Perioperative atrial fibrillation

Learning objectives.

By reading this article, you should be able to:

• •Discuss the prevalence of perioperative atrial fibrillation (AF) and the consequences that contribute to increased morbidity and mortality.
• •Explain the pathogenesis of perioperative AF.
• •Distinguish risk factors based on the perioperative phase.
• •Identify high-risk patients or surgical settings.
• •Summarise different strategies for prevention and treatment based on the clinical scenario, and be able to tailor perioperative care based on risk.

Key points.

• •Atrial fibrillation is the most common cardiac arrhythmia in the perioperative period and increases morbidity and mortality.
• •Understanding the pathogenesis can help the anaesthetist to reduce the risk of perioperative AF.
• •It is important to identify high-risk patients or surgical settings to reduce the risk of postoperative AF.
• •Risk factors depend on the perioperative phase and relate to the patient, surgery and anaesthesia.
• •Treatment for perioperative AF depends on the clinical condition of the patient and should be in consultation with the multidisciplinary team.

Atrial fibrillation (AF) is the most common serious cardiac arrhythmia in the perioperative period. The prevalence of postoperative AF (POAF) varies, depending on the type of surgery. Postoperative atrial fibrillation is defined as AF that occurs after any surgery, typically developing within 7 days, with the peak onset typically around 48 h. Although a vast majority of cases occur within the first week after surgery, with up to 70% occurring during the initial 4 days, late onset POAF (up to 30 days) is present in 4% of patients undergoing cardiac surgery after discharge from hospital. The incidence of POAF after non-cardiac surgery is 3–8% in adults aged ≥45 yrs, compared with rates of up to 30% and 80% after non-cardiac thoracic surgery and cardiac surgery, respectively (the highest risk is with combined coronary artery bypass [CABG] and valvular surgery). Postoperative AF is the most common cardiac dysrhythmia after CABG, and occurs in 10–40% of patients. Previous work has shown that the periods of highest risk for POAF after cardiac surgery are both immediately after surgery and at 48 h after surgery.

Postoperative atrial fibrillation is correlated with poorer outcomes. including increased morbidity and mortality, increased length of stay in hospital and the intensive care unit (ICU), a higher readmission rate, and therefore, increased total healthcare costs. Complications associated with POAF include thromboembolic events (including stroke), and perioperative cardiac complications, such as myocardial infarction (MI) and congestive heart failure (CHF).¹

In this review we discuss its pathophysiology, risk factors, prevention and management of POAF.

Pathophysiology and risk factors

The pathophysiology of POAF after non-cardiac surgery is not fully understood and is likely to result from a combination of factors. Increased sympathetic nervous system activation from surgical stress or pain, with release of catecholamines, is a significant mechanism. Other contributory factors may include direct surgical irritation or injury, hyper- or hypovolaemia, intraoperative hypotension, anaemia, hypoxaemia, hypercarbia and metabolic imbalances. Hypoxaemia acts via interrelated mechanisms to induce POAF. Hypoxaemia may lead to direct cellular ischaemia, altering atrial conduction velocities, refractoriness and increasing anisotropy. In addition, pulmonary vasoconstriction caused by hypoxaemia, or alternatively hypercarbia, leads to increased right heart pressures and atrial dilation which contributes to altered conduction through atrial tissue and dispersion of electrical activity.^1,2

After cardiac surgery, in addition to the above factors, POAF has been correlated with inflammatory changes occurring in cardiac tissues. Patients with increased inflammatory markers including C-reactive protein, white blood cells and interleukins have a higher incidence of POAF (Fig. 1).

Fig 1.

Pathogenesis of perioperative atrial fibrillation. There are multiple inciting events during the perioperative period that could potentially lead to the development of atrial fibrillation.

