OUP user menu

Pharmacodynamic assessment of platelet inhibition by prasugrel vs. clopidogrel in the TRITON-TIMI 38 trial

Alan D. Michelson, Andrew L. Frelinger, Eugene Braunwald, William E. Downey, Dominick J. Angiolillo, Nicholas P. Xenopoulos, Joseph A. Jakubowski, Youfu Li, Sabina A. Murphy, Jie Qin, Carolyn H. McCabe, Elliott M. Antman, Stephen D. Wiviott
DOI: http://dx.doi.org/10.1093/eurheartj/ehp159 1753-1763 First published online: 12 May 2009

Abstract

Aims To examine the extent of platelet inhibition by prasugrel vs. clopidogrel in a TRITON-TIMI 38 substudy.

Methods and results TRITON-TIMI 38 randomized acute coronary syndrome (ACS) patients undergoing percutaneous coronary intervention (PCI) to prasugrel or standard dose clopidogrel. Selected sites prospectively enrolled TRITON-TIMI 38 patients to evaluate adenosine diphosphate (ADP)-attenuated phosphorylation of platelet vasodilator-stimulated phosphoprotein (VASP) (n = 125 patients) and, in a subset (n = 31 patients), ADP-stimulated platelet aggregation. VASP platelet reactivity index (PRI) was lower in prasugrel-treated patients than in clopidogrel-treated patients at 1–2 h post-PCI (≥1 h after loading dose) (P < 0.001) and at 30 days (P < 0.001). Maximal platelet aggregation to 20 µM ADP was lower in prasugrel-treated patients than in clopidogrel-treated patients at 1–2 h (P = 0.004) and 30 days (P = 0.03). Results were similar with 5 µM ADP. Thienopyridine hyporesponsiveness, prespecified as VASP PRI >50%, was more frequent in clopidogrel-treated patients than in prasugrel-treated patients at 1–2 h (P < 0.001) and 30 days (P = 0.03).

Conclusions The TRITON-TIMI 38 platelet substudy shows that prasugrel results in greater inhibition of ADP-mediated platelet function in ACS patients than clopidogrel, supporting the hypothesis that greater platelet inhibition leads to a lower incidence of ischaemic events and more bleeding both early and late following PCI.

  • Platelets
  • Prasugrel
  • Clopidogrel
  • Clinical trials
  • Platelet function

Introduction

Platelets play a central role in the pathogenesis of the ischaemic complications of acute coronary syndrome (ACS) and following percutaneous coronary intervention (PCI).13 Clopidogrel, a thienopyridine prodrug whose active metabolite irreversibly inhibits P2Y12 receptors for adenosine diphosphate (ADP) on platelets, is well established as an antithrombotic therapy in ACS and PCI.4 Nevertheless, ACS and PCI patients treated with a combination of aspirin and clopidogrel continue to have atherothrombotic events, which may be related in part to the significant variability in the degree of inhibition of platelet P2Y12 receptors in response to clopidogrel.46 Indeed, platelet hyporesponsiveness (or ‘resistance’) to clopidogrel has been associated with poor clinical outcomes in patients with ACS and following PCI.711 Potential causes of platelet hyporesponsiveness to clopidogrel include pre-existent platelet hyperreactivity, poor absorption, poor metabolism (e.g. because of single nucleotide polymorphisms in cytochrome P450), increased clearance, increased platelet turnover, or decreased susceptibility of P2Y12 to inhibition as a result of single nucleotide polymorphisms.46,12

Prasugrel is an investigational thienopyridine prodrug13 that, like clopidogrel, requires metabolism to an active metabolite.14,15 Although the active metabolite of prasugrel irreversibly inhibits platelet P2Y12 receptors to a similar extent to the active metabolite of clopidogrel,14 there is much more efficient in vivo generation of the active metabolite of prasugrel than of the active metabolite of clopidogrel.15 Prasugrel has been shown to produce higher, more rapid, and more consistent levels of platelet inhibition compared with clopidogrel in healthy subjects, patients with stable coronary artery disease (CAD), and those undergoing elective PCI,4,1619 but this has not previously been studied in patients with ACS.

The Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition with Prasugrel–Thrombolysis in Myocardial Infarction 38 (TRITON-TIMI 38) randomized 13 608 ACS patients undergoing PCI to prasugrel or standard dose clopidogrel,20 and showed that prasugrel therapy was associated with significantly reduced rates of ischaemic events, including stent thrombosis, but with an increased risk of major bleeding.21 The presently described TRITON-TIMI 38 substudy was not powered to show a significant difference in clinical outcomes between patients on prasugrel and those on clopidogrel. However, this substudy was designed and sufficiently powered to show differences in platelet function between patients on prasugrel and those on clopidogrel. Previous studies have shown that such differences in platelet function affect clinical outcome.711

Methods

Patient population

TRITON-TIMI 38 enrolled patients with ACS with scheduled PCI. Inclusion and exclusion criteria for TRITON-TIMI 38 have been described previously.20,21 Patients meeting these enrolment criteria for TRITON-TIMI 38 and randomized at one of the 13 selected US sites were eligible to participate in the present platelet function substudy. All participating sites mailed samples to a central core laboratory for measurement of vasodilator-stimulated phosphoprotein (VASP) phosphorylation. A pre-selected subset of participating sites (n = 3) also performed light transmission aggregometry (LTA) on site. Additionally, for participation in the LTA component of the platelet function substudy, patients who had received abciximab within the previous 30 days or either eptifibatide or tirofiban within the previous 7 days were excluded. Patients receiving abciximab, eptifibatide, or tirofiban were not excluded from the VASP component of the platelet function substudy. The platelet function substudy complied with the Declaration of Helsinki was approved by the institutional review boards of all participating centres, and written informed consent was provided by all patients.

Study protocol

The study protocol for TRITON-TIMI 38, of which this is a substudy, has been described in detail elsewhere.20,21 Briefly, a loading dose of study medication (60 mg of prasugrel or 300 mg of clopidogrel) was to be administered, in a double-blind manner, as soon as possible between randomization and 1 h after leaving the cardiac catheterization laboratory. If the coronary anatomy was previously known or primary PCI for STEMI was planned, pretreatment with the study drug was permitted for up to 24 h before PCI. Randomization was to occur before PCI was performed. The choice of vessels treated, devices used, and adjunctive medication administered to support PCI was left to the discretion of the treating physician. After PCI, patients received maintenance doses of either prasugrel (10 mg) or clopidogrel (75 mg) daily. The use of aspirin was required, and a loading dose of 325–500 mg and a daily maintenance dose of 75–162 mg was recommended. Samples for the platelet function substudy were collected at baseline (pre-PCI, pre-study drug), 1–2 h post-PCI (at least 1 h after dosing), and 30 days post-PCI.

Laboratory procedures

Blood samples were drawn from venipuncture or indwelling catheters. The first 5 mL was to be discarded. For VASP assays, samples were drawn into a citrated tube, capped, stored, and shipped in insulated containers at ambient temperature to a central core laboratory (Center for Platelet Function Studies, University of Massachusetts Medical School). VASP phosphorylation in response to prostaglandin E1 with and without ADP was determined by whole-blood flow cytometry as described in the manufacturer’s brochure (BioCytex, Marseilles, France), using a single lot of fluorescently labelled monoclonal antibody. The FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA, USA) was calibrated daily with a single lot of calibration particles. VASP phosphorylation data were expressed as platelet reactivity index (PRI) in per cent, defined as: Embedded Image where MFI is mean fluorescence intensity and PGE1 is prostaglandin E1. A lower PRI indicates a greater antiplatelet effect.

For LTA, blood was drawn into a 4.5 mL, 3.2% citrate Vacutainer tube and centrifuged to prepare platelet-rich and platelet-poor plasma. ADP aliquots were prepared at the central core laboratory (Center for Platelet Function Studies, University of Massachusetts Medical School) and shipped to clinical sites. LTA was determined with 20 and 5 µM ADP. Maximal platelet aggregation (MPA) and final platelet aggregation (FPA, defined as the extent of aggregation 5 min after stimulation with ADP) were determined by laboratory personnel unaware of treatment assignment. Inhibition of platelet aggregation (IPA) at time t was defined as follows: (1 − [(MPA at time t after LD)/(MPA at baseline)]) × 100. IPA (final) at time t was defined as follows: (1 − [(FPA at time t after LD)/(FPA at baseline)]) × 100. Evidence of haemolysis, platelet count of the platelet-rich plasma, and procedural irregularities were reported. All tracings were read locally and then overread at the central core laboratory prior to the unblinding of the study by a single reader who was unaware of treatment assignment. Central readings were used for all analyses.

