Author + information
- Received June 8, 2010
- Accepted June 11, 2010
- Published online September 1, 2010.
- José M. de la Torre Hernández, MD, PhD⁎,⁎ (, )
- Fernando Alfonso, MD, PhD†,
- Federico Gimeno, MD, PhD#,
- Jose A. Diarte, MD, PhD⁎⁎,
- Ramón Lopez-Palop, MD, PhD††,
- Armando Pérez de Prado, MD‡‡,
- Fernando Rivero, MD‡,
- Juan Sanchis, MD, PhD§§,
- Mariano Larman, MD∥∥,
- Jose F. Diaz, MD¶¶,
- Jaime Elizaga, MD, PhD§,
- Javier Martín Moreiras, MD##,
- Alfredo Gomez Jaume, MD, PhD⁎⁎⁎,
- José M. Hernández, MD, PhD†††,
- Josepa Mauri, MD, PhD§§§,
- Angel Sánchez Recalde, MD∥,
- Juan A. Bullones, MD‡‡‡,
- Jose R. Rumoroso, MD, PhD∥∥∥,
- Bruno García del Blanco, MD¶¶¶,
- Jose A. Baz, MD###,
- Francisco Bosa, MD, PhD⁎⁎⁎⁎,
- Javier Botas, MD, PhD††††,
- Felipe Hernández, MD¶,
- ESTROFA-2 Study Group
- ↵⁎Reprint requests and correspondence:
Dr. Jose M. de la Torre Hernández, Unidad de Hemodinámica y Cardiología Intervencionista, Hospital Universitario Marques de Valdecilla, Santander 39012, Spain
Objectives This study sought to evaluate second-generation drug-eluting stent (DES) thrombosis in clinical practice.
Background First-generation DES are associated with a significant incidence of late thrombosis. There is paucity of data regarding real practice late thrombosis incidence and predictors with second-generation DES, zotarolimus-eluting stent (ZES), and everolimus-eluting stents (EES).
Methods A prospective, large-scale, non-industry-linked multicenter registry was designed. Complete clinical-procedural data and systematic follow-up of all patients treated with these stents was reported in a dedicated registry supported by the Spanish Working Group on Interventional Cardiology.
Results From 2005 to 2008, 4,768 patients were included in 34 centers: 2,549 treated with ZES, and 2,219 with EES. The cumulative incidence of definite/probable thrombosis for ZES was 1.3% at 1 year and 1.7% at 2 years and for EES 1.4% at 1 year and 1.7% at 2 years (p = 0.8). The increment of definite thrombosis between the first and second year was 0.2% and 0.25%, respectively. In a propensity score analysis, the incidence remained very similar. Ejection fraction (adjusted hazard ratio [HR]: 0.97; 95% confidence interval [CI]: 0.95 to -0.99; p = 0.008), stent diameter (adjusted HR: 0.37; 95% CI: 0.17to 0.81; p = 0.01) and bifurcations (adjusted HR: 2.1; 95% CI: 1.14 to 3.7; p = 0.02) emerged as independent predictors of thrombosis. In the subgroup of patients with bifurcations, the use of ZES was independently associated with a higher thrombosis rate (adjusted HR: 4; 95% CI: 1.1 to 13; p = 0.03).
Conclusions In a real practice setting, the incidence of thrombosis at 2 years with ZES and EES was low and quite similar. The incidence of very late thrombosis resulted lower than was reported in registries of first-generation DES. In the subset of bifurcations, the use of ZES significantly increased the risk of thrombosis.
Drug-eluting stents (DES) have proved to be an outstanding improvement in the interventional cardiology field. Their known impact on restenosis rates drop in almost every clinical and lesional profile has expanded percutaneous revascularization horizons. Nevertheless, both first-generation DES, paclitaxel-eluting stents (PES) and sirolimus-eluting stents (SES), have been associated with a certain incidence of late thrombosis in trials that was greater than that observed with bare-metal stents (1). This incidence was higher in real practice registries than in randomized trials due to a frequent off-label use of these DES (2–6).
