Author + information
- Received July 4, 2009
- Revision received September 4, 2009
- Accepted September 14, 2009
- Published online February 9, 2010.
- Dirk Sibbing, MD* (, )
- Julia Stegherr, MS,
- Siegmund Braun, MD,
- Julinda Mehilli, MD,
- Stefanie Schulz, MD,
- Melchior Seyfarth, MD,
- Adnan Kastrati, MD,
- Nicolas von Beckerath, MD and
- Albert Schömig, MD
- ↵*Reprint requests and correspondence:
Dr. Dirk Sibbing, Deutsches Herzzentrum München, Lazarettstrasse 36, 80636 München, Germany
Objectives The goal of this study was to assess whether a platelet rebound exists and whether it can be attenuated by clopidogrel tapering.
Background Clinical studies have reported a clustering of thrombotic events after stopping clopidogrel treatment. The hypothesis of a rebound phenomenon of platelets has been declared causative, but its existence has never been confirmed. Tapering of clopidogrel over a certain period of time before stopping the drug completely might provide a way to attenuate this supposed phenomenon.
Methods Patients (n = 69) receiving clopidogrel treatment due to prior drug-eluting stent placement and planning to stop clopidogrel were recruited in a double-blind, randomized trial. Patients were randomized to either receive a pre-specified tapering regimen (tapering group; n = 35) for 4 weeks with complete discontinuation of clopidogrel thereafter or continue a daily clopidogrel intake for 4 more weeks with abrupt discontinuation afterwards (off group; n = 34). Platelet aggregation (PA) was assessed with light transmission aggregometry (LTA) and multiple electrode aggregometry (MEA) simultaneously at study inclusion and at weeks 2 to 8 after randomization. The primary end point was the highest value of adenosine diphosphate–induced PA measured with LTA in the weeks after complete cessation of clopidogrel in both groups.
Results The highest values of adenosine diphosphate–induced PA after complete cessation of clopidogrel were similar between both groups (p = 0.21 with LTA, and p = 0.55 with MEA).
Conclusions Tapering of clopidogrel does not result in lower PA values after clopidogrel withdrawal. The course of PA values after clopidogrel cessation provides no evidence for the existence of a rebound phenomenon of platelets.
In coronary artery disease (CAD) patients with drug-eluting stent (DES) placement, a dual antiplatelet regimen of aspirin and clopidogrel is the mainstay treatment strategy to prevent thrombosis of the treated vessels and subsequent ischemic events (1–3). Whereas a lifelong treatment with aspirin is recommended for CAD patients with prior stent placement, current guidelines recommend antiplatelet treatment with clopidogrel for a time-period of 6 to 12 months after DES implantation (1–3). In daily clinical practice, patients who are scheduled to stop clopidogrel therapy do it abruptly.
Recently, different clinical studies have demonstrated a clustering of thrombotic events after discontinuation of long-term clopidogrel treatment (4,5). Ho et al. (5) reported, in both medically treated and percutaneous coronary intervention (PCI)-treated patients with an acute coronary syndrome, a clustering of all-cause mortality and acute myocardial infarction in the very first weeks after stopping treatment with clopidogrel. The hypothesis of a rebound phenomenon of blood platelets has been declared causative, but its existence in the context of clopidogrel withdrawal has never been addressed by specifically designed studies. In vitro, such a rebound phenomenon of platelets after clopidogrel cessation might become manifest in a significant peaking of adenosine diphosphate (ADP)-induced platelet aggregation values soon after stopping clopidogrel treatment completely. Because different studies have demonstrated that insufficient suppression of platelet reactivity to ADP is associated with an increased risk of thrombotic events after coronary stent placement (6–9), the observed clustering of adverse events reported in clinical studies might be related to an intermittent status of platelet hyper-reactivity or so-called platelet rebound with very high ADP-induced platelet aggregation levels. A tapering of clopidogrel treatment over a certain period of time before stopping the intake of the drug completely might provide a beneficial treatment strategy to attenuate this supposed rebound phenomenon of platelets (5). The goal of this study was to assess whether a rebound phenomenon exists after stopping long-term clopidogrel treatment and whether it can be attenuated by a clopidogrel-tapering regimen.
