Article Text

Download PDFPDF

Drug-eluting stents: a reappraisal
  1. Scot Garg,
  2. Patrick W Serruys
  1. Department of Interventional Cardiology, Thoraxcentre, Rotterdam, The Netherlands
  1. Correspondence to Professor Patrick W Serruys, Head of Interventional Cardiology, Ba583a, Thoraxcentre, Erasmus MC,'s-Gravendijkwal 230, 3015 CE Rotterdam, The Netherlands; p.w.j.c.serruys{at}

Statistics from


The introduction of drug-eluting stents (DES) in 2002 revolutionised the field of percutaneous coronary intervention (PCI), by significantly reducing rates of restenosis when compared to bare-metal stents (BMS).1 Following this there was a rapid and unprecedented uptake in their use, such that within a matter of years they were used in approximately 80–90% of PCI cases in the USA, while in the UK their use rose by over 300% between 2003 and 2006.2 3This honeymoon period was brought to an abrupt end following the well publicised concerns over their long-term safety that emerged during the ‘Barcelona firestorm’ in 2006.4 The consequences were a global reduction in their use, and only recently have rates of use recovered to what they were.3 It has been 8 years since the introduction of DES, and this would seem an appropriate time to reappraise their current position, and look to the future, as DESs try to maintain their unique position in interventional cardiology practice.

Are DESs effective?

Restenosis, the Achilles heel of both balloon angioplasty and stenting with BMS, with rates in published series of nearly 50% and 25%, respectively, was ultimately one of the driving forces behind the development of DES. The zero percentage restenosis rate published in the pivotal RAVEL study was unlikely to be maintained for all patient groups1; however, there can be little disappointment with the consistently low rates of restenosis observed in patients treated with DES. A large meta-analysis by Stettler et al reported a reduction in target lesion revascularisation (TLR) of 70% (p<0.0001) with sirolimus-eluting stents (SES) and 58% (p<0.001) with paclitaxel-eluting stents (PES) compared to BMS.5 This corresponded to an impressive number needed to treat, to prevent a single revascularisation, of seven and eight patients for SES and PES, respectively. Comparable meta-analyses, albeit on a smaller scale, by Spaulding et al, Stone et al, Kastrati et al and Mauri et al all produced similar impressive results.69 Although the benefit with DESs was greatest in the first year after stent implantation, the advantage was preserved out to 4-year follow-up. Moreover, recent data from the SIRIUS and TAXUS II studies also confirm that this benefit is maintained out to 5-year follow-up.10 11

Importantly, these favourable results are not solely confined to the select patients from randomised controlled trials treated for ‘on-label’ indications, but have been consistently reproduced in registries, and randomised controlled trials which have included those with patients receiving DES for ‘off-label’ indications.12 13 In fact, the benefit in terms of reduced restenosis from DES use appears to be greater in those patients treated for ‘off-label’ as opposed to ‘on-label’ indications. For example, in the large meta-analysis by Kirtane et al, which included over 9000 patients, the reduction in target vessel revascularisation (TVR) was 46% (HR 0.54, 95% CI 0.48 to 0.62, p<0.01) and 58% (HR 0.46, 95% CI 0.34 to 0.52, p<0.01) for ‘on-label’ and ‘off-label’ use, respectively.13

Some argue that restenosis is a benign phenomenon, and this benefit alone does not justify the use of a DES. These arguments are largely based on the false belief that in-stent restenosis presents with stable exertional angina owing to the slow proliferation and migration of smooth muscle cells leading to progressive neointimal hyperplasia. This was previously documented with restenosis following balloon angioplasty; however, the current reality is somewhat different. Retrospective registries reported that a stable presentation for in-stent restenosis occurs in only approximate a quarter of cases, with the commonest presentation being that of an acute coronary syndrome.14–16 This may be caused by the accumulation of macrophages, and extensive neovascularisation that has been seen in in-stent restenosis, together with the presence of tissue factor, both of which may act as a trigger for mural thrombus formation.17 18 In addition to the acute presentation, outcomes for PCI in patients with in-stent restenosis are poorer than PCI for de novo lesions. A mortality rate of 2.5% was reported in the Ontario registry in those having TVR within a month of BMS implantation,19 while De Labriolle et al reported respective rates of mortality and major adverse cardiovascular events (MACE) of 5.7% and 25.9% among 1958 BMS patients with in-stent restenosis.14 Similarly, Doyle et al also demonstrated a significantly higher mortality among those patients presenting with an MI caused by restenosis compared to no restenosis (HR: 2.37; p<0.001).20 These outcomes are certainly not benign, and although DES restenosis presents with similarly poor outcomes14 surely a lower overall risk of restenosis with the use of a DES is the ‘lesser of the two evils’?

