Drug-eluting versus bare-metal stents for treating saphenous vein grafts

Congestive Heart Failure

Drug-eluting versus bare-metal stents for treating saphenous vein grafts Mehdi H. Shishehbor, DO, MPH, a Riem Hawi, MD, b Inder M. Singh, MD, MS, c E. Murat Tuzcu, MD, a Deepak L. Bhatt, MD, MPH, d Stephen G. Ellis, MD, a and Samir R. Kapadia, MD a Cleveland, OH; Indianapolis, IN; and Boston, MA

Background Current data show conflicting results regarding safety and efficacy of drug-eluting stents (DES) versus baremetal stents (BMS) for treating saphenous vein grafts (SVG). Our objective was to compare DES with BMS for SVG intervention. Methods Patients undergoing stenting with DES or BMS to SVG from January 2000 to June 2007 were included. To eliminate any unobserved bias regarding stent selection, the BMS cohort was divided into pre- and post-2003 when DES became available. Adjusted Cox analysis compared DES with pre- and post-2003 BMS patients. The primary end point was a composite of all-cause mortality, myocardial infarction, or target lesion revascularization. Results Of the total 566 patients, 217 (38%) received DES, 110 (20%) received BMS post-2003, and 239 (42%) received BMS pre-2003. Median follow-up was 2.9 years (interquartile range 1.4-4.9 years). There was a trend toward lower primary end point with DES compared to post-2003 BMS (91 events, adjusted hazard ratio 0.61, 95% CI 0.35-1.07, P = .08). However, despite 179 events, DES use was not associated with lower primary end point compared with pre-2003 BMS (adjusted hazard ratio 0.61, 95% CI 0.28-1.35, P = .23). Conclusions Although DES showed a trend toward a lower primary end point when compared with BMS post-2003, this association was no longer present when DES was compared to pre-2003 BMS. These results are consistent with the preponderance of available data and indicate that unobserved bias in observational registries may explain the reported benefit of DES over BMS for treating SVG. (Am Heart J 2009;158:637-43.)

In randomized clinical trials, drug-eluting stents (DES) have been shown to significantly reduce the future rate of revascularization compared with bare-metal stents (BMS) when treating native coronaries.1 However, the safety and efficacy of DES in treating saphenous vein grafts (SVG) remains controversial.2-4 To date, 2 randomized clinical trials, one with sirolimus-eluting stents (SES) and the other with paclitaxel-eluting stents (PES), have been

From the aDepartment of Cardiovascular Medicine, Cleveland Clinic, Cleveland, OH, b Department of Internal Medicine, Cleveland Clinic, Cleveland, OH, cDivision of Cardiovascular Medicine, Indiana University School of Medicine, Indianapolis, IN, and d Veterans Affairs Boston Healthcare System and Brigham and Women's Hospital, Boston, MA. Dr Shishehbor is supported in part by the National Institutes of Health, National Institute of Child Health and Human Development, Multidisciplinary Clinical Research Career Development Programs grant K12 HD049091, and the National Institutes of Health Loan Repayment Program. Submitted May 9, 2009; accepted August 6, 2009. Reprint requests: Samir R. Kapadia, MD, 9500 Euclid Ave, F-25, Cleveland Clinic, Cleveland, OH 44114. E-mail: [email protected] 0002-8703/$ - see front matter © 2009, Mosby, Inc. All rights reserved. doi:10.1016/j.ahj.2009.08.003

published with a total of 155 patients.5 In one study, at 32 months, SES did not decrease repeat revascularization and was associated with increased mortality compared with BMS.6 In another, at 18 months, PES decreased ischemia-driven target lesion revascularization (TLR) but at the cost of a trend toward higher all-cause mortality when compared to BMS.7 Similarly, small observational studies have provided conflicting information on the utility of DES in SVG and many of these studies have lacked adequate power.8 Despite these shortcomings, DES are routinely used to treat SVG in real-world settings.1,9 Therefore, we sought to compare the safety and efficacy of DES versus BMS in patients undergoing percutaneous coronary intervention (PCI) of SVG, while accounting for unobserved selection bias that may have occurred in observational studies.

