Download Print - Circulation

413
Ventricular Proarrhythmic Effects of Ventricular
Cycle Length and Shock Strength in a Sheep
Model of Transvenous Atrial Defibrillation
Gregory M. Ayers, MD, PhD; Clif A. Alferness, BSEE; Marina Ilina, PhD;
Darrell 0. Wagner, ASET; William A. Sirokman, BS;
John M. Adams, BSEE; Jerry C. Griffin, MD
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Background Synchronized cardioversion is generally accepted as safe for the treatment of ventricular tachycardia and
atrial fibrillation when shocks are synchronized to the R wave
and delivered transthoracically. However, others have shown
that during attempted transvenous cardioversion of rapid
ventricular tachycardia, ventricular fibrillation (VF) may be
induced. It was our objective to evaluate conditions (short and
irregular cycle lengths [CL]) under which VF might be induced
during synchronized electrical conversion of atrial fibrillation
with transvenous electrodes.
Methods and Results In 16 sheep (weight, 62±7.8 kg), atrial
defibrillation thresholds (ADFT) were determined for a 3-ms/
3-ms biphasic shock delivered between two catheters each
having 6-cm coil electrodes, one in the great cardiac vein
under the left atrial appendage and one in the right atrial
appendage along the anterolateral atrioventricular groove. A
hexapolar mapping catheter was positioned in the right ventricular apex for shock synchronization. In 8 sheep (group A),
a shock intensity 20 V less than the ADFT was used for testing,
and in the remaining 8 sheep (group B), a shock intensity of
twice ADFT was used. With a modified extrastimulus technique, a basic train of eight stimuli alone (part 1) and with
single (part 2) and double (part 3) extrastimuli were applied to
right ventricular plunge electrodes. Atrial defibrillation shocks
were delivered synchronized to the last depolarization. In part
4, shocks were delivered during atrial fibrillation. The preced-
ing CL was evaluated over a range of 150 to 1000 milliseconds.
Shocks were also delayed 2, 20, 50, and 100 milliseconds after
the last depolarization from the stimulus (parts 1 through 3) or
intrinsic depolarization (part 4). The mean ADFT for group A
was 127±48 V, 0.71±0.60 J and for group B, 136+37 V,
0.79±0.42 J (NS, P>.15). Of 1870 shocks delivered, 11 episodes of VF were induced. Group A had no episodes of VF in
part 1, two episodes of VF in part 2 (CL, 240 and 230
milliseconds with 2-millisecond delay), and one episode each
in parts 3 (CL, 280 milliseconds with 2-millisecond delay) and
4 (CL, 240 milliseconds with 100-millisecond delay). Group B
had two episodes in part 1 (CL, 250 and 300 milliseconds with
20-millisecond delay), three episodes in part 2 (CL, 230, 230,
and 250 milliseconds with 2-millisecond delay), and one episode each in parts 3 (CL, 260 milliseconds with 2-millisecond
delay) and 4 (198 milliseconds with 100-millisecond delay). No
episodes of VF were induced for shocks delivered after a CL
>300 milliseconds.
Conclusions Synchronized transvenous atrial defibrillation
shocks delivered on beats with a short preceding ventricular
cycle length (<300 milliseconds) are associated with a significantly increased risk of initiation of VF. To decrease the risk
of ventricular proarrhythmia, short CLs should be avoided.
(Circulation. 1994;89:413-422.)
Key Words * atrial fibrillation * ventricles * fibrillation
* defibrillation * heart rate
t has been recognized since the early 1960s that
electrical reversion of cardiac arrhythmias is effec_ tive.12 Equally well known are the ventricular
arrhythmias, tachycardia and fibrillation, associated
with cardioversion of atrial arrhythmias.2-5 These proarrhythmic side effects are attributed to vulnerable-period
Although the timing relation and duration are somewhat
disputed, shocks delivered during some portion of the T
wave consistently result in ventricular fibrillation.3-5 The
atrial vulnerable period has also been documented. During normal sinus rhythm, the atrial vulnerable period is on
the downslope of the R wave of the surface ECG.'
With the advent of the R-wave synchronized defibrillator, the risk of ventricular arrhythmia induction during
cardioversion of atrial fibrillation was lessened. Recent
data have shown that atrial fibrillation is associated with
a 2.2% to 2.5% risk of arrhythmic death without concomitant antiarrhythmic therapy6'7 and 5% in patients
treated with antiarrhythmic agents.6 It is therefore
apparent that cardioversion of atrial fibrillation to sinus
rhythm is of significant importance and that all efforts
must be expended to minimize risk.
Although isolated empirical observations have been
made with respect to the induction of ventricular fibrillation during synchronized, transvenous cardioversion
of atrial fibrillation,8-10 no study has been undertaken to
systematically model and evaluate the conditions under
stimulation.
The existence of two vulnerable periods has been well
documented for the heart.1-5 During these periods, either
the atria or ventricles are susceptible to fibrillation. The
vulnerable period is directly related to the relative refractory period during which the heart is susceptible to
fibrillation induction by a stimulus. For ventricular myocardium, this period is near the apex of the T wave.
Received June 11, 1993; revision accepted August 3, 1993.
From InControl, Inc, Redmond, Wash (G.M.A., C.A.A., M.I.,
D.O.W., W.A.S., J.M.A., J.C.G.) and University of California San
Francisco (J.C.G.).
Correspondence to Gregory M. Ayers, MD, PhD, Senior Scientist, InControl, Inc, 6675 185th Ave NE, Suite 100, Redmond, WA
98052.
414
Circulation Vol 89, No 1 January 1994
4
Lk
-- 6CM.
-
~ ~
~
_
.
m
~
A~:_: aW
:E
h
p
RPLt
* UW
X
(a)
|4 -
6CM.
- -
(b)
FIG 1. Diagrams of the two defibrillation catheters used in this
study. a, Catheter used in the great cardiac vein; b, catheter
used in the right atrial appendage. Both were modified 6F
hexapolar catheters with a 6-cm silver-plated stainless steel
electrode coil attached to the sixth electrode ring (a) or the distal
tip electrode (b).
