COMPLICATION

Acute Ischemic Stroke During Deep Brain
Stimulation Surgery of Globus Pallidus Internus:
Report of 5 Cases
BACKGROUND: Cerebrovascular accident (CVA) is a potentially devastating complication of deep brain stimulation (DBS) surgery. Although there are substantial data reporting the incidence and cause of hemorrhagic CVA, reports of acute ischemic
infarctions during DBS implantation surgery are rare.
OBJECTIVE: To present a series of 5 patients who experienced clinically significant
ischemic CVA during microelectrode-guided globus pallidus internus (GPi) DBS, and
evaluate the potential risk factors and mechanisms.
METHODS: A retrospective analysis of GPi DBS surgeries performed between June 2010
and February 2015 at UCLA Medical Center and June 2010 and February 2014 at CedarsSinai Medical Centers was performed to identify stroke risk factors. Statistical analysis
was performed, comparing the stroke group with all patients undergoing GPi DBS.
RESULTS: All 5 patients developed acute onset of lethargy, dysarthria, and contralateral facial and/or hemibody weakness intraoperatively. Computed tomographic
scans in all cases were negative for hemorrhage. Magnetic resonance images obtained in 3 patients revealed infarction in the posterior limb of the internal capsule.
During the time period analyzed, a total of 234 GPi leads were placed in 129 patients,
yielding a 2.14% rate of ischemic stroke per lead. No statistically significant risk factors
were identified in the stroke group. Given the variability of symptom onset during
surgery, the mechanism is not clear, but it could be related to compression, compromise, or vasospasm of lenticulostriate arteries and/or anterior choroidal branches
near the GPi target.
CONCLUSION: Ischemic stroke in GPi DBS is a significant complication for clinicians to
be aware of and discuss with their patients preoperatively.

Angela E. Downes, MD*
Patrick Pezeshkian, MD‡
Eric Behnke, BS‡
Yvette Bordelon, MD, PhD§
Michele Tagliati, MD¶
Adam Mamelak, MDk
Nader Pouratian, MD, PhD‡
*Department of Neurosurgery, University
of South Florida Morsani College of
Medicine, Tampa, Florida; Departments
of ‡Neurosurgery, and §Neurology,
University of California, Los Angeles,
Los Angeles, California; Departments of
¶Neurology, and kNeurosurgery, Cedars
Sinai Medical Center, Los Angeles,
California
Correspondence:
Angela E. Downes, MD,
Department of Neurosurgery,
University of Colorado,
12631 E. 17th Avenue, C307,
Aurora, CO 80045.
E-mail: angela.downes@ucdenver.edu
Received, September 23, 2015.
Accepted, June 5, 2016.
Published Online, July 25, 2016.

KEY WORDS: Complications, Deep brain stimulation, Globus pallidus internus, Ischemia, Stroke
Operative Neurosurgery 12:383–390, 2016

Copyright © 2016 by the
Congress of Neurological Surgeons.

T

he efficacy of globus pallidus internus (GPi)
deep brain stimulation (DBS) for the treatment of Parkinson disease (PD) and dystonia is well documented.1-4 The Veterans Affairs
cooperative trial and several other trials have
concluded that GPi and subthalamic nucleus
(STN) DBS are equally effective for the management of the motor symptoms of PD.2 Because
STN DBS is possibly associated with greater rates
ABBREVIATIONS: DBS, deep brain stimulation;
GPi, globus pallidus internus; MER, microelectrode
recording; PD, Parkinson disease; PLIC, posterior
limb of the internal capsule; STN, subthalamic
nucleus

OPERATIVE NEUROSURGERY

DOI: 10.1227/NEU.0000000000001359

of neurocognitive decline and falls,1,2 GPi DBS
has regained relative interest.
Arguably, the most dangerous and clinically
significant complication of DBS is intracerebral
hemorrhage. In a large series of 567 electrode
placements, Sansur et al5 reported a 1.2% risk of
symptomatic hemorrhage and a 0.7% risk of
hemorrhage resulting in permanent neurological
deficit. The authors did not find a statistically
significant difference in hemorrhage rates between
STN, GPi, and thalamic targets. Although studies
have extensively analyzed hemorrhagic complications,5-8 few reports describe, detail, or quantify
subcortical ischemic adverse events related to DBS
implantation. Although it is mentioned in some
complication analyses,9 only 2 reports in literature

VOLUME 12 | NUMBER 4 | DECEMBER 2016 | 383

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

DOWNES ET AL

specifically describe nonhemorrhagic infarctions occurring during
DBS surgery.10,11
Herein, we report 5 cases of clinically significant ischemic events
occurring during stereotactic GPi DBS procedures. We quantify
the incidence of this complication across 2 centers and describe the
clinical presentation of each patient, their neurological outcomes,
and neuroradiological findings.

