Revascularization strategies for patients with stable coronary artery disease


  • J. Iqbal,

    1. South Yorkshire Cardiothoracic Centre, Northern General Hospital, Sheffield, UK
    2. Thorax Centre, Erasmus MC, Rotterdam, the Netherlands
    Search for more papers by this author
  • P. W. Serruys

    Corresponding author
    1. Thorax Centre, Erasmus MC, Rotterdam, the Netherlands
    2. International Centre for Circulatory Health, Imperial College London, London, UK
    • Correspondence: Patrick Serruys, Department of Interventional Cardiology, Thorax Centre, Erasmus Medical Centre, Rotterdam, the Netherlands.

      (fax: +31 107039154; e-mail:

    Search for more papers by this author


Patients with coronary artery disease who have prognostically significant lesions or symptoms despite optimum medical therapy require mechanical revascularization with coronary artery bypass grafting (CABG), percutaneous coronary intervention (PCI) or both. In this review, we will evaluate the evidence-based use of the two revascularization approaches in treating patients with coronary artery disease. CABG has been the predominant mode of revascularization for more than half a century and is the preferred strategy for patients with multivessel disease, especially those with diabetes mellitus, left ventricular systolic dysfunction or complex lesions. There have been significant technical and technological advances in PCI over recent years, and this is now the preferred revascularization modality in patients with single-vessel or low-risk multivessel disease. Percutaneous coronary intervention can also be considered to treat complex multivessel disease in patients with increased risk of adverse surgical outcomes including frail patients and those with chronic obstructive pulmonary disease. Improvements in both CABG (including total arterial revascularization, off-pump CABG and ‘no-touch’ graft harvesting) and PCI (including newer-generation stents, adjunctive pharmacotherapy and intracoronary imaging) mean that they will continue to challenge each other in the future. A ‘heart team’ approach is strongly recommended to select an evidence-based, yet individualized, revascularization strategy for all patients with complex coronary artery disease. Finally, optimal medical therapy is important for all patients with coronary artery disease, regardless of the mode of revascularization.


Coronary artery disease (CAD) remains the single most important cause of morbidity and mortality worldwide. There have been substantial advances in medical therapy to prevent and treat CAD. However, patients with prognostically significant disease or symptoms of angina despite optimal medical therapy (OMT) require mechanical revascularization: either coronary artery bypass grafting (CABG) or percutaneous coronary intervention (PCI). CABG has remained the predominant mode of revascularization during the second half of 20th century. PCI has become the preferred method of revascularization in patients with single- or double-vessel disease without left main stem (LMS) involvement. However, the optimal strategy in patients with multivessel disease and/or unprotected LMS disease has remained controversial and the subject of many clinical trials in recent years. In this review, we discuss the current status and the evidence-based use of these two revascularization strategies in the treatment of patients with stable CAD.

Optimal medical therapy

The main focus of this review is revascularization; however, it is essential to first highlight the importance of OMT, which encompasses pharmacotherapy for CAD, good control of cardiovascular disease risk factors and lifestyle modifications (Table 1).

Table 1. Optimal medical treatment for patients with stable coronary artery disease
  1. BMI, body mass index; LDL, low-density lipoprotein; HbA1c, haemoglobin A1c; ACS, acute coronary syndrome; PCI, percutaneous coronary intervention; ACE, angiotensin-converting enzyme; LV, left ventricular.

Lifestyle modificationSmoking cessationAdvice, encouragement and pharmacological aid with nicotine-replacement therapy
 Healthy dietLimit saturated fats, five servings of fruit/vegetables daily and fish twice per week
 Physical activityModerate-intensity aerobic exercise for 30 min at least three times per week
Control of risk factorsObesityAim for BMI <25 kg m−2 by healthy eating and exercise
 HypertensionAim for blood pressure lower than 140/85 mmHg
 DyslipidaemiaAim for LDL level below 1.8 mmol L−1; usually achievable with statin monotherapy
 Diabetes mellitusAim for HbA1c < 7.0%
Antianginal drugsFirst-line drugsBeta-blockers, nitrates and calcium channel blockers
 Second-line drugsLong-acting nitrates, ivabradine, nicorandil and ranolazine
Event preventionAntiplatelet drugsAspirin for all (unless contraindicated); add P2Y12 inhibitors for high-risk, post-ACS or post-PCI
 OthersStatins; ACE inhibitors if hypertension or LV systolic dysfunction

