11. AF and Special Populations

11.1 Device-detected AF

By convention, and on the basis of somewhat arbitrary definitions, the diagnosis of AF requires ECG documentation of an irregular rhythm with no discernible, distinct P waves, lasting at least 30 seconds. Contrary to this widely accepted threshold for AF diagnosis, the minimal duration of incessant AF that a patient should manifest before warranting OAC for stroke prevention remains a matter of debate, even in the presence of other stroke risk factors. The uncertainty relates to the few studies in which this was examined and the fact that the duration of subclinical AF associated with stroke differed widely across the studies. Specifically, AF episode durations associated with increased risk of stroke/systemic embolism ranged from 5 minutes in the Mode Selection Trial (MOST; HR, 2.79),^[99] 6 minutes in the Asymptomatic Atrial Fibrillation and Stroke Evaluation in Pacemaker Patients and the Atrial Fibrillation Reduction Atrial Pacing Trial (ASSERT; HR, 2.5),^[50] > 1 hour in Stroke Prevention Strategies on the Basis of Atrial Fibrillation Information From Implanted Devices (SOS AF; HR, 2.1),^[684] > 5.5-hour daily burden in the The Relationship Between Daily Atrial Tachyarrhythmia Burden From Implantable Device Diagnostics and Stroke Risk (TRENDS) study (HR, 2.4),^[100] and > 24 hours in a prospective observational study (HR, 3.1).^[101] It is not entirely clear why there is a discrepancy between arrhythmia durations and stroke risk, but there are several hypotheses. First, the studies used CIEDs, which specifically detect AHRE, meaning not all of these arrhythmia episodes were AF or AFL.^[685],[686] Second, the AF duration cut points across studies were not empirically derived. The 5-minute cutoff in the MOST study was chosen to avoid false positive results from oversensing^[99]; whereas the 5.5-hour threshold used in the TRENDS study corresponded to the median AHRE duration that was measured.^[100] Third, studies did not specifically adjudicate strokes as being cardioembolic and, in some cases, these were combined with TIA and systemic emboli.^[685] Fourth, the incidence rates for stroke were generally low and, notwithstanding significant HR, the absolute risk imparted by AF was often small or unreported. Fifth, differences in study populations, and specifically their stroke risk profiles, could have accounted for some of the discrepancies. Patients with higher AF burden are typically older and/or have other conditions that independently increase stroke risk.^[685],[686] Finally, it is unlikely that any influence of AF duration on stroke risk would manifest consistently across individual patients.More recent analyses have questioned the association between shorter AF episodes and stroke risk. A reanalysis of ASSERT showed that the risk of ischemic stroke/systemic embolism in patients with an AF episodes lasting between 6 minutes and 24 hours was comparable with those without subclinical AF.^[96] Conversely, episodes of AF lasting > 24 hours were associated with a greater than threefold increased risk of stroke/systemic embolism (HR, 3.2). Clear insight into the correlation between the risk of stroke and the duration of incidentally discovered AF is overdue. The issue has become increasingly important because of the rapid proliferation of wearable technologies that detect arrhythmias, some of which have been marketed specifically to detect AF.^[687] Putting aside the relative inaccuracy of many consumer devices,^[688],[689] the surge in individuals requiring more prolonged and diagnostically precise rhythm monitoring will uncover far greater numbers of people with silent AF than is presently the case. Discovery of subclinical AF will generate patient anxiety over having a condition with potential serious consequences and pressure on physicians to implement treatment. The clinical quandary remains the point at which the upfront hazard of antithrombotic therapy, in terms of major and fatal bleeding, is outweighed by the preventable stroke risk conferred by a specific amount and duration of paroxysmal AF. Addressing these gaps are the focus of 2 ongoing clinical trials, Apixaban for the Reduction of Thrombo-Embolism Due to Sub-Clinical Atrial Fibrillation (ARTESiA; NCT01938248) and Non-Vitamin K Antagonist Oral Anticoagulants in Patients With Atrial High Rate Episodes (NOAHAFNET 6; NCT02618577). In the absence of conclusive data, the CCS AF Guidelines Committee recognizes that it is reasonable to prescribe OAC for patients with AF who are aged 65 years or older or with a CHADS2 score ≥ 1 who have episodes of subclinical AF lasting ≥ 24 continuous hours. Last, rather than focus simply on the correlation between AF episode duration and stroke risk, a more useful exercise might be to incorporate AF burden into established stroke risk calculators to improve their predictive ability and better identify patients who might benefit from antithrombotic therapy. A 2009 retrospective study showed better stroke risk stratification as a result of combining AF episode duration (< 5 minutes, 5 minutes to 24 hours, and > 24 hours) with CHADS2 score (scores of 0, 1, 2, and ≥ 3).^[690] A similar study from 2019, which involved 28,032 patients improved stroke risk stratification by combining AF duration (no AF, 6 minutes to 23.5 hours, and > 23.5 hours) with CHA2DS2-VASc score (0-1, 2, 3-4, ≥ 5).^[691]

