The New England Journal of Medicine OCTOBER 31, 2016

Evaluation of XIENCE Versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization (EXCEL) Trial

Gregg W. Stone, A. Pieter Kappetein, Joseph F. Sabik, et al.

Bottom Line

The EXCEL trial demonstrated that in patients with unprotected left main coronary artery disease and low-to-intermediate anatomical complexity (SYNTAX score ≤32), percutaneous coronary intervention (PCI) with second-generation drug-eluting stents was noninferior to coronary artery bypass grafting (CABG) for the primary composite endpoint of death, stroke, or myocardial infarction at 3 years.

Key Findings

1. At 3 years, the primary endpoint (death, stroke, or myocardial infarction) occurred in 15.4% of patients in the PCI group versus 14.7% in the CABG group (P=0.02 for noninferiority; hazard ratio 1.00; 95% CI 0.79-1.26; P=0.98 for superiority).
2. At 30 days, the composite rate of death, stroke, or myocardial infarction was significantly lower with PCI (4.9%) compared to CABG (7.9%), P=0.008.
3. At 3 years, ischemia-driven revascularization was higher in the PCI group (12.6%) compared to the CABG group (7.5%), P<0.0001.
4. At 5 years, the primary endpoint occurred in 22.0% of patients in the PCI group and 19.2% in the CABG group (P=0.13), with numerically higher all-cause mortality in the PCI group (13% vs. 9.9%).

Study Design

Design
RCT
Open-Label
Sample
1,905
Patients
Duration
5 yr
Median
Setting
Multicenter, international
Population Patients with unprotected left main coronary artery disease and low-to-intermediate anatomical complexity (SYNTAX score ≤32).
Intervention PCI with second-generation everolimus-eluting stents.
Comparator Coronary artery bypass grafting (CABG).
Outcome Composite of death, stroke, or myocardial infarction at 3 years.

Study Limitations

The trial was open-label, which could introduce bias in the reporting of subjective endpoints or revascularization decisions.
The primary endpoint was powered for noninferiority at 3 years, limiting the formal statistical strength of long-term (5-year) superiority conclusions.
The results are strictly applicable only to patients with low or intermediate anatomical complexity (SYNTAX score ≤32), excluding complex multi-vessel disease patients.
Controversy exists regarding the definitions of periprocedural myocardial infarction used in the trial, which significantly influenced the primary endpoint results.

Clinical Significance

The EXCEL trial provides evidence that PCI is a viable and noninferior revascularization strategy to CABG for patients with unprotected left main coronary artery disease who have low-to-intermediate anatomical complexity. It supports the role of the 'Heart Team' approach in determining the most appropriate revascularization method based on individual patient anatomy, comorbidities, and clinical risk profiles.

Historical Context

Prior to the EXCEL trial, CABG was the gold standard for unprotected left main disease based on historical evidence. As stent technology evolved with second-generation drug-eluting stents (DES), the medical community sought to determine if less-invasive PCI could achieve outcomes comparable to surgery for specific anatomical subgroups, leading to trials like EXCEL and the concurrent NOBLE trial.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

Why is the left main coronary artery (LMCA) referred to as the most critical segment of the coronary anatomy, and how does its occlusion differ physiologically from a distal occlusion?

Key Response

The LMCA supplies approximately 75% to 100% of the left ventricular myocardium via the left anterior descending and circumflex arteries. Occlusion leads to massive myocardial ischemia, rapidly progressing to cardiogenic shock and lethal arrhythmias, whereas distal occlusions affect smaller territories with higher likelihood of collateral compensation.

Resident
Resident

The EXCEL trial focused on patients with a SYNTAX score of 32 or less. What is the clinical significance of this score in choosing between PCI and CABG for left main disease?

Key Response

The SYNTAX score quantifies coronary anatomical complexity. Scores ≤32 represent low-to-intermediate complexity where PCI is often competitive with surgery. For scores >32 (high complexity), CABG remains the standard of care due to the significantly higher risk of late adverse events and incomplete revascularization associated with PCI in complex multivessel disease.

