Passman and colleagues have proposed a series of nomograms to assess the degree of risk for postoperative AF based on multiple preoperative clinical and electrocardiogram (ECG) variables.¹¹ These investigators performed a chart review of 229 consecutive patients who underwent CABG and found that independent predictors for postoperative AF included advanced age; left main or proximal right coronary artery stenoses; history of AF or heart failure; or preoperative ECG findings of a P-R interval 185 milliseconds (ms) or longer or a P-wave duration in lead V1 of 110 ms or longer. Univariate analysis indicated that in this cohort, chronic obstructive pulmonary disease (COPD), left main or proximal right coronary artery stenoses, frequent premature atrial contractions, and left atrial abnormality on ECG were not significant predictors for postoperative AF. Initial reports indicated that a minimally invasive approach to CABG (vs. a conventional sternotomy) does not lessen the incidence of postoperative AF, when corrected for disease severity.^15,¹⁶ However, more recent investigations indicate that the incidence of postoperative AF is lessened with minimally invasive techniques for CABG and valvular surgery.¹⁷^–¹⁹

CABG without the use of cardiopulmonary bypass (CPB) has been associated with a decreased incidence of postprocedural AF in situations of minimally invasive techniques, reoperative (“redo”) single-vessel revascularization, and octogenarian patients.²⁰^–²² Ascione and colleagues investigated, in a prospective, randomized trial, whether the use of CPB and cardioplegic arrest influenced the incidence of postoperative AF.¹⁰ Two hundred patients were randomized to CPB (with normothermic CPB vs. cardioplegic arrest CPB) and off-pump “beating heart” surgery. In this study, a risk factor for postoperative AF was observed to be CPB with cardioplegic arrest, not CPB in general. Other risk factors for postoperative AF described by this study included postoperative inotropic support, intubation time, chest infection, and length of hospital stay. In summary, similar to patients with postoperative AF observed after conventional CABG, patients with AF after minimally invasive techniques have a higher in-hospital morbidity rate, length of stay, and mortality rate compared with patients without AF; however, AF can occur with a lower frequency following minimally invasive cardiac surgery or after cardiac surgery without CPB or cardioplegia.

Risk stratification for postoperative atrial arrhythmias can be performed using clinical characteristics (see Box 60-1) or by laboratory methods. One example of a stratification method is based on P-wave duration, as calculated directly from the surface ECG or from signal-averaged data.²³ An investigation by Buxton and Josephson assessed P-wave duration from standard electrocardiographic leads in 99 cardiac surgical patients and found that the mean total P-wave duration in patients who developed AF or atrial flutter was significantly longer than in patients who remained in sinus rhythm (mean total P-wave duration of 160 ms and 126 ms [P = .001], respectively).²³ A significantly prolonged P-wave duration was observed to be a sensitive (83%) but not specific (43%) predictor of postoperative atrial arrhythmias. Prolonged P-wave duration in patients who develop postoperative atrial arrhythmias may be a reflection of underlying preoperative atrial disease.

Prognosis

Postoperative AF usually arises 1 to 5 days following surgery, with a peak incidence on day 2, and usually has a self-limited course.²⁴^–²⁶ More than 90% of patients with AF following cardiac surgery who have no history of atrial arrhythmias are in sinus rhythm 6 to 8 weeks following their operation. Rubin and associates followed up postcardiac surgical patients for an average of 26 months and observed no differences in cardiovascular or cerebrovascular morbidity or mortality between patients with postoperative AF and patients without AF.²⁷ However, other studies have shown an increased rate of early and late postoperative stroke in association with AF (see later section on postarrhythmia therapy).^14,^20,^28,²⁹ Postoperative atrial arrhythmias are usually considered to increase morbidity, length of stay in the intensive care unit (ICU), length of hospitalization, and medical costs. Almassi and coworkers, in a series of 3855 cardiac surgical patients, found that postoperative AF was associated with a longer ICU stay (3.6 vs. 2 days for patients without AF, P = .001), an increased rate of ICU re-admission (13% vs. 4%), a greater incidence of perioperative myocardial infarction (7.4% vs. 3.4%), more persistent congestive heart failure (4.6% vs. 1.4%), and a higher rate of re-intubation (10.6% vs. 2.5%).¹⁴ An investigation by Abreu and colleagues reported a longer hospital stay of 4.9 days because of postcardiac surgical AF, and it was calculated to increase medical costs by $10,000.¹⁷ It should be noted that, in slight distinction, a study by Kim and associates observed a shorter hospital stay (1 to 1.5 days in this study) attributed to postoperative AF.³⁰ Villereal and colleagues found that the occurrence of AF after CABG identifies a subset of patients at an increased risk for both early and long-term cardiovascular events.³¹