11.2 Hypertrophic cardiomyopathy

HCM is transmitted as an autosomal dominant trait and has an annual incidence of 0.3-0.5 per 100,000.^[692],[693] The prevalence of AF in subjects with HCM varies between 18% and 40%, which is four- to sixfold more frequent than that in the non-HCM population.^{[300],[694]–[696]} The development of AF in subjects with HCM might precipitate a sudden clinical deterioration and has been associated with a significant increase in morbidity and mortality.^{[298],[302],[305],[695],[696]} ND-CCBs or β-blockers may be used for rate control in subjects with HCM. These agents are negative inotropes with bradycardic effects, which might improve symptoms related to obstruction, diastolic dysfunction, and myocardial demand. Digoxin should be avoided in subjects with HCM. Amiodarone and dofetilide are the preferential pharmacological rhythm control agents, with sotalol reserved for subjects with mild hypertrophy.^[697],[698] Flecainide and propafenone should be avoided because of risk of proarrhythmia. Select subjects may be considered for catheter ablation, typically after a trial of antiarrhythmic drug therapy. The AF-free survival is lower in subjects with HCM compared with the general population with AF. The safety outcomes were comparable with those in the general population that underwent catheter ablation for AF. Subjects with longstanding persistent AF, older subjects, and subjects with evidence of advanced atrial remodelling are less likely to benefit from catheter ablation.^{[699]–[706]} In subjects with HCM and AF who undergo surgical myectomy a concomitant Cox-Maze procedure might be considered.^[707],[708] Anticoagulation management of patients with HCM are discussed in section 8.3.7.

11.3 AF and athletic participation

The relationship between AF and athletic participation is complex. In individuals who exercise < 5-7 hours per week, it is entirely unclear whether sport or exercise bears any consistent contribution to the observation of AF, which is particularly true for recreational athletes who do not train for competition. For individuals with a history of participation in endurance sports, the data are inconsistent. Retrospective, poorly controlled cohort studies of former or current endurance athletes have suggested RR ratios for AF occurrence of between 3 and 5 compared with sedentary individuals; however, higher-quality data (eg, prospective cohort studies) have observed the lowest RR.^[709] The largest prospective cohort study followed 17,000 physicians (median age 50 years) for15-19 years.^[710] The absolute risk of AF in the highest exercise group (defined as exercising 5-7 days per week) was 0.7%- 0.8% per year, equivalent to an absolute increase in risk over sedentary persons of 0.1%-0.2% per year. In contrast, several large meta-analyses have shown an inverse relationship between physical fitness and AF incidence, with the lowest rates of AF occurrence observed in the most fit group compared with the least fit group.^{[711]–[713]} Similarly, the relationship between AF and exercise duration and intensity is complex. Mozaffarian et al. reported that the individuals older than 65 years with the greatest time spent in leisure activity had the lowest risk of AF, whereas individuals with the highest intensity of exercise had no such reduction in the risk of AF.^[714] Similarly, Ricci et al. reported that men with the highest intensity of exercise (> 20 MET hours per week) did not have a reduction in risk of AF compared with sedentary individuals, whereas those with < 20 MET hours per week did show a significant reduction.^[715] Conversely, the top quintile of exercisers in the Women’s Health Study (> 23 MET hours per week) experienced a decrease in the incidence of newly documented AF (1.87 vs 2.43/1000 person-years of follow-up in sedentary individuals).^[716] A similar sex-based difference was observed in a large study of 208,654 long-distance skiers in Sweden, in which female skiers had a lower incidence of AF than female nonskiers (HR, 0.55; 95% CI, 0.48-0.64).^[717] Conversely, male skiers had an AF incidence similar to that of nonskiers (HR, 0.98; 95% CI, 0.93-1.03). Moreover, there was an increase in incident AF in men who completed ≥ 3 races (HR, 1.24; 95% CI, 1.03-1.51). Taken together, these data indicate that levels of physical exercise that are less than intense competitive training are not risk factors and might actually be protective against AF. Conversely, competitive endurance sport is a risk factor for AF in men, however the absolute risk is very low. To our knowledge, there have been no randomized or controlled trials of AF management in athletes.^[718]

11.3.1 Careful attention to standard risk factors for AF

In addition to the factors outlined in section 6, particular attention should be placed on the assessment of hypertension, including overt and masked hypertension (not present in the clinic but present during daily life, assessed using ambulatory BP monitoring); unexpected OSA (in the absence of obesity); occult valvular disease such as mitral regurgitation; and alcohol consumption.

