Most patients with significant LMCA disease are symptomatic as a result of compromised blood supply to a large area of the myocardium. However, significant LMCA disease can be identified incidentally in stable patients undergoing coronary angiography. In the absence of significant LMCA stenosis or relevant clinical symptoms, the hemodynamic significance of incidental or intermediate LMCA lesions warrants further evaluation. Current clinical practice guidelines rely on angiographic lesion severity as the sole determinant of risk and for the de facto threshold for revascularization decision-making. However, this approach may be outdated in the current era of clinically proven noninvasive and invasive modalities and revascularization based solely on the angiographic appearance of LMCA stenosis of intermediate severity (50–70%) is not appropriate.11) In addition to the unwarranted surgical risk, premature CABG for potentially noncritical lesions may ultimately prove harmful to patients due to low graft patency rates and an accelerated rate of obstruction of bypassed native coronary vessels that may be up to 6-fold higher and which makes subsequent PCI technically challenging if required for symptom relief.12) In the contemporary clinical setting, more detailed evaluation of the anatomic severity and hemodynamic significance of clinically ambiguous LMCA lesions can be obtained by intracoronary imaging with intravascular ultrasound (IVUS), or physiologically using pressure wire assessment of the fractional flow reserve (FFR).

IVUS characterizes the vessel size and the distribution of the plaques within the LMCA and its daughter branches, enabling accurate minimal lumen area (MLA) measurements at a cross-sectional level. Evaluation of IVUS in a prospective study showed that a MLA ≥6 mm² assessed by this technique is a safe value for deferring revascularization of the LMCA.13) Another study proposed a smaller MLA of ≤4.5 mm² as a useful surrogate of functional significance in patients with isolated ostial and shaft intermediate LMCA stenosis.14) Ideally, a dual pullback from the left anterior descending artery (LAD) and left circumflex artery (LCX) should be used to fully characterize the LMCA bifurcation and avoid overestimation of, for example, the MLA at the LCX ostium from a LAD pullback.

Physiological guidance by means of FFR or instantaneous wave-free ratio (iFR) may be helpful for the evaluation of intermediate or ambiguous LMCA lesions.15) In this setting, the visual-functional mismatch between coronary angiography and FFR can be as high as 30–40%,11)16) and deferring LMCA PCI based on FFR values >0.75 or 0.80 has been shown to be safe.16)17) Recently, deferring revascularization of LMCA disease with an iFR >0.89 has also shown favorable outcomes.18)19) However, specific outcome studies evaluating iFR in LMCA disease are required to determine whether iFR could be widely adopted as the sole determinant of revascularization in patients with intermediate LMCA lesions.

LMCA PCI conducted by experienced operators (i.e., those who have performed at least 15 LMCA PCIs per year for at least 3 consecutive years) is associated with better short- and long-term outcomes than LMCA PCI performed by less experienced operators.34) PCI of the LMCA ostium or shaft is a straightforward procedure that is associated with a lower requirement for repeat revascularization than PCI of the distal LMCA bifurcation.35) The ostium of the LMCA lacks the tunica adventitia and is richer in smooth muscle cells and elastic tissue than any other portion of the LMCA and its branches, which requires attention to ensure that stent expansion is adequate.36)

For LMCA bifurcation disease, many challenging technical issues remain. Generally, provisional stenting has been advocated as the preferred approach in bifurcation lesions, as it is technically simpler with at least similar outcomes to a systematic 2-stent strategy.37)38)39) In practice, the single-stent crossover technique has been used more frequently, in as many as 60% of all LMCA bifurcation treatments.40) However, this was recently challenged by two RCTs that compared the double-kissing (DK) crush technique with the culotte and provisional stenting strategies for the treatment of true distal LMCA bifurcations.41)42)43) In both studies, the DK crush technique significantly reduced the primary composite ischemic endpoint.