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
which ventricular arrhythmias and fibrillation can be
induced during synchronized, low-energy, transvenous
cardioversion of atrial fibrillation. It was our objective to
evaluate conditions (short and irregular cycle lengths)
under which ventricular arrhythmias and fibrillation
might be induced during synchronized electrical conversion of atrial fibrillation with a transvenous electrode
system that has been shown"": 2 to be effective for atrial
defibrillation.
Methods
This study involving experimental animals conformed to the
positions of the American Heart Association on research
animal use adopted November 11, 1984, by the American
Heart Association.
Animal Preparation
Sixteen sheep (weight, 62±7.8 kg) of random breed and sex
were sedated with xylazine 0.2 mg/kg IM and morphine 10.0
mg/kg IM. Anesthesia was induced with sodium pentobarbital
25 mg/kg IV. Additional pentobarbital was used as needed to
maintain a constant plane of anesthesia. Additional morphine
5.0 mg/kg IM was given every 4 hours. A femoral artery was
cannulated to monitor arterial blood pressure and the femoral
vein for venous access for drug delivery. Femoral arterial
blood pressure and lead II ECG were monitored throughout
the study. Arterial blood gases were determined every 30
minutes, with adjustments made to ventilation or intravenous
sodium bicarbonate given on the basis of the results. Serum
electrolytes, including potassium, calcium, and magnesium,
were determined from blood drawn each hour, with supplemental electrolytes added to the intravenous fluids to maintain
normal serum levels. Sodium heparin 10 000 U IV was given
after initial surgical preparation. Succinyleholine 0.6 mg/kg IV
was administered before each shock delivery sequence.
Surgical Preparation
The thoracic cavity was opened through a right lateral
thoracotomy at the fourth interspace with the pericardium
isolated and opened. The heart was cradled in the pericardium. Three 9F catheter sheaths (USCI, Billerica, Mass) were
inserted, two into the right jugular vein and one into the left
jugular vein. A modified 6F hexapolar catheter (Elecath,
Rahway, NJ) with a 6-cm silver-plated stainless steel electrode
coil (Renton Coil Spring, Renton, Wash, and ASKO Selective
Plating, Seattle, Wash) attached to the sixth electrode ring
(Fig la) was inserted through one of the right jugular vein
sheaths, passed through the coronary sinus, and positioned in
the great cardiac vein (Fig 2). This catheter was advanced until
the catheter tip was in the anterior coronary vein. A second
modified 6F hexapolar catheter with a 6-cm coil electrode
attached to the distal tip electrode (Fig lb) was inserted
through the second jugular vein sheath on the right side and
advanced so that the tip was in the right atrial appendage (Fig
2). A third catheter, a 6F hexapolar mapping catheter, was
inserted through the sheath in the left jugular vein and
advanced to the right ventricular apex. Plunge electrodes were
placed into the right ventricular free wall (Fig 2).
Lead II surface ECG, systemic blood pressure, and bipolar
electrograms were recorded with a custom, isolated interface/
amplifier (InControl, Redmond, Wash), digital storage oscilloscope (Tektronix model 2232, Beaverton, Ore) and Teac
model RD130/T digital tape recorder (Teac America, Montebello, Calif) and digitized on a Macintosh TTfx computer
(Apple Computer, Cupertino, Calif) with custom LABVIEW
software (National Instruments, Austin, Tex). Bipolar electrograms were recorded from the great cardiac vein coil electrode
to right atrial coil electrode and right ventricular tip to a
proximal right ventricular electrode. Digital recordings of both
normal sinus rhythm and atrial fibrillation were made at the
FIG 2. Diagrams of frontal and lateral views of the heart showing the
locations of the three 6F catheter
locations, great cardiac vein (CS/
GCV), right atrial appendage (RA),
and right ventricular apex (RV). Also
shown are the two plunge electrodes
positioned on the right ventricular
free wall.
2)
RV Plunge
Electrodes
Ayers et al Rate Effect on Atrial Defibrillation Proarrhythmia
415
|3 ms.|
Vif
V 2i
V2j = -Vif
-
W
v3ms.
1
1
FIG 3. Diagram of the defibrillation shock waveform used in this
study. It is a single-capacitor 90-,uF biphasic waveform with both
phases having a 3-millisecond duration.
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
beginning of each study. Critical events were marked and later
retrieved from digital tape for analysis of electrograms.
After completion of the study, the animals were killed with
a premixed solution of pentobarbital, 6 grains/mL (Anthony
Products, Arcadia, Calif).
Atrial Defibrillation Threshold Determination
After surgical preparation, atrial defibrillation thresholds
were determined for a 3-ms/3-ms single-capacitor biphasic
shock (Fig 3).
Atrial fibrillation was electrically induced with 10-mA,
1-millisecond pulses at a frequency of 60 Hz with epicardial
wires attached to the right atrium. Before the first defibrillation attempt, atrial fibrillation was permitted to sustain for 5
minutes to demonstrate adequate stability. If atrial fibrillation
spontaneously terminated in less than 5 minutes, t3-methacholine, 50 ,ug/250 mL normal saline, was applied to the atrial
epicardium in a continuous infusion.13 Before each defibrillation shock, atrial fibrillation was induced and permitted to
continue for 30 seconds. For every atrial defibrillation attempt,
the great cardiac vein coil electrode served as the cathode and
the right atrial coil electrode as the anode, with respect to the
first phase of the shock. For each shock, a customized LABVIEW
software package running on a Macintosh IIfx computer
controlled the shock delivery from a custom defibrillator
(InControl) and acquired the delivered voltage and current
waveforms. From the acquired shock data, the software calculated the delivered energy and shock impedance. All shocks
were synchronized to the V wave sensed from the right
ventricular catheter.