METHODS
A retrospective analysis of GPi DBS surgeries performed between June
2010 and February 2015 at UCLA Medical Center and June 2010 and
February 2014 at Cedars-Sinai Medical Centers was performed. Cases in
which ischemic events occurred (based on acute clinical deficit without
gross radiographic evidence of hemorrhage) were reviewed in detail.
Clinical history, presentation, surgical indications, neurological outcomes, and neuroradiological outcomes were reviewed. Potential risk
factors were evaluated, including age, sex, diagnosis, smoking history,
history of cardiovascular and cerebrovascular disease (including history of
hypertension, stroke, and myocardial infarction), use of microelectrode
recording (MER), and intraoperative blood pressure.

Surgical Procedure
All patients underwent frame-based stereotactic GPi DBS implantation
with similar pre- and intraoperative steps. Before surgery, 3-T stereotactic
magnetic resonance (MR) imaging of the brain with contrast was
performed. On the morning of surgery, a stereotactic frame (Leksell
[Elekta, Stockholm, Sweden] at UCLA, CRW [Integra, Plainsboro, New
Jersey] at Cedars-Sinai) was placed under brief propofol sedation and local
anesthesia, after which a computed tomography (CT) scan of the head
with a localizer was obtained. CT and MRI data were registered using
a commercial software (BrainLab [Munich, Germany] at UCLA,
Framelink [Medtronic, Dublin, Ireland] at Cedars-Sinai), and the GPi
was targeted using a burr hole approximately at the coronal suture and 0 to
10 degrees lateral in the coronal plane, as indicated by a trajectory that
avoided the ventricle. Target coordinates are detailed below for each case.
Skin incision and burr holes were made with propofol-induced light
sedation. After reexamining the patient and confirming normal blood
pressure as measured by cuff (systolic less than 160 mm Hg), MER was
performed using single-use microelectrodes with the Alpha Omega
recording system (Alpha Omega, Nazareth, Israel) and microdrive with
a single pass in each case. After MER targeting, the DBS electrode
(Medtronic model number 3387) was implanted and intraoperative
macrostimulation testing performed.

RESULTS
A total of 234 GPi leads were stereotactically implanted in 129
patients, including first implants and revisions (Table 1).
Diagnoses include PD (n = 112, 86.8%), dystonia (n = 16,
12.4%), and Tourette disease (n = 1, 0.78%). The mean age at
the time of surgery was 60.3 6 14.3 years, and 59.7% were male.
Microelectrode recordings were used in 126 cases (97.7%).
Incidence of Ischemic Stroke
Ischemic strokes occurred during implantation of 5 of 234
leads, resulting in a 2.14% rate of ischemic stroke per lead

