Journal of Clinical Neuroscience (2005) 12(4), 399–404

0967-5868/$ - see front matter ª 2005 Published by Elsevier Ltd.

doi:10.1016/j.jocn.2004.07.013

Clinical study

Deep brain stimulation for phantom limb pain

Richard G Bittar1,3,4 MBBS PHD FRACS, Sofia Otero2 BSC, Helen Carter1 RN, Tipu Z Aziz1,2 MD DMEDSC FRCS
1
Department of Neurosurgery, Radcliffe Infirmary, 2University Department of Physiology, University of Oxford; Oxford, UK, 3Departments of Neurosurgery and
Surgery, The Alfred Hospital and Monash University, 4Australasian Movement Disorder and Pain Surgery (AMPS) Clinic; Melbourne, Australia

Summary Phantom limb pain is an often severe and debilitating phenomenon that has been reported in up to 85% of amputees. Its patho-
physiology is poorly understood. Peripheral and spinal mechanisms are thought to play a role in pain modulation in affected individuals;
however central mechanisms are also likely to be of importance. The neuromatrix theory postulates a genetically determined representation of
body image, which is modified by sensory input to create a neurosignature. Persistence of the neurosignature may be responsible for painless
phantom limb sensations, whereas phantom limb pain may be due to abnormal reorganisation within the neuromatrix. This study assessed the
clinical outcome of deep brain stimulation of the periventricular grey matter and somatosensory thalamus for the relief of chronic neuropathic
pain associated with phantom limb in three patients. These patients were assessed preoperatively and at 3 month intervals postoperatively.
Self-rated visual analogue scale pain scores assessed pain intensity, and the McGill Pain Questionnaire assessed the quality of the pain.
Quality of life was assessed using the EUROQOL EQ-5D scale. Periventricular gray stimulation alone was optimal in two patients, whilst a
combination of periventricular gray and thalamic stimulation produced the greatest degree of relief in one patient. At follow-up (mean 13.3
months) the intensity of pain was reduced by 62% (range 55–70%). In all three patients, the burning component of the pain was completely
alleviated. Opiate intake was reduced in the two patients requiring morphine sulphate pre-operatively. Quality of life measures indicated a
statistically significant improvement. This data supports the role for deep brain stimulation in patients with phantom limb pain. The medical
literature relating to the epidemiology, pathogenesis, and treatment of this clinical entity is reviewed in detail.
ª 2005 Published by Elsevier Ltd.

Keywords: pain, deep brain stimulation, phantom limb

INTRODUCTION
Pain referred to an absent extremity, known as phantom limb pain,
is a common sequel to the amputation of an arm or leg. It may fol-
low traumatic limb amputation, as a result of wartime conflicts,
terrorist attacks, and motor vehicle accidents, or surgical amputa-
tion for peripheral vascular disease and malignancy.
Whilst the term “phantom limb” was coined by Mitchell in
1871,1 it was the French surgeon Par
who first described the
phenomenon of post-amputation sensation some three centuries
earlier.2 It is now accepted that amputees almost invariably expe-
rience non-painful phantom limb sensations,3 and 55–85% suffer
from phantom limb pain.4–7
The treatment of phantom limb pain is frequently difficult, and
comprises pharmacotherapy, non-invasive techniques, and a vari-
ety of surgical procedures. We present the results of chronic deep
brain stimulation (DBS) in three patients with phantom limb pain,
and review the contemporary literature pertaining to this debilitat-
ing condition.
METHODS
Three patients (3 male) with phantom limb pain were studied
(Table 1). All patients described pain which included a strong
burning component. This pain was refractory to multiple pharma-
cological and non-pharmacological therapies, including epidural
morphine, TENS and ultrasonic therapy, as well as oral amitripty-
line, carbamazepine, and opioids.
Received 28 May 2004
Accepted 26 July 2004
Correspondence to: Richard G. Bittar MBBS PhD FRACS, Australasian
Movement Disorder and Pain Surgery (AMPS) Clinic, Suite 5, Level 4,
517 St Kilda Road, Melbourne, Victoria 3004, Australia. Tel.: +61 3 9821 5718;
E-mail: neurosurgeon@ampsclinic.com
Preoperative evaluation of the patients was carried out using the
SF-36 v-2 questionnaire (to assess general health) and the McGill
Pain Questionnaire (MPQ) (to assess pain).8 The patients kept a
pain diary to assess pain intensity on a self-rated visual analogue
scale (VAS) at certain points throughout the day. VAS pain scores
were taken after a period of rest in order to avoid distortion by
exercise or activity. The outcome of the preoperative analysis en-
abled comparisons to an analogous postoperative analysis that
would assess the clinical outcome of DBS on the patient’s pain.