Etiology

Many perioperative factors have been described in the pathogenesis of postoperative AF (Box 60-2), but no definitive data are available. The pathophysiology of postoperative AF is probably related to pre-existing age-related degenerative cardiac changes in many patients, coupled with perioperative abnormalities in several electrophysiological parameters such as dispersion of atrial refractoriness, atrial conduction velocity, and atrial transmembrane potential. Nonuniform atrial conduction is greatest on postoperative days 2 and 3, and the longest atrial conduction time is greatest on day 3.³² These abnormalities coincide with the time of greatest risk for AF, which has a peak incidence on days 2 and 3.³³ AF after CABG has been associated with increased expression and heterogeneity in distribution of connexin40, an intercellular gap junction protein, unlike in patients who do not develop AF.³⁴ These changes could result in differences in resistive properties and conduction velocity among spatially adjacent regions of the atrial myocardium. Postoperative pericarditis is generally felt to be an etiology, or at least a potentiating factor, for AF. Perioperative hypokalemia has been shown to be associated with atrial arrhythmias with an odds ratio of 1.7 (even after adjusting for confounding factors) in a multicenter trial that followed up over 2400 patients through cardiac surgery.³⁵ Potential mechanisms whereby hypokalemia might alter atrial electrophysiology include increased phase 3 depolarization, increased automaticity, and decreased conduction velocity.

Box 60-2 Potential Factors Implicated in the Pathogenesis of Post–Cardiac Surgery Atrial Arrhythmias

Pericarditis

Atrial injury from surgical handling or from cannulation

Acute atrial enlargement from pressure or volume overload

Inadequate cardioprotection during bypass

Atrial infarction or ischemia

Hyperadrenergic state

Pulmonary issues (e.g., infection, atelectasis, pleural effusion, hypoxia)

Prophylactic Therapy with Pharmacologic Agents

Given the high incidence of postoperative AF, it is strongly recommended that prophylactic treatment be considered, especially in the presence of the risk factors described in Box 60-1. The magnitude of benefit from β-blockers, sotalol, amiodarone, and pacing was evaluated in a 2004 meta-analysis of 58 randomized trials that included over 8500 patients and in which placebo or routine therapy was given to controls.³⁶ Despite the significant reduction in AF, prophylactic drug therapy was associated with a nonsignificant reduction in stroke (odds ratio [OR], 0.76; 95% confidence interval [CI], 0.43 to 1.31), which may be attributed to a low rate of events (1.2% vs. 1.4%).³⁶ The use of β-blockers, in the presence or absence of digitalis, has been demonstrated to decrease AF from 40% for CABG patients and 60% for valvular surgery patients to 20% and 30%, respectively.^25,^26,³⁷ The effect of β-blockade in reducing postoperative AF, both alone or with digoxin, has been demonstrated in multiple meta-analyses.^37,³⁸ Even though a preventive strategy of β-blocker administration might save both medical resources and decrease the length of hospital stay, such benefits have not yet been demonstrated. Specifically, in the β-Blocker Length of Stay Study (BLOSS) trial, 1000 patients undergoing cardiac surgery were randomized in a double-blinded fashion to receive either metoprolol (possibly in an up-titrated dose) or placebo.¹² In this trial, patients treated with β-blockers had a decreased incidence of atrial arrhythmias, but this did not translate into decreased length of hospital stay. Although investigators have found an association between postoperative AF and cardiovascular and other morbidities, broadly accepted therapies that reduce the incidence of AF have not yet been demonstrated to have long-term benefit or cost effectiveness.³¹

Digitalis given preoperatively or postoperatively has been shown only to be possibly helpful and not to the same extent or with the same reliability as β-blockers.³⁸ Postoperative verapamil given to patients in sinus rhythm has been observed to slow the rate of AF if it occurs but not to alter the prevalence.³⁹ Other antiarrhythmic agents such as procainamide have been studied in a prophylactic role but have been associated with varying benefits in different reports.⁴⁰ No comprehensive data on the effectiveness of propafenone or flecainide are available.