Atrial defibrillation thresholds were determined with the
algorithm shown in Fig 4. Starting at a peak leading-edge
voltage of 80 V, atrial defibrillation shocks were delivered to
the catheter electrode system. The philosophy of this threshold
method was to determine the lowest-intensity shock that
would defibrillate the atria and then determine the percent
successful defibrillation at this intensity. If the shock failed to
defibrillate, the shock intensity was increased by 10 V. Once
the shock was successful, atrial fibrillation was induced, and 20
shocks were delivered at this intensity to determine the
percent success. After each successful shock, atrial fibrillation
was initiated, and 30 seconds was allowed to elapse before the
next shock was delivered. If the percent success was < 10%, the
percent success was determined at a shock voltage 10 V higher
than just tested. If the percent success was >90%, the percent
success was determined at an intensity 10 V less than the
intensity just tested. Thus, the threshold was considered valid
if the resulting percent success was between 10% and 90%.
Ventricular Vulnerability Testing
The 16 sheep in this study were divided into two groups for
study of the effect of shock strength on the likelihood of
Yes
|ADFT
|
FIG 4. Algorithm used to define atrial defibrillation thresholds in
this study.
ventricular arrhythmia induction. Group A received subthreshold atrial defibrillation shocks (20 V less than threshold), and group B received suprathreshold shocks (twice
threshold). Both groups underwent a four-part ventricular
vulnerability testing protocol.
Pacing studies, as with the defibrillation threshold testing,
were performed with custom LABVIEW software running on a
Macintosh IIfx platform. The software permitted standard
programmed stimulation as well as combined pacing and
defibrillation control (Fig 5). Pacing threshold current was
determined for the bipolar right ventricular plunge electrodes
with a 1-millisecond square wave at a cycle length of 300
milliseconds. When a cycle length of 300 milliseconds did not
consistently capture the ventricles because of the rate, the
cycle length was increased in steps of 50 milliseconds until
consistent capture of a train of eight was achieved. Intensity of
twice pacing threshold current was used for all stimuli in the
remainder of the study.
A modified extrastimulus technique was used for the first
three parts (parts 1 through 3) of the protocol. These three
parts were performed in random order. In part 4, shocks were
delivered during atrial fibrillation. Shocks were delivered
synchronously to depolarizations sensed from the right ventricular apex catheter in all parts. Shock delay at each testing
cycle length started with 100 milliseconds and was decreased
to 50, 20, and 2 milliseconds. Incidence of postshock sustained
ventricular arrhythmias was documented, and the digital tape
recording of electrograms was marked for later retrieval.
416
Circulation Vol 89, No 1 January 1994
FIG 5. Diagram of the interconnections between
the various devices used in this study and the
animal. RV indicates right ventricular apex recording catheter; GCV, great cardiac vein defibrillation
catheter; and RA, right atrial appendage defibrillation catheter. Enable and Si, S2, S3 are the
computer control lines for the custom defibrillator
and programmable stimulator, respectively.
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
delivered 100 milliseconds after the sensed depolarization of
the last S, from the right ventricular apex catheter. If ventricular fibrillation was not induced, the sequence of S, stimuli and
shock was repeated, but the delay was reduced to 50, then 20,
and last, 2 milliseconds. If ventricular fibrillation was not
induced, the cycle length of the train was decreased by 50
Part 1
This part of the protocol modeled rapid but regular cycle
lengths (Fig 6a). A train of eight stimuli (S,) was delivered to
the right ventricular plunge electrodes starting at a cycle
length of 600 milliseconds. An atrial defibrillation shock was
Sl
Si
s1
(a)
A.
1---r
do
L----a
EI
S1
-
Si
51
£I
Sli
S1
SHOCK
r_
fSe
-e#
51
n
C
dA..
51
sI
51
S1
S2
SHOCK
(b)
it
i
1-
Si
51
li
----i
51
----------------------
---L
m
li
51
51
CC
-
51
1l
S2
S3
VSHOCK
(c)
a
-------
a
a
--
---- -------
-1-
-1
---------------------
L%- .<z
Ventricular depolarizations in AF
1v
V
\v,
(d)
'
10r
5
2.170 i2so
tt
32
40
30
V
ISHOCK
36
a
I~~~~
350-400 msec window
FIG 6. Schematic of each part of the protocol. A basic train of eight, part 1 (a), single, part 2 (b), and double, part 3 (c) extrastimuli were
applied with a shock delivered after the sensed depolarization from the last stimulus. The shock was delayed 100 milliseconds after this
depolarization. The shock delay was then decreased to 50, 20, and then 2 milliseconds. The pacing intervals, S,-S, in part 1, S1-S2 in
part 2, and S2-S3 in part 3, were decremented and all shock delays retested. In part 4 (d), ventricular depolarizations were sensed and
cycle lengths calculated and compared with a 50-millisecond shock window. When a preceding cycle length was within the window, a
shock was delivered delayed from the ventricular depolarization. This delay was then reduced from 100 to 50, 20, and then 2
milliseconds.
Ayers et al Rate Effect on Atrial Defibrillation Proarrhythmia
milliseconds and the sequence of delays repeated. This continued either until ventricular fibrillation was induced or until
the cycle length reached the effective refractory period so that
all eight S, stimuli were not capturing the ventricle.
Part 2
This part of the protocol modeled rapid and regular cycle
lengths followed by a premature complex (Fig 6b). A train of
eight stimuli (S,) at the shortest cycle length found during
threshold testing to produce consistent capture (300 milliseconds, seven sheep; 350 milliseconds, eight sheep; and 400
milliseconds, one sheep) was delivered to the plunge electrodes, followed by a premature stimulus (S2). The S1-S2
interval started at 290 milliseconds, with the shock delayed 100
milliseconds from the sensed depolarization. As in part 1, the
delay was reduced to 50, then 20, then 2 milliseconds at a
constant S1-S2 interval, then the S1-S2 interval was decreased in
10-millisecond steps and the sequence of delays repeated at
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
each S1-S2 interval either until ventricular fibrillation was
induced or until the S2 failed to capture the ventricle. An
epinephrine drip (2 mg/250 mL normal saline) was then
infused at a rate resulting in a mean blood pressure rise to
twice the preinfusion baseline value to further shorten ventricular refractoriness. The delay scanning and 10-millisecond
S1-S2 interval decreases were resumed until the S2 again failed
to initiate a depolarization or until ventricular fibrillation was
induced.
Part 3
This part of the protocol modeled a series of short intervals
followed by a long interval, at least twice the cycle length of the
basic train, then a short interval with a shock after the last
depolarization (Fig 6c).
