From Wiki Journal Club
Jump to navigation Jump to search
Farkouh ME, et al. "Strategies for multivessel revascularization in patients with diabetes". The New England Journal of Medicine. 2012. 367(25):2375-2384.
PubMedFull textPDF

Clinical Question

Among diabetic patients with multivessel coronary artery disease, how does revascularization with CABG compare to PCI in terms of rates of death, myocardial infarction, and stroke?

Bottom Line

Among diabetic patients with multivessel coronary artery disease (CAD), revascularization with CABG reduces the rates of death and myocardial infarction compared to PCI, but causes a modest increase in the rate of stroke.

Major Points

Patients with diabetes mellitus (DM) and CAD tend to have more extensive atherosclerotic disease than those without diabetes. As a result, these patients may derive a particular benefit from bypass grafting over percutaneous coronary revascularization in the setting of multivessel CAD. BARI (1996)[1] was one of the first trials to demonstrate a survival benefit with CABG over coronary angioplasty in patients with DM and multivessel CAD. Furthermore, the related BARI 2D (2009) suggested that CABG was superior to PCI with bare metal stents (BMS) in patients with DM although it did not compare the two approaches directly. Similarly, SYNTAX (2009), which randomized patients with multivessel CAD to CABG versus first-generation drug-eluting stents (DES), demonstrated a higher rate of major adverse coronary and cerebrovascular events, an effect that was more pronounced in patients with DM.

The FREEDOM trial was the first randomized comparison of PCI with first-generation paclitaxel/sirolimus DES PCI versus CABG which studied exclusively patients with DM and multivessel CAD. The FREEDOM trial demonstrated that in patients with DM and angiographically-confirmed multivessel CAD (≥70% stenosis in ≥2 major epicardial vessels) undergoing revascularization, CABG provided a 5% absolute reduction in all-cause mortality versus PCI and also reduced rates of both myocardial infarction and need for repeat revascularization. Consistent with previous studies, there was also a 3% absolute increase in stroke with CABG.

The recent BEST trial[2] corroborates these findings by demonstrating a lower rate of death, MI, or repeat revascularization with CABG even when compared to second-generation stents. The superiority of CABG was significantly more marked in the DM subgroup in this study. Moreover, in a recent large registry-based observational study of CABG versus second-generation PCI in multivessel CAD[3], CABG was associated with a lower risk of spontaneous MI, although this benefit did not appear to differ by DM status.

Taken together, these results suggest that CABG remains superior to PCI in multivessel CAD as it is associated with lower rates of MI and repeat revascularization. This difference appears more pronounced in patients with DM, making CABG an even more desirable option for revascularization in patients with DM if their surgical risk profile is acceptable.


2011 ACCF/AHA Guideline for Coronary Artery Bypass Graft Surgery[4]

  • CABG is probably recommended in preference to PCI to improve survival in patients with multivessel CAD and DM, particularly if a LIMA graft can be anastomosed to the LAD artery (class IIA, level B)


  • Multicenter, randomized, open-label, controlled trial
  • N=1900
    • PCI (n=953)
    • CABG (n=947)
  • Setting: 120 centers in the US and internationally
  • Enrollment: 2005-2010
  • Mean follow-up: 3.8 years
  • Analysis: Intention-to-treat
  • Primary outcome: Death from any cause, nonfatal MI, and nonfatal stroke
  • Secondary outcomes: Major adverse cardiovascular and cerebrovascular events (MACCE) at 30 days, MACCE at 12 months, all-cause mortality, cardiovascular mortality


Inclusion Criteria

  • Age ≥18 years
  • Diabetes mellitus defined by ADA
  • Angiographically confirmed multivessel CAD (≥70% stenosis in ≥2 major epicardial vessels and ≥2 separate territories)
  • Indication for revascularization (by symptoms and/or objective ischemia)

Exclusion Criteria

  • NYHA class III-IV heart failure or pulmonary edema
  • Prior CABG
  • Prior heart valve surgery
  • Prior PCI with stent placement within prior 6 months
  • Prior significant bleeding within prior 6 months
  • In-stent restenosis of target vessel
  • ≥2 CTOs in major coronary territories considered targets for revascularization
  • Left main stenosis ≥50%
  • STEMI within 72 hours before enrollment
  • CK >2x ULN
  • Planned simultaneous surgical procedure unrelated to coronary revascularization
  • Contraindication to CABG or PCI/DES
  • Significant cytopenias of bleeding diathesis
  • Contraindication to aspirin or clopidogrel and ticlopidine
  • Dementia (MMSE <20)
  • Extracardiac illness limiting survival to <5 years (eg, severe COPD)
  • Suspected pregnancy