384 | VOLUME 12 | NUMBER 4 | DECEMBER 2016

implanted. Details of each case are presented below. Of the
patients who had a stroke, 4 (80%) had a diagnosis of PD and 1
(20%) was diagnosed with dystonia, in comparison with 86.8%
and 12.4%, respectively, for the overall population. In the
cohort of patients who did not have a stroke, there was 1
hemorrhage noted on postoperative scans in a patient who also
experienced a neurological deficit (0.78% symptomatic hemorrhage rate), with symptoms that resolved within 3 months of
implantation.
Risk Factors
Potential risk factors in patients with ischemic stroke were
compared with the rest of the cohort of patients who received
implants in the GPi (Table 1). No significant age difference was
observed: 54.2 6 4.2 years for ischemic cerebrovascular accident
vs 60.3 6 14.3 years for all others (P = .08, t test with unequal
variance). Eighty percent of affected patients were female, and
only 1 (20%) had known stroke risk factors, including
hypertension and obesity (Table 2). None of the patients who
had an ischemic complication were past or current smokers.
Microelectrode recording was used in all cases in which stroke
occurred. Blood loss was minimal in each case.
Imaging
CT scan was performed immediately postoperatively in all 5
patients. MRIs were obtained in 3 of the 5 patients. There was no
evidence of blood on either CT or MRI in any patient. The
ischemic zones as seen on MRI are outlined in Table 2.
Case Report 1
A 64-year-old woman underwent bilateral GPi DBS for a 6year history of PD presenting with left-handed tremor,
bradykinesia, rigidity, and on-time dyskinesias. Vascular risk
factors included obesity and hypertension, for which she was
taking atenolol, Lisinopril, and terazosin. GPi was targeted at X =
19.0 mm, Y = 2.0 mm, and Z = 25.0 mm (AC-PC length =
25.9 mm). At surgery, left-hemisphere MER was uneventful, with
excellent recordings confirming the GPi target. Upon implantation
of the DBS lead, but before stimulation, the patient developed
acute-onset dysarthria and weakness of the right face, arm, and leg
(4/5). Stimulation testing demonstrated excellent clinical response.
The right-side DBS lead was rapidly implanted without MER.
Immediate postoperative head CT did not reveal intracerebral
hemorrhage (although a small hemorrhage obscured by the DBS
lead cannot be definitively excluded), but on retrospective review
revealed an area of infarction of the left posterior limb of the
internal capsule (PLIC), posterior and medial to the DBS lead tip.
At 3-month follow-up, a brain MRI revealed hyperintensity in the
left PLIC, consistent with chronic infarct. Wallerian degeneration
was also seen in the left corticospinal tract in the brainstem
(Figure 1). At 6-month follow-up, the patient had persistent slight
right facial droop and pronator drift with 41 strength in the

www.operativeneurosurgery-online.com

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

ACUTE ISCHEMIC STROKE IN GPI DBS

TABLE 1. Characteristics of All Patients Undergoing GPi DBS From June 2010 to February 2015 at Both Institutions in This Seriesa
Patients With Ischemic Stroke
During GPi DBS Surgery

All GPi DBS Patients
Without Ischemic Stroke

5
5
54.2 6 4.2
20

129
234
60.3 6 14.3
59.7

80
20
0

86.8
12.4
0.78

No. patients
No. electrodes
Mean age
Male, %
Diagnosis, %
Parkinson disease
Dystonia
Tourette
a

P
N/A
N/A
..08 (t test)
..05 (Fisher exact test)
..05 (2-way ANOVA)

ANOVA, analysis of variance; DBS, deep brain stimulation; GPi, globus pallidus internus.

right upper extremity. PD motor symptoms are well controlled
with bilateral GPi stimulation.
Case Report 2
A 54-year-old woman presented with a 30-year history of PD
with features of right-sided greater than left-sided rigidity,
bradykinesia, and on-time dyskinesia. At baseline, her examination was significant for dysarthric speech secondary to a history of
tonsillar carcinoma treated with radiation to the pharynx and
chemotherapy 5 years earlier. She did not have a history of tobacco
use, hypertension, stroke, or myocardial infarction. GPi was
targeted at X = 20.0 mm, Y = 2.6 mm, and Z = 23.4 mm
(AC-PC length = 28.2 mm). At surgery, MER and lead
placement in the left hemisphere proceeded uneventfully, as
well as the initial MER and subsequent implant of the right-sided
GPi lead. Upon initiation of macrostimulation testing of the
right-side lead, the patient became acutely somnolent and
developed a subtle left-sided facial droop, left arm weakness
(4/5), and worsened dysarthria. She remained arousable to voice
and followed commands. Head CT performed postoperatively
did not demonstrate hemorrhage but revealed a right PLIC

hypodensity, again posterior and medial to the DBS lead tip. At
6-month follow-up, left arm weakness had resolved, but her mild
left facial weakness and dysarthria were still present, associated
with dysphagia and sialorrhea from the left mouth. These
symptoms have been evaluated and managed on an outpatient
basis with speech and occupational therapy. PD motor symptoms
are well controlled with bilateral GPi stimulation.
Case Report 3
A 52-year-old man with a 12-year history of PD underwent
DBS for motor fluctuations with disabling on-time dyskinesias.
He did not have a history of tobacco use, hypertension, stroke, or
myocardial infarction. GPi was targeted at X = 20.0 mm, Y =
2.0 mm, and Z = 23.0 mm (AC-PC length = 27.2 mm). At
surgery, left-sided MER and lead placement proceeded uneventfully. During right-sided MER, approximately 6 mm from target,
the patient became somnolent, had decreased spontaneous
speech, dysarthria, and slight left facial droop, in the absence of
limb weakness, while rigidity and tremor in the left upper and
lower extremities completely resolved. Electrical silence on the
MER was observed. The lead was implanted, stimulation