Optimal medical therapy as first-line treatment of CAD

The Medical, Angioplasty, or Surgery Study-II (MASS-II) trial (= 611), though underpowered for outcomes, has shown no difference in survival between OMT and revascularization (OMT 69%, CABG 74.9%, PCI 75.1%, = 0.089) at the 10-year follow-up, despite differences in the rates of myocardial infarction (MI) and repeat revascularization [1]. In the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial (= 2287), it was shown that amongst patients with significant one-, two- and three-vessel CAD without left main stem (LMS) involvement randomly assigned to receive OMT either alone or plus PCI, there was no significant difference in the composite end-point of death or nonfatal MI at a median follow-up of 4.6 years [hazard ratio (HR) for the PCI group 1.05, 95% confidence interval (CI) 0.87–1.27, = 0.62] [2]. Despite frequent crossover to revascularization, the majority (67%) of patients remained on OMT alone for the duration of the trial. Both groups were also equivalent in terms of freedom from angina at 5 years [2]. Other studies comparing OMT with PCI have also reported no mortality benefit, increased nonfatal periprocedural MI and reduced need for urgent revascularization with PCI compared with OMT [3]. Therefore, OMT is the recommended initial choice for patients with stable CAD without significant disease of the LMS or proximal left anterior descending artery (LAD) [4].

Optimal medical therapy for all patients undergoing revascularization

Optimal medical therapy is important and complementary to revascularization therapy, whether PCI [2] or CABG [5]. Progression of atherosclerosis continues after revascularization and is associated with deterioration of left ventricular function. However, appropriate use of secondary prevention medications reduces mortality and the incidence of MI after revascularization [6].

Indications for revascularization

Careful evaluation is needed before a decision is made to embark on surgical or percutaneous revascularization. This decision is usually based on the patient's symptoms or ischaemic burden in the presence of significant obstructive coronary artery stenosis.

Resistant symptoms

Patients with significant CAD and symptoms despite OMT should be considered for revascularization [4]. Use of standardized questionnaires (e.g. the Seattle Angina Questionnaire) or objective assessment with exercise testing can be helpful. If the significance of a lesion is uncertain, assessment of the fractional flow reserve (FFR) is strongly recommended [7].

Large ischaemic burden

A large ischaemic burden has an adverse impact on clinical outcomes. In patients with a moderate or large degree of inducible ischaemia, the survival benefit (absolute and relative) was greater when treated with revascularization, as compared with OMT [8]. The nuclear substudy of the COURAGE trial has also shown that the degree of reduction in ischaemic myocardium was significantly greater with PCI than with OMT (2.7% vs. 1.6%, < 0.0001) [9].

Left ventricular dysfunction

Revascularization therapy is of benefit in patients with left ventricular systolic dysfunction and angina due to significant CAD [4]. However, amongst patients with ischaemic heart failure but no angina, revascularization therapy should be limited to those who have ischaemic but viable myocardium on noninvasive testing as revascularization in these patients has been shown to improve left ventricular function and survival [10].

Revascularization techniques

Revascularization can be achieved surgically (CABG) or percutaneously (PCI), or by a hybrid approach.

Coronary artery bypass grafting

Vasilii Kolesov is believed to have been the first surgeon to perform an anastomosis between the left internal mammary artery (LIMA) and the left circumflex artery in man in 1962. Rene Favaloro used a saphenous vein graft (SVG) as a bypass conduit in 1967. Reed was the first surgeon to perform the operation using cardiopulmonary bypass. Carpentier used radial artery graft for CABG in 1973. The history of CABG has been reviewed in detail elsewhere [11].

The superiority of CABG over OMT has been demonstrated in multiple studies and meta-analyses. CABG confers a survival benefit in patients with unprotected LMS or three-vessel CAD, particularly in those with severe symptoms, early positive exercise tests and/or impaired left ventricular function [12].

Percutaneous coronary intervention

Balloon angioplasty was first performed by Andreas Grüentzig in 1977. Coronary stents were developed subsequently to overcome the problems of dissection, elastic recoil and constrictive remodelling with angioplasty [13, 14]. However, the medium- and longer-term outcome using these bare metal stents (BMSs) was compromised by a high incidence of in-stent restenosis, and the drug-eluting stents (DESs) were therefore developed to reduce restenosis and the need for target vessel revascularization (TVR) [15, 16]. The concern of stent thrombosis with the first-generation DESs [17] led to the development of novel polymers, antiplatelet agents and the ‘newer generation’ of DESs [18]. The newer-generation DESs have been shown to reduce major adverse cardiac events (MACEs) [19, 20]. The historical development of coronary stents is reviewed elsewhere [18].

Hybrid revascularization

The concept of hybrid revascularization is to combine the prognostic benefits of the LIMA-to-LAD graft with minimal invasiveness and PCI of non-LAD significant lesions. This approach has largely been used for patients with limited conduit availability or predicted healing problems after sternotomy, or following primary PCI of a non-LAD culprit lesion. However, hybrid revascularization may provide a minimally invasive alternative to traditional CABG. In a single-centre study of 300 consecutive patients, hybrid revascularization resulted in reasonably good outcomes: 30-day mortality, stroke and nonfatal MI were observed in four (1.3%), three (1.0%) and four (1.3%) patients, respectively [21]. In a propensity matched study, consecutive patients undergoing one-stop hybrid revascularization (n = 141) were compared with patients who underwent isolated CABG or PCI (n = 141). After stratification by EuroSCORE or SYNTAX score, the cumulative major adverse cardiac and cerebrovascular event (MACCE) rates were similar in the low and medium tertiles amongst the three groups. However, in the top EuroSCORE tertile, the hybrid group had a lower MACCE rate than the CABG (= 0.030) and PCI (= 0.006) groups. Additionally, amongst patients with a high SYNTAX score, in the hybrid group, the MACCE rate was lower than in the PCI group (= 0.002), but similar to that of the CABG group (= 0.36) [22].