The blinded platelet core laboratory personnel used standardized criteria to assess each determination for validity. Conditions that invalidated LTA samples included, but were not limited to, haemolysis, low platelet-rich plasma platelet counts (<150 000/µL), and instability of baseline LTA tracings indicating instability of sample temperature or pre-agonist sample activation. Sample validity was determined prior to unblinding and analysis. Those samples considered to be invalid were not included in the analyses, and subjects were considered to be non-evaluable for the given measure and time point. For IPA analyses, for a subject to be evaluable at time t, both a valid baseline measurement and time t MPA measurement were required. For MPA or VASP analyses, a subject was evaluable at time t, regardless of the availability of a valid baseline measurement, if there was a valid measurement at time t.

Endpoints

Pharmacodynamic

Pre-specified primary endpoints for the VASP assay were: (i) mean VASP PRI, prasugrel vs. clopidogrel; (ii) prevalence of thienopyridine resistance, as defined by a VASP PRI of >50,7 among prasugrel- and clopidogrel-treated patients. Pre-specified primary endpoints for platelet aggregation were MPA, inhibition (absolute reduction) of MPA (compared with baseline before prasugrel or clopidogrel), and IPA, each at 1–2 h and 30 days after treatment.

Secondary endpoints for platelet aggregation included: FPA, which is of interest because of the role of P2Y12 in stabilizing ADP-induced platelet aggregates;8,22,23 per cent inhibition of FPA (compared with baseline) at 1–2 h and 30 days after treatment; and prevalence of thienopyridine resistance. The latter was defined as an absolute reduction of 20 µM ADP-induced platelet aggregation of <10% or inhibition, compared with baseline, of 20 µM ADP-induced platelet aggregation of <20% at 1–2 h or 30 days.24

Clinical

The primary endpoint of TRITON TIMI 38 was the composite of cardiovascular death, non-fatal MI, or non-fatal stroke. A key secondary endpoint was the composite of cardiovascular death, non-fatal MI, or urgent target vessel revascularization. The key safety endpoint was TIMI major bleeding. Because these events were infrequent overall, in this analysis we considered TIMI major and minor bleeding in addition to haemorrhagic adverse events. Event definitions have been previously reported.20 All components of the key clinical endpoints were adjudicated by an independent clinical events committee unaware of treatment assignment.

Statistical methods

All tests were two sided and conducted at an α = 0.05 level of significance. All CIs presented were two-sided 95% CIs. Standard two-sample comparisons between treatment groups were carried out using the following methods unless otherwise specified. Data were analysed using a mixed ANCOVA model with factors for study site, baseline platelet function results, and treatment received. Categorical data were compared by use of Pearson’s χ2 test, except when the absolute number of events in each group was <5, in which case Fisher’s exact test was used. Continuous data were compared with a two-sample t-test. Analyses were performed, using independent copies of the complete clinical and pharmacodynamic databases, by the Department of Statistics at Nottingham Clinical Research Ltd (Nottingham, UK) using SAS version 9.1 or S-PLUS 6.2 (SAS Institute Inc., Cary, NC, USA), the TIMI Study Group, and the Center for Platelet Function Studies at the University of Massachusetts Medical School. Key analyses carried out at one institute were independently verified by one of the other institutes. The TRITON–TIMI 38 study, and this platelet function substudy, was sponsored by Daiichi Sankyo and Eli Lilly and Company. All drafts of the manuscript were written by the academic authors who made the decision to publish these data and vouch for their accuracy; the sponsors had an opportunity to comment, but held no editorial authority.

Results

Patient population

There were 125 subjects enrolled at 13 sites. The demographics by treatment assignment of the TRITON-TIMI 38 patients enrolled in the VASP component of the platelet function substudy are given in Table 1, and the LTA component is given in Table 2. No significant demographic differences (including pharmacotherapy with aspirin, statin, beta-blocker, calcium blocker, angiotensin-converting enzyme inhibitor, and angiotensin-receptor blocker during index hospitalization [data not shown]) were observed between prasugrel- and clopidogrel-treated patients in either cohort of this substudy.

