Clinical Efficacy of Drug-Eluting Stents in Diabetic Patients

Transcription

1 Journal of the American College of Cardiology Vol. 51, No. 25, by the American College of Cardiology Foundation ISSN /08/$34.00 Published by Elsevier Inc. doi: /j.jacc CLINICAL RESEARCH Interventional Cardiology Clinical Efficacy of Drug-Eluting Stents in Diabetic Patients A Meta-Analysis Ehtisham Mahmud, MD, FACC, Guilherme Bromberg-Marin, MD, Vachaspathi Palakodeti, MD, FACC, Lawrence Ang, BS, Dana Creanga, PHD, Anthony N. DeMaria, MD, MACC San Diego, California Objectives Background Methods The purpose of this study was to compare estimates for revascularization and major adverse cardiac events (MACE) (death, myocardial infarction, repeat revascularization) in diabetic patients treated with paclitaxel- and sirolimus-eluting stents (PES and SES). Outcomes in diabetic patients treated with PES and SES have not been adequately evaluated. We searched MEDLINE/EMBASE from January 2002 to February 2007 and identified abstracts/presentations from this period at major cardiology conferences. Randomized controlled trials (RCTs) and registries were included if data for diabetic patients treated with PES or SES were available. Point estimates with 95% confidence intervals (CIs) were computed as summary statistics. Results In RCTs (13 trials; n 2,422) similar point estimates for target lesion revascularization (TLR) (PES: 8.6%, 95% CI 6.5% to 11.3%; SES: 7.6%, 95% CI 5.8% to 9.9%) and MACE (PES: 15.4%, 95% CI 12.4% to 19.1%; SES: 12.9%, 95% CI 8.5% to 19.2%) were observed. In head-to-head trials (4 RCTs), no difference in the likelihood of TLR (PES vs. SES) was observed (odds ratio [OR] 1.37, 95% CI 0.64 to 2.9, p 0.42). In registries (16 registries; n 10,156), point estimates for target vessel revascularization (TVR) (PES: 5.8%, 95% CI 3.9% to 8.5%; SES: 7.2%, 95% CI 4.6% to 11.2%) and MACE (PES: 10.1%, 95% CI 7.3% to 13.8%; SES: 11.9%, 95% CI 8.6% to 16.4%) were also similar. In registries reporting outcomes with both stents (8 registries for TVR and 7 registries for MACE), the likelihood of TVR (PES vs. SES) (OR 0.77, 95% CI 0.54 to 1.10, p 0.15) and MACE (OR 0.83, 95% CI 0.68 to 1.01, p 0.056) were nonsignificantly lower with PES. Conclusions This analysis of over 11,000 diabetic patients treated with drug-eluting stents demonstrates single-digit revascularization rates. Furthermore, revascularization and MACE estimates are similar with both PES and SES. (J Am Coll Cardiol 2008;51: ) 2008 by the American College of Cardiology Foundation From the Division of Cardiovascular Medicine, University of California San Diego Medical Center, San Diego, California. Dr. Mahmud has received research and clinical trial support in the past from Boston Scientific and Cordis Corporation, manufacturers of paclitaxel- and sirolimus-eluting stents, respectively. Dr. Bromberg- Marin has received educational assistance from both Boston Scientific and Cordis Corporation in the form of a training fellowship grant. Dr. Creanga is an independent biostatistician and consultant and has also provided statistical service for this project and was funded by Boston Scientific. Steven E. Nissen, MD, served as Guest Editor for this article. Manuscript received April 2, 2007; revised manuscript received February 21, 2008, accepted March 7, Epidemiological data have firmly established a relationship between coronary heart disease and diabetes mellitus (1). Although the precise mechanism behind this relationship remains uncertain, hyperglycemia, abnormal lipid metabolism, and insulin resistance, coupled with frequently occurring hypertension, result in acceleration of the atherosclerotic process (2). Coronary artery disease is currently responsible for three-quarters of diabetes-related deaths (3), and as the prevalence of diabetes increases, the number of diabetic patients requiring revascularization for advanced coronary artery disease will escalate. Outcomes with either percutaneous coronary intervention (PCI) or coronary artery bypass graft surgery are poorer for diabetic than nondiabetic patients (4). Percutaneous coronary intervention with bare-metal stents (BMS) was limited in diabetic patients, owing to higher restenosis rates (5). The paclitaxel-eluting stent (PES) Taxus (Boston Scientific, Natick, Massachusetts) and the sirolimus-eluting stent (SES) Cypher (Johnson & Johnson, Miami Lakes, Florida) are both effective in reducing the need for repeat revascularization compared with BMS (6 12), and this has led to PCI being used more frequently in diabetic patients (13,14). However, the

2 2386 Mahmud et al. JACC Vol. 51, No. 25, 2008 DES in Diabetic Patients June 24, 2008: Abbreviations and Acronyms BMS bare-metal stent(s) CI confidence interval DES drug-eluting stent(s) MACE major adverse cardiac event OR odds ratio PCI percutaneous coronary intervention PES paclitaxel-eluting stent(s) RCT randomized controlled trial SES sirolimus-eluting stent(s) TLR target lesion revascularization TVR target vessel revascularization use of DES in diabetic patients is considered off label by the Food and Drug Administration, because adequate numbers of diabetic patients have not been evaluated in clinical trials (15,16). Furthermore, data regarding the relative efficacy of the 2 stents in diabetic patients are less clear. There has only been 1 randomized controlled trial (RCT) comparing the 2 stents exclusively in diabetic patients, and this study included only 125 patients in each arm and also did not have a clinical end point (17). Because late luminal loss is greater with the PES than the SES (6,12), it could be hypothesized that diabetic patients, who have a higher proclivity to restenosis, would have worse clinical outcomes with the PES. However, an attenuation of the antimigratory effect of sirolimus described in the presence of hyperglycemia (18) could lead to lower efficacy of SES in diabetic patients. In addition, the variability in stent platforms, polymer technology, and delivery systems could also lead to differences in clinical outcomes between the 2 stents. The evaluation of diabetic patients treated with PES and SES in multiple RCTs and registries provides an opportunity to determine clinical outcomes of DES in diabetic patients and to compare the relative efficacy and safety of these 2 DES. With suggested guidelines (19), this meta-analysis was performed to estimate target lesion revascularization (TLR), major adverse cardiac events (MACE), and stent thrombosis, for diabetic patients receiving either PES or SES as part of an RCT and to compare the efficacy of the 2 DES from head-tohead RCTs. With registry data, we also estimated target