The limited available data on hybrid revascularization appear promising. However, little progress has been made in this field over recent years, and without an appropriately powered randomized trial of hybrid versus traditional revascularization, the uptake of this approach is unlikely to increase.

Selection of revascularization technique

The choice between CABG and PCI should be based on a careful evaluation of the extent of CAD, the expected completeness of revascularization, the presence of comorbidities and any associated significant valvular disease [4].

Extent of CAD

The anatomical extent and complexity of CAD play an important role in decision-making regarding revascularization strategy (Fig. 1).

Figure 1.

A potential revascularization pathway for patients with stable angina pectoris. OMT, optimal medical therapy; FFR, fractional flow reserve; CAD, coronary artery disease; LMS, left main stem; PCI, percutaneous coronary intervention; CABG, coronary artery bypass graft.

One- to two-vessel disease without LMS involvement

Patients with one- or two-vessel disease not involving the LMS can be treated with PCI, if technically feasible to achieve complete revascularization. PCI can provide good outcomes in such patients and may be the preferred revascularization strategy [4]. Percutaneous coronary intervention and CABG can provide similar levels of the long-term improvement in quality of life in this group of patients [23].

Historically, proximal LAD disease has been treated with CABG. However, recent studies have shown that PCI using DESs is noninferior to CABG [24, 25]. Therefore, international guidelines have accepted PCI as a reasonable alternative to surgery for the treatment of isolated proximal LAD disease [4].

Complex or triple vessel disease

The historical data showing clear superiority of CABG over balloon angioplasty or PCI with BMS for patients with multivessel disease have limited applicability in contemporary practice [26].

With the advent of DESs, indirect comparison with CABG was attempted by the addition of DES arms to the original PCI with BMS versus CABG trials, for example in the Arterial Revascularization Therapies Study II (ARTS-II) and Argentine randomized trial of coronary angioplasty with stenting versus coronary bypass surgery in patients with multiple vessel disease (ERACI-III) studies. In the ARTS-II study, higher rates of repeat revascularization were observed in the PCI arm (20.8% vs. 9.0%, < 0.001), but there was no difference in survival between PCI with DES and CABG (94.5% vs. 92.6%, respectively) at the 5-year follow-up [27]. In the ERACI-III study, similar event rates were also observed in the CABG and PCI with DES groups after follow-up for 3 years (5.7% vs. 9.8%; relative risk 0.59, 95% CI 0.31–1.14) [28]. The comparison between CABG and PCI with DES in various registries has produced mixed results: the Asan Medical Center-Multivessel Revascularization Registry found no difference in mortality at 5 years [29]; the New York State registry reported similar unadjusted survival (PCI 93.7% vs. CABG 93.4%) but different risk-adjusted survival for the two techniques (PCI 94.0% vs. CABG 92.7%, = 0.03) at 18 months [30]; and ASCERT (the ACCF-STS Database Collaboration on the Comparative Effectiveness of Revascularization Strategies) demonstrated no mortality difference at 1 year (PCI 6.55% vs. CABG 6.24%; HR 0.95, 95% CI 0.90–1.0) but lower mortality with CABG at 4 years (PCI 20.8% vs. CABG 16.4%; HR 0.79, 95% CI 0.76–0.82) [31].

The SYNTAX (SYNergy between percutaneous coronary intervention with TAXus and cardiac surgery) trial is the largest (= 1800) contemporary ‘all-comers’ study of patients with complex CAD. At the 5-year follow-up, CABG was superior to PCI with a first-generation DES: MACCE (PCI 37.3% vs. CABG 26.9%, < 0.001), all-cause death (PCI 13.9% vs. CABG 11.4%, = 0.10), MI (PCI 9.7% vs. CABG 3.8%, < 0.001), stroke (PCI 2.4% vs. CABG 3.7%, = 0.09) and repeat revascularization (PCI 25.9% vs. CABG 13.7%, < 0.001) [32]. The subgroup analysis of this trial showed that in the tertile of patients with the lowest SYNTAX scores (0–22), there was no significant difference in the MACCE rate between the two groups (PCI 32.1% vs. CABG 28.6%, = 0.43). However, CABG outperformed PCI in the intermediate (23–32 score) (PCI 36.0% vs. CABG 25.8%, = 0.008) and top (≥33 score) tertiles (PCI 44.0% vs. CABG 26.8%, < 0.001) for MACCE rate at 5 years [32]. These outcomes are also consistent with the findings of several other studies [30, 31]. It has been shown that combining anatomical SYNTAX score with clinical characteristics (SYNTAX score II) is useful for selecting a favourable revascularization strategy [33]. Although prospective validation of the SYNTAX score II is needed, we would advocate using this score for decision-making in a multidisciplinary setting (Fig. 1).

