Author + information
- Received April 4, 2007
- Revision received May 21, 2007
- Accepted June 3, 2007
- Published online July 31, 2007.
- Lisette Okkels Jensen, MD, PhD⁎,⁎ (, )
- Michael Maeng, MD, PhD†,
- Anne Kaltoft, MD, PhD†,
- Per Thayssen, MD, DMSci⁎,
- Hans Henrik Tilsted Hansen, MD‡,
- Morten Bottcher, MD, PhD†,
- Jens Flensted Lassen, MD, PhD†,
- Lars Romer Krussel, MD, DMSci†,
- Klaus Rasmussen, MD, DMSci‡,§,
- Knud Noerregaard Hansen, MD⁎,
- Lars Pedersen, MSc∥,
- Soeren Paaske Johnsen, MD, PhD§∥,
- Henrik Toft Soerensen, MD, DMSci, PhD∥,¶ and
- Leif Thuesen, MD, DMSci†
- ↵⁎Reprint requests and correspondence:
Dr. Lisette Okkels Jensen, Department of Cardiology, Odense University Hospital, Sdr. Boulevard 29, 5000 Odense C, Denmark.
Objectives The aim of the study was to examine outcomes subsequent to implantation of drug-eluting stents (DES) and bare-metal stents (BMS).
Background Use of DES might be associated with increased risk of stent thrombosis (ST), myocardial infarction (MI), and death.
Methods From January 2002 through June 2005, data from all percutaneous coronary interventions in western Denmark were prospectively recorded in the Western Denmark Heart Registry; 12,395 consecutive patients (17,152 lesions) treated with stent implantation were followed for 15 months. Data on death and MI were ascertained from the national databases. The Academic Research Consortium definition of ST was used.
Results The DES were implanted in 3,548 patients (5,422 lesions) and BMS were implanted in 8,847 patients (11,730 lesions). Definite, probable, or possible ST was found in 190 (2.15%) patients in the BMS group and in 64 (1.80%) patients in the DES. The risk of definite ST was similar in the 2 groups (DES: 0.65%; BMS: 0.61%). Very late definite ST (between 12 and 15 months after implantation) occurred more frequently in patients receiving DES (hazard ratio [HR] 10.93, 95% confidence interval [CI] 1.27 to 93.76). Also, the risk of MI between 12 and 15 months after implantation was higher in the DES group (HR 4.00, 95% CI 2.06 to 7.79). Mortality was similar in the 2 groups. Target lesion revascularization was reduced by 43% in patients treated with DES (HR 0.57, 95% CI 0.48 to 0.67).
Conclusions The minor risk of ST and MI within 15 months after implantation of DES seems unlikely to outweigh the benefit of these stents.
Coronary drug-eluting stent (DES) implantation is increasingly used for coronary revascularization. In patients with coronary artery disease, randomized controlled trials have consistently linked DES with reduction of neointimal hyperplasia, decreased risk of restenosis, and less frequent need for repeated revascularization, compared with bare-metal stents (BMS) (1–4). The beneficial effects of DES were achieved without raising any safety concerns initially (5). Recent studies, including a post hoc analysis, a meta-analysis of randomized trials of DES, and a large observational study, reported increased risks of stent thrombosis (ST), myocardial infarction (MI), and death associated with the use of this type of stent (6–9). In December 2006, the U.S. Food and Drug Administration (FDA) issued a statement concluding that, on the basis of available data, the small increase in the risk of ST with DES, compared with BMS, did not confer an increased risk of death or MI (10). According to the FDA, the lack of an observed association could be explained either by insufficient numbers of patients in currently available studies or by deaths or MIs being offset by a reduction in the rate of events associated with in-stent restenosis and with additional revascularization procedures. In contrast, 2 recently randomized trials demonstrated no increased risk with DES compared with BMS in terms of long-term survival and survival free of MI and ST (11,12).
We compared rates of ST, MI, death, and target lesion revascularization (TLR) after implantation of DES and BMS, with population-based medical databases in Denmark.
Setting and design
We performed this 15-month follow-up study with western Denmark’s health care databases, covering the region’s entire population (approximately 3 million inhabitants; 55% of the Danish population). All patients were followed for 15 months. The Danish National Health Service provides universal tax-supported health care, guaranteeing free access to general practitioners and hospitals, while refunding variable proportions of prescription medication costs. Our data came from the sources described in the following text.