vessel revascularization (TVR) and MACE for diabetic patients receiving either DES. Comparison of the 2 DES from registries that reported outcomes with both stents in similar populations was also separately performed. Because angiographic measures, such as binary restenosis and late loss, are debatable end points for clinical utility in individual patients and are not available for most registries, we specifically chose not to include them in this analysis. Society of Cardiology, Transcatheter Cardiovascular Therapeutics, and Paris Course on Revascularization) were also reviewed. The search terms drug, eluting, and stent were combined with the terms registry or registries and then clinical and trial or trials as 2 distinct published report searches. The RCTs and registries were included in the analysis if data for diabetic patients treated with a PES or SES were available for at least 1 of the clinical end points (TLR or MACE for RCTs; TVR or MACE for registries) being assessed and were not duplicative. Target lesion revascularization was defined as any procedure (PCI or coronary artery bypass graft surgery) performed to revascularize the index lesion, whereas TVR was defined as any procedure to revascularize the index vessel. Although the definition of MACE varied slightly in the various studies, it usually consisted of cardiac death, myocardial infarction, and repeat revascularization. RCTS. All studies and abstracts/presentations were identified and reviewed by at least 2 of the authors. The RCTs Methods Identification and selection of studies. A search was conducted of the English language published reports in MEDLINE and EMBASE from January 2002 to February Abstracts and presentations from this time period at major international cardiology conferences (American College of Cardiology, American Heart Association, European Figure 1 Clinical Event Rates for Diabetic Patients Treated With DES in RCTs Point estimates for target lesion revascularization and major adverse cardiac events for diabetic patients treated with drug-eluting stents (DES) in (A) all randomized controlled trials (RCTs) and (B) RCTs with Silber score 5. Bars indicate 95% confidence intervals.

3 JACC Vol. 51, No. 25, 2008 June 24, 2008: Mahmud et al. DES in Diabetic Patients 2387 Figure 2 Comparative Analysis of Target Lesion Revascularization With Drug-Eluting Stents in Diabetic Patients From Head-to-Head Randomized Controlled Trials Likelihood of target lesion revascularization (paclitaxel-eluting stent [PES] vs. sirolimus-eluting stent [SES]) (n 899 diabetic patients). Size of boxes indicates relative contribution of each trial. See Table 1 for further details regarding the individual trials included. CI confidence interval; ISAR-DIABETES Intracoronary Stenting and Angiographic Results: Do Diabetic Patients Derive Similar Benefit from Paclitaxel-Eluting and Sirolimus-Eluting Stents trial; pc personal communication from J. J. Goy, 2006; REALITY Sirolimus- vs Paclitaxel-Eluting Stents in De Novo Coronary Artery Lesions trial; SIRTAX Sirolimus-eluting and Paclitaxel-eluting Stents for Coronary Revascularization trial; TAXi A Prospective Randomized Comparison Between Paclitaxel and Sirolimus Eluting Stents in the Real World of Interventional Cardiology trial. were only included in the analysis if they involved de novo coronary lesions, the methodology and results of the trial had been published, and data for the diabetic subset were available. Published data were always given priority, and if the data were unavailable for the diabetic cohort, the principal investigator was contacted and asked to provide the necessary information. There were no numerical limitations regarding trial size for study inclusion. A method of assessing the quality of clinical trials, including RCTs with DES, has been proposed (20), adhering to the principles of evidence-based medicine. The Silber score evaluates various factors that constitute a well-designed RCT, including adequate power, being multicenter, having an independent events committee, and having a primary clinical end point. High scores (closer to 10) suggest a stronger basis for making an evidence-based decision, whereas low scores (closer to 0) provide hypotheses rather than confirmatory evidence. A Silber score was assigned to all of the RCTs. REGISTRIES. All registries with prospectively collected outcomes data after the use of either the PES or SES were identified. Registries were included if data for at least 1 of the end points (TVR or MACE) for diabetic patients could be extracted from the published manuscript or from abstracts/presentations. In addition, registries that were not published in complete form were also analyzed separately. Statistical methodology. All analyses were conducted on the basis of the intent-to-treat principle. Point estimates and 95% confidence intervals (CIs) were computed for all outcomes as summary statistics with both random effects (Der Simmonian and Laird method) (21) and fixed effects models (chi-square with normal approximation) (22). The null hypothesis of homogeneity of response across studies was tested with the Cochran Q statistic. If the null hypothesis was rejected, point estimates and 95% CI estimated on the basis of a random effects model were presented. Otherwise, the fixed effects solution was presented. All of the point estimates and comparative analyses from the head-tohead RCTs and registries are based on the random effects solution. The only outcomes presented with the fixed effects solution to derive point estimates and 95% CI are for TLR and MACE for each DES from the RCTs. Because stent thrombosis was a low-incident event, the estimated rate and 95% CIs in the analysis are based on the observed data and represent raw percentages computed as the number of reported stent thromboses out of the total diabetic cohort. Statistical analyses were performed with the Comprehensive Meta-analysis software, version 2.0 (Biostat Inc., Englewood, New Jersey). The results presented in this analysis are based primarily on studies that do not evaluate both PES and SES in the same population. Given the diversity of the populations and the differences between RCT and registry data, a direct head-to-head statistical comparison was not pursued for the entire dataset. However, a comparative analysis is presented for TLR from a subset of head-to-head Figure 3 Clinical Event Rates for Diabetic Patients Treated With Drug-Eluting Stents in Registries Point estimates for target vessel revascularization and major adverse cardiac events for diabetic patients treated with drug-eluting stents in all registries. Bars indicate 95% confidence intervals.

