Late mortality with sirolimus-eluting stents in diseased saphenous vein grafts

Compared with bare-metal stents (BMS), sirolimus-eluting stents (SES) have been shown to be effective in decreasing the rate of restenosis and repeated revascularization procedures at 6 months among patients with lesions in native coronary arteries.^1,2 The Reduction of Restenosis in Saphenous Vein Grafts with Cypher Stent (RRISC) trial was the first study to evaluate the use of SES in saphenous vein graft lesions, and it showed that the rate of restenosis and repeated revascularization was decreased at 6 months in saphenous vein graft lesions as well.³ Recent long-term data from randomized trials in native coronary arteries suggest that there is an increase in "hard" end points, mainly very late stent thrombosis but possibly also death and myocardial infarction (MI), with SES and other drug-eluting stents.^4-8 Because there are insufficient data on the long-term effects of SES in saphenous vein grafts, we conducted a secondary post-hoc analysis of the RRISC trial comparing the use of BMS with SES in saphenous vein grafts over a period of 3 years.

Subjects and methods

We randomly assigned 75 subjects who had previously undergone coronary artery bypass graft surgery to undergo stenting of diseased saphenous vein grafts with either SES or BMS from September 2003 to November 2004. Subjects had a total of 96 "de novo" target lesions in 80 saphenous vein grafts, with a reference vessel diameter >2.5 and <4.0 mm. All subjects were given clopidogrel (Plavix) and aspirin for at least 2 months. After 6 months, subjects underwent coronary angiography. The primary end point of the original study was 6-month in-stent late loss.³

We performed a secondary analysis to evaluate the incidence of new events, including new hospitalizations, repeated revascularizations, MI, death, and additional ambulatory visits. Between November and December 2006, we telephoned subjects and asked them to complete a questionnaire about any new events over the phone. Relevant records were checked for every new event. The end point for the follow-up analysis was all-cause mortality, with a secondary end point of major adverse cardiac events, including death, all nonfatal major MI (also periprocedural), target lesion revascularization, target vessel revascularization, and stent thrombosis (defined according to the Academie Research Consortium criteria).⁹ All clinical events were assessed by an independent clinical events committee blinded to the patients' treatment assignment.

Results

A total of 54 BMS were implanted in 49 lesions in 37 subjects, and 60 SES were implanted in 47 lesions in 38 subjects. As shown in Table 1, the baseline and clinical characteristics of the subjects were similar. The BMS group had a median follow-up period of 32 months, and the SES group had a median follow-up period of 30.5 months (P = .24).

Table 1. Baseline characteristics of the subjects in the 2 groups.

Eleven subjects (29%; 95% confidence interval [CI], 17%-45%) in the SES group died, whereas no subjects (0; 95% CI, 0-9%) in the BMS group died, with an absolute difference of 29% (95% CI, 14%-45%; P <.001). Seven of the deaths resulted from cardiac causes, with 3 sudden deaths (possibly due to late or very late stent thrombosis) and 1 death caused by very late stent thrombosis 13 months after SES implantation documented by angiography. Table 2 shows the clinical and angiographic data on the SES patients who died. A nonfatal MI occurred 30 months after the baseline procedure due to a very late SES thrombosis documented by angiography, which was managed successfully with percutaneous intervention. Overall, 2 of the 38 subjects (5%) in the SES group had definite angiographically documented (very late) stent thrombosis, whereas there were no stent thromboses in the BMS group (P = .49). The rate of any possible stent thrombosis was 0% after BMS and 13% (5 of 38; 2 late and 3 very late) after SES (Fisher exact test 2-sided P value = .054; log rank test = 0.022).

Table 2. Details on the specific time courses and causes of death in the sirolimus-eluting

stent subjects.

Fifteen subjects (40.5%) in the BMS group and 22 subjects (57.9%) in the SES group experienced major adverse cardiac events (MACEs; Table 3). After the 6-month follow-up, there was an increase in target lesion revascularization and target vessel revascularization, which cancelled out the 6-month benefit of decreased revascularization procedures previously shown with SES. Both the BMS and SES groups were similar with regard to confounding factors, such as the use of hypocholesterolemic and antiplatelet therapy and additional revascularization procedures not related to target lesion or target vessel revascularization.

Table 3. Clinical events and medical therapy in the 2 groups at the longest available

follow-up period.

Discussion

We compared the effects of SES with BMS in subjects with diseased saphenous vein grafts in a secondary analysis of the RRISC study. Findings showed that subjects in the SES group had a significantly higher mortality rate than those in the BMS group. Furthermore, the initial 6-month benefit of decreased revascularization procedures shown with SES was eliminated after a follow-up period of up to 3 years.

Recent studies have shown that there is a possible increase in "hard" adverse events, in particular, very late stent thrombosis, following the use of drug-eluting stents in native coronary arteries.^4-8 As a result, much attention has been paid to the long-term effects of drug-eluting stents. Although the Food and Drug Administration has declared the use of drug-eluting stents to be safe, it has cautioned against an increased risk of adverse events with off-label use.¹⁰ Based on the results of this study, and because percutaneous saphenous vein graft treatment with SES is currently off label, all potential problems with SES implantation in saphenous vein grafts should be carefully considered and thoroughly discussed with patients.

Results of this study showed that the previous benefit of decreased restenosis and revascularization procedures in saphenous vein grafts at 6 months was lost with a longer follow-up period. Other studies have not shown any benefit of drug-eluting stents over BMS in diseased saphenous vein grafts.^11,12

The mechanisms that would explain the results of this study are not well understood. The long-term prognosis of patients with diseased saphenous vein grafts is affected by disease progression in the nonintervened saphenous vein graft segments.¹³ The specific responses in a diseased saphenous vein graft may also lead to late or very late thrombosis and late restenosis more dramatically than in native coronary arteries. The plaques in saphenous vein grafts are more likely to rupture, being softer and more lipid-rich, than plaques in native coronary arteries. This can encourage development of atherosclerosis and may lead to greater inflammation and thrombotic reaction after the implantation of stents, which may be further increased when the stents are treated with drugs and polymers.¹⁴

For the treatment of diseased saphenous vein grafts, BMS should still be regarded as the reference percutaneous treatment. Bare-metal stents should be used as the control for future studies involving saphenous vein grafts, which should take into consideration the occurrence of late restenosis and very late thrombosis.

This study had several limitations. Our findings may be under-powered because of the small sample size and may be subject to type I and type II statistical error. Double antiplatelet therapy was only required for at least 2 months in our study, whereas today, it should be recommended for a period of at least 12 months for all patients undergoing drug-eluting stent implantation.¹⁵ At the time the RRISC study was conducted, SES patients traditionally received 2 to 3 months of double antiplatelet therapy.¹ Therefore, some of the events in our study may have occurred because of early discontinuation of dual antiplatelet therapy. Because this was a secondary analysis, the end point of all-cause mortality was not identified at the beginning of the original study (powered for a 6-month difference in angiographic late loss). Because studies had shown an increase in long-term mortality after SES in native coronary arteries,^4-8 death was selected as an end point in the current analysis. Although none of the studies assessing drug-eluting stents have been powered to evaluate death rates,^1,2 preliminary information on long-term results of prior trials may be helpful in determining future studies.

Conclusions

Findings from a secondary analysis of the RRISC trial showed that SES is associated with a higher long-term mortality rate than BMS in saphenous vein graft disease. Furthermore, the previous benefit of decreased restenosis and revascularization procedures shown in saphenous vein grafts at 6 months of follow-up was lost with a longer follow-up period. Further studies are required, as the results of the current study may have been caused by factors other than the type of stent used.