TABLE 2. Characteristics of Patients Experiencing Ischemic Stroke During GPi DBS Implantationa

a

Patient Age

Vascular Risk
Factors

Diagnosis

Onset Time of Ischemic
Symptoms

Symptoms of
Ischemia

Location of
Ischemia

1

64

Obesity, HTN

PD

Macroelectrode insertion
(no stimulation)

Dysarthria, weakness of right face, arm, leg

Left PLIC and lateral
putamen

2

54

None

PD

52

None

PD

4

51

None

PD

MER

5

52

None

Dystonia

MER

Somnolence, dysarthria, weakness of left
face and arm
Somnolence, dysarthria, left face weakness,
resolution of tremor
Left gaze preference, dysarthria, right arm
and leg weakness
Somnolence, resolution of right-sided
dyskinesias

Right PLIC

3

Macroelectrode
stimulation
MER

Right PLIC
Left GPi and PLIC
Left GPi and PLIC

GPi, globus pallidus internus; HTN, hypertension; MER, microelectrode recording; PD, Parkinson disease; PLIC, posterior limb of internal capsule.

OPERATIVE NEUROSURGERY

VOLUME 12 | NUMBER 4 | DECEMBER 2016 | 385

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

DOWNES ET AL

FIGURE 1. Case 1. A, axial FLAIR MRI obtained 3 months poststroke showing low-intensity signal medial and posterior to left
electrode representing chronic infarct in PLIC. B, axial T2-weighted MRI showing hyperintense area of encephalomalacia in this
region. C, axial T2-weighted MRI depicting Wallerian degeneration down to the left cerebral peduncle. FLAIR, fluidattenuated inversion recovery; MRI, magnetic resonance imaging; PLIC, posterior limb of the internal capsule.

testing was performed to ensure appropriate side-effect thresholds, and the lead was secured at the radiographic target.
Postoperative head CT was negative for hemorrhage but
revealed a right PLIC hypodensity, posterior and medial to
the DBS lead tip. At 6-month follow-up, left facial weakness
and dysarthria had resolved. PD symptoms were very well
controlled with DBS and medications. In particular, the left
arm continued to be free of tremor and rigidity with minimal
stimulation amplitude (1.5 V).
Case Report 4
A 51-year-old woman presented with a 10-year PD history
characterized by progressive bradykinesia, rigidity, and tremor
that eventually required bilateral GPi DBS implantation. GPi was
targeted at X = 20.0 mm, Y = 2.0 mm, and Z = 24.0 mm (ACPC length = 25.0 mm). At surgery, left-sided MER of GPi was
recorded from 7 mm above target to 1.5 mm below target.
During MER and just after entering GPi, the patient developed
left gaze preference and right-sided weakness (2/5) but remained
alert and able to follow commands. Within 5 minutes, the gaze
preference resolved and hemiparesis improved to 4/5. Given
quick symptom resolution, a seizure was suspected and surgery
proceeded. The original lead was removed and a second
trajectory was performed 2 mm lateral to the initial trajectory,
with macrostimulation confirmation of excellent therapeutic
thresholds. The right GPi DBS was deferred because of
a persistent, mild right hemiparesis and dysarthria. Postoperative
CT showed no evidence of hemorrhage. A brain MRI obtained
on the same day demonstrated restricted diffusion in the left
globus pallidus and PLIC just posterior and medial to the
electrode (Figure 2). Three months after surgery, she had marked
improvement in right-sided rigidity and tremor with some
persistent dysarthria and mild right hemiparesis.

386 | VOLUME 12 | NUMBER 4 | DECEMBER 2016

Case Report 5
A 52-year-old woman with a 13-year history of tardive dystonia
secondary to neuroleptic use presented for revision of bilateral GPi
DBS. After undergoing initial DBS surgery without resolution of
symptoms 2 years earlier, it was recommended that she undergo
revision of the GPi electrodes. Left GPi was targeted at X =
23.9 mm, Y = 2.3 mm, and Z = 24.8 mm (AC-PC length = 23.1
mm). Beginning with the left hemisphere, MERs were performed
but were poor quality without any clear localization, presumably
because of the gliosis created by the previous macroelectrode
placement. During MER, the patient became somnolent and
developed dysarthria and mild right-sided hemiparesis (4/5) along
with complete resolution of her right-sided dystonic movements.
The DBS electrode was then inserted to target, which was 1 mm
medial to the previously placed electrode. Because of her
somnolence and concern for possible hemorrhage, the rightside implantation was deferred. A postoperative CT scan
demonstrated no evidence of intracerebral hemorrhage. A brain
MRI performed on the same day revealed restricted diffusion in
the distribution of the globus pallidus and PLIC (Figure 3), again
posterior and medial to the DBS lead tip. Three months
postoperatively, the hemiparesis had resolved, but her speech
remained slightly dysarthric. Dystonic symptoms were initially
improved on the right side but subsequently returned.

DISCUSSION
The most commonly discussed complications of DBS surgery
are intracerebral hemorrhage and infection. Symptomatic hemorrhage has a reported incidence of 1.2%, with a 0.7% risk of causing
permanent neurological deficit, without a statistical difference
between targets.5 Acute ischemic stroke is a rare complication of
DBS surgery. Although it is mentioned in some complication

www.operativeneurosurgery-online.com

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

ACUTE ISCHEMIC STROKE IN GPI DBS

FIGURE 2. Case 4. A, axial DW MRI obtained within 1 hour postoperatively demonstrating acute infarct medial and
posterior to the left electrode, corresponding to B, axial ADC map. C, axial T2-weighted MRI showing increased signal of acute
stroke in left PLIC. DW, diffusion-weighted; MRI, magnetic resonance imaging; PLIC, posterior limb of the internal capsule.

analyses,9 only 2 reports in literature describe nonhemorrhagic
infarctions occurring acutely during DBS surgery.10,11 Several
studies have focused on the incidence and risk factors associated
with hemorrhage both before and during surgery,5-9,12 but little is
known about the causes of nonhemorrhagic stroke as a complication of DBS. A recent meta-analysis reported a 2.0% incidence
of nonhemorrhagic stroke with GPi DBS and 0.2% in STN
DBS. Unfortunately, only 1 of the 18 articles referenced in this
meta-analysis actually described an intraoperative stroke in
detail.13 The authors reported a 42-year-old male patient who
experienced hemiparesis and aphasia during the simultaneous
advancement of 5 parallel microelectrodes targeting GPi.13
In our series, the incidence of ischemic stroke during GPi DBS
was 2.14% per lead. Only 1 patient had known stroke risk
factors—hypertension and obesity—for which she was receiving
treatment. None of the patients were past or current smokers

(Table 2). MER was used in all cases in which strokes occurred,
but only 3 patients experienced symptom onset during MER,
suggesting the observed strokes were unlikely to be causally
related to the duration of mapping. Interestingly, we observed
electrical silence during the MER of patient 3. In patient 4,
a second MER tract was performed, for which recordings were
less optimal than the first. In patient 1, symptoms were not noted
until the DBS lead was implanted but before the initiation of
macrostimulation. Finally, in patient 2, symptom onset was
during macrostimulation. Although our patients developed
symptoms at various times during the case, including both
MER and macroelectrode placement, the first clear indication in
each instance of a problem was clinical deterioration.
Only 3 publications detail intraoperative ischemic events
during DBS surgery, across all targets.10,11,14 Three ischemic
strokes were reported in association with STN DBS surgery.

FIGURE 3. Case 5. A, axial DW MRI obtained within 1 hour postoperatively depicting acute infarct medial and posterior to
the left DBS electrode. B, axial ADC map with corresponding hypointensity. C, axial T2-weighted MRI again showing
hyperintense signal in globus pallidus and PLIC. DBS, deep brain stimulation; DW, diffusion-weighted; MRI, magnetic
resonance imaging; PLIC, posterior limb of the internal capsule.