4 2388 Mahmud et al. JACC Vol. 51, No. 25, 2008 DES in Diabetic Patients June 24, 2008: Randomized Controlled Trials Included in the Meta-Analysis Table 1 Randomized Controlled Trials Included in the Meta-Analysis Mean Age (Mean SD) Gender (% Male) Study Summary Study Site Total Patients (n) PES SES PES SES Follow-Up* (Months) ISAR-Diabetes (17) SES vs. PES comparison at 2 centers Munich, Germany 125 SES, 125 PES TAXi (23), J. J. Goy, personal communication, 2006 SIRTAX (24,26) REALITY (25,26) SIRIUS (6) DIABETES (28) SES-SMART (29) RAVEL (30) TAXUS II (31,32) TAXUS IV (13) TAXUS V (33,34) TAXUS VI (35) ISAR-Test (27), A. Kastrati, personal communication, 2006 SES vs. PES comparison at single center SES vs. PES comparison at 2 centers SES vs. PES comparison SES vs. BMS comparison SES vs. BMS comparison SES vs. BMS comparison SES vs. BMS comparison PES vs. BMS comparison at multiple centers PES vs. BMS comparison PES vs. BMS comparison PES vs. BMS comparison PES vs. SCS at single center Lausanne, Switzerland Bern, Switzerland 102 SES, 100 PES SES, 509 PES Global 684 SES, 669 PES U.S. Spain Italy Europe Europe U.S. U.S. Europe Munich, Germany 533 SES, 525 BMS 80 SES, 80 BMS 129 SES, 128 BMS 120 SES, 118 BMS 266 PES, 270 BMS 662 PES, 652 BMS 577 PES, 579 BMS 219 PES, 227 BMS SR MR 70SR 76 MR PES, 225 SCS *Follow-up time is reported for target lesion revascularization (TLR) (follow-up for other outcomes might vary); percent and actual count of events where reported; baseline characteristics reported for diabetic patients only; results and demographic data represent a combination of both the moderate- (MR) and slow-release (SR) paclitaxel-eluting stent (PES) formulations. BMS bare-metal stent; MACE major adverse cardiac events; SES sirolimus-eluting stent; SCS sirolimus-coated stent. RCTs in which the 2 stents were compared. Similarly, a comparative analysis of TVR and MACE from a subset of registries that reported outcomes with both DES in similar populations is also presented. Results RCTs. There were 13 RCTs that met the inclusion criteria in which 2,422 diabetic patients were enrolled (Table 1). Of these 13 studies, 4 involved a direct head-to-head comparison of the PES and SES (17,23 25), of which only 1 was a multicenter trial (25). The ISAR-DIABETES trial (Intracoronary Stenting and Angiographic Results: Do Diabetic Patients Derive Similar Benefit from Paclitaxel- Eluting and Sirolimus-Eluting Stents) was a 2-center study and is the only trial exclusively conducted in diabetic patients (n 250) with a direct head-to-head comparison of the 2 DES (17). The TAXi (A Prospective Randomized Comparison Between Paclitaxel and Sirolimus Eluting Stents in the Real World of Interventional Cardiology) trial was a single center comparison between the 2 stents in which one-third of the patients were diabetic (n 69) (23) (J. J. Goy, personal communication, 2006). The SIRTAX (Sirolimus-eluting and Paclitaxel-eluting Stents for Coronary Revascularization) trial, was a 2-center study that compared outcomes between these 2 DES with a primary clinical end point and enrolled approximately 20% diabetic patients (n 201) (24,26). The REALITY (Sirolimus- vs. Paclitaxel-Eluting Stents in De Novo Coronary Lesions) trial is the only multicenter clinical trial conducted to compare the 2 DES and enrolled 28% diabetic patients (n 379) (25,26). The limitations of the 4 aforementioned head-to-head trials include: a small number of centers (17,23,24), absence of an independent core laboratory (17,23,24), an inability to angiographically blind the investigators to the stent placed (17,23 25), and the use of angiographic primary end points (17,25). In the evaluation of these trials with the Silber score, all 4 trials were assigned lower Silber score values than the pivotal RCTs of these DES versus their BMS controls (6,12). In the other 9 RCTs, PES or SES were compared against their BMS control group, except for the ISAR-TEST (Intracoronary Stenting and Restenosis-Test Equivalence Between 2 Drug-Eluting Stents) in which a PES (26% diabetic patients; n 58) was compared against a BMS coated on-site with nonpolymeric rapamycin (not SES Cypher) (27) (A. Kastrati, personal communication, 2006). The 4 SES trials were all multicenter and included the pivotal SIRIUS (Sirolimus-eluting Balloon-Expandable Stent in the Treat-

