Revascularization treatment in patients with coronary artery disease

Abstract

The choice of optimal revascularization strategy in patients with coronary artery disease (CAD) is becoming more challenging lately, due to recent advances in percutaneous coronary intervention (PCI) and coronary artery bypass grafting surgery (CABG). On one hand, drug–eluting stents (DES) have emerged as a solution to the problem of restenosis after balloon angioplasty or bare–metal stent implantation, which was responsible for a higher rate of events (mainly repeat revascularization) in relation to CABG. On the other hand, off–pump bypass techniques and minimally invasive grafting of the left anterior descending artery appear to be safe and efficacious alternatives to traditional, on–pump CABG. Available literature includes studies outdated by current technologies, leaving the dilemma of best revascularization strategy unanswered in the general CAD population, but also in high–risk groups, such as diabetics and patients with chronic kidney disease. A number of ongoing trials, especially designed for this purpose, are set to end the debate, providing headto– head comparisons between DES–assisted PCI and contemporary bypass surgery.

As coronary artery disease (CAD) patients increase, so does the need for optimal choice of revascularization strategy, in the setting of both chronic stable angina and acute myocardial infarction. Moreover, the two available modes of coronary revascularization, i.e. percutaneous coronary intervention (PCI) and coronary artery bypass grafting (CABG), demonstrate recent advances, the effectiveness of which, if confirmed in ongoing trials, will change the current landscape in the management of CAD. Of course, it cannot be stressed enough that not all patients with CAD have an indication for revascularization; recommendations for these procedures (implying a benefit over medical therapy alone, in terms of either hard clinical endpoints or symptom control) have been clearly described in recent guidelines¹ and in general include multivessel, left main (MVD and LMD, respectively) and proximal left anterior descending artery (LAD) disease, symptom refractoriness to optimal medication and high risk patient characteristics. Interestingly enough, only CABG -and not PCI- offers a clear survival benefit over medical therapy alone, and this is true in certain patient populations. This article will focus away from this comparison between medical treatment and revascularization procedures and revisit the question: "PCI or CABG?" in multivessel disease, with a special interest on patients with diabetes or chronic kidney disease. A comparison between stent types will be attempted, and advances in the field of bypass surgery discussed. Emphasis will be given to trials currently under way, the results of which may affect clinical practice and help shape future guidelines.

Multivessel disease

The historical randomized trials of the 1990s that produced the first results regarding the comparison between CABG with percutaneous transluminal coronary baloon angioplasty (PTCA) have long been outdated due to the advents in Interventional Cardiology, mainly the introduction of stents, bare metal (BMS) and drug eluting (DES) ones, as well as the evolution of surgical techniques, such as arterial graft harvesting, off-pump procedures and minimally invasive approaches. The Randomized Intervention Treatment of Angina trial (RITA-1) included 1011 patients, 45% with single-vessel disease and 55% with MVD, and was the first to demonstrate, both at one-² and five-year follow-up³, a non-significant difference in mortality, myocardial infarction (MI) and total health-care costs between balloon angioplasty and CABG; the latter, however, was associated with fewer anginal symptoms and reinterventions. Another "classic" study was the Coronary Angioplasty versus Bypass Revascularisation Investigation (CABRI), which randomized 1054 symptomatic MVD patients to CABG or PTCA; similar one-year mortality was recorded, with more reinterventions and clinically significant angina in the PTCA arm⁴. A more recent analysis of the CABRI data, however, showed that, after adjustment for baseline variables, including CAD angiographic scores, mortality was significantly greater in the PTCA group⁵. The last of the large-scale conventional angioplasty vs CABG trials was the Bypass Angioplasty Revascularization Investigation (BARI), which consisted of a randomized group (1829 MVD patients equally suitable for CABG and PTCA) and the BARI registry (the remaining 2010 of the patients initially enrolled)⁶. According to the 10- year results that were recently published⁷, survival, MI occurrence and angina were comparable in the two arms of the randomized group, while repeat revascularization was higher for PTCA patients. An earlier analysis of the BARI data also demonstrated similar 7-year survival for CABG and PTCA in high-risk anatomic subsets, in which survival is prolonged by CABG versus medical therapy. Finally, the analysis of the BARI registry data revealed a two-times greater preference to PTCA compared to CABG and equivalent 7-year mortality rates⁸. All in all and with the exception of the CABRI reanalysis, the randomized trials of the 1990s agreed on similar mortality and MI rates for balloon angioplasty and CABG, the latter requiring fewer reinterventions during follow-up.

