OUP user menu

Effect of vorapaxar on myocardial infarction in the thrombin receptor antagonist for clinical event reduction in acute coronary syndrome (TRA·CER) trial

Sergio Leonardi, Pierluigi Tricoci, Harvey D. White, Paul W. Armstrong, Zhen Huang, Lars Wallentin, Philip E. Aylward, David J. Moliterno, Frans Van de Werf, Edmond Chen, Luis Providencia, Jan E. Nordrehaug, Claes Held, John Strony, Tyrus L. Rorick, Robert A. Harrington, Kenneth W. Mahaffey
DOI: http://dx.doi.org/10.1093/eurheartj/eht104 1723-1731 First published online: 25 March 2013


Aims The TRA·CER trial compared vorapaxar, a novel platelet protease-activated receptor (PAR)-1 antagonist, with placebo in 12 944 patients with high-risk non–ST-segment elevation acute coronary syndromes (NSTE ACS). In this analysis, we explored the effect of vorapaxar on myocardial infarction (MI).

Methods and results A blinded, independent central endpoint adjudication committee prospectively defined and classified MI according to the universal MI definition, including peak cardiac marker value (creatine kinase-MB [CK-MB] and/or troponin). Because the trial failed to meet its primary endpoint, these analyses are considered exploratory. During a median follow-up of 502 days, 1580 MIs occurred in 1319 patients. The majority (n = 1025, 64.9%) were type 1 (spontaneous) MI, followed by type 4a [percutaneous coronary intervention (PCI)-related] MI (n = 352; 22.3%). Compared with placebo, vorapaxar reduced the hazard of a first MI of any type by 12% [hazard ratio (HR), 0.88; 95% confidence interval (CI), 0.79–0.98; P = 0.021] and the hazard of total number of MIs (first and subsequent) by 14% (HR, 0.86; 95% CI, 0.77–0.97; P = 0.014), an effect that was sustained over time. Vorapaxar reduced type 1 MI by 17% (HR, 0.83; 95% CI, 0.73–0.95; P = 0.007). Type 4a MIs were not significantly reduced by vorapaxar (HR, 0.90; 95% CI, 0.73–1.12; P = 0.35). Vorapaxar effect was consistent across MI sizes defined by peak cardiac marker elevations and across key clinical subgroups; however, in patients not treated with thienopyridine at baseline (HR, 0.65; 95% CI, 0.46–0.92) compared with patients who received thienopyridine (HR, 0.91; 95% CI, 0.81–1.02), there was a trend towards a higher effect (Pint = 0.077).

Conclusion The PAR-1 antagonist vorapaxar was associated with a reduction of MI, including total number of infarctions. This reduction was sustained over time and was mostly evident in type 1 MI, the most common type of MI observed.

  • Myocardial infarction
  • Trials
  • Thrombosis

See page 1699 for the editorial comment on this article (doi:10.1093/eurheartj/eht127)


Vorapaxar is a novel, oral antiplatelet agent that potently inhibits thrombin-induced platelet aggregation via a selective antagonism of the protease-activated receptor (PAR)-1, and thus produces a potent and consistent inhibition of thrombin-receptor agonist peptide (TRAP)-induced platelet aggregation.1 Because thrombin plays a pivotal role in platelet aggregation, it has been hypothesized that vorapaxar could reduce thrombotic cardiovascular events, including myocardial infarction (MI). In a large phase III program, including the Thrombin Receptor Antagonist for Clinical Event Reduction in Acute Coronary Syndrome trial [TRA·CER—a study of vorapaxar vs. placebo in nearly 13 000 patients with non–ST-segment elevation acute coronary syndromes (NSTE ACS) that was terminated prematurely, after enrolment completion and the target endpoint was reached, following an unplanned independent safety review] and the Thrombin Receptor Antagonist in Secondary Prevention of Atherothrombotic Ischemic Events–Thrombolysis in Myocardial Infarction 50 trial (TRA 2P°-TIMI 50—a study of vorapaxar vs. placebo in more than 26 000 patients with documented atherosclerotic disease), vorapaxar was associated with a reduction of the composite endpoint of cardiovascular death, MI, or stroke.2,3 Although this composite was the primary endpoint in the TRA 2P°-TIMI 50 trial, in TRA·CER, this was the key secondary endpoint, with a non-significant primary endpoint. In both trials, a reduction of MI was the key driver of the composite outcome reduction.