View this table:
Table 1

Demographics of prasugrel- vs. clopidogrel-treated subjects enrolled in the VASP component of the TRITON TIMI 38 platelet substudy

Prasugrel (%)nNClopidogrel (%)nNP-value
Categorical variables
 Unstable angina or NSTEMI9460649357610.944
 STEMI646474610.944
 Gender (female)2717642817610.870
Medical history
 History hypertension8353648451610.906
 History hypercholesterolaemia7246647043610.864
 History of diabetes3824642616610.177
 Current tobacco use3824643924610.832
 Prior MI2818642113610.378
 Prior CABG11764159610.523
 CrCl <60216153610.309
Index procedure
 PCI performed98636410061610.327
 CABG performed during index hospitalization216400610.327
 Stent used for index PCI9459639357610.962
 BMS only used for index PCI2213591810570.544
 At least 1 DES used for index PCI7846598247570.544
 Multivessel PCI11762138610.757
Antithrombin to support PCI
 UFH6742636740600.296
 LMWH10663181160
 Bivalirudin64637460
 Other/combo1711638560
GPIIb/IIIa inhibitor during index hospitalization6340647043610.344
Timing of study-drug administration
 Pre-PCI171163159610.682
 During PCI835263855261
 Post-PCI00630061
Continuous variables
Age (years)
 Mean ± SD56.5 ± 10.86457.9 ± 10.2610.59
 25th percentile5051
 Median56.558
 75th percentile6464
Body mass index at baseline (mean ± SD)31.3 ± 6.56130.0 ± 6.2610.314
  • Abbreviations: BMS, bare metal stent; CABG, coronary artery bypass graft; CrCl, creatinine clearance; DES, drug-eluting stent; LMWH, low-molecular-weight heparin; MI, myocardial infarction; NSTEMI, non-ST-segment myocardial infarction; PCI, percutaneous coronary intervention; PE, pulmonary embolism; STEMI, ST-segment myocardial infarction; UFH, unfractionated heparin.

View this table:
Table 2

Demographics of prasugrel- vs. clopidogrel-treated subjects enrolled in the light transmission aggregometry component of the TRITON TIMI 38 platelet substudy

Prasugrel (%)nNClopidogrel (%)nNP-value
Categorical variables
 Unstable angina or NSTEMI8814168713150.945
 STEMI13216132150.945
 Gender (female)19316335150.354
Medical history
 History hypertension10016169314150.294
 History hypercholesterolaemia50816609150.576
 History of diabetes50816406150.576
 Current tobacco use44716406150.833
 Prior MI38616132150.124
 Prior CABG19316203150.930
 CrCl <6000130015
Index procedure
 PCI performed10016161001515
 CABG performed during index hospitalization00160015
 Stent used for index PCI8814168012150.570
 BMS only used for index PCI29414253120.838
 At least 1 DES used for index PCI711014759120.838
 Multivessel PCI1931671150.316
Antithrombin to support PCI
 UFH50816274150.417
 LMWH2541653815
 Bivalirudin61167115
 Other/combo1931613215
GPIIb/IIIa inhibitor during index hospitalization2541671150.165
Timing of study-drug administration
 Pre-PCI13216132150.945
 During PCI881416871315
 Post-PCI00160015
Continuous variables
Age (years)
 Mean ± SD54 ± 8.81656 ± 11.7150.752
 25th percentile47.548
 Median5555
 75th percentile6061
Body mass index at baseline (mean ± SD)33 ± 7.11330.3 ± 6.6150.394
  • LTA, light transmission aggregometry. Other abbreviations as in Table 1.

Figure 1 summarizes the evaluability determinations obtained at each time point for prasugrel- and clopidogrel-treated patients enrolled in the VASP and LTA components of the platelet substudies. A similar proportion of samples was non-evaluable in the prasugrel and clopidogrel-treated groups. In a small number of patients, contrary to protocol, baseline samples were collected after study drug was administered (three prasugrel-treated and one clopidogrel-treated subjects). Deviation from protocol also occurred in nine ‘1–2 h’ samples that were collected at <1 h after administration of study drug (three prasugrel-treated and six clopidogrel-treated). Results from these subjects at these times were included in intent-to-treat (ITT) analyses and were excluded from per protocol analyses.