DL–Sotalol has β-blocker and class III activities and may have a role in the prophylactic treatment of postoperative AF. Preliminary data indicate that oral sotalol may reduce the incidence of AF following cardiac surgery.^41,⁴² Sotalol may be more effective than metoprolol, speaking to a potential incremental benefit of the activity of sotalol class III.⁴³ Conversely, a relatively large study of 429 consecutive patients by Suttorp and associates demonstrated no dramatic difference in the benefit of low-dose sotalol versus high-dose sotalol compared with low-dose β-blockers versus high-dose β-blockers, which suggests that the potential benefit of DL-sotalol arises from its β-blocker effect.⁴⁴

The potential prophylactic role of sotalol was further examined in a recent prospective, randomized, double-blinded, placebo-controlled study of 85 post–cardiac-surgery patients by Gomes and coworkers.⁴⁵ In this group of patients, a significant reduction in postoperative AF was observed with sotalol treatment compared with either placebo or β-blocker treatment. Also, no increase in ventricular arrhythmias was detected, which suggests that the membrane effect of sotalol class III, when given in this study’s setting, was not a liability. However, this study excluded patients with heart failure or marked left ventricular dysfunction (characteristics that might predict a high risk for postoperative AF or for ventricular proarrhythmia), which suggests that although sotalol may be useful, it cannot be broadly applied.⁴⁶

Amiodarone—both oral and parenteral formulations—have been evaluated for prophylaxis against perioperative atrial arrhythmias. Daoud and colleagues, in a placebo-controlled trial, assessed the potential benefit of preoperatively administered amiodarone in 124 patients undergoing cardiac surgery.⁴⁷ Patients who received amiodarone, which had been initiated at least 7 days preoperatively, had a lower incidence of AF (25%) compared with patients who had received placebo (53%). Amiodarone administration was also associated with shorter duration of hospitalization and resultant decreased hospital costs. These data suggest a possible benefit of outpatient preoperative medication with oral amiodarone in decreasing the incidence of AF. However, despite a high rate of use of β-blockers—an approach shown to prevent at least 50% of AF in almost every trial in which it has been studied—a high incidence of atrial arrhythmias was observed in the control group, which suggests that these observations may not be broadly applicable. A more complete assessment of the effects of prophylactic amiodarone was provided by a 2005 meta-analysis of 10 trials.^2,^38,⁴⁸ Amiodarone therapy was associated with a significant reduction in the rate of AF or atrial flutter (22% vs. 35%; relative risk [RR], 0.64; 95% CI, 0.55 to 0.75). Also, significant reductions in the much less frequent complications of ventricular tachycardia (VT) or ventricular fibrillation (3.6% vs. 9.6%; RR, 0.42; 95% CI, 0.28 to 0.63) and stroke (1.5% vs. 4.%; RR, 0.39; 95% CI, 0.21 to 0.76) were seen. Only 61% of patients in these trials were treated with β-blockers. A similar magnitude of benefit was seen in the largest randomized trial of amiodarone in relation to cardiac surgery (PAPABEAR).⁴⁹ In this trial, which was published after the above-mentioned meta-analysis, 601 patients undergoing elective CABG or valve surgery were randomly assigned to oral amiodarone (10 mg/kg daily starting 6 days prior to surgery and continued until 6 days after surgery) or placebo. In the entire population, a 48% reduction was seen in perioperative atrial tachyarrhythmias. This benefit was consistent across a number of predefined subgroups, including patients 65 years of age or younger, those undergoing CABG or valve surgery, and those who were taking preoperative β-blockers, and those not taking β-blockers. Patients assigned to amiodarone who had an atrial tachyarrhythmia had a significantly lower average ventricular rate (105 vs. 131 beats/min with placebo). Patients randomly assigned to amiodarone had more adverse cardiac events compared with those taking placebo, including bradycardia requiring temporary pacing (5.7% vs. 2%) and QT prolongation (1.3% vs. 0%). Preoperative amiodarone treatment does have distinct disadvantages: the need to identify patients well in advance of their procedure, potential bradyarrhythmic hazards (especially in an outpatient setting and in older adults), and, although rare, a risk of perioperative pulmonary toxicity.⁵⁰ The latter may be caused by a potentiated risk, from amiodarone, of CPB-associated adult respiratory distress syndrome (ARDS), which has a poor prognosis.