OPERATIVE NEUROSURGERY

VOLUME 12 | NUMBER 4 | DECEMBER 2016 | 387

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

DOWNES ET AL

Larson et al14 described an ischemic event after unilateral STN
DBS: when the generator was activated, the patient experienced
marked decrease in premorbid bilateral tinnitus. A focal area of
ischemia of the anterior body of the caudate was seen on
postoperative MRI. The lead was adjacent to but not within the
lesion, which raises questions about the extent of electrically
induced vasoconstrictive effects. Novak et al11 reported 2 cases of
ischemic stroke associated with STN DBS. In the first case, the
56-year-old patient with vascular risk factors, including prior
ischemia, obesity, and mitral valve disease, developed slurred
speech and contralateral motor weakness during the final of
5 MER passes through the STN. There was a 10-second period of
electrical silence on the MER, which was present simultaneously
with the patient’s clinical deterioration. The authors theorized that
the high voltage used during microstimulation induced small vessel
vasospasm and caused ischemia, represented as electrical silence.
The second was less specific, with confusion and short-term
memory loss occurring in a 67-year-old patient approximately 3 to
6 days after surgery. Seven MER tracks were made to target the
STN in this patient, and after the third pass, the electrical
recordings were unusually quiet throughout the STN area. The
patient had no neurological deficits during surgery. Similar to the
pallidotomy patients who experienced delayed ischemia, this
patient had several vascular risk factors, including hypercholesterolemia, obesity, and previous ischemia. The authors hypothesize
that the delayed ischemia in the periventricular and subcortical
white matter in both cases is related to vasoconstriction. Electrical
stimulation, edema surrounding the leads, or simply mechanical
disruption of small vessels may cause irritation and thus vasospasm
of the perforating arteries, resulting in a lacunar infarct.
To our knowledge, only 1 ischemic stroke of the posterior limb of
the internal capsule during GPi DBS has been reported, which was in
a 34-year-old patient with cervical dystonia.10 The patient was
otherwise healthy and had no stroke risk factors. Right-sided
microelectrode stimulation was performed up to an amplitude of
8 mA to elicit side effects of parasthesias before lead insertion. Blood
pressure was elevated at 153/98 during the neurological examination at this time. The DBS lead was then implanted and secured,
and the contralateral DBS implantation was begun. It was not until
the left-side procedure was started that the patient was found to
have slurred speech and left facial weakness. MRI revealed ischemia
in the right posterior limb of the internal capsule. The authors
suggest small-vessel vasospasm secondary to high-amplitude stimulation as a possible etiology. Alternatively, one might consider that
ischemia was due to physical compression or interruption of an
anterior choroidal branch or a lenticulostriate artery, because there
was a several-minute delay before symptom onset.
Interestingly, ischemia occurring near a radiofrequency pallidotomy lesion has been reported to occur both in a delayed15 and
acute fashion.16 Lim et al15 described 3 of 50 patients who
experienced a delayed stroke in the posterior internal capsule. The
proximity to the lesion suggests that local changes such as
breakdown of the blood-brain barrier contribute to increased
susceptibility of the region to ischemia. The authors suggest that

388 | VOLUME 12 | NUMBER 4 | DECEMBER 2016

the higher vascularity in the gray matter is protective against
ischemia, and thus the white matter tracts of the internal capsule
are more vulnerable. The temporal delay of the onset of ischemia,
which ranged from 10 to 117 days, could be related to the
statistically significant vascular risk factors that existed in all
affected patients. Christoforidis et al16 described 1 case of
ischemic stroke occurring during radiofrequency pallidotomy
in a 54-year-old man with PD. The patient underwent intraoperative MER and stimulation uneventfully. As the thermocoagulation lesion was created, the patient experienced right facial
droop and right pronator drift, and the procedure was aborted.
MRI revealed ischemia in the left posterior limb of the internal
capsule and the radiofrequency lesion in the GPi. Considering the
area of ischemia was not in the immediate vicinity of the lesion,
the authors’ proposed etiology was electrical stimulus causing
thrombosis or vasospasm of a lenticulostriate artery that supplied
the area of infarction. Risk factors for stroke were not mentioned.
Because the clinical manifestation of ischemic damage occurred
at very diverse stages of DBS implantation (during MER, after
DBS lead implant before stimulation, and after macrostimulation)
both in this series and as reported in the literature, it is challenging
to pinpoint a single causative factor for ischemic events during GPi
DBS implantation. On retrospective review, none of the patients
had unusual anatomy in the region of the GPi, such as enlarged
Virchow-Robin spaces or tortuous vessels. Possible etiologies
include physical disruption or compression of the anterior
choroidal or lenticulostriate branches, trauma-induced vasospasm,
or electrically induced atraumatic vasoconstriction. Interestingly,
the location of ischemia was consistently posterior and medial to
the final DBS lead position (and stereotactic trajectory), supporting a likely vascular etiology (Figure 4). With respect to trajectory,
it is possible that the location of the burr hole near the coronal
suture could result in a more posterior entry to GPi, possibly
predisposing to an encounter with the anterior choroidal artery or
lenticulostriate arteries. Of note, we cannot definitively exclude
small hemorrhages associated with these ischemic events, because
the artifact from the DBS lead obscures clear visualization.
Nonthermal vasoconstriction and thrombosis of femoral and
mesenteric arteries and veins using short electric pulses has been
proposed as a novel treatment of traumatic noncompressible
hemorrhages.17 In vivo animal studies have demonstrated that
electrically induced vasoconstriction occurs as a function of the
electrical stimulus amplitude and duration, as well as the pulse
width. For example, hemorrhage from a complete transection of
the femoral artery was controlled by applying 100 microsecond
pulses of 150 V at a pulse width of 10 Hz for 30 seconds, which
resulted in 75% constriction of the artery. Possible mechanisms for
this profound effect are numerous: stimulation of sympathetic
innervation and direct contact stimulating smooth muscle
contraction, but not due to thermal mechanisms.17 The authors
also suggest that electrically induced vasoconstriction is even more
powerful than any reported pharmacological effects. Although
vasoconstriction via this mechanism is plausible for the cases
occurring during macroelectrode stimulation testing, 2 of our cases