Quality of life was assessed pre- and postoperatively using the
EUROQOL EQ-5D VAS (0 = worst imaginable health state,
100 = best imaginable health state). These procedures were ap-
proved by the Local Ethics Committee, Radcliffe Hospitals
NHS Trust, Oxford, UK.
A magnetic resonance imaging (MRI) scan of the patient’s
brain had been performed several weeks pre-operatively. A
CRWe base ring was fixed to the patient’s skull under local
anaesthetic, and stereotactic computed tomography (CT) scans
were obtained. The CT and MRI scans were fused using Image
Fusion
and Stereoplan
(Radionics Inc., Burlington, MA,
USA). The coordinates of the target stimulation site in the contra-
lateral periventricular gray (PVG) and sensory thalamus were cal-
culated, and an appropriate trajectory determined.
Surgery was carried out under local anaesthesia, as patient feed-
back during the procedure was paramount. The scalp was prepared
with an alcoholic chlorhexidine solution. The stereotactic arc sys-
tem was prepared by setting the coordinates for stimulation,
according to values previously determined in the pre-surgical
planning. The stereotactic arc was then attached to the base ring,
which was firmly attached to the operating table using an adaptor.
A frontal scalp incision, overlying the coronal suture in the mid-
pupillary line, was made. A twist drill was then used to perforate
the skull. Passage of a Radionics TM radiofrequency electrode
(with impedance monitoring to detect transgression of the lateral

399
400 Bittar et al.
Table 1 Demographic and clinical data
Patient Age (years) Sex Aetiology Amputated limb
1 53 Male Vascular insufficiency Lower
2 76 Male Trauma Lower
3 38 Male Trauma Upper
ventricle) was followed by insertion of a Medtronic 3387 deep
brain stimulation electrode (Medtronic Inc., Minneapolis, MN,
USA). This electrode has four exposed contacts, each 1.5 mm
long. These are arranged linearly with a space of 1.5 mm between
contacts. Once the electrode was in place its location was adjusted
to find the site of greatest pain relief. The frequency, voltage and
pulse width were altered to test the effect of the different param-
eters on the magnitude and spatial coverage of pain relief. When a
location was found where relief was thought to be maximal, the
electrode was implanted and secured to the skull. If no relief
was obtained, the electrode was removed. Temporary lead exten-
sions were then externalised for further testing under less stressful
conditions post-operatively.
After several days of extraoperative testing, the second stage
was carried out under general anaesthesia. A subcutaneous pulse
generator, Kinetra
, (Medtronic Inc., Minneapolis, MN, USA)
was implanted into a pectoral pocket below the clavicle.
Following surgery the patients were assessed at 3 month inter-
vals. Assessment was equivalent to that used preoperatively.
These measures enabled comparisons of the pre- and postopera-
tive data to be made to assess the clinical outcome of DBS.
RESULTS
The outcome data is presented in Table 2. Satisfactory pain relief
(P50%) was obtained in all patients. PVG stimulation alone was
optimal in two patients, whilst a combination of PVG and thalamic
stimulation produced the greatest degree of analgesia in one pa-
tient. At follow-up (mean 13.3 months, range 8–20 months) the
intensity of pain was reduced by an average of 61.7 € 7.4% (range
55–70%). In all three patients, the most troublesome burning com-
ponent of the PLP was completely alleviated. Opiate intake was re-
duced in the two patients requiring morphine pre-operatively.
Quality of life measures revealed a statistically significant
improvement following surgery using a paired sample t test (mean
pre-operative value 43, mean post-operative value 68; P = 0.02).
DISCUSSION
In this small group of patients with phantom limb pain, we have
demonstrated that chronic electrical stimulation of carefully se-
lected deep brain structures yields a significant improvement in
pain intensity and quality, subjective well-being, and frequently
reduces opiate requirements.
Historical aspects of phantom limb pain
The 16th century French military surgeon Ambroise Par
was the
first to medically document the phenomena of phantom limb sen-
sation and phantom limb pain in 1551:
Table 2 Outcome following deep brain stimulation for phantom limb pain in three patients
Patient Target % Burning sensation Follow-up
Improvement (months)
1 PVG 70 Relieved completely 20
2 PVG 55 Relieved completely 12
3 PVG/VPL 60 Relieved completely 8 Morphine 120 mg/day
Journal of Clinical Neuroscience (2005) 12(4), 399–404
“For the patients, long after the amputation is made, say they
still feel pain in the amputated part. Of this they complain
strongly, a thing worthy and almost incredible to people who
have not experienced this”
In 1830, the neurologist Charles Bell published a description of
the condition in The Nervous System of the Human Body.9 The
term ‘phantom’, however, is often credited to the American mili-
tary surgeon Silas Weir Mitchell, who in 1871 gave the first mod-
ern description of a post-surgical ‘ghost’ occurring in an
amputee:10
“There is something almost tragical, something ghastly, in the
notion of these thousands of spirit limbs haunting as many
good soldiers, and every now and then tormenting them...
when... the keen sense of the limb’s presence betrays the
man into some effort, the failure of which of a sudden reminds
him of his loss”
Epidemiology of phantom limb sensation and pain
Phantom limb sensation and pain are inextricably linked. How-
ever, in order to examine the literature that documents the preva-
lence of these phenomena, the two must be distinguished.