Between March 2008 and October 2008 patients were enrolled in this double-blind, randomized study at the Deutsches Herzzentrum München (Technische Universität München, Munich, Germany). The CAD patients on long-term (≥6 months) dual antiplatelet treatment with aspirin (100 mg twice/day) and clopidogrel (75 mg/day) after DES placement were eligible for this study if they were scheduled to stop clopidogrel treatment.
Patients were enrolled after a follow-up angiography and pre-angiography loading with 600 mg of clopidogrel, which is a routine follow-up procedure for patients with prior PCI in our clinic. Only patients who did not need a (re-) intervention were included. This is because a new intervention would have required an extension of dual antiplatelet therapy duration. Exclusion criteria were age ≤18 years and age >80 years, angiographic lesions requiring re-intervention, malignancies or other comorbid conditions that might result in protocol noncompliance, a history of stent thrombosis (ST), prior treatment with glycoprotein IIb/IIIa inhibitors during the 10 days before study inclusion, pregnancy, and relevant hematologic deviations (hemoglobin <100 g/l, platelet count <100 × 109cells/l or platelet count >600 × 109cells/l). The present study complies with the Declaration of Helsinki and was approved by the institutional ethics committee. All patients gave written informed consent before entering the study.
This study was a prospective, randomized, and double-blind platelet function study. Patients with the plan to stop clopidogrel treatment were eligible for study inclusion and for randomization. Patients were randomized to either receive a pre-specified tapering regimen (tapering group) over a pre-defined time period of 28 days with complete cessation of clopidogrel thereafter or continue treatment with 75 mg clopidogrel daily for a further 28 days with abrupt discontinuation afterwards (off group). Sealed envelopes that contained the treatment code and the study medication were used for the randomization process and were provided by the pharmacy of the Deutsches Herzzentrum München, Munich, Germany. The envelopes were numbered sequentially on the basis of a computer-generated randomization list and contained a medication blister with prepared medication in 28 numbered compartments for 28 days. For the off group, each compartment contained 1 tablet of 75 mg clopidogrel (verum). For the tapering group, a pre-specified tapering regimen was used consisting of 75-mg clopidogrel tablets (verum) or placebo tablets. The tapering regimen for the 28 days (4 weeks) was as follows (1 = verum, 0 = placebo): week 1: 1/0/1/0/1/0/1, week 2: 0/0/1/0/0/1/0, week 3: 0/1/0/0/0/1/0, and week 4: 0/0/1/0/0/0/1. Double-blinding was achieved by using similarly appearing tablets for clopidogrel and for placebo in the 2 treatment groups. Thereby, operators and patients were completely blinded to the treatment regimen. All patients received the treatment regimen to which they were randomized. The randomized treatment was started the day after the control angiography before discharge of patients. From this time point onward, all patients were under treatment with 100 mg aspirin (twice/day) and the assigned study medication (once/day) for 28 days. After 28 days the study medication was finished, and aspirin was continued indefinitely.
Study enrollment, randomization, and follow-up of patients for the present trial was completely independent of the currently ongoing double-blind and randomized ISAR-CAUTION (Abrupt Versus Tapered Interruption of Chronic Clopidogrel Therapy After DES Implantation) trial (NCT00640679).
Blood sampling and follow-up of patients
For all patients, peripheral venous blood samples for platelet function testing were drawn at all time points throughout the study in a fasting state through a short venous catheter inserted into a forearm vein. The first tube drawn was labeled as a discard and was not used for platelet function testing. Platelet function was measured with 2 different methods simultaneously—in platelet-rich plasma (PRP) with light transmission aggregometry (LTA) and in whole blood with multiple electrode platelet aggregometry (MEA). For platelet function testing with LTA, blood was placed in 5.0-ml plastic tubes containing the anticoagulant citrate. For platelet function testing with MEA, blood was placed in 4.0-ml plastic tubes containing the anticoagulant lepirudin (25 μg/ml, r-hirudin, Dynabyte, Munich, Germany). Blood samples were kept at room temperature for at least 30 min before platelet function testing.