11.3.2 Rate and rhythm management

In most athletes, AF presents as paroxysmal as opposed to persistent AF. Paroxysms often occur at rest or at nighttime and are often not present during exercise or training. In such patients, treatment should be directed at improving QOL rather than suppressing AF episodes or reducing AF burden because there is no convincing evidence that AF suppression will alter morbidity or stroke in the athletic population. Specific treatment might not be required in patients with infrequent AF episodes that occur at rest, particularly if the symptoms are mild, the episodes are self-limiting, and the spontaneous ventricular rate is < 110 bpm. Highly symptomatic patients with impaired QOL might warrant a “pill-in-the-pocket” approach if the episodes are infrequent or, alternatively, antiarrhythmic drug therapy if the episodes aremore significantly symptomatic or prolonged (as outlined in section 9.3.2). If “pill-in-the-pocket” therapy is used, then it is recommended that vigorous exercise be avoided within 6 hours of administration to avoid 1:1 conduction of AFL. For individuals who have very frequent AF episodes, continuous antiarrhythmic drug therapy may be considered. Class Ic antiarrhythmics such as flecainide or propafenone (if the individual does not have LV dysfunction or CAD, which is unusual in athletes), combined with an AV nodal blocker are typically preferred. Some authors favour CCBs as the AV nodal blocking agents of choice in light of empirical and anecdotal evidence that suggests that β-blockers are very poorly tolerated by athletes. Specifically, most studies of CCBs show no change or improved exercise tolerance in AF, whereas β-blockers lead to no change or decreased exercise tolerance despite effective rate control.^[577] Catheter ablation should be considered for individuals whose QOL is substantially impaired, particularly if the “pillin-the-pocket” strategy or continuous antiarrhythmic drug therapies are ineffective, poorly tolerated, or strongly not desired by the patient. Anecdotal evidence and small case series suggest that PVI is at least as effective in athletes as it is in sedentary individuals, particularly because risk factors for decreased efficacy such as longstanding persistent AF, obesity, OSA, and LV dysfunction are rarely present in athletes.^[719] Although a reduction in exercise intensity or duration (“detraining”) has been considered in the management of AF in athletes, most athletes are reluctant to contemplate this strategy. Moreover, there are no controlled or prospective studies to suggest that detraining can reduce the frequency, severity, or duration of AF episodes in athletes.^[720] Athletes might also frequently present with AFL as the initial presenting arrhythmia rather than AF,^[721] and AFL ablation can be safely performed as the primary treatment strategy in these individuals. One study suggests that the risk of AF after AFL ablation might be higher in athletes than in nonathletes, but AF still occurs in < 25% of patients at 1 year.^[721]

11.3.3 Stroke risk in athletes vs nonathletes with AF

There are few data to assess the relative and absolute risk of stroke in athletes with AF vs that in nonathletes. Confounding the assessment of a potential risk of stroke are the observations that athletes tend to be younger than nonathletes when they develop AF; they tend to have fewer stroke risk factors; and their AF episodes tend to be more often infrequent and paroxysmal. These observations suggest a lower risk of stroke in athletes when controlling for other stroke risk factors. The risk of stroke in cross-country skiers with AF (some of whom received anticoagulant therapy) was significantly lower than in nonskiers without AF (HR, 0.64; 95% CI, 0.60-0.69).^[717] These caveats notwithstanding, it seems prudent to manage athletes similarly to nonathletes, with the decision regarding OAC being on the basis of the “CCS algorithm.” Some athletes might be concerned about the possibility of injury leading to a higher risk of bleeding, but there are no data to support a different treatment approach in such individuals except possibly those engaged in competitive contact sports. Using the shared decision-making model, athletes should be counselled regarding stroke risk and advisability of OAC when warranted according to guidelines.

11.4 Congenital heart disease

Atrial tachyarrhythmias are highly prevalent in patients with CHD,^[722] are the leading cause of morbidity/hospitalizations, and have been linked to HF, sudden death, and stroke.^{[277],[280],[723]–[725]} Contemporary prevalence estimates for atrial arrhythmias in adults with CHD range from 10% to 15%,^[277] with > 50% of those with severe CHD projected to develop an atrial arrhythmia by 65 years of age.^[277] Whereas organized atrial macroreentrant arrhythmias are the most common supraventricular arrhythmia in CHD patients, the prevalence of AF is rising, particularly with atrial or AV septal defects, Ebstein anomaly, tetralogy of Fallot, univentricular hearts, and left-sided obstructive lesions.^[726],[727] The type and prevalence of arrhythmias depend, in part, on the subtype of CHD, nature of the surgical intervention, residual defects, and age.^[728],[729] Potential contributors include surgical incisions, natural conduction barriers (eg, valve orifices, venous structures, septal defects, crista terminalis), and sequelae of chronic hemodynamic or hypoxic stress (eg, fibrosis, hypertrophy). Defects associated with the highest prevalence of atrial arrhythmias include single ventricles with Fontan palliation, transposition of the great arteries with atrial redirection surgery (eg, Mustard or Senning baffle), tetralogy of Fallot, Ebstein anomaly, and atrial septal defects.^[730] Factors associated with IART with some consistency include older age, more complex CHD, later repair, coexisting sick sinus syndrome, and right atrial dilation.^[731],[732] In contrast, AF is predicted by factors such as older age, residual left-sided obstructive lesions, lower systemic ventricular EF, and LA dilation.^[731],[733] It is also noteworthy that acquired comorbidities associated with AF in the general population are applicable to adults with CHD, including hypertension, obesity, OSA, and male sex.^[734],[735] In a multicentre North American study of adults with CHD and atrial arrhythmias, AF accounted for 29% of presenting arrhythmias, IART for 62%, and focal atrial tachycardia for 9.5%.^[735] A strong association between AF and age was observed, with AF surpassing IART as the most common presenting arrhythmia in patients 50 years of age or older. Moreover, although the predominant pattern of AF was paroxysmal (> 60%), permanent AF (20%) more frequently occurred in older patients. The coexistence of AF with IART was frequently observed, with the most common scenarios being IART progressing toward AF, and paroxysmal forms transforming to permanent.^[735],[736]

11.4.1 Rate and rhythm management

11.4.1.1 Surgical considerations

The timing and type of surgery can have a major effect on arrhythmia outcomes.^[729] Knowledge of arrhythmic complications during follow-up has led to surgical modifications to improve outcomes in subsequent generations. Although preventive arrhythmia surgery could be considered in specific circumstances, there is currently no evidence to support a prophylactic left-sided Maze procedure in adults with CHD undergoing cardiac surgery for other indications.^[282] However, in patients with preexisting atrial arrhythmias, AF can be addressed surgically with a modified CoxMaze III procedure, albeit with scarce supportive data.^[737] A modified right atrialMaze in conjunction with an LACox-Maze III procedure has been performed in patients with biventricular hearts and AF undergoing Fontan conversion or revision. In general, these decisions are best guided by an interdisciplinary team that includes an electrophysiologist with expertise in the care of adults with CHD. Prophylactic arrhythmia surgery should not be performed if it substantially increases surgical morbidity or mortality.