Because MI represents a heterogeneous group of events with different pathophysiologies, sizes, and clinical and prognostic implications, it is important to characterize a drug's effect on the type of MI to better understand the potential for clinical benefit and to define a correlation between putative biology mechanisms and observed clinical phenotypes.

For these reasons, we analysed the effect of vorapaxar on different types of MI as currently classified by the Universal Definition of Myocardial Infarction document, analysed the effect of vorapaxar on total occurrences of MI, and explored the effect of vorapaxar across MIs of various sizes and over time.4


Study population

The TRA·CER trial's primary results and design have been previously reported.2,5 In summary, this was a multicentre, randomized, double-blinded, event-driven trial with a minimum follow-up duration of 1 year that compared placebo with vorapaxar administered as a 40 mg loading dose followed by a daily 2.5 mg maintenance dose in 12 944 patients with NSTE ACS at high risk for ischaemic events enrolled at more than 800 centres in 37 countries. The protocol recommended for the loading dose to be administered at least 1 h before any coronary revascularization. Previous vorapaxar platelet function studies revealed that 67.9% of patients reached adequate (>80%) ex vivo inhibition of TRAP-induced platelet aggregation 1 h following a 40 mg loading dose, and this proportion increased to 96.3% at 2 h.6 A maintenance dose of 2.5 mg daily sustained this effect during the 60-day treatment period.6

The TRA·CER trial complied with the Declaration of Helsinki and was approved by the appropriate national and institutional regulatory authorities and Ethics Committees. All patients provided written informed consent.

Endpoint definition

Myocardial infarction and other endpoints included in the primary outcome measure were prospectively adjudicated by a clinical events committee (CEC) blinded to treatment allocation. All suspected endpoints (death, MI, recurrent ischaemia with rehospitalization, urgent coronary revascularization, stent thrombosis, and bleeding) were systematically identified via an integrated assessment of investigator-reported events along with capture of electronic case report form (eCRF) data suggestive of recurrent events, including cardiac biomarker trends. During the adjudication process, every confirmed MI event was classified according to the universal MI definition (Table 1). The definition of MI used in TRA·CER (Supplementary material online, Appendix S1) adhered to the criteria proposed by the universal MI definition working group with some adaptation, particularly the use of percutaneous coronary intervention (PCI)-related MI (type 4a) by using creatine kinase-MB (CK-MB) as a preferred diagnostic biomarker.5 Additional characterization of MI was performed by the CEC with respect to electrocardiogram (ECG) (presence/absence of persistent ST-segment elevation; new Q-waves) and peak biomarker values (troponin and CK-MB). In TRA·CER, the three coronary ischaemic endpoints—MI, recurrent ischaemia with rehospitalization, and urgent coronary revascularization—were defined hierarchically and were mutually exclusive.5 In the presence of ischaemic myocardial necrosis, only an MI was assigned, regardless of a concurrent coronary revascularization or urgent hospitalization. Stent thrombosis was adjudicated using the Academic Research Consortium Classification,7 based on catheterization laboratory reports and additional source documents but without independent review of angiograms.

View this table:
Table 1

Universal MI definitiona

Statistical analyses

The outcome of interest for this analysis was MI, which was assessed from randomization to the last study visit. The primary analyses were performed based on the first occurrence of an MI. The effect of vorapaxar on MI itself and different types of MI was estimated by time-to-event analysis. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated using Cox proportional hazards regression modelling with adjustment of randomization stratification factors (planned use of glycoprotein IIb/IIIa inhibitor and planned use of direct thrombin inhibitor), as specified in the protocol. Cumulative incidence of MI was estimated considering the competing risk of death. We also performed an analysis of vorapaxar effect on total occurrences of MI, which included all MIs (first and subsequent) observed during follow-up, using the Andersen–Gill approach with robust standard error estimates. As per the study design, because the primary endpoint was not statistically significant, all secondary analyses on efficacy endpoints including MI are considered exploratory. To evaluate vorapaxar's effect on type 1 and type 4a MIs of increasing sizes, the first event of each type was classified by peak CK-MB according to the following categories: >3, >5, and >10 times the local upper limit of normal (ULN). We also described the rates of type 4a MI according to the time of administration of loading dose before PCI (>1 or ≤1 h; >2 or ≤2 h). Finally, to characterize the type of coronary lesion found on angiography in patients with MI, we presented investigator-reported angiographic data on treated coronary lesions in patients who experienced a non-procedural MI (i.e. excluding type 4a and type 5 MIs) after randomization and who underwent a coronary angiography with PCI within 7 days of the event. All hypothesis tests were two-sided, and a P-value of 0.05 was considered statistically significant. P-values were not adjusted for multiplicity of comparisons. All analyses were performed using SAS 9.1 (SAS Institute, Cary, NC, USA).


