Intended for healthcare professionals

Clinical Review

New approaches to preventing restenosis

BMJ 2003; 327 doi: (Published 31 July 2003) Cite this as: BMJ 2003;327:274
  1. Balram Bhargava, senior consultant (balrambhargava{at},
  2. Ganesan Karthikeyan, consultant1,
  3. Alexandre S Abizaid, consultant2,
  4. Roxana Mehran, consultant3
  1. 1Department of Cardiology, Cardiothoracic Sciences Center, All India Institute of Medical Sciences, New Delhi 110029, India
  2. 2Dante Pazzanese Institute of Cardiology, Sao Paulo, Brazil
  3. 3Cardiovascular Research Foundation, New York, NY, USA
  1. Correspondence to: B Bhargava
  • Accepted 3 July 2003

For over a quarter of a century percutaneous coronary interventions have been used to treat patients with coronary artery disease, yet restenosis continues to be a problem. This review discusses the advances being made to overcome restenosis, particularly the development of drug eluting stents


Andreas Gruntzig pioneered percutaneous transluminal coronary angioplasty in 1977.1 Since then, percutaneous coronary interventions have revolutionised the treatment of patients with symptoms of coronary artery disease, sparing them the need for surgical revascularisation. In the United States, the number of percutaneous interventions performed each year is nearly double that of coronary artery bypass surgery.w1 There have been three phases in the evolution of percutaneous interventions (see box 1), but in each restenosis has been a problem. It seems the next (and optimistically the final) phase is the development of drug eluting stents—that is, stents used for the local delivery of drugs (table).

Restenosis and instent restenosis, 1977-2002

View this table:

Sources and selection criteria

The information in this review is based on the results of a Medline search using the key words coronary angioplasty, restenosis, instent restenosis, drug eluting stents, and intracoronary brachytherapy, and articles from the authors' personal collection. Cross references and related articles were accessed when necessary.


Restenosis is the maladaptive response of the coronary artery to injury. The simplest and most widely used definition of restenosis is a diameter stenosis of 50% at follow up. Restenosis occurs in 20-50% of patients after balloon angioplasty and in 10-30% of patients receiving a stent. Three distinct processes are involved—recoil of the vessel, neointimal proliferation, and early thrombus formation. The relative contribution of each of these depends on the type of injury. About three quarters of the lumen loss after balloon angioplasty is due to vessel recoil and the rest to neointimal proliferation,2 whereas coronary stenting virtually eliminates vessel recoil, and restenosis is largely due to neointimal proliferation. Growth factors and cytokines are the major stimuli for proliferation of smooth muscle cells after an artery is injured, when deposition of platelets, leucocyte infiltration, expansion of smooth muscle cells, deposition of extracellular matrix, and re-endothelialisation occur. The platelets release platelet derived growth factor, transforming growth factor, epidermal growth factor, and thrombin, which stimulate the migration, growth, and division of smooth muscle cells.w2 w3 Within the smooth muscle cell, cyclin dependent kinase and cyclins (regulators) drive the cell from one phase to another: G0, resting phase; G1, growth phase; S, replication of DNA; G2, preparation for division; and M, cell division. Cyclin dependent kinase has been described as the engine and the cyclins as the gearbox, determining whether the engine will run in the steady state or drive the cell forward in the cell cycle. Together they regulate cell cycle progression and migration of smooth muscle cells (fig 1).


Inhibition of neointimal formation by sirolimus impregnated stent

Summary points

Restenosis is a common problem after percutaneous coronary interventions, depending on patient characteristics, lesion morphology, and the type of intervention

Coronary stenting is the only intervention that has reliably and consistently reduced restenosis in coronary lesions

Near zero per cent restenosis rates have recently been achieved with antiproliferative drug eluting coronary stents

Recurrent stenosis after treatment of instent restenosis is common

Adjuvant treatment with intracoronary γ and β radiation is the most effective currently available means to prevent recurrent stenosis after the treatment of instent restenosis

Modifying gene expression using antisense therapies or gene transfer will play an important part in the prevention of restenosis

Box 1 Evolution of percutaneous interventions

  • Phase 1 (1977-87): simple balloon angioplasty

  • Phase 2 (1988-92): new device angioplasty

  • Phase 3 (1993-2002): stent revolution

Clinically significant restenosis generally occurs between one and three months after balloon angioplasty. Most patients present with typical angina, usually within four months of the procedure—restenosis is much less likely to be the cause of symptoms after six months. With coronary stenting the presentation is delayed further by about one to three months.3 w4