Selection of a single- or 2-stent technique should be based on disease involvement of the LCX ostium, as side-branch compromise after crossover stenting is frequent. Therefore, to determine the choice of a single- or 2-stent strategy, IVUS provides accurate information for both main- and side-branch disease status in LMCA bifurcation lesions. After main-stent crossover from the proximal LAD to the LM, geometric changes in the LCX ostium were related mainly to carina shift, reduction of MLA, and increased eccentricity of the external elastic membrane and carina angle between the LAD and the LCX.44) In cases in which the LCX ostium is significantly compromised (>50%) after provisional stenting, FFR measurement should be considered first before further treatment of the LCX.

IVUS is also useful for procedural optimization of PCI of the distal LMCA bifurcation. After stent implantation, IVUS guidance ensures adequate expansion at the level of the ostial LAD, the ostial LCX, the polygon of confluence (e.g., the convergence zone of the LMCA, LAD, and LCX), and the distal LMCA.45) The best IVUS-MSA (minimal stent area) criteria that predicted angiographic in-stent restenosis were 5.0 mm² for the LCX ostium, 6.3 mm² for the LAD ostium, 7.2 mm² for the POC, and 8.2 mm² for the distal LMCA. The 2-year major adverse coronary event-free survival rate was significantly lower in patients with underexpansion of at least one segment vs. lesions with no underexpansion (90±3% vs. 98±1%, p<0.001); post-stenting underexpansion was an independent predictor for major adverse cardiac events.

Key clinical trials comparing PCI and CABG from the first-generation DES era to the second-generation DES era are summarized in Table 1. The SYNTAX study was a key pivotal trial; in the LMCA subgroup, no significant differences were seen in the 5-year rate of major adverse cardiac and cerebrovascular events (MACCEs), mortality, or MI between PCI and CABG.4) However, the 5-year rate of repeat revascularization was higher after PCI and the rate of stroke was higher after CABG. The first LMCA-specific RCT (RECOMBAT) showed that the 5-year rate of MACCEs, death, MI, or stroke was similar between PCI and CABG, but the rate of target-vessel revascularization was significantly higher after PCI.6)46) These trials prompted the initiation of two additional large-scale RCTs, EXCEL and NOBLE, which involved the use of contemporary DES.9)10) In the EXCEL study, the primary composite endpoint of death, stroke, or MI at 3 years was similar between PCI and CABG (p value for non-inferiority=0.02; p value for superiority=0.98).9) PCI was associated with a lower incidence of major periprocedural adverse events (i.e., major arrhythmias, infections, reoperations, bleeding, or transfusions). PCI was also associated with a more rapid recovery and greater improvement in quality of life at 30 days than was CABG, although both procedures resulted in similar quality of life and angina relief at 3 years.47) In the NOBLE trial, the primary composite endpoint of all-cause mortality, nonprocedural MI, stroke, or repeat revascularization at 5 years was significantly higher after PCI than after CABG (29% vs. 19% exceeding the limit for non-inferiority, respectively). The difference in favor of CABG was statistically significant (p value for superiority=0.007) and was driven by significantly higher rates of nonprocedural MI, repeat revascularization, and stroke in the PCI arm.10)

There may be several explanations for the inconsistent results seen in the EXCEL and NOBLE studies.48) First, different types of DES were used. In EXCEL a thin-strut, fluoropolymer-based CoCr-EES was employed, which was associated with the lowest risk of stent thrombosis of all available DES.49) The NOBLE study used first-generation, thicker-strut, stainless-steel, sirolimus-eluting Cypher stents (11%) or the biolimus-eluting Biomatrix Flex stent (89%). A substantial difference in the rate of definite stent thrombosis (0.7% in EXCEL vs. 3% in NOBLE) suggests the differential performance of stenting for LMCA disease. Secondly, the soft clinical endpoint of repeat revascularization was adopted as the key component of the primary endpoint in the NOBLE study. The majority of previous studies have consistently shown that the rate of repeat revascularization is significantly higher after PCI than after CABG. Therefore, the selection of this primary composite outcome may unfairly penalize the PCI stratum. The SYNTAX trial showed that the increase in repeat revascularization in the PCI group did not directly translate into an increase in the incidence of death or MI.50) Thirdly, the definitions used for components of the primary composite outcomes differed between the studies, particularly the definition of MI. The Society for Cardiovascular Angiography and Interventions -defined clinically relevant MI definition was used in EXCEL,51) while periprocedural MI was disregarded in NOBLE. Finally, in the NOBLE study, the rate of stroke was more than two times higher after PCI than after CABG, which is not in agreement with the findings of previous clinical trials comparing PCI and CABG. This observation lacks a clear explanation and biologic plausibility and is, therefore, likely to be due to a chance effect.52)

