Review

International Journal of Stroke

0(0) 1–13
Central poststroke pain: ! 2017 World Stroke Organization
Reprints and permissions:
A systematic review sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1747493017701149
journals.sagepub.com/home/wso

Jonathan Singer1, Alyssa Conigliaro2, Elizabeth Spina3,

Susan W Law4 and Steven R Levine5

Abstract
Background: Physical, psychological, and/or social impairment can result after a stroke and can be exacerbated by pain.
One type of pain after stroke, central poststroke pain, is believed to be due to primary central nervous system mech-
anisms. Estimated prevalence of central poststroke pain ranges widely from 8% to 55% of stroke patients, suggesting
a difficulty in reliably, accurately, and consistently identifying central poststroke pain. This may be due to the absence of
a generally accepted definition.
Aim: We aimed to clarify the role of thalamic strokes and damage to the spinothalamic pathway in central poststroke
pain patients. Also, we aimed to gain a current understanding of anatomic substrates, brain imaging, and treatment of
central poststroke pain.
Summary of review: Two independent reviewers identified 10,144 publications. Based on Preferred Reporting Items
for Systematic Reviews and Meta-Analysis guidelines, we extracted data from 23 papers and categorized the articles’ aims
into four sections: somatosensory deficits, pathway stimulation, clinical trials, and brain imaging.
Conclusions: Our systematic review suggests that damage to the spinothalamic pathway is associated with central
poststroke pain and this link could provide insights into mechanisms and treatment. Moreover, historical connection of
strokes in the thalamic region of the brain and central poststroke pain should be reevaluated as many studies noted that
strokes in other regions of the brain have high occurrence of central poststroke pain as well.

Keywords
Central poststroke pain, thalamic stroke, spinothalamic pathway, somatosensory, brain imaging, clinical studies, rando-
mized control trials

Received: 2 August 2016; accepted: 15 November 2016

definition.5 Clarifying the relationship between CPSP

Introduction
Stroke often leaves an individual with residual physical,
psychological, and/or social impairment, which may
compromise one’s capacity to carry out activities of
daily living.1 These impairments can result from pain
after a stroke that is categorized as either poststroke
pain (PSP) triggered by peripheral/mechanisms, or cen-
tral poststroke pain (CPSP) due to primary central ner-
vous system (CNS) mechanisms.2–4 Also, CPSP can
have an effect on quality of life (QoL) without there
being any disability or impairment from the stroke.
Prevalence of CPSP among stroke patients has been
found to range widely from 8% to 55%.1 This variabil-
ity suggests a difficulty in reliably identifying CPSP, in
part due to the numerous types of pain that can be
exhibited, including sensory deficits of thermal, tactile,
and/or pinprick experiences of pain.1 Further compli-
cating its diagnosis, CPSP currently lacks a universal
with strokes located in the thalamic region and
damage to the spinothalamic pathway (STP) may
help provide the groundwork for creating a universal
definition and constructing an effective treatment plan.
1
Department of Neurology, SUNY Downstate Medical Center & Stroke
Center, Brooklyn, NY, USA
2
Department of Emergency Medicine, Kings County Hospital Center &
SUNY Downstate Medical Center, Brooklyn, NY, USA
3
NYIT College of Osteopathic Medicine, Brooklyn, NY, USA
4
Department of Neurology, Kings County Hospital Center and SUNY
Downstate Medical Center, Brooklyn, NY, USA
5
SUNY Downstate Medical Center and Stroke Center and Kings County
Hospital Center, Brooklyn, NY, USA
Corresponding author:
Steven R Levine, SUNY Downstate Medical Center and Stroke Center
and Kings County Hospital Center, 450 Clarkson Avenue, Brooklyn, NY
11203, USA.
Email: steven.levine@downstate.edu

International Journal of Stroke, 0(0)

