International consensus guidelines from 2001 recommend giving specific antiarrhythmic drugs according to the patient's underlying pathology. Patients with no clinical heart disease have a wide variety of options. On the other hand, patients with hypertension, heart failure, or coronary disease have a much smaller choice of drugs.

These recommendations are an important step away from empirical treatment. However, this is not a move toward effective drugs for specific conditions. Rather, this is an attempt to avoid known side effects.

One safety concern is the incidence of torsades de pointes. Physicians in the United States and elsewhere no longer favor quinidine because of the potential for increased mortality. Researchers presume the underlying mechanism is torsades de pointes. Sotalol also causes torsades de pointes, but it may be a safer drug because its effect appears to be dose related whereas the adverse effect may not be dose related for quinidine.

Recent data suggest that clinicians could use adjunctive therapies to block the adverse effect of these two drugs. In a rabbit model, use of verapamil almost completely eliminated the risk of erythromycin-induced torsades de pointes.

In an unpublished European clinical trial, researchers tried use of verapamil in conjunction with quinidine. There were no cases of documented torsades de pointes in 88 placebo-treated patients compared with a 2.3% incidence in 380 patients who received sotalol. Interestingly, there were no cases of torsades de pointes in 377 patients on quinidine and verapamil.

One obstacle in moving away from empiricism is that many trials have lumped together patients with atrial fibrillation and those with atrial flutter. This is probably not appropriate for analysis, particularly with the newer antiarrhythmic drugs, Dr. Falk said.

One example comes from a study of dofetilide for the conversion and maintenance of sinus rhythm. The most effective dose had a 30% chance of converting arrhythmia into sinus rhythm. However, when the data are broken down by specific arrhythmia, the conversion rate for atrial flutter was 60% in contrast to just over 20% for atrial fibrillation.

In the dofetilide study, probability of patients remaining in sinus rhythm dropped rapidly in the first few weeks. After that, the loss of sinus rhythm was much slower. Atrial remodeling due to atrial fibrillation may be responsible for the early and rapid decline.

Verapamil appears to maintain sinus rhythm after cardioversion. In one study, patients received either propafenone alone or with verapamil. Patients who received propafenone alone had a much greater rate for loss of sinus rhythm than those treated with propafenone and verapamil.

Another study compared the maintenance of sinus rhythm in patients receiving amiodarone, propafenone, or sotalol. The propafenone and sotalol groups had a similar rapid loss of sinus rhythm early in the study, whereas amiodarone treatment seemed to mimic the effect researchers showed in the study with verapamil plus propafenone.

It is not clear why amiodarone would be superior to other antiarrhythmic agents. However, one researcher presented an abstract at AHA 2002 suggesting that amiodarone may be effective in preventing remodeling, and hence early loss of sinus rhythm in converted patients.

A more recent approach is to examine the effect of amiodarone plus irbesartan on maintenance of sinus rhythm. A small, unblinded study showed that patients who received amiodarone had a significantly higher recurrence rate than the patients who received amiodarone plus irbesartan. Dr. Falk said it is not clear yet whether the combination does indeed work better than amiodarone alone.

For the future, researchers will have to rethink the role of antiarrhythmic therapy for atrial fibrillation. One important goal is to conduct large, prospective, randomized studies of adjunctive therapies.