Prophylactic Oral Amiodarone for the Prevention of Arrhythmias That Begin Early After Revascularization, Valve Replacement, or Repair: Title and subTitle BreakPAPABEAR: A Randomized Controlled Trial FREE

Atrial tachyarrhythmias, usually atrial fibrillation or atrial flutter, are facilitated by atrial trauma, atrial stretch, atrial ischemia, epicardial inflammation, hypoxia, acidosis, electrolyte disturbances, and electrophysiological changes that accompany sympathetic nervous system discharge.¹ As these factors are common immediately after cardiac surgery, atrial tachyarrhythmia is the most common postoperative complication.^{1 - 5} The incidence of sustained atrial tachyarrhythmias after coronary artery bypass graft (CABG) surgery is approximately 30%; after valve surgery, approximately 40%; and after combined CABG and valve replacement/repair surgery, approximately 50%.¹ The consequences of these atrial tachyarrhythmias include discomfort or anxiety, hemodynamic deterioration, stroke, exposure to the risks of tachyarrhythmia treatments, prolongation of hospital stay, and increased health care costs.^{2 - 10} Atrial tachyarrhythmias have recently been reported to increase hospital stay after cardiac surgery by 1.4 days at an additional cost of $6356 per patient.⁹

Accordingly, substantial effort has been invested in finding an effective, safe, and widely applicable prophylactic treatment for atrial tachyarrhythmia occurring after cardiac surgery.¹ The most extensively studied prophylactic therapy is β-blockade.¹ However, the applicability of β-blocker therapy is limited by the frequency of contraindications to its use¹¹ and its modest efficacy in contemporary cardiac surgery populations.¹² Furthermore, clinical trials have not demonstrated that prevention of postoperative atrial tachyarrhythmias with β-blocker therapy reduces hospital stay or resource use.^{12 - 13}

There have been 4 randomized trials of preoperative oral amiodarone for atrial tachyarrhythmia prophylaxis after cardiac surgery.^{14 - 17} The largest trial¹⁷ randomized only 315 patients. Two trials^{14 - 15} reported a statistically significant reduction in the incidence of postoperative atrial tachyarrhythmias; 2 trials,^{16 - 17} including the largest trial,¹⁷ did not. None were statistically powered to evaluate the efficacy of amiodarone in patient subgroups. Nevertheless, potentially important subgroup differences have been suggested. None of these trials was powered to evaluate the tolerability and safety of oral amiodarone in this setting. Indeed, each trial reported the improbable absence of significant adverse symptoms or signs in patients receiving amiodarone.

Accordingly, the purpose of the Prophylactic Amiodarone for the Prevention of Arrhythmias that Begin Early After Revascularization, Valve Replacement, or Repair (PAPABEAR) trial was to test the hypothesis that amiodarone is an effective, well-tolerated, and safe therapy for prevention of atrial tachyarrhythmias after cardiac surgery in a placebo-controlled, randomized trial in a large patient population using stratified randomization to facilitate separate analysis in younger patients (<65 years) compared with older patients (≥65 years), patients undergoing CABG surgery compared with patients undergoing cardiac valve surgery or combined CABG and valve replacement/repair surgery, and in patients who received preoperative β-blocker therapy compared with those who did not receive preoperative β-blocker therapy.

Adult patients of either sex and of any age were considered for participation when listed for nonemergent CABG surgery and/or valve replacement/repair surgery without other concomitant procedures. All patients had a baseline 24-hour ambulatory electrocardiogram. Exclusion criteria included myocardial infarction within 2 weeks, any rhythm other than sinus, New York Heart Association class IV congestive heart failure despite treatment, persistent hypotension (systolic blood pressure <80 mm Hg), requirement for antiarrhythmic drug therapy, history of sustained atrial tachyarrhythmias, treatment with amiodarone within 3 months, sinus bradycardia while awake (<50/min^-1), PR interval greater than 220 milliseconds, second- or third-degree AV block, corrected QT interval of 480 milliseconds or higher, peripheral neuropathy, aspartate aminotransferase twice the upper limit of normal, interstitial pulmonary disease, clinical hypothyroidism or hyperthyroidism, bleeding diathesis, and women of child-bearing potential. Eligible patients participated after a full-disclosure, written, informed consent process. The study was approved by the Conjoint Medical Ethics Committee of the Calgary Health Region and the University of Calgary.