2 International Journal of Stroke 0(0)
Lesion location and affected neural pathways have
been at the cornerstone of multiple theories concerning
the underlying etiology of CPSP. The foundation of the
disinhibition theory states that CPSP is only related to
thalamic strokes.6 Formally referred to by Dejerine and
Roussy6 as thalamic pain, more than 100 years ago,
patients who had strokes affecting their thalamic
region were the ones having pain after their stroke
(Figure 1). Autopsies were performed on five patients
who complained of pain after stroke in the thalamic
region.6,7 Figure 1 illustrates the evolution of identify-
ing CPSP and how in the late 1950s two new theories
were developed. One theory following Dejerine and
Figure 1. Evolution of classifying CPSP.
Ed Linger first described
a paent who had pain
aer their stroke. No
term is developed
(1892)
The term ‘central pain’
is introduced to define
thalamic pain
(1914)
George Riddoch adds the following
descripon of ‘central pain’. “spontaneous
pain and painful overreacon to objecve
smulaon resulng from lesions confined to
the substance of the central nervous system
including dyaaesthesia of a disagreeable
kind.”
(1938)
Central Imbalance theory is introduced
stang that damage to the spinothalamic
pathway is the primary factor associated
with CPSP.
(late 1950’s)
International Journal of Stroke, 0(0)
Roussy’s findings and another (central imbalance
theory (CIT)) stating that the primary cause of CPSP
is by damage to the STP. The CIT acknowledges that
there could be alternative affected pathways that may
cause CPSP (e.g. medial lemniscus pathway, posterior
column-medial lemniscus pathway). Despite the intro-
duction of this new theory, research has largely focused
on the relationship between strokes in the thalamic
region and CPSP.
Differences in clinical presentation, brain imaging,
and specific pathway lesions have made it challenging
to treat patients with CPSP.5,9,10 To address the discre-
pancies in the literature, this systematic review will help
Dr. Roussy has first paent
who complained of pain aer
their stroke. Their symptoms
were associated with thalamic
strokes (e.g. slight hemiplegia,
persistent superficial
hemianeshesia, severe
hemiataxia). Aer five stroke
paents deaths, autopsies
concluded they had thalamic
strokes. This was known as
thalamic pain
(1906)
The term ‘central pain’ is
redefined as Central Post
Stroke Pain and new
theories are developed.
(mid 1950’s)
Disinhibion theory is derived stang that
thalamic strokes are the primary factor for
causing CPSP. The disinhibion theory has
been the widely known and used theory in
research since the 1950’s
(late 1950’s)
Singer et al.
the reader gain a great understanding of CPSP, while
also constructing a more coherent understanding of the
role of thalamic strokes and damage to the STP in
patients with CPSP. Also, it builds upon the current
understanding of anatomic substrates and brain ima-
ging studies related to CPSP. Reviewing the current
literature may also help to gain a greater understanding
of why CPSP does not improve over time, but rather
worsens. Also, this paper addresses evidence-based
results of clinical trials, including randomized con-
trolled trials (RCTs).
Methods
Search strategy
Eligibility assessment was performed independently
and in a standardized manner by two reviewers (JS
and ES) blinded to one another review. A third
reviewer (AC) resolved any disagreements. When
developing the search strategy, we consulted with a pro-
fessional information scientist at SUNY Downstate
library to generate the search terms and databases we
should utilize. We searched the PubMed, CINAHL,
and Web of Science databases starting on 11 January
2016 using keywords ‘‘pain,’’ ‘‘physical pain,’’ ‘‘psycho-
logical pain,’’ ‘‘mental pain,’’ or ‘‘poststroke pain’’ in
combination with the medical subject headings, ‘‘brain
stem infarction,’’ ‘‘infarction, anterior cerebral artery,’’
‘‘infarction, middle cerebral artery,’’ ‘‘cerebral hemor-
rhages,’’ ‘intracranial hemorrhages,’’ ‘‘thalamic,’’
‘‘putaminal hemorrhages,’’ ‘‘basal ganglia hemor-
rhages,’’ ‘‘cerebral vascular disorders,’’ ‘‘intracranial
hemorrhage, hypertensive,’’ ‘‘infarction, basilar
artery,’’ or ‘‘infarction brain stem.’’ We applied lan-
guage limitations (English only).
Eligibility criteria
We included only original, clinical studies, and RCTs.
The studies had to assess CPSP, using a single question
or pain scale. The studies needed a sample size of at
least 10 patients with either hemorrhagic or ischemic
strokes.
Studies were excluded if they (i) examined animals,
(ii) lacked primary data (e.g. editorials, review articles,
or protocol papers), (iii) investigated peripheral PSP
(e.g. shoulder pain, arm pain), (iv) were not published
in a peer reviewed journal, (v) were retrospective studies,
and (iv) or examined individuals under 18 years old.
Results
The initial search yielded 1074 articles, removal of
duplicates resulted in 731 selected articles. Two other
3
articles were added from reference sections of other
papers. All article titles and abstracts were screened
resulting in 111 relevant studies. After reading the full
articles, 88 articles were excluded (28 retrospective
studies, and 60 that only measured PSP). Thus, 23
papers met our criteria and were included in the
review (Figure 2, Preferred Reported Items for
Systematic Reviews and Meta-Analyses (PRISMA)).
We extracted data from 23 articles and categorized
the articles’ aims into four sections: somatosensory def-
icits (five studies), pathway stimulation (three studies),
clinical trials (eight studies) and brain imaging (seven
studies). The selected articles were published between
1988 and 2014.
Somatosensory deficits
Five studies specifically addressed somatosensory def-
icits in CPSP patients.11–15 The five studies reported 474
patients, and four of the five studies used the visual
analogue scale (VAS) to measure pain (see Table 1).
Allodynia, (the production of pain by a normally pain-
less stimulus) and hyperaesthesia (increased response to
stimulus) were the most common somatosensory def-
icits appearing in about 85% of CPSP patients and
were investigated in all five studies.11–15 Three of the
five studies12,14,15 exclusively examined patients who
had been diagnosed with CPSP, while the other two
studies reported 6% and 74% of participants having
CPSP,11–13 respectively.
Tactile and thermal allodynia were investigated in all
five studies using a sensory measure (see Table 1).
When examining patients who have been diagnosed
with CPSP, allodynia as a whole was found to range
from 23% to 85% across studies. Tactile allodynia,
including sensitivity to light or touch, was found to
range from 5% to 55% in CPSP patients, with a
mean of 40% across studies. Thermal allodynia was
found to range from 15% to 56% with the mean of
the five studies being 33%. The impact of allodynia
on stroke patients has been addressed,5 and the results
from these studies show a wide range of prevalence of
allodynia within CPSP. A number of factors may have
contributed to the large variability in prevalence rates
for reported throughout these studies, including the
lack of a universal definition of CPSP and wide vari-
ations in sample sizes utilized across studies.
Table 1 outlines the prevalence of hyperalagesia
(excessive pinprick, cold, and touch sensitivity) in all
five studies.11–15 Four of the studies reported data on
pinprick sensitivity. Three of these four studies found
rates of pinprick sensitivity to be between 69% and
72% of CPSP patients. However, the remaining study
found that only 6% of CPSP patients reported pin-
prick sensitivity.13 Cold sensitivity was reported in all
International Journal of Stroke, 0(0)
4 International Journal of Stroke 0(0)

Figure 2. PRISMA 2009 flow diagram.

