- Email: [email protected]
Effect of Prolonged Thienopyridine Use After Drug-Eluting Stent Implantation (from the TAXUS Landmark Trials Data)
Effect of Prolonged Thienopyridine Use After Drug-Eluting Stent Implantation (from the TAXUS Landmark Trials Data) Gregg W. Stone, MDa,b,*, Stephen G. Ellis, MDc, Antonio Colombo, MDd, Eberhard Grube, MDe, Keith D. Dawkins, MDf, Mark Friedman, MDf, and Donald S. Baim, MDf Previous observational studies have evaluated the utility of prolonged thienopyridine therapy in patients receiving drug-eluting stents, with conflicting results. A landmark analysis was therefore performed on the basis of the prospective, double-blind TAXUS-II SR, TAXUS-IV, and TAXUS-V trials. Of 2,736 randomized patients, 2,171 were event free at 1 year, of whom 964 (44.4%) at physician discretion were still taking thienopyridines at that time. Among the 1,141 event-free patients randomized to Taxus stents, a trend was present in those patients still taking compared with those having discontinued thienopyridines at 1 year to have numerically fewer very late stent thrombosis episodes at 2 years (0 vs 4 [0.7%] events, respectively, p ⴝ 0.07) and 5 years (4 [0.8%] vs 8 [1.4%] events, respectively, p ⴝ 0.43). However, in event-free Taxus patients taking compared with those not taking thienopyridines at 1 year, there were no significant differences in the 2- and 5-year rates of death (1.4% vs 0.6%, p ⴝ 0.22, and 6.0% vs 7.4%, p ⴝ 0.83, respectively) and death or myocardial infarction (1.8% vs 1.9%, p ⴝ 0.82, and 8.3% vs 9.8%, p ⴝ 0.75, respectively). There were no significant interactions between 1-year thienopyridine use status and stent type (Taxus vs bare-metal stent) on the individual or composite end points of stent thrombosis, death, or myocardial infarction at either 2 or 5 years (all interaction p values >0.50). In conclusion, thienopyridine use beyond 1 year after drug-eluting stent implantation may reduce stent thrombosis over the subsequent 12-month period, but the rates of death and myocardial infarction at 2 and 5 years are not prevented by extended thienopyridine use after either drug-eluting or bare-metal stent implantation. © 2008 Elsevier Inc. All rights reserved. (Am J Cardiol 2008;102:1017–1022) Current recommendations call for 1 year of thienopyridine use to minimize the risk for stent thrombosis (ST) after coronary artery drug-eluting stent (DES) implantation in patients not at high risk for bleeding.1 However, chronic thienopyridine use is costly and is associated with an approximate 1% risk for major bleeding per year.2– 4 No randomized trials have been performed to determine the optimal duration of dual-antiplatelet therapy after coronary stenting. A landmark analysis from the Duke database suggested that the 2-year rates of death and myocardial infarction (MI) after DES implantation (but not after bare-metal stent [BMS] implantation) were decreased in patients taking compared with those not taking thienopyridines at 6 months or 1 year.5 ST events were not assessed in this study, however, and in a large European multicenter study, thienopyridine discontinuation after 6 months in patients with DES was not a risk factor for subsequent ST.6 The conflicting findings from these studies may reflect inherent limita-
a The Cardiovascular Research Foundation, New York, New York; Columbia University Medical Center, New York, New York; cCleveland Clinic, Cleveland, Ohio; dSan Raffaele Hospital, Milan, Italy; eHELIOS Klinikum Siegburg, Siegburg, Germany; and fBoston Scientific Corporation, Natick, Massachusetts. Manuscript received April 10, 2008; revised manuscript received and accepted May 23, 2008. This study was funded by Boston Scientific Corporation, Natick, Massachusetts. *Corresponding author: Tel: 212-851-9304; fax: 212-851-9396. E-mail address: [email protected] (G.W. Stone).
b
0002-9149/08/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2008.05.050
tions of retrospective observational databases, including varying patient characteristics, end points, and definitions and incomplete monitoring and adverse event ascertainment. In addition, DES and BMS selection was not blinded in the Duke study, potentially introducing bias. To overcome many of these limitations, we examined the issue of prolonged thienopyridine use from the prospective, doubleblind, randomized trials of the Taxus DES (Boston Scientific Corporation, Natick, Massachusetts). Methods Study design: To evaluate the impact of long-term thienopyridine use in a uniform cohort of patients receiving DES and BMS in whom rigorous study procedures, monitoring, and consistent definitions were applied, we pooled the databases from the prospective, multicenter, doubleblind, placebo-controlled TAXUS-II SR, TAXUS-IV, and TAXUS-V, trials in which 2,736 total patients were randomized 1:1 to the commercialized slow-release Taxus stent or an otherwise identical BMS.7–10 The enrollment criteria and demographic and lesion characteristics of the patients enrolled in these trials have recently been summarized.10 In brief, patients presenting with mostly stable ischemic syndromes eligible for stent implantation in single de novo noncomplex lesions in native coronary arteries (reference vessel diameter 2.25 to 4.0 mm, lesion length ⱕ46 mm) were eligible for enrollment (Table 1). In addition to indefinite aspirin use, each protocol mandated thienopyridine use for ⱖ6 months (clopidogrel preferred); thereafter, extended www.AJConline.org
1018
The American Journal of Cardiology (www.AJConline.org)
Table 1 Selected features of the Taxus trials Variable Stent platform Diabetes mellitus Reference vessel diameter (mm) Lesion length (mm) Maximum planned stents Latest available follow-up (yrs)
TAXUS-II SR (n ⫽ 266)
TAXUS-IV (n ⫽ 1,314)
TAXUS-V (n ⫽ 1,156)
NIRx 11.3% 3.0–3.5
Express 24.2% 2.5–3.75
Express 30.8% 2.25–4.0
10–12 1 5
10–28 1 5
10–46 2 3
Figure 1. Patient flow in the study. Of 2,736 patients randomized to Taxus stents or BMS in the TAXUS-II SR, TAXUS-IV, and TAXUS-V trials, 2,171 were event free at 1 year, comprising the study population. Patients in each group were then stratified according to whether or not they were taking thienopyridines at 1 year, with subsequent event rates analyzed at 2 and 5 years. See text for details.
