Author + information
- Received December 28, 2007
- Revision received February 28, 2008
- Accepted March 15, 2008
- Published online June 1, 2008.
- Mehdi H. Shishehbor, DO, MPH,
- Reza Amini, MD,
- Leonardo P.J. Oliveria, MD,
- Inder M. Singh, MD, MS,
- Peter Kelly, MD,
- Deepak L. Bhatt, MD,
- Samir R. Kapadia, MD,
- Stephen G. Ellis, MD,
- Patrick L. Whitlow, MD and
- Sorin J. Brener, MD⁎ ()
- ↵⁎Reprint requests and correspondence
: Dr. Sorin J. Brener, Associate Professor of Medicine, Director, Cardiac Catheterization Laboratory and Interventional Cardiology, NY Methodist Hospital, 506 Sixth Street, Brooklyn, New York 11215.
Objectives We sought to examine the clinical outcomes of patients treated with drug-eluting stents (DES) compared with bare-metal stents (BMS) during primary angioplasty for ST-segment elevation myocardial infarction (STEMI).
Background Pathophysiologic studies suggest that the use of DES for STEMI may be associated with stent thrombosis and increased clinical events. However, although short-term data exist, long-term follow-up is lacking.
Methods Patients who presented with STEMI from January 2002 to May 2007 to our institution were included. In addition to multivariable adjusted analysis, propensity analysis for stent choice was performed. The primary end point was the composite of death or target lesion revascularization (TLR).
Results Of the 804 patients, 699 underwent stenting and met our study criteria. There were 152 composite events over a median follow-up of 1.7 years. In a multivariable Cox model, DES use was associated with a trend toward lower death or TLR compared with BMS (adjusted hazard ratio [HR] 0.72, 95% confidence interval [CI], 0.50 to 1.02, p = 0.06). However, this was mainly due to lower TLR (adjusted HR 0.60, 95% CI, 0.36 to 0.98, p = 0.043). Similarly, DES was associated with a trend toward lower death or TLR compared with BMS in the propensity-matched patients (adjusted HR 0.65, 95% CI 0.42 to 1.00, p = 0.05). This was mainly due to lower TLR in the DES patients (adjusted HR 0.52, 95% CI 0.28 to 0.96, p = 0.04).
Conclusions Both DES and BMS are effective in the setting of STEMI; however, DES is associated with lower TLR without an increase in all-cause mortality.
Currently 42% of patients who present with ST-segment elevation myocardial infarction (STEMI) are treated with primary percutaneous coronary intervention (PCI) and an additional 19% undergo rescue PCI (1). Drug-eluting stents (DES) became available in the U.S. in March 2003 and were initially used in 70% to 80% of patients undergoing PCI including those who presented with STEMI (2). However, the initial enthusiasm was tempered by data on late-stent thrombosis, which were first reported in October 2004 (3–5). Furthermore, plaque rupture leading to STEMI is accompanied by a large thrombus burden and may increase the risk for stent thromosis (6).
To date, there are 5 published randomized controlled trials (RCTs) that have evaluated the use of DES versus bare-metal stents (BMS) is the setting of STEMI (7–11). However, these studies were underpowered to detect small differences in major adverse cardiac events and lacked long-term follow-up (12). Therefore, we sought to examine the long-term outcomes of DES versus BMS for treating patients with STEMI in a large cohort of patients at a single referral center.
From January 1, 2002, onward, all patients presenting with STEMI to the Cleveland Clinic cardiac catheterization laboratory were included in a prospectively collected ongoing registry. Concomitantly, all patients undergoing PCI were included in an ongoing registry for which baseline characteristics, angiographic data, and medications are collected at the time of coronary intervention. These 2 databases were merged for this analysis. The decision to choose a particular stent was completely based on the operators' preference. Patients who presented with STEMI but did not undergo stenting (because of medical management alone or balloon angioplasty alone) were excluded. The Institutional Review Board of the Cleveland Clinic waived requirements for informed consent for the institutional PCI and STEMI registries.