Danish Civil Registration System
This system has kept electronic records on gender, date of birth, residence, date of emigration, and changes in vital status since 1968 (13). Information on vital status is updated daily. The records include a unique 10-digit civil registration number, which is assigned at birth and is used in all public registries, allowing accurate record linkage.
National Registry of Causes of Deaths
This registry has collected data on dates and causes of death since 1973.
National Patient Registry
This registry, established in 1977, collects data on all non-psychiatric hospital stays at Danish hospitals, including dates of admission and discharge, surgical procedure(s), and up to 20 discharge diagnoses assigned by the treating physician and coded according to the International Classification of Diseases (8th revision [ICD-8] until the end of 1993 and 10th revision [ICD-10] thereafter) (14).
Western Denmark Heart Registry
This regional registry, established in 1999, collects detailed patient and procedure data for all interventions carried out in western Denmark’s 3 coronary intervention centers (Odense University Hospital, Aarhus University Hospital [Skejby], and Aarhus University Hospital [Aalborg]) (15,16).
Patients and procedures
With the Western Denmark Heart Registry, we identified all percutaneous coronary interventions (PCIs) performed from January 1, 2002, through June 30, 2005. We included the patients’ first PCI occurring during the study period (the index procedure) and were further restricted to patients who received either a BMS or a DES. Patients with a PCI before the study period were not excluded. Those treated by balloon angioplasty without stent implantation or by a combination of BMS and DES were excluded.
Percutaneous coronary intervention was performed according to the standard practices of the participating centers. The post-intervention antiplatelet regimen included lifelong acetylsalicylic acid (75 to 150 mg once daily) and clopidogrel with a loading dose of 300 mg followed by maintenance with 75 mg daily. The recommended duration of clopidogrel treatment was 3 to 12 months until November 2002 and 12 months thereafter. In patients with definite ST, we reviewed medical records from the hospital admission to verify the use of antiplatelet therapy.
The primary end point was ST (definite, probable, or possible). Secondary end points were all-cause mortality, cardiac death, MI, and TLR. All end points were assessed within 15 months of the date of PCI and were ascertained from the Western Denmark Heart Registry, the National Patient Registry, and the National Registry of Causes of Deaths. All relevant records were reviewed by a specialist committee, which adjudicated the end points. In the case of ST, the medical records were retrieved and the catheterization films were reviewed.
We defined ST using the Academic Research Consortium (ARC) definition, with a modification for probable ST (17).
Angiographic confirmation of ST and at least 1 of the following signs present within 48 h: new onset of ischemic symptoms at rest, new electrocardiographic changes suggestive of acute ischemia, or typical rise and fall in cardiac biomarkers.
Any unexplained death within the first 30 days after intracoronary stenting.
Any unexplained death occurring from 30 days after intracoronary stenting until the end of the follow-up period.
Stent thrombosis was further characterized as acute (0 h to <24 h), subacute (≥1 day to <30 days), late (≥30 days to <1 year), and very late (≥1 year to 15 months).
Admissions and readmissions for MI (ICD-10 codes I21 to I21.9) were ascertained from the National Patient Registry (13). New MIs were defined as admissions with a MI occurring >28 days after the index PCI (18).
We ascertained deaths from the Danish Civil Registration System and then reviewed original death certificates obtained from the National Registry of Causes of Deaths. Deaths were classified according to their underlying causes, as recorded on the death certificate.
Target lesion revascularization within 15 months after the index stent implantation was ascertained from the National Heart Registry and defined as a repeat PCI of the index lesion or coronary artery bypass grafting (CABG). Repeat PCI or CABG of study patients in hospitals outside the study region virtually did not take place.
We retrieved data from the Western Denmark Heart Registry on a wide range of potential predictors for subsequent cardiovascular disease, including characteristics of the patients, procedures, and lesions. For each patient, we obtained data on all discharge diagnoses from the National Patient Registry from 1977 until the date of implantation. We then computed the Charlson comorbidity index (19), which covered 19 major disease categories, including diabetes mellitus, heart failure, cerebrovascular diseases, and cancer. The index value was a weighted summary of the diagnoses, with each weight calculated on the basis of 1-year mortality associated with each disease in the original Charlson dataset (19).