Methods Patient population Patients undergoing PCI to SVG from January 2000 through June 2007 were included. Because of the observational nature of the current study and to eliminate selection bias, we categorized our cohort into 3 groups: patients who received a BMS pre-2003

American Heart Journal October 2009

638 Shishehbor et al

Table I. Baseline and procedural characteristics based on stent type Characteristics Age (y) Male (%) Body mass index Left ventricular ejection fraction Heart rate Years since coronary artery bypass surgery Risk factors (%) Family history of early coronary disease Cigarette smoking Diabetes mellitus Insulin-dependent diabetes mellitus Non–insulin-dependent diabetes mellitus Medical history (%) Prior MI Peripheral arterial disease Stroke or transient ischemic attack Chronic obstructive pulmonary disease New York Heart Association class (%) 3 4 Medications at discharge (%) Aspirin Clopidogrel Heparin Glycoprotein IIb/IIIa inhibitors β-Blockers Angiotensin-converting enzyme inhibitors Statins Clinical presentation (%) Acute MI Unstable angina Location of culprit lesion (%) Graft to the left anterior descending artery Graft to the left circumflex artery Graft to the right coronary artery Angiographic characteristics Reference vessel diameter (mm) Lesion length (mm) Multivessel intervention American College of Cardiology lesion score A B1 B2 C Embolic protection device Procedural success (%) Emergency coronary artery bypass surgery

BMS pre-2003 (n = 239)

BMS post-2003 (n = 110)

DES (n = 217)

69 ± 9 184 (77) 29 ± 5 48 ± 11 71 ± 14 9±5

68 ± 10 84 (76) 29 ± 5 49 ± 12 71 ± 14 10 ± 6

70 ± 10 163 (75) 30 ± 6 48 ± 12 69 ± 14 10 ± 6

83 (35) 25 (10) 86 (36) 28 (12) 56 (23)

31 11 42 14 27

(28) (10) (38) (13) (25)

61 (28) 23 (11) 92 (42) 30 (14) 59 (27)

124 (52) 37 (16) 48 (20) 21 (9)

56 25 20 16

(51) (23) (18) (15)

114 (53) 43 (20) 34 (16) 33 (15)

6 (3) 2 (1)

23 (21) 16 (15)

41 (19) 31 (14)

239 (100) 239 (100) 99 (41) 179 (75) 68 (28) 91 (38) 105 (44)

110 (100) 110 (100) 45 (41) 39 (35) 89(81) 51 (46) 86 (78)

217 (100) 217 (100) 72 (33) 60 (28) 187 (86) 99 (46) 176 (81)

10 (4) 154 (64)

8 (7) 56 (51)

7 (3) 105 (48)

17 (7) 116 (49) 106 (44)

8 (7) 45 (41) 57 (52)

25 (11) 90 (42) 102 (47)

3.4 ± 0.8 13 ± 7 58 (24)

3.9 ± 0.8 15 ± 7 9 (8)

3.2 ± 0.5 16 ± 8 72 (33)

26 (11) 38 (16) 32 (13) 143 (60) 38 (16) 229 (96) 0 (0)

1 (1) 3 (3) 19 (17) 87 (79) 68 (62) 104 (95) 0 (0)

1 (1) 13 (6) 27 (12) 176 (81) 122 (56) 212 (98) 1 (1)

(before DES were commercially available); patients who received BMS post-2003 (after DES were commercially available); and patients who received DES. By comparing those who received DES with those who received a BMS pre-2003, we aimed to eliminate any unobserved selection bias regarding stent use. Baseline clinical and procedural data were prospectively obtained and recorded by trained research coordinators. No extramural funding was used to support this work. The authors are solely responsible for the design and conduct of this study, all study analyses, and the drafting and editing of the paper and its final contents. The institutional review board

waived requirements for informed consent for the institutional PCI registry.