Risk factors predisposing to the development of POAF may be categorised by perioperative phase. A meta-analysis by Seo and colleagues identified several preoperative patient-related factors that were associated with POAF after cardiac, and more specifically CABG surgery. These risk factor included advanced age, renal dysfunction, a depressed left ventricular ejection fraction (LVEF) and hypertension.³ Several studies have validated advanced age as a risk factor for POAF independently, regardless of the type of surgery. Advanced age was also found to be a risk factor in a prospective observational study of patients undergoing non-cardiothoracic surgery, with the average age for new-onset atrial arrhythmias (including AF or flutter, paroxysmal supraventricular tachycardia, or multifocal atrial tachycardia) being 67 yrs.⁴ The relationship between hypertension and AF is thought to result from higher left atrial pressure, associated with left ventricular diastolic dysfunction and subsequent atrial remodelling and dilation. Similarly, mitral valve disease increases left atrial size, which causes adverse effects on electrical conduction, such as dispersion of activation.⁵ Male sex has been shown to increase the risk of development of POAF after both cardiac and non-cardiac surgery.^1,6 Diabetes mellitus (DM) is also an important preoperative risk factor in both cardiac and non-cardiac surgical patients. As mentioned, increased inflammatory markers have been correlated with the development of POAF, and inflammation associated with DM is likely to play a role in the increased risk for development of POAF. A multivariate analysis showed DM to be independently associated with AF, with an odds ratio (OR) of 1.46.⁷ Higher body mass index (BMI) has been identified as a risk factor for new-onset AF after cardiac surgery. Pre-existing cardiac disease of any type, specifically coronary artery disease, cardiomyopathy or CHF, valvular heart disease and disease of the pericardium, also increases the risk of POAF, regardless of the type of surgery. Elevated model for end-stage liver disease (MELD) scores and fulminant hepatic failure were independent predictors of AF after liver transplantation.^4,8 In a retrospective cohort study, Wong and colleagues found that obstructive sleep apnoea increased the risk of AF after cardiac surgery.⁹ Atrial fibrillation is also the most common arrhythmia encountered in patients in the ICU. Triggers of arrhythmias typically seen during critical illness include systemic inflammation, sepsis, hypovolaemia, hypoxaemia and electrolyte imbalances.¹⁰ Moreover, disease severity and the therapeutic use of catecholamines have been linked to the development of a new-onset AF in the ICU. Adrenaline (epinephrine) and dopamine have the highest potential for triggering new-onset AF because of their significant chronotropic effects, which can augment atrial ectopic discharges.¹¹ Furthermore, increased left atrial size on bedside echocardiography is associated with increased incidence of a new-onset AF in critically ill patients.¹² Increased length of ICU stay has been associated with development of POAF.

Intraoperative factors are important in determining the risk of POAF. In cardiac surgery, the type and extent of the surgical procedure can have a substantial impact on the risk for POAF. Isolated CABG carries a risk of 15–40% while isolated valve surgery carries a risk ranging from 37% to 50%.¹³ Mitral valve surgery by itself should be considered a major risk factor and carries an incidence of POAF of ∼50%.⁵ Combined CABG and valve procedures carry a risk as high as 70%.¹³ The use of cardiopulmonary bypass (CPB) during cardiac surgery has been shown to increase the risk of POAF. Contributory factors include the duration of myocardial ischaemia (which may involve inadequate cardioprotection during the bypass period); adverse effects of cardioplegia; iatrogenic atrial injury from cannulation; the inflammatory response; and activation of the sympathetic nervous system.

In non-cardiothoracic surgery, abdominal surgeries have been found to have a greater incidence of POAF compared with other surgical procedures. Kazaure and colleagues found 12.5% of patients aged >65 yrs and 25% aged >85 yrs who underwent abdominal surgery developed AF afterwards.⁶ In another review of POAF, specifically after non-cardiothoracic surgery, 93.2% of patients who developed perioperative AF had undergone some type of abdominal or general surgery.¹ Other procedures with increased risk were transplant (4.8%), vascular (1.2%), orthopaedic (0.51%) and head and neck surgery (0.24%). Turaga and colleagues found laparoscopic abdominal surgery to have a lower incidence of AF compared with open surgery.¹⁴

