The Current State of Coronary Revascularization: Percutaneous Coronary Intervention versus Coronary Artery Bypass Graft Surgery

Abstract

The question of percutaneous coronary intervention (PCI) versus coronary artery bypass graft (CABG) surgery remains among the most important questions in the treatment of coronary artery disease. The leading North American and European societies largely agree on the current guidelines for the revascularization of unprotected left-main disease (ULMD) and multivessel disease (MVD) which are largely supported by the outcomes of several large randomized trials including SYNTAX, PRECOMBAT, NOBLE, and EXCEL. While these trials are of the highest quality, currently available, they suffer several limitations, including the use of bare metal and/or first-generation drug-eluting stents in early trials and lack of updated surgical outcomes data. The objective of this review is to briefly discuss these key early trials, as well as explore contemporary studies, to provide insight on the current state of coronary revascularization. Evidence suggests that in ULMD and MVD, there are similar mortality rates for CABG and PCI but PCI is associated with fewer “early” strokes, whereas CABG is associated with fewer “late” strokes, myocardial infarctions, and lower need for repeat revascularization. Additionally, studies suggest that CABG remains superior to PCI in patients with intermediate/high SYNTAX scores and in MVD with concomitant proximal left anterior descending (pLAD) artery stenosis. Despite the preceding research and its basis for our current guidelines, there remains significant variation in care that has yet to be quantified. Emerging studies evaluating second-generation drug-eluting stents, specific lesion anatomy, and minimally invasive and hybrid approaches to CABG may lend itself to more individualized patient care.

Coronary artery disease (CAD) is among the leading causes of morbidity and mortality worldwide. The prevalence of CAD in the United States is estimated at 7.2% for ages 20 years and greater. Treatment of CAD and myocardial infarction (MI) are two of the top-10 most costly diseases in the U.S. health care system, combining to consume an estimated $21.1 billion annually. ¹ While invasive treatment of symptomatic CAD via percutaneous coronary intervention (PCI) or coronary artery bypass graft (CABG) surgery has generally been considered superior to optimal medical management, this notion has recently been challenged by the well-designed ISCHEMIA trial published in 2020 which further begs the question “which patients benefit from PCI, CABG, or a combination of both modalities.” ^{2 3 4 5 6} This question is especially relevant given the overall burden on the health care system, and more importantly, in terms of what the revascularization options mean for any given patient with respect to short- and long-term outcomes. The most common framework for this discussion involves the location of coronary artery lesions with a focus on unprotected left-main CAD (ULMD), left-main equivalent disease (LMED), proximal left anterior descending (pLAD) disease, and multivessel disease (MVD). This anatomic framework is used purposefully, given the notable differences on the impact of myocardial territories affected by lesions in different locations. This is also the framework used within societal guidelines pertaining to coronary revascularization. The objective of this review is to discuss the core literature used to formulate the current leading societal guidelines with respect to PCI versus CABG for symptomatic CAD, as well as explore contemporary literature, which may impact future guidelines for coronary revascularization. Considerations for special populations, including those with diabetes, chronic kidney disease (CKD), and octogenarians, will also be discussed.

Advances in Percutaneous Coronary Intervention and Surgical Conduit Selection

Advances in Percutaneous Coronary Intervention

From 1986 when Dr. Jacques Puel successfully implanted the first coronary artery stent until to 2002 when first-generation drug-eluting stents (DES1) were introduced, only bare-metal stents (BMS) were available. ⁷ The advent of second-generation DES (DES2) in the 2010's found them to be superior to BMS and DES1, with lower rates of restenosis and in-stent thrombosis. ^{8 9 10} The type of stent used in each study during the following discussion is worth noting due to differences in performance and how that may affect interpretation of the studies. Other advances in PCI that may not be entirely reflected in the literature include the use of intracoronary imaging with intravascular ultrasound (IVUS), optical coherence tomography, and near-infrared spectroscopy; all of which allow real-time tomographic imaging and may permit more optimal stent placement and improve outcomes related to PCI. ¹¹ More current widespread use of IVUS may impact the relevance of findings from studies in which expertise in IVUS was not available.

Conduit Selection for Coronary Artery Bypass Graft

The left internal mammary artery (LIMA) is considered the gold standard for revascularization, given its longevity, lower incidence of adverse effects, and improved patient survival. Saphenous vein grafts (SVG) are the most frequently utilized conduits after the LIMA. Easy access and the ability to generate several conduits from a single saphenous vein make this a desirable option. However, studies indicate a high rate of graft failure, with estimates of 25% graft failure between 12 and 18 months. ¹² Multiple arterial conduits including the radial artery have been utilized as alternatives. Results have found the radial artery to be superior to SVGs for both rate of adverse cardiac events, as well as patency at 5 years, although arterial graft spasm and flow to the extremity must be considered. ¹³ The right internal mammary artery (RIMA) is another high-quality conduit. Using bilateral internal mammary arteries (BIMA) for CABG was recently evaluated in the Arterial Revascularization Trial (ART). In the ART trial, more than 3,000 patients with MVD were randomized to LIMA plus SVG or BIMA. Investigators found that rates of all-cause mortality, MI, and stroke were similar at 10 years. However, patients in the BIMA arm had a higher rate of sternal wound complications (3.5 vs. 1.9%). ¹⁴ It is worth noting that only 84% of patients who were randomized to BIMA actually received BIMA grafts. However, in the per-protocol analysis which only included patients who actually received BIMA, no difference in all-cause mortality, MI, and stroke was noted. With the many options as additional conduits for CABG, consideration should be given to which conduits were used in specific studies comparing CABG to PCI. This is analogous to analyzing the type of stent used in PCI treatment.

Unprotected Left-Main and Left-Main Equivalent Coronary Artery Disease

Percutaneous Coronary Intervention with Bare-Metal and First-Generation Drug-Eluting Stents versus Coronary Artery Bypass Graft

ULMD is defined as ≥50% stenosis of the left main coronary artery (LMCA) proximal to the bifurcation of the LAD and left circumflex arteries without a previous coronary artery bypass to a region of myocardium supplied by the left main coronary artery. Current guidelines for ULMD are based on a series of studies from the 2000s to late 2010s. A thorough understanding of these studies will clarify societal recommendations, as well as certain caveats ( Table 1 ).

Table 1. Studies evaluating PCI versus CABG in unprotected left main coronary artery disease.