thienopyridine use was per investigator discretion. The case report form for each trial collected whether the patient was taking a thienopyridine agent at the time of hospital discharge; at 1 month, 4 months, 9 months, and 1 year; and then yearly for 5 years. Clinical follow-up was performed concordantly. Previous studies have consistently shown increased rates of ST after DES implantation in patients with thienopyridine discontinuation before 6 months,6,11,12 and current guidelines recommend dual-antiplatelet therapy for 1-year in patients at low risk for bleeding.1 Whether longer term thienopyridine use is beneficial remains undetermined. We therefore performed a long-term landmark analysis5 confined to patients who remained free from death, MI, target vessel revascularization, or ST at 1 year and stratified those patients according to whether a thienopyridine was or was not being taken at exactly 1 year (the T⫹1yr and T⫺1yr groups, respectively; Figure 1). Subsequent clinical outcomes were then examined for the latest follow-up available from each trial (currently 5 years for TAXUS-II SR and TAXUS-IV and 3 years for TAXUS-V; median duration of follow-up 4.3 years for all studies), with all patients and study personnel having remained blinded to stent type. Definitions: All end points were adjudicated by an independent clinical events committee without knowledge of
the assigned stent type. Death was defined as due to any cause. MI was defined as either the development of new pathologic Q waves ⬎0.04 seconds in duration in ⱖ2 contiguous leads with an elevated creatine phosphokinase-MB level or, in the absence of Q waves, an elevation of creatine phosphokinase level to ⬎2.0 times normal with positive creatine kinase-MB (creatine phosphokinase ⬎5.0 times elevated after bypass graft surgery). Target vessel revascularization was defined as repeat revascularization (by either percutaneous coronary intervention or coronary artery bypass graft surgery) for ischemia with either angiographic restenosis of the target lesion or a ⬎50% diameter stenosis in a nontarget lesion in the target vessel with either electrocardiographic changes at rest or positive functional study results, or a ⬎70% diameter stenosis with recurrent symptoms only. ST was defined as either definite or probable using the Academic Research Consortium criteria.13 Statistical analysis: Categorical variables were compared using the chi-square or Fisher’s exact test. Continuous variables were compared using unpaired Student’s t tests and are expressed as mean ⫾ 1 SD. Event rates were determined using Kaplan-Meier time-to-event estimates and were compared using the log-rank test at 2 years and 5 years (representing 1- and 4-year follow-up, respectively, beyond the 1-year event-free landmark). Multivariate Cox proportional-hazards regression was performed, adjusting for the baseline covariates in Table 2, to determine whether thienopyridine use at 1 year was an independent determinate of late events at 2 and 5 years. Formal interaction testing was performed to determine whether thienopyridine use at 1 year had a differential impact on the occurrence of late adverse events in patients randomized to Taxus stents compared with BMS. All analyses are by intention to treat, including all patients randomized to each stent, and all p values are 2 sided. The investigators had full access to the data and take responsibility for its integrity. All investigators have read and agree to the report as written. Results Patients: Of the 2,736 randomized patients, 2,171 were event free at 1 year, including 1,141 patients randomized to Taxus stents and 1,030 patients randomized to BMS (Figure 1). Of the 2,171 event-free patients, 964 (44.4%) were taking thienopyridines at 1 year, including clopidogrel in 952 (98.8%) and ticlopidine in 12 (1.2%). Consistent with the double-blind nature of the study, similar proportions of patients with Taxus stents and BMS were taking thienopyridines at 1 year (Figure 1). However, thienopyridine continuation at 1 year was not randomized, and event-free patients still taking thienopyridines at 1 year were more likely to be diabetic and had longer lesions, more stents, and greater total stent length implanted than those not still taking such medications (Table 2). Other baseline clinical and angiographic characteristics were similar between the T⫹1yr and T⫺1yr patients, including the use of aspirin. Thienopyridine use over time: As seen in Figure 2, thienopyridine use in the T⫹1yr patients progressively decreased at each subsequent year, such that at the end of the 5-year study period, approximately 57% of patients with
Coronary Artery Disease/Long-Term Thienopyridine Use
1019
Table 2 Baseline features and aspirin use in patients randomized to Taxus stents or bare-metal stents according to thienopyridine use at 1 year Variable
Randomized to Taxus Stents T⫺1yr (n ⫽ 624)
Age (yrs) Men Diabetes mellitus Insulin requiring Current smokers Reference vessel diameter (mm) Minimal luminal diameter (mm) Diameter stenosis (%) Lesion length (mm) Total no. of stents Total stent length (mm) Aspirin use At 1 yr At 2 yrs At 3 yrs At 4 yrs At 5 yrs
T⫹1yr (n ⫽ 517)
63.1 ⫾ 10.9 443 (71.0%) 124 (19.9%) 34 (5.4%) 141 (22.6%) 2.75 ⫾ 0.50 0.91 ⫾ 0.33 66.8 ⫾ 10.6 13.7 ⫾ 7.0 1.1 ⫾ 0.4 22.0 ⫾ 9.4
62.3 ⫾ 11.2 378 (73.1%) 152 (29.4%) 41 (7.9%) 111 (21.5%) 2.74 ⫾ 0.50 0.90 ⫾ 0.37 67.4 ⫾ 11.3 15.5 ⫾ 8.0 1.3 ⫾ 0.5 25.4 ⫾ 11.8
600/624 (96.2%) 570/611 (93.3%) 528/583 (90.6%) 363/406 (89.4%) 354/394 (89.8%)
502/517 (97.1%) 475/498 (95.4%) 441/479 (92.1%) 190/215 (88.4%) 186/211 (88.2%)
Randomized to BMS p Value 0.21 0.47 0.0002 0.09 0.67 0.75 0.44 0.34 0.0001 ⬍0.0001 ⬍0.0001 0.42 0.16 0.45 0.69 0.58
T⫺1yr (n ⫽ 583)
T⫹1yr (n ⫽ 447)
61.8 ⫾ 10.8 417 (71.5%) 135 (23.2%) 48 (8.2%) 125 (21.4%) 2.78 ⫾ 0.51 0.94 ⫾ 0.35 66.2 ⫾ 10.9 13.9 ⫾ 7.0 1.1 ⫾ 0.4 21.6 ⫾ 9.2
62.5 ⫾ 10.4 327 (73.2%) 117 (26.2%) 29 (6.5%) 90 (20.1%) 2.73 ⫾ 0.51 0.91 ⫾ 0.37 66.9 ⫾ 11.7 15.0 ⫾ 8.3 1.2 ⫾ 0.5 24.5 ⫾ 11.4
558/583 (95.7%) 540/566 (95.4%) 509/545 (93.4%) 348/387 (89.9%) 340/375 (90.7%)
432/447 (96.6%) 412/430 (95.8%) 378/412 (91.7%) 159/175 (90.9%) 150/170 (88.2%)
p Value 0.29 0.58 0.27 0.34 0.64 0.08 0.13 0.33 0.02 0.03 ⬍0.0001 0.52 0.88 0.38 0.88 0.44
Data are expressed as mean ⫾ SD or number (percentage). Table 3 Event rates according to thienopyridine use at 1 year
Figure 2. Thienopyridine use during the 5-year study period in patients taking compared with those not taking thienopyridines at the 1-year landmark.