Regardless of the type of surgery, withholding or failing to restart beta-blockers in a timely manner can increase the risk of developing POAF in the immediate postoperative period. In contrast, continuing beta-blocker therapy throughout the perioperative period or initiating them after cardiac surgery, reduces the POAF by 51–68%. Similarly, Mathew and colleagues found that angiotensin converting enzyme inhibitors (ACEIs) started before or after cardiac surgery were also associated with reduced risk of POAF, while withdrawal increased the risk.² However, in a recent 2019 systematic review, the authors found the risk reduction of ACEI therapy to be insignificant in reducing the sequelae of AF including stroke, hospital length of stay, or perioperative mortality.¹⁵ Another study stated that it is best to avoid ACEIs in patients at risk of vasoplegic syndrome.¹⁶ Hypomagnesaemia, while known to cause supraventricular tachycardia after cardiac surgery, is controversial as a potential cause of POAF; the most recent European Society of Cardiology guidelines do not support the prophylactic use of magnesium to prevent AF.¹⁷ A recent meta-analysis of 20 RCTs including >2400 participants demonstrated that magnesium supplementation was effective in reducing POAF after CABG surgery¹⁸ (Table 1).

Table 1.

Risk factors for the development of perioperative atrial fibrillation. The table lists the multiple risk factors which play a role in the development of perioperative atrial fibrillation categorised by perioperative phase.

Risk factors for perioperative atrial fibrillation
Preoperative	•Advanced age •Male •High BMI •Hypertension •Diabetes mellitus •Obstructive sleep apnoea • Pre-existing cardiac disease o Low LVEF • Hepatic dysfunction o Elevated MELD score •Renal dysfunction
Intraoperative	• Type of surgery o Cardiac • Isolated CABG • Isolated valve • Mitral highest risk • Combined CABG + valve o Open abdominal •Use of cardiopulmonary bypass •Longer aortic cross-clamp time
Postoperative	•Discontinuation of previous beta-blocker therapy •Discontinuation of previous ACEI therapy • Electrolyte imbalances o Hypomagnesaemia •Critical illness •Prolonged ICU length of stay

Identification and risk prediction

More accurate prediction of individual risk is paramount to prevention and guided treatment because of the increased morbidity and mortality associated with POAF. Several scores such as the CHA₂DS₂-VASc, POAF score, HATCH, and COM-AF have been developed to help identify patients at risk, though there is no consensus on an accepted risk model to predict POAF.¹⁹ The CHA₂DS₂-VASc, which was designed to assign risk for thromboembolic complications in patients with AF, has been prospectively and retrospectively validated for also predicting AF after cardiac surgery. Segar and colleagues, in their retrospective single-centre analysis of several predictive score models, found that the POAF score performed best.²⁰ However, Burgos and colleagues recently developed the COM-AF score as a combination risk assessment model incorporating several highly predictive value variables from the POAF, CHA₂DS₂-VASc and HATCH models, and found it to have the best discrimination ability compared with the individual models.¹⁹

Future possibilities for risk identification and prediction include machine learning (ML) models. The sensitivity to multicollinearity of independent variables in classical regression methods make logistic regression analysis difficult to analyse large numbers of variables. A recent retrospective study examined two ML approaches (gradient-boosting decision tree and support vector machine) and logistic regression to build predictive models for AF after cardiac surgery, finding the support vector machine to have the best predictive efficiency, and highest accuracy, specificity and positive predictive value (PPV).²¹ All three models achieved appropriate net benefit by decision curve analysis, demonstrating their potential clinical use. Machine learning is further supported in a 2023 literature review by El-Sherbini and colleagues in which they looked at studies of >10 separate models. They concluded ML can aid in predicting the development of POAF after cardiac surgery and may offer an advantage over conventional risk scores because of their ability to analyse different correlations and their potential to incorporate several demographic and clinical variables.²²

Prevention

Maintaining homeostasis during the perioperative period is essential for preventing arrhythmias in general. Euvolaemia and electrolyte balance should always be the goal. As such, any electrolyte abnormalities should be corrected promptly, regardless of perioperative phase. Potassium and magnesium supplementation are especially critical in lowering the risk of perioperative arrhythmias including AF.

Beta-blocker therapy, most commonly with medications such as metoprolol or carvedilol, is a mainstay of the prevention of POAF. The Society of Cardiovascular Anesthesiologists and European Association of Cardiothoracic Anaesthetists both have a class I recommendation for continuation through the perioperative period, and early postoperative prophylactic initiation for cardiac surgical patients who are not already taking them before surgery.^23,24 Unfortunately, data supporting new preoperative or intraoperative initiation of beta-blockers are lacking.