A newer set of trials, with recruitment periods extending from the middle of the last decade to the beginning of the 2000s, attempted to evaluate the effectiveness of the newly applied, uncoated stents versus CABG. Since the inferiority of PTCA versus CABG had been attributed directly to restenosis and the subsequent need for reintervention in previous studies, the introduction of stents, aimed to protect from restenosis, should improve PCI outcomes. The Arterial Revascularization Therapies

Study (ARTS) randomized 1205 patients with MVD to stent implantation or bypass surgery, both modalities being able to revascularize the patient to the same extent. One–year results demonstrated that PCI with stenting had similar rates of death, MI and stroke with CABG, and a lower cost; it necessitated, though, repeat revascularization more often⁹. The Angina With Extremely Serious Operative Mortality Evaluation (AWESOME) trial¹⁰ initially enrolled 2431 patients, who were divided into a randomized group of 454 patients (suitability for PCI or CABG plus consent), a patient–choice registry of 327 (suitability for PCI or CABG without consent to randomization), and a physician–directed registry (the remaining 1650 patients, not equally suitable for PCI or CABG). All patients had MVD with symptoms refractory to medical treatment and high–risk features for outcome after CABG (one of following five: prior CABG, age greater than 70, left ventricular ejection fraction < 0.35, recent MI, and need for intraaortic balloon pump). Equivalent survival rates for up to 36 months were demonstrated for PCI, with or without stenting, and CABG. The ERACI II study (Argentine Randomized Study: Coronary Angioplasty with Stenting versus Coronary Bypass Surgery in patients with Multiple–Vessel Disease) randomized 225 patients to PCI with bare–metal stents (BMS) and 225 patients to CABG¹¹. Thirty–day and one–year analysis revealed increased survival and freedom from MI after PCI, however at 5–year follow–up CABG demonstrated similar survival and non–fatal MI rates with PCI, with fewer (65.3% vs. 76.4%) major cardiovascular events (MACE), less severe angina and need for repeat revascularization (71.5% vs. 92.4%)¹². A subgroup analysis of the ERACI II participants who additionally had proximal LAD stenosis arrived to the same conclusions with the five–year report¹³. The Stent Or Surgery trial (SoS) enrolled 988 patients in Europe and Canada and randomized them to stent–assisted PCI or CABG¹⁴. It differed from the previous studies in that repeat revascularization rather than death or MI was the primary outcome measure, with death or MI and all–cause–mortality as secondary endpoints. It showed that the need for reintervention was still higher for PCI versus CABG (21 vs 6%), but markedly lower than in the era of balloon angioplasty. However, it was the only study to demonstrate an increase in all–cause mortality for the PCI arm, while cardiac death or MI were comparable with CABG rates. This was attributed to increased cancer–related deaths in the stenting group. The medicine, angioplasty, or surgery study (MASS–II) was a trial of 611 patients with stable MVD, who were randomized to medical therapy alone, PCI or CABG¹⁵. Five–year follow–up results were recently published¹⁶, which show comparable overall mortality for the three study arms and statistically different primary endpoint rates (total mortality, Q–wave MI, angina requiring revascularization), with CABG being advantageous over medical treatment in the pairwise analysis, but not over PCI (26.2%, 36% and 32.7%, respectively); medical therapy was found non–inferior to the latter. This study, together with ARTS, ERACI II, and SoS, were included in the meta–analysis by Mercado et al¹⁷, where multiple stenting was no different from CABG (1513 versus 1533 patients respectively) in the primary endpoint (death, MI or stroke), but still appeared to necessitate repeat revascularization procedures relatively more frequently. This meta–analysis in a way sums up the results of all aforementioned randomized trials of BMS–assisted PCI versus CABG, conducted until the early 2000s.