A basic train of eight stimuli (S,), with a cycle length equal
to the train used in part 2, was delivered 10 times to determine
the time from end of the train to the first intrinsic ventricular
beat (V2). The minimum S,-V2 interval was then compared
with the drive train S,-S, interval, and if it was found to be at
least twice the S,-S, interval, this part of the study was
performed. A premature stimulus (S2) was added to the basic
train of eight with an S1-S2 interval equal to 50 milliseconds
less than the minimum S,-V2 interval. A second premature
stimulus (S3) was added to the pacing train at an initial S2-S3
interval of 290 milliseconds. Shocks were again delivered
delayed by 100, then 50, 20, and 2 milliseconds from the last
depolarization from the pacing stimuli. The S2-S3 interval was
decreased by 10 milliseconds, testing each delay in decremental manner either until S3 failed to capture the ventricle or
until ventricular fibrillation was induced.
Part 4
In this part of the protocol, shocks were delivered during
electrically induced atrial fibrillation. The defibrillation control software for shock delivery selected only ventricular
depolarizations whose preceding cycle lengths fell into a
50-millisecond window (Fig 6d). We tested 50-millisecond
windows from 150 to 200 milliseconds up to 950 to 1000
milliseconds.
Atrial fibrillation was induced and permitted to sustain for 5
minutes. The customized defibrillation control software was
configured to deliver a 100-millisecond delayed shock on the
first preceding cycle length that fell into a window of 950 to
1000 milliseconds. The other three delays were tested, 50, then
20, then 2 milliseconds, for the same cycle length window. The
window was then shifted down in steps of 50 milliseconds, with
all delays tested at each window of cycle lengths. A time limit
of 2 minutes was placed on the length of time the defibrillator
would wait for a given window. If no intervals were found
during this 2-minute period, the window was shifted by 50
milliseconds and testing was resumed. If the time-out period
was exceeded because of a lack of adequately short cycle
417
TABLE 1. Heart and Body Weight, Threshold Data and
Stimulation Data, and S, Cycle Length From the
Animals in This Study
Group A
Group B
Body weight, kg
58.4+4.9
64.6±9.1
Heart weight, g
361.9±57.1
399.3±120.9
ADFT, V
127.5±48.0
136.3±36.6
ADFT, J
0.71±0.60
0.79±0.42
Percent success at ADFT
52.0±31.3
48.1±29.3
S1 cycle length, msec
319±26
343±32
ADFT indicates atrial defibrillation threshold. With the exception of S, cycle length, there were no differences between
groups A and B (P>.1). Values are mean±SD.
lengths, an epinephrine infusion (2 mg/250 mL normal saline)
was started and delivered at a rate to double the mean
preinfusion blood pressure, and testing was resumed. Testing
continued until the time period at the window of 150 to 200
milliseconds expired or at a window and delay that generated
ventricular fibrillation. If ventricular fibrillation was induced,
the cycle length range was documented and the tape recording
of the electrograms was marked electrically. After the study,
electrograms were replayed to a Gould TA4000 paper recorder (Gould Inc, Cleveland, Ohio), and the preceding cycle
length was measured for the shocks that resulted in ventricular
fibrillation.
Statistical Analysis of Data
All statistical analyses were performed using JMP software
(SAS Institute, Cary, NC) running on a Macintosh Quadra 700
platform (Apple Computer). Continuous numerical data requiring comparison were subject to ANOVA for comparing
the effect of group differences. These data included body and
heart weights, atrial defibrillation thresholds, and stimulus
parameters. Comparisons in the episodes of ventricular fibrillation induced between groups and parts were made by x2
tests. Differences were considered significant at values of
P<.05.
Results
The mean atrial defibrillation threshold voltage and
energy are shown in Table 1. Also shown are the mean
S, interval used for the stimulus train in parts 2 and 3 of
the protocol. There was no significant difference in the
animal characteristic or threshold data when groups A and
B were compared, although the mean S1 for group B was
longer (P<.05). Not including the shocks used to determine threshold, 1870 test shocks were delivered in this
study. Eleven episodes (0.59%) of ventricular fibrillation
were induced in 8 of the 16 sheep with these shocks, 4
sheep in group A and 4 in group B. Table 2 details the
episodes of ventricular fibrillation induced. All of the
episodes of ventricular fibrillation were induced with a
preceding cycle length 5300 milliseconds.
In part 1, two episodes of ventricular fibrillation were
induced out of the 348 shocks delivered (0.57%). Both
episodes were induced for animals in group B, twice
atrial defibrillation threshold, although not statistically
significantly different from group A by a x2 test (P> .15).
Both episodes were induced with a right ventricular
depolarization-to-shock delay of 20 milliseconds. An
example of one ventricular fibrillation induction in this
part of the study is shown in Fig 7A.
418
Circulation Vol 89, No 1 January 1994
TABLE 2. Episodes of Ventricular Fibrillation, by Sheep and Protocol Part, With Corresponding
Preceding Ventricular Cycle Length and Shock Delay
Protocol Part
Sheep
Group A
1
2
3
4
5
6
7
8
1
2
3
4
S1-S1, Delay, ms
S1-S2, Delay, ms
S1-S2, 52-S3, Delay, me
Cycle Length, Delay, ms
n
n
n
n
n
n
n
n
240,2
x
740,280,2
x
n
n
x
x
n
n
n
n
n
n
n
n
240,100
n
n
n
n
n
n
n
198,100
x
n
n
n
n
n
n
n
n
n
n
n
230,2
n
n
Group B
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
9
10
11
12
13
14
15
16
n
x
300,20
250,2
n
n
n
n
n
n
250,20
230,2
550,260,2
x
n
230,2
Eleven episodes of ventricular fibrillation were induced by the ventricular vulnerability protocol in this study. The
letter n indicates lack of induction of ventricular fibrillation. In part 3, for seven sheep the intrinsic escape interval was
too rapid and therefore did not permit the delivery of a long S2-S3 interval, so this part was omitted, denoted by the
letter x.