Unprotected LMS disease

Left main stem disease has remained a class III indication for PCI (i.e. the procedure is generally not effective and may even be harmful) according to international guidelines, due to historical data showing high rates of complications and modest outcomes [34]. However, recent data from multiple registries have shown that PCI using DESs in patients with LMS disease has similar mortality and safety outcomes, but a higher rate of TVR, compared with CABG [35-37]. Several randomized trials, including LEMANS [38], SYNTAX left main [32], Boudriot [39] and PRECOMBAT [40], have compared PCI with DES against CABG for the treatment of LMS stenosis (Table 2). In a recent meta-analysis of four randomized trials including 1611 patients, it was shown that PCI, as compared with CABG, was associated with a lower risk of stroke, increased risk of repeat revascularization and similar risk of mortality or MI, resulting in a higher risk of MACEs but a similar risk of MACCEs [41]. Based on these recent data, PCI for LMS has been upgraded to a class IIb indication in current guidelines and may be considered for patients with coronary anatomy that is associated with a low risk of procedural complication if treated by PCI and/or clinical conditions that predict an increased risk of adverse surgical outcomes [42]. Therefore, we recommend that patients with LMS disease and SYNTAX scores between 0 and 32 can be treated with PCI using DESs when technically feasible, whereas CABG surgery should remain the standard treatment in patients with SYNTAX scores ≥33.

Table 2. Clinical trials comparing PCI with DES and CABG for left main stem disease
TrialNumber of patientsMean SYNTAX scoreMean logistic EuroSCOREFollow-upOutcomes (PCI vs. CABG),%
MortalityMIStrokeRepeat revascularizationMACE/MACCE
  1. PCI, percutaneous coronary intervention; DES, drug-eluting stent; CABG, coronary artery bypass graft; MI, myocardial infarction; MACE, major adverse cardiac event; MACCE, major adverse cardiac and cerebrovascular event; Pn, P value for noninferiority; NS, nonsignificant.

LEMANS105253.41 year

2 vs. 8,


2 vs. 6,


0 vs 4,


30 vs. 10,

= 0.01

32 vs. 26,


SYNTAX Left Main701303.93 years

7.3 vs. 8.4,

= 0.64

6.9 vs. 4.1,

= 0.14

1.2 vs. 4.0,

= 0.02

20.0 vs. 11.7,

= 0.004

36.9 vs. 31.0,

= 0.12

Boudriot et al. [39]201242.51 year

2 vs. 5,

Pn < 0.001

3 vs. 3,

Pn = 0.002

14 vs. 6,

Pn= 0.35

19 vs. 14,

Pn = 0.19

PRECOMBAT600252.72 years

2.4 vs 3.4,

= 0.45

1.7 vs 1.0,

= 0.49

0.4 vs 0.7,

= 0.56

9.0 vs. 4.2,

= 0.02

12.2 vs. 8.1,

= 0.12

Expected completeness of revascularization

Complete revascularization of complex CAD has been shown to improve clinical outcomes in patients with multivessel disease [43]. A recent meta-analysis has demonstrated that complete revascularization is associated with better outcomes [44]. It is, therefore, important to evaluate whether a given technique can achieve complete revascularization.

Patient comorbidities

Older age, frailty and the presence of multiple comorbidities including chronic obstructive pulmonary disease generally increase surgical risk, and PCI may be the preferred mode of revascularization in elderly, frail and/or comorbid patients [33, 45].

Diabetes mellitus is a risk factor for both revascularization strategies. There have been some conflicting reports on the effect of diabetes on outcome of revascularization [29, 46-49]. However, the Future Revascularization Evaluation in Patients With Diabetes Mellitus: Optimal Management of Multivessel Disease (FREEDOM) study, the largest dedicated contemporary trial in patients with diabetes and multivessel CAD randomly assigned to either PCI with DES or CABG, has clearly shown the superiority of CABG in treating these patients [50]. Therefore, CABG should be the revascularization option of choice for patients with multivessel CAD and diabetes mellitus. However, diabetic patients with less complex CAD can be treated with PCI.

There are conflicting data on revascularization in patients with chronic heart failure; the Asan Medical Center-Multivessel Revascularization Registry found no difference between PCI with DES and CABG in patients with or without abnormal ventricular function [29], whereas data from an Australian registry have shown that heart failure is an independent predictor of 30-day mortality after PCI but not after CABG [49]. The findings of the SYNTAX trial also suggest that patients with multivessel disease and heart failure with low ejection fraction may derive more benefit from surgical revascularization [33].

