Coronary artery revascularisation: selecting the appropriate strategy
- Funding: None.
- Conflict of interest: None.
Advances in interventional cardiology, particularly in better efficacy and safety of drug-eluting stents, have made percutaneous coronary revascularisation practical for most patients requiring revascularisation. While this has reduced the perceived need for coronary bypass surgery, it has also focused attention on the appropriate use of coronary stents and the complexity of choosing the right revascularisation strategy. To achieve the best outcomes, it would seem that collaboration rather than competition between cardiac surgeons and interventional cardiologists is necessary.
In the last few years, there have been major advances in the practice of interventional cardiology. Advances in stent design and adjunctive medical therapy, as well as improvements in techniques for tackling challenging lesions, have made percutaneous intervention a possible revascularisation option for almost all patients. Unfortunately, the rate of these developments has often been faster than randomised clinical trials that can be undertaken to test the relative efficacy of new treatment options, and although percutaneous coronary revascularisation is now possible in most cases where there is coronary stenosis or occlusion, it is not always the appropriate treatment. As a result, there have been controversies surrounding the appropriateness of its use instead of medical therapy or surgery. With constant advances in surgical and interventional technology and techniques, the treatment selection for individual patients is often very complex. These factors have led to a major shift in the way that interventional cardiologists and cardiac surgeons interact in assessing the challenges of coronary revascularisation as well as the treatment of structural heart disease. Many centres now have a ‘Heart Team’ that involves input from interventional cardiologists, cardiac surgeons, treating physicians, intensive care physicians, anaesthetists and often other medical specialists, such as geriatricians, to help make the appropriate choice for a patient. In some cases, this may involve percutaneous solutions, sometimes surgical and in some cases a hybrid approach. These issues are explored in more detail later.
Improvements in percutaneous coronary intervention
It is many years since percutaneous coronary intervention (PCI) was characterised by the frequent need for urgent bailout surgery, with the development of coronary stents seeing this falling from around 20% of cases to less than 0.5%. This has allowed percutaneous coronary interventions to be performed with less attention to the availability of cardiac surgical support, thus improving patient access, particularly in rural and regional centres.
Although stents revolutionised the utilisation of PCI, for many years, their clinical application has been greatly limited by restenosis. For example, a diabetic patient with a coronary vessel less than 2.5 mm in diameter, the rate of restenosis approaches 50%. The pathophysiology of this process is complex and involves endothelial disruption, platelet and leukocyte activation, and ultimately smooth muscle cell proliferation. The targeting of these processes, particularly smooth muscle cell proliferation and activation, has led to effective strategies for preventing restenosis. The most important of these developments has been the evolution of drug-eluting stents. These consist of a metallic stent, with a polymer coating containing an antiproliferative drug that elutes over a period of time. The results of these first-generation drug-eluting stents using paclitaxel and sirolimus were unprecedented. Reductions in neointimal hyperplasia within the stent were dramatic with often no detectable lumen reduction on follow up.
Although the first generation of drug-eluting stents greatly reduced the problem of restenosis, after their widespread use, a number of problems soon became evident. The major problem has been the effect that the drug has on normal healing of the vessel wall resulting in a higher risk of thrombosis if dual antiplatelet therapy is discontinued in the first 12 months. This had a major impact on the universal applicability of drug-eluting stents. Patients who may require surgery within 12 months, those who are unable to comply with or tolerate dual antiplatelet therapy or who may have bleeding problems should probably receive bare-metal stents. The other major concern with first-generation drug-eluting stents has been that of late stent thrombosis more than 12 months after their implantation, often after cessation of dual antiplatelet therapy. In 2007, Stone and colleagues performed a pooled analysis of 1748 patients randomised to either bare-metal stents or first-generation drug-eluting stents. They found that in the paclitaxel- or sirolimus-eluting stent group, there were 14 cases of late stent thrombosis between 1 and 4 years compared with two cases in the bare-metal stent group, even though the need for repeat revascularisation was much higher in this latter group mostly because of restenosis.
Early data from trials of the currently available generation of drug-eluting stents suggests that these problems have been at least partly addressed. The current generation of drug-eluting stents have much thinner stent struts, less polymer coating, and use the drugs zotarolimus and everolimus. Trials comparing this generation of drug-eluting stents have shown reduced levels of revascularisation and stent thrombosis compared with first-generation drug-eluting stents,[7, 8] as well as lower rates of revascularisation compared with bare-metal stents. One of the most important observations of the current generation of drug-eluting stents is the reduction in the risk of stent thrombosis and potentially in the required duration of dual antiplatelet therapy. For example, trials of everolimus-eluting stents have shown a lower rate of stent thrombosis than first-generation paclitaxel-eluting stents. In a study published by Smits et al. in 2011, everolimus-eluting stents had a probable stent thrombosis rate of 0.9% compared with 3.9% for paclitaxel-eluting stents at 2 years. A meta-analysis published this year also found a lower rates of stent thrombosis in everolimus-eluting stents compared with bare-metal stents at 2 years.[10, 11] Similarly, a pooled analysis of 919 patients presented this year found no stent thrombosis when dual antiplatelet therapy was interrupted between 3 and 12 months.