Procedural characteristics The index procedure was defined as PCI of a new SVG stenosis. As part of the institutional registry, information regarding pharmacotherapy, important angiographic features, and procedural characteristics were captured. Pre-2003 only BMS were used, but post-2003, the choice of stent type (DES vs BMS) was at the discretion of the operator

American Heart Journal Volume 158, Number 4

Shishehbor et al 639

Figure 1

Kaplan-Meier curves of DES versus pre- and post-2003 BMS cohorts for composite end point of death, MI, or TLR.

performing the procedure as was the use of distal embolic protection devices (EPDs).

Clinical end points The primary end point was a composite of all-cause mortality, myocardial infarction (MI), and TLR. Secondary end points were the individual components of the primary end point. Myocardial infarction was defined as occurrence of troponin elevation with electrocardiographic changes or angina. Patients were prospectively followed for adverse cardiovascular events by research coordinators through hospital record review, telephone contact, and the US Social Security Death Index. In addition, detailed chart review was performed on all patients (MHS, RH) to obtain data on graft's age, TLR, and to ensure the accuracy of all end points.

Statistical methods The study population was categorized into 3 prespecified groups. Differences in clinical outcomes among the 3 prespecified groups were tested via unadjusted Kaplan-Meier curves and adjusted multivariable Cox proportional hazards modeling that accounted for baseline demographic features, angiographic variables, treatment assignment, and other confounders. Specifically, each model was adjusted for age, gender, body mass index, ejection fraction, heart rate, blood pressure, history of previous MI, stroke, transient ischemic attack, and presence of diabetes. In addition, all models were adjusted for medications listed in Table I including statins, β-blockers, angiotensinconverting enzymes inhibitors, aspirin, clopidogrel, heparin,

and glycoprotein IIb/IIIa inhibitors. In addition to the use of EPD and years since bypass surgery, all models were also adjusted for important angiographic variables such as those listed in Table I. In total, all models were adjusted for 30 confounding variables as listed in Table I. To account for advances in PCI over time, all multivariable models were adjusted for the procedural date. Selected subgroups were chosen for further analysis because they have been previously reported to gain the most benefit from DES. These were diabetes, vessel size, and lesion length. Both post-2003 and pre-2003 BMS were compared with DES for each of these selected subgroups using multivariable adjusted Cox analyses (version 9.1, SAS Institute, Cary, NC).

Results Baseline characteristics A total of 588 patients with previous SVG underwent PCI with either BMS or DES. Twenty-two patients were excluded because they received both SES and PES on the day of intervention. Of the 566 remaining patients, 239 (42%) received BMS pre-2003, 110 (20%) received BMS post-2003, and 217 (38%) received DES. Median follow-up for death was 2.9 years (interquartile range 1.4-4.9 years). Baseline and target lesion characteristics of the 3 groups are presented in Table I. In general, patient and procedural characteristics were similar between both groups. However, patients receiving PCI post-2003 (for both the BMS and DES groups) had more chronic obstructive pulmonary disease and worse New York Heart Association functional

American Heart Journal October 2009

640 Shishehbor et al

Table II. Unadjusted and multivariable adjusted HRs for the primary and secondary end points DES, events (%)

BMS, events (%)

DES vs post-2003 BMS cohort Total population (N = 327) Composite of death, MI or TLR (n = 91) Unadjusted Adjusted All-cause mortality (n = 51) Unadjusted Adjusted Myocardial infarction (n = 23) Unadjusted Adjusted TLR (n = 36) Unadjusted Adjusted Target vessel failure (n = 42) Unadjusted Adjusted