5 JACC Vol. 51, No. 25, 2008 June 24, 2008: Mahmud et al. DES in Diabetic Patients 2389 Continued Table 1 Continued Silber Score (24) (0 10) Diabetic Patients (n) TLR, % (n/n) MACE, % (n/n) Stent Thrombosis at 8 12 Months Follow-Up, % (n/n) PES SES PES SES PES SES SES, 125 PES 12.0 (15/125) 6.4 (8/125) 0.8 (1/125) 0.0 (0/125) 5 33 SES, 36 PES 2.8 (1/36) 3.0 (1/33) 5.6 (2/36) 9.1 (3/33) SES, 93 PES (2/93) 0.0 (0/108) SES, 192 PES (6/192) 1.1 (2/187) SES, 148 BMS SES, 80 BMS (0/80) 6 29 SES, 45 BMS 17.2 (5/29) 20.7 (6/29) 3.4 (1/29) 7 19 SES, 25 BMS 0.0 (0/19) 10.5 (2/19) 0.0 (0/19) 7 32 PES, 40 BMS 3.1 (1/32) 15.6 (5/32) PES, 163 BMS PES, 173 BMS 9.6 (17/178) 16.9 (30/178) 0.6 (1/177) PES, 50 BMS 2.6 (1/39) 4 58 PES, 73 SCS 10.3 (6/58) 13.8 (8/58) ment of Patients with De Novo Native Coronary Artery Lesions) trial (6,28 30) but only 1 trial that exclusively enrolled diabetic patients (n 160) (28). The 4 PES trials were also all multicenter and included the pivotal Taxus IV study (13) in addition to the Taxus V and VI trials, which enrolled increasingly complex lesions and patients (31 35). Clinical follow-up was obtained in all studies at 8 to 12 months, whereas angiographic follow-up was obtained in 75% of patients in all studies except the SIRTAX (53.4% angiographic follow-up) and TAXi trials (no angiographic evaluation mandated). Many published sources only provided patient age and gender for the entire population (diabetic and nondiabetic), but when diabetic patient information was provided (13,17,28 30) no differences in age or gender were noted. The definition of MACE (cardiac death, myocardial infarction, ischemia driven TLR or TVR) was reasonably homogenous across the trials. However, subtle variations in the definition included all cause mortality and documented stent thrombosis or occlusion in the TAXi and ISAR-TEST trials as additional MACE end points (23,27). Three trials did not require revascularization to be ischemia driven (28,30,31). Data concerning stent thrombosis were only available for 8 studies (13,17,24 26,28 30,33,34) and were based on reported events similar to those in real world studies (36) and before the use of the recently agreed upon Academic Research Consortium definition of stent thrombosis (37). Dual antiplatelet therapy with aspirin and a thienopyridine was routinely recommended for at least 6 months after the placement of a PES and for at least 2 to 3 months after the placement of an SES in the various trials. Registries. There were 16 registries involving 10,156 diabetic patients (PES 5,597; SES 4,559) that met the inclusion criteria and were included in this meta-analysis (Table 2) (38 53). Six of these registries have been published in complete form (38,39,42,45,47,49), whereas data from the other 10 are derived from published abstracts and presentations. The majority of registries (10 studies) (40,41,43, 46,48 53) were multicenter, whereas data from 8 registries (38 45) were comparative between the 2 DES. Follow-up was between 6 and 12 months after the index procedure, and the major clinical outcomes of TVR and MACE (death, myocardial infarction, and TVR) were available for most studies. To maintain consistency, when MACE events were reported separately as TLR-MACE and TVR- MACE, we chose to report and include only TVR-MACE in the analysis. Patient age and gender data were only available for the diabetic patients from 9 registries (39

6 2390 Mahmud et al. JACC Vol. 51, No. 25, 2008 DES in Diabetic Patients June 24, 2008: Registries Included in Meta-Analysis Table 2 Registries Included in Meta-Analysis Mean Age (Mean SD) Study Summary Study Site Total Patients (n) PES SES Cosgrave et al. (38) 1 center Milan, Italy 248 SES, 281 PES RESEARCH/T-SEARCH (39) SOLACI (40) STENT (all) (41) C vs. T REWARDS (42) TC-WYRE (43) Prairie Institute (44) Stankovic et al. (45) SIRIUS-DIRECT (46) 1 center 20 centers 8 centers 1 center 19 centers 1 center 2 centers SES direct stenting at 16 centers Rotterdam, the Netherlands 145 SES, 148 PES Latin America 353 SES, 280 PES U.S. 3,421 SES, 3,142 PES Washington DC 873 SES, 447 PES U.S. 742 SES, 816 PES Springfield, Illinois 954 SES, 208 PES Italy 147 SES, 113 PES U.S. 225 SES Berenguer et al. (47) SES experience at 1 center Spain 231 SES 65 German Cypher-Stent SES experience at 376 centers Germany 5,878 SES 63.8 Registry (48) ARTS II (49) SES experience at 45 centers Europe 159 SES PORTO I (50) SES experience at 9 centers Portugal 120 SES MILESTONE II (51) PES experience at 164 centers Global 3,688 PES ARRIVE 1 (52,53) PES experience at 50 centers U.S. 2,585 PES ARRIVE 2 (53) PES experience at 53 centers U.S. 5,007 PES *Follow-up time is reported for target vessel revascularization (TVR) (follow-up for other outcomes might vary); percent and actual count of events where reported; baseline characteristics reported for diabetic patients only; estimates based on the reported demographic data for insulin- and non insulin-dependent diabetic patients. MACE major adverse cardiac events; SES sirolimus-eluting stent; PES paclitaxel-eluting stent. 