In part 2, five episodes of ventricular fibrillation were
induced out of 568 shocks delivered (0.88%). There
were three for group B and two for group A (NS,
P>.15). All five episodes were induced with a right
ventricular depolarization-to-shock delay of 2 milliseconds and preceding cycle lengths .250 milliseconds. An
example of one ventricular fibrillation induction in this
part of the study is shown in Fig 7B.
During part 3, two episodes of ventricular fibrillation
were induced out of 238 shocks delivered (0.84%). Each
group had one sheep with an episode of ventricular
fibrillation induced during this part of the protocol. For
one sheep, the S1-S2 interval was 740 milliseconds
followed by an S2-S3 of 280 milliseconds, and for the
other sheep, the S1-S2 interval was 550 milliseconds
followed by an S2-S3 interval of 260 milliseconds. Both
occurred with a right ventricular depolarization-toshock delay of 2 milliseconds. An example of one
ventricular fibrillation induction in this part of the study
is shown in Fig 7C.
In part 4, two episodes of ventricular fibrillation were
induced out of 716 shocks (0.28%). There was one
episode for each group. Because the preceding cycle
length was not paced in this part but rather was selected
by the defibrillator-computer system to fall within a
preselected range, the exact preceding interval was
measured after completion of the study from the taperecorded electrograms. For one sheep, the preceding
cycle length range was 150 to 200 milliseconds, with a
measured cycle length of 198 milliseconds, and for the
other sheep, the preceding cycle length range was 200 to
250 milliseconds, with a measured preceding cycle
length of 240 milliseconds. Both episodes of ventricular
fibrillation were initiated with shock delays of 100
milliseconds. An example of one ventricular fibrillation
induction in this part of the study is shown in Fig 7D.
To analyze the dependence of vulnerability on cycle
lengths, an arbitrary cycle length of 300-millisecond
cutoff was used because it was the longest cycle length
preceding a shock that induced ventricular fibrillation.
There were 11 episodes of induced ventricular fibrillation out of 975 shocks with a preceding cycle length
s300 milliseconds (1.1%) and no episodes of ventricular fibrillation out of 895 shocks with a preceding cycle
length >300 milliseconds. The episodes of ventricular
fibrillation were compared on a sheep-by-sheep basis by
a McNemar's test and found to be significantly different
at P<.05. A paired t test comparing the number of
shocks delivered to individual sheep showed that the
number of shocks delivered on intervals <300 milliseconds were not significantly different from the number of
shocks delivered on intervals >300 milliseconds, P>.05.
Discussion
New Findings of This Study
In this study, 11 episodes of ventricular fibrillation
were induced during the delivery of 1870 atrial defibrillation shocks. We have shown that shocks delivered with
Ayers et al Rate Effect on Atrial Defibrillation Proarrhythmia
si si si
Si
.S
Sl SlSl
S1
:mV
Si
Si
S1
SI
S
419
12
Surface: ECG
Surface ECG
i
1 mY
Atrium
IlmV
Atrium
A V
lAI
sticl
1secon7d*
shock
1 second
B
A
Surfecb
S
1 SI
SISIS
81SI
1SlS
S2 S3
SurfacECG
ECG
mV
Right
jntricle
Rght
'Ventricle
V
V
l
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Atriumi
Atrium:
'1 m~V
1mV
I~ -seco-nt
Cec'
D
shock
second
shock
D
FIG 7. Electrogram recordings from shocks that resulted in fibrillation induction, one for each part of the study. In each part, the three
channels shown from top to bottom are surface lead 11 ECG, right ventricular apex, and the electrogram recorded between the two
shocking coils, one in the right atrium and the other in the great cardiac vein under the left atrial appendage. A, Example of an episode
of ventricular fibrillation induced in part 1 of the study. There are two sinus beats to the left in the electrograms, followed by the initiation
of the stimulus train at a cycle length of 250 milliseconds. After the eighth stimulus, there is a ventricular depolarization and the shock.
The shock is delayed 20 milliseconds from the last depolarization. After the shock is ventricular fibrillation. B, Example of an episode of
ventricular fibrillation induced in part 2 of the study. There are two sinus beats to the left in the electrograms, followed by the initiation
of the stimulus train with eight paced beats at a cycle length of 300 milliseconds, followed 230 milliseconds later by the premature
stimulus. After the premature stimulus, there is a ventricular depolarization and the shock. The shock is delayed 2 milliseconds from the
last depolarization. After the shock is ventricular fibrillation. C, Example of an episode of ventricular fibrillation induced in part 3 of the
study. There are two sinus beats to the left in the electrograms, followed by the initiation of the stimulus train with eight paced beats at
a cycle length of 300 milliseconds, followed 550 milliseconds later by the first premature stimulus and 260 milliseconds later by the
second premature stimulus. After the second premature stimulus, there is a ventricular depolarization and the shock. The shock is
delayed 2 milliseconds from the last depolarization. After the shock is ventricular fibrillation. D, Example of an episode of ventricular
fibrillation induced in part 4 of the study. The animal is in atrial fibrillation, and after a cycle length of 240 milliseconds, a shock is
delivered, delayed 100 milliseconds from the V wave. After the shock is ventricular fibrillation.
a preceding ventricular cycle length .300 milliseconds
are associated with a significantly increased likelihood
of ventricular proarrhythmia. All episodes of ventricular
fibrillation seen during this study occurred when the
preceding cycle length was c300 milliseconds, whereas
no episodes of ventricular fibrillation were seen when
the preceding cycle length was >300 milliseconds. We
showed that with appropriate synchronization, which
has previously been thought adequate to avoid the
vulnerable period, ventricular fibrillation could occur if
the preceding cycle length was c300 milliseconds.
There was no correlation between shock strength and
the occurrence of ventricular fibrillation. Although
there were 4 episodes of ventricular fibrillation out of 24
testing protocols for group A and 7 out of 22 for group
B, this difference was not significantly different overall
(P>.15), nor was it significantly different when groups
within any part, 1, 2, 3, or 4, of the protocol were
compared (P>.15).
Atrial Defibrillation Thresholds
The atrial defibrillation thresholds seen in this study
are comparable to the thresholds seen by others for
similar left to right atrial vectors in sheep.10-'2 Like the
previous studies, this study shows that in the sheep
model of atrial fibrillation, atrial defibrillation can be
achieved with transvenous electrodes with mean shock
energies of < 1 J.