Key time intervals and concomitant treatments relevant for this analysis are listed in Table 2. During the index hospitalization, most patients underwent coronary angiography and guidelines-recommended treatments for NSTE ACS.

View this table:
Table 2

Revascularization strategy and medications during index hospitalization

After randomization, during a median follow-up of 502 days [interquartile range (IQR), 349–667], 1319 patients experienced an MI, with 1133 (85.9%) patients having one MI and 186 (14.1%) patients having two or more events. Overall, 1580 MIs occurred, including recurrent events. Type 1 (spontaneous) MIs were the most common MI type (n = 1025; 64.9%), followed by type 4a (PCI-related) MIs (n = 352; 22.3%). Other types of MIs were less common (Table 3).

View this table:
Table 3

Total number of MI events

During the index hospitalization, at least one value of troponin or CK-MB within 24 h of a coronary revascularization procedure was obtained in 93.4% of patients undergoing PCI (median, 5; IQR, 4–6) and in 70.2% of patients undergoing coronary artery bypass grafting (CABG) (median, 4; IQR, 0–4).

Effects of vorapaxar

Compared with placebo, vorapaxar reduced the hazard of a first MI of any type by 12% (HR, 0.88; 95% CI, 0.79–0.98; P = 0.021) (Figure 1 and Table 4). The effect of vorapaxar was similar when the endpoint included all MIs, including recurrent MIs after the first event (HR 0.86; 95% CI, 0.77–0.97; P = 0.014) (Supplementary material online, Appendix S2). The cumulative incidence of MI in patients with vorapaxar and placebo is presented in Figure 2, which shows that the two curves started to diverge ∼1 month after randomization and continued to separate over time. A type 1 (spontaneous) MI occurred in 382 of 6473 patients in the vorapaxar group (5.9%) and in 455 of 6471 patients of the placebo group (7.0%) (HR, 0.83; 95% CI, 0.73–0.95; P = 0.007). The effect of vorapaxar was directionally consistent across MIs of different sizes as assessed by troponin or CK-MB peak data indexed by the local MI decision limit (Figure 3A and B). Similar consistency of vorapaxar effect was observed using higher biomarker thresholds for type 1 MI (Figure 4, top panel).

View this table:
Table 4

Vorapaxar effect by universal MI type and by ECG criteria over time

Figure 1

Incidence of any first MI with vorapaxar or placebo. Cumulative incidence with 95% confidence intervals of a first MI of any universal MI type between vorapaxar and placebo. CI, confidence interval; HR, hazard ratio; MI, myocardial infarction.

Figure 2

Incidence of a spontaneous (type 1) MI with vorapaxar or placebo. Cumulative incidence with 95% confidence intervals of a first type 1 MI between vorapaxar and placebo. CI, confidence interval; HR, hazard ratio; MI, myocardial infarction.

Figure 3

(A) Incidence of MI categorized by degree of maximal troponin elevation with vorapaxar or placebo. On the Y-axis, the % of MI events among patients with troponin data available (n = 1244). MI, myocardial infarction; ULN, upper limit of normal. (B) Incidence of MI categorized by degree of maximal CK-MB elevation with vorapaxar or placebo. On the Y-axis, the % of MI events among patients with CK-MB data available (n = 989). CK-MB, creatine kinase-MB; MI, myocardial infarction; ULN, upper limit of normal.

Figure 4

Spontaneous and peri-PCI MI with vorapaxar or placebo by degree of CK-MB elevation. HR plot of a first occurrence of type 1 (upper panel) or type 4a (lower panel) MI by degree of CK-MB elevation. CI, confidence interval; CK-MB, creatine kinase-MB; HR, hazard ratio; MI, myocardial infarction; PCI, percutaneous coronary intervention; ULN, upper limit of normal.

Type 4a MI occurred in 166 (2.6%) patients in the vorapaxar group and in 183 (2.8%) patients in the placebo group (HR, 0.90; 95% CI, 0.73–1.12; P = 0.350) (Table 4). By selecting larger type 4a MIs, lower HRs for vorapaxar were observed but with broader CIs (Figure 4, bottom panel).