A ‘heart team’ approach for individualized risk stratification

The selection of revascularization strategy for patients with complex coronary disease remains challenging, and adopting a multidisciplinary ‘heart team’ approach for decision-making after careful consideration of the relevant data is strongly recommended (Fig. 2) [51]. It is best to avoid ad hoc PCI in stable patients with complex CAD, and each case should be discussed by the heart team before a deferred revascularization (PCI or CABG) procedure [4]. There are various tools available to help the heart team in selecting the optimal strategy. The EuroSCORE and the SYNTAX score have been shown to predict adverse outcomes in studies containing both PCI and CABG arms [32, 52]. The recently proposed SYNTAX score II, combining anatomical and clinical factors, may provide an evidence-based approach for decision-making [33]. It is important to acknowledge that all risk models have their limitations; informed patient consent and clinical judgement of the heart team remain vital.

Figure 2.

Factors influencing decision-making for coronary revascularization. CAD, coronary artery disease.

Factors to consider in patients undergoing CABG

Selection of conduit

The choice of conduits for surgical revascularization may influence the long-term outcome. Reduced mortality and morbidity have been demonstrated for the LIMA graft compared to other conduits [53]. It is therefore common practice to offer LIMA to LAD graft and supplemental SVG to other arteries. It is arguable that total arterial revascularization using the bilateral internal mammary arteries (BIMAs) or radial artery may offer additive benefits [54]. The radial artery has been shown to offer better long-term graft and patient survival [55]; however, concerns regarding vasospasm and neointimal hyperplasia remain. Use of the right internal mammary artery appears to be more attractive with patency rates equal to those of LIMA and better than those of the radial artery and SVG [56]. Use of the BIMAs has been shown to offer a survival advantage [57]. However, due to concerns about the risk of infection and impaired sternotomy healing, caution is warranted in the use of the BIMAs for patients with chronic obstructive lung disease and diabetes mellitus.

On-pump or off-pump surgery

Surgical revascularization on the beating heart (off-pump CABG) has a potential to reduce the detrimental effects of cardiopulmonary bypass. In the Surgical Management of Arterial Revascularization Therapies (SMART) trial, patients (= 197) randomly assigned to either off-pump or conventional CABG had similar rates of death, stroke, MI, angina, reintervention and graft patency, and similar quality of life, at 30 days and 1 year [58]. In another trial (= 308), off-pump CABG was associated with fewer grafts and a higher incidence of atrial fibrillation but reduced intensive care unit and hospital length of stay. However, after 5 years of follow-up, there was no difference in MACCE rate between the two groups (HR 0.71, 95% CI 0.41–1.22, = 0.21) [59]. These small studies providing the proof of concept for off-pump CABG led to two large-scale trials. The Randomized On/Off Bypass (ROOBY) trial (= 2203) showed lower graft patency rates with off-pump CABG, suggesting that this technique may not completely revascularize all major coronary arteries [60]. The CABG Off or On Pump Revascularization Study (CORONARY) is the largest (= 4752) multicentre trial comparing off-pump versus on-pump procedures. No significant differences between the two techniques in the rate of the primary composite end-point (death, nonfatal stroke, nonfatal MI or new renal failure requiring dialysis) (9.8% vs. 10.3%; HR 0.95, 95% CI 0.79–1.14, = 0.59) or in any individual component of the primary end-point were shown in this study at 30 days after randomization. However, the use of off-pump CABG significantly reduced the need for blood transfusion (50.7% vs. 63.3%, < 0.001), reoperation due to perioperative bleeding (1.4% vs. 2.4%, = 0.02), acute kidney injury (28.0% vs. 32.1%, = 0.01) and respiratory complications (5.9% vs. 7.5%, = 0.03) but increased the rate of early repeat revascularization (0.7% vs. 0.2%, = 0.01) [61]. A recent meta-analysis has shown a reduction in the incidence of stroke with off-pump CABG [62].

In summary, it may be reasonable to consider off-pump CABG in patients at increased risk of perioperative complications, especially those attributed to cardiopulmonary bypass (inflammation, infection, kidney injury, atrial fibrillation) and aortic manipulation (stroke), if a surgeon experienced in performing off-pump CABG is available. If off-pump CABG is performed, the degree of aortic manipulation should be reduced to a minimum to decrease the risk of neurological complications [63].

Minimally invasive direct CABG

Minimally invasive direct coronary artery bypass (MIDCAB) uses a small anterior left thoracotomy incision to harvest the LIMA and perform an anastomosis to the LAD with or without cardiopulmonary bypass. It is a safe procedure with good short- and long-term results [64]. However, in a small trial in patients (= 130) with isolated proximal LAD disease, MIDCAB was not superior to PCI with DES (in terms of MACE rate at 12 months), although TVR was higher in the PCI arm [25]. Therefore, MIDCAB is primarily performed for LAD lesions unsuitable for PCI or for repeat CABG when sternotomy or cardiopulmonary bypass could be at high risk, or for diffuse aortic calcification. Minimally invasive direct coronary artery bypass to the LAD in conjunction with PCI of other vessels, that is hybrid revascularization, as discussed above, remains an attractive option, but further data to support this approach are needed.