Figure 1

Evaluable samples. (A) All participating sites (n = 13) mailed samples to a central core for measurement of VASP phosphorylation. (B) A pre-selected subset of participating sites (n = 3) also performed LTA on site. Each box contains the number of subjects evaluable for a given measure at a given time point. Abbreviations: IPA, inhibition of platelet aggregation; LD, loading dose; LTA, light transmission aggregation; PCI, percutaneous coronary intervention; VASP, vasodilator stimulated phosphoprotein.

Prasugrel vs. clopidogrel effects on vasodilator-stimulated phosphoprotein platelet reactivity index

Prior to thienopyridine treatment, subjects assigned to prasugrel had similar VASP PRI to those assigned to clopidogrel (Figure 2). Mean VASP PRI was significantly lower in prasugrel-treated subjects than in clopidogrel-treated subjects at both 1 and 2 h post loading dose (51.8 ± 5.1 vs. 78.8 ± 2.5, P < 0.001) and during maintenance dose (30 day) (33.6 ± 2.9 vs. 47.9 ± 2.7, P < 0.001). Results were consistent when analysed on an intention to treat basis, rather than per protocol.

Figure 2

Distribution of individual VASP PRI values in prasugrel- and clopidogrel-treated patients. P-values are from ANCOVA with factors for site, baseline value, and treatment. Line represents mean for each group. Results shown are per protocol. Intent-to-treat analysis gave similarly significant results. Abbreviations: LD, loading dose; PRI, platelet reactivity index; VASP, vasodilator stimulated phosphoprotein.

Despite the presence of 38% diabetic patients in the prasugrel-treated group and 26% diabetic patients in the clopidogrel-treated group (Table 1), VASP PRI did not differ significantly for diabetic vs. non-diabetic subjects at either time point (1–2 h post-loading dose and 30 days), whether combining treatment groups or analysing prasugrel- and clopidogrel-treated subjects separately. However, no firm conclusions may be drawn from this given the low n in each subgroup.

The cumulative distribution of VASP PRI in prasugrel- vs. clopidogrel-treated subjects is shown in Figure 3. At 1–2 h post-loading dose, a greater proportion of prasugrel- than clopidogrel-treated subjects responded to thienopyridine treatment, when defined as VASP PRI <80, <50, or <20 (Figure 3A). At 30 days (i.e. during the maintenance phase period), a greater proportion of prasugrel- than clopidogrel-treated subjects responded to thienopyridine treatment, when defined as VASP PRI <50 or <20 (Figure 3B). A VASP PRI of 50 is a clinically relevant cut-off in certain patient populations,7,10 and thienopyridine hyporesponsiveness, prespecified as VASP PRI >50%, was more frequent in clopidogrel- than in prasugrel-treated patients at 1 and 2 h (96 vs. 57%, P < 0.001) and at 30 days (43 vs. 24%, P = 0.033).

Figure 3

Cumulative distribution of VASP PRI response to prasugrel vs. clopidogrel. P-values determined by χ2 analysis.

Prasugrel vs. clopidogrel effects on platelet aggregation

Figure 4 shows MPA and FPA results in individual subjects in response to 20 and 5 µM ADP. Prior to thienopyridine treatment, subjects assigned to prasugrel-treated group had similar ADP-induced platelet aggregation to subjects assigned to clopidogrel. Mean MPA with ADP 20 µM was significantly lower in prasugrel- than in clopidogrel-treated subjects at both 1 and 2 h post-loading dose (46.5 ± 7.7 vs. 73.7 ± 1.5%, mean ± SE, P = 0.004) and during maintenance dosing (30 days) (39.9 ± 3.2 vs. 55.2 ± 3.2%, P = 0.03) (Figure 4). At 1 and 2 h post-loading dose, prasugrel also resulted in significantly lower FPA in response to ADP 20 µM and lower MPA and FPA in response to ADP 5 µM. At 30 days, FPA with ADP 20 µM and MPA and FPA in response to ADP 5 µM, while numerically lower in prasugrel-treated patients, were not statistically different from clopidogrel-treated patients (but the study was underpowered to detect a significant difference of this magnitude).

