(Circulation. 1995;92:1954-1968.)
© 1995 American Heart Association, Inc.
Articles |
Atrial Fibrillation Begets Atrial Fibrillation
A Study in Awake Chronically Instrumented Goats
From the Department of Physiology, Cardiovascular Research Institute Maastricht (CARIM), University of Limburg, The Netherlands.
Correspondence to Prof Dr M.A. Allessie, Department of Physiology, Cardiovascular Research Institute Maastricht, University of Limburg, PO Box 616, 6200 MD Maastricht, The Netherlands.
 |
Abstract |
|---|
 Top Abstract IntroductionMethods Results Discussion References |
|---|
Background In this study we tested the hypothesis that atrial fibrillation (AF) causes electrophysiological changes of the atrial myocardium which might explain the progressive nature of the arrhythmia.
Methods and Results Twelve goats were chronically instrumented with multiple electrodes sutured to the epicardium of both atria. Two to 3 Weeks after implantation, the animals were connected to a fibrillation pacemaker which artificially maintained AF. Whereas during control episodes of AF were short lasting (6±3 seconds), artificial maintenance of AF resulted in a progressive increase in the duration of AF to become sustained (>24 hours) after 7.1±4.8 days (10 of 11 goats). During the first 24 hours of AF the median fibrillation interval shortened from 145±18 to 108±8 ms and the inducibility of AF by a single premature stimulus increased from 24% to 76%. The atrial effective refractory period (AERP) shortened from 146±19 to 95±20 ms (-35%) (S1S1, 400 ms). At high pacing rates the shortening was less (-12%), pointing to a reversion of the normal adaptation of the AERP to heart rate. In 5 goats, after 2 to 4 weeks of AF, sinus rhythm was restored and all electrophysiological changes were found to be reversible within 1 week.
Conclusions Artificial maintenance of AF leads to a marked shortening of AERP, a reversion of its physiological rate adaptation, and an increase in rate, inducibility and stability of AF. All these changes were completely reversible within 1 week of sinus rhythm.
Key Words: atrium • fibrillation • remodeling • wavelength
|
Introduction |
|---|
|
TopAbstractIntroductionMethods Results Discussion References |
|---|
Atrial fibrillation (AF) is a common arrhythmia in man with an overall prevalence of 0.4% to 0.9%. Its incidence increases with age with about 0.1% to 0.2% per year over the age of 40, resulting in a prevalence of 2% to 4% in the population over 60 years of age.1 2 3 Rheumatic and ischemic heart disease, hypertension and congestive heart failure are important risk factors for the development of atrial fibrillation leading to a prevalence of as high as 40% in patients with overt congestive heart failure.4 On the other hand, in 3% to 31% of documented atrial fibrillation no underlying cardiovascular disease can be detected (lone atrial fibrillation).2 5 6 7 8
Paroxysmal atrial fibrillation often progresses to chronic atrial fibrillation, the transition rate varying considerably with the underlying etiology.1 However 18% of patients with lone paroxysmal atrial fibrillation also develop sustained fibrillation.7 The duration of the paroxysms of AF was found to be of importance, transition to chronic fibrillation occurring in 31% of patients with paroxysms shorter than 2 days versus 46% if the episodes of AF were of longer duration.1 From these epidemiological data thus it seems that, independent of the underlying etiology, atrial fibrillation in itself is a progressive disease. The clinical experience that with time it becomes more and more difficult to keep a patient with AF in sinus rhythm, has been expressed by M. Rosenbaum by the term "domestication of atrial fibrillation" (personal communication, 1992).
Other evidence that atrial fibrillation promotes atrial fibrillation is the observation that chemical or electrical defibrillation has a higher success rate when atrial fibrillation has existed only for a short time. Cardioversion by intravenous administration of flecainide was successful in 76% to 93% of patients with recent onset fibrillation (<24 hours) compared with 0% to 83% in patients with AF of longer duration.9 10 11 12 Amiodarone cardioverted AF in 85% if lasting for less than 1 year versus 57% in patients who had AF for longer than 1 year. In patients with long-lasting AF the chances of maintaining normal sinus rhythm after cardioversion with amiodarone were also less.13
The success rate of electrical atrial defibrillation also depends on the duration of AF.14 15 In 186 patients whom were successfully defibrillated, the mean duration of atrial fibrillation was 16±27 months compared with 28±45 months in the total study group of 246 patients.16 The recurrence rate after successful DC countershock was also higher in patients with a longer history of atrial fibrillation.14 15 17 18 19
A possible explanation for all these epidemiological and clinical observations is that, apart from the progressive changes due to an underlying heart disease, atrial fibrillation itself causes progressive electrophysiological and/or structural changes to the atria, which promote the initiation or perpetuation of atrial fibrillation.
The present study was designed to test the hypothesis that atrial fibrillation begets atrial fibrillation and to explore the electrophysiological changes that may be responsible for this phenomenon.
|
Methods |
|---|
|
TopAbstractIntroductionMethods Results Discussion References |
|---|
A Conscious Goat Model of Sustained Atrial Fibrillation
Twelve goats weighing between 46 and 68 kg (55.3±7.4 kg) were used for this study. Animal handling was performed according to the guiding principles of the American Society of Physiology and approved by the Animal Investigation Committee of the University of Limburg. The goats were anaesthetized with Nesdonal (thiopental, 10 mg/kg) and ventilated by Halothane (1% to 2%) and a 1:2 mixture of O2 and N2O2. A left intercostal thoracotomy was made and the pericardium was opened to expose the heart. A silicon strip (Silastic, Dow Corning) of 10x1.2 cm, containing 15 unipolar platinum recording electrodes (diameter, 1.5 mm; interelectrode distance, 6 to 10 mm) was guided through the anterior transverse sinus between the atria and the aortic root and sutured to the tips of both atrial appendages. In addition, two smaller silicon strips of 3x1.2 cm, each containing 6 electrodes, were sutured to the lateral walls of the right and left atria (Fig 1
). After
approximation of the pericardium and closure
of the thorax, the electrode leads were tunneled subcutaneously to the
neck and exteriorized by a 30-pin connector (Lemosa; outer diameter, 10
mm). Three silver plates (diameter, 25 mm) were left subcutaneously to
serve as grounding and indifferent electrodes and to record a
precordial ECG. Before surgery, the animals received
buprenorfine (Temgesic, Reckitt & Colman) for 1 to 3 days. Ampicillin
(1000 mg) was given prophylactically once before and once
after surgery.