11.4.1.2 Pharmacological therapy

A paucity of data exist to inform optimal pharmacological strategies for AF in adults with CHD. Conversion rates with ibutilide and sotalol for the acute termination of AF have ranged from 50% to 80%, with hypotension, bradycardia, and TdP as reported complications.^[738],[739] Successful pharmacological cardioversion has also been observed with dofetilide in 41% of patients with CHD.^[740] There exists no efficacy or safety data on acute cardioversion of AF in patients with CHD treated with class Ia, Ic, or other class III agents. Regarding long-term management, in the absence of CHD specific data, rhythm control is generally preferred to rate control as an initial treatment strategy in those with moderate or complex CHD.^[282] Rhythm control is justified on the basis that AF can be poorly tolerated in the context of conditions such as a single ventricle, systemic right ventricle, cyanosis, concomitant pulmonary hypertension, and/or significant residual hemodynamic lesions.^[282] The global clinical scenario, including coexisting bradyarrhythmia and ventricular dysfunction should be taken into consideration in the selection of pharmacological agents.^[282] Current expert recommendations discourage the use of class I antiarrhythmic agents in patients with CAD or systolic dysfunction of a subaortic or subpulmonary ventricle.^[282],[741] Observational studies have suggested that class III antiarrhythmic agents are most effective in reducing recurrences of atrial arrhythmias in patients with CHD.^[742] Amiodarone is considered a drug of choice in the context of HF. However, particular to the CHD population, amiodarone-induced thyrotoxicosis is four- to sevenfold higher in those with cyanotic heart disease and Fontan palliation.^[743],[744] Dofetilide appears to be a reasonable alternative to amiodarone in the setting of ventricular dysfunction or refractory arrhythmias.^[745] In a multicentre study, dofetilide led to better control of atrial arrhythmias in 49% of patients with CHD at 3 years of follow-up.^[740]

11.4.1.3 Catheter ablation

Small studies have recently emerged on catheter ablation for AF in adults with CHD centred on PVI. A single-centre series of 36 patients, most of whom had atrial (61%) or ventricular (17%) septal defects, included 72% with paroxysmal and 28% with persistent AF.^[746] Freedom from recurrent AF without antiarrhythmic drugs was 42% at 300 days and 27% at 4 years after a single procedure. The largest series of AF ablation in patients with CHD included 57 patients, with a mean age 51 years, of whom 61% had simple, 18% moderate, and 21% severe forms of CHD.^[747] The pattern of AF was paroxysmal in 37% and persistent in 63%. Freedom from recurrent atrial arrhythmias with or without antiarrhythmic drugs was 63% at 1 year and 22% at 5 years. Recovery of PV conduction was observed in 65% of patients who underwent a second intervention. One case series described the feasibility and safety of cryoballoon ablation in 10 patients with AF and CHD.^[748] PVI was acutely successful in all, with no major complication. One year after a single ablation procedure, 60% remained free from AF. The totality of evidence thus far suggests that catheter ablation is feasible, safe, and modestly efficacious. A more thorough understanding of underlying mechanisms and extra-PV triggers carries the potential to further improve the selection of optimal candidates and procedural outcomes.^[729]

11.5 Secondary AF or AF due to reversible precipitants

As outlined in section 1.2, the likelihood of developing AF varies across physiological and pathological states. Conceptually, AF may be considered “primary” if the AF represents an established pathophysiological process or “secondary” if the AF is caused by a self-limited or at least partially reversible precipitant.^[18] Within secondary AF, these episodes can be conceptualized as arising from a combination of several complementary elements. The first is the patient’s underlying propensity to AF due to modifiable and nonmodifiable substrates, the second is the new substrate imparted by the acute/secondary precipitant, and the third is the reversible AF trigger provoked by the secondary cause (Fig. 3).^[749] This conceptual model suggests that secondary AF will be more likely in patients with a propensity to AF (eg, more baseline substrate) and, even when the secondary cause for the disease is completely reversible these patients might still be at higher risk for future events.