Figure 4

Distribution of individual maximal (A and C) and final (B and D) platelet aggregation values in prasugrel- and clopidogrel-treated patients. (A and B) 20 µM ADP. (C and D) 5 µM ADP. P-values are from ANCOVA with factors for site, baseline value, and treatment. Line represents mean for each group. Abbreviations: FPA, final platelet aggregation; LD, loading dose; MPA, maximal platelet aggregation.

Figure 5 shows the cumulative distribution of MPA and per cent IPA with 20 µM ADP in prasugrel- vs. clopidogrel-treated subjects. At both 1 and 2 h post-loading dose and at 30 days, a greater proportion of prasugrel- than clopidogrel-treated subjects responded to thienopyridine treatment, when defined as MPA <50% (Figure 5A and B). Inhibition of platelet aggregation by >50%, a clinically relevant cut-off in certain patient populations,25 occurred significantly more often in prasugrel- than in clopidogrel-treated patients at 1–2 h post-loading dose (Figure 5C). Similarly, at 1–2 h post-loading dose, a greater proportion of prasugrel- than clopidogrel-treated patients had >20% IPA (Figure 5C). The frequency of IPA at 30 days by >50% or by >20% was not significantly different in prasugrel- vs. clopidogrel-treated patients (Figure 5D). Thienopyridine hyporesponsiveness, prespecified as a <10% absolute decrease in platelet aggregation, was more frequent in clopidogrel- than in prasugrel-treated patients at 1–2 h (75 vs. 27%, P = 0.039).

Figure 5

Cumulative distribution of maximal platelet aggregation (MPA) and percent inhibition of platelet aggregation (IPA) with 20 µM ADP. P-values determined by χ2 analysis.

Comparison of vasodilator-stimulated phosphoprotein platelet reactivity index and platelet aggregation in TRITON-TIMI 38

Response to thienopyridine therapy by the VASP assay correlated with response to thienopyridine therapy as measured by MPA with ADP 20 µM (Spearman r = 0.7235, P < 0.0001, Figure 6) and 5 µM (Spearman r = 0.6550, P < 0.0001). Correlation of these independent assays, one (LTA) performed on-site, and one (VASP) performed at a central core facility, suggests that the current results are not due to assay or site artefacts. There was a significant association between the VASP assay and MPA with ADP 20 µM for the identification of responders (lower left quadrant) and non-responders (upper right quadrant) to thienopyridine treatment (relative risk 5.246, 95% CI 2.342–11.75).

Figure 6

Correlation of VASP PRI and maximal platelet aggregation (MPA) with 20 µM ADP. Results shown are for both prasugrel- and clopidogrel-treated subjects in whom both results were valid. Spearman r = 0.7235 (95% CI 0.5774–0.8247), P < 0.0001.

Clinical endpoints

This substudy was not powered to correlate platelet function assays with clinical outcomes. Consistent with this, death from cardiovascular causes, non-fatal MI, or non-fatal stroke (the primary endpoint of TRITON-TIMI 3821) occurred in a total of 18 subjects and was not significantly different between prasugrel- and clopidogrel-treated subjects in this platelet substudy. Likewise, the combined endpoint of death from cardiovascular causes, non-fatal MI, non-fatal stroke, or re-hospitalization for target vessel revascularization, and the haemorrhagic endpoints of TIMI major, TIMI minor, or TIMI major and TIMI minor, and investigator-reported non-serious haemorrhagic events were also not significantly different in prasugrel- vs. clopidogrel-treated subjects (data not shown). Nevertheless, reduced VASP PRI was significantly associated with haemorrhagic events in the follow-up period (combined serious and non-serious haemorrhagic events occurring ≥3 days post-PCI; logistic regression with adjustment for age, sex, and beta-blocker use, OR 0.971, CI 0.944–0.998, P = 0.033). However, this association was no longer significant when adjusted for study site and baseline result.

Discussion

The major findings of the TRITON-TIMI 38 platelet substudy are that at both the 1–2 h and 30 day time points: (i) VASP PRI was significantly lower in prasugrel- than in clopidogrel-treated patients; (ii) platelet aggregation in response to 20 µM ADP was inhibited to a significantly greater extent in prasugrel-treated patients than in clopidogrel-treated patients; (iii) thienopyridine hyporesponsiveness, prespecified as a <10% decrease in platelet aggregation and VASP PRI >50%, was significantly less frequent in prasugrel- than in clopidogrel-treated patients.

