Revascularization for unprotected left main coronary artery disease- an update from recent trials

In June 2010 and March 2014 BCS editorials provided updates on revascularization of left main coronary artery (LMCA) disease. A number of relevant trials have since been reported. Importantly 2 key trials, EXCEL and NOBLE, were published last month. Following is the current state of play on LMCA revascularization in view of recent evidence.
Left main coronary artery (LMCA) disease
About 4-7 % patients undergoing coronary angiogram have significant left main coronary artery (LMCA) disease^1,2; a finding that has prognostic implications because of the amount of myocardium at risk. In the Coronary Artery Surgery Study (CASS) registry survival without surgical revascularization in the presence of LMCA, even in asymptomatic patients, was 57% at 5 years, which was improved to 88% with coronary artery bypass surgery (CABG)². CABG has traditionally been the preferred mode of revascularization of LMCA; an early consensus that was further reinforced because of limited experience and sub-optimal outcomes with percutaneous coronary intervention (PCI) on LMCA in the balloon angioplasty and bare metal stent era. 
Early evidence
With the advances in the PCI techniques and adjunct pharmacotherapy, there has been a renewed interest in using PCI for LMCA revascularization over the last decade with reporting of several trials. This included reporting of early outcomes of LE MANS, PreCOMBAT, SYNTAX and other RCTs as well as meta-analyses and registry data, as described in the earlier BCS editorials. This early evidence supported a role of PCI in LMCA revascularization, mainly with low or intermediate coronary artery disease (CAD) complexity, which was reflected in the major society guidelines on myocardial revascularization.
Guidelines
In the European Society of Cardiology (ESC) guidelines on myocardial revascularization (2014) CABG has a class 1 recommendation for LMCA disease regardless of the SYNTAX score³. For PCI; patients with LMCA and SYNTAX score <22 have a class I and those with SYNTAX score 23-32, a class IIa indication. PCI is not indicated for patients with LMCA disease and SYNTAX score >33 (a class III indication).
Recent evidence
Long-term follow up data from earlier trials
The 10-year outcomes of the Study of Unprotected Left Main Stenting versus Bypass Surgery (LE-MANS) were reported earlier this year⁴. In this prospective, multicentre trial 105 patients with unprotected LMCA disease and low to medium complexity CAD were randomised to PCI or CABG. At 10 years PCI was associated with favourable primary (change in left ventricular ejection fraction) and secondary (major adverse cardiac and cerebrovascular events(MACCE)) outcomes. The Premier of Randomized Comparison of Bypass Surgery versus Angioplasty using Sirolimus-Eluting Stent in Patients with Left Main Coronary Artery Disease (PreCOMBAT) reported 5 years outcomes last year⁵. The trial had randomized 600 patients to PCI or CABG. At 5 years there was no significant difference in the primary endpoint (MACCE) between the 2 cohorts. These earlier trials were limited by small sample sizes and/or use of first generation DES. The landmark Synergy between PCI with Taxus and Cardiac Surgery (SYNTAX) trial randomised a subset of 705 patients with unprotected LMCA to PCI (with first generation drug eluting stent (DES) Taxus Express) or CABG⁶. At 5 years, overall there was no significant difference in MACCE between the two cohorts. However the outcomes from PCI were dependant on the complexity of the CAD. At low to medium complexity (objectively quantified by the SYNTAX tertiles of 0-22 and 22-32) the outcomes were similar between PCI and CABG. With the more complex CAD (SYNTAX tertile of score >32) CABG had better outcomes than PCI. Also post procedural stroke was more frequent with CABG (4.3% vs 1.5%) and repeat revascularization with PCI (26.7% 15.5%). The SYNTAX trial supported the idea of “heart team” and objective quantification of CAD complexity by SYNTAX score.
Although these early trials supported the use of PCI in select patients with LMCA, the exact role of PCI, using the contemporary DES, modern pharmacological agents and surgical techniques, remains somewhat undefined. Recently 2 key trials (EXCEL and NOBLE) comparing CABG with PCI for the treatment of unprotected left main disease have been published.
Recent trials 
The Everolimus-Eluting Stents or Bypass Surgery for Left Main Coronary Artery Disease (EXCEL) trial randomized 1905 patients with unprotected LMCA disease and low (SYNTAX score <22) or intermediate (SYNTAX score 23-32) complexity CAD to PCI with a DES (mean stents 2.4, mean stent length 49 per patient) vs. CABG (Internal mammary graft 99%, total arterial grafts 25%).  Two or three vessel CAD was present in 51% of the patients and distal left main bifurcation (or trifurcation) disease in 80%. The trial showed that PCI in this context is non-inferior (but not superior) to CABG. At median follow up of 3 years, primary end point (composite of all-cause death, myocardial infarction or stroke) occurred with similar frequency in both groups (15.4 % for PCI vs. 14.7% for CABG, P=0.02 for non-inferiority, P=0.98 for superiority).  Cardiac mortality was similar between the two groups. It was also evident that in the first 30 days, more primary end point events (primarily myocardial infarction and stroke) occurred following CABG than after PCI but between 30 days to 3 years there was an increase in primary end point events in PCI group with similar outcomes at 3 years. At 3 years ischemia-driven revascularization occurred more frequently after PCI (12.6%) than after CABG (7.5%, P<0.001). The incidence of definite stent thrombosis and stroke following PCI was 0.7% and 2% respectively at 3 years. This further demonstrated the lower peri-procedural morbidity of PCI (peri-procedural major adverse events 8% vs 23%) but a higher repeat revascularization rate compared to CABG.
The Percutaneous coronary angioplasty versus coronary artery bypass grafting in treatment of unprotected left main stenosis (NOBLE), a non-inferiority, trial randomised 1201 patients to PCI with a DES (11% first generation, 89% Biolimus-eluting second generation) or CABG (86% with arterial graft to left anterior descending artery). In both groups 81% patients had distal left main bifurcation disease. In patients undergoing PCI, 33% had 1 additional and 9% patients had 2 additional lesions treated.  The 5 year Kaplan-Meier estimates of composite primary end point (MACCE; all cause mortality, non procedural myocardial infarction, repeat revascularization or stroke) were supportive of superiority of CABG over PCI  (19% vs. 29% respectively, P=0.0066). For the individual components of the composite primary end point, the all cause mortality was not significantly different between CABG and PCI (9 % vs. 12% respectively, P=0.77) and the difference was primarily driven by higher non procedural myocardial infarction (2% vs. 7%, P=0.0040) and repeat revascularization (10% vs. 16%, P=0.032) estimates following PCI. Somewhat surprisingly, the PCI was also associated with increased 5 year estimates of stroke (5% vs. 2%, P=0.073). Considering that there were no incidents of stroke at 30 days in the PCI cohort, the 5-year estimate may be a chance finding.
Both EXCEL and NOBLE trials seem to have reached a conflicting conclusion, however some points are worth noting. The inclusion of repeat revascularization in composite primary end points in NOBLE trial may partly explain the increased MACCE following PCI. Also the incidence of stent thrombosis (3%) and stroke (5%) following PCI in the NOBLE trial was unexpectedly higher. The late increase (between 30 days and 3 years) in event rates in the PCI cohort in EXCEL trial also makes the availability of long-term data quite important.
Conclusions
The EXCEL and NOBLE trial are a valuable addition to the decision making in treatment of patients with unprotected left main coronary artery disease. The role of heart team enabling an informed patient decision remains paramount. Based on the available evidence, PCI is a suitable alternative to CABG in patients with unprotected LMCA and low to intermediate anatomic complexity.  Long term follow up data from the recent trials would be important.
References;
1. Main left coronary artery disease. Clinical experience from 1964-1974. Cohen MV, Gorlin R. Circulation. 1975 ;52(2):275-85.
 