Thrombosis of first-generation DES may occur up to 4 years after implantation, with an increment after the first year of 0.4% to 0.6% per year. Second-generation DES, zotarolimus-eluting stents (ZES) and everolimus-eluting stents (EES), are already available. They have already been evaluated in different clinical trials (7–14) and have shown in most of them a similar if not better safety profile than first-generation DES (7,8,10–13), but worse results have been also reported (9,14). However, once again, large independent registries, with long-term follow-up, are needed in order to assess late thrombosis incidence in the real-world clinical practice.
The ESTROFA-2 (Estudio Español Sobre Trombosis de Stents Farmacoactivos de Segunda Generacion-2) registry is an independent, multicenter, and prospective registry run by a “research hospital network” designed to assess DES thrombosis incidence in real practice. A total of 34 centers all around Spain have been involved. Economic costs and maintenance of the Web-based electronic clinical research forum were held thanks to the support given by the Spanish Working Group on Interventional Cardiology. All coordination between the 34 participating centers, the adjudication process, and analyses have been performed centrally at the coordination center Hospital Universitario Marques de Valdecilla in Santander, Spain.
All consecutive patients treated at the participating hospitals with either an EES or a ZES, since their introduction in the Spanish market in 2005 to 2006 up to 2008, were prospectively entered in an electronic CRF. For the cases to be considered for final analysis, all lesions should have been treated with EES or with ZES, with no mix. Clinical, angiographic, and procedure features of every single case were filled in by the local investigators. If a thrombosis event was detected in the established follow-up periods, a wider detailed application form had to be completed. Data collection met all national requirements regarding “data protection law” as well as anonymity.
All thrombosis recalled in this registry were indexed following the well-known Academic Research Consortium definitions (15). Hyperlipidemia was defined as plasmatic cholesterol greater than 200 mg/dl or previous treatment with statins. Hypertension was defined as either systolic pressure greater than 140 mm Hg, or diastolic pressure above 90 mm Hg or previous treatment with antihypertensive therapy. Renal function impairment was considered for plasmatic creatinine above 1.5 mg/dl. Angiographic success was defined as TIMI (Thrombolysis In Myocardial Infarction) flow grade 3 along with residual stenosis below 25%.
Information was obtained from medical reports and hospital or interventional cardiology department registries. In those cases with no events reported in hospital files and no recent outpatient clinical visit, phone contact was required. All data regarding vital status, clinical events, as well as new interventional procedures were, therefore, specifically collected for the purpose of this investigation. Nonparticipant centers were contacted when necessary.
Continuous variables are presented as mean ± SD. Categorical variables are expressed as percentages. Continuous variables were compared with the Student t test if the data followed a normal distribution and with Wilcoxon tests if the data were skewed. Categorical variables were compared with the chi-squared test or the Fischer exact test, where indicated. The survival curves were constructed with the use of Kaplan-Meier method to describe the incidence of thrombosis over time, and log rank tests were applied to evaluate differences between groups.
Because patients were not randomized to EES or ZES, propensity scores (conditional probability) for receiving an EES rather than a ZES for each patient were calculated by using baseline covariates within a logistic regression model. A baseline model was defined and included the demographic, clinical, and procedural covariates listed in Table 1. Patients were then matched according to their individual propensity scores. The individual propensity score as well as type of stent were incorporated into Cox proportional hazard regression models as a covariate to calculate the propensity adjusted hazard ratio (HR).
Identification of predictors of stent thrombosis was done through a Cox proportional hazard regression model including those variables with p < 0.2 in the univariate analysis. The variable of antiplatelet therapy discontinuation was not included because that information was not available for all cases without stent thrombosis. Only the prescribed period of dual therapy could be included in this analysis. All probability values were 2-sided and values of p < 0.05 were considered statistically significant. The statistical package SPSS version 15.0 (SPSS, Inc., Chicago, Illinois) was used throughout.