Phantom limb sensation may be described as the feeling that the
limb persists when it is not present, either as a result of amputation
or congenital limb deficiency, or when normal sensation from the
limb is absent. Excision of a limb (or body part) by trauma or sur-
gery11 is the typical setting. It has been reported that phantom sen-
sation is experienced by almost all amputees.12 Deafferentation in
the absence of amputation, e.g., spinal cord injury13 may yield
phantom sensations. Phantom limb sensation has also been in-
duced experimentally using an anaesthetic block on an intact
limb.14 Curiously, children with congenital limb deficiencies
may also experience a phantom.15
Phantom limb pain occurs in the majority of individuals follow-
ing amputation, with several large series describing a prevalence
of 55–85% in this patient group (4–7). Almost all amputees
who experience innocuous phantom limb sensations also suffer
from phantom limb pain, whilst phantom limb pain is rare in
the patient group without such phantom limb sensations.16 Exam-
ining risk factors for phantom limb pain using a multivariate anal-
ysis, Dijkstra and colleagues revealed the strongest risk factors to
be lower limb amputation and bilateral amputation. They demon-
strated a prevalence of phantom limb pain of 80% in lower limb
amputees, compared with around 40% in those who lost an upper
limb.17 An investigation of lower limb amputation observed that
the higher the level of the amputation, the greater the incidence
of moderate to severe phantom limb pain.18,19
Clinical features
Phantom limb pain may be constant, however is usually intermit-
tent (90% of cases). The majority only experience pain weekly, or
even less frequently. Daily pain is encountered by around 10% of
patients.16 It is important to distinguish between stump pain and
phantom limb pain, with the latter having a predilection for the
Meds (pre-op) Meds (post-op) EUROQOL EUROQOL
VAS (pre-op) VAS (post-op)
Gabapentin 1200 mg/day Unchanged 60 80
Morphine 10 mg/day Paracetamol 40 60
Morphine 80 mg/day 30 60
ª 2005 Published by Elsevier Ltd.

distal portion of the absent extremity.20,21 The character of phan-
tom limb pain is variable, and may include burning, stabbing, or
throbbing sensations, among others.3
Of importance to the patient and treating physician is the fact
that phantom limb pain can be extremely debilitating. Sherman
et al. found that over 50% of the study population experienced
pain sufficiently severe to hinder their lifestyle on more than 6
days each month. Approximately one in four patients experienced
pain for more than 15 h per day.22
In over three-quarters of cases, phantom limb pain develops in
the first few days following amputation.23,24 Its prevalence ap-
pears to decline over 1–2 years to around 60%.20,24 It must be
noted that both of these studies used an amputee population that
comprised primarily of elderly patients who had lost their limbs
as a result of vascular insufficiency. Studies examining younger
amputees with traumatic amputation following injury are some-
what conflicting. In one study, 44% of subjects reported that their
phantom pain had not diminished over a 30 year period,22 whilst
other authors found a progressive decrease in pain intensity or
severity,6 as well as frequency and duration of painful episodes.21
Warten et al. studied the long term course of phantom limb pain in
over 500 war veterans, revealing that the pain remained un-
changed in nearly half, diminished substantially in over one-third,
and resolved in around 15%. Less than 5% of patients in their ser-
ies experienced worsening of pain.7
It is thought that the incidence of phantom pain is lower in the
paediatric population than in adults.25 Melzack et al.26 studied pa-
tients with either congenital limb deficiency or amputation before
the age of 6 years. Twenty percent of the group with congenital
limb deficiencies and 50% of those who had undergone amputa-
tion experienced phantom limb sensation. Twenty percent of the
group with congenital limb deficiencies and 42% of the amputee
group reported pain. It may be, however, that epidemiological dif-
ferences between adult and paediatric groups arise partly from
underreporting of paediatric pain.25
Mechanisms of phantom limb pain
Theories relating to pathophysiology of phantom limb pain are of-
ten separated into peripheral, spinal and central components,
although all three areas are strongly interconnected.