The method of determining atrial defibrillation
threshold was different in this study than for others.10'11
In this study, we determined the lowest-intensity shock
required to defibrillate the atria and then delivered 20
shocks to estimate the percent successful defibrillation
at this intensity. Techniques used by other investigators
tend to estimate the 50% successful defibrillation point
on the dose-response curve. At the lowest intensity that
would defibrillate the atria, we found that the mean
percent success was 52% for group A and 48% for group
B (NS, P<.05). Although there was a rather large
standard deviation in the measurement of these values,
this method, on average, estimates the 50% point on the
dose-response curve. Additionally, the mean percent
success of 50% with a large standard deviation using a
technique designed to find the lowest-intensity shock
necessary to defibrillate the atria might suggest that the
dose-response curve for atrial defibrillation is quite
420
Circulation Vol 89, No 1 January 1994
steep in comparison to the 10-V steps used in this
protocol.
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Mechanisms of Ventricular Fibrillation Induction
Three possible mechanisms exist to account for induction of ventricular fibrillation by internally delivered
cardioversion shocks. One possible mechanism is lack of
synchronization of the cardioversion shock with respect
to intrinsic ventricular depolarization, resulting in the
shock being delivered during a T wave. All shocks in this
study were appropriately synchronized to ventricular
depolarizations; therefore, this mechanism does not
apply to the episodes of ventricular fibrillation induced
in this study.
A second potential mechanism is if a shock is synchronized to a ventricular depolarization that occurs
before complete ventricular repolarization from the
previous beat. From a surface electrogram, this may
appear as an R on T. In our study, 9 of the 11 episodes
of ventricular fibrillation appear to have been caused by
this mechanism. All of the episodes were in parts 1, 2,
and 3. Sample electrograms from these episodes are
shown in Fig 7A, 7B, and 7C with sensed V wave-toshock delay of either 2 or 20 milliseconds. For all these
episodes, it was evident from the right ventricular
electrogram that the ventricular depolarization that
triggered the shock was encroaching on the terminal
portion of the preceding T wave.
Others have observed a higher incidence of shockinduced ventricular fibrillation with shorter cycle
lengths in both dogs and humans during transvenous
cardioversion of ventricular tachycardia.1415 The investigators concluded that for tachycardia cycle lengths
<200 milliseconds, transvenous cardioversion shocks
would frequently accelerate the tachycardia or induce
ventricular fibrillation. Although the results in those
studies were similar to the present study, the mechanisms may be somewhat different because the ventricular tachycardia in the previous studies probably involved
intraventricular reentry. Nathan et a18 described the
induction of ventricular fibrillation from a 5-J shock
delivered to right heart catheters near the onset of the
QRS in a patient undergoing atrial defibrillation. The
patient had an apparent rapid ventricular response
based on the figure showing the intracardiac electrograms. They suggested that the mechanism of induction
may be similar to that described by Jackman and Zipes14
for ventricular tachycardia. These reports are under
conditions similar to part 1 of our study, in which the
ventricular response in tachycardia is rapid and regular.
Both the cycle length and dispersion of ventricular
refractoriness have been associated with the vulnerability of the ventricles to reentry. These effects could
increase the likelihood of ventricular proarrhythmia for
synchronized shocks delivered to the terminal recovery
phase of the preceding beat. In humans with long QT
syndrome, long/short cycle lengths frequently precede
episodes of ventricular arrhythmias.16 In patients with
suspected sudden death, long/short and short/long cycle
lengths frequently induce tachyarrhythmias and may
predict the method of induction of spontaneous episodes.17-20 El-Sherif et a12' showed, in canine infarcted
hearts, that reentry was more predominant when a
short/long/short stimulation sequence was used. In
atrial fibrillation, the irregular ventricular response and
the predominance of short cycle lengths are an ideal
setting for intraventricular reentry. In parts 1, 2, and 3
of this study, we systematically generated a spectrum of
cycle lengths that facilitated shock induction of ventricular fibrillation. This study showed that short cycle
lengths increased the likelihood of ventricular fibrillation induction in normal sheep hearts.
The third mechanism pertains to the delay of shock
delivery from the sensed ventricular depolarization and
the effect of irregular and rapid cycle lengths on the
refractory period of the ventricle and therefore the time
from depolarization to the vulnerable period. Dunbar et
a19 observed that transvenous defibrillation shocks delivered 116 to 180 milliseconds after the onset of the
QRS were associated with nine episodes of induced
ventricular fibrillation seen in their study. In part 4 of
our study, two episodes of ventricular fibrillation were
induced after short intervals with 100-millisecond
sensed ventricular depolarizations-to-shock delivery
delay. In these two episodes (Fig 7d), the shocks were
delivered during the early upstroke of the T wave, in the
right ventricular electrogram, with some of the myocardium apparently in a state of vulnerability. This relatively short recovery time of some of the myocardium
may be attributed to the dispersion of refractoriness
associated with the short/long/short preceding cycle
lengths.22
Relative Risk of Ventricular Fibrillation Induction
Previous studies done on low-energy transvenous
cardioversion of atrial fibrillation have reported only
minimal occurrence of ventricular fibrillation,8-10
whereas recent studies of higher-energy cardioversion
of atrial fibrillation with shocks delivered between an
intracardiac electrode and the chest wall have reported
no episodes of ventricular fibrillation.23 In our study, in
which producing ventricular proarrhythmia was the
objective, only 11 episodes of ventricular fibrillation
were induced. Eight of the 16 sheep in this study had no
episodes of ventricular fibrillation despite having half of
the test shocks delivered on intervals <300 milliseconds. It is apparent that even with conditions ripe for
the induction of ventricular fibrillation (short and irregular cycle lengths with a dispersion of ventricular refractoriness), the induction of ventricular fibrillation is
relatively rare. This study clearly showed that the risk of
ventricular fibrillation induction was significantly higher
when the preceding ventricular cycle length was 5300
milliseconds. In fact, no instances of ventricular fibrillation occurred when the preceding cycle length was
>300 milliseconds. It is logical to assume that if enough
time has elapsed since the last beat for all ventricular
cells to fully repolarize, none of the ventricular myocardium should be vulnerable. This hypothesis is supported
by the work of Elharrar and Surawicz,24 who showed, in
canine myocardium, that ventricular action potential
durations approached a maximum of 330 milliseconds at
long preceding cycle lengths.