Taxus stents and BMS were still taking thienopyridines. Conversely, in the T⫺1yr patients, thienopyridine use progressively increased at each subsequent year, such that at the end of 5 years, approximately 15% of patients with Taxus stents and BMS were taking thienopyridines. Event rates according to thienopyridine use at 1 year: As listed in Table 3 and shown in Figure 3, a trend was present in patients with Taxus stents taking compared with those not taking thienopyridines at 1 year for numerically fewer ST episodes at 2 years (0 vs 4 [0. 7%] events, respectively, p ⫽ 0.07). Of the 4 ST events in the T⫺1yr group, 3 were classified as definite ST and 1 as probable ST. Beyond 2 years, an equal number of ST episodes occurred in the T⫺1yr and T⫹1yr groups, such that numerically fewer ST episodes had occurred in the T⫹1yr group (4 [0.8%] vs 8 [1.4%] events, respectively) at 5 years, although this difference no longer approached statistical significance (p ⫽ 0.43). Similar to the DES group, there were slightly fewer ST episodes at 5 years in patients with BMS taking com-
Stent thrombosis TAXUS (T⫹1yr) TAXUS (T⫺1yr) BMS (T⫹1yr) BMS (T⫺1yr) Death TAXUS (T⫹1yr) TAXUS (T⫺1yr) BMS (T⫹1yr) BMS (T⫺1yr) Death or MI TAXUS (T⫹1yr) TAXUS (T⫺1yr) BMS (T⫹1yr) BMS (T⫺1yr) Death, MI, or ST TAXUS (T⫹1yr) TAXUS (T⫺1yr) BMS (T⫹1yr) BMS (T⫺1yr)
2-Year Rate, % (n)
p Value
5-Year Rate, % (n)
p Value
0.0% (0) 0.7% (4) 0.2% (1) 0.0% (0)
0.07
0.8% (4) 1.4% (8) 0.2% (1) 0.7% (3)
0.43
1.4% (7) 0.6% (4) 2.0% (9) 1.4% (8)
0.22
6.0% (25) 7.4% (37) 6.9% (21) 6.6% (35)
0.83
1.8% (9) 1.9% (12) 2.5% (11) 1.9% (11)
0.82
8.3% (35) 9.8% (51) 7.9% (25) 9.2% (48)
0.75
1.8% (9) 1.9% (12) 2.5% (11) 1.9% (11)
0.82
8.3% (35) 9.8% (51) 7.9% (25) 9.2% (48)
0.75
0.25
0.41
0.51
0.51
0.59
0.66
0.28
0.28
pared with those not taking thienopyridines at 1 year (1 [0.2%] vs 3 [0.7%], respectively, p ⫽ 0.59). There were no significant differences, however, in the 2- and 5-year rates of death, composite death or MI, and composite death, MI, or ST in patients with Taxus stents and BMS taking compared with those not taking thienopyridines at 1 year. Nor were there significant interactions between stent type (Taxus vs BMS) and thienopyridine use status at 1 year on the individual end points of ST or death or the composites of death or MI or death, MI, or ST at either 2 or 5 years (all interaction p values ⬎0.50). Finally, after adjustment for differences in baseline covariates, thienopyridine use status
1020
The American Journal of Cardiology (www.AJConline.org)
Figure 3. Rates of ST (top left), death (top right), composite death or MI (bottom left), and composite death, MI, or ST (bottom right) in patients randomized to Taxus stents or BMS, according to thienopyridine use at the 1-year landmark.
at 1 year was not an independent predictor of adverse events at either 2 or 5 years (Figure 4). Timing of thienopyridine discontinuation and Taxus ST: The exact temporal relation between ST and thienopyridine use in the 12 patients with Taxus stents in whom ST developed is shown in Figure 5. In the 8 patients with Taxus stents not taking thienopyridines at 1 year in whom ST subsequently developed, the ST episodes occurred 410 to 1,334 days after Taxus stent implantation, with a delay from thienopyridine discontinuation to ST of ⱖ228 days in all patients. In the 4 patients with Taxus stents taking thienopyridines at 1 year in whom ST subsequently developed, 2 patients had discontinued the thienopyridines (14 and 244 days before the ST episodes), and 2 were still taking thienopyridines at the time of the ST events. Discussion The present study provides limited support that thienopyridine continuation beyond 1 year is beneficial after DES implantation. The major conclusions from the present study can be summarized as follows: (1) In patients treated with DES in single noncomplex de novo native coronary artery lesions who were event free at 1 year, a trend was present for fewer ST episodes over the subsequent 12-month period in patients taking compared with those no longer taking thienopyridines at 1 year, although such patients were more frequently diabetic and had longer lesions and more stents
implanted; (2) however, no significant trends were apparent between thienopyridine use status at 1 year and the subsequent rates of death, composite death or MI, and composite death, MI, or ST at either 2 or 5 years in patients receiving either DES or BMS, and prolonged thienopyridine use after the second year did not further reduce ST; (3) there were no significant interactions between the effects of long-term thienopyridine use and stent type (Taxus vs BMS) on longterm cardiovascular outcomes, including ST death and death or MI; and (4) the interval between thienopyridine discontinuation and ST was prolonged in most cases of ST, which does not suggest a causal relation of thienopyridine withdrawal to ST. Elements from the present study do suggest a possible though limited benefit of prolonged thienopyridine use in patients with DES. The rate of ST in patients receiving Taxus stents from 1 to 2 years was 0.7% in the T⫺1yr compared with 0% in the T⫹1yr group (p ⫽ 0.07), whereas ST from 1 to 2 years was rare after BMS implantation and unaffected by continuing thienopyridine use after 1 year. Moreover, 10 of the 12 episodes of Taxus ST from 1 to 5 years occurred after thienopyridine discontinuation, including 2 of 4 patients in the T⫹1yr group, demonstrating the confounding from therapy crossover that may occur during follow-up in landmark studies. However, tempering these findings, the duration between thienopyridine discontinuation and Taxus ST was prolonged in 9 of 10 cases, ranging from ⱖ228 to ⬎1,000 days, making less certain a causal
Coronary Artery Disease/Long-Term Thienopyridine Use
2-Year Event Rates T+1yr T- 1yr
HR [95% CI] ST
Death
Death or MI
Death, MI, or ST
TAXUS
N/A
0.0%
0.7%
Cox P value N/A
BMS
N/A
0.2%
0.0%
N/A
TAXUS
0.36 [0.10-1.23]
1.4%
0.6%
0.10
BMS
0.71 [0.28-1.85]
2.0%
1.4%
0.49
TAXUS
1.10 [0.46-2.60]
1.8%
1.9%
0.84
BMS
0.78 [0.34-1.79]
2.5%
1.9%
0.55
TAXUS
1.10 [0.46-2.60]
1.8%
1.9%
0.84
BMS
0.78 [0.34-1.79]
2.5%
1.9%
0.55
1.0 0 2.0 3.0 T+ Worse T+ Better
HR [95% CI] ST
Death
Death or MI
Death, MI, or ST
5-Year Event Rates T+1yr T-1yr
Cox P value
TAXUS
1.65 [0.50-5.48]
0.8%
1.4%
0.41
BMS
1.79 [0.18-17.19]
0.2%
0.7%
0.62
TAXUS
1.08 [0.64-1.81]
6.0%
7.4%
0.78 0.48
BMS
1.22 [0.70-2.11]
6.9%
6.6%
TAXUS
1.06 [0.68-1.64]
8.3%
9.8%
0.81
BMS
1.34 [0.83-2.18]
7.9%
9.2%
0.24
TAXUS
1.06 [0.68-1.64]
8.3%
9.8%
0.81
BMS
1.34 [0.83-2.18]
7.9%
9.2%
0.24
1.0 3.0 0 2.0 T+ Worse T+ Better
Figure 4. Event rates, hazard ratios (HRs), and 95% confidence intervals (CIs) comparing patients taking with those not taking thienopyridines at 1 year, with p values adjusted for differences in baseline covariates. Top graph: outcomes at 2 years; bottom graph: outcomes at 5 years. N/A ⫽ not applicable.
Figure 5. Timing of ST in relation to thienopyridine use in the 12 patients in whom ST developed after Taxus stent implantation. See text for details. * Patients not taking thienopyridines at 1-year follow-up but in whom the exact dates of discontinuation were unknown.
relation between thienopyridine use at 1 year or later and the prevention of incipient ST. These data are consistent with and extend the observations made by Airoldi et al,6 who reported that in patients with DES, the median time from clopidogrel termination to ST was 13.5 days when clopidogrel was prematurely discontinued before 6 months, compared with 90 days when it was stopped after the first 6 months.