Of the patients who were treated with PCI (n = 7479), the majority received a loading dose as mandated by the protocol, with only 339 patients (4.5%) receiving the loading dose within 1 h or after PCI (non-adherent to the protocol). In total, 2208 patients (29.5%) received the loading dose within 2 h of PCI. No signal was observed suggesting a different effect of vorapaxar on type 4a MI according to the time from loading dose to PCI (Supplementary material online, Appendix S3). The rates of MI associated with stent thrombosis (type 4b) were similar between the two treatment groups (HR, 1.02; 95% CI, 0.66–1.57). Likewise, the occurrence of type 2 MI was similar between the vorapaxar and placebo groups (HR, 1.17; 95% CI, 0.74–1.8). With the adoption of the universal MI definition, events meeting the criteria for type 5 MI (CABG-related) were infrequent and the incidence was not significantly different between the vorapaxar and placebo groups.

The analysis on ST-segment elevation MIs and MIs associated with new Q-waves indicated that the effect of vorapaxar was directionally consistent with the overall effect on MI (Table 4).

Vorapaxar effect on MI was consistent across key subgroups (Supplementary material online, Appendix S4), and no interaction test was statistically significant. The only observed non-significant trend was on thienopyridine use at baseline, where a more pronounced efficacy with vorapaxar was observed in patients who were not treated with these drugs at randomization. In the subgroup of patients with non-procedural MI who underwent a coronary angiography with PCI within 7 days and had angiographic data available (n = 438), the treated culprit coronary lesion was mainly related to de novo lesion (n = 267; 61%), while restenosis (n = 119; 27.1%) and stent thrombosis (n = 89; 20.3%) were less common (Table 5).

View this table:
Table 5

Non-procedural MIs rates and associated type of coronary lesion with vorapaxar and placebo


In this study of patients with high-risk NSTE ACS who were mostly treated with concomitant dual antiplatelet therapy, vorapaxar was associated with a reduction of MI, most evident for spontaneous MI. This effect seemed directionally consistent across various MI sizes and was sustained over time. We also observed a reduction in the total occurrences of MIs in patients treated with vorapaxar.

In the TRA·CER trial, vorapaxar was associated with a non-significant reduction of the primary endpoint and a significant reduction of the key secondary composite endpoint of cardiovascular death, MI, or stroke.2 A reduction in MI was the main effect of vorapaxar observed in the composite endpoints. In composite endpoints of randomized clinical trials, MIs are typically computed as one type of event. However, from clinical and mechanistic perspectives, MIs are a heterogeneous group of events, and this is highlighted by the universal MI definition classification types.4,8 Infarct size is also important to define the clinical implications of an MI and potential benefits of such compounds. Because there are different underlying mechanisms leading to the various types of MI and their clinical consequences, it is important to determine the effects of drugs on each subtype, as a drug may not similarly affect all mechanisms.

Spontaneous MIs are arguably the most important type because they are the most commonly observed in clinical practice and are important predictors of future cardiovascular morbidity and mortality.911 In TRA·CER, 65% of observed MIs were spontaneous, and they were the most frequently observed coronary ischaemic endpoint. The risk of death after a spontaneous MI is estimated to increase by more than five- to seven-fold.911 In comparison, the association between peri-PCI MI and mortality is less robust.911 Spontaneous MIs are caused by plaque erosion or rupture with superimposed thrombosis. Plaque erosions or ruptures, while creating a thrombogenic milieu, are often asymptomatic because they do not lead to the formation of an ischaemic, flow-limiting coronary thrombus.12 Factors leading to thrombosis and acute coronary events that are clinically apparent are not well-understood, but modulation of platelet pathways is likely important.13 Therefore, it is critical to identify strategies that could reduce clinical thrombosis on spontaneous plaque rupture. Our study indicates a 17% reduction of spontaneous MI with vorapaxar, in addition to aspirin and thienopyridine, suggesting that (i) PAR-1 activation is important in the genesis of clinically evident coronary thrombosis in the presence of plaque disruption and (ii) PAR-1 antagonism is a viable mechanism to reduce symptomatic thrombosis when such plaque disruption occurs. These data also show that the effect on spontaneous MI was sustained over time, suggesting continuous antagonism of PAR-1 is potentially important long-term. In support of an effect persisting over time, we have also observed that use of vorapaxar was associated with a reduction of total occurrences of MI.