Study (year)	Design	Stent type	PCI ( n )	CABG ( n )	Endpoints	Pertinent Findings
LE MANS (2008)	Randomized control of patients with ULMD	BMS	52	53	•Change in LVEF 12 months after intervention •MACCE at 1 year	•Significant increase in LVEF at the 12-month follow-up in the PCI group (3.3 ± 6.7% after PCI versus 0.5 ± 0.8% after CABG; p  = 0.047) •MACCE at 1-year (PCI 20 vs. CABG 21%)
MAIN COMPARE (2008)	Registry-based matched cohort of patients with ULMD	BMS and first-generation DES	1,102	1,130	•Composite of death, Q-wave MI, stroke, TVR	•No difference in composite endpoint at 3 years •CABG showed lower mortality and serious composite outcome rates after 5 years •PCI was associated with higher risks of death (HR: 1.35; 95% CI: 1.00–1.81) and the composite outcome (HR: 1.46; 95% CI: 1.10–1.94) at 5 years
SYNTAX (2009)	Randomized control of patients with MVD and/or ULMD	First-generation DES	903	897	•Primary: death from any MACCE at 12 months •Secondary: death, stroke, MI, repeat revascularization, graft occlusion or stent thrombosis	•Primary endpoint: (PCI vs. CABG) 17.8 vs. 12.4% (RR: 1.44; 95% CI: 1.15–1.81; p  = 0.002; NNT = 19) •Secondary endpoint (PCI vs. CABG): Stroke: 0.6 vs. 2.2% ( p  = 0.003), Repeat Revascularization: 13.5 vs. 5.9% (RR: 2.29; 95% CI: 1.67–3.14; p  < 0.001) MACCE (CABG vs. PCI) results from subgrouping: •All LM: 13.7 vs. 15.8 •LM only ( n  = 91): 8.5 vs. 7.1 •LM + 1VD ( n  = 138): 13.2 vs. 7.5 •LM + 2VD ( n  = 218): 14.4 vs. 19.8 (statistically significant) •LM + 3VD ( n  = 258): 15.4 vs. 19.3 (statistically significant)
PRECOMBAT (2011)	Randomized control of patients with ULMD	First-generation DES	300	300	•MACCE at 1- and 2-year follow-up •MACCE at 5-year follow-up	•1 year: cumulative event rate, 8.7 vs. 6.7%; absolute risk difference, 2.0 percentage points; 95% CI: −1.6 to 5.6; p  = 0.01 for noninferiority) •2 years: cumulative event rate, 12.2 vs. 8.1%; HR with PCI, 1.50; 95% CI: 0.90–2.52; p  = 0.12 •5 years: cumulative event rates of 17.5 and 14.3%, respectively; HR: 1.27; 95% CI: 0.84–1.90; p  = 0.26 • Revascularization occurred more frequently in the PCI group than in the CABG group (11.4 and 5.5%, respectively; HR: 2.11; 95% CI: 1.16–3.84; p  = 0.012)
Boudriot et al (2011) 24	Randomized control of patients with ULMD	First-generation DES	100	101	•Freedom from MACCE at 12 months	•Primary end point was reached in 13.9% of patients after surgery, as opposed to 19.0% after PCI ( p  = 0.19 for noninferiority) •Combined rates for death and myocardial infarction were comparable •Stenting was inferior to surgery for repeat revascularization (5.9 vs. 14.0%; noninferiority p  = 0.35).
NOBLE (2016)	Randomized control for patients with ULMD	Second-generation DES	598	603	•MACCE at 5 years	•5 years estimates of MACCE were 28% for PCI (121 events) and 18% for CABG (80 events), HR: 1·51 (95% CI: 1·13–2·00), exceeding the limit for noninferiority, and CABG was significantly better than PCI ( p  = 0·0044) Comparing PCI with CABG, 5-year estimates were: •11 vs. 9% (1·08, 0·67–1·74, p  = 0·84) for all-cause mortality •6 vs. 2% (2·87, 1·40–5·89, p  = 0·0040) for nonprocedural myocardial infarction •15 vs. 10% (1·50, 1·04–2·17, p  = 0·0304) for any revascularization •5 vs. 2% (2·20, 0·91–5·36, p  = 0·08) for stroke •Actual 5-year follow-up affirmed the estimates
EXCEL (2016)	Randomized control of patients with ULMD	Second-generation DES	948	957	•MACCE at 3 years	•MACCE occurred in 15.4% of the patients in the PCI group and in 14.7% of the patients in the CABG group (difference, 0.7 percentage points; upper 97.5% confidence limit, 4.0 percentage points; p  = 0.02 for noninferiority; HR: 1.00; 95% CI: 0.79 to 1.26; p  = 0.98 for superiority) 5-year follow-up HR (PCI versus CABG) for the primary outcome varied in the three periods between: •0–30 days (HR: 0.61; 95% CI: 0.42–0.88) •30 days–1 year (HR: 1.07; 95% CI: 0.68–1.70) •1–5 years (HR: 1.61; 95% CI: 1.23–2.12) •Found statistical significant for LM + 2 (25% PCI vs. 17.8% CABG, OR: 1.55, CI: 1.06–2.36)
Giacoppo et al 29 (2017)	Meta-analysis of the SYNTAX, PRECOMBAT, NOBLE, and EXCEL trials	First-generation DES, and second-generation DES	2,203 first-generation DES: 657 second-generation DES: 1,546	2,208	•Composite of all-cause death, MI, and stroke	•No difference for primary endpoint by fixed effect (HR: 1.06; 95% CI: 0.90–1.24; p  = 0.48) and random effects (HR: 1.06; 95% CI: 0.85–1.32; p  = 0.60) in low-to-intermediate SYNTAX scores •PCI was associated with higher risk for PCI (HR: 1.70; 95% CI: 1.42–2.05; p  < 0.001) •No difference in rate of revascularization between first- and second-generation DES
Ahmad et al 31 (2020)	Meta-analysis of the SYNTAX, PRECOMBAT, NOBLE, EXCEL trials, and Boudriot et al	First-generation DES, second-generation DES (no subgroup analysis by stent generation)	2,303	2,308	• Primary: all-cause mortality • Secondary: cardiac death, MI, stroke, TVR	•No difference in all-cause mortality between PCI and CABG (RR: 1.08, 95% CI: 0.89–1.31, p  = 0.444) •No difference in long-term risk of cardiac death (RR: 1.08, 95% CI: 0.84–1.38, p  = 0.561) •No significant difference in the long-term risk of all MI (RR: 1.21, 95% CI: 0.95–1.55, p  = 0.114)  •PCI was associated with higher risk of unplanned revascularization (RR: 1.76, 95% CI: 1.52–2.04, p  < 0.001)
Al-Abcha et al 32 (2021)	Meta-analysis of long-term follow-up of the SYNTAX, PRECOMBAT, NOBLE, and EXCEL trials	First-generation DES, second-generation DES (no subgroup analysis by stent generation)	2,203	2,208	• Primary: all-cause mortality • Secondary: cardiac death, MI, stroke, and TVR	•Median weight follow-up period was 6.5 years •No difference in all-cause mortality (RR: 1.10; 95% CI: 0.92–1.31; p  = 0.28) •No difference in risk of cardiac death (RR: 1.08, 95% CI: 0.84–1.38; p  = 0.56) •No difference in total MI (RR: 1.22, 95% CI: 0.96–1.56; p  = 0.11) •No difference in stroke (RR: 0.85, 95% CI: 0.46–1.57; p  = 0.60). •PCI with DES was associated with higher risk of repeat revascularization (RR: 1.75, 95% CI: 1.50–2.03; p  < 0.00001)  •PCI with DES was associated with higher risk of nonperiprocedural MI (RR: 2.13, 95% CI: 1.53–2.97; p  < 0.00001)
Ullah et al 33 (2020)	Meta-analysis of 6 RCT's and 37 observational trials	BMS, first-generation DES, second-generation DES	13,709	15,478	• MACCE at 30 days, 5- and 10-year follow-up	• MACCE (OR: 0.56; 95% CI: 0.42–0.76; p  = 0.0002) and all-cause mortality (OR: 0.52; 95% CI: 0.30–0.91; p  = 0.02) favored PCI at 30 days • CABG was favored at 5 years due to a significantly lower rate of MACCE (OR: 1.67; 95% CI: 1.18–2.36; p  ≤ 0.04), MI (OR: 1.67; 95% CI: 1.35–2.06; p  ≤ 0.00001), and revascularization (OR: 2.80; 95% CI: 2.18–3.60; p  ≤ 0.00001) • Higher rate of MACCE with PCI regardless of SYNTAX score and stent generation

