Bayesian methods affirm the use of percutaneous coronary intervention to improve survival in patients with...

Area of Science:

• Cardiology
• Medical Statistics
• Interventional Cardiology

Background:

• Coronary artery bypass grafting (CABG) is recommended for unprotected left main coronary artery disease.
• Percutaneous coronary intervention (PCI) is suggested as equivalent to CABG, supporting guidelines for its use to improve survival.
• Bayesian statistical approaches were used to evaluate these recommendations.

Purpose of the Study:

• To assess whether Bayesian statistical methods support the current guidelines recommending PCI for unprotected left main coronary artery disease.
• To compare the effectiveness of PCI and CABG using Bayesian meta-analysis.

Main Methods:

• A Bayesian cross-design and network meta-analysis was conducted.
• Included 12 studies (4 RCTs, 8 observational) comparing CABG vs. PCI (4574 patients).
• Included 7 studies (2 RCTs, 5 observational) comparing CABG vs. medical therapy (3224 patients).

Main Results:

• Bayesian cross-design meta-analysis showed no significant difference in 1-year mortality between PCI and CABG across study types.
• Network meta-analysis indicated medical therapy is associated with significantly higher 1-year mortality compared to PCI.
• Odds ratios for PCI vs. CABG were 0.99 (RCTs), 1.10 (matched cohorts), and 0.93 (other cohorts).

Conclusions:

• Bayesian methods support current guidelines suggesting PCI improves survival for unprotected left main coronary artery disease, similar to CABG.
• Both PCI and CABG demonstrate improved survival compared to medical therapy.
• Integrating direct and indirect evidence via Bayesian approaches can enhance clinical practice.