All the above trial results are hardly generalizable in the modern era of restenosis–reducing DES and off–pump bypass surgery. Both PCI and CABG have evolved, therefore new studies had to be designed, in order to compare these novel modes of revascularization. An important ongoing trial is the SYNergy between percutaneous coronary intervention with TAXus and cardiac surgery (SYNTAX) study, which will enrol, across Europe and North America, 1500 patients with three–vessel disease and/or left main disease without traditional exclusion criteria (an "all–comers" trial)¹⁸. These patients will be randomized to receive Taxus (Boston Scientific) stents or contemporary bypass surgery, if an interventional cardiologist and a cardiac surgeon agree that coronary lesions can be equally revascularized by the two strategies. In case the patient is ineligible for PCI, he/she enters the CABG registry; if CABG is contraindicated, the patient is allocated to the PCI registry. Follow–up will extend to 5 years, but the primary endpoint will be one–year freedom from major adverse cardiac and cerebral events (MACCE). By relying on a randomization algorithm so close to everyday clinical decision–making, this trial hopes to compare revascularization strategies in a real–world setting. A runin–phase intended to project accrual rates demonstrated that CABG is the preferred method of revascularization (74%), and that Europeans tend to use PCI to a greater extent than Americans (29% vs 18%) - certainly a high rate of application for a strategy not supported by current guidelines¹⁹; CABG with totally arterial grafting represents 12% of the whole CABG procedures, while totally venous grafting is still used in 7% of patients. This study will inevitably include patients from various highrisk groups, too, for example diabetic patients (a formally predefined subgroup of the study), patients with left main disease or with suppressed left ventricular systolic function.

Revascularization in special patient subgroups Diabetes Mellitus

Diabetes Mellitus

Patients with Diabetes Mellitus (DM) constitute approximately 25% of patients undergoing coronary revascularization²⁰, and this percentage is bound to increase, given the current DM pandemic. Diabetes creates an especially atherogenic environment that encompasses impaired fibrinolysis, dyslipidemia, endothelial dysfunction and increased platelet aggregation, culminating in early and accelerated coronary lesions. Most importantly, due to more extensive and severe coronary atherosclerosis (more lesions in number and complexity), advanced systolic and diastolic dysfunction of the left ventricle and higher rates of silent ischemia and prior history of cardiovascular events, all revascularization–related outcomes in diabetic patients are severely compromised compared to non–diabetic groups. This occurs independently of revascularization mode: on one hand, CABG is confronted with bad quality of grafts and accelerated graft occlusion (especially venous), and with higher perioperative risks and subsequent mortality; on the other hand, PCI outcomes are adversely affected by markedly high restenosis rates in the era of either balloon angioplasty or uncoated stents. Despite the high incidence of CAD in diabetic patients and the high revascularization rates, no prospective, randomized trial, especially designed and powered for this purpose, has compared contemporary PCI and bypass surgery techniques. The only comparative information available is derived from earlier trials, namely the BARI diabetic subgroup, where both five–⁶ and tenyear⁷ survival was significantly reduced in the PCI group in relation to CABG (65.5 vs 80.6% and 45.5 vs 57.8% respectively). This has been for years considered to be the single most robust study result guiding clinical practice, however it demonstrates important limitations: a) CABG patients received more complete coronary revascularization than PCI patients (87 vs 76%); b) in the diabetic subgroup of the BARI registry⁸, no difference in 7–year mortality was found; and c) no stenting was available at the time of BARI, whereas optimal bypass technique, that is IMA grafting, was widely implemented in the bypass cohort (81% of patients). Another trial with diabetic representation was the AWESOME trial, where approximately 30% of patients in all 3 groups (randomized group, physician–directed registry, and patient–choice registry) suffered from DM and shared high–risk characteristics for bypass surgery²¹. Similar mortality up to 36 months was recorded in all groups between CABG– and PCI–treated patients. Finally, data regarding the diabetic cohort of the MASS–II trial, where medical therapy was compared to PCI and CABG, revealed a similar cumulative 5–year survival; medical treatment performed poorer beyond the first and up to the fifth year of follow–up²². All in all, a survival benefit has been demonstrated for CABG, though not consistently, in the few trials with data on diabetics available to date; no extrapolation to today's clinical practice can be attempted, of course.