In a meta-analysis of the four largest studies of LMCA revascularization with follow-up available at 3–5 years, incorporating data from the EXCEL and NOBLE trials, the hazard ratio (HR) for death, stroke, or MI with PCI compared with CABG was neutral (1.06) in a random-effects model (p=0.60).53) Based on individual patient data reconstruction, the Kaplan-Meier estimates of death, stroke, or MI at 5 years were 18.3% for PCI and 16.8% for CABG (p=0.52). No statistically significant subgroup interaction for this combined outcome was noted across studies based on the generation of DES used for PCI (p value for interaction=0.25). There were no significant differences in the pooled effects for death (HR, 1.04; p=0.77) and cardiac death (HR, 1.00; p=0.99). The endpoints of MI and stroke also did not differ between the PCI and CABG groups (HR, 1.48; p=0.17 and 0.87; p=0.72, respectively), but these outcomes were confounded by high heterogeneity across the trials. Repeat revascularization was consistently higher following PCI in all trials, leading to a pooled HR of 1.70 (p<0.001). In another meta-analysis, including all the six trials available to date, missing data were collected by the principal investigators, enabling further subgroup analyses.54) PCI was found to significantly reduce death, MI, or stroke by 36% within 30 days. PCI reduced periprocedural MI by 33%, but this effect was offset by 93% more spontaneous MIs beyond 30 days after the procedure. Cardiac death differed in relation to angiographic complexity in that it tended to be lower with PCI among patients with low SYNTAX scores and higher in patients with high SYNTAX scores.

A recent large-scale, pooled analysis of individual patient data reported a comparable treatment effect for PCI and CABG with regard to all-cause mortality up to 5 years in selected patients participating in RCTs.55) This analysis included 11 RCTs involving 11,518 patients who were assigned to undergo PCI (n=5,753) or CABG (n=5,765). The 5-year rate of all-cause mortality was 11.2% after PCI and 9.2% after CABG (HR, 1.20; 95% CI, 1.06–1.37; p=0.004). Interestingly, the 5-year all-cause mortality differed significantly between the two interventions in patients with multivessel disease (11.5% after PCI vs. 8.9% after CABG; HR, 1.28; 95% CI, 1.09–1.49; p=0.002), including in those with diabetes (15.5% vs. 10.0%, respectively; HR, 1.48; 95% CI, 1.19–1.84; p=0.0004), but not in those without diabetes (8.7% vs. 8.0%, respectively; HR, 1.08; 95% CI, 0.86–1.36; p=0.49). By contrast, the 5-year rate of all-cause mortality was similar between the two groups in patients with LMCA disease (10.7% after PCI vs. 10.5% after CABG; HR, 1.07; 95% CI, 0.87–1.33; p=0.52), regardless of diabetes status and SYNTAX score.

Large registries of patients undergoing revascularization for LMCA disease in different geographical regions are useful resources to generalize the findings from RCTs into the daily clinical scenario. The MAIN-COMPARE, IRIS-MAIN, PRECOMBAT-2, and DELTA-2 registries cover a treatment period of 2000–2015,2)56)57)58)59) and are summarized in Table 2. Compared with PCI patients enrolled in the EXCEL trial, the mean SYNTAX score in these registries tended to be higher. In the DELTA-2 study of LMCA PCI with second-generation DES (n=3,986), Kaplan-Meier estimates of events at 2 years were 9.5% and 16.7% for death or target-vessel revascularization, respectively.56) HRs for PCI vs. CABG with respect to the composite of death, stroke, or MI were similar between the PCI and CABG groups.