The flow of patients through the study appears in Figure 1. Participants were randomized to receive amiodarone or matching placebo in a double-blind fashion in a 1:1 ratio, blocked in groups of 10, and stratified for age (<65 years vs ≥65 years), surgical procedure (CABG surgery alone vs valve replacement/repair surgery with or without CABG surgery), and preoperative treatment (with β-blockers vs without β-blockers). The random allocation sequences were computer-generated and were implemented by a hospital pharmacist who was not otherwise involved in the trial. Study personnel were not aware of the allocation sequences or of patient allocation.

Treatment with amiodarone or placebo was administered 6 days prior to surgery through 6 days after surgery (13 days). The amiodarone dosage was 10 mg/kg daily, divided into 2 doses per day. Continuous telemetry electrocardiographic monitoring began intraoperatively and was continued for the subsequent 6 days. Trough serum amiodarone and desethylamiodarone levels were determined preoperatively on the day of surgery but were not provided to study personnel until after study closure. Patients receiving digoxin had its dosage halved.

The primary outcome was an atrial tachyarrhythmia during the first 6 days after surgery that lasted for 5 minutes or longer and prompted treatment by the attending physician. Prespecified secondary outcomes included the primary outcome in each subgroup defined by the stratified randomization scheme, the ventricular response rate of atrial tachyarrhythmias that did occur, the postoperative day of an atrial tachyarrhythmia occurrence, the number of atrial tachyarrhythmia episodes, the duration of the longest episode, the tachyarrhythmia “burden” (number of hours of atrial tachyarrhythmia during the first 6 postoperative days), and length of postoperative hospital stay. Adverse events potentially attributable to amiodarone were also prospectively recorded. Because the optimal treatment for atrial tachyarrhythmia after cardiac surgery is unknown, it was not prescribed by protocol.

Analyses were based on the intention-to-treat principle in all patients as randomized. The sample size required to provide 80% power to detect an absolute difference of 10% in the incidence of postoperative atrial tachyarrhythmia from an estimated control incidence of 30%, allowing for a 5% crossover using a 2-tailed test at the P<.05 level, was 600 patients.¹⁸ The observed difference was larger than this conservative estimate and provided greater than 80% power within each of the stratified subgroups.

Patient characteristics were summarized by standard descriptive statistics and were compared between groups using paired or unpaired t tests or the Fisher exact test as appropriate. The Kaplan-Meier method¹⁹ was used to construct time-to-event curves. Adjustment for differences between groups with respect to baseline characteristics was accomplished using the Cox proportional hazards model.²⁰ The proportional hazards assumption was confirmed to be valid using log-log plots. All comparisons were 2-tailed and significance was ascribed to P<.05. Analyses were performed using STATA statistical software version 8.2 (STATA Corp, College Station, Tex).

Between February 1, 1999, and September 26, 2003, a total of 601 patients were randomized in the PAPABEAR trial. One patient (assigned to the placebo group) died after study therapy initiation but before surgery subsequent to a monitored electromechanical dissociation arrest. The remaining 600 patients had surgery; of the total group, 299 were assigned to the amiodarone group and 302 were assigned to the placebo group. Patients in the 2 groups were well matched (Table 1). The mean (SD) age of the patients was 61.6 (11.2) years; 494 (82%) were male. The mean age, sex, type of heart disease, New York Heart Association functional class, and mean systolic and diastolic blood pressure values were similar in patients who agreed to participate compared with those who declined to participate. No patients were lost to follow-up.

The mean (SD) durations of preoperative treatment with amiodarone (5.6 [1.7] days) or placebo (5.6 [1.4] days) were equivalent. The mean (SD) duration of postoperative treatment with amiodarone (5.4 [1.8] days) was slightly shorter than with placebo (5.9 [1.7] days) (P<.001). On the day of surgery, patients receiving amiodarone had mean (SD) trough serum concentrations of amiodarone of 1.62 (0.78) μmol/L and of desethylamiodarone of 0.80 (0.09) μmol/L.

Of the 137 patients experiencing postoperative atrial tachyarrhythmia, 128 (93.4%) had atrial fibrillation and 9 (6.6%) had atrial flutter. Postoperative atrial tachyarrhythmia occurred in fewer amiodarone patients (48/299; 16.1%) than placebo patients (89/302; 29.5%) (hazard ratio [HR], 0.52 [95% confidence interval {CI}, 0.34-0.69]; P<.001; Figure 2). The absolute reduction in the incidence of postoperative atrial tachyarrhythmia of 13.4% translates into a number needed to treat to prevent 1 patient from developing a postoperative atrial tachyarrhythmia of 7.5 (95% CI, 4.8-14.7).