n = 217 58 (27) – – 28 (13) – – 16 (7) – – 29 (13) – – 34 (16) – –

n = 110 33 (30) – – 112 (16) – – 7 (6) – – 7 (6) – – 8 (7) – –

DES vs pre-2003 BMS cohort Total population (n = 456) Composite of death, MI, or TLR (n = 179) Unadjusted Adjusted All-cause mortality (n = 112) Unadjusted Adjusted Myocardial infarction (n = 36) Unadjusted Adjusted TLR (n = 65) Unadjusted Adjusted Target vessel failure (n = 79) Unadjusted Adjusted

n = 217 58 (27) – – 28 (13) – – 16 (7) – – 29 (13) – – 34 (16) – –

n = 239 121 (51) – – 84 (35) – – 20 (8) – – 36 (15) – – 45 (19) – –

class. In the post-2003 era, the use of heparin and glycoprotein IIB/IIIA inhibitors was lower compared to pre-2003 BMS, but the use of cardioprotective medications (β-blockers and statins) was higher. In addition, embolic protection devices were more often used in the post-2003 groups (both BMS and DES).

Clinical outcomes Drug-eluting stents versus post-2003 BMS. A cumulative 91 events (death, MI, or TLR) occurred in this cohort (Figure 1). Drug-eluting stents were associated with a trend toward a lower composite end point compared with those who received BMS post-2003 (adjusted hazard ratio [HR] 0.61, 95% CI 0.35-1.07, P = .08) (Table II). This benefit seen with drug-eluting stent implantation was mainly driven by the lower all-cause mortality (adjusted HR 0.49, 95% CI 0.20-1.16, P = .10) (Table II). Importantly, no significant difference in MI or TLR was seen when comparing DES versus those who received BMS post-2003 (Table II).

DES vs BMS, HR (95% CI)

P

– 0.89 (0.58-1.36) 0.61 (0.35-1.07) – 0.61 (0.35-1.06) 0.49 (0.20-1.16) – 1.09 (0.45-2.65) 0.48 (0.13-1.79) – 2.14 (0.94-4.88) 1.26 (0.43-3.67)

– .59 .08 – .08 .10 – .85 .27 – .07 .67

2.22 (1.03-4.81) 1.54 (0.57-4.21)

.04 .60

– 0.61 (0.44-0.84) 0.75 (0.42–1.33) – 0.72 (0.46-1.41) 0.72 (0.32-1.62) – 1.62 (0.77-3.38) 1.18 (0.28–5.01) – 1.07 (0.65-1.77) 0.83 (0.32-2.19)

– .002 .32 – .16 .43 – .20 .82 – .78 .71

0.98 (0.62-1.55) 0.84 (0.28-2.56)

.95 .76

Drug-eluting stents versus pre-2003 BMS. Because stent selection was not randomized, we compared the outcomes of those who received DES with individuals who received BMS pre-2003 when DES were not available (Figure 1). This would potentially eliminate any selection bias against BMS that could have occurred post-2003. Despite more events (n = 179), DES were not associated with lower composite end point of death, MI, or TLR when compared to those who received BMS pre-2003 (27% vs 51%, adjusted HR 0.61, 95% CI 0.28-1.35, P = .23). The secondary end points were also not significantly different when comparing DES versus those who received BMS pre-2003 (Table II). Subgroup analysis. Patients with diabetes and those without had better outcomes with DES compared to BMS in the post-2003 era (Table III). However, when DES use was compared with pre-2003 BMS, only patients with diabetes benefited (Table III). Patients with vessel size b3.5 mm benefited from DES both in the pre-2003 and post-2003 analysis (Table III), whereas this benefit was

American Heart Journal Volume 158, Number 4

Shishehbor et al 641

Table III. Multivariable adjusted HRs for composite of death, MI, or TLR in selected subgroups*

DES vs post-2003 BMS cohort (n = 237) Diabetes mellitus Present (n = 134) Absent (n =193) Vessel diameter (mm) ≥3.5 (n = 160) b3.5 (n = 167) Lesion length (mm) ≥15 (n = 148) b15 (n = 179 ) DES vs pre-2003 BMS cohort (n = 456) Diabetes mellitus Present (n = 178) Absence (n = 278) Vessel diameter (mm) ≥3.5 (n = 173) b3.5 (n = 283) Lesion length (mm) ≥15 (n = 278) b15 (n = 178)