Data on all essential patient and procedure characteristics were >95% complete, and ascertainment of end points (ST, death, and MI) was 100% complete.
We counted end point events that occurred during the follow-up period and compared their rates between the cohort of patients with DES and the cohort with BMS. Follow-up began on the date of the index PCI procedure. In analyses with ST as the outcome, follow-up continued until date of definite ST, death, or emigration, or after 15 months of follow-up, whichever came first. In analyses with death as the outcome, follow-up continued until date of death or emigration, or after 15 months of follow-up. In analyses with MI as the outcome, follow-up continued until date of MI, death, or emigration, or after 15 months of follow-up.
We constructed Kaplan-Meier curves for patients and lesions treated with DES or BMS. We used the life-table method to compute the 15-month cumulative incidence for each end point (proportion of the population at risk with the outcome of interest).
Cox proportional hazards regression was used to compute hazard ratios (HRs) as estimates for each end point. Because the hazards were not proportional throughout the follow-up period, we computed the estimates of HRs within separate time windows, where the proportionality assumption held. The HRs in these subanalyses reflected the risk among patients alive and at risk for the specific end point at the start of each time period (e.g., at 30 days or 1 year of follow-up). We controlled for age, gender, PCI indication, and procedure time in all regression analyses. To improve precision of the risk estimates, we used the change-in-estimate method, which entailed retaining variables that changed HR estimates for an outcome by more than 10% (20). Thus, in the lesion-specific analyses (ST and TLR) we also adjusted for diabetes mellitus, stent length, and reference vessel size. Distributions of continuous variables in the 2 groups were compared with either the 2-sample ttest or the Mann-Whitney test according to whether data followed the normal distribution. Distributions of categorical variables were compared with the chi-square test. We analyzed data with the SAS software version 9.13 (SAS Institute Inc., Cary, North Carolina).
The study encompassed 12,395 consecutive patients with 17,152 lesions. Of these, 8,847 patients with 11,730 lesions were treated with BMS and 3,548 patients with 5,422 lesions received DES (Cypher n = 3,429 [63.2%], and Taxus n = 1,993 [36.8%]). The patients’ mean age was 63.5 ± 11.4 years; 15.9% of the patients were older than 75 years, and 11.5% suffered from diabetes mellitus. The indications for PCI were ST-segment elevation MI (29.9%), non–ST-segment elevation MI/unstable angina (30.4%), stable angina (37.0%), and other (2.7%). Baseline patient and procedure characteristics (Table 1)and lesion characteristics (Table 2)differed substantially between the groups treated with BMS versus DES.
Overall rates of definite ST were similar for patients treated with BMS (n = 71; cumulative incidence, 0.65%) and DES (n = 35; cumulative incidence, 0.61%), p = 0.78 (chi-square test). Rates of acute, subacute, and late definite ST were also similar in the 2 groups (Fig. 1A).Very late definite ST occurred in 1 patient (cumulative incidence = 0.01%) in the BMS group and in 5 patients (cumulative incidence, 0.09%) in the DES group (adjusted HR 10.93; 95% confidence interval [CI] 1.27 to 93.76; p = 0.03) (Table 3).None of the 106 cases of definite ST occurred in saphenous vein grafts.
Definite, Probable, or Possible ST
Definite, probable, or possible ST was found in 190 (cumulative incidence, 2.15%) patients treated with BMS and in 64 (cumulative incidence, 1.80%) patients treated with DES (HR 0.83; 95% CI 0.63 to 1.12; p = 0.20) (Fig. 1B). After adjusting for covariates, the risk of ST (definite, probable, or possible) did not differ between the 2 groups (adjusted HR 0.91; 95% CI 0.67 to 1.24; p = 0.57) (Table 4).Late possible ST (i.e., beyond 1 year of follow-up) occurred in 11 patients treated with BMS and 1 patient treated with DES (adjusted HR 0.70; 95% CI 0.15 to 3.56; p = 0.65).
Predictors of Definite ST
ST-segment elevation MI (adjusted HR 3.80; 95% CI, 1.97 to 5.46; p < 0.001), stent length (adjusted HR 1.03; 95% CI 1.01 to 1.05; p < 0.001), and procedure time (adjusted HR 1.01; 95% CI 1.00 to 1.02; p < 0.001) increased the risk of definite ST.
