A dual antiplatelet regimen of aspirin and clopidogrel is the standard treatment for the prevention of stent thrombosis (1- 3), and retrospective studies have shown that the discontinuation of clopidogrel, even after 6 months or later after stent implantation, is associated with an increased risk of thrombotic events in patients with drug-eluting stents (DES) (4- 7). However, stent thrombosis also can occur in patients taking clopidogrel and aspirin, and it has been shown that patients who suffer stent thrombosis have a high in vitro post-treatment platelet reactivity despite the dual antiplatelet treatment, suggesting that platelet aggregation nonresponsiveness to clopidogrel is the main cause of the thrombotic event (8- 12). The definite demonstration of the association between low in vitro responsiveness to clopidogrel and thrombotic events is still lacking because the large majority of previous studies were retrospective or underpowered. Moreover, post-treatment platelet reactivity may interact with 1 or more established clinical and procedural predictors of stent thrombosis, making it difficult to define its role in precipitating thrombosis. In addition, studies have used different platelet reactivity assessments and definitions for determining the platelet responsiveness to clopidogrel. This prospective study sought to determine the impact of low responsiveness to clopidogrel on the clinical outcome of patients receiving DES.

This study is based on a cohort of 804 consecutive patients who received successful sirolimus- or paclitaxel-eluting stent implantation and for whom platelet reactivity after clopidogrel treatment was prospectively assessed. All patients were considered eligible for the study irrespective of clinical presentation or coronary anatomy. Thus, patients with acute coronary syndromes and ST-segment elevation acute myocardial infarction (AMI) were included, as well as patients with left main disease, chronic total occlusions, bifurcation lesions, or diffuse disease. The only exclusion criteria were: 1) in-hospital death that was not due to stent thrombosis; 2) anticipated noncompliance to dual antiplatelet treatment for at least 6 months; and 3) premature discontinuation of clopidogrel therapy. All patients gave informed consent. The study was approved by the local ethical committee.

All interventions were performed according to current standard guidelines, and the type of stent implanted and the use of IIb/IIIa inhibitors were at discretion of the operator. All patients received aspirin (325 mg) and a loading dose of 600 mg of clopidogrel followed by a maintenance dose of 75 mg daily. Patients on a maintenance dose of ticlopidine or clopidogrel at the time of admission received a loading dose of clopidogrel (600 mg).

Blood samples anticoagulated with 0.129 mol/l sodium citrate (ratio 9:1) for platelet reactivity assessment was obtained 12 to 18 h from clopidogrel loading. For patients receiving in the cath lab both the loading dose of clopidogrel and a IIb/IIIa inhibitor, blood samples were obtained after 6 days while the patient was on the 75-mg maintenance dose of clopidogrel. Platelet-rich plasma, obtained by centrifuging whole blood for 10 min at 200 g, was stimulated with 10 μmol/l adenosine 5′-diphosphate (ADP) (Mascia Brunelli, Milan, Italy) and residual aggregation was assessed using a APACT 4 light transmittance aggregometer (Helena Laboratories, Milan, Italy). The 100% line was set using platelet-poor plasma and the 0 baseline established with platelet-rich plasma (adjusted from 18 × 10⁹/l up to 30 × 10⁹/l). Platelet aggregation (according to the Born’s method) was evaluated considering the maximal percentage of platelet aggregation in response to stimulus. Control samples from 100 healthy volunteers were run to determine the normal reference laboratory aggregation value that was 68% (range 55% to 99%). Patients with platelet aggregation by 10-μmol ADP ≥90th percentile of controls (70%) were defined as nonresponders.

The primary end point of the study was definite or probable stent thrombosis during a 6-month follow-up. Definite stent thrombosis was defined as acute coronary syndrome and either angiographic confirmation of thrombosis or pathological confirmation of thrombosis. Probable stent thrombosis was defined as unexplained death or myocardial infarction in the territory supplied by a stented vessel without angiographic confirmation. The diagnosis of myocardial infarction was based on either the development of new Q waves on 2 or more electrocardiographic leads or an increase of creatine kinase-myocardial band isoenzyme or troponin T >3 times the upper limit of normal. Event time was categorized as acute (within 24 h from stent implantation), subacute (from 1 day to 30 days), and late (30 days to 6 months). The secondary end point was the composite of cardiac mortality and definite or probable stent thrombosis. All events were adjudicated by 3 observers (A.M., D.A., G.M.) who were blinded to patient responsiveness to clopidogrel and not involved in the follow-up process.

All patients had scheduled clinical and electrocadiographic examinations at 1, 3, and 6 months. All other possible information derived from hospital readmission or by the referring physician, relatives, or municipality live registries was entered into the database.