Blood for platelet function testing was taken at pre-specified time points during the study period. The first sample was taken in the hospital directly after study inclusion after the control coronary angiography. Further samples were taken at outpatient follow-up visits of the patients. These samples were taken at weeks 2, 3, and 4 after randomization, which corresponds to the time period when patients were under the randomized study treatment regimen, and further on at weeks 5, 6, 7, and 8 during the time period without any intake of clopidogrel in all patients. The empty medication blister was collected from all patients to assess compliance with the study protocol.
LTA was used to assess agonist-induced platelet aggregation in citrated PRP with the platelet aggregation profiler (PAP 8) aggregometer (Moelab, Berlin, Germany) with a constant stirring rate of 1,200 rpm at 37°C. The final platelet count was adjusted to 250 × 109/l with autologous platelet-poor plasma (PPP). The PRP (0% light transmission) and PPP (100% light transmission) served as references. As agonists, ADP (Moelab) was used in different concentrations (1.25, 2.5, 5, and 20 μmol/l) as well as collagen (0.19 mg/ml) (Moelab) and thrombin-receptor activating peptide (TRAP) (25 μmol/l) (Sigma, Deisenhofen, Germany). After baseline adjustment, the agonist was added and aggregation was recorded for 6 min. The analyzed parameter was maximal aggregation (%) within the first 6 min after addition of the agonist.
Aggregation in whole blood was assessed on a new generation impedance aggregometer with MEA on the Multiplate analyzer (Dynabyte). Details of this method have been reported previously (8,10). In brief, after dilution (1:2 with 0.9% sodium chloride solution) of hirudin-anticoagulated whole blood and stirring for 3 min in the test cuvettes at 37°C, the agonist was added and aggregation was continuously recorded for 6 min. Platelet aggregation was measured by MEA in response to 6.4 μmol/l ADP (ADPtest), in response to 6.4 μmol/l ADP in the presence of 9.4 nmol/l prostaglandin E1 (ADPtest HS), in response to 32 μmol/l TRAP-6 (TRAPtest), and in response to 3.2 μg/ml collagen (COLtest). Aggregation measured with MEA is quantified as aggregation units (AU) and area under the curve of AU (AU × min). All material used, including the different agonists, was obtained from the manufacturer (Dynabyte).
End points and sample size calculation
The primary end point of the study was the highest level of maximal ADP (5 μmol/l)-induced platelet aggregation measured with LTA in the time period after complete cessation of clopidogrel treatment (weeks 5 to 8). Sample size calculation for the present study was based on the assumption that the mean peak levels of ADP (5 μmol/l)-induced platelet aggregation were 80% in the off group versus 65% in the tapering group (absolute difference in means of 15%). Choosing a power of 80% and a 2-sided α-level of 0.05, we calculated that 29 patients in each group were needed (NQuery advisor, version 7.0, Statistical Solutions, Cork, Ireland). To compensate for a loss to follow-up, we aimed for the inclusion of 34 patients/group. The secondary end point of the study was the highest level of ADP (6.4 μmol/l)-induced platelet aggregation with MEA measured in the time period after complete cessation of clopidogrel treatment (weeks 5 to 8). In addition, we recorded the entire time course of platelet aggregation values in both groups in order to detect a possible rebound phenomenon of platelets and we also tested for significant differences of agonist-induced platelet aggregation values between the 2 study groups for each time point separately in the time period after complete cessation of clopidogrel treatment (weeks 5 to 8).
Variables of patients' baseline characteristics are presented as mean ± SD or counts and percentages. Categorical variables were compared with chi-square test, and continuous variables were compared with unpaired ttest. The Wilcoxon test was used to compare the previous clopidogrel treatment interval between both groups. Platelet function data obtained with LTA and MEA are presented as mean ± SEM and were compared with unpaired ttest for between group comparisons and paired ttest for in group comparisons. Analyses were performed with the software package S-PLUS version 4.5 (Insightful Corp., Seattle, Washington). The p values <0.05 were considered statistically significant. No corrections in the level of significance for multiple comparisons were undertaken, because comparisons at the different time points and for the different agonists were considered as exploratory.
A total of 69 patients were enrolled in this study, of whom 35 patients were randomized to the tapering group and 34 patients were randomized to the off group. Baseline characteristics and concomitant medical treatment of the patients according to the study treatment regimen applied are presented in Table 1.No significant differences were observed between the 2 groups. The controlling of the medication blister demonstrated full compliance with the study protocol in all patients. No ischemic or bleeding events occurred during the study period.
