Author + information
- Received October 2, 2008
- Revision received October 30, 2008
- Accepted November 3, 2008
- Published online March 17, 2009.
- Emmanouil S. Brilakis, MD, PhD*,†,* (, )
- Christopher Lichtenwalter, MD*,†,
- James A. de Lemos, MD*,†,
- Michele Roesle, RN*,
- Owen Obel, MD*,†,
- Donald Haagen, RCIS*,
- Bilal Saeed, MD‡,
- Chiranjeevi Gadiparthi, MD*,†,
- Joseph K. Bissett, MD§,
- Rajesh Sachdeva, MD§,
- Vassilios V. Voudris, MD, PhD∥,
- Panagiotis Karyofillis, MD∥,
- Biswajit Kar, MD¶,
- James Rossen, MD#,
- Panayotis Fasseas, MD**,
- Peter Berger, MD†† and
- Subhash Banerjee, MD*,†
- ↵*Reprint requests and correspondence:
Dr. Emmanouil S. Brilakis, Dallas VA Medical Center (111A), 4500 South Lancaster Road, Dallas, Texas 75216
Objectives The aim of this study was to compare the frequency of angiographic restenosis and clinical events between a paclitaxel-eluting stent (PES) and a similar bare-metal stent (BMS) in saphenous vein graft (SVG) lesions.
Background There are conflicting and mostly retrospective data on outcomes after drug-eluting stent implantation in SVGs.
Methods Patients requiring SVG lesion stenting were randomized to BMS or PES. The primary study end point was binary in-segment restenosis at 12-month follow-up quantitative coronary angiography. Secondary end points included death, myocardial infarction, ischemia-driven target vessel and lesion revascularization, and target vessel failure.
Results Eighty patients with 112 lesions in 88 SVGs were randomized to a BMS (39 patients, 43 grafts, 55 lesions) or PES (41 patients, 45 grafts, 57 lesions). Binary angiographic restenosis occurred in 51% of the BMS-treated lesions versus 9% of the PES-treated lesions (relative risk: 0.18; 95% confidence interval [CI]: 0.07 to 0.48, p < 0.0001). During a median follow-up of 1.5 years the PES patients had less target lesion revascularization (28% vs. 5%, hazard ratio: 0.38; 95% CI: 0.15 to 0.74, p = 0.003) and target vessel failure (46% vs. 22%, hazard ratio: 0.65; 95% CI: 0.42 to 0.96, p = 0.03), a trend toward less target vessel revascularization (31% vs. 15%, hazard ratio: 0.66; 95% CI: 0.39 to 1.05, p = 0.08) and myocardial infarction (31% vs. 15%, hazard ratio: 0.67; 95% CI: 0.40 to 1.08, p = 0.10), and similar mortality (5% vs. 12%, hazard ratio: 1.56; 95% CI: 0.72 to 4.11, p = 0.27).
Conclusions In SVG lesions, PES are associated with lower rates of angiographic restenosis and target vessel failure than BMS. (The SOS [Stenting Of Saphenous Vein Grafts] Randomized-Controlled Trial; NCT00247208)
- bare-metal stents
- coronary artery bypass graft surgery
- drug-eluting stents
- percutaneous coronary intervention
- saphenous vein grafts
Retrospective studies have provided conflicting results in the treatment of saphenous vein aortocoronary bypass graft (SVG) lesions (1). In the only published prospective randomized trial of drug-eluting stents (DES) versus bare-metal stents (BMS) in SVG lesions, a sirolimus-eluting stent (SES) had lower angiographic restenosis at 6 months but higher mortality during follow-up (2,3). No prospective trial has specifically examined the outcomes after paclitaxel-eluting stent (PES) implantation in SVGs.
We performed the multicenter, randomized SOS (Stenting Of Saphenous Vein Grafts) trial to compare the angiographic and clinical outcomes after implantation of a PES versus BMS in SVG lesions.
The SOS trial (NCT00247208) is a randomized, controlled, multicenter, prospective trial designed to test the hypothesis that implantation of a polymer-based PES (Taxus, Boston Scientific, Natick, Massachusetts) in SVG lesions is associated with a reduced 12-month binary angiographic restenosis rate compared with a BMS with similar design (Express2, Boston Scientific). The trial was single-blinded: patients and outcome assessors were blinded, but the interventional cardiologists and physicians treating the patients were not.