|
Automatic Induction of Atrial Fibrillation
About 2 to 3 weeks
after surgery, the goats were connected to an
external automatic atrial fibrillator. The goats were kept in separate
boxes (size, 1.5x0.7 m) with free access to food and water. A cable
from the ceiling was plugged into the connector in the neck of the
animals and the atrial electrodes were connected to a multichannel
recording unit (gain, 200 to 400; bandwidth, 1 to 500
Hz).20 A spring between the cable and the fixation point
at the ceiling allowed free movements of the goats in their boxes. The
atria could be stimulated through any of the epicardial electrodes. The
automatic fibrillator consisted of a personal computer (386 processor)
connected to a stimulator (Medtronic, SP3084). The computer program
continuously analyzed one of the recorded bipolar atrial
electrograms and determined the maximal length of the isoelectrical
segment in consecutive time windows of 1 second. During sinus rhythm
the duration of the isoelectrical segment was always longer than 300 to
400 ms, whereas during atrial fibrillation it was shorter than 80 to
120 ms. Because of this large difference this proved to be a simple and
reliable criterion to automatically distinguish between sinus rhythm
and atrial fibrillation. When the automatic fibrillator was turned on,
a 1-second burst of biphasic stimuli (interval, 20 ms; 4 times
diastolic threshold) was delivered as soon as sinus rhythm
was detected. As can be seen in Fig 2, this promptly
induced atrial fibrillation in a reproducible way. By giving automatic
bursts of stimuli immediately after atrial fibrillation converted to
sinus rhythm, atrial fibrillation could be maintained continuously
during 24 hours a day, 7 days a week. All moments of induction of AF
were stored in the computer and the duration of each episode was
calculated. During the maintenance of AF, the atrial
fibrillation interval was measured on line from a single bipolar atrial
electrogram. Additionally, at regular time intervals all electrograms
were stored on magnetic tape and atrial fibrillation interval
histograms were made from the unipolar electrograms recorded at
different sites.
|
Experimental Protocol
After the goats had recovered from
surgery and before they were
connected to the automatic fibrillator, first an extensive
electrophysiological study was done to
measure the atrial refractory period and conduction velocity at various
sites of the atria. The atrial effective refractory period (AERP) was
measured during a wide range of pacing frequencies
(S1S1 interval, 120 to 600 ms). A single
premature stimulus (4x threshold) was interpolated at every fifth
basic interval and, starting from well within the refractory period,
the S1S2 coupling interval was incremented in
steps of 1 ms. The shortest S1S2 interval
resulting in a propagated atrial response was taken as the AERP. This
method of measuring the refractory period is fast and reproducible and
has the advantage that the coupling interval of the test stimulus can
be incremented rapidly without disturbing the steady state of the paced
heart rate.21 Intra-atrial conduction velocity was
measured during regular pacing with intervals ranging between 120 and
600 ms either from the left or the right atrial appendage. The
conduction velocity was calculated from the conduction times
recorded at the row of electrodes positioned on Bachmann's bundle
between the right and left atrial appendages. The distance between the
electrodes on Bachmann's bundle used for the calculation of the
conduction velocity varied in different experiments between 2.2 and 7.2
cm.
After the fibrillation pacemaker had been turned on, the atrial refractory periods and conduction velocities were measured again after respectively 6 and 24 hours of maintained AF. In some goats these measurements could also be repeated after 2 to 4 days.
The vulnerability of the atria to fibrillation was compared during control and after 24 hours of AF. During regular pacing with a fixed interval of 400 ms single early premature stimuli of 4 times threshold were administered through the same stimulating electrodes as used for regular pacing. Atrial fibrillation was considered to be induced if the single premature stimulus was followed by rapid irregular atrial activity lasting for more than 1 second.
In 5 goats the reversibility of effects was studied after conversion of long-lasting AF (2 to 4 weeks) to sinus rhythm. In 4 goats AF converted spontaneously and in 1 goat AF was terminated by an infusion of Cibenzoline (Cipralan, 1.5 mg/kg). Measurements were made 6 hours, 24 hours, 1 week and 2 weeks after conversion to sinus rhythm.
Data are presented as mean±SD. Statistical analysis of the obtained data was performed by using the paired Student's t test or when the data did not have a normal distribution by the Wilcoxon matched pairs signed ranks test. In case of multiple statistical comparisons (eg, control versus 6 hours and 24 hours of AF) the probability value was corrected by multiplying it with the number of comparisons (Bonferroni's correction). A corrected probability value of less than .05 was considered to be statistically significant.
|
Results |
|---|
|
TopAbstractIntroductionMethods Results Discussion References |
|---|
Development of Sustained Atrial Fibrillation
In Fig 3
, a representative example
is shown of the induction of atrial fibrillation in a goat in sinus
rhythm and after 24 hours and 2 weeks of electrically maintained atrial
fibrillation. Typically in goats which had been in sinus rhythm all the
time, electrically induced AF lasted only for a short period and
usually terminated spontaneously within seconds (upper trace). However,
as shown by the second and third tracing recorded from the same
goat, the duration of paroxysms of AF increased after AF had been
maintained for some time. In this case after 24 hours of AF the bouts
of atrial fibrillation lasted for about 20 seconds (middle trace), and
after 2 weeks atrial fibrillation became sustained and no longer
terminated spontaneously (bottom trace). Actually, the bottom trace
shows induction of AF the last time it had converted spontaneously to
sinus rhythm. The repetitive induction of fibrillation not only
resulted in a dramatic prolongation of the duration of fibrillation but
also the rate of fibrillation increased significantly. This is
illustrated in Fig 4 in which a single unipolar
electrogram is shown recorded from the left atrial appendage during
various stages of AF. While during acute fibrillation (upper tracing)
the median fibrillation interval was 149 ms, already after 6 hours of
AF the interval had shortened to 132 ms. After 24 hours of maintained
AF the median fibrillation interval had further decreased to 117 ms to
become as short as 91 ms after 2 weeks when AF had become sustained
(lower tracing). Also the configuration of the atrial electrograms
changed during the development of chronic fibrillation. While during
the first days of AF most activations still showed a single steep
negative deflection of high amplitude separated by an isoelectric
segment, after 2 weeks of AF single steep deflections were less common
and many electrograms now were of low amplitude with a high degree of
fragmentation and no clear isoelectric segment (lower tracing).
