Classification system of atrial fibrillation
Samuel Lévy, MD, FACC, FESC
A number of publications and clinical trials on the management
of atrial fibrillation (AF) deal with this arrhythmia as if it represents a single entity. As a result, advances made in recent
years have not affected the way AF patients are treated in
general practice except, perhaps, for the use of warfarin in
anticoagulation. Therefore, there is a need for a classification
system and for uniformity in the nomenclature used. The two
terms currently used to describe AF, paroxysmal and chronic,
require a time frame. It is proposed that if an AF episode lasts
longer than 7 days, the condition should be considered
chronic. For the first symptomatic, non–self-terminating
episode that is fewer than 7 days long, the term recent onset
AF may be used, or recent discovery if the AF is asymptomatic
or if the duration cannot be determined. Attacks of paroxysmal
AF may differ in their duration, frequency, and functional tolerance. In the classification system described, three clinical
aspects of paroxysmal AF were isolated in such a way as to
have implications for therapy. This classification system was
found to be useful for characterizing different subsets of
patients with AF. Curr Opin Cardiol 2000, 15:54–57 © 2000 Lippincott
Williams & Wilkins, Inc.
Division of Cardiology, University of Marseille, Hopital Nord, Marseille, France
Correspondence to Samuel Lévy, MD, FACC, FESC, Division of Cardiology,
Hopital Nord, Marseille, 13015 France
Atrial fibrillation (AF) represents one of the most challenging problems of modern cardiology [1]. In order to
take into account recent advances, several scientific
societies, including the European Society of Cardiology,
have recently issued recommendations for the management of this troublesome arrhythmia [2]. However, the
management of AF patients remains difficult, and the
advances made in the mechanism and treatment of
arrhythmia have probably not affected the way AF
patients are treated in general practice, except for the
use of warfarin in anticoagulation. One of the possible
reasons is that the clinician does not know which subset
of patients this new knowledge applies to.
The medical community has not yet clarified the
nomenclature of AF. This arrhythmia is often dealt with
as if AF represents a single entity, while every physician
who treats AF patients knows that there are multiple
clinical presentations of the arrhythmia. Therefore, it is
necessary to classify the various subsets of patients with
AF in order to be able to properly address the problems
raised by the management of each subgroup. For
example, we may be tempted to compare the efficacy of
antiarrhythmic agents reported in the literature. We
know that such comparisons would be inappropriate,
because the differences in success rates may be mainly
related to differences in patient populations and in the
definitions used.
Current Opinion in Cardiology 2000, 15:54–57
Abbreviation
AF
atrial fibrillation
ISSN 0268–4705 © 2000 Lippincott Williams & Wilkins, Inc.
There are several possible classifications of AF. Some
may be based on the electrocardiographic presentation.
For example, Jais et al. [3] have identified an electrocardiographic aspect that is suggestive of “focal AF.” AF
has been classified by Allessie et al. [4] in their experimental model according to a mapping of the atria.
Saksena et al. [unpublished data, Taipei, Taiwan, 1998]
have attempted to use endocavitary recordings to identify various aspects of AF. We will focus our attention on
a clinical classification system and propose a nomenclature. This classification system was found to be useful
for characterizing different subsets of patients with AF.
Definitions and classifications of different
subsets of atrial fibrillation
We use the Bellet definition [5] for the electrocardiographic diagnosis of AF, excluding atrial flutter. In the
current literature, AF is usually divided into two types:
paroxysmal and chronic. However, the definitions of
these terms vary significantly from one article to the
54
Classification system of atrial fibrillation Lévy 55
other, and this may cause some confusion. For example,
Gold et al. [6] define “paroxysmal AF as AF of less than
one year and chronic AF as AF of more than one year.”
For Suttorp et al. [7], “paroxysmal AF is AF of less than
6 months and patients subdivided into AF of recent
onset (<24 hours) and AF long-lasting (>24 hours).” It is
not clear whether these authors wished to characterize
the current AF episode or the time since the first AF
episode. In the Framingham study [8], subjects had a
biennial electrocardiogram. They were defined to have
chronic AF if AF was present in 75% or more biennial
electrocardiograms. For these reasons, in most
Framingham reports, AF was not classified as either
paroxysmal or chronic.
Figure 1. Characterization of various subsets of atrial
fibrillation
AF > 2 min (unrelated to an acute cause)
Clinical history
Paroxysmal
<7d
Self-terminating
< 48 h
Chronic
>7d
Persistent
> 48 h
<7d
© 2000 Lippincott Williams & Wilkins
The term recent onset is often used to designate a
non–self-terminating episode of AF that cannot be classified as chronic or paroxysmal because of its duration.
In the literature, the duration varies from fewer than 24
hours to 1 month. Therefore, various definitions and
time frames are currently used to characterize different
subsets of patients with AF.
We propose the following definitions, which are in
keeping with the words most commonly used to characterize AF. The end point serves more to characterize
the pattern of AF in a given patient at a given time
than to characterize one episode of AF. Paroxysmal AF
is diagnosed in a patient with a history of recurrent
episodes of AF lasting more than 2 minutes and fewer
than 7 days. The first episode of self-terminating AF
lasting under 48 hours is also classified in this group.