The study was performed at 5 clinical sites: Dallas Veterans Affairs Medical Center, Dallas, Texas (which also served as the coordinating center); Little Rock Veterans Affairs Medical Center, Little Rock, Arkansas; Michael E. DeBakey Veterans Affairs Medical Center, Houston, Texas; Iowa Veterans Affairs Medical Center, Iowa City, Iowa; and Onassis Cardiac Surgery Center, Athens, Greece. The study was approved by the institutional review board of each participating institution.
The inclusion criteria for the study were: 1) age ≥18 years; 2) 1 or more 50% to 99% de novo or restenotic lesions in an SVG that were between 2.5 and 4.0 mm in diameter; 3) need for percutaneous coronary intervention (PCI) in the opinion of the attending cardiologist; and 4) willingness to return for repeat graft angiography at 12 months and be contacted after 1, 6, 12, and 24 months for clinical follow-up.
Exclusion criteria were: 1) prior brachytherapy in the target vessel; 2) left ventricular ejection fraction <25%; 3) hemorrhagic diatheses; 4) contraindications or allergy to aspirin, thienopyridines, paclitaxel, or stainless steel; 5) history of anaphylaxis to iodinated contrast medium; 6) use of paclitaxel within 12 months before study entry or current use of colchicine; 7) serum creatinine level >2.0 mg/dl; 8) leukocyte count <3,500/mm3; 9) platelet count <100,000/mm3; 10) recent positive pregnancy test, breast-feeding, or possibility of a future pregnancy; and 11) coexisting conditions limiting life expectancy to <24 months or that could affect a patient's compliance with the protocol.
After eligible patients provided written informed consent, they were randomized in a 1:1 ratio to treatment with either PES or BMS. Randomization was done with sealed, opaque, sequentially numbered envelopes. The allocation schedule was based on computer-generated random numbers and stratified according to study center, with block sizes of 4 to 8, which varied randomly.
Before stenting, all patients received oral aspirin (325 mg daily) and oral clopidogrel (a loading dose of 300 to 600 mg) as soon as possible but within 24 h of the procedure. Patients receiving daily clopidogrel for ≥72 h before stenting were not required to receive a clopidogrel loading dose.
Anticoagulation during PCI was achieved with heparin or bivalirudin at the operator's discretion. If heparin was used, additional intravenous heparin boluses were administered for a target activated clotting time (measured by the Hemochron device) of >250 s. Heparin and bivalirudin were discontinued immediately after the procedure. The use of intravenous glycoprotein IIb/IIIa inhibitors was left to the discretion of the operator.
Balloon dilation before stent placement was not required. One or more stents of the assigned type were used to treat the target lesion and edge dissections of types B through E or for otherwise suboptimal results. The use of embolic protection devices (EPDs) and intravascular ultrasound were strongly encouraged. It was recommended that cardiac biomarkers be measured before the intervention and 8 h and 16 to 24 h after the procedure.
Aspirin was administered indefinitely after stenting. Clopidogrel was initially recommended for 6 months after PES placement and for at least 1 month after BMS placement. Since December 2006, a minimum of 1 year of clopidogrel was recommended after PES placement.
Clinical and angiographic follow-up
Patients were asked to return for repeat coronary angiography 12 months after enrollment and contacted by phone at 1, 6, 12, and 24 months after enrollment to assess whether any cardiovascular events had occurred. Hospital records and coronary angiograms were obtained to determine whether any clinical or angiographic end points had occurred.
Quantitative coronary angiography
Quantitative angiographic analyses were performed blinded to the patient's identity, type of stent used, outcome, and film sequence. The stents used in the study are angiographically indistinguishable. Matched projections of coronary angiograms obtained at baseline, on completion of the stenting procedure, and at follow-up coronary angiography at 12 months were analyzed with the CAAS automated edge detection system version 5.4. (Pie Medical Imaging, Maastricht, the Netherlands). The projection that best showed the stenosis in its tightest view was selected. The intraobserver and interobserver correlation coefficients of the quantitative coronary angiography measurements ranged from 0.90 to 0.97 and 0.92 to 0.99, respectively.