11.5.1 Common causes of secondary AF

Common causes of secondary AF include surgery (cardiac and noncardiac), acute cardiac pathology (eg, MI and myopericarditis), acute pulmonary pathology (eg, COPD, pulmonary emboli, pneumonia), thyrotoxicosis, infection and sepsis, acute alcohol consumption and substance use, electrocution, and other metabolic disorders.^[18]

11.5.2 Rate and rhythm management of secondary AF

The decision to provide rate or rhythm control for secondary AF depends on the clinical context. In all cases, management of the underlying precipitant is paramount. For AF associated with acute medical illness, the risk-benefit balance must take into consideration the underlying medical condition (see section 9.1). In the case of AF associated with hyperthyroidism, the hyperadrenergic state encourages a rapid ventricular response and often necessitates β-blockade. In this case, propranolol might be the preferred agent because can effectively manage the ventricular rate, symptoms of hyperthyroidism (tremor, anxiety, palpitations), and might act to inhibit the monodeiodinase type I enzyme responsible for conversion of T4 to the more potent T3.^[750]

11.5.3 Prognosis and follow-up of secondary AF

An increased risk of adverse outcomes in association with new-onset AF secondary to MI, sepsis, or surgery has been reported.^{[751]–[753]} Increased mortality has been observed with recent-onset AF in the presence of ACS, sepsis, and COPD exacerbation.^[754] ACS-associated new-onset AF acts as an independent prognostic indicator for adverse cardiovascular events such as early reinfarction, stroke, and HF. Recent-onset AF in patients with sepsis is associated with increased risks of HF, ischemic stroke, and prolonged length of ICU stay.^{[288],[752],[755],[756]} AF in patients with severe sepsis has been associated with increased mortality and in-hospital stroke.^[752],[755] New-onset AF in the HF population has been associated with increased mortality.^[757] New-onset AF in the setting of an ACS has been associated with increased mortality, stroke, and reinfarction, resulting in a significantly higher in-hospital mortality compared with those with preexisting AF.^{[751],[758],[759]} Irrespective of the cause of secondary AF, long-term recurrence of AF is frequently observed.^{[18],[288],[756],[760]} At present, there are limited data to predict which patients will develop recurrent AF beyond an assessment of the traditional risk factors for AF outlined in section 3. As such, it is recommended that patients with secondary AF undergo careful reassessment and follow-up.

11.6 AF after cardiac and noncardiac surgery

11.6.1 Incidence of postoperative atrial tachyarrhythmias

AF commonly occurs in the perioperative setting because of the relationship between AF and atrial stretch, atrial ischemia, epicardial inflammation, hypoxia, acidosis, electrolyte disturbances, and high sympathetic tone.^{[761]–[767]} The incidence of POAF after cardiac surgery is approximately 30% for isolated CABG, approximately 40% for valve replacement or repair, and approximately 50% for combined CABG/valve procedures.^[768] The peak incidence of POAF is between postoperative days 2 and 4.^[767] Although POAF might be transient and cause little morbidity in many cases, POAF has been independently associated with increased in-hospital duration and health care costs.^[767]

11.6.2 Risk factors for postoperative atrial tachyarrhythmias

Independent factors for POAF after cardiac surgery include a preexisting history of AF, older age, male sex, a history of hypertension, the procedure performed, the number of bypass grafts, the duration of surgery, the duration of aortic cross-clamp time, a requirement for an intraoperative balloon pump, a requirement for ventilation > 24 hours, and withdrawal of β-blocker therapy.^{[761]–[767]} Of these, age has the highest predictive value.^[32]

11.6.3 Prevention of postoperative atrial tachyarrhythmias

11.6.3.1 β-Blockers

Current evidence suggests that β-blocker therapy reduces the incidence of POAF in patients who undergo cardiac and noncardiac surgery.^[769],[770] For patients who underwent cardiac surgery, use of β-blockers significantly reduced the occurrence of AF (RR, 0.50; 95% CI, 0.42-0.59; 40 studies, 5650 participants) and ventricular arrhythmias (RR, 0.40; 95% CI, 0.25-0.63; 12 studies, 2296 participants) without significant bradycardia, hypotension, or MI.^[770] In patients who underwent noncardiac surgery use of β-blockers significantly reduced the occurrence of AF (RR, 0.41; 95% CI, 0.21-0.79; 9 studies, 9080 participants) and MI (RR, 0.72; 95% CI, 0.60-0.87; 12 studies, 10,520 participants) at the expense of bradycardia (RR, 2.49; 95% CI, 1.74-3.56; 49 studies, 12,239 participants) and hypotension (RR, 1.40; 95% CI, 1.29-1.51; 49 studies, 12,304 participants).^[769] All-cause mortality at 30 days after cardiac or noncardiac surgery was not influenced by β-blocker use.^[769],[770] Relative to cardiac surgery, some trials required preoperative withdrawal of preexisting β-blocker therapy in patients randomized not to receive study β-blocker therapy (β-blocker withdrawal-mandated trials); other trials continued preexisting β-blocker therapy in patients randomized not to receive study β-blocker therapy (β-blocker withdrawal-not mandated trials). In the β-blocker withdrawal-mandated trials, study β-blocker therapy substantially reduced POAF (25 RCTs; 2600 patients; 10.5% vs 28.7%; OR, 0.30; 95% CI, 0.22-0.40; P < 0.001).^[771] In the β-blocker withdrawal-not mandated trials, study β-blocker therapy reduced POAF to a lesser extent (3 trials; 1163 patients; 31.5% vs 40.2%; OR, 0.69; 95% CI, 0.54-0.87; P ¼ 0.002). This observation might relate to preoperative β-blocker withdrawal increasing POAF in the control groups of the β-blocker withdrawal-mandated trials. Subgroup analyses did not identify any differences in outcomes according to the β-blocker used, or according to whether β-blocker therapy was initiated before, during, or immediately after surgery.^[772]

