Several concerns have emerged regarding the higher risk for stent thrombosis (ST) after drug-eluting stent (DES) implantation, especially in the setting of ST-segment elevation myocardial infarction (STEMI). Few data have been reported so far in patients with diabetes mellitus, which is associated with high rates of target vessel revascularization after bare-metal stent (BMS) implantation but also higher rates of ST after DES implantation. Therefore, the aim of this study was to perform a meta-analysis of individual patients’ data to evaluate the long-term safety and effectiveness of DES compared with BMS in patients with diabetes who undergo primary percutaneous coronary intervention for STEMI. Published reports were scanned by formal searches of electronic databases (MEDLINE and CENTRAL). All completed randomized trials of DES for STEMI were examined. No language restrictions were enforced. Individual patients’ data were obtained from 11 of 13 trials, including a total of 972 patients with diabetes (616 [63.4%] randomized to DES and 356 [36.6%] to BMS). At long-term follow-up (median 1,095 days, interquartile range 1,087 to 1,460), DES significantly reduced the occurrence of target vessel revascularization (hazard ratio 0.42, 95% confidence interval 0.29 to 0.59, p <0.0001), without any significant difference in terms of mortality, late reinfarction, and ST (>1 year) with DES. In conclusion, this meta-analysis, based on individual patients’ data from 11 randomized trials, showed that among patients with diabetes with STEMIs who undergo primary percutaneous coronary intervention, sirolimus-eluting stents and paclitaxel-eluting stents, compared with BMS, are associated with a significant reduction in target vessel revascularization at long-term follow-up, without any apparent concern in terms of mortality, despite the trend toward higher rates of reinfarction and ST.
Primary percutaneous coronary intervention (PCI) has been shown to improve survival compared with thrombolysis in the treatment of ST-segment elevation myocardial infarction (STEMI). Stent implantation has further improved outcomes by reducing the occurrence of restenosis compared with balloon angioplasty in selected patients with STEMIs. However, the results seem unsatisfactory in unselected patients. Several studies and meta-analyses have been performed on the use of drug-eluting stents (DES) in the setting of STEMI. Although clear benefits have emerged in terms of stent restenosis, there were concerns regarding a higher risk for mortality and stent thrombosis. Few data have been reported so far in patients with diabetes mellitus. In fact, these patients are associated with high rates of target vessel revascularization (TVR) after bare-metal stent (BMS) implantation but also a higher rate of stent thrombosis after DES implantation. The aim of the Drug-Eluting Stents in Primary Angioplasty (DESERT) cooperation was to evaluate outcomes of DES implantation in patients with STEMIs using individual patients’ data from randomized trials. Therefore, the aim of this study was to evaluate the safety and efficacy of DES in patients with diabetes who undergo primary PCI for STEMI.
Methods
We obtained the results of all completed randomized trials comparing DES with BMS in primary PCI for STEMI. The published research was scanned using formal searches of electronic databases (MEDLINE and CENTRAL) and the scientific session abstracts in Circulation , the Journal of the American College of Cardiology , the European Heart Journal , and the American Journal of Cardiology from January 1990 to June 2011. Furthermore, oral presentations and/or expert slide presentations were included (searched on the Transcatheter Cardiovascular Therapeutics [ http://www.tctmd.com ], EuroPCR [ http://www.europcr.com ], American College of Cardiology [ http://www.acc.org ], American Heart Association [ http://www.aha.org ], and European Society of Cardiology [ http://ww.escardio.org ] Web sites from January 2002 to June 2011). The following key words were used: “randomized trial,” “myocardial infarction,” “reperfusion,” “primary angioplasty,” “stenting,” “drug-eluting stents (DES),” “bare-metal stent (BMS),” “sirolimus-eluting stent (SES),” “Cypher,” “paclitaxel-eluting stent (PES),” and “Taxus.” Inclusion criteria were (1) randomized treatment allocation, and (2) availability of complete clinical data. Exclusion criteria were (1) follow-up data in <90% of patients, (2) follow-up data no longer than 1 year, (3) ongoing studies or irretrievable data, (4) trials including <50 patients, and (5) unwillingness to provide individual patients’ data. No language restrictions were enforced. All principal investigators were contacted to provide individual patients’ data, which were transferred without patients’ identifiers (initials and birth dates) to Eastern Piedmont University (Novara, Italy). The data set was checked for completeness and consistency and compared with the results of any publications. Queries were resolved by direct correspondence with the responsible study investigator. Data were managed according to the intention-to-treat principle. The diagnosis of diabetes mellitus was based on a history of diabetes at the time of admission.