The results of ongoing trials in the setting of CAD and DM are eagerly awaited. The SYNTAX trial has a predefined DM arm, as already mentioned, which will be further divided according to type of DM, mode of treatment and glycosylated haemoglobin level. The Future Revascularization Evaluation in Patients With Diabetes Mellitus: Optimal Management of Multivessel Disease (FREEDOM) study is a superiority trial aimed in comparing 5–year all–cause mortality rates in 2400 diabetic individuals with multivessel coronary artery disease (CAD) who undergo either CABG or PCI with serolimus–eluting stents (SES)²³. Another study especially designed for diabetic patients is The Coronary Artery Revascularisation in Diabetes (CARDia) trial, based in UK and Ireland, which will enrol around 600 patients with MVD or complex single–vessel disease and assign them to PCI with SES (plus aspirin, clopidogrel and abciximab) or bypass surgery with at least one arterial graft and off–pump technique, if available²⁴. One–year death, MI or stroke will be the primary endpoint. Finally, the Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) Trial has completed its randomization process, having recruited 2368 DM type 2 patients with at least one vessel amenable to revascularization and objective ischemia or typical angina. Five–year mortality will be assessed and two major comparisons made: the first between elective revascularization plus aggressive medical treatment versus aggressive medical treatment alone; and, the second, between an insulin–sensitizing and an insulin–providing strategy²⁵. These trials will probably end the debate regarding the optimal revascularization strategy in diabetics, an ever–growing group of patients with particularly adverse cardiovascular outcomes.

Chronic kidney disease

A patient population afflicted by even more frequent and severe coronary artery disease than diabetics (and equally increasing in numbers) are the patients with chronic kidney disease (CKD). Diabetes Mellitus, along with other traditional risk factors for CAD, clusters with specific for the disease conditions, such as subclinical inflammation, oxidative stress, malnutrition, and creates an extremely atherogenic milieu. It is well known that a creatinine clearance rate lower than 60 ml/min/1.73 m² is a marker of poor prognosis for a number of cardiovascular outcomes²⁶. Post-PCI outcomes are heavily affected by high restenosis rates, which have been significantly improved in the stent–era²⁷, but have never become equivalent to those of patients with normal renal function. Drugeluting stents have proved to be superior to Base Metal Stents (BMS) in terms of restenosis rates^28,29 and even cardiovascular morbidity and mortality³⁰ in patients with CKD. Contrast–induced nephropathy (CIN) can be a serious complication in the PCI setting, as pre–existing renal dysfunction is one of the best predictors of CIN development³¹; all precautions should be taken in this group of patients³², and staged procedures preferred³³. On the other hand, CKD is an adverse prognostic factor for every cardiac surgical procedure, including bypass surgery, after which outcomes are markedly worse than in patients with normal renal function³⁴. When comparing CABG to PCI in the setting of patients with CKD, randomized, prospective trials are lacking, due to a systematic exclusion of such patients from major trials in this field. A review of the literature³⁵ identified only retrospective trials and analyses of registry data; in general, bypass surgery superiority over balloon angioplasty was unequivocally demonstrated³⁶. However, uncoated stenting in a substudy of the ARTS trial proved equivalent to CABG in terms of death, MI, or stroke, and inferior only in reintervention rates³⁷. This has been the only, up to now, randomized, prospective trial in the CKD–CAD setting. Such prospective trials that will assess the relative effectiveness of contemporary bypass techniques with DES in CKD patients are urgently needed.

