Epilepsia, 50(4):928–932, 2009
doi: 10.1111/j.1528-1167.2008.01918.x
BRIEF COMMUNICATION
Late-onset periodic asystolia during vagus nerve
stimulation
*Jorge Iriarte, *Elena Urrestarazu, *Manuel Alegre, yAlfonso Macı́as, *Asier Gómez,
*Paola Amaro, *Julio Artieda, and *Cesar Viteri
*Epilepsy Section, Clinica Universitaria, University of Navarra. Pamplona, Spain;
and yCardiology, Clinica Universitaria, University of Navarra. Pamplona, Spain
SUMMARY
Cardiac changes may occasionally occur during
vagus nerve stimulation (VNS) used in epileptic
patients. As they can be potentially life-threaten-
ing, it is important to detect them, and this is
why an intraoperative test is performed during
the implantation. Few cases of asystole during
this test have been described. Only one patient
with late-onset bradyarrythmia caused by VNS
Vagus nerve stimulation (VNS) is a safe treatment for
epilepsy. The most frequent side effects are hoarseness,
cough, dysphonia, paresthesias, and superficial infections,
but others may also occur (Ben Menachem, 2001; Iriarte
et al., 2001; Smyth et al., 2003).
Cardiac changes related to VNS in epilepsy have also
been reported to be very unusual (0.1%) (Ali et al., 2004;
Stemper et al., 2008). Several cases of bradyarrhythmia
during implantation have been reported (Asconape et al.,
1999; Tatum et al., 1999; Ali et al., 2004). In these cases
the procedure was immediately terminated and the device
removed. However, studies have not found an excess of
cases of sudden death in patients with VNS (Annegers
et al., 2000). The first case of late-onset bradyarrythmia
related to VNS was reported in 2007 (Amark et al., 2007).
The device was stopped and the problems disappeared.
In this report we present a patient who developed brad-
yarrhythmias related to each run of stimulations 9 years
after starting VNS. The arrhythmia disappeared when the
device was stopped. A relationship with status epilepticus
and antiepileptic drugs has also been suggested.
Accepted September 19, 2008; Early View publication December 4,
2008.
Address correspondence to Jorge Iriarte M.D., Epilepsy Unit, Clnica
Universitaria, University of Navarra, Avda Po XII 36, Pamplona 31008,
Navarra, Spain. E-mail: jiriarte@unav.es
Wiley Periodicals, Inc.
ª 2008 International League Against Epilepsy
928
has been reported. This patient had been
implanted 2 years and 4 months before the
episode. We present another case of late
asystole in a patient whose VNS had been
implanted 9 years before the arrhythmia onset.
In our patient, each run of stimulation produced
bradyarrhythmias and very often severe asystolia
due to atrium-ventricular block.
KEY WORDS: Vagal stimulation, Asystole,
Epilepsy.
Case Report
The patient is a 47-year-old woman who has had epilepsy
since the age of 12. She had refractory epilepsy with simple
partial seizures (lightening feeling in the back) and complex
partial seizures with and without secondary generalization.
She has been occasionally admitted to hospital with convul-
sive or nonconvulsive status epilepticus. Pseudoseizures
have also been suspected in recent years due to unusual
symptoms. Electroencephalography (EEG) findings were
frequently abnormal, showing very frequent right and left
temporal slow waves and spikes. Brain magnetic resonance
imaging (MRI) was normal. In 1999, given the intractability
of her seizures, a VNS was implanted. The device was an
NCP type 101 (Cyberonics, Houston, TX, U.S.A.). The
implantation and its functioning for 9 years were unevent-
ful. The stimulation parameters were unchanged during the
last 3 years (30 s on, 5 min off, 1.75 mA, frequency 20 Hz,
and pulse width 500 ms).
The patient complained of new events described as brief
episodes of dizziness and unsteadiness without loss of
consciousness. They started approximately 2 months
before admission, and 9 years after VNS implantation,
shortly after she was discharged from the hospital follow-
ing the control of an episode of status epilepticus. At that
time she was on pregabalin (450 mg/day), clonazepam
(3 mg/day), and levetiracetam (2,500 mg/day, being
added to the previous treatment at the time of the status