42,44,45,47,49,50). With the exception of ARTS II (Arterial Revascularization Therapy Study II) (49), which was a registry-controlled trial, all of the other registries reported real-world experience that included evaluation of DES in simple and complex lesions during single and multivessel PCI. Mandatory angiographic follow-up was only required in 4 registries (46,47,49,50), and therefore the majority of outcomes events were clinically driven. Dual antiplatelet therapy with aspirin and a thienopyridine was routinely recommended for at least 6 months after the placement of a PES and for at least 3 months after the placement of an SES. Pooled analysis. The pooled analysis of the RCTs yields estimated rates of TLR (PES: 8.6%, 95% CI 6.5% to 11.3%; SES: 7.6%, 95% CI 5.8% to 9.9%) and MACE (PES: 15.4%, 95% CI 12.4% to 19.1%; SES: 12.9%, 95% CI 8.5% to 19.2%) in diabetic patients that were similar for both DES with overlapping CIs for the respective point estimates (Table 3)(Fig. 1A). The estimated rates of stent thrombosis (PES: 1.48%, 95% CI 0.74% to 2.63%; SES: 0.55%, 95% CI 0.11% to 1.59%) were low for both stents. Because studies with lower Silber scores more often generate hypotheses than resolve questions (20), we repeated the analysis for patients enrolled in trials with a Silber score 5, which resulted in exclusion of 4 studies (17,23,25,27). In the studies with higher Silber scores (6,13,24,28 36), the estimated rates of TLR (PES: 8.9%, 95% CI 6.2% to 12.6%; SES: 8.0%, 95% CI 5.4% to 11.6%), MACE (PES: 16.2%, 95% CI 12.8% to 20.4%; SES: 13.8%, 95% CI 8.7% to 21.0%), and stent thrombosis (PES: 1.23%, 95% CI 0.34% to 3.12%; SES: 0.42%, 95% CI 0.01% to 2.34%) were also similar for both DES (Table 3) (Fig. 1B). A head-to-head comparison of TLR from the 4 RCTs (17,24 27) (A. Kastrati, J. J. Goy, personal communication, 2006) that enrolled diabetic patients treated with both DES revealed no difference in the likelihood of requiring revascularization with either stent in diabetic patients (PES vs. SES; odds ratio [OR] 1.37, 95% CI 0.64 to 2.9, p 0.42) (Fig. 2). When evaluating the registries, the point estimates for TVR (PES: 5.8%, 95% CI 3.8% to 8.5%; SES: 7.2%, 95% CI 4.6% to 11.2%) and MACE (PES: 10.1%, 95% CI 7.3% to 13.8%; SES: 11.9%, 95% CI 8.6% to 16.4%) showed similar and low rates for both DES (Table 3) (Fig. 3). Because all of the registries had not appeared in a peerreviewed journal at the time of this study, the analysis was repeated separately for those published and revealed a similar rate of TVR and MACE with both stents, although the absolute event rates in the published studies were higher (Table 3). Eight registries (38 45) reported outcomes of TVR with both stents in similar populations, whereas 7 registries (38 42,44,45) reported outcomes of MACE with both stents in similar populations. A comparative analysis of these studies revealed likelihood of TVR (PES vs. SES; OR

7 JACC Vol. 51, No. 25, 2008 June 24, 2008: Mahmud et al. DES in Diabetic Patients 2391 Continued Table 2 Continued Gender (% male) TVR, % (n/n) MACE, % (n/n) PES SES Follow-Up* (Months) Diabetic Patients (n) PES SES PES SES SES, 67 PES 19.4 (13/67) 33.3 (20/60) 23.9 (16/67) 36.7 (22/60) SES, 148 PES SES, 280 PES SES, 805 PES 4.1 (33/805) 3.4 (30/875) 6.8 (55/805) 7.1 (62/875) SES, 447 PES 5 (11/221) 3 (19/630) 12 (27/221) 11 (69/630) SES, 289 PES SES, 208 PES SES, 113 PES 18.6 (21/113) 23.1 (34/147) 19.5 (22/113) 24.5 (36/147) SES SES 12.2 (12/98) SES 9.8 (45/458) SES SES 2.5 (3/120) 5.0 (6/120) ,254 PES 6.7 (74/1,098) 9.4 (103/1,098) PES 5.7 (43/750) 8.4 (63/750) ,236 PES , 95% CI 0.54 to 1.10, p 0.15) (Fig. 4A) and MACE (PES vs. SES; OR 0.83, 95% CI 0.68 to 1.01, p 0.056) (Fig. 4B) to be nonsignificantly lower with the PES. Discussion This report of over 12,000 diabetic patients is the largest study evaluating outcomes of DES in diabetic patients and provides a number of important findings. This metaanalysis demonstrates that DES are highly effective in reducing the need for repeat revascularization in diabetic patients, decreasing the incidence to single digits. More- Results of Pooled Analysis With Point Estimates for Clinical Events in RCTs and Registries Table 3 Results of Pooled Analysis With Point Estimates for Clinical Events in RCTs and Registries over, in addition to the results obtained in pivotal RCTs of DES, these results are also seen in high-risk real world populations. Furthermore, this analysis represents the most comprehensive and definitive data showing that the rates of repeat revascularization, MACE, and stent thrombosis are similar for PES and SES. The data also demonstrate that short-term stent thrombosis rates with both DES are low in diabetic patients and comparable to those reported for unselected populations (54). Diabetic patients are predisposed to more aggressive atherosclerosis and are at high risk for restenosis (1 5). Revascularization,* % (95% CI) MACE,* % (95% CI) Stent Thrombosis, % (95% CI) PES SES PES SES PES SES RCTs (13,17,23 37) 8.6 ( ) 7.6 ( ) 15.4 ( ) 12.9 ( ) 1.48 ( ) 0.55 ( ) RCTs with Silber score 5 (13,24,26,28 37) 8.9 ( ) 8.0 ( ) 16.2 ( ) 13.8 ( ) 1.23 ( ) 0.42 ( ) Registries (40 56) 5.8 ( ) 7.2 ( ) 10.1 ( ) 11.9 ( ) NA NA Registries (published) 10.5 ( ) 10.5 ( ) 16.9 ( ) 15.9 ( ) NA NA (40,41,44,45,48,50,52) Registries (presented) (42,43,46,47,49,51,53 56) 4.2 ( ) 5.7 ( ) 7.1 ( ) 8.2 ( ) NA NA *Point estimates; target lesion revascularization (TLR) for randomized controlled trials (RCTs) and target vessel revascularization (TVR) for registries; actual event rates. CI confidence interval; MACE major adverse cardiac events (death, myocardial infarction, and target lesion revascularization or target vessel revascularization); PES paclitaxel-eluting stent; SES sirolimus-eluting stent.