Chronic AF is defined as AF present for more than 7
days. Recent onset AF is defined as persistent
(non–self-terminating) AF lasting 48 hours or more but
fewer than 7 days. Classified in this group are the first
symptomatic attack of AF lasting more than 48 hours
and fewer than 7 days, an asymptomatic or mildly
symptomatic persistent AF of recent discovery, and an
AF episode for which the onset could not be determined. Should the physician opt for either pharmacologic or electrical cardioversion for this first episode of
persistent (48 hours or more) AF before 7 days, the
patient would be classified in the recent onset group.
We acknowledge the heterogenous nature of this
group, but it allows us to obtain two clean groups:
paroxysmal AF and chronic AF (Fig. 1).
The terms acute, recurrent, intermittent, and transient
have also been used in the literature to designate the
paroxysmal form of AF. Such uses, with their accompanying inconsistencies in definitions, contribute to the
confusion [9]. For this reason, the Working Group of
Arrhythmias of the French Society of Cardiology [10]
has recommended that acute be reserved for characterizing AF likely to be related to acute etiologies. For
Recent onset
First symptomatic episode
Asymptomatic, first discovery,
or unknown
Cardioversion
Chronic
indicated
<7d
Paroxysmal
<7d
example, AF occurring in the setting of acute myocardial infarction, acute pericarditis, acute myocarditis,
acute pulmonary embolus, hyperthyroidism, or acute
pulmonary disease represents a separate group, because
AF has no tendency to recur should the etiology disappear or be cured. In these cases, AF rarely represents
the major problem. The treatment of the cause or of the
acute episode may result in the disappearence of the
arrhythmia and an end to recurrences.
Classification of paroxysmal atrial fibrillation
The paroxysmal form of AF is often poorly defined in the
current literature. A classification system was recently
proposed that takes into account symptoms, frequency,
mode of termination of attacks, and response to treatment aimed at prevention of recurrences [11] (Tables 1
and 2) . Paroxysmal AF is subclassified into three groups
(Table 1). Class I includes the first symptomatic episode
of AF documented on an electrocardiogram, with spontaneous termination in under 48 hours. Class II consists of
recurrent attacks of AF that are asymptomatic (IIA),
symptomatic and infrequent (less than one attack every 3
months) (IIB), and symptomatic and frequent (more than
one attack every 3 months) (IIC). Class III includes AF
that does not respond to one or more antiarrhythmic
agents aimed at prevention of recurrences.
Table 1. Paroxysmal atrial fibrillation: proposal for a classification
Class I
A
B
Class II
A
B
C
Class III
A
B
C
First symptomatic attack of AF
Spontaneous termination
Pharmacologic or electrical cardioversion
Recurrent attacks of AF (untreated)
Asymptomatic
Less than 1 attack/3 mo
More than 1 attack/3 mo
Recurrent attacks of AF (treated)
Asymptomatic
Less than 1 attack/3 mo
More than 1 attack/3 mo
56 Arrhythmias
Table 2. Paroxysmal atrial fibrillation: therapeutic strategies
Class I
Class II
A
B
C
Class III
A
B&C
First attack of AF
No treatment for prevention
Asymptomatic
Role of antiarrhythmic agents (AA) to prevent
recurrences not established
Episodic treatment is an option
Prevention of recurrences may be warranted
Role of AA therapy for prevention unsettled
Several options
Further investigations, other AA trials, agents to slow the
rate, atrioventricular node modification or ablation
Validity of the proposed classification system
In fact, the definitions of the paroxysmal and chronic
forms of AF are far from universal. It is also not clear
from most reports whether these terms are used to characterize the current episode of AF or the pattern of AF
evolution. It is well known that chronic AF (the established form of arrhythmia unless successfully converted
to sinus rhythm) may be the end result of paroxysmal
AF or the initial presentation of the arrhythmia.
We propose that the current or last episode of AF be
defined according to its duration. When patients seek
medical attention, they may either be in AF at the time of
clinical examination or report a history of AF. By definition, electrocardiographic documentation is necessary to
distinguish AF from other related or nonrelated arrhythmias. As for ventricular arrhythmias, an arbitrary temporal
definition is needed. To define sustained ventricular
tachycardia as lasting 30 seconds or more is arbitrary but
accepted by the medical community. Why not define the
duration of AF? We adopted the temporal definition used
by Friedlander and Levine [12]. If the current or last
episode lasted 7 days or more, AF is considered chronic as
meant in current literature. Although some attacks of
paroxysmal AF may last longer than 7 days, the chances for
an episode of paroxysmal AF to terminate spontaneously
after 7 days are small. The clinical history is necessary to
define the pattern of AF over a period of time. If, on one
hand, the current episode has been present for fewer than
7 days and the patient gives a history of one or more
episodes, AF may still be called paroxysmal. In the paroxysmal form of AF, the episodes may be self-terminating in
fewer than 48 hours or persistent, lasting more than 48
hours. The difficulty arises when dealing with the first
episode or when cardioversion has been performed and has
succeeded within 7 days. As shown in Table 1, we propose
that the first episode of AF self-terminating within 48
hours or successfully cardioverted before 48 hours be
included in the paroxysmal AF classification.