1021
Moreover, a similar proportion of T⫹1yr and T⫺1yr patients developed ST from 2 to 5 years in the present study, and as a result, the correlation between thienopyridine use at 1 year and subsequent ST after Taxus stent implantation was less apparent by 5 years than at 2 years. Importantly, no significant differences were apparent in either the 2- or 5-year rates of death, composite death or MI, and composite death, MI, or ST in patients with Taxus stents and BMS taking compared with those not taking thienopyridines at 1 year. Despite the trend toward reduced ST at 2 years in patients with Taxus stents in the T⫹1yr group, the unadjusted rate of death tended to be paradoxically greater at 2 years in the T⫹1yr group compared with the T⫺1yr patients (p ⫽ 0.10) and then nonsignificantly lower in T⫹1yr compared with T⫺1yr patients at 5 years, for BMS and Taxus stents. This favorable very late trend may reflect a positive benefit of extended clopidogrel administration in patients with established cardiovascular disease,4,14 independent of stent type or even coronary intervention, although this hypothesis remains to be definitively proved in a randomized trial, and the likelihood of a chance finding cannot be excluded. Finally, in contradistinction to the Duke landmark study,5 there were no significant interactions between thienopyridine use status at 1 year and stent type (Taxus vs BMS) on the subsequent occurrence of death, MI, ST or their composite occurrence either at 2 or at 5 years. The performance of a landmark analysis within the framework of a prospective, randomized trial overcomes many but not all of the drawbacks inherent in earlier analyses from retrospective observational registry studies. Specifically, randomization and the double-blind nature of the present study design eliminate bias by study personnel and adjudication committees on the basis of stent type and ensure complete monitoring of all study end points in all patients. Despite these advantages, as in previous such studies, the duration of thienopyridine use was uncontrolled and nonrandomized, and prolonged clopidogrel administration was not surprisingly used more frequently in higher risk patients (those with diabetes) with longer lesions requiring more stents. Although multivariate analysis was used in an attempt to correct for such measured confounders, we were unable to adjust for unmeasured or more subtle factors that might affect thienopyridine use duration, such as bleeding propensity or the requirement for noncardiac surgery, the performance of which was not routinely collected. Second, although more detailed information regarding dual-antiplatelet therapy was collected in the trials of Taxus stents than is typically available in uncontrolled observational registries, the exact dates of thienopyridine use and reasons for discontinuation were not prospectively assessed, nor was thienopyridine use status at exactly 6 months measured, precluding analysis of the potential utility of clopidogrel from 6 to 12 months. Third, there was a nontrivial rate of therapy crossover between the T⫹1yr and T⫺1yr groups during follow-up, a phenomenon that to our knowledge has not previously been described, which makes it more likely that such an analysis will be most accurate when applied to adverse events occurring shortly after the landmark date. Fourth, information regarding bleeding events was not collected after hospital discharge, an important consider-
1022
The American Journal of Cardiology (www.AJConline.org)
ation when considering the risk/benefit ratio of prolonged clopidogrel use, because recent studies have clearly established that hemorrhagic complications from medications may otherwise offset the life-prolonging benefits of reduced ischemia.15–17 Finally, the Taxus trials excluded many categories of “real-world” patients, including those receiving stents for multiple lesions, bifurcations, left main coronary arter or saphenous vein graft disease, in-stent restenosis and acute MI; as such, the results of the present investigation apply only to stent implantation in single de novo lesions in native coronary arteries in patients with mostly stable ischemic heart disease. Follow-up for the most complex patients in the study, those enrolled in the TAXUS-V trial, was available only to 3 years. Similarly, although one of the largest studies of its kind, ST and death occurred relatively infrequently in the present study, even out to 5 years; a larger study may have elicited a smaller magnitude of benefit (or risk) of prolonged thienopyridine use than was possible in the present investigation. 1. King SB III, Smith SC Jr, Hirshfeld JW Jr, Jacobs AK, Morrison DA, Williams DO, Feldman TE, Kern MJ, O’Neill WW, Schaff HV, et al. 2007 focused update of the ACC/AHA/SCAI 2005 guideline update for percutaneous coronary intervention. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation 2008;117:261–295. 2. Clopidogrel in Unstable Angina to Prevent Recurrent Ischemic Events Trial Investigators. Effects of clopidogrel in addition to aspirin in patients with acute coronary syndromes without ST-segment elevation. N Engl J Med 2001;345:494 –502. 3. Steinhubl SR, Berger PB, Mann JT III, Fry ET, DeLago A, Wilmer C, Topol EJ; CREDO Investigators. Early and sustained dual oral antiplatelet therapy following percutaneous coronary intervention: a randomized controlled trial. JAMA 2002;288:2411–2420. 4. Bhatt DL, Fox KA, Hacke W, Berger PB, Black HR, Boden WE, Cacoub P, Cohen EA, Creager MA, Easton JD, et al; CHARISMA Investigators. Clopidogrel and aspirin versus aspirin alone for the prevention of atherothrombotic events. N Engl J Med 2006;354:1706 – 1717. 5. Eisenstein EL, Anstrom KJ, Kong DF, Shaw LK, Tuttle RH, Mark DB, Kramer JM, Harrington RA, Matchar DB, Kandzari DE, et al. Clopidogrel use and long-term clinical outcomes after drug-eluting stent implantation. JAMA 2007;297:159 –168. 6. Airoldi F, Colombo A, Morici N, Latib A, Cosgrave J, Buellesfeld L, Bonizzoni E, Carlino M, Gerckens U, Godino C, et al. Incidence and
7.
8.
9.
10.
11. 12.
13.
14.
15.
16.
17.
predictors of drug-eluting stent thrombosis during and after discontinuation of thienopyridine treatment. Circulation 2007;116: 745–754. Colombo A, Drzewiecki J, Banning A, Grube E, Hauptmann K, Silber S, Dudek D, Fort S, Schiele F, Zmudka K, et al; TAXUS II Study Group. Randomized study to assess the effectiveness of slow- and moderate-release polymer-based paclitaxel-eluting stents for coronary artery lesions. Circulation 2003;108:788 –794. Stone GW, Ellis SG, Cox DA, Hermiller J, O’Shaughnessy C, Mann JT, Turco M, Caputo R, Bergin P, Greenberg J, et al; TAXUS-IV Investigators. A polymer-based paclitaxel-eluting stent in patients with coronary artery disease. N Engl J Med 2004;350:221–231. Stone GW, Ellis SG, Cannon L, Mann JT, Greenberg JD, Spriggs D, O’Shaughnessy CD, DeMaio S, Hall P, Popma JJ, et al; TAXUS V Investigators. Comparison of a polymer-based paclitaxel-eluting stent with a bare metal stent in patients with complex coronary artery disease: a randomized controlled trial. JAMA 2005;294:1215–1223. Stone GW, Moses JW, Ellis SG, Schofer J, Dawkins KD, Morice MC, Colombo A, Schampaert E, Grube E, Kirtane AJ, et al. Safety and efficacy of sirolimus- and paclitaxel-eluting coronary stents. N Engl J Med 2007;356:998 –1008. Ong AT, McFadden EP, Regar E, de Jaegere PP, van Domburg RT, Serruys PW. Late angiographic stent thrombosis (LAST) events with drug-eluting stents. J Am Coll Cardiol 2005;45:2088 –2092. Jeremias A, Sylvia B, Bridges J, Kirtane AJ, Bigelow B, Pinto DS, Ho KK, Cohen DJ, Garcia LA, Cutlip DE, Carrozza JP Jr. Stent thrombosis after successful sirolimus-eluting stent implantation. Circulation 2004;109:1930 –1932. Cutlip DE, Windecker S, Mehran R, Boam A, Cohen DJ, van Es GA, Steg PG, Morel MA, Mauri L, Vranckx P, et al; Academic Research Consortium. Clinical end points in coronary stent trials: a case for standardized definitions. Circulation 2007;115:2344 –2351. Bhatt DL, Flather MD, Hacke W, Berger PB, Black HR, Boden WE, Cacoub P, Cohen EA, Creager MA, Easton JD, et al; CHARISMA Investigators. Patients with prior myocardial infarction, stroke, or symptomatic peripheral arterial disease in the CHARISMA trial. J Am Coll Cardiol 2007;49:1982–1988. Feit F, Voeltz MD, Attubato MJ, Lincoff AM, Chew DP, Bittl JA, Topol EJ, Manoukian SV. Predictors and impact of major hemorrhage on mortality following percutaneous coronary intervention: an analysis of the REPLACE-2 trial. Am J Cardiol 2007;100:1364 –1369. Manoukian SV, Feit F, Mehran R, Voeltz MD, Ebrahimi R, Hamon M, Dangas GD, Lincoff AM, White HD, Moses JW, et al. Impact of major bleeding on 30-day mortality and clinical outcomes in patients with acute coronary syndromes: an analysis from the ACUITY trial. J Am Coll Cardiol 2007;49:1362–1368. Wiviott SD, Braunwald E, McCabe CH, Montalescot G, Ruzyllo W, Gottlieb S, Neumann FJ, Ardissino D, De Servi S, Murphy SA, et al; TRITON-TIMI 38 Investigators. Prasugrel versus clopidogrel in patients with acute coronary syndromes. N Engl J Med 2007;357:2001– 2015.