8 2392 Mahmud et al. JACC Vol. 51, No. 25, 2008 DES in Diabetic Patients June 24, 2008: Figure 4 Comparative Analysis of Clinical Events for Diabetic Patients Treated With Drug-Eluting Stents in Registries Enrolling Similar Populations (A) Likelihood of target vessel revascularization (PES vs. SES) (n 5,542 diabetic patients). (B) Likelihood of major adverse cardiac events (PES vs. SES) (n 5,006 diabetic patients). Size of boxes indicates relative contribution of each study. See Table 2 for further details regarding registries included. C REWARDS Cypher Registry Experience at the Washington Hospital Center With Drug-Eluting Stents; RESEARCH/T-SEARCH Rapamycin-Eluting and Taxus Stent Evaluated At Rotterdam Cardiology Hospital registries; SOLACI Society of Latin American Interventional Cardiologists registry; STENT Strategic Transcatheter Evaluation of New Therapies group; T REWARDS Taxus Registry Experience at the Washington Hospital Center With Drug-Eluting Stents; TC-WYRE The Taxus Cypher What s Your Real-World Experience study; other abbreviations as in Figure 2. They usually have smaller vessels and longer lesions, which are additional predictors of restenosis (2 5). Despite these facts, this analysis demonstrates that single-digit rates of repeat revascularization can be achieved in diabetic patients with DES, which represents a significant improvement over the 20% to 40% rate of repeat revascularization previously seen with BMS (6,12,28,30). Importantly, low revascularization rates were observed not only in RCTs, which typically include selected and lower-risk individuals, but also in registries, which include a much broader patient population and lesion subset. Surprisingly, the estimated rate of repeat intervention in diabetic patients from registry data (TVR: PES 5.8%, SES 7.2%) was lower than the rate from the RCTs (TLR: PES 8.6%, SES 7.6%), despite including reintervention for any lesion in the treated artery. However, this is consistent with previous studies that have demonstrated a higher reintervention rate in RCTs with mandated angiographic follow-up (55), whereas reintervention is usually symptomdriven in registries. Therefore, even though registries typically include more higher-risk patients than are usually enrolled in RCTs, the actual rate of revascularization is often lower (2). The current analysis demonstrates that in diabetic patients the estimated rates of TLR, TVR, and MACE are similar for patients receiving a PES or an SES (Figs. 1 to 4). These data do not confirm the results of a single study involving 2 centers (17) that suggested a difference in clinical outcomes between PES and SES in diabetic patients. The analysis was repeated for multicenter RCTs with a Silber score exceeding 5 (20), to ensure that data were reliable and in adherence with the principles of evidencebased medicine. This revealed that the numerical differences in TLR and MACE rates between PES and SES were even less pronounced in these studies (Fig. 1B). Rates of MACE and stent thrombosis with either DES were also considered comparable to those previously reported for nondiabetic patients (56). Analysis of the 4 RCTs that reported a head-to-head comparison of PES and SES also demonstrated no difference in the rate of repeat revascularization between the 2 stents. These data provide the strongest evidence that any difference that might exist in late loss or binary restenosis between these 2 stents does not play a clinically meaningful role in the treatment of diabetic patients. These findings contrast with the results of a previous meta-analysis, not limited to diabetic patients, in which the TLR rate in populations was lower in patients treated with SES (5.1%) compared with PES (7.8%) (OR 0.64; 95% CI 0.49 to 0.84; p 0.001) (54). Because late lumen loss and the degree of neointimal tissue formation is greater with PES compared with SES (2,6,54), one might have expected clinically significant differences in revascularization rates in diabetic patients with the 2 stents. Our findings might be

9 JACC Vol. 51, No. 25, 2008 June 24, 2008: Mahmud et al. DES in Diabetic Patients 2393 due to a difference in the mechanism of action of paclitaxel and sirolimus in diabetic patients. Both drugs act by reducing smooth muscle cell proliferation, which is central to the development of neointimal hyperplasia and restenosis, through inhibition of the cell cycle. Sirolimus induces G 1 cell cycle inhibition, whereas paclitaxel predominantly leads to M-phase arrest (57). Smooth muscle cell cultures have been used to compare the activities of sirolimus and paclitaxel in a model of diabetes. Although both drugs activate mitogen-activated protein kinase pathways similarly, sirolimus potently activates AKT-dependent signaling, overriding the down-regulation of this pathway by insulin resistance. This effect is associated with attenuation of the antimigratory effects of sirolimus in the presence of hyperglycemia, which might account for decreased efficacy in diabetic patients as opposed to nondiabetic patients (18). Two-year outcomes for diabetic patients treated with PES or SES available from 1 single-center registry are consistent with this hypothesis, because TVR at 2 years was numerically lower for diabetic patients treated with PES (9.7% PES vs. 15.3% SES, p 0.06) (58). However, the results of our comparative analysis of TVR and MACE from registries reporting outcomes with both DES did not demonstrate a significant difference between the 2 DES (Fig. 4). The occurrence of late stent thrombosis with DES is an increasingly important topic. Several pathological studies have shown that neointimal coverage might be incomplete 3 to 6 months after DES implantation, and this has been associated with subclinical thrombus formation (59). The risk of stent thrombosis might be even greater in diabetic patients. In a porcine model, uncontrolled hyperglycemia resulted in increased thrombosis after coronary stent placement (60), and diabetes has also been associated with a higher rate of late stent thrombosis (36). In this metaanalysis, the reported rates of short-term stent thrombosis for both DES were low (PES 1.48%; SES 0.55%) and similar to those previously reported for mixed diabetic and nondiabetic populations (PES 1.1%, SES 0.9%) (54). However, these results should be interpreted with caution, given the small number of events reported, the nonuniform definitions used, and the intermediate follow-up period of 6 to 12 months in the studies included in this analysis. Study limitations. Owing to the differences in angiographic appearance of PES and SES, it is impossible to adequately blind observers to the 2 stents. All studies