www.operativeneurosurgery-online.com

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

ACUTE ISCHEMIC STROKE IN GPI DBS

a definitive cause is not known, and given the variability in the
temporal onset of neurological symptoms and the location of ischemic
lesions, it may be difficult to avoid such an undesirable event in the
future. Nonetheless, this is an important adverse event for clinicians to
be aware of and to discuss with their patients before GPi DBS surgery.
Disclosures
Nader Pouratian has fellowship support from Medtronic. The other authors
have no personal, financial, or institutional interest in any of the drugs, materials, or
devices described in this article.

REFERENCES

FIGURE 4. Axial T2 MRI depicting the typical location of the GPi target, the
area of infarct, and relative color-coded vascular territories of lenticulostriate and
anterior choroidal arteries. GPi, globus pallidus internus; MRI, magnetic resonance imaging.

and 2 cases described by Novak et al occurred during MERs. The
alternative explanation of compromise of branches of the anterior
choroidal artery due to adjacent tissue trauma from the stereotactic
procedure must therefore be considered the leading potential
etiology in this intraoperative series of acute ischemia.
In our series, all patients had CT- or MRI-verified evidence of
injury to the posterior limb of the internal capsule. This region is fed
by the anterior choroidal artery, and a trajectory through the GPi
could certainly lead to compression or spasm of this vessel or
a branch, leading to the observed outcomes. The absence of any
evidence of subarachnoid or intraparenchymal blood would argue
against direct puncture of the vessel, but compression due to microor macroelectrodes is a reasonable hypothesis as well. Given that
these strokes were therapeutic in these patients, this pathoetiology
would be concordant with the therapeutic effects of anterior
choroidal ligation originally described by Irving Cooper.18

CONCLUSION
Although ischemic stroke during GPi DBS implantation has not
been extensively described in the literature, we provide a detailed
account of this adverse event in 5 patients. Unfortunately, no unifying
or modifiable risk factor could be identified. Considering that