Platelet aggregation with LTA
For the primary end point of the study, the highest value of ADP (5 μmol/l)-induced platelet aggregation (in percentage) measured with LTA after complete cessation of clopidogrel during weeks 5 to 8, no significant differences were observed between the 2 study groups (73.0 ± 2.5% in the off group vs. 69.3 ± 1.5% in the tapering group; p = 0.21). Figure 1shows the time course of ADP-induced platelet aggregation values for the 2 study groups in the weeks after randomization. For all concentrations of ADP used (1.25 to 20 μmol/l), no significant differences were observed for between group comparisons of mean values (off group vs. tapering group) for the distinct time points after complete cessation of clopidogrel at weeks 5, 6, 7, and 8 (p ≥ 0.39 for 1.25 μmol/l ADP, p ≥ 0.40 for 2.5 μmol/l ADP, p ≥ 0.35 for 5 μmol/l ADP, p ≥ 0.54 for 20 μmol/l ADP). Figure 2shows the course of TRAP- and collagen-induced platelet aggregation values for the 2 study groups in the weeks after randomization. For both agonists used, no significant differences were observed for between group comparisons of mean values (off group vs. tapering group) for the distinct time points after complete cessation of clopidogrel at weeks 5, 6, 7, and 8 (p ≥ 0.15 for TRAP, and p ≥ 0.31 for collagen).
Due to administration of a 600-mg loading dose of clopidogrel on top of a clopidogrel maintenance treatment before the coronary angiography, platelet aggregation values were significantly lower at week 0 after randomization compared with week 2 after randomization (p < 0.01 for both LTA and MEA measurements).
Platelet aggregation with MEA
For the secondary end point of the study, the highest value of ADP (6.4 μmol/l)-induced platelet aggregation (in AU × min) measured with MEA after complete cessation of clopidogrel during weeks 5 to 8, no significant differences were observed between the 2 study groups (925 ± 43 AU × min in the off group vs. 890 ± 39 AU × min in the tapering group; p = 0.55). Figure 3shows the course of agonist-induced platelet aggregation values for the 2 study groups in the weeks after randomization. For all agonists used, no significant differences were observed for between group comparisons of mean values (off group vs. tapering group) for the distinct time points after complete cessation of clopidogrel at weeks 5, 6, 7, and 8 (p ≥ 0.32 for ADP, p ≥ 0.18 for ADP in the presence of prostaglandin E1, p ≥ 0.48 for TRAP-6, p ≥ 0.06 for collagen).
Here we report the first double-blind and randomized trial on the prevention and existence of a rebound phenomenon of platelets after cessation of clopidogrel treatment. The first main result of the study is that a tapering of clopidogrel treatment as compared with abrupt cessation of the drug does not result in significantly lower platelet aggregation values in the time period after complete cessation of clopidogrel treatment. The second main result of the study is that the time course of platelet aggregation values—regardless of the device, the agonist, or the agonist concentration used—after clopidogrel cessation provides no evidence for the existence of a rebound phenomenon of platelets after discontinuing clopidogrel.
Results of this study were obtained concordantly with 2 different devices used for platelet function testing. Specifically, we measured platelet aggregation in PRP with LTA, which is still considered to be the gold standard in this setting, but also in whole blood with MEA on the Multiplate analyzer (Dynabyte) (8,10). For both LTA and MEA, an association of post-treatment platelet reactivity in response to ADP and the occurrence of ischemic events has been demonstrated in prospective studies (6,8,11). Therefore major results of our study are strengthened by the fact that we found no evidence for the existence of a rebound phenomenon and for a possible benefit of clopidogrel tapering with any of the 2 different devices used. Additional LTA measurements were performed with low concentrations of ADP (1.25 and 2.5 μmol/l) to avoid extensive aggregation after complete cessation of clopidogrel and that could have masked a possible rebound phenomenon. These measurements did not show any evidence for platelet rebound after abrupt cessation of clopidogrel therapy. Besides P2Y12 pathway stimulation with ADP, we implemented the use of collagen and TRAP as agonists for platelet aggregation measurements. For all agonists used and regardless of the concentration that was used for ADP, evidence for the existence of a rebound phenomenon or for a possible benefit of tapering clopidogrel was not provided by any of our investigations.