2. Asymptomatic left main coronary artery disease in the Coronary Artery Surgery Study (CASS) registry. Taylor HA et al. Circulation. 1989; 79: 1171-1179. DOI: 10.1161/01.CIR.79.6.1171.
 
3. 2014 ESC/EACTS Guidelines on myocardial revascularization. Windecker S et al. Eur Heart J. 2014 Oct 1;35(37):2541-619. DOI: 10.1093/eurheartj/ehu278.
 
4. Left Main Stenting in Comparison With Surgical Revascularization: 10-Year Outcomes of the (Left Main Coronary Artery Stenting) LE MANS Trial. Buszman PE et al. J Am Coll Cardiol Intv. 2016;9(4):318-327. 
 
5. Randomized Trial of Stents Versus Bypass Surgery for Left Main Coronary Artery Disease: 5-Year Outcomes of the PRECOMBAT Study. Ahn J et al. J Am Coll Cardiol.2015;65(20):2198-2206. 
 
6. Five-year outcomes in patients with left main disease treated with either percutaneous coronary intervention or coronary artery bypass grafting in the synergy between percutaneous coronary intervention with taxus and cardiac surgery trial. Morice MC et al.  Circulation. 2014 Jun 10; 129 (23): 2388-94.
 
7. Everolimus-Eluting Stents or Bypass Surgery for Left Main Artery Coronary Disease. Stone GW et al. N Engl J Med. 2016 Oct 31. [Epub ahead of print]. PMID:27797291.DOI: 10.1056/NEJMoa1610227.
 
8. Percutaneous coronary angioplasty versus coronary artery bypass grafting in treatment of unprotected left main stenosis (NOBLE): a prospective, randomised, open-label, non-inferiority trial. Mäkikallio T et al.  Lancet. 2016 Oct 31. pii: S0140-6736(16)32052-9. doi: 10.1016/S0140-6736(16)32052-9. [Epub ahead of print].