BARI 2D

Published

Frye RL, et al. "A Randomized Trial of Therapies for Type 2 Diabetes and Coronary Artery Disease". The New England Journal of Medicine. 2009. 360(24):2503-2515.
PubMed • Full text • PDF

Clinical Question

Among patients with T2DM and stable CAD, how does revascularization with either CABG or PCI compare to OMT in reducing CV events and death?

Bottom Line

Among patients with T2DM and stable CAD that are CABG candidates, CABG and OMT reduced the rate of CV events compared to OMT alone. There was no difference in the PCI cohort.

Major Points

Revascularization with either CABG or PCI is recommended in patients with stable CAD and persistent angina refractory to OMT. The optimal revascularization strategy for stable CAD in diabetic patients, whom are more likely to have diffuse, multivessel disease, had not yet been studied rigorously. COURAGE (2007) demonstrated no difference between PCI and OMT, although only 32% were diabetic and CABG candidates were excluded. A post hoc subgroup analysis of the BARI study (1996) suggested that CABG improved survival compared to PCI among diabetic patients. However, a direct comparison of revascularization strategies had not yet been undertaken among diabetics with stable CAD.

The Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) trial enrolled 2,368 patients with T2DM and stable CAD. At enrollment, patients were assigned by their treating physician to either CABG or PCI. Patients were then randomized to OMT or OMT plus revascularization. The majority of CABG patients were treated on-pump (64%) and received an internal mammary graft (94.2%), while the majority of PCI patients underwent a single-vessel intervention (79.3%) and received a BMS (56%). At 5 years, there was no difference in survival (88.3% vs. 87.8%) and freedom from major CV events (77.2% vs. 75.9%). However, an analysis of the subgroup of patients assigned to CABG, those randomized to CABG and OMT had more freedom from CV events compared to those randomized to OMT alone (77.6% vs. 69.5%). Conversely, there were no differences in 5-year survival (89.2% vs. 89.8%) or freedom from major CV events (77.0% vs. 78.9%) among the subgroup of patients assigned to PCI.

BARI 2D was not designed to be a direct comparison of revascularization strategies, and the results of the BARI 2D subgroup analyses cannot be made generalizable to mean that CABG is superior to PCI among diabetic patients with stable CAD. Since patients chosen to the PCI cohort may have less severe CAD than those chosen to the CABG cohort, the physician-directed allocation of treatment groups inherently introduced selection bias. Specifically, among the subset of patients deemed appropriate for CABG, BARI 2D suggests CABG plus OMT may reduce the rate of CV events compared to OMT alone, whereas, PCI plus OMT does not reduce the rate of CV events among the subset of patients deemed appropriate for PCI.

Ultimately, FREEDOM (2012) randomized patients with diabetes and multivessel CAD to either PCI with DES or CABG and found that CABG was superior to PCI in reducing the rates of both MI and all-cause mortality, with a higher rate of strokes.

Guidelines

ACCF/AHA Guideline for Coronary Artery Bypass Graft Surgery (2011)

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

Design

• Multicenter, open-label, parallel group, randomized trial
• N=2,368
  • Revascularization (n=1,176)
  • Optimal medical therapy (n=1,192)
• Setting: 49 international sites
• Enrollment: 2001-2005
• Mean follow-up: 5.3 years
• Analysis: Intention-to-treat
• Primary outcome: All-cause mortality

Population

Inclusion Criteria

• T2DM requiring insulin or oral hypoglycemic agent
• Angiographic CAD, defined as:
  • ≥50% stenosis of a major coronary artery and a positive stress test, or
  • ≥70% stenosis of a major coronary artery and classic angina
• Eligible for elective CABG or PCI

Exclusion Criteria

• Patients requiring immediate revascularization
• Left main CAD
• Cr >2 mg/dL
• HbA1c >13%
• Class III-IV heart failure
• Liver dysfunction
• Revascularization within prior 12 months

Baseline Characteristics

• Mean age: 62 years
• Male: 70.4%
• Race: white (65.9%), black (16.8%), Hispanic (12.5%), Asian/other (4.8%)
• Region: US (63.3%), Canada (14.9%), Brazil (15%), Mexico (3.6%), Czech/Austria (3.2%)
• Mean HbA1c: 7.7%
• Mean duration of diabetes: 10.4 years
• Taking insulin: 27.9%
• PMH: prior MI (32%), CHF (6.6%), CVA (9.8%), prior coronary revascularization (23.6%)
• Angina category: stable angina 1-2 (42.5%), stable angina 3-4 (8.6%), unstable angina (9.5%), anginal equivalents but no angina (21.4%), no angina (17.9%)
• 3-vessel CAD: 30.7%
• Proximal LAD disease: 13.2%
• Mean LVEF: 57.2% (LVEF <50%: 17.5%)

Interventions

• Patients were enrolled and each had a revascularization strategy (either CABG or PCI) selected by their treating physician
• Patients randomized to either prompt revascularization or optimal medical therapy
  • Prompt revascularization group underwent revascularization within 4 weeks of randomization
  • Optimal medical therapy group could be revascularized only for angina progression or ACS
• A second randomization studied insulin sensitization vs. insulin provision therapy, but these results are not discussed here
• All patients received optimal medical therapy, targeting:
  • HbA1c <7%
  • LDL <100
  • BP <130/80
  • Smoking cessation
  • Weight loss
  • Regular exercise

Outcomes

Comparisons are revascularization vs. optimal medical therapy at 5 years.

Primary Outcome

5-year survival

88.3% vs. 87.8% (95% CI −2.0 to 3.1; P=0.97)

Secondary Outcomes

TIA

2.5% vs. 2.7% (P=0.75)

HF

21.3% vs. 21.2% (P=0.96)

Subgroup Analysis

Survival in CABG stratum

86.4% vs. 83.6% (P=0.33)

Freedom from major CV events in CABG stratum

77.5% vs. 69% (P=0.01)

Survival in PCI stratum

89.2% vs. 89.8% (P=0.48)

Freedom from major CV events in PCI stratum

77% vs. 78.9% (P=0.15)

Criticisms

• Patients were included based upon a coronary anatomy that was amenable to revascularization, limiting the generalizability of the results
• The physician assignment to PCI or CABG precludes drawing comparative conclusions from their outcomes
• The use of CK 10x the ULN as the definition for perioperative MI may have been too high, multiple groups use 5x the ULN as their definition
• A combination of metformin and insulin was not studied, some centers consider this the standard of care
• No data on the amount of screened but excluded patients was given
• Enrolled patients were relatively low risk for CV events based upon their anginal symptoms
• Low rate of events and fewer patients enrolled than original goal, potentially resulting in underpowering of outcome detection
• Low rate of DES and TZD use
• High rate of crossover

Funding

• National Heart, Lung, and Blood institute
• National Institute of Diabetes and Digestive and Kidney Diseases
• Multiple private sources of funding: GlaxoSmithKline, Lantheus Medical Imaging, Astellas Pharma, Merck, Abbott Laboratories, Pfizer, MediSense, Bayer, Becton Dickinson, J.R. Carlson Labs, Centocor, Eli Lilly, LipoScience, Novartis, and Novo Nordisk