Although our results further define the clinical efficacy of vorapaxar, it must be acknowledged that MI was a secondary endpoint and the TRA·CER trial failed to meet its primary objective. Therefore, our results are to be considered exploratory and cannot yet be applied to clinical practice to treat NSTE ACS patients; our results need to be further validated in future trials. In secondary analysis, particularly if the primary endpoint was not met, there is the risk of a chance finding due to multiple testing. However, our findings are validated by the TRA 2P°-TIMI 50 trial, in which a reduction of MI with vorapaxar was observed. In the TRA 2P°-TIMI 50 trial, a study of more than 26 000 patients with stable atherothrombosis (i.e. recent MI, history of stroke, or peripheral artery disease), vorapaxar was associated with a significant 13% reduction of cardiovascular death, MI, or stroke.3 This effect was also driven by lower MI rates observed in the vorapaxar arm. Notably, in the large subgroup of patients with a history of an MI (n = 17 779), vorapaxar reduced MI by 21% compared with placebo (P = 0.0003).14 As in TRA·CER, the differences were primarily driven by reduced spontaneous MI rates with vorapaxar. The consistency in MI reduction from these two trials, totalling almost 40 000 patients, supports that PAR-1 antagonism with vorapaxar is a viable mechanism to reduce MI rates in patients with coronary disease.2,3

In the analysis of vorapaxar effect by infarct size, defined as different degrees of troponin or CK-MB elevation or presence of new Q-waves or ST-segment elevation, the effect of vorapaxar by infarct size appears to be consistent. Nonetheless, because of the small number of events and wide CIs, cautious interpretation is warranted.

The efficacy signal on MI observed in TRA·CER must be put in the context of the overall results. In TRA·CER, despite the continued effect of MI reduction over time, the effect on composite outcomes does not appear to be incremental long-term and might have been counterbalanced by the continuous accrual of bleeding.2 A significant increase in major bleeding with vorapaxar was reported in TRA·CER. Overall, the efficacy and safety balance observed in TRA·CER was not favourable. Clinicians considering vorapaxar use in clinical practice or future studies will have to carefully weigh anti-ischaemic efficacy against risk of bleeding, which included uncommon but clinically important events like intracranial haemorrhage. Thus, although our exploratory results support the potential for clinical efficacy of vorapaxar and the opportunity for future confirmatory studies in ACS patients, future studies of vorapaxar must include strategies to reduce bleeding, which may include careful patient selection, dosing strategies, and combination therapies.

The effect of vorapaxar on type 4a (peri-PCI) MI was lower than observed in phase II clinical trials.6,15 These events are prognostically less important than spontaneous MI, yet reduction of peri-procedural myocardial necrosis is a goal of antiplatelet therapy.10,11 Peri-PCI myocardial necrosis may be caused by various mechanisms, including distal embolization, side-branch occlusion, coronary dissection, and reduced blood flow. It is possible that pathways different than TRAP-induced platelet aggregation, such as collagen-mediated platelet aggregation, are pivotal to the endothelial damage caused by balloon inflation or stenting. Other acute antiplatelet therapies, such as thienopyridines or glycoprotein IIb/IIIa inhibitors, were commonly used in TRA·CER patients. In addition, limitations exist in detecting peri-PCI MI in patients presenting with an acute, spontaneous MI. In TRA·CER, more than 90% had an MI as the qualifying event; therefore, challenges exist in detecting a new biomarker raise post-PCI during the evolution of a spontaneous MI. This may have reduced the ability to detect a drug effect. The CEC rigorously adhered to the universal MI definition for peri-PCI MI, which required normal or decreasing cardiac biomarkers before PCI for a post-PCI MI to be adjudicated. With these limitations, in addition to ongoing discussions on the appropriate definition of peri-PCI MI,16 it remains unclear whether there is a potential meaningful effect of vorapaxar in preventing peri-PCI MI.