Recently, an observation study of the MAIN-COMPARE registry reported 10-year comparative outcomes of PCI and CABG for LMCA disease.60) Overall, there was no significant difference in the adjusted risk of death and the composite outcome between the PCI group and the CABG group up to 10 years. The risk of target-vessel revascularization was significantly higher in the PCI group. However, in the cohort comparing DES and concurrent CABG, DES was associated with a higher risk of death (HR, 1.35; 95% CI, 1.00–1.81) and the composite outcome of death, Q-wave MI, or stroke (HR, 1.46; 95% CI, 1.10–1.94) than was CABG after 5 years. In the DES era, the application of PCI for LMCA disease was substantially expanded and widely performed in patients with a broader range of clinical and anatomical complexities.61) Therefore, the results of this study can be interpreted as follows: CABG is associated with superior long-term outcomes compared with multivessel PCI in patients with high clinical and anatomical complexity. It is interesting to note that the suggestion of superiority of CABG over PCI during long-term follow-up was also seen in the EXCEL and NOBEL trials, which reported a trend towards late catch-up or crossover in the rate of death or the composite endpoint of death, stroke, or MI favoring CABG over PCI during the late period of follow-up. Therefore, longer-term follow-up is necessary to examine additional differences between PCI and CABG over time.

Existing clinical practice guidelines continue to advocate CABG surgery as the singular class I indication for myocardial revascularization of LMCA disease. However, more recent RCTs and registry studies support PCI as a reasonable alternative in selected patients with less complex LMCA anatomy.

As new evidence has become available, guideline recommendations for LMCA revascularization have slowly evolved over time in both Europe and the US (Table 3). Recently, the 2018 European Society of Cardiology guidelines incorporated compelling data from the EXCEL and NOBLE trials, as well as the results of the pooled analysis.62) The 2018 European guideline indicates the same class of recommendation, but all evidence levels have been upgraded to level A. For PCI in LMCA with intermediate anatomical complexity, the previous class IIa recommendation was maintained in view of the incomplete 5-year follow-up in the two largest RCTs in this setting. In the future, the guideline will propose less restrictive indications for PCI, thereby expanding the patient pool that might be eligible for PCI. In addition, given that SYNTAX score was not an important factor for decision-making regarding optimal revascularization and to differentiate the comparative outcomes between CABG and PCI in the EXCEL and NOBLE studies, it may be debated whether the SYNTAX score can play a pivotal role in decision-making regarding LMCA revascularization.

Regardless of which method of revascularization is used, current guidelines highlight the importance of a ‘heart team’ approach to the management of LMCA disease. The heart team evaluates the risks and benefits of PCI, surgery, or medical treatment alone, taking into account the patient's informed preference (Figure 1). In general, PCI offers more rapid recovery and a lower early adverse event rate, whereas CABG offers more durable revascularization. However, the relative outcomes of PCI vs. CABG can be attributed to a complex interplay of patient comorbidities, coronary anatomic complexity, and ventricular function, in addition to other less tangible factors such as operator expertise and compliance with medication. The complexity and extent of coexisting CAD with the intention of achieving complete revascularization should also be considered by the heart team. Previous evaluation has shown that major adverse cardiovascular events are higher in patients with incomplete revascularization than in those with complete revascularization regardless of the revascularization strategy.63) The heart team approach is critical when evaluating the risks and benefits of surgery in high- and extreme-risk populations. Additional clinical factors that are not included in most risk models also need to be considered by the heart team when making management recommendations, including frailty, cognitive status, surgical recovery and social support, quality of life, life expectancy, patient preference, and any potential concerns regarding tolerance or adherence with long-term dual antiplatelet therapy.