At two years following randomization, revascularization was associated with a lower mortality and a reduction in the composite endpoint of death, myocardial infarction, and recurrent hospitalization [20]. An important study that evaluated treatment of silent ischemia is the Swiss Interventional Study on Silent Ischemia Type II (SWISSI II) trial which compared medical therapy with balloon angioplasty among patients who had suffered a myocardial infarction, and had one or two vessel disease [21]. A surprising finding was that cardiac death and myocardial infarction were significantly lower in the group randomized to balloon angioplasty. While the findings of the ACIP and SWISSI II trials are significant, they need to be interpreted with the knowledge that both enrolled relatively small number of patients and that optimal medical therapy, as defined in the COURAGE trial, was not implemented.
The studies to date have had significant crossover to revascularization in those originally randomized to medical therapy and hence have been trials of “initial treatment strategies” rather than specific treatments. Thus, based on the evidence from the COURAGE trial and the preceding randomized clinical trials, it is reasonable to conclude that medical therapy is an appropriate initial strategy for a substantial proportion of patients with mild to moderately severe stable angina. PCI is suitable for those patients who are significantly symptomatic despite optimal medical therapy, or as initial strategy for those with a positive stress test at low workload, or have a large ischemic territory. Aggressive secondary prevention is essential regardless of the treatment strategy utilized. The findings of the Atorvastatin versus Revascularization Treatment (AVERT) trial showed that PCI in combination with inadequate lipid lowering therapy is associated with worse outcomes in patients with angina when compared with a strategy of optimal lipid management and medical therapy alone [11].
In BARI 2D, 2368 patients with type 2 diabetes mellitus and stable coronary artery disease (defined as either a ≥50% stenosis of a major epicardial artery with a positive stress test or ≥70% stenosis and classic angina) were randomized to either revascularization (CABG or PCI) within 4 weeks together with intensive medical therapy or to intensive medical therapy alone [22]. The decision regarding CABG versus PCI was based on clinical judgment, and made prior to randomization. At five years, there was no difference in the primary endpoints of the rates of survival (88.3% vs 87.8%) or freedom from the composite of death, myocardial infarction, and stroke (77.2% vs 77.7%). In the PCI stratum, there was no significant difference in primary endpoints between the revascularization group compared to the medical‐therapy only group. However, in the CABG stratum, the rate of major cardiovascular events was significantly lower in the revascularization group. Patients selected for CABG had higher angiographic and clinical risk scores than those selected for PCI, and it was those with the highest clinical and angiographic risk profile who seemed to derive a benefit from CABG.
In a meta‐analysis from 12 randomized clinical trials with 37 548 patient‐years of follow‐up demonstrated that PCI compared with medical therapy alone was associated with a statistically significant 24% relative reduction in the risk of spontaneous non‐procedural myocardial infraction (MI), at the cost of a 317% relative increase in the risk of procedural MI, with no overall difference in the risk of all MI. The point estimate for PCI versus medical therapy for the outcome of all‐cause mortality and cardiovascular mortality paralleled that of spontaneous non‐procedural MI (incident rate ratio = 0.70; 95% CI 0.44–1.09), but was not statistically significant [23]. Another meta‐analysis of 15 trials with 14 669 patients with stable CAD that included data from the ISCHEMIA trial once again confirmed that PCI does not reduce the rates of overall mortality (relative risk = 0.98;95% CI 0.87–1.1), cardiovascular mortality or myocardial infarction, compared to medical therapy [24]. In a network meta‐analysis from 100 trials in 93 553 patients with 262 090 patient‐years of follow‐up, new generation drug eluting stents (everolimus: rate ratio 0.75, 95% CI 0.59–0.96; zotarolimus (Resolute): 0.65, 95% CI 0.42–1.00) were associated with improved survival compared with medical treatment [25]. However, balloon angioplasty (0.85, 95% CI 0.68–1.04), bare metal stents (0.92, 95% CI 0.79–1.05), or early generation drug eluting stents (paclitaxel: 0.92, 95% CI 0.75–1.12; sirolimus: 0.91, 95% CI 0.75–1.10; zotarolimus (Endeavor): 0.88, 95% CI 0.69–1.10) were not associated with improved survival compared with medical treatment. The findings suggest that there may be improved survival with new (second) generation DES but not with other PCI technology, compared with medical treatment. These reports are provocative and challenge the general dogma that PCI has no impact on mortality. The findings are especially notable because the randomized trials have generally limited the enrolment to patients with single vessel disease and have excluded high risk patients with left main disease or chronic total occlusion for whom revascularization can offer greater benefit. Thus, until further data are available and strategies for risk stratification are improved, therapeutic decisions ought to be based on guidelines, but tailored according to a combined assessment of the patient’s clinical presentation, severity of ischemia, and coronary anatomy.