PRISMA 2009 Flow Diagram

Idenficaon

Records idenfied through Addional records idenfied

database searching through other sources
(n = 10,144) (n = 2)

Records aer duplicates removed

(n = 733)
Screening

Records screened Records excluded

(n = 344) (n = 233)

Full-text arcles assessed Full-text arcles excluded,

Eligibility

for eligibility with reasons:

(n =111)
Post Stroke Pain
(n = 85)
Retrospecve
(n=28)
Included

Studies included in
qualitave synthesis
(n = 23)

five of the studies and was found to range from 38%

to 100%.11–15 Similar to allodynia, there was a large
variability of hyperalagesia in CPSP patients. The
same aforementioned reasoning for the large variabil-
ity may apply to hyperalegesia, but hyperalgesia was
found at a much higher rate than allodynia.
In an earlier review,5 patients with CPSP were found
to report allodynia more often, but from the results of
the present review, there is compelling evidence that
hyperesthesia is as likely and could possibly be more
prevalent than allodynia. All five of these studies show
high rates of some form of somatosensory deficit with
CPSP patients; however, the underlying causes of these
deficits are unclear. Most of the studies do not refer to
location of the stroke as playing a role of these deficits,
but the theme of the study’s findings were that path-
ways may play a role and future research is
needed11,12,14,15 (Table 1).
STP
The STP’s importance to the development of CPSP has
been highly debated and its role in the syndrome’s
development is still poorly understood.13 As mentioned
above, the CIT states that impairment to STP is asso-
ciated with pathogenesis of CPSP.12,13 Previous studies
have shown that patients who have CPSP can have
lesions outside the thalamic region.12–17 Moreover,
the theory states that the development of CPSP is not
determined by the location of the lesion within the
brain; rather it is dependent on the impairment of
the STP. Furthermore, even partial lesions affecting
the STP lead to a higher risk of developing CPSP.18
In our systematic review, there were three articles that
fit the criteria and specifically investigated the relation-
ship between CPSP and the STP.10,18,19 The three art-
icles had a total of 109 stroke patients, and two of the

International Journal of Stroke, 0(0)

 Table 1. Studies accessing somatosensory deficits in CPSP (n ¼ 5)
Singer et al.

Demographic
References Aim of study pain measures Sensory measures characteristics Outcomes
12
Boivie et al. To examine sensory def- – VAS – Touch, by strokes of – 20 Men and 7 women – The spino-thalamo-cor-
icits in patients who cotton, cold, by apply- – All patients had CPSP tical pathway playing a
have CPSP ing a round surface of a major role in CPSP
tuning fork of room
temp, pin-prick

Vestergaard et al.15 To investigate sensory
abnormalities in
patients with CPSP
– McGill pain question- – Sensation was graded as – 6 Men and 5 women – The spino-thalamo-cor-
naire either increased, – All patients had CPSP tical pathway plays a
– VAS normal or decreased role in CPSP
using the non-affected
side as references

Andersen et al.11 To investigate the inci-
dence of CPSP pro-
spectively in an
unselected stroke
population
– Patient must have com- – Allodynia was defined – 110 Men and 157 – Abnormal thermal sen-
plaint of pain occurring as pain due to a touch women sation is more frequent
after stroke or temperature stimu- – 87 Patients reported in the pain group with
lus which normally does somatosensory deficits somatosensory dys-
not provoke pain – 71 Had no pain and 16 function than in the
had pain corresponding non-pain
group

Bowsher13 To study the clinical and
pathophysiological fea-
tures of central pain
due to damage to the
CNS
– McGill Pain – Quantitative sensory – 156 Patients (112 had – The critical deficit
Questionnaire perception threshold an ischemic stroke, 19 seems to be thermal
– VAS testing was used had a hemorrhagic and pinprick sensations
stroke) – CNS plays a role in
– 102 Patients had CPSP CPSP

Greenspan et al.14 To evaluate the sensory
loss and sensory
abnormalities to dis-
cern submodalitiy rela-
tionships between
sensory features of
CPSP
VAS: visual analogue scale; CPSP: central poststroke pain; CNS: central nervous system.
– VAS – Warm and cold was – 13 Consecutive patients
derived using the –
Peltier stimulator
– Semmes-Weinstein
monofilament kit
– Tactile allodynia occurs
with CPSP when disturbances to
thermal/pain pathways
occur and this spares
the tactile-signaling
pathways. Therefore,
cold hypoesthesia is
neither necessary nor
sufficient for cold
allodynia

International Journal of Stroke, 0(0)