epilepticus). The frequency of these spells was 10–50
times a day. The frequency of her typical simple and com-
plex partial seizures also increased. EEG at this admission
showed very frequent sharp waves over the right temporal
area, very similar to electroencephalographic focal status.
However, no clear relationship with the new spells could
be established. Because she was having daily events, a
video-EEG monitoring study was undertaken.
The video-EEG was performed with a 72-channel
video-EEG system using Harmonie 6.0 software (Stellate,
Montreal, PQ, Canada). Interictal EEG showed very fre-
quent sharp waves over the right temporal area, and some
slow waves and occasional sharp waves over the left tem-
poral area. In 18 h, two types of episodes were recorded.
The most frequent type of episodes (n = 5) consisted of a
feeling of a seizure coming on for several seconds, with
neither loss of consciousness nor changes to the EEG. In
addition, two spells without loss of consciousness and
irregular jerking of the limbs, the eyes remaining open,
Figure 1.
Electrocardiography (ECG) samples. The figure shows five 80-s long ECG samples (upper part). The tracings are
5 min 34 s apart. In the last two traces we see two examples of the aystolia due to atrioventricular blocking (the p
wave is present) in more detail.
Epilepsia ILAE
929
Asystole and Vagal Stimulation
and with no EEG change, were captured. The first episode
type was considered likely to be simple partial seizures,
whereas the second type was diagnosed as pseudoseizures.
The most striking finding was the presence of periodic car-
diac changes with bradyarrhythmias every 5 min and
30 s, many of them with prolonged asystolia (up to 8 s).
These arrhythmias appeared during sleep and wakefulness
and were not clearly associated with clinical symptoms.
The electrocardiography (ECG, always included in the
EEG and video-EEG recordings) showed a complete atrio-
ventricular block with persistence of the p waves (Fig. 1).
This block started some seconds after the vagal stimula-
tion—a second-degree block—but sometimes the block-
ing became a full block with prolonged asystole (Fig. 1).
She was admitted to the cardiac intensive care unit and
underwent Holter monitorization. Because the periodicity
of the bradyarrhythmias was the same as that of the VNS
(around 30 s every 5.5 min), the VNS was stopped. In
Fig. 2 we see a 2-h period of the instant pulse rate (bpm).
Epilepsia, 50(4):928–932, 2009
doi: 10.1111/j.1528-1167.2008.01918.x
930
J. Iriarte et al.
Figure 2.
Instant pulse rate (bpm) in a 2-h period. Every 5.5 min there is a bradyarrythmia and/or asystole with very low pulse
rates (10–30 bpm). After the asystolia there is a compensatory tachycardia. Pulse in beats per minute (bpm) in the
corresponding time. The patient was sleeping, including REM and non-REM sleep stages. hh:mm:ss = hour: minutes:
seconds.
Epilepsia ILAE
The cardiac arrhythmia resolved immediately. No
abnormality was detected in the ECG during the next
24 h. During the week after this study, the patient did not
complain of presyncopal spells, and she was sent home.
We reviewed the routine EEG performed in recent years
and confirmed that the ECG was regular until 2 months
ago. The first time that changes in the cardiac rate were
seen was in an EEG performed the day she was admitted
because of status epilepticus, before the introduction of
levetiracetam. It was a decrease in the cardiac rate from
110 bpm to 85 bpm every 5.5 min, but no asystole was
noted (Fig. 3); it was never noted in previous EEGs. How-
ever, in the EEG on the same day, 2 h after the administra-
tion of 1 g of intravenous levetiracetam, the instant
cardiac rate fell to 50 bpm every 5.5 min (Fig. 3), but
again no asystole was observed.
Discussion
The vagus nerve regulates cardiac function, and this
fact has always raised concern that VNS might affect the
cardiac rhythm. This was the reason for using stimulation
of the left vagus (more related to the ventricles than to the
atria). However, the cardiac complications of VNS are
very infrequent (Ben Menachem, 2001; Galli et al., 2003).
Epilepsia, 50(4):928–932, 2009
doi: 10.1111/j.1528-1167.2008.01918.x
Usually patients with a predisposition for arrhythmia
are detected in the test during the implantation procedure
(Asconape et al., 1999; Ali et al., 2004)). Amark et al.
(2007) reported a patient with syncope and cardiac
arrhythmia related to the long-term use of VNS. The
patient was an adolescent who had been using the VNS for
2 years and 4 months. No inductors were found in this
patient.
In our patient, the correlation between the stimulation
and the arrhythmia was evident. The ECG showed com-
plete atrioventricular (AV) nodal block, a typical pattern
of the left VNS (Amark et al., 2007). The interval between
arrhythmias coincided with the activation of the VNS. The
duration of the arrhythmia was a few seconds longer than
the stimulation (35 vs. 30 s, respectively), and was very
similar in all the stimulations, as can be seen in Figs. 1 and
2. Finally, the arrhythmia disappeared immediately after
the deactivation of the VNS.
Why this problem appeared only years after implanta-
tion is not known, and a possible influence of status epi-
lepticus and the antiepileptic drugs cannot be excluded.
The first changes in the cardiac rate were detected during
status epilepticus, and it is a well-known phenomenon that
vegetative changes are frequent during epileptic seizures
and status epilepticus. It is possible that these changes
 931
Asystole and Vagal Stimulation