Peripheral mechanisms
Deafferentation results in degeneration of the distal peripheral
nerve, and neuromas form from regeneration of the surviving
proximal nerve section, which can generate abnormal discharges
that may assist in the maintenance of pain.27 Although physical
stimulation of neuromas can result in C-fibre activity and in-
creased phantom limb pain,28 pain is not usually eliminated by
conduction blockade of the peripheral nerves.29 Furthermore,
pain often persists when ectopic discharges in the peripheral
nerves resolve spontaneously, following surgery, or as a result
of pharmacotherapy.30 It is therefore suggested that peripheral
mechanisms at most modulate the phenomenon of phantom limb
pain.
Spinal mechanisms
Reorganisation of the receptive fields of the spinal cord can occur.
Peripheral nerve injury can lead to degeneration of the C-fibre ter-
minals in lamina II of the dorsal horn. A-fibre axons that terminate
in laminae III and IV may subsequently branch into lamina II, and
the A-fibre input may possibly be interpreted as noxious.31 Central
hyperexcitability occurs, mediated by the NMDA receptor and its
transmitter glutamate.32 In addition, a ‘phenotypic switch’ occurs
ª 2005 Published by Elsevier Ltd.
Phantom limb pain 401
in the expression of neuropeptides such as substance P. Usually
expressed by Ad-and C-fibres, following peripheral nerve injury
they may be expressed by Ab-fibres. Ab-fibres thus generate
spinal cord hyperexcitability, a phenomenon usually associated
with noxious input.32 Decreased inhibitory interneuron activity
may also play a role.29
As phantom limb pain is experienced by paraplegics with com-
plete spinal cord injury,12 mechanisms in addition to those operat-
ing at peripheral and spinal levels must presumably be involved.
Central mechanisms
Structural and functional changes in the primary somatosensory
cortex occur following amputation. These involve the sprouting
of neurons from somatotopically adjacent areas into the deaffe-
rented region.32,33 Functional magnetic resonance imaging (fMRI)
has shown a shift of the mouth representation zone into that of the
hand in the primary somatosensory cortex of individuals with arm
or hand amputations.34 The extent of cortical reorganisation and
the degree of phantom limb pain have been shown to be directly
proportional.35 Amputees may also retain neural activity and func-
tion of the thalamic representation of the amputated limb, as sup-
ported by studies showing that stereotactic microstimulation in the
ventrocaudal thalamus produces painful sensations in the phantom
limb.36
The neuromatrix theory proposes that an extensive network of
neurons connects the thalamus and the cortex and the cortex
and the limbic system.37 This network is a genetically determined
representation of the body image that is modified by sensory input
to create a neurosignature for each region of the body. The neuro-
signature of a particular body part determines how it is con-
sciously perceived. The persistence of the neurosignature,
despite the loss of the body part, may be the cause of phantom
sensations. Phantom pain may result from abnormal reorganisa-
tion in the matrix secondary to a pre-existing pain state or the
amputation process. The neuromagnetic field analysis conducted
by Flor et al.35 supports the theory that reorganisation of the neu-
romatrix correlates with subsequent development of phantom limb
pain.
The phantom limb open-circuit theory suggests that following
amputation the return signal from that limb disappears while the
neuromatrix continues to send an outgoing signal. Alteration of
these signals may be responsible for pain, which could be de-
creased by eliminating the outgoing signal from the neuromatrix
or by creating a normal return signal via electrical stimulation.38
Opposition to Melzack’s theory of the neuromatrix does exist.
The neuromatrix hypothesis does not explain why relief of phan-
tom sensations rarely alleviates phantom pain, why the sensations
can spontaneously cease, or why some amputees do not experi-
ence phantom limb pain. According to Melzack, the broad range
of sensations experienced by the phantom must be attributable
to the diffuse nature of the neuromatrix and hence it would not
be affected by lesions at specific sites. However, focal brain le-
sions of the parietal cortex, thalamus, or corticothalamocortical fi-
bres contralateral to the amputated limb can relieve phantom limb
pain.39 Canavero proposes a dynamic reverberation hypothesis,
whereby pain results from an unbalanced oscillatory mechanism
between somatosensory thalamus and somatosensory cortical
areas.