The minimal lumen diameter (MLD) and nearest normal reference diameter were measured in millimeters by using the catheter as a scaling factor. Percent stenosis was calculated as 100 × [1 − (MLD/reference vessel diameter)]. Acute gain was defined as the difference between the baseline MLD and the in-stent MLD after stent implantation. Late luminal loss was defined as the difference between the MLD at the end of the stenting procedure and at follow-up angiography. Quantitative angiographic measurements of the target lesion were obtained in the “in-stent” zone (including only the stent segment) and “in-segment” zone (the stented segment and the 5-mm proximal and distal to the stent).
For every study SVG, a degeneration score was calculated according to Coolong et al. (4), as an ordinal metric of the extent of lumen irregularities and ectasia (>20% of the reference normal segment) within the SVG, as follows: <25% of the total SVG length (SVG degeneration score: 0), 26% to 50% (SVG degeneration score: 1), 51% to 75% (SVG degeneration score: 2), or >75% (SVG degeneration score: 3).
The primary end point of the study was binary angiographic restenosis/lesion, defined as a stenosis of ≥50% of the MLD in the target SVG segment at 12-month angiographic follow-up. Comparison of angiographic restenosis was also performed for each saphenous vein graft and patient. If multiple lesions were treated in the same graft of the same patient, angiographic restenosis was considered to have occurred if any lesion developed in-segment restenosis.
Secondary end points included procedural success, death, myocardial infarction (MI), ischemia-driven target lesion revascularization (TLR) and target vessel revascularization (TVR), target vessel failure, overall major adverse cardiac events, cerebrovascular events, stent thrombosis, and intra-stent intimal hyperplasia accumulation at follow-up, as measured by intravascular ultrasound (the intravascular ultrasonography data have not yet been analyzed and are not included in this report).
Procedural success was defined as the achievement of residual in-stent stenosis of <30%, associated with Thrombolysis In Myocardial Infarction (TIMI) flow grade 3 (or at least similar flow to the baseline flow if the baseline flow was TIMI flow grade 1 or 2), in the absence of a dissection worse than D1, as assessed according to the National Heart, Lung, and Blood Institute classification, without the occurrence of death, MI, or repeat TLR during the index hospital stay (5). Post-PCI acute MI was defined as a post-procedural rise in creatine kinase-myocardial band (CK-MB) (measured after 8 h and 16 to 24 h from PCI) >3 times the upper limit of normal.
All deaths were considered to be cardiac unless an unequivocal noncardiac cause could be established. Myocardial infarction during follow-up was defined as typical rise and fall of troponin or CK-MB above the upper limit of normal, with either ischemic symptoms or electrocardiographic changes indicative of ischemia (ST-segment elevation or depression or development of pathologic Q waves) (6). Ischemia-driven TVR and TLR were defined as repeat PCI or coronary artery bypass grafting performed because of restenosis of the target vessel or lesion, respectively, in association with angina or objective evidence of myocardial ischemia.
Target vessel failure was defined as the composite end point of cardiac death, MI, and TVR. A composite end point of cardiac death, MI attributed to the target vessel, and TLR was also evaluated (device-oriented composite end point, as suggested by Cutlip et al. ). If an adverse event could not unequivocally be attributed to a nontarget vessel, the event was considered to represent target vessel failure. We also assessed the incidence of major adverse cardiac events defined as the composite of any death, any MI, or any coronary revascularization (similar to the patient-oriented composite end point suggested by Cutlip et al. ).
In the original protocol, stent thrombosis was defined as an angiographic thrombus within the stented vessel at a clinically driven angiographic restudy for documented myocardial ischemia. However, when the Academic Research Consortium (ARC) criteria were published after the initiation of the study, we included plans to perform an additional analysis of definite/probable stent thrombosis with the ARC definition (7,8). All events were reviewed and adjudicated by an independent clinical events committee blinded to treatment allocation.
Continuous variables were summarized as mean ± 1 SD and were compared between the 2 study groups with the Student ttest or the Wilcoxon 2-sample test. Discrete variables were presented as frequencies and group percentages and were compared with the likelihood-ratio chi-square or the Fisher exact test.