Previous studies on the occurrence of ventricular
fibrillation after atrial cardioversion have been performed with transthoracic cardioversion and found that
shocks that were poorly synchronized with the R wave
would consistently result in ventricular fibrillation when
the shock fell during the T wave, the ventricular vulner-
able period.1-5 With the advent of synchronized defibril~
Ayers et al Rate Effect on Atrial Defibrillation Proarrhythmia
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
lators, the fear of ventricular fibrillation was decreased
but not abolished for cardioversion of atrial fibrillation.
Rarely is the association made between transthoracic
cardioversion of atrial fibrillation and the occurrence of
ventricular fibrillation. There may be several reasons for
this low occurrence rate. First, as indicated previously,
the induction of ventricular fibrillation by atrial defibrillation shocks is indeed rare, making a systematic study
of its occurrence difficult. Second, most patients undergoing cardioversion are receiving medications to control
their ventricular response to the atrial fibrillation, making shorter cycle lengths less frequent. Last, the energy
typically used for transthoracic cardioversion is at or
near the energy required to defibrillate the entire heart
and is therefore unlikely to induce ventricular
fibrillation.
Implications
This research has implications for the feasibility of
transvenous atrial defibrillation with either an external
or implanted atrial defibrillator. This study confirms the
work of others1",12 that low-energy transvenous atrial
defibrillation is feasible when both atria are interposed
between the electrodes. The concept of an implantable
atrial defibrillator has been raised by Levy and Camm,25
who discussed the critical paths in development of this
device. One point discussed pertained to defibrillating
the atria without ventricular arrhythmia induction. This
study shows conditions under which transvenous cardioversion of atrial fibrillation can result in ventricular
fibrillation. Shocks synchronized to sensed ventricular
depolarizations preceded by a cycle length <300 milliseconds can cause ventricular fibrillation and therefore
should be avoided by a transvenous atrial defibrillator.
Fortunately, the distributions of cycle lengths during
atrial fibrillation provide frequent cycle lengths longer
than those shown in this study to be arrhythmogenic
even when the mean heart rate is rapid.26 Although
untested, a similar technique of precluding shocks when
preceding cycle lengths are short may be applicable for
transthoracic cardioversion as well.
Limitations
As with all animal studies, there are limitations to the
animal model. First, in this study, the sheep was used to
model human atrial fibrillation and ventricular vulnerability. It must be noted that although sheep and human
hearts are similar in size and atrial fibrillation characteristics,26 undetected differences may exist. In addition,
these were normal hearts not under the influence of any
disease states that cause atrial fibrillation in humans,
such as ischemia, atrial dilation, hypertension, etc. Also
absent was the effect of long-term atrial fibrillation on
the ventricular response, such as tachycardia-induced
cardiomyopathy or a decreased ability of the ventricles
to compensate for the rapid and irregular rates present
in some of the human atrial fibrillation population.
Additionally, the interactions of many of the drugs used
to treat atrial fibrillation may affect the vulnerability of
the ventricles or alter the atrial and ventricular defibrillation and fibrillation thresholds. Testing the influences
of these other factors was beyond the scope of this study
and requires additional evaluation.
Second, ventricular pacing was used to mimic supraventricular influences of atrial fibrillation during
421
three parts of this study. Although many of the aspects
of atrial fibrillation were simulated with ventricular
pacing, the activation sequence was different with pacing. The effects of ventricular pacing, which could be
considered analogous to premature ventricular complexes, should, in theory, increase the likelihood of
ventricular fibrillation induction, since ventricular pacing delays conduction, resulting in prolonged depolarization and an increased dispersion of refractoriness.27
Additional testing with supraventricular pacing or a
combination of supraventricular sensing and ventricular
pacing, S, sensed with S2 and S3 ventricular paced,
merits further study.
Last, there was only limited randomization of the
parts of this protocol. Although parts 1 through 3 were
performed in random order, they always preceded part
4. Intervals within any part were not randomized. This
lack of randomization was caused by our concern that
many episodes of ventricular fibrillation would increase
the risk of death of the animal, preventing completion of
the protocol, or that multiple episodes of ventricular
fibrillation would compromise the normal physiology of
the model. We therefore conducted the study according
to standard electrophysiological methodology in which
prematurity was gradually increased. For this reason,
the protocol for any of parts 1, 2, 3, or 4 was stopped
after induction of ventricular fibrillation. For episodes
of ventricular fibrillation that were probably caused by
an R-on-T situation and a shock delay of 20 milliseconds, it can only be inferred that a shock delay of 2
milliseconds would have also caused stimulation in the
vulnerable period and induced ventricular fibrillation,
although this was not actually observed. In part 4, the
short preceding cycle length and a shock delay of 100
milliseconds probably induced ventricular fibrillation by
vulnerable-period stimulation of the T wave after the
synchronized V wave. It is unknown in these two
episodes whether a shorter delay would have induced
ventricular fibrillation via an R-on-T shock mechanism,
since no shorter delays were tested.
This study shows that synchronized shocks with short
preceding cycle lengths (<300 milliseconds) have a
significantly greater risk of ventricular proarrhythmia.
This study also confirms that low-energy transvenous
atrial defibrillation of human-size sheep hearts can be
accomplished with energies <1 J.
Acknowledgment
We thank Dr Susan Bernard for her technical assistance.
References
1. Lown B, Amarasingham R, Neuman J. New method for terminating cardiac arrhythmias. JAMA. 1962;182:548-555.
2. Lown B. Electrical reversion of cardiac arrhythmias. Br Heart J.
1967;26:469-489.
3. Ferris LP, King BG, Spence PW, Williams HB. Effect of electric
shock on the heart. Elec Eng. 1936;55:498-515.