Treatment of phantom limb pain
Most treatments are ineffective and do not take into account the
mechanisms underlying phantom limb pain. Few controlled trials
of large patient samples have been conducted. A maximum benefit
of about 30% has been reported from many treatments, including
Journal of Clinical Neuroscience (2005) 12(4), 399–404
402 Bittar et al.
pharmacological agents (anticonvulsants, barbiturates, antidepres-
sants, neuroleptics and muscle relaxants) and surgical treatments
(sympathectomy, dorsal root entry zone lesions, cordotomy, rhi-
zotomy and neurostimulation). This does not exceed the placebo
effect.40
Pharmacological therapy
Tricyclic antidepressants (such as amitriptyline) and sodium chan-
nel blockers (notably carbamazepine) are presently the pharmaco-
logical agents of choice for neuropathic pain. Some amelioration
of pain may also be seen with opioids,41 ketamine,27 calcitonin,42
and lidocaine.36
Non-invasive non-medical therapy
As the degree of post-amputation cortical reorganisation is
approximately proportional to phantom limb pain, input to the
amputated region may theoretically reverse these changes. A myo-
electric prosthesis yields decreased cortical reorganisation and
phantom limb pain.34 Alternatively, discrimination training of
electric stimuli to the stump for 2 h per day leads to significant
improvement in pain and reversal of cortical reorganisation.43
Transcutaneous electrical stimulation (TENS), vibration therapy,
hyponosis and acupuncture may provide benefit to some patients
with phantom limb pain.3
A behaviour-based alternative is Ramachandran’s mirror box,
in which the normal and the phantom limb are placed side by
side with a mirror between the two. The position of the normal
limb is adjusted so that its reflection appears to be superimposed
on the phantom. The patient then sends motor commands as if to
make symmetrical movements of the limbs, while observing the
reflection of the normal limb. This visual feedback creates the
sensation of the phantom moving synchronously, thereby reliev-
ing ‘paralysis’ of the phantom limb and associated pain. In cer-
tain cases prolonged use of this treatment eventually led to
complete disappearance of the phantom and consequently the
pain.44
Neurosurgical stimulation
It is difficult to assess the effectiveness of neurosurgical stimula-
tion – spinal cord stimulation (SCS), DBS, and motor cortex stim-
ulation (MCS) – for treatment of central neuropathic pain, due to
the absence of long-term controlled trials. Assessment of neuro-
surgical stimulation for individuals with phantom limb pain is
even rarer.
Katayama et al. applied SCS to a group of 19 patients with
phantom limb pain.45 Of those, 6 (32%) had long-term pain con-
trol. Ten patients in which SCS was unsuccessful underwent DBS
of the somatosensory thalamus. Satisfactory pain control was ob-
tained in 60% of these. Five patients who failed to achieve satis-
factory pain reduction with SCS underwent MCS. Satisfactory
pain relief was only achieved in one (20%). Four patients under-
went both DBS and MCS, with one reporting better pain control
from MCS than DBS and two reporting the opposite. This study
shows that SCS and DBS can both be successful in the treatment
of phantom limb pain. However, the results are difficult to inter-
pret due to the fact that only those who failed SCS underwent
DBS and/or MCS. Not withstanding the small numbers in this
study and its methodological limitations, the majority receiving
DBS experienced satisfactory phantom limb pain control, whilst
the same desirable end-point was achieved in a minority of sub-
jects following MCS. Sol et al.46 instituted chronic MCS in 3 pa-
tients with phantom limb pain following upper limb amputation.
All patients had initial pain control and 2 of 3 continued to have
stable pain relief at 2-year follow-up.
Journal of Clinical Neuroscience (2005) 12(4), 399–404
Young and Chambi47 conducted a review of the world literature
documenting the effectiveness of electrical stimulation of the
brain for treatment of chronic pain of varying aetiologies. Of
the 964 patients, 59% had satisfactory pain relief. These authors
also concluded that electrical stimulation was more effective for
nociceptive pain than for neuropathic pain (70% vs 50% of pa-
tients relieved).
While it appears that neurosurgical stimulation is a valuable
treatment for phantom limb pain, it is currently impossible to
accurately predict which patients will respond to which type of
stimulation and how effective that response will be. Further
long-term, controlled, double-blind studies of larger samples of
patients must be carried out to resolve these issues.
Mechanisms of pain relief by deep brain stimulation
Thalamic stimulation
The ventral posterior nucleus of the thalamus regulates sensory in-
put to the cortex. Afferent projections are derived from the
ascending somatosensory pathways, and efferent fibres project
to the primary somatosensory and other regions of the cerebral
cortex. The ventroposterolateral (VPL) nucleus receives input
from the trunk and limbs, whereas the ventroposteromedial
(VPM) nucleus receives input from the head, neck and face.
Duncan et al.48 measured changes in cerebral blood flow (CBF)
using positron emission tomography (PET) scanning in patients in
whom successful long-term relief of chronic pain had been
achieved with somatosensory thalamic stimulation. Thalamic
stimulation of <100 Hz increased CBF in and near the thalamus.
