Author + information
- Received August 26, 2004
- Revision received September 23, 2004
- Accepted September 28, 2004
- Published online March 15, 2005.
- Andrew T.L. Ong, MBBS, FRACP,
- Angela Hoye, MBChB, MRCP,
- Jiro Aoki, MD,
- Carlos A.G. van Mieghem, MD,
- Gaston A. Rodriguez Granillo, MD,
- Karel Sonnenschein,
- Evelyn Regar, MD, PhD,
- Eugene P. McFadden, MBChB, MD, FRCPI, FACC,
- Georgios Sianos, MD, PhD,
- Willem J. van der Giessen, MD, PhD,
- Peter P.T. de Jaegere, MD, PhD,
- Pim de Feyter, MD, PhD, FACC,
- Ron T. van Domburg, PhD and
- Patrick W. Serruys, MD, PhD, FACC* ()
- ↵*Reprint requests and correspondence:
Dr. Patrick W. Serruys, Thoraxcenter, Bd-406, Dr. Molewaterplein 40, 3015-GD Rotterdam, Netherlands
Objectives We sought to determine the real-world incidence of angiographically confirmed and possible stent thrombosis (ST) in an unrestricted population during the first 30 days after bare-metal stent (BMS), sirolimus-eluting stent (SES), and paclitaxel-eluting stent (PES) implantation.
Background Current data on ST in drug-eluting stents (DES) have come from randomized trials with strict entry criteria, which limits their generalizability to daily practice.
Methods The study population comprised three sequential cohorts of 506 consecutive patients with BMS, 1,017 consecutive patients with SES, and 989 consecutive patients treated with PES.
Results In the first 30 days after stent implantation, 6 BMS (1.2%, 95% confidence interval [CI] 0.5% to 2.6%; p = 0.9), 10 SES (1.0%, 95% CI 0.5% to 1.8%), and 10 PES (1.0%, 95% CI 0.6% to 1.9%) patients developed angiographically proven ST. Multiple potential risk factors were identified in most patients with ST. Bifurcation stenting in the setting of acute myocardial infarction was an independent risk factor for angiographic ST in the entire population (odds ratio [OR] 12.9, 95% CI 4.7 to 35.8, p < 0.001). In patients with DES who had angiographic ST, 30-day mortality was 15%, whereas another 60% suffered a nonfatal myocardial infarction; no further deaths occurred during six months of follow-up. Including possible cases, 7 BMS (1.4%, 95% CI 0.7% to 2.8%), 15 SES (1.5%, 95% CI 0.9% to 2.4%), and 16 PES (1.6%, 95% CI 1.0% to 2.6%) patients had ST.
Conclusions The unrestricted use of SES or PES is associated with ST rates in the range expected for BMS. Stent thrombosis was associated with a high morbidity and mortality. Bifurcation stenting, when performed in patients with acute myocardial infarction, was associated with an increased risk of ST.
Drug-eluting stents (DES) reduce clinical events related to restenosis. Concerns have been raised regarding the incidence of stent thrombosis (ST) with the unrestricted use of these stents. Data from the bare-metal stent (BMS) era report a high morbidity and mortality with ST (1,2). Evidence for ST in DES has come from randomized controlled trials with strict entry criteria for the treatment of single lesions, limiting conclusions that are applicable to the real-world setting (3–6). Other information has come from electronic registries with inherent biases that preclude generalization of the findings. A single-center registry recently reported its results with sirolimus-eluting stents (SES) (7). The aim of this present study is to describe the incidence of ST (both angiographically proven and including possible cases) in three consecutive populations while analyzing the unrestricted use of a control BMS group, SES, and paclitaxel-eluting stents (PES).
Study design and patient population
Since April 2002, SES (Cypher; Cordis Corp., Miami Lakes, Florida, a Johnson & Johnson Company) have been the stents of choice for all percutaneous coronary interventions irrespective of their clinical presentation or clinical outcome (8). In the first quarter of 2003, PES (Taxus; Boston Scientific Corp., Natick, Massachusetts) replaced SES as the default stent.
This present study comprises three sequential cohorts: a control group of the last 506 consecutive patients treated with BMS before April 2002; 1,017 consecutive patients with SES treated between April 2002 and February 2003; and 989 consecutive patients with PES treated between February 2003 and December 2003.
Procedure and antiplatelet management
All interventions were performed according to current standard guidelines, and the final interventional strategy including periprocedural glycoprotein IIb/IIIa and intravascular ultrasound use, was left to the discretion of the operator. Patients were pretreated with aspirin and a loading dose of 300 mg of clopidogrel. After their procedure, all patients were prescribed a lifelong aspirin regimen. Clopidogrel was prescribed for at least one month in the BMS group, for at least three months in the SES group (8), and for at least six months in the PES group.