A total of 4,768 patients were prospectively included in the registry; 2,549 of the patients were treated with ZES and 2,219 with EES. As a rule, dual antiplatelet therapy was prescribed for up to 12 months (only 6.4% of EES- and 10.6% of ZES-treated patients had dual antiplatelet therapy prescribed for 6 months). Clinical and procedural features are described in Table 1. The EES-treated patients were slightly younger, more frequently diabetics, and had a greater rate of previous interventions. ZES were more often used during an acute coronary syndrome. As for the lesional profile, more restenotic lesions were treated with EES. Dual antiplatelet treatment recommended duration was slightly shorter for the ZES group.
Thrombosis incidence (definite; definite and probable; and definite, probable, and possible) curves are displayed in Figure 1. Cumulative stent thrombosis incidence for both kinds of stents is presented in Table 2. No statistically significant difference was found between the 2 groups. The increase in definite stent thrombosis incidence observed from the first to the second year was 0.2% to 0.25%.
Given the differences observed in the basal characteristics of the 2 groups, a propensity score matching was performed. Eventually, 1,300 patients from each group were selected and compared. Clinical and procedural features are detailed in Table 3. After adjusting, baseline characteristics were identical in the 2 groups. Cumulative stent thrombosis incidence curves for matched groups are presented in Figure 2 and incidences in Table 4. No difference in the incidence of stent thrombosis was found after the matching process.
Definite and probable thrombosis predictors found in univariate and multivariate analysis are presented in Table 5. Only bifurcations, ejection fraction, and stent diameter were found to be independent thrombosis predictors.
Inappropriate withdrawal of antithrombotic drugs was confirmed in up to 15% of definite thrombosis and 7% of probable thrombosis. However, dual antiplatelet treatment prescription (12 months vs. 6 months) was not found to be a predictor of stent thrombosis.
As previously described, there was a subgroup of patients with dual therapy indicated during 6 months. These patients were more frequently found in the ZES group. We therefore decided to analyze stent thrombosis incidence only in patients with antiplatelet dual treatment indicated for 12 months. As presented in Figure 3A, no difference was noted either.
Differences in stent thrombosis incidence between patients with prescription for 6 months and those for 12 months are shown in Figure 3B. This incidence tended to be greater in the 6-month treated group although the difference was concentrated in the early period.
Bifurcation-treated patients developed more cases of thrombosis (Fig. 4A). The use of complex techniques (2 stents) tended to be higher among those cases with thrombosis (27.5% vs. 13.8%; p = 0.06). The use of the double stent treatment for bifurcations was comparable in both stent groups (12.8% with EES and 14.9% with ZES; p = 0.4). As is shown in Figure 4B, a greater stent thrombosis rate was demonstrated in ZES-treated bifurcations. Bifurcation was the only lesion subset in which statistically significant differences in stent thrombosis were found between stent types (Fig. 5) in matched groups. This difference was observed both in the early and late period. Age (adjusted HR: 1.06, 95% CI: 1.01 to 1.12; p = 0.04), ejection fraction (adjusted HR: 0.96, 95% confidence interval [CI]: 0.92 to 0.99; p = 0.02), and ZES use (adjusted HR: 4, 95% CI: 1.1 to 13; p = 0.03) were found to be independent thrombosis predictors in bifurcation-treated patients. Based on our results, the double stent technique was not an independent predictor.
First-generation DES have proved to diminish the need for new revascularization in almost every clinical and lesional setting leading to a spread in its use, even in off-label indications. Low late thrombosis rates obtained from clinical trials have supported safety issues (1). Nevertheless, real practice registries frequently include off-label indications and have consistently revealed a higher late thrombosis rate than bare-metal stents, with a 1.2% to 1.7% first-year definite thrombosis rate and a 0.4% to 0.6% late thrombosis rate per year later on (2–6). Mechanisms involved in this process could be lack of long-term endothelialization, positive arterial remodeling leading to late acquired incomplete stent apposition, and persistent local hypersensibility reactions possibly due to residual polymer (16).