The greatest cortical increase in CBF was found in the ipsilateral
rostral insula. This region is also activated by non-painful thermal
stimuli applied to the skin.49 Cold is thought to inhibit pain pro-
cessing,50 and it has been proposed that the temperature projection
involving the VPM thalamus and anterior insula is an important
pain-inhibitory pathway. Many patients feel thermal sensations
during thalamic stimulation. It was hypothesised that in some pa-
tients, stimulation of the somatosensory thalamus activated thala-
mocortical circuits involved in thermal as well as tactile
processing.48
Periaqueductal grey/periventricular grey stimulation
Stimulation of the PVG has long been believed to activate the
descending anti-nociceptive opioid system. It is now thought that
ascending pathways from the PVG may also be involved in the
mechanism of pain relief. Increased activation of the medial dor-
sal nucleus of the thalamus, which is associated with the limbic
system and has extensive connections with the amygdala and cin-
gulate cortex, has been observed during stimulation of the PVG.51
Therefore, in addition to activating the descending opioid system,
stimulation of the PVG may also modify the patient’s emotional
response to pain.
Much of the work on periaqueductal grey (PAG) and PVG stim-
ulation has been carried out on the premise that stimulation of
these areas activates the descending opioid anti-nociceptive sys-
tem. This theory was corroborated by the finding that pain relief
from PAG–PVG stimulation was reversed by naloxone, an opioid
antagonist.52 Stimulating in the PAG/PVG has shown to cause ele-
vated levels of endogenous opioids and related compounds in the
CSF.53,54 Hosobuchi later suggested that these previous observa-
tions of elevations in endogenous opiate levels in the CSF could
have been an artefact resulting from cross-reaction of the iodin-
ated contrast agent used in ventriculography,55 but further studies
have confirmed early experiments. Following 20 min DBS of the
PVG in chronic pain patients, immunoreactivity for both b-endor-
ª 2005 Published by Elsevier Ltd.

phin and met-enkephalin in ventricular CSF significantly in-
creased over baseline levels. Stimulation from a second electrode
in the VPL did not cause similar changes.56,57 This strongly sup-
ports the theory that PAG/PVG stimulation activates a descending
pain-inhibitory pathway that mediates opioid analgesia.
Young and Chambi47 suggest non-opioid mechanisms play an
important role in pain relief by PAG and PVG stimulation. The
development of tolerance to chronic electrical stimulation of these
regions was taken by various authors as circumstantial evidence of
an opioid system at work. Although over one-third of patients
undergoing PAG/PVG stimulation developed tolerance, a similar
proportion of patients undergoing thalamic stimulation also devel-
oped tolerance.47 As this latter is assumed to be a non-opioid sys-
tem, tolerance per se does not seem to be dependent on opioids.
Administration of naloxone or a placebo to a group of patients
during PAG/PVG stimulation in a controlled double-blind trial
showed pain intensity, as assessed by a visual analogue scale, to
be increased to the same degree by both placebo and naloxone;
further evidence to support non-opioid mechanisms as the primary
cause of pain relief. Studies have found that PVG stimulation pro-
duces changes in electrical activity recorded from sensory thala-
mus.58 This immediate effect suggests the involvement of a
direct neural pathway rather than a descending opioid mechanism.
Ascending projections from PAG to midline, intralaminar and
medial dorsal thalamic nuclei have been observed in monkey.59
The medial dorsal nucleus of the thalamus has extensive ascend-
ing limbic connections, including the amygdala, and is therefore
involved with mood and the conscious emotional response to
events. Pain relief by PVG stimulation may in part be mediated
by ascending pathways from the PVG to the medial dorsal thala-
mus, cingulate cortex, and amygdala which may decrease the
emotional response to pain, rather than the pain itself. This is sup-
ported by the observations in humans that stimulation-induced
analgesia is frequently most effective for extremely noxious pain
accompanied by strong emotional overtones,52 and PVG stimula-
tion results in activation of the medial thalamus and cingulate
cortex.51
CONCLUSION
In a small group of patients with phantom limb pain, we have
demonstrated that deep brain stimulation yields a significant
improvement in pain severity and quality. Analgesic requirements
are frequently reduced, and a statistically significant quality of life
benefit is derived.
REFERENCES
1. Whitaker HA. A historical note on phantom limb. Neurology 1979; 29: 273.
2. Keil G. Sogenannte erstbeschreibung des phantomschmerzes von Ambroise
Par
. Fortschr Med 1990; 108: 62–66.
3. Nikolajsen L, Jensen TS. Phantom limb pain. Br J Anaesth 2001; 87: 107–116.
4. Sherman RA, Sherman CJ. Prevalence and characteristics of chronic phantom
limb pain among American veterans. Results of a trial survey. Am J Phys Med
1983; 62: 227–238.
5. Pohjolainen T. A clinical evaluation of stumps in lower limb amputees. Prosthet
Orthot Int 1991; 15: 178–184.
6. Houghton AD, Nicholls G, Houghton AL, Saadah E, McColl L. Phantom pain:
natural history and association with rehabilitation. Ann R Coll Surg Engl 1994;
76: 22–25.
7. Wartan SW, Hamann W, Wedley JR, McColl I. Phantom pain and sensation
among British veteran amputees. Br J Anaesth 1997; 78: 652–659.