Data on procedural characteristics and device use were captured prospectively. Location of the culprit lesion was defined as proximal left anterior descending, middle or distal left anterior descending, right coronary artery, or left circumflex coronary artery.
Clinical end points
The primary end point was the composite of death from any cause or target lesion revascularization (TLR) (PCI or coronary artery bypass grafting). All patients were prospectively followed for adverse cardiovascular events through hospital record review and the Social Security Death Index.
Baseline and angiographic characteristics of DES and BMS patients were compared with the use of the Wilcoxon rank-sum test for continuous variables and the chi-square test for categorical variables. Differences in outcome between the 2 stent types were tested in unadjusted Kaplan-Meier curves and multivariable Cox proportional hazards analysis. Adjusted models accounted for baseline demographic features, angiographic variables, treatment assignment, door-to-balloon time, peak creatinine kinase levels, and ultrasensitive C-reactive protein (Table 1). Stepwise selection was used to select statistically significant confounders. However, results were not substantially different when this method was compared with a nonparsimonious model. Therefore, the results shown represent nonparsimonious models. The proportional hazards assumption was confirmed by testing the weighted Schoenfeld residuals and by plotting hazard ratio against time plots for selected variables.
To address potential biases, propensity analysis was performed. Nonparsimonious logistic regression models were used to generate a propensity score for stent choice. We considered all variables listed in Table 1 for this model, excluding procedural success. Each subject from the DES group was matched with a patient who received BMS using the derived propensity score. Subsequently, we performed an adjusted multivariable analysis that incorporated all baseline characteristics in addition to the propensity score for the propensity-matched patients. Matching is performed to the nearest fifth digit. All analyses were performed using SAS, version 9.1 (SAS Institute, Cary, North Carolina).
A total of 804 patients presented to the cardiac catheterization laboratory after a presumptive diagnosis of STEMI. Of these, 24 had normal coronaries, 46 received balloon angioplasty only, and 35 patients were lost to follow-up. Therefore, 699 patients met our study criteria. Baseline and target lesion characteristics according to stent type are presented in Table 1. In general, patient and procedural characteristics were similar between both groups. However, patients undergoing DES were more likely to be in Killip class IV. Additionally, patients undergoing DES implantation were more likely to be treated with beta-blockers and statins. The DES were less often used in saphenous vein grafts. The predominant procedural anticoagulation was based on unfractionated heparin and glycoprotein IIb/IIIa inhibitors (Table 1).
A cumulative of 152 events occurred during a mean follow-up of 1.7 years. In the unadjusted analysis DES use was significantly associated with lower composite end point of death or TLR (Fig. 1,Table 2). In multivariable adjusted Cox proportional hazard model, there was a trend for lower events in the DES group, which was mainly due to a reduction in TLR (Table 2). Importantly, the rate of death was similar in both groups.
To account for the differences that existed in the nonpropensity-matched patients, propensity analysis was performed. After generating a propensity score (c-statistic = 0.72), 240 of the 344 patients who received DES were matched with a patient who received BMS. No significant differences were noted in more than 30 baseline characteristics for the propensity-matched subjects (Table 3).
A total of 95 events occurred during a mean follow-up of 1.8 years. Both in the unadjusted and multivariable adjusted models, there was a trend toward lower rates of death or TLR with DES, which again was mainly driven by a reduction in TLR (Fig. 2,Table 2).
In this cohort of consecutive, unselected patients who presented with STEMI, PCI with DES was associated with a lower TLR rate and a trend toward lower rate of the composite end point of death or TLR. These results persisted despite multiple statistical adjustments, including propensity analysis. Importantly, there was no difference in the incidence of all-cause mortality between the 2 groups.