5
6 International Journal of Stroke 0(0)
three used the VAS to evaluate pain (see Table 2). The
first study10 examined 27 stroke patients affected by
CPSP and calculated their somatosensory evoked poten-
tials (SEP) scores10 (see Table 2). They found that calcu-
lated SEPs in patients with CPSP varied between lesions
of the brain stem (n ¼ 8), supratentorial or extrathalamic
regions (n ¼ 6), thalamus (n ¼ 9), and unidentified loca-
tion (n ¼ 4). The authors also stated that there was no
direct correlation between CPSP and SEP response.
When SEPs scores were compared to abnormal sensory
findings (thresholds) in vibration (Vibrameter), touch
(von Frey hairs), and innocuous and noxious temperature
(Thermotest), several correlations were observed. A high
threshold for touch and vibration was correlated with a
more severe SEP score on the contralateral side of the
pain. This relationship with tactile and vibration sensa-
tion indicates an altered medial lemniscus pathway, that
normally plays an important inhibitory role in pain regu-
lation, further supporting the CIT. Moreover, all CPSP
patients were found to have reduced temperature sensi-
tivity. They concluded that damage to the STP has to
occur to develop CPSP.
The second study20 used diffusion tensor tractogra-
phy (DTT) to visualize the STP and measure its integ-
rity in 30 participants.18 Patients were placed into one
of two groups, CPSP (n ¼ 16) or non-CPSP (n ¼ 14). A
significant decrease in STP volume was reported in the
CPSP group when compared to the non-CPSP group.
The authors reported no significant correlations
between both CPSP and non-CPSP groups with the
two other DTT parameters, fractional anisotropy
(FA) and mean diffusivity (MD), representing the
degree of microstructure directionality and water diffu-
sion, respectively. They concluded that lack of signifi-
cant correlations with FA and MD could be due to
partial integrity of the STP, and that injury of the
STP is a requirement for development of CPSP.
The integrity of the STP using DTT was also inves-
tigated in the third study,19 but it incorporated patients
with complete lesions of STP. A total of 52 patients
were split into two groups, preserved (n ¼ 34) and dis-
rupted (n ¼ 18) STP and subgroups of CPSP (preserved
n ¼ 16, disrupted n ¼ 3) and non-CPSP (preserved
n ¼ 18, disrupted n ¼ 15). The preserved group had a
higher incidence of CPSP and despite the subgroup,
tactile sensation was decreased in the preserved
group, which may represent a dysfunctional regulation
system. Spinothalamic tract volume in the CPSP pre-
served group was significantly less than the volume of
the non-CPSP preserved group and the control group.
Overall, they concluded that the STP plays a role in the
development of CPSP and that patients with partial
injury of the STP might have an increased risk of CPSP.
These three studies display consistent evidence that
the STP plays a major role in CPSP. They also delineate
International Journal of Stroke, 0(0)
that CPSP can occur even when the stroke is not in the
thalamic region10,12,17,18 and when damage to the STP
occurs. Therefore, it may be important to focus on the
CIT when attempting to identify the etiology and treat-
ment of CPSP. The CIT supports the findings that any
disturbance to the pathway will result in pain after a
stroke. However, the reason damage to these pathways
cause CPSP is still unknown.
Brain imaging
This systematic review identified seven studies17,21–26
that addressed brain imaging and CPSP. Five of the
seven studies used MRI to examine brain lesions, one
used CT and one used MRI and SPECT. Four of the
seven studies17,21–23 found a high prevalence of CPSP
outside of the thalamic region. All four concluded that
patients who have strokes outside of the thalamic
region could develop CPSP. The other three studies24–26
found that CPSP was more likely to occur in the thal-
amic region. Several reasons (see Table 3) may have
caused these equivocal results in these seven studies.
One reason for these differing results could be due to
limitations in the studies (see Table 3). The four studies
that17,21–23 found that strokes outside of the thalamic
region had a high prevalence of CPSP did not have a
control group. Also, two of these studies examined only
patients who had a stroke in their thalamus, one exam-
ined patients who had strokes only in their ventropos-
terior thalamic nucleus and one examined patients who
only had medial medullary infarction. In the three stu-
dies24–26 that concluded that CPSP is related to thalamic
infarcts, two only examined thalamic strokes and one
examined only patients with diagnosed CPSP. All
seven studies had wide variability from time of diagnoses
of stroke (imaging) and diagnoses of CPSP (see Table 3).
CPSP often arises after initial stroke imaging and issues
of selection bias due to major time variability (from time
of imaging and diagnoses of CPSP) could cause both
type I and type II errors. If imaging was done after
CPSP was diagnosed, it may rule out other confounding
variables (e.g. new infarct) and reduce the risk of a type I
or type II errors. However, these studies only reported
baseline imaging even when CPSP was diagnosed three
years later. The results of the present systematic review
indicate that the major limitations have impeded reliable
and consistent findings about CPSP (Table 3). Future
studies with larger sample sizes, a control group, and no
selection bias are warranted to examine the role of CPSP
with lesion location.
Clinical trials
In this section, we summarized the findings of
studies investigating treatments for CPSP. As depicted
 Singer et al.

Table 2. Studies accessing pathways in the brain (n ¼ 3)

References Aim of study Measures Demographics Outcomes Conclusions

10
Holmgren et al. Correlation between lesion SEPs Vibration: Vibrameter – 20 Men and 7 women – No direct correlation Lesions of the STP have a
location and sensory Touch: von Frey hairs patients with CPSP, 9 of with CPSP and SEPs major role in the devel-
abnormalities Innocious and noxious tem- which that thalamic – All patients had reduced opment of CPSP.
perature: Thermotest lesions temperature sensibility
– High threshold for touch
and vibration had a posi-
tive correlation with high
SEP scores on the
contralateral side of pain

Hong et al.18 Investigated the relationship
between integrity of the
STP and CPSP in patients
with ICH.
Pain: VAS Total: 30 pts, 16 men, 14 – Relative tract volume of
Tactile: NSA women the STP in the partial
CPSP group was signifi-
cantly lower than the
non-CPSP group with no
significant difference in
FA or MD parameters
– Lesions in STP is neces-
sary for development of
CPSP
– Partial lesions of the STP
have increased preva-
lence of CPSP

Hong et al.19 (2012)
SEP: somatosensory-evoked potential; STP: spinothalamic pathway; ICH: intracranial hemorrhage; CPSP: central poststroke pain; VAS: visual analogue scale; FA: fractional anisotropy; MD: mean diffusivity
Examined the relationship
of CPSP with the integ-
rity of STP
Pain: VAS – 52 Patients, 34 men and – Tactile sensation thresh- – STP damage is necessary
Tactile: NSA 18 Women old was higher in patients for CPSP development.
with preserved STP – Prevalence of CPSP is
when compared to non- higher in patients with
preserved STP partial injury of STP than
– Volume of the CPSP sub- with complete injury
group of the preserved
group was smaller than
the non-CPSP subgroup
of the preserved group
– The prevalence of CPSP
was higher within the
preserved group in com-
parison to the non-pre-
served group.