8. Melzack R. The McGill Pain Questionnaire: major properties and scoring
methods. Pain 1975; 1: 277–299.
9. Furukawa T. Charles Bell’s description of the phantom phenomenon in 1830.
Neurology 1990; 40: 1830.
ª 2005 Published by Elsevier Ltd.
Phantom limb pain 403
10. Mitchell SW. Phantom limbs. Lippincott’s Magazine of Popular Literature and
Science 1871; 8: 563–569.
11. Simmel ML. On phantom pain. Arch Neurol Psychiat 1956; 75: 637–647.
12. Melzack R. Phantom limbs. Sci Am 1992; 266: 120–126.
13. Paeslack V, Spahn B, Sommer K. Pain symptoms in paraplegic patients.
Fortschr Med 1990; 108: 57–58.
14. Melzack R, Bromage PR. Experimental phantom limbs. Exper Neurol 1973; 39:
261–269.
15. Weinstein SM, Sersen EA. Phantoms in cases of congenital absence of limbs.
Neurology 1961; 10: 905–1111.
16. Kooijman CM, Dijkstra PU, Geertzen JHB, Elzinga A, Schans CP. Phantom
pain and phantom sensations in upper limb amputees: an epidemiological study.
Pain 2000; 87: 33–41.
17. Dijkstra PU, Geertzen JH, Stewart R, van der Schans CP. Phantom pain
and risk factors: a multivariate analysis. J Pain Symptom Manage 2002; 24:
578–585.
18. Roth YF, Sugarbaker PK. Pains and sensations after amputation: character and
clinical significance. Arch Phys Med Rehabil 1980; 61: 490.
19. Postone N. Phantom limb pain: a review. Int J Psychiatry Med 1987; 17: 57–70.
20. Jensen TS, Krebs B, Nielson J, Rasmussen P. Immediate and long-term
phantom limb pain in amputees: clinical characteristics and relationship to pre-
amputation limb pain. Pain 1985; 21: 407–414.
21. Nikolajsen L, Ilkjaer S, Christensen JH, Krøner K, Jensen TS. Randomised trial
of epidural bupivacaine and morphine in prevention of stump and phantom pain
in lower-limb amputation. Lancet 1997; 350: 1353–1357.
22. Sherman RA, Sherman CJ, Parker L. Chronic phantom and stump pain among
American veterans: results of a survey. Pain 1984; 18: 83–95.
23. Jensen TS, Krebs B, Nielsen J, Rasmussen P. Phantom limb, phantom pain and
stump pain in amputees during the first six months following limb amputation.
Pain 1983; 17: 243–256.
24. Parkes CM. Factors determining the persistence of phantom pain in the
amputee. J Psychosom Res 1973; 17: 97–108.
25. Krane EJ, Heller LB. The prevalence of phantom sensation and pain in
paediatric amputees. J Pain Symptom Manage 1995; 10: 21–29.
26. Melzack R, Israel R, Lacroix R, Schultz G. Phantom limbs in people with
congenital limb deficiency or amputation in early childhood. Brain 1997; 120:
1603–1620.
27. Nikolajsen L, Hansen CL, Nielsen J, Keller J, Arendt-Nielsen L, Jensen TS.
The effect of ketamine on phantom pain: a central neuropathic disorder
maintained by peripheral input. Pain 1996; 67: 69–77.
28. Nystrom B, Hagbarth K. Microelectrode recordings from transected nerves in
amputees with phantom limb pain. Neurosci Lett 1981; 27: 211–216.
29. Weinstein SM. Phantom pain. Oncology 1994; 8: 65–70.
30. MacFarlane BV, Wright A, O’Callaghan J, Bensen HA. Chronic neuropathic
pain and its control by drugs. Pharmacol Ther 1997; 75: 1–19.
31. Cook AJ, Woolf CJ, Wall PD, McMahon SB. Dynamic receptive field plasticity
in rat spinal cord dorsal horn following C-primary afferent input. Nature 1987;
325: 151–153.
32. Flor H. Phantom-limb pain: characteristics, causes, and treatment. Lancet
Neurol 2002; 1: 182–189.
33. Pons TP, Garrahty PE, Ommaya AK, Kaas JH, Taub E, Mishkin M. Massive
cortical reorganization after sensory deafferentation in adult macaques. Science
1991; 252: 1857–1860.
34. Lotze M, Flor H, Grodd W, Larbig W, Birbaumer N. Phantom movements
and pain: an fMRI study in upper limb amputees. Brain 2001; 124:
2268–2277.
35. Flor H, Elbert T, Knecht S, et al. Phantom limb pain as a perceptual correlate
of cortical reorganization following arm amputation. Nature 1995; 375:
482–484.