The primary end point was mortality, and the secondary end points were reinfarction, TVR, target lesion revascularization, and stent thrombosis (defined according to the Academic Research Consortium definition) at long-term follow-up.
Statistical analysis was performed using Review Manager version 4.27 (Cochrane Collaboration, Copenhagen, Denmark) and SPSS version 15.0 (SPSS, Inc., Chicago, Illinois). Continuous variables are expressed as median (interquartile range) and categorical variables as percentages. Pooled odds ratios were calculated using a fixed-effect model (the Mantel-Haenszel method), whereas weighted mean differences were used for continuous variables. Data were pooled using the fixed-effect method with generic inverse variance weighting. The weight of an individual study was measured as the inverse of the estimated variance of the log of the hazard ratio (HR) obtained using Cox proportional-hazards analysis. Potential publication bias was examined by visual estimation of a funnel plot, in which the SE of the natural log of the HR was plotted against the HR for mortality.
A stratified multivariate Cox regression analysis was used to evaluate whether the observed benefits in outcomes were confirmed after correction for major baseline confounding factors (age, gender, hypertension, previous myocardial infarction, previous coronary revascularization, smoking, dyslipidemia, anterior infarct location, Killip class at presentation, time to treatment, preprocedural Thrombolysis In Myocardial Infarction [TIMI] grade 3 flow, and multivessel disease), which were entered in block into the model.
Heterogeneity across trials was assessed using the I 2 statistic. Survival curves are presented as nonstratified Kaplan-Meier curves across trials. Events are reported as probabilities estimated from Kaplan-Meier curves. A subsequent landmark analysis was performed for patients who were event free at 1-year follow-up to define outcomes in terms of short-term events (≤1 year) and long-term events (>1 year). The study was performed in compliance with the Quality of Reporting of Meta-Analyses guidelines.
Results
As previously described, 11 trials were included in the meta-analysis, enrolling a total of 972 patients with diabetes (616 [63.4%] randomized to DES and 356 [36.6%] to BMS). Characteristics of the included trials are listed in Table 1 . In most trials, patients were treated with glycoprotein IIb/IIIa inhibitors combined with a 300- to 600-mg loading dose of clopidogrel. The length of clinical follow-up varied from 3 to 6 years. Baseline characteristics are reported in Table 2 , with no difference observed between the 2 groups.
| Study | Period | Study Design (No. of Patients With Diabetes) | Stent Type | Glycoprotein IIb/IIIa Inhibitors | Primary Study End Point | FU Available (yrs) | Routine Angiographic FU | Dual-Antiplatelet Therapy | |
|---|---|---|---|---|---|---|---|---|---|
| Aspirin | Clopidogrel | ||||||||
| SELECTION | 2004–2005 | SES (n =3) vs BMS (n = 7) | SES and Bx-Velocity | 100% | NIH measured by IVUS at 7 ± 1 mo | 6 | 95% | + | + |
| PASSION | 2003–2004 | DES (n= 38) vs BMS (n = 31) | PES vs Express-2 or Libertè | 73.8% | Combined death, reinfarction, or TLR at 1 yr | 5 | 0 | + | + |
| TYPHOON | 2003–2004 | DES (n = 55) vs BMS (n = 61) | SES or BMS ∗ | 71.5% | Target vessel failure at 1 yr | 4 | 24% | + | + |
| SESAMI | 2003–2004 | DES (n = 29) vs BMS (n = 36) | SES or BMS ∗ | 74.9% | Angiographic restenosis at 1-yr FU | 5 | 52% | + | + |
| PASEO | 2003–2004 | SES (n = 24) or PES (n = 21) vs BMS (n = 23) | SES, PES, BMS ∗ | 100% | Combined death, reinfarction, TVR, and stroke at 6 mos | 5 | 0 | + | + |
| BASKET-AMI | 2003–2004. | SES (n = 7) or PES (n = 10) vs BMS (n = 9) | SES, PES, Vision | 65% | Combined death, reinfarction, TVR, and stroke at 6 mos | 3 | 0 | + | + |
| HORIZONS-AMI | 2005–2007 | PES (n = 364) vs BMS (n = 114) | PES, Express | 52% | (1) 12-mo TLR; (2) death, reinfarction, stroke, and ST | 3 | 40% | + | + |
| MISSION | 2004–2006 | SES (n = 20) vs BMS (n = 10) | SES, BMS | 80% | Late loss at 9-mo FU | 3 | 82% | + | + |
| DEDICATION | 2005–2006 | DES (n = 29 vs BMS (n = 37) | SES, PES, Endeavor; Vision, Express, Libertè, Driver | 96% | Late loss at 8-mo FU | 3 | 100% | + | + |
| Díaz de la Lera et al | 2004–2006 | SES (n = 16) vs BMS (n = 16) | SES, BMS | 100% | Death, nonfatal MI, and recurrent ischemia at 1-yr FU | 6 | 100% | + | + |
| Pasceri et al | 2002–2003 | DES (n = 7), BMS (n = 7) | Cypher vs Bx-Velocity | 89% | Death, nonfatal MI, and recurrent ischemia at 1-yr FU | 6 | 100% | + | + |
∗ Choice of stent was left to the discretion of the operator.