Abbreviations: BMS, bare-metal stents; CABG, coronary artery bypass graft; CI, confidence interval; DES, drug-eluting stents; HR, hazard ratio; LVEF, left ventricular ejection fraction; MACCE, major adverse cardiovascular and cerebrovascular events; MI, myocardial infraction; MVD, multivessel disease; NNT, number needed to treat; OR, odds ratio; PCI, percutaneous coronary intervention; RCT, randomized controlled trial; RR, risk ratio; TVR, target vessel revascularization; ULMD, unprotected left main disease.

The Left Main Stenting in Comparison with Surgical Revascularization (LE MANS) and the MAIN COMPARE trials published in 2008 are the first modern trials examining this topic. The LE MANS Trial compared CABG to PCI with BMS making the findings relatively obsolete, given that BMS are generally not used in today's practice. ^{15 16} The MAIN COMPARE trial compared CABG with patients who underwent PCI with DES1. Investigators found that there was no difference in mortality or composite outcome (death, Q-wave MI, stroke, or TVR target vessel revascularization) at 3 years, though TVR was significantly higher in the PCI group. ¹⁷ At 5-year follow-up, DES1 was associated with higher risk of death and composite outcome but this difference disappeared at long-term follow-up. ¹⁸

The SYNTAX Trial in 2009 by Serruys and colleagues is perhaps the single, most important and well-cited dataset exploring this topic. Researchers randomized 1,800 patients with three-vessel disease and/or ULMD to either PCI with DES1 or CABG. Primary endpoint of MACCE from any cause and at 12 months was reached in 13.7% of patients who underwent CABG versus 15.8% who underwent PCI. ¹⁹ The rate of stroke was higher in the CABG group (2.7 vs. 0.3%) while PCI was associated with a higher rate of TVR (11.8 vs. 6.5%). Subgroup analysis based on lesion distribution and SYNTAX score can be found in Table 1 . At 5-year follow-up, rates of MI and the combination of death, stroke, MI, and repeat revascularization were higher in the PCI group. ²⁰ At 10-year follow-up, there was no survival difference in patients with LMD. ²¹

The PRECOMBAT study in 2011 randomized 600 patients with ULMD disease to PCI with DES1 or CABG. At 1-year follow-up, there was noninferiority of PCI compared with CABG for MACCE (8.7 vs. 6.7). ²² At 2 ears, MACCE occurred in 12.2% of PCI patients compared with 8.1% of CABG patients. ²² The 5-year follow-up found no difference in death from any cause, MI, or stroke. However, similarly to the previously discussed studies, TVR occurred more frequently in the PCI group (11.4 and 5.5%). ²³ It is worth noting that the margin for noninferiority was wide which does raise questions regarding the interpretation of this study.

The last of the commonly cited core studies involving DES1 was reported in 2011 by Boudriot and colleagues. In this study, 201 patients with ULMD disease were randomized to DES1 or CABG. Contrary to the PRECOMBAT trial, Boudriot and colleagues found PCI to be inferior to CABG, as the primary endpoint of death, MI, or TVR was reached in 19% of patients with DES1 versus 13.9% who underwent CABG. ²⁴ This outcome was largely driven by the rate of TVR.

Percutaneous Coronary Intervention with Second-Generation Drug-Eluting Stents versus Coronary Artery Bypass Graft

DES2 are the most commonly used stents in practice today. To evaluate DES2 versus CABG, the “Nordic-Baltic-British Left Main Revascularization (NOBLE)” trial and the “Everolimus-eluting stents or bypass surgery for left main CAD (EXCEL)” trial were performed.

The NOBLE trial, published in 2016 by Mäkikallio and colleagues that randomized 1,201 patients with ULMD to either DES2 or CABG. Investigator reported 5-year MACCE for PCI were 28 versus 18% for CABG, which did not qualify as noninferiority for PCI compared with CABG. The rate of MI and repeat revascularization were lower in the CABG group, although all-cause mortality was similar for both groups. ²⁵ The 5-year follow-up confirmed CABG to be superior to PCI in the composite endpoint, nonprocedural MI, and TVR. ²⁶

The EXCEL trial evaluated DES2 ( n  = 948) versus CABG ( n  = 957) in patients with ULMD of low-to-intermediate complexity (SYNTAX < 32). There was no difference in hazard ratio (HR) for the primary outcome of death, stroke, or MI for the period 30 days to 1 year. ²⁷ The rate of death, stroke, and MI at 3 years was 15.4% in the PCI group compared with 14.7% in the CABG group which met criteria for noninferiority. ²⁸ The 5-year follow-up found the HR for the primary outcome varied across different time periods, with a HR of 0.61 for the period 0 to 30 days in favor of PCI but an HR of 1.61 for the period of 1 to 5 years in favor of CABG. Subgroup analysis found that CABG was superior to PCI with respect to death, stroke, and MI in left main disease plus two vessels (PCI 25 vs. CABG 17.8%). ²⁷

The findings of the NOBLE and EXCEL trials largely correlate with the findings in trials comparing PCI with BMS and DES1 versus CABG. While these studies are sufficiently powered, we must wait for the 10-year follow-up.