This suggests that during recent onset fibrillation the atrium is still
activated uniformly by broad activation waves (type I
fibrillation) and that after 2 weeks of AF, activation of the atrium
had become more complex by the development of multiple arcs of
intra-atrial conduction block (type II or type III
fibrillation).22 23
|
|
Fig 5 shows
the time course of the prolongation of
episodes of AF during the first 2 weeks of fibrillation in 4 different
goats (A through D). As can be seen there existed a large
inter-individual variation in the time course of development of
sustained atrial fibrillation, defined as AF lasting longer than 24
hours. In panel A, sustained fibrillation occurred already after 3
days. In contrast, in goat D the time course of prolongation of AF was
much slower and after 2 weeks the average duration of AF was still not
longer than 50 minutes. B and C show two examples of an intermediate
time course in which sustained AF was reached after about 8 and 12
days.
|
In Fig 6 (A through D), the time course of changes
in
fibrillation interval is given for the same 4 goats as in Fig
5
.
Compared with the time course in AF duration, the time course of
changes in fibrillation interval showed much less inter-individual
variation. In all goats the fibrillation interval decreased rapidly
within the first 24 hours with a time course of about 2 ms/h. The
atrial fibrillation interval continued to shorten more slowly during
the next days with only a few milliseconds per day to reach a steady
state after about 6 days.
|
Table 1 gives the median
duration of the episodes of
induced AF and the fibrillation intervals of all goats both during
sinus rhythm and after respectively 24 hours, 48 hours, 1 week, and 2
weeks of atrial fibrillation. During sinus rhythm the episodes of
electrically induced atrial fibrillation were very short and terminated
already after 6±3 seconds. The median interval of these short runs of
AF ranged from 113 to 176 ms (mean, 145±18 ms). After 24 hours of
atrial fibrillation the average duration of AF had increased to
2.2±3.0 minutes and the fibrillation interval had shortened to
108±8
ms. After 48 hours in 2 goats, AF had become sustained (lasting >24
hours), while in the remaining 10 goats, fibrillation lasted for
7.8±9.7 minutes. The mean fibrillation interval was 105±8 ms.
After 1
week, in 5 of 11 goats AF had become sustained and in the 6 other
animals fibrillation lasted for 241±459 minutes. The fibrillation
interval was 100±5 ms (one goat (No. 8) dropped out due to the
development of a serious sepsis after 4 days of AF). After 2 weeks, in
9 of 11 goats and within 3 weeks in 10 of 11 goats fibrillation lasted
longer than 24 hours. In the goat in which AF did not become sustained
the longest episode of AF lasted for 13.5 hours. During sustained
atrial fibrillation (>24 hours) the average AF interval was 99±10 ms
compared with 145±18 ms for the short-lasting episodes of AF
induced during sinus rhythm.
|
Correlation Between Rate and Duration of Atrial
Fibrillation
Although the development of chronic fibrillation was
accompanied
by a progressive increase in the rate of fibrillation, there was a
discrepancy in the time course of these changes. This is illustrated in
Fig 7, in which the changes in both
parameters in goat No. 3 are plotted during the first 24
hours of electrically maintained AF. As usual fibrillation started as
short paroxysms lasting for only a few seconds (upper panel) with a
fibrillation interval between 150 and 164 ms (lower panel). However,
while the average fibrillation interval already commenced to shorten
within the first hours of maintained AF, it was not until after 15
hours of repetitive induction of AF that also the duration of the
episodes of AF started to get longer. At that time the average
fibrillation interval had already shortened to 123 ms. At the end of
the first 24 hours of atrial fibrillation the duration of AF had
increased to an average of 1.5 minutes and the fibrillation interval
had shortened to 100 ms. In Fig 8, for 6 goats the
duration of the episodes of AF is plotted against the corresponding
fibrillation intervals. From this plot it appears that at fibrillation
intervals longer than about 120 ms, fibrillation was short lasting and
usually terminated spontaneously within less than 10 seconds. However
when the median fibrillation interval became shorter than 120 ms an
exponential rise in the duration of AF was found. From these data thus
it seems that the cascade of cause and effect, finally leading to
chronic atrial fibrillation, is started by a shortening of the
fibrillation interval. As soon as the fibrillation interval passes a
critical threshold of 120 ms, obviously atrial fibrillation becomes
more stable and the duration of AF starts to increase. This in turn
will further shorten the fibrillation interval which will prolong the
duration of AF again, etc. Such a positive feedback mechanism will
continue until a new steady state is reached in which atrial
fibrillation has become the predominant atrial rhythm (domestication of
atrial fibrillation).
|
|
Effects of Fibrillation on Intra-atrial
Conduction
To test whether the prolongation of AF was caused by
disturbances in intra-atrial conduction, the
intra-atrial conduction velocity was determined both during sinus
rhythm and after 6 and 24 hours and 2 to 4 days of atrial fibrillation.