Seven patients in the placebo group and 5 patients in the amiodarone group died before the sixth postoperative day. One patient in each group had already had an atrial tachyarrhythmia. By way of a sensitivity analysis, if the 4 incompletely monitored patients in the amiodarone group and none of the 6 incompletely monitored patients in the placebo group were considered to have had postoperative atrial tachyarrhythmia, the difference in postoperative atrial tachyarrhythmia incidence remains significantly lower in patients randomized to amiodarone compared with those randomized to placebo (HR, 0.54 [95% CI, 0.38-0.75]; P<.001).

When atrial tachyarrhythmia did occur, the ventricular response rate was slower in amiodarone patients than in placebo patients but there were no differences in the postoperative day of atrial tachyarrhythmia onset, the duration of the longest atrial tachyarrhythmia episode, the mean number of atrial tachyarrhythmia episodes per patient, or the burden of atrial tachyarrhythmia per patient (Table 2).

Three patients from the placebo group and 1 patient from the amiodarone group were readmitted to the hospital within 6 months of surgery with an atrial tachyarrhythmia diagnosis. Of these, 1 patient (randomized to placebo) had not had an atrial tachyarrhythmia prior to hospital discharge.

Amiodarone reduced the incidence of postoperative atrial tachyarrhythmia in each of the patient subgroups created by stratified randomization and, in each instance, the treatment effect was significant (Table 3).

The incidence of postoperative atrial tachyarrhythmia was higher in patients aged 65 years or older, in patients who had valve replacement/repair surgery with or without CABG surgery, and in patients not receiving preoperative β-blocking therapy (Table 3).

The HRs for postoperative atrial tachyarrhythmia in amiodarone patients were comparable in the 6 prespecified subgroups. Subgroups with higher probabilities of postoperative atrial tachyarrhythmia had larger absolute risk reductions that translate into lower numbers of patients needed to treat (Table 3).

A total of 153 patients had valve replacement/repair surgery only. In this nonstratified subgroup, the HR for amiodarone prophylaxis was comparable with that of the prespecified subgroups at 0.56 (95% CI, 0.29-1.07). However, the difference did not reach statistical significance in this smaller subgroup (P = .08).

The durations of postoperative intensive care unit stays were equivalent in the amiodarone and placebo groups. Although the total mean (SD) length of postsurgical hospital stay was shorter for amiodarone patients (8.2 [7.4] days) than for placebo patients (8.9 [8.1] days), this difference did not reach statistical significance (P = .11).

Potential adverse effects of amiodarone were assessed in 4 ways: analysis of withdrawal or reduction in the dosage of blinded therapy; analysis of nonfatal postoperative complications; analysis of in-hospital mortality; and analysis of the frequency of hospital readmission after discharge.

Blinded therapy was more often withdrawn or reduced in dosage in amiodarone patients (34/299; 11.4%) than in placebo patients (16/302; 5.3%) due to an increase in bradycardia requiring pacing, QTc interval prolongation longer than 650 milliseconds, and skin rash (P = .008; Table 4). In contrast, postoperative sustained ventricular tachyarrhythmia occurred less frequently in amiodarone patients (1/299; 0.3%) than in placebo patients (8/302; 2.6%) (P = .04). Of these events, 3 were fatal and occurred in patients randomized to placebo (P = .25). Other major nonfatal postoperative complications and causes of operative mortality were not different between the 2 treatment groups (Table 5). There was no difference between the 2 treatment groups in the incidence of readmission to the hospital for any reason within 6 months of discharge (42 [14.0%] of 299 amiodarone patients vs 37 [12.3%] of 302 placebo patients; P = .55) or in 1-year mortality rates (8 [2.7%] of 299 amiodarone patients vs 12 [4.0%] of 302 placebo patients; P = .50).

The PAPABEAR trial demonstrates that a 13-day perioperative course of oral amiodarone therapy is associated with a halving of the postoperative incidence of atrial tachyarrhythmias overall, in those patients younger than 65 years, in those aged 65 years or older, in those having CABG surgery, in those having valve replacement/repair surgery with or without CABG surgery, in those receiving concomitant β-blocker therapy, and in those not receiving concomitant β-blocker therapy. The number needed to treat to prevent 1 patient from developing postoperative atrial tachyarrhythmia was only 7.5 overall and was even lower in older patients, in patients having valve surgery, and in patients not receiving concomitant β-blocker therapy. This benefit was associated with a reduction in the incidence of sustained ventricular tachyarrhythmia. Amiodarone had few adverse effects and was not associated with serious postoperative events. These results may be generalizable to any setting performing cardiac surgery.

