Economic Analysis of Intravenous Plus Oral Amiodarone, Atrial Septal Pacing, and Both Strategies to Prevent Atrial Fibrillation After Open Heart Surgery

Study Objectives. To compare the cost‐effectiveness of intravenous plus oral amiodarone, atrial septal pacing, and both strategies combined to prevent atrial fibrillation after open heart surgery. Secondary objectives were to compare the cost‐effectiveness of amiodarone versus no amiodarone and of pacing versus no pacing, and to compare hospitalization costs of the various strategies. Design. Piggyback cost analysis of a randomized, 2 times 2 factorial trial. Setting. Urban academic hospital. Patients. One hundred and sixty patients with coronary artery and/or valvular disease. Intervention. Patients were randomized to receive amiodarone or matching placebo and then further randomized to receive atrial septal pacing or no pacing. Measurements and Main Results. The economic analysis was conducted from a hospital perspective. Charges were converted to costs using cost: charge ratios. For the cost‐effectiveness analysis, a joint distribution of costs and effectiveness was performed using the nonparametric bootstrap method. Amiodarone plus pacing significantly decreased the frequency of atrial fibrillation after open heart surgery, compared with amiodarone alone, pacing alone, and placebo. Total costs (mean ± SD) were $27,026 ± 30,226 for the placebo group, $22,725 ± 17,661 for the amiodarone group, $33,868 ± 60,309 for the pacing group, and $18,697 ± 8174 for the amiodarone plus pacing group (p=0.27). In the joint distribution cost‐effectiveness analysis, when compared with placebo, the probability of lower cost but higher effect (superiority) was 67% for amiodarone, 15% for pacing, and 97% for amiodarone plus pacing. In the multivariate analysis, preoperative β‐blockers and amiodarone were negatively associated with hospital costs (p<0.05). Conclusions. Data suggest that both amiodarone alone and the combination of amiodarone plus pacing are cost‐effective compared with placebo. Additional comparative studies of these strategies are warranted to confirm these findings.