PCI (percutaneous coronary intervention) now outnumbers CABG (coronary artery bypass grafting) by more than 3:1 for the treatment of coronary heart disease. In this article, we discuss the current challenges faced by interventional cardiologists including restenosis and its treatment options and potential therapies for the future. The impact of stent geometry on restenosis and strategies to deal with challenging lesions such as bifurcations and lesions in the left main stem are also discussed.
ACSs (acute coronary syndromes)
ACSs, caused by coronary atherosclerotic plaque destabilization and interruption of coronary flow, are common, serious conditions that account for 15% of all deaths in the U.K. and are associated with considerable morbidity among survivors.
The past decade has seen a growth in treatment options available for ACSs as a consequence of better understanding of the pathophysiology of this condition. At a biological level, the process is driven by inflammatory mechanisms in the vessel wall where atherogenesis is initiated by lipids and other environmental agents [1,2]. Instability of the plaque atheroma so as to cause the disease to present is also viewed as an inflammatory process, predominantly where plaque macrophages destabilize the cap of the atheroma, resulting in the formation of an intra-coronary thrombus.
PCI (percutaneous coronary intervention)
Treatment of ACSs is done initially with pharmacological agents such as aspirin and statins, but when medical therapy fails to control symptoms, revascularization by either CABG (coronary artery bypass grafting) or PCI is indicated. Over the last few years, due to advancements in the equipment available and increased operator experience, the indications for PCI have changed, encompassing a wider remit. PCI comprises PTCA (percutaneous transluminal coronary angioplasty; ‘balloon angioplasty’) and coronary stent implantation. Andreas Gruentzig published his initial PTCA results in 1978 and pioneered the technique. Stent implantation has become commonplace since the mid-1990s and technology in this area continues to grow. Although catheter-based techniques of revascularization have an initial success rate that exceeds 95%, the immediate benefit is still attenuated by restenosis (the return of clinical symptoms following an apparently successful intervention) that occurs over the following 6 months.
Restenosis is a multifactorial process, and it features an excessive neointimal (smooth-muscle cell) response to injury, recoil of the artery and negative remodelling (in which the restenosing artery gets smaller rather than bigger to accommodate the intraluminal tissue). It therefore follows that forming and maintaining a large arterial lumen will reduce the impact of restenotic tissue on blood flow. One method of doing this is to implant a coronary stent, which scaffolds the artery, thereby eliminating recoil.
Over 90% of PCI procedures performed in the U.K. today involves the implantation of an intraluminal stent. Traditional coronary stents are bare-metal (usually stainless-steel) cages that scaffold the artery, stabilize the plaque and help to prevent arterial recoil and therefore restenosis. BENESTENT 1 and STRESS, two early but significant clinical trials, clearly showed that stenting in native vessels reduced the incidence of restenosis compared with PTCA alone [3,4]. Stents eliminate acute recoil of the vessel but, unfortunately, neointima formation still occurs and 20–30% of patients who have a conventional bare-metal stent still have ISR (in-stent restenosis) (depending on lesion and patient type).
DESs (drug-eluting stents)
During the last 2–3 years, interest has focused on the implantation of DESs to reduce the tissue response to injury. The two agents in common use are sirolimus and paclitaxel, both of which are cytostatic and reduce cell replication in the developing neointima [5,6]. There is abundant experimental, and now clinical, evidence that they are effective at doing so, reducing the incidence of clinically relevant restenosis to single figures, albeit in selected lesions and patients [6,7].
There are some emerging problems with DESs, however. One problem is that the agents loaded on the stents can interfere with the healing process; and detailed pathological examination of stented sites has revealed an exposed injury surrounding the stent struts, with adherent platelets and inflammatory cells, and absent endothelium for prolonged periods . Coronary thrombosis has only been avoided in patients receiving DESs by the prolonged administration of antiplatelet agents, usually two being given together for 6 months or longer .
