in accordance with these results [22, 45]. However, after 6 months of treatment, there was an increase in sTF levels that correlated with the levels of inflammatory markers. The progression of leucocyte–endothelial cell interaction induced by DTIs may generate an increase in cell activity and production of circulating TF-presenting microparticles, which could contribute to the results of the present study [29, 46]. DTIs do not appear to modulate the ADP-induced formation of platelet–leucocyte aggregates and TF production, and as the use of ADP inhibitors was prohibited in the ESTEEM trial, this could also influence the present results [22]. There is ongoing debate about the source of sTF and its contribution to coagulation [47]. The coagulation activity of sTF was not evaluated in this trial, but we found no correlation between sTF and markers of thrombin generation and fibrin turnover. Therefore, sTF seems to be a marker for cell activity.

Activated endothelial cells express P-selectin and other adhesion molecules, and a correlation between sP-selectin and soluble E-selectin levels has been demonstrated in some studies [8]. The DTI-induced increase in levels of inflammatory markers and sTF observed in the present study indicates an activation of both leucocytes and endothelial cells. DTIs did not increase the concentration of sP-selectin, and sP-selectin concentration was not correlated with that of the inflammatory markers. These findings may indicate that sP-selectin concentration better reflects platelet activation than endothelial cell activation.

Limitations of the present study include the small number of samples used to analyse TF, CRP and IL-10. In addition, IL-18-binding protein, which regulates free IL-18, has not been evaluated [48]. The duration of the study treatment (6 months) might be too short for evaluation of the effect of oral DTIs on inflammation. These subgroup analyses are considered to be exploratory and hypothesis generating only. Ximelagatran demonstrated ‘proof of concept’ in the treatment of thromboembolic diseases, but was withdrawn from the market because of risk of hepatotoxicity. However, new oral DTIs are in development and are being evaluated in clinical trials [14, 16, 49]. Thus, the results of the effect of ximelagatran on markers of platelet activity and inflammation will form a firm basis for future studies.

We conclude that the oral DTI ximelagatran dose-dependently decreased platelet activity early after the acute MI event. By contrast, this DTI increased markers of inflammation over a longer duration. The mechanism of the inflammatory response to prolonged oral treatment with DTIs is the subject of our ongoing studies.

Conflict of interest statement

L Wallentin and A Siegbahn have received research grants from AstraZeneca.


We are grateful to the investigators and centres in Sweden, Norway and Denmark for the inclusion of patients and the collection of blood samples, and Birgitta Fahlström for excellent technical assistance. AstraZeneca was the sponsor of the ESTEEM trial and supported the present study with a research grant. Financial support was provided by the Swedish Heart-Lung Foundation, the Swedish Research Council and the Uppsala County Association Against Heart and Lung Diseases.