High-Quality Meta-analyses That Include SYNTAX, PRECOMBAT, NOBLE, and EXCEL Trails

Independent trials are inherently limited by sample size, population character, regional differences in practice, and other factors. Several high-quality meta-analyses have been conducted that include not only the core studies discussed above but other lesser known trials as well.

Giacoppo and colleagues produced a meta-analysis in 2017 including only the SYNTAX, PRECOMBAT, NOBLE, and EXCEL trials, totaling 4,394 patients. Investigators found similar rates of all-cause death, MI, and stroke for PCI versus CABG in low-to-intermediate SYNTAX scores. ²⁹ Rate of revascularization was found to be higher in the PCI group. ²⁹ There no was difference in the rate of revascularization between first- and second-generation DES.

The results described by Giacoppo and colleagues were echoed by another 2017 meta-analysis by Palmerini et al. In the latter meta-analysis, six randomized control trials (RCTs) totaling 4,686 patients found no difference in risk of death, stroke, or MI at 30 days and at the longest follow-up. However, the meta-analysis did find a higher rate of revascularization in the PCI group. ³⁰ One difference reported by Palmerini and colleagues was a lower mortality associated with PCI in low SYNTAX scores but a lower mortality for CABG in patients with high SYNTAX scores. ³⁰

In 2020, Ahmad and colleagues found no significant difference in all-cause mortality, stroke, or MI, but more repeat revascularization in the PCI group. ³¹ Al-Abcha et al evaluated RCTs that used DES and had at least 5 years of follow-up. These investigators reported no difference in all-cause mortality, risk of cardiac death, MI, or stoke. ³² A higher rate of revascularization and non-periprocedural MI were noted in the PCI arms. ³² The largest meta-analysis by Ullah et al included six RCT's and 37 observational trials, totaling 29,187 patients. The 30-day rate of MACCE and all-cause mortality were lower in the PCI groups. However, CABG was favored at 5 years due to lower rate of MACCE, MI, and TVR. ³³ Authors also reported that PCI was associated with higher rates of MACCE at 5 years regardless of SYNTAX score or the generation of stent used. ³³

Registry Data and Current Trends of Practice

Data extrapolated from registries does not carry the same statistical weight as randomized trials. However, registry data may represent the current trends of practice better than RCT's.

A 2020 study based on the MAIN COMPARE Registry evaluated the long-term (mean follow-up: 12 years) outcomes after PCI or CABG for ULMD based on the location of the lesion within the LMCA. Investigators found higher risk of death, MI, or CVA (cerebrovascular accident), for patients who underwent PCI for distal bifurcation disease but noted this divergence emerged primarily after 5 years. No difference was noted in patients with ostial or shaft LMD but PCI was associated with significantly higher rates of TVR. ³⁴

A publication based on the “Multi-Vessel Coronary Artery Disease (MULTICAD)” Israeli Registry evaluated 256 patients with ULMD who underwent CABG or PCI. Investigators report that after 3 years, patients who underwent PCI had a significantly higher rate of all-cause mortality (26.9 vs. 8.7%; p  < 0.001). ³⁵ Of note, only 27% of patients underwent CABG. Another publication exploring the factors for referral to PCI or CABG within the MULTICAD registry noted higher SYNXTAX score as an independent factor associated with referral to CABG, whereas older age and renal insufficiency were associated with referral for PCI. ³⁶ Additionally, only half of patients with type-2 diabetes mellitus (DM2) plus MVD and/or ULMD underwent CABG despite societal guidelines recommending CABG in this population. ³⁶ This finding is in part echoed by a publication of the American-based National Cardiovascular Data Registry which reported the rate of PCI for ULMD disease has been increasing over time. ³⁷

Unfortunately, there is not sufficient evidence to determine whether CABG or PCI are being chosen for therapy along proper guideline parameters versus patient/physician preference. A concern for inappropriate use of PCI was noted in a 2012 publication which found that location of therapy within New York State dramatically affected the likelihood of receiving PCI for the same pattern of CAD. ³⁸ Additionally, a finding published in the European Journal of Cardiothoracic Surgery found that 78% of patients erroneously believed that PCI would extend life expectancy and 71% erroneously believed PCI would prevent MI. Perhaps even more concerning, 68% of PCI patients and 59% of CABG patients were not aware of alternative revascularization strategies to treat their CAD. ³⁹ Further studies are necessary to evaluate the appropriate use of these therapies.

Left-Main Equivalent Disease

LMED is defined as pLAD and left circumflex arterial lesions of at least 70% stenosis, effectively covering the same myocardial territory as the LMCA. One of the first descriptions of LMED was made by Babb and colleagues in 1980. In this publication, 70 patients with LMED were compared with 75 patients with full ULMD. Symptoms did not differ between the two groups nor did hemodynamic indices such as ejection fraction, cardiac index, or left-ventricular end-diastolic pressure. Surprisingly, the authors noted that patients with LMED did not carry the same morbidity or mortality rates seen in ULMD. Instead, the authors concluded that the sense of urgency caused by the phrase “left-main equivalent” was inappropriate, as patients with LMED behaved more like two- or three-vessel disease rather than ULMD. ⁴⁰ In 1984, Califf and colleagues described the nonoperative prognosis of 55 patients with LMED disease compared with the nonoperative prognosis of patients with ULMD. Patients with LMED disease were noted to have better survival at 1-year (78 vs. 65%) and at 5 years (55 vs. 40%) when compared with ULMD. ⁴¹

The largest study that specifically evaluated LMED disease comes from the Coronary Artery Surgery Study (CASS) registry long-term experience published in 1995. This study compared 630 patients with LMED who underwent CABG with 282 who were managed medically. The median survival for the CABG group was 13.1 versus 6.2 years for medical therapy; however, survival was not different for patients with a normal left ventricular ejection fraction. ⁴² Unfortunately, a robust search effort found a lack of recent studies, specifically evaluating this patient population using current state-of-the-art technologies. CABG for LMED disease was a class-I recommendation in the American College of Cardiology and American Heart Association guidelines as recently as the early 2000s. ⁴³ However, LMED is not explicitly mentioned in the most recent consensus guidelines. Instead, it is considered as two- or three-vessel CAD depending on the coronary dominance. ⁴⁴

Societal Guidelines for the Treatment of Unprotected Left-Main Coronary Artery Disease