International Journal of Stroke, 0(0)

7
 8

Table 3. Brain imaging studies (n ¼ 7)

References Demographics Aims Location of the strokes Findings Limitations

17
Leijon et al. (1989) – 20 Men 7 Women; – Characterize patients – No lesion: 7 patients – A small percentage of – CT scans; All CPSP
– Pain duration:44 months with CPSP with regard to – Cerebellum: 3 patients patients with CPSP have patients; selection bias;
– Time since stroke: 9–188 type and location – Thalamus: 9 patients thalamic lesions. The did not look at individual
months – Basal ganglia: 12 patients patients were found to parts of thalamus
– Cortical/sub-cortical: 13 have decreased tempera-
patients ture sensibility

Bowsher et al.23 – 73 CPSP participants – Correlate MRI, sensory – Brain stem: 27 CPSP; 2 – Different stroke sites – Two groups had different

International Journal of Stroke, 0(0)

– 13 No pain after stroke
changes and lesion control produce different pat- infarct locations; age and
location – Ventroposterior thalamic terns of sensory deficits gender were not
nucleus: 49 CPSP; 3 con- reported
trol
– Basal ganglia: 33 CPSP; 6
control
– Parietal cortex: 13 CPSP;
4 Control

Kim et al.27 (2009)
– 13 Men 7 women – Investigated MMIs and – All participants had lenti- – Symptoms included – No control group; selec-
– Pain average 5.6 (VAS CPSP culo-capsular hemor- numb, cold, burning, tion bias; did not address
Scale) rhages (LCH) aching swollen and thalamic region
squeezing. LCH should
be considered one of the
causes of CPSP.

Bowsher24 – 122 Participants – Compare lesion site to – All participants had an – CPSP is mostly found in – No control group;
– 39.5% of CPSP had CPSP while addressing ischemic stroke the thalamus region; selection bias
Allodynia sensory abnormalities pathways can play a role
– 29% Had CPSP as well. allodynia has
been found to correlate
with lesion location.

Kalita et al.21
– 20 Men 3 women, Age – Compare findings on – Right side: 10; left side:
mean 52 CPSP between MRI and 13
SPECT – Thalamic stroke: 9
patients; subcortical: 5,
parietal cortex: 6,
medulla:1, multiple
infarcts: 3
– MRI and SPECT were – Only CPSP patients; small
not different in CPSP sample size; selection
patient with and without bias; did not look at
allodynia no matter brain individual parts of
location. thalamus
(continued)
International Journal of Stroke 0(0)
Singer et al. 9

CPSP: central poststroke pain; VAS: visual analogue scale; MRI: magnetic resonance imaging; SPECT: single-photon emission computed tomography; LCH: lenticulocapsular hemorrhage; VPL: ventral
in Table 4 clinical trials on treatments for CPSP,

from stroke (imaging) to

– Only examined thalamic

– Small sample size; Time

CPSP diagnoses ranged

from 2 months to 132
including spinal cord and motor cortex stimulation, tri-

strokes; small sample

size; Selection bias
cyclic antidepressants, opioid antagonists, acupuncture,
and anticonvulsants. These trials range from 1989 to
2013 and represent clinical studies and RCTs identified
Limitations

in our systematic review that attempted to determine

months efficacious treatment options for CPSP.
The following trials include invasive as well as non-
invasive methods. To date, effective clinical trials for
CPSP are limited.29,30 Presently, tricyclic antidepressant
– VPL was present in both

CPSP participants. More amitriptyline is the only oral medication with a benefi-

pattern was evident in

anterior medial lesion
affected mainly in the
– Lateral and posterior
thalamic nuclei were
CPSP and non-CPSP

cial effect on CPSP. However, a number of studies have

reported high rates of patient non-responders and
intolerable side effects, which often result in medication
cessation.15,31,32
patients.

controls
Findings

Spinal cord and motor cortex stimulation

One clinical study (n ¼ 30) addressed CPSP with percu-
taneous spinal cord stimulation (SCS) during a continu-
– All participants had thal-

location of stroke (thal-

and 10 control patients
– 10 Control participants

ous treatment phase of two to seven days for CPSP

Location of the strokes

matched gender and

amus origin and not

patients.31 Global pain relief measured at follow-up indi-

thalamic origin)

cated that 30% (n ¼ 9) reported good relief, 20% (n ¼ 6)

amic strokes

reported fair relief, and 50% (n ¼ 15) reported poor

relief. VAS score decreased from 8 to 6 (p ¼ 0.001)
from baseline to post-treatment.31 Although the results
of this study offer compelling preliminary evidence that
SCS may provide improved pain relief for individuals
with CPSP, there are notable limitations including
small sample size, high percentages of individuals report-
– Investigate location of

ing poor to no relief, and lack of a control arm. As a

– Aimed to determine
mutual lesion site in
patients with CPSP
stroke and lesion

result, findings from this study must be interpreted with

caution until they can be replicated with larger samples
sizes with a more stringent research design.
location