As soon as an episode of induced atrial fibrillation terminated
spontaneously the atria were paced regularly at various pacing rates
either from the left or the right atrial appendage and the conduction
times along the row of electrodes sutured on Bachmann's bundle were
measured over a distance varying between 2.2 and 7.2 cm. In Fig
9, an example is given (goat No. 4). At a pacing
interval of 500 ms, the conduction velocity was 143 cm/s. Pacing at
shorter intervals first resulted in a slight slowing of the conduction
velocity whereas at pacing intervals of less than 250 ms conduction
velocity was depressed more markedly resulting in a velocity of less
than 110 cm/s at the maximum pacing rate. During the first days of AF,
no depressive effect on the intra-atrial conduction velocity could
be found. The slight shift of the velocity curve to the left was due to
a shortening of the atrial refractory period (see below). As can be
seen from Table 2, actually this resulted in an increase
in conduction velocity during pacing with short intervals (250 and 200
ms).
|
|
Effects of Fibrillation on the Atrial Refractory
Period
In contrast to the intra-atrial conduction velocity,
which did not seem to be affected during the first days of atrial
fibrillation, marked changes in the atrial refractory period occurred
within the first 24 hours of AF. Fig 10 gives a
representative example. During pacing with a fixed
interval of 400 ms, the atrial refractory period was measured by giving
an early interpolated stimulus (S2) after every fifth basic
stimulus (S1). During control (upper two tracings), the
shortest S1S2 coupling interval that resulted
in an atrial response was 127 ms. Already after 6 hours of AF (middle
tracings) the atrial refractory period had shortened considerably and a
premature stimulus of 104 ms elicited a premature beat which started a
short run of rapid atrial responses. After 24 hours of AF, the AERP had
shortened to 90 ms and a single early premature beat now induced a
short run of AF.
|
In Fig 11, four examples of the
changes in AERP at
different pacing intervals are shown (A through D). In A (goat No. 6),
because of the physiological rate adaptation of the
refractory period, during control the refractory period shortened from
150 ms during slow pacing to 132 ms at a pacing interval of 180 ms.
Already after 6 hours of fibrillation, the adaptation curve had clearly
shifted downward, indicating a general shortening of the refractory
period. After 24 hours of AF the curve was further shifted downward and
the refractory period at 500 ms pacing interval had shortened by about
50 ms to less than 100 ms. At the higher pacing rates the curve had
become flat and the normal prolongation of the refractory period upon
slowing of the heart rate was abolished. At the slower heart rates
(right part of the curve), now the refractory period actually became
shorter when the pacing interval was prolonged. In goat No. 5 (B) after
6 hours of AF only the right part of the curve had shifted downwards
and the refractory period during higher pacing rates was not yet
changed. However, after 24 hours, the whole curve was shifted downward
and at all heart rates the refractory period had become shorter than 80
ms. Also in this case the rate adaptation was reversed and the
refractory period during slow pacing was shorter than at higher pacing
rates. In goats No. 4 and No. 7 (C and D) already during control a
slight inversed adaptation of the AERP was present at slow heart
rates. Again, atrial fibrillation shortened the refractory period
markedly and in these cases the rate adaptation of the refractory was
maintained. As a result the AERP was short, both at very short and at
long pacing intervals. Whereas in C, after 24 hours of AF the
refractory period at short and long cycle length was similar (about 110
ms), in panel D the atrial refractory period at slow heart rates was
shorter (94 ms) compared to fast heart rates (107 ms). In summary,
within 24 hours of atrial fibrillation the atrial refractory period
became markedly shortened at all heart rates. Because this shortening
was more pronounced at slower heart rates the
physiological adaptation of the refractory period
to changes in heart rate was attenuated or even inversed.
|
In Table
3, the average values of the atrial refractory
period at different pacing rates are given for all experiments. Within
24 hours of atrial fibrillation, during pacing at 400 ms the refractory
period shortened from 146±19 ms to 95±20 ms (-35%)
(P<.001). At a pacing interval of 200 ms it shortened from
131±11 ms to 106±17 ms (-19%) (P<.001),
whereas during
the maximal pacing rate (Fmax) the refractory period
changed from 117±12 ms to 103±14 ms (-12%)
(P<.01).
Because of the abnormal (reversed) adaptation of the refractory period
to changes in heart rate, after 24 hours of fibrillation the refractory
period during slow heart rates actually had become shorter than during
the maximum pacing rate (95±20 ms versus 103±14 ms). Because the
atrial vulnerability to fibrillation progressively increased and also
the duration of the induced episodes of AF became longer, after 24
hours of maintained AF in some goats it was no longer possible to
measure the adaptation curve of the atrial refractory period. In 8
goats in which the changes in atrial refractoriness could be followed
for a longer period of time (2 to 4 days) the atrial refractory period
shortened further to 81±22 during slow pacing and 90±16 during
fast
pacing (Fmax)(Table 3).
|
Spatial Dispersion of Atrial Refractoriness
To assess whether
an increased spatial dispersion in atrial
refractoriness may play a role in the increased stability of
fibrillation, in four goats the differences in effective refractory
period between the right and left atrial appendage was measured during
the first two days of AF. In Fig 12, the differences in
AERP are given during pacing at various intervals during control (sinus
rhythm) and after 24 and 48 hours of AF. During sinus rhythm the
average differences between right and left atrial refractory period
during pacing with intervals of 400, 300, 250, and 200 ms, were 14, 22,
20, and 16 ms, respectively. After 48 hours of AF these values were 8,
8, 6, and 8 ms. This decrease in the spatial difference in atrial
refractoriness may be explained by the general shortening of the atrial
refractory period during the first 48 hours of AF.
|
In addition, in 6
goats the spatial differences in median atrial
fibrillation interval at 10 to 13 atrial sites were measured after
respectively 1 and 14 days of maintained AF. An example of the spatial
distribution of median AF intervals is given in Fig 13
(goat No. 7). In this case the fibrillation interval was measured at 12
sites located at the right and left atrial free wall, the right and
left atrial appendages and the bundle of Bachmann. After 24 hours of AF
the largest difference in fibrillation interval was 29 ms. At the
bundle of Bachmann the median AF interval was 135 ms compared with 106
ms at the free wall of the left atrium. After 2 weeks of AF the spatial
dispersion in AF interval was 35 ms (131 ms at Bachmann's bundle and
96 ms at the left atrial free wall). In all 6 goats the longest
fibrillation interval was mostly found at Bachmann's bundle. After 24
hours of maintained AF in all 6 goats the shortest fibrillation
interval was found in the left atrium. Differences in fibrillation
intervals between the left and the right atrium varied between 4 and 24
ms. After 2 weeks of AF still in 4 of 6 goats the shortest AF interval
was found in the left atrium (differences between right and left
atrium, 11 to 15 ms). In the other 2 goats AF intervals were similar in
both atria. In Table 4, the data of all goats are
listed. No statistically significant differences in spatial
distribution of the median fibrillation intervals were found between 24
hours and 2 weeks of atrial fibrillation. Thus, the measurements of
spatial differences in refractory period as well as the spatial
distribution of median AF intervals, do not support the hypothesis that
the increase in stability of AF is due to an increased spatial
dispersion in atrial refractoriness.
