Atrial fibrillation

AF is very common in the ICU, reflecting in the increasing number of past paper SAQs which discuss the fibrillating atrium, and what precisely you ought to do about it.

Chapter 22 of Oh's Manual (Management  of  cardiac  arrhythmias  by Andrew  Holt) was my first instinctive reference for this topic. The chapter is a marvellous resource, consisting of highly concentrated information compacted in layers upon layers into a dense sedimentary deposit. Nuggets of gold are embedded within.  Most of the end section (pages 216 to  224) deals almost exclusively with atrial fibrillation, and this summary chapter uses it and its references ad libitum to answer the SAQs. As far as useful  information goes, the AHA's 2014 guidelines statement can be viewed as the definitive document for management and investiations, in spite of the fact that its intended audience is largely ward-based and outpatient physicians. As is pointed out below, AF in the ICU is a completely different animal to "outpatient AF".  

Atrial flutter vs. atrial fibrillation

Oh's Manual pays a little token attention to flutter before engaging in a detailed discussion of AF. Yes, both are atrial arrhythmias, but totally different pathology is involved in each. Flutter is an intra-atrial re-entrant circuit which just goes around and around (usually, counter-clockwise and in the right atrium). It is apparently notoriously drug-resistant and Holt recommends ibutilide as the drug most likely to cardiovert this rhythm. Judging by the tone of Holt's chapter, there is nothing to recommend any specific agent to the task of controlling ventricular rate in flutter. A group is listed in a nonchalant manner ("Drugs such  as  digoxin,  diltiazem,  beta-adrenergic blockers, sotalol  and  amiodarone  may  be  tried;  the  choice depends on LV function"). We are warned against the use of Class 1A and 1C agents, as they may result in 1:1 conduction (and a ventricular rate in excess of 200). By comparison, the more recent article by Herzog et al (2017) is much more firmly prescriptive. It separates the management of flutter into four main areas, described by the RACE acronym: Rate control, Anticoagulation, Cardioversion and Electrophysiology. The article summarises the recommendations from the most recent guidelines from the AHA/ACC:

  • Same drugs for rate control as AF (but higher doses)
  • Same anticoagulants as AF
  • Cardioversion: chemical choices include ibutilide or amiodarone
  • Cavo-tricuspid isthmus ablation (EPS)

Generally, rate control is more difficult to achieve than with AF. UpToDate authors suggest calcium channel blockers (diltiazem or verapamil) but the level of evidence is somewhat sub-standard ("we prefer", the authors say). The problem with ICU patients is that frequently, the (higher) dose required to slow AV nodal conduction would produce an intolerable level of toxicity (i.e. the LV would not be grateful for the decreased contractility). Ergo, amiodarone becomes the natural choice in the ICU, because it covers both the "R" and the "C" of the RACE algorithm. 

Consequences of atrial fibrillation

Why do we even try to control AF?

  • Adverse haemodynamic effects due to the following factors:
    • loss of atrial systole (the "kick")
    • Decreased diastolic filling time during rapid AF
    • "Tachycardiomyopathy" - a global cardiomyopathy associated with the rapid rate
  • Systemic embolism and stroke

Causes of atrial fibrillation

Question 14 from the first paper of 2012 asked the candidates to list several non-cardiac causes of AF. The list of causes could be long, and it is possible to organise it in several ways. One way is the familiar VINDICATE framework which should theoretically make it easier to memorise the causes. A strategy which might be more useful (actually enhancing one's practice) would be to organise the causes according to their pathophysiology. Both are made available below.

Causes of Atrial Fibrillation Organised by System


  • Myocardial infarction
  • Pulmonary embolism
  • Pulmonary hypertension
  • Subarachnoid haemorrhage


  • Sepsis
  • Myocarditis
  • Pericarditis
  • Infective endocarditis


  • Cardiac mass, eg. myxoma


  • Catecholamines
  • Alcohol
  • Caffeine


  • Infiltrative disease, eg. amyloidosis
  • Age-related fibrotic changes


  • Infiltrative disease, eg. amyloidosis
  • Age-related fibrotic changes


  • Atrial septal defect
  • Familial AF


  • Autoimmune myocarditis


  • Cardiac contusion
  • Cardiac surgery


  • Hypothermia
  • Hyperthyroidism
  • Haemochromatosis/iron overload
  • Phaeochromocytoma
  • Electrolyte derangement
Causes of Atrial Fibrillation Organised by Pathophysiology

Catecholamine excess

  • Exogenous (eg. adrenaline infusion)
  • Endogenous:
    • Subarachnoid haemorrhage
    • Stress
    • Phaeochromocytoma
    • Thyrotoxicosis (indirectly)

Atrial distension

  • Pulmonary hypertension:
    • Obstructive sleep apnoea
    • Pulmonary embolism
    • Primary pulmonary hypertension
    • Pulmonic valve disease
  • Septal defects
  • Valvular disease, including infective endocarditis

Abnormality of conducting system

  • Congenital cardiac disease, eg. septal defect
  • Infiltrative cardiac disease, eg. amyloidosis
  • Ischaemic heart disease
  • Age-related fibrotic changes
  • Haemochromatosis/iron overload
  • Hypothermia

Increased atrial automaticity / irritation

  • Alcohol
  • Caffeine
  • Catecholamines
  • Electrolyte derangement
  • Myocarditis

Management of atrial fibrillation in a broad sense

"Outline your initial management of the tachycardia", asks Question 14 from the first paper of 2012. Treatment for AF is a pursuit of several goals: control of ventricular rate, protection from systemic emboli and restoration of sinus rhythm.