Are DESs Safe?

The events of the ESC meeting in 2006 live long in the memory; it was the presentation here of the two meta-analyses by Camenzind et al and Nordmann et al that first raised the safety concerns regarding the use of DES. In brief, Camenzind et al reported a statistically significant 2.4% increase in the risk of death and myocardial infarction (MI) in patients treated with DES compared to BMS,21 while Nordmann et al demonstrated a statistically significant increase in non-cardiac mortality 2–3 years after SES implantation.22 Although critics cited methodological problems with both studies—for example, the use of unusual endpoints by Camenzind, and the reliance in both studies on aggregate trial data, as opposed to patient-level data, the results had immediate effects on DES usage. Moreover the results from the 2003 to 2004 SCAAR registry, and the BASKET-LATE study reiterated these concerns by also indicating a higher risk of mortality in patients treated with DES.23 24 Interesting, re-analysis of the same studies used by Camenzind, but this time using patient-level data, indicated no significant difference in death/MI between both groups (11.4% SES vs 10.1% BMS, p=0.4).6 A similar result has also been seen in the extended follow-up of the SCAAR registry which now includes patients having PCI in 2005 and 2006.25 26

Perhaps the most important contribution that these four studies made was to stimulate additional research, which has been vital to ensure the survival and public acceptance of DESs. Data from registries, randomised controlled trials and meta-analyses have subsequently been published which demonstrate the overall comparable outcomes between DES and BMS in terms of death and MI, at both short-term and long-term follow-up.5–9 Most notably the largest of these meta-analyses by Stettler et al included 38 DES trials, and included over 18 000 patients, followed up for 4 years, and reported a similar risk of mortality between patients treated with SES, PES and BMS; while the risk of MI, although comparable between PES and BMS (p=0.99), was significantly lower with SES compared to BMS (p=0.03).5

Understandably, questions were asked as to how applicable these results were to ‘real-world’ practice considering that most of these trials assessed DES use for so-called ‘on-label’ indications (lesions were stable, de novo, with a mean lesion length of 23–24 mm, a mean vessel diameter of 2.7 mm and treated using an average of 1.2–1.4 stents), while conversely in routine practice 70–75% of DESs were implanted for ‘off-label’ indications, which can be associated with a worse outcome and a higher risk of stent thrombosis (ST).27 28

These concerns prompted numerous studies, many of which are observational, and suggest a higher rate of death, MI and TLR in patients treated with DES for ‘off-label’ compared to ‘on-label’ indications.27 29 30 Moreover, data also indicate that for ‘off-label’ indications the use of a DES is no worse than the use of a BMS 31 32 with some studies, such as the STENT registry actually demonstrating significantly improved outcomes with the use of a DES.29 The recent meta-analysis by Kirtane et al specifically included patients with both on-label and off-label indications and demonstrated similar rates of death and MI among 9420 patients from randomised trials followed up for a median of 2.9 years. On the other hand among observational studies, which included over 180 000 patients, significant reductions in death (22%), and MI (13%) were seen with the use of a DES.13 It seems likely that the overall poor outcome with off-label use is more related to patient or lesions characteristics, rather than to specific shortcomings of DES.