DES, events (%)

BMS, events (%)

DES vs BMS HR (95% CI)

P

28 (30) 30 (24)

14 (33) 19 (28)

0.12 (0.04-0.43) 0.29 (0.12-0.72)

b.001 .007

20 (25) 38 (28)

18 (22) 15 (50)

0.90 (0.34-2.38) 0.39 (0.16-0.92)

.82 .03

28 (27) 30 (27)

14 (33) 19 (28)

0.43 (0.14-1.33) 0.43 (0.18-1.00)

.14 .05

28 (30) 30 (24)

50 (58) 71 (46)

0.10 (0.02-0.45) 1.14 (0.41-3.19)

.003 .80

20 (25) 38 (28)

52 (56) 69 (47)

1.28 (0.26-6.46) 0.36 (0.12-1.05)

.76 .06

28 (27) 30 (27)

31 (42) 90 (54)

2.02 (0.46-8.90) 0.20 (0.06-0.67)

.35 .009

* Number in parenthesis represents the number of events in each subgroup divided by the number of patients receiving that particular stent in that subgroup.

absent in those with a vessel size N3.5 mm. In the pre-2003 and post-2003 cohorts, patients with lesion lengths b15 mm benefited from DES, but there was no difference in longer lesions (Table III).

Discussion We compared the safety and efficacy of DES to BMS in a large cohort of consecutive, unselected patients who underwent PCI of SVG. Our results showed that when compared to post-2003 BMS, DES had a trend toward a reduction in the primary end point of death, MI, or TLR, which was mainly driven by lower mortality. However, DES were not associated with a lower rate of death, MI, or TLR when compared with a cohort of patients who received BMS pre-2003 despite more events and greater statistical power. Importantly, TLR was not lower with DES use in either era (Table II). Consistent with data from the non-SVG population, our subgroup analysis indicate that patients with diabetes and smaller vessels benefit from DES.1 Saphenous vein graft intervention accounts for 15% of all PCI's.2,10,11 However, optimal treatment of SVG disease continues to be elusive with lack of benefit of balloon angioplasty or covered stents and up to 30% restenosis rates with BMS.2,3,11-13 To address the issue of restenosis and repeat revascularization, DES use has gained popularity as an “off-label” indication.1,9 However, recent concerns with the safety of DES14,15 led to a secondary post hoc analysis of the first randomized clinical trial of SES in the treatment of SVG.6 This showed that the initial benefit of SES on repeat

revascularization was lost because of a “late catch-up phenomenon” and SES was associated with higher mortality at 32 months, although most events were noncardiac death.6 More recently, a second dedicated randomized trial evaluated PES in 80 patients undergoing PCI for diseased SVG.7 The primary end point of binary angiographic restenosis and a secondary end point of ischemia-driven TLR both favored PES (P b .0001 and P = .003, respectively).7 There were 34 cumulative major cardiac adverse events with no significant difference between DES or BMS (P = .20).7 Similarly, there was no benefit for the hard end points of death or MI (P = .34).7 As with these 2 relatively underpowered randomized controlled trials to detect clinical benefit,6,7 conflicting data have plagued observational studies with 8 studies showing improved outcomes with DES16-23 and 10 suggesting no difference between BMS and DES.24-33 Observational studies showing superiority of DES included a total of 869 patients for a mean follow-up of 13.4 months, whereas those showing no significant difference between BMS and DES included 2558 patients for a mean follow-up of 24.1 months. It is not surprising that the longer followup was less favorable for DES given “late catch-up” restenosis seen with DES in SVG.6 However, as a group, these observational studies have been small with inadequate power (10 with b150 patients).16-21,31,33 Our study represents the largest cohort of patients with SVG disease treated with BMS versus DES, and our results indicate that unobserved biases in observational registries may explain the reported benefit of DES over BMS for treating SVG. These findings are important not