Only one study was identified that examined motor

cortex stimulation as a potential treatment for CPSP.
Aims

In this study, 11 patients with CPSP were treated at the

pre-central gyrus.32 Effective treatment (see Table 4)
was observed in 73% of patients with pre-central stimu-
– 22 Men 8 women, 18 had

– 7 Men 3 women in both

control group and CPSP
others were controlled

lation, while 100% reported no pain reduction with

diagnosed CPSP, the

post-central stimulation. Additionally, 45% continued

to achieve pain relief with pre-central stimulation at
their two-year follow-up. The authors32 proposed that
Demographics

posterior lateral; CT: computed tomography.

the mechanism of action is likely due to deafferented

groups

(partially or totally interrupted afferent signals) noci-

ceptive neurons using motor provocation, which was
more readily activated in the STP. Along with STPs,
coexisting tracts were also stimulated, suggesting the
Krause et al.28 (2012)

need for further investigation on anatomical delinea-

Table 3. Continued

tions to narrow down the therapeutic focus.

Sprenger et al.22
References

Tricyclic antidepressants
A double-blind RCT (n ¼ 15) to evaluate amitriptyline
(75 mg/d), cabamazepine (800 mg/d), and placebo’s

International Journal of Stroke, 0(0)

 10
Table 4. Clinical studies and RCT for CPSP (n ¼ 8)

No. of
Frequency and duration of CPSP Pain score baseline/pain
References Treatments treatment(s) patients Pain measure(s) score post-treatment p
17
Leijon et al. Amitriptyline (75 mg) Four weeks 15 – Verbal scale – Baseline/post-treat- <.05 for
Cabamazpine (800 mg) – Global assessment of ment (mean): Amitriptyline vs. pla-
Placebo change in pain scale Amitriptyline: (daily): cebo in global assess-
4.7/4.2 ment of change in pain
Cabamazepine: (daily):
4.6/4.2
Placebo: (daily): 5.5/5.3

International Journal of Stroke, 0(0)

Baintn et al.33 8 mg Intravenous nalox- Day cases on two occa- 20 – VAS – Differences in VAS 0.86
one in 20 ml manufac- sions, separated by – Five-word pain scale score:
turer’s vehicle two to three weeks Naloxone: 9.35
Saline: 10.05

Tsubokawa et al.32 Motor cortex stimulation One time 11 – VAS – Pre-central stimula- NR
tion ¼ 73% pain reduc-
tion
– Post-central
stimulation ¼ No pain
relief

Vestergaard et al.34 Lamotrigine (200 mg); Two eight-week treat- 30 – Global pain scale – Ongoing daily pain: 0.02 for global assessment
placebo ment periods sepa- – Ongoing pain scale baseline/post-treat- of pain for lamotrigine
rated by two-week ment vs. placebo
washout period Lamotrigine: 6/5,
Placebo:6/7
– Global pain:
Lamotrigine ¼ 3 (mod-
erate pain), Placebo ¼ 4
(strong pain)

Aly et al.30 Percutaneous spinal cord Continuous two to seven 30 – VAS VAS: <.001
stimulation days – Patient global impres- – Baseline: (median) ¼ 8
sion of change scale – Post: median ¼ 6

Kim et al.35 Pregabalin – 150–600 mg/ 12 Weeks (four-week 220 – Daily pain rating scale – Baseline/post-treat- .578
d, final maximum dose dose adjustment, eight- ment
placebo week maintenance) Pregabalin (mean): 6.5/
4.9
Placebo (mean): 6.3/5
(continued)
International Journal of Stroke 0(0)
Singer et al. 11

effect on pain intensity was completed over a four-week

Bee venom: .007; saline:

treatment period. Table 4 shows mean pain intensity
scores during the fourth week were identical for ami-
triptyline and cabamazepine. Global assessment of pain
at the end of each treatment phase demonstrated a dif-
ference in effect for amitriptyline vs. placebo (p ¼ 0.05)

.013
with improvement reported for 10 patients from the
NR

amitriptyline group and one from placebo.

p
Pain score baseline/pain

Saline (median): 64.50/

Opioid antagonists
Placebo (median): 7/7

Bee venom (median)

– Baseline/post-treat-

– Baseline/post-treat-
score post-treatment

One double-blind trial (n ¼ 20) assessed the effects of

Levertiracetam

8 mg intravenous naloxone in a 20 ml manufacturer’s

(median): 7/7

vehicle and saline on CPSP on two occasions separated

72/35.5

by two to three weeks33 (Table 4). The results indicated

ment

ment

no immediate effect of naloxone on pain. Analyses of

53

before/after differences between naloxone and saline for

all patients showed no significance (p ¼ 0.86). The
authors also noted that none of the subjects who experi-
enced immediate pain relief were able to maintain
– Likert pain scale

improvement when long-term analyses concluded at

questionnaire
Pain measure(s)

two weeks. This further suggests that naloxone is inef-

– Mcgill pain

fective for the treatment of CPSP pain.33

– VAS

Acupuncture
A single-blind RCT of apipuncture (see Table 4) and
patients

saline acupuncture on pain over a three-week treatment

No. of
CPSP

period found a significant reduction in the median VAS

42

20

score (36.5 vs. 11.50; p ¼ 0.009).29 These results suggest

that apipuncture might have an effect on reducing
Frequency and duration of

CPSP. Further studies including larger sample sizes,

Twice weekly, for three
lowed by two eight-
baseline period, fol-
24 Weeks: four-week

long-term follow-up, and double-blind trials would be

week treatment

beneficial in determining the efficacy of apipuncture for

this population.
treatment(s)

periods

weeks

Anticonvulsants
VAS: visual analogue scale; RCT: randomized controlled trial.