11.6.3.2 Amiodarone

Perioperative amiodarone therapy has been shown to reduce the occurrence of POAF after cardiac surgery compared with placebo or usual care (33 RCTs; 5402 patients; 19.4% vs 33.3%; OR, 0.43; 95% CI, 0.34-0.54; P < 0.001).^[773] However, a meta-analysis^[774] of head-to-head trials showed no significant difference between amiodarone and standard β-blockers for POAF prophylaxis after cardiac surgery, which might reflect the withdrawal of preexisting β-blocker therapy in the amiodarone group, which biases the comparison in favour of β-blocker prophylaxis. Results of 1 small comparative trial suggest that the combination of amiodarone and β-blocker prophylaxis is more effective than β-blockers alone.^[775] In a recent network meta-analysis^[776] the effect of timing and route of amiodarone administration on its efficacy and adverse effects were examined. Regimens that included only oral amiodarone reduced POAF similarly as did regimens that included I.V. administration. Regimens that included a least 1 day of preoperative amiodarone reduced POAF to an equivalent degree as did regimens that started amiodarone the day of or after surgery. This meta-analysis^[776] also showed that bradycardia, hypotension, or QT prolongation were less common with oral-only regimens than with I.V. regimens. In addition to reducing POAF, amiodarone has been shown to reduce postoperative hospital length of stay by approximately 1 day.^[773],[777]

11.6.4 Other apparently effective therapies less commonly used

11.6.4.1 Sotalol

A Cochrane intervention review^[773] showed that perioperative sotalol therapy reduced POAF after cardiac surgery compared with placebo or usual care (11 trials; 1609 patients; 18.1% vs 40.0%; OR, 0.34; 95% CI, 0.26-0.43; P < 0.001). Another meta-analysis^[778] showed a significantly higher efficacy of sotalol compared with standard β-blocker drugs for prevention of POAF (5 RCTs; 1043 patients; 14.3% vs 22.7%; RR, 0.64; 95% CI, 0.50-0.84; P < 0.001). It also showed similar efficacies when sotalol and amiodarone were compared.^[778] Thus, sotalol prophylaxis appears to be effective for prevention of POAF after cardiac surgery but is not considered a first-line therapy because of its risk of QTinterval prolongation and TdP.

11.6.4.2 I.V. magnesium

A meta-analysis^[779] indicated that postoperative I.V. magnesium supplementation care reduced POAF after cardiac surgery compared with placebo or standard care (21 RCTs; 3248 patients; RR, 0.69; 95% CI, 0.56-0.86; P ¼ 0.006). Although most magnesium trials reported no adverse effects of magnesium therapy, one trial^[36] showed more postoperative hypotension with a combination of I.V. magnesium and propranolol therapy than with propranolol therapy alone.^[780]

11.6.4.3 Overdrive atrial pacing

Trials have shown a significant reduction in POAF after cardiac surgery with overdrive atrial pacing (14 RCTs; 1885 patients; 17.7% vs 35.3%; OR, 0.60; 95% CI, 0.47-0.77; P < 0.001), and in particular with biatrial pacing.^[771] However, the Atrial Fibrillation Suppression Trial II (AFIST-II) showed that amiodarone is better at prevention of POAF than overdrive septal atrial pacing.^[781]

11.6.4.4 Colchicine

Colchicine has been shown to reduce POAF (5 RCTs; 1412 patients; 18.1% vs 26.8%; RR, 0.69; 95% CI, 0.57- 0.84; P ¼ 0.0002) and hospital stay after cardiac surgery (3 RCTs; 692 patients; mean difference, 1.2 days; 95% CI, 1.9 to 0.4 days; P ¼ 0.002) compared with placebo or standard care.^[782] However, these potential benefits were counterbalanced by an increased incidence of drug-related adverse effects (4 RCTs; 1196 patients; 21.0% vs 8.2%; RR, 2.52; 95% CI, 1.62-3.93; P < 0.0001), which was dominated by adverse GI side effects.

11.6.4.5 Posterior pericardial drainage

Drainage of the posterior pericardial space at the time of cardiac surgery with a posterior pericardiotomy and/or a posterior pericardial drainage tube reduces early and late pericardial effusion and cardiac tamponade, POAF (17 RCTs; 3245 patients; 12.6% vs 24.8%; OR, 0.42; 95% CI, 0.29-0.59; P < 0.001), hospital length of stay (13 RCTs; 2068 patients; mean difference, 0.82 days; 95% CI, 1.12 to 0.51 days; P ¼ 0.005), and death or cardiac arrest (10 RCTs; 2141 patients; 1.1% vs 2.3%; OR, 0.49; 95% CI, 0.25-0.94; P < 0.001).^[783] A large RCT, Posterior Left Pericardiotomy for Prevention of Atrial Fibrillation After Cardiac Surgery (PALACS; NCT02875405), is under way.^[784]

11.6.4.6 HMG-CoA reductase inhibitors

A meta-analysis^[785] of RCTs of an HMG-CoA reductase inhibitor vs placebo or standard care showed that perioperative statin therapy reduces POAF after cardiac surgery (16 RCTs; 3985 patients; 20.3% vs 23.7%; OR, 0.50; 95% CI, 0.34-0.73; P ¼ 0.0004) and shortens hospital stay (12 RCTs; 1185 patients; mean difference, -0.43 days; 95% CI, -0.70 to 0.15; P ¼ 0.002). However, no differences were observed with respect to stroke or short-term mortality.^[785]