Figure 3.
Instant pulse rate (bpm) in the initial electroencephalography (ECG) before and after levetiracetam. In the upper
part, we show the instant pulse rate in the ECG during status epilepticus, before levetiracetam. In the lower part of
the figure, electrocardiography analysis after levetiracetam is shown, still during the focal status epilepticus. The
analysis was performed in 16 min in both samples. In all the EEG and video-EEG studies, ECG is also recorded.
Epilepsia ILAE

predisposed the patient to VNS-related cardiac arrhyth-
mia. It is possible also that levetiracetam was involved in
the production of the arrhythmia. The introduction of lev-
etiracetam (1 g, i.v.) during the episode of status epilepti-
cus produced a fall in the cardiac rate and, afterward, AV
block developed, prompting new clinical spells. To our
knowledge this drug has not been associated previously
with significant cardiac changes, and, interestingly,
neither reduction nor withdrawal of levetiracetam
reversed the cardiac alterations. In the cases reported in
the literature, none of the patients was receiving this drug
(Asconape et al., 1999; Ali et al., 2004; Amark et al.,
2007).
In conclusion, this case illustrates that severe cardiac
arrhythmias associated with VNS can occur years after
implantation. Although we cannot confirm that the VNS
was wholly responsible for this new-onset asystolia, the
relationship with the VNS seems clear. It is also possible
that antiepileptic drugs can modify the susceptibility of
the vagus nerve to chronic stimulation. Therefore, new-
onset episodes of nonepileptic origin in patients with VNS
merit urgent cardiac study. Epileptologists must be aware
of this complication, even in patients who have had
excellent tolerability of the VNS for years.
Acknowledgments
Dr. Amaro is a Research Fellow supported by Alban Institution,
E07E402268CL.
We confirm that we have read the Journal’s position on issues involved in
ethical publication and affirm that this report is consistent with those
guidelines.
Disclosure of conflicts of interest: None of the authors has any conflict of
interest to disclose.

Epilepsia, 50(4):928–932, 2009

doi: 10.1111/j.1528-1167.2008.01918.x
932
J. Iriarte et al.

Galli R, Limbruno U, Pizzanelli C, Giorgi FS, Lutzemberger L,

References Strata G, Pataleo L, Mariani M, Iudice A, Murri L. (2003)
Analysis of RR variability in drug-resistant epilepsy patients chroni-
Ali II, Pirzada NA, Kanjwal Y, Wannamaker B, Medhkour A, Koltz MT,
cally treated with vagus nerve stimulation. Auton Neurosci
Vaughn BV. (2004) Complete heart block with ventricular asystole
107:52–59.
during left vagus nerve stimulation for epilepsy. Epilepsy Behav
Iriarte J, Artieda J, Alegre M, Schlumberger E, Urrestarazu E, Pastor MA,
5:768–771.
Viteri C. (2001) Spasm of the sternocleidomastoid muscle induced by
Amark P, Stodberg T, Wallstedt L. (2007) Late onset bradyarrhythmia
vagal nerve stimulation. Neurology 57:2319–2320.
during vagus nerve stimulation. Epilepsia 48:1023–1024.
Smyth MD, Tubbs RS, Bebin EM, Grabb PA, Blount JP. (2003) Compli-
Annegers JF, Coan SP, Hauser WA, Leestma J. (2000) Epilepsy,
cations of chronic vagus nerve stimulation for epilepsy in children.
vagal nerve stimulation by the NCP system, all-cause mortal-
J Neurosurg 99:500–503.
ity, and sudden, unexpected, unexplained death. Epilepsia 41:
Stemper B, Devinsky O, Haendl T, Welsch G, Hilz MJ. (2008) Effects of
549–553.
vagus nerve stimulation on cardiovascular regulation in patients with
Asconape JJ, Moore DD, Zipes DP, Hartman LM, Duffell WH Jr. (1999)
epilepsy. Acta Neurol Scand 117:231–236.
Bradycardia and asystole with the use of vagus nerve stimulation for
Tatum WO, Moore DB, Stecker MM, Baltuch GH, French JA, Ferreira
the treatment of epilepsy: a rare complication of intraoperative device
JA, Carney PM, Labar DR, Vale FL. (1999) Ventricular asystole dur-
testing. Epilepsia 40:1452–1454.
ing vagus nerve stimulation for epilepsy in humans. Neurology
Ben Menachem E. (2001) Vagus nerve stimulation, side effects, and
52:1267–1269.
long-term safety. J Clin Neurophysiol 18:415–418.

Epilepsia, 50(4):928–932, 2009

doi: 10.1111/j.1528-1167.2008.01918.x