Second-generation DES are made of new polymers and drugs. Pre-clinical studies show a promising faster endothelialization with EES, compared with SES or PES (17). In addition, the ultrasonographic analysis of patients with EES in the SPIRIT III (Clinical Evaluation of the Xience V Everolimus Eluting Coronary Stent System in the Treatment of Patients with de novo Native Coronary Artery Lesions) trials (11) showed only a 1.1% of late acquired incomplete stent apposition. The ultrasonographic studies performed in subgroups of patients in ENDEAVOR II (Randomized Controlled Trial to Evaluate the Safety and Efficacy of the Medtronic AVE ABT-578 Eluting Driver Coronary Stent in De Novo Native Coronary Artery Lesions) and III (Randomized Controlled Trial of the Medtronic Endeavor Drug-Eluting Coronary Stent System Versus the Cypher Sirolimus-Eluting Coronary Stent System in De Novo Native Coronary Artery Lesions) trials with ZES yielded a complete and uniform intimal coverage without late acquired incomplete stent apposition (18). These latter results have been confirmed using optical coherence tomography in the ODESSA (Optical Coherence Tomography for DES Safety) trial (19). Despite of these findings, these stents have already been evaluated in different clinical trials showing disparate results regarding safety comparison with first-generation DES (7–14). Nevertheless, the clinical trials carried out to date have not been powerful enough to point out differences in thrombosis rates between stent models.
However, EES and ZES significantly differ in their performance. Late lumen loss with EES is very low, around 0.15 mm (10,11), whereas it is greater with ZES, 0.6 mm (7,8). Hence, this different efficacy profile may potentially translate into safety issues.
Thrombosis of ZES in trials
Combined analysis of ENDEAVOR I, II, and III (Multicenter Evaluation of ABT-578 Elution From a Phosphorylcholine-Coated Stent) trials with a total of 1,317 ZES treated patients, obtained an absence of thrombosis after 14 days of deployment and up to 2 years, being the final thrombosis rate of only 0.3% (20). However, these results diverge from those found in ENDEAVOR IV (Randomized, Controlled Trial of the Medtronic Endeavor Drug-Eluting Coronary Stent System Versus the Taxus Paclitaxel-Eluting Coronary Stent System in De Novo Native Coronary Artery Lesions) (9) and SORT OUT III (Comparison of Zotarolimus-Eluting Stents and Sirolimus-Eluting Stents in Patients With Coronary Artery Disease) (14) trials. In the first study (9), involving 773 ZES- and 775 PES-treated patients, 1-year definite plus probable thrombosis rate was 0.1% with PES and 1% with ZES. In the SORT OUT III (14) study, definite thrombosis incidence at 18 months was 1.2% in 1,162 ZES-treated patients and 0.6% in the 1,170 SES-treated patients.
Thrombosis of EES in trials
In the SPIRIT III study, the 669 patients treated with EES showed a global stent thrombosis incidence similar to that of 332 PES-treated patients. However, a trend toward lower late thrombosis rate was noted in EES-treated patients after the first year and in those in whom thienopyridine withdrawal took place 6 month after the procedure (12).
In the pooled SPIRIT II and III analysis (21), late definite plus probable thrombosis rate at 1 year was 0.8% for both the 892 EES-treated patients and the 410 PES-treated patients. The larger SPIRIT IV study (22), not published yet, has shown in 2,459 EES-treated patients a definite plus probable thrombosis rate at 1 year of 0.29% compared with 1.06% in PE-treated patients.
The COMPARE (Second-Generation Everolimus-Eluting and Paclitaxel-Eluting Stents in Real-Life Practice) trial (13), which aimed to compare EES versus PES, found 1-year thrombosis rate to be significantly lower with EES. This difference was due to a lower subacute thrombosis rate, whereas late thrombosis was similar in both groups (13).
Thrombosis in second-generation DES registries
After these inconclusive results and taking into account the already mentioned limitations of clinical trials, it becomes necessary to evaluate the real practice outcomes including long-term follow-up to assess whether second-generation DES thrombosis rate remains similar to that of first-generation DES or if they differ. Moreover, it is essential to find out whether late thrombosis rate is comparable in the 2 currently available second-generation DES, taking into account their different behaviors (lower late lumen loss with EES).
To assess these questions, the Rotterdam team has reported its real practice experience with EES up to 6 months. Six-month outcomes in 649 EES-treated patients were compared with cohorts of patients treated with either bare-metal stents, PES, or SES (23). They found no statistically significant differences in thrombosis rates; of note, no thromboses were detected in EES-treated patients after 1 month. However, no conclusions regarding the safety of the stent profile could be done as the number of patients was relatively small. In addition, clinical follow-up was limited at 6 months.