The sample size was calculated assuming the incidence of the primary end point to be 2%, and we hypothesized a 3.5-fold increase in the incidence of stent thrombosis in nonresponders to clopidogrel. With these assumptions, to have a power of 0.80 to detect the hypotesized effect size with a 1-sided p value <0.05, a sample size of at least 800 was needed. Discrete data are summarized as frequencies, whereas continuous data as mean ± SD. The chi-square test was used for comparison of categorical variables. A 2-tailed Student t test was used to test differences among continuous variables. Forward stepwise Cox proportional hazards regression analysis was performed to identify independent correlates of the primary and secondary end point. The variables entered in the proportional hazards regression model were selected using stepwise regression analysis with an entry criterion of p < 0.10. Hazard ratio (HR) and 95% confidence intervals (CIs) were calculated. Survival curves were generated using the Kaplan-Meier method, and the difference between curves was assessed by log-rank test. A p value <0.05 was considered significant. Analyses were performed using the software package SPSS 8.0 (SPSS Inc., Chicago, Illinois).

The study cohort comprised 804 patients who were treated from July 2005 to February 2006 with sirolimus- (Cypher, Cordis Corp., Miami Lakes, Florida) or paclitaxel- (Taxus, Boston Scientific Corp., Natick, Massachusetts) eluting stents. There were 105 patients who were not responsive to clopidogrel (13%). (Table 1) summarizes the clinical and procedural characteristics of the patients. Nonresponsive patients were older as compared with responders; had a greater incidence of diabetes mellitus, unstable angina, and multivessel disease; and had a lower incidence of smoking and ST-segment elevation myocardial infarction. Left ventricular ejection fraction (LVEF) was significantly lower in the nonresponder group. Multivessel intervention was performed more frequently in nonresponders as compared with responders. Again, lesions at high risk of stent thrombosis, such as long lesions and the total chronic occlusions, were treated more frequently in nonresponder patient group. As a consequence, the mean stent length per patient of the nonresponder group was superior to the one of the responder group.

The follow-up rate was 100%. (Table 2) shows the clinical outcome at 6 months. The overall primary end point rate was 3.1%. Patients with ST-segment elevation AMI had a greater thrombosis rate as compared with non-AMI patients (5.1% and 2.3%, respectively, p = 0.052). There were no differences in the stent thrombosis rate among patients receving sirolimus- or paclitaxel-eluting stents or both types of stents (3.4%, 2.3%, and 5.6%, respectively, p = 0.406). Most of stent thromboses (64%) were subacute and occurred at a mean time of 54 ± 62 days from stent implantation (range 4 to 180 days), whereas no patient had acute stent thrombosis.

Definite or probable stent thrombosis occurred in 9 of 105 (8.6%) nonresponders and in 16 of 699 (2.3%) responders (p < 0.001). In both groups, patients who developed stent thrombosis had multiple high-risk features. Among the 16 responder patients with stent thrombosis, there were 7 patients with AMI, 10 with bifurcation lesions, 10 with multivessel percutaneous coronary interventions, 8 with a total stent length >30 mm, and 8 with an LVEF <30%. Similarly, among the 9 nonresponders with stent thrombosis, 3 had AMI, 5 multivessel intervention, 5 bifurcation stenting, 8 had a total stent length >30 mm, and 5 had a LVEF <30%. The event-free survival curves are shown in (Figure 1). The event-free survival rate from the primary end point was 91.3% in the nonresponders and 98.1% in the responders (p < 0.001). The cardiac mortality rate was 8.6% in the nonresponders and 1.4% in responders (p < 0.001). The incidence of the composite of cardiac death and definite or probable stent thrombosis was 10.5% in the nonresponders and 2.7% in the responders (p < 0.001). (Table 3) shows the results of univariate and multivariate analyses. By multivariate analysis, the predictors of the primary end point were as follows: nonresponsiveness to clopidogrel (HR 3.08, 95% CI 1.32 to 7.16; p = 0.009), ST-segment elevation AMI (HR 2.41, 95% CI 1.04 to 5.63; p = 0.041), total stent length (HR 1.01, 95% CI 1.00 to 1.02; p = 0.010), and LVEF (HR 0.95, 95% CI 0.92 to 0.98; p = 0.001). The same variables also were independently related to the risk of the secondary end point: nonresponsiveness to clopidogrel (HR 3.25, 95% CI 1.51 to 7.00; p = 0.003), ST-segment elevation AMI (HR 3.25, 95% CI 1.46 to 7.23; p = 0.004), total stent length (HR 1.01, 95% CI 1.00 to 1.02; p = 0.006), and LVEF (HR 0.94, 95% CI 0.91 to 0.96; p < 0.001).