a The Cardiovascular Research Foundation, New York, New York; Columbia University Medical Center, New York, New York; cCleveland Clinic, Cleveland, Ohio; dSan Raffaele Hospital, Milan, Italy; eHELIOS Klinikum Siegburg, Siegburg, Germany; and fBoston Scientific Corporation, Natick, Massachusetts. Manuscript received April 10, 2008; revised manuscript received and accepted May 23, 2008. This study was funded by Boston Scientific Corporation, Natick, Massachusetts. *Corresponding author: Tel: 212-851-9304; fax: 212-851-9396. E-mail address: [email protected] (G.W. Stone).
b
0002-9149/08/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2008.05.050
tions of retrospective observational databases, including varying patient characteristics, end points, and definitions and incomplete monitoring and adverse event ascertainment. In addition, DES and BMS selection was not blinded in the Duke study, potentially introducing bias. To overcome many of these limitations, we examined the issue of prolonged thienopyridine use from the prospective, doubleblind, randomized trials of the Taxus DES (Boston Scientific Corporation, Natick, Massachusetts). Methods Study design: To evaluate the impact of long-term thienopyridine use in a uniform cohort of patients receiving DES and BMS in whom rigorous study procedures, monitoring, and consistent definitions were applied, we pooled the databases from the prospective, multicenter, doubleblind, placebo-controlled TAXUS-II SR, TAXUS-IV, and TAXUS-V, trials in which 2,736 total patients were randomized 1:1 to the commercialized slow-release Taxus stent or an otherwise identical BMS.7–10 The enrollment criteria and demographic and lesion characteristics of the patients enrolled in these trials have recently been summarized.10 In brief, patients presenting with mostly stable ischemic syndromes eligible for stent implantation in single de novo noncomplex lesions in native coronary arteries (reference vessel diameter 2.25 to 4.0 mm, lesion length ⱕ46 mm) were eligible for enrollment (Table 1). In addition to indefinite aspirin use, each protocol mandated thienopyridine use for ⱖ6 months (clopidogrel preferred); thereafter, extended www.AJConline.org
1018
The American Journal of Cardiology (www.AJConline.org)
Table 1 Selected features of the Taxus trials Variable Stent platform Diabetes mellitus Reference vessel diameter (mm) Lesion length (mm) Maximum planned stents Latest available follow-up (yrs)
TAXUS-II SR (n ⫽ 266)
TAXUS-IV (n ⫽ 1,314)
TAXUS-V (n ⫽ 1,156)
NIRx 11.3% 3.0–3.5
Express 24.2% 2.5–3.75
Express 30.8% 2.25–4.0
10–12 1 5
10–28 1 5
10–46 2 3
Figure 1. Patient flow in the study. Of 2,736 patients randomized to Taxus stents or BMS in the TAXUS-II SR, TAXUS-IV, and TAXUS-V trials, 2,171 were event free at 1 year, comprising the study population. Patients in each group were then stratified according to whether or not they were taking thienopyridines at 1 year, with subsequent event rates analyzed at 2 and 5 years. See text for details.
thienopyridine use was per investigator discretion. The case report form for each trial collected whether the patient was taking a thienopyridine agent at the time of hospital discharge; at 1 month, 4 months, 9 months, and 1 year; and then yearly for 5 years. Clinical follow-up was performed concordantly. Previous studies have consistently shown increased rates of ST after DES implantation in patients with thienopyridine discontinuation before 6 months,6,11,12 and current guidelines recommend dual-antiplatelet therapy for 1-year in patients at low risk for bleeding.1 Whether longer term thienopyridine use is beneficial remains undetermined. We therefore performed a long-term landmark analysis5 confined to patients who remained free from death, MI, target vessel revascularization, or ST at 1 year and stratified those patients according to whether a thienopyridine was or was not being taken at exactly 1 year (the T⫹1yr and T⫺1yr groups, respectively; Figure 1). Subsequent clinical outcomes were then examined for the latest follow-up available from each trial (currently 5 years for TAXUS-II SR and TAXUS-IV and 3 years for TAXUS-V; median duration of follow-up 4.3 years for all studies), with all patients and study personnel having remained blinded to stent type. Definitions: All end points were adjudicated by an independent clinical events committee without knowledge of
the assigned stent type. Death was defined as due to any cause. MI was defined as either the development of new pathologic Q waves ⬎0.04 seconds in duration in ⱖ2 contiguous leads with an elevated creatine phosphokinase-MB level or, in the absence of Q waves, an elevation of creatine phosphokinase level to ⬎2.0 times normal with positive creatine kinase-MB (creatine phosphokinase ⬎5.0 times elevated after bypass graft surgery). Target vessel revascularization was defined as repeat revascularization (by either percutaneous coronary intervention or coronary artery bypass graft surgery) for ischemia with either angiographic restenosis of the target lesion or a ⬎50% diameter stenosis in a nontarget lesion in the target vessel with either electrocardiographic changes at rest or positive functional study results, or a ⬎70% diameter stenosis with recurrent symptoms only. ST was defined as either definite or probable using the Academic Research Consortium criteria.13 Statistical analysis: Categorical variables were compared using the chi-square or Fisher’s exact test. Continuous variables were compared using unpaired Student’s t tests and are expressed as mean ⫾ 1 SD. Event rates were determined using Kaplan-Meier time-to-event estimates and were compared using the log-rank test at 2 years and 5 years (representing 1- and 4-year follow-up, respectively, beyond the 1-year event-free landmark). Multivariate Cox proportional-hazards regression was performed, adjusting for the baseline covariates in Table 2, to determine whether thienopyridine use at 1 year was an independent determinate of late events at 2 and 5 years. Formal interaction testing was performed to determine whether thienopyridine use at 1 year had a differential impact on the occurrence