OPERATIVE NEUROSURGERY

1. Okun MS, Fernandez HH, Wu SS, et al. Cognition and mood in Parkinson’s
disease in subthalamic nucleus versus globus pallidus interna deep brain
stimulation: the COMPARE trial. Ann Neurol. 2009;65(5):586-595.
2. Weaver FM, Follett KA, Stern M, et al. Randomized trial of deep brain stimulation
for Parkinson disease: thirty-six-month outcomes. Neurology. 2012;79(1):55-65.
3. Follett KA, Weaver FM, Stern M, et al. Pallidal versus subthalamic deep-brain
stimulation for Parkinson’s disease. New Engl J Med. 2010;362(22):2077-2091.
4. Odekerken V, van Laar T, Staal MJ, Mosch A. Subthalamic nucleus versus
globus pallidus bilateral deep brain stimulation for advanced Parkinson’s disease
(NSTAPS study): a randomised controlled trial. Lancet Neurol. 2013;12:37-44.
5. Sansur CA, Frysinger RC, Pouratian N, et al. Incidence of symptomatic hemorrhage
after stereotactic electrode placement. J Neurosurg. 2007;107(5):998-1003.
6. Ben-Haim S, Asaad WF, Gale JT, Eskandar EN. Risk factors for hemorrhage
during microelectrode-guided deep brain stimulation and the introduction of an
improved microelectrode design. Neurosurgery. 2009;64(4):754-762; discussion
762-763.
7. Elias WJ, Sansur CA, Frysinger RC. Sulcal and ventricular trajectories in
stereotactic surgery. J Neurosurg. 2009;110(2):201-207.
8. Zrinzo L, Foltynie T, Limousin P, Hariz MI. Reducing hemorrhagic complications in functional neurosurgery: a large case series and systematic literature review.
J Neurosurg. 2012;116(1):84-94.
9. Videnovic A, Metman LV. Deep brain stimulation for Parkinson’s disease:
prevalence of adverse events and need for standardized reporting. Movement
Disord. 2008;23(3):343-349.
10. Kang DW, Kim HY, Chang JW. Cerebral ischemia related to globus pallidus
internus stimulation for cervical dystonia. Stereotact Funct Neurosurg. 2011;89(4):
201-204.
11. Novak KE, Nenonene EK, Bernstein LP, et al. Two cases of ischemia associated
with subthalamic nucleus stimulator implantation for advanced Parkinson’s
disease. Mov Disord. 2006;21(9):1477-1483.
12. Binder DK, Rau GM, Starr PA. Risk factors for hemorrhage during
microelectrode-guided deep brain stimulator implantation for movement disorders. Neurosurgery. 2005;56(4):722-732; discussion 722-732.
13. Anderson VC, Burchiel KJ, Hogarth P, Favre J, Hammerstad JP. Pallidal vs
subthalamic nucleus deep brain stimulation in Parkinson disease. Arch Neurol.
2005;62(4):554-560.
14. Larson PS, Cheung SW. A stroke of silence: tinnitus suppression following
placement of a deep brain stimulation electrode with infarction in area LC.
J Neurosurg. 2013;118(1):192-194.
15. Lim JY, de Salles AA, Bronstein J, Masterman DL, Saver JL. Delayed internal
capsule infarctions following radiofrequency pallidotomy. J Neurosurg. 1997;87
(6):955-960.
16. Christoforidis GA, Spickler EM, Papaioannou G, Junn F. Lacunar infarct during
pallidotomy: case report. Neuroradiology. 2001;43(4):321-324.
17. Mandel Y, Manivanh R, Dalal R, et al. Vasoconstriction by electrical stimulation:
new approach to control of non-compressible hemorrhage. Sci Rep. 2013;3:2111.
18. Cooper IS. Ligation of the anterior choroidal artery for involuntary movements;
parkinsonism. Psychiatr Q. 1953;27(2):317–319.

Acknowledgments
MRI anatomic illustration was created by Zach Folzenlogen, MD, of University of Colorado, Department of Neurosurgery.

VOLUME 12 | NUMBER 4 | DECEMBER 2016 | 389

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.

DOWNES ET AL

COMMENT

T

he authors have described 5 cases of ischemic stroke in 129 patients
who underwent (mostly) bilateral globus pallidus internus deep
brain stimulation (GPi DBS) implantation over a 4-year period at an
experienced medical center. Their examination of the circumstances in
each case is reasonable and does suggest that, more likely than not, they
were, in fact, seeing the complication of an ischemic infarction in these
cases rather than other potential explanations for the outcomes
(hemorrhage, circuitry disruption, medication alterations, poor electrode placement). This complication, I suspect, will be viewed as
surprising to many in the field, perhaps mostly because there were 5
and not just 1 or 2.
I think the most likely explanation is at the “systems” level, in that
the general trajectory this center takes for GPi is slightly posterior to
what many of us use. Their burr hole position alone suggests this, as

390 | VOLUME 12 | NUMBER 4 | DECEMBER 2016

they describe it at the coronal suture rather than just anterior to it.
While no one can be sure of this conclusion (authors, reviewers, or
readers), their rate of over 2% is higher than other reports of ischemic
stroke or even the more common hemorrhage rates that have been
found (well documented now to be about 1.8%), suggesting they
have a nuance in their technique that makes the possibility of injury
to a lenticulo-striate or anterior choroidal artery slightly more
probable.
In any case, they should be lauded for bringing these cases to the fore
and assessing them in a transparent way, because we all can try to learn
from them. DBS is often a highly beneficial treatment, but it has risks, and
our understanding of complication avoidance starts with the acknowledgement of the complication itself, as these authors have done.
Jeff Arle
Boston, Massachusetts

www.operativeneurosurgery-online.com

Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.