A detailed review of studies reveals some findings that are unambiguous while other findings are less consistent. Findings consistently showed that PCI is associated with higher risk of short-, mid-, and long-term repeat revascularization and both periprocedural and nonperiprocedural MI. The complexity of the lesion is also important for the clinician to consider. Studies are in agreement that lesions with high SYNTAX complexity (>32) benefit from CABG. For low and intermediate SYNTAX scores, the leading recommendation is for CABG, though both PCI and CABG may be appropriate, thus an individualized patient-care approach should be utilized. ^{44 45} PCI can be considered in patients with low anatomic complexity, low SYNTAX scores, and ostial or trunk lesions. ⁴⁴

Revascularization of Multivessel Coronary Artery Disease

MVD is defined as the presence of ≥50% stenosis of two or more epicardial coronary arteries. Additionally, lesions of the pLAD, defined as stenosis occurring proximal to any septal perforators or diagonal branches, are often present in MVD. Guidelines often differentiate whether or not there is a concomitant pLAD lesion. The remainder of this section will discuss the pertinent literature comparing PCI versus CABG in patients with MVD with and without concomitant pLAD lesions. Given the propensity for trials to avoid a subgroup analysis, specifically based on the presence of pLAD, the following discussion will be divided into trials with and without pLAD subgrouping, as well as a brief discussion comparing different CABG techniques in the setting of MVD ( Table 2 ).

Table 2. Select studies comparing PCI to CABG in multivessel coronary artery disease.

Study (year)	Design	Stent type	PCI ( n )	CABG ( n )	Endpoints	Pertinent findings
FREEDOM (2012)	Randomized control of diabetic patients with MVD	First-generation DES	953	947	•Death from any cause, MI, and stroke at a minimum of 2 years Late outcomes (median: 6 years)	•Composite PCI vs. CABG: 26.6 vs. 18.7% ( p  = 0.005) •Myocardial infarction: 13.9 vs. 6.0% ( p  < 0.001) •Stroke: 2.4 vs. 5.2% ( p  = 0.03) Late Outcomes: •CABG was associated with better survival and fewer nonfatal MI outcomes compared with PCI
ASCERT (2012)	Registry-based propensity matched retrospective of patients with MVD	BMS and first-generation DES	103,549	86,244	•All-cause mortality and 1 and 4 years	•At 1 year, there was no significant difference in adjusted mortality between the groups (6.24% in the CABG group as compared with 6.55% in the PCI group; RR: 0.95; 95% CI: 0.90–1.00) •At 4 years, there was lower mortality with CABG than with PCI (16.4 vs. 20.8%; RR: 0.79; 95% CI: 0.76–0.82).
BEST (2015)	Randomized noninferiority	Second-generation DES	438	442	•Composite of death, MI, and TVR at 2 years after randomization	•At 2 years: 11.0% (PCI) vs. 7.9% (CABG; absolute risk difference, 3.1 percentage points; 95% CI, −0.8 to 6.9; p  = 0.32 for noninferiority). •At longer term follow-up (median: 4.6 years): the primary end point had occurred in 15.3% (PCI) vs. 10.6% (CABG; HR: 1.47; 95% CI: 1.01–2.13; p  = 0.04).
Garp et al (2011)	Observational 5-year follow-up of the ARTS-I and ARTS-II study	BMS and first-generation DES	473 BMS: 184 First-generation DES: 289	209	•MACCE at 5-year follow-up	•MACCE occurred in 33.7% (BMS), 18.0% (CABG) and 24.9% (first-generation DES; BMS vs. first-generation DES p  = 0.04, CABG vs. first-generation DES p  = 0.07) •TVR was lower in CABG regardless of adjustment
Cavalcante et al 47 (2017)	Pooled analysis of SYNTAX and BEST Trials	first- and second-generation DES	577	589	•Composite of all-cause death, MI, and stroke at 5-year follow-up	• Primary endpoint occurred more frequently in PCI arm (16.3%) vs. CABG (11.5%; HR: 1.43; 95% CI: 1.05–1.95; p  = 0.026
Papadopoulos et al 52 (2017)	Prospective cohort study	Second-generation DES	70	70	•All-cause mortality, MI, TVR, and angina at 1-year follow-up	•For PCI vs. CABG, there was no difference in mortality for (5.7 vs. 11.4%, p  = 0.135), MI (0 vs. 4.3%), repeat revascularization (4.3 vs. 8.6%, p  = 0.115) and angina (10 vs. 18.6%, p  = 0.153) •PCI met criteria for noninferiority
Bangalore et al 53 (2015)	Propensity-matched observational registry study of PCI with second-generation DES or CABG	Second-generation DES	9,223	9,223	•Primary outcome: all-cause mortality •Secondary outcomes: rates of myocardial infarction, stroke, and repeat revascularization	•Similar risk of death (3.1% per year vs. 2.9% per year, respectively, HR: 1.04; 95% CI: 0.93–1.17; p  = 0.50) •PCI associated with higher risks of MI (1.9% per year vs. 1.1% per year; HR: 1.51; 95% CI: 1.29–1.77; p  < 0.001) and repeat revascularization (7.2% per year vs. 3.1% per year; HR: 2.35; 95% CI: 2.14–2.58; p  < 0.001) •PCI associated with lower risk of stroke (0.7% per year vs. 1.0% per year; HR: 0.62; 95% CI: 0.50–0.76; p  < 0.001)

Abbreviations: BMS, bare-metal stents; CABG, coronary artery bypass graft; CI, confidence interval; DES, drug-eluting stents; HR, hazard ratio; MACCE, major adverse cardiovascular and cerebrovascular events; MI, myocardial infraction; MVD, multivessel disease; OR, odds ratio; PCI, percutaneous coronary intervention; RR, risk ratio; TVR, target vessel revascularization.