Left main coronary artery disease

This constitutes a challenging group of patients, since unprotected LMD, especially in cases of distal bifurcation, is long known to be associated with worse long–term outcome compared to CAD patients with other sites of atherosclerotic lesions in the coronary tree. In this setting, CABG has traditionally been the revascularization therapy of choice, as the risk of subacute thrombosis with conventional angioplasty or uncoated stenting is high³⁸. However, the introduction of DES has urged researchers to investigate whether LMD can now be treated with a safety profile equivalent to that of CABG and the additional advantage of low perioperative risk, important for patients with comorbidities and ineligibility for bypass surgery. Until now, only observational studies are available. Palmerini et al³⁹ retrospectively compared 154 LMD patients that underwent CABG to 157 patients treated with PCI (94 with DES and the rest with BMS). Although DES reduced by 25% the rate of all endpoints (death, MI, target lesion revascularization -TLR) compared to BMS, they still demonstrated a worse event–free survival than CABG, a result, however, driven by a greater rate of reinterventions and not of death or MI. Another study reported a single–center experience about the outcomes of DES implantation versus CABG in 50 and 123 LMD patients respectively. Although the PCI group had more high–risk patients than the CABG group (expressed as a Parsonnet score of over 15), it had a lower 30–day MACCE rate (2 vs 17%), while on multivariate analysis CABG was found to be an independent prognosticator of MACCE; 6–month and 1–year event–free survival was similar for the two strategies⁴⁰. The above data underline the need for a randomized comparison between DES and CABG in this setting. In the ongoing SYNTAX trial, as already mentioned, LMD is one of the inclusion criteria, whereas the COMparison of Bypass surgery and AngioplasTy using sirolimus–eluting stent in patients with unprotected left main coronary artery disease (COMBAT) trial⁴¹ will only include patients with LMD, approximately 1730, equally randomized to SES or CABG; other vessel lesions have to be amenable to both modes, and prior history of LM–stenting or CABG along with highrisk features (low ejection fraction and NYHA class III-IV) cause exclusion from the study. Two–year all–cause mortality, MI and stroke will be the primary endpoint, with a projected follow–up of five years. The completion of these trials will give valuable insights in the optimal strategy for LMD.

Percutaneous coronary intervention: Which stent to use?

Drug-eluting stents undoubtedly represent a turning point in the field of percutaneous coronary interventions, as they promise to overcome all disadvantages of baremetal stents, mainly restenosis, which kept PCI in the shadow of bypass surgery for years. It cannot be disputed that DES indeed offer lower restenosis rates than uncoated stents. Among the most representative trials in this setting are the following: a) The SIRIUS trial, which examined the occurrence of target vessel failure (that is the composite endpoint of cardiac death, MI, and target vessel revascularization -TVR) 270 days after PCI with serolimus- eluting stents versus bare-metal stents, and found SES to offer significantly more protection than BMS (primary endpoint rates 8.6 vs 21% respectively), also true for diabetic patients (respective rates 12.2 and 27%)⁴²; and b) the TAXUS IV trial, where a paclitaxel-eluting stent (PES) was tested against BMS and was found to offer lower rates of TVR at 9 months (4.7 vs 12% respectively for the whole study population and 11.3 vs 24% for the subgroup of diabetics)⁴³. Furthermore, SES have demonstrated superiority over BMS in the management of total coronary occlusions, where their use has been associated with less binary stenosis, target lesion- and vessel- revascularization, along with fewer MACE⁴⁴. Similar findings resulted from the comparison of SES with BMS when implanted in diseased saphenous vein grafts: while death and MI rates were equivalent, late lumen loss was greater by BMS, as was median neointimal volume, measured with intravascular ultrasound⁴⁵. However, doubt has recently been cast over the long-term safety of DES, as a result of potentially higher rates of thrombosis compared to BMS^46,47; this could compromise long-term outcomes of DES, closing the gap with BMS. This was the case with ERACI III⁴⁸, a study that added a DES arm to the ERACI II trial of BMS versus CABG (see section "Multivessel disease"). Although DES was superior to both CABG and BMS at one-year, this initial advantage decreased over time up to the 3-year follow-up; moreover, late stent thrombosis was more frequent with DES. This, however, was not a randomised trial, but a registry of DES-treated patients, who generally had more high-risk features than the historical two arms of ERACI II. The demonstrated risk of late stent thrombosis in ERACI III has not been consistently supported by other studies (Table 1, where various stent types are compared in the setting of acute MI). Another concern is the cost-effectiveness of DES: immediate post-procedural excess cost versus BMS ranged between 1600 and 3200 US dollars in a study⁴⁹, with a significant long-term reduction, dependent on the cost per avoided revascularization. All in all, DES have shown greater efficacy over BMS in restenosis prevention, therefore in the need for repeat revascularization. However, the risk of excess late thrombosis, although not confirmed, should continue to be addressed in studies with longer follow-up periods. If this issue is settled, it will be easier to assess the real cost-effectiveness of DES, which in the short term appears very low.