11.6.4.7 Antioxidants

A meta-analysis^[786] of RCTs of vitamin C treatment vs placebo or standard care for prevention of POAF after cardiac surgery showed that vitamin C reduces POAF (10 RCTs; 1956 patients; 26.2% vs 36.4%; RR, 0.68; 95% CI, 0.54-0.87; P ¼ 0.002) and reduces hospital length of stay (9 RCTs; 1158 patients; mean difference, -0.95 days; 95% CI, -1.64 to -0.26 days; P ¼ 0.007). A meta-analysis^[787] of RCTs of N-acetylcysteine vs placebo or standard care for prevention of POAF showed that Nacetylcysteine reduces POAF (10 RCTs; 1026 patients; 20.5% vs 28.8%; OR, 0.56; 95% CI, 0.40-0.77; P ¼ 0.0005) without a reduction in hospital length of stay but with a possible reduction in postoperative mortality.

11.6.4.8 Other potential therapies

Class Ia (disopyramide, procainamide, quinidine) and class Ic (propafenone, flecainide) antiarrhythmic agents might prevent POAF^[788],[789] but are generally avoided in cardiac surgical patients because of concerns of proarrhythmia. Small RCTs have supported the use of dofetilide (1 RCT; 133 patients; 17.9% vs 36.4%; OR, 0.32; 95% CI, 0.16-0.83; P ¼ 0.017) or ranolazine (1 RCT; 102 patients; 8.8% vs 30.8%; RR, 0.29; 95% CI, 0.09-0.89; P < 0.001).^[790] Perioperative steroid use, renin-angiotensin system inhibitors, digoxin prophylaxis, calcium antagonist prophylaxis, n-3 polyunsaturated fatty acids, glucose-insulinpotassium therapy, vasopressin, and NSAIDs have been shown to lack promise.^{[771],[791]–[796]} Intraoperative injection of botulinum toxin into epicardial neural ganglia has been shown to be promising in small RCTs with a larger study in progress.^[797]

11.6.5 Treatment of postoperative atrial tachyarrhythmias

The management of POAF centres on the identification of potential contributors (pulmonary thromboembolism, volume overload, congestive HF, sepsis), prevention of thromboembolic events, slowing of the ventricular response rate, and consideration of sinus rhythm restoration. The natural history of POAF is dominated by self terminating but frequently recurrent AF episodes, and resolution in 6-12 weeks regardless of the therapy used.^{[798]–[800]} Furthermore, the hyperadrenergic state lessens the effectiveness of therapies that do not include β-blockade. There is an association between POAF, thromboembolic cerebrovascular events,^{[766],[767],[801],[802]} and cognitive impairment.^[803] Nevertheless, early anticoagulation might predispose the postoperative cardiac surgery patient to delayed pericardial bleeding and tamponade.^[804] In recognition of this risk and in the absence of controlled trials, the initiation of anticoagulation is generally not recommended in the first 72 hours after cardiac surgery. When initiated, anticoagulation is usually continued for > 6 weeks. Therapy for ventricular rate control is usually required for patients who experience POAF. Nevertheless, cardiac surgery might also predispose patients to bradyarrhythmia after conversion of AF. Accordingly, the availability of back-up ventricular pacing is important. Because the postsurgical state includes adrenergic discharge, β-blocker therapy is often effective. Alternatives to β-blocker therapy include ND-CCBs or amiodarone because digoxin is usually ineffective. The advantages and disadvantages of sinus rhythm restoration are similar to those in other settings. In the POAF population specifically, a recent RCT failed to show a significant difference between a rate control vs a rhythm control approach in the outcomes of total number of days in hospital, length of the index hospitalization, readmission rate, time to sustained sinus rhythm, or adverse events including the need for pacemaker.^[805] Accordingly, either the rate control or the rhythm control approach to the treatment of POAF seems appropriate. If rhythm control is pursued, antiarrhythmic drug therapy is preferred to isolated DCCV because early recurrence of POAF is common. Regardless of the approach chosen, therapy provided for POAF can usually be withdrawn 6-12 weeks later.