36. Davis KD, Kizz ZHT, Luo L, Tasker RR, Lozano AM, Dostrovsky JO.
Phantom sensations generated by thalamic microstimulation. Nature 1998; 391:
385–387.
37. Melzack R. Phantom limbs and the concept of neuromatrix. Trends Neurosci
1990; 13: 88–92.
38. Kingham DJ. On the general theory of neural circuitry. Med Hypoth 1994; 42:
291–298.
39. Canavero S. Dynamic reverberation. A unified mechanism for central and
phantom pain. Med Hypoth 1994; 42: 203–207.
40. Sherman RA, Sherman CJ, Gall NG. A survey of current phantom limb pain
treatment in the United States. Pain 1980; 8: 85–99.
41. Huse E, Larbig W, Flor H, Birbaumer N. The effect of opioids on phantom limb
pain and cortical reorganization. Pain 2001; 90: 47–55.
42. Jaeger H, Maier C. Calcitonin in phantom limb pain: a double-blind study. Pain
1992; 48: 21–27.
43. Flor H, Denke C, Schaefer M, Grsser M. Sensory discrimination training alters
both cortical reorganisation and phantom limb pain. Lancet 2001; 357:
1763–1764.
44. Ramachandran VS, Blakeslee S. Phantoms in the Brain. London: Fourth
Estate; 1999, pp. 46–49.
Journal of Clinical Neuroscience (2005) 12(4), 399–404
404 Bittar et al.
45. Katayama Y, Yamamoto T, Kobayashi K, Kasai M, Oshima H, Fukaya C.
Motor cortex stimulation for phantom limb pain: comprehensive therapy with
spinal cord and thalamic stimulation. Stereotact Funct Neurosurg 2001; 77:
159–162.
46. Sol JCH, Casaux J, Roux FE, et al. Chronic motor cortex stimulation for
phantom limb pain: correlations between pain relief and functional imaging
studies. Stereotact Funct Neurosurg 2001; 77: 172–176.
47. Young RF, Chambi VL. Pain relief by electrical stimulation of the
periaqueductal and periventricular gray matter. Evidence for a non-opioid
mechanism. J Neurosurg 1987; 66: 364–371.
48. Duncan GH, Kupers RC, Marchand S, Villemure JG, Gybels JM, Bushnell MC.
Stimulation of human thalamus for pain relief: possible modulatory circuits
revealed by positron emission tomography. J Neurophys 1998; 80: 3326–3330.
49. Craig AD, Bushnell MC, Zhang ET, Blomqvist A. A thalamic nucleus specific
for pain and temperature sensation. Nature 1994; 372: 770–773.
50. Bini G, Cruccu G, Hagbarth KE, Schady W, Torebjork E. Analgesic effect of
vibration and cooling on pain induced by intraneural electrical stimulation. Pain
1984; 18: 239–248.
51. Rezai AR, Lozano AM, Crawley AP, et al. Thalamic stimulation and
functional magnetic resonance imaging: localization of cortical and subcortical
activation with implanted electrodes. Technical note. J Neurosurg 1999; 90:
583–590.
Journal of Clinical Neuroscience (2005) 12(4), 399–404
52. Hosobuchi Y, Adams JE, Linchitz R. Pain relief by electrical stimulation of the
central gray matter in humans and its reversal by naloxone. Science 1977; 197:
183–186.
53. Hosobuchi Y, Adams JE, Bloom FE, Guilleum R. Stimulation of human
periaqueductal grey for pain relief increases immunoreactive b-endorphin in
ventricular fluid. Science 1979; 203: 279–281.
54. Akil H, Richardson DE, Hughes J, Barchas JD. Enkephalin-like material
elevated in ventricular cerebrospinal fluid of pain patients after analgesic focal
stimulation. Science 1978; 201: 463–465.
55. Hosobuchi Y. Subcortical electrical stimulation for control of intractable pain
in humans. Report of 122 cases (1970–1984). J Neurosurg 1986; 64: 543–553.
56. Young RF, Bach FW, Van Norman AS, Yaksh TL. Release of beta-endorphin
and methionine-enkephalin into cerebrospinal fluid during deep brain
stimulation for chronic pain. Effects of stimulation locus and site of sampling.
J Neurosurg 1993; 79: 816–825.
57. Young RF, Rinaldi PC. Brain stimulation. In: North RB, Levy RM, editors.
Neurosurgical Management of Pain. New York, NY: Springer; 1997. p.
283–301.
58. Nandi D, Aziz T, Carter H, Stein J. Thalamic field potentials in chronic central
pain treated by periventricular gray stimulation. Pain 2003; 101: 97–107.
59. Mantyh PW. Connections of the midbrain periaqueductal gray in the monkey.
J Neurophys 1983; 49: 567–581.
ª 2005 Published by Elsevier Ltd.