Table 1. Comparison of stent types in the setting of primary angioplasty for acute myocardial infarction with ST-elevation.

* denotes statistical significance at the <0.05 level. All numbers except sample size and hazard ratio (HR) express percentages %. MACE: major adverse cardiac events/ SES: serolimus-eluting stents/ BMS: bare-metal stents/ PES: paclitaxel-eluting stents/ TLR: target lesion revascularization/ TVR: target vessel revascularization/ TVF: target vessel failure/ FU: follow-up/ MI: myocardial infarction.

As for the relative efficacy of various types of DES, study results have shown SES to be superior to PES, in terms of either MACE rates in normoglycemic individuals -the SIRTAX trial–⁵⁰, or late lumen loss in diabetics -the ISAR–DIABETES trial⁵¹. A recent meta–analysis of 16 randomized trials of SES versus PES included 8695 patients and found SES to offer significantly lower rates of reintervention and stent thrombosis than PES; mortality was similar, while MI rates tended to be higher with PES⁵². Apart from sirolimus and paclitaxel, a novel antiproliferative agent, zotarolimus, has been applied to a phosphorylcholine polymer–based stent and compared head–to–head with: a) BMS, in the ENDEAVOR II trial⁵³, where it reduced rates of clinical and angiographic restenosis for up to 2 years, with similar occurrence of stent thrombosis; and b) SES, in the ENDEAVOR III trial, where it was found inferior to it in terms of late lumen loss, binary restenosis and TLR at 8 months⁵⁴. A randomized trial (PRISON III) comparing zotarolimus–eluting stents with SES in patients with total coronary occlusions is currently under way⁵⁵. It is obvious that in the DES era, more studies are needed to safely identify the most potent agent, with the best antiproliferative properties and the fewer adverse clinical outcomes.

Adequate clinical and angiographic data (Table 1) have been accumulated in the setting of stent-assisted PCI in patients suffering from acute myocardial infarction with ST-segment elevation (primary PCI). Two large-scale, randomized trials, the SESAMI⁵⁶ and the TYPHOON⁵⁷, both with follow-up angiographic data, demonstrated superiority of SES over BMS in terms of MACE and restenosis, while exhibiting no significant difference in stent thrombosis. These findings were not supported by a Spanish study, where only trends were revealed in favour of SES and a remarkably high rate of stent thrombosis; however, the number of participants was significantly lower than in the afore-mentioned studies⁶². As for PES, the PASSION trial showed only a trend towards PES superiority over BMS in MACE and TLR⁵⁸. It has to be pointed out that these studies report shortterm outcomes (maximum 1 year), and the possibility exists that cases of really late stent thrombosis might have been lost. The only long-term reports available come from two related dutch registries, the RESEARCH and the T-RESEARCH registries⁵⁹, where 3-year outcomes after SES, PES or BMS implantation are recorded: only a trend towards fewer events was found in favour of SES over BMS, with thrombosis rates being higher in the SES and PES arms, although not at a statistically significant level. All above data considered together, suggest that DES are quite safe and effective in treating coronary lesions in stable patients with CAD or in the setting of primary PCI for STEMI. Fine differences between agents used in stent coating and the issue of stent thrombosis remain to be elucidated by trials to come.