Amiodarone is commonly used to prevent POAF. Baeza-Herrera and colleagues showed amiodarone reduced the incidence of AF after cardiac surgery by 57% when compared with placebo, with a number needed to treat of 7.4.⁵ Amiodarone is typically reserved as a second-line drug when beta-blockers are contraindicated, and has a class IIa recommendation in patients at increased risk.²³ Previous comparisons of amiodarone with beta-blockers in preventing POAF showed no difference in incidence, efficacy and safety, but, while amiodarone has a high success rate of maintenance of sinus rhythm, it (like sotalol) is hindered by significant adverse effects. These include significant incidences of bradycardia and hypotension, and pulmonary toxicity, thyroid and liver dysfunction with more chronic use.²³

Non-dihydropyridine calcium channel blockers are another preventative medication and are a recommended alternative if beta-blockers are not tolerated or are contraindicated.¹ An older meta-analysis found non-dihydropyridine calcium channel blockers were associated with a decrease in supraventricular arrhythmias (OR 0.62; 95% confidence interval [CI] 0.41–0.93).²⁵ However, their use is limited by an association with increased atrioventricular block and low output syndrome.

The choice of vasopressor drug may influence the development of POAF. Specifically, in patients with vasoplegic shock after cardiac surgery, infusions of noradrenaline (norepinephrine) and vasopressin were compared in a double-blind RCT.²⁶ The incidence of AF was lower, with a significant decrease in patients needing renal replacement therapy for acute kidney injury in the vasopressin group. A recent follow-up meta-analysis confirmed these findings.²⁷

Several other preventative strategies have been investigated regarding the prevention of POAF with unclear results. In a small randomised controlled trial, atorvastatin, started 7 days before cardiac surgery requiring CPB, was associated with a relative risk reduction of 39%.²⁸ In a Cochrane review by Arsenault and colleagues, atrial pacing and posterior pericardiotomy may be useful in preventing POAF if there are contraindications for beta-blockers, amiodarone, or both.²⁹ Given that increased systemic inflammation has been implicated in the pathogenesis and risk of POAF, non-steroidal anti-inflammatory drugs (NSAIDs) have been shown to decrease the odds of AF.² Colchicine has been studied for this use in the COPPS trial and showed a reduction in the occurrence of POAF compared with placebo (12% vs 22%, relative risk reduction = 45%; 95% CI 34–94%).³⁰ Similarly, a large meta-analysis of 31 RCTs of >1900 patients found a significant reduction in the risk of POAF when steroids are used during cardiac surgery, and this has a class IIb recommendation in European guidelines.³¹ More recently, the COP-AF trial found colchicine did not considerably reduce the incidence of clinically important AF in patients having major non-cardiac thoracic surgery.³²

Research on various intraoperative and postoperative anaesthesia-specific strategies to prevent POAF has yielded interesting results. Liu and colleagues investigated using dexmedetomidine compared with propofol for sedation and its role in reducing the incidence of new-onset AF after cardiac surgery, and found a statistically significant difference (13.6% in the dexmedetomidine group compared with 36.4% in the propofol group).³³ This also resulted in a lower median ICU length of stay but was associated with an increased incidence of hypotension.^33,34 Possible mechanisms for this preventative effect include reducing the inflammatory response, lower sympathetic tone and an alteration in calcium currents across the cardiomyocyte which increases the effective refractory period.^33,34

Despite theoretical benefits, there are limited data to support the routine use of regional anaesthesia to reduce sympathetic tone and reduce the incidence of AF. Thoracic epidural anaesthesia has historically been the technique described and may play a role in non-cardiac surgery, but the need for systemic heparinisation limits its use in cardiac surgery. There are limited data on newer regional anaesthetic blocks in this setting²⁴ (Fig. 2).

Fig 2.

Prevention and risk reduction strategies. The figure shows different strategies to reduce the risk and attempt to prevent development of perioperative atrial fibrillation categorised by perioperative phase.