|
|
Inducibility of Atrial Fibrillation
The inducibility of
atrial fibrillation by single premature
stimuli was tested in 11 goats at a total of 17 pacing sites (1 to 2
sites in each goat). During control, at 4 of 17 sites (24%) (in 3 of
11 goats) the application of a single early premature stimulus induced
short paroxysms of atrial fibrillation. After 24 hours of maintained
AF, single premature stimuli produced atrial fibrillation at 13 of 17
sites (76%; P<.01) (9 of 11 goats) (Table 5). Thus
already after 24 hours of AF the vulnerability
of the atria to fibrillation was clearly increased.
|
After a total period of 19±5 days of maintained AF, in 5 goats the inducibility of AF was tested again after conversion to sinus rhythm. 6 Hours after conversion the inducibility of AF was still very high (100%). However after 24 hours the vulnerability to AF had already clearly decreased (43%), whereas after 1 week of sinus rhythm the inducibility of AF was comparable to control (29%).
Reversibility of Fibrillation-Induced
Electrophysiological Changes
In 5 goats in which atrial fibrillation
was maintained for 2 to 4
weeks the reversibility of the
electrophysiological changes by atrial
fibrillation was studied. In 4 goats sinus rhythm restored
spontaneously, whereas in 1 goat (goat No. 6) AF was terminated by
intravenous infusion of Cibenzoline (Cipralan 1.5 mg/kg).
In Table 6, the duration and interval of the paroxysms
of electrically induced atrial fibrillation are given together with the
intra-atrial conduction velocity and refractory period as measured
during control (before AF was chronically maintained) and 6 hours, 24
hours, 1 week, and 2 weeks after conversion of AF to sinus rhythm.
|
|
Six hours after conversion to sinus rhythm the median duration of electrically induced paroxysms of atrial fibrillation was already back to normal and lasted only 7±2 seconds. Also the atrial fibrillation interval was significantly prolonged from 105±10 to 139±7 ms. However in all goats some of the induced episodes of AF were still long-lasting (the 95th percentile of AF duration was 49.1 minutes compared with 13 seconds during control). In 2 goats, an episode of AF was induced still lasting longer than 1 and 6 hours respectively. After 24 hours of sinus rhythm only short-lasting episodes of AF could be induced, terminating spontaneously within 6±4 seconds. After 24 hours also the fibrillation interval was normalized to 151±25 ms.
Six hours after restoration to sinus rhythm, the AERP400 and the AERP200 were still shorter than during control, the atrial refractory period still showing a clear reversed adaptation to heart rate. Twenty-four hours after conversion the AERP200 had returned to control values, but at slower pacing rates (AERP400) the refractory period was still shortened. After 1 week of sinus rhythm the rate adaptation of the atrial refractory period was fully normalized and also at slow heart rates the refractory period was normal again.
After conversion of AF to sinus rhythm the conduction velocity along Bachmann's bundle remained slightly slower than it was before chronic atrial fibrillation.
Although from these observations no exact time constant for the reversibility of the various electrophysiological changes can be derived, they show that after cardioversion of AF, all electrophysiological changes induced by atrial fibrillation are completely reversible within a few days.
|
Discussion |
|---|
|
TopAbstractIntroductionMethods Results Discussion References |
|---|
Mechanisms of Chronic Atrial Fibrillation
Burst pacing between two closely spaced atrial electrodes induces atrial fibrillation which in normal goats terminates spontaneously within a few seconds. In this study we showed that, when atrial fibrillation is maintained artificially, the duration of these paroxysms progressively increases, until after 1 to 3 weeks atrial fibrillation becomes sustained. Around 1960, Moe and coworkers24 25 26 introduced the multiple wavelet hypothesis, postulating that the persistence of atrial fibrillation depends on the average number of wavelets being present in the atria. If the number of wavelets is high, the statistical probability that they will all extinguish at the same time will be small and atrial fibrillation will persist. On the other hand, when only a small number of wavelets is present, the chance that they will die out simultaneously is higher and atrial fibrillation will self-terminate. In 1985, Moe's multiple wavelet hypothesis was experimentally confirmed by mapping of the atrial activation pattern of atrial fibrillation in isolated canine hearts and the critical number of wavelets required to sustain AF was estimated to be between four and six.27 More recently high density mapping has been applied clinically and the presence of random reentry28 of multiple wandering wavelets during electrically induced AF could also be demonstrated in man.22 29
The number of wavelets that can coexist in the atria is determined both by the atrial tissue mass (or surface area), and the wavelength of the atrial impulse.30 31 It is well known that in larger hearts atrial fibrillation is more stable and of longer duration,32 and that in humans atrial dilatation is an important risk factor for atrial fibrillation.33 This can be easily understood by realizing that the number of circuits in the atria increases with the square of the atrial diameter and that in larger mammals the wavelength of the atrial impulse does not increase proportionally to the size of the atria.34 The well known profibrillatory effects of vagal stimulation, acetylcholine and adenosine are generally ascribed to a shortening of the atrial action potential without a noticeable effect on conduction velocity. Since the wavelength of atrial refractoriness is the product of conduction velocity and refractory period, the wavelength will become considerably shorter, thus allowing more wavelets to coexist in a given tissue mass. On the other hand drugs that prolong the wavelength of the atrial impulse have been shown to prevent or terminate atrial fibrillation.31 35 36 In patients with atrial fibrillation the wavelength may be shortened, at least locally, by the existence of intra-atrial conduction defects. Several clinical studies have indicated that long or biphasic P waves, late potentials, fragmented atrial electrograms, or increased conduction delays of premature beats are all associated with a higher vulnerability to atrial fibrillation.37 38 39 40 41 42 43 44 45 46 In addition to this pathophysiological triad of chronic fibrillation (atrial dilatation, shortened refractoriness and depressed conduction), also increased heterogeneity, either in intra-atrial conduction (enhanced nonuniform anisotropy,47 48 locally depressed action potentials49 50 51 ), or in recovery of excitability (increased spatial dispersion in atrial refractory periods)52 53 54 55 may be of crucial importance.