Another problem, and this is much more frequent than the former, is that DESs are being used in increasingly complex lesions (such as long plaques in small calibre vessels in patients with diabetes), with marked attenuation of their efficacy. These are known as ‘real worlD' lesions, to distinguish them from the simpler lesions studied in the early trials. The rate of MACE (major adverse cardiac events) in patients with such lesions is reduced from 21% (for bare-metal stents) to 15% (for paclitaxel-eluting stents) in a recent large randomized trial . Most of the MACE are repeat PCI procedures, undertaken to treat restenosis (rather than the other components of MACE, namely death, myocardial infarction or stroke). The improvement in MACE seen with DESs is due entirely to a reduction in restenosis. So, in the trial just mentioned, which more accurately represents contemporary treatment practices than the earlier, more ‘optimistic’ studies [6,7], attenuation of the reduction in restenosis clearly indicates that these antiproliferative drugs do not provide a complete solution to the restenosis in the ‘real worlD'.
The final problem relates to the durability of the reduction in restenosis found with these agents. The drug with the most complete set of supportive data is sirolimus. In a previous study, performed in the pig coronary artery, the initial suppression of neointimal growth was almost completely lost (a ‘catch-up’ effect) by 90 days after implantation (90 days in the pig model is the equivalent of >3 year follow-up in humans) .
Even in the era of DESs, ISR is likely to remain a significant problem. The quantity of neointima is directly proportional to the amount of injury inflicted on the artery by the stent itself . That injury is in the form of either a deep laceration or stretch of the arterial wall. We have previously shown that stretch is ubiquitous after stenting and, even in the absence of deep injury, is an important mediator of neointima formation . A few parameters of stent geometry have been shown to have an influence on neointima formation, including smaller strut number , greater strut thickness  and increased variation in the angular burden (pointedness) of the stent cross-section.
Role of inflammation in ACSs and stenting
The agents discussed are, in the main, cytostatic, yet ACSs are predominantly inflammatory processes. They may not, therefore, be expected to eliminate the problems associated with stenting in the context of ACSs. IL-1 (interleukin-1) is an apical cytokine in the inflammatory response and is unique in that it has its own naturally occurring antagonist, the IL-1ra (IL-1 receptor antagonist). IL-1 and IL-1ra are expressed in diseased human coronary arteries [15,16]. Coronary artery IL-1β is up-regulated in porcine coronary arteries after balloon injury . Genetic studies have indicated that polymorphic variants within the IL-1 cluster that have net pro-inflammatory effects are associated with angiographic CAD (coronary artery disease) [18,19].
In mouse models of atherosclerosis, IL-1ra inhibits the formation of intimal fatty streaks in ApoE (apolipoprotein E)-deficient mice , and IL-1β-deficient mice when crossed with an ApoE homozygous ‘knockout’ background have reduced experimental atherosclerosis in the animals . Relevant to vessel wall healing, we have results from mice and pigs that indicate that the vessel wall response to injury (neointima formation) is an intensely IL-1-dependent event . In the porcine coronary artery, the neointima response to balloon angioplasty (a model of arterial wall rupture in some way analogous to that which occurs in ACSs) is reduced in its amount by a 14 day subcutaneous infusion of IL-1ra. We have crossed the IL-1R1−/− mouse with the ApoE−/− mouse and found reduced atherosclerosis. This double knockout, unable to signal IL-1, has no rise in blood pressure with fat feeding (in contrast with the ApoE−/− control mice) and this is associated with reduced serum amyloid A levels [a murine surrogate for CRP (C-reactive protein)], reduced markers of oxidative stress in the arterial wall and preservation of endothelial cell function in resistance arteries. We have reproduced the blood pressure rise inhibition and the modulation of oxidative stress seen in the IL-1R1−/−, ApoE−/− double knockout by using subcutaneous IL-1ra treatment in ApoE−/− mice. IL-1 could therefore be responsible for the development of hypertension and atherosclerosis, and its therapeutic control could limit atherosclerosis.