In the present study we investigated platelet function parameters over a total time span of 8 weeks, including measurements over a period of 4 weeks after complete cessation of clopidogrel. It is highly unlikely that a rebound phenomenon of platelets would occur later than 4 weeks after complete cessation of the drug, for several reasons. Having left the bone marrow, platelets survive for approximately 8 to 10 days on average, and approximately 10% to 15% of platelets are exchanged/day (12,13). In line with this, it was reported in different studies that a full recovery of platelet function after discontinuation of clopidogrel is observed within 5 to 7 days after complete withdrawal of the drug (13,14). If a rebound phenomenon of platelets would indeed exist, it would be likely to occur at that time point, when the entire platelet pool is exchanged after clopidogrel withdrawal, with platelets in the circulation being completely uninhibited and in an assumed hyper-reactive state. This time point might be estimated—taking into account the platelet life span and the time of platelet function recovery after clopidogrel withdrawal—at week 2 after complete cessation of the drug, which corresponds to week 6 after randomization in the off group of patients in our study setting. For all platelet function parameters studied here, however, we did not observe a peaking of platelet aggregation values at week 6 or even thereafter (Figs. 1 to 3).
Limited data are available on proaggregatory or proinflammatory effects after withdrawal of thienopyridine treatment. Serebruany et al. (15) described a patient in whom relatively high markers of platelet activation and aggregation were measured after discontinuation of prasugrel due to noncompliance. In addition, Angiolillo et al. (16) observed proaggregatory and proinflammatory effects in diabetic patients after clopidogrel withdrawal. None of these studies, however, was designed to assess a possible rebound phenomenon of platelets or a possible benefit of tapering clopidogrel treatment.
A clustering of thrombotic events has previously been reported for different antithrombotic agents in the time course after their cessation, including aspirin (17), unfractionated heparin (18,19), and clopidogrel (5,20,21). For clopidogrel, however, evidence for the existence of a rebound phenomenon of platelets after drug cessation mainly stems from clinical studies investigating the occurrence of clinical events in relation to stopping clopidogrel treatment. Available clinical studies, however, have provided conflicting results in this context. In a retrospective cohort study of 3,137 acute coronary syndrome patients from Ho et al. (5), a clustering of death or myocardial infarction was observed in the very first weeks after stopping clopidogrel in both medically treated and PCI-treated patients. In line with this, other studies and case report collections in patients with prior stent placement have also reported on a clustering of ischemic events, including ST early after cessation of thienopyridine treatment (20–22). Whereas possible evidence for the existence of a rebound phenomenon of platelets is provided by these studies, results of other studies do not fit into this assumption. In an observational study including 10,778 patients with sirolimus-eluting stent implantation, discontinuation of thienopyridine treatment alone was not associated with an excess of ST in any of the time intervals studied (23). In a prospective observational cohort study including 3,021 patients with prior DES placement, opposing results were found for a possible association of adverse events and discontinuation of thienopyridine therapy (4). In this cohort study, discontinuation of thienopyridine treatment was only found to be a strong and independent predictor (p < 0.001) for the occurrence of ST within 6 months after coronary stenting. Thienopyridine cessation after 6 months of stenting did not predict the occurrence of ST (p = 0.92). Recently, similar results were reported by Lemesle et al. (24), who compared patients (n = 2,254) who stopped their clopidogrel treatment within the first 6 months, between 6 and 12 months, and after 12 months after coronary stent placement. This study demonstrated that clopidogrel cessation was only associated with an increased short-term risk of adverse events when clopidogrel was stopped within the first 6 months after PCI. Results of the latter studies (4,24) challenge the existence of a rebound phenomenon of platelets after clopidogrel cessation, which would be expected to be manifest at any time point after stopping the drug, and further strengthen the hypothesis that the clustering of thrombotic events after DES implantation and cessation of clopidogrel is mainly related to the extent of re-endothelialization and stent coverage.