Stent thrombosis

No discussions on the safety aspects of DESs are complete without considering ST. It is important to appreciate that ST has long been an adversary, and undesired complication of PCI, and is not, as is commonly misconceived, a problem exclusive to the use of DES. The rate of early and late stent occlusion (thrombotic or not) with the self-expanding Wall stent in 1986 was over 20%,33 and it wasn't until the realisation of the benefits of lifelong aspirin; the introduction of thienopyridines post-stenting; and techniques to ensure adequate stent deployment that acceptable rates of ST of 1–1.5% were achieved.34–38

At the present time ST remains fortunately rare, but nevertheless it is unpredictable, and is associated with a significant morbidity and mortality4; 70–80% of patients present with MI and up to one-third of patients with definite ST will die. The risks of acute (<24-hour), early (<30 days) and late (>30 days, <365 days) ST are similar between BMS and DES; however, of greatest concern is the excess risk of very late ST (>1 year) with the use of DES. Moreover, data from the Rotterdam-Bern registry of 8146 consecutive patients have demonstrated that use of a DES exposes patients to a persisting risk of very late ST of 0.6% per year, out to 5 years of follow-up.39 The adverse effect of ST on subsequent outcome cannot be overstated. For example, in the ARTS-II study 32% of all MACE events at 5-year follow-up were attributed to definite, probable or possible ST.40

It is not surprising that current generation DESs are associated with an increased risk of very late ST when considering that the elution from their surface of cytotoxic and cytostatic drugs, together with the presence of non-erodable polymers, are both associated with impaired endothelialisation, allergic reactions, inflammation, and vascular dysfunction; factors, among many, that have all been implicated in precipitating ST.41–44 Unfortunately, therefore, the very mechanism of action of DES is such that ST is always going to be a persisting risk with DES, and thus measures must be implemented to minimise this risk. For example, ensuring that patients are able, and likely, to comply with at least 12 months of dual anti-platelet therapy (DAPT) before implanting a DES, together with ensuring optimal stent deployment during PCI are two effective ways of reducing this incessant risk of ST.

Does the net clinical benefit favour DES?

In evaluating the overall performance of DES it is important to weigh up the overall net clinical benefit, which can be summarised after considering the stent's beneficial and adverse effects. The unquestionable benefit of DESs relates to their significant reduction in rates of restenosis; while their adverse effects clearly relate to the increased risk of late, and very late, ST. Of note the clinical sequelae and incidence of restenosis and ST are different—for example, in a study by Stone et al the rate of ST and TLR was 3.1% (DES:BMS 1.4:1) and 12.3% (DES:BMS 0.47:1), respectively, while the rate of death and MI within 7 days of these events was 91.1% for ST and 3.5% for TLR.45 Therefore to assess the overall benefit of DES it is not sufficient to simply consider the net benefit between the number need to prevent one revascularisation, and the number needed to harm—actual mortality must be considered. Numerous observational studies and randomised controlled trials have confirmed that compared to BMSs, DESs not only significantly reduce restenosis, but do so without increasing the risk of mortality or MI. Thus the overall net clinical benefit favours DES.

It is likely that the similar outcomes in terms of death and MI at long-term follow-up, despite higher rates of ST, occur as a result of the timing of the respective adverse events. The reduction in restenosis following implantation of a DES results in a survival benefit over BMS as early as 3 months after stent implantation. Unfortunately, however, this early survival benefit appears to be subsequently lost owing to a late catch-up in mortality observed after 6–12 months, which is possibly the result of ST, and in particular may relate to the discontinuation of DAPT around this time.46

Current status

The spectrum of coronary artery disease (CAD) requiring revascularisation has gradually progressed from predominantly single-vessel disease, to complex CAD (multivessel disease (MVD) and left main stem (LMS) disease), which is largely the result of the increasing age and multiple comorbidities of patients currently being investigated for angina pectoris.47 Historically complex CAD was treated exclusively with coronary artery bypass graft (CABG); however, the improved outcomes in revascularisation using DES, have made PCI a more attractive alternative. Nevertheless, 5-year results from the ARTS-II registry and observational data of ‘real-world’ practice have shown that despite similar rates of death and MI among patients with MVD treated with DES or CABG, rates of repeat revascularisation and major adverse cardiac and cerebrovascular events (MACCE) appear to be significantly higher with PCI, such that it remains the preferred method of revascularisation in only approximately one-third of patients.40 48 More recently these results have been supported by prospective randomised data from the 1800 patients with complex CAD enrolled in the SYNTAX study. At the 2-year follow-up, 23.4% and 16.3% of patients treated with PCI and CABG, respectively, reached the primary endpoint, a composite of death (PCI 6.2% vs CABG 4.9%; p=0.24), MI (5.9% vs 3.3%; p=0.01), stroke (1.4% vs 2.8%; p=0.03) and repeat revascularisation (17.4% vs 8.6%; p<0.001).49

