that the absolute CABG benefit is even higher. This CABG benefit was seen irrespective of symptom status and extended to asymptomatic patients. Note that the survival benefit in stable CAD does not emerge until 2 years after CABG, partly because of the early surgical hazard; thus, CABG is an appropriate therapy in patients who are otherwise likely to have a good longevity. CABG is expected to be beneficial sooner in patients with unstable CAD. Those trials were done in the 1970s, at a time when CABG technique was suboptimal (LIMA was not routinely used, which explains why the survival advantage of CABG gradually narrowed beyond 10 years). But in those trials medical therapy was also suboptimal (mainly consisting of β-blockers, with very limited use of aspirin and no statin). In fact, in the CABG stratum of the modern BARI 2D trial, initial CABG did not reduce the mortality of diabetic patients with multivessel, non-left main disease (vs. initial medical therapy). Same results were replicated in ISCHEMIA trial. Thus, the only absolute indication for CABG in the stable CAD setting is left main disease, not 3-vessel or proximal LAD disease.
The above trials excluded patients with severe LV dysfunction (EF <35%). In patients with severe ischemic LV dysfunction (EF <35%), no or mild angina, and no severe HF, CABG improved death and cardiovascular hospitalizations at 10 years in the STICH trial.69,70The benefit was, however, not dramatic (~20% reduction of cardiovascular death). Once again, the benefit did not emerge until 2 years after CABG. This benefit was irrespective of viability testing.
CABG may also be performed for single- or two-vessel CAD not involving the LAD, if PCI is not technically feasible and the patient has refractory, severe angina. The value of a single- or two-vessel CABG to a non-LAD vessel is mainly symptomatic.
The well-known benefit of CABG in diabetic patients is seen in CABG vs. PCI trials, rather than in the above trials of CABG vs. medical therapy.
VIII. PCI and PCI vs medical therapy
General PCI indications and pitfalls- Consider the following three settings:
PCI is a first-line consideration for one- or two-vessel CAD, along with severe angina.
PCI is an alternative to CABG in three-vessel CAD, or two-vessel CAD involving the proximal LAD, with an angiographic SYNTAX score ≤22 and no diabetes.
In three-vessel CAD, or complex two-vessel disease involving the LAD (especially proximal LAD), with either diabetes or SYNTAX score >22, CABG significantly improves survival in comparison with PCI (SYNTAX and FREEDOM trials).
PCI is flawed by a risk of restenosis of 20% with BMS, which is reduced to <10% with DES. At the 5-year follow-up of PCI populations without extensive CAD, 20-25% of patients had recurrent events, mainly recurrent symptoms with a need for revascularization. Approximately 50% of these events occur in the target vessel, mainly the target stented lesion, while 50% occur in remote vessels (non-culprit disease progression).67,71 Those events would be more common in patients with diffuse and complex CAD, who not only have a higher risk of restenosis but also a higher risk of progression of non-target lesions/vessels, as PCI only addresses the focal disease.
PCI vs medical therapy-The COURAGE trial evaluated patients with stable, rather mild angina, normal EF, and a good functional status (mean stress test exercise time = 7 min).72 It included a large population of patients with multivessel CAD (~70%, including 30% three-vessel CAD) and patients with proximal LAD disease (~35%). Initial PCI did not improve death or MI outcomes as compared to initial medical therapy ± delayed PCI (PCI was eventually performed in 31% of patients in the medical therapy arm). COURAGE results, however, do not dismiss the value of PCI in highly symptomatic patients with frequent anginal episodes (i.e., daily), those with severely limiting angina, or those with persistent angina despite medical therapy, patients in whom PCI has strong effects on quality of life and functional status, according to the COURAGE quality-of-life substudy.73
Similar results were replicated in the BARI 2D trial.74 In diabetic patients with mild angina, initial revascularization did not reduce death or MI in comparison to initial medical therapy, although 40% of patients eventually crossed over from medical therapy to revascularization over 5 years. More specifically, in patients with one- or two-vessel CAD, initial revascularization with PCI did not reduce death or MI in comparison with medical therapy (PCI stratum). In patients with extensive CAD randomized to initial CABG vs. medical therapy (CABG stratum), initial CABG reduced MI (14.6% vs. 7.4% at 5 years) but not death in comparison with medical therapy.