American Heart Journal October 2009

642 Shishehbor et al

only because of previously raised long-term safety issues with DES but also because of associated cost and need for prolonged dual antiplatelet therapy, which has inherent shortcomings.34 Pathophysiologic explanations for lack of consistent benefit of DES over BMS in SVG are not entirely clear. Thrombosis due to large burden of friable plaque or impaired endothelialization, intimal hyperplasia causing aggressive in-stent restenosis, and malignant atherosclerosis have been proposed as the events leading to graft failure in the acute, subacute, and late post-intervention periods, respectively.2,3,9,35 In addition, the long-term prognosis of patients with diseased SVG is influenced by progression of disease in non-intervened SVG segments, thus further reducing the local benefit of DES over BMS, if any.36,37 Given the aforementioned issues, there is an urgent need for an adequately powered randomized controlled trial to conclusively address the safety and efficacy of DES versus BMS for treatment of SVG disease. An ideal study would be powered for clinical end points instead of protocol-driven follow-up angiography, which is known to exaggerate the benefits of DES over BMS. Furthermore, the ideal trial would use “thin strut” BMS and possibly a similar duration of thienopyridine in both arms as both of these strategies have established efficacy.38 Clinical follow-up of at least 2 years would also clarify the phenomenon of “late catch-up’’ restenosis that has hindered the superiority of DES in this population. Our study had several limitations. It is an observational study on prospectively collected data and thus is prone to bias by unmeasured confounders. However, we made every effort to capture many variables that could potentially confound this analysis (Table I). Practice changed after the advent of EPD and also after publication of the early beneficial results of SES in SVG. However, the use of EPD was greater for the DES cohort compared to BMS pre-2003, which would have biased the results in favor of DES. The use of cardioprotecive medications such as β-blockers and statins was higher in the post-2003 cohort, but this should have also favored the DES cohort. Data regarding duration of dual antiplatelet therapy with aspirin and clopidogrel were not available; however, in our institution, we typically encourage dual antiplatelet therapy for at least 1 year for all patients undergoing DES intervention. Lastly, we do not have formal data on stent thrombosis which may be important in complex “offlabel” procedures.

Conclusion In this large cohort, DES were associated with improved clinical outcomes when compared with contemporary patients treated with BMS post-2003. However, this association was no longer present when DES use was compared with those that received a BMS pre-2003. These

results indicate that unobserved biases in observational registries may explain the reported benefit of DES over BMS for treating SVG, and potentially in other poorly studied lesion subsets as well. In addition, our results may aid in elucidating the conflicting results of the only 2 randomized controlled trials comparing DES with BMS in percutaneous treatment of diseased saphenous vein grafts. Adequately powered and clinical end point–based randomized controlled trials are urgently needed to clarify the optimal treatment strategy in this high risk population.