The primary study end point (binary angiographic restenosis at 12-month follow-up coronary angiography) was analyzed with the Fisher exact test. The incidence of death, TLR, TVR and non-TVR, target vessel failure, device-oriented composite end point, major adverse cardiac events, and stent thrombosis during the follow-up period were calculated with the Kaplan-Meier method, and the differences between the 2 study groups were assessed with the log-rank test. Cox proportional hazards methods were used to calculate the hazard ratios for the PES versus BMS groups for each of the aforementioned end points. All analyses were performed on an intention-to-treat principle with JMP 7.0.1 (SAS Institute, Cary, North Carolina) and Stata (Stata Corporation, College Station, Texas). A 2-sided p value <0.05 was considered statistically significant.
The sample size required for demonstrating a 66% reduction in binary angiographic restenosis (from 50% to 17%) at 12 months with a 2-sided test with an alpha error of 0.05 and a power of 80% was 31 patients/group. We aimed to enroll 40 patients/group to compensate for patients whose procedures were unsuccessful or were lost to follow-up.
Patient and procedural characteristics
Between 2005 and 2007, 80 patients were enrolled at 5 clinical sites and were randomized to a BMS (n = 39) or a PES (n = 41) (Fig. 1).The baseline and procedural characteristics of the 2 study groups were similar (Table 1).Primary stenting was used in most lesions (Table 1). A total of 124 stents were implanted, 64 in the BMS and 64 in the PES group. Procedural success was achieved in 77 patients (96%). One patient (PES group) required emergency coronary bypass surgery due to an entrapped Filterwire (Boston Scientific) and 2 patients (1 in the BMS and 1 in the PES group) developed stent thrombosis within the first 24 h after stent implantation. Two patients (1 in each study group) had an asymptomatic CK-MB rise after PCI. One patient (BMS group) had a femoral hematoma that did not require transfusion, and 1 patient (PES group) had a right external iliac artery dissection that was successfully stented. An EPD was used in 60 of 112 study lesions (54%). Filterwire (Boston Scientific) was used in 70% of EPD lesions, Spider (ev3, Plymouth, Minnesota) in 12%, Proxis (St. Jude, Maple Grove, Minnesota) in 12%, and Guardwire (Medtronic, Santa Rosa, California) was used in 7%.
The primary end point, binary angiographic restenosis, could be evaluated in 90 lesions. It could not be evaluated in 22 lesions (20%), 16 grafts (18%), and 14 patients (18%, 6 BMS and 8 PES patients) (Fig. 1). Angiographic restenosis occurred in 51% of the BMS-treated lesions versus 9% of the PES-treated lesions (relative risk: 0.18; 95% confidence interval [CI]: 0.07 to 0.48, p < 0.0001) (Table 2,Fig. 2).The pattern of restenosis was focal, intrastent, proliferative, and total occlusion in 9, 2, 2, and 11 lesions, respectively, in the BMS group and 2, 0, 0, and 2 lesions, respectively, in the PES group. Angiographic restenosis occurred in 20 of 33 BMS patients (61%) versus 4 of 33 PES patients (12%) (relative risk: 0.20; 95% CI: 0.08 to 0.52, p < 0.0001). Angiographic restenosis occurred in 21 of 37 grafts in the BMS group (57%) versus 4 of 35 grafts in the DES group (11%) (relative risk: 0.20; 95% CI: 0.08 to 0.53, p < 0.0001).
Clinical follow-up was available for all study patients for a median of 1.5 years (Table 3,Fig. 3).Seven patients died, 2 (5%) in the BMS and 5 (12%) in the PES group (p = 0.27) (Fig. 3). The cause of death in the BMS group was cardiac arrest in 1 patient and unknown in 1 patient. In the PES group death was due to MI in 1 patient and was noncardiac in the remaining 4 patients (lung cancer, small bowel obstruction, multiple strokes and pneumonia, and chronic obstructive pulmonary disease). Clopidogrel use was similar in the 2 groups (Table 3). Compared with BMS patients, PES patients had lower rates of TLR and target vessel failure (Table 3, Fig. 3). There were trends toward fewer MIs, TVRs, and ARC definite/probable stent thrombosis in PES patients (Table 3, Fig. 3). Stroke occurred in 1 patient (in the PES group) who subsequently died. In addition to the 2 acute stent thromboses, angiographically documented stent thrombosis occurred during follow-up in 3 patients, all in the BMS group after 11, 12, and 22 months, respectively, from stent implantation.