11.6.6 Prognosis of postoperative atrial tachyarrhythmias

Because patients with POAF are older and have more comorbidities than patients without POAF, the challenge has been to determine if POAF is an independent predictor of long-term adverse events. In one of the first studies^[806] to address these issues, the outcomes of consecutive survivors of CABG surgery from a single centre in Sweden who did not have a preoperative AF history or a pacemaker were retrospectively analyzed. Of these 571 patients, 165 (28.9%) experienced POAF. After a median follow-up of more than 6 years, patients with POAF had a higher probability of recurrent AF (25.4% vs 3.6%; adjusted OR, 8.31; 95% CI, 4.20-16.43; P < 0.001), a higher probability of late all-cause mortality (29.7% vs 14.8%; adjusted HR, 1.57; 95% CI, 1.05-2.34; P ¼ 0.027), and a higher probability of death from cerebral ischemia (4.2% vs 0.2%; P < 0.001) compared with patients without POAF. A very recent analysis^[807] took this observational comparison further with a larger sample size, prospectively collected data, longer follow-up, and a matched cohort evaluation. This report considered 7145 consecutive survivors of CABG surgery who did not have a preoperative AF history or a pacemaker from a single centre in Sweden. Of these 7145 patients, 2183 (30.6%) experienced POAF. Patients in the POAF group and in the non-POAF group were matched (on age, sex, and county) to 5 control patients who did not undergo cardiac surgery. The cumulative incidence of late AF (at 10 years of follow-up) was higher in patients with POAF compared with those without POAF (15.3% vs 4.9%; adjusted HR, 3.20; 95% CI, 2.73-3.76; P < 0.001) or matched, nonsurgical controls (15.3% vs 6.4%; P < 0.001). There was no difference in the incidence of late AF in patients without POAF compared with their matched, nonsurgical control participants (4.9% vs 5.7%; P ¼ 0.043). Patients with POAF had higher probabilities of late ischemic stroke, HF, cardiac mortality, cerebrovascular mortality, and all-cause mortality, compared with patients without POAF. Most other analyses of this sort have shown similar findings.^{[753],[808],[809]} Accordingly, POAF after cardiac surgery is associated with long-term risk of future AF (15%-25% over 10 years, which is higher than in patients without POAF and higher than in the general population). As such, this patient group should be followed more closely for the ultimate development of AF, particularly because of the observation that approximately half of the AF episodes in this setting are asymptomatic.^[810]

11.7 AF and supraventricular arrhythmia

AF might occur in conjunction with SVT. Previous cohort studies have observed AF in association with SVT in almost 20% of patients, with a gradient dependent on the underlying arrhythmia (58% with AFL, 27% with atrial tachycardia, 14% with AV reentrant tachycardia [AVRT], and 10% with AV nodal reentrant tachycardia [AVNRT]).^[811] Conversely, studies of AF patients who underwent electrophysiology studies before catheter ablation of AF have shown inducible SVT, in 4.3%-10.0%, including AFL (3.7%), typical AVNRT (1.7%), AVRT (1.2%), and atrial tachycardia (1%).^{[812]–[814]} In those with inducible SVT, there was a transition from SVT into AF during electrophysiology study in almost one-half of the patients.⁸¹⁴ AF patients with inducible SVT tend to be younger, more likely to have paroxysmal AF, less likely to have structural heart disease, and are more likely to be free of AF after a single ablation procedure (88% were AF-free after SVT ablation alone or SVT and AF ablation).^[812],[814] Enriched populations of younger AF patients, those free of structural heart disease, and those with a history of regular palpitations or narrow QRS tachycardia have been shown to have much higher rates of inducible SVT, with one series documenting AVNRT in 18% and AVRT in 9% of AF patients younger than 65 years,^[815] and another series inducing SVT in up to 57% of patients younger than 35 years of age without medical comorbidities.^[816]

11.7.1 Role of the electrophysiology study

Because of the variable rates of inducible SVT in patients with AF, an electrophysiology study should be primarily considered in AF patients who are younger, without medical comorbidities, have a history of regular palpitations or narrow complex tachycardia, and/or are planned to undergo catheter ablation of AF. It is reasonable to perform catheter ablation of the SVT alone. In younger patients, successful ablation of SVT can treat the underlying cause and might be associated with lower rates of AF recurrence.^[812],[814] Older patients generally have higher rates of AF occurring after SVT ablation (up to 20%-30% after several years) and may benefit from combined SVT and AF ablation.^{[69],[817],[818]}

11.8 Wolff-Parkinson-White syndrome

11.8.1 Acute management of preexcited AF

AF can occur in up to 50% of young patients with WolfParkinson-White (WPW) syndrome.819 Rapid conduction of AF from the atrium to the ventricle via the accessory pathway is rare but can result in ventricular fibrillation and sudden cardiac death.^[820],[821] In patients who present with preexcited AF, we recommend urgent synchronized cardioversion in the presence of hemodynamic instability^[822],[823] or when pharmacological cardioversion of hemodynamically stable preexcited AF with I.V. ibutilide or procainamide is ineffective.^[824],[825] The use of I.V. b-blockers, diltiazem, verapamil, digoxin, and amiodarone are not recommended for acute management of preexcited AF because of the potential to precipitate a life-threatening arrhythmia.^{[826]–[829]} Conduction over the accessory pathway is enhanced with use of drugs that block the AV node (β-blockers, diltiazem, verapamil, or amiodarone) and they might cause hypotension and a catecholamine release, resulting in an increase in the ventricular rate. Digoxin shortens the refractory period of the accessory pathway, further enhancing antegrade conduction of AF.

11.8.2 Chronic management of preexcited AF

The treatment of choice for patients with preexcited AF is catheter ablation. Ablation of an accessory pathway has a very high acute success rate (> 90%) with a low rate of complications (overall 3%; death 0.1%-0.3%; stroke 0.1%-0.2%; tamponade 0.4%-0.6%; AV block requiring pacemaker 0.3%-1.0%).^[649],[830] Certain high-risk features that increase the risk of ventricular fibrillation include: an accessory pathway with a short refractory period, arrhythmia-associated syncope, the presence of multiple pathways, and orthodromic AV reciprocating tachycardia that degenerates into preexcited AF. When present, these factors should prompt urgent catheter ablation.^{[821],[831]–[833]}

11.9 Screening for AF in patients with a history of atrial dysrhythmia

Patients with a history of atrial dysrhythmia (AFL, atrial tachycardia, AVRT, and AVNRT) should be followed for the development of AF.