References 1. King III SB, Smith Jr SC, Hirshfeld Jr JW, et al. 2007 focused update of the ACC/AHA/SCAI 2005 guideline update for percutaneous coronary intervention: a report of the American College of Cardiology/American Heart Association Task Force on Practice guidelines. J Am Coll Cardiol 2008;51:172-209. 2. Baim DS. Percutaneous treatment of saphenous vein graft disease: the ongoing challenge. J Am Coll Cardiol 2003;42:1370-2. 3. de Feyter PJ. Percutaneous treatment of saphenous vein bypass graft obstructions: a continuing obstinate problem. Circulation 2003;107: 2284-6. 4. Brilakis ES, Saeed B, Banerjee S. Use of drug-eluting stents in saphenous vein aortocoronary bypass graft lesions: a critical appraisal. J Interv Cardiol 2008;21:151-7. 5. Vermeersch P, Agostoni P, Verheye S, et al. Randomized double-blind comparison of sirolimus-eluting stent versus bare-metal stent implantation in diseased saphenous vein grafts: six-month angiographic, intravascular ultrasound, and clinical follow-up of the RRISC Trial. J Am Coll Cardiol 2006;48:2423-31. 6. Vermeersch P, Agostoni P, Verheye S, et al. Increased late mortality after sirolimus-eluting stents versus bare-metal stents in diseased saphenous vein grafts: results from the randomized DELAYED RRISC Trial. J Am Coll Cardiol 2007;50:261-7. 7. Brilakis ES, Lichtenwalter C, de Lemos JA, et al. A randomized controlled trial of a paclitaxel-eluting stent versus a similar bare-metal stent in saphenous vein graft lesions the SOS (Stenting of Saphenous Vein Grafts) trial. J Am Coll Cardiol 2009;53:919-28. 8. Blankenship JC. Parsing the subsets: when small studies create confusion. Catheter Cardiovasc Interv 2008;72:23-4. 9. Colombo A, Chieffo A. Drug-eluting stent update 2007: part III: technique and unapproved/unsettled indications (left main, bifurcations, chronic total occlusions, small vessels and long lesions, saphenous vein grafts, acute myocardial infarctions, and multivessel disease). Circulation 2007;116:1424-32. 10. Morrison DA, Sethi G, Sacks J, et al. Percutaneous coronary intervention versus repeat bypass surgery for patients with medically refractory myocardial ischemia: AWESOME randomized trial and registry experience with post-CABG patients. J Am Coll Cardiol 2002; 40:1951-4. 11. Savage MP, Douglas Jr JS, Fischman DL, et al. Stent placement compared with balloon angioplasty for obstructed coronary bypass grafts. Saphenous Vein De Novo Trial Investigators. N Engl J Med 1997;337:740-7. 12. Brilakis ES, Berger PB. Should bare metal or drug-eluting stents be used during PCI of saphenous vein graft lesions: waiting for Godot? Catheter Cardiovasc Interv 2008;72:815-8. 13. Hanekamp CE, Koolen JJ, Den Heijer P, et al. Randomized study to compare balloon angioplasty and elective stent implantation in venous

American Heart Journal Volume 158, Number 4

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

bypass grafts: the Venestent study. Catheter Cardiovasc Interv 2003; 60:452-7. Lagerqvist B, James SK, Stenestrand U, et al. Long-term outcomes with drug-eluting stents versus bare-metal stents in Sweden. N Engl J Med 2007;356:1009-19. Pfisterer M, Brunner-La Rocca HP, Buser PT, et al. Late clinical events after clopidogrel discontinuation may limit the benefit of drug-eluting stents: an observational study of drug-eluting versus bare-metal stents. J Am Coll Cardiol 2006;48:2584-91. Jeger RV, Schneiter S, Kaiser C, et al. Drug-eluting stents compared with bare metal stents improve late outcome after saphenous vein graft but not after large native vessel interventions. Cardiology 2009;112: 49-55. Assali A, Raz Y, Vaknin-Assa H, et al. Beneficial 2-years results of drug-eluting stents in saphenous vein graft lesions. EuroIntervention 2008;4:108-14. Kaplan S, Barlis P, Kiris A, et al. Immediate procedural and long-term clinical outcomes following drug-eluting stent implantation to ostial saphenous vein graft lesions. Acute Card Care 2008;10:88-92. Minutello RM, Bhagan S, Sharma A, et al. Long-term clinical benefit of sirolimus-eluting stents compared to bare metal stents in the treatment of saphenous vein graft disease. J Interv Cardiol 2007;20: 458-65. Hoffmann R, Pohl T, Koster R, et al. Implantation of paclitaxel-eluting stents in saphenous vein grafts: clinical and angiographic follow-up results from a multicentre study. Heart 2007;93:331-4. Wohrle J, Nusser T, Kestler HA, et al. Comparison of the slow-release polymer-based paclitaxel-eluting Taxus-Express stent with the baremetal Express stent for saphenous vein graft interventions. Clin Res Cardiol 2007;96:70-6. Lee MS, Shah AP, Aragon J, et al. Drug-eluting stenting is superior to bare metal stenting in saphenous vein grafts. Catheter Cardiovasc Interv 2005;66:507-11. Ge L, Iakovou I, Sangiorgi GM, et al. Treatment of saphenous vein graft lesions with drug-eluting stents: immediate and midterm outcome. J Am Coll Cardiol 2005;45:989-94. Lozano I, Garcia-Camarero T, Carrillo P, et al. Comparison of drugeluting and bare metal stents in saphenous vein grafts. Immediate and long-term results. Rev Esp Cardiol 2009;62:39-47. van Twisk PH, Daemen J, Kukreja N, et al. Four-year safety and efficacy of the unrestricted use of sirolimus- and paclitaxel-eluting stents in coronary artery bypass grafts. EuroIntervention 2008;4: 311-7.