Comparison of percutaneous and surgical revascularization
The available evidence suggests that PCI and CABG are equivalent for the treatment of single vessel disease. This was specifically investigated in the MASS trial at a single center in which patients with significant (>80%) proximal LAD stenosis were randomized to balloon angioplasty, CABG, or medical therapy. The data demonstrated that there was similar relief of symptoms with both forms of revascularization. However, revascularization resulted in a lower incidence of inducible ischemia compared to medical therapy alone, and all three strategies resulted in the effective treatment of limiting angina [17]. Similar findings have been reported from another small study of 134 patients with isolated proximal LAD stenosis in which angioplasty and CABG produced comparable results [26] also when the follow‐up is prolonged to 10 years [27]. Importantly, the need for repeat revascularization during follow‐up was greater with percutaneous revascularization using balloon angioplasty in both trials.
In the assessment of trials comparing surgery with PCI, an important premise is that none of these studies have specifically addressed the functional significance of the lesions treated. With respect to PCI in multivessel disease, the Fractional Flow Reserve vs Angiography for Multivessel Evaluation (FAME) trial demonstrated that a targeted strategy guided by measurement of FFR provides superior outcomes at one year compared with treatment of all vessels with visually estimated significant stenoses [28]. Several randomized clinical trials [29–34] and meta‐analyses of the data [35] have compared PCI directly with CABG for the treatment of single and multivessel disease. The studies, predominantly from the pre‐stent era, were among relatively low risk patients with multivessel disease and preserved LV function. By design, the inclusion criteria for these trials had mandated that the coronary anatomy was suitable for both forms of revascularization, thereby excluding most patients with very complex coronary anatomy or chronic total occlusions. These older trials provided important lessons, but are not directly relevant to current practice because they predated stents and the widespread use of internal mammary artery graft in CABG.
The largest of these studies is the Bypass Angioplasty Revasculari‐zation Investigation (BARI) trial in which 1829 patients were enrolled. PCI was performed with balloon angioplasty which was associated with similar frequency of death, myocardial infarction, and recurrent angina when compared with CABG during 10 years of follow‐up in the overall study population [36]. Not surprisingly, repeat revascularization was significantly less common in patients randomized to surgery. One notable finding that has influenced current practice is that cardiac mortality was significantly lower (19.4% vs 34.5%; p = 0.003) amongst diabetics requiring glucose lowering therapy who were randomized to CABG and received at least one mammary graft. Potential explanations for this observation are that CABG offers complete revascularization and hence mitigates some of the adverse impact of the greater atherosclerotic burden and “future culprit” lesions, as well as the greater likelihood of restenosis and disease progression in diabetic patients. This finding has been confirmed in a more recent study [37] but the clinical relevance and applicability of the BARI study results have been questioned by the findings from the BARI registry, which highlighted the importance of clinical judgment. The patients in the registry had clinical characteristics that were similar to those in the randomized trial, but the survival among diabetic patients was similar regardless of whether angioplasty or CABG