As part of the national health system, our institution as a tertiary referral center is the only interventional facility within our catchment area. The survival status of our patients at one and six months after discharge was obtained from the Municipal Civil Registries. Details of all repeat interventions (surgical and percutaneous) were collected prospectively during follow-up. Referring physicians and institutions were contacted whenever necessary for additional information. This protocol was approved by the Hospital Ethics Committee, and written, informed consent was obtained from every patient.
Stent thrombosis was considered to have occurred when confirmed angiographically: either Thrombolysis In Myocardial Infarction (TIMI) flow grade 0 or 1 or the presence of flow-limiting thrombus (TIMI flow grade 1 or 2) occurring in an acute (within 24 h of stent implantation) or subacute (between 1 and 30 days) time period after stent implantation (9). In addition, a clinical definition of “possible stent thrombosis” was used for patients who within the first 30 days experienced sudden death, who suffered a fatal out-of-hospital cardiac arrest, or who suffered a myocardial infarction (MI) that was not clearly attributable to another coronary lesion and who did not undergo repeat angiography. All deaths and MIs were reviewed independently by two interventional cardiologists (A.O., E.Mc.F) for “possible stent thrombosis.”
Categorical variables were compared using the Fisher exact test and continuous variables with the Student ttest or one-way analysis of variance where appropriate. Univariate and forward stepwise (entry criteria of 0.05 and exit criteria of 0.10) multivariate logistic regression analysis were performed to identify characteristics or variables independently associated with stent thrombosis. From the univariate analysis, the following baseline, clinical, angiographic and procedural variables were entered into the multivariate model: bifurcation stenting, diabetes, smallest stent diameter, multilesion stenting, and acute myocardial infarction (AMI) as the indication. All probability values are two-sided, and statistical significance was set at the 0.05 level. A cumulative event graph consisting of patients with angiographic stent thrombosis was generated plotting the proportion of patients with stent thrombosis (Y-axis) against time (X-axis) stratified by stent type. Incidences of stent thrombosis are reported as a percentage with associated 95% confidence intervals (CIs).
Baseline and procedural characteristics
The patients in our cohort were at high risk, with unstable angina or AMI being the indication in more than one-half of the cases (Table 1).Multivessel disease was present in more than one-half of the population. One-third of the population had a previous AMI, whereas one-quarter had previous coronary interventions. Glycoprotein IIb/IIIa use was lower in the SES and PES groups compared with the BMS group.
Angiographic ST was documented in 26 of 2,512 patients (Table 2).Six cases occurred in the BMS group (1.2%, 95% CI 0.5% to 2.6%), 10 cases occurred in the SES group (1.0%, 95% CI 0.5% to 1.8%), and 10 cases occurred in the PES group (1.0%, 95% CI 0.6% to 1.9%). The first two SES patients with ST have been reported previously (10). Most stent thromboses occurred in the first 11 days, regardless of stent type, with a mean time to event of 5.8 ± 5.4 days (Fig. 1).
In the BMS population, there were two acute stent thromboses and four subacute stent thromboses. Among the six patients, ST presented as AMI in five patients. None died during the six months of follow-up (Table 2). In the combined group of SES and PES (2,006 patients), there were 2 cases of acute ST and 18 cases of subacute ST (Fig. 1). A detailed description of these patients is given in Table 3.Analysis via intravascular ultrasonography was performed in four patients. In most patients, at least one recognized risk factor for ST (i.e., long stented length, use of small stents, use of multiple stents, and residual dissection after stent implantation) was present. Importantly, 2 of the 20 patients had not taken clopidogrel.
Mortality and morbidity
Overall, 20 of 26 patients (77%) re-presented with an AMI, whereas the other 6 re-presented with angina pectoris (Table 2). Of these 26 patients, 3 (Patients #12, #18, and #20 from Table 3—all in the DES population) died at days 11, 5, and 3, respectively. Two patients died during reintervention from intractable ventricular fibrillation, whereas the third underwent emergency surgery after a suboptimal reintervention and could not be weaned from bypass. The incidence of death at 30 days was 12%, whereas another 65% suffered a nonfatal MI. Among the survivors of ST, there were no further deaths in the six months after reintervention.