| Variable | DES (n = 616) | BMS (n = 356) | p Value |
|---|---|---|---|
| Age (yrs) | 63 (54–71) | 65 (56–72) | 0.50 |
| Men | 71.9% | 72.2% | 0.93 |
| Hypertension | 62.2% | 61.0% | 0.45 |
| Hypercholesterolemia | 55.5% | 50.2% | 0.74 |
| Symptom onset to balloon time (min) | 230 (165–335) | 240 (155–404) | 0.86 |
| Symptom onset to balloon time <4 h | 49.3% | 51.3% | 0.60 |
| Preprocedural TIMI grade 0 or 1 flow | 63.9% | 63% | 0.69 |
| Postprocedural TIMI grade 3 flow | 89.1% | 89.9% | 0.81 |
| Type of DES | <0.001 | ||
| Cypher | 28.4% | 0% | |
| Taxus | 70.6% | 0% | |
| Endeavor | 1% | 0% | |
| Glycoprotein IIb/IIIa inhibitors | 69.8% | 76.0% | 0.80 |
| Infarct-related coronary artery | 0.59 | ||
| Left main | 0.2% | 0% | |
| Left anterior descending | 45.4% | 48.4% | |
| Anterolateral branch | 0.2% | 0.3% | |
| Circumflex | 13.0% | 11.6% | |
| Right | 39.4% | 38% | |
| Saphenous vein graft | 0.2% | 0% | |
| Clopidogrel | |||
| 6 mos | 93.1% | 87.3% | 0.034 |
| 12 mos | 58.7% | 58.3% | 0.91 |
| 24 mos | 24.7% | 12.8% | 0.89 |
| 36 mos | 18.5% | 9.8% | 0.85 |
At long-term follow-up (median 1,095 days, interquartile range 1,087 to 1,460), a total of 127 patients had died. As depicted in Figure 1 , no significant benefit in mortality was observed with DES compared with BMS (17.6% vs 21.6%, HR 0.79, 95% confidence interval [CI] 0.54 to 1.15, p = 0.21, p for heterogeneity = 0.52). Similar outcomes in terms of death between DES and BMS were observed in the landmark analysis, when early (≤1 year) (5.9% vs 8.8%, HR 0.72, 95% CI 0.42 to 1.24, p = 0.23, p for heterogeneity = 0.66) and late (>1 year) outcomes (12.5% vs 14%, HR 0.91, 95% CI 0.52 to 1.57, p = 0.73, p for heterogeneity = 0.84) were separately analyzed ( Figure 2 ). As shown in Figure 3 , no publication bias was observed.
Reinfarction was observed in a total of 75 patients. As shown in Figure 4 , no significant difference in the occurrence of reinfarction was observed between DES and BMS (11.4% vs 8.7%, respectively, HR 0.83, 95% CI 0.49 to 1.41, p = 0.49, p for heterogeneity = 0.87). As shown in Figure 2 , landmark analysis showed a statistically nondifferent outcome between DES and BMS in the rate of reinfarction within 1 year (5.4% vs 3.8%, HR 0.72, 95% CI 0.35 to 1.48, p = 0.37, p for heterogeneity = 0.96) and later than 1 year (7.4% vs 5%, HR 1.10, 95% CI 0.49 to 2.45, p = 0.82, p for heterogeneity = 0.62).
