Prophylaxis Against Postoperative Atrial Fibrillation: Title and subTitle BreakCurrent Progress and Future Directions

Atrial fibrillation is by far the most common complication following cardiac surgery, with an incidence consistently reported to range between 27% and 40% and with little change over the past 2 decades.^{1 - 5} Postoperative atrial fibrillation is associated with greater resource use as well as neurological, renal, and infectious complications.⁴ Patients with recurrent episodes of postoperative atrial fibrillation appear to be at the greatest risk for these adverse outcomes.⁴ Although atrial fibrillation is often less costly than other postsurgical complications, it is clear that the high rate of occurrence and recurrence places substantial financial and clinical burdens on clinicians and patients.^{2 ,6}

Many preoperative predictors of postoperative atrial fibrillation have been defined and include advanced age, a history of atrial fibrillation or chronic obstructive pulmonary disease, and valve replacement/repair surgery.^{2 ,4 - 5} Advanced age has been the most consistent of these predictors with every 10-year increase in age associated with a 75% increase in the odds of developing atrial fibrillation.⁴ Thus, on the basis of age alone, any patient older than 70 years would be considered at high risk for atrial fibrillation. As the age at presentation for cardiac surgery increases, it is likely that the burdens associated with postoperative atrial fibrillation will also increase. Development of interventions that decrease the high incidence of this dysrhythmia are therefore potentially of great value.

The randomized, placebo-controlled trial of oral amiodarone (Prophylactic Amiodarone for the Prevention of Arrhythmias that Begin Early After Revascularization, Valve Replacement, or Repair [PAPABEAR]) reported in this issue of JAMA by Mitchell and colleagues⁷ is one such intervention. In the largest trial of amiodarone reported to date, these investigators randomized 601 patients undergoing coronary artery bypass graft surgery, valve replacement/repair surgery, or both to receive 10 mg/kg of oral amiodarone daily or placebo from 6 days prior to surgery through 6 days after surgery (13 days). The 29.5% incidence of atrial tachyarrhythmias reported in the placebo group is comparable with that of other studies but significantly greater than the 16.1% incidence in patients treated with amiodarone.

Importantly, Mitchell and colleagues stratified their randomization by age, surgical procedure, and preoperative β-blocker treatment to assess the efficacy of amiodarone in subgroups that are typically considered to be at high risk for postoperative atrial fibrillation. Among patients aged 65 years or older, those having valve replacement/repair surgery combined with coronary artery bypass graft surgery and those not receiving preoperative β-blocker therapy, larger absolute reductions in the incidence of atrial tachyarrhythmias were evident with amiodarone prophylaxis. In addition, amiodarone therapy lowered the ventricular rate when atrial tachyarrhythmias did occur and decreased the incidence of fatal and nonfatal postoperative-sustained ventricular tachyarrhythmias. Moreover, the safety and long-term follow-up data presented in this study add substantially to the existing literature.

Given these findings, should physicians routinely administer prophylactic amiodarone to preoperative cardiac surgery patients? As with all pharmacological therapy, the benefits of therapy must be balanced against adverse effects. Of note, blinded therapy in the PAPABEAR trial was more likely to be withdrawn in patients treated with amiodarone largely because of a 3-fold increase in bradycardia requiring pacing and an increase in QTc interval prolongation of longer than 650 milliseconds.

An important limitation of this study is the incomplete data on the interaction between amiodarone therapy and postoperative β-blocker therapy. Despite guidelines recommending that early postoperative administration of β-blockers should be the standard therapy for atrial fibrillation,⁸ only 50% of cardiac surgical patients are typically treated with β-blockers and β-blockers are actually withdrawn in 25% of patients.⁴ Such withdrawal has been shown to increase the risk of postoperative atrial fibrillation.^{4 ,9} Although data from the PAPABEAR trial demonstrate no difference in the postoperative use of β-blockade between treatment groups, rates of β-blocker therapy withdrawal are not reported. The possibility exists that more patients were withdrawn from β-blockers in the placebo group, thus exaggerating the observed effect of amiodarone. Similarly, the importance of other concomitant medications such as angiotensin-converting enzyme inhibitors and anti-inflammatory agents in reducing postoperative atrial fibrillation has recently been highlighted⁴ ; the interaction between these agents and amiodarone also was not assessed in the PAPABEAR trial.

Effective amiodarone prophylaxis appears to require a preoperative treatment period that limits the use of this protocol to nonemergency settings. Even among the nonemergency candidates in the PAPABEAR trial, a small risk of mortality was observed in the waiting period, and this risk must be weighed against the potential benefit of a reduction in postoperative atrial fibrillation and its associated complications. Although many may consider a 6-day waiting period unnecessary, it would appear that such a policy may be driven largely by concerns over legal liability and by the excess cardiac surgical capacity currently present within the US health care system. Further large-scale study is needed to determine if surgery should be deliberately delayed for the purpose of prophylaxis against atrial fibrillation in as many patients as possible. As with the increasingly common clopidogrel-free preoperative waiting period of 5 days aimed at reducing the risk of perioperative hemorrhage, it is possible that preventing atrial fibrillation will be well worth the wait.

Future study of atrial fibrillation prophylaxis should consider 3 important methodological issues. First, when addressing the relationship between postoperative atrial fibrillation and complications, it is essential to capture the precise time of onset of both atrial fibrillation and the complications,⁴ thus allowing for a detailed examination of the temporal relationship between the two. Second, patients experiencing multiple episodes of atrial fibrillation should be targeted for therapeutic intervention because the incidence of complications is much greater in these patients than in those experiencing a single episode.⁴ Amiodarone therapy has potential benefit in this setting as well, having been associated with a lower risk of recurrence in an observational trial.⁴ Third, a majority of the known risk factors for postoperative atrial fibrillation should be recorded and accounted for in statistical analyses. The use of a risk index⁴ offers statistical advantages and may be used to define patient selection criteria or identify patients for whom prophylactic therapy might be most effective.

Additional study also is needed to define potential genetic modifiers of postoperative atrial fibrillation. Despite numerous studies suggesting a genetic basis for atrial fibrillation in the ambulatory population,^{10 - 18} genetic studies among patients with postoperative atrial fibrillation are sparse and mainly implicate polymorphisms modulating the perioperative inflammatory response.^{19 - 21} Through application of genomic technologies to the study of atrial fibrillation, both previously unanticipated mechanistic pathways^{22 - 25} and novel therapeutic modalities²⁶ are beginning to emerge. Moreover, it appears that subtle genetic variants occurring frequently in the population may dramatically alter the response to drugs that act on ion channels.²⁷ With the recent introduction of screening tests for selected mutations in Na⁺ and K⁺ channels and in drug-metabolizing enzymes, some of the important pharmacogenomic determinants of antiarrhythmic drug response can now be assessed. Even though it has yet to directly affect medical management, using integrated genomic and biological systems approaches to dissect complex postoperative arrhythmia phenotypes has the potential to improve perioperative risk characterization, choice of prophylactic agents, identification of new drug targets, and patient safety.

In the meantime, to help prevent postoperative atrial fibrillation, more widespread use of amiodarone for patients undergoing elective cardiac surgery should be considered.