Prediction of restenosis

Restenosis depends on several factors (box 2). The risk of restenosis is increased in patients with diabetes mellitus, acute coronary syndromes, and, importantly, an inherent predisposition to restenosis.4 w5 w6 The increased restenosis rate among patients with diabetes is attributed to endothelial dysfunction, dysregulation of growth factor production, and increased platelet aggregability and thrombogenicity.w7 In acute coronary syndromes, vascular injury worsens the existing thrombogenic milieu and promotes cellular proliferation. Evidence of a predilection to restenosis comes from observations of a bimodal pattern of distribution of stenosis after angioplasty and the presence of restenosis at another treated site.5 w8 This predisposition might be mediated by genetic factors. Putative abnormalities are deletion polymorphisms of the gene encoding angiotensin converting enzyme, abnormalities in the expression of genes encoding the platelet Gp IIb-IIIa receptors, and some specific apolipoprotein E genotypes.w9 w10 w11 The most important lesion characteristics predicting restenosis are a small vessel diameter and a long lesion length.6 7w12 w13 A large minimum lumen diameter after stenting reduces the rate of restenosis—that is, “bigger is better.”w13

Prevention of restenosis


Several drugs, including those with antiplatelet, anti-thrombotic, antiproliferative, and anti-inflammatory properties, have been used to prevent restenosis, but except for probucol and prednisolone their contribution has been largely disappointing.w14 w15 The role of drugs in preventing restenosis is likely to be small because the principal mechanism of restenosis after balloon angioplasty concerns remodelling of the artery. These drugs might have a role in preventing neointimal proliferation after stenting.

Coronary stents

Coronary stents Stents produce the largest net gain in lumen diameter because they achieve an immediate large gain with modest later loss of diameter. The Belgium Netherlands stent study (BENESTENT) and the stent restenosis study (STRESS) showed that stents reduce restenosis rates after balloon angioplasty.8 9 These studies were conducted on “ideal” lesions, however, and higher restenosis rates have occurred with more complex lesions. Several materials, coatings, and designs have been developed in the quest for the ideal stent (see table A on>). Most have shown little success, except for drug eluting stents.

Drug eluting coronary stents

Using a coronary stent for local delivery of drugs combines scaffolding with targeted drug action. Although initial research and clinical trials concentrated on sirolimus and paclitaxel, numerous agents with antiproliferative properties are currently under investigation (see table B on>).

Box 2 Predictors of restenosis

Patient factors

  • Restenosis at another treated site

  • Diabetes mellitus

  • Intervention in acute coronary syndromes

Lesion characteristics and procedural variables

  • Diameter of vessel

  • Length of lesion or stent

  • Minimum lumen diameter before stenting

  • Minimum lumen diameter after stenting (“bigger is better”)

  • Ostial lesions

  • Saphenous venous grafts

  • Total occlusions


Sirolimus (Rapamycin; Wyeth-Ayerst, PA, USA), a macrolide antibiotic with potent antifungal, immunosuppressive, and antimitotic properties, is produced by Streptomyces hygroscopicus, an actinomycete found in soil. It is used to prevent the rejection of renal transplants, but its systemic use is limited by the side effects of hyperlipidaemia and thrombocytopenia.w16 Stents are coated with a polymer containing low dose sirolimus then a layer of drug-free polymer, which serves as a barrier to diffusion. The polymer acts as a reservoir for sirolimus, allowing gradual elution of the drug (see figure). Sirolimus downregulates cytokine dependent kinase mTOR (mammalian target of rapamycin), thus preventing progression from the G1 phase (cell growth) to the S phase (DNA replication).10 11 w17 Sirolimus is only effective when bound to sirolimus binding protein (FKBP), present only on smooth muscle cells and lymphocytes, thereby providing important targeting.