Multivessel Coronary Artery Disease with Concomitant Proximal Left Anterior Descending Artery Lesions

There are several key studies examining MVD with pLAD. The ARTS study group published 5-year follow-up data from the ARTS-I and ARTS-II studies of 682 patients with MVD involving the pLAD. At 5 years, MACCE occurred in 33.7, 24.9, and 18.0% of patients treated with BMS, DES, and CABG, respectively, with statistical significance favoring CABG. ⁴⁶ This 5-year superiority of CABG over PCI was echoed in a 2017 publication in which Cavalcante and colleagues pooled data from the SYNTAX and BEST trials. Specifically, the primary composite of all-cause death, MI, or stroke at 5 years was reached in 16.3% of PCI patients versus 11.5 patients who underwent CABG. ⁴⁷

Multivessel Coronary Artery Disease without Concomitant Proximal Left Anterior Descending Artery Lesion Subgrouping

The SYNTAX trial included a subgroup analysis of patients with MVD without concomitant ULMD. Of the 549 patients in the CABG arm, 11.5% experienced MACCE compared with 19.2% in the PCI arm at 12 months. ¹⁹ At the 5-year follow-up, there was no difference in MACCE in patients who had low SYNTAX scores, but CABG was favored in patients with intermediate/high SYNTAX scores. ²⁰ The 10-year follow-up validated the findings of the 5-year follow-up based on SYNTAX scores. ²¹

The ASCERT (ACCF and STS Database Collaboration on the Comparative Effectiveness of Revascularization Strategies) Trial published in 2012 used data from the ACCF National Cardiovascular Data Registry and the Society of Thoracic Surgeons Adult Cardiac Surgery Database for the years 2004 to 2008 to compare CABG versus PCI in patients with MVD. ⁴⁸ This trial included 86,244 patients who underwent CABG and 103,549 patients who underwent PCI. There was no difference in mortality at 1-year, but at 4 years, CABG was associated with a lower mortality (16.4 vs. 20.8). ⁴⁸

The BEST Trial of 2015 was an RCT that compared CABG versus PCI with DES2. At a median follow-up of 4.6 years; the primary outcome of the composite of death, MI, and TVR was higher in PCI versus CABG. ⁴⁹ Authors reported that CABG resulted in more complete revascularization. ⁴⁹

The CREDO-Kyoto PCI/CABG Registry evaluated 945 patients who underwent PCI with DES1 and 1,248 patients who underwent CABG for MVD including the LAD. At 5-year follow-up, composite of all-cause mortality, MI, and stroke was not different. However, the rates of TVR, MI, and revascularization were higher in the PCI arm. ⁵⁰

A 2014 meta-analysis included six RCT's examining CABG versus PCI in MVD. PCI was associated with a higher rate of TVR and lower rate of stroke at 1 year. However, at 5-year follow-up, PCI was associated with a higher rate of death and MI. ⁵¹ Unfortunately, quality meta-analyses including patients with MVD without ULMD are scarce.

Percutaneous Coronary Intervention with Second-Generation Drug-Eluting Stents versus Coronary Artery Bypass Graft for Multivessel Disease

A 2017 prospective cohort study of 140 patients found similar rates of mortality, MI, TVR, and angina which satisfied criteria for noninferiority. ⁵² Bangalore and colleagues performed a registry-based retrospective study comparing patients who underwent PCI with DES2 or CABG for MVD. Propensity-matched arms included 9,223 each. Authors reported that PCI was associated with a higher risk of MI and TVR but a lower risk of stroke per year. No difference in risk of death per year was reported. ⁵³ A 2020 retrospective study of 81 patients who underwent CABG were compared with 132 who underwent PCI with DES2. Propensity matching of 46 pairs found that MACCE occurred more frequently in the PCI arm but this was largely driven by TVR. ⁵⁴ Data evaluating PCI with DES2 versus CABG for MVD is limited in volume and quality. Additionally, there are no studies with subgroup analysis powered to evaluate MVD with and without concomitant pLAD lesions.

Off-Pump Coronary Artery Bypass, Robot-Assisted Coronary Artery Bypass Graft, and Hybrid Revascularization for Multivessel Disease

Alternative approaches to CABG have been explored aimed at minimizing the sources of complications associated with conventional CABG: the full median sternotomy, aortic cross-clamping, exposure to cardiopulmonary bypass, and vein graft failure. A 2010 meta-analysis by Edelman et al consisting of 10 studies totaling 4,821 evaluated patients who underwent with PCI or off-pump coronary artery bypass (OPCAB). Authors found that rate of MACCE and TVR was lower in the OPCAB arm at 12 months. ⁵⁵ A meta-analysis of 16 RCTs in 2015 by Fan and colleagues evaluated hybrid coronary revascularization with OPCAB or PCI. There was no difference in short-term mortality, MACCE, or TVR between hybrid revascularization and PCI. OPCAB was found to have lower rates of MACCE compared with hybrid revascularization at both 1- and 3-year follow-up. ⁵⁶ A propensity-matched analysis of studies comparing hybrid revascularization with conventional CABG with LIMA or BIMA found no difference in mortality at 30-day or mid-term follow-up. Hybrid revascularization was associated with shorter hospital stays and lower rates of in-hospital complications. ⁵⁷ A 2019 retrospective analysis evaluated robot-assisted CABG ( n  = 281) versus PCI ( n  = 357). Authors found similar long-term survival, but robot-assisted CABG was associated with lower rate of TVR. Long-term mortality was predicted by age, left ventricular ejection fraction (LVEF), and CKD. ⁵⁸

Societal Guidelines for the Treatment of Multivessel Coronary Artery Disease

For three-vessel disease, CABG is recommended over PCI. For two-vessel disease without pLAD, CABG and PCI are both reasonable options with CABG favored in the setting of extensive ischemia. ^{44 45} CABG is generally favored over PCI for two-vessel disease with pLAD involvement, although the EACTS guidelines suggest that CABG and PCI are more equivocal in this population. ^{44 45}

Revascularization of Isolated Proximal Left Anterior Descending Artery Lesions

The pLAD provides the main blood supply to the anterior ventricular septum and the majority of the anterior wall of the left ventricle making the significance of lesions in this vessel obvious. Societal guidelines have a special focus on isolated pLAD lesions, making a thorough review of revascularization strategies for this specific lesion mandatory. Table 3 contains more detailed information of selected studies discussed in the following section.

Table 3. Select studies examining PCI versus CABG in isolated proximal anterior descending coronary artery disease.