An anti-inflammatory strategy
Current work in our Unit has focused on the development of an anti-inflammatory strategy to treat restenosis using coronary stenting. We have combined an anti-inflammatory agent with a novel polymer and used SPR (surface plasmon resonance) to determine its binding capacity and release profile. We have then gone on to develop an anti-inflammatory DES that is currently undergoing formal preclinical testing using a porcine model of vascular injury. The pig is an ideal species in which to study restenosis because its size, anatomy and physiology make it amenable to manipulations, both biological and technical, that closely resemble those relevant to clinical practice. It is internationally accepted as the ‘prior to studies in human’ gold standard . It is a practical, realistic model, and it has been used in research into restenosis following balloon injury  and stent implantation in our laboratory, and worldwide, for over 10 years. The pig is economical, readily available and raises fewer ethical issues than some other models, such as primates (which in other respects would be ideal). We have developed specialized skills in working with this model, and in processing stented arteries.
Stem cells and stents
Another interest of our group is in the potential role of EPCs (endothelial progenitor cells) in restenosis following vascular injury. There is much evidence from both animal studies, and some from clinical studies, to indicate that this role is an important one. When EPCs, derived from mouse spleen, were infused intravenously after injury in the mouse carotid artery, neointimal formation was found to be reduced . Transplantation of GFP (green fluorescent protein)-transfected BM (bone marrow) prior to injury revealed the presence of green-stained ECs (endothelial cells) in the injured area, while administration of a statin enhanced the circulating pool of EPCs, accelerated re-endothelialization and reduced neointima formation . The benefits of EPCs extend to other forms of vascular injury, different mobilization signals and other species. For example, mobilization of EPCs by GCSF (granulocyte colony-stimulating factor) accelerated re-endothelialization and reduced inflammation after intravascular radiation in rabbits . Differing types of injury, however, seem to provoke differing degrees of involvement of EPCs. After wire injury of the murine carotid artery, for example, many BM-derived cells were found, whereas after ligation very few were seen . This difference probably arises from the well-documented different inflammatory responses in the two models. Where homing is observed, BM monocyte-lineage cells adhere in an MCP-1 (monocyte chemoattractant protein-1)-dependent manner and accelerate re-endothelialization as EPCs . It is, however, essential that any newly formed endothelium is functional. Transplantation of autologous circulating EPCs, cultured from peripheral monocytes, has been shown to restore endothelial function in denuded rabbit carotid arteries .
Turning to clinical data, in humans with coronary artery ISR, immunohistochemistry of 17 atherectomy samples for a range of surface markers, including CD34, revealed 7.1% EPCs compared with 0.6% in native disease . In another study, the number of circulating PBMC (peripheral blood mononuclear cell)-derived EPCs in peripheral blood was increased in a group with ISR, reduced in a group with widely patent stents and least in a group with no clinically evident CAD . As regards functionality of EPCs in ISR, 16 patients with angiographically evident ISR were compared with 11 without. While the number of circulating EPCs was similar in each group, EC adhesion in response to fibronectin was compromised in the ISR group. There were also fewer EPCs in patients with diffuse, compared with focal, ISR. Interestingly, and disturbingly from the point of view of DES implantation, sirolimus (one of the drugs used on DESs) also inhibits the proliferation and differentiation of human EPCs in vitro . A combination of antiproliferative (drug) and pro-endothelializing (cell) therapy may, therefore, need careful handling.
Taken together, these results suggest that EPCs participate in the formation of neointima, with possible phenotypic transformation into VSMCs (vascular smooth-muscle cells); but that they exert a negative controlling influence in the extent of tissue growth. Where their function is impaired, restenosis is more likely to occur. Overall, therefore, the role of EPCs in the context of restenosis appears to be protective, both in animal models and clinically, and we currently have a large programme of work studying this.
Injury and stent geometry
We examined the contribution of a wide panel of parameters of stent geometry to the development of neointima in stent sections that displayed stretch but not deep injury and showed that stent-induced stretch of coronary arteries results in a neointima whose magnitude is directly related to a number of simple parameters of stent geometry. In particular, neointimal thickness at a strut is related to strut protrusion, inter-strut distance, medial bowing, strut depth and angular burden. Neointimal area is related to major axis, mean medial bowing, mean angular burden and mean strut protrusion. It has been suggested that stent design is unimportant to restenosis. We have shown this to be incorrect. While novel stent designs are directed by industrial ingenuity towards enhancing deliverability, conformability and radial strength (all highly desirable attributes), control of the biological response may also be possible through careful manipulation of stent design, to enhance the beneficial effect of stent coatings and drugs .