Definite ST and Antiplatelet Therapy
Among the 106 patients who developed definite ST, 85.8% (n = 91) received treatment with dual antiplatelet therapy (aspirin and clopidogrel) at the time of the thrombotic event. In the 6 patients with very late ST, 2 patients were receiving dual antiplatelet therapy, 3 patients were treated with aspirin only, and 1 patient had discontinued both aspirin and clopidogrel after 12 months treatment.
All-cause 15-month mortality was 5.7% and was higher among patients with BMS than among patients with DES (6.2% vs. 4.4%; p < 0.001) (chi-square test) (Fig. 1C). After adjustment for covariates, this disparity disappeared (adjusted HR 0.90; 95% CI 0.75 to 1.09; p = 0.29). All-cause mortality 12 to 15 months after initial stenting was nearly similar in the 2 groups (adjusted HR 1.11; 95% CI 0.62 to 1.99; p = 0.22). Overall 15-month cardiac mortality was 3.4%, with higher cardiac mortality among BMS patients than among DES patients (3.8% vs. 2.4%) (HR 0.63; 95% CI 0.50 to 0.80; p = 0.002). After adjustment, risk of cardiac death no longer differed in the 2 groups (adjusted HR 0.88; 95% CI 0.68 to 1.13; p = 0.31). The non-cardiac 15-month cumulative mortality was 2.2% and was similar among BMS (2.3%) and DES (1.9%) stent patients (HR 0.81; 95% CI 0.61 to 1.06; p = 0.12). Covariate adjustment did not change this result (adjusted HR 0.95; 95% CI 0.71 to 1.27; p = 0.71).
The 15-month cumulative incidence of MI was similar in the 2 stent groups (3.2% [DES] vs. 3.0% [BMS]; p = 0.65). However, from 12 to 15 months after the index stent implantation, MI occurred in 0.64% of patients treated with DES versus 0.22% of patients treated with BMS (HR 3.13; 95% CI 1.69 to 5.79; p < 0.001) (Fig. 1D). The association became stronger after adjustment for covariates (adjusted HR 4.00; 95% CI 2.06 to 7.79; p < 0.0001).
Target lesion revascularization within 15 months occurred less frequently in DES patients than in BMS patients (cumulative incidence, 4.6% vs. 7.1%; p < 0.0001). After adjustment for age, gender, clinical presentation, diabetes mellitus, stent length, and reference vessel size, the risk reduction was 43% (adjusted HR 0.57; 95% CI 0.48 to 0.67; p = 0 < 0.001). Diabetes mellitus (adjusted HR 1.31; 95% CI 1.09 to 1.58; p = 0.005), ST-segment elevation MI (adjusted HR 1.41; 95% CI 1.19 to 1.66; p < 0.0001), and stent length (adjusted HR 1.02; 95% CI 1.01 to 1.03; p < 0.001) increased the risk of TLR within 15 months, whereas size of the reference vessel was inversely related to the risk of TLR (adjusted HR 0.69; 95% CI 0.61 to 0.79; p < 0.001).
In this population-based study we found that BMS and DES treatments were associated with low rates of cardiac events within 15 months after their implantation. The overall rates of ST, MI, and death were similar among the patients treated with the 2 stent types. Furthermore, use of DES reduced the risk of clinically necessary TLR by 43%.
The incidence rate of definite ST in our study was similar to that reported in a meta-analysis (5) of 10 randomized studies with 9 to 12 months of follow-up that compared DES and BMS. It was also comparable to the results of the e-Cypher Registry (21). Our findings, however, differed from those of the BASKET-LATE (BAsel Stent Kosten Effektivitäts Trial—LAte Thrombotic Events) study (6), in which there was a higher overall rate of ST. Furthermore, patients treated with DES experienced an earlier and increased occurrence of ST compared with patients with BMS than found in our study. This might be related to the longer duration of dual antiplatelet therapy in our study. This might also explain the lower risk of ST among DES patients in our study as compared with a patient-level meta-analysis presented at a December 2006 FDA hearing (10) and a recently published registry study on sirolimus- and paclitaxel-eluting stents (22).