Factors to consider in patients undergoing PCI

Selection of a stent

Although a variety of stents are available for clinical use, the main choice is between BMSs and DESs. Bare metal stents have higher incidence of in-stent restenosis, whereas DESs are expensive and healing may be delayed leading to late stent thrombosis [18].

A meta-analysis of clinical trials comparing DESs with BMSs reported similar rates of all-cause death, cardiac death and nonfatal MI, but a significant reduction in TVR with DESs [65]. By contrast, an unadjusted analysis of 182 901 patients in 34 observational studies of BMSs and DESs demonstrated a significant reduction in mortality (HR 0.78, 95% CI 0.71–0.86) and MI (HR 0.87, 95% CI 0.78–0.97) with DESs [66]. After multivariable adjustment, the benefits of DESs were significantly attenuated. A recent Cochrane review has shown that patients have similar rates of death and MI with BMSs and DESs [67].

It could be argued that all patients without bleeding problems should receive DESs. However, it is also acceptable that shorter lesions (≤15 mm) in bigger vessels (≥3 mm diameter) in nondiabetic patients may be treated with BMSs [68, 69]. Bare metal stents could also be the preferred choice for patients unwilling to take or unlikely to adhere to dual antiplatelet therapy (DAPT). Diabetes mellitus is an independent predictor of in-stent restenosis, and diabetic patients treated with DESs have significantly lower rates of death, acute MI and repeat revascularization than those treated with BMSs [70].

If DESs are to be used, then the ‘newer-generation’ DESs are highly preferable, compared with the ‘first-generation’ stents, as they may reduce cardiovascular complications and TVR [19, 20, 71]; however, there is no clear effect on mortality between the older and newer versions [72].

Selection of antiplatelet regimen

Unless contraindicated, all patients undergoing PCI with stents should receive DAPT to reduce the risk of stent thrombosis. For patients undergoing PCI for acute coronary syndromes, DAPT is recommended for 12 months. However, the duration and choice of antiplatelet agents for patients undergoing PCI for stable angina remain debatable and may depend on the choice of stent (BMS or DES). Dual antiplatelet therapy for a minimum of 1 month with BMSs and 6 months with DESs is generally recommended. However, 3 months of treatment with DAPT for the newer-generation DESs (e.g. Xience stents from Abbot Vascular Ltd, Santa Clara, CA, USA) have also been approved in Europe, based on data suggesting a low incidence of stent thrombosis with these stents [71].

Aspirin should be continued indefinitely, and a low dose (75–100 mg daily) is preferred over higher doses. Clopidogrel is still the most commonly used P2Y12 inhibitor; however, it is a pro-drug that requires hepatic activation by the cytochrome P450 system, and consequently, some patients are resistant to clopidogrel or respond poorly [73]. Therefore, the newer P2Y12 inhibitors prasugrel and ticagrelor have been developed in recent years [73]. These agents have proven benefits for patients with acute coronary syndromes [74, 75]. Currently, there is limited evidence to support the use of these potent antiplatelet agents in patient with stable angina undergoing PCI.

Patients receiving coronary stents who require warfarin are at high risk of bleeding if they also receive DAPT. Omission of aspirin may be advantageous in such patients [76]. Routine platelet function or genetic testing is currently not recommended to tailor antiplatelet therapy after PCI [77].

Adjunctive intracoronary imaging

Adjunctive intracoronary imaging during PCI can be used to guide appropriate sizing and deployment of stents and to exclude any local complication (e.g. dissection). Intravascular ultrasound-guided PCI has been shown to reduce the incidence of adverse outcomes [78]. Optical coherence tomography, which offers considerably higher resolution at the expense of limited penetration, is also a promising tool to optimize stent deployment [79], although further data are needed to establish its precise role in clinical practice.

Future of revascularization strategies

Advances in surgical techniques

Robotic surgery

Minimally invasive direct coronary artery bypass can also be performed with computer-assisted (robotic) surgery; the surgeon, seated at a computer console, introduces instruments through small incisions in the chest and manipulates them with robotic arms. Early results of robotic coronary surgery are promising, and it may play an important role in the coming years [80].

‘No-touch’ SVG harvest

The poor outcome with SVG can partially be due to trauma during harvesting. ‘No-touch’ harvesting can potentially eliminate this problem. The PATENT-SVG study compared markers of vascular injury in 17 patients who had SVGs harvested with the no-touch technique from one leg and using the conventional method from the other leg. Saphenous vein graft segments harvested using the no-touch technique exhibited preserved intimal, medial and adventitial architecture. Furthermore, vascular smooth muscle cell expression of key transcriptional genes involved in aberrant differentiation and phenotypic modulation was significantly reduced in the no-touch harvesting group [81]. Larger studies are required to determine whether no-touch vein harvesting can reduce the incidence of SVG blockage and improve clinical outcomes.