Study (year)	Design	Stent type	PCI ( n )	CABG ( n )	Endpoints	Pertinent findings
Kinnaird et al 60 (2016)	Meta-analysis of three RCTs and eight observational studies comparing PCI and CABG for isolated pLAD lesions	First-generation DES	2,237	2,793	• Mortality and MACCE	•No difference in mortality between CABG (5.2%) and PCI (4.7%; RR: 1.23; 95% CI: 0.90–1.69) •For MACCE, PCI was associated with significant increase in adverse events (RR: 1.41; 95% CI: 1.03–1.93) •There were no significant differences in the risk of MI (RR: 0.86; 95% CI: 0.58–1.26) or stroke (RR: 2.36; 95% CI: 0.54–10.43) •Increased risk of TVR events in the PCI group (10.7%) vs. CABG group (5.2%) (RR: 2.52; 95% CI: 1.69–3.77)
Matsoukis et al 61 (2020)	Cohort study comparing conventional CABG versus PCI with second-generation DES	Second-generation DES	631	379	• Composite and components of MACE (cardiac death, MI, TVR)	•No difference between the two groups at follow-up (mean:4.6 ± 2.5 years) for MACEs (HR: 1.45, 95% CI: 0.92–2.28, p  = 0.11; HR:1.43, 95% CI: 0.91–2.26, p  = 0.13), cardiac death (HR: 0.97, 95% CI: 0.46–2.05, p  = 0.93; HR: 0.79, 95% CI: 0.36–1.72, p  = 0.56), or MI (HR: 1.43, 95% CI: 0.49–4.13, p  = 0.51; HR: 1.57, 95% CI: 0.53–4.64, p  = 0.42) •PCI had a borderline significantly greater risk of repeat revascularization (HR: 1.99, 95% CI: 1.00–3.94, p  = 0.05; HR: 1.95, 95% CI: 0.98–3.9, p  = 0.06).
Karskamp et al (2014)	Meta-analysis of studies comparing MIDCAB and PCI with DES	First-generation DES	463	478	• Cardiac death, MI, stroke, TVR	• Incidence of cardiac mortality and MI was similar between MIDCAB and DES (OR: 1.05; 95% CI: 0.44–2.47; and OR: 0.83; 95% CI: 0.43–1.58, respectively) • TVR was significantly lower in MIDCAB (OR: 0.16; 95%CI: 0.08–0.30; p  < 0.0001; number needed to treat, 13) • Similar incidence of periprocedural death (OR: 0.85; 95% CI: 0.21–3.47; p  = 0.82), MI (OR: 0.98; 95% CI: 0.38–2.58; p  = 0.97), and stroke (OR: 1.36; 95% CI: 0.28–6.70; p  = 0.70) was noted
Hannan et al 65 (2021)	Retrospective observational study based on New York's cardiac registries compared PCI with DES to conventional CABG and MIDCAB	First- and second-generation DES	13,115	1001 (MIDCAB = 211)	• Mortality composite of mortality, MI, stroke, TVR	• No significant differences in mortality or in mortality, MI, or stroke after 7 years • Conventional CABG surgery was associated with a lower subsequent revascularization rate than PCI (adjusted HR: 0.45; 95% CI: 0.35–0.58) and MIDCAB (adjusted HR: 0.46; 95% CI: 0.32–0.66).
Indja et al 66 (2020)	Bayesian network meta-analysis including 37 studies comparing DES with OPCAB and MIDCAB in isolated pLAD	First- and second-generation DES	First-generation DES: 17,558 Second-generation DES: 7,223	MIDCAB: 2,308 OPCAB: 4,739	• Early and late mortality, MI, stroke, revascularization	• OPCAB had fewer late MI compared with those who had first-generation DES (OR: 0.38, 95% CI: 0.20–0.72) and MIDCAB (OR: 0.41, 95% CI: 0.17–0.97) • OPCAB had less late TVR compared with first-generation DES (OR: 0.17, 95% CI: 0.09–0.32) and second-generation DES (OR: 0.32, 95% CI: 0.14–0.72) • Rate of late MACE was lower with OPCAB compared with that with first-generation DES (OR: 0.33, 95% CI: 0.26–0.43) and second-generation DES (OR: 0.62, 95% CI: 0.45–0.90) • Rate of late MACE with MIDCAB was lower than that with first-generation DES (OR: 0.43, 95% CI: 0.31–0.62) as was that with first-generation DES compared with first-generation DES (OR: 0.53, 95% CI: 0.39–0.70)

Abbreviations: CABG, coronary artery bypass graft; CI, confidence interval; DES, drug-eluting stents; HR, hazard ratio; MACCE, major adverse cardiovascular and cerebrovascular events; MACE, major adverse cardiovascular events; MI, myocardial infraction; MIDCAB, minimally invasive direct coronary artery bypass MVD, multivessel disease; OPCAB, off-pump coronary artery bypass; OR, odds ratio; PCI, percutaneous coronary intervention; pLAD, proximal left anterior descending; RCT, randomized controlled trial; RR, risk ratio; TVR, target vessel revascularization.

Percutaneous Coronary Intervention versus Conventional Coronary Artery Bypass Graft for Isolated Proximal Left Anterior Descending Lesions

A 2008 meta-analysis evaluated nine RCTs totaling 1,210 patients who underwent with PCI or conventional CABG for isolated pLAD lesions. Investigators found no difference in mortality, procedural stokes, or MI at 30 days, 1 year, or 5-years. Repeat revascularization was more common in the PCI group. Patients in the CABG arm experienced higher rates of relief of angina. ⁵⁹ This meta-analysis limited in the DES was only used in two studies and balloon angioplasty without stenting was used as the primary PCI which is currently not the standard of care. A 2016 meta-analysis in the American Journal of Cardiology by Kinnaird and colleagues evaluated three RCTs and eight cohort studies, totaling 5,044 patients. This analysis demonstrated no difference in mortality, risk of MI, or stroke, but PCI was found to have increased rate of revascularization. ⁶⁰ More recently, a 2020 publication by Matsoukis and colleagues evaluated PCI with DES2 ( n  = 631) versus CABG ( n  = 379) in patients with pLAD lesions. Investigators found no significant difference in MACE major adverse cardiovascular events at the mean follow-up of 4.6 years ( ± 2.5 years). PCI was associated with a greater rate of repeat revascularization. ⁶¹

Percutaneous Coronary Intervention versus Off-Pump Coronary Artery Bypass or Minimally invasive direct coronary artery bypass for Isolated Proximal Left Anterior Descending Lesions

Minimally invasive direct coronary artery bypass (MIDCAB) is an alternative to conventional CABG which spares the patient a full sternotomy. MIDCAB can be performed on- or off-pump.

In 2009, Thiele and colleagues compared PCI with DES against MIDCAB in patients with isolated pLAD lesions. Authors found PCI to be noninferior to MIDCAB for death and MI. However, the rate of TVR was noted to be worse in the PCI arm. ⁶² A 2015 study by Blazek and colleagues examined the 7-year follow-up data of patients with isolated pLAD lesions who underwent either PCI with DES1 or MIDCAB. The composite endpoint of death, MI, or TVR occurred in 22% of DES1 patients and only 12% of CABG patients. ⁶³ Authors also noted TVR for PCI occurred in 20% of patients compared with 1.5% of patients in the CABG arm. ⁶³ These findings were echoed by a 2014 meta-analysis compared MIDCAB with PCI with DES in a total of 941 patients. ⁶⁴

A 2020 registry study based on New York's cardiac registry evaluated the outcomes of patients who underwent PCI, conventional CABG, or MIDCAB. A total of 14,327 patients were included, with 13,115 patients in the PCI arm, 211 undergoing MIDCAB, and 1,001 undergoing conventional CABG. Researchers found that conventional CABG was associated with a lower rate of revascularization as compared with both PCI and MIDCAB; however, there was no difference in mortality, MI, or stroke. ⁶⁵ Mean follow-up time was reported to be approximately 4 years.