Data from the Swedish Coronary Angiography and Angioplasty Registry showed rates of mortality and MI similar to those in our study (7). These registry data suggested that rates of death and MI were similar for patients with DES and BMS during the first 6 months after stent implantation. After 6 months, both mortality and MI rates increased in the DES group. Our finding of a minor excess of MI after 12 months in the DES group thus accords with the Swedish findings. However, we found no difference between the 2 groups in cardiac or non-cardiac mortality during 15 months of follow-up.
For most of our study period, the recommended duration of clopidogrel treatment was 12 months. This potentially lowered rates of ST, both in patients treated with DES and those treated with BMS. Thus, our data support the recent FDA recommendation for prolongation of antiplatelet therapy (10). Several studies have linked the risk of ST to discontinuation of clopidogrel treatment among patients with DES (23,24). It is notable that in our study the number of definite stent thromboses and acute MI events in patients with DES increased significantly after 12 months, even though the absolute event count was small.
We identified ST-segment elevation MI as a predictor of definite ST. As found in 2 recent randomized trials (25,26), ST-segment elevation MI did not affect the magnitude of the relative risk significantly when patients with the 2 stent types were compared.
Observational studies should be scrutinized for potential biases. Our relative risk estimates are based on population-based data, largely ruling out referral and diagnostic biases. Still, our study has limitations, including potential unmeasured confounding and bias due to lack of blinding. However, adjustment for the most important predictors of cardiovascular events reduced the likelihood of severe residual confounding by indication.
Our data were collected over a 3-year period, during which the proportion of DES increased from 0% to 75% of all stent implantations. To minimize study bias during this time of transition, we provided 15 months of follow-up data for all patients. Also, we reported recommended 12-month duration for antiplatelet combination treatment, although the actual duration of treatment with dual anti-platelet agents was only available for patients with definite ST. Another concern is that we were unable to group patients with MI by area of infarction and therefore could not include MI in the ARC definition of probable ST. With the ARC definition of probable and possible ST, any unexplained death is categorized as ST. Although this situation is attributable to some acute events, most likely thrombosis, it is not invariably associated with all cases of unexplained death. Finally, the follow-up time of 15 months, although longer than in most other studies, might still be inadequate to fully quantify possible risks associated with use of DES.
During the 15 months after stent implantation, overall rates of ST, MI, and death were low and of similar order of magnitude in patients receiving BMS or DES and dual antiplatelet therapy for 12 months. After 12 months, there was a small but significant excess of definite stent thromboses in the group with DES. Our findings lend support to the recommendation of at least 12 months of dual antiplatelet therapy for patients treated with DES. Our study also suggests that the low risk of ST and MI within 15 months after implantation of DES does not outweigh the benefit of these stents, materializing in a marked reduction in the risk of clinically driven TLR.
- Abbreviations and Acronyms
- bare-metal stent(s)
- confidence interval
- drug-eluting stent(s)
- Food and Drug Administration
- hazard ratio
- myocardial infarction
- percutaneous coronary intervention
- stent thrombosis
- target lesion revascularization
- Received April 4, 2007.
- Revision received May 21, 2007.
- Accepted June 3, 2007.
- American College of Cardiology Foundation
- Moreno R.,
- Fernandez C.,
- Hernandez R.,
- et al.
- Pfisterer M.,
- Brunner-La Rocca H.P.,
- Buser P.T.,
- et al.
- Nordmann A.J.,
- Briel M.,
- Bucher H.C.
- ↵Update to FDA Statement on Coronary Drug-Eluting Stents. Available at: http://www.fda.gov/cdrh/news/010407.html. Accessed December 2006.
- Frank L.
- Maeng M.,
- Jensen L.O.,
- Rasmussen K.,
- et al.
- Jensen L.O.,
- Maeng M.,
- Thayssen P.,
- et al.
- ↵New standard stent-thrombosis definition yields comparable event rates for DES and bare-metal stents. Available at: http://www.medscape.com/viewarticle/546575. Accessed October 2006.
- Tunstall-Pedoe H.,
- Kuulasmaa K.,
- Amouyel P.,
- Arveiler D.,
- Rajakangas A.M.,
- Pajak A.
- Urban P.,
- Gershlick A.H.,
- Guagliumi G.,
- et al.