Although we did not observe an effect on MIs related to stent thrombosis with vorapaxar, these data were based on site reports of stent thrombosis in clinical documentation that typically do not apply standard criteria for diagnosing stent thrombosis. Angiograms are being collected for central core laboratory adjudication of stent thrombosis in a subset of TRA·CER patients and may be informative. Moreover, because of the frequent use of dual antiplatelet therapy, the TRA·CER results do not allow conclusions to be drawn on how vorapaxar alone, or in addition to aspirin, affects stent thrombosis and how this compares with current standards of care.1719


In TRA·CER we observed that, among patients with NSTE ACS, vorapaxar compared with placebo was associated with a reduction of MI. This effect was mostly evident in type 1 MI, was sustained over time, and was also apparent in total occurrences of MI. Given the exploratory nature of our findings (i.e. secondary analysis in presence of non-significant primary endpoint), caution must be applied to the interpretation of these results. However, similar evidence from the TRA 2P°-TIMI 50 trial supports the validity of our findings. Overall, these results support the potential role of PAR-1 antagonism to reduce events mediated by coronary thrombosis following ACS and the opportunity for additional confirmatory studies, which will require careful consideration of strategies to minimize bleeding risk.


The TRA·CER trial was funded by Merck & Co., Inc.

Conflict of interest: S.L.: none. P.T.: consultancy for and grants from Merck & Co, Inc. H.D.W.: research support from Sanofi-Aventis, Eli Lilly, The Medicines Company, the National Institutes of Health, Pfizer, Roche, Johnson & Johnson, Merck Sharpe & Dohme, AstraZeneca, GlaxoSmithKline, Daiichi Sankyo Pharma Development, and Bristol-Myers Squibb; and advisory boards for Merck Sharpe & Dohme and Regado Biosciences.  

P.W.A.: research support and consulting fees/honoraria from Merck Sharp & Dohme Corp, consultancy for Eli Lilly, Regado Biosciences, F. Hoffmann-La Roche Ltd, GlaxoSmithKline, Sanofi-Aventis, Takeda Pharmaceuticals, and Merck & Co., Inc.; and research support from Boehringer Ingelheim, Sanofi-Aventis Canada, GlaxoSmithKline, AstraZeneca, Regado Biosciences, Amylin, and Novartis. Z.H.: research support from Merck & Co., Inc. L.W.: research support and consulting fees/honoraria from Merck & Co., Inc.; consultancy for Regado Biotechnologies, Protola, C.S.L. Behring, Athera Biotechnologies, Boehringer Ingelheim, AstraZeneca, GlaxoSmithKline, Bristol-Myers Squibb, Pfizer; and research support and lecture fees from AstraZeneca, Boehringer Ingelheim, Bristol-Myers Squibb, Pfizer, GlaxoSmithKline, and Schering-Plough.

P.E.A.: research support from Merck & Co., Inc., AstraZeneca, Eli Lilly, and Bayer/Johnson & Johnson; consultancy for AstraZeneca, Boehringer Ingelheim, Pfizer, Sanofi-Aventis, and Eli Lilly; and lecture fees from AstraZeneca, Boehringer Ingelheim, and Eli Lilly.

D.J.M.: research support and consulting fees/honoraria from Merck & Co., Inc. F.V.d.W.: research support from Merck & Co., Inc. and Boehringer Ingelheim; consultancy for AstraZeneca and Boehringer Ingelheim; and lecture fees from Boehringer Ingelheim and AstraZeneca. E.C.: employee of Merck & Co., Inc. L.P. and J.E.N.: none. C.H.: research support from Merck & Co., Inc., GlaxoSmithKline, AstraZeneca, and Bristol-Myers Squibb; advisory board for Pfizer and AstraZeneca; and lecture fees from AstraZeneca. J.S.: employee of Merck & Co., Inc. T.L.R.: none. R.A.H.: research support from Merck & Co., Inc., Novartis, Merck, Portola, Sanofi-Aventis, The Medicines Company, Bristol-Myers Squibb, and AstraZeneca; advisory board for Novartis, Portola, Merck & Co., Inc., Johnson & Johnson, Pfizer, and Regado; consultancy for Merck & Co., Inc., Novartis, Sanofi-Aventis, AstraZeneca, Eli Lilly, and Bristol-Myers Squibb; and lecture fees and payment for manuscript preparation from AstraZeneca.

K.W.M.: research support from Merck & Co., Inc., AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Eli Lilly, Johnson & Johnson, Novartis, Pozen, Regado, Sanofi-Aventis, Schering-Plough, The Medicines Company, Daiichi Sankyo, and GlaxoSmithKline; and consultancy for AstraZeneca, Bayer, Bristol-Myers Squibb, Boehringer Ingelheim, Daiichi Sankyo, Eli Lilly, GlaxoSmithKline, Johnson & Johnson, Merck & Co., Inc., Novartis, Pfizer, Polymedix, and Novartis.


View Abstract