Rate control vs. rhythm control

Holt's chapter for Oh's Manual mentions four studies (AFFIRM, RACESTAF and PIAF) in support of rate control instead of rhythm control. They were published between 2000 and 2003.  A more recent meta-analysis (Caldeira et al, 2012) identified four more (a total of eight) suitably high-quality studies, featuring data from 7499 patients. The mortality data from these was underwhelming. "No clear survival benefit is apparent", laments Holt. Rate control was found to be superior only in terms of the composite endpoint (death, stroke and recurrent hospitalisation). Theoretically, rhythm control should actually be better (particularly in the absence of significant structural heart disease) because it may prevent myocardial remodelling due to AF.

Methods of rate control

Multiple methods are available. Oh's Manual lists essentially the entire antiarrhythmic arsenal,  cautioning against the use of digoxin in patients with enhanced sympathetic tone (useless, apparently). Single-dose flecainide is apparently also good, but in patients with structural heart disease it tends to cause sudden cardiac death.

The 2014 AHA statement gives the following recommendations:

  • β-blocker or calcium channel blocker for paroxysmal AF, which can be given IV if they are haemodynamically stable (CCBs are favoured in COPD patients)
  • IV amiodarone for critically ill patients
  • If the patient has some sort of pre-excitation problem, these AV node blockers are not recommended (cardioversion is the better choice; the AHA also recommend procainamide for stable patients but this is not available in Australia)
  • If the patient is in decompensated heart failure, calcium channel blockers are not recommended (use digoxin or amiodarone instead). In the long term these people seem to benefit from a combination of digoxin and a β-blocker

They recommend a rate of 80 or so as the endpoint to aim for, but give a slightly weaker recommendation in favour of a more "lenient" rate (~110) provided the LV function is well-preserved. 

Methods of rhythm control

For the haemodynamically unstable patient, DC cardioversion is the gold standard. The arrhythmia management algorithm in the pre-arrest management section of the ARC ALS Handbook  (based on ILCOR Guidelines) recommends this approach. For the haemodynamically stable patient, there are a variety of chemical cardioversion options, among which amiodarone and vernakalant (the latter being a novel drug with rapid activity against AF and  a reasonably benign side-effect profile). Electrical cardioversion may still be an option, within 48 hours of onset or later following adequate anticoagulation and TOE.

Systemic anticoagulation

The options are:

  • Warfarin: relative risk reduction for stroke 62%; absolute risk reduction 2.8% per year
  • Aspirin: relative risk reduction for stroke 22%; absolute risk reduction 1.5% per year
  • Warfarin plus aspirin: no additional benefit over warfarin alone
  • Dabigatran: 35% reduction in stroke compared to warfarin

The 2014 AHA statement recommends:

  • No anticoagulation is a reasonable choice if the AF is within 48 hours and the CHA2DS2-VASc score is 0; the risk of stroke is 0.2%.
  • If the AF is within 48 hours and the score is anything but 0, IV heparin is recommended.
  • If the AF has been going on for longer than 48 hours (or, god knows how long) - IV heparin should be used.
  • If you are going to anticoagulate, anticoagulation with something should continue for at least 3 weeks before and 4 weeks after their TOE-cardioversion.
  • Aspirin is an alternative if the score is 1
  • Warfarin or similar if the score is 2 or greater
    • Essentially this means that if you're over 75 you automatically score some warfarin because that age bracket immediately gets you a score of 2 from the CHA2DS2-VASc scoring system.

Risk stratification

The CHA2DS2-VASc scoring system is the recommended method of determining the risk of stroke. In essence it comes down to three main categories: score 0, score 1 and any score of 2 or more.

 C   Congestive heart failure (or Left ventricular systolic dysfunction)
 H  Hypertension: blood pressure consistently above 140/90 mmHg (or treated hypertension on medication)
 A2  Age ≥75 years
 D  Diabetes Mellitus
 S2  Prior Stroke or TIA or thromboembolism
 V  Vascular disease (e.g. peripheral artery disease, myocardial infarction, aortic plaque)
 A  Age 65–74 years
 Sc  Sex category (i.e. female sex)

A score of 0 is hard to get, but confers a virtually negligible risk of stroke (~ 0%). A score of 1 equates to a risk of 1.3% and a score of 2 puts you in a high risk category (2.2%). The maximum score is 9, with an associated stroke risk of 15.2%. Mind you, these are annual risks. What's the daily risk in ICU patients? Nobody knows. Sibley et al (2015) mulls the problem over and concludes that whatever that risk is, it surely mist be higher in the ICU population, particularly among patients with sepsis. At present there does not appear to be either any sort of scoring system or any sort of guidelines statement to help direct the intensivist here, apart from generic guidelines like the 2014 AHA statement.