There is a wide spectrum of MVD and common sense and experience tells us that outcomes are not the same for all patients. It follows that not all patients with MVD can be appropriately revascularised with PCI, and therefore suitable patient selection is imperative. Perhaps one of the most important features of the SYNTAX study was the introduction of the SYNTAX score, a novel method of objectively assessing the complexity of CAD. Importantly the stratification of outcomes according to SYNTAX score demonstrated that outcomes between PCI and CABG were only comparable in those patients with MVD who had a SYNTAX score in the lowest score tercile (SYNTAX score ≤22).50 On the other hand, among those with LMS disease outcomes were similar in patients with SYNTAX scores in the lower and middle terciles (SYNTAX score ≤32, figure 1).51 It is important to acknowledge therefore that in patients with complex CAD, CABG is the standard of care; however, in a selection of patients PCI with DES offers a suitable alternative.

Figure 1

The major adverse cardiac and cerebrovascular events (MACCE) rates at 2-year follow-up in the SYNTAX trial stratified by SYNTAX score tercile. In patients with triple vessel disease, PCI only offers a comparable outcome to CABG for patients with an SYNTAX score between 0 and 22; in patients with left main stem disease, outcomes are comparable between PCI and CABG for SYNTAX scores up to 32.

Diabetes mellitus represents one of the most significant risk factors for experiencing an adverse event following revascularisation. Data from registries and randomised studies have confirmed the superior outcomes in diabetics following treatment with a DES as opposed to a BMS; however, it is disappointing that data are still lacking on a similar comparison between treatment with DES and CABG.

At present the CARDIA trial represents the only randomised assessment of PCI and CABG in diabetics with MVD. Unfortunately the study was underpowered following the early discontinuation of recruitment, such that only 510 of the desired 600 patients were enrolled. Preliminary results at 1-year follow-up demonstrated no significant difference in mortality (3.2% PCI vs 3.3% CABG, p=0.83) or the composite clinical endpoint of death, non-fatal MI or non-fatal stroke (10.2% in PCI vs 11.8% in CABG). Repeat revascularisation was required significantly more frequently in the PCI group (9.9% vs 2.0%, p=0.001), which was consistent even in the subgroup of patients treated using DES (7.3% vs 2.0%, p=0.013).52

Additional data on this important group of patients come from the SYNTAX study, which enrolled 452 diabetics (221 CABG, 231PCI). Results demonstrated an adverse effect of diabetes on mortality irrespective of revascularisation strategy. Moreover, diabetic patients had a higher rate of MACCE when treated with PCI which was driven by significant higher rates of repeat revascularisation.53 Interestingly the stratification of outcomes by SYNTAX score demonstrates higher rates of MACCE with PCI in all terciles; however, this difference only reached significance in those patients with the most complex disease (SYNTAX score ≥33).54

The optimal management of these diabetic patients with MVD remains to be established, and it is hoped the results of the ongoing FREEDOM study will provide much needed data on this important group of patients.55

Future directions

So where does the future of DESs lie? There is always room for improvement, and although additional reductions in restenosis may be hard to achieve, improvements in safety—namely, reductions in ST are clearly desired. With these thoughts in mind there have been extensive new developments in DES design, with changes in both polymers and stent platforms.