Advances in PCI

Novel DESs

Drug-eluting stents either with biodegradable polymer or with no polymer at all eliminate the long-term undesirable effects of the presence of polymer. These polymer-free DESs contain drugs incorporated into a microporous or nanoporous surface of the metallic stent. These stents are described in detail elsewhere [82]. A short summary is provided in Table 3.

Table 3. Summary of novel CE-approved drug-eluting stents
Stent (manufacturer)PlatformStrut thicknessSurfaceDrugDrug release (days)StudyLate lumen loss (follow-up)
  1. PLA, Poly-lactic acid; PLLA, Poly-L-lactic acid; PCL, Poly caprolactone; PVP, Poly vinylpyrrolidone; PLGA, poly (lactic-co-glycolic) acid.

Stents with biodegradable polymers
Biomatrix Flex (BioSensors)Stainless steel112 μmAbluminal PLABiolimus A945% (30 days)LEADERS0.13 mm (9 months)
Nobori (Terumo)Stainless steel112 μmAbluminal PLABiolimus A945% (30 days)NOBORI-10.11 mm (9 months)
Synergy (Boston Scientific)Platinum–chromium71 μmAbluminal PLGAEverolimus50% (60 days)EVOLVE0.10–0.13 mm (6 months)
Orsiro (Biotronik)Cobalt–chromium60 μmPLLA with silicon carbideSirolimus50% (30 days)BIOFLOW-10.05 mm (9 months)
DESyne BD (Elixir)Cobalt–chromium81 μmAbluminal PLANovolimus90% (90 days)EXCELA BD0.12 mm (6 months)
Supralimus (Sahajanand)Stainless steel80 μmPLLA-PLGA-PCL-PVPSirolimus100% (48 days)SERIES-10.09 mm (6 months)
Mi stent (Micell)Cobalt–chromium64 μmPLGASirolimus50% (30 days)DESSOLVE II0.27 mm (9 months)
Combo (OrbusNeich)Stainless steel100 μmAblumnial biodegradable polymer and luminal anti-CD34 antibodiesSirolimusN/AREMEDEE0.39 mm (9 months)
Polymer-free stents
Yukon  Choice (Translumina)Stainless steel87 μmAbluminal microporous surfaceSirolimus100% (25 days)ISAR TEST-10.48 mm (9 months)
BioFreedom (Biosensor)Stainless steel119 μmAbluminal microporous surfaceBiolimus A990% (3 days)FIM0.17–0.22 mm (12 months)
Cre8 (CID)Cobalt–chromium80 μmAbluminal reservoirs for drugSirolimus100% (90 days)NEXT-Cre80.14 mm (6 months)

Pro-healing stents

The antiproliferative drugs used in DESs not only inhibit proliferation of vascular smooth muscle cells underlying neointimal formation, but also compromise on endothelial regeneration and, hence, increase the risk of stent thrombosis. Different agents to promote endotheial healing have been investigated. Vascular endothelial growth factor-eluting stents were tested; they not only failed to promote endothelialization but also increased neointimal proliferation [83]. The GENOUS (OrbusNeich, Wanchai, Hong Kong) is a stainless steel stent coated with anti-CD34 antibodies to capture endothelial progenitor cells (EPC). A new-generation Combo (OrbusNeich) stent combines EPC-capturing anti-CD34 antibodies on the luminal surface and a sirolimus-eluting biodegradable polymer on the abluminal surface. The Combo stent has been compared in the Randomized Evaluation of an Abluminal sirolimus coated Bio-Engineered Stent (REMEDEE) trial with the first-generation TAXUS Liberté (Boston Scientific Corporation, Natick, MA, USA) stent and was found to be noninferior for angiographic in-stent late lumen loss at 9 months and MACE rates at 12 months [84].

Bioresorbable scaffolds

The bioresorbable stents (scaffolds) provide initial scaffolding similar to that of the metallic stents but undergo gradual bioresorption, so that after a predefined period, the vessel will be free of the metallic cage and could regain its normal function. The absence of any residual foreign material and restoration of endothelial coverage would also reduce the risk of stent thrombosis and the requirement for the long-term treatment with DAPT [85]. The current scaffolds are composed of either magnesium alloys or polymers of poly-lactic acids [85]. The ABSORB BVS (bioresorbable vascular scaffold) (Abbott Vascular Ltd, Santa Clara, CA, USA) and DESolve (Elixir Medical corporation, Sunnyvale, CA, USA) devices have achieved CE marking. An example of a complex PCI case with the use of intracoronary imaging is shown in Fig. 3. A number of bioresorbable devices are currently undergoing clinical and preclinical evaluation (Fig. 4).

Figure 3.