Interaction of PCI and coronary artery bypass grafting
Both PCI and coronary artery bypass grafting (CABG) have proven effective in controlling the symptoms of myocardial ischaemia, and in many cases, both may be reasonable treatment alternatives. Although stenting has become more universally applicable, coronary bypass grafting remains an excellent revascularisation technique. Several variables influence the choice of treatment, for example clinical presentation, patient preference, the extent and severity of coronary artery disease, left ventricular function, and other comorbidities. While in some clinical scenarios, the choice may be obvious; in a variety of others, the treatment choice is much more difficult, particularly in those areas where there is a paucity of randomised clinical trial results. Although guidelines for revascularisation are published by a number of bodies, such as the American College of Cardiology/American Heart Association, in many cases, the level of evidence for each strategy is low. Examples of some of these guidelines for different scenarios are given in Table 1.
Table 1. Examples of guidelines for revascularisation in various clinical scenarios
|Asymptomatic single-vessel disease||III (B)||III (B)||N/A|
|Symptomatic stable single- or two-vessel disease despite OMT||I (B)||I (B)||N/A|
|Two-vessel disease without proximal LAD lesion or extensive ischaemia (with or without symptoms)||IIb (B)||IIb (C)||N/A|
|Two-vessel disease with proximal LAD lesion (with or without symptoms)||IIb (B)||I (B)||I (C)|
|One-vessel (proximal LAD)||IIb (B)||IIa (B)||I (C)|
|Three-vessel disease (with or without symptoms)||IIb (B)||I (B)||I (C)|
|Unprotected LM or complex multivessel disease||IIa (B)||I (B)||I (C)|
|Previous CABG with symptomatic lesion despite OMT||IIa (C)||IIb (C)||I (C)|
|ST elevation myocardial infarction||I (A)||N/A||N/A|
PCI has very much become the treatment of choice in acute coronary syndromes, with most major hospitals offering 24-h access to primary PCI for ST elevation myocardial infarction (STEMI). Indeed, door-to-balloon times in these cases have become key performance indicators for cardiology units, leading to innovations such as ambulance triage of patients and emergent transfer from hospitals not able to offer PCI. Similarly, the use of PCI in non-STEMI and unstable angina with refractory symptoms where there is single or multivessel disease is clearly indicated. While revascularisation may also improve symptoms in stable coronary artery disease, it has a prognostic impact only in certain anatomical subsets. This includes three-vessel and left main disease, although some other cases, for example proximal left anterior descending artery disease where there is extensive ischaemia, may also benefit. The treatment of this group of patients has been the topic of much debate.
Although a large number of randomised trials have been conducted comparing PCI and CABG in stable patients with multivessel coronary artery disease, the data are often difficult to apply to current practice because of the rapid pace of technological advancement. For example, one of the most often quoted trials, the Bypass Angioplasty Revascularisation Investigation (BARI) study was performed before the routine use of stents and before the advent of modern antiplatelet agents. Nevertheless, this trial showed no survival difference between those randomised to PCI compared with CABG, even out to 10-year follow up, although the PCI group had a higher incidence of repeat interventions, and subgroup analysis showed that patients with diabetes fared better with CABG both in terms of survival and repeat interventions. Several other trials in the era prior to drug-eluting stents found similar results with no difference in survival between those randomised to CABG or PCI. Bravata and colleagues recently conducted a systematic review of 23 randomised controlled trials of almost 10 000 patients randomised to either PCI or CABG. They found that although CABG was more effective than balloon angioplasty or bare-metal stents at relieving angina, CABG had a higher rate of procedural stroke and no survival benefit at 10 years.
Although meta-analysis of the major randomised trials has consistently shown a higher rate of repeat procedures with PCI compared with CABG, there are few data from trials where modern drug-eluting stents have been used, and even less where the current generation of drug-eluting stents have been used. Particularly, considering the survival benefit seen in diabetic patients undergoing CABG in the BARI study, randomised, controlled data for drug-eluting stents in diabetic patients are of great interest. Initial results from the Arterial Revascularisation Therapies Study II were encouraging for an approach using drug-eluting stents in diabetic patients with multivessel disease. In this study, at 1-year postprocedure, the 89.2% of patients treated by PCI were free of major adverse cardiac events compared with 88.4% of historical CABG controls and 73.5% of historical PCI controls. Unfortunately, this trial did not have contemporaneous surgical controls, although it was suggested further trials were warranted in this area. Although several trials are addressing this area, only one large, randomised, controlled study comparing CABG with PCI using drug-eluting stents has been published. This was the synergy between percutaneous coronary intervention with TAXUS drug-eluting stent and cardiac surgery (SYNTAX) trial that randomised 1800 patients with complex multivessel or left main disease to either CABG or PCI using paclitaxel-eluting stents. At 3 years of follow up, there was no significant difference in death and stroke rate between the two groups; however, there was an excess of repeat procedures and myocardial infarction in the stent group, giving a combined major adverse cardiac event rate of 20.2% in the CABG group and 28.0% in the stent group. Perhaps most interesting of all in this trial was the development and use of the SYNTAX score, which aimed to give a measure of the extent and severity of coronary disease. A post-hoc analysis of the trial data revealed that those patients with a low SYNTAX score fared similarly and in some cases better with stenting, where as those with more complex disease clearly benefited from CABG. When there is left, main disease with a low SYNTAX score stenting may be an acceptable alternative to CABG; however, the presence of diabetes or complex disease may make CABG desirable at this stage.[2, 16] Having said this, it must be remembered that although the SYNTAX trial generated interesting data, it still used first-generation drug-eluting stents, which potentially have a higher rate of stent thrombosis than those now available.