A number of concerns have led to lower use of DES in patients with STEMI (4,12,13), including the risk of late stent thrombosis, impaired endothelial function distal to stent, and excessive platelet activation (14). Mechanistically, these abnormalities may be related to delayed arterial healing of the necrotic core characterized by lack of complete re-endothelialization and persistence of fibrin within the ruptured plaque when compared with BMS (14). However, despite these concerns all RCTs to date show lower rate of restenosis with no significant difference for death or myocardial infarction (7–10).
Recently, Kastrati et al. (15) published a meta-analysis of 5 published and 3 unpublished randomized clinical trials in the setting of STEMI (15). The follow-up duration was 12 to 24 months among 2,786 patients. These authors showed similar composite end points between DES and BMS for STEMI, with lower TLR in patients that received DES. However, concerns have been raised about the lack of power and long-term follow-up in each of these individual trials (14). Additionally, these meta-analyses are from randomized clinical trials and not the “real-world” setting. Therefore, we attempted to address this question in 699 patients who underwent PCI in the setting of STEMI. There were a total of 152 composite events with 77 deaths, which gave us adequate power to assess differences between the 2 stent types. Additionally, propensity analysis and matching were performed and demonstrated consistent results in the matched cohort.
Our study had several limitations. This study was an observational one on prospectively collected data; therefore, it is prone to a number of biases that are present in such study design. To address some of these limitations, we conducted propensity analysis. However, propensity analysis can only account for variables that are measured. Therefore, there may remain biases that are not accounted for in this analysis such as presence of cancer. It is not clear why some patients received DES whereas other received BMS. By using propensity analysis, we adjusted for factors that are known to impact the decision by a surgeon regarding the use of DES versus BMS in a patient, such as multivessel coronary artery disease, lesion length, vessel diameter, and the presence of diabetes mellitus and severe co-morbidity, such as malignancy. Even after these adjustments, patients who received a DES had significantly lower rates of reintervention.
The present study is from a tertiary care center and may not be generalizable. However, our results are consistent with previous randomized clinical trials and observational studies that have examined safety and efficacy of DES versus BMS. Another limitation is lack of knowledge on duration of dual antiplatelet therapy used in the 2 groups of patients and the intensity and extent of concomitant medical therapy. Finally, we began capturing the occurrence of stent thrombosis in our registry in mid 2006, and thus we did not have accurate data on stent thrombosis or myocardial infarction.
Our results extend the conclusions of previous randomized clinical trials and observational studies and confirm the superiority of DES in decreasing the need for clinically relevant TLR without increasing the risk of death. The use of DES in STEMI in this analysis was not associated with increased mortality; however, the impact of DES on subacute thrombosis and myocardial infarction in this study cannot be assessed. Data from ongoing RCTs should provide further guidance about the optimal approach to stenting in AMI.
Dr. Shishehbor is supported in part by the Case Western Reserve University/Cleveland Clinic National Institutes of Health CTSA (1KL2RR024990).
- Abbreviations and Acronyms
- bare-metal stent(s)
- confidence interval
- drug-eluting stent(s)
- hazard ratio
- percutaneous coronary intervention
- randomized clinical trial
- ST-segment elevation myocardial infarction
- target lesion revascularization
- Received December 28, 2007.
- Revision received February 28, 2008.
- Accepted March 15, 2008.
- American College of Cardiology Foundation
- Eagle K.A.,
- Goodman S.G.,
- Avezum A.,
- Budaj A.,
- Sullivan C.M.,
- Lopez-Sendon J.
- Holmes D.R. Jr..,
- Kereiakes D.J.,
- Laskey W.K.,
- et al.
- Joner M.,
- Finn A.V.,
- Farb A.,
- et al.
- Menichelli M.,
- Parma A.,
- Pucci E.,
- et al.
- van der Hoeven B.L.,
- Liem S.S.,
- Jukema J.W.,
- et al.
- Suryapranata H.,
- De Luca G.,
- van't Hof A.W.,
- et al.
- Luscher T.F.,
- Steffel J.,
- Eberli F.R.,
- et al.
- Kastrati A.,
- Dibra A.,
- Spaulding C.,
- et al.