The FAME 2 trial randomized patients who mainly had one- or two-vessel CAD (65% proximal/mid-LAD) to FFR-guided PCI vs. medical therapy. FAME-2 trial differs from the COURAGE trial mainly in the use of FFR guidance. PCI strikingly reduced the risk of urgent revascularization at 8 months and 5 years (6% vs 21%), as well as all future revascularizations (13% vs 51%), more so in lesions with FFR≤0.65.66 Yet PCI failed to show any mortality or MI reduction even at 5 years.28 While PCI significantly reduced the risk of late spontaneous MI, which occurred in 10% of medically managed patients vs. 6.5% of PCI patients at 5 years, it was associated with a risk of periprocedural MI (2%) and stent thrombosis (1%) which, beside non-target lesion progression, negated the overall MI benefit. Moreover, late STEMI was not reduced by PCI and was similarly low in both arms (1.6% at 5 years). It was speculated that long-term mortality rates might diverge in favor of PCI, as spontaneous MI has a larger impact on long-term mortality than periprocedural MI,75 but this was not seen at 5 years (not even a trend).
The ISCHEMIA trial selected stable patients with moderate or severe ischemia on stress testing (≥10% of the LV). Those patients were subjected to coronary CTA to confirm CAD diagnosis (≥50% in a major epicardial vessel) and to rule out left main disease; ~5% of screened patients had left main disease and were excluded. Patients with CAD were randomized to medical therapy vs. cardiac catheterization followed by revascularization; patients in the medical therapy arm only received CTA at baseline, not cardiac catheterization.18 Unlike COURAGE and BARI 2D, this trial randomized patients before any catheterization was done. In a way, this study did not just question the value of revascularization, but also the value of invasive coronary angiography after a severely abnormal stress test. At baseline, in both groups, 45% of patients had 3-vessel CAD on CTA, 31% had 2-vessel CAD, 47% had proximal LAD disease, and 42% had diabetes. In the invasive arm, only 13% did not have significant CAD on catheterization, and thus most patients underwent revascularization, of whom 76% received PCI and 24% received CABG. At 4 years of follow-up, invasive strategy and revascularization did not affect mortality (~6.5% in both arms), MI, or the combined endpoint of MI/mortality (11.7% for invasive vs 13.9% for conservative). Invasive strategy increased MI by ~2% in the first 6 months (large periprocedural MI) and reduced it by 4% between 6 months and 5 years (spontaneous MI), with a nearly neutral overall effect. This lack of benefit was consistent across all subgroups, including 3-vessel CAD, proximal LAD, and diabetic subgroups; and was not affected by ischemia or anatomic severity.63 Of note, only EF>35% was included, and EF was mostly normal (median 60%). The biggest pitfall is that >95% of patients had CCS 0-II angina and most patients had infrequent angina, monthly or less.
In sum, ISCHEMIA trial shifts 2 paradigms in stable CAD:
1 Revascularization: in stable CAD, revascularization does not reduce mortality or MI even in extensive CAD and severe ischemia (the only exclusion being left main disease).
2 Diagnostic strategy: this trial questions the value of functional stress testing in the management of stable CAD and rather emphasizes the value of anatomical rule-out of left main disease via CTA. Even invasive coronary angiography is not routinely necessary after a severely abnormal stress test, which is a paradigm shift. CTA takes a more central role.
CKD patients in ISCHEMIA-Only patients with CKD did not undergo CTA (GFR 30-60 ml/min/1.73 m2 in main ISCHEMIA trial, and GFR<30 ml/min/1.73 m2 or dialysis in ISCHEMIA-CKD trial).19 CKD patients