Shortening of Atrial Refractoriness by Atrial
Fibrillation
In the present study we found no changes in atrial
conduction
velocity during the first days of atrial fibrillation. In contrast
marked changes were observed in the atrial refractory period which,
during the first 24 hours of fibrillation, depending on the pacing
rate, shortened by as much as 12% to 35%. Since the conduction
velocity was not affected, the wavelength of the atrial impulse must
have been shortened by a similar amount. This progressive shortening of
the wavelength by atrial fibrillation provides a good explanation for
the observed stabilization of AF with time (domestication of AF).
Previous studies have shown that a high correlation exists between the
refractory period as measured by programmed electrical stimulation and
the average cycle length during fibrillation, suggesting that the local
fibrillation interval might be used as an index of local
refractoriness.53 56 57 The advantage of
such an index of
local refractoriness is that one can follow the time course of changes
in refractory periods during fibrillation without the need for
extensive pacing protocols, which are not only time limited but may
also disturb the experimental conditions. The relationship between the
atrial refractory period and the local fibrillation interval was
confirmed in our present study by the observation that the
progressive shortening of the fibrillation interval during the
development of chronic fibrillation was associated with a concomitant
shortening of the atrial refractory period as measured by programmed
electrical stimulation. However we want to emphasize that this does not
mean that the median fibrillation interval is equal to the local
refractory period. On the contrary we believe that during AF the
refractory period is somewhat shorter than the median fibrillation
interval and that a small partially excitable gap exists during atrial
fibrillation.58 59 Nevertheless, with these
limitations in
mind, in our opinion the median fibrillation interval can be used to
estimate changes in atrial refractory period by atrial fibrillation.
During the first 24 hours of atrial fibrillation the median cycle
length of AF progressively shortened with a time course of about 2 ms
per hour. During the following days this shortening of the atrial
"refractory period" continued at a much slower rate of only a few
milliseconds per day. A steady state in atrial fibrillation interval
was reached after about 6 days of AF.
Correlation Between Atrial Fibrillation Cycle Length and Stability
of AF
Immediately after the fibrillation pacemaker was turned on and
atrial fibrillation was maintained artificially, the refractory period
of the atria slowly started to shorten. Initially, this did not
increase the stability of fibrillation and the episodes of induced AF
remained short and self-terminating. The atrial refractory period
had to be shortened to a certain critical value before atrial
fibrillation got more stable and the paroxysms of AF started to last
for a longer period of time. On the average, atrial fibrillation
started to last longer after the median fibrillation interval had
shortened to about 120 ms. The median fibrillation interval continued
to shorten up to 99±10 ms during sustained atrial fibrillation (see
Table 1 and Fig 8). At this point we can only
estimate the critical
changes in wavelength associated with (or as we think partly
responsible for) the development of chronic AF. If we assume that the
atrial refractory period is 10% shorter than the median fibrillation
interval, and if we take an average conduction velocity of 61 cm/s as
recently measured during type I AF22 the wavelength during
the short episodes of acute AF would be in the order of 8 to 9 cm. The
critical wavelength at which AF starts to prolong would then be about 6
to 7 cm, whereas during chronic atrial fibrillation the wavelength
should be 5 to 6 cm. However when atrial conduction during chronic
atrial fibrillation in reality is slower than 61 cm/s, the wavelength
actually might be smaller. At a conduction velocity of 50 cm/s the
wavelength during sustained AF would be 4 to 5 cm, whereas at 40 cm/s
the multiple wavelets would be as short as 3 to 4 cm. At such short
wavelengths the diameter of intra-atrial reentrant circuits could
be as small as 1 cm.
There is reason to believe that besides the shortening of refractoriness also other factors may play a role in the development of chronic fibrillation. This is supported by the observation that the time course of changes in atrial refractoriness does not completely run parallel with the time course of development of sustained fibrillation. Whereas the median fibrillation interval usually already reached a steady state within a couple of days, it often took an additional 1 to 2 weeks for atrial fibrillation to become persistent. Possible candidates of additional changes in the atria requiring a longer time period to develop, might be atrial dilatation,60 a general depression of atrial conduction velocity, or the development of local areas of structural intra-atrial conduction block.37
Maladaptation of the Atrial Refractory Period
In 1982, Attuel
et al61 62 measured the atrial
refractory period in 39 patients during pacing at three or more basic
cycle lengths. They found that in patients in which sustained atrial
tachyarrhythmias could be provoked with 1 to 3 premature
stimuli, the atrial refractory period either failed to adapt or adapted
poorly to changes in heart rate. On the basis of these observations
they suggested that a poor or absent rate adaptation of the atrial
refractory period may constitute a clinical entity and might be a
marker of atrial pathology causing a propensity to atrial fibrillation.
These observations were extended by Le Heuzey et al who measured the
effects of changes in heart rate on the duration of the action
potential recorded from isolated strips of human atrial
myocardium.54 55 From these studies it was
suggested that a maladaptation of refractoriness might be the cause of
atrial fibrillation in humans. In our present study we made a
similar observation that maintenance of AF was associated with
maladaptation of the atrial refractory period to changes in heart rate.
While normal goats in sinus rhythm showed a clear shortening of the
atrial refractory period at shorter pacing intervals, goats which had
been artificially kept in atrial fibrillation, after one or more days
lost this physiological adaptation and showed
either a constant duration of the refractory period at different pacing
rates or an inverse adaptation curve, ie, instead of lengthening, the
atrial refractory period actually now shortened at slower heart rates.
After cardioversion to sinus rhythm the normal adaptation to changes in
heart rate was restored within a couple of days. From these experiments
thus it seems that the maladaptation of the atrial refractory period
rather is the result of atrial fibrillation than the cause of it.
However it can not be excluded that the changes in rate adaptation of
the refractory period is one of the factors that cause atrial
fibrillation to become sustained.
What Causes the Shortening of the Refractory Period During
AF?
