Atrial fibrillation (AF) is the most common sustained heart rhythm disorder and is associated with significant symptoms and health problems including an increased risk of stroke and death. Current drug therapies are often ineffective and associated with significant side effects. Abnormalities of calcium regulation in cells may lead to triggers for AF. Emerging data suggest that abnormal intracellular calcium regulation mediated through the ryanodine receptor in heart cells may contribute to AF. Recently the investigators have shown that the β-blocker carvedilol which is most commonly used to treat patients with heart failure, modifies calcium regulation mediated through the ryanodine receptor. At present this drug is not frequently used to treat AF. Therefore the investigators will conduct a randomized trial comparing carvedilol to metoprolol for prevention of paroxysmal AF. This may result in improved health and quality of life for people who suffer AF.

The study population will consist of patients with electrocardiographically documented symptomatic paroxysmal AF (≥ 2 episodes of ≥ 15 min duration within a six month period) while on stable antiarrhythmic drug therapy. Patients will be excluded if they have AF due to a reversible cause, persistent AF, significant heart failure or a life expectancy of less than one year. Eligible patients will be randomized to carvedilol or metoprolol and followed for 13 months. Randomization will be stratified based on the AF management strategy (rate or rhythm control). Carvedilol and metoprolol will be initiated over a one month drug titration period to achieve doses of 25mg bid and 50 mg bid respectively. AF occurrence will be documented using event recorders. The co-primary outcome measures are survival free from AF after one month blanking period for drug titration and the number of days in AF detected during follow-up. Secondary outcomes include event free survival to first symptomatic episode of AF, days in symptomatic AF, time between first and second episodes of AF, proportion of patients who develop persistent AF, AF Severity Scale, CCS-AF symptom score, ventricular rate during AF, proportion of patients with discontinuation of the assigned therapy and number of emergency department visits or hospitalizations for cardiovascular causes. Adverse effects and need to discontinue carvedilol or metoprolol will be documented. This study will determine whether carvedilol is safe and effective for prevention of recurrent paroxysmal AF in a general population with AF. Data will be analyzed on an intention to treat basis. Three hundred patients will be recruited over 4 years. The sample size is based on an estimated 20% reduction in event free survival from AF (power 0.90, α = 0.05). Patients will be recruited from our AF clinic population which averages 50 new referrals per month and over 2500 referrals since it's' inception in 2005.

Inclusion Criteria:

• Must be in sinus rhythm at enrollment
• ECG documented symptomatic PAF (> 2 episodes of > 15minutes duration over a 6 month period)

Exclusion Criteria:

• AF due to reversible causes
• Contraindication or previous significant adverse reaction to Beta blocker therapy
• Persistent AF
• NYHA Class II or greater CHF
• LVEF < or = 35%
• Life expectancy < 1 year
• Geographic isolation
• Unable to give informed consent