Fellow
Fellow

Critically analyze the impact of the 'Protocol-defined Myocardial Infarction' used in EXCEL versus the 'Universal Definition of MI' on the study's primary endpoint results.

Key Response

EXCEL utilized a protocol-specific definition for periprocedural MI (CK-MB >10x ULN), which was more sensitive in capturing surgical MI than the Universal Definition. Critics argue this definition favored PCI by inflating the number of periprocedural events in the CABG arm, thereby making PCI appear noninferior for the composite primary endpoint of death, stroke, or MI.

Attending
Attending

Considering the 3-year EXCEL results and the subsequent 5-year follow-up showing a divergence in all-cause mortality, how should the informed consent process change for a patient with left main disease?

Key Response

While PCI shows noninferiority at 3 years with lower early risks of stroke and faster recovery, 5-year data suggested a potential late mortality signal (13.0% PCI vs 9.9% CABG). Consent must balance early PCI benefits (reduced periprocedural morbidity) against the superior long-term durability and potential survival advantage of CABG, particularly in younger patients.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

Discuss the implications of using a noninferiority margin of 4.2% for the primary composite endpoint in EXCEL. How does the choice of this margin affect the trial's power and the risk of a Type I error in clinical interpretation?

Key Response

The noninferiority margin defines the maximum acceptable clinical difference to claim 'no worse than.' A relatively wide margin increases the likelihood of a positive noninferiority result. If the margin is larger than the clinically relevant benefit of the gold standard (CABG), the trial may conclude noninferiority despite a potentially significant disadvantage in 'hard' outcomes like mortality.

Journal Editor
Journal Editor

As a peer reviewer, how would you address the discrepancy between the primary endpoint results at 3 years and the secondary 5-year all-cause mortality data that emerged later, and what does this say about the choice of follow-up duration for revascularization trials?

Key Response

A seasoned reviewer would flag that 3 years is often too short to capture the 'late catch-up' phenomenon where the surgical benefits of graft patency outweigh the early risks of bypass. This discrepancy suggests that for revascularization trials involving left main disease, 5-to-10-year follow-up should be a mandatory requirement for definitive editorial acceptance of noninferiority claims.

Guideline Committee
Guideline Committee

Based on the EXCEL and NOBLE trials, should PCI for left main disease with low SYNTAX scores be elevated to a Class I recommendation in the next guideline update, or should it remain Class IIa?

Key Response

Current ACC/AHA and ESC guidelines generally maintain CABG as Class I due to long-term survival data. Despite EXCEL's 3-year noninferiority, the NOBLE trial (which used a different MI definition) did not show noninferiority for PCI, and the EXCEL 5-year mortality signal provides a cautious barrier to a Class I upgrade, favoring the continued recommendation of CABG for long-term durability.

Clinical Landscape

Noteworthy Related Trials

2009

SYNTAX Trial

n = 1,800 · NEJM

Tested

CABG

Population

Patients with de novo three-vessel disease or left main coronary disease

Comparator

PCI with paclitaxel-eluting stents

Endpoint

Major adverse cardiac or cerebrovascular events at 12 months

Key result: CABG remained the standard of care for patients with complex three-vessel or left main disease, as PCI was associated with higher rates of repeat revascularization.
2011

PRECOMBAT Trial

n = 600 · NEJM

Tested

PCI with sirolimus-eluting stents

Population

Patients with unprotected left main coronary artery stenosis

Comparator

CABG

Endpoint

Major adverse cardiac or cerebrovascular events at 12 months

Key result: PCI was non-inferior to CABG for the composite of death, myocardial infarction, stroke, or ischemia-driven target vessel revascularization at 1 year.
2016

NOBLE Trial

n = 1,201 · Lancet

Tested

PCI with biolimus-eluting stents

Population

Patients with unprotected left main coronary artery disease

Comparator

CABG

Endpoint

Major adverse cardiac or cerebrovascular events

Key result: CABG was superior to PCI in preventing major adverse cardiac or cerebrovascular events at 5 years, primarily due to higher rates of repeat revascularization in the PCI group.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis