Fifty-two trials (10,230 participants) were included.
Digoxin: the use of digoxin alone showed benefit in one study, but there was no evidence of any effect in a second. However, used in combination with beta-blockers it appeared to be effective in reducing AF (3 studies).
Beta-blockers: 2 studies showed a beneficial effect of propranolol and two showed no evidence of effect, while timolol and acebutolol showed a benefit in comparison with placebo. Comparisons between different beta-blockers showed sotalol to be superior to metoprolol, whilst both were superior to no beta-blocker. One large well-designed study showed metoprolol to be associated with a reduction in post-operative AF, but not length of hospital stay. The data suggested that beta-blockers were more effective when administered pre- as well as post-operatively.
Calcium-channel blockers: there was a lower incidence of arrhythmias with diltiazem when compared with nitroglycerin. However, evidence concerning the differences in benefit between beta-blockers and calcium-channel blockers was lacking.
Magnesium; 2 small studies showed a benefit with magnesium, whereas two others showed no benefit.
Class I anti-arrhythmic agents: 2 studies showed no benefit with quinidine or propafenone, while 2 small studies showed a non-statistically significant reduction in AF with procainamide.
Class III anti-arrhythmic agents: sotalol consistently showed a reduction in the incidence of post-operative AF (8 studies). However, other studies showed that dose-related adverse events could occur. Six studies showed a benefit of oral amiodarone, started pre-operatively, against placebo. In some of these studies the participants also received beta-blockers, and this appeared to enhance the effect. When intravenous (i.v.) amiodarone was used, 2 studies showed benefit whilst two showed no benefit. Other studies showed that a short course of i.v. amiodarone post-operatively, followed by oral therapy, was effective. This effectiveness may be improved by atrial pacing.
When events were counted across studies, the incidence of AF was 29.2% in controls, 19.3% with beta-blockers, 15.4% with sotalol, and 18% with oral or i.v. amiodarone (12.5% with both oral and i.v., 16.5% with oral alone, and 31.4% with i.v. alone). The authors also briefly discussed the use of non-steroidal anti-inflammatory agents and angiotensin antagonists.