of late adverse events in patients randomized to Taxus stents compared with BMS. All analyses are by intention to treat, including all patients randomized to each stent, and all p values are 2 sided. The investigators had full access to the data and take responsibility for its integrity. All investigators have read and agree to the report as written. Results Patients: Of the 2,736 randomized patients, 2,171 were event free at 1 year, including 1,141 patients randomized to Taxus stents and 1,030 patients randomized to BMS (Figure 1). Of the 2,171 event-free patients, 964 (44.4%) were taking thienopyridines at 1 year, including clopidogrel in 952 (98.8%) and ticlopidine in 12 (1.2%). Consistent with the double-blind nature of the study, similar proportions of patients with Taxus stents and BMS were taking thienopyridines at 1 year (Figure 1). However, thienopyridine continuation at 1 year was not randomized, and event-free patients still taking thienopyridines at 1 year were more likely to be diabetic and had longer lesions, more stents, and greater total stent length implanted than those not still taking such medications (Table 2). Other baseline clinical and angiographic characteristics were similar between the T⫹1yr and T⫺1yr patients, including the use of aspirin. Thienopyridine use over time: As seen in Figure 2, thienopyridine use in the T⫹1yr patients progressively decreased at each subsequent year, such that at the end of the 5-year study period, approximately 57% of patients with
Coronary Artery Disease/Long-Term Thienopyridine Use
1019
Table 2 Baseline features and aspirin use in patients randomized to Taxus stents or bare-metal stents according to thienopyridine use at 1 year Variable
Randomized to Taxus Stents T⫺1yr (n ⫽ 624)
Age (yrs) Men Diabetes mellitus Insulin requiring Current smokers Reference vessel diameter (mm) Minimal luminal diameter (mm) Diameter stenosis (%) Lesion length (mm) Total no. of stents Total stent length (mm) Aspirin use At 1 yr At 2 yrs At 3 yrs At 4 yrs At 5 yrs
T⫹1yr (n ⫽ 517)
63.1 ⫾ 10.9 443 (71.0%) 124 (19.9%) 34 (5.4%) 141 (22.6%) 2.75 ⫾ 0.50 0.91 ⫾ 0.33 66.8 ⫾ 10.6 13.7 ⫾ 7.0 1.1 ⫾ 0.4 22.0 ⫾ 9.4
62.3 ⫾ 11.2 378 (73.1%) 152 (29.4%) 41 (7.9%) 111 (21.5%) 2.74 ⫾ 0.50 0.90 ⫾ 0.37 67.4 ⫾ 11.3 15.5 ⫾ 8.0 1.3 ⫾ 0.5 25.4 ⫾ 11.8
600/624 (96.2%) 570/611 (93.3%) 528/583 (90.6%) 363/406 (89.4%) 354/394 (89.8%)
502/517 (97.1%) 475/498 (95.4%) 441/479 (92.1%) 190/215 (88.4%) 186/211 (88.2%)
Randomized to BMS p Value 0.21 0.47 0.0002 0.09 0.67 0.75 0.44 0.34 0.0001 ⬍0.0001 ⬍0.0001 0.42 0.16 0.45 0.69 0.58
T⫺1yr (n ⫽ 583)
T⫹1yr (n ⫽ 447)
61.8 ⫾ 10.8 417 (71.5%) 135 (23.2%) 48 (8.2%) 125 (21.4%) 2.78 ⫾ 0.51 0.94 ⫾ 0.35 66.2 ⫾ 10.9 13.9 ⫾ 7.0 1.1 ⫾ 0.4 21.6 ⫾ 9.2
62.5 ⫾ 10.4 327 (73.2%) 117 (26.2%) 29 (6.5%) 90 (20.1%) 2.73 ⫾ 0.51 0.91 ⫾ 0.37 66.9 ⫾ 11.7 15.0 ⫾ 8.3 1.2 ⫾ 0.5 24.5 ⫾ 11.4
558/583 (95.7%) 540/566 (95.4%) 509/545 (93.4%) 348/387 (89.9%) 340/375 (90.7%)
432/447 (96.6%) 412/430 (95.8%) 378/412 (91.7%) 159/175 (90.9%) 150/170 (88.2%)
p Value 0.29 0.58 0.27 0.34 0.64 0.08 0.13 0.33 0.02 0.03 ⬍0.0001 0.52 0.88 0.38 0.88 0.44
Data are expressed as mean ⫾ SD or number (percentage). Table 3 Event rates according to thienopyridine use at 1 year
Figure 2. Thienopyridine use during the 5-year study period in patients taking compared with those not taking thienopyridines at the 1-year landmark.
Taxus stents and BMS were still taking thienopyridines. Conversely, in the T⫺1yr patients, thienopyridine use progressively increased at each subsequent year, such that at the end of 5 years, approximately 15% of patients with Taxus stents and BMS were taking thienopyridines. Event rates according to thienopyridine use at 1 year: As listed in Table 3 and shown in Figure 3, a trend was present in patients with Taxus stents taking compared with those not taking thienopyridines at 1 year for numerically fewer ST episodes at 2 years (0 vs 4 [0. 7%] events, respectively, p ⫽ 0.07). Of the 4 ST events in the T⫺1yr group, 3 were classified as definite ST and 1 as probable ST. Beyond 2 years, an equal number of ST episodes occurred in the T⫺1yr and T⫹1yr groups, such that numerically fewer ST episodes had occurred in the T⫹1yr group (4 [0.8%] vs 8 [1.4%] events, respectively) at 5 years, although this difference no longer approached statistical significance (p ⫽ 0.43). Similar to the DES group, there were slightly fewer ST episodes at 5 years in patients with BMS taking com-
Stent thrombosis TAXUS (T⫹1yr) TAXUS (T⫺1yr) BMS (T⫹1yr) BMS (T⫺1yr) Death TAXUS (T⫹1yr) TAXUS (T⫺1yr) BMS (T⫹1yr) BMS (T⫺1yr) Death or MI TAXUS (T⫹1yr) TAXUS (T⫺1yr) BMS (T⫹1yr) BMS (T⫺1yr) Death, MI, or ST TAXUS (T⫹1yr) TAXUS (T⫺1yr) BMS (T⫹1yr) BMS (T⫺1yr)
2-Year Rate, % (n)
p Value
5-Year Rate, % (n)
p Value
0.0% (0) 0.7% (4) 0.2% (1) 0.0% (0)
0.07
0.8% (4) 1.4% (8) 0.2% (1) 0.7% (3)
0.43
1.4% (7) 0.6% (4) 2.0% (9) 1.4% (8)
0.22
6.0% (25) 7.4% (37) 6.9% (21) 6.6% (35)
0.83
1.8% (9) 1.9% (12) 2.5% (11) 1.9% (11)
0.82
8.3% (35) 9.8% (51) 7.9% (25) 9.2% (48)
0.75
1.8% (9) 1.9% (12) 2.5% (11) 1.9% (11)
0.82
8.3% (35) 9.8% (51) 7.9% (25) 9.2% (48)
0.75
0.25
0.41
0.51
0.51
0.59
0.66
0.28
0.28
pared with those not taking thienopyridines at 1 year (1 [0.2%] vs 3 [0.7%], respectively, p ⫽ 0.59). There were no significant differences, however, in the 2- and 5-year rates of death, composite death or MI, and composite death, MI, or ST in patients with Taxus stents and BMS taking compared with those not taking thienopyridines at 1 year. Nor were there significant interactions between stent type (Taxus vs BMS) and thienopyridine use status at 1 year on the individual end points of ST or death or the composites of death or MI or death, MI, or ST at either 2 or 5 years (all interaction p values ⬎0.50). Finally, after adjustment for differences in baseline covariates, thienopyridine use status
1020
The American Journal of Cardiology (www.AJConline.org)
Figure 3. Rates of ST (top left), death (top right), composite death or MI (bottom left), and composite death, MI, or ST (bottom right) in patients randomized to Taxus stents or BMS, according to thienopyridine use at the 1-year landmark.