Baseline Characteristics

Characteristics are PCI vs. CABG

  • Age (yrs) 63.2 vs. 63.1
  • Male (%) 73.2 vs. 69.5
  • A1c (%) 7.8 vs. 7.8
  • Smoker (%) 14.8 vs. 16.6
  • Previous MI (%) 26.2 vs. 25
  • Recent ACS (%) 31.9 vs. 29.5
  • 3-vessel disease (%) 82.3 vs. 84.5
  • LVEF < 40% (%) 3.3 vs. 1.7
  • Mean EuroSCORE 2.7 vs. 2.8
  • SYNTAX score 26.2 vs. 26.1
    • Low (<23) (%) 34.7 vs. 36.2
    • Intermediate (23-32) (%) 46.2 vs. 43.3
    • High (>32) (%) 19.2 vs. 20.5
  • Number of lesions (#) 5.7 vs. 5.7
  • Chronic total occlusion (%) 5.8 vs. 5.8
  • Use of insulin (%) 33.8 vs. 30.9

PCI-only characteristics

  • Staged procedure: 34.2%
  • Total number of lesions stented (#): 3.5
  • Total length of stents placed (mm): 26.1

CABG-only characteristics

  • Surgery off-pump: 18.5%
  • Number of grafts: 2.9
  • Use of LIMA graft: 94.4%


  • Patients were randomized in 1:1 fashion to receive either PCI with first-generation paclitaxel-eluting DES (51%) or sirolimus-eluting DES (43%) versus CABG.
  • For PCI, dual antiplatelet therapy with aspirin and clopidogrel was recommended for at least 12 months after DES.
  • For CABG, arterial revascularization was encouraged.
  • For both groups, achievement of guideline-driven risk factor targets was encouraged (LDL <70, HbA1c <7, BP <130/80).
  • Patients were followed and screened for stroke using the NIH Stroke Scale and modified Rankin scale at each follow-up visit.
  • Minimum follow-up was 2 years.
  • Events committee provided central independent adjudication of all occurrences of the primary endpoints in an unblinded fashion.


Comparisons are PCI vs. CABG.

Primary Outcome

Death, MI, or stroke
26.6% vs. 18.7% (P=0.005)

Secondary Outcomes

All-cause mortality
16.3% vs. 10.9% (P=0.049)
Myocardial infarction
13.9% vs. 6.0% (P<0.001)
2.4% vs. 5.2% (P=0.03)
Cardiovascular death
10.9% vs. 6.8% (P=0.12)

Subgroup Analysis

All subgroup outcomes reflect primary outcome of death, MI, or stroke. Comparisons are PCI vs. CABG.

SYNTAX score (interaction P=0.58)
<23: 23% vs. 17%
23-32: 27% vs. 18%
>32: 31% vs. 23%
LVEF (interaction P=0.37)
<40: 62% vs. 31%
≥40: 23% vs. 18%
Glycated hemoglobin (interaction P=0.99)
<7: 23% vs. 16%
≥7: 28% vs. 20%
Region (interaction P=0.05)
North America: 28% vs. 16%
Other 25% vs. 21%

Adverse Events

Comparisons are PCI vs. CABG.

Major adverse cardiovascular and cerebrovascular events at 30 days
4.8% vs. 5.2% (P=0.68)
Major adverse cardiovascular and cerebrovascular events at 12 months
16.8% vs. 11.8% (P=0.004)
MI at 30 days
1.8% vs. 1.7% (P=0.82)
MI at 12 months
5.8% vs. 3.4% (P=0.02)
Stroke at 30 days
0.3% vs. 1.8% (P=0.002)
Stroke at 12 months
0.9% vs. 1.9% (P=0.06)
Repeat revascularization at 30 days
3.3% vs. 1.1% (P=0.002)
Repeat revascularization at 12 months
12.6% vs. 4.8% (P<0.001)
Major Bleeding at 30 days
0.02% vs. 0.04% (P=0.13)
Acute Renal Failure at 30 days
0.001% vs. 0.008% (P=0.06)


  • Protocol was amended twice to account for slow enrollment. As a result, the study enrolled 500 fewer patients than originally intended.
  • Study was not blinded to either intervention or outcome assessment, introducing risk of bias in assessment and post-intervention therapy.
  • Results apply only to first-generation DESs which have been largely replaced by superior second-generation everolimus-eluting stents, which have yet to be studied specifically in the diabetic population with multivessel CAD.


Study was sponsored by the National Heart, Lung, and Blood Institute (NHLBI). Cordis, Johnson and Johnson, and Boston Scientific provided stents. Eli Lilly provided abiciximab and an unrestricted research grant. Sanofi-Aventis and Bristol-Myers Squibb provided clopidogrel.

Further Reading