The TRITON-TIMI 38 trial showed that in patients with ACS with scheduled PCI, prasugrel therapy was associated with significantly reduced rates of ischaemic events, including stent thrombosis, but with an increased risk of major bleeding.21 This TRITON-TIMI 38 substudy was not powered to show a significant difference in clinical outcomes between patients on prasugrel and those on clopidogrel. However, this substudy was designed and sufficiently powered to show differences in platelet function between patients on prasugrel and those on clopidogrel. Previous studies have shown that such differences in platelet function affect clinical outcome.711 Therefore, our findings support the hypothesis that greater inhibition of platelet function in ACS patients, as shown in the presently described TRITON-TIMI 38 substudy, is responsible for the reduction in platelet-mediated adverse thrombotic events, and the increase in bleeding, with prasugrel when compared with clopidogrel.

Assays used to measure the antiplatelet effects of prasugrel and clopidogrel

Two assays, LTA and VASP, were used to measure the antiplatelet effects of prasugrel and clopidogrel. ADP-stimulated platelet aggregation by LTA is presently considered to be the gold-standard method.26 VASP was chosen because: (a) it is the most specific assay available for the measurement of P2Y12 blockade; (b) it has been linked to clinical outcomes;79 (c) unlike LTA and other platelet function assays, VASP can monitor the inhibition of platelet function by thienopyridines in patients who also receive GP IIb/IIIa antagonists; and (d) it has great logistical advantages in a multicenter trial, because anticoagulated blood can be mailed at room temperature to a single core laboratory.19

Effects of prasugrel vs. clopidogrel on platelet function in an acute coronary syndrome population

Although prasugrel has previously been demonstrated to result in greater and more rapid platelet inhibition than clopidogrel in healthy subjects, patients with stable CAD, and those undergoing PCI,4,1619 the present data are the first to show that prasugrel results in greater and more rapid platelet inhibition than clopidogrel in a purely ACS population. Absolute levels of inhibition of both ADP-stimulated platelet aggregation and ADP-attenuated phosphorylation of VASP by prasugrel at 1–2 h post-loading dose in the present study were less than previously reported normal donor studies16,27 and less than in patients undergoing cardiac catheterization for planned PCI (Prasugrel in Comparison to Clopidogrel for Inhibition of Platelet Activation and Aggregation–Thrombolysis in Myocardial Infarction 44 [PRINCIPLE-TIMI 44]).19 There are several possible explanations for these differences. ACS patients (such as those in TRITON-TIMI 38) have more activated and hyperreactive platelets than patients with stable CAD (such as those in PRINCIPLE-TIMI 44) and normal donors.3 In addition, there were technical differences between the studies (e.g. US subjects were enrolled in the TRITON-TIMI 38 platelet substudy, whereas subjects from Europe, USA, and Israel were enrolled in PRINCIPLE-TIMI 44). However, platelet function assay artefact is an unlikely explanation because of the good agreement between two independent assays (Figure 6), especially because one (LTA) was performed at the clinical site and the other (VASP) was performed at a central core laboratory.

Variation in the degree of inhibition of platelet function

Although prasugrel resulted in greater inhibition of ADP-dependent platelet function than clopidogrel, both antiplatelet agents resulted in considerable variation in the degree of inhibition (Figures 2 and 4). This raises the question as to whether or not it would be clinically beneficial to assess the degree of inhibition of platelet function in patients receiving prasugrel or clopidogrel. The present platelet substudy was not powered to determine whether or not the degree of inhibition of platelet function by prasugrel or clopidogrel correlated with clinical outcomes in this patient cohort. However, other studies have reported that platelet hyporesponsiveness to clopidogrel is associated with poor clinical outcomes in patients with ACS and following PCI.79 The 2006 ACC/AHA PCI guidelines provide a Class IIb recommendation (based on Level C evidence) that, in patients in whom subacute stent thrombosis may be catastrophic or lethal, platelet aggregation studies may be considered and the maintenance dose of clopidogrel increased from 75 to 150 mg per day if <50% IPA is demonstrated.25 In a small study using a VASP PRI cut-off of >50%, Bonello et al.10 reported the first data to suggest that adjusting the clopidogrel loading dose according to the results of a platelet function test is safe and may significantly improve clinical outcomes after PCI. However, there has been no definitive prospective study to determine whether guiding therapy based on the effects of clopidogrel, or any other antiplatelet therapy, on platelet function assays results in clinical benefit to patients. The finding in the present study of a significant association of reduced VASP PRI with the occurrence of haemorrhage events is greatly limited by (a) the post hoc nature of this analysis, (b) the small sample size, and (c) the lack of this association after adjustment for covariables. Nevertheless, the possibility of an association between reduced VASP PRI and haemorrhagic events suggests that inclusion of VASP in appropriately powered future studies could lead to identification of platelet inhibition levels associated with risk of bleeding.