at 1 year was not an independent predictor of adverse events at either 2 or 5 years (Figure 4). Timing of thienopyridine discontinuation and Taxus ST: The exact temporal relation between ST and thienopyridine use in the 12 patients with Taxus stents in whom ST developed is shown in Figure 5. In the 8 patients with Taxus stents not taking thienopyridines at 1 year in whom ST subsequently developed, the ST episodes occurred 410 to 1,334 days after Taxus stent implantation, with a delay from thienopyridine discontinuation to ST of ⱖ228 days in all patients. In the 4 patients with Taxus stents taking thienopyridines at 1 year in whom ST subsequently developed, 2 patients had discontinued the thienopyridines (14 and 244 days before the ST episodes), and 2 were still taking thienopyridines at the time of the ST events. Discussion The present study provides limited support that thienopyridine continuation beyond 1 year is beneficial after DES implantation. The major conclusions from the present study can be summarized as follows: (1) In patients treated with DES in single noncomplex de novo native coronary artery lesions who were event free at 1 year, a trend was present for fewer ST episodes over the subsequent 12-month period in patients taking compared with those no longer taking thienopyridines at 1 year, although such patients were more frequently diabetic and had longer lesions and more stents
implanted; (2) however, no significant trends were apparent between thienopyridine use status at 1 year and the subsequent rates of death, composite death or MI, and composite death, MI, or ST at either 2 or 5 years in patients receiving either DES or BMS, and prolonged thienopyridine use after the second year did not further reduce ST; (3) there were no significant interactions between the effects of long-term thienopyridine use and stent type (Taxus vs BMS) on longterm cardiovascular outcomes, including ST death and death or MI; and (4) the interval between thienopyridine discontinuation and ST was prolonged in most cases of ST, which does not suggest a causal relation of thienopyridine withdrawal to ST. Elements from the present study do suggest a possible though limited benefit of prolonged thienopyridine use in patients with DES. The rate of ST in patients receiving Taxus stents from 1 to 2 years was 0.7% in the T⫺1yr compared with 0% in the T⫹1yr group (p ⫽ 0.07), whereas ST from 1 to 2 years was rare after BMS implantation and unaffected by continuing thienopyridine use after 1 year. Moreover, 10 of the 12 episodes of Taxus ST from 1 to 5 years occurred after thienopyridine discontinuation, including 2 of 4 patients in the T⫹1yr group, demonstrating the confounding from therapy crossover that may occur during follow-up in landmark studies. However, tempering these findings, the duration between thienopyridine discontinuation and Taxus ST was prolonged in 9 of 10 cases, ranging from ⱖ228 to ⬎1,000 days, making less certain a causal
Coronary Artery Disease/Long-Term Thienopyridine Use
2-Year Event Rates T+1yr T- 1yr
HR [95% CI] ST
Death
Death or MI
Death, MI, or ST
TAXUS
N/A
0.0%
0.7%
Cox P value N/A
BMS
N/A
0.2%
0.0%
N/A
TAXUS
0.36 [0.10-1.23]
1.4%
0.6%
0.10
BMS
0.71 [0.28-1.85]
2.0%
1.4%
0.49
TAXUS
1.10 [0.46-2.60]
1.8%
1.9%
0.84
BMS
0.78 [0.34-1.79]
2.5%
1.9%
0.55
TAXUS
1.10 [0.46-2.60]
1.8%
1.9%
0.84
BMS
0.78 [0.34-1.79]
2.5%
1.9%
0.55
1.0 0 2.0 3.0 T+ Worse T+ Better
HR [95% CI] ST
Death
Death or MI
Death, MI, or ST
5-Year Event Rates T+1yr T-1yr
Cox P value
TAXUS
1.65 [0.50-5.48]
0.8%
1.4%
0.41
BMS
1.79 [0.18-17.19]
0.2%
0.7%
0.62
TAXUS
1.08 [0.64-1.81]
6.0%
7.4%
0.78 0.48
BMS
1.22 [0.70-2.11]
6.9%
6.6%
TAXUS
1.06 [0.68-1.64]
8.3%
9.8%
0.81
BMS
1.34 [0.83-2.18]
7.9%
9.2%
0.24
TAXUS
1.06 [0.68-1.64]
8.3%
9.8%
0.81
BMS
1.34 [0.83-2.18]
7.9%
9.2%
0.24
1.0 3.0 0 2.0 T+ Worse T+ Better
Figure 4. Event rates, hazard ratios (HRs), and 95% confidence intervals (CIs) comparing patients taking with those not taking thienopyridines at 1 year, with p values adjusted for differences in baseline covariates. Top graph: outcomes at 2 years; bottom graph: outcomes at 5 years. N/A ⫽ not applicable.
Figure 5. Timing of ST in relation to thienopyridine use in the 12 patients in whom ST developed after Taxus stent implantation. See text for details. * Patients not taking thienopyridines at 1-year follow-up but in whom the exact dates of discontinuation were unknown.
relation between thienopyridine use at 1 year or later and the prevention of incipient ST. These data are consistent with and extend the observations made by Airoldi et al,6 who reported that in patients with DES, the median time from clopidogrel termination to ST was 13.5 days when clopidogrel was prematurely discontinued before 6 months, compared with 90 days when it was stopped after the first 6 months.