Prognosis and management

Historical data have shown that POAF carries a poor prognosis, with increased risks of complications, morbidity and mortality. The POISE trial of >8000 patients undergoing non-cardiac surgery randomised to receive beta-blockers or placebo for the reduction of perioperative cardiovascular events, found that the risk of postoperative stroke within 30 days was greater in patients where new-onset, clinically significant AF developed. This finding persisted after adjusting for other risk factors, with an OR of 3.51, 95% CI 1.45–8.52).³⁵ Furthermore, a larger cohort study, examining longer-term risk of stroke in >1.6 million patients undergoing non-cardiac surgery, found that 0.78% of patients developed POAF and among those who did, the cumulative risk of stroke at 1 yr after discharge was 1.47% compared with 0.36% in those who did not develop POAF, with a hazard ratio of 2.0, 95% CI 1.7–2.3.³⁶

In addition to stroke, POAF increases the risk of CHF (relative risk, RR 3.9, 95% CI 2.9–5.3), MI (RR 4.2, 95% CI 2.7–6.6), cardiac arrest (RR 8.0, 95% CI 3.9–16), bacterial pneumonia (RR 7.4, 95% CI 5.5–9.9) and increased hospital length of stay (increase of 2.5 days, 95% CI 1.9–3.1 days).¹¹

The initial goal of treatment should identification and management, if possible, of the likely mechanisms causing AF. Subsequent attempts at rhythm or rate control are less likely to be effective if modifiable risk factors have not been addressed.¹²

In the general population, with asymptomatic and haemodynamically stable patients, conversion of AF to sinus rhythm has been historically thought to be the treatment option of choice. Increasing evidence has shown no clinical benefit when compared with rate control. The AFFIRM trial in 2002 was the first and largest study to compare rate control with rhythm control in >4000 patients with non-valvular AF and increased risk of stroke or death. There were no differences in survival, complications or duration of hospital stay between the two approaches, whether using beta-blockers, calcium channel blockers, or digoxin.³⁷

Postoperative AF is commonly self-limited, averaging a duration of 11–12 h after surgery, with minimal acute alterations in haemodynamics, and 80% of patients who develop POAF convert back to sinus rhythm spontaneously within the first 24 h after onset. This further supports initial management with rate control to allow for possible spontaneous cardioversion.¹² Beta-blocker therapy for rate control is the standard treatment choice, particularly in patients with concomitant ischaemic heart disease.¹³ Other medications, such as amiodarone, non-dihydropyridine calcium channel blockers (such as diltiazem) and digoxin can also be considered to control heart rate depending on the patient's characteristics and comorbidities.¹² The optimal target heart rate is unclear, with recent guidelines suggesting <100 beats min⁻¹ in stable or chronic AF based on data in the AFFIRM and RACE II studies. However, a stricter control of <80 beats min⁻¹ may be beneficial in symptomatic patients or those with reduction in left ventricular systolic function. Specifically after cardiac surgery, <100 beats min⁻¹ should be targeted in asymptomatic patients.^17,37

When AF causes symptoms or haemodynamic instability, rhythm control with immediate cardioversion to sinus rhythm is indicated, using synchronised direct current (DC) cardioversion.^17,38,39 Electrical or pharmacological cardioversion may be considered if patients are haemodynamically stable, and either unable to attain adequate rate control or are intolerant of AF. Choices for pharmacological cardioversion, like the choice for rate control medication, should be tailored to patient-specific risk factors and interactions given each drug's safety profiles.⁴⁰ Antiarrhythmic medications such as amiodarone, procainamide, flecainide, propafenone and vernakalant are used based on specific comorbidities, route and time course to achieve cardioversion.¹² Vernakalant is an atrial-selective antiarrhythmic approved in Europe but not by the US Food and Drug Administration (FDA). It was studied in the setting of POAF after cardiac surgery specifically and carries a class IIb recommendation from the European Society of Cardiology.^31,41, 42, 43, 44 Anticoagulation has been a mainstay for the prevention of thromboembolic events associated with AF, especially as an initial measure in the acute setting. The decision to initiate anticoagulation and the choice of drug used (based on pharmacokinetics and reversibility) can be challenging in the setting of POAF given the need to balance this therapy in the hypercoagulable postoperative state, with the high risk of bleeding if there is inadequate postoperative surgical haemostasis.¹² Most guidelines recommend starting anticoagulation in POAF persisting more than 48 h; the balance of risks and benefit should be assessed and the decision to start anticoagulation should be made using the CHA2DS2-VASc score for thrombotic risk and HAS-BLED score for bleeding risk.^31,40,45 Left atrial appendage exclusion, via surgical ligation (which can be performed in conjunction with other cardiac surgeries in patients with known AF before surgery) or using minimally invasive catheter-based occlusion, is an option to reduce thromboembolic risk in patients who are at high risk or known to be intolerant of anticoagulation.

