Six RCTs were included in the review and meta-analysis (1,033 patients). Length of follow-up was either not reported (two trials) or was reported as being one month, the entire hospital stay or time spent in an intensive care unit. Total Jadad scores ranged from 2 to 5; four of the six trials scored a total score of 4 or 5. All six trials reported adequate randomisation. Completeness of follow-up and description of withdrawals were each reported by five trials and four trials reported random sequence generations. Three trials involved double-blinding.
Incidence of atrial fibrillation
No statistically significant difference in the incidence of atrial fibrillation was found between amiodarone and beta-blockers, using fixed-effect (OR 0.81, 95% CI 0.61 to 1.08; six trials) or random-effects models (OR 0.77, 95% CI 0.55 to 1.06; six trials). No significant statistical heterogeneity was indicated (Ι²=34%). Sensitivity analyses relating to type of beta-blocker administration, study quality, type of surgery, and use of beta-blockers in the amiodarone groups also demonstrated non-significant results (Ι² range 26 to 71%, results reported fully in the paper).
Subgroup analyses revealed a statistically significant lower incidence of postoperative atrial fibrillation with amiodarone, compared with propranolol (OR 0.46, 95% CI 0.27 to 0.78). Results were not statistically significant when amiodarone was compared with bisoprolol or metoprolol.
Other outcomes
No statistically significant differences were shown between amiodarone and beta-blockers for length of hospital stay (WMD -0.05 day, 95% CI -0.64 to 0.54; three trials; Ι²=0%), mean ventricular rate following cardiac surgery (WMD -2.31 beats/min, 95% CI -9.98 to 5.36; two trials; Ι²=0%) and overall risk of postoperative adverse events (OR 1.00, 95% CI 0.72 to 1.38; three trials; Ι²=0%).
Evidence of publication bias was not reportedly found.