Percutaneous coronary intervention of a chronically occluded left anterior descending artery (LAD) using the Absorb bioresorbable vascular scaffold (BVS). Coronary angiography revealed a chronically occluded LAD (A), which was successfully opened using the Absorb BVS (B). Illustrative images using intravascular ultrasound (C) and optical coherence tomography (D) in a patient who underwent revascularization using the Absorb BVS (in Absorb cohort B study) are shown at various follow-up time-points.

Figure 4.

Design of clinically tested bioresorbable scaffolds.

Novel revascularization approaches

Various novel revascularization approaches, mechanical and pharmacological, are currently being investigated. Transmyocardial revascularization is based on creating small (1-mm diameter) channels mechanically or using a laser along the left ventricular free wall. Although some studies have shown no effect [86], others have demonstrated an improvement in angina pectoris and functional capacity [87, 88]. This procedure is currently indicated only if PCI and CABG are not possible, and the patient has resistant symptoms on maximal therapy. Further studies of transmyocardal revascularization via a minimally invasive or endocardial approach are needed. Additionally, interest is increasing in the concept of pharmacological revascularization using local or systemic cell-, gene- or drug-based therapy to promote angiogenesis. This approach may prove to be useful for patients with diffuse disease not amenable to PCI or CABG. It is possible that expansion of the microvascular bed may also induce enlargement of upstream collateral arteries via gap junction-mediated retrograde signalling and increased shear stress, resulting in effective perfusion downstream of the occlusion and hence providing a ‘biological bypass’ [89]. Several agents have shown promising results in preclinical or first-in-man studies, and further clinical studies and randomized trials are warranted [90, 91].

Ongoing and planned clinical trials

EXCEL trial

The debate on the optimal revascularization strategy for LMS disease is not yet over. A large dedicated LMS study, the Evaluation of Xience Prime Versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization (EXCEL) trial, with planned enrolment of 2600 patients to revascularization with either a newer-generation everolimus-eluting stent (Xience V/PRIME; Abbott Vascular Ltd.) or CABG is currently underway (NCT 01205776). The results are awaited with great interest.

BEST trial

The purpose of the BEST (Bypass Surgery Versus Everolimus-Eluting Stent Implantation for Multivessel Coronary Artery Disease; NCT 00997828) study is to determine that the safety and efficacy of PCI with everolimus-eluting stents are noninferior to CABG for the treatment of patient with multivessel CAD. It is planned that 1776 patients will be enrolled, and the primary end-point is the composite of all-cause death, nonfatal MI and ischaemia-driven TVR.


In the ISCHEMIA (International Study of Comparative Health Effectiveness With Medical and Invasive Approaches; NCT 01471522) trial, patients are randomly assigned, before coronary angiography, to a conservative OMT strategy or to an invasive strategy if they have documented myocardial ischaemia. The primary end-point is death or MI.


Revascularization for Ischaemic Ventricular Dysfunction (REVIVED-BCIS2; NCT 01920048) is a multicentre, randomized, open, controlled trial to evaluate the efficacy and safety of PCI compared to OMT alone for ischaemic left ventricular dysfunction. The trial has just started with the intention of recruiting 700 patients, and the primary end-point is all-cause death or hospitalization due to heart failure.

Clinical trials of bioresorbable scaffolds

The second-generation magnesium scaffold (DREAMS, Biotronik, Berlin, Deutschland) is being tested in the BIOSOLVE-II study. ReZolve2 (REVA Medical, San Diego, CA, USA), a desaminotyrosine polycarbonate scaffold, is being tested in the RESTORE-II study. ABSORB EXTEND is an international prospective, single-arm study that will recruit more than 800 patients with more complex coronary disease than previously studied in the ABSORB cohort A and B. ABSORB-II (NCT 01425281) is a prospective, randomized, controlled trial to compare the safety and efficacy of the Absorb BVS versus the Xience stent in 501 patients with stable angina and one- to two-vessel disease. The primary end-points are the superiority of Absorb BVS for vasomotion of the treated segment and noninferiority for angiographic minimum lumen diameter at 2 years. There are a number of other ongoing studies with Absorb BVS, including ABSORB Japan (NCT 01844284), ABSORB-STEMI (NCT 01986803) and ABSORB-III (NCT 01751906). The potential advantages of bioresorbable scaffolds need to be proven in adequately powered trials with long-term follow-up to establish the role of this device in routine practice [85].


The optimal treatment strategy for patients with complex CAD will continue to be debated in the coming years. Recent studies including SYNTAX and FREEDOM have helped to improve therapeutic decision-making and also provided tools to identify the appropriate treatment for individual patients. EXCEL will further inform treatment decisions. Generally, as the anatomical complexity increases, surgical revascularization appears to be more beneficial; however, comorbidities in certain subset of patients are likely to make PCI a more attractive and practical choice. It remains reasonable to suggest that an open dialogue between members of the heart team and the individual patient is the key to deciding the OMT. Both surgeons and interventional cardiologists should note that OMT and management of risk factors are essential for the best prognosis after either PCI or CABG.

Conflict of interest statement

No conflicts of interest to declare.