An RCT to evaluate the effect Lamotrigine’s (200 mg/

d) effect over two eight-week treatment periods
Bee Venom Acupuncture

(n ¼ 30).34 Both the treatment and control group indi-

Levetiracetam (3000 mg/
d, maximum dose)

cated a baseline ongoing daily pain score of 6 (11-

point scale) and post-treatment scores of 5 and 7,
(.05 ml); saline

respectively. At the end of treatment, global pain

Treatments

placebo

scores were measured as 3 (moderate pain) for

Lamotrigine and 4 (strong pain, p ¼ 0.02) for pla-
cebo.34 The study also implemented a global assess-
ment of pain on daily activities, which did not indicate
significant differences between Lamotrigine and pla-
Table 4. Continued

36
Jungehulsing et al.

cebo (p ¼ 0.11). The results suggest that Lamotrigine

is not effective monotherapy for CPSP.
Cho et al.29
References

One double-blind RCT evaluated Pregabalin’s

(n ¼ 220) effect on pain intensity, was assessed in a
12-week trial.35 Results indicated that although the
weekly pain score decreased for both treatment

International Journal of Stroke, 0(0)

12
groups (pregabalin: 6.3–4.9, placebo: 6.3–5.0), there
were no significant differences (p ¼ 0.578).The authors
delineate that the majority of patients treated with
pregabalin did not achieve a 30% or 50% reduction
in pain at endpoint compared to baseline as hypothe-
sized. Additionally, 70% of pregabalin and 55% in pla-
cebo patients experienced adverse events.35 Serious
adverse events were experienced in six patients receiving
pregabalin and two patients in the placebo arm.
A double-blind RCT examined the effect of levetir-
acetam (n ¼ 42) on pain intensity over 24 weeks.36 Side
effects including, tiredness, increased pain, dizziness,
pruritus, and headache were substantially more
common in the treatment group than the control
group, during the first treatment period. The authors
reported that levetiracetam did not improve spontan-
eous or evoked pain and median pain scores did not
differ between groups. They concluded that levetirace-
tam was not effective in reducing pain for CPSP
patients.
Interpretation
Although a number of the above clinical trials29,30
reported reduction in CPSP symptoms, a substantial
number of limitations (e.g. not multi-centered; no repli-
cation studies with diverse populations; small sample
sizes; small effect sizes) were noted across studies, result-
ing in decreased generalizability of results. We are unable
at this time to report any gender, race, or age differences
in the trials, as they were not reported in the analysis and
would likely have been underpowered. The brain stem
and motor cortex stimulation studies show promise, but
they were preliminary studies with small samples, and no
significant values were reported. This systematic review
of clinical trials did not identify any study that gave a
clear picture of how to treat or identify CPSP.
Conclusions and limitations
There were three limitations in this systematic review.
The first is that we examined papers in English and the
second is we only used three major databases. However,
more than 10,000 articles were identified during the
screening process. The third is publication bias, as
negative studies about CPSP were not published and
could create bias with the findings in this systematic
review. Even though CPSP has been linked to the thal-
amic region of the brain since the early 1900s, the lit-
erature showed equivocal results when examining
location of stroke and CPSP. Overall, our review,
which used the highest protocol for systematic reviews
(PRISMA Guidelines), of the relevant literature indi-
cates that there is a salient relationship between damage
to the STP and CPSP.
International Journal of Stroke, 0(0)
International Journal of Stroke 0(0)
Future investigation is pivotal in creating a universal
definition of CPSP. Further investigating brain imaging
and CPSP with studies that have stronger methodology
may give way to the development of clinical trials that
can address treatment options for CPSP. Also, we were
unable to identify any research that addresses why
CPSP worsens over time, but hypothesize and hope
that future research will examine the role of damage
to the STP networks as the primary reason for this
progression. The results summarized in this systematic
review support the CIT, suggesting that there are alter-
native affected pathways that may cause CPSP.
Moreover, the link between strokes in the thalamic
region of the brain and CPSP should be reevaluated
as many studies indicated variability and/or a lack of
association.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with
respect to the research, authorship, and/or publication of this
article.
Funding
The author(s) received no financial support for the research,
authorship, and/or publication of this article.
References
1. Buono VL, Corallo F, Placido B and Silvia M. Coping
strategies and health-related quality of life after stroke.
J Health Psychol 2015; 10: 1–13.
2. Singer J and Levine SR. Central post-stroke pain: the-
ories, diagnosis and treatment. Future Neurol 2016; 11:
5–8.
3. Harno H, Hapaniemi E, Putaala J, et al. Central post-
stroke pain in young ischemic stroke survivors in the