The Amiodarone Reduction in Coronary Heart (ARCH) trial investigated, in a placebo-controlled, double-blind study of 300 patients, whether postoperative administration of intravenous (IV) amiodarone reduced the incidence of AF.⁵¹ Results showed a significant decrease in the incidence of AF in patients given amiodarone (35%) compared with those given placebo (47%), without significant risk from the active agent. However, the size of the benefit did not result in shorter hospital stay in this study. The relatively modest benefit of IV amiodarone in this report would probably have been even smaller had a greater number of patients received β-blockers.⁵¹

In any clinical situation, potential benefits have to be balanced against risks. Antiarrhythmic medications have possible proarrhythmic effects, particularly classes I and III agents in the context of structural heart disease, myocardial ischemia, and metabolic dysfunction. Caution is urged, especially with these agents and in these circumstances.

β-Blockers have been shown to be the most effective prophylactic agents and carry a lower risk relative to other antiarrhythmic agents.^2,³⁸ On the basis of these data, the authors of this chapter believe that β-blocker prophylaxis should be widely applied. Use of other drugs for prophylaxis needs to be further investigated.

Prophylactic Therapy with Pacing

The potential role of nonpharmacologic therapy in the prevention of postoperative AF has been examined in several studies. Single-site and multiple-site atrial pacing has been shown to be helpful in some cases of non-perioperative paroxysmal AF.⁵² Investigations into the potential role of single-site and multiple-site atrial pacing in the prevention of postoperative AF have shown varying benefits.

Initial reports indicated that atrial pacing might not be beneficial for postoperative AF. An investigation of 86 post-CABG patients found that atrial pacing via single-site atrial epicardial wires, at a rate of at least 80 beats/min and always above the intrinsic sinus rate (“overdrive pacing”), was not associated with a different incidence of postoperative atrial arrhythmias when compared with the absence of pacing. A recent study of 100 post-CABG patients, randomized to no atrial pacing versus atrial pacing at 10 beats/min or more above the resting heart rate, indicated that atrial pacing significantly increased atrial ectopy and did not attenuate the rate of AF occurrence.⁵³ The potential role of bi-atrial overdrive pacing in the prevention of postoperative AF has also been investigated in several studies. One prospective, randomized trial by Kurz and colleagues examined the effect of bi-atrial pacing in a group of post-CABG patients, assessing the incidence of AF and the possible proarrhythmic effects of pacing.⁵⁴ Unfortunately, after only 21 of the planned 200 patients were randomized, the study was terminated because this study’s pacing protocol was observed to promote AF, a possible consequence of undersensing of atrial signals by the epicardial pacing system leading to asynchronous atrial pacing. An investigation by Gerstenfeld and coworkers studied 61 post-CABG patients who were randomized to right atrial pacing, left and right atrial pacing, or no pacing, using epicardial wires.⁵⁵ No significant difference in the incidence of AF was observed among groups, although a trend toward less atrial arrhythmia in paced patients also receiving a β-blocker was seen.