The most important finding of this trial is that PES reduce angiographic restenosis rates and target vessel failure in SVG lesions compared with BMS.
Drug-eluting stents have been shown to reduce restenosis in many lesion types and clinical syndromes. However, there is a paucity of prospective data on DES in SVG interventions, although BMS is associated with high restenosis rates in SVG lesions. The results of mostly retrospective observational series comparing DES with BMS in SVGs have been conflicting: 6 studies showed better results with DES (9–14) and 9 suggested no difference between BMS and DES (15–23).
The SOS trial is the first prospective, randomized, controlled trial of PES in SVG lesions and the first multicenter DES clinical trial of any kind in SVG lesions. Two prospective, randomized, single-center studies have reported comparisons of DES and BMS in SVG lesions. In a post hoc subgroup analysis of the BASKET (Basel Stent Kosten Effektivitats Trial) study, the subgroup of 47 patients who underwent SVG PCI derived greater benefit from DES implantation during a follow-up of 18 months (24). In the BASKET study, major adverse cardiac events occurred in 62% in the BMS group versus 21% in the DES group (p = 0.007), mainly due to a lower TVR rate in the DES group (46% vs. 18%, p = 0.04).
In the only published, prospective, randomized trial of DES versus BMS in SVG lesions, the RRISC (Reduction of Restenosis In Saphenous vein grafts with Cypher sirolimus-eluting stent) trial, an SES (Cypher, Cordis, Warren, New Jersey) was compared with a similar BMS, the BxVelocity stent (Cordis) (2,3). Overall, 75 patients were enrolled, and 96 lesions were treated in 80 SVGs. Angiographic follow-up after only 6 months revealed less in-segment restenosis in the SES arm (13.6% vs. 32.6%, p = 0.031). Repeat TVR within 6 months was performed in 5.3% of patients in the SES versus 27% in the BMS group (p = 0.012) (2). However, after a median of 32 months, mortality was higher in the SES group (29% vs. 0%, p = 0.001), and the early benefits of SES were lost: no difference was seen in the incidence of MI (18% SES vs. 5% BMS, p = 0.15) or TVR (34% SES vs. 38% BMS, p = 0.74) (3). Of the 11 deaths in the SES group, 1 was due to angiographically confirmed late stent thrombosis and 3 were sudden.
As in the RRISC trial, a significant reduction in late loss was observed in the DES arm of SOS. The in-stent late loss in the DES group was similar in the 2 studies (0.38 ± 0.51 mm in the RRISC trial, and 0.42 ± 0.57 mm in the SOS trial) but was less in the BMS group in the RRISC study (0.79 ± 0.66 mm) than in the SOS trial (1.29 ± 1.03 mm). Accordingly, angiographic restenosis occurred more frequently in the BMS arm of the SOS trial (51% vs. 33% in the RRISC trial). This was likely in part due to the performance of follow-up coronary angiography at 12 months in the SOS trial versus at 6 months in the RRISC trial. The higher prevalence of diabetes (44% in the SOS trial vs. 15% in the RRISC trial) and smoking (26% in the SOS trial vs. 8% in the RRISC trial) might have also played a role. In contrast to the RRISC study, where most restenoses were nonocclusive, nearly one-half of the restenotic lesions in the SOS trial were occlusive, in which repeat PCI is often not attempted, explaining in part why the difference in TVR did not reach statistical significance in the SOS trial.
The PES group in the SOS trial had less TLR without attenuation of benefit during follow-up, as was seen in the RRISC trial. Both the Express2 BMS used in SOS and the BxVelocity BMS (Cordis) used in the RRISC trial are thick strut stents (strut thickness 132 and 140 μm, respectively); therefore, BMS design is unlikely to explain the higher in-stent restenosis rates in the BMS arm of the SOS trial.
Mortality was similar in the BMS and DES arms of the SOS trial. Although total mortality rates were numerically higher in the PES arm (5 vs. 2), most deaths were noncardiac; only 2 cardiac deaths occurred in the BMS group, and 1 cardiac death occurred in the PES group. A trend toward a lower incidence of MI in the PES group should further allay fears of a late increase in cardiac deaths as a result of placement of a PES in an SVG. Larger randomized trials are needed to evaluate the impact of DES (and differences between PES and SES or other DES) on mortality, MI, and stent thrombosis.