Shishehbor et al 643

26. Okabe T, Lindsay J, Buch AN, et al. Drug-eluting stents versus bare metal stents for narrowing in saphenous vein grafts. Am J Cardiol 2008;102:530-4. 27. Gioia G, Benassi A, Mohendra R, et al. Lack of clinical long-term benefit with the use of a drug eluting stent compared to use of a bare metal stent in saphenous vein grafts. Catheter Cardiovasc Interv 2008; 72:13-20. 28. Applegate RJ, Sacrinty M, Kutcher M, et al. Late outcomes of drugeluting versus bare metal stents in saphenous vein grafts: propensity score analysis. Catheter Cardiovasc Interv 2008;72:7-12. 29. Ramana RK, Ronan A, Cohoon K, et al. Long-term clinical outcomes of real-world experience using sirolimus-eluting stents in saphenous vein graft disease. Catheter Cardiovasc Interv 2008;71:886-93. 30. Vignali L, Saia F, Manari A, et al. Long-term outcomes with drugeluting stents versus bare metal stents in the treatment of saphenous vein graft disease (results from the REgistro Regionale AngiopLastiche Emilia-Romagna registry). Am J Cardiol 2008;101:947-52. 31. Bansal D, Muppidi R, Singla S, et al. Percutaneous intervention on the saphenous vein bypass grafts—long-term outcomes. Catheter Cardiovasc Interv 2008;71:58-61. 32. Ellis SG, Kandzari D, Kereiakes DJ, et al. Utility of sirolimus-eluting Cypher stents to reduce 12-month target vessel revascularization in saphenous vein graft stenoses: results of a multicenter 350-patient case-control study. J Invasive Cardiol 2007;19:404-9. 33. Chu WW, Rha SW, Kuchulakanti PK, et al. Efficacy of sirolimuseluting stents compared with bare metal stents for saphenous vein graft intervention. Am J Cardiol 2006;97:34-7. 34. Holmes Jr DR, Kereiakes DJ, Laskey WK, et al. Thrombosis and drugeluting stents: an objective appraisal. J Am Coll Cardiol 2007;50: 109-18. 35. Christiansen JF, Hartwig D, Bechtel JF, et al. Diseased vein grafts express elevated inflammatory cytokine levels compared with atherosclerotic coronary arteries. Ann Thorac Surg 2004;77:1575-9. 36. Ellis SG, Brener SJ, DeLuca S, et al. Late myocardial ischemic events after saphenous vein graft intervention—importance of initially “nonsignificant” vein graft lesions. Am J Cardiol 1997;79:1460-4. 37. Keeley EC, Velez CA, O'Neill WW, et al. Long-term clinical outcome and predictors of major adverse cardiac events after percutaneous interventions on saphenous vein grafts. J Am Coll Cardiol 2001;38: 659-65. 38. Pache J, Kastrati A, Mehilli J, et al. Intracoronary stenting and angiographic results: strut thickness effect on restenosis outcome (ISAR-STEREO-2) trial. J Am Coll Cardiol 2003;41:1283-8.