A 2020 meta-analysis compared DES1 and DES2 with OPCAB and MIDCAB in patients with either ULMD or pLAD disease. This large meta-analysis included 37 publications amounting to more than 31,000 patients. ⁶⁶ Researchers found no difference in 30-day mortality rates, strokes, or MI. OPCAB was associated with lower rates of MACCE when compared with both generations of DES. OPCAB was also associated with fewer late MI's when compared with DES1 and MIDCAB. ⁶⁶ Both OPCAB and MIDCAB were associated with lower rates of TVR when compared with PCI. There was no difference in MACCE when MIDCAB was compared with DES2. Authors concluded that surgical approaches to isolated ULMD and pLAD lesions are superior to both generations of DES with respect to MI, TVR, and long-term survival. ⁶⁶

Societal Guidelines for the Treatment of Isolated Proximal Left Anterior Descending Lesions

Studies suggest that there is minimal difference in mortality, MI, or stroke for PCI versus CABG, particularly when PCI is performed with DES2. Surgical revascularization is consistently associated with lower rates of repeat revascularization. The EACTS considers both CABG and PCI as 1a recommendations for isolated pLAD, whereas the ACC/AATS/AHA/ASE/ASNC/SCAI/SCCT/STS guidelines make CABG a 2a recommendation versus 2b for PCI. ^{44 45}

Special Populations

Patients with Diabetes Mellitus Type 2

Patients with DM2 are known to be at increased risk for MACCE. Several publications have focused specifically on patients with DM2 undergoing PCI or CABG. The BARI-2D trial of 2009 examined CABG and PCI separately versus optimal, noninvasive medical therapy (OMT). Freedom from MACCE was superior for CABG versus OMT (77.5 vs. 69%) but not for PCI when compared with OMT (77 vs. 78.5%). ⁶⁷ In this study, CABG was never directly compared with PCI. It is worth noting that the OMT groups had a notable difference in performance in terms of freedom from MACCE.

The FREEDOM trial in 2012 by Farkouh and colleagues compared CABG and PCI in patients with DM2. In total, 953 patients under PCI and 947 patients underwent CABG. At 5 years, CABG was favored in all-cause death and non fatal MI, but rate of stroke was higher in the CABG arm. MVD with concomitant LAD disease favored CABG. ⁶⁸

A meta-analysis of five RCTs totaling more than 3,000 patients with diabetes found CABG to be associated with a lower risk of MACCE at 3 years as compared with PCI. In this meta-analysis, PCI was associated with a lower risk of all-cause mortality than CABG. ⁶⁹ Currently, the EACTS recommends CABG for patients with three-vessel disease and DM2. No distinction is made with respect to two-vessel disease or any other pattern of CAD with or without DM2. ⁴⁵ The most recent ACC/AATS/AHA/ASE/ASNC/SCAI/SCCT/STS guidelines make no distinction between patients with or without DM2. ⁴⁴

Patients with Chronic Kidney Disease

CKD poses unique clinical challenges for both surgical and percutaneous intervention. A subgroup analysis of the EXCEL trial included 361 patients with CKD. At 3 years, event-free survival was less for patients with CKD compared with those without CKD for both PCI and CABG. PCI was associated with less acute kidney injury superimposed on CKD than CABG. ⁷⁰ Chan and colleagues performed a propensity-matched population-based study including 1,786 patients with CKD undergoing PCI with DES or CABG. CABG was associated with improved survival and greater 1-, 2-, and 3-year freedom from MACCE. Additionally, PCI was associated with greater hazard for late mortality and MACCE. ⁷¹ No specific guidelines are currently in place for patients with CKD. The EACTS and ACC/AATS/AHA/ASE/ASNC/SCAI/SCCT/STS recommend evaluation of these patients with a “Heart Team” approach. ^{44 45}

Octogenarians

Elderly patients represent a population worthy of extra considerations due to both comorbidities, as well as surgical risk and life expectancy. Šerpytis and colleagues examined this population in a 2018 publication including 411 patients who underwent either PCI or CABG. Authors found no difference in mortality at a median follow-up of 3 years. Authors also noted that chronic heart failure, atrial fibrillation, cardiogenic shock, and LMCA stenosis were independent predictors of mortality. ⁷² A study by Nicolini and colleagues found that PCI was an independent predictor of death at long-term follow-up. Additionally, authors found that CABG was most clearly reduced the risk of death in octogenarians, aged 80 to 85 years, had a previous MI, history of heart failure or CKD, or three-vessel disease. ⁷³ These findings were echoed in a 2016 meta-analysis of seven studies, totaling 1,879 patients who underwent CABG and 1,432 who underwent PCI. Authors reported a lower short-term mortality for the PCI arm but longer overall survival in the CABG arms. ⁷⁴

Conclusion

PCI versus CABG remains one of the most important questions in the management of symptomatic CAD. Numerous variables, such as the number and location of coronary lesions, the complexity of the coronary anatomy, and comorbidities, have been shown to impact outcomes. When comparing CABG and PCI findings are generally as follows: (1) PCI has fewer “early” strokes; (2) CABG has fewer “late” strokes; (3) long-term outcomes, including myocardial infarctions, and lower rates of revascularization favor CABG; and (4) CABG is superior to PCI in patients with intermediate or high SYNTAX scores. Current guidelines from leading societies are largely in consensus on recommendations and are based on the aforementioned findings. However, many of the studies from which these recommendations are derived are approaching or surpassing 10 years since publication and may not reflect the most current technologies and techniques for either method of revascularization. Additionally, it remains irrefutable that the greater “convenience” of PCI is attractive, especially as short-term outcomes are relatively similar and repeat PCI procedures are feasible. What remains to be seen is the consequences of repeat intervention and the overall durability and outcomes of such practices. This is likely contributing to variation in real-life practice trends. Furthermore, so called “high risk PCI” remains an alternative to CABG in those patients deemed to be of prohibitive surgical risk; procedures that may not be accounted for in the current literature and available data which may skew our understanding of the available subject matter. As it stands, there continues to be a need for large, properly subgrouped RCT's utilizing state-of-the-art techniques, as well as data to better explore practice trends, from region to region. The answer to the question of PCI versus CABG is less likely to be solved than it is to evolve over time.