The recent onset or recent discovery group refers to the
first presentation of AF that is either symptomatic and
persistent (lasting more than 48 hours), asymptomatic and
of recent discovery, or when the onset is unknown because
the patient is unable to give an appropriate history.
Other classification systems
Sopher and Camm [13] subdivided AF into acute (AF
present for fewer than 48 hours “distinguished by the
lack of need of anticoagulation”) and chronic AF. In
chronic AF, they recognize three forms: paroxysmal,
persistent, and permanent. For these authors, the term
chronic characterizes the history of AF or the cause of
AF and not the last episode as it is used in the rest of
the literature. In the paroxysmal form, episodes of AF
are self-terminating as opposed to the persistent form of
AF, which requires medical intervention. However,
such classification demonstrates the fact that certain
episodes of AF may be persistent, requiring medical
intervention, as opposed to those that are self-terminating. These three forms characterize the episode (the last
one) rather than the pattern over time of the arrhythmia
attacks. For example, a patient who has at different
times three self-terminating episodes (paroxysmal) per
week and three non–self-terminating episodes (persistent) per month will be difficult to classify. The permanent form of AF refers to AF resistent to cardioversion.
The decision to respect AF and not to make further
attempts at cardioversion obviously depends on the
physician’s judgment. Both AF and the physician’s therapeutic attitude are characterized by the permanent
form as well. Today, we have the capacity to terminate
nearly all episodes of AF with external or internal
cardioversion. Failure to maintain sinus rhythm for a
significant period of time may also justify the inclusion
of the patient in the so-called permanent form. The
maintenance of sinus rhythm also depends on the type
of pharmacological or nonpharmacological therapy used.
Therefore, the medical community needs to adopt
universal definitions for characterizing various subsets of
AF patients. A recent study has shown that important
differences exist between various subsets of AF [14] in
terms of underlying heart diseases and outcome.
References
1
Zipes D: Atrial fibrillation. From cell to bedside. J Cardiovasc Electrophysiol
1997, 8:927–938.
2
Lévy S, Breithardt G, Campbell WF, Camm AJ, Daubert JC, Allessie M, et
al.: Atrial fibrillation: current knowledge and recommendations for management. Eur Heart J 1998, 19:1294–1320.
3
Jais P, Haissaguerre M, Shah DC, Chouairi S, Gencel L, Hocini M,
Clémenty J, et al.: A focal source of atrial fibrillation treated by discrete
radiofrequency ablation. Circulation 1997, 95:572–576.
4
Allessie MA, Konings K, Kirchhof C: Mapping of atrial fibrillation. In Atrial
Fibrillation: Mechanisms and Therapeutic Strategies. Edited by Olsson SB,
Allessie MA, Campbell RWF. New York: Futura Pub Armonk; 1994:37–49.
5
Bellet S: Clinical Disorders of the Heart Beat. Philadelphia: Lea & Febiger;
1971:223–233.
6
Gold RL, Haffajee CL, Charos G, Sloan K, Baker S, Alpert JS: Amiodarone
for refractory atrial fibrillation. Am J Cardiol 1986, 57:124–127.
Classification system of atrial fibrillation Lévy 57
7
Suttorp MJ, Kingma HJ, Lie-A-Huen L, Mast EG: Intravenous flecainide
versus verapamil for acute conversion of paroxysmal atrial fibrillation or
flutter to sinus rythm. Am J Cardiol 1989, 63:693–696.
8
Kannel WB, Abbot RD, Savage DD, McNamara PM: Epidemiologic
features of chronic atrial fibrillation: the Framingham study. N Engl J Med
1982, 306:1018–1022.
9
Lévy S: Nomenclature of atrial fibrillation or the Tower of Babel. J Cardiovasc
Electrophysiol 1997, 9:S83–S85.
10
Lévy S, Attuel P, Fauchier JP, Medvedowsky JL, et al. on behalf of Le Groupe
de Travail en Rythmologie de la Société Française de cardiologie: Nosologie
de la fibrillation auriculaire. Essai de clarification. Arch Mal Coeur 1995,
88:1035–1038.
11
Lévy S: Epidemiology and classification of atrial fibrillation. J Cardiovasc
Electrophysiol 1998, 9:S78–S82.
12
Friedlander RD, Levine SA: Auricular fibrillation and flutter without evidence
of organic heart disease. N Engl J Med 1934, 211:624.
13
Sopher SM, Camm AJ: Therapy for atrial fibrillation: control of the ventricular response and prevention of recurrence. Coronary Artery Disease 1995,
6:106–114.
14
Lévy S, Maarek M, Coumel P, Guize L, Lekieffre J, Medvedowsky JL,
Sebaoun A.: Characterization of different subsets of atrial fibrillation in
general practice in France: The Alfa Study. Circulation 1999,
99:3028–3035.