4. Wiggers CJ, Wegria R. Ventricular fibrillation due to single
localized induction and condenser shocks applied during the vulnerable phase of ventricular systole. Am J Physiol. 1940;128:
500-505.
5. Lown B, Kaidbey S, Perlroth M, Abe T. Comparative studies of
ventricular vulnerability to fibrillation. J Clin Invest. 1963;42:953.
Abstract.
6. Flaker GC, Blackshear JL, McBride R, Kronmal RA, Halperin JL,
Hart RG, on behalf of the SPAF investigators. Antiarrhythmic
drug therapy and cardiac mortality in atrial fibrillation. JAm Coil
Cardiol. 1992;20:527-532.
422
Circulation Vol 89, No 1 January 1994
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
7. Kannell WB, Abbott RD, Savage DD, McNamara PM. Epidemiologic features of chronic atrial fibrillation: the Framingham study.
N Engl JMed. 1982;306:1018-1022.
8. Nathan AW, Bexton RS, Spurrell RA, Camm AJ. Internal transvenous low energy cardioversion for the treatment of cardiac
arrhythmias. Br Heart J. 1984;52:377-384.
9. Dunbar DN, Tobler G, Fetter J, Gornick CC, Benson DW, Benditt
D. Intracavitary electrode catheter cardioversion of atrial tachyarrhythmias in the dog. JAm Coll Cardiol. 1986;2:1015-1027.
10. Powell AC, Garan H, McGovern BA, Fallon JT, Krishnan SC,
Ruskin JN. Low energy conversion of atrial fibrillation in the
sheep. JAm Coll Cardiol. 1992;20:707-711.
11. Cooper RAS, Alferness CA, Smith WM, Ideker RE. Internal
cardioversion of atrial fibrillation in sheep. Circulation. 1993;87:
1673-1686.
12. Ayers GM, Ilina M, Wagner D, Kreyenhagen P, Bernard S,
Alferness C. Cardiac vein electrode locations for transvenous atrial
defibrillation. J Am Coll Cardiol. 1993;21:306A. Abstract.
13. Hoff HE, Geddes LA, McCrady JD. The maintenance of experimental atrial fibrillation by cholinergic factors. Cardiovasc Res Cent
Bull. 1965;3:117-129.
14. Jackman WM, Zipes DP. Low-energy synchronous cardioversion
of ventricular tachycardia using a catheter electrode in a canine
model of subacute myocardial infarction. Circulation. 1982;66:
187-195.
15. Zipes DP, Jackman WM, Heger JJ, Chilson DA, Browne KF,
Naccarelli GV, Rahilly GT, Prystowski EN. Clinical transvenous
cardioversion of recurrent life-threatening ventricular tachyarrhythmias: low energy synchronized cardioversion of ventricular
tachycardia and termination of ventricular fibrillation in patients
using a catheter electrode. Am Heart J. 1982;103:789-794.
16. Kay GN, Plumb VJ, Arciniegas JG, Henthorn RW, Waldo AL.
Torsade de pointes: the long-short initiating sequence and other
clinical features: observations in 32 patients. J Am Coll Cardiol.
1983;2:806-817.
17. Denker S, Lehmann MH, Mahmud R, Gilbert C, Akhtar M.
Divergence between refractoriness of His-Purkinje system and ventricular muscle with abrupt changes in cycle length. Circulation.
1983;68:1212-1221.
18. Denker S, Lehmann MH, Mahmud R, Gilbert C, Akhtar M. Facilitation of macroreentry within the His-Purkinje system with abrupt
changes in cycle length. Circulation. 1984;69:26-32.
19. Denker S, Lehmann M, Mahmud R, Gilbert C, Akhtar M. Facilitation of ventricular tachycardia induction with abrupt changes in
ventricular cycle lengths. Am J Cardiol. 1984;53:508-515.
20. Gomes JA, Alexopoulos D, Winters SL, Deshmukh P, Fuster V,
Suh K. The role of silent ischemia, the arrhythmic substrate and
the short-long sequence in the genesis of sudden cardiac death.
JAm Coll Cardiol. 1989;14:1618-1625.
21. El-Sherif N, Gough WB, Restivo M. Reentrant ventricular
arrhythmias in the late myocardial infarction period: mechanism by
which a short-long-short cardiac sequence facilitates the induction
of reentry. Circulation. 1991;83:268-278.
22. Janse MJ, van der Steen ABM, van Dam R, Durrer D. Refractory
period of the dog's ventricular myocardium following sudden
changes in frequency. Circ Res. 1969;24:251-262.
23. Levy S, Lauribe P, Dolla E, Kou W, Kadish A, Calkins H,
Pagannelli F, Moyal C, Bremondy M, Schork A, Shyr Y, Das S,
Shea M, Gupta N, Morady F. A randomized comparison of
external and internal cardioversion of chronic atrial fibrillation.
Circulation. 1992;86:1415-1420.
24. Elharrar V, Surawicz B. Cycle length effect on restitution of action
potential duration in dog cardiac fibers. Am J PhysioL 1983;
244(Heart Circ Physiol 13):H782-H792.
25. Levy S, Camm J. An implantable atrial defibrillator. Circulation.
1993;87:1769-1771.
26. Ayers GM, Alferness CA, Illina MI, Wagner DO, Kreyenhagen P.
Evaluation of a sheep model of atrial fibrillation. Am Heart J.
1992;124:839. Abstract.
27. Moore EN, Spear JF. Ventricular fibrillation thresholds: its physiological and pharmacological importance. Arch Intern Med. 1975;
135:446-453.
Ventricular proarrhythmic effects of ventricular cycle length and shock strength in a
sheep model of transvenous atrial defibrillation.
G M Ayers, C A Alferness, M Ilina, D O Wagner, W A Sirokman, J M Adams and J C Griffin
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Circulation. 1994;89:413-422
doi: 10.1161/01.CIR.89.1.413
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1994 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circ.ahajournals.org/content/89/1/413
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial
Office. Once the online version of the published article for which permission is being requested is located,
click Request Permissions in the middle column of the Web page under Services. Further information about
this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation is online at:
http://circ.ahajournals.org//subscriptions/