The E-Five Registry (a prospective registry) involved 8,314 ZES-treated patients in 118 centers. In this registry, with a follow-up limited to 1 year, definite/probable stent thrombosis rate was 1.1% (24). However, it was unclear whether patient inclusion was consecutive. In fact, the number of lesions treated per patient was only 1.2 (in our registry this was 1.45, which correlates better with numbers obtained in nonselective global series).
Our ESTROFA-2 registry aimed to evaluate definite plus probable and possible stent thrombosis rates for up to 2 years in an ample “all comers” group of consecutive cases treated with EES and ZES, with no exclusion criteria.
Clinical and lesional features revealed a standard real-practice profile. Being a registry, ZES and EES groups differ in some characteristics: ZES were used more often in acute coronary syndromes (maybe due to a pre-conceived higher safety profile), whereas EES, were more frequently used in diabetics and restenotic lesions (perhaps as the result of their lower late lumen loss).
Thrombosis rates following the different definitions were low and comparable between groups. This lack of difference was confirmed also after matching propensity score analysis. Definite thrombosis rate at 2 years was 1% with an increment between the first and second year of 0.2% to 0.25%.
Comparing these results with those obtained in first-generation DES registries is intricate as a great amount of noncontrollable factors could be playing a role. However, definite thrombosis rate at 1 year in first-generation DES registries was around 1.2% to 1.7% with a thrombosis increment after the first year of 0.4% to 0.6% (2–6). Different issues could help to explain this difference apart from those intrinsic to DES. A better selection of patients for DES took place after 2006 because of the “late thrombosis DES crisis,” but certainly clinical profile remains comparable and DES penetration (at least in Spain) has been even higher in ESTROFA-2 inclusion period (55% to 60%) than in ESTROFA inclusion period (35% to 40%).
Therapeutic compliance and dual antiplatelet treatment extent could also have increased. Indeed, mean time of dual antiplatelet treatment in ESTROFA registry was 8 months compared with almost 12 months in ESTROFA-2 (5). In ESTROFA, up to 31.6% of definite thromboses were related to nontreatment compliance compared with 15% in the present registry. This statement could be linked to a better compliance but also to a greater susceptibility of first-generation DES to antiplatelet treatment withdrawal. The lack of specific evaluation of antiplatelet treatment compliance in these registries remains a caveat that should be addressed in future, dedicated registries.
Regarding duration of dual antiplatelet treatment, most patients in this study were instructed to continue it during 12 months. There was a small group with only 6-month prescriptions. These patients were more frequently found in the ZES group, most probably due to an understood high safety profile. This group showed a nonsignificant trend toward more thrombosis but, paradoxically, it was concentrated in the early stage. These findings suggest that this group could be conformed of patients with worse biological conditions and more comorbidities that could lead to a greater bleeding risk. Anyway, this variable did not come out as a predictor in the analysis.
Finally, thrombosis predictors detected in our analysis—lower ejection fraction, lower stent diameter, and bifurcations—had already been documented in other studies (2,3). Bifurcations have been analyzed in depth, as it was the only subset in which statistically significant differences regarding thrombosis rates were observed in among second-generation DES. Thrombosis rates in ZES-treated patients with bifurcation lesions were greater in both early and late stages. Use of ZES was an independent predictor of thrombosis in the subgroup of patients with bifurcation lesions. The reason for this could be explained by means of various factors such as greater late lumen loss (favoring a restenosis leading to a prothrombotic condition) and different platform design. Anyway, the reason remains unknown and these findings warrant further research.
Being a registry implies observational and nonrandomized findings. Stents were assigned by the operator, and even with matching propensity score analysis, there remain underlying variables that could play a role in the results. That is why the comparison of both groups has to be done by taking into account this limitation. Moreover, the different outcome of the bifurcational subgroups in both stent groups should be considered cautiously.