Atrial tachyarrhythmia after cardiac surgery is often transient and causes little morbidity. However, atrial arrhythmia after cardiac surgery may have important consequences including hemodynamic deterioration, stroke, ventricular arrhythmia, permanent pacemaker implantation, longer hospital stays, and exposure to the risks of arrhythmia treatments including anticoagulation.^{2 - 10 ,21}

Meta-analyses^{13 ,22} of treatments to prevent atrial tachyarrhythmia after cardiac surgery have suggested benefit with standard β-blockers, sotalol, intravenous magnesium, and atrial pacing but not digoxin, verapamil, or procainamide. The most robust data supports treatment with β-blockers. Nevertheless, β-blocker prophylaxis is often not prescribed.²³ Recently, amiodarone has been evaluated as an alternative prophylactic treatment.

Four previous trials^{14 - 17} evaluated preoperative oral amiodarone for atrial tachyarrhythmia prophylaxis after cardiac surgery. Two trials^{14 - 15} showed a significant advantage for amiodarone; 2 trials^{16 - 17} did not. The largest trial¹⁷ involved 315 patients and showed no benefit. None were powered to evaluate the efficacy of amiodarone in subgroups. Nevertheless, potentially important subgroup differences have been suggested. With respect to age, 2 studies^{15 ,17} have reported that the benefits of amiodarone are limited to patients older than 70 years. With respect to the surgical procedure performed, it is noteworthy that the 2 trials^{16 - 17} reporting no benefit enrolled only patients having CABG surgery while the 2 trials^{14 - 15} reporting benefit also enrolled patients having valve replacement/repair surgery. This suggestion that the benefits of amiodarone are limited to those having valve replacement/repair surgery was supported by a post hoc analysis¹⁵ that reported significant benefits in patients without coronary heart disease (who presumably had valve replacement/repair surgery) but not in patients with coronary heart disease (who presumably had CABG surgery). With respect to the effects of concomitant β-blocker therapy, each of these 4 trials reported an interaction in post hoc analyses. However, the results reported were disparate. One trial¹⁴ reported significant benefits from amiodarone in patients receiving concomitant β-blocker therapy but not in patients not receiving β-blocker therapy while another trial¹⁵ reported the reverse interaction. The third trial¹⁷ reported a nonsignificant increase in risk of atrial tachyarrhythmia in patients receiving concomitant β-blocker therapy and a nonsignificant benefit in patients not receiving concomitant β-blocker therapy, while the fourth¹⁶ reported that combined amiodarone and β-blocker therapy was prophylactic but neither amiodarone nor β-blocker therapy was when administered alone. The PAPABEAR trial demonstrates that the atrial tachyarrhythmia risk reduction with amiodarone is comparable in each of these subgroups. However, because the baseline risk of postoperative atrial tachyarrhythmia is higher in older patients, in patients having valve replacement/repair surgery, and in patients not receiving concomitant β-blocker therapy, the absolute benefit in these subgroups is larger, which facilitates their detection in small patient populations.

After initiation of oral amiodarone, its electrophysiological effects slowly accumulate. With the dosages used in the PAPABEAR trial, the latency to suppression of atrial arrhythmias is 4 to 6 days²⁴ but the maximum atrial electrophysiological effects may require 2 weeks.²⁵ The 6-day preoperative treatment period in the PAPABEAR trial was a compromise between awaiting amiodarone’s treatment effects and permitting timely surgery. Although shorter preoperative oral amiodarone treatment periods have been used, it is not clear that they are effective. The 2 trials^{14 - 15} that provided preoperative therapy for 5 days or longer reported positive results for most of their patients while the 2 trials^{16 - 17} that provided preoperative therapy for fewer than 5 days reported negative results. Compared with preoperative amiodarone loading for fewer than 5 days, the effects of preoperative amiodarone loading over 5 days or longer yielded more positive effects.²⁶ Other trials have attempted to overcome the need for preoperative loading using higher dosages of postoperative oral amiodarone²⁷ or preoperative and/or postoperative intravenous amiodarone.^{28 - 37} These trials with alternative amiodarone dosing approaches almost exclusively enrolled patients having CABG surgery (96%) and ranged in size from 77 patients²⁹ to 300 patients.³⁰ Seven trials^{27 - 30 ,35 - 37} showed a significant advantage for amiodarone therapy; 4 trials^{31 - 34} did not. Although no study has directly compared preoperative oral amiodarone with any of these alternative amiodarone approaches, their reported effect sizes are comparable with those with preoperative oral amiodarone loading lasting 5 days or longer. Nevertheless, intravenous amiodarone loading may be associated with more adverse effects. None of these trials with alternative amiodarone dosing was large enough to evaluate efficacy in important subgroups.