Thirty-day survival data was complete for 98% of patients (Table 4).There were 12 patients who were judged with “possible stent thrombosis,” of which 9 died and 3 had nonfatal MIs. Of the nine deaths, four were out-of-hospital sudden deaths, three occurred in hospital with ventricular tachycardia as the initiating preterminal rhythm, and two had ST-segment elevation and died before they could undergo reangiography. Among those with MIs, one patient developed a postprocedural enzyme leak, and another developed ventricular fibrillation requiring multiple cardioversions the day after the procedure. Repeat coronary angiography six months later demonstrated occluded stents in both of these patients; whereas a third underwent coronary angiography 14 days after stent implantation because of an increase in cardiac enzyme levels, which demonstrated an in-stent filling defect which was treated with abciximab, and subsequently underwent repeat percutaneous coronary intervention two weeks later. Including the suspected cases, the combined incidence of angiographic and possible ST was 1.4% (95% CI 0.7% to 2.8%) in the BMS control group, 1.5% (95% CI 0.9% to 2.4%) in the SES group, and 1.6% (95% CI 1.0% to 2.6%) in the PES group. In the combined total of 38 documented and possible ST, there were 12 deaths (32%) and 20 nonfatal MIs (53%) in the first 30 days.
By univariate analysis, bifurcation stenting was the only significant factor (p = 0.01). Multivariate analysis was performed with the following covariates based on their significance on univariate analysis as well as their potential clinical impact: diabetes (p = 0.07), smallest stent diameter (p = 0.13), multilesion stenting (p = 0.17), AMI as the indication (p = 0.3), and bifurcation stenting. By multivariate analysis, bifurcation stenting was the only independent predictor of ST (odds ratio [OR] 3.0, 95% CI 1.3 to 6.8, p < 0.01). When the interaction of bifurcation stenting by AMI was entered as a covariate, it was highly significant (OR 12.9, 95% CI 4.7 to 35.8, p < 0.001), and bifurcation stenting as a covariate was no longer significant.
The main findings in this study can be summarized as follows: 1) the incidence of angiographic ST in an unselected, complex DES population was low (∼1.0%), within the same range as the corresponding BMS population and concordant with previously published results from the BMS era; 2) the inclusion of possible ST increases the overall incidence of ST to ∼1.5%; 3) angiographically proven ST was associated with a high mortality and morbidity; 4) patients who developed ST often had multiple high-risk features, regardless of stent type; and 5) the association of bifurcation stenting for AMI was a highly significant independent risk factor for ST.
The availability of DES as the default stent at our institution has allowed us to analyze this new technology in an unrestricted population (8), a population that would have comprised any BMS population in the pre-DES era. Therefore, this availability allows us to analyze incidences in an “all-comers” population because patients were enrolled irrespective of clinical presentation or outcome. In this population sample, angiographic ST rates in the first 30 days for both DES, i.e., SES and PES, occurred within the range as that reported in the BMS era (1,2,11,12).
The angiographic definition used is the most accurate for diagnosis but may underestimate the true incidence of ST because some patients who have a presumed ST may die before receiving medical attention. Conversely, the use of major adverse cardiac events (i.e., death and MI in addition to the angiographic findings) to define ST overestimates the true incidence because not all patients who die suddenly or suffer a MI do so because of ST (13). This consideration is important in our heterogeneous unrestricted population with multivessel disease, previous MI, and previous revascularization. Furthermore, not all patients who die will undergo autopsy studies to determine the cause of death. To attenuate this overestimation and to provide an accurate figure, we have adjudicated all deaths and noncatheterized, nonfatal MIs within the first 30 days in the three groups and included them with the angiographically proven patients to provide an overall incidence for each group.
The incidences of ST for both groups of DES are within the range reported in the larger randomized clinical trials of DES (3–6) despite longer total stent length, multivessel treatment, and a heterogeneous population (Table 5).This incidence complements information already available from the randomized trials regarding the safety of these new devices.
Angiographic ST was associated with a high mortality and morbidity in our study. Within the DES population, 15 patients (75%) experienced a MI as their diagnosis at the second presentation, and 3 (15%) died during the reintervention procedure. The inclusion of possible ST patients increased the mortality to 32%. Given the small number of events, the fact that no deaths occurred in the BMS group was most likely due to chance. These results are in concordance with the results of a large BMS registry (2).
Previous studies have demonstrated that residual dissection (1,11), long stents (1), small final lumen diameter (1), and use of multiple stents (2) are risk factors for the development of ST. In our series, multiple risk factors were identified in most patients who developed ST. Patients with ST had more multiple lesions treated, smaller minimum stent diameters, and longer stent lengths compared with those without ST; however, these factors were not significant on univariate analysis. What did emerge and which has not been previously reported is that patients undergoing bifurcation stenting had a higher incidence of ST compared with those without bifurcation stenting. A recent study on bifurcations reported a 3.5% incidence of ST, which is higher than the overall incidence in this population (14).