The mechanisms of the shortening of atrial refractoriness by AF
are as yet unclear and require further study. Possible causes are (1)
long-term changes in activity or sensitivity of the autonomic
nervous system, (2) stretch of the atrial wall due to the increased
intra-atrial pressure, (3) ischemia of the atrial
myocardium, (4) an increase in plasma atrial
natriuretic factor (ANF) levels, and (5) the high rate of
electrical activation of the atrial cells per se.
Several studies have emphasized the importance of the autonomic nervous system for the initiation and perpetuation of atrial fibrillation. Coumel et al63 have described two different subgroups of patients with atrial fibrillation. In one group the initiation of atrial fibrillation was dependent on a high vagal tone, whereas in another group the occurrence of atrial fibrillation seemed to be related to adrenergic stimulation of the heart.64 Indeed it is well known that a high vagal tone or the administration of acetylcholine is profibrillatory because it shortens the atrial action potentials and the wavelength31 due to activation of the IKAch channel.
There is disagreement in the literature about the effect of stretch on the refractory period, some studies reporting a shortening whereas others have found no change or even a lengthening of the refractory period.65 66 67 68 69 70 So far, nobody has measured the effect of prolonged changes in atrial wall stress on atrial refractoriness. Therefore, although a greater than twofold acute increase in atrial pressure was found to have no effect on the human atrial refractory period, it remains to be seen whether a chronic increase in atrial pressure above a certain value does exert important electrophysiological changes.70
When the atria become ischemic, activation of ATP-regulated potassium channels may result in a shortening of the atrial action potential. In 1982 White et al71 showed that immediately after induction of AF both atrial perfusion and oxygen consumption rise sharply. The oxygen consumption increased more than threefold while the blood supply increased with a factor of 2 to 3, actually resulting in a higher flow per gram in the fibrillating atria than in the pumping left ventricle. Since the reactive hyperemia response was significantly attenuated and in some dogs nearly abolished, the flow reserve during atrial fibrillation is clearly decreased. A further increase in atrial metabolism, for instance by adrenergic stimulation, could lead to a further increase in oxygen demand which now can no longer be met by the already maximally dilated coronary arteries. Whether AF actually causes atrial ischemia is at present unknown.
The increase in the production of ANF by the atrial cells when the atria are thrown into fibrillation is well documented.72 Recently, Stambler et al have demonstrated that the infusion of ANF in dogs may give rise to a shortening of the AERP and the monophasic action potential.73 If in the goat the plasma concentration of ANF increases by atrial fibrillation and if the ANF levels become high enough to shorten the AERP, this mechanism might be involved in the process of domestication of AF.
The fifth most intriguing possibility that the shortening of atrial refractoriness is mediated by the high rate of depolarization itself will be discussed below.
T Wave `Memory'
In 1982, Rosenbaum et al
described that rapid atrial or
ventricular pacing in humans could induce T wave changes
which developed to a maximum in about 24 hours of pacing. Because
repeated rapid pacing caused the repolarization changes to appear after
a shorter period of time, they concluded that "the myocardial cells
involved in this process seem to keep a `memory' of the previous
effect... ."74 As reviewed by Katz,75
primary repolarization abnormalities can be caused by three
fundamentally different causal mechanisms taking place at different
levels in the heart: altered structure (organ), altered
metabolism (cell), or altered ion channels (genes). A
distinction between these three different mechanisms can be made on the
basis of the time course and reversibility of the changes. While
structural changes are generally irreversible and may take weeks or
even years to develop, metabolic changes occur virtually
instantaneously and are rapidly reversible. It has been postulated that
the changes in repolarization referred to as cardiac "memory,"
which develop more slowly and persist longer than the transient changes
mediated by changes in cellular metabolism but are still
reversible and not associated with obvious organ damage, "arises
from molecular replacements involving the channel proteins of the
heart's plasma membrane."75 Similar to the
posttachycardia T wave changes, the shortening in atrial
refractory period by atrial fibrillation might be based on alterations
in synthesis and assembly of the potassium channels that control atrial
repolarization. As shown by Agnew76 and
Aldrich77 cells posses the ability to synthesize a
rich variety of potassium channels by "mixing and matching"
different subunits which can be expressed by a large family of genes.
Recent studies of Rosen et al78 79 have demonstrated
that
in the ventricles cardiac memory was abolished by
4-aminopyridine, which blocks both the transient
outward potassium current (Ito) as well as
IK. In nerve cells the mechanism of memory has been shown
to be caused by second messager activation of protein kinases which
modify ion channel functions of the cell membrane.80
Although it is not yet known whether prolonged alteration of the atrial
rate and activation sequence modulates protein synthesis and how this
could change structure and/or function of potassium channels, the idea
that the development of chronic atrial fibrillation may be based on
changes in gene expression is an intriguing one and certainly merits
more detailed studies using molecular biology techniques.
Other Models of Sustained Atrial Fibrillation
In 1985,
Salmon81 reported that atrial pacing (60 Hz)
for more than 90 days resulted in the development of progressive left
atrial enlargement and persistent AF in 4 of 6 dogs. However, in this
study no changes in atrial
electrophysiological properties, like
intra-atrial conduction velocity or refractory period, were
studied.
Just recently, Morillo et al82 published a canine study in which 6 weeks of continuous rapid atrial pacing (400/min), produced sustained atrial fibrillation (defined as AF lasting >15 minutes) in 82% of the animals. At pacing intervals of 400 and 300 ms the atrial refractory period had shortened from 150±8 to 127±10 ms (-15%) and from 147±11 to 123±12 ms (-16%), respectively. Together with a marked increase in atrial size this shortening of the atrial refractory period yielded a positive predictive value of 88% for the induction of sustained AF. In our study we found an even more marked shortening in atrial refractoriness measured at a wider range of pacing intervals. Already after 2 to 4 days of atrial fibrillation the refractory period during the maximal pacing rate still eliciting a 1:1 response (Fmax), had shortened from 123±13 to 90±16 ms (-27%). During slow pacing (400 ms interval) the shortening of AERP was as much as 45%, from 146±19 to 81±22 ms. Due to the more pronounced shortening of the AERP during slow heart rates, the normal rate adaptation of the refractory period was inverted and instead of getting longer, now the refractory period actually got shorter as a response to slowing of the heart rate.