Stents and complex lesion geometries
Bifurcation lesions also remain a significant challenge for the interventional cardiologist. A previous study by Louvard et al.  summarized no fewer than 11 different deployment techniques. These usually involve the deployment of two or more stents within the bifurcation and modifying them in situ to take on the geometry of the vessel. No studies on the haemodynamic effects of the different stent deployment techniques have been carried out, and Louvard did not speculate as to the haemodynamic effects of the techniques. Consequently, subjective criteria such as operator experience and simplicity remain the arbiters of which method is chosen. No consensus exists of a technical success other than angiographic appearance possibly combined with intravascular ultrasound. No regard is given to the level of blood flow post intervention. Our own studies are currently under way to examine the deployment characteristics of the different techniques. We are able to deploy stents in both rigid and flexible in vitro models of bifurcations, utilizing a range of bifurcation geometries. These stents can then be imaged using microCT (micro computed tomography). Preliminary results show highly variable deployment characteristics. These studies are being extended to include a flow modelling aspect, so that shear stress and flow patterns can be studied.
Stents and LMS (left main stem) disease
One of the final frontiers open to interventional cardiologists is treatment of the LMS. An LMS stenosis is a class I indication for CABG [ACC (American College of Cardiology)/AHA (American Heart Association) guidelines] and has been a relative contraindication to PCI, partly because of the catastrophic consequences of abrupt vessel closure and restenosis seen in the pre-stent and early stent era [35,36], and partly due to results showing improved survival (compared with conservative treatment) after CABG . In the last 10 years, due to the considerable evolution in stent technology, implantation techniques, antiplatelet therapy and drug-eluting coatings, LMS lesions are increasingly being treated by PCI. As yet, no single, widely accepted technique for treating LMS bifurcation lesions with stents exists. Our group has studied the ‘shotgun’ (‘kissing’ or ‘V stent’) technique as it seems particularly suited for such lesions. Since 2000, we have offered PCI, with a policy of full revascularization whenever possible, to all patients presenting to our team with LMS ± other vessel diseases as the default strategy. We have consistently used a single stent approach for disease of the ostium and body of the LMS, and SKS (simultaneous kissing stents) for disease of the bifurcation. Treating all-comers, we obtained a 100% technical success rate. There was one in-hospital death in a patient with cardiogenic shock. The ischaemia-driven TLR (target lesion revascularization) was 3.4% at 6 months. We also performed animal experiments to demonstrate re-endothelialization and healing [38,39].
In conclusion, despite huge technical strides made in the last three decades, the challenges facing the interventional cardiologist today have never been greater. Stents have contributed greatly to the safety of PCI, but their utility is limited by their physical properties, inflexibility and durability. DESs, while limiting restenosis, have created new problems in the form of a low, but continuing, risk of late stent thrombosis. Nor do they directly address the problem of the inflammation associated with ACSs. A much gentler, kinder approach is needed, in which the stenosed artery is supported, but in which inflammation is limited and healing is promoted without scar formation. There is plenty to be done in the fields of stent design, technology, material, polymer, drug, cell and molecule.
Cardiovascular Bioscience: A Focus Topic at Life Sciences 2007, held at SECC Glasgow, U.K., 9–12 July 2007. Edited by S. Kennedy (Strathclyde, Glasgow, U.K.), M. Lloyd (Bath, U.K.) and C. Wainwright (Robert Gordon University, Aberdeen, U.K.).
acute coronary syndrome
coronary artery bypass grafting
coronary artery disease
endothelial progenitor cell
IL-1 receptor antagonist
left main stem
major adverse cardiac events
percutaneous coronary intervention
percutaneous transluminal coronary angioplasty