In the initial pilot study of sirolimus (30 patients), the stent virtually eliminated angiographic and ultrasonic evidence of neointimal proliferation.12w18 No instent or edge restenosis (restenosis within the stent or within 5 mm proximal and distal to the stent edges) was noted, and no major complications occurred. In the randomised study with the sirolimus eluting Bx velocity balloon expandable stent in the treatment of patients with de novo native coronary artery lesions (RAVEL) study (238 patients), no restenosis was observed in the sirolimus stent group compared with 27% of patients in the uncoated stent group.13 In the US arm of the SIRolImUS eluting stent in de novo native coronary lesions (SIRIUS) study, 9% of patients in the sirolumus stent group showed angiographic evidence of restenosis at nine months compared with 37% in the uncoated stent group.w19 Revascularisation of the target lesion was reduced from 17% to 4%. There was no increase in subacute or late thrombosis. These results were all the more significant because many patients in this study were at a high risk of restenosis (patients with diabetes, those who had multivessel angioplasty, and those with overlapping stents). The Food and Drug Administration has recently approved the clinical use of sirolimus eluting stents in the United States.


Paclitaxel was originally isolated from the bark of the Pacific yew. It is an antineoplastic agent, currently used to treat ovarian cancer. Paclitaxel enhances the assembly of microtubules that interrupt proliferation, migration, and signal transduction. It is incorporated into a fast release carrier system on the stent. Trials of paclitaxel coated stents have shown significant reductions in restenosis. The pacliTAXel-elUting Stent (TAXUS I) trial showed no restenosis in the paclitaxel stent group at 12 months compared with 10% for the standard stent group.14 This difference was reflected in the revascularisation of the target lesion in both groups (0% v 10%). The major adverse cardiac event rate at one year was also considerably reduced (3% v 10%). Results were maintained at 24 months. The TAXUS II trial confirmed the reduction in restenosis.w20 Similar results were observed in the European evaluation of paclitaxel eluting stent trial (ELUTES).w21 In the recently published Asian paclitaxel eluting stent clinical trial (ASPECT), angiographic restenosis decreased from 27% in the uncoated stent group to 4% in the paclitaxel stent group, due to a dose dependent decrease in neointimal proliferation.w22 No increase in subacute stent thrombosis (thrombosis occurring between 48 hours and two weeks after stenting) or restenosis at the edges of the stent has been reported in published randomised trials. However, late thrombosis has been reported after stopping antiplatelet therapy.w23 Some concern has been expressed that inflammation of the intima might occur at higher doses. Nevertheless, it is clear that drug eluting stents will be increasingly used for the prevention of restenosis because they nearly eliminate neointimal proliferation with little untoward effect and without the need for special expertise or instrumentation.

Biodegradable stents

The Japanese Igaki-Tamai stent is made of poly-L-lactic acid and takes 18-24 months to biodegrade. The scaffolding prevents abrupt vessel closure in the acute phase and vessel recoil within six months. Clinical results are available for 15 patients, showing no major complications and a restenosis rate of about 10%.w24 The stent is being evaluated further.

Debulking devices

The occurrence of restenosis is reduced when larger minimum lumen diameters are achieved after stenting. This makes a case for the use of debulking devices to mechanically reduce plaque volume. Early trials comparing directional coronary atherectomy with balloon angioplasty showed an increased mortality and Q wave infarction with atherectomy.15 w25 Subsequently, strategies to optimally debulk combined with adjunctive balloon angioplasty to achieve the largest possible lumen diameters were tested against conventional balloon angioplasty.16w26 Less restenosis was shown with directional coronary atherectomy compared with plain balloon angioplasty, but there was no difference in revascularisation of the target vessel at one year. Further, this benefit came at the cost of a higher incidence of increases in periprocedural myocardial enzymes. This failure to show consistent benefit from debulking has been attributed to the late loss in lumen diameter being proportional to the acute gain after the procedure (device “taxing”). A synergistic approach using directional coronary atherectomy and coronary stenting might have a role in reducing restenosis rates in complex lesions.w27 w28

Gene therapy

Vascular gene therapy can potentially control cellular growth related restenosis by modifying gene expression. Antisense oligonucleotides can be used to block gene expression by binding segments of specific mRNA with complementary DNA sequences. In a porcine model of restenosis, intramural delivery of specific antisense oligonucleotides completely eliminated c-myc expression and dramatically reduced neointimal proliferation.w29 An alternative approach is by transfer of genes, which have beneficial antiproliferative effects (for example, genes for endothelial nitric oxide synthase (eNOS) and vascular endothelial growth factor). Gene transfer has mostly been achieved using adenoviruses as vectors, although plasmid DNA and retroviruses have also been used. Animal studies with gene transfer have shown marked inhibition of neointimal proliferation; phase I studies in humans are under way.w30

Intracoronary radiation

The role of intracoronary radiation in preventing restenosis in de novo lesions is not entirely clear. Although one study showed a dose dependent decrease in the rate of restenosis after angioplasty, data from two studies in the United States were not favourable.17 w31 w32

Instent restenosis

Instent restenosis is due largely to neointimal proliferation. A classification for instent restenosis has been developed, depending on the degree and distribution of the hyperplastic neointima.18 Broadly, instent restenosis can be focal or diffuse. Therapeutic success is dictated by the volume and extent of the neointima. Focal lesions respond much better than diffuse ones.