Although stent implantation for AMI was not significant on univariate analysis, the interaction of AMI and bifurcation stenting when entered as a covariate for ST on multivariate analysis emerged as a highly significant independent predictor, and bifurcation stenting as a covariate was no longer significant. This result confirms a clinical suspicion in our department regarding the increased risk of ST in patients treated with bifurcation stenting in the setting of AMI.
Mechanical reasons that predispose to ST can be modified by interventional technique. Optimizing stent placement including, if necessary, intravascular ultrasound-guided postdilation, kissing balloon postdilation with bifurcation stenting, and careful inspection for residual dissection after stent implantation, may further reduce the incidence of ST.
Pharmacologic reasons for ST, i.e., inadequate antiplatelet therapy, are patient-specific factors. Recent research literature has focused on “resistance” to either aspirin (15) or to clopidogrel (16). Currently, most laboratories do not routinely test for antiplatelet resistance. In our series, two patients who had not taken their prescribed clopidogrel after the procedure developed ST.
This report covers ST occurring in the first 30 days after stent implantation only, during which all patients received dual antiplatelet therapy. The duration of clopidogrel therapy differed among the three groups; in part, it reflects uncertainty with regards to re-endothelialization after DES implantation. Late ST has been reported to occur with BMS (17) and with DES (18), including a reported fatality (19) after clopidogrel discontinuation. At this stage, the incidence of late ST in the DES era is unknown, and further studies are required to clarify this potential late complication.
Comment on sample size and statistical comparisons
Because ST occurs at a low incidence (∼1.0 to 1.5%), a small sample size may underestimate or overestimate the true incidence. In a previously published report from our institution, we reported an angiographic incidence of 0.4% in 508 patients (8). In the present study we extended the population to incorporate the entire period of DES used to date at our institution (n = 2,006) to allow a more accurate analysis of the true incidence of ST in the DES population. Despite having 2,512 patients, the low and small/negligible absolute difference in incidence precludes formal statistical comparisons of ST rates among the three groups because it lacks sufficient statistical power. To achieve adequate power would require sample sizes in the order of >100,000 patients. To date, this study is the largest series of patients reported on in the DES era.
These single-center registry data complement available randomized data, as they reflect the results of unrestricted DES use.
Despite having a more complex cohort with high-risk inclusion criteria, longer stent lengths, and more complex procedural features, the incidence of ST with DES are in the same range as the BMS population observed in our present study. They also are in agreement with previously reported data by others from the BMS era and with those results reported on in the earlier randomized DES trials. Furthermore, the two groups of DES, i.e., SES and PES, share an incidence of ∼1.0% to 1.5%. Stent thrombosis is associated with a high morbidity and mortality.
As extensively documented in previous reports with BMS, mechanical reasons were observed to be frequent associations for ST with DES. In this study, bifurcation stenting in the setting of AMI was a highly significant independent predictor for angiographic ST.
The authors would like to thank Pedro A. Lemos, MD, PhD, Chourmouzios A. Arampatzis, MD, PhD, and Pieter C. Smits, MD, PhD, for their contribution to this work.
Supported by Erasmus Medical Center, Rotterdam, and by an unrestricted institutional grant from Cordis a Johnson and Johnson Company and Boston Scientific Corporation.
- Abbreviations and acronyms
- acute myocardial infarction
- bare-metal stents
- confidence interval
- drug-eluting stents
- myocardial infarction
- odds ratio
- paclitaxel-eluting stents
- sirolimus-eluting stents
- stent thrombosis
- Thrombolysis In Myocardial Infarction
- Received August 26, 2004.
- Revision received September 23, 2004.
- Accepted September 28, 2004.
- American College of Cardiology Foundation
- Cutlip D.E.,
- Baim D.S.,
- Ho K.K.,
- et al.
- Orford J.L.,
- Lennon R.,
- Melby S.,
- et al.
- Schampaert E.,
- Cohen E.A.,
- Schluter M.,
- et al.
- Jeremias A.,
- Sylvia B.,
- Bridges J.,
- et al.
- Lemos P.A.,
- Serruys P.W.,
- van Domburg R.T.,
- et al.
- Lemos P.A.,
- Lee C.H.,
- Degertekin M.,
- et al.
- Cheneau E.,
- Leborgne L.,
- Mintz G.S.,
- et al.
- Honda Y.,
- Fitzgerald P.J.
- Colombo A.,
- Moses J.W.,
- Morice M.C.,
- et al.
- Gum P.A.,
- Kottke-Marchant K.,
- Welsh P.A.,
- White J.,
- Topol E.J.
- Lau W.C.,
- Gurbel P.A.,
- Watkins P.B.,
- et al.
- Colombo A.,
- Drzewiecki J.,
- Banning A.,
- et al.
- Virmani R.,
- Guagliumi G.,
- Farb A.,
- et al.