Both clinical and experimental studies have shown that within the first week after open heart surgery there is a high incidence of atrial tachyarrhythmias due to the development of a sterile pericarditis.23 83 84 In 1986 Pagé et al83 described a new model of atrial flutter in dogs, in which a sterile pericarditis was deliberately produced by dusting generous amounts of talcum powder on the atria and by leaving a gauze on the free wall of the atria. During the first week following this procedure, episodes of atrial flutter could be induced reproducibly by programmed electrical stimulation. More recently the same technique has been used by Ortiz et al to produce atrial fibrillation in dogs with sterile pericarditis.85 In our study the animals also underwent a thoracotomy and multiple electrodes were sutured to the atria probably causing a sterile pericarditis. To avoid electrophysiological changes due to pericarditis the goats were allowed to recover from surgery for 2 weeks. The experimental protocol was started after an additional control period of 1 week during which the electrophysiological measurements of conduction velocity and refractory period were stable and long-lasting episodes of atrial fibrillation could not be induced. Therefore we believe that despite the presence of chronically implanted electrodes pericarditis did not play a major role in our present model of AF. This is further supported by the observation that all changes associated with maintained AF were found to be completely reversible after conversion to sinus rhythm.
Clinical Implications
The concept that "atrial
fibrillation begets atrial
fibrillation" might have some important clinical implications. First
of all it emphasizes that most of our
electrophysiological knowledge stems from
acute experiments and that we know relatively little about chronic
electrophysiological adaptation processes.
If it is true that the long-term shortening of atrial
refractoriness during fibrillation is based on a fundamental change in
composition of the ion channels responsible for repolarization of the
atrial cells, the action of antiarrhythmic drugs on fibrillating atria
may be different than the effects as measured during sinus rhythm. The
clinically observed diminished efficacy of chemical cardioversion after
a prolonged period of atrial
fibrillation9 10 11 12 might
be
explained by such a process of electrical remodeling. In fact it might
be imperative to reevaluate the effects of existing
anti-fibrillatory drugs in chronically fibrillating hearts. On the
other hand it opens the possibility to develop new drugs specifically
targeted at those ion channels that become expressed during atrial
fibrillation. At this moment however, these implications are still
speculative and more information is needed about the ionic mechanisms
of the fibrillation-induced shortening of repolarization before any
firm conclusions can be drawn.
The observed anomalous rate adaptation of the atrial refractory period may play an important role in the recurrences of AF which are so frequently seen clinically during the first week after electrical or chemical defibrillation.14 18 Directly after cardioversion the atrial interval suddenly prolongs from about 100 to 150 ms during atrial fibrillation to about 1000 ms during sinus rhythm. When the atrial refractory period fails to adapt to such a sudden slowing in heart rate by a prolongation of the refractory period, or even worse, when it becomes shorter due to an inversed rate adaptation, after conversion to sinus rhythm the atria will be left with a dangerously short refractory period. Without the natural protection of a long refractory period, the atrial wavelength will be very short and on first occasion an atrial premature beat may start fibrillation again. In the goat the shortening of the atrial refractory period and the maladaptation to heart rate was reversible within the first days of sinus rhythm. If this is also true in humans, protection against the fibrillation-induced maladaptation of the refractory period during the first week after conversion might help to prevent early recurrences of AF.
The question remains whether electrical remodeling by AF also occurs in humans and if so, how this process of electrical remodeling would interfere with atrial fibrillation. As pointed out above the shortening of the atrial refractory might explain the diminished success rate of chemical and electrical cardioversion in patients with long-lasting atrial fibrillation. And indeed the finding that the shortening of atrial refractoriness needs a few days to revert completely could explain the early recurrences seen after cardioversion. However the complete reversion of the electrophysiological changes within 1 week after restoration of sinus rhythm implies that the role of electrical remodeling in patients with paroxysmal atrial fibrillation with an incidence of less than once a week seem limited. Due to the reversibility of electrical remodeling of AF each paroxysm of AF is independent of the previous one. In general however our study implicates that the best prevention of atrial fibrillation is to terminate the arrhythmia as soon as possible, thus interrupting the electrophysiological sequelae which will lead to chronic atrial fibrillation.
Limitations of the Study
One of the limitations of this study
is that we could not follow
the changes in atrial refractory period and conduction velocity for
much longer than the first days of fibrillation. Due to the increased
vulnerability of the atria, by that time programmed electrical
stimulation induced periods of AF lasting for such a long time that it
became impossible to complete the protocol. For the same reason it was
impossible to measure the exact time course of reversibility of the
shortening of the refractory period. Directly after cardioversion the
atria are still so vulnerable that the administration of premature
stimuli will reinduce long-lasting episodes of atrial fibrillation,
which obviously interrupt the reversibility process. Therefore, our
reversibility measurements are limited and do not allow accurate
quantitative conclusions. Nevertheless, the data that could be
collected leave little doubt that the shortening of refractoriness and
the maladaptation to rate are both completely reversible. After 1 week
of sinus rhythm the atrial refractory period and the duration of
induced paroxysms of AF were normal again. Because of this, each animal
served as its own control and it was not necessary to include a control
group of sham operated animals. However, after conversion to sinus
rhythm the intra-atrial conduction velocity did not return to
control values. The lack of a control group of sham operated animals
makes it hard to decide whether the observed slowing in atrial
conduction velocity is caused by chronic atrial fibrillation or is a
long-term effect of the presence of the implanted electrodes.
The observations reported in this paper raise many questions that cannot be answered at the present time. Additional experiments will be needed to determine the possible role of neurohumoral changes, atrial dilatation, ischemia, ANF, and to determine the ionic channels responsible for the fibrillation-induced shortening of the action potential. It also remains to be determined whether on the long term, structural changes of the atrial wall and increased heterogeneity of excitability, refractoriness and conduction properties contribute to the development of chronic AF. Despite all these limitations, our study indicates that the concept of "domestication of atrial fibrillation" may have a clear pathophysiological basis that seems worthwhile to explore. More specifically, our study indicates that the AF-induced electrophysiological changes should also be studied in humans, and that the reversibility of these changes after conversion to sinus rhythm should be monitored.