Balloon angioplasty

In contrast to the atherosclerotic plaque in de novo lesions, the hyperplastic neointima is relatively less compressible, and therefore the emphasis has been on debulking rather than on plaque compression to prevent recurrent stenosis. Nevertheless, balloon angioplasty has been used extensively. The occurrence of restenosis in this setting largely depends on whether the restenosis is focal or diffuse. The recurrent restenosis rate with diffuse lesions has been shown to be twice that of focal lesions (63% v 31%, respectively).w33 However, the recent angioplasty versus rotational atherectomy for treatment of diffuse instent restenosis trial (ARTIST) suggests that even for diffuse lesions balloon angioplasty is better than rotational atherectomy.19 Plain balloon angioplasty is also likely to be successful in patients with restenosis due to poorly apposed stent struts, where luminal gain depends more on additional stent expansion than tissue compression.w34 A meta-analysis concluded that balloon angioplasty should be the procedure of choice when a good immediate outcome can be expected, because the nine month major adverse cardiac event rate is virtually identical with all forms of treatment.w35 Additional stenting within the region of instent restenosis (“stent sandwich”) improves immediate angiographic outcomes, but long term results are poor.w36 w37

Cutting balloon angioplasty

Cutting balloon angioplasty utilises a standard balloon catheter with three or four microblades (depth 0.18 mm, thickness 0.70-0.76 mm). The blades, attached longitudinally and folded within the balloon, are exposed on inflation and cut the neointimal tissue, facilitating plaque compression. They also help to anchor the balloon to the neointima (standard balloons tend to slip proximally or distally during inflation). Cutting balloon angioplasty for instent restenosis is associated with lower rates of revascularisation of target lesions and restenosis.w38-w40 At many centres, cutting balloon angioplasty combined with intracoronary radiation is the treatment of choice for instent restenosis.

Debulking devices

Rotational atherectomy

Rotational atherectomy utilises a high speed (150 000-200 000 rpm) diamond tipped drill, which effectively removes hyperplastic neointima and results in a larger minimum lumen diameter compared with balloon angioplasty. The promising early results (20% restenosis) of the randomised trial of percutaneous transluminal coronary angioplasty versus rotablator for diffuse instent restenosis (ROSTER) could not be reproduced in the angioplasty versus rotational atherectomy for treatment of diffuse instent restenosis trial (ARTIST), which showed a worse outcome for rotablation.19 w41 One possible reason for this discrepancy is that the low pressure used for inflation of the balloon after rotablation led to little increase in the stent diameter. Presumably high pressure inflation would resolve this.

Directional coronary atherectomy

Directional coronary atherectomy effectively removes neointima, with about a third of patients requiring revascularisation of target lesions at 10 months.w42 However one meta-analysis showed that at nine months the major adverse cardiac event rates with directional coronary atherectomy did not differ significantly from other procedures.w35 Further, activation of the cutter through stent struts may disrupt stent integrity or create sharp edges that puncture balloons.

Intracoronary radiation (brachytherapy)

A more recent approach to instent restenosis is intracoronary radiation or brachytherapy, on the grounds that it may control the excessive proliferative response and limit neointima formation, analogous to low doses of ionising radiation to treat benign proliferative disorders such as keloids, heterotropic bone formation, and pterygia.w43 Different isotopes on various platforms have been used to develop practical systems for endovascular brachytherapy. The platforms are catheter based systems such as line source wires, radioactive beads (delivered hydraulically), and balloons filled with radioactive gas or liquid, or stents incorporating γ (low energy, high penetration) or β (high energy, low penetration) emitters.