The definite or probable DES thrombosis rate of 3.1% in this study was greater as compared with randomized trials comparing DES with bare-metal stents and to previous observational studies that reported a thrombosis rate of 0.8% to 1.5% (6- 7,13- 15). Baseline and procedural characteristics of our patient cohort may explain the high stent thrombosis rate: as compared with these studies, our patients were older and had a greater incidence of unstable angina or ST-segment elevation AMI, chronic coronary occlusion, and bifurcation lesion.

This prospective study showed that post-treatment platelet aggregation after a 600-mg dose of clopidogrel is a strong marker of the risk of DES thrombosis. Patients with in vitro >70% post-treatment platelet aggregation had nearly 4-fold increase in definite or probable stent thrombosis as compared with clopidogrel responders.

Platelet reactivity assessment to address clinical outcome was based on the results of previous studies (10,16- 17). In vitro platelet reactivity after clopidogrel loading may be quite variable, and low or nonresponsiveness may be the result of inadequate generation of the active drug metabolite required to inhibit the P2Y₁₂ receptors. The mechanisms of this variability may lie in the hepatic CYP3A4 pathway, in the polymorphisms of CYP3A4 or P2Y₁₂ receptor, or differences in the rate of intestinal absorption of clopidogrel (18- 19). Again, responsiveness, as assessed by in vitro tests, also depends on the baseline platelet reactivity, which is higher in acute coronary syndromes as compared with stable coronary artery disease (20- 21). Furthermore, clopidogrel responsiveness is also dose dependent, and it has been shown that the incidence of low or nonresponders decreases dramatically after a 600-mg loading dose of the drug as compared with the original 300-mg dose (22- 26).

Despite these variables in the single patient that may affect his or her responsiveness to clopidogrel over time, our study, based on a large unrestricted cohort of patients receiving drug-eluting stents, shows that a single assessment of the in vitro responsiveness to clopidogrel is strongly related to the risk of drug-eluting stent thrombosis.

The clinical relevance of low or nonresponsiveness to clopidogrel in patients receiving DES has not been fully investigated. A prospective study based on a sample of 379 patients undergoing coronary stent implantation showed that low responsiveness to clopidogrel is associated with an increased risk of adverse events (HR 3.71, 95% CI 1.08 to 12.69; p = 0.037) at 3-month follow-up (27). This study used an arbitrary cutoff for the definition of low responsiveness (clopidogrel-dependent platelet inhibition <30%), had a very short follow-up, did not focus on DES thrombosis, and also included in the primary end point nonfatal ischemic stroke and any cardiovascular death.

Another study, based on a sample of 802 patients who received elective coronary stent implantation, showed that post-treatment platelet aggregation above the median value carried a 6.7-fold increase in major adverse cardiac events (death, myocardial infarction, and target vessel revascularization) at 1-month follow-up (17). This study enrolled mainly low-risk patients with stable coronary artery disease who received, in the large majority of cases, elective bare-metal stent implantation. Again, this study focused more on periprocedural events, whose rate was very low (1.9%) and mostly driven by target lesion revascularization and non–Q-wave myocardial infarction, than on acute or subacute stent thrombosis. Another small study, based on 60 consecutive patients with ST-segment elevation myocardial infarction who underwent angioplasty and stenting, showed that those patients in the lowest quartile in terms of responsiveness to clopidogrel were at an increased risk for a recurrent cardiovascular event during a 6-month follow-up (8).

Similarly, in a prospective study of 105 patients undergoing percutaneous coronary stenting, Muller et al. (28) found that the 2 patients who developed stent thrombosis were nonresponders to clopidogrel. The Platelet Reactivity in Patients and Recurrent Events Post-Stenting study, based on a sample of 192 patients, showed that post-treatment ADP-induced aggregation in patients with adverse events was higher as compared with patients without events. Also this study did not explore specifically the issue of DES thrombosis, and did not stratify patients before clinical events to different degrees of platelet reactivity (10).

The strong predictive value of stent thrombosis provided by a single platelet aggregation assessment, as revealed by our study, has important clinical implications since a unique bedside examination may have an impact on revascularization strategies in the DES era. Alternative revascularization strategies (coronary surgery or bare-metal stents), or pharmacologic strategies with the aim of a more complete P2Y₁₂ receptor inhibition using increasing doses of clopidogrel or other antiplatelet agents should be considered to reduce the risk of ischemic events (29- 30).

Our study is a a clinical trial whose primary end point is definite or probable stent thrombosis. Probable stent thrombosis is a clinical end point and may overestimate the true incidence of stent thrombosis. Conversely, definite stent thrombosis is an angiographic or pathological end point that surely underestimates the true incidence of stent thrombosis. However, we tried to overcome the intrinsic limitation in probable stent thrombosis definition assessing also the cardiac mortality not attributable to stent thrombosis. We did not report the relation between responsiveness to aspirin associated or unassociated with nonresponsiveness to clopidogrel that could be a variable related to definite or probable stent thrombosis.