Management of AF in the ICU

There are several important differences between "free-range" AF in the movie-going public and "ICU AF " in the critically ill patient.

  • AF in ICU is a transient phenomenon. Only about 15% of the fibrillating ICU patients will still be fibrillating at the time of their discharge from the ICU (Artucio et al, 1990).
  • The cause of AF in ICU is usually non-structural, reversible, and non-cardiac (i.e. related to the cause of the non-cardiac critical illness). A large epidemiological survey of non-cardiothoracic ICU patients by Kanji et al (2012) confirmed that most of the time AF in the ICU can be attributed to such causes as shock, sepsis, electrolyte disturbance, atrial stretch, hypoxia, inotropes, and so forth.  All of these are problems which can be expected to resolve at the end of the critical illness. 
  • The mortality and morbidity in the ICU population are different to those of the general fibrillating population. The ICU patient with AF is less likely to die of stroke, but is more likely to develop worsening multi-organ system failure due to poorly controlled shock (when they lose that atrial systolic kick). Among ICU patients it is difficult to say whether the AF is independently associated with increased mortality, or whether it is merely a marker of greater illness severity.
  • Given that "ICU AF" is likely to be associated with shock and a catecholamine excess, it makes sense that a beta-blocker may be the most appropriate choice of drug. However, choice of agents is clouded by methodologic heterogeneity among trials. Kanjii et al (2008) made valiant attempts to perform a meta-analysis of known methods, and came up against a wall of wildly differing methodologies and outcome measures. Standards are lacking. Unstable patients were excluded from all studies (which makes them completely useless in the ICU patients who are unstable almost by definition).
  • As far as chemical cardiversion goes, Kanjii et al (2008) reported on some figures. Esmolol and flecainide seem to give the best chance of cardioversion (60-100% success rate at 12 hours), with amiodarone far behind (40-76% success rate at 12 hours). Even magnesium was better than amiodarone (51-93% success rate).


Wyse, D. G., et al. "A comparison of rate control and rhythm control in patients with atrial fibrillation." The New England journal of medicine 347.23 (2002): 1825-1833.

Van Gelder, Isabelle C., et al. "A comparison of rate control and rhythm control in patients with recurrent persistent atrial fibrillation." New England Journal of Medicine 347.23 (2002): 1834-1840.

Jörg Carlsson, J., et al. "Randomized trial of rate-control versus rhythm-control in persistent atrial fibrillation: The Strategies of Treatment of Atrial Fibrillation (STAF) study." Journal of the American College of Cardiology 41.10 (2003): 1690-1696.

Hohnloser, Stefan H., et al. "Rhythm or rate control in atrial fibrillation—Pharmacological Intervention in Atrial Fibrillation (PIAF): a randomised trial." The Lancet 356.9244 (2000): 1789-1794.

Yoshida, Takuo, et al. "Epidemiology, prevention, and treatment of new-onset atrial fibrillation in critically ill: a systematic review." Journal of intensive care 3.1 (2015): 19.

Caldeira, Daniel, Cláudio David, and Cristina Sampaio.  "Rate versus rhythm control in atrial fibrillation and clinical outcomes: updated systematic review and meta-analysis of randomized controlled trials." Archives of cardiovascular diseases 105.4 (2012): 226-238.

ARTUCIO, HERNAN, and MAXIMO PEREIRA. "Cardiac arrhythmias in critically ill patients: epidemiologic study." Critical care medicine 18.12 (1990): 1383-1388.

Reddy, Madhu, et al. "VERNAKALANT FOR RAPID CARDIOVERSION OF RECENT ONSET ATRIAL FIBRILLATION: A META-ANALYSIS." Journal of the American College of Cardiology 63.12_S (2014).

Morrison, Laurie J., et al. "Part 8: advanced life support 2010 international consensus on cardiopulmonary resuscitation and emergency cardiovascular care science with treatment recommendations." Circulation 122.16 suppl 2 (2010): S345-S421.

Arrigo, Mattia, Dominique Bettex, and Alain Rudiger. "Management of Atrial Fibrillation in Critically Ill Patients." Critical Care Research and Practice 2014 (2014).

Kanji, Salmaan, et al. "Epidemiology and management of atrial fibrillation in medical and noncardiac surgical adult intensive care unit patients." Journal of critical care 27.3 (2012): 326-e1.

Kanji, Salmaan, et al. "Treatment of new-onset atrial fibrillation in noncardiac intensive care unit patients: A systematic review of randomized controlled trials*." Critical care medicine 36.5 (2008): 1620-1624.

January, Craig T., et al. "2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation." Circulation (2014): CIR-0000000000000041.

Sibley, Stephanie, and John Muscedere. "New-onset atrial fibrillation in critically ill patients." Canadian respiratory journal 22.3 (2015): 179-182.

Herzog, Eyal, et al. "Pathway for the Management of Atrial Fibrillation and Atrial Flutter.Critical pathways in cardiology16.2 (2017): 47-52.