Although tapering of clopidogrel treatment does not result in significantly lower absolute values of ADP-induced platelet aggregation in the weeks after stopping clopidogrel, a possible clinical benefit of tapering clopidogrel in CAD patients with prior stent placement cannot be excluded from the results of the present study. It could be hypothesized that, apart from the maximal values of platelet aggregation observed, a more gradual increase of platelet aggregation values achieved by a clopidogrel-tapering regimen is beneficial for the reduction of thrombotic events. In fact, we observed a relatively rapid increase of platelet aggregation values in the off group of patients in our study. Whether this rapid increase might be disadvantageous in case of stopping clopidogrel treatment remains uncertain. The clinical safety and efficacy of abrupt versus tapered interruption of chronic clopidogrel therapy after DES implantation will be addressed in a separate and independent study, the ISAR-CAUTION trial.
In this study extensive assessment of agonist-induced platelet aggregation was performed with different ADP concentrations and additional agonists with 2 different methods. We did not assess surface expression of activation-dependent platelet membrane receptors. However, ADP-induced platelet aggregation has been by far the most broadly used parameter to link platelet function to the risk of ischemic events after PCI. Only patients with stable coronary artery disease and concomitant aspirin treatment were included in our study. Therefore it cannot be excluded that a rebound phenomenon of platelets after clopidogrel cessation might occur in clinically unstable patients or in patients without any concomitant oral antiplatelet treatment. Here, we have evaluated a potentially clinically relevant issue, that of platelet function response after clopidogrel discontinuation in patients receiving chronic dual antiplatelet treatment after PCI. Obviously, it is unclear whether the phenomenon under investigation should be considered as “rebound” or simply as recovery of platelet function after 2 approaches of clopidogrel discontinuation.
Tapering of clopidogrel does not result in lower platelet aggregation values after complete cessation of clopidogrel treatment. The course of platelet aggregation values after clopidogrel cessation provides no evidence for the existence of a rebound phenomenon of platelets after discontinuation of clopidogrel.
This work was supported by a research grant from Cordis and Medtronic. Materials for platelet function analysis on the Multiplate device were provided free of charge from Dynabyte. The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation, review, or approval of the manuscript. Dr. Sibbing has received speaker fees from Dynabyte and fees for advisory board activities from Eli Lilly. Dr. Kastrati has received speaker fees from Eli Lilly, Sanofi-Aventis, and Bristol-Myers Squibb. Dr. von Beckerath has received speaker fees from Eli Lilly and fees for advisory board activities from Eli Lilly and Sanofi-Aventis.
- Abbreviations and Acronyms
- adenosine diphosphate
- aggregation units
- coronary artery disease
- drug-eluting stent(s)
- light transmission aggregometry
- multiple electrode platelet aggregometry
- percutaneous coronary intervention
- platelet-poor plasma
- platelet-rich plasma
- stent thrombosis
- thrombin-receptor activating peptide
- Received July 4, 2009.
- Revision received September 4, 2009.
- Accepted September 14, 2009.
- American College of Cardiology Foundation
- King S.B. III.,
- Smith S.C. Jr..,
- Hirshfeld J.W. Jr..,
- et al.
- Silber S.,
- Albertsson P.,
- Aviles F.F.,
- et al.
- Smith S.C. Jr..,
- Feldman T.E.,
- Hirshfeld J.W. Jr..,
- et al.
- Airoldi F.,
- Colombo A.,
- Morici N.,
- et al.
- Buonamici P.,
- Marcucci R.,
- Migliorini A.,
- et al.
- Sibbing D.,
- Braun S.,
- Morath T.,
- et al.
- Marcucci R.,
- Gori A.M.,
- Paniccia R.,
- et al.
- Hochholzer W.,
- Trenk D.,
- Bestehorn H.P.,
- et al.
- Angiolillo D.J.,
- Fernandez-Ortiz A.,
- Bernardo E.,
- et al.
- Ferrari E.,
- Benhamou M.,
- Cerboni P.,
- Marcel B.
- Granger C.B.,
- Miller J.M.,
- Bovill E.G.,
- et al.
- Valgimigli M.,
- Campo G.,
- Arcozzi C.,
- et al.
- Grines C.L.,
- Bonow R.O.,
- Casey D.E. Jr..,
- et al.
- Kimura T.,
- Morimoto T.,
- Nakagawa Y.,
- et al.
- Lemesle G.,
- Bonello L.,
- De Labriolle A.,
- et al.