Appropriate use of revascularisation
The enormous advances in interventional cardiology technology and techniques have made PCI possible in almost all clinical and anatomical situations. However, while percutaneous intervention may be universally applicable, it is not universally appropriate. The spotlight has been well and truly focused on this problem by the recent gaoling of an interventional cardiologist in the USA for inappropriate use of coronary stenting. In addition, there has been publication of data showing a possible lack of benefit of stenting in certain situations.[17, 18] The Clinical Outcomes Utilising Revascularisation and Aggressive Drug Evaluation (COURAGE) trial, a randomised trial conducted in the USA and Canada, found that as a treatment for stable angina PCI is effective at improving quality of life and reducing angina but compared with modern medical therapy has no mortality benefit. Around the same time, there was evidence from Hochman and colleagues that PCI for an occluded coronary artery following recent infarction in that territory was not beneficial. Recently, in response to a number of these factors, expert technical panels in the USA have developed appropriate use criteria ratings for PCI. Interestingly, when these criteria were tested against physicians, there was a 94% concordance where the panel had found that PCI was appropriate but more disagreement in those situations that the panel found inappropriate. An assessment of over 500 000 PCI cases undertaken in the USA between July 2009 and September 2010 found that by these criteria, almost all acute PCI undertaken was appropriate, but for non-acute indications, 12% of cases were inappropriate. These findings suggest that it is likely several inappropriate PCI are still being performed. Keeping appropriate use criteria for PCI relevant in the face of rapid technical developments and new clinical trial data will be challenging and will require constant review.
One of the most exciting of these developments is in the haemodynamic guidance of PCI. It is possible that this technology may once again change what is an appropriate use of PCI. The Fractional Flow Reserve (FFR) Versus Angiography for Multivessel Evaluation (FAME) study, which was published in 2010, studied patients with multivessel coronary disease who were thought to be suitable for PCI. The coronary lesions were assessed for haemodynamic significance using an intracoronary pressure measuring wire to compare the pressure in the vessel distal to the lesion with aortic pressure following vasodilatation with adenosine. This measure, known as the FFR, was found to be highly predictive of the need for PCI, with those patients undergoing FFR-guided PCI having a lower rate of mortality and myocardial infarction at 2 years. Perhaps of even more impact has been the FAME II study, which used the same technology but compared optimal medical therapy against stenting for non-acute lesions that were haemodynamically significant by FFR measurement. The data safety monitoring board recommended this year that enrolment in this trial cease after an interim analysis found a highly significant decrease in hospital readmission and urgent revascularisation in the group receiving FFR-guided PCI. If the results of this trial are confirmed, this could have a massive impact on what defines an appropriate use of PCI. Compared with the COURAGE trial that found no or minimal advantage to PCI compared with optimal medical therapy in stable disease, these results suggest that a haemodynamic guided approach to PCI may offer advantages over medical therapy alone.
Heart team approach
Despite the availability of guidelines to help choose between CABG and PCI, and the availability of appropriateness criteria for PCI, there are many situations that do not fit these categories. In many cases, this requires communication between surgeons, interventional cardiologist, referring cardiologist and imaging specialists, or in other words, the development of a heart team. This facilitates an individualised approach that might consider evidence from medical literature and recent advances. Indeed, at many centres, we are already seeing cooperation between surgeons and interventional cardiologists in the delivery of transcatheter valve treatments and hybrid revascularisation solutions. American Heart Association guidelines also suggest that a team approach is the best in dealing with unprotected left main diseases, as well as complex multivessel coronary disease. It is possible that these changing patterns of resource utilisation will need to be considered in future when planning hospital facilities and organising patient triage.
There have been great leaps forward in the few years in terms of the efficacy and safety of PCI, making percutaneous revascularisation a technically feasible treatment option in almost every case of coronary artery disease. While this has given more options for treatment, it has also led to increased complexity when choosing which option is right for an individual patient. The appropriate application of PCI for individual cases is vital to avoid unnecessary complications, and the increased use of haemodynamic guidance of stenting may prove useful in achieving this. Nevertheless, more evidence is needed to support stenting in preference to surgery in certain settings, most notably in diabetic patients with multivessel disease, patients with high SYNTAX scores, multiple chronic total occlusions, bifurcation lesions and proximal left anterior descending artery lesions. To achieve the best outcomes, the optimal approach would seem that of collaboration rather than competition between cardiac surgeons and interventional cardiologists.