Advances in coronary artery bypass surgery

In parallel to the developments in percutaneous coronary interventions, bypass surgery has recently demonstrated its own advances that challenge the traditional on pump coronary bypass surgery (ONCAB) and strengthen the opposition to drug–eluting stents. Off–pump coronary bypass surgery (OPCAB) has been around for years, gaining ground over ONCAB in everyday practice. Head–tohead comparisons of the two techniques have been conducted, either with a prospective design or retrospectively. Several issues have been resolved, while others need to be further investigated. Graft patency following OPCAB has been shown in two meta–analyses of randomized trials to be compromised in relation to ONCAB, suggesting an increased risk of graft occlusion^63,64; revascularization was also found to be less complete with OPCAB⁶³. Regarding mortality rates, the off–pump technique appears at least non–inferior to conventional surgery: in a New York database study, OPCAB demonstrated lower in–hospital and similar 3–year mortality with ONCAB⁶⁵, confirming an older finding of better mortality rates in high–risk patients with multi–vessel disease⁶⁶. As for the need for reinterventions, it has been found to be higher after OPCAB65, but not unanimously⁶⁷. Quality of life, as perceived by patients after bypass surgery, seems to be higher in the OPCAB groups^67,68, while at the same time costs are kept lower than in conventional surgery⁶⁷. Perioperative risk is lower for OPCAB, with fewer blood transfusions, less myocardial necrosis, less neurocognitive deficit, and, perhaps most importantly, better preservation of renal function, either in patients with pre–existing diabetic nephropathy⁶⁹ or in dialysis patients⁷⁰. Conclusively, both surgical techniques seem to offer an excellent outcome to patients referred for surgical revascularization. Off–pump surgery has a more favourable perioperative profile and offers comparable survival with conventional CABG, but is also characterized by a lower graft patency rate, which is translated into a more frequent need for repeat revascularization.

Totally endoscopic coronary artery bypass (TECAB) of the left anterior descending artery (LAD) with thoracoscopic harvesting of the left internal mammary artery (LIMA) has been already implemented with success, with low mortality, perioperative morbidity and satisfactory LIMA patency^71,72. It was initially intended for isolated high–grade stenosis of proximal LAD. In this setting, a meta–analysis comparing TECAB to bare–metal stenting was performed, demonstrating similar mortality and MI rates, but less TVR with TECAB⁷³; however, the only trial with DES was excluded from the analysis. Based on the efficacy of this novel, minimally invasive surgical technique, integrated coronary revascularization has been proposed: it combines TECAB to LAD with DES implantation to the non–LAD lesions. Experience is growing, even in high–risk groups, such as diabetic patients⁷⁴, and sufficient data exist to justify a randomized comparison of this integrated approach with traditional CABG. Still, optimal timing of PCI in relation to TECAB has to be found and a higher than usual reintervention rate for lesions corrected with stenting needs to be explained⁷².

Conclusions

Revascularization procedures in CAD patients who are truly in need of them (as medical therapy alone is sometimes as effective) have evolved significantly, making comparisons more difficult than they were in the era of balloon angioplasty without stenting, when bypass surgery was undoubtedly superior to PTCA in all aspects. The introduction of bare–metal stents closed the gap with CABG, and the addition of DES strengthened the argument for PCI. Mortality and recurrent MI rates do not differ and the higher rate of reinterventions is the cost a patient has to pay for the fewer periprocedural events that PCI offers in relation to CABG. Drug–eluting stents are not superior to BMS at all times, and recommendations concerning their use have to be followed, since high health–care costs and longstanding antiplatelet treatment are drawbacks to their unrestricted use; as for stent thrombosis in DES–treated patients, it does not appear now as menacing as it did before. Ongoing and future randomized trials, rather than registries or retrospective studies, will elucidate unresolved issues both in the general CAD population and in high–risk groups, such as diabetic patients.