In the PAPABEAR trial, patients assigned to amiodarone who developed postoperative atrial tachyarrhythmia had a 20% slower mean ventricular response rate. Similar observations were made in the previous preoperative oral amiodarone trials^{14 - 17} and in 6 of the 8 alternative amiodarone dosing trials that reported this measure.^{27 - 28 ,33 - 34 ,36 - 37} In the PAPABEAR trial, when atrial tachyarrhythmia occurred, the postoperative day of atrial tachyarrhythmia onset and the atrial tachyarrhythmia burden were the same in both treatment groups. Two^{15 - 16} of the 4 previous preoperative oral amiodarone studies and 3 studies^{30 - 32} with alternative amiodarone dosing reported that the day of atrial tachyarrhythmia onset was significantly delayed by treatment with amiodarone. However, a greater number of studies^{14 ,17 ,27 - 28 ,33 ,36} did not. One¹⁵ of the 4 previous preoperative oral amiodarone studies and 2 studies^{28 ,36} with alternative amiodarone dosing reported a decrease in postoperative atrial tachyarrhythmia burden. However, a greater number of studies^{14 ,16 - 17 ,27 ,31 ,35 ,37} did not.

None of the trials^{14 - 17} of preoperative oral amiodarone was powered to evaluate the tolerability and safety of the therapy in this setting. Indeed, each trial reported the improbable absence of statistically significant adverse symptoms or signs in patients receiving amiodarone. The PAPABEAR trial demonstrates that the incidence of adverse effects leading to a reduction in dosage or discontinuation of amiodarone was 11.4% (approximately double that of placebo: 5.3%). None of the excess adverse effects of amiodarone was serious. Four^{31 - 34} of the 11 trials of alternative amiodarone dosing approaches reported statistically significant adverse effects dominated by hemodynamic instability, sinus bradycardia, and excessive QT interval prolongation.

Early experience with amiodarone included case reports of a possible interaction between long-term amiodarone therapy and cardiac surgery manifest by excessive sinus bradycardia, hemodynamic deterioration, and adult respiratory distress syndrome.^{38 - 40} Subsequently, 2 case-control studies^{41 - 42} found no such interaction. The PAPABEAR trial is the first controlled evaluation of this risk and found no evidence of an association between these adverse outcomes and short-term amiodarone therapy.

The low incidence of adverse effects from a short course of oral amiodarone makes it attractive for atrial tachyarrhythmia prophylaxis. Other advantages of amiodarone include the absence of depression of ventricular function when given orally⁴³ and a low risk (<1%) of ventricular proarrhythmia.⁴⁴ No evidence of either was seen in the PAPABEAR trial.

The PAPABEAR trial also demonstrates that perioperative oral amiodarone reduces the incidence of postoperative sustained ventricular tachyarrhythmia. This observation also was reported in another trial.¹⁵

The PAPABEAR trial noted a trend toward a reduction in postoperative hospital stay in patients receiving amiodarone. Two trials^{14 ,36} reported significant reductions in postoperative hospital stay in patients receiving amiodarone. Evidence that amiodarone reduces postoperative hospital stay also has emerged from a meta-analysis.¹³

Although the PAPABEAR trial involved twice as many participants as previous amiodarone studies, it is still limited by its power to detect adverse events that occur infrequently.

Use of oral amiodarone prophylaxis for postoperative atrial tachyarrhythmia according to the PAPABEAR algorithm requires a 6-day preoperative treatment period. In many jurisdictions, the waiting period for nonemergent cardiac surgery exceeds this time. However, for those jurisdictions that provide nonemergent cardiac surgery within 6 days of referral, the advantages of amiodarone prophylaxis must be weighed against the risks of delaying surgery. The mortality risk during the preoperative treatment period in the present study was 0.17% (1/601). To justify this risk, the benefits of amiodarone therapy must dominate. In this regard, it is noteworthy that throughout this trial, including after 1 year of follow-up, the mortality difference favored the amiodarone treatment group. Although none of these small mortality differences was statistically significant, their direction suggests that the advantages of amiodarone prophylaxis may be sufficient to delay nonemergent surgery to permit initiation of amiodarone therapy.

The PAPABEAR trial demonstrates that a 13-day perioperative course of oral amiodarone is an effective, possibly safe, well-tolerated, and widely applicable therapy for the prevention of postoperative atrial tachyarrhythmia after cardiac surgery. This benefit was associated with a reduction in the probability of perioperative sustained ventricular tachyarrhythmia and a trend toward a reduction in postoperative hospital stay.