were open label in design and therefore might have been subject to observer bias. Because data were collected from a variety of studies, the length of follow-up varied and baseline data were heterogenous. Use of the random effects model should help overcome some of these limitations. Furthermore, data from registries are nonrandomized and suffer from selection bias. We did not include studies evaluating newer DES, because intermediate and long-term outcomes with these stents are not yet available. Owing to the low number of events, nonuniform definitions, and the absence of long-term follow-up, this study is unable to definitively address the issue of late stent thrombosis with DES in diabetic patients. Whereas individual patient data are preferred for any analysis, in the absence of access to such data, clinical inferences based on summarized results are informative. With over 12,000 diabetic patients (over 11,000 treated with DES) in this analysis, we believe reliable conclusions can be drawn regarding similar clinical outcomes with both PES and SES in diabetic patients. Conclusions This meta-analysis demonstrates that DES are associated with single-digit repeat revascularization rates in diabetic patients. No differences in clinical outcomes with the use of PES or SES were observed in these patients. Acknowledgments The authors would like to thank Ross J. Kay, PhD, for editorial assistance. Editorial support was provided by Boston Scientific. Reprint requests and correspondence: Dr. Ehtisham Mahmud, Director, Cardiovascular Catheterization Laboratories, University of California, San Diego Medical Center, 200 West Arbor Drive, San Diego, California emahmud@ucsd.edu. REFERENCES 1. Goldberg I. Why does diabetes increase atherosclerosis? I don t know! J Clin Invest 2004;114: King S, Dangas G, Moses JW. Surgery is preferred for the diabetic with multivessel disease. Circulation 2005;112: Roffi M, Topol EJ. Percutaneous coronary intervention in diabetic patients with non ST-segment elevation acute coronary syndromes. Eur Heart J 2004;25: Mathew V, Holmes DR. Outcomes in diabetics undergoing revascularization: the long and the short of it. J Am Coll Cardiol 2002;40: Kornowski R, Mintz GS, Abizaid A, Leon MB. Intravascular ultrasound observations of atherosclerotic lesion formation and restenosis in patients with diabetes mellitus. Int J Cardiovasc Intervent 1999;2: Moses JW, Leon MB, Popma JJ, et al. Sirolimus-eluting stents versus standard stents in patients with stenosis in a native coronary artery. N Engl J Med 2003;349: Schofer J, Schluter M, Gershlick AH, et al. Sirolimus-eluting stents for treatment of patients with long atherosclerotic lesions in small coronary arteries: double-blind, randomised controlled trial (E- SIRIUS). Lancet 2003;362: Schampaert E, Cohen EA, Schluter M, et al. The Canadian study of the sirolimus eluting stent in the treatment of patients with long de novo lesions in small native coronary arteries (C-SIRIUS). J Am Coll Cardiol 2004;43: Fajadet J, Morice MC, Bode C, et al. Maintenance of long-term clinical benefit with sirolimus-eluting coronary stents: three-year results of the RAVEL trial. Circulation 2005;111: Colombo A, Drzewiecki J, Banning A, et al. Randomized study to assess the effectiveness of slow- and moderate-release polymer-based paclitaxel-eluting stents for coronary artery lesions. Circulation 2003; 108: Stone GW, Ellis SG, Cox DA, et al. One-year clinical results with the slow release, polymer-based, paclitaxel-eluting TAXUS stent: the TAXUS-IV trial. Circulation 2004;109:

10 2394 Mahmud et al. JACC Vol. 51, No. 25, 2008 DES in Diabetic Patients June 24, 2008: Stone GW, Ellis SG, Cox DA, et al. A polymer-based, paclitaxeleluting stent in patients with coronary artery disease. N Engl J Med 2004;350: Hermiller JB, Raizner A, Cannon L, et al. Outcomes with the polymer-based paclitaxel-eluting TAXUS stent in patients with diabetes mellitus: the TAXUS-IV trial. J Am Coll Cardiol 2005;45: Holmes DR Jr., Leon MB, Moses JW, et al. Analysis of 1-year clinical outcomes in the SIRIUS trial: a randomized trial of a sirolimus-eluting stent versus a standard stent in patients at high risk for coronary restenosis. Circulation 2004;109: Maisel WH. Unanswered questions drug-eluting stents and the risk of late thrombosis. N Engl J Med 2007;356: Farb A, Boam AB. Stent thrombosis redux the FDA perspective. N Engl J Med 2007;356: Dibra A, Kastrati A, Mehilli J, et al. Paclitaxel-eluting or sirolimuseluting stents to prevent restenosis in diabetic patients. N Engl J Med 2005;353: Patterson C, Mapera S, Li HH, et al. Comparative effects of paclitaxel and rapamycin on smooth muscle migration and survival: role of AKT-dependant signaling. Arterioscler Thromb Vasc Biol 2006;26: Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of randomized controlled trials: the QUOROM statement. Lancet 1999;354: Silber S. A new and rapid scoring system to assess the scientific evidence from clinical trials. J Interv Cardiol 2006;19: Der Simonian R, Laird N. Metaanalysis in clinical trials. Control Clin Trials 1986;7: Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22: Goy JJ, Stauffer JC, Sieganthaler M, Benoit A, Seydoux C. A prospective randomized comparison between paclitaxel and sirolimus stents in the real world of interventional cardiology: the TAXi trial. J Am Coll Cardiol 2005;45: Windecker S, Remondino A, Eberli FR, et al. Sirolimus-eluting and paclitaxel-eluting stents for coronary revascularization. N Engl J Med 2005;353: Morice MC, Colombo A, Meier B, et al. Sirolimus- vs paclitaxeleluting stents in de novo coronary artery lesions: the REALITY trial: a randomized controlled trial. JAMA 2006;295: Windecker S. Cypher is preferred in diabetics. Presentation at the 2006 Transcatheter Cardiovascular Therapeutics Annual Meeting October, 24, 2006, Washington, DC. Available at: Accessed February 4, Mehilli J, Kastrati A, Wessely R, et al. Randomized trial of a nonpolymer-based rapamycin-eluting stent versus a polymer-based paclitaxel-eluting stent for the reduction of late lumen loss. Circulation 2006;113: Sabate M, Jimenez-Quevedo P, Angiolillo DJ, et al. Randomized comparison of sirolimus-eluting stent versus standard stent for percutaneous coronary revascularization in diabetic patients: the diabetes and sirolimuseluting stent (DIABETES) trial. Circulation 2005;112: Ortolani P, Ardissino D, Cavallini C, et al. Effect of sirolimus-eluting stent in diabetic patients with small coronary arteries (a SES-SMART substudy). Am J Cardiol 2005;96: Abizaid A, Costa MA, Blanchard D, et al. Sirolimus-eluting stents inhibit neointimal hyperplasia in diabetic patients. Insights from the