Helsinki Young Stroke Registry. Neurology 2014; 13:
1147–1154.
4. Chae J, Mascarenhas M, Yu DT, et al. Poststroke shoul-
der pain: its relationship to motor impairment, activity
limitation, and quality of life. Arch Phys Med Rehabil
2007; 3: 298–301.
5. Klit H, Finnerup NB and Jensen TS. Central post-stroke
pain: clinical characteristics, pathophysiology, and man-
agement. Lancet Neurol 2009; 9: 857–868.
6. Dejerine J and Roussy G. Le syndrome thalamique. Rev
Neurol 1906; 14: 521–532.
7. Henry JL, Lalloo C and Yashpal K. Central poststroke
pain: an abstruse outcome. Pain Res Manage 2008; 13:
41–49.
8. Oh H and Seo W. A comprehensive review of central
post-stroke pain. Pain Manage Nurs 2015; 16: 804–818.
9. Kumar G and Soni CR. Central post-stroke pain: current
evidence. J Neurol Sci 2009; 284: 10–17.
10. Holmgren H, Leijon G, Boivie J, Johansson I and
Lievska L. Central post-stroke pain – somatosensory
evoked potentials in relation to location of the lesion
and sensory signs. Pain 1990; 40: 43–52.
Singer et al.
11. Andersen G, Vestergaard K, Ingeman-Nielsen M and
Staehelin TS. Incidence of central post-stroke pain.
Pain 1995; 61: 187–193.
12. Boivie J, Leijon G and Johansson I. Central post-stroke
pain – a study of the mechanisms through analyses of the
sensory abnormalities. Pain 1989; 37: 173–185.
13. Bowsher D. Central pain: clinical and physiological char-
acteristics. J Neurol Neurosurg Psychiatry 1996; 61:
62–69.
14. Greenspan J, Ohara S, Sarlani E and Lenz FA. Allodynia
in patients with post-stroke central pain (CPSP) studied
by statistical quantitative sensory testing within individ-
uals. Pain 2004; 109: 357–366.
15. Vestergaard K, Nielsen J, Anderson G, et al. Sensory
abnormalities in consecutive, unselected patients with
central post-stroke pain. Pain 1995; 61: 177–186.
16. Boulos MN, Brewer AC, Karimkhani C, Buller DB and
Dellavalle RP. Mobile medical and health apps: state of
the art, concerns, regulatory control and certification.
Online J Public Health Inform 2014; 5: 229.
17. Leijon G, Boivie J and Johansson I. Central post-stroke
pain – neurological symptoms and pain characteristics.
Pain 1989; 36: 13–25.
18. Hong JH, Bai DS, Jeong JY, et al. Injury of the spino-
thalamo-cortical pathway is necessary for central post-
stroke pain. Eur Neurol 2010; 64: 163–168.
19. Hong JH, Choi, Chang CH, et al. The prevalence of cen-
tral poststroke pain according to the integrity of the
spino-thalamo-cortical pathway. Eur Neurol 2012; 67:
12–17.
20. Kumar A, Al-Bader M, Al-Thani H, et al. Multicenter
cross-sectional study of asymptomatic peripheral arterial
disease among patients with a single previous coronary or
cerebrovascular event in the Arabian Gulf. Curr Med Res
Opin 2014; 30: 1725–1732.
21. Kalita J, Kumar B, Misra UK and Pradhan PK. Central
post stroke pain: clinical, MRI, and SPECT correlation.
Pain Med 2011; 12: 282–288.
22. Sprenger T, Seifert CL, Valet M, et al. Assessing the risk
of central post-stroke pain of thalamic origin by lesion
mapping. Brain 2012; 135: 2536–2545.
23. Bowsher D, Leijon G and Thuomas KA. Central post-
stroke pain Correlation of MRI with clinical pain
13
characteristics and sensory abnormalities. Neurology
1998; 51: 1352–1358.
24. Bowsher D. Allodynia in relation to lesion site in central
post-stroke pain. J Pain 2005; 6: 736–740.
25. Kim JS. Central post-stroke pain or paresthesia in lenti-
culo-capsular hemorrhages. Neurology 2003; 61: 679–682.
26. Cho ST, Park JY, Jung WS, et al. Bee venom acupunc-
ture point injection for central post stroke pain: a prelim-
inary single-blind randomized controlled trial.
Complementary Ther Med 2013; 21: 155–157.
27. Kim JS, and Han YS. Medial medullary infarction: clin-
ical, imaging and outcome study in 86 acute, consecutive
patients. Cerebrovasc Dis 2009; 27: 6.
28. Krause T, Brunecker P, Pittl S, et al. Thalamic sensory
strokes with and without pain: differences in lesion pat-
terns in the ventral posterior thalamus. J Neurol
Neurosurg Psychiatry 2012; 83: 776–784.
29. Leijon G and Boivie J. Central post-stroke pain – a con-
trolled trial of amitriptyline and carbamazepine. Pain
1989; 36: 27–36.
30. Aly MM, Youichi S, Koichi H, et al. Spinal cord stimu-
lation for central poststroke pain. Neurosurgery 2010; 67:
206–212.
31. Kim JS. Pharmacological management of central post-
stroke pain: a practical guide. CNS Drugs 2014; 28:
787–797.
32. Tsubokawa T, Katayama Y, Yamamoto T, et al. Chronic
motor cortex stimulation in patients with thalamic pain.
J Neurosurg 1993; 78: 393–401.
33. Baintn T, Fox M, Bowsher and Wells C. A double-blind
trial of naloxone in central post-stroke pain. Pain 1992;
48: 159–162.
34. Vestergaard K, Andersen G, Gottrup H, Kristensen BT
and Jensen TS. Lamotrigine for central poststroke pain a
randomized controlled trial. Neurology 2001; 56:
184–190.
35. Kim JS, Bashford T, Murphy K, et al. Safety and efficacy
of pregabalin in patients with central post-stroke pain.
Pain 2011; 152: 1018–1023.
36. Jungehulsing G, Israel H, Safar N, et al. Levetiracetam in
patients with central neuropathic post- stroke pain – a
randomized, double-blind, placebo-controlled trial. Eur
J Neurol 2013; 20: 331–337.
International Journal of Stroke, 0(0)