Ablation therapy, most commonly accomplished in the cardiac surgical patient via a combined surgical (i.e. MAZE procedure) approach for pulmonary vein isolation is a beneficial option for rhythm control in patients with known AF in the perioperative period. While otherwise uncommon in the perioperative period, surgical vs catheter-based vs hybrid ablation therapy provides a long-term treatment option for symptomatic, drug-refractory AF.^41,42 Atrial fibrillation ablation has been compared with pharmacological control of rhythm in several randomised trials, and may be a more effective treatment choice for maintaining long-term sinus rhythm while also conferring a mortality benefit in patients with concomitant CHF^43,44 (Table 2).

Table 2.

Management of postoperative atrial fibrillation. The table lists the appropriate management strategies for treating postoperative atrial fibrillation once it has developed. Initial and long-term management strategies are presented.

Management of postoperative atrial fibrillation
Initial	Unstable haemodynamics, symptoms, or both •Synchronised DC cardioversion •Consider pharmacological cardioversion if the patient is haemodynamically stable, but unable to obtain adequate rate control or intolerant to rate control medication (recommend consultation with electrophysiologist; medications include amiodarone, procainamide, flecainide, propafenone, vernakalant ([Europe only])
	Stable haemodynamics and asymptomatic •Typically self-limiting •Pharmacological rate control (medications include beta-blockers, amiodarone, non-dihydropyridine calcium channel blockers, digoxin)
Long term	•Electrical or pharmacologic cardioversion • Ablation therapy o Pulmonary vein isolation (PVI) o Surgical, catheter-based or hybrid approach o More effective at maintaining longer-term rhythm control than pharmacological management •Anticoagulation for primary risk prevention of thromboembolic sequelae

Conclusions

Perioperative AF is one of the most common arrhythmias. Its effects on individual morbidity and mortality, and on the healthcare system are far-reaching. Appropriate risk identification followed by institution of prevention strategies is imperative, and in the perioperative setting should include a multidisciplinary approach in coordination with anaesthetists, surgeons and cardiac electrophysiologists aimed at reducing the incidence of AF, especially in high-risk patients. Machine learning risk prediction models provide promising tools for increasing accuracy of predicting POAF. If POAF develops, acute management typically involves rate control, unless the arrhythmia leads to haemodynamic instability or causes significant symptoms. In the case of haemodynamically significant AF, early synchronised DC cardioversion should be done and pharmacologic antiarrhythmic therapy can be instituted in consultation with cardiology colleagues. Long-term management of persistent AF may also include the need for anticoagulation and ablation therapy.

MCQs

The associated MCQs (to support CME/CPD activity) will be accessible at www.bjaed.org/cme/home by subscribers to BJA Education.

Declaration of interests

JH is an editor and editorial board member of BJA Education. The other authors declare that they have no conflicts of interest.

Biographies

Jeremy Haysley MD is assistant professor of cardiothoracic anesthesiology in the Department of Anesthesiology and Perioperative Medicine at the University of Louisville School of Medicine.

Hatem Soliman-AboumarieMSc EDICM FRCP FEACVI FASE FHEA is a consultant and honorary senior lecturer in the Department of Cardiothoracic Anaesthesia, Intensive Care, Mechanical Circulatory Support and Transplantation at Harefield and Royal Brompton Hospitals, and at the School of Cardiovascular, Metabolic Medicine and Sciences, King's College London.

Jiapeng Huang MD PhD is professor of cardiothoracic anesthesiology in the Department of Anesthesiology and Perioperative Medicine at the University of Louisville.

Dinesh Kalra MD is professor in the Division of Cardiovascular Medicine in the Department of Medicine at the University of Louisville.

Matrix codes: 1A01, 1A02, 2A03, 2A04, 2A06, 2A07, 2A12, 2C01, 2C07, 3A03, 3A04, 3A05, 3A08, 3C00, 3F00, 3G00, 3I00, 3A13