RAVEL Trial. Eur Heart J 2004;25: Colombo A, Drzewiecki J, Banning A, et al. Randomized study to assess the effects of slow- and moderate-release polymer-based paclitaxel-eluting stents for coronary artery lesions. Circulation 2003;108: Colombo A. Twelve month clinical follow-up of TAXUS II paclitaxel-eluting stent study. Paper presented at: 2003 American College of Cardiology Annual Scientific Session; March 30 April 2, 2003; Chicago, IL. 33. Stone GW, Ellis SG, Cannon L, et al. Comparison of a polymerbased paclitaxel-eluting stent with a bare metal stent in patients with complex coronary artery disease: a randomized trial. JAMA 2005;294: Ellis SG. TAXUS V trial global results: expanding the randomized data. Paper presented at: 2005 American College of Cardiology Annual Scientific Session; March 6 9, 2005; Orlando, FL. 35. Dawkins KD, Grube E, Guagliumi G, et al and Taxus VI Investigators. Clinical efficacy of polymer-based paclitaxel-eluting stents in the treatment of complex, long coronary artery lesions from a multicenter, randomized trial: support for the use of drug-eluting stents in contemporary clinical practice. Circulation 2005;112: Iakovou I, Schmidt T, Bonizzoni E, et al. Incidence, predictors, and outcome of thrombosis after successful implantation of drug-eluting stents. JAMA 2005;293: Cutlip DE, Windecker S, Mehran R, et al. Clinical end points in coronary stent trials: a case for standardized definitions. Circulation 2007;115: Cosgrave J, Agostoni P, Ge L, et al. Clinical outcome following aleatory implantation of paclitaxel-eluting or sirolimus-eluting stents in complex coronary lesions. Am J Cardiol 2005;96: Ong AT, Aoki J, van Mieghem CA, et al. Comparison of short- (one month) and long- (twelve months) term outcomes of sirolimus- versus paclitaxel-eluting stents in 293 consecutive patients with diabetes mellitus (from the RESEARCH and T SEARCH registries). Am J Cardiol 2005;96: Sousa AG, Abizaid A, Hayashi EB, Buitron F, Pena I, Ribeiro E. Outcomes of percutaneous interventions in diabetics treated with drug eluting stents or bare metal stents: SOLACI registry results. Am J Cardiol 2005;96 Suppl 1:31H Simonton C, Brodie B, Cheek B, Late et al. clinical outcomes of paclitaxel and sirolimus-eluting coronary stents in diabetic patients: results from the Strategic Transcatheter Evaluation of New Therapies (STENT) group. Presentation at the 2006 American College of Cardiology/i2 Summit Annual Scientific Session March 12, 2006, Atlanta, Georgia. Available at: Accessed February 4, Kuchulakanti PK, Chu WW, Torguson R, et al. Sirolimus-eluting stents versus paclitaxel-eluting stents in the treatment of coronary artery disease in patients with diabetes mellitus. Am J Cardiol 2006;98: Kandzari D, O Neill W, Strauss W. One-year outcomes from the TC-WYRE registry. TAXUS Express stent vs. Cypher stent: what s your real world experience? Paper presented at: 2006 Transcatheter Cardiovascular Therapeutics Annual Meeting; October 22 27, 2006; Washington, DC. 44. Mishkel G, Moore A, Slater C. Comparison of paclitaxel versus sirolimus drug eluting stents in patients with diabetes mellitus in a real-world setting. Paper presented at: 2006 Transcatheter Cardiovascular Therapeutics Annual Meeting; October 22 27, 2006; Washington, DC. 45. Stankovic G, Cosgrave J, Chieffo A, et al. Impact of sirolimus-eluting and paclitaxel-eluting stents on outcome in patients with diabetes mellitus and stenting in more than one coronary artery. Am J Cardiol 2006;98: Moses JW, Leon MB, Popma JJ. DIRECT: Direct stenting using the sirolimus- eluting Bx Velocity stent. Paper presented at: 2004 American College of Cardiology Annual Scientific Session; March 7 10, 2004; New Orleans, LA. 47. Berenguer A, Mainar V, Bordes P, Valencia J, Gomez S. Efficacy of sirolimus-eluting stents in diabetics with complex coronary lesions. Rev Esp Cardiol 2006;59: Hamm C, Schneider S, Senges J. German Cypher-Stent Registry. Paper presented at: 2004 Transcatheter Cardiovascular Therapeutics Annual Meeting; September 27 October 1, 2004; Washington, DC. 49. Macaya C, Garcia-Garcia HM, Colombo A, et al. One-year results of coronary revascularization in diabetic patients with multivessel coronary artery disease. Sirolimus stent vs. coronary artery bypass surgery and bare metal stent: insights from ARTS-II and ARTS-I. EuroIntervention 2006;2: Seabra-Gomes R. PORTugese prospective, non-randomised, multicenter Outcomes Trial of the Sirolimus eluting stent in small native coronary arteries of diabetic (PORTO I) and non-diabetic (PORTO II) patients. Paper presented at: 2005 Transcatheter Cardiovascular Therapeutics Annual Meeting; October 16 21, 2005; Washington, DC. 51. Niemela K, Chevalier B, Stoerger H, Bilodeau L, Brito F, Russell ME. The TAXUS paclitaxel-eluting stent in high-risk patients: global usage patterns and clinical outcomes in the Real-World MILE- STONE II registry. Am J Cardiol 2005;96 Suppl 7A:130H. 52. Lasala JM. First report of the TAXUS ARRIVE 1-Year outcomes. Paper presented at: 2005 Transcatheter Cardiovascular Therapeutics Annual Meeting; October 16 21, 2005; Washington, DC.

11 JACC Vol. 51, No. 25, 2008 June 24, 2008: Mahmud et al. DES in Diabetic Patients Lasala JM. Snapshot of DES use and outcomes in the US Program: 57. Wessely R, Schomig A, Kastrati A. Sirolimus and paclitaxel on ARRIVE Program. Presentation at the DES Summit at the 2006 polymer-based drug-eluting stents: similar but different. J Am Coll American College of Cardiology Annual Meeting. March 11, 2006, Cardiol 2006;47: Atlanta, Georgia. Available at: Accessed February 4, value of sirolimus- and paclitaxel-eluting stents over bare metal 58. Daemen J, Garcia-Garcia HM, Kukreja N, et al. The long-term 54. Kastrati A, Dibra A, Eberle S, et al. Sirolimus-eluting stents vs stents in patients with diabetes mellitus. Eur Heart J 2007;28:26 paclitaxel-eluting stents in patients with coronary artery disease: 32. meta-analysis of randomized trials. JAMA 2005;294: Kotani J, Awata M, Nanto S, et al. Incomplete neointimal coverage of 55. Rugrok P, Melkert R, Morel M. Does angiography six months after sirolimus eluting stents: angioscopic findings. J Am Coll Cardiol coronary intervention influence management and outcome? Benestent 2006;47: II Investigators. J Am Coll Cardiol 1999;34: Carter AJ, Bailey L, Devries J, Hubbard B. The effects of uncontrolled 56. Eisenberg MJ, Konnyu KJ. Review of randomized clinical trials of hyperglycemia on thrombosis and formation of neointima after coronary stent placement in a novel diabetic porcine model of restenosis. drug-eluting stents for the prevention of in-stent restenosis. Am J Cardiol 2006;98: Coron Artery Dis 2000;11:473 9.