1021
Moreover, a similar proportion of T⫹1yr and T⫺1yr patients developed ST from 2 to 5 years in the present study, and as a result, the correlation between thienopyridine use at 1 year and subsequent ST after Taxus stent implantation was less apparent by 5 years than at 2 years. Importantly, no significant differences were apparent in either the 2- or 5-year rates of death, composite death or MI, and composite death, MI, or ST in patients with Taxus stents and BMS taking compared with those not taking thienopyridines at 1 year. Despite the trend toward reduced ST at 2 years in patients with Taxus stents in the T⫹1yr group, the unadjusted rate of death tended to be paradoxically greater at 2 years in the T⫹1yr group compared with the T⫺1yr patients (p ⫽ 0.10) and then nonsignificantly lower in T⫹1yr compared with T⫺1yr patients at 5 years, for BMS and Taxus stents. This favorable very late trend may reflect a positive benefit of extended clopidogrel administration in patients with established cardiovascular disease,4,14 independent of stent type or even coronary intervention, although this hypothesis remains to be definitively proved in a randomized trial, and the likelihood of a chance finding cannot be excluded. Finally, in contradistinction to the Duke landmark study,5 there were no significant interactions between thienopyridine use status at 1 year and stent type (Taxus vs BMS) on the subsequent occurrence of death, MI, ST or their composite occurrence either at 2 or at 5 years. The performance of a landmark analysis within the framework of a prospective, randomized trial overcomes many but not all of the drawbacks inherent in earlier analyses from retrospective observational registry studies. Specifically, randomization and the double-blind nature of the present study design eliminate bias by study personnel and adjudication committees on the basis of stent type and ensure complete monitoring of all study end points in all patients. Despite these advantages, as in previous such studies, the duration of thienopyridine use was uncontrolled and nonrandomized, and prolonged clopidogrel administration was not surprisingly used more frequently in higher risk patients (those with diabetes) with longer lesions requiring more stents. Although multivariate analysis was used in an attempt to correct for such measured confounders, we were unable to adjust for unmeasured or more subtle factors that might affect thienopyridine use duration, such as bleeding propensity or the requirement for noncardiac surgery, the performance of which was not routinely collected. Second, although more detailed information regarding dual-antiplatelet therapy was collected in the trials of Taxus stents than is typically available in uncontrolled observational registries, the exact dates of thienopyridine use and reasons for discontinuation were not prospectively assessed, nor was thienopyridine use status at exactly 6 months measured, precluding analysis of the potential utility of clopidogrel from 6 to 12 months. Third, there was a nontrivial rate of therapy crossover between the T⫹1yr and T⫺1yr groups during follow-up, a phenomenon that to our knowledge has not previously been described, which makes it more likely that such an analysis will be most accurate when applied to adverse events occurring shortly after the landmark date. Fourth, information regarding bleeding events was not collected after hospital discharge, an important consider-
1022
The American Journal of Cardiology (www.AJConline.org)
ation when considering the risk/benefit ratio of prolonged clopidogrel use, because recent studies have clearly established that hemorrhagic complications from medications may otherwise offset the life-prolonging benefits of reduced ischemia.15–17 Finally, the Taxus trials excluded many categories of “real-world” patients, including those receiving stents for multiple lesions, bifurcations, left main coronary arter or saphenous vein graft disease, in-stent restenosis and acute MI; as such, the results of the present investigation apply only to stent implantation in single de novo lesions in native coronary arteries in patients with mostly stable ischemic heart disease. Follow-up for the most complex patients in the study, those enrolled in the TAXUS-V trial, was available only to 3 years. Similarly, although one of the largest studies of its kind, ST and death occurred relatively infrequently in the present study, even out to 5 years; a larger study may have elicited a smaller magnitude of benefit (or risk) of prolonged thienopyridine use than was possible in the present investigation. 1. King SB III, Smith SC Jr, Hirshfeld JW Jr, Jacobs AK, Morrison DA, Williams DO, Feldman TE, Kern MJ, O’Neill WW, Schaff HV, et al. 2007 focused update of the ACC/AHA/SCAI 2005 guideline update for percutaneous coronary intervention. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation 2008;117:261–295. 2. Clopidogrel in Unstable Angina to Prevent Recurrent Ischemic Events Trial Investigators. Effects of clopidogrel in addition to aspirin in patients with acute coronary syndromes without ST-segment elevation. N Engl J Med 2001;345:494 –502. 3. Steinhubl SR, Berger PB, Mann JT III, Fry ET, DeLago A, Wilmer C, Topol EJ; CREDO Investigators. Early and sustained dual oral antiplatelet therapy following percutaneous coronary intervention: a randomized controlled trial. JAMA 2002;288:2411–2420. 4. Bhatt DL, Fox KA, Hacke W, Berger PB, Black HR, Boden WE, Cacoub P, Cohen EA, Creager MA, Easton JD, et al; CHARISMA Investigators. Clopidogrel and aspirin versus aspirin alone for the prevention of atherothrombotic events. N Engl J Med 2006;354:1706 – 1717. 5. Eisenstein EL, Anstrom KJ, Kong DF, Shaw LK, Tuttle RH, Mark DB, Kramer JM, Harrington RA, Matchar DB, Kandzari DE, et al. Clopidogrel use and long-term clinical outcomes after drug-eluting stent implantation. JAMA 2007;297:159 –168. 6. Airoldi F, Colombo A, Morici N, Latib A, Cosgrave J, Buellesfeld L, Bonizzoni E, Carlino M, Gerckens U, Godino C, et al. Incidence and
7.
8.
9.
10.
11. 12.
13.
14.
15.
16.
17.
predictors of drug-eluting stent thrombosis during and after discontinuation of thienopyridine treatment. Circulation 2007;116: 745–754. Colombo A, Drzewiecki J, Banning A, Grube E, Hauptmann K, Silber S, Dudek D, Fort S, Schiele F, Zmudka K, et al; TAXUS II Study Group. Randomized study to assess the effectiveness of slow- and moderate-release polymer-based paclitaxel-eluting stents for coronary artery lesions. Circulation 2003;108:788 –794. Stone GW, Ellis SG, Cox DA, Hermiller J, O’Shaughnessy C, Mann JT, Turco M, Caputo R, Bergin P, Greenberg J, et al; TAXUS-IV Investigators. A polymer-based paclitaxel-eluting stent in patients with coronary artery disease. N Engl J Med 2004;350:221–231. Stone GW, Ellis SG, Cannon L, Mann JT, Greenberg JD, Spriggs D, O’Shaughnessy CD, DeMaio S, Hall P, Popma JJ, et al; TAXUS V Investigators. Comparison of a polymer-based paclitaxel-eluting stent with a bare metal stent in patients with complex coronary artery disease: a randomized controlled trial. JAMA 2005;294:1215–1223. Stone GW, Moses JW, Ellis SG, Schofer J, Dawkins KD, Morice MC, Colombo A, Schampaert E, Grube E, Kirtane AJ, et al. Safety and efficacy of sirolimus- and paclitaxel-eluting coronary stents. N Engl J Med 2007;356:998 –1008. Ong AT, McFadden EP, Regar E, de Jaegere PP, van Domburg RT, Serruys PW. Late angiographic stent thrombosis (LAST) events with drug-eluting stents. J Am Coll Cardiol 2005;45:2088 –2092. Jeremias A, Sylvia B, Bridges J, Kirtane AJ, Bigelow B, Pinto DS, Ho KK, Cohen DJ, Garcia LA, Cutlip DE, Carrozza JP Jr. Stent thrombosis after successful sirolimus-eluting stent implantation. Circulation 2004;109:1930 –1932. Cutlip DE, Windecker S, Mehran R, Boam A, Cohen DJ, van Es GA, Steg PG, Morel MA, Mauri L, Vranckx P, et al; Academic Research Consortium. Clinical end points in coronary stent trials: a case for standardized definitions. Circulation 2007;115:2344 –2351. Bhatt DL, Flather MD, Hacke W, Berger PB, Black HR, Boden WE, Cacoub P, Cohen EA, Creager MA, Easton JD, et al; CHARISMA Investigators. Patients with prior myocardial infarction, stroke, or symptomatic peripheral arterial disease in the CHARISMA trial. J Am Coll Cardiol 2007;49:1982–1988. Feit F, Voeltz MD, Attubato MJ, Lincoff AM, Chew DP, Bittl JA, Topol EJ, Manoukian SV. Predictors and impact of major hemorrhage on mortality following percutaneous coronary intervention: an analysis of the REPLACE-2 trial. Am J Cardiol 2007;100:1364 –1369. Manoukian SV, Feit F, Mehran R, Voeltz MD, Ebrahimi R, Hamon M, Dangas GD, Lincoff AM, White HD, Moses JW, et al. Impact of major bleeding on 30-day mortality and clinical outcomes in patients with acute coronary syndromes: an analysis from the ACUITY trial. J Am Coll Cardiol 2007;49:1362–1368. Wiviott SD, Braunwald E, McCabe CH, Montalescot G, Ruzyllo W, Gottlieb S, Neumann FJ, Ardissino D, De Servi S, Murphy SA, et al; TRITON-TIMI 38 Investigators. Prasugrel versus clopidogrel in patients with acute coronary syndromes. N Engl J Med 2007;357:2001– 2015.