Aims: The efficacy of cardioversion (DCCV) for restoration of sinus rhythm (SR) in persistent atrial fibrillation (AF) is limited by a high relapse rate. Relapse may be reduced by amiodarone but no placebo-controlled trials of efficacy have been performed and the appropriate duration of therapy is unknown. Method and results: In this double-blind study, 161 subjects with persistent AF were randomized to one of three groups - placebo (n=38); amiodarone 400 mg BD for 2 weeks prior to DCCV and 200 mg daily for 8 weeks followed by placebo for 44 weeks (n=62, short-term amiodarone); amiodarone 400 mg BD for 2 weeks then 200 mg daily for 52 weeks (n=61, long-term amiodarone). Spontaneous reversion to SR occurred before DCCV in 21% (26/123) patients on amiodarone and none of the 38 patients on placebo (absolute difference 21%, 95% confidence interval (CI): 10 to 29%, P=0.002). At 8 weeks following DCCV, 51% (63/123) patients on amiodarone remained in SR compared to 16% (6/38) taking placebo (difference - 35% 95% CI: -48 to -18%, P<0.001). At 1 year, 49% (30/61) patients on long-term amiodarone were in SR compared to 33% (21/62) taking short-term amiodarone (difference - 15%, 95% CI: -31 to 2%, P=0.085). There was no difference in adverse event rate or quality of life scores between groups. Conclusions: Amiodarone pre-treatment before electrical DCCV for persistent AF allows chemical conversion in one-fifth of patients without altering the efficacy of subsequent DC conversion. Amiodarone is more effective than placebo in the maintenance of SR when continued for 8 weeks following successful DCCV. More patients taking long-term amiodarone remained in SR at 52 weeks, but more had serious adverse effects requiring discontinuation of therapy. Eight weeks of adjuvant therapy with amiodarone following successful DCCV may be the preferred option. © 2003 Published by Elsevier Ltd on behalf of The European Society of Cardiology.