The first intracoronary brachytherapy procedure in humans was performed in Caracas, Venezuela in 1994.w44 Endovascular brachytherapy has since been tested in over 5000 patients, leading to approval by the Food and Drug Administration for the treatment of instent restenosis with both β and γ radiation.w45

Of the trials for the treatment of coronary instent restenosis by brachytherapy (see table C on>), the important ones are catheter based radiotherapy to inhibit restenosis after coronary stenting (SCRIPPS I), intracoronary gamma radiation therapy after angioplasty inhibits recurrence in patients with instent restenosis (WRIST), and localised intracoronary gamma radiation therapy to inhibit the recurrence of restenosis after stenting (GAMMA I).20 21 w46 Initiated at the Scripps Clinic in March 1995, SCRIPPS I tested the efficacy of iridium-192 (γ source) in 55 patients with restenosis. At follow up, the patients who received radiation had better mean minimum lumen diameters, less late loss of lumen, and lower restenosis rates. The WRIST trial randomised 100 patients with instent restenosis and 30 patients with instent restenosis in saphenous vein grafts to 192Ir or placebo. Restenosis was significantly reduced, particularly in patients with diabetes and those with vein grafts. Similarly, in the GAMMA I study on 252 patients, the primary end point—death, myocardial infarction, and the need for repeat revascularisation of the target lesion during nine months of follow up—occurred in 53 patients assigned to placebo (44%) and 37 patients assigned to 192Ir (28%). Similar results have been observed with β radiation in the proliferation reduction with vascular energy trial (PREVENT) and intimal hyperplasia inhibition with beta in-stent trial (INHIBIT).22 w47

Two important complications related to brachytherapy are edge restenosis (“candy wrapper” effect) and the risk of late thrombosis. Edge restenosis was first noted with radioactive stents and is possibly the result of radiation dose “fall-off” at the edges of the stent combined with vessel wall injury from inflation of the balloon outside the stent margins. Edge restenosis may be prevented by avoiding the “geographic miss,”—that is, a portion of the vessel that has undergone injury during the procedure, but has not received the intended dose of radiation. The likelihood of geographic miss can be decreased by radioactive coverage of an adequate length of the vessel.w48 w49

Delayed (after one month) thrombosis may occur in up to 9% of patients receiving additional stents at the time of radiation.23 Therefore, such repeat stenting (stent sandwich) is discouraged. Further, a long course (more than six months) of antiplatelet therapy is recommended for all patients undergoing brachy-therapy. Such treatment has reduced the risk of late thrombosis to that of the control population.w50

Drug eluting coronary stents

Preliminary data on a paclitaxel coated stent has shown a high major adverse cardiac event rate, which was largely due to the need for repeat revascularisation of the target lesion.w51 The potential for improved results with drug eluting stents in the treatment of instent restenosis is being explored.

Additional educational resources


O'Neill WW, Leon MB. Drug-eluting stents: costs versus clinical benefit. Circulation 2003;107: 3008-11

Lemos PA, Serruys PW, Sousa JE. Drug-eluting stents: cost versus clinical benefit.Circulation 2003;107:3003-7

Kandzari DE, Mark DB. Intracoronary brachytherapy:time to sell short? [Editorial.] Circulation 2002;106: 646-8


tctmd (>)—an industry supported website, providing useful information on topics in interventional cardiology.

Information for patients

Cypher, sirolimus eluting coronary stent (>)—Patientsection of Cordis Corporation's sirolimus eluting stent website. Provides useful general information about restenosis and drug eluting stents.

The future

Drug eluting coronary stents will be increasingly used in clinical practice despite their high initial cost (about three times that of a bare metal stent) because the cost is likely to be compensated for by the reduced need for repeat revascularisation. The costs of stents should go down as more manufacturers enter the market, especially from Asia. It is anticipated that under 5% of patients who undergo multivessel angioplasty will need repeat revascularisation compared with the current 14-16%. This is expected to at least equal, if not better, the outcome with coronary artery bypass surgery for patients with multivessel disease, even in those with diabetes. This will be tested in the arterial revascularisation therapies (ARTS II) study.24 Intracoronary brachy-therapy will remain the best treatment for diffuse or refractory instent restenosisw52 until drug eluting stents are proved useful for this indication or until an efficient and reliable antiproliferative genetic therapy becomes available.

Embedded ImageAdditional web references (w1-w53) and tables appear on>


  • Contributors All authors were involved in the preparation of the manuscript and in the revision process. BB will act as guarantor.

  • Funding None declared

  • Competing interests None declared


View Abstract