Some studies, however, have shown a potential benefit from atrial pacing in the prevention of postoperative AF. Greenberg and associates performed an investigation of 154 patients following CABG or CABG plus aortic valve replacement.⁵⁶ Patients were randomized to no pacing, right atrial pacing, left atrial pacing, or bi-atrial pacing for 72 hours postoperatively, and efforts were made to administer β-blocker medications. Any pacing modality reduced the incidence of AF from 37.5% to 17% and the length of hospitalization from 7.8 to 6.1 days. In this study, multivariate analysis indicated that the most effective sites of pacing were the right atrial, left atrial, and bi-atrial sites, in that order. Of note, patients in this study did not have significant left ventricular dysfunction (average ejection fraction 53 ± 10%). In contrast, an investigation by Blommaert and colleagues examined the course of 96 postoperative patients who had a wide range of left ventricular function.⁵⁷ Patients were randomized to no pacing versus 24 hours of atrial pacing using a dynamic overdrive algorithm.⁵⁷ Attention was paid to the use of β-blocker medication. Pacing was associated with a lower incidence of AF (10%) compared with no pacing (27%). Multivariate analyses showed that the beneficial effect of atrial pacing was observed particularly in patients with preserved left ventricular function and older patients.

Findings regarding the potential benefit of bi-atrial pacing in the prevention of postoperative AF have been varied. In contrast to the findings of Greenberg and coworkers, Fan and colleagues observed a greater benefit with bi-atrial versus single-site atrial pacing.^56,⁵⁸ Fan and associates studied 132 postoperative patients without a history of AF and randomized them to no pacing, bi-atrial pacing, left atrial pacing, or right atrial pacing. After overdrive atrial pacing for 5 days, the incidence of AF was 41.9%, 12.5%, 36.4%, and 33.3%, respectively. Reductions in the rates of postoperative AF translated into shorter hospital stays in this study. Also, patients who remained in sinus rhythm had significant reductions in P-wave duration and variability in P-wave duration following pacing therapy.

One component of the potential benefit of postoperative pacing in the prevention of AF might be its allowance for more thorough dosing with β-blockers. Clearly, additional studies are necessary to investigate the potential role of overdrive pacing in the prevention of postoperative AF.

Other Prophylactic Agents

Angiotensin Inhibition

Although angiotensin converting enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs) have not previously been considered a specific therapy in patients with AF, a number of observations suggest their benefit in nonsurgical settings. A reduction in the incidence of postoperative AF with ACE inhibitors was also seen in a multicenter analysis of 4657 patients undergoing CABG.⁵⁹ Postoperative AF occurred significantly less often in patients who were treated preoperatively and postoperatively with ACE inhibitors compared with those who were not (20% vs. 34%; OR, 0.62). Patients who had previously been taking ACE inhibitors and were withdrawn from therapy had an increase in risk (46%; OR, 1.69).

Statins

Statins may reduce the incidence of perioperative AF. This was illustrated in the ARMYDA-3 (Antiplatelet Therapy for Reduction of Myocardial Damage During Angioplasty Study–3) trial with 200 patients who underwent CABG without a prior history of statin treatment.⁶⁰ Patients were randomly assigned to 40 mg of atorvastatin or placebo daily, starting 7 days prior to surgery. Atorvastatin significantly lowered the incidence of AF (35% vs. 57% with placebo). On the basis of the established benefits of statin therapy in patients with coronary heart disease, patients should be on a statin prior to elective CABG. The possible suppression of perioperative AF may be an added benefit, but the validity of this finding does not impact the recommendation for statin use in this setting.

Glucocorticoids

On the basis of the hypothesis that perioperative inflammation may contribute to the development of AF, glucocorticoids have been suggested as prophylactic therapy. In a multicenter trial, 241 patients undergoing CABG, valve surgery, or both were randomly assigned to 100 mg of hydrocortisone every 8 hours for 3 days after surgery or to placebo.⁶¹ During a follow-up period of 84 hours after surgery, the incidence of AF was significantly reduced with hydrocortisone therapy (30% vs. 48% with placebo). No difference in adverse events, including infections, was observed between the groups. Because of the wide range of the physiological effects of glucocorticoids, additional data on both the efficacy and safety of this approach are necessary before it can be considered for routine use.

Fish Oil

Consumption of fish that induce high plasma levels of n-3 polyunsaturated fatty acids (PUFAs) may be associated with a moderate reduction in the risk of AF.⁶² A randomized controlled trial of 160 patients assessed whether the administration of PUFAs (2 g per day), compared with placebo, would reduce the incidence of postoperative AF after CABG.⁶³ The development of AF was significantly reduced in patients receiving PUFAs (15.2% vs. 33.3%).