Study limitations

The number of patients in this study was small (n = 125 for the VASP component and n = 31 for the LTA component) and not powered for clinical outcomes, and there were a limited number of time points. However, it is logistically difficult to conduct large platelet function studies as part of prospective, randomized, double-blind, multicentre clinical trials. The reason that the number of subjects enrolled in the LTA component of the study was lower than in the VASP component of the study was the smaller number of participating clinical sites and administration of GP IIb/IIIa antagonists was an exclusion criterion for LTA, but not VASP—because GP IIb/IIIa antagonists interfere with the endpoint of LTA but not VASP. There were a significant number of non-evaluable subjects, due to prespecified preanalytical variables (haemolysis, technical problems). However, evaluable subjects who received prasugrel did not have significantly different baseline characteristics from evaluable subjects who received clopidogrel (Tables 1 and 2).

Conclusion

The results of the TRITON-TIMI 38 platelet substudy support the hypothesis that greater inhibition of ADP-mediated platelet function, both early and late following PCI in ACS patients, led to the lower incidence of ischaemic events, both early and late, and the increase in bleeding, observed in the TRITON-TIMI 38 study.21

Funding

This work was supported by Daiichi Sankyo Company, Limited (Tokyo, Japan) and Eli Lilly and Company (Indianapolis, IN, USA).

Conflict of interest: A.D.M. and A.L.F.: research grant support from Accumetrics, Arena Pharmaceuticals, Dade Behring, GLSynthesis, Lilly/Daiichi Sankyo, McNeil Consumer Healthcare, and sanofi-aventis/Bristol-Myers Squibb. A.D.M.: consulting fees from Lilly/Daiichi Sankyo. E.B.: consulting fees from or serving on paid advisory boards for Daiichi Sankyo and sanofi-aventis, lecture fees from Eli Lilly and sanofi-aventis, research funding to the TIMI Study group from Astra Zeneca, Beckman Coulter, Bristol-Meyers Squibb, Cardiovascular Therapeutics, Genentech, Johnson and Johnson, Merck and Co., Novartis, Pfizer, Roche Diagnostics, and Schering Plough Research. D.J.A.: honoraria/lecture fees from Bristol-Myers Squibb, Daiichi Sankyo, Eli Lilly, sanofi-aventis; honoraria/advisory board from Accumetrics, Arena Pharmaceuticals, Astra Zeneca, Bristol-Myers Squibb, Daiichi Sankyo, Eli Lilly, Medicure, Novartis, Portola, sanofi-aventis, The Medicines Company; research grants from Accumetrics, Astra-Zeneca, Daiichi Sankyo, Eisai, Eli Lilly, GlaxoSmithKline, Otsuka, Portola, Schering-Plough, The Medicines Company. J.A.J.: employee of Eli Lilly & Co. S.D.W., E.M.A. and E.B.; S.A.M. and C.H.M.: research grant support for this and other studies from Eli Lilly and Daiichi Sankyo and for other studies from sanofi-aventis. E.M.A.: consulting fees from or serving on paid advisory boards for sanofi-aventis, lecture fees from Eli Lilly and Sanofi-Aventis. S.D.W.: research support from Schering-Plough, honoraria from Lilly/Daiichi Sankyo, Astra-Zeneca, Accumetrics and Pfizer; consulting fees from sanofi-aventis. W.E.D., N.P.X., and Y.L.; S.A.M. and J.Q. have no conflict of interest.

Acknowledgements

We thank the TRITON-TIMI 38 platelet substudy clinical sites, who are listed in the Supplementary material online, Appendix.

Footnotes

  • These authors contributed equally to this manuscript.

References

View Abstract