Using these results to obtain conclusions regarding second-generation DES safety as opposed to first-generation DES is even tougher. This does not nullify the fact that the second-generation DES safety profile in the settings in which they are used nowadays is better than the profile revealed by first-generation DES in the previous years.
Antiplatelet treatment compliance was not possible to analyze. This would have required a very frequent and exhaustive monitoring of every single patient, which was beyond the potential of the registry. The antiplatelet therapy cessation would probably be a major determinant of thrombosis, though we do not know how many patients without thrombosis had discontinuation of therapy in the study period.
Patients with previous percutaneous coronary intervention procedures (whatever the stent type) were included. This fact could be a problem in the adjudication of a probable thrombosis event; however, every single case of thrombosis was carefully analyzed, and we did not have such a conflictive case. Last but not least, the earlier introduction of ZES in the Spanish market implies a longer follow-up of these patients than for EES-treated patients.
In a real practice setting with frequent off-label indications, the incidence of thrombosis at 2 years with ZES and EES was low and quite similar. The incidence increase of definite thrombosis between the first and second years was 0.2% to 0.25%, which is lower than the rates previously reported for SES and PES (0.4% to 0.6%). This finding suggests that the incidence of very late thrombosis could be lower for second-generation DES. In the subset of bifurcations, the use of ZES could be associated with a higher risk of thrombosis, though this finding warrants further research.
The authors thank Dr. Tamara Garcia Camarero for her assistance with the statistical analysis.
Appendix List of Other Participating Investigators and Corresponding Centers
Juan C. Fernández Guerrero (Complejo Hospitalario de Jaén, Jaén, Spain), José Moreu (Hospital Virgen de la Salud, Toledo, Spain), Rafael Melgares (Hospital Virgen de las Nieves, Ganada, Spain), José M. Vázquez (Hospital de la Coruña, La Coruña, Spain), Adolfo Rincon (Hospital Dr. Peset, Valencia, Spain), Jose D. Cascon (Hospital Santa Maria del Rosell, Cartagena, Spain), Antonio Serra (Hospital del Mar, Barcelona, Spain), Iñigo Lozano (Hospital Central de Asturias, Oviedo, Spain), Ramiro Trillo (Hospital Clínico de Santiago, Santiago de Compostela, Spain), Jesús M. Jiménez Mazuecos (Hospital General de Albacete, Albacete), Eduardo Pinar (Hospital Virgen de la Arrixaca, Murcia, Spain).
- Abbreviations and Acronyms
- drug-eluting stent(s)
- everolimus-eluting stent(s)
- paclitaxel-eluting stent(s)
- sirolimus-eluting stent(s)
- zotarolimus-eluting stent(s)
- Received June 8, 2010.
- Accepted June 11, 2010.
- American College of Cardiology Foundation
- Kuchulakanti P.K.,
- Chu W.W.,
- Torguson R.,
- et al.
- De la Torre-Hernández J.M.,
- Alfonso F.,
- Hernández F.,
- et al.,
- ESTROFA Study Group
- Wenaweser P.,
- Daemen J.,
- Zwahlen M.,
- et al.
- Fajadet J.,
- Wijns W.,
- Laarman G.J.,
- et al.,
- ENDEAVOR II Investigators
- Kandzari D.E.,
- Leon M.B.,
- Popma J.J.,
- et al.,
- ENDEAVOR III Investigators
- Leon M.B.,
- Mauri L.,
- Popma J.J.,
- et al.,
- ENDEAVOR IV Investigators
- Stone G.W.,
- Midei M.,
- Newman W.,
- et al.,
- SPIRIT III Investigators
- Cutlip D.E.,
- Windecker S.,
- Mehran R.,
- et al.,
- Academic Research Consortium
- Jaffe R.,
- Strauss B.H.
- Joner M.,
- Nakazawa G.,
- Finn A.V.,
- et al.
- Guagliumi G.
- Stone G.W.
- Onuma Y.,
- Kukreja N.,
- Piazza N.,
- et al.,
- Interventional Cardiologists of the Thoraxcenter (2000 to 2007)
- Lotan C.,
- Meredith I.T.,
- Mauri L.,
- Liu M.,
- Rothman M.T.,
- E-Five Investigators