The second generation DESs were developed with more biocompatible polymers, which appear to have had a beneficial effect on rates of late/very-late ST. Although no study has been adequately powered to detect a difference in ST, significantly lower rates of ST were observed at 1-year follow-up with the Xience V everolimus eluting stent (EES, Abbott) compared to the PES in the recently presented SPIRIT IV and COMPARE study (p<0.05 for both), while in the SPIRIT II study numerically lower rates of ST were seen with the EES compared to the PES at 3-year follow-up.56–58 Similarly a pooled analysis of the Endeavour zotarolimus eluting stent (Medtronic) trials demonstrated consistently lower rates of ST compared to other DES out to 5-year follow-up.59

More recently the need for a permanent polymer has been rightly questioned, as these polymers serve no purpose once drug elution has been completed. Consequently newer DES stents have been developed that are coated with biodegradable polymers, which slowly degrade into carbon dioxide and water within 6–9 months of stent implantation. Moreover, to reduce the total dose of both drug and polymer, these polymers have been applied to only the abluminal surface of the stent. An early proof of concept comes from the LEADERS study, which has demonstrated comparable outcomes in terms of clinical efficacy between the Cypher SES (Cordis) with a permanent polymer, and the Biomatrix biolimus eluting stent (Biosensors Interventional Technologies) with a biodegradable polymer.60 Notably at 2-year follow-up, there was a lower rate of very late ST (0.2% vs 0.5%) and overall ST (2.2% vs 2.5%) with the biodegradable polymer stent.61 Stent designers have realised the potential of this new technology, and consequently several other biodegradable polymer stents, such as the NEVO SES (Cordis) and the JACTAX PES (Boston Scientific Corporation), will commence clinical trials in 2010.

A step forward from biodegradable polymers stent is the concept of a DES without a polymer. Currently the sirolimus-eluting YUKON stent (Translumina) is the only polymer-free stent available commercially, and uniquely, the dose of sirolimus on this stent can be customised in the catheter laboratory, just before stent implantation. Clinical data come from the randomised ISAR-TEST study, which indicates non-inferiority of the YUKON stent when compared with PES at 9–12-month follow-up.62 63 Notably, long-term data from an angiographic observational study of 1331 patients have recently reported a significantly lower change in late loss between 6 months and 8 months and 2 years for the YUKON stent, when compared with SES and PES (YUKON 0.01±0.42 mm, SES 0.17±0.50 mm, and PES 0.13±0.50 mm, p<0.001).64 This observation has important implications as it suggests that these polymer-free stents may not be subject to the ‘late catch-up’ phenomena that have been seen with permanent polymer DES.65–67 Other polymer-free DES stents include the biolimus-eluting BioFreedom stent (Biosensors), and the VESTAsync SES stent (MV Therapeutics Inc), which are both currently undergoing first-in-human studies and have so far demonstrated encouraging safety results.68 69

Completely biodegradable stents (BDS) represent the ultimate in new DES technology. They rely on the relatively simple concept that once healing has been completed, the need for a permanent scaffold—that is, a stent, disappears. BDS was first studied in 200070; however, it has only been in recent years that this technology has fully developed. Many BDS are currently still in the pre-clinical phase of development, and as yet only two have entered clinical trials, the most prominent of which has been the poly-l-lactide Bioabsorbable Vascular Solutions EES stent. This was studied in the landmark ABSORB trial which has recently reported the absence of ST out to 3-year follow-up, and a MACE rate of 3.6% with no events reported between 6 months and 36 months of follow-up.71 Importantly, intravascular imaging at 2 years demonstrated complete biodegradation of the stent, while acetylcholine testing demonstrated the restoration of vasomotor function.72 The second phase ABSORB study has completed enrolment, and a large registry, the ABSORB EXTEND is planned in 2010, together with an application for regulatory approval.


There is no denying the association of DES with an increased risk of ST; however, this risk appears to be offset by the reduced risk of repeat revascularisation such the overall net clinical benefit favours the use of a DES over a BMS. The impressive results from the early trials of DES certainly allowed a degree of complacency to set in, and the safety concerns that subsequently emerged should be considered a timely and welcome eye-opener, which have ultimately resulted in research leading to overall improvements in safety.

In recent times the use of DES has expanded to more complex patient groups; however, patient selection is key, and for the majority of patients with complex disease CABG is still the standard of care, although PCI offers a viable alternative in certain selected patients. Looking to the future there is much anticipation and excitement from the newer DESs and biodegradable stents that have shown promising early safety results.


View Abstract


  • Competing interests None.

  • Provenance and peer review Not commissioned; not externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.