Although the SOS trial is the largest and first multicenter trial comparing DES and BMS in SVGs, it is nonetheless limited by the relatively small number of patients enrolled and the losses to angiographic follow-up, although the loss to angiographic follow-up in the SOS trial (18%) was similar to that achieved in the pivotal trials of all 4 currently Food and Drug Administration–approved DES. The SOS trial was powered for detecting angiographic restenosis and was therefore underpowered to detect differences in clinical outcomes, although significant differences in clinical outcomes were observed between the study groups. All patients enrolled were male; even though sex differences have not surfaced in any of the other DES versus BMS trials, the results of this study might not apply to women. Although all repeat revascularizations were driven by clinical symptoms, the protocol-required angiographic follow-up might have increased rates of repeat coronary revascularizations in both study groups and particularly in the BMS arm, because late loss was so much greater in that arm. The study was single-blinded, although quantitative coronary angiography and the adjudication of the clinical events were performed blinded to stent allocation. The median follow-up of SOS patients was 1.5 years; it is unknown whether the outcomes would change with longer-term follow-up. The outcomes seen with PES might not apply to other DES. Although the post-PCI MI rates were low in our study and the use of EPDs was similar in both study groups, more frequent use might have led to better outcomes in both groups.
The use of PES in SVG lesions is associated with lower rates of angiographic restenosis and target vessel failure than BMS. Large, prospective, multicenter, randomized-controlled clinical trials that use a clinical rather than angiographic end point are needed to confirm the beneficial role of DES in SVG lesions (25).
The SOS (Stenting Of Saphenous Vein Grafts) trial was funded by a Veterans Affairs VISN-17 Startup Award, and by the Clark R. Gregg fund of the Harris Methodist Foundation to Dr. Brilakis. Dr. Brilakis has received speaker honoraria from St. Jude. Dr. Obel works predominantly with cardial rhythm devices and has speaker agreements with St. Jude, Medtronic, and Boston Scientific. Dr. Rossen participated in multicenter clinical studies supported by Boston Scientific. Dr. Berger has spoken at CME-approved scientific symposia supported by Bristol-Myers Squibb/Sanofi-Aventis, the Medicines Company, AstraZeneca, Medtronic, and Eli Lilly/Daiichi-Sankyo (each for <$10,000); served as a consultant to PlaCor, Eli Lilly, Accumetrics, The Medicines Company, and Eli Lilly/Daiichi-Sankyo (each for <$10,000); and owns equity in Lumen, Inc. (a company that is developing an embolic protection device) (>$10,000). Dr. Banerjee has served on the Speakers' Bureau for St. Jude Medical Center, Medtronic Corp., and Johnson & Johnson and has received a research grant from Boston Scientific. The SOS trial was presented as a late-breaking trial at the TCT 2008 meeting in Washington, DC, in October 2008.
- Abbreviations and Acronyms
- Academic Research Consortium
- bare-metal stent(s)
- confidence interval
- creatine kinase-myocardial band
- drug-eluting stent(s)
- embolic protection device
- myocardial infarction
- minimal lumen diameter
- percutaneous coronary intervention
- paclitaxel-eluting stent(s)
- sirolimus-eluting stent(s)
- saphenous vein graft
- target lesion revascularization
- target vessel revascularization
- Received October 2, 2008.
- Revision received October 30, 2008.
- Accepted November 3, 2008.
- American College of Cardiology Foundation
- Vermeersch P.,
- Agostoni P.,
- Verheye S.,
- et al.
- Vermeersch P.,
- Agostoni P.,
- Verheye S.,
- et al.
- Coolong A.,
- Baim D.S.,
- Kuntz R.E.,
- et al.
- Smith S.C. Jr..,
- Feldman T.E.,
- Hirshfeld J.W. Jr..,
- et al.
- Antman E.,
- Bassand J.-P.,
- Klein W.,
- et al.
- Cutlip D.E.,
- Windecker S.,
- Mehran R.,
- et al.
- Ge L.,
- Iakovou I.,
- Sangiorgi G.M.,
- et al.
- Hoffmann R.,
- Pohl T.,
- Koster R.,
- Blindt R.,
- Boeckstegers P.,
- Heitzer T.
- Brodie B.R.,
- Stuckey T.,
- Downey W.,
- et al.